betaFeedback

WorksPhysics and Astronomy

Introducing the step Monte Carlo method for simulating dynamic properties

Abstract

In this work, we introduce a simple modification of the Monte Carlo algorithm, which we call step Monte Carlo (sMC). The sMC approach allows to simulate processes far from equilibrium and obtain information about the dynamic properties of the system under investigation. In the approach proposed here the probability of accepting the final (trial) state depends on the activation energy, not on the relative energy between the final and initial state. This barrier height is probed on an ongoing basis, by generating intermediate states along the path connecting the initial and trial positions. Importantly, to calculate the activation energy, our model only requires knowledge of the Hamiltonian without having to introduce additional input parameters such as transition rates etc. The details of sMC are explained for the case of a simple spin model. The comparison of its results with the ones obtained within the frame of stochastic Landau-Lifshitz-Gilbert indicates the correctness of sMC. In our opinion, the proposed here method can be applied to simulate other processes, for example dynamics of classical atoms and complex fluids, diffusion, nucleation, surface adsorption and crystal growth processes.

Introduction

The Monte Carlo (MC) method has proved to be a valuable simulation tool in many branches of science such as physics, chemistry [1], biology [1], computer science [1], economics - finance [2, 3] and engineering [4]. In the field of condensed matter physics and materials science MC can be used to study, among many others, system of classical particles [5, 6], classical spin systems [7, 8, 9, 10], nucleation and crystal growth processes [11, 12, 13, 14, 15], polymer solutions [16], hopping transport [17, 18, 19], percolation [20] and fractals problems. The MC algorithm is the natural choice for studying the static properties of a system, where dynamical effects are not required. Then, the advantages of Monte Carlo are the relative ease of implementation and the rapid convergence to steady state. In general, the theory of equilibrium properties is well developed for a wide variety of models and materials. Here, however, we shall consider situations far from equilibrium. For example, the MBE growth of thin films proceeds by both deposition and diffusion processes. The probability of an atom to hop to a nearest empty site, does not depend only on the relative energies of the configuration before and after the hopping event, but rather on the barrier height. The Kinetic Monte Carlo (KMC) method [21, 22] was developed for evolving systems dynamically from state to state. KMC simulates how the occupation of the sites changes over time in a system with a known transition rates (these rates are inputs to the KMC algorithm).

The modified MC approach proposed here (similarly to the KMC approach) the probability of accepting the final state depends on the activation energy, not on the relative energy between the final and initial state. However, the barrier height is calculated on an ongoing basis, by generating intermediate states with a predefined step Δ\Delta. Therefore, we name this method step Monte Carlo (sMC). Importantly, the sMC method correctly takes into account the presence of various local barriers and it obeys the detailed balance condition, even if the system is not in equilibrium. As a result, the appropriate dynamics of the tested system is simulated. The details of sMC algorithm are explained for the case of magnetization process. To test the correctness of sMC, we compare its results with those obtained by stochastic Landau-Lifshitz-Gilbert (sLLG) approach.

Step Monte-Carlo method

Here we outline the details of step Monte Carlo method, that allows us to simulate processes far from equilibrium and obtain information about the dynamic properties of the system under investigation. The differences between sMC and standard MC are shown schematically in Figure 1 for the case of the magnetization process. In sMC, the probability of accepting the final magnetic direction mfinal\mathbf{m}_{final} depends on the activation energy (barrier height) EAEE_{AE}, not on the relative energy between the final and initial state Efinal−EinitalE_{final}-E_{inital}. To calculate the barrier height, we probe energies Ek=f(mk)E_k=f(\mathbf{m}_k) of intermediate magnetic states m2\mathbf{m}_2, m3,…,mN−1\mathbf{m}_3,\ldots,\mathbf{m}_{N-1}. These intermediate states are generated on the arc connecting the initial minit=m1\mathbf{m}_{init}=\mathbf{m}_1 and the final (trial) mfinal=mN\mathbf{m}_{final}=\mathbf{m}_N magnetization direction. This means that mk+1\mathbf{m}_{k+1} is obtained using a predefined step Δ\Delta in such a way that ∣mk+1−mk∣∝Δ\lvert\mathbf{m}_{k+1}-\mathbf{m}_k\rvert\propto\Delta.

