Molecular Modeling of the Physical Properties for Aqueous Amine

Subscriber access provided by University of Sussex Library

Article

Molecular Modeling of the Physical Properties for Aqueous Amine Solution Containing a CO2 Hydration Catalyst Wei Shi, Leland R. Widger, Moushumi Sarma, Cameron A. Lippert, David E. Alman, and Kunlei Liu Ind. Eng. Chem. Res., Just Accepted Manuscript • DOI: 10.1021/acs.iecr.7b03224 • Publication Date (Web): 18 Sep 2017 Downloaded from http://pubs.acs.org on September 23, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Industrial & Engineering Chemistry Research is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

Molecular Modeling of the Physical Properties for Aqueous Amine Solution Containing a CO2 Hydration Catalyst Wei Shi,∗,†,‡,¶ Leland R. Widger,§ Moushumi Sarma,§ Cameron A. Lippert,§ David E. Alman,† and Kunlei Liu§ †U. S. Department of Energy, National Energy Technology Laboratory, Pittsburgh, PA 15236, USA ‡AECOM, South Park, PA 15129, USA ¶Department of Chemical and Petroleum Engineering, University of Pittsburgh, Pittsburgh, PA 15261, USA §Center for Applied Energy Research, University of Kentucky, Lexington, KY 40511, USA E-mail: [email protected] Phone: +1 412 386 4406. Fax: +1 412 386 5990 Abstract The effects of an amphiphilic CO2 hydration catalyst (C3P) on the physical properties of aqueous monoethanolamine (MEA) solutions were studied using molecular simulations and verified experimentally. Adding 2.7−27.7 g/L of C3P in 30 wt% MEA aqueous solution did not significantly affect the solution viscosity, surface tension, or CO2 diffusivity. These results confirm that the previously reported increase in CO2 mass transfer by C3P is due to CO2 hydration catalysis, and not due to changes in the physical properties of the MEA solution. Additional simulations indicate that the

1

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

catalyst molecules tend to aggregate in MEA solution, and are preferentially adsorbed at the gas-liquid interface region. For the catalyst molecules remaining in the bulk solution, the local concentrations of CO2 and MEA in the area immediately around the catalyst are increased while the local water concentration is decreased, relative to their concentrations in the rest of the bulk MEA solution.

1

INTRODUCTION

Reduction of global CO2 emissions will require the widespread use of carbon capture and sequestration (CCS) technologies. 1 As the most widely studied for over 30 years, amine solvent-based CCS systems are the mostly likely to be adopted commercially. 2–4 Industrially relevant solvents often incorporate additional components to improve solvent stability or performance, such as corrosion inhibitors, radical scavengers, surfactants, catalysts, etc. However, additives can alter the physical properties of a solvent and can have significant impact on solvent performance. For example, the mass transfer coefficient (KG ) for CO2 in the aqueous monoethanolamine (MEA) solution, which determines the height of a packed gas absorption tower and the absorption capital cost, is a lumped parameter that depends on the solution physical properties (i.e. solution viscosity, surface tension, and CO2 diffusivity in the solution), the liquid superficial velocity, the size of the packing material, and the CO2 reaction rate. 5,6 Establishing structure-function relationships for the effects of additives on the solvent physical properties would be desirable, but is difficult due to the complex matrix of these concentrated aqueous solutions. The development of molecular modeling computational tools for predicting these relationships is valuable that would aid significantly in screening and selecting additives based on fundamental properties. Recent work from our lab has shown the first reported examples of zinc- and cobaltcontaining coordination complexes, inspired by carbonic anhydrase (CA), that are able to enhance the KG of CO2 absorption into aqueous MEA solution under industrially-relevant conditions. 7–10 The zinc-containing catalyst C3P was previously shown to increase the CO2 2


Page 2 of 32

Page 3 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


KG in 30 wt% MEA solution by 23−34%. 7 The increased KG was attributed to the enhancement of the CO2 hydration reaction rate, catalyzed in a manner analogous to that of CA. 7 In turn, this increased KG has a significant impact on CO2 capture cost; it could reduce the carbon capture capital cost by 25 million dollars for a 550 MW power plant, relative to the MEA reference case without catalyst. 7 In spite of the above experimental progress, there are still some important questions to be addressed computationally, such as the effects on solution physical properties when adding the C3P catalyst, the catalyst molecule absorption behavior in the solution. Herein we report molecular modeling studies of the physical properties (viscosity and surface tension) and CO2 diffusivity for aqueous MEA solution with the C3P catalyst additive, which are verified experimentally where possible. Modeling of the solution behavior of C3P shows that catalyst molecules aggregate in the solution and that they are preferentially adsorbed at the gas-liquid interface. The local concentrations for CO2 , MEA and H2 O around the C3P molecule in the bulk solution are also studied, and may affect the CO2 absorption rate.

2

Experimental Details

Reagents for solvent mixtures, catalyst synthesis, and kinetics studies were purchased from Sigma Aldrich, monoethanolamine was purchased from Univar. Solutions of monoethanolamine (MEA, 5 M, 30 wt%) were prepared by weight % and adjusted to an alkalinity of 5.0 mol N/Kg. The surface tension data was acquired at 25 ◦ C on a Biolin Scientific Optical Tensiometer, using Oneattension software. The viscosity of the solutions were determined by using the Brookfield DVI viscometer.

