Molecular simulation and experimental validation of ...

5 downloads 0 Views 2MB Size Report
Apr 27, 2018 - Corresponding author at: Department of Civil Engineering, University of Louisiana at Lafayette, P. O. Box 43598, Lafayette, LA, 70504, USA.

Journal of Hazardous Materials 354 (2018) 258–265

Contents lists available at ScienceDirect

Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat

Molecular simulation and experimental validation of resorcinol adsorption on Ordered Mesoporous Carbon (OMC)

T

Zaki Uddin Ahmada,c, Bing Chaoa, Mas Iwan Konggidinatab, Qiyu Liana,c, Mark E. Zappib,c, ⁎ Daniel Dianchen Ganga,c, a

Department of Civil Engineering, University of Louisiana at Lafayette, P. O. Box 43598, Lafayette, LA, 70504, USA Department of Chemical Engineering, University of Louisiana at Lafayette, P. O. Box 43675, Lafayette, LA, 70504, USA c Center for Environmental Technology, The Energy Institute of Louisiana, P. O. Box 43597, Lafayette, LA, 70504, USA b

G R A P H I C A L A B S T R A C T

A R T I C LE I N FO

A B S T R A C T

Keywords: Ordered mesoporous carbon Rhombic OMC model Molecular simulation Adsorption Validation

Numerous research works have been devoted in the adsorption area using experimental approaches. All these approaches are based on trial and error process and extremely time consuming. Molecular simulation technique is a new tool that can be used to design and predict the performance of an adsorbent. This research proposed a simulation technique that can greatly reduce the time in designing the adsorbent. In this study, a new Rhombic ordered mesoporous carbon (OMC) model is proposed and constructed with various pore sizes and oxygen contents using Materials Visualizer Module to optimize the structure of OMC for resorcinol adsorption. The specific surface area, pore volume, small angle X-ray diffraction pattern, and resorcinol adsorption capacity were calculated by Forcite and Sorption module in Materials Studio Package. The simulation results were validated experimentally through synthesizing OMC with different pore sizes and oxygen contents prepared via hard template method employing SBA-15 silica scaffold. Boric acid was used as the pore expanding reagent to synthesize OMC with different pore sizes (from 4.6 to 11.3 nm) and varying oxygen contents (from 11.9% to 17.8%).



Corresponding author at: Department of Civil Engineering, University of Louisiana at Lafayette, P. O. Box 43598, Lafayette, LA, 70504, USA. E-mail address: [email protected] (D.D. Gang).

https://doi.org/10.1016/j.jhazmat.2018.04.072 Received 2 February 2018; Received in revised form 25 April 2018; Accepted 26 April 2018 Available online 27 April 2018 0304-3894/ © 2018 Elsevier B.V. All rights reserved.

Journal of Hazardous Materials 354 (2018) 258–265

Z.U. Ahmad et al.

Based on the simulation and experimental validation, the optimal pore size was found to be 6 nm for maximum adsorption of resorcinol.

model of MCM-41, and studied the adsorption of CO2, N2 and flue gas on MCM-41 by the Grand Canonical Monte Carlo (GCMC) simulation. Chen et al. [17] constructed ZSM-5-MCM-41 model using Materials Visualizer module to investigate the structural information and toluene adsorption. The simulated adsorption isotherm followed Langmuir isotherm and was consistent with the experimental results. Cao et al. [18] carried out the GCMC simulation to investigate the adsorption of nitrogen onto MCM-41 by employing Tjatjopoulos–Feke–Mann (TFM) potential to understand the interaction between fluid molecule and MCM-41 pore wall. In addition, Jentys et al. [19] studied the type and concentration of hydroxyl groups present on MCM-41 as a function of Si/Al ratio and the pore size using infrared spectroscopy. The internal surface of MCM-41 was found to be partially covered with hydroxyl groups based on the results from spectroscopy and simulation. Yuan et al. [20] studied the adsorption equilibrium and kinetics of CO2, CH4 and N2 onto OMC for the separation of CO2 and CH4 from air for the purpose of controlling greenhouse gas emission. Higher adsorption capacity was shown by OMC for the separation of binary mixtures and separation selectivities were estimated based on Ideal Adsorbed Solution Theory (IAST) model. Based on literature survey, limited number of research effort has been dedicated to understand the adsorption of organic compounds onto OMC utilizing the simulation technique and to explore the relationship between adsorption capacity and physiochemical properties of OMC. In this paper, a study to explore the relationship between the structural information of OMC and adsorption capacities is presented in order to find the optimal pore size and oxygen content for resorcinol adsorption by simulation and the obtained results were validated experimentally. In order to investigate the structural information and to find the optimal structural conditions for adsorption, the Materials Studio software was used in this study. Materials Studio is an effective tool for studying the structural details and properties of porous materials on atomistic level. OMC models with different pore sizes and oxygen contents were constructed and simulated using ZSM-5 as the repetitive unit cell. ZSM-5 is the most studied and used zeolite due to its distinct features such as, unique framework topology, higher hydrothermal stability, and easy modification of physicochemical properties such as acidity and texture [21]. ZSM-5 framework contains two intersecting channel systems, with one straight channel parallel to [0 1 0] and the other sinusoidal running parallel to [0 0 1] defined by 10membered ring openings of 0.53 nm × 0.56 nm and 0.51 nm × 0.55 nm, respectively [22,23]. The typical organic contaminant, resorcinol, was selected to evaluate the adsorption capacity.

