Degradation of meropenem by heterogeneous photocatalysis using TiO2/fiberglass substrates

Alejandro Altamirano Briones, Iván Cóndor Guevara, David Mena, Isabel Espinoza, Christian Sandoval-Pauker, Luis Ramos Guerrero, Paul Vargas Jentzsch* and Florinella Muñoz Bisesti 1 Departamento de Ciencias de la Tierra y la Construcción, Universidad de las Fuerzas Armadas ESPE, Av. Gral. Rumiñahui s/n, 171103 Sangolquí, Ecuador 2 Departamento de Ciencias Nucleares, Escuela Politécnica Nacional, Ladrón de Guevara E11-253, 170525 Quito, Ecuador. 3 Centro de Investigación de Alimentos_CIAL, Universidad UTE, Av. Mariscal Sucre y Mariana de Jesús, 170527 Quito, Ecuador. 4 Laboratorio de Fisicoquímica, Departamento de Química, Universidad Técnica Federico Santa María, Avenida España 1680, 2390123 Valparaíso, Chile.


Introduction
The occurrence of some contaminants of emerging concern (CECs) in the environment has been the focus of intense studies in recent years. The massive use of pharmaceuticals, personal care products, pesticides and others used in everyday life has contributed to the release of this type of compounds into the environment [1]. Among CECs, antibiotics are compounds largely applied to control bacterial infections and allow the treatment of many diseases. Indeed, the development of antibiotics has contributed to reducing mortality and morbidity rates [1,2].
In general, it is a well known fact that pharmaceutical compounds are not completely metabolized, are excreted from the human and/or animal body and then released to the environment [1,[3][4][5]. Moreover, they are not efficiently removed by conventional wastewater treatment plants [6], mostly

Characterization of the TiO 2 /Fiberglass Substrates
In Figure S1 (Supplementary material) are depicted photos of the fiberglass disc before and after the immobilization of TiO 2 . The quantity of TiO 2 immobilized on the substrates was determined by gravimetric analysis (see Table S1-Supplementary material). On average, 2.1578 ± 0.3319 mg/cm 2 of TiO 2 were supported on each substrate. It is important to remark that, probably, not all the TiO 2 particles were available for the photocatalysis since a portion of them could be encapsulated by silicone. However, considering that this method of immobilization allowed a stronger retention of particles on the supporting material (fiberglass) compared to other methods (e.g., dip coating), the loss of activity due to the encapsulation of TiO 2 nanoparticles is a minor disadvantage.
SEM images of the TiO 2 /fiberglass substrates were obtained to determine the morphological structure of the catalyst (Figure 1). The images showed that TiO 2 was homogeneously distributed over the fiberglass matrix. A certain degree of agglomeration of the catalyst was observed. These agglomerates presented sizes between 200 and 400 nm approximately (see Figure S2-Supplementary material).
Catalysts 2020, 10, x FOR PEER REVIEW 3 of 11 TiO2 were supported on each substrate. It is important to remark that, probably, not all the TiO2 particles were available for the photocatalysis since a portion of them could be encapsulated by silicone. However, considering that this method of immobilization allowed a stronger retention of particles on the supporting material (fiberglass) compared to other methods (e.g., dip coating), the loss of activity due to the encapsulation of TiO2 nanoparticles is a minor disadvantage. SEM images of the TiO2/fiberglass substrates were obtained to determine the morphological structure of the catalyst (Figure 1). The images showed that TiO2 was homogeneously distributed over the fiberglass matrix. A certain degree of agglomeration of the catalyst was observed. These agglomerates presented sizes between 200 and 400 nm approximately (see Figure S2-supplementary material). The determined point of zero charge (PZC) of the substrate is 5.7 (see the plot in Figure S3 -Supplementary material) which is close to PZC values reported in other studies for anatase TiO2 (5.8 to 6.0) [31][32][33]. The PZC value represents the pH at which the surface of the catalysts has the same number of positive and negative charges. This parameter is used to assess the charge of the catalyst surface and infer whether or not the adsorption can be favored by the pH of the solution. When the pH value is below the PZC, the surface is positively charged, whereas when the pH value is above the PZC, it is negatively charged. In the case of TiO2, this behavior can be explained with Equations 1 and 2 [24,31]: TiOH + HO ‾ → TiO ‾ + H2O (pH > PZC),