3



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

3

Page 4 of 32

SIMULATION DETAILS

3.1

Classical Force Field (FF)

The classical FF potential used to simulate C3P, H2 O, MEA, and CO2 molecules and their interactions is given by

V(r) =

X

kr (r − r0 )2 +

bonds

+

X

kθ (θ − θ0 )2

angles

X

kχ [1 + cos(n0 χ − δ0 )] +

+

i=1

kψ (ψ − ψ0 )2

impropers

dihedrals

 N −1 X N  X

X



σij 4ǫ   ij rij j=i+1

!12

σij − rij

!6 



qq  + i j , rij 

(1)

where the symbols represent their conventional meanings. 11 Standard Lorentz-Berthelot combining rules were used to calculate the mixed Lennard-Jones (LJ) interaction parameters. The LJ potential was switched from 10.5 to 12.0 ˚ A. A Verlet neighbor list with a 13.5 ˚ A radius was used. The intramolecular electrostatic and LJ interactions for atoms separated by exactly three consecutive bonds were scaled by 0.5 and were neglected for atoms separated by less than three consecutive bonds. The classical FF parameters for CO2 , 12 H2 O, 12 and MEA 13 were taken from the previous work. Bond and angle potentials were added to Zn in the C3P catalyst molecule (Figure 1) to maintain the Zn metal coordination with the O, N and Cl atoms. These types of bond and angle potentials have also been used by others to model the flexibility of metal of framework. 14 The atomic partial charges for C3P were obtained from quantum ab initio calculations (see below). The details to set up the classical FF parameters for C3P are given in the Supporting Information. In order to calculate transport properties such as gas diffusivity and solution viscosity, 20 ns NV E molecular dynamics (MD) simulations were run by following the NP T and NV T MD simulations. 15,16 The 0.5 fs time step was used in NV E MD simulation. Coordinates and

4


Page 5 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


pressure tensors were dumped every 1000 and 1 fs, for diffusivity and viscosity calculations, respectively. All MD simulations in this work were performed by using the NAMD program. 17 The smooth particle-Mesh Ewald method 18 was used to calculate the electrostatic interaction in MD simulations.

3.2 3.2.1

Quantum Ab Initio (AI) Calculations AI Gas Phase Calculation

The optimized structure for C3P was obtained from the quantum AI gas-phase optimization calculation at the B3LYP level of theory by using the Gaussian 09 program; 19 the initial C3P structure of the AI calculation was obtained from the experimental crystal data. 7 The 6-31G(d,p) basis set was used for the C, H, N, O, Cl, and P non-metal atoms; an effective core potential and the LANL2DZ basis set were used for the Zn(II) transition metal atom. The atomic partial charge for each atom of the C3P molecule was also obtained from AI calculations by using the CHELPG protocol 20 at the same level of theory as optimization. The atomic radii used in CHELPG calculations are: 1.39 ˚ A for Zn, 1.55 ˚ A for N, 1.52 ˚ A for A for Cl atom. 21 A for P, and 1.75 ˚ A for H, 1.80 ˚ A for C, 1.20 ˚ O, 1.70 ˚ 3.2.2

Ab Initio Molecular Dynamics (AIMD) Condensed-Phase Simulation

In order to investigate the C3P stability in water and aqueous MEA solution, AIMD simulations were performed by using the CPMD 22,23 (version 3.15.3) program. For the first system, one C3P molecule is solvated by 230 water molecules in a 19.912 ˚ A×19.912 ˚ A×19.912 ˚ A cubic box, which contains 802 atoms. The NV T simulation was run for 60 ps at 298 K. The second system contains 632 atoms, which consists of 1 C3P, 100 H2 O and 20 MEA molecules. The cubic simulation box length was set to be 18.952 ˚ A. A 11-ps NV T production run was conducted. The simulation box length and the initial molecular configuration in AIMD simulation were obtained from the classical FF NP T MD simulation. The classical FF NP T MD simulations were conducted for 20 ns by starting from a configuration in which the C3P, 5



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 32

water, and MEA molecules were randomly put in the simulation box. The C3P molecular structure which corresponds to the experimental crystal data was fixed during the whole classical FF simulation. As indicated by the simulation snap shot, the 20-ns classical FF simulation is long enough to mix water and MEA molecules. The BLYP exchange-correlation functional and norm-conserving pseudo-potentials of the Troullier-Martins form were employed in AIMD simulations. The Grimme dispersion correction 24 implemented in the CPMD program was also applied. More AIMD simulation details can be found in our previous work. 25 Considering the computing resources and the expensive AIMD simulations, we have run 60 ps and 11 ps NVT simulations for the above two systems although longer simulations will give better statistics. Further root-meansquare deviation (RMSD) analysis between the NVT production snap shots and the experimental C3P crystal structure indicates that the RMSD values fluctuate around 0.75 ˚ A after 3 ps, which suggests that the equilibrated C3P structure was obtained in AIMD simulations.

