Organic–inorganic hybrid polyaspartimide polymers were synthesized via Michael addition by using octa(aminophenyl)silsesquioxane (OAPS) (2.5–30 wt %), leading to a family of porous polymers. The formation of NH structure during Michael addition assists in CO2 capture. The CO2 capture capacity increased with increasing concentration of OAPS from 2.5 to 25 wt % in the PAI/OAPS polymers.
Carbon capture and sequestration (CCS) technologies offer a great potential for diminishing CO2 emissions in the atmosphere. Accordingly, four key approaches have been widely investigated over the last decade for CCS: cryogenic processing,1 membrane purification,2 absorption with liquids,3 and adsorption using solids.4 Among these approaches, adsorption on solid media is considered a promising technology for CO2 capture, offering possible energy savings as compared to other more established absorption technologies.5, 6 In recent years, a number of porous solids have been developed as potential adsorbents for CO2 capture,7–9 including activated carbons,10 zeolites,11 calcium oxides,12 hydrotalcites,13 hybrid crystalline solids (e.g., metal organic frameworks, zeolitic imidazolate frameworks, covalent-organic frameworks, monochlorophenols, pentachlorophenols),14 amino-polymers,15 and organic–inorganic hybrid polymers.16 Among them, zeolites, activated carbons, and organic–inorganic hybrids work at room temperature.
Zeolites and organic–inorganic hybrids show higher CO2 adsorption capacities than activated carbons.17 To desorb CO2 adsorbed on the adsorbent for subsequent reuse, zeolites require higher energy consumption than organic–inorganic hybrids because of the requirement of controlled pressures and temperatures. Hence, significant development is underway by several groups toward utilization of organic–inorganic hybrid polymers as CO2 adsorbents.18 Indeed, CO2 capture by adsorbents produced through grafting amines onto nanoporous silica gels and mesoporous silica has been intensely studied.19 The nature of the silica supports for the amine modified adsorbents impacts their adsorption properties. For example, SBA-15 shows the highest adsorption capacity and fastest adsorption kinetics.20 In earlier studies,21, 22 it was reported that polymer pores smaller than five times the molecular size of the adsorbate were effective in gas adsorption at atmospheric pressure and pores less than 1.0 nm were still effective for CO2 capture at atmospheric pressure. Herein, we made an attempt to use an organic–inorganic hybrid porous polymer, that is, polyaspartimide (PAI),23 as a solid CO2 adsorbent, which was synthesized via the Michael addition reaction by using a kind of nano-sized polyhedral oligomeric silsesquioxane (POSS),24 that is, octa(aminophenyl)silsesquioxane (OAPS), as a cross-linker.
OAPS, a well defined nano-sized cage, is one of the derivatives of the organic–inorganic hybrid POSS, that has attracted many researchers in the field of nanomaterial science.25–27 The octa-amino functionalized POSS (OAPS) consists of a rigid cubic silica core of a 0.53 nm side length and can have organic functional groups connected at each vertex of the cubic core for specific functionalities.28, 29 In addition, OAPS possesses excellent thermal stability, good chemical resistance, and mechanical strength,30 which enable a wide range of applications. Hence, in this study, we selected OAPS to prepare cross-linked PAI via Michael addition. This process offers an amine enriched porous silica based structure, which greatly facilitates CO2 capture and shows results comparable to other silica based materials.
Phenyltrichlorosilane (97%, Aldrich) and 1,3,5-benzenetriamine (BTA, 98%, Ruiyi) were used as received. Benzyltrimethylammonium hydroxide in methanol solution (40%), palladium (5% on carbon), p-toluene sulfonic acid (p-TSA, 97%), p-phenylenediamine (PDA, 97%), and triethylamine (TEA, 99%) were purchased from Alfa Aesar. Maleic anhydride (98%, Showa), acetic anhydride (99.5%, TEDIA), formic acid (98%, Aldrich), benzene (99%, TEDIA), dimethylformamide (DMF, 99.5%, TEDIA), dimethylacetamide (DMAc, 99.85%, TEDIA), and tetrahydrofuran (THF, 100%, TEDIA) were purchased and used for the synthesis. Tetrahydrofuran, acetic anhydride, and TEA were further dried and maleic anhydride was further purified with the recrystallization method before use. 1,3,5-tris(maleimido)benzene (TMIB) was synthesized from BTA according to a reported method.31
Synthesis of POSS-PAI Polymers
The preparation procedures for linear PAIs were followed for the preparation of PAI/OAPS.23 A 250 cm3 three-necked flask was equipped with a magnetic stirrer, a reflux condenser, a thermometer, and a nitrogen inlet. The aromatic diamine, that is, PDA and DMAc (50 cm3) were added to the flask and the mixture was stirred. Once the PDA moiety was completely dissolved, the synthesized compounds TMIB and OAPS (2.5–30 wt %, synthetic procedures offered in the Supporting Information) were added to the flask with 0.1 g of p-TSA. The molar ratios of OAPS:TMIB and PDA:TMIB were 3:4 and 3:2, respectively in all cases. The reaction mixture was then stirred at 100 °C for 96 h, during which an appreciable increase in the solution viscosity and darkening in colour were noticed. Pouring the viscous reaction mixtures into excess ethanol with vigorous stirring isolated the polymer. The pure polymer was extracted with hot ethanol using a Soxhlet extractor and subsequently dried at 70 °C in vacuo for 24 h. FTIR analyses for all the samples were carried out to identify their structures.
