Reports: DNI652494-DNI6: Dynamics and Kinetics of Carbon Dioxide in Nano-Porous Materials Environment from Quantum Molecular Dynamics Simulations

Yosuke Kanai, PhD, University of North Carolina (Chapel Hill)

Figure 1. Power spectrum of CO2 in water and in vacuum, calculated via velocity-velocity auto-correlation function from the FPMD trajectories.

Text Box: Figure 1. Power spectrum of CO2 in water and in vacuum, calculated via velocity-velocity auto-correlation function from the FPMD trajectories.

Several key progresses were made in few different areas during the second year of the grant period (9/1/2013-8/31/2014).

CO2 in H2O

We have been investigating how water solvation influences the CO2 dynamics. As the first step, we have calculated the vibrational spectrum of CO2 in vacuum and in water. From the trajectories of first principles molecular dynamics (FPMD), the power spectrum can be obtained by taking the Fourier transform of the velocity-velocity auto-correlation function between atoms. In vacuum, the well-known splitting at around ~1300 cm-1 due to Fermi resonance is apparent (Figure 1). However, with water solvation, much of the detailed features are washed out, and there is a prominent broad peak centered around 3000 cm-1. This broad peak was identified to derive from the hydrogen bonding with surrounding water molecules.

           We also started to investigate how the presence of a solvated electron influences the dynamics of CO2 in water. From the trajectory of the neutral CO2 in water, we arbitrarily selected three configurations (separated by the minimal of 2 picosecond) and re-optimized the electronic structure of the system with an extra electron before proceeding with the molecular dynamics. The CO2 bending angle was monitored during the simulation, and the initially fluctuation of CO2 angle around 170¡ was quickly and rather abruptly reduced and began fluctuating around 135¡ in all three cases (Figure 2).

Figure 2. CO2 bending angle change after an extra electron was introduced in the water solvated system at three instances.

Text Box:
Figure 2. CO2 bending angle change after an extra electron was introduced in the water solvated system at three instances.

H2O in confinement

FPMD simulations were used to gain an atomistic-level insight into how the molecular behavior of interfacial water is influenced by specific surface adsorbates. Although the overall hydrophobic versus hydrophilic character of a given surface is widely recognized to be important in determining the behavior of interfacial water molecules, we have shown that subtle molecular details may also play a role in determining the dynamical behavior of water. By comparing water diffusivity at three different nonpolar surfaces, we find that specific surface features can lead to a suppression of hydrogen bond network ring structures by enhancing hexagonal spatial distributions of water molecules near the surface. Such a distinct molecule-dependent behavior of the interfacial water was found to persist well into the liquid, while most structural properties are noticeably influenced in only the first water layer.

In particular, water molecules diffuse significantly faster at the H−Si surface than at CH3− or CF3−Si surfaces because the hydrogen bond ring network formation is suppressed at those surfaces (Figure 3). The electrostatic potential generated by the adsorbate species appears to dictate the spatial water distribution in the first water layer. Such behavior is clearly noted for both metal- and carbon-based surfaces in the literature. Interestingly, such a distinct long-range molecular behavior in the first water layer extends farther away (1 nm) because HB ring sizes

Figure 3. Mean square displacement of water molecules at Si surfaces passivated with different surface functional groups.

Text Box:
Figure 3. Mean square displacement of water molecules at Si surfaces passivated with different surface functional groups.
are appreciable (4−5) in water. Our present study shows that long- range structural arrangements of water molecules are also important to dictate the dynamical behavior of water molecules at surfaces passivated by different hydrophobic molecules. The observed structural dependence on the dynamics of interfacial water is likely to be a universal character for a wide range of surfaces.

Role of Charge Transfer in describing Water Diffusivity

Figure 4. Mean square displacement of water molecules with several ionic species using the three different classical MD potentials and using FPMD simulation.

Text Box:
Figure 4. Mean square displacement of water molecules with several ionic species using the three different classical MD potentials and using FPMD simulation.
During the investigation of the CO2 hydrolysis to bicarbonate ion in bulk water, it has become clear that classical molecular dynamics simulations are necessary to address some important dynamical properties such as diffusion constant of the water molecule when CO2 molecule is solvated. First-principles molecular dynamics (FPMD) turns out to be computationally too expensive for computing the diffusion constant accurately in terms of statistical sampling. At the same time, most classical molecular dynamics is known to suffer from the inadequate treatment of charge transfer. Building on the recent work by Rick and co-workers (1), we investigated the performance of the new potential that can take into account both electronic polarization and charge transfer. Classical molecular dynamics simulations were performed using three different force fields: a fixed charge force field (SPCE/HMN), a polarizable force field that includes explicit polarization (SWM4-NDP), and also a recently developed force field that includes polarization and charge transfer (QF-DCT). These simulations were then compared to FPMD simulations. We performed molecular dynamics simulations on four types of systems containing ion and solvating water. Two systems contained a cation (Na+ or K+), and two other systems an anion (Cl or I) as representative cases. While the first-principles simulations showed that the anions accelerated water translational diffusion, the cations slowed it down. In simulations with the classical force fields, only the force field that incorporates explicit charge transfer reproduced this ion-specific behavior (Figure 4). Additional simulations performed to understand the effect of charge transfer demonstrated that two competitive factors determine the behavior of water translational diffusion: the ions diminished charge accelerates water, while the net charge acquired by water either accelerates or slows down its dynamics. Our results showed that charge transfer plays a crucial role in governing the water dynamics in aqueous ionic solutions.

References

1. Soniat, M.; Rick, S. W. The Effects of Charge Transfer on the Aqueous Solvation of Ions. J. Chem. Phys. 2012, 137 (4), 044511.