Activated carbon–clay composite as an effective adsorbent from the spent bleaching sorbent of olive pomace oil: Process optimization and adsorption of acid blue 29 and methylene blue

Activated carbon–clay composite as an effective adsorbent from the spent bleaching sorbent of olive pomace oil: Process optimization and adsorption of acid blue 29 and methylene blue

Accepted Manuscript Title: Activated carbon–clay composite as an effective adsorbent from the spent bleaching sorbent of olive pomace oil: Process opt...

944KB Sizes 0 Downloads 23 Views

Accepted Manuscript Title: Activated carbon–clay composite as an effective adsorbent from the spent bleaching sorbent of olive pomace oil: Process optimization and adsorption of acid blue 29 and methylene blue Authors: F. Marrakchi, M. Bouaziz, B.H. Hameed PII: DOI: Reference:

S0263-8762(17)30580-4 https://doi.org/10.1016/j.cherd.2017.10.015 CHERD 2856

To appear in: Received date: Revised date: Accepted date:

8-5-2017 5-9-2017 12-10-2017

Please cite this article as: Marrakchi, F., Bouaziz, M., Hameed, B.H., Activated carbon–clay composite as an effective adsorbent from the spent bleaching sorbent of olive pomace oil: Process optimization and adsorption of acid blue 29 and methylene blue.Chemical Engineering Research and Design https://doi.org/10.1016/j.cherd.2017.10.015 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Activated carbon–clay composite as an effective adsorbent from the spent bleaching sorbent of olive pomace oil: Process optimization and adsorption of acid blue 29 and methylene blue

F. Marrakchia,b, M. Bouazizb,c, B.H. Hameeda,*

a

School of Chemical Engineering, Engineering Campus,

University of Science Malaysia, 14300 Nibong Tebal, Penang, Malaysia

b

Laboratoire d’Electrochimie et Environnement, Ecole Nationale d’Ingénieurs de

Sfax, Université de Sfax, BP 1173, 3038 Sfax, Tunisia

c

Institut Supérieur de Biotechnologie de Sfax,

Université de Sfax, BP 1175, 3038 Sfax, Tunisia

* Corresponding author. Tel.: +6045996422; Fax: +6045941013 E-mail address: [email protected] (B.H. Hameed)

Highlights 

Activated carbon–clay composite from the spent bleaching sorbent of pomace oil.



Langmuir model best fitted the isotherm data.



Pseudo-second-order model fits the kinetic data for both dyes.

1

Abstract An activated carbon–clay (ACC) composite was prepared using spent bleaching sorbent generated from refined olive pomace oil through carbonization followed by K2CO3 activation. The adsorption-removal efficiencies for acid blue 29 (AB 29) and methylene blue (MB) of the developed ACC were examined. The K2CO3 activation process was optimized using Box– Behnken design. The optimum conditions were a K2CO3 impregnation ratio of 1:1, an activation temperature of 800 °C, and an activation time of 120 min, under which the optimized ACC achieved 83.81% AB 29 and 96.20% MB removal. The surface properties of the optimized ACC were characterized by different physicochemical measures and techniques, including surface area, point of zero charge, scanning electron microscopy, and Fourier transform infrared spectroscopy. Pseudo-second-order and Langmuir models best fitted the adsorption kinetics and isotherm experimental data. The maximum monolayer adsorption capacity of the ACC was 104.83 and 178.64 mg/g for AB 29 and MB, respectively, at 30 °C. All these results indicated the potential application of ACC in dye adsorption.

Keywords Spent bleaching sorbent, Carbonization, K2CO3, Batch adsorption, Methylene blue, Acid blue 29

1. Introduction The food industry is the largest industry worldwide, thus making it a significant contributor to the generation of massive waste, which poses severe harm to the environment (Otles et al., 2015). Edible oil refineries are part of the food industry that produce stable and pleasantly tasting refined oil by removing undesirable materials (i.e., wax, gum, free fatty 2

acids, trace metals, color pigments, pesticide residues, soap trace, and odoriferous materials) (Marrakchi et al., 2015). Crude oil refining processes (including degumming, neutralization, washing, drying, bleaching, and deodorization) generate numerous by-products, such as soap stocks, acidic water, spent bleaching sorbent, and deodorizer distillates. Such wastes are poisonous and hazardous to the environment (Dumont and Narine, 2007). Tunisian olive pomace oils, which account for 64,040 tons of the total production, are mostly intended for refining or industrial uses (Lahyani et al., 2015). The bleaching of olive pomace oils is conducted with both acid-activated clay and activated carbon in the range of 0.5–1 wt.% and 0.05–0.1wt.% of olive pomace oil, respectively (Petrakis, 2006). Acidactivated clay, which is also called bleaching earth, consists mainly of bentonite or montmorillonite (Al2O3.4SiO2.nH2O). The main purpose of using bleaching earth is to adsorb color matters (chlorophylls and carotenoids) and other contaminants (soap residues, phospholipids, heavy metal traces, and hydroperoxides) (Beshara and Cheeseman, 2014; Loh et al., 2013). Acid-activated clay exhibits a hydrophilic nature and has very low adsorption capacities toward nonpolar organic molecules. Thus, activated carbon is typically used as an admixture with acid-activated clay to entrap polycyclic aromatic hydrocarbons (PAHs) occasionally present in crude oils (Petrakis, 2006). After being used, the spent bleaching sorbent (SBS) composite of bleaching earth and activated carbon contains up to 30% of retained oil (Beshara and Cheeseman, 2014; Huang and Chang, 2010; Loh et al., 2013). Disposal of untreated SBS in landfills or farmlands in proximity to factories leads to the oxidation of residual unsaturated oil to the point of auto-ignition, thereby causing fire hazards (spontaneous combustion) and greenhouse gas (GHG) emission. Therefore, SBS discarded from olive pomace refineries is considered a hazardous industrial waste (i.e. flammable waste) (Beshara and Cheeseman, 2014; Loh et al., 2013; Mana et al., 2011). For this reason,

3

researchers have attempted to develop treatment strategies for SBS waste generated from olive pomace oil refineries and transform it into valuable products. Carbonization of SBS yields a carbon–clay (CC) product as a result of the conversion of the retained olive pomace oil into valuable commodity carbon. The bifunctional sides, namely, carbon and clay, of the carbonized material can be texturally and chemically enhanced by alkaline activation, which can be achieved by using potassium carbonate K2CO3 (Tsai et al., 2004). K2CO3 has been used as activating agent in the preparation of activated carbons from agricultural wastes (Foo and Hameed, 2012a, 2012b, 2012c, 2012d). Therefore, reusing SBS and enhancing its properties as an adsorbent for wastewater treatment address the problem of SBS disposal. This work focuses on the conversion of SBS into an activated carbon–clay composite (ACC) by carbonization followed by chemical activation with K2CO3. The effects of K2CO3 impregnation ratio, activation temperature, and time on ACC preparation were investigated by applying a response surface methodology (RSM) with the Box-Behnken design (BBD). The potential application of ACC as an adsorbent for liquid phases was tested for the uptake of acid blue 29 (AB 29) and methylene blue (MB).

