This work presents the study of tris(2-butoxyethyl) phosphate advanced oxidation in TiO2-containing systems. Titania was immobilized on aluminum surfaces from recyclable materials and the results were compared with the suspension system. The initial concentration of photocatalyst and the oxidizing agent was optimized in a 23 experimental design and a kinetic study of the reactions was performed in the selected conditions. The experimental data were fitted to the pseudo-first-order model (rate constants estimated at 0.0129 ± 0.0009 and 0.0079 ± 0.0006 min−1 for the systems with TiO2 in suspension and immobilized, respectively). Artificial neural networks were also employed to model the experimental data and they presented correlation coefficients superior to 0.98 in all the training operations. After five cycles of degradation, the TiO2–aluminum meshes exhibited a very low decrease in photocatalytic activity (inferior to 2%). Acute phytotoxicity assays demonstrated that the byproducts of the oxidation of TBEP molecules are less toxic than the raw samples regarding lettuce seeds. For both TiO2 systems, COD decreased considerably as a consequence of the degradation. The immobilized TiO2 system achieved similar degradation rates when compared with the suspension system.

Organophosphate esters (OPEs) are widely employed in coating rubber, polymeric materials, textiles, metal products and packaging as flame retardants. OPEs can contaminate the environment by leaching, abrasion and volatilization mechanisms, turning into recalcitrant species (Greaves et al. 2016; Vykoukalová et al. 2017). Many of the chemicals from this class are known to cause adverse effects to water ecosystems and human health due to carcinogenic, mutagenic and neurotoxic properties. Among these compounds, it is worth highlighting tris(2-butoxyethyl) phosphate (TBEP), tributyl phosphate (TBP) and tris(2-chloroethyl) phosphate (TCEP), which present huge levels of toxicity even at very low concentrations (Greaves et al. 2016).

Physical, chemical and biological conventional methods have been demonstrated to be inefficient in the removal of OPEs in wastewater because these pollutants are highly stable, and therefore, require the search for alternative techniques for their treatment (Charamba et al. 2018). Advanced oxidation processes (AOPs) are considered an alternative to treating non-biodegradable organic chemicals, acting non-selectively, reaching high degradation rates and in most of the cases, forming less harmful by-products (Nascimento Júnior et al. 2018).

However, when mineralization is not achieved, a range of by-products may be formed and among them, more harmful species than the parent chemicals. For this reason, biotoxicity assays using living organisms are suggested in order to evaluate comparatively the acute toxicity of treated AOP samples regarding the raw samples (Utzig et al. 2019).

AOP methods target the complete or partial mineralization of organic compounds from the generation of free radicals with high oxidant potential (Nasirian et al. 2017). In heterogeneous photocatalysis, titanium dioxide (TiO2) is the most employed photocatalyst due to its high performance, low toxicity and low cost (Barbosa et al. 2019). As a semiconductor, titania is excited by a radiation equal or of superior energy to its band-gap, generating an electron–hole pair able to oxidize water molecules and generate free radical species with high oxidation potential. Water and oxygen molecules can be adsorbed onto the surface of titania, interact chemically with the semiconductor and free radicals may be formed due to redox reactions (Wiedmer et al. 2016). The rutile phase of titania (3,2 eV) is able to absorb energy and form a hole in the UV electromagnetic spectrum from 200 to 380 nm (Bayarri et al. 2007).

Many reports in the literature dealing with TiO2 in AOP processes operate directly with it in suspension since it is insoluble in water and some of them have achieved great results working with the powder immobilized on the surface of a range of materials, such as glass (Xing et al. 2018), ceramic (Veréb et al. 2014), chitosan (Huang et al. 2014) and polymers (Santos et al. 2018). These low-cost materials as supports make the catalytic systems more environmentally friendly and aim to reduce further costs of separation (Aquino et al. 2019). Aluminum is a metal that encompasses these characteristics, in addition to presenting suitable mechanical properties (Verran et al. 2005) to be a photocatalytic carrier and being a worldwide recyclable good.

One of the most promising techniques in modelling kinetic data is based on the use of artificial neural networks (ANNs). ANNs are able to predict the behavior of chemical processes even with no knowledge of the chemical routes (Barbosa et al. 2019). Nascimento Júnior et al. (2018) and Aquino et al. (2019) have implemented this method in predicting organic pollutant removal in AOP systems. The ANN learns the behavior trends of a particular process, including the chemical ones, and it does not require the comprehension of the chemical routes, which makes it feasible to evaluate complex reactions such as AOPs.

The present study deals with heterogeneous photocatalysis for the removal of TBEP flame retardant in suspended TiO2 and immobilized over aluminium meshes manufactured from can seals. The novelty of the work relies on the usage of low-cost and mechanically resistant aluminum material, besides it being a product largely employed in recycling processes. Sample analyses were carried out by high-performance liquid chromatography (HPLC) and degradation kinetics was evaluated according to the pseudo-first-order (PFO) model and ANN method.

