Article Cite This: Langmuir 2018, 34, 9349−9358
pubs.acs.org/Langmuir
Effect of Surface Chemistry on Confined Phase Behavior in Nanoporous Media: An Experimental and Molecular Modeling Study Evan Lowry* and Mohammad Piri College of Engineering and Applied Sciences, University of Wyoming, Laramie, Wyoming 82071, United States
Downloaded via UNIV OF NORTH DAKOTA on November 14, 2018 at 03:22:39 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
S Supporting Information *
ABSTRACT: It is well accepted that nanopore size is a controlling parameter in determining the phase behavior of confined adsorbate molecules. Despite this knowledge, the quantitative effect of surface chemistry on the confined phase behavior is a factor that remains obfuscated. Obtaining a complete understanding of the variables controlling confined phase behavior is a critical step in developing more complete equations of state for predictive modeling. To this end, a combined experimental and molecular modeling study was conducted to investigate the effects of surface chemistry and wetting on the confined phase behavior of propane and nbutane in modified and unmodified silica MCM-41. Isotherms were measured in four types of silica MCM-41 modified with varying sizes of alkyl groups to determine the effects of increasing surface modification. Results showed that increased pore surface coverage of carbon resulted in a notable change in the capillary condensation pressures, adsorption enthalpy, and confined critical temperature of the adsorbate. Correlations between the surface coverage of carbon and the confined critical temperature were presented and supported by thermodynamic arguments. The primary conclusions were partially supported by hybrid molecular dynamics-Monte Carlo simulations of propane adsorption in models of the four types of experimental adsorbents. Several differences were noted and explained between the experimental and modeling results. Energetic heterogeneity on the surface of the modified MCM-41 adsorbents as well as differences in adsorbate entropy induced by surface features and chemistry were suggested as primary driving factors for the observed trends. The results of this work have direct implications for improving understanding of confined phase behavior in materials of varying surface chemistries.
■
and Monte Carlo simulation studies.16−20 Using an empirical approach, Tan and Piri devised an equation of state for confined fluids based on coupling of perturbed-chain statistical associating fluid theory and the Young−Laplace equations.21 Despite these notable efforts, there remains a large amount of uncertainty regarding the key variables controlling phase behavior in confinement. Silica-based adsorbents have received much attention in research associated with capillary condensation primarily due to the relative ease of production via chemical templating and the regular, nanometric pore sizes associated with such materials.22 Adsorbents such as MCM-41 and SBA-15 have been routinely used to investigate confinement phenomena.23,24 Morishige et al. investigated the so-called capillary critical points of argon, nitrogen, ethylene, and carbon dioxide in MCM-41. They showed that the capillary critical temperature of these confined gases is much lower than the bulk critical point.3 In subsequent studies, the hysteresis temperature was differentiated from the confined critical temperature
INTRODUCTION The effects of decreasing pore sizes within microporous adsorbents have been investigated at length in the recent literature.1−4 This phenomenon is critically relevant to a number of industrial applications and fundamental research areas. Careful characterization of the type and extent of the effects of confinement on phase behavior is vital for engineering design in areas such as catalysis, oil and gas applications, and air pollution remediation.5−11 Previous studies have suggested that confinement-induced phase behavior, also termed and associated with capillary condensation, appears when pore sizes are in the range of 2−100 nm and when the pore diameters approach the mean free path of the adsorbate molecules.1,2 This behavior results in a first-, or nearly-first-, order phase transition that often occurs at lower pressures than the bulk vapor−liquid phase transition.12 Travalloni and colleagues produced a series of modified equations of state based on previous knowledge regarding capillary condensation.13−15 The fundamental variables in their models were based on the relationship between the molecular diameter and the pore radius, with the effects of fluid−pore wall interactions increasing as the pore size decreased. This correlative observation has been upheld in molecular dynamics © 2018 American Chemical Society
Received: March 28, 2018 Revised: June 12, 2018 Published: July 14, 2018 9349
DOI: 10.1021/acs.langmuir.8b00986 Langmuir 2018, 34, 9349−9358
Article
Langmuir
toward the bulk behavior.35,36 Experimentally, most previous investigations of surface−fluid interactions have focused on maximizing the adsorption capacity for the removal of volatile organic compounds (VOCs) from industrial waste effluent streams.37−41 Both Kim et al. and Wang et al. demonstrated that MCM-41 and SBA-15 modified with organic surface groups showed superior adsorption performance with common industrial solvents, including hexane, gasoline, benzene, and toluene.37,38 In spite of the aforementioned studies, there is still a lack of understanding regarding the fundamental mechanisms and driving factors that govern the interplay between adsorbent− adsorbate interactions and capillary condensation. To this end, experiments and simulations were conducted to develop a better understanding of these complex interactions. In the following section, the experimental methods and theory are presented. This is followed by detailed presentation and discussion of the results, which lead to the major conclusions of this work.