3.3

Viscosity

The viscosity was calculated as previously reported 16 at 298 K and 1 bar by using the following Einstein relation, 1 V d η = lim t→∞ 2 kb T dt

*Z

t0 +t

1 V d = lim t→∞ 2 kb T dt

*Z

t0 +t

t0

0

′

′

Pζν (t )dt

2 +

t0

′

′

Pζν (t )dt −

Z

0

t0

′

′

Pζν (t )dt

2 +

,

(2)

t0

where V is the volume of the simulation box, kb is the Boltzmann constant, T is the temperature, Pζν is the symmetrized pressure tensor, and hit0 indicates the average obtained from multiple time origins. Viscosities were calculated for three systems: 1) 30 wt% MEA aqueous solution (500 MEA and 3950 water molecules; cubic simulation box length 54.715 ˚ A); 2) 2.77 g/L C3P in 30 wt% MEA solution (1 C3P, 2000 MEA and 15800 water molecules, cubic simulation box 6


Page 7 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


length 86.918 ˚ A); 3) 27.7 g/L C3P in MEA solution (10 C3P, 2000 MEA and 15800 water molecules, cubic simulation box length 87.492 ˚ A). For each system, the viscosity uncertainty was estimated from three sets of viscosity calculations, each of which corresponds to an independent NV E MD simulation. The linear regime for at 100 ps ≤ t ≤ 2 ns were used to calculate the viscosity.

3.4

R

t0 +t t0

Pζν (t′ )dt′

2

versus time

t0

Surface Tension

There are two types of methods to compute the surface tension from molecular modeling, mechanical 26–29 and test-area methods. 30 In this work, the test-area method as previously implemented with our in-house software package 31 was used to compute the surface tension by using configurations obtained from the NV T MD simulation, which involves the vaporliquid equilibrium and the interface between the vapor and the liquid phases. Along with the γsim obtained from the test-area calculation, a tail correction (γtail ) is added to obtain the total surface tension (γsim,total ). More details for γsim and γtail calculations can be found in our previous work. 31 To speed up the calculation, the Fennel and Gezelter shift force method 12,32 was used to compute the electrostatic interaction in γsim calculations. Similar to viscosity calculation, surface tensions were also computed for three systems: 1) 30 wt% MEA solution (2000 MEA and 15800 water molecules, simulation box dimensions 86.854 ˚ A × 86.854 ˚ A × 160 ˚ A); 2) 2.77 g/L C3P in 30 wt% MEA (1 C3P, 2000 MEA and 15800 water molecules, simulation box dimensions 86.918 ˚ A × 86.918 ˚ A × 160 ˚ A); 3) 27.7 g/L C3P in 30 wt% MEA (10 C3P, 2000 MEA and 15800 water molecules, simulation A). The box length in the z-dimension was A × 160 ˚ A × 87.492 ˚ box dimensions 87.492 ˚ elongated to create a vapor-liquid interface. For all three systems, 50,000−110,000 configurations corresponding to the last 50−110 ns production runs were saved for surface tension calculation. For each system, three independent simulations were performed and they give consistent surface tension values. These results suggest that simulations are long enough to give precise surface tension values. 7



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

3.5

CO2 Diffusivity

The CO2 self-diffusivity in solution was calculated from the NV E MD simulation using the Einstein relation. More diffusivity calculation details can be found in our previous work. 15,33 Similar to the above viscosity and surface tension calculations, CO2 diffusivities were calculated for three solutions, that is, 30% MEA aqueous solution, C3P-MEA solution mixtures with 2.77 g/L and 27.7 g/L C3P. To improve the simulation statistics, 50−200 CO2 molecules were put in the solutions. The CO2 mean square displacements (< ∆r 2 (t) >) in the linear region of < ∆r 2 (t) > versus time at 0.5 ns ≤ t ≤ 3 ns were used to calculate CO2 diffusivity. In this time regime, d(log < ∆r 2 (t) >)/d log(t) was found to be close to 1 and CO2 is in the diffusion regime.

4 4.1

RESULTS AND DISCUSSION Optimized C3P Structure Obtained from Quantum AI Gas Phase Calculations

The optimized C3P structure, which contains 112 atoms, is shown in Figure 1. The bond lengths between Zn and the coordinated O, N, and Cl atoms are close to the experimental crystal structure data, 7 with differences between 0.07−0.1 ˚ A (Table 1). All calculated bond lengths and angles are also in good agreement with the experimental structure, with maximum difference of 0.1 ˚ A and 3.76◦ , respectively. The root-mean-square deviation for all atoms of the C3P molecule between the optimized structure and the experiment was found to be 0.86 ˚ A. All these results suggest that the AI gas phase calculations give optimized structure close to the experiment and serve as validation of our computational method.

8


Page 8 of 32

Page 9 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


4.2

Stability of C3P in Water and Aqueous MEA Solution

In order to determine C3P stability in solution, radial distribution (g(r)) was used to calculate the relative probability of atoms in proximity to each other. A large g(r) at small distance r indicate a stable configuration. The radial distributions for C3P (Zn−O, Zn−N, and Zn−Cl, Figure 2) in water are very sharp. The distances obtained from AIMD simulations at which g(r) exhibits the maximum peaks are: 1.99 ˚ A for Zn−O, 2.03 ˚ A for Zn−N, and A for Zn−Cl, close to the values obtained from AI gas phase calculations (Table 1). In 2.31 ˚ contrast, the g(r) between Zn and O atom of water exhibits a much broader shape. The first peak occurs at r = 3.8 ˚ A, much larger than the peak distance values between Zn and the coordinating O, N, and Cl atoms. During the whole simulation time of 60 ps, the Zn atom remains coordinated by O, N and Cl atoms. Water molecules are not coordinated with the Zn metal and do not displace O, N, and Cl atoms in the Zn coordination sphere. For C3P in aqueous MEA solution, simulations also indicate that the Zn atom is coordinated by O, N, and Cl atoms and not by water or MEA. These simulation results suggest that C3P is stable in water or aqueous MEA solution, and it is reasonable to add bond potentials for Zn−O, Zn−N, and Zn−Cl in classical FF simulations to maintain the Zn coordination. Note that although water coordination with metal is important for CO2 hydration reaction, it is not observed in this study. This is partly due to the fact that the pH value is not accounted for in this work, which may affect water coordination. Additionally, the AIMD simulation length may not be long enough to observe water coordination with metal.

