1: 

2: \documentclass[12pt]{article}

3: \usepackage{graphicx}

4: \begin{document}

5: \centerline{\bf Ising model spin $S=1$ on directed Barab\'asi-Albert networks}

6:  

7: \bigskip

8: \centerline{F.W.S. Lima}

9:  

10: \bigskip

11: \noindent

12: Departamento de F\'{\i}sica, 

13: Universidade Federal do Piau\'{\i}, 57072-970 Teresina - PI, Brazil

14:  

15: \medskip

16:   e-mail: wel@ufpi.br

17: \bigskip

18:  

19: {\small Abstract: On directed Barab\'asi-Albert networks with two and

20:  seven  neighbours selected by each added site, the Ising model with

21:  spin $S=1/2$ was seen

22:  not to show a spontaneous magnetisation. Instead, the decay time for 

23:  flipping of the magnetisation followed an Arrhenius law for Metropolis

24:  and Glauber algorithms, but for Wolff cluster flipping the

25:  magnetisation  decayed exponentially with time. On these networks the

26:  Ising model spin $S=1$ is now studied through Monte Carlo simulations.

27:  However, in this model,  the order-disorder phase transition 

28:  is well defined in this system. We have obtained a first-order phase

29:  transition for values of connectivity $m=2$ and $m=7$ of the 

30:  directed Barab\'asi-Albert network.}

31:  

32:  Keywords:Monte Carlo simulation, Ising , networks, desorden.

33:  

34: \bigskip

35: 

36:  {\bf Introduction}

37:  

38:  Sumour and Shabat \cite{sumour,sumourss} investigated Ising models with

39:  spin $S=1/2$ on directed Barab\'asi-Albert networks \cite{ba} with

40:  the usual Glauber dynamics.  No spontaneous magnetisation was 

41:  found (and we now confirmed this effect), in contrast to the case of 

42:  undirected  Barab\'asi-Albert networks

43:  \cite{alex,indekeu,bianconi} where a spontaneous magnetisation was

44:  found lower a critical temperature which increases logarithmically with

45:  system size. More recently, Lima and Stauffer \cite{lima} simulated

46:  directed square, cubic and hypercubic lattices in two to five dimensions

47:  with heat bath dynamics in order to separate the network effects  form

48:  the effects of directedness. They also compared different spin flip

49:  algorithms, including cluster flips \cite{wang}, for

50:  Ising-Barab\'asi-Albert networks. They found a freezing-in of the 

51:  magnetisation similar to  \cite{sumour,sumourss}, following an Arrhenius

52:  law at least in low dimensions. This lack of a spontaneous magnetisation

53:  (in the usual sense)

54:  is consistent with the fact

55:  that if on a directed lattice a spin $S_j$ influences spin $S_i$, then

56:  spin $S_i$ in turn does not influence $S_j$, 

57: % following remark added

58:  and there may be no well-defined total energy. Thus, they show that for

59:  the same  scale-free networks, different algorithms give different

60:  results. Now we study the Ising model for spin $S=1$ on directed 

61:  Barab\'asi-Albert network and different from the Ising model for

62:  spin $S=1/2$, the order-disorder phase transition of 

63:  order parameter is well defined in this system. We have 

64:  obtained a first-order phase transition for values of 

65:  connectivity $m=2$ and $m=7$ of the directed Barab\'asi-Albert network.

66:  

67: \bigskip

68:  

69: \begin{figure}[hbt]

70: \begin{center}

71: \includegraphics[angle=-90,scale=0.60]{lima1.eps}

72: \end{center}

73: \caption{Plot of spontaneous magnetization versus temperature for various network

74: sizes.

