Ground-state characterizations of systems predicted to exhibitL11orL13crystal structures

Abstract

Despite their geometric simplicity, the crystal structures L11 (CuPt) and L13 (CdPt3) do not appear as ground states experimentally, except in Cu-Pt. We investigate the possibility that these phases are ground states in other binary intermetallic systems, but overlooked experimentally. Via the synergy between high throughput and cluster expansion computational methods, we conduct a thorough search for systems that may exhibit these phases and calculate order-disorder transition temperatures when they are predicted. High throughput calculations predict L11 ground states in the following systems: Ag-Pd, Ag-Pt, Cu-Pt, Pd-Pt, Li-Pd, Li-Pt, and L13 ground states in the following systems: Cd-Pt, Cu-Pt, Pd-Pt, Li-Pd, Li-Pt. Cluster expansions confirms the appearance of these ground states in some cases. In the other cases, CE predicts unsuspected derivative superstructures as ground states. The order-disorder transition temperatures for all L11/L13 ground states were found to be sufficiently high that their physical manifestation may be possible.

Surface Segregation and Chemical Ordering Patterns of Ag–Pd Nanoalloys: Energetic Factors, Nanoscale Effects, and Catalytic Implication

“…Fig. 1 shows the heats of formation ∆H for homogeneous solid solutions and imaginary intermetallics AgPd 3 (L1 2 or D0 22 ), AgPd (L1 0 or [64][65][66][67][68][69][70] It was theoretically suggested that stable intermetallics might exist at low temperature, in contrast to experimental phase diagram. For example, DFT calculations indicated that D0 23 , D0 24 , L1 2 (all for Ag 3 Pd), L1 1 (for AgPd), and L1 3 (for AgPd 3 ) are the possible ground states but may be overlooked experimentally as the low order-disorder transition temperatures.…”

Automated Computation of Materials Properties

“…This approach has also been used to calculate the solubility of elements in titanium alloys [71], to study the effect of hydrogen on phase separation in iron-vanadium [72], and to find new superhard tungsten nitride compounds [73]. The data has been employed to generate structure maps for hcp metals [74], as well as to search for new stable compounds with the Pt 8 Ti phase [75], and with the L1 1 and L1 3 crystal structures [76]. Note that even if a structure does not lie on the ground state convex hull, this does not rule out its existence.…”

Revisiting the revised Ag-Pt phase diagram

“…Based on the early work of Schneider and Esch [35] and subsequent refinements, the phase diagram of DF is surprising. It eliminates two of the intermediate compounds (at 45% and 75% platinum), and the 50%-Pt phase, computationallypredicted [3,27,38,33,28,4] in several works to be L1 1 (isostructural to Cu 1 Pt 1 ), is replaced by a compound at 53%. Based on x-ray diffraction (XRD) and EPMA data, the DF authors speculate that this compound has a concentration of 53%-Pt (atomic %), with a cubic cell with at least 32 atoms and a stoichiometry of 15:17.…”

“…Several computational studies predict [27,38,33,28,4] the L1 1 structure (isostructural to Cu 1 Pt 1 ), a 50%-Pt compound, in the middle of the phase diagram instead of DF's 53%-Pt structure. Since ab initio calculations are limited to T = 0 K predictions and the DF measurements were made at finite temperatures, one is left to wonder whether the discrepancy is genuine or merely due to finite temperature effects.…”