Solar Light-Induced Photocatalytic Degradation of Sulfamethoxazole by Cobalt Phosphide-Promoted Bismuth Vanadate

Abstract

The pursuit of low-cost, high-efficiency co-catalysts that are free of noble metals has become an area of considerable interest in the field of photocatalysis over the past few years. In this work, a series of cobalt phosphide (CoP 0.125–1.00 wt.%)-promoted bismuth vanadate (BiVO₄) photocatalysts was synthesized and physicochemical characterized by means of X-Ray diffraction, nitrogen isotherm absorption diffuse-reflectance spectroscopy, and high-resolution transmission electron microscopy. The efficiency of the as prepared photocatalytic materials was investigated for sulfamethoxazole (SMX) destruction in ultrapure water under simulated solar light irradiation. Results showed that the deposition of small amounts (0.50 wt.%) of CoP on BiVO₄ enhances SMX degradation. Moreover, SMX removal increased by increasing 0.50 CoP/BiVO₄ loading (up to 1 g/L) and decreasing SMX loading (1000–250 μg/L). Further tests were carried out in real and synthetic matrices, such as wastewater secondary effluent and bottled water, revealing the existence of hindering effects on SMX removal. The efficiency of 0.50 CoP/BiVO₄ photocatalyst was further investigated in a pilot plant configuration where the examined system was able to remove >99% of 300 μg/L SMX in deionized water utilizing 80 kJ/L of solar irradiation.

1. Introduction

Over the last few decades, scientists have drawn attention to the imminent danger linked with the identification of a great number of complex compounds belonging to pharmaceuticals, cosmetics, pesticides, and so on in groundwater, surface water, and wastewater treatment plants [1]. The substances under consideration are classified as Emerging Contaminants (ECs) due to their known or suspected ability to elicit adverse ecological and human health outcomes [2]. Specific pharmaceuticals, including antibiotics, analgesics, steroids, and others, have been detected in secondary effluents, surface, and groundwater in very low concentrations of ng/L to low μg/L range. They tend to bioaccumulate, resulting in significant ecosystem and human health consequences, such as developing antibiotic-resistant genes in soil bacteria [1,2]. Conventional treatment schemes (chemical and biological treatments, membrane filtration, and adsorption) have not shown very good efficiency towards their elimination from wastewater and water bodies. As a result, new technologies have arisen for efficient wastewater remediation. Photocatalysis has emerged as a green, sustainable, and promising method for ECs-rich wastewater treatment. Photocatalytic systems involve the photoexcitation of the photocatalyst, usually a semiconductor, resulting in the generation of electrons and holes pairs that can initiate degradation reactions for the complete elimination of ECs in aqueous media.

Based on this technology, many groups have reported efficient pharmaceuticals degradation in environmentally relevant matrices. For example, Jin et al. examined the degradation of norfloxacin using N-doped TiO₂, paying particular attention to the toxicity of intermediates produced during the treatment, providing beneficial results concerning the treatment of antibiotics-contaminated water [3]. Tran et al. studied metronidazole degradation in a UV/TiO₂ photocatalytic system, pointing out the significant impact of aqueous matrix on the observed performance [4]. In addition, sulfamethoxazole, another agent from the group of antibiotics, was chosen as the target compound by Tomara et al. in different environmental samples with Ag₂O photocatalyst [5]. Sulfamethoxazole removal was also studied by different combinations of ZnWO₄ and carbon nitride nanocomposites showing high efficiency and great stability [6].

As evidenced by the aforementioned studies, many semiconductors have been examined as potential photocatalytic materials in wastewater treatment [7]. Among them, bismuth vanadate (BiVO₄) is among the most studied materials in photocatalytic water remediation [8,9]. Due to several key factors, Monoclinic BiVO₄ is considered a promising candidate for photocatalytic applications. These include its narrow band gap of 2.4–2.8 eV, which facilitates visible light activation, its non-toxic nature, high chemical stability, and low cost of synthesis. Despite its many favorable characteristics, the observed efficiency of BiVO₄ remains relatively low, primarily due to the high recombination rate of the photoproduced carriers. One way to overcome this problem is the addition of a proper co-catalyst [10]. This approach is widely recognized as one of the most effective means of enhancing the activity of photocatalysts [11]. The role of co-catalysts is to “pump” electrons away from the photocatalyst, thus preventing the spontaneous reaction of photoproduced electrons and holes. The activity of the co-catalysts depends not only on its type and characteristics but also on the quantity deposited on the photocatalyst surface. In general, it is widely acknowledged that a critical threshold exists beyond which co-catalyst particles serve as electron-hole recombination centers, ultimately leading to a reduction in the overall rate of the photocatalytic reaction. Metal co-catalysts have been observed to enhance the photocatalytic activity of organic pollutant degradation via two primary mechanisms: firstly, by facilitating the electron transfer to dissolved oxygen molecules, and secondly, by enhancing the catalytic rate of various dark reactions [12].

Although co-catalysts based on noble metals such as Pt and Au have demonstrated high levels of efficiency, their prohibitive cost and limited availability render them unsuitable for practical applications [13]. Consequently, developing co-catalysts based on earth-abundant elements represents an appealing strategy for producing a cost-competitive photocatalyst.

Metal phosphides are solid-state compounds with a general formula, M_xP_y, produced by the combination of phosphorus with metallic or semi-metallic elements. They can be categorized depending on the item that is in excess, to metal-rich (x > y in M_xP_y), stoichiometric (x = y = 1 in M_xP_y), or phosphorus-rich transition metal phosphides (y > x in M_xP_y) [14]. Metal phosphides falling into the first two categories are frequently semiconducting and, occasionally, even metallic or superconducting, owing to the significant metal–metal bonding within them. The M–P bonds in transition metal phosphides are relatively robust and provide high thermal stability and hardness. In addition, they are resistant to chemicals and oxidation. Conversely, phosphorus-rich transition metal phosphides display notable phosphorus–phosphorus bonding, as phosphorus can form diverse oligomers and clusters by bonding with itself. Consequently, numerous phosphorus-rich phosphides are thermally unstable and are prone to be disproportionate at elevated temperatures, resulting in the formation of elemental phosphorus and a more metal-rich phase [14]. Transition metal phosphides are utilized to catalyze the hydrodesulfurization process and have demonstrated considerable activity in the electrocatalytic water reduction. Recently, Chen et al. reported the production of CoP nanoparticles as a co-catalyst for enhancing the efficiency of porous ZnSnO₃ in the removal of tetracycline under visible light. The researchers observed that the 2% CoP/ZnSnO₃ composite displayed outstanding photocatalytic efficiency (96.4%, 1 h), 3.38 times greater than that of pure ZnSnO₃ [15]. Moreover, Shi et al. fabricated Cu₃P/BiOCl and tested their photocatalytic activity to decompose different antibiotics (including tetracycline, oxytetracycline, and ofloxacin). Their investigation illustrated that Cu₃P is an effective co-catalyst in facilitating the oxidation process of persistent pollutants [16].

In the present study, cobalt phosphide (CoP) was deposited on a BiVO₄ surface and it was for the first time incorporated in a photocatalytic system for pharmaceuticals degradation in water. Conducting a systematic study of the physicochemical properties along with the observed activity of the synthesized materials is anticipated to provide substantial insights regarding the utilization of metal phosphides as co-catalysts in photocatalytic oxidation. A series of CoP/BiVO₄ samples of various CoP loadings (0.12–1.00 wt.%) were synthesized and physicochemical characterized. The activity of the as-prepared composites under simulated solar irradiation was examined for the oxidation of sulfamethoxazole (SMX), a representative antibiotic often detected in environmental samples. SMX has been detected in raw wastewater [17], surface water, and groundwater [18], as well as wastewater treatment effluents [19] and even in drinking water resources [20]. The concentration of SMX in various water matrices is usually in the 226 to 3000 ng/L range, and elevated levels are often reported in hospital wastewater [21]. Although potential adverse effects, which stem from direct exposure to trace levels of pharmaceuticals, are considered unlikely in humans, the effects of long-term, low-level exposure to antibiotics might be hazardous for sensitive subpopulations (e.g., children and pregnant women). Moreover, the presence of antibiotics in the natural environment is a threat to aquatic and soil microorganisms [22]. Another concern associated with the dissemination of antibiotics in the environment is the progression of antibiotic resistance [23].

The effect of operating conditions such as catalyst and SMX loading was studied using a batch photocatalytic reactor. Moreover, to acquire significant information concerning the suitability of CoP/BiVO₄ photocatalysts in practical applications, further experiments were conducted using the best performing 0.50 CoP/BVO sample in a stainless steel (SS) flat plate reflector (FPR) photo-reactor under solar radiation. To the best of our knowledge, this is the first study to report on the data obtained from a pilot plant scale employing CoP/BiVO₄.

2. Materials and Methods

2.1. Chemicals and Matrices

Ammonium vanadium oxide (NH₄VO_3, 99% metal basis, CAS: 7803-55-6), cobalt nitrate hexahydrate (Co(NO₃)₂∙6H₂O, 99.999% trace metals basis, CAS: 10026-22-9), and sodium hypophosphite monohydrate (NaH₂PO₂∙H₂O, ≥99%, CAS: 10039-56-2) was purchased by Alfa Aesar. Tri-sodium citrate dihydrate (Na₃C₆H₅O₇∙2H₂O, CAS: 6132-04-3) and sodium hydroxide (NaOH, CAS: 1310-73-2) were purchased from Sigma Aldrich. Bismuth (III) nitrate pentahydrate (Bi(NO₃)₃∙5H₂O, 99.999% trace metals basis, CAS: 10035-06-0), urea powder (NH₂CONH_2, CAS: 57-13-6), and nitric acid (nitric acid 65%, CAS: 7697-37-2) were purchased from Sigma Aldrich. Sulfamethoxazole (SMX, C₁₀H₁₁N₃O₃S) was also supplied by Sigma Aldrich. N,N-Dimethylformamide (DMF) (HCON(CH₃)₂, CAS: 68-12-2), n-hexane (CH₃(CH₂)₄CH₃, CAS: 110-54-3), ethanol absolute (EtOH) (CH₃CH₂OH, CAS: 64-17-5), and toluene (C₆H₅CH₃, CAS: 108-88-3) were supplied from Sigma Aldrich.

Ultrapure water (UPW, pH~6), was provided by a Millipore Milli-Q laboratory system. The main characteristics of bottled water (BW) were pH = 7.5, 0.4 mS/cm conductivity, 10 mg/L chloride, 17 mg/L sulfate, 215 mg/L bicarbonate, 7 mg/L nitrate, and 75 mg/L. The University of Patras, campus treatment plant, provided us with secondary effluent before chlorination (WW, COD = 25 mg/L, pH = 8.1).

Humic acid (HA, CAS:1415-93-6) was also purchased from Sigma-Aldrich (Merck KGaA, Darmstadt, Germany and/or its affiliates).

2.2. Photocatalyst Preparation

2.2.1. Cobalt Phosphide (CoP)

Cobalt phosphide was synthesized as follows [24]: A pre-weighted amount of Co(NO₃)₂∙6H₂O was added to an aqueous solution containing Na₃C₆H₅O₇∙2H₂O. An appropriate amount of 0.5 M NaOH solution was added dropwise until Co(OH)₂ precipitation occurred. Finally, Co(OH)₂ was filtered, washed with water and ethanol, and dried at 60 °C. The as-prepared Co(OH)₂ and excess quantity of NaH₂PO₂∙H₂O calcined under argon (Ar) atmosphere at 300 °C for 2 h.

2.2.2. Bismuth Vanadate (BiVO₄)

Bismuth vanadate was synthesized using the following procedure: [10]: Aqueous solutions of Bi(NO₃)₃·5H₂O (0.4 M) and NH₄VO₃ (0.2 M) containing HNO₃ (2 M) were prepared separately. After the Bi(NO₃)₃·5H₂O (100 mL) and NH₄VO₃ (200 mL), solutions were mixed together, 7.5 g of urea was added, and the final solution was kept under stirring for 4 h at 90 °C. The BiVO₄ powder with a characteristic yellow color formed after the hydrolysis procedure was filtered off, washed with water and ethanol, and dried in an oven at 70 °C for 12 h. The final product was ~13 g. Details are available in a former study of our group [25].

2.2.3. Cobalt Phosphide Promoted Bismuth Vanadate (CoP/BiVO₄)

A solution-phase method was employed to incorporate CoP nanoparticles onto BiVO₄ by means of mixing two immiscible solutions (CoP in hexane and BiVO₄ in DMF) via constant sonication [26]: one hundred and sixty mg of BiVO₄ was added in 20 mL of DMF (S1) while appropriate amounts of CoP (0.20 mg, 0.40 mg, 0.80 mg, and 1.6 mg in case of 0.12, 0.25, 0.50, and 1.00 wt.%) was dispersed in 20 mL of hexane (S2). Both solutions were sonicated for 30 min. After this period, S2 was added in S1 and the final solution remained under sonication for 60 min to ensure the proper dispersion of the BiVO₄/CoP in DMF/hexane mixture. After that, 10 mL EtOH was added for further dispersion and homogeneity of the as prepared material, and sonication continued for 20 min. The product was separated from this mixture under vacuum filtration, dispersed in 35 mL of toluene to improve the interfacial contact between the two semiconductors, and remained under sonication for 60 min. Finally, the suspension was aged for 24 h. Thereafter, by carefully removing and discarding the supernatant, the resulting precipitate was separated by centrifugation and washed with excessive ethanol for three times before it was placed in a vacuum oven at 120 °C for 8 h to remove the residual solvents prior to the materials characterization. The photocatalysts thus prepared are denoted in the following as x CoP/BVO, where x = 0.12, 0.25, 0.50 or 1.00 denotes the wt.% cobalt phosphide loading wt.%.

2.3. Photocatalyst Characterization

Photocatalysts were characterized considering: (a) Their specific surface area (SSA) with the Brunauer–Emmett–Teller (BET) method employing nitrogen physisorption at the temperature of liquid nitrogen (77 K) nitrogen physisorption, using a Micromeritics Gemini III (2375, Norcross, GA, USA). Prior to each measurement, the sample was outgassed under a dynamic vacuum at 250 °C for 2 h. (b) Their crystallographic structure by X-ray diffraction (XRD) was also considered. Powder X-ray diffraction patterns were obtained using a Brucker D8 Advance instrument (Billerica, MA, USA) equipped with a Cu Kα source operated at 40 kV and 40 mA. Data were collected in the 2θ range of 2° to 85° at a scan rate of 0.05° s⁻¹ and a step size of 0.015°. Phase identification was based on JCPDS cards. (c) Their optical properties by UV-Vis diffuse reflectance spectroscopy (DRS) was also considered. The diffuse reflectance spectra were recorded on a UV-vis spectrophotometer (Varian Cary 3, (3E Palo Alto, CA, USA)) equipped with an integrating sphere, using BaSO₄ as a reference. The catalyst powder was loaded into a quartz cell, and the spectrum was obtained at room temperature in the wavelength range of 200–800 nm. The DR measurements were converted into the equivalent absorption coefficient by applying the transformation based on the Kubelka–Munk function:

$$\mathit{F}\left({\mathit{R}}_{\mathit{\infty }}\right)=\frac{\mathit{K}\left(\mathit{\lambda }\right)}{\mathit{S}\left(\mathit{\lambda }\right)}=\frac{{\left(1-{\mathit{R}}_{\mathit{\infty }}\right)}^2}{2{\mathit{R}}_{\mathit{\infty }}}$$

(1)

where K and S are the absorption and scattering coefficients, respectively, and $R_{\infty }=R/R_{\mathrm{ref}}$ is the reflectance. The optical band gaps of the semiconductors were evaluated based on the following expression.

$${\left(\mathit{\alpha }\mathit{h}\mathit{\nu }\right)}^{1/\mathit{n}}=\mathit{B}\left(\mathit{h}\mathit{\nu }-{\mathit{E}}_{\mathrm{bg}}\right)$$

(2)

where a is the absorption coefficient, hν is the incident photon energy, E_bg is the band gap energy, B is a constant related to the effective masses of charge carriers associated with valance and conduction bands, and n is a factor that depends on the kind of optical transition induced by photon absorption. Band gap energies (absorption thresholds) were estimated assuming that F(R) values are proportional to the optical absorption coefficients and that the synthesized materials are indirect semiconductors for which n = 2. Thus, the values of E_bg were obtained from the plot of [F(R)hν]^1/2 versus hν (Tauc plot) in the region of high absorption and the extrapolation of the linear region to the horizontal axis at zero F(R). (d) Their morphology by high-resolution transmission electron microscopy (HR-TEM) and scanning electron microscopy (SEM) was also considered. The morphological characteristics of the synthesized samples were examined using transmission electron microscopy on a JEOL JEM-2100 system operated at 200 kV (point resolution 0.23 nm). TEM Images were recorded employing an Erlangshen CCD Camera (Gatan Model 782 ES500 W), while films (Kodak SO-163) were used for HRTEM images. The specimens were prepared by dispersion in water and spread onto a carbon-coated copper grid (200 mesh). (e) SEM images were obtained using a JEOL 6300 scanning electron microscope equipped (Peabody, MA, USA) with an Energy Dispersive Spectrometer (EDS) to determine the chemical composition and element distribution of the samples. Details can be found elsewhere [27].

2.4. Photocatalytic Tests

2.4.1. Batch Photocatalytic Reactor

The degradation experiments of sulfamethoxazole were conducted using a cylindrical glass photocatalytic reactor (120 mL). At first, an aqueous solution of known SMX concentration (typically 500 μg/L) was added to the photoreactor, followed by the photocatalyst. The suspension was left under stirring for 15 min in the dark. An Oriel LCS-100 solar simulator connected with a 100 W xenon lamp was utilized as the radiation source, which had an estimated intensity of 7.3 × 10⁻⁷ Einstein/(L.s) as determined by chemical actinometry [28]. For the experiments conducted under visible light, irradiation was reduced using a 420 nm cut-off filter. Samples were taken at a predetermined time from the irradiated suspension and filtered (0.2 μm). Samples were then analyzed employing liquid chromatography. In all cases, experiments were repeated at least twice, and mean values are quoted as results. It is worth noting that the variance between measurements never exceeded 5%.

2.4.2. Pilot-Scale Demonstration of the Process

Although optimizing the system was not the primary objective of this study, pilot-scale photocatalytic oxidation experiments were conducted on the optimal synthesized catalyst to examine the system performance under more representative conditions. For this purpose, a stainless steel flat plate reflector (SS-FPR) was employed, consisting of seven borosilicate tubes mounted on a fixed platform tilted at an angle of 45°, along with one tank and one centrifugal pump. A characteristic photo can be found in Figure S1. The reactor was located at the University of Patras, Greece.

Each tube’s outer diameter equals 32 mm; the length was 1.5 m. The flow rate was 12 L/min, the total volume was V_T = 17 L, the total irradiated surface ≈1.1 m², and the irradiated volume (Vi) ≈ 7.9 L. SMX solution (300 μg/L) containing 125 mg/L 0.50CoP/BVO was mixed in the feeding tank using air sparging and the flow provided by the magnetic centrifugal pump. Considering the relationship between solar radiation and treatment time, the variable Q_UV,n was utilized instead of the actual treatment time [29].

2.5. SMX Measurement

SMX was estimated using a Waters Alliance 2695 HPLC system, as reported in previous studies of our group [30,31].

3. Results

3.1. Catalyst Characterization

The XRD patterns of CoP/BVO samples are shown in Figure 1. In the same Figure, data concerning pure BiVO₄ are also included. It is observed that all diffraction peaks of promoted materials are indexed to monoclinic BiVO₄ (JCPDS No. 14-0688) with no additional peaks of CoP [32]. This can be attributed to the very low content of CoP nanoparticles and their low crystallinity. Moreover, CoP/BVO are characterized by low SSA < 2 m²/g, typical for BiVO₄ [33]. The specific surface area was not affected by CoP addition. This was expected, considering both the low content of CoP and the equally low specific surface area of CoP (<1 m²/g). Figure 2A presents the DRS spectra of CoP/BVO and pure BiVO₄. It is observed that they are all characterized by an absorption threshold of ca 520 nm in accordance with previously published information for BiVO₄. However, adding CoP marginally enhances the absorption at higher wavelengths [34,35]. The optical band gaps of the samples were estimated using the Tauc method and were found to be equal to ~2.3 eV as shown in Figure 2B. Both sample structure and morphology were further investigated using HRTEM (Figure 3a,b). The lattice fringe distance of 0.246 nm assigned to the (111) plane of CoP is clearly discerned [15] in intimate contact with BiVO₄ with the characteristic interplanar 0.47 nm spacing of the (110) plane. Figure 3c shows a representative SEM image with EDS mapping of Co and P for the sample with the highest CoP loading (1.00 wt.% CoP). It is observed that both cobalt and phosphorus are homogeneously dispersed on BiVO₄ surface. HRTEM and EDS/SEM analyses confirm the successful incorporation of CoP on BiVO_4′s surface, suggesting that CoP nanoparticles are closely attached to BiVO₄, favoring the electron transfer between the two semiconductors.

3.2. Photocatalytic Efficiency

The performance of CoP/BVO catalysts for sulfamethoxazole (SMX) elimination in ultrapure water (UPW) under simulated solar radiation was studied, and the obtained data are depicted in Figure 4A. The findings suggest that introducing a minor proportion of cobalt phosphide (<0.25 wt.%) onto the surface of BiVO₄ leads to reduced destruction times for SMX. The optimal results were achieved with the 0.5 CoP/BVO sample.

In contrast, higher CoP loading (1.00 wt.%) results in lower photocatalytic activity. The observed phenomenon can be rationalized by the excess presence of metal phosphides on the surface of BiVO₄, particularly at higher loadings, leading to the formation of aggregates. These aggregates can mask BiVO₄ active sites, thereby acting as photogenerated charges recombination locations [36]. The results depicted in the same figure also include data collected in the absence of irradiation (indicated by open symbols) to assess the influence of adsorption phenomena on SMX removal. As demonstrated, adsorption accounted for less than 8% in all the examined samples.

The decomposition of SMX can be accurately represented by pseudo-first-order kinetics (Equation (3)) with linear regression (R²) values exceeding 0.99.

$$\frac{dC}{dt}={-k}_{app}C\to \mathit{ln}\left(\frac{C}{C_o}\right)={-k}_{app}t\to \frac{C}{C_o}=e^{{-k}_{app}t}$$

(3)

where C and C₀ is SMX concentration (mg/L) at time t = t and t = 0, respectively, and k_app (min⁻¹) is an apparent rate constant incorporating the relatively constant concentration of oxidizing species.

The observed apparent constants (k_app in min⁻¹) can be estimated using the linearized form of Equation (1), and results obtained using data from Figure 4A, are shown in Figure 4B. Evidently, the observable kinetic constant increases from 0.0038 min⁻¹ for pure BiVO₄ (BVO) to 0.0045 min⁻¹ for 0.25 CoP/BVO and maximizes to 0.0080 min⁻¹ for the 0.50 CoP/BVO sample.

The improved photocatalytic efficacy of CoP-facilitated BiVO₄ photocatalytic materials can be elucidated by examining the relative positions of the conduction band (CB) and valence band (VB) of both semiconductors, along with their energy band gap (E_bg) and Fermi level (E_F), as depicted in Scheme 1. Irradiation of the semiconductors results in the generation of photo-produced charge carriers [37]. The close proximity of the semiconductors facilitates the creation of p-n heterojunctions at their interfaces, generating an internal electric field that equilibrates their Fermi level (E_F) according to thermodynamic constraints [38,39]. As shown in Scheme 1, photogenerated electrons (e⁻) end up in BiVO₄ CB, whereas photogenerated holes (h⁺) are transferred in CoP VB, preventing their recombination. SMX degradation can now take place directly, by reacting with holes, while electrons may react with adsorbed O₂, thus forming secondary reactive species for SMX degradation.

3.3. Impact of Catalyst Loading

The impact of the catalyst loading on SMX removal was studied by varying the loading of 0.50 CοP/BVO from 0 to 1000 mg/L, and the results are illustrated in Figure 5. Increasing photocatalyst loading in the suspension was found to accelerate SMX removal. In specific, 77%, 80%, and 89% SMX removal are achieved after 180 min of irradiation using 500, 750 και 1000 mg/L 0.50 CοP/BVO correspondingly. In the inset of Figure 5, k_app for each catalyst concentration is presented. Increasing 0.50 CoP/BVO concentration is reported to raise the value of the apparent kinetic constant, with k_app having the greatest value 0.0116 min^−1, for the highest catalyst loading.

Furthermore, the reusability of 0.50 CoP/BVO was examined. Reusability tests were carried out according to the following procedure: 1000 mg/L of 0.50 CoP/BVO was added to 120 mL of ultrapure water containing 500 μg/L of SMX, and the mixture was irradiated for 180 min. At that time, SMX conversion was measured, and the photocatalyst was recovered by vacuum filtration. After that, the photocatalyst obtained was added again to 120 mL of ultrapure water containing 500 μg/L of SMX, and the mixture was irradiated for another 180 min. This cycle was repeated twice. It was observed that SMX removal reduced to ~90% and 85% after the second and third cycles, respectively, showing the high stability of the prepared photocatalyst.

3.4. Impact of SMX Initial Loading

Further tests were undertaken to determine the impact of the initial SMX loading on photo process efficiency. Specifically, additional evaluations were conducted at 250, 500, and 1000 μg/L SMX in the presence of 1 g/L 0.50 CoP/BVO (Figure 6A).

Interestingly, for the lowest SMX concentration studied (250 μg/L), the time required for its complete degradation is only 180 min. Increasing the concentration to 500 μg/L has practically no impact on its degradation. However, in the case of the highest SMX concentration studied (1000 μg/L), only 60% SMX removal was observed in the same period. The corresponding k_app were computed (Figure 6B). It is evident that for the lowest SMX loading, k_app remained unchanged (0.0116 min⁻¹), implying that degradation kinetics was first order. In contrast, for the highest SMX values, k_app declined to 0.0061 min^−1, showing that kinetics changed to zero, thus confirming the hypothesis of pseudo-first-order kinetics used in the present study.

3.5. Impact of Matrix on SMX Removal

To determine the impact of the aqueous matrix on the efficacy of the current system, another group of tests was carried out utilizing WW and BW as alternatives to UPW. In addition, experiments were performed with the addition of HA, which imitates the naturally occurring organic matter in environmental samples in UPW. Results in the form of kapp are depicted in Figure 7. It is noticed that k_app decreases from 0.0116 min⁻¹ in the case of UPW to 0.0017 min⁻¹ in BW and to 0.0008 min⁻¹ in WW. This behavior is typical in most photocatalytic systems due to scavenging phenomena deriving from inorganic and organic matter in aqueous solutions [40]. Indeed, data acquired using HA in UPW, showed that the organic matter probably reacts with the photogenerated species in the photocatalyst surface, thus lowering SMX removal [41,42].

However, it is observed that the k_app value in the case of HA is 0.0068 min⁻¹, significantly higher than those obtained in BW and WW (0.0017 min⁻¹, 0.0008 min⁻¹), thus implying that inorganic species, such as bicarbonates and chlorides present in both, significantly decrease SMX degradation.

3.6. Impact of Irradiation Type on SMX Decomposition

The performance of 0.50 CoP/BVO under visible light alone was evaluated using a suitable cut-off filter, and the findings are illustrated in Figure 8. As depicted in the Inset of Figure 8, approximately 40% removal of SMX is achieved within 180 min under visible light alone. In contrast, nearly complete SMX removal is accomplished within the same duration under simulated solar irradiation. Indeed, k_app under solar (0.0116 min⁻¹) is practically twice that of k_app under visible light (0.0056 min⁻¹) (Figure 8). These results confirm the efficiency of CoP/BVO photocatalyst under visible light, with the dominant role, however, belonging to the UVA part of the solar spectrum in accordance with DRS results of the as-prepared materials.

3.7. Pharmaceuticals Degradation by Other Photocatalytic Systems

The results presented in this work are in agreement with previous studies of our group investigating the photocatalytic degradation of SMX with (i) BiOCl photocatalysts, yielding complete removal after 90 min [43], and (ii) Cu₃P/BVO suspensions leading to SMX degradation after 120 min [30]. Moreover, in two recent studies, Zangeneh et al. [44] achieved complete degradation of 50 mg/L Metronidazole at pH = 4 and catalyst concentration equal to 1.5 g/L under 90 min of solar light irradiation, using TiO₂ doped photocatalytic materials, while Goh et al. [45] studied the photocatalytic degradation of Carbamazepine by HF-Free-Synthesized MIL-101(Cr)@Anatase TiO₂ composite under UV-A irradiation. They found that under optimal conditions, 99.9% degradation can be achieved after 60 min under UV-A irradiation. Carbamazepine photocatalytic degradation was also studied by Levakov et al. [46], using clay-based materials as photocatalysts. Their study showed that kinetics was significantly lower, with more extended periods required for carbamazepine degradation. One step ahead from the other systems, CoP/BVO photocatalytic activity was further demonstrated in a pilot plant, showing very promising results for the adoption of metal phosphides as co-catalysts in water remediation photocatalytic systems.

3.8. Pilot Plant Demonstration

A flat stainless steel reactor equipped with seven borosilicate glass tubes was selected. A study by Gomes et al. [29], highlighted the ease of construction and durability of such reactors made of reflective materials such as SS, which due to the high ratio of tubes/area can be economically competitive in specific conditions compared with more expensive and selective reflective materials, as well as more complex and efficient geometries such as CPC reactors. The pilot plant study focused on the decomposition of 300 μg/L of SMX antibiotic in deionized water using 125 mg/L 0.50 CoP/BVO synthesized photocatalyst and solar photolysis of the same amount of antibiotic without the catalyst in inherent pH. As illustrated in Figure 9, the solar photolysis of SMX could not eliminate SMX at 80 kJ/L, thereby confirming the results obtained from experiments conducted in the laboratory’s solar radiation simulator and SMX’s reported long half-life in the environment. However, when the 0.50 CoP/BVO photocatalytic material was added, >99% of 300 μg/L SMX was removed for absorbed energy equivalent to 80 kJ/L, demonstrating the material’s efficacy even on a semi-pilot scale using non-optimized conditions.

3.9. Limitations of the Proposed System

The results presented above provide only preliminary indications of the potential of the proposed photocatalytic system. Thus, it is suggested that further research should concentrate on evaluating the system’s performance in pilot-scale reactors under more realistic conditions, such as using real effluents, varying concentrations of pharmaceutical compounds, and the ability of the process to eliminate pathogenic microorganisms. Future work should point out the experimental conditions that will counterbalance the hindering phenomena observed in real water matrices, while photocatalytic disinfection experiments need to be carried out in various water matrices. At the same time, the potential formation of toxic by-products resulting from the oxidation of SMX and the constituents of water matrices should be investigated. Moreover, catalyst’s recovery and reuse after photocatalytic tests is another critical issue that is of great concern in pilot-scale systems and was out the scope of the present study.

4. Conclusions

To summarize, in the present study, CoP-promoted BiVO₄ photocatalytic materials were synthesized and examined for SMX destruction in water under simulated solar irradiation. It was found that deposition of 0.5 wt.% on BiVO_4′s surface enhances SMX removal. Specifically, the observable kinetic constant increased from 0.0038 min⁻¹ for pure BiVO₄ to 0.0080 min⁻¹ for the 0.50 CoP/BVO sample. This was mainly attributed to the effective separation of photo-produced carriers, thus proving the suitability of CoP nanoparticles as co-catalysts in photocatalytic water treatment technologies. In studies of this nature, the quantity of co-catalyst employed and the relative positions of the band edges of the semiconductors are critical factors that demand careful consideration. On the other hand, a further increase in CoP content led to lower SMX removal in the same period.

Varying the experimental parameters, SMX degradation was highly affected, with the best results obtained for the highest photocatalyst concentration (1000 mg/L) and the lowest SMX dosage (250 μg/L). Moreover, SMX degradation kinetics decreased significantly in real environmental samples such as BW and WW due to the presence of both inorganic and organic matter in them. Kinetic constants decreased from 0.0116 min⁻¹ in the case of UPW to 0.0017 min⁻¹ in BW and to 0.0008 min⁻¹ in WW. In addition, k_app under solar (0.0116 min⁻¹) was twice that of k_app under visible light (0.0056 min⁻¹). Reusability tests showed that SMX removal only slightly reduced to ~90% and 85% after the second and third cycle, respectively, showing the high stability of the as-prepared photocatalyst. The experiments conducted on a pilot scale showed that solar photolysis of SMX was minor. However, with the use of 0.50 CoP/BVO more than 99% of 300 μg/L SMX was removed for absorbed energy equivalent to 80 kJ/L. The above results confirmed the moderate level of SMX photolysis and the synthesized photocatalyst’s ability to effectively remove SMX from deionized water.

Although the results are promising, future work in different directions is needed. In particular, future research needs to be directed towards (i) the evaluation of the system’s performance in pilot-scale reactors under more realistic conditions, (ii) disinfection experiments and toxicity estimation, (iii) and catalyst recovery after photocatalytic tests on the pilot scale and reuse.