1: \documentstyle[12pt]{article}
2: \textwidth=17cm
3: \textheight=23cm
4: \topmargin=-0.5cm
5: \oddsidemargin=-0.5cm
6: \begin{document}
7: \begin{center} {\Large {\bf
8: Nonequilibrium Multicritical Behaviour in Anisotropic
9: Heisenberg Ferromagnet driven by Oscillating Magnetic Field }} \end{center}
10:
11: \vskip 1 cm
12:
13: \begin{center}{Muktish Acharyya}\end{center}
14: \begin{center}{\it Department of Physics, Krishnanagar Government College,
15: PO-Krishnanagar, Dist-Nadia, PIN-741101, WB, India}\end{center}
16: \begin{center}{\it E-mail: muktish@vsnl.net}\end{center}
17: \vskip 2cm
18:
19: The Heisenberg ferromagnet (uniaxially anisotropic along z-direction), in the presence of time dependent (but uniform over space)
20: magnetic field, is studied by Monte Carlo simulation. The time dependent magnetic field was taken as elliptically
21: polarised in such a way that the resulting field vector rotates in the XZ-plane. In the limit of low anisotropy, the
22: dynamical responses of the system are studied as functions of temperature and the amplitudes of the magnetic field.
23: As the temperature decreases, it was found that the system undergoes multiple
24: dynamical phase transitions. In this limit, the multiple transitions
25: were studied in details and the phase
26: diagram for this observed multicritical behavior was drawn in field amplitude and temperature plane.
27: The natures (continuous/discontinuous)
28: of the transitions are determined by the temperature variations of fourth-order Binder cumulant ratio and the distributions
29: of order parameters near the transition points.
30: The transitions are supported by finite size study. The temperature variations of the variances of dynamic order
31: parameter components (for different system sizes) indicate the existence of diverging length scale near the dynamic
32: transition points. The frequency dependences of the transition temperatures
33: of the multiple dynamic transition are also studied briefly.
34:
35: \vskip 0.5cm
36:
37: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
38: %\twocolumn
39: \begin{center} {\bf INTRODUCTION} \end{center}
40:
41: The nonequilibrium dynamical phase transition \cite{rmp}, particularly in the kinetic Ising model, has drawn immense
42: interest of the researchers working in the field of nonequilibrium statistical physics. The dynamic transition in the
43: kinetic Ising model in the presence of sinusoidally oscillating magnetic field was noticed \cite{tome} first in the
44: mean field solution of the dynamical equation for the average magnetisation. Due to the absence of fluctuations, in
45: the mean field study, the dynamic transition would be possible also even in the static (zero frequency) limit. But the
46: true dynamic transition should never occur in the static limit due to the presence of nontrivial
47: thermodynamic fluctuations. The occurence of the
48: true dynamic transitions for models, incorporating the thermodynamic fluctuations, was later shown in several Monte
49: Carlo studies \cite{rmp}.
50:
51: Since, the Ising model is quite simple and a prototype to study the nonequilibrium
52: dynamical phase transitions, a considerable amount
53: of work was done on it \cite{rmp}. Among these, several studies were done to establish that all these transitions
54: have some features which are similar to
55: well known thermodynamic phase transitions. The divergence of "time scale" \cite{ma1} and the "dynamic specific heat"
56: \cite{ma1} and the divergence of "length scale" \cite{rikpre} are two very important observations to establish that the
57: dynamic transition, in kinetic Ising model driven by sinusoidally oscillating magnetic field,
58: is indeed a thermodynamic phase transition.
59:
60: Although, the dynamic phase transition in kinetic Ising model is an interesting phenomenon which acts as a simple example to grasp the
61: various features of nonequilibrium phase transitions, it has several limitations. Since the orientations of the spins
62: are limited by only two (up/down) directions, some interesting features of dynamic transitions (related to the dynamical
63: transverse ordering) cannot be observed in Ising model. The classical vector spin model \cite{matis} would be
64: the proper choice to study such interesting phenomena which are missing in Ising model. The "off-axial" dynamic transition
65: \cite{ijmpc}, recently observed in Heisenberg ferromagnet, is an example of such interesting phenomena. In the "off-axial"
66: dynamic transition, the dynamical symmetry along the axis of anisotropy can be broken dynamically by applying an
67: oscillating magnetic field along any perpendicular direction. There are several important studies done on classical
68: vector spin models driven by oscillating magnetic field. The dynamic phase transition in an anisotropic XY ferromagnet
69: driven by oscillating magnetic field was recently studied \cite{yasui} by solving Ginzburg-Landau equation. The dynamic
70: phase transition and the dependence of its behaviour on the bilinear exchange anisotropy of classical Heisenberg ferromagnet
71: (thin film), was studied by \cite{jang} Monte Carlo Simulation. The dynamic transitions alongwith hysteresis scaling
72: in Heisenberg ferromagnet was also studied \cite{huang} both by Monte Carlo simulation and mean field solution.
73:
74: All these
75: studies (discussed in above paragraph) of dynamic transitions in classical vector spin model give single dynamic transitions. There
76: are a few studies in Heisenberg ferromagnets where the multiple dynamic transition was reported \cite{mdt1,mdt2}. A recent
77: study \cite{mdt1} of dynamical phase transition in thin Heisenberg ferromagnetic films with bilinear exchange
78: anisotropy has shown multiple phase transitions for the surface and bulk layers of the film at different temperatures.
79: However, another study \cite{mdt2} shows triple dynamic transitions (at three different temperatures), for bulk dynamic ordering only,
80: in uniaxially anisotropic Heisenberg ferromagnet driven by elliptically polarised magnetic field. In the present paper, the multiple
81: dynamic transition (bulk only) \cite{mdt2}
82: in uniaxially anisotropic Heisenberg ferromagnets driven by elliptically polarised magnetic field, is studied extensively.
83:
84: In this paper, the multiple dynamic transition, in the uniaxially anisotropic classical Heisenberg ferromagnet
85: in presence of elliptically polarised magnetic field, is studied by Monte
86: Carlo simulation. The various dynamical phases were found. The multiple dynamic phase transition was observed by studying the
87: temperature variation of the dynamic specific heat and the derivatives of dynamic order parameter components.
88: The nature (continuous/discontinuous) of the transition was detected by studying the distribution of order parameter and the
89: temperature variation of Binder cumulant ratio near the transition point.
90: A finite size
91: study was done supporting the multiple dynamic phase transitions. A length scale was found to diverge near the dynamic transition
92: points. And finally and most importantly, the phase diagram for this multiple dynamic phase transition was drawn. The frequency
93: dependence of these multiple dynamic transitions was also studied briefly.
