Effect of Polarity of Activated Carbon Surface, Solvent and Adsorbate on Adsorption of Aromatic Compounds from Liquid Phase

Abstract

In this study, introduction of acidic functional groups onto a carbon surface and their removal were carried out through two oxidation methods and outgassing to investigate the adsorption mechanism of aromatic compounds which have different polarity (benzene and nitrobenzene). Adsorption experiments for these aromatics in aqueous solution and n-hexane solution were conducted in order to obtain the adsorption isotherms for commercial activated carbon (BAC) as a starting material, its two types of oxidized BAC samples (OXs), and their outgassed samples at 900°C (OGs). Adsorption and desorption kinetics of nitrobenzene for the BAC, OXs and OGs in aqueous solution were also examined. The results showed that the adsorption of benzene molecules was significantly hindered by abundant acidic functional groups in aqueous solution, whereas the adsorbed amount of nitrobenzene on OXs gradually increased as the solution concentration increased, indicating that nitrobenzene can adsorb favourably on a hydrophilic surface due to its high dipole moment, in contrast to benzene. In n-hexane solution, it was difficult for benzene to adsorb on any sample owing to the high affinity between benzene and n-hexane solvent. On the other hand, adsorbed amounts of nitrobenzene on OXs were larger than those of OGs in n-hexane solution, implying that nitrobenzene can adsorb two adsorption sites, graphene layers and surface acidic functional groups. The observed adsorption and desorption rate constants of nitrobenzene on the OXs were lower than those on the BAC due to disturbance of diffusion by the acidic functional groups.

In the past, numerous studies for the adsorption of aromatic compounds onto activated carbon have been conducted. It is recognized that adsorption mechanisms depend on various factors including the pore structure¹⁾ and surface chemistry such as the presence of acidic functional groups²⁾ on activated carbons as well as polarity³⁾ and solubility⁴⁾ of solvents and adsorbates, respectively. Among these factors, the polarity of carbon surface is considered to significantly influence the adsorption of organic pollutant such as aromatic compounds.⁵⁾ However, it is not well understood how the surface polarity affects the adsorption of aromatic compounds. Although the activated carbon surface is generally hydrophobic and has no polarity, acidic functional groups could be introduced to the carbon surface by oxidation to give polarity on the surface.^6,7) Acidic functional groups on the carbon surface would drastically improve the adsorption ability for cationic solutes such as lead(II) and cadmium(II) ions.^8,9) It is also reported that aromatic compounds can adsorb onto acidic functional groups.¹⁰⁾ Therefore, it is assumed that adsorption ability for aromatic compounds might improve in the presence of acidic functional groups.

In this study, various activated carbons possessing different pore structure and amount of acidic functional groups were prepared through by several oxidation methods and adsorption experiments of benzene and nitrobenzene in aqueous solutions and n-hexane solutions were conducted for each prepared activated carbon. Adsorption and desorption kinetics in aqueous solutions were also carried out for nitrobenzene, and the effects of polarity of the activated carbon, solvent and adsorbate on the adsorption of aromatic compounds were investigated.

Experimental

Preparation of Activated Carbon

Commercially available bead-shaped activated carbon (BAC), purchased from Kureha Corporation, Japan, was employed for the experiments. BAC was boiled in de-ionized water to thoroughly remove impure fine powder. In this study, two oxidation methods were used to obtain activated carbons possessing different properties. Approximately 3 g of BAC was soaked in 90 mL of concentrated H₂SO₄ overnight at 30°C under stirring condition. Then, 15 mL of concentrated nitric acid (HNO₃) or 9 g of potassium permanganate (KMnO₄) were added into the solution. BAC was oxidized for 3 h at 60°C with HNO₃ and for 72 h at 30°C with KMnO₄ to introduce acidic functional groups onto the carbon surface. The oxidation process with KMnO₄ is known as Hummers’ method.¹¹⁾ After the oxidation process of Hummers’ method, 30% (w/w) H₂O₂ was added into the solution to reduce remaining MnO₄⁻ and the sample was washed with 1 M HCl to remove MnO₂ formed during oxidation. The samples obtained in the two oxidation processes were washed several times with hot distilled water until the pH value of the washing solution was no longer changed. For the sample oxidized with HNO₃, heat treatment for 2 h at 250°C was performed to allow the decomposition of the nitrate ions still remaining on the carbon surface. The samples oxidized by HNO₃ and KMnO₄ were respectively denoted as NOX and POX.

