62nd Annual Report on Research 2017

We carry out geometry optimizations and parallel tempering simulations to determine the impact of a neutral lithium atom, and a lithium ion on the structural and thermodynamic properties of a cluster comprised of a mixture of two dipolar particles. A total of four systems are investigated, the Li⁺@(Nitromethane)₆(THF)₁₂, the neutral Li@(Nitromethane)₆(THF)_12, and the solvent mixture (Nitromethane)₆(THF)₁₂ are investigated so that their behaviors can be compared. The nitromethane and THF molecules are treated as point dipoles – Lennard – Jones particles. One would like to use a mixtures of two types of dipolar molecules to better understand the role of the second "solvation shell", with a first solvation layer mimicking the additive used to dissolve the ionic compounds typically used in lithium ion batteries. As proposed originally, we make use of a coarse grained representation of the interactions to be able to scan the parameters, and in the present ongoing work to determine the range of temparatures where the quantum nature of the Lithium ion may have an impact on the solvation energy and dynamics. The composition is selected based on the previously reported solvent dissociation energies of Li⁺@(Nitromethane)_n (n=1 → 20) where we have identified the first "magic number" that marks the completion of the second solvation layer for the lithium ion - nitromethane clusters as the n = 18 system, a solvated ion with the first sheath having octahedral symmetry. We have chosen to build a second solvation layer with 12 weaker dipoles with Stockmayer parameters selected to approximate a tetrahydrofuran molecule as its dipole moment is typical of solvents used in lithium ion batteries.

We find that the structure of the global minimum, the minima energy density, and the thermodynamic properties are vastly different among all the species. For the structural optimization we employ a parallel implementation of the genetic algorithm, and for the largest system we follow this with an equally long basin hopping calculation. A total of 12 generations are sufficient to reach convergence for the smallest systems, whereas the optimizations of the largest system is ongoing. Though similar in composition, the structural differences among the three clusters are profound. The hexacoordinated lithium ion is surrounded by the most polar particles, as expected. For the remaining THF particles there is a competition between two stabilizing configurations. The first is to coordinate the lithium ion from the second solvent shell while aligning THF dipoles with the nitromethane ones. This competes with the stabilization obtained by aligning THF like dipoles. For the present model, the vertices of the octahedral first layer are coordinated asymmetrically, leaving the octahedral layer on the surface of the (THF)₁₂ cluster. The latter is not sufficiently large to observe the braid to ring structure we have seen in our work with Li⁺(Nitromethane)_n - like clusters.[J. Chem. Phys. 143, 214301 (2015)]. When the charge is removed, the (Nitromethane)₆ cluster forms a hexagonal ring embedded in an icosahedral – like network of THF. Surprisingly, the neutral lithium atom occupies the volume inside the icosahedral cage formed by the latter.

We use structural comparison data, coupled with visualization by graphical software of selected configurations to gain insight into the most prominent cold features in the heat capacities of the three species. We note that at the coldest temperatures, the THF - like dipoles are fluxional in the two species, Li@(Nitromethane)₆(THF)₁₂ and the solvent mixture (Nitromethane)₆(THF)₁₂ .The THF dipoles solvate the neutral lithium atom by drawing it into a icosahedral cage and this has the effect of stabilizing the THF component . The solid - solid like features of (Nitromethane)₆(THF)₁₂ and Li⁺@(Nitromethane)₆(THF)₁₂, involve only the first few minima along with the global one, and the random walkers at the low temperatures are predominately near the global minimum and the next isomer. The (THF)₁₂ framework melts in the Li@(Nitromethane)₆(THF)₁₂ system at around 95 K, whereas the (Nitromethane)₆ framework (albeit quite different in the neutral species compared to the ionic cluster) melts well above 100 K. These observations explain quite well the presence of three peaks below 200 K in their heat capacity of the ion and the base system, and only two peaks in the neutral lithium- mixed dipoles cluster.

The data allows us to conclude that the presence of the ion in particular, drastically enhances the fluxional nature of the less polar components which occupy the second solvation layer, whereas the neutral atom has the effect of reducing it. These effects are investigated in detail by analyzing the simulations structurally. Though quite complex, the structure and thermodynamic behaviors of Li⁺@(Nitromethane)₆(THF)₁₂, is exquisitely predictable. The structure of the minima tell us that the core remains intact as the system freezes. The weaker dipole, coupled with the shielding of the first solvation layer allow for several solid - solid like phase changes to exist at extremely cold temperatures, and cause the THF component to melt at relatively cold temperatures compared to the nitromethane component. The structures and thermodynamic behaviors of the bare mixed Stockmayer cluster and the Lithium - mixed Stockmayer cluster are much less predictable. Inspecting the minima of the two neutral species we conclude that the nitromethane component segregates and crystallizes separately from the THF component, whereas the lithium atom co-crystallizes with it. The Li@(Nitromethane)₆(THF)₁₂ system is surprisingly rigid at cold temperatures compared to the lithium free cluster indicating that the neutral lithium atom plays a pronounced stabilizing role. This can perhaps be appreciated best by the adiabatic solvation energy of lithium which takes negative values at temperatures above 160 K. The next question for these systems is to determine the size of the quantum effects at different temperatures. Fourier Path Integral calculations are in progress. A total of four undergraduate students worked on these related projects. All four of them will be submitting an abstract for a poster at the next National ACS meeting in March 2018.