Rejection of low molecular weight solutes by mean of cnts: A quantum mechanics and atomistic study

The rejection of charged and neutral molecules with low Molecular Weight, as well as the membrane fouling are fundamental aspects that should be considered during the waste water treatment. The material choice, used in the membrane preparation, is of great importance (material design) as well the optimization of parameters related to the integrated separation processes (process design). The carbon nanotubes (CNTs) have shown amazing hydrodynamic properties. Thus, CNTs-composite membranes can be considered promising solutions for the waste water treatment. The special fluiddynamics of CNTs has been thoroughly studied using experimental and modelling approaches1,2. Although the CNT embedding in the correct orientation constitutes largest drawback to the development of CNTs-composite membranes, the ability of CNTs to reject low MW solutes remains a relevant issue. A computational study on the rejection of interesting solutes by CNTs is presented here. After the optimization of the CNT diameters as a function of their rejection capability, the analysis of the water flow in the identified nanotubes was carried out. Thus, rejection and water flow were connected. The originality of this study lies in the suitably combination of different computational methods: Quantum Mechanics (QM), Monte Carlo (MC) and Molecular Dynamics (MD). Whilst QM allows a sub-nano scale investigation, MC and MD simulations permit analysis in the nano-scale. QM gives accurately description of the noncovalent interactions among solutes, water and CNTs, regardless of use of ad-hoc Force Fields3,4 while MC and MD simulations allow to analyse the solute dynamics in the CNTs. Thus, the combination of these different approaches provides an overview on the CNTs selectivity. Thirteen charged and neutral solutes of large, medium and small molecular weight were considered, such as tyrosol, vanillic acid, EDTA, octylphenol ethoxylate and etc. Their geometries were optimized at quantum mechanics level in the frame of Density Functional Theory. As regards the large molecules, conformer research was performed before QM optimizations. The ab-initio geometries permit to be released from experimental parameters without losing any generality. The calculated geometries were then used to evaluate the effective diameter, Deff, and cross-sections of each single molecule including the vdW radii of the solute atoms. The Deff calculation was performed using an home-made algorithm, which also enables to give the maximum and minimum projections of the solute atoms on the CNTs opening (Dproi-max and Dproi-min). Dproi-max was used to draw the solute arrangement into a CNT according to the conformation shown in Figure 1. The CNTs with internal diameter ranging between 1.1 nm and 10 nm were used in this study. It was assumed that the conformation, shown in Figure 1, may be indicative about the molecular packing in CNTs. Aware that this simple consideration is reductive, MC and MD simulations were carried out to investigate more accurately the solute packaging in the CNTs with diameter smaller than 2.77 nm. These simulations were carried out in the grand canonical ensemble with a very efficient Monte Carlo algorithm and they allowed to study the sorption of small molecules inside smooth single-wall nanotubes. The obtained results reveal highly-ordered structures, their degree of ordering depending strongly on the CNT diameter. Representative configurations form Figure 1. Arrangement of tyrosol molecule with the phenyl parallel to the main axis of (8,8) CNT the MC studies were then subjected to MD simulations in the isothermal-isobaric statistical ensemble at temperature T=298K and P=1atm with the LAMMPS code, using the DREIDING forcefield5. The outcome, here, was the calculation of the mean res