Chapter 9
Reactive Barrier Treatment Technology
ORGANIC COMPOUNDS NITROAROMATIC AND NITROAMINE COMPOUNDS (NACS )
Nitroaromatic and nitroamine compounds (NACs) are major groundwater and soil contaminants at military ammunition fabrication and demilitarization facilities. NACs in groundwater can disperse widely and undergo microbial transformation to additional hazardous products. A promising remedy is Fenton’s reagent in-situ chemical oxidation (FRISCO). Fenton’s reagent produces a hydroxyl free radical (OH•) by reaction of H2O2 and Fe(II). Oxidative systems (in addition to Fenton’s reagent) that generate OH•(H2O2/ozone, H2O2/UV, or ozone/UV) rapidly destroy NACs via side-chain oxidation and ring cleavage .
Fenton’s reagent oxidation of 2,4-DNT generates intermediate products including 1,3-DNB, benzaldehyde, benzoic acid,and nitroaniline . Fenton’s reagent oxidation of TNT and RDX in soil results in formation of NO 3-2and oxalic acid (from TNT), and ammonia, NO 3-2 formic acid, and methylene dinitramine (from RDX)
Formation of these compounds indicates that side-chain oxidation and ring cleavage also occur with Fenton’s reagent. Fenton’s reagent also oxidizes 2-A-4,6-DNT and 4-A-2,6-DNT (natural TNT degradation products) with rate constants approaching diffusion-controlled limits (>1.6 x 10 9 M -1 s -1 )
The source of NACs in groundwater at PCD was a drainage ditch and leach bed at a former TNT Washout Facility and Discharge System. Contaminated soil from the leach bed, drainage ditch, and associated areas was removed in 1997-1998. A groundwater plume follows paleochannels and around a bedrock high The shallow aquifer is a heterogeneous deposit of sand with sporadic clay, silt, and gravel (eolian, alluvial and colluvial origin), which unconformably overlies Pierre Shale bedrock. Average hydraulic conductivity is 2 x 10 -2 cm/sec.
Two existing monitoring wells (TNTMW04 and CSPDPW171) were located in Area 1.
The depth to groundwater and bedrock were 3.4 m and 4.0 m, respectively. Area 1 had the highest NAC concentration. Baseline samples yielded NACs (and their maximumconcentrations) as follows: TNB (1,160 ug/L), TNT (678 ug/L), 2-A-4,6-DNT (212ug/L), 4-A-2,6-DNT (164 ug/L), TNX (56.0 ug/L), 2,4-DNT (33.6 ug/L), 1,3-DNB (16.1ug/L), MNX (5.94 ug/L), RDX (4.90 ug/L), and HMX (1.51 ug/L). Soil sampling for preliminary bench tests detected TNT (1.73 mg/kg) and TNB (3.14 mg/kg) in soil. NO 3-2 in groundwater ranged from 1.9 to 15 mg/L.
Area 1. The field pilot test was conducted from September 6-7, 2000.
Approximately 1,500 L of 25% H2O2 and 4,200 L of catalyst solution were injected to IW-1 and IW-2, and 2,700 L of 25% H2O2 and 5,800 L of catalyst were injected to IW-3.Groundwater pH, Fe, and H2O2 concentrations in monitoring wells adjacent to the injectors were analyzed to determine radius of influence (. The difference between potential and observed NO 3 -2 production represents destruction of soil-sorbed NACs.
The size of the treatment area was estimated at approximately 204 m 2 from the observed ROI and distribution of injectors. Assuming an aquifer thickness of 0.9 m and 30% porosity, the water volume within the treatment area was approximately 55,000 L.
The total injectate volume (H2O2 plus catalyst) was approximately 20,000 L, hence themaximum total volume over which the NO 3 -2 increase was integrated was 75,000 L. This is a minimum estimate that assumes that advection and dispersion were negligible.
The mean baseline concentration of total NACs was 888 ug/L, with weighted average nitrogen content (incorporating the contribution of each compound to the total NAC concentration) of 19.8%. The corresponding mass of nitrogen that could be produced by 100% mineralization of the dissolved explosives, therefore, is approximately 9.7 g (as N) or 43 g (as NO3 - ). The mean total NAC concentration in the first post-test sampling event was 853 ug/L with weighted average nitrogen content of 19.2%.Incorporating the total volume of groundwater plus injected reagents (75,000 L), the corresponding mass within the treatment area is, therefore, 12 g (as N) or 54 g (as NO3 - ). Thus there is a net negative NO 3 -2balance within the treatment area (i.e., post-test NO 3-2 mass is greater than pre-test mass), indicating release of NO 3-2from sources (e.g., soil-sorbed explosives) other than the dissolved NACs.
The average NO 3-2concentration was 5.8 mg/L prior to the pilot test and 34.3 mg/L during the first post-test sample round, for an average increase of 28.5 mg/L. A28.5 mg/L increase integrated through 55,000 L of groundwater yields a minimum production of 1.6 kg of NO 3-2. The maximum production for 75,000 L (groundwater plus injected reagents) corresponds to 2.1 kg of NO 3-2. Thus the increase in NO 3-2 concentration is far greater than expected from 100% mineralization of the dissolved NACs, further indicating that soil-sorbed NACs were oxidized during the pilot test.
TheNAC mineralized mass was estimated from the observed NO 3-2increase of 1.6-2.1 kg (equivalent to 0.4-0.5 kg as N). The weighted average nitrogen content of the dissolved NACs was 19.8%, thus approximately 2.0-2.5 kg of NACs were mineralized to generate the observed NO 32increase. The conclusion from the NO 3-2mass balance is that the observed NO 32increase is too large to be explained by mineralization of NACs in groundwater. Treatment must have also oxidized soil-sorbed explosives. 1,3-DNB is formed as an intermediate Fenton oxidation product of 2,4-DNT (and possibly other NACs) by side-chain oxidation ). The concentration of 1,3-DNB in Area 1 groundwater increased following treatment , which could be interpreted as either an intermediate oxidation product or soil desorption. 1,3-DNB is not a known contaminant disposed at PCD, thus formation as an intermediate oxidation product is most likely. Theincrease in 1,3-DNB concentrations was, therefore, also used to estimate corresponding NAC destruction, assuming all of the1,3-DNB was a 2,4-DNT oxidation product. Theaverage concentration of 1,3-DNB in Area 1 groundwater during baseline sampling was4.1 ug/L. The average 1,3-DNB concentration in the first post-injection sampling round was 43 ug/L, for an average increase of 38.9 ug/L. Integrating a 38.9 ug/L increase through a 55,000-75,000 L treatment area volume yields an estimated 1,3-DNB mass increase of 2.1-2.9 kg. The formula weight of 1,3-DNB is 168.1 g per mole, thus the increased mass corresponds to production of 12.5-17.3 moles of 1,3-DNB. Assuming that 1,3-DNB is derived only from 2,4-DNT, then 12.5-17.3 moles of 2,4-DNT were oxidized. The formula weight of 2,4-DNT is 182.2 g per mole, thus 2.3-3.2 kg of 2,4-DNT were oxidized.
The estimate of NAC contaminant mass destruction from 1,3-DNB production has greater uncertainty than estimates derived from NO 32production, because more than one compound may produce 1,3-DNB as an oxidation product, the 1,3-DNB is itself subject to oxidation, and the possible pre-treatment presence in the soil-sorbed fraction is unknown. However, the large (10-fold) increase in 1,3-DNB during treatment, coupled with the known susceptibility of 1,3-DNB to Fenton reagent oxidation and production as an oxidation product, support the interpretation of 1,3-DNB formation as an intermediate product.
Significance for Explosives Remediation. The Area 1 results demonstrate that FRISCO is an effective remedy for dissolved RDX, HMX, 2-A-4,6-DNT, 4-A-2,6-DNT, 2,4-DNT,MNX, and TNX. Due to enhanced soil desorption, dissolved concentrations of strongly-sorbedNACs such as TNT and TNB may not decrease following treatment, despiteevidence (such as NO 32or 1,3-DNB production) of significant mass destruction.
Effectiveness of FRISCO cannot be evaluated from only dissolved NAC concentrations if a soil-sorbed phase exists. Such application requires evaluation of soil NAC concentrations and oxidation products to determine effectiveness. Soil-sorbed NACs will supply groundwater plumes as the NACs and their degradation products are slowly desorbed, and effectiveness of any groundwater remedy will be reduced if the soil-sorbedsource is addressed. FRISCO can be applied as a source-area remedy to address soil-sorbed NACs. Coupling FRISCO with other technologies may also improve performance.
Increased desorption of TNB and TNT following FRISCO (in addition to the overall contaminant mass destruction) may enhance performance of existing pump-and-treat systems or other remediation technologies, thus reducing time and cost for site cleanup.
REFERENCES
J. Daniel Bryant and James T. Wilson,Paper 2G-01, in: A.R. Gavaskar and A.S.C. Chen (Eds.), Remediation of Chlorinated and Recalcitrant Compounds—2002.
Proceedings of the Third International Conference on Remediation of Chlorinated and Recalcitrant Compounds
(Monterey,CA; May 2002). ISBN 1-57477-132-9, published by Battelle Press, Columbus, OH, .