1. Introduction The conventional porous carbon materials used as adsorbents, for example activated carbon, usually possess relatively wide pore size distributions in micropore range (pore size less than 2 nm) [1]. Limitations such as slow mass transfer and low specific surface area that lead to lower adsorption capacity have been reported for these adsorbents [2]. To overcome such limitations, synthetically produced adsorbent such as Ordered Mesoporous Carbon (OMC) with tunable pore size (3–10 nm), high specific surface area (1300–2000 m2 g−1) and large pore volume (1–2 cm3 g−1) was studied [3]. Previous studies have demonstrated that OMCs with well-controlled pore structure and narrow pore size distribution have shown good potential for the adsorption of organic and inorganic pollutants. Wang et al. [4] suggested that OMC is a potential adsorbent for the removal of volatile organic compounds (VOCs). Good adsorption capacities of OMC were reported for the removal of resorcinol, BTEX and synthetic azo dyes from aqueous solutions [5,6,7,8,9]. Chen et al. [10] reported the synthesis of OMC with large accessible pores (22 nm). The obtained OMC was found to be an effective adsorbent for Cr(VI) removal. In addition, OMC functionalized with 4-acetophenone oxime reported by Tian et al. [11] demonstrated promising adsorption capacity for U(VI). Several different approaches have been proposed and experimentally studied to synthesize OMC and develop the relation between structural characteristics of OMC and its adsorption capacity. Yuan et al. [12] investigated the effect of pore volume on the adsorption capacity of Methylene Blue onto OMC. It was concluded that the volume of mesopores with the pore size larger than 3.5 nm have large impact and was found to be important in the effective adsorption of Methylene blue onto OMC. Kim et al. [13] synthesized OMC at three different pyrolysis temperatures to understand the effects of acidic function groups and porosity on adsorption capacity. It was found that the adsorbed amount of methyl mercaptan onto OMC was strongly influenced by the acidic functional group. Hence, the pore size and surface chemistry of OMC are two important factors that can affect the adsorption capacity [14,15]. Although widespread experimental research works have been undertaken to synthesize OMC with different pore sizes and different function groups to enhance the adsorption capacity and to investigate the adsorption mechanism, the interactions between OMC and adsorbates are still not completely understood. An alternative technique to study the adsorption behavior of mesoporous materials is molecular simulation on atomistic level. Zhou et al. [16] constructed a hexagonal

Fig. 1. (a) The model of amorphous carbon unit cell (b) Repetitive rhombic unit for OMC modeling. 259

Journal of Hazardous Materials 354 (2018) 258–265

Z.U. Ahmad et al.

geometry of the structure until the relative distances between a group of atoms are fixed. Smart algorithm was chosen to run the iteration process to adjust both the cell parameters and atomic coordinates. Smart algorithm is a cascade of methods using successively Steepest Descent and Conjugate Gradient Algorithms. The convergence quality was set as Medium, in which energy was 0.001 kcal mol−1, maximum force was 0.5 kcal mol−1 per Å, and maximum displacement was 0.015 Å. The cell parameters and atomic coordinates were adjusted until the total energy of the structure was minimized [27]. The small-angle X-ray diffraction patterns were also acquired using Forcite module. The physical properties of the OMC models could be obtained by embedding the models into the Materials Studio Quantitative Structure-Activity Relationships (QSAR) module, in which, the properties (molecular mass, volume, surface area and crystal dimensions) of the model are united and listed. The specific surface area (SSA) was calculated by the following equation:

