Mesoporous carbon nitride as a metal-free catalyst for the removal of aniline

Abstract

Mesoporous carbon nitride (MCN) was used for the first time as a thermal catalyst for the removal of aniline. While bulk carbon nitride (BCN) and mesoporous silica (MCM-41) did not show any catalytic activities, the MCN showed aniline removal (33%) at 398 K. It was revealed that the much higher activity of the MCN originated from the presence of mesoporosity and suitable adsorption sites for the reaction to occur. The nature of aniline adsorption on MCN was found to fit well to the Freundlich adsorption model. The evidence on the interactions between aniline and MCN was supported by fluorescence and FTIR spectroscopies. It was suggested that the interactions between MCN and aniline involved both weak π to π stacking and strong hydrogen bonding, which were important to initiate the catalytic removal of aniline.

---

Introduction

Aniline is a primary aromatic amine that is recognized to be an essential precursor in the production of isocyanates,¹ azo dyes^2,3 and rubbers.⁴ As a result, aniline effluents are widely generated and discharged into the environment. Unfortunately, aniline has been identified as a highly toxic amine-based compound. High exposure to aniline can cause damage to human's DNA.⁵ Hence, removal of aniline is critical to reduce the harmful effect of aniline to human health.

There are numerous reported approaches to remove aniline. Even though adsorption process is an efficient method that has been commonly used for the removal of aniline,^6–8 catalytic process would be among the best way to achieve complete removal of aniline. Generally, the employed catalyst must have good adsorption ability towards aniline so that the catalytic process can proceed efficiently. For such reason, various porous heterogeneous catalysts have been investigated for the catalytic removal of aniline, such as mesoporous MCM-41,⁹ activated carbons,¹⁰ mesoporous carbon xerogel¹¹ and multi-walled carbon nanotube.¹² Unfortunately, their catalytic performances were constrained by the adsorption sites on the materials, which could only provide weak electrostatic interactions with the aniline, thus affecting the catalytic activity.^9–12 In order to enhance the adsorption and improve the activity, loading of transition metal (copper)⁹ or noble metal (platinum)^10–12 was widely suggested. The later would cause the high cost of the catalysts, which should be taken into consideration when applying the catalysts. Moreover, some stable organic intermediates, such as azo compounds were detected after the catalytic reaction on the aforementioned catalysts.^9–11 Therefore, development of active metal-free catalysts having good adsorption capability and without forming the stable intermediates at the end of the reaction is highly demanded.

One of the best candidates for such sustainable and active metal-free catalysts is mesoporous carbon nitride (MCN). The MCN was reported able to adsorb toxic aniline⁶ and N-nitrosopyrrolidine (NPYR).¹³ Apart from being an adsorbent, previous studies showed that the MCN has excellent catalytic activity for transesterification of keto ester,¹⁴ cyclisation of functional nitriles and alkynes¹⁵ and Friedel–Craft reaction of benzene.¹⁶ The carbon nitride (CN) was also particularly reported acting as thermal catalyst to decompose nitrogen monoxide (NO).¹⁷

In this study, the MCN was used for the first time as the catalyst for removal of aniline. Various catalytic parameters were examined in terms of the effect of reaction temperature, reaction time and amount of catalyst loading in the catalytic reaction. For comparison purpose, the activity of the MCN was compared to the bulk carbon nitride (BCN) and the mesoporous silica MCM-41. Moreover, the adsorption sites and its interactions with aniline were further clarified and discussed in this work.

Experimental

Preparation of materials

The BCN and MCN were synthesized utilizing thermal polymerization approach as previously reported.^{6,13,15,16,18} Pale yellow powder of BCN was obtained when the cyanamide precursor was calcined at a rate of 2.2 K min⁻¹ to reach the temperature of 823 K and it was held for 4 h. The MCN was synthesized by addition of an appropriate amount of cyanamide to the colloidal silica solution where mass ratio was fixed to 1. The mixture was subsequently stirred at 353 K until dried. The resulted white powder was calcined at a rate of 2.2 K min⁻¹ to reach the temperature of 823 K, and was tempered at this temperature for 4 h. The yellow calcined powder was dispersed in ammonium hydrogen difluoride solution (4 M) for 3 h. The powder was filtered and washed with double distilled water and ethanol for several times, followed by drying at 353 K in an oven overnight.

