- Email: [email protected]
Adsorption of trichlorophenol on zeolite and adsorbent regeneration with ozone
Journal of Hazardous Materials 271 (2014) 178–184
Contents lists available at ScienceDirect
Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat
Adsorption of trichlorophenol on zeolite and adsorbent regeneration with ozone Yongjun Zhang ∗ , Raoul Georg Mancke, Marina Sabelfeld, Sven-Uwe Geißen Technische Universität Berlin, Department of Environmental Technology, Chair of Environmental Process Engineering, Secr. KF 2, Straße des 17. Juni 135, 10623 Berlin, Germany
h i g h l i g h t s • • • •
FAU zeolite can efficiently adsorb trichlorophenol. The adsorption follows Freundlich model and Pseudo-second-order kinetics. Ozonation can significantly increase the specific surface of FAU zeolite. Regeneration of loaded zeolite with ozonation is highly feasible.
a r t i c l e
i n f o
Article history: Received 14 November 2013 Received in revised form 28 January 2014 Accepted 17 February 2014 Available online 24 February 2014 Keywords: Phenol Chlorophenol Molecular sieve Ozonation Wastewater
a b s t r a c t A FAU-type zeolite was studied as an adsorbent to remove 2,4,6-trichlorophenol (TCP), a frequently detected recalcitrant pollutant in water bodies. Both adsorption isotherm and kinetics were studied with TCP concentrations from 10 to 100 mg/L. It was observed that TCP was effectively adsorbed onto the zeolite with a high adsorption capacity and a high kinetic rate. Freundlich model and pseudo-secondorder kinetics were successfully applied to describe the experimental data. The influence of solution pH was also studied. Furthermore, ozone was applied to regenerate the loaded zeolite. It was found that an effective adsorption of TCP was kept for at least 8 cycles of adsorption and regeneration. The ozonation also increased the BET specific surface of zeolite by over 60% and consequently enhanced the adsorption capacity. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Chlorophenols (CPs) are widely used in the production of various industrial products, such as wood preservatives, pesticides, disinfectants, etc. As a result, CPs have been widely detected in water bodies. For example, 2,4,6-trichlorophenol (TCP) was found at a maximal concentration of 28.65 g/L in Yellow River, China [1]. When the contaminated water bodies are used for the production of drinking water, it is highly possible that CP would pass through the treatment facilities and finally appear in drinking water [2]. Many studies have demonstrated the ecological toxicity and potential impact of CPs on human health. Researchers also found the potential genotoxcity of CPs, e.g., with point mutations of p53 gene in the liver genome of zebrafish which was exposed to 5 g/L
∗ Corresponding author. Tel.: +49 30 31425298; fax: +49 30 31425487. E-mail addresses: [email protected], [email protected] (Y. Zhang). http://dx.doi.org/10.1016/j.jhazmat.2014.02.020 0304-3894/© 2014 Elsevier B.V. All rights reserved.
TCP for 10 d [3]. Consequently, CPs are listed as priority pollutants in many regions [4]. Various removal technologies (e.g. chemical oxidation, biodegradation, membrane separation, etc.) have been so far developed and applied to remove CPs from water [5]. Adsorption, in which pollutants (adsorbates) are adhered onto a solid material (adsorbent), has some advantages compared with other technologies, such as low investment, high flexibility of design, ease of operation and high tolerance to toxicity. It is widely applied for the removal of a broad range of organic and inorganic pollutants from water [6]. Many adsorbents have been used to adsorb CPs from water, including activated carbons, agricultural and industrial wastes, natural materials, and synthetic resins [7–9]. However, adsorption is a separation process where pollutants are only transferred to adsorbents but not degraded to harmless substances. Furthermore, an adsorbent has a limited adsorption capacity and will be saturated after a certain time of operation. Therefore, the used adsorbent will be either regenerated or disposed and replaced with new one. The popular regeneration
Y. Zhang et al. / Journal of Hazardous Materials 271 (2014) 178–184
methods include thermal calcination, decreasing pressure, wave irradiation, and chemical elution (e.g., with a different pH or an organic solvent) [10,11]. Recently, some researchers start to make efforts to apply a chemical oxidation method to regenerate the adsorbents via oxidizing the organic adsorbates. This method can in situ restore the adsorption capacity of used adsorbents and furthermore can also degrade or mineralize organic pollutants. Nevertheless, the method is not suitable for all adsorbents. For instance, it was found that ozone can transfer basic sites of activated carbon into acid sites and also can decrease the specific surface area, which resulted in a declined adsorption of methylene blue [12]. Therefore, it is of great interest to study a material which can adsorb recalcitrant pollutants like TCP and whose adsorption capacity cannot be compromised during the regeneration process. Zeolite is a group of porous materials based on aluminosilicate and could be a promising candidate for the purpose. The hydrophobicity of zeolite can be easily adjusted during the manufacturing process to adsorb different pollutants, e.g. by changing the ratio of Si/Al, the pore size and structure [13]. However, the adsorption of CPs on zeolite has not been fully studied and only several studies were reported [14–16]. Their adsorption kinetics on zeolite is rarely available in literature. On the other hand, the regeneration of loaded zeolite with chemical oxidation is only reported in limited studies, most of which used the Fenton reaction. Koryabkina et al. coated Fe2+ on zeolite by a reaction with sodium borohydride and achieved a stable adsorption of disinfection byproducts in 4 cycles of adsorption and regeneration with the Fenton reaction [17]. Wang et al. used the homogeneous Fenton oxidation to regenerate MCM-22 zeolite loaded with methylene blue but only 60% of its adsorption capacity was recovered [18]. However, the Fenton reaction normally requires a strong acidic condition and leaching of immobilized Fe2+ would lead to a decreased oxidation rate [19]. Ozone can be an alternative oxidant to regenerate zeolites, which was rarely reported in literature. To our knowledge, only one study was reported in scientific journal by Reungoat et al., who obtained a complete recovery of the adsorption capacity for nitrobenzene [20]. Therefore, the objective of this study was firstly to investigate the adsorption of TCP onto a commercial zeolite, including both adsorption isotherm and kinetics, and secondly to apply ozone to regenerate the used zeolite. 2. Experimental 2.1. Materials
179
membrane and analyzed with HPLC to determine the concentration of TCP. The kinetic tests followed a similar procedure except that samples were taken in a given time interval during 6 h. The effect of initial pH was studied between pH 3 and pH 10, which was adjusted by adding 10 mM HCL or 10 mM NaOH at the beginning. The test solution had a volume of 50 mL and contained 1 g of zeolite and 100 mg/L of TCP. Other conditions were the same as above. A sample was taken after 24 h. All tests were conducted in triple and an average value was used. The TCP adsorption on the zeolite at given time was calculated according to: qt = V
C0 − Ct W
where qt is the adsorbed TCP at time t (mg/g); C0 initial TCP in water (mg/L); Ct TCP in water at time t (mg/L); V the solution volume (L); W the adsorbent weight (g). The equilibrium data (qe , Ce ) in the isotherm tests were calculated with the same method. 2.3. Regeneration with ozone A glass bottle was used to contain 10 g of the zeolite and 400 mL of 50 mg/L TCP solution. The zeolite was suspended by a magnetic stirrer. After 2-h adsorption, a sample was taken and then the ozone-containing gas (45–50 ppm) was bubbled into the solution for 30 min to regenerate the zeolite. The ozone in the water phase (7 mg/L) was measured with the Indigo colorimetric method [21]. Ozone was produced from pure oxygen with an ozone generator (SORBIOS GSF 010.2). After the ozonation, air was used to strip out the ozone residue in the solution. After air stripping, ozone in water phase was below 0.07 mg/L, which was almost the detection limit. A new adsorption test was started by replacing half of the old solution with a new TCP solution (100 mg/L) and the final TCP concentration was 50 mg/L again. A lot of intermediates may be formed during the above test, which could influence the adsorption of TCP. Therefore, FAU was washed and dried after 10 adsorption cycles, 1 g of which was added into 40 mL TCP solution (50 mg/L) for a clean adsorption test. A sample was taken after 2 h to compare the adsorption with the previous adsorption/regeneration cycles. The specific surface areas of FAU before and after 10 cycles of ozonation were determined with the BET method (Autosorb-1, Quantachrome) with nitrogen gas. 2.4. HPLC conditions
The adsorbent used in this study was a high-silica FAU zeolite ˚ diameter 1.6–2.5 mm), supplied by CWK Chemiew(pore size 9 A, erk Bad Köstritz GmbH. The chemical composition of zeolite is (H,Na)2 O·Al2 O3 ·xSiO2 ·nH2 O (x ≥ 30), as described by the supplier. 2,4,6-Trichlorophenol (98% purity) and methanol of HPLC grade were purchased from Sigma–Aldrich, Germany.