Image not extracted

Figure 1. Schematic illustration of (a) Monte Carlo and (b) step Monte Carlo (sMC) methods for the case of magnetization process (one Monte Carlo move). The magnetization direction m\mathbf{m} (bottom panels) can take any point on the unit sphere with energy E=f(m)E=f(\mathbf{m}) (top panels). minit\mathbf{m}_{init} and mfinal\mathbf{m}_{final} are initial and final (trial) direction of magnetic moment with energy EinitE_{init} and EfinalE_{final} respectively. In the sMC approach the probability of accepting the final random state mfinal\mathbf{m}_{final} depends on the activation energy EAE=Emax−EinitE_{AE}=E_{max}-E_{init}, not on the relative energy Efinal−EinitE_{final}-E_{init}. This barrier height is probed by generating intermediate magnetic states mk\mathbf{m}_k (k=2,3,…,N−1k=2,3,\ldots,N-1) on the arc connecting the initial minit=m1\mathbf{m}_{init}=\mathbf{m}_1 and the final mfinal=mN\mathbf{m}_{final}=\mathbf{m}_N magnetization direction with a predefined magnetization step Δ\Delta. This means that mk+1\mathbf{m}_{k+1} is obtained in such a way that ∣mk+1−mk∣∝Δ\lvert\mathbf{m}_{k+1}-\mathbf{m}_k\rvert\propto\Delta.

fined step Δ\Delta in such a way that ∣mk+1−mk∣∝Δ\lvert\mathbf{m}_{k+1}-\mathbf{m}_k\rvert\propto\Delta. Details of the generation of intermediate states are shown in Figure 2. After obtaining energies EkE_k, the barrier height is obtained from EAE=max⁡{E2,E3,…,EN}−EinitE_{AE}=\max\{E_2,E_3,\ldots,E_N\}-E_{init}, where EN=EfinalE_N=E_{final}. Finally, we proceed in a similar way to the Monte Carlo Metropolis algorithm for a classical spin system. The complete procedure is shown below:

Image not extracted

Figure 2. Schematic illustration of the generation of intermediate states by shifting the magnetization direction on the unit sphere (on the arc connecting the initial minit\mathbf{m}_{init} and the final mfinal\mathbf{m}_{final} magnetization direction).

  1. Select a random atom jj with the initial direction of the magnetic moment mj,init\mathbf{m}_{j,init}.

  1. Choose randomly a new trial direction mj,final\mathbf{m}_{j,final}.

  1. Generate intermediate magnetic states mj,2,mj,3,…,mj,N−1\mathbf{m}_{j,2},\mathbf{m}_{j,3},\ldots,\mathbf{m}_{j,N-1} with the predefined step Δ\Delta on the arc connecting the initial mj,init=mj,1\mathbf{m}_{j,init}=\mathbf{m}_{j,1} and the final mj,final=mj,N\mathbf{m}_{j,final}=\mathbf{m}_{j,N} magnetization direction (see Figure 2).

  1. Calculate energies Ej,k=f(mj,k)E_{j,k}=f(\mathbf{m}_{j,k}) and the activation energy EAE=max⁡{Ej,2,…,Ej,N=Ej,final}−Ej,initE_{AE}=\max\{E_{j,2},\ldots,E_{j,N}=E_{j,final}\}-E_{j,init}

  1. The new trial direction mj,final\mathbf{m}_{j,final} is accepted with the probability P=exp⁡(−EAEkBT)P=\exp\left(-\frac{E_{AE}}{k_BT}\right), that is:

a. Draw a random number rr uniformly distributed between 0 and 1.

b. If r≤exp⁡(−EAEkBT)r\leq\exp\left(-\frac{E_{AE}}{k_BT}\right), accept the new state, otherwise reject it.

  1. Repeat the procedure.

One complete iteration of sMC consists of NCN_C such trial moves (one move corresponds to the whole procedure 1–5 presented above), where NCN_C is the number of atoms in the system. In sMC approach we use 4⋅1074 \cdot10^7 equilibration and averaging iterations.

