(136d) Wet Air Oxidation Batch Autoclave Testing with Savannah River Site Tank 48H Actual Waste

Tank 48H at the Department of Energy's (DOE) Savannah River Site (SRS) in Aiken, South Carolina currently holds legacy material containing organic tetraphenylborate (TPB) compounds from the operation of the In-Tank Precipitation process. This material is not compatible with the waste treatment processes at SRS and must be removed or undergo treatment to destroy the organic compounds before the tank can be returned to Tank Farm service. In addition, the material can decompose to benzene and lead to potentially flammable concentrations in the tank headspace. Tank 48H currently holds approximately 250,000 gallons of alkaline slurry containing 21,800 kg of potassium and cesium tetraphenylborate (KTPB and CsTPB) solids. The tank has been isolated from Tank Farm service, and its return to service is a high priority to the DOE.

Wet Air Oxidation (WAO) is one of the two technologies currently under consideration for the treatment of TPB in Tank 48H. WAO is an aqueous phase process in which soluble and/or suspended waste components are oxidized using oxygen or oxygen in air. In general, the process operates at elevated temperatures and pressures ranging from about 150 to 320 °C and 7 to 210 atmospheres, respectively. The products of the reaction are carbon dioxide, water, and low molecular weight oxygenated organics (e.g. acetate, oxalate, etc.)

Batch bench-scale autoclave testing with actual (radioactive) Tank 48H waste is among the tests required in the DOE WAO Technology Maturation Plan. The objective of the testing is to confirm the ability of the WAO process to destroy TPB and its associated compounds using actual (radioactive) Tank 48H waste. Specifically: (i) Confirm the TPB destruction efficiency obtained from comparable simulant tests at same reaction times, (ii) Determine the destruction efficiency of other organics including biphenyl, (iii) Identify and quantify the reaction byproducts, and (iv) Determine off-gas composition.

Batch bench-scale stirred autoclave tests using both simulated and actual Tank 48H wastes have been conducted. The study was approached in three ways. (1) Replication of the 2006 batch bench-scale shaking autoclave simulant tests at the same conditions as those identified in 2006; (2) Batch bench-scale stirred autoclave simulant tests at nearly the same (i.e, conservative) conditions identified during the continuous-flow pilot-scale testing conducted in the January 2009; and (3) Replication of item # 2 above using actual or radioactive Tank 48H waste.

The test conditions for the first approach were a temperature of 300 °C, a reaction time of 3 hours, CuSO4.5H2O solution (500 mg/L Cu) as catalyst, and 1:1 Tank 48H simulant dilution with 2M NaOH/CuSO4.5H2O solutions, and 1 mL/L of antifoam agent.

The test conditions for the second and third approach were a temperature of 280 °C, a reaction time of 1 hour, CuSO4.5H2O solution (250 mg/L Cu) as catalyst, and 1:1 Tank 48H material (simulant or radioactive waste) dilution with 2M NaOH/CuSO4.5H2O solutions, and 0.21 mL/L of antifoam agent.

The test conditions for the second and third approach are conservative because the continuous-flow pilot-scale testing used a dilution ratio of 2:1. Note that even though a catalyst concentration of 100 mg/L Cu was used in the pilot-scale testing, the 250 mg/L Cu used in approaches 2 and 3 testing is not expected to impact the performance. Only the soluble form of the Cu performs the catalytic action. The amount of Cu that remains soluble is less than 100 mg/L.

The tests of the first approach demonstrated a good replication of the results of the 2006 testing. TPB, its daughter compounds (triphenylborane [3PB], diphenylborinic acid [2PB], phenylboronic acid [1PB]), phenol, and biphenyl, in the simulant were all destroyed to below their respective detection limits. It must be noted that biphenyl was only partially destroyed in the 2006 shaking autoclave testing. The almost total destruction of biphenyl is attributed to the better mixing characteristic of stirred autoclaves. Other results in terms of conversion of nitrite to nitrate, solid compounds in the treated simulant identified by x-ray diffraction, constituents in the off-gas, etc. were largely comparable.

The tests of the second and third approach demonstrated WAO is a viable technology for the treatment Tank 48H radioactive waste. TPB, 3PB, 2PB, 1PB, phenol, nitrobenzene, nitrosobenzene, etc. in the radioactive waste were all destroyed to below their respective detection limits. In fact, biphenyl is the only organic compound of interest that was not completely destroyed. The partial destruction of biphenyl stems from its propensity to go to the vapor phase during the oxidation reaction.

The undestroyed biphenyl is essentially biphenyl vapor that solidifies (because of its low melting point ? about 68 oC) when the autoclave is cooled at the end of the WAO reaction. The undestroyed biphenyl is expected to be in the off-gas in the actual WAO process, which is a continuous-flow system.

Destruction efficiency of TPB was > 99.997% while that of the other organic compounds (except biphenyl) were for the most part > 85%. The fairly low destruction efficiencies (compared to that of TPB) for most of the other compounds are due to their relatively high detection limits and/or low initial concentrations in the waste.

The autoclave radioactive waste tests were a confirmation to earlier simulant tests performed at identical conditions. The radioactive waste tests and the simulant tests were fairly comparable particularly, the destruction efficiencies of the major organic compounds (TPB and phenol).

All the above results should be viewed in the context that the test conditions are conservative compared to the conditions identified in the continuous-flow pilot-scale testing. By all accounts, had the tests been conducted at a 2:1 diluent/Tank 48H waste volume ratio as identified in the pilot-scale testing (versus the 1:1 diluent/Tank 48H waste volume ratio used for the current tests), the destruction efficiency of all the organic constituents in the Tank 48H waste especially biphenyl most likely would have been greater.

The common compounds identified by x-ray diffraction (XRD) in all four tests (i.e., two radioactive waste tests and two simulant tests) were CuO (tenorite), carbon (graphite), and Na2Ti3O7 (sodium titanium oxide). Graphite is an artifact of the tests. It is the result of autoclave graphite gasket breakage and possibly erosion from graphite stirrer bushing.

Additional compounds identified by XRD in radioactive tests 1 and 2 were H2(Ti2SiO7)(H2O)1.5 (hydrogen titanium silicate hydrate) and FeO(OH) (goethite) respectively. Even though these compounds were not identified in the simulant tests, they are all believed to present in all the treated material. Smaller loss of solids during pouring of treated material in the simulant tests compared to the radioactive tests precluded their identification by XRD.

The percent conversion of nitrite to nitrate was higher in the simulant tests (17%) than radioactive tests (5%). The organics-oxygen reaction from the increased number of organic constituents or compounds (oxalate, formate, nitrobenzene, etc.) in the radioactive waste seems to supercede the nitrite-oxygen reaction.

The major radionuclide in the waste, Cs-137, is present in the treated waste as soluble species even though it is in the untreated waste as 137CsB(C6H5)4 solids. The same is true for B and K which are also present in the untreated waste as insoluble 137CsB(C6H5)4 and KB(C6H5)4 solids. Cs-133, the non radionuclide cesium in the waste behaves in a similar manner. Np-237, U-233, U-233, U-234, U-235, U-236, U-238, and Pu-238 are largely present in the treated waste as soluble species while Sr-90 is present in insoluble form.

The untreated off-gas from both the radioactive waste and the simulant tests indicated the major constituents are nitrogen, oxygen, benzene, and hydrogen in that order. All other gases (methane/other volatile organic compounds, carbon dioxide, carbon monoxide, nitric oxide, and nitrious oxide) were either below detection limits or were not detected at all. Benzene and hydrogen in the off-gas for all the tests were relatively low. The highest concentrations of benzene and hydrogen in the off-gas obtained for all four tests were 0.096 and 0.0063 vol% respectively. These translate to 6.9% of benzene lower flammability limit (LFL) of 1.4 vol% at 25 oC and 0.16% of hydrogen LFL of 4 vol% at 25 oC.