See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/222814191
Electrocoagulation/electroflotation in an external-loop airlift reactor - Application to the decolorization of textile dye wastewater: A case study Article in Chemical Engineering and Processing · August 2008 DOI: 10.1016/j.cep.2007.03.013
CITATIONS
READS
79
183
6 authors, including: Abdel hafid Essadki
Ch. Vial
École de Technologie Supérieure
Université Blaise Pascal - Clermont-Ferrand II
17 PUBLICATIONS 402 CITATIONS
86 PUBLICATIONS 1,809 CITATIONS
SEE PROFILE
SEE PROFILE
Mohammed Azzi
Henri Delmas
Université Sidi Mohamed Ben Abdellah
Institut National Polytechnique de Toulouse
29 PUBLICATIONS 853 CITATIONS
204 PUBLICATIONS 4,082 CITATIONS
SEE PROFILE
SEE PROFILE
All content following this page was uploaded by Abdel hafid Essadki on 16 December 2016.
The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document and are linked to publications on ResearchGate, letting you access and read them immediately.
Available online at www.sciencedirect.com
Chemical Engineering and Processing 47 (2008) 1211–1223
Electrocoagulation/electroflotation in an external-loop airlift reactor—Application to the decolorization of textile dye wastewater: A case study A.H. Essadki a,∗ , M. Bennajah a , B. Gourich a , Ch. Vial b , M. Azzi c , H. Delmas d a
Ecole Sup´erieure de Technologie de Casablanca, BP 8012, Oasis Casablanca, Morocco Laboratoire de G´enie Chimique et Biochimique, LGCB-UBP/ENSCCF, 24 avenue des Landais, BP 206, 63174 Aubi`ere Cedex, France c Facult´ e des Sciences A¨ın Chock, Laboratoire d’Electrochimie et Chimie de l’Environnement, BP 5366 Maarif, Casablanca, Morocco d Laboratoire de G´ enie Chimique, ENSIACET-INPT, 5 rue Paulin Talabot, 31106 Toulouse, France
b
Received 20 August 2006; received in revised form 28 March 2007; accepted 28 March 2007 Available online 19 April 2007
Abstract A 20 L external-loop airlift reactor was used as an electrochemical cell in order to carry out water depollution using batch electrocoagulation (EC) without mechanical agitation, pumping requirements or air injection. Mixing and complete flotation of the pollutants were achieved using only the overall liquid recirculation induced by H2 microbubbles generated by water electrolysis. A red dye from the Moroccan textile industry was used in a case study to validate this innovative application of airlift reactors. Experimental results showed that the axial position of the Al electrodes and the residence time in the separator section were the key parameters to achieve good mixing conditions, to avoid bubbles/particles recirculation in the downcomer and to prevent floc break-up/erosion by hydrodynamic shear forces. Such optimum conditions corresponded to an optimum liquid overall recirculation velocity that was correlated to current, electrode position and dispersion height. Operation time required to achieve 80% COD and 80% color removal efficiencies was modeled as a function of current density. Similarly, specific energy and electrode consumptions were correlated to current, electrode gap and conductivity, which provided the necessary tools for scale-up and process optimization. Operation time and removal efficiencies were similar to those reported in conventional EC cells, but specific energy and electrode consumptions were even smaller without the need for mechanical agitation, pumping requirements and air injection, which could not be achieved in other kinds of conventional gas–liquid contacting devices. © 2007 Elsevier B.V. All rights reserved. Keywords: Electrocoagulation; Electroflotation; External-loop airlift reactor; COD removal; Decolorization
1. Introduction Wastewater from dyeing and finishing processes in the textile manufacturing industry constitute a substantial source of pollution which exhibits intense color, high chemical oxygen demand, fluctuating pH and suspended particles. Indeed, the textile industry utilizes about 10,000 pigments or dyes, but most of them are toxic substances to human and aquatic life [1] and it has been reported that up to 15% of the dyes used are released into wastewaters [2]. These must therefore be treated before final discharge to achieve legal and aesthetic
∗
Corresponding author. Tel.: +212 60878671; fax: +212 22252245. E-mail address:
[email protected] (A.H. Essadki).
0255-2701/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.cep.2007.03.013
standards. This is a particularly critical problem in Morocco where the textile industry is highly developed. Conventional methods for removing dyes from industrial wastewater consist mainly of biological and physicochemical treatments and their various combinations [3–7]. Biological treatments are cheaper than other methods, but dye toxicity usually inhibits bacterial growth and limits therefore the efficiency of the decolorization [7]. Physicochemical methods include adsorption (e.g. on active carbon), coagulation–flocculation (using inorganic salts or polymers), chemical oxidation (chlorination, ozonisation, etc.) and photodegradation (UV/H2 O2 , UV/TiO2 , etc.) [8–19]. However, these technologies usually need additional chemicals which sometimes produce a secondary pollution and a huge volume of sludge [2,20,21]. Water treatments based on the electrocoagulation technique have been recently proved to circumvent
1212
A.H. Essadki et al. / Chemical Engineering and Processing 47 (2008) 1211–1223
most of these problems, while being also economically attractive [1,2,21–25]. Electrocoagulation (EC) is an electrochemical method for treating polluted water which has been successfully applied for treatment of soluble or colloidal pollutants, such as wastewater containing heavy metals, emulsions, suspensions, etc., but also drinking water for lead or fluoride removal [26–28]. In EC, coagulants are delivered in situ using the corrosion of sacrificial anodes when a DC voltage is applied. Simultaneously, electrolytic gases (typically H2 ) are generated at the cathode. Aluminum and iron materials are usually used as anodes, the dissolution of which produces hydroxides, oxyhydroxides and polymeric hydroxides. These are usually more effective coagulants than those used in chemical dosing: they are able to destabilize colloidal suspensions and emulsions, to adsorb, neutralize or precipitate dissolved polluting species, and finally to form flocs that can be removed either by settling/filtration or flotation. In EC, settling is the most common option, while flotation may be achieved by H2 (electroflotation) or assisted by air injection [28]. A typical EC unit includes therefore an EC cell/reactor, a separator for settling or flotation, and often a filtration step. EC was seen as a promising technology in the 19th century, but had nearly disappeared by the 1930s. However, EC has been experiencing a renaissance in the 1990s [28]. Indeed, the benefits of EC include simplicity, efficiency, environmental compatibility, safety, selectivity, flexibility and cost effectiveness [21–31]. In particular, the main points involve the reduction of sludge generation [1], the minimization of the addition of chemicals and little space requirements due to shorter residence time [1,20–25], especially when EC is compared to biological treatments. Although EC may be as cost-effective as chemical dosing [21,32], its main deficiency is the lack of dominant reactor design and modeling procedures. Mollah et al. [27] described six typical configurations for industrial EC cells and report their respective advantages and drawbacks. However, the literature reveals any systematic approach for these configurations for design and scale-up purpose. This