Dynamical Structure, Bonding, and Thermodynamics of the Superionic Sublattice inα−AgI

Abstract

Using extensive first-principles molecular dynamics calculations, we characterize the superionic phase transition and the lattice and electronic structures of the archetypal Type-I superionic conductor α-AgI. We find that superionicity is signalled by a phase transition of the silver ions alone. In the superionic phase, the first silver shell surrounding an iodine displays a distinct dynamical structure that would escape a time-averaged characterization, and we capture this structure in a set of ordering rules. The electronic structure of the system demonstrates a unique chemical signature of the weakest-bound silver in the first shell, which in turn is most likely to diffuse. Upon melting, the silver diffusion decreases, pointing to an unusual entropic contribution to the stability of the superionic phase.PACS numbers: 71.15. Pd,66.30.Dn Key advances in energy research have prompted a surge of interest in superionic materials, as a crucial enabling technology for a variety of nanotechnological devices, including sensors, switches, batteries, and fuel cells. Of the superionics, AgI and related silver halides and sulfides have attracted particular attention because of the unusually high levels of ionic conductivity they exhibit, and as such are finding increased and varied technological implementation [1,2,3]. At normal pressure, AgI enters its superionic α phase above T c = 420 K, at which temperature a phase transition to a body-centered cubic structure is accompanied by an increase in the silver conductivity of nearly three orders of magnitude, to a value of 1.31. Previous molecular dynamics studies using classical pair potentials have successfully reproduced experimental characteristics of the α and β phases, as well as the α → β transition [5,6,7,8,9,10], but these are unable to describe the electronic structure in a dynamic environment, or to capture the phenomenology of the melting transition. In this regard, first-principles simulations provide unique and unbiased predictive power.We perform Car-Parrinello molecular dynamics simulations in the canonical NVT ensemble at temperatures ranging from 200 K to 1250 K for a total of 800 ps. All simulations were performed with a 54-atom unit cell and a 0 = 5.174Å, except the Wannier function calculations, which were performed in a 32-atom unit cell [21].First, we find evidence of a phase transition of the silver ions near the experimental T c that is independent of the conformation and dynamics of the iodine sublattice and signals the transition into the superionic α phase. The silvers exhibit a sharp decrease in their diffusion behavior upon cooling below T c , although cubic boundary conditions forbid the iodine structural transition to the hexagonal wurtzite β phase. Results from a series of simulations in which we immobilized the iodines in a fixed bcc configuration provide a secondary, stronger indicator of the independence of the silver transition from any iodine dynamics. Fig. 1 displays the associated silver ion diffusion coefficients D Ag for the fixed...

Unsupervised landmark analysis for jump detection in molecular dynamics simulations

“…Classical/empirical force fields were chosen in several studies [6][7][8][9][10][11][12][13] due to their computational efficiency and access to the time and length scales required to characterize ionic transport. Accurate, yet expensive, firstprinciples simulations have also been employed for selected systems [14][15][16][17][18][19][20][21][22][23][24]. The necessary compromise between the transferability of first-principles potential energy surfaces and the computational efficiency of force * These two authors contributed equally to this work.…”

Modeling lithium-ion solid-state electrolytes with a pinball model

“…This observation resulted in the bond-valence method [35] which accounts, in a static single-particle picture, for volume and energy effects and has already been used in large-scale screening for ionic conductors [36,37]. Wang et al [20] could relate superionicity to the bcc-like topology of the underlying anionic sublattice, as was also discussed by Wood and Marzari [8] for AgI. Work by Adelstein and Wood [38] showed how the mixed ionic-covalent nature of lithium bonds and frustration of the bonding during transition can explain superionic behavior observed in Li 3 InBr 6 .…”

Comparison of Anion Reorientational Dynamics in MCB9H10 and M2B10H10 (M = Li, Na) via Nuclear Magnetic Resonance and Quasielastic Neutron Scattering Studies

“…closo-Borate-based salts of Li and Na have been shown to exhibit impressive superionic conductivities in their highertemperature, entropy-driven, disordered phases. 1−4 Although these solid-state phases retain a translationally rigid facecentered-cubic (fcc), body-centered-cubic (bcc) or hexagonal stacking of large polyhedral anions (such as B 12 , and CB 9 H 10 − ; see Figure 1 for a depiction of the latter two), 2−6 the individual anions are orientationally extremely mobile, with typical reorientational jump frequencies in the range of 10 10 −10 11 s −1 above the compound disordering temperatures. 3−5,7−9 Moreover, Li + and Na + cations travel with liquid-like mobilities (typically >10 8 jumps s −1 ) 1,2,4,7,10 throughout the partially vacant interstitial space afforded by these anions.…”

“…Further insights into the nature and importance of such dynamical effects will require even more comprehensive theoretical investigations. Of course, these extra dynamical effects are not operational in the more traditional solid ionic conductors such as α-AgI, 12 which possesses a monatomic anion, or cubic Na 3 , and PO 4 3− anions exhibit no orientational disorder and likely possess much less orientational mobility. In these cases, anion dynamical effects can be ignored, and cation conduction depends predominantly on the physics of diffusive translations through an orientationally static array of anions.…”