Biochar Nanoparticles over TiO 2 Nanotube Arrays: A Green Co-Catalyst to Boost the Photocatalytic Degradation of Organic Pollutants

: Biochar nanoparticles (BC NPs), produced by low temperature pyrolysis (350 ◦ C) of microalgae ( Nannochloropsis sp.) and nutshells, are proposed as low-cost and sustainable co-catalysts to promote the photocatalytic activity of TiO 2 nanotube (NT) arrays towards the degradation of methylene blue (MB) used as an organic pollutant model molecule. BC NPs (size < 25 nm) were obtained by treating bulk BC (i.e., biomass after pyrolysis) by sonication–centrifugation cycles in a water solution. The ﬁltered BC NPs dispersion was deposited by simple drop-casting on the TiO 2 NT support. The BC loading was varied by performing multiple depositions. Photocatalytic experiments under UV light (365 nm) revealed that the decoration with BC NPs signiﬁcantly improves the TiO 2 photoactivity. Such enhancement is mainly inﬂuenced by the amount of BC deposited; upon optimizing the BC deposition conditions, the rate of photocatalytic degradation of MB increases approximately three times with respect to bare TiO 2 , almost irrespective of the nature of the raw material. The greater photocatalytic activity of BC-TiO 2 can be attributed to the synergistic combination of reactant/product adsorption and catalytic degradation of the adsorbed organic pollutant, as well as an improved charge carrier separation and electron transfer.


Introduction
With the accelerated expansion of worldwide industrialization and urbanization, environmental pollution is increasingly becoming a serious global issue faced by humankind [1]. Therefore, green, cost-effective, and efficient ways to resolve this adverse situation have been extensively studied by researchers in recent decades. Among the various strategies conceived to solve this global problem, heterogeneous photocatalysis, in which the naturally available and clean energy of the sun can be exploited by semiconductors (e.g., TiO 2 , α-Fe 2 O 3 , ZnO), has emerged as one of the most promising technologies, especially for water purification [2,3]. Photocatalytic materials have been widely studied for the degradation and removal of many inorganic [4][5][6] and organic [7,8] pollutants in natural and waste waters.
In this context, biochar (BC), a porous carbonaceous material produced from biomass feedstocks via pyrolysis, has recently been proposed as a greener supporting carbon material for powdered photocatalytic platforms [26][27][28] and, in a fewer reports, as cocatalytic nanoparticles supported on nanostructured TiO 2 [29]. As one main advantage, the employment of BC allows for strongly reducing the environmental footprint and costs associated with the photocatalyst synthesis compared to traditional noble metal-and carbon-based materials; its production is highly sustainable, as waste biomasses can be used as raw materials [30,31]. Many other advantages of using biochar are reported in the literature concerning the implementation of renewable energies as easy metal recovery [32] and its application in fuel cells [33]. The unique physicochemical properties of BC make it an ideal candidate for (photo)catalytic applications due to its high specific surface area, easily tunable functional groups, chemical stability, and electrical conductivity [27,34,35]. The coupling of BC with TiO 2 can therefore offer a synergistic combination of reactant/product adsorption and catalytic degradation of the adsorbed pollutants, as well as an improved charge carrier separation and electron transfer system, which result in a significantly enhanced overall photocatalytic efficiency [27,36].
The properties of BC can be tuned by changing pyrolysis temperature and biomass feedstocks [37][38][39] in order to reach the best (photo)catalytic performances. Highly conductive, graphitized BC produced at high temperatures (800-1000 • C), characterized by a high specific surface area and porosity, is often preferred for catalytic applications, as BC is commonly employed as a support for photocatalyst nanoparticles [27,40]. Activating agents (e.g., KOH, H 3 PO 4 ) are also employed to further promote the creation of pores and to enable the formation of surface functional groups, which are partially lost at high temperatures due to the graphitization of the material [41]. At temperatures above 500-600 • C, in fact, the degradation of organic matter with stronger chemical bonds occurs, leading to a loss of functionalization. Despite their high performances, the required harsh pyrolysis conditions drastically reduce (i) the environmental friendliness of BC fabrication, (ii) BC yield [37], and (iii) the natural self-doping of BC (which strictly depends on the feedstock chemical composition and pyrolysis temperature) [42]. In contrast, the employment of milder conditions, e.g., chemical-free slow pyrolysis at low temperatures (<400 • C), can provide a solution to these limitations leading to high production yields of naturally functionalized and self-doped BC with improved adsorptive properties [43][44][45][46]. Considering that functionalized materials are highly sought in (photo)catalytic applications, the combination of low temperature pyrolysis with heteroatom-rich biomass feedstocks, e.g., microalgae, should be beneficial, as highly functionalized biochar with improved adsorption properties is produced. However, the limited specific surface area of BC produced at low temperature hinders its typical employment as supporting material for powdered photocatalysts, and that is why it is more rarely studied. Downsizing of BC to quasi-spherical nanoparticles, e.g., by forming carbon dots by sonication, centrifugation, and filtration [47]), may instead open the possibility to exploit such highly functionalized BC nanoparticles as co-catalytic material deposited onto nanostructured photocatalytic layers; the BC downsizing process results in a greatly increased specific surface area and enhanced electronic and optical properties [47,48]. Moreover, the production of BC NPs should be not considered less green than bulk materials, as the employment of chemicals or harsh conditions (that may hinder the sustainability of the production process) is not involved. In this light, BC NPs can be deemed as a green co-catalyst that can enable the possibility to fabricate a supported photocatalytic layer rather than conventional powdered photocatalytic platforms produced by using bulk biochar.
In light of these considerations, BC nanoparticles produced by the slow pyrolysis at a low temperature (350 • C) of microalgae (Nannochloropsis sp.) and nutshell were used in this work as a co-catalyst on the surface of anodic TiO 2 nanotube arrays (TiO 2 NTs) to boost the photocatalytic activity towards the degradation of methylene blue, selected as an organic pollutant model. These biomass feedstocks were selected because of their markedly different composition, with microalgae-derived BC being noticeably more polar than the nutshell-derived one. The investigation of these diverse materials aims at evaluating the effect of the nature of the starting biomass feedstock on the promotion of the catalytic performances. The influence of other parameters such as BC loading and photocatalytic conditions were evaluated, too.

Bulk BC Characterization
Microalgae-and nutshell-derived biochar used for the fabrication of BC-TiO 2 photocatalysts were produced and fully characterized in a previous work [46]. Briefly, BC having high surface rugosity (see SEM images in Figure S1) was produced with quite high yield by using both lignocellulosic and microalgae feedstocks (34 ± 2% and 41 ± 4% yields, respectively). Concerning the elemental composition of these materials, heteroatom content (P and N) of MBC is tenfold of that observed for NBC ( Figure S2): this evidence is coherent with the higher amount of protein and lipids in microalgae. Furthermore, a more marked presence of aromatic C-H, as well as of -OH, C=O, C-O-C, and N-H functional groups, was found in MBC compared to NBC, in which no carbonyl bands were detected. Such features can be ascribed to the diverse nature of the feedstocks and to the catalytic effect of alkali (Na and K) and alkaline earth (Mg and Ca) metals during pyrolysis (see Figure S2 for alkali and alkaline earth contents); these chemicals, massively present in microalgae, can promote the formation of light oxygenates via ring fragmentation reactions of carbohydrates [46]. Finally, XRD analysis showed that no crystalline phases were formed during pyrolysis, most likely due to the low temperature of the thermal treatment [41].

Characterization of TiO 2 and BC-TiO 2 Photocatalysts
After pyrolysis, bulk BC powder was dispersed in ultrapure water (10 g/L) and sonicated for 1 h to induce the partial breaking of BC to nanosized particles. Then, the so obtained BC suspension was centrifuged, and the supernatant was filtered (0.2 µm pore size membrane) and collected. A schematic representation of the BC NPs fabrication route is depicted in Figure 1. As shown in Figure 2, BC NPs having an average size of ~2 nm are present in the filtered aqueous solution. This suspension was drop-casted onto the TiO2 support ( Figure  1) with the aim of forming composite BC NPs-TiO2 materials.   As shown in Figure 2, BC NPs having an average size of ~2 nm are present in the filtered aqueous solution. This suspension was drop-casted onto the TiO2 support ( Figure  1) with the aim of forming composite BC NPs-TiO2 materials. Cross-section and top-view SEM images of TiO2 nanotube arrays used in this work are shown in Figure 3. The morphology of tubes appears well-ordered: an inner diameter of 104 ± 10 nm and a length of 5.6 ± 0.5 µm were measured over the entire surface of the photocatalyst. Such length, easily tunable by adjusting the anodization time, was selected as it proved to be a good compromise between high light harvesting and limited charge carrier recombination [49]. Cross-section and top-view SEM images of TiO 2 nanotube arrays used in this work are shown in Figure 3. The morphology of tubes appears well-ordered: an inner diameter of 104 ± 10 nm and a length of 5.6 ± 0.5 µm were measured over the entire surface of the photocatalyst. Such length, easily tunable by adjusting the anodization time, was selected as it proved to be a good compromise between high light harvesting and limited charge carrier recombination [49]. As shown in Figure 2, BC NPs having an average size of ~2 nm are present in the filtered aqueous solution. This suspension was drop-casted onto the TiO2 support ( Figure  1) with the aim of forming composite BC NPs-TiO2 materials. Cross-section and top-view SEM images of TiO2 nanotube arrays used in this work are shown in Figure 3. The morphology of tubes appears well-ordered: an inner diameter of 104 ± 10 nm and a length of 5.6 ± 0.5 µm were measured over the entire surface of the photocatalyst. Such length, easily tunable by adjusting the anodization time, was selected as it proved to be a good compromise between high light harvesting and limited charge carrier recombination [49]. These one-dimensional TiO 2 nanostructures were decorated with MBC or NBC nanoparticles produced by multiple sonication-centrifugation steps, as described in the Materials and Methods section. As shown in Figure 4a,e, BC nanoparticles (diameter < 25 nm) are present at the surface of the TiO 2 nanotubes, already after a single drop-casting deposition (samples 1-MBC-TiO 2 and 1-NBC-TiO 2 ). This evidence suggests the suitability of the employed strategy for the production and deposition of BC nanoparticles.
As can be observed in Figure 4, the number and density of carbonaceous nanoparticles on the surface of the NTs arrays increase along with the number of depositions. This is particularly clear for the NBC-TiO 2 samples (see Figure 4e These one-dimensional TiO2 nanostructures were decorated with MBC or NBC nanoparticles produced by multiple sonication-centrifugation steps, as described in the Materials and Methods section. As shown in Figure 4a,e, BC nanoparticles (diameter < 25 nm) are present at the surface of the TiO2 nanotubes, already after a single drop-casting deposition (samples 1-MBC-TiO2 and 1-NBC-TiO2). This evidence suggests the suitability of the employed strategy for the production and deposition of BC nanoparticles.  No significant difference of the nanoparticle morphology was observed as a function of the number of depositions (see size distributions in Figure 5), since the shape of the BC co-catalysts remained nearly spherical. Only the nature of the feedstock seems to influence the BC NPs' morphology; as depicted in Figure 5, NBC NPs are larger and exhibit a wider dimensional distribution (24.2 ± 7.8 nm) with respect to MBC ones (15.3 ± 4.0 nm). Such NP size indicates the occurrence of an agglomeration process during drop-casting deposition, as NPs suspended in the aqueous solution have a much lower dimension (~2 nm, see Figure 2). The difference between NBC and MBC NPs in terms of agglomeration degree can be ascribed to the diverse hydrophilicity of the materials. In detail, NBC is more hydrophobic than MBC, most likely due to the absence of a polar surface functional group [46]; according to the literature, the more hydrophobic BC NPs are, the higher their tendency to agglomerate in clusters is [50]. Such a consideration is also well in line with the much higher BC loading observed in NBC-TiO 2 materials. As a confirmation of the occurrence of the agglomeration process, only a few distinct NPs are discernable when carrying out the seven drop-casting depositions (7-MBC-TiO 2 and 7-NBC-TiO 2 ); most of the deposited carbonaceous material clustered in a homogeneous thin film covering the surface of the TiO 2 nanotubes (see Figure 4d,h).  upon prolonged thermal treatments or at higher temperatures [51,52]. As expected, XRD performed on BC-decorated samples ( Figure 6, green and red lines) showed that no TiO 2 phase modifications are induced by the deposition of biochar. Moreover, no peaks that could be ascribed to the presence of crystalline carbon phases are visible, likely owing to the low amount and amorphous nature of the synthesized biochar.  Diffusive reflectance UV-vis spectra were also recorded for all the samples (Figures 7 and S4) to investigate the potential effect of BC NPs in inducing a redshift in the light absorption spectrum from UV to the visible region: this feature is sometimes reported for bulk BC when used as supporting material for powdered TiO 2 photocatalysts [36]. As expected, all the samples show high absorbance values for wavelengths lower than 380 nm, in line with the anatase TiO 2 composition of the NTs. Moreover, periodic absorption signals are always present even in the undecorated TiO 2 sample: this spectroscopic effect is most likely due to scattering inside the nanotube array [53] as already reported in previous works [54,55]. Interesting differences were found when decorating TiO 2 with BC NPs. In detail, 5-NBC-TiO 2 presents a broad and weak absorption peak in the 400-600 nm range, not present in any other sample, and the 5-MBC-TiO 2 sample presents a peculiar peak with a relative maximum absorbance at 331 nm. The presence of this peak can be ascribed to the π-π* absorption typical of the aromatic system [56]; this assignment is also strengthened by the evidence of aromatic absorptions in the ATR spectrum recorded on such material [46]. Regarding the broad and weak absorption band present in the nutshell-derived sample, an analogous relatively weak absorbance in the visible region is widely reported in the literature [28,36,57,58], and this signal is generally attributed to the biochar content and to the interaction of the TiO 2 with BC (Ti-O-C bonds). The low intensity of this absorption signal is reasonably ascribed to the low BC loading on the photocatalyst. Despite these differences in the UV absorption spectra, the energy band gaps (E g ), estimated using the Kubelka-Munk method (see inset in Figure 6) for TiO 2 and BC-TiO 2 materials, show no significant differences, ruling out the occurrence of shifts in the E g , which remains in the UV light region even after decoration with BC NPs (E g~3 .3 eV for all the samples).
attributed to the biochar content and to the interaction of the TiO2 with BC (Ti-O-C bonds). The low intensity of this absorption signal is reasonably ascribed to the low BC loading on the photocatalyst. Despite these differences in the UV absorption spectra, the energy band gaps (Eg), estimated using the Kubelka-Munk method (see inset in Figure 6) for TiO2 and BC-TiO2 materials, show no significant differences, ruling out the occurrence of shifts in the Eg, which remains in the UV light region even after decoration with BC NPs (Eg ~ 3.3 eV for all the samples).

Assessment of the Photocatalytic Performances
Prior to proceeding with photocatalytic experiments, MB stability under UV irradiation was preliminary carried out. A 10 mg/L dye solution was stirred and irradiated by UV light in the absence of any photocatalyst; as a result, no decrease of MB concentration was observed after three hours, demonstrating that MB does not undergo decomposition

Assessment of the Photocatalytic Performances
Prior to proceeding with photocatalytic experiments, MB stability under UV irradiation was preliminary carried out. A 10 mg/L dye solution was stirred and irradiated by UV light in the absence of any photocatalyst; as a result, no decrease of MB concentration was observed after three hours, demonstrating that MB does not undergo decomposition under these experimental conditions. Then, the presence of no light-assisted adsorption/degradation processes at the photocatalyst surface was investigated; tests were carried out by stirring the MB solution under dark conditions in the presence of bare TiO 2 and BC-TiO 2 photocatalysts. Even in these cases, no appreciable decreases of MB concentrations were observed for all tested catalysts after three hours using the bare or five times decorated materials (TiO 2 , 5-NBC-TiO 2 and 5-MBC-TiO 2 ). The absence of remarkable absorption phenomena can be ascribed to the low amount of BC deposited on the TiO 2 nanotubes.
The photocatalytic tests were then performed to assess the effects of the BC NPs' loading on the photocatalytic performances. It should be underlined that mild photocatalytic conditions were intentionally used to enhance differences between each system and to better analyze the kinetic of MB degradation; experiments under boosted conditions are discussed later.
A kinetic study over short-term experiments (three hours) was preliminary carried out. As shown in Figure 8, the presence of BC NPs significantly enhances the photocatalytic activity towards the degradation of MB, whatever their origin and amount. This evidence clearly demonstrates the suitability of using microalgae-and nutshell-derived BC NPs as co-catalysts.
Significant differences in photocatalytic efficiency were found by varying both the nature of the biomass feedstock and the loading of BC NPs on TiO 2 nanotubes. For both systems, the degradation rate raises with the amount of co-catalyst (i.e., with the number of drop-casting depositions), reaching a maximum after five depositions (see Figure 8a,b); compared to undecorated TiO 2 , 2.5-and 2.8-times larger amounts of MB were removed after three hours by 5-MBC-TiO 2 and 5-NBC-TiO 2 , respectively. When loading more carbonaceous material, i.e., by performing the drop-casting procedure seven times, a decrease in photocatalytic efficiency was observed (see samples 7-MBC-TiO 2 and 7-NBC-TiO 2 in Figure 8). These experimental findings are well in line with the typical bell-shaped behavior reported for the photocatalytic efficiency when varying the loading of co-catalysts [14]. compared to undecorated TiO2, 2.5-and 2.8-times larger amounts of MB were removed after three hours by 5-MBC-TiO2 and 5-NBC-TiO2, respectively. When loading more carbonaceous material, i.e., by performing the drop-casting procedure seven times, a decrease in photocatalytic efficiency was observed (see samples 7-MBC-TiO2 and 7-NBC-TiO2 in Figure 8). These experimental findings are well in line with the typical bell-shaped behavior reported for the photocatalytic efficiency when varying the loading of co-catalysts [14]. Concerning the reaction pathway, since it was impossible to observe MB degradation via adsorption mechanism (i.e., in the absence of light irradiation) or photodecomposition of MB itself (i.e., in the absence of the photocatalyst), it appears clear that the MB degradation mechanism involves a photocatalytic process on the surface of BC-TiO2 materials. It is widely accepted that the overall process involves the formation of superoxide radicals, generated by a reaction with holes at the semiconductor surface, which quickly react with the dye molecules, breaking them down into smaller molecular fragments and eventually to CO2 and water [36]. This process is greatly enhanced by the sorption of MB on Concerning the reaction pathway, since it was impossible to observe MB degradation via adsorption mechanism (i.e., in the absence of light irradiation) or photodecomposition of MB itself (i.e., in the absence of the photocatalyst), it appears clear that the MB degradation mechanism involves a photocatalytic process on the surface of BC-TiO 2 materials. It is widely accepted that the overall process involves the formation of superoxide radicals, generated by a reaction with holes at the semiconductor surface, which quickly react with the dye molecules, breaking them down into smaller molecular fragments and eventually to CO 2 and water [36]. This process is greatly enhanced by the sorption of MB on the BC NPs' surfaces. Additionally, as metal particles (e.g., Pt) do when decorated on a semiconductor surface, biochar can shuttle and trap electrons [27]; this results in an improvement of charge carrier separation in the semiconductor, thus reducing the charge carrier recombination rate [28]. Based on the information present in literature, it can be reasonably supposed that the photocatalytic process occurring at the surface of the prepared materials is the combination of an adsorption process on BC nanoparticles and an oxidation reaction happening due to the presence of holes on the surface of the TiO 2 nanotubes array. In this light, it is reasonable to suggest that the decomposition process is enhanced for samples which show isolated BC nanoparticles, while the formation of a BC film might lead to a decrease in photocatalytic activity due to a reduction of TiO 2 surface available for the oxidation process. This interpretation explains the lower photocatalytic activity of 7-BC-TiO 2 samples compared to that of 5-BC-TiO 2 ones; in the former, a dense carbonaceous film completely covers the surface of TiO 2 (where the oxidation of MB most likely takes place), whereas the latter shows well-defined BC nanoparticles and a more available TiO 2 surface.
Based on these results, the most active materials, i.e., 5-NBC-TiO 2 and 5-MBC-TiO 2 , were studied in view of their photocatalytic activity during longer experiments (up to 51 h) to assess their (photo) stability and gain insights into the degradation kinetics. MB degradation profiles are shown in Figure 9.
As can be observed for the long kinetics experiments, the MBC-decorated TiO 2 is the most efficient photocatalyst for the degradation of MB, doubling the reaction rate compared to bare TiO 2 (see Figure 10) and completely mineralizing the dye within 48 h. In contrast, undecorated and nutshell-derived materials were unable to completely degrade the azo dye and exhibited a less efficient behavior, even though 5-NBC-TiO 2 performs even better than 5-MBC-TiO 2 within the first 3 h (see Figures 8 and 10). This behavior might be ascribed to a poisoning effect induced by methylene blue which is more marked for TiO 2 and 5-NBC-TiO 2 rather than for 5-MBC-TiO 2 . A second 3-h MB degradation was then performed on the catalysts used to assess if a partial deactivation could be observed. As can be seen in Figure 10, both TiO 2 and 5-NBC-TiO 2 show a marked deactivation, while 5-MBC-TiO 2 shows only a slight decrease in its photocatalytic efficiency. carbonaceous film completely covers the surface of TiO2 (where the oxidation of MB most likely takes place), whereas the latter shows well-defined BC nanoparticles and a more available TiO2 surface.
Based on these results, the most active materials, i.e., 5-NBC-TiO2 and 5-MBC-TiO2, were studied in view of their photocatalytic activity during longer experiments (up to 51 h) to assess their (photo) stability and gain insights into the degradation kinetics. MB degradation profiles are shown in Figure 9. As can be observed for the long kinetics experiments, the MBC-decorated TiO2 is the most efficient photocatalyst for the degradation of MB, doubling the reaction rate compared to bare TiO2 (see Figure 10) and completely mineralizing the dye within 48 h. In contrast, undecorated and nutshell-derived materials were unable to completely degrade the azo dye and exhibited a less efficient behavior, even though 5-NBC-TiO2 performs even better than 5-MBC-TiO2 within the first 3 h (see Figures 8 and 10). This behavior might be ascribed to a poisoning effect induced by methylene blue which is more marked for TiO2 and 5-NBC-TiO2 rather than for 5-MBC-TiO2. A second 3-h MB degradation was then performed on the catalysts used to assess if a partial deactivation could be observed. As can be seen in Figure 10, both TiO2 and 5-NBC-TiO2 show a marked deactivation, while 5-MBC-TiO2 shows only a slight decrease in its photocatalytic efficiency. The reason behind the different behavior of the biochar materials can be ascribed to their different chemical composition; as reported in Section 2.1, the microalgae derived biochar presents a higher content in the oxygenated functional group and has a higher heteroatoms content altogether that results in a higher photocatalytic performance (likely due to a reduced recombination rate for electron-holes pairs). For the lignocellulosic-derived nanoparticles, it can instead be supposed that the enhanced photocatalyst activity is due to a higher MB adsorption compared to bare TiO2; however, the lack of oxygenated functional groups and heteroatoms results in an overall less active material.
Concerning the kinetic model, data fit with a first order model ( Figure 9b); thus, we reasonably assumed that the mechanism is comparable to what is reported in the literature: MB molecules are adsorbed on BC, and MB degradation proceeds via reaction with superoxide radicals produced from the reaction between water and photogenerated holes [36].

Comparison of MB Degradation Performances with Literature Data
The two best performing materials (5-MBC-TiO2 and 5-NBC-TiO2) and pristine TiO2 NTs were tested under conditions closer to those most commonly reported in the literature for MB degradation, i.e., higher light power density and lower MB/catalyst ratios. These experiments were carried out to compare the performance of our materials against similar ones reported in the literature. It can be observed in Figure 11 that both MBC-and NBC-derived materials could completely degrade methylene blue within 3 h and did not exhibit a remarkable difference in photocatalytic performances. The reason behind the different behavior of the biochar materials can be ascribed to their different chemical composition; as reported in Section 2.1, the microalgae derived biochar presents a higher content in the oxygenated functional group and has a higher heteroatoms content altogether that results in a higher photocatalytic performance (likely due to a reduced recombination rate for electron-holes pairs). For the lignocellulosicderived nanoparticles, it can instead be supposed that the enhanced photocatalyst activity is due to a higher MB adsorption compared to bare TiO 2 ; however, the lack of oxygenated functional groups and heteroatoms results in an overall less active material.
Concerning the kinetic model, data fit with a first order model ( Figure 9b); thus, we reasonably assumed that the mechanism is comparable to what is reported in the literature: MB molecules are adsorbed on BC, and MB degradation proceeds via reaction with superoxide radicals produced from the reaction between water and photogenerated holes [36].

Comparison of MB Degradation Performances with Literature Data
The two best performing materials (5-MBC-TiO 2 and 5-NBC-TiO 2 ) and pristine TiO 2 NTs were tested under conditions closer to those most commonly reported in the literature for MB degradation, i.e., higher light power density and lower MB/catalyst ratios. These experiments were carried out to compare the performance of our materials against similar ones reported in the literature. It can be observed in Figure 11 that both MBC-and NBC-derived materials could completely degrade methylene blue within 3 h and did not exhibit a remarkable difference in photocatalytic performances. A comparison of the performances of our materials against the BC-based ones reported in the literature is shown in Table 1. It should be underlined that, at present, only few articles report on MB degradation over BC-based photocatalysts. As a matter of fact, we did not find any data on BC-based planar photocatalysts, and thus only materials in powder form were considered for this comparison. In this respect, our systems represent the first examples of BC-decorated photocatalytic layers for this kind of application. Both microalgae-and nutshell-derived materials proposed in this work provide MB abatement rates comparable to or even higher than those reported for other BC-based photocatalysts. This is particularly interesting considering the much lower amount of co-catalyst present on our planar photocatalyst (in the range of micrograms) compared to that on reported materials. In any case, such an evaluation should be carefully considered owing to the quite different experimental conditions and photocatalyst form (i.e., planar platform vs. powder).
With the aim of further highlighting the competitive performances of BC-TiO2, different TiO2-based systems used under similar testing conditions were also considered in the comparison of the photocatalytic performances. Even when compared with these less green systems, the photocatalytic activity exhibited by TiO2 nanotube arrays decorated with BC NPs proposed in this work is comparable to or even higher than that reported for other TiO2-based photocatalysts.  Table 1. It should be underlined that, at present, only few articles report on MB degradation over BC-based photocatalysts. As a matter of fact, we did not find any data on BC-based planar photocatalysts, and thus only materials in powder form were considered for this comparison. In this respect, our systems represent the first examples of BC-decorated photocatalytic layers for this kind of application. Both microalgae-and nutshell-derived materials proposed in this work provide MB abatement rates comparable to or even higher than those reported for other BC-based photocatalysts. This is particularly interesting considering the much lower amount of co-catalyst present on our planar photocatalyst (in the range of micrograms) compared to that on reported materials. In any case, such an evaluation should be carefully considered owing to the quite different experimental conditions and photocatalyst form (i.e., planar platform vs. powder). With the aim of further highlighting the competitive performances of BC-TiO 2 , different TiO 2 -based systems used under similar testing conditions were also considered in the comparison of the photocatalytic performances. Even when compared with these less green systems, the photocatalytic activity exhibited by TiO 2 nanotube arrays decorated with BC NPs proposed in this work is comparable to or even higher than that reported for other TiO 2 -based photocatalysts.

TiO 2 Nanotube Arrays Fabrication
Ti foils (99.945% purity, 0.03 mm thickness, William Gregor Ltd., London, UK) were degreased by sonicating in solvents of increasing polarity: acetone, isopropanol, ethanol, and ultrapure water. Each sonication step was conducted for 15 min and the samples were then dried under a N 2 stream. Subsequently, cleaned Ti foils (15 × 15 mm) were anodized at room temperature in an ethylene glycol solution containing ultrapure water (2% v/v, produced by a Sartorius Arium mini UV Lab Water System) and a fluoride concentration of 0.09 M (obtained by dissolving NH 4 F, Merck, Kenilworth, NJ, USA). A two-electrode configuration was used for the anodization, where the Ti foil was the working electrode, and a graphite plate was the counter electrode. The anodization was performed by applying a potential of 40 V (for 60 min) using a DC power supply (Wanptek NPS1230W). The anodized samples were rinsed and soaked in ethanol for at least 30 min. The as-formed (amorphous) TiO 2 nanotubes were then annealed and crystallized at 450 • C for 1 h in air with a heating ramp of 5 • C/min.

Biochar Production
Biochar was produced by slow pyrolysis at low temperature (350 • C) of two different commercially available feedstocks: nutshells and the microalgae Nannochloropsis sp. Biomasses were treated as detailed in previous work [46]. Briefly, approximately 10 g of dried raw biomass feedstock were put in a quartz tubular furnace and heated at 350 • C with a ramp of 10 • C/min under N 2 stream (0.5 L/min). Isothermal conditions at 350 • C were kept for 1 h. The obtained biochar was pulverized using a Fritsch Pulverisette 0 agate ball mill. BC produced from nutshells and microalgae is labelled as NBC and MBC, respectively.

Fabrication of the Composite BC-TiO 2 Photocatalyst
The fabrication of the composite BC-TiO 2 photocatalysts is summarized in Figure 1. After pyrolysis, the bulk BC powder was dispersed in ultrapure water (10 g/L) and sonicated for 1 h at room temperature using a Branson 5800 ultrasonic bath; this process induces the partial breaking of BC to nanosized particles [47,63]. The so obtained BC suspension was then centrifugated at 3000 rpm for 30 min. The supernatant was collected and centrifuged again to further disperse the smaller particles. Finally, the suspension obtained by the second centrifugation was filtered using a 0.2 µm pore size membrane; the collected liquid, containing nanosized BC, was deposited by drop-casting (120 µL) onto the TiO 2 support. Then, the sample was dried under a laminar flow hood at room temperature for at least 6 h. The loading of BC was varied by performing multiple depositions (1, 3, 5, and 7 times): photocatalysts are accordingly labelled as n-MBC-TiO 2 and n-NBC-TiO 2 , where "n" is the number of performed depositions.

Characterization of BC and BC-TiO 2
The complete characterization of the morphology and chemical composition of BC (as obtained by pyrolysis) is reported in detail in our previous work [46].
The morphology of TiO 2 NTs and BC-TiO 2 was studied using an XL30 Environmental Scanning Electron Microscope (ESEM FEG Philips) at 20kV in high vacuum condition. Transmission electron microscopy (TEM) and high-resolution TEM (HR-TEM) analysis were performed by a ZEISS LIBRA 200FE microscope. A small volume (20 µL) of the suspension containing BC NPs was deposited onto a TEM grid to observe the morphology of as-formed NPs, i.e., before the deposition on the TiO 2 nanotubes substrate. The particlesize distribution of BC nanoparticles was determined by processing SEM images using the software ImageJ (National Institutes of Health (NIH), Bethesda, MD, USA). X-ray diffraction (XRD) data were recorded on a Siemens D5000 system using Cu-Kα radiation, 40 kV, 40 mA, with a step of 0.03 • (2θ) and a scanning speed of 0.06 • ·s −1 . Diffuse reflectance UV-visible spectroscopy (DR-UV) measurement on solid samples and UV-vis analysis to determine the methylene blue concentration (in photocatalytic tests) were carried out using a Shimadzu UV-2600i spectrometer.

Photocatalytic Degradation of Methylene Blue
Methylene blue (MB) was selected as target molecule as this dye is widely studied and adopted as a pollutant model in the literature for the development of photocatalysts [27,43,64].
Photocatalytic tests under mild conditions were performed using a 250 mL quartz beaker placed on a magnetic stirrer to ensure the continuous mixing of the solution. Each experiment was performed using 100 mL of a 10 mg/L MB solution (obtained by dilution of a concentrated 1 g/L solution prepared by dissolving solid MB in ultrapure water) and an irradiation power of 40 mW/cm 2 .
Boosted photocatalytic tests were performed using 10 mL of a 10 mg/L MB solution stirred in a quartz test tube with an irradiation power of 100 mW/cm 2 .
An UV LED light source (λ = 365 nm, UVWave) with tunable power was used for both types of photocatalytic tests. Power density was measured by means of a thermal sensor (OptoSigma ® ).
MB concentration was monitored during the degradation tests by spectrophotometric means, recording absorbance values at 665 nm. A 30-min equilibration without light irradiation was performed before each photocatalytic test.

Conclusions
In this paper we demonstrated the feasibility of producing biochar-decorated TiO 2 photocatalysts by means of simple drop-casting. The composite BC-TiO 2 materials prepared in the present work proved to be more efficient than undecorated TiO 2 nanotubes towards the degradation of methylene blue under UV light irradiation; this evidence clearly indicates the ability of BC NPs to act as co-catalysts on the TiO 2 surface. The greatly enhanced sorption of organic molecules, together with improved charge carrier separation provided by BC NPs led to almost tripled photocatalytic performances.
It was observed that two types of biochar derived from markedly different biomasses (i.e., microalgae and nutshells) can promote MB decomposition to a similar extent, even if microalgae-derived BC is much less prone to deactivation than the nutshell-derived one. Further studies will be carried out to deeply understand the nature of the deactivation process and restore the original activity of the material.
Future perspectives might involve the optimization of reaction conditions to obtain the best photocatalytic performances and the application of biochar nanoparticle-decorated photocatalysts for the abatement of contaminants of emerging concern (e.g., pharmaceuticals such as acetaminophen).