In this study, MCM-41 was synthesized by hydrothermal method.^7,9,19,20 The molar composition of tetraethylorthosilicate, sodium hydroxide, cetyltrimethylammonium bromide, ammonium hydroxide and distilled water was thereby fixed to 6, 3, 1, 0.3, and 250, respectively. The sodium silicate solution was prepared by mixing tetraethylorthosilicate and sodium hydroxide in water. The mixture was placed in a propylene bottle and stirred at 353 K for 2 h. Meanwhile, a template solution of cetyltrimethylammonium bromide and ammonium hydroxide was mixed in another bottle and stirred at 353 K for 1 hour. The prepared sodium silicate solution was then added simultaneously to the template solution. The mixture was aged at 370 K for 4 days. The pH of the mixture was monitored each day and adjusted to 10.2 by using a few drops of acetic acid. The obtained precipitate was filtered and washed with distilled water. The precipitate was later dried in an oven for another 24 h, followed by calcination in air at 823 K for 10 h.

Characterizations of materials

The X-ray diffraction (XRD) patterns were recorded on a Bruker D8 Advance diffractometer at room temperature, using Cu-Kα radiation where λ is 1.5406 Å at 40 kV and 40 mA. The Brunauer–Emmett–Teller (BET) specific surface area and Barret–Joyner–Halenda (BJH) pore size distributions of the MCN and MCM-41 were determined from nitrogen adsorption–desorption isotherms obtained at 77 K by using a Quantachrome Autosorb-1 instrument. The thermogravimetric analyser (TGA) profile was performed at a heating rate of 10 K min⁻¹ under nitrogen atmosphere on a Mettler TGA/SDTA 851°.

The CO₂-temperature programmed desorption (CO₂-TPD) profiles were measured on a Micromeritics AutoChem II chemosorption analyser. Before the CO₂ adsorption, the sample was heated at 393 K under He gas flow (30 mL min⁻¹) for 1 h. The sample was then cooled to 353 K and purged with He gas (30 mL min⁻¹) for 2 h. Desorption of CO₂ was measured in the range of 323–573 K with heating rate of 10 K min⁻¹.

Thermal catalytic removal of aniline

Thermal catalytic removal of aniline was carried out in a 100 mL round bottom flask connected with a gas sampling bag and a Liebig condenser with water circulation to ensure the reaction occurred in a closed reactor within the monitored temperature. Typically, 0.2 g of catalyst was added to 10 mL of aniline in acetonitrile solution (100 mg L⁻¹). Three small boiling chips were added to the mixture. The mixture was then stirred and heated at 398 K for 24 h. The residue of the aniline solution was analysed by gas chromatography-flame ionization detector (GC-FID, Shimadzu 2014) to determine the amount of aniline removed after the reaction. The gas product was analysed by gas chromatography-thermal conductivity detector (GC TCD, Shimadzu 2014). Various parameters were investigated for the removal of aniline on MCN, which were the catalyst loading amounts (0.1–0.4 g), the reaction temperatures (298–398 K), and reaction times (6–48 h) under the similar experimental procedure. For reusability test, the MCN catalyst was used for five cycles. The MCN was separated from the solution after the first reaction. The collected MCN was washed with ethanol and dried in an oven at 353 K overnight. The MCN was then reused as the catalyst under the similar conditions.

Adsorption and interactions studies

Adsorption processes of aniline on the BCN, the MCN and the MCM-41 were performed at room temperature for 24 h. The adsorbent (0.2 g) was added to 10 mL of aniline in acetonitrile solution (100 mg L⁻¹) for each process. The residue of aniline solution was analysed by the GC-FID after 24 h. The amount of aniline adsorbed on the adsorbents (Q_e) was calculated according to the following eqn (1).where V (L) represents the volume of aniline solution used for the adsorption process, m (g) is denoted as the amount of adsorbent used for the adsorption process, while C₀ (mg L⁻¹) and C_e (mg L⁻¹) are the initial and equilibrium concentrations of aniline, respectively.

The interactions between aniline and the adsorption sites of the MCN were explored in this work. Various concentrations of aniline (0.017–0.043 μmol) were added on the MCN at room temperature by measurements of the excitation and emission spectra of the MCN by using a fluorescence spectroscopy (JASCO, FP-8500). The emission wavelength at 463 nm was measured by monitoring at two excitation wavelengths (275 and 370 nm). The changes on the emission spectra of the MCN were then measured at both excitation wavelengths, subsequently.

The type of interactions between aniline and MCN was determined by Fourier transform infrared spectroscopy (FTIR, Nicolet iS50). The aniline spectrum was acquired by dropping the aniline solution on the sodium chloride pellet. On the other hand, the MCN spectrum was measured after mixing the MCN with potassium bromide. The FTIR spectra of the MCN after the addition of various concentrations of aniline solution (0.017 and 0.043 μmol) were also recorded. Similar experiments were also carried out using the BCN and the MCM-41.

