Zirconium carbide, ZrC, is made by reacting Zr02 or zircon with graphite in an electric furnace. It has a very high melting point, 3420°C. It reacts with chlorine at 500°C to produce ZrCV

1.5 Zirconium Nitride

Zirconium nitride, ZrN, is formed when zirconium metal is heated with nitrogen above 1200°C.

1.6 Zirconium Hydrides

The phase diagram for the system zirconium-hydrogen is shown in Fig. 7.3. The maximum hydrogen content of the solid phase approaches that of ZrH2, 6б| a/о H.

The 5 and e phases have been used extensively as moderator in several types of nuclear reactor. The compact SNAP1’ reactors [D2] employed in power plants for space vehicles use as fuel and moderator a mixture of 7 to 10 w/o enriched (90 a/о) HSU with zirconium hydride

1Systems for Nuclear Auxiliary Power.

Figure 7.3 Zirconium-hydrogen phase diagram. (Reprinted with per­mission from Dr. W. M. Mueller [MS] and the copyright holder, Academic Press, Inc., New York.)

containing from 63 to 64.9 a/о hydrogen. The TRIGA research reactor developed by General Atomic [Wl] uses a mixture of uranium, zirconium, and hydrogen in atomic proportions 0.03:1:1. The uranium contains 20 a/o 235U. Zirconium hydride has been suggested as moderator for high-temperature power reactors producing superheated steam.

The desirable properties of zirconium hydride for these systems are (1) the low equilibrium pressure of hydrogen at temperatures up to 650° C, and (2) the high atomic density of hydrogen in them. Because the є phase has a density of 5.62 g/cm3, the atomic density of hydrogen in zirconium hydride containing 64.9 a/о hydrogen is

(5.62 g/cm3X6.02252 X 1023 molecules/g-mol)[(0.649/0.351) atoms H/molecule]

[91.22 + (0.649/0.351)1.008] g/g-mol

= 6.7 X 1022 atoms H/cm3 (7.3)

This hydrogen density in e zirconium hydride is as high as in water at room temperature and is appreciably higher than in water at the 300°C used in power reactors. Another advantage of the uranium-zirconium hydride fuel-and-moderator mixture is its high prompt negative temperature coefficient of reactivity, a consequence of the intimate thermal contact between 235 U and hydrogen atoms.

The left side of Fig. 7.3 shows that hydrogen is appreciably soluble in a zirconium, up to a maximum concentration of 6.1 a/о at 550°C. Dissolved hydrogen reduces the impact strength of zirconium [M4] and has been responsible for the failure of fuel cladding through hydrogen embrittlement.

Zirconium hydride is made by reacting zirconium metal with hydrogen.

Massive zirconium hydride is stable in air at temperatures below 600°C, but finely divided hydride will ignite at 430°C [H2] and should be kept out of contact with air.

For additional information on zirconium hydrides, see reference [В1].

Fuel elements discharged from PWRs also contain radionuclides formed by neutron activation in the zircaloy cladding, stainless steel end fittings, and Inconel spacers. A typical 3-year irradiation of the metallic structure produces the radionuclides listed in Table 8.12, calculated for fuel elements discharged from a LWR and stored for 150 days [B3]. Neutron capture in stable 94 Zr forms 65-day 95 Zr and its decay daughter, 35-day 95 Nb. The radioactivity produced is large, but it is still smaller than the radioactivity of these two nuclides formed as fission products (cf. Table 8.1). Other large contributors to the cladding radioactivity are “Co, resulting from neutron capture in stable 59 Co, and slCr, S5Fe, 58 Co, and 68 Ni.

After 10 years of decay there is still appreciable radioactivity remaining, so irradiated cladding must be treated as a long-lived radioactive waste. The only species that persist after about 1000 years of decay are 1.5 X 106 year 93Zr and 2.12 X 10s year "Tc. The activity of 93Zr in irradiated cladding is about the same as the activity of fission-product 93Zr (cf. Table 8.1), but the activity of "Tc in cladding is about 1000 times less than the activity of fission-product "Tc.

The fast-breeder fuel cladding and structure, typically of 316 stainless steel, result in the radionuclides listed in Table 8.12 [B3]. Because the structure is entirely an austenitic alloy, the most radioactive nuclides are 54Mn, ssFe, and “Co.

