Re: Criticality Accident

[baumbaugh <baumbaug@nosc.mil>]

      RADSAFERS,

      This is "circulating"
around the Navy - unofficially of course...

      Joel Baumbaugh
      SSC-SD (baumbaug@nosc.mil)

Here is the latest summary of information about the accident in
Japan.
I    My summary
II   Dr Komura
 most recent email messages

=================================
I.  ACCIDENT SUMMARY TO DATE by 
(see emails and URLs for sources)
=================================

DISCLAIMER:  This summary is necessarily based on news articles,
others experience with fuel fabrication in the U.S., and email messages
from technical personnel who are not directly involved with the plant,
incident response, or incident investigation.  Available news
articles are mostly non-technical and often sensationalized and sometimes
contradictory, as expected.  In addition, there are some questions
arising from different nomenclatures and possible translation
problems.

For convenience, low enriched uranium (LEU) and intermediate enriched
uranium(IEU) are here respectively defined here as less than 10% enriched
and 10-60% enriched.

A criticality accident occurred at 10:35am (0135 GMT) on September 30,
>1999.  It
occurred in a nuclear fuel fabrication plant at the JCO Co. site in
Tokaimura, Japan.  The city is about 140km northeast of Tokyo, in
the Ibaraki Prefecture.

Public areas are rather close to this plant.  However, plant areas
require little or no shielding because process material is
unirradiated.  [FYI:  There is a separate, nuclear fuel
reprocessing plant in Tokaimura.  It apparently was shutdown after a
1997 fire/explosion accident.  If I remember correctly, that as a
chemical or industrial accident but it released radioactive
material.  I speculate BBC video of a building with a hole in its
roof is actually file footage of this other plant, taken during or after
that earlier accident.]

The plant had no criticality accident response plans because it was
assumed critical fission chain reactions could not occur.  [More
probably, "Emergency planning was not required because management
and authorities assumed a criticality accident was not credible, if an
appropriate criticality safetyprogram was implemented."]

This fabrication plant converts UF6 gas to UO2 powder.  It primarily
produces LEU (5%) fuel for PWRs.  However, this time the product was
18.8% enriched fuel for JOYU, a fast-breeder research reactor.  News
articles indicate IEU processing was not new but it was apparently last
conducted three years ago. Operational limits might be based on the
highest enrichment allowed, but are more likely changed when needed to
accommodate the LEU and IEU enrichments.  Apparently this was either
the first campaign, or first campaign with IEU, for two operators. 
In addition, the third operator had only a few months experience with the
subject operation, with handling IEU, or both.

The subject operation apparently was a precipitation process involving
uranium oxide and uranyl nitrate.  [One report indicates the
material is UF6 but most others specify uranium, uranium oxide, and or
uranium and nitric acid.]  Apparently uranyl nitrate is added in
batches to a sedimentation (precipitation?) tank.  Emails indicate
the tank was a vertical cylinder, with a 50cm diameter, dished bottom,
3mm stainless steel walls, and an approximate 2.5cm-thick-water cooling
jacket around the sides and bottom.

A regulator-approved manual controls plant operations.  For this
operation material must first be weighed and added to a separate, small
dissolution tank.  Apparently material is uranium oxide and nitric
acid, but it might be concentrated uranyl nitrate.  Resultant
solution is then transferred to the sedimentation vessel via plant
piping.  The batch size is operationally limited to a maximum 2.3 or
2.4kg U or, less likely, 235U.

However, two to five years ago JCO officials approved a manual change
without regulator concurrence or notification.  This change
significantly accelerates the process (3 hours vs. 30 minutes, if a
churning device [stirrer?] is used in the sedimentation tank).  It
allows operators to use steel buckets and funnels to add materials
directly to the sedimentation tank.  Some reports indicate the
buckets are 50L mop buckets, considerably larger than the aforementioned
dissolution tank.  One official indicated the change would not have
been approved because uranium oxide, nitric acid, and stainless steel
react to produce a toxic gas [can anybody tell me what the reaction and
products are?].

On Wednesday, operators poured about 9.2kg uranium from four buckets into
the sedimentation tank.  On Thursday, operators added about 6.9kg
uranium from three
buckets, triggering the critical excursion.  The uranium was
"condensed" which I
think refers to uranium oxide powder, although it might refer to highly
concentrated uranyl nitrate.  The resultant solution, or reflected
slurry, went "flash critical" [possibly prompt
critical?].  One email indicates the critical material was
approximately 370g/L uranium with, possibly, 1mole/L nitric acid.

