We study heat transport in a one-dimensional inhomogeneous quantum spin-1/2 system. It consists of a finite-size
XX spin chain coupled at its ends to semi-infinite
XX and
XY chains at different temperatures, which play the role of heat and spin reservoirs. After using the Jordan-Wigner transformation we map the original spin Hamiltonian into a fermionic Hamiltonian, which contains normal and pairing terms. We find the expressions for the heat currents and solve the problem with a nonequilibrium Green's-function formalism. We analyze the behavior of the heat currents as functions of the model parameters. When finite magnetic fields are applied at the two reservoirs, the system exhibits rectifying effects in the heat flow.
We study the thermal transport through a quantum spin-[Formula: see text] heterojunction, which consists of a finite-size chain with two-site anisotropic XY interaction and three-site XZX+YZY interaction coupled at its ends to two semi-infinite isotropic XY chains. By performing the Jordan–Wigner transformation, the original spin Hamiltonian is mapped onto a fermionic Hamiltonian. Then, the fermionic structure is discussed, and the heat current as a function of structural parameters is evaluated. It is found that the magnetic fields applied at respective chains play different roles in adjusting the heat current in this heterojunction. Moreover, the interplay between the anisotropy of the XY interaction and the three-site spin interaction assists to further control the thermal transport. In view of the numerical results, we propose this heterojunction to be an alternate candidate for manipulating the heat current in one-dimensional (1D) systems.
We have proposed a cluster heat bath(CHB) method of a quasi-one dimensional Ising model and demonstrated that it reproduces the exact results of the two dimensional Ising model with axially anisotropic exchange interactions. The point of the method is to select one of the equilibrium spin configurations of each of the chains which are subjected by effective fields of surrounding spins. Thus the chains are always in their equilibrium state during the simulation and the
d dimensional model effectively becomes as a
d−1 dimensional model of fictitious spins with their freedom of
2N and the simulation time is drastically reduced, where
N is the number of the spins on the chain.