14.2.1 History of Landfill Leachate Quality

Table 14.1 summarizes the current understanding of leachate quality which is critical in the selection of treatment processes during design process.

14.2.2 Leachate Characteristics

An anaerobic degradation scheme for organic material in a sanitary landfill is shown in Figure 14.1. In young landfills containing large amounts of biodegradable organic matter (OM), rapid anaerobic fermentation takes place, resulting in volatile fatty acids (VFA) as the main fermentation products. Acid fermentation is enhanced by high moisture or water content in the solid waste. This early phase of a landfill’s lifetime is referred to as the acidogenic phase and leads to the release of large quantities of free VFA, as much as 95% of the organic content. As a landfill matures, the methanogenic phase occurs. Methanogenic microorganisms develop in the waste, and the VFA are converted to biogas (CH₄, CO₂). The organic fraction in the leachate becomes dominated by refractory (nonbiodegradable) compounds such as humic substances (Figure 14.1).

Landfills undergo at least five stages of waste degradation, accompanied by the formation of various compounds and emissions: (i) for aerobic, water and carbon dioxide are the main products, with carbon dioxide released as gas or absorbed into water to form carbonic acid; (ii) for acidogenic, carbon dioxide, hydrogen, ammonia, and organic acids; (iii) for acetogenic, acetic acid and derivatives, carbon dioxide, and hydrogen; (iv) for methanogenic, typical landfill gas has approximately 60% methane and 40% carbon dioxide; and (v) for aerobic, carbon dioxide and water. The acidogenic phase occurs when the landfill containing large amounts of biodegradable OM is a few years old, resulting in the so‐called young leachate with high values of chemical oxygen demand (COD) and biologic oxygen demand (BOD) levels, while the ratio BOD/COD is higher than 0.7 and pH is at low value due to the high concentrations of VFA. Acidic fermentation is enhanced by a high moisture or water content in the solid waste. The methanogenic phase is specific to landfills older than 10 years, and the leachate generated is referred to as old. Methanogenic microorganisms developed in the waste degrade VFA, which are converted to biogas (CH₄, CO₂), while pH >7 and the ratio BOD/COD is stabilized on levels. Refractory (nonbiodegradable) compounds such as humic substances become the dominant organic fraction in the leachate.

There are many factors affecting the quality of leachates, such as landfill age, precipitation, seasonal weather variation, waste type, and composition. In particular, the composition of landfill leachates varies greatly depending on the age of the landfill (Silva et al., 2004). The characteristics of the landfill leachate can usually be represented by the basic parameters COD, BOD, the ratio of BOD/COD, pH, suspended solids (SS), ammonium nitrogen (NH–N), total Kjeldahl nitrogen (TKN), and heavy metals. Table 14.2 presents typical physicochemical characteristics of leachate. The leachate is usually slightly alkaline pH (>7), dark in color, low in BOD₅ (151 mg/l), and high in COD (2183 mg/l) with a BOD₅/COD ratio of 0.07, consistent with typical characteristics of stabilized leachate.

Using the EDS analysis of a transmission electronic microscope, the elements of raw leachate are presented in Table 14.3. The energy‐dispersive spectroscopy (EDS) analysis for the elemental composition of leachate sample (Table 14.3) showed highest concentration for the elemental carbon (35 ± 11% (w/w)) that included inorganic as well as organic carbon. It shows that carbon, oxygen, sodium, and chlorine occupy 35.3, 21.5, 27.4, and 12.4% of major weight of leachate sample, respectively.

14.4.1 Kinetic Model of DMPO–OH EPR Signal

Hydroxyl radicals generated from the Fenton reaction (Reaction 14.1) could also react with Fe²⁺ and H₂O₂ as described in termination Reactions (14.2) and (14.3). However, when DMPO is in excess, all the hydroxyl radicals generated by H₂O₂ and Fe²⁺ would quickly attack the electron‐rich carbon of the DMPO in Reaction (14.4) to form the classic paramagnetic compound 2‐hydroxy‐5,5‐dimethyl‐1‐pyrrolidinyloxy (DMPO–OH; see Scheme 1, compound 2), thus limiting side Reactions (14.2) and (14.3):

(14.6)

where k_D is the formation rate constant of paramagnetic DMPO–OH compound.

The stability of DMPO under attack from a hydroxyl radical depends upon the relative amount of DMPO and Fenton reagent concentrations. In the presence of high concentration of Fe³⁺ under acidic conditions, it was previously reported that DMPO–OH could lead to the formation of two diamagnetic intermediates (Makino et al., 1992): 1‐hydroxy‐5,5‐dimethyl‐1‐pyrrolid‐2‐one (HDMPN, Scheme 1, compound 3) and its tautomer 2‐hydroxy‐5,5‐dimethyl‐1‐pyrroline‐N‐oxide (HDMPO, Scheme 1, compound 4), as described in Reaction (14.7):

(14.7)

where k_I is the formation rate constants for diamagnetic tautomers HDMPN and HDMPO.

Moreover, if there is an excess of hydroxyl radicals in the system, these EPR‐silent compounds quickly react with •OH to form the stable paramagnetic radicals 5,5‐dimethyl‐2‐oxopyrroline‐N‐oxyl (DMPOX, Scheme 1, compound 5) and 2‐dihydroxy‐5,5‐dimethyl‐1‐pyrrolidinyloxy (HDMPO–OH, Scheme 1, compound 6) (Reactions 14.8 and 14.9):

(14.8)

(14.9)

where k_H and k_X are the formation rate constants for the paramagnetic products HDMPO–OH and DMPOX, respectively.