by noting a change in the linearity of the plot of logarithmic pressure versus temperature over a series of measured isotherms.4 The hysteresis temperature, confined critical temperature, and differential enthalpy of adsorption were observed to be dependent on the pore size in a number of other studies as well.25−27 Although there is a relatively significant amount of data available regarding the impact of pore size on capillary condensation, the effects of variation in adsorbate−adsorbent interactions are still poorly understood. The effects of wettability on the qualitative shape of adsorption isotherms were codified in a recent IUPAC report.28 Isotherms are generally classified by type. A “wetting” isotherm (type IV) is indicated by a downward concave shape in the low-pressure region, whereas nonwetting behavior (type V) is classified by an isotherm that exhibits upward concavity in the low-pressure region. Despite having evidence of the effects of surface chemistry on capillary condensation via isotherm shape, there remains very little explanation of the fundamentals governing this phenomenon. Attempts to model the effects of these interactions have been made by Gubbins et al. by introducing a wetting parameter to the traditional expressions for the grand partition function. This parameter is based on the ratio between the adsorbate−adsorbent and the adsorbate− adsorbate interactions. Although the parameter is rather esoteric, it was shown to fit experimental data fairly well for adsorption in carbon nanotubes with a variety of adsorbates and for MCM-41 with the adsorption of water.29 Other studies have focused on the industrial applications of surface modification in MCM-41. One such study showed that a hybrid organic/inorganic MCM-41 matrix enhanced the loading and release characteristics of ibuprofen. The discussion incorporated an investigation of both pore size and surface modification with aminopropyl groups.30 The outcome showed that the surface chemistry leveraged first-order effects on loading and desorption, whereas pore size was not a significant controlling factor. Mello et al. demonstrated that aminomodified MCM-41 served as a more efficient low-pressure adsorbent for CO2 capture due to the increased enthalpy of adsorption compared to that of unmodified MCM-41.31 The surface groups induced chemisorption at lower pressure that enhanced the low-pressure loading capacity of the modified MCM-41 material. Xu et al. demonstrated that adsorption capacity and kinetics are directly affected by the degree of surface modification in a silica MCM-41, which was modified with polyethylenimine (PEI) to varying degrees. It was observed that although the PEI decreased the pore size, it led to greater loading capacity for CO2 compared to pure MCM-41 and pure PEI adsorbents.32 Several Monte Carlo studies have also investigated the effects of interactions between surface and adsorbate. Puibasset et al. performed simulations in the grand canonical ensemble to determine the phase transition of a Lennard-Jones fluid within three different pore types. Using a regular, geometrically undulated, and chemically undulated pore, the adsorption behavior and coexistence regions were extracted from simulation. Results indicated that the geometrical undulation did not significantly impact the phase behavior of the fluid, whereas chemical undulations resulted in a larger hysteresis region as well as created apparent intermediate phases within the vapor−liquid coexistence region.33,34 Other simulations showed that changes in the attractive parameters to make the solid less wetting resulted in a shift of the phase behavior
■
METHODS
Preparation of Adsorbents and Fluids. Four types of adsorbents were synthesized based on standard templated mesoporous silica MCM-41. Pure silica MCM-41 as well as silica MCM-41 possessing surface modification with C1, C8, and C18 alkyl functionalities were synthesized by Galantreo, Ltd. The adsorbents were modified such that all had nearly the same pore diameter and resulted in alkyl bonding densities between 0.475 and 1.44 μmol/m2. Elemental analysis was provided by the manufacturer and was also independently confirmed using electron-dispersive spectroscopy (EDS). EDS was conducted with a Bruker X-Flash detector with an energy resolution of 121 eV. Measurements were taken over multiple sample particles with accelerating voltage of 10 kV and current of 0.4 nA. Spectra were acquired and subjected to matrix corrections before integrating and determining the averages. The results from Brunauer− Emmett−Teller (BET) and the average wt % of carbon from elemental analysis for each adsorbent are listed in Table 1.
Table 1. Results of BET Analysis for Silica MCM-41 with Various Surface Modifiers MCM-41 surface modifier surface area (m2/g) pore diameter (Å) pore volume (cm3/g) wt % C
none
C1
C8
C18
596 40 0.49 0
345 36 0.34 6.15
183 42 0.22 12.3
122 38 0.15 17.3
Pure n-butane and propane (99.995% purity) were obtained for use as the fluid phases during all experiments. The gravimetric adsorption apparatus was designed and fabricated to produce highly accurate temperature and mass measurements.42,43 Differential balances manufactured by Mettler Toledo were used to read the mass of the adsorption cells to an accuracy of 0.00001 g. Temperature was controlled within 0.1 K using a Thermotron environmental chamber. High-precision Rosemount pressure transducers monitored the positive pressure of the fluid in contact with the titanium adsorption cell, whereas Leybold vacuum gauges were used to read values below atmospheric pressure. Data was digitally logged and recorded for averaging after the experiment. More details about the apparatus used in this study can be found elsewhere.43 Experimental Procedure. Four adsorption cells were first cleaned and weighed before packing with each adsorbent and weighed again. The adsorption cells were installed into the apparatus, and the system was pressure tested at 400 psi to eliminate leaks. Special care was taken not to increase the temperature to a point 9350
DOI: 10.1021/acs.langmuir.8b00986 Langmuir 2018, 34, 9349−9358
Article
Langmuir
alkyl surface groups as well as for the propane adsorbate molecules.50 Once modified, each model was subjected to a 5 ns NVT equilibration sequence with a time step of 1 fs and temperature coupling every 10 steps to allow the surface alkyl chain molecules to achieve a minimum energy configuration. The void volume of each model was calculated using simulated helium porosimetry with a probe radius of 1.2 Å.51 To simulate adsorption, a hybrid NVTGCMC framework was implemented using the LAMMPS platform.52 This framework allows the surface groups and the adsorbate molecules to undergo translational and rotational motions between Monte Carlo steps. Each chemical potential and temperature condition was simulated for over 1 × 106 Monte Carlo steps until equilibrium was reached, as implied by stability of the total inserted particles. Isotherms, as well as thermodynamic data, were extracted from the raw simulation data and averaged over the last 50 000 simulation steps. Only dispersion interactions were considered in the simulations, and consequently surface hydrogen atoms were ignored.
