Enhanced Selectivity for CO2 Adsorption on Mesoporous Silica with

Samsung Advanced Institute of Technology, Samsung Electronics Co., Ltd., 130 Samsung-ro, Yeongtong-gu, Suwon-si, Gyeonggi-do 443-803, Republic of Kore...
1 downloads 14 Views 1MB Size
Subscriber access provided by UNIVERSITY OF ADELAIDE LIBRARIES

Article

Enhanced Selectivity for CO2 Adsorption on Mesoporous Silica with Alkali Metal Halide due to Electrostatic Field: A Molecular Simulation Approach Soonchul Kwon, Hyuk Jae Kwon, Ji Il Choi, Ki Chul Kim, Jeong Gil Seo, Jung Eun Park, Su Jin You, Eun Duck Park, Seung Soon Jang, and Hyun Chul Lee ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.7b04508 • Publication Date (Web): 22 Aug 2017 Downloaded from http://pubs.acs.org on August 22, 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.

ACS Applied Materials & Interfaces 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 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

ACS Applied Materials & Interfaces

Enhanced Selectivity for CO2 Adsorption on Mesoporous Silica with Alkali Metal Halide due to Electrostatic Field: A Molecular Simulation Approach

Soonchul Kwon,†,‡ Hyuk Jae Kwon,† Ji Il Choi,§ Ki Chul Kim,§ Jeong Gil Seo,†,∈ Jung Eun

Park∫, Su Jin You∫, Eun Duck Park∫, Seung Soon Jang, §,* Hyun Chul Lee†,*

† Samsung

Advanced Institute of Technology, Samsung Electronics Co., Ltd., 130 Samsung-ro,

Yeongtong-gu, Suwon-si, Gyeonggi-do 443-803, Republic of Korea ‡

Department of Civil and Environmental Engineering, Pusan National University, 2,

Busandaehak-ro 63beon-gil, Geumjeong-gu, Busan 46241, Republic of Korea §

School of Materials Science and Engineering, Georgia Institute of Technology, Atlanta, GA

30332-0245, USA ∈ Department of Environmental Engineering and Energy, Myongji University, 116 Myongji-ro, Cheoin-gu, Yongin-si, Gyeonggi-do 449-728, Republic of Korea ∫ Department of Energy Systems Research and Department of Chemical Engineering, Ajou University, 206, Worldcup-ro, Yeongtong-gu, Suwon 443-749, Republic of Korea

ACS Paragon Plus Environment

1

ACS Applied Materials & Interfaces

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 2 of 29

ABSTRACT

Since adsorption performances are dominantly determined by adsorbate–adsorbent interactions, accurate theoretical prediction of the thermodynamic characteristics of gas adsorption is critical for designing new sorbent materials as well as understanding the adsorption mechanisms. Here, through our molecular modeling approach using newly-developed quantum-mechanics-based force field, it is demonstrated that the CO2 adsorption selectivity of SBA-15 can be enhanced by incorporating crystalline potassium chloride particles. It is noted that the induced intensive electrostatic fields around potassium chloride clusters create gas trapping sites with high selectivity for CO2 adsorption. The newly developed force field can provide a reliable theoretical tool in accurately evaluating the gas adsorption on given adsorbents, which can be utilized to identify good gas adsorbents.

KEYWORDS: CO2 adsorption, density functional theory, molecular dynamics, gas trap, mesoporous silica

1. INTRODUCTION The atmospheric concentration of carbon dioxide (CO2) has been steadily increasing since the start of the industrial revolution, which is a main contributor to challenging climate change issues.1 Over 80% of the anthropogenic CO2 emissions in the world are generated through energy production, and in particular, around 40% of the total CO2 emissions are released from coal-fired power plants.1-2 Consequently, immense effort has been devoted to the capture and