75: } 

76: \end{figure}

77:  

78: \begin{figure}[hbt]

79: \begin{center}

80: \includegraphics[angle=-90,scale=0.30]{lima2a.eps}

81: \includegraphics[angle=-90,scale=0.30]{lima2b.eps}

82: \end{center}

83: \caption{

84: Plot of the Binder parameter $B_{i}$ versus $K$ for severals systems sizes ($N=250$, $500$, $1000$, $2000$, $4000$, $8000$ and $16000$). In the  Part (a) $m=2$ and Part (b) $m=7$.}

85: \end{figure}

86:   

87: \bigskip

88:  

89: \begin{figure}[hbt]

90: \begin{center}

91: \includegraphics[angle=-90,scale=0.60]{lima3.eps}

92: \end{center}

93: \caption{

94: Plot of the Binder parameter $B_{i}$ versus $1/N$ for $m=2$ (circle) and $m=7$ (square), and severals systems sizes ($N=1000$, $2000$, $4000$, $8000$ and $16000$).}

95: \end{figure}

96:  

97: \begin{figure}[hbt]

98: \begin{center}

99: \includegraphics[angle=-90,scale=0.60]{lima4.eps}

100: \end{center}

101: \caption{Plot of the Binder parameter $2/3-B_{i}(K)$ versus $1/N$ for $m=2$ (circle) and $m=7$ (square), and severals systems sizes ($N=1000$, $2000$, $4000$, $8000$ and $16000$).

102: } 

103: \end{figure}

104:  

105: \begin{figure}[hbt]

106: \begin{center}

107: \includegraphics[angle=-90,scale=0.30]{lima5a.eps}

108: \includegraphics[angle=-90,scale=0.30]{lima5b.eps}

109: \end{center}

110: \caption{Plot of the specific heat $C_{max}$ (circle) and susceptibility $\chi_{max}$ (square) versus $N$. In the  Part (a) $m=2$ and Part (b) $m=7$. }

111: \end{figure}

112: 

113: \bigskip

114: 

115: {\bf Model and Simulation}

116: 

117: We consider the Ising model with spin $S=1$, on directed 

118: Barab\'asi-Albert Networks, defined by a set of

119: spins variables ${S}$ taking the values $-1$, $0$ and

120: $+1$, situated on every site of a directed 

121: Barab\'asi-Albert Networks with $N$ sites. 

122: 

123: The Ising interation energy is given by

124: \begin{equation}

125: %\begin{center}

126: E=-J\sum_{i}\sum_{k}S_{i}S_{k}

127: %\end{center}

128: \end{equation} 

129: where $k$-sum runs over all nearest neighbors of $S_{i}$. In this network, each new site

130: added to the network selects $m$

131: already existing sites as neighbours influencing it; the newly

132: added spin does not influence these neighbours. 

133: To study the critical behavior of the model we define the variable $e=E/N$

134: and $m=\sum_{i=1}^{N}S_{i}/N$ .

135: >From variable of the energy measurements we can compute, the average energy and specific heat and energetic fourth-order parameter,

136: %

137: \begin{equation}

138:  u(K)=[<E>]_{av}/N,

139: \end{equation}

140: %

141: \begin{equation}

142:  C(K)=K^{2}N[<e^{2}>-<e>^{2}]_{av},

143: \end{equation}

144: %

145: \begin{equation}

146:  B_{i}(K)=[1-\frac{<e^{4}>}{3<e^{2}>^{2}}]_{av},

147: \end{equation}

148: where $K=J/k_BT$, with $J=1$, and $k_B$ is the Boltzmann constant. 

149: Similarly, we can derive from the magnetization measurements

150: the average magnetization, the susceptibility, and the magnetic

151: cumulants,

152: %

153: \begin{equation}

154:  m(K)=[<|m|>]_{av},

155: \end{equation}

156: %

157: \begin{equation}

158:  \chi(K)=KN[<m^{2}>-<|m|>^{2}]_{av},

159: \end{equation}

160: %

161: \begin{equation}

162:  U_{4}(K)=[1-\frac{<m^{4}>}{3<|m|>^{2}}]_{av}.

163: \end{equation}

164: where $<...>$ stands for a thermodynamics average and $[...]_{av}$ square brackets

165: for a averages over the 20 realizations. 