Photocatalytic Removal of Meropenem
The variation of the MER concentration due to the heterogeneous photocatalytic process with TiO2/fiberglass substrates is depicted in Figure 2. This process was evaluated at three pH values: 4.0, The determined point of zero charge (PZC) of the substrate is 5.7 (see the plot in Figure S3-Supplementary material) which is close to PZC values reported in other studies for anatase TiO 2 (5.8 to 6.0) [31][32][33]. The PZC value represents the pH at which the surface of the catalysts has the same number of positive and negative charges. This parameter is used to assess the charge of the catalyst surface and infer whether or not the adsorption can be favored by the pH of the solution. When the pH value is below the PZC, the surface is positively charged, whereas when the pH value is above the PZC, it is negatively charged. In the case of TiO 2 , this behavior can be explained with Equations (1) and (2) [24,31]: TiOH

Photocatalytic Removal of Meropenem
The variation of the MER concentration due to the heterogeneous photocatalytic process with TiO 2 /fiberglass substrates is depicted in Figure 2. This process was evaluated at three pH values: 4.0, 5.7 and 7.9. The degradation of MER followed a pseudo-first order kinetics according to the coefficients of determination, R 2 (0.933-0.976). 5.7 and 7.9. The degradation of MER followed a pseudo-first order kinetics according to the coefficients of determination, R 2 (0.933-0.976). The pseudo-first order rate constants of the MER photocatalytic degradation calculated for the different pH values tested are depicted in Table 1. A statistical analysis was carried out to determine the effect of the pH value on the removal of MER ( Figure S4-Supplementary material) and the results showed that no significant differences between the rate constants at pH 4.0 and 5.7 can be observed. Moreover, a decrease in the rate constant was observed at pH 7.9. The removals of MER attained after 60 min of photocatalytic reaction were 83.79%, 83.48% and 81.05% when the process was carried out at pH 4.0, 5.7 and 7.9, respectively. The decrement in the rate constant at a pH value of 7.9 could be explained in terms of the PZC of the substrate and the pKa values of MER. MER has two pKa values, the first one (2.9-3.1) is associated with the carbonyl group while the second one (7.4-7.5) is associated with the pyrrolidone group [34]. As mentioned before, the charge of catalyst surface at a specific pH value can be predicted based on the PZC of the material. In this case, when the pH value is above 5.7 (PZC of the material) the adsorption of MER on the substrate is not favored because of the repulsive electrostatic forces between the surface and MER (both are negatively charged). Considering that the adsorption phenomenon is an important step for the photocatalytic degradation of the pollutants [35], a decrease of the adsorption capacity of the material could also affect the photocatalytic performance of the catalyst.
The direct photolysis of MER under UV irradiation was tested. Figure 3 shows the variation of the MER concentration due to the direct photolysis and heterogeneous photocatalysis at pH 5.7. This pH value was selected by considering the results of the photocatalytic degradation of MER. Taking into consideration that no significant differences between the treatments carried out at pH 4.0 and 5.7 were observed, a pH value of 5.7 was selected for the treatment since it implies a lower operational costs at an industrial scale (less acid is required to adjust the pH). The photolysis experiments showed that MER can be degraded by direct photolysis, as reported previously [11]. Indeed, a MER removal of 66.44% was achieved after 60 min of reaction. Nevertheless, when the substrate was employed, an The pseudo-first order rate constants of the MER photocatalytic degradation calculated for the different pH values tested are depicted in Table 1. A statistical analysis was carried out to determine the effect of the pH value on the removal of MER ( Figure S4-Supplementary material) and the results showed that no significant differences between the rate constants at pH 4.0 and 5.7 can be observed. Moreover, a decrease in the rate constant was observed at pH 7.9. The removals of MER attained after 60 min of photocatalytic reaction were 83.79%, 83.48% and 81.05% when the process was carried out at pH 4.0, 5.7 and 7.9, respectively. The decrement in the rate constant at a pH value of 7.9 could be explained in terms of the PZC of the substrate and the pKa values of MER. MER has two pKa values, the first one (2.9-3.1) is associated with the carbonyl group while the second one (7.4-7.5) is associated with the pyrrolidone group [34]. As mentioned before, the charge of catalyst surface at a specific pH value can be predicted based on the PZC of the material. In this case, when the pH value is above 5.7 (PZC of the material) the adsorption of MER on the substrate is not favored because of the repulsive electrostatic forces between the surface and MER (both are negatively charged). Considering that the adsorption phenomenon is an important step for the photocatalytic degradation of the pollutants [35], a decrease of the adsorption capacity of the material could also affect the photocatalytic performance of the catalyst.
The direct photolysis of MER under UV irradiation was tested. Figure 3 shows the variation of the MER concentration due to the direct photolysis and heterogeneous photocatalysis at pH 5.7. This pH value was selected by considering the results of the photocatalytic degradation of MER. Taking into consideration that no significant differences between the treatments carried out at pH 4.0 and 5.7 were observed, a pH value of 5.7 was selected for the treatment since it implies a lower operational costs at an industrial scale (less acid is required to adjust the pH). The photolysis experiments showed that MER can be degraded by direct photolysis, as reported previously [11]. Indeed, a MER removal of 66.44% was achieved after 60 min of reaction. Nevertheless, when the substrate was employed, an improvement on the degradation of MER was observed. For comparative purposes, the degradation of MER by direct photolysis was adjusted to pseudo-first order kinetics (k = 0.0189 min −1 and R 2 = 0.992).
improvement on the degradation of MER was observed. For comparative purposes, the degradation of MER by direct photolysis was adjusted to pseudo-first order kinetics (k = 0.0189 min −1 and R 2 = 0.992). The adsorption of MER over the TiO2/fiberglass substrates at pH = 5.7 was also studied. According to the results, the time required to attain the adsorption equilibria was 10 min. Moreover, Langmuir and Freundlich isotherm models were applied to describe the adsorption data. Based on the coefficients of determination a better fit was found with the Langmuir model (R 2 = 0.925). The maximum adsorption capacity (qmax) and the Langmuir constant (KL) calculated for the MER adsorption were 0.052 mg/g and 0.073 L/mg, respectively.

Mineralization and Biodegradability Studies
The mineralization of MER by the photocatalytic treatment was studied in terms of chemical oxygen demand (COD) and total organic carbon (TOC). These experiments were conducted with an initial MER concentration of 100 mg/L, a pH value of 5.7 and 60 min of reaction. COD and TOC removals of 25.80% and 29.60% were achieved, respectively. These results suggest that the treatment could trigger the formation of less complex species. Nevertheless, the low COD and TOC removals in comparison to the removal of MER (83.79%) may be explained by the presence of different reaction intermediates which do not easily degraded in comparison with MER molecules [17,36].

Reuse and Regeneration of the Material
One of the main drawbacks of the application of heterogeneous photocatalysis is the deactivation of the catalysts generally due to the adsorption of the pollutant and/or its degradation products. For industrial purposes, it is desirable that the photocatalysts can be reused and/or regenerated. in this sense, the recyclability of the substrates was studied. The substrates were reused for five subsequent cycles of MER degradation. Figure 4 shows the MER photocatalytic degradation kinetics for each cycle. The activity of the substrates was slightly affected until the fourth cycle of reuse. However, a decrease in MER removal was observed during the fifth cycle (see Figure S5 -Supplementary material). Other studies have also reported the progressive fouling of TiO2 after several uses [37][38][39][40]. The adsorption of MER over the TiO 2 /fiberglass substrates at pH = 5.7 was also studied. According to the results, the time required to attain the adsorption equilibria was 10 min. Moreover, Langmuir and Freundlich isotherm models were applied to describe the adsorption data. Based on the coefficients of determination a better fit was found with the Langmuir model (R 2 = 0.925). The maximum adsorption capacity (q max ) and the Langmuir constant (K L ) calculated for the MER adsorption were 0.052 mg/g and 0.073 L/mg, respectively.

Mineralization and Biodegradability Studies
The mineralization of MER by the photocatalytic treatment was studied in terms of chemical oxygen demand (COD) and total organic carbon (TOC). These experiments were conducted with an initial MER concentration of 100 mg/L, a pH value of 5.7 and 60 min of reaction. COD and TOC removals of 25.80% and 29.60% were achieved, respectively. These results suggest that the treatment could trigger the formation of less complex species. Nevertheless, the low COD and TOC removals in comparison to the removal of MER (83.79%) may be explained by the presence of different reaction intermediates which do not easily degraded in comparison with MER molecules [17,36].

Reuse and Regeneration of the Material
One of the main drawbacks of the application of heterogeneous photocatalysis is the deactivation of the catalysts generally due to the adsorption of the pollutant and/or its degradation products. For industrial purposes, it is desirable that the photocatalysts can be reused and/or regenerated. in this sense, the recyclability of the substrates was studied. The substrates were reused for five subsequent cycles of MER degradation. Figure 4 shows the MER photocatalytic degradation kinetics for each cycle. The activity of the substrates was slightly affected until the fourth cycle of reuse. However, a decrease in MER removal was observed during the fifth cycle (see Figure S5-Supplementary material). Other studies have also reported the progressive fouling of TiO 2 after several uses [37][38][39][40].
After the fifth cycle of reuse, a simple regeneration method for the substrates was tested. This regeneration process consisted of washing the substrates with a 1% H 2 O 2 solution in an ultrasonic bath for 15 min and drying them at 80 • C for 5 h. Then, the regenerated substrate was employed for another MER degradation cycle. The results demonstrate that this simple procedure can restore the activity of the substrates (see Figure S5-Supplementary material). The regeneration of the catalyst could not only be explained by the removal of the substances adsorbed on the surface of the substrate by washing. Perhaps, the combined action of ultrasound irradiation and H 2 O 2 (which is a well-known advanced oxidation process) can generate •OH radicals which could oxidize the adsorbed pollutants and thus facilitate their desorption. Other works have also proven the capacity of H 2 O 2 to aid in the regeneration process of photocatalysts [40]. The reuse and regeneration of the substrates make them suitable for industrial wastewater applications by reducing the cost of the treatment and the amount of the spent catalyst discarded to the environment.  After the fifth cycle of reuse, a simple regeneration method for the substrates was tested. This regeneration process consisted of washing the substrates with a 1% H2O2 solution in an ultrasonic bath for 15 min and drying them at 80 °C for 5 h. Then, the regenerated substrate was employed for another MER degradation cycle. The results demonstrate that this simple procedure can restore the activity of the substrates (see Figure S5-Supplementary material). The regeneration of the catalyst could not only be explained by the removal of the substances adsorbed on the surface of the substrate by washing. Perhaps, the combined action of ultrasound irradiation and H2O2 (which is a well-known advanced oxidation process) can generate •OH radicals which could oxidize the adsorbed pollutants and thus facilitate their desorption. Other works have also proven the capacity of H2O2 to aid in the regeneration process of photocatalysts [40]. The reuse and regeneration of the substrates make them suitable for industrial wastewater applications by reducing the cost of the treatment and the amount of the spent catalyst discarded to the environment.

Reagents
Meropenem trihydrate was purchased from Chongqing Pharmaceutical (71.6% meropenem; impurities were 7.1% NaCl, 100 ppm acetone, 500 ppm ethanol, 1000 ppm isopropanol, <10 ppm metals, and water). Nanosized titanium dioxide (TiO 2 ) was provided by Nanostructured and Amorphous Materials Inc. (size of grain 15 nm, 99% anatase). Moreover, tetrabutylammonium hydroxide (20%), acetonitrile (LiChrosolv®) and methanol (LiChrosolv®) were bought from Merck. Phosphoric acid (85%) was obtained from Fisher. The structure of Meropenem (MER) is provided in Figure 5. After the fifth cycle of reuse, a simple regeneration method for the substrates was tested. This regeneration process consisted of washing the substrates with a 1% H2O2 solution in an ultrasonic bath for 15 min and drying them at 80 °C for 5 h. Then, the regenerated substrate was employed for another MER degradation cycle. The results demonstrate that this simple procedure can restore the activity of the substrates (see Figure S5-Supplementary material). The regeneration of the catalyst could not only be explained by the removal of the substances adsorbed on the surface of the substrate by washing. Perhaps, the combined action of ultrasound irradiation and H2O2 (which is a well-known advanced oxidation process) can generate •OH radicals which could oxidize the adsorbed pollutants and thus facilitate their desorption. Other works have also proven the capacity of H2O2 to aid in the regeneration process of photocatalysts [40]. The reuse and regeneration of the substrates make them suitable for industrial wastewater applications by reducing the cost of the treatment and the amount of the spent catalyst discarded to the environment.

Preparation of TiO 2 /Fiberglass Substrates
TiO 2 was immobilized on a fiberglass disc with an inner and outer diameter of 20 and 80 mm, respectively. Before the immobilization, the discs were washed with ethanol (50% v/v) and sonicated for 5 min. Subsequently, they were dried at 100 • C for 24 h. To guarantee a homogeneous weight distribution, only the discs with a mass of 1.8 ± 0.2 g were selected as support for the photocatalyst.
For the immobilization of TiO 2 , 0.75 g of liquid silicone was mixed with 15 mL of ethanol (96% v/v) at constant stirring. Once dissolved, 0.25 g of TiO 2 were added and stirred for 5 min. Then, the TiO 2 suspension was carefully spilled onto the fiberglass discs, spread to achieve a uniform distribution and then dried at 80 • C for 24 h. The amount of catalyst present in the prepared substrates was determined by gravimetric analysis using an analytical balance (CX220 Citizen).
The point of zero charge (PZC) of the TiO 2 /fiberglass substrates was determined using a methodology previously described [31,41]. For the morphological characterization, the prepared substrates were analyzed by a Scanning Electron Microscope (JEOL, Model: JSM-IT300).

Photocatalytic Degradation of Meropenem
Photocatalytic assays were carried out with solutions of MER (100 mg/L). The solutions were treated in a batch lab-scale system which consisted of a stirred reaction vessel, two UV lamps (8 W) and a cooling system to maintain the temperature at approximately 20 • C. The TiO 2 /fiberglass substrates were submerged into 25 mL of MER solutions. Then, the solutions were stirred for 35 min in the darkness (to ensure the adsorption equilibria). After the pre-equilibration step, the lamp was turned on to start the photocatalytic reaction. The tests were performed at different times of reaction (0, 2, 5, 10, 20, 40, 60 min) and three pH values (4.0, 5.7 and 7.9). A clean substrate was employed for each assay. The pH value that allowed the highest MER removal was chosen for further experiments. Moreover, to measure the degree in which MER can be degraded due to photolysis, MER solutions (100 mg/L) were irradiated without the substrates.
After each run, the remaining MER in the solutions (concentration) was measured by HPLC (Agilent Technologies 1120 compact LC) equipped with a variable wavelength UV-Vis detector and a Zorbax Eclipse Plus C-18 column from Agilent Technologies (4.6 mm × 150 mm; 5 µm particle size). The mobile phase consisted of a mixture of methanol, acetonitrile, and a buffer solution of tetrabutylammonium hydroxide adjusted to a pH of 7.0 with phosphoric acid (volume ratio 10:15:75). The flow rate of the mobile phase was set to 1.5 mL/min with an isocratic regime. The detection was carried out at 300 nm.
In addition, the adsorption of MER on the TiO 2 /fiberglass substrates was evaluated. The equilibrium time of adsorption was first determined. The substrates were submerged in 25 mL of MER solutions (20 mg/L) with stirring at 200 rpm in the absence of light. The assays were carried out at different times (5, 10, 20, 30, 40, 50 min). For each assay, the remaining MER in the solutions was measured by HPLC.
Once the equilibrium time was determined, the adsorption isotherm of MER on the substrates was obtained at room temperature (~18 • C). For this, the TiO 2 /fiberglass substrates were submerged in 25 mL of MER solutions (40, 50, 60, 70, 80, 100 mg/L) with stirring at 200 rpm in the absence of light until the adsorption equilibrium was reached. The concentration of MER after the tests was measured by HPLC. The resulting adsorption isotherm was adjusted to the linear form of Langmuir and Freundlich models which are presented in Equations (3) and (4), respectively [42].
ln(q e ) = 1/n · ln(C e ) + ln(K F ) where C e is the equilibrium concentration (mg/L), q e is the equilibrium amount of MER adsorbed on the substrate (mg/g), q max is the maximum adsorption capacity of the substrate (mg/g), K L (L/mg) is the Langmuir constant, K F (mg 1−n ·L n ·g −1 ) is the Freundlich constant, and 1/n is a constant related to the intensity of the adsorption in the Freundlich model.

Mineralization of MER
The mineralization of MER was studied in terms of the chemical oxygen demand (COD) and total organic carbon (TOC). For this, MER solutions (100 mg/L) were treated using the TiO 2 /fiberglass substrate. After 60 min of treatment, COD and TOC were measured according to the standard methods 5220 D and 5310 B, respectively [43].

Reuse and Regeneration of the TiO 2 /Fiberglass Substrates
The reuse and regeneration of the TiO 2 /fiberglass substrates were studied. Five subsequent cycles of (re)use were carried out using the same substrate. For these assays, portions of 25 mL of MER solutions (100 mg/L) were treated for 60 min. After the treatment, the used substrate was separated and dried at room temperature before its reuse in another treatment (this procedure was repeated four times). Additionally, after the fifth cycle, the substrates were regenerated and employed for an additional degradation run. The regeneration process consisted of washing the substrates with a 1% H 2 O 2 solution in an ultrasonic bath for 15 min and then drying them at 80 • C for 5 h.

Conclusions
The degradation and mineralization of MER using TiO 2 /fiberglass substrates as heterogeneous photocatalysts were studied. The photocatalysts were prepared by immobilizing TiO 2 on a fiberglass matrix. The morphology of the substrates was characterized by Scanning Electron Microscopy. A homogeneous distribution of TiO 2 over the fiberglass matrix was observed by SEM. The quantity of TiO 2 immobilized on the plates was on average 2.1578 ± 0.3319 mg/cm 2 . The photocatalytic degradation of MER was carried out at three pH values (4.0, 5.7 and 7.9). The statistical analysis showed that there were no significant statistical differences between the rate constants at pH 4.0 and 5.7 (0.032 min −1 in both cases). Nevertheless, a decrease in the rate constant was observed at pH 7.9 (0.028 min −1 ). The direct photolysis of MER with UV light was also tested. The rate constant for the photolysis of MER was 0.0189 min −1 which was lower than the rate constant of the photocatalysis. Thus, the addition of the substrates had a positive effect on the removal of this compound. The adsorption isotherm of MER on the substrates was adjusted to the Langmuir model with a maximum adsorption capacity (q max ) and Langmuir constant (K L ) of 0.052 mg/g and 0.073 L/mg, respectively. To provide insight into the mineralization of MER by the treatment, COD and TOC parameters were measured. COD and TOC removals were 25.80% and 29.60%, respectively. These values suggest that the treatment could trigger the formation of a less complex species. Moreover, the reuse and regeneration of the substrates were studied. For this, the substrates were reused for five subsequent cycles of MER degradation, the activity of the substrates was maintained until the fourth cycle of reuse. However, a decrease in MER removal was observed during the fifth cycle. After the fifth cycle of reuse, the activity of the substrates was recovered by the regeneration procedure.
Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4344/10/3/344/s1, Table S1: Average content of TiO 2 immobilized on the fiberglass substrates, Figure S2: SEM images of the TiO 2 /fiberglass substrate, Figure S3: Point of zero charge of the TiO 2 /fiberglass substrate, Figure S4: Statistical analysis of the effect of the pH value on the pseudo-first order rate constant of the photocatalytic degradation meropenem, Figure S5