*5.2. Anion Exchange Resin*

The anion exchange column is an alternative process for separating metal ions from the contaminating elements where chelating resins such as Dowex [99] and Amberlite [100–106] are employed. The efficiency of the separation depends on both the anion exchange resin and the type of acidic eluent [99–101,106]. A malonic acid eluent in methanol is employed to circulate impurities through the Dowex ion-exchange column, resulting in the adsorption of uranium with an efficiency of over 99%. However, thorium recovery was inefficient, and it remained with REE in the eluate. The Amberlite XAD-4 is another anion exchange resin that is applicable in a wide range of pH [100] and could achieve 99% recovery of uranium [107]. Compared to the XAD-4, the Amberlite XAD-2 has less surface area and larger pore diameter, and extracts uranium when impregnated with Cyanex 302; however, it also partially co-extract thorium (separation factor U/Th <sup>=</sup> 1.2 <sup>×</sup> <sup>10</sup>4) [101]. In some cases, methanol is employed to elute the collected uranium in the column [101,107].

The anion exchange resin column is effective for uranium separation from REE, whereas it is inefficient for thorium separation [108]. Therefore, it can be used as the last step of REE purification. The Amberlite IRA 402 Cl resin was applied as a final purification step in a successive separation of thorium and uranium by precipitation. The low concentration uranium remained with the REE was then separated by anion chromatography, where REE recovery of 99% was achieved by the elution with NaCl. Next, 99% of uranium was recovered by water elution, Figure 8 adapted from [106].

In terms of the potential for the scale-up, in anion exchange resins, similar to cation exchange resins, the small flow rate of the feed solution, e.g., around 1 mL/min, is a significant limitation for an industrial-scale application [99,100,106].

**Figure 8.** Flow diagram of uranium removal from REE.

#### **6. Separation by Membranes**

Membranes have emerged as a new method for recovering thorium and uranium from the REE. At first glance, membranes were used due to their selectivity to individually recover thorium and uranium from other metal ions in liquid solutions [109–111], e.g., recovery of thorium by either graphene oxide (GO) or a silica membrane [112,113]. Such a recovery occurred by the formation of a complex between the element of interest and the membrane. In general, membranes have proved to be efficient and selective in acidic conditions at a pH of 4 to 5.5 for recovery of uranium and lower than 4 to recover thorium. However, to increase the recovery of radioactive elements, it is required to functionalize the membrane using an additive with a higher selectivity towards target elements [114–119]. Membrane functionalization is done by polymer adsorption at the membrane surface. Several functionalization methods are described in detail by Xu et al. [120]. The functionalized membrane has two main purposes, increasing the membrane resistance to the acid media, i.e., at pH lower than 4, and allowing for a better affinity of the membrane toward some ions by modifying some of the membrane parameters, e.g., surface rugosity, conductivity or hydrophobicity [120,121]. In this case, membrane functionalization is thus synonym of a increasing the selectivity of the membrane towards thorium and uranium versus REE (neodymium, europium, and samarium) at ambient conditions and pH below 2. For instance, Li et al. [109] functionalized graphene oxide with PDA (GO-PDA) and showed that, as in Figure 9 adapted from [109], it is more efficient for simultaneous recovery of thorium and uranium as compared with a graphene oxide membrane. The selectivity increase was caused by the surface modification of the functionalization that generated porous channels in the membrane that allowed the REE to pass through at low pH while being impermeable to thorium and uranium [109].

Despite the encouraging performance of the functionalized membranes for simultaneous recovery of radionuclides from the REE, further study is essential to evaluate the effect of various functionalized groups towards selectivity or the regeneration capability of the membranes, especially at high acidic conditions.

**Figure 9.** Separation factor of actinides/Nd by graphene oxide (GO) and functionalized graphene oxide with PDA (GO-PDA) membranes. (Reproduced with permission from ref. [109], copyright (2012), Elsevier).

#### **7. Selection of a Separation Process and Potential Waste Management Approaches**

Selection of an appropriate process from the presented methods for separating radioactive elements from REE requires various considerations such as the composition of the fresh ore, the applied upstream processes, operating conditions of the reagent, temperature and pH, scale of the process, purification range, and economy of the process, as well as advantages and limitations in each separation techniques. Table 7 summarizes these considerations for leaching, precipitation, solvent extraction, and ion chromatography.

As the separation of the radioactive elements from the REE is required in the REE supply chain, a new problem appears after this step since the presence of radioactive elements in tailings issues a waste management problem. Even though several waste management techniques exist such as water dilution [122], used when thorium and uranium are in the aqueous phase, or the safe storage when they are in the solid phase [123], those seem unreasonable due to high water or space occupation as well as health problem they may generate. One of the most recent propositions was to use thorium as a co-product of the REE industry as it would reduce the waste management problems, and could be used as a feed in the new generation of nuclear reactors [124,125]. As a matter of the fact, thorium is already a co-product of the titanium industry, and its recovery from the REE industry would represent the third most important thorium resource after titanium and uranium [124,126].


*Metals* **2020** , *10*, 1524

([136–138]).

#### **8. Conclusions**

Radioactive elements (thorium and uranium) are commonly associated with REE-bearing minerals, where the concentration depends on the mineral, the formation of the rocks, and the geographical position of the deposit. The presence of radioactive elements causes various problems in the environment and waste management. The extraction of thorium and uranium is necessary to ensure having a low radioactive product, while the loss of REE is minimized. The utilization of conventional hydrometallurgical processes such as selective precipitation, leaching, and solvent extraction for the extraction of radioactive elements is often conducted using complex industrial processes. There is a new trend of investigating new separation processes, such as the ion-exchange chromatography and membrane separation. These processes are yet at lab-scale development stage, but they seem to result in a more selective separation of Th and U from REE. Further research and development activities are required to maturate such processes and to evolve new technologies for economically viable applications at an industrial scale.

The most critical parameters to be controlled in these methods are the operating conditions (pH, and temperature), reagent type, and upstream processes. Depending on the process requirements and limitations, either one- or multi-steps processing would be applied to efficiently separate radioactive elements from REE. Considering the efficiency and the cost of the process, a specific process can be selected with regard to the advantages and limitations in each process.

**Author Contributions:** A.C.G. reviewed literature, wrote the first draft, and revised the manuscript. M.L. scientifically reviewed and edited the manuscript. A.A. scientifically reviewed and edited the manuscript. J.C. was the supervisor and also scientifically reviewed the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work is part of a NSERC/CRD funding supplied by Natural Science and Engineering Research Council of Canada and Niobec company in Province of Quebec of Canada.

**Acknowledgments:** Authors would like to sincerely acknowledge the support of NSERC and Niobec.

**Conflicts of Interest:** The authors declare no conflict of interest.