2. Materials and methods 2.1. Preparation of activated carbon–clay adsorbent SBS was obtained by bleaching olive pomace oil from a refinery in Sfax, Tunisia. Oven-dried (105 °C, 24 h) SBS was used throughout the preparation process. Approximately 10 g of SBS was placed in a ceramic crucible and subjected to carbonization under nitrogen (99%) flow (0.1 L/min) at 500 ºC for 60 min. The carbonization temperature was determined using thermogravimetric analysis. The resulting carbonized

4

material, named CC, was washed with 0.1 M HCl solution (R&M Chemicals Company, Malaysia) for 1 h and then with hot distilled water until the pH became neutral. The CC product was subjected to K2CO3 activation, which was optimized by applying RSM (Duan et al., 2014). A BBD was established to investigate the empirical relationships among total yield, dye removal percentage, and three potential factors, namely, x1 (impregnation ratio, IR; 𝑊𝐾2 𝐶𝑂3 : 𝑊𝐶𝐶 ), x2 (activation temperature; °C), and x3 (activation time; min). For each factor, three different coded levels corresponded to low (−1), intermediate (0), or high (+1) levels. Table 1 lists the factors corresponding to each coded level. The BBD matrix was generated and analyzed on a Design-Experts-Software (version 7.0.0, Stat-Ease, Inc, Minneapolis, USA). Run number (N) in the matrix was expressed by 𝑁 = 2𝑝 (𝑝 − 1) + 𝑛𝑐 , where p and nc denoted the number of independent variables and central points, respectively (Ferreira et al., 2007). Therefore, 15 runs were generated, which included 12 factorial points and 3 central points, as indicated in Table 1. For each run, 5 g of CC was mixed with K2CO3 pellets (R&M Chemicals Company, Malaysia) at the desired IR (x1 ). Distilled water (100 mL) was then added while stirring to dissolve all K2CO3 pellets. The mixture was oven-dried overnight at 105 °C. The dried material was activated in a stainlesssteel reactor surrounded by a furnace under continuous nitrogen gas flow (0.1 L/min) at a heating rate of 10 °C/min until the activation temperature (x2 ) was reached at a precise activation time (x3 ). After activation, the furnace was cooled under nitrogen gas. The produced sample was washed with HCl (1.0 M) and repeatedly washed with hot distilled water until the pH became neutral, and the sample was then dried overnight at 105 °C. The final product was designated as activated carbon–clay (ACC).

5

The relationship among responses Y1 (the total yield of ACC, %), Y2 (MB removal percentage, %), Y3 (AB 29 removal percentage, %), and the independent variables can be approximated by the following quadratic model: ̂𝑖 = 𝑏0 + 𝑏1 𝑥1 + 𝑏2 𝑥2 + 𝑏3 𝑥3 + 𝑏12 𝑥1 𝑥2 + 𝑏13 𝑥1 𝑥3 + 𝑏23 𝑥2 𝑥3 + 𝑏11 𝑥12 + 𝑏22 𝑥22 𝑌 + 𝑏33 𝑥32

(1)

where 𝑌𝑖 = 𝑌̂𝑖 + 𝑒 (𝑌̂𝑖 is the predicted response output, and 𝑌𝑖 , the measured response); e is the error, b0 is model constant, b1 to b33 are the regression coefficients; x1 , x2 , and x3 are the individual effects; x1 x2 , x1 x3 , and x2 x3 are the interaction effects; and x21 , x22 , and x23 are the quadratic effects. The fitness and significance of the models were analyzed by ANOVA based on BBD. The model and the model terms were considered significant when the p-values were less than 0.05 (Das and Mishra, 2017). Then, the optimum conditions for the three variables, namely, impregnation ratio, activation temperature, and time, were determined by implementing a desirability approach on BBD, which searched for the combination of factors that simultaneously maximized the three outputs (Y1, Y2, and Y3) (Ghani et al., 2017). The obtained optimum parameters were then also applied to the single-step activation of SBS. 2.2. Characterization of the materials Scanning electron microscopy (SEM) was performed using a Zeiss Supra 35VP to identify the topographic characteristics of the materials. BET analysis was conducted using a Micromeritics ASAP 2020 apparatus. A Fourier transform infrared (FTIR) spectrophotometer Model 2000 was used to obtain IR spectra in the intervals of 4000–400 cm-1. The zero point of charge (pHzpc) of the ACC was defined according to a previously reported drift method (Benhouria et al., 2015).

6

2.3. Acid blue 29 and methylene blue dyes adsorption analysis Batch adsorption analyses of AB 29 (Sigma-Aldrich Chemical Company) and MB (Merck Company, Malaysia) were conducted in 250 mL conical flasks by adding 1 g/L of adsorbent (ACC) dosage with various initial dye concentrations (25–400 mg/L) while maintaining the solution pH without any adjustments. Afterward, the flasks were agitated at 150 rpm in a thermostatic water bath shaker at 30 °C, 40 °C, and 50 °C. The effect of the solution pH (3 to 11) on the dye adsorption uptake was examined using an initial concentration of 100 mg/L at 30 °C while the other parameters were kept constant. The solution pH was adjusted by adding a few drops of HCl or NaOH (0.1/1.0 M) solutions. Then, the pH was measured by a EUTECH pH meter. The residual AB 29 and MB concentrations in the solution were determined with a UV-vis spectrophotometer (Shimadzu, Model UV 1601, Japan) at a maximum wavelength of 602 and 665 nm for AB 29 and MB, respectively. The equilibrium adsorption capacity qe (mg/g) of AB 29 and MB can be expressed as follows: 𝑞𝑒 =

