Dissolution and Clarification of Rertr Silicide Fuels*

DISSOLUTION AND CLARIFICATION OF RERTR SILICIDE FUELS*

George C. Rodrigues E. 1. du Pont de Nemours Co. Savannah River Laboratory Aiken, SC 29808 XA04C1501

Abstract

SRP may reprocess spent REM fuel as early as FY-85. Candidate RERTR fuels include uranium oxides, aluminides, and sicides. Much less is known about the behavior of uranium sicides during Purex reprocessing than either uranium oxides or aluminides. Studies were performed at SRL with RERTR silicide fuels to ensure their reprocessability. The results of dissolution and head-end clarification studies with both unirradiated and irradiated fuel are presented.

Introduction

Highly enriched uranium fuels from a number of foreign and domestic research and test reactors are currently reprocessed at the Department of Energy's (DOE) Savannah River Plant (SRP). The objective of the Reduced Enrichment Research and Test Reactor (RERTR) program is to demonstrate on- version of these reactors to low enriched uranium (LEU) fuels to reduce the potential for diversion of highly enriched uranium. Ideally, this conversion should be accomplished without changing the fuel configuration or de- grading reactor performance. Conversion to LEU under these constraints requires that the new fuel elements contain up to 41/2 times as much total uranium. Advanced powder metallurgy fabrication techniques utilizing high density oxides, aluminides, and sicides should produce reliable fuels with uranium loadings in the desired range. Candidate silicide intermetallic compounds are U3Si, U3Si2, and U3SiAl. While a good deal is known about reprocessing oxides and aluminidesl,2, information on reprocessing sicides is limited and dissolution of some uranium sicides has proven difficult. In addition, dissolved silicon can disrupt solvent extraction2j3. Therefore, reprocessing studies on silicide fuels were required to demonstrate compati- bility with current SRP operations.

Aluminum clad fuels are reprocessed at SRP using a modified PUREX process. The irradiated fuel and cladding is first codissolved in mercury catalyzed nitric acid Trace fluoride is sometimes required for high burnup fuels. Fluoride catalytically attacks silicon which grows into the aluminum cladding by neutron capture during irradiation. Dissolved silicon above 100 ppm can severely disrupt solvent Sxtraction by stabilizing emulsions of the organic and aqueous phases Dissolved fuel is treated with gelatin (referred to as a gelatin strike), which polymerizes the dissolved silicon to a solid which can be separated by centrifugation. The dissolved fuel is also treated with a reverse permagnianate strike. This step lowers the gamma radiation level by scavaging Zr-95Nb in the MnO2

The information contained in this article was developed during the course of work under Contract No. DE-AC09-76SROOOOI with the U.S. Department of Energy.

134 precipitate formed. Uranium and plutonium are then separated from the dissolved fission products and cladding by solvent extraction into a tributyl phosphate-normal paraffin mixture.1

Studies at Savannah River Laboratory (SRQ examined dissolution rate, silicon removal,-and nuclear safety aspects of dissolving and clarifying REM silicide fuels.

Summary

REM silicide fuels were successfully dissolved in mercury catalyzed nitric acid in bench-scale tests. Trace fluoride was necessary for a con- sistent and reasonable dissolution rate of irradiated silicide fuel. The dispersed fuel phase dissolved at least as fast as the aluminum matrix and cladding. Solids remaining after dissolution were amorphous silica contain- ing negligible uranium. Average composition of the dissolved fuel solution was 24 gm/l U, 3.OM H, 1.3M A1+3, and 6.9M NO3_. A reverse permanganate strike gave satisfactory results. A gelatin strike reduced dissolved silicon to less than 70 ppm, which is acceptable to solvent extraction. Irradi- ated fuel was dissolved and clarified as feed for bench-scale solvent extraction studies.

Discussion

Headend dissolution and clarification studies were conducted at SRL with both unirradiated and irradiated RERTR miniplates provided by Argonne National Laboratories (ANL). These developmental fuel elements were pre- pared for fabrication and irradiation studies. Depleted miniplates were produced during fabrication and uranium loading studies at ANL. Enriched miniplates were produced at ANL and irradiated to about 30% 235U burnup in the Oak Ridge Reactor (ORR) for post-irradiation examination at ANL.

Dissolution Studies

Bench-scale studies demonstrated that RERTR silicide fuels show no significant difference in dissolution from aluminum clad fuels now reprocessed at SRP. Dissolution procedures for REM silicide fuels are based on current SRP practices and previous SRL studies on dissolution of aluminum clad fuels.

REM dissolution studies were conducted in bench-scale glassware. The dissolver apparatus consisted of an electrically heated flask fitted with a water cooled reflux condenser. The flask was also fitted with a burette for metering in reagents during dissolution. Studies with the intensely radio- active irradiated fuel were conducted in SRL's heavily shielded High Level Cells (HLC) facility.

