The explosives 2,4,6-trinitrotoluene (TNT) and hexahydro-1,3,5-trinitro-1,3,5-triazine (RDX) are common munitions constituents. Both compounds and their derivatives are Environmental Protection Agency (EPA) priority pollutants and are persistent in the environment. In addition to legacy contamination at numerous Department of Defense (DoD) facilities, continued use of RDX and TNT in live fire ranges indicates the likelihood that environmental exposure is ongoing. Within the contiguous 48 United States, there are approximately 41 active DoD installations located within the coastal zone. Exposure of marine/estuarine ecosystems at some sites is well documented, while other installations have a high potential for exposure but limited data on RDX or TNT concentrations in marine end members. Coastal habitats are highly productive, nitrogen-limited, and economically valuable ecosystems. Their response to munitions compounds and their effect on munitions cycling, persistence, bioaccumulation, and mineralization are largely unknown.

The objective of this project was to quantify ecosystem processing of the legacy munitions compounds RDX (hexahydro-1,3,5-trinitro-1,3,5-triazine) and TNT (2,4,6-trinitrotoluene) in coastal marine environments. Sorption to sediments, mineralization, and uptake/elimination by macrobiota was examined. 

The objective of this project is to quantify the pathways and rates of RDX and TNT processing in three typical coastal ecotypes: subtidal vegetated, subtidal unvegetated, and intertidal salt marsh. The following technical questions will be addressed: (1) What are the uptake rates of these compounds at the organismal to ecosystem scales, and which ecosystem components are important regulators of processing? (2) What ecosystem components act as zones of storage for munitions compounds versus those that promote metabolism? (3) Do ecosystem characteristics (e.g., mineralization, autotrophy, redox profile, trophic structure) relate to processing or accumulation of munitions compounds? and (4) What is the ecosystem-scale residence time or clearance rate of these compounds?

Technical Approach

Stable isotope (15N) -labeled and unlabeled RDX and TNT were used in benchtop, aquaria, and large mesocosm (ecocosm) tracer experiments (Figure 1).  The stable isotope tracer was tracked into biota, sediments, major organic derivatives, and into inorganic nitrogen species that constituted complete mineralization of the munitions constituent.  The magnitudes and rates of isotope tracer appearance in each of these pools permitted quantification of fate of RDX and TNT in typical temperate coastal marine ecotypes.

Figure 1. Experimental approach.

Isotope labeled TNT and RDX were used in separate experiments. Ecocosm experiments were several weeks in duration to simulate a steady state input of RDX or TNT consistent with chronic legacy contamination.


Uptake and Elimination by Macro-organism: Exposure of common coastal marine flora and fauna to RDX and TNT yielded bioconcentration factors (BCFs) that were similar to freshwater organisms of similar trophic level and lifestyle. Uptake into tissues was rapid. Primary nitroso derivatives for RDX were a small fraction of the tissue RDX concentrations, but amino-dinitrotoluenes (dominated by 4-ADNT) were 20% of the TNT concentration. Highest BCF was measured in lipid rich crab eggs, but generally RDX, TNT and their primary derivatives yielded low BCFs (<50 ml g-1) that were reasonably well-predicted by the octanol-water partition coefficient (Kow) of the compounds.

First-order modeling of RDX and 15N tracer concentrations in tissues revealed high rates of uptake offset by rapid elimination and redistribution of tracer into bulk biomass.  Tissue 15N tracer exceeded intact 15N-RDX by 10-fold.  Modeled uptake rates were similar to published values, but elimination rates were several orders of magnitude smaller (0.05 to 0.7 day-1). Low elimination rates were offset by high rates of retention of 15N in some unknown form. Four different biotransformation pathways were proposed to explain the 15N retention.  Some of these pathways indicate photosynthetic utilization of RDX derived N, while others suggest the formation of undetected metabolites and/or adduct formation between metabolites and cellular constituents.

The uptake, retention, and transformation of 15N-RDX measured in 13 marine species were independent of coastal ecotype.  Despite different levels of RDX loss and mineralization between sand, vegetated silt, and intertidal marsh mesocosms, the variance in the tissue RDX and 15N concentrations were similar. Limited correlation between aqueous and tissue RDX concentrations combined with an order-of-magnitude more isotope tracer in biomass relative to tissue RDX suggested that post uptake transformations were as important as aqueous RDX concentrations in setting tissue RDX levels (Figure 2). Extrapolating mesocosm results to ecosystem scales revealed that macrobiotic RDX retention and processing scaled linearly with expected species biomass, with “hot spots” of high retention and/or transformation in marsh macrophyte roots.

Figure 2. Distribution of isotope tracer derived from 15N-labeled RDX between aqueous and biomass fractions (A); and the total isotope tracer in tissues that was attributable to RDX and its nitroso-derivatives (B).  Approximately 10% of the RDX derived tracer was found in biota, only 5-10% of that existed as RDX + nitroso-derivatives.

Removal and Mineralization - The presence of sediments had a major influence on the removal kinetics of all compounds detected.  RDX degraded only in the presence of sediment, and TNT degraded significantly faster in the presence of sediment.  RDX and TNT removal from the overlying water increased with decreasing grain-size.  Photodegradation at light levels typical of temperate coastal waters was not a factor in RDX or TNT removal from the water column.

Sediment effects on aqueous RDX and TNT concentrations resulted partly from high sorption to marine sediment. Abiotic sorption tests in anaerobic sediment slurries showed that marine sediments had significantly higher uptake rates for both compounds relative to freshwater sediment of similar grain size.  Equilibrium partition constants were on the same order of magnitude for marine and freshwater silt (1.1 – 2.0 Lkg-1sed), but lower for marine sand. Anaerobic slurry conditions promoted mineralization of both TNT and RDX as evidenced by conversion of 15N labeled parent compound to various dissolved inorganic nitrogen (DIN) constituents.  Two percent of the TNT added was mineralized through denitration, and deamination which resulted in primarily NH4+.  Multivariate analyses suggested that iron and sulfate reduction facilitated mineralization.  Six times more RDX was mineralized than TNT.

More realistic aquaria experiments using 15N-RDX with intact sediment with aerobic / anaerobic gradients showed a linkage between RDX mineralization and denitrification. Modeling the evolution of 15N2O and 15N2 showed mineralization of 11% of the added RDX after 22 days, and 29% of the total removed RDX-N was identified as 15N2.  Denitrifiers were not responsible for the RDX degradation but rather that they used RDX mineralization products as a substrate.  The result documented that RDX mineralization supplied N generally to the marine N cycle.    Aerobic/anaerobic aquaria-scale 15N-TNT experiments indicated that the dominant TNT mineralization product remained as NH4+ even under mixed redox conditions. The 15N tracer mass balance revealed that the majority of TNT was not mineralized, but instead transformed into an unidentified derivative.  We used principal components analysis to constrain the steps leading to the derivative production and to infer that the major transformation pathway was the deamination of TNT, promoted by sorption to SPM and oxic surface sediments.

Larger mesocosms containing water, sediment, biota, and 15N-RDX were used to evaluate RDX mineralization among three different marine ecotypes:  1) subtidal low carbon sand, 2) subtidal higher organic carbon vegetated silt, 3) intertidal salt marsh.  Uptake and retention of parent compound and primary organic derivatives by macrobiota were minor fates for both RDX and TNT.  RDX processing was largest (50%) in the high carbon and redox variable intertidal marsh, and smallest in the low carbon sand mesocosm (25% of added RDX).  The nitroso derivatives MNX, DNX, and TNX, comprised only 2-3% of the RDX loss. Mineralization of RDX, resulted in primarily N2O and N2, and was the fate of 47-70% of the RDX lost from the water column.  Similar mesocosm experiments with 15N TNT showed ~78% loss of TNT from the overlying water that was independent of ecotype. TNT-15N partitioning to sediment was greater than RDX, but with little mineralization.  Poor isotope mass balances in the TNT mesocosms, similar to that found in the TNT aquaria experiments, suggested that TNT processing largely resulted in the formation of an unidentified organic derivative.  Analysis of the dissolved organic nitrogen pool did not further constrain the nature of the derivative.  

Figure 3. The fate of isotope tracer distributed through biotic and abiotic ecosystem pools during three-week steady state 15N-RDX (upper panel) and 15N-TNT (lower panel) ecocoms experiments. Mineralization, affected by sediment organic content, was the dominant fate of RDX. Unknown organic derivative(s) was the likely dominant fate of TNT. Dashed pie wedges in the lower panel indicate confirmation of the unknown fraction in the aqueous phase.


Results provide end-user benefits in two categories. First, the project yielded rate constants for direct parameterization of models that support risk assessments. The water column half-lives and sorption isotherms measured across range of sediment types can be incorporated into transport models that predict RDX and TNT mobility and migration off-site. The BCFs and associated uptake and elimination constants measured for over a dozen marine biota species, including those consumed by humans, provide calibration factors for human exposure models.

Second, the project vetted a technology that can be used to quantify in situ natural attenuation of RDX and TNT at aquatic sites under field conditions. The detection of mineralization species unique to the isotope labeled munitions constituent, coupled with high recovery of total labeled RDX products in large mesocosms identified a sensitive tool for quantifying natural attenuation (NA) rates in situ.  The in situ mineralization rates can be evaluated relative to transport times of contaminants off-site and/or to sensitive human health/ecological targets to determine suitability of NA as a remediation option. Implementation of this technology in contaminated ground or surface waters would require tagging existing contamination with small amounts of isotope labeled contaminant, accounting for dilution during transport, and analyses of labeled mineralization products over time.

Transfer of the RDX results to end users with respect to both categories can be done with high confidence. Data gaps remain for transfer of the TNT results whose primary fate was an unmeasured aqueous organic derivative. Identification of this aqueous derivative is a critical missing piece of information needed to provide end-users with guidance on TNT mineralization or human health/ecological risk.


Ariyarathna, T., M. Ballentine, P. Vlahos, R.W. Smith, C. Cooper, J. K. Bohlke, S. Fallis, T.J. Groshens, and C. Tobias. 2019. Tracing the Cycling and Fate of the Explosive, Hexahydro-1,3,5-trinitro-1,3,5-trazine in a Simulated Sandy Coastal Marine Habitat with a Stable Isotopic Tracer, 15N-[RDX]. Science and the Total Environment, 647:369-378.

Ariyarathna, T., P. Vlahos, R.W. Smith, J.K. Bohlke, S. Fallis, T.J. Groshens, and C. Tobias. 2017. Biodegradation and Mineralization of Isotopically Labeled TNT and RDX in Anaerobic Marine Sediments. Environmental Toxicology and Chemistry, 36(5):1170-1180.

Ariyarathna, T., P. Vlahos, C. Tobias, and R. Smith. 2015. Sorption Kinetics of Explosives in Anaerobic Sediments. Environmental Toxicology and Chemistry, 35(1): 47–55, 2016

Ballentine, M.L., T. Ariyarathna, R.W. Smith, C. Cooper, P. Vlahos, S. Fallis, T.J. Groshens, and C. Tobias. 2016. Uptake and Fate of Hexahydro-1,3,5-trinitro-1,3,5-triazine (RDX) in Coastal Marine Biota Determined using a Stable Isotopic Tracer, 15N [RDX]. Chemosphere, 153:28-38.

Ballentine, M., C. Tobias, P. Vlahos, R. Smith, and C. Cooper. 2014. Bioconcentration of TNT and RDX in Coastal Marine Biota. Archives of Environmental Contamination and Toxicology, 68(4):718-728.

Smith, R.W., T. Ariyarathna, M. Ballentine, C. Cooper, P. Vlahos, J.K Böhlke, S. Fallis, T. J. Groshens, and C. Tobias. 2015. Tracing the Cycling and Fate of the Explosive 2,4,6-trinitrotoluene in a Simulated Coastal Marine Habitat with a Stable Isotopic Tracer, 15N – [TNT]. Environmental Science and Technology, 49(20):12223-12231.

Smith, R., C. Tobias, P. Vlahos, C. Cooper, M. Ballentine, and T. Ariyarathna. 2015. Mineralization of RDX-Derived Nitrogen to N2 via Denitrification in Coastal Marine Sediments. Environmental Science and Technology, 49(4):2180-2187.

Smith, R. W., P. Vlahos, C. Tobias, M. Ballentine, T. Ariyarathna, and C. Cooper. 2013. Removal Rates of Dissolved Munitions Compounds in Seawater. Chemosphere, 92(8):898-104.