Rupa Shantamal Madyal and
Jyotsna Sudhir Arora*
Department of Chemical Engineering, Institute of Chemical Technology, Nathalal Parekh Marg, Matunga, Mumbai 400019, India. E-mail: josharora@gmail.com; Fax: +91-022-33611020; Tel: +91-022-33612013
First published on 23rd April 2014
The interaction of carbon dioxide with four amine functionalized polystyrene based adsorbents was investigated by means of density functional theory (DFT). The structures of the adsorbents and CO2–adsorbent complexes were optimized using the B3LYP/6-311++G(d,p) method. The interaction energies, equilibrium distances, charge transfer from the functionalized adsorbent to CO2, highest occupied molecular orbital (HOMO)–lowest unoccupied molecular orbital (LUMO) energy gap and thermochemical parameters of the complexes were evaluated. Weak intermolecular forces were responsible for the interaction between the active centers of functionalized adsorbents and CO2 as confirmed from experimental studies. The interaction energies and vibrational frequencies confirmed stronger interaction of CO2 towards imidazole functionalized polystyrene (PS-imidazole) followed by N-methylpiperazine loaded polystyrene (PS-piperazine) and were found to be the least for the bare adsorbent chloromethylated polystyrene (CMPS). Steric hindrance was found to play a major role in the case of dimethylamine (PS-DMA) and diethanolamine (PS-DEA) functionalized polystyrene during their interaction with CO2. On the basis of the HOMO–LUMO energy gap and calculated density of states, a negligible change in the electronic properties of CO2–adsorbent complexes was observed, indicating a physisorption process. The outcomes of the present theoretical studies are in good agreement with the experimental results and provide detailed insights for understanding the interaction of CO2 with the active center of the functionalized adsorbent.
CO2 capture by liquid amines (primary and secondary) are amongst the most commonly used processes.1,4–6 However, the main drawbacks of the solvent based CO2 absorption are high energy consumption for solvent regeneration accompanied by degradation of the amine leading to generation of waste, thus resulting in the capital loss.1,4,7 Hence, adsorption is considered to be an alternative promising technology for separating CO2 from gaseous mixtures due to its low energy requirement, easy operation and low maintenance.8 It involves binding of a molecule onto the solid surface by strong chemical or weak physical interactions or at times even both and adsorbent regeneration is carried out by either pressure or temperature swing cycles.8–15 However, the development of low cost adsorbent with high CO2 selectivity and adequate adsorption capacity under ambient conditions is an active area of research.16 The solid sorbents such as activated carbon,17 zeolites,18 zeolitic imidazole frameworks (ZIFs),19 single/multi-walled carbon nanotubes,1 nanoporous silica-based molecular basket,20 polymers21–25 and metal–organic frameworks (MOFs)1,7,10 have been used for adsorption of CO2. The recent studies have focused on use of amine functionalized adsorbents for selective separation of CO2 from flue gas mixtures.7,17,20,22–24,26–28 The incorporation of amine functionality on the surface of the adsorbent significantly increases CO2 selectivity and uptake capacity because CO2 is an acidic gas.2
As the quest for the development of CO2-specific adsorbent is continuously increasing, the use of computational studies has become an indispensable tool for understanding the parameters which govern CO2–adsorbent interactions at the molecular level. Panda et al. performed computational studies with a series of ZIFs using different 4,5-functionalized imidazole units, viz., –CH3, –OH, –Cl, –CN, –CHO and –NH2 by applying Universal Force Field (UFF).19 The symmetry and polarizability of the functional group exhibited significant influence on CO2 uptake. MOF,1,7,10 CNT1 and zeolites18 have also been investigated to predict the selective CO2 adsorption using ab initio calculations. Thakur et al. theoretically studied the adsorption of CO2 on graphene surface and reported π interaction of CO2 with the electron rich benzene ring.29 Hussain et al. performed ab initio calculations to study the affinity of CO2 towards 20 naturally occurring amino acids and also proposed amino acids as alternatives to chemical dissolution of CO2.30 Baei et al. theoretically studied the interaction of CO2 with B12N12 nanoclusters using DFT calculations and reported interaction energy of −2.30 kcal mol−1 on physisorption.12 Valenzano et al. reported 1:1 adsorption complex of CO, N2, and CO2 with the Mg2+ adsorption site of Mg-MOF-74 by weak (dispersion) interactions.11
Extensive efforts are being directed from a variety of fields including polymer technology to search better CO2 capturing materials.21–25,31–33 The surface properties of the polymer can easily be tailored with different organic or inorganic ligands. After functionalization of the polymer surface, the accessibility of the amine to capture CO2 is the most important factor to be considered during adsorption. Primary22–24 and secondary amine22,23 functionalized polymeric adsorbents have already been used for CO2 capture which result in the formation of carbamate species. The regeneration of such adsorbents involve energy intensive temperature swing process.22,24,34 However, tertiary amines (in the absence of moisture) do not support formation of carbamate and thus bind CO2 by physical interactions only.28 Recently, Khot et al. carried out CO2 sorption studies with tertiary amine functionalized polystyrene as well as with bare polystyrene.8 The amine functionality is covalently attached to the macroporous surface of polymeric adsorbent after the bead formation. Hence, the surfaces of the pores are expected to be available for adsorption and a molecule like CO2 diffuses through the pores of the adsorbent beads. Further, these amines are neither mobile nor there will be any loss of functionality from the surface during the adsorption and desorption process.