Fuel cladding hulls will also contain uranium, plutonium, and other transuranic radio-

nuclides as contaminants on the inner surfaces of the cladding. These transuranics can be removed by chemical treatment of the cladding-hull surfaces, or the cladding hulls can be classified as transuranic wastes.

The HTGR fuel contains no metallic structure, but impurities in the graphite fuel blocks result in the production of relatively small amounts of radioactive cobalt and nickel, as listed in Table 8.12 [HI, P3]. The total activity from metallic contaminants in HTGR fuel is considerably lower than that in the fuels from light-water and breeder reactors.

At about the same time that the Redox process was being developed in the United States, a group of Canadian chemists [Cl] at the Chalk River Laboratory were developing the Trigly

process, for recovering plutonium from natural uranium irradiated in the NRX reactor. This process used triglycol dichloride (C1C2 H4 OC2 H4 OC2 H4 Cl) as solvent and nitric acid and ammonium nitrate as salting agents. Hexavalent plutonium has a higher distribution coefficient than uranium in this solvent; seven batch extractions of the aqueous phase, each with one-fourth its volume of trigjy, recovered 97 percent of the plutonium, 5 percent of the uranium, and only 0.01 percent of the fission products. Further purification was with hexone as in the Redox process.

The equilibrium distribution of thorium nitrate and nitric acid between their solutions in water and in 30 v/o TBP in a hydrocarbon diluent at temperatures between 30 and 60°C has been reported by Siddall [S13] and Weinberger et al. [W6]. Siddall’s diluent was Ultrasene, a mixture of normal, iso-, and cycloparaffins with an average molecular weight of 175. Weinberger et al. used practical-grade n-dodecane, molecular weight 170. Rainey and Watson [R4] modified the SEPHIS computer program to represent the distribution coefficients of nitric acid and thorium nitrate between an aqueous phase and 30 v/o TBP. Figures 10.24 and 10.25 display the distribution coefficients predicted by the 1978 version of the SEPHIS code [VI] for equilibria at 30°C. Distribution coefficients agree with measurements of Siddall [S13] and Weinberger et al. [W6] except for thorium at aqueous concentrations below 0.06 M within the dashed line, where the code predicts lower values than observed.

Adequate data on distribution coefficients of uranyl nitrate between 30 v/o TBP and aqueous solutions of thorium nitrate and nitric acid are not available. Examination of concentrations of coexistent phases in Thorex process mixer-settler runs reported in references [Rll], [01], and [02] indicate that the distribution coefficient of uranium Du when present at uranium concentrations below 0.02 M in Thorex systems at thorium concentrations above 0.1 M is given approximately by

Du = 20DTh (10.15)

Source: L. Kiichler, L. Schafer, and B. Wojtech, “The Thorex Two-Stage Process for Reprocessing Thorium Reactor Fuel with High Burnup,” Kemtechnik 13: 319 (1971).

However, column separation performance in the Hanford Thorex campaign correlates better with a Z>u/Z>Th ratio of 14 (Prob. 10.5). Because the distribution coefficient of thorium is so much less than that of uranium, the Thorex process requires a much higher organic/aqueous flow ratio than the Purex process.

Figure 10.26 compares the low-concentration distribution coefficients of uranium, thorium, plutonium, protactinium, and the principal fission products. The spread between thorium and fission-product zirconium is greatest between 1 and 2 M HN03, the range used in the decontamination step of the acid Thorex process. Because the distribution coefficient of protactinium is close to that of thorium, it is necessary to remove protactinium or complex it with fluoride or phosphate ion to prevent its extraction with thorium.

A complication of the Thorex process is appearance of a second organic phase at high concentrations of thorium nitrate and nitric acid. To obtain reproducible separation, Thorex process systems are designed to stay below the thorium concentrations at which the second organic phase forms. Figure 10.27 shows these conditions for п-dodecane diluent and Ultrasene at 30°C. Siddall [SI4] has pointed out that substitution of triamyl phosphate for TBP would essentially eliminate formation of a second organic phase with thorium.

Next Page

FUEL REPROCESSING 527

In terms of special radioactive waste three radionuclides will be discussed, which are collected separately in the reprocessing plant: tritium, 1291, and мКг. 14C, as mentioned before, is presently not considered waste in the sense that attempts are made to develop techniques for recovery and final disposal.

