1 Introduction

The rapidly increasing concentration of atmospheric carbon dioxide generated from combustion of fuels (include coal, petroleum and natural gas) has aroused environmental concern, as a result, carbon capture and sequestration (CCS) gradually gain their growing popularity (D’Alessandro et al. 2010; Sumida et al. 2011; Goeppert et al. 2012; Lu et al. 2013; Wang and Xu 2014; Romanov et al. 2015). Chemical absorption by aqueous solutions of ethanolamines (Brennecke and Gurkan 2010), which is most widely used CCS technique, suffer from several fatal defects including solvent loss, corrosion, and tremendous energy cost for the regeneration (Service 2004). So more efforts have been devoted to solid porous adsorbent. Various categories of novel materials with extraordinary porosity have been discovered, including silicas (Suhendi et al. 2013), zeolites (Wakihara et al. 2010), metal–organic frameworks (Li et al. 2014), zeolitic imidazolate frameworks (Cai et al. 2014) and carbon materials (Sevilla and Fuertes 2011). With the advantages of outstanding BET surface area and stability, tremendous porous organic polymers (POPs) have been designed (Jiang et al. 2009; Dawson et al. 2012; Zhao et al. 2012; Han et al. 2013; Liu et al. 2013; Zhu et al. 2013; Thompson et al. 2014; Zhu and Zhang 2014; Puthiaraj et al. 2015). More importantly, the organic skeletons of POPs give rise to their easy-to-functionalize nature so that enhancement of CO2/N2 selectivity is able to be realized via the incorporation of CO2-philic groups into networks (Thomas 2010). Typically, PPNs (porous polymer networks) (Ben et al. 2009; Yuan et al. 2011) was synthesized with ultrahigh Brunner–Emmet–Teller (BET) surface area, the specific surface area of materials calculated by using adsorption theory developed by three scientists, namely Brunner, Emmet and Teller), some of which can be further incorporated by various groups, such as sulfonic acid, lithium sulfonate and ammonium sulfonate, to significantly improve the CO2 uptake (3.6–3.7 mmol/g, 295 K, 1 bar) (Lu et al. 2011, 2013). However, in spite of their impressive performance, these polymers are difficult to be used for industrial scale-up applications because of the harsh synthesis conditions.

Alternatively, a facile and efficient method has been reported, known as “knitting” aromatic building blocks to a polymeric skeleton. A large number of heterocyclic rings were picked as monomers to afford crosslinking networks by applying the new technique (Li et al. 2011; Woodward et al. 2014; Zhu et al. 2014; Xu et al. 2015). However, their modest CO2 uptakes need great improvement for the future CO2 capture applications (Luo et al. 2012).

Herein, a single-step procedure was adopted to facilely realize “knitting” strategy to afford a porous organic polymer via Friedel–Crafts alkylation (FC-POPs). Triptycene, which may allow for a high degree of internal free volume, was chosen as monomer, in order to deliver FC-POPs with abundant mesopores. The mesopores of FC-POPs shows a facility for further modifications, to investigate the influence of pore structure and nitrogen-site density on CO2 adsorption. To realize this fancy, various groups with nitrogen-site including amine-group, diethylenediamine (DETA) and polyethylenimine (PEI) were grafted into the network to produce task-specific microenvironments by modifying microporosity and nitrogen-site density (Scheme. 1). Accordingly, the CO2 adsorption capacity and selectivity of CO2 over N2 were significantly improved after the post-synthesis modifications.

Scheme 1
scheme 1

Synthesis of FC-POP and its derivatives. The reaction conditions in each step; a formaldehyde dimethyl acetal, FeCl3, dichloroethane, 45 °C, 5 h; 80 °C, 19 h; b paraformaldehyde, H3PO4, AcOH, HCl, 90 °C, 72 h; c i potassium phthalimide, DMF, 100 °C, 8 h; ii hydrazine monohydrate, EtOH, reflux, 20 h; d DETA, 90 °C, 72 h; e PEI, 90 °C, 72 h

Full size image

2 Experiments

2.1 Materials and synthesis

Solvents, reagents and chemicals were purchased from Aldrich and TCI. All were used without any further purification.

