Introduction

Since 1970, the atmospheric CO2 concentration has steadily increased from 310 to 400 ppm. The responsible factors for such sharp rise in CO2 concentration are industrial stacks and fugitive emissions, auto exhausts, thermal power stations, natural wetlands and ruminates. The single largest source of CO2 emission is thermal power plants (Edenhofer 2011). During the past decade, different kinds of dry adsorbents such as zeolites, metal–organic frameworks and microporous organic polymers have been explored as alternatives to the current commercial absorption technology (Xu and Hedin 2014; Abanades et al. 2015). In designing solid adsorbents for CO2 capture, the adsorbents should have (1) an adsorption capacity greater than 3.0 mmol g−1 at pressure 1 bar/temperature between 313 and 353 K, (2) reasonable rate of CO2 adsorption, and (3) high selectivity toward CO2. Some of the classical examples of polymer adsorbents reported in the literature are: (1) Hyper-cross-linked polymers containing tetraphenylmethane showed CO2 adsorption capacity of 2.5 mmol g−1; and the hyper-cross-linked polymers formulated with binaphthol reported CO2 adsorption capacity of 4.00 mmol g−1 (Errahali et al. 2014). (2) Benzimidazole-linked showed very good CO2 uptake, i.e., 5.34 mmol g−1 at 1.0 bar (Altarawneh et al. 2014). (3) Poly (melamine-formaldehyde) materials showed a CO2 uptake capacity of 4.54 mmol g1 at 273 K (Tan et al. 2013). (4) A coordination polymer network with exceptional CO2 uptake (5.54 mmol g−1) is reported (Choi and Suk 2009).

The high-adsorption-capacity adsorbents with non-selectivity restrict application of adsorbents for real-life situations. The non-selective materials are difficult to regenerate, and the adsorbed gases cannot be reused due to the presence of non-targeted gases. For example, stack emissions from thermal power plants contain 14–16% CO2, 75–80% N2 and 10% other gases. Such situations demand precisely designed materials with exceptionally high capability to recognize target molecule in complex mixtures/matrices. Adsorbents prepared with metal ions and organic linkers emerged as highly selective CO2 adsorption adsorbents at ambient conditions (Li et al. 2013). Lu et al. (2011) developed a group of high-surface-area polymers by grafting with sulfonic acid and its lithium salt exhibited high selective CO2 adsorption capacities. The iminebenzothiazole polymer prepared via Schiff’s base condensation reaction showed good CO2 uptake (7.8 wt% at 273 K/1 bar). The polymer also reported selectivity for CO2 in the gaseous mixtures N2 (51) and CH4 (6.3) (Rabbani et al. 2017). The benzimidazole-linked polymers are another group of adsorbents that showed CO2 selectivity in the presence of N2 (70) and H2 (10) (Zhu et al. 2013a, b). McDonald et al. (2014) reported CO2 capture by amino acid ionic liquids, and impact of water on CO2 capture. The ionic liquids—tetramethylammonium glycinate ([N1111][Gly]) and tetraethylammonium prolinate ([N2222] [Pro])—with CO2 under wet conditions form carbamate by the amine-functionalized anions of these salts. By adding CO2 to these salts, the carbamate releases the covalently bound CO2. Liu et al. (2018) reported 3D network structures of nanofibers of hydrogels with adsorption capacity of 1.78 mmol g−1 and regenerable for about 10 times.

Another desired parameter of adsorbents for real-life application is thermodynamic efficiency of regeneration. Polyethyleneimine adsorbents are regenerable; however, formation of carbamate reduces the adsorbent performance in thermal swing regeneration (Drage et al. 2009). Triazine-based benzimidazole-linked polymers showed CO2 uptake capacity (5.19 mmol g−1) (Sekizkardes et al. 2014). The porous azo-linked polymers showed high CO2 uptake of 5.37 mmol g−1, and these polymers reported Qst values (28–30 kJ/mol) for CO2 capture (Arab et al. 2015).

