University of New Haven Digital Commons @ New Haven Civil Engineering Faculty Publications Civil Engineering 6-16-2018 Characterization of Mg-based Bimetal Treatment of Insensitive Munition 2,4-dinitroanisole Emese Hadnagy University of New Haven, ehadnagy@newhaven.edu Andrew Mai Stevens Institute of Technology Benjamin Smolinski United States Army Washington Braida Stevens Institute of Technology Agamemnon Koutsospyros University of New Haven, akoutsospyros@newhaven.edu Follow this and additional works at: https://digitalcommons.newhaven.edu/civilengineeringfacpubs Part of the Civil Engineering Commons Publisher Citation Hadnagy, E., Mai, A., Smolinski, B., Braida, W., & Koutsospyros, A. (2018). Characterization of Mg-based bimetal treatment of insensitive munition 2,4-dinitroanisole. Environmental Science and Pollution Research, 1-14. Comments This is the authors' accepted manuscript of the article published in Environmental Science and Pollution Research. The article of record can be found at http://dx.doi.org/10.1007/s11356-018-2493-1. 1 Characterization of Mg-based Bimetal Treatment of Insensitive Munition 2,4- 2 dinitroanisole 3 4 Emese Hadnagy1,*, Andrew Mai2, Benjamin Smolinski3, Washington Braida2, Agamemnon 5 Koutsospyros1 6 7 * Corresponding Author. Email address: EHadnagy@newhaven.edu Department of Civil and Environmental Engineering, University of New Haven 2 Department of Civil, Environmental, and Ocean Engineering, Stevens Institute of Technology 3 RDECOM-ARDEC 1 Distribution A: Approved for Public Release; Distribution is Unlimited 2 8 Abstract 9 The manufacturing of insensitive munition 2,4-dinitroanisole (DNAN) generates waste streams 10 that require treatment. DNAN has been treated previously with zero-valent iron (ZVI) and Fe- 11 based bimetals. Use of Mg-based bimetals offers certain advantages including potential higher 12 reactivity and relative insensitivity to pH conditions. This work reports preliminary findings of 13 DNAN degradation by three Mg-based bimetals: Mg/Cu, Mg/Ni, and Mg/Zn. Treatment of 14 DNAN by all three bimetals is highly effective in aqueous solutions (>89% removal) and 15 wastewater (>91% removal) in comparison to treatment solely with zero-valent magnesium 16 (ZVMg; 35% removal). Investigation of reaction byproducts supports a partial degradation 17 pathway involving reduction of the ortho or para nitro- to amino- group, leading to 2-amino-4- 18 nitroanisole (2-ANAN) and 4-amino-2-nitroanisole (4-ANAN). Further reduction of the second 19 nitro group leads to 2, 4-diaminoanisole (DAAN). These byproducts are detected in small 20 quantities in the aqueous phase. Carbon mass balance analysis suggests near complete closure 21 (91%) with 12.4% and 78.4% of the total organic carbon (TOC) distributed in the aqueous and 22 mineral bimetal phases, respectively. Post treatment surface mineral phase analysis indicates 23 Mg(OH)2 as the main oxidized species; oxide formation does not appear to impair treatment. 24 25 Keywords: bimetal, magnesium, insensitive munition, reduction, DNAN, wastewater Distribution A: Approved for Public Release; Distribution is Unlimited 3 26 1. Introduction 27 The quest for safe munitions has led to the development of new formulations, designated 28 as insensitive munitions (IMs), based on components that are less prone to accidental detonation. 29 Manufacturing and handling of these IMs generate waste streams containing mixtures of IMs and 30 their manufacturing and transformation byproducts that require further treatment. One specific 31 IM component, 2,4-dinitroanisole (DNAN) has seen heavy use, and thus has garnered research 32 interests in different treatment methods to degrade this target compound in waste streams. 33 Degradation of pure DNAN by ZVI (Hawari et al., 2015) and its photodegradation (Rao 34 et al., 2013a; Arthur et al., 2017) have been reported. In addition, extensive research on the 35 degradation of DNAN in IM wastewater has been conducted. These studies have evaluated 36 various technologies including: phytoremediation (Shih et al., 2009), aerobic biodegradation 37 (Fida et al., 2014), ZVI/Fenton treatment (Liu et al., 2015), ZVI/anaerobic digestion (Ahn et al., 38 2011), Fe/Cu bimetal/Fenton treatment (Shen et al., 2013), and reduction by Fe/Cu 39 (Koutsospyros et al., 2012; Kitcher et al., 2017). Treatment of DNAN by bimetals typically 40 exhibits several advantages when compared to other technologies including extremely rapid 41 degradation kinetics leading to high removal efficiency. DNAN degradation with Fe-based 42 bimetals exhibited fast degradation with complete removal in several minutes (Kitcher et al., 43 2017). Although treatment of DNAN with Fe-based reagents (ZVI or bimetals) has been 44 demonstrated, the potential use of a similar reagent (i.e. Mg-based bimetals) has not yet been 45 explored. 46 The bimetal technology is based on enhancing the reactivity of a zero-valent base metal 47 by close contact (i.e. coating) with a catalytic metal to create a galvanic cell. Both Mg and ZVI 48 have been combined with various catalytic metals to produce reductive bimetal systems that have Distribution A: Approved for Public Release; Distribution is Unlimited 4 49 treated effectively halogenated compounds and nitro-based explosives (Morales et al., 2002; 50 DeVor et al., 2009; Begum and Gautam, 2011; Koutsospyros et al., 2012; Liu et al., 2015). 51 Specifically, Mg-based bimetals are an emerging technology for the treatment of various organic 52 (Gautam and Suresh, 2007; DeVor et al., 2008; Agarwal, Al-Abed and Dionysiou, 2009; Ghauch 53 and Tuqan, 2009) and inorganic contaminants (Ramavandi et al., 2011). Magnesium has 54 attracted additional interest due to its greater electrode potential than iron. In the hydrogenation 55 of phenol to cyclohexane and cyclohexanone, Mg/Pd was found more effective than Fe/Pd, and 56 Mg0 was more effective than Fe0 (Morales et al., 2002). In addition to the selection of a base 57 metal, the choice of catalytic metal can improve treatment effectiveness by increasing the 58 galvanic potential difference between the pair. Some researchers have utilized noble metals to 59 increase the galvanic potential difference such as Pd, Ag, and Au (Cwiertny et al., 2006; DeVor 60 et al., 2008; Patel and Suresh, 2008; Coutts et al., 2011; Saitta et al., 2015). Catalytic metal 61 selection criteria may be expanded to include economic (e.g. cost), sustainability (e.g. relative 62 abundance, available deposits) and environmental (e.g. regulatory levels) considerations. In this 63 respect, other more inexpensive and readily available metals, such as Cu, Ni, and Zn may be 64 attractive alternatives for use in bimetal formulations. 65 Similar to many other organic compounds, treatment of DNAN by chemical or biological 66 methods may generate transformation byproducts. Identification of byproducts is critical for 67 unveiling the contaminant degradation pathway and establishing that transformed products are 68 toxicologically and environmentally more benign than the parent contaminant. Treatment 69 methods such as photodegradation (Rao et al., 2013b; Hawari et al., 2015; Taylor et al., 2017) 70 and aerobic biodegradation (Fida et al., 2014; Karthikeyan and Spain, 2016) are typically 71 oxidative. Conversely, typical transformation pathways in treatment with ZVI, ZVMg or Fe- and Distribution A: Approved for Public Release; Distribution is Unlimited 5 72 Mg-based bimetals indicate reductive chemistry, as observed for example in the reduction of 73 nitrate to nitrite (Ileri, Ayyildiz and Apaydin, 2015; Khalil et al., 2016), Cr(VI) to Cr(III) 74 (Rivero-Huguet and Marshall, 2009), and the reductive dechlorination of PCBs (Hadnagy, Rauch 75 and Gardner, 2007; Agarwal et al., 2009; Coutts et al., 2011). Reduction of nitro groups by ZVI 76 or Fe-bimetals in various energetics has been demonstrated in the literature. Examples include 77 treatment of 1,3,5-trinitroperhydro-1,3,5-triazine (RDX) by Fe/Cu (Koutsospyros et al., 2012), 78 dinitrophenol (DNP) and dinitrochlorobenzene (DNCB) by Fe/Cu (Liu et al., 2015) and DNAN 79 by ZVI and Fe/Cu (Ahn et al., 2011; Hawari et al., 2015; Kitcher et al., 2017). 