2.1.1 Synthesis of FC-POP

Similar to Tan’s method (Li et al. 2011), Triptycene (0.64 g, 2.5 mmol) and formaldehyde dimethyl acetal (1.14 g, 15.0 mmol) dissolving in 5 mL dichloroethane was added to a 25 mL flask. Anhydrous FeCl3 (2.44 g, 15.0 mmol) was then added as a catalyst, and the mixture was stirred at 45 °C for 5 h and then heated to 80 °C for another 19 h. After cooling to room temperature, the mixture was filtered, washed by methanol for 3 times. The solid was collected, and extracted with methanol by Soxhlet apparatus for 24 h, brown solid was obtained after drying (yield ~95%).

2.1.2 Synthesis of FC-POP–CH2Cl

A mixture of 0.30 g FC-POP, 1.50 g paraformaldehyde, 9.0 mL acetic acid, 4.5 mL phosphoric acid and 30.0 mL concentrated hydrochloric acid were charged in a flask, by following a well-known method (Lu et al. 2012). The flask was sealed, then the mixture was heated to 90 °C and maintained for 3 days. After cooling to room temperature, the solid was collected and washed with water and methanol for three times, then dried to afford FC-POP–CH2Cl.

2.1.3 Synthesis of FC-POP–CH2NH2

A mixture of FC-POP–CH2Cl (0.20 g), potassium phthalimide (1.19 g) and 20.0 mL N,N-dimethylformamide (DMF) was stirred at 100 °C for 8 h under an argon atmosphere, similar to a well-known method (Ilhan et al. 1999). The cooled mixture was filtered and cursorily washed by DMF. The resulting phthalimide derivative was mixed with 1 mL of hydrazine monohydrate in 12 mL ethanol. The reaction was refluxed for 20 h. The reaction mixture was filtered and washed subsequently by DMF, water and MeOH, then dried to afford FC-POP–CH2NH2.

2.1.4 Synthesis of FC-POP–CH2DETA

A mixture of 0.10 g FC-POP–CH2Cl and 10.0 mL diethylenetriamine (DETA) was charged in a flask, according to a well-known method (Lu et al. 2012). The flask was sealed then heated to 90 °C and maintained for 3 days. After cooling to room temperature, the solid was collected and washed with water and methanol for several times, then dried to afford FC-POP–CH2DETA.

2.1.5 Synthesis of FC-POP–CH2PEI

In a fashion similar to the synthesis of FC-POP–CH2DETA, 0.10 g FC-POP–CH2Cl and 10.0 g polyethylene imine (PEI, mw = 600) was reacted to afford FC-POP–CH2PEI.

2.2 Characteristics

Thermogravimetry analysis (TGA) were performed under N2 on a NETZSCH STA449F3, with a heating rate of 10 °C /min. 13C NMR measurements were performed on a 9.4 T Bruker Avance spectrometer at a Larmor frequency of 100.6 MHz. Measurements were made with a 4 mm MAS probe spinning at 15 kHz. Chemical shifts were externally referenced to TMS (δ = 0 ppm) using the methyl resonance of hexamethylbenzene (17.5 ppm relative to TMS). Nitrogen adsorption isotherms were measured at 77 K using Micromeritics ASAP 2020 static volumetric analyzer. Before adsorption measurements the polymer was degassed at 110 °C under vacuum. The BET surface area was calculated within the relative pressure range 0.05–0.30. Total volume was calculated at p/p 0 = 0.98 and micropore volume was calculated by t-plot method. FTIR data were obtained using a Nicolet Magna-IR 550 spectrometer. Elemental analysis was determined using a Vario EL III Elemental Analyzer (Elementar, Germany).

