Article pubs.acs.org/EF
Oxidative Stability of Amino Polymer−Alumina Hybrid Adsorbents for Carbon Dioxide Capture Sumit Bali, Thomas T. Chen, Watcharop Chaikittisilp,† and Christopher W. Jones* School of Chemical & Biomolecular Engineering, Georgia Institute of Technology, 311 Ferst Drive, Atlanta, Georgia 30332-0100, United States ABSTRACT: Amine/oxide hybrid carbon dioxide adsorbents prepared via impregnation of low molecular weight polymeric amines into porous oxide supports are among the most promising solid adsorbents developed for postcombustion CO2 capture or CO2 extraction from ambient air. The oxidative stability of adsorbents prepared by impregnation of poly(ethylenimine) (PEI) or poly(allylamine) (PAA) into mesoporous γ-alumina under humid oxidation conditions is evaluated in this work. The PEIbased adsorbents, which contain primary, secondary, and tertiary amines, are shown to degrade drastically at elevated temperatures (110 °C) and in high oxygen concentrations (21%, akin to air), with these effects reduced by both reductions in temperature (70 °C) and oxygen concentration (5%, akin to flue gas). The oxidation behavior of PEI-based adsorbents supported on alumina is qualitatively similar to past work on silica-supported PEI adsorbents. In contrast, the alumina-supported PAA adsorbents that contain only primary amines show significantly improved oxidative stability, losing only 10% or less of their original CO2 capacity after prolonged oxidative treatment under a variety of conditions. Analysis of the fresh and thermally treated samples by Fourier transform (FT) IR, FT-Raman, and 13C NMR spectroscopies demonstrates the clear formation of carbonyl functionalities over the oxidized PEI-based adsorbents, whereas no significant changes in the spectra for PAA samples are observed after oxidative treatments. The collected data demonstrate that secondary-amine-free, primary-amine-rich polymers such as PAA may be used to formulate supported amine adsorbents with improved oxidative stability compared to adsorbents based on PEI, which is used ubiquitously in the field today. organic frameworks,21,22 zeolites,23,24 and supported amine materials.25−28 Due to their ease of synthesis and high adsorption capacities, supported amine materials have attracted much attention for capturing carbon dioxide from both flue gas and ambient air. In these materials, amine groups can be introduced on supports by a variety of methodologies, such as wet impregnation of amines onto the supports (class 1 materials),29−31 grafting aminecontaining molecules onto the surface of supports via covalent bonds (class 2 materials),32−34 and in situ polymerization of amine-containing monomers on the surface of supports (class 3 materials)12,25,26 Class 1 adsorbents offer the advantage of being simple to synthesize and they often exhibit the highest loadings of amine groups, thereby offering high adsorption capacities and excellent CO2 selectivity. Consequently, impregnation of amine-based organic polymers such as poly(ethylenimine) (PEI)29,30,35,36 poly(allylamine) (PAA)37 and oligomeric species such as tetrethylenepentamine (TEPA)38−41 onto a variety of supports has been extensively evaluated. Whereas the adsorption capacities of such materials are routinely investigated, only recently have researchers begun to focus on the stability of such materials under practical operating conditions. The adsorbent materials may degrade via a range of pathways25 including evaporation of volatile amines, such as molecular or oligomeric amines, oxidation associated with the
1. INTRODUCTION The increase in the levels of CO2 in the earth’s atmosphere is being touted as the primary reason for climate change and is resulting in an increase in the average global temperature.1 Because of the impact of climate change on the human condition,2−4 considerable efforts are being dedicated to the mitigation of these climate change effects by capturing carbon dioxide produced from combustion of fossil fuels at large point sources in an effort to significantly curb the release of carbon dioxide into the atmosphere.5,6 Many novel techniques are being developed to separate and capture carbon dioxide from the flue gases generated during fossil fuel combustion, which is the foremost contributor to CO2 emissions.7,8 Once separated and captured from the flue gas, CO2 can be permanently or semipermanently sequestered underground.9,10 More recently efforts are also being directed toward the capture and storage of carbon dioxide directly from the already existing pool of CO2 in the ambient atmosphere.11,12 Such a process for CO2 capture is commonly referred to as air capture. Processes for capture of CO2 from flue gases as well as from air primarily rely on the use of chemical systems that have moieties that can selectively bind to CO2 and remove it from dilute streams. The benchmark technology for large-scale carbon dioxide capture is absorption by use of aqueous amines.13,14 Increasingly, researchers are investigating a variety of solid adsorbent materials as components of hypothetical adsorption processes for CO2 capture and concentration as well.5 For example, CO2 can be chemically absorbed by capturing species such as alkali and alkaline earth oxides and hydroxides, resulting in the formation carbonates.5,15−17 Alternatively, a variety of porous solids have been used, including carbons,18−20 metal− © 2013 American Chemical Society
