[n}@wfi@ Vi 7

FALL
RIDGE NATIONAL LAE

Pt evievw,

=== OAK RIDGE NATIONAL LABORATORY

" VOLUME 3, NUMBER 2 FALL 1969

THE COVER: Silhouetted against the
dome aof Johns Hopkins University Med-

ical Center in Baltimore, C. D. Scott 1 The Chemistry ofa Man

prepares a body fluid sample for detailed
scrutiny in one of his two chromato-
graphic analyzers. For more about this
fabulous new diagnostic tool, see story

on page 1. 9 The Consulting Statistician:
Who Needs Him?

By MARVIN A. KASTENBAUM

By C. D. Scorr

Editor

BarBArA LyoN | 12 25 Years of Creative Support

Consulting Editors By H. E. SEAGREN

Davip A. SUNDBERG

A. H. SNELL 22 Benefits vs. Risks in Nuclear Power

Graphic assistance is provided by By WALTER JORDAN
Graphic Arts and Photography
Departments of the ORNL Division

of Technical Information. 35 The INOR-8 STORY
By H. E. McCoy

The Review is published quarterly and dis-
tributed to employees and others associated
with the Oak Ridge National Laboratory.

The editorial office is in Room 283, Building FEATURES
4500-North, Oak Ridge National Lakoratory,
P. O. Box X, Oak Ridge, Tenn. 37830. AMW, 20
Telephone: 483-8611, Extension 3-6510
(FTS No. 615-483-6510). Books, 31

OAK RIDGE NATIONAL LABORATORY [union

OPERATED BY UNION CARBIDE CORPORATION e FOR THE U.S. ATOMIC ENERGY COMMISSION

CARBIDE

000000000000000000000000000000000000000060000000000000

T e

. am

—

— |

— —

: el o 5 -,
00000C000000Q00000000000000C

\
4 ——-V—J

;
|

!

=¥ __d —

0000000000 C0O0000QO00QO000000

Paim - 4aY
l' " .' - o
. ;."-" " '\ ¢ l ' 1)
The Chemistry of A Y
Here is a new diagnostic to dega ‘
needle-sharp profile of a patient’s chienijcal \
[y \ .\\-,\ - - ‘.} ny
By CHARLEsS D. Scorr X 3
\ 7
. ‘ “
UPPOSE YOU WER N an ext ) nes that h n disassembled after they
intricate machine tha pable of pe lost their u has failed to reveal all the
ing numerous complicated ing only co ts of the cordpl erating sequences. Most of
chemicals for fuel. Comp d reactions nowledge fegard to the status of the ma-
the machine convert this to useful | as well 4 hecessily for repair, depends
Usually this machine requirga little attentigh; = h§t we can m sampling the fuel, waste,
ever, occasionally it malfurigtsons and 'ay er vital . Access to these streams is
repaired. It is a very expensivé abtl ex : ' imple r.
able piece of equipment; thergfére, you 1 iE'® a brief and simple analogy to the problems
willing to expend great effort }and large ) e medi fession in attempting to main-
money to correct any ohvious'delégts a \ \ com chine that is the human body
probably also schedule routine of% P n b Bdalthy is relatively easy for a physi-

feature that does not permit its tofal dis ply;
this configuration, in turn, makes maintendfce and
correction of malfunctions very difficult 4nd
determination of the exact cause of a mal 0
may be impossible. Even the study of other,

FaLL 1969 l-

1an, of the If. to determine when the

nZhi is o ioning properly, at least when

there i§ a gross disorder, because of the partial or

| disahility that ensues. For example, high fever,

r a¢veére headache or pain, can temporarily reduce

ar] L9 the state of an invalid. Frequently remedies

se conditions are readily available which, in

fi?n'p es, provide effective relief without actually
A |
/ |

J
|
/ 1
’

Chuck Scott, who received his doctorate in
chemical engineering from the University
of Tennessee, has been in the Chemical
Technology Division since 1957. Originally
in nuclear fuel production and reprocess-
ing, he has since turned to biomedical
engineering. At present he directs the
Body Fluids Analyses Program which spans
four divisions. Besides Chem Tech, the
program's personnel represent the Mole-
cular Anatomy (MAN) Program, Instru-
mentation and Controls, and Analytical
Chemistry. The development of the re-
markable machinery described here has
attracted the attention of a number of
medical centers. The two analyzers, ultra-
violet and carbohydrate, still in the process
of constant improvement, are being tested
out by medical research teams at Duke,
Johns Hopkins and the National Institutes
of Health. The latest addition to the rig
is a PDP-8 digital computer, designed to
enhance the analyzer's evaluation capa-
city. Further improvements in the offing
include miniaturization (by a factor of
eight), and an attempt to combine the two
analyzers into one unit. Scott is shown
here at Johns Hopkins University Medical
Center with Dr. R. Rodney Howell of the
Center’s department of pediatrics.

repairing the malfunction. But how do we find the
basic cause of a malfunction? Or, better yet, how do
we maintain or service this machine to prevent
malfunction?

Abnormal Body Function

When man is ill, his body usually functions in an
abnormal manner. Often these abnormalities—
aches, pains, fever, etc. —can be observed clinically.
Clinical symptoms can be clues to the source of the
trouble just as the change in the sound of an oper-
ating automobile motor may be a clue to which part
has gone wrong. However, in order to solve more
difficult problems that may be related only indi-
rectly to clinical signs, it may be necessary to ana-
lyze the more delicate operating sequences of the
machine or to perform tests that can be used to
evaluate individual portions of the machine.

Since man’s body is a machine whose operation
depends on many very complex chemical reactions,
the capability of determining the level of chemical
activity in the body at a given time is invaluable.

To some extent, this can be done from accurate
analyses of the waste streams (urine, breath, sweat,
feces) and other, vital streams (blood, poral fluids,
spinal fluids, etc.) coupled with the knowledge of
the composition of the fuel stream (food intake).
Medical science has, in fact, gained vast amounts
of information about the relationship between ex-
cretion products and abnormal body function. In a
recent bibliography on that. subject prepared at
ORNL, reference to over 1,000 molecular constitu-
ents in human urine was found, and most of these
compounds have pathologic significance. For ex-
ample, the excretion of excessive amounts of uric
acid is known to be associated with gout and also
occurs in some types of mental retardation; exces-
sive quantities of glucose in the blood or urine may
indicate diabetes mellitus, etc. In some eases the
biochemical reasons for the abnormal molecular
pattern in body fluids have been determined to be
the result of disease or a metabolic deficiency. In
fact, it is now reasonably well accepted that most
diseases will ultimately be understood and con-
trolled on the molecular level.

OAK RiDGE NATIONAL LABORATORY Review

000000000000 0000000000000000O00O0OCOO0O0VOODOOOOOO0CO0000O0OC OO0

i

g

.:] i
j } Y Y "_r_

es

dy Flui ses P

© BTl gk BeeRpBcted, Aot YRS oF 942 © © © HCRRL 1B VAP IRK RiFhOrSsPI R RR SiiPyi-

ous body fluids can be valuable diagnostic aids for
the clinician. Most clinical laboratories can accu-
rately analyze for as many as a dozen different
chemical constituents of a blood serum or urine on
a routine basis, while some laboratories associated
with research hospitals will frequently be able to
analyze for as many as 30. Typically, each analytical
method used in a clinical laboratory will result in
the analysis of only a single constituent or of a
single group of constituents in a complicated physio-
logical mixture. Considerable developmental effort
has been directed toward automating many of these
methods and, in some cases, in combining several
analyses into a single automated instrumental
array that requires a minimum of operator time.
This arrangement has allowed some clinical labora-
tories in large hospitals to performm more than a
million analyses per year.

Unfortunately most of this development has been
concentrated on obtaining more rapid analyses of
the compounds routinely detected instead of in-
creasing the number of constituents being analyzed.
Recently, there has been some work on the develop-
ment of automated, high-resolution analytical sys-
tems in which many of the individual constituents
of a physiologic sample can be quantified by a single
analytical step. Such techniques, if successfully
developed, could be used for in-depth analysis of
physiologic material. Also, they might be used to
establish a relatively complete chemical prefile of
a particular individual. In this context, the term
“high-resolution analysis” has been chosen to de-
scribe an analysis in which many, or all, of the con-
stituents of a sample mixture are separated and
quantified.

FALL 1969

cal tools, ORNL has established a Body Fluids
Analyses Program that is funded by the National
Institute of General Medical Sciences. This pro-
gram, a part of the Molecular Anatomy Program,
is a multidivisional effort which involves members
of the Chemical Technology, Analytical Chemistry,
and Instrumentation and Controls Divisions. In a
sense, the program is typical of many ORNL projects
in that the efforts of many scientists and engineers,
representing a wide range of disciplines and talents,
have been combined to solve a specific problem. In-
cluded are chemical engineers, instrument engi-
neers, analytical chemists, biochemists, organic
chemists, and a heavy complement of support
personnel.

The development program includes basic studies
on high-resolution separation techniques and de-
tection methods that will have application in the
automated analysis of body fluids. However, its ulti-
mate objective is the development of totally inte-
grated, automated analyzers capable of high-resolu-
tion analyses of body fluids. To this end, prototype
systems are being built and tested in collaboration
with various research and clinical laboratories.

