Application of TiO2-Based Photocatalysts to Antibiotics Degradation: Cases of Sulfamethoxazole, Trimethoprim and Ciprofloxacin

The extensive application of antibiotics in human and veterinary medicine has led to their widespread occurrence in a natural aquatic environment. Global health crisis is associated with the fast development of antimicrobial resistance, as more and more infectious diseases cannot be treated more than once. Sulfamethoxazole, trimethoprim and ciprofloxacin are the most commonly detected antibiotics in water systems worldwide. The persistent and toxic nature of these antibiotics makes their elimination by conventional treatment methods at wastewater treatment plants almost impossible. The application of advanced oxidation processes and heterogeneous photocatalysis over TiO2-based materials is a promising solution. This highly efficient technology has the potential to be sustainable, cost-efficient and energy-efficient. A comprehensive review on the application of various TiO2-based photocatalysts for the degradation of sulfamethoxazole, trimethoprim and ciprofloxacin is focused on highlighting their photocatalytic performance under various reaction conditions (different amounts of pollutant and photocatalyst, pH, light source, reaction media, presence of inorganic ions, natural organic matter, oxidants). Mineralization efficiency and ecotoxicity of final products have been also considered. Further research needs have been presented based on the literature findings. Among them, design and development of highly efficient under sunlight, stable, recyclable and cost-effective TiO2-based materials; usage of real wastewaters for photocatalytic tests; and compulsory assessment of products ecotoxicity are the most important research tasks in order to meet requirements for industrial application.


Introduction
Pharmaceuticals (analgesics and anti-inflammatories, antibiotics, anticonvulsants, β-blockers, endocrine active pharmaceuticals, lipid regulators, psychiatric drugs [1]) have become an essential part of our life, benefiting society's health and wellbeing. Improved life expectancy, the better prevention of diseases and significantly reduced mortality could not be imagined without appropriate medication [2]. The consumption of pharmaceuticals continues to rise because of the ongoing demographical changes: growing world population, population density and aging societies (especially in developed countries) [1,3]. A growing population also means that more food is needed, leading to the increase in number and density of livestock, especially in intensive farming [1]. As a result, the demand for pharmaceuticals used for veterinary purposes also rises. Besides, the continuous progress of medical science, improvements in research and development, high health-care investments and the availability of the global market contribute to the development of more and more new drugs and their wider availability [2,3].
However, a huge environmental concern arises when pharmaceutical substances enter natural aquatic ecosystems [1,2,4,5]. Approximately 2000 pharmaceutically active compounds are administered globally in prescription and non-prescription medicines, as well as veterinary drugs [1,6]. These active pharmaceutical ingredients (APIs) are complex molecules that belong to micropollutants, as they are found in the range of ng/Lµg/L in water bodies [7]. APIs are being found worldwide ( Figure 1) in surface water, groundwater, seawater, drinking water, sediment, soil, manure and biota [3,6,8,9]. Such a widespread occurrence of pharmaceuticals is a result of their high consumption and increased production [1].

Antibiotics Occurrence in Aqueous Environment
Among all the pharmaceuticals, antibiotics pose one of the most significant concerns. They are widely used to treat human infectious diseases, as well as for the treatment and prevention of infections in animals, apart from promoting their growth [10,11]. Approximately 100,000 tons [12,13] of antibiotics are manufactured annually in the world, and 50-60% of this amount is spent on veterinary purposes [8,13,14].
Antibiotics are chemotherapeutic agents that kill or inhibit the metabolic activity of various microorganisms, such as bacteria, fungi, viruses, microalgae or protozoa [9,11,15,16]. They can be natural, synthetic or semi-synthetic compounds, which show antibacterial, antiparasitic or antifungal activities [11,16]. There are different groups of antibiotics, depending on their mechanism of action or chemical structure. They are, for example, β-lactams (amoxicillin, cefalexin), aminoglycosides (kanamycin, neomycin, streptomycin), diaminopyrimidines (trimethoprim), glycopeptides (teicoplanin, vancomycin), macrolides (azithromycin, erythromycin, tylosin), quinolones (enrofloxacin, ciprofloxacin, levofloxacin, norfloxacin, ofloxacin), sulfonamides (sulfamethoxazole, sulfathiazole, sulfamethazine), tetracyclines (tetracycline, oxytetracycline) and others [4,10,11,14]. According to the assessment of antibiotic consumption in 76 countries, global consumption grew to 42.3 billion defined daily doses (DDDs) between 2000 and 2015, which is a 39% increase [6,16] and is still growing. The reason for that is obviously growth of population and, therefore, increasing demand of the animal protein [16]. As a result, antibiotics are being found in natural water bodies across the whole world [4,9,12]-concentrations of different antibiotics from various classes found in surface waters, in areas away from wastewater treatment plants (WWTPs) and industrial production sites, are depicted in Figure 2. Although individual antibiotics are always detected at quite low concentrations, ranging from ng/L to µg/L [10,17], the simultaneous presence of several antibiotics even at low concentrations might result in their much stronger cumulative effect [14,16]. Properties of antibiotics clearly demonstrate why their presence in natural aquatic environment is highly undesirable: antibiotics are, first of all, persistent, so that they are non-degradable and are able to withstand natural decomposition [1,6,18]. Secondly, they are highly mobile, bioaccumulative and toxic [1,2,19]. Antibiotics are deliberately designed to affect living organisms and cause pharmacological response even at low doses. This means that aquatic organisms exposed to such pharmaceuticals may develop negative chronic effects, which affect their behavior, growth and reproduction [16].

Sulfamethoxazole, Trimethoprim and Ciprofloxacin as Antibiotics of Great Concern
Among the antibiotics, sulfamethoxazole (SMX), trimethoprim (TMP) and ciprofloxacin (CIP) are the most often detected substances [3,12,15] (Table 1, Scheme 1, their physicochemical properties are presented in Table 2). These pharmaceuticals have been found at high concentrations in influents and effluents of wastewater treatment plants, surface waters (rivers and seas), groundwater, and drinking water, not only across Europe, but worldwide [15,20,21]. Apart from that, maximum concentrations of SMX, TMP and CIP detected in Africa were~100 times,~54 times and~125 times, respectively, higher than those in Europe [22]. This happens because these antibiotics are widely used in human and veterinary medicine, have the largest numbers of manufacturers [22], are toxic and persistent in the aquatic environment [4,16]. Table 1. Concentrations of SMX, TMP and CIP detected in surface waters worldwide.