ACS Paragon Plus Environment

2

Page 3 of 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

ACS Applied Materials & Interfaces

sequestration of CO2 during energy-production processes. Among various technologies developed for this purpose, the scrubbing process using amine solution is the only technology commercially used nowadays.3 However, there are known drawbacks such as the reduction of CO2 capacity due to the contaminants in the gas stream, and the intensive energy consumption for regeneration of used amine solution.4,5 To overcome these drawbacks, alternative technologies such as membrane based separation method and porous material based sorption method have been considered.6-8 The adsorption on porous solid materials is a promising technique because of the benefits such as energy effectiveness for regeneration, high efficiency and high selectivity of CO2 capture, and alleviation of equipment corrosion problems.6-8 To date, various porous solid materials such as carbon materials, zeolites, metal-organic frameworks (MOFs), zeolitic imidazolate frameworks (ZIFs), hydrotalcite-like materials (HTCs) and mesoporous silica materials such as MCM-41 and SBA-15 have been studied as adsorbents for CO2 capture technologies.4-13 While carbons, zeolites, MOFs, ZIFs, and HTCs have been intensively studied,9-12 mesoporous silica nanoparticles have not been thoroughly investigated although they are representatively known as a high adsorptive material, owing to their high surface area and high active-site density (e.g., silanol groups). Wang and Yang reported that the CO2 adsorption capacity of amine-grafted SBA-15 is increased as a function of amine graft density owing to the strong interactions between CO2 and amine groups where SBA-15 provides many sites for amine loading.13 Meanwhile, computational studies of adsorption-based CO2 capture have been widely used to mechanistically scrutinize experimental results and to assist in designing better adsorbent candidates.14-18 As mentioned in these computational studies, the grand canonical Monte Carlo (GCMC) method is a reliable method for the simulation of CO2 adsorption in order to screen candidate adsorbent materials. Since GCMC simulations are

ACS Paragon Plus Environment

3

ACS Applied Materials & Interfaces

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 4 of 29

performed using potential energy functions (force field), it should be emphasized that a highquality force field for the description of the adsorptive characteristics is an essential prerequisite for successful investigation of the CO2 adsorption, confirmed by experimental data: in the studies of adsorptive behaviors of CO2 or H2 molecules in various MOFs,17,18 it was demonstrated that a specifically developed force fields for specific systems provides better description in comparison with generic force fields such as universal force field (UFF) and Dreiding.19,20 In this study, we investigated the effects of alkali-metal chlorides on the selectivity of CO2 adsorption on ordered siliceous SBA-15 within a simulation-experiment collaborative approach to verify the developed force field. In order to achieve fundamental understanding of adsorption property at atomic level, density functional theory (DFT) calculations were initially utilized to provide a comprehensive analysis of such CO2, N2, H2, CH4 binding properties to adsorbents. Then, multi-scale simulation approach consisting of DFT for developing force field and GCMC for the molecular simulation of gas adsorption enabled us to accurately describe the adsorption properties of various gas molecules such as CO2, CH4, N2, H2, O2, H2S, NO, and NO2 on the surface of bare SBA-15 and KCl incorporated SBA-15 (KCl/SBA-15) through using a newlydeveloped quantum-mechanics-based Dreiding force field for molecular dynamics, derived by DFT calculation with dispersion corrections. These approaches, in turn, allow a comprehensive analysis of the favorable molecular interactions of KCl/SBA-15 with gas molecules, as compared to bare SBA-15 at the atomistic scale. This was accomplished by simulation of KCl clusters over SBA-15, which investigates the energy-minimized geometry of the morphology and structure, as well as electronic properties of the atomistic units. Theoretical studies suggest the nature of inorganic materials influences the attractive intermolecular interactions with gas molecules to capture undesirable pollutant gases. The MD simulation was performed using the newly

ACS Paragon Plus Environment

4

Page 5 of 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

ACS Applied Materials & Interfaces

developed force field to investigate extensively binary gas adsorption (e.g., CO2/CH4, CO2/N2, and CO2/H2) on adsorbents, especially, focusing on the effect of electrostatic field generated by the atomic charges of the adsorbent surface on the adsorption mechanism.

2. MATERIALS AND METHODS 2.1. Syntheses of alkali-metal/SBA-15s. The alkali-metal chloride/SBA-15 samples were prepared in an aqueous solution with EO20PO70EO20 block-copolymer (Pluronic P123, SigmaAldrich) as a structure-directing agent. For the synthesis, P123 (10.0 g) was dissolved in deionized water (320 ml) and hydrochloric acid (60 ml). To this mixture, tetraethylorthosilicate (TEOS) was quickly added (22 ml) under stirring at 313 K. The alkali-metal precursor (e.g., LiCl, NaCl, KCl, or KOH) was then introduced at a 2:1 molar ratio in alkali-metal/Si. The mixture was aged at 313 K for 24 h, and subsequently heated for 24 h at 353 K under static conditions. The solid complex was then filtered and calcined at 823 K for 5 h to remove the template (Figure S1). Table S1 lists the physico-chemical properties of alkali-metal chloride/SBA-15s to determine the alkali-metal composition and structure information. To determine the degree of alkali-metal loading over the SBA-15 support, we characterized the adsorbent composition using inductively coupled plasma atomic emission spectroscopy (ICPAES). Though alkali-metal precursors were introduced during the synthesis at a 2:1 molar ratio of alkali-metal/Si, the amount of loaded alkali-metal in the resultants decreased from K > Na> Li, as shown in Table S1. 2.2. CO2 isotherm experiment. To investigate the adsorptive capacity of CO2, the adsorption isotherm experiment is performed at 298 K on a volumetric apparatus by BEL SORP-mini (BEL Japan). The prepared samples were degassed at 623 K for 12 h before measurements.

