Characterization of ruthenium tetroxide capture from dry gas streams by silica gel and metal surfaces

Processing used nuclear fuel releases volatile and semivolatile radionuclides into the off-gas streams associated with each processing step. Efforts to develop methods for controlling the release of gaseous radionuclides to the environment have identified ¹⁰⁶Ru as one of the semivolatile nuclides requiring a high degree of abatement when processing fuel cooled less than 20 years. The tritium pretreatment process will release ruthenium as volatile ruthenium tetroxide (RuO4). There are few studies characterizing ruthenium removal from a dry gas stream, and many of the engineering parameters required to design a ruthenium abatement system for this application have not been determined. Testing examined the capture of RuO4 by silica gel and steel wool/metal mesh. This testing was intended to provide key engineering data required for the design of a capture bed to remove RuO4 from a tritium pretreatment off-gas stream. Two methods were used to produce gaseous RuO4: dry oxidation of powdered ruthenium metal and oxidation of ruthenium dioxide (RuO2) with sodium periodate in an aqueous phase. The second method was more successful in producing volatile ruthenium in the quantities needed for experimentation. Initial tests showed that RuO4 deposited on the test equipment. This deposition was minimal at ambient temperature but was observed to be rapid and complete at 150°C. Scoping tests found RuO4 retention from a dry air stream by silica gel at 40°C required a greater bed length than RuO4 deposition on metal. RuO4 retention on silica gel was found to be reversible, with RuO4 diffusing from the Ru-bearing-silica gel during storage. Deposition of RuO4 onto metal surfaces was characterized across a temperature range of 50-250°C. At a bed temperature of 150°C, penetration of the RuO4 into the bed varied with the packing density of the metal, ranging from 3.8-15.3 cm. Increasing the bed temperature to 250°C decreased the penetration into the bed to less than 1.8 cm. The deposition zone did not increase in length as additional RuO4 was delivered to the sorbent bed, indicating that the depth of the deposition layer on the metal sorbent will continuously increase until the bed is removed from service. Analysis of the RuO4 concentration in the bed effluent resulted in estimated decontamination factors in excess of 10⁶ for sorbent beds composed of steel wool.