Synthesis of nitrogen doped mesoporous carbons for carbon dioxide capture

Abstract

Nitrogen doped mesoporous carbons were prepared by nanocasting method at varying carbonization temperatures followed by characterization in terms of their structural, textural and chemical properties. Melamine-formaldehyde resin and mesoporous silica were used as the polymeric precursor and hard template respectively. Meso-structural ordering of the template was retained by the prepared carbons as suggested by the structural analysis. Evolution of nitrogen and oxygen functionalities along with textural properties of nitrogen doped carbons were regulated by the carbonization temperature. The prepared carbons obtained by carbonization at 700 °C exhibited a maximum surface area of 266 m² g⁻¹ along with a nitrogen content up to 21 weight%. CO₂ adsorption was studied in a fixed-bed column at several temperatures (30 to 100 °C) and CO₂ concentrations (5 to 12.5%). Adsorbent reusability was examined by carrying out multiple adsorption–desorption cycles. MF-700 showed the highest CO₂ adsorption capacity of 0.83 mmol g⁻¹ at 30 °C. CO₂ adsorption kinetics were investigated by fitting experimental CO₂ uptake data to different adsorption kinetic models, out of which the fractional order model was found to fit over the complete adsorption range with the error% between experimental and model predicted data within the range of 5%. In addition, the isosteric heat of adsorption was estimated to be around 17 kJ mol⁻¹, confirming the occurrence of the physiosorption process.

---

1. Introduction

Carbon dioxide (CO₂) emissions, resulting from fossil fuel combustion for power generation, have lead to growing concerns on global warming and climate change. On account of our strong dependence on fossil fuels and the inability of other energy sources, CO₂ non-emitting sources like solar energy, nuclear energy etc., to cope up with the energy requirements, there is an urgent requirement in the development of effective CO₂ capture techniques to reduce these CO₂ emissions.^1–3 Technologies proposed to capture CO₂ include absorption, adsorption, membrane separation and cryogenic separation.⁴ Absorption of CO₂ with alkanolamine solvents like mono-ethanol amine, di-ethanol amine etc. is used industrially but suffers from several drawbacks like high energy demand, corrosive nature and oxidative degradation of solvents.^5,6 Separation technology based on adsorption process is associated with low energy requirement, ease of regeneration etc. and hence is regarded as an attractive option for CO₂ capture.

Various adsorbents such as zeolites,^7,8 carbon based adsorbents,^9–11 metal organic frameworks,^12,13 and amine-modified silicas^14–17 have been studied. Among these available adsorbents, carbon materials are comprised of desired features for CO₂ capture like high surface area, fast kinetics, tailored pore structure, hydrophobic character, low energy for regeneration and high stability.¹⁸ In addition carbon adsorbents can be prepared from variety of raw materials and various techniques namely sol–gel process, nanocasting and pyrolysis of polymeric precursor. Among these nanocasting is an effective method for the fabrication of mesoporous carbons with controlled pore structure that are otherwise difficult to be prepared by the conventional methods. It consists of three steps (a) infiltration of carbon precursor into the pores of the template, (b) heat treatment in controlled atmosphere and (c) template removal.^19,20 The affinity of carbon adsorbents towards CO₂ can be increased by surface modification with amines or incorporation of nitrogen groups into the carbon matrix.^21–23 Nitrogen functional groups increase the surface polarity of carbon materials and basicity which enhances the interaction between acidic CO₂ molecules and carbon surface. Amine modified carbons have high CO₂ adsorption capacities but they suffer from amine degradation during regeneration. Nitrogen incorporated carbons can be prepared by pyrolysis of nitrogen containing polymers like melamine, acetonitrile, etc. with various templates such as zeolite, silica etc.^24–26 Hao et al.²⁷ reported the synthesis of porous carbon materials from resorcinol and formaldehyde with lysine as catalyst which exhibited static CO₂ adsorption capacity up to 3.13 mmol g⁻¹ at 25 °C. Zhao et al.²⁸ prepared nitrogen-doped carbons from IBN-9 mesoporous silica hard template by nanocasting technique and chemical activation. These materials showed equilibrium CO₂ uptake of 4.5 mmol g⁻¹ at 25 °C and 1 bar. Porous carbon materials, obtained from melamine-formaldehyde resin by nanocasting route, showed CO₂ capacity of 2.25 mmol g⁻¹ at 25 °C and 0.86 mmol g⁻¹ at 75 °C under 100% CO₂ flow.²⁴ Nitrogen enriched carbon material was prepared from urea-formaldehyde resin by carbonization followed by chemical activation with KOH and showed adsorption capacity of 3.21 mmol g⁻¹ at 25 °C.²⁹

Most of the carbon materials obtained from nanocasting technique are given post-synthesis treatment (physical or chemical activation) that are energy intensive and time consuming processes. Furthermore these materials are evaluated for CO₂ capture under static conditions at room temperature which are not pertinent to flue gas application. In the present study, nitrogen doped mesoporous carbons were prepared from nanocasting technique with the aim to evaluate them for CO₂ capture from flue gas. Nitrogen doping was employed in order to improve the selectivity of the adsorbents. They were characterized with X-ray diffraction (XRD), transmission electron microscopy (TEM), scanning electron microscopy (SEM), nitrogen adsorption–desorption, elemental analysis (CHN), Fourier transform infrared (FTIR) spectroscopy and X-ray photoelectron spectroscopy (XPS) for their physico-chemical properties. Fixed-bed column adsorption studies under simulated flue gas conditions have been carried out to evaluate their CO₂ adsorption capacity, regenerability and kinetics. Adsorption capacity at high temperatures like 75 °C and 100 °C was also evaluated keeping in view of their application to CO₂ capture from flue gas.

2. Experimental

All the reagents and solvents were procured from M/s S. D. Fine Chemicals India Ltd. Mesoporous silica (MCM-41 type) with specific surface area of 450 m² g⁻¹ and average pore diameter of 3.5 nm was used as template and was purchased from M/s Tianjin Chemist Scientific Ltd., Tianjin, China. Dry helium, nitrogen and carbon dioxide gases of grade-1 purity (99.999%) were obtained from M/s Sigma Gases and Services, India.

2.1 Preparation of nitrogen doped mesoporous carbons

Mesoporous carbons were prepared by nanocasting technique using melamine-formaldehyde resin as polymeric precursor and mesoporous silica as template. It included templated resin synthesis, followed by carbonization and then template removal in the last. To a mixture of 200 ml methanol (5 weight%) and melamine (46.6 g), heated to 70 °C, 200 ml of formaldehyde solution (37% w/v) was slowly added over a period of 3 h to form hexamethylomelamine. Potassium carbonate was added to adjust the pH in the range of 8–9. Advanced polymerization was carried out by raising the reaction temperature to 75 °C and maintaining the solution pH to 7–8, by the addition of sodium hydroxide (N/10) and borax (0.4 g), for further 2 h. This was followed by cooling down of the reaction mixture to room temperature and addition of 6 ml of H₂SO₄ solution (48%). 15.0 g of mesoporous silica (template) was added and mixed thoroughly to this reaction mixture. After this, curing of templated resin was carried out at 60 °C for 2 h and then left at room temperature until complete solidification had occurred.