2.3 Gas adsorption experiments

2.3.1 Gas adsorption and desorption isotherms

Both of gas adsorption and desorption isotherms of POPs were measured using a Micromeritics ASAP 2020 static volumetric analyzer at the setting temperature. Prior to each adsorption experiment, the samples were degassed for 12 h at 110 °C ensuring that the residual pressure fell below 0.2 Pa and then cooled down to the target temperatures, followed by introduction of a single component gas (CO2 or N2) into the system. Once the adsorption process finished, the desorption experiment was automatically initiated. In the desorption stage, gas pressure gradually decreased and the corresponding amount of residue adsorbed gas was measured and calculated by the instrument.

2.3.2 Fits of isotherms

For FC-POP and FC-POP–CH2Cl, single-site Langmuir fit is appropriate. Their isotherms can be described by the Eq. (1).

$${q} = {q}_{{\text{sat}}} \frac{{b}p}{{\text{1} + {b}p}}$$
(1)

where, b is a parameter in the pure component Langmuir isotherm (Pa−1), p represents gas pressure (Pa), q is molar loading of gas components (mol/kg) and q sat is saturation capacity of gas components (mol/kg).

For FC-POP–CH2NH2, FC-POP–CH2DETA and FC-POP–CH2PEI, both physical and chemical interaction should be taken into account so that dual-site Langmuir (Eq. 2) fitting is fine.

$${q} = {q}_{{\text{sat,1}}} \frac{{b}_{\text{1}} p}{{\text{1} + {b}_{1} p}} + {q}_{{\text{sat},\text{2}}} \frac{{b}_{\text{2}} p}{{\text{1} + {b}_{\text{2}} p}}$$
(2)

where two distinct adsorption sites are assumed to be existed, so there are four parameters in the equation, namely q sat,1, b 1, q sat,2 and b 2.

2.3.3 Adsorption enthalpy

For CO2 adsorption, fits of isotherms were used to calculate adsorption enthalpy by employing Clausius–Clapeyron equation (Eq. 3).

$$\text{ln}\left( {\frac{{p_{1} }}{{p_{2} }}} \right) = \Delta {H}\left( {\frac{\text{1}}{{{T}_{1} }} - \frac{\text{1}}{{{T}_{2} }}} \right)$$
(3)

where p i is the pressure for isotherm i, T i is the temperature for isotherm i, R is 8.315 J/(K mol).

2.3.4 IAST calculation

Pure-component isotherm fitting parameters were used for calculating ideal adsorbed solution theory (IAST) binary-gas adsorption selectivity (Myers and Prausnitz 1965), defined as the Eq. (4):

$${S} = \frac{{{q}_{\text{1}} /{q}_{2} }}{{p_{1} /p_{2} }}$$
(4)

where p i represents partial pressure of gas i and q i is the corresponding pure-component gas uptake amount at p i . The IAST calculations were carried out for flue gas model (binary mixture containing 15% CO2 and 85% N2).

3 Results and discussions

3.1 General characteristics

FC-POP was synthesized in good yield via a simple Friedel–Crafts reaction. Triptycene was selected to be the monomer because of its rigid molecular structure, which would lead to a microporous skeleton with the extraordinary porosity and high stability. The resulting FC-POP was found to be insoluble in common organic solvents, such as water, methanol, dichloromethane, tetrahydrofuran, acetone and hexane. The chemical structure of FC-POP was confirmed by 13C cross-polarization magic-angle spinning (CP/MAS) NMR (Fig. 1). In the NMR spectrum, the resonance peaks at 139 and 130 ppm correspond to the substituted aromatic carbons of triptycene, while the resonance peak at 120 correspond to their non-substitute counterparts. Resonance peaks at 46–50 belong to the bridgehead carbons of triptycene. The peaks around 32 ppm ascribes to the carbons of methylene linkers. Besides, high thermo-stability of FC-POP was proved by thermogravimetric analysis (TGA) that little apparent weight loss below 350 °C (Fig. 2a).