The calculation of the concentration of extractable components in a countercurrent cascade of equilibrium solvent extraction stages is first developed for the simple countercurrent extraction section of Fig. 4.3. The theory is then extended to the extracting-scrubbing system of Fig. 4.4 for fractional extraction and is illustrated by a numerical calculation for the separation of zirconium from hafnium, using TBP in kerosene as solvent.

0.1 Extracting Cascade

Here we consider an extracting cascade in which a feed solution containing one or more extractable components is contacted countercurrently with an organic solvent. Nomenclature for flow rates, concentrations, and stage numbers is shown in Fig. 4.8. It will be assumed that equilibrium is reached between the aqueous and organic phases leaving each stage. Changes in the volume flow rates of the aqueous and organic phases will be neglected. Consider the portion of the cascade below stage n. A material balance on one of the extractable components is

Ey0 + Fxn = Eyn_ , + Fxi

F. .

Уп — і — Уо = j(x„ — Xi)

Concentrations in the organic and aqueous phases leaving a stage are related by the equilibrium relation

Уп Dnxn

where D„ is the distribution coefficient at the conditions of the nth stage.

The meaning of Eqs. (4.27) and (4.28) may be visualized on a plot of у versus x, in the McCabe-Thiele diagram, Fig. 4.9. The material-balance relation (4.27) is represented by the operating line that passes through the point (xf, jf) and has the slope F/E. The equilibrium relation (4.28) is represented by the equilibrium line. When D is constant, the equilibrium line is a straight line, as would occur for the extraction of trace quantities of solutes in the presence of nonextractable salting agents, with constant concentration of uncombined complexing agent. More generally, as has been demonstrated in Sec. 4, Dn varies from stage to stage, resulting in a curved equilibrium line. Figure 4.9 illustrates the equilibrium line typical for the extraction of a single component in the presence of a nonextractable salting agent.

The McCabe-Thiele diagram is useful for constructing a graphic solution for the stagewise compositions. The operating line is the locus of points xn, yn_j of adjacent interstage flows. The vertical projection of any such point intersects the equilibrium line at x„, yn, thereby defining the compositions of the aqueous and organic phases leaving the equilibrium stage n.

Assume that the cascade is to reduce the concentration of the extractable component from xF to xі by extraction with organic of relative volume E/F. The point Xj, y0 is thereby specified. Beginning at x1; ya and projecting upward in vertical and horizontal steps, the compositions for all of the other equilibrium stages are determined. The number of vertical projections between the operating line and equilibrium line necessary to step from xx to x*7 gives the required number of equilibrium stages.

Given this number of stages, construction of a similar McCabe-Thiele diagram for other components in the feed, such as impurities, allows the calculation of the extent to which these impurities extract into the organic phase. If two or more extractable components are each in sufficient concentration to affect the distribution coefficient of the other species, e. g., TBP extraction of U02(N03)2 and HN03, the equilibrium lines for the two components cannot be specified in advance but must be calculated by an iterative procedure, similar to that to be illustrated in Sec. 6.6 for the zirconium-hafnium separation.

From Eq. (4.27) or from the construction of Fig. 4.9 it is apparent that the ratio of organic flow rate to aqueous flow rate is given by

I = (4.29)

F Ум~У 0

because xF is the virtual aqueous effluent concentration from stage N + 1. If the overall fractional recovery p of the extractable component is specified as

For given compositions Xі7, Xi, and y0, or for given Xе, y0, and p, reducing the relative amount of organic flow brings the operating line nearer to the equilibrium line and increases

[a) [b)

the required number of stages. The minimum flow ratio (E/F)m-m occurs when the operating line intersects the equilibrium line at xF, requiring an infinite number of stages as illustrated in Fig. 4.10b. In this limiting condition, denoted by asterisks, the concentration y*. in the organic stream leaving the cascade is

so that (4.29) becomes

and (4.32) becomes

If the equilibrium line is locally concave upward, as is possible in the extraction of a self-salting component with excess extracting agent, with sufficiently low xt the operating line may intersect the equilibrium at x <rF. In this event Eqs. (4.33) through (4.35) are invalid unless x! is increased to allow intersection at xF, as illustrated in Fig. 4.10h.