where the surface functional groups may be chemically altered or removed from the pore surfaces. Jaroniec et al. performed a systematic study of the surface modification effects in silica MCM-41 using organosilane modification reactions. Thermogravimetric analysis showed that, unlike standard MCM-41, the organo-modified MCM41 had excellent thermal stability up to 100 °C, at which point the organosilane bonds began to decompose.44 Therefore, the system was regulated at 50 °C for 2 weeks while under high vacuum to remove contaminants from the surface of the adsorbents. The vacuum level reached 1 mbar, at which point it became stable. The initial isotherms at 16 °C for propane in the unmodified MCM-41 were compared to those recently published by Barsotti et al.45 The isotherms were found to match nearly identically, and therefore it was concluded that no significant adsorbed water remained within the system. Experiments were conducted at multiple different temperatures with propane and n-butane. To construct each isotherm, the pressure, temperature, and mass were recorded in real time while small amounts of fluid were injected to each adsorption cell. The pressure response was monitored for stability, which was taken as an indicator of thermodynamic equilibrium. Doses of fluid were progressively administered to each cell until after the bulk condensation point of the fluid was observed. At this point, the desorption isotherm was established by progressively placing the system under short periods of vacuum pressure to remove incremental amounts of fluid. The mass and pressure data points on the isotherm were extracted by taking an average over 100 time-series data points prior to the time at which a dose was administered. By doing this, each isotherm point was taken as close to equilibrium as possible. Between experimental temperatures, the system was placed under high vacuum (1 mbar) to return the adsorbents as close to the initial conditions as possible. NVT-Grand Canonical Monte Carlo (GCMC) Simulation. In an attempt to verify the results obtained from experiments at the molecular level, coupled grand canonical Monte Carlo (GCMC) and molecular dynamics (NVT) simulation was used to study the adsorption of propane on several different models of silica MCM41. GCMC simulation uses traditional Monte Carlo moves as well as particle insertion−deletion steps to match the model system to the thermodynamic characteristics of an imaginary reservoir. The grand canonical ensemble fixes the chemical potential (μ), system volume, and temperature during particle insertion−deletion steps and follows the metropolis criterion for determining probabilities in Monte Carlo moves.46 GCMC has been traditionally used for predicting phase behavior and adsorption.47,48 By coupling GCMC with NVT simulation, particles are allowed to translate and rotate according to the traditional molecular dynamics framework between GCMC moves. This allows the system to reach a minimum energy state more quickly and provides for more complex particle−surface interactions. Four different model pores were prepared to mimic the substrates used in the experimental work. Initially, a base silica MCM-41 pore was prepared as described previously.49 An algorithm was used to place different surface-modifying groups within the pores. First, undercoordinated silica was removed from the interior surface of the pore. Next, each undercoordinated surface oxygen atom was considered for bonding to a surface-modifying group. The surfacemodifying molecule with appropriate chain length was progressively “grown”, starting from an oxygen atom at the surface of the pore toward the central pore axis. Surface-modifying groups were placed randomly with the constraint that each progressive addition must be a maximal distance from other nearby groups. The algorithm was terminated once the desired level of surface carbon was reached, as determined by matching the characterization data in Table 1 as closely as possible. This procedure was used to create four pore types: unmodified MCM-41 and MCM-41 with methyl, octyl, and octadecyl group surface modifications. It was desirable to use a very simple model to approximate the adsorbents to aid in determining the controlling effects in the adsorption process. As a result, charge effects were not considered and only Lennard-Jones dispersive forces were used for simulation using a united-atom approach. The TRAPPE-UA model was used for the
■
RESULTS AND DISCUSSION Isotherms were extracted from the raw data and are reported in terms of the average fluid density within the pore. This quantity was calculated by first determining the pore volume from BET data in Table 1. Next, the amount of bulk phase fluid was subtracted from the total amount of fluid prior to dividing by the pore volume. The bulk phase density was extracted from data published by NIST at the equilibrium pressure for each point.53 By reporting the amount adsorbed in terms of average fluid density within the pores, direct comparisons were possible between each isotherm. On first observation, it is clear that tests conducted on the unmodified, C1-modified, and C8-modified MCM-41 adsorbents resulted in type IV isotherms for both propane and nbutane. All isotherms showed a downward concavity, which indicates favorable or wetting adsorption.28 The C18-modified MCM-41 did not present classically type IV isotherm behavior. Instead, the C18 MCM-41 isotherms appeared to induce continuous condensation along the lower region of the isotherm prior to reaching a capillary saturation plateau. For all isotherms in the modified MCM-41, the general trend appeared to be a decrease in the capillary condensation pressure as the degree of surface modification of the MCM-41 increased.40 This is attributable to the increased adsorbent− adsorbate interactions resulting from the alkyl surface groups. Additionally, the unmodified MCM-41 induced the largest average density in the adsorbed fluid within the pores, followed by the C18-, C8-, and C1-modified MCM-41 adsorbents. This observation is in agreement with the previous literature.54,55 One hypothesis for the lack of a distinct capillary condensation point in the C18-modified MCM-41 is that the adsorbate experienced premature condensation during the pore-filling region across the entire pressure range preceding the pore saturation plateau and bulk condensation. One surprising result is the fact that the unmodified MCM41 substrate resulted in the lowest capillary condensation pressure for both the propane and n-butane adsorbates. Similar observations were reported by Zhao and Lu for adsorption of benzene on MCM-41 and silylated MCM-4155 and by two other studies with toluene and trifluoromethane adsorbates.54,56 The explanation provided by Zhao and Lu was related to increased diffusive resistance in the pore-filling process due to the addition of methyl group surface modification.55 This is a plausible explanation for the difference observed between the capillary condensation pressures in the modified MCM-41 and those of the 9351
DOI: 10.1021/acs.langmuir.8b00986 Langmuir 2018, 34, 9349−9358
Article
Langmuir
Figure 1. Propane isotherms for standard, C1-, C8-, and C18-modified silica MCM-41. Isotherms for MCM-41-C8, MCM-41-C18, and MCM-41 are shifted vertically by 0.15, 0.3, and 0.45, respectively. MCM-41 isotherms are denoted by triangles, C1-modified MCM-41 is denoted by circles, C8-modified MCM-41 is denoted by squares, and C18-modified MCM-41 is denoted by pentagons. Open symbols denote desorption data.
