The Nuclear Fuel Cycle

Undoubtedly the most extraordinary things that happen to reactor fuel happen within the
core of an operating reactor. But a great deal happens to it outside the reactor, both before
and after its sojourn in the core. The odyssey of the fuel material, from its origin in the
earth's crust, takes it from a mine, to a mill, possibly through a specialized facility called
an enrichment plant, and through a fuel fabrication plant before it enters the reactor.
When it emerges from the reactor it may go into storage or into another specialized
facility called a reprocessing plant. Some of the material thereafter reaches a theoretically
final resting place, while the rest may re-enter the process at an earlier stage. The whole
succession of processes, with the transport which links them, is called the nuclear fuel
cycle. In practice it is not very cyclic; but the possibility exists of making it much more
so, provided certain problems - both technical and otherwise - can be overcome. Present
policies within the nuclear industry are generally directed towards this end. But, cyclic or
otherwise, the nuclear  fuel cycle outside the reactor gives rise to many of the most
controversial aspects of nuclear technology. In the following pages we discuss the fuel
cycle, and some of the problems which arise in it.
Uranium Production
Uranium is found in nature as mineralization in sandstones, in quartz pebble
conglomerate rocks, and in veins, and to a smaller extent in other types of deposit. There
are significant uranium reserves in the USA, Canada, Southern Africa, Australia, France
and elsewhere. High-grade uranium ores contain up to 4 per cent uranium; but known
reserves of this quality have been largely worked out, and ore grades ten times lower, 0.4
per cent and less, are now being worked. Still lower grades - down to 0.01 per cent and
less - are also being noted for development.
Uranium ore deposits are found by a variety of exploratory techniques. The folklore
image of the uranium prospector with his Geiger counter, picking his way over the
hillside listening for clicks, has little to do with contemporary uranium prospecting.
Uranium exploration usually begins in the air, looking for abnormal traces of airborne
radioactivity given off by the decay products - so-called 'daughters'  - of uranium.
Airborne instruments look for tell-tale gamma rays and other evidence of radioactivity.
More evidence is assembled on the ground, by studying likely geological formations, by
testing samples chemically, and ultimately by drilling.

Uranium ore is extracted by surface or underground mining. The crude ore is fed into a 

series of crushing mills, which grind it to the consistency of fine sand. Chemical solvents 

then dissolve out the uranium, which emerges from the process in the form of a mixture 

of uranium oxides with a chemical formula equivalent to U308 This oxide mixture,

usually called 'yellowcake', forms the raw material for all the succeeding processes that 

lead eventually to the reactor core and the chain reaction. Yellowcake contains 85 per 

cent uranium by weight. Besides the yellowcake, there remains after extraction some one 

hundred times its weight of residual sand, called 'tailings' - which also contains the 

radium which had accompanied the uranium. There also remains, per tonne of ore, over 

3700 litres of liquid waste, which is both chemically toxic and radioactive. A uranium 

mine and associated mill may produce over 1000 tonnes of uranium per year, from at 

least 250 000 tonnes of ore.

Hazards arise at several stages in the uranium production process. The first of these arises 

from the uranium ore itself, in situ and subsequently. When uranium-238 under goes 

alpha decay, it produces a succession of further alpha-emitters, including radium-226 and 

its immediate daughter-product, the chemically inert but radioactive gas radon-222. Any 

aggregation of uranium which has remained for some time undisturbed - such as a 

geological deposit  - therefore exudes this radioactive gas. When a uranium ore deposit is 58

broken up in mining the escape of the radon is facilitated. Radon-222 is an alpha emitter 

with a half-life of less than four days, which produces its own radioactive 'daughters'. 

These radon daughters are however solids. When a radon-222 nucleus in the air emits an 

alpha particle, the resulting nucleus of polonium-218, being momentarily electrically 

charged, adheres to any dust particle nearby. Accordingly, air containing radon also

contains dust particles laden with intensely radioactive radon daughters. Underground 

uranium miners who are permitted to inhale such air have proved appallingly susceptible 

to lung cancer.

The first evidence of this effect was established by 1930, after medical investigations of 

miners working deposits in Joachimsthal, in Germany. A similar effect has since

appeared in the miners working deposits in the south-western USA after the Second 

World War. Inadequate ventilation and insufficient expenditure on mine safety were 

blamed for the lung cancer deaths of over 100 American uranium miners; out of a total of 

some 6000 men who had worked in American underground uranium mines in the boom 

years of the 1950s the US Public Health Service estimated at the end of the 1960s that 

from 600 to 1100 would die of lung cancer because of radiation exposure on the job. In 

Canada the Royal Commission on the Health and Safety of Workers in Mines - the Ham 

Commission, after its head  - in 1976 devoted an entire chapter of its long report to 'Lung 

Cancer and Ionising Radiation in the Uranium Mines', making twenty-three substantive 

recommendations about desirable improvements.

Uranium mine tailings also present a problem. The military rush for uranium in the USA 

led to accumulation of vast piles of tailings; estimates range as high as 90 million tonnes, 

much of this piled on riverbanks in the south-western USA. The consequent radioactive 

pollution of waterways has represented a serious problem; at one stage inhabitants 

downstream in the Colorado river basin were exposed, through their drinking water, to 

three times the ICRP maximum permissible intake of radium - which is a bone-seeking 

radio-nuclide even more dangerous than strontium-90. Canadian tailings piles likewise 

caused concern; indeed, the Ontario Royal Commission on Electric Power Planning  - the 

Porter Commission  - observed in 1978 that uranium mine tailings might be much the 

most serious long-term radioactive waste problem. Dry tailings are still blown freely by 

the wind across many inhabited areas of the southwestern USA. The tailings piles will 

remain dangerously radioactive for tens of thousands of years.

Meanwhile it was discovered in the 1960s that the sandy tailings had been used as fill 

beneath the foundations of many buildings in many communities, notably Grand

Junction, Colorado: buildings including homes, schools and hospitals. The radon gas 

emanating from these building structures into the air indoors now exposes the local

inhabitants - including children - to exactly the same radon daughter-products which have 

already been responsible for thousands of lung-cancer deaths of miners from

Joachimsthal to Grand Junction. Sorne government aid has been available to reconstruct 

these radioactive buildings; but many residents would prefer that nothing be said about it, 

because of the effect on property values.

Uranium Enrichment59

As indicated in Chapter 1, the fissile uranium-235 nuclei in natural uranium  - 7 out of 

1000 nuclei - are too dilute to support a chain reaction. Their effectiveness can be 

increased by interspersing the uranium fuel with a moderator to improve the neutron 

economy, as already described. Alternatively, or in addition, it is possible to increase the 

proportion of uranium-235 nuclei in the material. This process is called 'uranium

enrichment'. Indeed, for weapons applications, it is possible, and often necessary, to have 

uranium which consists almost wholly of the 235 isotope; such weapons applications use 

uranium which is at least 90 per cent uranium-235.

