Efficient Photocatalytic Degradation of Aqueous Atrazine over Graphene-Promoted g-C₃N₄ Nanosheets

Abstract

Atrazine is a systemic herbicide widely used in weed control. In recent years, it has been largely detected in surface and groundwater in several locations all over the world. Photocatalysis is a green and sustainable technology with huge application prospects in pollution control and the degradation of organic water pollutants. In this work, photodegradation of aqueous atrazine was investigated over pristine graphitic carbon nitride (g-C₃N₄) synthesized via urea pyrolysis and graphene/g-C₃N₄ composite synthesized via the in situ growth method involving direct deposition of g-C₃N₄ nanosheets on the graphene surface. The obtained photocatalysts were characterized using transmission and scanning electron microscopy, Fourier-transformed infrared spectroscopy, UV-visible spectroscopy, photoluminescence spectroscopy, X-ray diffraction, and surface area measurements. It was demonstrated that the composite material exhibited remarkable photocatalytic properties for the efficient degradation of aqueous atrazine under visible light at ambient temperature. After 5 h of reaction, atrazine conversion reached 100% in the presence of graphene/g-C₃N₄ composite, while the pristine g-C₃N₄ allowed 40% conversion under the same conditions, thus demonstrating the positive effect of graphene on the photocatalytic activity of g-C₃N₄. Moreover, graphene/g-C₃N₄ was shown to keep its activity even when it was recycled five times, thus proving its stability and its potential to be used at the industrial scale.

1. Introduction

With the continuous growth of the world’s population, the demand for crops has increased, and the use of chemical fertilizers and pesticides has seriously impacted global soil and water quality [1,2]. Pesticide residues in large quantities occurring in the environment have received much attention as new emerging contaminants. Herbicides, a class of pesticides, remain the most efficient and economical way to control weeds. With the development of herbicide-tolerant crops, the use of herbicides is increasing around the world, resulting in severe contamination of the environment [3]. Atrazine is a commonly used herbicide in several countries where it is among the most used and sold [4]. This herbicide is used on agricultural crops, in urban areas, in forestry and in railroad or energy transportation outlets. Many pesticides, mostly herbicides, including atrazine, have been detected in surface and groundwater [5]. Due to its high chemical stability, slow hydrolysis, low volatility and poor biodegradability, atrazine is often found and accumulates in soil ecosystems [6,7]. Despite atrazine’s low water solubility, there is much concern about the contamination of water resources [5,7]. Atrazine is recognized as an endocrine disruptor [6]. The metabolism of atrazine has been investigated in surrogate animals and humans [8]. Several investigations reported that atrazine’s half-life in the aquatic environment and groundwater resources ranges from 41 to 237 days and 15 months to 20 years, respectively [9]. The maximum contaminant level for atrazine in drinking water established by the United States Environmental Protection Agency (US EPA) and the World Health Organization (WHO) is 3.0 and 2.0 μg·L⁻¹, respectively [10]. It is, therefore, essential to find effective solutions for the degradation of such water pollutants.

Photocatalysis, an advanced oxidation process (AOP), is considered as a cost-effective means to degrade toxic and harmful pollutants. In recent years, graphitic carbon nitride (g-C₃N₄), a metal-free visible-light-activated semiconductor, has attracted significant attention as a promising material for various environmental remediation applications and pollutant degradation [11]. Due to its unique electronic band structure, g-C₃N₄, a low-cost and non-metal photocatalyst, shows great application prospects in the decomposition of water into hydrogen and oxygen, photocatalytic degradation of organic pollutants and carbon dioxide reduction. However, g-C₃N₄ exhibits some shortfalls, such as small specific surface area, low light quantum yield, and high recombination of electron–hole pairs, thus reducing its application in the field of photocatalysis [12]. To improve the photocatalytic activity of g-C₃N₄, many strategies have been developed. They mainly included developing various nanostructures (such as nanosheets, nanoparticles, nanorods and nanospheres) [13], as well as morphological [14,15,16,17], molecular [18,19,20,21,22], and electronic [23,24,25] modifications. Also, surface engineering strategies such as cocatalyst loading and heterogeneous structure construction have been reported [26,27]. Graphene is a two-dimensional structure of sp²-hybridized carbon atoms arranged hexagonally in a honeycomb framework with a high strength-to-weight ratio, unique and extraordinary electrical properties, high thermal conductivity, and high surface area [28,29,30,31,32]. The electronic–band structure of graphene combines semiconducting and metallic characteristics [28]. This property stems from graphene’s unit cell, comprised of two carbon atoms, which makes the π and π* bands indistinguishable. Moreover, the close carbon–carbon distance in graphene leads to a pronounced interatomic overlap [33,34]. Thus, for application purposes, graphene is not used as pristine, but it must be employed as reduced graphene oxide, activated graphene, doped graphene, graphene/metal oxide composites or graphene/polymer composites [30]. Since these carbon materials exhibit high electron mobility, the photocatalytic activity of g-C₃N₄ can be significantly enhanced by coupling with these carbon materials. However, it has been shown that it is difficult to directly use pristine graphene as a material due to its long and difficult synthesis as well as the formation of agglomerations in solution due to van der Waals interactions [35,36,37]. An alternative approach is the recourse to graphene oxide (GO) compound with a similar structure of graphene (G) and presenting hydroxyl (–OH), alkoxyl (C–O–C), carbonyl (C=O), carboxylic acid (–COOH) and other oxygen-based surface functional groups [38]. Graphene oxide has many advantages and disadvantages. Among the advantages, GO is easily synthesized, and its oxygen groups allow not only a very high wettability but also the possibility of functionalization of its surface. Also, it is feasible to synthesize reduced graphene oxide (rGO) or graphene (G) using different methods, with the aim of minimizing the number of oxygen groups and obtaining properties closer to those of pristine graphene [39]. However, these electrical, thermal, and mechanical properties sometimes limit its application. Therefore, when synthesizing GO, the excessive addition of functional groups in its surface impairs its electrical conductivity, making GO electrically resistive (1.64 × 10⁴ Ω·m). In addition, GO has a low thermal conductivity of 0.5 to 1 W/m·K [40]. Given these modest properties, several researchers have explored various techniques to improve the performance of GO. For instance, the reduction of GO to graphene (G) or reduced graphene oxide (rGO) seems to be an effective alternative to obtain excellent material properties. Several published works reported a very clear overview of the properties of graphene compared to those of GO. Graphene has a high electron mobility (25 m²/V·s) as well as a high electrical conductivity (6500 S/m) [41,42]. Indeed, the reduction of GO via a pyrolysis method at very high temperatures is based on the decomposition of oxygen groups into CO and CO₂ gases [39]. These gases effectively exfoliate the graphene oxide into nanosheets.

The combination of graphene with graphitic carbon nitride into composites with improved photocatalytic efficiency was previously reported for the photodegradation of dyes such as methylene blue, methyl-orange, and rhodamine [43]. However, graphene/graphitic carbon nitride composites were hardly reported for the photocatalytic degradation of pesticides in general and atrazine in particular [44]. Moreover, few g-C₃N₄-based photocatalysts were reported for atrazine degradation, such as AgCl@pg-C₃N₄ nanocomposites [45], ternary MoS₂-loaded ZnO-g-C₃N₄ nanocomposites [46], Z-scheme CdS/RGO/g-C₃N₄ hybrid photocatalysts [47], activated carbon/g-C₃N₄ composites with peroxymonosulfate activation [48], oxygen-doped g-C₃N₄ [49] and p-p stacked step-scheme PDI/g-C₃N₄/TiO₂@Ti₃C₂ [18].

The aim of the present work is the investigation of the visible light-driven photocatalytic degradation of aqueous atrazine in the presence of g-C₃N₄ and the composite graphene/g-C₃N₄ for the first time in the literature. The specific objectives are (i) the synthesis and characterization of the photocatalytic materials and (ii) the investigation of the effect of operating conditions on the reaction advancement.

2. Results and Discussion

2.1. Photocatalysts Characterization

The XRD patterns of the synthesized photocatalysts, i.e., pure g-C₃N₄ and graphene/g-C₃N₄ composites, are shown in Figure 1. There are two peaks of reflections for both materials. The peaks are most pronounced and centered around 13.1° and 27.4°, indicating very clearly that the graphitic-type crystal structure of g-C₃N₄ was formed [50,51]. The low-intensity peak appearing at 2θ = 13.1° can be indexed to the (100) plane and corresponds to a structural rearrangement pattern between the tri-s-triazine (heptazine) continuous planes with an interlayer distance of 0.676 nm [52,53]. Moreover, the strong peak at 2θ = 27.4° is attributed to a (002) plane, reflecting the periodic stacking of the conjugated aromatic structure with an interlayer distance of 0.33 nm [54,55,56]. Also, another low-intensity diffraction peak is observed at 2θ = 21.8° with a (002) plane, probably due to the low amount of graphene oxide remaining after the reduction of GO to graphene. Interestingly, Sengupta et al. [57] also reported the occurrence of two characteristic peaks corresponding to (002) and (100) planes in their investigation of the effect of temperature on the reduction of graphene oxide to graphene. These findings are comparable with the XRD data reported in the present work; therefore, it can be concluded that the photocatalysts under study were well synthesized and had a high structural crystallinity corresponding to pure g-C₃N₄ and graphene/g-C₃N₄. The presence of two diffraction peaks always existed in graphene/g-C₃N₄ with 2θ = 13.1° and 2θ = 27.4° implying a strong interaction between graphene and g-C₃N₄ [58].

Figure 2 depicts the SEM and TEM images of g-C₃N₄ and graphene/g-C₃N₄ photocatalysts. The SEM analysis of the g-C₃N₄ material revealed a typical flaky and ultrafine nanosheet structure (see Figure 2a). These nanosheets are curled to reduce the surface tension. This particular property is identical to the construction of the exfoliated carbon scrolls during graphite peeling [59,60]. The graphene/g-C₃N₄ composite exhibited a sandwich structure, which could explain the very close contact between the g-C₃N₄ and graphene (Figure 2b) [50,61]. Also, in the enlarged image of the targeted location (Figure 2b inset), it is clearly seen that the g-C₃N₄ lamellas are dispersed on the graphene surface, which may not only minimize the aggregation of g-C₃N₄ but may also increase the active sites of graphene/g-C₃N₄ for the photocatalytic reaction [58]. As for the TEM analysis, g-C₃N₄ exhibited a typical small, thin and flat irregular two-dimensional shape (see Figure 2c), while graphene/g-C₃N₄ showed no free graphene sheets (see Figure 2d) apart from the g-C₃N₄ sheets. This may suggest that the in situ growth method successfully integrated g-C₃N₄ with graphene.

Figure 3 depicts the FTIR spectra belonging to g-C₃N₄ and graphene/g-C₃N₄ materials. In the case of g-C₃N₄ material, a broad band around 3600 cm⁻¹ is observed and assigned to N–H stretching vibrations of non-condensed amino groups in the structure of g-C₃N₄ [54,62]. Moreover, the broad, strong peak around 3350 cm⁻¹ was attributed to hydroxyl groups (–OH) and physically adsorbed water [58]. The strong peaks at 1235, 1325, 1403, 1572 and 1637 cm⁻¹ correspond to the typical stretching vibrations of (C=N) and (C–N) [52]. In addition, the band around 812 cm⁻¹ corresponds to the bending modes of the tri-s-triazine backbone units [52]. For graphene/g-C₃N₄, the peak at 3350 cm⁻¹ (–OH) flattened slightly, thus signifying a partial removal of hydroxyl groups (–OH) [59]. Since the synthesis of graphene/g-C₃N₄ involves an in situ thermal reduction of the graphene oxide, the disappearance of characteristic absorption peaks belonging to graphene oxide may indicate the effective GO reduction, as previously reported [57]. The absorption peak at 1734 cm⁻¹ corresponding to the stretching vibrations of carbonyl groups (–C=O–) of GO was not observed. However, a weak peak at 1637 cm⁻¹ attributed to an epoxy group (–CO–) known by its high binding energy occurred, thus demonstrating that it is difficult to remove all the epoxy in graphene oxide following the thermal reduction [58]. The weak intensity of the band at 1403 cm⁻¹ was attributed to a partial removal of the carboxylic groups (–COOH) [57]. Furthermore, the peak at 1325 cm⁻¹ is responsible for the C=C backbone vibration in graphene. In summary, it can be concluded that the characteristic peaks of g-C₃N₄ did not change after graphene deposition.

Textural properties of the g-C₃N₄ and graphene/g-C₃N₄ materials were investigated using N₂ adsorption–desorption volumetric analysis. As illustrated in Figure 4, the obtained isotherms were attributed to the type IV with H3 hysteresis loop according to the IUPAC classification, thus indicating the occurrence of a mesoporous structure. The BET-specific surface areas of g-C₃N₄ and graphene/g-C₃N₄ were 54.4 and 74.8 m²/g, respectively. The occurrence of mesopores was corroborated by the pore diameters evaluated to be 3.4 and 4.9 nm for g-C₃N₄ and graphene/g-C₃N₄, respectively. The pore volumes reached 0.37 and 0.68 cm³/g for g-C₃N₄ and graphene/g-C₃N₄, respectively.

PL spectroscopy is a well-known technique to study both the interface charge carrier transfers and the electron–hole pair recombination. The PL emission spectra were thus employed to investigate the combination and separation of the photogenerated charges. Generally, the stronger PL intensity indicated the faster recombination of photogenerated charge carriers [63]. Figure 5a shows the PL spectra of g-C₃N₄ and graphene/g-C₃N₄. As seen, the PL intensity of graphene/g-C₃N₄ was lower compared with pure g-C₃N₄, implying that the recombination rate of electron–hole pairs is substantially slowed, thus improving charge separation. Meanwhile, the emission peaks maximum red-shifted from 494 nm to 528 nm, which could promote the π-electron delocalization of the graphene/g-C₃N₄ network. This is associated with the decreased band gap energy [63]. Time-resolved fluorescence spectroscopy of g-C₃N₄ and graphene/g-C₃N₄ was also performed to study the charge separation dynamics (See Figure 5b). The lifetime of the charge carriers is 4.18 ps for g-C₃N₄ material. Interestingly, this lifetime increased to 39.89 ps for graphene/g-C₃N₄, which is almost 10 times that of the charge carriers of g-C₃N₄, indicating a more efficient separation of electrons and holes in graphene/g-C₃N₄ compared to pure g-C₃N₄. Therefore, these photogenerated carriers of graphene/g-C₃N₄ were more stable and able to initiate redox reactions prior to recombination [64]. This behavior impacts the photocatalytic performance, which depends mainly on the number of active sites, the extent of visible light absorption and the efficiency of charge carrier generation–separation and transfer. As reported by Dong et al. [65], the improvement in the photocatalytic activity brought by the heterojunctions such as graphene/g-C₃N₄ is attributed to the improvement of the separation of electrons and photogenerated holes. The large number of active sites and direct transport or transfer pathways for fast charge carriers within nanostructures could significantly improve charge carrier separation.

The registered UV-Vis spectra are depicted in Figure 6a. These UV-Vis spectra were performed to characterize the visible light absorption capacity and optical band gap energy of g-C₃N₄ and graphene/g-C₃N_4, respectively. From Figure 6a, it is found that g-C₃N₄ effectively absorbs in the visible light region, and the absorption edge was at about 494 nm. After the incorporation of graphene into g-C₃N₄, the graphene/g-C₃N₄ photocatalyst showed a significantly enhanced absorption in the visible light region at about 528 nm. As shown in (Figure 6b), Tauc plots based on the following equation were used to evaluate the band gap energies:

$$\left(\mathsf{\alpha }\mathrm{h}\mathsf{\nu }\right)=\mathrm{A}{\left(\mathrm{h}\mathsf{\nu }-{\mathrm{E}}_{\mathrm{g}}\right)}^{\mathrm{n}/2}$$

where α represents the absorption coefficient, h is Planck’s constant, ν is the frequency of light, A is a constant that it is independent of energy, E_g is the band gap energy of a photocatalyst, and n = 4 for direct band gap of g-C₃N₄ [66].

Given that the value of n of graphitic carbon nitride is 4, the E_g of photocatalysts could be determined by the tangent lines versus (hν) [66]. Therefore, the obtained results indicated that the band gap energies of g-C₃N₄ and graphene/g-C₃N₄ were 2.51 and 2.35 eV, respectively. The band gap of g-C₃N₄ compares relatively well with that found in the literature [12]. These findings confirm that the combination of g-C₃N₄ semiconductor with graphene can decrease the band gap energy of the composite material and significantly increase the absorption in the visible light region, which could promote higher photocatalytic activity.

2.2. Photocatalytic Reaction

To assess the adsorption and photocatalytic performances, the removal of aqueous atrazine was carried out under dark and visible light irradiation in the presence of the photocatalysts. The obtained results indicated that the equilibrium was reached within ca. 20 min. Performances of g-C₃N₄ and graphene/g-C₃N₄ were investigated under similar conditions, as presented in Figure 7. Noticeably, the adsorption under dark conditions allowed the removal of ca. 25 and 50% of atrazine in the presence of g-C₃N₄ and graphene/g-C₃N₄, respectively. The subsequent photocatalytic degradation performed under visible light allowed to reach ca. 50 and 100% removal of atrazine within 5 h of reaction over g-C₃N₄ and graphene/g-C₃N₄, respectively. Clearly, the composite containing graphene largely outperformed the pristine g-C₃N₄. This improved photocatalytic activity under visible light may be attributed to the important role of graphene in the enhanced photoinduced electron conduction through the graphene sheets. Additionally, its two-dimensional, planar π-conjugation structure gives graphene excellent electron conductivity, allowing for rapid transport that results in efficient charge separation [67].