unmodified MCM-41 in Figures 1 and 2. Another contributing factor could be related to the smoothness of the interior pore surface in the modified versus unmodified MCM-41.57 Nonuniform modification of the surface with alkyl groups may have resulted in a less homogeneous adsorption surface
that did not favor uniform adsorbate nucleation and simultaneous onset of capillary condensation in all areas of the adsorbent. These hypotheses are additionally supported by a qualitative review of the slope of the condensation step for each adsorbent. Isotherms in the modified MCM-41 9352
DOI: 10.1021/acs.langmuir.8b00986 Langmuir 2018, 34, 9349−9358
Article
Langmuir
Figure 2. n-Butane isotherms for standard, C1-, C8-, and C18-modified silica MCM-41. Isotherms for MCM-41-C8, MCM-41-C18, and MCM-41 are shifted vertically by 0.15, 0.3, and 0.45, respectively. MCM-41 isotherms are denoted by triangles, C1-modified MCM-41 is denoted by circles, C8-modified MCM-41 is denoted by squares, and C18-modified MCM-41 is denoted by pentagons. Open symbols denote desorption data. 9353
DOI: 10.1021/acs.langmuir.8b00986 Langmuir 2018, 34, 9349−9358
Article
Langmuir adsorbents appear to have a progressively smaller slope over the condensation region. This would indicate that capillary condensation in the modified substrates continuously occurred over a wider pressure range than in the unmodified MCM-41. Hysteresis was not pronounced for the propane isotherms but was clearly present in experiments with n-butane. For nbutane, the C18-modified MCM-41 displayed a hysteresis region at 2 °C, which extended to low-pressure, further supporting the hypothesis of adsorbate trapping and impediment within the long alkyl chains on the surface. Hysteresis rapidly disappeared for the C8- and C18-modified MCM-41 after 8 °C. The disappearance of hysteresis seemed to be weakly correlated to the degree of surface modification as well. The unmodified MCM-41 displayed some amount of hysteresis even above the estimated pore critical temperature. To evaluate the driving thermodynamic factors for the qualitative observations from the isotherms, the differential enthalpy of adsorption was calculated between two subcritical temperatures for each adsorbent−adsorbate pair. Classically, the differential enthalpy of adsorption may be calculated using the following thermodynamic relation Δads hT , Γ ij ∂ y jj ln(P*)zzz = − RT 2 k ∂T {Γ
Figure 4. Differential enthalpy of adsorption for n-butane in MCM-41 (triangles), MCM-41-C1 (circles), MCM-41-C8 (squares), and MCM-41-C18 (pentagons). Dashed line indicates the bulk value for standard enthalpy of vaporization of n-butane.
(1)
where P* is the equilibrium pressure divided by the standard pressure. By integrating over paths of constant loading (Γ), the differential enthalpy may be estimated from several adsorption isotherms. It is important to note that the estimate is only a good approximation over the low loading region of the adsorption isotherm because it is considered to be thermodynamically reversible. As such, the data in Figures 3 and 4 are only reported for low loading. For both propane and n-butane, the unmodified MCM-41 resulted in the largest values of ΔadshT,Γ at loading prior to
capillary condensation. This indicates that the adsorbate experienced stronger physisorption on the unmodified substrate, resulting in a larger reduction in entropy (as Δadss0T,Γ = ΔadshT,Γ − R ln(P/P*)) of the confined fluid compared to that of the other modified adsorbents. The lower entropy was a result of favorable siting of the adsorbate molecules in the pores of the unmodified surfaces. Although counterintuitive, this phenomenon could be explained by the smoothness of the interior pore surface of the unmodified substrate. The uniform decline with increased loading, which is observable in all cases, is indicative of an energetically heterogeneous adsorption and is due to the progressive filling of high-energy surface adsorption sites. A coupled effect of larger decrease in entropy and greater change in enthalpy upon adsorption in the unmodified MCM-41 is the best explanation for the lower condensation pressures observed in the experimental isotherms for the unmodified MCM-41.58 The results of the GCMC simulation are also shown in Figure 3. The simulations did not match the experimental data for the unmodified MCM-41. For the modified adsorbents, the modeling results showed better agreement with the experimental counterparts but were not able to account for the significant heterogeneity that was present in the experimental systems. Confined Critical Temperature. The depression of the critical point in confinement is a phenomenon that has been well documented in the literature.3,4,25−27 As such, the confined critical points of propane and n-butane were calculated in all types of adsorbent materials used. The method of Nardon and Larher was used to calculate the confined critical temperature (Tcp) based on the inverse slope of pressure versus loading over the capillary condensation region of the isotherm.3,59 The inverse slope changes drastically once the critical point has been reached, allowing for the estimation of the critical point from multiple isotherms over a temperature range that includes the critical point.59 Only the points that were decidedly above the critical point,
Figure 3. Differential enthalpy of adsorption for propane in MCM-41 (triangles), MCM-41-C1 (circles), MCM-41-C8 (squares), and MCM-41-C18 (pentagons). Dashed line indicates the bulk value for the standard enthalpy of vaporization of propane. The open symbols are the values obtained from the GCMC simulation of propane adsorption in the four model adsorbents (with linear fit line). 9354
DOI: 10.1021/acs.langmuir.8b00986 Langmuir 2018, 34, 9349−9358
Article
Langmuir
For both propane and n-butane, Tcp decreases with increase in surface modification. This observation is consistent with previous GCMC simulation results.35,49 Tcp for propane followed a linear trend described by eq 2, whereas Tcp for nbutane was described by eq 3, where Xcarbon represents the wt % of surface carbon and Tcp,MCM‑41 is the confined critical temperature for unmodified MCM-41.