The current work includes theoretical studies of these tertiary amine functionalized adsorbents for interaction with CO2. Since quantum mechanical calculations involving the entire polymeric structure would be difficult and computationally expensive, only the amino functional group covalently attached to the repeating structure of the polymer viz. styrene moiety, was taken for the current study. The calculations were carried out with the objective to gain fundamental insights into equilibrium geometry of the adsorbents and the complexes, interaction energy, charge transfer, band gap, vibrational frequencies and understand the dominant factors responsible for CO2 complexation. This study is expected to be useful for future designing of CO2 specific adsorbents.
Corrections to the interaction energy for all the complexes were evaluated by considering the basis set superposition error (BSSE) with counterpoise correction (CPC) calculation.4,11 Natural population analysis using the Natural Bond Order (NBO) method was performed.4 The charge transfer (QT) from the donor atoms of the amine loaded adsorbents to CO2 during the interaction was evaluated by summing up the electrostatic potential (ESP) charges on CO2 molecule.30 To compare the calculated vibrational frequencies with the experimental values, the calculated frequencies were scaled using a factor of 0.964 ± 0.023.37 The contribution of a group to a molecular orbital was calculated by Mulliken population analysis (MPA). The density of states (DOS) spectra were created by convoluting the molecular orbital information, allowing identification of amine and CO2 centered orbitals with Gaussian curves of full width at half maximum (FWHM) = 0.3 eV. The DOS diagrams for all the adsorbents and the complexes presented in this work were obtained using the Multiwfn program.38
ΔEnoCPC = TEcomplex − (TEadsorbent + TECO2) | (1) |
ΔECPC,corr = ΔEnoCPC + BSSE | (2) |
Fig. 2 Optimized structures of PS-imidazole + CO2 (A), PS-piperazine + CO2 (B), PS-DMA + CO2 (C), PS-DEA + CO2 (D), and CMPS + CO2 (E) complexes. The calculated values of d1, d2, d3 and θ are given in Table 2. |
Adsorbents | ΔEnoCPC | ΔEBSSE | ΔECPC,corr | Δ(ZPE) | ΔET/R/V | TΔS | ΔG | ΔGa |
---|---|---|---|---|---|---|---|---|
a Interaction between Lewis base, pentafluoro phenolate anion and CO2, ΔE = −23.1 kJ mol−1 and the distance between the oxygen of the anion and ‘C’ of CO2, 2.43 Å (Teague et al.).4 | ||||||||
PS-imidazole | −12.80 | 0.95 | −11.85 | 1.67 | 5.74 | −24.18 | 17.25 | 12.7 |
PS-piperazine | −8.24 | 1.58 | −6.66 | 2.41 | 5.94 | −28.56 | 30.25 | |
PS-DMA | −6.57 | 1.49 | −5.08 | 2.78 | 6.02 | −30.38 | 34.10 | |
PS-DEA | −4.22 | 1.55 | −2.67 | 3.13 | 6.18 | −30.56 | 37.19 | |
CMPS | −3.56 | 1.51 | −2.05 | 3.24 | 6.29 | −30.77 | 38.26 |
The smaller values of ΔECPC,corr implies that CO2 experiences weak van der Waals interaction with the adsorbent, thus it is identified as physical in nature.2,40 The thermal and entropic contributions were considered to calculate the free energy of the system which is given by eqn (3),4
ΔG = ΔECPC,corr + Δ(ZPE) + ΔET/R/V – RT − TΔS | (3) |
The enthalpic contributions to ΔE were indicative of the specific interactions, but the entropic factors seem to negate the advantage of specific interactions. The ΔG values for the functionalized and the bare adsorbent is positive indicating that the interaction between the adsorbent and CO2 is weak as the complexes are destabilized by entropic effects and hence the gas phase reaction is unfavorable at 303 K and 1 bar. A positive change in free energy, however, should not be taken as a sufficient reason for not pursuing a potentially useful reaction involving CO2, the kinetics might indeed be favorable. Hence to achieve significant adsorption, increasing the pressure and/or lowering the temperature may enhance the binding of CO2 with amine functionality. The optimized parameters (ΔE and distances between the interacting centers) from the current study are comparable to the positive free energy results reported by Teague et al.4 (Table 1).