situation stems mainly from the complex interactions between electrochemistry, colloidal forces and hydrodynamics that govern the behavior of EC reactors. While the recent literature focuses usually on the two first aspects (such as electrode design [26,27], electrode material [20–22], etc.), the hydrodynamics of the three-phase gas–liquid–solid flow is still disregarded, the presence of gas being usually considered as an unnecessary complication [33]. This stems mainly from the fact that most papers use laboratory-scale EC cells in which magnetic stirring is adjusted experimentally and the separation step by floatation/sedimentation is not studied. The objective of this work is to demonstrate that airlift reactors can be suitable EC units. Airlift reactors constitute a particular class of bubble columns in which the difference in gas hold-up between two sections (namely, the riser and the downcomer) induces an overall liquid circulation without mechanical agitation [34,35]. They have been extensively applied in the process industry to carry out chemical and biochemical slow reactions, such as chemical oxidation using O2 , Cl2 or aerobic fermentation, but never as EC cells, as far as the authors
know. Airlift reactors present two main designs: external-loop and internal-loop configurations. External-loop airlift reactors offer the advantage to allow various designs of the separator section, which favors gas disengagement at the top of the reactor and maximizes consequently the overall recirculation velocity at the expense of more complex reactor geometries. Their hydrodynamics has also been extensively studied in two-phase gas–liquid [36] and three-phase gas–liquid–solid flows [37,38]. In this study, the aim is to demonstrate that such reactors can be used as an EC cell in which complete flotation can be achieved without air injection and good mixing conditions can be obtained without mechanical agitation. This means that pollutants should be floated to the surface only by tiny bubbles of hydrogen and oxygen gases generated from water electrolysis; similarly, the overall liquid circulation responsible for mixing in airlift reactors should result only from electrochemically-generated bubbles. In this way, the removal of a commercial red dye used in a Moroccan factory by electrocoagulation/electroflotation (EC/EF) has been investigated to validate this innovative application of external-loop airlift reactors. 2. Materials and methods 2.1. Reactor design An external-loop airlift made of transparent plexiglas was used for this study. The reactor geometry is illustrated by Fig. 1. By definition, the riser is the section in which the gas phase is sparged and flows upwards. The diameters of the riser and the downcomer are respectively 94 mm and 50 mm in this work. Consequently, the riser-to-downcomer cross-sectional ratio (Ar /Ad ) is about 3.5. This is a typical value when reaction takes place only in the riser section. Both are 147 cm height (H2 + H3 ) and are connected at the bottom by a junction of 50 mm diameter and at the top by a rectangular gas separator (also denoted gas disengagement section) of HS = 20 cm height. The distance between the vertical axes of the riser and the downcomer is 675 mm, which limits the recirculation of bubbles/particles from the riser into the downcomer. At the bottom, the curvature radius of the two elbows is 12.5 cm in order to minimize friction and avoid any dead zone. The liquid volume depends on the clear liquid height (h) and can be varied between 14 L and 20 L, which corresponds to a clear liquid level between 2 cm and 14 cm in the separator section. All the experiments were conducted at room temperature (20 ± 1 ◦ C) and atmospheric pressure in the semi-batch mode (reactor open to the gas, closed to the liquid phase). Contrary to conventional operation in airlift reactors, no gas phase was sparged at the bottom in the riser. Only electrolytic gases induced the overall gas recirculation resulting from the density difference between the fluids in the riser and the downcomer. Two readily available aluminum flat electrodes of rectangular shape (250 mm × 70 mm × 1 mm) were used as the anode and the cathode, which corresponds to S = 175 cm2 electrode surface area (Fig. 1). The distance between electrodes was e = 20 mm, which is a typical value in EC cells (see, e.g., [21,33]). They were treated with a HCl aqueous solution for cleaning prior use
A.H. Essadki et al. / Chemical Engineering and Processing 47 (2008) 1211–1223
1213
Fig. 1. External-loop airlift reactor (1: downcomer section; 2: riser section; 3: conductivity probes; 4: conductimeter; 5: analog output/input terminal panel (UEIAC-1585-1); 6: 50-way ribbon cable kit; 7: data acquisition system; 8: electrodes; 9: separator; 10: electrochemically-generated bubbles).
to avoid passivation. The electrodes were placed in the riser, parallel to the main flow direction to minimize pressure drop in the riser and maximize the recirculation velocity. The axial position of the electrode could also be varied in the column. The distance (H1 ) between the bottom of the electrodes and the bottom of the riser ranged between 7 cm and 77 cm (Fig. 1). EC was conducted in the intensiostat mode, using a digital DC power supply (Didalab, France) and recording potential during the experiments. The width of the electrodes was maximized by taking into account riser diameter and electrode inter-distance. Electrode length L should be maximized to enhance mixing of dissolved Al and favor the formation of H2 microbubbles, but the condition L < H2 /4 was retained to avoid limiting the possible H1 values. Current density values (j) between 5.1 and 51 mA/cm2
were investigated, which corresponds to current (I = jS) in the range 1.0–10 A. 2.2. Chemicals and methods The average liquid velocity in the downcomer (ULd ) was measured using the conductivity tracer technique [39]. Two conductivity probes placed in the downcomer section were used to record the tracer concentration resulting from the injection of 5 mL of a saturated NaCl solution at the top of the downcomer using a data acquisition system based on a PC computer equipped with RTI-815 A/D converter. The distance between the probes was 90 cm (Fig. 1). Liquid velocity was estimated using the ratio of the mean transit time between the tracer peaks
1214
A.H. Essadki et al. / Chemical Engineering and Processing 47 (2008) 1211–1223
Fig. 4. UV–vis absorption spectrum of the red dye. Fig. 2. Example of experimental data from conductivity tracer experiments in the downcomer section when current density j = 21.4 mA/cm2 .
detected successively by the two electrodes and the distance between the probes. The superficial liquid velocity in the riser (ULr ) was deduced from a mass balance on the liquid phase: ULr =
Ad ULd Ar
(1)
An example of experimental data provided by the conductivity tracer technique is reported in Fig. 2. Experiments were carried out using a red dye solution consisting of a mixture of 2-naphthoic acid and 2-naphtol (Fig. 3) with a total concentration C0 = 20 mg/L. This mixture belongs to the dispersive dye class [20]. This class groups organic, non-ionic compounds nearly insoluble although they are applied in aqueous solution using simple immersion technique [20]. This dye and its concentration are typical of wastewater from a Moroccan textile factory. Synthetic solutions were prepared by dissolving the dye in tap water. Solution conductivity and pH were measured using a CD810 conductimeter (Radiometer Analytical, France) and a ProfilLine pH197i pHmeter (WTW, Germany).
Fig. 3. Molecular structure of the constituents of the red dye.
