To carry out a specified separation in an actual cascade with a finite number of stages, S/E must be greater than the minimum ratio given by (4.69) and (F + S)/F must be less than the maximum ratio given by (4.70). For such an actual cascade it is important to know the number of extracting stages N and stripping stages M needed to effect a specified separation with given values of these flow ratios. When the distribution coefficients of each of the species to be separated are not affected by the concentration of the other extractable species, the number of
stages can be calculated either by the graphic procedure illustrated in Fig. 4.15 or, for constant distribution coefficients, by means of the equations to be derived in this section. When the distribution coefficients of each of the extractable species are affected by the concentrations of the other extractable species, it is usually preferable to use the numerical, stage-to-stage calculation procedure to be described in Sec. 6.6. The graphic procedure is usually not useful under these conditions because the location of the equilibrium lines cannot be set in advance of the calculation of the stagewise concentrations of extractable species throughout the cascade.

When D is constant, equations for the number of stages can be outlined by applying the Kremser equation (4.45) derived in Sec. 6.2. For the extracting section the appropriate flow ratio is E/(F + S), and the extraction factor [}E is

~ F + S i4-72)

so that, from (4.45), the concentration in the organic leaving the extracting section is

У% =-jr—T (Dxf -Уо) + Уо PE “ 1

The extraction factor f$s in the scrubbing section is

(4.75)

Although derived for an extracting section, the Kremser equation (4.45) can be applied to the scrubbing section by the following transformation:

x[ = xsM )

Уо = Ум+1

y, N=ySi

where the primed quantities are to be substituted in Eq. (4.45). This substitution results, with some rearrangement, in

To eliminate xfj, a material balance is written around the scrubbing section:

S(*M — 4) = E(y$f+1 — ysx)

Here we assume that x$ = 0 and combine (4.78) with (4.75) to obtain

By applying the continuity Eq. (4.53) for a cascade with aqueous feed, (4.74) and (4.79) are equated to result in

1^ y’ = (Dx? "} + ^ (4-80)

There is one equation of the form of (4.80) for each of the extractable components. When flow ratios S/E and E/F have been specified, and when any three of the four terminal concentrations xf, y§, yf, have been specified for two components, values of (3E and f°r the two components can be calculated. Two equations of the form of (4.80) can be written, one for each component. These can then be solved for the number of stages N and M in the extracting and scrubbing sections, respectively.

Scrub

3.5 Af NaN03 *^0“З Af HNO3 S-48 liters

Feed

3.5 Af NaN03 *£-3AfHN03 *£-0.123 AfZr *£-0.00246 Af Hf F-48 liters

Aqueous raffinate

3.5 Af NaN03 *нд-3 Af HNO3 x£ j-0.00123 Af Zr *£n-0.00122 Af Hf &+F-96 liters

The above equations can be transposed to relatively simple equations for the recovery p of an extractable component and the decontamination factor /, as defined by Eqs. (4.64) and (4.65), respectively. To eliminate Xе from (4.64), an overall material balance is written:

Лг’7 = (S + F)xf + Efrf -y£)

Combining (4.64) and (4.81):

=_______ A________

p (M/fe) + yf-yf

Now assume, for simplicity, that yf = 0, so that Eq. (4.80) becomes

(4.83)

Substituting (4.83) in (4.82):

________ feKfig — i)/(fe -1)1______________

O/fls ЖйГ1 — i)/(& — i)l + fe[G$ — i)/(fe -1)]

There is one equation of the form of (4.84) for each extractable component. The values of (3 for each component are determined by the value of the distribution coefficient for that component and the flow ratios. If the recovery pA of one component is given and the overall decontamination fAB for two extractable components is specified, the recovery pB of the second component is obtained from (4.65). Thus, for known pA and pB and known /3’s Eq. (4.84) can be solved twice to obtain the necessary numbers M and N of equilibrium stages.

As an example of the use of these equations for an extracting-scrubbing cascade, consider the addition of a scrubbing section to the hafnium-zirconium separation example, which was first analyzed in Sec. 6.2 as a simple extraction problem. The modified flow sheet is shown in Fig. 4.17, and desired recoveries and decontamination are given in Table 4.7.

As noted in Sec. 6.2, and as will be demonstrated in Sec. 6.6, actual distribution coefficients will depend on concentration. Constant distribution coefficients are assumed for the purpose of this illustration, the same as used in the minimum-flow-ratio example of Sec. 6.4. Top and bottom concentrations are as follows.

Organic extract:

Zirconium: ysZl i = -^ x£r = (0.98) (0.48)(0.123) = 0.0578

* E

S + F

x§M = FX*fc ~ = (0.5) (0.00246) — (1.04) (0.00000578) = 0.00122

* S "Ь г

With the above quantities the extraction factors are as follows.

Zirconium:

Ae. Zr = Iff = (1.20) (1.04) = 1.248

fczr=^=h20 = ZS0

5,Zr S 0.48
Hafnium:

ftr. Hf = = (0.12) (1.04) = 0.1248

a _ £>HfE _ 0.12 _ n

fe. Hf — 0.250

Equation (4.80) written for hafnium is

°мГ’ ~ 1 m 0-00000578 = T 1 (0.12)(0.00122)

0.25M+1 — 0.25" 0.1248 — 1 v ‘

or (4.0)"+1 — 1 = 86.8 [1 — (0.1248)^]

Since (0.40 )"+1 and (0.1248)^ are small relative to unity, approximate solutions of these equations are

In 17.2 _ 2.84 * In 1.248 0.221

Although slightly more accurate values could be obtained by using these results for a second approximate solution, it is evident that terms that have been neglected would not change the result appreciably.

The McCabe-Thiele diagrams of Fig. 4.18 for the zirconium-hafnium separation example illustrate the principles of the extracting-scrubbing separation system. In the extracting section the flow ratios are such that there are large changes in concentration of the more extractable component (Zr), but relatively little change in the aqueous concentration of the less extractable component (Hf). In the scrubbing section the flow ratios are such that the larger changes in concentration occur for the less extractable component (Hf).

Because of the great variety of natural sources of uranium, no one process is uniquely suited to concentration of uranium from all ores. General methods that have been found useful for certain types of ores will be discussed first and then a few examples will be given of specific processes used for specific ores. Detailed accounts of uranium-concentration processes developed in the United States have been given by Marvin et al. [Ml], Gegg and Foley [Cl], and Merritt [М3].

Gravity concentration. Only the richest pitchblende ores can be concentrated by the specific gravity methods commonly used for other metals. Pitchblende particles are much denser than accompanying nonuraniferous rock (called gangue). In rich ores pitchblende particles are large enough to be concentrated by gravity methods such as were used at Shinkolobwe. Gravity methods are seldom used elsewhere because few other uranium ores have uranium-bearing particles large enough for selective concentration by this method. Most ores contain secondary minerals such as autunite or carnotite, which are soft and form such small particles as to be unrecoverable by gravity methods.

Flotation. Flotation, which is an extremely selective method of concentrating ores of many common elements, is seldom applicable to uranium because few uranium minerals can be selectively floated.

Leaching. Because of the unsuitability of physical methods such as specific gravity or flotation separation, the extraction of uranium from its ores is effected almost exclusively by chemical leaching. The principal leaching reagents are sulfuric acid or alkali carbonate solution. The type of leaching agent depends on the nature of the uranium mineral and of the gangue. When the gangue is silica or some other material insoluble in acid, sulfuric acid leaching is preferred, because it costs less than carbonate and dissolves uranium minerals more rapidly. When the uranium is combined as a silicate or titanate, as in brannerite, it is necessary to leach for longer times, or at higher temperatures, or in stronger acid than when it is more readily dissolved, as in carnotite or autunite. Add consumption for U. S. ores treated for uranium recovery only is from 40 to 120 lb/t+ of ore. When vanadium is also extracted, acid consumption may reach 300 lb/t. Leaching of ores containing tetravalent uranium requires addition of an oxidant to convert uranium to the soluble hexavalent condition. Manganese dioxide or sodium chlorate are the principal oxidants. Iron must be present in solution to catalyze the oxidation.

ft = short ton, 2000 lb.

When the gangue is limestone or some other rock that consumes acid, leaching with sodium or ammonium carbonate is preferred, to reduce chemical consumption and produce a cleaner solution containing lower concentrations of impurities than when acid leaching is used. The reactions that occur in carbonate leaching are

U02 + j02 -+U03

U03 + Na2 C03 + 2NaHC03 -> Na4U02(C03)3 +H20

As carbonate leaching is slower than acid leaching and carbonate solutions do not attack and penetrate gangue particles, it is usually necessary to grind the ore finer and to leach for longer times and at higher temperatures when leaching with carbonate solutions than with sulfuric acid.

Recovery of uranium from leach liquors. Uranium may be recovered from leach liquors by precipitation, ion exchange, or solvent extraction. Precipitation with sodium hydroxide was the recovery method used in the first uranium mills. When used on sodium carbonate leach liquors, the uranium precipitate is fairly free of other metallic contaminants, because sodium carbonate dissolves few other metals beside uranium. However, when used in sulfuric acid leach liquors, the uranium precipitate contains other metals, such as iron dissolved from the ore by the acid, and is no longer commercially acceptable. Consequently, in the United States, uranium mills employing acid leaching now follow it with selective recovery by either solvent extraction or ion exchange. These processes are described in Secs. 8.5 and 8.6, respectively.

Table 6.15 summarizes two sources of information on the annual rate of thorium production, by country. The first three columns give the production rate of monazite concentrates for the more recent years of 1976, 1977, and 1978 [El], We have estimated total thorium production from a typical monazite thorium content of 6 weight percent (w/o). These columns do not include monazite production in the United States or Soviet Union, nor the small production of other thorium minerals. The last two columns give the U. S. Bureau of Mines figures [Ul] for total thorium production in 1973 and an estimate of total thorium production capacity in 1980, if demand were such as to support it.

U. S. consumption has averaged about 70 MT thorium/year for all uses, nonnuclear included.

1.11 Thorium Requirements

Thorium makeup requirements for one reactor system, the HTGR (high-temperature gas-cooled reactor), may be estimated from Fig. 3.33. A 1000-MWe HTGR requires 7.4 MT of thorium as feed pet year. Reprocessing recovers 6.8 MT, which can be recycled after storage for 20 to 30 years to permit excess 228 Th to decay. The net thorium consumption of a 1000-MWe reactor then is 0.6 MT/year. Thus, the 441,000 MT of U. S. Th02 thorium reserves listed in Table 6.14 would provide thorium fuel for

of HTGR operation. The point to be made is that U. S. thorium resources are more than ample for any likely use of thorium to supplement uranium as a nuclear fuel.

It should be noted, however, that Fig. 3.33 indicates that more than 80 MT of natural uranium are consumed in an HTGR for every 0.6 MT of net thorium consumption. But in

Table 6.IS Thorium production rate

other reactor systems not treated in this text, such as the Canadian thorium-heavy water reactor, or thorium-fueled liquid-metal fast-breeder reactors, natural uranium consumption would be much lower and could be reduced to zero when and if self-sustaining breeding is reached.

Irradiated fuel discharged from nuclear power reactors can be stored or it can be reprocessed to recover and recycle the fissile and fertile nuclides. In the latter case the discharge fuel is stored at the reactor site for several months until the intense radioactivity has decayed to a level suitable for shipment and reprocessing. Criteria for such cooling times are considered in Sec. 4. The fission products are the dominant contributors to radioactivity in discharge fuel, and their intense radioactivity persists in stored fuel or in radioactive wastes from fuel reprocessing for periods of several hundred years after the fuel is discharged from the reactor. Some fission products, especially 12SI, contribute appreciable radioactivity for even longer periods, but the principal longer-term radioactive nuclides are the actinides and their decay daughters, as discussed in Sec. 3. The radioactivity of the actinides is also important in determining the necessary fission-product decontamination to be achieved in fuel reprocessing and in determining the shielding and containment required for fabricating recycled fuel. As discussed in Sec. 3, activation products in reactor structural materials and fuel cladding are important contributors to the radioactive wastes, and 14 C, formed by neutron activation of nitrogen impurities in the fuel, may be important as a potential environmental contaminant.

In aqueous solutions neptunium exists in the five oxidation states Np(III), Np(IV), Np(V), Np(VI), and Np(VII), although the heptavalent Np(VII) is stable only in alkaline solutions. In the absence of complexing agents the first four oxidation states exist as Np3+, Np4*, Np02+, and Np022+, usually in the hydrated form, whereas in strong alkaline solutions the heptavalent state is Np05 3 ‘ [K2].

Pentavalent neptunium is the most stable state in solution. It hydrolyzes only in basic solu­tions, disproportionates only at high acidity, and forms no polynuclear complexes. As shown by the oxidation-reduction potentials of Table 9.6, hexavalent neptunium is much less stable in solution than is hexavalent plutonium; in fact, hexavalent neptunium is a strong oxidizing agent and is easily reduced in the presence of oxidizable substances, such as those present in ion-exchange and solvent extraction separations [K2].

The data in Table 9.6 demonstrate that all oxidation states of neptunium are stable to dis­proportionation in 1 M HC104, in the absence of complexing agents. Only in very strong acids (pH < — 1) does Np02+ disproportionate to Np4+ and Np022+ [Al]. The disproportionation of Np(V) is promoted by addition of complexing agents, because Np4+ and Np022+ form more stable complexes than does Np02+.

Trivalent neptunium is stable only in the absence of oxygen, being oxidized to Np(IV) in aqueous solutions exposed to air [Al ].

Tetravalent neptunium forms strong complexes with anions, but Np(V) forms only weak complexes, evidently a result of the low charge of the neptunyl ion Np02+ and its small size [K2]. This may account for the relatively low distribution coefficients for Np(V) in solvent extraction and in ion exchange. To adsorb neptunium onto anion exchange resins it is necessary to reduce neptunium to Np(IV). Pentavalent neptunium is only weakly adsorbed, and Np(VI) is reduced by most exchange resins to Np(V). Because of the relatively weak organic complexes of Np02+, the distribution coefficient of its TBP complex is much lower than the distribution coefficients of the TBP complexes of Np4+ and Np02 2+ ions.

The next step in the Purex process after primary decontamination is separation of plutonium from uranium. This is done by reducing plutonium to the trivalent state, in which it is inextractable by TBP, while leaving the uranium in the extractable hexavalent condition. Reductants that have been used for this purpose include Fe2+, U4*, hydroxylamine, or cathodic reduction.

With ferrous ion or cathodic reduction, conversion of plutonium from Pu4+ to Pu3+ is so rapid that back extraction of plutonium to the aqueous phase and reduction there to Pu3+ can be carried out simultaneously in a single multistage contactor. With tetravalent uranium, reduction of plutonium is slower, so that additional contactor volume is desirable to complete back extraction. With hydroxylamine, reduction of plutonium is so much slower that it is preferable first to return both uranium and plutonium to the aqueous phase by stripping with dilute nitric acid and then to reduce the plutonium in equipment providing sufficient residence time for reduction to proceed to completion. Finally, the uranium is reextracted by TBP.

The kinetics of the reduction of Pu4+ to Pu3+ with U(IV) have been studied by Newton [N3]. Reduction kinetics with hydroxylamine have been studied by Barney [B2] and Koltunov et al. [K5].

Reduction with ferrous ion was the reaction used in the first Purex flow sheets, at Hanford and Savannah River. The specific reductant used was ferrous sulfamate Fe(S03NH2)2, a compound selected because it stabilized ferrous ion against oxidation in a nitric acid-nitrous acid system. The process was satisfactory in all respects except its addition of extraneous, nonvolatile components to the wastes.

The other three reductants are free of this disadvantage, but introduce process complica­tions. Simultaneous cathodic reduction and partitioning has been patented by Allied-General [G13] and is proposed for use in the Barnwell Nuclear Fuel Plant (Sec. 4.14), but it has not yet been used commercially. It requires a novel, electrolytic reduction extraction contactor.

Tetravalent uranium has been used in French reprocessing plants and in the German WAK plant. It requires auxiliary equipment for reducing uranium to the tetravalent form.

Hydroxylamine nitrate (HAN) requires addition of hydrazine as a holding reductant, to prevent destruction of hydroxylamine by the nitrous acid present as a result of radiolysis, in the reaction

3NH2OH + HN02 — 2N2 + 5H20 Hydrazine removes HN02 in the reaction

HN02 + N2H4 — HN3 + 2H20

Hydrazoic acid HN3 is a volatile, potentially explosive compound, but it is extracted by Purex solvent and can be removed safely in the solvent wash system.

A nuclear fission chain reaction in a reprocessing plant is an accident that must be carefully guarded against. Although such a critical reaction is not likely to generate sufficient energy to be mechanically destructive, it emits intense neutron and gamma radiation that can kill nearby plant personnel and may release radioactive fission products outside the plant.

This section outlines the methods for preventing nuclear criticality and gives some background for the conditions placed on reprocessing plant design and operation by criticality considerations. The brief discussion here and the limited examples to be cited should be used only to suggest conditions for safe design and operation. Greater detail is given in the U. S. Nuclear Safety Guide [T5], which is the source of the examples of this section, and in earlier reports [C3, C9, PI, P2, T4], which contain experimental data also. Nuclear criticality safety has been codified in American National Standards published by the American Nuclear Society, of which the ones most applicable to reprocessing plants are [A4, A5, and A6]. Even after using these standards, the design or operation of equipment in which fissile material is to be processed should be reviewed for criticality safety by an expert. And even after such review, some reprocessing systems may have novel features whose safety can be verified only by experiment.

1 URANIUM ISOTOPES

Table 5.1 lists the isotopes of uranium that are important in nuclear technology and their most important nuclear properties.

1.1 Natural Uranium

The nominal isotopic content of natural uranium is 99.274 atom percent (hereafter a/o) 438 U, 0.7205 a/o 135U, and 0.0054 a/o 234U. Slight variations in 235U content from this nominal value are discussed in Chap. 14. The 234 U:238!! ratio equals the ratio of their life-lives, as 234U, a decay product of 238 U, is in secular equilibrium with its parent. 235 U is the only naturally occurring fissile nuclide.

1.2 232 U and 233 U

233 U is a synthetic fissile nuclide produced by neutron capture in natural thorium, followed by two successive beta decays, as described in Fig. 3.2. 232U is a short-lived (72 years) alpha-emitting contaminant that is always present in 233 U from the fast-neutron reactions in thorium and 233 U described in Chap. 8. The hard gamma rays emitted by daughters of 232 U makes nuclear fuel containing 233 U more difficult to handle than fuel enriched with 23SU.

UO3 is converted to U02 by reduction with cracked ammonia gas (3H2:1N2) around 590°C in two fluidized reactors through which solids and reducing gases flow countercurrently. Exhaust gases are filtered to remove entrained dust, cooled to condense steam formed in the reaction

U03 + H2 -» U02 + H20

and the unreacted hydrogen is burned. Conditions must be carefully controlled to prevent sintering of the oxides, to obtain a product that will react satisfactorily with HF in the next hydrofluorination step. If the U02 is to be used directly as reactor fuel, as in CANDU reactors, reduction is carried out at a higher temperature to make a denser oxide.

Zirconium and hafnium have very similar chemical properties, invariably occur together in nature, and are difficult to separate. Yet their absorption cross sections for thermal neutrons are very different:

Absorption cross section for 2200 m/s neutrons Zirconium, Zr 0.185

Hafnium, Hf 102

The thermal absorption cross section of zirconium is the lowest of all mechanically strong, high-melting, corrosion-resistant metals. For this reason, zirconium and zirconium-based alloys are the materials preferred for cladding and structural materials in water-cooled, thermal — neutron power reactors.

When this type of reactor was under development for the U. S. nuclear submarine program in the early 1950s, the good chemical and mechanical properties of zirconium were recognized, but its low neutron absorption was obscured in the zirconium then available commercially by the hafnium present in natural zirconium. This caused the neutron-absorption cross section reported for commercial zirconium to be high and variable. Workers at Oak Ridge National Laboratory deduced that the variability was due to the presence of small amounts of hafnium, with its high cross section. They devised processes for removing hafnium and showed that the cross section of pure zirconium was 0.18 b. Reactor-grade zirconium has less than 100 ppm hafnium by weight.

When separated from zirconium, hafnium also has valuable nuclear applications. The high cross section, good mechanical strength, and corrosion resistance of hafnium make it an excellent material for control elements in water-cooled reactors, where it can be used without cladding.

The amount of hafnium-free zirconium used in nuclear applications is much smaller than the ordinary commercial uses of zirconium metal and compounds, for which costly removal of hafnium is not required. Zirconium metal is used in corrosion-resistant equipment for chemical plants, refractory alloys, and photo flashbulbs. The mineral zircon is used extensively in foundry sands, abrasives, and ceramics. Zircon and zirconia are widely used as refractories.

Hafnium metal is used in refractory alloys and in photo flashbulbs where especially high light output is wanted.