Fig. 1
figure 1

Solid State 13C cross-polarization magic-angle spinning (CP/MAS) NMR of FC-POP and FC-POP–CH2Cl

Full size image
Fig. 2
figure 2

Thermogravimetric analysis (TGA) (a) and Fourier Transform infrared spectroscopy (FTIR) (b) of FC-POP, FC-POP–CH2Cl, FC-POP–CH2NH2, FC-POP–CH2DETA and FC-POP–CH2PEI

Full size image

In order to modify the microenvironment, FC-POP was covalently incorporated by various moieties to obtain a series of amine-functionalized FC-POPs (Scheme 1). FC-POP–CH2Cl, which containing vivacious chloromethyl groups, was first been prepared as an intermediate and chloromethyl groups were then substituted by various amines to afford FC-POP–CH2NH2, FC-POP–CH2DETA and FC-POP–CH2PEI. The peaks at 75–93 ppm in the solid-state NMR spectrum of FC-POP–CH2Cl suggested the graft of chloromethyl to the polymeric skeleton (Fig. 1). The appearance of peaks around 200 ppm for the spectrum of FC-POP–CH2Cl ascribed to the carbonyl group of residual acetic acid and paraformaldehyde that involved in the synthesis. Successful incorporation of functional amine groups were confirmed by Fourier Transform infrared spectroscopy (FTIR) that prosperous graft of amine groups can be proved by enhancement of the band around 1650 cm−1 (Fig. 2b). The amine-modified POPs also exhibit quite good thermal stability according to the thermogravimetric analysis results (Fig. 2a).

To quantitatively describe the density of functional groups, element analysis was conducted for the polymers. The calculated density of nitrogen element (mmol/g) shows a significant growth of nitrogen density in sequence of FC-POP–CH2NH2 (3.46 mmol/g), FC-POP–CH2DETA (8.43 mmol/g) and FC-POP–CH2PEI (11.47 mmol/g) (Table 1). Therefore, the tailor-made functional group would make the nitrogen-site density controllable in the microenvironment of these POPs.

Table 1 Elemental analysis and calculated N-site density for FC-POP, FC-POP–CH2Cl, FC-POP–CH2NH2, FC-POP–CH2DETA and FC-POP–CH2PEI
Full size table

Porosity of these materials was measured by nitrogen adsorption at 77 K (Fig. 3a) and the surface areas were calculated by the BET model. FC-POP exhibits extraordinary porosity, with the BET surface area as high as 1540 m2/g. The adsorption/desorption isotherms for FC-POP is not closed, mainly because of a swelling of polymer matrix at 77 K (Zhang et al. 2012). After the incorporation of functional groups, the BET surface area dropped correspondingly. The S BET data drops to 939 m2/g for FC-POP–CH2NH2, 636 m2/g for FC-POP–CH2DETA and only 129 m2/g for FC-POP–CH2PEI. For the adsorption isotherms of these FC-POPs, swift climb at low pressure (p/p 0 < 0.001) is contributed by micropores of these FC-POPs, while hysteresis at higher pressure proves the existence of mesopores. Pore size distribution caculated by non-local density functional theory (NLDFT) method also confirmed the presence of both micropores and mesopores (Fig. 3b). The mesopores, facilitating the mass transfer during the gas uptake process, can act as the transport channels (Liao et al. 2014). On the other hand, micropores, enhancing the interaction between the wall and CO2 molecules, can act as the adsorption active points (Islamoglu et al. 2013). As a result, both of them are crucial to CO2 adsorption applications. For FC-POP, V micro/V total value (the ratio of micropore volume to the total pore volume) is as low as 0.0842, demonstrating the existance of impressive mesopores. After grafting amines into the network, the microenvironment of pore structure has altered. For example, V micro value of FC-POP–CH2NH2 (0.251 cm3/g) doubles that of FC-POP (0.113 cm3/g), and its V micro/V total value upsurges to 0.311. Accordingly, the incorporation of functional groups may spatially separate mesopores into several micropores, result in the modified microenvironment of these FC-POPs through the redistribution of pore volumes, making the resulting POPs more suitable for CO2 capture applications.

