Source Characteristics and Permanganate Remediation of Chlorinated Ethenes in Fractured Sedimentary Media

Permanganate has been selected for full-scale design-phase assessment for in situ remediation of a source area in fractured shale at the Watervliet Arsenal near Albany, New York. The contamination, which comprises PCE and TCE and their breakdown products, occurs in this organic and sulfide-rich shale from near ground surface to depths greater than 150 ft bgs. The selection of permanganate was based on results of (a) a major site characterization effort conducted in 1998-2001 to determine VOC distribution and delineation of major hydraulic transmission zones and (b) follow-up site characterization and two pilot-scale KMnO4 injection tests conducted in 2002. These studies formed the basis for the site conceptual model in which the contamination in the source area occurs at all depths in the fracture network. This network is comprised of interconnected fractures with two major hydraulically transmissive zones and many less permeable fractures. Small samples of rock were collected at one to two-foot intervals from continuous core from two vertical holes and analyzed for VOC concentrations. The results indicate that most of the VOC mass now resides as dissolved and sorbed mass in the low permeability rock matrix blocks between numerous fractures. The rock matrix porosity determined from core measurements ranges from 2-4%. The VOC contamination was caused initially by DNAPL distribution in the fractured shale, which has very small bulk fracture porosity (<0.01%). The DNAPL has undergone dissolution and diffusion-driven mass transfer into the rock matrix so that very little DNAPL now remains. Permanganate was identified as the most promising in situ technology for source zone remediation for several reasons, including its propensity to diffuse from fractures into the rock matrix causing chlorinated ethene destruction in both the fractures and the rock matrix. A short-term permanganate pilot test was conducted in March 2002 in which a near saturation KMnO4 solution was injected into a major transmissive zone. The test showed rapid lateral distribution in this zone across the source area. Then, episodic permanganate injections were conducted in two holes open to both the high and low transmissive parts of the fracture network from March through June, 2002. With depth-discrete multi-level monitoring using Westbay and Solinst CMT systems, these tests showed that permanganate was distributed throughout the fracture network in much of the source zone including low-permeability zones. After injections ceased, KMnO4 disappeared slowly. Chloride increases and changes in the carbon 13/12 isotope ratio of the remaining VOC mass indicate that VOC mass is being destroyed. These field results combined with laboratory tests for diffusion in the rock matrix and reactive transport modeling are being used to assist the design of the full-scale permanganate remediation system to prevent continued migration of VOCs from the source zone and to accomplish source mass reduction.