TBEP (C18H39O7P, 94%, purchased from Sigma Aldrich) samples were prepared at 1.0 mg/L in deionized water. Titanium dioxide (TiO2-P25: 80% anatase, 20% rutile) employed in this study was acquired from Degussa Evonik Industries and hydrogen peroxide (30% v/v) was acquired from Solvay Peróxidos Brasil.

Manufacturing of aluminum meshes and TiO2 immobilization

Aluminium meshes were manufactured from can rings obtained from commercial drinks. The meshes (8.0 × 4.5 cm, 1 mm depth) were calcined at 490 °C for 26 hours before the immobilization procedure. The immobilization of photocatalyst on aluminum surface was carried out as described by Barbosa et al. (2019). After immobilization of titania, the meshes were calcined once more for 2 hours at 490 °C in order to promote the adhesion of the photocatalyst particles.

Aluminum mesh characterization

The TiO2-mesh was characterized by X-ray diffraction (XRD) and scanning electron microscopy (SEM). XRD used radiation of Cu Kα, 2 kVA power, 30 kV voltage and 30 mA current (XRD-600 Shimadzu diffractometer). The scanning was performed at 2θ between 20° and 80° at a pace of 0.02° and speed of 2°.s1. Crystallographic letters were used to evaluate the characteristic peaks exhibited by the material (21-1272 and 21-1276 from International Center of Diffraction Data) and the data were compared with the raw aluminum material data. Morphology analyses of TiO2-mesh were performed with samples before and after the thermal treatment (calcination) and immobilization procedure in a Tescan MIRA3 Microscope. Images at a magnification of 500× were accessed.

Analytical methods (HPLC-MS/MS)

TBEP samples were analysed by HPLC (Acquity, Waters, USA) coupled with a single quadrupole mass spectrometer (SQ Detector 2, Waters, EUA) equipped with an electro-sprayer for ionization. The ionizing source operated at 3.5 kV of capillary voltage, 25 V cone voltage, and dewatering temperature of 150 °C and gas flow at a rate of 550 L·h−1. The TBEP raw sample (1.0 mg·L−1) was quantified according to its molecular weight (398.48 g·mol−1) by selected ion recording (SIR) mode.

Preliminary tests

The TBEP synthetic samples were treated in UVC, UVC/H2O2, UVC/TiO2(suspended), UV/TiO2(immobilized), UVC/H2O2/TiO2(suspended) and UVC/H2O2/TiO2(immobilized) systems. The treatment systems were also performed in the absence of radiation. The assays were carried out in cylindrical glass vessels (7.0 cm diameter and 5.5 cm height) inside a photoreactor in batch mode. The inner chamber of the reactor was coated with reflective material and the reactor contained a germicidal lamp (20 W, Tovalight T8, λ = 253.7 nm). The samples underwent magnetic stirring during the operation and TiO2-meshes were submerged in the reaction medium during the experiments.

Samples of 300 mL of TBEP (1.0 mg·L−1) were treated in each system for 240 minutes. Hydrogen peroxide was added at a concentration of 5.47 mmol·L−1 according to the stoichiometric balance of the degradation reaction between the pollutant and the oxidizing agent. The titania concentration at 500 mg·L−1 was based on the Santos et al. (2018) experiments. TBEP samples were brought into contact with TiO2 under stirring in the preliminary systems for 30 minutes in order to reach adsorption equilibrium in the absence of the radiation source (Nasirian et al. 2017).

Experimental planning design

A 23 experimental design was performed to investigate the operational conditions in the selected UVC-based system from the preliminary tests. Concentration of TiO2 (333.3, 500 and 666.7 mg·L−1), concentration of H2O2 (2.73, 5.47 and 8.20 mmol·L−1) and time (120, 180 and 240 minutes) were investigated. The central point was carried out in triplicate to determine the pure error of the experiment. Pareto charts were used to investigate the statistical significance of the parameters and their interaction in Statistica Experimental Design 6.0.

Kinetic study of TBEP degradation

The operational conditions were selected according to the experimental planning assays by the highest removal rates achieved. The experiments were carried out with TiO2 in suspension and immobilized on aluminium meshes with eight sampling times for 360 minutes. The kinetic profiles were fitted to non-linear pseudo-first-order models in Origin Pro 9.0. Rate constants (k) and half-life reaction times (t1/2) were estimated for both profiles.

ANNs were also employed to fit the experimental data provided by the UVC/H2O2/TiO2(suspension) and UVC/H2O2/TiO2(immobilized) systems. The software was developed in C# language in Unity 3D© according to Barbosa et al. (2019) and Göb et al. (1999). The software was able to receive input data, set a network type, run the training operation and display the resultant network in several forms. The network structure was 1:2:1, with time as the input variable, two neurons and concentration at time t as the output variable (response). The neural network was of type feed-forward with a sigmoidal function and all the data were normalized between −1 and 1 in the input in order to improve quality and speed of the networks. The performance of the network in describing a set of data could be evaluated by the mean absolute error and the correlation coefficient (R2). Training times were estimated at around 2,600 seconds on a desktop computer.