To date, emphasis has been centered on the analy-
sis of low-molecular-weight (less than 1000) con-
stituents of urine and other body fluids, with two
different analytical systems: an analyzer for de-
tecting and quantifying the ultraviolet-absorbing
constituents (UV analyzer), and a carbochydrate
analyzer for detecting and quantifying carbchy-
drates. Both systems use high-pressure ion ex-
change chromatography (up to 5000 psi) for separat-
ing the constituents of the physiologic sample, and
continuous photometry or colorimetry for determin-

Elutriation system, for separating ion-exchange resins into
specified particle sizes needed.

Analytical chemist Ken Warren
adjusts the reagent flow rate.

ing the separated constituents. In fact, the develop-
ment of high-pressure ion exchange chromatography
for use as a routine analytical tool has been one of
the major technological contributions of this pro-
gram. The high-pressure technique has also found
application in other fields at ORNL; for example, it
is now used extensively in the transuranium ele-
ment program for separating some of the actinides.

The results from these analyzers are graphically
presented on a strip-chart as a plot of the absorbance
of light by the chromatographic column exit stream
as a function of time. These chromatograms contain
a series of peaks, each of which indicates a molec-
ular constituent. The UV analyzer resolves up to
150 such peaks from a 1.0-cc body fluid sample,
whereas the carbohydrate analyzer will yield up
to 48 chromatographic peaks from a typical sample.
Both systems require a relatively long period of
time for analysis (40 hours for the UV analyzer
and 20 hours for the carbohydrates); however, ac-
tual operator time is equivalent to only 15 to 20
minutes per sample.

Oak RiDGE NATIONAL LABORATORY Review

Exportable Hardware

As in many of the programs at ORNL, we can al-
most say that “paper is our most important product,”
since the end results of our research or development
inevitably call for the preparation of progress re-
ports and myriad other writings. However, “hard-
ware” is also a result of our work.

As new, automated, analytical systems are de-
veloped, prototype systems are built and sent to
other laboratories for testing and evaluation. This
is an important part of our program since, in addi-
tion to collecting clinical research data, these eval-
uations furnish us valuable information concerning
feasibility and operability that will prove useful in
designing future systems. Ultimately, we would
like to develop systems that a relatively inexperi-
enced technician can be trained to operate in a
short time.

Four prototypes of our model Mark II series ana-
lyzers are currently being evaluated in other lab-
oratories: one UV analyzer, at Duke University
Medical School, is at work on inborn errors in
metabolism; another is being tested in the Depart-

FaLL 1969

Norman Lee, technician from Chem Tech,
adjusts the flow rate of the chromatographic system.

ment of Pediatrics at Johns Hopkins University
Medical Center on abnormalities in infants; and
both UV and carbohydrate analyzers have been
placed in the Clinical Center of the National In-
stitutes of Health for in-depth analyses of body
fluids of selected patients.

Already these machines have produced some in-
teresting results. For example, very complex carbo-
hydrate chromatograms have been obtained from
the urine of diabetics, indicating that diabetes is
characterized by abnormal quantities of carbohy-
drates other than glucose. Also, the effects of drugs
for the treatment of gout are being monitored by
the UV analyzer.

Various other items of hardware that were de-
signed and fabricated during the development of
our analytical systems have also been made avail-
able to interested groups. Two of the most important
of these are a small sample-injection valve that per-
mits the injection of a liquid sample into a flowing
stream at pressures up to 5000 psi, and a small,
inexpensive UV photometer that can be used with
other liquid chromatographic systems.

Identification of
Body Fluid Constituents

Although the chromatographic patterns them-
selves might provide the basis for determining an
abnormal situation, the biochemical significance of
the abnormality will be entirely lost unless the ab-
normal chromatographic peaks can be identified.
Therefore, identification of the separated chemicals
constitutes a significant effort in our program.

In general, one must first isolate the chemical in
a column fraction, then find a means to purify the
chemical, and, finally, establish its identity by
various spectral and chemical techniques. This se-
quence of tasks must be accomplished with only a
few micrograms or, in some cases, less than a micro-
gram of material. Such work has necessitated the
development of additional competence in our Ana-
lytical Chemistry Division, which now boasts a well-
staffed and equipped analytical biochemical labora-
tory. Over 40 UV-absorbing constituents and 16
carbohydrates have been at least tentatively identi-
fied in body fluids; some of these have not been pre-
viously reported as components of body fluids.

What Is Normal?

Relatively early in our program it was necessary
to determine whether the chromatograms of body
fluids of “normal” persons are similar, that is,
whether the body fluids of “normal” persons contain
similar amounts of the various molecular constitu-
ents. The term “normal” as used by us and others in
medical science represents a paradox, because by
placing enough restrictions on the population to
rule out many abnormal conditions, we are left with
only a very small portion of the population that is
“normal.” Actually our “normal” represents an
idealized composite and in no way denotes the aver-
age member of society. Nevertheless, we did find
eight “normal” persons at ORNL and samples of
their urine and blood were obtained periodically and
analyzed by both systems.

Persons who met these conditions of normalcy
were found to have very similar chromatographic
patterns even though no attempt was made to con-
trol their diet or physical activity. Therefore we
believe that our systems can detect pathologic con-
ditions that result in abnormal chemical levels. This
has already been confirmed by analyses of several
body fluids from patients suffering from both physi-
cal and emotional disabilities.

UV analyzer’s spectrophotometer gets a minor adjustment.

Oak RipGe NaTioNAL LABORATORY Review

JO0000000C0O00OOO0OCO0O0C0O00OOO000O

] R —— TN.*...M e

- . S N T I T Wew—

p—
e

egrercanaTs

o

1d samples

e

at were ished by Oak Ridg : -
QRS Bastern®State PsGcRiBtric AGHRAL g

the University of Tennessee Memorial Hospital.

Some interesting results have been observed; for-

example, the analysis of urine of patients with acute
lymphocytic leukemia showed an almost total ab-
sence of hippuric acid; two large, unidentified
chromatographic peaks were found to be present in
urine samples from schizophrenics; a large quantity
of homovanillic acid was detected in the urine of a
patient with a neuroblastoma; and, of course, large
quantities of glucose and other carbohydrates were
found in the blood serum and urine of diabetics.

The Future and Preventive Medicine

It is our aim to streamline our present analytical
systems so that eventually they will be more highly
automated, enabling analyses to be performed in
shorter time. Already, work is progressing on the
combination of several chromatographic systems
with a small, on-line computer to permit computer-
evaluation of the data.

We have also built and are now testing advanced
prototype systems that are more compact, less ex-
pensive, and easier to operate than the older models.
An attempt is also being made to combine the two
analytical systems, along with additional detection
systems, into one instrument package. Our ultimate
goal is to develop a sort of “black box” that can
rapidly analyze body fluids for literally hundreds
of their molecular constituents.

The advantages that such tools will offer the
medical profession are many and obvious. As the
equipment becomes more reliable and as analysis
time is reduced, it is possible that the high-resolu-
tion analyzers will be used in clinical laboratories

FaLrL 1969

V. E. Walker of Instrumentation and Controls demonstrates the
automatic injection valve he developed (shown above)

with Scott and W. F. Johnson to inject controlled samples

into the high-pressure stream of the analyzers.

3
"

N

on a routine basis, as an aid in diagnoses, as well
as in monitoring the effects of drugs. Another use
can be in the screening laboratories, where periodic
testing of healthy individuals is done to determine
if abnormal conditions are developing. Use of high-
resolution analyzers by such laboratories would
accumulate a wealth of useful information for de-
tecting incipient disease.

Possibly one of the most exciting future exten-
sions of this work is in its ability to acquire data
that would allow physicians to detect abnormalities
in infancy or childhood, before clinical manifesta-
tions are present. Some pediatricians say (perhaps
with bias) that infancy or childhood is the only
useful time for practicing true preventive medicine,
since at that time any body damage that might
result from abnormal conditions is at its minimum.
Physicians are finding that an increasing number
of abnormalities (principally metabolic deficiencies),
if detected in infancy, can be treated successfully
to prevent an early death or severe body damage.
For example, early detection of phenylketonuria

Bill Butts, analytical chemist, evaluates
one of the chromatograms with Scott.

(PKU), by analyzing for excessive amounts of
phenylalanine in blood and urine, has made it
possible to treat infants suffering from this condi-
tion with proper diet and thus prevent severe brain
damage that could result in mental retardation
or early death.

To carry this type of thinking a step further, the
ideal time to detect, treat, and correct some types
of abnormalities may be during the development of
the fetus. One way to determine certain metabolic
deficiencies would be to analyze the amniotic fluid
for its molecular constituents. We are now begin-
ning to analyze amniotic fluid samples and initial
results have shown that the low-molecular-weight
constituents of normal amniotic fluids are present
in about the same quantities as in blood serum. This
is a very new technique and we feel that such analy-
ses have a large potential for the future.

Great strides are being made in the biomedical
sciences, and we in the Body Fluids Analyses Pro-
gram are looking forward to a productive and ex-
tremely exciting time.

OAK RipGE NATIONAL LABORATORY Review

THE CoNSULTING
STATisTiCiaN:
WHO NEEDS HiM?

By MArRvIN A. KASTENBAUM

“FIVE DOLLARS will get you
a hundred if I'm wrong.”

How often have you seen these
odds offered when there is real
money involved? Not very often.
But eavesdrop sometime on a group
of kids talking about their favorite
ball players, and you’ll hear it a

FaLL 1969

lot. Or, better still, pick up any
scientific journal in which quanti-
tative measurements are being dis-
cussed, and you’ll find sophisticated
adults of great repute making this
type of statement frequently about
the results of their experiments.

