Multivalent ion batteries (MVIBs) provide an inexpensive and energy-dense alternative to Li-ion batteries when portability of the battery is not of primary concern. However, it is difficult to find cathode materials that provide optimal battery characteristics such as energy density, adequate charge/discharge rates, and cyclability when paired with a multivalent ion. To address this, we investigate six MnO 2 polymorphs as cathodes for MVIBs using density functional theory calculations. We find voltages as high as 3.7, 2.4, 2.7, 1.8, and 1.0 V for Li, Mg, Ca, Al, and Zn, respectively, and calculate the volume change due to intercalation. We then focus specifically on Ca and compute the energy barriers which are associated with the diffusion of the ion throughout the materials. Our findings show that the α-phase displays the most rapid diffusion kinetics for a Ca ion, with a diffusion barrier as low as 190 meV. We then investigate the potential for the five polymorphs exhibiting the highest voltage to intercalate additional atoms and demonstrate that it is energetically favorable for each to accept at least one additional Ca ion; furthermore, two of the phases can accept more than two Ca ions. However, in each case, there is also a corresponding drop in the voltage as further atoms are intercalated. We also utilize a crystal-chemistry approach to detail the structural evolution of each phase by computing the bond valence sum and effective coordination of the Mn 4+ ions upon intercalation of increasing numbers of Ca ions. Finally, by computing the electronic density of states, Bader charges, and real space charge density, we describe how the additional electrons from the Ca ions are distributed throughout the unit cell. These insights provide guidance in selecting a MnO 2 polymorph with the traits necessary for the realization of MVIBs.

One of the major bottlenecks to the development of alternatives to existing Li ion battery technology, such as Li metal or multivalent ion (Mg, Ca, Zn, or Al) batteries, has to do with the layer of inorganic and organic compounds that forms at the interface between the metallic anode and electrolyte via solvent and salt decomposition (the solid–electrolyte interphase or SEI). In Li metal batteries the growth of dendrites causes continual formation of new SEI, while in multivalent ion batteries the SEI does not allow for the diffusion of the ions. Finding appropriate electrolytes for such systems and gaining an understanding of SEI formation is therefore critical to the development of secondary Li metal and multivalent ion cells. In this work, we use ab initio molecular dynamics simulations to investigate the initial stages of decomposition of organic electrolytes based on ethylene carbonate (EC) and formation of the SEI on Li, Ca, and Al metal surfaces. We first find that pure EC only decomposes to CO and C2H4O2 2– species on each type of surface. However, when a salt molecule is introduced to form an electrolyte, a second EC decomposition route resulting in the formation of CO3 2– and C2H4 begins to occur; furthermore, a variety of different inorganic compounds, depending on the chemical composition of the salt, form on the surfaces. Finally, we find that EC breaks down more quickly on Li and Ca surfaces than on Al and show that this is because the rate of charge transfer is much faster owing to their lower electronegativity and ionization energies. The molecular level understanding of decomposition and SEI formation generated by this computational modeling can lead to the design of new electrolytes for beyond-Li ion batteries.

Hybrid density functional theory (DFT) is used to study the Chevrel phase MoX (X = S, Se, Te) as a promising cathode material intercalated with various metal ions (M = Li, Na, Be, Mg, Ca, Sr, Ba, Zn, Al). Electronic properties and voltages are calculated for each case. Ca ions are predicted to produce a voltage output ranging from 1.8-2.1 V, comparable to the voltage calculated for Li ions while providing two electrons per transferred ion. The highest voltage is determined to result when the chalcogen X in MoX is S, over Se or Te. Additionally, a comparison of the local-density approximation (LDA), the Perdew-Burke-Ernzerhof (PBE), the Hubbard U corrected GGA-PBE (PBE+U), the meta-GGA modified Becke-Johnson (mBJ), and the hybrid Heyd-Scuseria-Ernzerhof (HSE) functionals are made. The electronic structure determined with HSE is taken as the most reliable, and PBE and LDA can provide reasonable approximations. The PBE+U approach yields an erroneous band gap and should be avoided. The voltages calculated with HSE are in excellent agreement with available experimental data.

No abstract