94:
95: The paper is organized as follows: in the next section the description of the system, i.e., the Heisenberg ferromagnet in
96: polarised magnetic field, is given, section III describes the Monte Carlo simulation technique employed here to study
97: the dynamical steady state of this type of classical
98: vector spin model, in detail. The numerical results with diagrams are reported in section IV and the paper ends with a
99: summary in section V.
100: \vskip 1 cm
101:
102: \begin{center} {\bf HEISENBERG FERROMAGNET IN POLARIZED MAGNETIC FIELD} \end{center}
103:
104:
105: The classical anisotropic (uniaxial and single-site) Heisenberg model, with nearest neighbour ferromagnetic interaction
106: in the presence of a magnetic field can be described by the following Hamiltonian:
107: \begin{equation}
108: H = -J\Sigma_{<ij>} {\bf S_i} \cdot {\bf S_j} - D \Sigma_i (S_{iz})^2 - {\bf h} \cdot \Sigma_i {\bf S_i}.
109: \label{hamiltonian}
110: \end{equation}
111: \noindent In the above expression, ${\bf S_i}[S_{ix},S_{iy},S_{iz}]$ represents a classical spin vector
112: (situated at the $i$th lattice site) of magnitude
113: unity, i.e., $|{\bf S}|=1$ or $S^2_{ix}+S^2_{iy}+S^2_{iz} = 1$. The classical spin vector ${\bf S_i}$ may take any (unrestricted)
114: angular orientation in the vector spin space. The first term, in the Hamiltonian, represents the nearest -neighbour ($<ij>$)
115: ferromagnetic ($J>0$) interaction. The factor $D> 0$, in the second term, represents the strength of uniaxial ($z$ axis here)
116: anisotropy which is favoring the spin vector to be aligned along the $z$ axis. This is readily seen since the second term
117: minimizes energy for maximization of the value of $S^z_i$. Here, it may be noted that for $D \rightarrow \infty$ the system
118: goes to the Ising limit and for $D=0$ the system is in the isotropic Heisenberg limit. The last term stands for the interaction with
119: the externally applied time dependent magnetic field $[{\bf h}(h_x,h_y,h_z)]$. The magnetic-field components are sinusoidally
120: oscillating in time, i.e., $h_{\alpha} = h_{0\alpha}{\rm cos}(\omega t)$, where $h_{0\alpha}$ is the amplitude and $\omega$ is
121: the angular frequency ($\omega = 2 \pi f; f$ is the frequency) of the $\alpha$th component of the magnetic field. In the present
122: study, the field is taken elliptically polarised. In general, the field vector is represented as
123: \begin{equation}
124: {\bf h} = {\hat x}h_x + {\hat y}h_y + {\hat z}h_z \\
125: = {\hat x}h_{0x}{\rm cos}(\omega t) + {\hat z} h_{0z} {\rm sin} (\omega t).
126: \end{equation}
127:
128: \noindent The time eliminated relation between $h_x (= h_{0x} {\rm cos} (\omega t))$ and $h_z (=h_{0z} {\rm sin} (\omega t)$
129: is
130: \begin{equation}
131: {h_x^2 \over h_{0x}^2} + {h_z^2 \over h_{0z}^2} = 1.
132: \end{equation}
133:
134: \noindent This is the equation of an ellipse which indicates that the magnetic field vector is
135: elliptically polarised (for $h_{0x} \neq h_{0z}$) and lies on the $X-Z$ plane. For $h_{0x}=h_{0z}=h_0$ (say), the field
136: will be circularly polarised and $h_x^2 +h_z^2 = h_0^2$. The magnetic fields and the strength of anisotropy $D$ are
137: measured in units of $J$. The model is defined on a simple cubic lattice of linear size $L$ with periodic boundary
138: conditions applied in all three ($x, y, z$) directions.
139:
140: \vskip 1 cm
141:
142: \begin{center} {\bf MONTE CARLO SIMULATION METHOD} \end{center}
143:
144:
145: Monte Carlo (MC) simulation method was employed to study the above described model. The algorithm is \cite{stauffer} described below.
146: The system is slowly cooled down from a random initial spin configuration \cite{uli}, to obtain the steady state spin configuration
147: at a particular temperature $T$ (measured in units of $J/k_B$, where $k_B$ is the Boltzmann constant).
148: The initial random spin configuration was generated as follows \cite{ijmpc,uli}: two different
149: (uncorrelated) random numbers $r_1$ and $r_2$ (uniformly distributed between -1 and +1), are chosen in such a way that
150: $R^2 = r_1^2 + r_2^2$ becomes less than or equal to unity. The set of values of $r_1$ and $r_2$, for which $R^2 > 1$, are rejected.
151: Now, $S_{ix} = 2ur_1$, $S_{iy} = 2ur_2$ and $S_{iz} = 1- 2R^2$, where $u = \sqrt{1-R^2}$. After preparing initial random configuration
152: of spins (this is the proper choice of the spin configurations,
153: corresponding to very high temperature), one has to find a steady state configuration
154: for any fixed temperature $T$.
155: For any fixed set of values of $h_{0x}$, $h_{0z}$, $\omega$ and $D$ and at any particular temperature $T$, a lattice site $i$
156: has been chosen randomly (random updateing scheme). The value (random direction) of the spin vector at this randomly chosen site
157: is ${\bf S_i}$ (say). The energy of the system is given by the Hamiltonian (Eq. \ref{hamiltonian}).
158: A test spin vector ${\bf S'_i}$ is then chosen (for a trial move)
159: at any random direction (following the same algorithm described above). For this choice of ${\bf S'_i}$, at site $i$ the energy will
160: be
161: $H = -J\Sigma_{<ij>} {\bf S'_i} \cdot {\bf S_j} - D \Sigma_i (S'_{iz})^2 - {\bf h} \cdot \Sigma_i {\bf S'_i}$.
162: The change in energy, associated with this change in the direction of spin vector (from ${\bf S_i}$ to ${\bf S'_i}$ at lattice site
163: $i$), is $\Delta H = H'-H$. The Monte Carlo method \cite{stauffer} will decide whether this trial move is acceptable or not.
164: The probability of this move (chosen here) is given by Metropolis rate \cite{metro}
165: \begin{equation}
166: W({\bf S_i} \rightarrow {\bf S'_i}) = {\rm Min}{\Large[}1,{\rm exp}{\Large (}-{{\Delta H} \over {k_B T}}{\Large )}{\Large ]}.
167: \end{equation}
168: \noindent Now, this probability will be compared with a random number $R_p$ (say) (uniformly distributed between zero and one).