TCP in all water samples was determined with a HPLC system (Agilent 1200) equipped with a Phenomex Gemini-NX 5 m column. 50 L of sample was injected into the system and the elution was conducted with a mixture of methanol and Milipore water (85:15) at a flow rate of 1 mL/min. The column oven was set at 30 ◦ C. The eluted TCP was monitored with a UV detector at a wavelength of 290 nm. The quantification limit of TCP was 0.2 mg/L.
2.2. Adsorption tests
3. Results and discussion
All adsorption kinetics and isotherm tests were conducted with a batch equilibrium procedure. In the isotherm test, 1 g of zeolite was added to 50 mL of TCP solution with deionsed water at different concentrations (10, 25, 50, 75, and 100 mg/L). The adsorption was carried out at room temperature in a 100-mL Erlenmeyer flask which was shaken at a speed of ∼110 rpm. No buffer was applied in order to avoid the interference of ions. However, pH values in all test stayed around 6.2–6.5. Samples (1.5 mL) were taken out of the solution after 24 h when the equilibrium was assumed. The samples were filtrated through a syringe filter with a 0.45 m polycarbonate
3.1. Effects of initial pH The solution pH can influence the adsorption process via altering the ionic states of both adsorbent and adsorbate. In addition, the excessive H+ or OH− ions can also compete with an adsorbate for the active sites of adsorbent. In this study, the adsorption of TCP on FAU zeolite was tested in pH 3–10. As shown in Fig. 1, a clear declination of the adsorption capacity can be found with the increase of initial pH. The adsorption capacity dropped down from 4.3 mg/g at pH 3.5–2.6 mg/g at pH 10. The similar effect was also found in
180
Y. Zhang et al. / Journal of Hazardous Materials 271 (2014) 178–184
Fig. 1. The adsorption capacity of FAU for TCP at different initial pH values.
the adsorption of TCP onto different activated carbons [22,23] and activated clay [24]. The observed phenomenon can probably be explained by the adsorption mechanisms, mainly including hydrophobic interactions, electrostatic interactions, electron donor–acceptor complexes, metal coordination. The pore size of FAU is too small for TCP to be hold inside and therefore, the above mentioned interactions may mainly occur on the surface of FAU. TCP is a weak acidic compound and stays in the undissociated state at pH below its pKa value (6.23, SRC PhysProp Database). Its solubility in water tends to be elevated from acidic pH to basic pH [14]. Therefore, the dispersion interaction between TCP and the hydrophobic sites of zeolite may play a predominant role at acidic pH. On the contrary, the zeolite surface is negatively charged in the basic conditions and thus would repulse the TCP anions by electrostatic interactions. Nevertheless, there was still a certain amount of TCP adsorbed onto FAU in the basic condition, indicating that other mechanisms might take effect. The presence of 3 chlorine atoms makes TCP tending to be electron-deficient. Therefore, an electron donor–acceptor complex might form between TCP and nonbonding electrons at the siloxane surface of FAU. This complex was also proposed for the interaction of TCP and another synthetic zeolite SBA-15 [25]. In addition, TCP anions could also form a coordination complex with metal sites of FAU, which have been reported between chlorophenol and the surface of natural zeolite [14]. 3.2. Adsorption isotherm The isothermal tests were conducted at room temperature. The equilibrium was assumed after 24 h. Three models were applied to simulate the obtained TCP concentrations in solution (Ce , mg/L) and on FAU (qe , mg/g) at equilibrium, as follows: (a) Langmuir: qe =
qm KL Ce 1 + KL Ce
where qm is the Langmuir adsorption capacity (mg/g); KL , the Langmuir constant (L/mg). (b) Freundlich: 1/n
qe = KF Ce
where KF and n are Freundlich constants. (c) Temkin: qe =
RT ln(ACe ) bT
where R is the universal gas constant (m2 kg s−2 K−1 mol−1 ); T is the absolute solution temperature (K); A is the equilibrium binding constant (L/mg). Langmuir model failed to describe the experimental data since a negative adsorption capacity was obtained with the linear regression, which indicates that some assumptions of Langmuir model may not be applicable in the studied system. For instance, FAU contains both metallic and silica sites and many adsorption mechanisms may occur as discussed above, which cannot satisfy the homogeneous assumption of Langmuir. Nevertheless, Langmuir model was still successfully applied to describe the adsorption of 2-CP on zeolitic materials derived from bagasse fly ash (SiO2 /Al2 O3 = 4) [16] and 2-CP on Na-zeolite and Na/Ni-zeolite [26]. On the other hand, both Freundlich and Temkin can describe the adsorption isotherm data at a different extent of accuracy. The accuracies of both models were evaluated by the coefficient of determination (R2 ) for the linear regression and the coefficient of nonlinear chi squire (X2 ) for the accordance of the experimental and simulated data, as proposed by Ho [27]. X2 was calculated with: X2 =
N (qe,sim − qe,exp )2 i=1
qe,exp
where N is the number of data points; qe,exp , the experimental adsorption capacity at equilibrium (mg/g); qe,sim , the simulated adsorption capacity at equilibrium (mg/g). Freundlich model resulted in a better accuracy than Temkin in terms of both R2 and X2 coefficients (Fig. 2). The simulated curves with both models showed a comparable accordance with the experimental data in the middle concentration range. However, the Freundlich curve presented a much better accordance at both low and high concentrations than the Temkin curve (Fig. 2a). Especially, Temkin model predicted a negative value of qe,sim at low concentrations which significantly increased its value of X2 . Nevertheless, X2 of Temkin was still higher than that of Freundlich even after excluding the negative qe,sim . Therefore, it can be concluded that Freundlich described the experimental data with the best accuracy among the studied three models. The calculated values of Freundlich constants were 0.43 and 1.80 × 10−3 mg/g/(mg/L)1/n for n and Kf , respectively. The obtained Kf value equals to 1.99 mmol/g/(mmol/L)−1/n which is comparable with most of adsorbents for phenolics in literature, such as low-cost activated carbon and synthetic resins [9]. The adsorption capacity obtained in this study is approximately 10 times lower than that of TCP adsorbed on activated clay [24], which however was conducted at pH 4. As shown before, the acidic conditions could significantly promote the adsorption of TCP. Nevertheless, consideration should also be given to the large particle size of FAU (∼2 mm), which is beneficial for large scale applications but would decrease the external specific surface. For instance, activated carbons could have a specific surface area more than 1000 m2 /g [9], while the zeolite applied in this study had a value below 300 m2 /g as shown later. It can also be noticed in Fig. 2a that the experimental data tend to follow a typical S-shape curve. This curve was also observed in the adsorption of phenol onto n-octadecyl alcohol crystals and onto wool [28]. Moreover, it seems that the tested concentrations were in the low range and the saturation point of FAU had not been reached. 3.3. Adsorption kinetics The adsorption kinetic tests were conducted in shaken flasks at room temperature and samples were continually taken from the solution at given intervals. The adsorbed TCP (qt , mg/g) was
Y. Zhang et al. / Journal of Hazardous Materials 271 (2014) 178–184
181
Fig. 2. The experimental and simulated adsorption isotherm data (a) and the regression with Freundlich (b) and Temkin (c) models (qe , mg/g; Ce , mg/L).