Here for generation of the trial direction mj,final\mathbf{m}_{j,final} we choose randomly one of the two different trial moves: Gaussian or random (c.f. Figure 3.b-c and 4.b in Ref. 9). The Gaussian trial move creates mj,final\mathbf{m}_{j,final} by shifting the initial direction mj,init\mathbf{m}_{j,init} on the unit sphere according to the formula:

Image not extracted

Figure 3. Investigated magnetic clusters composed of up to four magnetic ions (blue circles) coupled by a ferromagnetic exchange interaction (red lines).

mj,final=mj,init+σgΓ∣mj,init+σgΓ∣,(1)\mathbf{m}_{j,final} = \frac{\mathbf{m}_{j,init}+\sigma_g\mathbf{\Gamma}}{\lvert\mathbf{m}_{j,init}+\sigma_g\mathbf{\Gamma}\rvert}, \tag*{(1)}

where Γ\mathbf{\Gamma} is a 3D random vector characterized by the Gaussian distribution with a mean of zero and a standard deviation of 1. The width of the cone σg\sigma_g takes the following form σg=25(kBTμB)1/5\sigma_g = \frac{2}{5}\left(\frac{k_BT}{\mu_B}\right)^{1/5} [9], which aims to obtain a move acceptance rate of around 50 %. The random trial move creates a random point on the unit sphere mj,final=ΓU/∣ΓU∣\mathbf{m}_{j,final} = \mathbf{\Gamma}_U/\lvert\mathbf{\Gamma}_U\rvert, where ΓU\mathbf{\Gamma}_U is a 3D random vector with each components uniformly distributed between −1 and 1. In order to correctly (densely enough) arrange intermediate states (see Figure 2) all trial states meeting the following condition mj,init⋅mj,final≤−0.97\mathbf{m}_{j,init}\cdot\mathbf{m}_{j,final} \leq-0.97 are rejected.

We can see that the implementation of the sMC algorithm is a little more difficult than the standard MC one. Also the execution time for one sMC iteration is longer than in MC case, especially for small values of Δ\Delta. However, the advantage of using the new method is that MC provides proper Boltzmann distribution of states in relation to the global energy minimum, whereas sMC (in the short run) provides Boltzmann distribution of states relative to the local energy minimum. This guarantees the correct dynamics of sMC, as energy barriers much higher than kBTk_BT prevent from too fast global thermalisation of spins.

It is important to note, that the sMC approach fulfills two general conditions for validity of Monte Carlo algorithms

: ergodicity and the condition of detailed balance. Ergodicity expresses the requirement that all possible configurations of the system can be reached from any other configuration in a finite number of Markov steps. The detailed balance condition requires that, in the long run, the system approaches the correct thermal equilibrium. In sMC, for any two configurations BB and CC we have

W(B→C)W(C→B)=e−(Emax⁡−EB)/(kBT)e−(Emax⁡−EC)/(kBT)=e−(EC−EB)/(kBT)=p(C)p(B)(2)\frac{W(B\to C)}{W(C\to B)}=\frac{e^{-(E_{\max}-E_B)/(k_BT)}}{e^{-(E_{\max}-E_C)/(k_BT)}}=e^{-(E_C-E_B)/(k_BT)}=\frac{p(C)}{p(B)} \tag*{(2)}

where W(B→C)W(B\to C) describes the probability to go from state BB to state CC in one move of the Markov process. p(B)=e−EB/(kBT)/Zp(B)=e^{-E_B/(k_BT)}/Z is the Boltzmann probability distribution, in which EBE_B is the energy of configuration BB and ZZ is the partition function. Emax⁡E_{\max} corresponds to the energy of the barrier (see Figure 1 with B=minitB=\mathbf{m}_{init} and C=mfinalC=\mathbf{m}_{final}). Eq. (2) is also true in the case with no barrier, that is, when EB=Emax⁡E_B=E_{\max} or EC=Emax⁡E_C=E_{\max} (then sMC moves coincide with MC ones).

One should note, that the sMC approach presented here for spin systems can be generalized in a straightforward way to simulate other processes, for example diffusion, nucleation, surface adsorption, crystal growth processes and dynamics of classical atoms. Then we just generate intermediate states along the line connecting the initial and final (trial) positions. Moreover, to simulate the crystal growth, our model only requires knowledge of the Hamiltonian without having to assume a specific discrete structure of the crystal lattice.

