Removal of Nitramine Explosives in Aqueous Solution by UV-Mediated Advanced Oxidation Process in Near-Neutral Conditions

Explosive compounds are hazardous to the environment, posing a serious risk to human and animal health and the ecosystem. The primary goal of research was to compare the efficiency of UV/H2O2, photo-Fenton, electro (EO)/ UV/H2O2 processes at near-neutral pH (pH=6) on the degradation of nitramine explosives (NAs), such as hexogen (RDX), octogen (HMX), and tetryl (TET), in an aqueous solution. The effect of operational conditions, likely pH of the solution, initial H2O2 concentration, initial Fe 2+ concentration, and solution temperature, was observed. The removal kinetics fit with first-order kinetics and were in the order: photo-Fenton >EO/UV/H2O2 > UV/H2O2. The results showed higher rate constant values for TET, RDX and HMX removal by UV/H2O2 (k = 0.07778, 0.03791 and 0.03786 min-1), EO/UV/H2O2 (k = 0.16599, 0.1475 and 0.08674 min -1) and photo-Fenton (k = 0.18018, 0.1501 and 0.09336 min-1) processes. Furthermore, TET, RDX and HMX were mineralized at 59.7%, 45.1%, and 25.1 %, respectively, under optimum conditions after 60 min of the photo-Fenton process. From the economic perspective, photo-Fenton only requires 2.132–4.113 kWh m-3 to completely reduce NAs. Finally, acute toxicity towards Vibrio fischeri was defeated after usage of near-neutral photo-Fenton. Thus, photo-Fenton at circum-neutral is promising for low-cost, eco-friendly and efficient processes for treating nitramine explosives in aqueous solutions.

