Category Archives: NUCLEAR CHEMICAL ENGINEERING

35S, 33P, and “a in HTGR Fuel

The graphite fuel blocks of the HTGR contain sulfur contaminant, which originates from the pitch used to form the fuel-rod matrix material. Neutron activation of the 4.22 percent 34 S in natural sulfur results in 88-day 35 S, according to the reaction

(8.66)

for which the 2200 m/s cross section is 0.24 b. Assuming that sulfur is present at 193 ppm in the HTGR fuel [HI], it is estimated that 215 Ci of 35S are present in the fuel discharged yearly from a 1000-MWe HTGR, after 150 days of storage. In HTGR fuel reprocessing the stable and radioactive sulfur will volatilize to follow the carbon dioxide from graphite incineration. The radioactive sulfur is a potential environmental contaminant that must be recovered. The amount of 35 S activity is greater than that of 14 C, and the radioactivity concentration limit for inhalation is more than an order of magnitude lower for35 S. The stable sulfur may interfere chemically with some of the recovery processes in the off-gas system.

Natural sulfur also contains 0.76 percent 33 S, which undergoes (n, p) reactions to form 25-day 33P according to

(8.67)

with a 2200 m/s cross section of 0.14 b. The estimated activity of 33P in the fuel discharged annually from a 1000-MWe HTGR, after 150 days of storage, is 1.1 Ci.

Another volatile radionuclide formed in HTGR fuel is 3.1 X 10s year 36 Cl, formed by neutron activation of chlorine contaminant in the fuel, according to the reaction

??C1 + -► $C1 + g?

Natural chlorine contains 75.77 percent 3SC1, for which the 2200 m/s activation cross section is 43 b. Assuming 3 ppm chlorine in the fabricated HTGR fuel [HI], the estimated yearly production of “Cl from a 1000-MWe reactor is 1.02 Ci.

These additional radionuclides volatilized in HTGR fuel reprocessing are summarized in Table 8.11.

Redox Process

As soon as sufficient plutonium was available for pilot-plant studies of more efficient processes, research was initiated on solvent extraction processes, which were already in large-scale use for recovery of uranium from ore leach liquors. Compared with the bismuth phosphate process, solvent extraction had the advantages that it could be operated continuously rather than batchwise, and that it could recover both uranium and plutonium in good yield and with high decontamination factors. The first solvent extraction process used in the United States for large-scale separation of uranium and plutonium from irradiated fuel was the Redox process. This process was developed by Argonne National Laboratory, tested in a pilot plant at Oak Ridge National Laboratory in 1948 to 1949, and installed by the General Electric Company at the Hanford plant of the U. S. Atomic Energy Commission (AEC) in 1951 [LI].

The solvent used in the Redox process was hexone, methyl isobutyl ketone, an extractant already in use for purifying uranium ore concentrates (Chap. 5).. Hexone is immiscible with water and will extract uranyl nitrate and plutonyl nitrate selectively from fission-product nitrates if the aqueous solution has a sufficiently high nitrate ion concentration. In the Redox process, aluminum nitrate was used as salting agent because high concentrations of nitric acid would decompose the hexone solvent.

Figure 10.1 is a material flow sheet for the first cycle of one form of the Redox process [F3]. Plutonium in the feed was converted to hexavalent plutonyl nitrate PuVI02(N03)2, by oxidation with dichromate ion Cr2072”, as this is the plutonium valence with highest distribution coefficient into hexone. In the decontamination contactor, hexavalent uranium and plutonium nitrates were extracted into hexone solvent, and fission-product nitrates were removed from the solvent by a scrub solution containing aluminum nitrate, sodium nitrate, and sodium dichromate.

In the partition contactor, plutonium was converted to inextractable, trivalent Punl(N03)3 by a reductant solution of ferrous sulfamate containing aluminum nitrate to keep uranium in the hexone phase. Plutonium was thus separated from uranium and transferred back to the aqueous phase along with the aluminum nitrate. Impure plutonium nitrate was purified by additional cycles of solvent extraction, not shown.

In the uranium stripping contactor, uranyl nitrate was transferred back to the aqueous phase by 0.1 M nitric acid strip solution. Fission products were separated from the impure uranyl nitrate by additional cycles of solvent extraction and by adsorption on silica gel, not shown.

A modification of the Redox process, the 235 U-hexone process, was used at the Idaho Chemical Processing Plant of the U. S. AEC, to recover highly enriched uranium from M5U-A1 alloy fuel elements irradiated in the Materials Testing Reactor. The aluminum nitrate needed as salting agent was provided when the fuel was dissolved in nitric acid. The plutonium content of the fuel was too low to warrant recovery. Plutonium was made trivalent and inextractable before solvent extraction and thus routed to the aqueous high-level waste.

Disadvantages of the Redox process were the volatility and flammability of the hexone solvent and the large amount of nonvolatile reagents such as A1(N03)3 added to the radioactive wastes.

JUPITER Pilot Plant

The 2-kg HM/day JUPITER pilot plant being built by Merz and associates [M7] at Julich will test a flow sheet for reprocessing high-burnup graphite-clad fuel generally similar to the two-stage process just described. In addition, the pilot plant will demonstrate fluidized-bed combustion of crushed gTaphite-clad fuel assemblies and refabrication of fuel from recovered uranium and thorium. Radiokrypton in combustion off-gases will be recovered by cryogenic absorption in liquid C02, and tritiated water vapor by adsorption on molecular sieves.

Recovery of Transuranium Elements

To begin with, it is necessary to measure the transuranic content of large volumes of rather heterogeneous wastes. Basically there are three ways to do so, namely, by making use of gamma — and x-ray spectra accompanying the alpha decay, of radiation produced by spontaneous fission, and of radiation produced by induced fission.

Recovery of transuranium elements is mainly of interest for solid waste from the refabrication of mixed-oxide fuel. Plutonium is the major element to be recovered, and 241 Am may be recovered as a by-product. Other transuranium elements are usually present in minor quantities. The treated wastes are seldom decontaminated to levels of plutonium that would permit unrestricted release.

The most rigorous recovery technique is burning of plutonium-impregnated material in a plutonium scrap-recovery incinerator followed by grinding and leaching the ash with a mixture of hot nitric and hydrofluoric acids. Undissolved plutonium in the ashes may be recovered by fusion with a suitable salt, such as a 10:1 K^SO^-NaF melt, to get a product soluble in nitric acid. Nonbumable solids can be leached directly with a HN03-HF mixture.

Once the plutonium is in solution, it can be recovered and purified for recycle by well-established solvent extraction and ion-exchange techniques. Aluminum nitrate is added to the feed to complex the fluoride and thus decrease its interference with plutonium recovery.

Figure 11.22 presents a scheme of typical plutonium recovery operations. The Plutonium Reclamation Facility (PRF) [Rl] at Hanford incorporates many of these options. Geo­metrically favorable process equipment and storage tanks are used to ensure criticality safety.

The PRF also recovers 241 Am from the raffinate of the TBP-solvent extraction plutonium — purification process. The process employs 30 v/o dibutyl butylphosphonate in ССЦ as the solvent to extract both americium and residual plutonium from the high-nitrate feed solution, adjusted to about pH 1 by the addition of NaOH. Americium is selectively stripped from the solvent and purified by a cation-exchange procedure.

3.1 Immobilization

Volume reduction as described above usually leads to a product that still contains considerable quantities of water or that is quite easily leached or dissolved by water. The policy as to the degree of immobilization required for final disposal varies in different countries. As yet, there is no official regulation in the United States requiring that non-high-level waste be immobilized before disposal. It is, however, practiced in many places. In West Germany, by regulation, any non-high-level waste has to be immobilized before disposal in such a way that low leachability is warranted over a sufficient period of time.

There is no doubt that immobilization at least of alpha-bearing waste must generally be required and will be in the future. It has been mentioned before that the total transuranic inventory of alpha-bearing non-high-level waste will be in the same order of magnitude as that of HLW.

The range of suitable immobilization products for non-high-level waste is broader than that for HLW because there will be no significant heat generation. It includes glasses as well as cement, bitumen, and polymers.

Hydraulic cement. Immobilization of radioactive waste by incorporation into hydraulic cements, as typified by portland cement, has been practiced for many years. The optimum proportions of cement and waste vary with the type of waste to be immobilized.

Several additives have been used to improve the setting properties, fission-product

Key

—- Solid Liquid

Figure 11.22 Typical plutonium-recovery operations (Hanford Engineering Development Labora­tory). (From Richardson [R1J.)

retention, or the volumetric efficiency of cement. A useful mixture is the Portland cement sodium silicate system developed by United Nuclear Industries, Inc. The sodium silicate additive produces a quick set with no free water, readily solidifies pressurized-water-reactor boric acid solutions, which set poorly with cement alone, and provides a significant reduction in the solidified volume [H3].

Another way to improve cement products is polymer impregnation. The process being developed in Italy consists of preparation of the cement product, thermal dehydration of the cement (165°C), impregnation with a catalyzed organic monomer, such as styrene, and polymerization by heating at 75 to 85°C [D3].

In spite of experience, solidification with cement is still an art. Each new waste application must be considered individually because of possible interactions between cement and the waste constituents.

Bitumen. Bituminization systems for immobilizing liquid and solid wastes are used in several countries. Bitumen, or asphalt, has certain advantages for immobilizing LLW and MLW. It is highly leach-resistant, it has good coating properties, and it possesses a certain degree of plasticity. Perhaps the greatest advantage is that at the operating temperature of 150 to 250°C, 99 percent of the water evaporates, resulting in a volume reduction of up to fivefold compared with conventional cementing techniques for products made with evaporator concentrates. Typical bitumen products contain 40 to 60 w/o waste solids.

One of the major drawbacks of bitumen is its potential fire hazard, particularly if used to encapsulate oxidants such as nitrates. The combustion problem is minimized by using bitumen grades with high flash points (>290°C). Improved safety can also be obtained by substituting more expensive polyethylene for bitumen. Fires have occurred in bituminization facilities, but they have been readily controlled.

Another problem to be observed is the radiation resistance of bitumen. There may be some radiolysis resulting in the release of hydrogen, methane, carbon dioxide, and ethylene. In the order of 0.5 cm3 H2/g product has been found to be generated per 108 rad. This is of significance primarily for alpha-bearing products.

The bituminization process is performed basically by adding a concentrated waste slurry or even a predried waste mixture to the molten bitumen. Residual water is evaporated and the solids are evenly distributed in the bitumen. After solidification a homogeneous product is obtained. Figure 11.23 shows a flow sheet of a screw extruder plant for bituminization; Figure

11.24 is a photograph of the screw extruder evaporator.

Glass. For liquid non-high-level alpha-bearing wastes with sufficiently high activity concentra­tion, glass may be a suitable fixation product as it is for high-level waste.

In terms of radiation stability, glass is superior to cement and particularly to bitumen. The leach rates, however, are about the same for glass and for bitumen, both being smaller than that of cement by up to three orders of magnitude depending on the type of cement. The fire hazard is a disadvantage of bitumen compared with both glass and cement. The costs of immobilization decrease in the order glass, bitumen, cement.

SOLVENT REQUIREMENTS

Generally speaking, a solvent suitable for separation of metals by fractional extraction should meet the following requirements:

1. It should be selective; i. e., the ratio of distribution coefficients should be high.

2. It should have good capacity for extraction; i. e., distribution coefficients in the extracting section should be of the order of unity or higher.

3. It should be readily stripped; i. e., distribution coefficients in the stripping section should be no greater than unity.

4. It should be relatively immiscible with water, to reduce solubility losses.

5. Its density should be appreciably different from water, and it should have low viscosity and fairly high interfacial tension. These physical properties are important in promoting separation of phases following contact.

6. For safety reasons it should be relatively nonvolatile, nonflammable, and nontoxic.

7. It should be readily purified, preferably by fractional distillation.

8. It should be stable in the presence of chemical agents used in the process, such as nitric acid. Solvents used for radioactive materials should also have good radiation stability.

TBP meets most of these requirements except those of low viscosity and a density different from water. These deficiencies are corrected by diluting TBP with a light, saturated hydro­carbon, such as an aromatic-free kerosene. This solvent is the one most commonly used at present in fractional extraction of metals. The physical properties of TBP are summarized in Table 4.5 [FI, S4].

Although TBP is a relatively stable organic compound, it does undergo slow hydrolysis to form di-n-butyl phosphate (DBP). Although the presence of DBP increases the distribution coefficients of uranium, plutonium, and other actinides, it interferes with the separation of plutonium from uranium, and it makes complete stripping of these elements difficult. DBP forms an insoluble compound with thorium. DBP formation is appreciable only when the

solvent is held for long periods at temperatures as high as 50 to 60°C, but it can be removed by periodically scrubbing the solvent with a basic solution [S4].

TBP can be decomposed explosively when heated to above 120°C in the presence of extractable nitrates [S4].

SOURCES OF URANIUM

1.13 Principal Uranium-containing Minerals

At current prices of uranium ($30 to $50/lb U308) ores containing around 0.1 percent U308 or more are being mined primarily for their uranium content. The principal uranium-bearing minerals found in such ores are listed in Table 5.15, classified according to the type of treatment needed to extract the uranium.

The first group, minerals containing high concentrations of uranium, mostly in the tetravalent state, can be concentrated by specific gravity methods when in massive form. Frequently, however, the particle size is so small that the uranium-bearing mineral must be dissolved in sulfuric acid or sodium carbonate leach liquors. In either case, an oxidant must be added to bring uranium to the soluble, hexavalent state.

Uraninite, or pitchblende as it is more commonly known, is the form in which uranium was first discovered, at Joachimsthai, Czechoslovakia. Later, very rich deposits of massive uraninite were discovered at the Shinkolobwe mine in the Belgian Congo (Zaire) and were the principal source of uranium for the Manhattan Project. Leaner ores containing finely divided

Table 5.15 Principal uranium minerals

Table 5.16 Low-grade sources of uranium^

Material

Location

Concentration range, g U/MT

Commercial sources

Lignite

South Dakota

10-4000

Shale

Sweden

250-325

Gold tailings

South Africa

60-300

Phosphate rock

Florida

100-200

Copper tailings

Western United States

2-50

Sources presently not economic

Shale

Chattanooga, Tennessee

50-70

Bostonite

Colorado

33

Granite

Conway, New Hampshire

10-30

Earth’s crust

Average

1.7

Seawater

0.003

^Most of this information is from [B2].

uraninite are now being mined in the Lake Athabasca district of Canada, in the White Canyon and Big Indian Wash districts of Utah, and the Jackpile mine of New Mexico.

Coffmite is a major mineral in the important Ambrosia Lake district of New Mexico. Uranothorite is mined in the Bancroft district of Ontario.

The second group of minerals in Table 5.15, hydrated minerals containing hexavalent uranium, are usually soft, finely divided, of relatively low density, and readily soluble in dilute sulfuric acid or sodium carbonate solution. Their uranium cannot be concentrated by specific gravity or flotation methods and must be recovered by leaching and concentration by selective precipitation, solvent extraction, or ion exchange. These ores are usually of secondary origin, precipitated from uranium-bearing groundwaters.

Camotite is a major ore of the Colorado plateau. Tyuyamunite occurs near Grants, New Mexico, in Utah, and in the Soviet Union. Autunite is found near Marysvale, Utah, and in Washington and Wyoming. Torbemite is found in the White Canyon district of Utah and in the upper zones of ore bodies in Zaire.

Minerals of the third group of Table 5.15 contain relatively small proportions of tetravalent uranium combined with a refractory oxide of titanium, niobium, or tantalum. To free the uranium from these minerals, they must be leached with hot, concentrated sulfuric acid. Davidite is one of the principal ores at Radium Hill in South Australia. Brannerite is found in the Blind River district of Ontario. Pyrochlore occurs in the Lake Nipissing district of Ontario and in Nigeria.

Thorium Hydrides [М2]

Thorium forms two hydrides: ThH2, density 9.50 g/cm3; and ТІцН^, density 8.28 g/cm3. The limited information on temperature-composition relations for condensed phases in the system thorium-hydrogen is shown in Fig. 6.3. Approximate values for the equilibrium pressure of hydrogen in the system aTh-ThH2_x obtained from measurements of Mallett and Campbell [Ml] made with impure metal are given in Table 6.9.

Like other metal hydrides, thorium hydride is pyrophoric and must be handled with care.

1.8 Thorium Halides

The tetrahalides are the thorium halides of greatest practical importance. The tetrafluoride ThF4 is the preferred starting material for large-scale production of thorium metal (Sec. 10.4). ThF4 has been proposed as fertile material in the fuel mixture of the molten-salt reactor. The tetraiodide has been used as feed material in the iodide process for making very pure thorium metal (Sec. 10.4).

Figure 6.3 Thorium-hydrogen phase diagram [М2]. (Reprinted with permission from the copyright holder, Academic Press, Inc., New York, and Dr. G. G. Libowitz.)

Table 6.9 Hydrogen pressure in system thorium-thorium hydride, Th/H =1.6

Temperature, °С

651

676

701

726

751

776

801

826

851

876

Hydrogen

pressure, Torr

20

35

55

90

130

195

280

400

580

800

The more important properties of the tetrahalides, from reference [II], are listed in Table 6.10. Many of these properties, especially for ThCl4, ThBr4, and Thl4, are known only semiquantitatively.

Anhydrous ThF4 is made by passing an excess of HF vapor over Th02 or ThOFj at temperatures between 550 and 600°C. The anhydrous double fluoride KThF5 is precipitated from aqueous solutions of thorium nitrate by addition of an excess of KF. It has been used for electrolytic production of thorium metal.

Table 6.10 Thorium Tetrahalides

Property

ThF4

ThCL,

ThBr4

Thl4

Color

White

White

White

Yellow

Density at 25°C, g/cm3

6.12

4.62

5.72

6.00

Crystal system

Low-temperature form

Monoclinic

Tetragonal

Orthorhombic

Monoclinic

High-temperature form

Orthorhombic

Tetragonal

Orthorhombic

Transition temperature, °С

(406)

~420

?

Melting temperature, °С

1110

770

679

570

Normal boiling point, °С

1782

942

905

853

Vapor pressure equation, logio p(atm) = A — B/T

Solid, A

9.345

9.426

9.498

9.747

В, К

17,089

10,630

10,151

9,894

Liquid, A

6.395

5.229

5.260

5.714

В, К

13,080

6,346

6,187

6,425

Heat capacity equation, Cp=A + 10*3 BT — 10SC/7’S

Solid, A, cal/(g-mol-K)

26.75

28.75

30.5

31.0

B, cal/(g-mol’K2)

5.854

5.561

3.6

3.1

C, (cal-K)/g-mol

1.805

1.470

1.47

1.47

Liquid, cal/fg-mol’K)

36.5

40.0

41.0

42.0

Heat of transition, cal/g-mol

1,200

1,000

?

Heat of fusion, cal/g-mol

10,510

14,700

(13,000)

11,500

Heat of vaporization at normal boiling point,

cal/g-mol

55,700

26,500

27,500

Heat of formation at 25 C,

cal/g-mol

-504,600

-283,600

-230,800

-158,800

Free energy of formation, AG=A+BT

Solid, A, cal/g-mol

-502,140

-281,100

-243,450

-187,040

Solid, B, cal/(g-mol-K)

70.11

67.70

70.09

69.46

Liquid, A, cal/g-mol

-485,080

-261,120

-225,190

-171,000

B, cal/(g-mol‘K)

57.38

48.09

50.55

50.25

ThCl4 can be made by reacting Th02 with chlorine mixed with CC14 or COCl2.

All tetrahalides react with water to form oxyhalides:

ThJCt + H20 ->• ThOX2 + 2HX

For this reason, ThF4, precipitated from aqueous solution, cannot be dried without contamina­tion by oxygen. When ThCl4 is dissolved in water, soluble ThOCl2 is formed and crystallizes out on evaporation. The oxyhalides are stable against disproportionation into oxide and tetrahalide at pressures near atmospheric and temperatures under 2000 K, as may be seen from the positive free-energy change AGdjsp in the reaction

2ThOX2 ->• Th02 + ThX4

The free-energy change ДС^р may be calculated from the enthalpy change A#disp and entropy change ASdisp f°r the disproportionation reaction given in Table 6.11 by Eq. (6.6):

AGdisp = Д^/disp T ASdisp (6.6)

Thorium di — and triiodides have been prepared by Scaife and Wylie [SI] and are of some practical significance in the iodide process for making thorium metal (Sec. 10.4). Other lower halides have only limited stability and are not well known.

Selective Reduction of Double Fluorides by Aluminum

Two recent patents [М2, М3] by J. A. Megy describe a process in which zirconium metal is reduced from a salt and separated from hafnium in the same step, thus shortening the long series of steps in present processes for producing reactor-grade zirconium from natural zircon.

The Megy process has evolved from the finding of Petenev and Ivanovskii [PI ] that when a mixture of K3ZrF6 and K2HfF4 dissolved in molten alkali chlorides was reduced electro — lytically at a molten zinc cathode, the metal phase was enriched in zirconium relative to the
residual salt. Megy found that when a mixture of Na2ZrF6 and Na2HfF6 is reduced by aluminum dissolved in liquid zinc, a very high separation factor between hafnium and zirconium is obtained, with very little contamination of the zirconium by aluminum. Examples given in Megy’s patent [М2] indicate that the ratio of zirconium to hafnium in the metal phase may be as high as 328 times the ratio of these elements in the residual salt phase. This high separation factor, a = 328, makes possible production of reactor-grade zirconium containing less than 0.01 w/o hafnium from typical natural zirconium containing 2 w/o hafnium in two stages of salt-metal contact, such as shown in Fig. 7.9.

Material quantities in Fig. 7.9 are based on production of 1.000 mol zirconium at point 6. Feed to this process (point 1) consists of 1.058 mol Na2ZrF6 and 0.011 mol Na2HfF6, corresponding to 2 w/o hafnium in zirconium + hafnium. This feed is combined with 0.0558 mol Na2ZrF6 and Na2HfF6 of the same composition (point 5) recycled from a previous batch. In step A the salt feed is reacted in a graphite-lined container at 900° C with a metallic solution of 4 w/o aluminum in molten zinc containing 1.4078 mol aluminum. Reactions taking place are

3Na2 ZrF6 + 4Al(Zn) -*• 4(NaF)1<s A1F3 + 3Zr(Zn)

and 3Na2HfF6 + 4Al(Zn) -* 4(NaF)i. sAlF3 + 3Hf(Zn)

At the reaction temperature the products, (NaF)1.sAlF3 and the solution of zirconium and hafnium in zinc, are two immiscible liquids. The lighter salt phase, enriched in hafnium to 27 w/o, is drawn off at point 4, leaving a heavier metallic zinc solution of zirconium containing

0. 1126 w/o hafnium at point 3. These compositions are consistent with the separation factor of 328:

(27X100-0.1126) (?

(100-27X0.1126)

To reduce the hafnium content below the 0.01 w/o specified for reactor-grade zirconium, the solution of zirconium and hafnium in molten zinc at point 3 is contacted in step В with a liquid mixture of 0.1116 mol ZnF2 + and 0.1116 mol NaF. Reactions

Zr + 2ZnF2 + 2NaF -*• Na2ZrF6 + 2Zn

+Alternatively, an equivalent amount of natural sodium fluozirconate could be used, with only slight increase in hafnium content of product zirconium.

Table 7.12 Flow rates and compositions in TBP solvent extraction pilot plant for hafnium- zirconium separation

Solvent feed, 60 v/o TBP in heptane

Aqueous

Stream

Feed

Raffinate

Scrub

Strip

Product

Flow rate, liters/h

5.0

1.0

2.2

1.18

1.62

1.56

Mol HN03/liter

0.65

5.1

2.65

5.4

0

3.96

g (Zr, Hf)02 /liter

0

127

3.5

0

0

76.4

w/o hafnium in zirconium + hafnium

2.2

43.8+

<0.01

+This reported composition does not satisfy a hafnium material balance, probably because of unsteady cascade conditions; 30.9 w/o hafnium would balance.

Source: R. P. Cox et al., Ind. Eng. Chem. 50: 141 (1958).

0.0 5 8 265 Na2Zr F6 0.011006 Na2Hf Fg 15.89 m/o Hf 26.99 w/o Hf I 4078 (NoR и A I Fj

0.0552 Na2Zr F6 0.0005759 Na2 Hf Fg 1.0 3 22 m/о Hf 2 0000 w/o Hf

Reactor — grade Zr in Zn

1.0000 Zr

0.0000318 Hf

0.003 18 m/o Hf

000622 w/o Hf

14.0570 Zn Solvent (Recycled!

Figure 7.9 Megy process for producing reactor-grade zirconium from natural sodium fluozircon — ate. Material quantities in moles. Basis, 1 mol zirconium product.———————————————————————————- salt;——- metal.

and Hf + 2ZnF2 + 2NaF -*■ Na2HfF6 + 2Zn

take place, with most of the hafnium and a small fraction of the zirconium reacting. The hafnium content of the residual zirconium metal at point 6 is 0.00622 w/o, thus meeting the

0. 01 w/o specification of reactor-grade zirconium.

The hafnium content of the salt at point 5 is at the feed level of 2 w/o, again satisfying the separation factor condition:

(2X100 — 0.00622) _

(100 — 2X0.00622) ~ 328 (7-5)

The salt at point 5 is recycled to step A of a later batch. The zinc solvent at point 6 is distilled from the zirconium product and recycled to a later batch at point 2.

The mixture of (NaF)i. sAlF3, Na2ZrF6, and Na2HfF6 is converted to more useful by-products by reduction with 4 w/o aluminum in zinc in step C. This produces a zirconium and hafnium-free fluoride salt by-product 7, (NaF)i.5AlF3, and a solution of 27 percent Hf, 73 percent Zr in zinc, 8. The (NaF)1.sAlF3 can be sold as a substitute for cryolite Na3AlF6, in electrolytic production of aluminum. The zinc can be distilled from the 27 percent Hf, 73 percent Zr and recycled to point 9, and the hafnium-zirconium alloy can be sold for metallurgical applications in which the high cross section of hafnium is not harmful.

Figure 7.10 shows how the Megy selective reduction process can be combined with K2SiF6
fusion to produce reactor-grade zirconium from zircon ore. The ore is fused with K2SiF6 in a graphite-lined arc furnace A at 1000°C to convert zirconium to K2ZrF6 and K2HfF6:

K2SiF6 + (Zr, ffi)Si04 ->■ K2 (Zr, Hf)F6 + 2Si02

Potassium is preferred to sodium because the potassium complex fluorides are more stable at this temperature. The K2(Zr, Hf)F6 is dissolved in water and filtered from insoluble Si02 at В and crystallized at C. To recover the relatively expensive potassium, the K2(Zr, Hf)F6 crystals are dissolved in heated NaCl brine and cooled to precipitate the less soluble Na2(Zr, Hf)F6 at D:

K2(Zr, Hf)F6 + 2NaCl -*■ Na2(Zr, Hf)F6 + 2KC1

NaCl Solution

Figure 7.10 Production of reactor-grade zirconium from zircon by combination of K2SiF6 fusion and Megy process.

The Na2(Zr, Hf)F6 is converted to zirconium metal by the Megy process described earlier. The KC1 is recycled and converted to K2SiF6 by countercurrent metathesis with purchased Na2SiF6 at E.

2KC1 + Na2SiF6 -*• K2SiF6 + 2NaCl

In this way zircon, Na2SiF6, and aluminum are converted to zirconium metal, hafnium-rich zirconium, and by-product (NaF)i.5AlF3 and Si02.

Protactinium Compounds

Protactinium oxides. Protactinium forms the fee dioxide Pa02, which, like U02, adds additional oxygen to form the hyperstoichiometric Pa02+x. The most stable oxide is the pentavalent Pa205, which exists in five different crystalline forms. The pentoxide results from heating any binary Pa(IV) or (V) compound in oxygen to temperatures above 650°C. Hydrogen reduction at 1550°C transforms Pa205 to black Pa02. Like thoria, high-fired Pa205 is only slightly soluble in mineral acids [K2].

Protactinium halides. Presently known halides of protactinium are PaF4, Pa2F9, PaFs, РаСЦ, PaCl5) PaBr4, PaBr5, Pal3, Pal4, and Pal5. The pentafluoride is formed by the high-temperature reactions of fluorine with protactinium compounds, Hydrogen-HF mixtures stabilize the tetrafluoride PaF4 [B4].

The volatilization of PaCl5 has been used in the analytic separation of 231 Pa in wastes, which are heated in ССЦ vapor in a closed vessel at 400 to 500°C [C7].

Decladding

The method used for decladding depends on the composition of the cladding and fuel and the bond between them, if any. The two general decladding methods are chemical and mechanical.

Chemical decladding. In chemical decladding, the clad is removed and the fuel exposed by dissolving the clad, to leave the fuel as a separable solid. Chemical decladding has the disadvantage that the clad reaction products require more storage space than the original cladding. However, it sometimes is the more practical method. For example, fuel for the first U. S. production reactors consisted of a uranium metal slug bonded with aluminum-silicon alloy to an aluminum can. Because of the bond, mechanical decladding was impractical. Chemical decladding consisted in dissolving the aluminum can and the bond in hot, aqueous, 10 w/o sodium hydroxide solution containing about 20 w/o sodium nitrate to prevent evolution of hydrogen. The overall reaction was [B12]

A1 + 0.85NaOH + 1.05NaNO3 -+ NaA102 + 0.9NaNO2 + 0.15NH3 + 0.2H2O

With this reactant, uranium metal fuel is relatively unattacked.

Another example of chemical decladding is afforded by the Zirflex process, which was proposed for zircaloy-clad U02 fuel before mechanical decladding was fully developed. In the Zirflex process [S17], zirconium or zircaloy cladding is dissolved as ammonium fluozirconate in a boiling solution of ammonium fluoride containing ammonium nitrate, the latter added to reduce hydrogen evolution. Overall reaction is approximately

Zr + 6NH4F + 0.5NH4NO3 -+ (NH4>2ZrF6 + 5NH3 + 1.5H20

Because of limited solubility of the ammonium fluozirconate product, there is an optimum NH4F concentration, around 5.5 Af, with an initial molar ratio of fluoride to zirconium between 6.5 and 7.0.

A third example of chemical decladding is afforded by the Sulfex process, which was proposed for stainless steel-clad U02 or Th02 fuel before mechanical decladding was fully developed. In the Sulfex process [F2], stainless steel cladding is dissolved in hot 4 to 6 M sulfuric acid. Disadvantages of the process are the slow and variable dissolution rate, passivation of the steel by nitrate ion unavoidably present if the same dissolver is used alternately to dissolve cladding and fuel, some attack of U02 by H2SC>4, and evolution of hydrogen.

Electrolytic dissolution in nitric acid has been used at the Savannah River [B22] and Idaho Chemical Processing plants [A10, All] to dissolve a wide variety of fuels and cladding materials, including uranium alloys, stainless steel, aluminum, zircaloy, and nichrome. The electrolytic dissolver developed by du Pont [B22], pictured in Fig. 10.4, uses niobium anodes and cathodes, with the former coated with 0.25 mm of platinum to prevent anodic corrosion. Metallic fuel to be dissolved is held in an alundum insulating frame supported by a niobium basket placed between anode and cathode and electrically insulated from them. Fuel surfaces facing the cathode undergo anodic dissolution in a reaction such as

Fe -► Fe3+ + 3e~

At nitric acid concentrations above 2 M, the cathode reaction is

N03~ + 3H+ + 2e~ -»■ HN02 + H20

so that hydrogen evolution is suppressed. The reverse reaction takes place at the platinum — coated anode, together with evolution of some oxygen in

2H20 ->■ 4H+ + 02 + 4e~

Anodic corrosion of the niobium basket that supports the fuel is inhibited by an electrically conducting oxide film, which forms on the niobium.

Advantages of this method are its wide applicability, the absence of anions other than the nitrate ion, and the fact that little hydrogen is evolved. A disadvantage is the presence of cladding metal nitrate in the dissolver solution and its eventual routing to the high-level wastes. When applied to zircaloy cladding, most of the zirconium is converted to hydrous Zr02, which can be filtered from the dissolver solution.

Wind scale, England

British Nuclear
Fuels, Ltd.

Magnox

Oxide

1964

Operating

Planned

5

4

Natural to 1%

4,000

37,000

130

Mechanical

360

Shear-

leach

Continuous

Batch

Decanter

Centrifuge

Karlsruhe, W. Germ­any

KFK/GWK

Hessen (WA 350) W. Germany DKW

Oxide

Oxide

1971

Operating

1992

Planned

0.17

2t

3

3.5

39,000

40,000

250

Shear-

leach

2500

Shear-

leach

Batch

Batch

Filter

Centrifuge

Contactors

Mixer-settlers +

Mixer — Centrifugal +

Mixer-

pulse

settlers mixer-

settlers

columns

settlers

v/o TBP

30

ЗО зо

20

Number of code-

contamination

cycles

Pu reductant*

1

U(IV)

2 3 U(IV)

2

FeSAm

Pu purification

1 cycle TBP +

1 cycle TBP +

2 cycles

oxalate ppt.

oxalate ppt.

TBP +

oxalate

PPt.

U purification

1 cycle

None

1 cycle

TBP

TBP

Waste treatment

CH20

CH20 denitration

Evaporation

denitration

Maintenance

Direct

Direct

Direct

References

[Cl4, C15, J3]

[C5, Cl 2, Cl 4, D2]

[W3]

t Expected to operate 175 days/yr.

^Also processed 93% 235 U at reduced rate and lower v/o TBP. § U(IV), tetravalent uranium; FeSAm, ferrous sulfamate.

Pulse

Mixer-

Pulse

Pulse

Mixer-

columns

settlers

columns +

columns +

settlers

mixer-

mixer-

settlers

settlers

30

30

30

30

30

1

1

1

1

2

U(IV)

U(IV)

U(IV)

U(IV)

U(IV) +

n2h4

2 cycles

1 cycle

2 cycles

1 cycle

1 cycle TBP

TBP +

oxalate

ppt.

TBP +

anion

exchange

TBP

TBP +

oxalate

ppt.

2 cycles

1 cycle

2 cycles

1 cycle

1 cycle

TBP (in M-S)

TBP + Si02

TBP

TBP + Si02

TBP

Evaporation

Evaporation

Evapora­tion + denitration

Evaporation

HCOOH

denitration

Direct

Direct

Direct

Direct

Remote and direct

[B17]

[S3]

[D3]

[11, Ul]

Figure 10.3 Cogema reprocessing plant at La Hague, France. (With permission of Cogema.)

Figure 10.4 Electrolytic dissolver at Savannah River Plant. (Photo courtesy of E. I. duPont de Nemours & Company.)

Mechanical dedadding. The objective of mechanical decladding is to break or cut the cladding so as to expose the fuel to reaction with a dissolvent that does not attack the cladding. The best decladding method is one that requires minimum disassembly of fuel elements, minimizes production of fines, produces fuel fragments that can be readily and completely leached, and uses equipment with minimum maintenance requirements. Methods that have been used include transverse chopping with a shear, transverse cutting with a saw or abrasion wheel, longitudinal slitting, and longitudinal extrusion.

Transverse methods have the advantage of not requiring disassembly of a fuel element into individual rods before decladding. Transverse chopping with a specially designed shear is the method now generally favored for decladding fuel bundles from U. S. boiling — and pressurized — water reactors. This method, developed at Oak Ridge National Laboratory [W4] and Hanford [КЗ], uses a stepped blade such as shown in Fig. 10.5 moved horizontally past a circular or V-shaped anvil. Blade wear is minimized by rounding the cutting edge to a ^j-in (0.8-mm) radius before use. Blade life is from 10,000 to 50,000 cuts at a stroke rate of from 1 to 2 in/s (2.5 to 5 cm/s). For a 36-rod boiling-water reactor (BWR) fuel assembly, forces of from 45 to 80 t are required. Optimum length of segments is from j to 2 in (1.25 to 5 cm), depending on degree of oxide fragmentation, dissolving conditions, blade life, and cost of blade replacement. Such shearing leaves the cut lengths open for subsequent leaching and produces little metal fines. With zircaloy cladding, the shear must be operated in an inert atmosphere to prevent zirconium fires. The first production use of such a shear, at the West Valley plant of Nuclear Fuel Services, was very satisfactory, with necessary maintenance carried out by remote means.

Fuel bundles have been declad by transverse sawing with a hacksaw blade operated under water to provide cooling and prevent zircaloy fires. The preferred saw consists of a hardened tool-steel cutting edge welded to a tough-steel blade [H2]. More fines are produced than in shearing.

Unbonded aluminum jackets are removed from uranium metal fuel rods in the British [C4] and French Magnox reactors by longitudinal extrusion through a hardened steel die. The hole in the die is large enough to admit the uranium rod but small enough to reject and peel off the jacket, which is scored lengthwise before meeting the die. The method is not applicable to oxide fuel because the fuel would crumble and jam the die.

A longitudinal cut with a remotely operated milling cutter was used to remove stainless steel jackets from uranium metal fuel rods in the first core of the Experimental Breeder Reactor-1 [Cl 8].

Gas evolution in decladding. Fuel elements with unvented cladding contain fission-product gases under pressure. Some fuel elements also contain helium charged during fabrication to improve heat transfer in subsequent reactor operation. In decladding, this helium is evolved, together with around 10 percent of the fission-product krypton and xenon and a small fraction of the iodine, tritium, and 14 C. Gas evolved in decladding is routed to the off-gas treatment system for

Figure 10.5 Stepped blade used for shearing metal-clad uranium oxide fuel bundles. (From J. T. Long, Engineering for Nuclear Fuel Re­processing, Gordon & Breach, New York, 1967, with permission.)

retention of iodine and entrained solids. Processes have been developed for retention of krypton and xenon. Processes are being developed for tritium and 14C, but are not yet in general use.

Voloxidation. If tritium is to be separated in the plant, it is highly desirable to do so before dissolution, when the tritium would be diluted isotopically with the large amount of hydrogen added as water and nitric acid in the dissolver. The voloxidation process has been developed by Oak Ridge National Laboratory for this purpose [FI, G8]. In this process, sheared or cut fuel is oxidized in a rotating kiln to convert U02 to U308. As U308 is less dense than U02, the fuel swells and is pulverized, thus exposing occluded tritium to oxidizing gases and converting it to tritiated water. More than 99 percent of the tritium and the remaining krypton and xenon escape from the fuel. The gases are filtered, passed over heated copper oxide to convert any unreacted hydrogen to water, and cooled, after which the tritiated water is absorbed by a molecular sieve or anhydrous CaSO,».

In voloxidation, stainless steel-clad U02 is oxidized with flowing air or oxygen at 575 to 650°C. With zircaloy-clad fuel, these gases may be unsafe because of the danger of a zirconium fire. Less reactive N204 has been proposed as an oxidant for such fuel.