Effect of Pore Structure on CO2 Adsorption Characteristics of

Mar 26, 2015 - It was determined that the introduction of amine polymer deteriorates the CO2 ... Aminopolymer Mobility and Support Interactions in Sil...
1 downloads 0 Views 1MB Size
Download PDF
Article pubs.acs.org/Langmuir

Effect of Pore Structure on CO2 Adsorption Characteristics of Aminopolymer Impregnated MCM-36 Christopher F. Cogswell, Hui Jiang, Justin Ramberger, Daniel Accetta, Ronald J. Willey, and Sunho Choi* Department of Chemical Engineering, Northeastern University, 313 Snell Engineering Center, 360 Huntington Avenue, Boston, Massachusetts 02115-5000, United States

ABSTRACT: The CO2 adsorption characteristics of a pillared 2-dimensional porous silicate material impregnated with amine containing polymers have been investigated. It was determined that the introduction of amine polymer deteriorates the CO2 capture kinetics of the MCM-36 supported amine adsorbents compared to that of the bare material, due to the fact that with the addition of a higher loading of amine polymer the diffusion of CO2 through the 2-dimensional interlayer mesoporous channels of MCM-36 becomes greatly hindered. This pore blocking sets an upper limit to the CO2 capture performance of the polymer impregnated MCM-36 and greatly reduces the utility of using this sort of amine−solid adsorbent for carbon capture. Interestingly, these results suggest that for 2-D channel solid supports there is an optimal amine loading which is not likely to be equal to the maximum loading, and which can be determined and utilized to obtain the maximum improvement over the original materials. The study performed in this work for the MCM-36 material could therefore be applied to other porous supports to determine these optimum loadings and be used to more easily compare and contrast the alterations to capture characteristics which occur upon amine loading for a wide range of materials. It is believed this will make it more straightforward to determine which solid supports hold the promise for greatly improved capture characteristics upon amine loading and allow the field to more quickly determine avenues for fruitful development. These results also suggest the need for a new sort of support structure for amine loaded solids, one which can allow us to decouple amine loading from increasing diffusion resistance so that high amine efficiency can be maintained throughout the range of increased amine loadings.

■

INTRODUCTION Porous solid adsorbents such as zeolites and metal−organic frameworks have attracted widespread attention for their potential use as carbon dioxide capture systems due to their high specific surface areas and low heats of adsorption.1−10 For solid adsorbents to become competitive with currently utilized capture systems such as aqueous amine absorbents, they must be capable of achieving high selectivity for carbon dioxide over other gas stream components as well as showing improved capture capacity.1,11 In order to achieve these performance requirements, many groups have looked at the incorporation of amine-containing groups into porous solids either by impregnation or covalent functionalization.1,2,10−32 These amine-incorporated solid adsorbents often achieve higher adsorption capacity and enhanced selectivity for carbon dioxide even in low partial pressure streams such as those for flue gases (CO2 partial pressure of 10−15%) compared to bare solid adsorbents without amines. On the other hand, as the amine © 2015 American Chemical Society

loading in the solid supports increases, the CO2 capture capacity and adsorption kinetics of the supported amine adsorbents often become depreciated. For instance, previous reports indicated that the amine efficiency (a measure of adsorption capacity defined as the CO2 amount adsorbed divided by the amine quantity) and the adsorption half time (a measure of adsorption kinetics, defined as the time to reach half the equilibrium capacity) of supported amine adsorbents tend to be degraded especially at high amine loadings.1,11,26,33 It is believed this occurs probably because increased amine loading onto the solid support limits the ability for entering molecules to diffuse throughout the solid material to internal amine sites.1,2,11,13,15,16,22,26 Despite this large literature background, quantitative analyses of the structure− Received: January 2, 2015 Revised: March 2, 2015 Published: March 26, 2015 4534

DOI: 10.1021/la505037f Langmuir 2015, 31, 4534−4541

Article

Langmuir

Figure 1. X-ray diffraction patterns of (a) MCM-22-P, (b) swollen MCM-22-P, (c) MCM-36, (d) MCM-22 zeolite, and (e) 14.5 wt % PEI loaded MCM-36.

property relationships of supported amine adsorbents, for example those between the amine loading and the adsorption capacity as well as those between the amine loading and the adsorption kinetics, have not been performed in detail so far. This gap in the literature has caused many in the field to attempt increasing the amine loading on the support to increase capture performance, without taking into account the changes to the transport properties of carbon dioxide through the material when designing these capture systems. In our group’s previous study, we have shown that pore blocking effects occur within the 1-dimensional channels of SBA-15, with increased amine impregnation leading to greatly reduced capture performance.33 In this paper, the influence of the amine loading on adsorption capacity and kinetics of adsorption for a material with 2-dimensional pore channels is studied systematically in order to determine a potential route to achieve a final material with both high amine loading and high adsorption capacity. The ultimate goal of this work is to optimize a potential route toward a final material that has both high amine loading and high adsorption capacity through the study of the changes to the diffusion and kinetic characteristics of the amine-loaded material. MCM-36 was chosen as the material for study due to the presence of a large pore interlayer space, which allows diffusion of CO2 molecules through the aand b-axis, and due to its well-known synthetic scheme.5,11,34−37 This interlayer space can also be preferentially loaded with amines due to the presence of surface silanol groups via covalent bonding or through selective impregnation due to the size difference between the micropores present in the layers versus the large pore channels created via pillaring. The polymeric amine polyethylenimine (PEI) is used as an organic component of the supported amine adsorbents, with varying polymer loadings within the adsorbents used. PEI was chosen as the polymer to be loaded because of the large literature background on the impregnation of PEI onto porous silica substrates, especially for carbon dioxide capture. These sets of polymer loaded adsorbents are characterized in terms of structural parameters such as surface area and effective pore size (pore size measured by physisorption instrument after amine incorporation) to be correlated with property parameters

including amine efficiency, adsorption half time, and initial capture rate.

■

EXPERIMENTAL SECTION

Synthesis of the MCM-22-P Precursor. MCM-22-P was prepared hydrothermally according to methods similar to those found in the literature.2,4,5,34−39 MCM-22-P was then used for the synthesis of MCM-36 as described below, and the resulting crystals were then impregnated with aminopolymers to create the adsorbents used to capture CO2 capture in this work. To create the precursor synthesis gel, first 155.5 g of DI water, 1.24 g of sodium hydroxide (NaOH Pellets, FisherSci), and 0.361 g of sodium aluminate (NaAlO2, Sigma) were mixed together to form a cloudy solution. Following this, 11.8 g of Cab-O-Sil M5 (Fisher) and 9.55 g of hexamethylenimine (HMI, 99% Sigma) were added to generate the reactant gel. The reactant gel was then stirred for 6 h vigorously at room temperature and reacted hydrothermally at 130 °C for 9 days. The resulting solids of MCM-22-P were recovered via repeated washing, filtration, and drying, and then used for MCM-36 synthesis or to generate fully zeolitic MCM-22 as a comparison for surface characteristics and pore connectivity. MCM-22 zeolites were prepared via calcination at 853 K for 3 h under air. Synthesis of MCM-36 as a Support for CO2 Adsorbent. Silicapillared MCM-36 was synthesized via swelling and pillaring of MCM22-P following a known procedure.34−38,40 For instance, swelling was carried out by the intercalation of cetyltrimethylammonium bromide (CTAB, 99% Fisher) in the interlayer space of the MCM-22-P while simultaneously removing the hexamethylenimine structural directing agent by the addition of tetrapropylammonium hydroxide (TPAOH, 20−25% in water, Fisher). Generally 0.5 g of the wet MCM-22-P cake was mixed with 2 g of DI water and then introduced to a cetyltrimethylammonium bromide solution composed of 2.83 g of CTAB dissolved in 4.025 g of DI water. Finally, 6.175 g of the TPAOH solution was added dropwise at 30 °C with vigorous stirring, and the solution was then allowed to react with stirring for 24 h, after which time the solid was collected via a series of washing with methanol and decanting steps. Afterward, intercalation and calcination of tetraethyl orthosilicate (TEOS, Sigma) were performed to yield the silica pillared MCM-36 material.36,37 Specifically, the swollen MCM22-P was added to TEOS in a 1:6 weight ratio and allowed to stir at 80 °C for 24 h in a nitrogen environment to yield the swollen derivative which has been intercalated with TEOS. Following this, the sample was allowed to hydrolyze in DI water in a 1:10 solid to water weight ratio for 5 h at a controlled pH of 8 and at room temperature. This solid was then collected via washing and decanting, dried overnight, 4535

DOI: 10.1021/la505037f Langmuir 2015, 31, 4534−4541

Article

Langmuir then calcined in air for 6 h at 450 °C and then at 580 °C for 3 h (temperature ramp rate: 2 °C/min) to produce MCM-36. Synthesis of CO2 Adsorbent via Aminopolymer Impregnation. After pillaring, the MCM-36 samples were impregnated with polyethylenimine (PEI, average MW ∼ 800 by LS, Sigma) to create a supported-amine adsorbent. Usually 0.2 g of MCM-36 solid (which had been held under vacuum for 24 h at 100 °C before reaction) was added to varying amounts of a 11.11 wt % PEI in methanol solution to form differing concentrations of MCM-36 to PEI-solution (for instance, 0.2 g of MCM to 7.2 g of solution, 0.2 to 14.4 g of solution, etc.). The MCM-36 was added to this solution and allowed to stir for 6 h at room temperature with reflux. After reaction the sample was collected via washing with methanol and decanting, followed by vacuum filtration and drying overnight at 80 °C. Samples were named as PEI-MCM-x, where x is the weight percent (wt %) of polymer loaded as determined by thermogravimetric analysis outlined below. For instance, the material with a 16.3 wt % of polymer is named PEIMCM-16.3. Characterization and Carbon Dioxide Capture. For all samples multiple tests were performed in order to characterize the surface and porous features of the materials. X-ray diffraction patterns were obtained using a Rigaku Ultima IV with a Cu Kα source. Fourier transform infrared resonance spectroscopy (FTIR) data were obtained using a Bruker Vertex system with OPUS analysis software. Nitrogen adsorption/desorption isotherms were collected using a Quantachrome NOVA 2200 e Series pore size analyzer and were analyzed using the NOVA software to determine the BET surface area and BJH pore size and volume. Carbon dioxide capture data were collected using a TA Q500 thermogravimetric analysis (TGA) system. Capture capacities were obtained by first degassing the samples at 120 °C in nitrogen for 3 h, followed by cooling back to 25 °C. Following this, pure carbon dioxide gas was flown into the furnace for 5 h. This data was used to obtain the total capture capacity of carbon dioxide, the initial capture rate, and the equilibrium time for each sample. The total amine content of each impregnated sample was obtained by heating the samples at a ramp rate of 10 °C/min up to 800 °C. This was compared with the bare material weight percent curve to determine the total polymer loading.

of PEI has not significantly altered the layer structure but only the interlayer space. Figure 1d, which represents the MCM-22 derivative obtained by calcining the precursor MCM-22-P crystal, is given as a means of further verifying the identity of the precursor crystal used in this study. FTIR analysis shows evidence for the physical impregnation of amines into the pore channels of the MCM-36 crystal. As can be seen in Figure 2, all samples show the characteristic

Figure 2. FTIR patterns for (a) MCM-22-P swollen, (b) MCM-36, and (c) 14.5 wt % PEI-loaded MCM-36.

peaks below ∼1250 cm−1 of an aluminosilica porous solid framework.6 Specifically, the peaks at 1225, 1077, 830, and 651 cm−1 are unchanged following functionalization. Figure 2b represents the bare MCM-36 crystal after pillaring with silica, and all peaks present can be attributed to the porous structure or hydroxyl groups on the surface or as entrapped water. The peak at 1675 cm−1 is characteristic of hydroxyl groups in the porous space, with the peak at 2430 cm−1 being assigned to CO2 captured from the air during handling of the sample. The broad peak spanning from around 3200−3600 cm−1 is characteristic of hydroxyl groups in the crystal, with the peak at 990 cm−1 and the sharp drop at 3700 cm−1 being attributed to surface silanol groups which are present on the outer surface of the microporous layers. Figures 2 and 3a show the FTIR of the swollen MCM-22-P derivative, which contains cetyltrimethylammonium bromide covalently bonded to the surface silanols. As can be seen in Figure 3a, which gives more detail for the IR bands of interest, this sample shows a weak amine peak near ∼1340 cm−1 and has strong peaks at 1470 and 2849 cm−1, characteristic of N−CH3 bond stretching vibrations. There is also a sharp peak at ∼2920 cm−1 characteristic of C−H bonds. The surface silanol band at 990 cm−1 has also become less pronounced, while the entrapped water peak at 1650 cm−1 appears to become slightly broader. The broad hydroxyl peak is also present, although the silanol drop at 3700 cm−1 is not observed. By comparing the FTIR patterns of the swollen (Figures 2 and 3a) and pillared (Figures 2 and 3b) MCM derivatives, it can be seen that the pillaring procedure completely removes the covalently bonded amine groups from the surface silanols. Figures 2 and 3c represent a sample with high loading (∼14.5 wt %) of PEI on the solid surface and show the presence of weak amine peaks at 2833 and 2970 cm−1 and at 1470, 1370, and 1315 cm−1 as well as characteristic C−H

■

RESULTS AND DISCUSSION As can be seen in Figure 1, the X-ray diffraction patterns obtained for the bare samples are in good agreement with those found in the literature.5,35−38,41,42 The PEI loaded sample, shown as Figure 1e, is the 14.5 wt % sample, which had one of the highest loadings of PEI polymer and was chosen as characteristic of the PEI loaded samples. Of particular interest are the c-axis peaks, shown in Figure 1a as the (002), (101), and (102) peaks, and which appear at two-theta (2θ) angles lower than 10°.36,37,39 As expected, upon swelling these peaks either disappear or merge together. This can be seen in Figure 1b, where the (002) peak can be seen to disappear, with the (100) peak becoming broader and the (101) and (102) peaks becoming merged and broadened significantly. Upon calcination to form the MCM-36 crystal the (100) peak becomes much more well-defined as seen in Figure 1c, although the broad peak near 9° corresponding to the (101) and (102) peaks becoming merged together is still retained. The peaks at (220) and (310) however are not disturbed significantly in the swollen or pillared samples, suggesting the success of the pillaring procedure in keeping the zeolite layers from becoming damaged. Upon the introduction of PEI into the MCM-36 sample, represented as Figure 1e, it can be seen that the low angle peaks below 10°, which are characteristic of the interlayer pore space, become slightly broadened. It is believed that this is due to the filling of the interlayer pore space upon the introduction of PEI into the framework.1 There is no apparent change in the high angle peaks, suggesting that the introduction 4536

DOI: 10.1021/la505037f Langmuir 2015, 31, 4534−4541

Article

Langmuir

Figure 3. FTIR analysis of the regions from 1200 to 2000 cm−1 and 2700 to 4000 cm−1 for the (a) swollen MCM-22-P, (b) MCM-36, and (c) PEIloaded MCM-36 structures.

Table 1. Surface Characteristics and PEI Weight Percent Data Obtained via NOVA and TGA Analysis sample name MCM-22 MCM-36 sample A MCM-36 sample B PEI-MCM-2.7 PEI-MCM-3.7 PEI-MCM-3.8 PEI-MCM-7.5 PEI-MCM-14.5 PEI-MCM-16.3

solid support

BET surface area (m2/g)

BJH pore volume (cm3/g)

micropore volume (cm3/g)

BJH pore radius (Å)

PEI loading (wt %)

A A A A B B

463 765 506 14.1 21.7 28.8 18.6 5.00 4.18

0.144 0.199 0.178 0.025 0.062 0.050 0.132 0.098 0.052

0.21 0.39 0.26 0.017 0.020 0.030 0.034 0 0

18.2 18.0 20.4 18.1 18.1 16.1 20.3 18.1 16.1

0 0 0 2.70 3.72 3.85 7.52 14.5 16.3

peaks at 1500 cm−1 and a string of weak C−H peaks in the 2800−3000 cm−1 region. This sample also shows weak silanol peaks near 900 and 1600 cm−1. These peaks suggest the success of the amine polymer impregnation procedure. To determine the total loading of amine on the sample surface, as well as the surface characteristics of the samples, TGA and BET characterizations were performed. BET surface characteristics are reported in Table 1, with the nitrogen adsorption isotherms given in Figure 4. As can be seen from

Table 1, the BET surface area of the adsorbents decreases significantly at the instant of PEI impregnation along with a significant reduction in pore volume. For instance, the surface area of the support had a maximum value of 765 m2/g for MCM-36 and shows a sharp decrease to 14.1 m2/g upon polymer addition, before reaching the minimum of 5.00 m2/g for the highest loadings. Interestingly, it can be seen that the surface area of the samples appears to increase initially with increasing amine loading from the lowest weight percent samples, before dropping again to the minimum value at the highest loadings. It is hypothesized that this is due to two affects working in tandem. First, the addition of amines fills the interlayer mesopore space and immediately reduces the available surface area. Second, the impregnation of amines may increase slightly the pore space available for adsorption by the creation of small pores in the polymer ligands as they become interpenetrated with each other inside of this large pore space. As more PEI is added, these small pores are decreased to the point of no longer being accessible for entering adsorbate molecules, and eventually the surface area decreases to that of a nonporous material. BJH pore volume and DR micropore volume were seen to increase with increasing PEI loading before seeing a large decrease with higher amine amounts, further supporting the hypothesis that interpenetration of impregnated polymer molecules may be increasing the amount of available micropores with increasing

Figure 4. Nitrogen adsorption isotherms for the samples prepared in this study. 4537

DOI: 10.1021/la505037f Langmuir 2015, 31, 4534−4541

Article

Langmuir Table 2. Carbon Dioxide Capture Data Obtained via TGA Analysis sample name MCM-36 sample A MCM-36 sample B PEI-MCM-2.7 PEI-MCM-3.7 PEI-MCM-3.8 PEI-MCM-7.5 PEI-MCM-14.5 PEI-MCM-16.3

solid support MCM-36 MCM-36 MCM-36 MCM-36 MCM-36 MCM-36 MCM-36 MCM-36

sample A sample B sample A sample A sample A Sample A sample B sample B

PEI loading (wt %) adsorption half time (min) 0 0 2.70 3.72 3.85 7.52 14.5 16.3

58.5 33.5 309 296 233 720 559 559

polymer content until eventually amine impregnation completely blocks the pore space. As can be seen in Figure 4, the isotherms change significantly upon pillaring with silica and the introduction of amine molecules. The bare MCM-22 material displays a type I isotherm with a slight hysteresis, indicating a material with high micropore content. Upon pillaring, the isotherm transitions to type II, suggesting a material with mixed micro- and mesoporosity as expected. Upon the introduction of amines into the material, a large decrease in surface area is indicated by the dramatic shift of the isotherms to lower adsorbed volumes. The isotherms for the PEI loaded samples are generally in the shape of type II isotherms with very large hysteresis regions, with the isotherms shifting down to lower volumes adsorbed and the hysteresis becoming smaller with increasing polymer loading. Based on the functionalization method, it is expected that the PEI loaded samples should contain both a microporous region with no polymer species and an interlayer region which contains the PEI polymer molecules. This is due to the fact that the PEI molecule used (with MW of ∼800 g/mol) is far too large to fit within the micropore channels in the MCM layers and so can only fit in the interlayer space created upon pillaring. Therefore, after impregnation it is expected that these interlayer channels should not contain polymer molecules. In the isotherms shown, we believe that the degradation of the isotherms seen for the middle range of loadings is indicative of the filling of the PEI loaded interlayer space, with the slight porosity and hysteresis observed being due to the presence of the empty microporous region within the layers. As more polymer chains are introduced, we see that the isotherms eventually become those for a primarily nonporous material with slight hysteresis as seen for the 14.5 and 16.3 wt % PEI samples, suggesting some micropore access. This further supports the claim that higher amine loadings will completely block the large pore access in the material. While more characterization is required to fully make these claims, we believe they are a reasonable hypothesis to explain the porous character displayed by these isotherms. Table 2 shows the CO2 capture data for these samples obtained via TGA analysis. Plots of these data are also given as Figures 5−8, which give the obtained values normalized as a percentage of the bare derivative values obtained via analysis. For instance, the normalized capture capacity is given as the capture obtained for the PEI-loaded sample divided by that for the bare derivative used to generate the PEI-loaded sample. This is done to account for changes which may be present between the two samples which were created from different MCM-36 bare materials. This also gives a bare material normalized value standard of 1 for each of the graphs shown (capture capacity, adsorption half time, and initial capture rate)

initial adsorption rate (mmol/min) capture capacity (mmol/g) 0.06 0.05 0.002 0.007 0.009 0.01 0.003 0.003

1.03 1.07 0.323 0.494 0.466 0.616 0.218 0.266

Figure 5. Normalized carbon dioxide capture versus PEI weight percent loaded on each sample.

and allows for comparison directly to the bare material property. Figure 5 shows the normalized CO2 capture capacity versus the PEI weight percent loaded obtained via TGA analysis. It appears that the capture capacity increases slightly at low PEI loadings below 8 wt % and then decreases again significantly for the highest loadings achieved. It seems that this curve shape can be attributed to the changing porous environment within the material as PEI polymer is loaded in higher weight percentages. The bare material (MCM-36) comprises two different pore systems, i.e., interlayer mesopores generated via pillaring the layers to have two-dimensional pore connectivity as well as intralayer micropore channels passing through the individual layers along the c-axis direction of the MCM-36 crystal. Carbon dioxide can access internal mesopore spaces via the directions along the a- and b-axis, but not through the layers along the direction of the c-axis due to the size of intralayer micropores that are smaller than the kinetic diameter of carbon dioxide molecules. Upon the introduction of amine polymer into the interlayer mesopore spaces, and the corresponding decrease in the surface area for the material, access to the interlayer space is greatly hindered, and the capture capacity decreases. As amine polymer is further impregnated, the amount of chemisorption sites increases, leading to an increase in the capture capacity observed. However, at a critical loading of amine the capture capacity decreases significantly again. It is believed this is due to the complete blockage of the entrance of mesopore spaces with the high polymer loading, preventing carbon dioxide from diffusing to the available amine sites.33 This claim is supported by the nitrogen adsorption isotherms and surface properties discussed above, which showed a similar trend of increasing surface area and pore volume and then greatly decreased surface properties upon the introduction of the highest polymer loadings. 4538

DOI: 10.1021/la505037f Langmuir 2015, 31, 4534−4541

Article

Langmuir The adsorption equilibrium half time results, shown in Figure 6, show a similar trend. Half the time to reach equilibrium,

Figure 7. Normalized initial capture rate versus PEI weight percent loaded. Figure 6. Normalized adsorption half time versus PEI weight percent loaded on each sample.

defined here as a weight percent versus time slope of less than 0.0001 mg/min in 100 min, was found to be 58.5 and 33.5 min for the bare MCM-36 samples. This represents adsorption of CO2 onto the material with no pore blockage or diffusion limitations due to the impregnation of PEI polymer. Upon amine impregnation the adsorption half time increase significantly, up to approximately 720 min for the 7.52 wt % amine loaded sample. This sample also had the highest CO2 capacity for all amine loaded samples, suggesting that carbon dioxide can diffuse through the large pore space, although this diffusion is very slow. The adsorption half time is then seen to decrease down to approximately 559 min for the 14.5 and 16.3 wt % amine samples. However, these samples showed the highest normalized adsorption half time, since they represent a much greater increase in the half time over the bare derivative which was slightly lower than that for the other samples. This further supports the idea that the capture capacity is decreased for these samples due to the inability of the carbon dioxide to adequately diffuse into the pore space. The amount of carbon dioxide captured in this time is also significantly less than for the other samples, suggesting the inability for gas molecules to diffuse into the total volume of the material. The idea that chemisorption can no longer occur due to the blocking of amine moieties along the impregnated groups when the PEI weight percent becomes too high is further supported by the initial capture rate data, shown in Figure 7. This shows the highest initial capture rate for the 8.4 wt % sample, which also showed the highest capture capacity. The initial increase in capture rate was expected, since with increasing polymer content there is also an increase in the amount of chemisorption sites within the material. However, the samples with the highest amine loadings show initial capture rates closer to those for the sample with the lower amine loadings, suggesting that the amine sites are no longer accessible to entering gas molecules as suggested. Figure 8 shows even more clearly the effect that higher polymer loading has on the ability of carbon dioxide to become adsorbed onto the chemisorption sites introduced to the material. If it is assumed that each carbon dioxide mole binds to 2 mol of amine in a nonhumid environment and 1 mol per site in the presence of water vapor as has been suggested in the literature,11 the maximum amine efficiency for an impregnated adsorbent should be 0.5 mol CO2/mol of amine in dry

Figure 8. Carbon capture efficiency (given as moles of carbon dioxide captured per mole of PEI loaded onto the sample) versus PEI weight percent loaded.

conditions and as high as 1 mol CO2/mol amine in humid conditions. However, for the lowest amine loading the maximum efficiency was calculated to be 1.08 mol CO2/mol amine, in what is assumed to be a nonhumid environment. This high efficiency shown for the lower weight percent samples is believed to be due to capture occurring at both the amine sites through chemisorption as well as throughout the remaining volume of the material through physisorption. With increasing amine content, this efficiency drops below the maximum value of 0.5 to a minimum of 0.12 mol CO2/mol amine. This is indicative of there being a large population of amine sites within the support which do not capture any carbon dioxide, supporting the hypothesis that increasing polymer loading limits diffusion to the amine sites within the large pore space. These results show clearly that increasing the amine polymer content in the MCM-36 material cannot be used to increase the capture capacity of the support, as the ability of each mole of amine to be useful toward the capture of carbon dioxide decreases with increasing polymer loading. This suggests that higher loading within other 2-dimensional porous solids may also be unsuccessful in increasing the capture performance, as higher loading leads to greatly decreased carbon dioxide diffusion. Therefore, it appears the optimum amine loading value is found in the middle loading regime, before pore blocking occurs and where the highest kinetic improvements can be made, which in this case is near 8.4 wt %. 4539

DOI: 10.1021/la505037f Langmuir 2015, 31, 4534−4541

Article

Langmuir

■

CONCLUSIONS Solid amine CO2 adsorbents have been spotlighted in past decades owing to their advantages over other CO2 capture technologies such as the liquid amine absorption process, including recyclability, tunable adsorption properties, moderate heat of adsorption, etc. The major direction of supported amine adsorbents research has been towards making their adsorption capacity as high as possible by maximizing the amount of amines loaded in the support. The results of this study show that, on the other hand, excessive addition of amine moieties such as PEI into the supports such as MCM-36 that comprise two-dimensionally interconnected mesoporous networks can cause the degradation of the CO2 capture performance, i.e., the adsorption capacity and the kinetics, of the supported amine adsorbents. We believe this is due to partial or complete blockage of the mesoporous networks upon aminopolymer impregnation, as observed in the supported amine adsorbents made of the supports with one-dimensional pore channels. Our previous reports revealed that the supports containing onedimensional pore system, e.g., SBA-15, cannot avoid degradation of CO2 capture capacity and kinetics, especially at high amine loading. It was attributed to partial or complete blockage of the pore channels by impregnated amines that causes hindered diffusion of CO2 toward the internal adsorption sites. To prevent this performance degradation upon amine integration, we attempted in this work to change the pore connectivity of the support from one-dimensional pore channels to two-dimensional pore networks, using pillared silica such as MCM-36. However, despite the presence of the pore systems interconnected to each other, the CO2 capture capacity and kinetics of the supported amines with twodimensional pore networks were degraded upon high amine loading, similarly to what happened to the solid amines with one-dimensional pore channels. We believe this is because current materials design of the supported amine adsorbents utilizes the pore systems of the supports for both the space for amine accommodation and the path for CO2 diffusion. In the present work, since the pore dimension of the intralayer micropore channels along the c-axis of the MWW layers in the MCM-36 material is larger than the kinetic diameter of carbon dioxide, access of the CO2 molecules to the adsorption sites located within interlayer spaces (impregnated amines) is only possible by diffusion through interlayer mesopores along the direction of the a- and b-axis of the MCM-36. Integration of amines within these mesopore spaces is followed not only by an increasing number of adsorption sites but also by the reduction in available pore spaces, which eventually leads to partial or complete blockage of the pores. Since these are the very same channels being filled upon aminopolymer impregnation, there is a critical loading value of amines at which the diffusion of carbon dioxide throughout the pores is hindered or obstructed, greatly decreasing the capture efficiency of the loaded polymers because carbon dioxide cannot access the internal chemisorption amine sites. This optimum loading is notably to be found not with the highest amine loadings onto the support, but in the middle region where the addition of amine has not yet caused diffusion to become the dominant factor determining capture results. The knowledge of the structure−property relationship of the supported amine adsorbents gained through the present work suggests that rational materials design of the support materials, e.g., systematic consideration of the pore characteristics such as

pore dimension and connectivity of the supports, is required to develop high performance supported amine CO2 adsorbents. In the sorts of porous solids investigated so far, the increase in amine content has been shown to lead necessarily to decreased diffusion of entering gas molecules as the pores become filled. Even more troubling, as amine is added the efficiency of each amine to capture carbon dioxide eventually becomes degraded, leading to a maximum capture capacity in general which does not successfully utilize each loaded amine site. In order to overcome these challenges, it appears that a new sort of solid support which allows for the decoupling of the loading from diffusion of entering molecules needs to be created.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail [email protected] (S.C.). Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors acknowledge the contributions to this work of Rebecca Chinn, Jillian Zummo, and Yang Lin, who worked as undergraduate research assistants on these experiments. We also acknowledge Dr. Sanjeev Mukerjee for the use of his facilities in the analysis of XRD and FTIR data.

■

REFERENCES

(1) Choi, S.; Drese, J. H.; Eisenberger, P. M.; Jones, C. W. Application of Amine-Tethered Solid Sorbents for Direct CO2 Capture from the Ambient Air. Environ. Sci. Technol. 2011, 45, 2420−2427. (2) Yang, S.-T.; Kim, J.-Y.; Kim, J.; Ahn, W.-S. CO2 Capture over Amine-Functionalized MCM-22, MCM-36 and ITQ-2. Fuel 2012, 97, 435−442. (3) Jeong, H.-K.; Nair, S.; Vogt, T.; Dickinson, L. C.; Tsapatsis, M. A Highly Crystalline Layered Silicate with Three-Dimensionally Microporous Layers. Nat. Mater. 2003, 2, 53−58. (4) Na, K.; Jo, C.; Kim, J.; Cho, K.; Jung, J.; Seo, Y.; Messinger, R. J.; Chmelka, B. F.; Ryoo, R. Directing Zeolite Structures into Hierarchically Nanoporous Architectures. Science 2011, 333, 328−332. (5) Kim, W.; Choi, S.; Nair, S. Swelling, Functionalization, and Structural Changes of the Nanoporous Layered Silicates AMH-3 and MCM-22. Langmuir 2011, 27, 7892−7901. (6) Choi, S.; Coronas, J.; Sheffel, J. a.; Jordan, E.; Oh, W.; Nair, S.; Shantz, D. F.; Tsapatsis, M. Layered Silicate by Proton Exchange and Swelling of AMH-3. Microporous Mesoporous Mater. 2008, 115, 75−84. (7) Choi, S.; Coronas, J.; Jordan, E.; Oh, W.; Nair, S.; Onorato, F.; Shantz, D. F.; Tsapatsis, M. Layered Silicates by Swelling of AMH-3 and Nanocomposite Membranes. Angew. Chem., Int. Ed. 2008, 47, 552−555. (8) Sumida, K.; Rogow, D. L.; Mason, J. a; McDonald, T. M.; Bloch, E. D.; Herm, Z. R.; Bae, T.-H.; Long, J. R. Carbon Dioxide Capture in Metal-Organic Frameworks. Chem. Rev. 2012, 112, 724−781. (9) Bae, Y.-S.; Snurr, R. Q. Development and Evaluation of Porous Materials for Carbon Dioxide Separation and Capture. Angew. Chem., Int. Ed. 2011, 50, 11586−11596. (10) Schell, J.; Casas, N.; Blom, R.; Spjelkavik, A. I.; Andersen, A.; Cavka, J. H.; Mazzotti, M. MCM-41, MOF and UiO-67/MCM-41 Adsorbents for Pre-Combustion CO2 Capture by PSA: Adsorption Equilibria. Adsorption 2012, 18, 213−227. (11) Choi, S.; Drese, J. H.; Jones, C. W. Adsorbent Materials for Carbon Dioxide Capture from Large Anthropogenic Point Sources. ChemSusChem 2009, 2, 796−854. (12) Serna-Guerrero, R.; Belmabkhout, Y.; Sayari, A. Influence of Regeneration Conditions on the Cyclic Performance of Amine-Grafted 4540

DOI: 10.1021/la505037f Langmuir 2015, 31, 4534−4541

Article

Langmuir

Adsorbent Based on Polyethylenimine-Modified Molecular Sieve MCM-41. Ind. Eng. Chem. Res. 2005, 44, 8113−8119. (33) Drese, J. H.; Choi, S.; Lively, R. P.; Koros, W. J.; Fauth, D. J.; Gray, M. L.; Jones, C. W. Synthesis-Structure-Property Relationships for Hyperbranched Aminosilica CO2 Adsorbents. Adv. Funct. Mater. 2009, 19, 3821−3832. (34) Maheshwari, S.; Martínez, C.; Teresa Portilla, M.; Llopis, F. J.; Corma, A.; Tsapatsis, M. Influence of Layer Structure Preservation on the Catalytic Properties of the Pillared Zeolite MCM-36. J. Catal. 2010, 272, 298−308. (35) Chlubná, P.; Roth, W. J.; Zukal, A.; Kubů, M.; Pavlatová, J. Pillared MWW Zeolites MCM-36 Prepared by Swelling MCM-22P in Concentrated Surfactant Solutions. Catal. Today 2012, 179, 35−42. (36) Corma, A.; Diaz, U.; Fornes, V.; Guil, J. M.; Martinez-Triguero, J.; Creyghton, E. J. Characterization and Catalytic Activity of MCM-22 and MCM-56 Compared with ITQ-2. J. Catal. 2000, 191, 218−224. (37) Corma, A.; Forn, V.; Mart, J.; Pergher, S. B. Delaminated Zeolites: Combining the Benefits of Zeolites and Mesoporous Materials for Catalytic Uses. J. Catal. 1999, 63, 57−63. (38) He, Y. J.; Nivarthy, G. S.; Eder, F.; Seshan, K.; Lercher, J. A. Synthesis, Characterization, and Catalytic Activity of the Pillared Molecular Sieve MCM-36. Microporous Mesoporous Mater. 1998, 25, 207−223. (39) Lawton, S. L.; Fung, A. S.; Kennedy, G. J.; Alemany, L. B.; Chang, C. D.; Hatzikos, G. H.; Lissy, D. N.; Rubin, M. K.; Timken, H. C.; Steuernagel, S.; Woessner, D. E. Zeolite MCM-49: A ThreeDimensional MCM-22 Analogue Synthesized by in Situ Crystallization. J. Phys. Chem. 1996, 100, 3788−3798. (40) Maheshwari, S.; Jordan, E.; Kumar, S.; Bates, F. S.; Penn, R. L.; Shantz, D. F.; Tsapatsis, M. Layered Structure Preservation during Swelling, Pillaring, and Exfoliation of a Zeolite Precursor. J. Am. Chem. Soc. 2008, 130, 1507−1516. (41) Kornatowski, J.; Barth, J.-O.; Erdmann, K.; Rozwadowski, M. Effect of Various Pillaring Oxides on Adsorption Behaviour of Novel MCM-36 Derivatives. Microporous Mesoporous Mater. 2006, 90, 251− 258. (42) Santos Marques, A. L.; Fontes Monteiro, J. L.; Pastore, H. O. Static Crystallization of Zeolites MCM-22 and MCM-49. Microporous Mesoporous Mater. 1999, 32, 131−145.

Mesoporous Silica for CO2 Capture: An Experimental and Statistical Study. Chem. Eng. Sci. 2010, 65, 4166−4172. (13) Heydari-Gorji, A.; Sayari, A. CO2 Capture on PolyethylenimineImpregnated Hydrophobic Mesoporous Silica: Experimental and Kinetic Modeling. Chem. Eng. J. 2011, 173, 72−79. (14) Kumar, P.; Kim, S.; Ida, J.; Guliants, V. V. PolyethyleneimineModified MCM-48 Membranes: Effect of Water Vapor and Feed Concentration on N2/ CO2 Selectivity. Ind. Eng. Chem. Res. 2008, 47, 201−208. (15) Klinthong, W.; Chao, K.; Tan, C. CO2 Capture by AsSynthesized Amine-Functionalized MCM-41 Prepared through Direct Synthesis under Basic Condition. Ind. Eng. Chem. Res. 2013, 52, 9834− 9842. (16) Gil, M.; Tiscornia, I.; de la Iglesia, Ó .; Mallada, R.; Santamaría, J. Monoamine-Grafted MCM-48: An Efficient Material for CO 2 Removal at Low Partial Pressures. Chem. Eng. J. 2011, 175, 291−297. (17) Gao, F.; Zhou, J.; Bian, Z.; Jing, C.; Hu, J.; Liu, H. Dynamic Properties in CO2 Adsorption on Amine-Modified MCM-41. Proc. Inst. Mech. Eng., Part E 2013, 227, 106−116. (18) Alkhabbaz, M. a.; Khunsupat, R.; Jones, C. W. Guanidinylated Poly(allylamine) Supported on Mesoporous Silica for CO2 Capture from Flue Gas. Fuel 2014, 121, 79−85. (19) Belmabkhout, Y.; Serna-Guerrero, R.; Sayari, A. Amine-Bearing Mesoporous Silica for CO2 Removal from Dry and Humid Air. Chem. Eng. Sci. 2010, 65, 3695−3698. (20) Ciftja, A. F.; Hartono, A.; Svendsen, H. F. Experimental Study on Carbamate Formation in the AMP− CO2−H2O System at Different Temperatures. Chem. Eng. Sci. 2014, 107, 317−327. (21) Roth, E. A.; Agarwal, S.; Gupta, R. K. Nanoclay-Based Solid Sorbents for CO2 Capture. Energy Fuels 2013, 27, 4129−4136. (22) Liu, F.-Q.; Wang, L.; Huang, Z.-G.; Li, C.-Q.; Li, W.; Li, R.-X.; Li, W.-H. Amine-Tethered Adsorbents Based on Three-Dimensional Macroporous Silica for CO2 Capture from Simulated Flue Gas and Air. ACS Appl. Mater. Interfaces 2014, 6, 4371−4381. (23) Aruldoss, D.; Saigoanker, R.; Savarimuthu, J.; Das, R. J. AmineGrafted Zeolites-Mesoporous Ceramics: Synthesis and Adsorption Characteristics. Ceram. Int. 2014, 40, 7583−7587. (24) Yue, M. B.; Chun, Y.; Cao, Y.; Dong, X.; Zhu, J. H. CO2 Capture by As-Prepared SBA-15 with an Occluded Organic Template. Adv. Funct. Mater. 2006, 16, 1717−1722. (25) Hiyoshi, N.; Yogo, K.; Yashima, T. Adsorption Characteristics of Carbon Dioxide on Organically Functionalized SBA-15. Microporous Mesoporous Mater. 2005, 84, 357−365. (26) Drese, J. H.; Choi, S.; Didas, S. a.; Bollini, P.; Gray, M. L.; Jones, C. W. Effect of Support Structure on CO2 Adsorption Properties of Pore-Expanded Hyperbranched Aminosilicas. Microporous Mesoporous Mater. 2012, 151, 231−240. (27) Li, W.; Bollini, P.; Didas, S. a; Choi, S.; Drese, J. H.; Jones, C. W. Structural Changes of Silica Mesocellular Foam Supported AmineFunctionalized CO2 Adsorbents upon Exposure to Steam. ACS Appl. Mater. Interfaces 2010, 2, 3363−3372. (28) Son, W.-J.; Choi, J.-S.; Ahn, W.-S. Adsorptive Removal of Carbon Dioxide Using Polyethyleneimine-Loaded Mesoporous Silica Materials. Microporous Mesoporous Mater. 2008, 113, 31−40. (29) Xu, X.; Song, C.; Andrésen, 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. (30) Xu, X.; Song, C.; Miller, B. G.; Scaroni, A. W. Adsorption Separation of Carbon Dioxide from Flue Gas of Natural Gas-Fired Boiler by a Novel Nanoporous “Molecular Basket” Adsorbent. Fuel Process. Technol. 2005, 86, 1457−1472. (31) Xu, X.; Song, C.; Andresen, J. M.; Miller, B. G.; Scaroni, A. W. Novel Polyethylenimine-Modified Mesoporous Molecular Sieve of MCM-41 Type as High-Capacity Adsorbent for CO2 Capture. Energy Fuels 2002, 16, 1463−1469. (32) Xu, X.; Song, C.; Miller, B. G.; Scaroni, A. W. Influence of Moisture on CO2 Separation from Gas Mixture by a Nanoporous 4541

DOI: 10.1021/la505037f Langmuir 2015, 31, 4534−4541