4.3

Viscosity

Viscosity is an important factor to determine CO2 mass transfer coefficient KG ; the smaller solution viscosity, the larger KG . We have computationally investigated the effect of adding 2.77 and 27.7 g/L C3P in 30 wt% MEA aqueous solution on solution viscosity, and the simulation results were compared to the experimentally determined viscosities. The results show that in all cases, addition of C3P has little effect on the solution viscosity (Figure 3). 9



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 32

The simulated viscosity variation for MEA solution after adding C3P is comparable to the simulated viscosity uncertainty, indicating that variations are likely within the error of the simulation. The simulated viscosity for the 30 wt% MEA baseline solution was 1.22(±0.22)× 10−3 Pa.s, only slightly lower than the simulated value 1.50(±0.15) × 10−3 Pa.s for 2.77 g/L C3P in MEA. The viscosity difference between these two cases is 0.28×10−3 Pa.s, comparable to the simulation uncertainties (0.15−0.22 ×10−3 Pa.s). At a higher C3P concentration (27.7 g/L C3P in 30 wt% MEA), the viscosity was calculated to be 1.27(±0.12) × 10−3 Pa.s, close to the MEA baseline and lower C3P concentration viscosities. The experimental data are in line with our computational results, showing that addition of C3P in 30 wt% MEA solution only slightly increases the solution viscosity (Figure 3). Measured viscosities for solutions of 30% MEA containing 2.7 and 27.7 g/L C3P are 2.87 ×10−3 Pa.s and 3.12 ×10−3 Pa.s, a 3% and 12% increase, respectively, compared with the measured viscosity of 2.78 ×10−3 Pa.s for the MEA baseline. The simulated viscosities are all about 2 times smaller than the experimental data, but these differences between simulation and experiment are comparable to other simulation studies. 16,34 The classical FF parameters for H2 O, MEA, and C3P used in this work need to be improved to reproduce the experimental viscosities. In an effort to understand the slight increase in viscosity (experimental) upon addition of C3P in solution, the rotation relaxation time (τrlx,1 ) for the largest principle axis of water and MEA was calculated. The largest principal axis exhibits the slowest rotation and the largest τrlx,1 value, which in turn determines the solution viscosity. A larger τrlx,1 therefore corresponds to a larger viscosity. More calculation details for τrlx,1 can be found in our previous work. 16 Note that rotational relaxation is a surrogate for viscosity calculation as opposed to providing a reason for increase or decrease in the viscosity. The viscosity calculation involves the collective pressure tensor calculation. In contrast, the relaxation time calculation involves an average for individual molecules in the system. As a result, the relaxation time calculation is expected to exhibit a smaller uncertainty compared with the viscosity calculation. Due to this reason, we calculated the relaxation time to differentiate small viscosity

10


Page 11 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


differences between different systems. For 30 wt% MEA solution, the τrlx,1 values are 3.5 ps for water and 5.6 ps for MEA, respectively. When adding 2.77−27.7 g/L C3P in 30% MEA, the τrlx,1 values for water and MEA increase slightly by 6−9%. These small increases in τrlx,1 for water and MEA suggest that adding C3P will only slightly increase MEA solution viscosity, which is consistent with the experiment.

4.4

Surface Tension and C3P Concentration at the Gas-Liquid Interface

The simulated surface tensions for water, pure MEA, 30 wt% MEA aqueous solution, and the C3P-MEA solutions at 298 K are shown in Table 2. For pure water and MEA, the simulated surface tension values are comparable to the experimental data, 35 with differences of 14−26%. For 30 wt% MEA aqueous solutions w/o adding 2.77−27.7 g/L C3P, the simulated surface tensions are very close to the experimentally measured values (Table 2 and Figure 4), with small differences less than 2%. Both simulation and experiment suggest that adding C3P in the aqueous MEA solution does not affect the solution surface tension. Because of the small variations in surface tension, one tends to expect that the C3P molecules will be absorbed in the bulk aqueous MEA solution. On the contrary, our simulations show that the C3P molecules are preferentially adsorbed to the liquid surface. As shown in Figure 5, C3P molecules were initially distributed randomly in the bulk solution, and only one C3P molecule is located at the vapor-liquid interface. After 112 ns of NV T MD simulations, 6 out of 10 C3P molecules migrate to the interface (Figure 5). This preferential migration of catalyst molecules to the interface region was interesting, so another set of MD simulation was conducted where 10 C3P molecules were initially placed in an aggregated state at the center of simulation box (the bulk solution region). After 62 ns NV T MD simulations, the entire aggregate structure containing all 10 C3P molecules had migrated to the interface. These simulation results suggest that C3P is preferentially concentrated in the interface region instead of in the bulk solution, likely because the C3P molecule (Figure 1) 11



