More is different: How chemical complexity influences stability in high entropy oxides

Abstract

Tailoring the chemical composition of a high entropy oxide (HEO) is a powerful approach to enhancing desirable material properties. However, the targeted synthesis of HEO materials is often hindered by competing stabilizing and destabilizing factors, which are difficult to predict. This work examines the effects of increased configurational entropy on the phase formation and stability of four notable complex oxide families: perovskite ($AB$O$_3$), pyrochlore ($A_2B_2$O$_7$), Ruddlesden-Popper ($A_2B$O$_4$), and zirconium tungstate ($AB_2$O$_8$). Each of these structures has a tetravalent cation site, which we attempt to substitute with an entropic mixture of four cations, benchmarked by the parallel synthesis of a non-disordered reference compound. While all four target high entropy materials can be expected to form based on ionic radii criteria, only the high entropy perovskite Ba(Ti,Zr,Hf,Sn)O$_3$ is successfully synthesized. In the case of the pyrochlore, an entropy-stabilized defect fluorite is formed instead, while the Ruddlesden-Popper phase co-exists with multiple competing phases. For the tungstate, an unexpected deep eutectic point between the precursors results in melting that precedes the formation of a high entropy phase. Our case studies therefore illustrate that the stability of HEOs cannot be straightforwardly predicted based on ionic radii, lattice geometry, and charge-balancing considerations alone due to the underlying complexity of the interactions between the many chemical constituents.

Figures (4)

• Figure 1: Successful synthesis of a high entropy perovskite.(a) The reference cubic perovskite BaSnO$_3$ with space group $Pm\overline{3}m$ was successfully prepared as a single-phase material via solid state synthesis, as shown by (b) Rietveld refinement of the synchrotron powder XRD pattern ($R_{\text{wp}} = 9.10\%$). (c) The high entropy perovskite, Ba(Ti,Zr,Hf,Sn)O3 was also successfully synthesized as confirmed by the (d) Rietveld refinement of its synchrotron powder XRD pattern in the same $Pm\overline{3}m$ structure ($R_{\text{wp}} = 4.61\%$). (e) Comparison of the peak shape for the (310) Bragg peak in BaSnO$_3$ and Ba(Ti,Zr,Hf,Sn)O3for data collected on a laboratory diffractometer with a Cu $K_{\alpha}$ source. The two patterns have been normalized according to the highest intensity (110) Bragg peak. Disorder effects and strain result in a suppression of the total intensity and peak broadening in the case of the high entropy analog. (f) Thermal stability of Ba(Ti,Zr,Hf,Sn)O3 showing the formation of weak impurity peaks when slow cooled or annealed below the synthesis temperature. Minimal impurities are observed at 1100° C, indicating kinetic stability at that temperature.
• Figure 2: Formation of an entropy-selected defect fluorite pyrochlore phase.(a,b) Crystal structure of the reference pyrochlore compound Y$_2$Sn$_2$O$_7$, where Y and Sn occupy distorted cubic and octahedral oxygen environments, respectively. (c) Rietveld refinement of the synchrotron powder XRD pattern for Y$_2$Sn$_2$O$_7$ confirming the formation of a pyrochlore phase in the cubic $Fd\overline{3}m$ space group ($R_{\rm{wp}} = 7.34\%$). The inset shows the (111) and (311) superlattice peaks at 1 and 2 Å$^{-1}$, respectively. (d,e) Crystal structure of the defect fluorite-structured high entropy phase Y$_2$(Ti,Zr,Hf,Sn)$_2$O$_7$, where all the cations share a single crystallographic site with cubic coordination to a partially vacant oxygen sublattice. (f) Rietveld refinement of the synchrotron powder XRD pattern for Y$_2$(Ti,Zr,Hf,Sn)$_2$O$_7$ in the cubic $Fd\overline{3}m$ space group ($R_{\rm{wp}}=12.01\%$). The superlattice reflections, as shown in the inset, are almost fully suppressed, indicating substantial site mixing that more closely resembles a defect fluorite structure.
• Figure 3: Instability of a high entropy Ruddlesden-Popper phase.(a) Crystal structure of Sr$_2$SnO$_4$, which is an orthorhombic Ruddlesden-Popper phase with space group $Pccn$. (b) Rietveld refinement of room temperature laboratory powder XRD data showing the successful solid state synthesis of phase pure Sr$_2$SnO$_4$ ($\chi^2 = 1.71$). (c) XRD patterns for the attempted synthesis of Sr2(Ti,Zr,Hf,Sn)O4as a function of reaction temperature showing a large $A_3B_2$O$_7$ impurity that persists to the highest reaction temperatures.
• Figure 4: Unexpected eutectic point preempts the formation of a high entropy tungstate.(a) Crystal structure of ZrW2O8, which is a cubic phase with space group $P2_13$. (b) Photo of resultant ZrW2O8 sample, showing a uniform pellet which has maintained its shape throughout annealing. (c) Rietveld refinement of powder XRD data showing the successful solid state synthesis of phase pure ZrW2O8 ($\chi^2 = 1.60$). (d) Photo of synthesis attempt of (Ti,Zr,Hf,Sn)W2O8 showing the adverse reaction outcome and evidence of melting occurring. (e) Results of DSC-TGA measurement on (Ti,Zr,Hf,Sn)W2O8 precursors, showing a melting phase transition at approximately 970° C before the target phase can form.