The molecular size of resorcinol is found to be 0.56 nm × 0.47 nm [24]. To validate the simulation results, OMCs were synthesized experimentally via hard template method with sucrose as the carbon precursor and SBA-15 as the silica scaffold. OMCs with tunable pore sizes ranging from 4.6 to 11.3 nm were obtained experimentally by utilizing boric acid as the pore expanding reagent [25]. In addition, a series of characterization techniques, such as nitrogen adsorption-desorption isotherm, pore size distribution (PSD), specific surface area (SSA), smallangle XRD, transmission electron micrograph (TEM) and resorcinol adsorption experiments, were undertaken in the experimental phase to validate the simulation results. 2. Details of simulation 2.1. OMC model construction Zhao [26] reported a rectangular shaped model using amorphous carbon unit cell (Fig. 1(a)) as the repetitive unit. The rectangular model had several limitations in the pore size control and formation of the hexagonal structure if the ratio of the side length and diagonal is only 2. Additionally, the density of the proposed model was 3.5 g cm−3 which was too dense and there was no channel connectivity. For this reason, the specific surface area of this rectangular shaped model could only reach approximately 100 to 200 m2 g−1, which was way off from that of the OMC prepared in this study. In addition, OMC prepared in this study possessed two dimensional hexagonal array (p6mm) pore structure as shown in Fig. 1(b). Therefore, the rectangular shaped model could not be directly used in this study. In this study, a new OMC model was proposed using rhombus shaped repetitive cell and ZSM-5 as the unit cell as shown in Fig. 2(a). ZSM-5, imported from the database of Materials Studio, is an aluminosilicate zeolite. It has a density of approximately 1.78 g cm−3 and has abundant and uniform micropore structure with channel connectivity as shown in Fig. 2(b). The structure unit of ZSM-5 model was 2.007 × 1.992 × 1.342 nm3 in size. The unit cell was modified by changing all the other atoms to carbon atoms. The obtained structures were optimized by Forcite module. The lattice parameter of angle was set to α = 90°, β = 90° and γ = 120° in order to build a hexagonal array structure. The OMC model as depicted in Fig. 2(c) was constructed by carving out straight cylindrical pores in a ZSM-5 silica super cell which was imported from the database of Materials Studio. The dimension of the supper cell was determined by the pore size of the OMC model. Next, two and a half cylindrical pores were created by removing the atoms inside the pore of the diameter selected. The pores and internal surface were saturated with hydrogen by Hydrogen Adjust Module. Variations of oxygen contents were achieved by adding oxygen atoms to the unsaturated carbon atom sites on the pore surface. All the oxygen atoms existed in the form of hydroxyl and carboxyl groups. OMC models with different oxygen contents were generated through the following procedure: (1) the carbon atoms located inside the pore of the selected diameter were removed; (2) the vacant bonds on carbon atoms on the surface of the pore were saturated by oxygen atoms; (3) saturation of all the carbon and oxygen atoms with hydrogen atoms. Different pore sizes were obtained by changing the radius of the carving cylindrical channel.

SSA =

A ρ×V

(1)

Where A is the BET surface area obtained from surface area analysis module; ρ is the density of the OMC model which calculated by the Forcite Analysis module; V is the bulk volume of the OMC model calculated by the lattice length. 2.3. Adsorption simulation The resorcinol molecular model built in material studio is depicted in Fig. 3. A six-membered ring was drawn by Sketch Ring tool. Then two oxygen atoms were added in the symmetrical position in the phenyl ring. The hydrogen atoms were filled with the empty valences by Adjust Hydrogen tool. The small-angle X-ray diffraction pattern and

2.2. Structure optimization The structures of the final models generated were geometrically optimized by Forcite Module in Materials Studio packages using Condensed-Phase Optimized Molecular Potentials for Atomistic Simulation Studies (COMPASS) force field. Therein, Forcite Module is a tool to perform a wide range of molecular mechanics calculations (such as single-point energy calculation, geometry optimization and molecular dynamics, etc.) using classical force field based on simulation techniques. Geometry Optimization task was used to refine the

Fig. 2. (a) The model of ZSM-5 unit cell (b) Channel connectivity of ZSM-5 (c) Line model of OMC. 260

Journal of Hazardous Materials 354 (2018) 258–265

Z.U. Ahmad et al.

Fig. 4. Comparison of the XRD patterns between experimental and simulation (based on Grand Canonical Monte Carlo theory) results.

Fig. 3. Resorcinol molecular model built by Materials Studio.

adsorption of resorcinol onto OMC was based on the Grand Canonical Monte Carlo (GCMC) theory and was simulated using Scattering tool in Forcite Analysis Module and Sorption Module of Material Studio packages. The adsorption capacity of resorcinol onto OMC was calculated by the following equation:

Adsorption Capacity =

n × Mresorcinol NA × V × ρ

(2)

Where n is the number of resorcinol molecules adsorbed per cell; Mresorcinol is the modular weight of resorcinol; NA is the Avogadro’s constant, which is 6.02 × 1023; V is the volume of the OMC cell; ρ is the density of OMC.