Received: January 19, 2013 Revised: February 13, 2013 Published: February 15, 2013 1547
dx.doi.org/10.1021/ef4001067 | Energy Fuels 2013, 27, 1547−1554
Energy & Fuels
Article
molecular oxygen in flue gas (ca. 5%) or air (ca. 21%),42−44 degradation from extensive or high-temperature cycling in CO2,45,46 or degradation due to exposure to SOx and NOx.47,48 The existing studies on the stability of class 1 adsorbents have focused exclusively on PEI-impregnated silica materials, which now stand as the most well-studied amine-based CO 2 adsorbents. Drage et al.45 reported in 2008 that a dramatic loss in CO2 adsorption capacity was observed when temperature swing desorption was carried out with concentrated CO2 as the regeneration purge gas at temperatures of 110−140 °C. This was attributed to chemical reaction of CO2 with amines on the PEI polymer, resulting in the formation of covalent linkages in the form of urea at elevated temperatures. Urea formation was irreversible under the conditions applied in that study and resulted in the degradation of the CO2 uptake in successive cycles. The PEI adsorbents were found to be stable only under nitrogen at elevated temperatures, conditions commonly used in academia to regenerate adsorbents but which are wholly impractical for production of concentrated CO2 product, as required for sequestration.25,38 In 2011, Bollini et al.42 and others43,44 demonstrated that supported amine materials, including materials with ethylenimine structures, as found in PEI, degraded upon exposure to oxygen at elevated concentration and temperatures. Bollini’s work showed that while primary and tertiary amines were quite stable in class 2 materials, secondary amines degraded quite easily upon exposure to concentrated oxygen at elevated temperatures (ca. 100 °C). Additionally, the degradation of secondary amines in ethylenimine structures simultaneously degraded the primary amines, leading to depressed CO2 adsorption characteristics due to decreased basicity of the formed species (imines, amindes, imides, etc.). Later work by Heydari-Gorji and Sayari46 evaluated the exposure of PEI-based adsorbents to air, simulated flue gas conditions, and different CO2/O2/N2 mixtures. It was observed that although the PEI-based adsorbents were thermally stable under mild temperatures, long-term exposure to dry CO2 as well as use of dry CO2 in multiple adsorption desorption cycles resulted in deactivation of PEI and loss of carbon dioxide uptake, which was attributed to the irreversible formation of urea under such conditions. Under humid oxidation conditions, however, urea formation was inhibited and the adsorbent exhibited high stability in CO2 capture adsorption−desorption cycles. Furthermore, it was observed that the presence of CO2 along with oxygen in the oxidation medium resulted in enhanced stability of the material toward oxygen. This was attributed to the competitive adsorption or faster reaction of amines with CO2 than with oxygen, resulting in the formation of carbamate and bicarbonate species, which exhibit much better oxygen stability compared to amines. Thus an important message from the work of Heydari-Gorji and Sayari46 is that class 1 materials can be somewhat stabilized upon exposure to oxygen in the presence of water and carbon dioxide and that materials can be stabilized in the presence of high concentrations of carbon dioxide by exposure to humidity. While the studies of adsorbent stability to oxygen and other species discussed above have begun to establish an operating window for class 1, PEI-based adsorbents, it should be noted that stability studies of other polymers have not been reported and the role of support composition, which can impact CO2 adsorption properties,49,50 has also not been evaluated. Previously, on the basis of the observation that primary amines in 3-aminopropyl-functionalized silica (class 2) adsorbents were
stable under oxidizing conditions, we investigated class 1 adsorbents based on PAA impregnated in porous silica.37 In that work, PAA was demonstrated as a novel primary-aminerich polymer that, when combined with an appropriate porous support, gave adsorbents with CO2 adsorption properties similar to sorbents with similar loadings of PEI. However, the oxidative stability of PAA has not yet been explored, and extrapolations about its oxidative stability from class 2 sorbents based on aminopropyl groups grafted to the silica surface may not be warranted. To this end, in this work, we explore the stability of both PEI and PAA supported on mesoporous alumina under varying oxygen concentrations and at two temperatures. More specifically, a comprehensive oxidation study was carried out at elevated temperature under 5% oxygen (flue gas conditions) and 21% oxygen (air capture conditions) by volume.
2. EXPERIMENTAL SECTION 2.1. Materials. The following chemicals were used as received from the supplier: pseudoboehmite (Catapal B, 74.3% Al2O3, Sasol North America), allylamine hydrochloride (AAHCl, TCI), anhydrous 2-propanol (99.5%, Alfa Aesar), methanol (200 proof, J. T. Baker), ethanol (99.5%, Acros), 2,2-azobis(isobutyric acid) dimethyl ester (MAIB, 98%, AK Scientific), strongly basic ion-exchange resin [Ambersep-900-(OH), Fluka], Pluronic P-123 EO20−PO70−EO20 triblock copolymer (Sigma−Aldrich), 1,3,5-trimethylbenzene (TMB, 97%, Sigma−Aldrich), tetraethyl orthosilicate (TEOS, 98%, Sigma− Aldrich), ammonium fluoride (>96%, Alfa Aesar), hydrochloric acid (37%, J.T. Baker), nitric acid (68%, J. T Baker), and branched poly(ethylenimine) (Mw = 800, Sigma−Aldrich) 2.2. Synthesis of Mesoporous γ-Alumina. γ-Alumina was synthesized according to an earlier reported procedure51 by surfactant P-123-mediated self-assembly of pseudoboehmite nanoparticles. In a typical procedure, 13.75 g of commercial pseudoboehmite from Sasol North America (Catapal B, 74.3% Al2O3) was peptized in a mixture of 1.27 g of nitric acid (Fischer Scientific, ∼70%) and 200 mL of deionized water. The suspension obtained was further sonicated for 90 min at room temperature. The sonicated suspension was then stirred at 60 °C for 17 h, after which it was cooled to room temperature. The peptized alumina thus obtained was slowly added to a solution of 15.30 g of Pluronic P123 in 200 mL of ethanol (200 proof). The resulting solution was further stirred at room temperature for 24 h. Subsequently the solvent was evaporated completely at 60 °C. The resulting P-123−alumina composite was further dried at 75 °C for 24 h. The white sol−gel-derived mesoporous γ-alumina was obtained by calcination of this composite at 700 °C for 4 h with a heating ramp of 1 °C/min and an intermediate step of 150 °C for 1 h to remove the water and ethanol. 2.3. Synthesis of Poly(allylamine). Allylamine hydrochloride (6.0 g, 0.06 mol), was mixed with 2-propanol (3.99 g) and MAIB (0.79 g, 3.43 mmol). The resulting solution was deaerated by purging with helium for 1 h. The solution was then heated at 60 °C for 48 h to carry out the free radical polymerization. The resulting polymer hydrochloride (PAA-HCl) was washed with excess methanol to remove unreacted monomer, filtered, and dried under vacuum at room temperature for 24 h to give 4.50 g of white powder (yield ∼70%). The polymer PAA in the free amine form was then obtained by using a strongly basic ion-exchange Ambersep-900-(OH) resin to remove the hydrogen chloride salt. To a degassed solution of PAA-HCl in deionized water (30.0 mL) was added strongly basic ion-exchange resin (Ambersep-900-OH, 16.0 g), and the resulting suspension was stirred for 1 h. The resulting polymer solution (pH 12) was filtered, the solvent was removed, and the polymer was dried under vacuum for 24 h to give 4.0 g of PAA product (yield ∼60%). 1H NMR (D2O, ppm): 1.05 (2H, -CH2-), 1.41 (1H, -CH-), 2.45 (2H, -CH2-). 2.4. Preparation of Poly(allyamine) and Branched Poly(ethylenimine) Sorbents. Poly(allylamine) (PAA) and branched poly(ethylenimine) (PEI) impregnated mesoporous alumina sorbents 1548
dx.doi.org/10.1021/ef4001067 | Energy Fuels 2013, 27, 1547−1554
Energy & Fuels
Article
were prepared by wet impregnation via a method described previously.51,52 In a typical synthesis, 1.0 g of alumina support was dispersed in 15.0 mL of methanol (Sigma−Aldrich, ACS reagent). To the resulting suspension was added dropwise a solution of the calculated amount of polymer (branched PEI/PAA) in 20 mL of methanol. The resulting solution was stirred at room temperature for another 24 h, after which the methanol was evaporated on a rotary evaporator. The obtained PEI/PAA impregnated sorbents were further dried in a high vacuum line (∼20 mTorr) at room temperature. 2.5. Oxidation of PAA and PEI Impregnated Mesoporous Alumina Sorbents. Evaluation of the oxidative stability of synthesized sorbent materials was carried out in a fixed-bed reactor. In a typical oxidation experiment, the PAA and PEI impregnated alumina sorbents (400 mg) were packed into a Pyrex tube, 1 cm in diameter, with a frit at the center to allow the flow of gas through the sample without loss of the adsorbent from the reactor. To remove residual water from the system, the sorbent was treated at 110 °C under flowing nitrogen at 15 mL/min for 2 h prior to switching to oxidation gas stream. The temperature was then set to the desired oxidation temperature (110 °C, 70 °C), and the flow was switched to the desired O2 concentration (21% or 5% by volume in N2) for the predetermined time of oxidation of 20 h. The oxidation gas was maintained at 15 mL/min through the reactor during the course of oxidation, after which the reactor was cooled and the adsorbent samples were recovered for further characterization and CO2 uptake experiments by thermnogravimetric analysis (TGA). The oxidized alumina impregnated PAA/PEI samples have been designated as ALPAA/PEI_PercentOxygen_Temperature of oxidation. 2.6. Sorbent Characterization. The organic loading relative to alumina support was determined by thermogravimetric analysis performed on the adsorbents by use of a Netzsch STA409PG thermogravimetric analyzer (TGA). The organic groups (amines) on the inorganic support alumina were combusted while the change in total mass was measured. The ramp rate was 10 °C/min under a mixed gas stream comprising air flowing at 90 mL/min and nitrogen flowing at 30 mL/min. Nitrogen physisorption measurements were carried out on a Micromeritics Tristar II 3020 instrument. Before the measurement, the samples were degassed under vacuum at 110 °C for at least 15 h. Surface areas, pore diameters, and pore volumes were calculated from the collected isotherm data. Surface areas were calculated by the Brunauer−Emmett−Teller (BET) method,53 and pore diameters and pore volumes were calculated by the Broekhoff−de Boer/Frenkel− Halsey−Hill (BdB-FHH) method.54 Powder X-ray diffraction (XRD) patterns were collected on a PANalytical X’pert diffractometer with a Cu−K-alpha X-ray source. The 13C cross-polarization magic-angle spinning (CP-MAS) solid-state nuclear magnetic resonance (NMR) measurements were carried out on a Bruker DSX-300 spectrometer. The samples were spun at a frequency of 10 kHz, and 16 000−18 000 scans were taken for each sample. 2.7. CO2 Adsorption on Fresh and Oxidized Adsorbents. A TA Instruments Q500 TGA was used to measure the adsorption capacities of the materials under dry CO2 capture conditions. The adsorbent materials were loaded into the platinum sample pan and helium was flowed through the sample chamber, while its temperature was ramped to 110 °C. The temperature was held constant at 110 °C for 3 h to remove residual water, and CO2 was potentially adsorbed from the atmosphere. The sample chamber was then cooled to 50 °C. After stabilization at 50 °C for 1 h, the gas flow was switched to 10% CO2 in helium, and the subsequent weight gain because of adsorption of CO2 was measured. The adsorption was done for 6 h to approach equilibrium capacities for all the tested adsorbents.