Templated resin samples were carbonized in a quartz tubular furnace under a nitrogen flow by heating at 10 °C min⁻¹ from room temperature up to 400–700 °C for 1 h. The mesoporous carbons were obtained after dissolution of the silica framework in 40 weight% NaOH solution followed by washing with water and then drying at 100 °C. Prepared materials were marked as MF-x, where x denotes the carbonization temperature (400–700 °C).

2.2 Characterization of nitrogen doped mesoporous carbons

The powder X-ray diffraction (XRD) patterns of the prepared carbons were collected using a PANalytical X'Pert Pro diffractometer using Cu–Kα radiation at 45 kV and 40 mA. The low angle XRD patterns were recorded over the scanning range (2θ) of 0.5 to 10° while the wide angle patterns were collected over the scanning range (2θ) of 10 to 80°. Transmission electron microscopy of carbon materials was carried out on a Philips CM200 transmission electron microscope at an accelerating voltage of 200 kV. Specimen preparation included the dispersion of carbon sample in toluene on carbon-coated copper grid followed by evaporation of toluene at room temperature. Scanning electron micrographs were recorded on a JEOL JSM-7600F field emission gun-scanning electron microscope (FEG-SEM) in the back scattering mode with an accelerating voltage of 10.0 kV. The samples were coated with a gold film of 50 μm thickness in an automatic sputter coater (Polaron) in order to avoid charging under electron beam.

Nitrogen adsorption–desorption isotherms were measured on a Micromeritics ASAP 2010 sorption analyzer at −196 °C. Prior to adsorption measurements, the samples were outgassed at 220 °C for 6 h under vacuum. The Brunauer–Emmett–Teller (BET) method was used to calculate the specific surface area. The pore size distribution and pore volume were obtained from the adsorption branch using the Barrett–Joyner–Halenda (BJH) model.

Elemental (CHN) analysis was done using a Thermo Finnigan Flash EA 1112 Series elemental analyzer. Oxygen content was determined by difference. Fourier transform infrared (FTIR) spectroscopic studies of prepared carbons were carried out using a Perkin Elmer Spectrum 100 FTIR spectrometer (Perkin Elmer, USA) in the range of 4000–625 cm⁻¹ with a resolution of 4 cm⁻¹. Surface chemistry of the carbons was analyzed by carrying out X-ray photoelectron spectroscopy (XPS) on a SPECS system equipped with a hemispherical electron analyzer using Mg Kα X-ray source (energy 1253.6 eV) operating at an emission current of 10 mA and anode potential of 15 kV. In the analysis chamber, the pressure was maintained below 2 × 10⁻⁹ torr. The pass energies for low and high resolution scans were 50 and 20 eV respectively. XPS peak 4.1 software was used to analyze the data and all the core level spectra were peak fitted to a mixed Gaussian–Lorentzian convoluted function (80/20) after Shirley background subtraction.

2.3 Dynamic adsorption–desorption measurements

The experimental setup used for dynamic CO₂ adsorption–desorption experiments is represented schematically in Fig. S1.† The adsorption column is made up of stainless steel with a length of 300 mm and an internal diameter of 9.39 mm. The adsorption column is filled with 2 g of dry adsorbent mixed with dry inert glass beads and placed in a temperature-controlled oven. Both the ends of the column are filled with temperature resistant glass wool to minimize the adsorbent loss. A K-type thermocouple, placed in the middle of the bed, monitors the adsorbent bed temperature. CO₂ and N₂ are mixed in a gas mixing section, made up of stainless steel, prior to adsorption column. Feed concentration is modulated by regulating the flow rates of pure CO₂ and N₂ with the aid of two mass flow controllers (Bronkhorst, Netherlands). The concentration of these gases is measured by a gas chromatograph (GC 7890A, Agilent Technologies, US), equipped with a thermal conductivity detector. Fused silica capillary column (HP Plot Q, Agilent Technologies, USA) having a length of 30 m, inner diameter of 0.32 mm and film thickness of 20 μm is used for the analysis.

In a typical experiment, the adsorbent is treated under a N₂ flow at 200 °C for 2 h to remove pre-adsorbed gases or moisture. The temperature is then lowered to the adsorption temperature (30 to 100 °C) and the nitrogen flow is switched to simulated flue gas at a flow rate of 80 ml min⁻¹. CO₂ concentration in the simulated flue gas was varied from 5 to 12.5 volume%. CO₂ concentration at the bed exit is continuously monitored until the saturation is reached. After adsorption, the simulated flue gas is switched back to pure N₂ and temperature is increased to 200 °C to carry out the desorption study. Four adsorption–desorption cycles were carried out to check the cyclic stability of the adsorbents.

The CO₂ adsorption capacity of the adsorbent is calculated using the following equation:

where q_t (mmol g⁻¹) is the CO₂ adsorption capacity, m (g) is the mass of the adsorbent, Q (ml min⁻¹) is the gas flow rate, C_o and C are the inlet and effluent CO₂ concentrations (percent in volume) respectively and t (minutes) is the time.

2.4 Equilibrium CO₂ sorption measurements

CO₂ adsorption and desorption experiments on prepared carbon (MF-700) was carried out on a Micromeritics ASAP 2010 volumetric analyzer at four different temperatures 30 °C, 50 °C, 75 °C and 100 °C and at CO₂ pressures from 0 to 1 bar. Before adsorption measurements, samples were degassed at 200 °C for 12 h under vacuum.

2.5 Adsorption kinetic models

Fast kinetics is one of the essential parameters expected in a good adsorbent. In a dynamic process like adsorption in a fixed-bed system, kinetic study of the adsorption process provides a base to determine the adsorbent's efficiency and its ability to withstand high adsorbate flow. Typically the experimental data is fitted with a series of the established models and the best fitting model is selected. In the present work, pseudo-first order, pseudo-second order and fractional order adsorption kinetic models have been used to study the adsorption process of the synthesized adsorbents.

