Los Angeles, London, New Delhi and Singapore http://www.sagepub.com
Copyright © ISWA 2009 ISSN 0734–242X Waste Management & Research 2009: 27: 534–541
DOI: 10.1177/0734242X08096974
Treating landfill leachate by electrocoagulation Job Contreras, Mario Villarroel, Rodrigo Navia Departamento de Ingeniería Química, Universidad de La Frontera, Casilla 54-D, Temuco, Chile
Margarita Teutli Facultad de Ingeniería, Benemérita Universidad Autónoma de Puebla, Edificio 123, Cd. Universitaria, Puebla, México, C.P. 72570
The determination of the optimal conditions for using electrocoagulation as a treatment for landfill leachate was carried out using surface response methodology. A central composite design was applied to investigate the effect of four control factors, namely current density, pH, time and fluid conductivity, as well as the interaction among (between) them to get an optimal turbidity removal. The independent variables were each coded at three levels and their values were selected on the basis of preliminary experimental results. The central composite design consisted of 29 experimental points with five replications at the centre point. A second order polynomial model was used for predicting the response. Regression analysis showed that more than 95% of the variation was explained by the model wherein current density with a 60.1% contribution turned out to be the factor with the most significant influence. Analysis of variance showed that time, pH, current density and the interaction time/current density had a significant influence on the turbidity removal, The optimal conditions obtained for the removal of turbidity were time 38.8 min, pH 7.6, current density 109.9 A m–2 and NaCl 2.9 g L–1. Experimental results showed that for a 96.9% turbidity removal, similar reduction in Al (97.0%) and Fe (99.5%) concentrations; as well as 66% total Kjeldahl nitrogen removal were obtained. Furthermore, the sludge formed exhibited a good floc size, which separated in less than 10 min by classical sedimentation. The results analysis provided evidence of reduction of chemical pollutants, although the electrocoagulated leachate could not satisfy regulations for the maximum Total Kjeldahl nitrogen leachate discharge level to public wastewater collection systems in the town of Freire, Chile. Keywords: Electrocoagulation, factorial design, landfill leachate, wmr 1234–1
Introduction Landfill leachate consists of liquids entering the landfill from external sources, as well as that produced by solid waste decomposition. Leachate is a complex mixture that contains biological and chemical constituents, and its composition varies with landfill age. Leachate management is essential to eliminate any potential risk of polluting underground aquifers. Leachate management options can be simple, such as recycling it back through the waste (Jun et al. 2006), or complicated, with multiple treatment and disposal steps involved (Tchobanoglous et al. 1994). Most leachate treatment processes focus on pH control and removal of organics, nitrogen, metals and anions. Recently, some successfully applied treatments for landfill leachate have included the process sequence of coagulation–flocculation, adsorption, and chemical oxidation (Rivas et al. 2005a, b).
Chemical coagulation is based on adding either iron or aluminium sulfate for charging suspended/dissolved chemical species. On the one hand, once electric forces are destabilized, ions can be agglomerated, forming particles that may settle. On the other hand, electrocoagulation generates coagulant ions via electrolytic reactions at the electrodes by means of a voltage difference. Electrocoagulation has been shown to be an adequate treatment for removal of fluoride in water conditioning (Mameri et al. 1998), for reducing high organic loads in wastewater (Chen et al. 2000, Koparal & Ögütveren 2002), and for removing non-biodegradable compounds such as nonylphenol or textile dyes from wastewater (Ciorba et al. 2002, Bayramoglu et al. 2004). In electrocoagulation, electrolytic reactions at each electrode provide not only coagulant ions (either iron or alumin-
Corresponding author: Prof. Dr. Rodrigo Navia, 1Departamento de Ingeniería Química, Universidad de La Frontera, Casilla 54-D, Temuco, Chile. Tel: +56 45 325477; fax: +56 45 325053 ; e-mail:
[email protected] Received 23 May 2007; accepted in revised form 23 July 2008 Figure 1 appears in color online: http://wmr.sagepub.com
534
Treating landfill leachate by electrocoagulation
