Dried H-Y zeolite was first converted to its sodium form by means of an ion-exchange method [7]. The zeolite was washed with 1 M NaCl at 80˚C for 4 h in such a way that the ratio of zeolite to sodium chloride solution was 1:80. The slurry was then filtered and washed with deionized water to remove the excess amount of NaCl. In order to achieve complete exchange of sodium into the zeolite, the above cycle was repeated twice. The sample was then dried for 12 h at a temperature of 80˚C and then calcined in a muffle furnace at 550˚C for 5 h.

The heterogeneous catalyst was prepared using the procedure reported by Li et al. [8]. One gram of NaY was suspended in 50 mL water and known amounts of Fe (NO₃)₃ (Fe³⁺ loading from 0.2 - 0.6 wt%) and fixed amount of titanium precursor (the theoretical proportion of TiO₂ in the catalyst is 15 wt% TiO₂) was added into the suspended solution and mixed by mechanical stirrer for 24 h at ambient temperature. The resulting samples were then dried using a rotary evaporator and placed in an oven at 100^ºC for 4 h. Then, the produced solids were calcined at 550^ºC for 5 h. The produced catalysts were denoted as x Fe%/15%TiO₂-NaY where x is the wt% of Fe. In order to investigate the effect of TiO₂ with fixed amount of Fe, the same preparation method was repeated with TiO₂ loading from 5 - 20 wt% and fixed amount of Fe.

The XRD pattern of NaY, potassium titanium oxalate (PTO) calcined at 550^◦C, 0.4% Fe-NaY, 15% TiO₂-NaY and 0.4% Fe/15%TiO₂-NaY are shown in Figure 1. The

parent NaY shows high crystalline structure while the calcined PTO showed an amorphous structure with small anatase peak at 2θ 48.1˚. However, after the incorporation of Fe and TiO₂ with NaY for the case of 0.4% Fe/ 15% TiO₂-NaY, the crystallinity of zeolite disappeared and a single wide peak with low intensity appeared at 2θ of 27.7˚ ascribed to rutile phase. The formation of new material after the incorporation of Fe/TiO₂ with zeolite was not suggested since the diffraction pattern of 0.4% Fe/15% TiO₂-NaY was similar to that of 15% TiO₂-NaY catalyst.

No characteristic peak for iron oxide phase that could be formed after the calcination step was detected either for 0.4% Fe-NaY or 0.4% Fe/15% TiO₂-NaY. This could be ascribed to the low loading of Fe used in this study. Castano et al. [9] reported that the unclear diffraction for some metals could be attributed to the small crystal size (<4 nm) or to the well-dispersion of amorphous phases that caused unclear detection by XRD. Wang et al. [10] did not indicate any clear peak for Fe phases after the doping of TiO₂ and Fe into a natural zeolite even at higher concentrations up to 7% mol of Fe.

It has been reported in literature that the transformation phase temperature from anatase to rutile usually occurs at temperature between 700˚C - 800˚C [11,12]. Therefore, the appearance of the rutile phase for both catalysts i.e. 15% TiO₂-NaY and 0.4% Fe/15% TiO₂- NaY should not be attributed to the calcination temperature used in this study. However, the titanium precursor could be the reason behind the appearance of rutile phase since no anatase peak at 25.1˚ (highest peak for anatase phase) was detected in the bare potassium titanium oxalate (PTO) diffraction calcined at 550˚C.

In order to investigate the effect of titanium precursor on the crystallinity of zeolite, the titanium precursor was changed with another organic precursor i.e. tetra butyl orthotitanate (TBT) using the same preparation method and the catalyst was denoted as 0.4% Fe/15% TBT-NaY as shown in Figure 1. The result indicated that zeolite support was able to retain its crystallinity even after the loading of both Fe and TiO₂ and a small anatase peak was detected at 25.1˚. As a result, it seemed that the dispersion of TiO₂ clusters with zeolite depend on the type of titanium precursor and consequently affecting on the crystallinity of the support. The partial incorporation between Ti species and the framework of zeolite might be occurred during the preparation method since the zeolite structure have super cages and pore size of 1.3 nm and 0.74 nm, respectively. The excess amount of Ti species could be anchored on the surface of the support after the calcination step thus lead to a significant reduction in the crystallinity of catalyst due to the accumulation of TiO₂ phase on it.

