Interconversion and Removal of Inorganic Nitrogen Compounds via UV Irradiation

: Dissolved inorganic nitrogen (DIN) species are key components of the nitrogen cycle and are the main nitrogen pollutants in groundwater. This study investigated the interconversion and removal of the principal DIN compounds (NO − 3 , NO − 2 and NH + 4 ) via UV light irradiation using a medium-pressure mercury lamp. The experiments were carried out systematically at relatively low nitrogen concentrations (1.5 mM) at varying pHs in the presence and absence of oxygen to compare the reaction rates and suggest the reaction mechanisms. NO − 3 was fully converted into NO − 2 at a pH > 3 in both oxic and anoxic conditions, and the reaction was faster when the pH was increased following a ﬁrst-order kinetic at pH 11 ( k = 0.12 min − 1 , R 2 = 0.9995). NO − 2 was partially converted into NO − 3 only at pH 3 and in the presence of oxygen and was stable at an alkaline pH. This interconversion of NO − 3 and NO − 2 did not yield nitrogen loss in the solution. The addition of formic acid as an electron donor led to the reduction of NO − 3 to NH + 4 . Conversely, NH + 4 was converted into NO − 2 , NO − 3 and to an unidentiﬁed subproduct in the presence of O 2 at pH 10. Finally, it was demonstrated that NO − 2 and NH + 4 react via UV irradiation with stoichiometry 1:1 at pH 10 with the total loss of nitrogen in the solution. With these results, a strategy to remove DIN compounds via UV irradiation was proposed with the eventual use of solar light.


Introduction
The chemical element nitrogen (N) is a key component of life.Its most abundant form on Earth is atmospheric dinitrogen (N 2 ).However, N 2 is unavailable for most organisms, making N a limiting nutrient.Therefore, N compounds in nature can be classified into two groups: nonreactive N (N 2 ) and reactive N (N r ).N r includes all biologically, photochemically, and radiatively active N compounds in the Earth's atmosphere and biosphere [1].N r can be found in both organic and inorganic compounds.Organic compounds include urea, amines, proteins, and nucleic acids.Inorganic compounds can be gaseous (e.g., nitrous oxide (N 2 O), nitrogen oxide (NO) and nitrogen dioxide (NO 2 ), with the last two known as NO x ) and also can be present in soil and water.The most common dissolved inorganic nitrogen (DIN) compounds in water are nitrate (NO − 3 ), nitrite (NO − 2 ) and ammonium (NH + 4 ).In particular, NO − 3 is the main nitrogen groundwater pollutant [2].In the natural cycle of N, N 2 is converted into N r by lightning and by biological nitrogen fixation.The reverse reaction in the cycle converts organic N and NH + 4 into NO −  2 and NO − 3 (nitrification) and then to N 2 (denitrification).A few decades ago, these opposite processes were approximately equal.However, in recent times, anthropogenic activity [3] has led to an imbalance between these processes, resulting in the accumulation of N r in the environment.This includes the industrial Haber-Bosch process that converts N 2 into NH 3 used in the synthesis of fertilizers for food production and the fossil-fuel combustion that converts fossil N into NO x [4].To overcome the N r accumulation, the natural nitrification/denitrification process has been implemented in wastewater treatment plants [5].Recently, another biological process, autotrophic anaerobic ammonium oxidation (Anammox), has been used on a large scale [6].In this process, NH + 4 is oxidized with NO − 2 to yield N 2 via a biologically mediated process.DINs can also be removed using traditional physicochemical treatments, such as ion exchange, adsorption, etc. [7].Therefore, DINs are key compounds in the nitrogen cycle and water treatment technologies.In particular, the photocatalytic reduction of NO − 3 into N 2 is emerging as an alternative to the removal of nitrogen from water [7].TiO 2 is the most studied photocatalyst for this purpose but the interconversion of inorganic nitrogen compounds can also be achieved without it via simple UV irradiation [8].The absorption spectra of NO − 3 and NO −  2 are dominated by intense π → π * bands at 200 nm (ε = 9900 M −1 cm −1 ) and 205 nm (ε = 5500 M −1 cm −1 ), respectively.But they also present weaker n → π * bands at 310 nm (ε = 7.4 M −1 cm −1 ) and 360 nm (ε = 22.5 M −1 cm −1 ), respectively.These last bands can absorb solar radiation (λ > 295 nm), and they are important not only for solar-based technologies but also for processes initiated by sunlight in natural waters and in atmospheric aerosols [9].Brown carbon aerosols are light-absorbing particles primarily originating from biomass combustion.Comprising near-UV-absorbing organic compounds mixed with NO − 3 within atmospheric aerosols, these aerosols are thought to contribute to the presence of nitrous acid in the atmosphere through the photolysis of NO − 3 into NO 2 and NO − 2 [10].Although NH + 4 does not absorb significantly within the UV-vis spectra, it can be oxidized through photocatalytic methods [11].The removal of NH + 4 /NH 3 via oxidation poses a significant challenge due to its high stability and water solubility.Photocatalytic oxidation using TiO 2 (P25), particularly at high pH conditions, has been successful at converting NH + 4 /NH 3 into N 2 [12].In this investigation, we explored NH + 4 /NH 3 oxidation via irradiation in the presence of oxygen, omitting the use of a catalyst.
Usually, irradiation experiments have been undertaken to study DIN species separately.This work attempts to describe systematically the main features arising from the irradiation of DIN species by applying similar conditions but also through the simultaneous irradiation of NO − 2 and NH + 4 .In this specific scenario, prior research [13] demonstrates that photolysis exhibits superior performance compared to the photocatalytic process for nitrogen removal.The focus was placed on the main DIN compounds (NO − 3 , NO − 2 and NH + 4 ) in a relatively low total nitrogen concentration (1.5 mM).
The irradiation of DINs produces nitrogen radicals and OH • , which subsequently facilitates the oxidation of organic matter.This approach has garnered attention for its potential in addressing environmental challenges, particularly in the oxidation of emerging contaminants of concern (CECs) through UV irradiation when residual NO − 2 is present [14].In this study, formic acid (HCOOH) was selected as the target organic compound for oxidation due to its convenient mineralization and traceable degradation process.
The irradiation was aimed to excite the n → π * absorption bands of the compounds, which have implications in solar applications.For this purpose, a medium-pressure mercury lamp (main emission bands between 313 and 436 nm) was used in contrast with other studies that employ low-pressure mercury lamps (maximum emission at 254 nm) that excite the π → π * bands principally [15].A discussion was completed about the interconversion of DINs, the possible intermediates and the potential conversion into N 2 for the nitrogen removal from water.

Irradiation Experiments
Irradiation experiments were performed in an immersion well reactor (Figure 1) consisting of a thermostatted (298 K) annular reactor (130 mm length, 57 mm external diameter, 48 mm internal diameter, 160 mL total cell volume).The irradiation source was a mediumpressure mercury lamp installed inside the reactor (Photochemical Reactors Ltd., Reading, UK).This lamp emits radiation between 300 and 440 nm with main peaks at 313, 366, 404 and 436 nm (Figure S1 in Supplementary Material includes the emission spectrum of the medium-pressure Hg lamp together with the absorption bands of NO − 3 and NO − 2 ).Actinometric measurements were performed using the ferrioxalate method.An incident photon flux per unit volume (q p /V, where q p is the incident photon flux, and V is the irradiated volume) of 3.2 × 10 −4 µeinstein s −1 L −1 was calculated.In all cases, 200 mL of NO − 3 (1.2mM) + HCOOH (30 mM) solutions was adjusted to the desired pH using dropwise addition of H 2 SO 4 (0.5 M) or NaOH (0.5 M) into the reservoir where argon or oxygen (0.2 L min −1 ) was constantly bubbled.All the experiments were performed at 25.0 ± 0.1 • C. Solutions in the reservoir were magnetically stirred throughout the reaction time.Samples (1 mL) were periodically extracted for analysis, and the pH was measured.All runs were performed at least in duplicate, and the results were averaged.The experimental error was never higher than 10%.
tion standards of KNO , NaNO and NH Cl were obtained from ChemLab.All ments were performed with MilliQ water (resistivity = 18 MΩ cm), and all me chemicals were analytical grade or superior.