calculated as mentioned before. The relation of qt and adsorption time t was simulated with following two models [6]: (a) Pseudo-second-order model: dqt = k2 (qe − qt )2 , or dt
t 1 1 = + t qt qe k2 q2e
where k2 is the reaction constant (g/mg/min). (b) Elovich equation: qt = ˇ ln ˛ˇ + ˇ ln t where ˛ is the initial sorption rate (g/mg/min); ˇ is an adsorption constant. The results of both kinetic models are presented in Table 1. The pseudo-second-order model obtained a higher R2 and a lower X2 than Elovich equation. Therefore, the former model is more suitable to simulate the studied adsorption process. With the obtained parameters, the pseudo-second-order model generated smooth curves which fit well with the experimental data, as presented in Fig. 3. In addition, the pseudo-first-order model was also tried but a poor linearity was obtained. It might be due to the low TCP concentrations in this study. It was previously found that the adsorption process with a low adsorbate concentration had an inclination to follow the pseudo-second-order model [29]. In fact, the pseudosecond-order model was also successfully applied to simulate the adsorption of 4-CP on a surfactant-modified natural zeolite [30] and TCP on activated carbon [31]. Another benefit of the pseudosecond-order model is that the experimental adsorption capacity at equilibrium is not required, which can even be deduced from the model [6]. As shown in Table 1, there is a very good correlation (X2 = 0.06) of qe,exp and qe,sim in the pseudo-second-order model. The adsorption rates (k2 ) in Table 1 are [33] in a comparable range with the data obtained for TCP adsorbed on activated carbon [31] and activated clay [24] but much higher than the rates of phenols adsorbed on zeolitic tuff [32]. One can notice that the adsorption rates (k2 ) of pseudo-secondorder kinetics varies with the initial adsorbate concentration (Table 1). A complex correlation between them has been
theoretically deduced and two empirical models were previously used in literature, as follows [29]: k2 =
C0 , aC0 + b
k2 = cC0d ,
or
or
Co = aCo + b k2
ln k2 = d ln C0 + ln C
where a, b, c, d are constants. Fig. 4 presents the regressions of k2 and C0 with above models. Both of them obtained a comparable coefficient of determination but the exponential model showed a much better coefficient of nonlinear chi squire between simulated k2 and experimental k2 . Therefore the exponential model is more suitable to predict the adsorption rate k2 for TCP adsorbed onto FAU in similar conditions.
Fig. 3. Time profiles of adsorbed TCP on FAU for different initial concentrations: Experimental data and simulated curves with the pseudo-second-order kinetics.
182
Y. Zhang et al. / Journal of Hazardous Materials 271 (2014) 178–184
Table 1 Results of simulating the kinetic data with pseudo-second-order model and Elovich equation. TCP
100 mg/L 75 mg/L 50 mg/L 25 mg/L 10 mg/L
Pseudo-second-order kinetics
Elovich equation
k2 (g/mg/min)
qe,sim (mg/g)
qe,exp (mg/g)
R2 (–)
9.82 × 10−3 1.57 × 10−2 2.35 × 10−2 5.71 × 10−2 6.13 × 10−2
3.18 2.29 1.33 0.45 0.15
3.06 2.09 1.19 0.40 0.11
0.998 0.999 0.995 0.999 0.995
3.4. Adsorbent regeneration When the adsorption equilibrium is reached, the adsorbent cannot adsorb the target adsorbate anymore. A regeneration step is in need to restore the adsorption capacity. Ozone, as one of the strongest oxidants, holds a standard electrode potential of 2.08 V and has been widely used to oxidize organic pollutants. Ozone was applied in this study to regenerate FAU loaded with TCP. For this purpose, two basic requirements should be fulfilled: (1) ozone can oxidize the adsorbate; (2) the adsorption capacity of adsorbent should not be compromised during the attack of ozone. Many studies have confirmed the oxidation of TCP by ozone [33]. Therefore, the adsorption capacity of FAU was emphasized in this study. The adsorption/regeneration cycle started with fresh FAU. After 2 h of adsorption, a sample was taken to calculate the removal of TCP and ozone gas was bubbled for 30 min into the solution where TCP both on FAU and in water was oxidized. The ozone residue was stripped out with aeration. After this regeneration treatment, TCP in the solution was almost eliminated (an average removal efficiency 97%). It was found that TCP adsorption onto FAU was significantly enhanced after the first regeneration with ozone: TCP removal from water was elevated from 52% up to 87%. As shown in Fig. 5, this elevated adsorption capacity was remained until 8th adsorption cycle. However, a dramatic dropdown of TCP removal was found
Fig. 4. Correlations between the pseudo-second-order rate and the initial adsorbate concentration in water: (a) exponential model; (b) division model (k2 , g/mg/min; C0 , mg/L).
X2 (–)
˛ (g/mg/min)
ˇ (–)
R2 (–)
0.06
0.63 0.89 1.12 3.06 5.24
0.62 0.46 0.28 0.094
0.976 0.969 0.946 0.975
X2 (–)
0.97
in 9th/10th adsorption, which might be due to two effects: matrix carbon and pH decrease as discussed below. Since only half of the solution was replaced in every cycle, the adsorption would be influenced by the intermediates formed during ozonation. Those matrix compounds could compete with TCP for the ozone consumption and finally might lead to an incomplete oxidation of adsorbed TCP. Consequently, a partial regeneration of FAU might occur. Moreover, the matrix compounds could also take over the adsorption site of FAU, which can decrease the adsorption of TCP. On the other hand, it is well known that some acids can be generated during the ozonation of TCP [33]. Firstly the chlorine atoms on TCP can be replaced with a hydroxyl group to form 2,6-dichlorobenzo-1,4-quinone and HCl. Secondly, the cleavage of the aromatic ring would generate aliphatic products which in turn may react with ozone to form organic acids, such as formic acid and acetic glycolic acids. The pH in the starting solution in our study was ∼6.3 and after the 10th adsorption cycle the pH decreased to 4.1. This acidic environment will promote the TCP adsorption as shown before. However, a significant decrease of TCP adsorption was observed at last adsorption cycles. It indicates that the effects of matrix compounds generated during ozonation might be more influential than the pH effect. More interestingly, the adsorption of TCP was kept at an elevated state in the clean adsorption with FAU after 10 cycles of use: 96% of TCP was adsorbed. Therefore, some modification should have happened on FAU, which could elevate its adsorption capacity. It was proposed that ozone can clean the pores of a FAU zeolite and thus lead to an elevated adsorption of nitrobenzene [20]. However, the researchers did not measure the specific surface of zeolite after ozonation. In this study, the BET specific surface area was analyzed
Fig. 5. Removal of TCP from water in the adsorption/regeneration cycles and in the clean adsorption test.
Y. Zhang et al. / Journal of Hazardous Materials 271 (2014) 178–184
183
process with a high accuracy. And the reaction rate k2 showed an exponential correlation with the initial TCP concentration. Both the BET surface area and TCP adsorption of FAU were significantly elevated by the regeneration with ozone. The adsorption capacity was kept stable for at least 8 cycles of adsorption and regeneration. Therefore, it is highly feasible to use FAU as an adsorbent to remove TCP from water and to apply ozone for the adsorbent regeneration. However, further studies are required to clarify the influence of other ingredients in real wastewater, such as metals, organics, which may compete with TCP for the adsorption sites. A column study is necessary to optimize the ozone regeneration and to step toward real applications. Acknowledgement
Fig. 6. BET surface area plots of studied zeolite before and after ozonation.