Boiling 3M HN03 - 0.002M Hg+2 completely dissolved sections 4 gms each) of unirradiated miniplates of all three silicide compositions. Initial dissolution was vigorous, but with controllable foaming. Boiling 8M HN03 0.002M Hg+2 exhibited initial dissolution rates so high that foaming was barely controlled in the oversized condenser used. Dissolution in IM HN03 0.002M Hg+2 was slow and inconsistent. One dissolution was made wit-h- reduce(i-Hg+2 (0.0002M). This run reduced the dissolution rate by almost 12

135 and showed that Hg+2 is necessary to the dissolution. The dispersed fuel particles dissolved as fast as the aluminum matrix and cladding in all tests. The range of total dissolution times of 7 to 15 hours for all tests compares favorably with times for fuels now processed at SRP.

The acid consumed during the course of a dissolution was replenished periodically with makeup HN03- 16M HN03 makeup acid constantly sustained a higher dissolution rate than 8M for the same total makeup H added. A two hour boiling digestion period issolved any fuel particles remaining after all large cladding fragments were dissolved. This digestion also precipitated most of the dissolved silicon as fluffy, white amorphous silica. The dissolution rate of unirradiated fuel was unaffected by .01M KF.

Trace fluoride (0.01M KF) was necessary to provide vigorous and con- sistent dissolution of 307. burnup miniplates. With no fluoride present, the irradiated fuel dissolved slowly and inconsistently. The dissolution rate of irradiated fuel was 36% higher with fluoride.

Analysis by neutron activation and plasma source emission spectroscopy showed that solid residues remaining after dissolution contained negligible uranium (see Table 1). The solid residues from unirradiated fuel appeared to be amorphous silica. The solids from irradiated fuel were darker and more granular in appearance, but again appeared to be precipitated silica.

Table I

Uranium Loss to Dissolution Residues

Composition % Total U Lost

Unirradiated U3Si2 0.07% U3Si 0.12% U3SiAl 0.30%

Irradiated U3Si 0.01%

Clarification Studies

Current SRP clarification steps were successfully tested on dissolved unirradiated and irradiated RERTR silicide fuels.

The dissolved fuel was first treated with a reverse permanganate strike. Digested dissolver solution was brought to 75% and 0.014M Mn(NO3)2 was added. This was followed by sufficient KMnO4 to react wiTh 90% of the Mn+2 present. This reaction is shown below:

3Mn(NO3)2 + 2KMnO4 + 2H20 --- 0- 5MnO2 + 2KNO3 4 HN03

The precipitated MnO2 cake centrifuged easily from the solution in less than five minutes at 1100 G's in a bottle centrifuge.

A gelatin strike successfully removed dissolved silicon to less than 70 ppm. 50 mg/l of gelatin was added to the dissolver solution and held at 75% for 30 minutes. Dissolved fuels initially contained as much as 180 ppm

136 dissolved silicon. A visible quantity of gelatin precipitated silica was observed only three times during these studies. This solid easily centrifuged from the solution in less than 10 minutes at 1100 G's in a bottle centrifuge.

Feed Preparation Demonstrations

Integrated dissolution and clarification was demonstrated with whole REM miniplates. The resulting solutions provided feed for bench-scale solvent extraction studies. An unirradiated miniplate of each silicide composition was dissolved and clarified by the procedures tested above with no problems observed. Two each U3Si and U3SiAl irradiated miniplates were processed in the same way with good results.

Conclusions

These studies demonstrate that spent RERTR silicide fuels can be successfully dissolved and clarified for reprocessing at SRP. Trace fluoride will probably be necessary for dissolution. A gelatin strike will be necessary to ensure satisfactory solvent extraction performance. All dissolving and feed preparation procedures dveloped and tested are within the present technical standards for SRP reprocessing.

References

1. M. L. Hyder, W. C. Perkins, M. C. Thompson, G. A. Burney, E. R. Russell, H. P. Holcomb, and L, F. Landon, Processing of Irradiated, Enriched Uranium Fuels at the Savannah River Plant, USDOE Report DP-1500, E. du Pont de Nemours and Co., Inc., Savannah River Laboratory, Aiken, SC (April 1979).

2. B. E. Paige, G. W. Gibson, and K. L. Rohde, The Effect of Silicon on Fabrication and Reprocessing of Aluminum Alloy Reactor Fuels, USDOE Report IN-1194, Idaho Nuclear Corporation, National Reactor Testing Station, Idaho Falls, ID (November 1968).

3. H. J. Groh, Removal of Silica from Solutions of Nuclear Fuels, USAEC Report DP-293, E. I. du Pont de Nemours Co., Inc., Savannah River Laboratory, Aiken, SC (June 1958).

137