ACS Paragon Plus Environment

5

ACS Applied Materials & Interfaces

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 29

2.3. Characterization. The adsorbent samples were characterized with several structural and morphological methods. The bright-field transmission electronic microscopy (TEM) image was obtained using a Tecnai G2 TEM (FEI) operated at 200 kV (Figure S2). For element mapping on the surface, energy dispersive spectroscopy (EDS) with TEM was performed (Table S2). Bulk crystalline structures of the catalysts were determined with X-ray diffraction (XRD) (Figure S3). The XRD patterns of small-angle X-ray scattering (SAXS) were obtained with Cu K radiation (40 kV) on a Phillips XPert PRO instrument at room temperature (Figure S4B). We then analyzed the physical and chemical properties of the adsorbents. The BET surface area and the N2 adsorption/desorption isotherm data are obtained at 77 K on a volumetric apparatus by BEL SORP-Max (BEL Japan) (Figure S4A). The prepared samples were degassed at 473 K for 24 h before measurements. The pore size distributions were determined from the adsorption isotherm using the BJH (Barrett-Joyner-Halenda) model with cylindrical pore geometry, and the pore volumes were taken at P/P0 > 0.97. The external surface area of the adsorbents was calculated using the t-plot method.21,22 Internal surface area was calculated using the subtraction of total surface to external surface area. The concentrations of Si, K, and Cl were determined by ICPAES on a Shimadzu ICPS-8100 sequential spectrometer (Table S1). The element surface distribution was investigated by X-ray photoelectron spectroscopy (XPS) using Quantum 2000 (Physical Electronics) (Figure S5). The Fourier transform infrared spectroscopy (FT-IR) measurement was performed using FTS-6000&UMA600 (Bio Rad) (Figure S6A). The

29

Si

magic-angle spinning nuclear magnetic resonance (MAS NMR) measurement was performed using Bruker AVANCE III 600 at 600 MHz (Figure S6B). The spinning rate was 4 kHz, and the repetition delay and pulse width used for the measurements were 60 s and 3 µs, respectively.

ACS Paragon Plus Environment

6

Page 7 of 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

ACS Applied Materials & Interfaces

2.4. DFT calculation. For quantum mechanical computations, the DFT calculations are used to determine adsorption energies and electronic properties of the interaction between the adsorbed CO2 molecules and the alkali-metal chloride/SBA-15 surface in three-dimensional periodic-slab models. The geometry optimizations were performed using the DMol3 module of Materials Studio software23. For calculating the functional options, the gradient-corrected approximation of the exchange-correlation functional, the generalized gradient approximation (GGA) with nonempirical, local functionals, and the Perdew-Burke-Ernzerhof correlation (PBE)24,25 were used with the double numerical basis set with polarization (DNP)26,27 to develop the electron exchange-correlation energy of the interaction. For the computation of the charge and geometry of an individual gas molecule, we used the popular B3LYP functional and 6-31G** basis set, showing that each gas molecule is neutral charge, as it is implemented in the DFT tool Jaguar.28 (See Fig. S10 and Table S3 in supplementary information) The determined charges are assumed to remain constant in the force field during molecular simulations. We performed an energy correction of the non-bonded interaction, described by the van der Waals interaction for a dispersive interaction. Conventional DFT fails to correctly describe the R−6 asymptotic potential for dispersive intermolecular interactions. However, the recently developed DFT with dispersion corrections (DFT-D3) can correct dispersive interactions.29,30 The London dispersion interaction, or the dispersion interaction, can be empirically defined as the attractive part of a van der Waals (vdW)-type interaction between atoms and molecules. To achieve chemical accuracy, the inclusion of the interactions in theoretical simulations is indispensable. Therefore, in this study, we apply DFT-D3 with Becke-Johnson (BJ) damping function for the better description of all inter-atomic interactions including KCl/SBA-15 surfaces. The corrected atomic binding energies

