In the past decades, one of the greatest threats to human health is global warming due to cause a significant climate change . To mitigate CO2 emissions, carbon capture and sequestration (CCS) is considered to be one of the most promising solutions . Fossil fuel-burning power plants are the single largest anthropogenic sources globally among all CO2 emission sources, accounting for approximately one-third of CO2 emissions . For the capture of CO2, amine scrubbing is currently the mature technology, utilizing the chemical reactions between amines and CO2 . However, the regeneration of the amine solution needs the heating of the system and generates serious energy consumption . In recent years, solid adsorbent materials using chemisorption or physisorption have been intensively investigated such as carbons , zeolites , metal-organic frameworks (MOFs) [8,9], covalent-organic frameworks (COFs) , and porous polymers .
Porous carbon as gas sorbent has a long history due to has various advantages such as high surface area, low cost, high conductivity, fast kinetics, diverse availability, facile preparation strategies, outstanding stability, and low manufacturing cost. Polyaniline (PANI) is a conductive polymer with high nitrogen content. It has been intensively investigated in materials science due to its versatile applicability (e.g., sensors [12,13] or conducting wires ). These applications are based on their macroscopic properties like electrical conductivity or charge by redox processes . Another potential application of polyaniline concerns its conversion to activated carbons by physical activation or chemical activation with various activating agents such as KOH, ZnCl2, Na2CO3, K2CO3, NaHCO3, NaOH, and H3PO4 for CO2 capture due to its large N content. Among these, KOH is most frequently reported in the literature, because ACs activated by KOH possess well-defined micropore size distributions and ultrahigh specific surface areas . During KOH activation, some of the nitrogen groups reacted with KOH to facilitate KOH penetration into deeper layers of the carbon, leading to a greater development of the micropore structure, while other nitrogen groups were transformed into more thermally stable species and built into the carbon structure. The resulting porous carbons thus featured a high fraction of fine micropores (<1 nm) and some degree of basic nitrogen-containing groups.
Nitrogen doping is a very powerful tool for enriching the functionality of carbon materials. Nitrogen-doped carbons have been widely investigated in different fields . Promising results on their CO2 adsorption performances have also been reported, nitrogen doping exerts positive influences on both the CO2 uptake and the adsorption selectivity over other gases [18-20]. The preparation of nitrogen-doped carbon from nitrogen-containing precursor is thus of particular significance.
With this in mind, this study reports the preparation of microporous activated carbon prepared from polyaniline as a carbon source. The adsorption equilibrium data of pure CO2 on prepared adsorbent was measured volumetrically at different temperatures (298, 308 and 318K) and pressures ranging 1 to 10 bar. The equilibrium and thermodynamic study of the CO2 adsorption were studied to describe the adsorption processes. To the best of our knowledge, no study has been done on the production of NDC for CO2 capture from PANI that prepared with DBSNa as a dopant.
Carbon dioxide and nitrogen were employed with a purity of 99.99% and 99.995%, respectively. Hydrochloric acid 37% (HCl) and potassium hydroxide (KOH) were obtained from Scharlau. Aniline (AR grade), sodium dodecylbenzenesulfonate (DBSNa), ammonium persulphate (APS) and ethanol were purchased from Merck. Also, the water used throughout the work was distilled water.
Synthesis of the N-doped organic porous activated carbon from PANI
The preparation of nano polyaniline was explained in our previous work . For the preparation of PANI, the first aniline was distilled twice under atmospheric pressure to eliminate the oxidation impurities. In this procedure, 1g APS as an oxidant and 0.5 g DBSNa each was dissolved in 50 mL of 1M H2SO4 and were stirred for 30 min and then mixed together stirred for 15 min. Then, 1 mL freshly distilled aniline monomer was added to the mixture under vigorous stirring. The reaction vessel was placed in an ice bath during addition oxidant. The reaction mixture was left stirring for about 3 hours at low temperature (0–5 °C). Consequently, the precipitated polymer was filtered and washed with ethanol and distilled water in order to remove oligomers and other non-polymeric impurities until the washing liquid was colorless. The polymer was dried in a vacuum oven at 60 °C for 24 h.
Chemical activation of the PANI was done with impregnating the precursor with KOH. 1 g of the dried precursor was impregnated with 2g KOH and mixed at 80 °C for 3 h. After mixing, the slurry was placed in the vacuum drying at 110 ⁰C for 20 h. The impregnated precursor was placed in tubular quartz reactor and carbonized up to 800⁰C at a rate of 5 ⁰C/min in an electrical furnace for 1h under N2 ﬂow (100 ml/min STP). After pyrolysis, the activated carbon was cooled down to room temperature in a ﬂow of nitrogen and then removed from the reactor. In all experiments, nitrogen ﬂow rate and heating rate were kept constant. Pyrolysis under inert gas breaks down the cross-link between carbon atoms and avoids the dilution of gas products. The product was washed sequentially with 0.5 M HCl solution to remove the activating agent from the material. The NDC was separated from the solution by ﬁltration and was washed several times with hot and finally cold distilled water until the pH of the ﬁltrate reached around 6.0 and then dried overnight in vacuum oven. The preparation interaction of NDC is schematically illustrated in Fig. 1.