Fig. 3
figure 3

a Nitrogen adsorption isotherms at 77 K, solid symbols for adsorption curves and hollow symbols for desorption ones; b pore size distribution calculated by NLDFT method

Full size image

3.2 CO2 adsorption performance

Excellent porosity and affluent nitrogen density are beneficial to boost CO2 adsorption capacity for these amine-modified POPs. Indeed, significant growth of CO2 uptake was observed after amine-grafting (Fig. 4). For FC-POP–CH2NH2 and FC-POP–CH2DETA, under the pressure of 1 bar, CO2 adsorption capacity reaches 4.2/2.5 and 4.5/3.4 mmol/g at 273/298 K, much higher than that of pristine FC-POP (3.1/1.8 mmol/g at 273/298 K). Notably, FC-POP–CH2PEI also exhibits impressive CO2 storage ability (3.2/2.6 mmol/g at 273/298 K) in spite of its poor BET surface area. There are two dominant contributions for CO2 adsorption capacity, the chemical adsorption contributed by amine groups (denoted as the nitrogen-site density) and the physical adsorption contributed by porosity (denoted as the BET surface area). For these FC-POPs with diverse incorporated amines, the hysteresis loops between their adsorption/desorption isotherms could be clear evidences of chemical interactions between amine groups and CO2 molecules. However, in our work, the growth of the nitrogen site density can significantly increase the CO2 adsorption enthalpy but decrease the BET surface area (Table 2). FC-POP–CH2DETA exhibited the excellent CO2 adsorption enthalpy (54.0 kJ/mol) yet maintained considerable BET surface area (636 m2/g). That means, FC-POP–CH2DETA reached a compromise between these two factors, hence its high CO2 adsorption capacity.

Fig. 4
figure 4

CO2 adsorption isotherms measured at 273 K (a) and 298 K (b), solid symbols for adsorption curves and hollow symbols for desorption ones

Full size image
Table 2 Characteristics of POPs in this work
Full size table

To further understand the influence of various functional groups on adsorption behavior, CO2 adsorption enthalpies were calculated by the dual-site Langmuir fitting of CO2 adsorption isotherms at 273 and 298 K based on Clausius–Claperyron equation. Figure 5a illustrates a plot of the CO2 adsorption enthalpies as a function of loading. It is obvious that FC-POP–CH2NH2, FC-POP–CH2DETA and For FC-POP–CH2PEI exhibit dramatically high CO2 adsorption enthalpy values as 37.9, 54.0 and 59.6 kJ/mol at low loadings, respectively; much higher than that of FC-POP (26.4 kJ/mol). The high adsorption enthalpy values ascribe to strong interactions between amine-functional groups in FC-POPs and CO2 molecules, for the presence of amines gives rise to chemical adsorption towards CO2. It is worth mentioned that the adsorption enthalpy value of FC-POP–CH2DETA remains a high value even when the CO2 loading reaches 1.5 mmol/g, suggesting its abundant chemical adsorption sites, hence the extraordinary CO2 adsorption capacity. The adsorption enthalpy values of amine-functionalized POPs show a similar sequence of nitrogen content, so that the nitrogen-site density are the dominant contribution for the strong affinity to CO2 in the microenvironment of these FC-POPs.

