[en] This report provides an assessment of the use of nitrogen trifluoride for removing oxide and water-caused contaminants in the fluoride salts that will be used as coolants in a molten salt cooled reactor. The Pacific Northwest National Laboratory, in support of the Oak Ridge National Laboratory's program to investigate an advanced molten salt cooled reactor concept for the U.S. Department of Energy, evaluated potential nitrogen trifluoride (NF₃) use as an agent for removing oxide and hydroxide contaminants from candidate coolants. These contaminants must be eliminated because they increase the corrosivity of the molten salt to the detriment of the materials of containment that are currently being considered. The baseline purification agent for fluoride coolant salts is hydrogen fluoride (HF) combined with hydrogen (H₂). Using HF/H₂ as the reference treatment, we compare HF and NF₃ industrial use, chemical and physical properties, industrial production levels, chemical, toxicity, and reactivity hazards, environmental impacts, effluent management strategies, and reaction thermodynamic values. Because NF₃ is only mildly toxic, non-corrosive, and non-reactive at room temperature, it will be easy to manage the chemical and reactivity hazards during transportation, storage, and normal operations. Industrial experience with NF₃ is also extensive because NF₃ is commonly used as an etchant and chamber cleaner in the electronics industry. In contrast HF is a highly toxic and corrosive gas at room temperature but because of its significance as the most important fluorine-containing chemical there is significant industrial experience managing HF hazards. NF₃ has been identified as having the potential to be a significant contributor to global warming and thus its release must be evaluated and/or managed depending on the amounts that would be released. Because of its importance to the electronics industry, commercial technologies using incineration or plasmas have been developed and are used to destroy the NF₃ in a facility's gaseous effluent stream. A process has been developed and used to recover and recycle NF₃. The electronics industry is actively pursuing alternative methods to control NF₃ releases. In comparison, HF has not been identified to be a potential global warming gas nor has it been determined to have any other environmental affect. Also because of the high solubility of HF in water and aqueous caustic solutions, the HF industry has developed and used aqueous scrubbers to effectively prevent its release into the environment. Care appears to be necessary when using NF₃ in a plant. Precautions must be taken to prevent adiabatic compression and make sure that NF₃ thermal decomposition does not occur in unplanned locations. The system must be engineered to avoid the use of ball valves and sharp bends. The materials of construction that will be required to contain NF₃ and anhydrous HF will be similar. If water is present such as in the process effluent, HF is more corrosive than NF₃ and its containment would require nickel or nickel-based alloys. Both of these fluorinating agents become more reactive with increasing temperature and would require pure nickel or nickel-based alloys for containment until the gas stream has cooled. With respect to the cost of the fluoride, HF is about one third the cost of NF₃ on a fluorine basis. Of the fluorine-containing chemicals, more HF is produced than any other. NF₃ is produced on an industrial scale and its capacity has grown each year since being identified as a useful etchant. Both NF₃ and HF have been demonstrated to be effective at removing oxide, hydroxide, and water contamination from fluoride salts during melt processing of fluoride glasses while HF in combination with H₂ has been demonstrated to be effective for some of the candidate coolant salts and some of their individual constituents such as beryllium oxide (BeO). HF has a limited solubility in molten 66 mol% LiF-33 mol% BeF₂ indicating that treatment with HF will result in free F- in HF-treated fluoride salts. H₂'s flammability and potential explosivity introduces additional hazards to its use. With respect to chemical viability, as measured by reaction free energies, NF₃ is the stronger fluorinating agent when compared to HF. For all postulated contaminants the calculated free energies for treatment by NF₃ were negative, indicating that the reactions were favorable and should occur provided there are no kinetic barriers. In contrast, HF's fluorinating power declined with increasing temperature, and in a couple of instances the reaction free energy became slightly positive (e.g., BeO above 700 C), indicating that use of excess HF would be required for the fluorination to occur or that the product water would have to be removed to force the reaction to occur.