for the tested FAU zeolite before and after zonation (Fig. 6). It can be clearly found that ozonation largely increased the BET surface of FAU by 63% (from 289.29 to 470.75 m2 /g). However, further study is required to clarify the ozone effect on zeolite in terms of ozone dosage, reaction time, stability, etc. Another benefit with ozone is that the presence of zeolite could promote the decomposition of ozone with the generation of free radicals, which can unselectively oxidize more organic compounds than ozone itself [34]. The ozone decomposition mainly occurs on the surface of zeolite where the adsorbed compounds also locate. This localized reaction could avoid the radical consumption by scavenger compounds and the self-decay of radicals. Thus the free radicals can be used more efficiently. However, the regeneration with ozone was conducted in this study at acidic pH which is not favorable for the decomposition of ozone. Therefore, less free radicals could be expected at the 9th/10th cycles of adsorption and regeneration and partially contributed to an incomplete regeneration of zeolite. Since ozone can oxidize a broad range of organics with high kinetic rates [35], the adsorbent regeneration can be completed with ozone in a short time: 30 min in this study vs. several hours with the popular thermal calcination (e.g., 1 h at 540 ◦ C for a synthetic zeolite loaded with methylene blue [18]; 2 h at 450 ◦ C for coal fly ashes loaded with methylene blue [36]; 12 h at 450 ◦ C for FAU with nitrophenolics [37]). Actually, the ozonation time was not optimized in this study. It is highly possible that the time can be further reduced, which, however, should be evaluated in another study. Moreover, ozone decomposition can be promoted by some simple methods (e.g. increasing reaction pH or adding H2 O2 ) and then more free radicals can be produced. It could be used to regenerate zeolites loaded with organics which have a low reaction rate with ozone. However, further studies are required to evaluate the influence of free radicals on zeolites and other materials. In addition, various types of natural zeolites, which are generally cheaper than the synthetic ones, have been applied to adsorb pollutants from water [8]. However, the adsorption capacity of natural zeolites might be reduced during regeneration. Wang et al. tried thermal calcination and Fenton reaction separately to regenerate a type of natural zeolite loaded with methylene blue [18]. They found that only 60% of adsorption capacity was recovered by both methods. 4. Conclusions and outlook TCP can be effectively removed from water by adsorbing onto FAU zeolite, especially at lower pH. The adsorption isotherm can be well described with Freundlich model within the studied conditions. Pseudo-second-order kinetics can simulate the adsorption
We appreciate the help of Prof. Dietmar Stephan and Mr. Gerrit Land for the analysis of BET specific surface. References [1] J. Gao, L. Liu, X. Liu, H. Zhou, S. Huang, Z. Wang, Levels and spatial distribution of chlorophenols – 2,4-dichlorophenol, 2,4,6-trichlorophenol, and pentachlorophenol in surface water of China, Chemosphere 71 (2008) 1181–1187. [2] A.O. Olaniran, E.O. Igbinosa, Chlorophenols and other related derivatives of environmental concern: properties, distribution and microbial degradation processes, Chemosphere 83 (2011) 1297–1306. [3] D. Yin, H. Zhu, P. Hu, Q. Zhao, Genotoxic effect of 2,4,6-trichlorophenol on P53 gene in zebrafish liver, Environ. Toxicol. Chem. 28 (2009) 603–608. [4] L. Xing, H. Liu, J.P. Giesy, H. Yu, pH-dependent aquatic criteria for 2,4-dichlorophenol, 2,4,6-trichlorophenol and pentachlorophenol, Sci. Total Environ. 441 (2012) 125–131. [5] G. Busca, S. Berardinelli, C. Resini, L. Arrighi, Technologies for the removal of phenol from fluid streams: a short review of recent developments, J. Hazard. Mater. 160 (2008) 265–288. [6] Y.-S. Ho, Review of second-order models for adsorption systems, J. Hazard. Mater. 136 (2006) 681–689. ˛ [7] A. Dabrowski, P. Podko´scielny, Z. Hubicki, M. Barczak, Adsorption of phenolic compounds by activated carbon – a critical review, Chemosphere 58 (2005) 1049–1070. [8] M. Ahmaruzzaman, Adsorption of phenolic compounds on low-cost adsorbents: a review, Adv. Colloid Interface Sci. 143 (2008) 48–67. [9] S.-H. Lin, R.-S. Juang, Adsorption of phenol and its derivatives from water using synthetic resins and low-cost natural adsorbents: a review, J. Envrion. Manage. 90 (2009) 1336–1349. ´ [10] R. Cherbanski, E. Molga, Intensification of desorption processes by use of microwaves – an overview of possible applications and industrial perspectives, Chem. Eng. Process. 48 (2009) 48–58. [11] D. Bathen, Physical waves in adsorption technology – an overview, Sep. Purif. Technol. 33 (2003) 163–177. [12] H. Valdés, M. Sánchez-Polo, J. Rivera-Utrilla, C.A. Zaror, Effect of ozone treatment on surface properties of activated carbon, Langmuir 18 (2002) 2111–2116. [13] M. Khalid, G. Joly, A. Renaud, P. Magnoux, Removal of phenol from water by adsorption using zeolites, Ind. Eng. Chem. Res. 43 (2004) 5275–5280. [14] R.I. Yousef, B. El-Eswed, The effect of pH on the adsorption of phenol and chlorophenols onto natural zeolite, Colloids and Surf. A 334 (2009) 92–99. [15] B. Koubaissy, G. Joly, I. Batonneau-Gener, P. Magnoux, Adsorptive removal of aromatic compounds present in wastewater by using dealuminated faujasite zeolite, Ind. Eng. Chem. Res. 50 (2011) 5705–5713. [16] B. Shah, R. Tailor, A. Shah, Sorptive sequestration of 2-chlorophenol by zeolitic materials derived from bagasse fly ash, J. Chem. Technol. Biotechnol. 86 (2011) 1265–1275. [17] N. Koryabkina, J.A. Bergendahl, R.W. Thompson, A. Giaya, Adsorption of disinfection byproducts on hydrophobic zeolites with regeneration by advanced oxidation, Microporous Mesoporous Mater. 104 (2007) 77–82. [18] S. Wang, H. Li, S. Xie, S. Liu, L. Xu, Physical and chemical regeneration of zeolitic adsorbents for dye removal in wastewater treatment, Chemosphere 65 (2006) 82–87. [19] R. Gonzalez-Olmos, F.-D. Kopinke, K. Mackenzie, A. Georgi, Hydrophobic Fezeolites for removal of MTBE from water by combination of adsorption and oxidation, Environ. Sci. Technol. 47 (2013) 2353–2360. [20] J. Reungoat, J.S. Pic, M.H. Manéro, H. Debellefontaine, Adsorption of nitrobenzene from water onto high silica zeolites and regeneration by ozone, Sep. Sci. Technol. 42 (2007) 1447–1463. [21] H. Bader, J. Hoigné, Determination of ozone in water by the indigo method, Water Res. 15 (1981) 449–456. [22] B.H. Hameed, I.A.W. Tan, A.L. Ahmad, Adsorption isotherm, kinetic modeling and mechanism of 2,4,6-trichlorophenol on coconut husk-based activated carbon, Chem. Eng. J. 144 (2008) 235–244.
184
Y. Zhang et al. / Journal of Hazardous Materials 271 (2014) 178–184
[23] I.A.W. Tan, A.L. Ahmad, B.H. Hameed, Adsorption isotherms, kinetics, thermodynamics and desorption studies of 2,4,6-trichlorophenol on oil palm empty fruit bunch-based activated carbon, J. Hazard. Mater. 164 (2009) 473–482. [24] B.H. Hameed, Equilibrium and kinetics studies of 2,4,6-trichlorophenol adsorption onto activated clay, Colloids and Surf. A 307 (2007) 45–52. [25] Q. Qin, K. Liu, D. Fu, H. Gao, Effect of chlorine content of chlorophenols on their adsorption by mesoporous SBA-15, J. Environ. Sci. 24 (2012) 1411–1417. [26] B. Okolo, C. Park, M.A. Keane, Interaction of phenol and chlorophenols with activated carbon and synthetic zeolites in aqueous media, J. Colloid Interface Sci. 226 (2000) 308–317. [27] Y.-S. Ho, Selection of optimum sorption isotherm, Carbon 42 (2004) 2115–2116. [28] C.H. Giles, T.H. MacEwan, S.N. Nakhwa, D. Smith, 786. Studies in adsorption. Part XI. A system of classification of solution adsorption isotherms, and its use in diagnosis of adsorption mechanisms and in measurement of specific surface areas of solids, J. Chem. Soc. 0 (1960) 3973–3993. [29] S. Azizian, Kinetic models of sorption: a theoretical analysis, J. Colloid Interface Sci. 276 (2004) 47–52. [30] A. Kuleyin, Removal of phenol and 4-chlorophenol by surfactant-modified natural zeolite, J. Hazard. Mater. 144 (2007) 307–315.
[31] M. Radhika, K. Palanivelu, Adsorptive removal of chlorophenols from aqueous solution by low cost adsorbent – kinetics and isotherm analysis, J. Hazard. Mater. 138 (2006) 116–124. [32] R.I. Yousef, B. El-Eswed, A.a.H. Al-Muhtaseb, Adsorption characteristics of natural zeolites as solid adsorbents for phenol removal from aqueous solutions: kinetics, mechanism, and thermodynamics studies, Chem. Eng. J. 171 (2011) 1143–1149. [33] N. Graham, W. Chu, C. Lau, Observations of 2,4,6-trichlorophenol degradation by ozone, Chemosphere 51 (2003) 237–243. [34] H. Valdés, V.J. Farfán, J.A. Manoli, C.A. Zaror, Catalytic ozone aqueous decomposition promoted by natural zeolite and volcanic sand, J. Hazard. Mater. 165 (2009) 915–922. [35] U. von Gunten, Ozonation of drinking water. Part I. Oxidation kinetics and product formation, Water Res. 37 (2003) 1443–1467. [36] Z. Sun, C. Li, D. Wu, Removal of methylene blue from aqueous solution by adsorption onto zeolite synthesized from coal fly ash and its thermal regeneration, J. Chem. Technol. Biotechnol. 85 (2010) 845–850. [37] B. Koubaissy, G. Joly, P. Magnoux, Adsorption and competitive adsorption on zeolites of nitrophenol compounds present in wastewater, Ind. Eng. Chem. Res. 47 (2008) 9558–9565.