169: If $R_p$ does not exceed $W$, the move (${\bf S_i} \rightarrow {\bf S'_i}$) will be accepted. In this way the spin vector ${\bf S_i}$
170: is updated. $L^3$ numbers of such random updates of spins, defines one Monte Carlo step per site (MCSS) and this is considered as the unit
171: of time in this simulation.
172: The efficiency of the MC technique can be increased by choosing the direction of the spin for the trial move close to the
173: present direction which will increase the probability of acceptance.
174: So, 50 MCSS are required to have one complete cycle of the oscillating magnetic field (of frequency $f = 0.02$) and hence the time
175: period ($\tau$) of the field becomes 50 MCSS. Any time dependent macroscopic quantity (i.e., any component of magnetisation at any
176: instant $t$) is calculated as follows: Starting with an initial random spin configuration (corresponding
177: to the high-temperature disordered phase),
178: the system is allowed to become stable (dynamically) up to $5\times10^4$ MCSS (i.e., 1000 complete cycle of the oscillating field).
179: The average value of various physical quantities are calculated from further $5\times10^4$ MCSS (i.e., averaged over another 1000
180: cycles). This was checked carefully that the number of MCSS mentioned above is sufficient to achieve dynamical
181: steady state value of the
182: measurable quantities, etc. which would clearly show the dynamic transition points within limited accuracy. But to describe the
183: critical behaviour very precisely (i.e., to estimate critical exponents etc) a much longer run is required. The total length of the
184: simulation becomes $10^5$ MCSS. The system is slowly cooled down ($T$ has been reduced by a small interval
185: $\delta T = 0.02$ here) to get the values
186: of the statistical quantities in the lower temperature phase.
187: Here, the last spin configuration corresponding to the previous
188: temperature is employed to act as initial configuration for the new (lower) temperature. The CPU time required for $10^5$ MCSS is
189: approximately 30 min on an Intel Pentium-III processor.
190:
191: One important point may be noted here regarding the dynamics chosen in this simulational study. Since the spin components do not
192: commute with the Heisenberg Hamiltonian, it has intrinsic
193: quantum mechanical dynamics. Considering the intrinsic dynamics, there was a study \cite{krech}
194: of the structure factor and transport properties in XY model. However, the present paper aims to study the nonequilibrium phase transition
195: governed by thermal fluctuations only. Keeping this in mind, one should choose the dynamics that arising solely only due to
196: the interaction with thermal
197: bath. Since the objective is different, in this paper, the dynamics chosen (arise solely due to the interaction with a thermal
198: bath) were Metropolis dynamics. The effect of intrinsic spin dynamics is therefore not considered here.
199:
200: \vskip 1 cm
201:
202: \begin{center} {\bf NUMERICAL RESULTS} \end{center}
203:
204: The Monte Carlo simulations, in the present study, were done on a simple cubic lattice of linear size $L = 20$.
205: For a fixed set of values of amplitudes ($h_{0x}, h_{0z}$) and frequency ($f$) of
206: polarised magnetic field, strength of anisotropy ($D$) and temperature ($T$), the instantaneous
207: magnetisation components (per lattice site) were calculated as follows: $m_x(t)=\Sigma_i S^x_i/L^3 $, $m_y(t)=\Sigma_i S^y_i/L^3 $
208: and $m_z(t)=\Sigma_i S^z_i/L^3 $. The dynamic order parameter is defined as the time averaged magnetisation over a full
209: cycle of the oscillating magnetic field. In this case the dynamic order parameter ${\bf Q}$ is a vector (since the
210: magnetisation is vector). The components of the dynamic order parameter are calculated as $Q_x=(1/\tau)\oint m_x(t) dt$,
211: $Q_y=(1/\tau)\oint m_y(t) dt$ and $Q_z=(1/\tau)\oint m_z(t) dt$.
212: Here, four different dynamical phases are identified. The high temperature disordered phase is denoted as
213: $P_0 : (Q_x = Q_y = Q_z = 0)$. Other three ordered phases are $P_1 : (Q_x = 0, Q_y \neq 0, Q_z =0)$, $P_2 :
214: (Q_x \neq 0, Q_y = 0, Q_z \neq 0)$ and $P_3 : (Q_x = 0, Q_y = 0, Q_z \neq 0)$. The phase diagram, obtained from multiple
215: dynamic transitions, is plotted in $h_{0x} - T$ plane and is shown in fig \ref{phase}. The methods of getting the phase
216: boundaries are discussed in the following paragraphs.
217:
218:
219: The signature of such successive phase transitions are also observed by studying the temperature variations of
220: the magnitude of the
221: dynamic order parameter, energy, specific heat and the temperature derivatives of dynamic order parameter components.
222: For the same set of values $D=0.2$, $h_{0x} = 0.3$ and $h_{0z} =1.0$ the temperature variations of various dynamic
223: quantities are studied and shown in Fig. \ref{pt3}. The temperature variations of the dynamic order parameter components
224: $Q_x$, $Q_y$ and $Q_z$ are shown in Fig. \ref{pt3}(a). As the system is cooled down, from a high-temperature dynamically
225: disordered (${\bf Q} = \vec 0$) phase, it was observed that first the system undergoes a transition from dynamically disordered
226: $P_0$ phase
227: (${\bf Q} = \vec 0$) to a dynamically Y-ordered ($Q_y \neq 0$ only) phase $P_1$ and the transition temperature is $T_{c1}$ (say).
228: It may be noted here that the resultant vector of elliptically polarised magnetic field lies in the $x-z$ plane and the dynamic
229: ordering occurs along the $y$-direction only ($Q_x =0$, $Q_y \neq 0$, $Q_z = 0$). This is
230: clearly an off-axial transition \cite{ijmpc}. In the case
231: of this type of off-axial transition the dynamical symmetry (in any direction; $y$ here) is broken by the application of the
232: magnetic field in the perpendicular (lies in the $x-z$ plane here) direction. As the system cools down further, this phase $P_1$
233: persists over a considerable range of temperature and at a temperature $T_{c2}$ a second transition was observed. In this phase, the
234: system becomes dynamically ordered both in the $X$ and $Z$ directions at the cost of $Y$ ordering. Usually one gets
235: the dynamically ordered phase of second kind $P_2: (Q_x \neq 0,
236: Q_y = 0, Q_z \neq 0)$. Here, the dynamical ordering is planar (lies on $x-z$ plane) and the dynamical ordering occurs in the same
237: plane on which the field vector lies. This transition is not of off-axial type. As the temperature decreases further one ends up with
238: a low-temperature dynamically ordered phase of third kind
239: $P_3 : (Q_x = 0, Q _y =0, Q_z \neq 0)$ via a transition occurs at temperature $T_{c3}$.
240: The system continues to increase the dynamical $Z$ ordering ($Q_z \neq 0$ only)
241: as the temperature decreases further. The low-temperature phase is only dynamically $Z$ ordered. No further transition was
242: observed as one cools the system further. From Fig.\ref{pt3}(a), it is quite clear, if one observes carefully, that the
243: sizes of the errorbars of $Q_x$, $Q_y$ and $Q_z$ are maximal near the transition points, indicating the growth of
244: fluctuations near the transiton points.
245:
246: One gets qualitative ideas of multiple dynamic phase transitions from the above mentioned studies. However, to estimate precisely the
247: transition temperatures $T_{c1}$, $T_{c2}$ and $T_{c3}$ further studies are required. For this reason, the temperature variations
248: of the derivatives (with respect to temperature) of the dynamic order parameter components are studied.
249: The derivatives were calculated numerically by using central difference formula \cite{mth}
250: \begin{equation}
251: {{df(x)} \over dx} = {{f(x+\delta x) - f(x-\delta x)} \over {2\delta x}}.
252: \label{derivative}
253: \end{equation}
254: Fig. \ref{pt3}(b) shows
255: such variations studied as a function of temperature. The derivative ${{dQ_y}/{dT}}$, shows a sharp minimum at
256: $T=T_{c1} \simeq 1.22$. The second transition temperature $T_{c2}$ was estimated from the
257: position of sharp maximum of ${{dQ_y}/{dT}}$ and the corresponding sharp minima of
258: ${{dQ_x}/{dT}}$ and ${{dQ_z}/{dT}}$. This gives $T_{c2} \simeq 0.96$. At slightly lower temperature
259: than $T_{c2}$, one observes
260: another maximum of ${{dQ_x}/{dT}}$ and minimum of ${{dQ_z}/{dT}}$ both
261: at the same position $T=T_{c3} \simeq 0.88$. From this study
262: one gets the quantitative measure of the transition temperatures of multiple dynamic transition \cite{mdt2}.
263: The maximum error, in estimating the transition temperatures, is 0.01.
264:
265: Similar values of the transition temperatures ($T_{c1}$, $T_{c2}$, $T_{c3}$) for the multiple dynamic transition can be
266: estimated independently from the study of the temperature variations of dynamic energy and specific heat. The dynamical energy ($E$)
267: is defined as the time averaged value of the instantaneous energy over a full cycle of the oscillating magnetic field.
268: From the definition, $E = (1/\tau)\oint H dt$ ($H$ is given in Eqn. \ref{hamiltonian}). The temperature variation of $E$ is
269: studied and shown in Fig. \ref{pt3}(c). This shows two inflection points and one (the middle one) discontinuity
270: at the same location of transition temperatures
271: estimated and mentioned above. This will be very clear if one studies the derivative of the dynamic energy, namely the dynamic
272: specific heat $C = {dE/dT}$ (calculated by using central difference formula \ref{derivative}). The temperature variation
273: of dynamic specific heat $C$ was studied and shown in Fig. \ref{pt3}(d). It indicates three peaks at three different temperatures
274: supporting $T_{c1} \simeq 1.22$, $T_{c2} \simeq 0.96$ and $T_{c3} \simeq 0.88$.
275: Thus the estimation of transition temperatures for the multiple
276: dynamic transition was reexamined by another independent study. The study of the temperature variation of dynamic specific heat
277: has another importance. It independently supports multiple transitions as well as it indicates that these transitions
278: are indeed thermodynamic phase transitions.
279: Now one may employ both the methods of studying the temperature
280: variations of derivatives of dynamic order parameter components
281: and the dynamic specific heat to estimate the transition points.
282: One gets three different dynamic transitions for the parameter values $D = 0.2$,
283: $h_{0x} =0.3$ and $h_{0z} = 1.0$.
284: This three transitions senario is observed for a range of values of $h_{0x}$ (keeping other parameters fixed $D=0.2$ and
285: $h_{0z}=1.0$) between $h_{0x} =0.1$ to $h_{0x} = 0.5$ (within the precision considered here).
286: The temperature variations of dynamic specific heat $C$ are shown in fig \ref{oth2} for two different values
287: of $h_{0x}$ (= 0.1 and 0.5). Both show the three dynamic transitions.
288: It may be noted from fig. \ref{oth2} that the transition temperature decreases as the amplitude $h_{0x}$ increases.
289: Now, let us see what happens if one takes the value of the $x$-amplitude of elliptically polarised magnetic
290: field, i.e., $h_{0x}$, outside this range, keeping all other parameter values unchanged.
291:
292: For $h_{0x} =0.0$ (effectively the field is now linearly polarised along $z$ direction only), to estimate the transition
293: points the temperature variations of dynamic specific heat was studied and plotted in fig. \ref{pt0}(a). This shows
294: two distinct and well separated peaks indicating two
295: dynamic transitions, one at $T \simeq 1.24$ and other at $T \simeq 0.98$. Now to characterise the different phases the temperature
296: variations of dynamic order parameter components were studied. This study is shown in fig. \ref{pt0}(b).
297: This shows that the system starts to get dynamically ordered from a high-temperature disordered phase.
298: So, the high-temperature {\it ordered phase}
299: ($Q_x \neq 0$, $Q_y \neq 0$, $Q_z = 0$) is quite different from $P_1$ (described above for $h_{0x} \neq 0$). The second
300: (low-temperature ordered phase) is $P_3$ type. This transition (from high-temperature ordered phase to low-temperature
301: ordered phase) occurs at $T_{c3} \simeq 0.98$. So, one gets two distinct transitions for $D=0.2$, $h_{0x}=0.0$ and $h_{0z} =1.0$.
302: It may be noted here that the temperature variations of the order parameter components, mainly $Q_x$ and $Q_y$,
303: are quite scattered (near the low-temperature transition).
304: Here, $Q_x$ and $Q_y$ are calculated by averaging over 2000 cycles discarding first 2000 cycles ($10^5 MCSS$).
305: For this reason the derivatives $dQ_x/dT$ and $dQ_y/dT$ do
306: not give smooth variation with respect to temperature and are not shown.
307:
308: Now if the value of $h_{0x}$ is higher (say around $h_{0x} =0.6$) with all other parameter values fixed (i.e.,
309: $f=0.02$, $D=0.2$ and $h_{0z} = 1.0$) the triple dynamic transitions (observed for $h_{0x} =0.3$) reduces to double transitions.
310: It gives two distinct peaks of $C$ indicating two different dynamic transitions at $T \simeq 1.20$ and $T \simeq0.82$.
