IOP PUBLISHING JOURNAL OF RADIOLOGICAL PROTECTION

J. Radiol. Prot. 27 (2007) 211–215 doi:10.1088/0952-4746/27/3/E02

EDITORIAL

The Windscale reactor accident—50 years on

The policy of the government of the United Kingdom to independently manufacture nuclear
weapons in Great Britain was formulated in the mid-1940s and implemented in the late-1940s
and early-1950s; full details are to be found in the monumental treatise on the subject by
Margaret Gowing and Lorna Arnold [1]. A significant component of this implementation was
the construction of a plutonium production factory on the remote coast of the then county of
Cumberland (now part of Cumbria) in north-west England. The site chosen was that of the
former ordnance factory at Sellafield, which, with its sister factory just down the road at Drigg,
had produced TNT during the Second World War. Construction began in September 1947,
and the site was renamed Windscale to avoid confusion with Springfields Works, the uranium
processing and fuel manufacturing establishment near Preston. Plutonium was initially created
in the uranium fuel of two nuclear reactors (the Windscale Piles), chemically separated from
untransmuted uranium and the waste by-products of nuclear reactions in a reprocessing plant,
and then converted to metallic form before being sent to the Atomic Weapons Research
Establishment at Aldermaston, near Reading, for machining and assembly as a weapon. The
speed with which the policy was put into practice is truly remarkable: Windscale Pile No. 1
was operational in October 1950 followed by Pile No. 2 in June 1951, the first reprocessing run
took place in 1952 and the extracted plutonium provided the explosive material in the UK’s
first nuclear weapons test (in Australia) on 3 October 1952, just five years after building work
commenced at Windscale. This speed, however, was achieved at a premium, as we shall see.
The Windscale Piles were each fuelled by 180 t of uranium metal fabricated (at Springfields
Works) into >70 000 aluminium-clad elements positioned in 3440 horizontal channels within
nearly 2000 t of graphite moderator. The reactor core was cooled by blowing a large volume of
environmental air through the channels and out of a 120 m high chimney—in contrast to power
reactors, in the Piles the heat generated by nuclear fission was purely incidental to the creation
of plutonium for military use. Although the primary purpose of the Piles was the production of
weapons-grade plutonium, the reactors were also used to generate other nuclides through the
neutron irradiation of appropriate materials fabricated as ‘isotope cartridges’ that were suitably
placed in channels within the core. Thus, the α-particle emitter 210 Po, used in combination with
9
Be as a neutron source to trigger nuclear fission chain reactions, was produced in bismuth oxide
cartridges (codenamed ‘LM cartridges’) and tritium was manufactured using magnesium–
lithium alloy cartridges (codenamed ‘AM cartridges’). (Other nuclides, such as 232 Th, 237 Np
and 59 Co, were also irradiated in the Piles at various times during their operational lives.)
These isotope cartridges depressed the neutron flux in the core, and, when uranium enriched in
235
U became available from Capenhurst Works (near Chester) in 1953, low enriched uranium
fuel was used in the Piles to counter the adverse effects of the isotope cartridges on the neutron
economy of the reactors.
The Windscale Piles posed problems to their operators throughout their service. Indeed,
even before construction was completed Sir John Cockcroft, on the basis of information
received from the USA, insisted that filters be installed to remove radioactive material
potentially present in the exhaust cooling air, which, since construction of the stacks had
already commenced, necessitated the building of filter galleries (‘Cockcroft’s follies’) towards