The Lagergren's first order equation is one of the most widely used model describing adsorption kinetics based on adsorption capacity. The pseudo-first order model can be written as:³⁰

where q_e and q_t (mmol g⁻¹) are the adsorption capacities at equilibrium and time t (minutes) respectively and k₁ (min⁻¹) is the rate constant of the pseudo-first order model. The integration of eqn (2) with initial conditions (q_t = 0 at t = 0 and q_t = q_t at t = t) leads to following equation:which can be rearranged to:

q_t = q_e(1 − e^−k₁t) (4)

The pseudo-second order model is represented as:³¹

where k₂ (g mmol⁻¹ min⁻¹) is the pseudo-second order rate constant. Integration of eqn (5) for above mentioned boundary conditions gives the following form of the equation:which can be rearranged to

The fractional order kinetic model assumes adsorption rate to be influenced by n^th power of driving force and m^th power of the adsorption time^32,33 and is represented as eqn (8):

where k_n is the fractional order rate constant and m and n are the model constants. For m = 1 and n = 1 or 2, eqn (8) is reduced to pseudo-first order or pseudo-second order models respectively. Integrated form of eqn (8) for the above given boundary condition is:

Linear regression analysis of pseudo-first order and pseudo-second order kinetic models has been widely used to determine the best fitted adsorption kinetic model. Pseudo-second order kinetic model equation can be presented in four different linear forms, each resulting in different model parameters. This is because of the change in the error structure upon linearization of the non-linear form of the kinetic model. Hence nonlinear regression is the appropriate way to predict adsorption kinetics.^34–37 Thus parameters of the above adsorption kinetic models were calculated by nonlinear regression on the experimental data with the help of OriginPro 8 software.

The adequacy of each model is determined by calculating an error function based on the normal standard deviation. The consistency between the experimental values and the model predicted values are expressed by this error function.

where error (%) is the error function, q_t(exp) and q_t(pred) are the amounts of CO₂ adsorbed at a given time determined experimentally and predicted by model respectively and N is the total number of experimental points.

2.6 Isosteric heat of adsorption

The isosteric heat of adsorption is defined as the difference between the activation energy required for adsorption and desorption process and it represents the strength between the adsorbate and the adsorbent.³⁸ It was calculated from the experimental adsorption data by using Clausius–Clapeyron equation:where Q_st (kJ mol⁻¹) is the isosteric heat of adsorption at a given q_e. Q_st is calculated by plotting ln P against 1/T at a constant amount of CO₂ adsorbed (q_e) which yields a straight line with a slope equal to −Q_st/R.

3. Results and discussion

3.1 Characterization of nitrogen doped mesoporous carbons

Fig. 1a shows the low angle powder XRD patterns of the silica template and the prepared carbons. The typical structure of silica exhibits two well resolved peaks at 2θ = 0.87° and 1.99° corresponding to (100) and (110) reflections respectively of two dimensional hexagonal mesostructure with lattice constant a_o = 11.7 nm. All the prepared carbons showed similar XRD patterns implying that the prepared materials exhibit mesostructural ordering.

Fig. 1 (a) Low angle XRD patterns of silica template and prepared mesoporous carbons, and (b) wide angle XRD patterns of prepared carbons.

The wide angle XRD patterns of the prepared carbons are presented in Fig. 1b. A broad diffraction peak is obtained at 2θ = 26–27° corresponding to the (002) diffraction planes of graphitic carbon having hexagonal structure (JCPDS X-ray Powder Diffraction Database no. 75-1621) with a d-spacing of 0.34 nm and is very similar to that of ideal graphite.³⁹ Also the resulting carbon materials are not completely crystalline as suggested by the absence of strong diffraction patterns. The peak intensity decreased with increasing carbonization temperature suggesting an increase in irregularity of layer structures. Line broadening of the most intense peak indicates very small crystallite size suggesting the prepared carbons to be nanostructured.

Transmission electron micrographs (Fig. S2a and S2b†) of carbon prepared at 700 °C show the formation of nanostructured materials with particle size in the range of 30 to 50 nm. Heterogeneous types of pores can be clearly seen in SEM images (Fig. S2c and S2d†). External pores between the particles are also present which help in the gas diffusion from bulk phase to adsorbent surface.

Textural parameters of the synthesized carbons and mesoporous silica are obtained by nitrogen sorption measurements. Fig. 2 presents the N₂ adsorption–desorption isotherms of mesoporous silica and MF-x materials prepared at different carbonization temperatures. Mesoporous silica and carbons other than MF-700 show a typical type IV isotherm with a hysteresis loop caused by capillary condensation which is characteristic of mesoporous materials. While MF-700 displays a combination of I and IV type profile, demonstrating the presence of both micropores and mesopores. Mesoporosity of the carbons is obtained on the dissolution of the silica template while small microporosity in MF-700 is generated during the pyrolysis of polymeric precursor of its own accord. Nitrogen adsorption increases with increase in carbonization temperature and is highest for 700 °C sample. Carbonization up to 600 °C is found to be not sufficient enough for the porosity development as indicated by small surface area values. MF-700 exhibits surface area of 266 m² g⁻¹ and pore volume of 0.25 cm³ g⁻¹ with very small micropore volume of 0.03 cm³ g⁻¹. Width of the hysteresis loop is more pronounced for parent silica template as compared to the porous carbons. Additionally large difference in pore diameters of silica template and prepared carbons is observed which could be ascribed to occurrence of hysteresis loop for silica at lower relative pressure as compared to that of the replicated carbons. Moreover, it is worth mentioning that the pore size of the prepared carbons is larger than the wall thickness of the silica template (ca. 8.4 nm) and this could be due to either partial filling of polymeric precursor in the template's pores or shrinkage of polymeric precursor filled in the pores of the template during high thermal treatments. Porous properties of parent silica and obtained porous carbons are congruous with already reported literature.^40–42 Hence employment of nanocasting technique for the development of carbon materials produces tailor made porous structure by the silica template having felicitous pore size and wall thickness.

Fig. 2 N₂ adsorption (closed symbols) and desorption (open symbols) isotherms of prepared carbons and mesoporous silica.

Pore size distribution of carbon materials from adsorption branch of BJH method is shown in Fig. S3† and all the samples show average pore diameter in the mesopore region i.e. 2 to 50 nm. Textural properties of these carbons and silica are given in Table 1.

Table 1 Textural parameters of mesoporous carbons and silica template^a

Sample	SBET (m^2 g^−1)	VBJH (cm^3 g^−1)	dp (nm)
a SBET: Surface area obtained from BET method; VBJH: Pore volume obtained from BJH method using adsorption branch; dp: Average pore size.
MF-400	6	0.022	14.80
MF-500	23	0.067	9.89
MF-600	42	0.117	9.41
MF-700	266	0.249	10.04
Silica	450	0.697	3.5

---

Fig. S4† depicts the elemental composition of the prepared carbon materials. Nitrogen content of the prepared adsorbents decreases continuously from 35 weight% to 21 weight% with increase in carbonization temperature from 400 to 700 °C. This reduction in nitrogen amount at higher carbonization temperatures results in increase in carbon amount of the prepared materials from 41 weight% to 52 weight%.

Investigation of the chemical functionalities developed in carbon adsorbents from melamine-formaldehyde resin is important in order to understand the adsorption process of these materials. It is difficult to determine the exact structure of the melamine-formaldehyde resin because of its complex structure. Moreover the exact structure of carbonized MF resin based carbon adsorbents is also difficult to be determined on account of development of wider functional groups during the carbonization process. Possible reaction equations for the polymerization of melamine and formaldehyde are demonstrated in Fig. 3. Hexamethylomelamine is formed during the addition of formaldehyde to melamine. Advanced polymerization step involves the formation of two kinds of bridges between the hexamethylomelamine molecules namely ether linkage (–N–CH₂–O–CH₂–N) and methylene linkage (–N–CH₂–N–).