Irradiation Experiments
Irradiation experiments were performed in an immersion well reactor (Figur sisting of a thermostatted (298 K) annular reactor (130 mm length, 57 mm extern eter, 48 mm internal diameter, 160 mL total cell volume).The irradiation sourc medium-pressure mercury lamp installed inside the reactor (Photochemical React Reading, UK).This lamp emits radiation between 300 and 440 nm with main peak 366, 404 and 436 nm (Figure S1 in Supplementary Material includes the emission sp of the medium-pressure Hg lamp together with the absorption bands of NO an Actinometric measurements were performed using the ferrioxalate method.An photon flux per unit volume ( /, where  is the incident photon flux, and irradiated volume) of 3.2 × 10 μeinstein s −1 L −1 was calculated.In all cases, 20 NO (1.5 mM), NO (1.5 mM), NH /NH (1.5 mM), NO (0.75 mM) + NH /N mM) or NO (1.2 mM) + HCOOH (30 mM) solutions was adjusted to the desired p dropwise addition of H2SO4 (0.5 M) or NaOH (0.5 M) into the reservoir where a oxygen (0.2 L min −1 ) was constantly bubbled.All the experiments were performe ± 0.1 °C.Solutions in the reservoir were magnetically stirred throughout the reacti Samples (1 mL) were periodically extracted for analysis, and the pH was measu runs were performed at least in duplicate, and the results were averaged.The mental error was never higher than 10%.

Irradiation of 𝑁𝑂
Figure 2 shows the removal of aqueous NO (dashed lines) and the paral formation (solid lines) due to the irradiation of the nitrate solutions in a pH rang to 11 with a medium-pressure mercury lamp in anoxic conditions.The reaction creased when varying from an acid to alkaline solution, yielding NO as the ma uct (Equation ( 1)).[17] and SM−4500−NH 3 −F [18], respectively, employing an HP8453A spectrophotometer (Hewlett−Packard) with UV detection.HCOOH concentration was determined using a Shimadzu 5000 A TOC analyzer in the non-purgeable organic carbon (NPOC) mode.

Irradiation of NO −
At a neutral and acidic pH, the nitrogen mass balance of NO , NO and NH in the solution during irradiation indicates that they did not account for 100% of the nitrogen introduced in the system as NO .This discrepancy is likely attributed to a loss of nitrogen from the aqueous phase.As presented in Table 1, the nitrogen recovery considers the tota nitrogen found in the solution after the reaction (i.e., nitrogen originating from NO NO and NH ).The observed behavior concerning the influence of the solution pH is consistent with findings reported by other authors who have utilized a low-pressure mer cury lamp [15].The total reduction of NO to NO was also confirmed in oxic conditions at an al kaline pH (Table 1), indicating that the presence of oxygen in the system did not affect the reduction of NO .The observed pH dependence could be explained by the formation of peroxynitrous acid (ONOOH)/peroxynitrite (ONOO ) species considered to be one of the main intermedi ates of the photo reduction of NO [7].ONOOH is unstable, and it isomerizes to NO (Equation ( 2)), thereby reversing the reaction (1) [19].This acid has a pKa = 6.5; therefore at pH higher than 6.5, ONOO predominates, and the reaction continues to yield NO without reversing to NO .

ONOOH → NO + H
(2 At a neutral and acidic pH, the nitrogen mass balance of NO − 3 , NO − 2 and NH + 4 in the solution during irradiation indicates that they did not account for 100% of the nitrogen introduced in the system as NO − 3 .This discrepancy is likely attributed to a loss of nitrogen from the aqueous phase.As presented in Table 1, the nitrogen recovery considers the total nitrogen found in the solution after the reaction (i.e., nitrogen originating from NO − 3 , NO − 2 and NH + 4 ).The observed behavior concerning the influence of the solution pH is consistent with findings reported by other authors who have utilized a low-pressure mercury lamp [15].

Irradiation of 𝑁𝑂
Table 1.Kinetics of the photo reduction of NO − 3 at varying pH conditions.The data presented were calculated by performing a mathematical adjustment of the kinetic curves presented in Figure 2. Nitrogen recovery was calculated as 100 The total reduction of NO − 3 to NO − 2 was also confirmed in oxic conditions at an alkaline pH (Table 1), indicating that the presence of oxygen in the system did not affect the reduction of NO − 3 .The observed pH dependence could be explained by the formation of peroxynitrous acid (ONOOH)/peroxynitrite (ONOO − ) species considered to be one of the main intermediates of the photo reduction of NO − 3 [7].ONOOH is unstable, and it isomerizes to NO − 3 (Equation ( 2)), thereby reversing the reaction (1) [19].This acid has a pK a = 6.5; therefore, at pH higher than 6.5, ONOO − predominates, and the reaction continues to yield NO − 2 without reversing to NO − 3 .

Purge
ONOOH The oxidation of NO − 2 into NO − 3 through irradiation was partially achieved at pH 3 in the presence of O 2 .In these conditions, 30% of NO − 2 (1.5 mM) was converted into NO − 3 within 2 h (Figure 3).Initially, when the NO − 2 solutions were prepared at pH 3, a rapid dark reaction leading to the oxidation of NO − 2 into NO − 3 was observed.Therefore, experiments at this pH were initiated with the presence of NO − 3 , as shown in Figure 3.In contrast, at a higher pH or in the absence of O 2 , this reaction was not significant, indicating the stability of NO − 2 in the dark.
contrast, at a higher pH or in the absence of O , this reaction was not significant, indicat ing the stability of NO in the dark.

HNO ⇌ NO + H (3
When exposed to 355 nm irradiation, both species undergo a similar photolytic pro cess, with the acid form exhibiting a higher quantum yield ( = 0.4 vs.  0.025 respectively) [7,20].This disparity in quantum yields could explain the observed oxidation only in acidic conditions (Equations ( 4) and ( 5)).

Irradiation of 𝑁𝐻 /𝑁𝐻
The removal of aqueous NH /NH via irradiation was observed at a high pH (Figur 4) in the presence of oxygen in contrast to the NO oxidation that only took place at a low pH (Figure 3).At pH 10, 72% of NH / NH (1.5 mM) was converted mainly into an un identified compound.The concentrations of NO and NO were low and constan throughout the experiment, suggesting that these compounds were reaction intermedi ates in a steady-state mechanism.The reaction was not observed at a pH lower than (kinetic curves not shown) or in the absence of O .
Unlike NO and NO , NH /NH do not absorb light in the emission range of th medium-pressure mercury lamp.Therefore, the direct photolysis of these species wa ruled out.Nevertheless, hydroxyl radical (HO • ) can be generated via direct photolysis o water at 254 nm with a quantum yield of 0.08 for the HO • formation [21].The oxidation of NH via a reaction with HO • was reported before, and it led to the formation of NO It was suggested that the initial reaction proceeds with the hydrogen abstraction with HO yielding the key intermediate NH • (Equation ( 6)) [22].In an aqueous solution, nitrous acid (HNO 2 ) exists in equilibrium with NO − 2 with a pK a = 3.4 (Equation (3)).
The removal of aqueous NH 3 /NH + 4 via irradiation was observed at a high pH (Figure 4) in the presence of oxygen in contrast to the NO − 2 oxidation that only took place at a low pH (Figure 3).At pH 10, 72% of NH 3 / NH + 4 (1.5 mM) was converted mainly into an unidentified compound.The concentrations of NO − 3 and NO − 2 were low and constant throughout the experiment, suggesting that these compounds were reaction intermediates in a steady-state mechanism.The reaction was not observed at a pH lower than 7 (kinetic curves not shown) or in the absence of O 2 .
It is postulated that after the initial oxidation of NH into NO /NO , both com pounds reacted to produce N2/NOx (see next section), which would explain the total ni trogen decrease.This could be described with a two-step consecutive reaction mechanism where the first reaction (NH oxidation) is much slower than the second reaction (N2/NO generation), producing a steady-state low concentration of NO and NO during the re action.

