Journal of Hazardous Materials 187 (2011) 264–273 Contents lists available at ScienceDirect Journal of Hazardous Materials journa l homepage: www.e lsev ier .com MnCl2 and MgCl for the removal of reactive dye synthe agg A.Z. Bouy A.G a Nancy Univer euvre b University of c University of 0 Tou d Lebanese Uni routh, e Nancy Univer ex, Fr a r t i c l Article history: Received 29 Ju Received in revised form 27 December 2010 Accepted 5 January 2011 Available online 12 January 2011 Keywords: Azo-dye pigm Levafix Brillian MgCl2 MnCl2 Brucite Feitknechite , mag ing t were performed in the presence or absence of auxiliary dyeing chemicals. They proved that (i) both divalent cation-based coagulants were effective in the treatment of those alkaline effluents, (ii) better performances in terms of color removal, residual turbidity, and settled volume, were achieved with manganese chloride, and (iii) the presence of dyeing auxiliaries significantly increases the required coagulant demand for treating the textile effluent. The dye removal mechanisms were investigated 1. Introdu Textile d the ecologi water cons that are par characteriz ity, a high [2]. In addi effluents re treatment p and to mic to be toxic of aquatic ∗ Correspon ment Toulouse E-mail add bruno.lartiges 0304-3894/$ – doi:10.1016/j. ent t Blue EBRA by combining observations of freeze-dried sediments with transmission electron microscopy coupled with energy-dispersive X-ray spectroscopy and selected area electron diffraction, Fourier transform infrared spectroscopy, adsorption experiments, and aggregates size measurements with a laser sizer under cyclic shear conditions. The results show that brucite (Mg(OH)2) particles are formed when apply- ing MgCl2 to the textile wastewaters, whereas a mixture of feitknechite (�-MnOOH) and hausmannite (Mn3O4) is obtained when using MnCl2. More poorly crystallized particles are formed in presence of auxiliary dyeing chemicals. The adsorption experiments suggested that the azo-dye pigment adsorbs onto the surface of precipitating phases, whereas the aggregation dynamics indicated that a charge- neutralization mechanism underlies the formation of aggregates. The dye removal is then consistent with a precipitation/adsorption mechanism. © 2011 Elsevier B.V. All rights reserved. ction yeing is one of the main contributing categories to cal footprint of textile production [1]. Besides a high umption, this process generates colored wastewaters ticularly difficult to treat. Indeed, textile effluents are ed by a large amount of dyestuffs, a strong alkalin- chemical oxygen demand, and a low biodegradability tion, the chemical structure of dyes contained in the sists degradation in most conventional wastewater rocesses because of their stability to oxidizing agents roorganisms [3–5]. As most dyes have been shown to some organisms and to cause direct destruction communities [6], the discharge of partially treated ding authors at: University of Toulouse (UPS) - Geosciences Environ- UMR 5563, 14 Av. E. Belin, F-31400 Toulouse. resses:
[email protected] (A.Z. Bouyakoub), @lmtg.obs-mip.fr (B.S. Lartiges). effluents often severely impacts the receiving water systems [7]. Various methods have been proposed for the treatment of col- ored wastewaters, namely, oxidation [8–10], electrolysis [11–13], biodegradation [14,15], adsorption [6,16–18], chemical coag- ulation [5,6,19–25] and membrane filtration [1,5,26–28]. The oxidationprocesses certainly represent thebest techniques toelim- inate the total organic carbon, but they have been shown to be effective only at very low dye concentrations [5,25]. Electrochem- ical processes can be easily adapted to the polluant load but the formation of an abundant hydroxide sludge represents a major drawback [24]. Conventional biological process is generally shown to be less efficient at degrading dye pigments because of their complicated aromatic structure andmolecular size [25]. Both coag- ulation and adsorption processes have been proven to be highly effective for decoloring textile effluent, but the costs associated with actived carbon requirement and sludge treatment appear pro- hibitive [24,25]. Finally, the performance of membrane processes for the removal of reactive dyes is often limited by severe fouling [24]. see front matter © 2011 Elsevier B.V. All rights reserved. jhazmat.2011.01.008 2 tic textile wastewaters: An adsorption/ akouba,b,∗, B.S. Lartigesa,c,∗, R. Ouhibb, S. Kachab, sity - LEM-ENSG/INPL-CNRS, Pôle de l’Eau, 15 Avenue du Charmois, BP 40 54501 Vando DjillaliLiabes - LMSR, BP 89, 22000 Sidi Bel Abbes, Algeria Toulouse (UPS) – Geosciences Environment Toulouse UMR 5563, 14 Av. E. Belin, F-3140 versity – Graduate School of Sciences & Technologies, Campus Rafic Hariri-Hadath Bey sity - SCMEM-FST/UHP, 7137 boulevard des Aiguillettes, BP 239 54506 Vandoeuvre Ced e i n f o ly 2010 a b s t r a c t Two divalent cation-based coagulants synthetic textile wastewaters contain / locate / jhazmat Levafix Brilliant Blue EBRA from regation mechanism . El Samranid, J. Ghanbajae, O. Barresa Cedex, France louse, France Lebanon ance nesium chloride and manganese chloride, were used to treat he azo-dye pigment Levafix Brilliant Blue EBRA. The jar-tests A.Z. Bouyakoub et al. / Journal of Hazardous Materials 187 (2011) 264–273 265 Table 1 Nature of dyeing auxiliary chemicals. Product Chemical nature Use Migrasol MV Amide acrylic polymer/sodium Anti-migrating agent Meropan XR Subitol RNC Urea Chemica to be an ef tile effluent alum or PA textile was ered from t by which t treatment h ticular, the by Leentvaa The purpos into the abi MnCl2, to e from synth a pigment pound is ab are availabl more, the i of dye rem fore, we us dye in the gate the im elimination UV–vis spec by the chem