Fig. 3 Possible reaction equation for melamine-formaldehyde resin synthesis.

FTIR spectra of the mesoporous carbon adsorbents show the development of similar functional groups in the materials (Fig. S5†). A small peak at 1059 cm⁻¹ is observed indicating the presence of C–O linkage of ether, alcohol or ester group. A peak at 1246 cm⁻¹ is observed attributing to the stretching vibration of aromatic C–N. A band around 1604 cm⁻¹ can be assigned to the in-plane deformation vibrations of N–H band and/or stretching vibration of aromatic C=C bonds suggesting the presence of aromatic rings. Peak at 1770 cm⁻¹ corresponds to stretching vibration of C=O bonds of esters, carbonyls, anhydrides. A small peak around 3000 cm⁻¹ corresponds to stretching vibration of C–H bond in –CH₂–NH–CH₂– or –CH₂–NH–CH₃ groups. The weak bands ca. 3400 cm⁻¹ is credited to symmetric stretching vibration of N–H group.^27,43

The nature of the functional groups present on the surface of melamine-formaldehyde based carbons was evaluated by X-ray photoelectron spectroscopy (XPS). In the survey spectra (Fig. S6†) of the carbon materials, three distinct peaks ca. 285, 400, 532 eV were observed attributing to carbon (C1s), nitrogen (N1s) and oxygen (O1s) respectively. No peak attributing to Si (2s or 2p) was detected demonstrating the complete removal of silica template from the mesoporous carbons by dissolution in sodium hydroxide solution. High resolution spectra were analysed to determine the chemical binding states in the prepared carbons.

The high resolution XPS spectra of the C1s region of MF-based carbons are shown in Fig. 4. The deconvolution of each C1s core level spectrum resulted in five well resolved peaks with binding energies around 284.9, 285.7, 286.8, 288.6 and 290.0 eV labelled as C1, C2, C3, C4 and C5 respectively. C1 peak corresponds to sp² (C=C) and sp³ (C–C) hybridized graphitic carbon while C2 ascribes to C–O bond of phenol, alcohol or ether and/or C=N bond. C3 peak is assigned for carbonyl or quinone groups and/or C–N linkage, C4 for carboxyl or ester linkages, and C5 for shake up satellite peaks because of π–π* transitions in aromatic rings.⁴⁴ The corresponding binding energy (B. E.), full width at half maximum (FWHM) and relative area contribution (A%) of each deconvoluted peak are reported in Table 2. Relative area percentage of C1 peak increased from ca. 3.7% to 23.4% on increasing the carbonization temperature from 400 to 700 °C suggesting enhanced graphitization at higher temperatures. MF-400 and MF-500 display similar envelopes for C1s region with carbonyl or quinone and/or C–N linkages being the dominant components. Carbonization at or above 600 °C resulted in a significant change in the surface functional groups as seen in Fig. 4. There is a considerable increase in ether, phenol, alcohol and/or C=N linkages for MF-600 and MF-700 as compared to MF-400 and MF-500. Hence with increase in carbonization temperature from 400 to 700 °C, the relative area percentages of graphitic carbon, carbon–nitrogen structures and carbon–oxygen structures have varied appreciably suggesting the cleavage of C–N, C–O and C=O structures and also the formation of more stabilised structures.

Fig. 4 Deconvoluted high resolution XPS spectra of C1s region of mesoporous carbons.

Table 2 Deconvolution results of C1s core level spectra of mesoporous carbons^a

Sample		C1	C2	C3	C4	C5
a BE: Binding energy (eV); FWHM: Full width at half maximum (eV); A%: relative area percentage.
MF-400	BE	284.93	285.89	286.71	288.58	289.81
	FWHM	1.02	1.29	1.87	1.92	1.91
	A%	3.72	21.76	47.50	16.59	10.43
MF-500	BE	284.98	285.89	286.75	288.58	289.86
	FWHM	1.29	1.33	1.96	1.86	2.23
	A%	5.80	23.68	42.46	15.54	12.52
MF-600	BE	284.84	285.58	286.87	288.65	290.22
	FWHM	0.95	1.36	2.27	1.83	1.93
	A%	21.233	39.23	29.80	6.65	3.09
MF-700	BE	284.84	285.54	286.87	288.63	290.22
	FWHM	0.99	1.71	2.29	1.98	2.37
	A%	23.42	48.93	19.70	4.75	3.20

---

Fig. 5 Deconvoluted high resolution XPS spectra of N1s region of mesoporous carbons.

High resolution XPS spectra of N1s region (Fig. 5) was deconvoluted into four different peaks with binding energies around 398.8, 400.0, 401.1, and 403.0 eV referred as N1, N2, N3, and N4 respectively with corresponding B. E., FWHM and relative area contributions listed in Table 3. Peak N1 is attributed to the pyridinic form of nitrogen and is reported to be stable up to even 1000 °C. Peak N2 corresponds for pyrrolic and/or pyridonic nitrogen. Pyridonic nitrogen is not found at same binding energy as pyridinic nitrogen though it contains nitrogen in a six membered ring structure. As a consequence of tautomeric equilibrium in pyridone, nitrogen in pyridone and pyrrole exhibits similar chemical environment and it contributes two p-electrons to the aromatic π-system. In contrast, pyridinic nitrogen contributes only one p-electron to the π-system. Thus within the precision of XPS measurements, pyridonic-N and pyrrolic-N cannot be distinguished. Peak N3 ascribes to quaternary nitrogen and peak N4 to pyridinic nitrogen oxides. Several structures can be possible for quaternary nitrogen however only nitrogen atoms substituting carbon in aromatic graphene structure at center or valley positions are stable during carbonization process.^44–47 Binding energy of N2 peak increased from 399.84 to 400.14 eV on increasing the carbonization temperature suggesting the presence of other nitrogen structures like amine, amide or imide in addition to pyrrolic and pyridonic nitrogen. Amide groups are found around binding energy of 399.6–399.9 eV while imides are reported to be found around 399.7 eV and hence it becomes difficult to separate these peaks.⁴⁸

Table 3 Deconvolution results of N1s core level spectra of mesoporous carbons^a

Sample		N1	N2	N3	N4
a BE: Binding energy (eV); FWHM: Full width at half maximum (eV); A%: relative area percentage.
MF-400	BE	398.80	399.86	401.16	403.05
	FWHM	1.65	1.73	1.56	4.45
	A%	10.76	58.70	24.89	5.66
MF-500	BE	398.83	399.84	401.14	403.04
	FWHM	0.98	1.76	1.57	4.41
	A%	8.25	63.04	23.38	5.33
MF-600	BE	398.73	400.09	401.07	403.13
	FWHM	1.54	1.71	1.44	4.52
	A%	53.22	28.85	12.70	5.23
MF-700	BE	398.73	400.14	401.14	403.12
	FWHM	1.55	1.64	1.33	4.48
	A%	46.66	30.19	15.84	7.31