Outgassing (OG) was carried out for each prepared sample to remove acidic functional groups on the carbon surface by heating for 1 h at 900°C in a tubular furnace under nitrogen flow. The sample was cooled down to room temperature under nitrogen atmosphere. NOX and POX treated by this process were referred to as NOG and POG, respectively.

Properties of the Activated Carbons

Surface area (S_BET) and micro pore volume (V_micro) for the prepared activated carbons were measured by the adsorption–desorption isotherms of nitrogen gas at −196°C with a surface area and pore size analyzer (Beckman Coulter, SA-3100). The amounts of acidic and basic surface functional groups were quantified by Boehm titration method.¹²⁾ Approximately 0.1 g of the prepared activated carbon and 25 mL of 0.1 M NaHCO₃, 0.05 M Na₂CO₃, 0.1 M NaOH, or 0.1 M HCl solution were put into an Erlenmeyer flask and agitated at 100 rpm for 4 d at 25°C. After that, 5 mL aliquot of each solution was separately drawn from the flask and titrated with 0.05 M HCl. For the HCl solutions, 10 mL of 0.1 M NaOH was added into the solution and then back titrated with 0.05 M HCl.

Adsorption Experiments

All chemicals were purchased in reagent grade from Kanto Chemical Co., Inc. and used without further purification. For the adsorption experiment, benzene and nitrobenzene were employed as adsorbates. The properties of these compounds are displayed in Table 1. In preparation of the solution for adsorption experiments, 1.8 g of benzene or 1.9 g of nitrobenzene was accurately taken in 1 L deionized water, while exactly 3 g of benzene or nitrobenzene was added in 1 L n-hexane solutions, and thoroughly solved in each solution at room temperature. Reproducibility for concentrations of benzene and nitrobenzene was confirmed and the relative error for each was ca. 5%. Fifty milligrams of the prepared activated carbons were added into 25 mL of aqueous solution or 20 mL of n-hexane solution, and stirred at 100 rpm for 1 d at 25°C. To examine the higher equilibrium concentration for the adsorption isotherms of benzene and nitrobenzene, after removing the first 25 mL equilibrium solution by decantation, another 25 mL saturated aqueous solution was added to the flask to overcome the solubility limits of adsorbates and then agitated again to achieve the next equilibrium. No pH adjustment of aqueous solution was done for all adsorption experiments. The experimental data were analyzed using Langmuir model,¹⁴⁾ which is represented by the following equation:   

(1)

where Q_e (mmol/g) is the adsorptive amount of aromatics at equilibrium, C_e (mmol/L) is the equilibrium concentration of aromatics in the solution, and X_m (mmol/g) and K_e (L/mmol) are the maximum adsorption capacity and the adsorption affinity, respectively.

Table 1. Properties of Adsorbates¹³⁾

Adsorbate	Solubility [g/100 mL]	Dipole moment [D]
Benzene	0.18	0.00
Nitrobenzene	0.19	4.22

After the adsorption reached equilibrium, the solutions were diluted as required and analyzed for the remaining concentration of benzene and nitrobenzene using an ultraviolet spectrophotometer (Shimadzu, UV-2550). To check reproducibility, two samples of the diluted solution were analyzed and confirmed that the relative error for each was less than 5%. The calibration curve for benzene and nitrobenzene ranged approximately 100–1800 mg/L and 1–25 mg/L, respectively, and the coefficient of determinations were 0.991–0.999. The equilibrium pH (pH_eq) of the solution was measured with pH meter (HORIBA, D-51).

Adsorption and Desorption Kinetics in Aqueous Solutions

Adsorption and desorption kinetics in aqueous solutions were examined for nitrobenzene. About 100 mg of either NOX or POX was added into 50 mL of 15 mmol/L nitrobenzene solution and sampled at several intervals. All experimental results were fitted to the pseudo-second-order kinetic model as represented in the following equation¹⁵⁾:   

(2)

where Q_t1 and Q_e1 (mmol/g) are the adsorption amount at time t (h) and at the equilibrium, respectively, and k_ads (g/mmol·h) is the pseudo-second-order rate constant for adsorption. The solution was removed by decantation and 50 mL of distilled water was added and then pipetted out at several intervals. All experimental results were fitted to the pseudo-second-order kinetic adsorption as,   

(3)

where Q₀, Q_t2, and Q_e2 (mmol/g) are the desorption amount at time 0 (h), t (h), and at the equilibrium, respectively, and k_des (g/mmol·h) is the pseudo-second-order rate constant for desorption.

