Exponential decay of equal-time four-point correlation functions in the Hubbard model on the copper-oxide lattice (Q398722)

The paper proves a theorem that equal-time 4-point correlation functions in the three-band Hubbard model on the copper-oxide lattice (or the CuO Hubbard model) at positive temperature decay exponentially in the thermodynamic limit if the coupling constants on both the cooper and the oxygen sites are smaller than some power of the temperature. A consequence of the theorem is the exponential decay of pairing-pairing correlation functions in the distance between the center of two-electrons and that of two holes, excluding long-range correlations between singlet Cooper pairs at high temperatures or in long-temperature weak-coupling regimes. The analiticity of the covariance enables the author to reformulate the correlation function multiplied by the distance between the electrons and the holes into a multi-contour integral of the correlation function with respect to new complex variables inserted in the covariance. The practical role of the multi-scale integration over the Matsubara frequency is to establish a volume-independent upper bound on the perturbed correlation function inside the multi-contour integral. Due to a self-contained nature of the multi-scale Matsubara expansion, the presented proofs merely rely on the repeated use of the free formula for logarithm of the Grassmann integral. The author deals with the Grassmann integral formulation, which is flexible to mathematical manipulations, as the rigorous counterpart of the correlation function. By sending the finite-dimensional formulation to the limit, the conclusion on the original correlation function defined by trace operations over the Fermionic Fock space is withdrawn. The paper provides an alternative method to solve the Matsubara ultraviolet problem. Though the paper involves a multi-scale analysis concerning the Matsubara sum as the main technical ingredient, it does not treat any infrared multi-scale analysis around zero points of the dispersion relation. The paper has no improvement on the temperature dependence of the allowed magnitude of the interaction over the single-scale analysis and cannot study the behavior of correlation functions at zero temperature. It is straightforward to adapt the proofs presented in this paper to conclude the same result for the Hubbard model on the honeycomb lattice as claimed for the CuO Hubbard model.