Contents lists available at ScienceDirect
Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat
Adsorption of trichlorophenol on zeolite and adsorbent regeneration with ozone Yongjun Zhang ∗ , Raoul Georg Mancke, Marina Sabelfeld, Sven-Uwe Geißen Technische Universität Berlin, Department of Environmental Technology, Chair of Environmental Process Engineering, Secr. KF 2, Straße des 17. Juni 135, 10623 Berlin, Germany
h i g h l i g h t s • • • •
FAU zeolite can efficiently adsorb trichlorophenol. The adsorption follows Freundlich model and Pseudo-second-order kinetics. Ozonation can significantly increase the specific surface of FAU zeolite. Regeneration of loaded zeolite with ozonation is highly feasible.
a r t i c l e
i n f o
Article history: Received 14 November 2013 Received in revised form 28 January 2014 Accepted 17 February 2014 Available online 24 February 2014 Keywords: Phenol Chlorophenol Molecular sieve Ozonation Wastewater
a b s t r a c t A FAU-type zeolite was studied as an adsorbent to remove 2,4,6-trichlorophenol (TCP), a frequently detected recalcitrant pollutant in water bodies. Both adsorption isotherm and kinetics were studied with TCP concentrations from 10 to 100 mg/L. It was observed that TCP was effectively adsorbed onto the zeolite with a high adsorption capacity and a high kinetic rate. Freundlich model and pseudo-secondorder kinetics were successfully applied to describe the experimental data. The influence of solution pH was also studied. Furthermore, ozone was applied to regenerate the loaded zeolite. It was found that an effective adsorption of TCP was kept for at least 8 cycles of adsorption and regeneration. The ozonation also increased the BET specific surface of zeolite by over 60% and consequently enhanced the adsorption capacity. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Chlorophenols (CPs) are widely used in the production of various industrial products, such as wood preservatives, pesticides, disinfectants, etc. As a result, CPs have been widely detected in water bodies. For example, 2,4,6-trichlorophenol (TCP) was found at a maximal concentration of 28.65 g/L in Yellow River, China [1]. When the contaminated water bodies are used for the production of drinking water, it is highly possible that CP would pass through the treatment facilities and finally appear in drinking water [2]. Many studies have demonstrated the ecological toxicity and potential impact of CPs on human health. Researchers also found the potential genotoxcity of CPs, e.g., with point mutations of p53 gene in the liver genome of zebrafish which was exposed to 5 g/L
∗ Corresponding author. Tel.: +49 30 31425298; fax: +49 30 31425487. E-mail addresses: [email protected], [email protected] (Y. Zhang). http://dx.doi.org/10.1016/j.jhazmat.2014.02.020 0304-3894/© 2014 Elsevier B.V. All rights reserved.
TCP for 10 d [3]. Consequently, CPs are listed as priority pollutants in many regions [4]. Various removal technologies (e.g. chemical oxidation, biodegradation, membrane separation, etc.) have been so far developed and applied to remove CPs from water [5]. Adsorption, in which pollutants (adsorbates) are adhered onto a solid material (adsorbent), has some advantages compared with other technologies, such as low investment, high flexibility of design, ease of operation and high tolerance to toxicity. It is widely applied for the removal of a broad range of organic and inorganic pollutants from water [6]. Many adsorbents have been used to adsorb CPs from water, including activated carbons, agricultural and industrial wastes, natural materials, and synthetic resins [7–9]. However, adsorption is a separation process where pollutants are only transferred to adsorbents but not degraded to harmless substances. Furthermore, an adsorbent has a limited adsorption capacity and will be saturated after a certain time of operation. Therefore, the used adsorbent will be either regenerated or disposed and replaced with new one. The popular regeneration
Y. Zhang et al. / Journal of Hazardous Materials 271 (2014) 178–184
methods include thermal calcination, decreasing pressure, wave irradiation, and chemical elution (e.g., with a different pH or an organic solvent) [10,11]. Recently, some researchers start to make efforts to apply a chemical oxidation method to regenerate the adsorbents via oxidizing the organic adsorbates. This method can in situ restore the adsorption capacity of used adsorbents and furthermore can also degrade or mineralize organic pollutants. Nevertheless, the method is not suitable for all adsorbents. For instance, it was found that ozone can transfer basic sites of activated carbon into acid sites and also can decrease the specific surface area, which resulted in a declined adsorption of methylene blue [12]. Therefore, it is of great interest to study a material which can adsorb recalcitrant pollutants like TCP and whose adsorption capacity cannot be compromised during the regeneration process. Zeolite is a group of porous materials based on aluminosilicate and could be a promising candidate for the purpose. The hydrophobicity of zeolite can be easily adjusted during the manufacturing process to adsorb different pollutants, e.g. by changing the ratio of Si/Al, the pore size and structure [13]. However, the adsorption of CPs on zeolite has not been fully studied and only several studies were reported [14–16]. Their adsorption kinetics on zeolite is rarely available in literature. On the other hand, the regeneration of loaded zeolite with chemical oxidation is only reported in limited studies, most of which used the Fenton reaction. Koryabkina et al. coated Fe2+ on zeolite by a reaction with sodium borohydride and achieved a stable adsorption of disinfection byproducts in 4 cycles of adsorption and regeneration with the Fenton reaction [17]. Wang et al. used the homogeneous Fenton oxidation to regenerate MCM-22 zeolite loaded with methylene blue but only 60% of its adsorption capacity was recovered [18]. However, the Fenton reaction normally requires a strong acidic condition and leaching of immobilized Fe2+ would lead to a decreased oxidation rate [19]. Ozone can be an alternative oxidant to regenerate zeolites, which was rarely reported in literature. To our knowledge, only one study was reported in scientific journal by Reungoat et al., who obtained a complete recovery of the adsorption capacity for nitrobenzene [20]. Therefore, the objective of this study was firstly to investigate the adsorption of TCP onto a commercial zeolite, including both adsorption isotherm and kinetics, and secondly to apply ozone to regenerate the used zeolite. 2. Experimental 2.1. Materials
179
membrane and analyzed with HPLC to determine the concentration of TCP. The kinetic tests followed a similar procedure except that samples were taken in a given time interval during 6 h. The effect of initial pH was studied between pH 3 and pH 10, which was adjusted by adding 10 mM HCL or 10 mM NaOH at the beginning. The test solution had a volume of 50 mL and contained 1 g of zeolite and 100 mg/L of TCP. Other conditions were the same as above. A sample was taken after 24 h. All tests were conducted in triple and an average value was used. The TCP adsorption on the zeolite at given time was calculated according to: qt = V
C0 − Ct W
where qt is the adsorbed TCP at time t (mg/g); C0 initial TCP in water (mg/L); Ct TCP in water at time t (mg/L); V the solution volume (L); W the adsorbent weight (g). The equilibrium data (qe , Ce ) in the isotherm tests were calculated with the same method. 2.3. Regeneration with ozone A glass bottle was used to contain 10 g of the zeolite and 400 mL of 50 mg/L TCP solution. The zeolite was suspended by a magnetic stirrer. After 2-h adsorption, a sample was taken and then the ozone-containing gas (45–50 ppm) was bubbled into the solution for 30 min to regenerate the zeolite. The ozone in the water phase (7 mg/L) was measured with the Indigo colorimetric method [21]. Ozone was produced from pure oxygen with an ozone generator (SORBIOS GSF 010.2). After the ozonation, air was used to strip out the ozone residue in the solution. After air stripping, ozone in water phase was below 0.07 mg/L, which was almost the detection limit. A new adsorption test was started by replacing half of the old solution with a new TCP solution (100 mg/L) and the final TCP concentration was 50 mg/L again. A lot of intermediates may be formed during the above test, which could influence the adsorption of TCP. Therefore, FAU was washed and dried after 10 adsorption cycles, 1 g of which was added into 40 mL TCP solution (50 mg/L) for a clean adsorption test. A sample was taken after 2 h to compare the adsorption with the previous adsorption/regeneration cycles. The specific surface areas of FAU before and after 10 cycles of ozonation were determined with the BET method (Autosorb-1, Quantachrome) with nitrogen gas. 2.4. HPLC conditions
The adsorbent used in this study was a high-silica FAU zeolite ˚ diameter 1.6–2.5 mm), supplied by CWK Chemiew(pore size 9 A, erk Bad Köstritz GmbH. The chemical composition of zeolite is (H,Na)2 O·Al2 O3 ·xSiO2 ·nH2 O (x ≥ 30), as described by the supplier. 2,4,6-Trichlorophenol (98% purity) and methanol of HPLC grade were purchased from Sigma–Aldrich, Germany.