Just what kind of game are they
playing? Presumably they are mak-
ing strong inferences concerning
some unknown quantities on the
basis of the results of one or more

.of their experiments. The interest-

ing thing is that unless they are
willing to give at least 19 to 1 odds
on their results, nobody in the sci-
entific community will pay much
attention. Isn’t this strange, when
most of the readers and editors of
these scientific journals would not
be caught dead risking 6 to 5 on a
sure thing, if six of their hard-
earned dollars were involved.

So what is this phenomenon that
we observe daily in scientific cir-
cles? Why is it that editors of sci-
entific journals are more concerned
with the odds and probability state-
ments of their contributing authors
than with the credibility and au-
thenticity of their experiments?

The odds against “crapping out”
when you roll a pair of fair dice are
8 to 1. You can bet even money
that, in a room containing 23
people, two will have the same
birthday. If you flip a coin 10 times
and observe seven heads and three
tails, would you conclude that the
coin is biased? No. But if you should

flip the same coin 1,000 times and
observe 700 heads and 300 tails,
would you then conclude that the
coin is biased? Yes. How do these
experiments differ? “In the number
of observations made,” you will
answer and your intuition would
be perfect. Why not apply it to
experimental situations? For in-
stance, why 1is it so difficult to ac-
cept the fact you may bet 9 to 1 but
not 19 to 1 that a 2.3% incidence
of a disease among 500 animals in

one experimental group is different
from a 4.6% incidence of the same
disease in a second group of 500
animals? That a similar difference
based on 1,000 animals in each
group would allow you to bet 19
to 1?

Because he can quote these odds
the consulting statistician could be
referred to as the resident bookie.
He has been called many things by
many people, and the term “bookie”
is not the worst. The fact of the
matter is that the statistician can
quote odds on most quantitative
studies, before they are performed.
What’s more, he can tell how many
observations are required to detect
differences of specified magnitude
for preassigned levels of risk.

You'd think that this ability
might be of great interest to scien-
tific administrators who are con-
stantly faced with the problem of
funding projects in a period of tight
budgets. Yet a great number of
large and expensive experiments
continue to be initiated with little
chance of success in terms of the
accepted scientific definition of
proof. Any consulting statistician
worth his salt can save the scien-
tific and industrial community the
equivalent of at least twice his
annual salary every year by saying
forcefully, “Do not run that experi-
ment. You are not planning to
make enough observations. The
chance that you will detect a true
difference is about 0.5, and you can
do just as well by flipping a coin.”

If you think it’s presumptuous of
a statistician to believe that he
can help sophisticated scientists
and engineers in planning their
experiments, then you should re-
alize that “experimental design” is
the statistician’s art. When a stat-

10

istician refers to “experimental
design,” he’s talking about plan-
ning experiments which yield opti-
mum results at reduced costs and

™
1T W

v
v

statisticians have been perfecting
this art for over 40 years.

Designing experiments should
not be confused with proposing ex-
periments to demonstrate or prove
certain physical phenomena. One
could safely say that some of the
most outstanding scientific experi-
ments have been proposed and per-
formed by investigators who had
little or no knowledge of “experi-
mental design.” This point was
brought home to me very vividly
by a scientific administrator who
said, “Enrico Fermi probably knew
absolutely nothing about experi-
mental design.” I could hardly deny
this statement, but I could respond
with the same degree of confidence
by asking, “How many Enrico
Fermis do you have on your staff?”
The fact is that in spite of what we
might think about ourselves, there
are precious few Nobel laureates
around. By the same token there
are a great many good scientists
and engineers who are performing
experiments. Some do achieve
greatness; most continue to do
good work. The statistician’s atti-
tude, based on numerous post mor-
tem observations, is that better
planning can only enhance the
relative frequency of successful
experiments.

The consulting statistician has
often been compared with the psy-
chiatrist who contributes to society
by guiding his patients through
the difficult problems of life. This
aspect of a statistician’s role in
science and engineering is most
easily understood and appreciated.
It is therefore most readily ac-
cepted. My young children explain
it simply by the statement, “Daddy
is a number doctor. He takes care
of sick numbers.” But the statisti-
cian’s own concept of his reason-
for-being is much broader than
this. If he is indeed practicing a
form of medicine, then his emphasis
is on insurance and prevention
rather than on emergency correc-
tive treatment.

It has been said that the statisti-
cian will soon be replaced by the
computer. Nothing is further from
the truth. The etymology of the
word statistician suggests an in-
dividual who is concerned with
matters of state. Traditionally, the
statistician was a person involved
in enumeration and tabulation of
data collected by governments.
More recently his scope has been
expanded to institutions other than
governments, and to concepts such
as estimation and induction. He
now plays an important role in
planning the collection of data and
in their interpretation. The di-
versified applications of the statis-
tician’s stock in trade are best

OaAxk RiDGE NATIONAL LABORATORY Review

illustrated by the specialized sci-
entific journals in which he pub-
lishes: Biometrika, Psychometrika,
Econometrica, Technometrics, Bio-
metrics, Qualitatiskontrolle, Skan-
denavisk Aktuvarietidskrift, Bio-
metrische Zeitschrift, Journal of
Combinatorial Theory and Teoriya
Veroyatnostei i ee Primeneniya.
The computer is merely a tool
which some statisticians use to
supplement the calculating ma-
chines on their desks. Very often
the computer is used for Monte
Carlo studies (once referred to
as “random sample studies”) to
achieve numerical results to theo-
retical problems which have no
“closed-form” solutions. But gene-
rally they are used by statisticians
to get more rapid answers, using
standard analytical techniques, to
problems involving the reduction
of data.

The statistician is not the keeper
of stores of data. This responsi-
bility, which was once in the hands
of bookkeepers and auditors, has
now been taken over by computer
centers. It doesn’t mean that the
statistician is no longer interested.
On the contrary, the statistician is
often looked upon as a queer duck
who, for some strange reason,
shows an interest in the great
quantities of data which he sees
collected all about him. Usually his

FaLL 1969

attitude is that these data have
been collected with good reason
and that someone should pay at-
tention to their analysis. This
doesn’t mean that he believes the
data are good simply because
they’re there. On the contrary, he’s
probably the most skeptical guy
you can find to give you an evalua-
tion, retrospectively, of a mass of
data which has been collected for
an entirely different purpose. He
will help with the salvage and re-
pair of old and shopworn data, but
he will adamantly refuse to sup-
port any inferences about them
with strong probability statements.
Can you blame him? He knows
about biases and he also knows
that the mathematics of probabil-
ity is not to be sprinkled like the
ashes of the red heifer, which ac-
cording to Old Testament accounts,
made the pure impure as it puri-
fied the impure.

The consulting statistician en-
joys his contacts with the scien-
tific and engineering communities.
Some of the problems he encounters
are old hat and trivial; others are
challenging. He is never selective
before the fact, because he cannot
predict when he will be challenged.
He usually spends hours with a
client asking stupid questions just
so he can get a better feel and un-
derstanding of the problem. His
caution should not be interpreted
as stupidity, but rather as igno-
rance of the subject of his client’s
field. By the same token, his client
should not feel badgered and over-

whelmed by the probing questions
which the statistician is asking.
Honest exchanges of this type have
been known to produce some out-
standing results in the scientific
and engineering communities.
These are the dynamic, person-to-
person encounters that have also
proven so useful in business, in-
dustry, and government. They are
the types of situations in which
sincerely interested people par-

(:/I._’
'\ |
7

ticipate and they bring fulfillment
and often joy to the statistician.
0. Henry touched on this phenome-
non in the following passage from
“The Handbook of Hymen:”

“‘Let us sit on this log at the
roadside, says I, ‘and forget the
inhumanity and ribaldry of the
poets. 1t is in the glorious columns
of ascertained facts and legalized
measures that beauty is to be
found. In this very log we sit upon,
Mrs. Sampson,’ says 1, ‘is statistics
more wonderful than any poem.
The rings show it was 60 years old.
At the depth of 2,000 feet it would
become coal in 3,000 years. The
deepest coal mine in the world is
at Killingworth, near Newcastle.
A box four feet long, three feet
wide, and two feet eight inches
deep will hold one ton of coal. If
an artery is cut, compress it above
the wound. A man’s leg contains
thirty bones. The Tower of London
was burned in 1841.°

“‘Go on, Mr. Pratt,’ says Mrs.
Sampson. “Them ideas is so original
and soothing. I think statistics are
just as lovely as they can be.””

11

25 Years of Creative Support

The history of the growth of Plant and Equip-
ment into ORNL’s largest division

By H. E. SEAGREN

ABORATORIUM SUPPORTANS might serve as
the official motto for the Plant and Equipment
Division. Many organizations at ORNL contribute
toward the total effort of supporting the Laboratory.
The Plant and Equipment Division performs a broad
spectrum of these functions and activities in as-
sisting Laboratory programs.

12

Harry Seagren presides over the iarges'
most complex operation in the Laboratory.
Plant and Equipment Division employs
over a thousand people and assumes the
responsibility of providing for an almost
unending variety of Laboratory needs. He
is peculiarly fitted for this, having v
atory deve e 1S
e to Oak Ri
the U. S. Arm
ob was pile ¢

How does a “kiloman” (and -woman) organization
such as P & E support the Laboratory? The answer
varies kaleidoscopically with time and program
fluctuations. The multiple impressions of P & E
have been compared to those of the blind men with
the elephant. Indeed, the backgrounds, experiences
and philosophies of Laboratory staff members, as

OAxk RIDGE NATIONAL LABORATORY Review

well as their specific needs for support services,
all affect their individual viewpoints.