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

contains hydrophobic phenyl groups that may disfavor C3P interaction with H2 O and MEA and orient outward to the gas phase. Simulation snap shots indeed show that the phenyl groups point outward to the gas phase. On the other hand, the C3P molecule also contains electropositive Zn atoms and coordinated electronegative O, Cl, and N atoms (Figure 1); this polar part of C3P will favor interactions with the polar H2 O and MEA molecules. As a result, the C3P molecules are adsorbed at the liquid-vapor interface instead of escaping to the vapor phase. Finally, we calculated local density distributions for C3P, H2 O and MEA in different regions of the aqueous MEA solution. The C3P density at the interface region is ∼ 5 times larger than in the bulk solution region (Figure 6). The MEA exhibits slightly lower concentration at r ∼ ±39 ˚ A (the interface region) compared with bulk solution, while water exhibits slightly larger concentration at r ∼ ±39 ˚ A. These different C3P, water and MEA concentrations in the interface regions compared with the bulk solution may affect CO2 absorption. For example, it is expected that the concentrated C3P at the interface would increase CO2 hydration rate at the surface. Additionally, these concentrated C3P at the interface may affect CO2 diffusion from the gas phase to the bulk solution phase. 36 Having a lower MEA concentration at the interface also decreases CO2 reaction rate at the surface. Note that the slightly smaller MEA concentration and slightly larger water concentration at the vapor-liquid interface are in contrast to their respective concentrations around the C3P catalyst molecule in the bulk solution (see below).

4.5

CO2 Diffusivity

The CO2 diffusivity in 30 wt% MEA aqueous solution almost remains unchanged upon the addition of C3P (Figure 7), with a small difference of 3.2%. Although CO2 will react with the aqueous MEA solution, the chemical reaction is not included in CO2 diffusivity calculation. As a result, the simulated CO2 diffusivity represents only the CO2 physical diffusion in MEA solution before any chemical reaction occurs. 12


Page 12 of 32

Page 13 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


4.6 4.6.1

Structure Analysis C3P Aggregation in Aqueous MEA Solution

As shown in Figure 8, the C3P molecules tend to aggregate in aqueous MEA solution even though they are evenly dispersed in the solution at the start of simulation. This aggregation is partly due to the hydrophobic phenyl groups of C3P that disfavor interactions with the hydrophilic MEA and water molecules,and orient towards each other or the gas phase. The C3P molecules are aggregated to minimize the contact area between phenol and water and MEA and to minimize the free energy of the entire system. This aggregation may be expected to hinder the accessibility of some C3P molecules to CO2 or other reactants, and decrease the catalyst performance. This catalyst aggregation, combined with the possibility of surface blocking of CO2 diffusion, 36 is likely the source of decreased CO2 reaction rate when the C3P concentration is increased to 27.7 g/L. 4.6.2

CO2 , MEA and Water Local Concentrations Around C3P

The CO2 , MEA, and water concentrations in bulk aqueous MEA solution containing C3P are shown in Figure 9. As shown in Figure 9 (b) , CO2 and MEA exhibit maximum peak densities at r = 7.5 ˚ A, which is roughly the size of one C3P molecule (Figure 9 (a)). The A, the immediate area around the C3P molecule, is about 3 CO2 peak density at r = 7.5 ˚ times larger compared with the density in the rest of bulk aqueous MEA solution. Similarly, the MEA peak density at r = 7.5 ˚ A is about 1.3 times larger than in the rest of bulk solution. Additionally, CO2 and MEA also exhibit small density peaks close to the C3P center. In contrast, H2 O exhibits 1.2 times smaller density at r = 7.5 ˚ A compared with its density in the rest of bulk solution region. All these simulation results suggest that C3P tends to attract CO2 and MEA molecules while repelling water. The increased local CO2 and MEA concentrations around C3P may contribute, at least in part, to the observed enhanced CO2 reaction rates.

13



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

5

Page 14 of 32

CONCLUSIONS

The behavior of a CO2 hydration catalyst (C3P) was modeled computationally in aqueous MEA solution, and the results were verified experimentally where possible. Both simulations and experiments show that adding C3P catalyst in the MEA solution has little effects on the solution physical properties, such as viscosity, surface tension and CO2 diffusivity. Simulations do indicate that the local concentrations of MEA and CO2 around C3P are 1.3−3 times larger compared with their concentrations in the bulk aqueous MEA solution. In contrast, water concentration around C3P is 1.2 times smaller compared with its concentration in the bulk MEA solution, meaning the amine solvent and CO2 are artificially concentrated in the immediate environment around the catalyst. These findings further support assertion that the previously reported increase in CO2 mass transfer coefficient observed upon addition C3P in aqueous MEA solution is likely due to enhanced CO2 chemical reactions catalyzed by C3P and the increased concentrations of CO2 and MEA around C3P catalyst, not due to changes in the physical properties of the aqueous MEA solution. Simulations also show that C3P molecules tend to aggregate together in aqueous MEA solution, and are preferentially adsorbed at the liquid-vapor interface region. This aggregation is likely due to the hydrophobic phenyl groups of C3P, and may hinder the accessibility of some C3P catalyst molecules, possibly impeding the catalytic performance. This simulation finding is consistent with the experimental observation that CO2 mass transfer coefficient is not increased when the C3P catalyst concentration is increased beyond 2.3 g/L, and even decreases slightly when C3P concentration is increased by 10 times. It may be possible to improve C3P dispersion in aqueous MEA solution, and catalyst performance, by replacing the hydrophobic phenyl groups on C3P with a more hydrophilic functional group. The development of these computational methods may be used in future catalyst development by screening catalyst candidates computationally before undertaking costly synthetic work.