ium), but also gas that favours a flotation process. Thus, target contaminants can be recovered by both sedimentation and flotation (Bayramoglu et al. 2004). A variety of factors have been reported to influence electrocoagulation removal efficiencies. One is the nature of the electrode material. Electrodes can employ either aluminium or iron plates, or in some cases both electrode types can be used (Ilhan et al. 2008, Gomes et al. 2007). Current density is the current applied by one unit of electrode surface (A m–2), and its magnitude determines the amount of anode dissolution that will occur. Optimal current density reported values for turbidity removal can range between 80 and 100 A m–2 in the case of Fe electrodes and between 150 and 200 A m–2 when using Al electrodes (Mameri et al. 1998, Bayramoglu et al. 2004). On the other hand, Al electrodes showed a better performance regarding chemical oxygen demand (COD) removal from leachate for a current density range between 348 and 631 A m–2 (Ilhan et al. 2008). Conductivity, the reciprocal of the electrical resistance (S cm–1), is needed in the target matrix to ensure electron movement through the solution. Typically, when the leachate conductivity is low (< 500 µS cm–1), a supporting electrolyte, such as CaCl2, NaCl, Na2SO4 and/or KCl, is added (Bayramoglu et al. 2004). To optimize the operational cost of electrocoagulation, such supporting electrolytes may be added until leachate conductivities of 3500 µS cm–1 (Bayramoglu et al. 2004). Another influential factor with regard to electrocoagulation success is pH, which affects both the current efficiency and hydroxide solubility. Experimental electrocoagulation conducted at a pH of between 6 and 8 has yielded high solid removal efficiency and minimal aluminium residue (Jiang et al. 2002). Finally, the amount of time that electrocoagulation is allowed to proceed influences contaminant removal. Electrocoagulation time depends on the current density applied in order to obtain the defined turbidity efficiency. For a typical current density range (100–300 A m–2), electrocoagulation time varies between 2 and 5 min (Mameri et al. 1998) In traditional process optimization, experimental factors are varied one at a time, with the remaining factors being held constant. For such a technique to have general relevance it is necessary to assume that the effect exhibited by the variable in question would remain unchanged in the presence of other variables. Certainly there remains a high degree of uncertainty regarding this assumption (Zhang et al. 2008). Alternatively, a factorial design approach, such s a two-level-factorialdesign which can be used to overcome the problem of intervariable interaction (Anderson & Whitcomb, 1996), will have better reliability. There are a few advantages in two-level-factorial-design over the traditional method (Anderson & Whitcomb 1996). By initially restricting the tests to only two levels, the number of experiments can be minimized. The two-level factorial design requires only two runs per factor studied, for example, low and high levels. This statistics-based method involves simultaneous adjustment of experimental factors at only two levels, assuming linearity in the factor effects (Zhang et al. 2008). Even though two-level factorial design is unable
to fully explore a wide range in the factor space, it can indicate major trends. A promising direction for further experimentation can be determined because the few critical factors are separated from the insignificant factors. Furthermore, they can detect and estimate interactions among variables. There have been several reports on the application of surface response methodology (SRM) to wastewater treatment processes in the last few years (Benatti et al. 2006, Chakinala et al. 2007, Catalkaya & Kargi 2007) but few reports of its application to the treatment of landfill leachate are available (Zhang et al. 2008). As each leachate matrix is likely to be unique, multiple trials that vary these parameters (electrode material, current density, conductivity, pH, and time) are required for each leachate for which an electrocoagulation process is considered. Certain modelling tools are available that can be used to predict the particular combination of factor values that will provide optimum turbidity removal. The predictions from the model can then be analysed statistically to generate an equation that relates the parameters and suggests the most potent factors among them. SRM is one such statistical analysis that can assess the validity of a prediction and describe the interactions among the parameters. As such, it can be used to optimize a target response based on an accounting of interactions between independent variables (Montgomery 2001). This dependence can be expressed by equation (1): Y = f (x1, x2, …, xn) + ε
(1)
where ε is the error, xi is the ith independent variable, and Y the observed response. If the expected response is denoted by E(Y), then the corresponding surface response is: E (Y) = f (x1, x2, …, xn) = η
(2)
In SRM problems, the relationship between independent variables and a target observed response is unknown. Therefore, the first step in SRM is to find the lowest degree of a polynomial expression that can represent the relationship between the observed response (Y) and the independent variables, obtaining a model for predicting electrocoagulation performance in leachate treatment. The purpose of this research was to use an experimental factorial design process to predict the conditions that would optimize electrocoagulation treatment for a particular landfill leachate. In addition, model predictions were analysed by SRM.
Materials and methods Materials Landfill leachate
Leachate was obtained from Coipúe’s, a 6-year-old landfill, at Freire’s community, located in the Araucanía Region of southern Chile. The leachate had fairly high levels of total suspended solids (TSS), aluminium, total Kjeldahl nitrogen
535
J. Contreras, M. Villarroel, M. Teutli, R. Navia
Fig. 1: Reactor configuration.
Sedimentation unit
Table 1: Freire’s leachate physicochemical characterization. Parameter
A classic sedimentation unit was used for floc separation and further supernatant analysis.
Unit
Value
Biochemical oxygen demand (BOD5)
mg L–1
132
Total Suspended solids (TSS)
mg L–1
305
Experimental design
Total Phosphorus (P)
mg L
–1
7.0
Total Kjeldahl nitrogen (TKN)
mg L–1
371
NTU
1338
According to Montgomery et al. (2003), an experimental design must define the parameters that constitute the important independent variables, as well as the parameters that represent the dependent variables, or the expected response. For electrocoagulation, the selected independent variables were (Mameri et al. 1998, Gürses et al. 2002, Jiang et al. 2002, Bayramoglu et al. 2004): time, pH, current density, and conductivity (expressed as NaCl concentration), and the dependent output response variable is turbidity removal efficiency. For the experimental factorial design process, the experimental domains were defined at the following levels: a centre point value (0), the minimal (–1) and maximum (+1) values; as well as projections on the minimal (–α) and maximum (+α) values (Table 2). Experimental work was designed on a 2k factorial basis, where ‘k’ corresponds to the independent variables (k = 4) time, pH, current density, and conductivity (expressed as NaCl concentration). The assigned values for each independent variable were selected taking into account previous research work in this field (Mameri et al. 1998, Gürses et al. 2002, Jiang et al. 2002, Bayramoglu et al. 2004): experimental time (X1) from 15 to 65 min; pH (X2) from 6.5 to 9.5; current density (X3) from 20 to 160 A m–2; and conductivity (NaCl concentration) (X4) from 0 to 5 g L–1. The use of base 2 means that the work was developed on two levels (maximum and minimal). This type of design required running 16 experiments to determine the variables interactions and their influence on
Turbidity (T) Conductivity (C) Aluminium (Al) Iron (Fe)
mS cm
–1
mg L–1 mg L
–1 –1
6.19 156 82 < 0.01
Lead (Pb)
mg L
Cadmium (Cd)
mg L–1
< 0.001
Copper (Cu)
mg L–1
0.12
–
7.6
pH
(TKN), turbidity, and iron (Table 1). Analytical grade sodium chloride was used as the supporting electrolyte for modifying leachate conductivity. Electrocoagulation cell
An experimental cell was designed according to literature descriptions (Yousuf et al. 2001, Gürses et al. 2002), and electrodes built from high purity aluminium (surface area approx. 54 cm2) were used in a 1.0 L glass beaker. Stirring of cell contents was done with a Heidolph MR 3001K stirrer plate (Figure 1). Electrical instrumentation
Current was imposed with a DC UNILAB power source; electrical parameters (current and voltage) were registered with two HC 3037N ammeters (see Figure 1).
536
Treating landfill leachate by electrocoagulation
Table 2: Experimental range and levels of process variables. Range and level Independent variable X1: Time (min)
–α
–1
0
1
α
15.0
20.0
40.0
60.0
65.0
X2: pH
6.5
7.0
8.0
9.0
9.5
X3: Current density (A m–2)
20.0
30.0
90.0
150.0
160.0
X4: NaCl concentration (g L–1)
0.0
0.5
2.5
4.5
5.0
Table 3: Experimental design matrix and obtained turbidity removal efficiency. Time (min)
pH
Current density (A m–2)
NaCl conc. (g L–1)
Turbidity
X1
X2
X3
X4
(NTU)
Actual
Prediction
1
–1
–1
–1
–1
94.20
92.96
92.86
2
1
–1
–1
–1
47.80
96.43
96.04
3
–1
1
–1
–1
170.00
87.29
89.82
4
1
1
–1
–1
110.00
91.78
92.78
5
–1
–1
1
–1
18.80
98.59
98.99
6
1
–1
1
–1
7.92
99.41
99.52
7
–1
1
1
–1
44.10
96.70
96.96
8
1
1
1
–1
23.70
98.23
97.27
9
–1
–1
–1
1
113.00
91.55
92.46
10
1
–1
–1
1
44.10
96.70
96.56
11
–1
1
–1
1
241.00
81.99
91.28
12
1
1
–1
1
58.50
95.63
95.17
13
–1
–1
1
1
14.80
98.89
98.01
14
1
–1
1
1
6.40
99.52
99.46
15
–1
1
1
1
33.30
97.51
97.84
16
1
1
1
1
15.20
98.86
99.08
17
–α
0
0
0
35.90
97.32
96.62
18
a
0
0
0
15.70
98.83
99.38
19
0
–α
0
0
22.20
98.34
98.48
20
0
a
0
0
51.70
96.14
95.91
Experiment
Removal efficiency (%)
21
0
0
–α
0
73.10
94.54
93.85
22
0
0
a
0
11.00
99.18
99.71
23
0
0
0
–α
31.20
97.67
97.44
24
0
0
0
a
23.70
98.23
98.32
25
0
0
0
0
20.30
98.48
98.54
26
0
0
0
0
19.40
98.55
98.54
27
0
0
0
0
19.40
98.55
98.54
28
0
0
0
0
20.00
98.51
98.54
29
0
0
0
0
22.90
98.29
98.54
turbidity removal. Additional experiments included: eight trials corresponding to the extreme conditions and five repetitions of the central points for each independent variable, so that the total number of experiments was 29. Table 3 reports the selected values for each independent variable (X1, X2, X3, X4) as well as the observed turbidity removal (desired response). The removal efficiencies for experiments 3 and 11 were less than 90%. All other efficiencies ranged between 90 and 100%. It is remarkable that repetitions in the central
point (experiments 25 to 29) reached turbidity removals higher than 98%. Once the removal efficiency was obtained, an SRM statistical analysis of the results in Table 3 and parameter interactions was performed with Design Expert 6.0.6 software. The obtained output provided the surface response variable and its model equation. Software results provided three options: linear, quadratic and cubic models, with a suggestion for using the quadratic one.
537
J. Contreras, M. Villarroel, M. Teutli, R. Navia
Results and discussion Analysing the trend of maximum and minimum values for each independent variable, the following could be deduced. 1. An increase in time produced an increase in turbidity removal efficiency, with a soft slope evident for the maximum values and a variable slope for the minimal ones. The average values show that maximum removal could be obtained in 40 min. 2. An increase in pH produced contradictory trends in the efficiency of turbidity removal. Increasing pH lowered the maximum removal efficiency value, while the minimal value became optimal at pH 8. The average values predict that maximum turbidity removal efficiency would take place at pH 8, which is very close to the original leachate pH (7.6). This suggests that the pH would not be suitable for modification. 3. The current density plot shows a rising trend in turbidity removal efficiency for both minimal and maximum values, with a retarding effect after 90 A m–2. This suggests that the cost associated with using a higher current would yield only minimal improvement in turbidity removal. 4. For NaCl concentration, the observed maximum values were practically constant whereas the minimal values exhibit a maximum removal at a dosage of 2.5 g L–1. NaCl addition could be omitted, because turbidity removal efficiencies were higher than 90% even without NaCl addition, as evident in experiment 23 in Table 3. When each of the terms of the quadratic equation of the model predictions were subjected to variance analysis (data not shown), the obtained R2 value was 0.9525 (Table 4), and the predictive R2 was 0.5654 These values reflect the accuracy and suitability of the model used, as well as the influence of each independent Table 4: Model R 2 analysis. Variation source
R2
Model
0.9525
X1
0.1198
X2
0.0790
X3
0.6008
X4
0.0156
X12
0.0055
2
0.0287
X32
0.0662
2
0.0082
X1X2
0.0002
X1X3
0.0338
X1X4
0.0054
X2X3
0.0049
X2X4
0.0221
X3X4
0.0022
X2 X4
538
variable in the model. The current density (X3), with R2 = 0.6008 was the factor that had the most significant impact. Based on the 29 experimental results, a global average turbidity removal of 97.2% was obtained, which reflects that the experimental configuration was very close to an optimal one. The prediction of turbidity removal capacity was based on an assessment of the influence of the independent variables and their interactions. All of the processes are accounted for in equation (3), the regression equation for modeling a surface response: η = 60.82494 + 0.18092X1 + 7.87744X2 + 0.098065X3 – 1.28741X4 – 8.57449 × 10–4X12 – 0.59916X22 – 3.59315 × 10–4X32 – 0.10547X42 – 2.72909 × 10–3X1X2 – 5.52411 × 10–4X1X3 + 5.78869 × 10–3X1X4 + 4.19718 × 10–3X2X3 + 0.23362X2X4 – 1.21554 × 10–3X3X4
(3)
The current density (X3) has the lower coefficient in the linear terms of equation (3); whereas in the quadratic terms, X3 has the higher coefficient. This quadratic behaviour is clearly exhibited in the corresponding surface response graph (Figure 2). In Figure 2a, predictions of turbidity removal efficiencies are functions of current density and time. The current density lines exhibit a bend, which agrees with the corresponding quadratic coefficient in the regression equation. Otherwise, the main contribution of time is in the linear terms, which is reflected by the lines of variable slope in Figure 2. SRM predictions suggest that time–pH, time–NaCl concentration and current density–NaCl concentration interactions were not meaningful. The latter may be due to the fact that leachate conductivity required no modification and NaCl addition was not warranted.
Model predictions and validation One of the main advantages of using Design Expert software to analyse the predictions is that it provides a model equation, and a set of predictions for optimal turbidity removal efficiencies with values higher than 98% (Table 5). From Table 5, prediction 6 was selected for laboratory verification. When the electrocoagulation of the leachate sample was conducted using the values from Table 6, the experimental turbidity removal efficiency (99.2%) was very close to the one predicted by the model (99.0 %), thus validating model predictions. With regard to the other parameters, it can be seen that this set-up allowed removal efficiencies higher than 90% for biochemical oxygen demand (BOD5), total suspended solids (TSS), Al and Fe. Regarding pH, a slight rise is observed
Treating landfill leachate by electrocoagulation
Fig. 2: Response surface for interaction (a) current density-time, b) pH-time and c) NaCl concentration-time.
Table 5: Model predictions for turbidity removal efficiencies higher than 98%. Prediction
Time (min) X1
pH
1
X2
Current density (A m–2) X3
NaCl concentration (g L–1) X4
Turbidity removal efficiency (%)
46.40
7.10
145.90
3.80
99.70
2
36.00
8.10
130.80
3.10
99.46
3
59.90
7.30
136.50
4.10
100.00
4
32.10
7.80
120.30
1.90
99.17
5
37.70
8.50
103.00
2.30
98.30
6
30.80
7.60
109.90
2.90
99.00
7
31.40
7.00
127.80
4.20
98.75
8
58.20
8.50
137.00
2.70
99.52
9
30.77
7.52
109.62
2.89
99.00
10
30.00
7.60
130.00
0.00
99.01
539
J. Contreras, M. Villarroel, M. Teutli, R. Navia
Table 6: Experimental data obtained with recommendation 6 from Table 5. Parameter
Unit
Initial value
Final value
Removal efficency (%)
mg L
–1
132
11
91.7
Total suspended solids (TSS)
mg L
–1
305
17
94.4
Total phosphorus (P)
mg L–1
7.0
n.d.
–
Biochemical oxygen demand (DBO5)
Total Kjeldahl nitrogen (TKN)
mg L
Turbidity (T)
–1
NTU
Conductivity (C )
mS cm
–1
–1
371
126
66.0
1,338
11.2
99.2
6.19
10.07
–
Aluminium (Al)
mg L
156
4.82
97.0
Iron (Fe)
mg L–1
82
0.38
99.5
Lead (Pb)
mg L–1
< 0.010
n.d.
–
–1
Cadmium (Cd)
mg L
< 0.001
n.d.
–
Copper (Cu)
mg L–1
0.12
n.d
–
–
7.6
8
–
pH n.d., not detected.
Table 7: Experimental design matrix and obtained turbidity removal for aluminium. Run
Current density
Time
A m–2 A1
Response 1: Turbidity
Response 2: Aluminium
min
NTU
Removal (%)
g L–1
Removal (%) 96.03
–1
20.0
–1
15.0
35.5
97.30
6.18
A2
1
160.0
–1
15.0
13.0
99.02
5.44
96.51
A3
–1
20.0
1
65.0
30.2
97.74
5.64
96.38
A4
1
160.0
1
65.0
12.4
99.07
9.11
94.16
(from 7.6 to 8), but the final value still falls between allowed limits in the Chilean Norms for wastewater discharge. Although TKN removal efficiency was 66%, the final value exceeded the value allowed by the above-cited norms, and it will be a matter of study in further experiments. According to Chilean Norms, aluminium concentration must be < 5 mg L–1 in a wastewater to be authorized for discharge or infiltration. The applied treatment allowed a removal efficiency of 97% (from 156 to 4.82 mg L–1), and thus brought aluminium concentration to within the allowed range. Considering that electrical set-up was implemented with aluminium electrodes, it is possible that electrode dissolution influenced the residual concentration. Therefore a factorial design was planned specifically for aluminium removal. According to the variance analysis reported in Table 4, time and current density were found to be the most significant variables affecting the efficiency of turbidity removal; therefore, time and current density were considered as variables, whereas NaCl concentration was set at 2.5 mg L–1, and the pH value at 8. The corresponding factorial design matrix, and responses obtained for turbidity and aluminium concentration are reported in Table 7. The difference between this latter analysis and the one neglecting electrode effects suggests that anode dissolution may contribute to the residual aluminium concentration remaining after leachate electrocoagulation.
540
Conclusions 1. The application of surface response methodology provided a model equation that predicted conditions of optimum turbidity removal efficiency. The prediction was easily tested, and the results confirmed the model prediction. 2. The factorial design was a useful tool for optimizing leachate treatment by electrocoagulation, allowing correct choices of the appropriate experimental domains for time, pH, current density and sodium chloride concentration. The predicted values proved to be the best option, since turbidity removal efficiency was higher than 90% in 27 of the 29 experiments. 3. Although the final concentration of total Kjeldahl nitrogen was higher than the allowed concentration in Chilean Norms for wastewater discharge, a reduction of 66% was obtained, which was considered to be a significant removal efficiency.
Acknowledgements This work was supported by an FNDR Grant from the Regional Government of the Araucanía Region of Chile, and partially supported by DIUFRO project 120519 from the Universidad de La Frontera, Temuco, Chile.
Treating landfill leachate by electrocoagulation
References Anderson, M.J. & Whitcomb, P.J. (1996) Optimize your process-optimization effects. Chemical Engineering Progress, 92, 51–60. Bayramoglu, M., Kobya, M., Taner, O. & Sozbir, M. (2004) Operating cost analysis of electrocoagulation of textile dye wastewater. Separation and Purification Technology, 37, 117–225. Benatti, C.T., Tavares, C.R.G. & Guedes, T.A. (2006) Optimization of Fenton’s oxidation of chemical laboratory wastewaters using the response surface methodology. Journal of Environmental Management, 80, 66–74. Catalkaya, E.C. & Kargi, F. (2007) Effects of operating parameters on advanced oxidation of diuron by the Fenton’s reagent: a statistical design approach, Chemosphere, 69, 485–492. Chakinala, A.G., Bremner, D.H., Gogate, P.R., Namkung, K.C. & Burgess, A.E. (2007) Multivariate analysis of phenol mineralisation by combined hydrodynamic cavitation and heterogeneous advanced Fenton processing. Applied Catalysis, B78, 11–18. Chen, G., Chen, X. & Yue, P.L. (2000) Electrocoagulation and electroflotation of restaurant wastewater. Journal of Environmental Engineering, 126, 858–863. Ciorba, G.A., Radovan, C., Vlaicu, I. & Masu, S. (2002) Removal of nonylphenol ethoxylates by electrochemically-generated coagulants. Journal of Applied Electrochemistry, 32, 561–567. Gomes, J.A.G., Daida, P., Kesmez, M., Weir, M., Moreno, H., Parga, J.R., Irwin, G., McWhinney, H., Grady, T., Peterson, E. & Cocke, D.L. (2007) Arsenic removal by electrocoagulation using combined Al–Fe electrode system and characterization of products. Journal of Hazardous Materials, 139, 220–231. Gürses, A., Yalçin, M. & Dogar, C. (2002) Electrocoagulation of some reactive dyes: a statistical investigation of some electrochemical variable. Waste Management, 22, 491–499. Ilhan, F., Kurt, U., Apaydin, O. & Gonullu, M.T. (2008) Treatment of leachate by electrocoagulation using aluminum and iron electrodes. Journal of Hazardous Materials, 154, 381–389.
Jiang, J. Q., Graham, N., Andre, C., Kelsall, G. H. & Brandon, N. (2002) Laboratory study of electrocoagulation-flotation for water treatment. Water Research, 36(16), 4064-4078. Jun, D., Yongsheng, Z., Henry, R.K. & Mei, H. (2006) Impacts of aeration and active sludge addition on leachate recirculation bioreactor. Journal of Hazardous Materials, 147, 240–248. Koparal, A.S. & Ögütveren, Ü.B. (2002) Removal of nitrate from water by electroreduction and electrocoagulation. Journal of Hazardous Materials, 89, 83–94. Mameri, N., Yeddou, A.R.., Lounici, H., Belhocine, D., Grib, H. & Bariou, B. (1998) Defluoridation of septentrional Sahara water of North Africa by electrocoagulation process using bipolar aluminium electrodes. Water Research, 32, 1604–1612. Montgomery, D.C. (2001) Design and Analysis of Experiments, 5th edn. Wiley & Sons Ltd., NY, USA. Montgomery, D.C., Runger, G.C. & Faris Hubele, N. (2003) Engineers Statistics, 3rd edn. Wiley & Sons Ltd., NY, USA. Rivas, F.J., Beltrán, F.J., Carvalho, F. & Alvarez, P.M. (2005a). Ozonepromoted wet air oxidation of landfill leachates. Industrial and Engineering Chemistry Research, 44, 749–758. Rivas, F.J., Beltrán, F.J., Carvalho, F., Gimeno, O. & Frades, J. (2005b) Study of different integrated physical-chemical + adsorption processes for landfill leachate remediation. Industrial and Engineering Chemistry Research, 44, 2871–2878. Tchobanoglous, G., Theisen, H. & Vigil, S.A. (1994) Gestión integral de residuos sólidos. Mc Graw Hill, Mexico. Yousuf, M., Mollah, A., Schennach, R., Parga, J.R. & Cocke D.L. (2001) Electrocoagulation (EC)-science and applications. Journal of Hazardous Materials, 84, 29–41. Zhang, H., Choi, H.J., Canazo P. & Huang, C.-P. (2008) Multivariate approach to the Fenton process for the treatment of landfill leachate. Journal of Hazardous Materials, (in press).
541