Fig. 5. Loading curve of resorcinol onto OMC model by sorption module.

3.2. Fabrication of OMC Different amounts of boric acid were added as the pore expanding reagent and sucrose was used as the carbon precursor. The synthesis procedure followed the methods reported by Lee et al. [1] and Wang et al. [29]. The obtained OMC samples were denoted as OMC- x . The x value indicates the molar ratio of boric acid to carbon precursor, which is varied from 0 to 12. In a typical synthesis of OMC-1, 0.27 g of boric acid (99.8 wt%), 1.5 g of sucrose and 15 drops of concentrated sulfuric acid were dissolved in 15.0 ml distilled water. After adding 2.0 g of SBA-15, the mixture was heated at 100 °C for 6 h and subsequently at 160 °C for another 6 h. The resulting composite was impregnated again with an aqueous solution of 0.09 g of boric acid, 0.51 g of sucrose and 4 drops of concentrated sulfuric acid. Additional distilled water of 10.0 mL was added in this stage. Following the same heat treatment method as the previous step, the composite was carbonized at 700 °C for 8 h under the nitrogen atmosphere. Finally the OMC-1 was obtained by the removal of silica template using 150 mL of 48% HF solution at room temperature.

3. Details of experiment 3.1. Synthesis of SBA-15 SBA-15 silica template was synthesized under acidic condition using Pluronic P123 (EO20PO70EO20, BASF) as the surfactant and tetraethyl orthosilicate (TEOS, 98%) as the silica source. The synthesis procedure followed the method reported by Zhao et al. [28] with modifications on the reaction temperature and reaction time. Typically, 100 mL of concentrated hydrochloride acid (HCl, 37%) was added into the 525 mL of distilled water. 20 g of triblock copolymer Pluronic P123 was added into the aqueous solution. The solution was stirred for approximately 2 h until all P123 was completely dissolved. Next, 46.5 mL of tetraethyl orthosilicate (TEOS, 98%, Aldrich) was added to the homogenous solution with vigorous stirring for 10 min. The resulting mixture was placed in a constant temperature water bath (Premiere Thermostatic water bath HH-4) for 4 h at 40 °C, followed by aging for 24 h at 90 °C. After the aging was done, the solid product was washed with 80–90 °C hot distilled water, and dried in an oven at 105 °C overnight. After drying, the product was calcined in a muffle furnace at 550 °C for 8 h. The white silica template SBA-15 was obtained and stored in the desiccator for the preparation of OMC.

3.3. Structural characterization Small angle X-ray diffraction patterns were obtained using a PANalytical X-ray diffractometer model Empyream operated at 45 kV

Table 1 Lattice parameters and elemental composition of OMC model and OMC-0 sample. Sample

Pore size (nm)

Wall thickness (nm)

Oxygen content (%)

Carbon content (%)

Density (g cm−3)

Pore volume (cm3 g−1)

SSA (m2 g−1)

OMC model OMC-0 sample

4.6 4.6

7.1 7.1

11.2 13.3

85.3 83.7

1.13 1.12

1.36 1.58

969.30 1404.2

261

Journal of Hazardous Materials 354 (2018) 258–265

Z.U. Ahmad et al.

Fig. 6. Energy variation in the resorcinol adsorption process.

Fig. 8. Comparison of the simulation results and experimental results: Specific Surface Area vs. Pore Size. Fig. 7. Density distribution of resorcinol in the OMC structure (red points denote the resorcinol molecules). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article). Table 2 Structural parameter and adsorption capacity of the OMC models. Pore size (nm)

Density (g cm−3)

BET Surface Area (m2 g−1)

Wall Thickness (nm)

Adsorption Capacity (mg g−1)

3 4 5 6 7 8 9 10 11

1.060 0.955 0.857 0.770 0.730 0.666 0.613 0.579 0.543

808.53 973.20 1132.20 1347.60 1148.50 954.30 963.20 986.70 990.30

8.7 7.7 6.7 5.7 4.7 3.7 2.7 1.7 0.7

31.5 45.6 56.1 68.3 53.3 42.8 51.2 57.3 52.2

Fig. 9. Comparison of the simulation results and experimental results: Adsorption Capacity vs. Specific Surface Area.

and 40 mA and using Cu Kα1 radiation (k = 1.5406 Å). Measurements were carried out to obtain resolved XRD patterns at 2θ angles from 0.5° to 5°. The step size was 0.0066° and the soller slit was 0.04 rad. The specific surface area (SSA), pore volume and pore size distribution were determined using the Micromeritics ASAP 2020 surface area and porosimetry physisorption system. The nitrogen adsorption

and desorption isotherm was measured in the relative pressure range of 0 and 0.99. TEM images were obtained using a Hitachi 7600 Transmission 262

Journal of Hazardous Materials 354 (2018) 258–265

Z.U. Ahmad et al.

