Evaluation of Ferric Chloride as a Coagulant for Cork Processing

Ind. Eng. Chem. Res. 2005, 44, 6539-6548

6539

Evaluation of Ferric Chloride as a Coagulant for Cork Processing Wastewaters. Influence of the Operating Conditions on the Removal of Organic Matter and Settleability Parameters Joaquı´n R. Domı´nguez,* Jesu ´ s Beltra´ n de Heredia, Teresa Gonza´ lez, and F. Sanchez-Lavado Chemical Engineering Department, Extremadura University (UEx), Avda, de Elvas S/N, E-06071 Badajoz, Spain

This is a two-part work on the chemistry of iron(III) as coagulant in the treatment of cork processing wastewater. The main aim of the first part was to determine the removal of organic matter as measured by reductions in chemical oxygen demand (COD), polyphenols (P), and aromatic compounds (A) that can be obtained using this physicochemical process. To this end, jar-test experiments were used to determine the optimal conditions for the process, in particular, the effective iron dosage, temperature, contamination level of the wastewater, coagulant mixing time, stirring speed, and pH. The ranges of parameters tested for the coagulation process were coagulant dose (200-1500 mg/L), contamination of the wastewater (water I, COD 800 mg/L; water II, COD 1100 mg/L; water III, COD 2000 mg/L; water IV, COD 4000 mg/L), mixing time (5-30 min), stirring speed (60-300 rpm), temperature (20-80 °C), and pH (4-11). The resulting removal capacities were in the ranges 35-65% for COD, 55-90% for polyphenols, and 40-90% for aromatics. The best results were obtained with the shortest coagulant mixing time, 5 min, a stirring speed of 300 rpm, and the lowest temperature studied, 20 °C, although the difference at higher temperatures was only slight. The optimal choices of pH and coagulant dose fundamentally depended on the contamination level of the wastewater. In the second part of the work, the main aim was to examine the influence of the operating conditions on the system’s settleability parameters. It is well-known that it is just as important to achieve good settleability parameters in the physicochemical treatment of wastewaters as it is to attain a high level of decontamination. These parameters will determine the dimensions of the required equipment and hence the costs of the installation. This second part of the study therefore analyzes the influence of the different operating variables on the following settleability parameters: sediment volumetric percentage, settling velocity, sludge volume index, and residual conductivity in the clarified water. The optimal conditions found for the settling process were not the same as those that had been determined in the first part of the study for the elimination of organic matter. Finally, taking into account economic and practical reasons, the Talmadge-Fitch method is used to apply the results to the design of a clarifier-thickener unit to treat 2 m3/h of wastewater. The required minimum area of the unit would be 4.01 m2. 1. Introduction One of the classical problems of water treatment is that the smallness of the colloidal particles of organic matter present in the water and the negative charges distributed over their surfaces makes colloidal suspensions that are highly stable. Indeed, the smaller the particles, the more difficult it becomes to separate them by natural sedimentation. One therefore has to apply technological processes that promote this settling to improve the characteristics of the water that will be passed on to the decantation stage. Coagulation is a commonly used process in water treatment in which compounds such as iron salts are added to effluents to destabilize the colloidal material and cause the small particles to agglomerate into larger, settleable flocs. How effective this process will be depends on the coagulating agent used, the dosage, the solution pH and ionic strength, and the concentration and nature of the organic compounds present in the water.1 In aqueous * To whom correspondence should be addressed. E-mail: [email protected]. Fax number: +34 924289385.

solution, when ferric salts dissolve, the metallic ion (Fe3+) hydrates and is hydrolyzed to form monomeric, [Fe(H2O)6]3+, [Fe(H2O)5(OH)]2+ (pK1 ) 2.2), [Fe(H2O)4(OH)2]+ (pK2 )3.5),Fe(OH)3(s)(pK3 )6),and[Fe(H2O)2(OH)4](pK4 ) 10), and polymeric, [Fe2(H2O)8(OH)2]4+ or [Fe2(H2O)7(OH)3]3+, species.2,3 Under very acidic conditions (pH < 2), [Fe(H2O)6]3+ remains in solution, but as the pH or the coagulant concentration rises, hydrolysis occurs to form ferric hydroxide, Fe(OH)3(s). In general, the hydrolysis reaction of trivalent iron is as follows:2

xFe3+ + yH2O f Fex(OH)y(3x-y) + yH+

(1)

The resulting metal hydroxide polymers have amorphous structures with very large surface areas and positive charges.1 They are hydrophobic, causing them to sorb onto the organic anionic particle surfaces and become insoluble.2 Iron has a strong tendency to form insoluble complexes with a number of ligands, especially with polar molecules and with oxygen-containing functional groups such as the hydroxyl or carboxyl groups.4 These provide a local negative charge, which reacts with

10.1021/ie0487641 CCC: $30.25 © 2005 American Chemical Society Published on Web 07/08/2005

6540

Ind. Eng. Chem. Res., Vol. 44, No. 17, 2005

Figure 1. Schematic illustration of the concept of the PCN model.

the iron cations. Charge neutralization leads to colloid destabilization with the consequent precipitation of the iron cations and organic anions. This induces sweepfloc coagulation (also known as sweep flocculation or sweep coagulation), the adsorption and bridging enmeshment of both particulate organic and inorganic solids to form large, amorphous flocs.5 Dissolved organic compounds are removed primarily by adsorption onto the hydroxide surface. 1.1. Classical Model. There are various mechanisms that can destabilize colloids: (a) an increase in the concentration of ionic species, thereby destabilizing the colloid particles by compression of the electric bilayer; (b) a reduction of the ζ potential due to adsorption of polynuclear anionic species, iron(III)-hydroxo complexes, onto the colloid surface; (c) sweep-floc coagulation, in which the colloidal particles are swept out of suspension by becoming enmeshed in the ferric hydroxide precipitate. The single most determining factor in the coagulation process is undoubtedly the pH, since it affects all the hydrolysis equilibria resulting from the addition of the metal cation. At low pH and low coagulant dose, the predominant mechanism is charge adsorption-neutralization, and at high pH and high coagulant dose, it is sweep-floc coagulation.6,7 The precipitation rate of the metal hydroxide will depend on the degree of supersaturation (SS) of the solution and can be described by eq 2.3 One can also predict that the greater the degree of supersaturation, the faster the rate of nucleation relative to the floc growth rate, thus yielding a sediment that is also more resistant to compaction.