Results and discussion

Characterization of materials

XRD patterns were recorded to study the structural properties of the synthesized materials. Fig. 1 shows the wide angle of XRD patterns for the BCN and the MCN, and small angle for the MCM-41 (inset). As shown in the Fig. 1, the synthesized BCN and MCN have two clear diffraction peaks at 2θ of 13.10 and 27.30°.^6,13,18 Both materials have a graphite-like structure with an intense type of interlayer stacking peak (002) at 2θ of 27.30°, corresponding to an interlayer distance of 0.33 nm, while the weak peak immersed at 2θ of 13.10 was denoted as in-planar repeating units in distance of 0.68 nm. It can be observed that the BCN exhibited higher intensity than the MCN. The stronger peak intensities indicated that the BCN has a better crystalline structure than the MCN. The reduced graphitic order of the MCN might be due to introduction of porosity into the structure and formation of defects during the synthesis process. On the other hand, MCM-41 (inset) synthesized using hydrothermal approach showed an intense peak at 2θ of 2.40° and three weak peaks at 2θ of 4.02, 4.58, and 6.22°, which can be indexed as (100), (110), (200), and (210) planes.^7,9,19,21 The most intense diffraction peak at 2θ of 2.40° reflected the short range hexagonal ordering of the MCM-41, while the other three weak peaks represented quasi-regular arrangement in hexagonal symmetry. All these peaks positions agreed well with the previously reported peaks for BCN, MCN, and MCM-41, suggesting the successful preparation of all the materials.^{6,7,9,18,19,21}

Fig. 1 XRD patterns of the (a) BCN (b) MCN and MCM-41 (inset).

The BET specific surface areas of the materials and mesoscale textural properties of the MCN and the MCM-41 were obtained by nitrogen adsorption–desorption measurements. Fig. 2 illustrates the isotherms and pore size distributions of the MCN and the MCM-41. From the Fig. 2[A], it was disclosed that both MCN and MCM-41 exhibited typical type IV isotherms.^6,13,16,18 It was observed that the MCN has larger capillary condensation than the MCM-41 due to the presence of combination of type H1 and H3 hysteresis loop with slit-shape pores present in the structure. On the other hand, the MCM-41 has narrow type H1 hysteresis loop with characteristic pores of cylindrical geometry. BJH pore size distributions are shown in Fig. 2[B]. The MCN and the MCM-41 have average pore sizes of around 7 and 3 nm, respectively. It was apparent that both of the MCN and the MCM-41 have narrow and sharp pores distribution, indicating the uniformity of the pores distribution. However, bimodal pore distribution could be observed for MCN with minor fraction of pores larger than 7 nm as shown in Fig. 2[B](a). The BCN possesses the lowest specific surface area (11 m² g⁻¹), followed by the MCN (287 m² g⁻¹) and the MCM-41 (1012 m² g⁻¹). The difference of specific surface area among the three materials was due to introduction of mesoporosity into the structure and the different pore diameter of the materials. Owing to these reasons, the BCN has the lowest specific surface area while the MCM-41 possessed the highest surface area.

Fig. 2 [A] Nitrogen adsorption–desorption isotherms and [B] the BJH pore size distributions of the (a) MCN and (b) MCM-41.

A thermal catalyst should possess high thermal stability property at the reaction temperature. Hence, thermal stability of the MCN was investigated using a thermogravimetric analyser (TGA). From Fig. 3(a), it can be seen that there was only 7% of weight losses observed below the temperature of 427 K, which might be due to the loss of adsorbed moisture in the pores and or on the surface of the MCN. The MCN has thermal stability up to around 700 K. The low thermal stability might suggest the incomplete condensation of tri-s-triazine units of MCN, generating minor C=N or N=C=N groups during polymerisation process, which affected the thermal stability of MCN. Since the complete decomposition occurred at 880 K, it was confirmed that the silica template used during the synthesis process was completely removed after washing with ammonium hydrogen difluoride solution. In contrast, the MCM-41 has the larger mass loss of 25% than the MCN below 380 K as shown in Fig. 3(b), reflecting the presence of water adsorbed on the silanol groups of MCM-41. Meanwhile, MCM-41 was reported to act with much higher thermal stability, which was over 1000 K.²⁰ Although the MCN has lower thermal stability than the MCM-41, it would not affect its application as thermal catalyst in this study since the applied temperature was considered to be low (298–398 K).

Fig. 3 The TGA thermograms of the (a) MCN and (b) MCM-41.