based on the inverse slope of the capillary condensation region, were used to fit the supercritical line. The Tcp estimates from the desorption branch for the nbutane isotherms rendered values that were typically within 1 °C of the adsorption branch. The critical temperatures of propane and n-butane in all of the adsorbent types are shown in Figure 5. From the figure, it is clear that there is a trend of decreasing confined critical temperature with increasing surface modification.
Tcp = Tcp,MCM ‐ 41 − 0.565Xcarbon
(2)
Tcp = Tcp,MCM ‐ 41 − 0.982Xcarbon
(3)
From Figure 5, there appears to be a correlation between the slope of the decline in Tcp with increased modification and the size of the n-alkane adsorbate. Although more adsorption data for larger n-alkanes would be needed to confirm a general trend, this observation may possess a qualitative explanation. One possible hypothesis is that n-butane presents a steeper decline in Tcp as compared to propane due to the additional postadsorption entropy retained through conformational isomerism. This additional, albeit small, rotational entropy retained by n-butane may become much more significant in confinement where translational motion is restricted. As a result, n-butane may advance toward confined criticality more quickly with the introduction of surface-modifying alkyl groups. NVT-GCMC Simulations. Propane adsorption was simulated over a range of temperature and chemical potential values in the four types of model pores shown in Figure 6. A hybrid NVT-GCMC scheme was used that allowed the surface groups and the adsorbate molecules to move throughout the simulations. The raw data from the simulations were used to extract the grand potential using eqs 4 and 520,49,60 ij ∂ϕG yz jj z = −⟨N ⟩ j ∂μ zz k {T , V
Figure 5. Confined critical temperature of propane and n-butane in silica MCM-41 with as a function of the amount of surface modification (surface bonded carbon). Dotted lines are shown for reference.
(4)
Figure 6. Model silica MCM-41 pores with (a) no surface modification, (b) C1 surface groups, (c) C8 surface groups, and (d) C18 surface groups. Pores are shown during an NVT equilibration sequence. 9355
DOI: 10.1021/acs.langmuir.8b00986 Langmuir 2018, 34, 9349−9358
Article
Langmuir
jij ∂ϕG /T zyz = ⟨E − μN ⟩ jj zz k ∂1/T { μ , V
unmodified MCM-41 may not adequately represent the real material. This discrepancy is also apparent in Figure 3 where the differential enthalpy of adsorption does not match well with the experimental data for the unmodified MCM-41. This might be attributed to the effects of Coulomb charge on the surface of the unmodified MCM-41 in the adsorption process, which could not be adequately represented by the relatively simple Lennard-Jones model employed here. The computed confined critical temperatures of propane are presented in Figure 8 as a function of the pore type. The
(5)
The grand potential was used to locate the phase transition point using the relation in eq 6 jij ∂ϕG zyz = 0 jj zz k ∂N {T , μ
(6)
The critical temperature can be estimated from the grand potential using methods previously presented in the literature.60 The grand potential is related to the average pressure for a homogeneous fluid by the relation in eq 7 and therefore is a useful way to present isotherm data from GCMC simulation. ϕG = −⟨P⟩V
(7)
Figure 7 shows simulated propane isotherms for all pore types. It can be noted that, with the exception of the unmodified
Figure 8. Confined critical temperature of propane estimated by GCMC in four types of model pores with varying surface modifications. Experimental results are shown for reference.
experimental values are shown in the plot for reference. Both experimental and simulated data show the same, nearly linear, negatively sloped trend. The simulations produced generally lower values of Tcp for the modified MCM-41 models. This is likely due to the lack of heterogeneity in the model that was present in the actual substrates. The magnitude of this discrepancy increased for the more modified surfaces, indicating that significant heterogeneity may have been present in the modified MCM-41 materials used in the experiments. This heterogeneity, likely caused by uneven modification of the pore surfaces, would result in large energetic differences in the adsorption surface. The simulated value of Tcp for the unmodified silica, MCM-41, was interestingly similar to the experimental value despite the differences in the qualitative isotherm behavior and the differential enthalpy of adsorption.
Figure 7. Propane isotherms at 210 K from GCMC simulation with four types of modified and unmodified silica MCM-41. Isotherms are plotted in terms of the normalized grand potential.
MCM-41 model, the simulated isotherms follow the same trend as the experimental data. The capillary condensation point occurs at lower grand potential values for the pore systems with more alkyl group surface modification. The density of the fluid is larger in the C1-modified MCM41 compared to that in the C8- and C18-modified MCM-41, likely due to the better packing configurations achieved with the lack of the long alkyl surface modifiers. The larger values of normalized grand potential in the vapor adsorption region of the unmodified MCM-41 simulated isotherms indicate a combination of higher entropy of the adsorbate and lower surface interaction with the substrate, which is different from what was observed experimentally. The fact that the simulated isotherms for MCM-41 resulted in the largest capillary condensation pressure of the four model pores, in contrast to the experiment, indicates that the potentials used to model
■
CONCLUSIONS In this study, both experimental and modeling techniques were employed to investigate the effects of surface chemistry on the adsorption and thermodynamic behaviors of propane and nbutane. Four types of MCM-41 were created with varying degrees of surface modification. The MCM-41 was modified using silylation reactions to bond methyl, octyl, and octadecyl groups to the surface of pure MCM-41. These adsorbents were used to measure adsorption isotherms of both propane and n9356
DOI: 10.1021/acs.langmuir.8b00986 Langmuir 2018, 34, 9349−9358
Article
Langmuir
SPE Annual Technical Conference and Exhibition; Society of Petroleum Engineers, 2013. (3) Morishige, K.; Fujii, H.; Uga, M.; Kinukawa, D. Capillary Critical Point of Argon, Nitrogen, Oxygen, Ethylene, and Carbon Dioxide in MCM-41. Langmuir 1997, 13, 3494−3498. (4) Morishige, K.; Nakamura, Y. Nature of Adsorption and Desorption Branches in Cylindrical Pores. Langmuir 2004, 20, 4503−4506. (5) Jia, M.; Seifert, A.; Thiel, W. R. Mesoporous MCM-41 Materials Modified with Oxodiperoxo Molybdenum Complexes: Efficient Catalysts for the Epoxidation of Cyclooctene. Chem. Mater. 2003, 15, 2174−2180. (6) Gelb, L. D.; Gubbins, K. E.; Radhakrishnan, R.; SliwinskaBartkowiak, M. Phase separation in confined systems. Rep. Prog. Phys. 2000, 63, 727. (7) Ye, G.; Zhou, X.; Yuan, W.; Coppens, M.-O. Probing pore blocking effects on multiphase reactions within porous catalyst particles using a discrete model. AIChE J. 2016, 62, 451−460. (8) Arogundade, O.; Sohrabi, M. A. In Review of Recent Developments and Challenges in Shale Gas Recovery, SPE Saudi Arabia Section Technical Symposium and Exhibition; Society of Petroleum Engineers, 2012. (9) Chen, J.-H.; Mehmani, A.; Li, B.; Georgi, D.; Jin, G. In Estimation of Total Hydrocarbon in the Presence of Capillary Condensation for Unconventional Shale Reservoirs, SPE Middle East Oil and Gas Show and Conference; Society of Petroleum Engineers, 2013. (10) Nojabaei, B.; Johns, R. T.; Chu, L. Effect of Capillary Pressure on Phase Behavior in Tight Rocks and Shales. SPE Reservoir Eval. Eng. 2013, 16, 281−289. (11) Shim, W. G.; Lee, J. W.; Moon, H. Heterogeneous Adsorption Characteristics of Volatile Organic Compounds (VOCs) on MCM48. Sep. Sci. Technol. 2006, 41, 3693−3719. (12) Thommes, M.; Findenegg, G. H. Pore Condensation and Critical-Point Shift of a Fluid in Controlled-Pore Glass. Langmuir 1994, 10, 4270−4277. (13) Travalloni, L.; Castier, M.; Tavares, F. W.; Sandler, S. I. Thermodynamic modeling of confined fluids using an extension of the generalized van der Waals theory. Chem. Eng. Sci. 2010, 65, 3088− 3099. (14) Travalloni, L.; Castier, M.; Tavares, F. W. Phase equilibrium of fluids confined in porous media from an extended Peng-Robinson equation of state. Fluid Phase Equilib. 2014, 362, 335−341. (15) Travalloni, L.; Castier, M.; Tavares, F. W.; Sandler, S. I. Critical behavior of pure confined fluids from an extension of the van der Waals equation of state. J. Supercrit. Fluids 2010, 55, 455−461. (16) Pitakbunkate, T.; Balbuena, P. B.; Moridis, G. J.; Blasingame, T. A.; et al. Effect of confinement on pressure/volume/temperature properties of hydrocarbons in shale reservoirs. SPE J. 2016, 21, 621− 634. (17) Zarragoicoechea, G. J.; Kuz, V. A. Critical shift of a confined fluid in a nanopore. Fluid Phase Equilib. 2004, 220, 7−9. (18) Jin, B.; Nasrabadi, H. Phase behavior of multi-component hydrocarbon systems in nano-pores using gauge-GCMC molecular simulation. Fluid Phase Equilib. 2016, 425, 324−334. (19) Shirono, K.; Daiguji, H. Molecular simulation of the phase behavior of water confined in silica nanopores. J. Phys. Chem. C 2007, 111, 7938−7946. (20) Page, K. S.; Monson, P. A. Monte Carlo calculations of phase diagrams for a fluid confined in a disordered porous material. Phys. Rev. E 1996, 54, 6557. (21) Tan, S. P.; Piri, M. Equation-of-state modeling of associatingfluids phase equilibria in nanopores. Fluid Phase Equilib. 2015, 405, 157−166. (22) Hanrahan, J. P.; Copley, M. P.; Ziegler, K. J.; Spalding, T. R.; Morris, M. A.; Steytler, D. C.; Heenan, R. K.; Schweins, R.; Holmes, J. D. Pore Size Engineering in Mesoporous Silicas Using Supercritical CO2. Langmuir 2005, 21, 4163−4167.