80 In the present work, reductive degradation of DNAN is reported using Mg-based bimetals 81 containing relatively inexpensive and readily available secondary (i.e. catalytic) metals. Three 82 bimetal formulations are evaluated, namely Mg/Cu, Mg/Ni, and Mg/Zn and are compared to 83 degradation with ZVMg. The treatment process is evaluated in laboratory prepared DNAN 84 aqueous solutions and in wastewater. Additionally, SEM imaging, EDS, and XRD analyses are 85 used for characterization of the bimetal reagent surface of unused and used particles (i.e. before 86 and after treatment). Furthermore, identification and quantification of byproducts in the 87 dissolved and particulate phases are performed to facilitate carbon mass balance analysis. Distribution A: Approved for Public Release; Distribution is Unlimited 6 88 2. Methods 89 2.1 Chemicals and Materials 90 Solid magnesium particles (20-230 mesh, reagent grade, 98% purity), nickel(II) chloride (98% 91 purity), zinc chloride (98% purity) and glacial acetic acid (99%+) were purchased from Sigma 92 Aldrich (St. Louis, MO). Copper(II) chloride (99%), acetonitrile (99.5%, ACS grade), glass fiber 93 filter paper (<1 micron, 55 mm) and nylon filter paper (0.45 micron, 55 mm) were purchased 94 from Fisher Scientific (Waltham, MA). Syringe filters (0.45 micron, nylon) were purchased from 95 Achemtek (Worcester, MA). DNAN solids and DNAN, RDX, and NQ (nitroguanidine) 96 standards dissolved in acetonitrile were obtained from Picatinny Arsenal (Wharton, NJ). DNP 97 standard dissolved in methanol, DAAN (2,4-diaminoanisole) solid standard, and 2-ANAN (2- 98 amino-4-nitro-anisole, 98%) were purchased from Sigma Aldrich (St. Louis, MO) and 4-ANAN 99 (2-nitro-4-amino-anisole, 97%) was purchased from Fisher Scientific. TOC standards were 100 purchased from Fisher Scientific (Waltham, MA). Chemical oxygen demand (COD) kits (TNT 101 821) and total nitrogen (TN) kits (TNT 826) were purchased from Hach (Loveland, CO). 102 Photometric analyses using these test kits were performed on a HACH spectrophotometer DR 103 6000 (Loveland, CO). 104 The composition of the IM wastewater, obtained from an industrial munitions facility, is 105 reported in Table 1. In addition to DNAN, RDX, NQ, and DNP were also identified and 106 quantified in the wastewater. Additional wastewater characterization included pH, inorganic 107 nitrogen species (NH3, NO2, NO3), TN, COD and TOC. Information on wastewater composition 108 is provided for completeness of information. The target compound for the present work is only 109 DNAN. Distribution A: Approved for Public Release; Distribution is Unlimited 7 110 Table 1. DNAN Wastewater Characteristics1 RDX NQ DNAN DNP NH3-N NO2-N NO3-N TN COD TOC pH -1 -1 -1 -1 -1 -1 -1 -1 -1 (mg L ) (mg L ) (mg L ) (mg L ) (mg L ) (mg L ) (mg L ) (mg L ) (mg L ) (mg L-1) 5 0.5 110 150 7.08 B.D.L. B.D.L. 3.08 47 470 120 B.D.L. = below detection limit 111 1 112 2.2 Treatment Experiments 113 Completely mixed laboratory batch experiments were conducted to evaluate the efficacy 114 of DNAN treatment using select Mg-based bimetals. All experiments were carried out in 40 mL 115 VOA vials using a 22 mL reaction volume at 0.5 % solids/liquid (S/L) ratio and 10:1 Mg to 116 secondary metal (i.e. Cu, Ni, Zn) ratio. Additional experiments for byproducts and mass balance 117 used different reaction volumes (15-132 mL) with the same S/L ratio and all other identical 118 conditions. The 0.5% S/L ratio was chosen based on previously reported work on DNAN 119 degradation using Fe-based bimetals (Koutsospyros et al., 2012). The 10:1 Mg to secondary 120 metal ratio was decided based on several other studies using Mg in bimetal formulations. Mg/Cu 121 was used in 10:1 ratio to treat azo dye (Asgari, Ramavandi and Farjadfard, 2013), and endosulfan 122 and lindane were treated with Mg/Pd at 7.5:1 and 5:1 ratios (Begum and Gautam, 2011) and at 123 50:1 ratio (Aginhotri, Mahidrakar and Gautam, 2011). In the present study, 0.11 g of Mg 124 granules, 10 mL of water and 1 mL of catalytic metal solution (22.27, 24.29, 22.94 g/L for 125 CuCl2, NiCl2, and ZnCl2, respectively) prepared in deionized water were combined and mixed on 126 a magnetic stirrer plate (Color Squid model, IKA, Wilmington, NC) at a mixing speed of 500 127 rpm for 5 min. In experiments with ZVMg, the volume of catalytic metal was replaced with 128 additional DI water. The treatment was initiated by adding 10 mL of DNAN wastewater or a 250 129 mg L-1 pure DNAN aqueous solution. After 2.5 h treatment, an aliquot of the dissolved phase 130 was analyzed by filtering the mixture with a nylon syringe filter (0.45 micron, Achemtek). Distribution A: Approved for Public Release; Distribution is Unlimited 8 131 2.2.1. Mass Balance Experiments 132 Mass balance experiments were performed to measure dissolved, adsorbed and volatilized TOC 133 and were carried out under identical treatment conditions (aqueous solutions, 0.5% S/L, 2.5 hr 134 treatment time, and 10:1 Mg to catalytic metal ratio). However, dissolved TOC measurements 135 were performed in experiments scaled by 3 (i.e. 66 mL total rather than 22 mL in previous 136 experiments). Adsorbed TOC was measured indirectly by acid digestion of the entire reaction 137 mixture (i.e. treated solution and bimetal together). Therefore, the TOC adsorbed to the bimetal 138 could be determined by subtracting the dissolved TOC from the combined adsorbed and 139 dissolved TOC measurements. Acid digestions were performed by the addition of 1 mL of 140 sulfuric acid (technical grade, 95% purity). Additionally, experiments for adsorbed TOC 141 measurements used an adjusted synthesis step that used less water (i.e. 5 mL instead of the 142 previous 11 mL). Volatilized TOC were analyzed qualitatively by GC-MS by capturing the gas 143 in multilayer foil gas bags (Supelco, Bellefonte, PA), however gaseous species were not 144 detected. 145 2.3 Analytical Methods 146 DNAN was analyzed by reversed phase high pressure liquid chromatography (HPLC) on an 147 Agilent 1260 HPLC instrument (Santa Clara, CA) equipped with a Grace Alltech Adsorbosphere 148 HS C-18 (5μm, 250x4.6mm) and a DAD detector (i.e. HPLC-DAD). The mobile phase was an 149 isocratic mixture of methanol:water at 70:30 (v/v), pumped at 1 mL min-1; the injection volume 150 was 30 μL of sample; the analytical wavelength was 300 nm (optimal absorbance wavelength for 151 DNAN). At these conditions, DNAN eluted at 4.1 min. 152 153 Quantification of 2,4-dinitrophenol (DNP) on HPLC was based on an isocratic flow using a solvent of 20% methanol and 80% water at a flow rate of 1 mL min-1; DNP eluted at 2.5 min. Distribution A: Approved for Public Release; Distribution is Unlimited 9 154 Simultaneous measurements of DNAN, 2-ANAN, 4-ANAN and DAAN were performed on the 155 same column and detector. A separate analytical method was developed for this analysis, which 156 used a 5 min hold of 90:10 water-methanol mobile phase, followed by a 50 min gradient to 10% 157 water, 90% methanol and with a 5 min hold of 10% water, 90% methanol pumped at a flow rate 158 of 1 mL min-1. At these conditions, the elution times were: DAAN at 5.5 min, 4-ANAN at 19 159 min, 2-ANAN at 26 min, and DNAN at 32 min. The analytical wavelength used was 254 nm (a 160 wavelength at which all four compounds of interest absorb well). The gradient method was used 161 due to greatly different hydrophobicity and, therefore, significantly different retention times of 162 DAAN and DNAN. TOC was measured via a UV-Persulfate TOC Analyzer Phoenix 8000 instrument from 163 164 Teledyne Tekmar (Mason, OH). Identification of byproducts was performed using electrospray 165 ionization tandem mass spectrometry (ESI-MS/MS) on a Waters Quattro Ultima (Milford, MA), 166 i.e. through direct injection of the sample without any separation. Analyses were performed in 167 both positive and negative ionization modes. Tandem mass spectrometry (MS/MS) was essential 168 due to the injection of mixtures (e.g. treated pure compound generating several byproducts) and 169 also in acquiring the necessary daughter spectra for compound identification. ESI-MS, combined 170 with front-end separation with HPLC (i.e. HPLC-ESI-MS), was used for additional confirmation 171 (HPLC: Agilent 1100 Series, Santa Clara, CA; MS: Waters Micromass ZQ instrument, Milford, 172 MA). 173 2.4 Surface Characterization 174 XRD patterns were acquired on a Rigaku Ultima IV X-Ray diffractometer (The Woodlands, 175 TX). Scans acquired were from 5 to 65 2θ with an increment of 0.03 θ and scan speed of 2 176 seconds. The x-ray conditions were 40 kV and 40 mA. These were the recommended standard Distribution A: Approved for Public Release; Distribution is Unlimited 10 177 method and conditions according to the manufacturer. SEM images were obtained with focus ion 178 beam scanning electron microscopy (FIB-SEM), and EDS analyses were performed with a 179 silicon drift detector (SDD) both on a Zeiss Auriga instrument (Oberkochen, Germany). 