Our earlier studies report that the naturally occurring solid amino acids with CO2-philic functionality are good candidates of CO2 adsorption (Chatterjee et al. 2016). In this study, the polymer has been imprinted with amino acid taurine to increase surface area considerably, and the cavities are decorated with CO2-philic functionalities. The size of the cavities in the polymer ranges between 1 and 3 nm. The size of CO2 is 0.21 nm, and the pores with size 1.0 nm are desired for effective CO2 adsorption (Wilmer et al. 2012; Hudson et al. 2012). It is also presumed that the Qst for CO2 can be improved by incorporating polar functionality of amino acids that interact with CO2.

Experimental

Chemicals

Vinylbenzyl chloride and divinylbenzene were purchased from Sigma-Aldrich (St. Louis, USA). Amino acid taurine and the solvents dichloromethane and acetonitrile were purchased from Merck (Potsdam, Germany). High-purity (99.999%) gases CO2, N2 and CH4 were purchased from Nikhil Gases (Nagpur, India). Helium high-purity gas was used as the purge gas in simultaneous thermal analyzer during CO2 adsorption–desorption experiments. All other chemicals were used as purchased.

Preparation of amino acid-imprinted and non-imprinted polymers

The hyper-cross-linked polymers prepared with the composition of vinylbenzylchloride (VBC) and divinylbenzene (DVB) reported high surface area (Errahali et al. 2014), and the same composition is used in imprinting of taurine to create CO2-philic functionality. This is a new approach to create specific functionality in high-surface-area polymers. The imprinted polymers were prepared using the composition given in Table 1. The procedure followed for synthesis of the non-imprinted polymer is described in Krupadam et al. (2014). Briefly, the polymer precursors with DVB given in Table 1 were heated to 340 K for about 8 h under nitrogen, and the polymerization was controlled with methanol. The polymer monolith was separated from the solvent by filtration, and the monolith was thoroughly washed with methanol. Using analytical ball mill, the polymeric monolith was made into fine particles of size between 70 and 100 µm. The polymer particles were cleaned by Soxhlet extraction using methanol for 24 h and then dried in vacuum at 70 °C overnight. The other polymer, imprinted polymer, was prepared by following the same procedure; however, taurine was added before the addition of formaldehyde dimethyl acetal (FDA) and FeCl3. After polymerization, taurine was removed from the polymer using Soxhlet extraction using methanol for 24 h, and then, water was used as the solvent for Soxhlet extraction for another 24 h. After the removal of template, the polymer was dried in vacuum at 70 °C overnight. These polymers were used for adsorption of CO2 and other gases.

Table 1 Polymer composition and surface properties
Full size table

Characterization of polymers

Surface area, pore volume and average pore diameter of the polymers were determined using N2-adsorption isotherms in Micromeritics ASAP 2420 at 77 K. Nitrogen adsorption surface area of IP was computed by Brunauer–Emmett–Teller equation (Walton and Snurr 2007), and pore volumes were calculated from the Dubinin–Radushkevich equation based on N2-adsorption isotherms (Carrasco-Martin et al. 1993). Polymers prepared were kept under vacuum at 373 K for about 7 h. Surface topology of the polymers was viewed using atomic force microscope (FlexAFM, Nanosurf AG, Switzerland). Tapping tips were procured from Silicon AFM Probes (Tap 300, Ted Pella, Inc) whose resonant frequency was between 290 and 410 kHz. The micrograph of the imprinted polymer was analyzed with Easyscan 2 Software. Furthermore, laser Raman microscope (Alpha 300 RA, WITec, Germany) equipped with a frequency-doubled Nd:YAG laser for 532 nm excitation to observe morphology of the polymers. The Raman imaging was integrated with spectral lines of CH2 stretching using 1 s with accumulation of 10. The measurements were acquired using the WITec Control Software. The attenuated total reflectance infrared spectra of polymers were recorded on a Digilab FTS equipped with an HgCdTe detector. The wave numbers scan ranges from 600 to 4000 cm with 4 cm−1 of spectral resolution.