311: So, in this case one gets
312: two dynamically ordered phases, namely $P_1$ (higher temperature ordered phase) and $P_3$ (lower temperature ordered phase). The
313: transition from $P_1$ phase to $P_3$ phase occurs at $T \simeq 0.82$. The high-temperature ordered phase ($P_1$) transition
314: (from dynamically
315: disordered phase) occurs at $T \simeq 1.20$. These two transitions were reexamined by studying the temperature variations of the
316: derivatives of the dynamic order parameter and obtained the same transition temperatures.
317:
318: From the above discussion it is quite clear that if one studies the dynamic phase transitions by varying $h_{0x}$ only (keeping
319: $D=0.2$ and $h_{0z} = 1.0$ fixed), one would get two transitions for $h_{0x} =0.0$. Just above $h_{0x} =0.0$, i.e., starting from
320: $h_{0x}=0.1$ upto $h_{0x}=0.5$ one would get three transitions (and three ordered phases). Above this value, say $h_{0x} = 0.6$
321: one would get again two transitions (two ordered phases). If one continues further, it was
322: observed that the two transitions feature continues.
323: The transition points (temperatures) are estimated by studying specific heat and the derivatives of the dynamic order parameter
324: components. These are not shown in figure, only the results (obtained from the peak positions of the specific heat
325: and positions of maximum or minimum of ${{dQ_{\alpha}}/{dT}}$) are taken.
326: It was
327: observed further that all the transition points shift towards lower temperatures for higher values of $h_{0x}$. The whole results
328: of these multiple transitions temperatures and plotted in $T-h_{0x}$ plane (fig. \ref{phase}) to get the multiple dynamic phase boundary.
329:
330: In the phase diagram (fig. \ref{phase}), the outermost boundary separtes the dynamically
331: $Y$-ordered phase $P_1:(Q_x = 0, Q_y \neq 0, Q_z =0)$ from the disordered phase $P_0: (Q_x = 0, Q_y = 0, Q_z = 0)$.
332: The transition temperature decreases as the value of the field amplitude $h_{0x}$ increases.
333: For lower values of $h_{0x}$, the
334: region bounded by the boundaries marked by the symbols circle and bullet is the ordered phase characterised as
335: $P_2: (Q_x \neq 0, Q_y = 0, Q_z \neq 0)$. This region is quite small (in area) but very clear and was observed very
336: carefully.
337: The transition temperatures for both the boundaries (left and right) of this phase decrease as the field amplitude
338: $h_{0x}$ increases.
339: The innermost (in the $T-h_{0x}$ plane) region, whose boundary is partly marked by circle (for higher values
340: of $h_{0x}$) and partly marked by bullet (for lower values of $h_{0x}$), represents the low-temperature phase characterised
341: by $P_3:(Q_x = 0, Q_y = 0, Q_z \neq 0)$. Here also, the transition temperature decreases as the value of the field
342: amplitude $h_{0x}$ increases.
343:
344: % Nature of the transitions
345: The question which should arise now,
346: what will be the nature (continuous/discontinuous) of these multiple dynamic transitions ? To get the answer of
347: this questions, the distributions of the magnitudes of the dynamic order parameter components are studied near the
348: transition points. On the phase boundary (Fig.\ref{phase}), a particular value of $h_{0x}(=0.7)$ is chosen. This
349: choice is not arbitrary. In this region, the phase boundaries are well separated to study comfortably
350: the temperature variations
351: of fourth-order Binder cumulant ratio \cite{binder} $U_y = 1 - <Q_y^4>/(3<Q_y^2>^2)$.
352: Since the transition from $P_0$ to $P_1$
353: phase is indicated by $Q_y$ the distribution of the magnitude of $Q_y$ is studied near the transition point. The
354: histogram is shown in Fig.\ref{binder2} for four different temperatures around the transition temperature. This figure
355: shows that as one goes through the results of temperature from $T=1.16$ to that of higher temperature,
356: the peak of the distributions shifts towards $Q_y = 0$ continuously
357: and around $T=1.20$ this gets peaked near $Q_y \simeq 0$.
358: For slightly higher value of the temperature, say $T=1.22$, the distribution is peaked around $Q_y =0$, corresponding to
359: the disordered phase.
360: This indicates the transition is continuous with $T_{c1} \simeq 1.20$ (for the transition from $P_0$ to $P_1$ phase)
361: and one could guess it
362: from the temperature variation of $Q_y$ and ${{dQ_y}/{dT}}$.
363: This same technique was applied to study the nature of the transition
364: from $P_1$ to $P_3$ phase. The distribution of $Q_y$ is shown (in Fig.\ref{binder1}) for three different temperatures.
365: Here, the different (from the previous case just discussed above) senario was observed and
366: may be noted, that the value of $Q_y$ drops to zero from a nonzero value as the system gets cooled.
367: At $T=0.80$ the distribution is singly peaked near $Q_y \simeq 0.63$. At slightly lower
368: temperature $T=0.79$, the distribution get doubly peaked. One additional peak appears near $Q_y \simeq 0$. This simultaneous
369: appearance of two peaks at very close to the transition temperature is a clear signature of discontinuous transition
370: with $T_{c} \simeq 0.79$ (for the transition from $P_1$ to $P_3$ phase). From this study, one can get the idea about the maximum error (0.01 here)
371: in estimation of the transition
372: temperature.
373: As one lowers the temperatures, $T =0.78$, one gets only one peak of the distribution around $Q_y=0$ after the
374: transition.
375: Here also, this could be guessed from the temperature variation of $Q_y$.
376: These studies indicate that the high-temperature transition is a continuous one and the low-temperature
377: transition is discontinuous type. These natures of the transitions were reexamined by studying the temperature variation
378: of Binder cumulant ratio $U_y$.
379: For a continuous transition, $U_y$ should change monotonically from 0 to 2/3 as one tunes the system from the disordered
380: to the ordered phase. On the other hand, for a discontinuous transition, $U_y$ develops sharp minimum, whose location
381: corresponds to the transition point.
382: The results, obtained here, indeed show this (Fig.\ref{cum12}). Fig.\ref{cum12} shows, that as one tunes the system from
383: high temperature, the Binder cumulant ratio $U_y$ first grows monotonically from 0 to 2/3 around $T_{c1} \simeq 1.20$
384: indicating that the high-temperature transition is continuous. As the temperature decreases further, it shows a very
385: sharp minimum around $T_{c} \simeq 0.80$ (for the transition from $P_1$ to $P_3$ phase)
386: indicating the discontinuous nature of the low-temperature transition.
387: These two (multi-histogram analysis and Binder cumulant ratio)
388: studies together confirm that the high-temperature transition is a continuous type and the low-temperature transition
389: is discontinuous type.