To test the correctness of sMC, we compare its results with those obtained by stochastic Landau-Lifshitz-Gilbert (sLLG) equation [9, 23, 24]. Using the sLLG approach, it was possible to simulate the dynamic properties of magnetic materials [23] including ferromagnetic hysteresis loops [9], time evolution of a system of spins [9], spin waves [25], domain wall motion [26, 27] and magnetization switching processes [28, 29]. The sLLG equation, applied at the atomistic level, reads

∂mi∂t=−γ1+αG2[mi×Heffi+αGmi×(mi×Heffi)].(3)\frac{\partial\mathbf{m}_i}{\partial t}=-\frac{\gamma}{1+\alpha_G^2}\left[\mathbf{m}_i\times\mathbf{H}_{eff}^i+\alpha_G\mathbf{m}_i\times\left(\mathbf{m}_i\times\mathbf{H}_{eff}^i\right)\right]. \tag*{(3)}

Here γ\gamma is the gyromagnetic ratio, αG\alpha_G is the damping parameter. The symbol mi\mathbf{m}_i describes normalized (∣mi∣=1\lvert\mathbf{m}_i\rvert= 1) direction of the local magnetic moment of atom ii and Heffi\mathbf{H}_{\mathrm{eff}}^i corresponds to effective magnetic field acting on this spin. Heffi\mathbf{H}_{\mathrm{eff}}^i is calculated from the first derivative of the total spin Hamiltonian H\mathcal{H} and incorporates also the instantaneous thermal magnetic field HThi\mathbf{H}_{Th}^i, according to the following formula

Heffi=−1μS∂H∂mi+HThi(4)\mathbf{H}_{\mathrm{eff}}^i = -\frac{1}{\mu_S}\frac{\partial\mathcal{H}}{\partial\mathbf{m}_i} + \mathbf{H}_{Th}^i \tag*{(4)}

with

HThi=Γ(t)2αGkBTγμSΔt(5)\mathbf{H}_{Th}^i = \mathbf{\Gamma}(t)\sqrt{\frac{2\alpha_G k_B T}{\gamma\mu_S\Delta t}} \tag*{(5)}

The instantaneous magnetic field HThi\mathbf{H}_{Th}^i represents thermal effects (fluctuations) in the system, and allows to simulate the magnetization at a finite temperature TT. Γ(t)\mathbf{\Gamma}(t) is a random vector in 3D space characterized by the Gaussian distribution with a mean of zero and a standard deviation of 1. μS=gμBS\mu_S = g\mu_B S is an actual value of atomic moment with spin S=2S = 2, electron g-factor g=2g = 2 and Δt=5⋅10−5\Delta t = 5 \cdot10^{-5} ns is the time step of simulation. Additionally, in sLLG method we use αG=1\alpha_G = 1 and 2⋅10102 \cdot10^{10} equilibration and averaging iterations.

Comparison with stochastic LLG approach

Having outlined the details of the step Monte Carlo method we now proceed to a comparative study of the magnetization dynamics obtained within the frame of the sMC and standard sLLG approach. We analyze a simple model consisting of few interacting magnetic atoms possessing the uniaxial anisotropy along the zz axis. The investigated small magnetic clusters are shown in Figure 3. The Hamiltonian of the system reads

H=−kz∑imiz2−μSH∑imi−J∑i>jmimj(6)\mathcal{H} = -k_z \sum_i m_{iz}^{2} - \mu_S \mathbf{H}\sum_i \mathbf{m}_i - J\sum_{i>j}\mathbf{m}_i\mathbf{m}_j \tag*{(6)}

with appropriate terms relating to uniaxial anisotropy energy, Zeeman energy and the exchange coupling between atoms. Here kz=0.4k_z = 0.4 meV is the anisotropy constant, J=2J = 2 meV describes the exchange parameter. The symbol H\mathbf{H} corresponds to the external magnetic field.

