CHAPTER 20 COMPOSITION AND PROPERTIES OF LIQUID-METAL FUELS* 20-1. Core Furn ComrosiTioN In Chapter 18, the advantages and disadvantages of liquid metal fuels were discussed in a general way. The point was made that a liquid-metal fuel has no theoretical limitation of burnup, suffers no radiation damage, and is easily handled for fission-product poison removal. In this chapter, the results of research and development on various liquid-metal fuels are presented. This work has been largely concentrated on uranium dissolved in bismuth. At the contemplated operating temperatures of approximately 500°C, it was found that uranium has adequate solubility in bismuth when present by itself. However, as the work progressed, it soon became evident that other materials would have to be added to the solution in order to obtain a usable fuel. The present fuel system contains uranium as the fuel, zir- conium as a corrosion inhibitor, and magnesium as an oxygen getter. An LMI'R operating on the contemplated Th**? to U?* breeding cycle an be designed with an initial U3 concentration of 700 to 1000 ppm in bismuth. The actual figure, of course, 18 dependent upon the specifie de- sign and materials used. In Chapter 24, in the design studies, such figures are given. The concentrations of zirconium and magnesium are each ap- proximately 300 ppm. It is contemplated that these concentrations will have to be varied depending upon desired operating conditions. In their use as corrosion inhibitor and antioxidant there is enough leeway for this purpose., The fuel deseribed in the previous paragraph is the clean fuel which would be charged initially. During reactor operation, however, fission products will build up in the fuel and would be maintained at a level dictated by the economics of the chemical reprocessing system used. It has been found that the fission products and other additives to the bismuth have an im- portant effect on the solubility of uranium in bismuth. These have been arefully investigated in order to permit selection of reactor temperatures that will ensure that all the uranium remains in solution during reactor operation. Likewise, the solubility of steel corrosion products has been in- vestigated to determine their effect on uranium solubility in bismuth. *Based on contributions by D. H. Gurinsky, D. G. Schweitzer, J. R. Weeks, J. 8. Bryner, M., B. Brodsky, C. J. Klamut, J. G. Y. Chow, . A. Meyer, R. Bour- deau, and O. F. Kammerer, Brookhaven National Laboratory. 722 20-2] SOLUBILITIES IN BISMUTH 723 635° C 560 496 441 394 352 315 5.0 ' I [ | I I [ K - 1.0 = B2 - § 5 }I: > 10 & 05 L | I | | | | 1.0 i 1.2 1.3 1.4 1.5 1.6 1.7 1000/7 K Fra. 20-1. Solubility of uranium in bismuth. It 1= important to note that although the basie fuel is a simple one, the uranium used for liquid metal fuel reactors using the Th—U?233 cycle must be almost completely enriched 233 or 235 in the initial charge. Ifurther, since the concentrations are measured in parts per million by weight, it is not an easy matter to maintain a striet accounting of all fuel. When deal- g with such small amounts, losses due to reaction of uranium with carbon and adsorption of uranium on steel and graphite walls can be sig- mificant, The fuel for the LMFR is still under extensive study. At present, most of the major information for the design of an LMFR experiment is at hand. Thix information is primarily solubility data and other fuel information, presented in the following pages. 20-2. SOLUBILITIES IN BIisMUTH 20-2.1 Uranium. The experimental techniques used to measure solu- bilities in liquid bismuth have been described previously [1,2]. Several workers [3-7] have investigated the solubility of uranium and bismuth. Recently, with improvements in analytical techniques, redetermination of the solubility curve has been undertaken. The latest results are at variance with the older work of Bareis [5], as shown In I'ig. 20-1. It can be seen that the recent data obtained at Brookhaven National Laboratory are, at some temperatures, as much as 20 to 259, lower than those obtained some vears ago. This variance in solubility determinations may be due to several factors, but 1t 1= believed that the improved techniques are more reliable, and that the newer values are consequently more precise. The presence of such 724 PROPERTIES OF LIQUID-METAL FUELS [cHAP. 20 Temperature, °C 1200 1000 8GO 700 600 500 400 300 IR ' 10.0 o S Solubility, Wt % o o 010 /T K x 104 Fig. 20-2. Solubility of uranium and thorium in bismuth. 