Table 3 Results of adsorption simulation for OMC model with different oxygen contents. Pore Size (nm)

Oxygen Content (%)

Density (g cm−3)

SSA (m2 g−1)

Adsorption Capacity (mg g−1)

4.6 4.6 4.6 4.6 4.6 4.6

0 2.9 4.4 6.2 9.6 12.3

0.87 0.92 1.03 1.28 1.49 1.69

1028.3 1030.5 1031.4 1033.5 1035.7 1036.8

41.3 43.6 45.1 45.8 46.7 47.2

Fig. 10. Pore size distribution curves of synthesized OMC.

Electron Microscopy (Hitachi America Ltd., Tarrytown, New York). The instrument was operated at the acceleration voltage of 100 kV. The samples were prepared by dispersing a large number of particles in ethanol in an ultrasonic water bath for 45 min and a drop of the resulting suspension was placed on a 400 mesh Cu grid. 3.4. Batch adsorption study Resorcinol, a typical TOC model compound, was selected to evaluate the adsorption capacity of OMC. The adsorption experiments were carried out by analyzing 100 mL of 7.5 mg L−1 resorcinol solution using 10 mg of OMC at the pH of 6.1. One sample of the same concentration solution without the adsorbent (blank) was prepared and treated under the same conditions as the solutions containing adsorbent. This blank was used as a reference to establish the initial concentration of the solutions containing the adsorbent. After the addition of adsorbent, these solutions were placed in a mechanical shaker at 250 rpm for 24 h at 25 °C. The conical flasks were then removed and solutions were filtered using a 0.45 μm filter paper. The equilibrium concentrations of each solution were measured by Cary 50 UV–vis spectrophotometer (Varian) at wavelength of 500 nm in room temperature. The amount of adsorbed resorcinol at equilibrium, q (mg L−1) was calculated by using the following equation:

q=

(C i−C f ) × V M

Where q is the adsorption capacity (mg g

Fig. 12. Comparison of simulation results and experiment results: Adsorption capacity vs. oxygen content.

phase initial and final concentration of resorcinol (mg L−1), respectively; V is the volume of the solution (L) and M is the mass of adsorbent used (g). 4. Results and discussion 4.1. Verification of OMC model In order to investigate the effects of pore size and surface chemistry, the ideal models of OMC with different properties were constructed and their adsorption capacities were simulated. The OMC model was optimized with the same unit cell parameter of OMC-0 sample. The pores were arranged hexagonally with pore radius of 4.6 nm and wall thickness of 7.1 nm.

(3) −1

); Ci and Cf are the liquid-

Fig. 11. TEM images of (a) OMC-0 (b) OMC-12. 263

Journal of Hazardous Materials 354 (2018) 258–265

Z.U. Ahmad et al.

which was the same as the experimental OMC. The simulation results are shown in Table 2. The relationships between the specific surface area and pore size are plotted and compared with the experimental data as shown in Fig. 8. The simulation results and experimental data show a similar trend with an average relative error of 0.1938. This indicates that the simulation results are well fitted with the experimental results. The highest adsorption capacity could be obtained at the diameter of 6 nm. At 6 nm, largest specific surface area was also obtained. The larger pore size did not indicate a better adsorption capacity. From the simulation results, it was observed that the specific surface area was the main factor affecting the adsorption capacity. The result was in line with the conclusion drawn by You et al. [30], where the specific surface areas were more closely related to the immobilization capacity of porous materials compared to the pore sizes. In order to explore the effect of the specific surface area, the relationship between adsorption capacity and specific surface area is plotted in Fig. 9. The adsorption capacity obtained from simulation results increase linearly with the increase in the specific surface area. The larger surface area provides more interaction sites for the adsorbates. Based on the simulation results, the specific surface area varied with the pore diameter, OMC with pore diameter of 6 nm yielded larger surface area. However, when the pore size was increased to 11 nm, the wall thickness was too thin to retain the ordered structure. Therefore, the optimal pore size of OMC made from SBA-15 is 6 nm for resorcinol removal. Boric acid was used as an effective pore expanding reagent to synthesize OMCs with tailored pore size. With the increment of pore size, the specific surface area increases upto a certain point and then decreases (Fig. 8). The adsorption capacity increases with the increment of specific surface area. So, the increment of pore size cannot be translated into better adsorption capacity. Based on the simulation results, the optimal pore sizes are found to be 6 nm and 10 nm with the corresponding adsorption capacity of 68.3 and 57.3 mg g−1, respectively. However, OMCs modified with boric acid have greater tendency to structural collapse compared to OMC without boric acid. Considering the stability of the OMC framework, the pore size of 10 nm was excluded because of the susceptible structural collapse. The differences between the simulated and experimental data were observed. This divergence is mainly caused by the assumption that the structure in the model is ideally ordered. Additionally, the extent of crystallization, the coarseness of the surface, and the absence of ideal pore system in the actual OMC can be the reasons for the reduction of resorcinol adsorption capacity under experimental conditions. The increase in the pore size with the increase in boron content was confirmed by pore size distributions as depicted in Fig. 10. The OMC-4, OMC-8 and OMC-12 had flatter peaks indicating wider ranges of pore size and less ordered structure. In addition, the largest pore size of 11.3 nm was obtained by OMC-12. The pore size was close to the unit cell length. For these three samples, the wall thickness was too thin to support the rigid carbon structure and therefore, less ordered carbon structure was obtained as shown in Fig. 11. The same observation was reported by Ryoo et al. [31]. In general, the OMC made with different ratio of boric acid to sucrose have tailored pore sizes ranging from 4.6 nm to 11.3 nm.