The experimental investigations of CO2 adsorption on these four functionalized adsorbents executed by Khot et al.41 confirmed the physical nature of adsorption. Also, the IR frequencies of the adsorbents after CO2 adsorption showed no changes in the C–N stretching frequency towards 1609 cm−1 and the asymmetric stretching (ν3) of CO2 (2349 cm−1), indicating no carbamate and bicarbonate formation.8 An improved adsorption of CO2 on the amine functionalized adsorbents was observed by Khot et al.8 during the experimental studies performed by increasing the pressure (∼40 bar). Thus, the above methodology correctly predicts the presence of physisorbed CO2 on amine functionalized polystyrene.
The observed trend in interaction energies can be well correlated to the distance between CCO2–donor atom of the functionalized adsorbent (d1). PS-imidazole which exhibits the strongest interaction with CO2 has the lowest CCO2–Namine distance (2.8533 Å), followed by PS-piperazine (2.9933 Å) and PS-DMA (3.1346 Å). This is because the planar structure of imidazole experiences less steric hindrance allowing the sp2 nitrogen atom to move closer to the ‘C’ of CO2 as compared to PS-piperazine and PS-DMA. In the case of PS-DEA, due to the inaccessibility of the ‘N’, electron deficient ‘C’ of CO2 interacts with the electronegative –OH groups of DEA. The ‘O’ atoms of CO2 strongly interact with the electropositive ‘H’ atoms of the alcoholic group thus leading to the formation of a hydrogen bonded structure.42 However with CMPS, CO2 exhibits π-quadrupolar interaction with the π-electron rich phenyl ring. In CMPS–CO2 complex, CO2 is parallel to the ring but slightly displaced from the centre of mass of phenyl ring.25,43
The interaction between CO2 and the donor atom of the amine functionality leads to varying degrees of distortion in carbon dioxide molecule as well as the interacting centers of the adsorbents. The calculated bond length of free CO2 is 1.1608 Å, which is comparable to the experimental bond length (1.161 Å).44 On analyzing the structure of adsorbed CO2, we see lengthening of the CO bond which is closer to the active center (d2 − d0) in all systems when compared to the free CO2 bond length (1.1608 Å) while, the CO bond farther away from the active center (d3) is shortened in all the adsorbents (d3 − d0). Summing up the two CO bond lengths (d2 and d3), yields the overall length (l) of the CO2 molecule, which can be used in determining elongation of the molecule (l − l0) after interaction, for all the adsorbents. PS-imidazole and PS-piperazine display a higher increase in bond length of physisorbed CO2 as compared to that with PS-DMA and PS-DEA (Table 2) owing to the strong interaction with CO2. Also, the bond lengths of amine functional group changes during the complexation. The Camine–donor atom bond lengths before and after complexation are listed in Table 3.