Catalytic removal of aniline

For the catalytic removal of aniline, the reactions were conducted at the initial concentration of 100 mg L⁻¹ over the BCN, the MCN and the MCM-41 at 398 K for 24 h. It is worthy noted that only the MCN successfully removed 33 mg L⁻¹ of aniline, while BCN and MCM-41 did not show any catalytic performances towards aniline removal as shown in Fig. S1.† It was confirmed that there was no reaction occurred in the absence of catalyst, indicating that aniline was stable and could not be decomposed thermally at the temperature of 398 K.

Since only MCN showed activity for catalytic removal of aniline, the effect of catalyst loading on the removal of aniline was investigated using the MCN in this study. The experiment was conducted using various amounts of the MCN (0.1–0.4 g), while the initial concentration of aniline, temperature and reaction time were 100 mg L⁻¹, 398 K and 36 h, respectively. The long reaction time (36 h) was used here to obtain clearer results. The amount of removed aniline increased when the catalyst loading increased up to 0.2 g (Fig. S2†). The increase in the amount of the removed aniline would be a result of the increase in the amount of active sites on the MCN. The increase of the active sites amount has led to the increase of the interaction with the aniline molecules and thus, promoted the removal of aniline. Further loading of MCN seems to have less influence in the removal of aniline and saturation was observed when 0.4 g of MCN was loaded into the system. This result suggested that the loading of 0.2 to 0.3 g of catalyst was sufficient for the catalytic removal of aniline in this system.

The effect of reaction temperature towards removal of aniline at various reaction temperatures (298–398 K) over 0.2 g of the MCN loading in 10 mL of aniline solution (100 mg L⁻¹) for 24 h was studied in order to distinguish the catalytic reaction from the adsorption process and reveal the activation energy required for the reaction. Fig. 4 demonstrates the removal of aniline on the MCN at various reaction temperatures. There was only 6.60 mg L⁻¹ of aniline was removed when the reaction temperature was low (298–335 K). This value was very close to the amount of aniline adsorbed on the MCN, indicating that the adsorption would be the main process occurred at the temperature range of 298–335 K. This result also confirmed that aniline was stable and could not be decomposed thermally at this temperature range. The catalytic activity was observable when the reaction temperature was 348 K or above, suggesting that the reaction temperature of 348 K was able to provide good heat transfer from the MCN to the adsorbed aniline, thus, overcome the activation energy required for breaking of bonds. The increase in the reaction temperature (348–398 K) was found to promote the catalytic removal of aniline, indicating that reaction would occur endothermically and the reaction temperature played an important factor for the removal of aniline.

Fig. 4 Effect of reaction temperature and the Arrhenius plot (inset) for the removal of aniline.

The activation energy (E_a) required for the catalytic removal process was revealed by applying the Arrhenius equation, which is commonly represented as eqn (2).²²

where k is rate constant (s⁻¹), which was calculated based on eqn (3).where A is a pre-exponential factor, E_a is activation energy (kJ mol⁻¹), R is universal gas constant (0.008314 kJ mol⁻¹ K⁻¹), and T is the temperature in Kelvin (K).

The logarithmic form of the Arrhenius equation can be transformed as eqn (4), which was also plotted in Fig. 4.

Fig. 4 (inset) depicts the Arrhenius plot of −ln k versus 1/T. The data fitted well the Arrhenius plot with R value of 0.997. From the slope of the best fit line, the E_a required for the catalytic removal of aniline was determined to be 28.4 kJ mol⁻¹. This value was found to be smaller as compared to the E_a value obtained for the removal of aniline on multi-walled carbon nanotube (36.8 kJ mol⁻¹) at the temperature range of 298–323 K.⁸ This result suggested that MCN would be a potential catalyst for the catalytic removal of aniline.

It was confirmed that after the thermal catalytic reaction the solution was colourless, indicating that there was no azo intermediate compounds formation in the catalytic removal of aniline. These results were in contrast with the works reported by other groups where the final solution obtained has an intense orange colour due to the formation of azo compounds such as quinone derivatives and azoxybenzene.^9–11 Besides, it was confirmed that there was no detection of any organic intermediates by GC-FID. Based on the GC-TCD, CO₂ was detected as the product of the reaction. The ratio of the produced CO₂ to the calculated one based on the amount of removed aniline was 0.96. Based on these results, it can be suggested that the MCN is a highly potential thermal catalyst to oxidize the aniline without the formation of any organic intermediates at the end of the reaction.

Reusability test was carried out to support the potential application of the MCN by reusing the MCN catalyst for the catalytic removal of aniline at 398 K for 24 h. After five cycles, it was obtained that the reused MCN still showed similar level of catalytic activity as shown in Fig. S3.† This was also supported by the FTIR spectra that showed the same functional groups of the MCN before and after reaction (Fig. S4†). Moreover, there were no remarkable new peaks due to the formation of any organic intermediates. These results showed the stability of the MCN as the catalyst for the oxidation of the aniline.

