Synthesis and characterization of rGO/Fe⁰/Fe₃O₄/TiO₂ nanocomposite and application of photocatalytic process in the decomposition of penicillin G from aqueous

Abstract

In this study, we synthesized rGO/Fe⁰/Fe₃O₄/TiO₂ nanocomposite according to Hummer's, and straightforward sol-gel method. The FESEM, EDX, TEM, FT-IR, XRD, BET, UV spectra, and VSM analysis were applied to determine the catalyst properties. Optimization of influence parameters on photocatalytic process performance to penicillin G degradation in aqueous media. pH (4–8), nanocomposite dose (10–20 mg/L), reaction time (30–60 min), and penicillin G concentration (50–100 mg/L) were optimized via central composite design. In the optimum condition of PCP, supplementary studies were done. As a result of the analysis, the nanocomposite was well synthesized and displayed superior photocatalytic properties for degrading organic pollutants. In addition to being magnetically separable, the synthesized rGO/Fe⁰/Fe₃O₄/TiO₂ nanocomposite exhibits high recyclability up to 5 times. The quadratic model of optimization is based on the adjusted R²(0.99), and predicated R²(0.97) suggested. According to the analysis of variance test, the model was significant (F-Value = 162.95, P-Value = 0.0001). Photocatalytic process is most efficiently decomposed at pH = 6.5, catalyst dose = 18.5 mg/L, reaction time = 59.1 min, and penicillin G concentration = 52 mg/L (efficiency = 96%). The chemical oxygen demand and total organic carbon decrease were 78, and 65%. The photolysis and adsorption mechanism as a single mechanism had lower performance in penicillin G degradation. Benzocaine had the greatest effect on reducing the efficiency of the process as a radical scavenger. The °OH, h^+, and O₂^●– were the main reactive oxidant species in penicillin G removal. Phenoxyacetaldehyde, Acetanilide, Diacetamate, Phenylalanylglycine, N-Acetyl-l-phenylalanine, Diformyldapsone, and Succisulfone were the main intermediates in penicillin G degradation. The results indicated the photocatalytic process with rGO/Fe⁰/Fe₃O₄/TiO₂ nanocomposite on a laboratory scale has good efficiency in removing penicillin G antibiotic. The application of real media requires further studies.

Graphical abstract

1. Introduction

Due to its resistant organic compounds (ROCs), high use, and sustainability, industrial wastewater, such as pharmaceutical wastewater, has been considered a major water pollutant in recent decades [1,2]. Predictably, significant levels of antibiotics (as parent compounds or metabolized compounds) will enter the environment from pharmaceutical, hospital, and animal husbandry wastewater since antibiotics are widely used in medicine, veterinary medicine, and aquaculture [3]. Because, Antibiotics cause human pathogenic bacteria, the development of metabolites and hazardous residues have received special attention from researchers. Some reports indicate that worldwide 10⁵ to 20⁶ tons of antibiotics are consumed each year [1,4,5]. There is no difference in the solubility order among antibiotics; water is more soluble than ethanol, and ethanol is more soluble than 2-propanol and acetone. In comparison to acetone, water has a three-orders-of-magnitude higher solubility [6]. One of the popular categories of antibiotics is the penicillin family. The structure of this family contains a lactam ring. Penicillin G (C₁₆H₁₇N₂NaO₄S, PNG) is one of the members of the penicillin family that is produced from the fungus called penicillium [7,8]. PNG is used to treat a wide range of diseases in humans, animals, and aquatic organisms [9]. Its mechanism of action involves the inhibition of the production of peptidoglycans, which destroys the bacterial cell wall. It is soluble in water (pKa = 2.75) and can destroy the cell wall of bacteria. In some countries, more than 12,000 tons of this antibiotic are produced annually. It has the highest antimicrobial activity among natural antibiotics [10]. The concentration of PNG in the environment have been reported in the range of ng to mg [11]. To date, several processes have been applied to the treatment of wastewater containing PNG at the lab and pilot scale, such as adsorption [12], catalytic ozonation process [13], electro-oxidation process [14], electro-Fenton process [15], membrane bioreactor [16], and photocatalytic process (PCP) [7]. As one of the most advanced oxidation processes (AOPs), PCP utilizes light energy (UV–C, UV-A, and visible light) in conjunction with a photocatalyst that consists of metal oxides (semiconductors) to provide degradation or mineralization of organic pollutants. Nowadays, PCP has attracted attention because of its facile process, stability, and ability to decompose organic structures like an antibiotic to non-hazardous byproducts, and inorganic compounds like H₂O and CO₂ [17]. The main drawback of classic photocatalysts such as TiO₂, and ZnO is the large gap band, which makes only UV-C can activate them. In addition to the potential dangers for humans, UV-C contains only 4% of the sun's waves, which limits the application of the PCP via classic photocatalysts. Today, researchers are looking to use new photocatalysts with short-band gaps that can be activated in UV-A and visible light [18]. In the field of purification, PCP is a modern, green, and promising technique. The photocatalyst converts light energy into chemical energy on the surface of the catalyst. When the photon energy exceeds/equals the band gap of the photocatalyst, electrons in the valence band move into the conduction band, and holes are generated in the conduction band. As a result of the excited electron-hole pairs, free radicals are produced directly or indirectly, which can cause organic compounds to decompose into inorganic compounds [19]. In the PCP, carrier separation and transfer are estimated to take approximately 10⁻¹²-10⁻⁶ s, which is longer than the time required for the recombination of photogenerated electron-hole pairs, which take 10⁻¹²-10⁻³ s. Increased carrier transfer rates are therefore one of the most effective strategies for enhancing photocatalytic efficiency [20]. In contrast to ZnO, SnO₂, ZrO₂, CdS, and other semiconductor materials that suffer from their own losses, TiO₂ is a popular photocatalyst because of its non-toxicity, low cost, excellent chemical stability, and outstanding corrosion resistance [21]. Despite this, it is extremely difficult for photo-generated electron-hole pairs to be recombined on TiO₂'s surface when they are caused by light-excited electrons. Furthermore, TiO₂ is an inefficient photocatalyst due to a high aggregation rate, which limits its application to practical photocatalysis [22]. As a result, TiO₂ is usually modified with other materials in order to enhance its photocatalytic properties in order to overcome its lack of photocatalytic activity when used alone. The combination of graphene-based materials with TiO₂ appears to be a promising candidate since it can serve as a buffer against particle aggregation as well as facilitate the transfer of electron-hole pairs generated by photons [23]. As a result of its high surface area, low mass, high electrical conductivity, and a large number of functional groups, such as carboxyls, hydroxyls, and epoxy groups, graphene oxide (GO) is a very attractive compound because of its excellent properties [24,25]. However, after the photocatalyst process, the used TiO₂ is commonly recovered through tedious filtration or high-speed centrifugation in order to prevent secondary pollution in the source, which is highly inconvenient, not economically feasible, and difficult to operate. In light of this, magnetite particles have been increasingly used in lab-scale magnetic separations [26]. Additionally, TiO₂ combined with iron oxide family materials has been shown to be more effective in destroying organic pollutants. As an example, Xu et al. reported high performance photocatalytic recovery of Cr(VI) in TiO₂/FeO versus pour TiO₂ [27]. Alternatively, Ag, Pt, Pd, Ru, Rh, and Fe have been added to TiO₂ in order to prevent recombination of electron-hole pairs. As reported in previous studies, Fe⁰ traps electrons produced by TiO₂ photocatalysis by the two half-reactions of Fe³⁺/Fe²⁺ and Fe²⁺/Fe⁰, preventing electron-hole recombination [28]. Bibi et al. applied rGO/Fe₃O₄/TiO₂ in a photocatalytic process to degrade malachite green and methylene blue. According to this study, from 3.2 eV to 2.6 eV, the band gap decreased. For malachite green (MG) degradation, the photocatalyst process via rGO/Fe₃O₄/TiO₂ showed the photocatalytic process is commendable (99%) in comparison to pure TiO₂ (67%) under visible light in 55 min [29]. Based on the above explanations, the presence of pharmaceutical compounds, especially antibiotics, in aquatic environments has become an environmental problem, and the aim of this study is to synthesize the catalyst with high photocatalytic ability to degrade the antibiotic PNG from aquatic environments. Various analyzes were performed to measure the quality of the synthesized catalyst. And, the test design method has been used to reduce human error, reduce the number of test steps, and optimize the process. Photocatalysis process depend on several parameters. We applied to central composite design (CCD) to investigation of effect of parameters on performance, so, in this study, rGO/Fe⁰/Fe₃O₄/TiO₂ nanocomposite was synthesized via hydrothermal method. To investigate the synthesized rGO/Fe⁰/Fe₃O₄/TiO₂ nanocomposite, some spectroscopic and analytical techniques were used.

2. Material and methods

2.1. Analytical chemicals and reagents

The chemicals were purchased from Sigma-Aldrich and were not refined further. Ethanol (C₂H₅OH: 97%), methanol (CH₃OH: 98%), titanium (IV) butoxide [Ti(C₄H₉O)4: 98%], sulfuric acid (H₂SO4: 95–97%), sodium hydroxide (NaOH: 99%), sodium nitrate (NaNO₃: 98%), sodium borohydride (NaBH₄, 98%), graphite powder (99%), potassium permanganate (KMnO₄: 99%), hydrochloric acid (HCl: 37%), sodium sulfate (Na₂SO₄: 98%), Ammonium (NH₄⁺: 99%), acetonitrile (C₂H₃N: 99%), ferric chloride (FeCl₃.6H₂O:97%), ferrous chloride (FeCl₂.4H₂O: 97%), PNG (C₉H₁₁N₂O₄S), with analytical grade bought. COD vial (Low range: 0–150 mg/L) bought from Laviband Co. Deionized water was used to prepare all solutions.

2.2. Preparation of nanocomposite

2.2.1. Preparation of rGO

We prepared graphene oxide according to a modified Hummers' method [30]. Briefly, it was prepared by mixing 10 g of graphite powder (GP) and 5 g of sodium nitrate in 230 mL of sulfuric acid in a flask and stirring the solution vigorously at an ice bath temperature (0 to 5 °C) for 60 min. Moreover, 30 g of potassium permanganate was gently added to the mixture and stirred continuously until green color developed and the temperature was maintained at 15 °C. As soon as the ice bath has been removed, stir the mixture at 35 °C for 24 h until the color of the mixture becomes light brown. It was added gradually to the suspension 460 mL of deionized water, and the reaction temperature increased to 98 °C, and a brownish color was observed. After adding 200 mL of warm water under stirring, H₂O₂ (10 mL) was added to terminate the reaction. To purify the solution, the pH of the solution was neutralized by adding 10% HCl and then washing with deionized water. This was followed by oven drying at 60 °C for 2 h. An ultrasonic bath was used to exfoliate 0.1 g of GO in 100 mL of water. The GO suspension was stirred at 50 °C for 12 h after adding 0.15 g of ascorbic acid. Lastly, the rGO was obtained. After washing with ethanol, centrifuging, and drying at 80 °C in a vacuum oven for 8 h, an rGO sample was obtained.

2.2.2. Preparation of rGO/Fe⁰, and rGO/Fe⁰/Fe₃O₄

As a starting point, 1.2 g of synthesized rGO was dissolved in 240 mL of distilled water in a round flask. Then, 60 mL of FeCl₃.6H₂O (1 M) was dissolved in a solution of rGO at 25 °C, and 140 mL of NaBH₄ (1.6 M) was added to the solution and stirred for 2 h under N₂ gas pressure. Under N₂ gas, the reaction solution was stirred for 1 h. With the aid of a magnet, the solution was collected and washed with ethanol to remove the excess NaBH₄ and iron salt. To obtain rGO/Fe⁰, the nanocomposite was dried at 60 °C in a vacuum oven for 12 h. Under an atmosphere of N₂, the dried material was heated to 400 °C for 2 h to obtain rGO Fe⁰/Fe₃O₄ [24].

2.2.3. Preparation of rGO/Fe⁰/Fe₃O₄/TiO₂

To synthesize rGO/Fe⁰/Fe₃O₄/TiO₂ nanocomposites, the straightforward sol-gel method was used [31]. In brief, we dispersed 0.5 g of rGO/Fe⁰/Fe₃O₄ in 350 mL of ethanol. For 30 min, the above solutions were sonicated. After that, 65 mL of titanium (IV) butoxide was added. The solution was stirred for another 2 h. A 250-mL Teflon-coated stainless steel autoclave with nitrogen purging was used, in which the solution was heated for 12 h at 160 °C under nitrogen purging. Upon completion of the reaction, a precipitate of greyish color was acquired, and it was cleaned several times using ethanol and water. Afterward, the as-prepared nanocomposite was baked in a vacuum oven at 60 °C for 18 h. In (Fig. 1), the schematic of the catalyst synthesis steps is presented.

Fig. 1.

Schematic of rGO, rGO/Fe⁰, rGO/Fe⁰/Fe₃O₄, and rGO/Fe⁰/Fe₃O₄/TiO₂ synthesis.

2.3. Characterization of rGO, rGO/Fe⁰, rGO/Fe⁰/Fe₃O₄, and rGO/Fe⁰/Fe₃O₄/TiO₂

To assess of as-synthesized rGO, rGO/Fe⁰, rGO/Fe⁰/Fe₃O₄, and rGO/Fe⁰/Fe₃O₄/TiO₂, FT-IR spectrophotometer (Spotlight 200i FT-IR Microscopy Systems; 4000-400 cm⁻¹) was utilized to explore the functional groups. X-ray diffraction (XRD) patterns were recorded on XRD diffractometer Rigaku- ZSX Primus 400; radiations source: Cu Kα [(λ = 1.54056 Å) monochromatic incident beam in the range of 10° to 80° with the step interval of 0.02°, and rate of 0.05°/s)] for the assess crystal structure of the structures. Also, the average crystallite size (D) of the prepared nanocomposite can be calculated from the Debye–Scherrer equation (Eq. 1) [32].

$$D=\frac{0.9\lambda }{\beta \mathit{cos}\theta }$$ (1)

A UV–vis spectrophotometer (DRS, Lambda 950, USA) was used to determine the diffuse reflectance absorption spectra of the rGO, rGO/Fe⁰, rGO-Fe⁰/Fe₃O₄, and rGO-Fe⁰/Fe₃O₄/TiO₂. To calculate the band gap, we used the Tauc formula (Eq. 2) [33].

$${\left(\mathrm{ὰ}\mathrm{h}\mathrm{ʋ}\right)}^2=\mathrm{A}(\mathrm{h}\mathrm{ʋ}-{\mathrm{E}}_{\mathrm{g}})$$ (2)

The surface morphology of the samples was observed with a field-emission scanning electron microscope (FE-SEM) (UN41219SEM) under vacuum conditions ≥1.3 × 10⁻⁴ mbar. To analyze the purity and elemental composition of samples, energy dispersive spectroscopy (EDS) was applied. Additionally, the samples were analyzed using a transmission electron microscope (TEM) operating at 200 kV (TEM Model CM200, SUPERTWIN). Quantachrome Autosorb analyzer (BET) was used to measure the special surface areas, volumes, and distribution pore sizes of catalysts at 77 K. Before the degassing process, samples were dried in situ at 200 °C under vacuum for 12 h. Brunauer–Emmet–Teller (BET) methods were used for calculating surface areas over a p/p₀ range. Magnetization curves of the samples were measured using an HH-18 vibrating sample magnetometer (VSM, MPMS-3 model, USA).

2.4. rGO-Fe⁰/Fe₃O₄/TiO₂ point of zero charge measurements

A total of 6 Erlenmeyer containing 25 mL of distilled water with pH values ranging from 2 to 12 were prepared to determine the potential of zero charges (pH_pzc) of rGO/Fe⁰/Fe₃O₄/TiO₂ nanocomposite. To the solutions was added 0.25 g of the nanocomposite. A second pH reading was performed after 24 h had passed.

2.5. Experiments set-up

To evaluate the performance of the synthesized nanocomposite, PNG was selected as the target pollutant. All test steps were performed at ambient pressure and temperature (25 ± 2 °C). In this study, a horizontal rectangular cubic glass photocatalytic reactor with a useful volume of 500 mL was used. In the middle of the reactor, a quartz sheath contained the light source. The reactor was placed inside a polyvinyl chloride (PVC) box to control the emission of light. The stock solution (1000 mg/L) was prepared from PNG powder and different concentrations of antibiotics were prepared from the stock solution. The pH of the samples was adjusted by sulfuric acid and sodium hydroxide (0.1 N). A pH meter (HI-110 - Canada) was used to adjust the pH of the solutions. The nanocomposite was added at the end of sample preparation and the process was started. Homogeneous mixing inside the reactor was performed by a mechanical stirrer. At room temperature and atmospheric pressure, all experiments were conducted. The quartz-glass jacket needed to be equipped with a continuous flow of condensation to lower the lamp's temperature during irradiation. At the end of each step of the process, sampling was performed by pipette and catalyst separated via centrifuge (XM-100, ARA co.), then the concentration of PNG was determined by high-performance liquid chromatography (HPLC). Finally, PNG removal, chemical oxygen demand (COD), and total organic carbon (TOC) performance were calculated via (Eq. (3)).

$$\mathrm{D}\mathrm{e}\mathrm{c}\mathrm{o}\mathrm{m}\mathrm{p}\mathrm{o}\mathrm{s}\mathrm{i}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{f}\mathrm{o}\mathrm{r}\mathrm{m}\mathrm{a}\mathrm{n}\mathrm{c}\mathrm{e}(\%)=\frac{\left(C0-Ce\right)}{C0}\times 100$$ (3)

The initial and residual concentrations (mg/L) of PNG are given by C₀ and C_e.

2.6. Influence parameters, and optimization by CCD

Several factors were considered in this study to assess the effectiveness of PCP in PNG removal, including initial solution pH, nanocomposite dose, reaction time (RT), and PNG concentration. CCD was applied as an approach in response surface methodology (RSM) to determine the effect of these main parameters [34] on PCP efficiency. The parameters and their levels are illustrated in Table 1. To investigate the influence of the experimental factors on the degradation of PNG, we designed 30 runs of experiments using the CCD design.

Table 1.

Parameters and their levels in PCP optimization.

Factors	Names	Units.	Min	Max	Coded low	Coded high	Mean	Std. Dev.
A	pH	–	2	10	−1 ↔ 4	+1 ↔ 8	6	1.82
B	Dose	mg/L	5	25	−1 ↔ 10	+1 ↔ 20	15	4.55
C	PNG Con	mg/L	25	125	−1 ↔ 50	+1 ↔ 100	75	22.74
D	Time	min	15	75	−1 ↔ 30	+1 ↔ 60	45	13.65

According to (Eq. (4)), a second-order polynomial model was developed empirically to describe the removal of PNG via PCP [35].:

$$Y={\beta }_0+\sum _{i=1}^k{\beta }_i{x}_i+\sum _{i=1}^k{\beta }_{ii}{x}_i^2+\sum _{i=1}^{k-1}\sum _{j=i+1}^k{\beta }_{ij}{x}_i{x}_j+\epsilon$$ (4)

There is a response variable Y (%) corresponding to the removal of PNG (%), an intercept b₀, and coefficients i, ii, and ij of the linear, quadratic, and interaction effects of each variable, respectively, xi and xj representing codes representing experimental levels. As part of the comparison between factors with different units, coded values were obtained using (Eq. (5)).

$$x_i=\frac{X_i-X_0}{\Delta x}$$ (5)

It has coded values of xi, a difference between high and median values of x, actual values of the variables, and median values of X₀. To assess the statistical significance of the fitted models and to determine the interaction between response and factors studied, analysis of variance (ANOVA), p-value and F-value were performed. A correlation coefficient, including R², R²_adj, and R²_predict was also used to evaluate. After optimization of the process, the pseudo-first-order kinetic of reaction (Eq. (6)) was used to determine and evaluate the constant rate (k, min⁻¹) of kinetics reaction profiles at different PNG concentrations, and COD and TOC studies, synergistic effect (SF) (Eq. (7)), changes occurred in during of optimum condition on scanning wavelength in the range of 190–400 nm, recovery, and reusability of the nanocomposite, the effect of different light source (UV–C, UV-A, and visible light), determine of adsorption share, the effect of organic, and inorganic radical scavengers, and by-products, probably pathways, and intermediates were investigated.

$$In\left(\frac{Ct}{C0}\right)=-Kt$$ (6)

C_t, C₀, K, and t are PNG concentration at t, PNG concentration at the start of the reaction (mg/L), rate constant (min⁻¹), and reaction time (min).

$$SF=\frac{PCPperformance}{Sumofsinglemechanismperformancr}$$ (7)

2.7. Analytical methods

An HPLC system with a UV–vis detector (Shimadzu LC-20) and a C18 column (Zorbax SB-C18, Agilent Technologies) was used to determine PNG concentration. 0.4 mL/min flow rates were used for mobile phases A and B, respectively, consisting of acetonitrile and water with 0.1% formic acid. 20 mL of each sample after the process was injected. A 220 nm wavelength was used for the detection of PNG. For the PNG, the retention time is 5.8 min. The detection and quantification limits (LOD and LOQ) were calculated based on the standard deviation of the response and the slope of the calibration curves. Analytikjena's multi-C/N 3100 TOC analyzer was used to measure the concentrations of TOC. The COD was measured using a spectrophotometer (DR6000-HACH, USA (5220-C; closed-reflux) [36].

3. Results and discussion

3.1. Characterization of rGO, rGO/Fe⁰, rGO/Fe⁰/Fe₃O₄, and rGO/Fe⁰/Fe₃O₄/TiO₂

3.1.1. FT-IR

Based on FT-IR analysis, the functional groups on the surface of samples were investigated between 4000 and 500 cm⁻¹ (Fig. 2).

Fig. 2.

FT-IR spectra of GO, rGO/Fe⁰, rGO/Fe⁰/Fe₃O₄, rGO/Fe⁰/Fe₃O₄/TiO₂.

An application of FT-IR spectroscopy is to analyze materials qualitatively, in particular, to identify and detect functional groups in organic compounds. The results indicate that the rGO FT-IR spectra exhibit peaks at 3440 cm⁻¹, which correspond to the O–H tensile band, and peaks at 1724, 1626, and 1034 cm⁻¹, which correspond to the C O, C C, and C–O–C vibrations, respectively. The peak band intensities of the oxidized groups were reduced in rGO/Fe⁰ FT-IR spectra. In this case, sodium borohydride may have reduced oxygen-containing groups. In the rGO/Fe⁰/Fe₃O₄ spectrum, all bands related to rGO/Fe⁰ were also observed, including the peak of Fe–O tensile vibration at 582 cm⁻¹. The FT-IR spectra of rGO/Fe⁰/Fe₃O₄/TiO₂ show wide peaks in the range of 750-700, which are indicative of TiO₂ coating on rGO/Fe⁰/Fe₃O₄. In previous studies, similar results have been reported. A study by Cakmak and coworkers reported that the GO surface contains several functional groups, such as hydroxyls, epoxides, and carboxyls. An overlap between the Fe–O and Ti–O vibrations at 500-700 cm⁻¹ was observed on the titanium dioxide/magnetite nanocomposite [31].

3.1.2. XRD

XRD analysis was applied to determine the crystalline properties of the samples (rGO, rGO/Fe⁰, rGO/Fe⁰/Fe₃O₄, and rGO/-Fe⁰/Fe₃O₄/TiO₂) (Fig. 3).

Fig. 3.

XRD pattern of GO, rGO/Fe⁰, rGO/Fe⁰/Fe₃O₄, rGO/Fe⁰/Fe₃O₄/TiO₂.

An apparent sharp diffraction peak was observed at the value of 10.50 in the rGO XRD pattern. According to the (JCPDS 41–1487), based on the modified hummer's method, this peak could be assigned to the (001) plane of graphite oxide. As a result of the oxidation, the graphite structure was destroyed, confirming that graphite powder had been converted to the graphitic oxide by this peak [37]. During the oxidation of graphite powder, numerous functional groups are formed, which are bonded on the edges and both sides of the basal plane. The results of Pudukudy and coworkers' study confirm the results of this study [38]. An XRD analysis confirmed the successful incorporation of Fe⁰/Fe₃O₄/TiO₂ onto GO sheets. The diffraction peaks at 25.68, 37.62, 53.96, 62.46, 68.48, 71.01, and 75.12 were indexed to the (101), (004), (105), (213), (116), and (220) planes of anatase TiO₂ (JCPDS 21–1272), respectively [39]. The result of this study was similar to Nguyen and coworkers' study [40]. Also, reflections were observed at 2θ = 31.5, 35.9, 54.3, 56.78, and 62.8 corresponding to the respective (220), (311), (422), (511), and (440) planes of cubic Fe₃O₄ and (JCPDS NO.19-0629), and reflections at 45.69(100) and 64.31(200) associated with Fe⁰(JCPDS NO.87-0721) [24].

3.1.3. FESEM, EDX, and TEM

The morphologies information of GO and rGO/Fe⁰/Fe₃O₄/TiO₂ were illustrated in (Fig. 4).

Fig. 4.

FESEM, TEM, and EDX of rGO and rGO/Fe⁰/Fe₃O₄/TiO₂.

The FESEM image of rGO (A-1, and A-2) has revealed a flaky surface, an ordered layer of pristine graphite is oxidized, resulting in this condition. More details of the morphology of rGO can be presented by TEM image (A-3). The TEM image confirms the rGO sheet in structure. The Energy-dispersive X-ray spectroscopy (EDX) image (A-4), showed that carbon and oxygen were the main elements in rGO. The obtained results were similar to Baghchesara and coworkers' study [41]. At rGO/Fe⁰/Fe₃O₄/TiO₂ FESEM images (B-1, and B-2), and TEM image (B-3), it was found that well-distributed crystalline TiO₂ was formed on rGO sheets and a core-shell nanocomposite was fabricated. Also, EDX result (B-4), showed that carbon (6.62 W%), oxygen (32.45 W%), titanium (43.47 W%), and iron (17.46 W%) were major elements in rGO/Fe⁰/Fe₃O₄/TiO₂. Boruah et al. also found similar results [42].

3.1.4. VSM, BET, and UV–vis spectra

At room temperature, rGO/Fe⁰/Fe₃O₄ and rGO/Fe⁰/Fe₃O₄/TiO₂ were analyzed for their magnetic properties (Fig. 5a). rGO/Fe⁰/Fe₃O₄/TiO₂ had saturation magnetizations of approximately 40.5 and 15.5 emu. g⁻¹, respectively. The phenomenon is primarily due to the presence of the rGO wrapped around Fe₃O₄ and the subsequent anchoring of TiO₂ nanoparticles on the surface. While rGO/Fe⁰/Fe₃O₄/TiO₂ had a lower saturation magnetization value than rGO/Fe⁰/Fe₃O₄, it was strong enough to collect from aqueous solutions. In previous studies, the magnetic properties of nanocomposite materials formed with TiO₂ and other particles have been reported to be reduced when coated with magnetic or paramagnetic particles. Fan and coworkers' study found that the magnetic saturation values of pure Fe₃O₄, Fe₃O₄TiO₂/RGO-1, 2, and 3 nanocomposites were 50.09, 11.37, 39.2, and 42.11, respectively [43]. A spectrum of diffuse reflectance (220–720 nm) for pure TiO₂, rGO, and rGO/Fe⁰/Fe₃O₄/TiO₂ is shown in (Fig. 5b). There was significant absorption in the UV-C region for pure TiO₂, rGO, and rGO/Fe⁰/Fe₃O₄/TiO₂, but at UV-A, TiO_2, and rGO-Fe⁰/Fe₃O₄/TiO₂ were remarkably absorbed, while at the visible range, only rGO/Fe⁰/Fe₃O₄/TiO₂ was remarkably absorbed. According to the Tauc formula, the bandgap for pure TiO₂ and rGO/Fe⁰/Fe₃O₄/TiO₂ is 3.2 eV, whereas that for rGO/Fe⁰/Fe₃O₄/TiO₂ is 2.2 eV. Compared to pure TiO₂, rGO/Fe⁰/Fe₃O₄/TiO₂ composite has a significantly lower band gap energy. In the rGO/Fe⁰/Fe₃O₄/TiO₂ composite, the band gap narrowed to a greater extent due to the formation of the Ti–O–C bond and core-shell structure, which created a new molecular orbital. Under UV-A and visible light irradiation, the band gap between the rGO/Fe⁰/Fe₃O₄/TiO₂ composite may narrow, resulting in better photocatalytic activity. The bandgaps of TiO₂, TiO₂–Fe₃O₄/graphene, and TiO₂–TiO₂–Fe₃O₄/graphene were reported by Li and coworkers to be 3.12 eV and 2.12 eV, respectively [44].

Fig. 5.

a) VSM of rGO/Fe⁰/Fe₃O₄, and rGO/Fe⁰/Fe₃O₄/TiO₂. b) UV–vis diffuse reflectance spectra of rGO, pure TiO₂, and rGO/Fe⁰/Fe₃O₄/TiO₂.

The specific surface area of the composite rGO/Fe⁰/Fe₃O₄/TiO₂ was calculated based on N₂ adsorption-desorption measurements and BET analysis (Fig. 6). rGO/Fe⁰/Fe₃O₄/TiO₂ reveals a type IV N₂ adsorption-desorption isotherm with H₃ hysteresis loop. Table 2 lists the parameters of pore structure. rGO/Fe⁰/Fe₃O₄/TiO₂ composite has a maximum specific surface area of 74.42 m² g⁻¹. There have been different results reported in the last literature. In Fan and coworkers' study area of Fe₃O₄–TiO₂/RGO nanomaterial, 183.33 m² g⁻¹ was reported [43].

Fig. 6.

Adsorption/desorption isotherm of rGO/Fe⁰/Fe₃O₄/TiO₂.

Table 2.

Summary of BET, Langmuir, t, and BJH test.

BET plot
Vm	17.099	[cm3(STP) g−1]
as,BET	74.422	[m2.g−1]
C	223.23
Total pore volume(p/p0 = 0.990)	0.1142	[cm3.g−1]
Mean pore diameter	6.1384	[nm]
Langmuir plot
Vm	21.308	[cm3(STP) g−1]
as,Lang	92.741	[m2.g−1]
B	0.7995
t plot
Plot data	Adsorption branch
a1	79.11	[m2.g−1]
V1	0	[cm3.g−1]
a2	22.178	[m2.g−1]
V2	0.045765	[cm3.g−1]
2t	1.5979	[nm]
BJH plot
Plot data	Adsorption branch
Vp	0.1031	[cm3.g−1]
rp,peak(Area)	1.29	[nm]
ap	47.29	[m2.g−1]

3.2. Development of the design, the fitting process, and the statistical analysis of the CCD

To determine the condition conducive to optimal performance of the decomposition of PNG by PCP, a statistical analysis of the responses presented in Table 3 was performed. The variables of pH, nanocomposite dose, initial PNG concentration, and reaction time were considered the main variables, and the UA lamp with an intensity of 16 (λ = 385 nm) was considered constant in all runs. According to the results, PCP is effective in removing PNG in a range of 57.6 to 95.3% (mean 77.37%). Fisher variance ratios and F-values are substantiated statistical indexes used in the ANOVA analysis to adjust how to fit factors representing deviations. A statistically significant model based on these indices, namely an F-value (>0.9) and a P-value (0.05), was selected. Design experts presented a quadratic model based on tests that describe the relationship between predicted and experimental values (Table 4).

Table 3.

Experiments designed by the CCD method and its results in the PCP.

Factors	Factor 1	Factor 2	Factor 3	Factor 4	Experimental performance
	A: pH	B: Dose	C: PNG con	D: Reaction time
Unit	–	mg/L	mg/L	min	%
1	8	10	50	30	62
2	6	25	75	45	82.3
3	4	20	50	30	78
4	6	15	75	45	83
5	10	15	75	45	59.1
6	6	15	75	45	83
7	4	20	100	60	78.8
8	4	10	50	30	61.4
9	6	5	75	45	57.6
10	8	10	100	60	62.3
11	4	10	100	30	92.5
12	4	20	50	60	93.2
13	4	10	100	60	61.3
14	6	15	75	45	83
15	6	15	75	45	83.1
16	6	15	125	45	93.4
17	6	15	75	45	83.1
18	4	20	100	30	93.5
19	2	15	75	45	66.7
20	4	10	50	60	66
21	8	10	100	30	77.6
22	6	15	75	15	77.7
23	6	15	75	75	77.4
24	8	20	50	60	95.3
25	8	10	50	60	78.6
26	6	15	25	45	93
27	8	20	100	60	70.1
28	8	20	100	30	75.1
29	6	15	75	45	83
30	8	20	50	30	70.1

Table 4.

Fit summary of the model.

Source	Sequential p-value	Lack of Fit p-value	Adjusted R2	Predicted R2
Linear	0.1211	<0.0001	0.1242	−0.1550
2FI	0.0085	<0.0001	0.4963	0.4643
Quadratic	< 0.0001	< 0.0001	0.9931	0.9794	Suggested
Cubic	0.0038	<0.0001	0.9988	0.9571	Aliased

According to Table 5, the ANOVA results for the quadratic regression model indicate a high reliability and a low probability value, suggesting that the coded relationship between actual data and predicted values can be justified. The correlation coefficient values (i.e., R2, R2adj, and R2predict) were much higher than those achieved for other models, confirming the model's extraordinary validity and efficiency. With an R2 of 0.98, the quadratic model can explain more than 98% of the total variability, and less than 2% of the data are descriptive. There is a perfect correlation between R2, R2adj, and R2predict, indicating that the model is capable of predicting responses accurately [45,46].

Table 5.

ANOVA for the quadratic model of PCP in PNG removal.

Source	Sum of squares	df	Mean square	F-value	p-value
Model	823.78	14	58.84	162.95	<0.0001	significant
A-pH	160.17	1	160.17	443.54	<0.0001
B-Dose	240.67	1	240.67	666.46	<0.0001
C-PNG Con	73.50	1	73.50	203.54	<0.0001
D-Time	121.50	1	121.50	336.46	<0.0001
AB	36.00	1	36.00	99.69	<0.0001
AC	20.25	1	20.25	56.08	<0.0001
BC	30.25	1	30.25	83.77	<0.0001
CD	90.25	1	90.25	249.92	<0.0001
A2	15.43	1	15.43	42.73	<0.0001
B2	3.86	1	3.86	10.68	0.0052
C2	3.86	1	3.86	10.68	0.0052
D2	27.43	1	27.43	75.96	<0.0001
Residual	5.42	15	0.3611
Lack of Fit	5.42	10	0.5417	3.2	0.981	Not significant
Pure Error	0.0000	5	0.0000
Cor Total	829.20	29

Accordingly, the F-value is 162.95 and the P-value is 0.0001, which indicates the model is well suited to explain the amplification of the relationship between response values. In addition, the model discernment is satisfactory if the precision ratio is greater than 4.0 [47].

As described below, the polynomial equation corresponding to PNG removal (%) is represented as coded factors (Eq. (8)):

$$\mathrm{P}\mathrm{N}\mathrm{G}\mathrm{r}\mathrm{e}\mathrm{m}\mathrm{o}\mathrm{v}\mathrm{a}\mathrm{l}(\%)=63.12+(4.5\times \mathrm{p}\mathrm{H})+(0.4\times \mathrm{D}\mathrm{o}\mathrm{s}\mathrm{e})–(0.3\times \mathrm{P}\mathrm{N}\mathrm{G})+(0.07\times \mathrm{T}\mathrm{i}\mathrm{m}\mathrm{e})–(0.15\times \mathrm{p}\mathrm{H}\times \mathrm{D}\mathrm{o}\mathrm{s}\mathrm{e})–(0.02\times \mathrm{p}\mathrm{H}\times \mathrm{P}\mathrm{N}\mathrm{G})+(0.011\times \mathrm{D}\mathrm{o}\mathrm{s}\mathrm{e}\times \mathrm{P}\mathrm{N}\mathrm{G})+(0.006\times \mathrm{P}\mathrm{N}\mathrm{G}\times \mathrm{T}\mathrm{i}\mathrm{m}\mathrm{e})–(0.1875\times {\mathrm{p}\mathrm{H}}^2)+(0.015\times {\mathrm{D}\mathrm{o}\mathrm{s}\mathrm{e}}^2)+(0.0006\times {\mathrm{P}\mathrm{N}\mathrm{G}}^2)–(0.004\times {\mathrm{T}\mathrm{i}\mathrm{m}\mathrm{e}}^2)$$ (8)

As shown in (Fig. 8), the studied mediators had both positive and negative effects on the removal of PNG. Even though the coefficient of variation (CV) was low for the quadratic model, this confirms the precision of the experiments and the reproducibility of the model [48].

Fig. 8.

a) COD and TOC removal efficiency, b) The synergistic effect of different mechanisms in PCP (pH = 6.5, RT = 59.1 min, PNG concentration = 52 mg/L, and catalyst dose = 18.5 mg/L).

3.3. The impact of parameters on the PNG removal at PCP

Fig. 7, illustrated the effect of pH, nanocomposite dose (Fig. 7a), initial concentration of PNG (Fig. 7b), and reaction time, values influence the PCP performance.

Fig. 7.

Effect of parameters into PNG removal in PCP.

As can be observed, all parameters' influence on the removal performance is significant, and the highest removal performance of PNG (more than 90%) occurred. By increasing the pH of the solution to near 6.0, the degradation degree was increased considerably, but then clearly declined when the pH value was increased to 10.0 (Fig. 7a). As described in the literature, pH at PCP affects the absorption capacity and decomposition of organic compounds, the distribution of electricity on the catalyst's surface, and the valence band oxidation potential. Photocatalytic oxidation is dependent on ph. The catalyst's surface has various charges when pH is below or above zero. It was determined that pH_zpc is approximately 6.5 in this study. As pH_zpc increases and decreases, the catalyst surface has a negative and positive charge. Catalyst surfaces are often colonized by H⁺ ions at pH levels below 6.5. At pH levels above 6.5, the catalyst's surface is colonized by the OH⁻ ion. In the same way, a different charge is illustrated when pH is lower or higher than the pka of the material (pka PNG = 2.74). Therefore, both the catalyst and pollutants have different ionization states at different pH levels. Due to the nature of the catalyst employed, and PNG, the decomposition performance of PNG favored a near-neutral solution. It is important to note that the change in pH affects the isoelectric point or the surface load of the photocatalyst. PNG tends to become an anion if pH exceeds pka [49]. The electrostatic repulsion force between the catalyst and the PNG is generated if pH is greater than 6.5, given the negative surface charge of the catalyst and the anionic nature of PNG. Accordingly, the degradation of PNG is reduced. Nevertheless, other studies have demonstrated that pH directly influences radical production. In strongly acidic conditions, PNG degradation is decreased due to hydroxyl radical scavenging (Eq. (9)) [1].

$${\mathrm{H}}^{+}+°\mathrm{O}\mathrm{H}+{\mathrm{e}}^{-}\to {\mathrm{H}}_2\mathrm{O}$$ (9)

Moreover, due to a reduction in the hydroxyl radical oxidation potential, and the formation of insoluble compounds, PNG degradation at alkaline pH is decreased. Therefore, increasing pH will result in hydrogen peroxide decaying and reducing ultraviolet radiation and radical hydroxyl production. Berkani and coworkers also reported that Penicillin G was best degraded by photocatalysis at a pH of 6.8 in an aqueous solution [7]. One of the other parameters investigated was the nanocomposite dose. A catalyst's dose affects the active site and the absorption of light [50]. As can be seen in (Fig. 7a), the PNG removal efficiency was improved by increasing the nanocomposite dose to 18.5 mg/L and then enforcing it. Increasing catalyst dosage increases the area of active sites, but increasing solution turbidity can reduce light penetration into the solution due to increasing catalyst dosage. As a result, the rate of PNG removal decreased. An optimum dose of ZNS of 0.5 g/L was reported in a study by Pouretedal et al. [51]. Pollution concentrations play an important role in the operation of processes. A study of the effect of initial concentrations of pollutants on the efficiency of a process is necessary due to the presence of different concentrations of pollutants, such as pharmaceutical compounds in the environment. PNG concentrations between 50 and 100 mg/L were investigated in this study. According to the results, the efficiency of the PCP decreased with increasing the initial concentration of PNG in the reaction medium (Fig. 7b). This is because as PNG molecules are concentrated in the environment, the likelihood of exposure to direct or indirect oxidizing agents decreases. There is a decrease in the probability of degradation of PNG as the number of PNG molecules oxidizing free radicals increases. During PCP, the presence of intermediate and organic compounds increases as there is a decreasing chance of complete oxidation. Similarly, Kamranifar and coworkers examined the photocatalytic capabilities of penicillin G at various concentrations (10–100 mg/L) in their study. At 10 mg/L, the highest performance was achieved (70.7%) [1]. PCP is also influenced by reaction time. This study examined the degradation of PNG over 30–60 min. In the study, it was found that the removal efficiency of PNG increases with an increase in reaction time (Fig. 7c). By increasing the reaction time, more hydroxyl radicals and electron-hole pairs are formed in the PCP, increasing the rate of PNG degradation by direct and indirect oxidation, and therefore increasing the efficiency of the process. The effect of reaction time has been investigated in previous studies. According to Al-Musawi, 60 min is the optimum reaction time for Ampicillin degradation [52]. Finally, after the optimization of PCP via CCD, optimal condition was determined as pH∼ 6.5, catalyst dose 18.5 mg/L, reaction time 59.1 min, and initial PNG concentration 52 mg/L. In these condition, the PCP was able to achieve 96% of the initial PNG concentration degradation.

3.4. Additional studies in optimal conditions

3.4.1. COD, TOC removal, changing in wavelength scan, determine the catalyst components efficiency on PNG degradation, and synergistic effect

COD and TOC tests are relatively suitable indicators for determining the mineralization of organic pollutants in advanced oxidation processes. In this study, these two indicators were investigated. Under optimal conditions, the photocatalytic process can remove 78% of the COD and 65% of the TOC from the PNG (Fig. 8a). Since degradation of PNG into intermediates may occur in the PCP, it is obvious that the rate of reduction of COD and TOC indicators was always less than the removal of PNG. To more efficient removal of COD and TOC, a higher reaction time was required. Within 120 min reaction time, more than 99% of PNG, more than 90% of COD, and TOC were decreased. In Aanchal Rathi and coworkers' study, metronidazole, COD, and TOC declined 72, 68, and 62%, respectively [53]. Absorption changes in wavelength scanning during the reaction time are presented in (Fig. 8 a*). According to the results, the rate of absorption varied at 0–60 min in the PCP process. Pirzada and coworkers study the Rhodamine B and p-Chlorophenol photodegradation under solar light in the presence of LaFeO₃/Ag₂CO₃ nanocomposites. The results show that in the period from 0 to 45 min, noticeable changes in absorption spectra have occurred and the concentration of pollutants has decreased [54].

Herein, the role of all applied mechanisms in PCP in the PNG degradation in the two systems was conducted separately and concurrently under optimal conditions, and the related results are illustrated in (Fig. 8b). Adsorption and photolysis (UV–C, UV-A, and visible light) were influence mechanisms in PCP. As you can see, all of mechanisms had very lower efficiency in PNG degradation separately. Sorption, UV-A, UV-C, visible light photolysis, PCP(UV-C), PCP (UV- A), and PCP(vis-light) were 10, 12, 25, 7, 98, 96, and 72% efficiency respectively. The SF coefficient to PCP(UV-A), PCP(UV-C), and PCP(Vis) were investigated 4.36, 2.57, and 4.23 respectively. The results of determining the synergistic effect showed with the simultaneous presence of light source and catalyst in the PCP, the main mechanism of PNG degradation was indirect oxidation by free radicals and electron-hole pairs. The reason for the higher efficiency of the process in the presence of UV-C was the further oxidation by photolysis. In Yu and coworkers' study, SF in photocatalyst process for Oxytetracycline (OTC) removal 4.64 reported. The electrons accumulated on the catalyst, and generating reactive oxidant species (ROS_s) were reasons for the synergistic degradation of OTC by this process [55].

3.4.2. Photocatalytic degradation mechanism of PNG, radical scavengers effect on PCP, recovery, and reusability of catalyst

An explanation for the degradation of PNG by PCP-rGO/Fe⁰/Fe₃O₄/TiO₂ systems has been proposed, and a schematic is shown in (Fig. 9). Different components in the synthesized nanocomposite have unique properties and specific activities. Graphene, a unique sp² hybrid carbon network material, including graphene oxide (GO) and reduced graphene oxide (rGO), has been used as an ideal cocatalyst or support for improving photocatalytic efficiency due to its high electron mobility, large surface area, and high transparency. Photo-induced electrons are quickly transferred from the conduction band of TiO₂ to its surfaces in rGO/TiO₂ due to the excellent electron acceptor properties of rGO sheets [56]. As Fe⁰ in the composite can trap electrons generated from TiO₂ photocatalysis via Fe³⁺/Fe²⁺ and Fe²⁺/Fe⁰ half-reactions, it prevents electron-hole recombination. Fe⁰ with TiO₂ was expected to preserve Fe⁰'s reduction reactivity and improve its stability and mobility by preventing the formation of oxide layers on its surface [28]. The separation of TiO₂ from aqueous solutions has been reported as an important drawback in studies. To facilitate photocatalytic recollection in magnetic fields, doping of magnetic materials has been proposed. For this purpose, ferriferous oxide (Fe₃O₄) can be doped on semiconductors [57]. Furthermore, the first step is to adhere PNG molecules to the surface of the rGO/Fe⁰/Fe₃O₄/TiO₂ photocatalyst in the reaction medium (Eq. (10)). Meanwhile, a hole (h⁺) has been created within the valence band (VB) of rGO/Fe⁰/Fe₃O₄/TiO₂ nanocomposite as a result of the absorption of energy photons. (Eq. (11)). From the oxygen presence in aqueous, electrons were immediately transferred to reduce e⁻/h⁺ recombination (Eq. (12)). As a result of the holes (h⁺) in the VB band, H₂O molecules decompose into oxygen and water (Eq. (13)). All ROSs produced, including °OH, h^+, and O₂^●–, attack the aliphatic and aromatic rings of PNG on the nanocomposite reaction site. PNG molecules are oxidized by all radicals to intermediates, CO₂, and H₂O (Eq. (14)) [58]. It is believed that the enhanced photocatalytic activity results from a synergistic interaction between nanostructured TiO₂ and Fe₃O₄ with rGO and their catalytic activity. In this case, PNG degradation can occur through several different methods, including the direct attack by the hydroxyl radical, direct reduction by photo-induced electrons, or direct oxidation in presence of holes. Due to strong van der Waal interactions, rGO sheets tend to agglomerate back into the graphitic structure. In this case, TiO₂ is incorporated on the surface of the graphene sheet to reduce agglomeration. Additionally, the interlink between TiO₂ and rGO during photocatalysis allows easy electron transfer, which improves the separation of charge carriers and hence the activity of the catalyst. Eventually, Fe₃O₄ increases degradation rate by adsorbing pollutants onto nanocomposite surfaces. The magnetic properties of Fe₃O₄ also make it useful for magnetic separation of nanocomposite photocatalysts from dispersions. As a result, the incorporation of TiO₂, Fe⁰, and Fe₃O₄ into rGO enhances its photocatalytic activity. Similar results were reported in Benjwal and coworkers' study. In this study use of rGO/Fe₃O₄/TiO₂ nanocomposite to Methylene Blue degradation. Based on the results, the ternary nanocomposite (rGO–Fe₃O₄–TiO₂) exhibits the highest dye degradation efficiency (100% within 5 min) when compared to single and binary nanocomposite. A synergetic interaction between rGO and Fe₃O₄ and TiO₂ is responsible for this enhancement [37]. In Banerjee and coworkers' study, similar results were reported [59].

$$\mathrm{r}\mathrm{G}\mathrm{O}/{\mathrm{F}\mathrm{e}}^0/{\mathrm{F}\mathrm{e}}_3{\mathrm{O}}_4/{\mathrm{T}\mathrm{i}\mathrm{O}}_2+\mathrm{P}\mathrm{N}\mathrm{G}\to \mathrm{P}\mathrm{N}\mathrm{G}-\mathrm{r}\mathrm{G}\mathrm{O}/{\mathrm{F}\mathrm{e}}^0/{\mathrm{F}\mathrm{e}}_3{\mathrm{O}}_4/{\mathrm{T}\mathrm{i}\mathrm{O}}_2$$ (10)

$$\mathrm{r}\mathrm{G}\mathrm{O}/{\mathrm{F}\mathrm{e}}^0/{\mathrm{F}\mathrm{e}}_3{\mathrm{O}}_4/{\mathrm{T}\mathrm{i}\mathrm{O}}_2+\mathrm{h}\mathrm{ʋ}\to \mathrm{r}\mathrm{G}\mathrm{O}/{\mathrm{F}\mathrm{e}}^0/{\mathrm{F}\mathrm{e}}_3{\mathrm{O}}_4/{\mathrm{T}\mathrm{i}\mathrm{O}}_2({\mathrm{e}}_{\mathrm{C}\mathrm{B}}^{-}+{\mathrm{h}}_{\mathrm{V}\mathrm{B}}^{+})$$ (11)

$$\mathrm{r}\mathrm{G}\mathrm{O}/{\mathrm{F}\mathrm{e}}^0/{\mathrm{F}\mathrm{e}}_3{\mathrm{O}}_4/{\mathrm{T}\mathrm{i}\mathrm{O}}_2\left({\mathrm{e}}_{\mathrm{C}\mathrm{B}}^{-}\right)+{\mathrm{O}}_2\to {\mathrm{O}}_2^{●–}+\mathrm{r}\mathrm{G}\mathrm{O}/{\mathrm{F}\mathrm{e}}^0/{\mathrm{F}\mathrm{e}}_3{\mathrm{O}}_4/{\mathrm{T}\mathrm{i}\mathrm{O}}_2$$ (12)

$$\mathrm{r}\mathrm{G}\mathrm{O}/{\mathrm{F}\mathrm{e}}^0/{\mathrm{F}\mathrm{e}}_3{\mathrm{O}}_4/{\mathrm{T}\mathrm{i}\mathrm{O}}_2\left({\mathrm{h}}_{\mathrm{V}\mathrm{B}}^{+}\right)+{\mathrm{H}}_2\mathrm{O}\to °\mathrm{O}\mathrm{H}+{\mathrm{H}}^{+}$$ (13)

$${\mathrm{O}}_2^{●–}+°\mathrm{O}\mathrm{H}+{\mathrm{h}}^{+}+\mathrm{P}\mathrm{N}\mathrm{G}\to \mathrm{o}\mathrm{r}\mathrm{g}\mathrm{a}\mathrm{n}\mathrm{i}\mathrm{c}\mathrm{a}\mathrm{n}\mathrm{d}\mathrm{i}\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{g}\mathrm{a}\mathrm{n}\mathrm{i}\mathrm{c}\mathrm{i}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{m}\mathrm{e}\mathrm{d}\mathrm{i}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{s}+{\mathrm{C}\mathrm{O}}_2+{\mathrm{H}}_2\mathrm{O}$$ (14)

Fig. 9.

Mechanism of PNG degradation by PCP- rGO/Fe⁰/Fe₃O₄/TiO₂ system.

To investigate the main active radicals produced by the nanocomposite during PCP for the degradation of PNG, radical trapping experiments were conducted (Fig. 10a). A quencher for superoxide radicals is p-benzoquinone (BQ), whereas ethylenediaminetetraacetate acid disodium (EDTA-2Na) and tertbutanol (t-BuOH) are applied as quenchers for the hole radical and the hydroxyl radical (°OH). The degradation efficiency of PNG was negligibly affected by EDTA-2Na, Accordingly, h⁺ plays a virtually insignificant role in the reaction system. A reduction in the degradation rate of PNG was observed following the addition of t-BuOH, indicating that °OH plays a critical role in the catalytic reaction. Nevertheless, the PCP performance was extremely shown with the addition of BQ, suggesting that O₂^•− has a powerful role in PNG decomposition. PNG was more immediately removed with the O₂^•− and °OH radicals in the PCP. A similar result was reported in a study by Saravanakumar and co-workers [60]. For the application of photocatalysts, stability, reusability, and recyclability are vital parameters. According to (Fig. 10 b), the degradation activity of the rGO/Fe⁰/Fe₃O₄/TiO₂ nanocomposite did not significantly decrease at the fifth cycle under visible light irradiation, indicating that the nanocomposite exhibits suitable photocatalytic stability. In Albornoz and coworkers' study, the stability of SnO₂-doped TiO₂ catalyst was investigated. Based on the results of this study, the catalyst had 44 h stability in the photocatalyst process [61].

Fig. 10.

a) Scavengers effect on PCP performance, b) Recovery and reusability of nanocomposite in PCP (pH = 6.5, RT = 59.1 min, PNG concentration = 52 mg/L, and catalyst dose = 18.5 mg/L).

3.4.3. Study of the kinetics of reaction

An empirical model of PNG degradation in PCP is constructed to describe the experimental data (Eq. (6)) was developed under optimized conditions and at different concentrations of antibiotic. A linear pseudo-first-order kinetic model for PNG degradation over PCP is shown in (Fig. 11) and Table 6. At 100 mg/L concentration, there was a significant correlation between the experimental data and the pseudo-first-order kinetic (R² > 0.98). It was found that the apparent rate constant of PNG decomposition was 1.8 min⁻¹ under these conditions. In Sekaran and coworkers' study, the degradation of acid blue 113 dye via photocatalysis process followed pseudo-first-order kinetic (R² > 0.99) [62].

Fig. 11.

The kinetics for the degradation of PNG in PCP (pH = 6.5, RT = 59.1 min, PNG concentration = 50,75, and 100 mg/L, and catalyst dose = 18.5 mg/L).

Table 6.

The kinetic constants for the degradation of PNG over PCP.

PNG Concentration (mg/L)	K × 10 −2(min−1)	R2
50	4.9	0.83
75	2.4	0.92
100	1.8	0.98

3.4.4. Mineralization of PNG and by-products studies

To investigate the intermediate and by-products and probably reaction pathways of PNG degradation at optimal conditions, the GC-MS was used according to the instructions provided in the materials and methods section. According to the results of GC-MS analysis and (Fig. 12), the PNG and several by-products were detected.

Fig. 12.

The analysis of the produced intermediates during PNG degradation via GC–MS (pH = 6.5, RT = 59.1 min, PNG concentration = 50 mg/L, and catalyst dose = 18.5 mg/L).

As a result of the oxidation of PNG, seven important intermediate products were Phenoxyacetaldehyde, Acetanilide, Diacetamate, Phenylalanylglycine, N-Acetyl-l-phenylalanine, Diformyldapsone, and Succisulfone. PNG degradation over a photocatalytic process has previously identified those by-products as intermediate compounds [7]. It was conceivable to attempt the transmutation pathways of PNG over the PCP process as an AOPs based on Free radicals, and electron-hole pairs (Fig. 13). This figure illustrates two possible pathways for PNG photodegradation in PCP. PNG consists of a five-membered thiazolidine ring fused to a strained lactam ring. There is substantial research indicating that this antibiotic ring is chemically unstable and susceptible to extreme conditions, including heat, acids, bases, UV light, nucleophiles, and enzymatic reactions in which the hydrolytic ring opens, resulting in a loss of antibiotic activity. The similar result was reported in Berkani and coworkers' study [7]. Based on the review of previous studies, it was found that the reaction pathway and intermediates produced in the degradation of PNG can be different based on the type of process, the catalyst used, and the degradation mechanism. Also, various intermediate materials have been reported in the photocatalytic process with different catalysts. Studies such as Berkani and coworkers [7], Chavoshan and coworkers [63], and Baladi and coworkers [64] are among the studies that reported intermediate substances produced during the degradation of PNG.

Fig. 13.

Proposed degradation pathway of PNG via PCP in optimum condition (pH = 6.5, RT = 59.1 min, PNG concentration = 50 mg/L, and catalyst dose = 18.5 mg/L).

Lastly, Table 7 presents several studies involving the application of different AOPs processes to the degradation of PNG from solution media.

Table 7.

Different AOPs studies in PNG degradation.

AOPs process	Results	Ref
CoFe2O4@CuS nanocomposite Synthesized and photocatalytic degradation of PG	Nanocomposite images obtained by SEM and TEM indicate that its size is less than 100 nm. Photocatalytic degradation of PG resulted in the maximum degradation of PG of 70.7% (pH = 5, nanocomposite dose of 0.2 g/L, PG concentration of 10 mg/L, and contact time of 120 min). Moreover, the photocatalytic reaction exhibited a pseudo-first-order kinetic behavior. Further, no significant decrease in catalyst efficiency was observed after five consecutive runs.	Kamranifar [1]
MMO-Ti/RuO2IrO2 anode made and Photoelectrolysis process to PG removal	As a result of photo-electrolysis, the toxicity and antibiotic effect on urine of the bacteria were reduced.	Gonzaga [65]
Fenton-like degradation of penicillin G	A Fenton-like reaction was found to degrade Penicillin G depending on pH, FeSO4 and CaO2.	Nisapai [66]
Heat-activated/PMS degradation of PNG in solutions	At pH = 5, the maximum removal rate of PNG was obtained and the decomposition of PNG was enhanced by increasing SPS doses from 0.05 to 0.5 mM.	Norzaee [67]
Photocatalytic degradation of Penicillin G (PG)	PG can be completely degraded by PPS at 500 mg L−1. Study results validated the safety and efficiency of the implemented system via Bacillus subtilis, which has demonstrated significant activity. It is highly efficient, low cost, and provides facilities benefits.	Berkani [7]
Synthesis and Characterizations of rGO/Fe0/Fe3O4/TiO2 Nanocomposite and Optimization of Photocatalytic Process in the Degradation of Penicillin G	The quadratic model of optimization is based on adjusted R2(0.99), and predicated R2(0.97) suggested. According to the ANOVA test, the model was significant (F-Value = 162.95, P-Value = 0.0001). The optimum condition of PCP obtained at pH = 6.5, catalyst dose = 18.5 mg/L, reaction time = 59.1 min, and PNG concentration = 52 mg/L (efficiency = 96%). The COD and TOC decrease were 78, and 65%. The photolysis and adsorption mechanism as a single mechanism had lower performance in PNG degradation. BQ had the greatest effect on reducing the efficiency of the process as a radical scavenger. The °OH, h+ and O2●– were the main ROS in PNG removal. Phenoxyacetaldehyde, Acetanilide, Diacetamate, Phenylalanylglycine, N-Acetyl-l-phenylalanine, Diformyldapsone, and Succisulfone were the main intermediates in PNG degradation.	This study

4. Study limitation

Advanced oxidation processes, including photocatalytic processes, have been investigated over the past few decades, but unfortunately, many of them remain at the laboratory scale. Application on a real scale depends on conducting more extensive studies and also investigating the influencing factors on the treatment process. The main drawback of the photocatalytic process is the cost of the process and its efficiency in real media like industrial wastewater. The presence of interfering factors in wastewater, including other organic substances (as a scavenger), heavy metal ions, and interfering ions are among the factors that limit the application of this process. It is recommended to be used in the real scale, reaction reactor design, the catalyst used (with determining its complete characteristics), environmental influencing factors, etc. should be carefully investigated.

5. Conclusion

This research work studied the synthesized rGO/Fe⁰/Fe₃O₄/TiO₂ and optimization of PCP to PNG degradation in aqueous media. In the optimum condition of PCP, supplementary studies were done. Briefly, rGO/Fe⁰/Fe₃O₄/TiO₂ nanocomposite was successfully synthesized according to Hummer's, and straightforward sol-gel method. The FESEM, EDX, TEM, FT-IR, XRD, BET, UV spectra, and VSM analysis were applied to determine the catalyst properties. The results of the analysis revealed that the nanocomposite was well synthesized and had superior properties for use in the PCP to organic pollutant degradation. PNG can be detected at very low concentrations using a LOD of 0.001–0.03 mg/L. As an added benefit, the synthesized rGO/Fe⁰/Fe₃O₄/TiO₂ nanocomposite is magnetically separable, leading to further degradation along with recyclability of up to five times. The pH, dose of the nanocomposite, and initial concentration of PNG. And reaction time optimized via CCD. The quadratic model of optimization is based on adjusted R²(0.99), and predicated R²(0.97) suggested. According to the ANOVA test, the model was significant (F-Value = 162.95, P-Value = 0.0001). The optimum condition of PCP obtained at pH = 6.5, catalyst dose = 18.5 mg/L, reaction time = 59.1 min, and PNG concentration = 52 mg/L (efficiency = 96%). The COD and TOC decrease were 78, and 65%. The photolysis and adsorption mechanism as a single mechanism had lower performance in PNG degradation. BQ had the greatest effect on reducing the efficiency of the process as a radical scavenger. The ROSs (°OH, h^+, and O₂^●–) were the main space in PNG decomposition. Phenoxyacetaldehyde, Acetanilide, Diacetamate, Phenylalanylglycine, N-Acetyl-l-phenylalanine, Diformyldapsone, and Succisulfone were the main intermediates in PNG degradation. This study was done on a laboratory scale and synthetic samples containing PNG. To properly judge the use of the photocatalytic process as well as the synthesized catalyst on a real scale, it is recommended that additional studies be carried out. These studies include surveying the performance of the process with real wastewater containing antibiotics, especially PNG, determining the toxicity of the effluent from the process, examining the economic aspect of the process, etc. By completing these studies, it is possible to decide with certainty regarding the use of this process practical application.

Declarations

Author contribution statement

Jamal Mehralipour: conceived and designed the experiments, analyzed and interpreted the data, materials, contributed reagents, materials, analysis tools or data, wrote the paper.

Susan Bagheri: performed the experiments, materials, contributed reagents, materials, wrote the paper.

Mitra Gholami: conceived and designed the experiments, analyzed and interpreted the data, wrote the paper.

Data availability statement

Data will be made available on request.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.