infrared spe binedwith e the structur laser sizer u 2. Experim 2.1. Chemic The com (EBRA), wa turer from (Patent no. dichloroqui dine. Dyein Subitol RN (99%) and N Aldrich, res their role u which fixat auxiliary ch Reagent obtained fr as coagulan stock soluti 2.2. Prepara Two typ SW1 was p NaHCO3 (1 to SW1 the various dyeing auxiliary chemicals, i.e. urea (50g/l), Subitol RNC (5g/l), Rapidoprint XR (3g/l), and Migrasol MV (10g/l). Prior to coagulation treatment, the synthetic waters were first heated at 50–70 ◦C for 15min to provide hydrolysis conditions sim- those o roo efflue y DEN n rat greg rder e con tted out ight was dditi on u e sur oced 60 rp n. he en n gra red a idual ed su llulo , con bsor 01 sp tedb t and greg ddit gate l coa eflo sizer ment as u reme tic w of th d a 5 easu e dis ean term actio sion dime natu was acrylate Pearls Mononitrobenzene sulphonate of sodium Oxidizing agent Special sulfonates Dampening agent Diamide carbonic acid Stabilizing agent l coagulation using magnesium salts has been shown fective alternative to conventional treatments of tex- s, especially since (i) conventional coagulants such as C are not operational at pHs encountered in typical tewaters [41] and (ii) magnesium can be easily recov- he sludge and reused [25]. However, the mechanisms he dye pigments are amenable to removal in such a ave only been roughly examined [21,23,29,30]. In par- adsorptive-coagulating mechanism initially suggested r and Rebhun [30], needs to be further substantiated. e of this work is then to provide some new insights lity of two divalent cation-based coagulants, MgCl2 and liminate the reactive dye Levafix Brilliant Blue EBRA etic textile wastewaters. Levafix Brilliant Blue EBRA is dye mainly used for jeans dyeing. Although this com- undantly present in textile effluents, only scarce data e regarding its removal fromwastewaters [23]. Further- nfluence of auxiliary dyeing chemicals on the process oval is commonly overlooked in the literature. There- ed two types of synthetic effluents, i.e. EBRA reactive presence or absence of dyeing auxiliaries, to investi- portance of those chemicals on dye coagulation. The of pigment dye was assessed by jar-test coupled with troscopymeasurements, themineral species generated ical treatment were identified using Fourier transform ctroscopy and transmission electron microscopy com- nergy-dispersiveX-ray spectroscopy (TEM–EDXS), and al characteristics of aggregates were examined with a nder cyclic shear conditions. ental als mercial dye used in this study, Levafix Brillant Blue s kindly supplied by DENIM, a jeans textile manufac- Algeria. The patent protecting its chemical formula F 51942), states that the chromophore part can be either noxaline, monofluorotriazine, or difluorochloropyri- g auxiliary chemicals, Migrasol MV, Meropan XR, and C, were graciously provided by CHTR (France). Urea aHCO3 (99.7%) were acquired from Acros organics and pectively. The chemical nature of these products and pon dyeing, are indicated in Table 1. Unlike EBRA dye ion rate on the textile fiber is about 70%, the dyeing emicals end up in the effluent. grade MnCl2·4H2O (99%) and MgCl2·6H2O (99%) om Acros organics and Sigma, respectively, were used ts. They were both freshly prepared as 1M coagulant ons with deionized water (MilliQ-plus 18.2M�). ilar to down t thetic used b fixatio 2.3. Ag In o ieswer high) fi carried the he waters wise a agitati the fre test pr ringat additio At t settle i measu for res collect size ce meter) EBRA a UV-25 calcula effluen 2.4. Ag In a investi optima tion. Th Master experi [31] w measu synthe the pH to yiel Size m floc siz their m was de ume fr dimen 2.5. Se The tration tion of synthetic waters es of synthetic effluents were used in this study. repared by dissolving EBRA reactive dye (30mg/l) and g/l) in deionized water. SW2 was prepared by adding 5804–2860 Sentry) and and then fr excess salt microscopy TEM (200k encountered during dyeing. They were then left to cool m temperature (25 ◦C). The EBRA concentration in syn- nts was determined from the initial dye concentration IM in industrial operation (100mg/l) and from the dye e (70%). ation procedure and supernatant characterization to limit the consumption of reagents, coagulation stud- ducted in 150ml reactors (60mmdiameter and 80mm with 4 Plexiglas baffles (8mm×70mm). Stirring was with a 3.15 cm×1.5 cm blade positioned at one-third of the reactor from the bottom. The pH of synthetic adjusted to 12, pH of DENIM textile effluents, by drop- on of NaOH 1N. The coagulant was then added under sing a micro-pipette (Eppendorf) at a point just below face of the effluent. Mixing followed a conventional jar urewith rapidmixingat 250 rpmfor3minandslowstir- m for20min. ThepHwasnot readjustedafter coagulant d of mixing, the coagulated suspension was allowed to duated Imhoff cones for 2h. The sediment volume was nd 20ml of supernatant was siphoned with a syringe turbidity assessment (Hach 2100N turbidimeter). The pernatant was then filtered through a 0.22�m pore se-acetate filter (Macherey-Nagel) for pH (LPH330T pH- ductivity (CD 810 Tacussel), and color measurements. bance was measured at �=590nm using a Shimadzu ectrophotometer. The percentage of color removal was y comparing theabsorbancevaluesof original synthetic supernatant. Ultrapure water served as reference. ate size measurements ion to classical jar-tests, the coagulation dynamics was d by monitoring the floc size distribution obtained at gulant concentrationunder cyclic step changes in agita- c sizedistributionwasmeasuredon-linewithaMalvern using a particle size detection range of 1.2–600�m. An al set-up similar to that described in Chaignon et al. sed in this study. To avoid multiple scattering in the nt cell, the