The chemical composition of raw PANI and prepared activated carbon were measured using a CHNS elemental analyzer (Costech ECS 4010, Italy). Functional groups present in the NDC was analyzed using Fourier transform infrared spectroscopy (Bruker Vertex-70, Germany). The samples were ground and then mixed with KBr, and the spectra were recorded in the range of 400–4000 cm-1. The textural properties of the prepared sorbents were determined from N2 adsorption isotherm at 77 K using an adsorption apparatus (BELSORP 28, BEL Inc., Japan). Before each measurement, the samples were degassed under a high vacuum at 393 K for 15 h in order to out-gas any remaining moisture or organic species. The specific surface area of the samples was calculated by the application of the Brunauer–Emmett–Teller (BET) method.
Gas adsorption measurement
The CO2 adsorption performance of the prepared sample was evaluated using volumetric method. The schematic diagram of volumetric apparatus is shown in Fig. 2. The apparatus consisted of two high-pressure stainless steel vessels including the pressure and adsorption cells. Both cells were placed into thermostatic water circulating bath to keep the temperature constant during CO2 adsorption. The high precision pressure gauge measured the changes in pressure in gas and adsorption cells in each CO2 adsorption experiment. Prior to CO2 adsorption experiments, activated carbon was degassed at 100 °C for about 24 h and the system was evacuated by vacuum pump. Helium gas was utilized as non-adsorbing gas to determine the dead volume. The CO2 adsorption experiments were conducted at pressures ranging from 0 to 10 bar at different temperatures (298–318 K). Using the SRK equation of state in MATLAB program the amount of CO2 adsorbed on NDC was calculated.
Adsorption isotherm simulation
The adsorption isotherm shows how the gas molecules distribute between the solid phase and the gas phase when the adsorption process reaches an equilibrium state. An important step to ﬁnd the appropriate model that can be used for design purposes is an analysis of the experimental isotherm data by ﬁtting them to different isotherm models.
The Langmuir model is based on the assumption that the adsorption occurs at homogeneous sites at the adsorbent surface and adsorption energy is constant and independent of surface coverage. When the surface is covered by a monolayer of adsorbate, the maximum adsorption occurs. The Langmuir model is the most commonly employed equilibrium relation and given by :
where q is the amount of CO2 adsorbed per unit mass of adsorbent (mmol.g-1); qm is the maximum CO2 adsorption capacity (mmol.g-1) and b is b=ka/kb, ka and kb are adsorption and desorption constants, respectively.
The empirical Freundlich equation is derived by assuming that there is an exponential variation in site energies of adsorbent and also surface adsorption is not rated limiting step. Freundlich isotherm can be expressed by :
The constant KF is the adsorption capacity of the adsorbent and n is a measure of the deviation of the model from linearity of the adsorption and indicates sorption intensity. For n equal to unity, the adsorption is linear. If the value is below unity, this confirms that the adsorption process is chemical and the surface is relatively homogeneous. For the value greater than unity, adsorption is a physical process and the sorbent is relatively heterogeneous. In a recent case, adsorbate is favorably adsorbed on the adsorbent. The higher n value shows a stronger adsorption intensity [24, 25].
Sips or Langmuir–Freundlich isotherm  can be considered as a combination of the Langmuir and Freundlich isotherm models and describe the heterogeneity of the surface adsorbent. This isotherm model is expressed as follow:
where qm, b, and n were previously described. As mentioned before, for the values of 1/n less than one, the adsorption process is heterogeneous whereas values closer to or even one confirms that there are relatively more homogeneous binding sites in the adsorbent .
Thermodynamic of adsorption
Information about the isosteric heat of adsorption as a critical design variable in estimating the performance of an adsorptive gas separation process is necessary for all industrial adsorption processes. The heat of adsorption expresses the changing of enthalpy before and after adsorption between gas molecules and the surface of adsorbent.
The isosteric heat of adsorption of a gas at a specific adsorbate loading can be estimated from the temperature dependence of the isotherm using the Clausius–Clapeyron equation as follows:
where q is the constant adsorbate loading. The isosteric heat of adsorption of CO2 onto adsorbents can be derived from the slope of the lines in the plots of lnP versus 1/T at a constant adsorbed amount.