ACS Paragon Plus Environment

7

ACS Applied Materials & Interfaces

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 8 of 29

between the gas molecules and surfaces (KCl, SiO2) are listed in Table S4. More details are described in Supplementary Information. 2.5. Development of the force field. The force field is a set of parameters for a given equation used to calculate the energy and force acting on atoms. The contributing energy terms in the force field are represented by unique potential functions with configuring parameters, as shown in equation (1). For this, we utilize the format of DREIDING force field20:

E total = E vdW + E Q + E bond + E angle + E torsion + E inversion

(1)

where Etotal, Evdw, EQ, Ebond, Eangle, Etorsion, and Einversion are the energies for the total, van der Waals, electrostatic, bond stretching, angle bending, torsion, and inversion components, respectively. The potential energies could be utilized for all molecular interactions regardless of the specific molecular components and structures. In particular, non-bonded interaction energy includes contributions from both the van der Waals and coulombic pair interactions. The Morse function was used to describe the van der Waals interaction of CO2 with KCl/SBA15 surfaces. All other van der Waals interactions were described using the LJ potential with geometric mixing rule, and the covalent bond interactions (Ebond in equation (1)) were described by harmonic potential functions employed in Dreding force field. In the Dreding force field, the van der Waals pair potential (UvdW(rij)) is given by 



  =     





− 2    

(2)

where ϵij is the magnitude of the minimum well depth, σij denotes the minimum energy of interaction between the two atoms, and rij indicates the distance between the nuclei of the two atoms.

ACS Paragon Plus Environment

8

Page 9 of 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

ACS Applied Materials & Interfaces

In general, a classical force field is comprised of Morse potential parameters.31 We adopt the Morse potential for force field calculations of the van der Waals forces between the atoms with the same force field type that is defined by diagonal van der Waals interaction, and between hybrid atom types that includes the same atomic species with different force field type, defined by the off-diagonal van der Waals. The Morse potential function is used to newly optimize the off-diagonal van der Waals parameters between the gas molecules and surfaces. Furthermore, non-covalently bonded atoms are described by Morse potential:   = !1 − # $%$&



(3)

where D is the depth of the potential, α is the width of the potential, and re is the equilibrium distance. The Morse potential function is used to newly optimize the off-diagonal van der Waals parameters between the gas molecules and surfaces, non-covalently bonded interaction, and Coloumbic pair interactions to improve the accuracy of simulation. The obtained inter-atomic binding energy variation from the DFT calculation can be utilized in the classical molecular dynamics simulations by projecting the DFT binding energy into the van der Waals energy (EvdW) and/or electrostatic interaction (EQ) as the function of distance. The electrostatic energy is determined by separate calculation and analysis using Mulliken analysis, and the EQ is calculated using the pairwise Coulombic potential in short-range and the particle-particle particle-mesh (PPPM) method in long-range interactions. Since EQ was calculated using Coulombic potential function and PPPM method, causing the difference between DFT energy (EDFT) and EQ, we developed new parameters for EvdW in order to make EvdW= EDFT - EQ. The DFT binding energy is decomposed to electrostatic and vdW interactions. Since the electrostatic interaction is not dependent on force field, we calculated the electrostatic interaction first. Then the Morse potential energy was determined to fill up the gap between total DFT

ACS Paragon Plus Environment

9

ACS Applied Materials & Interfaces

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 29

energy and the electrostatic interaction. In other words, the sum of electrostatic and Morse vdW potential energy is the same with the DFT binding energy. All other inter-molecular interactions are described by the total energy terms in equation (1). The off-diagonal van der Waals interactions are plotted for both the KCl and SiO2 surfaces. The geometry of each gas molecule was optimized to produce the maximal binding energy by DFT calculation. From the optimized position, each molecule was moved upward and downward and their geometry was optimized to calculate the binding energy at each position. After including the dispersive intermolecular interaction correction, the total binding energy can be re-written as '(()*_,--. = '/01_,--. + '3-_44(-

(4)

where EDFT_binding is the binding energy determined by DFT computations and Edispersion_correction indicates the correction of dispersive interactions. Etotal_binding, obtained by DFT calculation, was used to fit the Morse potential energy, EMorse to make the fitting energy, EFitting in the equation (5). The fitting energy is written as E0((-. = E + E*4(()(4_-()4(-