Kinetic study was investigated in order to enhance the understanding of the factors that might influence the rate of the reaction. The study was carried out where 0.2 g of MCN was added into 10 mL of aniline (100 mg L⁻¹) and heated at 398 K for different reaction times (6–48 h). The reaction times have a significant effect on the oxidation of aniline (Fig. S5†). The oxidation of aniline increased almost linearly when the reaction time increased, suggesting the reaction occurred catalytically. There was 54 mg L⁻¹ of aniline has been successfully removed after 48 h.

The kinetic order on the catalytic removal of aniline was reinforced by plotting the first-order kinetics plot of ln C versus time (t) based rate law in the eqn (5).

Upon integration, eqn (6) was obtained.

where C is denoted as equilibrium concentration of aniline in solution, C₀ is the initial concentration of aniline (100 mg L⁻¹), k is rate constant (h⁻¹) and t is reaction time (h).

Plotting the ln C versus t gave a straight line with linear regression coefficient of 0.997 (inset of Fig. S5†), indicating that the data of thermal catalytic removal of aniline on MCN fitted well the first-order kinetic reaction. This result suggested that the catalytic removal of aniline is a unimolecular reaction and the rate depends directly to the initial concentration of aniline in the reaction. The rate constant obtained from the slope of the plot was determined to be 0.02 h⁻¹.

Adsorption of aniline

It has been generally accepted that a catalytic reaction would occur after an adsorption process occurs. In order to understand the parameters affecting the high activity of the MCN, several adsorption tests were carried out on the BCN, the MCN, and the MCM-41 under the same experiment conditions. The adsorption of aniline (100 mg L⁻¹) was carried out at room temperature after stirring for 24 h. The adsorption of aniline on these samples is also shown in Fig. S1.† Among the examined adsorbents, the MCN exhibited the highest adsorption capability. The MCN successfully adsorbed 6.60 mg L⁻¹ of aniline, which was three times higher than that adsorbed on the BCN. This might be related to the porosity in the MCN structure that could provide more efficient diffusion of aniline molecules and thus, enhance the adsorption process on the MCN. It is worthy to highlight here that despite of a larger surface area of the MCM-41 (1012 m² g⁻¹), the MCM-41 did not show good adsorption. This result clearly suggested that the surface area was not the only decisive factor for the adsorption process. It can be therefore suggested that the MCN might have favourable adsorption sites for the adsorption of aniline.

Langmuir and Freundlich adsorption models were implemented in this study to investigate the adsorption nature of aniline on the MCN. The linear forms of the Langmuir and Freundlich adsorption equations are presented in eqn (7) and (8), respectively.²³

where Q_e is derived as equilibrium adsorption capacity of aniline on the MCN (mg g⁻¹), Q_m (mg g⁻¹) is the maximum monolayer adsorption capacity of MCN and C_e is the equilibrium aniline concentration in solution (mg L⁻¹). The Langmuir and Freundlich adsorption constants represented with K_L (L mg⁻¹) and K_F (mg g⁻¹) are ascribed as the free energy of adsorption and adsorption capacity of the MCN, respectively. The 1/n is denoted as the nature of the adsorption, which can be determined from the reverse value of the slope of Freundlich plot. If a value for n is above unity, the adsorption is a physical process and when the value of n is below unity, it is attributed as a chemical process.

Fig. 5(a) and (b) show the Langmuir and Freundlich plots of aniline adsorption on the MCN, respectively. The linear regression coefficient (R) of Langmuir plot was only 0.943, while the Freundlich plot was 0.993. The Langmuir adsorption model could not explain the adsorption of aniline on MCN owing to the low linearity of Langmuir plot and the negative value of adsorption capacity (Q_m). In contrast, good linearity plot obtained for Freundlich model indicated that the adsorption isotherm of aniline on MCN could be explained by using the Freundlich model, where the adsorption of aniline occurred in the heterogeneous adsorption sites on the MCN. From the n value calculated from the reverse value of the slope in the Freundlich plot, the nature of aniline adsorption on the MCN was determined to be chemical interaction since the n value was below the unity, which was 0.447.

Fig. 5 (a) Langmuir and (b) linear Freundlich plots of the MCN.