coagulated suspensions were diluted with ater prepared without EBRA dye and were adjusted to e corresponding suspension. The dilution was adjusted % volume concentration for all coagulated suspensions. rements were averaged over 1 s and taken every 2 s. The tributions are monomodal, and can be represented by floc diameter dm. The fractal structure of aggregates ined from the equation � ∼dm(3−Df), where � is the vol- n given by the particle sizer, and Df is the mass fractal of aggregates [31]. nt characterization re of aggregates obtained at optimal coagulant concen- examined by (i) TEM–EDXS on centrifuged (Eppendorf g for 10min) and then freeze-dried sediment (EL 105 (ii) FTIR on dialyzed (Spectro Por 6 – MWCO 2000) eeze-dried sediment. Dialysis was required to remove originating from dyeing auxiliary chemicals. Electron observationswere performedwithwith a Philips CM20 V) coupled with an Princeton Gamma Tech energy dis- 266 A.Z. Bouyakoub et al. / Journal of Hazardous Materials 187 (2011) 264–273 50 60a b y ( N T U ) MnCl2 MgCl2 40 50 60 0,1 ( m L ) Mgocc Mnocc 0,1 Mnocc n (mo Fig. 1. Compa rbidit after coagulati persive X-r re-suspend pension wa Elemental a X-ray emiss tified after with a Bruk 1mg of free grade,Merc a vacuum p range with 2.6. Adsorp The adso itate obtain optimal coa dyewere th of adsorbed the total ad after 2h set 3. Results 3.1. Jar test Figs. 1 an tions in term conductivit and SW2. S residual tur centration coagulant a are defined bidity value a MnCl2 tre appl g SW gges com rmor coa auxi rved arac ed i ed w 0 10 20 30 40 c d 0 0,1 0,2 0,3 0,4 0,5 R es id u al T u rb id it Mgocc Mnocc 0 10 20 30 0 S et tl ed v o lu m e 6 7 8 9 10 11 12 13 0 0,1 0,2 0,3 0,4 0,5 p H af te r co ag u la ti o n ) MnCl2 MgCl2 Mgocc Mnocc 0 2 4 6 8 10 12 0 C o n d u ct iv it y ( m s/ cm ) Coagulant concentratio rison of MnCl2 (�) and MgCl2 (©) coagulation performances for SW1: (a) residual tu on. The arrows indicate the optimal coagulant concentrations. ay spectrometry. The freeze-dried sediment was first ed in ethanol under ultrasonication, and a drop of sus- s then evaporated on a carbon-coated copper grid. nalysis was carried out with a 20nm probe size, K� ion lines ofNa, S, Cl,Mg,Mn, being integrated andquan- a 30 s counting time. Infrared analysis was conducted er IFS 55 spectrophotometer in the transmission mode. ze-dried sediment was mixed with 250mg KBr (FTIR k), and a pelletwas prepared using a press connected to ump. The spectra were recorded in the 4000–400 cm−1 200 scans collected at 2 cm−1 resolution. MgCl2 treatin This su iliaries Furthe ganese dyeing is obse ties ch increas obtain tion isotherms rption experiments were performed using the precip- ed from the treatment of SW1 without EBRA dye at gulant concentration. Various amounts of the pigment enaddedunder slowmixing for 30min, and the amount EBRA dye was calculated from the difference between ded concentration and the supernatant concentration tling. and discussion results d2 compare theperformance ofMnCl2 andMgCl2 solu- s of turbidity removal, sediment volume, final pH, and y after coagulation, for the two synthetic effluents SW1 imilar patterns are observed for both coagulants: thus, bidities (Figs. 1a and 2a) increase at low coagulant con- to reach a maximum, and then decrease with further ddition. The optimum coagulant concentrations (OCC) as the minimum dosages to attain the low residual tur- s (arrows in Figs. 1a and 2a). The OCC values found with atment are slightly lower than those obtained with a residual tur tions. All the s trations to r noting that sludgeprod the first cas cals does no by coagula until OCC to respectively all cases. On underdosag concentrati as abateme SW2, the pe coagulant c removal can 3.2. Fate of At first a ume, pH, a hydrolysis: both residu 0,2 0,3 0,4 0,5 0,2 0,3 0,4 0,5 MnCl2 MgCl2 Mgocc l/L) MnCl2 MgCl2 y, (b) sediment volume, (c) pH after coagulation, and (d) conductivity ication, whereas the required coagulant demands for 2 are almost double than those determined for SW1. ts that the negatively charged molecules of dyeing aux- pete with the reactive dye for the coagulant species. e, the residual turbidity at OCC is lower for the man- gulant in the case of SW1, whereas in presence of all liary chemicals (SW2), an equivalent turbidity removal for both chemicals. Above OCC, the residual turbidi- terizing SW1 treatment are stable, and then strongly n the case of MnCl2 coagulation. For turbidity curves ith SW2, the range of optimal dosing is reduced, as the bidities slightly increase at higher coagulant concentra- ediments build-up from the lowest coagulant concen- each a plateau around OCC (Figs. 1b and 2b). It is worth MnCl2 is clearly more effective than MgCl2 in terms of uction, sedimentvolumesatOCCbeing25–30% lower in e. Furthermore, the presence of dyeing auxiliary chemi- t significantly increase the sediment volume generated tion. The pH measurements show a marked decrease reach a pH of about 9.4 and 8.3 for MgCl2 and MnCl2, . Above OCC, the pH diminishes with a lesser slope in the other hand, the conductivity slightly decreases at es, and then increases proportionally with coagulant on above OCC. The color removal is illustrated in Fig. 3 nt rate vs coagulant concentration. For both SW1 and rcentage of EBRA dye removal strongly increases at low oncentration and reaches about 90% at OCC. A complete even be observed at slight overdosages. dye and nature of coagulating species pproximation, the variations in turbidity, settled vol- nd conductivity, may simply be related to coagulant the formation of an hydroxide precipitate can increase al turbidity and sediment volume, the associated A.Z. Bouyakoub et al. / Journal of Hazardous Materials 187 (2011) 264–273 267 0 10 20 30 