14


Page 15 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Supporting Information Available Classical force field parameters for C3P. This material is available free of charge via the Internet at http://pubs.acs.org/.

6

ACKNOWLEDGEMENT

This technical effort was performed in support of the National Energy Technology Laboratory’s on-going research in computational chemistry under the RES contract DE-FE0004000. This project was funded by the Department of Energy, National Energy Technology Laboratory, an agency of the United States Government, through a support contract with AECOM. Neither the United States Government nor any agency thereof, nor any of their employees, nor AECOM, nor any of their employees, makes any warranty, expressed or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.

References (1) Rogelj, J.; den Elzen, M.; Hoehne, N.; Fransen, T.; Fekete, H.; Winkler, H.; Chaeffer, R. S.; Sha, F.; Riahi, K.; Meinshausen, M. Paris Agreement Climate Proposals Need a Boost to Keep Warming Well Below 2 Degrees C. Nature 2016, 534, 631–639.

15



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 32

(2) Desideri, U.; Corbelli, R. CO2 Capture in Small Size Cogeneration Plants: Technical and Economical Considerations. Energy Conv. Manag. 1998, 39, 857–867. (3) Rao, A.; Rubin, E. A Technical, Economic, and Environmental Assessment of AmineBased CO2 Capture Technology for Power Plant Greenhouse Gas Control. Environ. Sci. Technol. 2002, 36, 4467–4475. (4) Parson, E.; Keith, D. Climate Change - Fossil Fuels without CO2 Emissions. Science 1998, 282, 1053–1054. (5) Onda, K.; Takeuchi, H.; Okumoto, Y. Mass Transfer Coefficients Between Gas And Liquid Phases in Packed Columns. J. Chem. Eng. Jpn 1968, 1, 56–62. (6) Onda, K.; Sada, E.; Takeuchi, H. Gas Absorption With Chemical Reaction in Packed Columns. J. Chem. Eng. Jpn 1968, 1, 62–66. (7) Lippert, C. A.; Liu, K.; Sarma, M.; Parkin, S. R.; Remias, J. E.; Brandewie, C. M.; Odom, S. A.; Liu, K. Improving Carbon Capture from Power Plant Emissions with Zinc- and Cobalt-based Catalysts. Catal. Sci. Technol. 2014, 4, 3620–3625. (8) Lippert, C. A.; Widger, L. R.; Sarma, M.; Liu, K. Catalyst Development for Rate Enhanced Acid Gas (CO2 ) Scrubbing. Energy Procedia 2014, 63, 273–278. (9) Kelsey, R. A.; Miller, D. A.; Parkin, S. R.; Liu, K.; Remias, J. E.; Yang, Y.; Lightstone, F. C.; Liu, K.; Lippert, C. A.; Odom, S. A. Carbonic Anhydrase Mimics for Enhanced CO2 Absorption in An Amine-based Capture Solvent. Dalton Trans. 2016, 45, 324–333. (10) Widger, L. R.; Sarma, M.; Bryant, J. J.; Mannel, D. S.; Thompson, J. G.; Lippert, C. A.; Liu, K. Enhancements in Mass Transfer for Carbon Capture Solvents Part I: Homogeneous Catalyst. Int. J. Greenh. Gas Control 2017, 63, 249–259.

16


Page 17 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(11) Allen, M. P.; Tildesley, D. J. Computer Simulation of Liquids; Clarendon: Oxford, 1987. (12) Shi, W.; Maginn, E. J. Atomistic Simulation of the Absorption of Carbon Dioxide and Water in the Ionic Liquid 1-n-hexyl-3-methylimidazolium bis(trifluoromethylsulfonyl) imide ([hmim][Tf2 N]). J. Phys. Chem. B 2008, 112, 2045–2055. (13) Hwang, G. S.; Stowe, H. M.; Paek, E.; Manogaran, D. Reaction Mechanisms of Aqueous Monoethanolamine With Carbon Dioxide: A Combined Quantum Chemical And Molecular Dynamics Study. Phys. Chem. Chem. Phys. 2015, 17, 831–839. (14) Amirjalayer, S.; Tafipolsky, M.; Schmid, R. Molecular dynamics simulation of benzene diffusion in MOF-5: Importance of lattice dynamics. Angew. Chem.-Int. Edit. 2007, 46, 463–466. (15) Shi, W.; Sorescu, D. C.; Luebke, D. R.; Keller, M. J.; Wickramanayake, S. Molecular Simulations and Experimental Studies of Solubility and Diffusivity for Pure and Mixed Gases of H2 , CO2 , and Ar Absorbed in the Ionic Liquid 1-n-Hexyl-3-methylimidazolium Bis(Trifluoromethylsulfonyl)amide ([hmim][Tf2 N]). J. Phys. Chem. B 2010, 114, 6531– 6541. (16) Shi, W.; Siefert, N. S.; Baled, H. O.; Steckel, J. A.; Hopkinson, D. P. Molecular Simulations of the Thermophysical Properties of Polyethylene Glycol Siloxane (PEGS) Solvent for Precombustion CO2 Capture. J. Phys. Chem. C 2016, 120, 20158–20169. (17) Phillips, J. C.; Braun, R.; Wang, W.; Gumbart, J.; Tajkhorshid, E.; Villa, E.; Chipot, C.; Skeel, R. D.; Kale, L.; Schulten, K. Scalable Molecular Dynamics with NAMD. J. Comput. Chem. 2005, 26, 1781–1802. (18) Essmann, U.; Perera, L.; Berkowitz, M.; Darden, T.; Lee, H.; Pedersen, L. A Smooth Particle Mesh Ewald Method. J. Chem. Phys. 1995, 103, 8577–8593.