To carry out a specified separation in an actual extracting cascade with a finite number of stages, the flow ratio E/F must be greater than the minimum ratio E*/F given in Eq. (4.34) or (4.35). The application of these equations for the case of constant distribution coefficients will be illustrated in Sec. 6.2.

In addition to the well-characterized minerals containing uranium listed in Table 5.15, uranium occurs as a minor constituent of many other materials, some of which have been used as commercial sources. Table 5.16 lists low-grade sources of uranium and gives the range of their uranium content. Uranium from the five commercial sources is being produced as a by-product or co-product of other materials whose value helps pay for the cost of producing the uranium.

Some lignites contain sufficient uranium so that they can be mined for the uranium alone; most are so lean that they must be used as fuel as well as a source of uranium.

Swedish shales contain 300,000 MT+ of recoverable uranium, together with organic matter from which fuel can be made and pyrites that can be converted to sulfuric acid. The combined value of these products makes the operation economic.

Tailings from many South African gold mines contain sufficient uranium to permit its recovery at competitive costs, because the cost of mining and crushing the ore has been paid for by the gold previously extracted. Most South African uranium is produced in this way.

In the wet process for converting phosphate rock to fertilizer, the rock is first converted to phosphoric acid. In this process around 95 percent of the uranium goes into solution, and around 85 percent could be recovered if a uranium extraction step were added. Some uranium is now being produced from this source. A total of 65,000 MT of uranium could be recovered in this way by the year 2000 in the United States.

Pilot-plant studies confirm the feasibility of recovering uranium from the sulfuric acid solutions used to extract copper from tailing piles in the Western United States. One plant producing around 250 MT/year went into operation at Twin Buttes, Nevada, in 1975. It is estimated that total U. S. production from copper tailings might reach 800 MT/year.

Uranium produced from the uneconomic sources listed in Table 5.16 would cost several hundred dollars per pound and is not economic at present. If the uranium-fueled fast-breeder reactor becomes economic, it would generate so much electricity per ton of natural uranium that Chattanooga shale and even Conway granite might be used as economic uranium sources. These sources are estimated to contain 5 million and 6 to 9 million MT of uranium, respectively. Environmental problems from the large amount of earth disturbed in mining these low-grade sources would be severe.

Although seawater contains only 3.34 д/g of uranium/liter, the oceans of the world are so vast that their total uranium content is estimated to be around 4 billion MT [Dl]. Extraction of uranium from seawater is discussed in Sec. 8.8.

Thorium forms relatively stable tetravalent salts of many of the oxyacids. These can be prepared by reacting thorium hydroxide or basic carbonate with the appropriate acid.

Thorium nitrate is very soluble in water, to the extent of 65.6 g Th(N03)4/100 g solution at 20°C. It can be crystallized from solution as the nominal tetrahydrate. Thorium nitrate solutions are used for purifying thorium by solvent extraction, Sec. 9.#

Although anhydrous thorium sulfate dissolves in water at 0°C to the extent of 20 w/o, the solution is metastable and deposits hydrates on standing. Stable solutions at higher temperature require the presence of free sulfuric acid, as in solutions used to leach thorium minerals, Sec. 8.5.

A solution of ThOCl2 is produced when ThCl4 reacts with water. Evaporation to dryness produces a succession of ill-defined hydrates that can be converted to anhydrous ThOCl2 by heating to 250°C.

Hydrated thorium fluoride is precipitated when a soluble fluoride is added to a solution of thorium nitrate. Precipitation can be prevented by addition of aluminum nitrate to complex the fluoride ion, an expedient used in the Thorex process (Chap. 10, Sec. 5).

Source: International Atomic Energy Agency, “Thorium: Physico­chemical Properties of Its Compounds and Alloys,” Atomic Energy Rev., Special Issue No. 5, 1975.

Thorium hydroxide Th(0H)4 is precipitated from solutions of thorium salts by addition of alkali hydroxides.

Thirty percent hydrogen peroxide precipitates thorium peroxide Th207 from solutions of thorium salts. As few cations other than thorium and uranium precipitate under these conditions, this method has been used to purify thorium.

Thorium orthophosphate Th3(P04)4 is precipitated by phosphate ion from neutral or slightly acid solutions of thorium nitrate or sulfate. It is soluble in concentrated phosphoric or sulfuric acid, such as is present when monazite is dissolved in sulfuric acid.

Addition of an alkali carbonate to an aqueous solution of a thorium salt first precipitates a basic thorium carbonate of variable composition. like uranyl carbonate, thorium carbonate dissolves in an excess of alkali carbonate, in this case forming the complex ion [Th(C03 )4 (OH) 2 ] * ".

Thorium oxalate is precipitated when a solution of oxalic acid is added to a solution of a thorium salt. This is a commonly used intermediate step in producing thorium dioxide from thorium nitrate solution (Sec. 10.1). Quantitative precipitation of thorium from solutions up to

1.8 N in nitric acid can be obtained by use of five times the stoichiometric amount of oxalic acid [A2]. Thorium oxalate can be dissolved by concentrated nitric acid or by sodium oxalate solution, with which it forms a double oxalate [B7].

8.1 Difficulties

The high melting point of these metals and their reactivity make their production in pure form very difficult. They form oxides, hydrides, nitrides, and carbides that are soluble in the metal, diffuse through it, and make it hard and brittle even at concentrations of a few tenths of a percent. The oxides are especially stable; once the metal has been contaminated by oxygen, no reducing agent can remove it completely. The metals react with air or nitrogen at temperatures above 300°C and, when finely divided, react with water even at room temperature. Conse­quently, they must be protected by helium, argon, or a vacuum during high-temperature reduction operations, casting, or hot forming, and finely divided metal cannot be cleaned by washing with water or aqueous solutions. The molten metal reacts with all known refractories, even graphite or lime, which can be used for uranium.

All these difficulties require that chemicals for producing these metals be exhaustively purified, especially of oxygen, water, and nitrogen, and limit the number of processes that can be used.

8.2 Available Processes

The principal processes that have been used for producing zirconium and hafnium metal of the requisite purity are as follows:

1. The Kroll process, involving reduction of tetrachloride vapor by molten magnesium

2. The hot-wire process, involving thermal decomposition of the iodide

3. Electrolysis of the double potassium fluoride dissolved in fused salts

In aqueous solutions trivalent protactinium is unknown. Tetravalent protactinium is stable in the absence of air, but it is rapidly oxidized to Pa(V) by oxygen. It can be prepared from Pa(V) by using strong reducing agents such as zinc dust, amalgamated zinc, or Cr(II) salts, or by electrolytic reduction. In moderately acidic (H+ < 1 M) aqueous solutions Pa(IV) appears to exist as the protactinyl ion, Pa02+ or Pa(OH)22+, which forms moderately stable complexes [L2].

Pentavalent protactinium is the most stable oxidation state in aqueous solution. It shows a strong tendency toward irreversible hydrolysis in solution, but it differs from the other pentavalent actinides in that it does not hydrolyze to form an actinyl ion of the form M02*. Instead, the postulated ionic species in noncomplexing solutions are PaOOH2+ and PaO(OH)2+. With strong complexing agents, even in aqueous solutions, Pa(V) can form a nonoxygenated complex, such as PaF6 ‘ [A1, L6].

The distribution coefficients of Pa(V) between nitric acid solutions and solvents containing TBP are less than those of uranium [C6], and are less than those of thorium except at high concentrations of HN03 [H2]. In extraction measurements it is found that the fraction of Pa(V) that can be extracted decreases with time, due evidently to the slow polymerization of Pa(V) colloids. The more highly condensed forms cannot be depolymerized by acid treatment [К2].

Protactinium can be recovered from irradiated thorium, after fission-product decontamina­tion, by exchange onto Dowex 1-X8 anion-exchange resin from a 9 M HC1 solution. Thorium is then eluted with 9 M HC1, followed by elution of Pa(V) with 9 M HC1-0.25 M HF. Uranium is eluted with 0.25 M HF [Kl, K4].

Protactinium can be recovered from an aqueous nitrate solution of fission products and protactinium by adding sodium chromate, which brings down protactinium on the aluminum chromate precipitate. After dissolution of the precipitate in acid, protactinium may be recovered by solvent extraction, or it may be allowed to decay to 233 U, which is more easily extracted [G4]. Protactinium can also be recovered by adsorption on powdered Vycor glass.

Specific extractants for Pa(V) from fairly concentrated (6 M) HC1 solutions are branched-chain ketones and alcohols such as diisobutyl carbinol or diisobutyl ketone [H2, K4].