Quenching tests were performed in this work to study the interactions between aniline and adsorption sites of the MCN by introducing various concentrations of aniline solution (0.017–0.043 μmol) to the MCN and analysed using fluorescence spectroscopy. The quenching effect of aniline on the MCN was investigated by monitoring the changes in the emission intensities of the MCN in the absence and presence of aniline molecules. From the fluorescence spectra as shown in Fig. 6, the MCN was confirmed to possess two maximum excitation peaks at 275 and 370 nm, which was an analogy to the DR UV-Vis spectrum as reported previously.¹³ An emission spectrum of the MCN at 463 nm was obtained when it was monitored at excitation wavelengths of 275 or 370 nm. The former excitation wavelength was originated from the excitation of electrons in the N=C groups (π to π*) in the aromatic 1,3,5-triazine, while the latter one was due to the excitation of electrons in the terminal N–C groups (n to π*) of the MCN. Fig. 6(a) and (b) show the changes in the emission spectra of the MCN after addition of various concentrations of aniline (0.017–0.043 μmol) monitored at excitation wavelengths of 275 and 370 nm, respectively. The emission intensities of the MCN were found to be decreased when aniline was added on the MCN. These results indicated that both emission sites on the MCN, which were excited at 275 and 370 nm, interacted with the aniline molecules and it led to the decrease in the emission intensities. The emission intensities spectra of MCN were found to be decreased as the concentration of aniline added onto the MCN increased up to 0.043 μmol. Further increase in aniline concentration did not decrease the emission intensities, indicating that there was only limited amount of active sites on the MCN that could interact with the aniline molecules.

Fig. 6 Changes in the emission intensity of the MCN monitored at excitation wavelengths of (a) 275 nm and (b) 370 nm with the addition of various concentrations of aniline.

Recovery test was also conducted by observing the emission spectra of the MCN after the addition of 0.043 μmol of aniline. The emission intensities of the MCN monitored at both excitation wavelengths were recovered up to 95% of the initial intensity after 5 minutes exposure in the air, indicating that the main interactions between aniline and the MCN were weak, such as π to π stacking interactions between the aromatic ring of aniline with the conjugated 1,3,5-tri-s-triazine ring of the MCN. The recovered emission intensity of MCN (95%) was unchanged even after 30 minutes exposure in the air. This result suggested that some small parts (5%) of the interactions between aniline and MCN might also involve a stronger bonding, for instance, hydrogen bonding. However, the full recovery of the intensity can be achieved by heating the MCN at 323 K for 5 minutes.

It was shown that the aniline molecules interacted with the MCN at both excitation wavelengths at 275 and 370 nm. The quenching efficiency was determined using Stern–Volmer plots by plotting the relative emission intensity to the aniline concentration monitored at both excitation wavelengths. When constant illumination intensity and the same amount of MCN were taken into account, the relative emission intensity was expressed as a function of aniline concentration based on the Stern–Volmer equation shown at eqn (9).²⁴

where I₀ and I are fluorescence intensities observed in the absence and presence of aniline, respectively, Q (μmol) is the aniline concentration, and K_SV is the Stern–Volmer quenching constant.

Fig. 7 represents the Stern–Volmer linear plots of the relative emission intensity of MCN with the addition of increasing aniline concentrations monitored at 275 and 370 nm excitation wavelengths. Linear relationship between relative emission intensity and the aniline concentration in the range of 0.017–0.043 μmol could be observed, suggesting that the MCN has good interactions with aniline. The quenching efficiency of aniline on the emission sites of the MCN can be determined via quenching rate constant from the slope value. It was found that the quenching rate constant for the emission sites monitored at 370 nm was only slightly higher (K_SV = 5.88 μmol⁻¹) than that of emission sites monitored at 275 nm (K_SV = 4.96 μmol⁻¹). This result indicated that aniline molecules interacted almost equally with both N=C groups in the aromatic 1,3,5-triazine and terminal N–C groups on the MCN.

Fig. 7 Stern–Volmer plots between relative emission intensities of the MCN and increased concentrations of aniline monitored at excitation wavelenghts of (a) 275 (b) 370 nm.

