Synthesis, characterization and photocatalytic activity of a composite based on TiO2 and active carbon prepared from mango pit

Instituto Federal de Educação, Ciência e Tecnologia Baiano, 48110-000, CatuBahia, Brazil Instituto Federal de Educação, Ciência e Tecnologia de Goiás, 76400-000 Uruaçu-Goiás, Brazil Instituto de Química/Laboratório de Fotoquímica e Ciências de Materiais, Universidade Federal de Uberlândia, 38400-462, Uberlândia-Minas Gerais, Brazil Departamento de Ciências Exatas e da Terra, Universidade do Estado da Bahia-Campus I, 41200-105, SalvadorBahia, Brazil


INTRODUCTION
Pollution of water resources is a socio-environmental problem of global importance in modern society. Eminently, dyes have a high potential for contamination. Crystal violet (Colour Index 42555), molecular formula C25H30ClN3, is widely used in the industry for dyeing paper, acrylic fibers and leather, identifying bacteria by the Gram staining method, treating burns and skin lesions, and identifying prints digital. It is a cationic dye that has harmful effects on life, causing irritation to the eyes and digestive tract and, in extreme cases, leading to respiratory and renal failure and carcinogenic potential. Furthermore, this dye can cause long-term effects in aquatic environments [1].
In general, the contamination of residual water by dyes, many of them recalcitrant to classic treatments, results in numerous environmental problems, compromising ecosystems [2]. The presence of colored species in water bodies, even in small concentrations, may present harmful effects to aquatic biota since it hinders the passage of light, compromising photosynthesis and oxygenation of water [1]. Considering the environmental impacts, together with the concerns about human health, the study of effluent treatment alternatives that promote the efficient removal or elimination of dyes becomes a priority [3].
In this context, we emphasize advanced oxidative processes (AOP) based on the action of free radicals and other active species on organic matter, leading to degradation or even complete mineralization [4,5]. A particular case of AOP is a heterogeneous photocatalysis employing primarily titanium dioxide (TiO2) as a catalyst, which has been shown to be effective in the degradation of dyes [6,7]. However, although thoroughly studied, TiO2 presents limitations to large scale applications, such as the need for radiation in the ultraviolet (UV) region for a band gap of 3.2 eV and the rapid recombination of the electron/hole pair beyond the difficulty of separation and regeneration after the process. Thus, it is necessary to develop alternative systems based on TiO2 with better photocatalytic properties in relation to commercial oxide, thus increasing the possibility of use in the treatment of effluents [8][9][10].
Several studies have been published focusing on the development of systems based in the combination of active carbon (or other carbonaceous materials) and semiconductor oxides in order to obtain more active catalysts for different applications [11][12][13]. Among the different solids, it is important to highlight those obtained from the association between semiconductor and activated carbon nanomaterials, which have advantages such as high specific surface area and the adsorption capacity of organic matter, which tends to favor the photocatalytic process, and the ease of regeneration of these materials [14][15][16][17].
According to the literature, active carbon can be obtained from agro-industrial waste such as mango [18,19], one of the most produced fruits in Brazil. Due to the nutritional and sensorial characteristics of the fruit, there is an expressive consumption of mango in the country, both in natura and in the form of pulp for juice, jellies, and sweets. The industrial processing of the fruit generates a large amount of bark and pits that are discarded as waste, generating a substantial environmental problem [20,21]. This residue, besides being used in the production of coal (adsorbents), presents a potential resource of raw material in the production of biofuel and as a nutritional by-product [20][21][22][23].
In the present study, mango pit was used for the preparation of coal, TiO2-carbon composites, being combined directly with titanium dioxide (TiO2-residue) in order to improve the characteristics of the semiconductor in photocatalytic processes.

Preparation of active carbon from mango pit
Mango pit (obtained from a fruit pulp mill located in the city of Salvador, Bahia, Brazil) were washed with water to remove impurities and fruit residues. Afterwards, they were sun-dried and then in an oven at 70 ºC for 24 hours. The residues were cut and the lumps were oven dried for another 24 hours. Afterwards, they were crushed in a knife mill with the powder obtained and separated in an 80 mesh sieve (sample B). This material was mixed with the activating agent ZnCl2 and submitted to a thermal treatment at 500 °C under nitrogen atmosphere, producing the activated carbon (AC). The active carbon was then washed with a solution of hydrochloric acid at 70 °C under stirring for 60 minutes and then washed with hot water until pH around 8.0. The material was again oven dried at 80 °C for 8 hours. The sample B, without the activating agent, was heat treated under the same carbonization conditions of the active carbon, producing the sample CB.

Modification in commercial titanium dioxide
Commercial TiO2 (P25 Degussa), 0.6 g, was dispersed in 29 mL of NaOH (10 mol L -1 ) solution under vigorous stirring. The obtained mixture was transferred to a teflon beaker, which was inserted into a steel reactor and heated in an oven at 130 °C for 3 hours. At room temperature, the resulting white powder was separated from the aqueous phase by centrifugation and washed with distilled water until a pH close to 8.0. After that, 60 mL of a 0.1 mol L -1 HCl solution was added to the powder and the resulting suspension was stirred for 2 hours. To promote ion exchange, the acid solution was renewed after 1 hour of stirring. The modified titanium dioxide (NT) was washed with distilled water and dried at 70 °C for 12 hours [24].

Synthesis of composites TC e TB
The active carbon (AC) prepared from mango pit was mixed with the modified titanium dioxide (NT) in the 1:1 mass ratio. The mixture was stirred for 12 hours and then suspended in deionized water, being subsequently centrifuged. After removing the liquid, the material was oven dried at 120 °C for 24 hours and reheated at 500 °C for 1 hour under nitrogen flow. The resulting composite was named TC. The same procedure was repeated using the sample B, thus obtaining the TB composite.

Infrared with Fourier Transform (FTIR)
In FTIR analyses, a Perkin Elmer FTIR Spectrum 100S model was used. The pellets were prepared by mixing the samples to be analyzed with potassium bromide (KBr). The spectra were obtained in the region between 4000 and 400 cm -1 with 4 cm -1 resolution.

X-ray diffraction (XRD)
The diffractograms were obtained using a Shimadzu XRD6000 equipment with Cu Kα radiation (λ=1.5418 Å). The experiments were conducted at 2θ intervals between 20 and 80° with a scanning speed of 2° min -1 .

Thermal analysis (TG/DTA)
TG/DTA curves were obtained using a Shimadzu TA-60 WS equipment at 25-900 °C min -1 under N2 atmosphere. Approximately 10 mg of sample per measurement were packed in a platinum crucible.

Nitrogen adsorption/desorption isotherms
Samples were pretreated for 2 hours at 250 °C under vacuum (6.6 x 10 4 Pa). The nitrogen adsorption / desorption isotherms at -196 °C were obtained using a Micrometrics ASAP 2020 equipment.

Scanning electron microscopy
The scanning electron microscopy (SEM) analysis was performed using a Tescan Model VEGA 3 LMU. The samples were placed on a carbon fiber covered stub, and then metallized in a Quorum Technologies Model Q150 RES equipment, where the coating was made by deposition of metallic ions of gold. The micrographs were obtained using an acceleration voltage of 10 kV and a secondary electron detector.

Transmission electron microscopy
The morphology and particle size distribution of the samples was characterized by, by transmission electron microscopy (TEM), using a model JEM1400 microscope.

Diffuse reflectance spectroscopy
The optical properties of the materials were determined by diffuse reflectance spectroscopy. The spectra were obtained at room temperature using a SHIMADZU UV-1201 spectrophotometer using potassium bromide as the reference.

Photocatalytic assays
The photocatalytic assays were performed using a reactor equipped with a cooling system which allows the solution to be kept under irradiation at a temperature of approximately 30 °C. A high pressure mercury vapor lamp (125 W) was used as source of irradiation ( Figure 1). Initially, aiming to evaluate the photodegradation kinetics of crystal violet (CV), dye used as model of oxidizable substrate in the absence of catalyst, its photolysis was carried out by adding only an aqueous solution of the dye (350 mL, 100 mg L -1 , pH 6.8) to the reactor.
In the photocatalytic assays, 350 mL of the CV solution (100 mg L -1 , pH 6.8) and 400 mg of the catalyst were added to the reactor. The temperature and pH of the reaction medium was the same as that the previously used. The suspension was stirred in the dark for approximately 15 minutes to establish the adsorption equilibrium between the solution containing the dye and the surface of the catalyst. Then, the mercury vapor lamp was turned on and aliquots were collected at predefined time intervals, centrifuged, being the reaction monitored at 589 nm with the aid of a Perkin Elmer UV-VIS spectrophotometer.
Photocatalytic degradation of the dye was also monitored using measurements of dissolved organic carbon (DOC) using a Shimadzu TOC-VCPH/CPN carbon analyzer equipped with an ASI-V autosampler. The samples were all filtered using 0.45 μm medium porosity filters prior the analysis.

RESULTS AND DISCUSSION
The infrared spectra of the samples of mango residue before (B) and after heat treatment (CB), and of the activated carbon (AC) are shown in Figure 2.
All samples present wide bands between 3000 and 3500 cm -1 related to stretching frequencies of OH groups at the surface of these materials [11]. In the infrared spectrum of the mango residue (B), absorption bands can be observed at 2922 and 2845 cm -1 . These bands are attributed to methylene (C-H) group vibrations [25]. It can also be observed an absorption band at 1720 cm -1 related to vibrational stretch of carbonyl groups [26], and bands between 1610 and 1500 cm -1 associated to C-C bond in aromatic rings, present in the lignin structure [26]. In addition, a band at 1119 cm -1 can be observed, attributed to C-O-C bonds. When analyzing the spectra of the samples of the thermally treated biomass and the active carbon (CB and AC), the disappearance of some bands is observed, such as for example those observed at 2922 cm -1 , 2845 cm -1 , 1720 cm -1 , and 1119 cm -1 in the biomass spectra. The observed changes in the spectra of these samples are a consequence of the decomposition of the precursors combined with the structural change that occurs during heat treatment [27]. Infrared spectra (Figure 3) of commercial titanium dioxide (TiO2 P25), modified titanium dioxide (NT), and composites (TC and TB) showed similar wide bands between 3000 and 3500 cm -1 . Bands around 1630 cm -1 were also observed, associated to hydroxyl group elongation, indicating the presence of water in the samples [28,29]. The solids also present a band around 2300 cm -1 and 1380 cm -1 , suggesting C-O and C=O stretching, respectively [30]. The absorption bands around 464 cm -1 and 700 cm -1 can be attributed to Ti-O bonding vibrations, characteristic of Ti-O-Ti [30,31]. The presence of these bands in TC and TB suggest that the titanium dioxide particles should be covering the surface of the carbonaceous materials in the composites [11].  The NT sample profile exhibits peaks in the same positions as those observed for TiO2 P25 suggesting that the phase composition in these solids was not modified. However, the peaks in NT are smaller and wider, indicating the formation of structures containing smaller crystals. In the profile of TC a decrease in intensity with wider peaks is also observed, more pronounced for rutile [32]. The diffractogram of TB show broader and less defined peaks, characteristic of the anatase phase. The absence of peaks characteristic of rutile indicates a different structure for the other materials, suggesting that the carbonization of the biomass generating reducing gases, such as CO and H2, provided more favorable conditions for stabilization of the anatase phase than for rutile. Another possibility for this is the coating of TiO2 by charcoal during the carbonization of the mango residue. The TG/DTA curves of the mango residue (B) and the titanium dioxide/mango residue mixture (TBBC, BC -before carbonization) are shown in Figures 5 and 6. Three regions of mass loss are observed in the TG curve of precursors B and TBBC, with the first one below 100 ºC (mass loss of approximately 20%), related to water and volatile losses, and the others between 200 and 500 ºC, related to the decomposition of biomass components (loss of mass of approximately 40%). However, the TG curve of sample B shows that there is a mass reverberation at temperatures above 500 o C, remaining to a temperature higher than 800 o C, while in the TG curve of the TBBC a temperature stability is observed above 500 o C. The DTA curve of the mango residue (B) shows an exothermic peak with ashoulder at about 340 °C and a maximum at 400 °C. Such thermal effects may correspond to the decomposition of lignocellulosic materials present in the biomass associated with changes in the structure [33]. In the DTA curve of TBBC only one peak is observed with a maximum at approximately 350 °C. This suggests that the decomposition of lignin, cellulose, and hemicellose present in the mango residue occurs at lower temperatures during the carbonization of B, indicating the influence of titanium dioxide in the process. The absence of other peaks at higher temperatures indicates that there was no phase transformation from anatase to rutile, which occur at temperatures of 600 -700 °C [34]. Thus, based on the TG-DTA information, it can be inferred that the carbonization of the precursor TBBC at a temperature of 500 °C should favours the anatase phase.
Analyzing the curves of the composite titanium dioxide/activated carbon before heat treatment (TCBC, BC -before carbonization) (Figure 7), a mass loss up to 100 °C is observed, decreasing between 100 and 200 °C, accompanied with an endothermic peak in the DTA curve, which may be associated to the removal of water from the material (mass loss of approximately 20%). Between 200 and 500 °C, the stability of the sample can be observed. From these results, it can be inferred that the TiO2 content in TC composite remains in 50% (content of the starting material) while in the composite TB the titanium dioxide content is about 62% due the biomass decomposition during the heat treatment temperature (500 o C). Between 500 and 700 °C, a mass variation can be observed in the TG curve, probably due the decomposition of carbon. This is accompanied with two exothermic peaks in the DTA curve at 500 °C, indicating a possible transformation of anatase in rutile, a process that occurs at temperatures higher than 600 °C [35]. From these results, it can be inferred that with thermal treatment of modified titanium dioxide/biomass precursors or modified titanium dioxide/activated carbon at 500 °C in an inert atmosphere, the anatase phase is stabilized with respect to rutile, corroborating with results of X-ray diffraction. The N2 adsorption/desorption isotherms of the composites follow the type IV, present hysteresis cycles at high pressures (P/Po) (Figure 8), suggesting the presence of mesoporous materials. In the isotherm of the TC sample the occurrence of adsorption is observed at low pressures, suggesting the existence of micropores, which can be attributed to the presence of carbon [36].

Figure 8. N2 adsorption/desorption isotherms obtained for the photocatalysts: NT is the modified TiO2 P25; TC is the active carbon with modified titanium dioxide (NT); TB is the precursor B with modified titanium dioxide (NT) and P25 is commercial TiO2.
The calculated specific surface area, mean pore diameter, pore volume and porosity are presented in Table 1. When analyzing these results, it was observed that the alkaline hydrothermal treatment produced a material with a high specific surface area (NT, 268 m 2 g -1 ) when compared to its precursor (TiO2), which may be related to the reduction in size of the crystals observed by XRD [37]. From the value of the mean pore diameter, it can be concluded that all materials have essentially mesoporous structure (2-50 nm). The TB composite showed higher porosity which may be a consequence of the thermal decomposition of the lignocellulosic materials present before heating, resulting in a porous carbonaceous material. The composite TC presented the largest specific surface area, suggesting the positive effect of the association of TiO2 with activated carbon. The pore size distributions were estimated using the method of Barrett-Joyner-Halenda (BJH) [38], shown in Figure 9. It can be seen a wide distribution of pore sizes in the mesoporous region with average values below 50 nm.  Figure 9: Distribution of pore volume of the materials under study, estimated using the BJH method [38]. NT is the modified TiO2 P25; TC is the active carbon with modified titanium dioxide (NT); TB is the B precursor with modified titanium dioxide (NT) and P25 is a commercial TiO2 (P25). Figure 10 shows the SEM images of the mango residue after heat treatment (CB), and active carbon (AC). Thermal treatment of the biomass at 500 °C led to a carbonaceous material with a characteristic morphology of aggregated particles with cavities. Activation using zinc chloride gave rise to the sample (AC), where an increase in the size of cavities was observed, in agreement with the literature [39].  Figure 11a shows that the NT sample consists of combined crystallites in larger clusters. A possible cause for the formation of agglomerates is the chemical etching and the type of heat treatment that TiO2 was submitted for formation of the NT sample. It can be inferred that the NT agglomerates can be distributed on the surface of the carbon and deposited in the opening of the pores of the active carbon [40], which can explain the micro-composition of TC (Figure 11b) and its smaller pore diameter when compared to the other materials. This agrees with X-ray diffraction data which show that there was no significant change in NT structure after heat treatment of the mixture between NT and active carbon, and with FTIR results that indicated a possible coverage of the carbon by TiO2 particles. The Figure 12 shows the photomicrographs of the composite obtained by association of modified TiO2 with the mango residue and submitted to carbonization at 500 °C (TB). It is possible to notice a structure with elongated morphology, evidencing the fibrous nature of the plant materials. It has been observed that titanium dioxide particles are able to partially cover the surface of the carbonaceous material while still maintaining pores, possibly justifying its adsorption in these materials [41]. TEM image of the TB composite evidence the formation of irregular spherical roughly dispersed nanoparticles in the carbonaceous material (Figure 13a). The estimated mean size from TEM image is 5.7 nm, according to the histogram (Figure 13b). This value is close to that estimated for crystallite size from X-ray diffractogram (Figure 4), which is 4.1 nm. As discussed previously, when compared to other materials obtained, TB presents significant differences in its structural properties ( Figure 4). Such differences should be associated with the carbonization process of the biomass/dioxide and titanium blend (TBBC). According to the literature, the decomposition of solids with morphological irregularities, as is the case of biomass derivatives, occurs through complex mechanisms and the formation of gaseous products is controlled by nucleation mechanisms involving the formation and growth of nuclei, without restrictions, or through the absorption and fusion thereof. However, in some situations, the diffusion process associated with the production of gases and volatiles may predominate [42]. Comparing the thermal analysis results (TG/DTA) of the precursor of the titanium dioxide/biomass (TBBC) composite ( Figure 6) and the pure biomass (B) (Figure 5), it is observed that in the presence of titanium dioxide the decomposition of lignin, cellulose, and hemicellose occurs at lower temperatures, suggesting a kinetic process. In this case, titanium dioxide would act as catalyst in the carbonization of the biomass, altering the intermediate state, which could be responsible for the differences of properties of the obtained composite (TB), when compared to the other materials.
The diffuse reflectance spectra of the studied materials, expressed in terms of the Kubelka-Munk F(R) function [43] are shown in Figure 14. As can be observed in Figure 14, all the materials with the exception of TiO2 display similar UV-vis spectra and present an increase in the radiation uptake capacity, mainly in the visibleregion of the electromagnetic spectrum. This is due to chemical modifications and the subsequent addition of active carbon material (TB and TC) that increases the absorption capability in the visible. As active carbon is active for UV-vis light and changes the transition of TiO2 [44], it is not possible to estimate the band gap energy of these composite materials.
An inspection in the spectra suggests that the absorption of radiation is only null when F(R) → 0. This suggests the existence of permitted states with energies below the TiO2 P25 band gap, estimated by the Kubelka-Munk function [43] as being 3.25 eV (381 nm) [45]. TC and TB stands out, presenting an expressive absorption of visible radiation, which extends to the infrared. These results are important because suggest the possibility of using these materials in photocatalytic processes using virtually the entire solar spectrum, one of the limitations for the use of pure titanium dioxide. However, as this property is not determinant in the photocatalytic performance, it is necessary to analyze a set of factors. Figure 15 show the photocatalytic activity of TiO2, NT, TB, and TC, evaluated against the discoloration of crystal violet as a function of the irradiation time. It can be observed that the association between TiO2 and the carbonaceous materials resulted in a synergism capable of increasing the photocatalytic activity of the oxide, especially using the composite TB, most probably due the higher production rate of hydroxyl radicals and other species active on the surface of these catalysts. The results shown in Figure 15 show that TB was the most effective photocatalyst in promoting the discoloration of aqueous solutions containing the crystal violet dye. A set of characteristics differentiates TB from the others: morphology with elongated structure (with tube appearance), higher porosity and expressive pore volume (Table 1) and an essentially mesoporous structure, also having a high specific area. In addition, the XRD results suggest that the carbonization process that resulted in this composite provided favourable conditions for the stabilization of the anatase phase, with smaller peaks, suggesting a greater dispersion and/or smaller size, inhibiting the rutile phase, present in both P25 and NT. It is well known that anatase is more photoactive than rutile [46]. However, this does not contribute to better performance since the anatase/rutile combination in P25 delays the recombination of the charge carriers formed at the moment of TiO2 excitation, which favours the photocatalytic action of this oxide [46]. On the other hand, the improved textural properties associated with the morphological properties should contribute to a more efficient diffusion of the reactive species involved in the photocatalytic process. This should compensate the recombination of the charge carriers ensuring a differentiated photocatalytic performance for this material, since such parameters are directly associated with the number of active sites.
Another aspect that deserves to be emphasized for these composites is the extension of the absorption of radiation to the visible and the near infrared. It is known that TiO2 P25 has a limited ability of radiation uptake limited to wavelengths shorter than 400 nm [46]. Although the TB composite exhibits a different absorption spectrum from the other, showing a decline in the photon uptake capability from 400 nm, its photon uptake extends to the near infrared. Despite the fact that TC has practically constant photon uptake capability in the same region covered by TB, TB showed a higher porosity, a preferential anatase phase, and TiO2 dispersion in the carbonaceous material matrix.
The differentiated performance presented by the composites under study, especially TB, was confirmed by analyzing the efficacy of dye mineralization through dissolved organic carbon (DOC) measurements. A mineralization of more than 60% was achieved with 30 minutes of reaction (DOCinitial = 73 mg L -1 ; DOCfinal = 44 mg L -1 ), using the catalysts NT and TC, while TB presented better performance in dye mineralization (99%) after 60 minutes of reaction (DOCinitial = 73 mg L -1 ; DOCfinal = 72 mg L -1 ). In other words, the photocatalytic activity of these materials was considerably improved when compared with the results achieved using untreated TiO2 P25.

CONCLUSION
The composites produced by combination of nanoparticles of a commercial titanium dioxide with active carbon or mango residue, resulted in materials with improved photocatalytic properties when compared to the pure catalyst. The composites, especially TB, presented better performances in the photocatalytic degradation of violet crystal, used as model of oxidizable substrate, confirmed by solution discoloration and DOC measurements. A mineralization of more than 60% of the dye was achieved with 30 minutes of reaction time, using the catalysts TB and TC. The TB composite presented better performance for dye mineralization (99%) after 60 minutes of reaction.

ACKNOWLEDGMENTS
The authors would like to thank the Applied Chemistry Postgraduate Program (PGQA) of Universidade do Estado da Bahia (UNEB) and Federal University of Uberlândia for the granted infrastructure.