1,2-dimethylhydrazine is known to cause cancer in rats (Fiala, 1977). Previous research showed that methanol, 1,1-dimethylhydrazine, hydrazine, and formaldehyde are derivatives of RDX (Freedman and Sutherland, 1998). Thus, by-products from RDX and HMX must be closely monitored and remediated in the environment (Liou et al. 2003).
In recent years, the treatment of NAs-contaminated water has been carried out using a variety of methods, such as adsorption (Lingamdinne et al., 2015), hydrolysis (Bishop et al.2000), wet oxidation (Hao and Phull, 1993), and biological treatment (Cyplik et al. 2012). However, these processes have many disadvantages, including long duration for biological treatment, high pressure and temperature demands for wet air oxidation, and incomplete contaminant removal for the adsorption process (Bhaskar Raju et al. 2011). For instance, complete denitrification took 13 days using anaerobic bacteria DeNitrifying Consortium 2 (Cyplik et al. 2012). However, advanced oxidation processes (AOPs) are promising technologies that aim to decompose and mineralize potential-contaminants into non-toxic products (Andreozzi et al., 1999). AOPs involve various methods leading to the production of free radicals, such as hydroxyl radicals (HO * ) and sulphate radicals (SO 4 *-), O 2 * , HO 2 * etc., which can oxidize contaminants and treat wastewater (Radovic et al. 2015). Many researchers have applied electrochemical technology (EO) as a simple, safe, and environmentally-friendly process for the removal of various pollutants from aquatic environments (Dalvand et al. 2011, Martínez-Huitle et al. 2012. Furthermore, the benefits of electro-oxidation compared to other AOPs are limited sludge and salt formation, noticeably safe and simple operation, fast reaction, and significant reduction of total organic carbon (TOC). Electrochemical technology, using electrolysis, can be coupled with UV light for methanol removal (Odetola, Trevani and Easton, 2017) or amoxicillin degradation in water (Fallahzadeh et al. 2019). This phenomenon explains the potential of the products to remove pollutants by way of free radicals from the electrooxidation process.
Nonetheless, there are some disadvantages, including an acidic pH value (2.8-3.5), which is needed to ensure that Fe(II) and Fe(III) ions enhance the catalytic effect, thus limiting iron oxyhydroxides Fe(OH) 3 and exaggerating the concentration of photo-promoted electrons. The feasibility of working at near-neutral pH (5.0-6.0) has provoked interest from researchers (Dantas and Esplugas, 2014). According to Zhai et al. (2018), the reduction of natural gas extraction wastewater was achieved at the optimal pH at 6 using a photo-Fenton process. The using of "low cost reactors" at neutral pH for elimination of emerging pollutants was checked out by Sarina-Molina et al. and it also uncovered that they could evaculate 80% of them within 15 min of therapy. In order to being used by more rural societies for drinking water, it is nescessary for optimization of the system at a neutral or near neutral. A various of chemical methods, which would request time, noticeable attempt and reduce chemical's costs, will be implemented by the communitites Use of a photo-Fenton reaction, an AOP, leads to total pollutant mineralization. However, partial mineralization can generate by-products which are more noxious than the original species. In some cases, the increase in toxins appeared after pre-treatment using a photo-Fenton reaction of toxic effluent (Carra et al. 2015). Since a biological response reflects the level of toxicity, various organisms can be applied to improve the reliability of the results.
Organisms applied in bioassays for toxicity involve typical samples of seawater, freshwater, or terrestrial ecosystems, including microorganisms, plants, invertebrates and fish (Tothill and Turner 1996). It is beneficial to control the processes and compare various AOPs when evaluating the toxicity of treatments using bioassays with different types of organisms (Rizzo, 2011). Luminescent microorganisms, namely Vibrio fischeri (V. fischeri) bacteria, have often been applied in toxicity test instruments (Rizzo, 2011). These tests are based on the shift in bacterial luminescence when exposed to toxic chemicals. Simplicity and short timeframe are the main benefit of these tests.
There are only a few studies regarding NA degradation in water by AOPs assisted by UV light and an EO process in circum-neutral conditions. This research aims to assess the efficiency of three AOPs processes in the degradation of NAs compounds: UV/H 2 O 2 , UV/EO/H 2 O 2 and photo-Fenton. The influence of parameters such as H 2 O 2 concentration, ratio H 2 O 2 /Fe 2+ , and temperature of the experiments on the degradation of NAs was investigated. Also, the efficiency and electric energy consumption of the three AOPs was compared. Finally, mineralization and toxicity were evaluated with V.fischeri bacteria of the circum-neutral photo-Fenton process. 30%) and NaCl, supplied by Sigma-Aldrich (Merck, Germany) and Acros Organics (Thermo Fisher Scientific, USA), respectively. All solvents (acetonitrile, ethanol, methanol, and hexane) used in high-performance liquid chromatography (HPLC) were of analytical grade.

Continuous-flow experiments
All experiments were carried out in a circulatory photoreactor system, following Khue et al. (2014), as shown in Figure 1.
All solutions in the AOP experiments were prepared using distilled water and the pH was maintained at 6.0, applying either NaOH 0. Fenton's reagent was placed into the reactor tank/ electrochemical reactor (500 mL) with a heating magnetic stirrer Are-6 (Velf scientifica, Italy). All NAs solution samples were measured by a pH meter (Milwaukee 102, Poland). The solution was mixed by heating, with the magnetic stirrer at 300 rpm during the experiment, and circulated through a UV reaction chamber with a 15W UV light lamp (Philips, Netherland) (radiation wavelength 254 nm) applying a dosing pump (Hanna BL 20-2, Hanna Instruments, Romania) at a flow rate of 700 mL min -1 . All the experiments were conducted in triplicate.

Electro-oxidation reactor
In the case of UV/EO/H 2 O 2 , an electrochemical reactor (500 mL) (Fig. 1b) in which the anode and cathode were 1 mm thick Ti plates (4 cm × 20 cm) covered with TiO 2 and RuO 2 /IrO 2 (Inox) was used. The interelectrode distance between anode and cathode was 2 cm. AC power supply (Model 1655A, B&K Precision, USA) with a current density of 15 mA cm -2 was supplied for this electro-oxidation system. 3 g L -1 NaCl was used as electrolyte concentration in the electrooxidation process.

NAs determination
Three NAs compounds: TET, RDX, and HMX were measured by HPLC system (HP-Agilent 1100 series, diode array detector, Agilent, USA), applying Hypersil C18 column (200 mm × 4 mm). The mobile phase was made up of 67% acetonitrile and 33% water (v/v) at a flow rate of 0.6 mL/min, with a pressure of 280 Bars. The analytical signal was determined at a wavelength of 227 nm and retention times (t R ) for TET, RDX and HMX at 5.7 min, 5.11 min and 4.7 min, respectively (Khue et al. 2014).

TOC determination
The total organic carbon (TOC) was measured by Torch Combustion Analyzer (Torch, Teledyne Tekmar, USA).

V. fischeri acute toxicity
The standard ISO 1348-3:2007 (International Organization for Standardization, 2007) was used for the luminescence inhibition test with V. fischeri at 30 min after treatment 0, 15, 30 and 60 min with the photo-Fenton process, with the Microtox SOLO freeze-dried bacteria kit. The bacteria were carefully preserved at -20°C and activated by hydration according to the kit's standard operation procedure. All samples were tested in triplicate and kept on a thermostatic plate at 15°C throughout the test period. The control sample used a solution of 20 g L -1 NaCl (Acros, USA). After 30 min, bioluminescence was measured with a Delta ATP illuminometer (Modern Water, USA).
The bioluminescence inhibition ratio (BIR) (%) was calculated using Eq. (3): where: L Blank and L sample are the bioluminescence signals after 30 min of exposure for the sample without NAs and tested samples, respectively.
The removal efficiency (H(%)) of NAs and TOC removal by three Aops is calculated by equation Eq. (4): where: H(%) is the efficiency of NAs removal C (0) -initial concentration of NAs and TOC; C (t) -concentration of NAs or TOC at time t.
Description of the kinetic reaction and activated energy E a of the TET, HMX and RDX oxidation using three AOPs was determined based on the concentration of Nitramines (C NAs ) and reaction time. The pseudo-first-order rate equation could be used: where: C NAs(0) and C NAs -the concentrations of nitramine compounds at initial time and at time instant t; k 0(NAs) -the pseudo-first-order rate constant. The activation energy of the reaction was performed by Arrhenius equation Eq. (6) (Bose, Glaze and Maddox, 1998) as follows: k (0)NAs = Aexp(-Ea/RT) (6) where: A -pre-exponential factor; Ea -activation energy; R -gas constant; T -temperature.

Energy consumption and current efficiency
The energy consumption performance per order for the three AOPs can be determined by equation Eq. (7) (Asaithambi, Saravanathamizhan and Matheswaran, 2015).
EE/O= (kWh m -3 /order ) (7) where: P is the electrical power input (kW), V is the volume of effluent used (L), k -rate constant of first-order kinetics (min -1 ).

UV/H 2 O 2 process: effect of the initial concentration of H 2 O 2
The initial dosage of H 2 O 2 is the essential operation influencing reduction efficiency in the Moreover, apart from the main effect of HO * , it is necessary to take into account the simultaneous effects of UV radiation even though direct quantum photonic activity is not high. The influence of UV makes bonds in the NA molecule weaker and more flexible, and therefore more susceptible to attack by free radicals HO*.

EO/UV/H 2 O 2 process: effect of the initial concentration of H 2 O 2
As shown in Figure 3, for EO/UV/H 2 O 2 processes, a faster increase in the rate constant was found when the initial dose of H 2 O 2 increased from 7.5 mM to 14.5 mM. The rate constant increased from 0.06031 min -1 to 0.08674 min -1 , 0.0914 min -1, to 0.1475 min -1 , and 0.12577 min −1 to 0.16599 min −1 for HMX, RDX and TET removal, respectively. This is perhaps due to the greater generation of free radicals, such as OH * and Cl * , by the electrochemical system. Firstly, HO* was produced by water oxidation and physically adsorbed at the anode surface, via the reaction in Eq.(9) (Alfonso-Muniozguren et al. 2020). Furthermore, in EO/UV/H 2 O 2 , hypochlorous acid (HClO) is produced from the oxidation of chloride (Cl -), which is included in the effluent of 3 mg/L at the anode via the reactions in Eq. (12)(13)(14) at pH 6 (Panizza and Cerisola, 2009 HClO + hv → HO * + Cl * Moreover, an increase in the dosage of H 2 O 2 to 29 mM led to a decrease in the reaction rate. The reason is that the accumulation of HO * as well as the reaction of hydroxyl radicals with H 2 O 2 provide the hydroxyl radical scavenging, following Eq. (10,11). These results show that the optimal initial H 2 O 2 concentration is 14.5 mM. At this concentration, complete degradation of TET was reached after only 20 min, and for RDX at 25 min. Simultaneously, the maximum removal of HMX concentration of 89.85% was observed in the EO/UV/H 2 O 2 process, with an initial dose of H 2 O 2 14.5 mM.

Photo-Fenton process: influence of the ratio H 2 O 2 / Fe 2+ on removal of NAs
The generation of HO * includes a different stage in the photo-Fenton system: the primary stage being the photo-Fenton reaction (Eq. (1)), photolysis of Fe 2+ (catalyst), and photolysis of H 2 O 2 . Thus, the ratio H 2 O 2 /Fe 2+ constitutes the main condition of this system. As shown in Figure 4, NAs reduction efficiency and the reaction rate constant (k) increased, with an increase in the ratio of H 2 O 2 /Fe 2+ ; the ratio H 2 O 2 / Fe 2+ increased from 3.65/0.34 (m/m) to 29/0.624 (m/m), constant k value increased from 0.02158 min -1 to 0.09336 min -1 for HMX removal; from 0.07004 to 0.1501 min -1 RDX reduction; and 0.12331 to 0.17882 min -1 for TET degradation. Conversely, a further increase in the H 2 O 2 /Fe 2+ ratio to 29/1.35 (m/m), produced lower HMX, RDX, and TET removal effi ciency improvements (k values) at 0.17522, 0.12818 and 0.0834 min -1 , respectively. This was due to the scavenger eff ect of excess Fe 2+ , where HO * reacts with redundant ferrous iron, as shown in Eq. (15). The optimal ratio of H 2 O 2 /Fe 2+ for the photo-Fenton reaction was observed at 29/0.624. In comparison, the traditional photo-Fenton process in acidic condition consumes less H 2 O 2 and requires only catalytic amounts of Fe 2+ ; this study recommends a relatively higher H 2 O 2 concentration and lower catalyst (Fe 2+ ). This may explain the photo-Fenton activation mechanism in near neutral (pH=6) conditions, which exist in a photo-congestion region, requiring defi nition of the participation of precipitated iron. First, in the photo-Fenton system, soluble iron ion (Fe 2+ ) is observed as iron hydroxides at near-neutral pH (pH=6) (Eq. 16

Eff ect of temperature on the removal of NAs by UV/H 2 O 2 , EO/UV/H 2 O 2 , and photo-Fenton
The reaction temperature is a critical condition. Figure 5 illustrates the influence of the reaction temperature on NA removal using the three AOPs: UV/H 2 O 2 , EO/UV/H 2 O 2 , and photo-Fenton. Experiments were conducted at various temperatures: 20 o C, 25 o C, 30 o C, 35 o C, and 40 o C. It can be observed that by raising the temperature, the rate of reaction increases significantly for all three AOPs processes. For example, the reaction rate constant for HMX removal in UV/H 2 O 2 increases from 0.0065 min -1 to 0.098 min -1 by raising the temperature from 20 o C to 40 o C. Similarly, the k value of RDX reduction by EO/UV/H 2 O 2 increases from 0.0067 min -1 to 0.2393 min -1 . In the degradation of TET by the photo-Fenton system, the highest k value is 0.2833 min -1 at 40 o C. Further increasing temperature, these are complete removal for NAs with electron-hole recombination due to free radical production increase (Ali et al. 2019).
The activation energy Ea values for NAs reduction by the UV/H 2 O 2 , EO/UV/H 2 O 2 and photo-Fenton system were calculated (Table  2). It can be concluded that TET degradation by the three AOPs is the most efficient, which can be explained by the structure of TET, which is a compound containing aromatic rings with less chemical resistance than RDX and HMX. The electronic substitute group of aromatic compounds often depends on the group's tendency to attract or repel electrons. Electronic repelling groups, such as -CH 3 , -NH 2 and -OH, activate the aromatic ring, while the electronic attracting groups, such as -NO 2 , -COOH, -NH 3 and -CN, reactivate the aromatic ring. The nitramine ring compounds RDX and HMX do not have the same electronic state as the benzene rings of aromatic nitro compounds, or of TET; therefore, even in the three AOPs processes, the efficiency of TET reduction is significantly higher. It is observed that the internal bond angle of HMX is much greater than that of RDX, and so it is easier to oxidize RDX than HMX. Consequently, the degradation potential determined by comparing the molecule structure and orbital hydride is in line with the experimental results of Liou et al. (2003).

Effi ciency of NA removal
All the reactions follow the pseudo-fi rst-order kinetic model. Rate constant k reactions were calculated from the linear regression Eq. (5) with R2 higher than 0.98, as reported in Table 3. NAs removal from aqueous solutions by the three AOPs was observed in descending order as follows:   With the same process treatment, the values of TET, RDX and HMX were diff erent, in some cases even of one order of magnitude of diff erence. Considering the reactions with HO * radical are indiscriminate and it is also highly reactive with organics hence there no distinction about k. In term of the wayward feature of both contaminants, their similar structure does not have any big diff erence. In the real water effl uent, the large diff erence about the concentration of TET (38.46 mg L -1 ), RDX (19.22 mg L -1 ) and HMX (5.28 mg L -1 ) since the variation of k values. It is necessary to consider that HO * radical is vying exhausted by the several of organics in solution, even though there is the pseudo-fi rst order kinetics non self-reliant to the original concentration. As the results, these toxic waste in higher concentration will be privileged degraded.

Electrical energy consumption
The electrical energy consumption of the three AOPs system is shown in Table 3. It can be seen that higher electrical energy consumption corresponds to lower effi ciency of NA removal. From the research, the lowest energy consumption for the greatest pollutant removal is the photo-Fenton process. Thus, with the photo-Fenton process and a rate constant of 0.18018, 0.1501 and 0.09336

Mineralization and toxicity evaluation on NA degradation of photo-Fenton at pH=6
It is true that mineralization is a complete degradation of organics and relevant by-products to CO 2 , H 2 O, and other mineral oxides. In-situ mineralization of organic pollutants is a crucial advantages of AOPs compared with physical and biological methods. This test was shown on treated NA solution (by photo-Fenton under optimum conditions). Figure 6 determines that the photo-Fenton process at pH=6 is able to mineralize TET, RDX and HMX solutions, by around 59.7%, 45.1% and 25.8%, respectively, after 60 min.
Following Drzyzga et al. (1995), the EC 50 after 30 min for TET, RDX, and HMX with V. fi scheri were determined to be 0.45 mg L -1 , 74.56 mg L -1 , and 25 mg L -1 , respectively. The toxicities for V. fi scheri of NAs were investigated during the photo-Fenton process (Fig. 7). The inhibition ratio of TET was highest: 100% for the control sample (TET = 38.46 mg L -1 ), then dropped signifi cantly with increasing reaction time of the photo-Fenton process. After 30 and 60 min, no toxicity was observed. This result indicated that the toxicity for V. fi scheri was attributed to decreasing TET concentration and the transformation products of TET becoming less toxic than TET itself. Similar results were observed for photo-Fenton treatment at pH=6 of RDX and HMX using the luminescent bacteria test (V. fi scheri). The acute test with V. fi scheri showed no sensitivity to the after 30-min  and 60-min treatment by the photo-Fenton system. Eventhough V. fischeri is usually described to be sensitive to organic pollutants in effluent, toxicity occurs in the parts-per-million range concentration (Larsson et al. 2007).