Therefore, if DMPO is not present in extra‐excess, additional EPR signals might appear, thus preventing a reliable detection/quantification of hydroxyl radicals. The critical question to elucidate is therefore at which concentration DMPO could be considered to be in excess. For this purpose, a kinetic model was developed based on the previous elementary reactions. With the aid of this model, experiments were designed to answer this critical question by monitoring DMPO–OH and additional signals using an EPR and UV‐Vis spectrophotometer. When the ratio of H₂O₂ and Fe²⁺ was kept at an optimal 10 (Tang, 2004), the conditions under which DMPO is in excess, i.e. trapping all the hydroxyl radicals in the system, without further Reactions (14.8) and (14.9), can be found starting from Reaction (14.10):

(14.10)

At a steady‐state concentration of hydroxyl radical, Equation 14.10 should equal zero to derive the steady‐state hydroxyl radical concentration. The final form of hydroxyl radical steady‐state concentration can then be expressed as follows:

(14.11)

To accurately quantify DMPO–OH concentration, the analytical conditions must be found and defined so that a stable quartet EPR signal can be monitored over the analytical time without further degradation to HDMPN. In terms of elementary reactions, the rate of Reactions (14.8) and (14.9) should approach zero as follows:

(14.12)

(14.13)

Therefore, the steady‐state hydroxyl radical concentration can be simplified as follows:

(14.14)

Moreover, the reaction rate of DMPO–OH can be expressed as follows:

(14.15)

The generation of a constant quartet EPR signal relies upon two conditions: the first is that the rate of reaction (14.7) approaches zero. Therefore,

(14.16)

The second condition is to make sure the change rate of d[DMPO–OH]/dt is not a function of [DMPO]. The only possible way to satisfy this condition is for the term k_D[DMPO] to be significantly greater than k_t1[Fe²⁺] + k_t2[H₂O₂] so that the latter can be negligible. Mathematically, the following conditions must be satisfied:

(14.17)

This equation can be divided by k_D[H₂O₂] at both sides, and the suitable analytical conditions become

(14.18)

The sign of significantly greater means that the right‐hand ratio of reaction (14.18) should be negligible compared with the left‐hand ratio. In other words, the left‐hand ratio should be at minimum two to three orders smaller than the right‐hand ratio for Equations (14.17) and (14.18) to truly hold. Therefore, at the initial H₂O₂ and Fe²⁺ concentrations of 1 and 0.1 mM (ratio of 10), respectively, the required ratio of initial DMPO to H₂O₂ concentrations, [DMPO] : [H₂O₂], should be at minimum 1.68–16.8. As a result, the system should generate a stable quartet EPR signal without any additional noise. Fontmorin et al. (2016) designed their experiments based upon these theoretical analyses, and results were used to determine the optimal [DMPO] : [H₂O₂] ratio more accurately. In this study, H₂O₂ : Fe²⁺ ratio was kept constant; however, the aforementioned model would help to predict an optimal concentration of DMPO for any other concentration of the Fenton reagent since initial conditions are included in Equation (14.18).

Fontmorin et al. (2016) showed the impact of the Fenton reagent concentration with H₂O₂ and Fe²⁺ concentrations varying from 1 to 100 mM and from 0.1 to 10 mM, respectively, with [DMPO] : [H₂O₂] ratio varying from 1 to 100. A typical spectrum of DMPO–OH is presented in Figure 14.2a: it is composed of a characteristic 1 : 2 : 2 : 1 quartet with hyperfine couplings a_N = a_βH = 1.50 mT (computer simulation, Figure 14.2b).

The amplitude of the DMPO–OH signal was followed over time at different concentrations of Fenton’s reagent, and results are presented in Figure 14.3.

As depicted, the concentration of Fenton’s reagents had a significant impact on both the intensity and stability of DMPO–OH signal. Indeed, when H₂O₂ and Fe²⁺ concentrations were 1 and 0.1 mM, respectively (i.e. [DMPO] : [H₂O₂] = 100 > 16.8), the peak amplitude of DMPO–OH signal was statistically stable because no clear downward trend appears over the time of experiment (60 min). The variations from one point to another could be attributed to the instability of the reaction (no control of pH after the initial adjustment). Nevertheless, the general trend of the system together with the nature of the EPR background (DMPO–OH quartet, data not shown) suggested an appropriate concentration in DMPO compared with the concentration of Fenton reagent for the detection of radicals, as predicted by the theoretical model. The increase in Fenton reagent concentration to 10 mM H₂O₂ and 1 mM Fe²⁺ (i.e. [DMPO] : [H₂O₂] = 10 > 1.68) led to a different behavior. The higher intensity measured during the very early minutes suggested the generation of higher concentration of hydroxyl radicals, as expected after the increase of the Fenton reagent. However, an important decrease in the DMPO–OH peak amplitude was then observed during the first 20 min, followed by a slower decrease until the end of the experiment. The fast decrease measured during the early stage of the reaction was unlikely to have been related to the lifetime of DMPO–OH, generally reported around 30 min. Even though the typical DMPO–OH quartet was the only signal detected in EPR spectra without the apparition of any additional line, the fast decrease of the signal intensity suggested the possible quenching or degradation of the spin adduct. According to the kinetic model developed, this observation would also suggest that the left‐hand ratio of Equation (14.18) should be more than two orders greater than the right‐hand ratio to provide a stable quartet signal. To confirm the possible quenching of DMPO–OH, H₂O₂ and Fe²⁺ concentrations were increased again to 100 and 10 mM, respectively ([DMPO] : [H₂O₂] = 1). In this case, the quenching effect of DMPO–OH was very obvious, with a lower signal amplitude from the beginning of the reaction and becoming hardly detectable after 10 min, thus clearly suggesting the oxidation of the spin adduct due to the excess of hydroxyl radicals. Moreover, the EPR revealed the appearance of additional lines consisting a 1 : 1 : 1 triplet. The triplet’s intensity increased along with the decrease of DMPO–OH quartet (Figure 14.4a). After a 35‐min reaction, the peak amplitude of the triplet reached its maximum, whereas DMPO–OH had completely disappeared. The corresponding EPR spectrum is presented in Figure 14.4b (a_N = 1.51 mT, simulation depicted in Figure 14.4c).