390: Since, the evidence of the nature of the high-temperature transition is found to be of continuous type, it will be quite
391: interesting to study the scaling behaviour (if any) with the estimations of critical exponents. But for this, one has to
392: estimate the transition temperatures more precisely which will require larger sizes of the system.
393:
394: What will be the nature of the transition from $P_2$ to $P_3$ phase ? For lower values of
395: $h_{0x}$ (=0.3 chosen here)
396: one would be able to see this transition. From Fig.\ref{pt3}(a), this transition is indicated by the temperature
397: variation of $Q_z$. Since, this transition corresponds to very small change of $Q_z$, it is quite difficult to observe
398: any change of the distribution of $Q_z$ near the transition temperature.
399: However, the temperature variation
400: of the Binder cumulant ratio $U_z = 1-<Q_z^4>/(3<Q_z^2>^2)$ indeed shows (see Fig.\ref{cum3})
401: smooth variation near the transition point and it
402: grows monotonically from 0 to 2/3 which indicates the transition is continuous with $T_{c3} \simeq 0.88$ (for the
403: transition from $P_2$ to $P_3$ phase).
404: All these natures of the transitions are marked in the
405: phase diagram (Fig.\ref{phase}).
406: % Nature of the transitions
407:
408: The finite size study is also done to confirm that the observation is not an artifact of limited size of the system.
409: This study was done for a particular set ($f=0.02$, $D=0.2, h_{0x} =0.3, h_{0z} = 1.0)$ of values of the parameters.
410: Different statistical quantities were studied as functions of temperatures for different linear sizes ($L =
411: 10, 20, 30$ here) of the system. This particular choice of the values of $D$, $h_{0x}$ and $h_{0z}$ is meaningful
412: in the sense that the three transitions phenomenon was observed clearly for this particular values of parameters
413: and the most important part of the phase boundary.
414: Figure \ref{fnt1} shows the temperature variations of ${{dQ_{\alpha}}/{dT}}$ for $L = 10, 20$ and 30. From the
415: figure it is clear that both the sharpness and the height (depth) of maximum (minimum) indicating the transition points
416: increase as the system size increase. The dynamic specific heat is also studied as a function of temperature for different
417: $L$ and same set of other parameter values. This is shown in fig. \ref{fnt3}. This also indicates that the height of
418: the peaks (indicating the transitions) increases as the system size increases. The finite size study, done
419: and briefly reported here, at
420: least may indicate that the multiple dynamic transitions mentioned above is not an artifact of finite size effect.
421:
422: The variances of the dynamic order
423: parameter components i.e.,
424: $L^3 {\rm Var}(Q_{\alpha})=L^3[<Q_{\alpha}^2>-<Q_{\alpha}>^2]=L^3\delta Q_{\alpha}^2$
425: are studied as function of temperature and for different
426: system sizes. The values of other parameters are $f=0.02$, $D=0.2$, $h_{0x} = 0.3$ and $h_{0z} = 1.0$.
427: Figure \ref{fnt2}(a) shows temperature variation of
428: $L^3 {\rm Var}(Q_{x})=L^3[<Q_{x}^2>-<Q_{x}>^2]=L^3 \delta Q_x^2$ for $L = 10, 20$ and 30. It gets sharply
429: peaked at around $T = T_{c2} \simeq 0.96$
430: (the transition temperature from $P_1$ phase to $P_2$ phase).
431: The height of the peak increases for larger system sizes.
432: In fig. \ref{fnt2}(b) the $L^3 {\rm Var}(Q_y)$ is
433: plotted against temperature for different system sizes. It shows two peaks. The high-temperature peak (very short in height
434: but visible in this scale) is located at $T \simeq 1.22$. This corresponds to the transition from disordered ($P_0$) to first
435: Y-ordered ($P_1$) phase. The low-temperature peaks are positioned at around $T \simeq 0.96$ corresponding to the transition
436: from $P_1$ to $P_2$.
437: Here, also the heights of the peaks increases systematically as the system size increases.
438: Similar things are observed for $L^3 {\rm Var}(Q_z)$ and shown in fig. \ref{fnt2}(c). Here, an important thing
439: should be noted that this study indicates
440: that their exists a diverging length scale at the dynamic transition
441: points \cite{rikpre}, for the multiple dynamic transitions in this system. It should also be noted that the transition
442: (from $P_2$ to $P_3$) temperature $T_{c3}$ cannot be resoved from $T_{c2}$ in the present study. It was already indicated
443: by the results shown in Fig.\ref{pt3}(a). The sizes of the errorbars of $Q_x$, $Q_y$ and $Q_z$ becomes maximum near $T_{c1}$ and $T_{c2}$.
444: However, the growth of the errorbars of $Q_z$ is not quite clear near $T_{c3}$.
445:
446: % Frequency dependence of multiple dynamic transitions.
447: The frequency dependence of these multiple dynamic transitions are also studied very briefly. For a lower frequency
448: ($f = 0.01$) of the polarised magnetic field, the multiple dynamic transition was studied with $D=0.2$, $h_{0x}=0.3$
449: $h_{0z}=1.0$. The temperature variations of the magnitudes of the dynamic order parameter components are plotted in
450: Fig.\ref{frq1}(a). This shows similar kind of temperature variations as that was obtained for $f=0.02$ ($D=0.2$,
451: $h_{0x}=0.3$ and $h_{0z}=1.0$),
452: with considerable shifting of transition points towards lower temperatures. From the study of the temperature
453: variations of the derivatives of the dynamic order parameter components (shown in Fig.\ref{frq1}(b)) one gets
454: $T_{c1} \simeq 1.06$, $T_{c2} \simeq 0.76$ and $T_{c3} \simeq 0.62$. A similar study was done for higher values of
455: $h_{0x} (=1.5)$ ($f=0.01$, $D=0.2$ and $h_{0z}=1.0$). The results are shown in Fig.\ref{frq2}. In this case
456: only two transitions were observed and the transition temperatures estimated are $T_{c1} \simeq 0.60$ and $T_{c3} \simeq 0.30$.
457: From these studies, one can get the indication
458: that as the frequency decreases (or as one approaches the static limit) the transition points shifts towards lower temperatures
459: keeping the qualitative nature of the phase diagram (for $f=0.02$) unchanged.
460: This is quite expected result, since as frequency decreases
461: the magnetisation components can follow the field in time
462: and the dynamic transition disappears. This is why the transition is
463: called {\it truely dynamic} and discussed in details in earlier studies \cite{rmp} in the case of Ising model.
464: % Frequency dependence of multiple dynamic transitions.