Results and Discussion

Properties of the Activated Carbons

The textural and surface properties of the prepared activated carbons are shown in Table 2. The highest surface area and the lowest amount of total surface functional groups were obtained for BAC. In the oxidized samples, especially NOX, approximately 30 times larger amount of acidic functional groups than that of BAC was introduced on the carbon surface, although the surface area and pore volume significantly decreased. These reductions by the oxidation would be due to the presence of acidic functional groups inhibiting accessibility of the nitrogen gas molecule into the pore at −196°C. On the other hand, the surface area and pore volume for the outgassed samples (OGs) increased again by heat treatment. These results can be attributed to the removal of surface acidic functional groups resulting in releasing steric hindrance in the pores. The increase in basic sites for OGs might occur due to the reduction of electron-withdrawing and bulky acidic groups such as carboxy groups in the pores, as well as the increase in π electron density and oxygen-containing basic groups on the carbon surface.¹⁶⁾

Table 2. Textural and Surface Properties of Each Prepared Activated Carbon

Sample	SBET [m2/g]	Vmicro [mL/g]	Surface functional groups [meq/g]
			Carboxy	Lactone	Phenol	Basic
BAC	1360	0.53	0.00	0.16	0.10	0.60
NOX	140	0.05	3.75	0.85	3.30	0.00
NOG	650	0.27	0.00	0.00	0.06	0.48
POX	1060	0.43	2.90	1.30	1.90	0.00
POG	1280	0.53	0.00	0.00	0.25	0.27

Adsorption of Aromatic Compounds in Aqueous Solutions

Adsorption isotherms of benzene and nitrobenzene for each prepared activated carbon are shown in Fig. 1 and the Langmuir parameters obtained from the isotherms are summarized in Table 3. The approximation curves for each activated carbon depicted by Langmuir model were well fitted to each experimental data except for the benzene adsorption on NOX and POX. The poor coefficient values for NOX and POX imply that heterogeneous adsorption might be caused by irregular assembly and orientation of benzene molecules on the different acidic functional groups present on the carbon surface.¹⁷⁾ The pH values in the equilibrium solution after adsorption of the two adsorbates on all samples were around 4.0–6.0. Since both adsorbates are present in molecular species in this pH range, the effect of solution pH on the irregular assembly of benzene could be neglected. The K_e values of OXs were lower than those of OGs for both adsorbates, suggesting that different adsorption sites could be expected between OXs and OGs. Aromatic compounds are considered to adsorb onto activated carbon by π–π interactions between the aromatic ring of the adsorbates and the graphene layers of the activated carbons.⁴⁾ Aromatics can also adsorb onto acidic functional groups and water molecules in aqueous solution are preferentially adsorbed onto acidic functional groups as well and form rigid clusters via hydrogen bond leading to increasing extent of diffusion hindrance in pores.¹⁸⁾

Fig. 1. Langmuir Isotherms of (a), (b) Benzene and (c), (d) Nitrobenzene on BAC (■), NOX (○), NOG (●), POX (△) and POG (▲) for 1 d at 25°C in Aqueous Solutions without pH Adjustment

Amount of adsorbent: 50 mg. Solution volume: 25 mL.

The X_m values of OXs for benzene were lower than those of OGs, as well as the K_e values. Within OXs, POX exhibits approximately 8 times larger surface area than NOX, the adsorption amount of benzene on POX was only twice as much as on NOX. In the case of nitrobenzene, the X_m values of OXs were surprisingly greater than that of OGs, in spite of the opposite order for K_e values as clearly seen in Table 3. Ramis et al.¹⁹⁾ proposed that nitrobenzene could adsorb perpendicularly to polar carbon surfaces with its nitro group attaching to the surface, namely end-on adsorption opposite to flat-on adsorption by π–π interactions. OXs have polar surfaces owing to the presence of acidic functional groups,²⁰⁾ hence dipole–dipole interactions will be working between the polar surface of OXs and high dipole moment of nitrobenzene. Based on the observation of non-polar benzene and polar nitrobenzene adsorption onto OXs, both of them were hard to access to the graphite surface because of physical inhibition in the pores by acidic functional groups. However, nitrobenzene could alternatively adsorb onto polar surface via dipole–dipole interactions, whereas non-polar benzene could adsorb only onto graphite surface remained. The K_e values of NOX are smaller than those of POX for both adsorbates. Acidic functional groups could reduce π-electron density of the graphene layer and decrease the adsorption affinity by π–π interactions.^21,22) This could cause greater decline in K_e values for NOX than POX.

