Solid state reduction of nickelate thin films

The square-planar nickelates are a class of superconductors analogous to the cuprates and promise to provide insight into the pairing mechanism in high-temperature superconducting oxides. The parent phase of doped superconducting films is ${\mathrm{NdNiO}}_2$, which is prepared by reducing $3^{+}$ Ni in ${\mathrm{NdNiO}}_3$ films. In this paper, we develop an ultrahigh vacuum reduction method using aluminum deposited on top of the $3^{+}$ nickelates and monitor the reduction process in real time through in situ crystal truncation rod measurements and diffraction x-ray absorption near edge spectroscopy measurements across the Ni $K$ edge. We establish a relation between Ni valence and the lattice constant of $\mathrm{NdNi}{\mathrm{O}}_{3-x}$ and show that the process can precisely control the oxygen content in the films. Finally, we extend the reduction process to quintuple square-planar nickelates.

Figure 1

In situ characterization during reduction. (a) Experimental diffraction geometry and in situ reduction reaction. The reaction shown is exothermic with a formation energy of 2.7 eV per oxygen atom removed from ${\mathrm{NdNiO}}_3$. (b) and (c) Crystal truncation rod (CTR) and diffraction x-ray absorption near edge spectroscopy (dXANES) taken during a shutter-controlled reduction. (b) shows a two-dimensional plot of 37 CTR ($00L$) scans on a log scale on the high-temperature, shutter-controlled (HTS) reduced sample. Each slice in the bright zone is a CTR scan, with the yellow dots locating the nickelate Bragg peak. Aluminum deposition is started and stopped after 1 monolayer (ML) is deposited seven times during the HTS process. The white number indicates the total Al thickness in MLs. The nickelate Bragg position shifts from $L=1.07$ ($c=3.64\AA$) to $L=1.185$ ($c=3.30\AA$) after reduction. In each dark blue region of the plot, an energy scan of the scattered intensity is done at the nickelate Bragg position determined from the previous CTR scan. A strong anomalous scattering signal is observed as shown in (c). The labels in (c) correspond to those shown in (b) and indicate the total Al thickness in MLs.

Figure 2

Scripts describing sample preparation and characterization for the low-temperature continuous (LTC), high-temperature continuous (HTC), and high-temperature, shutter-controlled (HTS) processes. The number in each decision panel indicates the typical number of iterations before the next step.

Figure 3

Experimental characterization before and after reduction. (a) Reflection high-energy electron diffraction (RHEED) pattern of the ${\mathrm{NdNiO}}_3$ thin film before reduction. (b)–(e) Characterization of $\mathrm{NdNi}{\mathrm{O}}_{3-x}$ thin films before, during, and after reduction. (b) The ($00L$) crystal truncation rod (CTR) scans for ${\mathrm{NdNiO}}_3$ and ${\mathrm{NdNiO}}_2$ thin films. The $x$ axis is in units of STO reciprocal lattice units (r.l.u.). (c) Resistivity of 15 u.c. ${\mathrm{NdNiO}}_2$ and ${\mathrm{NdNiO}}_3$ thin films. Note that, for ${\mathrm{NdNiO}}_2$, the resistivity measured overlaps while heating and cooling. (d) Off-specular (103) reciprocal space map (RSM) of a ${\mathrm{NdNiO}}_2$ film. (e) Surface topography measured using atomic force microscopy (AFM) of an ${\mathrm{Al}}_2{\mathrm{O}}_3$-capped ${\mathrm{NdNiO}}_2$ thin film.

Figure 4

Comparison of three reduction conditions. (a) follows the change of the nickelate lattice constant and thickness as a function of time and Al coverage. The light gray lines indicate when Al is being deposited during high-temperature, shutter-controlled (HTS) reduction. The film thickness is calculated by the full width at half maximum (FWHM) of diffraction peaks using the Scherrer equation (see Supplemental Material Fig. S4 [16] for details). (b) Average Ni valence $\overline{V}$ as a function of ${\mathrm{NdNiO}}_2$ lattice constant. The red curve is from a linear fitting, and the patched region indicates the 95% confidence interval. (c) plots the change in rate of the average Ni valence during reduction for the three processes.

Figure 5

Physical and electronic structure during reduction. (a) shows fits of Ni valence and $c$-axis lattice constant vs Al coverage ML($n$). The fitting excludes the $n=7$ point because the film is completely reduced when $n=6$, judging from Ni valence and measured lattice parameter. (b) and (c) show diffraction x-ray absorption near edge spectroscopy (dXANES) and crystal truncation rod (CTR) data from the 15 u.c. $\mathrm{NdNi}{\mathrm{O}}_{3-x}$ thin films before and during reduction. All data are taken at $T=270{}^{\circ }\mathrm{C}$. The data for a ${\mathrm{NdNiO}}_3$ standard is taken prior to high-temperature continuous (HTC) reduction, and the rest are measured during the high-temperature, shutter-controlled (HTS) process. The labels indicate Ni valence determined by principal component analysis (PCA) of dXANES measurements. (b) Imaginary part of the Ni anomalous scattering factor $f_{\mathrm{Ni}}^{\prime \prime }$ across the $K$ edge, taken at the corresponding diffraction peak marked in (c). The white line energy is marked in each curve, and the dashed lines are guides for the eye marking the energy of a pre-edge feature that grows in intensity during reduction. (c) shows CTR scans along the ($00L$) direction for 15 u.c. $\mathrm{NdNi}{\mathrm{O}}_{3-x}$ thin films before and during reduction. Note that the initial lattice parameter in HTS reduction is 3.64 Å, deviating from strained ${\mathrm{NdNiO}}_3, c=3.76\AA$.

Figure 6

Microstructure of reduction process. (a) Atomic-resolution high-angle annular dark-field (HAADF) image showing the cross-section of the ${\mathrm{NdNiO}}_2$ film. The crystallographic orientations marked on the image correspond to the cubic ${\mathrm{SrTiO}}_3$ structure. (b) Atomic-resolution HAADF image [same as shown in (a)] with labeled Nd and Sr atomic columns. (c) The corresponding average bond distances along [001] and [010]. The dotted red line marks the interface. The error bars correspond to the standard deviation across each atomic plane.

Figure 7

(a) X-ray diffraction (XRD) of ${\mathrm{Nd}}_6{\mathrm{Ni}}_5{\mathrm{O}}_{16}$ (top) and ${\mathrm{Nd}}_6{\mathrm{Ni}}_5{\mathrm{O}}_{12}$ (bottom) thin films. The superlattice diffraction peaks are labeled. (b) Resistivity of 3 s.l.u.c. ${\mathrm{Nd}}_6{\mathrm{Ni}}_5{\mathrm{O}}_{16}$ and 3 s.l.u.c. ${\mathrm{Nd}}_6{\mathrm{Ni}}_5{\mathrm{O}}_{12}$ thin films.