RESULTS AND DISCUSSION
SEM imaging was conducted to reveal the evolution of the morphology and surface from NDC. The morphology of nano polyaniline and NDC is shown in Fig. 3. As can be observed from Fig. 3(a), PANI synthesized by rapidly-mixing the solution of oxidant, monomer, and surfactant had thin nanofibrillar morphology with diﬀerent lengths and diameters. Large pores with different shapes and sizes are randomly distributed in the particle of adsorbent. This observation is associated with the chemical reaction between KOH and the carbon taking place during the activation process, this reaction playing an important role in the development of the porosity in the activated carbon .
Fig. 4 shows the N2 adsorption-desorption isotherms of the NDC prepared from PANI. It can be seen from Fig. 4 that the NDC demonstrates the type I isotherm according to the Brunauer classiﬁcation, indicating that there are mainly micropores and a relatively small external surface area in the sample. The most uptake of N2 adsorption was determined in at low relative pressures region, which is considered due to the formation of highly microporous material with narrow pore-size distribution and monolayer adsorption on the NDC. The textural parameters of the prepared sample were shown in Table 1.
The presence of the nitrogen-containing groups can be conﬁrmed by FTIR spectroscopy. Fig.5(a) shows the FTIR spectra of PANI, recorded using KBr pellet. The peaks observed at 1118 and 1295 cm-1 are assigned to the C-N stretching vibration. The characteristic peaks at 1467, 1562 and 1650cm-1 are assigned to the C=C in the benzene ring and the quinoid ring. The adsorption at 3214 cm-1 is due to the N–H stretching vibration. FTIR absorbance spectrum of NDC shows peaks at 1057 cm-1 for C−N stretching vibration. The band at 1558 can be identiﬁed as originating from the N-H in-plane deformation vibration. The weak bands at 2854 and 2923 cm-1 correspond to C-H bond, and the weak band of 1651 cm-1 is attributed to the distinctive absorbance of C-H bonds of benzene rings as well as the C=N bonds from the carbon framework. The peak at 1700 cm-1 corresponds to the–C=O stretching vibration. The peak at about 3444 cm-1 assigned to the N-H symmetric stretching vibration of –OH stretching vibration [29-33]. The FTIR analysis, therefore, confirms the existence of N-H and C-N species in the carbon sample.
Fig. 6 demonstrates the equilibrium adsorption isotherms of CO2 on the prepared activated carbon from PANI. Based on the Brunauer classiﬁcation, all of the CO2 adsorption isotherms are Type I. This phenomenon is predictable for adsorption of small adsorbates such as CO2 molecules on a microporous adsorbent. Also, it exhibits a steep increase at low pressure due to existence vacancies or sites available for adsorption. At higher pressure, the amount of gas adsorbed amounts increases slightly because at high pressure most of the sites are occupied. Results demonstrated that the adsorption capacity of CO2 decreased with increasing the temperature from 298K to 318K, indicating the exothermic nature of the adsorption. In fact, as temperature increased, the physical bonding between the gas molecules (including CO2) and the active sites of the adsorbent weakened. Therefore, gas molecules adsorbed on the AC surface obtain enough energy to overcome the van der Waals forces and go back to the gas phase. Consequently, the sorption values decreased with temperature increment.
The experimental equilibrium data for CO2 adsorption on NDC were ﬁtted with Langmuir, Freundlich and Sips isotherm to determine which model gives the best correlation to experimental data. These isotherm models were ﬁtted to experimental data and presented in Fig.6. The parameter values of the three isotherm equations are summarized in Table 2. According to Table 2, the amount of qm, b, and k for all adsorbents was decreased by increasing the temperature from 298 to 318 K due to existing an inverse relationship between CO2 adsorption and temperature which approve an exothermic nature of the adsorption process. The Freundlich exponent (n) for AC was more than 1, indicating that CO2 was favorably adsorbed onto this adsorbent at all temperatures. Based on the regression (R2) coefﬁcients, the Sips model had the best ﬁt with the experimental data, suggesting the heterogeneous surface of AC . The value of qmax obtained from the Langmuir model was slightly higher than that obtained with the Sips model for all adsorbents.
CO2 adsorption mechanism
The improvement in the adsorption capacity of the NDC can be attributed to the chemisorption interaction of CO2 with amine groups. The reactions occurring between the N-doped AC are basically the same as that between the liquid amine and the CO2 gas. The overall reaction between the CO2 molecule and the secondary amine is formation of carbamate. The proposed reaction can be expressed as follows :
As can be seen from the reaction, in order to form the carbamate, as the ﬁnal product, two nitrogen atoms from the amine groups (close enough to each other) are needed. Base on the Dankwerts’ zwitterion mechanism the reaction includes a formation of a zwitterion intermediate. Actually, the above reaction is constituted from two reactions:
Breaking of amine N–H bond and trapping of a proton by another amine group cause to form carbamate. Hence, the process depends on the ability of the amine to split/bind the proton.
