Miljøprojekt, 1066 – Kemisk oxidation med permanganat

Kemisk oxidation med permanganat

Summary and conclusions

This report presents the results from laboratory experiments and field investigations regarding design and transport parameters for remediation with permanganate. The laboratory experiments are carried out with clay till sediments from two Danish sites (Dalumvej, Odense, and Stavnsbjerggård, Hvidovre) to illustrate the consumption of permanganate and PCE in clay till sediments, and, among these, to determine the kinetic parameters.

Parallel to this, core sampling has been conducted at Dalumvej, as a follow-up on a full scale project on remediation of a PCE contaminated site. The site is characterized by complex geology with clay till and sand lenses, where pollution and remediation is transported via the sandy lenses and diffusion into the clay till.

These results and experiences have been used to build a box experiment to simulate clean up of PCE with permanganate under controlled conditions. The set-up of the box experiment simulated conditions in the field with regards to transport of PCE and subsequent remediation with permanganate. The experiment has given basic insight in the controlling processes, among these consumption of permanganate in sand layers and the clay matrix and transport, diffusion and oxidation of PCE. Together with the other experiences, this is used to provide a conceptual understanding of the most essential mechanisms of remediation in the field.

The results from the laboratory experiments show that over 50 % of the total consumption of permanganate takes place within the first 10 hours. The total consumption of permanganate for the four investigated sediments shows NOD-MnO₄^- values between 3 and 20 g/kg DW for permanganate concentrations at 5,000 mg/l and 20,000 mg/l. The permanganate consumption for Dalumvej matches fairly well with earlier studies of clay till sediments, whereas the results for Hvidovre are generally higher than earlier findings.

The results of this investigation confirm that the natural oxidant demand (NOD) of the sediment increases with higher initial concentrations of permanganate. The dependence of the initial concentration means in practice that NOD values should be reported with the used concentration of permanganate. The consumption of permanganate also depends on the redox status of the sediment, since reduced sediment has a considerable higher consumption of permanganate. This means in practice that the location of the oxidation/reduction front is important to know when planning full scale remediations.

The results from the field studies at Dalumvej, compared with laboratory studies, show that it can be difficult to transfer laboratory measurements to consumption in a clay matrix. The results show a higher consumption in the batch experiments compared to results from the field and the box experiments, which can be explained by better contact between the sediment and permanganate in the batch experiments. Oxidant consumption determined by batch experiments with clay till gives an upper limit for consumption rather than the consumption seen in the field. For practical purposes with project proposals on chemical oxidation, it is suggested to base them on experiences with NOD values. If full scale remediation with chemical oxidation is decided, actual determination of NOD is recommended.

The consumption of permanganate does not follow first order reaction kinetics over the full experimental period of 20 days. The consumption follows a first order kinetic over the first eight hours, where the kinetic rates are determined to be in order of 0.02 to 3 hours^-1. The rates are dependent on the initial concentration of permanganate, as the rates increase with a lower concentration of permanganate.

The results show an effective mass removal of PCE in many cases. The consumption of PCE takes place in an interaction with the sediment consumption of oxidant, which has been shown to limit the PCE consumption, as the sediment consumes permanganate as fast as or faster than PCE. The kinetics following a first order expression is in accordance with literature data. Rates found in this experiment are in the order of 0.2-4 hours -1, which agree to rates for PCE oxidation in aqueous systems. There is a dependency on the permanganate concentration, the rates for the consumption of PCE being highest at a high concentration of permanganate.

In theory, the concentrations of PCE play, from a stoichiometric point of view, a limited role for the consumption of permanganate, which the experiments have confirmed. This is also observed/calculated for the core samples from Dalumvej and in the box experiments. It is important, though, to recognize that an ongoing supply of PCE can take place, while the permanganate-consuming parts of the sediment is fixed. With long-term remediation, where the parts of the sediments that can be oxidized gradually are consumed and PCE is supplied continuously due to diffusion, the concentration of PCE can be of importance.

The experiments show that the main part of the consumption of permanganate is due to the reactive constituents in the sediment. Earlier experiments looking into the relation between the total content of organic carbon (TOC) and NOD, showed an increase in NOD when TOC increased, though a significant scatter was observed. In this investigation a correlation for the four sediments has not been observed, but by comparing two clay till sediments from Dalumvej, respectively Hvidovre, a higher consumption of permanganate was observed with higher TOC.

A model for the reaction between permanganate and organic carbon, based on a second order reaction, was tested without success. The experiments show a larger decrease in TOC than the consumption of permanganate can explain, which can be due to variation of TOC in the sediment in the experiments and the reaction proportions from the model compound C₇H₈O₄. The results indicate that the oxidant consumption is governed by organic carbon, and the highest decrease in TOC takes place within the first 8-10 hours. This indicates that consumption of permanganate after the first hours does not adhere to the reaction with organic carbon. A possible explanation for the consumption of permanganate can be auto-destruction.

The results from the core samples from Dalumvej indicate that there has been a very large consumption of permanganate within the first year with a limited effect on the concentration of PCE, not only in the clay matrix, but also in the sand lenses. Transport of permanganate into the clay matrix via diffusion, identified as manganese dioxide in the cores (app. 15 cm in oxidised and 2-3 cm in the reduced zone) was in accordance with the expectations for the transport for one year. However, the box experiments show that diffusion of permanganate into the clay matrix is retarded due to reaction with PCE, organic matter and other reduced components. Only very low concentrations were detected within the reaction front in the clay. Thus, limited concentrations of permanganate must be expected to have been present in the clay matrix in the field site, which was not as expected. Diffusion from the central part of the clay matrix to the zone within the reaction front and the sand lenses is anticipated to have a higher importance for the remediation than expected earlier.

Surprisingly, high PCE concentrations (200 – 10.000 µg/kg) were found within the reaction front in the field studies. This is seen as an evidence of the back diffusion of PCE from the clay matrix after the permanganate was consumed. The box experiments confirm that back diffusion happens, but they also show that PCE and permanganate can co-exist. Thus, PCE and permanganate could co-exist in the field as well.

At first, the limited transport of permanganate in the reduced clay till, compared to the oxidised clay till, was surprising. The difference in NOD values from the oxidised and reduced zone was limited, but there was a significant difference in the COD values. Both the box experiments and the field investigations show that only a limited part of the permanganate (based on MnO₂ content in clay) was consumed. The limited transport of permanganate in the reduced zone could not only be due to higher TOC values but also to the content of reduced minerals (especially reduced iron compounds).

The consumption of permanganate is both due to reaction with the sediment and PCE, but auto-destruction also plays a significant role if the initial concentration of permanganate is high (> 5000 mg/l). The applied strategy at Dalumvej with a high load of permanganate at very high initial concentrations is not recommended in clay till with preferential flow in sand lenses. It is probably more appropriate to carry out a long-term or repeated injection with lower concentrations of permanganate. Thus, a low concentration of PCE in the sand lenses can be maintained, which in turn takes advantage of the back diffusion of PCE. The permanganate is used more for the oxidation of PCE rather than oxidation of the clay till. Though, the competition between PCE and organic carbon demands a certain concentration of permanganate (1000-5000 mg/l), dependent on the oxidant consumption from the sediment. Transport of permanganate in a sand lens will be faster when/if there is limited retardation in the sand lens.

Complete remediation of PCE in a clay till matrix will be very time-consuming and result in a high consumption of permanganate. However, partial remediation of the matrix could result in reduction of the flux of PCE to the sand lenses and, hereby, the concentrations in groundwater. Remediation of chlorinated solvents in a clay till matrix with other potential technologies will normally also be limited by matrix diffusion.