180 The surface of the bimetals was examined by SEM, XRD, and EDS analyses both before and 181 after treatment, referred to as unused and used particles, respectively. One sample for each 182 bimetal (i.e. Mg/Cu, Mg/Zn, and Mg/Ni) was prepared. Unused particles were synthesized under 183 the same conditions as used ones except without the addition of DNAN. Treated bimetal solids 184 were separated from the liquid by vacuum filtration and allowed to dry on glass slides for 30-60 185 min. Unused samples were decanted and dried overnight to ensure complete dryness. 186 187 188 189 3. Results and Discussion 3.1 Bimetal Synthesis and Characterization Bimetal particles were synthesized in this work and, therefore, surface characterization 190 was required to ensure that the catalytic metal had coated the base metal. SEM imaging with 191 backscatter detection allowed the detection of the heavier catalytic metals (i.e. Cu, Ni, and Zn), 192 which appeared brighter than the less heavy base metal Mg on the images. For each bimetal 193 configuration, successful coating of the base metal by the catalytic metal was observed. Solid Cu 194 nanoparticles (<100nm) coated the Mg (Figure 1a). A contrasted and zoomed-in image allowed 195 better observation of the bright Cu nanoparticles (Figure S.1). Other studies that evaluated the 196 Mg/Pd bimetal found small islands of Pd deposits (i.e. 50-100 nm) on the Mg surface using the 197 same imaging technique (Agarwal, Al-Abed and Dionysiou, 2007). In the present work, solid Zn 198 was coated on the Mg in the form of larger micron-sized particles (Figure 1b). This figure is 199 presented at smaller magnification in order to optimally view the Zn deposits against the Mg Distribution A: Approved for Public Release; Distribution is Unlimited 11 200 base metal. In contrast to the Cu and Zn particles, Ni was observed to coat the Mg uniformly, an 201 observation supported by the lack of distinct structures on the surface of the Mg/Ni bimetal 202 (Figure 1c). 203 204 Fig. 1 SEM images of catalytic metal coating on the Mg base metal: (a) Mg/Cu: Cu nanoparticles 205 (‘bright spots’), (b) Mg/Zn: micron-sized Zn deposits, and (c) Mg/Ni: uniform Ni coating 206 207 208 3.2 Degradation of Pure DNAN in the Aqueous Phase The degradation of the target compound DNAN was first examined in laboratory-made 209 aqueous solution. This was done to isolate the behavior of the compound from the wastewater 210 matrix. Treatment of aqueous solutions of pure DNAN with any of the three bimetals resulted in 211 significantly higher extent of removal compared to that of ZVMg alone, i.e. without the addition 212 of a catalytic metal (Figure 2). The treatment efficiency of the Mg/Cu, Mg/Zn and Mg/Ni 213 bimetal configurations was 100%, 95% and 89% removal, respectively. ZVMg performed poorly 214 at a removal efficiency of 35%. Poor removal efficiencies of ZVMg and ZVI with systems near 215 neutral pH have been reported for nitrate and Cr(VI) reduction, by Khalil et al. (2016) and 216 Rivero-Huguet et al. (2009). Furthermore, enhancement of reductive degradation by addition of a 217 catalytic salt has been reported for Fe-based bimetals (Rivero-Huguet and Marshall, 2009; Xiong 218 et al., 2015; Khalil et al., 2016) and Mg-based bimetals (Solanki and Murthy, 2011; Saitta et al., 219 2015) for various inorganic and organic contaminants. In the present work, degradation by 220 ZVMg was evidently similarly enhanced with the addition of the catalytic metal. 221 222 Fig. 2 Pure DNAN removal (%) in the aqueous phase after bimetal treatment (0.5% S/L, 10:1 223 Mg to catalytic metal ratio, and 2.5 h treatment time) compared to ZVMg treatment Distribution A: Approved for Public Release; Distribution is Unlimited 12 224 225 After treatment with ZVMg and the bimetals, the final pH was higher than that of the 226 original DNAN aqueous solution. This was likely due to the consumption of protons (H+) during 227 the chemical reduction (Begum and Gautam, 2011; Khalil et al., 2016). Furthermore, treatment 228 with any of the three bimetals equilibrated to a final pH in the range of 9.9-10.2, while treatment 229 with ZVMg resulted in a higher final pH of 10.7 (Table 2). Reduction of nitrate by ZVI also 230 generated a higher pH than treatment by Fe/Cu (Khalil et al., 2016). Oxidized species of Cu and 231 Fe (i.e. CuFe2O4*Fe3O4) had formed, and these side reactions likely generated protons; therefore, 232 the final pH was lower in the Fe/Cu treated system. Similar reactions may have occurred during 233 reduction with ZVMg versus a Mg-bimetal in the present study, i.e. side reactions during the 234 formation of oxidized metal species of the base metal Mg and/or catalytic metal may generate H+ 235 thereby reducing the pH. 236 Table 2. Parameters of DNAN Treatment in Two Matrices* Matrix System Initial pH** Final pH Control 5.70 (±0.10) 5.01 (±0.34) ZVMg 9.77 (±0.64) 10.72 (±0.10) Mg/Cu 9.42 (±0.53) 10.20 (±0.06) Mg/Ni 7.92 (±0.08) 9.91 (±0.08) Mg/Zn 7.03 (±0.36) 10.22 (±0.05) Control 6.64 (±0.01) 7.15 (±0.09) ZVMg 10.18 (±0.06) 10.86 (±0.06) Mg/Cu 10.43 (±0.07) 9.99 (±0.08) Mg/Ni 7.96 (±0.07) 9.93 (±0.09) Mg/Zn 8.32 (±0.57) 10.15 (±0.10) 237 238 239 240 241 Initial Final DNAN ORP** ORP Removal (mV) (mV) (%) 76 (±44) 171 (±34) 1.6 (±0.5) 82 (±36) 24 (±8) 35.1 (±4.0) -96 (±20) -96 (±12) 100.0 (±0.0) -16 (±27) -108 (±22) 88.8 (±3.6) 27 (±11) -7 (±7) 94.7 (±1.6) 196 (±64) 199 (±10) 6.2 (±0.3) -131 (±20) 36 (±4) 12.9 (±2.0) -295 (±37) -225 (±8) 100.0 (±0.0) -231 (±27) -188 (±20) 97.2 (±0.4) -108 (±55) -158 (±18) 90.5 (±3.5) *Treatment time of 2.5 h **Initial pH and initial ORP obtained for treated samples were measurements taken immediately after contact between DNAN and the reagents had been established Distribution A: Approved for Public Release; Distribution is Unlimited 13 242 243 3.3 Reaction Byproduct Identification in the Dissolved Phase To shed light to the Mg-based reductive degradation of DNAN, it is critical to identify 244 and quantify the reaction products formed. Mass spectra were acquired from treated samples in 245 both the aqueous solution and wastewater experiments using ESI-MS/MS and HPLC-ESI-MS in 246 positive and negative ionization modes. Detection of products at the attempted initial DNAN 247 concentration was not possible due to low concentrations close to detection levels. Since higher 248 initial DNAN concentrations could not be pursued due to aqueous solubility limitations, 249 experiments were set up using an alternative solvent. Products were, however, identified under 250 different conditions: 1) treatment of pure DNAN in an acetonitrile solvent matrix and 2) 251 treatment of an identified byproduct, 4-ANAN, in DI water. Products were characterized only for 252 Mg/Cu treatment. 253 3.3.1 DNAN Treated in a Solvent Matrix 254 A solvent matrix (i.e. ACN) was used to produce higher initial DNAN concentrations, which led 255 to higher, detectable concentrations of byproducts. Treatment conditions were identical to those 256 of previously mentioned experiments (i.e. 0.5% S/L ratio, 10:1 Mg to Cu ratio) except that the 257 initial stock solution contained 1,350 mg L-1 DNAN in ACN as opposed to the earlier 250 mg L-1 258 DNAN in water. In ACN, the daughter spectrum of m/z 169 in positive ionization mode 259 produced fragmentation that indicated the production of 2-ANAN and 4-ANAN during treatment 260 (Figure 3a). The daughter spectrum of m/z 139 produced fragmentation that indicated the 261 formation of DAAN (Figure S.2a). Reference spectra were acquired from aqueous solutions of 2- 262 ANAN, 4-ANAN (Figure 3b, c) and DAAN (Figure S.2b) to confirm their presence in the treated 263 DNAN sample. The peaks at m/z 169 for the 2-ANAN and 4-ANAN spectra and at m/z 139 for 264 the DAAN spectrum were identified as the protonated species, i.e. [M+H]+, because the nominal 265 masses of the neutral molecules are 168 and 138 Da, respectively. In negative ionization mode, Distribution A: Approved for Public Release; Distribution is Unlimited 14 266 no other significant peaks, other than those corresponding to 2-ANAN, 4-ANAN or DAAN, 267 were observed. 