465:
466: \begin{center} {\bf SUMMARY} \end{center}
467:
468: The classical Heisenberg ferromagnet (uniaxially anisotropic) in presence of a time-varying polarised magnetic field
469: and in contact with a thermal bath is studied by Monte Carlo simulation using Metropolis dynamics. The magnetic field
470: is elliptically polarised and the field vector
471: rotates on XZ - plane. For a very weakly anisotropic (uniaxial) system and particular
472: set of parameter values (amplitudes of field in x and z directions) if the system cools down it undergoes successive
473: phase transitions. For a range of values of field amplitude along x-direction, three dynamic phase transitions were
474: observed. Keeping the values of anisotropy strength and amplitude of field along z direction fixed, the phase transitions
475: were studied for different values of the field amplitude along the x direction. In a plane formed by the temperature
476: and the ratio of field amplitudes the phase diagram for the multiple dynamic phase transition was drawn.
477: The nature (continuous/discontinuous) of the transitions are investigated by the temperature variations of the
478: Binder cumulant ratio and the distribution of order parameter near the transition points.
479: The
480: finite size study was done to check that the transitions are not artifacts of limited system size. The variances of
481: dynamic order parameter components are studied as a function of temperature taking system size as parameter. This
482: particular study indicates the existence of a diverging length scale at the transition points.
483: The choice of the parameter values,
484: in the present study, are obtained from various trials. In other parameter range the multiple transitions
485: are also possible to observe, however this particular choice gives quite distinct results of
486: the observed {\it multicritical behaviour} \cite{multicritical}. Since, the transition across the
487: outermost boundary of the phase diagram is found to be continuous, it will be much interesting to study the
488: scaling behaviour and to estimate the critical exponents.
489: The frequency variations of these multiple dynamic transitions are also studied very briefly. This shows that the
490: qualitative nature of the phase diagram remains unchanged however the transition temperatures are lowered by lowering
491: the frequency of the polarised magnetic field. This is, of course, an expected result
492: experienced from the earlier studies \cite{rmp}, particularly, in
493: the case of the dynamic transition in Ising
494: model.
495:
496: The dynamic transition in Ising ferromagnet can be explained simply by spin reversal and nucleation \cite{nucl}.
497: The multiple dynamic transition occurs in Heisenberg ferromagnet possibly due to the coherent
498: spin rotation \cite{uli}. To get
499: the clear idea about it one has to study also the dynamic configurations of spins in details. The study of the
500: relaxation dynamics of the system in the low anisotropic limit may help to analyse few results.
501:
502: The alternative methods of studying this multiple dynamic phase transitions in anisotropic Heisenberg ferromagnet
503: driven by polarised magnetic field are: (i) to study Landau-Lifshitz-Gilbert equation of motion \cite{hinzke} with
504: Langevin dynamics, (ii) spin wave analysis of anisotropic Heisenberg ferromagnet in presence of polarised magnetic
505: field.
506:
507: \begin{center} {\bf ACKNOWLEDGMENTS}\end{center}
508:
509: The library facility provided by Saha Institute of Nuclear Physics, Calcutta, India, is
510: gratefully acknowledged.
511:
512: \vskip 1 cm
513: %%%%% LIST OF REFERENCES %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
514: \begin{center} {\bf REFERENCES}\end{center}
515:
516: \begin{enumerate}
517:
518: \bibitem{rmp} M. Acharyya, Int. J. Mod. Phys. C 16 (2005) (in press) Cond-mat/0508105,
519: and the references therein; See also
520: B. K. Chakrabarti and M. Acharyya, Rev. Mod. Phys. {\bf 71}, 847 (1999).
521:
522: \bibitem{tome} T. Tome and M. J. de Oliveira, Phys. Rev. A {\bf 41}, 4251 (1990).
523:
524: \bibitem{ma1} M. Acharyya, Phys. Rev. E {\bf 56}, 1234 (1997); Phys. Rev. E {\bf 56}, 2407 (1997)
525:
526: \bibitem{rikpre} S. W. Sides, P. A. Rikvold and M. A. Novotny, Phys. Rev. Lett. {\bf 81}, 834 (1998)
527:
528: \bibitem{matis} D. C. Mattis, {\it The Theory of Magnetism I: Statics and Dynamics}, Springer Series
529: in Solid State Sciences No. 17 (Springer- Verlag, Berlin, 1988).
530:
531: \bibitem{ijmpc} M. Acharyya, Int. J. Mod. Phys. C {\bf 14}, 49 (2003); Int. J. Mod. Phys. C {\bf 12}, 709 (2001).
532:
533: \bibitem{yasui} T. Yasui {\it et al}., Phys. Rev. E {\bf 66}, 036123 (2002); {\bf 67}, 019901(E) (2003)
534: See also, Fujiwara {\it et al}, Phys. Rev. E {\bf 70}, 066132 (2004).
535:
536: \bibitem{jang} H. Jang, M. J. Grimson and C. K. Hall, Phys. Rev. E {\bf 68}, 046115 (2003).
537:
538: \bibitem{huang} Z. Huang, F. Zhang, Z. Chen, Y. Du, Eur. Phys. J. B, {\bf 44}, 423 (2005)
539:
540: \bibitem{mdt1} H. Jang, M. J. Grimson and C. K. Hall, Phys. Rev. B {\bf 67}, 094411 (2003).
541:
542: \bibitem{mdt2} M. Acharyya, Phys. Rev. E {\bf 69}, 027105 (2004).
543:
544: \bibitem{stauffer} D. Stauffer {\it et al.}, {\it Computer Simulation and Computer Algebra} (Springer-Verlag, Heidelberg, 1989);
545:
546: \bibitem{uli} U. Nowak, in {\it Annual Reviews of Computational Physics}, edited by D. Stauffer (World Scientific,
547: Singapore, 2001), Vol. 9, p. 105; D. Hinzke and U. Nowak, Phys. Rev. {\bf B 58}, 265 (1998).
548:
549: \bibitem{metro} N. Metropolis, A. W. Rosenbluth, M. N. Rosenbluth, A. H. Teller, E. Teller, J. Chem. Phys.
550: {\bf 21}, 1087 (1953).
551:
552: \bibitem{krech} M. Krech and D. P. Landau, Phys. Rev. B 60, 3375 (1999)
553:
554: \bibitem{mth} K. F. Riley, M. P. Hobson, and S. J. Bence, in {\it Mathematical Methods for Physics and Engineering},
555: Cambridge University Press, (2004) pp. 1179
556:
557: \bibitem{binder} K. Binder and D. W. Heermann, in {\it Monte Carlo Simulation in Statistical Physics,}
558: Springer Series in Solid-State Sciences,
559: (Springer, New York, 1997); D. P. Landau and K. Binder,
560: in {\it A guide to Monte Carlo Simulations in Statistical Physics}
561: (Cambridge University Press, Cambridge, 2000), pp. 145.
