Photocatalytic Degradation of Polyethylene Microplastics and Disinfection of E. coli in Water over Fe- and Ag-Modified TiO2 Nanotubes

In this study, Feand Ag-modified TiO2 nanotubes were synthesized via an anodization method as photocatalysts for degradation of polyethylene microplastics and disinfection of Escherichia coli (E. coli). The anodization voltage, as well as the Fe3+ or Ag+ concentrations on TiO2 nanotubes were evaluated and correlated to their corresponding photocatalytic properties. TiO2 nanotubes were firstly synthesized by anodization of Ti plates in a glycerol-based electrolyte, followed by incorporation of either Fe or Ag via a Successive Ionic Layer Adsorption and Reaction (SILAR) method with Fe(NO3)3 and AgNO3 as Fe and Ag precursors, respectively. UV-Vis DRS shows that the addition of Fe or Ag on TiO2 nanotubes causes a redshift in the absorption spectra. The X-ray diffractograms indicate that, in the case of Fe-modified samples, Fe3+ was successfully incorporated into TiO2 lattice, while Ag scatters around the surface of the tubes as Ag and Ag2O nanoparticles. A microplastic degradation test was carried out for 90 mins inside a photoreactor with UVC illumination. TiO2 nanotubes that are anodized with a voltage of 30 V exhibit the best degradation results with 17.33% microplastic weight loss in 90 mins. Among the modified TiO2 nanotubes, 0.03 M Ag-TiO2 was the only one that surpassed the unmodified TiO2 in terms of microplastic degradation in the water, offering up to 18% microplastic weight loss in 90 min. In terms of E. coli disinfection, 0.03M Ag-TiO2 exhibit better performance than its unmodified counterpart, revealing 99.999% bactericidal activities in 10 mins. Copyright © 2022 by Authors, Published by BCREC Group. This is an open access article under the CC BY-SA License (https://creativecommons.org/licenses/by-sa/4.0).


Introduction
The prevalent usage of plastics, particularly those with the size of less than 5 mm, (known as microplastics) has become a major environmental problem. Microplastics have been found on tap water, beverages, and commercial salt [1]. Along with the presence of bacteria, such as E. sary potentials of charge carriers (electrons and holes), degradation of organic compounds (such as microplastics and microorganisms) can be achieved through the formation of reactive oxygen species (ROS) in a humid, aerobic environment. The degradation of polyethylene microplastics using TiO2-based photocatalysts had been conducted and reported by several authors [11][12][13][14]. However, most of the studies used powdered TiO2, the downside of which is the need for post-process filtration to recover the used photocatalysts. Immobilization of Titania in a substrate through the formation of films is hence deemed more practical for the sustainability of the process. Titania with ordered nano-tubular arrays is one of the most attractive nanostructures of titania because of its high specific surface area, photon absorption efficiency, and compatibility with fast electron transport [15]. Previous studies conducted by Ali et al. [16] showed that titania nanotubes could effectively degrade polyethylene film. However, to the best of our knowledge, the degradation of polyethylene microplastics over photo-driven TiO2 nanotube arrays has yet to be explored.
The use of TiO2-based photocatalysts, however, is by no means without drawbacks. Two of the most quoted disadvantages of TiO2 are its wide bandgap and high rate of electron-hole recombination that limits its photocatalytic performance [12][13][14]17,18]. One of the considerations that can be taken to overcome this weakness is by incorporating metals ions into TiO2 structure [19]. Among the wide range [17] of available choices, Fe 3+ was of tremendous interest owing to its similar ionic radii to Ti 4+ that facilitates incorporation into the crystal lattice of TiO2 [20][21][22]. On the other hand, Ag is also a promising promotor for TiO2 as it is a noble metal that could reduce the recombination rate of the photogenerated electrons and holes through the formation of the Schottky Barrier on the Ag-TiO2 interface. It is also well documented that Fe-and Ag-modified TiO2 nanotubes were successfully fabricated through electrochemical anodization of Ti plates to grow nanotubular structure, which is followed by a successive ionic layer adsorption and reaction (SILAR) method to incorporate Fe and/or Ag [23][24][25][26]. Interestingly, besides being assigned as photocatalysts [23,25], Fe-and Agmodified TiO2 nanotubes fabricated via the anodization-SILAR combined methods were suitable for other applications, e.g. as electrocatalysts [24] and supercapacitors [26]. Despite these virtues, nonetheless, there remain limited studies that compare Fe and Ag as promotors for TiO2 nanotubes-, especially in degrading microplastics and disinfection of E. Coli. Furthermore, the titania nanotubes used by most researchers are mainly immobilized on rectangular titanium (Ti) plates [27][28][29][30].
In this paper, we report our study on the photocatalytic degradation of microplastics using Ag + -and Fe 3+ -decorated TiO2 nanotubes, as well as the antibacterial test of the composite photocatalysts. The nanotube arrays were formed on a curved plate -mimicking the shell of a cylinder-via an anodization method. Ag + and Fe 3+ were subsequently incorporated into TNT using the SILAR method to improve the photocatalytic properties. In particular, the effects of anodization voltage, Fe 3+ concentration, and Ag + concentration on the resultant features of the composites and their performance in terms of microplastic degradation and bacteria disinfection are emphasized.