17



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 32

(19) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09, Revision A.1. Gaussian, Inc., Wallingford, CT, 2009. (20) Breneman, C. M.; Wiberg, K. B. Determining Atom-centered Monopoles from Molecular Electrostatic Potentials. The Need for High Sampling Density in Formamide Conformational Analysis. J. Comput. Chem. 1990, 11, 361–373. (21) Elements, Atomic Radii and the Periodic Table. http://crystalmaker.com/support/ tutorials/crystalmaker/atomic-radii/index.html, Accessed: 2016-08-26. (22) Car, R.; Parrinello, M. Unified Approach for Molecular Dynamics and DensityFunctional Theory. Phys. Rev. Lett. 1985, 55, 2471–2474. (23) CPMD, http://www.cpmd.org/. Copyright IBM Corp 1990-2008, Copyright MPI f¨ ur Festkörperforschung Stuttgart 1997-2001. (24) Grimme, S. Semiempirical GGA-type density functional constructed with a long-range dispersion correction. J. Comput. Chem. 2006, 27, 1787–1799.

18


Page 19 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(25) Shi, W.; Myers, C. R.; Luebke, D. R.; Steckel, J. A.; Sorescu, D. C. Theoretical and Experimental Studies of CO2 and H2 Separation Using the 1-Ethyl-3-methylimidazolium Acetate ([emim][CH3 COO]) Ionic Liquid. J. Phys. Chem. B 2012, 116, 283–295. (26) Rowlinson, J. S.; Widom, B. Molecular Theory of Capillarity; Clarendon: Oxford, 1982. (27) Kirkwood, J.; Buff, F. The Statistical Mechanical Theory Of Surface Tension. J. Chem. Phys. 1949, 17, 338–343. (28) Irving, J.; Kirkwood, J. The Statistical Mechanical Theory Of Transport Processes .4. The Equations Of Hydrodynamics. J. Chem. Phys. 1950, 18, 817–829. (29) Walton, J.; Tildesley, D.; Rowlinson, J.; Henderson, J. The Pressure Tensor at The Planar Surface of A Liquid. Mol. Phys. 1983, 48, 1357–1368. (30) Gloor, G.; Jackson, G.; Blas, F.; de Miguel, E. Test-area Simulation Method for the Direct Determination of the Interfacial Tension of Systems with Continuous or Discontinuous Potentials. J. Chem. Phys. 2005, 123, 134703. (31) Shi, W.; Siefert, N. S.; Morreale, B. D. Molecular Simulations of CO2 , H2 , H2 O, and H2 S Gas Absorption into Hydrophobic Poly(dimethylsiloxane) (PDMS) Solvent: Solubility and Surface Tension. J. Phys. Chem. C 2015, 119, 19253–19265. (32) Fennell, C. J.; Gezelter, J. D. Is the Ewald Summation Still Necessary? Pairwise Alternatives to the Accepted Standard for Long-Range Electrostatics. J. Chem. Phys. 2006, 124, 234104. (33) Shi, W.; Sorescu, D. C. Molecular Simulations of CO2 and H2 Sorption into Ionic Liquid 1-n-Hexyl-3- methylimidazolium Bis(trifluoromethylsulfonyl)amide ([hmim][Tf2 N]) Confined in Carbon Nanotubes. J. Phys. Chem. B 2010, 114, 15029–15041. (34) Yan, T. Y.; Burnham, C. J.; Pópolo, M. G. D.; Voth, G. A. Molecular dynamics

19



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

simulation of ionic liquids: The effect of electronic polarizability. J. Phys. Chem. B 2004, 108, 11877–11881. (35) Vázquez, G.; Alvarez, E.; Navaza, J.; Rendo, R.; Romero, E. Surface Tension of Binary Mixtures of Water Plus Monoethanolamine And Water Plus 2-Amino-2-methyl-1propanol And Tertiary Mixtures of These Amines With Water From 25 Degrees C To 50 Degrees C. J. Chem. Eng. Data 1997, 42, 57–59. (36) Bryant, J. J.; Lippert, C.; Qi, G.; Liu, K.; Mannel, D. S.; Liu, K. Enhanced Carbon Capture through Incorporation of Surfactant Additives. Ind. Eng. Chem. Res. 2016, 55, 7456–7461.