268 Fig. 3 Daughter spectrum of m/z 169 from ESI-MS/MS in positive mode from (a) after DNAN 269 treatment (solvent matrix, 0.5% S/L, 10:1 Mg to Cu ratio, 2.5 hour treatment time), (b) pure 2- 270 ANAN reference, and (c) pure 4-ANAN reference. The difference in maximum intensities are 271 neglected as comparison of relative intensities were required 272 273 DAAN was also detected when DNAN was reduced with ZVI by Hawari et al. (2015) and Ahn 274 et al. (2011) and with Fe/Cu by Liu et al., 2015 (Ahn et al., 2011; Hawari et al., 2015; Liu et al., 275 2015). Ahn et al. (2011) also identified both 2-ANAN and 4-ANAN similarly to the present 276 work (Ahn et al., 2011), while Hawari et al. (2015) only detected 2-ANAN as byproducts of 277 DNAN degradation. Hawari et al. (2015) attributed the fact that only 2-ANAN was detected in 278 their study to the regioselectivity of reduction in the ortho position, i.e. reduction more favorably 279 produced 2-ANAN over 4-ANAN; therefore, 4-ANAN generated was below the detection limit 280 (Hawari et al., 2015). Kitcher et al. (2017) confirmed reduction of nitro groups to amino groups 281 by in source deuterium exchange, which indicated m/z 169 as an amino product; it was assumed 282 that m/z 169 meant both ortho and para reduction occurred, i.e. both 2-ANAN and 4-ANAN 283 were formed (Kitcher et al., 2017). 284 In the present study, the analysis of daughter spectra revealed that 2-ANAN produced 285 mass spectral peaks of m/z (in positive mode): 169, 154, and 123. 4-ANAN produced mass 286 spectral peaks of m/z: 169, 154, 152, 123, 122, and 94. The low intensity of m/z 152 in the 287 treated sample (Figure 3a) indicates that a mixture of both 4-ANAN and 2-ANAN was present. 288 For example, if only 4-ANAN were present, the intensity of m/z 152 peak would be higher, Distribution A: Approved for Public Release; Distribution is Unlimited 15 289 while if only 2-ANAN were present, there would be no peak at m/z 152. However, there is a 290 minute peak of m/z 152 in the mass spectrum (Figure 3a). In addition, the same was observed 291 with the relative intensity of the m/z 123 and 122 peaks. If only 4-ANAN were present, the m/z 292 122 peak in Figure 3a would be relatively higher to that of m/z 123, i.e. the two peaks should 293 match the relative heights of the m/z 122 and 123 peaks in the 4-ANAN reference. Ultimately, 294 the presence of 2-ANAN effectively suppressed the intensity of the m/z 122 peak when mixed 295 with 4-ANAN. Olivares et al. (2016) was also able to differentiate between the 2-ANAN and 4- 296 ANAN isomers based on their unique fragmentation patterns. Furthermore, Hawari et al. (2015) 297 detected a hydroxylamino intermediate (i.e. 2-HA-NAN) and small amounts of a nitroso 298 intermediate (i.e. 2-NO-NAN). According to Hudlicky (1984), reduction of the nitro group in 299 nitroaromatic compounds follows the scheme: nitro (-NO2) > nitroso (-NO) > hydroxylamino (- 300 NHOH) > amine (-NH2), according to the degree of reduction. Nitroso and hydroxylamino 301 compounds are typically rarely observed under this scheme (Hudlicky, 1984). In this study, no 302 nitroso or hydroxylamino intermediates in either the ortho or para positions were identified. 303 3.3.2 Reduction of 4-ANAN to DAAN 304 The high aqueous solubility of 4-ANAN (4,400 mg L-1 at 25°C (Hawari et al., 2015)), an 305 identified byproduct of DNAN degradation in the ACN matrix, allowed investigation of further 306 byproduct formation directly in aqueous solutions, i.e. as opposed to work performed in an ACN 307 matrix. The reduction of 4-ANAN to DAAN was confirmed; treatment of 4-ANAN (800 mg L-1 308 initial concentration, 0.5% S/L, 10:1 Mg to Cu ratio, 1 hour treatment time) yielded DAAN (m/z 309 139) (Figure 4). The presence of DAAN was additionally confirmed on HPLC-ESI-MS for the 310 same treatment condition (Figure S.3). Adding front-end separation with HPLC to the MS 311 method allowed the matching of elution times of pure DAAN (observed as m/z 139) to that of Distribution A: Approved for Public Release; Distribution is Unlimited 16 312 the m/z 139 peak in treated 4-ANAN. Although the reduction of 4-ANAN to DAAN has been 313 proposed as part of the DNAN degradation pathway using ZVI (Ahn et al., 2011; Hawari et al., 314 2015) and Fe/Cu (Kitcher et al., 2017), the same has not yet been reported for Mg-based bimetal 315 treatment in the literature. Additionally, the detection of DNAN reduction to DAAN by any 316 bimetal has not yet reported until now; Kitcher et al. (2017) proposed DAAN as a probable final 317 product by Fe-bimetal reduction, but could not confirm analytically. 318 319 Fig. 4 Daughter spectrum of m/z 139 from ESI-MS/MS in positive mode of 4-ANAN treated 320 with Mg/Cu (DI water matrix, 0.5% S/L, 10:1 Mg to Cu ratio, 1 hr treatment time) showing 321 DAAN (m/z 139) with matching fragmentation of pure DAAN (Figure S.2b) 322 323 324 3.4 Reaction Pathway The proposed partial reaction pathway for DNAN reduction by Mg/Cu is: 1) reduction of 325 one nitro group, either ortho or meta position, which forms 2-ANAN or 4-ANAN, and then 2) 326 subsequent reduction of the other nitro group, which forms DAAN (Figure 5). This pathway was 327 identical to the one identified for DNAN reduction with ZVI (Ahn et al., 2011) and Fe/Cu 328 (Kitcher et al., 2017). In the present study, the byproducts and reaction pathways characterized 329 to-date were from DNAN treatment with Mg/Cu, and it is likely that identical pathways exist for 330 treatment with Mg/Ni and Mg/Zn. The proposed pathway excludes the unstable nitroso and 331 hydroxylamine derivatives discussed earlier since these compounds were not detected. 332 333 Fig. 5 Partial reaction pathways of DNAN degradation by Mg/Cu 334 Distribution A: Approved for Public Release; Distribution is Unlimited 17 335 The reduction of DNAN to byproducts 2-ANAN, 4-ANAN and DAAN were made possible by 336 electrons released by the dissolution of Mg0 as illustrated in Figure 6. In the bimetal pair, Mg/Cu, 337 Cu is the cathode while Mg is the anode. Mg thus preferentially corrodes to Mg2+ with the 338 concurrent release of 2 electrons. Electrons become available for reduction illustrated by the 339 dashed lines. In addition, Mg(OH)2 forms by the oxidation of Mg in water (Figure 6). The same 340 reactions occur in the case of Mg/Ni and Mg/Zn, where Ni and Zn are cathodic relative to the 341 Mg. 342 343 Fig. 6 Diagram illustrating the galvanic corrosion of Mg in the Mg/Cu bimetal pair resulting in 344 reduction of DNAN and oxidation of Mg to Mg(OH)2 345 346 347 3.5 Quantification of Reaction Byproducts and Mass Balance Mass balance experiments are necessary in order to fully characterize any treatment 348 system. A significantly open mass balance means undetected compounds or unknown fates, with 349 potential to toxic exposure. In the present study, preliminary mass balance data were obtained for 350 treatment of pure DNAN in water and in ACN with Mg/Cu. Since byproducts of pure DNAN 351 were not detected in the aqueous (i.e. dissolved) phase, the TOC mass balance was evaluated 352 (see Section 3.5.1). Next, experiments with pure DNAN were performed in ACN and the mass 353 balance of DNAN and its known reaction products was evaluated in the combined dissolved and 354 adsorbed phase (obtained by acid digestion of the reaction mixture), with subsequent 355 quantification of individual known compounds in the dissolved phase (see Section 3.5.2). 356 3.5.1 Overall Carbon Mass Balance Closure Using Aggregate TOC Measurements Distribution A: Approved for Public Release; Distribution is Unlimited 18 357 After treatment of pure DNAN with Mg/Cu, a 91% carbon mass balance closure was attained by 358 addition of the dissolved and adsorbed phases (i.e. adsorbed to the bimetal surface). The 359 dissolved phase and adsorbed phase contained 12.4% and 78.4% of TOC, respectively, compared 360 to the control. On the other hand, treatment of DNAN by ZVI resulted in 95% of the initial 361 DNAN mass recovered as DAAN in the aqueous phase after 1 hour of treatment (Ahn et al., 362 2011). This observed difference of mass balance between ZVI in the study of Ahn et al. (2011) 363 and Mg/Cu in the present work might stem from increased oxidization of the Mg-bimetal from 364 galvanic corrosion. The enhanced corrosion of the Mg surface during treatment may have 365 increased the total surface area and therefore increased the number of sites available for 366 adsorption. 367 This significant adsorption was validated by analysis of the sealed reactor headspace with 368 GC-MS in experiments conducted with aqueous solutions of pure DNAN. A significant amount 369 of gas (approximately 0.5-1.0L) was generated (with reaction volumes of 132 mL), but no 370 reaction byproducts were identified. The gas captured was speculated to be mainly H2 gas, an 371 expected gas formed in the dissolution of Mg or Fe in water (Patel and Suresh, 2007; Lee and 372 Park, 2013; Nie et al., 2013). 