---

Up to 500 °C, nitrogen functionalities are present mainly in the form of pyrrolic and/or pyridonic nitrogen in addition to amines, amides and imides. On increasing the temperature from 500 to 600 °C, a considerable increase in area percentage of N1 peak and decrease in area percentage of N2 peak is observed. This could be attributed to conversion of pyrrolic nitrogen to pyridinic nitrogen under severe carbonization conditions and also to loss of primary amines that are present in melamine-formaldehyde resin and are reported to decompose at 600 °C. Pyridinic nitrogen oxides are present as minor component in all the prepared carbon adsorbents. Pyridinic nitrogen and quaternary nitrogen are reported to be the most stable forms of nitrogen formed during the carbonization process. In the present study, contribution of these two functionalities increased from ca. 35.6% to 62.5% with increase in temperature from 400 to 700 °C. Moreover, as most of the nitrogen functionality in the prepared carbons is present in the form pyridinic and pyrrolic or pyridonic nitrogen, which are reported to have electron donating capability and hence exhibit Lewis basicity, they may serve as active sites for adsorption of CO₂.

Fig. 6 shows the high resolution O1s core level spectra of the prepared carbons and their deconvolution lead to four separate peaks with binding energies around 531.3, 532.4, 533.6, and 535.0 eV labelled as O1, O2, O3 and O4 respectively (Table 4). Peak O1 is ascribed to C=O of carbonyl or ketone groups, O2 to carbonyl oxygen of esters, anhydrides, amides, and oxygen atom of phenol, alcohol or ether groups, O3 to oxygen of carboxylic groups, and O4 to oxygen in water.⁴⁴ In MF-400, peak O3 has the highest relative area contribution and hence oxygen is mainly present in the form of carboxylic group. But with increase in carbonization temperature, relative peak area of O1 and O2 peaks have increased appreciably at the expense of carboxylic groups. Development of oxygen functionalities with carbonization temperature as predicted by O1s spectra are in accordance with that obtained from C1s core level spectra of the carbons.

Fig. 6 Deconvoluted high resolution XPS spectra of O1s region of mesoporous carbons.

Table 4 Deconvolution results of O1s core level spectra of mesoporous carbons^a

Sample		O1	O2	O3	O4
a BE: Binding energy (eV); FWHM: full width at half maximum (eV); A%: relative area percentage.
MF-400	BE	531.35	532.49	533.64	535.01
	FWHM	1.05	1.33	2.41	1.27
	A%	3.46	13.92	70.85	11.77
MF-500	BE	531.36	532.47	533.68	535.02
	FWHM	1.30	1.98	2.29	1.30
	A%	6.62	40.18	46.90	6.31
MF-600	BE	531.30	532.45	533.65	535.06
	FWHM	1.69	1.73	1.77	2.59
	A%	36.98	33.18	25.00	4.85
MF-700	BE	531.26	532.44	533.67	535.16
	FWHM	1.78	1.73	1.75	3.11
	A%	39.84	26.40	21.03	12.73

---

3.2 CO₂ adsorption measurements

3.2.1. CO₂ adsorption with carbonization temperature. Synthesized carbon adsorbents were evaluated for CO₂ adsorption capacity at 30 °C in 10% CO₂ feed concentration and results are shown in Fig. 7a. As carbonization temperature increases so does the CO₂ adsorption capacity, being highest for the 700 °C carbon. Carbonization up to 600 °C lead to an increase in adsorption capacity but a significant increase is observed for MF-700.

Fig. 7 (a) Effect of carbonization temperature on CO₂ adsorption capacity of synthesized carbons for 10% CO₂ feed concentration at 30 °C (b) N₂ (open symbols) and CO₂ (closed symbols) breakthrough curves on MF-600 and MF-700 for adsorption at 30 °C and 5% CO₂ feed concentration.

Highest adsorption capacity of 0.78 mmol g⁻¹ is achieved under 10% CO₂ feed gas by the adsorbent prepared at 700 °C. This could be attributed to the improvement of textural properties like surface area and pore volume in addition to increase in basic functional groups on the adsorbent surface. As observed from the nitrogen sorption isotherms at 77 K, there is significant improvement in the textural properties of the prepared carbons with development of some micropores in MF-700 with increase in the carbonization temperature. Basicity of carbons is due to presence of heteroatoms like nitrogen and oxygen. As observed from the XPS analysis of prepared carbons, nitrogen moieties present on MF-400 and MF-500 are mainly pyrrolic/pyridonic nitrogen followed by quaternary nitrogen. Pyridinic nitrogen content is very small on these carbons. On the other hand, most of the nitrogen functionality present on carbons obtained at 600 °C and 700 °C is in the form of pyridinic nitrogen followed by pyrrolic/pyridonic nitrogen. The relative percentage of quaternary nitrogen decreased in these carbons. Both pyridinic and pyrrolic/pyridonic nitrogen groups have lone pair of electrons and exhibit Lewis basicity thereby providing active sites for acidic CO₂ adsorption. However pyridinic nitrogen is found to have higher affinity for CO₂ molecules than pyrrolic nitrogen attributable to its more basic nature.^43,49 In contrast quaternary nitrogen is found to have Lewis acidity.^50,51 Hence MF-600 and MF-700 samples have higher basic character than MF-400 and MF-500. It is also important to note that CO₂ adsorption capacity is not directly associated with total nitrogen content of the carbon materials but is dependent on the kind of nitrogen moieties. Consequently CO₂ uptake of MF-600 and MF-700 is higher than that of MF-400 and MF-500 although there is decrease in total nitrogen content with carbonization temperature. Kind of oxygen functionalities present also affect the surface chemistry of the carbon material. Carbonyl, pyrones and ether groups are responsible for basicity of the carbon surface while carboxyl, lactones or lactols and phenol groups contribute towards surface acidity. With increase in carbonization temperature, there is an increase in relative percentage of carbonyl and pyrone groups and a decrease in carboxyl and lactone/lactol groups as suggested by O1s core level spectra. But the contribution of ether and phenol groups cannot be distinguished from XPS spectra as both appear around 532.4 eV binding energy. Enhancement in surface basicity of the porous carbons is observed with rise in carbonization temperature. Consequently, both the textural properties and surface chemistry of carbon materials play important role in CO₂ adsorption performance. As MF-700 demonstrated best textural properties and high basic character along with highest CO₂ uptake, this adsorbent is selected for further studies.