The effect of photocatalyst loading on the photodegradation of atrazine is presented in Figure 8a. The results showed that the photocatalytic conversion of atrazine ranged from 54% to 100% within 5 h of the reaction when the photocatalyst loading was increased from 0.05 to 0.4 g/L. The effect is due to the increase in the amount of photocatalyst, which increases the number of active sites of graphene/g-C₃N₄ photocatalyst. Interestingly, it was found that the loading of 0.4 g/L was optimum, leading to the best performances (see Figure 8b). Indeed, the photocatalytic degradation of atrazine decreased when the photocatalyst loadings were set beyond the optimum loading of 0.4 g/L, i.e., 0.6, 0.8 and 1 g/L. This effect may be attributed to the fact that the degradation reaction was hampered by the shielding effect occurring in the reactor when high photocatalyst loadings are used. Indeed, under these conditions, agglomeration of the photocatalyst particles may occur, thus generating an obstacle preventing light from reaching the surface of the photocatalyst. Therefore, an optimal photocatalyst loading is likely able to generate more hydroxyl radicals and superoxide reactive species, leading to better photocatalytic degradation of atrazine than high loadings [68,69]. Sharma et al. also noted that the excessive photocatalyst loadings during the reaction led to a decrease in photocatalytic activity [70].

pH is among the most important factors affecting the efficiency of many chemical and biological processes [71]. It also plays an important role in the production of hydroxyl radicals (OH•), which have been shown to oxidize many different recalcitrant organic pollutants into mineral end products [72]. The photocatalytic reaction medium pH referring to a specific concentration of protons (H⁺) or hydroxyl (OH⁻) ions in aqueous solutions can have an impact on the protonation or deprotonation of pollutants as well as photocatalysts surfaces, thus influencing the ability of the photocatalyst to adsorb the reactants, its activity, and the course of the reaction [73]. Therefore, the effect of pH on the degradation of atrazine in the presence of graphene/g-C₃N₄ composite was investigated. The original atrazine solution had a pH of 5.33, and the addition of hydrochloric acid and sodium hydroxide in solution regulated the other initial pH values (3.3, 2.5, 7.0, 8.6, 9.0 and 11.7). As shown in Figure 9, the conversion of atrazine decreased with increasing initial pH except for pH = 5.3, where there was no noticeable difference in the degradation of atrazine under acidic conditions. Rapid degradation occurred at low initial acidic pH at (pH = 3.3) and high at (pH = 5.3), where pH was the optimum, and where 85.2% and 100% of atrazine, respectively, was removed within 5 h of exposure time. Therefore, it can be concluded that acidic conditions are advantageous for the photocatalytic degradation of atrazine. Indeed, more H⁺ will bind to the surface of the graphene/g-C₃N₄ photocatalyst under acidic conditions, positively charging the surface. The single chloride atom of atrazine has an electron-withdrawing group and high electronegativity, allowing stronger adsorption by the photocatalyst. Xue et al. reported a BiOBr/UiO-66 composite which was very effective for the degradation of atrazine under visible light; they reported that this composite has the same reaction impact in view of the degradation with pH effect as the graphene/g-C₃N₄ composite. Their results show that the positively charged composite could enhance the electron transport on the surface of the photocatalyst and thus react with oxygen to produce a superoxide radical (O₂•⁻). In addition to this advantage, the recombination of charge carriers (e⁻/h⁺) could be decreased, and thus, the electron reacts with the with O₂ from the solution, obtaining (O₂•⁻) and efficiently oxidizing atrazine in the presence of graphene/g-C₃N₄ [74].

The heterogeneous photocatalysis allows the conversion of photonic energy (solar or artificial radiation) into chemical energy [75]. The photocatalyst (graphene/g-C₃N₄) absorbs visible light photon energy greater than or equal to its band gap energy. The electrons are excited to move from the valence band (BV) to the conduction band (BC), leading to the formation of electron–hole pairs (see Figure 10). The free charge carriers of the photocatalyst are positively charged holes (h⁺) in the valence band (BV) and electrons (e⁻) in the conduction band (BC). These charge carriers, the electron–hole pairs in the BC and BV, respectively reduce and oxidize the compounds adsorbed on the surface of the photocatalyst. However, the photogenerated electron–hole pairs have a very short lifetime of the order of 10⁻⁹ s, which means that the probability of the chemical reaction time of the photocatalyst with atrazine is very low. The rapid recombination strongly affects the photocatalytic activity. The mechanism of heterogeneous photocatalysis of a g-C₃N₄ deposition on graphene can be explained as follows: (1) increased absorption of visible light, (2) separation and transfer of photoinduced charge carriers to the surface of the photocatalyst, and (3) reaction between the charge carriers (e⁻/h⁺) and atrazine molecules. A schematic diagram of the mechanism of photodegradation of atrazine is shown in proposed in Figure 10). According to Du et al. [67], electrons can be directly excited from the valence band of graphene to the conduction band of g-C₃N₄. Indeed, graphene is a conductor that allows electrons to move freely. The energy of g-C₃N₄ in the conduction band is more negative than that of graphene. Therefore, photoinduced electrons can be transferred from the conduction band of g-C₃N₄ to graphene. The electrons of g-C₃N₄ can reduce the oxygen O₂ present in the solution into active superoxide O₂•⁻ since the redox potential NHE (O₂/O₂•⁻) is −0.33 eV). These electron–hole interlayer transfers between g-C₃N₄ and graphene extend the lifetime of the charge carriers. The valence band energy of g-C₃N₄ is more positive than the valence band energy of graphene, but the valence band energy of g-C₃N₄ (+1.57 eV) is less positive than the potential of (•OH/H₂O = +1.99 eV), which does not allow the EVB holes of g-C₃N₄ to oxidize H₂O and produce OH• hydroxyls. The adsorbed atrazine on the surface of the photocatalyst decomposes by transferring electrons (e⁻) to the positive holes (h⁺) on the valence band of graphene/g-C₃N₄. These charge carriers trigger oxidation/reduction reactions that degrade the adsorbed contaminants (atrazine) and transform them into harmless products such as CO₂, H₂O, and other by-products such as inorganic acids. The equations of the photocatalytic reaction of atrazine over graphene/g-C₃N₄ are proposed together with the illustration of this mechanism (Figure 10), in agreement with the findings reported in the literature [50].

$$Graphene+g-C_3{N}_4+\mathbf{h}\mathsf{\nu }\to (e^{-})+\left(h^{+}\right)\hspace{1em}\hspace{1em}\mathrm{Generation} \mathrm{of} \mathrm{electron}–\mathrm{hole} \mathrm{pairs}$$

$$\left(e^{-}\right)+O_2\to O_2^{•-}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\mathrm{Oxygen} \mathrm{reduction} \mathrm{to} \mathrm{superoxide}$$

$$g-C_3{N}_4\left(e^{-}\right)+O_2\to g-C_3{N}_4\left(e^{-}\right)+O_2^{•-}\hspace{1em}\hspace{1em}\mathrm{Reduction} \mathrm{of} \mathrm{oxygen} \mathrm{to} \mathrm{superoxide}$$

$$\begin{gathered} Atrazine+O_2^{•-}+\to C{O}_2+H_{2}O+other endproducts\hspace{1em}\mathrm{atrazine} \mathrm{degrades} \mathrm{photocatalytically} \\ \mathrm{with} \mathrm{superoxide} \mathrm{radicals}. \end{gathered}$$

Stability and reusability are crucial for a heterogeneous photocatalyst. To assess the stability of the graphene/g-C₃N₄ photocatalyst in the photocatalytic degradation of atrazine, five reaction cycles were carried out under the optimum conditions of the above-mentioned reactions, i.e., photocatalyst loading of 0.4 g/L in contact with 10 mg/L atrazine solution, reaction duration 5 h and pH = 5.3, using the same photocatalyst batch. After each run, the photocatalyst was collected and rinsed several times with distilled water and dried for 24 h before the beginning of the next cycle. As depicted in Figure 11, it was found that the graphene/g-C₃N₄ composite, as designed, reveals stability and reusability after five cycles of recycling, as there was no significant loss of photocatalytic activity for atrazine degradation. It can be concluded that the graphene/g-C₃N₄ photocatalyst is of good quality as it maintains excellent stability and photocatalytic activity. Additionally, the conditions of its reuse require only a simple rinsing with pure water followed by drying, thus facilitating its long-lasting and efficient reuse.

Table 1. Different photocatalysts performances in the degradation of aqueous atrazine under visible light at ambient temperature.

Initial Concentration (mg/L)	Photocatalyst (Loading g/L)	Time (min)	Degradation (%)	Ref.
100	g-C3N4 (0.4)	60	10	[45]
100	AgCl/g-C3N4 (0.4)	60	100	[45]
10	ZnO–g-C3N4/MoS2 (0.3)	300	80	[46]
10	CdS/RGO/g-C3N4 a (0.2)	300	90	[47]
10	AC/g-C3N4 b (1)	120	90	[48]
10	Vc OCN c (1)	120	100	[49]
1	TEA-CN d (0.3)	60	90	[77]
10	PDI-g-C3N4/g-C3N4 e (0.8)	60	75	[18]
10	Graphene/g-C3N4 (0.3)	300	100	This work

^a RGO: reduced graphene oxide. ^b AC: Activated carbon. ^c V_c-OCN: Carbon vacancy-oxygen doped graphitic carbon nitride. ^d TEA-CN: Triethanolamine modified carbon nitride. ^e PDI-g-C₃N₄/g-C₃N₄: pyromellitic diimide modified g-C₃N₄/g-C₃N₄ homojunction.

presents recent investigations reported in the literature regarding the photocatalytic degradation of aqueous atrazine using graphitic carbon nitride-based photocatalysts. Compared to the reported findings, the graphene/g-C₃N₄ photocatalyst performed very well, as complete degradation of atrazine was reached within a reasonable reaction time. Moreover, even though the investigations reported in Table 1 were promising, the procedures used for the modification of the graphitic carbon nitride were either complicated and time-consuming or involved the use of toxic metals such as molybdenum and cadmium, which may leach in the reaction medium, thus leading to serious environmental threats. Additionally, it is worth mentioning that the solubility of atrazine in water is around 30 mg/L at 20 °C [76]; therefore, the initial concentration of 100 mg/L reported by Mohamed (2015) is certainly inaccurate. Moreover, in the case of the pyromellitic diimide-modified g-C₃N₄/g-C₃N₄ homojunction, peroxymonosulfate was used to activate the photocatalyst (see reference [18]).

3. Experimental Methods

3.1. Materials

The chemicals used in this study were of a very high analytical quality and without the need for further purification. Atrazine (C₈H₁₄ClN₅, 99%) and terbuthylazine (C₉H₁₆ClN₅, 99%) were supplied by Sigma-Aldrich (Oakville, ON, Canada). Urea (CH₄N₂O, 99%) was purchased from BDH Chemicals (Toronto, ON, Canada). The solvent acetonitrile (C₂H₃N, 99%), anhydrous magnesium sulfate (MgSO₄, 98%), hydrogen chloride (HCl, 1N) and sodium hydroxide (NaOH, 1N) were acquired from Fisher ScientificChemicals (Ottawa, ON, Canada). For GO (graphene oxide) synthesis, natural graphite flakes (325 mesh, 99.8%) from Alfa Aesar (Ottawa, ON, Canada); potassium persulfate (K₂S₂O₈ ≥ 99%) from Anachemia (Montréal, QC, Canada); phosphorus pentoxide (P₂O₅ ≥ 99%) from JT Baker (Mississauga, ON, Canada); sulfuric acid (H₂SO₄, 98%) and potassium permanganate (KMnO₄ ≥ 99%) from Fisher Scientific Chemicals (Ottawa, ON, Canada); hydrogen peroxide solution (H₂O₂, 30% by weight in water) from EMD Millipore (MilliporeSigma, Oakville, ON, Canada); hydrochloric acid (HCl, 36%) from Fisher ScientificChemicals (Ottawa, ON, Canada) and ethylene glycol (≥99%) from BDH (West Chester, PA, USA) were used. Millipore water (18 MΩ·cm) was used in the present investigation.

3.2. Synthesis of Pure g-C₃N₄ and Graphene/g-C₃N₄ Composite

GO was first prepared as reported by Yamaguchi et al. [78]. The composite graphene/g-C₃N₄ was synthesized as reported by Zhu et al. [79]. In a typical synthesis, 4 g of urea were dissolved in 50 mL of deionized water and put under magnetic stirring for 1 h. Then, 2 mg of GO was suspended in 50 mL of deionized water and ultrasonicated for 4 h so that a complete dispersion of the GO sheets was obtained. The urea solution and the GO suspension were then mixed. The mixture was ultrasonicated for 3 h and magnetically agitated for 3 h, and then dried in the oven at 80 °C for 24 h. The obtained solid was then calcinated at 550 °C in a covered crucible for 2 h at a heating rate of 10 °C·min⁻¹. The final product of a gray color was denoted graphene/g-C₃N₄. The synthesis of pristine g-C₃N₄ was performed using urea as a precursor. Typically, 5 g of urea was dissolved in 100 mL of deionized water under ultrasonic and magnetic stirring for 1.5 and 1 h, respectively. The solution was dried at 60 °C for 12 h, then calcinated at 550 °C in a covered crucible for 2 h at a heating rate of 10 °C·min⁻¹. The obtained yellow powder was designated g-C₃N₄.

3.3. Photocatalysts Characterization

The crystal structure of the as-prepared samples was analyzed using X-ray diffractometer Rigaku MiniFlex 6G (The Woodlands, TX, USA) operating with copper radiation (Kα = 1.5406 Å) at a power of 600 W (40 kV, 15 mA). Powder diffraction patterns were obtained between 5° and 60° with a scan speed of 5°/min. The morphology and size were examined via scanning electron microscopy (SEM) using a JEOL 6360LV microscope (Tokyo, Japan) with an acceleration voltage ranging from 5 to 25 kV, with the Secondary Electron Imaging signal (SEI). For the Transmission Electron Microscopy (TEM), a JEOL 1230 (Tokyo, Japan) equipped with a Gatan 1000XP bottom-mounted camera was used with an acceleration voltage of 80 kV. A HP UV-8453 UV-visible spectrophotometer using MgO as a background reference was used to record the UV–visible diffuse spectra of the synthesized samples. The photoluminescence (PL) spectra were obtained using a fluorescence spectrometer Horiba Quanta-Master 8000 (Edison, NJ, USA). Horiba scientific software was used to determine the rate of electron transfer. A decay mathematical approach was used, where τ represents the measured lifetime value. Fourier-transform infrared spectra (FTIR) were recorded between 300 and 4000 cm⁻¹ on a Varian 1000 spectrometer Scimitar series (Foster City, CA, USA) with a resolution of 2 cm⁻¹ using KBr pellets at room temperature. The textural properties of the samples were studied using nitrogen adsorption/desorption analyses at 77 K using a volumetric adsorption analyzer (model Autosorb-1, Quantachrome Instruments, Boyton Beach, FL, USA). The specific surface area was evaluated from the linear part of the Brunauer–Emmett–Teller (BET) plot (P/P₀ = 0.05–0.30). Pore size distribution curves were determined from the desorption branch using the Barrett–Joyner–Hallenda (BJH) method. The total pore volume was estimated from the amount adsorbed at a relative pressure P/P₀ = 0.99 single point.

3.4. Photocatalytic Activity Tests

The aqueous atrazine (10 mg/L) was photodegraded at ambient temperature in a 48 W photocatalytic reactor operating under visible light at a wavelength λ > 420 nm. The tests were operated batchwise under the following ranges of conditions: photocatalyst loading 0.05–1 g/L, pH 3.3–11.7 and time 0–300 min. The reacting volume was 5 mL for all the tests performed under agitation (750 rpm). To avoid any disturbance due to multiple liquid sampling on kinetic measurements, a separate experiment was used for each measurement. Furthermore, tests were performed in the dark to assess the adsorption occurring on the photocatalysts. At pre-set reaction times, aliquots of the aqueous solution were withdrawn, filtered with a microsyringe (0.2 μm), solvent extracted and analyzed for atrazine content using a HP 6890 Series gas chromatograph equipped with a flame ionization detector (Thermo Fisher Scientific, Markham, ON, Canada). The column used was an HP-5 (30 m × 0.25 mm ID × 0.25 μm film thickness). The injector and detector temperatures were set to 250 °C and 1 μL sample was injected with a split ratio of 50:1. The oven temperature profile was: initial temperature 40 °C for 2.5 min, ramp at 25 °C/min to 200 °C, hold for 5 min; ramp at 25 °C/min to 250 °C, hold for 1 min.

In addition, recycling experiments were performed for five consecutive cycles to test the stability and reusability of the photocatalysts. After each cycle, the photocatalyst was separated via centrifugation and gently washed with pure water several times to remove any residual atrazine impurities and then dried at 80 °C for the next test.

All the tests reported in the present investigation were performed in duplicate and analyzed twice.

4. Conclusions

Photodegradation of atrazine was achieved using urea-based graphitic carbon nitride (g-C₃N₄) and graphene from graphene oxide pyrolysis. The photocatalytic activity of graphene/g-C₃N₄ was found to be better than that obtained via bare g-C₃N₄ towards the degradation of atrazine. This was attributed to the fact that graphene increases the lifetime of electron–hole charge carriers (e⁻/h+) and improves the creation of adsorption sites in the graphene/g-C₃N₄ network, thus favoring the reduction of the energy gap by increasing the population of electrons present in the solution. It was also noticed that this reduces the recombination of electron–hole pairs and improves the charge separation. The overall results showed a complete degradation of 100% of atrazine within 300 min at room temperature and under visible light irradiation, even after five cycles of photocatalyst recycling. The effects of pH conditions and photocatalyst loading on the degradation extent of atrazine using optimally synthesized graphene/g-C₃N₄ were investigated. Interestingly, the normal (uncontrolled) pH value of the atrazine solution and a moderate catalyst loading of 0.3 g/L led to the best degradation performances.