Abstract: Biological methods are not appropriate for mature leachate (5-10 years) due to the presence of bio refractory materials and toxic compounds such as heavy metals. Now-a-days advanced oxidation processes such as Fenton reagent are used for the treatment or biodegradability enhancement of refractory wastewater. In this study, the biodegradability enhancement of mature landfill leachate using Fenton oxidation process in lab scale has been investigated. To optimize Fenton reaction for biodegradability enhancement of leachate, Response surface methodology was used. Design Expert7 was used for statistical analysis of experimental data. The leachate samples have been taken from a 10 years old landfill and oxidation experiments in lab scale were done in different conditions from operational points of view (reaction time, pH, Hydrogen peroxide to Fe (II) ratio). The BOD5, COD, pH, H2O2, Fe2+ oxidation time were evaluated in 32 runs, 12 run as pretests and 20 run for optimization. Results from numerical optimization showed that in pH = 3.3, H2O2/Fe2+ = 15.82 and oxidation time of 130 min, COD removal, BOD5 and BOD5/COD ratio variation were 93.1, 0.56 and 0.49%, respectively. In optimal condition that BOD5/COD ratio will be 0.45, the oxidation time is 130 min, pH = 3-3.5 and H2O2/Fe2++ ratio = 14-17. This study indicates that Fenton oxidation enhances the biodegradability of the mature leachate. In the optimal condition of Fenton oxidation the BOD5/COD ratio increased to 80 times greater than initial ratio. Results from experiments will agree to results from model. This agreement confirms accuracy and approach of obtained model.
INTRODUCTION
leachate generated from landfills can be defined as a high strength organic wastewater which is containing the high concentration of recalcitrant organics and toxic mineral matter such as heavy metals (Englehardt et al., 2006). When untreated leachate discharged directly to the environment may cause many environmental and health adverse effects (Zhang et al., 2005). Management of all types of wastewater can play an important role in the protection of public health and the environment (Farrokhi et al., 2008, 2014).
There are many important effective factors on the quality and quantity of leachate, include; seasonal changes, type of landfill method, method of compacting, characteristics of solid waste, structure of landfill and finally the age of landfill (Lopez et al., 2004). The COD of young leachate (leachate from landfill of less than 1-2 years age) is 30-40 times greater than municipal wastewater with the high BOD5/COD ratio (>0.6) and high concentrations of low molecular weight organics. Therefore biological treatment methods (anaerobic and aerobic) are commonly applied for treatment of young leachate (De Morais and Zamora, 2005). However, biological methods do not have enough efficiencies for treatment of mature leachates (leachate from landfills of more than 5-10 years age), due to their low BOD5/COD ratios (<0.3) and high fraction of high molecular weight, refractory organics (Lopez et al., 2004). For this reason, there has been increased interest in physicochemical alternative treatment methods for pretreatment and full treatment of mature leachate (Di Iaconi et al., 2006).
The Fenton oxidation process has been extensively studied in recent years and these studies indicate Fenton oxidation process is one of the most cost-effective alternatives for treatment of mature leachate (Deng, 2007; Kajitvichyanukul and Suntronvipart, 2006; Rivas et al., 2003). Typically there are four main stages for Fenton wastewater treatment process includes: oxidation, neutralization, coagulation and solid separation (Gogate and Pandit, 2004).
Fenton reagent is a mixture of H2O2 and ferrous iron, which generates hydroxyl radical according to the following reaction (Kitts et al., 1999):
The ferrous iron initiates and catalyzes the decomposition of H2O2, resulting in the generation of hydroxyl radical. The generation of hydroxyl radical involves a complex reaction sequence in an aqueous solution (Jiang and Waite, 2003).
However, mineralization of contaminant by AOPs often requires long time of reaction and strong doses of oxidant, often less reactive intermediates are produced that inhibit the degradation of parent compound. Once the recalcitrant compounds have been transformed by means of the chemical treatment, their properties including biodegradability are modified. Often the intermediates have reduced recalcitrance with respect to the parent compound (Marco et al., 1997). Since biological treatment is 10-20 times less expensive than chemical processes, integration of chemical oxidation and biological processes can improve the economy of single process. Response Surface Methodology (RSM) is a statistical modeling technique employed for multiple regression analysis. In this method influences of individual factors as well as their complex interactive influences can be evaluated. One of the most important benefits of RSM is the reduction of the number of tests and experiments (Moradi and Ghanbari, 2014; Ahmad et al., 2007). In recent years optimization of Fenton process for different wastewater and synthetic wastewater using RSM have been studied, but only few researchers report about optimization of Fenton process for leachate treatment using RSM, their report focused on COD, BOD5 and TOC reduction or level of mineralization of leachate (Amiri and Sabour, 2014). It is rarely seen a report about biodegradability improvement of leachate by Fenton process that optimized using RSM.
