TiO 2 and TiO 2 -Carbon Hybrid Photocatalysts for Diuron Removal from Water

: TiO 2 and TiO 2 -activated carbon (AC) photocatalysts have been prepared (by sol-gel synthesis), characterized, and tested in the removal of diuron from water under simulated solar light. The preparation variables of the two series of catalysts are: (i) heat-treatment temperature of bare TiO 2 (350, 400, 450 and 500 ◦ C) and (ii) activated carbon content (0.5, 1, 5, and 10 wt.%) in TiO 2 -AC samples heat-treated at 350 ◦ C. The activated carbon was previously prepared by hydrothermal carbonization of saccharose and has spherical shape. The heat-treatment temperature does not determine the efﬁciency of TiO 2 for diuron photocatalytic degradation, but clearly inﬂuences the diuron adsorption capacity. The capacity of TiO 2 -AC samples for diuron removal increases with the carbon content and it is the result of combined diuron adsorption and photodegradation. Thus, the sample with highest carbon content (10 wt.% nominal) leads to the highest diuron removal. The TiO 2 -AC photocatalysts have proved to be capable of degrading diuron previously adsorbed in dark conditions, which allows their regeneration.


Introduction
As intensive agriculture involves the use of large quantities of herbicides, they are becoming emerging pollutants whose concentration in waters and soils is alarmingly increasing all over the world. Herbicides are toxic and usually persistent compounds, being thus very harmful for natural ecosystems [1]. Moreover, the problem of available water resources scarcity is growing in many areas of the Earth. Hence, the development of efficient and cheap methods to treat aqueous effluents is a scientific and technical challenge. Although biodegradation of organic/inorganic pollutants is often viable, a variety of pollutants, such as some pesticides, are recalcitrant to conventional biological treatments [2]. For this reason, more powerful degradation technologies are required.
Diuron (3-(3,4-dichlorophenyl)-1,1-dimethylurea) is one of the most commonly used herbicides [3], employed to control a wide variety of weeds and mosses, to protect many agricultural crops and to keep clean non-crop areas [4]. Dispersion of this compound leads to aquatic environments pollution by soil leaching [5,6], and it appears to be moderately persistent in water [7]. Diuron has been reported to be present in 70% of the rivers in 27 European countries, with concentration peaks up to 0.826 µg/L [8]. It is considered a real environmental hazard, demonstrated in 2013 by an issued law of the European Commission that included diuron in the list of priority pollutants in water treatment policy [9].
Advanced oxidation processes, such as heterogeneous photocatalysis, is one of the promising techniques for the abatement of water pollutants [1,[10][11][12][13][14]. Photocatalysis can be considered advantageous over existing disinfection methods, such as chlorination or ozonation, because the process is self-sustaining and the regular replenishment of treatment products can be avoided [1]. Titanium dioxide is one of the most used photocatalysts due to properties such as good efficiency, low cost, nontoxicity, and stability. Anatase is usually regarded as the most photoactive TiO 2 phase, mainly because of the lower e − /h + pairs recombination rate and the 3.20 eV wide band-gap [15][16][17]. Solar photocatalysis on TiO 2 has proved to allow total detoxification of aqueous diuron solutions in parallel to a complete mineralization of the pollutant [18,19]. However, it is still necessary to increase its photoactivity and efficiency.
As the photocatalytic activity of TiO 2 depends on physicochemical properties such as surface area, surface chemistry, crystallinity degree, phase composition, and crystal size, it is expected that an appropriate combination of these parameters would lead to very efficient photocatalysts [17,[20][21][22]. The mentioned properties can be modified, in a certain range, varying the preparation conditions. Moreover, the incorporation of other chemical elements or compounds has been also considered to be a way to improve the TiO 2 photocatalytic efficiency. Visible-light-induced photoreactivity of TiO 2 was improved by doping with metals [23][24][25][26], coupling with other semiconductors [24,27] or adding non-metal component(s), such as carbon. Carbonaceous materials have attracted attention as additives or components to prepare hybrid TiO 2 -C photocatalysts, mainly because of their unique and controllable structural and electrical properties [21,28,29]. The review work of Leary and Atwood [28] highlights the progress and development of nanocarbon-TiO 2 photocatalysts and states that they have the potential to address the three aspects that may bring an enhancement of the photocatalytic activity: (i) band-gap tuning or extension of excitation wavelength through photosensitization, (ii) retardation of electronhole recombination, and (iii) provision of high-surface area for adsorption of reactants and of active sites.
The Ngo's group has investigated the photocatalytic degradation of metsulfuronmethyl herbicide using TiO 2 and a small amount of powdered activated carbon (AC) as catalyst. These authors reported that hybrid TiO 2 -AC samples lead to significant better performance in herbicide removal than the independent adsorption by AC or the photocatalytic degradation by TiO 2 [30,31]. Bamba et al. studied the activity of TiO 2 prepared by sol-gel deposited on carbon materials such as AC, graphite, and carbon aerogel, for diuron photodegradation [32]. They found that TiO 2 /AC and TiO 2 /carbon aerogel samples were promising under sunlight irradiation, attributing the good behavior to a high anatase crystallinity, high specific surface area, and large pore volume.
The photocatalysts' regeneration efficiency is also another important property from the point of view of the practical applicability of the prepared photocatalysts. A few studies have been carried out on the recyclability of photocatalysts used in diuron degradation. For example, Anirudhan et al. proved that photocatalysts like ZnO nanorod decorated carboxylic graphene/polyaniline composite photocatalysts [33] and TiO 2 -based ones [34] can be used and regenerated at least for five cycles. However, no studies have been found to date to prove the recyclability of TiO 2 -AC materials as photocatalysts for diuron degradation.
In general, the photoactivity of the carbon-TiO 2 hybrid photocatalysts depends on the carbon content but also, as in the case of pure TiO 2 , on physicochemical parameters such as crystal structure, crystallite size and surface area [35,36] which, in turn, depend, among others, on the temperature of the post-synthesis crystallization heat treatment [37].
Considering the presented ideas, this work addresses the preparation by sol-gel of two series of samples: pure TiO 2 samples submitted to a post-synthesis heat treatment at a temperature ranging from 350 to 500 • C; TiO 2 -AC hybrid samples containing different amounts of a spherical activated carbon for which the post-synthesis heat-treatment temperature was 350 • C, and the study of their use for the removal of diuron from water under simulated solar light, including regeneration efficiency. In fact, comparison between these Catalysts 2021, 11, 457 3 of 16 two series, prepared in the same conditions, lets us analyze the importance of the carbon presence. Their regeneration has also been studied.

Determination of the Carbon Content in TiO 2 -AC Samples
The thermogravimetry (TG) curves of the TiO 2 -ACx (350) series (x = 0, 0.5, 1, 5, and 10) ( Figure S1a in Supplementary Materials) show three small weight losses at, approximately, 80, 180 and 330 • C, and a main weight loss from 450 to 650 • C. Differential thermogravimetry (DTG) profiles ( Figure S1b in Supplementary Materials) have allowed to precisely locate the temperature of the weight losses. The first two weight losses are due, respectively, to adsorbed water and organic residues remaining from the TiO 2 synthesis [38,39]. The third one is only evident in the TiO 2 -AC10 (350) sample, and corresponds to the decomposition of surface oxygen groups of the activated carbon [40]. The largest weight loss is due to combustion of carbon and the removal of oxygen-containing groups [41]. The weight loss related to the phase transformation from amorphous TiO 2 to anatase can also occur in the 450 to 650 • C range [42], but it is small compared to that due to the carbon combustion. Such a weight loss has been quantified for sample TiO 2 (350) in a similar TG experiment and it amounts about 0.2%.
Thus, the carbon content of the TiO 2 -ACx (350) samples has been calculated by subtracting the weight loss corresponding to the transformation of amorphous TiO 2 to anatase from the weight loss observed between 450 and 650 • C. The calculated carbon content was: 0.0, 0.4, 0.5, 3.6 and 8.3 wt.% for TiO 2 -ACx (350) samples (x = 0, 0.5, 1, 5, and 10, respectively). The complete quantification of the TG profiles can be seen in Table S1 (Supplementary Materials). Figure 1a shows the N 2 adsorption-desorption isotherms of TiO 2 (nt), TiO 2 (T) and P25 samples. All the photocatalysts showed an isotherm type IV according to the IUPAC (International Union of Pure and Applied Chemistry) classification [43], with a hysteresis loop typical of the presence of mesoporosity. The hysteresis loop in the isotherms of the TiO 2 (nt) and TiO 2 (T) samples is type H2 [43], indicating the presence of cylindrical pores with narrow and wide necks, and a certain contribution of "ink bottle" or bottle-type pores [44], while it is type H3 in the isotherm of P25, usually given by non-rigid aggregates of plate-like particles forming slit-like pores [44]. The calculated values of surface area and pore volume are collected in Table S2 (Supplementary Materials) showing that most of them are higher for the prepared samples than for P25, and decrease as the temperature of the heat treatment increases (Brunauer-Emmett-Teller (BET) surface area decreases from 296 to 45 m 2 /g, mesopore volume decreases from 0.17 to 0.07 cm 3 /g and total pore volume decreases from 0.32 to 0.09 cm 3 /g). In the literature this is usually explained as a consequence of the crystallite size growth upon heating [45,46]. The prepared TiO 2 materials have surface areas in the same order of magnitude, or even larger, than those of other materials prepared in similar conditions [46,47]. Figure 1b shows the N 2 adsorption-desorption isotherms of the prepared AC and of the TiO 2 (350) and TiO 2 -ACx (350) samples, while the calculated surface area and pore volume are included in Table S2. The N 2 adsorption-desorption isotherms of AC is type I, characteristic of an essentially microporous solid [43], for which the S BET is 491 m 2 /g, and those of the TiO 2 -ACx (350) samples are type IV with H2-type hysteresis loops, revealing that these samples are mesoporous. All of them show a rather similar adsorption capacity, slightly higher than that of the bare TiO 2 samples (from 144 for TiO 2 (350) up to 173 m 2 /g for the TiO 2 -AC10 (350) sample). The surface area of the TiO 2 -ACx samples is the expected one according to the proportion of TiO 2 and AC.

Textural Properties
The pore size distributions of the photocatalysts are compiled in Figure S2. They are very similar for all the prepared samples, showing a monomodal distribution ranging surface area of the TiO2-ACx samples is the expected one according to the proportion of TiO2 and AC.
The pore size distributions of the photocatalysts are compiled in Figure S2. They are very similar for all the prepared samples, showing a monomodal distribution ranging between 3 and 18 nm (mainly based on mesopores), and with a mean pore diameter similar for most samples, around 8-9 nm, except for TiO2 (nt).

X-ray diffraction (XRD) Analysis
The XRD patterns of P25, TiO 2 (nt), TiO 2 (T) and TiO 2 -ACx (350) samples ( Figure 2) show that P25 contains anatase and rutile, while the prepared photocatalysts contain only anatase [48,49]. surface area of the TiO2-ACx samples is the expected one according to the proportion of TiO2 and AC. The pore size distributions of the photocatalysts are compiled in Figure S2. They are very similar for all the prepared samples, showing a monomodal distribution ranging between 3 and 18 nm (mainly based on mesopores), and with a mean pore diameter similar for most samples, around 8-9 nm, except for TiO2 (nt).

X-ray diffraction (XRD) Analysis
The XRD patterns of P25, TiO2 (nt), TiO2 (T) and TiO2-ACx (350) samples (Figure 2) show that P25 contains anatase and rutile, while the prepared photocatalysts contain only anatase [48,49].    TiO2  TiO2 350  TiO2 400  TiO2 450  TiO2 500   10  20  30  40  50  60  70    Quantification of crystalline phases and amorphous TiO 2 in each sample was performed following the procedure reported by Jensen et al. [50], which is described in detail in a previous publication [23]. The mean crystallite size has been calculated as indicated in the experimental section. Table 1 shows the obtained results. Table 1. Crystalline and amorphous TiO 2 contents, as well as mean crystallite size, for P25, TiO 2 (nt) and TiO 2 (T) samples (Set 1) and for TiO 2 -ACx (350) (x = 0, 0.5, 1, 5, and 10 wt.%) samples (Set 2). Data of Set 1 in Table 1 show that TiO 2 (nt) and TiO 2 (T) samples present high crystallinity, although they are less crystalline than P25, which is consistent with their surface areas since a lower surface area could be associated with: higher crystallinity and/or larger crystallite size (grain coarsening) or with some agglomeration. Even the not thermally treated sample (TiO 2 (nt)) shows high crystallinity (62%). It is important to mention that the anatase content of the TiO 2 (T) samples is similar, or even slightly higher, than that of P25. As expected, the increase of the heat-treatment temperature leads to an increase of the anatase content (from 62 to 78%) and of the anatase crystallite size (from 6 to 16 nm). This increase in crystalline size could be accompanied by a reduction in the BET surface area (Table S2) values.

Set
Samples TiO 2 -ACx (350) ( Table 1, Set 2) contain, as well, only anatase. The anatase content and crystallite size is similar in TiO 2 -ACx (350) samples and in TiO 2 (350), meaning that the incorporation of AC does not significantly influence the crystallinity of the synthesized TiO 2 , independently of the AC wt.% content. The XRD pattern of the synthesized AC (reported in Supplementary Materials, Figure S3) corresponds to an amorphous material.

Scanning Electron Microscopy (SEM)
SEM images obtained for AC and for TiO 2 -ACx samples before the heat treatment at 350 • C are shown in Figure 3. Figure 3a shows that the obtained AC presents spherical morphology (with 6.25 ± 1.02 µm average diameter size). TiO 2 -ACx samples before the heat treatment are shown in Figure 3b-f. It is expected that this aspect does not significantly change after the heat treatment because it leads to very small loss of the carbon content. AC cannot be observed in the SEM images of the TiO 2 -AC5 (nt) and TiO 2 -AC1 (nt) samples, while it is clearly seen in the case of TiO 2 -AC5 (nt) and TiO 2 -AC10 (nt). It can be noted that TiO 2 particles are deposited on the AC spheres, and these two materials do not mix intimately (in agreement with the preserved TiO 2 crystallinity in the TiO 2 -AC samples). By comparison of the images in Figure 3a,e,f it can be inferred that the carbon spheres are covered by a TiO 2 layer in TiO 2 -ACx (nt) materials. AC5 (nt) and TiO2-AC10 (nt). It can be noted that TiO2 particles are deposited on the AC spheres, and these two materials do not mix intimately (in agreement with the preserved TiO2 crystallinity in the TiO2-AC samples). By comparison of the images in Figure 3a,e,f it can be inferred that the carbon spheres are covered by a TiO2 layer in TiO2-ACx (nt) materials.  Figure 4 shows the plots of the relative diuron concentration C/Cs versus time (where C and Cs are, respectively, diuron concentration at a certain time t and the initial one, at t = −120 min) measured at certain time intervals under dark and illumination conditions. Figure 4a shows data obtained in experiments carried out using TiO2 (nt), TiO2 (T) samples and P25 and in a blank experiment carried out to evaluate diuron photolysis, if any, while Figure 4b shows the results obtained with TiO2-ACx (350) sample, including TiO2 (350) for comparison. The plots are divided in two zones: the first 120 min (negative numbers in x-axis) correspond to dark conditions and from the time identified as 0, the system is irradiated. Diuron concentration at t = 0 is identified as C0. Plots of Ln(C0/C) vs. irradiation time are shown in Figure S4, Supplementary Materials.  Figure 4 shows the plots of the relative diuron concentration C/C s versus time (where C and C s are, respectively, diuron concentration at a certain time t and at the initial one, at t = −120 min) measured at certain time intervals under dark and illumination conditions. Figure 4a shows data obtained in experiments carried out using TiO 2 (nt), TiO 2 (T) samples and P25 and in a blank experiment carried out to evaluate diuron photolysis, if any, while Figure 4b shows the results obtained with TiO 2 -ACx (350) samples, including TiO 2 (350) for comparison. The plots are divided in two zones: the first 120 min (negative numbers in x-axis) correspond to dark conditions and from the time identified as 0, the system is irradiated. Diuron concentration at t = 0 is identified as C 0 . Plots of Ln(C 0 /C) vs. irradiation time are shown in Figure S4, Supplementary Materials.

Diuron Removal in Water
It can be observed that, in some cases, the relative concentration of diuron decreases under dark conditions, which is due to diuron adsorption on the catalysts surface and depends on the photocatalysts' properties. On one hand diuron molecular size and, on the other hand, the surface areas, pore volumes and, especially, pore size distribution of the photocatalysts (pore size in the range 3-18 nm, with average value around 8-9 nm) allow consideration that diffusional problems will not negatively affect the adsorption kinetics [51]. When the light is turned on, diuron concentration decreases much faster because of the photodegradation process. However, adsorption is, probably, also occurring and, therefore, two phenomena can simultaneously occur: diuron adsorption (in darkness and under illumination) and diuron photodegradation (only under illumination). It is also important to mention that as diuron adsorption depends on the sample, the herbicide concentration when the illumination is switched on is not the same in each experiment. Figure 4a shows that for the TiO 2 (nt) sample diuron concentration is reduced about 20% due to adsorption in the dark period. Such a reduction is about 10% in the case of TiO 2 (350) and P25, whereas it is almost negligible for TiO 2 samples treated at higher temperatures. Figure 4b shows that diuron adsorption in dark conditions notably increases with the carbon content and, in the case of the TiO 2 -AC5 (350) and TiO 2 -AC10 (350) samples, the adsorption equilibrium is not reached in the 2 h dark period.
The amount of adsorbed diuron is related with the photocatalysts' surface areas. A direct relationship between both parameters has been found for bare TiO 2 and TiO 2 -AC samples (Supplementary Materials, Figure S5), being the diuron adsorption capacity of the TiO 2 -ACx samples higher than that of bare TiO 2 samples with similar or even larger S BET . It can be pointed out that TiO 2 -AC10 (350) adsorbs 90.5% of diuron present in solution  It can be observed that in some cases, the relative concentration of diuron decreases under dark conditions, which is due to diuron adsorption on the catalysts surface and depends on the photocatalysts' properties. On one hand diuron molecular size and, on the other hand, the surface area, pore volumes and, especially, pore size distribution of the photocatalysts (pore size in the range 3-18 nm, with average value around 8-9 nm) allow consideration that diffusional problems will not negatively affect the adsorption kinetics [51]. When the light is turned on, diuron concentration decreases much faster because of the photodegradation process. However, adsorption is, probably, also occurring and, therefore, two phenomena can simultaneously occur: diuron adsorption (in darkness and under illumination) and diuron photodegradation (only under illumination). It is also important to mention that as diuron adsorption depends on the sample, the herbicide concentration when the illumination is switched on is not the same in each experiment. Figure 4a shows that for the TiO2 (nt) sample the diuron concentration is reduced about 20% due to adsorption in the dark period. Such a reduction is about 10% in the case of TiO2 (350) and P25, whereas it is almost negligible for TiO2 samples treated at higher temperatures. Figure 4b shows that diuron adsorption in dark conditions notably increases with the carbon content and in the case of the TiO2-AC5 (350) and TiO2-AC10 (350) samples, the adsorption equilibrium is not reached in the 2 h dark period.
The amount of adsorbed diuron is related with the photocatalysts' surface areas. A direct relationship between both parameters has been found for bare TiO2 and TiO2-AC samples (Supplementary Materials, Figure S5), being the diuron adsorption capacity of the TiO2-ACx samples higher than that of bare TiO2 samples with similar or even larger SBET. It can be pointed out that TiO2-AC10 (350) adsorbs 90.5% of diuron present in solution after 2 h in darkness. However, diuron adsorption is similar on P25 and on TiO2 (350), despite their different SBET values. This highlights that diuron adsorption extent in not only determined by the surface area.  Diuron photodegradation does not occur in absence of catalyst (Figure 4a), meaning that diuron photolysis can be neglected in these experimental conditions. The TiO 2 (nt) catalyst has shown to be inactive, while all TiO 2 (T) samples are photoactive, leading to similar diuron degradation rates. This means, on one hand, that the post-synthesis heat treatment is necessary to make the samples active and, on the other hand that either the differences between samples are too small to show any effect in the rate of diuron removal, or that there is a compensation of effects, i.e., larger surface area and lower crystallinity is equivalent to lower surface area and higher crystallinity. Many works claim that crystallinity, surface area, and size of TiO 2 particles affect the photocatalytic activity; however it is difficult to estimate the separate contribution of each of them [52,53].
As an example, products of diuron photodegradation obtained when using TiO 2 -AC10 (350) were analyzed by ionic chromatography. Short-chain organic acids (1.12 mg/L oxalic acid, 0.24 mg/L acetic acid, 0.06 mg/L formic acid) and chloride ions (0.3 mg/L) were detected in the solution, in agreement with the literature [1,54]. The low concentration of inorganic chloride (product of diuron dechlorination during the early stage of its photocatalytic degradation [55]) suggests that probably most of the chloride species remained adsorbed on the photocatalyst's surface.
Diuron photodegradation over illuminated TiO 2 has been modelled using the Langmuir-Hinshelwood model [56]. For highly diluted solutions, the kinetic model can be simplified to an apparent first-order equation: Ln(C 0 /C) = k app ·t [57], in which C 0 is the diuron concentration at t = 0, C is the diuron concentration at a certain time t and k app is the apparent rate constant (if photodegradation and adsorption take place simultaneously, k app is an apparent rate constant comprising both phenomena). Table 2 shows the k app values determined from the plots of Figure S4. It can be observed that k app for P25 is clearly higher than those determined for TiO 2 (T) samples. The higher activity of P25 could be related to its higher degree of crystallinity (87% vs. 76-78% of TiO 2 (T) samples), although TiO 2 (T) samples had smaller mean crystallite size and larger surface area. This underlines the importance of an accurate determination of the crystalline/amorphous proportion of the TiO 2 photocatalysts, and not only of the crystalline phase proportion within the crystalline fraction. Table 2. Apparent first-order rate constant (k app ) for diuron removal with the corresponding linear regression coefficients R 2 and diuron removal efficiency (%) in experiments shown in Figure 4. Regarding TiO 2 -ACx samples, data of Table 2, Set 2, show that k app increases with the carbon content and, consequently, with the adsorption capacity of these photocatalysts. The k app value is highest for TiO 2 -AC10 (350), 78.2 × 10 −3 min −1 , which surpasses the k app value obtained for P25 (63.4 × 10 −3 min −1 ). However, as the k app values have been calculated with different initial diuron concentrations (C 0 ), it is clear that the comparison between samples is not completely fair. Because of that, the photocatalysts behavior has also been measured in terms of diuron removal efficiency after 120 min in darkness and 20 min under irradiation (at 20 min irradiation time diuron removal is complete in one of the catalysts) and calculated as the decrease percentage of the initial diuron concentration (C s ) (by adsorption and photodegradation). Data are presented in Table 2, where it can be observed that for samples of Set 1, the highest diuron removal is achieved with P25 and TiO 2 (T) samples are more efficient than TiO 2 (nt), being TiO 2 (350) the one leading to the highest diuron removal, which can be related to its higher surface area. From samples of Set 2, TiO 2 -ACx (350) catalysts are more efficient for diuron removal than TiO 2 (T) counterparts, which is due to the important diuron adsorption in the carboncontaining hybrid photocatalysts. The percentage of diuron removed from water increases with the carbon content: samples TiO 2 -AC5 (350) and TiO 2 -AC10 (350) are more efficient than P25 (82.1 and 99.8 vs. 76.4% efficiencies, respectively, after 120 min in darkness and 20 min under irradiation). In the literature, such high photodegradation efficiencies in so short irradiation times have not been reported before for TiO 2 -based photocatalysts. Other research groups have studied the photocatalytic behavior of metal-TiO 2 catalysts in diuron photodecomposition. Menzi et al. [58] proved that after 60 min irradiation time, diuron photodegradation was 51% and 75% for pure TiO 2 and for Au-TiO 2 , respectively, while it reached 100% for the hybrid Pt-TiO 2 photocatalyst. Foura et al. [59] prepared a W-TiO 2 (3 wt.% W) that exhibited better photoactivity than pure TiO 2 (k app were 43.6 × 10 3 and 21.9 × 10 3 min −1 , respectively). Cruz et al. [60] studied the photodegradation of diuron using graphene oxide-TiO 2 catalyst and showed that 79 min of irradiation were needed to achieve 50% diuron conversion. Bamba et al. [32] studied TiO 2 photocatalysts containing activated carbon and reported a diuron removal efficiency of 99.95 and 98.61% after 90 min in darkness and 180 min under irradiation for P25 and TiO 2 -AC, respectively (TiO 2 -AC was prepared by sol-gel method, with a surface area of 206 m 2 /g and 99.2% of anatase). The apparent first-order rate constant for their TiO 2 -AC catalyst (with Ti/C ratio of 2/1) was 26 × 10 −3 min −1 , which is well below the k app values obtained in this investigation.

Set
In our work, photocatalysts with nominal AC loading beyond 10 wt.% were not studied because, on one hand, 10 wt.% AC leads to an important diuron adsorption and it is much more difficult to analyze the photocatalytic process and, on the other hand, a high AC loading could result in an important covering of the TiO 2 surface, limiting its capacity to absorb light.
In the case of the TiO 2 -AC10 (350) sample, diuron removal was also evaluated after adsorption periods in darkness of variable duration, ranging from 1 min to 360 min, and keeping the same irradiation time indicated above, 20 min. Figure 5 shows the evolution of the relative diuron concentration, C/C s , vs. time, and Table 3 compiles the calculated apparent first-order rate constant values. Data of Table 3 show that k app increases with the adsorption time in darkness up to 120 min (at this time more than 90% of diuron is adsorbed at the catalyst surface) and, then, it decreases for the sample remaining 360 min in darkness. In the case of the TiO2-AC10 (350) sample, diuron removal was also evaluated after adsorption periods in darkness of variable duration, ranging from 1 min to 360 min, and keeping the same irradiation time indicated above, 20 min. Figure 5 shows the evolution of the relative diuron concentration, C/Cs, vs time, and Table 3 compiles the calculated apparent first-order rate constant values. Data of Table 3 show that kapp increases with the adsorption time in darkness up to120 min (at this time more than 90% of diuron is adsorbed at the catalyst surface) and then, it decreases for the sample remaining 360 min in darkness.   Thus, to compare experiments with similar diuron concentration at t = 0 (C 0 ), the dark period should not be the same. Such a comparison has been done keeping samples TiO 2 (450) and TiO 2 (500) 2 h in darkness prior to illumination and sample TiO 2 -AC10 (350) only one minute ( Figure S7 (Supplementary Materials)). For TiO 2 (450) and TiO 2 (500) samples, adsorption equilibrium was reached before the light was turned on, meaning that under illumination diuron was removed from solution exclusively by its photocatalytic degrada-tion. In contrast, in the case of TiO 2 -AC10 (350) diuron adsorption and photodegradation simultaneously occurred under irradiation. Comparison of the calculated k app for these experiments (18.5 × 10 −3 , 20.5 × 10 −3 , 45.2 × 10 −3 min −1 for TiO 2 (450), TiO 2 (500) and TiO 2 -AC10 (350), respectively) shows that the k app for TiO 2 -AC10 (350) is clearly higher and, considering that k app for most of the bare or low-carbon TiO 2 samples is around 20 × 10 −3 , it can be estimated that the contribution of adsorption to the combined process is more that 50%.
Comparing all the tested samples, it can be concluded that the selection of an optimum material would depend on the operational procedure. If the system continuously operates under Ultraviolet A (UV-A) light irradiation, P25 would be the most efficient material. By contrast, if the operational procedure includes a period of time in darkness, in which adsorption can occur independently of degradation, TiO 2 -ACx (350) samples, particularly TiO 2 -AC10 (350), would be more effective for diuron removal.
Another point to highlight is that TiO 2 -AC samples can be considered quite robust, since the samples remain undamaged after the different processes in which they are used.

Total Organic Carbon (TOC) Measurements
Considering the relevance and extent of diuron adsorption on TiO 2 -ACx (350) hybrid photocatalysts, additional experiments were conducted to study the potential desorption of adsorbed diuron (or other intermediates of diuron degradation) and their subsequent photooxidation. Samples TiO 2 -AC1 (350) and TiO 2 -AC10 (350) were treated as follows: they were put in contact with 100 mL of aqueous solution of diuron (10 mg/L), maintained for 2 h in darkness and then, irradiated for 1 h. In these conditions the complete removal of diuron from solution occurred with both samples (see Supplementary Materials, Figure S8). Afterwards, the solids were recovered by filtration and mixed with 100 mL ultrapure water under stirring. The suspension was further irradiated, and the total organic carbon (TOC) in the solution was periodically measured (Supplementary Materials, Figure S9). For both samples, an increase of the TOC value was observed, in the first 30 min in the case of TiO 2 -AC1 (350), and in a period of about 180 min in the case of TiO 2 -AC10 (350), followed by a subsequent decrease. This reveals that diuron, or other partially oxidized organic intermediate compounds, desorb from the catalysts, and are further photocatalytically degraded. As expected, the number of desorbed species is larger for TiO 2 -AC10 (350) than for TiO 2 -AC1 (350). These results indicate that illumination could be used to regenerate TiO 2 -carbon hybrid photocatalysts and so, they could be used again. Additional tests would be needed to state up to which extent the photocatalytic activity is kept after several regeneration cycles.

Preparation of TiO 2 and TiO 2 -AC Photocatalysts
TiO 2 and TiO 2 -AC photocatalysts were prepared by sol-gel as follows [61]: 9.3 mL titanium tetraisopropoxide (TTIP) (97%, Sigma-Aldrich, St. Louis, MO, USA) were mixed with 17.5 mL glacial acetic acid (HAc) (99%, Sigma-Aldrich) at 0 • C. Then, to prepare TiO 2 materials, 197.5 mL distilled water were added dropwise under vigorous stirring. To obtain hybrid TiO 2 -AC photocatalysts, a certain amount of activated carbon (see next section) (0.01, 0.02, 0.13 or 0.27 g, to obtain catalysts with nominal 0.5, 1, 5, and 10 wt.% AC, respectively) was added to 5 mL water, and this mixture was incorporated dropwise to the TTIP/HAc solution; then, water was added dropwise up to a total volume of 197.5 mL. Afterwards, in both preparations the mixture was ultrasonicated for 30 min, stirred for 5 h and the formed suspension was aged in an oven at 70 • C for 12 h. The solid product was dried at 100 • C for 12 h and crushed into a fine powder. Then, the obtained TiO 2 was heat-treated in air at 350, 400, 450 or 500 • C for 2 h (muffle, 5 • C/min) to increase the crystallinity and to remove carbon residues from the synthesis process. The nomenclature used is TiO 2 (T), where T is the temperature of the heat treatment. A non-treated sample named TiO 2 (nt), where nt means not treated, was also studied. The TiO 2 -AC catalysts were submitted to a post-synthesis heat treatment in air at 350 • C for 2 h (muffle, 5 • C/min). This temperature has been chosen because it allows obtaining a good compromise between the crystallinity of the TiO 2 -AC materials and their carbon content remaining after treatment. The samples were named TiO 2 -ACx (350), where x refers to the nominal AC wt.% indicated above. Degussa Aeroxide P25 TiO 2 has been used as a reference material.

Preparation of Activated Carbon
The activated carbon (AC) was prepared by hydrothermal carbonization of saccharose (99.5%, Sigma-Aldrich) as follows [62]: 12 g saccharose were mixed with 21 mL water (1.6 M solution). Then, the mixture was transferred to a 50 mL Teflon-lined stainlesssteel autoclave and heated in an oven at 180 • C for 12 h. After cooling down to room temperature, the solid product was recovered by filtration, washed with distilled water and dried (110 • C, 5 h). Finally, it was activated with CO 2 (80 mL/min) at 800 • C (5 • C/min heating rate) for 5 h.

Photocatalysts' Characterization
Porosity and specific surface area of the prepared photocatalysts were characterized by N 2 adsorption-desorption at −196 • C (after outgassing at 250 • C, 4 h) in a Quantachrome Autosorb-6B equipment (Quantachrome Instruments, Boynton Beach, FL, USA). The specific BET surface area (S BET ) and the micropore volume (V N2 ) were determined by applying the Brunauer-Emmett-Teller (BET) and the Dubinin-Radushkevich equations to the N 2 adsorption isotherms, respectively. The mesopore volume (V meso ) was estimated by the difference of the volume of N 2 adsorbed as liquid at P/P 0 = 0.9 and P/P 0 = 0.2, while the total pore volume (V T ) was determined as the volume of N 2 adsorbed at P/P 0 = 0.99 [63][64][65]. The pore size distribution was calculated using the adsorption branch from the N 2 isotherm, by using BJH (Barrett-Joyner-Halenda) method.
XRD analysis was performed in a SEIFERT 2002 equipment (Rich. Seifert & Co., Ahrensburg, Germany), using Cu Kα (1.54 Å) radiation, in the 6-80 • 2θ range, with a scanning velocity of 2 • /min. The average crystallite size, referred to as crystal size, was calculated by Scherrer's equation (Equation (1)) [66]: where B is the average crystallite size (nm); K is a constant with a value of 0.9 [67]; λ is the wavelength of the radiation source; β is the full width at half maximum intensity (FWHM) and θ is the angle associated with the main peak corresponding to the studied phase (2θ values of 25.3 • and 27.5 • for anatase (A) and rutile (R), respectively). The amount of anatase and rutile and the percentage of amorphous TiO 2 have been calculated as explained in a previous work [23]. The actual carbon content of the TiO 2 -ACx (350) samples was determined by thermogravimetry (TA Instruments, New Castle, DE, USA) (TG SDT Q600, air flow (50 mL/min), 10 • C/min up to 900 • C).
Scanning Electron Microscopy (SEM, JSM-840 microscope (JEOL, Tokyo, Japan)) with scintillator-photomultiplier type secondary electron detector was used to analyze the morphology of the obtained AC and TiO 2 -ACx (nt) samples.

Photocatalytic Activity Measurements
The experiments were carried out, under stirring, inside a Suntest XLS+ reaction chamber (Atlas Material Testing Technology BV, Gelnhausen, Germany) equipped with a 1700 W Xenon arc lamp and a Solar ID65 filter to limit the UV radiation at 320 nm for simulating solar exposition, according to ICH Q1B guidelines. The spectral distribution is reported in Figure S10 in Supplementary Materials.
A batch beaker-type glass reactor was used at atmospheric pressure. In each experiment, 100 mg of photocatalyst were dispersed under stirring in 100 mL of a 10 mg/L aqueous solution of diuron (C 9 H 10 Cl 2 N 2 O, 98%, Sigma-Aldrich). The suspension was exposed to simulated solar light at 250 Wm −2 , which corresponds to an average solar radiation in a summer day in southern Europe [68]. The distance between the light source and the reactor was 32 cm. All the tests were performed using 1 g/L photocatalyst concentration since it has been shown to be the optimum one for P25 [69].
Prior to irradiation, the catalyst suspension was stirred for 2 h in darkness to ensure the adsorption/desorption equilibrium (unless a different specific protocol was mentioned). 1 mL solution was periodically sampled, filtered (0.20 µm filter (Chromafil ® PET-20/25)), and analyzed by UV-Vis spectrophotometry (CARY 100 Scan UV-visible spectrophotometer, Agilent Technologies, Santa Clara, CA, USA). The process was followed by monitoring the intensity of the main absorption peak, at λ = 248 nm. Total organic carbon (TOC) was measured using a Shimadzu TOC-L analyzer. In some cases, solutions were also analyzed by ionic chromatography (Metrohm 790 IC) using a conductivity detector.

Conclusions
TiO 2 and TiO 2 -carbon photocatalysts have been synthetized by sol-gel, being the first submitted to a post-synthesis heat treatment at different temperatures, and the second prepared with different amounts of activated carbon (AC) and heat-treated at 350 • C.
The bare TiO 2 samples contain only anatase (about 62-78%) and, as the temperature of the heat treatment increases, the surface area decreases and the anatase crystallite size increases, being the anatase content for samples treated between 350 and 500 • C essentially constant. The surface area of the TiO 2 -carbon samples increases with the carbon content, and the incorporation of AC does not affect TiO 2 crystallinity.
Diuron adsorption plays an important role in the photocatalysts behavior. Results obtained with bare TiO 2 samples reveal that diuron adsorption extent is not only determined by the surface area. In contrast, a clear relationship between surface area and diuron adsorption has been observed in TiO 2 -ACx (350) catalysts. P25 shows the highest photocatalytic activity for diuron degradation, which can be explained by its higher crystallinity. However, the largest diuron removal, 99.8%, was achieved with the TiO 2 sample prepared with 10 wt.% AC. It is due to the combination of adsorption and photodegradation, being this an interesting result in comparison with those previously published in the literature, especially for the low illumination time used.
Our results highlight that for this application, the selection of an optimum material would depend on the operational procedure: P25 shows the best performance in illumination conditions, but carbon-containing hybrid samples are very efficient if the process includes dark periods in which adsorption takes place, even more considering that adsorbed diuron can afterwards be desorbed and oxidized. This points out to a potential regeneration procedure that merits being further investigated.