butane at multiple temperatures. Additionally, hybrid NVTGCMC modeling was used to simulate adsorption in four model pores that were analogous to the experimental adsorbents using a united-atom, Lennard-Jones interaction scheme. The surface modification directly impacted the shape of the isotherms as well as the differential enthalpy of adsorption. Interestingly, the unmodified MCM-41 presented the lowest capillary condensation pressures and the highest differential enthalpy of adsorption, pointing to energetically favorable adsorption behavior and larger reductions in entropy upon adsorption. These findings were explained by diffusive resistance caused by the surface-modifying groups and heterogeneity caused by nonuniform modification of the surface. The combination of these two factors resulted in less reduction in entropy upon adsorption as well as causing capillary condensation to occur over a wider pressure range as compared to that in the unmodified MCM-41. The experimental and simulated isotherms were used to calculate the confined critical temperature (Tcp) of both fluids. Both calculations supported a negatively sloped trend of Tcp with increasing alkyl surface modification. The slope of this trend also appeared to be correlated to the adsorbate size; however, more data are needed to confirm a more general behavior. The results of the simulations did not closely agree with their experimental counterparts for the unmodified MCM-41 but showed encouragingly similar results for the modified substrates in terms of qualitative capillary condensation pressure behavior and the calculated differential enthalpy of adsorption. These results suggest that impacts of surface charge effects are non-negligible for adsorption on unmodified MCM41, even when the adsorbate is a nonpolar molecule. The results presented in this work have direct implications for developing a better understanding regarding the effects of surface chemistry on confined phase behavior.
■
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.langmuir.8b00986. Adsorption and desorption isotherms for propane and nbutane (PDF)
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. ORCID
Evan Lowry: 0000-0001-5375-9852 Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS The authors gratefully acknowledge the financial support of Saudi Aramco and the School of Energy Resources at the University of Wyoming.
■
REFERENCES
(1) Barsotti, E.; Tan, S. P.; Saraji, S.; Piri, M.; Chen, J.-H. A review on capillary condensation in nanoporous media: Implications for hydrocarbon recovery from tight reservoirs. Fuel 2016, 184, 344−361. (2) Jin, L.; Ma, Y.; Jamili, A. In Investigating the Effect of Pore Proximity on Phase Behavior and Fluid Properties in Shale Formations, 9357
DOI: 10.1021/acs.langmuir.8b00986 Langmuir 2018, 34, 9349−9358
Article
Langmuir (23) Neimark, A. V.; Ravikovitch, P. I. Capillary condensation in MMS and pore structure characterization. Microporous Mesoporous Mater. 2001, 44−45, 697−707. (24) Gor, G. Y.; Paris, O.; Prass, J.; Russo, P. A.; Ribeiro Carrott, M. M. L.; Neimark, A. V. Adsorption of n-Pentane on Mesoporous Silica and Adsorbent Deformation. Langmuir 2013, 29, 8601−8608. (25) Qiao, S. Z.; Bhatia, S. K.; Nicholson, D. Study of Hexane Adsorption in Nanoporous MCM-41 Silica. Langmuir 2004, 20, 389− 395. (26) Russo, P. A.; Ribeiro Carrott, M. M. L.; Carrott, P. J. M. Trends in the condensation/evaporation and adsorption enthalpies of volatile organic compounds on mesoporous silica materials. Microporous Mesoporous Mater. 2012, 151, 223−230. (27) Tanchoux, N.; Trens, P.; Maldonado, D.; di Renzo, F.; Fajula, F. The adsorption of hexane over MCM-41 type materials. Colloids Surf., A 2004, 246, 1−8. (28) Thommes, M.; Kaneko, K.; Neimark, A. V.; Olivier, J. P.; Rodriguez-Reinoso, F.; Rouquerol, J.; Sing, K. S. Physisorption of gases, with special reference to the evaluation of surface area and pore size distribution (IUPAC Technical Report). Pure Appl. Chem. 2015, 87, 1051−1069. (29) Gubbins, K. E.; Long, Y.; Asliwinska-Bartkowiak, M. Thermodynamics of confined nano-phases. J. Chem. Thermodyn. 2014, 74, 169−183. (30) Muñoz, B.; Ramila, A.; Perez-Pariente, J.; Diaz, I.; Vallet-Regi, M. MCM-41 Organic Modification as Drug Delivery Rate Regulator. Chem. Mater. 2003, 15, 500−503. (31) Mello, M. R.; Phanon, D.; Silveira, G. Q.; Llewellyn, P. L.; Ronconi, C. M. Amine-modified MCM-41 mesoporous silica for carbon dioxide capture. Microporous Mesoporous Mater. 2011, 143, 174−179. (32) Xu, X.; Song, C.; Andresen, J. M.; Miller, B. G.; Scaroni, A. W. Preparation and characterization of novel CO2 molecular basket adsorbents based on polymer-modified mesoporous molecular sieve MCM-41. Microporous Mesoporous Mater. 