166: 

167: In the order to verify the order of the transition this model, we apply finite-size scaling (FSS) \cite{fss}. Initially we search for the minima of energetic fourth-order cumulant in eq. (4). This quantity gives a qualitative as well as a quantitative description of the order the transition \cite{mdk}. It is known \cite{janke} that this

168: parameter takes a minimun value $B_{i,min}$ at the effective transition temperature 

169: $T_{c}(N)$. One can show \cite{kb} that for a second-order transition $\lim_{N\to \infty}$

170: $(2/3-B_{i,min})=0$, even at $T_{c}$, while at a first-order transition the same limit measures the latent heat $|e_{+}-e_{-}|$:

171: 

172: \begin{equation}

173: \lim_{N\to \infty}(2/3-B_{i,min})=\frac{1}{3}\frac{(e_{+}-e_{-})^{2}(e_{+}+e_{-})^{2}}

174: {(e_{+}^{2}-e_{-}^{2})^{2}}.

175: \end{equation}

176: 

177: A more quantitative analysis can be carried out throught the FSS of the specific heat 

178: $C_{max}$, the susceptibility maxima $\chi_{max}$ and the minima of the Binder parameter $B_{i,min}$. If the hypothesis of a first-order phase transition is correct, we should then expect, for large systems sizes, an asymptotics FSS behavior of the form

179: \cite{wj,pbc},

180: \begin{equation}

181: C_{max}=a_{C} + b_{C}N +...

182: \end{equation}

183: %

184: \begin{equation}

185: \chi_{max}=a_{\chi} + b_{\chi}N +...

186: \end{equation}

187: \begin{equation}

188: B_{i,min}=a_{B_{i}} + b{B_{i}}N +...

189: \end{equation}

190: 

191: We have performed Monte Carlo simulation on directed Barab\'asi-Albert networks with

192: values of connectivity $m=2$ and $7$. For a given $m$, we used systems

193: of size $N=250$, $500$, $1000$, $2000$, $4000$, $8000$, and $16000$ sites. We waited $10000$ Monte Carlo

194: steps (MCS) to make the system reach the steady state, and the time averages were

195: estimated from the next $ 10000$ MCS. In our simulations, one MCS is accomplished

196: after all the $N$ spins are updated. For all sets of parameters, we have generated

197: $20$ distinct networks, and have simulated $20$

198: independent runs for each distinct network.

199: 

200: \bigskip

201: 

202: {\bf Results and Discussion}

203: 

204: Our simulations, using the HeatBath algorithm, indicate that the model displays a first order phase transition. Fig.1 shows the overall behaviour of the 

205: spontaneous magnetisation.

206: In Fig. 2 we show the dependence of the Binder parameter $B_{i}(K)$ for

207: connectivity $m=2$ and $7$

208: on the inverse of temperature $K$ and various

209:  systems size.

210: Part (a) shows the  curves for $m=2$ of bottom to top of $N=250$ to $16000$, part (b)  the same as part (a) for $m=7$. The Binder parameter clearly goes to a value which is different from $2/3$. This is a sufficient condition to characterize a first-order transition.

211: In Fig. 3 we plot the Binder parameter $B_{i}$ versus $1/N$ for $m=2$ (circle) and $m=7$ (square), and severals systems sizes ($N=1000$, $2000$, $4000$, $8000$ and $16000$). We show the scaling of the Binder parameter minima, and again the first order phase transition is verified. 

212: The order of the transitions can be confirmed by ploting the values of $2/3-B_{i,min}$

213: again versus $1/N$. For a second-order transition the curves goes to zero as we increase the system size. Here, the quantity $2/3-B_{i,min}$ approaches a nonvanishing value in the

214: limit of small $1/N$ as for $m=2$ than as $m=7$, see Fig. 4.

