Summary and conclusions

Monitering af PCE-afværge ved kemisk oxidation (permanganat) i moræneler - Hovedrapport

In 2001 a remediation project was started to treat PCE contamination at Dalumvej 34B, Odense SV. The contamination originated from a dry cleaning facility at the site and had migrated into the top 10 meters of the clay till and to the groundwater downstream of the site. The remediation project included excavation of contaminated soil - using large diameter augers - and subsequent in-situ chemical oxidation using potassium permanganate.

The overall remediation project was implemented and financed by the County of Funen, and part of the monitoring was financed by the Danish EPA (development fund related to soil and groundwater contamination issues). The remediation project was carried out by a consortium comprising Hedeselskabet A/S (currently Orbicon A/S) and NIRAS A/S.

The purpose of the remediation project was to reduce the contamination in the target treatment zone and thereby reduce the potential for future migration to soil vapor and groundwater. Additionally the purpose was to acquire monitoring experience related to the performance of chemical oxidation using potassium permanganate.

On the basis of the contamination at Dalumvej 34B, Odense, SV Environment & Resources, the Technical University of Denmark investigated the transport parameters and the effect of permanganate oxidation in clay till and sand lenses associated with preferential transport. Additionally Orbicon A/S set up a transport model to illustrate the effect of chemical oxidation using potassium permanganate in clay till, and thereby identified the most critical hydrogeological, physical and chemical parameters for the remediation process.

This report is to present the most relevant experience from the overall work related to the remediation project at Dalumvej 34B, Odense SV.

Remediation strategy and remediation project performed

After excavation of the contaminated soil (using large diameter augers) the bore casings were backfilled with a mixture of sand, gravel, water and potassium permanganate. A total mass of 12,000 kg potassium permanganate (solid phase) was installed at the site. This mass generated a very high initial KMnO₄-concentration in the groundwater (>30,000 mg/l). Upon backfill the potassium permanganate was distributed via the groundwater flow to sand lenses in the clay till. From these sand lenses potassium permanganate was transported into the clay matrix via diffusion. As a part of the remediation project a supply and injection system was set up to facilitate future injection and distribution of aqueous phase permanganate. This system was tested with a positive result (water injection), but an aqueous phase permanganate injection was never performed.

Monitoring program

In order to track the distribution of potassium permanganate and in order to investigate the remediation effect, a monitoring program was set up. The monitoring program included measurement of hydraulic head and conductivity as well as water sampling to analyze the concentrations of potassium permanganate, chloride, chlorinated solvents, chlorinated daughter products and aqueous phase chromium (chromium^VI).

The monitoring program included 27 monitoring wells (82 screened intervals), 5 large diameter wells (in the target treatment zone) and a nearby stream. To facilitate sampling from various water-bearing depths, 9 of the 27 monitoring wells were fitted with screened intervals at 5-7 depths (multi level sampling).

To track possible soil settlings caused by oxidation of the reactive constituents in the soil, the monitoring program also included control leveling to 25 newly established and well-defined benchmarks.

Sampling according to the monitoring program was started in January 2003 (3 months after the installation of potassium permanganate) and continued up until February 2005 (28 months after the installation of potassium permanganate).

Distribution of oxidant

Already after the first monitoring (January 2003) it was evident that the distribution of oxidant in the sand lenses occurred much faster than expected and to a greater extend than expected. The greater oxidant distribution was due to changes in the flow conditions at the site, including:

Large hydraulic conductivity in the sand lenses.
A vacuum effect caused by the large diameter auger excavation.
A density effect due to the larger density for aqueous phase permanganate compared to groundwater with no permanganate.
A rise in water level due to the water injection (injection test for supply and injection system).

Overall potassium permanganate was distributed in the sand lenses over a fairly large area (app. 600 m²) and closest to the target treatment zone potassium permanganate was observed in several water-bearing depths.

After the initial very large distribution of potassium permanganate a significantly weakened distribution of potassium permanganate was observed. This weakening in distribution was not expected. One year after the installation of potassium permanganate only a limited quantity remained in the target treatment zone, and in the downstream area the oxidant distribution was reduced to a narrow plume. The reduced oxidant distribution could be due to:

An initial fast and extensive distribution of oxidant and a subsequent consumption of oxidant.
A greater consumption of oxidant than expected to oxidize PCE and the reactive constituents in the soil.
Auto-destruction of potassium permanganate.

Regarding the distribution of oxidant at Dalumvej 34B, specific investigations have showed that permanganate transport into the clay matrix via diffusion was limited to 15 cm and 2-3 cm in the oxidized zone and the reduced zone, respectively.

Finally, evaluation of the permanganate distribution - and the risk of unintended off-site distribution - has emphasized the importance of a risk assessment prior to the use of potassium permanganate. Furthermore this evaluation has identified a need for a contingency plan in order to address unintended off-site distribution of potassium permanganate.

Remediation effect

At Dalumvej 34B, local oxidation of PCE in soil (clay till) and groundwater (water-bearing sand lenses) was documented. The oxidation, which was most evident in the downstream monitoring wells, was linked to actual or previous presence of potassium permanganate. However, the remediation project did not result in a complete remediation of the site.

In the evaluation of the remedial effect at the site, it is likely that the PCE oxidation has been limited by the installed amount of potassium permanganate (too small amount) and by back diffusion of PCE from the clay matrix to the sand lenses.

The remedial effect is also likely to be influenced by a new source zone discovered at the north western part of the site.

Model simulations with site-specific parameters demonstrated that continuous injection of permanganate over a period of 10-20 years would not lead to remediation (oxidation) of the total mass of PCE in the clay till at the site. Back diffusion of PCE from the clay matrix to the sand lenses would be a limiting factor, and due to this limitation, the effect of a continuous injection of permanganate would not differ substantially from the situation with no injection.

Model simulations of in-situ chemical oxidation in the clay till with preferential transport in sand lenses has demonstrated that the distance between different sand lenses is the most important parameter in mass removal. In order to obtain an effective remediation, the distance between the sand lenses must be no larger than 20 cm. Furthermore model simulations have demonstrated that the downstream contaminant concentration can be reduced by an increase in the water flow in the sand lenses, by an increase in the injection of permanganate or by a reduction in the distance between the oxidant injection points.

Future use of chemical oxidation with permanganate

It is recommended that a future chemical oxidation project with permanganate is not performed with solid phase installation of oxidant (potassium permanganate). As an alternative, it is recommended to inject liquid phase permanganate in concentrations of 1,000-5,000 mg/l. Additionally it is recommended to use an appropriate sequential injection (pulse injection) of permanganate. These recommendations are based on the fact that oxidant consumption is dependant on the oxidant concentration and that injection/installation of high concentrations of permanganate is associated with a risk of auto-destruction of permanganate.

It is recommended that future chemical oxidation remediation projects using permanganate include a risk assessment related to the use of permanganate. Furthermore it is recommended to prepare a contingency plan addressing unintended off site distribution of permanganate.

In relation to the choice of monitoring parameters, the experience from Dalumvej 34B (clay till with preferential transport in sand lenses) has indicated that:

Visual assessments can be used to estimate approximate permanganate concentrations in groundwater samples.
Visual assessments can identify permanganate concentrations above 3 mg/l and distinguish concentrations up to 2,000 mg/l.
Conductivity measurements cannot be used as a reliable indicator for permanganate concentration.
Chloride analysis cannot be used as a reliable indicator for previous oxidation of PCE (and subsequent formation of chloride) using permanganate.
Interrelated data series of PCE and KMnO₄concentrations are required in order to evaluate the remediation effect.
Accurate levelling can be used to document that no soil settlings occur, due to the oxidation of reactive constituents in the soil.

In the clay till oxidation of contamination will occur (reduction of contaminant mass). The reaction conditions (reaction zone and back diffusion) will, however, at the same time result in a situation where contamination located within the clay matrix cannot be completely removed. For this reason in-situ chemical oxidation in low permeable layers must be considered only as a cut off remediation technique for reduction of the off site migration of contaminants.

Based on the experience from Dalumvej 34B, the use of chemical oxidation with permanganate in low permeable media is evaluated to have a limited remediation effect. During the remediation project (active phase) chemical oxidation with permanganate will be able to reduce leaching of contaminants from the site, and based on this statement in-situ chemical oxidation should be considered only as a cut off remediation technique.

However, in high permeable media, the results from other remediation projects indicate that in-situ chemical oxidation using permanganate can result in a complete and successful remediation.