Cluster Mean-field Approximations Research Articles

Cluster truncated Wigner approximation in strongly interacting systems

We present a general method by which linear quantum Hamiltonian dynamics with exponentially many degrees of freedom is replaced by approximate classical nonlinear dynamics with the number of degrees of freedom (phase space dimensionality) scaling polynomially in the system size. This method is based on generalization of the truncated Wigner approximation (TWA) to a higher dimensional phase space, where phase space variables are associated with a complete set of quantum operators spanning finite size clusters. The method becomes asymptotically exact with increasing cluster size. The crucial feature of TWA is fluctuating initial conditions, which we approximate by a Gaussian distribution. We show that such fluctuations dramatically increase accuracy of TWA over traditional cluster mean-field approximations. In this way we can treat on equal footing quantum and thermal fluctuations as well as compute entanglement and various equal and non-equal time correlation functions. The main limitation of the method is exponential scaling of the phase space dimensionality with the cluster size, which can be significantly reduced by using the language of Schwinger bosons and can likely be further reduced by truncating the local Hilbert space variables. We demonstrate the power of this method analyzing dynamics in various spin chains with and without disorder and show that we can capture such phenomena as long time hydrodynamic relaxation, many-body localization and the ballistic spread of entanglement.

Role of spin-orbit coupling in the Kugel-Khomskii model on the honeycomb lattice

We study the effective spin-orbital model for honeycomb-layered transition metal compounds, applying the second-order perturbation theory to the three-orbital Hubbard model with the anisotropic hoppings. This model is reduced to the Kitaev model in the strong spin-orbit coupling limit. Combining the cluster mean-field approximations with the exact diagonalization, we treat the Kugel-Khomskii type superexchange interaction and spin-orbit coupling on an equal footing to discuss ground-state properties. We find that a zigzag ordered state is realized in the model within nearest-neighbor interactions. We clarify how the ordered state competes with the nonmagnetic state, which is adiabatically connected to the quantum spin liquid state realized in a strong spin-orbit coupling limit. Thermodynamic properties are also addressed. The present work should provide another route to account for the Kitaev-based magnetic properties in candidate materials.

Ground-state phase diagram in the Kugel-Khomskii model with finite spin-orbit interactions

We study ground-state properties in the Kugel-Khomskii model on the two-dimensional honeycomb lattice. Using the cluster mean-field approximations, we deal with the exchange and spin-orbit couplings on an equal footing. We then discuss the stability of the ferromagnetically ordered states against the nonmagnetic state, which is adiabatically connected to the quantum spin liquid state realized in a strong spin-orbit coupling limit.

Cluster mean-field approximations with the coherent-anomaly-method analysis for the driven pair contact process with diffusion

The cluster mean-field approximations are performed, up to 13 cluster sizes, to study the critical behavior of the driven pair contact process with diffusion (DPCPD) and its precedent, the PCPD in one dimension. Critical points are estimated by extrapolating our data to the infinite cluster size limit, which are in good accordance with recent simulation results. Within the cluster mean-field approximation scheme, the PCPD and the DPCPD share the same mean-field critical behavior. The application of the coherent anomaly method, however, shows that the two models develop different coherent anomalies, which lead to different true critical scaling. The values of the critical exponents for the particle density, the pair density, the correlation length, and the relaxation time are fairly well estimated for the DPCPD. These results support and complement our recent simulation results for the DPCPD.

Role of diffusion in branching and annihilation random walk models

Different branching and annihilating random walk models are investigated by the cluster mean-field method and simulations in one and two dimensions. In the case of the A-->2A , 2A--> 0 model the cluster mean-field approximations show diffusion dependence in the phase diagram as was found recently by the nonperturbative renormalization group method [Phys. Rev. Lett. 92, 255703 (2004)]]. The same type of survey for the A-->2A , 4A--> 0 model results in a reentrant phase diagram, similar to that of the 2A-->3A , 4A--> 0 model [Phys. Rev. E 69, 036112 (2004)]]. Simulations of the A-->2A , 4A--> 0 model in one and two dimensions confirm the presence of both the directed percolation transitions at finite branching rates and the mean-field transition at zero branching rate. In two dimensions the directed percolation transition disappears for strong diffusion rates. These results disagree with the predictions of the perturbative renormalization group method.

Pecularities of Bethe-like approximations and long-range-interaction Ising models.

The mean-field approximation and the Bethe approximation are two of the most often used approximations when one wants to obtain approximations of the phase diagrams and the critical temperature of lattice spin systems. Both can and have often been generalized to produce what are known as cluster mean-field and Bethe approximations. Generally, three characteristics are associated with these approximations. First, they give upper bounds to the critical temperature; second, considering larger clusters will result in better approximations; and third, the Bethe approximation is better than the corresponding mean-field approximation. We show what we believe to be a rather surprising result that, for one-dimensional Ising models with algebraically decaying interactions falling off slowly enough, the Bethe cluster approximations violate all three of these characteristics.

Critical behavior of the one-dimensional diffusive pair contact process.

The phase transition of the one-dimensional diffusive pair contact process is investigated by N cluster mean-field approximations and high precision simulations. The N=3,4 cluster approximations exhibit smooth transition line to absorbing state by varying the diffusion rate D with beta(2)=2 mean-field order parameter exponent of the pair density. This contradicts with former N=2 results, where two different mean-field behavior was found along the transition line. Extensive dynamical simulations on L=10(5) lattices give estimates for the order parameter exponents of the particles for 0.05<or=D<or=0.7. These data may support former two distinct class findings. However, the gap between low- and high-D exponents is narrower than previously estimated and the possibility for interpreting numerical data as a single class behavior with exponents alpha=0.21(1), beta=0.40(1) assuming logarithmic corrections is shown. Finite-size scaling results are also presented.

Dynamical mean-field approximation for a pair contact process with a particle source.

The one-dimensional pair contact process with a particle source is studied by using dynamical cluster mean-field approximations with sites up to n=12. The results obtained for different levels of approximation become convergent especially for n>or=6 and allow us to derive reliable extrapolations to the limit n--> infinity. At the zero source limit, the critical point exhibits a discontinuity whose magnitude vanishes with 1/n. Coherent anomaly analysis of the data supports the conclusion that the vanishing of the order parameter and the density of isolated particles have the same critical behavior. In contrast to an earlier prediction, the present approximation does not support the existence of critical behavior in the inactive phase where the frozen density of isolated particles depends on the initial state.

Phase transition of a two-dimensional binary spreading model.

We investigated the phase transition behavior of a binary spreading process in two dimensions for different particle diffusion strengths (D). We found that N>2 cluster mean-field approximations must be considered to get consistent singular behavior. The N=3,4 approximations result in a continuous phase transition belonging to a single universality class along the D subset (0,1) phase transition line. Large scale simulations of the particle density confirmed mean-field scaling behavior with logarithmic corrections. This is interpreted as numerical evidence supporting the bosonic field theoretical prediction that the upper critical dimension in this model is d(c)=2. The pair density scales in a similar way but with an additional logarithmic factor to the order parameter. At the D=0 end point of the transition line we found directed percolation criticality.

Phase transition of the one-dimensional coagulation-production process.

Recently an exact solution has been found by M. Henkel and H. Hinrichsen [J. Phys. A 34, 1561 (2001)] for the one-dimensional coagulation-production process: 2A-->A, AØA-->3A with equal diffusion and coagulation rates. This model evolves into the inactive phase independently of the production rate with t(-1/2) density decay law. This paper shows that cluster mean-field approximations and Monte Carlo simulations predict a continuous phase transition for higher diffusion/coagulation rates as considered by the exact solution. Numerical evidence is given that the phase transition universality agrees with that of the annihilation-fission model with low diffusions.

Critical behavior of the one-dimensional annihilation-fission process 2A-->O, 2A-->3A

Numerical simulations and cluster mean-field approximations with coherent anomaly extrapolation show that the critical line of the 1d annihilation-fission process is separated into two regions. In both the small and high diffusion cases, the critical behavior is different from the well known universality classes of nonequilibrium phase transitions to absorbing states. The high diffusion region seems to be well described by the cyclically coupled directed percolation and annihilating random walk. Spreading exponents show nonuniversal behavior.

Skeletonization, Fluctuating Mean-Field Approximations and Coherent Anomalies in Critical Phenomena

Concepts of skeleton order parameters and their conjugate symmetry breaking skeleton fields are introduced in order to describe the phase transition of the relevant system. It is shown that the skeletonization of Hamiltonians is very useful in understanding the essence of phase transitions. Singularities of skeleton susceptibilities are studied in general. There are given some examples in spin systems. Fluctuating mean-field approximations are introduced to study phase transitions qualitatively, namely to find more accurate values of critical points even in frustrated spin systems. This idea gives the basis of the coherent anomaly method (CAM). This new method is explained briefly to obtain non-classical critical exponents in cluster mean-field approximations on the basis of the concept of coherent anomalies.