215: 

216: As decipted in Figure 5, our results for scaling of the specific heat and susceptibility are consistent with equations (9,10). Part a shows $m=2$, and part b, $m=7$. The same occurs with the plot the Fig. 4 for the Binder  parameter minima, equation (11). In the part a for $m=2$ and $m=7$ part b.

217: 

218: \bigskip

219:  

220: {\bf Conclusion}

221:  

222: In conclusion, we have presented a very simple equilibrium model on

223: directed Barab\'asi-Albert network \cite{sumour,sumourss}. Different from

224: the spin 1/2 Ising model, in these networks, the spin 1 Ising model presents a

225: the first-order phase transition which occurs in model with 

226: connectivity $m=2$ and $m=7$ here studied. We also verific that occur 

227: a phase transition for  Potts Model for $q=3$ and $q=8$  on directed Barab\'asi-Albert network \cite{lima2}.

228: 

229: 

230: 

231:   The F.W.S. Lima  is a pleasure to thanks D. Stauffer for many suggestions and fruitful

232: discussions during the development this work and also for the revision of this paper.

233: I also acknowledge the Brazilian agency FAPEPI

234: (Teresina-Piau\'{\i}-Brasil) for  its financial support

235: and also the Fernando Whitaker for help in the support the system SGI Altix 1350 the computational park CENAPAD.UNICAMP-USP, SP-BRASIL.

236:  

237: \begin{thebibliography}{99}

238: 

239: \bibitem{sumour} M.A. Sumour and M.M. Shabat, Int. J. Mod. Phys. C 16,

240:  585 (2005) and cond-mat/0411055 at www.arXiv.org. 

241:  

242: \bibitem{sumourss} M.A. Sumour, M.M. Shabat and D. Stauffer, Islamic University 

243: Journal (Gaza) 14, 209 (2006) (cond-mat/0504460 at www.arXiv.org).

244: 

245: \bibitem{ba} R. Albert and A.L. Barab\'asi, Rev. Mod. Phys. 74, 47

246:  (2002).

247:  

248: \bibitem{alex} A. Aleksiejuk, J.A. Ho\l yst and D. Stauffer, Physica A

249:  310, 269 (2002).

250:  

251: \bibitem{indekeu} J.O. Indekeu, Physica A 333, 451 (2004); 

252: C.V. Giuraniuc, J.P.L. Hatchett, J.O. Indekeu, M. Leone, I. P\'{e}rez Castillo, 

253: B. Van Schaeybroeck and C. Vanderzande, Phys. Rev. Lett. 95, 098701 (2005).

254:   

255: \bibitem{bianconi} G. Bianconi, Phys. Lett. A 303, 166 (2002).

256:  

257: \bibitem{lima} F.W.S. Lima and D. Stauffer, Physica A 359, 423 (2006).

258:  

259: \bibitem{wang} J.S. Wang and R. H. Swendsen, Physica A 167, 565 (1990).

260: 

261: \bibitem{fss} See Finite Size Scaling and Numerical Simulation of Statistical Systems, edited by V. Privman (World Scientific, Singapore, 1990).

262: 

263: \bibitem{mdk} M.S.S. Challa, D. P. Landau, K. Binder, Phys. Rev. B, 34, 1841 (1986).

264: 

265: \bibitem{janke} W. Janke, Phys. Rev. B 47, 14757 (1993).

266: 

267: \bibitem{kb} K. Binder, D. W. Heermann, in Monte-Carlo Simulation in Statistical

268: Phys., edited by P. Fulde (Springer-Verlag, Berlin, 1988), p. 61-62.

269: 

270: \bibitem{wj} W. Janke, R. Villanova, Phys. Lett. A 209, 179 (1995).

271: 

272: \bibitem{pbc} P. E. Berche, C. Chatelain, B. Berche, Phys. Rev. Lett. 80, 297 (1998).

273: 

274: \bibitem{lima2} In preparation.

275: 

276: \end{thebibliography}

277: \end{document}

278: 

279: 

280: