2.4.2. Efficient Nitrate Removal from Wastewater over Different Materials

Photocatalytic nitrate reduction is one of the emerging transformative technologies capable of yielding harmless gaseous products. The eCB−/hVB<sup>+</sup> pair recombination is the main drawback of photocatalytic processes and affects their overall efficiency. The strategies for improving the charge carrier separation have already been reported and discussed thoroughly in several reviews [117,135–138]. It is widely accepted that NO2 − is the first stable intermediate product obtained from nitrate reduction, and it can remain in solution as NO2 − or undergo further reduction to N2 or to NH4 +. Some authors have been unable to detect quantifiable amounts of NO2 − at the end of the photocatalytic reduction of NO3 − because it can be easily reduced after its formation [139–143]. Because of its faster reduction, several works study the direct reduction of NO2 −. The reduction of nitrite is the divergent point that defines the selectivity towards harmless N2 or undesired NH4 +.

Few studies quantify the direct yield towards N2. The quantification of N2 by gas chromatography was reported by Kominami et al. [141] and by Zhang et al. [144]. Experimental work by Zhang et al. [144] concluded that N2 was the only gas product released using an Ag/TiO2 photocatalyst. N2O has been identified as an intermediate released in other reductive treatments, such as hydrogenation [145–147]. Even though N2 is an inert species, other nitrogen-containing gas species such as N2O, NO, and other NxOy are hazardous species with high environmental implications in atmospheric chemistry [148,149]. The last major product obtained during NO3 − and NO2 − reduction is the ammonium cation, released from an alternative pathway to the one of the HNO intermediate. The selectivity of nitrate depends on the ratio of surface coverage of N species to reductant species. A low coverage or high concentration of reducing mediators could deteriorate the selectivity for the formation of NH4 +. An appreciable pseudo-concentration of adsorbed nitrogen intermediates, mainly HNO and NO•, would favor the pathway leading to N−gas species. Many reactions are highly pH−dependent; therefore, acidic pH is necessary to ensure sufficient H+ to allow fast kinetic rates. Therefore, the pH dependence is not only related to the surface charge that modulates the adsorption of species on the photocatalyst surface but also to the H+ source to ensure the complete reduction.

Doped semiconductor photocatalysts can provide higher conversion of nitrate and selectivity to nitrogen gases than pristine TiO2, but results are comparable to composite catalysts, raising the question of whether interstitial/substitutional doping is necessary compared to photo deposition methodologies [150]. An excessive metal loading in the composite can be detrimental, becoming a recombination center instead of the desired electron sink [151]. Optimizing metal loads to about 1.0% *w/w* typically minimizes this detrimental effect [152]. Monometallic composites are the most prevalent types reported in the literature. The metals' performance was also related to the intrinsic capability of each platinoid to stabilize Hads because materials with higher overpotential for H2 evolution presented a predominant yield of NH4 +. The most influential factors affecting the efficiency of proton reduction on the metal surface are (i) the hydrogen overpotential during water splitting and (ii) the Hads stabilization [153]. In support of this hypothesis, Hamanoi et al. [154] proved experimentally that a decrease in NO3 − conversion is observed when hydrogen evolution is increased. Furthermore, bubbling H2 enhances the reduction of NO3 − to NH4 +, demonstrating that the adsorption of H2 on platinoids surface catalytic sites as Hads contributes to the reduction process [117,127].

Photocatalysts based on pristine TiO2 and related mono/bimetallic composites or bimetallic systems are intensively studied [155–158], but conventional metal-modified photocatalysts usually suffer from metal leaching, aggregation, and gradual deactivation and need to be significantly improved in terms of N2 selectivity. New materials were developed or used for the first time for photocatalytic denitration (perovskite-based photocatalysts, layered double hydroxides (LDHs) with hydrotalcite-like structures, nonlinear optical material LiNbO3) [159].

It is well known that photocatalytic oxidation has been investigated extensively for its capability of producing highly oxidative •OH, but little attention has been paid to the photocatalytic reduction of oxidative pollutants such as nitrate in water. Photocatalytic denitrification appeared as a feasible approach to accomplish this aim since it was first reported by Schlögl and co−workers in 1999 [160]. During the photocatalytic denitrification process activated by light irradiation, the photocatalyst generates electrons (eCB−) in the conduction band (CB) and holes (hVB+) in the valence band (VB) of the semiconductor. Then, the nitrate is reduced through direct interaction with eCB− or reaction with reductive CO2•<sup>−</sup> radicals produced from the reaction between hVB<sup>+</sup> and hole scavengers (e.g., formic acid) [161–163]. According to literature, the latter mechanism generally rules the photocatalytic denitrification for several materials like conventional TiO2, ZnO, ZnS, CdS, and SrTiO3 [128,164–168]. It is difficult to control the formation of CO2•<sup>−</sup> radicals due

to the dependence on the used hole scavengers. Liu et al. [169] reported the photocatalytic denitrification by nonlinear optical (NLO) material, i. e. lithium niobate (LiNbO3) in the presence of formic acid (FA) serving as a hole scavenger. A 110 W high−pressure Hg lamp was employed as a 365 nm UV light source. LiNbO3 achieved 98.4% total nitrate removal and 95.8% N2 selectivity under neutral pH conditions. During the process, the nitrate may be reduced by (i) reductive CO2•<sup>−</sup> radicals produced from the reaction between hVB+/•OH and hole scavengers, (ii) electrons generated at CB, as well as (iii) hydrogen produced from water splitting at CB. They concluded that photocatalytic denitrification should be dominated by reactions involving the conduction band (CB) either through interaction with electrons or hydrogen produced from water splitting. They showed that the role of H2 is very limited, and more than 98% of NO3 − is reduced directly by electrons at the conduction band of LiNbO3.

Photocatalytic denitration using various non-toxic hole scavengers is the most common technique reported in the literature. The photocatalytic degradation of nitrates in an aqueous solution has been examined by Anderson and co-workers [118], using different Au/TiO2 photocatalysts and oxalic acid as a hole scavenger. It has been shown that oxalic acid and nitrate can be simultaneously degraded over Au/TiO2 to produce predominantly CO2 and nitrogen, but complete nitrate removal was not achieved.

Luiz et al. [139] studied TiO2 and TiO2 doped with Zn2+, Cu2+ and Cr3+ (metal doped-TiO2 (Cu-TiO2, Cr-TiO2 and Zn-TiO2).) The prepared materials were used as photocatalysts to reduce nitrate and oxidize formic acid under the irradiation of a low-pressure mercury lamp (UV radiation at 254 nm, output power of 17 W). The results obtained from the nitrate photoreduction experiments indicated that the metal-doped TiO2 activity decreases in the order: 4.4% Zn-TiO2 > 4.4% Cu-TiO2 > 4.4% Cr-TiO2. Zn-TiO2 exhibits the greatest selectivity towards N2 (95.5%), a nitrate conversion up to 92.7%, and a high reaction rate (14.2 μmol NO3 <sup>−</sup> (min gcatalyst) <sup>−</sup>1).

Doudrick et al. [128] examined the photocatalytic reduction of nitrate in water using titanium dioxide (Evonik P90) loaded with silver nanoparticles and formate as a hole scavenger (electron donor). Photocatalytic experiments were performed using a UV light source (450-W medium-pressure mercury–vapor lamp). Under acidic conditions (pH = 2.5), nitrogen gases (~85%) and ammonium (~15%) were the final by-products. The authors evidenced that radicals are unlikely to be responsible for nitrate reduction, so a photocatalyst with the proper Fermi level must be selected to meet the thermodynamic requirements. Because the pH was a factor in their experiments, proton localization at the reaction sites was important for treatment at ambient pH and for achieving harmless by-products, which can be accomplished by selecting the proper co-catalysts (e.g., Ag, Cu). Although photocatalysis is not fully suitable for drinking water applications yet, P90/Ag removes nitrate efficiently and with high selectivity.

There are some reports presenting the photocatalytic reduction of NO3 − in the absence of sacrificial agents, but their activity is not sufficiently satisfactory [130,157,170–180]. Wei et al. [181] aimed to develop photocatalysts for the chemical reduction of NO3 − in visible light (fluorescent lamps irradiated with a power intensity of 2.64 mW/cm2, λ = 419 nm) and in the absence of sacrificial agents. The use of Ni2P as a potential base-metal alternative to precious metals as a catalyst for the hydrogenation of NO3 − under mild, near-ambient conditions (1 atm, 60 ◦C) has been demonstrated [182]. This potential catalyst exhibited complete NO3 − reduction with very high selectivity for ammonia (NH3) [182]. Considering that the light source and the photocatalyst are two key factors in the photocatalytic reduction of NO3 −, Wei et al. [181] synthesized Ni2P/semiconductors (Ni2P/Ta3N5, Ni2P/TaON, and Ni2P/TiO2) and used these heterostructures as photocatalysts for the reduction of NO3 − in water (Figure 15a,b).

Starting with a 2 mM (28 g/mL NO3 −-N) solution at pH 2, Ni2P/Ta3N5 and Ni2P/TaON achieve 79% and 61% NO3 <sup>−</sup> conversion, respectively, and conversion rates of 196 μmol g−<sup>1</sup> h−<sup>1</sup> and 153 μmol g−<sup>1</sup> h−1, respectively, after 12 h under 419 nm irradiation. Control experiments confirmed that Ni2P/semiconductor heterostructures and light illumination are

requisites for the photocatalytic reduction of NO3 −. Based on these findings, Wei et al. proposed two possible electron migration pathways and assumed that the dominant pathway in these heterostructures is light absorption by the semiconductor followed by electron injection into Ni2P (Figure 15c,d) [181].

**Figure 15.** (**a**) Half and overall reactions for photocatalytic reduction of NO3 − to the most desirable product(s) N2 and/or NH3. (**b**) Mechanistic scheme showing energy flow during the photocatalytic reduction of NO3 − over a Ni2P-modified semiconductor. (**c**,**d**) Schematic illustrations of two possible charge separation pathways during photocatalysis. In mechanism (**a**), light absorption by the semiconductor results in photo-generated electrons and holes, with the electrons getting trapped by Ni2P. The resulting Ni2P Fermi level upshift leads to a higher driving force for NO3 − reduction. In mechanism (**b**), photogenerated "hot" electrons from Ni2P are injected into the semiconductor. Reproduced and adapted with permission from ref. [181]. Copyright 2020 John Wiley and Sons.

Silveira and co-workers [134] presented the promising use of FeTiO3 and oxalic acid as reducing agents for the selective photo-reduction of nitrate to N2. They studied the feasibility of using natural ilmenite as a catalyst for NO3 − photo-reduction with oxalic acid as a reducing agent. The generation of NOx(g) via NO3 − and NO2 − reduction is also observed. The complete NO3 − and C2O4 <sup>2</sup><sup>−</sup> removal and a selectivity towards N2 > 93% was achieved by using the stoichiometric C2O4 <sup>2</sup><sup>−</sup> amount after 210 min, without the generation of undesirable NH4 +.

Formic acid is known as one of the most efficient hole scavengers for nitrate reduction. The deep reduction to N2 is quite difficult because the process requires a significantly high density of electrons at the catalytic sites. Yue et al. [120] proposed that the reactions occur on the surface of the catalyst particles, as shown in Figure 16.

The authors [120] systematically investigated the performance of CuInS2 in photocatalytic nitrate reduction under visible light irradiation by loading co-catalysts. A 300 W Xe lamp was used to provide visible light irradiation. Band-pass or cut-off filters were applied to obtain monochromatic beam light (λ = 400, 450, 500, 550, 600, or 650 nm) or pure visible light (λ > 400 nm), respectively. In particular, with the assistance of the LSPR effect of Au, the high record of the nitrate conversion rate of 8.32 mg N h−<sup>1</sup> was achieved under pure visible light. Overall, CuInS2 holds high potential in the application of photocatalytic nitrate removal under solar irradiation. Yue et al. [120] advanced the idea that the reaction mechanism takes place via adsorption–reduction reactions where nitrate ions are reduced directly by photo-generated e−. This mechanism is supported by the fact that the introduc-

tion of additional halide anions in an aqueous solution reduces the photocatalytic efficiency due to the competition between the adsorption of ion species. Yue and co-workers [120] The efficiency of solid formate as a hole scavenger was evaluated. It was shown that both glucose and sucrose appear to be effective, with only a slight decrease in photocatalytic efficiency. On the contrary, harmful molecules (including benzene, phenol, and benzoic acid) and other typical h+ scavenger agents (such as methanol and ethylene glycol) were also applicable with the rationale of simultaneously decomposing two pollutants.

**Figure 16.** Photocatalytic nitrate reduction on CuInS2 loaded with co-catalysts in the presence of sacrificial agents. The abbreviations (aq), (ads), and (g) mean ions in an aqueous solution, adsorbed on the surface and in the gas form, respectively. Reproduced with permission from ref. [120]. Copyright 2016 Royal Society of Chemistry.

Titania (TiO2) and metal-loaded titania using Pt [123,151,165], Pd [141,151,165], Rh [151], Ru [151], Au [118,121], and Ag [127,128,144] are widely used and effective in the reduction of nitrate with high selectivity toward N2 [125,127,128,144,165].

Zhang et al. [144] obtained high conversion (98%) and almost 100% selectivity for nitrogen for nitrate photocatalytic reduction by using as catalyst nontoxic fine Ag clusters obtained by photo-deposition of silver precursors on nano-sized titanium dioxide particles (denoted as Ag/TiO2), formic acid as hole scavenger, and a 125-W high-pressure Hg lamp, main wavelength around 365 nm, as a light source. The formation of more detrimental products, nitrite and ammonium, was thereby avoided, and residual formic acid can be completely decomposed into a harmless CO2 by further irradiation. Hou et al. [183] presented the novel core-shell structured Ag/SiO2@cTiO2 composites for photocatalytic reduction of high-concentration nitrate (2000 mg L<sup>−</sup>1). Photocatalytic denitrification experiments were performed with the light source of a 500 W high-pressure mercury lamp (main wavelength around 365 nm). Due to the electron sink effect, Ag NPs in the TiO2 shell could trap the photogenerated electrons and prolong the lifetime of charge carriers. The photogenerated electrons could be transferred from the CB of the TiO2 shell to Ag NPs for prevention of its oxidation to Ag+. Therefore, Ag/SiO2@cTiO2 could reduce high-concentration nitrate to N2 effectively.

Lin et al. [119] developed a bio electro-photocatalytic system under UV irradiation (Figures 11 and 17a) which exhibits a high selectivity for photocatalytic reduction of nitrate to N2.

A 30-W low-pressure mercury lamp was used as the light source. The proposed nonconventional bio electro-photocatalytic system has the advantage of a greater denitrification rate, higher selectivity to N2, absence of harmful by-products formation (nitrite or ammonium), the introduction of hole scavenger in nitrate solution is avoided, and costeffectiveness (Figure 17b–d). Compared with the conventional denitrification mechanism shown in their work [119], this type of a bio electro-photocatalytic reaction pathway has a lower energy barrier (Ea) (Figure 17b), suggesting that the complete photocatalytic reduction of nitrate to N2 without cumulation of harmful byproducts is energetically possible.

**Figure 17.** (**a**) Scheme of the bio electro-photocatalytic denitrification system. (**b**) Comparison of energy barrier (Ea) for photocatalytic denitrification in a bioelectronic-assisted way and the conventional pathway. The Ea of each step for NO3 − reduction on the TiO2 (101) surface with the transition state (TS) structures of Steps IV and VI in the bio electro-photocatalytic system. Step I in the bio electro-photocatalytic system is the adsorption of NO3 − on the TiO2 (101) surface. (**c**,**d**) Performance of the bio electro-photocatalytic denitrification system (c) the change of nitrate, nitrite, and ammonium concentrations in the denitrification process; (**d**) the relative concentration profiles of nitrate during the denitrification process in the bio electro-photocatalytic system (•: with bioanode to supply the bio-electrons but without irradiation; -: with UV irradiation but without connecting to the bioanode; -: with both UV irradiation and bio-electrons supply from the bioanode). Reproduced with permission from ref. [119]. Copyright 2017 Elsevier.

The photocatalytic denitrification reaction can be described below (Equation (40)):

$$2\,\mathrm{10e}\_{\mathrm{cb}}\,^{-} + 2\,\mathrm{NO}\_{3}\,^{-} + 12\mathrm{H}^{+} \to \mathrm{N}\_{2} + 6\mathrm{H}\_{2}\mathrm{O} \tag{40}$$

The absence of nitrite generation in the bio electro-photocatalytic denitrification setup is indicative that the reaction pathway is different compared to the conventional denitrification reaction mechanism. The authors provided a valuable solution to increase the efficiency and selectively of photocatalytic denitrification by coupling an electron generation device with a photocatalytic denitrification process and simultaneous activation of nitrate atom pairs for the final formation of N2 from nitrate [119,120].

In the study conducted by Liu et al. [184], a novel two-step reduction process was constructed for the selective removal of nitrate in an aqueous solution of Na2SO3 using Cu/Fe bimetal photocatalyst. The produced nitrite by the reduction of nitrate on the Cu0 surface could not be converted to ammonia rapidly on the surface of iron oxide layer, leading to the accumulation in time of nitrite. In the next step, the accumulated nitrite was

efficiently and easily reduced to nitrogen by Na2SO3, which worked as an efficient electron donor for nitrite reduction. The selectivity for N2 was over 90%, and the yield of ammonia was below 10% during the two-step reduction process.

Shang et al. [185] investigated the influence of exposed facets of silvered TiO2 photocatalysts on photo denitrification. They found that the nitrate reduction percentage and selectivity to N2 for Ag2O/Ag/101-TiO2 reached 99.1 and 81.1%, respectively, due to the formation of the junction at TiO2-metallic Ag0 interfaces and to Z-scheme charge transfer pathway mediated by adjacent Ag.

Recently, Silveira et al. [186] presented a study of the capability of the natural ilmenite (FeTiO3) to reduce nitrate from ultra-pure and mineral water. They claim that natural ilmenite can be a great applicant for reducing NO3 − in contaminated water. In ultra-pure water, the nitrate is totally converted to NOx (2%) and N2 (98%) after 210 min. If using oxalate in the mineral water, the nitrate is removed, but NO2 −, NOx, and N2 appear as products. In another study [187], the nitrate reduction in saline waters was explored for the first time employing the FeTiO3/oxalic acid photocatalytic process. A 150 W medium mercury lamp was used. Acidic pH values must be maintained to avoid oxalic acid precipitation by Ca2+ present in the water matrix. Under those conditions and compared to ultrapure water, salinity (in the range of 5–33 g/L) has a small influence on nitrate reduction, which is related to the evolution of C2O4 <sup>2</sup><sup>−</sup> concentration.

Wang et al. [188] prepared a novel SiW9/TiO2/Cu composite catalyst and studied the impact of catalyst loading, initial nitrate concentration, polyoxometalate loading, formic acid and O2 on the removal of nitrate under UV light. Nitrate removal up to 76.53% and 82.09% of N2 selectivity was obtained under specific experimental conditions: initial nitrate concentration of 30 mg/L, concentration of formic acid, 30 mmol/L, SiW9/Cu loading level 8.33%, the catalytic dosage of 0.8 g and presence of N2.

Graphitic carbon nitride (g−C3N4) has been broadly used in the area of photocatalysis due to its suitable features such as very good stability, graphene-like structure, ease of synthesis and the capability to produce photocarriers. Liu et al. [189] obtained AgyPd10−y/g-CxN4 Mott−Schottky heterojunction by growing AgPd nanowires (NWs) on the surface of nitrogen-rich g-CxN4. Their strategy opens a new way for making photocatalytic hydrogen production in tandem with the reduction of NO3 − and NO2 − in water, also extending it to remove metal ions. The Ag3Pd7/g-C1.95N4 catalyst exhibited the highest photocatalytic activity and selectivity for photocatalytic reduction of NO3 − and NO2 −, and the removal rates of NO3 − and NO2 − are 87.4% and 61.8%, respectively, under 365 nm irradiation, at 25 ◦C.

Soliman et al. [190] studied the reduction of nitrate in water under solar radiation using activated carbon prepared from date palm stone decorated with single and bimetallic nanoparticles. In their work, acetic acid, formic acid, oxalic acid, and ammonium oxalate have been investigated as holes scavengers for nitrate reduction. The obtained results for activated carbon modified with Pd-Ag (using formic acid as a hole scavenger with 0.05 M) showed that the conversion of nitrate (85% after 35 h of natural solar irradiation) takes place mainly through nitrogen gas (N2) rather than nitrite (NO2 −) or ammonium (NH4 +).

The photocatalytic activity strongly depends on the applied experimental conditions, including the mass of the photocatalyst, the incident beam intensity, the type of sacrificial agents, the nitrate concentration in the starting aqueous solution, and so on. Table 4 shows some photocatalytic performance results of different materials reported in the literature, with an emphasis on selectivity towards harmless nitrogen, although an accurate comparison is difficult because of variations in experimental conditions.

*Catalysts* **2023**, *13*, 380



Having a high solubility in water, the nitrate anion is recognized as one of the most widespread contaminants. That is why research is still needed for the development of efficient technologies in the purification of contaminated waters [192].

To completely clarify the nitrate and nitrite photocatalytic reduction mechanism, future studies should quantify the gaseous reaction products in order to elucidate which gaseous species are released during photocatalytic treatment. However, the precise assessment of gases released in the photocatalytic reduction process is extremely challenging from an experimental point of view. It is highly advised to look into reactors that can provide high mass transfer, efficient nitrate reduction, and, on the other hand, a good recovery of N-gases. Combining with other technologies would be a wise choice for improving photocatalytic processes.