The recovery test and Freundlich plot showed that the interaction of aniline on the MCN might involve strong bonding such as hydrogen bonding. Hence, FTIR analysis was undertaken in this study to investigate the interactions between the aniline and the MCN. Fig. 8 shows the FTIR spectra of aniline solution, the MCN, and the MCN with addition of aniline solution (0.017 and 0.043 μmol). A shown in Fig. 8(a), the absorption peaks of aniline in acetonitrile solution can be measured in the region of 750–3530 cm⁻¹. The typical stretching vibration of N–H bond in a primary amine of aniline can be clearly seen at 3542 and 3618 cm⁻¹. On the other hand, absorption peaks for the MCN shown in Fig. 8(b) can be observed in the 810–1630 cm⁻¹ region, which was consistent with the reported graphitic CN single and double bond characters.^6,13,16,25 The peak at 810 cm⁻¹ was assigned to 1,3,5-substituted aromatic rings, while the peaks at 1253 cm⁻¹, 1329 cm⁻¹ and 1430 cm⁻¹ were denoted as C–N heterocycles. The C=N conjugated peak was appeared at 1630 cm⁻¹. There was a small absorbance at 2176 cm⁻¹, confirming that minor C=N or N=C=N groups were generated during the polymerization process. This result suggested the incomplete condensation process as mentioned above. The MCN also showed a broad band at 3330 cm⁻¹, which can be ascribed to the stretching vibrations of NH₂ or NH groups. It was obvious that absorption peaks of aniline can be observed after the addition of 0.017 μmol of aniline in the region of 750–3530 cm⁻¹, as shown in Fig. 8(c). In Fig. 8(d), it could be observed that the aniline peaks became more intense with further increase in the concentration of aniline to 0.043 μmol. A sharp peak above 3500 cm⁻¹ can be observed, which can be assigned as N–H bond. This additional peak might be due to the NH₂ group stretching originated from the aniline molecules, and/or formation of new N–H bonds.

Fig. 8 FTIR spectra of (a) aniline solution, (b) MCN, (c) MCN with 0.017 μmol, and (d) MCN with 0.043 μmol of aniline solution.

Similar results were also observed for the BCN and the MCM-41 as shown in Fig. S6.† For comparison purpose, the ratio for the improved absorption intensity around 3618 cm⁻¹ on all the samples after addition of the aniline solution is also shown in Table S1.† It was obvious that the ratio of peak intensity for the N–H bonds increased in the order of the MCM-41 < the BCN < the MCN. This result is in good agreement with the aniline adsorption tests discussed previously.

In order to further confirm the active sites of the catalysts, the TPD profiles for the desorbed CO₂ are shown in Fig. 9 and quantified in Table 1. Both the BCN and the MCN exhibited two desorption peaks with total quantities of desorbed CO₂ were 1.92 and 1.95 mmol g⁻¹, respectively. The similar TPD profiles and the desorbed CO₂ values showed that the BCN and the MCN have similar basic sites, in which the MCN showed slightly larger amount of basic sites. Despite of the existence of the basic sites on the BCN, the BCN showed poor adsorption capability, which in turn caused the BCN did not show activity for removal of aniline under the investigated conditions. The poor adsorption of the CN without porosity was also reported previously for the adsorption of Rhodamine B.²⁶

Fig. 9 CO₂-TPD profiles of (a) BCN, (b) MCN, and (c) MCM-41.

Table 1 Distribution of CO₂ desorption peaks and their quantities on the BCN, the MCN, and the MCM-41

Sample	Temperature at maximum (K)	Quantity (mmol g^−1)	Total quantity (mmol g^−1)
BCN	377.9	0.76	1.92
	503.0	1.16
MCN	377.2	0.77	1.95
	488.2	1.18
MCM-41	352.4	0.03	0.45
	512.1	0.42

---

Meanwhile, it was obvious that the MCM-41 gave lower total quantity of desorbed CO₂ (0.45 mmol g⁻¹) than the BCN and the MCN. These results suggested that the MCN has the largest amount of active sites to interact with the aniline molecules, while the MCM-41 has the least amount, in good agreement with the adsorption results shown above.

Owing to the strong interaction between the aniline and MCN as supported by the Freundlich adsorption model and quenching results, it can be suggested that hydrogen bonding might be formed between the nitrogen atom from the N=C and terminal N–C groups of MCN and hydrogen atoms from NH₂ groups of aniline molecules. The proposed interactions both via hydrogen bonding and π to π stacking are illustrated in Fig. 10.

Fig. 10 The proposed interactions of N=C and terminal N–C groups on the MCN with aniline molecules.

Conclusions

Under the same reaction conditions, the MCN gave 33% removal of aniline at 398 K, while the BCN and the MCM-41 did not show any catalytic activities. The relatively low reaction temperature at 398 K was able to overcome the activation energy required for breaking the bonds of the aniline. It was confirmed that the MCN gave no formation of any organic intermediates. The outstanding catalytic performance of the MCN would be attributable to the high adsorption of aniline on the MCN, which facilitated the catalytic reaction. It was revealed that the presence of both mesoporosity and favourable N=C and N–C groups adsorption sites on the MCN gave stronger interactions with the aromatic aniline as compared to the BCN and the MCM-41.

Acknowledgements

This work was financially supported by the Ministry of Higher Education (MOHE, Malaysia) and the Universiti Teknologi Malaysia (UTM, Malaysia) through a Flagship Research University Grant (cost center code: Q.J130000.2426.00G07) and partially supported by the Ministry of Science, Technology and Innovation (MOSTI, Malaysia) through a Nanotechnology Research Grant Scheme (Nano Fund, cost centre code: R.J130000.7926.4H007). M.S.S acknowledged financial support from Mybrain MyPhd scholarship.