(5)

where EMorse is the Morse potential energy, and the binding energy by electrostatic interaction is given by Eelectrostatic_interaction. To obtain the total binding energy, the Morse potential parameters in equation (3) are determined by fitting equation (5). The optimized DFT-based force fields with vdW parameters between the gas molecules and surface, or the improved DREIDING force field (IDFF), were used to calculate the total energy. To validate the developed IDFF, Figure S14 and S15 show the plots of the binding energy variation as a function of distance on the KCl and SiO2 surfaces for various gas molecules (CO2, CH4, N2, H2, O2, H2S, NO, and NO2), respectively, in comparison to those obtained from quantum mechanical calculations. The binding energy profiles of IDFF obtained via Cerius232 or Forcite-plus33 are in good agreement with those

ACS Paragon Plus Environment

10

Page 11 of 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

ACS Applied Materials & Interfaces

obtained via DFT-D3 calculations. Different molecules may contain the same elements in their components (e.g. nitrogen in NO and NO2), but each of these elements may have a different partial charge and show different reactivity when binding to the surface. Thus, the elements should be individually identified in the force field. In the IDFF, we are able to manually assign a specific “type” to the individual elements in each molecule (Table S5). For the hydroxideterminated SiO2 surface, the hydroxide is also given a specific type. Using the typing in Table S5, the fitted force field parameters (coefficients of equation (3)) are listed in Tables S6 and S7, where γ = 2αγe. Such newly optimized off-diagonal van der Waals parameters between the gas molecules and surfaces could be used to calculate the corresponding energy. To develop the force field, we performed following tasks: (i) determine an energy function describing the molecular motion, (ii) optimize parameters for the potential function, (iii) compare the adsorption properties obtained by the force field using MD simulations with an affordable DFT calculation to evaluate the force field, and (iv) if the comparison gives acceptable agreement, the force field will be used for large-scale MD simulation; otherwise, the procedure is repeated from step (ii). 2.6. Molecular dynamics (MD) simulations. MD computations provide data that can be directly compared with experimental observations such as the adsorption isotherm measured at various thermodynamic conditions (temperature and pressure). As mentioned above, the MD simulations require accurate force fields to describe the interaction between CO2 and metal oxide surface. Because the accuracy of the force field determines the success of the simulation, we developed a DFT-based force field. The MD simulations were performed using the MD code LAMMPS.34 The MD simulations of pure CO2 molecules or CO2 molecules balanced with N2, H2, and CH4 molecules over the surfaces of SBA-15 and KCl/SBA-15 were shown in Figure 5.

ACS Paragon Plus Environment

11

ACS Applied Materials & Interfaces

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 12 of 29

3. RESULT AND DISCUSSION 3.1. Structure design of adsorbent. We synthesized SBA-15 incorporated with alkali-metal chloride nanoparticles LiCl, NaCl, and KCl (A/SBA-15s) (Figure S1 to S7, and tables S1 and S2)35 in which the SBA-15 provides a highly ordered pore structure with vast amount of surface area and excellent thermal stability. In order to investigate CO2 adsorption on such systems using DFT method, we also constructed atomistic model systems of SBA-15 (20 Å × 20 Å × 100 Å) and A/SBA-15s (Figure 1A insets and Figure S8). Interestingly, it turns out that the adsorption of CO2 on the A/SBA-15s is not strong (less than −0.3 eV for all adsorbents in Figure 1A) enough to form stable complexes (carbonates). Note that the reported energies of chemisorbed CO2 on alkali-metal oxides such as MgO, CaO, and BaO exceed −0.5 eV typically.36 Additional examination of the electronic properties of CO2-adsorbent systems using Mulliken population analysis37 verifies that KCl/SBA-15 exhibited a relatively large charge transfer (−0.43e; inset of Figure 1A) compared to NaCl/SBA-15 (−0.11e) and LiCl/SBA-15 (−0.08e) and bare SBA-15 (−0.08e).

ACS Paragon Plus Environment

12

Page 13 of 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

ACS Applied Materials & Interfaces

A

LiCl/SBA-15 NaCl/SBA-15 KCl/SBA-15

B

C ~1 0 nm

50n m

D

5n m

E

Figure 1. Alkali-metal chloride incorporated SBA-15 adsorbent. (A) Integrated CO2 adsorption profile from DFT calculations along with TEM images. C.T. denotes the charge transfer from the adsorbent (alkali-metal chloride) to the CO2 molecule. (B, C) TEM images of KCl/SBA-15. (D) Surface configuration with incorporated KCl clusters on SBA-15. (E) Structural model of the pore geometry of KCl/SBA-15. Note that the bond distance corresponds the distance between alkali-metal atom and oxygen atom of carbon dioxide (Inset in Figure 1A) and the O-C-O angles of the absorbed CO2 on LiCl/SBA-15, NaCl/SBA-15, KCl/SBA-15 are 176.2°, 160.7°, and 153.1°, respectively.