Adsorbents | d1 | d2 | d3 | (d2 − d0) × 103 | (d3 − d0) × 103 | (l − l0) × 103 | θ (deg) | QT |
---|---|---|---|---|---|---|---|---|
a All distances are given in Å. d0 = 1.1608 Å (CO bond length of unadsorbed CO2). d2 and d3 are CO bond length of physisorbed CO2. l0 = 2 × d0 = 2.3216 Å (bond length of CO2 before interaction). l = d2 + d3 (bond lengths of CO2 after interaction). | ||||||||
CMPS | 1.1615 | 1.1602 | 0.72 | −0.60 | 0.12 | 179.54 | ||
PS-imidazole | 2.8533 | 1.162 | 1.1607 | 1.47 | −0.01 | 1.46 | 176.05 | −0.009 |
PS-piperazine | 2.9933 | 1.1616 | 1.1615 | 0.80 | 0.75 | 1.55 | 176.74 | −0.012 |
PS-DMA | 3.1346 | 1.1613 | 1.1609 | 0.55 | 0.15 | 0.70 | 177.53 | −0.005 |
PS-DEA | 2.9013, 2.2023, 2.7061 | 1.1616 | 1.1599 | 0.84 | −0.90 | −0.06 | 178.06 | −0.011 |
Adsorbents | Bonds | Before complexation | After complexation |
---|---|---|---|
PS-imidazole | C(11)–N(12) | 1.3137 | 1.3157 |
PS-piperazine | C(20)–N(24) | 1.4623 | 1.4662 |
C(21)–N(24) | 1.4638 | 1.4669 | |
C(25)–N(24) | 1.4539 | 1.4571 | |
PS-DMA | C(9)–N(10) | 1.4635 | 1.4706 |
C(11)–N(10) | 1.4581 | 1.4611 | |
C(12)–N(10) | 1.4565 | 1.4601 | |
PS-DEA | O(15)–H(35) | 0.9618 | 0.9626 |
O(16)–H(36) | 0.9618 | 0.9641 | |
C(12)–O(15) | 1.4387 | 1.4413 | |
C(14)–O(16) | 1.4251 | 1.4275 |
Further, CO2 molecule undergoes nonlinear distortion, and the calculated deviation of ∠O–C–O from its linear geometry (180°) in the adsorbed state is relatively small, ∼1–5° (Table 2). The extent of bending of CO2 during complexation is generally considered to be a measure of the basicity of the amine.4 Amongst the amine functionalized adsorbents, PS-imidazole and PS-piperazine reveal maximum change in the ∠O–C–O bond angle as compared to PS-DMA and PS-DEA, which justifies superior contact with the former adsorbents. CMPS which shows the least interaction with CO2 at atmospheric pressure exhibits the lowest bent of ∠O–C–O (179.54°) as compared to the functionalized adsorbents (Table 2). Thus, the extent of distortion of CO2 from the linear geometry and shorter CCO2–donor atom distances (d1) can be correlated to stronger CO2–adsorbent interactions.
Adsorbents | Bonds | Before (cm−1) | After (cm−1) | |||||
---|---|---|---|---|---|---|---|---|
C-donor atom | ν3a | ν3b | ν3c | Δν3 | ||||
Exp. | Cal. | |||||||
a Calculated frequency of CO.b Calculated frequency of CO multiplied with the scaling factor (0.964).c Experimental frequency.8d Simulated frequency of CO2.46 | ||||||||
CO2 | CO | 2420 | 2333 | 2349d | — | — | ||
PS-imidazole | C(11)–N(12) | 1535 | 1537 | 2413 | 2326 | 2335 | 14 | 7 |
PS-piperazine | C(20)–N(24) | 1219 | 1221 | 2411 | 2324 | 2336 | 13 | 9 |
PS-DMA | C(9)–N(10) | 1115 | 1116 | 2415 | 2328 | 2348 | 1 | 5 |
PS-DEA | O(15)–H(35), C(12)–O(15) | 3315, 1039 | 3345, 1043 | 2420 | 2333 | 2348 | 1 | 0 |
The simulated frequency values are slightly higher than the experimental values which is due to the fact that the experimental values are anharmonic frequencies while the calculated values are harmonic frequencies.47 Hence, the calculated vibrational frequencies of CO2 (v3) are compared with experimentally determined FTIR frequency values8 of physisorbed CO2 after multiplying with the scaling factor (Table 4). Increase in the CO bond length of CO2 results in decrease in v3 (red shift).46 As mentioned earlier, PS-imidazole and PS-piperazine show a considerable increase in bond length of CO2 and thus result in the higher v3 shift as compared to PS-DMA and PS-DEA, respectively. A fairly consistent trend is observed between experimental and DFT calculated v3 values.8 The stretching frequency of C(11)N(12) bond in PS-imidazole exhibited a 2 cm−1 shift on complexation with CO2. Similarly, the C(20)–N(24) symmetrical stretching shifted from 1219 to 1221 cm−1 in PS-piperazine, while only 1 cm−1 shift in C(9)–N(10) symmetrical stretching was observed for PS-DMA adsorbent. However, in PS-DEA adsorbent the electron rich oxygen center (O(15)) polarizes more on complexation as indicated by increase in its symmetric stretching frequency (C(12)–O(15) exhibits 4 cm−1 while O(15)–H(35) reveals 30 cm−1 shift).