The triplet observed remained stable until the 60‐min reaction, although its amplitude slightly decreased. It should be noted that the appearance of background EPR signals in case of an excessively high initial concentration of Fe²⁺ as described (Bosnjakovic and Schlick, 2006). However, the signals reported and already attributed to HDMPO–OH and DMPOX are different than the triplet observed by Fontmorin et al. (2016), thus implying the generation of a different by‐product. To understand the mechanisms leading to the decrease of DMPO–OH in the aforementioned two systems and to the formation of the triplet in the last system, similar analyses were performed by UV‐Vis spectrophotometry. The contribution of each reagent (H₂O₂, Fe²⁺ and DMPO) was first analyzed, and, as shown in Figure 14.5a, the peak measured in visible part (λ_max = 520 nm) only appeared in the presence of both the Fenton reagent and DMPO. Time series of UV‐Vis absorbance were then followed for the last two Fenton systems described previously, and the corresponding results are presented in Figure 14.5b and c. With 10 mM H₂O₂ and 1 mM Fe²⁺, EPR and UV‐Vis signals are following opposite trends. Indeed, the DMPO–OH EPR signal intensity decreased during the first 20 min, whereas the UV‐Vis signal increased during the same period of time (Figure 14.5b). This result suggests the generation of EPR‐silent by‐products (diamagnetic) related to the quenching of DMPO–OH, which can most likely be explained by the reaction of DMPO–OH with the generated Fe³⁺ to form the tautomers HDMPN and HDMPO (Bosnjakovic and Schlick, 2006). This result also confirms that a [DMPO] : [H₂O₂] ratio of 1.68 is not high enough to provide a stable EPR signal. It implies that under these conditions, Reactions (14.8) and (14.9) barely took place in the system, while reaction (14.7) is not negligible. Increasing the concentration of Fenton reagent to 100 mM H₂O₂ and 10 mM Fe²⁺ led to the increase of the absorbance in the visible range over the first 20 min (Figure 14.5c). It can reasonably be assumed that the higher concentration in Fe²⁺, and therefore in the generated Fe³⁺, enhanced the production of the two tautomers.

In the presence of excessive hydroxyl radicals, the oxidation of HDMPN and HDMPO can lead to the formation of HDMPO–OH and DMPOX (Scheme 1). Since DMPOX generates a 1 : 1 : 1 triplet (a_N = 0.72 mT) of 1 : 2 : 1 triplet (a_γH = 0.41 mT, 2H) and that HDMPO–OH generates a 1 : 1 : 1 triplet (a_N = 1.53 mT) of 1 : 3 : 3 : 1 quartet (a_γH = 0.12 mT, 3H) (Bosnjakovic and Schlick, 2006), the presence of HDMPO–OH and DMPOX should be excluded. The computer simulation of the triplet (a_N = 1.51 mT, Figure 14.4b) suggests the absence of the contribution of hydrogen nuclei in β and γ positions. The possibility of degradation of DMPO–OH spin adduct via C–N bond cleavage and the opening of the pyrroline ring have also been suggested and assigned to a similar triplet signal. However, neither a structure nor a mechanism was proposed. Such a degradation pathway was also discussed in a thermodynamic study stating that this mechanism may play a crucial role in the stability of spin adducts. The detailed mechanism related to the DMPO–OH ring opening remains unclear; moreover, the degradation product that could be assigned to the triplet observed (Scheme 1, compound 7) because of the absence of hydrogen nuclei in α and β positions is thermodynamically highly unstable and is thus unlikely to generate a stable signal over time. In the case of 2,5,5‐trimethyl‐1‐pyrroline‐N‐oxide (M₃PO) tautomers, the appearance of the EPR triplet signal could be explained by the dimerization reaction occurring between HDMPN and HDMPO, thus leading to the formation of a stable diamagnetic dimer, as described in the proposed mechanism presented in Scheme 2. In the presence of excessive hydroxyl radicals compared with the initial concentration of DMPO, the dimer could be further oxidized to generate the corresponding paramagnetic compound 8 associated with the additional triplet. Indeed, as shown in Scheme 2, the paramagnetic dimerization product does not present hydrogen nuclei in β position, and the contribution of hydrogen in γ position is highly unlikely due to steric effect after dimerization (Figure 14.6).

According to these experimental results, the theoretical model proposed can be adjusted as follows: to ensure a reliable detection of hydroxyl radicals, the left‐hand ratio of Equation (14.18) should be three orders higher than the right‐hand ratio. In other words, the [DMPO] : [H₂O₂] ratio should be at least 16.8. The observed reactions described earlier suggest that the DMPO concentration should therefore be carefully selected according to Fenton’s reagent concentration and expected magnitude of hydroxyl radical concentration as demonstrated in the previous theoretical kinetic analysis. The condition illustrated by Equation (14.18) could be applied to any concentration of the Fenton reagent during EPR quantification hydroxyl radical concentration as standard calibration procedure. Therefore, the impact of DMPO concentration on EPR signals was validated at a constant ratio of H₂O₂/Fe²⁺ but different the Fenton reagent concentrations. Initial concentration of H₂O₂ and Fe²⁺ was chosen at 5 and 0.5 mM, respectively. DMPO concentration varied from 5 to 100 mM, corresponding to the ratio of [DMPO] to [H₂O₂] varying from 1 to 20. The time series signals of DMPO–OH are presented in Figure 14.7a. As observed, DMPO concentration had a significant impact on DMPO–OH signal intensity. Using DMPO 100 mM, which is 20 times the initial H₂O₂ concentration, the signal intensity remained statistically stable until 25 min (apart from the instability of the system itself). The peak‐to‐peak amplitude then decreased slowly until the end of the experiment, which is consistent with the lifetime of DMPO–OH. As shown in Figure 14.7b, the corresponding EPR background after 10 min only composted of the DMPO–OH quartet, showing that no other paramagnetic by‐product was formed and that the adduct was not further oxidized. Decreasing DMPO concentration from 100 to 50 mM ([DMPO] : [H₂O₂] = 10) also led to a decrease in DMPO–OH signal intensity, even though the 1 : 2 : 2 : 1 quartet was the only signal detected (Figure 14.7c).

Therefore, under these conditions, it can be assumed that DMPO–OH is partially degraded to HDMPN and HDMPO and that hydroxyl radicals are not in excess to generate the paramagnetic dimer. Finally, when DMPO concentration was lowered to 5 mM (DMPO] : [H₂O₂] = 1), the EPR background evolved into a combination of quartet and triplet (Figure 14.7d). These results are consistent with both the theoretical model and the results presented in Figure 14.7. It confirms that in order to achieve a reliable detection of hydroxyl radicals, the concentration ratio between DMPO and Fenton reagent should be carefully chosen, which is true for any concentration of Fenton reagent. Indeed, it should be noted that with the most concentrated system, increasing DMPO concentration from 0.1 to 0.5 M led to a significant decrease in the triplet and an increase in the quartet. The influence of DMPO concentration on the stability of EPR signal shows that the “excess concentration” in DMPO should be carefully adopted to each system, even in the absence of additional lines. The absence of additional lines does not necessarily imply a reliable detection of hydroxyl radicals, since EPR‐silent compounds might be formed. Therefore, three conclusions can be drawn: (i) the [DMPO] : [H₂O₂] : [Fe²⁺] ratio holds the key in terms of the oxidation kinetics of DMPO–OH, and at a fixed ratio, the kinetics are the same; (ii) the stability of DMPO–OH depends upon Fe³⁺ and ^•OH concentrations wherein the higher these concentrations, the shorter the lifetime of DMPO–OH; and (iii) to ensure a reliable detection of hydroxyl radicals, DMPO concentration should be set so that [DMPO] : [H₂O₂] > 16.8.