Reuse of aluminum meshes

The reuse of TiO2-meshes was evaluated in five consecutive cycles for 360 minutes in order to estimate the photocatalytic activity loss through the UVC/H2O2/TiO2(immobilized) system. Samples of 1.0 mg·L−1 of TBEP were applied in the degradation test cycles. The TiO2-meshes were washed with distilled water and dried at 35 °C for 1 h at the end of each cycle. The temperature conditions were mild to not cause any change in the mesh structure and, consequently, alter the degradation efficiency.

Toxicity bioassays and water quality parameters

Phytotoxicity bioassays in Lactuca sativa (lettuce) seeds were adapted from Aquino et al. (2019) for raw TBEP samples and treated samples in the UVC/H2O2/TiO2(suspension) and UVC/H2O2/TiO2(immobilized) systems. Inhibition concentration of root growth for 50% of the organisms (IC50) was calculated from the root elongation measurements. The relative growth index (RGI) was also estimated (Equation (1)):
formula
(1)
where ARS represents the average root elongation in treated samples and ARC is the average root elongation in the negative control (deionized water).

Chemical oxygen demand (COD), pH and conductivity of raw and treated samples in the UVC/H2O2/TiO2(suspension) and UVC/H2O2/TiO2(immobilized) systems were investigated. The methods were carried out as recommended by Standard Methods for the Examination of Water and Wastewater (APHA 2012).

Aluminum mesh characterization

Structural modifications occurred in the surface of the aluminum-can rings after the calcination step and TiO2 immobilization (Figure 1).

Figure 1

SEM images of (a) the raw aluminum mesh, (b) calcined aluminum mesh and (c) TiO2-immobilized aluminum mesh (500×).

Figure 1

SEM images of (a) the raw aluminum mesh, (b) calcined aluminum mesh and (c) TiO2-immobilized aluminum mesh (500×).

Close modal

According to Figure 1(a) and 1(b), it can be seen that the thermal treatment caused an increase in surface roughness, as expected, and in Figure 1(c), a smoother surface is observed when compared with Figure 1(b), indicating titania partially filled the pores of the treated aluminum material.

Figure 2 shows the XRD diffractogram of titania over the aluminum mesh. The qualitative analysis of the crystallographic composition after TiO2 immobilization identified aluminum presence by the peaks at 43.2°, 44.8° and 78.0° in the support, which is in accordance with the results presented by Ayieko et al. (2015).

Figure 2

X-ray diffractogram of (a) titania immobilized over aluminum seals and (b) titania P25 powder (raw photocatalyst). (A – anatase; R – rutile; Al – aluminum.)

Figure 2

X-ray diffractogram of (a) titania immobilized over aluminum seals and (b) titania P25 powder (raw photocatalyst). (A – anatase; R – rutile; Al – aluminum.)

Close modal

As can be seen in Figure 2(a), the peaks at 2θ = 25.5° 37.2°, 38.6°, 48.3°, 55.0°, 62.6° and 75.1° represent the anatase phase of TiO2 while the rutile phase is identified by the peaks at 2θ = 27.9°, 41.2°, 54.5°, 65.2° and 69.5° for titania crystallographic planes. The content of each crystalline phase of titania was also measured in the aluminum meshes coated with TiO2-P25 (80% anatase and 20% rutile). The calcined material exhibited a decrease in anatase phase content to 70% after immobilization, which may be associated with the thermal treatment. In Figure 2(b), some of the anatase peaks are observed at 53.8°, 68.7° and 70.2° and they are no longer observed in the calcined titania, some peaks even being replaced by rutile ones. Nadarajan et al. (2015) obtained similar results when they evaluated the crystallographic properties of TiO2 at 500 °C and achieved 70:30 proportion of the anatase and rutile phases, respectively.

Thermal treatment is believed to promote/cause an increase in rutile phase content regarding the commercial TiO2-P25. The heat treatment is responsible for enhancing roughness, pointed out to be related to amorphous aluminum oxide formation, which enhances the surface area and eases the adherence of powder in the pores (Figure 1(a) and 1(b)). After the immobilization (Figure 1(c)), smaller particles are observed on the mesh's surface in relation to Figure 1(b), indicating that the TiO2 crystals (Nadarajan et al. 2015) are appropriately immobilized on the metallic surface.

Preliminary tests

No considerable degradation rates were achieved in any of the systems run in the absence of UVC radiation. Since favourable degradation rates were achieved in the UVC systems, the breaking of organic molecules is linked to photon absorption rate by the oxidizing agents, promoting the generation of free radicals and electron–hole pairs (Velo-Gala et al. 2017).

