Antiferromagnetic domain walls in lightly doped layered cuprates

Abstract

Recent ESR data shows rotation of the antiferromagnetic (AF) easy axis in lightly doped layered cuprates upon lowering the temperature. We account for the ESR data and show that it has significant implications on spin and charge ordering according to the following scenario: In the high temperature phase AF domain walls coincide with (110) twin boundaries of an orthorhombic phase. A magnetic field leads to annihilation of neighboring domain walls resulting in antiphase boundaries. The latter are spin carriers, form ferromagnetic lines and may become charged in the doped system. However, hole ordering at low temperatures favors the (100) orientation, inducing a π/4 rotation in the AF easy axis. The latter phase has twin boundaries and AF domain walls in (100) planes. The ESR data 1 also shows that AF domains are present and that their relative intensity is controlled by magnetic fields. In the high temperature phase with field in the [100] direction, the [010] polarized domains are preferred since the canting of AF spins has a higher susceptibility. This defines a "depinning" field at which the unpreferred domains are diminished. E.g. at high temperatures and above a field of ≈ 1T 80% of the domains are polarized perpendicular to the field. In the present work we suggest that the transition in the AF polarization correlates with a lattice distortion and with condensation of the holes into a charge density wave (CDW). Our reasoning involves a scenario for spin and charge ordering with the following steps: (i) At high temperatures and weak magnetic fields the [100] polarized AF leads to an orthorhombic structure; the AF domain walls (DW's) coincide then with twin boundaries at (110) planes. (ii) Strong magnetic fields favor one of the AF domains so that DW's annihilate. We show that this process results in antiphase boundaries (APB's) which consist of ferromagnetic lines. (iii) At low temperatures hole ordering can occur by binding states localized at the APB's. However, hole ordering favors the (100) orientation, i.e. a CDW with maxima on (100) planes and an (incommensurate) wavevector in the [100] direction. This CDW favors an AF polarization in a diagonal [110] direction which is weakly orthorhombic with (100) twin boundaries and DW's.An AF polarized in the [100] axis favors an orthorhombic phase since dipoles interact differently for → ← and ↑ ↓ pairs. This magnetoelastic coupling is expected to be rather small in the cuprates, leading to a change in lattice constant of order 3 10 −5 −10 −6 . In principle ESR analysis can detect the resulting change in crystal field 4 , however, the effect seems too weak in the case AF Y Ba 2 Cu 3 O 6 . Yet, this coupling can be sufficiently strong in samples of size ≈ 1mm so that twin boundaries are necessary to relieve macroscopic strains.We proceed to present evidence for steps (i) , (iii) and then derive step (ii). We claim that an orthorhombic structure is in fact supported by the mere observation of AF domains. The presence of AF domains and the associated domain walls 5...

Elastic Domains in Antiferromagnets on Substrates

“…This argumentation is close in spirit to arguments of Ref. [6] (see also [7][8][9]). The elastic interaction between domains and point defects was discussed in Ref.…”

“…Finally, we note that the paper by Horovitz [8] deals with the transition in the antiferromagnetic polarization due to a microscopic mechanism: condensation of the holes into a charge-density wave. In our description, all microscopic details, leading to a specific antiferromagnetic polarization without magnetic field, are hidden in a few parameters (anisotropy and constants of magnetoelastic interaction ).…”

“…j . The possibility of such plain defects appearing in antiferromagnets was considered by Horovitz [8]. If the field is oriented along the x direction, the behavior of the domain walls in Figs.…”