Dye concentration was estimated from its absorbance characteristics (Fig. 4) in the UV–vis range at maximum wavelength A450 (λmax = 450 nm) using a UV–vis spectrophotometer (Pye Unicam, SP8-400, UK). Chemical oxygen demand (COD) was measured using the standard closed reflux colorimetric method. Initial COD was about 2500 mg O2 /L. Initial pH was varied between 5 and 10 using minute addition of 0.1 M H2 SO4 or NaOH solutions [1,5,20,22]. The conductivity κ (i.e. the ionic strength) of dye solutions was adjusted by sodium chloride addition in the range 1.0–29 mS/cm, which covers the range usually explored in the literature (see, e.g., [21,31,33]). NaCl is a salt exhibiting low toxicity at moderated level, reasonable cost, high conductivity and high solubility, but it played also the role of supporting electrolyte. This addition had a negligible effect on the initial pH of the solutions. pH, absorbance and COD were measured over time on filtered samples recovered from two locations in the downcomer section in order to check reactor homogeneity. Experimental values reported in this work correspond to the average of these two samples. COD, color removal and turbidity efficiencies (YCOD , YCOL and YABS ) were expressed as percentage and defined as: YCOD =
COD(t = 0) − COD(t) COD(t = 0)
(2a)
YCOL =
A450 (t = 0) − A450 (t) A450 (t = 0)
(2b)
YCOL and YABS were obtained using the same equation (Eq. (2b)), but YCOL was based on absorbance measurements at 450 nm after filtration, while YABS was measured without filtration/decantation. YABS was used to analyze qualitatively the evolution of turbidity over time. This parameter shows whether the flocs flotate or are destroyed and driven by the liquid flow. The specific electrical energy consumption per kg dye removed (Edye ) and the specific electrode consumption per kg dye (μAl ) were calculated as follows: kWh UIt Edye (3) = kg dye 1000V (C0 YCOL ) kg Al 3600MAl ItφAl 1 (4) = μAl kg dye 3F V (C0 YCOL )
A.H. Essadki et al. / Chemical Engineering and Processing 47 (2008) 1211–1223
using initial dye concentration C0 (kg/m3 ), current intensity I (A), cell voltage U (V), electrolysis time t (h), liquid volume V (m3 ), molar weight of aluminum MAl = 0.02698 kg/mol, Faraday’s constant F (96,487 C/mol e− ) and the Faradic yield φAl of Al dissolution. φAl was estimated as the ratio of the weight loss of the aluminum electrodes during the experiments mexp and the amount of aluminum consumed theoretically at the anode mth : φAl =
mexp 3F = mexp mth 3600MAl It
(5)
This parameter depends upon the pH and the amount of other species present in solution, for example co-existing anions [40]. 3. Results and discussion 3.1. Airlift reactors as EC cells In airlift reactors, the driving force of the overall liquid circulation results from the gas hold-up difference between the riser (εr ) and the downcomer (εd ), and also from the dispersion height. Gas hold-up is defined as the ratio of volume occupied by the gas phase over the total volume of the corresponding section. Dispersion height (hD ) corresponds to the distance from the surface in which a gas phase can be observed in the riser (Fig. 1). Conversely, wall friction effects oppose to the overall liquid circulation. They are usually taken into account by two empirical coefficients KT (that takes the effects of pressure drop in the riser and the separator section into account) and KB (that accounts for pressure drop in the downcomer and the junction). The overall liquid circulation velocity in the riser ULr can therefore be predicted from an energy balance using the following Equation [34,35]: 0.5 2ghD (εr − εd ) (6) ULr = KT /(1 − εr )2 + (Ar /Ad )2 KB /(1 − εd )2 When the clear liquid height h is high enough to achieve total gas disengagement in the separator (this is often the case in external-loop airlift reactors when recirculation velocity is not too high), εd remains negligible. In this work, the situation is however more complex because the gas phase is not injected, but electrochemically-generated. This means that both hD and εr depend on the axial position of the electrodes in the riser. At constant current density, εr should vary approximately as dispersion height hD when electrode position H1 is modified, provided εr 1. One can deduce from Eqs. (1) and (6) that ULd ∼ hD . At constant H1 , the evolution of dispersion height due to gas flow should be negligible when εr 1. On the basis of empirical correlations from the literature [34–36], εr should α with α between 0.5 vary with superficial gas velocity UGr as UGr and 0.6 in external-loop airlift reactors. As gas flow rate should be proportional to current (provided the evolution with I of the Faradic yield for H2 formation remains negligible), this gives ULd ∼ I␣/2 . As the distance between electrodes is e = 20 mm and the riser diameter 94 mm, it is obvious that a fraction of the liquid phase
1215
does not circulate between the electrodes. As a result, it is compulsory to reduce the circulation time (i.e. the time necessary for the liquid to accomplish one loop) and achieve perfect mixing as quickly as possible. This may be obtained easily by increasing ULd . Conversely, the objective of complete flotation may be achieved only if hydrodynamic shear forces remain weak in the riser to avoid floc break-up and in the separator to limit break-up and erosion, which means low ULd values. A simple solution would consist in increasing e, but this would reduce electrode surface S at constant L. Additionally, electrical energy supply is known to increase steeply with electrode distance (see, e.g., [41]). At constant electrode geometry, the solution consists of an adequate selection of ULd resulting from a compromise between mixing and floc stability. This can be obtained first by optimizing the axial position of the electrodes using both the conductivity tracer technique for ULd estimation and turbidity measurements to estimate the amount of dispersed Al particles in the downcomer. Experimental results showed that no liquid overall circulation could be detected when the electrodes were placed in the upper part of riser, for H1 approximately higher than 60 cm (Fig. 1). For 7 cm < H1 < 60 cm, ULd decreased when H1 increased, as expected from Eq. (6). Overall liquid velocities in the downcomer for two axial positions are reported in Fig. 5 at various current densities. Using the fact that ULd ∼ hD I␣/2 , a two-parameter model was used to fit the data using the Levenberg–Marquardt method for parameter optimization. The results are given in the following equation: hD ULd (cm/s) = 5.8 j 0.20 (7) hD max in which hDmax is the maximum dispersion height corresponding to H1 = 7 cm. A good agreement was obtained between experiments and predictions (Fig. 5). This equation confirms the key role of the axial position of the electrodes on reactor hydrodynamics and mixing properties, and the weaker influence of current. The fitted parameter α = 0.40 was a bit lower than expected from theoretical considerations, but this
Fig. 5. Influence of the axial position of the electrodes (H1 ) and current density (j) on the overall liquid recirculation ULd (h = 14 cm; initial pH 8.3; initial conductivity: κ = 2.4 mS/cm).