40 50 60 70 80 a b c d 0 0,1 0,2 0,3 0,4 0,5 R es id u al t u rb id it y ( N T U ) MnCl2 MgCl2 Mgocc Mnocc 9 10 11 12 13 H a ft er c o ag u la ti o n MnCl2 MgCl2 Mgocc Mnocc 5 6 7 8 9 10 11 12 0, o n d u ct iv it y ( m s/ cm ) Mgocc 0 10 20 30 40 50 60 70 80 0 0,1 0,2 0,3 0,4 0,5 S et tl ed v o lu m e (m L ) MnCl2 MgCl2 Mgocc Mnocc ol/L) Fig. 2. Compa rbidi after coagulati OH− consu decrease in the release tion of hyd SW1 coagu platelets an (Fig. 4a), w particles ar phous phas MnCl2 or M lating speci particles. EDXS m particles ob fies a Mg(O Mg/O is som of the auxil A slightly h hydrolysis amounts of associated ation the gs a 4 A˚ s –pe togra ce in , Mn hat c nd h coun The p Fig. 3. Pe 6 7 8 0 0,1 0,2 0,3 0,4 0,5 p 2 3 4 0 C Coagulant concentration (m rison of coagulation performances for SW2 (SW1+dyeing auxiliaries): (a) residual tu on. The arrows indicate the optimal coagulant concentrations. mption implying a drop in pH, and the corresponding conductivity being, in our case, likely compensated by of chloride ions from the added coagulants. Examina- roxide precipitates by TEM–EDXS shows that, during lation, both well-crystallized though irregular shaped d small pseudo-cubic units are obtained with MnCl2 hereas monodispersed, but more poorly crystallized e formed with MgCl2 (Fig. 4c). In contrast, nearly amor- es are precipitated during SW2 coagulation with either gCl2 (Fig. 4b and d), the crushed sheet texture of coagu- es being nevertheless reminiscent of previous plate-like coagul confirm (spacin and 2.0 to MgO diffrac presen 2.64 A˚) 1.9 A˚) t chite, a may ac EDXS. icroanalyses yield an atomic ratio Mg/O close to 0.5 for tained with the MgCl2 coagulant, which likely identi- H)2–brucite precipitate. In the case of SW2 sediments, etimes lower than 0.5 because of the oxygen content iary dyeing chemicals associated with the precipitate. igher Mn/O of 0.6 characterizes the particles formed by of MnCl2. Elemental analyses also indicate that minor sulfur, originating either from Subitol or Meropan, are with the hydroxide precipitate generated during SW2 nite is actu at basic pH The infra verify the p (Fig. 5a). Br at 3698 cm 3212 cm−1 quency reg [34]. On the 0 10 20 30 40 50 60 70 80 90 100 a b 0 0,1 0,2 0,3 0,4 0,5 C o lo r re m o v al ( % ) MnCl2 MgCl2 Mgocc Mnocc 0 10 20 30 40 50 60 70 80 90 100 0 0, C o lo r re m o v al ( % ) Coagulant concentration (mol/L rcentage of color removal as a function of coagulant concentration. (�) MnCl2 and (©) M 1 0,2 0,3 0,4 0,5 MnCl2 MgCl2 Mnocc ty, (b) sediment volume, (c) pH after coagulation, and (d) conductivity . The electron diffraction patterns (insets Fig. 4c and d) presence of brucite in both SW1 and SW2 sediments t 2.39, 4.72, 1.82, 1.55 A˚), but additional spacings at 3.05 uggest that part of original brucite has been converted riclase under the electron beam [32]. Similarly, electron ms of Mn-precipitates (insets Fig. 4a and b) reveal the SW1 of feitknechite �-MnOOH (spacings at 4.62 and O2–ramsdellite (weaker reflections at 4.04, 2.54 and ertainly results from electron beam damage to feitkne- ausmannite Mn3O4 (peak at 3.09 A˚). The latter mineral t for the Mn/O mean atomic ratio of 0.6 measured by recipitation of a mixture of feitknechite and hausman- ally expected when divalent manganese is hydrolyzed under aerated conditions [33]. red spectra taken from freeze-dried sediments allow to resence of both brucite and hydrous manganese oxide ucite is characterized by a very sharp and intense peak −1 (OH stretching vibration), a broad shoulder around (strongly bonded water), and bands in the low fre- ion at 638, 568, and 436 cm−1 (MgO translation modes) other hand, the absorption bands around 629, 525 and 1 0,2 0,3 0,4 0,5 MnCl2 MgCl2 Mgocc ) Mnocc gCl2. The arrows indicate the optimal coagulant concentrations. 268 A.Z. Bouyakoub et al. / Journal of Hazardous Materials 187 (2011) 264–273 Fig. 4. Electron micrographs and electron diffraction patterns of freeze-dried coagulated sediments: (a) SW1 at MnCl2 optimal concentration; well-crystallized platelets correspond to feitknechite whereas small pseudo-cubic minerals identify hausmannite. (b) SW2 at MnCl2 optimal concentration. (c) SW1 at MgCl2 optimal concentration. (d) SW2 at MgCl2 optimal concentration. 418 cm−1 may be assigned to stretching vibrations of Mn–O bonds [35,36]. The bands of inorganic species largely dominate all spectra, even though EBRA dye (e.g. bands at 2924 and 2854 cm−1) can be recognized in SW1 sediments, and urea and Migrasol (see Table 2 for peaks assignment) can be readily distinguished in SW2 sed- iments. It should be noted that part of EBRA dye molecules and auxiliary dyeing chemicalsmayhave been removedduring dialysis. The infrared spectra of various dyeing auxiliaries are presented in Fig. 5b. Peak positions and assignments are listed in Table 2. Closer examination of EBRA dye infrared features provides supplemen- 40 36 99 16 36 44 7 a Fig. 5. (a) FTIR MgCl2 optimal Subitol; (9) Ra 34 18 29 33 28 59 42 (1) 17 17 21 0628 54 29 25 30 78 34 16 33 52 16 74 34 44 19 2228 6630 963 45 8 29 453 20 933 85 16 39 17 30 20 9228 54 29 27 29 66 34 37 19002400290034003900 wavenumber (cm -1 ) (5) (9) (8) (7) (6) 36 20 34 30 78 29 29 28 55 16 28 37 03 34 39 33 51 32 16 29 25 28 53 16 67 28 53 29 24 32 123 43 8 36 27 16 64 (2) (3) (4) b spectra of dialyzed and freeze-dried coagulated sediments: (1) SW1 at MgCl2 optimal concentration; (4) SW2 at MnCl2 optimal concentration. (b) FTIR spectra of EBRA dye and pidoprint. The peak assignments are given in Table 2. 