In this study, application of Response Surface Methodology (RSM) and Central Composite Design (CCD) for optimization of Fenton oxidation process for biodegradability enhancement of a mature landfill leachate have been investigated. Central composite design is one of the common methods of RSM model that is useful and applicable for design of experiments (Mason et al., 2003). In this method every parameter defined in five levels and in comparison to other experiments design methods have more ability to predict responses.
The main objectives of this study were the determination of Fenton process potential for biodegradability enhancement of mature landfill leachate and to optimize Fenton process for biodegradability improvement of mature leachate using RSM.
MATERIALS AND METHODS
Landfill leachate was collected from the Saravan County Landfill in the city of Rasht in northern Iran which was 10 years landfill. The collected leachate was stored in a plastic bottle at 4°C until use.
For the removal of suspended materials, all samples were filtered through a 40 micron glass filter paper.
All reactions were performed in batch reactor. The reactor was 1 L cylindrical glass, mixing was done by a plastic paddle jointed to an epoxy coated shaft, the shaft was connected to a variable speed electromotor and agitation rates of 180-190 rpm were used in all testes. The pH was controlled by Orion pH meter. Each test began with 500 mL of mature leachate. Each initial pH was adjusted using sulfuric acid.
Required FeSO4.7H2O was added in to the solution and Fenton reaction was started up after adding H2O2. Effluent was collected at a predetermined reaction time. Potential of hydroxyl radical reaction in the collected effluent was quenched by 6N NaOH and 10% Na2SO3.
For the study on Fenton oxidation potential for biodegradability enhancement of leachate many parameters such as BOD5, COD, pH, H2O2, Fe++, oxidation time in 32 runs were evaluated, 12 run as pretests and 20 run for optimization. The BOD5/COD ratio has been used as biodegradability indicator. Definitively BOD5/COD is indicator of biodegradability (Singh, 2012).
Design Expert7 was used for statistical analysis of experimental data. To optimize three main variables (pH, H2O2/Fe+2 ratio and reaction time) Response Surface Methodology (RSM) and Central Composite Design (CCD) were used. As shown in Table 1 independent variables including the pH, H2O2/Fe+2 ratio and reaction time were coded in five level as A, B and C, respectively. Preliminary tests were done for determination of levels before designing the research. The COD (Y1), BOD5 (Y2) and BOD5/COD ratio (Y3) were evaluated as important parameters for the evaluation of biodegradability enhancement. The number of tests was calculated from equation as k is the number of factors:
CCD = 2k+2k+6
Six tests were repeated in central point for evaluation of pure error. Quadratic model as Eq. 1 has been used for analysis of experimental data and prediction of optimum condition:
| (1) |
| Y | = | Response |
| I | = | Linear constant |
| j | = | Quadratic constant |
| β | = | Regression constant |
| k | = | Number of studied factors and |
| e | = | Random error |
| Table 1: | Independent variables and their levels in the experimental design as actual and coded |
Analysis of variances (ANOVA) was used for graphical analyses of the data to obtain the interaction between the process variables and the responses.
The quality of model fitness determined by R2 (coefficient of determination) and p-value with confidence level of 95% was used for influences of individual variables as well as their interactive influences evaluation. To control the significance and Adequacy of model, calculated F-value (Fisher variation ratio), probability value and Adequate precision were evaluated. Finally to find optimum condition of Fenton process for biodegradability enhancement of leachate, response variables have been compared in an overlaying three dimensional graph.
Analytical methods: All chemicals were purchased from MERK and ALDRICH Company. All chemicals were of analytical grade, 98% and higher (except for H2O2 which was 33%). The Chemical Oxygen Demand (COD) and Biochemical Oxygen Demand (BOD5) were measured to investigate the biodegradability enhancement of leachate by Fenton reagent. The pH was measured by Orion pH meter 420 A model and COD and BOD5 were analyzed according to the standard, 5220-C and 5210-B methods, respectively (APHA et al., 1995).
RESULTS AND DISCUSSION
Landfill leachate characteristics are given in Table 2. It can be seen that the mean BOD5/COD is less than 0.01.
The low BOD5/COD ratio and high concentration of NH3-N and alkalinity as shown in Table 2 indicate that it is a mature, stabilized and non biodegradable leachat (Zhang et al., 2012). Since the Fenton oxidation efficiency is strongly dependent to H2O2/COD ratio therefore this parameter was evaluated and the optimal H2O2/COD ratio of 2.12 g g–1 was obtained which was according to the results of puangrant and Nattapol study on biodegradability of hospital wastewater (Kajitvichyanukul and Suntronvipart, 2006). In all experiments this ratio was constant. In this experimental study 20 run of CCD were performed that results are shown in Table 3.
Table 4 shows regression results of response surface model for COD reduction, BOD5 and BOD5/COD variation. Results from Table 4 have been used to determine models coefficients in Eq. 2-4.
| Table 2: | Landfill leachate characteristics |
| Table 3: | Central composite face-centered design and experimental results |
| COD: Chemical oxygen demand, BOD5: Biochemical oxygen demand | |
| Table 4: | Regression results of response surface model for COD reduction, BOD5 and BOD5/COD variation |
| Table 5: | ANOVA of quadratic model for COD reduction, BOD5 and BOD5/COD variation |
Significant at p<0.05,nsNot significant at p<0.05, AP: Adequate precision, PRESS: Predicted residual error sum of squares |
|
Table 5 shows ANOVA results for predicted parameters regression of response level of quadratic model for COD removal, BOD5 and BOD5/COD ratio variations.
As shown in Table 3 the lake of fit test for COD, BOD5 and BOD5/COD variations were 1.68, 2.1 and 0.59, respectively, therefore, it is founded that F test insignificant, implying a significant model correlation between variables and COD, BOD5 and BOD/COD percentage variations as process responses.
In this study 0.04-17% of total variation did not explain by model. High R2 coefficient (R2>0.8) shows a satisfactory adjustment of quadratic models to experimental results (Montgomery, 2009). Adjusted R2 values for COD, BOD5 and BOD5/COD variations were 0.93, 0.68 and 0.83 it can be seen a strong significance of model.
| Fig. 1(a-c): | Actual and predicted removal (a) COD, (b) BOD5 and (c) BOD5/COD |
Based on the sequential model sum of squares, the models COD removal, BOD5 and BOD5/COD ratio variations were selected based on the highest order polynomials where the additional terms were significant and the models were not aliased. The models were coded as Y1, Y2 and Y3 COD, BOD5 and BOD5/COD ratio, respectively. The quadratic model for all three terms Y1, Y2 and Y3 were selected as suggested by the software and are shown in Eq. 2-4. The independent variables in the models were pH, H2O2/Fe+2 ratio and reaction time. All insignificant terms from statistical point of view were removed and the empirical models for each variable resulted as follows:
| (2) |
| (3) |
| (4) |
As shown in Table 5 all obtained quadratic models were significant (F<0.007, prob>F). Adequate Precision (AP) of model for COD removal, BOD5 and BOD5/COD ratio variations were 14.3, 7.32 and 9.8, respectively. They were greater than 4 that indicated adequate signal and it is concluded that predicted data could be used to navigate the design space defined by the CCD (Mason et al., 2003). As shown in Table 5 the lake of fit test for COD, BOD5 and BOD5/COD variations were 1.68, 2.1 and 0.59, respectively, therefore it is founded that F-test insignificant, implying a significant model correlation between variables and COD, BOD5 and BOD5/COD variations as process responses Adequacy of model could be assessed by diagnostic plot such as normal probability plots of Studentised of residual and predicted versus actual values.
| Fig. 2(a-c): | Normal probability versus studentized residuals (a) COD, (b) BOD5 and (c) BOD5/COD |
Figure 1 shows actual and predicted values for COD removal, BOD5 and BOD5/COD variations. Figure 2 shows normal probability versus studentized residues. As shown in Fig. 1 and 2, there are a strong correlation between real data and the ones obtained from quadratic model about COD removal, BOD5 and BOD5/COD variations and so adequacy of the model was confirmed for prediction of COD removal and BOD5, BOD5/COD variations. Furthermore outlier t-value was about 3.5± that was indicated approximation of model to response surface is acceptable.
| Fig. 3(a-c): | Residuals versus Runs for (a) COD removal, (b) BOD5 variation and (c) BOD5/COD ratio |
Figure 3 shows residuals versus runs for COD removal BOD5 and BOD5/COD variations. These plots can be used to study independence of residuals. If there were sinusoidal variations therefore, data independence would be acceptable (Amr et al., 2013; Mason et al., 2003). As shown in Fig. 3, there are no evidences for rejection of data independence hypothesis and this is concluded that selected model is suitable for data analysis.
Optimization of operating conditions: In this study factors including the pH, H2O2/Fe2+ and oxidation time were selected in range and responses including COD removal, BOD5 removal and BOD5/COD ratio were selected as maximize. Results from numerical optimization showed that in pH = 3.3, H2O2/Fe2+ = 15.82 and oxidation time of 130 min, COD removal, BOD5 and BOD5/COD ratio variation were 93.1, 0.56 and 0.49%, respectively. In this condition desirability was 100% (Fig. 4). These results about pH were consistent with those of (Yoo et al., 2001) but about H2O2/Fe2++, results were not adapted to results of them (Yoo et al., 2001).
| Fig. 4: | Optimal conditions for COD removal, BOD5 and BOD5/COD variations |
| Table 6: | Verification experiments at optimal conditions |
Five excess tests were performed at the optimum point to confirm adequacy and validity of optimization procedure. The results of these tests are given in Table 6. As shown in Table 6 results from experiments will agree to results from model. This agreement confirms accuracy and approach of model.
Response surface of quadratic models are given in Fig. 5-9. To determine effect of experimental parameters on process response, in optimal condition one parameter was selected at midrange and variation of the tow other parameters was evaluated. Results of response surface of quadratic model are given in Fig. 5.
Figure 5 shows contour plot and 3D plot of COD removal as a function of H2O2/Fe2+ ratio and pH at the oxidation time of 90 min it can be seen that pH significantly influenced the COD removal and the more pH from optimal point (3-3.5) the less COD removal achieved (p<0.05). This supports results reported in many previous studies which supported by Zhang et al. (2012), Ghafari et al. (2009) and Farrokhi et al. (2003).
At the pH below optimum conditions COD was not reduced significantly by Fenton reagent because H2O2 was not decomposed by ferrous ion at the pH below optimal and so there would be no H2O2 decomposition, no hydroxyl radical generation. This fact proved by constant oxygen level (DO = 0.61 at the initial of reaction and DO = 0.73 at the end of reaction). Also hydrogen ions could be played a scavenging role for hydroxyl radical according to the following reaction (Barbusinski et al., 2006):
| Fig. 5(a-b): | (a) COD removal as a function of H2O2/Fe2+ ratio and (b) Counter plot and 3D plot of ph at the oxidation time of 90 min |
| Fig. 6(a-b): | Contour plot and 3D plot of (a) Oxidation time at the pH = 3.5 and (b) COD removal as a function of H2O2/Fe2+ ratio |
When the pH reaches to above optimal the COD removal significantly decrease that it can be due to decrease dissolved fraction of iron, resulting relatively slow reaction between Fe2+ and H2O2. Furthermore the oxidation potential of hydroxyl radical will be decrease with increasing pH. This was consistent with other researchers report that Fenton oxidation had high removal efficiency at the pH = 3-3.5 (Farrokhi et al., 2003; Ghafari et al., 2009; Singh and Tang, 2013).
| Fig. 7(a-b): | (a) BOD5 variation as a function of H2O2/Fe2+ and (b) pH at the oxidation time of 90 min |
| Fig. 8(a-b): | (a) BOD5 variation as a function of H2O2/Fe2+ and (b) Oxidation time at the pH = 3.5 |
As shown in Fig. 5, H2O2/Fe2+ ratio had similar effect on COD removal. At the optimal condition of H2O2/Fe2+ = 17.5, COD removal was 95% and above this ratio COD removal was decreased significantly. There is a transparent optimal zone in plot. This phenomenon could be explained by scavenging effect that occur in the presence of excess reagent (Amiri and Sabour, 2014). As shown in this plot the interactive effect of pH and H2O2/Fe2+ ratio is significant for COD removal. This result was consistent with previous studies that have indicated H2O2/Fe2+ ratio has an important role in Fenton oxidation efficiency (Lak et al., 2012).
| Fig. 9: | BOD5 /COD variation as a function of H2O2/Fe2+ and pH at the oxidation time of 90 min |
It may be observed that H2O2 doses significantly influenced the degradation of COD and the maximum COD removal was obtained in the molar ratio of H2O2/Fe2+ of 17.5. Optimal molar ratio of H2O2/Fe2+ of 17.5 was not adapted to results of Yang Deng study which was 3 (Deng, 2007). It can be observed that COD removal efficiency dropped when H2O2/Fe2+ ratio increase from 17.5 to above. This is due to the scavenging effects of peroxide and Fe2+ on hydroxyl radicals according to the below reactions, respectively:
Interactive effects of oxidation time and H2O2/Fe2+ ratio at the constant pH of 3.5 are given in Fig. 6. it can be seen that COD removal will increase until H2O2/Fe2+ reached to 17.5 and oxidation time reached to 120 min in optimal condition COD removal efficiency was 92%. With attention to Fig. 6 after optimal zone there is a revers relationship between COD removal and H2O2/Fe2+ ratio and oxidation time. There is transparent lump in plot that shows interactive effect of independent variables (oxidation time and H2O2/Fe2+) is significant. All p-values are given in Table 3.
Figure 7 indicates BOD5 variation as a function of H2O2/Fe2+ and pH at the oxidation time of 90 min as shown in Fig. 7, the optimum condition for BOD5 variation was pH = 3-3.5 and H2O2/Fe2+ = 17.5 as BOD5 concentration was maximized. It can be seen clearly that interactive effect of pH and H2O2/Fe2+ on BOD5 variation is significant. All P values are given in Table 3.
As shown in Fig. 8 BOD5 will increase when oxidation time increase to 90 min, after this oxidation time BOD5 will decrease with increasing the time.
The study on biodegradability enhancement of mature leachate was the main objective of this research, therefore, the changes of BOD5/COD ratio as biodegradability indicator was evaluated in the Fenton oxidation process in the optimal H2O2/Fe2+ molar ratio of 17 and different oxidation time.
| Fig. 10: | Design expert overlay plot for optimal region using fenton reagent |
As shown in Fig. 6, BOD5/COD ratio increased with the time and the best BOD5/COD ratio of 0.45 is achieved at the time of oxidation of 130 min. In comparison to the results of Derco et al. (2001), it is observed that Fenton oxidation is more efficient than ozonation for biodegradability enhancement of landfill leachate.
Derco results indicated that BOD5/COD ratio increased from zero up to a maximum value of 0.21 using ozonation (Derco et al., 2001). The BOD5/COD ratio and COD removal efficiency are reached to above 0.35 and 80%, respectively that these observations are consistent with the shrawan study (Singh and Tang, 2013).
More biodegradability enhancement may be achieved at the longer oxidation time and the higher concentration of oxidation reagent (more than H2O2/COD ratio of 2.12), but it should be considered that it will be decrease the organic availability of leachate for the next biological treatment step and the cost of treatment will be strongly increased.
Figure 9 indicates 3D plot and contour plot of BOD5/COD ratio variation as a function of pH and H2O2/Fe2+ ratio at the oxidation time of 90 min. It can be seen BOD5/COD ratio will be increased when pH reached to 3.6, after this point BOD5/COD ratio will be decreased. The similar influence was observed about H2O2/Fe2+ ratio, the maximal BOD5/COD ratio was achieved at the H2O2/Fe2+ ratio of 17.5 and after this, it was decreased. There is a transparent zone about BOD5/COD variation in these plots that show interactive influences of pH and H2O2/Fe2+ ratio on response surface are significant. Overlying the super imposing contour plots of response surfaces can be used to find optimum condition for all parameter (Montgomery, 2009). Graphical optimization shows the possible value of response levels in operative region as zone of fitness with optimization criteria as darker. Approximated selective response for each parameter as the minimum removable values that have overlying zone with other response was selected 70, 0.30 and 0.45% for COD removal and BOD5and BOD5/COD ratio variation, respectively. These zones are near to points that have maximum values of removal efficiency and variations, therefore, a small zone will be selected as optimal zone.
Darker zone of Fig. 10 shows the optimum condition of Fenton oxidation for biodegradability enhancement of leachate. In this optimal condition it can be seen that oxidation time is 130 min, pH = 3-3.5 and H2O2/Fe2++ ratio = 14-17. This is complies with the results of other study that report oxidation time of 120 and pH = 3.5 (Singh and Tang, 2013).
CONCLUSION
From this study it was concluded that mature landfill leachate biodegradability can be improved by Fenton oxidation process. In the optimal condition of Fenton oxidation as: H2O2/Fe2+ = 14-17, pH = 3-3.5 and oxidation time = 130 min, the BOD5/COD ratio increased to 90 time greater than initial ratio (0.005-0.45).
Results from experiments will agree to results from model. This agreement confirms accuracy and approach of obtained model, therefore, it was concluded that response surface methodology can be used to optimize Fenton oxidation process for leachate treatment.
ACKNOWLEDGMENT
The authors gratefully thank the experts of the Laboratory of Wastewater of Health School of Guilan province for their valuable contribution.