(𝐶0 − 𝐶𝑒 )𝑉 𝑊

(2)

where C0 and Ce (mg/L) are the initial and equilibrium concentrations of AB 29 and MB, respectively, in aqueous phase; W (g) is the mass of ACC; and V (L) denotes the volume of the AB 29 and MB solutions. The adsorption capacity qt (mg/g) at time t (min) was calculated as follows: 𝑞𝑡 =

(𝐶0 − 𝐶𝑡 )𝑉 𝑊

(3)

3. Results and discussion

7

3.1. Effect of the activation process parameters on the yield and dye removal of activated carbon–clay The total yield was determined by Y1=Yc × Ya, where Yc is the yield obtained after the carbonization of SBS, and Ya is the K2CO3 activation yield. As revealed by the total yield data in Table 1, Y1 varied between 27.07% and 39.52%. The total yield (Y1) can be represented by Eq. (4): 𝑌1 = +38.22 − 2.77 𝑥1 − 0.56 𝑥2 − 1.21 𝑥3 − 2.15 𝑥1 𝑥2 − 1.09 𝑥12 − 4.13 𝑥22 − 2.27 𝑥32

(4)

The adequacy and accuracy of Eq. (4) were validated by the correlation coefficient (R2 = 0.99) and ANOVA. The p-value of the quadratic model (Y1) was 0.0005 less than 0.05, thus proving the significance of the model (Table S1). The perturbation plot was obtained by assuming constant values for two of the factors at the center points (x1 = 3 g/g; x2 = 700 °C; x3 = 90 min), whereas the third factor varied along the adopted range (Fig. S1 (a)). The total yield (Y1) of the ACC decreased with an increase in the three parameters (x1 , x2 , and x3 ), with the highest yield corresponding to the lowest point among all three factors. ANOVA revealed that the activation time significantly affected the ACC yield (p = 0.0054) (Table S1). Specifically, an increase in the activation time (x3 ) promoted the volatilization of the volatile matter, leading to the reduction of yield amount. Moreover, the greater is the increase in x1 from 1.0 to 5.0 and in x2 from 600 °C to 800 °C, the higher are the reaction rates between CC and K2CO3. Such increase in x1 and x2 also causes an increased amount of volatile compounds to be released because of the dehydration reaction. Consequently, the ACC weight loss would be intensified. Similar findings were observed by Gurten et al. in preparing activated carbon from tea waste by K2CO3 treatment (Gurten et al., 2012). Compared with AB 29, MB was preferentially adsorbed onto the ACC adsorbent. The robustness of the model (Eq. 5) was verified by the correlation coefficient R2 of 0.91 and the 8

insignificant lack of fit (p-value > 0.05), suggesting that the experimental values corresponded with the model prediction. 𝑌2 = +75.48 + 8.14 𝑥1 + 17.08 𝑥2 + 5.17 𝑥3

(5)

ANOVA revealed that MB adsorption was significantly affected in decreasing order by activation temperature (p <0.0001) > chemical impregnation ratio (p =0.0012) > activation time (p = 0.0187) (Table S1). The effects of K2CO3 activation parameters on MB removal via ACC were illustrated by 3D response surface (Fig. S1 (b)) and main factor (Fig. S1 (c)) plots. MB removal from aqueous solution increased markedly and linearly with the increase in x1 , x2 , and x3 , as indicated in Figs. S1(b)–S1(c). At the same time, an increase in activation temperature and time under inert conditions reduced K2CO3 into K, K2O, CO, and CO2 (Foo and Hameed, 2012a; McKee, 1983). The K formed during K2CO3 activation would diffuse into the internal structure of the CC, thus forming new pores and broadening existing micropores to mesopores; consequently, the pore volume and the surface area were enhanced, ultimately further intensifying response Y2 (Foo and Hameed, 2012a). AB 29 removal increased from 8.86% to 75.58% (Table 1). The experimental and theoretical values were assumed to have satisfactory correlation, given that R2 agreed well with R2adj . According to the ANOVA results (Table S1), 97% of the AB 29 removal was well predicted by the model (Eq. 6), indicating that the terms considered in the proposed models were significant enough to make adequate predictions. 𝑌3 = +12.15 − 6.84 𝑥1 + 25.69 𝑥2 + 1.87 𝑥3 + 9.16 𝑥12 + 17.23 𝑥22 + 7.69 𝑥32

(6)

ANOVA results of Y3 indicated that AB29 removal was significantly affected by the IR (p = 0.0083), the activation temperature (p < 0.0001), the quadratic effect of these factors (𝑥12 , 𝑥22 ), and the activation time (𝑥32 ). Fig. S1(d) shows the adverse effects of the IR (x1 ) and the activation temperature (x2 ) on the adsorption of AB 29 at an activation time of 90 min. As shown, Y3 sharply increased with increased the activation temperature (x2 ). Nonetheless, AB 9

29 removal from aqueous solution was negatively affected by the IR in a quadratic manner. The reason is that the high IR of K2CO3 led to intensified attacks of silicon dioxide (SiO2). Consequently, soluble silicates (SiO23 ), which give more negative charges, were formed on the surface of the ACC. Thus, AB29 adsorption was decreased, whereas MB adsorption was enhanced (Tsai et al., 2004). 3.1.3. Optimization of K2CO3 activation To simultaneously maximize the three responses, a maximum value (0.92) of the desirability function was obtained at an IR of 1:1 (𝑊𝐾2 𝐶𝑂3 : 𝑊𝐶𝐶 ), an activation temperature of 800 °C, and an activation time of 120 min. Similar optimum parameters have been reported when K2CO3 was utilized as an activating agent (Adinata et al., 2007; Gurten et al., 2012; Mestre et al., 2011). Under these conditions, the predicted responses for the produced ACC, MB, and AB29 removal were 33.88%, 89.59%, 80.60%, respectively. To validate the adequacy of the model equations in predicting the optimum response values, an experiment was conducted under the optimal conditions, and an ACC yield of 34.56% as well as MB and AB 29 removal rates of 96.20% and 83.81% were achieved. The experimental values were close to the predicted ones, with small errors. Therefore, these optimal operation conditions can be applied in the preparation of optimal ACC composites derived from SBS. 3.2. Characterization of the materials 3.2.1. Surface morphology Figs. 1(a)–1(c) show the SEM images of SBS, CC, and ACC, respectively. Fig. 1a displays the compacted and dense morphological structure of SBS. Such structure can be attributed to the entrapment of residual pomace oil, which contain impurities, such color pigments, free fatty acid, oxidation products, on the surface of the virgin adsorbent used in the 10

bleaching step (a mixture of bleaching earth and activated carbon). In Fig. 1b, the surface of the CC was still saturated with impurities and the appearance of a few pores resulting from the loss of volatile compounds because of the brunt of the residual oil and the release of carbon in the form of CO and CO2. Fig. 1c shows that the chemical activation with K2CO3 can effectively widen pores and create new ones on the CC surface, thereby enhancing the surface area and porous structure of the ACC composite. 3.2.2. BET surface area and pore volume of CC and ACC N2 adsorption-desorption isotherms (Fig. 2a) and pore size distributions (Fig. 2b) were investigated to examine the surface area and pore structure of the CC and the ACC. In Figs. 2a–2b, both the CC and the ACC present isotherms of types IV (IUPAC system). In addition, the presence of H2 hysteresis loops indicated the presence of mesopores. Table 2 presents the surface physical parameters of the CC and the ACC. As shown, after the K2CO3 activation of the CC, the BET surface area was enhanced from 168.45 m2/g to 355.11 m2/g. Moreover, the micropore and mesopore pore volumes increased from 0.000637 cm3/g to 0.048 cm3/g and from 0.29 cm3/g to 0.37 cm3/g, respectively. The development of pores and the increase in pore volume were ascribed to the dehydrating effect of K2CO3 that penetrated deep into the ACC structure, causing the creation of pores. Furthermore, the average pore diameter slightly increased to 5.83 nm (Table 2), suggesting that the ACC adsorbent was mainly mesoporous in accordance with the IUPAC classification. 3.2.3. Surface chemistry The SBS infrared spectrum (Fig. 3) confirmed the characteristics of the edible oil characteristics and the composite structure consisting of bleaching earth and activated carbon. The bands at 3470.09–3547.24 cm-1 were assigned to the bleaching earth interlayer water molecule stretching vibrations. The 1046.43 cm-1 band was attributed to Si-O stretching. The 11

798.56 cm-1 band was ascribed to Si-O quartz impurity vibration. The bands at 523 and 463.90 cm-1 were assigned to Si-O-Al bending. The small peaks at 1628 and 1457.28 cm-1 corresponded to the C-H of aromatic stretching and the conjugated C=C stretching of the aromatic ring of the activated carbon on the surface of SBS. The presence of residual pomace olive oil was detected by the peaks at 2927.10 and 2858.63 cm-1 corresponding to the C-H stretching of the saturated carbonaceous chains of oil, 1724.40 cm-1 assigned to the C=O of ester carbonyl vibration, and a band at 3417.12 cm-1 ascribed to the hydroxyl OH stretching vibrations of carboxylic acids (free fatty acids). These findings are consistent with the reported results on spent bleaching earth derived from vegetable oil refining process (Mana et al., 2008, 2007; Meziti and Boukerroui, 2011). The disappearance of hydroxyl groups from silica, the Si-O vibration of quartz impurities, the C-H stretching, and the carbonyl bands were observed after carbonization. The peaks at 1628 and 1457.28 cm-1 shifted to 1634.37 and 1384.95 cm-1, respectively. This shifting was presumably due to the additional aromaticity of the carbon surface of the resulting material (CC). K2CO3 activation of the CC at 800 °C led to the disappearance of the peak at 523.70 cm1

and the decrease in the intensity of the Si-O vibration at 466.79 cm-1. This process also

displaced other peaks in greater frequencies as a result of the deformation of the bleaching earth structure. In turn, this deformation led to the appearance of the band at 796.67 cm-1, which matched the deformation of the Si-O-Si and Si-O-Al vibrations in the amorphous material (ACC). The FTIR results indicated that the SBS surface was modified after carbonization followed by K2CO3 activation. Comparison of the FTIR spectra of the ACC and the MB-adsorbed ACC revealed that the peaks at 3458.52 and 1057.90 cm−1 shifted to 3457.55 and 1055.11 cm−1, respectively, with the 795.67 cm−1 band disappeared. For the AB 29-loaded ACC, the peaks at 466.79 and

12

795.67 cm−1 disappeared, and the peaks at 3458.52 and 1057.90 cm−1 shifted to 3459.59 and 1053.17 cm−1. The shifting and disappearance of peaks indicated that AB 29 was adsorbed by silanol groups. 3.3. Adsorption studies 3.3.1. Effect of initial dye concentration and contact time Figs. 4a and 4b present the effect of initial dye concentrations at 30 °C for various time intervals. As shown, when the initial dye concentration (C0) was increased from 25 mg/L to 400 mg/L, the residual concentration increased from 1.67 mg/l to 286.07 mg/L (93.12%– 39.50 %) for AB 29 (Fig. 4a) and from 0.15 mg/l to 179.25 mg/L (99.32%–52.36 %) for MB (Fig. 4b). Therefore, dye removal was highly dependent on the initial concentration of the adsorbate. The decrease in the removal percentages of AB 29 and MB as the initial dye concentration increased reflected the increase in the concentration gradient driving forces (Altıntıg et al., 2017; Seyahmazegi et al., 2016). Moreover, the adsorption of AB 29 and MB was a time-dependent process. Figs. 4a and 4b shows that the adsorption of both dyes onto the ACC composite was rapid at the initial time intervals and then gradually increased until equilibrium was attained. For Ab 29, the required time to reach equilibrium was estimated at 5 h for 25 and 50 mg/L, and 26 h for 100, 200, 300, 400 mg/L. In comparison, equilibrium was achieved for MB at around 5 h for all studied concentrations, demonstrating the better affinity of the ACC to cationic dyes. The rapid adsorption at an initial step was due to the accessibility of the active sites on the ACC surface. After a fast uptake, a transitional stage occurred, wherein the adsorption rate slightly increased as a result of the diffusion of the slow dyes into the pores of the ACC composite adsorbent, leading to equilibrium (Sayğılı and Güzel, 2016). 3.3.2. Kinetic modeling 13

AB 29 and MB kinetic adsorption on the ACC composite were investigated using the Lagergren's pseudo-first-order (PFO) (Lagergren, 1898) and Ho's pseudo-second-order (PSO) models (Ho and McKay, 1999), which are expressed as follows: 𝑞𝑡 = 𝑞𝑒 (1 − 𝑒 −𝑘1 𝑡 ) 𝑞𝑡 =

(7)

𝑞𝑒2 𝑘2 𝑡 1 + 𝑞𝑒 𝑘2 𝑡

(8)

where k1 (1/min) and k2 (g/(mg min)) are the PFO and PSO constant rates. The suitability of the kinetic models (Figs. S2(a) and S2(b)) was examined by calculating R2 and RMSE, as well as by visualizing the correspondence between the calculated (qe,cal) and experimental (qe,exp) adsorption capacities. 2

𝑅 =1−

∑𝑛𝑛=1(𝑞𝑒.𝑒𝑥𝑝.𝑛 − 𝑞𝑒.𝑐𝑎𝑙.𝑛 )

2

∑𝑛𝑛=1(𝑞𝑒.𝑒𝑥𝑝.𝑛 − ̅̅̅̅̅̅̅̅̅) 𝑞𝑒.𝑒𝑥𝑝.𝑛

(9)

2

1

𝑅𝑀𝑆𝐸 = √𝑛−1 ∑𝑛𝑛=1(𝑞𝑒.𝑒𝑥𝑝.𝑛 − 𝑞𝑒.𝑐𝑎𝑙.𝑛 )2

(10)

For both PFO and PSO models (Table 3), the R2 values were close to unity. However, the RMSE values of the PSO kinetic model were inferior to those of the PFO model. In addition, qe,cal was significantly closer to qe,exp for PSO than for PFO. Therefore, the PSO kinetic model was more suitable for the adsorption of AB 29 and MB onto ACC. Previous studies focusing on regenerated spent materials and porous adsorbents also support the use of PSO kinetic models for basic and acid dye adsorption (Mana et al., 2011, 2007; Sayğılı and Güzel, 2015; Wang et al., 2016).

3.3.3. Isotherm modeling

The equilibrium data (Ce vs. qe) were fitted to two isotherm models. The nonlinear Langmuir isotherm model (Langmuir, 1916) is expressed by Eq. (11):

14

𝑞𝑒 =

𝑞𝑚𝑎𝑥 𝑘𝑙 𝐶𝑒 1 + 𝑘𝑙 𝐶𝑒

(11)

where kl (L/mg) and qmax (mg/g) are the Langmuir constant and maximum adsorption capacities of AB 29 and MB per unit mass of ACC, respectively. The nonlinear Freundlich isotherm model (Freundlich, 1906) is expressed by Eq. (12): 1/𝑛

𝑞𝑒 = 𝑘𝑓 𝐶𝑒

(12)

where kf ((mg/g) (L/mg)1/n) and n denote the Freundlich adsorption capacity and intensity, respectively. Table 4 presents the isotherm parameters, R2, and RMSE for the Ab29 and Mb dyes, which were determined by fitting the equilibrium experimental data (plots not shown) at 30 °C, 40 °C, and 50 °C). The R2 values close to unity and the minimum RMSE values indicated that the Langmuir model was more appropriate than the Freundlich model for both dyes. Thus, the ACC surface was homogenous at areas where monolayer adsorption occurred (Mana et al., 2011, 2007). When the temperature was increased from 30 °C to 50 °C, the maximum adsorption capacity, qmax, decreased from 104.83 mg/g to 96.67 mg/g for AB 29 and from 178.64 mg/g to 171.18 mg/g for MB, indicating the exothermicity nature of the adsorption process. Furthermore, the decrease in the Freundlich constant (kf ) values from 40.67 ((mg/g) (L/mg)1/n) to 32.68 ((mg/g) (L/mg)1/n) for AB 29 and from 65.09 ((mg/g) (L/mg)1/n) to 58.22 ((mg/g) (L/mg)1/n) for MB indicated that lower temperatures were favorable for adsorption. Table 5 lists the maximum adsorption capacities (qmax) for different anionic and cationic dyes on different adsorbents. In this study, the qmax values of the ACC for MB and AB 29 were comparable to those reported in the literature (Mana et al., 2011, 2007; Santos and

15

Boaventura, 2016; Srivastava and Sillanpää, 2017; Tsai et al., 2005, 2004; Xin-hui et al., 2014). 3.3.5. Effect of solution pH Fig. 5 shows the effects of solution pH on the ACC adsorption uptake at equilibrium of AB 29 and MB. The adsorption uptake was pH dependent. The equilibrium adsorption uptake of AB 29 slightly increased from 75.40 mg/g to 79.15 mg/g with an increase in pH from 3 to 7, and then decreased to 68.77 mg/g at pH 11. For MB, the equilibrium adsorption uptake increased from 76.28 mg/g to 104.67 mg/g with an increase in pH from 3 to 9, and then decreased to 78.68 mg/g at pH 11. The pHzpc value for ACC was 7.72. Given that pH < pHpzc, the protonation of H+ activities dominated the adsorbent surfaces, causing them to become positively charged. Thus, an electrostatic interaction can occur between the protonated SiOH groups and the anionic AB 29 dye, as well as a repulsion reaction between the MB molecules and the ACC adsorbent active surfaces. These repulsive activities reduced the adsorption of MB onto the ACC at lower pH ranges. By contrast, when pH > pHpzc, the ACC surface became negatively charged, thus enhancing the adsorption of the cationic MB dye by silanol and the hydroxyl group via hydrogen bonds.

4. Conclusion The two-step process of carbonization followed by K2CO3 activation was found to be an effective method for converting SBS to a potential adsorbent (ACC) for the removal of AB 29 and MB from aqueous solution. An ACC composite with bifunctional sides, namely, carbon and clay, was obtained at an impregnation ratio of 1:1, an activation temperature of 800 °C, and an activation time of 120 min. The optimized ACC composite had a surface area 16

of 355.11 m2/g, mesopore volume of 0.37 cm3/g, and removal rates of 83.81% and 96.20% for AB 29 and MB. The adsorption analysis of both dyes revealed that the kinetic data fit the pseudo-second-order model and that the adsorption isotherm data fit the Langmuir model with a maximum adsorption capacity of 104.83 mg/g for AB 29 and 178.64 mg/g for MB at 30 °C.

17

References Adinata, D., Wan Daud, W.M.A., Aroua, M.K., 2007. Preparation and characterization of activated carbon from palm shell by chemical activation with K2CO3. Bioresour. Technol. 98, 145–149. Altıntıg, E., Altundag, H., Tuzen, M., Sarı, A., 2017. Effective removal of methylene blue from aqueous solutions using magnetic loaded activated carbon as novel adsorbent. Chem. Eng. Res. Des. 122, 151–163. Benhouria, A., Islam, M.A., Zaghouane-Boudiaf, H., Boutahala, M., Hameed, B.H., 2015. Calcium alginate–bentonite–activated carbon composite beads as highly effective adsorbent for methylene blue. Chem. Eng. J. 270, 621–630. Beshara, A., Cheeseman, C.R., 2014. Reuse of spent bleaching earth by polymerisation of residual organics. Waste Manag. 34, 1770–1774. Das, S., Mishra, S., 2017. Box-Behnken statistical design to optimize preparation of activated carbon from Limonia acidissima shell with desirability approach. J. Environ. Chem. Eng. 5, 588–600. Duan, X., Zhang, Z., Srinivasakannan, C., Wang, F., Liang, J., 2014. Regeneration of spent catalyst from vinyl acetate synthesis as porous carbon: Process optimization using RSM. Chem. Eng. Res. Des. 92, 1249–1256. Dumont, M.-J., Narine, S.S., 2007. Soapstock and deodorizer distillates from North American vegetable oils: Review on their characterization, extraction and utilization. Food Res. Int. 40, 957–974. Ferreira, S.L.C., Bruns, R.E., Ferreira, H.S., Matos, G.D., David, J.M., Brandão, G.C., da Silva, E.G.P., Portugal, L.A., dos Reis, P.S., Souza, A.S., dos Santos, W.N.L., 2007. Box-Behnken design: An alternative for the optimization of analytical methods. Anal. Chim. Acta 597, 179–186.

18

Foo, K.Y., Hameed, B.H., 2012a. Mesoporous activated carbon from wood sawdust by K2CO3 activation using microwave heating. Bioresour. Technol. 111, 425–432. Foo, K.Y., Hameed, B.H., 2012b. Factors affecting the carbon yield and adsorption capability of the mangosteen peel activated carbon prepared by microwave assisted K2CO3 activation. Chem. Eng. J. 180, 66–74. Foo, K.Y., Hameed, B.H., 2012c. Porous structure and adsorptive properties of pineapple peel based activated carbons prepared via microwave assisted KOH and K2CO3 activation. Microporous Mesoporous Mater. 148, 191–195. Foo, K.Y., Hameed, B.H., 2012d. Preparation, characterization and evaluation of adsorptive properties of orange peel based activated carbon via microwave induced K2CO3 activation. Bioresour. Technol. 104, 679–686. Freundlich, H.M.F., 1906. Over the adsorption in solution. J Phys Chem 57, e470. Ghani, Z.A., Yusoff, M.S., Zaman, N.Q., Zamri, M.F.M.A., Andas, J., 2017. Optimization of preparation conditions for activated carbon from banana pseudo-stem using response surface methodology on removal of color and COD from landfill leachate. Waste Manag. 62, 177–187. Gurten, I.I., Ozmak, M., Yagmur, E., Aktas, Z., 2012. Preparation and characterisation of activated carbon from waste tea using K2CO3. Biomass Bioenergy 37, 73–81. Ho, Y.S., McKay, G., 1999. Pseudo-second order model for sorption processes. Process Biochem. 34, 451–465. Huang, Y.-P., Chang, J.I., 2010. Biodiesel production from residual oils recovered from spent bleaching earth. Renew. Energy 35, 269–274. Lagergren, S., 1898. About the theory of so-called adsorption of soluble substances 24, 1–39.

19

Lahyani, R., Coelho, L.C., Khemakhem, M., Laporte, G., Semet, F., 2015. A multicompartment vehicle routing problem arising in the collection of olive oil in Tunisia. Omega 51, 1–10. Langmuir, I., 1916. The constitution and fundamental properties of solids and liquids. J. Am. Chem. Soc 38, 2221–2295. Loh, S.K., James, S., Ngatiman, M., Cheong, K.Y., Choo, Y.M., Lim, W.S., 2013. Enhancement of palm oil refinery waste–Spent bleaching earth (SBE) into bio organic fertilizer and their effects on crop biomass growth. Ind. Crops Prod. 49, 775–781. Mana, M., Ouali, M.-S., de Menorval, L.C., 2007. Removal of basic dyes from aqueous solutions with a treated spent bleaching earth. J. Colloid Interface Sci. 307, 9–16. Mana, M., Ouali, M.S., de Menorval, L.C., Zajac, J.J., Charnay, C., 2011. Regeneration of spent bleaching earth by treatment with cethyltrimethylammonium bromide for application in elimination of acid dye. Chem. Eng. J. 174, 275–280. Mana, M., Ouali, M.S., Lindheimer, M., de Menorval, L.C., 2008. Removal of lead from aqueous solutions with a treated spent bleaching earth. J. Hazard. Mater. 159, 358– 364. Marrakchi, F., Kriaa, K., Hadrich, B., Kechaou, N., 2015. Experimental investigation of processing parameters and effects of degumming, neutralization and bleaching on lampante virgin olive oil’s quality. Food Bioprod. Process. 94, 124–135. McKee, D.W., 1983. Mechanisms of the alkali metal catalysed gasification of carbon. Fuel, Fundamentals of Catalytic Coal and Carbon Gasification 62, 170–175. Mestre, A.S., Bexiga, A.S., Proença, M., Andrade, M., Pinto, M.L., Matos, I., Fonseca, I.M., Carvalho, A.P., 2011. Activated carbons from sisal waste by chemical activation with K2CO3: Kinetics of paracetamol and ibuprofen removal from aqueous solution. Bioresour. Technol. 102, 8253–8260.

20

Meziti, C., Boukerroui, A., 2011. Regeneration of a solid waste from an edible oil refinery. Ceram. Int. 37, 1953–1957. Otles, S., Despoudi, S., Bucatariu, C., Kartal, C., 2015. Chapter 1 - Food waste management, valorization, and sustainability in the food industry A2 - Galanakis, Charis M., in: Food Waste Recovery. Academic Press, San Diego, pp. 3–23. Petrakis, C., 2006. 9 - Olive Oil Extraction A2 - Boskou, Dimitrios, in: Olive Oil (Second Edition). AOCS Press, pp. 191–223. Santos, S.C.R., Boaventura, R.A.R., 2016. Adsorption of cationic and anionic azo dyes on sepiolite clay: Equilibrium and kinetic studies in batch mode. J. Environ. Chem. Eng. 4, 1473–1483. Sayğılı, H., Güzel, F., 2016. Effective removal of tetracycline from aqueous solution using activated carbon prepared from tomato (Lycopersicon esculentum Mill.) industrial processing waste. Ecotoxicol. Environ. Saf. 131, 22–29. Sayğılı, H., Güzel, F., 2015. Performance of new mesoporous carbon sorbent prepared from grape industrial processing wastes for malachite green and congo red removal. Chem. Eng. Res. Des. 100, 27–38. Seyahmazegi, E.N., Mohammad-Rezaei, R., Razmi, H., 2016. Multiwall carbon nanotubes decorated on calcined eggshell waste as a novel nano-sorbent: Application for anionic dye Congo red removal. Chem. Eng. Res. Des. 109, 824–834. Srivastava, V., Sillanpää, M., 2017. Synthesis of [email protected] nanocomposite for rapid scavenging of cationic and anionic dyes from synthetic wastewater. J. Environ. Sci. 51, 97–110. Tsai, W.T., Chang, Y.M., Lai, C.W., Lo, C.C., 2005. Adsorption of ethyl violet dye in aqueous solution by regenerated spent bleaching earth. J. Colloid Interface Sci. 289, 333–338.

21

Tsai, W.T., Hsien, K.J., Yang, J.M., 2004. Silica adsorbent prepared from spent diatomaceous earth and its application to removal of dye from aqueous solution. J. Colloid Interface Sci. 275, 428–433. Wang, Y., Xie, Y., Zhang, Y., Tang, S., Guo, C., Wu, J., Lau, R., 2016. Anionic and cationic dyes adsorption on porous poly-melamine-formaldehyde polymer. Chem. Eng. Res. Des. 114, 258–267. Xin-hui, D., Srinivasakannan, C., Jin-sheng, L., 2014. Process optimization of thermal regeneration of spent coal based activated carbon using steam and application to methylene blue dye adsorption. J. Taiwan Inst. Chem. Eng. 45, 1618–1627.

22

Figure captures Fig. 1:

SEM images of (a) SBS, (b) CC, and (c) ACC at 9000× magnification

Fig. 2:

(a) Nitrogen adsorption isotherm and (b) pore distribution for CC and ACC

Fig. 3:

FTIR spectra of SBS, CC, ACC, MB-loaded ACC and AB 29-loaded ACC

Fig. 4:

Effect of initial dye concentration on (a) AB 29 and (b) MB removal efficiency (agitation speed: 125 rpm; ACC dosage: 1 g/L; adsorption time: 26 h)

Fig. 5:

Solution pH effect on AB 29 and MB adsorption uptake; dye concentration (100 mg/L; agitation speed: 125 rpm; ACC dosage: 1 g/L; adsorption time: 26 h; adsorption temperature: 30 °C)

a)

(

(

(

c)

b)

Fig. 1

23

Quantity adsorbed (cm³/g STP)

300

(

Adsorption

Desorption

250

a)

200

150

ACC 100

CC 50

0 0.0

0.2

0.4

0.6

0.8

1.0

Relative Pressure (P/P0)

dV/dlog(D) Pore volume (cm³/g·Å)

0.8

(

CC

ACC

0.7 0.6b) 0.5 0.4 0.3 0.2 0.1 0 0

200

400

600

800

Pore diameter (Å)

Fig. 2

24

1000

1200

1400

4000

3600

3200

2800

2400

Fig. 3

25

2000

Wavenumber (cm-1)

1600

1046.43

1457.28

798.56

1725.40 1627.99

2927.10 2858.63

3547.24 3470.09 3415.12

1200 523.70 467.76

1053.18

1384.95

1634.74

3453.69

800

466.79

795.67

1087.90

1388.81

1636.67

3458.52

Transmittance (%)

466.79

1055.11

1392.66

1634.74

3457.55

1053.18

1397.49

1637.64

3459.48

AB 29-loaded ACC composite

MB-loaded ACC composite

Activated Carbon-Clay (ACC)

Caron-Clay (CC)

Spent Bleaching Sorbent (SBS)

523.70 463.90

400

400

Ct (mg/L)

300

(

25 mg/L 150 mg/L 400 mg/L

50 mg/L 200 mg/L

100 mg/L 300 mg/L

a)

200

100

0 0

400

800

1200

1600

Time (min)

400

(

25 mg/L 150 mg/L 400 mg/L

50 mg/L 200 mg/L

100 mg/L 300 mg/L

Ct (mg/L)

300b)

200

100

0 0

400

800

Time (min) Fig. 4

26

1200

1600

100

100

80

80 60 60 40 40

MB

20

20

AB 29 0

0 2

4

6

8

pH Fig. 5

27

10

12

qe, AB 29 (mg/g)

qe, MB (mg/g)

120

List of Tables Table 1:

Box-Behnken design matrix for K2CO3 activation together with the ACC total yield and the removal responses of AB 29 and MB

Factors

Levels

Unit

Low (-1)

Center (0)

High (+1)

𝑊𝐾2 𝐶𝑂3 /𝑤𝐶𝐶

1

3

5

x2 : Activation temperature

°C

600

700

800

x3 : Activation time

min

60

90

120

x1 : Impregnation ratio

Factors x1

x2

x3

ACC composite yield (%)

1

3

800

120

30.38

97.86

72.00

2

3

600

120

31.27

67.96

12.04

3

5

700

120

31.36

88.50

21.30

4

5

700

60

33.06

70.60

22.91

5

1

700

120

35.52

69.46

34.28

6

1

600

90

34.66

43.31

13.46

7

3

800

60

32.86

91.88

52.00

8

5

600

90

33.19

70.66

10.73

9

3

600

60

32.79

51.37

12.25

10

1

700

60

39.52

68.57

39.13

11

5

800

90

27.07

98.43

54.40

12

1

800

90

37.13

81.76

75.58

13c

3

700

90

39.01

78.62

8.86

14c

3

700

90

37.52

82.30

14.82

3 700 90 : Center points

38.20

70.88

12.77

Runs

15c c

Table 2:

MB removal (%)

AB 29 removal (%)

BET and Langmuir surface area, micropore and mesopore volumes, and average pore sizes of CC and ACC 28

Surface physical parameters BET surface area

Carbon-clay

Activated carbon-clay

(CC)

(ACC)

168.45

355.11

262.09

487.31

0.000637

0.048

0.29

0.37

5.52

5.83

(m2/g) Langmuir surface area (m2/g) Micropore volume (cm3/g) Mesopore volume (cm3/g) Average pore diameter (nm)

29

Table 3:

Kinetic parameters for the batch adsorption of AB 29 and MB onto ACC (agitation time: 26 h, agitation speed: 125 rpm, initial dye concentration: 25400 mg/L, dye solution volume: 0.2 L and ACC mass: 0.2 g)

C0 AB 29 (mg/L)

qe,exp (mg/g)

25

Pseudo- First-Order 𝑞𝑡 = 𝑞𝑒 (1 − 𝑒 −𝑘1 𝑡 )

Pseudo- Second-Order 𝑞𝑒2 𝑘2 𝑡 𝑞𝑡 = 1 + 𝑞𝑒 𝑘2 𝑡 qe,cal k2×10-4 g (mg/g) R2 RMSE ( ) mg.min

qe,cal (mg/g)

k1×10-2 (1/min)

R2

RMSE

22.59

22.08

1.03

0.99

0.50

24.34

6.13

0.98

0.95

50

49.13

48.28

1.61

0.99

1.24

52.03

5.05

0.98

1.97

100

87.08

80.95

0.97

0.96

4.54

90.32

1.49

0.99

2.27

150

93.59

83.92

1.16

0.93

6.64

93.52

1.70

0.98

3.81

200

98.62

89.19

1.40

0.92

6.90

97.85

2.12

0.98

3.08

300

105.32

96.33

1.01

0.93

7.13

107.19

1.33

0.99

3.38

400

108.93

97.64

2.07

0.90

8.13

108.71

2.79

0.97

4.13

C0 MB (mg/L)

qe,exp (mg/g)

25

Pseudo- First-Order 𝑞𝑡 = 𝑞𝑒 (1 − 𝑒 −𝑘1 𝑡 )

Pseudo- Second-Order 𝑞𝑒2 𝑘2 𝑡 𝑞𝑡 = 1 + 𝑞𝑒 𝑘2 𝑡 qe,cal k2×10-4 g (mg/g) R2 RMSE ( ) mg.min

qe,cal (mg/g)

k1×10-2 (1/min)

R2

RMSE

22.00

21.75

5.60

0.99

0.34

22.23

72.30

0.99

0.22

50

43.10

42.98

6.90

0.99

0.44

43.66

54.50

0.99

0.28

100

91.15

89.27

6.30

0.99

1.81

91.19

19.70

0.99

0.31

150

129.85

123.42

4.70

0.98

4.27

127.74

7.98

0.99

1.97

200

154.25

149.11

3.20

0.98

3.90

156.12

4.04

0.99

2.73

300

168.50

163.71

5.18

0.98

5.33

168.93

6.84

0.99

1.33

400

197.00

187.92

3.13

0.97

7.58

197.80

2.98

0.99

1.76

30

Table 4:

Langmuir and Freundlich adsorption isotherm parameters for the batch adsorption of AB 29 and MB onto ACC at 30, 40, and 50 °C

Isotherms

D

Constants

Temper ature °C

qmax (mg/g)

30

104.83

40

102.03

Langmuir:

50

96.67

𝑞𝑚𝑎𝑥 𝑘𝑙 𝐶𝑒 𝑞𝑒 = 1 + 𝑘𝑙 𝐶𝑒

30

178.64

40

173.07

50

171.18

yes

A B 29

M B

8

A

Freundlich B 29

: 1/𝑛

𝑞𝑒 = 𝑘𝑓 𝐶𝑒

M B

40.67

40

35.29

50

32.68

30

65.09

40

59.29

31

7

5

0.9

0.2

0.9

0.2

0.9

0.2

0.9

R2

5.4

0.9 2

5.0

MSE 1

0.9

4.9

1 2.35

0.8 9

4.6

1 3.49

0.9 7

4.5 3

R

2.53

1

6

9 .71

n

6

9 .50

8

6

1 7.39

8

5

5 .96

5

3

4 .41

8

5

7

0.9

0.1

9

MSE

.27

9

3

R

0.9

0.1

kf ((mg/g)(L/ mg)1/n) 30

R2

kl (L/ mg) 0.2

1 4.36

0.9 4

1 8.56

50

58.22

32

4.5 1

0.9 4

1 7.76

Table 5:

Efficiency assessment of different adsorbents used for anionic and cationic dye adsorption from aqueous solutions

Adsorbents

Process

Dye

Activated

Thermal treatment +

Methylene

Carbon-clay

K2CO3 activation

Activated

Thermal treatment +

Carbon-clay

29

Cethyltrimethylammonium bromide

Basic treated

Silica from

Acid Black 10B Methylene

Sodium hydroxide

spent bleaching earth

Blue

Thermal treatment +

diatomaceous earth

NaOH activation

Sepiolite

Without treatment

Sepiolite

Without treatment

[email protected]

Methylene Blue Basic Red 46 Direct Blue 85 Methylene

Malachite +bentonite

nanocomposite [email protected]

Blue Congo

Malachite +bentonite

nanocomposite Regenerated

Red Ethyl

KOH activation

spent bleaching earth Regenerated spent coal

Acid Blue

K2CO3 activation

Modified spent bleaching earth

Blue

Violet

Thermal regeneration with steam

Methylene Blue

33

qmax (mg/g)

References

178.65

This work

104.83

This work

100.00

85.40

56.20

(Mana et al., 2011) (Mana et al., 2007) (Tsai et al., 2004)

(Santos 110.00 and Boaventura, 2016) (Santos 232.00 and Boaventura, 2016) (Srivastava 277.77 and Sillanpää, 2017) (Srivastava 238.09 and Sillanpää, 2017) (Tsai et al., 40.70 2005) 375.93

(Xin-hui et al., 2014)