RE: 1988 Irradiator Incident

[GrayStarNJ@aol.com]

You can quote me on this:  

The dimensions of the capsule are: Length- 20.775" OD - 2.625"

OK, now that the details are out of the way...I have been studying the 
Decatur incident for many years and can provide you with detailed info.  If 
anyone has any specific questions or wants me to go into more detail than the 
limited stuff to follow, just let me know.

1)  There was probably little or no moisture in the CsCl at time of 
encapsulation.  Even if there was, it most likely was not a contributor to 
the incident.  The CsCl was poured into the encapsulation at approximately 
730 degrees C.  Prior to welding the endcap, the source maintained a high 
temperature due to decay heat.

2)  The only improper manufacturing of that particular tube #1502 is that it 
was probably overfilled with CsCl.  In my opinion the overfilling of the 
source did not significantly contribute to the incident.

3) Co-60 sources have failed in the past with much greater release of curie 
levels to pool water...but that's a different subject.

4)  The source that leaked had 55,800 curies at time of manufacture (July 
1982).  I have hear verbally that only 8 curies leaked into the pool but 
wouldn't be surprised if that figure were as great as "10-20 curies".

5)  My simplified take of the "cause" of the failure.  This is based on both 
reports and recreation experiments that we have performed.

First a little history.  The WESF sources were originally designed for 
storage and NOT for use as a source of radiation in an industrial irradiator. 
 They are very inefficient for irradiator use (i.e. Too much cross section, 
tube wall thickness too great).  These inificnienties lead to a relatively 
high temperature due to heat transfer and self absorption of the gamma 
photons as well as all of the beta particles.  The DOE had an interest at the 
time of exploring food irradiation by setting up pilot plants.  They were 
unaware of the extent of the radiation processing industry and would have 
competed with it.  Therefore, it was decided that they should lease the 
sources to the industry itself on an experimental program basis.  At the same 
time there was a severe cobalt-60 shortage.  The NRC granted a conditional 
"sealed-source" approval to use these sources under a highly monitored 
program including temperature studies and destructive examinations.  4 
facilities took part in the program.  Two of these facilities (if not all) 
were specifically designed for cobalt-60 which included wet storage and dry 
irradiation.  The Decatur facility was one of these facilities.

CsCl is an interesting material.  If pure, it will go through a solid to 
solid phase transition at 467 degrees C.  This transition will increase the 
volume by 18%.  The WESF material has impurities which will tend to lower 
this temperature.  In air, the centerline temperature can be above 400 
degrees C.  The WESF was manufactured by melting the CsCl and pouring it into 
an encapsulation.  The overflow continued to a second encapsulation, and that 
overflow continued into a third encapsulation.  This process would fill 7 
sources at one time.  The first 6 sources were filled to the brim with the 
final one being only a partial pour.  At the beginning of the next session, 
the 7th. capsule would be placed in the front of the line.  The idea was that 
if the material was poured in molten and filled the entire volume, then any 
expansion would fill the same volume and the sources could not burst by 
having too much volume.  I believe that #1502 was one of the 7th. sources.  
By definition, a 7th. source would have been overfilled because after the 
initial fill, it is allowed to cool and contract.  That extra volume would 
have been displaced by molten CsCl on the next pour.  

At molten temperatures, CsCl is corrosive to stainless steel.  The molten 
pour process introduced a potentially significant amount of iron as an 
impurity to the CsCl.  This will lead to a lowering of the phase transition 
temperature.  In one incidence the CsCl broke through the crucible desolving 
a large amount of Fe into the CsCl.  It is possible that this included #1502. 
 All of the above might have contributed to the incident.  More likely it 
probably allowed for #1502 to fail prior to other source under similar 
physical conditions.

Now for the physical conditions.  In Decatur the sources were stored in a 
vertical position.  They were subjected to over 7,300 thermal cycles.  The 
source bulged (over time) with it's maximum deformation 4 inches from the 
bottom of the encapsulation.  (Outer encapsulation bulged 0.224 inches in 
diameter.)  Interestingly there was approximately 6 inches of void space in 
the top part of the encapsulation.  Also there was the CsCl was broken into 
two parts with a 1 inch thick slug located about 1 inch above the main body 
of the CsCl.  My conclusion is that during thermal cycling the CsCl either in 
total or more likely in part expand radially and contract linearly....leaving 
the top part of the source void and the bottom bulged.  

Test recreations:  We have made CsCl material similar in chemical composition 
(impurities) to that of the WESF material.  However, we used non-radioactive 
CsCl (133) in lieu of CsCl (137).  After a molten pour the material shrinks 
away from the sides of the encapsulation and leave most of the void volume at 
the top of the encapsulation as one would expect.  The material is very 
strong and resilient.  Virtually hammer proof!  It is also very dense at 4.0 
g/cc.  Upon cyclic heating of the material it would significantly migrate and 
fill the annulus between the CsCl and the lower portion of the encapsulation. 
 Once this occurs you have a solid cross section of non-compressive CsCl on 
the lower portion of the encapsulation.  As cycling continues, a significant 
bulge is created at that location.

Conclusions:  CsCl should not be used in an irradiator which subjects it to 
repetitive thermal cycling.  Also, the maximum temperature should be 
maintained significantly below the phase transition temperature.  Most 
importantly, CsCl should not be used in a molten-pour form (solid CsCl).

At this point in time, CsCl is not used for commercial food 
irradiation...however, it is used for most research irradiators.  GRAY*STAR, 
Inc. is presently designing a CsCl irradiator which is specifically designed 
to mitigate against the problems identified in the Decatur facility.

I hope this helps,

Russell N. Stein
GRAY*STAR, Inc.
http://www.GrayStarInc.com

ps  This is the short answer!
************************************************************************
The RADSAFE Frequently Asked Questions list, archives and subscription
information can be accessed at http://www.ehs.uiuc.edu/~rad/radsafe.html