TCP in all water samples was determined with a HPLC system (Agilent 1200) equipped with a Phenomex Gemini-NX 5 m column. 50 L of sample was injected into the system and the elution was conducted with a mixture of methanol and Milipore water (85:15) at a flow rate of 1 mL/min. The column oven was set at 30 ◦ C. The eluted TCP was monitored with a UV detector at a wavelength of 290 nm. The quantification limit of TCP was 0.2 mg/L.
2.2. Adsorption tests
3. Results and discussion
All adsorption kinetics and isotherm tests were conducted with a batch equilibrium procedure. In the isotherm test, 1 g of zeolite was added to 50 mL of TCP solution with deionsed water at different concentrations (10, 25, 50, 75, and 100 mg/L). The adsorption was carried out at room temperature in a 100-mL Erlenmeyer flask which was shaken at a speed of ∼110 rpm. No buffer was applied in order to avoid the interference of ions. However, pH values in all test stayed around 6.2–6.5. Samples (1.5 mL) were taken out of the solution after 24 h when the equilibrium was assumed. The samples were filtrated through a syringe filter with a 0.45 m polycarbonate
3.1. Effects of initial pH The solution pH can influence the adsorption process via altering the ionic states of both adsorbent and adsorbate. In addition, the excessive H+ or OH− ions can also compete with an adsorbate for the active sites of adsorbent. In this study, the adsorption of TCP on FAU zeolite was tested in pH 3–10. As shown in Fig. 1, a clear declination of the adsorption capacity can be found with the increase of initial pH. The adsorption capacity dropped down from 4.3 mg/g at pH 3.5–2.6 mg/g at pH 10. The similar effect was also found in
180
Y. Zhang et al. / Journal of Hazardous Materials 271 (2014) 178–184
Fig. 1. The adsorption capacity of FAU for TCP at different initial pH values.
the adsorption of TCP onto different activated carbons [22,23] and activated clay [24]. The observed phenomenon can probably be explained by the adsorption mechanisms, mainly including hydrophobic interactions, electrostatic interactions, electron donor–acceptor complexes, metal coordination. The pore size of FAU is too small for TCP to be hold inside and therefore, the above mentioned interactions may mainly occur on the surface of FAU. TCP is a weak acidic compound and stays in the undissociated state at pH below its pKa value (6.23, SRC PhysProp Database). Its solubility in water tends to be elevated from acidic pH to basic pH [14]. Therefore, the dispersion interaction between TCP and the hydrophobic sites of zeolite may play a predominant role at acidic pH. On the contrary, the zeolite surface is negatively charged in the basic conditions and thus would repulse the TCP anions by electrostatic interactions. Nevertheless, there was still a certain amount of TCP adsorbed onto FAU in the basic condition, indicating that other mechanisms might take effect. The presence of 3 chlorine atoms makes TCP tending to be electron-deficient. Therefore, an electron donor–acceptor complex might form between TCP and nonbonding electrons at the siloxane surface of FAU. This complex was also proposed for the interaction of TCP and another synthetic zeolite SBA-15 [25]. In addition, TCP anions could also form a coordination complex with metal sites of FAU, which have been reported between chlorophenol and the surface of natural zeolite [14]. 3.2. Adsorption isotherm The isothermal tests were conducted at room temperature. The equilibrium was assumed after 24 h. Three models were applied to simulate the obtained TCP concentrations in solution (Ce , mg/L) and on FAU (qe , mg/g) at equilibrium, as follows: (a) Langmuir: qe =
qm KL Ce 1 + KL Ce
where qm is the Langmuir adsorption capacity (mg/g); KL , the Langmuir constant (L/mg). (b) Freundlich: 1/n
qe = KF Ce
where KF and n are Freundlich constants. (c) Temkin: qe =
RT ln(ACe ) bT
where R is the universal gas constant (m2 kg s−2 K−1 mol−1 ); T is the absolute solution temperature (K); A is the equilibrium binding constant (L/mg). Langmuir model failed to describe the experimental data since a negative adsorption capacity was obtained with the linear regression, which indicates that some assumptions of Langmuir model may not be applicable in the studied system. For instance, FAU contains both metallic and silica sites and many adsorption mechanisms may occur as discussed above, which cannot satisfy the homogeneous assumption of Langmuir. Nevertheless, Langmuir model was still successfully applied to describe the adsorption of 2-CP on zeolitic materials derived from bagasse fly ash (SiO2 /Al2 O3 = 4) [16] and 2-CP on Na-zeolite and Na/Ni-zeolite [26]. On the other hand, both Freundlich and Temkin can describe the adsorption isotherm data at a different extent of accuracy. The accuracies of both models were evaluated by the coefficient of determination (R2 ) for the linear regression and the coefficient of nonlinear chi squire (X2 ) for the accordance of the experimental and simulated data, as proposed by Ho [27]. X2 was calculated with: X2 =
N (qe,sim − qe,exp )2 i=1
qe,exp
where N is the number of data points; qe,exp , the experimental adsorption capacity at equilibrium (mg/g); qe,sim , the simulated adsorption capacity at equilibrium (mg/g). Freundlich model resulted in a better accuracy than Temkin in terms of both R2 and X2 coefficients (Fig. 2). The simulated curves with both models showed a comparable accordance with the experimental data in the middle concentration range. However, the Freundlich curve presented a much better accordance at both low and high concentrations than the Temkin curve (Fig. 2a). Especially, Temkin model predicted a negative value of qe,sim at low concentrations which significantly increased its value of X2 . Nevertheless, X2 of Temkin was still higher than that of Freundlich even after excluding the negative qe,sim . Therefore, it can be concluded that Freundlich described the experimental data with the best accuracy among the studied three models. The calculated values of Freundlich constants were 0.43 and 1.80 × 10−3 mg/g/(mg/L)1/n for n and Kf , respectively. The obtained Kf value equals to 1.99 mmol/g/(mmol/L)−1/n which is comparable with most of adsorbents for phenolics in literature, such as low-cost activated carbon and synthetic resins [9]. The adsorption capacity obtained in this study is approximately 10 times lower than that of TCP adsorbed on activated clay [24], which however was conducted at pH 4. As shown before, the acidic conditions could significantly promote the adsorption of TCP. Nevertheless, consideration should also be given to the large particle size of FAU (∼2 mm), which is beneficial for large scale applications but would decrease the external specific surface. For instance, activated carbons could have a specific surface area more than 1000 m2 /g [9], while the zeolite applied in this study had a value below 300 m2 /g as shown later. It can also be noticed in Fig. 2a that the experimental data tend to follow a typical S-shape curve. This curve was also observed in the adsorption of phenol onto n-octadecyl alcohol crystals and onto wool [28]. Moreover, it seems that the tested concentrations were in the low range and the saturation point of FAU had not been reached. 3.3. Adsorption kinetics The adsorption kinetic tests were conducted in shaken flasks at room temperature and samples were continually taken from the solution at given intervals. The adsorbed TCP (qt , mg/g) was
Y. Zhang et al. / Journal of Hazardous Materials 271 (2014) 178–184
181
Fig. 2. The experimental and simulated adsorption isotherm data (a) and the regression with Freundlich (b) and Temkin (c) models (qe , mg/g; Ce , mg/L).