211
212 Editorial

the top of the chimneys. It was predicted at the design stage that occasional failures in the
aluminium cladding of fuel elements could lead to releases of fission products into the cooling
air, and radiation detectors were installed to locate channels where a ‘burst’ had occurred so
that the affected channel could be cleared of fuel before the core was contaminated. Such
bursts did occur throughout the period that the Piles operated, and kept the workforce busy. It
was also anticipated that the flow of cooling air would be sufficiently great that fuel elements
would be buffeted and might move along the channels, and steps were taken to attempt to
prevent this; but it was found that, in practice, elements were being blown out of the core,
leading to a re-design of the arrangement of elements in a channel.
The fuel element ‘blow outs’ were accompanied by other, unforeseen, events: some
elements were found to have become stranded, on discharge, in locations where the irradiated
metallic uranium fuel became oxidised in the cooling air and radioactive particles were being
released from the chimneys into the environment. The magnitude of these particulate releases
varied over the lifetime of the Piles, but they were a constant problem for the operators; these
fuel particle releases have been described in detail by Andrew Smith and his colleagues in the
June issue of this Journal [2]. Another unexpected operational challenge was Wigner energy
stored within the graphite moderator. When graphite is bombarded by neutrons, carbon nuclei
are displaced in the lattice, which, at the relatively low operating temperature of the Piles,
increased the potential energy of the graphite. This stored Wigner energy could, if released
in an uncontrolled manner, lead to localised high temperatures and the possibility of a fire.
The first Wigner energy release in the Windscale Piles took the operators by surprise, but once
the process was understood, controlled releases of Wigner energy were conducted in regular
annealing procedures. It was the ninth anneal in Pile No. 1 that led to a fire in the core during
10–11 October 1957 and the consequent release of radioactive material from the Pile chimney
that is the worst accidental discharge of radionuclides that has been experienced in the UK; a
comprehensive description of the accident has been provided by Lorna Arnold in her highly
impressive book on the subject [3].
The Windscale fire had profound political effects and the UK Atomic Energy Authority
(UKAEA) that ran the British nuclear facilities was never to be the same again. The
two Windscale Piles were permanently closed, although this did not greatly influence the
weapons production programme as eight UKAEA-owned Magnox reactors—of a much more
sophisticated design than the Piles, and which were also used to generate electricity—were
coming on-line at Calder Hall, adjacent to Windscale Works, and at Chapelcross, in southern
Scotland. An inquiry into the Windscale accident, chaired by Sir William Penney, was instituted
by the UK Government within days of the accident, and the Penney Committee submitted its
report to Government on 26 October, a remarkably short time after the accident. The Prime
Minister, Harold Macmillan, whose government was involved in delicate negotiations to re-
establish nuclear weapons cooperation with the USA, decided that just a summary of the
Penney Report should be published [4], and the full report was only made public 30 years
later (and is included as an appendix in Lorna Arnold’s book [3]). A committee chaired by
Sir Alexander Fleck then investigated the wider implications of the accident, which led to,
among other things, the establishment of the National Radiological Protection Board (NRPB)
in 1971 (since 2004, subsumed within the Health Protection Agency as the Radiation Protection
Division).
The Penney Committee guardedly concluded that an uncontrolled localised release of
Wigner energy during the ninth anneal had led to a fire in a fuel element that had then spread
to involve about 10 t of uranium. At the time, some senior and experienced people in the
UKAEA expressed their doubts over this explanation, and pointed to evidence that an AM
cartridge (made of magnesium–lithium alloy) was likely to have been the initiator of the fire
Editorial 213

[3]. Evidence that accumulated after the Penney Inquiry and which was presented to the Fleck
Committee, such as the seriously damaged AM cartridges that were removed from Pile No. 2
in 1958, tended to support this alternative view; but one gains the impression that the somewhat
battered UKAEA wanted to ‘move on’ after the accident, and that the cause of the accident
as identified by the Penney Inquiry should be regarded, if at all possible, as ‘the final word’.
Whatever the actual cause of the fire, it is difficult to disagree with Lorna Arnold’s view that
the operation of the Windscale Piles was ‘an accident waiting to happen’ [3].
The first reports of the activities of radionuclides released during the accident, and of
their travels, were published during 1958–59. It was clear from these reports that the primary
radiological hazard arose from 131 I, although the major emissions of other fission products were
also quantified. Three reports [5–7] made mention of the release of 210 Po (from the affected LM
cartridges), although no quantification of the activity discharged was offered, and no reference
was made to tritium having been released (from the affected AM cartridges) although this
was likely to have been of relatively minor radiological significance. Given the sensitivity
surrounding the fire and, in particular, the involvement of the LM and AM cartridges, it may
be that the 210 Po discharge was only acknowledged because it was known that the radionuclide
had been detected in the Netherlands [8]. The release of 210 Po was not even mentioned in the
next official report, published in 1960, of the environmental aspects of the accident [9]—an
omission that Lorna Arnold describes as ‘incomprehensible’ [3]—encouraging the inference
that the authorities did not want to unnecessarily shine a spotlight on difficult issues that
might conveniently be considered ‘closed’. Without doubt, the high security classification
assigned to the production of weapons materials at that time, together with the ‘need-to-know’
principle, would have offered little assistance to any comprehensive ‘external’ investigation of
the radioactive materials discharged during the fire. Against this conspiratorial interpretation
is the unclassified UKAEA report published in 1959 [10] which examined the α-activity found
on air filters at Windscale, at the Harwell nuclear research establishment (south of Oxford) and
in Belgium and concluded that this was principally due to 210 Po, and a further Harwell report
written in 1961 and declassified in 1962 [11] which makes extensive reference to 210 Po activity
in air concentrations measured in the UK and the rest of Europe using data gathered under the
auspices of the Advisory Committee on Nuclear Radiation of the International Geophysical
Year (IGY; July 1957 to December 1958). These two UKAEA documents make the failure to
estimate the magnitude of the 210 Po activity discharged during the accident in reports published
during the years immediately following the fire even more perplexing.
J R Beattie [12] and Roger Clarke [13] later re-evaluated the activities of the fission
products released from the uranium fuel during the accident, but the next thorough examination
of the quantities of all the radionuclides emitted during the Windscale fire was conducted
almost a quarter of a century after the accident by Arthur Chamberlain of Harwell [14], who
was heavily involved in the original assessment of the environmental impact of the accident.
In addition to fission product activities, Chamberlain quantified the releases of 210 Po and 3 H.
Unfortunately, Chamberlain’s report relied on some material which was still classified at that
time, so that his report was also classified (it was declassified in 1983) and not known to
Malcolm Crick and Gordon Linsley, two scientists from the NRPB who were investigating the
risks to public health posed by the Windscale accident. As a consequence, their first assessment
[15] did not consider 210 Po, a fact that was pointed out by John Urquhart [16]. In an extension
of their original study, Crick and Linsley [17, 18] examined the risks resulting from the release
of both 210 Po and 3 H, as well as a number of minor radionuclides. Interestingly, Crick and
Linsley [18] concluded that although the risk of thyroid cancer from exposure to 131 I remained
the greatest radiological impact of the fire, the predicted health effects of exposure to 210 Po
came in a close second. Roger Clarke [19], using updated cancer risk coefficients, estimated
214 Editorial

that the accident had caused, or would cause, ∼100 fatal cancers (of which <10 are thyroid
cancers due to exposure to 131 I and ∼70, mainly lung cancers, are due to exposure to 210 Po)
and ∼90 non-fatal cancers (of which ∼55 are thyroid cancers due to exposure to 131 I and ∼10
are due to exposure to 210 Po)—the release of the now notorious polonium-210, which was
largely ignored in early environmental assessments, was considered by Clarke to have had the
greatest radiological impact of the radionuclides discharged during the Windscale accident.
Recently, John Garland, the late Arthur Chamberlain’s long-time colleague at Harwell, has
refined the estimates of the quantities of radionuclides released during the fire [20], using
original documents and information on the travel of radioactive material provided by the Met
Office using the NAME atmospheric dispersion model and detailed meteorological data for
October 1957 [21].
The half-century that has elapsed since the Windscale fire has provided some perspective
on the accident—the quantity of 131 I released was 1000 times less than that released from the
Chernobyl accident almost 30 years later. Nonetheless, the Windscale accident can hardly be
considered as trivial—it is rated as a Level 5 accident on the International Nuclear Event Scale
(INES) [22]—and it could have been a lot worse. The extensive environmental monitoring that
took place during and after the Windscale fire provided the evidence upon which the authorities
decided that a milk distribution ban should be enforced in the west Cumbrian coastal strip
running from 10 km north of Windscale Works to some 20 km to the south. Iodine-131 had
been quickly identified as the major radiological hazard arising from the accident, although
the health physicists had little guidance available as to what constituted an acceptable limit
for the level of 131 I activity in milk, and they derived, essentially from first principles, such a
limit (0.1 µCi/L) to constrain thyroid doses, particularly to infants and young children. A milk
ban based on these ad hoc calculations was a courageous but wise decision, which prevented
a significant enhancement of the local collective thyroid dose and limited individual thyroid
doses. The environmental monitoring programme was described in detail by John Dunster
and his UKAEA colleagues from Windscale, Huw Howells and Bill Templeton, at a large
international conference organised by the United Nations and held in Geneva in 1958 [7]; but
this conference paper is not now readily accessible. Hence, it has been decided to reproduce
the paper in this issue of Journal of Radiological Protection as a tribute to the substantial
efforts of John Dunster, Huw Howells, Bill Templeton and their many co-workers to swiftly
understand the potential radiological consequences of the fire and, where possible, limit its
impact. This reproduction has been made possible by the goodwill of the United Nations (and
the good offices of Malcolm Crick) and Rose Dunster, John Dunster’s widow, to whom thanks
are due.

References

[1] Gowing M and Arnold L 1974 Independence and Deterrence: Britain and Atomic Energy 1945–1952 2 vols
(Basingstoke: Macmillan)
[2] Smith A D, Jones S R, Gray J and Mitchell K A 2007 A review of fuel particle releases from the Windscale Piles,
1950–1957 J. Radiol. Prot. 27 115–45
[3] Arnold L 1995 Windscale 1957. Anatomy of a Nuclear Accident 2nd edn (Basingstoke: Macmillan)
[4] Atomic Energy Office 1957 Accident at Windscale No. 1 Pile on 10th October, 1957. Cmnd. 302 (London: Her
Majesty’s Stationery Office)
[5] Stewart N G and Crooks R N 1958 Long-range travel of the radioactive cloud from the accident at Windscale
Nature 182 627–8
[6] Crabtree J 1959 The travel and diffusion of the radioactive material emitted during the Windscale accident Q. J.
R. Meteorol. Soc. 85 362–70
[7] Dunster H J, Howells H and Templeton W L 1958 District surveys following the Windscale incident, October
1957 Proceedings of the Second United Nations International Conference on the Peaceful Uses of Atomic
Editorial 215

Energy (Geneva, 1–13 September 1958). Volume 18: Waste Treatment and Environmental Aspects of Atomic
Energy (Geneva: United Nations) pp 296–308
[8] Blok J, Dekker R H and Lock C J H 1958 Increased atmospheric radioactivity in the Netherlands after the
Windscale accident Appl. Sci. Res. 7 150–2
[9] Loutit J F, Marley W G and Russell R S 1960 The nuclear reactor accident at Windscale—October, 1957:
environmental aspects The Hazards to Man of Nuclear and Allied Radiations: A Second Report to the Medical
Research Council. Cmnd. 1225. Appendix H (London: Her Majesty’s Stationery Office) pp 129–39
[10] Crooks R N, Glover K M, Haynes J W, Osmond R G and Rogers F J G 1959 Alpha Activity on Air Filter Samples
Collected After the Windscale Incident. UKAEA Report AERE-R2952 The National Archives Catalogue Ref.
AB 15/6193 (Kew: The National Archives)
[11] Stewart N G, Crooks R N and Fisher E M R 1961 Measurements of the Radioactivity of the Cloud from
the Accident at Windscale: Data Submitted to the I.G.Y. UKAEA Report AERE-M857 (Declassified 1962)
(Harwell: UK Atomic Energy Authority)
[12] Beattie J R 1961 An Assessment of Environmental Hazard from Fission Product Releases. UKAEA Report
AHSB(S)R9 (Declassified and reprinted in 1963 as UKAEA Report AHSB(S)R64) The National Archives
Catalogue Ref. AB 7/11617 (Kew: The National Archives)
[13] Clarke R H 1974 An analysis of the 1957 Windscale accident using the WEERIE code Ann. Nucl. Sci. Eng. 1
73–82
[14] Chamberlain A C 1981 Emission of Fission Products and Other Activities During the Accident to Windscale Pile
No. 1 in October 1957. UKAEA Report AERE-M3194 (Revised and declassified 1983) (Harwell: UK Atomic
Energy Authority)
[15] Crick M J and Linsley G S 1982 An Assessment of the Radiological Impact of the Windscale Reactor Fire,
October 1957. Report NRPB-R135 (Chilton: National Radiological Protection Board)
[16] Urquhart J 1983 Polonium: Windscale’s most lethal legacy New Scientist 1351 873–5
[17] Crick M J and Linsley G S 1983 An Assessment of the Radiological Impact of the Windscale Reactor Fire,
October 1957. Report NRPB-R135 Addendum (Chilton: National Radiological Protection Board)
[18] Crick M J and Linsley G S 1984 An assessment of the radiological impact of the Windscale reactor fire, October
1957 Int. J. Radiat. Biol. 46 479–506
[19] Clarke R H 1990 The 1957 Windscale accident revisited The Medical Basis for Radiation Accident Preparedness
ed R C Ricks and S A Fry (New York: Elsevier) pp 281–9
[20] Garland J A and Wakeford R 2007 Atmospheric emissions from the Windscale accident of October 1957 Atmos.
Environ. 41 3904–20
[21] Johnson C A, Kitchen K P and Nelson N 2007 A study of the movement of radioactive material released during
the Windscale fire in October 1957 using ERA40 data Atmos. Environ. 41 3921–37
[22] Webb G A M, Anderson R W and Gaffney M J S 2006 Classification of events with an off-site radiological
impact at the Sellafield site between 1950 and 2000, using the International Nuclear Event Scale J. Radiol.
Prot. 26 33–49

Richard Wakeford
Editor-in-Chief