SS )

[Fe3+][OH-]3 Ks

(2)

Ks is the solubility constant of the metal hydroxide. 1.2. Precipitation Charge Neutralization Model (PCN). This model was introduced by Dentel8 to explain coagulation in water treatment by the hydrolysis of metal salts. A schematic illustration of the processes involved is given in Figure 1. Although the PCN model has been presented in a quantitative form,9 this aspect will not be covered here. According to the PCN model, coagulation with iron salts involves three steps: (a) Destabilization begins after addition of a dose of coagulant that exceeds the operational solubility of iron. (b) Iron hydroxide species are then deposited onto colloidal surfaces, as shown in Figure 1. (c) Under typical conditions, the metal hydroxide is positively charged, while the original colloidal particles are negatively charged. So the deposition process can result in charge neutralization or charge reversal of the colloidal particles.

If the positively charged adsorbed species are in the form of isolated regions, then a form of “electrostatic patch” attraction may be important. It has to be noted that the PCN model does not cover bulk hydroxide precipitation sweep-floc coagulation. There is no doubt that, at the correct dosage, charge neutralization by the adsorbed hydrolysis products and hydroxide precipitate can cause negatively charged particles to become destabilized and hence to coagulate. When charge-neutralization is the predominant destabilization mechanism, then there should be a stoichiometric relationship between the particle concentration and the optimal coagulant dosage.10 At low particle concentrations, only low coagulant dosages should be required. Under such conditions, coagulation rates can be very low, thereby causing problems in water treatment. Another practical difficulty is that the optimal coagulant dosage range can be quite narrow, which means that fairly precise dosing control is needed. These difficulties can be overcome by using higher coagulant dosages, for which extensive hydroxide precipitation occurs, giving rise to sweep-floc coagulation. 1.3. The Phenomenology of Floc Sedimentation. It is well-known that it is just as important to achieve good settleability parameters in the physicochemical treatment of wastewaters as it is to attain a high level of decontamination. These parameters will determine the dimensions of the required equipment and hence the costs of the installation. The second part of the study analyzes the influence of the different operating variables on the following settleability parameters of the system: sediment volumetric percentage, settling velocity, sludge volume index, and residual conductivity in the clarified water. The phenomenon of the sedimentation of particles that have formed or are in the process of forming flocs is different from that of discrete particles. As they descend, flocs merge with each other by adsorption or coalescence. As they increase in size, their settling velocity increases. In a suspension, the manner in which the particles settle depends on the concentration of the suspension and the characteristics of the particles. Fitch11 describes the different types of sedimentation: clarification, flocculent sedimentation, zonal settling and compression or compaction. In the sedimentation of a concentrated suspension of solids, the particles are so close to one another that there is interference between their respective velocity fields. Furthermore, as the liquid is displaced upward, it acts to brake the descent of the particles.12 2. Materials and Methods 2.1. Characterization of the Wastewater. In the industrial processing of cork, when the cork slabs arrive at the plant, those of best quality are parboiled in cookers (calderas in Spanish) for 1-11/2 hours in nearboiling water (temperature around 98 °C). This has the 2-fold purpose of disinfecting the cork (elimination of fungi, insects, reptiles, and tannins and other detrimental substances) and improving its mechanical properties (elasticity, texture, consistency, etc.). The calderas are filled with clean water on the first day of the working week (usually Monday), and 8-10 batches (known as “calderadas”) of raw cork are treated per day. The water is changed twice weekly, so it may used to treat between 15 and 25 calderadas, depending on the requirements and the rate of work.

Ind. Eng. Chem. Res., Vol. 44, No. 17, 2005 6541 Table 1. Main Physicochemical Characteristics of the Cork Process Wastewatera parameter

water I

water II

water III

water IV

pH Fe (mg/L) conductivity (µS/cm) TSS (g/L) turbidity (NTU) COD (mg/L) absorbance at 254 nm of 1/50 diluted simples (ua) aromaticity (mg/L of phenol) polyphenols (mg/L of caffeic acid)

5.45 1.80 477 0.087 79.9 810 0.193

5.10 2.60 568 0.120 115 1106 0.289

5.22 5.20 894 0.250 230 2228 0.537

5.62 8.00 1520 0.430 381 3722 0.922

1877

2809

5233

8969

113

153

380

549

a

All the values are affected by an error of (5%.

Sometimes agro-industrial wastewaters have a constituent that can flocculate as the result of simple agitation or of the addition of a flocculant. At other times, however, it is necessary to add a coagulant that will form a precipitate, which can then be flocculated. Such is the case with cork industry wastewaters, which contain large quantities of suspended and colloidal material and thus need a coagulant to facilitate sedimentation. These waters also vary greatly over the course of the week, depending on the number of calderadas that they have been used to treat. Furthermore, they contain very high organic loads, including chlorinated organic compounds, tannins, and other harmful substances. The wastewaters of the present study were provided by the company Corchos de Me´rida S.A. (Extremadura Autonomous Community, Spain). The samples were taken at different times during the week, corresponding to different numbers of calderadas and hence with different organic matter concentrations. Table 1 lists the nomenclature that will be used for these waters with their principal physicochemical characteristics. 2.2. Materials and Methods. Coagulation studies were performed in a conventional model Velp Scientifica JLT4 jar-test apparatus, equipped with four 1-L beakers (height 17.5 cm, diameter 10 cm) equipped with an external water jacket surrounding the beaker to maintain a constant temperature. The calculated quantity of coagulant solution (FeCl3, 37.8% w/v) was added to the wastewater, and the pH was adjusted to the required value with Ca(OH)2. The pH adjustment process required a time of 1 min. This time is not considered in the stirring time. The mixture was stirred (with one single-palette stirrer, 7.5 cm × 2.5 cm) at the rate and for the time fixed for each experiment. It was then transferred to a 1-L graduated cylinder (height 40 cm, diameter 6.5 cm) for the sedimentation test during a settling period of 1 h. After that test, the samples were taken from the clarified liquid and were assayed for aromatics, polyphenols, and COD. 2.3 Analysis. The aromatic content was determined by measuring the absorbance in the ultraviolet region at 254 nm wavelength, at which aromatic and unsaturated compounds present the maximum absorption.13,14 The results are expressed in terms of absorbance units of 1/50 diluted samples and as a reference aromatic compound,15 milligrams per liter of phenol. The polyphenol content of a water can be determined colorimetrically using the Folin-Ciocalteau reagent (a mixture of molybdophosphoric and tungstophosphoric

