Computational assessment, against criteria specified in advance, of a Na+/Ag+ ion-exchange route to Ag2MgSiO4, a predicted wurtzite-derived silver silicate

Abstract

Many metastable inorganic compounds exist only because a synthetic route delivers them, and database screening is poorly placed to find them. Topotactic ion exchange is such a route: at low temperature it swaps the mobile cation of a known parent for another, ideally while the framework stays intact, yet it is rarely assessed computationally before synthesis. We test whether a simple exchange-energy criterion, specified in advance and checked against independent calculations, can nominate a credible target. Computed with a machine-learned interatomic potential, the criterion reproduced the sign of 42 published DFT exchange enthalpies (36/36 for |∆H| ≥ 10 kJ mol−1) and selected Ag2MgSiO4, to be made from Na2MgSiO4 in molten AgNO3. Density functional theory gives an exchange energy of about −58 kJ mol−1 per Ag, within the ±10 kJ mol−1 window around three reported preparations of silver silicates and germanates from sodium parents in molten AgNO3, and an energy above the hull (0.038–0.042 eV atom−1) like that of known silver silicates. The Pna21 structure is dynamically stable in the harmonic approximation (0 K). Calculated Ag+ hop barriers are comparable to or lower than those of a reported Ag+ conductor, and the HSE06 gap, calibrated on one literature value, is about 3 eV, at the violet edge of the visible range. These are 0 K predictions; no synthesis was attempted, and the closest precedent is Ag2MnSiO4, so a single exchange experiment can confirm or refute them (incomplete exchange or changes in particle size and shape are reported in some precedents).