Bringing about this increase in the concentration of the 235 isotope is not, however, easy. 

Chemically it is very difficult; in chemistry uranium-235 and uranium-238 are virtually 

identical. Only their minute difference in mass - 3 units in 235 - can be used as a basis for 

separation. There are several physical phenomena in which this minute mass difference 

produces a measurable difference in behaviour between the two isotopes. Of these the 

phenomenon of earliest large -scale interest was the rate of diffusion through a thin porous 

membrane. The lighter uranium-235 diffuses just slightly more swiftly through such a 

membrane. this effect is the basis for what are arguably the largest industrial

establishments in the world, the 'gaseous diffusion plants'. There are three such plants in 

the USA; one in the UK, at Capenhurst in Cheshire, now shut  down; two in France, at 

Pierrelatte and Tricastin; two in the Soviet Union. and at least one in China.

The details of gaseous diffusion technology are  - because of their military implications  -

still to a considerable extent secret. The basis of a gaseous diffusion plant is very simple: 

a metal-walled cell, with a thin membrane of porous metal dividing it in two. (Fabrication 

of such membranes, which must withstand lateral pressures and chemical corrosion while 

providing a selective diffusion barrier, is one area where much detail is still not publicly 

available.) In order to utilize the different diffusion rates of the two uranium isotopes, it is 

necessary to convert the original yellowcake, solid uranium oxides, into uranium

hexafluoride, UF6. This compound, called 'hex' for short, is the simplest compound of 

uranium which can be easily vaporized. Furthermore, fluorine has only one stable

isotope; so the different diffusion rates of hex molecules will depend only on the 

difference between the uranium isotopes involved. It must be added that hex is a

viciously corrosive, reactive gas, requiring very careful handling and high-quality 

metallurgy in the vessels through which it travels.

Under controlled pressure hex enters one chamber of a diffusion cell. It diffuses through 

the membrane into the other chamber, the lighter 235 isotopes diffusing slightly faster 

than the heavier 238 isotopes. In a given cell the concentration of 235 can be increased, 

however, only by about one part in a thousand. Accordingly, the diffusion process must 

be repeated thousands of times. A cascade arrangement is set up. Gas from the

high-pressure chamber of a cell, slightly depleted of the 235 isotope, is piped back to 

earlier cells; gas from the low-pressure chamber, slightly enriched in the 235 isotope, is 

piped onwards to later cells. By this means, using thousands of pumps and condensers, it 

is possible to raise the proportion of 235 isotope to more than 99 per cent. Since pumping 

heats the gaseous hex the plant must also include large-scale cooling systems.60

The uranium whose share of 235 nuclei has been reduced is called, as mentioned earlier, 

depleted uranium. One factor affecting the performance of a gaseous diffusion plant is the 

'tails assay' - the level at which the percentage of 235 is so low that it is no longer worth 

trying to extract any more 235 from the hex. This tails assay is usually somewhere 

between 0.2 and 0.3 per cent 235, compared with 0.7 per cent in natural uranium. If the 

depleted hex is discharged from the plant when it still contains 0.3 per cent 235, more 

yellowcake will be required to produce a given amount of uranium enriched to a given 

level; on the other hand if the depleted hex is not discharged until its tails assay is down 

to 0.2 per cent 235, part of the plant operates with very depleted hex, from which it is 

even more difficult to extract a useful amount of 235.

In the early stages of enrichment, diffusion cells must be comparatively large; the 

desirable uranium-235 nuclei are accompanied by a compa ratively cumbersome cloud of 

uranium-238 fellow-travellers. As the proportion of 235 increases, the total mass of hex 

which must pass through successive cells decreases, the high-enrichment end of the plant 

uses comparatively small cells, in which only the 235 nuclei remain, with a few stragglers 

of 238. For this reason the early stages of enrichment, to 3 or 4 per cent uranium-235, 

require as much pumping as all the stages from this level onwards. The effort expended 

in the enrichment process is measured  in units of 'separative work'; the throughput 

capacity of the plant is measured in units of 'separative work per year.' Separative work is 

loosely correlated with the total energy required to carry out an operation  - energy to run 

pumps, etc. In general a comparatively large amount of hex enriched to a few per cent 

requires the same amount of separative work as a comparatively small amount of hex 

enriched to 90-plus per cent.

All the first generation of gaseous diffusion plants were built under military aus pices. 

Their electrical requirements are awesome. The Oak Ridge plant, in full operation,

requires some 2000 megawatts of electricity - enough to power a sizeable city.

(Electricity for the Oak Ridge plant is largely provided by fossil-fuel power plants 

burning strip-mined coal, a nicely ironic touch.) A gaseous diffusion plant likewise takes 

up an impressive area, as much as half a square kilometre. However, because of the 

differences between the low-enrichment end of the plant and the high-enrichment end,

such plants are not easy to convert from production of strictly military weapons-material, 

involving enrichment to more than 90 per cent uranium-235, to production of fuel for 

power reactors. Water reactors and AGRs require for their fuel a level of enrichment of 

only 2 to 4 per cent. In consequence, although the first generation gaseous diffusion 

plants are in theory available to service present power reactors, other approaches to

enrichment are now attracting attention.

France, in partnership with several other countries under the corporate name of Eurodif, 

built a large diffusion plant at Tricastin, designed especially for production of reactor fuel 

and different in detail from the military plant at Pierrelatte. France also planned a further 

diffusion  plant to sell enrichment services to foreign customers; but this consortium, 

called Coredif, came a cropper with the fall of the Shah of Iran, whose government had 

put up 20 per cent of the investment capital for the plant. The consequent financial 61

dispute remains unresolved.

Meanwhile an alternative enrichment technology is making its first contribution to the 

present-day nuclear fuel cycle. Like the gaseous diffusion process this alternative requires 

thousands of stages in cascade; the stages this time consist of gas centrifuges. When 

uranium hexafluoride gas enters a spinning centrifuge, the uranium-238 hex molecules 

tend to drift to the outer perimeter of the centrifuge chamber, leaving the lighter uranium-

235 hex molecules closer to the axis of the chamber. Piping channels the axial hex, 

slightly enriched, onwards to successive centrifuges, and the perimetral hex, slightly 

depleted, backwards  - just as in the cascades of a diffusion plant. It is claimed that the 

centrifuge method consumes only one tenth of the energy required for diffusion, a major 

advantage for the centrifuge approach. The Almelo Treaty in 1970, signed by the UK, the 

Federal Republic of Germany and the Netherlands, created two tripartite consortia, each 

country holding a one-third partnership: URENCO, to sell enrichment services, and 

CENTEC, to manufacture the hundreds of thousands of centrifuges required. URENCO 

centrifuge enrichment plants at Capenhurst in Cheshire and at Almelo in the Netherlands 

delivered their first separative work  in the late 1970s. Both plants continue to expand, as 

new contracts require. A third plant is planned for Gronau in the Federal Republic.

