Metastable pathways allow high rates: In batteries that allow for fast charging and discharging, lithium usually forms a solid solution with the anode so that the only limiting factor is the ionic diffusion. However, for a lithium titanate (Li 4 Ti 5 O 12 ) anode, the lithium ions interact with two phases and the diffusion is slow in both, but it still shows high-rate capabilities. Zhang et al. used electron energy-loss spectroscopy combined with density functional theory calculations to probe the anomalous behavior. They found that a diffuse interface forms between the starting and ending compositions, Li 4 Ti 5 O 12 and Li 7 Ti 5 O 12 , and this is what allows the lithium ions to travel quickly. Science , this issue p. 1030
27 October 2010 in The Journal of Chemical Physics
We describe an ab initio approach to compute the optical absorption spectra of molecules and solids, which is suitable for the study of large systems and gives access to spectra within a wide energy range. In this approach, the quantum Liouville equation is solved iteratively within first order perturbation theory, with a Hamiltonian containing a static self-energy operator. This procedure is equivalent to solving the statically screened Bethe–Salpeter equation. Explicit calculations of single particle excited states and inversion of dielectric matrices are avoided using techniques based on density functional perturbation theory. In this way, full absorption spectra may be obtained with a computational workload comparable to ground state Hartree–Fock calculations. We present results for small molecules, for the spectra of a 1 nm Si cluster in a wide energy range (20 eV), and for a dipeptide exhibiting charge transfer excitations.
Intermolecular interactions in the van der Waals bonded benzene crystal are studied from first principles, by combining exact exchange energies with correlation energies defined by the adiabatic connection fluctuation-dissipation theorem, within the random phase approximation. Correlation energies are evaluated using an iterative procedure to compute the eigenvalues of dielectric matrices, which eliminates the computation of unoccupied electronic states. Our results for the structural and binding properties of solid benzene are in very good agreement with experimental results and show that the framework adopted here is a very promising one to investigate molecular crystals and other condensed systems bound by dispersion forces.
Simulations of excited state properties, such as spectral functions, are often computationally expensive and therefore not suitable for high-throughput modeling. As a proof of principle, we demonstrate that graph-based neural networks can be used to predict the x-ray absorption near-edge structure spectra of molecules to quantitative accuracy. Specifically, the predicted spectra reproduce nearly all prominent peaks, with 90% of the predicted peak locations within 1 eV of the ground truth. Besides its own utility in spectral analysis and structure inference, our method can be combined with structure search algorithms to enable high-throughput spectrum sampling of the vast material configuration space, which opens up new pathways to material design and discovery.
18 July 2002 in The Journal of Physical Chemistry B