The optimized structural parameter and elemental composition of the OMC model and OMC-0 sample are listed in Table 1. The specific surface area of the OMC model was estimated using the atom volumes and surface tools in Materials Studio package. The specific surface area of the OMC model was found to be 969.3 m2 g−1, which was less than the BET surface area of OMC-0. The decrease in surface area compared to BET surface area of OMC-0 may be caused by the ideally ordered structure assumed in the OMC model. In addition, some complications such as the coarseness of the surface and the extent of crystallization were ignored in order to simplify the model. In a study carried out by Chen et al. [17], lower surface area and lower total pore volume were also observed for the MCM-41 model based on ZSM-5 zeolite structure. However, there was no significant effect on the description of the pore structure of the samples, with all the other parameters similar to the parameters observed for the OMC produced experimentally. Based on the experimental XRD analysis, the same analysis conditions such as the wavelength (k = 1.5406 Å) and scan angle range (from 0.5° to 5°) were selected. The simulated XRD pattern and experimentally obtained XRD pattern are compared in Fig. 4. The small-angle XRD estimated by simulation shows three peaks at (100), (110) and (200) reflection which is a typical pattern for hexagonally ordered array structure. The (100) diffraction peak of the experimental pattern bifurcates to two small peaks, while the (110) and (200) diffraction peaks are indiscernible. The results indicate that the OMC obtained experimentally has less ordered hexagonal mesopore structures. The model in the simulation represented an ideal material that was constructed with an ordered configuration and uniform orientation. Therefore, the model could only describe the pore structure and crystal reflection up to a certain extent. The adsorption process was simulated by Sorption Module which can be used to predict fundamental properties, such as sorption isotherms and Henry’s constants. During the course of simulation, sorbates are randomly created, deleted, translated and rotated until a stage is accepted or rejected according to the selection rules of the Monte Carlo method. The adsorption isotherm was carried out in a pressure range from 20 kPa to 101 kPa and temperature at 298 K. The COMPASS force field was used and the system was equilibrated at 298 K to obtain minimum energy configuration. The isotherm is shown in Fig. 5. The initial configuration was generated by loading fifty resorcinol molecules randomly in the pore structure. The OMC model rapidly took up resorcinol molecules in the first stage, followed by a slow continuous uptake for adsorption. The system was then equilibrated for 60,000 GCMC steps and the average amount of adsorption was reported. The energy variation during the adsorption process is shown in Fig. 6. The energy variation results demonstrate that the system is equilibrated and in a stable condition for each step. The slight increase of the intermolecular energy and decrease of the Van der Waals energy and electrostatic energy indicate that the interaction of the adsorbate and OMC model is not strong. So, the adsorption of resorcinol on OMC model was more likely a physical adsorption process. The density distribution of the adsorbate is depicted in Fig. 7. The channels were found to be fully saturated with the adsorbate molecules. The loading of resorcinol at equilibrium was found to be 157 adsorbed resorcinol per cell. The simulated adsorption capacity of OMC model was found to be 41 mg g−1. The simulated adsorption capacity was close to the experimentally obtained adsorption capacity of 36.5 mg g−1. Therefore, the OMC model developed using ZSM-5 as the basic template was a reliable model for adsorption simulation.