Operators saw a blue flash.  Reportedly they felt sick
immediately.  [I?d feel sick that quickly just knowing what the blue
flash means, whether or not physical symptoms were evident that
quickly.]  Presumably everybody in the room left as quickly as they
could; reports do not indicate if radiation alarms were installed or
activated.  35-year-old Hisashi Ouchi received 17Sv, 39-year-old
Masato Shinohara received 10Sv, and 55- or 54-year-old Yutaka Yokokawa
received some unspecified lower radiation dose.  [Based on doses
reported for other accidents, I assume Ouchi and Shinohara were next to
or leaning over the tank, and Yokokawa was a few feet away from the
vessel.]  These three operators were hospitalized and, as of Oct.
4th, all three were alive.  In fact, Yokokawa apparently was well
enough to answer investigator questions by Oct. 2nd.
However, doctors planned to give Ouchi a bone marrow transplant, at last
report (Oct. 4th).

News reports indicate an additional 37 or 38 people also received some
non-trivial radiation doses, including three fireman who responded, five
nearby residents, and two employees of a nearby golf-course.  [Just
reinforces what I?d always heard: Japan is densely populated, with little
if any space between industrial/nuclear/chemical plants and the
public.]  Reports either do not quantify, or provide wildly
discrepant indications of, radiation levels.  Many articles report
radiation levels anywhere from 1000 to 17000 times above normal
levels.  One embassy official told INEEL personnel that responders
measured a fairly steady 0.84mSv/hr dose rate at the facility boundary
for several hours after the accident, and further indicated that fission
neutrons were detected during that time.  Until the system was
safely subcritical, about 160 people
within a 350m radius were evacuated, and people between 350m and 10km
were advised to stay inside (with doors and windows closed). 
Official notifications were reportedly slow but news articles did not
indicate most elapsed times.  City officials were apparently told
almost an hour after the first critical excursion.  Radiological
responders were apparently  from the plant and presumably activated
immediately.  Other in-field and city responders were apparently
from the civil police, civil firefighters, and army.  Naturally they
weren't activated until official notifications were made.   One
or more Japanese authorities requested US experiential information but it
seems doubtful it was very useful in shutting down the reaction.

One would expect the system to oscillate (or pulse) between super- and
sub-critical states, or to shut itself down quickly.  However, the
system apparently settled into a quasi-steady-state for almost 20
hours.

Tokyo Electric Power Company rushed some 880 lbs of borated material
[boric acid?] to the JCO plant.  However, responders could not use
it right away because they had no readily apparent means of remotely
adding the neutronabsorber to the tank.

Dr. Komuro wrote:  It took about 3 hours to drain cooling water from
a water jacket around the tank. At 02:35 Oct. 1st, two workers with
radiation protection coveralls took some pictures of a
cooling-water-drain-valve outside the building in which the nuclear
criticality accident occurred.  The area where the valve was highly
contaminated [I think he means highly radioactive].  Although they
stayed in the area only 3 minutes, one of them received 11.92mSv gamma
and 91.2mSv neutron.  At 03:00 another two workers went to a pump
yard(?) and confirmed a cooling-water-circulation-pump(s) was working
well.  At 03:22 another two workers opened the valve.  However
responders could not confirm that water was successfully discharged from
the water jacket.  Finally other workers cut a drain pipe outside at
the valves up-stream side, and injected gas.  At 06:00 responders
confirmed cooling water was exhausted.  The tank was finally
subcritical.  18 workers total were involved in draining the water
jacket.

Apparently responders added boron to the system.  Officials declared
the system
safely subcritical at 9:20am Oct. 1st.  I do not know what prompted
premature news reports that the reaction had terminated at various times
after 6:30pm Sept. 30th.

Fission yields are not specified for any pulse or for the reaction
duration.  After the first pulse, the maximum neutron dose reported
was 4.5mSv/hr, but detector location with respect to the tank is not
indicated.

Apparently regulators declared this a Level 4 incident using the IAEA
guidelines.  Their basis was radiation exposures to two operators
and to two off-facility people.  The building suffered no mechanical
damage but filters did not trap fission products.  Essentially all
fission products were released to the atmosphere.  Officials warned
people they should not eat produce or drink milk from the area until
testing was complete.  [However, I?m not certain why they thought
this accident would release enough to be harmful.]

Accident investigation for response decisions began immediately. 
Further investigation apparently began the afternoon of October
1st.  JCO conducted an
in-house investigation and apparently police are conducting much of the
official
investigation.  One news article indicated police raided JCO offices
[is this anything like the USA FBI raiding DOE's Rocky Flats site years
ago?].

MISCELLANEOUS consequences and reactions:
Tokaimura resident attitudes range from calm to angry about nuclear
safety, but few    are concerned about long-term
radiological effects Greenpeace is sending people to survey Tokaimura
Japanese, USA, and UK programs will be reviewed in light of the
accidentIf USA experience is indicative, the JCO plant's entire safety
basis will be
scrutinized and reworked Stock in JCO's parent company is apparently
dropping.  Japan might review its nuclear policy but dramatic
changes are doubtful US President Clinton threatened (or did) veto a bill
regarding Yucca Mountain and nuclear waste

=================================
II.  Dr. KOMURA?s 2nd, 3rd and 4th EMAIL MESSAGES
=================================