Page 27 - ILL Annual Report 2019

SCIENTIFIC HIGHLIGHTS
24-25
dependence. These magnetic moments have no temperature trend; however, as we increase the pressure their magnitude decreases and approaches zero at around 30–40 GPa (figure 1).
X-ray emission spectroscopy is a powerful technique for detecting local magnetic moments. However, it provides no information on how these individual magnetic moments globally order, i.e. whether they have long-range magnetic order. Therefore, to detect the global magnetic order in epsilon iron we turned to neutron diffraction.
Previous theoretical studies of epsilon iron predict an antiferromagnetic magnetic structure where the magnetic moments alternate in direction between neighbouring atoms [1]. Antiferromagnetic magnetic structures have a larger unit cell than that of the crystal structure. Therefore, magnetic reflections are expected to occur at smaller scattering angles (2θ) that are separate from the nuclear diffraction peaks.
We performed high-pressure low-temperature diffraction on the D20 diffractometer using a Paris–Edinburgh cell at truly extreme P/T conditions. The iron sample was pressurised past the structural bcc–hcp transition and then cooled down while collecting diffraction data. The diffraction pattern shown in figure 2 was measured at 20.2 GPa and 1.8 K, which is a record-setting high-pressure/low-temperature measurement [2]. Low temperature is critical in this study since long-range magnetic order can only occur well below room temperature.
Although the data contain strong diffraction peaks from the diamond anvils used in the Paris–Edinburgh cell, we were able to accurately refine the data as shown by the blue line in figure 2. There are no reflections evident at scattering angles below 30 degrees; therefore, a 30x zoom of this region is also shown in figure 2. It is important to note that all the diffraction peaks shown in the refinement in this region are due to the small (0.2 %) harmonic contamination from the monochromator in the
Figure 2
Neutron powder diffraction pattern of epsilon iron at 20.2 GPa/1.8 K performed with λ = 1.3-Å neutrons. The diffraction measurement (black line) is shown with its Rietveld refinement (blue line). The peaks due to epsilon iron and diamond are shown as magenta and green ticks respectively, where the lighter peaks are due to secondary reflections from λ/2 contamination. A 30x zoom
of the low-scattering angle region is shown to emphasise the lack of diffraction peaks due to long-range magnetic order.
high-flux configuration of D20. In this zoom, we can see that the refinement corresponds well with our data and that there are no reflections due to long-range magnetic order. Through simulations, we placed an upper limit
on the magnitude of the magnetic moment that was five times smaller than theoretical estimation [1]. Furthermore, the diffraction patterns showed no change from 1.8 K to 260 K. Theory estimates a critical temperature of around 75 K [3]. Therefore, if long-range magnetic order exists in epsilon iron we should have detected a change.
We detected local magnetic moments without global magnetic order. To reconcile these observations, as well as those of previous studies, we propose a new, spin-smectic, magnetic structure. We discovered this unique structure through extensive density functional theory and classical Monte Carlo simulations. As shown in figure 1, we predict that epsilon iron is paramagnetic but that if it is cooled below Tm it forms this spin-smectic state. The spin-smectic state is characterised by alternating magnetic and non- magnetic bilayers. The interaction between the magnetic bilayers is extremely weak, leading to very small interlayer correlations. Therefore, we predict an orientationally disordered state with no long-range magnetic order which would be undetectable by neutron diffraction, while the local magnetic moments are still detectable by X-ray emission spectroscopy.
Our measurements find local magnetic moments in epsilon iron at around 30–40 GPa without long-range order, which we suggest is because they form a spin-smectic state. Epsilon iron is known to superconduct at low temperatures from 15 to 32 GPa (Tc in figure 1). This pressure region coincides with where we observed a local magnetic moment. Therefore, we predict that magnetic fluctuations of the spin-smectic state play an important role in the superconductivity of epsilon iron. Future experiments to confirm the existence of the spin-smectic state in epsilon iron are planned.
www.ill.eu