From calculating the Madelung potential energies ordered as LiCl (8.57 eV) > NaCl (7.96 eV) > KCl (6.63 eV), we found that the relatively high affinity of CO2 molecules to KCl/SBA-15 may result from this relatively low Madelung potential energy caused by the relatively long ionic distance between the K cation and Cl anion (Li+-Cl-: 2.98Å, Na+-Cl-: 2.99Å, K+-Cl-: 3.01Å).

ACS Paragon Plus Environment

13

ACS Applied Materials & Interfaces

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 14 of 29

We think such low Madelung potential energy of KCl leads to the relatively strong basicity of the K+ cation facilitating active charge transfer from the adsorbent surface to the acidic CO2.38 Therefore, KCl/SBA-15 was selected for further examination. All of the A/SBA-15s in this study are composed of crystalline alkali-chloride clusters over SBA-15 which is partially covered by hydroxyl groups.35 Through physico-chemical analyses (Figure S2 to S6, and tables S1 and S2), it was observed that KCl/SBA-15 has a well-defined two-dimensional p6mm hexagonal mesoporous structure with an average pore size of 10 nm (Figure 1, B and C). Through the DFT geometry optimization of CO2 on the surface of the KCl/SBA-15 model with the DFT-D3 correction (Figure 1, D and E, and Figure S9 and S10, and tables S3 and S4),39 it was found that the top and edge sites of the KCl clusters are the strongest adsorption sites for the CO2 molecules (Figure S11 and S12), which is in a good agreement with those of a previous report for pristine KCl (100).40 3.2. GCMC simulations using newly developed force field. Next, we developed a new force field (improved DREIDING-type force field, IDFF, listed in tables S5 to S7) that can reproduce the DFT results for the CO2 adsorption in the bare SBA-15 and KCl/SBA-15 systems (Figure S13). To validate the newly developed force field (FF) parameters, we compared the binding energy curve obtained from newly developed FF with those obtained from DFT with DFT-D3 correction for various gas molecules (CO2, CH4, N2, and H2), as shown in Figure 2, confirming that the binding energies27 calculated using newly developed FF are in good agreement with the DFT binding energies. We also extended the development of the force field parameters to other potential gas components including O2, H2S, NO, and NO2, as described in the Supplementary Information (Figure S14 and S15).

ACS Paragon Plus Environment

14

Page 15 of 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

ACS Applied Materials & Interfaces

Figure 2. Binding energy profiles of CO2, CH4, N2, and H2 gas molecules. Profiles (A) on the KCl/SBA-15 surface and (B) on the SBA-15 (SiO2) surface. The x-axis represents interaction coordinates (Å), and the y-axis represents the binding energy (kcal/mol). The binding energies were predicted by the dispersion-corrected DFT-D3 method.

ACS Paragon Plus Environment

15

ACS Applied Materials & Interfaces

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 29

Figure 3. Prediction of gas adsorption properties on SBA-15. (A) Structure model of SBA-15. (B) Fitting of experimental isotherm and GCMC-simulated isotherms using IDFF and UFF of CO2 adsorption on SBA-15 at 273 K. CO2 adsorption capacity (molCO2/gads.) measured from experiment was converted to same unit from the simulations (molecules/nm2) using Avogadro constant (1 mol = 6.02 × 1023 molecules) and external surface area (m2/gads.) of the adsorbent. (C) Prediction of the adsorption of various gases on SBA-15 using GCMC simulation at 273 K. (D) Predicted adsorption profiles of 40% CO2 balanced with balanced gases (CH4, N2, and H2) at various operating temperatures using GCMC simulation.

Grand Canonical Monte Carlo (GCMC) simulations of pure and mixed gas systems were performed using IDFF to predict the CO2 adsorption at a range from zero to 1.0 bar of CO2 partial pressure. We found that the simulated CO2 adsorption isotherms for bare SBA-15 using

ACS Paragon Plus Environment

16

Page 17 of 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

ACS Applied Materials & Interfaces

IDFF are well agreed with the experimental isotherms (the heat of adsorption was about 23 kJ/mol; Figure S16A), in comparison to those resulted from the simulations using UFF (discrepancy range between

experiment

and simulation: