Rare-earth nitride solid solutions

Abstract

<p>Improvements to computing efficiency are required to offset the energy costs of increasing demand on cloud computing data centres (the cost of which is currently borne by fossil fuel energy production, and thus, the environment) [4]. The 20th-century technologies that currently support our heavy use of computers are unable to keep providing efficiency improvements due to the fundamental limits associated with the physics that they rely on to operate, meaning that new physical mechanisms are required [5–8]. Improvements to computing efficiency are offered by superconducting computing, but the implementation of this technology requires cryogenically-viable memory, which is not currently available [9, 10]. The research field of spintronics promises devices with plausible improvements over current memory technologies, but these devices require materials in which complex properties can be controlled [11, 12]. The rare-earth nitrides are a series of intrinsic ferromagnetic semiconductors, that have been included in device architectures with some success [13–15]. Their combination in a solid solution offers control over the magnetic and electrical properties of a homogenous single-layer thin film, allowing the fine-tuning of the properties to meet the requirements of application. Understanding the material properties of rare-earth nitride solid solutions is the central topic of this thesis.</p> <p><br></p> <p>This thesis documents research into the properties of Gd<sub><em>x</em></sub> Sm<sub><em>1</em></sub><sub><em>−x</em></sub> N, a solid solution of the rare-earth nitrides GdN and SmN. We present magnetometry, x-ray magnetic circular dichroism (XMCD), and Hall effect measurements on a series of Gd<sub><em>x</em></sub>Sm<sub><em>1−x</em></sub>N films. Magnetometry results show a linear dependence of the saturation magnetisation on the Gd fraction of the films, <em>x</em>, over much of the range 0 ≤ <em>x</em> ≤ 1. Further results show that Gd<sub><em>x</em></sub>Sm<sub><em>1−x</em></sub>N films with a low Gd fraction (<em>x</em> ≤ 0.3) exhibit perpendicular magnetic anisotropy. Films for which <em>x</em> ≤ 0.2 show a deviation from the linear dependence of the saturation magnetisation on <em>x</em>, and extremely low magnetic moments. This low moment results, in part, from the opposition of the net moment of the Gd<sup>3+</sup> and Sm<sup>3+</sup> ions. The opposition of the net moment comes from the orbital-dominant magnetism of the Sm<sup>3+</sup> ion and the alignment of the Gd<sup>3+</sup> and Sm<sup>3+</sup> spins by the exchange interaction (which is confirmed by XMCD).</p> <p><br></p> <p>This combination of the orbital-dominant magnetism of the Sm<sup>3+</sup> ion and the alignment of the Gd<sup>3+</sup> and Sm<sup>3+</sup> spins by the exchange interaction also causes there to be compositions of Gd<sub><em>x</em></sub>Sm<sub><em>1−x</em></sub>N for which the net magnetic moment or the net angular momentum of the solution is zero. These “compensation points” are familiar from other classes of material generally held to be promising for similar applications [16, 17]. Measurements of XMCD show an increasing spin polarisation of Sm<sup>3+</sup> as a function of <em>x</em>, changing where one expects to find the angular momentum compensation point <em>x</em><sub><em>a</em></sub>. The value of <em>x</em><sub><em>a</em></sub> is estimated with reference to the XMCD measurements to be <em>x</em><sub><em>a</em></sub> = 0.245. Furthermore, measurements of the anomalous Hall effect are used to locate the compensation point of the net magnetic moment, <em>x</em><sub><em>c</em></sub> , which is found to be <em>x</em><sub><em>c</em></sub> = 0.05.</p>