In addition to adsorption capacity of the adsorbents, selectivity of CO₂ over N₂ is an important parameter to be studied. Breakthrough curves of N₂ and CO₂ on MF-600 and MF-700 for adsorption at 30 °C under 5% inlet CO₂ concentration are depicted in Fig. 7b. Both the porous carbons exhibit very low N₂ adsorption capacity as suggested by presence of nitrogen gas right away in the exit gas stream. Alternatively, CO₂ gas concentration gradually increased for MF-600 but for MF-700 CO₂ gas concentration at the column exit was zero for around 3.5 minutes.

3.2.2. CO₂ adsorption with adsorption temperature. Fig. 8 illustrates CO₂ adsorption capacity of MF-700 at different CO₂ feed concentrations (5 to 12.5%) as a function of adsorption temperature (30 to 100 °C). Adsorption capacity decreased gradually with increasing adsorption temperature at all gas compositions. This is attributed to physiosorption process where surface adsorption energy and diffusion rate of molecules tend to increase with temperature resulting in instability of adsorbed gas on the carbon surface thereby leading to desorption of adsorbed CO₂ molecules. The values dropped to 0.29 mmol g⁻¹ at 100 °C from 0.83 mmol g⁻¹ at 30 °C for CO₂ concentration of 12.5%. This fall in capacity is more significant at higher CO₂ partial pressures. Availability of CO₂ per adsorption site is limited at low CO₂ partial pressure and hence has lower effect on reduction in adsorption capacity.

Fig. 8 CO₂ adsorption capacity of MF-700 at different CO₂ partial pressures as a function of adsorption temperature.

Activated carbon obtained from carbonization of coal tar pitch and furfural followed by activation with steam showed CO₂ adsorption capacity of 0.61 mmol g⁻¹ under 15% CO₂ rest N₂ flow at 30 °C. The capacity of this material reduced to 0.30 and 0.15 mmol g⁻¹ at 50 °C and 80 °C respectively.⁵² Equilibrium CO₂ capacity of MF-700 at 50 °C is ∼2 times more than this synthetic activated carbon. Carbon from soybean by chemical activation with ZnCl₂ and physical activation with CO₂ exhibited equilibrium capacity of 0.93 and 0.51 mmol g⁻¹ at 30 °C and 75 °C respectively under 15.4% CO₂ atmosphere. But for adsorption at 75 °C, this carbon material could not be regenerated completely in the second cycle resulting in zero capacity in the next adsorption cycle.⁵³ Fly ash derived activated carbons exhibited CO₂ uptake of 0.93 and 0.42 mmol g⁻¹ at 30 °C and 75 °C respectively.⁵⁴ Hence the CO₂ adsorption capacity achieved for the prepared carbons are in good agreement with the values reported in literature.

3.2.3. CO₂ adsorption with CO₂ concentration. To study the effect of CO₂ concentration on adsorption capacity of MF-700, fixed-bed experiments were conducted in four varying CO₂ concentrations (5 to 12.5%). As seen in Fig. 8 adsorption capacities under various CO₂ concentrations demonstrate similar temperature dependent behavior. Adsorption capacity rises with CO₂ partial pressure at a given adsorption temperature. This is ascribed to increased concentration gradient and hence the rate of CO₂ mass transfer across the boundary layer and within the adsorbent pores at high CO₂ partial pressure. However this increase in capacity decreases with CO₂ concentration. At 30 °C, capacity increased by only ∼6% when CO₂ concentration was increased from 10 to 12.5% over an increase of ∼32% for concentration increase from 7.5 to 10%.

3.2.4. Multi-cycle adsorption study. Reusability of the adsorbent is another important parameter for its long-term use in industries and to assess the same cyclic adsorption–desorption experiments were carried out. Desorption study was carried out by switching the inlet gas from simulated flue gas to pure N₂ and increasing the bed temperature to 200 °C. Fig. 9a illustrates the multi-cycle CO₂ adsorption–desorption curves of MF-700 for adsorption at 30 °C and 12.5% CO₂ concentration. It can be seen that complete saturation is achieved within 10 minutes. Adsorbed CO₂ was easily removed from the carbon surface as suggested by the sudden fall in CO₂ concentration profile.

Fig. 9 (a) Multi-cycle adsorption–desorption concentration profile (b) Cyclic regenerability after 4 consecutive adsorption–desorption cycles of MF-700 for adsorption at 30 °C under 12.5% CO₂ concentration.

CO₂ adsorption capacity of MF-700 at different temperatures under 12.5% CO₂ atmosphere during repeated cycles is depicted in Fig. 9b. CO₂ adsorption capacities for four runs were nearly identical implying adsorption process to be completely reversible, being assigned to weak van der Waal forces acting between the adsorption sites and CO₂ molecules. This confirms the stability of CO₂ adsorption performance of MF-700 for long-term cyclic operations.

3.2.5. Equilibrium CO₂ sorption measurements. As mesoporous carbon obtained at 700 °C presented the best textural properties among the prepared carbons, we investigated the equilibrium CO₂ uptake on this adsorbent at various temperatures ranging from 30 °C to 100 °C. Fig. 10 illustrates the CO₂ adsorption–desorption isotherms of MF-700 as a function of both pressure and temperature. CO₂ uptake was found to decrease with increase in adsorption temperature but it increased with increase in pressure. CO₂ uptake decreased from 1.15 mmol g⁻¹ to 0.52 mmol g⁻¹ on increasing the adsorption temperature from 30 °C to 100 °C. A hysteresis loop is observed in the adsorption–desorption isotherms at all the temperatures suggesting small energy requirement for reuse of the adsorbent.

Fig. 10 CO₂ adsorption–desorption isotherms of MF-700 at different temperatures.

CO₂ uptake of 1.4 mmol g⁻¹ at 25 °C and 1 bar was demonstrated by carbon nitride materials obtained from polymerization of ethylenediamine and carbon tetrachloride in the pore of mesoporous silica FDU-12 and carbonization at 600 °C for 5 h.⁴² Direct pyrolysis of resorcinol, formaldehyde and lysine at 500C (RFL-500) produced carbon that exhibited very high CO₂ uptake of 3.13 mmol g⁻¹ at 25 °C as compared to our material.²⁷ But the CO₂ uptake at higher temperatures are comparable with 0.62 mmol g⁻¹ at 120 °C for RFL-500 and 0.52 mmol g⁻¹ at 100 °C for MF-700. Adsorption capacities at higher temperatures are more relevant with respect to CO₂ capture from flue gas.

3.2.6. Adsorption kinetics. Fig. 11 shows the CO₂ adsorption capacity of MF-700 as a function of time, in 10% CO₂ balance N₂, at various adsorption temperatures and the corresponding adsorption profiles as predicted by different kinetic models. Most of the adsorption occurred within few minutes of contact with CO₂. Later on adsorption rate became slower until equilibrium condition was attained. During the adsorption process, there is an increase in diffusion resistance and reduction in the unoccupied active sites which leads to a decrease in the adsorption rate. Kinetic model parameters along with correlation coefficients (R²) and associated errors (as calculated by eqn (10)) are reported in Table 5. Among the studied kinetic models, fractional order model provides the best explanation of CO₂ adsorption process on mesoporous carbon adsorbent on the fixed bed.