14.6.1 Fenton Oxidation of Landfill Leachate

FO of leachate is a multistep process. At first, leachate pH is lowered to the range of 2.5–4.0 because in this pH range ^•OH is more efficiently generated in addition to keeping Fe²⁺ and Fe³⁺ ions soluble (Kochany and Lipczynska‐Kochany, 2009). Sulfuric acid is generally used to lower the pH and added at the start of reaction or continuously controlled the pH at the desired level (Zhang et al., 2006). Afterward, Fe²⁺ salt (ferrous sulfate) is added to the leachate and mixed at 80–400 rpm followed by gradual addition of H₂O₂ while continuing the rapid mixing for 30 s to 60 min. H₂O₂ is commonly added in single step (Lopez et al., 2004; Mohajeri et al., 2010). The effect of multistep H₂O₂ addition on leachate oxidation has also been reported (Rivas et al., 2003; Goi et al., 2010). At the end of this stage, pH is increased to the neutral pH range using sodium hydroxide or lime solution. Under neutral pH conditions, the Fe³⁺ forms ferric‐hydroxo complexes, and a slow mixing (20–100 rpm) of solution for 10–30 min flocculates ferric‐hydroxo complexes that subsequently precipitate for approximately 30 min to several days (Kim and Huh, 1997; Cotman and Gotvajn, 2010). The analysis of supernatant provides the treatment efficiency of the process.

FO has great potential in removing OM that contributes to the COD and color of landfill leachate. A statistical analysis of the data of leachate treatment efficiency and the optimum operating conditions as obtained from peer‐reviewed publications on FO landfill leachate is presented in Table 14.5 by Singh and Tang (2013). Raw and pretreated leachates have been subjected to FO. Biological treatment and coagulation are the most common pretreatment processes used with FO. Biological treatment processes effectively remove low MW OM, whereas FO removes high MW refractory OM (Gulsen and Turan, 2004). FO has been generally used as a polishing step with coagulation treatment due to its effectiveness in removing similar types of OM (Goi et al., 2010).

Raw and biological‐ and coagulation‐treated leachate showed a median COD removal of 66, 71, and 73%, respectively. Reduced COD₀ values of pretreated leachates lead to an increased percentage of COD removal as compared with raw leachate (Kang and Hwang, 2000); however, the percentage of COD removal does not necessarily translate to the amount of COD removed. Zhang et al. (2006) observed increased amounts of COD removal at higher COD₀, but the percentage of COD removal efficiency decreased under the same Fenton‐operating conditions. An average of 89% color removal was observed. FO greatly improved biodegradability of leachates, which can be represented by the ratio of BOD₅ and COD (BOD/COD). An increase from less than 0.01 to a range of 0.15–0.70 with a median increase of 0.50 was observed for BOD/COD values, which shows FO suitability as a pretreatment process for the biological treatment of landfill leachate.

FO may also help reduce the toxicity of leachate. Goi et al. (2010) observed a twofold toxicity reduction for Daphnia magna, and Cotman and Gotvajn (2010) observed 24 and 32% reduction for heterotrophic and nitrifying microorganisms, respectively; however, the removal was not significant enough, and leachate remained extremely toxic. Although ^•OH has a strong oxidizing capacity, the generated ^•OH during FO cannot oxidize ammonia‐N that largely contributes to the toxicity of leachate (Gotvajn et al., 2011). Therefore, FO cannot be used as a primary method to remove the toxicity of leachate. Addition of Fe²⁺ salt and chemicals used for pH adjustment also increases the total dissolved solids (TDS) of treated leachate (Deng and Englehardt, 2006).

14.6.2 Optimization Fenton Oxidation of Leachate

Landfills are designed to dispose high quantities of waste at economical cost with potentially fewer environmental effects; however, improper landfill management may pose serious environmental threats through the discharge of high‐strength leachate. Although modern landfill engineering and operating practices minimize leachate generation, proper leachate discharge remains one of the great challenges for landfill operators due to its complicated and changing characteristics over time. Multiple factors influencing leachate characteristics include waste age, climatic conditions, waste composition, landfill design, and operating practice; however, selecting leachate treatment method, which is often a combination of multiple processes, is mostly associated with the state of landfill stabilization.

Leachate treatability is evaluated as a function of landfill age or by the ratio of biochemical oxygen demand (BOD₅) to COD with values close to 1, representing young leachate and values of 0.1 or less described as stabilized leachate (Diamadopoulos et al., 1997). Young leachate generally contains low molecular weight and more biodegradable OM, so it is subjected to standard biological treatment processes. Once waste decomposition reaches the stable methanogenic phase, landfill leachate OM is dominated by biologically refractory OM (such as humic and fulvic‐like OM), reducing the effectiveness of biological treatment processes and necessitating other physicochemical processes.

The most common physicochemical processes used for landfill leachate treatment are coagulation and flocculation (Amokrane et al., 1997; Tatsi et al., 2003), activated carbon (AC) adsorption (Maranon et al., 2009; Singh et al., 2012a), chemical oxidation (Rivas et al., 2004; Tizaoui et al., 2007), and membrane‐based technologies (Trebouet et al., 2001; Tabet et al., 2002). These physicochemical processes are often used as pretreatment or posttreatment steps prior to biological processes. Although FO reactions were discovered more than 100 years ago, their application for oxidizing toxic OM in leachate was first reported only in the early 1990s (Huang et al., 1993). Landfill leachate treatment using FO should be carried out under optimal operating conditions such as pH, reaction time, reaction temperature, and reagent dose for a specific leachate to achieve the most cost‐effective treatment. Optimal conditions were generally determined based on OM or COD removal efficiencies from leachate (Deng and Englehardt, 2006); however, a wide range of optimum conditions have been reported in literature primarily due to variation in leachate characteristics and experimental conditions.

Singh and Tang (2013) developed a database for analysis of the optimum FO parameters such as pH, reaction time, reaction temperature, and reagent doses as determined by previous studies along with the experimental conditions such as initial BOD, COD (COD₀), and color and COD removal efficiencies. The reagent weight ratios (H₂O₂/Fe²⁺), reagent doses (H₂O₂/COD₀ and Fe²⁺/COD₀), COD loading factor (L_COD = COD₀/H₂O₂), and oxidation efficiencies (η = COD_ox/H₂O₂) were calculated for each study. The mean, maximum, minimum, standard deviation, and median for each parameter were calculated. A relationship between dimensionless parameters η and L_COD was developed using the optimum operating conditions obtained by previous researchers and can be used to predict η under optimum operating conditions for landfill leachate treatment. The relationship between η and L_COD was cross‐validated using the leave‐one‐out cross‐validation technique, in which one observation is left out at a time, while regression is carried out. The uncertainty in the relationship was evaluated by applying Monte Carlo simulation using Crystal Ball (v11.1) software. During the uncertainty analysis, the mean (η_mean) and standard deviation (σ) of η at a 95% confidence interval for a range of L_COD were determined. The L_COD range of 0.01–34 was used as observed in previous peer‐reviewed studies. The validation and uncertainty analysis procedures were adopted from Tang and Wang (2010).

14.6.3 Optimum Operating Conditions

14.6.3.1 pH

Obtaining an optimum pH is the first most important task in FO process, and typically it varies in a narrow range of 2.5–4.0. The pH below the optimum range affects the oxidation process by producing fewer ^•OH, scavenging ^•OH by H⁺, and reducing Fe²⁺ regeneration (Tang and Huang, 1997). The pH above the optimum range affects the oxidation reaction by reducing the production of ^•OH radicals, self‐decomposition of H₂O₂, formation of ferric‐hydroxo complexes, scavenging of ^•OH radicals by alkalinity (CO₃²⁻ and HCO₃⁻), and reduction in oxidation potential of ^•OH radicals (Buxton et al., 1988).

A statistical analysis of optimum pH conditions determined from the previous studies showed a range of 2.5–6.0 with a median of 3.0 (Table 14.5). An optimum pH range of 2.5–4.5 was observed for raw and coagulation‐treated leachate with a median of 3.0, whereas biologically treated leachate ranged from 2.5 to 6.0 with a median of 3.5. The probable reason for a higher optimum pH of biologically treated leachate is the variability in OM characteristics as compared with raw and coagulation‐treated leachate. Lau et al. (2001) and Li et al. (2010) observed that coagulation dominates the FP as compared with oxidation while using Fenton reagents for biologically treated leachate possibly due to the presence of a higher relative concentration of negatively charged OM as compared with raw and coagulation‐treated leachate.

Figure 14.11 shows the effect of optimum pH conditions on COD removal efficiencies for FO of raw and pretreated leachates. At each optimum pH condition in the range of 2.5–6.0, a wide variability in COD removal efficiency was observed, so one conclusive pH condition is difficult to determine; however, a pH range of 2.5–4.0 showed maximum COD removal efficiencies and varied in the range of 31–95%. The variability in optimum pH conditions among different studies is most likely caused by the differences in experimental conditions and the type of OM present in leachate.

14.6.3.3 Effect of Reaction Time on Fenton Oxidation

The kinetics of OM removal from landfill leachate using FO was evaluated by conducting a Fenton experiment on leachate COD₀ 679 mg/l and L_COD 0.75 at initially adjusted pH of 2.0, 3.5, and 6.0 by Singh et al. (2013). The percentage of COD removal obtained during the FO is shown in Figure 14.12 with a fit to the exponential rise to maximum. Fenton reaction kinetics showed a two‐stage reaction at each pH conditions, and a COD removal of 26–46% was observed within 1 min of reaction that later stabilized. For an initial 1 min (first stage) of FO reaction, the catalyst Fe²⁺ helped in rapidly generating ^•OH from H₂O₂, and ^•OH rapidly oxidized OM present in leachate. However, after 1 min of reaction (second stage), the COD removal was not as fast, and only 5–9% of additional COD removal was observed. At this stage, the concentration of ^•OH reached a steady state due to limited availability of Fe²⁺. Although Fe²⁺ are regenerated during the Fenton reaction by Fe³⁺, the regeneration reaction rate of Fe²⁺ (10⁻³ to 10⁻²/Ms) is approximately four times slower than Fe²⁺ consumption (76/Ms) (Umar et al., 2010). Similar OM removal kinetics have been reported by Zhang et al. (2006). In addition, they observed the first stage of FO reaction for approximately 60 min when the pH was consistently maintained at 3.5. Singh et al. (2013) did not maintain the initially adjusted pH during the oxidation reaction, and a change in pH was reported for an initial 1 min of reaction that later stabilized in the range of 2.3–2.7. The change in pH during the initial 1 min of reaction might have led to shorter FO duration as compared with 60 min as observed by Zhang et al. (2006). The OM removal results also show that the most of the OM removal by oxidation occurred when pH was changing. Very few previously published studies have reported a change in pH during FO of leachate even when pH was not maintained at an initially adjusted level (Wang et al., 2009; Cortez et al., 2011).

The kinetics of OM removal by OH^• can be expressed by pseudo‐first‐order kinetics as shown in Equations (14.19) and (14.20):

(14.19)

(14.20)

where

• C is the concentration of OM represented by COD
• k_obs is the pseudo‐first‐order rate constant
• C₀ and C_t are the concentration of OM at time 0 and t, respectively
• K is a constant dependent on k_obs and amount of OM present in samples

The COD removal kinetic parameters at initially adjusted pH 2.0, 3.5, and 6.0 are shown in Table 14.6. The experimental results show a strong correlation with pseudo‐first‐order kinetics with a coefficient of determination (r²) of 0.99. Zhang et al. (2006) also described a similar observation while studying FO in a continuously stirred tank reactor. Among the three pH conditions evaluated, the greatest rate constant was for pH 3.5, which also complemented what have observed a pH range of 3.0–4.0 as the most common optimum pH condition for the Fenton treatment of leachate (Singh and Tang, 2013).

Initial pH	k	kobs (min−1)	r2
2.0	48.02	2.132	0.995
3.5	50.44	2.691	0.996
6.0	42.49	1.967	0.997

COD removal of 50, 53, and 43% was obtained when pH was adjusted at the start of experiment to 2.0, 3.5, and 6.0, respectively. Although COD removal efficiency was observed to be dependent on the initially adjusted pH in the range of 2–6, the reaction is rather independent of pH when pH falls below 3. In the range of pH 3–4, the reaction rate increases until the reaction reaches a steady state due to the formation of ferrous hydroxides (Fe(OH)₂), which react 10 times faster than Fe²⁺ (Pignatello et al., 2006). Based on the results of kinetic experiments and total COD removal, further experiments were conducted at an initially adjusted pH 3.5.

The increase in reaction time did not always translate into higher COD removal. Gotvajn et al. (2011) obtained 86% COD removal for COD₀ of 3064 mg/l within 10 min of reaction time, whereas Kim and Huh (1997) observed 67% COD removal efficiency for COD₀ of 2000 mg/l in 2 days. The characteristics of raw and pretreated leachates also did not show any conclusive impact on change in the reaction time of the FP for maximum OM removal; however, a reaction time of 10–30 min showed COD removal efficiencies in the range of 60–95%.

14.6.3.6 Generalized Fenton Dosing for Landfill Leachate Treatment

Although studies on FO of landfill leachate in the past were generally focused on determining the optimum Fenton operating conditions, these studies report a large variability in terms of COD removal efficiencies under optimum conditions, so a generalized relationship between the reagent dose and COD removal efficiency was not obtained. Theoretically, 1 mg/l H₂O₂ can oxidize 0.47 mg/l COD, and theoretical η can be defined as Equation (14.21) (Kang and Hwang, 2000):

(14.21)

where η is the oxidation efficiency.

However, this relationship does not always hold true for landfill leachate. When COD₀ is considerably higher than H₂O₂, the observed η may be greater than 100%, but when concentration of H₂O₂ and COD₀ is comparable, η gradually decreases due to the remaining refractory OM that cannot be oxidized or to the scavenging of generated OH• by leachate alkalinity (Kang and Hwang, 2000). Therefore, the selection of an optimum H₂O₂ dose is very much dependent upon the COD₀, and, to determine the η in landfill leachate, it is important to consider the H₂O₂ dose with respect to COD₀.

The impact of L_COD (COD₀/H₂O₂) on η was evaluated from the data obtained from previous studies published in peer‐reviewed literature. L_COD and the oxidation efficiency, η, were linearly correlated, and as L_COD increased from 0.01 to 34, η increased from 0.01 to 24. The linear relationship obtained between L_COD and η as shown in Equation (14.21) had a coefficient of determination (r²) of 0.99:

(14.22)

More importantly, the relationship (Equation 14.22) between L_COD and η holds for FO of raw and biological and coagulation‐treated leachates. Ideally, the slope of Equation (14.22) should be 1; however, due to the presence of •OH scavengers in landfill leachate, the slope of Equation (14.22) was less than 1.

14.6.3.7 Total COD Removal Under Different L_COD

The effectiveness of FO and FP treatment for total COD removal from landfill leachate can be seen for leachate of three different COD₀ 1092, 546, and 273 mg/l under Fenton operating conditions: L_COD in the range of 0.25–1.0, reagent dose ratio (H₂O₂/Fe²⁺) of 1.8 (w/w), and initially adjusted pH 3.5. Figure 14.13 shows that COD removal increased with a decrease in L_COD for all three leachates. Lower L_COD corresponds to higher amount of H₂O₂ used in the reaction that provided more O₂ to oxidize OM; therefore, increased COD removal was observed with the decrease in L_COD. FO in all three leachates showed COD removal in the range of 38–52% at L_COD 0.25, which was approximately twice as much as at L_COD 1.0. COD removal due to FO also showed dependence on COD₀, and leachate with COD₀ 1092 and 273 mg/l showed a COD removal of 52 and 38%, respectively, at L_COD 0.25, i.e. higher OM oxidation was obtained with increased availability of OM. Zhang et al. (2006) also observed a similar trend of increased COD oxidation with the increase in H₂O₂ dose and at higher COD₀.

FP includes oxidation as well as coagulation–flocculation processes. During FO, OM was mineralized or transformed from high molecular weight OM (such as humic and fulvic like) to smaller OM. The smaller OM remains in the solution removed by coagulation and flocculation processes under neutral pH conditions. FP treatment of all three leachate showed more than 70 and 43% of COD removal at L_COD of 0.25 and 1.0, respectively. The highest OM removal was observed for leachate with the highest COD₀. Additionally, the COD/COD₀ after FO and FP treatment of all three leachates was observed to be linearly dependent upon L_COD with a slope of 0.29 and 0.37, respectively, for L_COD in the range of 0.25–1.0. The slope <1 for FO and FP treatments shows that the decrease in L_COD does not proportionally remove OM from leachate because of the presence of OM that are harder to be oxidized by OH radicals, and the amount of H₂O₂ applied at lower L_COD is less efficiently utilized than those at higher L_COD. Although none of the previous studies selected the Fenton reagent doses based on L_COD, mathematical conversion of doses used by Zhang et al. (2006) into L_COD also showed a linear dependence of COD removal on L_COD.

The effect of COD removal by FO, FP, and coagulation–flocculation during the Fenton treatment of three leachates with COD₀ 1092, 575, and 373 mg/l, respectively, was also evaluated for L_COD in the range of 0.25–1.0. As shown in Figure 14.14 by the average COD removal for all three leachates by each process, FO was a more dominant process than coagulation–flocculation at lower L_COD of 0.25, and an average of 46% COD was removed by FO, whereas 25% COD was removed by the coagulation–flocculation process. The possible reason for greater oxidation efficiency at low L_COD was the increased availability of ^•OH within the system. To maintain the constant H₂O₂/Fe²⁺ ratio, a lower amount of Fe²⁺ was used at higher L_COD as compared with Fe²⁺ used at lower L_COD, which caused less OM removal by coagulation processes at higher L_COD. The L_COD 1.0 showed a COD reduction of 15%, which was 10% less COD removal than at L_COD 0.25. Kang and Hwang (2000) also observed increased OM removal at higher Fe²⁺ doses in the coagulation process.

14.6.3.8 Effect of L_COD on COD Removal Efficiency

Leachate with the highest COD₀ (1092 mg/l) showed η_FO in the range of 0.15–0.26 for L_COD variation in the range of 0.25–1.0, in addition to showing relatively greater η_FO among the three leachates at each L_COD (Figure 14.15). For all three initial COD₀ concentrations of leachate, η_FO increases from 0.11 ± 0.03, 0.15 ± 0.05, 0.20 ± 0.02, to 0.22 ± 0.02 at L_COD 0.25, 0.50, 0.75, and 1.0, respectively. The higher COD₀, which corresponds to a relatively greater amount of OM present to be oxidized and higher L_COD, which corresponds to a lower amount of available O₂ or ^•OH for oxidation of OM, leads to the increased possibility of ^•OH radical reaction with OM. As a result, η_FO increases with L_COD. The average η_FO for all three initial COD₀ concentrations of leachate also showed a linear relationship with the increase in L_COD in the range of 0.25–1.0, which implies that the parameter L_COD derived from the COD₀ and applied H₂O₂ dose can help in predicting the OM oxidation efficiency for any leachate.

The COD removal efficiency due to FP (η_FP) also showed greatest utilization of available O₂ at higher COD₀ and at COD₀ 1092 mg/l; the η_FP varied in the range of 0.20–0.49 for L_COD ranging from 0.25 to 1.0. Similar to the FO, the FP also showed an increase in η_FP with the increase in L_COD. At L_COD 0.25, 0.50, 0.75, and 1.0, the average η_FP increases from 0.17 ± 0.02, 0.30 ± 0.03, 0.36 ± 0.03, to 0.47 ± 0.01, respectively. As discussed earlier, the higher the L_COD, the greater the possibility of OM reacting with ^•OH radicals, leading to the completely mineralization of OM or its transformation into smaller OM. These smaller OM may coagulate during the coagulation–flocculation process, thereby showing increased η_FP with the increase in L_COD. Additionally, η_FP was observed to be linearly increasing with L_COD ranging from 0.25 to 1.0 with a slope of 0.38, so the application of L_COD for selecting Fenton oxidant H₂O₂ dose can be useful in predicting the treatment efficiency of landfill leachate.

L_COD and η_FP were plotted against the relationship between L_COD and (Equation 14.25) as obtained by Singh and Tang (2013) for L_COD in the range of 0.25–1.0. For L_COD in the range of 0.25–0.75, was lower as compared with η_FP obtained by experimental data; however, for L_COD in the range of 0.75–1.0, the was higher than the η_FP. The experiments were conducted at pH 3.5 and the H₂O₂/Fe²⁺ dose ratio of 1.8, which is the median of previous studies as reported by Singh and Tang (2013) without determining the optimum Fenton conditions of each leachate. Figure 14.15 shows that the line of η_FP versus L_COD intercepts the line of versus L_COD at L_COD 0.75. This suggests that η_FP is the same as the obtained at the optimum conditions by previous studies at L_COD 0.75. Therefore, L_COD 0.75 could be used as an optimum dose for all three different initial COD₀ of leachate. As a result, it could be concluded that the universal relationship between and L_COD under optimum conditions could be applied in determining the optimum Fenton oxidant dose without performing multiple experiments according to its initial COD₀ and the desired .

14.6.3.9 Effect of L_COD on Biodegradability

Lopez et al. (2004) found the effectiveness of the Fenton reagent process for pretreating landfill leachate in order to improve its biodegradability (defined as BOD₅/COD) to be ≥0.5, a value compatible with subsequent aerobic biological treatment. In their study, the leachate analyzed had a low concentration of heavy metals, a pH value of 8.2, a BOD₅/COD ratio of 0.2, and high contents of NH₄–N (5210 mg/l) and alkalinity (2174 mg/l) and therefore was classified as “old” and nonbiodegradable. To maximize the effectiveness of the FP (i.e. minimize scavenging effects), it was necessary to determine the optimal operational H₂O₂/Fe²⁺ mass ratio using a constant concentration of H₂O₂ of 6300 mg/l. This ratio was determined to be 12 (6300/500). Maintaining this mass ratio, the leachate was treated with increasing amounts of both reagents in order to determine the maximum amount of COD removed, which was determined to be 60% removal for dosages of 10 000 mg/l of H₂O₂ and 830 mg/l of Fe²⁺. Also maintaining the mass ratio of 12, experiments were conducted in order to determine the achievement of a BOD₅/COD ratio of ≥0.5, which was achieved with dosages of 3300 mg/l of H₂O₂ and 275 mg/l of Fe²⁺ as indicated in Figure 14.16. Lopez et al. (2004) concluded that their results were encouraging and should be tested on a larger scale as the results indicated a technologically simple and cost‐effective method for improving the biodegradability of leachate.

A common approach to landfill leachate treatment is to transfer leachate to a wastewater treatment facility where it is subjected to low‐cost biological treatment after mixing with municipal sewage (Renou et al., 2008). However, the low biodegradability of stabilized leachate as compared with municipal sewage affects the treatment performance of WWTPs that are often built for the treatment of municipal wastewater (Diamadopoulos et al., 1997). Therefore, a single or combination of physicochemical processes that help improve leachate biodegradability is often used as a pretreatment step in biological processes (Lin and Chang, 2000). The most common physicochemical processes that help reduce refractory OM from landfill leachate treatment are coagulation (Comstock et al., 2010), anion exchange (Singh et al., 2012a), AC adsorption (Maranon et al., 2009; Singh et al., 2012b), and chemical oxidation (Rivas et al., 2004; Tizaoui et al., 2007).

In the past decade, the applicability of AOPs for removing OM and improving the biodegradability of leachate has attracted significant interest from various researchers (Umar et al., 2010). Singh et al., 2013 evaluated the biodegradability of leachate by analyzing the BOD₅ of raw and treated leachate samples. Figure 14.17 showed that a twofold increase in BOD₅ at each L_COD in the range of 0.25–1.0 during Fenton oxidation (FO), with the maximum increase of 2.4 times at L_COD of 0.25. When L_COD range varied from 0.5 to 1.0 of 0.50–1.0 BOD₅/BOD_5,0 was increased from 1.9 to 2.4. The smaller L_COD corresponds to a higher amount of available O₂ or ^•OH for oxidation, so more OM may be transformed into biodegradable form as compared with the higher L_COD. OM that were resistant to reaction with ^•OH did not transform into biodegradable form, so the application of a higher amount of H₂O₂ at lower L_COD would not provide additional benefit in terms of improving the biodegradability of leachate. FP showed smaller BOD₅/BOD_5,0 as compared with FO at each L_COD and varied in the range of 1.4–2.1. The increase in BOD₅ after FO was greater than after FP because after FO, the larger OM that are transformed into smaller OM coagulate with each other at an increased pH. Although most of the coagulated OM was precipitated during the precipitation process, the coagulated OM that are still present in the sample are harder to biologically degrade, thus showing less BOD₅ by FP than FO.

14.6.3.10 Effect of L_COD on Cost of Fenton Process Treatment for Landfill Leachate

The amounts of H₂O₂, FeSO₄⋅7H₂O, acid (H₂SO₄), and base (NaOH) used for pH adjustment for oxidation of leachate can be used to estimate the required amount of each chemical in leachate treatment. The unit price for each chemical was collected from the commercial grade chemical suppliers, and the cost associated with each chemical and total cost was calculated using Equations (14.23) and (14.24):

(14.23)

(14.24)

where total and chemical cost was in $/m³ of leachate to be treated, unit chemical cost in $/m³ of chemical to be used for leachate treatment, and chemical consumption in m³ of chemical to be used /m³ of leachate to be treated.

The average commercial grade cost of H₂O₂ (35% w/w), FeSO₄⋅7H₂O, H₂SO₄ (98%), and NaOH (97%) was found to be $510, 75, 200, and 325/ton, respectively. The cost of H₂O₂ (35% w/w) was found to be the major cost‐contributing chemical in FP. The chemical consumption cost at each L_COD for the three leachates was calculated as shown in Figure 14.18. Leachate with the highest COD₀ and lowest L_COD (COD₀ 1092 mg/l at L_COD of 0.25) was observed to be the most expensive treatment case with a total chemical cost of $17.11/m³, whereas leachate with low COD₀ (273 mg/l) and the highest L_COD (1.0) required a chemical cost of $1.30/m³. Additionally, the total chemical cost for all three leachates exponentially decreased with an increase in L_COD. The L_COD 0.25 required an average of $10.3/m³ of leachate, whereas at L_COD 1.0, the estimated chemical cost was $2.9/m³. The high cost at low L_COD was primarily due to the amount of H₂O₂ used for the Fenton reaction and relatively high usage of H₂SO₄ and NaOH for pH adjustment as compared with the Fenton treatment of L_COD in the range of 0.50–1.0.

The average cost associated with treatment of three leachates using FP was also evaluated with respect to η_FP. For L_COD smaller than approximately 0.47, the η_FP decreased, while the treatment cost exponentially increased. However, for L_COD greater than 0.47, the treatment cost associated with chemical usage decreased and η_FP increased. Therefore, L_COD less than 0.47 would not be cost effective due to decrease in η_FP and increase in treatment cost. The total chemical cost was also calculated in terms of per m³ of leachate treated per mg/l COD removed. The average cost for all three initial leachate COD_o varied in the range of 8.2 × 10⁻³–2.0 × 10⁻² ($/m³ of leachate/mg/l COD removed) for L_COD in the range of 0.25–1.0. Although not enough information has been published on the cost analysis of leachate treatment using FP, Rivas et al. (2004) observed a comparable treatment cost of $0.008/m³ of leachate/mg/l of COD removed for H₂O₂ at a cost of $300/ton.

Dimensionless parameter L_COD could be used to predict the optimal Fenton oxidant doses without performing multiple experiments. To establish a universal dimensionless equation between the optimal efficiency of Fenton and the L_COD, Singh et al. (2013) conducted Fenton treatment of three leachates for L_COD in the range of 0.25–1.0. The kinetics of FO showed that the oxidation process reached stable stage within 1 min at the start of reaction and showed a maximum COD removal of 46% at pH 3.5, depicting FO as a very quick efficient process for OM removal from landfill leachate. The lower the L_COD, the greater the COD removal observed, and a maximum of 52 and 70% COD removal was observed at L_COD 0.25, by FO and FP, respectively. However, the usability of Fenton oxidation was greater at L_COD 1.0 than at L_COD 0.25 as observed by higher η_FO and η_FP. The comparison of experimental results with the model equation as developed by Singh and Tang (2013) showed L_COD 0.75 as an optimum L_COD for any leachate. Therefore, the selection of Fenton oxidant dose in terms of L_COD and comparison of experimental results with the equation may eliminate the practice of performing multiple experiments to determine optimum Fenton operating conditions. Leachate biodegradability with respect to L_COD showed an increase in BOD₅ in the range of 1.9–2.4 after FO for L_COD in the range of 0.25–1.0. Additionally, FO showed greater biodegradability improvement than FP. The evaluation of leachate treatment cost (chemical) showed an exponential increase with the decrease in L_COD, and L_COD greater than 0.47 showed cost‐effective treatments by decreasing cost and increasing COD removal efficiencies.

14.8.1 Questions

1. List at least four factors that influence leachate generation rate.
2. Why is leachate control necessary?
3. List at least four conditions that will influence leachate chemical characteristics.
4. Please explain why leachate is toxic.
5. What does it imply when the theoretical molar ratio of H₂O₂/Fe²⁺ should be 11?
6. Should leachate be treated with wastewater in WWTP?
7. Why is equation considered as an universal equation? What are the limits of this equation?
8. What type of organic compound would be oxidized the most at this ratio at pH 3.5?

14.8.2 Calculations

1. If H₂O₂ and Fe²⁺ were dosed at 1 and 0.1 mM, what would be the steady‐state concentration of hydroxyl radical at pH 3.5? If a leachate has the same water quality as listed in Table 14.2 and has COD of 1500 mg/l, what treatment efficiency would you expect at pH 3.5?
2. In Miami‐Dade Southern WWWTP, there are two treatment systems of leachate. One treats 1.2 MGD that is discharged to WWTP of 75 MGD WWTP, while 0.32 MGD leachate was treated and sent to the 35 MGD WWTP. If the leachate quality is the same as those listed in Table 14.2, please design a leachate treatment process of 1.52 MGD to meet the following standards:
   1. BOD₅ less than 20 mg/l, TSS less than 20 mg/l, total N less than 5, and total P less than 0.4 mg/l for deep well injection.
   2. BOD₅ less than 10 mg/l, TSS less than 10 mg/l, total N less than 2, and total P less than 0.02 mg/l for water reuse, what are the additional unit processes should be added to the above treatment train?
   3. For water reuse, UV disinfection is usually used to reduce the production of disinfection by‐products (DBPs). What are the major barriers if the treated leachate is going to be disinfected with UV? What are the technologies you would recommend to improve the system so that UV disinfection would be effective?

14.8.3 Projects