Reductive Transformation of Insensitive Munitions Compounds by Reactive Iron-Based Minerals

Abstract

Insensitive munitions compounds (IMCs) are emerging contaminants widely applied by the U.S. Armed Forces to prevent unintended detonations from legacy munitions due to their insensitivity to mechanical and thermal shock. The toxicity of IMCs towards aquatic organisms and mammals has recently been reported, raising concerns about potential adverse impacts of IMC emissions from contaminated sites at military training sites or storage areas, and from IMC manufacturing wastewaters. Therefore, the objective of this research is to develop effective remediation technologies for IMCs using reactive minerals. The IMCs 3-nitro-1,2,4-traizol-5-one (NTO), 2,4-dinitroanisole (DNAN), and nitroguanidine (NQ) contain nitro groups which are susceptible to reductive transformation. Therefore, the effectiveness of iron-based reactive materials, including zero-valent iron (ZVI), iron sulfide (FeS) minerals (synthesized mackinawite (FeS), and commercial FeS identified as pyrrhotite and troilite) and sulfidated ZVI (SZVI), as strong reductants to remove IMCs from contaminated water was investigated in batch and column experiments. The applied iron-based materials effectively reduced NTO, DNAN, and NQ and their transformation pathways were elucidated. NTO was reduced to its daughter product 3-amino-1,2,4-triazl-5-one (ATO). The selective reduction of the para- and ortho-nitro groups of DNAN led to the aromatic amine intermediates, 4-methoxy-3-nitroaniline (iMENA) and 2-methoxy-5-nitroaniline (MENA), which were further reduced to 2,4-diaminoanisole (DAAN), respectively. On the other hand, the initial product of the reactive transformation of NQ by ZVI and FeS was nitrosoguanidine (NsoQ). The nitroso intermediate was further reduced by ZVI to aminoguanidine (AQ), guanidine and cyanamide, whereas reduction of NsoQ by FeS resulted in the formation of guanidine, NH4+, and small amounts of urea. Sulfidation of nano-sized ZVI effectively reduced the anoxic corrosion of water, a reaction that is expected to compete with the reductive degradation of NTO. However, the suppressed anoxic corrosion did not lead to improved NTO reduction by SZVI than ZVI in pH ranging from 3.0 to 6.0. Mackinawite reacted with NTO and DNAN was oxidized to goethite (α-FeO(OH)) and elemental sulfur (S0). On the other hand, reduction of NTO by ZVI led to the concomitant formation of oxidized iron minerals, including magnetite (Fe3O4), lepidocrocite (γ-FeO(OH)), and goethite (α-FeO(OH)). Precipitation of oxidized iron minerals on the ZVI surface can hamper electron transfer, decreasing the performance of ZVI as confirmed in experiments with two ZVI materials with differing in the thickness of their surface iron oxide layer (~ 880 vs. 300 nm). This study demonstrated that NTO removal by ZVI could be greatly enhanced by removing the passivating layer using different pretreatments (in order of increasing effectiveness): washing with 1 M HCl > 60 mM NaHCO3 > 1 M acetic acid. Acid treatment using 1 M HCl was also shown to be an effective approach to fully reactive and extend the service life of ZVI in a continuous-flow packed-bed column treating NTO-contaminated water. Similarly, acidic conditions (pH 3.0) were beneficial to achieve faster NTO and NQ reduction by ZVI compared to circumneutral to moderately alkaline conditions (NTO transformation: pH 6.0 − 8.0; NQ transformation: pH 5.5 – 7.0) by preventing the precipitation of iron oxides. As an example, when the pH was kept constant at 3.0 during the reaction, the rate of NTO degradation by ZVI was 316-fold faster than at pH 6.0. In contrast, FeS (i.e., mackinawite and commercial FeS) was insensitive to changes in pH values in the range of pH tested for NQ (5.5 – 10.0) and DNAN (6.5 – 7.6). Additionally, sulfidated ZVI (SZVI) with a molar S/Fe ratio of 0.03 in alkaline conditions (pH 8.0) had 3-fold higher NTO reduction rate compared to acidic conditions (pH 3.0). Therefore, ZVI application in NTO and NQ remediation technologies is recommended when the influent is acidic, whereas FeS and SZVI are more applicable to treat water with circumneutral to alkaline pH values. Sulfidation increased the selectivity of ZVI towards the contaminants by suppressing the production of hydrogen gas resulting from ZVI corrosion by anoxic water. However, the enhanced selectivity did not improve the NTO reduction, presenting up to11-fold lower pseudo first-rate order constants for NTO reduction compared to those of ZVI at initial pH values ranging from 3.0 to 6.0. Only under mildly alkaline conditions (pH 8.0), the rate of NTO reduction by SZVI (S/Fe= 0.03) was higher (3-fold) compared to that of ZVI since SZVI (S/Fe= 0.03) was more negatively charged than ZVI above pH 7.6. This study demonstrated the feasibility of treating NTO and NQ in flow-through columns packed with ZVI or FeS. The service life of the ZVI-packed bed was strongly dependent on the influent pH. When treating an acidic NTO influent (pH 3.0) the service life until the breakthrough point (i.e., < 85% NTO removal) was 11-fold higher (2,930 pore volumes, PVs) compared to the treatment of a pH 6.0 influent (250 PVs). In spite of the long period of operation, the hydraulic behavior of the ZVI bed was good during most of the experiment and a moderate decrease in flow rate was only observed at the end of the experiment (7,200-10,912 PVs). ZVI is a more effective packing material for the remediation of NQ compared to FeS. Full removal of NQ was attained in a flow-through column packed with ZVI/quartz sand (1:1, v/v) until the end of the experiment (390 PVs). In contrast, NQ breakthrough was observed in the reactor packed with FeS/corundum (1:1, v/v) after only 100 PVs. The hydraulic conductivity of the two columns did not deteriorate during the experiment, suggesting that supplementation of quartz sand or corundum (Al2O3) in the column bed could diminish some of the hydraulic problems observed in ZVI-packed column treating acidic NTO influent. Taken together these results indicate that ZVI and FeS are promising materials for the reductive remediation of IMC-contaminated water with potential application in packed bed reactor systems and permeable reactive barriers (PRBs). Our findings can facilitate the selection of the most effective Fe-based reductant for implementation in full-scale remediation systems by providing detailed analysis of the impact of aqueous chemistry conditions (pH, presence of co-contaminants) on the reactivity of the different iron-based materials.