Fourier Transform Infrared (FTIR) Spectra
Spectra were recorded on a Perkin–Elmer RX I FTIR spectrometer. About 100 mg of optical-grade KBr was ground in a mortar with a pestle, and enough solid samples were ground with KBr to make a 1 wt % mixture for making the KBr pellets. After the sample was loaded, a minimum of 16 scans was collected for each sample at a resolution of ±4 cm−1.
Nuclear Magnetic Resonance (NMR) Spectra
All 1H and 13C NMR analyses were done in acetone-d6 and recorded on a Bruker AMX 400 spectrometer. 1H NMR spectra were collected at 400 MHz using a 8000 Hz spectral width, a relaxation delay of 3.5 s, a pulse width of 45°, 32 K data points, and acetone-d6 (2.0 ppm) as an internal reference. 13C NMR spectra were obtained at 100.6 MHz using a 25,000 Hz spectral width, a relaxation delay of 1.68 s, a pulse width of 45°, 32 K data points, and acetone-d6 (30.3 and 206.4 ppm) as an internal reference.
Assessment of the CO2 Capture Capacity
The CO2 adsorption capacity of the samples at atmospheric pressure was evaluated in a thermogravimetric analyser (SDT 600 TGA). In a typical experiment, a sample (ca. 50 mg) was loaded in the TGA, dried at 125 °C under inert atmosphere of nitrogen at a rate of 20 °C/min and kept at the isothermal condition for 30 min before the adsorption experiment. The system was then cooled to ambient temperature and, after stabilization of the sample mass and temperature (25 °C), the nitrogen flow was changed to 5% CO2 (CO2:N2 = 5:95) or 100% CO2 (60 mL/min) at the isothermal condition for 60 min. The CO2 adsorption capacity at 25 °C was assessed from the maximum mass increased of the samples when exposed to a pure CO2 and mixed CO2 atmosphere.
Nitrogen adsorption and desorption isotherms were measured at 77 K on a NOVA e1000 (BEL Japan) equipped with an accelerated surface area and porosimetry system. Before each adsorption measurement, the samples were evacuated at 100 °C under vacuum (p < 76 torr) in the degas port of the adsorption analyzer. The specific area, SBET, was determined from the linear part of the BET equation; the pore volume was calculated using the BET plot from the amount of nitrogen gas adsorbed at the last adsorption point (P/Po = 1); and the pore size distribution was estimated with the Barrett-Joyner-Halenda (BJH) method.
RESULTS AND DISCUSSION
Structure of PAI/OAPS
OAPS was synthesized in a three-step reaction. The first step involved the synthesis of octaphenylsilsesquioxane (OPS) via the hydrolysis and condensation of phenyltrichlorosilane and the subsequent rearrangement reaction catalysed by benzyltrimethylammonium hydroxide. The second and third steps (Scheme 1) were nitration of OPS to obtain octa(nitrophenyl)silsesquioxane (ONPS) and hydrogen-transfer reduction of ONPS to OAPS, respectively. FTIR, 1H and 13C spectra of the three POSS monomers, including OPS, ONPS, and OAPS, are shown in Supporting Information Figures S1–S3.
PAI/OAPS polymers were synthesized via the Michael addition between aromatic diamine and aromatic trimaleimide using various weight ratios of the OAPS in the presence of p-TSA (Scheme 2), and the structure of the polymers was identified from the FTIR study. The FTIR spectra of the PAI/OAPS are shown in Figure 1. The presence of OAPS was revealed by the strong band at about 1110 cm−1 because of the asymmetric stretching vibrations of the SiOSi bonds.30 The peaks at 3307 and 3378 cm−1 are assigned as the νNH vibrations of the free amino groups attached to the phenyl ring in the PAI/OAPS polymer, whose intensity increased with increasing the OAPS loading. The absorption bands near 1790 cm−1 (asym. CO str.), 1720 cm−1 (sym. CO str.), and 1380 cm−1 (CN str.) were assigned to the characteristic absorption bands of the PAI's imide groups. The peaks at 697 and 746 cm−1 are related to the out-of-plane deformation of the phenyl groups. Bands at 1137 cm−1 (partially overlapped with the absorption peaks of the SiO bonds) and 1432 cm−1 were caused by the Si-phenyl group bonds deformation. The stretching vibration of the CC bonds was revealed by the peak at 1595 cm−1, while bands around 3000 cm−1 were attributed to the CH stretching modes. There were no significant maleimide ν CH bands at 666 cm−1 in the PAI/OAPS polymers, which indicated that most of the maleimide groups had reacted with the amino groups in the Michael addition reaction.
