Radioactive Contamination over Geologic Time
Source: Ruth Tinnacher, Dan Hawkes
Environmental engineers need to predict the mobility of radioactive contaminants in soils and groundwater systems to assess the environmental risks associated with nuclear waste repositories and contaminated field sites. For this assessment, it is essential to know how radioactive contaminants, such as uranium, plutonium or neptunium, are distributed between the groundwater solution and the bulk mineral phase of a soil or host rock. Only contaminants that are dissolved or suspended in groundwater will be transported with the natural groundwater flow; contaminants that are precipitated or (ad)sorbed onto large soil mineral particles are typically immobilized or retarded (Fig. 1). However, contaminant (ad)sorption onto small mineral particles (mineral colloids) can increase the contaminant fraction that remains suspended in the groundwater solution and is potentially mobile.
Fig.1: The radioactive contaminant neptunium (Np) can only be transported with the natural groundwater flow if Np remains dissolved or suspended in the groundwater solution. Neptunium (ad)sorption onto the bulk mineral phase removes Np from the groundwater and slows down or immobilizes the contaminant. However, Np (ad)sorption onto mineral colloids potentially increases the fraction of Np that remains suspended in solution and is mobile due to ‘colloid-facilitated’ transport. The relevance of colloid-facilitated transport is ultimately dependent on Np (ad)sorption and desorption kinetics, as well as on aging, hysteresis and irreversible sorption behavior, since these characteristics determine over what time frame the contaminant remains associated with mineral colloids.
The relevance of this “colloid-facilitated” transport, however, is ultimately dependent on the kinetics of contaminant (ad)sorption and desorption reactions, since kinetics determine the time-frame over which contaminants remain associated with mineral colloids. In addition, various sorption phenomena, such as aging, hysteresis, and irreversible sorption, can cause differences between contaminant (ad)sorption and desorption behavior. These characteristics further affect the extent or kinetics of contaminant desorption from colloids, and as a consequence, the overall mobility of contaminants. In this context, (1) aging is defined as a series of chemical processes that change the chemical form (speciation) of a contaminant sorbed onto a mineral surface over time; (2) hysteresis is described by a difference in sorption equilibria, depending on the net direction of surface reactions (adsorption versus desorption); and (3) irreversible sorption occurs if a contaminant fraction remains sorbed after complete desorption equilibration with a solute-free solution. For the successful, long-term prediction of contaminant mobility in the subsurface, engineers need to determine which of these sorption characteristics are relevant for a specific system, and then conceptually include them in contaminant transport models.
ESD’s Ruth Tinnacher, in collaboration with scientists at LLNL and Clemson University, recently evaluated the relevance of these characteristics for neptunium(V) (Np(V)) sorption/desorption onto an iron-oxide colloid (goethite), using a 34-day flow-cell experiment and kinetic modeling. Based on the experimental results (recently published in Geochimica et Cosmochimica Acta), the Np(V) desorption rate is much slower than the (ad)sorption rate, and appears to decrease over the course of the experiment. The best model fit with a minimum number of fitting parameters was achieved with a multireaction model, which assumes a fast initial adsorption step followed by slow, consecutive surface reactions (Fig. 2). Furthermore, the new modeling concept includes a parameter related to transition state theory. This parameter allows the researchers to simulate differences in adsorption and desorption kinetics based on an assumption of different overall reaction pathways for net (ad)sorption and desorption processes.
Fig. 2: Comparison between experimental flow-cell data (circles) with best model fit (solid line). The kinetic multi-reaction model includes a fast (equilibrium) Freundlich site, a consecutive, kinetically-limited, first-order site, and the fitting parameter ψ2,de related to transition state theory. Dashed lines represent calculated model outputs using the upper and lower limit values of ψ2,de (UL=1.62 ´ 10-2, LL=3.75 ´ 10-3).
Based on modeling results, the researchers concluded that aging processes are relevant in this system, while hysteresis and irreversible sorption behavior can be neglected within the time frame (desorption over 32 days) and the chemical solution conditions evaluated. Neptunium desorption is very slow, possibly due to a complex reaction pathway, but it is not irreversible.
Hence, these results suggest that after the initial Np(V) sorption onto goethite colloids, the contaminant might slowly desorb from colloid surfaces over extended time frames. Therefore, this study does not support an assumption of “irreversible Np(V) sorption” onto goethite colloids in contaminant transport models. As a result, the relevance of colloid-facilitated Np(V) transport might be limited to small time frames and short distances from the initial source of contamination, provided that the conditions at a field site and in this lab experiment are comparable.
Reference: Tinnacher, R.M., M. Zavarin, B.A. Powell, and A.B. Kersting (2011), Kinetics of neptunium(V) sorption and desorption on goethite: An experimental and modeling study. Geochimica et Cosmochimica Acta, 75 (21), 6584–6599; published online, DOI: 10.1016/j.gca.2011.08.014.
