Active oxygen species (AOS) are effective green oxygenants for environmental-pollution control. AOS have many types, including superoxide radical (^·O₂^–), hydrogen peroxide (H₂O₂), and hydroxyl radical (^·OH) [1-8]. To remove environmental pollution, many methods such as electrochemistry, Fenton reaction, and photocatalysis have been developed to produce ^·OH. Electrochemistry method requires substantial electric energy, and Fenton reaction requires extra H₂O₂ [9-11]. Moreover, these two methods may harm the environment. For example, electrochemistry method may release metal ions and harmful gas, and Fenton reaction may produce iron sludge [12, 13]. Photocatalytic technology has sufficient energy source (sunlight) and is environmentally friendly [14]. Meanwhile, photocatalytic technology could produce ^·OH under mild conditions. Thus, photocatalytic ^·OH production has the potential for practical applications [15]. However, the ^·OH production efficiencies of photocatalysts remain at a low level because the recombination rate of photogenerated electron and hole is relatively high. Many methods have been developed to solve this problem. For example, the photocatalyst can be modified by various modification methods (e.g., doping, heterojunction formation, and noble-metal deposition) to suppress the recombination of photogenerated electron and hole, or a sacrificial agent may be added to improve the redox half-reaction [12, 16-18]. However, photocatalyst modification is a complicated process [19-21], and adding a sacrificial agent may cause environmental pollution. For example, Xiao's team attempted to improve the photocatalyst with ozone (O₃) as a sacrificial agent [5, 7, 8, 15]. O₃ is known to be harmful to human health and the environment because of the lung damage it inflicts. Thus, developing a new method of improving the photocatalytic production of ^·OH is very important.

Nitrogen oxide (NO_x), a major pollutant in the atmosphere, originates from the combustion of fossil fuel in boilers and motor vehicles (exhaust gas) [22]. The N atom of NO has an unpaired electron, so NO possesses strong reducing property (Fig. S1 in Supporting information) [23, 24]. Therefore, we believe that NO can be used as a sacrificial agent for photogenerated hole, meaning that photocatalytic AOS production can be improved by NO bubbling. Meanwhile, NO can be oxidized to NO₂ by using a photogenerated hole, and then the generated NO₂ is absorbed by water (became NO₃⁻) [25]. In this case, the NO in the exhaust gas can be used and further removed. To verify the above analysis, the visible-light photocatalyst g-C₃N₄ was used to produce ^·OH through NO bubbling. We found that the introduction of NO could improve the photocatalytic ^·OH generation activity of g-C₃N₄. In this process, the structure and activity of g-C₃N₄ are very stable [22, 26]. The improvement mechanism was studied systematically. The details can be found as below.

We detected the generation of active oxygen species (AOS) with or without NO gas (500 ppb, diluted by air stream) bubbling. Figs. 1a and b show the EPR signal of ^·O₂^–. No clear ^·O₂^– signal was observed in darkness. After the light was turned on, ^·O₂^– can be produced in g-C₃N₄ suspension and g-C₃N₄/NO suspension. However, the bubbling of NO did not significantly affect the intensity of generated ^·O₂^–. Figs. 1c and d display the generated H₂O₂ under dark and light-irradiation conditions. H₂O₂ can be produced only under light irradiation in the presence of g-C₃N₄ sample. Similarly, with ^·O₂^–, NO bubbling did not significantly affect the intensity of generated H₂O₂. As for ^·OH, the effect of NO bubbling was highly significant. Fig. 1e shows that ^·OH was undetected in darkness. After the light was turned on, ^·OH can be produced in g-C₃N₄ suspension and g-C₃N₄/NO suspension (Fig. 1f). Surprisingly, the concentration of generated ^·OH in the g-C₃N₄/NO system was nearly twice as high as that in the g-C₃N₄ suspension. Therefore, NO bubbling helped g-C₃N₄ produce more ^·OH under light irradiation (see Supporting information for details).

To understand the mechanism of NO in inducing g-C₃N₄ to produce more ^·OH under light irradiation, we analyzed the converted products of NO during photocatalysis. In the photocatalytic reaction of NO in the gas-solid phase (Eq. 1 and Fig. 2a), most reaction products were NO₂ [3]. However, the NO₂ signal disappeared when the hole scavenger was added, suggesting that NO can be oxidized to NO₂ by hole. This can be further proved by the hole capture experiment (Fig. S3a in Supporting information). Interestingly, NO₂ cannot be detected when the photocatalytic reaction proceeded in the liquid-solid phase (Fig. 2b). This result suggested that the generated NO₂ decomposed (Eq. 2) or absorbed (Eq. 3) in water. If NO₂ decomposes, the generated oxygen free radicals (^·O) would react with water to produce ^·OH [27]. To verify this mechanism, NO_x was bubbled in water (without the presence of photocatalysts) under light irradiation (λ = 420 nm). The results are displayed in Fig. 2c. ^·OH can be detected in the NO₂-bubbled water but not in the NO-bubbled one. These results confirmed Eqs. 2 and 4. Accordingly, NO₂ photolysis can contribute to the increase in ^·OH yield. Fig. 2d shows that NO₂ photolysis (70 ppb) could produce only 3.33% of the ^·OH increment. Thus, other reasons for the ^·OH increment besides NO₂ photolysis must exist. The photolysis-mechanism diagram of NO₂ is shown in Fig. 2e. Extra NO₂ would be absorbed by water and became NO₃⁻. Besides, the conversion rate of NO to NO₃⁻ is 47%. (Fig. S4 in Supporting information).

(1)

(2)

(3)

(4)

Since NO could consume hole and inhibit the recombination of photogenerated carriers. Interestingly, NO bubbling obviously improved the photocurrent signal of g-C₃N₄ under the same condition. And the further confirmation could be found in Fig. S3b (Supporting information). This phenomenon ensured that NO could inhibit the recombination of photogenerated carriers. Consequently, the valence band of g-C₃N₄ can provide more excitation electrons for the higher fluorescent signal (Fig. S3c in Supporting information). For a visual expression, the reaction-mechanism diagram is shown in Fig. S3d (Supporting information). The more consumption of photogenerated hole could be created, the more adsorbed O₂ molecules could be reduced.

Based on the redox potential of H₂O₂/O₂ (0.659 V) [28] and the valence-band potential of g-C₃N₄ (1.4 eV) [29], H₂O₂ may be oxidized to O₂ by the photogenerated holes (Eq. 5). This speculation can be proven by H₂O₂ catalytic-decomposition experiment. (Figs. 3a and b and S2.3 in Supporting information). To investigate the effect of NO on H₂O₂ catalytic decomposition, NO was introduced in the above experiments. Fig. 3c shows that the amount of produced O₂ significantly decreased after NO introduction, indicating that NO can inhibit the oxidation reaction of H₂O₂ caused by photogenerated holes consumed. Meanwhile, the redox potential of NO/NO₂ (0.428 V) was smaller than that of H₂O₂/O₂ (0.659 V), which can be proven by the cyclic voltammetry curve in Fig. 3d. In this case, H₂O₂ can be protected by NO from the oxidation of photogenerated holes, just like the drawing in Fig. 3e. Therefore, more H₂O₂ can be used to produce ^·OH (Eq. 6).

(5)

(6)

After ^·OH generation, it diffused in water. However, NO is insoluble in water, so NO cannot consume the ^·OH generated in our system. When NO was oxidized to NO₂ by photogenerated holes, NO₂ was adsorbed by water and became NO₃^– and NO. However, NO₃^– cannot be further oxidized, so it cannot consume the generated ^·OH. In this case, NO₂ did not consume the generated ^·OH in our system. Based on the above experimental results and discussion, although NO was used as the hole sacrificial agent, and it did not consume ^·OH. For this reason, our system could accumulate more ^·OH for the degradation of organic pollutants [30] (Fig. S5 and S2.4 in Supporting information). Herein, Fig. 4 can be well matched with this theme.

Stability is an important evaluation indicator for the application value of one technology or sample [31]. Accordingly, cycling experiments were conducted to investigate the stabilities of this technology and g-C₃N₄ sample [32, 33]. Fig. S6a (Supporting information) shows that the ^·OH yield did not decrease after five cycles of use. Meanwhile, the degradation rate of tetracycline performed using our technology was very stable throughout the five cycles of repeat experiments (Fig. S6b in Supporting information). These results suggested that our technology was stable and recyclable. The stability can be further confirmed by the XRD, FT-IR, UV-DRS, and SEM characterizations. A detailed analysis can be found in Figs. S6c, S6d, S7 and S8 (Supporting information).

In conclusion, the introducing NO could significantly enhance the photocatalytic ^·OH production ability of g-C₃N₄. After a series of mechanism studies, we found that there were four reasons for this improvement: (1) The ^·O produced by the NO₂ photolysis can react with H₂O to form ^·OH; (2) NO can be used as a hole-capture agent, which makes g-C₃N₄ generate more photo-generated electrons to reduce O₂; (3) H₂O₂ can be protected by NO from being oxidized by holes; (4) ^·OH cannot be consumed by NO or NO₂. Since NO could enhance the ^·OH production, NO was introduced in the photocatalytic tetracycline removal process. It was found that NO could enhance the mineralization rate of tetracycline from 59% to 86%. Moreover, the presence of NO in the photocatalytic process does not affect the stability of g-C₃N₄.

Declaration of competing interest

The authors report no declaration of competing interest.

Acknowledgments

Financial support by the National Nature Science Foundation of China (Nos. 21876104 and 21603271) is gratefully appreciated. The work described in this paper was also partially supported by the General Research Fund (Nos. 18300920 and 18301117) of Research Grants Council, the Department Collaborative Fund (No. 04490), Dean's Research Fund (No. FLASS/DRF 04554) of the Faculty of Liberal Arts and Social Sciences, The Education University of Hong Kong, Hong Kong Special Administrative Region, China.

Supplementary materials

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2021.12.071.