The FTIR spectra for the parent NaY, 0.4% Fe-NaY, 15% TiO₂-NaY and 0.4% Fe/15% TiO₂-NaY are presents in Figure 2. The absorption features of NaY resulted in different absorption peaks in the range between (4000 - 400) cm⁻¹. The peak at 3439 cm⁻¹ and 1626 cm⁻¹ represents the presence of water molecules or hydroxyl groups attached to zeolite structure. Hassan and Hameed [13] reported that the region from 1600 to 3700 cm⁻¹ is ascribed to the existence of zeolite water molecules. These water molecules attached to the silicate minerals can be found in two types i.e. intact molecules or as OH groups formed from the dissociation of water interact with metals such as (Si⁴⁺, Al³⁺). The peak at 1138 cm⁻¹ was ascribed to the asymmetric stretching modes of internal tetrahedral. While the absorption peak at 514 and 450 cm⁻¹ represent the T-O bending vibrations of internal tetrahedral in zeolite [14]. The zeolite framework also present peak around 1050 cm⁻¹ ascribed to Si-O-Si asymmetric stretching mode of external linkages bond in zeolite framework. It should be mentioned that the terms internal and external vibrations have been used in FTIR spectroscopy for zeolites materials to describe the vibrations in tetrahedral building units and between them [15].

A significant change in stretching vibration for 15%

TiO₂-NaY and 0.4% Fe/15% TiO₂-NaY in comparative to NaY was observed at peak value around 1650 cm⁻¹ ascribed to the existence of water molecules in zeolite [13]. Nikazar et al. [16] reported that the changes occurs in this regions after the doping of metal indicated that a certain amount of OH was demolished by this metal i.e. Ti and Fe. The peaks at 1138 cm⁻¹ and 1050 cm⁻¹ for NaY were attributed to the asymmetric stretching modes of internal tetrahedral and asymmetric stretching of external linkage for (T-O-T ) respectively [17]. However, the peak at 1138 cm⁻¹ (NaY) was disappeared in the case of 15% TiO₂-NaY and 0.4% Fe/15% TiO₂-NaY due to the effect of TiO₂ and Fe loading. Accordingly, the change occurred in peak at 1050 cm⁻¹ was ascribed to the sensitivity of this region to the structure changes.

No peak could be detected at 960 cm⁻¹ for 15% TiO₂- NaY and 0.4% Fe/15% TiO₂-NaY after the loading of TiO₂ which was assigned to the anti-symmetric stretching vibration of the Ti-O-Si [18]. This peak was ascribed to the interaction of Ti species with the internal framework of zeolite. However, the absence of this peak in this catalyst suggests the dispersion of TiO₂ to be more on the external surface of zeolite. This result is in agreement with the results obtained by XRD.

Quintero Ortíz et al. [19] reported that the peak in NaY at 820 cm⁻¹ was assigned to the stretching vibrations of external bond between tetrahedral of zeolite. However, this peak showed a reduction in stretching vibration and shifting to lower wavelength after the incurporation of TiO₂ for the case of 15% TiO₂-NaY and 0.4% Fe/15% TiO₂-NaY. While no changes were detected for the case of 0.4% Fe-NaY so, this change was ascribed to the incorporation of TiO₂. On the other hand, the stretching vibration from 500 to 700 cm⁻¹ was ascribed to the sensitive change in the pseudolattice vibration of zeolite after the loading of Metal/Zeolite [20]. From the above analysis, the differences in FTIR spectra were evident especially for TiO₂ loading due to its higher amount in comparative to Fe.