20


Page 20 of 32

Page 21 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Table 1: The bond lengths between Zn and the coordinated Cl, O, and N atoms of C3P obtained from quantum ab initio (AI) gas phase optimization calculations. The experimental (exp.) data 7 obtained from the C3P crystal structure are also shown for comparison. bond Zn−O Zn−N Zn−Cl

˚) difference (˚ AI (˚ A) exp. (A A) 2.031 1.941 0.09 2.144 2.043 0.101 2.312 2.245 0.067

21



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 32

Table 2: Simulated surface tensions (unit in 10−3 N/m) at 298 K for neat water, MEA, 30 wt% aqueous MEA solution, and the 30 wt% aqueous MEA solution systems containing 2.77 g/L and 27.7 g/L of C3P. The total simulated surface tension is given by γsim,total = γsim +γtail . Simulated uncertainties in the last digit are given in parentheses. For comparison, also shown are the experimental surface tension values for neat water and neat MEA, 35 30 wt% aqueous MEA solution and the C3P-MEA aqueous solution systems obtained in this work. system neat water neat MEA 30 wt% MEA C3P(2.77 g/L)-30 wt% MEA C3P(27.7 g/L)-30 wt% MEA

simulated surface tension γsim γtail γsim,total 59.2 (3) 2.8 62.0 (3) 53.3 (4) 8.4 61.7 (4) 57.4 (5) 5.1 62.5 (5) 58.0 (4) 5.0 63.0 (4) 57.9 (6) 5.0 62.9 (6)

22


γexp 72.01 48.95 63.36 64.08 63.76

Page 23 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 1: The optimized C3P structure obtained from quantum ab initio gas-phase calculations. Also shown are the atomic partial charges for Zn, Cl, O, N, and P atoms obtained from quantum calculations by using the CHELPG protocol. 20 The C3P molecule consists of two methylphenol groups, two triphenylphosphane (P(C6 H5 )3 ) groups, two -CH2 groups, two -CH groups, and two Zn metals, each of which is coordinated by two Cl atoms, one N, and one O atom. The C3P molecule contains 112 atoms.

23



Zn-O/C3P Zn-N/C3P Zn-Cl/C3P Zn-O/H2O (X100)

1000 g(r)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 32

500

0

2

4

6

8

r (Å) Figure 2: The radial distributions (g(r)) obtained from ab initio molecular dynamics simulations for the system of one C3P molecule solved in 230 water molecules at 298 K. For clarity, the g(r) values between Zn and the O atom of water are enlarged by a factor of 100.

24


Page 25 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 3: The simulated viscosity for 30 wt% aqueous MEA solution at 298 K and 1 bar before and after adding C3P. The simulated viscosity and error bar (vertical line) were obtained from three independent simulations. For comparison, the experimental viscosity values obtained in this work are also shown.

25



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 4: The simulated surface tension for 30 wt% aqueous MEA solution at 298 K before and after adding 2.77 g/L and 27.7 g/L of C3P. The simulation error bar (vertical line) was estimated from the standard block average calculation. For comparison, the experimental surface tension values obtained in this work are also shown.

26


Page 26 of 32

Page 27 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 5: Two representative snap shots obtained from the NV T molecular dynamics simulation of vapor-liquid equilibrium, which is used in surface tension calculation. The simulation was performed at 298 K and the simulation system contains 10 C3P (red VDW spheres), 2000 MEA (white lines), and 15800 water molecules (silver lines). Figure (a) indicates the starting configuration (t = 0 ns) and figure (b) corresponds to the simulation at t = 112 ns. The simulation boxes (blue lines) are also shown.

27


Page 28 of 32

3

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

local mass density (g/cm ) 0 0.2 0.4 0.6 0.8


water MEA C3P catalyst

liquid

-50

0 Z (Å)

interface gas

50

Figure 6: Local density distributions for MEA, water, and C3P obtained from the NV T molecular dynamics simulation of vapor-liquid equilibrium. One interface region, which separates the gas and bulk liquid regions, is denoted by two vertical dashed lines. The simulation system is the same as in Figure 5.

28


Page 29 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 7: Simulated CO2 physical diffusivity values in 30 wt% aqueous MEA solution before and after adding C3P. The simulation error bar (vertical line) was estimated from the standard block average calculation. Simulations were performed at 298 K and they correspond to low CO2 mole fractions of ∼ 0.01. CO2 chemical reactions in aqueous MEA solution are not included in diffusivity calculations.

29



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 8: Two representative snap shots obtained from the NP T molecular dynamics simulation at 298 K and 1 bar for 30 wt% aqueous MEA solution containing C3P. The system consists of 10 C3P, 2000 MEA, and 15800 water molecules. For clarity, only the C3P molecules are shown. Figure (a) indicates the starting configuration at t = 0 ns, in which all 10 C3P molecules are randomly dispersed in the solution. Figure (b) corresponds to the simulation at t = 160 ns, for which all 10 C3P molecules are aggregated in the aqueous MEA solution. The simulation boxes (red lines) are also shown.

30


Page 30 of 32

Page 31 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 9: Figure (a): A snap shot from molecular dynamics simulation to show water, MEA, and CO2 absorption near the C3P molecule at 298 K and 1 bar. The simulation system contains 1 C3P, 15800 water, 2000 MEA, and 200 CO2 . For clarity, only the nearby water, MEA, and CO2 molecules are shown. A circler is drawn around the C3P to roughly indicate its size. Figure (b): The local density distributions for water, MEA, and CO2 . The CO2 density was enlarged by a factor of 10 to get a better view.

31



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Table of content. The snap shot from molecular simulation to show C3P molecules are preferentially adsorbed at the gas-liquid interface regions. 102x135mm (96 x 96 DPI)


Page 32 of 32

Molecular Modeling of the Physical Properties for Aqueous Amine

Recommend Documents