Table 3. Langmuir Parameters of Benzene and Nitrobenzene at 25°C in Aqueous Solutions

Sample	Benzene			Nitrobenzene
	Xm [mmol/g]	Ke [L/mmol]	R2	Xm [mmol/g]	Ke [L/mmol]	R2
BAC	11.4	0.50	0.99	5.49	8.79	0.99
NOX	1.03	0.05	0.61	3.19	0.15	0.97
NOG	6.64	0.12	0.97	2.33	3.26	0.99
POX	2.14	0.14	0.67	4.94	0.37	0.98
POG	9.74	0.48	0.99	4.66	3.83	0.99

Comparing BAC with OGs, adsorption amounts of benzene and nitrobenzene seemed to be proportional to the surface area, while adsorption affinities of the adsorbates varied depending on the surface nature, probably due to the presence of basic sites/groups on OGs.

Adsorption of Aromatic Compounds in n-Hexane Solutions

Adsorption results of benzene and adsorption isotherms of nitrobenzene for each prepared activated carbon are shown in Figs. 2 and 3, respectively. Langmuir parameters obtained from the isotherms are summarized in Table 4. In n-hexane solutions, only a little adsorptive amount of both adsorbates, especially benzene, was observed for all activated carbons compared to those in aqueous solutions. High hydrophobicity of benzene and nitrobenzene would cause high affinity with n-hexane, therefore adsorption onto activated carbon hardly occurred. Since the affinity between non-polar benzene and n-hexane is assumed to be larger than that between high polarity nitrobenzene and n-hexane, benzene can adsorb only a little amount regardless of the kind of activated carbon. On the other hand, the adsorption amounts of nitrobenzene onto the OXs were always larger than those onto corresponding OGs at any equilibrium concentration. Nitrobenzene can be estimated to adsorb onto two adsorption sites, π-electron of graphene layers and acidic functional groups. Especially, the adsorption amount of nitrobenzene onto POX was much higher than even that for BAC since POX has relatively high surface area and a lot of acidic functional groups.

Fig. 2. Adsorption of Benzene on the Each Prepared Activated Carbon for 1 d at 25°C in n-Hexane Solutions

Initial benzene concentration: 38.5 mmol/L. Amount of adsorbent: 50 mg. Solution volume: 20 mL.

Fig. 3. Langmuir Isotherms of Nitrobenzene on AC (■), NOX (○), NOG (●), POX (△) and POG (▲) for 1 d at 25°C in n-Hexane Solutions

Amount of adsorbent: 50 mg. Solution volume: 20 mL.

Table 4. Langmuir Parameters of Nitrobenzene at 25°C in n-Hexane Solutions

Sample	Xm [mmol/g]	Ke [L/mmol]	R2
BAC	1.33	0.22	0.94
NOX	0.58	0.27	0.95
NOG	0.54	0.12	0.89
POX	1.89	0.28	0.99
POG	1.36	0.15	0.97

Unlike in aqueous solutions, the K_e values of OXs were about twice as high as the corresponding OGs. The nitrobenzene adsorption onto OGs would be inhibited by non-polar n-hexane solvent due to only a little amount of acidic functional groups. The adsorption inhibition could become weaker for polar natured OXs, therefore the K_e values of OXs were considered to become larger than that of OGs.

Adsorption and Desorption Kinetics of Nitrobenzene

Figure 4 shows the adsorption and desorption kinetics of nitrobenzene for BAC, NOX, and POX, and the calculated parameters from the plots are summarized in Table 5. The value of adsorption rate constant k_ads for NOX was the smallest among the three adsorbents, indicating that nitrobenzene was less diffusive in the pores of NOX due to a plenty of acidic functional groups in the smaller surface area as given in Table 2. Desorption kinetics parameters of nitrobenzene for BAC, NOX, and POX are represented in Table 5. The results were obtained by adding distilled water into pre-adsorbed activated carbons as described previously. Similar to adsorption, the k_des value of BAC was larger than those of OXs, which was also caused by the inhibition of diffusion by plenty of acidic functional groups in the pores. For the present study, fitting by pseudo-second-order kinetic model was applied for BAC and OXs, revealing that the diffusivity would govern the adsorption and desorption kinetics. The orders of k_ads and k_des in Table 5 can be interpreted in terms of congestion degree of acid functional groups in the pores as can be estimated by surface area and the amount of acidic functional groups as given in Table 2.