The morphological characteristics was investigated based on the values of mean roughness that were presented in two and three dimensional AFM images of NaY, 15% TiO₂-NaY, 0.4% Fe-NaY and 0.4% Fe/15% TiO₂-NaY as in Figure 3. The variation occurred in surface morphology of the prepared catalysts suggested the significant effect of Ti and Fe incorporated with zeolite. The results for the degree of roughness was in the sequence of NaY > 0.4% Fe-NaY > 0.4% Fe/15% TiO₂-NaY > 15% TiO₂- NaY with roughness values of 74 nm, 72 nm, 48 nm and 44 nm, respectively. The similarity in the roughness of the catalyst surface between NaY and 0.4% Fe-NaY seemed to be due to the low loading of the Fe₂O₃ which might be formed after the calcination step. Thus, showed this slight effect on the uniformity on the catalyst surface.

However, it was clear that the roughness of the catalyst was significantly decreased after the loading of TiO₂ for both cases of 15% TiO₂-NaY and 0.4% Fe/15% TiO₂- NaY. This extreme reduction in the roughness of the catalysts surfaces were ascribed to the small crystal size of the supported TiO₂ which was in agreement with XRD results (low intensity and wide peak). Although, the AFM technique can not differentiate between the existing metal on the surface of the catalyst. However, small clusters were significantly detected for the case of 0.4% Fe/15% TiO₂-NaY which could be ascribed to Fe oxide formed after the calcination step. While for the case of 15% TiO₂-NaY, it was clear that TiO₂ phase formed a layer covered the surface of zeolite and no small clusters were observed on its surface comparing to that of 0.4% Fe/15% TiO₂-NaY.

The TEM imaging was carried out for parent NaY, and 0.4% Fe/15% TiO₂-NaYcatalysts. The unmodified zeolite structure can be observed in Figures 4(a) and (b) for two different magnifications i.e. 8 kX and 45 kX, respectively. The hexagonal crystalline geometry of the zeolite remained virtually unchanged for the case of 0.4% Fe/ 15% TiO₂-NaY using impregnation method as was observed in Figure 4(c). These results proved that the zeolite was generally able to maintain its original structure after the loading of TiO₂ and Fe since no significant damages were detected on its hexagonal structure. The TEM image suggested that the position of TiO₂ clusters were mainly positioned on the external surface of zeolite due to appearance of many white clusters accumulated on the external surface of zeolite structure and it was not shielded by zeolite particles. However, it was not possible to detect the Fe metal due to the higher loading of TiO₂ that was significantly occupied the surface of zeolite.

In order to obtained higher decolorization activity of the catalyst, preliminary experiments were carried out to investigate the synergistic effect between different wt% of Fe with fixed amount of TiO₂ 15 wt% as shown in Figure 5. The experimental runs were carried out with 10 mg/L initial dye concentration, catalyst loading of 1.5 g/L and natural pH about 5.5 without the addition of H₂O₂ in the reaction. The results showed that the catalytic activity significantly increased with increasing the wt% of Fe loading from 0.2% to 0.4% Fe with catalytic activities of 54% and 76%, respectively. In this reaction, the Fe acted as electron acceptor from the e⁻-h⁺ pairs generated through the activation of TiO₂ in the zeolite.

Increasing the loading amount of Fe would provide more electron traps to hinder e-h⁺ ions recombination leading to an increase in the sonocatalytic activity. In a complementary role, the zeolite support worked as a hole trap thus enhance efficiently the inhibition of the electron hole recombination. Therefore, the lower catalytic activity for the catalyst at 0.2% Fe loading was ascribed to the insufficient amount of Fe ions to inhibit the e⁻ – h⁺ recombination. However, with increasing loading of Fe up to 0.6 wt%, the catalytic activity was found to decrease to 64%. In this case, as the wt% Fe was further increased, the e⁻ – h⁺ recombination rate would increase also due to the existence of excess amount of Fe which would act as a charge carrier to cause this reduction in the catalytic activity. Bajamundi et al. [21] investigated the effect of Fe loading from 0 - 3.0 wt% Fe on the activity of TiO₂/Al₂O₃ and the higher degradation efficiency of methylene blue was at 0.1% Fe. As a conclusion it seems that increasing the concentration of Fe up to 0.4 wt% showed significant enhance in catalytic activity and the addition of higher concentration affected negatively on catalytic activity.