Antibiotic Sulfamethoxazole Trimethoprim Ciprofloxacin
Number of countries where antibiotic was detected 47 29 20 Average concentration (µg/L) 0.095 0.037 18.99 Maximum concentration (µg/L) 53.8 13.6 6500 Scheme 1. Molecular structure of SMX, TMP and CIP and their 3D models. pK a1 , pK a2 1.7; 5.6 [23] 3.2; 6.7 [23] 5.9; 8.9 [24] Solubility in water (mg/L) [8] 610 400 30000 Sulfonamides are one of the oldest antibiotics groups that is still widely used [6]. All substances from this group have been detected in WWTPs effluents, surface water and groundwater, as well as in drinking water [10]. Sulfamethoxazole is the most often detected sulfonamide in the environment [6,10,23]. It is usually prescribed together with trimethoprim to treat various infectious diseases in humans [10,25,26]. Fluoroquinolones are also very often found in water bodies, especially ciprofloxacin. Hydrophilic properties of this antibiotic class make them highly mobile in the aquatic environment. This, together with their high consumption both in human and veterinary medicine, contributes to wide occurrence of ciprofloxacin, from WWTP effluents to drinking water [6,27,28]. Sulfamethoxazole, trimethoprim and ciprofloxacin have been identified in hospital wastewater with concentrations up to 35.5 µg/L [17,29,30]. Besides, sulfanomides and fluoroquinolones have demonstrated very high toxicity of aquatic organisms, especially to cyanobacteria, freshwater algae and duckweeds [10]. These organisms play a crucial role in aquatic ecosystems [10], and therefore whole ecosystems can be disturbed when antibiotics are released into them [31]. Such a wide spreading of these antibiotics resulted in adding ciprofloxacin to the second EU watch list of substances for union-wide monitoring in the field of water policy (EU Decision, 2018/840 of June 5, 2018), and sulfamethoxazole and trimethoprim were selected to be included in the third watch list [32].

Pathways of Antibiotics Release into Aquatic Environment
Discharge of antibiotics into the natural aquatic environment occurs as a result of (1) industrial production of antibiotics; (2) consumption and excretion of antibiotics by humans and animals; (3) improper disposal of unused or expired antibiotics [1]. Hotspots of antibiotic discharge in the environment include wastewater treatment plants (receiving effluents from households and hospitals), antibiotic manufacturing plants, agriculture and aquaculture ( Figure 4) [1,6,9,16]. Disposed antibiotics from households and hospitals end up in landfill, where they can eventually leak to pollute ground water. The usage of farming animal manure and WWTP sludge contaminated with antibiotics as fertilizer also results in the occurrence of antibiotics in ground water [1]. This is because 30 to 90% of oral antibiotic doses taken both by humans and animals are excreted in its unmetabolized form [1,6,17]. This means that antibiotics can be released into the environment as entirely biologically active substances [37,38]. In groundwater, under anoxic conditions, antibiotics remain unchanged for a very long time or might undergo minor degradation and produce even more toxic metabolites [6]. The consumption of groundwater (most often used as a source of drinking water) contaminated with antibiotics, ARB or ARGs might lead to the development of antibiotic resistance in humans [6]. Industrial and agricultural runoff (rainfall and snowmelt runoff) also results in antibiotic occurrence in natural water bodies [2,8,13,39]. Wastewater treatment plants are often recognized as the main source of pharmaceuticals discharge into the environment [37,40]. The reason for this lies in the common design of WWTPs.
Wastewater treatment at WWTP usually comprises a combination of physical, chemical and biological treatment processes. First, physical methods are utilized: large sediments, grit and grease are removed during pre-treatment and primary treatment stages using physical barriers and gravity forces [35]. Physical processes include, for example, sedimentation, flocculation, filtration, etc. [35,41,42]. Afterwards, secondary treatment place-biological processes are employed in order to remove organic content [35,39,42] through aerobic and anaerobic processes. For this purpose, biological reactors are utilized, in which the biochemical degradation of organic substances takes place in a similar way to the one occurring in natural aquatic systems [35]. Biological processes employ various microorganisms (bacteria, fungi, algae) and plants to transform biodegradable organic substances into simple products [41,42]. Biological methods utilize biomass in different forms: sus-pended (activated sludge), fixed (trickling filters) or suspended in aquatic systems (aerated lagoons) [35,42]. Tertiary treatment takes place in exceptional cases, only when additional water purification is required; for example, the disinfection or removal of specific nutrients or toxic compounds [35]. For this purpose, different chemical and physical methods are used: chlorination (most often), ozonation, UV treatment, membrane filtration, adsorption processes [35,43].
Although common treatment processes are very efficient to remove suspended solids, biodegradable organic compounds, inorganic matter and pathogens, they fail to remove pharmaceuticals and antibiotics, in particular [1,35,44,45]. Biological treatment, which is the foundation of any WWTP, is simple and easy to operate, cost-effective and does not generate toxic by-products [39,41,46]. However, the degradation efficiency of nonbiodegradable antibiotics by this method is low [41,42,[45][46][47]. Additionally, what is worse, antibiotics present in wastewater are able to change the composition of microbial communities in bioreactors [8,35] and make bacteria there multi-resistant to antibiotics. As a result, discharge from WWTPs containing antibiotic resistant bacteria and genes, contributes to the wide spread of AMR in the aquatic environment. This poses a significant threat to human health if receiving water bodies are used as a source of drinking water or for recreational purposes [48].
In this review, different advanced water treatment techniques, which are capable of the effective removal of antibiotics when applied in WWTP, are briefly introduced and compared. Among them, various advanced oxidation processes are analyzed, and heterogeneous photocatalysis is highlighted. The fundamentals and mechanism of heterogeneous photocatalysis over TiO 2 are discussed in detail. Surface modification strategies of TiO 2 nanomaterials in order to improve their photoactivity are presented. With this in view, the focus of the present review is on the removal efficiency of SMX, TMP and CIP from the aqueous environment by photocatalysis over various recently developed unmodified and modified TiO 2 materials. The role of different operating parameters on the removal efficiency is outlined. The ecotoxicity of the final products is also considered. Finally, further research needs are presented. We expect that the present review will provide some new ideas for the design and development of TiO 2 -based photocatalysts to be successfully applied in antibiotics elimination from the aquatic environment.

Water Treatment Techniques for Degradation of Antibiotics
It is clear from the previous section that conventional wastewater treatment plants have not been designed to eliminate pharmaceuticals from wastewater, and therefore they are unable to provide the efficient treatment of water contaminated with antibiotics [35,41,46,47,49]. Therefore, the application of advanced water treatment technologies is required in order to achieve the effective removal of pharmaceuticals in aqueous systems. Among these technologies, adsorption, membrane filtration and advanced oxidation processes gain a lot of interest. Short descriptions, advantages and drawbacks are compiled in Table 3. Table 3. Advanced technologies applied to the treatment of wastewater contaminated with antibiotics.

Technology Description Advantages Drawbacks
Adsorption A mass transfer process of accumulation of chemicals from the liquid phase into the solid phase of adsorbent [23,42]. A wide variety of adsorbents is utilized (clays and minerals, metal oxides, polymers, nanocomposites, carbon nanotubes; activated carbon and biochars are among the most commonly used) [23,46].
High efficiency of pollutants degradation and possibility of complete mineralization in a short period of time [43,46,47] High operation costs [43,46]. In case of incomplete mineralization post-treatment is required to remove toxicity [43] Include environmentally friendly, safe and sustainable processes [52] Disinfection properties [43] AOPs: Ozonation Ozone molecule O 3 has high oxidation capability (E 0 = 2.07 V), and therefore is able to efficiently oxidize organic pollutants [41,42].

Technology Description Advantages Drawbacks
AOPs: Electrochemical processes Processes of oxidizing organic substances using electric current [53]. Most widely employed anode materials include graphite and TiO 2 , as well as Ti-based alloys, Ru and Ir oxides [41].
Easy to operate, safe and highly efficient process [23,43] High operation costs [23,42,45] No chemical reagents required and no generation of secondary wastes [23,41] Expensive electrodes [41], short electrode life time, electrode fouling [23] Suitable for waste waters with high concentrations of pharmaceuticals [45] It is required that wastewater is highly conductive (otherwise electrolytes should be added) [23,41,42] Mass transfer resistance [41] Applicable to wastewater with low flow rate [45] Water oxidation occurs faster than oxidation of organic pollutants [42] AOPs: Ultrasonication Processes employing sound waves for formation, growth and collapse of bubbles in liquid media [23].
Fast and energy-efficient process [23] Toxicity of intermediates in mineralization process should be considered [23] No chemical reagents required [23] Various factors affect degradation efficiency (dose of radiation, pH, water matrix composition) [51] AOPs: Catalytic wet peroxide oxidation Processes of pollutants degradation in aqueous media through catalytic H 2 O 2 reduction to OH − and • OH under extreme pressures and temperatures [42].
Fast and efficient process [23] High operation and material costs [23] Some substances (containing nitro functional groups and halogens) are difficult to degrade [42] AOPs: Photolysis Processes employing light (artificial or natural) for the generation of reactive species and subsequent degradation of organic pollutants [41].
Easy to operate, highly efficient and environmentally friendly process [18,42,51] Fast recombination of photogenerated charge carriers decreases process efficiency [42] Performed under ambient temperature and pressure and utilizes atmospheric oxygen as oxidant [23,44] Limited visible light response [42] Complete mineralization of organic pollutants is possible, no waste generation [42,43] Laboratory scale [43] Efficient recovery and reuse of photocatalysts is possible [52,54] Not applicable to water with high concentrations of organic pollutants [42,51] Losses of photocatalyst under long-term operation [51] Toxicity of by-products should be considered [42] As follows from the information in Table 3, physical processes, such as adsorption and membrane filtration, have one significant limitation-the production of secondary waste. The problem is that pollutants are transferred from the liquid phase of wastewater to the solid phase of adsorbents in the case of adsorption process, or are collected on the membrane surface in the case of membrane filtration without any decomposition [44]. This means that pharmaceuticals do not undergo any changes in chemical structure and therefore, still retain their harmful potential [35]. Moreover, the obtained waste should be properly disposed of, which makes a treatment technology more expensive. Thus, the main advantage of advanced oxidation processes is the obvious, efficient degradation of pollutants with the possibility of their complete mineralization.
Currently, AOPs are widely studied by different research groups, as these methods give very promising results and have high potential for broad application. Carried out studies state that the best performance is associated with the antibiotic removal strategies that involve combined processes [41,46]. They include, for example, a combination of AOPs with biological treatment [43,57] or membrane filtration [42,57]. However, further improvements are needed, mainly in the development of new materials for AOPs [46], in order to make these processes cost-effective while retaining high efficiency.

Advanced Oxidation Processes
Advanced oxidation processes (AOPs) are powerful technologies used in wastewater treatment to degrade organic pollutants, including various pharmaceuticals [58]. Most common AOPs include the Fenton process, heterogeneous photocatalysis, ozonation, electrochemical oxidation, etc., and are briefly described and compared in Table 3. As can be seen from the table, AOPs include many different processes, and the only feature that unites all of them is the production of highly reactive species, for example, hydroxyl radicals ( • OH), ozone molecules (O 3 ), superoxide radicals (O 2 •− ), hydrogen peroxide molecules (H 2 O 2 ), etc.
The efficient generation of hydroxyl radicals ( • OH) in situ is the purpose of AOPs, as it defines their performance. These radicals are the second highest powerful species after fluorine with oxidizing potential E 0 = 2.80 V [52,55]. • OH radicals attack organic molecules rapidly and non-selectively, oxidizing them and thus transforming them into more biodegradable and less toxic substances. What is more, • OH radicals are also capable of mineralizing organic contaminants, yielding CO 2 , H 2 O and inorganic ions. In this case, organic pollutants are completely destroyed and therefore no further post-treatment is required.
The combination of different AOPs is gaining lots of interest due to enhanced process efficiency. For example, the performance of homogeneous Fenton reaction, being the most widely applied AOP [55], can be significantly increased through combination with UV light (photo-Fenton [23,41]), with electrochemical process (electro-Fenton [55]), or even with both (photo-electro-Fenton [42]).
After Fenton process, heterogeneous photocatalysis is the second most widely applied AOP [55]. This technology has numerous advantages over the other types of AOPs, mainly simplicity and sustainability. The possibility of the utilization of solar energy as an abundant, clean and renewable light source makes solar photocatalysis a green and sustainable technology for wastewater treatment. It is also very cost-efficient compared to the technologies utilizing artificial sources of light or electrodes. Besides, it was found that solar photocatalysis-based AOPs have the lowest global warming potential compared to the other wastewater treatment methods [57].
What makes TiO 2 such an effective photocatalyst, is the favorable thermodynamic processes taking place in TiO 2 photocatalysis. It is known that the conduction band (CB) energy of TiO 2 anatase (with a band gap E g = 3.2 eV) is -0.51 V [63] at neutral pH, while valence band (VB) energy is 2.69 V under the same conditions. This means that CB energy level is more negative than potential of oxygen reduction (E 0 (O 2 /O 2 •− ) = −0.33 V), and that VB energy level is more positive than the potential of water oxidation (E 0 (H 2 O/ • OH) = 2.29 V). This makes it possible to obtain highly reactive oxidizing species-superoxide radicals O 2 •− and hydroxyl radicals • OH in photocatalytic process over TiO 2 , which non-selectively degrade and mineralize organic pollutants.
Adsorption of organic pollutants on TiO 2 surface; 2.
The photocatalytic degradation of the adsorbed organic pollutants via oxidationreduction reactions with photogenerated electrons, holes and reactive species (depicted in Figure 5); 3.
Desorption of degradation products. Therefore, photocatalytic process itself is initiated when TiO 2 absorbs a photon from the light source with the energy (hv), equal to or greater than the band gap (E g ) of the photocatalyst [65]. This makes the photocatalyst electrons e − excited, so that they move from the valence band to the conduction band, and holes h + are generated in the valence band (Equation (1)) [45,54].
After excitation, separated charge carriers (e − and h + ) might recombine in the bulk with an energy release (Equation (2)) [45]. Otherwise, they migrate to the surface of TiO 2 , where they participate in redox reactions with organic pollutants previously adsorbed on the photocatalyst surface. The recombination of the charge carriers on the catalyst surface might also take place.
Besides, holes oxidize OH − at the catalyst surface, and more • OH radicals are produced (Equation (4)).
Hydrogen peroxide H 2 O 2 might react with superoxide radicals O 2 •− (Equation (9)) or is decomposed by the light (Equation (10)), and • OH radicals are produced in both cases.
In TiO 2 photocatalysis, the oxidation of organic substances might take place either through aliphatic oxidation ( • OH radicals remove H + from the molecule), or through aromatic oxidation (O 2 and O 2 •− oxidize aromatic molecule) [62].

Doping and Modification of TiO 2
Although TiO 2 has high photocatalytic activity, there are two main drawbacks limiting its wide practical application [67,68]. These limitations include unfavorable dynamics of photogenerated charge carriers (fast recombination of electrons and holes in bulk and on the photocatalyst surface) and poor visible light harvesting [63]. Wide band gap of TiO 2 (3.0-3.2 eV [63]) limits TiO 2 usage to UV light only (wavelengths shorter than 390 nm) according to Equation (13) [63]: where h is the Planck constant, c is the speed of light, E g is the band gap of the photocatalyst (3.2 eV for TiO 2 anatase) [63]. It is known that UV light accounts for only 3-5% of the solar light, while visible light constitutes 45% of the solar spectrum [18,54]. It is very important that process is cost-effective and sustainable when implemented on an industrial scale. If heterogeneous photocatalysis over TiO 2 cannot be carried out under visible light, expensive UV light sources have to be used [18,44]. It is obvious that for successful TiO 2 employment under visible light, the wavelength adsorption of the photocatalyst should be extended into the visible region, thus making TiO 2 active in visible light [44,66].
Therefore, in order to prevent the recombination of electrons and holes and ensure the effective utilization of visible light, particular measures should be taken. The separation of charge carriers should be enhanced, their lifetime should be prolonged, photocatalyst band gap should be narrowed and TiO 2 surface area should be increased [63,69]. For this purpose, various strategies of TiO 2 surface modification have been proposed [55]. They include mainly metal and non-metal doping [18,44,54,70], metal and non-metal co-doping [44], coupling with other semiconductors [18,44,66], constructing heterojunctions [54,63] and Z-schemes [45], dye sensitization [44].
In metal doping, metal ions (transition metals, noble metals, rare earth metals) [18,69] substitute for Ti 4+ . This suppresses electron-hole recombination, effectively reduces the photocatalyst band gap, enhances visible light harvesting and improves photocatalyst surface morphology [18,45,71]. Nonmetal dopants (N, C, S, B, F) also reduce the band gap and are stated to be more effective in making TiO 2 active in visible light [18,71]. In codoping, TiO 2 is doped with two or more metals, nonmetals or both metals and nonmetals, in order to overcome the limitations of a single-doped photocatalyst [18].
Coupling TiO 2 with another semiconductor (SnO 2 , ZnO, SiO 2 , CdS) to obtain heterojunctions also contributes to the better separation of the charge carriers and improves utilization of visible light [18,45,64]. In the case of Z-scheme photocatalysts, band structure remains similar to the one of heterojunctions. However, the transfer mechanism of electrons and holes changes. In particular, recombination of the charge carriers with weaker redox abilities take place, while electrons and holes with stronger redox abilities are separated and preserved [45]. Another way to enhance the separation of electrons and holes is to create vacancy engineered photocatalysts containing crystal defects-oxygen vacancies. Thus, a new energy level is formed between the valence band and conduction band and, as a result, band gap is reduced, and visible light response is improved [45]. Dye-sensitized photocatalysts are also used to generate more electrons and therefore enhance photocatalytic performance [70,71].
Still, there are many other parameters affecting the efficiency of the photocatalytic degradation process; for example, organic pollutant concentration, photocatalyst dosage, light intensity, pH of reaction solution, etc. [51]. Thus, process conditions should be properly designed, numerously tested and optimized, so that maximum performance can be achieved while keeping costs to a minimum.

Effect of Antibiotic Concentration
The initial concentration of a contaminant is a very important parameter in a photocatalytic process. Researchers study its effect on the photodegradation of pharmaceuticals by varying the initial concentration of target antibiotics under other similar conditions.
The effect of initial concentrations of SMX was studied by Tiwari et al. [72] in photocatalytic removal over thin film photocatalyst Ag0(NP)/TiO 2 (B). Researchers reported a decrease in photodegradation efficiency from 57% to 20% when SMX concentration was increased from 0.5 mgL −1 to 15.0 mgL −1 .
Researchers attribute the hindered decomposition of pharmaceuticals when antibiotic concentration is increased to a number of reasons. First of all, higher contaminant concentration means that higher number of contaminant molecules are adsorbed on active sites of the photocatalyst [75,77]. As a result, generation of reactive species is suppressed [75]. Secondly, as antibiotic concentration increases, more byproducts and intermediates are produced that compete with antibiotic molecules for a limited number of active sites on photocatalyst surface and reactive species present [24,73,77,78]. In addition, solution transmittance might be reduced when pharmaceutical concentration in the solution is increased. Weaker transmittance results in longer pathways for photons to get to photocatalyst, so that lower number of photons is adsorbed on photocatalyst surface and, therefore, a lower number of reactive species are generated [77,78]. Moreover, some antibiotics (for example, CIP) adsorb photons themselves, so that a lower number of photons are available for the photocatalyst [24].

Effect of Catalyst Concentration
The influence of photocatalyst concentration on removal efficiency of different antibiotics has been studied by many researchers. For example, Xie et al. [79] performed photocatalytic experiments under visible light irradiation utilizing different concentrations of a photocatalyst (Zn-TiO 2 /pBC): 0.625 gL −1 , 1.25 gL −1 and 2.5 gL −1 . The optimal concentration was found to be 1.25 g, resulting in 81.21% (k app = 0.0087 min −1 ) removal efficiency of 10 mgL −1 SMX after 3 h of irradiation. For comparison, 0.625 gL −1 and 2.5 gL −1 catalyst concentrations could achieve 57.73% (k app = 0.0043 min −1 ) and 76.28% (k app = 0.0075 min −1 ) SMX degradation under the same conditions, respectively. The same results were obtained previously by Zhang et al. [80] who tested TiO 2 /pBC photocatalyst for SMX removal under the same conditions but utilizing UV light instead of visible.
Obviously, increased catalyst concentration results in improved photocatalytic performance only up to a point, and any catalyst dosage beyond the optimal value leads to a decrease in removal efficiency. Such results were obtained also in the works of Ioannidou et al. [74], Kim and Kan [81], Chiang and Doong [82], Sarafraz et al. [24], Karim and Shriwastav [76], Hassani et al. [78] and Abellan et al. [83].
Researchers explain this phenomenon as follows: an increase in photocatalyst concentration increases the number of active sites available on the catalyst surface [82]. A higher number of active sites absorb more photons and, as a result, a higher number of reactive species (for example, hydroxyl radicals) are produced, which take part in antibiotic mineralization [24,74,79]. As a result, higher antibiotic removal and process efficiency can be achieved. However, excessive catalyst dosages limit the utilization of light because of greater photocatalyst aggregation and increased turbidity of the reaction solution [78][79][80]. The first one leads to a decrease of the catalyst active surface area, while the latter, to the reduced penetration of light through the light scattering effect [24,76,80,82]. This results in the generation of smaller number of reactive species and, therefore, photocatalytic efficiency decreases.
However, the situation is the opposite when it comes to mineralization efficiency. Kim and Kan [81] studied the effect of biochar/TiO 2 catalyst concentration (2.5-10 gL −1 ) on the removal efficiency of 0.1 L SMX solution (10 mgL −1 ). Although increased catalyst dosage led to a slight decrease in photocatalytic SMX removal from 86% to 76%, COD removal efficiency improved from 35% to 65%. Authors explain this phenomenon as follows: SMX can be better degraded by photolysis in UV-C light (SMX absorbs light at 250-270 nm [81,84]), than by photocatalysis. High photocatalyst loading results in high UV light absorption on TiO 2 surface. As a result, interaction between UV light and SMX decreases, and therefore photolysis efficiency also decreases. However, higher photocatalyst dosage contributes to the enhanced generation of • OH radicals and, therefore, significantly improves COD removal efficiency that is impossible under photolysis only.

Effect of pH
pH plays a very important role in the photocatalytic decomposition of organic pollutants. Its change affects the concentration of H + and OH − ions in a solution [88], which has a direct influence on the generation of reactive species (photogenerated holes h + , hydroxyl radicals • OH, superoxide radicals O 2 •− ) and surface chemistry (surface charge of photocatalyst, reactivity of pollutants and electrostatic interaction between pollutants and photocatalyst) [76,77,88].
Photocatalysis efficiency greatly depends on the surface properties of the photocatalyst. The surface charge of TiO 2 in aqueous solutions depends on the solution pH and zerocharge point of TiO 2 . The point of zero charge (pH PZC ) is pH at which catalyst surface is uncharged and is calculated as follows [63]: Ti-OH 2 + ↔ Ti-OH + H + (pK a1 ) Ti-OH ↔ Ti-O − + H + (pK a2 ) (16) pH PZC of TiO 2 P25 is reported to be~6.2 (pK a1 = 4.5 and pK a2 = 8) [63]. Thus, TiO 2 is positively charged at pH < 6.2 and negatively charged at pH > 6.2 [75]. Alfred et al. [73], Wang et al. [89], Yuan et al. [84], Xie et al. [79], Tiwari et al. [72], Zhang et al. [80], Cai and Hu [75] studied the influence of pH on the photocatalytic degradation of SMX and observed decrease in SMX removal efficiency with an increase in initial pH. For example, the kinetic parameters of SMX photocatalytic degradation gradually decreased from k = 0.0087 min −1 to k = 0.0049 min −1 when pH increased from 5.03 to 10.97, respectively [79]. This can be explained when the ionic states of both pharmaceutical compound and photocatalyst are considered. The pK a values of SMX are reported to be as follows: pK a1 = 1.85 ± 0.30 and pK a2 = 5.60 ± 0.04 [75,84]. It means that SMX exists in cationic form when pH < 1.85, in anionic form when pH > 5.6, and in neutral form when 1.85 < pH < 5.6. When SMX and TiO 2 surfaces are both positively or negatively charged, there is no electrostatic attraction between them. As a result, adsorption cannot contribute to SMX removal, and only photocatalytic degradation takes place [90]. Besides, when initial pH is close to pH ZPC of photocatalyst, catalyst aggregation might occur, resulting in reduced surface area and decreased photodegradation efficiency [84]. Results suggest that the neutral form of SMX (pH 3-5) is more easily photodegraded than its charged species [73,75], as SMX shows stronger light absorption and higher photochemical reactivity in its neutral form [84].
On the contrary, Mourid et al. [90] and Dlugosz et al. [91] reported higher rates of photocatalytic SMX removal at higher pH than at lower pH. Such contradicting results of pH effect on SMX photocatalytic removal demonstrate importance of taking into account the role of reactive species. When the main reactive species in the degradation process are holes (h + ), low pH values favor efficiency of photocatalytic decomposition [89]. On the other hand, if the main reactive species are • OH, then an increase in pH leads to an • OH content increase and, therefore, enhances photodegradation efficiency [84,89,90]. For example, in their research, Wang et al. [89] found that h + played major role in the photocatalytic degradation of SMX, while de Matos Rodrigues et al. [92] reported • OH and O 2 •− to be the main reactive species in SMX decomposition. This clearly demonstrates that all the experimental conditions should be considered in order to evaluate the pH effect on SMX photocatalytic removal.
Cai and Hu [75] found that the photocatalytic removal rate of TMP (pK a = 6.7) was almost similar at pH 3, 5, and 9 (~90%) and lower at pH 7 (68%). The authors attributed such a performance to the effect of pH on the absorption spectrum of the antibiotic. Luo et al. [93] also did not observe any significant influence of pH on TMP photodegradation.
The influence of pH on the decomposition rate of CIP was studied by Sarafraz et al. [24] who carried out photocatalytic degradation of CIP at initial pH of 4, 5.5, 7, 8.5 and 10 and obtained the following kinetic parameter values: 0.0237, 0.0417, 0.0763, 0.028 and 0.0196, respectively. The highest removal rate of CIP was observed at pH 7. A similar trend (increase in removal efficiency with pH increase up to a point and a subsequent decrease in efficiency with further pH increase) was also reported by Hu et al. [94] (highest efficiency at pH 6.3), Karim and Shriwastav [76] (highest efficiency at pH 7), Hassani et al. [78] (highest efficiency at pH 5), Gad-Allah et al. [95] (highest efficiency at pH 5.8). Such results can be explained by the ionic values of CIP and the surface charge of different photocatalyst used. It is known that CIP has two pK a values: pK a1 = 5.9 and pK a2 = 8.9 [24,77,78]. It means that CIP is present in cationic form CIP + when pH < 5.9, in zwitterionic form CIP ± when 5.9 < pH < 8.9 and in anionic form CIP − when pH > 8.9 [24,77]. If TiO 2 with pH PZC = 6.2 is used as a photocatalyst, then, in alkaline conditions when both TiO 2 surface and CIP are charged negatively (in acidic conditions, both charged positively), electrostatic repulsion occurs between them, decreasing removal efficiency [24,78,94]. Results obtained in the abovementioned studies demonstrated that photocatalytic decomposition efficiency was significantly improved at neutral pH, when CIP was present in its zwitterionic form [24,76]. Besides, predominant reactive species in the process should also be taken into account [77]. Sarafraz et al. [24], Hu et al. [94], Li and Hu [96] found that h + and • OH played a major role in the CIP degradation process. At the same time, Li et al. [97], Huang et al. [98] and Du et al. [99] confirmed h + and O 2 •− to be the main reactive species, while Gan, Zhang and Xiong et al. [100] reported all three radicals (h + , • OH and O 2 •− ) to play a significant part in CIP photocatalytic decomposition.

Effect of Presence of Inorganic Ions
Inorganic anions naturally present in water bodies have a significant impact on the photocatalytic removal of organic pollutants. Their presence affects both working pH and ionic strength of a reaction solution that eventually leads to the change in concentration of active sites on the photocatalyst surface, catalyst adsorption capacity and electrostatic interaction between organic molecules and the catalyst surface [84]. Besides, inorganic anions compete with antibiotic molecules for the same active sites on the photocatalyst surface [79,84]. These anions are also able to react with reactive species-photogenerated holes and hydroxyl radicals, decreasing their amount in the system. When h + are captured by inorganic anions, the production of • OH is hindered and, therefore, the decomposition rate of antibiotics decreases [24,78,79,89]. The presence of some inorganic ions might also result in photocatalyst agglomeration, reducing the photocatalytic efficiency of the process [79].
Yuan et al. [84]  − anions were added, respectively. Researchers attributed inhibited SMX decomposition to the suppression of reactive species (h + , • OH) by anions. On the contrary, the slightly accelerated removal of SMX in case of H 2 PO 4 − addition was attributed to the effect of a phosphorus-containing anion. It was reported to be able to enhance e − transfer between h + and water molecules to form hydroxyl radicals • OH and, therefore, improve SMX decomposition. Similarly, Xie et al. [79] reported the weakened photocatalytic performance of TiO 2 in SMX removal when SO 4 2− , Cl − , NO 3 − were present in the reaction system, and Wang et al. [89] when HCO 3 − anions were added.
Photocatalytic oxidation of ciprofloxacin was affected in a similar way. Sarafraz et al. [24] found that HCO 3 − , SO 4 2− , Cl − , and NO 3 − anions slightly decreased the efficiency of CIP photocatalytic decomposition. Among the studied anions, HCO 3 − anions were reported to cause the most significant inhibitory effect, while Cl − anions were the least significant. Hassani et al. [78] also assessed the influence of HCO 3 − , SO 4 2− , I − , and Cl − anions on CIP photocatalytic removal and obtained different results regarding the highest inhibition. Researchers reported a decrease in decomposition efficiency from 61.70% to 52.87%, 42.94%, 39.86% and 25.83%, respectively.

Effect of Presence of Natural Organic Matter
Natural organic matter (NOM) is widely present in natural water bodies in relatively high quantities (0.1-50 mgL −1 ) [96]. It consists of different aromatic and aliphatic organic compounds, including humic substances [84,89,96] and negatively affects the efficiency of photocatalytic processes. The photocatalytic decomposition of antibiotics is reduced for a number of reasons. First of all, NOM competes with antibiotic molecules for the same active sites on the photocatalyst surface and for the same number of reactive species ( • OH, h + and O 2 •− ) present in reaction media [84,87,89,96,101]. Secondly, NOM can absorb light both in the UV and visible range that is also harvested by the photocatalyst, thus creating an inner filter effect [84,96,101]. As a result, photocatalyst performance is significantly reduced. However, the inhibitory impact of natural organic matter depends greatly on the type and concentration of NOM, photocatalyst properties and target antibiotics [84,96,101].
Awfa et al. [101] assessed the photocatalytic oxidation of SMX in the presence of NOM using pure TiO 2 P25 and magnetic carbon nanotube-TiO 2 P25 composites (MCNT-TiO 2 ). Researchers reported the significantly reduced rate constants of SMX removal from 4.7 × 10 −2 min −1 to 5 × 10 −3 min −1 over TiO 2 P25; and from 4.9 × 10 −2 min −1 to 7 × 10 −3 min −1 over MCNT-TiO 2 in the absence and in the presence of NOM, respectively. Yuan et al. [84] studied the effect of three different types of NOM on SMX photocatalytic decomposition and observed decreased antibiotic removal when NOM concentration was increased. Photodegradation rates of SMX oxidation dropped from 0.544 min −1 in the absence of NOM to 0.081-0.141 min −1 in the presence of NOM, depending on the NOM type. Similar results were obtained by Li and Hu [96], who observed a negative influence of different NOM isolates on CIP photocatalytic removal. Sarafraz et al. [24] also carried out research on the influence of NOM different concentrations on CIP photocatalytic removal. Researchers reported the inhibition of CIP decomposition rate from 91.5% to 59% when the concentration of NOM was increased from 5 to 30 mgL −1 .

Effect of Reaction Media
At present, antibiotics can be widely found in natural water bodies and different wastewater effluents [79,102]. These water matrices are complex systems-they contain large amounts of organic matter or various inorganic anions that significantly affect photocatalysis efficiency due to the reasons described above [80,103]. Many researchers have carried out studies on the photocatalytic degradation of pharmaceuticals in different water matrices in order to evaluate their inhibitory effect.
Wang et al. [89] studied the photocatalytic degradation of SMX in natural water bodies under solar light. Authors reported decreased removal efficiency (86% in river water and 90.6% in lake water) compared to 100% removal efficiency in deionized water. Inhibited photocatalytic performance was attributed to the presence of organic matter and high turbidity in natural water that resulted in the lower utilization efficiency of the photocatalyst. Xie et al. [79] also observed a significant decrease in photocatalytic degradation of SMX from~81% in ultrapure water to~54% in river water under visible light [79]. Carbajo et al. [102] used natural ground water with reduced carbonates to study the photocatalytic degradation of SMX. The values of apparent kinetic constants decreased from 0.095 min −1 in deionized water to 0.061 min −1 in natural ground water when the home-made TiO 2 photocatalyst was utilized. Hassani et al. [78] also observed the inhibited photocatalytic decomposition of CIP when well water was used as a matrix.
Porcar-Santos et al. [104] carried out the photocatalytic decomposition of SMX using TiO 2 P25 under simulated solar irradiation, both in deionized water and simulated seawater. As a result, SMX degradation was twice as fast in deionized water than in saline water (pseudo-first order rate constants were 0.041 min −1 and 0.020 min −1 , respectively). The investigation of the degradation mechanisms revealed a major contribution of reactive halogen species on the photocatalytic degradation of SMX in seawater, while hydroxyl radicals ( • OH) played their main role in photocatalytic SMX decomposition in deionized water. Yang et al. [105] utilized seawater (34‰ salinity), diluted seawater (20‰ salinity) and distilled water to study the effect of salinity on the photocatalytic degradation of SMX. Authors observed a significant decrease in photocatalysis efficiency with an increase in salinity: the remaining amount of SMX increased from 7.7% in distilled water to 34% in diluted seawater and to 49% in seawater. Apart from that, the inhibitory effect of cations and anions present in seawater was studied and reported to be as follows: Ca 2+ > Mg 2+ > K + > Na + for cations, I − > SO 4 2− > Cl − > Br − for anions. Interestingly, the mineralization rate of the antibiotic was also severely affected. The remaining COD concentration increased from 34% in freshwater experiments to 75% and 84% when salinity was 20‰ and 34‰, respectively.
Malesic-Eleftheriadou et al. [103] carried out photocatalytic experiments with SMX as a target antibiotic using effluent from an urban wastewater treatment plant. The increased complexity of the water matrix resulted in decreased kinetic rates from 0.015 min −1 in distilled water to 0.03 min −1 in wastewater effluent. Ioannidou et al. [74] studied SMX photocatalytic degradation in three different matrices: ultrapure water, secondary treated wastewater and drinking water. Researchers reported slightly decreased removal efficiency in drinking water (apparent rate constant~0.1 min −1 ) and significantly inhibited the photocatalytic process in wastewater (apparent rate constant~0.015 min −1 ) compared to the results obtained when ultrapure water was used as a matrix (apparent rate constant 0.13 min −1 ). Karim and Shriwastav [76] observed a decrease in CIP photocatalytic removal from 77% to 41% when secondary treated wastewater was used as a matrix instead of distilled water. Large amounts of non-target organic compounds in wastewater effluents consumed reactive species and thus interfered with the photocatalytic decomposition of target antibiotics [74,103].

Effect of Dissolved Oxygen and Oxidants
The concentration of dissolved oxygen is an important parameter that affects the efficiency of photocatalytic processes. High contents of dissolved oxygen enhance the photocatalytic oxidation of organic pollutants for a number of reasons. First of all, oxygen acts as a trap for photoexcited electrons, decreasing the recombination rate of charge carriers [106]. Secondly, the reaction between oxygen and electrons results in the generation of reactive species: superoxide radicals O 2 •− , hydroxyl radicals • OH and peroxide radicals • HO 2 that improve the oxidation rate of pharmaceuticals [107]. As the concentration of dissolved oxygen constantly changes in natural water bodies [90], many researchers have studied its influence on the photocatalytic oxidation of antibiotics.
Mourid et al. [90] studied the effect of dissolved oxygen on SMX removal by carrying out photocatalytic tests in aerated and de-aerated media. The decomposition of the antibiotic decreased from 80% in aerated medium to 45% in de-aerated medium, proving the significance of dissolved oxygen in the photocatalytic process. Diao et al. [107] obtained similar results when CIP was treated under aerobic and anoxic conditions, as removal efficiency was reduced from 97.2% to 56.1%, respectively. The purging of nitrogen (N 2 ) gas through the reaction system to remove dissolved oxygen was reported by Khan et al. [108] to have an inhibitory effect on the efficiency. CIP photodegradation dropped from 85.91% to 70.98%. Xekoukoulotakis et al. [106] performed photocatalytic tests under two different conditions. In the first case, an open to air reactor was used, while in the second case, additional oxygen supply was provided. Researchers did not observe a significant influence of mode on CIP photocatalytic degradation. However, mineralization rate was greatly affected, as TOC reduction increased from 76% to 96% when aeration was applied.
The presence of different chemical oxidants can also enhance the efficiency of the photocatalytic process. For example, as Hassani et al. [78] observed in their research, the addition of hydrogen peroxide, potassium iodate and potassium persulfate increased the decomposition efficiency of CIP from 61.70% to 83.36%, 86.43% and 97.88%, respectively. Improved photocatalytic performance was attributed to the increased generation of reactive radicals, mainly • OH, through various mechanisms.

Effect of Light Source
The efficiency of the photocatalytic decomposition of target pharmaceuticals greatly depends on the wavelength of light applied and light intensity [45,88]. An increase in light intensity increases the number of photons in a reaction system. Thus, more charge carriers are generated that take part in photocatalytic oxidation and contribute to an improved removal rate of antibiotics [45]. Taking into account that pure TiO 2 has a band gap of 3.0 eV (rutile phase) and 3.2 eV (anatase phase) [47], it can utilize light in the UV region only (wavelengths below 390 nm) [54]. However, various TiO 2 -based photocatalysts can be successfully applied under visible light [109]. Many researchers have studied the performance of TiO 2 -based photocatalysts under UV-light [87,[110][111][112], visible light [77,79,113,114], simulated solar light irradiation [94,101,104,115] and natural sunlight [116][117][118]. Eskandarian et al. [119] studied the photocatalytic removal of SMX using LED lamps emitting light in different ranges of UV: UV-A (365 ± 10 nm), UV-B (300 ± 5 nm) and UV-C (260 ± 10 nm). Researchers reported 58%, 80% and 100% SMX decomposition efficiency after 3 h of irradiation when UV-A, UV-B and UV-C LED lamps were employed, respectively. The mineralization of the antibiotic followed the same trend, increasing from 35.1% (UV-A) to 59.9% (UV-C). An increase in CIP removal efficiency was observed in the same order (UV-A < UV-B < UV-C) by Hassani et al. [78]. Authors attributed such results to the increase in energy of photons emitted. A higher number of reactive species was produced under UV-C light compared to UV-B and UV-A. Lin et al. [120] compared the photocatalytic performance of the synthesized TiO 2 -based photocatalysts under UV irradiation and visible light source. The removal efficiency of SMX was significantly improved when UV light was used due to the higher utilization rate of photons. Although higher rates of antibiotic decomposition can be achieved under UV light [120,121], TiO 2 -based photocatalysts have the potential to be efficient enough under solar irradiation [116,118,122] that is a clean and sustainable source of energy [45] and, therefore, is more preferable.

Degradation Pathways of Antibiotics
Many researchers have studied degradation pathways of sulfamethoxazole in photocatalytic processes over TiO 2 and TiO 2 -based materials, for example, Yuan et al. [ Photodegradation routes of trimethoprim in TiO 2 photocatalysis are less studied than those of sulfamethoxazole. Sirtori et al. [125] and Samy et al. [126] investigated phototransformation products of TMP when TiO 2 P25 and S-TiO 2 were employed as catalysts, respectively. Both research groups reported hydroxylation, demethylation, and cleavage to be the main transformation pathways in the TMP (C 14  , m/z 305) is a product of TMP oxidation. Based on the identification of these phototransformation products, a suggestion of possible TMP degradation pathway is presented in Scheme 3 (products are labeled as P followed by the corresponding mass of the compound). The photodegradation pathways of ciprofloxacin over TiO 2 and TiO 2 -based materials have been studied by many research groups: Wang et al. [114], Sarafraz et al. [24], Manasa et al. [118], Karim and Shriwastav [76], Huerta-Aguilar et al. [109], Li and Hu [96], and Calza et al. [127]. Obtained photoproducts varied greatly and are summarized in an excellent review by Hu et al. [128]. In CIP photodegradation, four main transformation reactions can be outlined: cleavage of piperazine ring, defluorination, decarboxylation and oxidation of the cyclopropyl group. Possible photoproducts in the CIP degradation process are presented in Scheme 4 (products are labeled as P followed by the corresponding mass of the compound).

Ecotoxicity of Photoproducts
Considering a wide variety of transformation products occurring during the photocatalytic degradation of antibiotics, evaluation of their toxic effects on ecosystems is highly important for the practical application of the process in water treatment.
The phytotoxicity of photocatalysis products towards plant Lepidium sativum was studied by Borowska et al. [123]. After 60 min of solar irradiation of TiO 2 catalysts doped with noble metals, toxicity was eliminated: germination index increased from 43 ± 4% for initial 1 mgL −1 SMX solution to 93 ± 10% and 110 ± 16% for solutions treated with Pd/TiO 2 and Pt/TiO 2 , respectively. Decreased ecotoxicity was attributed to very low concentrations of remaining antibiotics. Negligible toxicity towards Vibrio fischeri and fully removed antibiotic activity against E. coli was achieved by Durán-Álvarez et al. [121], when 30 mgL −1 CIP solution was treated using TiO 2 modified with bi-metallic nanoparticles. This was a result of the complete mineralization of the parent compound after 360 min of photocatalytic treatment under simulated sunlight irradiation.
The antimicrobial activity of 15 mgL −1 CIP solution towards Vibrio fischeri was diminished significantly upon a photocatalytic process with graphitized mesoporous carbon-TiO 2 nanocomposite under UV-C irradiation in the study carried out by Zheng et al. [129]. Resulting non-toxicity was attributed to the almost complete mineralization of the antibiotic after 120 min of reaction. Cai and Hu [75] applied UV/TiO 2 P25 system and observed no acute toxicity (30 min of incubation) of 1 ppm SMX and 1 ppm TMP towards Vibrio fischeri. However, significant chronic toxicity (24 h of incubation) was reported. Sarafraz et al. [24] treated CIP solutions using a N-TiO 2 /visible LED system that resulted in significantly decreased toxicity towards Daphnia magna compared to the parent compound. A reduction in antibacterial activity of CIP against Staphylococcus aureus was also obtained by Gan et al. [130], who applied synthesized TiO 2 under artificial sunlight.
Gong and Chu [86] also observed a significant reduction of toxic activity against aquatic organisms when a UV-C/CoFe 2 O 4 /TiO 2 system for 100 µM SMX degradation is utilized. In the case of green alga Chlorella vulgaris, SMX was suggested to have been transformed into nutrients (source of carbon and inorganic salts) upon photocatalytic process, favoring algae growth. In contrast, SMX and its photoproducts inhibited the feeding of the brine shrimp Artemia salina, demonstrating that although the toxic effect of transformation products was significantly decreased, it was not completely eliminated. High rates of SMX decomposition and mineralization contributed to negligible toxicity towards Daphnia magna and Escherichia coli in the study carried out by Kim and Kan [81]. Researchers utilized UV-C/biochar/TiO 2 system for the photocatalytic degradation of 10 mgL −1 SMX and thus, obtained nontoxic products (nitrates, sulfates and some organic acids) after 6 h of reaction. However, when nitrate and bicarbonate were added into reaction solution, toxicity was highly increased up to 100%, suggesting the cautious application of the given process to the treatment of natural water bodies.
Ioannidou et al. [74] reported a small decrease in antimicrobial activity after photocatalytic treatment with the WO 3 -doped TiO 2 P25 catalyst, attributed to the slow mineralization of by-products after 180 min of simulated solar irradiation. Ecotoxic effects of SMX and its transformation products in that case were assessed towards bacterial strains: Escherichia coli and Enterococcus faecalis. During the oxidation process, toxicity first increased from a 50% population reduction of both E. coli and E. faecalis (20 mgL −1 initial SMX solution) to 64% and 57%, but eventually decreased to 45% and 37%, respectively. Murgolo et al. [131] also observed only a partial decrease in toxicity when a synthesized TiO 2 catalyst was utilized under UV-C irradiation for the photocatalytic degradation of a mixture of pharmaceuticals containing TMP. Acute toxicity tests were performed with green alga Selenastrum capricornutum, embryo of fish Danio rerio (FET) and AMES fluctuation test.
However, Nasuhoglu et al. [85] observed the formation of more toxic products towards Daphnia magna than initial 60 mgL −1 SMX solution when the target antibiotic was treated with UV-C/TiO 2 P25 for 13 h. Sirtori et al. [125] also reported the increased inhibition of Vibrio fischeri by TMP transformation products compared to parent compound (20 mgL −1 TMP), both in demineralized water and seawater, when the photocatalytic process with TiO 2 P25 under solar irradiation was carried out. In addition, the toxicity of CIP solutions treated by UV-A/TiO 2 P25 increased towards Vibrio fischeri compared to the original compound in the work of Silva et al. [132]. Although toxicity decreased after 15 min of the photocatalytic process, CIP solutions became very toxic (70% of luminescence inhibition) after 45 min of treatment due to higher quantities of transformation products being generated. An interesting observation was made by Yang et al. [105]: the antimicrobial activity of SMX transformation products against bacteria Vibrio fischeri increased compared to the parent compound (50 mgL −1 SMX) when photocatalytic tests were carried out in distilled water (UV-A/TiO 2 P25 system, 510 min). However, no significant toxicity was observed when tests were performed in different media: seawater with 20‰ and 34‰ salinity. Explanation for such findings might be in decreased photocatalytic decomposition of antibiotic when salinity was increased. Thus, smaller quantity of transformation products resulted in decreased biotoxicity. Xing et al. [133] utilized Toxicity Estimation Software Tool (T.E.S.T.) software to assess the toxicity of initial CIP solution and its photoproducts. Although most of the transformation products were reported to be less toxic than untreated antibiotic, some degradation products were found to be more toxic.
The photocatalytic performance of various TiO 2 and TiO 2 -based materials in degradation processes of sulfamethoxazole, trimethoprim and ciprofloxacin, along with photocatalysts characterization and process description, is presented in Table 4.

Challenges and Future Research Needs
Among the different advanced oxidation processes, heterogeneous photocatalysis over TiO 2 -based materials is one of the most efficient techniques for the removal of antibiotics. However, practical application on a large scale remains a challenge. This process has great potential, but it needs to be improved and optimized to be implemented in industry. The most significant issues are outlined as follows: • Huge variety of TiO 2 -based materials with unique features have been synthesized, as well as many different techniques having been reported to synthesize efficient TiO 2 -based photocatalysts. However, as the operational cost is crucial for practical applications, there is still a strong need for simple and cost-effective synthesis and modification processes to decrease the cost of the photocatalytic process. This can be achieved by implementing low-cost synthesis methods or by using low-cost materials, or both. The introduction of multiple cheaper modificators might be more costeffective than introducing one expensive dopant, while retaining high photocatalytic activity. For example, low-cost alternatives to photocatalysts modified with expensive and scarce noble metals could be TiO 2 -doped with cocatalysts like Ni and Cu [154,155] or Co and Ni [156]; Mo-doped TiO 2 [157], graphene-doped TiO 2 [158], Cu nanowires decorated with TiO 2 [159]. • Special emphasis should be placed on the development of TiO 2 -based photocatalysts active under solar light. Considering that high energy costs are associated with the usage of visible and ultraviolet light sources, solar photocatalysis is not only a costefficient solution, but also a sustainable one. Besides, further cost reduction can be achieved through the development of highly stable photocatalysts that can be easily separated from aqueous solution (immobilized/supported catalysts) and successfully reused over multiple cycles; • Further research is needed in studying the potential of photocatalytic process over TiO 2 -based materials for the successful removal not only of antibiotics, but also of ARB and ARGs from wastewaters; • In order to fulfill industrial application, photocatalytic tests should be carried out on real wastewater samples. The composition of wastewater is complicated: it contains a mixture of different pharmaceuticals and other organic pollutants, as well as inorganic substances. This decreases the degradation efficiency of antibiotics compared to the degradation efficiency of a single antibiotic in purified water. Moreover, there is a need to design and develop pilot plant installations that work under solar irradiation and in the continuous mode of operation. These are essential requirements for industrial treatment systems; • Further investigations should be performed on the efficiency of combined methods. Heterogeneous photocatalysis combined with conventional treatment methods, for example with biological treatment, should be studied in order to develop a technically feasible and cost-effective solution for wastewater treatment; • Assessment of eco-toxicity should be an essential part of degradation tests over TiO 2based photocatalysts. The complete mineralization of antibiotics remains a challenge, and incomplete mineralization leads to the formation of various intermediates and byproducts, sometimes more toxic than the parent compound. Degradation mechanisms require deeper studies.

Conclusions
The current review focuses on the recent progress in the photocatalytic removal of antibiotics over TiO 2 and TiO 2 -based materials. Heterogeneous photocatalysis has shown huge potential as an efficient, cost-effective and energy-efficient advanced oxidation process that is able to utilize the green and sustainable source of light -solar irradiation. The global occurrence of antibiotics in the environment, their fate and harmful effects on ecosystems and human health have been covered by the present review. Special emphasis is given to sulfamethoxazole, trimethoprim and ciprofloxacin as the most commonly detected antibiotics in different aquatic systems, raising significant concerns. The huge need for antibiotics removal from contaminated waters has been highlighted. Since conventional wastewater treatment fails to remove such complicated pharmaceuticals pollutants, various antibiotic removal technologies have been briefly discussed. The fundamentals of photocatalytic process over TiO 2 have been presented, along with the modification methods of TiO 2 to overcome its limitations and enhance efficiency. The photocatalytic performance of various TiO 2 and TiO 2 -based materials and the effect of different process parameters (amounts of pollutant and photocatalyst, pH, light, reaction media, presence of inorganic ions, natural organic matter and oxidants) have been the main focus of this review. Summarizing recent progress in the photocatalytic removal of the three widespread antibiotics, heterogeneous photocatalysis over TiO 2 -based materials has demonstrated its high potential for industry applications. We believe that the present review will contribute to the further progress in the design and development of efficient and cost-effective TiO 2 -based photocatalytic systems to meet the outlined challenges of large-scale application.