Fig. 11 Experimental CO₂ uptake on MF-700 and corresponding fit to kinetic models at 10% CO₂ and different adsorption temperatures.

Table 5 Kinetic model parameters for adsorption at 10% CO₂ on MF-700

Kinetic model	Parameter	Temperature (°C)
		30	50	75	100
Pseudo-first order	k1 (min^−1)	0.244	0.203	0.158	0.155
	qe (mmol g^−1)	0.773	0.589	0.484	0.263
	R^2	0.996	0.993	0.979	0.961
	Error%	4.480	4.026	7.791	10.919
Pseudo-second order	k2 (g mmol^−1 min^−1)	0.383	0.398	0.344	0.640
	qe (mmol g^−1)	0.871	0.675	0.569	0.308
	R^2	0.976	0.996	0.995	0.985
	Error%	10.414	2.480	2.997	6.571
Fractional order	kn (mmol^1−m g^m−1 min^−n)	0.307	0.272	0.111	0.072
	qe (mmol g^−1)	0.783	0.615	0.493	0.269
	n	1.755	1.302	0.521	0.445
	m	1.614	0.943	0.565	0.503
	R^2	0.999	0.998	0.996	0.996
	Error%	0.749	1.579	2.223	2.594

---

Pseudo-first order and pseudo-second order kinetic models underestimated the CO₂ uptake at the initial adsorption phase for adsorption at all temperatures except 30 °C. Pseudo-second order model also overestimated the equilibrium CO₂ capacity of the adsorbent while pseudo-first order model was found to be in good accord with the experimental data at the final adsorption phase. Also higher values of error%, based on normal standard deviation, for these two kinetic models suggest that these models are not strictly followed by the experimental CO₂ uptake for mesoporous carbon.

On the other hand, adsorption behavior as predicted by fractional order model was in good agreement with the experimental data over the entire adsorption region with maximum error of 2.6%. Lower values of error% from fractional order model compared to pseudo-first and pseudo-second order models, within the range of adsorption temperature and CO₂ feed concentration, recommend its applicability for mesoporous carbon adsorbent.

As CO₂ adsorption is an exothermic process and is favored at lower temperatures, there was a decrease in adsorption rate constant k_n with increase in adsorption temperature. Parameters n and m of fractional order model indicate the effect of the driving force and diffusion resistance respectively. Increase in adsorption temperature led to a decrease in both of these parameters. Pseudo-order with respect to driving force i.e. n had the maximum value at 30 °C indicating the large driving force at this temperature. The system under study followed kinetics with n around 1.3–1.8 at low temperatures (30 and 50 °C) but for adsorption at higher temperature of 100 °C, the order reduced to 0.45 with respect to driving force. Predicted equilibrium adsorption capacity was also found to be in good agreement with the experimental value.

Fig. 12 shows the experimental and fractional order model predicted uptake of CO₂ vs. time as a function of CO₂ feed concentration at 30 °C. Fractional order model was found to predict the CO₂ adsorption behavior of mesoporous carbon adsorbent, MF-700, appropriately for all the CO₂ concentrations. CO₂ uptake is smaller at low partial pressure of CO₂ because of the smaller driving force.

Fig. 12 CO₂ adsorption capacity vs. time as a function of CO₂ feed concentration at 30 °C (symbols represent experimental data and curves represent model predicted data).

3.2.7. Isosteric heat of adsorption. Isosteric heat of adsorption was calculated from the slope of the isosteres (logarithmic pressure of CO₂ vs. inverse of adsorption temperature at constant surface coverage). Fig. 13 shows the isosteric heat of adsorption as a function of amount of CO₂ adsorbed on MF-700. During the gradual covering of thermodynamically favourable adsorption sites, most favourable sites are filled first followed by filling of rest of the sites. This leads to decrease in isosteric heat of adsorption with increase in surface coverage, which further suggests the adsorbent surface to be energetically heterogeneous. Also adsorption process with heat of adsorption above 80 kJ mol⁻¹ indicates chemisorption process whereas the lower values indicate physiosorption. Average heat of adsorption is estimated to be 17.01 kJ mol⁻¹ showing the occurrence of physical adsorption process. Values between 16–30 kJ mol⁻¹ have been reported for adsorption on various carbon based adsorbents.^11,55,56

Fig. 13 Isosteric heat of adsorption of CO₂ on MF-700.

4. Conclusions

Nitrogen doped mesoporous carbons, prepared from melamine-formaldehyde resin by nanocasting technique, have been characterized for their structural, textural and chemical properties. Carbonization temperature is found to have significant effect on their physico–chemical properties. Elemental analysis, FTIR and XPS analysis confirmed the development of nitrogen doped mesoporous carbons with nitrogen species being present majorly in the form of pyridinic and pyrrolic/pyridonic nitrogen. CO₂ adsorption capacity of MF-700 decreased with increase in adsorption temperature but increased with CO₂ feed concentration. Decline in capacity with temperature is attributed to exothermic adsorption process. Furthermore these prepared carbons exhibited complete regeneration and stability for long-term use. Kinetic analysis of CO₂ adsorption on nitrogen doped mesoporous carbons was carried out by validating different adsorption kinetic models with experimental data of CO₂ adsorption measurements at various CO₂ feed concentrations. Fractional order kinetic model successfully predicted the dynamic adsorption of CO₂ as signified by error% values between experimental and predicted data within the range of 5%.

Acknowledgements

The financial support from Science and Engineering Research Board under Department of Science and Technology (DST), and All India Council of Technical Education (AICTE), New Delhi, is acknowledged. Chitrakshi Goel acknowledges the financial support from DST-INSPIRE under its Assured Opportunity for Research Careers (AORC) scheme.