4.2.2. Effect of oxygen content The effects of oxygen content on the surface of the OMC model were evaluated by constructing the OMC model doped with different amount of oxygen atoms. OMC model with pore size of 4.6 nm was selected as the basic structure. Oxygen content was the only variable in this iteration of simulation. The oxygen atoms existed in the form of hydroxyl and carboxyl groups and were distributed on the surface of the pore structure. The simulation results are shown in Table 3. The plot of oxygen content as a function of adsorption capacity is presented together with the experimental result as depicted in the

4.2. Simulation of resorcinol adsorption 4.2.1. Effect of pore size In order to investigate the effect of pore size on the adsorption performance, a series of OMC models with different pore sizes (from 3 nm to 11 nm) were built using Materials Studio. The unit cell parameter was maintained at 11.7 nm (calculated from the XRD pattern), 264

Journal of Hazardous Materials 354 (2018) 258–265

Z.U. Ahmad et al.

Fig. 12. It was observed that the adsorption capacity of resorcinol increased with the increase in oxygen atoms present on the pore surface. With the increase in oxygen content from 0% to 12%, the adsorption capacities increased from 41.3 mg g−1 to 47.2 mg g−1. The difference in the adsorption capacity of OMC model and OMC prepared experimentally may be attributed to the ideal ordered structure assumed in the OMC model. Comparing the influence of specific surface area with the influence of oxygen-containing functional groups, the specific surface area increased the adsorption capacity up to approximate 116% while the influence of oxygen-containing functional group for the adsorption of resorcinol was nearly negligible. Many researchers have proved that the hydroxyl functional group has the ability to enhance the affinity with the acid adsorbates while the carboxylic functional groups are more beneficial for the basic adsorbates [32]. However, based on the simulation results obtained, the oxygen-containing functional groups show no significant improvement on the adsorption capacity on resorcinol removal. Therefore, the adsorption process of resorcinol is more of a physical process rather than a chemical interaction. It can also be confirmed by the energy variation diagram as shown in Fig. 6. The main attractive force between the adsorbates and carbon is Van der Waals force.

