Karen F. Myers

Comparison of solid phase extraction with salting-out solvent extraction for preconcentration of nitroaromatic and nitramine explosives from water

Abstract Residues of high explosives are a significant pollution problem at U.S. military facilities. Because TNT, RDX and HMX, as well as several manufacturing impurities and environmental transformation products, are mobile in the soil and have caused groundwater pollution, there is an increasing demand for low-concentration analysis of these compounds in water from installation boundary wells. Because RDX and HMX are polar, conventional liquid-liquid extraction with nonpolar solvents yields poor recovery. Two techniques have been reported that appear to offer improved recovery and adequate preconcentration: solid phase extraction (SPE) and salting-out solvent extraction (SOE). This paper compares resin based cartridge-SPE, membrane-SPE, and SOE using fortified reagent grade water samples and a set of 58 groundwater samples from an explosives-contaminated military facility. The three methods were comparable with respect to low-concentration detection capability, which ranged from 0.05 to 0.30 μg/l. Percent recoveries generally exceeded 80%, except for HMX and RDX by membrane-SPE. Interferences were found in extracts from half of the groundwater samples preconcentrated using the two SPE procedures, but were not found in any of the extracts from the SOE. These interferences were traced to matrix interaction of the polymeric resins with low-pH groundwater containing high levels of dissolved solids.

Sampling for Explosives-Residues at Fort Greely, Alaska

Fort Greely, Alaska has an extensive complex of weapon training and testing areas located on lands withdrawn from the public domain under the Military Lands Withdrawal Act (PL106-65). The Army has pledged to implement a program to identify possible munitions contamination. Because of the large size (344,165,000 m2) of the high hazard impact areas, characterization of these constituents will be difficult. We used an authoritative sampling design to find locations most likely to contain explosives-residues on three impact areas. We focused our sampling on surface soils and collected multi-increment and discrete samples at locations of known firing events and from areas on the range that had craters, pieces of munitions, targets, or a designation as a firing point. In the two impact areas used primarily by the Army, RDX was the most frequently detected explosive. In the impact area that was also used by the Air Force, TNT was the most frequently detected explosive. Where detected, the explosives concentrations generally were low (<0.05 mg/kg) except in soils near low-order detonations, where the explosive-filler was in contact with the soil surface. These low-order detonations potentially can serve as localized sources for groundwater contamination if positioned in recharge areas.

in situ thermal desorption of VOCs in vadose zone soils

A volatile organic compound (VOC) sampler developed for the site characterization and analysis penetrometer system (SCAPS) program to provide semi-quantitative field screening data was evaluated in the laboratory and in the field. The device is based on thermal desorption principles and is capable of sampling in the vadose and capillary zones. Approximately 5 g of soil is desorbed in situ, and the volatilized compounds transferred to the surface, where they are trapped and analyzed by ion trap mass spectroscopy (ITMS). Ex situ validation samples from a Mostap™ core taken 9 in. away from the in situ sample are of two types. One sample is placed in the sampler and desorbed and analyzed in the same manner as the in situ sample. The second validation sample is preserved in methanol and analyzed off site by GC/MS. Comparisons of in situ data to ex situ validation data show that subsurface heterogeneity strongly affects correlation, especially in zones of bedded sand and silt. When the two ex situ samples from the Mostap™ core are compared, the subsurface heterogeneity factor is removed and the co-efficient of determination (r2) increases to 85%.