11 14 5 15 58 15 27 14 42 12 69 12 33 10 28 84 7 62 9 52 5 8 15 70 10 231 23 0 12 03 11 02 55 3 63 2 73 4 80 0 94 6 89 1 10 52 12 67 14 10 14 441 53 316 18 14 63 10 57 10 03 77 8 57 5 52 6 11 53 16 27 42 5 56 91 53 3 13 50 11 92 66 7 61 9 10 45 14 29 16 33 73 3 80 687 9 10 851 23 9 13 19 16 68 16 16 15 64 14 56 14 08 13 25 11 19 5 63 10 36 11 57 13 78 12 5014 69 4009001400 15 69 14 53 14 28 12 66 12 33 10 66 10 24 94 6 87 7 81 5 41 41 8 16 23 8 761 04 0 11 681 45 3 13 52 13 19 58 2 53 5 58 1 87 710 40 11 69 13 29 13 53 14 58 16 24 concentration; (2) SW1 at MnCl2 optimal concentration; (3) SW2 at dyeing auxiliary chemicals: (5) EBRA dye; (6) urea; (7) Migrasol; (8) A.Z. Bouyakoub et al. / Journal of Hazardous Materials 187 (2011) 264–273 269 Table 2 Main infrared absorption bands of EBRA dye pigment and dyeing auxiliaries. Peaks are assigned according to [42–44]. Band (cm−1) Peak assignment Urea 1158 1463 1627 1674 Subitol 903 1000–110 1100–125 1378 1469 1639 Migrasol 1177 1349 1408 1456 1564 1616 1668 Rapidoprint 1000–110 1100–125 1350–155 800–1650 EBRA dye 157, 1444 1000–110 1100–120 1300–140 1400–165 tary inform 1504, 1449 the remain rings. No obvio the EBRA d evidenced u normangan of EBRA dy remains br MnCl2 treat of the prese illustrated from both i with a sligh redissolved absorption concentrati that results 3.3. Charac Eventua icals has be species, th described a ticles. The a manganese from the re in Fig. 7. Fo of EBRA dye centration, 0 ,2 ,4 ,6 ,8 1 ,2 ,4 1,6 1,8 2 200 V–vis ight ov ed wi A d so rb ed a m o u n t (m g ) 0 20151050 Equilibrium concentration (mg/L) dsorption isotherms of EBRA dye onto hydrous manganese oxide at pH 8.3 C–N stretching N–H deformation NH2 deformation C O stretching S–O stretching 0 SO3 symmetrical stretching 0 SO3 asymmetrical stretching CH3 deformation CH2 deformation O–H deformation C–N stretching C–H deformation N–H deformation CH2 deformation NH2 deformation C C stretching C O stretching 0 SO3 symmetrical stretching 0 SO3 asymmetrical stretching 0 Ar–NO2 Ø meta substituted , 1410 Triazine 0 Aromatic rings stretching 0 Ar–F 0 NH–Ar 0 Aromatic rings deformation ation on the molecule structure. Indeed, peaks at 1560, and 1410 cm−1 can be assigned to a triazine structure, 0 0 0 0 1 1 ab so rb an ce ( A ) Fig. 6. U with a sl coagulat pH 7. Fig. 7. A ing bands corresponding to a substitution of aromatic us interaction between the hydroxide precipitates and ye pigment or the dyeing auxiliary chemicals could be sing infrared spectroscopy. Actually, neither Mg(OH)2 ese hydrous oxide precipitation should alter the nature e. In the case of MgCl2 addition, the sludge obtained ight blue, whereas the sediments generated by the ment process of SW1 and SW2 turn greenish because nce of strongly colored manganese minerals. Indeed, as in Fig. 6, almost identical UV–vis spectra are obtained nitial SW1 solution at pH 7 and sediments coagulated t overdose in MgCl2 or MnCl2 (0.32mol/l) and then by acidification at neutral pH. The slight change in around 600nm is attributed to a small difference in dye on between the original and the redispersed sediments from the acidification. teristics of the precipitation/adsorption mechanism lly, though the treatment with MgCl2 and MnCl2 chem- en attributed hitherto to a coagulation with hydrolyzed e actual dye removal mechanism might better be s a sorption phenomenon onto newly precipitated par- dsorption isotherms of EBRA molecules onto hydrous oxide and magnesium hydroxide precipitates obtained spective optimal coagulant concentrations, are shown r magnesium hydroxide surfaces, the adsorbed amount pigment increases steadily with the equilibrium con- whereas it shows two well-defined steps for hydrous (�) and brucit manganese precipitate. a larger ad observed on with the be ment. Such a p posed by Le of MgCl2 on dyeing auxi ing process and the ele appearance the treatme an inhibitio EBRA mole readily evid precipitated MnCl2 conc SW1 solutio removal cu tions when thus implyi removal effi the adsorp of the prec hence incre capacity. In 300 400 500 600 700 800 wavelenght(nm) (c) (b) (a) spectra of (a) EBRA dye solution acidified at pH 7; (b) SW1 coagulated erdose inMgCl2 and then redissolvedbyacidificationat pH7; (c) SW1 th a slight overdose in MnCl2 and then redissolved by acidification at 10 12 14 MnCl2 MgCl2 2 4 6 8 e at pH 9.4 (©). oxide particles thus confirming the dual nature of that In addition, at any given equilibrium concentration, sorbed amount of EBRA dye pigment is consistently manganese-based sorbents, which can be then related tter removal efficiency obtained with the MnCl2 treat- recipitation/adsorption mechanism was initially pro- entvaar and Rebhun [30] when investigating the action domestic sewage. However, both EBRA molecule and liary chemicals necessarily interfere with the hydrolyz- es of MgCl2 and MnCl2. Indeed, the TEM micrographs ctron diffraction patterns reveal a drastic change in the and the cristallinity of precipitates generated during nt of SW1 and SW2 effluents. This can be attributed to n of hydroxide precipitate growth in the presence of cule and/or dyeing auxiliaries. The latter effect can be enced by comparing EBRA dye removal onto already Mg(OH)2 or MnOOH particles at various MgCl2 or entrations, with the original jar-tests carried out with ns. As shown in Fig. 8, both the turbidity and the color rves are shifted towards higher coagulant concentra- the dye is added after the formation of precipitates, ng a poorer elimination of the contaminant. The higher ciency observed during jar-tests can be explained by tion of EBRA dye pigment to the active growth sites ipitate, thus limiting the size of particles formed, and asing both its specific surface area and its adsorption the case of SW2 effluents, the presence of a larger 270 A.Z. Bouyakoub et al. / Journal of Hazardous Materials 187 (2011) 264–273 80 90 100 Coagulation Adsorption a 70 80 90 100 Coagulation Adsorption c 30 40 50 60 70 occ 30 40 50 60 70 occ 0 10 20 0 0,1 0,2 0,3 0,4 0,5 R es id u al T u rb id it y ( N T U ) R es id u al T u rb id it y ( N T U ) occ 0 10 20 0 0,1 0,2 0,3 0,4 0,5 occ 70 80 90 100 l (% ) b occ occ 70 80 90 100 al ( % ) d occ occ 30 40 50 60 C o lo r re m o v a Coagulation occ 20 30 40 50 60 C o lo r re m o v a occ 0 10 20 0,50,40,30,20,10 Coagulation Adsorption 0 10 20 0,50,40,30,20,10 Coagulation Adsorption MnCl2 concentration (mol/L) MgCl2 concentration (mol/L) Fig. 8. Comparison of EBRA dye removal by coagulation (�) and adsorption onto brucite or hydrous manganese oxide precipitated at various concentrations (©). (a) Residual turbidity vs MnCl2 concentration; (b) color removal vs MnCl2 concentration; (c) residual turbidity vs MgCl2 concentration; (d) color removal vs MgCl2 concentration. amount of (Fig. 4b–d). 