Electronic superlattice revealed by resonant scattering from random impurities in Sr3Ru2O7

Abstract

Resonant elastic x-ray scattering (REXS) is an exquisite element-sensitive tool for the study of subtle charge, orbital, and spin superlattice orders driven by the valence electrons, which therefore escape detection in conventional x-ray diffraction (XRD). Although the power of REXS has been demonstrated by numerous studies of complex oxides performed in the soft x-ray regime, the cross section and photon wavelength of the material-specific elemental absorption edges ultimately set the limit to the smallest superlattice amplitude and periodicity one can probe. Here we show--with simulations and REXS on Mn-substituted Sr3Ru2O7--that these limitations can be overcome by performing resonant scattering experiments at the absorption edge of a suitably-chosen, dilute impurity. This establishes that--in analogy with impurity-based methods used in electron-spin-resonance, nuclear-magnetic resonance, and Mössbauer spectroscopy--randomly distributed impurities can serve as a non-invasive, but now momentum-dependent probe, greatly extending the applicability of resonant x-ray scattering techniques.

Figure 1. Scattering from Sr₃(Ru_1−xMn_x)₂O₇ along the [110] plane; one can see the Mn-substituted RuO₂ bilayers, separated by SrO planes.

(a) When the x-ray energy is tuned to the Ru L₃ edge (2.96 KeV), the scattering signal is determined by the Ru atoms (green); (b) on the converse, at the Mn L₃ edge (641 eV), only the Mn atoms (red), occupying ~10% of the sites, contribute to the scattering signal.

Figure 2. Ru resonance REXS results from 10% Mn substituted Sr₃(Ru_1−xMn_x)₂O₇ (x = 0.1).

(a) Energy profile of the () peak across the Ru L_2,3 edge at T = 20 K. (b) Temperature dependence of the integrated intensity of the () peak at the Ru L₂ edge; the energy position (2968 eV) of the measurement is shown by the arrow in panel (a).

Figure 3

(a) Mn resonance profile for the () superlattice diffraction peak measured at 20 K on Sr₃(Ru_1−xMn_x)₂O₇ with x = 0.1 (the arrow at 641 eV indicates the energy used in the Mn-edge REXS experiments). Inset: full reciprocal space map of the superlattice peak. (b) Temperature dependence of the Mn L₃ edge () momentum scans for x = 0.1. (c) Temperature dependence of the integrated intensity of the () peak for 5% and 10% Mn-substitution.

Figure 4

(a) Square lattice (40 × 40 sites) with an up-up-down-down zig-zag spin order (red/up, blue/down). (b) Randomly selected 5% lattice sites from panel (a), with the same spin correlations as in (a). (c) Same 5% lattice sites as in (b), but now with random spin orientation for the selected sites. (d,e,f) Reciprocal space map of the diffraction peaks.

Figure 5

Temperature dependence of the (a) normalized integrated intensity and (b) correlation length of the () superlattice order, measured at the Mn (red) and Ru (green) L-edges for 10% Mn-substitution.

Figure 6

(a) Zig-zag antiferromagnetic islands (red/spin-up, blue/spin-down) around Mn impurities, just below the metal-insulator transition temperature T_MIT, i.e. the percolation threshold before the onset of long-range order – the islands are not correlated with one-another yet. (b) Randomly selected 5% lattice sites from panel (a), representing the 5% Mn impurities, with the same spin correlations. (e,f) Reciprocal space map of the diffraction peaks for Ru and Mn edges, calculated from the corresponding top panels. (c,d,g,h) Same as (a,b,e,f), now below the long-range antiferromagnetic ordering temperature T_order, with partial magnetic correlations having been established between different antiferromagnetic islands.