Adsorption of cationic and anionic azo dyes on sepiolite clay: Equilibrium and kinetic studies in batch mode

Accepted Manuscript Title: Adsorption of cationic and anionic azo dyes on sepiolite clay: Equilibrium and kinetic studies in batch mode Author: S´ılvia C.R. Santos Rui A.R. Boaventura PII: DOI: Reference:

S2213-3437(16)30053-7 http://dx.doi.org/doi:10.1016/j.jece.2016.02.009 JECE 976

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

10-12-2015 2-2-2016 3-2-2016

Please cite this article as: S´ilvia C.R.Santos, Rui A.R.Boaventura, Adsorption of cationic and anionic azo dyes on sepiolite clay: Equilibrium and kinetic studies in batch mode, Journal of Environmental Chemical Engineering http://dx.doi.org/10.1016/j.jece.2016.02.009 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.

Adsorption of cationic and anionic azo dyes on sepiolite clay: equilibrium and kinetic studies in batch mode

Sílvia C.R. Santos* and Rui A.R. Boaventura

Laboratory of Separation and Reaction Engineering - Laboratory of Catalysis and Materials (LSRE-LCM), Faculdade de Engenharia da Universidade do Porto, Rua Dr. Roberto Frias, 4200-465 Porto, Portugal

* Corresponding author: Tel. +351 220414976; Fax: +351 225081674. E-mail address: [email protected]

1

Abstract In the present work, sepiolite clay adsorptive properties were investigated for textile azo dyes in aqueous solution. The clay (78 wt.% sepiolite) was characterized in terms of physical, textural, chemical and mineralogical properties. Basic Red 46 and Direct Blue 85 azo dyes were selected as adsorbates. Adsorption equilibrium data were successfully fitted to Freundlich and Langmuir equations. For Basic Red, the maximum adsorption capacities, predicted by the Langmuir model, were 110 mg/g (at 25 ºC) and 310 mg/g (at 35 ºC). The value obtained at 25 ºC, matched perfectly the cation exchange capacity of the clay. For Direct Blue, the adsorption capacity was 332 mg/g. Adsorption equilibrium was also evaluated for Direct Blue dye in a synthetic effluent, containing salt and auxiliary dyeing chemicals. In this mixture, the amount of dye adsorbed by sepiolite decreased, but very considerable values were still reached (106 mg/g). Adsorption kinetics was studied and modeled. Homogeneous solid diffusion coefficient and effective pore diffusivities were calculated for both dyes. Sepiolite clay showed to be an advantageous adsorbent in terms of price and versatility, being effective under a wide pH range, for both anionic and cationic dyes. Its use as adsorbent does not require further purification or chemical modifications, which is beneficial for the environment and for the economy of the treatment.

Keywords: Adsorption, Sepiolite, Textile dye, Equilibrium, Kinetics

2

1. Introduction Textile industries are one of the largest consumers of water and hence producers of liquid effluents. The dyeing and finishing processes generate volumes of wastewaters in the range 45-450 m3 per ton of product [1], containing salts, acids, bases, additives and unfixed dyes. A suitable treatment of these complex wastewaters should lead to a strong decline on the organic load, and desirably to a complete elimination of the residual color. The entrance of dyes into water bodies harms their aesthetic condition, compromising many water uses and affects aquatic life [2]. In particular, azo dyes and their derivatives are mutagenic and carcinogenic [3, 4], posing serious risks to aquatic life and human health. Biological and coagulation/flocculation processes are usually the most economical options to treat high volumes of wastewaters. Aerobic biological treatment is however ineffective against many dyes [5, 6], and decolorization by anaerobic via can yield dangerous aromatic amines [7]. Color removal by coagulation using traditional iron and aluminium salts is not consistently effective [8] and produces a huge volume of sludge. Many alternative treatments have been proposed to remove effectively the dyes from the wastewaters, such as chemical reduction, membrane separation, chemical oxidation, advanced oxidation processes (AOPs). In spite of the good decolorization results achieved and the specific advantages or disadvantages of each one, the cost of implementation or operation is not attractive for the industry. Adsorption has been viewed as an economic and effective option for the removal of contaminants from aqueous solution. Adsorption by activated carbon is especially useful as post treatment, after activated sludge or coagulation/flocculation [9, 10]. Combining a conventional process with a tertiary level treatment, wastewater characteristics can be polished in the final adsorptive removal and a safe discharge can be achieved. However, the use of activated carbon (granular or powdered) entails considerable acquisition and regeneration costs. From an environmental point of view, it is also worth noting the high energy that is spent in the production of activated carbons. As a contribution to overcome these problems and to achieve color removal in a cost-effective manner, a wide range of natural and waste materials, raw and chemically modified, 3

have been proposed as alternatives [6, 11-16]. In spite of the considerable research on this topic, the state of knowledge still presents limitations, such as low adsorption capacities for anionic adsorbates, safety risks related to matrices not completely immobilized (heavy metal or organic matter possible leaching) [15], inadequate grain size or mechanical properties and the need of chemical modifications to give or improve the sorption capacity [17]. Sepiolite is a natural hydrated magnesium silicate. Its structure is composed by alternation of blocks and cavities (tunnels) that grow up in the fibre direction [18]. Rectangular channels contain some exchangeable Ca and Mg cations and two types of water: bound water (molecules coordinating Mg atoms at the broken bond surfaces of the channels) and zeolitic water (clusters filling the empty space in the channels and hydrogen-bonded to the bound water) [19]. As the silica sheets are discontinuous, silanol groups (SiOH) are present at the border of each block in the external surface of the silicate. The ability of sepiolite to uptake cationic dyes has been demonstrated in some works [20-23], although these studies have been mostly focused on methylene blue and crystal violet dyes, which are not typically used in textile industry. The adsorption of monovalent cationic species is a result of the cation exchange capacity of sepiolite, but also of its neutral sites, where basic dyes can be adsorbed [22]. Further treatments such as saturation with sodium or calcium [24] showed to enhance a little more the adsorption capacity of sepiolite for methylene blue. Regarding anionic dyes, natural sepiolite provided relatively low adsorbed amounts of an azo acid dye [25]. In its untreated form, sepiolite was not also able to remove reactive dyes, requiring a chemical modification with a quaternary amine (hexadecyltrimethylammonium bromide, HTAB) in order to create a positive surface charge and provide suitable adsorption [17]. Literature is relatively restricted concerning direct dyes removal by sepiolite. This kind of dyes are mainly applied in cellulosic fibres and are still commonly used in industry (instead of reactive dyes), due to their lower cost and easy application [6]. In the present work, a sepiolite clay material, usually commercialized as absorbent for industrial spillages and cat and pet litters, was studied as adsorbent for cationic and anionic textile azo dyes. 4

The clay was characterized (in terms of physical, mineralogical and chemical properties). Adsorption kinetics and equilibrium were studied and modeled.

2. Materials and Methods 2.1. Dyes Two commercial textile dyes, kindly supplied by Dystar, were studied in this work: Basic Red 46 (BR) and Direct Blue 85 (DB). Fig. 1 illustrates the molecular structures of the dyes. BR is a cationic dye, usually used for dyeing acrylic fibres. DB is a direct anionic dye, suitable for cotton, viscose and modal dyeing, but also applicable for polyamide, wool and silk. Both dyes are from the azo chemical class. Aqueous solutions were simply prepared by dissolving the required mass of dye (dried at 105 ºC) in distilled water.

