All reagents were analytical grade and used without further purification. The reducing agent for GO was obtained from Capsicum annuum using solvent extraction. Diclofenac sodium powder (C₁₄H₁₀Cl₂NNaO₂, ≥98%), phosphoric acid (H₃PO₄, ≥95%), potassium permanganate powder (KMnO₄, ≥99%), liquid hydrogen peroxide (H₂O₂, ≥35%), titanium dioxide powder (TiO₂, 100%), graphite powder (C, ≥99%) and ethanol (C₂H₅OH, ≥99.5%) were supplied by DKL Laboratory Supplies, Philippines. Concentrated nitric acid (HNO₃, ≥90%) and sulfuric acid (H₂SO₄, ≥95%) were readily available in the laboratory. Prepared hydrochloric acid (HCl, 1 M) and sodium hydroxide (NaOH, 1 M) solutions were used for pH control. All solutions were freshly prepared using deionized water. Moreover, laboratory centrifuge, hot plate with magnetic stirrers, laboratory oven, ultrasonicator, and UV-spectrophotometer were used as equipment to carry out the synthesis, adsorption, and degradation processes.

GO was synthesized via modified Hummers’ method [19] . Concentrated H₂SO₄ and deionized water (in 1:9 volume ratio) were mixed with 1.5 g of graphite powder under constant stirring for 15 min. A constant amount of 6 g of KMnO₄ was slowly added to the acid mixture. It was then heated to 50˚C while being magnetically stirred for 12 h. Four (4) milliliters of H₂O₂ were added to the mixture to terminate the oxidation reaction. A color change was noted from purple to dark brown solution [20] . The resulting mixture was centrifuged at 4000 rpm for 15 min. The supernatant solution was then decanted and washed with deionized water to remove any adhered supernatant solution on the GO particles. The washing process was repeated twice to ensure complete separation of the GO from the by-products. The solid material was then oven-dried at 80˚C for 4 h. For the reduction process, the synthesized GO was dispersed in 40 mL ethanol to facilitate material exfoliation. The mixture was then ultrasonicated for 20 min. After which, a constant volume of 30 mL extract was added to the ultrasonicated GO solution. The mixture was refluxed at 80˚C for 5 h. The successful reduction of GO was indicated by the change in color from brown to black with precipitation [20] .

The TiO₂-rGO composite was synthesized via wet-impregnation method. Twenty-five (25) grams of the prepared rGO was mixed with 0.5 g of TiO₂ powder in a 50 mL ethanol solution. The mixture was heated and magnetically stirred for 1 hr to allow the reaction to proceed. The solids formed were then filtered using a Whatman filter. The recovered solids were oven-dried and transferred to an air-tight glass container.

Batch adsorption experiments were done to evaluate the removal of DCF with the use of TiO₂ and TiO₂-rGO composite under the absence of UV light. Thirty (30) milligrams of each material were mixed with 100 mL of 10 mg∙L⁻¹ DCF solution and allowed to adsorb for 60 min. After, samples were filtered using a 0.45 µm Whatman filter and analyzed for DCF residual concentration using a UV-vis spectrophotometer.

The photocatalytic performance of the catalysts was evaluated by exposure under UV-a light irradiation using a 365 nm UV-a lamp positioned at the roof of the photoreactor. The satisfactory suspension of the photocatalyst and the uniformity of the reacting mixture were guaranteed by employing magnetic stirring. The pH of the sample was adjusted by adding 1 M NaOH or 1 M HCl. At varying time intervals, 3 mL of the suspension was sampled and centrifuged at 4000 rpm for 5 min to separate the photocatalyst residue from the solution. The residual DCF concentration on each trial was analyzed using UV-vis spectrophotometer. Removal efficiency (R) was quantified using the Equation (1);

R ( % ) = C 0 − C t C 0 × 100 (1)

where R (%) represented the DCF removal, C₀ (mg∙L⁻¹) and C_t (mg∙L⁻¹) were the initial and final concentration of DCF at time t min, respectively.

The optimization of photocatalytic degradation under UV-a irradiation was done using central composite design (CCD) method. Three (3) operating parameters were considered namely catalyst loading, pH, and UV irradiation time at constant DCF concentration. The operating parameters together with their factor levels are shown in Table 1. Overall, 19 experimental runs were conducted. The experimental design was executed using Design-Expert® software for statistical analysis and data visualization.

It can be observed in Figure 1 that an increase in catalyst loading increases the degradation of DCF. The maximum DCF removal of 92.75% was noted on 50 mg catalyst dosage. The increase in degradation is possibly due to the increase in the availability of surface catalytic sites and adsorption sites in the composite. Results also showed that TiO₂@rGO composite had the highest adsorptive and photocatalytic removal compared to TiO₂ alone.

As the solution approaches pH < 6 and pH > 6 the percent removal of DCF decreases. As depicted in Figure 2, TiO₂@rGO composite exhibited the highest percent removal in photocatalysis. The maximum DCF degradation of 90.98% was obtained in the presence of composite and 76.80% in the TiO₂ catalyst alone at a nearly neutral pH of 6.0, after an irradiation time of 40 min. The pH is known to have a significant effect on the parameters influencing heterogeneous photocatalytic reactions in aqueous systems [21] . Consequently, the interaction of the substrate as well as the intermediates with the in-situ generated reactive

oxygen species (ROS) also would be affected.

UV irradiation provided a significant leap on the degradation efficiency of DCF as shown in Figure 3. The first 60 min showed the performance of adsorption alone and the second half showed the effect of UV irradiation. The removal of DCF was evident by adsorption alone, but with low removal efficiency at a slower

rate reaching 20% and 35% for TiO₂ and TiO₂@rGO, respectively. The presence of UV irradiation allowed the degradation kinetics to increase, as judged by the steep slope on the graph. Consequently, higher degradation efficiency was achieved reaching nearly 100% removal after subjecting to UV light. The curve flattened after 40 min of irradiation reaching the maximum degradation for the set-up.

The removal performance could be explained by the conduction-band electrons and valence-band holes generated during illumination [14] . Electron-hole pairs prevent the recombination rate which reduces the degradation performance. The electrons react with the dissolved oxygen in the solution to form superoxide while the holes oxidize donor molecules and react with water molecules to form hydroxyl radicals responsible for attacking the DCF [22] .

One-way analysis of variance (ANOVA) was used to determine whether there were any statistically significant differences between the means of three or more independent groups. The summary of p-values is presented in Table 2.

The result showed that there was a significant difference in every operating parameter wherein the corresponding p-values were less than 0.05.

Design Expert^® software was used to model the degradation of DCF in aqueous solution with varying parameters. The succeeding table summarizes the results of ANOVA analyses for the model, parameters, and interaction of parameters. Terms in consideration with p-values < 0.05 alpha were considered significant in the process. In this case, the model generated was significant which correlated the parameters and the response effectively.

The data in Table 3 showed the applicability of the suggested model for the process studied. The best process fit model order was cubic, with the coefficient of determination R² of 0.9292, predicted R² of 13.3035, and adjusted R² of 0.7452 which has the most acceptable fit among all tested models.

R ( % ) = 68.07 + 23 A − 0.4784 B + 18.27 C − 0.1529 A B + 5.05 A C +   1.91 B C              − 4.56 A 2 − 8.62 B 2 − 3.44 C 2 − 1.73 A B C − 1.81 A 2 B              − 6.27 A 2 C − 14.43 A B 2 (2)

The equation described by Equation (2) in terms of coded factors can be used to make predictions about the response for given levels of each factor. By default, the high levels of the factors are coded as +1 and low levels are coded as −1. The coded factors are useful to determine the relative impact of the factors by comparing the factor coefficients.

The perturbation plot helps to compare the effects of all the factors at a particular point in the design space. Referring to Figure 4, a straight-line C and two curved lines A and B intersected from the center point of the experimental region, and from there it measured response in each of the three-dimensional axes.

The contour plots presented in this subsection discuss the effects of the active terms in the model equation on the response values. These correlates only to those non-one-factor terms and other significant factors. The one-factor terms were not included because their involvement in the interactions have a great

effect on the responses.

As observed in Figure 5, the satisfactory removal percent, i.e., >80% could be attained at a pH value range of ~6 to 10 and composite dosage at >50 mg employing a constant UV irradiation time of 40 min.

The effect of the interaction between pH and UV irradiation time at constant catalyst is depicted in Figure 6. The desired % removal of DCF (red region), i.e., ~80% removal and higher could be achieved at a pH value of 6 - 10 and ~55 min irradiation time or higher. The catalyst performance would be unsatisfactory when the conditions were below the said values.

When both the UV irradiation time and catalyst dosage were considered simultaneously while holding the pH constant at 8.0, the parameter values that would give the most satisfactory degradation performance were >45 min UV irradiation time and ~47 mg catalyst dosage or higher as reflected in Figure 7. The irradiation time-composite dosage combination was more sensitive compared to the two previous combinations. Even at a high composite dosage, the degradation would still be low or shorter irradiation time. The case would be the same when the irradiation time is very high, but the catalyst dosage is not that high.