Figure 1. Wide-angle XRD pattern of synthesized mesoporous γalumina.
distributions calculated by the Broekhoff−de Boer method with Frenkel−Halsey−Hill modification (BdB-FHH) method are shown in Figure 2b. The synthesized γ-alumina exhibited a typical type IV isotherm that showed a type H1 hysteresis loop, having sharp uptakes at P/P0 = 0.8−0.9, which indicated the synthesized γalumina contains well-defined, large mesopores.
3. RESULTS AND DISCUSSION The wide-angle XRD pattern of the synthesized alumina shown in Figure 1 confirms the γ-alumina structure.51,55,56 The mesoporous structure of the synthesized alumina was confirmed by the nitrogen physisorption measurements. Nitrogen adsorption−desorption isotherms for the alumina are shown in Figure 2a, and the corresponding pore size
Figure 2. (a) Nitrogen adsorption−desorption isotherms of synthesized mesoporous γ-alumina. (b) Pore size distribution of synthesized mesoporous γ-alumina, calculated from the adsorption branch by the BdB-FHH method. 1549
dx.doi.org/10.1021/ef4001067 | Energy Fuels 2013, 27, 1547−1554
Energy & Fuels
Article
Table 1. Oxidation of PEI and PAA Impregnated Alumina Adsorbents under Humid Conditions and CO2 Adsorption Behavior oxidation parameters amine polymer
impregnated sorbent
organic loading (wt %)
PEI PAA PEI PEI PEI PEI PAA PAA PAA PAA
ALPEI fresh ALPAA fresh ALPEI_21_110 ALPEI_21_70 ALPEI_5_110 ALPEI_5_70 ALPAA_21_110 ALPAA_21_70 ALPAA_5_110 ALPAA_5_70
41.6 45.8 39.0 41.7 41.4 40.0 44.4 45.3 42.0 44.8
O2 concn (%)
21 21 5 5 21 21 5 5
oxidation temp (°C)
110 70 110 70 110 70 110 70
The BET surface area of the synthesized alumina was 223 m2/g, while the pore volume (at P/P0 = 0.99) was 1.19 cm3/g. The average mesopore diameter of the synthesized alumina was calculated to be 17.3 nm. On the synthesized alumina support, the amino polymers PAA and PEI were impregnated via wet impregnation with methanol as the solvent at room temperature. The organic loadings for all the synthesized adsorbent materials were kept close to 40% by weight. 3.1. Oxidation of Impregnated PAA and PEI Adsorbents. The impregnated PEI- and PAA-based adsorbents were oxidized in flowing oxygen in a fixed-bed reactor. The CO2 uptake for the fresh samples and those oxidized under humid conditions are shown in Table 1. For PEI/alumina adsorbents, the CO2 uptake capacity was reduced drastically by about 70% when the adsorbent was treated under a humid flow of 21% oxygen at 110 °C for 20 h (ALPEI_21_110; from 1.87 to 0.56 mmol/g). For samples treated at the lower oxidation temperature (70 °C) with 21% oxygen flowing under humid conditions, the capacity of the oxidized adsorbent (ALPEI_21_70) decreased by ∼35% to 1.24 mmol/g. There was considerably less oxidative degradation for the PEI adsorbents, as evidenced by the measured CO2 adsorption capacities, when the oxidation was carried out at the much lower oxygen concentration of 5% under humid conditions. The CO2 adsorption capacity for the oxidized sample was found to be 1.73 mmol/g when the oxidation was performed at 110 °C for 20 h (ALPEI_5_110). This was a reduction of only 7.5% relative to the fresh, nonoxidized PEI-impregnated adsorbent. At the same O2 concentration of 5% but at the lower temperatures of 70 °C, it was observed that no reduction in CO 2 capture capacity occurred after oxidation (ALPEI_5_70). The collected results are presented in Figure 3. From the above results, it can be inferred that, at the oxygen concentration typical of flue gas from the coal-fired power plants (5%) under humid conditions, PEI/alumina sorbents demonstrate good stability over the temperature range studied (