Synthesis of TNT
Titanium plate (Shaanxi Yunzhong Metal Technology Co., Ltd.) with a thickness of 0.3 mm was cut into 17 cm  5 cm and curved to shape like the shell of a cylinder. It was then mechanically polished with sandpapers and chemically polished in a solution containing hydrogen fluoride (Merck, 40%), nitric acid (Merck, 65%), and distilled water (1:3:6). The polished plate was then cleaned using distilled water, sonicated for 10 min, and subsequently dried at ambient temperature. Nanotubular ar-rays of TiO2 were generated on Ti plate via an electrochemical anodization method [31], by assigning Ti plate as the anode and platinized Ti (Shaanxi Yunzhong Metal Technology Co., Ltd.) as the cathode, situated with 1-cm apart between each other. The anodization was carried out under continuous stirring for 45 mins in a 160-mL electrolyte solution containing glycerol (PT. Wiloso), 0.5% wt NH4F (Merck, 98%), and 25% volume of distilled water at different voltages: 10 V, 20 V, 30 V, and 40 V (denoted as TNT 10V, TNT 20V, TNT 30V, and TNT 40V respectively).

Doping Method
The introduction of either Fe or Ag dopant into TiO2 nanotubes was done after anodization by a Successive Ionic Layer Adsorption and Reaction (SILAR) method [23][24][25]. The amorphous TNT was firstly immersed in a 140-mL solution of either Fe(NO3)3 (Merck) as Fe 3+ precursor or AgNO3 (Merck) as Ag + precursor under continuous stirring for 5 mins. The concentration of the precursor salts were varied as follows: 0.03 M, 0.06 M, or 0.09 M (denoted as 0.03M Fe-TNT, 0.06M Fe-TNT, 0.09M Fe-TNT for Fe-modified TNT and 0.03M Ag-TNT, 0.06M Ag-TNT, 0.09M Ag-TNT for Ag-modified TNT). It was followed by natural drying at the ambient temperature for 30 s before being transferred to distilled water for 30 s. This procedure constitutes one SILAR cycle and the operation was repeated for 20 cycles. Finally, the modified TNT was cleaned with distilled water and then annealed at 500 °C for 3 h to induce crystallization.

Characterizations
The as-calcined samples were then cut and characterized. The superstructure and morphology of the nanotubes were investigated by Field Emission Scanning Electron Microscope (FESEM, JEOL JIB-4610F). The bandgap is obtained by extrapolating the linear region of the plot to intersect with the x-axis, which marks the estimated bandgap value of any given sample. The crystallite structure of the photocatalyst was identified using an X-ray diffractometer (XRD) with Cu anode (Kα 1.5406 Å) as a beam source, operating under 40 kV and 30 mA. The crystallite size was then estimated using the Debye-Scherrer equation. The photoabsorption spectra of the samples were measured using UV-Vis diffuse reflectance spectroscopy (UV-Vis DRS, Agilent Technologies Cary 60 UV-Vis, DRS from Hartrick Scientific with a scan rate of 600 nm/min). Accordingly, the band gap of the photocatalyst samples was estimated using the Kubelka-Munk function: (1) with: (2) where, R is reflectance, hv is the photon energy, K is a constant for TiO2, and Eg is the band gap. The parameter [F(R)hv] 1/2 is assigned as the y-axis and hv as the x-axis.

Microplastic Photocatalytic Degradation and E. coli Disinfection Test
The effectiveness of the photocatalyst samples was evaluated in terms of degradation of polyethylene microplastics and disinfection of E. Coli. The tests were carried out inside a photoreactor with UVC (Yamano 11W, λ = 254 nm) as the photon source. The total plate count (TPC) method was used to analyze the E. coli colony over time. On the other hand, An amount of 25 mg of polyethylene scrub microplastics (100 -150 microns) and 250 mL of distilled water was added into the reactor prior to irradiation. During irradiation, the photoreactor was wrapped with aluminum foil, while continuously stirred. The degradation of polyethylene microplastics was observed by filtering them using Whatmann Filter 42, then measuring its weight loss over time using a gravimetry method by weighing the mass of microplastics before and after photodegradation, according to the Equation (1). (1)

Photocatalyst Characterization
Figures 1 and 2, respectively, show the corresponding surface morphology and crosssectional view of titania nanotubes at different anodization voltages. At the low anodization voltage of 10 V, the nanotube arrays have not yet uniformly formed. The nanotubular structure is gradually formed as the voltage becomes higher. It was also found that the uniformity of the tubes is better at high voltages, as highly ordered nanotube arrays grow on the plate. Important parameters of the superstructure, namely tube diameter, wall thickness, and tube length, can accordingly be de-

Bulletin of Chemical Reaction Engineering & Catalysis, 17 (2), 2022, 268
Copyright © 2022, ISSN 1978-2993 duced from the micrographs, summarized in Tables 1 and 2. It is evident that all of nanotubes diameter, wall thickness, and height are almost linearly correlated with the anodization voltage. In terms of nanotube inner diameter, increasing anodic voltage from 10 V to 40 V leads to larger pores from 30 nm to 153 nm, while the wall becomes thicker from 10 nm to 29 nm. The tube also elongates from 140 nm to 1.69 µm following the increase of anodization voltage from 10 V to 40 V. The effect of anodization voltage on the average diameter of nanotubes can be related to the number of pits formed in the initial stage of anodization. As the anodizing voltage becomes higher, the electric field between the anode and the cathode increases so that the chemical dissolution (i.e. the formation of (TiF6) 2− complex) shall also increase, forming more pits. These pits form larger pores, resulting in larger nanotubes diameter. Concurrently, at higher anodizing voltages, the high chemical dissolution becomes more prominent, producing longer nanotubes as a result of the faster pore deepening rate [32]. The elongation is due to the increased driving force for ionic species (H + , F − , and O 2− ) to transport through the barrier layer at the bottom of the nanotube under electric field, which results in faster movement of the Ti/TiO2 interface into the Ti metal [33]. As discussed in the next section, the TNT samples prepared at 30 V and 40 V show superior activities than the samples anodized at lower voltages. For the purpose of further modifications and characterizations, the sample anodized at 30 V will be used because, at this voltage, the photocatalytic properties are sufficiently high and the use of higher voltage (40 V) does not offer further improvements. Figure 3 displays FESEM micrographs of Ag-doped TiO2 nanotubes at different concentrations of the precursor salt, along with the elemental mapping of Ag. In all cases, the Ag signals are observed which suggest the successful deposition of Ag particles on the TNT substrate. It can be inferred from Figure 6   modified TiO2 samples could be explained by the introduction of impurity levels in the band structure of TiO2 as a result of the incorporation of doping ions in the crystal lattice. In addition, in the case of Ag, the visible light absorption could also be originated from the collective oscillations of free electrons on Ag surfaces when interacting with visible electromagnetic radiation [34]. Table 3 summarizes the band gap of the prepared samples, in which it is informative that TNT has a bandgap of 3.15 eV, which is close to the known band gap for anatase (3.2 eV). Importantly, the addition of either Fe or Ag on TiO2 nanotubes leads to slight bandgap narrowing and causes a redshift in the light absorption capacity.
X-ray diffractometry (XRD) analysis ( Figure  5 [35,36]. An additional peak at 2θ = 32.4° is also observed, attributable to Ag2O based on JCPDS 41-1104. Unlike Fe 3+ , which has a comparable radius to that of Ti 4+ , the ionic radius of Ag + (1.26 Å) is higher than that of Ti 4+ . The incorporation of Ag + in TiO2 is hence difficult to achieve because of these inherent properties [37]. Consequently, the majority of Ag is expected to be dispersed on the photocatalyst surface. On the other hand, the XRD patterns for the Fe-modified samples reveal no characteristic signals attributed to Fe species. This could be attributed to the limitation of XRD detection, well dispersion of Fe, or because Fe 3+ is successfully incorporated into TiO2 lattice. Applying the Scherrer equation to estimate the crystallite size of anatase with (101) plane, the TiO2 crystallite size for TNT, 0.09M Fe-TNT, and 0.09M Ag-TNT are found to be 25.49 nm, 23.97 nm, and 44.25 nm, respectively [38].

Microplastic Degradation
The photocatalytic degradation of polyethylene microplastics is illustrated in Figure 6, presenting the degree of degradation (% weight loss) as a function of time for TNT samples anodized at different voltages. The overall performance of the photocatalysts is expected to be driven by the degree of nanotube for-  Table 3. Bandgap estimation of the synthesized photocatalysts. Copyright © 2022, ISSN 1978-2993 mation, namely the pore diameter and the tube length. As informative in Figures 1 and 2, applying a higher voltage of anodization results in more established and ordered nanotube arrays [32]. The photocatalytic properties seem to be consistent with this finding as the samples prepared under high anodization voltages offer better performance in terms of microplastic degradation. It is inferred from Figure 6 that TNT samples prepared at 30 V and 40 V could induce weight loss of microplastics up to 17% after 90 mins, while those prepared at 10 V and 20 V could only maintain degradation of 15% and 12%, respectively, for the same duration. Expectedly, larger pore diameter and longer nanotubes constitute higher specific surface area, resulting in better photon absorption and more effective carrier transport which eventually ensures higher photocatalytic performance. Although the size of microplastics is too large to diffuse into TNT pores, their removal could be driven by the formation of reactive oxygen species (ROS). ROS are produced by means of redox reactions driven by photogenerated charge carriers. The first step of microplastic degradation involves the generation of electrons and holes due to interaction between photons and TiO2 (Equation 3), which are subsequently responsible for the production of hydroxyl radicals (Equation 4) and superoxides (Equation 5). What follows are a series of reactions producing more ROS [39], which might involve the processes described in Equations 5-8. Once ROS are generated, the abstraction of polyethylene alkyl chains by hydroxyl radicals to form alkyl radicals occurs, as shown by Equation 9. Afterward, the propagation is in effect, leading to chain breaking of polyethylene groups. Finally, intermediate groups are generated, which can react further to produce CO2 and H2O as the final products, as shown by Equations 10-17 [16,[40][41][42]. Initiation: Propagation: Formation of carboxylic acid, aldehyde, and ketone: Norrish Type I reaction: (15) Norrish Type II reaction (formation of vinyl from ketone): Mineralization: (17) Figure 7 shows FTIR spectra of a microplastic sample before and after photocatalytic degradation. Four strong peaks associated with C−H stretching (2915 cm −1 and 2848 cm −1 ) and C−H bending (1463 cm −1 and 719 cm −1 ) confirm the characteristics of polyethylene FTIR spectra. For the sample examined after the photocatalytic reactions, a signal attributable to the formation of carbonyl groups is observed in the range of 1665-1760 cm −1 , indicating the existence of intermediate products from the photocatalytic degradation of microplastics. Furthermore, the surface morphology of microplastics before and after photocatalytic degradation was also elucidated using an optical microscope. It is observable that, after degradation, the surface of microplastics is ruptured and becomes more irregular in shape as compared to its original conditions. This result is consistent with the observed weight loss of microplastics over time, which can be   expected due to the conversion of polyethylene into harmless products (CO2 and H2O). As the final product of the degradation is in the gas phase, it does not contribute to the final weight of the microplastics. The photographs of microplastics before and after photocatalytic reaction can be seen in Figure 8. Figure 9(a) depicts the weight loss of microplastics as a function of illumination time on Fe-modified TNT. It can be seen that the photocatalytic performances of all Fe-modified TNT are inferior to that of their unmodified counterpart. In the use of the unmodified TNT, around 17% weight loss of microplastics was obtained, while the Fe-modified samples could only maintain 6-9% of microplastic degradation. This is presumably because the concentration of Fe-modified on TNT is too high. Fe 3+ can act as both electron trap and hole trap by forming Fe 2+ or Fe 4+ , respectively, which may potentially improve photocatalytic performance (Equations 18-20). However, if  the amount is too abundant, Fe may rather act as recombination centers [43], hindering photocatalytic activities (Equations 21 and 22). In addition, as displayed in Table 3, increasing Fe content leads to slight decreases in bandgap. Indeed, a lower bandgap is beneficial for visible light absorption, but it also induces higher recombination [44]. Since the photon source used in the present experiments illuminates UV, the effect of visible light absorption is superimposed, while the side effect of recombination retains. The schematic diagram elaborating the competitive processes between trapping and recombination offered by the presence of Fe in Fe-modified TiO2 is illustrative in Figure 9(b).
According to the crystal field theory, both Fe 2+ and Fe 4+ are relatively unstable compared to Fe 3+ [43]. They tend to revert to Fe 3+ due to the loss of d 5 electronic configuration. These reactions can compete with either hydroxyl or superoxide formation. The high concentration of Fe dopant in the photocatalyst increases the possibility of multiple trappings [45]. The high amount of Fe 3+ doped on TiO2 reduces the distance between trapping sites [46], which is Fe 3+ in this case. Fe 3+ is then believed to mediate the recombination between electrons and holes before they can reach the photocatalyst surface, depriving the photocatalytic degradation of microplastics. The negative effects caused by Fe dopant were also found by Zhu et al. [45] and Sood et al. [47]. At suitable concentrations, the addition of Fe 3+ dopant can improve the performance of the photocatalyst by increasing the radical/ROS species formed, which had been reported by other authors [48][49][50]. The beneficial effects of Fe 3+ as a dopant are expressed in Equations 23 to 26. However, the interfacial charge involving Fe 3+ is valid only when Fe 3+ is located close to the surface sites [48].
Electron transfer from Fe 2+ to adsorbed oxygen:  as an electron trapper by forming Schottky Barrier. The Schottky barrier hinders the reverse flow of electrons to the semiconductor, thus the electron-hole recombination can be prevented [51]. The schematic diagram in Figure 10(b) is intended to attempt at elucidating this effect. Furthermore, the presence of Ag nanoparticles on TiO2 surface contribute to the formation of •O2 − , as expressed in Equations (27) to (29): Ag-TiO2 + hv (UV) → Ag-TiO2 (e − + h + ) (27) Electron transfer from TiO2 conduction band to Ag through Schottky Junction : Reduction of adsorbed oxygen on Ag surface: Importantly, further increase in Ag content for the cases of 0.06 M Ag-TNT and 0.09 M Ag-TNT is detrimental to microplastic degradation performances, which are even poorer than that achieved by the unmodified TNT. The diminished photocatalytic activities at higher Ag concentrations are thought to be due to the formation of Ag clusters on the surface of the photocatalyst, thereby reducing the active site of Titania in absorbing photons. Furthermore, as confirmed by XRD patterns, the presence of Ag2O could cause photocatalysts to suffer from poor activities. It is well documented that the conduction band of Ag2O lies below (more positive than) the standard O2/•O2 − redox potential, so that the electrons in the Ag2O conduction band are lack thermodynamic requirements to reduce the adsorbed O2 on the Ag2O surface [52][53][54][55]. With the same line of arguments, holes in the Ag2O valence band are thermodynamically incapable of oxidizing the adsorbed H2O on the Ag2O surface because •OH/H2O is higher (more negative) than the Ag2O valence band. On top of this, the formation of a large number of Ag2O clusters can reduce the active sites of the photocatalyst in producing radical species.

Antibacterial Test
From the results of photocatalytic tests on microplastic degradation, we then further expound the virtues of Ag-modified TNT as photocatalysts for disinfection of E. coli. Figure 11 displays the profiles of the E. coli colony as a function of time for three cases: 1) in the absence of UV illumination and photocatalysts; 2) in the presence of UV illumination and the unmodified TNT; 3) in the presence of UV illumination and 0.03 M Ag-TNT. In the first case, there are essentially no bactericidal activities as the E. coli colony seems invariant with time, as expected. Under illumination, the unmodified sample (TNT 30V) could reduce the E. coli colony by 97.94% within 10 minutes and 99.66% within 60 minutes. Importantly, the 0.03 M Ag-TNT photocatalyst exhibit much better antibacterial properties, as it performs E. coli disinfection of up to 99.999% within 10 minutes under UV-C irradiation. It should be noted that the value of 10 CFU/mL was used on 0.03 M Ag-TNT profiles to indicate the detection limit (<10 CFU). The bactericidal activity could be rationalized by considering the synergetic effect of UVC illumination and photocatalysis. The UVC spectrum is strongly absorbed by the nucleic acids of a microorganism and, therefore, is lethal for microorganisms. The light-induced damage to the DNA and RNA of a microorganism often results from the dimerization of pyrimidine molecules [56]. Hydroxyl radicals generated during the photocatalytic process further enhanced the microorganism disinfection properties. Other ROS such as atomic oxygen and hydrogen peroxide can also be actively involved in the oxidation of cellular components, membrane leakage of the microbial cell wall, and other processes [57,58]. Moreover, it is also well documented that Ag inherently possesses an antimicrobial capacity. Besides playing a role in increasing photocatalytic activities by performing an electron-trapping function to retard electron-hole recombination, Ag could also react with water to form Ag + and interact with the SH (retinol) Figure 11. Profile of E. coli colony as a function of time under dark and illuminated conditions, with two different photocatalysts. group in the bacteria, which eventually kills them [59].

Conclusions
In this paper, we report the synthesis of TiO2 nanotube arrays that are further modified by either Fe or Ag via a SILAR method to promote photocatalytic activities in degrading microplastics and disinfecting E. Coli. Based on our findings, it is suggested that Fe 3+ is more easily incorporated into TiO2 lattice than Ag + ; the latter tends to scatter around the surface of TiO2. The addition of Fe 3+ into TiO2 nanotubes is apparently detrimental to the photocatalytic degradation of microplastics, as it causes the decrease of microplastic degradation from 17% in the use of unmodified TNT to 6% -9% in the case of Fe-modified samples. This is presumably because the concentrations are too high for Fe 3+ to act as an electron trapper and rather become a recombination center. In contrast, modification by Ag is evidently beneficial for the removal of microplastics with delicate adjustment on the loading content. A suitable concentration of Ag (prepared in 0.03 M of precursor solution) added to TiO2 nanotubes may induce a Schottky barrier that is essential for effective separation of the photo-charge carriers, while an excessive amount (0.06-0.09 M) would create clusters of Ag and Ag2O particles that block the photon absorption by photoactive TiO2. In terms of disinfection of E. coli, the Ag-modified TiO2 nanotube sample exhibit exceptional bactericidal properties, by killing 99.999% of the E. coli colony in 10 mins. This research is expected to pave the way for further development in maintaining water quality from microplastic contaminants and bacteria via a sustainable process route.