3.4. Behavi In accor suggested a centration alkalinity so (8.3) determ treatment [ should carr tion, and th arge mea ty, fo and decr ggre ersed e in a rovi al do cs br Fig. 9. Evoluti floc mass fract additives further inhibits the formation of particles or of aggregated suspensions under agitation dance with the precipitation/adsorption mechanism bove, the pHs reached at the optimal coagulant con- are (i) close to the pH of Mg(OH)2 formation in high lutions [37] and (ii) similar to the point of zero charge ined for hydrousmanganese oxide in the case ofMnCl2 38]. Therefore, the surfaces of precipitated particles to a ch tion in intensi MnCl2 slowly those a redisp decras stage p materi the flo y no net charge at the optimal treatment concentra- e aggregates thus obtained should behave according agitator spe form rapidl on of average floc size as a function of time under consecutive cyclic step changes in agita al dimension. (a) SW1 treated with MnCl2; (b) SW2 treated with MnCl2; (c) SW1 treated neutralisation mechanism [45]. Fig. 9 shows the varia- n aggregate size under cyclic step changes in agitation r SW1 and SW2 treated with optimal concentrations of MgCl2, respectively. In all cases, the size of aggregates eases at the initial constant stirrer speed of 100 rpm. As gates were prepared from the coagulated suspensions in their respective synthetic solutions, the observed ggregate size can certainly be attributed to that dilution ded that a dynamic exchange between the aggregated es exist [31]. Upon an increase in stirring (300 rpm), eak up first rapidly and then more slowly. When the ed is reset to its initial value (100 rpm), theaggregate re- y but to a size about half smaller than that obtained just tion intensity. The insets show examples of the determination of the with MgCl2; (d) SW2 treated with MgCl2. A.Z. Bouyakoub et al. / Journal of Hazardous Materials 187 (2011) 264–273 271 0 10 20 30 40 50 60 70 80 90 100 0 0,1 0,2 0,3 0,4 0,5 S et tl ed v o lu m e (m L ) Coagulation Agitation at 300 rpm Centrifugation 0 10 20 30 40 50 60 70 80 90 100 0 0,1 0,2 0,3 0,4 0,5 S et tl ed v o lu m e (m L ) Coagulation Agitationat at 300 rpm Centrifugation 0 10 20 30 40 50 60 70 80 90 100 0 0,1 0,2 0,3 0,4 0,5 R es id u al t u rb id it y ( N T U ) Coagulation Agitation at 300 rpm occ a 0 10 20 30 40 50 60 70 80 90 100 0 0,1 0,2 0,3 0,4 0,5 R es id u al t u rb id it y ( N T U ) Coagulation Agitation at 300 rpm occ c occ b occ d MnCl2 concentration (M/l) Fig. 10. Effect of further agitation at 300 rpm and centrifugation at 2860× g on residual turbidity and sediment volume. (a and b) SW1 treated with MnCl2; (c and d) SW2 treated with MnCl2. before initiating thefirst increase in shearing. Further cyclingof agi- tation conditions between high and low levels of intensity, reveals that floc break-up andfloc re-aggregation becomenearly reversible for aggrega consistent w [31]. In con only partial nature of b like intergrown agglomerates that may not rebuild once broken. The insets of Fig. 9 show typical log–log plots of aggregate vol- ume fraction vs mean floc size obtained during the consecutive step term ip to nt to ynth c Fig. 11. Effect treated with M tes generated with a MnCl2 treatment. Such behavior is ith a charge-neutralisation destabilisation mechanism trast, the aggregates obtained after a MgCl2 application ly reforms after shearing. This may be attributed to the rucite precipitate which, as illustrated in Fig. 4c, looks cyclic size de tionsh sufficie of the s 90 100 Coagulation 90 100 ) a 0 10 20 30 40 50 60 70 80 0 0,1 0,2 0,3 0,4 0,5 R es id u al t u rb id it y ( N T U ) Agitation at 300 rpm 0 10 20 30 40 50 60 70 80 90 100 0 0,1 0,2 0,3 0,4 0,5 S et tl ed v o lu m e (m L ) b occ 0 10 20 30 40 50 60 70 80 0 0 R es id u al t u rb id it y ( N T U occ 0 10 20 30 40 50 60 70 80 90 100 0 S et tl ed v o lu m e (m L ) Coagulation Agitation at 300 rpm Centrifugation d MgCl2 concentration (M/l) of further agitation at 300 rpm and centrifugation at 2860× g on residual turbidity and gCl2. changes in stirring. Obviously, the range of mean floc inedhere limits the applicability of thepower–law rela- barelyonedecade in length,which cannotbe considered define a true fractal dimension. In the particular case etic effluent SW2 treated with the MgCl2 solution, such Coagulation Agitation at 300 rpm ,1 0,2 0,3 0,4 0,5 occ 0,1 0,2 0,3 0,4 0,5 Coagulation Agitation at 300 rpm Centrifugation occ sediment volume. (a and b) SW1 treated with MgCl2; (c and d) SW2 272 A.Z. Bouyakoub et al. / Journal of Hazardous Materials 187 (2011) 264–273 exponent cannot even be determined (Fig. 9d). Nevertheless, the Df values thus obtained are consistent with classical aggregation mechanisms reported in the literature. Thus, the Df value close to 1.7 calculated for aggregates resulting from the application of MgCl2 to th cluster agg nent slightl b) could be the latter c likely influe the aggrega At first a changes in has occurre