Other techniques are also under development. One is based on deflection of gas sprayed 

from a nozzle: the lighter hex 235 molecules are more easily deflected. A small prototype 

nozzle plant has been built in the Federal Republic, and a larger plant employing the 

same principle is in operation under conditions of great secrecy at Valindaba in South 

Africa. Undoubtedly the most exotic technique for isotope separation is based on lasers. 

A laser can be tuned so finely that its radiation ionizes uranium-235 atoms or hex

molecules while not ionizing uranium-238 atoms or hex molecules. It is then necessary 

somehow to use the ha ndle provided by the electric charge on the ionized atoms or 

molecules to sift the 235 atoms or molecules out of the cloud. Major research

programmes are known to be under way at the US nuclear weapons laboratories at Los 

Alamos and Livermore, at Harwell in the UK, and possibly in Israel and elsewhere. The 

curtain of secrecy around laser enrichment is akin to that surrounding the Manhattan 

Project itself; but oblique reports indicate that the technology is proving entirely feasible. 

The secrecy is all too understandable: unlike the other technologies mentioned, a single 

stage of laser enrichment could conceivably bring about almost complete separation of 

fissile uranium-235 and non-fissile uranium-238, offering an alarmingly short cut to 

weapons-material, even from uranium ore.

Heavy Water Production

Uranium isotopes are not the only ones requiring separation for nuclear applications. At 

the other end of the table of elements come the isotopes of hydrogen - of which the 

second, deuterium, is the best neutron moderator of all. The American and Canadian 

reactor designs offer in this context a tidy contrast: whereas the Americans enrich the fuel 

and take the moderator as it comes, the Canadians take the uranium as it comes and, so to 

speak, enrich the moderator.62

The difference in mass between an ordinary hydrogen nucleus and a nucleus of heavy 

hydrogen or deuterium is proportionally very large; a deuterium nucleus is about twice as 

massive as a nucleus of ordinary hydrogen. As a result certain types of chemical

interchange can be used to separate the light and heavy hydrogen nuclei. The

Girdler-Sulphide (GS) process now in large-scale use employs the two chemically similar 

molecules water and hydrogen sulphide. The former consists of two hydrogen atoms 

bonded to an oxygen atom, the latter of two hydrogen atoms similarly bonded to a

sulphur atom. In a mixture of water molecules and hydrogen sulphide molecules, the 

distribution of the hydrogen isotopes between the oxygen and sulphur atoms depends on 

the temperature. At low temperatures - about 25 C - the liquid water contains

proportionally more deuterium than it does at higher temperatures  - about 100 C. This 

shift of equilibrium can be used to transfer deuterium atoms out of one batch of water and 

into another, using hydrogen sulphide as a sort of conveyor belt.

First the water and hydrogen sulphide are mixed together at the lower temperature; 

deuterium shifts from hydrogen sulphide into water. Some of the enriched water is led off 

for further enrichment. The rest is fed into a tower at the higher temperature; deuterium 

now shifts from this water into the hydrogen sulphide. This enriched hydrogen sulphide 

in turn shuttles back to enrich more water. The depleted water can he discarded, and the 

enriched water fed onward through a cascade, successively boosting its percentage of 

deuterium.

Whereas the enrichment of uranium becomes easier the higher it gets, as far as mass 

transport is concerned, the enrichment of water gradually gets too cumbersome, as the 

water-hydrogen sulphide exchange reaction becomes inefficient. By this time it is

however possible to carry out fractional distillation, utilizing the significantly higher 

boiling temperature of deuterium oxide  - about 1014 C - to boil away much of the 

remaining ordinary water. Electrolysis can refine this to a final composition of 99.75 per 

cent deuterium oxide; by this stage electrolysis is a comparatively inexpensive and

efficient way to dispose of the remaining ordinary hydrogen.

There are perhaps a dozen heavy water production plants in all in the USA, Canada, 

France, India and elsewhere. The continuing international interest in heavy water reactors 

- and the high efficiency of such reactors for plutonium production - seem likely to 

continue to keep current production capacity, of the order of 300 tonnes of deuterium 

oxide per year per plant, fully occupied. On the other hand, heavy water is intended to be 

a permanent part of a reactor system; unlike enriched fuel it is not 'consumed'. Once a 

reactor is equipped with its operating complement of heavy water its only requirement 

from then on is enough to replace losses in refuelling and inevitable leakage. Since heavy 

water now costs upwards of £100 per kilogram operators strive to minimize such losses.

Fuel Fabrication

Fabrication of fuel for reactors is now a major  - and complex  - industrial process in its 

own right. In Chapter 2 we encountered some of the determinants affecting reactor fuel 63

and its cladding: case of heat removal, durability when subject to radiation damage, 

chemical stability, and physical and mechanical properties which lend themselves to

economical fabrication. An additional requirement, at every stage, is establishing and 

maintaining high purity in the materials, to keep them free of neutron-absorbing 

impurities. Fuel fabrication facilities accordingly strive to carry out the relevant industrial 

processes in conditions of cleanliness like those of an operating theatre.

Among the present range of power reactors the only large ones using uranium  metal fuel 

are the UK Magnox reactors, and their French cousins. The awkward metallurgy of 

uranium has already been mentioned. Nevertheless uranium can be fabricated by

common metalworking techniques.

The uranium fuel of water-cooled reactors - PWRs, BWRs, and CANDUs - is in the form 

of uranium dioxide. Uranium dioxide powder is made from uranyl nitrate solution, which 

may originate either in a uranium mill (from natural uranium), an enrichment plant (from 

enriched uranium hexafluoride), or a reprocessing plant. Fabrication by powder

techniques is employed to form the desired shapes  - for instance the short cylindrical 

pellets described in Chapter 2. Baking at high temperature produces stable, dense pellets -

the denser the better; high density facilitates  the chain reaction, improves the generally 

poorer thermal conductivity, and also helps to retain the gaseous fission products which 

accumulate in the fuel material.

If uranium metallurgy is awkward, plutonium metallurgy is positively fiendish. The metal 

occurs in six different crystal phases, whose properties change drastically with

temperature; two phases even contract, rather than expanding, as temperature increases. 

Its thermal conductivity is low, its melting point is low, it oxidizes violently on contact 

with air, and - when its peculiarly vicious radio-toxicity is added to the mix - all in all it is 

a material without many redeeming virtues. Of course, in the one -shot chain reaction of a 

fission bomb most of these problems are overcome in a flash. But for the controlled chain 

reaction in a reactor the choice falls not on the metal but on the dioxide.

Its fabrication is a much more demanding process than that of uranium, requiring much 

tighter control of quantities and ancillaries. Precautions must prevent not only the escape 

of the toxic material but also inadvertent juxtaposition in undesirable geometries; unlike 

most power reactor uranium, power reactor plutonium is mostly fissile nuclei, and may 

easily come together in quantity in such a way as to achieve criticality. The consequent 

barrage of neutrons and gamma rays could cause serious injury or death to anyone near 

by. This is particularly hazardous in the case of aqueous solutions of plutonium

compounds, since the water acts as a moderator. However, with appropriate precautions 

plutonium dioxide can be processed like uranium dioxide; indeed, the two oxides, mixed 

in suitable proportions, are an effective fuel material.

Fuel  rods or fuel pellets, once fabricated, are clad as described in Chapter 2, and where 

appropriate arranged in assemblies for transport to the reactor.

Transport64

One of the major advantages claimed for central electricity generation with a nuclear heat 

source is the relatively small bulk and mass of fuel and waste that must be transported to 

and from the station. A fossil-fuel station requires so much coal or oil that it is

economically advisable to situate the station near the fuel supply. A nuclear power 

station, on the other hand, requires at most one or two shipments a week; furthermore the 

shipments away from the station are far more massive and bulky than those to the station. 

Fresh fuel elements, only minimally radioactive, can be and are shipped in ordinary cases 

like any other cargo. But once irradiated they must be heavily shielded, so that a 

shipment of two tonnes of irradiated fuel requires a fifty-tonne steel shipping cask.

Fresh reactor fuel and fuel materials are shipped by rail, by road, by water and by air, in 

increasing quantities every year, between various parts of the fuel cycle. Apart from the 

usual protections against low-level radioactivity around the shipping cases, the main

technical consideration is to guard against the stacking of cases so close together that the 

aggregation of fissile material can reach criticality. Elaborate codes of practice are 

published to provide appropriate technical guidelines.

The shipping of irradiated fuel is something else again. The irradiated fuel must be 

handled remotely at every stage. For short journeys irradiated fuel elements usually travel 

in massive water-filled casks which are vaned on the exterior to help dissipate the decay 

heat. For longer journeys, especially those by sea, the casks must be coupled to cooling 

circuitry. Since an accident involving a shipment of irradiated fuel could release

dangerous amounts of radioactivity, shipping casks must pass severe tests, typically a 

thirty-minute fire after a ten-metre fall.

Spent Fuel

One feature distinguishes nuclear power technology from all others: the left-overs. Unlike 

the ash, say, from a coal-fired power station, the used fuel from a nuclear power station 

contains both potentially valuable material and uniquely troublesome waste. Recall that 

the first large reactors were built expressly so that, under neutron  bombardment, the 

uranium-238 in the fuel would be transmuted into plutonium-239. This plutonium had to 

be recovered from the fuel. When the first power reactors came into operation in the mid 

1950s, uranium was still in short supply; so it made sense to recover also the unused 

uranium-235 which was still left in the fuel, after poisoning of the chain reaction by 

fission products and other effects had made it necessary to remove the fuel from the 

reactor. It thus became customary to assume that spent fuel from power reactors, like that 

from plutonium production reactors, should be 'reprocessed'.

A nuclear fuel reprocessing plant is a chemical plant  - but no ordinary chemical plant. 

Because its raw material, irradiated reactor fuel, is intensely radioactive, all the

operations must be carried out by remote control, behind heavy shielding. The process 

equipment must be highly reliable, and require a minimum of maintenance. Once in 

operation it is contaminated by the radioactivity, and any malfunction necessitates 65

months, or indeed years, of decontamination before it can be set right. Accordingly, the 

process line uses as few mechanical parts as possible, and depends instead on gravity 

flow and simple valves.

Different designs of fuel require different handling. The British reprocessing plant at 

Windscale was originally set up to process metal fuel elements from plutonium

production and Magnox reactors. Magnox fuel is stored in a cooling pond adjacent to the 

reprocessing plant. When ready for reprocessing it is  transferred under water, by

operators watching on closed-circuit television, into the building and up into the first of a 

series of 'caves' or 'hot cells'. The walls of the caves are of concrete some two metres 

thick, to intercept the gamma radiation from the fission products in the fuel. Once in the 

caves the fuel can be observed through special windows built into the cave walls. Each 

window is like a large aquarium, filled with a solution of a chemical such as zinc 

bromide, which is virtually transparent to visible light but strongly absorbs the very short 

wavelengths of gamma radiation.

A Magnox fuel element entering the reprocessing caves is picked up by remote control 

and dropped on a stripping machine which chops off the ends of the element and unzips 

the Magnox cladding as easily as peeling a banana. The contaminated cladding drops on 

to a conveyor belt to be transported to another building near by, which looks like an 

aircraft hangar but is actually a heavy concrete storage bin. The bare fuel rod is loaded 

into a transfer magazine and then dropped into a vat of nitric acid, which dissolves it 

ready for reprocessing. The nitric acid is mixed with a solution of an organic solvent. In 

the Windscale plant this solvent has a polysyllabic name which is unceremoniously 

abbreviated to TBP/OK. The uranium and plutonium cross over into the TBP/OK,

leaving behind about 99.96 per cent of the fission products in the water-based acid. This 

acid stream carries these fission products out of the reprocessing plant; its  subsequent 

progress will be described in the following section.

Almost all the uranium and plutonium (not quite all) are now in the TBP/OK stream, 

flowing under gravity from one section of the plant to the next. After another pass 

through a similar 'solve nt extraction' stage, to remove lingering fission products, this 

stream now encounters another water-based solution; the plutonium and uranium at this 

point part company, the plutonium returning to the water solution leaving the uranium 

behind in the TBP/OK. Various chemical manoeuvres occur en route, to bring about 

these shifts of loyalty. Eventually, after repetitions of some steps, the uranium emerges in 

one stream and the plutonium in another, as uranyl nitrate solution and plutonium nitrate 

solution respectively. The uranyl nitrate is converted back into solid uranium oxide and 

stored for possible future use; the plutonium nitrate may be likewise converted back to 

solid oxide, or may be kept in nitrate solution, ready to be made into fast reactor fuel (or 

weapons). Process streams of concentrated fissile material especially plutonium - must be 

designed to guard against accidental criticality. Reprocessing plants have elaborate alarm 

systems to warn personnel in the event of a criticality accident, which  may of course be 

quite invisible, despite the fusillade of neutrons and gamma rays.

Reprocessing oxide fuel is more difficult than reprocessing Magnox fuel, both because of 66

the design of the fuel and because spent oxide fuel, with its longer 'burn-up', is usually 

much more radioactive than metal Magnox fuel. When the Windscale plant was

modernized in the early 1960s, with the construction of the chemical separation plant 

B205, the original reprocessing plant in building B204 was converted into a 'Head End 

Plant' to prepare oxide fuel for reprocessing. During their sojourn in the reactor the 

pellets of oxide fuel swell, and wedge themselves inextricably inside their tubular

cladding. Accordingly no attempt is made to strip off the metal cladding mechanically. In 

the Windscale Head End Plant, an entire fuel element was fed gradually into a cave 

containing an awesome ram-powered shear. This shear would chop through the entire 

element consisting of perhaps more than 100 fuel pins. Each chop produced a burst of 

pulverized pellets and a barrage of centimetre-long rings of cladding which dropped into 

nitric acid. The acid dissolved the remnants of the pellets. The rings of cladding were left 

behind to be stored, with the Magnox cladding, indefinitely. The acid stream then passed 

on to the chemical separation plant in the adjacent building B205.

In September 1973 an unexpected chemical reaction in the Head End Plant led to a 

leakage of radioactivity which slightly contaminated thirty-five employees; the plant was 

shut down and eventually abandoned. The accident was caused by tiny granules of fission 

products, insoluble even in nitric acid and intensely radioactive, accumulating in a

process vessel. The handling of these granules is one of several technical problems which 

arise in the reprocessing of high burn-up oxide fuel. The fierce radiation from fission 

products tends to tear apart the molecules of solvent, especially in the first stage of 

chemical separation; and the replacement and maintenance of key components, like the 

shear for chopping up the fuel, which operate in parts of the plant filled with searing 

radiation, continues to raise questions about the cost and feasibility of reprocessing oxide 

fuel. The operators of the Windscale works, British Nuclear Fuels Ltd (BNFL) put 

forward a proposal in the mid-1970s to build a large new Thermal Oxide Reprocessing 

Plant (THORP). The resulting controversy was prolonged and bitter. Several attempts to 

reprocess oxide fuel have proved unsuccessful. The only civil oxide fuel reprocessing 

plant presently in operation is the French UP-2 plant at Cap la Hague, but the Cap la 

Hague plant too has run into technical difficulties.

Regardless of whether spent fuel is to be reprocessed or not, it must be stored for a time 

after discharge from the reactor, to allow the shortlived fission products to decay. As we 

have noted, the usual procedure had long been to discharge the spent fuel into a

water-filled cooling pond. Recently, however, partly as a result of growing doubts about 

the desirability and feasibility of reprocessing, various methods of longer-term storage of 

spent fuel have been developed. At the Wylfa Magnox station of the CEGB in Wales, 

spent fuel is discharged not into a pond but into a storage magazine cooled by carbon

dioxide  - exactly the same environment as the fuel experiences inside the reactors. 

Magnox cladding deteriorates rapidly in contact with water, becoming unreliable after a 

year or so; but it can be left in a dry gas-cooled storage magazine more or less

indefinitely. The CEGB has recently added two more dry storage magazines at Wylfa, 

cooled by natural circulation of ordinary air. The stored fuel is reported to be still sound 

and safe in storage after some years, and experts conclude that such a storage facility 

would be suitable not only for gas-cooled but also for more hardy watercooled fuel, for 67

periods extending into decades if desired. They note also that the longer fuel of any kind 

is stored, the lower is its remaining radioactivity; reprocessing, if eventually undertaken, 

would therefore be easier the longer the fuel had been stored.

Radioactive Waste

Throughout the nuclear fuel cycle, the materials involved share one common property: 

they are all to some extent radioactive, that is, they  emit radia tion. Natural radioactive 

materials are encountered in mining and milling; the materials remain radioactive

throughout enrichment, fuel fabrication and transport - but their activity is not

particularly intense. This, however, changes dramatically once the y find themselves 

inside an operating reactor: neutrons from reactor cores tend to make their entire

neighbourhood radioactive. So long as the materials in this neighbourhood remain within 

the biological shielding all is well, but radioactivity inevitably  finds a number of escape 

routes from the confines of reactors, however well buttoned up. The most important of 

these is via refuelling, when the entire radioactive inventory of the spent fuel is removed 

from the core of the reactor. We shall discuss in a  moment the eventual fate of this 

concentrated or 'high-level' radioactivity. But less concentrated radioactivity also makes 

its way directly out of an operating reactor, and must be dealt with.

Low-level Waste

Any radioactivity which emerges into the environment outside the biological shield in the 

course of routine reactor operation is called a 'running release'. The simplest kind of 

running release originates just inside the biological shield itself. In reactors with concrete 

shielding close to the core, it is desirable to keep the concrete from being exposed

directly to the heat of the core. Accordingly, except in the case of prestressed concrete 

pressure vessels, a thin layer of air is blown up the inside wall of the concrete and 

discharged to the atmosphere from a stack atop the reactor building. Some of the nuclei in 

the air absorb neutrons and become radioactive, a process called 'neutron activation'. The 

most notable of the activation products is argon-41, a radioisotope of the inert gas argon. 

Some  reactors are known to discharge hundreds of thousands of curies of argon-41 

annually. Fortunately, however, argon-41 has a very short half-life, only some 1.8 hours; 

so this apparently enormous output decays to a very low activity before drifting from the 

stack down to ground level. Other atoms in the air also become activated, but only in 

small amounts and/or for very short half-lives.

Reactor coolant may carry radioactivity out of the biological shield. Impurities in water 

or graphite moderator are susceptible to neutron activation. Carbon - from graphite 

moderator or carbon dioxide coolant or both  - can become radioactive carbon-14. But 

since normal carbon is carbon 12 the transmutation requires the absorption of not one but 

two neutrons and is accordingly infrequent. Heavy water coolant can absorb neutrons, 

turning the deuterium (hydrogen 2) into hydrogen 3, or tritium, which is radioactive. But 

the one coolant which does respond readily to neutron activation is the sodium coolant in 

liquid metal fast  breeder reactors. As already indicated it becomes sodium-24, so

intensely gamma-active that it must be kept entirely within the biological shield.68

The fuel cladding too may contribute to the activity in the cooling circuit, as the cladding 

suffers gradual  corrosion by the hot coolant. Again it is primarily a consequence of 

impurities in the cladding, which has, of course, been made as little susceptible to neutron 

absorption as possible, for reasons of neutron economy. The worst offenders in this 

category are impurities in zircaloy cladding on water-cooled reactor fuel. Corrosion of 

this cladding is enhanced by the intimate contact with the fast-moving fluid at high 

pressure, which quickly carries surface corrosion into the coolant flow.

Much more serious are leaks from the fuel cladding, to which some reactors seem prone. 

The build-up of gaseous fission products inside a fuel rod imposes increasing strain on 

the cladding; if for any reason the cladding develops a flaw the fission gas seeks it out 

and escapes into the coolant. A more sizeable leak also releases the volatile fission

products, among them the dangerous iodine-131. 'Burst can detection gear' in a Magnox 

reactor sniffs out tell-tale radioactivity in the coolant, and locates faulty fuel elements. If 

the reactor can be refuelled on load, it is possible to remove leaking fuel without a 

shutdown. If the reactor, like most water-cooled designs, can only be refuelled off load, a 

shutdown would be necessary. Furthermore fuel which is leaking slowly may be difficult 

to locate. In any case, early replacement of the fuel disrupts the fuel programme and 

distorts the planned pattern of neutron density in the core. For all these reasons leaking 

fuel is frequently left in a reactor until routine refuelling.

These effects make it necessary to decontaminate the cooling circuits of  a reactor. 

Otherwise the unavoidable leakage of radioactivity through valve seals and other

permeable points becomes a potential hazard to personnel, and may interfere with

maintenance. Decontamination is usually done routinely, by bleeding off a small portion 

of the coolant and replacing it with fresh uncontaminated coolant. Of course, each time a 

refuelling machine is coupled to a reactor vessel for on-load refuelling the machine 

acquires a share of the activity in the coolant; this activity must be discharged  - and kept 

track of  - when the machine is depressurized. In the case of carbon dioxide coolant the 

gas bled off - the 'off gas' - is passed through a variety of filters and delay stages.

Boiling water reactors, which pass the primary coolant directly through turbines, are 

especially prone to leakage of active coolant. One possible procedure in such a case is to 

provide storage tanks for coolant bled off. In such tanks the coolant activity can be 

allowed to decay for some months before it is released. Similar hold-up tanks may also be 

provided for drainage from floors, for water from sinks used in decontamination, and for 

water from the laundry in which contaminated clothing is cleaned. Wastes which require 

interim storage in hold-up tanks or the like are called medium-level wastes. The cooling 

ponds for storing irradiated fuel prior to shipment for long-term storage or reprocessing 

usually pick up activity from the exterior of the fue l elements and also from any leaking 

elements,  so this cooling water, too, has to be dealt with. A common procedure is to 

cycle the water slowly through the ponds, continually diverting a small fraction, mixing 

and diluting it with the much greater mass of cooling water discharged from the turbine 

condensers back into the river or coastal water from which it has been abstracted. Water 

treatment systems may also use ion exchangers and other standard forms of purification 69

installations to collect and segrega te radioactive compounds in sludges.

In the course of everyday business in a reactor plant a certain amount of solid material 

also becomes contaminated with radioactivity - floor mops, paper towels, broken

glassware from sampling labs, etc. The volume of contaminated solids ought to be 

reduced by, say, incineration; but at present they are usually simply buried in designated 

burial grounds, or dumped at sea in prepared containers, as are active ion-exchange 

sludges.

All of the radioactivity which reaches the outside world directly from a reactor

installation by these routes may be lumped together as 'low -level' radioactivity. Until the 

mid-1970s it was not looked upon as much of a problem. Since then, however, two 

aspects of low-level waste have begun to attract concern. One is the sheer bulk of dry 

low-level waste; adequate dumping sites, away from ground-water, and able to isolate the 

radioactivity, are proving more and more difficult to find. More worrying is the presence 

in some low-level wastes, both  solid and liquid, of traces of plutonium and other

biologically hazardous transuranic elements. Removal of these traces would be

troublesome and prohibitively expensive; but they may nevertheless make the wastes too 

dangerous to release uncontrolled into the environment. Such wastes -  in the UK called 

PCM, plutonium-contaminated material, and in the USA called TRU for transuranic  -  

seem likely to become more of a problem, especially if reprocessing and the fast breeder 

reactor become more common.

Decommissioning

Just as reactor fuel eventually becomes no longer usable, so too the reactor itself will in 

due course, for one reason or another, be shut down permanently. Unlike a fossil-fuelled 

power station, however, a reactor cannot simply thereafter be dismantled and the ground 

cleared for future use. As indicated earlier, some parts of the reactor - the solid moderator 

and other core materials in gas-cooled reactors, the pressure vessel and possibly other 

parts of the primary cooling circuit, possibly also the concrete biological shielding, and 

possibly also the spent fuel cooling pond  - will be contaminated with radioactivity. This 

complicates both the process of dismantling and the disposal of the resulting remains. 

The question has recently, albeit belatedly, received a great deal of study. The general 

consensus is that such 'decommissioning' of a reactor can be carried out in three stages. 

The first and least expensive stage would involve removal of the spent fuel and drainage 

of the cooling circuits, leaving all the fixed structure of the power station in place; the 

reactor building would be locked and the plant isolated from access by the casual public, 

probably under physical surveillance of some kind to prevent unauthorized entry. The 

second stage would involve removal by dismantling and demolition of all the fixed

structure of the power station except the reactor itself. The third stage would complete the 

process by removing the reactor itself  - pressure vessel, core internals, pipework, steam 

generators, the lot - and clearing the remaining concrete away to leave a site available for 

any future use desired: anything from building a new nuclear power station to growing 

Brussels sprouts.70

Such, at least, is the theory of decommissioning. Unfortunately,  it remains to a large 

extent unsupported by practice. No one anywhere has ever decommissioned a large 

power reactor after its normal working life. Small research reactors have been

decommissioned, and so have a handful of experimental power reactors; but the power 

reactors which have been decommissioned are, not surprisingly, those which did not long 

remain in service after start-up. Their core materials had received only brief exposure to 

neutron radiation, and had little opportunity to accumulate radioactive contamination. 

Even so, the task of decommissioning them has been demanding.

The first power reactor to be decommissioned after even moderate service was the Elk 

River reactor in Minnesota, a 22-MWe BWR which operated only intermittently from 

1964 until its permanent shutdown in 1968. The first and second stage decommissioning 

was more or less straightforward, but removal of the reactor vessel itself required some 

striking imagination. In order to be able to cut apart the heavy steel vessel, without 

exposing workers to any more radiation than necessary from activation products in the 

steel, the demolition team filled the vessel with water and lowered divers into it with 

underwater cutting torches like those normally used on offshore oil installations.

Needless to say the entire exercise was not cheap. Whether it would be feasible at all for 

a much larger reactor, after much more radiation exposure of the materials, is as yet 

unclear. Further important information about techniques and costs will be gathered from 

the decommissioning of the Shippingport reactor  - the first nuclear power plant in the 

USA - which began in 1981. Until someone decommissions a full-scale power reactor, 

the problems will remain uncertain - and the costs frankly 'guesstimates'. Some 

guesstimates have put the cost as high as the original cost of the reactor; others are much 

more optimistic. It seems fairly safe to assume that no one will be in a hurry to embark 

even on second-stage decommissioning, much less third-stage, any sooner  than 

absolutely necessary. There will undoubtedly be a good many mothballed nuclear power 

stations dotted about the landscape in the early twenty-first century.

High-level Waste

Because running releases are as a rule dilute and not very radioactive it is  often said that 

reactors discharge very little radioactivity to their surroundings. This is, strictly speaking, 

true; but it is slightly misleading. Almost all the radioactivity which leaves a reactor does 

so within the used fuel removed from its core. Since the reactor has created virtually the 

whole of this radioactivity, it is somewhat special pleading to imply that the radioactivity 

thenceforth bears no relation to the reactor. On the contrary: the radioactivity from used 

reactor fuel is one of the most challenging problems posed by operation of nuclear 

reactors.

When a fresh fuel element enters a reactor it is as sleek and glossy as a surgical implant. 

When it emerges again after radiation it is discoloured, possibly even swollen, caked with 

what the nuclear engineers bluntly call 'crud'. Inside the cladding the fuel now contains 

unused uranium-235 and  -238, plus a wide assortment of other nuclei created by the 

fission reactions, neutron absorption and radioactive decay: uranium-237, plutonium-239,  

-240,  -241 and   -242, americium-241 and other so-called 'actinides', and literally71

hundreds of different fission product nuclei and their decay and neutron-activation 

products, including krypton-85, strontium-89 and   -90, iodine-129 and   -131, and

caesium-137. Some of these species have short half-lives., while the irradiated fuel sits in 

the cooling pond or travels from the reactor to long term storage or a reprocessing plant, 

the short-lived isotopes like uranium-237 and iodine-131 decay to insignificance. After, 

say, a hundred days of cooling, its radioactivity arises mainly from radioisotopes of only 

about a dozen elements.

If the spent fuel is left in long-term storage the decay processes continue and the 

radioactivity, still confined, continues to decrea se, albeit less rapidly. If, however, the 

fuel is reprocessed comparatively soon after its removal from the reactor - as is the case 

for instance with most Magnox fuel - the radioactive contents then follow various 

different paths to various different destinations. Assuming that the cladding has been 

gas-tight until it is stripped or chopped open, the gaseous fission products, particularly 

krypton-85, thereupon emerge into the atmosphere of a hot cell. Krypton-85 has a 

half-life of about 10.8 years. It is a chemically inactive inert gas, and is accordingly 

difficult to recapture. Present practice is simply to discharge it from a stack into the 

outside air. (Until 1971 US authorities regarded the amount of krypton-85 discharged 

from their reprocessing plants  as classified  - that is, secret - because it might reveal how 

much fissile material they had produced.) Radiobiologists do not consider that the

consequent gradual build-up of gamma-emitting krypton in the global atmosphere offers 

any present hazards. But if nuclear power programmes expand as widely as some

anticipate, some form of krypton retention will have to be installed in reprocessing plants 

before the turn of the century. (Liquid air already contains enough concentrated krypton-

85 to represent a minor safety hazard to users.) Some similar considerations also apply to 

iodine-129. It has a half -life of 16 million years, and is therefore not very radioactive; but 

since it is concentrated, like all isotopes of iodine, in the human thyroid any significant

build-up in the environment must be viewed with caution.

Within a reprocessing plant there is also an inevitable accumulation of low-level liquid 

and solid radioactive wastes exactly like those which collect in a reactor plant  - indeed 

probably more copious. Solid wastes are buried, or dumped at sea, as before. At the 

Windscale reprocessing plant low-level liquid wastes are discharged into the Solway 

Firth through twin pipelines emptying under water more than three kilometres off shore, 

at the rate of some 500 000 litres per day.

All such routine releases of radioactivity are carried out in accordance with standards laid 

down by national authorities, usually based on the guidelines of the International

Commission on Radiological Protection (see Appendix  B). In the UK, for instance,

discharges of radioactive effluent are monitored by the Ministry of Agriculture, Fisheries 

and Food, as well as by the dischargers themselves, to ensure compliance with standards 

imposed as conditions of operation in licences issued by, among others, the Inspectorate 

of Nuclear Installations, under national legislation. As we describe in Appendix B, such 

standards continue to be the subject of protracted controversy.

If reprocessing takes place, the fuel, apart from the gaseous  or volatile fission products 72

and the cladding - and, in some reprocessing plants, including the cladding - is dissolved 

in nitric acid for first-stage separation. The witches' brew left behind when the uranium 

and plutonium pass over into the organic solvent is called 'high-level waste'. Without 

doubt it is the most daunting waste material produced in any industrial process. At 

Windscale, the reprocessing of one tonne of fuel produces about five cubic metres of 

high-level waste  - that is, about enough to fill five or six bathtubs. The waste contains 

nitric acid, fission products which are both thermally hot and intensely radioactive, iron 

from corrosion of plant vessels, chemical impurities from the original fuel, and a dash of 

carried-over organic solvent. As may be imagined, it requires delicate treatment, to avoid 

unpleasant side-effects like reactions between solvent and nitric acid at high

temperatures. The volume is reduced by evaporation - under vacuum, to keep down 

temperatures. The procedure must be carried out under careful control  - always remotely 

of course - to avoid crystallization or precipitation where it might prove embarrassing 

(such as in process lines) and to keep the fission products at a low concentration so that 

the rate of heat output does not overwhelm the cooling system.

After evaporation the concentrated waste is led to the storage facility near the main 

reprocessing plant. At Windscale this is a concrete building containing an array of special 

storage tanks, eight each of 70 cubic metres capacity and six  - so far  -  of 150 cubic 

metres capacity. The smaller tanks are fitted with cooling coils; each of the larger tanks 

has seven independent cooling circuits, external water jackets which include leak

detectors, and an internal system of agitators to prevent solids from settling on to the 

bottom of the tank. The cooling circuits on one of the larger tanks can remove up to 2 

megawatts of heat  - that is, about 13 watts per litre; this in turn limits the permissible 

concentration of the fission-product stream flowing into the tank. The temperature in the 

tank is kept about 50 C. Gradual evaporation of the water from the solution is

accompanied by gradual decrease of the radioactive heat output; evaporation can be kept 

in step with heat output to maintain sufficiently low concentration. It is also necessary to 

prevent a buildup of hydrogen, produced by the breakdown of water molecules by

radiation  - so-called 'radiolytic hydrogen'. Tanks can be interconnected, to prevent

overloading of cooling circuits with incoming waste of comparatively high output, and to 

provide transfer facilities in case of a leak. Tanks in use are permanently sealed into 

massive steel-lined concrete shielding vaults, never to be seen again. A programme of 

construction of new tanks keeps spare capacity available. At the end of 1981 the total 

volume of liquid waste stored at Windscale was about 1000 cubic metres.

Similar tank storage installations are located at reprocessing facilities in the USA, France, 

Belgium, the Soviet Union, India, China and elsewhere. The most famous - or notorious -

is at the Hanford Reservation in Washington state. Here, in 151 very large tanks, is stored 

the high-level liquid waste - nearly 250 000 cubic metres resulting from recovery of the 

plutonium from the Hanford production reactors, for the US nuclear weapons

programme. It is generally reckoned that tank storage of high-level wastes can only allow

for a useful life of twenty to twenty-five years per tank, albeit perhaps somewhat longer 

for tanks of stainless steel. Many of the Hanford tanks are ordinary carbon steel; more 

than a dozen leaks have already occurred, including at least one very large leak indeed. 

Between 20 April and 8 June 1973 tank 106T leaked some 435 000 litres of high-level 73

liquid waste into the earth beneath it, while plant personnel continued to pour more liquid 

into the tank, oblivious to the falling level recorded on measuring instruments. The  leak 

released approximately 40,000 curies of  caesium-137, 14,000 curies of strontium-90 and 

4 curies of plutonium. Investigators later declared that the radioactivity would not reach 

the level of the ground-water below the tank, but a drilling programme to locate the hot 

waste had to be curtailed, lest new drill holes facilitated the downward migration of 

radioactivity. The leak was the eleventh recorded at Hanford; it was not the last.

Clearly the hazardous life-span of some constituents of highlevel waste far outreaches 

that of a storage tank. Strontium-90 has a half-life of 28 years, caesium-137 one of 30 

years. It takes ten half-lives to reduce the radioactivity of a sample a thousandfold 

(1/2 x 1/2 x 1/2 x 1/2 x 1/2 x 1/2 x 1/2 x 1/2 x 1/2 x 1/2 equals 1/ 1024). Accordingly, it 

takes about 300 years for the radioactivity of 1  curie of strontium-90 or caesium-137 to 

drop to 1 millicurie. The high-level waste from 1 tonne of irradiated fuel includes about 

100 000 curies of each. A 1000-MWe PWR produces at least 25 tonnes of irradiated fuel 

per year  - that is, well over 2 million curies of strontium-90 and another 2 million curies 

of caesium-137. Some 300 years hence this particular contribution will have dwindled by 

a factor of a thousand, to only 2000 curies of each: except that 2000 curies of strontium-

90 is not very 'only'. Multiply such figures by the number of reactors now in operation, 

under construction or planned, and the magnitude of the consequent problem becomes 

numbingly apparent. Furthermore present methods do not  - for economic reasons, if not 

for technical reasons - extract all the actinides from the fission product waste;  perhaps 1 

per cent of the plutonium, with its half-life of 24 400 years, is left behind to add to the 

unpleasantness of the residue.

It is evident that such quantities of potentially dangerous radioactivity require scrupulous 

stewardship. While tank storage is regarded as a satisfactory interim measure, efforts 

continue to devise a long-term solution to the problem. To mitigate this burden

somewhat, it is now accepted that high-level waste ought at least to be in solid rather than 

in liquid form, to immobilize the waste, reducing the possibility of its spreading via leaks 

or vaporization. Although reprocessing was long considered an essential stage in the 

management of spent fuel, opinion is now divided as to its desirability. Some consider 

that spent fuel itself  - a solid structure engineered with high integrity to withstand the 

severe conditions inside an operating reactor  - might eventually prove to be the best 

available waste form for final disposal. It is also pointed out that even should

reprocessing prove in due course to be a desirable stage of waste management, it

becomes progressively easier the longer the fuel has been stored, because of the decay of 

the radioactivity. In any case, as we shall describe in later chapters, spent fuel storage is 

de facto  an important interim stage in the waste management process, since oxide fuel 

reprocessing is neither cheap nor readily available.

Be that as it may, a substantial inventory of high-level waste in liquid form already exists 

and will require appropriate treatment before final disposal. Several approaches for 

solidification are under development. In the USA, at Hanford, wastes are simply allowed 

to boil themselves dry inside storage tanks, to be  left as solid cake in the tanks. At the 

National Reactor Testing Station in Idaho, high-level waste is 'calcined' (baked at high 74

temperature) into granules like coarse white sand, which are stored in huge concreteshielded stainless steel bins underground. Another approach, favoured particularly by the 

UK and France, is to evaporate and fuse the high-level waste into dense glass  - a process 

called vitrification. The French AVM (Atelier Vitrification Marcoule) came into

operation in the late 1970s, producing pillars of borosilicate glass impregnated with

high-level waste; British Nuclear Fuels Ltd is building a pilot plant at Windscale based 

on the French technology. The glass pillars will be stored for the foreseeable future, 

pending establishment of suitable final disposal facilities in the two countries - and 

indeed elsewhere, since some of the vitrified waste will eventually be returned to

customer countries, under existing contracts.

The favourite approach to final disposal has long been insertion of solidified highlevel 

waste into a stable geological formation. Finding such an ideal formation has proved, 

however, difficult. For a time rock salt, in beds or domes, was the preferred geological 

stratum. A hole would be drilled into the floor of an underground gallery in a salt dome. 

A waste canister would be lowered into the hole -  by remote control as usual -  and loose 

salt would be poured in after it. The salt would become soft under the intense heat from 

the canister and would snuggle close around it, sealing it permanently in place and

conducting heat away at an adequate rate to keep the solid waste from melting. But 

experience has cast some doubt upon the suitability of salt. 'Salt Vault', a pilot scheme for 

salt-formation storage, was carried out in the USA in the late 1960s near Lyons, Kansas. 

But, despite earlier official pronouncements to the contrary, the site was eventually 

abandoned as unsuitable. A company mining salt on a nearby location pumped several 

thousand cubic metres of water down a drill-hole to bring up dissolved salt; but the water 

disappeared, casting doubt on the alleged impermeability of the salt formation.

Fortunately no high-level waste had yet been buried in it. The search for more reliable 

formations and locations continues - not without controversy.

It is worth noting in passing that the nuclear industry refers, as a matter of course, to 

'waste management'. It looks like a career with a future - a long future.