Magnetic susceptibility and low-temperature structure of the linear chain cuprateSr2CuO3

Abstract

Magnetic susceptibility measurements for ${\mathrm{Sr}}_2$ ${\mathrm{CuO}}_{3\pm \mathrm{\delta }}$ were made from 2 to 800 K, and a strong dependence upon oxygen content (δ) was observed. Samples synthesized under oxygen, followed by various nitrogen treatments, exhibited markedly different Curie-Weiss-type terms, and we discuss possible origins for this behavior. High-temperature magnetic susceptibility measurements for the sample with the smallest Curie-Weiss-type term clearly show the increase with temperature expected from the Bonner-Fisher model for a spin-1/2 one-dimensional (1D) Heisenberg antiferromagnet. This is a direct experimental observation of 1D magnetic behavior in this system. The in-chain superexchange coupling constant, as determined by a fit to the Bonner-Fisher model, is ‖J‖/${\mathit{k}}_{\mathit{B}}$≊${1300}_{\mathrm{-}200}^{+100\mathrm{}}$ K, comparable to the values observed in the two-dimensional layered cuprates. Estimates of the interchain magnetic interaction indicate this material may be the best realization of a 1D spin-1/2 Heisenberg antiferromagnet reported to date. Low-temperature neutron and synchrotron x-ray powder-diffraction studies of ${\mathrm{Sr}}_2$ ${\mathrm{CuO}}_3$ show that the low-temperature structure of this system has Immm space-group symmetry, the same structure reported at room temperature, indicating that this material, in contrast to ${\mathrm{La}}_2$ ${\mathrm{CuO}}_4$, does not undergo any structural transformations upon cooling. The absence of crystallographic distortions precludes a magnetic anisotropy contribution from a Dzyaloshinsky-Moriya interaction, implying that ${\mathrm{Sr}}_2$ ${\mathrm{CuO}}_3$ should be a nearly ideal spin-1/2 antiferromagnetic Heisenberg chain compound, in agreement with the magnetic susceptibility results. A search for the presence of long-range three-dimensional antiferromagnetic order by magnetic neutron powder diffraction at temperatures as low as 1.5 K was not successful, although we estimate an upper limit for the size of the ordered moment which could have been detected to be ∼0.1${\mathrm{\mu }}_{\mathit{B}}$ per ${\mathrm{Cu}}^{2+}$ ion.