Fig. 4. (a) Adsorption and (b) Desorption Kinetics of Nitrobenzene on AC (■), NOX (○), and POX (△) at 25°C in Aqueous Solutions

Solid curves are drawn by pseudo-second-order kinetic model. Amount of adsorbent: 100 mg. Solution volume: 50 mL.

Table 5. Pseudo-second-order Kinetic Parameters for Nitrobenzene at 25°C in Aqueous Solutions

Sample	Adsorption			Desorption
	Qe1 [mmol/g]	kads [g/mmol·h]	R2	Q0 [mmol/g]	Qe2 [mmol/g]	kdes [g/mmol·h]	R2
BAC	5.36	3.74	0.99	5.18	4.78	21.6	0.99
NOX	2.16	1.15	0.99	2.17	1.22	3.03	0.99
POX	3.71	3.17	0.99	3.69	2.37	3.48	0.99

Conclusion

Based on the experimental results and the data analysis, the conclusions of the study can be summarized as follows.

Acknowledgment

This study was funded in part by the Japan Society for the Promotion of Science (JSPS) under Grant-in-Aid for Scientific Research (C) (No. 26340058).

Conflict of Interest

The authors declare no conflict of interest.

References

• 1) Derylo-Marczewska A., Goworek J., Swiatkowski A., Buczek B., Carbon, 42, 301–306 (2004).
• 2) Radovic L. R., Silva I. F., Ume J. I., Menendez J. A., Leon C. A., Scaroni A. W., Carbon, 35, 1339–1348 (1997).
• 3) Li D., Wu Y., Feng L., Zhang L., Bioresour. Technol., 113, 121–126 (2012).
• 4) Derylo-Marczewska A., Swiatkowski A., Biniak S., Walczyk M., Colloids Surf. A Physicochem. Eng. Asp., 327, 1–8 (2008).
• 5) Dowaidar A. M., El-Shahawi M. S., Ashour I., Sep. Sci. Technol., 42, 3609–3622 (2007).
• 6) Shafeeyan M. S., Daud W. A., Houshmand A., Shamiri A., J. Anal. Appl. Pyrolysis, 89, 143–151 (2010).
• 7) Chen J., Chen Q., Ma Q., J. Colloid Interface Sci., 370, 32–38 (2012).
• 8) Song X., Liu H., Cheng L., Qu Y., Desalination, 255, 78–83 (2010).
• 9) Rangel-Mendez J. R., Streat M., Water Res., 36, 1244–1252 (2002).
• 10) Tarasevich Y. I., Aksenenko E. V., Colloids Surf. A Physicochem. Eng. Asp., 215, 285–291 (2003).
• 11) Hummers W. S. Jr., Offeman R. E., J. Am. Chem. Soc., 80, 1339 (1958).
• 12) Boehm H. P., Carbon, 40, 145–149 (2002).
• 13) Dean J. A., “Lange’s Handbook of Chemistry 15th Edition,” Vol. V, McGraw-Hill, Washington DC, U.S.A., 1998, pp. 105–129.
• 14) Langmuir I., J. Am. Chem. Soc., 40, 1361–1403 (1918).
• 15) Ho Y. S., Mckey G., Water Res., 34, 735–742 (2000).
• 16) Machida M., Amano Y., Aikawa M., TANSO, 247, 62–69 (2011).
• 17) Coasne B., Alba-Simionesco C., Audonnet F., Dosseh G., Gubbins K. E., Langmuir, 25, 10648–10659 (2009).
• 18) Kato Y., Machida M., Tatsumoto H., J. Colloid Interface Sci., 322, 394–398 (2008).
• 19) Ramis G., Busca G., Lorenzelli V., J. Electron Spectrosc. Relat. Phenom., 64–65, 297–305 (1993).
• 20) Ahnert F., Arafat H. A., Pinto N. G., Adsorption, 9, 311–319 (2003).
• 21) Mourão P. A. M., Laginhas C., Custódio F., Nabais J. M. V., Carrott P. J. M., Ribeiro Carrott M. M. L., Fuel Process. Technol., 92, 241–246 (2011).
• 22) Oskouie A. K., Miura Y., Furuya E. G., Noll K. E., Carbon, 40, 1199–1202 (2002).