Incompletely ordered phases and phase transitions in the three-dimensional general clock model

Abstract

In our previous studies of the antiferromagnetic Potts and Q =6 clock models in three dimensions, we have argued the existence of an incompletely ordered phase (IOP) which is characterized by soft rigidity with a nonintegral stiffness exponent and an incomplete order (though long ranged) such that some of the spin states are exclusively dominant. To investigate the IOP s further, we study Q-state general clock models with one varying energy parameter which can include the ferromagnetic Potts models, by means of our Monte Carlo twist method and analytically in the pair approximation. The twist method gives the following results for Q =6. There are two kinds of IOP's (IOP1 and IOP2) on the opposite side of the ferromagnetic Potts model. Their stiffness exponents are about 1.2 less than 2 (for the highest rigidity) in good agreement with those previously obtained. A pair of adjacent clock states are dominantly well mixed in the IOP1, as are a set of three adjacent ones in the IOP2. Thermal fluctuations in the IOP s can be characterized by the spin configurations of the soft structures with buffers that prevent spins in different states by more than the least angle from getting with each other, which is consistent with the nonintegral stiffness exponent. Large entropy contributions to the IOP s are revealed in both approaches. The phase boundary, which extends from the ferromagnetic Potts point, is clearly of first order, while among those relevant to the IOP's one is discontinuous and the other is continuous but the last two are not evident. It is strongly suggested that there is a transition without symmetry breaking between the IOP2 and the rigid phase. In the pair approximation applied for Q =4-12, various properties for Q =6 are consistent with those obtained by the simulation, except some other properties. Peculiar Q dependence of the highest critical temperature of the IOP in the extreme case is also found.

Ordered phase and phase transitions in the three-dimensional generalized six-state clock model

“…The existence of intermediate phase have attracted interest for the following models: three-dimensional three-state antiferromagnetic Potts model (3D-3AFP model) [5], and stacked triangular antiferromagnetic Ising model (STAFI model) [6]. Although, the existence of various type of intermediate phases for these models have been proposed, there is no general understanding on the nature of the intermediate phase(s) [3][4][5][6][7][8][9][10][11][12]. If we consider only one universality class for models with Z 6 symmetry, we conclude that the intermediate phase is an apparent phase as we found in the regular six state model.…”

“…However, it has been clarified that the situation is not so simple and interesting new type of phase exists in the generalized case. In fact, the 3D-6GCL model was introduced in order to understand general properties of other Z 6 models [4]. In the present paper, we study the 3D-6GCL model with a set of the parameters which is considered to correspond to the STAFI model and the 3D-3AFP model.…”

“…Ueno et al have introduced three-dimensional generalized six-state clock model (3D-6GCL model) [4]. The 3D-6GCL model can be regarded as prototypes of wide range of Z 6 models.…”

“…In this study, we choose the energy parameters of the 3D-6GCL model as ε 0 = 0, ε 1 = 0.1 and ε 2 = ε 3 = 1.0 which are the same as those in [4].…”

“…Since the COP has six degenerate states, the order parameter space of this model is illustrated by a hexagon as in Ueno rigorously proved that for ε 1 = 0, at least one kind of IOPs can exist at T < ε 2 = ε 3 [13]. For ε 1 > 0 , Ueno and Kasono proposed that this model has two incompletely ordered phases (IOP1, IOP2) besides the COP by using the Monte Carlo twist method (MCTM) [4]. In this paper, we reexamine the nature of the intermediate phase(s) in detail.…”