For the last 25 years, the U.S. Army has been searching for the most appropriate technology to treat pink water other than granular activated carbon (GAC). The major constituents of pink water are 2,4,6-trinitrotoluene (TNT), hexahydro-1,3,5-trinitro-1,3,5-triazine (RDX), octahydro-1,3,5,7-tetranitro-1,3,5,7-tetrazocine (HMX), 2,4-dinitrotoluene (DNT), trinitrobenzene (TNB), and dinitrobenzene (DNB). Among these toxic compounds, TNT and RDX are the major constituents in pink water and both of them are potential human carcinogens. The National Defense Center for Environmental Excellence prioritized all available pink water treatment technologies and selected five technologies for further development: large aquatic plants, thermophilic biological regeneration of GAC, Fenton’s oxidation, electrolytic oxidation, and anaerobic fluidized bed reactor. However, it is now known that both biological and chemical oxidations of pink water are inefficient due to the oxidized nature of these compounds. As a result, GAC adsorption is the most commonly used method to remove energetic compounds from pink water. The GAC process is expensive and also generates explosive-laden spent carbon, a hazardous waste that needs to be regenerated or disposed of properly. The long-term goal of this project was to develop a more cost-effective technology for the treatment of wastewaters containing energetic compounds.

During the first phase, a study was conducted to demonstrate the feasibility of using zero-valent iron (ZVI) to treat energetic compounds. ZVI was evaluated as a reductant for the degradation of TNT, RDX, HMX, and nitroglycerin (NG). It was hypothesized that reduction of these energetic compounds with ZVI would rapidly convert them to products that are either non-toxic or more readily mineralized in a subsequent oxidative treatment. This was demonstrated successfully in a series of laboratory-scale experiments with batch and flow-through reactors, using ZVI treatment alone or ZVI treatment followed by Fenton oxidation.

The reductive degradation of TNT, RDX, and NG in aqueous solution was carried out under anaerobic conditions in batch reactors with ZVI containing commercial cast iron or high-purity iron powder. All three compounds were removed rapidly from solution using either cast iron or high-purity iron. The fitted pseudo-first-order rate constants for the disappearance of TNT, RDX, and NG were 4.4 (± 0.3) x 10^-2, 6.3 (± 0.1) x 10^-2, and 1.65 (± 0.3) x 10^-2 L•m^-2•h^-1, respectively. Initial removal of TNT and RDX was partly due to adsorption to iron particles, but the adsorption was reversible and all adsorbed TNT and RDX, as well as their reaction intermediates, were further transformed. The extent of adsorption with cast iron was greater than that with high-purity iron, presumably due to the presence of carbon (mostly graphite) in cast iron. In contrast, adsorption of NG was negligible because of its high water solubility.

Using either cast iron or high-purity iron, TNT was quantitatively reduced to the end product 2,4,6-triaminotoluene (TAT) through dinitroamino- and nitrodiamino-toluenes as intermediates. In contrast to TNT, RDX reduction with iron involved ring cleavage and yielded formaldehyde (69.1%) as the dominant carbonaceous product and NH₄⁺ (35.5%) and N₂O (26.9%) as major nitrogenous products. Methylenedinitramine (MDNA) was identified as an intermediate product of RDX and was itself reduced completely to formaldehyde (~100%), NH₄⁺ (19.2%), and N₂O (25.4%). Reduction of NG with iron occurred through sequential reductive denitration reactions via 1,2- and 1,3-dinitroglycerins (DNGs) and 1- and 2-mononitroglycerins (MNGs) to glycerol. Each of the denitration reactions released a nitrite, which was further reduced to NH₄⁺. A reaction pathway and a kinetic model were proposed to describe the reduction of NG and nitrite with cast iron.

The rapid and quantitative reduction of NG to innocuous and biodegradable glycerol and NH₄⁺ suggests that ZVI alone may be a viable method to treat NG-laden wastewaters. For TNT and RDX, an oxidative process following ZVI treatment may be necessary due to potential product toxicity and incomplete mass balance. For this study, Fenton oxidation was selected to demonstrate the effectiveness of ZVI treatment to enhance the rate and extent of TNT and RDX oxidation. An integrated treatment system consisting of a ZVI column and a Fenton reactor was proposed, and a bench-scale unit was constructed for the demonstration. A solution of TNT and RDX was passed through a cast iron-packed column (9.7 min residence time). The column effluent, which contained the reduction products of RDX and TNT, was sent to a completely-stirred tank reactor where Fenton reagent (H₂O₂ and Fe2+) was added. Compared to direct Fenton oxidation of TNT and RDX, ZVI treatment enhanced the TOC removal by 20% and 60% for TNT and RDX, respectively. Complete TOC removal for both compounds was achieved under relatively mild conditions following ZVI treatment, whereas mineralization was not achievable without ZVI treatment even at high H₂O₂ and Fe²⁺ concentrations. The result suggests that a sequential ZVI treatment-Fenton oxidation process may be a feasible technology for pink water treatment.

In the second phase, the ZVI-Fenton process was further evaluated against real pink water from the Iowa Army Ammunition Plant (AAP). Results of the batch and column experiments show that TNT, RDX, and HMX in the pink water were completely degraded and TAT and NH₄⁺ were recovered after ZVI treatment. Using the integrated system, 85% of the TOC in the pink water, corresponding to the TOC of TNT, RDX, and HMX combined, was removed under optimal conditions. Laboratory column experiments performed with “real” wastewater samples from Holston AAP showed that complete removal of RDX can be achieved with a column hydraulic residence time of 6.2 minutes. An additional column test carried out at an elevated temperature of 75°C suggested that the iron contact time and thus ZVI reactor size may be further reduced by pre-heating the wastewater with steam.

Laboratory results indicate that elemental iron rapidly reduces nitroaromatics, nitramines, and nitrate esters to non-toxic or more oxidizable products and thus represents a promising new approach to treat energetic compounds in wastewater. The results strongly justified a pilot-phase study evaluation and demonstration of the proposed ZVI process.

A trailer-mounted ZVI reactor was built and pilot-phase evaluation was conducted at the University of Delaware. The pilot-scale demonstration (50 gallons/hour) was conducted with RDX solution containing four different RDX concentrations (10-40 mg/L) and two different pH values. The results from the pilot-scale iron column study showed that RDX and DNT were readily degraded by ZVI under a wide range of influent conditions.

This pilot study demonstrated that the ZVI-based treatment process is an efficient way to convert RDX in wastewater into products that are easily degradable through chemical or biological oxidation. For example, about 65% of carbon in influent RDX was recovered as formaldehyde, which can be easily oxidized in conventional aerobic wastewater treatment plants (WWTP) or by Fenton oxidation.