In Figure 4 we present the magnetization MM of a single ion, computed as a function of the external magnetic field H\mathbf{H} applied along the magnetic easy axis, namely H∥z\mathbf{H} \parallel\mathbf{z}. The simulations are obtained using the sMC method with different steps Δ\Delta. The standard MC approach is represented by Δ=3\Delta= 3. Only in this case we observe the rapid convergence to equilibrium state with no hysteresis loop seen in M−HM-H curve. By decreasing Δ\Delta critical dynamic behaviors far from equilibrium emerge. Now the magnetization has to overcome the energy barrier for switching, what creates the hysteresis loop. By decreasing Δ\Delta this barrier height is probed with higher resolution leading to the increase of the coercive field HCH_C. However, for Δ≤10−1\Delta\le10^{-1} we observe saturation of HCH_C and a clear presence of thermal fluctuations.

Image not extracted

Figure 4. The magnetization MM of single ion, computed as a function of the external magnetic field H\mathbf{H} applied along the magnetic easy axis at T=0.04T = 0.04 K. The simulations are performed using the sMC method with different steps Δ\Delta. The standard MC approach is represented by Δ=3\Delta= 3.

In Figure 5 the comparison of numerical results obtained using the sMC (Δ=10−3\Delta= 10^{-3}) and sLLG method is displayed. First, in Figure 5a we plot the magnetization MM of a single ion with the magnetic field H\mathbf{H} applied at different angles from the easy axis z\mathbf{z} at the temperature T=0.04T = 0.04 K. As we see both approaches give very similar quantitative results. Small discrepancies are due to the presence of thermal fluctuations. These discrepancies decrease with lowering the temperature as shown in Figure 5b. An additional source of the observed discrepancy between the

Image not extracted

Figure 5. The comparison of numerical results between sMC (Δ=10−3\Delta= 10^{-3}) and sLLG methods. a) The magnetization MM of a single ion with the magnetic field H\mathbf{H} applied at different angles θ\theta from the easy axis z\mathbf{z} at temperature T=0.04T = 0.04 K. b) The magnetization MM of a single ion at different temperatures TT with the magnetic field H\mathbf{H} applied along the easy axis H∥z\mathbf{H}\parallel\mathbf{z}.

results of sLLG and sMC is the fact, that the sLLG equation consists of two processes: precessional motion of the magnetization along the effective magnetic field and the damping term parametrized by αG\alpha_G. In the MC method only the last term, related to the relaxation of the system to equilibrium, is present.

Finally, magnetizations for larger clusters are presented in Figure 6. At finite temperature, the coercive field increases slightly with the size of the cluster NCN_C, as expected. Again, the consistency between the sLLG and sMC results is very well, indicating a correct assumptions and implementation of the sMC algorithm.

Image not extracted

Figure 6. The comparison of numerical results between sMC (Δ=10−3\Delta= 10^{-3}) and sLLG methods for different clusters with size NCN_C. The magnetization MM of a pair (a), closed triplet (b) and square quartet (c) at the temperature T=0.04T = 0.04 K. The magnetic field H\mathbf{H} is applied at two different angles θ\theta from the easy axis z\mathbf{z}, namely θ=0\theta= 0 and θ=45∘\theta= 45^\circ.

Interestingly, the coercive field obtained from sLLG is slightly smaller than the one obtained from the sMC approach. This is especially noticeable for the magnetic filed applied along the easy axis H∥z\mathbf{H}\|\mathbf{z} (θ=0\theta= 0). In our opinion, it is connected with the fact that in sMC alogirthm the Landau-Lifshitz (LL) term is absent. This LL term producing precessional motion of magnetization along the effective magnetic field seems to be important in such switching processes.

Conclusions

Here, we present a modification of MC algorithm, named step Monte Carlo (sMC), that allows to simulate dynamic properties of various classical systems. In sMC the probability of accepting the final state depends on activation energy (barrier height), not on the relative energy between the final and initial state. This barrier height is probed by generating intermediate states with a predefined step Δ\Delta. To calculate the activation energy, our model only requires knowledge of the Hamiltonian without having to introduce additional input parameters such

as transition rates etc. Importantly, the sMC method correctly takes into account the presence of various local barriers. As a result, the appropriate dynamics of the tested system is simulated. To test the correctness of sMC, we simulate dynamic magnetic properties for interacting spin system and compare its results with those obtained by stochastic Landau-Lifshitz-Gilbert (sLLG) approach. The obtained excellent correspondence of both methods indicates the correctness of the proposed approach. In our opinion, sMC method can be applied to simulate other processes, for example dynamics of classical atoms and complex fluids, diffusion, and crystal growth processes.