A synthetic dyeing effluent (designated as SDE) was also prepared using DB, salt (NaCl) and auxiliary chemicals usually added to dyebaths in the industrial operations: a wetting agent (Sera Wet C-AS), which is anionic, a lubrificant agent (Sera Lube M-CF), non-ionic nature, and a sequestering agent, anionic (Sera Quest M-PP). The preparation of SDE was as follows [6]: 0.25 g of each one of the three auxiliary chemicals (wetting, lubrificant and sequestering) were dissolved in 250 mL warm distilled water; 150 mg of DB dye were added and the solution was heated to 100 ºC; after 15 min, 2.5 g NaCl were added and the temperature was kept constant for 45 minutes, more; after cooling, the volume was made up to 1L; this last dilution (250 mL to 1L) simulates 3 washing baths (2 rinsing and 1 conditioning/softening baths) after dyeing. The whole procedure was based on a typical scheme used for cellulosic fibres dyeing with direct dyes and in other details provided by the dye supplier.

2.2. Sepiolite characterization The sepiolite clay used as adsorbent is a commercial product with a grain size in the range 0.250– 0.600 mm. According to its supplier it was composed by 80 wt.% sepiolite. Some properties of the 5

clay (mean particle size, porosimetry, apparent and real densities, surface area by N2 adsorption, cation exchange capacity, mineralogical and chemical analysis) were reported in a previous study [23]. Grain size distribution was obtained by laser diffraction, using water as dispersive medium. Mercury porosimetry (Quantachrome, Poremaster 60), helium picnometry (Quantachrome UPY100) and nitrogen adsorption at 77 K (Coulter Omnisorp 100 CX) were used to assess physical and textural properties. It is known that surface areas measured by dry (N2 adsorption) and wet methods (water or methylene blue adsorption) can render different values [26]. Although sepiolite is not a particular swelling mineral, the methylene blue (MB) adsorption method was used in this work to assess the surface area of sepiolite and to compare with the previous reported value (obtained by N2 adsorption). Sepiolite (50 mg) was stirred with methylene blue solutions of different concentrations (45-175 mg/L) for 2 days. Liquid and solid phases were then separated by centrifugation and the remaining MB dye was measured spectrophotometrically (660 nm). The amount of MB added and MB adsorbed were plotted and the “end-point” (complete cation replacement) was then identified. Surface area (SMB, m2/g) was calculated by Eq. 1 [26]:

S MB  QEP 

L (1)  M  103

In Eq. 1, QEP (mg/g) is the MB adsorbed amount at the “end-point”, L the Avogadro constant, M the molar mass of MB dye (g/mol) and σ (m2/molecule) the surface area covered by one MB molecule, typically assumed to be 130 Å2/molecule. Points of zero net proton charge (PZNPC) of sepiolite clay were determined at three different ionic strengths, by pH-drift method [21]. The clay (3.0 g/L) was stirred with NaCl solutions 0.1, 0.01 and 0.001 mol/L, at different initial pH values (from 2 to 12). After 48 h under orbital shaking, equilibrium pH values were recorded. Blanks were also carried out in order to discount the effect of CO2 from air. Equilibrium pH values were plotted against the initial ones, for each electrolyte concentration.

6

The intersection point with a straight line of slope equal to 1 (initial pH equals the final pH), gives the pH at which a zero net adsorption of H+/OH-, designated as PZNPC.

2.3. Analytical methods BR and DB concentrations in aqueous solution were analyzed by UV-Vis spectrometry (Helios Alpha Unicam) at the wavelengths of maximum absorbance, 525 nm and 590 nm, respectively. Calibration straight lines (R2=1.0) were obtained in the ranges 0.3-11 mg/L (for BR) and 2-30 mg/L (DB), and with limits of detection for BR and DB analysis (determined based on blank measurements) of 0.1 and 0.6 mg/L, respectively. Before analyses, all samples from adsorption tests were centrifuged for 5 min (Mini Spin Eppendorff, 13 400 rpm) in order to completely separate the liquid from the sepiolite particles.

2.4. Equilibrium adsorption studies The pH is usually referred to as an important factor affecting the extent of adsorption. In the case of sepiolite, it was previously observed (based on the outcome of PZNPC assays) that suspensions would provide high pH values (around 9), due to the high buffering capacity of the clay. In order to evaluate the effect of pH on the adsorption, preliminary adsorption tests were performed using 100 mL of dye solution (200 mg/L, for BR and 150 mg/L for DB) at different initial pH values (adjusted with NaOH or HCl solutions) and 0.10 g of sepiolite. The suspensions were stirred at 25 ºC in Erlenmeyer flasks for 24 h (BR) or 4 days (DB). No further adjustments were done in the pH during the contact time. The amount of dye adsorbed on sepiolite in equilibrium (qeq, mg/g) was calculated by a mass balance equation (Eq. 2), and plotted as a function of initial and final pH values. q eq 

(C 0  C eq )  v

(2)

m

where C0 (mg/L) is the initial dye concentration, Ceq the final (equilibrium) dye concentration, m (g) the adsorbent mass used and v (L) the volume of dye solution.

7

Adsorption equilibrium isotherms were determined in a similar procedure, using the same initial dye concentrations, but variable sepiolite dosages (m/v), in the range 0.3-2.7 g/L. The pH was kept approx. constant (pH 8.7±0.5) by regular checking and readjustment. Two different temperatures were studied, 25 and 35 ºC. For DB adsorption, an additional isotherm was determined using the synthetic effluent (SDE, prepared as reported in 2.1.), instead of the pure dye aqueous solution (dye plus distilled water). The aim was to evaluate the equilibrium relation in conditions closer to real ones.

2.5. Adsorption kinetic studies Adsorption kinetics of BR and DB dyes were studied in batch mode, at constant temperature (T=25ºC), initial pH 7 (adjusted with HCl or NaOH solutions) and different initial dye concentrations: 15-100 mg/L for BR, and 50-150 mg/L for DB. The sepiolite dosages were 0.25 g/L (for BR) or 1.0 g/L (DB). Experiments were performed using an acrylic vessel, containing 500 mL of dye solution and the known quantity of sepiolite, under a stirring rate of 400 rpm, inside a thermostatic cabinet. Samples (approx. 2 mL) were withdrawn at specific time intervals and the dye concentration in the liquid phase (C, mg/g) analyzed as a function of contact time (t).

2.6. Desorption studies In order to study the regeneration of the adsorbent and evaluate the mechanism involved in BR adsorption, some desorption experiments were performed, using NaOH and HCl solutions as eluents. Sepiolite clay, loaded with BR dye, was shaken (4.0 g/L) with solutions at different pH (range 2-12). After 48 h of contact time, the dye concentration in the liquid phase was analyzed and the percentage of the dye desorbed from the sepiolite clay was calculated.

8

3. Results and discussion 3.1. Sepiolite characterization Physical, textural, chemical and mineralogical properties of sepiolite are summarized in Table 1. Mineralogical analysis indicate 78% of sepiolite (close to the 80% value, given by the supplier), 9% of K-feldspar, 5% of quartz, 4% of mica and 4% of calcite (weight percentages) [23]. The mean particle size (equivalent spherical diameter) was determined as 0.576 mm. Results obtained by mercury porosimetry indicated a total porosity of 46.5% (Table 1). The curve of incremental mercury intrusion vs. pore diameter (figure not shown) presented a unimodal and relatively symmetric pore size distribution, with mean of 20.1 nm and median of 29.9 nm (mesopores). The intrusion of mercury in the pores of a material is however limited to a size of approx. 4 nm. The structural channels of sepiolite, with a rectangular shape of approx. 0.4 nm x 1.1 nm [18], and occupied by water molecules, contribute to the microporosity, but are not included in the mercury intrusion data. Nitrogen adsorption, able to quantify microporosity, indicated a micropore volume of 13 mm3/g and a surface area of 108 m2/g. This value was 42% higher than the value based on mercury intrusion (76 m2/g), which means that in addition to the mesopores, smaller mesopores (2-4 nm) and micropores (< 2 nm) are also significant as regards the porous structure of sepiolite. The results obtained in the MB method are illustrated in Fig. 2. Surface area was estimated (Eq. 1) using the complete replacement point (“end-point”) as 175 m2/g. This value is in general agreement with the surface area measured by nitrogen adsorption (108 m2/g), confirming the non-swelling nature of the clay. In montmorillonitic clays (swelling minerals) wet methods render much higher (18-45 times greater) surfaces areas than dry methods [27]. Fig. 3 illustrates the results obtained by the pH-drift method, at three different ionic strengths. This method has been applied to different materials [15, 28] to estimate the pH of zero charge (PZC) and the ranges where the surface charge is positive (below PZC) and negative (above PZC). When a common intersection point is obtained for the curves resulting from different electrolyte concentrations, the intersection point is not dependent on the ionic concentration and it can be 9

designated as a zero point of charge. The results depicted in Fig. 3, show however three distinct intersection points for different ionic concentrations: 10.4, 10.0 and 9.8, respectively for NaCl 0.001 M, 0.01M and 0.1M. This observation is due to the dual nature of surface charge in minerals, where two different surface charges are present: a permanent negative charge (resulting from the structural isomorphous substitutions) and a pH-dependent charge (primarily related to the dissociation of hydroxyl groups). This means that the intersection points here obtained, corresponding to zero net adsorption of H+/OH-should be named as PZNPC, instead of PZC, since they represent the conditions where the pH-dependent charge is null [29] and not the conditions under which the total charge is null. The high PZNPC values (Table 1) allow to predict a higher propensity of sepiolite to adsorb the anionic dye in pH conditions up to 10.

3.2. Adsorption equilibrium 3.2.1. Effect of pH The pH is usually an important parameter affecting the adsorption extent. Its effect on BR and DB uptake by sepiolite was evaluated performing tests at different initial pH. The results are presented in Fig. 4.

The results show that under initial strong acidic conditions (pH 2), the sepiolite suspensions keep the pH constant throughout the contact time. The same behavior was observed for initial pHs 9-10, with final pH values very close to the initial ones. If neutral or moderate acidic/alkaline conditions were used, the pH of the suspension quickly changes and equilibrium pH values around 8-9 were obtained. This strong buffering capacity was also reported in other works [30]. Regarding BR adsorption, a general tendency to be favored by higher initial and final pH values was observed. A slight increase (27%) in the adsorbed amount was verified from initial pH 2 to initial pH 7. The most visible increase occurred from initial pHs 7-8 to pHs 9-10, under which optimum removals were achieved (approx. 180 mg/g). 10

Regarding DB dye, the adsorbed amount clearly decreased with the increase in the initial pH (from 4 to 10) and with the equilibrium pH (from 8 to 10). Minimum (24 mg/g) and maximum adsorbed amounts (140 mg/g) were obtained for initial pH 4 and 10, respectively. The different pH effects here observed for BR (cationic) and DB (anionic) have been also reported for other cationic [20] and anionic species [20, 25, 31], and are explained by the electrostatic attraction. With the increase in pH, the dependent charge tends to be more negative and hence amenable to adsorb cationic species (BR dye) and less to adsorb anionic ones (DB). The PZNPC of sepiolite occurs at pH 10 (Table 1), and then a significant increase in BR dye adsorption was observed from an equilibrium pH 8.5-8.7 to 9.2-9.5; in turn, a sharp decrease in the DB uptake was observed from final pH 9.3 to 9.8 (Fig. 4). It is important to mention that BR dye is removed extensively by sepiolite, even under unfavorable pH conditions, showing that permanent charge is mostly involved. In the case of DB dye, the effect of unfavorable pH conditions (pH 10) was more evident and this means that the electrostatic attraction assumes an important role.

3.2.2. Adsorption equilibrium isotherms Dye adsorption equilibrium isotherms in aqueous solution were determined at pH approx. 9 (8.7±0.5) and temperatures of 25 and 35 ºC. The results are presented in Fig. 5. As it can be seen, very considerable adsorbed amounts of both dyes were obtained. Langmuir [32] and Freundlich [33] models were fitted to experimental data, by non-linear regression. In Langmuir equation (Eq. 3), Qm denotes the maximum adsorption capacity, corresponding to a monolayer coverage, and KL is a constant related to the energy of adsorption. In Freundlich equation (Eq. 4), KF is a constant related to the adsorption capacity and n a constant related to adsorption intensity. The parameters obtained for both models are shown in Table 2, and the model curves illustrated in Fig. 5. qe q 

Qm K L Ceq

(3)

1  K L Ceq

11

qe q  K F Ceq

1/ n

(4)

Determination coefficients (R2) indicate a good correlation between experimental data and both models, especially the Langmuir model. The normalized standard deviations, denoted as Δqeq (%), were also calculated (Eq. 5) [34], in order to compare the accuracy of the predictions.

qeq  %   100 

  q

eq

 qeq pred  / qeq  N

2

(5)

In Eq. 5, N represents the number of experimental points, qeq the experimental values of equilibrium adsorbed amounts and qeq pred the corresponding Langmuir or Freundlich predicted values. For BR, the normalized standard deviations of Freundlich fittings (7.5 and 10%) were better than those generated from Langmuir model (8.2 and 20%). In the case of DB, slightly lower deviations were found for Langmuir. In the conditions studied, Langmuir model predicted a maximum adsorption capacity of 110 mg/g (25 ºC) for BR dye. At this temperature, the isotherm is almost defined by a constant q eq value, with a plateau observed for equilibrium concentrations higher than 25 mg/L. Sepiolite showed a great affinity for BR dye, which is visible from the initial portion of the isotherm (high adsorbed amounts observed at low concentrations). Considering the molar mass of BR dye, the maximum adsorption amount obtained at 25 ºC (110±6 mg/g) is equivalent to 0.27±0.02 mmol/g, which is in perfect agreement with the measured CEC of the sepiolite (CEC=0.27±0.01 mmol/g, Table 1). Cation exchange is then confirmed as the main mechanism for BR removal. The relation between the amount of basic dyes adsorbed and CEC of clay materials is variable, according to the literature. Some works reported a much higher adsorbed amount than the CEC [22], suggesting that most of the adsorption occurred in neutral sites of the clay; other works reported an adsorbed amount matching CEC for clay materials [35, 36], as also observed in the present work. The increase in temperature, from 25 to 35

12

ºC, clearly favored the BR adsorption. The adsorbed amounts increased in almost the entire concentration range (Fig. 5) and the maximum adsorbed amount, predicted by Langmuir modeling, was 310 mg/g at 35 ºC. Regarding DB adsorption, sepiolite also presented a good performance, achieving very considerable adsorbed amounts in the entire range. The temperature rise from 25 to 35 ºC showed no significant effect in the equilibrium relation, as can be ascertained by the similarity of the model parameters for both conditions (Table 2). For DB, Langmuir equation fitted quite well the experimental data, but is important to mention that the monolayer coverage (hypothesized in the model) was not attained in the Ceq range studied (which was selected to cover the values of practical interest). The maximum adsorbed amount predicted (232 mg/g at 25 ºC) is then slightly higher than the experimental observed value (185 mg/g). Although higher adsorbed amounts have been reached for DB than for BR, the initial slope of the isotherm is not as high as it is for BR, which means a lower sepiolite affinity for DB. As suggested by the results obtained in section 3.2.1, the electrostatic interaction that can occur between protonated SiOH groups (at low pH) and the anionic DB dye, plays an important role in the adsorption mechanism. However, other type of interactions can be hypothesized between the active sites of sepiolite and hydroxyl, amino and sulphonate groups of the DB dye (Fig. 1). Amines can interact with (i) zeolitic or bound water, via hydrogen bonds, (ii) octahedral magnesium exposed to the channel, and (iii) siloxane groups at the surface of the channels, via hydrogen bonds [19]. Tabak et al. (2009) [31] concluded that the interaction between a reactive dye and sepiolite was through the dye N-H group and via hydrogen bond with bound water at channels sites, indicating that dye species replaced partly the zeolitic water. In addition, sulphonate groups of the DB dye can interact with sepiolite, by complexation with Mg ion at the octahedral sheet and hydrogen bonding between its Oand H+ of the bound or zeolitic water, as proposed for anionic surfactants removal [30]. Although seeming to exist these interaction possibilities, the accessibility to the sorption sites located in the tunnels (0.4 nm x 1.1 nm) depends essentially on the size of the adsorbate. The molecular dimensions 13

of DB dye are not accurately known, but distances between peripheral atoms reaching 2 nm are likely, based on literature estimates for other direct dyes [37]. Clearly, DB penetration in the narrow sepiolite tunnels is not possible. Eventual interactions via hydrogen bonding with water molecules or via octahedral magnesium are possible, but limited to the entrance of the tunnels (ends of the fibres) and imply blocking of tunnels access by the adsorbed dyes. Therefore, DB interaction with sepiolite clay occurs mostly in the external surface. Considering that silanol groups are located in sepiolite surface under a separation of 0.5 nm [38], it becomes also clear the good degree of accommodation that DB molecules have to acquire in order to maximize the number of interactions.

3.2.3. Adsorption equilibrium in synthetic effluents In the previous section, the equilibrium results obtained in the adsorption of basic and direct dyes in pure dye aqueous solutions were presented. In the case of DB adsorption, sepiolite performance was also evaluated using a synthetic textile effluent, SDE, prepared as explained in section 2.1. The aim was to evaluate the effect of the presence of auxiliary dyeing chemicals and salts (NaCl) on DB dye removal. In order to evaluate the process near the real worst-conditions, it was hypothesized that the whole amounts of auxiliary dyeing chemicals and salt (used in the dyeing bath) are washed out and appear in the final effluent (100% lost to the water). The adsorption equilibrium isotherm obtained for SDE is presented in Fig. 5(b). In the concentration range studied, this isotherm shows decreases in the adsorbed amounts between approx. 40 and 80%, in comparison to the values observed in “pure” dye solutions. Even with this limitation, very significant removals were still found, with experimental adsorbed amounts reaching 106 mg/g. Langmuir and Freundlich parameters for SDE isotherm are listed in Table 2. Both models described similarly and quite well the experimental data. The maximum predicted adsorbed amount (Langmuir model) was 201 mg/g, which is only 13% lower than the Qm value obtained in pure dye aqueous solution (232 mg/g). Considering that the amount of NaCl used in SDE corresponds to a final concentration of 0.04 mol/L (low value) and the results previously obtained with DB dye adsorption on a waste material [6], which 14

showed no significant effect of this salt, it was hypothesized that the auxiliary dyeing chemicals were the responsible for the decrease in sepiolite performance. In fact, the sequestering and the wetting agents, used as dyeing auxiliary chemicals, have anionic nature and can compete to the same active sites of sepiolite. In order to confirm this, an aqueous solution containing these compounds was prepared (following the same procedure for SDE, but with no addition of DB). This solution was analyzed and presented a total organic carbon (TOC) content of 98.5 mg/L. It was then put in contact with different sepiolite dosages (0.2-3.0 g/L) and final values of TOC determined again. Results have confirmed the capacity of sepiolite to adsorb these chemicals: final TOC values in the liquid phase were in the range 68.2-91.9 mg/L (higher TOC values were found for lower sepiolite dosages), which correspond to removals between 7% and 31% (Fig. 6). TOC removals, expressed as mg of organic carbon adsorbed per g of sepiolite, were also calculated. Values were found to be in the range 29-317 mg/g (Fig. 6). These results confirm the competitive behavior between some auxiliary chemicals usually used in the dyeing process and the dye itself.

3.2.4. Comparison of sepiolite with other adsorbents The maximum adsorption capacity (Qm) is usually the key parameter used to evaluate and compare the performance of different adsorbents. This kind of comparison should consider the type of dye used as adsorbate, the pH and temperature conditions (in the interest ranges). Table 3 summarizes some Qm values reported in literature for the adsorption of basic and direct dyes. In comparison to commercial and lab-synthetized activated carbons, sepiolite performance for basic dyes removal can be considered similar or, in some circumstances, up to 5 times lower. For direct dyes adsorption, sepiolite clearly presented higher adsorptive performance (at least 2.5 times greater), even comparing the Qm obtained at pH 9 with Qm values obtained under more favorable pH conditions (see Table 3). Depending on the properties, activated carbons can be acquired at a price ranging from US$ 700 to US$ 5000 per ton [50]. Sepiolite price per ton ranges from US$ 90 to as much as US$ 800, for very fine highly refined material [51]. As it can be inferred, the cost-effective ratio is more 15

favorable for sepiolite, which also presents the advantage of being highly effective for both basic and anionic dyes. Some waste materials, such as coffee husks, fruit seeds and industrial residues (Table 3), also presented interesting adsorption capacities; considering they are available at almost no cost, their use can present advantageous cost-effective ratios. In these cases, and in a practical application, other important criterions (kinetics, leachability, mechanical strength, and toxicity) should be considered. The cost-effective ratio can also be evaluated for other materials, such as chitosan/Fe composites [16, 46], which presented outstanding adsorption capacities. Apart from this, besides the direct economic benefit, sepiolite use as adsorbent provides several other environmental advantages: it does not need any chemical treatment in order to be used as adsorbent; it is available at different grain sizes, allowing to choose between finer and coarser particles, depending on the practical application (stirred adsorbers or fixed-bed columns); no risks related to leaching toxic compounds; and it provides a direct pH adjustment to values 8-9, usually allowable for discharge, and under which adsorption is feasible. This direct pH adjustment avoids additional expenses in acids or bases.

3.2.5. Thermodynamic Parameters Equilibrium data obtained at different temperatures (25 and 35 ºC) were used to estimate the thermodynamic parameters of adsorption. The change in free energy (∆Gº) was calculated by Eq. 6, where R is the gas constant, T the temperature and Kc the equilibrium constant, considered as the Langmuir constant KL, expressed in mol/L [52]. The change in entropy (∆Sº) and change in enthalpy (∆Hº) were obtained from Eq. 7 (Van´t Hoff equation). The thermodynamic parameters obtained are presented in Table 4. Gº   R  T  Ln K c (6)

Ln K c 

S º H º (7)  R RT

The negative values of ∆Gº indicated feasibility of adsorption process for both dyes. For BR, a more negative enthalpy variation was found, which indicated the exothermic nature of the adsorption. This 16

is due to the higher affinity observed at 25ºC, in comparison with 35ºC. The ∆Sº and ∆Hº values obtained in the present work are similar to the ones obtained by Eren and Afsin [53] for the adsorption of crystal violet on bentonite (∆Hº=-83,81 kJ/mol and ∆Sº =-0,26 kJ/mol/K). For DB dye adsorption, ∆Hº is much less negative, which means the low influence of temperature on the adsorption equilibrium. In this case, the calculated value for ∆Sº was slightly positive (0.03 kJ mol -1 K-1), but with a too high uncertainty (0.1 kJ mol-1 K-1), so it was considered not statistically significant.

3.3. Adsorption kinetic studies 3.3.1. Experimental results The effect of contact time in BR and DB adsorption on sepiolite clay, was studied for different initial dye concentrations. Results are presented in Fig. 7.

The uptake of BR dye (Fig. 7a) occurred rapidly during the initial stage. In the first hour of contact time, the decay in the BR concentration reached more than 70 % of the maximum decay observed in the equilibrium. After 5-8 hours, the equilibrium was considered achieved, with adsorbed amounts between 50 mg/g (for C0=15 mg/L) and 90 mg/g (C0=100 mg/L). The adsorption of DB dye on sepiolite clay is presented in Fig. 7b. Equilibrium was achieved after a contact time of 1 day (50 mg/L) or 2 days (100 and 150 mg/L). 80% of the maximum removal was attained after 4 h (C0=50 mg/L) or 20 h (C0=100 or C0=150 mg/L). Direct dyes usually have big molecular structures (larger size than basic dyes, see Fig. 1). The high molecular weight and dimension cause difficulties to intraparticle diffusion, and hence higher contact times are necessary to attain steady-state and slower kinetics is observed. In the conditions tested, the dye removals attained in equilibrium were 91-93%, in respect to initial concentrations.

17

3.2.2. Pseudo-first and pseudo-second-order modeling Lagergren's pseudo-first order [54] and pseudo-second order models [55] are represented by Eq. 8 and Eq. 9, respectively, and were used in the present work to model the experimental data (by nonlinear regression, using Fig.P software from Biosoft).

q  qeq  1  exp  k1  t 

q  qeq 

k 2  qe  t 1  k 2  qe  t

(8) (9)

In these expressions, q and qeq symbolize the dye adsorbed amount per mass unit of adsorbent, at time t and at equilibrium, respectively, and k1 and k2 are the kinetic constants. The kinetic parameters obtained are presented in Table 5. Normalized standard deviation was also calculated (Eq. 5) in order to compare both models. Modeled curves are shown in Fig. 7. High determination coefficients (R2≥0.94) were obtained for both models and dyes. For BR adsorption, pseudo-second order model fittings presented higher R2 values and lower normalized standard deviations, describing better the experimental data (particularly in the initial and final adsorption stages, see Fig. 7). The kinetic constant (k2) did not show a significant variation with BR concentration between 15 and 50 mg/L, but increased when the initial concentration raised to 100 mg/L, justified by the higher driving force to the adsorption. For DB adsorption, both models presented very high correlation coefficients (R2≥0.98). The pseudofirst order model predicted equilibrium adsorbed amounts with higher accuracy (see Fig. 7). However, the normalized standard deviations (Table 5) for pseudo-second order model are slightly better. Kinetic constants tended to decrease with the increase in the initial DB concentration, which is opposite to the behavior usually observed. However, if we consider the space limitations of DB molecules to assess the active sites in sepiolite (as discussed, section 3.2.2.), and the need to acquire certain degree of accommodation, it is expected that when in higher concentrations, a higher resistance is observed and lower kinetic constants are obtained.

18

3.2.3. Linear Driving Force approximation Linear Driving Force (LDF) [56], originally proposed for adsorption chromatography, assumes that the uptake of the adsorbate is linearly proportional to a driving force (Eq. 10), defined as the difference between the adsorbed amount at the surface (q*) and the average adsorbed amount in the adsorbent (q).

q  k LDF q *  q  t

(10)

Considering q*=Qm given by the Langmuir model (Eq. 3), the dimensionless variables ξ and y and the additional variables a, b, α and β, given by Eq. 11 and 12, respectively, Eq. 10 can be integrated and an explicit equation, in order to the contact time (Eq. 13), is obtained (other details in [15]):



Qm ; y C m/ v Cin Cin a   1

t

(11)

1 ; 1  a  a 2  4b ;  a  a 2  4b ;   b  2 2 K L  Cin K L  Cin

 y2  a  y  b  1  1 a  1   1    y       ln  ln       1  k LDF  2  b  a  b  1   2  b       1    y    

(12)

(13)

Solver tool in Excel was used to obtain kLDF constants from Eq. 13 by minimizing the sum of squared residuals. The values are presented in Table 6. Coefficients of determination were calculated to evaluate how well experimental data fit linear driving force model. In well stirred batch systems, external resistance is usually negligible and intraparticle diffusion is usually the rate controlling step. In these circumstances, for a spherical porous particle, two different mechanisms contribute to the overall intraparticle diffusion: (i) diffusion within the pore volume and (ii) diffusion along the surface of pores. Two simple models can then be derived: Homogeneous solid diffusion model (HSDM), characterized by a homogeneous diffusion coefficient (Dh), and diffusion into the pores, characterized by the effective pore diffusivity (DPe). Considering a parabolic profile of dye concentration as a function of radial position inside the particle, the relations between the LDF 19

kinetic constant and the diffusivity coefficients are respectively given by Eq. 14 and Eq. 15, where Rp is the average sepiolite particle radius, Ω is a constant related to the geometric shape of particles (15 for spherical particles) and dq/dC is the mean slope of the equilibrium isotherm. k LDF 

k LDF 

15  Dh Rp

2

(14) 15  D pe

R p ( p   ap  dq * / dC ) 2

(15)

From the LDF kinetic constants, the effective pore diffusivity and homogeneous solid diffusivity coefficients were estimated and presented in Table 6. As expected, Dh values for DB are lower than for BR, due to the higher molecular size. For BR dye adsorption, the variations of Dh and DPe with the initial concentration are clear. The homogeneous diffusion coefficient increased linearly with C0, as a result of a higher surface coverage, and as observed by other researchers [15, 57]. For the DB dye, the variations observed are only slightly significant, seeming to be preferable to consider an average Dh (4×10-13 m2/s) and DPe (7×10-10 m2/s) for the conditions studied. As previously referred, opposite aspects influencing the adsorption kinetics result from increasing the DB initial concentration: a higher driving force and coverage, which would contribute to a higher Dh, but a probable greater difficulty to accommodate these dyes and a higher intraparticle resistance.

3.3. Desorption studies Regeneration of a spent adsorbent (e.g.: activated carbon) is usually carried out in order to reuse it, avoiding the cost of a new acquisition and minimizing the amount of waste. When low-cost adsorbents are used, regeneration is not usually economically justifiable. Desorption studies were however conducted for BR-loaded sepiolite, as a function of pH, and results are shown in Fig. 8. As it can be seen from Fig. 8, the maximum BR desorbed percentage was 9.4% and it was obtained for strong acidic conditions (pH 2). For higher pH values, desorption was very low and limited to 23%, which reflects the establishment of strong chemical bonds between the basic dye and the sepiolite

20

in the adsorption. The results showed no feasible regeneration by using acid or alkaline solutions. The reduced color leachability from the dye-loaded sepiolite is, on the other side, a good feature considering the disposal in landfill or reuse for other purposes, such as incorporation in construction materials or in polymeric composites [58].

4. Conclusions A sepiolite clay showed to be a good adsorbent for two azo dyes, from different types: Basic Red 46 and Direct Blue 85 dyes. The obtained maximum adsorption capacities, at 25 ºC and pH 9, were 110 and 232 mg/g, respectively. The amount of basic red dye adsorbed matched the cation exchange capacity of the clay (0.27 mmol/g). In the case of basic red adsorption, the increase in temperature and pH favored the adsorption extent (from 25 to 35 ºC). Regarding the direct blue dye, the adsorption was not influenced by temperature (15-35 ºC), but a significant effect of pH was observed, with a sharp decrease in the uptake capacity for strong alkaline conditions. A synthetic wastewater, simulating the effluent obtained in cotton dyeing with direct dyes was prepared. Auxiliary dyeing chemicals were found to compete with direct blue dye to the sepiolite adsorption sites. Adsorption kinetics was studied and modeled successfully by pseudo-first (for direct blue) and pseudo-second order (basic red) rate equations. Homogeneous solid and pore effective diffusivities of the basic and direct dyes were estimated using the LDF approximation model. The results obtained in the present work indicate good perspectives for the use of sepiolite clay as adsorbent for textile colored wastewaters.

Acknowledgments This

work

was

financed

by

FCT

and

FEDER

under

Programe

PT2020

(Project

UID/EQU/50020/2013). S. Santos acknowledges her postdoctoral scholarship financed by Project UID/EQU/50020/2013.

21

References [1] Citeve, Estudo das dificuldades das empresas do setor têxtil e vestuário no cumprimento da legislação ambiental, in: Centro Tecnológico das Indústrias Têxtil e do Vestuário de Portugal, 2012. [2] B.D. Waters, The regulator's view, in: P. Cooper (Ed.) Colour in Dyehouse Effluent, The Society of dyers and Colourists, The Alden Press, Oxford, 1995, pp. 22-29. [3] K.T. Chung, C.E. Cerniglia, Mutagenicity of Azo Dyes - Structure-Activity-Relationships, Mutat Res, 277 (1992) 201-220. [4] A. Garg, K.L. Bhat, C.W. Bock, Mutagenicity of aminoazobenzene dyes and related structures: a QSAR/QPAR investigation, Dyes Pigments, 55 (2002) 35-52. [5] B.G. Hazel, Industry evaluation of colour reduction and removal - the DEMOS project. , in: P. Cooper (Ed.) Colour in Dyehouse Effluent, The Society of dyers and Colourists, The Alden Press, Oxford, 1995, pp. 59-72. [6] S.C.R. Santos, R.A.R. Boaventura, Treatment of a simulated textile wastewater in a sequencing batch reactor (SBR) with addition of a low-cost adsorbent, J Hazard Mater, 291 (2015) 74-82. [7] S.V. Mohan, P.S. Babu, S. Srikanth, Azo dye remediation in periodic discontinuous batch mode operation: Evaluation of metabolic shifts of the biocatalyst under aerobic, anaerobic and anoxic conditions, Sep Purif Technol, 118 (2013) 196-208. [8] A.K. Verma, R.R. Dash, P. Bhunia, A review on chemical coagulation/flocculation technologies for removal of colour from textile wastewaters, J Environ Manage, 93 (2012) 154-168. [9] G.M. Ayoub, A. Hamzeh, L. Semerjian, Post treatment of tannery wastewater using lime/bittern coagulation and activated carbon adsorption, Desalination, 273 (2011) 359-365. [10] S. Papadia, G. Rovero, F. Fava, D. Di Gioia, Comparison of different pilot scale bioreactors for the treatment of a real wastewater from the textile industry, Int Biodeter Biodegr, 65 (2011) 396-403. [11] S. Banerjee, G.C. Sharma, M.C. Chattopadhyaya, Y.C. Sharma, Kinetic and equilibrium modeling for the adsorptive removal of methylene blue from aqueous solutions on of activated fly ash (AFSH), Journal of Environmental Chemical Engineering, 2 (2014) 1870-1880. [12] N.F. Cardoso, E.C. Lima, I.S. Pinto, C.V. Amavisca, B. Royer, R.B. Pinto, W.S. Alencar, S.F.P. Pereira, Application of cupuassu shell as biosorbent for the removal of textile dyes from aqueous solution, J Environ Manage, 92 (2011) 1237-1247. [13] F.A. Pavan, E.S. Camacho, E.C. Lima, G.L. Dotto, V.T.A. Branco, S.L.P. Dias, Formosa papaya seed powder (FPSP): Preparation, characterization and application as an alternative adsorbent for the removal of crystal violet from aqueous phase, Journal of Environmental Chemical Engineering, 2 (2014) 230-238. [14] M. Sadeghi-Kiakhani, M. Arami, K. Gharanjig, Preparation of chitosan-ethyl acrylate as a biopolymer adsorbent for basic dyes removal from colored solutions, Journal of Environmental Chemical Engineering, 1 (2013) 406-415. [15] S.C.R. Santos, V.J.P. Vilar, R.A.R. Boaventura, Waste metal hydroxide sludge as adsorbent for a reactive dye, J Hazard Mater, 153 (2008) 999-1008. [16] L.C. Zheng, C.G. Wang, Y.H. Shu, X.M. Yan, L.S. Li, Utilization of diatomite/chitosan-Fe (III) composite for the removal of anionic azo dyes from wastewater: Equilibrium, kinetics and thermodynamics, Colloid Surface A, 468 (2015) 129-139. [17] O. Ozdemir, B. Armagan, M. Turan, M.S. Celik, Comparison of the adsorption characteristics of azoreactive dyes on mezoporous minerals, Dyes Pigments, 62 (2004) 49-60. [18] E. Ruiz-Hitzky, Molecular access to intracrystalline tunnels of sepiolite, J Mater Chem, 11 (2001) 86-91. [19] S. Yariv, Thermo-IR-spectroscopy analysis of the interactions between organic pollutants and clay minerals, Thermochim Acta, 274 (1996) 1-35.

22

[20] Z.X. Han, Z. Zhu, D.D. Wu, J. Wu, Y.R. Liu, Adsorption Kinetics and Thermodynamics of Acid Blue 25 and Methylene Blue Dye Solutions on Natural Sepiolite, Synth React Inorg M, 44 (2014) 140-147. [21] S. Lazarevic, I. Jankovic-Castvan, D. Jovanovic, S. Milonjic, D. Janackovic, R. Petrovic, Adsorption of Pb2+, Cd2+ and Sr2+ ions onto natural and acid-activated sepiolites, Appl Clay Sci, 37 (2007) 47-57. [22] G. Rytwo, S. Nir, L. Margulies, B. Casal, J. Merino, E. Ruiz-Hitzky, J.M. Serratosa, Adsorption of monovalent organic cations on sepiolite: Experimental results and model calculations, Clay Clay Miner, 46 (1998) 340-348. [23] S.C.R. Santos, R.A.R. Boaventura, Adsorption modelling of textile dyes by sepiolite, Appl Clay Sci, 42 (2008) 137-145. [24] C. Bilgic, Investigation of the factors affecting organic cation adsorption on some silicate minerals, J Colloid Interf Sci, 281 (2005) 33-38. [25] F. Tumsek, O. Avci, Investigation of Kinetics and Isotherm Models for the Acid Orange 95 Adsorption from Aqueous Solution onto Natural Minerals, J Chem Eng Data, 58 (2013) 551-559. [26] J.C. Santamarina, K.A. Klein, Y.H. Wang, E. Prencke, Specific surface: determination and relevance, Can Geotech J, 39 (2002) 233-241. [27] A.P. Magnoli, L. Tallone, C.A.R. Rosa, A.M. Dalcero, S.M. Chlacchiera, R.M.T. Sanchez, Commercial bentonites as detoxifier of broiler feed contaminated with aflatoxin, Appl Clay Sci, 40 (2008) 63-71. [28] V. Marjanovic, S. Lazarevic, I. Jankovic-Castvan, B. Jokic, D. Janackovic, R. Petrovic, Adsorption of chromium (VI) from aqueous solutions onto amine-functionalized natural and acid-activated sepiolites, Appl Clay Sci, 80-81 (2013) 202-210. [29] M.J. Avena, C.P. De Pauli, Proton adsorption and electrokinetics of an Argentinean montmorillonite, J Colloid Interf Sci, 202 (1998) 195-204. [30] O. Ozdemir, M. Cinar, E. Sabah, F. Arslan, M.S. Cefik, Adsorption of anionic surfactants onto sepiolite, J Hazard Mater, 147 (2007) 625-632. [31] A. Tabak, E. Eren, B. Afsin, B. Caglar, Determination of adsorptive properties of a Turkish Sepiolite for removal of Reactive Blue 15 anionic dye from aqueous solutions, J Hazard Mater, 161 (2009) 1087-1094. [32] I. Langmuir, The Adsorption of Gases on Plane Surfaces of Glass, Mica and Platinum, J Am Chem Soc, 40 (1918) 1361-1403. [33] H.M.F. Freundlich, Over the adsorption in solution, Journal of Physical Chemistry, 57 (1906) 385-471. [34] R.K. Singh, S. Kumar, S. Kumar, A. Kumar, Development of parthenium based activated carbon and its utilization for adsorptive removal of p-cresol from aqueous solution, J Hazard Mater, 155 (2008) 523-535. [35] Y. El Mouzdahir, A. Elmchaouri, R. Mahboub, A. Gil, S.A. Korili, Adsorption of methylene blue from aqueous solutions on a Moroccan clay, J Chem Eng Data, 52 (2007) 1621-1625. [36] Z.H. Li, P.H. Chang, W.T. Jiang, J.S. Jean, H.L. Hong, Mechanism of methylene blue removal from water by swelling clays, Chem Eng J, 168 (2011) 1193-1200. [37] S.A. Figueiredo, J.M. Loureiro, R.A. Boaventura, Natural waste materials containing chitin as adsorbents for textile dyestuffs: Batch and continuous studies, Water Res, 39 (2005) 4142-4152. [38] J.M. Serratosa, Surface properties of fibrous clay minerals (playgorskite and sepiolite), in: M.M. Mortland, V.C. Farmer (Eds.) Developments in Sedimentology, Elsevier, 1979, Volume 27, Developments in Sedimentology, , Pages 99-109, 1979, pp. 99-109. [39] E.N. El Qada, S.J. Allen, G.A. Walker, Adsorption of basic dyes from aqueous solution onto activated carbons, Chem Eng J, 135 (2008) 174-184. [40] M.J. Martin, A. Artola, M.D. Balaguer, M. Rigola, Activated carbons developed from surplus sewage sludge for the removal of dyes from dilute aqueous solutions, Chem Eng J, 94 (2003) 231-239.

23

[41] E. Demirbas, M. Kobya, M.T. Sulak, Adsorption kinetics of a basic dye from aqueous solutions onto apricot stone activated carbon, Bioresource Technol, 99 (2008) 5368-5373. [42] B. Acevedo, R.P. Rocha, M.F.R. Pereira, J.L. Figueiredo, C. Barriocanal, Adsorption of dyes by ACs prepared from waste tyre reinforcing fibre. Effect of texture, surface chemistry and pH, J Colloid Interf Sci, 459 (2015) 189-198. [43] M. Turabik, Adsorption of basic dyes from single and binary component systems onto bentonite: Simultaneous analysis of Basic Red 46 and Basic Yellow 28 by first order derivative spectrophotometric analysis method, J Hazard Mater, 158 (2008) 52-64. [44] O. Duman, S. Tunç, T.G. Polat, Determination of adsorptive properties of expanded vermiculite for the removal of C. I. Basic Red 9 from aqueous solution: Kinetic, isotherm and thermodynamic studies, Appl Clay Sci, 109-110 (2015) 22-23. [45] L.S. Oliveira, A.S. Franca, T.M. Alves, S.D.F. Rocha, Evaluation of untreated coffee husks as potential biosorbents for treatment of dye contaminated waters, J Hazard Mater, 155 (2008) 507-512. [46] S. Saber-Samandari, S. Saber-Samandari, N. Nezafati, K. Yahya, Efficient removal of lead (II) ions and methylene blue from aqueous solution using chitosan/Fe-hydroxyapatite nanocomposite beads, J Environ Manage, 146 (2014) 481-490. [47] L.D.T. Prola, F.M. Machado, C.P. Bergmann, F.E. de Souza, C.R. Gally, E.C. Lima, M.A. Adebayo, S.L.P. Dias, T. Calvete, Adsorption of Direct Blue 53 dye from aqueous solutions by multi-walled carbon nanotubes and activated carbon, J Environ Manage, 130 (2013) 166-175. [48] A. Khaled, A. El Nemr, A. EI-Sikaily, A. Abdelwahab, Treatment of artificial textile dye effluent containing Direct Yellow 12 by orange peel carbon, Desalination, 238 (2009) 210-232. [49] L.L. Lian, L.P. Guo, C.J. Guo, Adsorption of Congo red from aqueous solutions onto Ca-bentonite, J Hazard Mater, 161 (2009) 126-131. [50] H. Saygili, F. Guzel, Y. Onal, Conversion of grape industrial processing waste to activated carbon sorbent and its performance in cationic and anionic dyes adsorption, J Clean Prod, 93 (2015) 84-93. [51] H. Murray, Industrial Clays Case Study, Mining, Minerals and Sustainable Development, no. 64, in: I.I.f.E.a.D. (IIED) (Ed.), 2002, pp. 1-9. [52] Y. Liu, Some consideration on the Langmuir isotherm equation, Colloid Surface A, 274 (2006) 34-36. [53] E. Eren, B. Afsin, Investigation of a basic dye adsorption from aqueous solution onto raw and pre-treated bentonite surfaces, Dyes Pigments, 76 (2008) 220-225. [54] S.Y. Lagergren, Zur theorie der sogenannten adsorption gelöster stoffe, Kungliga Svenska Vetenskapsakademiens, Handlingar, Band, 24 (1898) 1-39. [55] G. Blanchard, M. Maunaye, G. Martin, Removal of Heavy-Metals from Waters by Means of Natural Zeolites, Water Res, 18 (1984) 1501-1507. [56] E. Glueckauf, Theory of chromatography. Part 10. Formulae for diffusion into spheres and their application to chromatography, Tans. Faraday Soc., 51 (1955) 1540-1554. [57] V. Meshko, L. Markovska, M. Mincheva, A.E. Rodrigues, Adsorption of basic dyes on granular acivated carbon and natural zeolite, Water Res, 35 (2001) 3357-3366. [58] K.Q. Zhou, Q.J. Zhang, B. Wang, J.J. Liu, P.Y. Wen, Z. Gui, Y. Hu, The integrated utilization of typical clays in removal of organic dyes and polymer nanocomposites, J Clean Prod, 81 (2014) 281-289.

24

Fig. 1. Molecular structures of (a) Basic Red 46 and (b) Direct Blue 85 dyes.

25

MB adsorbed (mg/g)

100

75

← "End-Point" 50

25

0 0

50

100

150

200

MB added (mg/g)

Fig. 2. Results obtained for MB adsorption by sepiolite.

26

13.0

Final pH

10.0

7.0 0,001 M NaCl PZNPC=10.4 0,01 M NaCl PZNPC=10.0

4.0

0,1 M NaCl PZNPC=9.8

1.0 1.0

4.0

7.0 10.0 Initial pH

13.0

Fig. 3. Determination of PZNPC of sepiolite, for different ionic strengths, by pH-drift method.

27

200 9.2 9.5 8.4

qeq (mg/g)

150

8.7 8.5

2.0

100

8.1

50

9.3

BR 9.8

DB 0 1.0

4.0

7.0

10.0

initial pH

Fig. 4. Effect of pH on the amount of dye adsorbed by sepiolite clay (final equilibrium pH values are presented near each marker).

28

(b)

(a)

300

200

T=35ºC

qeq (mg/g)

qeq (mg/g)

T=25ºC

150

100 T=25ºC T=35ºC T=25ºC - SDE 0

0 0

50

100

150

0

Ceq (mg/L)

50

100

150

Ceq (mg/L)

Fig. 5. Equilibrium isotherms for (a) BR and (b) DB adsorption on sepiolite, at different temperatures: experimental data, Langmuir (—) and Freundlich (---) modeling.

29

30

200

20

100

←

→

10

0

TOC Removal (%)

TOC Removal (mg/g)

300

0 0.0

1.0

2.0

3.0

m/v (g/L)

Fig. 6. Total organic carbon removal from an aqueous solution containing dyeing auxiliary chemicals, as a function of sepiolite dosage.

30

1.0

1.0 T=25ºC m/v=0.25 g/L

(a) 0.8

0.6

0.6

50 mg/L 100 mg/L

C/C0

C/C0

0.8

150 mg/L

0.4

0.4

0.2

0.2 15 mg/L 50 mg/L

0.0 0

3

30 mg/L 100 mg/L

6 time (h)

(b)

T=25ºC m/v=1.0 g/L

0.0 9

12

0

12

24

36

48

60

72

time (h)

Fig. 7. Effect of the contact time on the adsorption of (a) BR and (b) DB by sepiolite, using different initial dye concentrations: experimental data, pseudo-first (—) and pseudo-second order (---) modeling.

31

15

11.0

Desorption %

9.0 10

9.4%

← →

Desorption % Final pH

7.0 pH 5.0

5

2.9%

3.2%

3.0%

2.3%

0

3.0 1.0

2

4

6

7

11

initial pH

Fig. 8. Desorption percentages from BR-loaded sepiolite (4.0 g/L) as a function pH conditions.

32

Table 1. Summary of properties of sepiolite clay. Physical and Textural Properties Mean particle size (mm) a Apparent density Real density

(g/cm3) a

(g/cm3) a

Porosity (%) a

1.22 2.39 46.5

Micropore volume

(mm3/g) a

Mean pore size (nm) SHg

0.576

(m2/g) a,b

13 20.1 76

SBET a (m2/g) a,b

108

(m2/g) c

175

SMB

Mineralogical composition Sepiolite (wt.%) a

78

Chemical properties CEC (meq/g) a,d

0.27 9.8; 10.0; 10.4 e

PZNPC a

b

reported in [23]; Surface area, measured by N2 adsorption (B.E.T. method); c surface area, measured by MB adsorption; d Cation Exchange Capacity; e for NaCl 0.1 M, 0.01 M and 0.001 M, respectively.

33

Table 2. Equilibrium models for BR adsorption on sepiolite clay: parameters (±standard error) and determination coefficients. Langmuir

BR DB DB-SDE

T (ºC) 25 35 25 35 25

Freundlich

kL (L/mg)

Qm (mg/g)

R2

0.23±0.08 0.05±0.02 0.04 ±0.01 0.03±0.02 0.009±0.003

110±6 310±30 232±31 239±42 201±37

0.95 0.94 0.91 0.88 0.98

Δqeq (%) 8.2 20 14 17 16

n

KF (mg1-1/ng-1L1/n)

R2

7±2 2.9±0.5 2.7±0.6 2.5±0.7 1.5±0.2

56±8 54±13 35±12 30±14 4±2

0.95 0.92 0.88 0.83 0.97

Δqeq (%) 7.5 10 16 21 16

34

Table 3. Maximum adsorption capacities reported in literature for basic and direct dyes. Adsorbent

Adsorbate

Conditions

Qm (mg/g)

Reference

Commercial Activated Carbon

BR 22

273 K; pH 7

556

[39]

Commercial Activated Carbon

BR46

273 K; pH 7

106

[40]

Activated carbon (apricot stone)

BY 21

298 K; pH 10

135

[41]

Activated carbons (waste tyre)

BY 21

pH 7

91-567

[42]

Bentonite

BR 46

298 K; pH 6

333

[43]

Expanded vermiculite

BR 9

298 K; pH 6.8

25

[44]

Chitosan-ethyl acrylate (biopolymer)

BB 41

318K; pH 9

217

[14]

Untreated coffee husks

BB 9

303 K; pH 8

90

[45]

Chitosan/Fe-hydroxyapatite nanocomposite beads

MB

293 K

1324

[46]

Activated fly ash

MB

293 K; pH 9

14

[11]

Formosa papaya seed powder

MB

298 K; pH 8

86

[13]

Sepiolite

BR 46

298 K; pH 9

110

this work

Commercial Activated Carbon

DB 168

293 K; pH 7.4

18.7

[40]

Commercial Activated Carbon

DB 53

298 K; pH 2

92

[47]

Orange peel activated carbon

DY 12

pH 1.5

75.8

[48]

Ca-Bentonite

CR

293 K; pH 6.9

107

[49]

Industrial waste sludge

DB 85

298 K; pH 7

339

[6]

Diatomite/chitosan–Fe(III) composite

DO 2GL

298 K; pH 6

1250

[16]

Cupuassu shell

DB 53

298 K; pH 2.0

66

[12]

Sepiolite

DB 85

298 K; pH 9

232

this work

Basic dyes

Direct dyes

BR – Basic Red; BY – Basic Yellow; BB-Basic Blue; DBr – Direct Brown; DY – Direct Yellow; DR – Direct Red; CR – Congo Red; DO – Direct Orange

35

Table 4. Thermodynamic parameters estimated for BR and DB adsorption in sepilite.

BR

∆Gº25 ºC (kJ/mol)

∆Sº (kJ/mol/K)

∆Hº (kJ/mol)

-28.4 ± 0.8

-0.3 ± 0.1

-120 ± 37

n.s.

-15 ± 4

DB -24.0 ± 0.9 n.s. – not significant

36

Table 5. Kinetic model parameters (±standard error) for BR and DB adsorption on sepiolite. Pseudo-1st order C0 (mg/L) 15 30 BR 50 100 50 DB 100 150 Dye

Pseudo-2nd order

k1. 102(min-1)

qeq(mg/g)

R2

Δqeq (%)

k2.104(g.mg-1.min-1)

qeq(mg/g)

R2

Δqeq (%)

2.1±0.2 2.9±0.4 3.8±0.6 7±1 0.84±0.07 0.16±0.01 0.13±0.02

48±1 65±2 72±3 84±3 44±1 101±2 149±5

0.99 0.97 0.94 0.95 0.99 1.0 0.99

16 16 17 14 20 14 26

5.4±0.3 6.5±0.6 7±1 14±2 2.2±0.3 0.14±0.02 0.07±0.02

53±1 71±1 78±2 90±2 50±1 120±4 181±11

1.0 0.99 0.98 0.98 0.99 0.99 0.98

9.0 7.4 10 7.7 20 8.9 23

37

Table 6. LDF kinetic constants, homogeneous solid diffusivities and effective pore diffusivities calculated for BR and DB adsorption on sepiolite.

BR

DB

C0 (mg/L) 15 30 50 100 50 100 150

kLDF (s-1) 6.4×10-4 1.4×10-3 2.7×10-3 5.3×10-4 7×10-5 5×10-5 1×10-4

R2 0.95 0.94 0.92 0.92 1.0 0.99 0.98

Dh (m2/s) 3.5×10-12 8.0×10-12 1.5×10-11 2.9×10-11 3.7×10-13 2.8×10-13 5.5×10-13

DPe (m2/s) 1.9×10-8 6.6×10-9 3.8×10-9 1.5×10-9 1.3×10-9 4.3×10-10 5.8×10-10

38