[5] W. Shou, B. Chao, Z.U. Ahmad, D.D. Gang, Ordered mesoporous carbon preparation by the in situ radical polymerization of acrylamide and its application for resorcinol removal, J. Appl. Polym. Sci. 133 (2016) 43426, http://dx.doi.org/10.1002/app.43426. [6] M.I. Konggidinata, B. Chao, Q. Lian, R. Subramaniam, M. Zappi, D.D. Gang, Equilibrium, kinetic and thermodynamic studies for adsorption of BTEX onto Ordered Mesoporous Carbon (OMC), J. Hazard. Mater. 336 (2017) 249–259, http://dx.doi.org/10.1016/j. jhazmat.2017.04.073. [7] J. Goscianska, M. Marciniak, R. Pietrzak, Mesoporous carbons modified with lanthanum (III) chloride for methyl orange adsorption, Chem. Eng. J. 247 (2014) 258–264, http:// dx.doi.org/10.1016/j.cej.2014.03.012. [8] J. Goscianska, M. Marciniak, R. Pietrzak, Ordered mesoporous carbons modified with cerium as effective adsorbents for azo dyes removal, Sep. Purif. Technol. 154 (2015) 236–245, http://dx.doi.org/10.1016/j.seppur.2015.09.042. [9] J. Goscianska, N.A. Fathy, R.M. Aboelenin, Adsorption of solophenyl red 3BL polyazo dye onto amine-functionalized mesoporous carbons, J. Colloid Interface Sci. 505 (2017) 593–604, http://dx.doi.org/10.1016/j.jcis.2017.06.052. [10] T. Chen, T. Wang, D.J. Wang, H.R. Xue, J.Q. Zhao, X.C. Ding, S.C. Wu, J.P. He, Synthesis of ordered large-pore mesoporous carbon for Cr(VI) adsorption, Mater. Res. Bull. 46 (2011) 1424–1430, http://dx.doi.org/10.1016/j.materresbull.2011.05.009. [11] G. Tian, J. Geng, Y. Jin, C. Wang, S. Li, Z. Chen, H. Wang, Y. Zhao, S. Li, Sorption of uranium (VI) using oxime-grafted ordered mesoporous carbon CMK-5, J. Hazard. Mater. 190 (2011) 442–450, http://dx.doi.org/10.1016/j.jhazmat.2011.03.066. [12] X. Yuan, S.P. Zhuo, W. Xing, H.Y. Cui, X.D. Dai, X.M. Liu, Z.F. Yan, Aqueous dye adsorption on ordered mesoporous carbons, J. Colloid Interface Sci. 310 (2007) 83–89, http://dx.doi.org/10.1016/j.jcis.2007.01.069. [13] D.J. Kim, H.I. Lee, J.E. Yie, S.J. Kim, J.M. Kim, Ordered mesoporous carbons: implication of surface chemistry, pore structure and adsorption of methyl mercaptan, Carbon 43 (2005) 1868–1873, http://dx.doi.org/10.1016/j.carbon.2005.02.035. [14] H. Ren, W. Shou, C. Ren, D.D. Gang, Preparation and post-treatments of ordered mesoporous carbons (OMC) for resorcinol removal, Int. J. Environ. Sci. Technol. 13 (2016) 1505–1514, http://dx.doi.org/10.1007/s13762-016-0990-7. [15] Z.U. Ahmad, Q. Lian, M.E. Zappi, P.R. Buchireddy, D.D. Gang, Adsorptive removal of resorcinol on a novel ordered mesoporous carbon (OMC) employing COK–19 silica scaffold: kinetics and equilibrium study, J. Environ. Sci. (2018), http://dx.doi.org/10. 1016/j.jes.2018.04.014. [16] S. Zhuo, Y. Huang, J. Hu, H. Liu, Y. Hu, J. Jiang, Computer simulation for adsorption of CO2, N2 and flue gas in a mimetic MCM-41, J. Phys. Chem. C 112 (2008) 11295–11300, http://dx.doi.org/10.1021/jp803428n. [17] C. Huiyong, X. Hongxia, C. Xianying, Q. Yu, Experimental and molecular simulation studies of a ZSM-5-MCM-41 micro-mesoporous molecular sieve, Microporous Mesoporous Mater. 118 (2009) 396–402, http://dx.doi.org/10.1016/j.micromeso.2008.09.020. [18] D. Cao, Z. Shen, J. Chen, X. Zhang, Experiment, molecular simulation and density functional theory for investigation of fluid confined in MCM-41, Microporous Mesoporous Mater. 67 (2004) 159–166, http://dx.doi.org/10.1016/j.micromeso.2003.11.001. [19] A. Jentys, K. Kleestorfer, H. Vinek, Concentration of surface hydroxyl groups on MCM-41, Microporous Mesoporous Mater. 27 (1999) 321–328, http://dx.doi.org/10.1016/S13871811(98)00265-0. [20] B. Yuan, X. Wu, Y. Chen, J. Huang, H. Luo, S. Deng, Adsorption of CO2, CH4, and N2 on ordered mesoporous carbon: approach for greenhouse gases capture and biogas upgrading, Environ. Sci. Technol. 47 (2013) 5474–5480, http://dx.doi.org/10.1021/ es4000643. [21] B. Yilmaz, U. Müller, Catalytic applications of zeolites in chemical industry, Top. Catal. 52 (2009) 888–895, http://dx.doi.org/10.1007/s11244-009-9226-0. [22] Y. Wei, P.E. de Jongh, M.L.M. Bonati, D.J. Law, G.J. Sunley, K.P. de Jong, Enhanced catalytic performance of zeolite ZSM-5 for conversion of methanol to dimethyl ether by combining alkaline treatment and partial activation, Appl. Catal. A-Gen. 504 (2015) 211–219, http://dx.doi.org/10.1016/j.apcata.2014.12.027. [23] J. Li, S. Liu, H. Zhang, E. Lü, P. Ren, J. Ren, Synthesis and characterization of an unusual snowflake-shaped ZSM-5 zeolite with high catalytic performance in the methanol to olefin reaction, Chin. J. Catal. 37 (2016) 308–315, http://dx.doi.org/10.1016/S18722067(15)60979-2. [24] J. Huang, K. Huang, C. Yan, Application of an easily water-compatible hypercrosslinked polymeric adsorbent for efficient removal of catechol and resorcinol in aqueous solution, J. Hazard. Mater. 167 (2009) 69–74, http://dx.doi.org/10.1016/j.jhazmat.2008.12.120. [25] B. Chao, M.I. Konggidinata, L. Lin, M. Zappi, D.D. Gang, Effect of carbon precursors and pore expanding reagent on ordered mesoporous carbon for resorcinol removal, J. Water Process. Eng. 17 (2017) 256–263, http://dx.doi.org/10.1016/j.jwpe.2017.05.002. [26] J. Zhao, Simulation on the Self-Assembly Behavior of Ordered Mesoporous Carbons and Their Performance in Adsorption (Master Dissertation) (Master dissertation), South China University of Technology, China, 2014. [27] Dassault Systèmes BIOVIA, Materials Studio, Dassault Systèmes, San Diego, 2016. [28] D. Zhao, P. Yang, N. Melosh, J. Feng, B.F. Chmelka, G.D. Stucky, Continuous mesoporous silica films with highly ordered large pore structures, Adv. Mater. 10 (1998) 1380–1385, http://dx.doi.org/10.1002/(SICI)1521-4095(199811)10:16

Suggest Documents