ual turbidit increased a treated with to more com increased s [40]. In our length scale suggest tha amount of MgCl2. Inde diminish th 4. Conclud The resu anisms invo for removin removal me pigment on iliary chem particles, an demand for Previous nomically v recovered a to the sea, water is us MnCl2 for t sent a prom and sludge with MgCl2 by atomic 0.37ppm fo values com ent disposa wastes can hydrous m color remov Acknowled A.Z.B. an cation and S References [1] I. Petrinic´ reactive d (2007) 51 [2] M.M.Kari try in Ba Anal. Chim [3] D.J. Joo, W.S. Shin, J.H. Choi, S.J. Choi, M.C. Kim, M.H. Han, T.W. Ha, Y.H. Kim, Decolorization of reactive dyes using inorganic coagulants and synthetic poly- mer, Dyes Pigments 73 (1) (2007) 59–64. [4] I. Koyuncu, Influence of dyes salts and auxiliary chemicals on the nanofiltra- tion of reactive dye baths: experimental observations and model verification, alinat . Ahm byco . J. 13 apic, syn cess, D . Dos ogies techn ygur, hlorot rs Co rslan, genou dy, Dy Muru dye w . Vlys mical ter. 70 . Körb stewa ation 07) 42 Rajku e 19 i h iden . Sudar mical thodo ’Neill treatm , Wat . Yu-L wast 95) 55 oush orptio olph ctive 2) (20 apic, stewa hi, G. perfo (1–2 . Gao, stewa Golob thod f 97. . Tan, magne Allègr antati (200 . Lee, perfo oval Al-Bas tic wa 1–15 rter, A ng a p llègre ing ef Judkin m carb eentva gulati haign trans floc e Wang opez- pared t. 35 ( Hem, ngane e SW1 effluent, may be ascribed to a diffusion-limited regation mechanism [39]. Likewise, the scaling expo- y above 2 found with MnCl2 treatments (Fig. 9a and attributed a reaction-limited aggregation. However, in ase, the aspect ratio of the hydrous MnOOH particles nces the arrangement of elementary particles within tes. pproximation, the constant Df values during the cyclic stirring suggest that no major aggregate restructuring d. However, as shown in Figs. 10 and 11, both resid- y and sediment volume decreased significantly after gitation at 300 rpm except in the case of SW2 effluent MgCl2 (Fig. 11c and d). A rearrangement of aggregates pact fractal structures when the flocs are exposed to hear has already been demonstrated in the literature case, such restructuring certainly occurs at a smaller than that investigated by the laser sizer. Figs. 10 and 11 t such phenomenon can be exploited to decrease the sediment generated by the application of MnCl2 or ed, simple centrifugation at 2860× g allows to further e sludge volume after treatment. ing remarks lts presented in this paper shed new light on the mech- lved when either MnCl2 or MgCl2 solutions are used g dye pigment from textile effluents. In both cases, the chanism can be attributed to the adsorption of the dye to the forming hydroxyde precipitate.Most dyeing aux- icals compete with the reactive dye for the precipitated d hence significantly increase the required coagulant optimal dye elimination. literature has shown that the use of MgCl2 is eco- iable [23], especially since magnesium can be easily nd reused [29]. In textile manufacturing plants close further cost reduction can certainly be achieved if sea ed as a source of magnesium ion. The application of he treatment of textile dye wastes appears to repre- ising alternative approach since both optimal dosage production are significantly lower than those achieved . The residual manganese concentrations measured absorption spectroscopy are less than 0.01ppm and r the treatment of SW1 and SW2, respectively. Such ply with the environmental standards on liquid efflu- l in surface waters. The application of MnCl2 to textile also certainly be optimized to preferentially yield the anganese oxide mineral that is the most efficient for al. gement d S.K. would like to thank the Ministry of Higher Edu- cientific Research of Algeria for financial support. , P.R.A. Andersen, S. Sˇostar-Turk, A.M. Le Marechal, The removal of ye printing compounds using nanofiltration, Dyes Pigments 74 (3) 2–518. m,A.K.Das, S.H. Lee, Treatmentof coloredeffluentof the textile indus- ngladesh using zinc chloride treated indigenous activated carbons, . Acta 576 (1) (2006) 37–42. Des [5] A.L tion Eng [6] S. P from pro [7] A.B nol bio [8] A. U dic Dye [9] I. A ero stu [10] M. azo [11] A.G che Ma [12] B.K wa miz (20 [13] D. Blu wit 1–7 [14] G. che me [15] C. O bio dye [16] Y.R dye (19 [17] A. B ads [18] R. D Rea (1– [19] S. P wa [20] B. S the 143 [21] B.Y wa [22] V. me 93– [23] B.H by [24] C. dec 116 [25] J.W the rem [26] N. the 156 [27] J. Po usi [28] C. A dye [29] J.F. siu [30] J. L coa [31] V. C and into [32] J.A. E. L pre Let [33] J.D. ma ion 154 (1) (2003) 79–88. ad, S.W. Puasa, Reactive dyes decolourization from an aqueous solu- mbinedcoagulation/micellar-enhancedultrafiltrationprocess, Chem. 2 (1-3) (2007) 257–265. N. Koprivanac, A.L. Bozic, A. Metes, Removal of some reactive dyes thetic wastewater by combined Al(III) coagulation/carbon adsorption yes Pigments 62 (2004) 291–298. Santos, F.C.J. Cervantes, J.B. van Lier, Review paper on current tech- for decolourisation of textile wastewaters: perspectives for anaerobic ology, Bioresour. Technol. 98 (12) (2007) 2369–2385. E. Kok, Decolorisation treatments of azo dye wastewaters including riazinyl reactive groups by using advanced oxidation method, J. Soc. lour 115 (1999) 350–354. I.A. Balcioglu, Degradation of commercial reactive dyestuffs by het- s and homogenous advanced oxidation processes: a comparative es Pigments 43 (1999) 95–108. ganandham, M. Swaminathan, Photochemical oxidation of reactive ith UV–H2O2 process, Dyes Pigments 62 (2004) 269–275. sides, M. Loizidou, P.K. Karlis, A.A. Zorpas, D. Papaioannou, Electro- oxidationof a textile dyewastewater using aPt/Ti electrode, J. Hazard. (1999) 41–52. ahti, A. Tanyolac, Electrochemical treatment of simulated textile terwith industrial components and Levafix Blue CA reactive dye: opti- through response surface methodology, J. Hazard. Mater. 