Comparing hexacyanoferrate loaded onto silica, silicotitanate and chabazite sorbents for Cs extraction with a continuous-flow fixed-bed setup: Methods and pitfalls

Abstract

Radioactive ¹³⁷Cs is one of the most common and problematic radionuclides in nuclear wastes. Decontamination typically involves passing the waste in continuous flow through an agitated or fixed bed reactor containing an efficient sorbent. There are many articles in the literature describing a broad spectrum of highly efficient sorbents. However,their properties is often difficult, mainly because the experimental conditions used differ. We describe the series of experiments that need to be performed to characterize Cs sorbents and illustrate by comparing three of these that, for the extraction of trace elements, the kinetics and selectivity of the exchange process are far more important than the maximum extraction capacity of the material.

Introduction

Most studies of solid phase extraction processes for effluent decontamination focus on the maximum sorption capacity (Q_max) as the main parameter to optimize (Zhang et al., 2019; Olatunji et al., 2015; Alby et al., 2018). The sorption capacity (Q) of a material is the amount or mass of contaminant captured per unit weight; the higher the sorption capacity of a material is, the less is required for a given purification or recycling process. Maximizing the sorption capacity is therefore an understandable objective. However, this maximum sorption capacity is never reached in processes for the treatment of solutions such as radioactive effluents in which the contaminants are at trace level and there are high concentrations of competitive species. In this context, comparing candidate sorbents in terms of their maximum sorption capacity at equilibrium is inappropriate.

A second important parameter for sorption materials, which is only sometimes considered (Alby et al., 2018), is their selectivity. This can be estimated by performing experiments in the presence of ions that compete for adsorption with the target species. To be relevant, these experiments have to be carried out with a ratio of competitive to target ions that is representative of actual effluents, i.e. of several order of magnitude, because the target ion (e.g. of a radioactive element) is typically at trace concentration. The distribution coefficient (K_d) is the amount of target species adsorbed relative to the amount remaining in the solution. In other words, it represents the concentration distribution of the contaminant between the solid and liquid phases. In solid phase extraction processes, the distribution coefficient is not dimensionless and depends on experimental conditions used for a given experiment. This makes the results of different experiments difficult to compare.

The sorption capacity and distribution coefficient are measured using batch mode sorption experiments. The sorption isotherm plotted as a function of the remaining concentration of target ions (the Q-mode curve) typically plateaus at Q_max at high concentrations and can be fitted by a Langmuir model. The curve representing the distribution coefficient as a function of the remaining concentration of the targeted ions (the Kd-mode isotherm) is typically flat at low concentrations and then decreases linearly. The plateau value of K_d is thus the highest that can be obtained with the investigated sorbent under the experimental conditions used, with higher values being favorable for the decontamination of trace ions. For K_d estimates to be realistic and comparable however, the experiments have to be performed under representative conditions, with the same solid/solution ratio, the same kind of effluent composition (salinity, pH) and trace level contaminants.

For continuous processes, it is also important to consider the sorption kinetics and the time dependence of the sorption capacity. Here, the key parameter is the time taken to reach equilibrium. For fixed bed processes furthermore, the geometry of the setup and the initial concentration of the contaminant need to be taken into account when studying of performance of a sorbent. This is typically assessed by measuring the concentration of the target element at the outlet of a column filled with the sorbent material as a function of the volume passing through. The shape of the resulting breakthrough curve depends on the geometry of the column, the flow rate and the initial concentration of the contaminant. It has been shown that the dynamic capacity of a column is close to the maximum batch adsorption capacity, meaning that the column process is at equilibrium, provided the height/diameter ratio is greater than five (Michel et al., 2018). For optimal performance moreover, the diameter of the column should be at least 40 times greater than the average particle diameter (DePaoli et al., 2001). The flow rate is important because fixed bed processes are only considered industrially viable at a Darcy velocity of 1 m·h⁻¹ of higher. Finally, although this is often neglected in the literature, the shape of the breakthrough curve also depends on the contaminant concentration because decreasing it to trace levels may affect the sorption kinetics (Mahendra et al., 2015).

This paper compares the sorption efficiency of three ion exchangers: Sorbmatech®, a K-Cu hexacyanoferrate; a Na-Chabazite (Herschelite) type zeolite; and a Nb substituted Na crystalline silicotitanate (CST). Crystalline silicotitanates are commercially available and have been developed for a long time (Miller and Brown, 1997). Their specific crystalline structure (Celestian et al., 2008) with several ion exchange sites allows to be used for Cs or Sr decontamination, depending on the CST composition (Chitra et al., 2011; Zhao et al., 2018). The Nb substituted Na-CST used here has been shown to offer improved Cs sorption (Clearfield et al., 2006). Zeolites are among the most widely used inorganic materials for water treatments because of their high cation exchange capacity and they have been widely studied for the removal of radionuclides (Van Speybroeck et al., 2015). While their selectivity is often low in saline solutions, they are inexpensive (naturally available) and their composition can be adjusted (Al/Si ratio, amount and nature of guest ions) to extract Cs or Sr (Van Speybroeck et al., 2015). Chabazite can accommodate various monovalent cations in its unit cell of adaptable size (Kong et al., 2016). In a recent comparison of three zeolites with different Si/Al ratios (chabazite, stilbite and heulandite), Baek et al. (2018) found that chabazite, in powder form, with the lowest Si/Al ratio, captured Cs the most rapidly with the highest sorption (exchange) capacity.

We measured sorption isotherms at two contact times (2 and 48 h) to compare the maximum sorption capacity of the three sorbents at high Cs concentration and their distribution constants at trace Cs concentration. We also measured the sorption kinetics of the three materials and performed breakthrough experiments with different Darcy velocities and inlet concentrations. This study highlights the careful experiment design and interpretation that is needed to choose a sorbent for the continuous-mode column decontamination of trace elements from effluents. This approach is applied here for Cs removal but can be easily transposed to the removal of other trace level radioactive species such as Sr.