calculated as mentioned before. The relation of qt and adsorption time t was simulated with following two models [6]: (a) Pseudo-second-order model: dqt = k2 (qe − qt )2 , or dt
t 1 1 = + t qt qe k2 q2e
where k2 is the reaction constant (g/mg/min). (b) Elovich equation: qt = ˇ ln ˛ˇ + ˇ ln t where ˛ is the initial sorption rate (g/mg/min); ˇ is an adsorption constant. The results of both kinetic models are presented in Table 1. The pseudo-second-order model obtained a higher R2 and a lower X2 than Elovich equation. Therefore, the former model is more suitable to simulate the studied adsorption process. With the obtained parameters, the pseudo-second-order model generated smooth curves which fit well with the experimental data, as presented in Fig. 3. In addition, the pseudo-first-order model was also tried but a poor linearity was obtained. It might be due to the low TCP concentrations in this study. It was previously found that the adsorption process with a low adsorbate concentration had an inclination to follow the pseudo-second-order model [29]. In fact, the pseudosecond-order model was also successfully applied to simulate the adsorption of 4-CP on a surfactant-modified natural zeolite [30] and TCP on activated carbon [31]. Another benefit of the pseudosecond-order model is that the experimental adsorption capacity at equilibrium is not required, which can even be deduced from the model [6]. As shown in Table 1, there is a very good correlation (X2 = 0.06) of qe,exp and qe,sim in the pseudo-second-order model. The adsorption rates (k2 ) in Table 1 are [33] in a comparable range with the data obtained for TCP adsorbed on activated carbon [31] and activated clay [24] but much higher than the rates of phenols adsorbed on zeolitic tuff [32]. One can notice that the adsorption rates (k2 ) of pseudo-secondorder kinetics varies with the initial adsorbate concentration (Table 1). A complex correlation between them has been
theoretically deduced and two empirical models were previously used in literature, as follows [29]: k2 =
C0 , aC0 + b
k2 = cC0d ,
or
or
Co = aCo + b k2
ln k2 = d ln C0 + ln C
where a, b, c, d are constants. Fig. 4 presents the regressions of k2 and C0 with above models. Both of them obtained a comparable coefficient of determination but the exponential model showed a much better coefficient of nonlinear chi squire between simulated k2 and experimental k2 . Therefore the exponential model is more suitable to predict the adsorption rate k2 for TCP adsorbed onto FAU in similar conditions.
Fig. 3. Time profiles of adsorbed TCP on FAU for different initial concentrations: Experimental data and simulated curves with the pseudo-second-order kinetics.
182
Y. Zhang et al. / Journal of Hazardous Materials 271 (2014) 178–184
Table 1 Results of simulating the kinetic data with pseudo-second-order model and Elovich equation. TCP
100 mg/L 75 mg/L 50 mg/L 25 mg/L 10 mg/L
Pseudo-second-order kinetics
Elovich equation
k2 (g/mg/min)
qe,sim (mg/g)
qe,exp (mg/g)
R2 (–)
9.82 × 10−3 1.57 × 10−2 2.35 × 10−2 5.71 × 10−2 6.13 × 10−2
3.18 2.29 1.33 0.45 0.15
3.06 2.09 1.19 0.40 0.11
0.998 0.999 0.995 0.999 0.995
3.4. Adsorbent regeneration When the adsorption equilibrium is reached, the adsorbent cannot adsorb the target adsorbate anymore. A regeneration step is in need to restore the adsorption capacity. Ozone, as one of the strongest oxidants, holds a standard electrode potential of 2.08 V and has been widely used to oxidize organic pollutants. Ozone was applied in this study to regenerate FAU loaded with TCP. For this purpose, two basic requirements should be fulfilled: (1) ozone can oxidize the adsorbate; (2) the adsorption capacity of adsorbent should not be compromised during the attack of ozone. Many studies have confirmed the oxidation of TCP by ozone [33]. Therefore, the adsorption capacity of FAU was emphasized in this study. The adsorption/regeneration cycle started with fresh FAU. After 2 h of adsorption, a sample was taken to calculate the removal of TCP and ozone gas was bubbled for 30 min into the solution where TCP both on FAU and in water was oxidized. The ozone residue was stripped out with aeration. After this regeneration treatment, TCP in the solution was almost eliminated (an average removal efficiency 97%). It was found that TCP adsorption onto FAU was significantly enhanced after the first regeneration with ozone: TCP removal from water was elevated from 52% up to 87%. As shown in Fig. 5, this elevated adsorption capacity was remained until 8th adsorption cycle. However, a dramatic dropdown of TCP removal was found
Fig. 4. Correlations between the pseudo-second-order rate and the initial adsorbate concentration in water: (a) exponential model; (b) division model (k2 , g/mg/min; C0 , mg/L).
X2 (–)
˛ (g/mg/min)
ˇ (–)
R2 (–)
0.06
0.63 0.89 1.12 3.06 5.24
0.62 0.46 0.28 0.094
0.976 0.969 0.946 0.975
X2 (–)
0.97
in 9th/10th adsorption, which might be due to two effects: matrix carbon and pH decrease as discussed below. Since only half of the solution was replaced in every cycle, the adsorption would be influenced by the intermediates formed during ozonation. Those matrix compounds could compete with TCP for the ozone consumption and finally might lead to an incomplete oxidation of adsorbed TCP. Consequently, a partial regeneration of FAU might occur. Moreover, the matrix compounds could also take over the adsorption site of FAU, which can decrease the adsorption of TCP. On the other hand, it is well known that some acids can be generated during the ozonation of TCP [33]. Firstly the chlorine atoms on TCP can be replaced with a hydroxyl group to form 2,6-dichlorobenzo-1,4-quinone and HCl. Secondly, the cleavage of the aromatic ring would generate aliphatic products which in turn may react with ozone to form organic acids, such as formic acid and acetic glycolic acids. The pH in the starting solution in our study was ∼6.3 and after the 10th adsorption cycle the pH decreased to 4.1. This acidic environment will promote the TCP adsorption as shown before. However, a significant decrease of TCP adsorption was observed at last adsorption cycles. It indicates that the effects of matrix compounds generated during ozonation might be more influential than the pH effect. More interestingly, the adsorption of TCP was kept at an elevated state in the clean adsorption with FAU after 10 cycles of use: 96% of TCP was adsorbed. Therefore, some modification should have happened on FAU, which could elevate its adsorption capacity. It was proposed that ozone can clean the pores of a FAU zeolite and thus lead to an elevated adsorption of nitrobenzene [20]. However, the researchers did not measure the specific surface of zeolite after ozonation. In this study, the BET specific surface area was analyzed
Fig. 5. Removal of TCP from water in the adsorption/regeneration cycles and in the clean adsorption test.
Y. Zhang et al. / Journal of Hazardous Materials 271 (2014) 178–184
183
process with a high accuracy. And the reaction rate k2 showed an exponential correlation with the initial TCP concentration. Both the BET surface area and TCP adsorption of FAU were significantly elevated by the regeneration with ozone. The adsorption capacity was kept stable for at least 8 cycles of adsorption and regeneration. Therefore, it is highly feasible to use FAU as an adsorbent to remove TCP from water and to apply ozone for the adsorbent regeneration. However, further studies are required to clarify the influence of other ingredients in real wastewater, such as metals, organics, which may compete with TCP for the adsorption sites. A column study is necessary to optimize the ozone regeneration and to step toward real applications. Acknowledgement
Fig. 6. BET surface area plots of studied zeolite before and after ozonation.