151 (2–3) 2–431. mar, B.J. Song, J.G. Kim, Electrochemical degradation of Reactive n chloride medium for the treatment of textile dyeing wastewater tification of intermediate compounds, Dyes Pigments 72 (1) (2007) janto, B. Keller-Lehmann, J. Keller, Optimization of integrated –biological degradation of a reactive azo dye using response surface logy, J. Hazard. Mater. 138 (2006) 160–168. , F.R. Hawkes, D.L. Hawkes, S. Esteves, S.J. Wilcox, Anaerobic–aerobic ent of simulated textile effluent containing varied ratios of starch azo er Res. 34 (2000) 2355–2361. i, A. Thomas, Color difference measurement and color removal from ewaters using different adsorbents, J. Chem. Technol. Biotechnol. 63 –59. er, X. Shen, R.G.J. Edyvean, Removal of coloured organic matter by n onto low-cost waste materials, Water Res. 31 (1997) 2084–2092. en, N. Sakkayawong, P. Thiravetyan, W. Nakbanpote, Adsorption of Red 141 from wastewater onto modified chitin, J. Hazard. Mater. 145 07) 250–255. N. Koprivanac, A. Loncaric Bozic, Removal of reactive dyes from ter using Fe(III) coagulant, J. Soc. Dyers Colour 116 (2000) 352–358. Li, D. Wang, C. Feng, H. Tang, Removal of direct dyes by coagulation: rmance of preformed polymeric aluminum species, J. Hazard. Mater. ) (2007) 567–574. Q.Y. Yue, Y. Wang, W.Z. Zhou, Color removal from dye-containing ter bymagnesiumchloride, J. Environ.Manage. 82 (2) (2007) 167–172. , A. Vinder, M. Simonicˇ, Efficiency of the coagulation/flocculation or the treatment of dyebath effluents, Dyes Pigments 67 (2) (2005) T.T. Teng, A.K. Mohd Omar, Removal of dyes and industrial dye wastes sium chloride, Water Res. 34 (2) (2000) 597–601. e, M. Maisseu, F. Charbit, P. Moulin, Coagulation–flocculation– on of dye house effluents: concentrated effluents, J. Hazard. Mater. 4) 57–64. S.P. Choi, R. Thiruvenkatachari, W.G. Shim, H. Moon, Evaluation of rmance of adsorption and coagulation processes for the maximum of reactive dyes, Dyes Pigments 69 (3) (2006) 196–203. taki, Removal of methyl orange dye and Na2SO4 salt from syn- ste water using reverse osmosis, Chem. Eng. Process. 43 (12) (2004) 67. .C. Gomes, The rejection of anionic dyes and salt fromwater solutions olypropylene microfilter, Desalination 128 (2000) 81–90. , P. Moulin, M. Maisseu, F. Charbit, Treatment and reuse of reactive fluents, J. Membr. Sci. 269 (1–2) (2006) 15–34. s, J.S. Hornsby, Color removal from textile dye waste using magne- onate, J. Water Pollut. Control Fed. 50 (11) (1978) 2446–2456. ar, M. Rebhun, Effect of magnesium and calcium precipitation on on–flocculation with lime, Water Res. 16 (1982) 655–662. on, B.S. Lartiges, A. El Samrani, C.Mustin, Evolution of size distribution fer of mineral particles between flocs in activated sludges: an insight xchange dynamics, Water Res. 36 (2002) 676–684. , O. Novaro, X. Bokhimi, T. Lopez, R. Gomez, J. Navarette, M.E. Llanos, Salinas, Characterizations of the thermal decomposition of brucite by sol–gel technique for synthesis of nanocrystalline MgO, Mater. 1998) 317–323. C.J. Lind, Nonequilibrium models for predicting forms of precipitated se oxides, Geochim. Cosmochim. Acta 47 (1983) 2037–2046. A.Z. Bouyakoub et al. / Journal of Hazardous Materials 187 (2011) 264–273 273 [34] R.L. Frost, J.T. Kloprogge, Infrared emission spectroscopic of brucite, Spec- trochim. Acta, Part A 55 (1999) 2195–2205. [35] T. Kohler, T. Armbruster, E. Libowitzky, Hydrogen bonding and Jahn–Teller dis- tortion in Groutite �-MnOOH and Manganite �-MnOOH, and their relations to the manganese dioxides Ramsdellite and Pyrolusite, J. Solid State Chem. 133 (1997) 486–500. [36] C.M. Julien, M. Massot, C. Poinsignon, Lattice vibrations of manganese oxides Part I. Periodic structures, Spectrochim. Acta, Part A 60 (2004) 689–700. [37] E.L. Berg, C.A. Brummer, W.T. Williams, Single stage lime clarification, Water Waste Eng. 7 (1970) 3–42. [38] M. Ramstedt, B.M. Andersson, A. Shchukarev, S. Sjöberg, Surface properties of hydrous manganite (�-MnOOH). A potentiometric, electroacoustic, and X-ray photoelectron spectroscopy study, Langmuir 20 (2004) 8224–8229. [39] P Meakin, Fractal aggregates, Adv. Colloid Interface Sci. 28 (1998) 249–331. [40] P.T. Spicer, S.E. Pratsinis, J. Raper, R. Amal, G. Bushell, G.Meesters, Effect of shear schedule on particle size, density, and structure during flocculation in stirred tanks, Powder Technol. 97 (1998) 26–34. [41] A.Z. Bouyakoub, S. Kacha, B.S. Lartiges, S. Bellebia, Z. Derriche, Treatment of reactive dye solutions by physicochemical combined process, Desalinat. Water Treat. 12 (1–3) (2009) 202–209. [42] C.J. Pouchert, The Aldrich Library of FTIR Spectra, 1st ed., Aldrich Chemical Company, Inc., 1985. [43] N.B. Colthup, L.H. Daly, S.E.Wibertey, Introduction to Infrared andRaman Spec- troscopy, 2nd ed., Academic Press, 1975, pp. 376–381. [44] C.N.R. Rao, Chemical Application of Infrared Spectroscopy, Academic Press, Inc., 1963, pp. 324–325. [45] W. Stumm, J.J. Morgan, Aquatic Chemistry, Chemical Equilibria and Rates in Natural Waters, 3rd ed., John Wiley & Sons, Inc., New York, 1996, pp. 349–424. MnCl2 and MgCl2 for the removal of reactive dye Levafix Brilliant Blue EBRA from synthetic textile wastewaters: An adsorpt... Introduction Experimental Chemicals Preparation of synthetic waters Aggregation procedure and supernatant characterization Aggregate size measurements Sediment characterization Adsorption isotherms Results and discussion Jar test results Fate of dye and nature of coagulating species Characteristics of the precipitation/adsorption mechanism Behavior of aggregated suspensions under agitation Concluding remarks Acknowledgement References