727 636 560 494 441 394 352 315 283 10% | ] [ [ B — ].05/0 [—— — = 2 = ; Los Alamos S01% - > a . 01% | 1 | | 1 | { 1 1 0% 110 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 1000/T °K Fra. 20-3. Solubility of plutonium and uranium in bismuth. 20-2] SOLUBILITIES IN BISMUTH 725 other materiils as nickel, copper, manganese, ete., in the bismuth in quan- tities large enough to affect the uranium solubility still remains to be in- vestigated. For example, nickel has been shown to markedly reduce the uranium =olubility in hismuth [1]. It 1= obvious that even slight variations of the solubility of uranium in bismuth nmight be of considerable importance in LMI'R reactor design. The solubility of uranium, according to the preferred data (the solid curve i g, 20-1), allows a rather small leeway in uranium concentration in the reactor cycle when the lowest temperature of 400°C in the heat exchangers 1s taken into account. 20-2.2 Thorium and plutonium. The solubility of thorium in bismuth, as determined by Bryner, is compared with the solubility of uranium in Iig. 20-2. In the temperature range 100 to 500°C, the solubility of thorium 1= markedly lower than that of uranium. In fact, it is =0 low that a breeding cvele using only thorium in solution with bismuth cannot be arried out, To fill out the mformation on fissionable fuel solubility in bismuth, Fig. 203 shows the solubility of plutonium in bismuth, as determined at the Los Akunes National Laboratory. In comparing plutonium with uranium, it 1= =een that plutonium is significantly more soluble. 20-2.3 Fission-product solubility. The solubilities of most of the impor- tant fis 2Nt T o B Aee Cae e - "R i * Lo Tl e Tl T N R N Fra. 20-13. 5 w/o Th-95 w/o Bi. Dispersion of ThBis platelets in Bi. Alloy heated to 1000°C and quenched by pouring into graphite crucible at 25°C. (150x) - L.y S Fic. 20-14. 10 w/0 Th-90 w/o Bi. Dispersion of reconstituted ThBi2 particles in Bi. Produced by heating fine-platelet dispersion to 800°C for 20 min. (150%) ultrasonic energy to platelet dispersions causes the platelets to break up into essentially equiaxed fragments. 20-6.5 Engineering studies of slurries. The intermetallic compound ThBig is quite soft, having a Rockwell 15-T hardness of approximately 60 at room temperature. It is brittle at room temperature but appears to exhibit some ductility at 400°C. The compound is pyrophoric and must be protected against oxidation. When slurries of equiaxed bismuthide in bismuth are prepared, they are fluid at temperatures above the melting point of bismuth, 271°C. In these slurries the solid phase is in thermodynamic equilibrium with the 20-6] THORIUM BISMUTHIDE BLANKET SLURRY 739 liquid phase and is perfectly wetted by it. At the proposed reactor tem- peratures (350 to 550°C) practically all the thorium in the slurry appears in the solid phase, since the solubility in the liquid is very low. The ideal slurry composition represents a balance between a desire for a high thermal neutron utilization factor (i.e., a high thorium content) and the necessity for high fluidity. IFluidity studies have shown that the upper limit of thorium concentration for high fluidity at reactor temperatures is approximately 10 w/o of thorium. This corresponds to 24.9% by volume of ThBis2, and a thermal neutron utilization factor of 0.957. Although the viscosity of Th-Bi slurries has not been measured, ealculations based on the viscosity of liquid bismuth and the behavior of similar systems indicate that at 530°C the viscosity of a 10 w/o0 Th suspension of 50-micron, equi- axed ThBi. particles should be approximately 2.5 centipoises. It has been observed that increasing the thorium content beyond 10 w/o Th causes a disproportionately large increase in the viscosity, so that the consistency approaches that of a mud or paste. The maximum thorium concentration for high fluidity decreases when the ThBi» particle shape departs signifi- cantly from an equiaxed shape. The density of liquid bismuth varies from 9.97 at 350°C to 9.72 at 550°C, and should not be changed appreciably by the small amount of thorium dix=olved at these temperatures. Therefore the solid particles should sink in the liquid. Although settling rates have not been measured, the mag- nitude of expected settling rates can be caleulated. The setiling rate for 100-micron spheres at 330°C, as caleulated by Stokes’” Law, is 0.030 fps. The settling rate in a 10 w/o Th-B1 dispersion of 100-micron spheres at 550°C, us ealeulated by the hindered settling equation, 1s 0.0026 fps. It has been observed in small systems that equiaxed ThBia particles . This suggested that mass transfer of the steel may have been instrumental in starting the ThBis deposition, perhaps by roughening the walls or perhaps by altering the composition of the tube surface. Specimens of 5 w/o Th slurries have been cycled for 300 hr between 350 and 580°C in graphite tubes with no evidence of plug formation. In these experiments, a relatively rapid increase in ThBis particle size (from 50 to 225 microns in 500 hr) was observed. This increase was due to particle agglomeration rather than growth of single erystals. No evidence of graph- 1te erosion was observed. Specimens of slurries containing 10 w/o thorium and 0.10 w/o tellurium have been cyceled between 350 and 580°C in graphite, and between 350 and 550°C in 239, Cr-19, Mo steel for 500 hr with no evidence of ThBiy plug formation or mass transfer of the steel. The speeimens showed no inerease in the maximum particle dimension and no particle agglomeration. When a specimen of slurry containing 10 w/o Th, 0.10 w/o Te was cycled at higher temperatures in a 239 Cr-19, Mo stecl tube, mass transfer of steel and deposition of ThBis in the cooler end were observed after less than 100 hr. Slurries containing up to 7 w/o Th and minor additions of zirconium have been circulated through small 239, Cr-19,, Mo steel loops by means of a propeller pump. Isothermal circulation at 450°C has been carried out for more than 450 hr at velocities between 0.3 and 1.5 fps, with no difficulty in circulation or maintaining suspension. Attempts to circulate these slurries through a temperature differential, however, have resulted in the formation of ThBiz> deposits in the coldest section of the loop. In a modified loop containing a graphite liner in the finned-cooler section, 1sothermal cireulation was maintained without difficulty. ThBiz, however, again deposited in the finned-cooler section when a temperature differential was applied. When a slurry containing 7 w/o Th. 0.025 w/0 Zr, and 0.10 w/0 Te was 20-7] THORIUM COMPOUND SLURRIES 741 circulated in a 219, Cr-19,, Mo steel loop through a temperature differ- ential, ThI3iz deposited in the finned-cooler section. The rate of buildup of the deposit was markedly less than in the case of slurries containing no tellurtum, The problem of ThBis deposition during circulation through a tempera- ture differential is one which must be solved before the Th-Bi slurry is aceeptable as a {luid breeder-blanket material. The favorable results ob- tained by tellurium additions in the capsule experiments offer hope that the problem can be solved. 20~7. Tuoritm CoMPOUND SLURRIES 20-7.1 Thorium oxide. Probably the best blanket material, next to the thorium bismuthide slurry, 1s the suspension of thorium oxide in bismuth. The thorium-oxide shurry should be compatible with the graphite and steel in the reactor structure. Iixperiments have shown that ThOs is wetted by the liquid bismuth if some zirconium or thorium iz dissolved i the bis- muth. =lurries of 10 w/o thorium oxide have been prepared. The separation of thorium oxide from the liquid bismuth for processing could be achieved by mechanical means, and the oxide could then be proceszed by the existing Thorex process. The thorium-oxide hlanket slurry is gaining increased attention. A loop of several pounds per minute capacity has been completed for forced cireulation of the oxide slurries at BNL and an 800 Ib/min loop is ready at Babeock' & Wilcox. 20-7.2 Other thorium compounds. A small amount of attention has been directed toward ThCs, ThS, and Thl'y slurries in bismuth, How- ever, the major effort is on the thorium bismuthide and thorium-oxide