On January 12, 1956, an opportunity arose at Oak Ridge
National Laboratory for a meeting of representatives of several
groups to discuss the results of the ORNL inspection of certain dis-
mantled reactor components, and in particular to hear information
and recommendations concerning steps that might be taken to avoid
in the future reactor the various corrosion problems of the first
reactor. Any appreciable corrosion presents a severe health and
safety problem.

A meeting was arranged through Dr. K. Z. Morgan who, un-
fortunately, was called away by an emergency just before the meeting
started. Those present were as follows:

ORNL	AEC	BABCOCK & WILCOX	N. C. STATE
Ed Bohlman	W. H. Behrman	C. C. Cardwell	F. P. Pike
Arnie Olson	(for H. M. Roth)	William Foley
George E. Moore		Durand
Unidentified

The core was stuck so tightly that the aluminum jacket
had to be cut and torn away. The core was then quartered longitudin-
ally in such a way as to do the least possible cutting of the coils
inside. Six metal pieces were then cut from strategic areas. From
each piece a number of metallurgical samples were taken for careful
inspection.

• 1. While the core had been tightly stuck inside the aluminum thinble,
  it appeared that this must have been because of the adhesive effect
  of encrusted uranium salts. By visual examination of the core
  when freed there was no evidence of any pressure distortion. The
  top surface, the weak area in regard to internal pressures, was
  unbowed. None of the welds appeared cracked or distorted.
•  
• 2. The core wall (1/16") appeared to be in good condition except
  for a row of 5 families of pin holes. This row was horizontal
  and about 2 inches long, situated just above a coloration change
  that apparently represented the solution level. The level of the
  solution was somewhat indeterminate. Viewed from the inside and
  at a distance, there seemed to be 5 single pits. Closer examina-
  tion revealed many pits in each family. From inside to outside,
  each of the five groups of pits expanded rapidly to areas of
  l-3 cm2 (my guess) outside.
•  
• 3. Metallographic techniques of examination of the wall in the pit
  regions revealed that the wall was spongy because of many internal
  passageways. This corrosive effect was severe.

• 4. Careful examination disclosed no evidence of corrosion on the
  bottom or around the bottom weld.
•  
• 5. The sample of a cooling coil bend revealed intergranular and
  transgranular (stress) corrosion inside the coil (water side).
  My recollection was that the indicated corrosion of one sample was
  considered to be severe enough to cause failure within not more
  than one or two years at the most. At least, one metallurgist
  advanced this opinion and no one disagreed.
•  
• 6. Except for a few suspicious crystals, of discredited identification,
  there was no evidence of carbide precipitation. The metallurgists
  were agreed on this point.
•  
• 7. On the outside of the core there was a thin (about 1 mil), adhering
  scale or film of iron, uranium and chromium oxides.
•  
• 8. No leak could be demonstrated in the coil sections by air pressure.
  However one section could not be tested (it was destroyed by
  examination) and the test procedure on the other sections was
  rather inadequate. It was tentatively judged that the leaks in
  the tested sections were at most small leaks.
•  
• 9. The metallographic inspection to date is reported in ORNL 55-
  12-70, by C. E. Moore, classified. There will be a subsequent
  report.
•  
• 10. Radiation damage to the metals was not evident nor was it to be
  expected under our conditions. According to Dr. Secoy, radiation
  damage begins to show in a reasonable period only if the power
  level is above about 10 KW per liter of solution fuel.

It was agreed by all that the corrosion was the consequence
of the high chloride content (>200 ppm). Pitting is characteristic
of chloride attack on stainless steels under comparable conditions
(oxidizing media), as is also the location just above the solution
level.

It was not characteristic that the corrosion outside
should be so much more extensive than inside, particularly since
on the outside the solution contact with aluminum should have given
protection. The metallurgists were at a loss for a possible ex-
planation of this anomalous behavior.

The localization of all pits within a 2 inch streak was
also puzzling. However, while no one could say why this occurred,
there was no lack of possible explanations. It could readily have
derived from some past incident, such a a linear occlusion before
or during rolling. Another suggestion was that the welder could
have accidentally struck an arc there.

It seemed clear that the coils failed because of the
corrosive action of the cooling water, probably through the action
of chlorides.

There was no information on corrosion resistance under
exactly our conditions. Most of the available information pertained
to uranyl sulfate solutions under more severe conditions of temperature
and solution velocity. One conclusion was that the net effect of
the simultaneous presence of hydrogen, oxygen, and peroxide was
that of an oxidizing solution.

By extrapolation of available (classified) data, these deductions were
made:

Nickel	No good.
Incoloy	Quite good at high temperature.
Nionel	No information, but question the high nickel content.
Chlorimet 3	Much too high in nickel.
Carpenter No. 20	Probably O.K.
Hastalloy C	Not bad, almost as good as 347.
Stainless 316	Poor.
Stainless 347	Reasonably good, but not very good.
Titanium	Good
Zirconium	Good

For sulfuric and other acids on Stainless 347, data
are available from unclassified sources. We were referred to C. P.
Larrabee as a start.

• 1. If the chloride situation could he controlled, use Stainless 347.
•  
• 2. If the chloride situation could not be controlled, use titanium.
•  
• 3. By all means, control the chloride situation. For Stainless 347,
  the extrapolation of available data indicated (at pH = 2.5 ?)
  no "appreciable" corrosion at 3.5 ppm chloride, but "appreciable"
  corrosion at the next higher chloride content tested. A limit
  of 5.0 ppm, or perhaps 3.0 ppm, was suggested if Stainless 347
  was used.
•  
• 4. If a Stainless Steel is used:
•  
• • I. During fabrication,
  • • (a) Have a camera record made of the entire welding operation.
    • (b) Take and store samples of the metal.
  • II. After fabrication,
  • • (a) Specify treatment with trisodium phosphate, or other
      alkaline, solution.
  • III. After assembly (or installation)
  • • (a) Flush completely with distilled water until certified
      free of chloride.

• • IV. During operation,
  • • (a) Analyze initial batch of fuel.
    • (b) Monitor fuel for chloride at regular intervals, say
      once a month.

It is pertinent that the Babcock and Wilcox metallurgist
said that it was their experience that chloride was a severe danger
for Stainless Steels, and that there were a small but disturbing
number of instances where stainless equipment had become accidentally
contaminated with chloride between early fabrication and first use.
Usually the cause of the contamination remained unknown. To combat
this situation, they have frequently adopted a procedure of flushing
with trisodium phosphate before items are permitted to leave the
plant. A check by the receiver before the use was considered only
common sense. In important cases of interest to them, they would
insist upon a camera recording of the welding operations, and on
the storage of suitable metal samples.

In some way the cooling water must be made less corrosive.
If the cooling water is used once, then discarded, it must be chemi-
cally treated, such as in accordance with the suggestions of our
Sanitary Engineers. Alternately the cooling water loop could be
isolated by a heat exchange unit, and distilled water used as the
media. The closed loop was apparently the preferred alternate of
those present.

The decision to adopt a closed-loop of distilled water will
not be a simple one. The decomposition of water into hydrogen and
oxygen will be a hazard that might be aggravated by the purity. The
circulating water should be monitored and maintained at a specified
specific resistance, which would be about 1/4 million reciprocal ohms
if there are no corrosion inhibitors or agents to reduce the
radiation damage to the water.