for the tested FAU zeolite before and after zonation (Fig. 6). It can be clearly found that ozonation largely increased the BET surface of FAU by 63% (from 289.29 to 470.75 m2 /g). However, further study is required to clarify the ozone effect on zeolite in terms of ozone dosage, reaction time, stability, etc. Another benefit with ozone is that the presence of zeolite could promote the decomposition of ozone with the generation of free radicals, which can unselectively oxidize more organic compounds than ozone itself [34]. The ozone decomposition mainly occurs on the surface of zeolite where the adsorbed compounds also locate. This localized reaction could avoid the radical consumption by scavenger compounds and the self-decay of radicals. Thus the free radicals can be used more efficiently. However, the regeneration with ozone was conducted in this study at acidic pH which is not favorable for the decomposition of ozone. Therefore, less free radicals could be expected at the 9th/10th cycles of adsorption and regeneration and partially contributed to an incomplete regeneration of zeolite. Since ozone can oxidize a broad range of organics with high kinetic rates [35], the adsorbent regeneration can be completed with ozone in a short time: 30 min in this study vs. several hours with the popular thermal calcination (e.g., 1 h at 540 ◦ C for a synthetic zeolite loaded with methylene blue [18]; 2 h at 450 ◦ C for coal fly ashes loaded with methylene blue [36]; 12 h at 450 ◦ C for FAU with nitrophenolics [37]). Actually, the ozonation time was not optimized in this study. It is highly possible that the time can be further reduced, which, however, should be evaluated in another study. Moreover, ozone decomposition can be promoted by some simple methods (e.g. increasing reaction pH or adding H2 O2 ) and then more free radicals can be produced. It could be used to regenerate zeolites loaded with organics which have a low reaction rate with ozone. However, further studies are required to evaluate the influence of free radicals on zeolites and other materials. In addition, various types of natural zeolites, which are generally cheaper than the synthetic ones, have been applied to adsorb pollutants from water [8]. However, the adsorption capacity of natural zeolites might be reduced during regeneration. Wang et al. tried thermal calcination and Fenton reaction separately to regenerate a type of natural zeolite loaded with methylene blue [18]. They found that only 60% of adsorption capacity was recovered by both methods. 4. Conclusions and outlook TCP can be effectively removed from water by adsorbing onto FAU zeolite, especially at lower pH. The adsorption isotherm can be well described with Freundlich model within the studied conditions. Pseudo-second-order kinetics can simulate the adsorption
We appreciate the help of Prof. Dietmar Stephan and Mr. Gerrit Land for the analysis of BET specific surface. References [1] J. Gao, L. Liu, X. Liu, H. Zhou, S. Huang, Z. Wang, Levels and spatial distribution of chlorophenols – 2,4-dichlorophenol, 2,4,6-trichlorophenol, and pentachlorophenol in surface water of China, Chemosphere 71 (2008) 1181–1187. [2] A.O. Olaniran, E.O. Igbinosa, Chlorophenols and other related derivatives of environmental concern: properties, distribution and microbial degradation processes, Chemosphere 83 (2011) 1297–1306. [3] D. Yin, H. Zhu, P. Hu, Q. Zhao, Genotoxic effect of 2,4,6-trichlorophenol on P53 gene in zebrafish liver, Environ. Toxicol. Chem. 28 (2009) 603–608. [4] L. Xing, H. Liu, J.P. Giesy, H. Yu, pH-dependent aquatic criteria for 2,4-dichlorophenol, 2,4,6-trichlorophenol and pentachlorophenol, Sci. Total Environ. 441 (2012) 125–131. [5] G. Busca, S. Berardinelli, C. Resini, L. Arrighi, Technologies for the removal of phenol from fluid streams: a short review of recent developments, J. Hazard. Mater. 160 (2008) 265–288. [6] Y.-S. Ho, Review of second-order models for adsorption systems, J. Hazard. Mater. 136 (2006) 681–689. ˛ [7] A. Dabrowski, P. Podko´scielny, Z. Hubicki, M. Barczak, Adsorption of phenolic compounds by activated carbon – a critical review, Chemosphere 58 (2005) 1049–1070. [8] M. Ahmaruzzaman, Adsorption of phenolic compounds on low-cost adsorbents: a review, Adv. Colloid Interface Sci. 143 (2008) 48–67. [9] S.-H. Lin, R.-S. Juang, Adsorption of phenol and its derivatives from water using synthetic resins and low-cost natural adsorbents: a review, J. Envrion. Manage. 90 (2009) 1336–1349. ´ [10] R. Cherbanski, E. Molga, Intensification of desorption processes by use of microwaves – an overview of possible applications and industrial perspectives, Chem. Eng. Process. 48 (2009) 48–58. [11] D. Bathen, Physical waves in adsorption technology – an overview, Sep. Purif. Technol. 33 (2003) 163–177. [12] H. Valdés, M. Sánchez-Polo, J. Rivera-Utrilla, C.A. Zaror, Effect of ozone treatment on surface properties of activated carbon, Langmuir 18 (2002) 2111–2116. [13] M. Khalid, G. Joly, A. Renaud, P. Magnoux, Removal of phenol from water by adsorption using zeolites, Ind. Eng. Chem. Res. 43 (2004) 5275–5280. [14] R.I. Yousef, B. El-Eswed, The effect of pH on the adsorption of phenol and chlorophenols onto natural zeolite, Colloids and Surf. A 334 (2009) 92–99. [15] B. Koubaissy, G. Joly, I. Batonneau-Gener, P. Magnoux, Adsorptive removal of aromatic compounds present in wastewater by using dealuminated faujasite zeolite, Ind. Eng. Chem. Res. 50 (2011) 5705–5713. [16] B. Shah, R. Tailor, A. Shah, Sorptive sequestration of 2-chlorophenol by zeolitic materials derived from bagasse fly ash, J. Chem. Technol. Biotechnol. 86 (2011) 1265–1275. [17] N. Koryabkina, J.A. Bergendahl, R.W. Thompson, A. Giaya, Adsorption of disinfection byproducts on hydrophobic zeolites with regeneration by advanced oxidation, Microporous Mesoporous Mater. 104 (2007) 77–82. [18] S. Wang, H. Li, S. Xie, S. Liu, L. Xu, Physical and chemical regeneration of zeolitic adsorbents for dye removal in wastewater treatment, Chemosphere 65 (2006) 82–87. [19] R. Gonzalez-Olmos, F.-D. Kopinke, K. Mackenzie, A. Georgi, Hydrophobic Fezeolites for removal of MTBE from water by combination of adsorption and oxidation, Environ. Sci. Technol. 47 (2013) 2353–2360. [20] J. Reungoat, J.S. Pic, M.H. Manéro, H. Debellefontaine, Adsorption of nitrobenzene from water onto high silica zeolites and regeneration by ozone, Sep. Sci. Technol. 42 (2007) 1447–1463. [21] H. Bader, J. Hoigné, Determination of ozone in water by the indigo method, Water Res. 15 (1981) 449–456. [22] B.H. Hameed, I.A.W. Tan, A.L. Ahmad, Adsorption isotherm, kinetic modeling and mechanism of 2,4,6-trichlorophenol on coconut husk-based activated carbon, Chem. Eng. J. 144 (2008) 235–244.
184
Y. Zhang et al. / Journal of Hazardous Materials 271 (2014) 178–184
[23] I.A.W. Tan, A.L. Ahmad, B.H. Hameed, Adsorption isotherms, kinetics, thermodynamics and desorption studies of 2,4,6-trichlorophenol on oil palm empty fruit bunch-based activated carbon, J. Hazard. Mater. 164 (2009) 473–482. [24] B.H. Hameed, Equilibrium and kinetics studies of 2,4,6-trichlorophenol adsorption onto activated clay, Colloids and Surf. A 307 (2007) 45–52. [25] Q. Qin, K. Liu, D. Fu, H. Gao, Effect of chlorine content of chlorophenols on their adsorption by mesoporous SBA-15, J. Environ. Sci. 24 (2012) 1411–1417. [26] B. Okolo, C. Park, M.A. Keane, Interaction of phenol and chlorophenols with activated carbon and synthetic zeolites in aqueous media, J. Colloid Interface Sci. 226 (2000) 308–317. [27] Y.-S. Ho, Selection of optimum sorption isotherm, Carbon 42 (2004) 2115–2116. [28] C.H. Giles, T.H. MacEwan, S.N. Nakhwa, D. Smith, 786. Studies in adsorption. Part XI. A system of classification of solution adsorption isotherms, and its use in diagnosis of adsorption mechanisms and in measurement of specific surface areas of solids, J. Chem. Soc. 0 (1960) 3973–3993. [29] S. Azizian, Kinetic models of sorption: a theoretical analysis, J. Colloid Interface Sci. 276 (2004) 47–52. [30] A. Kuleyin, Removal of phenol and 4-chlorophenol by surfactant-modified natural zeolite, J. Hazard. Mater. 144 (2007) 307–315.
[31] M. Radhika, K. Palanivelu, Adsorptive removal of chlorophenols from aqueous solution by low cost adsorbent – kinetics and isotherm analysis, J. Hazard. Mater. 138 (2006) 116–124. [32] R.I. Yousef, B. El-Eswed, A.a.H. Al-Muhtaseb, Adsorption characteristics of natural zeolites as solid adsorbents for phenol removal from aqueous solutions: kinetics, mechanism, and thermodynamics studies, Chem. Eng. J. 171 (2011) 1143–1149. [33] N. Graham, W. Chu, C. Lau, Observations of 2,4,6-trichlorophenol degradation by ozone, Chemosphere 51 (2003) 237–243. [34] H. Valdés, V.J. Farfán, J.A. Manoli, C.A. Zaror, Catalytic ozone aqueous decomposition promoted by natural zeolite and volcanic sand, J. Hazard. Mater. 165 (2009) 915–922. [35] U. von Gunten, Ozonation of drinking water. Part I. Oxidation kinetics and product formation, Water Res. 37 (2003) 1443–1467. [36] Z. Sun, C. Li, D. Wu, Removal of methylene blue from aqueous solution by adsorption onto zeolite synthesized from coal fly ash and its thermal regeneration, J. Chem. Technol. Biotechnol. 85 (2010) 845–850. [37] B. Koubaissy, G. Joly, P. Magnoux, Adsorption and competitive adsorption on zeolites of nitrophenol compounds present in wastewater, Ind. Eng. Chem. Res. 47 (2008) 9558–9565.