Band gap (Eb) = ELUMO − EHOMO | (4) |
On analysing the FMO's of the adsorbents, the HOMO of PS-imidazole is the most stable followed by PS-DMA > PS-DEA and least is for PS-piperazine. Further, the HOMO's of all the adsorbents are more stable as compared to the LUMO of CO2. Since CO2 is electrophilic in nature, the HOMO of the adsorbent reacts with the LUMO of CO2 molecule.49
The molecules which exhibit a large HOMO–LUMO gap are considered as hard molecules and are relatively more stable than soft molecules which have a small HOMO–LUMO gap.50 To understand the interactions taking place during the adsorption of CO2, the total and partial density of states (TDOS and PDOS, respectively) of the CO2–adsorbent systems were determined and were compared with the TDOS of adsorbents without CO2 (Fig. 3). During the interaction of CO2 with the functionalized adsorbent, the electronegative donor atoms of the adsorbent attract the electron deficient ‘C’ of CO2 increasing the band gap which leads to stabilization of system, except in the case of PS–DEA–CO2 complex (Table 5). For the PS–imidazole–CO2 complex (Fig. 3(a)), the red curve (PDOS of imidaole fragment) is high and nearly approaches the black line (TDOS of complex) in the region of −6.8 to −6.1 eV. Hence it can be concluded that imidazole moiety has significant contribution towards HOMO. In contrast, the LUMO is mostly localized on the styrene moiety as there is no such clear contribution of imidazole or CO2 fragments found in the PDOS diagram. However, for the remaining amine functionalized adsorbents, the % contribution of amine towards HOMO decreases indicating lesser overlap between the amine and electron deficient ‘C’ of CO2. The PDOS of CO2 in the LUMO for PS-piperazine (Fig. 3(b)) occurs at slightly higher energy than PS-DEA (Fig. 3(d)), PS-DMA (Fig. 3(c)) and CMPS (Fig. 3(e)). However, the difference is very small (0.01–0.02 eV). The HOMO's of the adsorbent–CO2 complexes are shown in the Fig. 4.
Fig. 4 HOMO's of CO2 with PS-imidazole (A), PS-piperazine (B), PS-DMA (C), PS-DEA (D), and CMPS (E) complexes. |
The stability of the complex is measured in terms of %ΔEg which is defined by eqn (5)
(5) |
The values of Eb,complex and %ΔEg are reported in Table 5. As the band gap Eb,complex increases after complexation, the ability of the adsorbent to attract CO2 increases, which indicates increased stability of the system.40 The above discussion is in trend with the values of %ΔEg which is highest for PS-imidazole and lowest for PS-DEA. Since the HOMO of all the complexes presents no contribution from interacting CO2 molecule, the interactions are considered as physisorption process.
The current work involves the use of specific amines which are loaded on polystyrene and are highly selective in its interaction with CO2. The quadrupole moment of CO2 (4.3 × 10−31 C m2) induces strong and specific Lewis acid–base type interactions with the functionalized adsorbents.8 In these complexes, the ‘C’ of CO2 behaves as an electron acceptor and the ‘N or O’ of the amine as the donor. The above studies indicate PS-imidazole to be a superior adsorbent for complexing with CO2 possessing ΔE of −11.85 kJ mol−1 and shortest CCO2–Namine distance (2.8533 Å). The objective of the current study was to understand the interaction and binding of CO2 with amine functional group at the molecular level. The above study highlights the molecular level information about charge transfer, FMO contribution and vibrational frequency analysis occurring during complexation. The results indicate that the CO2–amine interaction is mainly of physisorption augmented by hydrogen bond type chemical interactions. As a result, the CO2 is adsorbed reversibly on these adsorbents and its desorption can be facilitated by thermal/pressure swing for regeneration as studied by Khot et al.8 This study aims at understanding the interaction and binding types of CO2 with amine functionalized adsorbents which can provide useful guidance for the designing of newer ligands for CO2 selective adsorption.
This journal is © The Royal Society of Chemistry 2014 |