As the optimum loading of Fe was 0.4 wt%, further experiment runs were carried out to investigate the interacttion between fixed loading of 0.4% Fe with different loadings of TiO₂ from 5 - 20 wt% and the results are shown in Figure 6. The sonocatalytic activity was found to increase with increasing the loading of TiO₂ from 5 - 15 wt% reaching its maximum activity of 76% at the highest TiO₂ loading. Increasing the loading of TiO₂ would provide more active sites to enhance the generation of ^•OH radicals during the sonocatalytic reaction [11]. Further increase in TiO₂ loading above the optimum value i.e. 20 wt%, the decolorization efficiency of amaranth was significantly decreased to 34%. This drastic reduction in catalytic activity might be ascribed to the aggregation of TiO₂ clusters on the surface of zeolite [20]. Furthermore, with increasing amount of TiO₂, the amount of available Fe sites would not be sufficient to inhibit the e⁻-h⁺ recombination thus, reduced the catalytic activity of the catalyst.

Another factor that can negatively effect on the catalytic activity might be referring to the formation of a separate TiO₂ phase with increasing the TiO₂ loading. This phase will no longer be incorporated with zeolite framework consequently, cause this reduction in catalytic activity. This result was suggested based on the results obtained from the AFM analysis since, there was a thin layer covered the surface of zeolite after the loading of 15 wt% of TiO₂. However, the incorporation of 0.4% of Fe/15% TiO₂ with zeolite showed the higher catalytic

activity and further increase in TiO₂ was negatively affected.

A preliminary study was carried out to investigate the role of catalyst under both silent (mechanical stirring only) and ultrasonic reaction. The adsorption study without the presence of ultrasonic irradiation was carried out to check the adsorption capacity of 0.4% Fe/15% TiO₂- NaY catalyst for amaranth dye under continuous mechanical stirring for 120 min. The experimental runs were at an initial dye concentration of 10 mg/L, a catalyst loading of 1.5 g/L and at original pH of the dye solution. The results presented in Figure 7 showed that only 4% of amaranth dye was adsorbed after 120 min of continuous stirring.

The decolorization efficiency under ultrasonic irradiation without the presence of catalyst was also investigated under the same reaction condition and the efficiency of the process was only about 12%. This low process efficiency was mainly ascribed to the limited generation of the hydroxyl radicals by the dissociation of water molecules during the ultrasonic reaction. In addition, Mahamuni et al. [18] suggested that the recombinetion of the ^•OH radicals within the cavity during the sonolysis process could eliminate the generation of the active species thus reduce the efficiency of the sonolysis process.

On the other hand, the decolorization efficiency visibly increased after the addition of catalyst under ultrasonic irradiation i.e. 17.4%, 30%, 68% and 76% for PTO (potassium titanium oxalate) calcined at 550˚C, 0.4% FeNaY, 15% TiO₂-NaY and 0.4% Fe/15% TiO₂-NaY, respectively. The synergistic between the catalysts and the ultrasonic process led to increase the production of more hydroxyl radicals and enhance the decolorization efficiency. From Figure 7, it is clear that the incorporation

of Fe (III) and Ti species with zeolite could enhance the decolorization efficiency of amaranth as compared to that of 0.4 % Fe-NaY and 15% TiO₂-NaY catalysts.

The effect of reaction acidity on the sonocatalytic degradation of amaranth was investigated within the pH range of 2.5 to 9.0 under ultrasonic reaction and the results are present in Figure 8. Before starting the sonocatalytic reaction, all the reactions were adjusted to the desire pH using 0.1 M NaOH or HCl with an initial dye concentration of 10 mg/L, 0.65 mM of H₂O₂ and 1.5 g/L catalyst loading for up to 120 min. It was noticed that the decolorization efficiency of the catalyst in the first 60 min was increased from 9% to 76% for pH 2.5 and 5.5, respectively. After longer reaction time, the color removal at different pH diminished and the reaction was almost complete with decolorization efficiency of about 98% for pH ranging from 3.5 to 9.0. Meanwhile, very low sonocatalytic activity of about 22.4% was obtained at the acidic condition of pH 2.5 after 120 min of reaction. The potassium (K) which was originally exists in the chemical structure of the titanium precursor was the reason behind this low catalytic activity at pH 2.5. The trace of such highly basic element exists at the catalyst in addition to the Fe and Ti will significantly affect the environment of the catalyst surface. Huang et al. [2] reported that the decolorization efficiency for methyl orange dye using Pt/ TiO₂-zeolite catalyst was higher in high acidic condition.