CO2 adsorption experiments

The adsorption experiments for CO2 adsorption and desorption were conducted in a simultaneous thermal analyzer (STA; PerkinElmer 6000), and the detailed procedure is reported in our earlier article (Chatterjee et al. 2016). The adsorption isotherms were drawn at 273 and 303 K at different pressures (up to 10 bar). After adsorption of CO2, the adsorbed CO2 was flushed with He gas at 473 K for 5 min. Once the polymer was degassed, CO2 gas was passed till saturation to record the adsorption capacity. These adsorption–desorption cycles were repeated for 50 cycles. For selectivity measurements, equimolar concentration of gaseous mixtures (CO2 and N2; and CO2 and CH4) was passed through the simultaneous thermal analyzer. The Qst of CO2 was measured in the Micromeritics ASAP 2420, and data from the adsorption isotherm were also used.

Results and discussion

Molecular imprinting is used to create molecule-specific cavities in terms of both geometry and functionality in the polymer to selectively recognize and capture the targeted molecule (Wulff 1995). In this study, molecular imprinting technique was used to create cavities of size between 1 to 3 nm, and these cavities are decorated with CO2-philic functionality in the polymers. Earlier reports indicate that pores bigger than 5 times of the adsorbate molecular size are required criteria for effective adsorption of gases, and this conclusion was derived from the study of mathematical simulations by Salles et al. (2008). The molecular size of CO2 is 0.21 nm, and the pores with size less than 1.0 nm are required for CO2 adsorption (Wahby et al. 2010). The introduction of CO2-philic functional groups improves the Qst, which influences capacity and selectivity of adsorbents for CO2. Banerjee et al. (2009) reported a drastic increase in Qst of CO2/N2 adsorption selectivity by introducing polar functionalities; however, this procedure reduces the surface area of the adsorbents significantly. The probable strategy to address this issue is to use high-surface-area polymers as the base materials, and then, the surfaces can be functionalized with CO2-philic functional groups. The vinylbenzyl chloride-co-divinylbenzene polymer was imprinted with taurine which forms small pore (size 1–3 nm) with sulfonate and amine functionalities. Such specific imprinted pores selectively adsorb CO2 compared to N2 and CH4 in gaseous mixtures.

The adsorption of N2 was performed at 273 K to evaluate the existence of permanent porosity and nature of isotherms of imprinted and non-imprinted polymers. The adsorption isotherms are type I, and the imprinted polymer has permanent microporosity (Fig. 1a). Hence, imprinted polymer showed rapid N2 adsorption at low pressures. Adsorption isotherms of CO2 for imprinted and non-imprinted polymers have been plotted for 273 and 303 K, at different pressures up to 1 bar (Fig. 1b). The CO2 adsorption capacity of imprinted polymer is compared with other adsorbents such as porous polymers (6.0 mmol g−1) (Popp et al. 2015), benzimidazole-linked polymers (5.3 mmol g−1) (Zhu et al. 2013a, b) and Tröger’s base-derived covalent organic polymers (5.16 mmol g−1) at 273 K (Byun et al. 2014). It would be clear from the adsorption isotherms that the taurine-imprinted polymer has the functionality of amine and sulfonic acid in the micro/nanocavities. Infrared spectra of the imprinted polymer before and after adsorption of CO2 provide further evidence of role of taurine-imprinted polymer (Fig. 2a). The stretching vibrations of functional groups, –CH2 and –CH, in the polymer showed peaks 2900 and 2850 cm−1 that correspond to C=C stretching vibration in the aromatic ring of divinylbenzene and vinylbenzyl chloride of the non-imprinted polymer. A quite small band at 1260 cm−1 corresponds to C–Cl stretching in vinylbenzyl chloride. The –S=O functionality peak at 1190 cm−1 and peaks at 1410 and 1510 cm−1 for –C–N (stretch) and –N–H (bend) are due to taurine functionality left during molecular imprinting (Ohno et al. 1992). The decrease in peak intensity at 1190, 1410 and 1510 cm−1 after adsorption of CO2 would be seen. Furthermore, laser Raman microscopy provided clear evidence of existence of nanocavities created during taurine imprinting in the polymer, and also the Raman spectra of the polymer shows the specific functionality of –C–S–, –C–N–, –C–O– and –CH2 vibrational spectra at 1080, 1610, 1760 and 2970 cm−1, respectively, evidenced taurine imprinting (Fig. 2b).

Fig. 1
figure 1

Gas adsorption and desorption isotherms a N2 adsorption and desorption isotherms of imprinted and non-imprinted polymers at 273 K and b CO2 adsorption and desorption isotherms of imprinted and non-imprinted polymers at 273 and 303 K. The adsorption isotherms are type I which represents that imprinted polymers contain permanent micro/nanoporosity. The adsorption capacity of imprinted polymer for CO2 is 5.67 mmol g−1 at 303 K

Full size image
Fig. 2
figure 2

a Infrared spectra, b Raman image with spectra, c pore size distribution in imprinted and non-imprinted polymers and d atomic force microscope image of imprinted polymer (inset: 3D view showing the existing imprinted cavities in the polymer)

Full size image

Morphologically, the size of imprinted pores is about 5 times higher than that of CO2 (0.209 nm) which could be another probable reason to retain CO2 molecules in the pores followed by forming number of double and multiple interactions between the polymer pore functionality and CO2. The IP showed surface area of 1020 ± 10 m2 g−1 and pore volume 0.316 cm3 g−1 compared with P (non-imprinted polymer). The size distribution curves were calculated by non-local density functional theory (Jagiello and Olivier 2009), where the median value of the pore size of taurine-imprinted polymers was 1 nm. An increase in the surface area 370 m2 g−1 was noticed due to molecular imprinting of taurine. As shown in Table 1, the total pore volume of imprinted polymer was 0.316 cm3 g−1 which is considerably more compared with porous polymer networks (0.07 cm3 g−1), hyper-cross-linked polymer (0.08 cm3 g−1) and molecularly imprinted polymers (0.067 cm3 g−1) (Zhao et al. 2012). The pores formed in the non-imprinted polymer were larger (average pore diameter 40 nm) than the pores in the imprinted polymer (average pore diameter, 1 nm); this lower pore diameter in imprinted polymer would be responsible for increase in surface area and pore volume (Fig. 2c). The morphology of imprinted polymer has been viewed using atomic force microscope. The external surface of this polymer is seen to be a continuous distribution of micro/nanocavities (Fig. 2d). Remarkably, at 273 K, imprinted polymer adsorbs 5.67 mmol g−1 of CO2; this quantity is about 40% higher than the hyper-cross-linked polymers formulated with bis(chloromethyl) monomers (Fontanals et al. 2015).

Practically, the fuels combusted in thermal power plants emit flue gases at ambient conditions typically containing CO2 concentration which is lower than 15%. Similarly, natural gas reservoirs constitute about 20% of CO2 and 80% of CH4. Such instances require selective CO2 separation from CH4 and other gases which is a challenging task to provide clean fuel and reusable CO2. The selective adsorption of CO2 from the gaseous mixture under low thermodynamic conditions is a great challenge. In this study, imprinted and non-imprinted polymers were used to evaluate the selectivity toward CO2 in the presence of N2 and CH4. The performance of imprinted polymer for uptake of CO2 in the presence of N2 and CH4 is depicted in Fig. 3a. The imprinted polymer showed significantly higher selectivity and capacity compared with P, which would be due to small pore size and functionality created in the imprinted sites. Compared to non-imprinted polymer, the pore diameter of imprinted polymer is quite small (i.e., 1 nm) and would be the cause of high-selectivity adsorption of CO2. The selectivity of imprinted polymer at 303 K and 10 bar for the adsorption of CO2 in the presence of N2 and CH4 is 91 and 87%, respectively. The selectivity reported by the imprinted polymer is better than the high-selectivity materials reported in the literature such as benzimidazole-linked polymers (Popp et al. 2015). The benzimidazole-linked polymers with the highest CO2 adsorption showed lowest selectivity (CO2/N2, 32:1). With the surface area 1020 m2 g−1, imprinted polymer showed a significant CO2 adsorption at 1 bar and 273 K (up to 5.67 mmol g−1); however, the CO2 adsorption at 10 bar by the imprinted polymer was 26.3 mmol g−1. The high CO2 adsorption at low pressure could be due to well-formed pore in imprinted polymer which condenses more CO2 molecules in the polymer. When compared with 3D coordination polymers, the imprinted polymer showed lower adsorption of CO2 at high pressures. This would be due to the existence of vacant space in 3D-coordinated polymer and other such materials (Choi and Suk 2009).

Fig. 3
figure 3

a Adsorption capacity of imprinted and non-imprinted polymers for CO2, N2 and CH4, b isosteric heat of adsorption for CO2 calculated from the isotherms measured at 273 and 303 K for imprinted and non-imprinted polymers and c CO2 adsorption after regeneration of imprinted polymer

Full size image

The isosteric heat of adsorption (Qst) is an indicative of adsorption equilibrium achieved under controlled conditions such as pressure and temperature for a given adsorbate. The Qst of adsorption of polymer was calculated form Langmuir isotherm fits of CO2 adsorption at 273 and 303 K. The Qst value calculated for CO2 adsorption for imprinted polymer was 39.8 kJ mol−1, which represents that CO2 uptake by imprinted polymer was predominantly by physical adsorption. Figure 3b shows a plot of Qst as a function of CO2 capture. The presence of sulfur and nitrogen atoms in the polymer would be responsible for reduction in the Qst value (4.0 kJ mol−1), thereby improving the affinity toward CO2 of imprinted polymer compared with non-imprinted polymer. Initially, the Qst was decreased and then stabilized with the CO2 uptake which represents that imprinted sites with high energy are filled initially, and then reached a saturation level after adsorption of CO2 to a certain amount. It is important to observe the existence of amine and sulfonic acid functionalities in the imprinted cavities of imprinted polymer that had showed slightly higher Qst which represents stronger CO2 interaction with imprinted polymer compared with non-imprinted polymer. The heat of adsorption Qst of imprinted polymer is higher than the microporous polycarbazoles (27–31 kJ mol−1) or benzimidazole-linked polymers (20–27 kJ mol−1) or conjugated microporous polymers (25–33 kJ mol−1) or hyper-cross-linked polymers (20–24 kJ mol−1) (Martin et al. 2011; Zhu et al. 2013a, b). The Qst values of imprinted polymer are below 50 kJ mol−1, the interaction between CO2 and polymers are reasonably weak, and this facilitates regeneration of polymers with quite low energy.

Regeneration of adsorbent is an important aspect which influences the affordability, and multiple cycles of use reduces the cost of CO2 capture (Abanades et al. 2004). The regeneration of imprinted polymer was evaluated in simulated temperature and pressure-controlled ASAP 2020 analyser. It was found that even after 50 cycles, there was no considerable change in capacity (Fig. 3c). The imprinted polymer reported lower regeneration energies compared to amine solutions. The polymer designed with synergistic approach of hyper-cross-linking and molecular imprinting has high potential for commercial applications due to its selective adsorption of CO2 and physical–chemical robustness and stability.

Conclusion

In conclusion, taurine-imprinted polymer formulated with divinylbenzene and vinyl benzyl chloride showed high adsorption capacity of 5.67 mmol g−1 and selectivity. Furthermore, the adsorbent showed no apparent loss in the CO2 adsorption capacity till 50 cycles. The polymer prepared using molecular imprinting of CO2-philic templates demonstrated applicability of adsorbents for real-life CO2 capture applications. Further studies are underway to establish physical–chemical robustness and stability under different industrially relevant conditions.