Acknowledgments

I would like to thank M. Załuska-Kotur, C. Sliwa, C. Simserides and M. Sawicki for proofreading of the manuscript and valuable suggestions. The work is supported by the National Science Centre (Poland) through project OPUS (2018/31/B/ST3/03438).The calculations were made with the support of the Interdisciplinary Center for Mathematical and Computational Modeling of the University of Warsaw (ICM UW) under the computational grant no GB77-6.

References

  1. [1]J. S. Liu. Monte Carlo Strategies in Scientific Computing. Springer Series in Statistics. Springer-Verlag, NY, 2004. doi:https://doi.org/10.1007/978-0-387-76371-2.
  2. [2]P. Glasserman. Monte Carlo Methods in Financial Engineering. Stochastic Modelling and Applied Probability. Springer Science+Business Media, NY, 2003.DOI
  3. [3]T. Gerstner and P. Kloeden. Recent developments in computational finance : foundation, algorithms and applications. Interdisciplinary Mathematical Sciences. Singapore World Scientific, 2013.
  4. [4]G. S. Fishman. Monte Carlo : Concepts, Algorithms, and Applications. Springer Series in Operations Research and Financial Engineering. Springer New York, NY, 2003. doi:https://doi.org/10.1007/978-1-4757-2553-7.
  5. [5]D. Frenkel and B. Smit. Understanding Molecular Simulation : From Algorithms to Applications. Academic Press, San Diego, 2002.DOI
  6. [6]A. R. Leach. Molecular Modelling : Principles and Applications. Longman, Harlow, 1996.
  7. [7]M. Sawicki, T. Devillers, S. Galeski, C. Simserides, S. Dobkowska, B. Faina, A. Grois, A. Navarro-Quezada, K. N. Trohidou, J. A. Majewski, T. Dietl, and A. Bonanni. Origin of low-temperature magnetic ordering in Ga1−xMnxN. Phys. Rev. B, 85:205204, 2012. doi: 10.1103/PhysRevB.85.205204.
  8. [8]C. Simserides, J.A. Majewski, K.N. Trohidou, and T. Dietl. Theory of ferromagnetism driven by superexchange in dilute magnetic semiconductors. EPJ Web of Conferences, 75:01003, 2014. doi: http://dx.doi.org/10.1051/epjconf/20147501003.
  9. [9]R. F. L. Evans, W. J. Fan, P. Chureemart, T. A. Ostler, M. O. A. Ellis, and R. W. Chantrell. Atomistic spin model simulations of magnetic nanomaterials. J. Phys.: Condens. Matter, 26:103202, 2014. doi: http://dx.doi.org/10.1088/0953-8984/26/10/103202.
  10. [10]H. Sato, P. Chureemart, F. Matsukura, R. W. Chantrell, H. Ohno, and R. F. L. Evans. Temperature-dependent properties of CoFeB/MgO thin films: Experiments versus simulations. Phys. Rev. B, 98:214428, 2018. doi: 10.1103/PhysRevB.98.214428.
  11. [11]H. Sitter. MBE growth mechanisms; studies by Monte Carlo simulation. Thin Solid Films, 267:37, 1995. doi: https://doi.org/10.1016/0040-6090(95)06593-8. Workshop on molecular beam epitaxy-growth physics and technology.
  12. [12]P. J. C. M. van Hoof, W. J. P. van Enckevort and M. Schoutsen. The growth of extremely thin crystals: a Monte Carlo study and an application to n-paraffins. J. Cryst. Growth, 193:679, 1998. doi: https://doi.org/10.1016/S0022-0248(98)00532-6.
  13. [13]D. L. Woodraska and J. A. Jaszczak. A Monte Carlo simulation method for {111} surfaces of silicon and other diamond-cubic materials. Surf. Science, 374:319, 1997. doi:https://doi.org/10.1016/S0039-6028(96)01201-0.
  14. [14]M. A. Zaluska-Kotur, F. Krzyżewski, S. Krukowski, R. Czernecki, and M. Leszczyński. Structures Built by Steps Motion during Sublimation from Annealed GaN(0001) Surface. Crystal Growth & Design, 13:1006, 2013. doi:10.1021/cg301521a.