Simultaneous Irradiation of 𝑁𝑂 and 𝑁𝐻 /𝑁𝐻
The simultaneous removal of aqueous NO and NH was achieved via irradiation at pH 10 without the detection of NO (Figure 5).Both compounds (0.75 mM) were re moved by 75% in 2.

Unlike NO −
3 and NO − 2 , NH 3 /NH + 4 do not absorb light in the emission range of the medium-pressure mercury lamp.Therefore, the direct photolysis of these species was ruled out.Nevertheless, hydroxyl radical (HO • ) can be generated via direct photolysis of water at 254 nm with a quantum yield of 0.08 for the HO • formation [21].The oxidation of NH 3 via a reaction with HO • was reported before, and it led to the formation of NO − 2 .It was suggested that the initial reaction proceeds with the hydrogen abstraction with HO • yielding the key intermediate NH • 2 (Equation ( 6)) [22].
Besides light, O 2 is a necessary reagent, so Equation ( 7) was proposed to be the overall reaction here [23].Also, the NH + 4 /NH 3 acid-base equilibrium pKa is 9.25, which explains that the oxidation was observed at a more alkaline pH where NH 3 is the main species.
It is postulated that after the initial oxidation of NH 3 into NO − 3 /NO − 2 , both compounds reacted to produce N 2 /NO x (see next section), which would explain the total nitrogen decrease.This could be described with a two-step consecutive reaction mechanism where the first reaction (NH 3 oxidation) is much slower than the second reaction (N 2 /NO x generation), producing a steady-state low concentration of NO − 3 and NO − 2 during the reaction.The simultaneous removal of aqueous NO − 2 and NH 3 was achieved via irradiation at pH 10 without the detection of NO − 3 (Figure 5).Both compounds (0.75 mM) were removed by 75% in 2.5 h.The kinetic curves obtained for NO − 2 and NH 3 /NH + 4 were almost identical, indicating that the reaction stoichiometry between NO − 2 and NH 3 / NH + was 1:1.Moreover, the experimental data adjusted very well to a rate equation with a partial first order for NO − 2 and also partial first order for NH At pH 6, the reaction was significantly slower, and the formation of NO was evi denced as a reaction product without a significant total nitrogen removal.This reactio dependence on pH suggested that NH , not NH , is the reactant that leads to nitroge removal.

Simultaneous Irradiation of NO
The mechanism of the photochemical reaction between NH and NO can be ana lyzed considering the mechanism of the thermal reaction.One of the key intermediate proposed in the thermal reaction is dinitrogen trioxide (N O ).NH reacts with N O (Equation ( 8)) through a Substitution Nucleophilic Bimolecular reaction (SN2) (Figure 6A [24] where NH is the nucleophile, N O the electrophile and NO the leaving group.

H NNO ⟶ H NNO + H (9
And H NNO isomerizes into HNNOH, which dissociates into the final products N and H O (Equation ( 10)) [24] (Figure 6B).As in the thermal reaction, N O could also be postulated as an intermediate for th photochemical reaction [25].The final steps of the mechanisms leading to N would b At pH 6, the reaction was significantly slower, and the formation of NO − 3 was evidenced as a reaction product without a significant total nitrogen removal.This reaction dependence on pH suggested that NH 3 , not NH + 4 , is the reactant that leads to nitrogen removal.The mechanism of the photochemical reaction between NH 3 and NO − 2 can be analyzed considering the mechanism of the thermal reaction.One of the key intermediates proposed in the thermal reaction is dinitrogen trioxide (N 2 O 3 ).NH 3 reacts with N 2 O 3 (Equation ( 8)) through a Substitution Nucleophilic Bimolecular reaction (S N 2) (Figure 6A) [24] where NH 3 is the nucleophile, N 2 O 3 the electrophile and NO − 2 the leaving group.At pH 6, the reaction was significantly slower, and the formation of NO was evidenced as a reaction product without a significant total nitrogen removal.This reaction dependence on pH suggested that NH , not NH , is the reactant that leads to nitrogen removal.

H NNO ⇌ HNNOH ⟶ N + H O (10
The mechanism of the photochemical reaction between NH and NO can be analyzed considering the mechanism of the thermal reaction.One of the key intermediates proposed in the thermal reaction is dinitrogen trioxide (N O ).NH reacts with N O (Equation ( 8)) through a Substitution Nucleophilic Bimolecular reaction (SN2) (Figure 6A) [24] where NH is the nucleophile, N O the electrophile and NO the leaving group.
And H NNO isomerizes into HNNOH, which dissociates into the final products N and H O (Equation ( 10)) [24] (Figure 6B).As in the thermal reaction, N O could also be postulated as an intermediate for the photochemical reaction [25].The final steps of the mechanisms leading to N would be The transient compound, H 3 NNO + , quickly converts into nitrosamine (H 2 NNO) (Equation ( 9)) [24].

H NNO ⇌ HNNOH ⟶ N + H O
And H 2 NNO isomerizes into HNNOH, which dissociates into the final products N 2 and H 2 O (Equation ( 10)) [24] (Figure 6B).As in the thermal reaction, N 2 O 3 could also be postulated as an intermediate for the photochemical reaction [25].The final steps of the mechanisms leading to N 2 would be the same as stated above, but the difference would lie in how N 2 O 3 is formed.In the thermal reaction, N 2 O 3 is postulated to be generated by the condensation of two molecules of HNO 2 (Equation ( 11)) [24].
In fact, HNO 2 and NH 3 are the postulated reagents for the thermal reaction where the reaction rate increases at a lower pH [24].On the contrary, in the photochemical reaction, the postulated reagents are NO − 2 and NH 3 , and as shown here, the reaction takes place in alkaline media.
The formation of N 2 O 3 via irradiation could be explained by the reaction between NO • and NO • 2 (Equation ( 12)) [8].
Another important difference is that in the thermal process, the kinetic mechanism is near-second-order in the HNO 2 concentration [24], whereas we showed here that in the irradiation process, the reaction was first-order in the NO − 2 concentration.We did not observe a significant reaction in the dark.This could be due to the relatively low total nitrogen concentration (about 1.5 mM) used in these experiments.The thermal reaction is reported to have occurred mainly at concentrations above 100 mM 11 .This agrees with the different reaction mechanisms indicated above.

Irradiation of NO −
3 in the Presence of Formic Acid As shown in previous sections, the irradiation of NO − 3 and NO − 2 with a mediumpressure mercury lamp produces reactive nitrogen species (RNS) as intermediates, which in turn yields NO − 2 as the main product regardless of the initial conditions of pH and oxygenation.To achieve further nitrogen reduction and eventually eliminate it, it is necessary to add another reactant as an electron donor, such as an organic compound.This is possible because the irradiation of NO − 3 and NO − 2 , besides RNS, produces reactive oxygen species (ROS), which are scavenged by organic compounds that are oxidized and mineralized in this way.HO • is the main ROS produced as a consequence of the irradiation of NO − 2 and NO − 3 (Equations ( 13) and ( 15), respectively).Therefore, this irradiation process has also been proposed for the removal of contaminants of emerging concern (CECs) [14].
In particular, it is known that formic acid is oxidized by HO • , yielding CO •− 2 as the first intermediate (Equation ( 16)) [28].
The irradiation of NO − 3 in the presence of formic acid yielded NH 3 and a loss of total nitrogen at pH 10 (Figure 7).In the beginning of the reaction, a transient peak of NO − 2 concentration was observed along with a quick fall of the NO − 3 concentration.Then, NH 3 appeared after a lag period in which a minimum of total nitrogen was observed.This suggested the presence of another nitrogen intermediate that was different from the three main DINs and that was converted into NH 3 more slowly.
The oxidation of formic acid was also confirmed with the measurement of total or ganic carbon (TOC).During the reaction, the TOC was reduced steadily along with the formation of inorganic carbon (IC), which showed the mineralization of the formic acid (Figure 7).Formate has already been tested for the removal of nitrate from groundwater [29] indicating that the environmental application is potentially possible.

Conclusions
The irradiation of the principal DIN compounds (NO , NO and NH /NH ) wa performed separately at a relatively low concentration (1.5 mM) with a medium-pressure mercury lamp with main emission bands between 313 and 436 nm.The interconversion of DIN compounds was observed mainly in alkaline media (pH 10-11).In these condi tions, it was found that NO is fully converted into NO following first-order reaction kinetics, and NH was removed from the solution in the presence of oxygen, also yielding a minimum amount of NO and NO .Interestingly, when NO and NH were irradi ated simultaneously, both were removed from the solution in a 1-to-1 mol proportion without the requirement of any other reactant.Finally, NO was reduced to NH only after the addition of formic acid as an electron donor, also yielding the removal of the DINs from the solution.
The required alkaline media suggested NH as the active compound instead o NH , and in the case of NO irradiation, it suggested the presence of a reaction interme diate, probably peroxynitrite (ONOO−), which is unstable in an acidic pH.
It was demonstrated that the DINs could be removed from the solution without in troducing a catalyst and by applying a light source emitting in the UV A/B range; there fore, solar irradiation could eventually be applied also.The strategy for DIN removal via NO − 3 removal with formic acid was faster than in the simple irradiation (Section 3.1).Also, the total nitrogen was removed faster than in the simultaneous irradiation of NO − 2 and NH 3 (Section 3.4), evidencing a different mechanism for nitrogen removal.Similar reaction rates were obtained at pH 3 with formic acid.This was another significant difference compared to the irradiation without the electron donor, which had a marked dependence on the pH as shown before.
It is suggested that this new mechanism and the nitrogen intermediates could be initiated by the reaction between NO • and CO •− 2 (Equation ( 17)), which are the products of the photolysis of NO − 2 (Equation ( 13)) and the oxidation of formic acid (Equation ( 16)), respectively [28]. NO The oxidation of formic acid was also confirmed with the measurement of total organic carbon (TOC).During the reaction, the TOC was reduced steadily along with the formation of inorganic carbon (IC), which showed the mineralization of the formic acid (Figure 7).
Formate has already been tested for the removal of nitrate from groundwater [29], indicating that the environmental application is potentially possible.

Conclusions
The irradiation of the principal DIN compounds (NO − 3 , NO − 2 and NH 3 /NH + 4 ) was performed separately at a relatively low concentration (1.5 mM) with a medium-pressure mercury lamp with main emission bands between 313 and 436 nm.The interconversion of DIN compounds was observed mainly in alkaline media (pH 10-11).In these conditions, it was found that NO − 3 is fully converted into NO − 2 following first-order reaction kinetics, and NH 3 was removed from the solution in the presence of oxygen, also yielding a minimum amount of NO − 3 and NO − 2 .Interestingly, when NO − 2 and NH 3 were irradiated simultaneously, both were removed from the solution in a 1-to-1 mol proportion without the requirement of any other reactant.Finally, NO − 3 was reduced to NH 3 only after the addition of formic acid as an electron donor, also yielding the removal of the DINs from the solution.
The required alkaline media suggested NH 3 as the active compound instead of NH + 4 , and in the case of NO − 3 irradiation, it suggested the presence of a reaction intermediate, probably peroxynitrite (ONOO−), which is unstable in an acidic pH.
It was demonstrated that the DINs could be removed from the solution without introducing a catalyst and by applying a light source emitting in the UV A/B range; therefore, solar irradiation could eventually be applied also.The strategy for DIN removal via irradiation would be to have NO −

Figure 1 .
Figure 1.Scheme of the irradiation setup.

Figure 1 .
Figure 1.Scheme of the irradiation setup.

3 Figure 2
Figure2shows the removal of aqueous NO − 3 (dashed lines) and the parallel NO − 2 formation (solid lines) due to the irradiation of the nitrate solutions in a pH range from 3 to 11 with a medium-pressure mercury lamp in anoxic conditions.The reaction rate

Figure 2 .
Figure 2. NO (dashed lines) and NO profiles (solid lines) vs. time obtained after the irradiation of NO (1.5 mM) at varying pH conditions.The solutions were continuously purged with argon.

Figure 2 .
Figure 2. NO − 3 (dashed lines) and NO − 2 profiles (solid lines) vs. time obtained after the irradiation of NO − 3 (1.5 mM) at varying pH conditions.The solutions were continuously purged with argon.

Figure 3 .
Figure 3. NO (dashed lines) and NO (solid lines) profiles vs. time obtained after the irradiation of NO (1.5 mM).The reaction was performed at pH 3 and pH 10 with and without O .

Figure 3 .
Figure 3. NO − 3 (dashed lines) and NO − 2 (solid lines) profiles vs. time obtained after the irradiation of NO − 2 (1.5 mM).The reaction was performed at pH 3 and pH 10 with and without O 2 .

Figure 4 .
Figure 4. NH (dotted lines), NO (dashed lines) and NO (solid lines) profiles vs. time for th irradiation of NH /NH (1.5 mM) in oxic and anoxic conditions.The solutions were purged with oxygen and argon, respectively.
5 h.The kinetic curves obtained for NO and NH /NH were almos identical, indicating that the reaction stoichiometry between NO and NH / NH wa 1:1.Moreover, the experimental data adjusted very well to a rate equation with a partia first order for NO and also partial first order for NH (ν = k [NO ] [NH ], k = 0.5 M − seg −1 , R 2 = 0.9994).

Figure 5 .
Figure 5. NO (dashed lines), NO (solid lines) and NH /NH (dotted lines) profiles vs. time fo the irradiation of NO and NH /NH at varying pH conditions.The initial total nitrogen concen tration (C0) was 1.5 mM.The solution was purged with argon.

Figure 6 .
Figure 6.Substitution Nucleophilic Bimolecular reaction between NH and N O (A) and H NNO isomerization and dissociation (B).

Figure 5 .
Figure 5. NO − 3 (dashed lines), NO − 2 (solid lines) and NH 3 /NH + 4 (dotted lines) profiles vs. time for the irradiation of NO − 2 and NH 3 /NH + 4 at varying pH conditions.The initial total nitrogen concentration (C 0 ) was 1.5 mM.The solution was purged with argon.

Figure 5 .
Figure 5. NO (dashed lines), NO (solid lines) and NH /NH (dotted lines) profiles vs. time for the irradiation of NO and NH /NH at varying pH conditions.The initial total nitrogen concentration (C0) was 1.5 mM.The solution was purged with argon.

Figure 6 .
Figure 6.Substitution Nucleophilic Bimolecular reaction between NH and N O (A) and H NNO isomerization and dissociation (B).

Figure 7 .
Figure 7. Irradiation of NO (1.5 mM) at pH 10 with 30 mM formic acid.The solution was contin uously purged with argon.

Figure 7 .
Figure 7. Irradiation of NO − 3 (1.5 mM) at pH 10 with 30 mM formic acid.The solution was continuously purged with argon.
3 and/or NO − 2 mixed with NH 3 at pH 10-11.If NO − 3 and NO − 2 are initially absent, they can be obtained via irradiation of NH 3 in the presence

Table 1 .
Kinetics of the photo reduction of NO at varying pH conditions.The data presented were calculated by performing a mathematical adjustment of the kinetic curves presented in Figure2Nitrogen recovery was calculated as 100 × ( NO + NO + NH )/ NO .