373 3.5.2 Mass Balance of Dissolved Byproducts: Treatment in Solvent Matrix 374 Byproducts in the dissolved phase have been qualitatively detected. However, quantification of 375 dissolved byproducts was necessary to close the mass balance in the liquid phase to ensure that 376 the products detected (2-ANAN, 4-ANAN and DAAN) accounted for all dissolved TOC. 377 Significant open mass balance in the liquid phase would indicate formation of additional 378 unidentified products. Byproducts in the dissolved phase after DNAN treatment in aqueous 379 solutions could not be detected by MS or HPLC. However, byproducts 2-ANAN and 4-ANAN Distribution A: Approved for Public Release; Distribution is Unlimited 19 380 were detected in small amounts for DNAN treatment by Mg/Cu in ACN (Figure 7). Furthermore, 381 dissolved DAAN was detected on ESI-MS, but was not detected on the HPLC-DAD used for 382 quantification, and therefore the bar for DAAN was marked with an asterisk (*) to indicate future 383 work is needed for analytical method development for this compound. Mass balance data from 384 aggregate measurements after treatment in the solvent matrix could not be supplemented with 385 COD and TOC removals due to the significant addition of ACN. Ultimately, the small amounts 386 of 2-ANAN and 4-ANAN found after treatment in the ACN matrix and the undetectable amounts 387 of the same compounds in the aqueous matrix could indicate that DAAN may further degrade to 388 other compounds. Furthermore, the higher concentration of 2-ANAN than that of 4-ANAN on 389 Figure 7 maybe suggestive of regioselectivity of the reduction reaction for the ortho rather than 390 the para isomer. This remains to be resolved and quantified in future work. 391 392 Fig. 7 Recovered mass of DNAN and products in the dissolved phase after treatment of DNAN 393 with Mg/Cu (ACN matrix, 0.5% S/L, 10:1 Mg to Cu ratio, 2.5 hr treatment). Zero amounts of 394 DAAN marked with (*) indicate quantification of this compound requires additional work. 395 Corresponding chromatogram of treated DNAN in upper right; DAAN was not detected on 396 HPLC-DAD Distribution A: Approved for Public Release; Distribution is Unlimited 20 397 398 3.6 Characterization of DNAN Treatment in Wastewater Parameters examined for DNAN-laden wastewater treatment included DNAN removal, 399 pH, ORP, TOC and COD reduction, and removal of dinitrophenol (DNP) (DNP was only 400 evaluated qualitatively), another major wastewater constituent. Observed DNAN removals in 401 wastewater were similar to those of in the pure DNAN solutions. In the wastewater, high 402 removals were achieved by Mg/Cu, Mg/Ni, and Mg/Zn (i.e. 100%, 97% and 91%, respectively), 403 while treatment with ZVMg resulted in poor degradation efficiency of 12.9% (Figure 8). 404 However, treatment with Mg/Zn resulted in lower DNAN removal than treatment with Mg/Ni in 405 wastewater, whereas the opposite trend was observed in aqueous solutions (Table 2). 406 407 Fig. 8 DNAN removal (%) in wastewater after bimetal treatment (0.5% S/L, 10:1 Mg to catalytic 408 metal ratio, and 2.5 h treatment time) compared to ZVMg 409 410 Based on removal efficiency, the reductive activity of the various bimetal systems was 411 ranked in the following order (treatment of aqueous solutions of pure DNAN): 412 Mg/Cu>Mg/Zn>Mg/Ni>ZVMg (Table 2). There was no correlation between the final ORP 413 values and the DNAN removal efficiencies. On the other hand, a correlation has been observed 414 between these variables for wastewater experiments. In the wastewater matrix, the final ORP 415 values showed that the Mg/Cu bimetal pair exhibited the most negative value (-225 mV), 416 indicating that this system generated the most reductive environment. Based on ORP values, the 417 bimetal systems were ranked in the following order (most negative shown first): 418 Mg/Cu<Mg/Ni<Mg/Zn<ZVMg, which correlated with the DNAN removal efficiency (i.e. higher 419 DNAN removal was achieved under more reducing conditions). The reduction of nitrate by ZVI Distribution A: Approved for Public Release; Distribution is Unlimited 21 420 and Fe/Cu also created a reductive environment as indicated by negative ORP values of around - 421 700 mV (Khalil et al., 2016); the difference in ORP values between reduction with ZVI and 422 Fe/Cu was not significantly different. Final pH values in wastewater were similar to those 423 observed in pure DNAN aqueous solutions: pH of 9.9-10.2 for the bimetals and somewhat higher 424 pH of 10.9 for ZVMg (Table 2). 425 426 3.7 Removal of TOC and COD The effectiveness of wastewater treatment using bimetals was also measured by removal 427 428 of aggregate parameters TOC and COD. Aggregate parameters were selected because although 429 the major constituents were DNAN, DNP, and RDX, small amounts of other unknown 430 compounds accounted for approximately 10% of wastewater TOC. All three bimetal systems were capable of high DNAN removal of 90-100% (Figure 8) 431 432 and TOC and COD removal of 60-70% (Figure S.4). Empirical measurements of TOC and COD 433 from solutions of DNAN at various concentrations allowed calculation of DNAN-derived TOC 434 and COD. The majority of the remaining TOC and COD after treatment were, therefore, DNAN 435 byproducts and/or byproducts derived from DNP and RDX. In addition, the overall amount of 436 removed TOC and COD exceeded that originating from DNAN. Therefore, the system treated 437 the other organic contaminants as well as DNAN. The remaining TOC and COD may also 438 include species that were more resistant to the bimetal treatment. This was corroborated by a 439 similar level of TOC and COD removal regardless of bimetal configuration and despite the larger 440 DNAN removal capability of Mg/Cu. COD removal was not measured for treatment with 441 ZVMg. 442 Furthermore, the degradation of DNP by Mg/Cu, Mg/Ni and Mg/Zn, was qualitatively 443 observed by superimposing HPLC chromatograms of wastewater at before and after treatment Distribution A: Approved for Public Release; Distribution is Unlimited 22 444 (0.5% S/L, 10:1 Mg to catalytic metal ratio) (Figure S.5). The intensity of the chromatographic 445 peak that correlated to DNP (2.1 min) decreased drastically indicating almost complete removal 446 after treatment with each bimetal. The same result was observed for DNAN (4.2 min). 447 448 449 3.8 Effect of Treatment on the Bimetal Surface The characterization of the reagent bimetal surface after use was performed to assess any 450 changes to the reagent. For bimetal systems and especially Fe-based bimetals, oxidation of the 451 bimetal surface has led to passivation and inhibition of treatment (Rivero-Huguet and Marshall, 452 2009; Fu, Cheng and Lu, 2015; Sun et al., 2016). In this work, surface analyses of used Mg- 453 bimetal particles included SEM imaging, elemental mapping through EDS, and XRD analysis. 454 EDS results indicated significant elemental oxygen on the surface of all tested bimetal 455 configurations. Therefore, the identification and relative quantification of the mineral species 456 containing oxygen was performed by XRD analysis. 457 3.8.1 Elemental Analyses of Oxidation with EDS 458 The EDS map of an unused particle of Mg/Cu (Figure 9) showed that oxygen was distributed 459 according to the topography of the Mg surface, which suggested oxidation of the Mg surface 460 (Figure 9b, c). Therefore, the surface of Mg/Cu particles had become oxidized during the bimetal 461 synthesis step before treatment. Oxidation of the surface was likely due to galvanic corrosion 462 after deposition of Cu0. The Cu nanoparticles were not clearly detectable on EDS mapping 463 (Figure 9d) but were observed on SEM imaging (Figure 1a). 464 465 Fig. 9 EDS mapping of (a) a sample region of an unused Mg/Cu particle pictured by SEM of an 466 elucidating distribution of (b) primary metal Mg to (c) oxygen, (d) and catalytic metal Cu Distribution A: Approved for Public Release; Distribution is Unlimited 23 467 468 After treatment (i.e. on used particles), identification of Cu nanoparticles on SEM images and 469 EDS mapping was not as certain due to the ‘rougher’ particle appearance resulting from 470 corrosion (Figure 10a). However, EDS elemental analysis still showed the presence of Cu on the 471 treated particle (Figure 10d, Table S.1). Additionally, oxidation of the bimetal surface had 472 increased slightly after treatment according to EDS compositional relative quantification (Table 473 S.1). While EDS compositional analyses are generally used for smoother and flat sample 474 surfaces (whereas the analyzed bimetal surfaces have varying topography), the compositions 475 obtained by EDS were still considered one effective measure of overall oxidation as seen in other 476 bimetal literature (Shih et al., 2009; Xu et al., 2012; Nie et al., 2013). In the present study, 477 similar results to that of EDS analyses of Mg/Cu particles were found for Mg/Ni and Mg/Zn 478 particles; i.e. oxidation was evident to some extent on the surface of unused and was more 479 pronounced on used particles (Table S.1; EDS elemental mappings of these particles are 480 provided in the supplemental data, Figures S.6-S.9). 481 482 Fig. 10 EDS mapping of (a) SEM of sample region of a used Mg/Cu particle elucidating 483 distribution of (b) primary metal Mg to (c) oxygen, and (d) catalytic metal Cu 484 485 3.8.2 Identification and Relative Quantification of Oxidized Species 486 The major oxidized species was identified as Mg(OH)2 through XRD analysis for all three 487 bimetal configurations. Furthermore, MgO formation was not observed. The oxidation of the 488 base metal Mg was consistent with galvanic corrosion when in contact with the catalytic metal 489 (i.e. Cu, Ni or Zn). Distribution A: Approved for Public Release; Distribution is Unlimited 24 490 The presence of Mg(OH)2 was corroborated on the surface of all three unused bimetal 491 particles, however, to a lesser extent on Mg/Ni and Mg/Zn than on Mg/Cu. The reduced 492 formation of Mg(OH)2 on Mg/Ni and Mg/Zn is in agreement with the lower galvanic potential 493 difference for these bimetal systems compared to Mg/Cu (Figure S.10a,b-11a,b). After treatment 494 of wastewater, the amount of Mg(OH)2 on the surface of Mg/Cu increased significantly (Figure 495 10a,b), while the opposite trend was observed on the surface of used Mg/Ni and Mg/Zn particles. 496 This may be due to the unknown ions in the wastewater, which may have solubilized part of the 497 Mg(OH)2 formed on the latter two bimetals. However, in the case of Mg/Cu, the formation of 498 hydroxide from high galvanic potential difference likely exceeded the rate of Mg(OH)2 499 dissolution due to unknown ions, therefore resulting in the expected increase of hydroxide 500 formation. This is also supported by observations made on the bimetal surfaces after treatment of 501 aqueous solutions of pure DNAN (i.e. a matrix lacking additional ions), where the trend in 502 oxidation of Mg for each bimetal configuration showed the expected outcome, i.e. the amount of 503 Mg(OH)2 increased after 2.5 hours of treatment (Figures 11c, S.10c, S.11c). Further analysis on 504 the ionic character and composition of the wastewater is required. 505 506 Fig. 11 XRD patterns of Mg/Cu (a) before treatment, (b) after treatment in wastewater, and (c) 507 after treatment in aqueous solutions 508 509 The decrease of Mg(OH)2 on used particles of Mg/Ni and Mg/Zn may contradict the 510 findings from EDS, which showed that overall oxygen content/oxidation increased. However, it 511 is possible that XRD could not detect other oxidized species since XRD is a technique that 512 detects crystalline materials. Distribution A: Approved for Public Release; Distribution is Unlimited 25 513 In addition, there were unidentified species (small peaks at 33 and 43 2Ө) that were not 514 matched to any oxidized forms of Cu or Mg with Cu. In the treatment of nitrate by Fe/Cu (Khalil 515 et al., 2016), species of Fe and Cu oxides were detected, specifically Fe3O4 and CuFe2O4●Fe3O4. 516 However, no other oxidized species of Cu were detected in the present work. Furthermore, 517 micron-sized deposits of Zn0 were observed, likely due to their larger particle size and increased 518 detection on XRD. Zn0 was the only zero-valent catalytic metal detected in the present work. 519 Despite significant oxidation of the Mg surface, treatment with the Mg-bimetals produced 520 high removals of DNAN. During treatment with ZVI or Fe-bimetals, the formation of 521 (oxy)hydroxides or oxides of iron passivate the surface causing inhibition of treatment (Fu, 522 Cheng and Lu, 2015; Sun et al., 2016). In the treatment with various Fe-bimetal configurations, 523 the passivated Fe-bimetals led to separate modelling of a slower kinetics and “inhibited” reaction 524 rate after an initial period of faster kinetics (Rivero-Huguet and Marshall, 2009). Generally, a 525 low pH condition for various Fe-based bimetals have been required in order to overcome 526 passivation by oxidation (Xu et al., 2005; Rivero-Huguet and Marshall, 2009; Luo et al., 2010). 527 However, in the present work, an unadjusted pH condition (Table 2) in each bimetal treatment 528 system still produced effective DNAN removal. 529 530 4. Conclusions 531 The preliminary work performed in this study indicates that Mg-based bimetals are effective 532 reagents for the degradation of DNAN and other energetic compounds often present in munitions 533 facilities waste streams. The use of inexpensive catalytic metals to generate effective Mg-bimetal 534 reagents was demonstrated by the high removals of DNAN, TOC and COD. Oxidation of the 535 Mg-bimetal surface did not inhibit the treatment, as opposed to oxidation on ZVI or Fe-bimetals 536 passivating the surface and inhibiting treatment. Future work may be performed on the impact of Distribution A: Approved for Public Release; Distribution is Unlimited 26 537 catalytic metal dose relative to the base metal and the coating method during bimetal synthesis. 538 These parameters subsequently may be correlated to overall removals and removal kinetics. In 539 addition, the reaction occurred at neutral to basic pH, i.e. without the need to lower the pH of the 540 treated solution as is commonly the case with Fe-based bimetals. This eliminates the need for an 541 additional chemical in a treatment scenario. Byproduct identification and subsequent 542 determination of the reaction pathways demonstrated that DNAN was reduced to amino 543 derivatives 2-ANAN, 4-ANAN and subsequently to DAAN. The 91% closure for the carbon 544 mass balance indicated that the dissolved and adsorbed phases contained 12.4% and 78.4% TOC, 545 respectively. The significant carbon adsorbed to the Mg-bimetal surface requires further 546 investigation in either the extraction of these adsorbed compounds or direct analysis of used Mg- 547 bimetal particles. In both the dissolved and adsorbed phases, after complete product 548 identification, closure of the mass balance should be obtained. Furthermore, the proposed end- 549 product DAAN requires additional confirmation (i.e. evidence that further treatment of DAAN 550 does not form another compound). 551 552 Acknowledgements 553 This work was supported by the Consortium for Energy, Environment and Demilitarization 554 (CEED) contract number SINIT-15-0013. Electrospray ionization mass spectra were obtained in 555 the Center for Mass Spectrometry of Department of Chemistry and Chemical Biology of Stevens 556 Institute of Technology. 557 558 References 559 Agarwal, S. et al. (2009) ‘Reactivity of substituted chlorines and ensuing dechlorination 560 pathways of select PCB congeners with Pd/Mg bimetallics’, Environmental Science and Distribution A: Approved for Public Release; Distribution is Unlimited 27 561 Technology, 43(3), pp. 915–921. doi: 10.1021/es802538d. 562 Agarwal, S., Al-Abed, S. R. and Dionysiou, D. D. (2007) ‘Enhanced corrosion-based Pd/Mg 563 bimetallic systems for dechlorination of PCBs’, Environmental Science and Technology, 41(10), 564 pp. 3722–3727. doi: 10.1021/es062886y. 565 Agarwal, S., Al-Abed, S. R. and Dionysiou, D. D. (2009) ‘Impact of organic solvents and 566 common anions on 2-chlorobiphenyl dechlorination kinetics with Pd/Mg’, Applied Catalysis B: 567 Environmental, 92(1–2), pp. 17–22. doi: 10.1016/j.apcatb.2009.07.029. 568 Aginhotri, P., Mahidrakar, A. B. and Gautam, S. K. (2011) ‘Complete dechlorination of 569 endosulfan and lindane using Mg0/Pd+4 bimetallic system’, Water Environ. Res., 570 83(September), pp. 865–873. doi: 10.2175/106143011X12928814445096. 571 Ahn, S. C. et al. (2011) ‘Detoxification of PAX-21 ammunitions wastewater by zero-valent iron 572 for microbial reduction of perchlorate’, Journal of Hazardous Materials, 192(2), pp. 909–914. 573 doi: 10.1016/j.jhazmat.2011.05.104. 574 Arthur, J. D. et al. (2017) ‘Batch soil adsorption and column transport studies of 2,4- 575 dinitroanisole (DNAN) in soils’, Journal of Contaminant Hydrology, 199, pp. 14–23. doi: 576 10.1016/j.jconhyd.2017.02.004. 577 Asgari, G., Ramavandi, B. and Farjadfard, S. (2013) ‘Abatement of Azo Dye from Wastewater 578 Using Bimetal-Chitosan’, 2013. 579 Begum, A. and Gautam, S. K. (2011) ‘Dechlorination of endocrine disrupting chemicals using 580 Mg0/ZnCl2 bimetallic system’, Water Research. Elsevier Ltd, 45(7), pp. 2383–2391. doi: 581 10.1016/j.watres.2011.01.017. 582 Coutts, J. L. et al. (2011) ‘The use of mechanical alloying for the preparation of palladized 583 magnesium bimetallic particles for the remediation of PCBs’, Journal of Hazardous Materials. Distribution A: Approved for Public Release; Distribution is Unlimited 28 584 Elsevier B.V., 192(3), pp. 1380–1387. doi: 10.1016/j.jhazmat.2011.06.052. 585 Cwiertny, D. M. et al. (2006) ‘Exploring the influence of granular iron additives on 1,1,1- 586 trichloroethane reduction’, Environmental Science and Technology, 40(21), pp. 6837–6843. doi: 587 10.1021/es060921v. 588 DeVor, R. et al. (2008) ‘Dechlorination comparison of mono-substituted PCBs with Mg/Pd in 589 different solvent systems’, Chemosphere, 73(6), pp. 896–900. doi: 590 10.1016/j.chemosphere.2008.07.006. 591 DeVor, R. et al. (2009) ‘Mechanism of the degradation of individual PCB congeners using 592 mechanically alloyed Mg/Pd in methanol’, Chemosphere. Elsevier Ltd, 76(6), pp. 761–766. doi: 593 10.1016/j.chemosphere.2009.05.007. 594 Fida, T. T. et al. (2014) ‘Aerobic biodegradation of 2,4-dinitroanisole by Nocardioides sp. strain 595 JS1661’, Applied and Environmental Microbiology, 80(24), pp. 7725–7731. doi: 596 10.1128/AEM.02752-14. 597 Fu, F., Cheng, Z. and Lu, J. (2015) ‘Synthesis and use of bimetals and bimetal oxides in 598 contaminants removal from water: a review’, RSC Adv., 5(104), pp. 85395–85409. doi: 599 10.1039/C5RA13067K. 600 Gautam, S. K. and Suresh, S. (2007) ‘Complete dechlorination of DDE/DDD using 601 magnesium/palladium system’, Water Environment Research, 79(Ddd), pp. 430–435. doi: 602 10.2175/106143006x115336. 603 Ghauch, A. and Tuqan, A. (2009) ‘Reductive destruction and decontamination of aqueous 604 solutions of chlorinated antimicrobial agent using bimetallic systems’, Journal of Hazardous 605 Materials, 164(2–3), pp. 665–674. doi: 10.1016/j.jhazmat.2008.08.048. 606 Hadnagy, E., Rauch, L. M. and Gardner, K. H. (2007) ‘Dechlorination of polychlorinated Distribution A: Approved for Public Release; Distribution is Unlimited 29 607 biphenyls, naphthalenes and dibenzo-p-dioxins by magnesium/palladium bimetallic particles.’, 608 Journal of environmental science and health. Part A, Toxic/hazardous substances & 609 environmental engineering, 42(6), pp. 685–695. doi: 10.1080/10934520701326222. 610 Hawari, J. et al. (2015) ‘Environmental fate of 2,4-dinitroanisole (DNAN) and its reduced 611 products’, Chemosphere. Elsevier Ltd, 119, pp. 16–23. doi: 10.1016/j.chemosphere.2014.05.047. 612 Hudlicky, M. (Virginia P. I. and S. U. (1984) Reductions in Organic Chemistry - Hudlicky 613 M..pdf. John Wiley & Sons. 614 Ileri, B., Ayyildiz, O. and Apaydin, O. (2015) ‘Ultrasound-assisted activation of zero-valent 615 magnesium for nitrate denitrification: Identification of reaction by-products and pathways’, 616 Journal of Hazardous Materials, 292(3), pp. 1–8. doi: 10.1016/j.jhazmat.2015.03.004. 617 Karthikeyan, S. and Spain, J. C. (2016) ‘Biodegradation of 2,4-dinitroanisole (DNAN) by 618 Nocardioides sp. JS1661 in water, soil and bioreactors’, Journal of Hazardous Materials. 619 Elsevier B.V., 312, pp. 37–44. doi: 10.1016/j.jhazmat.2016.03.029. 620 Khalil, A. M. E. et al. (2016) ‘Promoting nitrate reduction kinetics by nanoscale zero valent iron 621 in water via copper salt addition’, Chemical Engineering Journal. Elsevier B.V., 287, pp. 367– 622 380. doi: 10.1016/j.cej.2015.11.038. 623 Kitcher, E. et al. (2017) ‘Characteristics and products of the reductive degradation of 3-nitro- 624 1,2,4-triazol-5-one (NTO) and 2,4-dinitroanisole (DNAN) in a Fe-Cu bimetal system’, 625 Environmental Science and Pollution Research, 24(3), pp. 2744–2753. doi: 10.1007/s11356-016- 626 8053-7. 627 Koutsospyros, A. et al. (2012) ‘Degradation of high energetic and insensitive munitions 628 compounds by Fe/Cu bimetal reduction’, Journal of Hazardous Materials. Elsevier B.V., 219– 629 220, pp. 75–81. doi: 10.1016/j.jhazmat.2012.03.048. Distribution A: Approved for Public Release; Distribution is Unlimited 30 630 Lee, G. and Park, J. (2013) ‘Reaction of zero-valent magnesium with water: Potential 631 applications in environmental remediation’, Geochimica et Cosmochimica Acta. Elsevier Ltd, 632 102, pp. 162–174. doi: 10.1016/j.gca.2012.10.031. 633 Liu, J. et al. (2015) ‘Selective removal of nitroaromatic compounds from wastewater in an 634 integrated zero valent iron (ZVI) reduction and ZVI/H2O2 oxidation process’, RSC Advances. 635 Royal Society of Chemistry, 5, pp. 57444–57452. doi: 10.1039/C5RA08487C. 636 Luo, S. et al. (2010) ‘Reductive degradation of tetrabromobisphenol A over iron-silver bimetallic 637 nanoparticles under ultrasound radiation’, Chemosphere. Elsevier Ltd, 79(6), pp. 672–678. doi: 638 10.1016/j.chemosphere.2010.02.011. 639 Morales, J. et al. (2002) ‘Hydrogenation of Phenol by the Pd/Mg and Pd/Fe Bimetallic Systems 640 under Mild Reaction Conditions’, Industrial & Engineering Chemistry Research, 41(13), pp. 641 3071–3074. doi: 10.1021/ie0200510. 642 Nie, X. et al. (2013) ‘Rapid degradation of hexachlorobenzene by micron Ag/Fe bimetal 643 particles’, Journal of Environmental Sciences (China). The Research Centre for Eco- 644 Environmental Sciences, Chinese Academy of Sciences, 25(3), pp. 473–478. doi: 645 10.1016/S1001-0742(12)60088-6. 646 Patel, U. D. and Suresh, S. (2007) ‘Dechlorination of chlorophenols using magnesium-palladium 647 bimetallic system’, Journal of Hazardous Materials, 147(1–2), pp. 431–438. doi: 648 10.1016/j.jhazmat.2007.01.029. 649 Patel, U. D. and Suresh, S. (2008) ‘Effects of solvent, pH, salts and resin fatty acids on the 650 dechlorination of pentachlorophenol using magnesium-silver and magnesium-palladium 651 bimetallic systems’, Journal of Hazardous Materials, 156(1–3), pp. 308–316. doi: 652 10.1016/j.jhazmat.2007.12.021. Distribution A: Approved for Public Release; Distribution is Unlimited 31 653 Ramavandi, B. et al. (2011) ‘Experimental investigation of the chemical reduction of nitrate ion 654 in aqueous solution by Mg/Cu bimetallic particles’, Reaction Kinetics, Mechanisms and 655 Catalysis, 102(2), pp. 313–329. doi: 10.1007/s11144-010-0274-z. 656 Rao, B. et al. (2013a) ‘Photochemical transformation of the insensitive munitions compound 2,4- 657 dinitroanisole’, Science of the Total Environment, 443, pp. 692–699. doi: 658 10.1016/j.scitotenv.2012.11.033. 659 Rao, B. et al. (2013b) ‘Photochemical transformation of the insensitive munitions compound 2,4- 660 dinitroanisole’, Science of the Total Environment. Elsevier B.V., 443, pp. 692–699. doi: 661 10.1016/j.scitotenv.2012.11.033. 662 Rivero-Huguet, M. and Marshall, W. D. (2009) ‘Reduction of hexavalent chromium mediated by 663 micro- and nano-sized mixed metallic particles’, Journal of Hazardous Materials, 169(1–3), pp. 664 1081–1087. doi: 10.1016/j.jhazmat.2009.04.062. 665 Saitta, E. K. H. et al. (2015) ‘Case study of a non-destructive treatment method for the 666 remediation of military structures containing polychlorinated biphenyl contaminated paint’, 667 Journal of Environmental Management. Elsevier Ltd, 158, pp. 40–47. doi: 668 10.1016/j.jenvman.2015.04.038. 669 Shen, J. et al. (2013) ‘Pretreatment of 2 , 4-dinitroanisole ( DNAN ) producing wastewater using 670 a combined zero-valent iron ( ZVI ) reduction and Fenton oxidation process’, Journal of 671 Hazardous Materials, 260, pp. 993–1000. doi: 10.1016/j.jhazmat.2013.07.003. 672 Shih, Y. hsin et al. (2009) ‘Dechlorination of hexachlorobenzene by using nanoscale Fe and 673 nanoscale Pd/Fe bimetallic particles’, Colloids and Surfaces A: Physicochemical and 674 Engineering Aspects, 332(2–3), pp. 84–89. doi: 10.1016/j.colsurfa.2008.09.031. 675 Solanki, J. N. and Murthy, Z. V. P. (2011) ‘Reduction of 4-chlorophenol by Mg and Mg-Ag Distribution A: Approved for Public Release; Distribution is Unlimited 32 676 bimetallic nanocatalysts’, Industrial and Engineering Chemistry Research, 50(24), pp. 14211– 677 14216. doi: 10.1021/ie2022338. 678 Sun, Y. et al. (2016) ‘The influences of iron characteristics, operating conditions and solution 679 chemistry on contaminants removal by zero-valent iron: A review’, Water Research, 100, pp. 680 277–295. doi: 10.1016/j.watres.2016.05.031. 681 Taylor, S. et al. (2017) ‘Photo-degradation of 2,4-dinitroanisole (DNAN): An emerging 682 munitions compound’, Chemosphere, 167, pp. 193–203. doi: 683 10.1016/j.chemosphere.2016.09.142. 684 Xiong, Z. et al. (2015) ‘Comparative study on the reactivity of Fe/Cu bimetallic particles and 685 zero valent iron (ZVI) under different conditions of N<inf>2</inf>, air or without aeration’, 686 Journal of Hazardous Materials. Elsevier B.V., 297, pp. 261–268. doi: 687 10.1016/j.jhazmat.2015.05.006. 688 Xu, F. et al. (2012) ‘Highly active and stable Ni-Fe bimetal prepared by ball milling for catalytic 689 hydrodechlorination of 4-chlorophenol’, Environmental Science and Technology, 46(8), pp. 690 4576–4582. doi: 10.1021/es203876e. 691 Xu, X. et al. (2005) ‘Catalytic dechlorination kinetics of p-dichlorobenzene over Pd/Fe 692 catalysts’, Chemosphere, 58(8), pp. 1135–1140. doi: 10.1016/j.chemosphere.2004.07.010. 693 Distribution A: Approved for Public Release; Distribution is Unlimited 33 694 (a) 695 (c) 696 697 698 Fig 1 699 Images output by SEM Analysis. 700 701 Distribution A: Approved for Public Release; Distribution is Unlimited (b) 34 DNAN Removal (%) 100% 80% 60% 40% 20% 0% ZVMg 702 703 Fig 2 704 Created on Microsoft Excel. Mg/Cu Mg/Ni Mg/Zn 705 706 707 Distribution A: Approved for Public Release; Distribution is Unlimited 35 1.6E+06 (a) Intensity (counts) 169 1.2E+06 8.0E+05 4.0E+05 152 154 123 122 0.0E+00 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 m/z 708 Intensity (counts) 5.0E+07 169 (b) 4.0E+07 3.0E+07 2.0E+07 123 1.0E+07 152 122 154 0.0E+00 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 m/z 709 Intensity (counts) 5.0E+08 (c) 169 4.0E+08 3.0E+08 2.0E+08 1.0E+08 122 123 154 152 0.0E+00 60 70 80 90 100 110 120 130 140 150 160 170 180 190 m/z 710 711 712 Fig 3 713 Figure generated by Microsoft Excel. 714 715 Distribution A: Approved for Public Release; Distribution is Unlimited 200 36 5.E+07 Intensity (counts) 139 4.E+07 124 3.E+07 2.E+07 107 108 122 1.E+07 0.E+00 60 70 80 90 100 110 120 130 140 m/z 716 717 Fig 4 718 Figure generated by Microsoft Excel. 719 720 Distribution A: Approved for Public Release; Distribution is Unlimited 150 37 721 722 723 Fig 5 724 Figure generated by ChemCad Free Version. 725 726 Distribution A: Approved for Public Release; Distribution is Unlimited 38 727 728 Fig 6 729 Generated on Microsoft Powerpoint 730 731 Distribution A: Approved for Public Release; Distribution is Unlimited 39 732 10 8 300 - 2-ANAN Mass of Compound (mg) 12 DNAN 4-ANAN 13.2 Absorbance (mAU) 14 150 - 10 20 6 30 40 Time (min) 4 DNAN 1.1 2 2-ANAN 4-ANAN 1.2E-01 7.2E-02 DAAN 0* 0 Control DNAN Dissolved Phase 733 734 735 Fig 7 736 Figure generated by Microsoft Excel and Word. 737 738 Distribution A: Approved for Public Release; Distribution is Unlimited 40 DNAN Removal (%) 100% 80% 60% 40% 20% 0% ZVMg Mg/Cu Mg/Ni Mg/Zn 739 740 Fig 8 741 Created on Microsoft Excel. 742 743 744 Distribution A: Approved for Public Release; Distribution is Unlimited 41 745 SEM: Mg/Cu (a) 746 O 747 Fig 9 748 Images generated by EDS mapping. (c) Mg (b) Cu (d) 749 750 Distribution A: Approved for Public Release; Distribution is Unlimited 42 751 752 SEM: Mg/Cu Mg (a) O (c) 753 Fig 10 754 Images generated by EDS mapping. Cu (b) (d) 755 756 Distribution A: Approved for Public Release; Distribution is Unlimited Intensity (Counts) 43 757 2500 (a) 2000 1500 1000 500 0 Intensity (Counts) 2500 (b) 2000 1500 1000 500 0 Intensity (Counts) 758 2500 (c) 2000 1500 1000 500 0 35 35.5 36 36.5 37 37.5 38 38.5 39 39.5 2θ (°) 759 760 Fig 11 761 Figure generated by Microsoft Excel. 762 Distribution A: Approved for Public Release; Distribution is Unlimited 40 Supplemental Information Characterization of Mg-based Bimetal Treatment of Insensitive Munition 2,4dinitroanisole Emese Hadnagy1,*, Andrew Mai2, Benjamin Smolinski3, Washington Braida2, Agamemnon Koutsospyros1 Submitted to: Environmental Science and Pollution Research * Corresponding Author. Email address: EHadnagy@newhaven.edu Department of Civil and Environmental Engineering, University of New Haven 2 Department of Civil, Environmental, and Ocean Engineering, Stevens Institute of Technology 3 RDECOM-ARDEC 1 Distribution A: Approved for Public Release; Distribution is Unlimited Table S.1 Elemental Compositions (% Mass) of Bimetal Surfaces Bimetal Mg/Cu Mg/Ni Mg/Zn Unused Used Unused Used Unused Used O Mg 53.1 65.6 44.7 54.4 41.9 53.8 34.7 34.4 14.7 18.8 23.4 31.5 Cu, Ni or Zn 12.2 0.6 40.6 26.8 34.7 14.7 Distribution A: Approved for Public Release; Distribution is Unlimited Fig. S.1 Contrasted SEM image of sample surface of unused Mg/Cu to more easily observe Cu nanoparticles Distribution A: Approved for Public Release; Distribution is Unlimited 4.E+07 139 (a) Intensity (counts) 3.E+07 3.E+07 2.E+07 2.E+07 1.E+07 5.E+06 108 107 122 124 0.E+00 60 70 80 90 100 110 120 130 140 m/z 7.E+07 139 (b) Intensity (counts) 6.E+07 5.E+07 4.E+07 3.E+07 2.E+07 1.E+07 107 108 124 122 0.E+00 60 70 80 90 100 110 120 130 140 m/z Fig. S.2 Daughter spectrum of m/z 139 from ESI-MS/MS in positive mode from (a) after DNAN treatment (solvent matrix, 0.5% S/L, 10:1 Mg to Cu ratio and 2.5 hr treatment) and (b) pure DAAN solution reference Distribution A: Approved for Public Release; Distribution is Unlimited Relative Intensity (a) (b) Time (min) Fig. S. 3 Mass chromatograms of selected ion m/z 139 obtained from HPLC-ESI-MS of (a) pure DAAN, and (b) treated 4-ANAN sample (aqueous solution, 0.5% S/L, 10:1 Mg to Cu ratio and 1 hr treatment) where the elution of m/z 139 was identical. The slight difference in elution times (<1min) was due to peak shifts on HPLC Distribution A: Approved for Public Release; Distribution is Unlimited 250 TOC TOC from DNAN COD COD from DNAN TOC, COD (mg/L) 200 150 100 50 0 Initial ZVMg Mg/Cu Mg/Ni Mg/Zn Fig. S.4 TOC, COD (mg L-1) and DNAN contribution to TOC and COD in treated wastewater (0.5% S/L, 10:1 Mg to catalytic metal ratio, and 2.5 h treatment time, COD not measured for ZVMg) Distribution A: Approved for Public Release; Distribution is Unlimited DNAN Absorbance (mAU) DNP Control Mg/Cu Mg/Ni Mg/Zn Time (min) Fig. S.5 Visualization of degradation of DNP (2.1 min) and DNAN (4.2 min) in the wastewater control (top chromatogram) versus wastewater treated with Mg/Cu, Mg/Ni and Mg/Zn using overlaid chromatograms after 150 minutes of treatment (wastewater matrix, 0.5% S/L, 10:1 Mg to secondary metal ratio) Distribution A: Approved for Public Release; Distribution is Unlimited SEM: Mg/Ni O Mg ( a ) ( c )  Ni ( b ) ( d )  )LJ6('6PDSSLQJRI D VDPSOHUHJLRQRIDQXQXVHGSDUWLFOHRI0J1LSLFWXUHGE\6(0 HOXFLGDWLQJGLVWULEXWLRQRI E SULPDU\PHWDO0JWR F R[\JHQDQG G FDWDO\WLFPHWDO1L  Distribution A: Approved for Public Release; Distribution is Unlimited SEM: Mg/Ni O Mg (a) (c) Ni (b) (d) Fig. S.7 EDS mapping of (a) sample region of a used particle of Mg/Ni pictured by SEM elucidating distribution of (b) primary metal Mg to (c) oxygen, and (d) catalytic metal Ni Distribution A: Approved for Public Release; Distribution is Unlimited SEM: Mg/Zn (a) Mg (b) O (c) Zn (d) Fig. S.8 EDS mapping of (a) sample region of an unused particle of Mg/Zn pictured by SEM elucidating distribution of (b) primary metal Mg to (c) oxygen, and (d) catalytic metal Zn Distribution A: Approved for Public Release; Distribution is Unlimited SEM: Mg/Zn O (a) (c) Mg (b) Zn (d) Fig. S.9 EDS mapping of (a) sample region of a used particle of Mg/Zn pictured by SEM elucidating distribution of (b) primary metal Mg to (c) oxygen, and (d) catalytic metal Zn Distribution A: Approved for Public Release; Distribution is Unlimited 9000 Intensity (Counts) (a) 6000 Mg0 Mg(OH)2 3000 Zn0 0 35 36 37 38 39 40 41 42 43 44 45 2θ (°) 9000 Intensity (Counts) (b) 6000 3000 0 35 36 37 38 39 40 41 42 43 44 45 2θ (°) Intensity (Counts) 9000 (c) 6000 3000 0 35 36 37 38 39 40 41 42 43 44 45 2θ (°) Fig. S.10 XRD patterns of Mg/Zn (a) before treatment, (b) after treatment in wastewater, (c) and after treatment in the pure aqueous phase Distribution A: Approved for Public Release; Distribution is Unlimited Intensity (Counts) 9000 (a) 6000 Mg0 Mg(OH)2 3000 0 35 35.5 36 36.5 37 37.5 38 38.5 39 39.5 40 2θ (°) Intensity (Counts) 9000 (b) 6000 3000 0 35 35.5 36 36.5 37 37.5 38 38.5 39 39.5 40 2θ (°) Intensity (Counts) 9000 (c) 6000 3000 0 35 35.5 36 36.5 37 37.5 38 38.5 39 39.5 40 2θ (°) Fig. S.11 XRD patterns of Mg/Ni (a) before treatment, (b) after treatment in wastewater, and (c) after treatment in the pure aqueous phase Distribution A: Approved for Public Release; Distribution is Unlimited