The UVC system presented the lowest degradation rate achieving only 15.9% after 240 minutes. UVC/TiO2(suspended) and UVC/TiO2(immobilized) systems exhibited 58.3% and 49.2% degradation rates, respectively, while UVC/H2O2 exhibited 61.0% for the same photocatalytic reaction time. The most efficient systems were UVC/H2O2/TiO2(suspended) and UVC/H2O2/TiO2(immobilized), which reached 85.0% and 71.6% degradation, respectively.

UV radiation is able to break organic bonds in complex structures, although, effective mineralization of the pollutants is normally not possible to achieve, even after long operation times. The combined effect of UV and H2O2 leads to the formation of highly oxidizing radicals that transform large molecules in smaller structures non-selectively (Raizada et al. 2017). Hydrogen peroxide is homolitically broken at λ = 200–280 nm, which makes feasible the application of UVC radiation to the generation of hydroxyl radicals in situ (Jia et al. 2018). TiO2 is photoactivated generating hydroxyl (•OH), superoxide (•O2) and hydroperoxide (•OOH) radicals by oxidizing water molecules and being restored at the end of the operation (Wiedmer et al. 2016).

A synergistic effect is often observed when H2O2 and TiO2 are employed in photoabsorption and generation of radicals by enhancing the global rate of degradation. H2O2 presence is also pointed out to suppress the recombination of electron–holes in the photocatalytic activity of titania, thus, favouring the removal of pollutants from water (Borges et al. 2016). Other works have reported similar results (Nascimento Júnior et al. 2018; Barbosa et al. 2019). In addition to the mass transfer limitations associated with the immobilized system, the decrease of anatase content, as observed in the XRD analysis, may also be responsible for the lower efficiency of immobilized systems when compared with the usage of titania in suspension.

Experimental planning design

The degradation results varied from 67.18% (in the experiment with the three negative levels) to 96.88% (with higher levels of time and titania concentration), although maximum degradation rates (100%) were attained only at the assay in which the three positive levels were employed ([TiO2] = 666.7 mg·L−1; [H2O2] = 8.20 mmol·L−1; time = 240 min). The central point was carried out in triplicate and exhibited around 85% degradation rates and the experimental pure error was estimated at 0.071, corroborating the reproducibility of the experiments.

The influence of the parameters and their interaction was investigated statistically. The Pareto chart for the statistical significance of the independent variables is exhibited in Figure 3.

Figure 3

Pareto chart for the 23 experimental design in UVC/H2O2/TiO2(suspension).

Figure 3

Pareto chart for the 23 experimental design in UVC/H2O2/TiO2(suspension).

Close modal

According to the chart, all the independent variables and their interactions exhibited statistical significance. Santos et al. (2018) and Barbosa et al. (2019) obtained similar results in the degradation of emergent contaminants in heterogeneous photocatalytic systems. The individual independent variables and the interaction among the three proved to influence the degradation rate positively. Thus, the range of time and H2O2 and TiO2 concentrations used to perform this study did not reach an excess and are directly proportional to the degradation rates due to a higher content of free radicals (Nascimento Júnior et al. 2018).

Kinetic study of TBEP degradation

From the selected conditions in the experimental design (8.20 mmol·L−1 of H2O2 and 666.7 mg·L−1 of TiO2), concentration profiles were built for the degradation of TBEP in UVC/H2O2/TiO2(suspension) and UVC/H2O2/TiO2(immobilized) (Figure 4).

Figure 4

Concentration profiles of TBEP degradation in UVC/H2O2/TiO2(suspension) and UVC/H2O2/TiO2(immobilized). [TiO2] = 666.7 mg·L−1, [H2O2] = 8.20 mmol·L−1.

Figure 4

Concentration profiles of TBEP degradation in UVC/H2O2/TiO2(suspension) and UVC/H2O2/TiO2(immobilized). [TiO2] = 666.7 mg·L−1, [H2O2] = 8.20 mmol·L−1.

Close modal

As can be seen in Figure 4, the UVC/H2O2/TiO2(suspension) and UVC/H2O2/TiO2(immobilized) systems reached removal rates of 97.45% and 71.26%, respectively, after 180 minutes. Complete degradation of TBEP was attained in the TiO2-suspended system in 240 minutes and in the TiO2-immobilized system in 360 minutes. The pseudo-first-order kinetic model is pointed out to fit the degradation data of organophosphate esters by AOP processes in other works (Antonopoulou et al. 2016; Sharma et al. 2016) reporting the degradation of very diluted samples of the flame retardant.

The kinetic rate constants were estimated at k = 0.0129 ± 0.0009 min−1 (R2 = 0.9716) and 0.0079 ± 0.0006 min−1 (R2 = 0.9721) for the UVC/H2O2/TiO2(suspension) and UVC/H2O2/TiO2(immobilized) systems, respectively, by the pseudo-first-order model. As a consequence, the half-life reaction time was increased from 53.73 to 87.74 minutes from the TiO2-suspended to the TiO2-immobilized system. Antonopolou et al. (2016) found similar results (k = 0.0141 min−1) in the degradation of tris(1-chloro-2-propyl) phosphate in TiO2 heterogeneous photocatalytic systems.

In the estimated residual distributions for the UVC/H2O2/TiO2(suspension) and UVC/H2O2/TiO2(immobilized) results, variances of 0.0134 and 0.0177 were assessed, correspondingly. The low values of variance indicate a good fit of the model to the experimental data along with the R2 results, similarly to the work reported by Barros Neto et al. (2010).

The TiO2-suspended systems are, in general, able to achieve higher removal rates since they provide higher contact area between the catalyst, the radiation, and the water and pollutant molecules. In TiO2-immobilized systems, there is a reduction in the photon harvesting rates due to the lower surface area of titania and mass-transfer limitations regarding the TiO2 suspension samples (Rostami-Vartooni et al. 2016). The same effect was observed by Aquino et al. (2019) when the efficiency of titania was compared in suspension samples and immobilized over polyethylene terephthalate plates. However, the TiO2-immobilized systems reduce costs in additional separation steps, which represent an economic advantage. Moreover, the use of a recyclable and low-cost material remodels the process to be more environmentally friendly.

The training of the ANN from the experimental data obtained is exhibited in Figure 5 for both the UVC/H2O2/TiO2(suspension) and UVC/H2O2/TiO2(immobilized) systems.

Figure 5

ANNs applied to the degradation data of TBEP in (a) UVC/H2O2/TiO2(suspension), (b) UVC/H2O2/TiO2(immobilized) systems and (c) fitting to the experimental data of both systems (1: UVC/H2O2/TiO2(suspension) and 2: UVC/H2O2/TiO2(immobilized)).

Figure 5

ANNs applied to the degradation data of TBEP in (a) UVC/H2O2/TiO2(suspension), (b) UVC/H2O2/TiO2(immobilized) systems and (c) fitting to the experimental data of both systems (1: UVC/H2O2/TiO2(suspension) and 2: UVC/H2O2/TiO2(immobilized)).

Close modal

From Figure 5(a) and 5(b), the input layers (normalized from −1 to 1) can be seen that represent degradation times and are linked to the output ones, where concentration is calculated at time t (also normalized from −1 to 1). The network multiplies the input values by their respective weights (numbered on the lines that are linked to the neurons), summed to the bias (numbered on the neurons), computed by the sigmoidal activation function between −1 and 1 and finally multiplied by the weights to provide a numeric response.

The pseudocode for the UVC/H2O2/TiO2(suspension) is displayed below:

//Input normalization

Input = 2*(Input)/360 − 1;

//Computation

Output = 2.283514*(2/(1 + exp(−1*(2.252896*(2/(1 + exp(−1*(−1.872869*Input − 2.2727))) − 1) − 1.136663*(2/(1 + exp(−1*(0.2178829*Input − 1.568911))) − 1) + 0.5732782))) − 1);

//Output denormalization

Output = 0.499*Output + 0.501;

For the UVC/H2O2/TiO2(immobilized) system the pseudocode is as follows:

//Input normalization

Input = 2*Input/360 − 1;

//Computation

Output = 2.848087*(2/(1 + exp(−1*(0.3007855*(2/(1 + exp(−1*(0.6244544*Input − 1.322824))) − 1) − 1.500366* (2/(1 + exp(−1*(2.905448*Input + 2.782597))) − 1) + 0.8577532))) − 1);

//Output denormalization

Output = 0.5*Output + 0.5;

From Figure 5(c), it is possible to conclude that the ANN models provided a satisfactory fitting to the experiments. The average error was estimated at 0.0199 for the TiO2-suspended system and 0.0138 for the TiO2-immobilized system and R2 values were estimated at 0.988 and 0.994 for both systems, respectively. Both correlation coefficients are presented to be superior than that obtained by the fitting to the pseudo-first-order kinetic model. Thus, the ANN model is able to predict the degradation rates at a range of conditions with lower deviation than the previous model.

Reuse of aluminum meshes

To evaluate the reuse capacity of the TiO2–aluminum meshes, cycles of degradation were performed in the same conditions as the kinetic study. In the first three cycles, 100% of TBEP was attained in 360 minutes indicating no significant loss of photocatalytic activity. In the fourth and fifth cycles, TBEP degradation rates reached 99.8% and 98.7% respectively, demonstrating a slight decrease in the activity of the meshes. These results may be associated with the adsorption of byproducts of the degradation on the surface of the titania, occupying active sites and hindering photon absorption (Bel Hadjltaief et al. 2016). However, the use of TiO2–aluminum meshes presents itself as a feasible alternative for reuse in scale-up application since the decrease in photocatalytic activity is minor after five cycles.

Toxicity bioassays and water quality parameters

From the phytotoxicity assay, IC50 was estimated for raw and treated TBP samples in the UVC/H2O2/TiO2(suspension) and UVC/H2O2/TiO2(immobilized) systems. The germination and root growth rates, RGI and IC50 are exhibited in Table 1. The root growth and RGI were calculated based on TBEP samples at a concentration of 1.0 mg·L−1.

Table 1

Phytotoxicity parameters for the raw and treated TBEP samples by AOP systems with Lactuca sativa seeds

SampleGermination rate (%)Average root growth (cm)RGIIC50 (mg·L1)
Raw TBEPa 90 1.6 ± 0.3 0.69 0.8 ± 0.1 
UVC/H2O2/TiO2(suspension) 97 2.1 ± 0.2 0.91 1.2 ± 0.3 
UVC/H2O2/TiO2(immobilized) 95 1.9 ± 0.1 0.82 1.0 ± 0.3 
Negative controlb 100 2.3 ± 0.3 – – 
Positive controlc – – – 
SampleGermination rate (%)Average root growth (cm)RGIIC50 (mg·L1)
Raw TBEPa 90 1.6 ± 0.3 0.69 0.8 ± 0.1 
UVC/H2O2/TiO2(suspension) 97 2.1 ± 0.2 0.91 1.2 ± 0.3 
UVC/H2O2/TiO2(immobilized) 95 1.9 ± 0.1 0.82 1.0 ± 0.3 
Negative controlb 100 2.3 ± 0.3 – – 
Positive controlc – – – 

aTBEP (1.0 mg·L−1).

bDistilled water.

cHg2SO4 (0.5 mmol·L−1).

Since germination rates were higher than 90% in the negative control, the phytotoxicity assay was considered valid (Utzig et al. 2019). From Table 1, it is possible to conclude that both photocatalytic treatments led to lower acute phytotoxicity levels of the TBEP samples since RGI was considerably magnified. According to the RGI results, lettuce radicles did not demonstrate growth inhibition by the treatment (RGI values > 0.8) while the raw sample demonstrated the existence of inhibitory effects (RGI < 0.8). The IC50 index corroborates that byproducts of TBEP photocatalytic removal were not more toxic than the parent compound. A slight increase in its value for both treatment systems indicates the inhibition concentration to 50% of the organisms was higher after treatment. The positive control did not present a considerable germination rate, as expected, and a growth was noticed of 0.3 cm for the UVC/H2O2/TiO2(immobilized) and 0.5 cm for the UVC/H2O2/TiO2(suspension) system.

Table 2 displays the water quality parameters determined for raw and treated samples after 240 minutes of operation in the UVC/H2O2/TiO2(suspension) and UVC/H2O2/TiO2(immobilized) systems.

Table 2

Water quality parameters for the raw and treated TBEP samples by AOP systems

SampleCOD (mg·L1)pHConductivity (μS·cm−1)
Raw TBEP 28.07 7.11 3.73 
UVC/H2O2/TiO2(suspension) 5.52 6.62 7.56 
UVC/H2O2/TiO2(immobilized) 9.64 6.11 6.89 
SampleCOD (mg·L1)pHConductivity (μS·cm−1)
Raw TBEP 28.07 7.11 3.73 
UVC/H2O2/TiO2(suspension) 5.52 6.62 7.56 
UVC/H2O2/TiO2(immobilized) 9.64 6.11 6.89 

COD values decreased around 80.3% and 65.6% in the AOP systems. The slight decrease in pH after the treatments may be associated with the formation of carbonic acid, often reported in AOP operation (Rayaroth et al. 2015). Conductivity is also associated with the byproducts of the degradation since inorganic ions are expected to be formed (Aquino et al. 2019). In the case of TBEP mineralization, phosphate anions are expected to remain in the samples after the treatment.

In conclusion, high rates of degradation were achieved in UVC/H2O2/TiO2 systems employing either suspended or immobilized titania powder for 360 minutes of operation (100% and 98.7%, respectively), even after five cycles for the immobilized titania treatment. From pseudo-first-order kinetic fitting, 0.0129 and 0.0079 min−1 were achieved as rate constants for titania in suspension and TiO2–aluminum mesh AOP systems, respectively. ANN was demonstrated to be a feasible method to describe the experimental data of TBEP degradation. Acute toxicity of the degradation byproducts was demonstrated to be lower for treated samples when compared with raw TBEP solution and along with water quality analyses made TiO2–aluminum meshes a suitable alternative in the treatment of emergent contaminants to minimize operation costs, besides being environmentally friendly.

Antonopoulou
M.
Karagianni
P.
Konstantinou
I. K.
2016
Kinetic and mechanistic study of photocatalytic degradation of flame retardant tris (1-chloro-2-propyl) phosphate (TCPP)
.
Applied Catalysis B, Environmental
192
,
152
160
.
10.1016/j.apcatb.2016.03.039
.
APHA
2012
Standard Methods for the Examination of Water and Wastewater
, 22nd edn.
American Public Health Association, American Water Works Association, Water Environment Federation
,
Washington, DC, USA
.
Aquino
R. V. S.
Barbosa
A. A.
Ribeiro
L. B.
Oliveira
A. F. B.
Silva
J. P.
Azoubel
P. M.
Rocha
O. R. S.
2019
Degradation of leaf green food dye by heterogeneous photocatalysis with TiO2 over a polyethylene terephthalate plate
.
Chemical Papers
73
(
10
),
2501
2512
.
https://doi.org/10.1007/s11696-019-00804-y
.
Ayieko
C. O.
Musembi
R. J.
Ogacho
A. A.
Aduda
B. O.
Muthoka
B. M.
Jain
P. K.
2015
Controlled texturing of aluminum sheet for solar energy applications
.
Advances in Materials Physics and Chemistry
5
,
458
466
.
http://dx.doi.org/10.4236/ampc.2015.511046
.
Barbosa
A. A.
de Aquino
R. V. S.
Oliveira
A. F. B.
Dantas
R. F.
Silva
J. P.
Duarte
M. M. M. B.
Rocha
O. R. S.
2019
Development of a new photocatalytic reactor built from recyclable material for the treatment of textile industry effluents
.
Desalination and Water Treatment
151
,
82
92
.
https://doi.org/10.5004/dwt.2019.23905
.
Barros Neto
B.
Scarminio
I. S.
Bruns
R. E.
2010
Como Fazer Experimentos: Pesquisa e Desenvolvimento na Ciência e na Indústria (How to Do Experiments: Research and Development in Science and Industry)
, 4th edn.
Editora da Unicamp
,
Campinas
,
Brazil
.
Bayarri
B.
Abellán
M. N.
Giménez
J.
Esplugas
S.
2007
Study of the wavelength effect in the photolysis and heterogeneous photocatalysis
.
Catalysis Today
129
,
231
239
.
https://doi.org/10.1016/j.cattod.2007.08.006
.
Bel Hadjltaief
H.
Ben Zina
M.
Galvez
M. E.
Da Costa
P.
2016
Photocatalytic degradation of methyl green dye in aqueous solution over natural clay-supported ZnO–TiO2 catalysts
.
Journal of Photochemistry and Photobiology A: Chemistry
315
,
25
33
.
http://dx.doi.org/10.1016/j.jphotochem.2015.09.008
.
Borges
S. S.
Xavier
L. P. S.
Silva
A. C.
Aquino
S. F.
2016
Immobilized titanium dioxide (TiO2) in different support materials to use in heterogeneous photocatalysis
.
Química Nova
39
,
836
844
.
Charamba
L. V. C.
Santana
R. M. R.
Nascimento
G. E.
Charamba
B. V. C.
Moura
M. C.
Coelho
L. C. B. B.
Oliveira
J. G. C.
Duarte
M. M. M. B.
Napoleão
D. C.
2018
Application of the advanced oxidative process on the degradation of the green leaf and purple açaí food dyes with kinetic monitoring and artificial neural network modelling
.
Water Science and Technology
78
(
5
),
1094
1103
.
Göb
S.
Oliveros
E.
Bossmann
S. H.
Braun
A. M.
Guardani
R.
Nascimento
C. A. O.
1999
Modeling the kinetics of a photochemical water treatment process by means of artificial neural networks
.
Chemical Engineering and Processing: Process Intensification
38
,
373
382
.
https://doi.org/10.1016/S0255-2701(99)00028-8
.
Greaves
A. K.
Letcher
R. J.
Chen
D.
McGoldrick
D. J.
Gauthier
L. T.
Backus
S. M.
2016
Retrospective analysis of organophosphate flame retardants in herring gull eggs and relation to the aquatic food web in the Laurentian Great Lakes of North America
.
Environmental Research
150
,
255
263
.
http://dx.doi.org/10.1016/j.envres.2016.06.006
.
Huang
K.
Grumezescu
A. M.
Chang
C.
Yang
C.
Wang
C.
2014
Immobilization and stabilization of TiO2 nanoparticles in alkaline-solidificated chitosan spheres without cross-linking agent
.
International Journal of Latest Research in Science and Technology
3
(
2
),
174
178
.
Jia
P.
Zhou
Y.
Zhang
X.
Zhang
Y.
Dai
R.
2018
Cyanobacterium removal and control of algal organic matter (AOM) release by UV/H2O2 pre-oxidation enhanced Fe(II) coagulation
.
Water Research
131
,
122
130
.
https://doi.org/10.1016/j.watres.2017.12.020
.
Nadarajan
R.
Bakar
W. A. W. A.
Ali
R.
2015
Effect of calcination temperature on metal oxides and their photocatalytic activity
.
Advanced Materials Research
1107
,
73
78
.
Nascimento Júnior
W. J. D.
da Rocha
O. R. S.
Dantas
R. F.
da Silva
J. P.
Barbosa
A. A.
2018
Kinetic study of food dyes removal from aqueous solutions by solar heterogeneous photocatalysis with artificial neural networks and phytotoxicity assessment
.
Desalination and Water Treatment
104
,
304
314
.
Nasirian
M.
Bustillo-Lecompte
C. F.
Mehrvar
M.
2017
Photocatalytic efficiency of Fe2O3/TiO2 for the degradation of typical dyes in textile industries: effects of calcination temperature and UV-assisted thermal synthesis
.
Journal of Environmental Management
196
,
487
498
.
http://dx.doi.org/10.1016/j.jenvman.2017.03.030
.
Raizada
P.
Shandilya
P.
Singh
P.
Thakur
P.
2017
Solar light-facilitated oxytetracycline removal from the aqueous phase utilizing a H2O2/ZnWO4/CaO catalytic system
.
Journal of Taibah University for Science
11
(
5
),
689
699
.
http://dx.doi.org/10.1016/j.jtusci.2016.06.004
.
Rayaroth
M. P.
Aravind
U. K.
Aravindakumar
C. T.
2015
Sonochemical degradation of Coomassie Brilliant Blue: effect of frequency, power density, pH and various additives
.
Chemosphere
119
,
848
855
.
http://dx.doi.org/10.1016/j.chemosphere.2014.08.037
.
Rostami-Vartooni
A.
Nasrollahzadeh
M.
Salavati-Niasari
M.
Atarod
M.
2016
Photocatalytic degradation of azo dyes by titanium dioxide supported silver nanoparticles prepared by a green method using Carpobrotus acinaciformis extract
.
Journal of Alloys and Compounds
689
,
15
20
.
http://dx.doi.org/10.1016/j.jallcom.2016.07.253
.
Santos
M. M. M.
Duarte
M. M. M. B.
Nascimento
G. E.
Souza
N. B. G.
Rocha
O. R. S.
2018
Use of TiO2 photocatalyst supported on residues of polystyrene packaging and its applicability on the removal of food dyes
.
Environmental Technology
40
(
12
),
1494
1507
.
https://doi.org/10.1080/09593330.2017.1423396
.
Sharma
A. K.
Tiwari
R. K.
Gaur
M. S.
2016
Nanophotocatalytic UV degradation system for organophosphorus pesticides in water samples and analysis by Kubista model
.
Arabian Journal of Chemistry
9
,
S1755
S1764
.
http://dx.doi.org/10.1016/j.arabjc.2012.04.044
.
Utzig
L. M.
Lima
R. M.
Gomes
M. F.
Ramsdorf
W. A.
Martins
L. R. R.
Liz
M. V.
Freitas
A. M.
2019
Ecotoxicity response of chlorpyrifos in Aedes aegypti larvae and Lactuca sativa seeds after UV/H2O2 and UVC oxidation
.
Ecotoxicology and Environmental Safety
169
,
449
456
.
https://doi.org/10.1016/j.ecoenv.2018.11.003
.
Velo-Gala
I.
Pirán-Montaño
J. A.
Rivera-Utrilla
J.
Sánchez-Polo
M.
Mota
A. J.
2017
Advanced oxidation processes based on the use of UVC and simulated solar radiation to remove the antibiotic tinidazole from water
.
Chemical Engineering Journal
323
,
605
617
.
http://dx.doi.org/10.1016/j.cej.2017.04.102
.
Veréb
G.
Ambrus
Z.
Pap
Z.
Mogyorósi
K.
Dombi
A.
Hernádi
K.
2014
Immobilization of crystallized photocatalysts on ceramic paper by titanium(IV) ethoxide and photocatalytic decomposition of phenol
.
Reaction Kinetics, Mechanisms and Catalysis
113
,
293
303
.
https://doi.org/10.1007/s11144-014-0734-y
.
Verran
G. O.
Kurzawa
U.
Pescador
W. A.
2005
Reciclagem de latas de alumínio visando melhor rendimento e qualidade metalúrgica no alumínio obtido (Recycling of aluminum cans aiming at better efficiency and metallurgical quality in the obtained aluminum)
.
Revista Matéria
10
(
1
),
72
79
.
Vykoukalová
M.
Venier
M.
Vojta
Š.
Melymuk
L.
Bečanová
J.
Romanak
K.
Prokeš
R.
Okeme
J. O.
Saini
A.
Diamond
M. L.
Klánová
J.
2017
Organophosphate esters flame retardants in the indoor environment
.
Environment International
106
,
97
104
.
https://doi.org/10.1016/j.envint.2017.05.020
.
Wiedmer
D.
Sagstuen
E.
Welch
K.
Haugen
H. J.
Tiainen
H.
2016
Environmental oxidative power of aqueous non-irradiated TiO2-H2O2 suspensions: methylene blue degradation and the role of reactive oxygen species
.
Applied Catalysis B: Environmental
198
,
9
15
.
http://dx.doi.org/10.1016/j.apcatb.2016.05.036
.
Xing
X.
Du
Z.
Zhuang
J.
Wang
D.
2018
Removal of ciprofloxacin from water by nitrogen doped TiO2 immobilized on glass spheres: rapid screening of degradation products
.
Journal of Photochemistry and Photobiology A: Chemistry
359
,
23
32
.
https://doi.org/10.1016/j.jphotochem.2018.03.026
.