Bounds on strong field magneto-transport in three-dimensional composites

TL;DR

This paper establishes bounds on the effective non-symmetric conductivity of 3D composites under strong magnetic fields, revealing limitations and behaviors of antisymmetric parts and deriving new Hashin-Shtrikman bounds.

Contribution

It introduces novel bounds for non-symmetric conductivities in 3D composites, including a variational principle and Hashin-Shtrikman type bounds for microstructures.

Findings

Antisymmetric part of effective conductivity can be arbitrarily large in general composites.

Antisymmetric part of effective conductivity tends to zero as local antisymmetric bounds decrease.

Elementary bounds are derived for transversely isotropic conductivities.

Abstract

This paper deals with bounds satisfied by the effective non-symmetric conductivity of three-dimensional composites in the presence of a strong magnetic field. On the one hand, it is shown that for general composites the antisymmetric part of the effective conductivity cannot be bounded solely in terms of the antisymmetric part of the local conductivity, contrary to the columnar case. So, a suitable rank-two laminate the conductivity of which has a bounded antisymmetric part together with a high-contrast symmetric part, may generate an arbitrarily large antisymmetric part of the effective conductivity. On the other hand, bounds are provided which show that the antisymmetric part of the effective conductivity must go to zero if the upper bound on the antisymmetric part of the local conductivity goes to zero, and the symmetric part of the local conductivity remains bounded below and above. Elementary bounds on the effective moduli are derived assuming the local conductivity and effective conductivity have transverse isotropy in the plane orthogonal to the magnetic field. New Hashin-Shtrikman type bounds for two-phase three-dimensional composites with a non-symmetric conductivity are provided under geometric isotropy of the microstructure. The derivation of the bounds is based on a particular variational principle symmetrizing the problem, and the use of Y-tensors involving the averages of the fields in each phase.