  15. [15]M. S. Song, T. Koren, M. Zaluska-Kotur, R. Buczko, N. Avraham, P. Kacman, H. Shtrikman, and H. Beidenkopf. Sub-band spectrum engineering via structural order in tapered nanowires. Nano Letters, 21:10215, 2021. doi:10.1021/acs.nanolett.1c03071.
  16. [16]N. Urakami and M. Takasu. Multicanonical Monte Carlo Simulation of a Polymer with Stickers. J. Phys. Soc. Jap., 65:2694, 1996. doi:10.1143/JPSJ.65.2694.
  17. [17]K. Murayama and J. Kuwabara. Monte Carlo simulation of electron transport due to hopping on band edge long-range-fluctuating with self-affine fractal dimension. Sol. State Comm., 103:591, 1997. doi: https://doi.org/10.1016/S0038-1098(97)00256-1.
  18. [18]S. Ma, H. Xu, Y. Li, and Z. Song. Hopping transport and electrical conductivity in one-dimensional systems with off-diagonal disorder. Physica B: Cond. Matter, 398:55, 2007. doi:https://doi.org/10.1016/j.physb.2007.04.067.
  19. [19]M. Dimakogianni, C. Simserides, and G.P. Triberis. Density of states and extent of wave function: two crucial factors for small polaron hopping conductivity in 1D. Phil. Mag., 93:2729, 2013. doi:10.1080/14786435.2013.785639.
  20. [20]D. Dhar and M. Barma. Monte Carlo simulation of directed percolation on a square lattice. J. Phys. C: Solid State Phys., 14:L1, 1981. doi:10.1088/0022-3719/14/1/001.
  21. [21]W. M. Young and E. W. Elcock. Monte Carlo studies of vacancy migration in binary ordered alloys: I. Proc. Phys. Soc., 89:735, 1966. doi:10.1088/0370-1328/89/3/329.
  22. [22]A. B. Bortz, M. H. Kalos, and J. L. Lebowitz. A new algorithm for Monte Carlo simulation of Ising spin systems. J. Comp. Phys., 17:10, 1975. doi: https://doi.org/10.1016/0021-9991(75)90060-1.
  23. [23]R. F. L. Evans, U. Atxitia, and R. W. Chantrell. Quantitative simulation of temperature-dependent magnetization dynamics and equilibrium properties of elemental ferromagnets. Phys. Rev. B, 91:144425, 2015. doi: http://dx.doi.org/10.1103/PhysRevB.91.144425.
  24. [24]Y. K. Edathumkandy and D. Sztenkiel. Comparative study of magnetic properties of Mn3+ magnetic clusters in GaN using classical and quantum mechanical approach. J. Magn. Magn. Mater., 562:169738, 2022. doi: https://doi.org/10.1016/j.jmmm.2022.169738.
  25. [25]S. A. Bender, H. Skarsvåg, A. Brataas, and R. A. Duine. Enhanced spin conductance of a thin-film insulating antiferromagnet. Phys. Rev. Lett., 119:056804, 2017. doi: 10.1103/PhysRevLett.119.056804.
  26. [26]F. Schlikeiser, U. Ritzmann, D. Hinzke, and U. Nowak. Role of entropy in domain wall motion in thermal gradients. Phys. Rev. Lett., 113:097201, 2014. doi: 10.1103/PhysRevLett.113.097201.
  27. [27]D. Hinzke and U. Nowak. Domain wall motion by the magnonic spin seebeck effect. Phys. Rev. Lett., 107:027205, 2011. doi:10.1103/PhysRevLett.107.027205.
  28. [28]M. O. A. Ellis and R. W. Chantrell. Switching times of nanoscale FePt : Finite size effects on the linear reversal mechanism. Appl. Phys. Lett., 106:162407, 2015. doi: 10.1063/1.4919051.
  29. [29]W. Scholz, T. Schrefl, and J. Fidler. Micromagnetic simulation of thermally activated switching in fine particles. J. Magn. Magn. Mater., 233:296, 2001. doi: https://doi.org/10.1016/S0304-8853(01)00032-4.

Paper details

vlnc:W_rzjm6k44
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 10

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents