Catalytic Oxidation of Tartrazine in Aqueous Solution Using a Pillared Clay with Aluminum and Iron

In this work, pillared bentonite with Al−Fe (Al−Fe−PILC) was synthesized and used as a heterogeneous Fentonlike catalyst in the oxidation of tartrazine azo-dye in aqueous solution. The modification of bentonite with the AlFe mixed system in a concentrated medium, with ultrasound assisted intercalation was carried out, and the obtained catalyst was characterized by XRF, XRD, and N2 adsorption at 77 K. The oxidation of tartrazine with Al−Fe−PILC, using different amounts of H2O2, expressed as a multiple (1, 3, 6, and 9) of a stoichiometry amount required to completely oxidize the dye was evaluated. The reaction of catalytic wet peroxide oxidation (CWPO) of the dye with 400 mg of Al−Fe−PILC and 6 times the stoichiometric amount of H2O2 at 25 °C, reached 98.2±1.8% of decolorization, 51.9±1.9% of TOC removal and 71.5±1.8% of TN removal. Results of this study show that the oxidation of tartrazine increased with the amount of H2O2 up to a certain limit. This oxidation process can be considered as an alternative for treating wastewater containing azo-dye because the reaction takes place under mild experimental conditions (room temperature and atmospheric pressure). Copyright © 2021 by Authors, Published by BCREC Group. This is an open access article under the CC BY-SA License (https://creativecommons.org/licenses/by-sa/4.0).


Introduction
The different sources of water pollution contain hazardous compounds which transfer adverse effects on the environment [1]. Some of the organic contaminants present in water biodegradable, carcinogenic and mutagenic for aquatic systems and human health [7,8].
Tartrazine (known as E102 or FDC Yellow 5) is a synthetic yellow azo-dye used primarily as a food coloring and also found in some pharmaceutical products and cosmetics [9,10]. This dye is found in much foodstuff as desserts, ice cream, soft drinks, confectionery, instant puddings, gelatin, sauces, etc. Food processing facilities require water for activities that include washing, processing, and clean-in-place operations. As a result, they generate a great amount of wastewater, which often contain very high levels of total suspended solids, chemical oxygen demand, fats, oils and dyes [11,12]. Dyes are visible to the human eye even in a low concentration (< 1 mg/L) [13,14] and their inadequate disposal in aqueous ecosystems leads to reduction of sunlight penetration which in turn diminishes photosynthetic activity, resulting in acute toxic effects on the aquatic flora/fauna and dissolved oxygen concentration [15].
Different physical, chemical and biological treatment methods for dye removal have been adopted. These methods include adsorption, coagulation/flocculation, chemical oxidation (classical chemical treatments and advanced oxidation processes), electrochemical treatment and microbial or enzymatic degradation [6]. Advanced Oxidation Processes (AOPs) have been reported to be some of the most effective methods for the degradation of azo dyes, due to their complex chemical structure and recalcitrant nature (non-biodegradable) [16]. AOPs have been defined as water treatment processes which involve the generation of hydroxyl radicals (HO•) by chemical, photochemical and/or radiolytic methods. The hydroxyl radicals are very reactive species that attack most of the organic molecules and not only offer complete decolorization of aqueous solutions but also promise a considerable degree of mineralization and detoxification of the dyes and their oxidation/hydrolysis byproducts [17,18].
Among the AOPs, the Fenton process (H2O2/Fe 2+ ) is one of the most effective methods of organic pollutant oxidation. Disadvantages in using the Fenton process include the production of a substantial amount of Fe(OH)3 sludge (precipitate) and additional water pollution caused by the homogeneous catalyst that added as an iron salt. To solve the previous drawback, a water-insoluble solid catalyst can be used. This heterogeneous Fenton process is called catalytic wet peroxide oxidation (CWPO) [19][20][21]. The CWPO has been mainly used to oxidize phenols and their derivatives [20,22]. Its application in the degradation of other organic pollutants such dyes is not often reported [8,[23][24][25][26]. The development of heterogeneous Fenton and Fenton-like catalysts has been extensively explored, using materials based on zeolites, clays and activated carbon loaded with iron and copper [24,[27][28][29][30].
The pillaring procedure involves the formation, intercalation and subsequent fixation of polynuclear cations among clay layers. Thus, the lamellar spacing and specific area increase, making these materials attractive catalysts for various reactions [31]. Pillared clays with Al−Fe (Al−Fe−PILC) are promising catalysts for the CWPO because they combine porous support with active sites capable of degrading organic compounds [32]. In the process of catalytic oxidation, the initial step is the adsorption of the dye upon the active support, which is the fundamental degradation mechanism [33]. The second step is the reaction between the adsorbed dye and HO• radicals in the support surface. Thus, Al−Fe−PILC offers a combination of properties such as adsorption and catalytic activity, it is highly efficient for mineralization of pollutants [8,33]. In general, Al−Fe−PILCs have been shown to be efficient catalysts for phenol removal under mild conditions (atmospheric pressure and room temperature) without considerable leaching of metals [34,35]. Regarding the use of Al−Fe−PILC as catalysts for the CWPO of dyes (heterogeneous photo-Fenton oxidation is not included), there are reports just for acid chrome dark-blue (C16H9O9Na2ClS2N2) between 25 and 75 °C [26], tartrazine (C16H9N4Na3O9S2) between 25 and 75 °C [23], orange acid II (C16H11N2NaO4S) at 60 °C [25], methyl orange (C14H14N3NaO3S) at 18±2.0 °C [24] and Congo red (C32H22N6Na2O6S2) at 25 °C [8].
In this work, pillared bentonite with Al−Fe in a concentrated medium was synthesized (assisting the intercalation with ultrasound) [36], characterized and evaluated in the CWPO of tartrazine at 25 °C and atmospheric pressure (78 kPa). In addition to the degree of decolorization of tartrazine, the total organic carbon (TOC) and total nitrogen (TN) removals were also quantified, in relation to the amount of H2O2 used in the oxidation reaction.
Tartrazine (azo food dye, molecular formula: C16H9N4Na3O9S2, 534.3 g/mol, CAS registry number: 1934-21-0) was a product of nonpurified industrial quality (purity 62%, 38% NaCl and NaSO4 combined), purchased from Retema S.A.S.-Colombia). The chemical structure of this dye is shown in Figure 1. Stock solution (100 mg/L) was made up by accurately dissolving a weighed quantity of the dye in double-distilled water. Experimental dye solution of different concentrations was prepared by diluting the stock solution with a suitable volume of double-distilled water.

Synthesis and Characterization of Materials
The starting material was a natural bentonite mined by Bentocol S.A. from Valle del Cauca (Colombia). This clay has been previously characterized, having found that dioctahedral smectite (or montmorillonite) is the main component of the natural material [24,37]. The bulk clay was ground in a ball mill, and clay powder passed through a 100 mesh. The particle size separation of the clay fraction (< 2 µm) was performed by gravitational sedimentation based on the Stokes' Law [38]. To achieve it, 100 g of the powder sample were suspended in 10 L of distilled water. The suspension was magnetically stirred for 20 min, and transferred to a graduated cylinder for gravitational sedimentation. To obtain a fraction smaller than 2 µm, the suspension was allowed to rest for 16 h, and the first 20 cm of the suspension removed. Subsequently, the suspension was centrifuged at 5000 rpm for 10 min to recover the clay fraction, which was dried at 60 °C, ground and sieved in a 100 mesh [39]. The clay fraction was homoionized with 0.5 M CaCl2, washed with distilled water until the leachate showed a negative test for chloride ions, dried at 60 °C and, finally, ground and sifted in an ASTM 100 mesh.
The pillaring agent of Al−Fe purified (Al13+Fe nitrate) was synthesized by using an OH/metal hydrolysis molar ratio of 2.4, a previously published procedure [36]. The amount of Fe used in the synthesis of the pillaring agent was 5% molar because this amount favors the formation of larger and better-distributed pillars, which determines the good catalytic performance of these solids in the CWPO. For the modification, both homoionized clay fractions and the solid pillaring agent were mixed (20 meq of metal/g clay) and placed into a dialysis membrane, which was then immersed in distilled water and agitated for 3 h. After that time, the mix underwent an ultrasound bath (50 kHz) for 30 min. The modified material was washed and centrifuged until reaching a conductivity close to that of distilled water, dried at 60 °C, ground and sifted in an ASTM 100 mesh, and calcined for 2 h at 400 °C [36].
Both the clay fraction homoionized with calcium (labeled as Ca-Bent) and pillared clay with Al−Fe (labeled as Al-Fe-PILC) were characterized by X-ray fluorescence (XRF), X-ray diffraction (XRD) and adsorption-desorption of N2 at 77 K. XRF was performed using a Magix Pro Philips PW2440 instrument with samples prepared as glass pearls. For the X-ray diffraction, a Shimadzu LabX XRD-6000, which operates with Cu-K radiation (=1.5406 Å, steps of 0.02 °2θ and 2 s/step) was used. Nitrogen adsorption-desorption isotherms were determined in a Micromeritics ASAP 2020 instrument at 77 K after outgassing the samples for 2 h at 90 °C followed by 3 h at 400 °C. The specific surface area (SBET) was measured by means of the BET equation and, the total pore volume, evaluated for nitrogen uptake at a relative pressure of 0.99. Microporous specific surface area and micropore volume were calculated using the t-method of the Harking-Jura equation [40]. out, and were also established in accordance with literature data. The initial concentration of 25 mg/L is in the range of azo dyes' concentrations, usually found in industrial waste streams (between 10 and 50 mg/L) [41,42]. In Fenton-like reactions of phenol and dye oxidation with Al−Fe−PILC, the pH is sustained between 3.5 and 4.0 [24,43], and the optimal value reported in the literature has been 3.7 [24,44,45]. The average catalyst load reported for chemical oxidation of azo dyes varies between 2-5 g/L [24,26,41], so the lowest dose was used for this study. Stirring speeds of 200 rpm guarantee a complete mixing of the solid (adsorbent or catalyst) in the solution [24]. The oxidation reaction with a powdered catalyst was performed in a batch glass reactor, open to the atmosphere, thermostated at 25 °C, under constant magnetic stirring at 200 rpm. For each test, the reactor was loaded with 200 mL of aqueous solution at 25 mg/L and 400 mg of the catalyst. In this study, the pH of the medium was adjusted to 3.6 by using diluted 0.1 M H2SO4 acid or 0.1 M NaOH. Once the adsorption equilibrium time was reached, 8 mL of an H2O2 solution (2.0 mL/h) was added to the reactor. Although the addition of the total H2O2 dose is frequent at the beginning of the reaction, the gradual peroxide dosing had positive effects on the increase of TOC removal [46]. The time of reaction was 5 h (not including the time of dye adsorption-desorption equilibrium). The amount of H2O2 was varied in multiples of stoichiometry amount (1, 3, 6, and 9), which is theoretically required to completely oxidize one mole of tartrazine into CO2, H2O, and mineral acids, according to Equation (1):

Measurement of Catalytic Activity
Dye decolorization (dye removal efficiency) of tartrazine was measured by monitoring the absorbance of dye in the aqueous medium at its respective maximum absorption wavelength (max = 429 nm) using a UV-Vis spectrophotometer (Mapada V-1200, China). The dye concentration was determined from aliquots (1 mL of sample filtered in 0.45 µm millipore paper) measured at time intervals and by using a calibration curve relating absorbance vs. sample concentration. The concentration interval went from 0.5 to 20 mg/L, with a correlation coefficient (R 2 ) of 0.9919. Detection limit (DL) and quantification limit (QL) were 0.21 mg/L and 0.53 mg/L, respectively.
The dye decolorization was calculated from the following Equation (2): where C0 is the dye concentration at time zero (t = 0). This concentration corresponds to the solution in the adsorption-desorption equilibrium, and Ct is the dye concentration at time t.
Kinetic studies were performed by monitoring the change in the tartrazine concentration (Ct/C0) as a function of time (t). Apparent firstorder rate constants (kapp) for the different amounts of H2O2 were determined from the plot of −ln(Ct/C0) versus time (t); and the kapp was calculated with the slope of the line obtained [47].
The contents of the total organic carbon (TOC) and total nitrogen (TN) at the beginning and end of the reaction were determined in filtered aliquots of the reaction mixture using a TOC/TN analyzer (Multi N/C 3100, Analytik Jena AG, Germany). The Fe ions leaching from the catalyst at the end of the reaction were measured using an atomic absorption spectrophotometer (Thermo Scientific iCE 3000 Series). All of the oxidation tests were performed by triplicate. Table 1 shows the chemical analysis of starting bentonite and Al−Fe−PILC. A recent study proposes the use of a practical chart to identify the predominant clay mineral based on oxide composition in clay soils [48]. In accordance with the chemical composition of SiO2 (56.79 wt%), Al2O3 + Fe2O3, (  Al2O3 and thus a decrease in the Si/Al ratio respect to Ca-Bent. Although the content of Fe2O3 decreased in Al−Fe−PILC with respect to Ca-Bent, the Si/Fe ratio increased slightly. This last result can be associated with the uncertainty of the measurement made by XRF. Given the conditions of the synthesis of the solid pillaring agent and the low incorporation of Fe [36], this result appears reasonable. The above happened since the introduction of aluminum is very high in the pillared clay compared to iron, making the proportion of Fe2O3 less important. Powder X-ray diffraction patterns of the calcium bentonite and pillared clay with Al−Fe are shown in Figure 2. The shift in d001 reflection for the basal spacing of 13.5 Å (2θ = 6.52°) for the Ca-Bent to higher values in Al-Fe-PILC (18.5 Å, 2θ = 4.77°) confirms the effective introduction of metal polyhydroxocations and the subsequent formation of pillars in the interlaminar spacing. The dimensions of an Al13 Keggin ion have been estimated to be 1:09 nm  0:98 nm  0:97 nm [49,50]. After heating with air at 500 °C, Keggin ions lost their water ligands forming shorter Al13-blocks, with a height of 0.84 nm, becoming the structure supporting pillars [50]. The basal spacing of 18.5 Å for Al−Fe−PILC corresponded to the thickness of a montmorillonite-type clay sheet (~10 Å) plus the aluminum pillar (8.4 Å). The amount of iron incorporated in the solid pillarizing agent was very small (0.17 wt%) and did not contribute to the formation of Fe2O3 pillars. The iron incorporated in the pillaring agent was decorating the alumina pillars and not as iron oxide on the sheets of clay (cluster) [36].

Characterization of Materials
Adsorption-desorption isotherms of Ca-Bent and Al−Fe−PILC are shown in Figure 3. There is an increase in the adsorption capacity of N2 due to the modification of the clay by pillaring. According to the IUPAC classification [51], Ca-Bent presented an isotherm type IVa while Al-Fe-PILC is a combination of types Ia and IVa, both with H3 hysteresis loop. Type IVa isotherms are characteristics of mesoporous adsorbents. The H3 hysteresis loop is the result of non-rigid aggregates of plate-like particles. The adsorption isotherm of Al−Fe−PILC between 0.05 and 0.2 of relative pressure has the shape of a type Ia isotherm, characteristic of microporous materials. Thus, the combination of isotherms type Ia and IVa with H3 hysteresis loop indicates that both slit-like mesopores and micropores are formed.
The specific surface area (SSA) was 41.8 m 2 /g and 150.3 m 2 /g for Ca-Bent and Al−Fe−PILC, respectively. The pore volume at p/p∘ = 0.99 is of 0.0605 cm 3 /g for Ca-Bent and 0.1093 cm 3 /g for Al-Fe-PILC. The t-plot method was used to determine the specific area and the volume developed by micropores of the samples. These values are 2.2 m 2 /g and 0.0008 cm 3 /g for Ca-Bent and 107.0 m 2 /g and 0.0414 cm 3 /g for Al−Fe−PILC, respectively. The specific surface area of Ca-Bent is fundamentally the external surface (approx. 95% of SSA), characteristic of a structure of closed sheets, attributed to the heterogeneous arrangement of the aluminosilicate sheets. For Al−Fe−PILC, the specific surface area is basically microporous (approx. 71% of SSA), which is a typical feature of pillared clays [8,24].
The above results (XRF, XRD, and N2 sortometry) are similar to those found in the  literature for the same system (pillared bentonite with Al−Fe in concentrated medium, intercalation assisted with ultrasound), indicating that the synthesis methodology of pillared bentonite was reproducible [36].

Catalytic Activity of Al−Fe−PILC
Prior to the CWPO tests, adsorption tests of tartrazine on the materials were carried out. The maximum adsorption occurred during the first 30 minutes and stabilized after 1 h, reaching values of 4.1±0.2 and 12.5±0.5% for Ca-Bent and Al−Fe−PILC, respectively. The low adsorption of the dye in Ca-Bent was due to the fact that tartrazine is an anionic dye and, bentonite, a 2:1 layered silicate with a negative charge due to ionic substitution in its structure [52,53]. This charge in Ca-Bent is balanced by cations Ca 2+ present in the interlayer space, which cannot be interchanged by the anionic species of the colorant in solution (C16H9N4O9S2 3− ). Pillared clays are materials which have lower hydrophilicity than their parent clays [8,54]. It has been established that pillaring improves the efficiency in processes of adsorption of the anionic dye [55], therefore its greater adsorption capacity of tartrazine in an aqueous medium.
For all the oxidation tests, an adsorption time of 1 h was established, guaranteeing conditions of adsorption-desorption equilibrium. The catalytic activity of calcium bentonite and pillared bentonite with Al−Fe in the oxidation of tartrazine is shown in Figure 4. For the reaction blank (only 6H2O2), Ca-Bent + 6H2O2 and Al−Fe−PILC + 1H2O2, decolorization shows a slight increase as a function of the reaction time. The color removal of Al−Fe−PILC + 3H2O2 in the first 2 h of reaction increased slightly, and then at a higher speed (between 2 and 5 h). Al−Fe−PILC (with 6 and 9H2O2) evidenced a decolorization curve with two zones, one of increase in color removal (during the first 3 h of reaction) and then a stabilization zone (between 3 and 5 h).
Decolorization achieved with only H2O2 (blank, equivalent to 6 times the stoichiometric amount, without catalyst) was 5.81±0.79%. The unmodified clay (Ca-Bent) with 6 times the stoichiometric of H2O2 amount showed a low activity, reaching color elimination close to 19.4±1.7%. This catalytic activity can be related to the iron content of the raw clay mineral (6.85 wt% Fe2O3, Table 1). Catalytic activity of the pillared bentonite with Al−Fe was increased with respect to the starting clay and varied with the dose of H2O2 used in the reaction. Although the amount of iron incorporated into Al−Fe−PILC was very low compared to the iron content of Ca-Bent, the results indicate that Fe incorporated by pillaring is more active than that originally present in the clay, results similar to those reported in the literature [24,35].
The oxidation reaction of tartrazine carried out with a dose of H2O2 of 1 and 3 times the stoichiometric amount reached a decolorization of 29.4±1.6 and 78.3±2.3%, respectively. When the stoichiometric amount of H2O2 was increased to 6 and 9 times, the decolorization obtained after 5 h of reaction decreased from  98.2±1.8 to 94.7±2.1%. Excess H2O2 in a reaction medium (9 times the stoichiometric amount in this investigation) increased the chance for recombination of the radicals (i.e., scavenging effect) and production of less active radicals such as HO2• in the solution to generate inactive species as H2O and O2, which can be expressed by the reactions described by Equations (3) and (4) [20,46,56]: Data of change in the tartrazine concentration (Ct/C0) as a function of time (t) were adjusted to first-order kinetics and kapp values, for each amount of H2O2. For the four quantities of H2O2 used in the CWPO of tartrazine, the coefficient of determination (R 2 ) was higher than 0.965, confirming that data fit the first order model. Figure 5 shows the effect of H2O2 on the rate constant of decolorization (kapp) of tartrazine in aqueous solution. When the amount of H2O2 is increased from 1 to 3 times the stoichiometric value, kapp value increased 4.9 times. Additions of 6 times the stoichiometric amount of H2O2 increased kapp 14 times compared to the value obtained with 1 time of stoichiometric H2O2. Finally, the amount of H2O2 was 9 times the stoichiometric value, kapp decreased by 23.8% in relation to the value obtained with 6 times the stoichiometric H2O2. A similar behavior, that is, the existence of an optimum hydrogen peroxide dose has been found during photo-Fenton degradation of the reactive brilliant orange over iron-pillared montmorillonite. When increasing the amount of H2O2 from 2.9 to 9.8 mmol/L, degradation efficiency of the dye went up from 65.4 to 95.6% at 90 min. However, when the H2O2 dosage was 19.6 mmol/L, the removal of dye decreased [57]. In general, an excess of H2O2 is used with respect to the theoretical [H2O2]/[organic compound] molar ratio to reach the maximum degradation of organic compound [46,58].
UV-Vis spectra of tartrazine oxidation using Al−Fe−PILC and 6 times the stoichiometric amount of H2O2 are shown in Figure 6. At 2 h of reaction, a considerable decrease in the absorbance at 429 nm is observed (close to 67% of decolorization) and after 3 h of reaction, the color removal is greater than 92% so the yellow tone of the solution no longer it is noticeable to the naked eye.
There is no criterion for determining the amount of hydrogen peroxide necessary for the oxidation of an azo dye since it depends on the structure of the compound and the reaction temperature. An excess of H2O2 is usually used with respect to the stoichiometric amount. For example, for the oxidation of tartrazine (50 mg/L) with Al,Fe-pillared clay (0.5 g of solid catalyst in 100 mL of aqueous solution of dye), a dose of H2O2 equal to 9.3 times the stoichiometric amount was used. Tartrazine removal at 4 h of reaction at 75 °C reached about 97.5% [23]. For the oxidation of methyl orange (100 mg/L) with Al/Fe-PILCs, the dose of H2O2 was 0.9 times the stoichiometric amount. Removal of the dye was close to 80% in only 1 h of reaction at 18±2.0 °C [24]. However, in two previous investigations, the TOC removal was not quantified. Methyl orange is an excellent model chemical molecule for the catalytic assessment of active solids in Fenton-like reactions, as, for instance, CWPO [24]. Nevertheless, it is an azo dye of simple structure, of low molecular weight (327.33 g/mol, topological polar surface area of 93.5 A 2 ), in comparison with tartrazine (534.36 g/mol, a topological polar surface area of 229 A 2 ). These differences in the two monoazo dyes, as well as in the reaction by-products, mean that the results for similar catalysts are not comparable.
TOC and TN removal efficiency in the oxidation of tartrazine using Ca-Bent and Al−Fe−PILC is shown in Figure 7. The H2O2 only (6 times the stoichiometric amount, without catalyst) achieves very low removal of TOC and TN. When the Al−Fe−PILC catalyst is used and the H2O2 dose is increased from 1 to 6 times the stoichiometric amount, the removal of TOC and TN increases considerably. By increasing the dose of H2O2 (9 times the stoichiometric amount), the removal of TOC and TN decrease with respect to the reaction with fewer H2O2 (6 times the stoichiometric amount). In all tartrazine oxidation tests, it was found that TN removal was greater than TOC, reaching TN conversion of up to 71.5±1.8% (with 6 times the stoichiometric amount of H2O2). Similar results for the oxidation of Congo red were obtained with Al−Fe−PILC. At 4 h of treatment, NO3 -concentration (measured by ion chromatography) was practically zero, demonstrating that the catalyst gives a major loss of the initial nitrogen as volatile N-compounds, probably NxOy and N2 or NH4 + [8]. The influence of the H2O2 dose on the TOC or TN removal ( Figure  7) was similar to that obtained for decolorization ( Figure 4), with an optimum oxidant concentration. For an H2O2 dose above that value, the final TOC or TN removals decreased a little, an effect similar to those reported by other researchers [59][60][61].
The reaction path for the generation of reactive species and mineralization of dye using Al−Fe−PILC starts by reduction of Fe 3+ on the surface of Al−Fe−PILC to Fe 2+ . Then the Fe 2+ formed accelerates the decomposition of H2O2 in solution (Fenton reagent), generating hydroxyl (HO•) and hydroperoxyl (HO2•) radicals. These radicals attack the dye molecule leading to reaction intermediates and, finally, the reaction intermediates are mineralized into CO2, H2O, NO3 − and SO4 2− . The simplified reactions that schematize this process are shown in Equations (5) and (6)  When the HO• and HO2• radicals formed during the reactions attack the azo group (−N=N−) and break it, there is a color decay of the solution, which is used to measure the efficiency of the catalytic oxidation of the azo dye [24]. In this study, the oxidation by-products of tartrazine (other than CO2 and NOx) were not quantified by chromatographic techniques. In solutions of tartrazines pre-treated by electrocoagulation (EC) and, subsequently, by advanced oxidation by photoelectro-Fenton (PEF), 18 compounds have been identified by GC-MS. The dehydroxylated by-products (in the form of phenols and quinones) are the intermediates formed by the action of HO• radicals, as well as by aliphatic carboxylic acids (cannot be mineralized) [62].
In the CWPO process, the concentration of oxidant (H2O2) is a vital factor that considerably affects the removal of organic pollutants. The amount of generated hydroxyl radicals is directly related to the concentration of H2O2 [20]. According to the results of the catalytic tests carried out for the oxidation of the azotartrazine colorant with Al−Fe−PILC, the most efficient dose of H2O2 corresponds to 6 times the stoichiometric amount. With this dose, a decolorization greater than 98% and a TOC removal close to 52% is achieved. It is important to note that these results correspond to tests performed at 25 °C. It is well known that the mineralization of organic pollutants usually increases by elevating the temperature from 25 °C to around 80 °C. However, the CWPO process at elevated temperatures (> 40 °C) will increase the total cost of the treatment and the thermal decomposition of H2O2 [20,63]. For stability of the catalyst, the leached iron concentration was measured at the end of the oxidation tests. For all oxidation tests with Ca-Bent and Al−Fe−PILC, the leached iron concentration was less than 0.2 mg/L. Thus, it can be concluded that only the heterogeneous Fenton reaction happened for the degradation of tartrazine. The low iron leaching indicated that the active phase of these catalysts is strongly fixed to the clay support and pillars, and that it is highly stable under oxidizing conditions of the reaction [34,64].

Conclusions
Pillared bentonite with Al−Fe (Al−Fe−PILC) was effectively synthesized in a concentrated medium and used as heterogeneous Fenton-like catalyst in the oxidation of tartrazine azo-dye in an aqueous solution. Despite of the amount of iron incorporated into Al−Fe−PILC was very low compared to the iron content of Ca-Bent, its catalytic performance indicate that Fe incorporated by pillaring is more active than the purified clay.
The effect of H2O2 amount for tartrazine catalytic oxidation was evaluated. Increasing the dose of H2O2 (1, 3, and 6 times the required stoichiometric amount to completely oxidize the dye) increases decolorization and removal of TOC and TN. A high dose of H2O2 (9 times the required stoichiometric amount to completely oxidize the dye) does not improve the performance of the reaction.
The catalytic oxidation of tartrazine carried out with Al−Fe−PILC and 6 times the stoichiometric amount of H2O2 at 25 °C, reached 98.2±1.8% of decolorization, 51.9±1.9% of TOC removal and 71.5±1.8% of TN conversion. Additionally, the leached iron concentration in all CWPO tests was less than 0.2 mg/L, which guarantees the catalyst stability and that tartrazine oxidation occurred by a heterogeneous Fenton-like reaction.