Notes and references

1. H. J. Twitchett, Chem. Soc. Rev., 1974, 3, 209–230.
2. C. L. Gáspár, I. Bâldea and I. Panea, Dyes Pigm., 2006, 69, 45–53.
3. R. Zhao, C. Tan, Y. Xie, C. Gao, H. Liu and Y. Jiang, Tetrahedron Lett., 2011, 52, 3805–3809.
4. B. G. Soares, G. S. Amorim, F. G. Souza Jr, M. G. Oliveira and J. E. Pereira da Silva, Synth. Met., 2006, 156, 91–98.
5. H. Ma, J. Wang, S. Z. Abdel-Rahman, P. J. Boor and M. Firoze Khan, Toxicol. Appl. Pharmacol., 2008, 233, 247–253.
6. M. S. Sam, H. O. Lintang, M. M. Sanagi, S. L. Lee and L. Yuliati, Adv. Mater. Res., 2014, 925, 135–139.
7. X. Yang, Q. Guan and W. Li, J. Environ. Manage., 2011, 92, 2939–2943.
8. H. Al-Johani and M. A. Salam, J. Colloid Interface Sci., 2011, 360, 760–767.
9. H. T. Gomes, P. Selvam, S. E. Dapurkar, J. L. Figueiredo and J. L. Faria, Microporous Mesoporous Mater., 2005, 86, 287–294.
10. S. Morales-Torres, A. M. T. Silva, F. J. Maldonado-Hodar, B. F. Machado, A. F. Perez-Cadenas, J. L. Faria, J. L. Figueiredo and F. Carrasco-Marin, Appl. Catal., B, 2011, 105, 86–94.
11. H. T. Gomes, P. V. Samant, P. Serp, P. Kalck, J. L. Figueiredo and J. L. Faria, Appl. Catal., B, 2004, 54, 175–182.
12. J. Garcia, H. T. Gomes, P. Serp, P. Kalck, J. L. Figueiredo and J. L. Faria, Catal. Today, 2005, 102, 101–109.
13. M. S. Sam, H. O. Lintang, M. M. Sanagi, S. L. Lee and L. Yuliati, Spectrochim. Acta, Part A, 2014, 124, 357–364.
14. X. Jin, V. V. Balasubramanian, S. T. Selvan, D. P. Sawant, M. A. Chari, G. Q. Lu and A. Vinu, Angew. Chem., Int. Ed., 2009, 48, 7884–7887.
15. F. Goettmann, A. Fischer, M. Antonietti and A. Thomas, New J. Chem., 2007, 31, 1455–1460.
16. F. Goettmann, A. Fischer, M. Antonietti and A. Thomas, Angew. Chem., Int. Ed., 2006, 45, 4467–4471.
17. J. Zhu, P. Xiao, H. Li and S. A. C. Carabineiro, ACS Appl. Mater. Interfaces, 2014, 6, 16449–16465.
18. A. Thomas, A. Fischer, F. Goettmann, M. Antonietti, J. O. Muller, R. Schlogl and J. M. Carlsson, J. Mater. Chem., 2008, 18, 4893–4908.
19. A. S. Araujo, M. B. Souza, A. S. Silva, A. G. Pedrosa, J. F. B. Aquino and A. S. L. S. Coutinho, Adsorption, 2005, 11, 181–186.
20. A. S. Araujo and M. Jaroniec, Thermochim. Acta, 2000, 363, 175–180.
21. S. Jana, B. Dutta, H. Honda and S. Koner, Appl. Clay Sci., 2011, 54, 138–143.
22. A. K. Galwey and M. E. Brown, Thermochim. Acta, 2002, 386, 91–98.
23. U. F. Alkaram, A. A. Mukhlis and A. H. Al-Dujaili, J. Hazard. Mater., 2009, 169, 324–332.
24. E. Z. Lee, S. U. Lee, N. S. Heo, G. D. Stucky, Y. S. Jun and W. H. Hong, Chem. Commun., 2012, 48, 3942–3944.
25. Y. Li, J. Zhang, Q. Wang, Y. Jin, D. Huang, Q. Cui and G. Zou, J. Phys. Chem. B, 2010, 114, 9429–9434.
26. T. Wang, K. Kailasam, P. Xiao, G. Chen, L. Chen, L. Wang, J. Li and J. Zhu, Microporous Mesoporous Mater., 2014, 187, 63–70.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra04829j

---