1216
A.H. Essadki et al. / Chemical Engineering and Processing 47 (2008) 1211–1223
result is however satisfactory when one considers the following points: • The range of gas flow rates used to establish correlations from the literature was widely different from that of this work; • The mechanisms of gas dispersion, the physicochemical properties of the gas phase (H2 in this work) and bubble size differed; • The Faradic yield of H2 generation was not taken into account; • Other geometrical parameters, such as the distance between riser and downcomer, were not directly accounted for. Consequently, Eq. (7) is of utmost importance for scale-up purpose at constant aspect ratio (Ar /Ad ) and L/H2 ratio. Nevertheless, if Ar /Ad had to be varied, one can benefit from the available know-how on external-loop airlift reactors to account for this effect [36]. The corresponding turbidity data based on A450 without filtration is reported in Fig. 6 for an electrolysis time of 10 min and two axial positions of the electrodes. In all cases, flocs occupied nearly 1 cm thickness at the free surface of the disengagement section and no settling was reported in the junction and in the separator. Fig. 6 shows however that turbidity rose when H1 = 7 cm for current density higher than 15 mA/cm2 . This corresponded to ULd values in the range 9–10 cm/s in Fig. 5. Conversely, complete flotation was always observed for H1 = 47 cm, as A450 remained low, about 0.06. As a result, electrode position must be chosen in order to maintain ULd always lower than 9 cm/s, regardless of current density in the range of j studied (Fig. 7). Such a condition was achieved for H1 = 47 cm, as shown in Fig. 5, which corresponds nearly to mid-height in the riser. However, H1 = 47 cm is only a coarse approximation of the optimum electrode height, but the simple way to optimize mixing conditions does not consist in adjusting precisely H1 because it is easier from a practical point of view to adjust the clear liquid height h. Indeed, h affects simultaneously hD and hDmax in Eq. (7), but its range (2 cm < h < 14 cm in this work) is usually far smaller than H1 (between 7 and 77 cm). Its expected impact
on ULd is therefore limited to ±7% on the basis of Eq. (7) for H1 = 47 cm, which imposes a preliminary determination of the optimum H1 range. Additionally, h plays an important role on gas disengagement and solid flotation. When h was equal to 2 cm, the sludge occupied almost the total volume of the disengagement section, allowing the flocs to be driven by the liquid flow. As far as h was increased, the flotation layer thickness at the surface occupied a decreasing volume fraction of the disengagement section. As ULd varies slightly, the residence time of the liquid phase in the separator (τ) should increase steeply with h. This should reduce both the direct entrainment of solid particles from the riser into the downcomer, but also the erosion of flotation layer by the overall liquid circulation in the separator. Fig. 8 highlights the strong influence of the clear liquid height on the evolution of turbidity at constant j and H1 . Experimental data was plotted as a function of τ that is nearly proportional to h. Initial A450 value was 0.25; in Fig. 8, absorbance after 10 min was divided by three when τ increased from 20 s to 55 s. In this work, it is difficult to quantify the respective effects of h and τ, as τ depends also on the distance between the riser and the down-
Fig. 6. Evolution of absorbance A450 of non-filtered samples after 10 min operation with 1 cm of floc thickness already formed as a function of the axial position of the electrodes (H1 ) and current density j (initial pH 8.3, conductivity: κ = 2.4 mS/cm).
Fig. 8. Influence of the residence time of the liquid phase in the separator τ on turbidity A450 after 10 min operation (initial pH 8.3; current density: j = 28.6 mA/cm2 ; conductivity: κ = 2.4 mS/cm).
Fig. 7. Evolution of absorbance A450 of non-filtered samples after 10 min operation with 1 cm of floc thickness already formed as a function of superficial liquid velocity ULd in the downcomer (initial pH 8.3, conductivity: κ = 2.4 mS/cm).
A.H. Essadki et al. / Chemical Engineering and Processing 47 (2008) 1211–1223
1217
comer. One can however assume that when h is not too small, the distance between riser and downcomer reduces mainly the direct recirculation of the solid phase, while h acts more effectively on floc erosion. As a conclusion, experimental data showed that an overall liquid circulation with good mixing conditions and complete flotation could be achieved using only the electrochemicallygenerated bubbles, provided: • The electrodes were adequately placed in riser: in this work, this means that ULr should be close, but lower than 2.5 cm/s in order to avoid floc break-up by the hydrodynamic shear forces in the riser; • The clear liquid height was carefully selected in the separator in order to minimize solid recirculation in the downcomer and floc erosion by the liquid phase, while maximizing solid entrainment by the bubbles and the formation of a stable flotation layer at the free surface of the reactor. The second point must take two constraints into account: h is limited both by the separator height (HS ) and by the increase of ULr that should remain lower than 2.5 cm/s. This justifies the choice of an external-loop airlift reactor because its geometry permits large distances between riser and downcomer in order to favor complete flotation, which is not possible in internal-loop airlift reactors. 3.2. Application to decolorization of textile dye wastewater The objective was to achieve A450 in the exhaust water lower than 0.07. This gave a condition on color removal, but also on turbidity removal. Additionally, 80% COD abatement was requested to achieve the Moroccan limits for COD in textile wastewater. Using the optimized position of the electrode defined in Section 3.1 (H1 = 47 cm and h = 14 cm), the respective influences of the operating conditions of EC (such as current density operation time. . .) and of the physicochemical properties of wastewater before treatment (such as initial pH and conductivity) were investigated in order to achieve the objectives by complete flotation. 3.2.1. Influence of initial pH pH is known to play a key role on the performance of EC [2,21–25,27]. In this work, the influence of the initial pH on COD and turbidity removals is illustrated in Fig. 9 at constant current density and initial conductivity. An optimum was found for the initial pH, which was between 7.0 and 8.0, although it differed slightly between COD and color removal yields. However, the pH changed during batch EC, as already mentioned in the above-mentioned papers. Its evolution depended on the initial pH. EC process exhibits some buffering capacity because of the balance between the production and the consumption of OH− [30], which prevents high change in pH (Fig. 10). The buffering pH seems just above 7: when the initial pH is above this value, pH decreases during EC; otherwise, the opposite behavior is observed.
Fig. 9. Influence of the initial pH on COD removal and decolorization after 8 min operation (conductivity: κ = 2.4 mS/cm, current density: j = 28.5 mA/cm2 ).
The effect of pH can be explained as follows. The main reactions are during EC are: Anode : Cathode :
Al0(s) → Al3+ + 3e− 2H2 O + 2e− → H2(g) + 2OH−
(8) (9)
At low pH, such as 2–3, cationic monomeric species Al3+ and Al(OH)2 + predominate. When pH is between 4–9, the Al3+ and OH− ions generated by the electrodes react to form various monomeric species such as Al(OH)2 + , Al2 (OH)2 2+ , and polymeric species such as Al6 (OH)15 3+ , Al7 (OH)17 4+ , Al13 (OH)34 5+ that finally transform into insoluble amorphous Al(OH)3(s) through complex polymerization/precipitation kinetics [23]. When pH is higher than 10, the monomeric Al(OH)4 − anion concentration increases at the expense of Al(OH)3(s) . In addition, the cathode may be chemically attacked by OH− ions generated together with H2 at high pH values [1]: 2Al + 6H2 O + 2OH− → 2Al(OH)4 − + 3H2
(10)
Two main mechanisms are generally considered: precipitation for pH lower than 4 and adsorption for higher pH. Adsorption may proceed on Al(OH)3 or on the monomeric Al(OH)4 −
Fig. 10. Evolution of pH values during EC for different values of initial pH (conductivity: κ = 2.4 mS/cm, current density: j = 28.6 mA/cm2 ).
1218
A.H. Essadki et al. / Chemical Engineering and Processing 47 (2008) 1211–1223
anion depending on the dye chemical structure. The formation of Al(OH)3(s) is therefore optimal in the 4–9 pH range, which corresponds to the optimum pH values investigated in this work. Consequently, one can conclude that the key mechanism for dye removal is adsorption in this work. However, pH affects also bubble size [43]. Typical bubble sizes in electroflotation always fall in the range of 20–70 m [44], far smaller than those observed in conventional airassisted flotation, which provides both sufficient surface area for gas–liquid–solid interfaces and mixing efficiency to favor the aggregation of tiny destabilized particles. Hydrogen bubbles, which obey usually to a lognormal size distribution, are known to be the smallest about neutral pH [45]. The difference between maximum pH for COD and turbidity confirms this result and can be attributed to the formation of more stable flocs when pH is about 7, although it does not correspond to the optimum for dye removal. As a conclusion, pH adjustment is compulsory before EC in airlift reactors, as in conventional EC cells. pH may be adjusted in the optimum range in order to achieve a compromise between best coagulation and best flotation. The optimum range may however vary as a function of electrode material and dye structure. In the following sections, initial pH will be fixed at about 8 to maximize COD removal efficiency. 3.2.2. Influence of current density and operation time The current density is expected to exhibit a strong effect on EC [21–31], especially on the kinetics of COD abatement and color removal: higher the current, shorter the treatment. This is ascribed to the fact that at high current density, the extent of anodic dissolution of aluminum increases, resulting in a greater amount of precipitate for the removal of pollutants. Moreover, bubble generation rate increases and the bubble size decreases with increasing current density. These effects are both beneficial for high pollutant removal by H2 flotation [25]. Experimental results show that COD removal efficiencies increase faster when current density increases (Fig. 11a). Dye removal seems therefore more rapid, as expected. The curves reach however the same asymptotic value as a function of time, around 83%, regardless of current density. As a result, the objective can be achieved on the whole range of current density, but the minimum operation time tN for which YCOD > 80% decreases from 36 min to 9 min when current increases from 1 A to 6 A. Color removal curves follow exactly the same trends, which was expected because the dyes constitute the only source of pollution in this work (data not reported). Color removal proceeds during EC because the constituents responsible for the red color (2-naphthoic acid and 2-naphtol) are adsorbed. For turbidity (Fig. 11b), the behavior is identical to YCOL , except when current is 6 A. In this case, when EC is carried out for a too long time, YABS decreases probably because of the too rapid generation of Al(OH)3 particles that seem to have difficulties to flocculate and flotate. As a result, there is a maximum time for optimum batch EC operation that is a function of current density. As a first approximation, the amount of Al released is proportional to the product φAl It. The Faradic yield φAl is reported in Fig. 12 as a function of time and current density: the values are
Fig. 11. Evolution of COD (a) and turbidity removal efficiency (b) during EC as a function of current density j (initial pH 8.3, conductivity: κ = 2.4 mS/cm).
between 100% and 160%; they decrease with increasing time in the first minutes of the run, but also with higher current density. These trends have already been reported in the literature [46]. This mass overconsumption of aluminum electrodes may be due to the chemical hydrolysis of the cathode (Eq. (10)), but it also often explained by the “corrosion pitting” phenomenon which causes holes on the electrode surface. The mechanism suggested for “corrosion pitting” involves chloride anions and
Fig. 12. Evolution of the Faradic yield φAl as a function of current density j (initial pH 8.3, conductivity: κ = 2.4 mS/cm).
A.H. Essadki et al. / Chemical Engineering and Processing 47 (2008) 1211–1223
1219
can be summarized as follows [33]: 2Al + 6HCl → 2AlCl3 + 3H2
(11a)
AlCl3 + 3H2 O → Al(OH)3 + 3HCl
(11b)
This mechanism can therefore produce more aluminum hydroxide flocs and H2 bubbles than the equivalent current supplied should. However, the time tN necessary to achieve 80% removal for COD and color did not follow a first-order kinetics [33]. The amount of pollution removed was not proportional to the amount of Al released by the electrodes, as tN did not vary as 1/(φAl I). Similarly, taking into account a minimum dissolved Al concentration before EC starts [46] did not fit the data in the range 1–6 A. This may be due to the complexity of dye removal: the dye is formed by two constituents, the second one being an oxidized form of the first one. Consequently, their affinity for flocculation may differ and vary with current density. Additionally, at high current density, other reactions may be encountered at the anode, such as the direct oxidation of one or the two constituents of the dye, but also oxygen formation at the anode, which plays a negative role on EC: 4OH− → O2 + 2H2 O + 2e−
(12)
Conversely, high current density allows the passivation of the cathode to be reduced [26], but an increase in energy consumption that induces heating by Joule effect. As a result, too high current densities have generally to be avoided. From an energetic point of view, energy consumption during EC is known to vary as the product UIt. Energy requirements per kg of dye removed (Edye ) to achieve 80% COD and color removal efficiency are reported in Fig. 13 as a function of current density. Energy requirements per kg COD removed can be estimated approximately by dividing Edye by about 125 (2467 mg/L O2 divided by 20 mg/L dye), as YCOL ≈ YCOD in this work. Edye values in Fig. 13 are similar and even lower than those reported by several authors [24,25]. Fig. 13 shows a continuous increase of Edye with j, which has already been observed in the literature for textile dye wastewater [2,21–25,27]. This means that the decrease in tN does not compensate the impact of current density increase on energy requirements (Edye ∼ I2 ). However, Edye represents only a fraction of the operating costs of EC, from 20% to
Fig. 13. Influence of current density j on electrode and energy consumption (μAl and Edye ) at minimum time tN for which YABS ≥ 80% and YCOD ≥ 80% (initial pH 8.3, conductivity: κ = 2.4 mS/cm).
Fig. 14. Influence of current density j on electrode and energy consumption (μAl and Edye ) at constant operation time of 8 min (initial pH 8.3, conductivity: κ = 2.4 mS/cm).
50% according to Bayramoglu et al. [23]. For these authors, the main cost is generally electrode material. As a result, the specific electrode consumption μAl (kg Al required per kg dye removed) has been also reported in Fig. 13. This parameter is roughly proportional to It and exhibits a maximum for I about 5 A, which is due to the counterbalancing effects of the decrease of tN when current increases. When one considers μAl at a constant time (8 min), regardless of YCOL and YCOD , an abrupt change appears between 1 A (5.7 mA/cm2 ) and higher current densities (Fig. 14) that is probably characteristic of a change in EC mechanisms described above. The fact that μAl (t = 8 min) is nearly constant and independent of j when j is higher than 10 mA/cm2 means that dye removal is nearly proportional to the amount of aluminum released, but only when j > 15 mA/cm2 . The amount of Al released can be estimated as follows [46]: t MAl I 0 φAl dt CAl (t) = (13) 3 · 3600FV For the red dye, the following relation has been established: tN = k
0.80C0 V 2 (R = 0.995) I
(14)
with k = 17.2 × 104 (A min)/(kg dye removed). As a conclusion, the influence of current density on EC in airlift reactors is similar to that observed in conventional EC cells. For the application studied in this work, one can distinguish four key points: • from the point of view of operating costs that account both for specific energy and electrode consumption, the best conditions correspond to a low current and a high operation time, such as I = 1 A and tN = 36 min; • considering investment costs, it may be interesting to limit the size of the equipment and therefore operation time by increasing current density; • considering scale-up procedure, it is preferable to operate at j > 15 mA/cm2 (about 2 A), as a simple model tN ∼1/I may be applied; • to avoid unnecessary Al(OH)3 release in wastewater, it is finally compulsory to avoid too high current density and to operate at j lower than 30 mA/cm2 .
1220
A.H. Essadki et al. / Chemical Engineering and Processing 47 (2008) 1211–1223
3.2.3. Influence of wastewater conductivity The increase of the conductivity by the addition of sodium chloride is known to reduce the cell voltage U at constant current density due to the decrease of the ohmic resistance of wastewater [2,23,25]. Energy consumption, which is proportional to UI, will therefore decrease. Chloride ions could significantly reduce the adverse effects of other anions, such as HCO3 − and SO4 2− , for instance by avoiding the precipitation of calcium carbonate in hard water that could form an insulating layer on the surface of the electrodes and increase the ohmic resistance of the electrochemical cell [2,41]. Chloride anions can also be oxidized and give active chlorine forms, such as hypochlorite anions, that can oxidize dyes. The main mechanism is as follows: Cl2 + 2e− → 2Cl−
(15a)
Cl2 + H2 O → Cl− + ClO− + 2H+
(15b)
3.2.4. Complementary rules for scale-up and process optimization The position of the electrodes in the airlift reactor can be optimized as a function of hydrodynamic parameters and current density using Eq. (7). The necessary time to achieve 80% removal efficiencies can be estimated using Eq. (14), though it is valid only for the red dye used in this work in a limited range of j values. Complementary rules should include the influences of electrode gap and operating conditions on voltage (and consequently on energy consumption). The measured potential is the sum of three contributions, namely the kinetic overpotential, the mass transfer overpotential and the overpotential caused by solution ohmic resistance. Kinetic and mass transfer overpotentials increase with current density, but mass transfer is mainly related to mixing conditions: if mixing is rapid enough, mass transfer overpotential should be negligible. In this case, the model described by Chen et al. [41] is often recommended for non-passivated electrodes [27]:
However, an excessive amount of NaCl (higher than 3 g/L) induces overconsumption of the aluminum electrodes due to “corrosion pitting” described above (Eq. (11)); Al dissolution may become irregular [42]. Additionally, it is important to monitor how the presence of NaCl affects COD and color removal efficiencies. YCOD and YCOL usually increase with the addition of NaCl [2], but this result is not general [23,25]. In this work, Fig. 15 shows an increase of YCOD and YCOL with κ for the red dye between 2 and 28 mS/cm. YCOL enhancement becomes however slight when κ is higher than 15 mS/cm in Fig. 15. The decrease in specific energy consumption Edye due to the increase of conductivity is illustrated by Fig. 16. This figure indicates that Edye can be divided roughly by a factor 13 when conductivity is multiplied by a factor 12. However, high chloride content exceeds usually the maximum concentration tolerated in waters destined for further reuse. As a conclusion, a κ value of 5 mS/cm seems a reasonable compromise because it offers a moderate value of electrical consumption by Joule effect, while preventing a rapid degradation of the electrode surface [1,42].
e U = −0.76 + j + 0.20 ln(j) κ
Fig. 15. Influence of conductivity κ on COD and color removal efficiencies (initial pH 8.3; current density: j = 28.6 mS/cm).
Fig. 16. Influence of conductivity κ on energy consumption Edye at constant current density j and operation time (initial pH 8.3).
(16)
This equation is compared to experimental U values (with e = 0.02 m) in Fig. 17. It exhibits a good agreement with experimental data, both as a function of conductivity and current density. This result shows that the overall liquid circulation induced by hydrogen bubbles in the external-loop airlift reactor achieves good mixing conditions and makes mass transfer overpotential negligible, but without the need for mechanical stirring, which corresponds to additional energy savings. As a result, Eq. (16) seems able to estimate quantitatively the effect of inter-electrode distance. Energy savings and heating by Joule effect could therefore be optimized by decreasing this gap. However, a small gap could hinder transport and suspension of solid particles that are currently found in textile dye wastewater [42] and it could also induce short-circuits [33]. As already mentioned in Section 2, electrode gap must obey a compromise for the regular flow of the multiphase media and acceptable energy dissipation. The value e = 0.02 m corresponds to this compromise and should maintained in a scale-up procedure. This is confirmed by the fact that this value has already been used in
A.H. Essadki et al. / Chemical Engineering and Processing 47 (2008) 1211–1223
1221
adequate selection of the axial position of the electrodes (H1 ) and the liquid height in the separator section (h) in order to avoid floc break-up in the riser and floc erosion at the free surface. A limiting value of the liquid velocity in the downcomer was defined, while ULd was correlated to dispersion height hD and current density j (Eq. (7)). These can be used at constant j and Ar /Ad ratio for scale-up purpose. Batch EC experiments on a red dye from the textile industry, with 80% COD and color removal as the objectives, confirmed that external-loop airlift reactors behaved as a conventional EC cells. However, specific energy and electrode consumptions seemed even lower than in the literature on treatment of textile dye wastewater. As expected, a preliminary pH adjustment was compulsory and the respective influences of current density and conductivity exhibited trends similar to those reported in the literature for conventional EC cells. Simple models for predicting the impact of operation time (Eq. (14)), current density and conductivity (Eq. (16)) were derived in order to estimate voltage and operating costs. and optimize process conditions. As this study was carried out in pilot-scale airlift reactor (20 L), scaleup seems easier on the basis of above-mentioned equations using both the available know-how on EC and airlift reactor design. In the future, the use of airlift reactors for EC under steadystate conditions seems also promising. The main disadvantage of airlift devices is however that their behavior is close to that of a perfectly mixed reactor, which implies larger volumes and longer residence time than in most conventional EC cells [27]. This will however avoid systems with an external recycle [24] because of the internal liquid recirculation of airlift reactors. Appendix A Fig. 17. Validation of theoretical models relating potential, current and conductivity: (a) j = 28.6 mA/cm2 , initial pH 8.3; (b) κ = 2.4 mS/cm, initial pH 8.3.
many other works on the economical evaluation of treatment of textile wastewater [21], oil recovery in O/W emulsions [46] and more generally in industrial electrochemical processes [33] at several scales, both under batch and continuous conditions. 4. Conclusions Electrocoagulation is already known as an efficient process for color removal in dye textile wastewater, but also for the removal of soluble and colloidal pollutants in wastewater and drinking water. In this work, external-loop airlift reactors have been shown to be versatile tools to carry out EC with complete flotation, using only electrochemically generated H2 bubbles to achieve an overall liquid circulation and good mixing conditions. Consequently, the use of mechanical agitation, pumping or compressed air was not necessary. This could not be achieved in other kinds of conventional gas–liquid contacting devices than airlift reactors. External-loop devices are particularly adapted because they offer specific designs for the disengagement section that allow large distance between riser and downcomer. This improves flotation by minimizing the recirculation of Al particles in the downcomer. These results were obtained by the
A450 absorbance measured at 450 nm (AU) cross-sectional area of the downcomer (m2 ) Ad cross-sectional area of the riser (m2 ) Ar C0 dye concentration (kg/m3 ) CAl concentration of dissolved aluminum (kg/m3 ) COD chemical oxygen demand (mg O2 /L) e electrode gap (m) Edye specific energy consumption (kWh/kg dye removed) EC electrocoagulation EF electroflotation F Faraday’s constant (96,487 C/mol e− ) g acceleration of gravity (m/s2 ) h liquid height in the separator section (m) hD dispersion height in Fig. 1 (m) hDmax maximum dispersion height (m) axial position of the electrodes in Fig. 1 (m) H1 H2 , H3 , HS geometrical characteristics of the reactor in Fig. 1 (m) I current (A) j current density (A/m2 ) k constant in Eq. (14) (min A/(min kg dye removed)) KB , KT friction factors in Eq. (6) L electrode length (m) mexp experimental weight loss of Al electrode (kg)
1222
mth MAl R2 S t tN U UGr ULd ULr V YABS YCOD YCOL
A.H. Essadki et al. / Chemical Engineering and Processing 47 (2008) 1211–1223
theoretical weight loss of Al electrode (kg) molar mass of aluminum (0.02698 kg/mol) correlation coefficient electrode surface area (m2 ) time (h) necessary time to achieve the objectives (min) measured potential (V) superficial gas velocity in the riser (cm/s) overall liquid recirculation in the downcomer (cm/s) overall liquid recirculation in the riser (cm/s) reactor volume (m3 ) turbidity removal efficiency (%) COD removal efficiency (%) color removal efficiency (%)
Greek letters εd gas hold-up in the downcomer εr gas hold-up in the riser κ wastewater conductivity (S/m) λmax maximum wavelength of the absorption spectrum (nm) specific consumption of the anode (kg Al/kg dye μAl removed) τ liquid residence time in the separator (s) φAl Faradic yield of Al dissolution References [1] A. Alinsafi, M. Khemis, M.N. Pons, J.P. Leclerc, A. Yaacoubi, A. Benhammou, A. Nejmeddine, Electro-coagulation of reactive textile dyes and textile wastewater, Chem. Eng. Process. 44 (2005) 461– 470. [2] N. Daneshvar, A. Oladegaragoze, N. Djafarzadeh, Decolorization of basic dye solutions by electrocoagulation: an investigation of the effect of operational parameters, J. Hazard. Mater. 129 (2006) 116–122. [3] R.W. Peters, T.J. Walker, J.E. Eriksen, T.K. Cheng, Y. Ku, W.M. Lee, Wastewater treatment-physical and chemical methods, J. Water Pollut. Control Fed. 57 (1985) 503–517. [4] J.S. Do, M.L. Chen, Decolorization of dye containing solutions by electrocoagulation, J. Applied Electrochem. 24 (1994) 785–790. [5] J.Q. Jiang, J.D. Graham, Enhanced coagulation using Al/Fe(III) coagulants: effect of coagulant chemistry on the removal of color-causing NOM, Environ. Technol. 17 (1996) 937–950. [6] Y.M. Slokar, A.M. Le Marechal, Methods of decoloration of textile wastewaters, Dyes Pigments 37 (1998) 335–356. [7] A.J. Greaves, D.A.S. Phillips, J.A. Taylor, Correlation between the bioelimination of anionic dyes by an activated sewage sludge with molecular structure. Part 1: Literature review, JSDC 115 (1999) 363–365. [8] P. Thebault, J.M. Cases, F. Fiessinger, Mechanism underlying the removal of organic micropollutants during flocculation by an aluminium or iron salt, Water Res. 15 (1981) 183–189. [9] F. Gahr, F. Hermanutz, W. Oppermann, Ozonation—an important technique to comply with German laws for textile wastewater treatment, Water Sci. Technol. 30 (1994) 255–263. [10] J.H. Churchly, Removal of dyewaste color from sewage effluent—the use of full-scale ozone plant, Water Sci. Technol. 30 (1994) 275–284. [11] S.F. Kang, H.M. Chang, Coagulation of textile secondary effluents with Fenton’s reagent, Water Res. 36 (1997) 215–222. [12] C. Hachem, F. Bocquillon, O. Zahraa, M. Bouchy, Decolourization of textile industry wastewater by the photocatalytic degradation process, Dyes Pigments 49 (2001) 117–125. [13] W. Chu, S.M. Tsui, Modeling of photodecoloration of azo dye in a cocktail photolysis system, Water Res. 36 (2002) 3350–3358.