2003, 62, 29−45. (33) Puibasset, J. Grand Potential, Helmholtz Free Energy, and Entropy Calculation in Heterogeneous Cylindrical Pores by the Grand Canonical Monte Carlo Simulation Method. J. Phys. Chem. B 2005, 109, 480−487. (34) Puibasset, J. Capillary Condensation in a Geometrically and a Chemically Heterogeneous Pore: A Molecular Simulation Study. J. Phys. Chem. B 2005, 109, 4700−4706. (35) Singh, S. K.; Saha, A. K.; Singh, J. K. Molecular Simulation Study of Vapor- Liquid Critical Properties of a Simple Fluid in Attractive Slit Pores: Crossover from 3D to 2D. J. Phys. Chem. B 2010, 114, 4283−4292. (36) Singh, S. K.; Sinha, A.; Deo, G.; Singh, J. K. Vapor- liquid phase coexistence, critical properties, and surface tension of confined alkanes. J. Phys. Chem. C 2009, 113, 7170−7180. (37) Kim, S.-H.; Park, J.-H.; Lee, C.-Y. Surface-functionalized mesoporous silica nanoparticles as sorbents for BTEX. J. Porous Mater. 2013, 20, 1087−1093. (38) Wang, H.; Wang, T.; Han, L.; Tang, M.; Zhong, J.; Huang, W.; Chen, R. VOC adsorption and desorption behavior of hydrophobic, functionalized SBA-15. J. Mater. Res. 2016, 31, 516−525. (39) Dou, B.; Hu, Q.; Li, J.; Qiao, S.; Hao, Z. Adsorption performance of VOCs in ordered mesoporous silicas with different pore structures and surface chemistry. J. Hazard. Mater. 2011, 186, 1615−1624. (40) Hu, Q.; Li, J. J.; Hao, Z. P.; Li, L. D.; Qiao, S. Z. Dynamic adsorption of volatile organic compounds on organofunctionalized SBA-15 materials. Chem. Eng. J. 2009, 149, 281−288. (41) Zandavi, S. H.; Ward, C. A. Contact angles and surface properties of nanoporous materials. J. Colloid Interface Sci. 2013, 407, 255−264. (42) Barsotti, E.; Saraji, S.; Piri, M. Nanocondensation Apparatus. U.S. Patent US15/588,094, 2017. (43) Barsotti, E.; Saraji, S.; Tan, S. P.; Piri, M. Capillary Condensation of Binary and Ternary Mixtures of n-Pentane,
Isopentane, and CO2 in Nanopores: An Experimental Study on the Effects of Composition and Equilibrium. Langmuir 2018, 34, 1967− 1980. (44) Jaroniec, C. P.; Kruk, M.; Jaroniec, M.; Sayari, A. Tailoring Surface and Structural Properties of MCM-41 Silicas by Bonding Organosilanes. J. Phys. Chem. B 1998, 102, 5503−5510. (45) Barsotti, E.; Tan, S. P.; Piri, M.; Chen, J.-H. Phenomenological Study of Confined Criticality: Insights from the Capillary Condensation of Propane, n-Butane, and n-Pentane in Nanopores. Langmuir 2018, 34, 4473−4483. (46) Adams, D. J. Grand canonical ensemble Monte Carlo for a Lennard-Jones fluid. Mol. Phys. 1975, 29, 307−311. (47) Peterson, B. K.; Gubbins, K. E. Phase transitions in a cylindrical pore: grand canonical Monte Carlo, mean field theory and the Kelvin equation. Mol. Phys. 1987, 62, 215−226. (48) Orkoulas, G.; Panagiotopoulos, A. Z. Phase behavior of the restricted primitive model and square-well fluids from Monte Carlo simulations in the grand canonical ensemble. J. Chem. Phys. 1999, 110, 1581−1590. (49) Lowry, E.; Piri, M. Effects of chemical and physical heterogeneity on confined phase behavior in nanopores. Microporous Mesoporous Mater. 2018, 263, 53−61. (50) Martin, M. G.; Siepmann, J. I. Transferable potentials for phase equilibria. 1. United-atom description of n-alkanes. J. Phys. Chem. B 1998, 102, 2569−2577. (51) Willems, T. F.; Rycroft, C. H.; Kazi, M.; Meza, J. C.; Haranczyk, M. Algorithms and tools for high-throughput geometrybased analysis of crystalline porous materials. Microporous Mesoporous Mater. 2012, 149, 134−141. (52) Plimpton, S. Fast Parallel Algorithms for Short-Range Molecular Dynamics. J. Comput. Phys. 1995, 117, 1−19. (53) Linstrom, P. J.; Mallard, W. G. The NIST Chemistry WebBook: A Chemical Data Resource on the Internet. J. Chem. Eng. Data 2001, 46, 1059−1063. (54) Russo, P. A.; Carrott, M. M. L. R.; Carrott, P. J. Hydrocarbons adsorption on templated mesoporous materials: effect of the pore size, geometry and surface chemistry. New J. Chem. 2011, 35, 407−416. (55) Zhao, X. S.; Ma, Q.; Lu, G. Q. M. VOC Removal: Comparison of MCM-41 with Hydrophobic Zeolites and Activated Carbon. Energy Fuels 1998, 12, 1051−1054. (56) Schreiber, A.; Bock, H.; Schoen, M.; Findenegg, G. H. Effect of surface modification on the pore condensation of fluids: experimental results and density functional theory. Mol. Phys. 2002, 100, 2097− 2107. (57) Long, Y.; Xu, T.; Sun, Y.; Dong, W. Adsorption behavior on defect structure of mesoporous molecular sieve MCM-41. Langmuir 1998, 14, 6173−6178. (58) Pascal, T. A.; Goddard, W. A.; Jung, Y. Entropy and the driving force for the filling of carbon nanotubes with water. Proc. Natl. Acad. Sci. U.S.A. 2011, 108, 11794−11798. (59) Nardon, Y.; Larher, Y. Critical temperatures of two-dimensional condensation of methane and rare gases on the cleavage face of layerlike halides. Surf. Sci. 1974, 42, 299−305. (60) Panagiotopoulos, A. Z. Molecular simulation of phase coexistence: Finite-size effects and determination of critical parameters for two-and three-dimensional Lennard-Jones fluids. Int. J. Thermophys. 1994, 15, 1057−1072.
9358
DOI: 10.1021/acs.langmuir.8b00986 Langmuir 2018, 34, 9349−9358