References

1. B. P. Spigarelli and S. K. Kawatra, J. CO₂ Util., 2013, 1, 69.
2. Q. Wang, J. Luo, Z. Zhong and A. Borgna, Energy Environ. Sci., 2011, 4, 42.
3. K. M. K. Yu, I. Curcic, J. Gabriel and S. C. E. Tsang, ChemSusChem, 2008, 1, 893.
4. R. Thiruvenkatachari, S. Su, H. An and X. X. Yu, Prog. Energy Combust. Sci., 2009, 35, 438.
5. D. Aaron and C. Tsouris, Sep. Sci. Technol., 2005, 40, 321.
6. F.-Y. Chang, K.-J. Chao, H.-H. Cheng and C.-S. Tan, Sep. Purif. Technol., 2009, 70, 87.
7. J. Merel, M. Clausse and F. Meunier, Ind. Eng. Chem. Res., 2008, 47, 209.
8. R. V. Siriwardane, M.-S. Shen and E. P. Fisher, Energy Fuels, 2005, 19, 1153.
9. A. Wahby, J. M. Ramos-Fernández, M. Martínez-Escandell, A. Sepúlveda-Escribano, J. Silvestre-Albero and F. Rodríguez-Reinoso, ChemSusChem, 2010, 3, 974.
10. N. Sun, C. Sun, H. Liu, J. Liu, L. Stevens, T. Drage, C. E. Snape, K. Li, W. Wei and Y. Sun, Fuel, 2013, 113, 854.
11. S. Himeno, T. Komatsu and S. Fujita, J. Chem. Eng. Data, 2005, 50, 369.
12. A. R. Millward and O. M. Yaghi, J. Am. Chem. Soc., 2005, 127, 17998.
13. Q. Yang, C. Xue, C. Zhong and J.-F. Chen, AIChE J., 2007, 53, 2832.
14. K. S. N. Kamarudin and N. Alias, Fuel Process. Technol., 2013, 106, 332.
15. M. R. Melloa, D. Phanon, G. Q. Silveira, P. L. Llewellyn and C. M. Ronconi, Microporous Mesoporous Mater., 2011, 143, 174.
16. X. Wang, X. Ma, C. Song, D. R. Locke, S. Siefert, R. E. Winans, J. Möllmer, M. Lange, A. Möller and R. Gläser, Microporous Mesoporous Mater., 2013, 169, 103.
17. Y. Belmabkhout, R. Serna-Guerrero and A. Sayari, Ind. Eng. Chem. Res., 2010, 49, 359.
18. M. Sevilla and A. B. Fuertes, J. Colloid Interface Sci., 2012, 366, 147.
19. T. Valdés-Solís and A. B. Fuertes, Mater. Res. Bull., 2006, 41, 2187.
20. C. Zhao, W. Wang, Z. Yu, H. Zhang, A. Wang and Y. Yang, J. Mater. Chem., 2010, 20, 976.
21. M. L. Gray, Y. Soong, K. J. Champagne, J. Baltrus, R. W. Stevens, P. Toochinda and S. S. C. Chuang, Sep. Purif. Technol., 2004, 35, 31.
22. M. G. Plaza, C. Pevida, A. Arenillas, F. Rubiera and J. J. Pis, Fuel, 2007, 86, 2204.
23. D. Bezerra, R. Oliveira, R. Vieira, C. Cavalcante Jr and D. S. Azevedo, Adsorption, 2011, 17, 235.
24. C. Pevida, T. C. Drage and C. E. Snape, Carbon, 2008, 46, 1464.
25. Y. Huang, F. Yang, Z. Xu and J. Shen, J. Colloid Interface Sci., 2011, 363, 193.
26. A. Lu, A. Kiefer, W. Schmidt and F. Schüth, Chem. Mater., 2004, 16, 100.
27. G.-P. Hao, W.-C. Li, D. Qian and A.-H. Lu, Adv. Mater., 2010, 22, 853.
28. Y. Zhao, L. Zhao, K. X. Yao, Y. Yang, Q. Zhang and Y. Han, J. Mater. Chem., 2012, 22, 19726.
29. Z. Liu, Z. Du, H. Song, C. Wang, F. Subhan, W. Xing and Z. Yan, J. Colloid Interface Sci., 2014, 416, 124.
30. S. Lagergren, Ksver. Veterskapsakad. Handl., 1898, 24, 1.
31. Y. S. Ho and G. McKay, Process Biochem., 1999, 34, 451.
32. A. Zhao, A. Samanta, P. Sarkar and R. Gupta, Ind. Eng. Chem. Res., 2013, 52, 6480.
33. A. Heydari-Gorji and A. Sayari, Chem. Eng. J., 2011, 173, 72.
34. K. V. Kumar and S. Sivanesan, J. Hazard. Mater., 2006, 134, 277.
35. Y.-S. Ho, Water Res., 2006, 40, 119.
36. Y.-S. Ho and A. E. Ofomaja, J. Hazard. Mater., 2006, 137, 1796.
37. K. V. Kumar and S. Sivanesan, J. Hazard. Mater., 2005, 123, 288.
38. B. Guo, L. Chang and K. Xie, J. Nat. Gas Chem., 2006, 15, 223.
39. Y. Xia and R. Mokaya, Chem. Mater., 2005, 17, 1553.
40. A. Vinu, K. Ariga, T. Mori, T. Nakanishi, S. Hishita, D. Golberg and Y. Bando, Adv. Mater., 2005, 17, 1648.
41. A. Vinu, S. Anandan, C. Anand, P. Srinivasu, K. Ariga and T. Mori, Microporous Mesoporous Mater., 2008, 109, 398.
42. K. S. Lakhi, W. S. Cha, S. Joseph, B. J. Wood, S. S. Aldeyab, G. Lawrence, J.-H. Choy and A. Vinu, Catal. Today, 2015, 243, 209.
43. C. Chen, J. Kim and W.-S. Ahn, Fuel, 2012, 95, 360.
44. K. László, E. Tombácz and K. Josepovits, Carbon, 2001, 39, 1217.
45. L. Wang and R. T. Yang, J. Phys. Chem. C, 2012, 116, 1099.
46. J. R. Pels, F. Kapteijn, J. A. Moulijn, Q. Zhu and K. M. Thomas, Carbon, 1995, 33, 1641.
47. A. Olejniczak, M. Lezanska, J. Wloch, A. Kucinska and J. P. Lukaszewicz, J. Mater. Chem. A, 2013, 1, 8961.
48. C. Pevida, M. G. Plaza, B. Arias, J. Fermoso, F. Rubiera and J. J. Pis, Appl. Surf. Sci., 2008, 254, 7165.
49. B. Zhu, K. Li, J. Liu, H. Liu, C. Sun, C. E. Snape and Z. Guo, J. Mater. Chem. A, 2014, 2, 5481.
50. M. Sevilla, P. Valle-Vigón and A. B. Fuertes, Adv. Funct. Mater., 2011, 21, 2781.
51. J. Zhou, W. Li, Z. Zhang, W. Xing and S. Zhuo, RSC Adv., 2012, 2, 161.
52. M. Balsamo, T. Budinova, A. Erto, A. Lancia, B. Petrova, N. Petrov and B. Tsyntsarski, Sep. Purif. Technol., 2013, 116, 214.
53. J. A. Thote, K. S. Iyer, R. Chatti, N. K. Labhsetwar, R. B. Biniwale and S. S. Rayalu, Carbon, 2010, 48, 396.
54. M. M. Maroto-Valer, Z. Lu, Y. Zhang and Z. Tang, Waste Manag., 2008, 28, 2320.
55. I. A. A. C. Esteves, M. S. S. Lopes, P. M. C. Nunes and J. P. B. Mota, Sep. Purif. Technol., 2008, 62, 281.
56. M. D. Hornbostel, J. Bao, G. Krishnan, A. Nagar, I. Jayaweera, T. Kobayashi, A. Sanjurjo, J. Sweeney, D. Carruthers, M. A. Petruska and L. Dubois, Carbon, 2013, 56, 77.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra05684e

---