[14] B. Neppolian, S. Sakthivel, B. Arabindoo, V. Murugesan, Solar/UV induced photocatalytic degradation of three commercial textile dyes, J. Hazard. Mater. 89 (2002) 303–317. [15] A. Pala, E. Tokat, Color removal from cotton textile industry wastewater in an activated sludge system with various additives, Water Res. 36 (2002) 2920–2925. [16] M. P´erez, F. Torrades, X. Dom´enech, J. Peral, Fenton and photo-Fenton oxidation of textile effluents, Water Res. 36 (2002) 2703–2710. [17] T. Robinson, B. Chandran, P. Nigam, Removal of dyes from a synthetic textile dye effluent by biosorption on apple pomace and wheat straw, Water Res. 36 (2002) 2824–2830. [18] M.V.B. Zanoni, J. Sene, M.A. Anderson, Photoelectrocatalytic degradation of Remazol Brilliant Orange 3R on titanium dioxide thin-film electrodes, J. Photochem. Photobiol. A: Chem. 157 (2003) 55–63. [19] B. Zielinska, J. Grzechuslka, A.W. Morawski, Photocatalytic decomposition of textile dyes on TiO2 Tytanpol A11 and TiO2 -Degussa P25, J. Photochem. Photobiol. A: Chem. 157 (2003) 65–70. [20] T.-H. Kim, C. Park, E.-B. Shin, S. Kim, Decolorization of disperse and reactive dyes by continuous electrocoagulation process, Desalination 150 (2002) 165–175. [21] M. Bayramoglu, M. Eyvaz, M. Kobya, Treatment of the textile wastewater by electrocoagulation: economical evaluation, Chem. Eng. J. 128 (2007) 155–161. [22] M. Kobya, O.T. Can, M. Bayramoglu, Treatment of textile wastewaters by electrocoagulation using iron and aluminum electrodes, J. Hazard. Mater. 100 (2003) 163–178. [23] M. Bayramoglu, M. Kobya, O.T. Can, M. Sozbir, Operating cost analysis of electrocoagulation of textile dye wastewater, Sep. Purif. Technol. 37 (2004) 117–125. [24] O.T. Can, M. Kobya, E. Demirbas, M. Bayramoglu, Treatment of the textile wastewater by combined electrocoagulation, Chemosphere 62 (2006) 181–187. [25] M. Kobya, E. Demirbas, O.T. Can, M. Bayramoglu, Treatment of levafix orange textile dye solution by electrocoagulation, J. Hazard. Mater. 132 (2006) 183–188. [26] M.Y.A. Mollah, R. Schennach, J.R. Parga, D.L. Cocke, Electrocoagulation (EC)—science and applications, J. Hazard. Mater. 84 (2001) 29–41. [27] M.Y.A. Mollah, P. Morkovsky, J.A.G. Gomes, M. Kesmez, J.R. Parga, D.L. Cocke, Fundamentals, present and future perspectives of electrocoagulation, J. Hazard. Mater. 114 (2004) 199–210. [28] P.K. Holt, G.W. Barton, C.A. Mitchell, The future for electrocoagulation as a localised water treatment technology, Chemosphere 59 (2005) 355–367. [29] K. Rajeshwar, J.G. Ibanez, G. Swain, Electrochemistry and the environment, J. Appl. Electrochem. 24 (1994) 1077–1091. [30] G. Chen, Electrochemical technologies in wastewater treatment, Sep. Purif. Technol. 38 (2004) 11–41. [31] M.Y.A. Mollah, S.R. Pathak, P.K. Patil, M. Vayuvegula, T.S. Agrawal, J.A.G. Gomes, M. Kesmez, D.L. Cocke, Treatment of orange II azo-dye by electrocoagulation (EC) technique in a continuous flow cell using sacrificial iron electrodes, J. Hazard. Mater. 109 (2004) 165–171. [32] P.K. Holt, G.W. Barton, M. Wark, C.A. Mitchell, A quantitative comparison between chemical dosing and electrocoagulation, Colloid Surf. A: Physicochem. Eng. Aspects 211 (2002) 233–248. [33] N. Mameri, A.R. Yeddou, H. Lounici, D. Belhocine, H. Grib, B. Bariou, Defluoridation of septentrional Sahara water of North Africa by electrocoagulation process using bipolar aluminium electrodes, Water Res. 32 (1998) 1604–1612. [34] Y. Chisti, M. Moo-Young, Airlift reactors: applications and design considerations, Chem. Eng. Sci. 43 (1987) 451–457. [35] Y. Chisti, Airlift Bioreactors, Elsevier, London, 1989. [36] J.B. Joshi, V.V. Ranade, S.D. Gharat, S.S. Lele, Sparged loop reactors, Can. J. Chem. Eng. 68 (1990) 705–741. [37] A. Livingston, S.F. Zhang, Hydrodynamic behaviour of three-phase (gas–liquid–solid) airlift reactors, Chem. Eng. Sci. 48 (1993) 1641– 1654. [38] C. Freitas, M. Fialov´a, J. Zahradnik, A. Texeira, Hydrodynamic model for three-phase internal- and external-loop airlift reactor, Chem. Eng. Sci. 54 (1999) 5253–5258.
A.H. Essadki et al. / Chemical Engineering and Processing 47 (2008) 1211–1223 [39] C. Vial, E. Camarasa, S. Poncin, G. Wild, N. Midoux, J. Bouillard, Study of hydrodynamic behaviour in bubble columns and external loop airlift reactors through analysis of pressure fluctuations, Chem. Eng. Sci. 55 (2000) 2957–2973. [40] C.Y. Hu, S.L. Lo, W.H. Kuan, Effects of co-existing anions on fluoride removal in electrocoagulation (EC) process using aluminum electrodes, Water Res. 37 (2003) 4513–4523. [41] X. Chen, G. Chen, P.L. Yue, Investigation on the electrolysis voltage of electrocoagulation, Chem. Eng. Sci. 57 (2002) 2449–2455. [42] L. S´anchez Calvo, J.-P. Leclerc, G. Tanguy, M.-C. Cames, G. Paternotte, G. Valentin, A. Rostan, F. Lapicque, An electrocoagulation unit for the purification of soluble oil wastes of high COD, Environ. Progress 22 (2003) 57–65.
View publication stats
1223
[43] V.A. Glembotskii, A.A. Mamakov, A.M. Ramanov, V.E. Nenno, Proceedings of the 11th International Mineral Processing Congress, Cagliari, Italy, 1975, pp. 562–581. [44] N. Adhoum, L. Monser, N. Bellakhal, J.-E. Belgaied, Treatment of electroplating wastewater containing Cu2+ , Zn2+ and Cr(VI) by electrocoagulation, J. Hazard. Mater. 112 (2004) 207–213. [45] O.J. Murphy, S. Srinivasen, B.E. Conway, Electrochemistry in Transition: from the 20th to the 21st Century, Plenum, New York, 1992; Y. Fukui, S. Yuu, Removal of colloidal particles in electroflotation, AIChE J. 31 (1985) 201–208. [46] M. Khemis, J.-P. Leclerc, G. Tanguy, G. Valentin, F. Lapicque, Treatment of industrial liquid wastes by electrocoagulation: experimental investigations and an overall interpretation model, Chem. Eng. Sci. 61 (2006) 3602–3609.
