Effect of Activating Agent on the Properties of TiO2/Activated Carbon Heterostructures for Solar Photocatalytic Degradation of Acetaminophen

Several activated carbons (ACs) were prepared by chemical activation of lignin with different activating agents (FeCl3, ZnCl2, H3PO4 and KOH) and used for synthesizing TiO2/activated carbon heterostructures. These heterostructures were obtained by the combination of the activated carbons with a titania precursor using a solvothermal treatment. The synthesized materials were fully characterized (Wavelength-dispersive X-ray fluorescence (WDXRF), X-ray diffraction (XRD), Scanning electron microscopy (SEM), N2 adsorption-desorption, Fourier transform infrared (FTIR) and UV-visible diffuse reflectance spectra (UV-Vis DRS) and further used in the photodegradation of a target pharmaceutical compound (acetaminophen). All heterostructures were composed of anatase phase regardless of the activated carbon used, while the porous texture and surface chemistry depended on the chemical compound used to activate the lignin. Among all heterostructures studied, that obtained by FeCl3-activation yielded complete conversion of acetaminophen after 6 h of reaction under solar-simulated irradiation, also showing high conversion after successive cycles. Although the reaction rate was lower than the observed with bare TiO2, the heterostructure showed higher settling velocity, thus being considerably easier to recover from the reaction medium.


Introduction
In recent decades, the treatment of contaminants of emerging concern in water effluents, such as pharmaceuticals and personal care products (PPCPs), is receiving special attention because of their commonly recalcitrant and toxic character [1][2][3]. Those species enter to water bodies through industrial discharges but also from municipal wastewaters [4,5]. In wastewater treatment plants (WWTPs), organic matter and suspended solids can be efficiently removed, but most PPCPs are highly resistant to conventional biological treatments, such as the activated sludge. In this context, there is a growing demand for technologies that can deal with these emerging contaminants in cost-effective terms. Advanced oxidation processes (AOPs), which include Fenton-based, UV-and sunlight-assisted, ozonation or heterogeneous catalytic systems, among others, have been so far widely investigated technologies for that purpose [6][7][8]. Heterogeneous photocatalysis offers the advantage of operating at mild conditions and the opportunity of using solar light as a sustainable and cost-effective energy source, which is a main challenge regarding the economy of this potential solution. In heterogeneous photocatalysis, the irradiation of a semiconductor induces the generation of charges (electrons, e − and

Preparation of Activated Carbons
Activated carbons were obtained by chemical activation of lignin using different activating agents. Table 1 summarizes the activation conditions of the different carbons [20,[25][26][27]. Firstly, 5 g of lignin and the corresponding mass of activating agent were physically mixed. For KOH, the lignin was previously carbonized at 800 • C for 2 h, avoiding the fragmentation and solubilization of the lignin caused by strongly nucleophilic hydroxyl ions from the activating agent [28]. Then, the mixtures were dried at 60 • C overnight and heat-treated at the desired temperature for 2 h under N 2 flow (100 Ncm 3 ·min −1 ) in a horizontal stainless-steel tube furnace, using a heating rate of 10 • C ·min −1 . Then, the samples were cooled down to room temperature under N 2 flow. The resulting solids were further washed in two steps. Firstly, with HCl (0.1 M) at 70 • C for 2 h to remove the residual activating agent and secondly, with deionized water at room temperature up to neutral pH. The final materials were dried overnight in an oven at 60 • C. The resulting activated carbons were denoted as Fe-C, Zn-C, P-C and K-C, according to the activating agent used. The activating agent to lignin mass ratio was established in each case after previous experiments where different values were tested.  [29]. Preliminary trials were conducted where different TiO 2 /AC mass ratios were checked prior to select the most suitable to achieve heterostructures based on anatase phase. After these studies, the amount of TiO 2 was fixed at 80% in all cases. Thus, 58 mg of activated carbon were suspended into 45 mL of EtOH at room temperature for 5 min, leading to the solution A. At the same time, 1 mL of Ti(OBu) 4 was diluted in 15 mL of EtOH for 5 min (solution B). Then, solution B was added dropwise to solution A under continuous stirring until complete homogenization. A solution of 3 mL of ultrapure water in 15 mL of EtOH was incorporated dropwise to produce the hydrolysis of the Ti precursor. The mixture was stirred for 5 min, transferred to a 125 mL Teflon-lined stainless-steel autoclave, and heated at 160 • C for 3 h. After the reaction, the solid was separated by centrifugation (5300 rpm, 10 min), washed three times with deionized water and finally with ethanol. The resulting grey materials were dried at 60 • C overnight. The heterostructures were labelled as TiO 2 /x-C, namely TiO 2 /Fe-C, TiO 2 /Zn-C, TiO 2 /P-C and TiO 2 /K-C, depending on the activated carbon used to form the heterostructure. For comparison, bare TiO 2 was also obtained under the same conditions in the absence of activated carbon.

Characterization Techniques
A Bruker S8 TIGER spectrometer (Bruker, Billerica, MA, USA) under inert atmosphere (He) (maximum voltage of 60 kV and maximum current of 170 mA) was used to determine the percentage of TiO 2 in the final samples by wavelength-dispersive X-ray fluorescence (WDXRF). X-ray diffraction (XRD) patterns were recorded on a Bruker D8 diffractometer (Bruker, Billerica, MA, USA) with a scintillation detector, using Cu-Kα source, with a scan step of 1 • ·min −1 between 5 and 70 • of 2θ. Scherrer's equation was used to estimate the average crystal size (D) from the most intense diffraction peak (101) of anatase phase. A Quanta 3D Field Emission Gun (FEG) microscope (FEI Company, Hillsboro, OR, USA) was used to obtain the scanning electron microscopy (SEM) images of the samples. The particles size distributions were obtained from these images by using the ImageJ software (National Institutes of Health, Bethesda, MD, USA). In each case, 20 particles per image (out of a total of five images) were analyzed to achieve a good representation of the particle size.
The porous texture was characterized by N 2 adsorption-desorption at −196 • C using a Micromeritics TriStar 123 static volumetric system (Micromeritics Instrument Corp., Norcross, GA, USA). The samples were previously outgassed under vacuum at 150 • C overnight in a Florprep 060 Micromeritics device (Micromeritics Instrument Corp., Norcross, GA, USA). The specific surface area (S BET ) was determined by the Brunauer-Emmett-Teller (BET) method [30], while the external or non-microporous surface area (S EXT ) and micropore surface area (S MP ) were calculated using the t-plot method [31]. Fourier Transform Infrared (FTIR) spectra (wavenumber range 4000-400 cm −1 ) were recorded on a Bruker iFS 66VS spectrometer (Bruker, Billerica, MA, USA) using a resolution of 2 cm −1 . Samples were previously prepared using KBr pellets. The pH drift method [32] was used to determine the pH at the point of zero charge (pH pzc ). Briefly, 50 cm 3 of 0.01 M NaCl solution at initial pH (adjusted between 3-11 using 0.1 M HCl or NaOH) were placed in a closed titration vessel. Then, 20 mg of the sample were suspended and nitrogen was bubbled before starting the test in order to stabilize the initial pH by removing dissolved gasses. The final pH, measured after 5 h, was plotted versus the initial one. The pH pzc is given by value where the curve crosses pH initial = pH final .
A Shimadzu 2501PC UV-vis spectrophotometer (Shimadzu Corporation, Kyoto, Japan) was used to record UV-vis diffuse reflectance spectra (UV-vis DRS) in the 250-800 nm region using BaSO 4 as reference material. The band gap values were estimated using the Tauc Plot standard technique [33]. Considering that all heterostructures are indirect semiconductors, as TiO 2 [34], this method uses the equation αhν = α·(hν−E g ) 1/2 , where α, h, ν and E g are the absorption coefficient, Planck constant, light frequency and the energy gap of the semiconductor, respectively. Plotting (αhυ) 1/2 vs. hυ results in a curve with a linear region. The extrapolation of this linear branch to the X-axis provides the band gap value of the material.

Photocatalytic Tests
The photocatalytic degradation of acetaminophen (ACE) was performed in a 500 mL Pyrex jacketed reactor (Segainvex UAM, Madrid, Spain) at a controlled temperature of 25 • C. The experiments were carried out inside a Suntest solar simulator (Suntest XLS+, ATLAS, Mount Prospect, IL, USA) equipped with a 765-250 W·m −2 Xe lamp. Solar radiation was simulated using a "Daylight" filter (cuts off λ ≤ 290 nm), selecting 600 W·m −2 (107.14 klx) as irradiation intensity. In each test, the concentration of photocatalyst was set so that all the experiments were performed with the same amount of TiO 2 , 250 mg·L −1 . For this purpose, the results obtained after characterization by WDXRF were used to know the amount of TiO 2 of each sample. The calculated amount of photocatalyst was dispersed in 150 mL deionized aqueous solution containing the contaminant. Prior to the photocatalytic tests, the adsorption capacity of each heterostructure was estimated since the heterostructures showed very different porous textures. Hereby, each catalyst was contacted with ACE solutions of different concentrations, measuring the amount adsorbed at equilibrium after 16 h. That value was used to adjust the initial concentration of ACE for each catalyst, which was in all cases 5 mg·L −1 . Further, the suspension was exposed to simulated solar light for 6 h. Samples of 450 µL were collected at different times and filtered using PTFE syringeless filters (Scharlau, Scharlab S.L., Barcelona, Spain) (Whatman 0.2 µm). The liquid phase was analyzed by HPLC (Shimadzu Prominence-I LC-2030C, Shimadzu Corporation, Kyoto, Japan) equipped with a diode array detector (SPD-M30A) and a reverse phase C18 column (Eclipse Plus 5 µm, Agilent Technologies, Santa Clara, CA, USA) to measure the ACE concentration (detection wavelength set at 246 nm). A mixture of acetonitrile/acetic acid 0.1% v/v (gradient method: 10/90-40/60% (0-17 min)) was used as the mobile phase, with a constant flow of 0.7 mL·min −1 . Total organic carbon (TOC) was measured at the beginning and end of the reaction using a Shimadzu TOC-L analyzer (Shimadzu Corporation, Kyoto, Japan). The experiments were carried out in duplicate. Settling tests were performed with catalysts suspensions of 1 g·L −1 . The settling profiles versus time were recorded using a Shimadzu 2501PC UV-vis spectrophotometer (Shimadzu Corporation, Kyoto, Japan) (double beam), measuring the absorbance at 600 nm (in which the extinction of light is mainly due to the scattering caused by the particles of the suspension). In these tests, 4 mL of the suspension were placed in a quartz cuvette and the absorbance were recorded continuously for 2 h. Deionized water in absence of the catalyst was used as blank. In parallel, 100 mL of the catalysts suspensions with the same concentration were placed in graduated cylinders allowing the natural sedimentation of the particles. Pictures at different time intervals were taken for comparison.

Characterization of TiO 2 /Activated Carbon Heterostructures
The XRD diffractograms of the TiO 2 /x-C heterostructures synthesized are depicted in Figure 1 together with that of the TiO 2 prepared as reference. All the samples show the characteristic peaks of the anatase phase (JCPDS file No. 78-2486), whose (hkl) planes are marked in the Figure 1. No peaks assigned to other crystal phases of TiO 2 (e.g., rutile and brookite) or other from the AC were detected. That was also the situation with TiO 2 /C heterostructures synthesized by a similar procedure using glucose as carbon precursor [29]. Thus, solvothermal synthesis at 160 • C allowed the development of the crystalline structure of titania without the need of further heat treatment. Comparing the diffractograms, the most intense peak of anatase (101) differs in width depending of the samples. Usually, the widening of the diffraction peaks is associated with the reduction of the crystal size and the creation of defects in the crystalline structure. The anatase crystal size (D) was calculated by using the Scherrer's equation and the position and width of the (101) peak. The values are collected in Table 2. The bare TiO 2 prepared by this solvothermal treatment exhibits a small crystal size, lower than that reported for anatase TiO 2 prepared by other methods [35]. In the TiO 2 /x-C samples, even smaller TiO 2 crystal sizes are observed, the values depending on the nature of the activating agent used to prepare the AC. The lowest anatase crystal size was obtained with Zn-C. These values are analogous to those reported for other TiO 2 /AC hybrids prepared by the sol-gel method that required a calcination step in air [36,37]. The amount of TiO 2 incorporated in the heterostructures was determined by WDXRF and the results are included in Table 2. Most of the samples have a TiO 2 content similar to the expected (80%), with only the sample TiO 2 /Zn-C showing some significant deviation. Additional samples were prepared by decreasing the percentage of TiO 2 , but the results obtained in terms of photocatalytic activity for ACE degradation were considerably poorer and thus, those samples were discarded.  All synthesized heterostructures showed an external spherical-like morphology, as it can be observed from the SEM images (Figure 2), analogous to the bare TiO 2 . Figure 2 also includes the histograms of the particle size distribution, obtained by measuring ca. to 100 particles for each sample. The average diameter of TiO 2 particles is higher than the obtained for the heterostructures, which fall within the range of 0.24-0.28 µm. However, the particles of TiO 2 /x-C show a higher degree of agglomeration. This suggests that the presence of activated carbon inhibits the growth of the TiO 2 particles, as previously reported by Wang et al. [38]. The size distribution is unimodal in all cases except for TiO 2 /Zn-C, which shows a bimodal profile centered at 0.21 and 0.30 µm.  Nitrogen adsorption-desorption isotherms of activated carbons and TiO 2 /x-C heterostructures are represented in Figure 3. K-C and Fe-C ACs (Figure 3a) show typical type I isotherms, characteristic of microporous materials, according to the IUPAC classification [39]. In both cases, the major uptake occurs at low relative pressures followed by an almost horizontal branch, this shape being indicative of highly microporous solids. In contrast, Zn-C and P-C describe hybrid I/II type isotherms with hysteresis loops at P/P 0 > 0.4, suggesting the existence of a well-developed microporous structure but with significant contribution of mesoporosity. Figure 3b shows that the different TiO 2 /x-C heterostructures present type II isotherms, with a remarkable decrease of adsorption, indicating that the porous structure of the activated carbons has been partially blocked by the incorporation of TiO 2 .  Table 3 summarizes the surface area values (BET, microporous and external or non-microporous) characterizing the porous texture of the synthesized materials. The development of surface area depends on the chemical activating agent used, following the order K-C >> Zn-C >> P-C > Fe-C. It should be mentioned the high development of mesoporosity upon ZnCl 2 -activation (as showed by the high value of the so-called external surface area). According to literature [28,40,41], the chemical activation mechanisms consist in the oxidation and/or dehydration of the precursor, including complex different reactions depending on the agent. However, a general mechanism involved in this type of activation is still not very clear. The dehydration during the heat treatment of the lignin seems to be the most determinant effect in the chemical activation with ZnCl 2 and H 3 PO 4 . In the case of H 3 PO 4 , a strong Brønsted acid, it yields a partial depolymerization, followed by dehydration and condensation processes; whereas in the case of ZnCl 2 , a Lewis acid, the reaction with lignin is proton-catalyzed producing the dehydration and further aromatization of the carbon skeleton [42]. KOH reacts with the solid precursor by means of redox reactions, resulting in the microporosity development after oxidizing carbon to CO and CO 2 . This predominant formation of micropores has also been observed with FeCl 3 -activation, in contrast to ZnCl 2 , ascribing a similar behavior to that of alkali agents [20]. In the case of TiO 2 , the entire surface area corresponds to mesopores. Further introduction of activated carbon in the solvothermal synthesis step allowed obtaining heterostructures including microporosity.
The surface functional groups of all activated carbons and synthesized heterostructures were assessed from the FTIR spectra ( Figure 4). The activated carbons (Figure 4a) show the characteristic bands due to the presence of adsorbed water, appearing the -OH stretching and bending bands at 3400 and 1600 cm −1 , respectively [29,43,44]. In the K-C carbon, the peaks located at 1537 and 1408 cm −1 were assigned to the stretching of -C=C in the skeletal aromatic ring and -CO stretching of carbonate groups, respectively [43,45,46]. Stretching of -COC can be observed in the absorption bands centered in the range of 1170-1030 cm −1 . This vibration can be produced from diverse oxygenated groups, thus varying the main absorption peak [29,43]. In P-C, the weak bands centered at 1060 and 980 cm −1 can be attributed to -PO and -POC groups. TiO 2 and TiO 2 /x-C FTIR spectra (Figure 4b), are very similar among them, indicating the homogeneous distribution of TiO 2 over the heterostructure. There are three main absorption bands. Those centered at 3400 and 1625 cm −1 are associated with adsorbed water, as previously indicated for activated carbons. The wide band located between 800 and 400 cm −1 corresponds to the characteristic -Ti-O-Ti stretching band of TiO 2 [47,48]. This band appears in all samples, with a maximum ca. to 700 cm −1 , which corroborates the generation of titania phase in all the heterostructures. The high intensity and width of this band overlaps those characteristic bands described for the activated carbons, taking into account that the TiO 2 percentage of these heterostructures is fairly high (ca. 80%). In the case of TiO 2 /P-C, around 1030 cm −1 appears a weak shoulder that can be ascribed to the stretching bands of -PO and -POC groups previously described for the corresponding activated carbon.  An important parameter regarding the acid-base behavior of the samples is the pH pzc . When the pH of the solution is lower than the pH pzc , the surface of the solid is positively charged, whereas it is negatively charged at pH above the pH pzc of the sample. Table 2 summarizes the pH pzc values of the synthesized heterostructures. The pH pzc of bare TiO 2 is almost neutral, which means that its surface is not charged in neutral water [49,50]. In the case of the heterostructures, it is clear that the pH pzc values depend on the activating agent used during the preparation of the activated carbon, corresponding the lowest value to TiO 2 /P-C. The low acidity of FeCl 3 (a weak Lewis acid) can result in the introduction of poor acidic groups in the heterostructure surface, giving a pH pzc analogous to that of TiO 2 , while the use of H 3 PO 4 (a strong Brønsted acid) yields a heterostructure with a low pH pzc probably due to the presence of phosphates that leave OH groups on the surface [17]. Regarding to TiO 2 /K-C, despite the basic character of KOH, the solid has a relatively neutral PZC (even lower than that prepared with FeCl 3 ). This may be due to the fact that after activation the solid was washed with HCl to remove the remaining KOH, thus eliminating the basic sites in the final solid.
The light absorption in the UV and visible region of the synthesized photocatalysts was investigated by UV-vis DRS technique, the resulting spectra are shown in Figure 5a. The absorption band observed in the UV range (below 360 nm) is very similar in all cases, typical of TiO 2 . In the visible region, the spectra of the heterostructures do not fall to zero because of their grey color. The effect is more evident in TiO 2 /Fe-C and TiO 2 /Zn-C, because these samples have lower percentages of TiO 2 ( Table 2), yielding the color change from light-to-dark grey. As mentioned above, the band gap values were determined from the UV-vis DRS spectra (according to the Tauc plot method, Figure 5b) and are included in Table 2. The values are fairly similar to that of TiO 2 [51,52], only somewhat higher except in the case of TiO 2 /Fe-C. Although it has been reported that the combination of TiO 2 with carbonaceous supports can produce a significant red-shifted displacement of absorption edge, in our heterostructures this has not been observed. This may be due to a more limited interaction between TiO 2 and AC compared to other carbon supports, like graphene or carbon nanotubes [11,12].

Photocatalytic Tests
The photocatalytic performance of the synthesized heterostructures in the degradation of acetaminophen (ACE) under solar light is depicted in Figure 6, which includes the results with bare TiO 2 for the sake of comparison, as well as a blank experiment, showing the complete stability of ACE under solar light irradiation in absence of catalyst. A comparative study using TiO 2 /Fe-C as photocatalyst was also carried out for 6 h with and without simulated solar radiation ( Figure S1) after the adsorption period, showing no variation in the ACE concentration in absence of light. As depicted in Figure 6, the light-assisted tests were preceded by a 16-h step in dark to allow the adsorption equilibrium. The differences in the adsorbed quantity of ACE can be explained by means of multiple interactions, such as electrostatic forces or the porous texture. Negligible ACE is adsorbed on the TiO 2 surface, probably due to its low porous development. In contrast, the interaction between the heterostructures and the contaminant seems to be more influenced by electrostatic interactions. The reaction was carried out at an initial pH of 6.9. Due to that the pKa of ACE is 9.9, the molecules of the contaminant are neutrally charged [53]. In contrast, pH pzc of the different heterostructures is below than the reaction pH and, thus, the surface of these photocatalysts is partially negatively charged (increasing the negative charge with decreasing the pH pzc ). As a consequence, the quantity of adsorbed ACE decreases as pH pzc diminishes because of partial repulsive electrostatic forces between the neutral molecule and the negative character of the surface of the heterostructures. For example, TiO 2 /P-C has the lower pH pzc value of synthesized heterostructures (4.86) and show the lowest adsorption of ACE. It can also be seen in Figure 6 that bare TiO 2 exhibits better photocatalytic performance than the synthesized TiO 2 /x-C heterostructures, which is consistent with the easier accessibility of TiO 2 [11], but also with the higher opacity of the suspension caused by the black-grey color. Furthermore, the higher concentration of the heterostructures contributes to their poorer behavior (all the tests were performed with the same amount of TiO 2 , and therefore, the amount of solid in the suspension is higher with the TiO 2 /x-C heterostructures). Regarding the photocatalytic performance of the synthesized heterostructures, TiO 2 /Fe-C appears as the most active, allowing complete ACE conversion after 6 h of reaction, probably due to the lowest band gap of this photocatalyst. The reduction of the photocatalytic activity of TiO 2 /activated carbon heterostructures with respect to bare TiO 2 has been previously reported [54,55]. However, other publications claimed the opposite behavior, where the highest photocatalytic activity was found for TiO 2 /activated carbon materials [29,56,57]. In some of these works, dark-adsorption was extended for no more than 1 h, which was not probably enough for the complete adsorption equilibrium. Therefore, when the photocatalytic tests started, a combination between adsorption and photocatalytic reaction was carried out. In this sense, the current work focuses only in the photocatalytic performance of the synthesized heterostructures, after reaching adsorption equilibrium.
Since the degradation pathway is of high interest, the final solution was evaluated by ion-exchange chromatography, founding different small carboxylic acids (acetic, formic and malonic) although their total concentration was really low, below 0.07 mg/L. Following the degradation pathway described in the literature [58,59], the presence of other proposed aromatic compounds was investigated but no one could be elucidated. Mineralization of the target pollutant, i.e., complete conversion into CO 2 and H 2 O, was followed by measuring the total organic carbon (TOC) in the solution after the 6 h-irradiation time ( Table 4). The TiO 2 /x-C heterostructures yielded significantly lower degradation of TOC than the bare TiO 2 , being again TiO 2 /Fe-C the most effective among them. The potential application of a photocatalyst must consider not only its activity but also the recovery from the reaction medium. Settling tests were conducted at neutral pH with bare TiO 2 and TiO 2 /Fe-C and the results demonstrated the significantly easier separation of this last (as depicted in Figure 7), probably due to the higher degree of agglomeration observed by SEM. In addition to this, Figure S2 shows pictures of the settling process for both photocatalysts at different times, in which the highest settling velocity of the TiO 2 /Fe-C can also be observed, thus being considerably easier to recover the heterostructure from the medium. Absorbance evolution profiles (600 nm) during settling test of TiO 2 /Fe-C and TiO 2 photocatalysts.
The stability of the photocatalyst with the best recoverability was investigated upon four consecutive cycles. After each cycle, the used TiO 2 /Fe-C was filtered, washed with deionized water and dried at 60 • C overnight. Each new cycle was carried out with identical conditions as previously described for the degradation of ACE ( Figure 6). As depicted in Figure 8, the heterostructure showed a good performance in the photocatalytic oxidation of ACE after four successive cycles with a slight decrease in the final conversion (92% removal after 6 h of irradiation). The porous texture of the used photocatalyst was also analysed (Figure 9), showing a decrease of the porous network, probably due to the partial blocking of the microporous structure by the adsorbed contaminant or even the oxidized intermediates. Furthermore, the leaching of titania was not detected in the solution after four recycles by using inductive coupled plasma methodology.

Conclusions
Solvothermal synthesis of TiO 2 /activated carbon (TiO 2 /AC) heterostructures was successfully achieved. Lignin has been used as starting material for the ACs following chemical activation with four agents (FeCl 3 , ZnCl 2 , H 3 PO 4 and KOH). The activated carbons showed a well-developed porous texture and different surface functional oxygenated groups and acid-basic character depending of the activation procedure. XRD patterns of the TiO 2 /AC heterostructures confirmed the presence of anatase phase with crystal size close to 10 nm in all cases after the solvothermal synthesis, without the need for further heat-treatment. These materials showed a spherical morphology with a particle size close to 0.27 µm. The presence of activated carbon in the heterostructures increased somewhat the band gap with respect to bare TiO 2 , except for TiO 2 /Fe-C. TiO 2 and TiO 2 /Fe-C showed the best efficiency in the degradation of acetaminophen under solar light, being higher in the case of TiO 2 . However, settling experiments demonstrated the easier recovery of the heterostructured material in spite of the lower size of individual particles, because of the higher aggregation observed by SEM. TiO 2 /Fe-C also showed a good performance in the photocatalytic oxidation of ACE after four successive cycles.