Fig. 5
figure 5

a CO2 adsorption enthalpy for POPs; b The selectivity of POPs for CO2 over N2 isotherms obtained from IAST calculations; c Cyclic CO2 adsorption for POPs on an ASAP 2020

Full size image

The excellent CO2 uptake performance and outstanding adsorptive affinity of these FC-POPs ensure their adsorption selectivity. The ideal adsorption solution theory (IAST) proposed by Myers and Prausnitz (1965) was used to calculate the adsorption selectivity of CO2/N2 for these materials. Fits of pure component isotherms were plotted to determine the molar uptakes at specified partial pressures in the bulk gas mixture and the selectivity was calculated under the simulated flue gas model (CO2/N2 = 15/85). The results suggested that the presence of incorporated amines in FC-POPs prompts the CO2/N2 selectivity (Fig. 5b). The selectivity value of FC-POP–CH2NH2 (31.58) is about 2.5 times higher than that of FC-POP (12.62) at 298 K. Significantly, the CO2/N2 selectivity value of FC-POP–CH2DETA and FC-POP–CH2PEI soar to 167.8 and 736.1 (298 K), respectively, thanks to their advantageous microenvironment including the microporosity and outstanding nitrogen-site density within the pores. For FC-POP–CH2PEI, although CO2 uptake seems modest comparing to the reported values [MCTP-1 (Puthiaraj et al. 2015), 2.7 mmol/g, 298 K, BILP-4 (Rabbani and El-Kaderi 2012), 3.59 mmol/g, 298 K; FCTF-1-600 (Zhao et al. 2013), 3.41 mmol/g, 273 K], the selectivity value is comparable to many other excellent reported POPs [azo-COP-2 (Patel et al. 2013), 130.6, 298 K; Py-1 (Luo et al. 2012), 117, 273 K; PPN-6-CH2DETA (Lu et al. 2012), 442, 295 K] (Table 3). The excellent selectivity of FC-POP–CH2PEI ascribe to the strong adsorption enthalpy towards CO2 (59.6 kJ/mol). Furthermore, the introduction of PEI moieties can effectively enhance the selectivity, because flexible PEI can block the pores of POPs to interfere the N2 adsorption while CO2 can infiltrate into the void of PEI chains and even swell the pores due to its high polarizability and quadrupole moment (Sung and Suh 2014). Although recently reported PEI (40 wt%) ⊂ PAF-5 (Sung and Suh 2014) exhibited even higher selectivity (1200, 298 K) than FC-POP–CH2PEI, the latter still stands out in regards to the physicochemical stability arising from covalent bonding amine to the network. Overall, the selectivity value relies on pore structure and CO2 adsorption enthalpies of polymers, so the adjustment of task-specific microenvironment would be an effective approach to improve CO2/N2 selectivity.

Table 3 Summary of surface area, CO2 uptake, selectivity and adsorption enthalpy (Qst) in selected POPs (with excellent reported results)
Full size table

To test the regeneration of the FC-POPs, cycling experiments were conducted on an ASAP 2020 analyzer (Fig. 5c). For each cycle, the adsorbents were saturated with CO2 up to 1 bar at 298 K followed by a high vacuum (0.2 Pa) at 80 °C for 90 min. The CO2 adsorption capacity in each cycle almost reached the same point, with no apparent loss of uptake amount in the cycling test.

4 Conclusions

A highly porous FC-POP, which was facilely synthesized via Freidel–Crafts reaction, was incorporated by various amines to afford FC-POP–CH2NH2, FC-POP–CH2DETA and FC-POP–CH2PEI. Both microporosity and excellent amine-affinity towards CO2 ensured those amine-functionalized polymers extraordinary CO2 selective adsorption performance. The CO2 adsorption capacity reached as high as 4.5 mmol/g at 273 K and 3.4 mmol/g at 298 K (1 bar) for FC-POP–CH2DETA; and the CO2/N2 selectivity soared to 736.1 at 298 K (1 bar) for FC-POP–CH2PEI. The reason is, that the incorporation of task-specific functional groups would work as a dual regulations for the microenviroment of POPs. The dual regulations, namely the modifications of pore structure as well as the basic amine density, improved microporosity and adsorption enthalpy of CO2 apparently and further enhanced the CO2 capture performance.