The replacement of conventional liquid electrolytes with solid electrolytes has the potential to safely enable energy-dense Li-metal anodes. Because of the challenges surrounding solid-solid interfaces, it is crucial to better understand the Li-metal-solid-electrolyte interface. This work utilizes stack pressure to correlate mechanics with the electrochemical behavior of Li-electrolyte cells during galvanostatic cycling. Symmetric cells are constructed using Li 7-La 3 Zr 2 O 12 and tested using AC and DC techniques under dynamic stack pressure conditions. It is demonstrated that significant polarization occurs during galvanostatic cycling at a current-dependent ''critical stack pressure.'' Using reference electrodes, this effect is isolated to the Li stripping electrode. This suggests that at low pressures, the Li stripping rate exceeds the rate at which mechanical deformation replenishes the interface, inducing the formation of voids and ultimately increasing resistance. This analysis not only motivates the need for further understanding of the Li-metal-solid-electrolyte interface but also provides guidelines for the future design of all-solid-state batteries.

PACS 31.15. Ew, 71.15.Mb While the success of density functional theory (DFT) has led to its use in a wide variety of fields such as physics, chemistry, materials science and biochemistry, it has long been recognised that conventional methods are very inefficient for large complex systems, because the memory requirements scale as N 2 and the cpu requirements as N 3 (where N is the number of atoms). The principles necessary to develop methods with linear scaling of the cpu and memory requirements with system size (O(N) methods) have been established for more than ten years, but only recently have practical codes showing this scaling for DFT started to appear. We report recent progress in the development of the CONQUEST code, which performs O(N) DFT calculations on parallel computers, and has a demonstrated ability to handle systems of over 10 000 atoms. The code can be run at different levels of precision, ranging from empirical tight-binding, through ab initio tight-binding, to full ab initio, and techniques for calculating ionic forces in a consistent way at all levels of precision will be presented. Illustrations are given of practical CONQUEST calculations in the strained Ge/Si(001) system.

Various aspects of the implementation of pseudo-atomic orbitals (PAOs) as basis functions for the linear scaling CONQUEST code are presented. Preliminary results for the assignment of a large set of PAOs to a smaller space of support functions are encouraging, and an important related proof on the necessary symmetry of the support functions is shown. Details of the generation and integration schemes for the PAOs are also given.

The Ge͑105͒ surface has attracted attention recently, both from interest in the reconstruction itself and because the facets of three-dimensional hut clusters which form during heteroepitaxy of Ge on Si͑001͒ are strained Ge͑105͒ surfaces. We present density functional theory ͑DFT͒ studies of this surface using local basis sets as a preparation for O͑N͒ DFT studies of full hut clusters on Si͑001͒. Two aspects have been addressed. First, the detailed buckling structure of the dimers forming the surface reconstruction is modeled using DFT and tight binding; two different structures are found to be close in stability, the second of which may be important in building hut-cluster facets ͓as opposed to perfect Ge͑105͒ surfaces͔. Second, the accuracy that can be achieved using local basis sets for DFT calculations is investigated, with O͑N͒ calculations as the target. Two different basis sets are considered: B splines, also known as blips, and pseudoatomic orbitals; B splines are shown to reproduce the result of plane-wave calculations extremely accurately. The accuracy of different modes of calculation ͑from non-self-consistent ab initio tight binding to full DFT͒ is investigated, along with the effect of cutoff radius for O͑N͒ operations. These results all show that accurate, linear-scaling DFT calculations are possible for this system and give quantitative information about the errors introduced by different localization criteria.

Electronic structure methods based on density-functional theory, pseudopotentials, and local-orbital basis sets offer a hierarchy of techniques for modeling complex condensed-matter systems with a wide range of precisions and computational speeds. We analyze the relationships between the algorithms for atomic forces in this hierarchy of techniques, going from empirical tight-binding through ab initio tight-binding to full ab initio. The analysis gives a unified overview of the force algorithms as applied within techniques based either on diagonalization or on linear-scaling approaches. The use of these force algorithms is illustrated by practical calculations with the CONQUEST code, in which different techniques in the hierarchy are applied in a concerted manner.