An experiment run was carried out to investigate the effect of titanium precursor on the decolorization efficiency of amaranth at pH value of 2.5. The TiO₂ precursor (potassium titanium oxalate) used in this study was replaced with tetrabutyl orthotitanate using the same preparation method used in this study and the produced catalyst was denoted as 0.4% Fe/15% TBT-NaY. The dark experiment run showed an adsorption rate of 68% after 30 min of mechanical stirring without ultrasonic irradiation. However, the absorbed dye was desorbed

from the catalyst surface at the beginning of ultrasonic reaction due to the mechanical vibration provided by ultrasonic. The catalytic activity was 83% after 120 min of reaction without any addition of H₂O₂. It was clear that this large difference in catalytic activity between the original catalyst 0.4% Fe/15% TiO₂-NaY and 0.4% Fe/ 15% TBT-NaY at pH 2.5 was ascribed to the difference in chemical nature of TiO₂ precursor.

It is important to highlight that the zero point charge of the catalyst is depend on dispersion nature and the difference in catalyst diameter [6]. Furthermore, the surface properties are affected by the synergistic between Fe, TiO₂ and zeolite in addition to the trace of potassium oxide. This properties is complicated to be understand due the different component interact together. As a result, one can observe that the synthesized catalyst had high catalytic activity in a wide range of pH from 3.5 to 9.0 which could affect by the chemical nature of the catalyst.

The decolorization efficiency of amaranth was demonstrated to obtain the optimum loading with respect to the amount of catalyst used from 1.0 to 2.5 g/L at an initial dye concentration of 10 mg/L, 0.65 mM of H₂O₂ and pH of 5.5 for up to 120 min. In Figure 9, it was notice that the difference in catalytic activity with different loading of catalyst was insignificant after 120 min of reaction and the activity was about 98% for the catalyst loadings variation from 1.0 to 2.5 g/L. This behavior could be attributed to the synergistic effect between the heterogeneous catalyst and the ultrasonic irradiation that provide mechanical mixing between the generated radicals and the dye molecules which enhance the catalytic activity [6]. However, there was a significant difference in decolorization efficiency at the first 30 min. The catalytic efficiency was found to increase with increasing catalyst loading to about 83% at catalyst loading of 2.0 g/L. Further increase in the amount of catalyst up to 2.5 g/L caused an activity drop to 74%.

The difference in sonocatalytic activity with different loading of catalyst at the first 30 min of reaction was

ascribed to multiple reason such as increase the number of the active sites of such catalyst that generate more ^•OH radicals during the sonocatalytic reaction. Furthermore, the presence of Fe ions provided higher decomposition rate for H₂O₂ to generate more hydroxyl radicals during the reaction [22]. Therefore, the reaction becomes faster with increasing the time of reaction. As such, the difference in catalytic activity with different loading of catalyst was diminished after 120 min of reaction. Wang et al. [23] reported that the optimum amount of catalyst loading for the Fe doped TiO₂ was 1 g/L for the degradation of azo fuchsine under ultrasonic reaction. Meanwhile, Zhong et al. [24] reported that the maximum catalyst loading of Fe₂O₃/Al₂O₃ for the sonocatalytic degradation of acid orange 7 was 0.3 g/L. As a conclusion, the optimal amount of catalyst loading varied depending on the process conditions, the nature of dye molecules and the properties of catalyst.

The decolorization efficiency of amaranth dye as a function of different H₂O₂ concentration was investigated at initial dye concentration of 10 mg/L, 2.0 g/L of catalyst loading, original pH of the dye about 5.5 and with different concentrations of H₂O₂ from (0.32 - 2.62) mM. The decolorization efficiency was found to increase from 72 % to 83 % at the first 30 min with increasing the concentration of H₂O₂ from 0.32 to 0.65 mM as presented in Figure 10. However, by increasing the concentration of H₂O₂ above 0.65 mM, the decolorization efficiency was decreased to 69% at 2.61 mM of H₂O₂ concentration. This behavior could be ascribed to the excess amount of H₂O₂ added to the reaction solution that led to the generation of hydroperoxyl radicals (HO₂•) which was considered as the less reactive radicals compared to ^•OH radicals [25]. Aravindhan et al. [26] reported that excess amount of H₂O₂ added into the reaction system can negatively affect the efficiency of the process due to the scavenging of ^•OH.

Figure10. Effect of H₂O₂ on decolorization efficiency amaranth.

Pang et al. [27] reported that the optimum amount of H₂O₂ needed to add into the ultrasonic reaction is depends on the type of the organic compound to be degraded and the ultrasonic operating condition. However, the presence of heterogeneous catalyst in synergistic with ultrasonic irradiation can be consider another factor that effect on the amount of H₂O₂ added into the reaction. In section 3.2.2 the sonocatalytic activity of 0.4% Fe/15% TiO₂-NaY without the addition of H₂O₂ was 74% thus, only 0.65 mM was needed to enhance the catalytic activity to 97% after 120 min of reaction and that no desirable effect in the color removal was observed with the addition of higher concentration of H₂O₂.

The decolorization efficiency of 0.4% Fe/15% TiO₂-NaY catalyst as a function of initial dye concentration was also investigated. The experimental runs were carried at 2 g/L of catalyst loading, original pH of the dye solution i.e. about 5.5 and 0.65 mM of H₂O₂. It was noted in Figure 11 that nearly complete decolorization was obtained at the end of the reaction at different initial dye concentrations from 10 to 100 mg/L. However, the rate of decolorization in the first 30 min of reaction was directly proportional to the decrease in initial dye concentration.

As the time of reaction was further increased, the color removal was relatively fast for the case of 30 mg/L of initial dye concentration. This behavior can be explained based on the collusion theory [28]. With increasing the concentration of dye molecules per unit volume of reaction, the chance of collisions between these reactants molecules increased thus, the decolorization efficiency steadily increased. Above the optimum value of 30 mg/L of the initial dye concentration, the decolorization efficiency after 60 min decreased from 100% to 85% for 30 and 100 mg/L of initial dye concentration, respectively. Jamalluddin Abdullah [29] reported the low decolorizetion efficiency with increasing the concentration of dye could be affected by the mutual screening between the

organic dyes molecules at higher concentration of dye. However, with increasing the time of reaction more decolorization rate was observed and this mutual screening with higher dye concentration was significant only in earlier time of reaction.

It was found that the leaching of Fe³⁺ ions increased with increasing the concentration of dye molecules due to the generation of more acidic-by products that were formed during the reaction thus enhance the dissolving of Fe³⁺ ions during the reaction. The concentrations of Fe³⁺ found in the solution after 120 min of reaction were 0.011, 0.017, 0.033, 0.051 and 0.06 mg/L for 10, 30, 50, 70 and 100 mg/L, respectively. The more Fe³⁺ leached out from the catalyst, the more loss in active sites of catalyst was expected [30]. However, this small amount of Fe³⁺ ions separated from the catalyst did not affect the decolorization efficiency even for the high concentration of dye as it was clear from Figure 11.

Our previous report [5] investigates the degradation of amaranth using the heterogeneous catalyst of Fe loaded on encapsulated TiO₂. The higher decolorization efficiency of amaranth was 98% at 10 mg/L initial concentration, 2.5 pH, 0.2 M H₂O₂ and 2 g/L catalyst loading. However, in this catalyst the maximum decolorization efficiency was obtained at 30 mg/L initial dye concentration, 5.5 pH, 0.65 mM H₂O₂ and 2 g/L catalyst loading. The difference in optimum conditions between the two catalysts was significant especially for amount of H₂O₂ and pH of reaction solution. It seems that position of TiO₂ with respect to zeolite structure, preparation method and the amount of TiO₂ affect extremely on chemical environment of the catalyst and consequently effect on the optimum condition of the reaction.