The N2 adsorption/desorption isotherms for the PAI/OAPS polymers (2.5, 5, 10, 20, 25, and 30 wt % OAPS loadings) gave information about the textural properties of the materials. The specific surface area, pore volume, and average pore size for all the samples were summarized in Table 1. A typical nitrogen adsorption/desorption isotherm of the sample, 25 wt % of OAPS in PAI/OAPS, was shown in Figure 2, exhibiting a type IV isotherm and type H1 hysteresis loop according to the IUPAC classification. The sharp increase in nitrogen uptake between partial pressures P/Po = 0.83–0.95 indicates capillary condensation in mesopores. The specific surface areas were found to increase significantly upon the OAPS crosslinking, and increased with increasing the OAPS loading of up to 25 wt %. Such a significant increase in the specific surface area can be attributed the inclusion of the cubic silica cages of the OAPS into the PAI/OAPS polymer. The specific surface area of the PAI/OAPS polymer of 30 wt % OAPS loading decreased significantly. This may be caused by the possible agglomerations of OAPS at higher OAPS loadings during polymerization.
Table 1. Textural Characteristic Properties of PAI/OAPS Polymers and Corresponding CO2 Capture Capabilities
CO2 adsorption assessments for PAI/OAPS (2.5–30 wt % of OAPS) at 25 °C were carried out using TGA under a 5% CO2 flow with atmospheric pressure. The mass gain in the TGA plot on feeding CO2 gas was taken as the CO2 adsorption capacity. The CO2 adsorption capacities with respect to different OAPS loadings in the PAI/OAPS polymers for both the cases of 5 and 100% CO2 atmosphere were summarized in Table 1.
A slow and continual CO2 uptake was observed up to a maximum of 2.31 wt % for the PAI/OAPS sample of 2.5 wt % OAPS [Supporting Information Fig. S4(a)], whereas fast and higher CO2 uptakes with increasing loading of OAPS up to 25 wt % OAPS for the samples were obtained. This implied that the interactions between the acidic CO2 and basic amine groups of the polymer might be a weak chemisorption reaction. The interactions of CO2 with the unreacted primary amine of OAPS and formation of secondary amines through the Michael addition reaction in the polymer are illustrated in Supporting Information Scheme S1. Herein, the presence of the secondary amines and free primary amines in the polymer provided selective affinity for CO2 via the formation of ammonium carbamate species under atmospheric conditions. A realistic mechanism behind the carbamate formation is the weak bonding attributed to the nucleophilic attack of the NH2 group on CO2. The primary NH2 group in the OAPS is basic, and the lone pair of electrons on the N atom can attack the C atom in the acidic gas CO2. Subsequently, the bonds between the H atom on NH2 and O atom on CO2 are established to form the carbamate. Thereafter, the active H atom in carboxyl may then form a hydrogen bond with the nearby amine group, and thus stabilize the chemisorption of CO2.
Pure CO2 gas (60 mL/min) was also used instead of 5% CO2 gas to evaluate the CO2 adsorption capacities of the polymer samples, which were then compared to the results obtained in 5% CO2 atmosphere. We observed an increase in the CO2 uptake of 1–1.5 wt % in 100% as compared to 5% CO2. The maximum CO2 adsorption capacity was obtained for the PAI/OAPS polymer of 25 wt % OAPS loading [Supporting Information Fig. S4(b)] in 100% CO2 atmosphere. Further increasing the OAPS loading to 30 wt % led to a decrease in the CO2 adsorption capacity.
Porous organic–inorganic hybrid PAIs were successfully prepared from octa(aminophenyl)silsesquioxane (OAPS) via the Michael addition reaction and were determined to possess a maximum CO2 adsorption capacity of 5.59% for the PAI/OAPS polymer of a 25 wt % OAPS loading in 100% CO2 atmosphere. The formation of the secondary amines during the Michael addition reaction and the presence of unreacted free amines in the PAI/OAPS polymers effectively facilitated the CO2 capture by forming carbamate species. The increasing loading of OAPS in the polymers increased the amine group content in the polymers and thus enhanced the CO2 adsorption capacity through the traditional carbamate mechanism. Too high an OAPS loading however led to agglomeration of OAPS and thus significantly lower specific surface area and a decrease in the CO2 capture capacity.
This work was financially supported by the National Science Council of the Republic of China (Taiwan) under grant NSC-98-2221-E-034-MY3. The authors thank the Department of Chemistry, National Tsing Hua University, Hsinchu 30013, Taiwan for assistances in sample characterizations.