Thermodynamic of adsorption
Determination of the heat of adsorption and its variation with coverage can give practical information about surface characteristics and the adsorbed phase. Surface characteristics are measurable by determining the heat of adsorption released from the interaction of adsorbate with the adsorbent at a certain loading.
The isosteric heat of adsorption of CO2 onto NDC was obtained from the slope of the lines of ln P versus 1/T plot at a constant adsorbed amount. The variation of isosteric heat of adsorption with surface loading was illustrated in Fig. 7. The negative value of the isosteric heat of adsorption suggests the exothermic nature of the process of CO2 adsorption by adsorbents. The heat of adsorption is interpreted as an indicator of the interaction strength between the gas molecules and the adsorbent. For heat of adsorption of 80 kJ/mole or more, the adsorption process is chemisorption and smaller values are representative of physisorption . According to the ideal Langmuir model, the heat of adsorption is independent of surface coverage. However, this necessity is seldom satisﬁed in actual systems due to the surface heterogeneity and strong eﬀects of adsorbate - adsorbate interaction.
At lower surface coverage (from 0.2 mmol.g-1 up to 0.6 mmol.g-1), the isosteric heat of adsorption of CO2 exhibits almost constant, thereby indicating a homogeneous adsorption system and there is no lateral interaction between the adsorbed CO2 molecules. By increasing CO2 loading, (from 0.8 mmol.g-1 up to 3 mmol.g-1), the isosteric heat of adsorption of CO2 decreased. The weak interaction between CO2 and adsorbent occurs. Since the strong adsorption sites are occupied ﬁrst and upon increased loading, the weaker adsorption sites will be occupied. Hence, the heat of adsorption decreases. With increasing CO2 loading (from 3.5 mmol.g-1 up to 4.5 mmol.g-1), the isosteric heat increases slightly due to the adsorbate-adsorbate interactions at higher pressure. By increasing CO2 loading, (from 5mmol.g-1 up to 6 mmol.g-1), the isosteric heat of adsorption of CO2 decreased. In the present study, the values of isosteric heat of adsorption obtained for NDC were lower than 80 kJ.mol-1 indicating that the interaction between NDC and CO2 molecules is mainly physical interactions involving weak van der Waals and electrostatic interactions.
From an industrial point of view, an effective adsorbent should be regenerable through several adsorption-desorption cycles. In this work, the study of sorbent regenerability was tested in 4 consecutive cycles at 298 K. After the adsorbent was exposed to the CO2 adsorption, it was recycled. The recycled adsorbent was then degassed at 100 °C for 24 h before the next cycle of CO2 adsorption. According to Fig. 8, the adsorption capacities of the sample decrease slightly after the first regeneration and are more or less constant after subsequent regenerations. The decrease in capacity after first regeneration can be attributed to the unreleased CO2 resulted from the first saturation. Moreover, no significant accumulation of CO2 is observed during each step of adsorption after first regeneration. This capability of satisfactory regeneration observed from the figure may be considered as a promising sign for stable performance in practical cyclic operations .
Comparison of CO2 adsorption capacity with other adsorbents
The CO2 uptake performance of the NDC sample synthesized in this work was compared to other adsorbents reported in the literature as summarized in Table 3. Based on Table 3, shows that there are many studies on the adsorption of CO2 using various types of adsorbent. Nevertheless, the adsorbent capacity for each adsorbent is different due to the different types of raw material and preparation procedures of adsorbents which affect the properties of them. It is clear that the NDC prepared in this study oﬀer a lot of promising beneﬁts for commercial purposes.
In conclusion, nitrogen-doped carbon was obtained by chemical activation of the polyaniline as a precursor with KOH. The carbon material activated at 800 °C showed high CO2 uptakes of 1.9 mmol/g-1 at 298 K and 1bar. Experimental results were analyzed through Langmuir, Freundlich and Sips models which lead to estimate relative performances of these three isotherm models, describe the experimental results, quantify the gas adsorption on prepared AC and classify them. Prepared adsorbent showed higher adsorption capacity in the whole pressure range and temperature investigated due to the existence of great affinity between CO2 molecules as a Lewis acid and basic amine sites on the NDC. Values of isosteric heat of adsorption were evaluated based on Clausius–Clapeyron equation. In the present study, the values of isosteric heat of adsorption obtained for NDC were lower than 80 kJ.mol-1 indicating that the interaction between NDC and CO2 molecules is mainly physical interactions involving weak van der Waals and electrostatic interactions. The high amounts of CO2 capture by NDC render it as a promising carrier for practical applications such as gas separation.
The authors gratefully acknowledge Babol Noshirvani University of Technology for ﬁnancial support of this work through Grant program No. BNUT/370302/98.
CONFLICTS OF INTEREST
There are no conflicts to declare.