562:
563: \bibitem{multicritical} P. M. Chaikin and T. C. Lubensky, in {\it Principles of Condensed Matter Physics},
564: Cambridge University Press, (2004) pp. 172-187.
565:
566: \bibitem{nucl} M. Acharyya and D. Stauffer, Eur. Phys. J. B {\bf 5}, 571 (1998)
567:
568: \bibitem{hinzke} D. Hinzke, U. Nowak and K. D. Usadel, in {\it Structure and Dynamics of Heterogeneous Systems}, Eds.
569: P. Entel and D. E. Wolf ( World Scientific, Singapore, 1999 ), pp. 331-337.
570:
571: \end{enumerate}
572:
573: \newpage
574: %%%%%%%%%%% LIST OF FIGURE CAPTIONS %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
575: \begin{figure}
576: %Fig-1
577: \caption {The phase diagram in $T-h_{0x}$ plane for the multiple dynamic transitions for $D=0.2$
578: $f = 0.02$ and $h_{0z} =1.0$. C.T. stands for continuous transition and D.T. stands for discontinuous
579: transition. The natures (continuous/discontinuous) of the transitions across the different phase boundary
580: are marked by C.T and D.T.}
581: \label{phase}
582: \end{figure}
583:
584: \begin{figure}
585: %Fig-2
586: \caption {The temperature variations of different dynamical quantities for
587: $f=0.02$, $D=0.2$, $h_{0x} =0.3$ and $h_{0z} = 1.0$.
588: (A) The magnitudes of components of dynamic order parameter $Q_x$, $Q_y$ and $Q_z$. Vertical lines are errorbars.
589: Note that the sizes of the errorbars become maximum near the transition points.
590: (B) derivatives of dynamic order parameters, (C) the dynamic energy and the (D) dynamic specific heat.
591: Solid lines are guide to the eye. The vertical arrows in (C) and (D) indicates the transition points.
592: The maximum error in estimating the transition temperature is 0.01.}
593: \label{pt3}
594: \end{figure}
595:
596: \begin{figure}
597: %Fig-3
598: \caption {The temperature variations of dynamic specific heat for $f=0.02$, $D=0.2$, $h_{0z} =1.0$.
599: (A) for $h_{0x} = 0.1$ and (B) for $h_{0x} = 0.5$.
600: Solid lines are guide to the eye. The maximum error in estimating the transition temperature is 0.01.}
601: \label{oth2}
602: \end{figure}
603:
604: \begin{figure}
605: %Fig-4
606: \caption {The temperature variations of (A) dynamic specific heat and
607: (B) magnitudes of the order parameter components. For $f=0.02$, $D=0.2$, $h_{0x} =0.0$ and $h_{0z} = 1.0$.
608: Vertical lines are errorbars. Note that the sizes of the errorbars becomes maximum near the transition points.
609: Solid line in (A) is guide to the eye.}
610: \label{pt0}
611: \end{figure}
612:
613: \begin{figure}
614: %Fig-5
615: \caption {The unnormalised distributions of the magnitudes of dynamic order parameter component $Q_y$ for
616: different temperatures. Here, $f=0.02$, $D = 0.2$, $h_{0x} = 0.7$, $h_{0z} = 1.0$ and $f = 0.02$.}
617: \label{binder2}
618: \end{figure}
619:
620: \begin{figure}
621: %Fig-6
622: \caption {The unnormalised distributions of the magnitudes of the dynamic order parameter component $Q_y$
623: for different temperatures. Here, $D = 0.2$, $h_{0x} = 0.7$, $h_{0z} = 1.0$ and $f = 0.02$.}
624: \label{binder1}
625: \end{figure}
626:
627: \begin{figure}
628: %Fig-7
629: \caption {The temperature variation of the Binder cumulant ratio $U_y$ near the transition points.
630: Here, $D=0.2$, $h_{0x}=0.7$, $h_{0z}=1.0$ and $f=0.02$.}
631: \label{cum12}
632: \end{figure}
633:
634: \begin{figure}
635: %Fig-8
636: \caption {The temperature variation of the Binder cumulant ratio $U_z$ near the transition point.
637: Here, $D=0.2$, $h_{0x}=0.3$, $h_{0z}=1.0$ and $f=0.02$.}
638: \label{cum3}
639: \end{figure}
640:
641: \begin{figure}
642: %Fig-9
643: \caption {Temperature variations of derivatives of dynamic order parameter components for different system sizes
644: ($L = 10, 20$ and 30) for $f=0.02$, $D=0.2$, $h_{0x} = 0.3$ and $h_{0z} = 1.0$.
645: (A) ${{dQ_x}/{dT}}$ versus $T$. (B) ${{dQ_y}/{dT}}$ versus $T$.
646: (C) ${{dQ_z}/{dT}}$ versus $T$. Solid lines are guide to the eye.}
647: \label{fnt1}
648: \end{figure}
649:
650: \begin{figure}
651: %Fig-10
652: \caption {Temperature variations of dynamic specific heat for different system sizes ($L$ = 10, 20 and 30)
653: for $f=0.02$, $D$ = 0.2, $h_{0x}$ = 0.3 and $h_{0z}$ = 1.0. Solid lines are guide to the eye.}
654: \label{fnt3}
655: \end{figure}
656:
657: \begin{figure}
658: %Fig-11
659: \caption {Temperature variations of variances of the dynamic order parameter components for different system sizes
660: ($L$ = 10, 20 and 30) for $D$ = 0.2, $h_{0x}$ = 0.3 and $h_{0z}$ = 1.0.
661: (A) $L^3 \delta Q_x^2$ versus $T$. (B) $L^3 \delta Q_y^2$ versus $T$. (C) $L^3 \delta Q_z^2$ versus $T$. Solid
662: lines are guide to the eye.}
663: \label{fnt2}
664: \end{figure}
665:
666: \begin{figure}
667: %Fig-12
668: \caption {Temperature variations of the (A) magnitudes of dynamic order parameter components
669: (B) the derivatives of the order parameter components. Here, $D=0.2$, $h_{0x}=0.3$, $h_{0z} = 1.0$
670: and $f = 0.01$.}
671: \label{frq1}
672: \end{figure}
673:
674:
675: \begin{figure}
676: %Fig-13
677: \caption {Temperature variations of the (A) magnitudes of dynamic order parameter components
678: (B) the derivatives of the order parameter components. Here, $D=0.2$, $h_{0x}=1.5$, $h_{0z} = 1.0$
679: and $f = 0.01$.}
680: \label{frq2}
681: \end{figure}
682: \end{document}
683: