Cationic xylan–METAC copolymer as a flocculant for clay suspensions

Shoujuan Wangabc, Mohan K. R. Konduric, Qingxi Hou*a and Pedram Fatehi*c
aPapermaking Science & Technology College, Tianjin University of Science and Technology, Tianjin, 300222, China. E-mail: qingxihou@tust.edu.cn
bKey Laboratory of Pulp and Paper Science and Technology of Education, Ministry of China, Qilu University of Technology, Jinan 250348, China
cChemical Engineering Department, Lakehead University, 955 Oliver Road, Thunder Bay, Ontario, Canada P7B 5E1. E-mail: pfatehi@lakeheadu.ca; Fax: +1-807-346-7943; Tel: +1-807-343-8697

Received 27th February 2016 , Accepted 11th April 2016

First published on 14th April 2016


Abstract

The work presented herein focused on the flocculation of kaolin and bentonite clay suspensions using cationic copolymerized xylan under controlled conditions. Cationic xylan-2-(methacryloyloxy)ethyl]trimethyl ammonium chloride (METAC) copolymers were produced under the conditions of pH 7, 3 h reaction time, 80 °C and 2 or 3 mol mol−1 METAC/xylan ratio. The charge densities of the produced cationic xylan copolymers, CMX1, and CMX2 were +1.8 and +2.4 meq. g−1 while their molecular weights were 88[thin space (1/6-em)]986 and 102[thin space (1/6-em)]545 g mol−1, respectively. The attachment of METAC to xylan was confirmed by elemental and gel permeation chromatography analyses. CMX2 was a more efficient flocculant than CMX1 as it adsorbed and removed more clay particles from the clay suspensions. CMX2 also changed the zeta potential and turbidity of the clay suspensions more remarkably than CMX1, which was attributed to its higher charge density and molecular weight. CMX2 was more efficient in flocculating bentonite than kaolin suspensions. The presence of CMX on the clay particles was confirmed by Fourier Transform Infrared Spectroscopy (FTIR) analysis. The size of kaolin particles increased from 3.2 to 9.8 and 11.7 μm, and that of bentonite particles from 5.8 to 13.8 and 15.5 μm by having 16 mg L−1 of CMX1 and CMX2 copolymers in the clay suspensions, respectively. The reversibility of the flocculation process was assessed with a photometric dispersion analyzer. Furthermore, unmodified xylan was ineffective in flocculating clay particles.


1. Introduction

The wastewater of mineral industries mainly consists of clay wastes, whose removal from suspensions is challenging due to their colloidal and charged properties.1 Flocculation using synthetic polymers has been regarded as a promising process to address this problem.2–5 However, there is a growing concern about the use of synthetic polymers in wastewater. To tackle this problem, natural based flocculants (i.e. biodegradable flocculants) should be utilized.

Xylan is a major constitute of woody and non-woody species. In the past, several studies reported the conversion of xylan to value-added products through chemical modifications.6–8 In one study, carboxymethylated xylan was produced via modifying wheat straw xylan with sodium chloroacetate and used in packaging films.9 In another work, sulfonated xylan was produced via treating corncob xylan with chlorosulfonic acid and dimethylformamide, and the product was used as an anticoagulant for human blood plasma.10 Alternatively, xylan–trimethyl ammonium butyrate chloride copolymer was produced through modifying birchwood xylan with 3-carboxypropyltrimethyl ammonium chloride, and the product was used as an additive for pulp fibers.11 Xylan based hydrogels were also successfully prepared via modifying hardwood xylan with 2-hydroxyethyl methacrylate, and the products were used as adsorbents for the removal of heavy metal ions, e.g. Pd, Cd, Hg, from aqueous solutions.12,13 However, limited studies were conducted on the preparation and application of xylan-based flocculants for clay suspensions.

In the past, the utilization of polysaccharides based flocculants in colloidal suspensions was reported.14–16 The addition of 5 mg L−1 of sulfonated cellulose to a clay suspension decreased the turbidity of kaolin suspension from 5000 NTU to 45 NTU.17 Wei and co-authors rendered starch cationic via grafting glycidyl octyl dimethyl ammonium chloride (GODAC) on starch, which resulted in 90% removal of clay from its suspension at a dosage of 0.29 g L−1.18 In another study, cationic guar gum was synthesized with N-(3-chloro-2-hydroxypropyl)-trimethyl ammonium chloride (CHPTAC) and the product exhibited 90% turbidity removal at a dosage of 3.6 mg L−1 in clay suspensions.19 However, the limited availability of guar gum as well as the heavy use of starch, chitosan and cellulose in medicine, agriculture, papermaking, food and other related industries would reduce the availability of starch, chitosan, and cellulose for producing flocculants. This created a driving force for generating flocculants from other biomaterials that have currently limited industrial applications. Xylan, a natural polymer, can be obtained as a by-product of several sugar platform industries. Xylan is also the main component of prehydrolysis liquor of kraft-based dissolving pulp process but currently under-utilized.20 Xylan is a linear polymer with abundant hydroxyl group that could be used for functionalization (e.g. to produce flocculants). Previously, Wang and co-authors produced a cationic xylan via grafting [2-(methacryloyloxy)ethyl]trimethyl ammonium chloride (METAC) on xylan, which removed 98% of orange 16 azo dye from a solution at the xylan–METAC concentration of 160 mg L−1.21 In that study, azo dye solution (a homogeneous colloidal solution) was used as a model for a flocculation analysis. However, the clay suspensions are heterogeneous colloidal suspensions with a broad range of particle size distribution and varied charge densities. The dye segments are very small and all dissolved in the solution, but clay particles are large particles and cannot dissolve in water. In other words, the dye segments and clay particles have very different densities and surface chemistries, and these factors impact the overall chemistry of the solutions and suspensions. The interaction of copolymer with clay suspensions is distinct from that with dyes.22 Thus, the mechanisms involved in the flocculation of copolymers with dyes and with clay suspensions would be very different.23,24 The first objective of this work was to investigate if xylan–METAC copolymer (CMX) can be an effective flocculant for clay suspensions and to understand the mechanisms of such flocculation.

Generally, polymers induce flocculation via charge neutralization, patching, polymer bridging, or combination of these mechanisms.23,24 Previous studies on polymer–clay interactions suggested that charged polymers adsorbed on the surface of clay particles mainly through electrostatic interactions.25,26 Electrostatic interactions and polymer conformation on the clay surface were strongly affected by the surface charge density of clay as well as the charge density and molecular weight of the polymer.27 The influence of these parameters on the adsorption of polyacrylamide polymers on clay particles was studied in the past.28–31 The increase in the cationic group of polyacrylamide from 10 to 50 mol% increased the turbidity removal of clay suspension by 90%.32 In another report, the reduction in the charge density of cationic dialdehyde cellulose from 4.27 to 4.03 mmol g−1 decreased the efficiency of the polymer in turbidity removal of calcium carbonate suspension from 55 to 35%.33 Based on these studies, it can be inferred that the properties of polymers significantly impacted their flocculation performance.34 The properties of cationic xylan (e.g. CMX) and these polymers, e.g. polyacrylamide and cationic dialdehyde cellulose, are different. Therefore, the results reported on the performance of these polymers cannot predict the performance of cationic xylan in flocculating clay in suspensions. The second objective of this work was to study the effect of charge density and molecular weight of CMX on the properties of clay suspensions.

The main novelty of this work was the study on the flocculation performance of CMX on bentonite and kaolin clay suspensions. In the present work, cationic xylan–METAC copolymers, CMX, with two charge densities and molecular weights were produced by grafting METAC monomer onto xylan under neutral conditions using potassium persulfate as an initiator.20 The chemical and physical properties of the CMX were characterized. The factors affecting the performance of cationic xylan–METAC copolymers as flocculants in clay suspensions were discussed for the first time in this work.

2. Experimental

2.1. Materials

Beechwood xylan, potassium persulfate (>99.0%), sodium nitrate (>99.0%), [2-(methacryloyloxy)ethyl]trimethyl ammonium chloride solution (METAC), 80 wt% in H2O and kaolin and bentonite clay were obtained from Sigma-Aldrich Company and used as received. Ethanol (99.5%) was obtained from Sigma-Aldrich Company and used as received or diluted to 80 wt% prior to use. Cellulose acetate dialysis membrane (molecular weight cut-off of 1000 g mol−1) was obtained from Spectrum Labs. Inc, USA. Polydiallyldimethylammonium chloride (PDADMAC) was obtained from Sigma Aldrich Company and diluted to 0.005 M prior to use. Potassium polyvinyl sulfate (PVSK) was purchased from Wako Pure Chemical Industries Ltd., Japan.

2.2. Preparation and purification of CMX

CMX was synthesized as described in the previous study.21 In this set of experiments, 1 g of xylan was dissolved in 40 mL of deionized water in a 250 mL three neck round bottom flask and kept in water bath at 30 °C with purging nitrogen for deoxygenating the solution. After incubation, two different ratios of 2 and 3 mol mol−1 METAC/xylan were maintained in the solutions, and the pH of the solutions was adjusted to 7 with 0.1 M NaOH. Then, 0.03 g of potassium persulfate was added as an initiator, and the copolymerization reaction was performed at 80 °C in a nitrogen environment for 3 h. After the reaction, the solution was cooled to room temperature and CMX was precipitated from the solutions via mixing solutions with 80 wt% ethanol. Afterward, the precipitated CMX was washed thrice with 99.5% ethanol. The solution was further centrifuged using Sorvall ST16 centrifuge (Thermofisher) at 3000 rpm for 10 min to separate the CMX from the solutions. The precipitated copolymers were dried at room temperature overnight and stored at 4 °C until further use.

2.3. Charge density analysis

In this study, unmodified xylan (UX) and xylan–METAC copolymers (CMX) were initially dried at 105 °C overnight to remove any moisture. Approximately, 0.2 g of sample was dissolved in 20 mL of deionized water and incubated for 2 h at 30 °C in a water bath shaker (Innova 3100, Brunswick Scientific Edition, NJ, USA) at 150 rpm. After incubation, the samples were centrifuged at 2000 rpm for 10 min to remove any undissolved sample and the charge density of the samples present in the supernatant was measured using a particle charge detector (Mutek, PCD 04, Germany). The titration analysis of CMX and UX was conducted using PVSK (0.0055 M) or PDADMAC standard solutions (0.0056 M) to determine the charge density of CMX or UX.21 All the experiments were performed in triplicates, and the average values were reported.

The surface charge density analysis of bentonite and kaolin clay particles was determined via following a back titration method with a Mutek, PCD 04, particle charge detector (Germany). Approximately, 0.2 g of clay was suspended in 50 mL of PDADMAC (0.005 M) solution and the suspension were incubated at 30 °C for 2 h at 150 rpm. After the incubation, the samples were filtered using Whatmann#1 filter membranes and the filtrates were titrated against PVSK (0.0055 M) solution. Similarly, the titration analysis was conducted for control (i.e. PDADMAC solution), and the difference was employed to quantify the surface charge density of clay particles.

2.4. Surface area analysis of clay particles

The surface area of bentonite and kaolin clay particles was determined by using Quantachrome surface area analyzer Nova2200e instrument. In this experiment, the samples were initially dried in an oven at 105 °C overnight, and approximately 0.05 g of sample was pretreated for 4 h at 250 °C prior to analysis. The specific surface area of the samples was then analyzed according to Brunauer–Emmett–Teller (BET) method via adsorption–desorption isotherms using nitrogen gas at −180 °C within relative pressure range of 0.01 to 0.99.35

2.5. Elemental analysis and grafting ratio

The elemental analysis of CMX was performed using an Elemental Vario EL elemental analyzer by the combustion method.36 The samples were dried at 60 °C in an oven, and approximately 0.02 g of dried samples was transferred into the carousel chamber of the elemental analyzer. The combustion was carried out at 1200 °C, and the generated gasses were reduced to analyze carbon, hydrogen, oxygen and nitrogen contents of the samples. As the nitrogen element is initially absent in UX,37 the nitrogen contents of the CMX copolymers are derived from METAC that is attached to xylan in the copolymers. The grafting ratios of METAC to xylan correspond to the nitrogen content of the CMX copolymers. The grafting ratio of METAC to xylan was calculated using eqn (1):38
 
image file: c6ra05223a-t1.tif(1)
where N is the nitrogen content of the samples in weight percentage, and Mw is the molecular weight of METAC, 207.7 g mol−1.

2.6. Molecular weight analysis

The molecular weight of CMX was measured using a gel permeation chromatography (Malvern, GPCmax VE2001 Module + Viscotek TDA305) equipped with RI and viscometer detectors. Polyanalytic PAC000101 and PAC000103 columns were used, and 5% acetic acid solution was used as an eluent or mobile phase. The temperature of the column was maintained at 30 °C. Polystyrenes were used as standards in this system. Approximately, 50 mg of sample was dissolved in 5 mL of 5% acetic acid solution by stirring at 200 rpm for 24 h at room temperature. The samples were filtered using a nylon 0.2 μm filter (13 mm diameter) and used for analysis at an eluent flow rate of 0.7 mL min−1.

The molecular weight of UX was measured by using Malvern GPCmax VE2001 Module + Viscotek TDA305 equipped with RI and viscometer detectors using polyethylene oxide as standard samples. The UX was dried overnight at 105 °C to remove any moisture and 100 mg of samples was dissolved in 10 mL of a 0.1 mol L−1 NaNO3 solution and filtered with a nylon 0.2 μm filter (13 mm diameter). The filtered solutions were used for molecular weight analysis. Polyanalytic columns, PAA206, and PAA203, were used at 35 °C. 0.1 mol L−1 NaNO3 solution was used as solvent and eluent. The flow rate of eluent was set at 0.70 mL min−1.

2.7. Particle size distribution analysis

The size distribution of bentonite and kaolin clay particles was analyzed by using a MasterSizer 2000 particle size analyzer (Malvern Instruments), which was equipped with a light scattering detector in the presence of UX or CMX. In this study, 4 g of clay suspension (25 g L−1) was added to 100 mL of deionized water or 100 mL of UX or CMX solution (4 mg g−1) and stirred at 150 rpm for 30 min at room temperature. After stirring, the samples were analyzed for their particle size distribution.

2.8. Adsorption studies

In this set of experiments, the clay suspensions of 1 g L−1 was prepared in 50 mL of deionized water, and UX or CMX was added to the suspensions in order to study the adsorption of UX or CMX on clay particles at different polymer dosages (2 to 64 mg g−1). The suspensions were stirred at 150 rpm for 30 min and room temperature. After mixing, the mixture was filtered using Whatman filter papers in order to prevent the interference of clay particles in the subsequent assessment.39 The concentration of UX or CMX before and after mixing with clay suspensions was determined in order to determine the adsorption of UX or CMX on clay particles.40 All the filtrates were hydrolysed with 4 wt% H2SO4 in an oil bath for 1 h at 121 °C.41 The concentration of UX or CMX present in the filtrates was measured with a high-performance liquid chromatography, HPLC 1260, Agilent Technologies, equipped with a refractive index detector. Deionized water was used as a mobile phase, and all the filtered samples were allowed to pass through ligand exchange Pb column at a flow rate of 0.6 mL min−1. The temperature of the column was maintained at 70 °C. All the experiments were performed in triplicates, and the average values were reported in the study.

2.9. Impact of UX or CMX on clay removal

In this set of experiments, UX or CMX (based on clay) was mixed with clay suspensions (50 mL) under the conditions of 1 g L−1 clay concentration at 30 °C for 30 min. A 10 mL sample was collected from the upper half of the suspension before and after the incubation and dried in an oven at 105 °C overnight. The concentration of clay particles in the samples was determined, which helped measure the removal of clay from suspensions. This analysis was repeated at different polymer dosages (2–64 mg g−1 based on clay). All the experiments were performed in triplicates, and the mean values were reported in the study.

2.10. FTIR analysis

The Fourier transform infrared spectrophotometer (FTIR) analysis was conducted on unmodified and CMX-treated kaolin and bentonite samples. The samples were dried at 105 °C overnight, and 0.02 g of the sample was then used for structural characterization using FTIR (Bruker Tensor 37, Germany, ATR accessory). The spectra were recorded in the transmittance mode in the range of 500 cm−1 and 4000 cm−1 with a 4 cm−1 resolution, and 32 scans per sample were conducted.

2.11. Zeta potential analysis

The zeta potential analysis of clay particles was performed using a compact automatic zeta meter (Laval labs Inc). One objective of this study was to relate the adsorption, zeta potential and flocculation performance of clay suspensions with copolymers. In this study, 4 g of clay suspension (25 g L−1) was added to 100 mL of deionized water and stirred at 150 rpm for 30 min at room temperature, and then its zeta potential was assessed. Similarly, the zeta potential of clay particles in the presence of UX or CMX was measured at different polymer dosages (2 to 64 mg L−1) under the same conditions as described above. All the measurements were carried out at room temperature with a constant electric field (8.4 V cm−1). The experiments were conducted three times, and the mean values were reported in this study.

2.12. Dispersion analysis

The flocculation of clay suspensions was investigated using a photometric dispersion analyzer (PDA 3000, Rank Brothers Ltd) connected to a dynamic drainage jar (DDJ) fitted with a 70 mm mesh screen.41,42 In this study, 500 mL of distilled water was first poured into the DDJ container and circulated from the DDJ to the PDA through a 3 mm plastic tube until a steady flow rate of 20 mL min−1 was achieved. The flow rate was regulated by a peristaltic pump throughout the experiment. Then, 20 mL of a 25 g L−1 clay suspension was added into DDJ (to make a 1 g L−1 clay concentration in DDJ) while stirring at 300 rpm. This caused a decrease in the initial DC voltage (V0) to a new DC voltage (Vi) in PDA analysis. After 100 s, 4 mg g−1 of UX or CMX (based on clay) was added to DDJ. The effect of polymer dosage on the flocculation of the clay suspension was studied. The increase in DC voltage was represented as the DC voltage (Vf) of the final suspension. The flocculation performance of the clay suspension, which was served by relative turbidity τr, was measured from the variation in the direct current (DC) voltage of the PDA instrument.43 The relative turbidity of the clay suspensions was measured using eqn (2):44
 
image file: c6ra05223a-t2.tif(2)
where τf is denoted as the final suspension turbidity, and τi is denoted as initial suspension turbidity.

To study the strength of clay flocs, 20 mL of 25 g L−1 clay suspension was added into DDJ (to make a 1 g L−1 clay concentration in DDJ) while stirring at 300 rpm. After 100 s, UX or CMX was added into the suspension present in the DDJ. After 500 s, agitation intensity was increased from 300 to 3000 rpm and maintained for another 100 s, during which flocs would break. Subsequently, the agitation intensity was decreased to 300 rpm and maintained for another 500 s. This process is repeated twice to study the flocculation behavior of clay suspensions in the presence of CMX. The degree of breakage and reflocculation was represented by flocculation index (FI), which is the ratio of the mean square root (RMS) of the voltage that flocculation occurs to the direct (DC) voltage of the PDA instrument.45 Thus, flocculation index provides a direct indication of flocculation and breakage of clay particles in the suspensions.46 All the experiments were performed in triplicates, and the mean values were reported in the study.

3. Results and discussion

3.1. Properties of CMX

The chemical structure of xylan–METAC copolymer (CMX)21 is shown in Fig. 1. As seen, the quaternary ammonium group in METAC provides the cationic charge density of METAC and thus its attachment to xylose unit of xylan polymer will render xylan cationic.
image file: c6ra05223a-f1.tif
Fig. 1 Chemical structure of xylan–METAC copolymer, CMX. “n” represents the number of xylose unit and “m” represents the number of METAC group attached to xylose.

The properties of CMX used in this study are listed in Table 1. The UX had a charge density of −0.6 meq. g−1. CMX1 and CMX2 had the charge densities of +1.8 and +2.6 meq. g−1, respectively. The nitrogen content of the UX, CMX1 and CMX2 were 0, 3.03, and 3.52 wt%, respectively. The nitrogen content of CMXs originates from the METAC segments attached to xylan, and the results showed that the more METAC in the reaction, the higher nitrogen content (the higher grafting ratio), the charge density and molecular weight of CMX would be.21,38 Previously, the molecular weight of starch was increased from 5 × 106 to 46 × 106 g mol−1 via copolymerizing starch with 2-hydroxy trimethyl ammonium chloride in the presence of potassium persulphate.47 The polydispersity of CMX1 (1.6) and CMX2 (1.53) was found to be lower than UX (1.7).

Table 1 Properties of UX, CMX1 and CMX2
Sample UX CMX1 CMX2
Time, h 3 3
Temperature, °C 80 °C 80 °C
pH 7 7
METAC/xylan, mol mol−1 2 3
Xylan concentration, g L−1 25 25
Charge density, meq. g−1 −0.6 (±0.1) +1.8 (±0.1) +2.6 (±0.1)
Nitrogen, wt% 0 3.03 (±0.02) 3.52 (±0.02)
Grafting ratio 0 81.7 110.5
Mw 20[thin space (1/6-em)]657 88[thin space (1/6-em)]986 102[thin space (1/6-em)]545
Mn 12[thin space (1/6-em)]119 55[thin space (1/6-em)]232 66[thin space (1/6-em)]854
Mw/Mn 1.71 1.6 1.53


3.2. Characterization of bentonite and kaolin clay

Table 2 lists the surface charge density and area of bentonite and kaolin clay particles. Bentonite had a larger particle size but smaller surface area than did kaolin particles. The large surface area of kaolin is due to its smaller particle size. Bentonite particles had slightly higher negative charge density (−9 μeq. g−1) than kaolin particles (−6 μeq. g−1). The high negative charge density of bentonite is probably due to the presence of a greater number of oxide anions on the bentonite surface than on kaolin surface.48
Table 2 Characteristics of bentonite and kaolin clay
Sample Surface charge, μeq. g−1 BET surface area, m2 g−1 Particle size, μm Total pore volume, cm3 g−1
Bentonite −9.0 20.121 5.8 0.012
Kaolin −6.3 55.645 3.2 0.027


3.3. Adsorption

Fig. 2 presents the adsorption of UX or CMX on the surface of kaolin (Fig. 2a) and bentonite (Fig. 2b) particles as a function of polymer concentration at pH 7. UX had a limited adsorption of 0.05 and 0.02 mg g−1 on kaolin and bentonite particles, respectively. Therefore, the hydrogen bonding of xylan to clay or bentonite particles was not a major factor for the adsorption. CMX1 and CMX2 exhibited a maximum adsorption of 3.1 and 4.6 mg g−1 on kaolin particles, respectively. The adsorption of CMX1 and CMX2 was 0.73 and 0.85 mg g−1 on bentonite particles, respectively. It was reported that polymers with a higher charge density would adsorb more on clay particles than those with a lower charge density.49 In another study, with the increase in the cationic starch concentration from 30 to 260 mg L−1 in a clay suspension, the adsorption of cationic starch on the clay surface was increased from 2 to 18 mg g−1 at pH 5.50
image file: c6ra05223a-f2.tif
Fig. 2 (a) Adsorption of UX and CMX on kaolin as a function of polymer dosage, (b) adsorption of UX and CMX on bentonite as a function of polymer dosage, (c) theoretical surface charge density of kaolin and bentonite clay particles as a function of adsorbed CMX conducted under the conditions of pH 7, 25 °C, 0.5 h and 1 g L−1 clay concentration.

Based on the amount of UX or CMX adsorbed on kaolin and bentonite particles (Fig. 2a and b) and the charge density of UX and CMX (Table 1), the surface charge density of particles can be theoretically calculated. Fig. 2c shows the impact of adsorbed UX or CMX on the theoretical surface charge density of kaolin and bentonite particles. The surface charge density of unmodified clay particles was experimentally determined as described in Section 2.3., but it was not possible to experimentally measure the surface charge density of the particles after CMX adsorption. It is seen in Fig. 2c that the surface charge density of kaolin particles would still be anionic when the adsorbed amount was less than 3.1 for CMX1 and 2.5 mg g−1 for CMX2. The surface charge density of bentonite particles was anionic even after reaching the saturation level of adsorption for CMX1 (0.73 mg g−1) and CMX2 (0.85 mg g−1). These results confirm that CMX partly covered the surface of clay particles even though unadsorbed CMX was available in the suspensions for adsorption when the adsorbed amount of CMX1 or CMX2 was less than 3.1 or 2.5 mg g−1 for kaolin and 0.73 or 0.85 mg g−1 for bentonite (Fig. 2c). This partial coverage may imply that not all of the anionic charges of clay particles were accessible to CMX, or the charges were not evenly distributed on the surface of clay particles.51

The adsorption results (Fig. 2a and b) were fitted into Langmuir (3) and Freundlich (4) isotherm models.52,53

 
image file: c6ra05223a-t3.tif(3)
 
image file: c6ra05223a-t4.tif(4)
where Ce is a concentration of CMXs at equilibrium (mg L−1), Qe is the amount of CMXs adsorbed on clay particles at equilibrium (mg g−1), Qmax is the maximum adsorption of CMXs on clay particles (mg g−1), Kl, Kf and n are Langmuir and Freundlich constants. Table 3 summarized the modeling parameters that were obtained from experimental results. It is clear that the adsorption isotherms of CMX1 and CMX2 fitted into Langmuir isotherm model better than Freundlich isotherm model.54

Table 3 Parameters of Langmuir and Freundlich adsorption isotherms of CMX1 and CMX2 on kaolin and bentonite
Sample Kaolin Bentonite
Langmuir Freundlich Langmuir Freundlich
Kl Qmax R2 Kf n R2 Kl Qmax R2 Kf n R2
CMX1 0.28 3.42 0.987 0.53 1.67 0.785 0.25 0.79 0.984 0.12 1.60 0.812
CMX2 0.35 4.92 0.995 0.93 1.87 0.800 0.44 0.89 0.996 0.19 1.96 0.804


3.4. FTIR analysis

The presence of CMXs on kaolin and bentonite particles was analyzed with FTIR spectrophotometry. Fig. 3a presents the FTIR spectra of CMX-modified kaolin particles. The bands displays between 3690 and 3618 cm−1 corresponds to O–H stretching of Al–O–H groups existing between the kaolin layers.55 The bands at 1113, 1030 and 1005 cm−1 correspond to Si–O, Si–O–Si, and Si–O–Al stretching, respectively.56 The bands at 911, 806 and 621 cm−1 correspond to Al–O–Al and Al–O stretching.57 Furthermore, the presence of other characteristic bands in the regions of 1718 and 1470 cm−1 represents C–N stretching of amide groups present in CMX adsorbed on kaolin.21,55
image file: c6ra05223a-f3.tif
Fig. 3 FTIR spectra of unmodified and CMX modified- (a) kaolin and (b) bentonite.

Fig. 3b presents the FTIR spectra of CMXs adsorbed bentonite particles. The bands present between 3712 and 3620 cm−1 constitutes the OH stretching vibration of Al–Al–OH and Al–OH–Mg groups.58 The strong band at 1020 cm−1 corresponds to stretching of Si–O groups.59 Furthermore, the bands at 918, 883 and 778 cm−1 correspond to the bending vibrations of Al–Al–OH, Al–OH–Mg and Si–O–Al groups, respectively.60 The spectra of bentonite treated with CMX1 and CMX2 exhibited characteristic bands in the region of 2986 cm−1 and 1718 cm−1, which corresponds to C–H stretching of N–CH3 group55 and C–N stretching of amide groups present in the CMX.21,60

These results suggest that the adsorption of CMX altered the surface chemistry of clay particles. The alternation was also reported by Zaman and co-authors for the adsorption of polyacrylic acid on kaolin.61 Furthermore, the characteristic bands of kaolin and bentonite treated with CMXs did not show any marked shifting,55 which might suggest that CMX adsorbed on the clay surface via electrostatic interactions (i.e. physical adsorption).

3.5. Zeta potential of clay particles

Fig. 4 presents the zeta potential of kaolin and bentonite particles as a function of UX or CMX dosage at pH 7. UX did not impact the zeta potential of the clay suspensions, which is due to the limited anionic charge density and adsorption of UX. However, the zeta potential of the kaolin and bentonite suspensions increased significantly via adding more CMX. The zeta potential of clay suspensions reached plateau levels at higher CMX dosages than 3 mg g−1. In one study, the zeta potential of a clay suspension was increased from −37.5 to +42.1 mV when the concentration of PDADMAC was increased from 0.16 to 3.2 mg g−1 in the clay suspension.62 Furthermore, the clay particles attained a higher zeta potential in the presence of CMX2 than CMX1, which is due to higher charge density and adsorption of CMX2 (Fig. 2a and b),30 and this probably made a more compact diffuse layer on clay particles.
image file: c6ra05223a-f4.tif
Fig. 4 Zeta potential of (a) kaolin and (b) bentonite particles in the presence and absence of CMX or UX as a function of polymer dosage conducted under the conditions of pH 7, 0.5 h, 25 °C and 1 g L−1 of clay concentration.

The results also depict that, at a CMX concentration higher than 16 mg L−1, the zeta potential of kaolin particles remained unchanged due to the saturation of negative charges located on the surface of the particles with the adsorbed CMX (Fig. 2a).63 In case of bentonite particles, at a CMX concentration higher than 8 mg L−1, the zeta potential remained unchanged due to limited adsorption of CMX (Fig. 2b).

To understand if the zeta potential of clay particles is affected by the charges of CMX adsorbed on the particles or by diffused CMX layers surrounding the particles in the suspension, the zeta potential of kaolin and bentonite suspensions was plotted against the theoretical surface charge of clay particles (Fig. 2c). Fig. 5 shows the zeta potential of the kaolin and bentonite suspensions as a function of the theoretical surface charge density of clay particles. The zeta potential of kaolin particles was −24 mV when its surface charge density was −6 μeq. g−1 (clay particles with no CMX adsorption). With an increase in the adsorption of CMX1 or CMX2 to 1 mg g−1 or 0.5 mg g−1, respectively, the surface charge density of kaolin particles was anionic, but its zeta potential significantly increased to a positive value. The zeta potential of bentonite particles was −30.08 mV while its surface charge density was −9 μeq. g−1 (bentonite particles with no CMX adsorption). With an increase in the adsorption of CMX1 and CMX2 to 0.73 and 0.85 mg g−1, respectively, the surface charge density of bentonite particles was increased only to −7.6 and −6.8 μeq. g−1 while the zeta potential was raised to −10 and −3 mV, respectively. This analysis confirmed that the surface charge density and zeta potential of the particles might not be directly related. The positive zeta potential of kaolin particles revealed that a diffuse layer of CMX existed on the adsorbed layer of CMX on the particles. In other words, the interaction among layers of CMX on the particles played an important role in the zeta potential of clay particles. A full surface coverage of CMX on kaolin particle was not required for a positive zeta potential, as the surface of the particles were still anionic (Fig. 5). When the theoretical surface charge density of clay particles became very positive for CMX2 on kaolin, the adsorbed layer probably was larger than the diffuse layer, which insignificantly impacted the zeta potential (50 mV) of kaolin clay particles. The overall anionic zeta potential of bentonite particles show that the overall diffused layer on bentonite particle had an anionic charge density, which is due to the limited adsorption of CMX on bentonite (Fig. 2b).


image file: c6ra05223a-f5.tif
Fig. 5 Zeta potential of kaolin and bentonite suspensions as a function of the theoretical surface charge density of particles, which was conducted under the conditions of pH 7, 0.5 h, 25 °C and 1 g L−1 of clay concentration.

3.6. Flocculation of clay suspensions

Fig. 6 shows the impact of CMXs on the settlement of clay particles and clarity of the suspensions. In this figure, samples 2 and 5 are bentonite and kaolin suspensions. It is seen that by adding 8 mg L−1 of CMX1 to bentonite and kaolin (samples 1 and 4), the particles settled after 30 min respectively. Similarly, by adding 4 mg L−1 of CMX2 to bentonite and kaolin suspensions (sample 3 and 6), the particles were settled after 30 min, respectively. However, this figure does not differentiate between the effectiveness of CMX1 and CMX2.
image file: c6ra05223a-f6.tif
Fig. 6 Impact of CMX1 and CMX2 on kaolin and bentonite suspensions, samples 2 and 5 are bentonite and kaolin suspensions, samples 1 and 3 are CMX1 and CMX2 treated bentonite suspensions, samples 4 and 6 are CMX1 and CMX2 treated kaolin suspensions conducted under the conditions of pH 7, 0.5 h, 25 °C and 1 g L−1 of clay concentration.

Relative turbidity and percentage removal of clay are the parameters studied in the literature to understand the flocculation behavior of clay suspensions.64,65 The influence of CMX on the relative turbidity and removal of kaolin and bentonite particles from the suspensions was presented in Fig. 7. This study is useful for determining the optimum polymer dosage required for flocculating clay particles. As can be seen, with the increase in CMX concentration from 0.25 to 16 mg L−1, the relative turbidity of the kaolin suspension was decreased to 0.58 and 0.41 in the presence of CMX1 and CMX2, respectively. The removal of kaolin for CMX1 and CMX2 were found to be 67.6 and 78.5 wt%, respectively. With increasing the dosage of CMX from 0.25 to 8 mg L−1 in bentonite suspension, the relative turbidity was decreased to 0.21 and 0.07 in the presence of CMX1 and CMX2, respectively. The removal of bentonite particles for CMX1 and CMX2 were found to be 85.3 and 97.1 wt%, respectively. This higher removal of bentonite particles is possibly due to the large size of bentonite particles that could be removed more easily from the suspensions (Table 2). Further increase in CMX concentration did not change the relative turbidity and clay removal. In one study, with an increase in the cationic pullulan dosage from 2 to 8 mg L−1, the removal of clay was increased from 10 to 90%.66 In another report, with an increase in the dosage of cationic polyvinyl alcohol from 0.5 to 3.0 wt%, the relative turbidity of the clay suspension was decreased from 1.0 to 0.8.67 Similarly, 2 mg L−1 concentration of cationic polyacrylamide led to 50 wt% clay removal from a 30 g L−1 suspension.32 The results in Fig. 7 also depicts that the CMX2 was a more effective flocculant, which could be due to its higher adsorption (Fig. 2), and/or its higher charge density and molecular weight.


image file: c6ra05223a-f7.tif
Fig. 7 Effect of xylan dosage on the relative turbidity of (a) kaolin and (b) bentonite as well as the removal of (c) kaolin and (d) bentonite particles in the presence of UX or CMX, which was conducted under the conditions of pH 7, 0.5 h, 300 rpm, 25 °C and 1 g L−1 of clay concentration.

Fig. 8a presents the relative turbidity of the kaolin and bentonite suspensions as a function of adsorbed CMX. It is evident that there is a close correlation between the relative turbidity and the amount of CMX adsorbed for both kaolin and bentonite particles, regardless of the CMX type. By increasing the amount of adsorbed CMX on particles, their relative turbidity dropped more significantly. At the same amount of CMX adsorbed, a smaller relative turbidity was obtained for CMX2 than CMX1, which is due to the large size and thus more effectiveness of CMX2 than CMX1 in flocculating particles. The last points on each curve represent the relative turbidity of clay suspensions at the maximum adsorption amount (i.e. saturation adsorption) that was achieved in Fig. 2. For these points (point s in Fig. 8a), the adsorption amount was close, but the amounts of unadsorbed CMX remained in the suspensions were different (Fig. 2). These results confirm that the amount of adsorbed CMX (and not the amount of unadsorbed CMX in suspensions) played a major role in the relative turbidity of clay suspension. At the same amount of CMX adsorption, smaller relative turbidity was obtained for bentonite than for kaolin particles. As bentonite had larger particles (Table 2), it needed less adsorption of CMX for flocculation and particle removals.


image file: c6ra05223a-f8.tif
Fig. 8 (a) Effect of adsorbed CMX on the relative turbidity of kaolin and bentonite suspensions, (b) effect of total charges introduced to particles on the relative turbidity of kaolin and bentonite clay suspensions under the conditions of pH 7, 0.5 h, 25 °C and 1 g L−1 of clay concentration.

Fig. 8b presents the relative turbidity of kaolin and bentonite suspensions as a function of total charges introduced on particles via CMX adsorption (based on results in Fig. 2). The results showed that a slightly lower relative turbidity was obtained for clay suspensions treated with CMX1 than for those treated with CMX2 while similar charges were introduced to clay particles. To introduce the same charge to clay particles, more CMX1 than CMX2 should be adsorbed as CMX1 had a lower charge density. Therefore, the higher adsorption and thus interaction of CMX and particles were the reasons for better performance of CMX1 than CMX2 in flocculating clay particles when the total charges introduced to clay particles were kept constant. However, as CMX2 was generally adsorbed more than CMX1 (Fig. 2a and b), a lower relative turbidity was obtained for CMX2 than CMX1 for both kaolin and bentonite particles (Fig. 8b). Furthermore, the relative turbidity of kaolin suspension was reduced less significantly compared to bentonite suspension via adsorbing CMX.

Fig. 9 shows the relative turbidity of kaolin and bentonite suspensions as a function of zeta potential. This study helps to determine the mechanism of CMX in removing clay and bentonite removal particles from suspensions. The relative turbidity of the kaolin suspension was reduced gradually when the zeta potential of kaolin suspension was positive indicating that CMX aid at removing kaolin particles via a patch and bridging mechanisms. Interestingly, the relative turbidity of bentonite suspension was reduced linearly with respect to its zeta potential change. This indicates that CMX possibly flocculated bentonite particles with mainly charge neutralization and patch mechanisms. These results imply that different mechanisms played major roles in the removal of kaolin and bentonite particles from suspensions.


image file: c6ra05223a-f9.tif
Fig. 9 Effect of the zeta potential of kaolin and bentonite suspensions on the relative turbidity of clay suspensions under the conditions of pH 7, 0.5 h, 25 °C and 1 g L−1 of clay concentration.

3.7. Particle size distribution of clay particles

Fig. 10 presents the size distribution of kaolin and bentonite particles in the presence and absence of UX or CMXs (16 mg L−1). In the absence of any polymer, the size of the kaolin and bentonite clay particles were 3.2 and 5.8 μm. The addition of UX showed an insignificant decrease in kaolin and bentonite size to 3.17 and 5.75 μm, respectively. Both CMX1 and CMX2 increased the size of kaolin particles from 3.2 to 9.8 and 11.7 μm, respectively. Furthermore, the size of the bentonite particles was increased to 13.81 and 15.53 μm in the presence of CMX1 and CMX2, respectively. As stated earlier, CMX2 had a larger molecular weight. Therefore, the large size of clay particles treated with CMX2 may be due to the larger size of adsorbed CMX2, or to the more adsorption of CMX2 on clay particles (Fig. 2). The adsorption of CMX also broadened the size distribution of clay particles (Fig. 10a and b). The high adsorbed amount of CMX incorporated number of clay particles together via charge neutralization and patch mechanisms contributing to the particle size and size distribution increases.68 This is because neutralized particles will not repel each other, in opposition to untreated particles, and they may collide together as they possess partly cationic charges. In another report, the size of clay particles increased from 0.1 to 100 μm via treating with 0.8 mg L−1 of PDADMAC (1.2 × 106) in a clay suspension.63 In another study, the addition of 4 mg g−1 polyacrylamide (1.3 × 106) increased the size of clay particles from 9 to 28 μm at the clay concentration of 20 g L−1 and pH 5.8.69
image file: c6ra05223a-f10.tif
Fig. 10 Size distribution of (a) kaolin and (b) bentonite particles in the presence and absence of UX or CMX conducted under the conditions of pH 7, 16 mg L−1 of polymer dosage, 25 °C and 1 g L−1 of clay concentration.

3.8. Effect of mixing conditions on floc strength

Fig. 11 presents the flocculation index of kaolin and bentonite particles as a function of time in the presence of CMX in the suspension. At 300 rpm, with the addition of CMX (16 mg L−1), the flocculation index of kaolin increased from 0.02 to 0.6 in the presence of CMX1 and from 0.02 to 0.8 in the presence of CMX2. When the shear rate increased and held for 100 s at 3000 rpm, the flocculation index dropped to 0.02 for both CMX/clay systems. Similarly, when the shear rate reduced to 300 rpm and held for some time, the flocculation index increased to 0.5 for CMX1 and 0.65 for CMX2. When the shear rate was increased for the second time to 3000 rpm and decreased to 300 rpm, a similar behavior was achieved, but the flocculation index was lower than that achieved in the first 300 s (Fig. 11a). In the case of bentonite suspension, the flocculation index was increased after the addition of CMXs (16 mg L−1) at a shear rate of 300 rpm. After 500 s, the shear rate was increased to 3000 rpm, which resulted in a decrease in the flocculation index to 0.01. When the shear rate decreased to 300 rpm no significant change in flocculation index was observed for CMX1 and CMX2. This behavior suggests that the flocs were poorly adhered to each other. The results also indicated that the flocculation process was fast, and CMX2 led to a higher flocculation index (i.e. more effective flocculant). The use of 0.13 mg g−1 PDADMAC in a clay suspension showed a decrease in the flocculation index from 3.0 to 0.8 when the shear rate was increased from 50 to 400 rpm and again rose to 2.5 when shear rate decreased to 50 rpm.70 These results suggest that during breaking and reflocculation of clay particles, CMX was not probably able to bridge clay particles efficiently, which implies that the flocculation process was mainly an irreversible process.71
image file: c6ra05223a-f11.tif
Fig. 11 Flocculation index of (a) kaolin and (b) bentonite suspensions as a function of time in the presence of 16 mg L−1 CMX conducted at 25 °C, 1 g L−1 of clay concentration.

4. Conclusions

CMX1 and CMX2 had more adsorption on kaolin particles (4.6 and 3.1 mg g−1) than bentonite (0.85 and 0.68 mg g−1). The presence of CMXs on kaolin and bentonite particles was confirmed with FTIR studies by absorption peaks at 1718 and 1470 cm−1. The zeta potential of the kaolin clay suspensions increased significantly from −24.3 to 43.9 and 56.7 mV due to the adsorption of CMX1 and CMX2 while the zeta potential of bentonite particles was increased slightly from −30.08 to −9.9 and −2.4 mV due to the low adsorption of CMX1 and CMX2, respectively. The relative turbidity of bentonite suspension was reduced more than that of kaolin due to the removal of more particles. The amount of charges introduced to the kaolin clay as a result of CMX adsorption on particles played a significant role in the reduction of relative turbidity. At the same charges introduced to clay particles via CMX adsorption, CMX1 was more effective in reducing the relative turbidity, but the overall efficiency of CMX2 was higher than that of CMX1 due to its higher adsorption on clay particles. Regardless of the charge density of CMX, the relative turbidity of kaolin and bentonite suspensions was directly related to the amount of CMX adsorbed. The particle size and distribution of kaolin and bentonite clay particles increased via adsorbing CMX and they were higher for CMX2 than CMX1. The floc breakage and regrowth studies also showed that the flocculation of kaolin particles was partly reversible, and that of bentonite particles were completely irreversible, and CMX2 was more effective in flocculating clay particles.

Acknowledgements

The authors would like to thank NSERC, Canada Research Chair, NOHFC-Industrial Research Chair and Canada Foundation for Innovation programs for supporting this research.

References

  1. M. J. Pearse, S. Weir, S. J. Adkins and G. M. Moody, Miner. Eng., 2001, 14(11), 1505–1551 CrossRef CAS.
  2. P. Kulkarni, R. A. Gundiah and B. M. Moudgil, Int. J. Miner. Process., 1991, 32, 259–263 CrossRef.
  3. T. Tripathy and B. R. De, J. Phys. Sci., 2006, 10, 93–107 Search PubMed.
  4. X. Guo, Z. Wu and M. He, Water Res., 2009, 43(17), 4327–4335 CrossRef CAS PubMed.
  5. E. Sabah, H. Yuzer and M. S. Celik, Int. J. Miner. Process., 2004, 74(1–4), 303–315 CrossRef CAS.
  6. K. P. Welcke, K. Schwikal, S. Daus and T. Heinze, Carbohydr. Polym., 2014, 100, 80–88 CrossRef PubMed.
  7. Y. Xu, M. A. Hanna and L. Isom, Open Agric. J., 2008, 2, 54–61 CrossRef CAS.
  8. T. Heinze, K. Petzold and S. Horing, Cellul. Chem. Technol., 2007, 41, 13–18 CAS.
  9. M. Alekhina, K. S. Mikkonen, R. Alen, M. Tenkanen and H. Sixta, Carbohydr. Polym., 2014, 100, 89–96 CrossRef CAS PubMed.
  10. W. Cai, Q. Chen, L. Xie, L. Yang and R. Zhang, Eur. Food Res. Technol., 2015, 240, 969–975 CrossRef CAS.
  11. B. Vega, K. P. Welcke, P. Fardim and T. Heinze, Carbohydr. Polym., 2012, 89, 768–776 CrossRef CAS PubMed.
  12. T. C. F. Silva, Y. Habibi, J. L. Colodette and L. A. Lucia, Soft Matter, 2011, 7, 1090–1099 RSC.
  13. X. W. Peng, L. X. Zhong, J. L. Ren and R. C. Sun, J. Agric. Food Chem., 2012, 60(15), 3909–3916 CrossRef CAS PubMed.
  14. B. R. Sharma, N. C. Dhuldhoya and U. C. Merchant, J. Polym. Environ., 2006, 14, 195–202 CrossRef CAS.
  15. L. Ghimici and M. Nichifor, J. Macromol. Sci., Part B: Phys., 2009, 48(1), 106–113 CrossRef CAS.
  16. R. Divakaran and V. N. S. Pillai, Water Res., 2001, 35(16), 3904–3908 CrossRef CAS PubMed.
  17. H. Liimatainen, J. Sirvio, O. Sundman, O. Hormi and J. Niinimaki, Water Res., 2012, 46, 2159–2166 CrossRef CAS PubMed.
  18. Y. Wei, F. Cheng and H. Zheng, Carbohydr. Polym., 2008, 74, 673–679 CrossRef CAS.
  19. X. Wan, Y. Li, X. Wang, S. Chen and X. Gu, Eur. Polym. J., 2007, 43(8), 3655–3661 CrossRef CAS.
  20. P. Fatehi and Y. Ni, Integrated Forest Biorefinery-Prehydrolysis/dissolving pulping process, in Sustainable Production of Fuels, Chemicals, and Fibers from Forest Biomass, ed. J. Zhu, X. Zhang and X. Pan, ACS Symposium Series, 2011, vol. 1067, ch. 18, pp. 475–506 Search PubMed.
  21. S. Wang, Q. Hou, F. Kong and P. Fatehi, Carbohydr. Polym., 2015, 124, 229–236 CrossRef CAS PubMed.
  22. C. Fleischmann, M. Leivenbruck and H. Reitter, Polymers, 2015, 7, 714–746 CrossRef.
  23. J. Gregory, The use of polymeric flocculants, American Institute of Chemical Engineers, New York, 1985, pp. 253–263 Search PubMed.
  24. L. T. Lee, R. Rahbari, J. Lecourtier and G. Chauveteau, J. Colloid Interface Sci., 1991, 147, 351–357 CrossRef CAS.
  25. J. Lyklema, Fundamentals of Interface and Colloid Science, Academic press, New York, 2000 Search PubMed.
  26. S. Pack, T. Kashiwagi, C. Cao, C. S. Korach, M. Lewin and M. H. Rafailovich, Macromolecules, 2010, 43(12), 5338–5351 CrossRef CAS.
  27. Y. Zhou, Y. Gan, E. J. Wanless, G. J. Jameson and G. V. Franks, Langmuir, 2008, 24, 10920–10928 CrossRef CAS PubMed.
  28. G. Atesok, P. Somasundaran and L. J. Morgan, Powder Technol., 1998, 54, 77–83 CrossRef.
  29. B. Ersoy, Int. J. Miner. Process., 2005, 75, 207–216 CrossRef CAS.
  30. L. Nabzar, E. Pefferkon and R. Varoqui, Colloids Surf., 1998, 30, 345–353 CrossRef.
  31. M. S. Nasser and A. E. James, Sep. Purif. Technol., 2006, 52, 241–252 CrossRef CAS.
  32. W. J. Chen, Sep. Sci. Technol., 1998, 33, 569–590 CrossRef CAS.
  33. J. Sirvio, A. Honka, H. Liimatainen, J. Ninimaki and O. Hormi, Carbohydr. Polym., 2011, 86(1), 266–270 CrossRef.
  34. S. Y. Yoon and Y. Deng, J. Colloid Interface Sci., 2004, 278(1), 139–145 CrossRef CAS PubMed.
  35. F. Oveissi and P. Fatehi, Environ. Technol., 2014, 35(20), 2589–2596 CrossRef PubMed.
  36. X. Peng, J. Ren, F. Peng and R. Sun, Adv. Mater. Res., 2011, 236–238, 292–296 CrossRef CAS.
  37. D. K. Buslov, F. N. Kaputski, N. I. Sushko, V. I. Torgashev, L. V. Solov'eva, V. M. Tsarenkov, O. V. Zubets and L. V. Larchenkov, J. Appl. Spectrosc., 2009, 76(6), 801–805 CrossRef CAS.
  38. R. Fang, Nat. Sci., 2009, 1(01), 17–22 CAS.
  39. A. S. Michaels and O. Morelos, Ind. Eng. Chem. Res., 1955, 47(9), 1801–1809 CrossRef CAS.
  40. P. Marco and J. Llorens, Colloids Surf., A, 2005, 270–271, 291–295 CrossRef CAS.
  41. P. Fatehi, J. Shen, F. C. Hamdan and Y. Ni, Carbohydr. Polym., 2013, 92, 2103–2110 CrossRef CAS PubMed.
  42. S. R. Ramphal and M. S. Sibiya, Drinking Water Eng. Sci., 2014, 5(3), 1895–1907 Search PubMed.
  43. F. Xiao, J. C. H. Huang, B. J. Zhang and C. W. Cui, Desalination, 2010, 237, 201–213 CrossRef.
  44. L. J. Wang, J. P. Wang, S. J. Zhang, Y. Z. Chen, S. J. Yuan, G. P. Sheng and H. Q. Yu, Sep. Purif. Technol., 2009, 67(3), 331–335 CrossRef CAS.
  45. D. Wang, J. Gregory and H. Tang, Drying Technol., 2008, 26, 1060–1067 CrossRef CAS.
  46. W. Yu, J. Gregory and L. Campos, Sep. Purif. Technol., 2010, 74, 305–309 CrossRef CAS.
  47. J. P. Wang, S. J. Yuan, Y. Wang and H. Q. Yu, Water Res., 2013, 47(8), 2643–2648 CrossRef CAS PubMed.
  48. C. U. Schmidt and G. Lagaly, Clay Miner., 1999, 34, 447–458 CrossRef CAS.
  49. G. Durand-Piana, S. Lafuma and R. Audebert, J. Colloid Interface Sci., 1987, 119(2), 474–480 CrossRef CAS.
  50. Y. Chen, S. Liu and G. Wang, Chem. Eng. J., 2007, 133(1–3), 325–333 CrossRef CAS.
  51. G. Zhou, Clays Clay Miner., 1992, 40(3), 365–368 Search PubMed.
  52. Q. Li, Q. Y. Yue, Y. Su, B. Y. Gao and L. Fu, J. Hazard. Mater., 2007, 147(1–2), 370–380 CrossRef CAS PubMed.
  53. N. Teki, O. Demirbas and M. Alkan, Microporous Mesoporous Mater., 2005, 85(3), 340–350 CrossRef.
  54. G. Durand-Piana, S. Lafuma and R. Audebert, J. Colloid Interface Sci., 1987, 119(2), 474–480 CrossRef CAS.
  55. S. Wei, L. Wen-li and H. Yue-hua, J. Cent. South Univ. Technol., 2008, 15, 373–377 CrossRef.
  56. M. R. Sheikh and A. G. Barua, Indian J. Pure Appl. Phys., 2013, 51, 745–748 CAS.
  57. S. Unlam, U. Irawati, Y. Arryanto and U. Sutarno, Indones. J. Chem., 2011, 11(1), 96–102 Search PubMed.
  58. F. G. Alabarse, R. V. Conceicao, N. M. Balzaretti, F. Schenato and A. M. Xavier, Appl. Clay Sci., 2011, 51, 202–208 CrossRef CAS.
  59. F. Ayari, E. Srasra and M. Trabelsi-Ayadi, Desalination, 2005, 185, 391–397 CrossRef CAS.
  60. H. Tahir, M. Sultan and Z. Qadir, Int. J. Chem., 2013, 5(3), 19–32 Search PubMed.
  61. A. A. Zaman, R. Tsuchiya and B. M. Moudgil, J. Colloid Interface Sci., 2002, 256(1), 73–78 CrossRef CAS.
  62. G. Petzold, M. Mende, K. Lunkwitz, S. Schwarz and H. M. Buchhammer, Colloids Surf., A, 2003, 218(1–3), 47–57 CrossRef CAS.
  63. J. Yu, D. Wang, X. Ge, M. Yan and M. Yang, Colloids Surf., A, 2006, 290(1–3), 288–294 CrossRef CAS.
  64. S. Farrokhpay, Adv. Colloid Interface Sci., 2009, 151, 24–32 CrossRef CAS PubMed.
  65. A. Bouvy, Eur. Coat. J., 1996, 11, 822 Search PubMed.
  66. L. Ghimici, M. Constantin and G. Fundueanu, J. Hazard. Mater., 2010, 181(1–3), 351–358 CrossRef CAS PubMed.
  67. Y. Sang and H. Xiao, J. Colloid Interface Sci., 2008, 326(2), 420–425 CrossRef CAS PubMed.
  68. M. L. Taylor, G. E. Morris, P. G. Self and R. S. Smart, J. Colloid Interface Sci., 2002, 250, 28–36 CrossRef CAS PubMed.
  69. Y. Zhou and G. V. Franks, Langmuir, 2006, 22(16), 6775–6786 CrossRef CAS PubMed.
  70. Z. Zhu, T. Li, J. Lu, D. Wang and C. Yao, Int. J. Miner. Process., 2009, 91, 94–99 CrossRef CAS.
  71. J. Gregory, Water Sci. Technol., 2004, 50(12), 163–170 CAS.

This journal is © The Royal Society of Chemistry 2016