acids). The blue complex that is formed is measured at a wavelength of 725 nm.16 The procedure of the present experiments was the following. An aliquot of the sample was transferred to a 50-mL graduated flask containing distilled water. Then 2.5 mL of Folin-Ciocalteau reagent and 5 mL of saturated sodium carbonate solution were added, and the volume was made up to the mark with distilled water. The solution was left to stand for 1 h, and then the absorbance was measured in a 1-cm optical path length cuvette. The results are expressed in milligrams per liter of caffeic acid, since this acid is a very common phenolic compound in a large variety of waters. For the COD determination, we used a Selecta mod. Tembloc oven, a PF-10 Macherey-Nagel spectrophotometer, and test cuvettes pre-prepared for the desired measurement range (the range of concentrations selected was 100-1500 mg/L of O2). The COD determination was carried out using LCK-114 Hach-Lange COD tests. This method follows the ISO 15705:2002 standard method. 2.4. Determination of the Settleability Parameters. Residual conductivity (Kf). The final conductivity of the supernatant after 60 min of settling was determined using a Crison (Spain) model GLP-32 conductimeter equipped with a CAT probe. Sediment Percentage (SP). The sediment percentage is the ratio between the volume beneath the supernatant-suspension interface after 60 min of sedimentation, V60, and the initial wastewater volume, V0, expressed as a percentage. Settling Velocity (vs). The mean settling velocity corresponds to the first 15 min of the settling process and is expressed in centimeters per minute. It is given by the following expression:17

vs )

mh0 V0

(3)

where h0 is the height (cm) of the initial column of wastewater, V0 is the initial wastewater volume (cm3), and m is the slope obtained from a plot of the data volume beneath the interface (cm3) vs time (min). Sludge Volume Index (SVI). The sludge volume index (cm3/g) is the ratio between the volume and the weight of sludge formed after 30 min of settling. It is given by the following expression:17

SVI )

V30 V0TSS

(4)

where V30 is the volume below the supernatantsuspension interface after 30 min of sedimentation (cm3), V0 is the initial wastewater volume expressed in liters, and TSS is the total suspended solids content of the wastewater in grams per liter. Total Suspended Solids (TSS). This parameter is the concentration of total suspended solids of the wastewater after the addition and mixing of the coagulant. It is determined gravimetrically from the solid fraction retained on a 0.45 µm pore-size glass-fiber filter.18 3. Results and Discussion 3.1. Removal of Organic Matter. Compared to alum, ferric chloride coagulates effectively over a broader

6542

Ind. Eng. Chem. Res., Vol. 44, No. 17, 2005

Figure 2. Influence of stirring speed on COD removal for waters I and II. Experimental conditions were pH 7, mixing time 5 min, coagulant dose 400 mg/L of FeCl3, and 20 °C.

pH range, forms a stronger, heavier floc, is less sensitive to problems with filtrate quality in instances of overdosing, and avoids the problem of the aluminum residues in the finished effluent. Hydrolyzing metal salts of iron are widely used as primary coagulants to promote the formation of floc in water treatment. The present experiment studied the use of iron salts to purify highly polluted wastewaters. The parameters monitored were chemical oxygen demand (COD), total polyphenols (P), and aromatic compounds (A). As can be seen in Table 1, waters I-IV had different organic loads. In all the experiments, the volume of water treated was 1 L, and the reaction temperature was 20 °C except for the water III experiments in which different temperatures were used: 20, 40, 60, and 80 °C. 3.1.1. Influence of Stirring Speed. Adequate mixing is necessary both when the coagulant is added and during the formation and growth of the floc. As particle sizes increase, stirring may break up existing flocs as a result of disruptive forces, and the collision efficiency of the particles in a shear field decreases.19,20 A dynamic balance between floc growth and breakage often leads to a steady-state floc-size distribution, where the limiting size depends on the applied shear rate.21 If the effective shear rate is increased, preformed flocs can be broken in a manner that depends on the floc size relative to the turbulence microscale.21 Flocs formed by hydrolyzing coagulants tend to be rather weak, so breakage occurs readily. In the case of sweep-floc coagulation, this breakage is not fully reversible, and flocs do not completely re-form when the original shear conditions are restored. The stirring speed in the present experiments had a moderate influence. Figure 2 shows the effect on COD removal (waters I and II). The optimal speed for COD and aromatic removal was 300 rpm. For the polyphenol removal, however, one observes that there was no optimal rate for water I, but a rate of 60 rpm gave the best results for water II. The removals of COD, aromatics, and polyphenols for water I (COD ≈ 800 mg/L) were far greater than those for water II (COD ≈ 1100 mg/L), indicating that the coagulation was more effective in reducing these three parameters in waters with lower organic matter content. 3.1.2. Influence of Coagulant Mixing Time. Some experiments were carried out on waters I and II to study the influence of the coagulant mixing time. The shortest time (5 min) was best for all three parameters, COD, polyphenols (P), and aromatic compounds (A) (see Figure 3). This therefore indicates that prolonged stirring leads to the rupture of the connections formed between the coagulant and the colloidal particles and the process’s consequent loss of efficiency.21 One also

Figure 3. Influence of coagulant mixing time on the removal of polyphenols (P) and aromatics (A) for waters I and II. Experimental conditions were pH 7, 300 rpm, coagulant dose 400 mg/L of FeCl3, and 20 °C.

Figure 4. Effect of the coagulation pH on COD removal for waters I-III. Experimental conditions were mixing time 5 min, 150 rpm, coagulant dose 400 mg/L of FeCl3, and 20 °C.

Figure 5. Effect of the coagulation pH on the removal of polyphenols for waters I-III. Experimental conditions were mixing time 5 min, 150 rpm, coagulant dose 400 mg/L of FeCl3, and 20 °C.

observes from the figures that, for this 5-min mixing time, the COD, aromaticity, and polyphenol removals in water I were 25%, 5%, and 3% greater, respectively, than those in water II, further confirmation of the greater efficacy of the process for waters with lower organic matter content. 3.1.3. Influence of the Coagulation pH. The coagulation process depends on various physicochemical characteristics of the water, above all on the pH. Indeed, the pH and the coagulant dose are, of all the factors involved, those that most affect the process since they influence all the hydrolysis equilibria that arise. Another reason for the importance of the pH is that the addition of the metallic cation automatically lowers the pH, the decrease being greater the higher the dose of coagulant and the lower the alkalinity of the water. Experiments on waters I-III were carried out to study the effect of modifying the pH. Figures 4-6 show the results for the final COD, polyphenol, and aromatic compound removals, respectively. One observes that in the range of pH studied, there appear to be two values of maximum reduction, this being clearest for water III. The explanation could be the following: at low pH, below the isoelectric point (which in the case of ferric

Ind. Eng. Chem. Res., Vol. 44, No. 17, 2005 6543

Figure 6. Effect of the coagulation pH on the removal of aromatics for waters I-III. Experimental conditions were mixing time 5 min, 150 rpm, coagulant dose 400 mg/L of FeCl3, and 20 °C.

hydroxide is pH 8), the charge adsorption-neutralization mechanism is of considerable importance (the first maximum), while at higher pHs (the second maximum), the sweep-floc coagulation mechanism predominates.6,7 Whichever the case, the net result is that the removal of COD can occur by a combination of the two mechanisms over a wide range of pH (pH 4-11), removals varying in the ranges 56-63% for water I, 46-57% for water II, and 45-53% for water III. This two-peaked curve has also been observed by Lefebvre and Legube22 for total organic carbon (TOC) removal and for the reduction of the absorbance at 254 nm in the coagulation of a solution of GFA (Gartempe fulvic acid) with Fe3+ at 20 °C, and similar results were obtained by Stephenson23 for the removal of color [PtCo] and turbidity by the coagulation of a mechanical pulping effluent with ferric chloride. One observes that as the contamination level of the wastewater increases from water I (COD 800 mg/L) to water III (COD 2000 mg/L), the maxima in the curve shift rightwards. The case is similar for the removal of polyphenols (Figure 5) and of aromatic compounds (Figure 6). For example, in the COD removal, the first maximum occurs at pH 4 for water I, pH 6 for water II, and pH 7 for water III. The second maximum occurs at pH 9 for waters I and II, and pH 11 for water III. One can draw two conclusions from an observation of Figures 4-6: first, although the pHs for the maximum removal of COD, polyphenols, and aromatics coincided in most cases (e.g., water III), this was not always so; and second, the values of the removal of COD, polyphenols, and aromatics were always greater in the less contaminated waters (water I > water II > water III). 3.1.4. Effect of Coagulant Dose. The coagulant dose to apply to a wastewater fundamentally depends on the concentration of organic material in suspension and on the pH. Similarly, the most appropriate interval of pH is closely related to the coagulant dosesthe greater the coagulant dose, the wider the effective pH interval for coagulation. Experiments were carried out on waters III and IV, which were those that had been used for the most “calderadas” and therefore had the highest levels of contamination, to study the influence of the coagulant dose. In these experiments, the dose of FeCl3 was varied between 200 and 1500 mg/L, maintaining the following variables constant: pH 7, stirring speed 150 rpm, and mixing time 15 min. Figures 7 and 8 show the results of the analysis of the supernatant. One observes that increasing the coagulant dose favored the removal of COD, aromaticity, and polyphenols in both types of wastewater. There are two mechanisms by which high coagulant dosages can increase the coagulation rate: (a) by

Figure 7. Effect of the coagulant dose on the removal of COD for waters III and IV. Experimental conditions were pH 7, mixing time 15 min, 150 rpm, and 20 °C.

Figure 8. Effect of the coagulant dose on the removal of polyphenols (P) and aromatics (A) for waters III and IV. Experimental conditions were pH 7, mixing time 15 min, 150 rpm, and 20 °C.

increasing the concentration of metal hydroxide precipitate and thus the aggregation rate and (b) by enmeshing particulates into ever larger aggregates by sweep-floc coagulation.2 Furthermore, in this wastewater, the addition of a high coagulant dose favors the formation of a greater number of flocs, since increasing the supersaturation (SS), see eq 2, increases the nucleation rate considerably relative to the floc growth rate. The result is a suspension with flocs that are smaller in size and greater in number, and that will hence remove a larger amount of organic matter because of the larger surface area available for adsorption. In contrast, low coagulant doses favor the formation of larger, but fewer, flocs because of the faster growth rate relative to nucleation rate, resulting in a smaller surface area available for the adsorption of organic matter. Doses of FeCl3 of 1000 mg/L for water III and 1400 mg/L for water IV yielded similar values of the removal of COD (approximately 65%), aromaticity, and polyphenols (approximately 90%) in the two types of water. One observes in Figure 7 that there is a relative shift between the two curves of about 400 mg/L of coagulant over the entire range of doses. That is, to achieve the same COD removal in water IV as in water III, it is necessary to increase the concentration of the coagulant by 400 mg/L. Similar conclusions can be drawn for the polyphenol and aromatic compound removals (see Figure 8), the required increases in coagulant concentration being again about 400 mg/L for the polyphenols and 600 mg/L for the aromatic compounds. One deduces that there is a stoichiometric or quasi-stoichiometric relationship between the COD, polyphenol, or aromatic content of the wastewater and the dose of coagulant required to attain a given level of elimination. Other

6544

Ind. Eng. Chem. Res., Vol. 44, No. 17, 2005

Table 2. Inflence of Temperature on the Organic Matter Removala expt

T (°C)

COD removal (%)

aromatics removal (%)

polyphenols removal (%)

T1 T2 T3 T4

20 40 60 80

67 66 64 58

92 83 82 81

89 87 87 86

a Experimental conditions: water III, 150 rpm, mixing time 15 min, coagulant dose 1000 mg/L FeCl3, pH 7.

workers22

report a stoichiometric ratio of 2 mg ofs Fe/mg of TOC in the process of coagulation of various humic substances. In these conditions, it is not easy to determine the principal destabilization mechanism. Although the values of the pH and the coagulant dose are high, which leads one to think of a sweep-floc coagulation mechanism, the observed stoichiometric or quasi-stoichiometric dependence also makes one think that the charge adsorption-neutralization mechanism is acting with a certain importance. It has to be noted that, although high coagulant doses gave the best results for the elimination of organic matter, doses of only 600 ppm of FeCl3 for water III and 800 ppm for water IV already yielded a clarified liquid with satisfactory characteristics. From a practical standpoint, this would therefore allow reagent costs to be reduced. 3.1.5. Effect of Temperature. Experiments were carried out on water III at 20, 40, 60, and 80 °C to study the influence of temperature on the coagulant process. As one observes from the results presented in the Table 2, increasing the temperature in this range of 20-80 °C had a negative, although minimal, effect. Specifically, an increase in temperature from 20 to 80 °C reduced the efficacy of the process from 67% to 58% for COD, from 92% to 83% for the aromatic compounds, and from 89% to 86% for the polyphenols. Other workers have studied the effect of temperature in the range 1-25 °C, but we find no work on coagulation above 25 °C in the literature. At low temperatures, it is commonly found that hydrolyzing metal coagulants perform less well,24,25 and it has been observed that the optimal coagulation pH shifts to a higher value.24,27,28 Hanson and Cleasby24 conclude that the effects of temperature on coagulation cannot be explained by the effect on such parameters as energy dissipation and turbulence microscale. The difference observed in alum coagulation between 20 and 5 °C is related to floc strength and not to turbulent flow field characteristics. The chemical influence of varying the water temperature on coagulation by hydrolyzing metal salts may be related to the effect on hydrolysis reactions, precipitation, and the solubility of the metal hydroxide. Kang and Cleasby25 report that decreasing the water temperature from 25 to 5 °C lowers the minimum solubility of Fe(OH)3 by 0.2 log units, shifting it approximately 0.4 pH units to the alkaline side. Water temperature may also affect the rate of the metal ion hydrolysis reactions and the establishment of equilibrium of the solid phase with dissolved species in solution. With increasing temperature, the hydrolysis of Fe(III) salts is accelerated, and the formation time of soluble polymeric iron species decreases rapidly.28 Morris and Knocke26 report, however, that the rate of aluminum or iron(III) precipitation is not significantly affected over a temperature range of 1-23 °C.

Figure 9. Floc formation: floc nucleation mechanism (left) versus floc growth mechanism (right).

3.2. Settleability Parameters. It is difficult to establish a theoretical or empirical formula that it is applicable to the settling process in wastewaters due to the great complexity of the system and to the variety of conditions that are recorded during the process. Nonetheless, one can state the following: (a) The larger the particles, the greater their settling velocity is. There also seems to be a relationship between the size and the number of particles (see Figure 9), since the process is itself the result of two phenomena, nucleation and floc growth.29 A greater number of small particles would normally imply the predominance of nucleation over growth. Such a situation is more likely the greater the supersaturation (SS) of the solution. (b) The sludge volume index, SVI, measures the compaction of the sediment. As will be seen below, there is a close relationship between the value attained by this parameter and particle size.29 (c) The sediment percentage, SP, will also be related to the particle size and to all the other parameters.29 This parameter is, of course, the complement of the volume percentage of clarified liquid (100 - SP) and should therefore be as small as possible. (d) The residual conductivity reflects the concentration of salts (mostly due to the coagulant) that remain in the clarified liquid. It therefore provides information about how much coagulant has been left over, either because it formed no flocs or flocs that were too small to settle or because it came from broken flocs. 3 (e) With respect to the total suspended solids (TSS), the greater the concentration, the more effective is the elimination of suspended organic matter. (f) Temperature is another parameter that affects the process.24-26 The higher the temperature, the lower the density of the liquid is and therefore the faster the sedimentation is, that is, the greater the yield is for the same retention time in the settling tank. (g) The longer the retention time in the settling tank, the greater is the efficiency attained in the clarification. In the present study, a sequence of experiments was carried out using ferric chloride as a coagulant to determine the optimal conditions of stirring speed, mixing time, pH, coagulant dose, and temperature for good sedimentation (in speed and in percentage of liquid clarified), producing the smallest volume of sludge and the lowest possible residual conductivity. The operating variables used in the trials were stirring speed (60, 150, and 300 rpm), mixing time (5, 15, and 30 min), pH (pH 4-11), coagulant dose (2001000 mg/L of FeCl3), and temperature (20, 40, 60, and 80 °C). Because testing the influence of these variables on all the types of wastewater (waters I-IV) would have involved an unfeasibly large number of experiments, the study was performed using only one or two of the water types for each variable. The choice of which waters to use in each series was partially determined by the availability of the samples. 3.2.1. Influence of the Stirring Speed on the Settleability Parameters. As was observed in the first

Ind. Eng. Chem. Res., Vol. 44, No. 17, 2005 6545

Figure 10. Effect of stirring speed on the settleability parameters in water I. Coagulant dose was 400 mg/L of FeCl3; mixing time was 5 min; pH ) 7; T ) 20 °C. Figure 12. Effect of pH on the settleability parameters in water III. Coagulant dose was 400 mg/L of FeCl3; stirring speed was 150 rpm; mixing time was 5 min; T ) 20 °C.

Figure 11. Effect of mixing time on the settleability parameters in water I. Coagulant dose was 400 mg/L of FeCl3; stirring speed was 300 rpm; pH ) 7; T ) 20 °C.

part of this work, the stirring speed was found to have only a moderate influence on the coagulation process. Figure 10 shows the influence of this factor on the four settleability parameters in the water I. The results for water II were similar. One observes that the most vigorous stirring, 300 rpm, gave the lowest sediment percentage (SP ≈ 20%), and the greatest settling velocity (vs ≈ 1.3 cm/min). The sludge volume index remained practically constant (SVI ≈ 320 mL/g). The lowest residual conductivity in the clarified liquid (∼1800 µS/ cm) was also obtained at 300 rpm. This parameter is proportional to the concentration of dissolved salts, and as was observed in the first part of this work, low values usually coincide with the conditions that give the best coagulation/flocculation of the organic matter. 3.2.2. Influence of Mixing Time on the Settleability Parameters. The mixing time of the coagulant with the wastewater was also a factor with only a moderate influence on the process. Figure 11 shows the results of the trials performed with water I aimed at determining the optimal mixing time. One observes that the shortest time, 5 min, corresponded to the formation of the largest flocs (maximum vs of 1.5 cm/min) and the smallest volume of sludge (SP ) 19%). There were two minima of the sludge volume index, the first at 5 min (SVI ≈ 340 cm3/g) and the second at 30 min (SVI ≈ 320 cm3/g). The lowest residual conductivity also was found for a 5 min mixing time. It therefore seems that, although the effect is only moderate, longer mixing times lead the floc to break up,21 with a consequent reduction in settling velocity and a rise in sediment percentage and residual conductivity. Bearing all the factors in mind including the elimination of organic matter discussed in the first part of the work, the optimal mixing time would be 5 min. 3.2.3. Influence of pH on the Settleability Parameters. The pH is one of the most influential factors affecting the performance of the process. The reason is

that it affects all the hydrolysis equilibria that the coagulation produces, the formation of mono- and polynuclear hydroxo complexes, and the concentrations of the different iron species in solution. All these factors will determine which type of sedimentation takes place (sweep-floc coagulation, adsorption, a combination of the two, etc.) and therefore the effectiveness of the treatment. It must also be taken into account that the wastewater will ultimately be discharged or reused, so the choice of an extreme pH would require a subsequent neutralization treatment. The influence of pH on the sedimentation was studied with water III (see Figure 12). One sees that from pH 6 onward the different parameters are interrelated. From pH 6 to 7, the sediment percentage (SP) and sludge volume index (SVI) increase, while the settling velocity (vs) decreases, that is, more numerous and smaller flocs are produced, which undergo less compression and hence occupy a greater volume. Surprisingly, it was at this point (pH 7) that the removals of COD, polyphenols, and aromatic compounds were greatest (see water III, Figures 4-6), probably because of the greater specific area of the solid adsorbent. In the following interval of pH 7-9, the values of SP and SVI decrease, and those of vs and Kf increase. In the first part of the work, pH 9 corresponded to a minimum in the removal of COD. Here, however, we obtained the best values for the settleability parameters (minima for SP and SVI, while vs has increased). As the pH rises from 7 to 9, the flocs increase in size and the sludge is smaller in volume and more compacted probably due to a greater compression, but the removal of COD diminishes because of the reduction in active adsorption area. Furthermore, from pH 7 onward, there is an increase in the residual conductivity, indicating a greater presence of the coagulant salt in the clarified liquid, probably in the form of Fe(OH)4-. In the interval pH 9-11, the mechanism must be exclusively sweep-floc coagulation, since at these values of the pH the supersaturation (SS) would be very high, and consequently, the floc nucleation rate would be far greater than the floc growth rate. One observes that the values of SP, SVI, and residual conductivity increase, while vs decreases (a worsening of the settleability parameters). The values of COD removal were higher, however, with an improved level of decontamination (see the first part of the work). 3.2.4. Influence of Coagulant Dose on the Settleability Parameters. The influence of the coagulant dose was studied with water III, varying the dose from

6546

Ind. Eng. Chem. Res., Vol. 44, No. 17, 2005

Figure 13. Effect of coagulant dose on the settleability parameters in water III. pH ) 7; stirring speed was 150 rpm; mixing time was 15 min; T ) 20 °C.

200 to 1000 mg/L FeCl3 with the other variables maintained constant. The first part of this work presented a discussion of how the coagulation dose affected the floc nucleation and growth rates and the relative importance of the different mechanisms of organic matter removal (adsorption/neutralization or sweep-floc coagulation). The predominant removal mechanism at low doses is adsorption and charge neutralization, and at high doses, it is that of being swept along by enmeshment in the ferric hydroxide precipitate, sweepfloc coagulation.6,7 Figure 13 shows the influence of the coagulant dose on the settleability parameters (SVI, SP, and vs) and on the residual conductivity in the clarified supernate (Kf). One observes that there is a practically linear increase in SVI and SP over the entire range of doses and also a linear increase in the residual conductivity over the range 400-1000 mg/L FeCl3. The settling velocity, however, shows a practically linear decline over the entire range of doses. The results are coherent with the interpretation being put forward in this work. A greater coagulation dose increases the nucleation rate and reduces the floc growth rate. With more numerous but smaller flocs, the settling velocity (vs) decreases, while SP and SVI increase. Furthermore, since more coagulant salt has been added, the residual conductivity is higher. It may therefore be concluded that while a high coagulant dose is beneficial for the removal of organic matter (see Figure 7), it is detrimental for the settling process since it reduces the settling velocity, generating a larger volume of less consolidated sludge and a higher residual conductivity. 3.2.5. Influence of Temperature on the Settleability Parameters. It was mentioned in the Introduction that the temperature is another parameter that affects the settling process. At higher temperatures, settling is faster because the liquid is less dense, and the minimum solubility of Fe(OH)3 is greater. Also, with changing water temperature there are variations in the chemical effect of hydrolyzing metal salts on coagulation, which may be related to the dependence of the hydrolysis reactions, precipitation, and solubility of the metal hydroxide on temperature.28,30,31 The effect of temperature was studied with water III, varying it over the range 20-80 °C while maintaining the other conditions constant. The results are shown in Figure 14. One observes that raising the temperature from 20 to 80 °C leads to increases in the settling velocity (by a factor of almost three) and the residual conductivity (by 8%) and to decreases in SP (by ap-

Figure 14. Effect of temperature on the settleability parameters in water III. pH ) 7; stirring speed was 150 rpm; mixing time was 15 min; coagulant dose was 1000 mg/L of FeCl3.

proximately a factor of 2) and SVI (by approximately a factor of 3). Unlike the findings of the first part of the work where low temperatures (20 °C) favored the organic matter removal process, the settling process is favored by increasing the temperature (greater settling velocity and lower sediment percentage and sludge volume). Also, the results shown in Figure 14 reflect the greater minimum solubility of the metal hydroxide at higher temperatures. A significant increase in temperature should favor the rate of growth of the flocs relative to their nucleation rate, as was indeed observeds increasing the temperature produced more compact flocs with a far greater settling velocity. It was observed in the first part of the work, however, that the greater size of the flocs was detrimental to the removal of organic matter, since the adsorption area was smaller. 3.2.6. Design of a Clarifier-Thickener Unit. On the basis of the data from the settling trials and other relevant factors, we calculated the design of a clarifierthickener unit according to the Talmadge-Fitch method,32-34 improved by Eckenfelder and Milbinger.35 The operating conditions chosen were pH 7, 600 ppm of FeCl3, T ) 20 °C, mixing time 5 min, and stirring speed 300 rpm. The choice was made taking into account economic factors (coagulant dose and temperature), possible discharge of the water (pH), and optimal reaction conditions (mixing time and stirring speed). We shall assume for the present case that there are two deposits of cork wastewater that supply an inflow to the tank of 2 m3/h. The areas required for clarification (Ac) and for thickening (Ae) were Ac ) 2.36 and Ae) 4.01 m2. The thickening area was therefore that used for the design. Some workers36 apply a scale factor in calculating these two areas. That recommended by Eckenfelder is 1.5 for Ac and 2 for Ae. 3.2.7. Analytical Profile of the Sample Output. Table 3 gives the analytical profile of a wastewater sample of water III before and after treatment under the conditions we used in the design of the clarifierthickener unit. One observes that there is a major removal of turbidity, COD, aromatics, and polyphenols. The greatest negative impact of this treatment may be the high value of the residual conductivity in the clarified water. Also, as was noted in the discussion above, the drawback of working at pH 7 is that the values of SP and SVI are high. 4. Summary and Conclusions The following are the principal conclusions of this work: (1) The removal capacities obtained in the various trials were in the ranges 35-65% for COD, 55-90% for

Ind. Eng. Chem. Res., Vol. 44, No. 17, 2005 6547 Table 3. Analysis of the Sample before and after Treatmenta parameter pH Fe (mg/L) Kf (µS/cm) turbidity (NTU) COD (mg/L) aromatics (mg/L) polyphenols (mg/L) SP (%) SVI (cm3/g) Vs (cm/min)

initial water

treated water

5.6 9.6 894 229 2228 5233 380

7.2 < 0.1 2620 6.40 1043 821 43 31 214 0.82

% removal 99 97 53 84 89

a Conditions: mixing time 5 min; stirring speed, 300 rpm; pH 7; T ) 20 °C, coagulant dose 600 ppm FeCl3, water III. All the values are affected by an error of (5%.

polyphenols, and 40-90% for aromatics. In general, the coagulation process gave greater percentage organic matter removal for the least contaminated waters. (2) Vigorous stirring at the rate of 300 rpm favored the coagulation process. (3) The shortest coagulant mixing time, 5 min, was best suited to the process. Longer times led to the breakup of the floc and the consequent reduction in the process’s effectiveness. (4) In view of the possible discharge/reuse of the water, a pH of 4-5 seemed to be the most favorable for water I, shifting to pH 6 for water II, and pH 7 for water III. That is, the greater the organic matter content of the cork wastewater, the higher the pH required by the coagulation process is. (5) The optimal coagulant dose depended on the contamination level of the wastewater. That is, to attain the same COD removal in water IV as in water III, it was necessary to increase the coagulant concentration by 400 ppm. The case was similar for the polyphenols and aromatic compounds. In particular, doses of 1000 ppm for water III and 1400 ppm for water IV gave similar removals of COD (approximately 65%), aromaticity, and polyphenols (both approximately 90%). (6) The effect of increasing the temperature in the range 20-80 °C was negative, albeit minimal. In particular, with an increase from 20 to 80 °C, the efficacy of the process declined from 67% to 58% for COD, from 92% to 83% for the aromatic compounds, and from 89% to 86% for the polyphenols. (7) The optimal conditions for the removal of organic matter do not coincide with the optimal conditions for sedimentation. (8) At 300 rpm, one obtains a smaller sediment percentage, a greater settling velocity, and a lower residual conductivity in the clarified liquid. (9) A mixing time of 5 min leads to the formation of the largest flocs and the smallest sediment volume. This time also gives the lowest residual conductivity. (10) The pH is one of the most influential factors. In the range pH 6-7, the sediment percentage is high, as is the sludge volume index. At pH 7, there is a maximum in the removal of COD, polyphenols, and aromatic compounds, while the settleability parameters are less favorable. In the range pH 7-9, the values of SP and SVI decrease, and those of vs and Kf increase. At pH 9, there is a minimum in COD removal, but the settleability parameters attain their best values. In the range pH 9-11, the values of SP, SVI, and the residual conductivity increase, as also does the organic matter removal.

(11) Increases in the dose of coagulant, in the range 200-1000 mg/L FeCl3, worsen the settleability parameters (increase in SVI, SP, and Kf). However, the removal of organic matter increases with increasing coagulant dose (very markedly up to 600 mg/L FeCl3). (12) While an increase of temperature from 20 to 80 °C is beneficial for the settleability parameters (decreases of SP and SVI by factors of approximately two and three, respectively), it reduces the removal of organic matter by about 10%. (13) According to the method of Talmadge-Fitch, a clarifier-thickener unit to treat a wastewater flow of 2 m3/h would require a minimum area of 4.01 m2. Acknowledgment This project was supported financially by the Comision Interministerial de Ciencia y Tecnologia, CICYT (“proyecto subvencionado PPQ 2001-0744”), and by the Junta de Extremadura (“proyecto subvencionado 2PR01A113”). Literature Cited (1) Randtke, S. J. Organic contaminant removal by coagulation and related process combinations. J. Am. Water Works Assoc. 1988, 80, 40. (2) Ching, H. W.; Tanaka, T. S.; Elimelech, M. Dynamics of coagulation of kaolin particles with ferric chloride. Water Res. 1994, 28, 559. (3) Aguilar, M. I.; Sa´ez, J.; Llorens, M.; Soler, A.; Ortun˜o, J. F. Tratamiento fı´sico-quı´mico de aguas residuales. Coagulacio´ nfloculacio´ n; Universidad de Murcia: Murcia, Spain, 2002. (4) Licsko´, I. Dissolved organics removal by solid-liquid-phase separation (adsorption and coagulation). Water Sci. Technol. 1993, 27, 245. (5) Jekel, M. R. Interactions of humic acids and aluminum salts in the floculation process. Water Res. 1986, 20, 1535. (6) Licsko´, I. On the types of bond developing between the aluminium and iron (III) hydroxides and organic substances. Water Sci. Technol. 1993, 27, 249. (7) Rebhun, M.; Curie, M. Control of organic matter by Coagulation and floc separation. Water Sci. Technol. 1993, 27, 1. (8) Dentel, S. K. Coagulation control in water treatment. Crit. Rev. Environ. Control 1991, 21, 41. (9) Dentel, S. K.; Gossett, J. M. Mechanisms of coagulation with aluminum salts. J. Am. Water Works Assoc. 1988, 80, 187. (10) Stumm, W.; O’Melia, C. R. Stoichiometry of Coagulation. J. Am. Water Works Assoc. 1968, 60, 514. (11) Fitch, B. Sedimentation of flocculent suspensions: State of the art. AlChE J. 1979, 25, 913. (12) Company, J. Coagulantes y Floculantes aplicados en el Tratamiento de Aguas; Gestio´ y Promocio´ Editorial: Barcelona, Spain, 2000. (13) Edwards, A. C.; Cresser, M. S. Relationships between ultraviolet absorbance and total organic carbon in two upland catchments. Water Res. 1987, 21, 49. (14) Mrkva, M. Evaluation of correlation between absorbance at 254 nm and COD of river waters. Water Res. 1983, 17, 231. (15) Schmauch, L. J.; Grubb, H. M. Determination of phenols in wastewaters by ultraviolet absorption. Anal. Chem. 1954, 26, 308. (16) Box, J. D. Investigation of the Folin-Ciocalteau phenol reagent for the determination of polyphenolic substances in natural waters. Water Res. 1983, 17, 511. (17) Beltra´n, J.; Torregrosa, J.; Gonza´lez, T. Experimentacio´ n en el tratamiento de aguas residuales; Universidad de Extremadura: Badajoz, Spain, 2004. (18) Beltra´n, J.; Torregrosa, J.; Gonza´lez, T.; Domı´nguez, J. R. Ana´ lisis quı´mico de aguas residuales; Abecedario: Badajoz, Spain, 2004. (19) Matsuo, T.; Unno, H. Forces acting on floc and strength of floc. J. Environ. Eng. Div. (Am. Soc. Civ. Eng.) 1981, 107, 527.

6548

Ind. Eng. Chem. Res., Vol. 44, No. 17, 2005

(20) Brakalov, L. B. A connection between the orthokinetic coagulation capture efficiency of aggregates and their maximum size. Chem. Eng. Sci. 1987, 42, 2373. (21) Mu¨hle, K. Floc Stability in Laminar and Turbulent Flow. In Coagulation and Flocculation; Decker, D. B., Ed.: New York, 1993; p 355. (22) Lefebvre, E.; Legube, B. Iron (III) coagulation of humic substances extracted from surface waters: Effect of pH and humic substances concentration. Water Res. 1990, 24, 591. (23) Stephenson, R. J.; Duff, S. J. B. Coagulation and precipitation of a mechanical pulping effluent-I Removal of carbon, colour and turbidity. Water Res. 1996, 30, 781. (24) Hanson, A. T.; Cleasby, J. L. Effects of temperature on turbulent flocculation. Fluid dynamics and chemistry. J. Am. Water Works Assoc. 1990, 82, 56. (25) Kang, L. S.; Cleasby, J. L. Temperature effects on flocculation kinetics using Fe(III) coagulant. J. Environ. Eng. 1995, 121, 893. (26) Morris, J. K.; Knocke, W. R. Temperature effects on the use of metal-ion coagulants for water treatment. J. Am. Water Works Assoc. 1984, 76, 74. (27) Van Benschoten, J. E.; Edzwald, J. K. Chemical aspects of coagulation using aluminum salts-I Hydrolytic reactions of alum and polyaluminum chloride. Water Res. 1990, 24, 1519. (28) Van der Woude, J. H.; De Bruyn, P. L. Formation of colloidal dispersions from supersaturated iron (III) nitrate solutions. I Precipitation of amorphous iron hydroxide. Colloids Surf. 1983, 8, 55. (29) Garcı´a, H. M. Tratamiento de las aguas residuales de la industria productora de corcho mediante procesos fisicoquı´micos de coagulacio´ n-floculacio´ n con sulfato de aluminio; Proyecto Fin de Carrera, Universidad de Extremadura: Badajoz, Spain, 2003.

(30) Vik, E. A.; Eikebrokk, B. Coagulation process for removal of humic substances from drinking water. In Aquatic humic substances. Influence on fate and treatment of pollutants; Suffet, I. H., MacCarthy, P., Eds.; American Chemical Society: Washington, DC, 1989. (31) Hem, J. D.; Roberson, C. E. Aluminum hydrolysis reactions and products in mildly acidic aqueous systems. In Chemical Modeling of Aqueous Systems II; Melchior, D. C., Bassett, R. L., Eds.; ACS Symposium Series 416; American Chemical Society: Washington, DC, 1990. (32) Degre´mont. Manual Te´ cnico del Agua; Degre´mont: RueilMalmaison, France, 1979. (33) Smth-Palmer, T.; Campbell, N.; Bowman, J. L.; Dewar, P. Flocculation behaviour of some cationic polyelectrolytes. J. Appl. Polym. Sci. 1994, 52, 1317. (34) Talmadge, W. P.; Fitch, B. Determining thickener unit areas. Ind. Eng. Chem. 1955, 47, 38. (35) Walter, R. H.; Sherman, R. M.; Downing, D. L. Reduction in oxygen demand of abattoir effluent by precipitation with metal. J. Agric. Food Chem. 1974, 22, 1097. (36) Sastre, H. Tratamientos secundarios fisico-quı´micos de efluentes lı´quidos. In Contaminacio´ n de las aguas, Contaminacio´n e Ingenierı´a Ambiental, FICYT: Oviedo, Spain, 1997.

Received for review December 22, 2004 Revised manuscript received June 3, 2005 Accepted June 7, 2005 IE0487641