Basically, our support activities can be placed in
three classes:

Supply of materials, which includes equipment,
utilities, process materials or supplies, and modi-
fications to facilities;

Application of Manpower, i.e., craft support and
other “extra hands” necessary to convert the ma-
terials class into the form or configuration required
for the particular program; and

Technical Direction and Supervision—the trans-
lation of ideas into verbal or written instruction for
the performance of work and the supervision of
that work for effective performance. Written in-
structions include orders, drawings, specifications,
bills-of-materials, sketches, and procedures.

In a complex, government-financed operation
there are obvious needs for additional groups to
handle the requirements for planning and cost
control, official records, systems data, and inspection.

How should support functions be organized? The
answers are myriad, and depend on the people in-
volved. One researcher may prefer personal contact
and direction of all work affecting his own parti-
cular project. Others may prefer to be relieved of
all involvement with the details of craft supervision
and job coordination.

FaLL 1969

Left, 1949 oscillator

at Graphite Reactor

adapted from a Maytag

washing machine for

measuring neutron cross sections; right,
neutron beam hole at HFIR.

The operator may prefer that all groups affecting
his operation report directly to him.

The designer may feel that his personal and pro-
fessional contribution would be more meaningful if
he followed his design through fabrication or pro-
curement, installation, and check-out of the “hard-
ware” for a “turn-key” job.

The accountant may insist on precise definition
of work increments and the establishment of accu-
rate budgets and costs for each.

The support supervisor naturally desires to con-
trol all contributors of material, manpower, and
equipment necessary to complete his assignment on
schedule within the allotted funding.

The current philosophy in P & E is best expressed
in Ovid’s advice: “. . treat a thousand dispositions
in a thousand ways.”

Evolution of hot-cell techniques:
Left, isotope handling in 1947;
middle, "over the wall with tongs;”
right, current method of processing
HFIR fuel element.

We are encouraging a team approach, where
practical, of field engineers, crafts, supervisors, and
aides on a local basis to provide a responsive sup-
port to programs. Priority decisions, job plans,
estimates and manpower assignments are made
where the precise knowledge of the actual require-
ments exists.

For purposes of budgeting, personnel functions,
administration of the company-union contract,
training, and employee benefits, the team members
are assigned to the more traditional organizational
departments. However, the structure is not rigid
and must remain flexible to adjust to changing
Laboratory programs.

Over the past quarter century a continuing
change in support functions has followed the evolu-
tion of the Laboratory. The transition is reflected
in the division’s successive names. Originally
Engineering and Maintenance, the title included
the Construction function as well for a brief period
in 1948 during “Plan H,” the first approved plan for
permanent facilities at the Laboratory. In 1954 it

14

was changed to Engineering and Mechanical shortly
after instrument-oriented members of E & M joined
groups from several research divisions to form the
Instrumentation and Controls Division. The word
“Maintenance” in E & M was changed to “Mechani-
cal” as it was alleged, by one who shall be granted
anonymity, that “maintenance is not a valid func-
tion in a research laboratory.”

The present title, conferred in 1963, is the most
recent attempt to define the division’s role. It be-
came permanent when the functions of engineering
design and coordination of construction became or-
ganized as the General Engineering and Construc-
tion Division. The Materials Department from the
former Finance and Materials Division joined the
P & E organization. The initials posed a potential
handicap when a few wags suggested facetiously (I
hope) that these letters stood for Promises and
Excuses.

Each of the division’s functions has undergone
an evolution of activities.

Supply of Materials

For most of the first decade in the Laboratory’s
history the inventory control techniques included
handwritten stores slips, hand-posted. ledger cards

OAK RiDGE NATIONAL LABORATORY Review

for each item, stores replenishment based on order
points that were set pretty much intuitively. In
the ’50s our ORNL materials people fostered the
establishment of a computer-assisted inventory con-
trol using pre-punched cards. The computer provides
the inventory control reports. In recent years the
computer has also been analyzing replenishment
needs using a formula for economic order quantity
(EOQ) which considers warehousing costs, safety
stock levels, lead time for procurement, quantity
discounts, packaging standards, past usage rate,
and cost of inventory funds. The EOQ formula works
well most of the time, but requires close scrutiny of
critical items when the stores’ withdrawal rate
fluctuates under program budget pressures. The
slope of the stores withdrawal curve plotted against
time is a fair barometer of the immediate budget
situation.

ORNL personnel contributed to the compilation of
a four-plant stores catalog for the Oak Ridge and
Paducah installations. Besides the benefits of com-
bined procurement of common items, the catalog
eases the transferring of materials and parts be-
tween plants. Approximately one-fourth of all
ORNL (X-10) stores issues for P & E shop and field
work come from stocks at other plants.

Veteran material handlers at the Laboratory re-

FaLL 1969

call ruefully the early years when a large stock of
lumber was physically relocated five times: mostly
by hand, but with the assistance of a steam-powered
crane in moving heavy bridge timbers. Today, most
bulk stock is handled in palletized form with the
help of fork-lift trucks.

In the early days No. 316 stainless steel and the
particulate forms of graphite were the critical
fabrication materials. In the intervening years the
transition has been through a special HRT grade
of stainless, a low-carbon 304 stainless, INOR-8
and various refractory metals. The nuclear codes
and material specifications have made the in-
specting groups an essential part of the “receiving”
function. Also, continuous-identification-marking
techniques and paper “pedigrees” of metal stocks
are now commonplace.

A bill-of-materials program ensures that the ma-
terials and supplies are available at the start of
the work. Chemical and electronic stores provide
access adjacent to most of the labs of staff members
using these items. “Prepaid, low cost items” mini-
mize the clerical requirements of many stores issue
records. Special auxiliary storerooms in some
research divisions, called “substores,” reduce delay
time in filling specialized supply needs for three
physical research groups.

15

One man, Sam Croft, working almost full time in
establishing contacts around the country, in re-
viewing lists of surplus items, and in personally
inspecting this material, has obtained almost $10
million (book value) of equipment, supplies and
material, ranging from precision optical equipment
to armorplate and naval gunmounts. These three
items are now in use as viewing devices for in-cell
operations, as special shielding chambers, and as
the rotating base for reactor beam experiments as
well as for many heavy machines in our shops.

Application of Manpower

Our craft groups support Laboratory programs in
shop and field activities. The originally separate
research shops and mechanical shops have been
united to offer a range of services from staff, area,
and speciality shops to the large machine and multi-
craft shops. The traditions of the department shop
of a university are coupled with the latest tech-
niques of industrial metal forming, shaping, cutting
and joining.

A fascinating analogy to Parkinson’s Law can be
seen in the way ingenuity exercised in the solving
of some “unique” problem becomes standard shop
procedure for the resolution of many other hardware
problems. The capability of repairing microscopes
acquired during W W II has developed into a broad
range of optics work. Many devices and techniques
viewed originally as novelties have progressed into
proven elements of shop capability.

The ancient art of glassblowing (even now, the
personal expertise of many scientists) continues to
find new areas of application. Currently of major
interest is the joining of glass to a variety of graph-
ites, ceramics, and refractories.

One shop group in Building 2506 concentrates on
the development of innovative practices and
problem solving. An intriguing challenge they met
recently was the application of powder-forming
techniques to teflon bag molding in meeting the
exacting specifications for NASA moon dust col-
lector components.

Most craftsmen working with research groups
develop a keen interest and many special skills,
particularly in adapting normal shop or field
techniques to resolving the immediate problems
in equipment development or operation. Years ago
we were troubled by requests for such items as
“rotating equipment with zero backlash on revers-

16

ing direction and absolute reproducibility of specific
settings.” Now, we work to meet such specifications
without question. Several craftsmen have been in-
cluded in the listing of co-authors of research reports
in recognition of their contributions to the ex-
periments.

In 1943 a stainless-steel welding certification was
rare for welders entering the Manhattan Project.
Now we use 45 separate certifications or qualifica-
tions on welding procedures developed for the metal
joints used at ORNL. To keep his qualification for
a specific technique, a welder must demonstrate his
capability on an inspected job assignment or take
a “bench test” in a 30-day period. This means he
must perform a specific weld job in 30 days or go
back to be recertified.

This growth in welding capabilities results from
a cooperative program with Metals and Ceramics. As

Oax RiDGE NATIONAL LABORATORY Review

the Laboratory searches for improved materials in
developing reactor and process systems, this rela-
tionship with M & C becomes even more meaningful.
One of the salient features of our support work is
the application of approved “bench-top” procedures
to field conditions. An inert-gas-shielded welding
technique becomes more complicated when it must be
performed on a leaking joint or flange under a tank
near the floor of a radioactive chemical processing
hot cell, possibly with the aid of a dental mirror.

In the field support work, electricians, mill-
wrights and pipefitters form a basic core in three
major areas. Other crafts are assigned as the work
requires. Most of the work assignments involve
assembly, installation, repair, modification or dis-
assembly of program-related equipment. Having the
right types and numbers of crafts available to fit all
work requests is somewhat idealistic. At one time

FaLL 1969

Maintenance of exhaust system in attic
over Metals & Ceramics Division in 45008.

Apprentice training in optical evaluation
of machine accuracy, H. Seaman instructing.

we postulated a quota system based on average sup-
port manpower commitments from program staffs.
However, the available manpower in a given week
proved to be considerably out of phase with the
immediate manpower requirements. The current
system is to assign a normal allotment of crafts to
various program areas, subject to weekly revision.
In this way we can accommodate to peak demands
like reactor shutdown schedules or the start-up of a
new processing facility. We are, of course, governed
by our obligation to distribute overtime equitably
as well as the need to limit radiation exposure.
The simple “over-the-wall-with-tongs” approach
to remote handling is remembered by some of our
pioneer researchers. Today, a manipulator main-
tenance crew cares for 267 manipulators (plus
several in-cell remote handling systems) with the
shop processing about five a week. Contamination

17

control  requirements  spurred a  development
program for “manipulator booting” which would
permit a longer operating life before replacement.

In the past decade we have differentiated be-
tween non-programi-related activities and program
support. Those activities which must continue in the
{strictly hypothetical) event the entire K & ) staff
goes on vacation for the month of August gualify as
“standard” plant services. They, in shovt, guarantee
the operability of the Laboratory when September
rolls around. These activities lend themselves more
readily to normal industrial approaches and systems
control, and are conducted in such a manner as to
minimize interference with the research and
administrative staffs.

“Programmed” Maintenance

At one fime it may have been acceptable to place
a bucket under the leaking roof of a wonden, war-
tirmne building and concentrate the available labor
on meeting military deadlines. Today, in permanent,
multiple-cccupancy buildings, someone must oon-
cern himself about the preservation of the capital
investment, Ray Ruel, & maintenance management
consultant here from Industrial Engineering Insti-
tute, told us last February that the ultimate goal of
the maintenance staff should he o put itself cutl of
business. ¥While this objective is remote, we have
proceeded with a number of work-simplifving, prob-
lenmidentifving measures to pinpeint costs, schedule
assignments and extend the operating life of
buildings and eguipment.

The foreman’s pocket notebock with penciled |

comments on spare parts, equipment history, and
reminders for future checks is no longer the basic
maintenance file. & “programmed-maintenance”
system has operated for over five years. Components
of utility systems, building service eguipment,
mobile equiprent, and some operating eguipmendt
for process systems have been located, identified,
and numbered. The servicing of over 8,000 plant
equipment items is based on computer-scheduled
cards that direct the specific items to be checked and
specify the scheduled period by week, month or year,
Al lubricants are coded to simplify procurement
and minimize any confusion in their apphcation.
The identification of bearings and V-belts now offers
some simplification of stores stock.

In sccumulating cost and observing repair fre-
guencies we can provide data fo justify higher
guality materials and parts. As John Ruskin once

18

said, “There is hardly anything in the world that
someone cannot make g lttle worse and sell a little
cheaper—and the people who consider price alone
are this man’s lawful prey.” The faucet in a lavatory
sink is a simple device, but the tirme spent main-
taining a thousand faucets is appreciable. We are
ohserving the total, long-term costa of a faucet with
washer and valve seat as opposed to the total cost
of a washerless faucet over the same time period.

The operability of a safety shower is vital to a
person who has just been splashed with an acid. We
have sought out each device, installed a plug valve
to permit maintenance without shutting off the area
water supply, provided a portable catch tank, and
now operate a semi-annual test program. Two men
work for six weeks on this routine to operate each
of the 544 safety showers and 44 evewash fountains,
correct any defects and reseal the valves for in-
stant use.

Progress in Waste Disposal

Hadioactive solid waste handling in the WWildavs
was a necessarily imprompin activity because of the
security situation, the urgency of demonsirating
reactor and chemical processing systems, and the
supposition that the entire project was short-term.
The physical boundaries of the first two "burial”
sites have been identified tentatively by “eve-
witness participants.” We are now opening up Solid
Waste Storage Area #6. To date, our disposal oper-
ations have handled five million cubic feet of solid
waste and used 135 ascores of ground.

In 1845 the 2300-V power lines to the Graphite
Reactor fans were the critical electrical distribu-
tion itern, Most of the plant was served by 480-V
circuits. The total plant load in 1947 was 2800 kVa.
A conceptual study at that timme set a maximum
future load of 5 MVa for 2 proposed new line from
¥-12. Today we have eight 13,800-V circuits serving
the general X-10 valey with a peak plant load of
36 MVa.

In the late ’40s a few air condifioning units
provided more stable conditions for counting rooms
and other instrumentation systems. “Comfort”
systems were taboo. The requirement for controlled
exhaust from labs and pressure differentials be-
tween office and other work areas for radiation
safety led {0 a necessity for general air conditioning
svatems. The X-10 site now has approximately
9,000 tons of refrigeration capacity in major sys-
tems and another 500 tons in window and wall anits.

Oax Rioce Namonal Laporarory Review

One of the goals of Laboratory management in the
early '50s was to convert the appearance of the site
into one of permanence. Prior to this time, spreading
crushed rock served to ease the mud problems.
Extensive grass-growing campaigns, the laying of
sidewalks and planting of shrubbery and ornamental
trees have all served to improve pedestrian travel
and the general landscape. Today, a Laboratory
Landscape Review Committee has approved long-
term plans to upgrade the general appearance of the
entire site.

These plans comprise a gradual conversion to a
“university campus” look appropriate to an East
Tennessee foothills location. Some areas will be per-
mitted to revert to their natural state, while others
will be maintained with “golf course fairway” care.

Organizational Structure

In the decentralized team approach to providing
these varied support activities, five area managers
act as deputies of the division director in coordi-
nating the intra-division efforts, determining sup-
port requirements, arranging priorities, and serving
as spokesmen for the P & E Division.

The role of the field engineer is most difficult to
define. It defies standardization. A 20-year evolu-
tionary process began when the engineer’s most
time-consuming role was to acquire the manpower
assignments he needed each morning from a “bull-
pen labor pool” and then to coordinate the various
craft efforts; his fiscal role was to write authori-
zations channeling costs into one of many money
pockets to which he had access; and the engineering
was whatever was necessary to get from the spoken
request to the finished job. Today he seldom coordi-
nates crafts, as we now have multi-craft foremen;
he has only a controlled access to financial authori-
zations and is called to account for expenditures;
but the classic engineering aspect of his job has
steadily increased.

Each field engineer assigned to a specific program,
operation or function has a different “job descrip-
tion,” depending on the needs and desires of the
requester. Broadly, his roles include that of inter-
face between requester and P & E, technical expert,
estimator, red tape handler, cost reviewer, inspector,
designer, scheduler, supervisor, problem solver, ex-
pediter, planner, and peacemaker. His responsi-
bilities range from serving essentially as a staff
member of the local operation, to simply being
available when needed. Each building at the X-10

FaLL 1969

site is assigned to a field engineer for surveillance
of maintenance effectiveness.

The primary purpose of the “on-paper” organiza-
tion of the division is to ensure a communications
flow among all groups affected by any problem, as-
signment, or proposal and to provide a common
interpretation for those Laboratory policies or proce-
dures requiring uniformity. There is no point in
describing administrative patterns.

However, a number of our activities have at-
tracted attention at other AEC laboratories and
plants. One is a foreman selection program that was
developed and demonstrated here with the assistance
of a University of Tennessee consultant, Cabot
Jaffee. This method employs hypothetical situations
and role-playing exercises designed to permit evalu-
ation of the candidates’ capabilities in leadership,
planning, organization, decision-making, and per-
suasion. The program has now been adapted for use
at the Y-12 plant and ORGDP.

Our “programmed-maintenance” system has
served as a model for similar programs at several
other laboratories. The ORNL Apprentice Training
Program administered by the Personnel Division
was the first such application of this craft training
technique in Union Carbide Corporation. In the P &
E Division 177 indentured apprentices have “topped
out” into journeyman status. Of this number, 132
are still employed here. Nine have advanced to
supervisory or technical aide positions.

Although industrial engineering may seem inap-
propriate at a research laboratory, the standard
research techniques of observation, data collection,
evaluation and analysis have proven beneficial in
improving many layouts and procedures. This is the
work of the Operations Analysis Group, which has
the simply stated assignment of “finding a better
way.”

As a majority of our people have a daily exposure
to mechanical, electrical, and chemical systems and
equipment, we have worked with the Safety Depart-
ment in fostering the Wise Owl, Golden Shoe, and
Turtle Clubs to spur interest in working safely.

To sum up, the entire subject of laboratory sup-
port could be considered to be fair game for the
organization consultant, the systems analyst, or a
psychiatrist. Although many changes and improve-
ments have taken place, we concede that we have
not achieved what may be desired in making every-
one in all programs satisfied all the time. However,
in old-time shop talk, our men and women have no
need to “back up to the pay window.”

19

We hear much these days about
priorities in science. With
the fraction of GNP for research
and development having fallen in
the past year from 2.5% to 2% (the
dollar amount has changed little),
everyone is being squeezed; hardly
a scientist or engineer is un-
affected by a decision taken some-
where about what should and
what should not be supported.

| became involved in this debate
about priorities quite a few years
ago when | was a member of the
President’s Science Advisory Com-
mittee. At that time people were
already talking about the nec-
essity to establish priorities in
science; but the discussion had
a rather theoretical flavor because
we were still enjoying budgets that
increased as much as 15% each
year. Having agreed in 1962 to
speak at UT before the annual
meeting of Phi Kappa Phi (the
land grant equivalent of Phi Beta
Kappa), | decided to order my
thoughts on priorities in science
and to use the UT talk as my ve-
hicle for so doing. The result, after
some re-edijting by Professor Ed-
ward Shils of Chicago and Cam-
bridge, was a series of papers on
“Criteria for Scientific Choice” in
the British magazine Minerva. (/I
had entitled the first paper “An
Agenda for Science,” but Profes-
sor Shils, editor of Minerva, sug-
gested the title under which the
papers finally appeared.)

| have gotten a lot of mileage out

20

of “‘Criteria for Scientific Choice.”
It seems that the question of pri-
orities in science had hardly oc-
curred to anyone prior to 1962;
but ever since, in crescendo, it has
become a central question in every
discussion of what is now known
as ‘“‘science policy”.

COMMENTS

My own formulation distin-
guishes between ‘“internal” and
external” criteria for scientific
choice. The internal criteria arise
from within the scientific field
under scrutiny: Are its practi-
tioners competent? Is it ripe for
exploitation? Do the people in the
field seem to have great enthusi-
asm? The external criteria arise
from the impact the field of science
has on universes outside itself—
on technology, on other science,
on sociely. In a general way, |
have argued that external criteria
are most important when what is
being judged is the support society
should give to a particular scienti-
fic field.

A few years ago | got into a
squabble with some of my friends
in high-energy physics because,

according to my scale, high-energy
physics did not rate high enough
on the external criterion scale.
Vicki Weisskopf, former Director
of CERN, was impelled to propose
his own criteria for scientific
choice, and | think his views are
well worth considering. Weisskopf
divides science into ‘“‘extensive”
and “intensive’ science. Exten-
sive science uses known basic
principles to enlarge our knowil-
edge of the things around us.
Solid state physics, most of chem-
istry, probably all of the environ-
mental sciences are examples of
extensive science. Intensive sci-
ence probes a very narrow seg-
ment of science so deeply as to
uncover completely new and un-
expected basic principles. High-
energy physics and modern cos-
mology are Professor Weisskopf's
best examples of intensive science.
In discussing which sciences de-
serve the most support, Weisskopf
urges that the intensive sciences,
even though they may bear rather
weakly on the rest of science,
nevertheless must not be over-
looked. To this | rejoin no, they
certainly should not be overfooked,
but neither should they be pushed
more strongly than anything else.
(The .Batavia 200 BeV accelerator
was just beginning to draw atten-
tion at the time of our discussion.)

These debates on scientific
choice have affected some of the
more formal deliberations about
the support of science. For ex-

0Ak Ripce NATIONAL LABORATORY Review

ample, in the 1968 report of the
AEC Panel on High Energy Physics
we read: “At the present there are
a number of important instances
which show the influence of high-
energy physics on the rest of
science. ... Thus, we find high-
energy physicists making many of
the most important contributions
to theoretical techniques in han-
dling many-body problems; to
computer technology; to the tech-
niques of dealing with ultrashort
time intervals; and to supercon-
ductivity technology. Not only the
methods but also the discoveries
themselves begin to have their im-
pact on other sciences.” Thus the
high-energy physicists invoke ex-
ternal as well as internal criteria
in arguing for their science.
During the past year, and espe-
cially during this time of budget
restrictions, | have come to realize
that priorities in science are set by
a political interplay much more
than by a priori philosophic wis-
dom. In this respect, science is no
different from any other enterprise
that operates outside the feedback
of the market place. The issue,
then, is not whether there is a
politics of pure science (some of
the modern scientific muckrakers,
such as Daniel Greenberg, seem
to be filled with righteous indigna-
tion by their discovery that such a
politics exists). It is whether this
politics is enlightening and trans-
cends the taints of venality, or at
least self-interest, that we some-

FaLL 1969

times associate with the word
politics.

| think the politics of science is,
on the whole, an enlightened poli-
tics; and | believe the widespread
debate on scientific choice has
affected this politics. Whether or
not particular philosophic formula-
tions prove to be acceptable, or
even directly applicable, the philo-
sophic debate has provided a lan-
guage and a framework in which
to conduct the political bargain-
ing. This was brought home to me
a month ago in a rather amusing
way when | was invited to speak
on natijonal science policy before
the Science Advisory Council of
Sweden.

Sweden, with only 8,000,000
people, is a much cozier place
than is the United States. The
Swedish Prime Minister, His Ex-
cellency Tage Erlander, takes an
active interest in formulating
science policy. He therefore
chaired the meeting in the ornate
nineteenth century Swedish Parlia-
ment Building at which | outlined
my views on science policy, the
politics of science, and the rela-
tion between the philosophy and
practice of scientific choice and
scientific decision-making. Cer-
tainly the questions that concern
the Swedes are very much like
those that beset Americans.
Should Sweden contribute to the
proposed 300 BeV CERN acceler-
ator? Should Sweden have a De-
partment of Science? What about

redeployment of Sweden’s large
laboratories, particularly its nu-
clear energy establishment at
Studsvik?

| used our experience in desalt-
ing research to illustrate how Oak
Ridge has achieved a measure of
redeployment. And in describing
our experience, | naturally men-
tioned —and possibly even boasted
—about the skill of our many
chemical and mechanical engi-
neers who have contributed so
much to improving the technology
of desalting. Altogether, | thought
my remarks were non-controver-
sial and that they were rather well
received. They were, except when
one Swedish biologist in the audi-
ence arose to ask why | had not
discussed, as he put it, “the re-
markable group of biologists at
Oak Ridge.” “Everyone knows,”
he continued, ‘“that biology is
much more important than desalt-
ing.” To this | responded that | like
both biology and desalting, and |
wasn’'t about to make a choice
between them.

All of which illustrates how diffi-
cult it is to make choices between
scientific fields, even when the
political atmosphere is as respect-
able as the Swedish Halls of Par-
liament, and the discussion has
been enlightened by a spirited
philosophic dialogue.

%«%7447

21

Benefits vs. Risks in Nuclear Power

A logical, facts-and-figures comment on the current anti-atom literature

By WALTER JORDAN

UST A FEW YEARS AGO almost everyone
looked forward to the coming age of nuclear
energy as a boon to mankind. Of course the coal
interests have always been less than enthusiastic
but that was to be expected. Recently, however, sev-
eral freelance writers have undertaken the role of
professional critics. Perhaps best known is the book,
“The Careless Atom” by Novick, but the writing
that recently stung me was an article in the March
issue of Natural History by Curtis & Hogan entitled
“The Myth of the Peaceful Atom,” since expanded
into a book, “The Perils of the Peaceful Atom.” 1
felt particularly betrayed in this instance, for I have
long considered myself a conservationist and have
read with satisfaction of some of the crusades for
conservation of resources, wildlife, or beauty that
magazines such as Natural History frequently carry.
Certainly one of my strongest motives in pro-

22

moting nuclear energy has always been the con-
serving of the fossil fuels —coal, oil and gas. It is on
these valuable and irreplaceable resources that
most of the aspects of modern life we take for granted
depend: the production of steel, lubrication of ma-
chinery, synthetic fibers, plastics, the bulk of the
chemical industry. These fast-diminishing fossil
fuels are indispensable. And so, knowing that they
can be conserved by the employment of a new source
of power that at the same time reduces atmospheric
pollution, I feel a real global mission in persuading
people of these benefits of nuclear energy.

But the professional critics, who are being joined
by some conservationists, say that all these fine
benefits are not worth the risk. I strongly disagree.
In fact I believe that more lives have already been
saved by the advent of nuclear energy than will be
lost as a consequence of it in the next 100 years.

OAK RIDGE NATIONAL LABORATORY Review

Left: "The demand for electricity has almost doubled
in the past ten years. Another doubling is
projected for the next decade.”

FALL 1969

A native of Montana, Walt Jordan taught
physics for six years at the University of
South Dakota after getting his Ph.D. at
Cal Tech. From there he joined the MIT
Radiation Laboratory in 1941, where he
worked on the development of radar for
use in WWIl. In 1946 he moved to the
Physics Division of Clinton Laboratories,
as ORNL was originally called, to work
with P. R. Bell in developing nuclear elec-
tronic equipment. He was director of the
Nuclear Propulsion Project in the 1950's,
and in 1959 became Assistant Director of
the Laboratory, the title he now holds. He
is currently Advisory Editor of the AEC
Nuclear Safety Quarterly, and a part-time
professor in the nuciear engineering de-
partment of the University of Tennessee.
A version of this article was first delivered
as a talk to Oak Ridge Associated Uni-
versities’ third Science for Clergymen
Conference last July.

A Swarm of Controversy

There does indeed seem to be developing a swarm
of controversy over the growing nuclear technology.
If it were just an occasional book or article I would
be inclined to hold my peace. Unfortunately, it is
deeper than that. Part of the licensing procedure for
a nuclear power plant (though not for any other
kind) stipulates that a public hearing be held at
which individuals may intervene. In some cases
these hearings have been so protracted that the
power company has withdrawn its application rather
than face the continued publicity. A power plant
planned for construction at Bodega Bay has been
abandoned. The opposition was concerned mainly
with the natural beauty of the proposed site but the
issue of earthquake damage was the deciding factor.
New York Gas & Electric Co. has decided to with-
draw its application to build a nuclear power plant
at Ithaca. In this instance the interveners protested
the possible thermal pollution to Lake Cayuga. The
utility will build a coal-fired plant instead.

23

“The reduction of this
outpouring of noxious
gases is imperative.”

First of all let me summarize some of the benefits.
One reason why power reactors are being installed
in so many places in the United States (some 50 nu-
clear power plants have been ordered; 10 are in
operation) is that they save money. If you are in an
area served by one of these power reactors your elec-
tricity is costing less than it did with fossil fuel
plants.

The demand for electricity has almost doubled

within the past ten years. Another doubling is pro-
jected for the next decade. If this accelerating de-
mand is to be met, more and bigger power stations

24

will have to be built. The alternative is an enforced
rationing of electricity.

Although nuclear energy is beginning to supply
some of the growing demand for power, fossil fuels
are being burned at an ever-increasing rate. Whether
they will be gone in 50 years or 200 years is not
certain, but the time is short in comparison with the
hundreds of millions of years that it took to form
those deposits of coal and oil. Qur limited reserves
are fast going up in smoke.

And smoke there is! From a large, coal-fired power
plant such as Bull Run near Oak Ridge, hundreds of

0OAK RIDGE NATIONAL LABORATORY Review
tons of noxious sulfur oxides pour out every day.
What’s more, thousands of tons of CO, are emitted
daily by Bull Run. (It has been observed that the
CO, concentration in the world’s atmosphere is
increasing at the rate of about 2% per decade, a
change that may have implications for long-term
effects on climate.) No longer is the air clean and
pure in Tennessee Valley: or, indeed, in most of the
United States. Eyes burn, lung diseases worsen,
pine trees drop their needles, and in extreme cases
people sicken and die. This happened in 1942 in
London, where a massive atmospheric inversion was

FaLL 1969

Manhattan from Riker’s Island.

At right are the stacks of Con Edison’s

Ravenswood plant, which manufactures power for all parts of city.
To left, the company’s Waterside plant.

followed by an increased mortality of an estimated
3,500 to 4,000 victims; and again in 1948 in Don-
norra, Pa., when 48% of the population were affected
by the smoke-filled atmosphere to which were at-
tributed 20 deaths by the third day.

The reduction of this outpouring of noxious gases
is imperative. It can be done by removing them from
the smokestacks (thereby increasing the cost of
electricity) or by installing nuclear power plants.
Nor are coal-fired power plants the chief contrib-
utors to the air pollution of the country: auto-
mobiles and trucks represent a major source as does
the heating of homes and buildings. To reduce this
pollution caused by combustion, a general con-
version to electricity will have to ensue. Homes
must be heated electrically and automobiles and
trains driven electrically. This will require that we
generate at least three times as much electricity as
we now produce, a challenge that can only be met
economically with nuclear power plants.

Nuclear power offers a virtually inexhaustible
supply of cheap electricity. Moreover it offers a
chance to clean up the atmosphere. But there is in
addition a third major benefit that has come with
the nuclear age and that is the myriad uses of radio-
isotopes. These isotopes which are produced so
copiously in every nuclear power plant—and, in-
deed, represent the chief danger in the operation
of these plants—have already proven to be a great
boon to mankind. Estimates of the benefits of these
isotopes to industry are of the order of a billion
dollars a year. Every major industry has found
dozens of uses for them. Radioisotopes can measure
the thickness of paper in the paper mill or the thick-
ness of sheet steel in a rolling mill, the level of a
liquid in a tank or the flow of oil through a pipeline.
Isotopes are used in the exploration for oil. A slow
leak in a water main or a gas line can be found with
an isotopic tracer. The gamma rays from cobalt-60
are used for “x-raying” welds; they are also used in a
chemical plant to produce new plastics. The dra-
matic uses of radioisotopes have caught the eye of

25

gvervones, Some insect pests have been overcome by
the ingenious method of sterilizing a large number
of male insects with radiocisoctopes before turning
them loose during the mating season. Irradiating
wheat before it is stored can greatly reduce the
losses due to insects. Spoilage of potatoes or straw-
berries can be greatly reduced by gamma radiation,
We are just beginning to sse the introduction of
radioisotopic powey plants for use in space satellites,
undersea beacons, and heart pacers, to mention a
few. Just to list the applications of radicisotopes
would make a thick dovument.

The felds of binlogy and medicine have been rev-
olutionized by the introduction of radinisotopes.
The number of shipments of radicisctopes to hos-
pitals in this country is in the many thousands
every day. At Johns Hepking University Hospital
about 400 patients a day are given diagnosis with
this new technigue. Thyroid tumors are spotted by
the radiciedine that is concentrated there, brain
tumors by their affimty for radicactive fechnetivm.
Heart funciion can be measured by the length of
titne it takes for chromium-51 to become diffused
into the blood siream. These medical procedures
are typical of the hundreds of spplications of iso-
topes in modern hospitals in the diagnosis and treat-
ment of dizease. And although nuclear power plants
are not required for the produciion of these valuable
isotopes, nuclear reactors are. And | think that the
critics would be hard put to argue that research re-
actors for isotope production are somehow safer than
reactors for production of electricity.

1 could easily devote many pages to the benefits of
nuclear energy. However, there are also risks. Those
radivactive isotopes that are so useful when prop-
grly purified and diluted also represent a major
hazard. The possibility, no matter how remote, of

26

spreading millions of curies of radivactivity over
the countryside i3 not a pleasant one to contem-
plate. The critics present a gloomy picture. How
likely is it to happen? Before discussing that ques.
tion, I would like to recall for you some of the risks
that you encounter in everyday life.

Hazards of Living

In order to get g feeling for the numbers involved,
consider the probability that an average member of
the popuistion will die during the next hour due to
disease such as heart failure, cancer, sic. The figure
is about one in a million, so 1 would write it

P = 10"%hy

it appears that people are willing to accept risks
of aboat that same magnitude provided it is volun-
tary and the benefits are personal and real For
example, the risk of being killed while riding in an
auto is about 107%hr of exposure, about one-tenth
of what it was 8 generation ago. There have indesed
been significant advances in asutomobile safety. The
risk of riding in a commercial airplane is also about
1075 hr, which means that air travel is some ten
times safer than auio travel on a mileage basis.
Alr travel in private planes iz 2 much more dan-
gerous undertaking, fatalities in these fights arve
some 20 ® 107%hr of exposure; 20 times gs risky as
commercial air travel. And yet many people wil-
lingly take the risk. But they do it of their own free
will! No one imposes the risk upon them. It would
appesy that so long as the individual has a choice
he is willing to accept risks considerably greater
than his normal risk of dying by disease, provided
the henefit to him is very real and immediate. On
the other hand, if the risk is imposed upon him

Oax Bipor Namional Lanorarory Review
“The situation is remarkably similar to the
controversy over the budding electric
power industry in the latter part

of the last century.”

(such as an airplane falling on a busy street, as
recently happened in Miami, or the explosion of gas
mains in New York City) he will insist that the
probability of death be much less than the normal
disease death rate. He will live below a dam if he is
convinced that the chance of the dam collapsing is
very remote (perhaps 10~¢ per hour of exposure) and
there is good reason (benefit) for him to live with
the exposure to a small but not zero hazard. He may
protest if a chemical plant or a nuclear power sta-
tion is built near his home —suggest that it be built
in another location—but if he is convinced that the
risk is small, he won’t move. A small risk is, as we
have seen, something less than 10~%/hr of exposure,
or 10~4/yr. In other words, if he is convinced that a
major catastrophe will happen only about once in
every 10,000 years, he will feel that the risk is ac-
ceptable. Will Los Angeles and San Francisco be
spared a major earthquake for that long? Less than
50 years ago 150,000 people were killed in Japan as
a result of an earthquake.

Nuclear Risks

How can we demonstrate that the risk of living
near a nuclear plant is small? Only by experience.
The situation is indeed remarkably similar to the
controversy over the budding electric power in-
dustry in the latter part of the last century. There
was a great deal of opposition to the introduction of
electricity into the home. The critics pointed out
that electricity was dangerous, that people would be
electrocuted, innocent children would stick their
fingers into electric sockets and die a horrible death,
that wires would become overheated and burn down
the homes. Of course they were right! A thousand

Fari 1969

people in the United States are accidentally elec-
trocuted every year. Moreover, it has been estimated
that 16% of the fires are electrical in origin and
1,200 people lost their lives last year in the United
States due to these electrically originated fires.
However, this risk of being killed by electricity is
low: about 10~*/hr of exposure, well below the “ac-
ceptable” risk of 10~%, and the benefits of electricity
are so apparent to everyone that no one wants to
turn back the clock.

Let us now turn to the risks of operating nuclear
power plants. These can currently be classified
into three categories:

1) Thermal pollution of the rivers and lakes.

2) Low-level release of radioactivity into the air
and ground waters due to the normal operation of
nuclear power plants and reprocessing facilities.

3) The accidental release of large amounts of
radioactivity.

To my mind item 3 is the risk of most concern, but
the critics (Chauncey Starr calls them “nuclear
hypochondriacs”) have been equally vociferous about
items 1 and 2.

Thermal pollution is not a new phenomenon, nor
is it confined to nuclear power plants. Many indus-
trial plants generate a large amount of heat and it
is much less expensive to dump the waste heat into
a river than it is to release it to the atmosphere.
The rivers that flow through Pittsburgh, for ex-
ample, are raised in temperature by 20 or 30 de-
grees. This has had an adverse effect on the fish and
has in general upset the ecology. Federal standards
are needed and enforcement by the states is most
desirable. Such legislation is now pending in Con-
gress. These regulations should apply to any plant,
be it nuclear, fossil-fueled or chemical. Nuclear
plants should conform, no more or less than any

27

R —

W “ ..I believe that more lives have already been saved

other type. It is true that a nuclear electric plant
dumps more heat into a stream than a fossil-fueled
plant of corresponding electric power output. But it
doesn’t make sense to raise a storm of protest over a
nuclear plant of 500 Mw electric capacity while a
1,000 Mw electric fossil-fueled plant escapes almost
unnoticed. Indeed, New York has passed legislation
limiting nuclear plants but not conventional plants.

It is not surprising that a nuclear power plant
which generates millions of curies of radioactivity
may discharge a very small amount of radioactivity
into the atmosphere or waste stream. The whole ar-
gument has to do with defining a “small amount” of
radioactivity. The nuclear critics insist that it
should be zero for a nuclear plant, whereas they rec-
ognize that a coal plant does emit some radioac-
tivity owing to the presence of a small amount of
uranium and its daughter products in the coal. In-
deed, a large coal plant releases to the environment
more radioactivity than any present or planned nu-
clear power plant, but the argument is made that
since it only redistributes the activity from the coal
mine to the atmosphere, it is somehow less harmful.

Actually we know very much more about the ef-
fects of radiation on the human body than we do
about the effects of various chemical pollutants that
occur in ever-increasing amounts in the air we
breathe and the water we drink. In full acknowl-
edgement of the potential hazards, the AEC has
spent hundreds of millions of dollars in biological

28

by the advent of nuclear energy than will be lost
as a consequence of it in the next 100 years.”

research aimed at establishing not only the effects
on man but also on the environment so that we can
be certain that the ecological effects will be minimal.
This concern is almost without precedent: Certainly
the automobile industry has not expended very
much money on the effects of smog on the popu-
lation, or the tobacco industry on lung cancer, or the
chemical industry on the effects of DDT on the eco-
logical cycle. One of the favorite expressions of the
nuclear critics is that there is enough radioactivity
in a reactor to irradiate everyone in the United
States with a lethal dose. There is also enough in-
secticide manufactured to poison every U. S, citizen;
moreover, the insecticides are meant to be widely
distributed, whereas the radioactivity is carefully
confined.

As a result of the intensive research that has gone
on over the past 25 years on the effects of radiation,
the Federal Radiation Council has developed a set
of radiation protection guides. The levels that have
been set, even for workers in the nuclear industry,
are meant to be at least an order of magnitude below
that where somatic effects on the individual would
be observed. (This is in contrast to the ozone level
in Los Angeles, which is set just barely below the
level where eye irritation will be noticed.)

If everyone in the nuclear industry were to get
the maximum level of 5 rem per year, there probably
would be a small increase in the observed number of
deaths due to leukemia after a number of years —but

OAK RIDGE NATIONAL LABORATORY Review

“Nuclear power offers

a virtually inexhaustible

supply of cheap electricity

. . . a chance to clean up the atmosphere.”

still a lower number than those from other indus-
trial accidents. Actually it is rare for anyone to get
5 rem during a year and most of us get very much
less. Even though 5 rem is considered to be a con-
servative figure, much less for example than radiol-
ogists used to take, it is felt that an additional
factor of 30 reduction should be made when con-
sidering the dosage levels to the population at large.
Hence the protection guides on the escape of radio-
activity limit the amount of activity to such a low
level that the general population will receive no
more than a fraction of a rem per year. Actually
everyone receives something like a tenth of a rem
per year of radiation due to cosmic rays and natural
radioactivity in the earth and air —everyone, that is,
but those who live in certain high-level radiation
areas, like some in India, where they receive eight
times as much. When one adds to this the radiation
due to medical x rays it is apparent that the amount
contributed by nuclear energy is small in com-
parison. I do not hesitate to take several rem of x
rays when it is needed for diagnosis or treatment of
disease. Here is a very real example of the benefits
far outweighing the risk. On the other hand I am
opposed to taking even medical x rays needlessly.
Some of the older machines for dental x rays sprayed
the whole body; the use of a filter and cone, which
produces better pictures with less radiation, is now
in practice. X-ray machines in hospitals have also
been greatly improved; good, clear lung radiograms

FaLL 1969

can be obtained with a dose of 0.1 rem rather than
the several rems required with poor equipment and
antiquated procedures.

It has been estimated that the maximum exposure
received by any person in the neighborhood of the
Dresden nuclear power station was less than 0.0001
rem per year, which is less than 1% of the exposure
they receive from natural radioactivity and cosmic
rays. But in spite of the ultra-conservatism in the
federal radiation protection guides, the Minnesota
Pollution Control Agency responsible for water pu-
rity has recently protested the granting of a license
to operate a reactor unless the operator guarantees
to maintain a level of activity release a factor of 100
below the values recommended by the Federal
Radiation Council. This may induce the utility to
put in a coal-fired station which will put out more
radioactivity than the Minnesota agency is at-
tempting to legislate plus the products of combus-
tion, with their known and unknown hazards. All in
the name of “safety.” I believe it is demonstrable
that the hazard from the presently regulated amount
of radioactivity released in normal operation of a
nuclear power station is much less than that from
the pollutants emitted by the operation of a fossil-
fueled station.

However, the risk of releasing a large amount of
activity inadvertently (item 3) is quite another
matter. The hypothetical consequences of such an
accident were the subject of a much publicized

29

e

Brackhaven report some 10 years ago. The authors
assumed the worst possible combination of circum-
stances. They gave no credit for containment in es-
timating that half of the fission products would be-
come atrborne; they assumed that the accident
would cocur during an atmospheric inversion and
low wind velocity, so that the fission products were
carried straight toward a population center with
very little dilution or mixing. Under these catas-
trophic but highly unlikely circumstances up o
3,000 people could be killed, assuwming evacuation
were not possible. For some reason it is much more
acceptable 1o the public to kill 10,000 people in 8
series of small accidents than to kill 3,000 1n a single
event. Nevertheless it is the stated mission of the
nuclear industry and the regulating agency to make
the possibility of such an accident exceedingly re-
mote. How do we go about it?

First, the fission products are contained in fuel
elements which would melt ondy if cooling were {0
fail. Second, the fuel elements are contained within
a primmary conlant circuit which undergoes the most
thorough series of tests and inspections that any
pressure vessel has ever been subjected to. Then the
whole works is contained within a large steel or con-
crete containment vessel. Finally there iz an ex-
clusion area surrounding the power plant and a low
population zone outside of that. This should resulf
in considerable dilution of the radicactive fission
products before they reach the population center as
well as introduce a delay so thal evacuation can
begin.

In order for the radivactive Hssion products to es-
cape, the fuel elements must melt, the primary
vesgel must burst and the containment vessel fail
And should all these failures coour coincidentally, it
appesrs that probably no more than 5% of the fission
products would become airborne rather than the
50% sssumed in the Brookhaven report. Even so,
the release of 5% of the radivactive products under
unfavorable atmospheric conditions would be sert-
ous, And we can see ways that that might happen.
However, bear in mind that when & mechanism for
an event can be posiulated, then the design can be
modified to make that particular mods of scourrence
most unlikely. It is {ruoe that fate has a way of fi-
guring out another path to an incident that was not
foreseen. But the designers and builders of nuclear
power plants have exercised sophisticated ingenuity
and expended large sums of money to make the
plants as safe as they know how to make them.

S

There have been accidents and releasses from ax-
perimental reactors. The releases have been small
and no member of the public has been injured. A
power reactor in Windscale, England, caught fire
and did spread a considerable amount of radiciodine
over the countryside, thereby contaminating milk
supplies and crops. The reactor was not in a con-
tainment vessel so perhaps 2% of the fission prod-
ucts did escape. No power reactor in the United
States has been similarly involved. There were some
fuel elemends that melied in the Fermi reactor but
neither the primary nor the secondary containment
were viclated, The prophets of doom have heavily
dramatized these reactor incidents, pointing out
that it can happen in spite of our best efforts. How-
ever, don't these accidents prove, rather, that a
fairly major release of radivactivity can ocour, such
as Windscale, or SL-1, without anyone outside the
reactor building receiving a tolerance dose of
radiation?

The $64 million question still remains, and that
is whether we have succeeded in reducing the risk
to a tolerable level —ie., something less than one
chance in ten thousand that a reactor will have a
serious accident in any year., Thus when we have
one hundred nuclear power stations in operation,
not oo far in the future, an accident once every
hundred years might be expected. And if 2 hundred
people were to be killed, such as now happens in a
major airline disaster, it is 8 lower calculable visk
than that taken by many facets of U. 8. industry
today.

Have ws succeeded in reducing the hazards to
such a low level? We can only say that we have ac-
cumuiated so far some 100 reactor years of accident-
free operation. That is a long way from 16,000 so
it doesm’t tell us much,

The only way we will know what the odds really
are is by continuing to accumulaie experience in
operating reactors. There i5 some risk but it is surely
worth it. 1 am impatient with those whe cry "wolf”
when there is so much to be achieved. Un the other
hand it is a mistake to use the head-in-the-sand ap-
proach and say it can never happen to us, We and
the public should be prepared to face the possibility
of a nuclear incident, just as we live with the pos-
sibility of major earthguskes which will exact a
large toll in property and lives. Ondy a few people
atdvocate abandoning the West Coast. 1 hope only g
few advocate abandoning nuclear energy that prom-
ises so much for mankind.

Oax Bmor Nationarn Lasorarony Eevisw
By Alex Zucker

GUILTY GAMES

Black Rain, by Masuji Ibuse, translated by John
Bester. Kodansha International, Ltd., Tokyo and Palo
Alto (1969). 300 pages, $6.95.

The Game of Science, by Garvin McCain and Erwin
M. Segal. Brooks/Cole Publishing Company, Belmont
(1969). 171 pages + index, $2.95.

ANNIVERSARIES are a form of social conscience.
Nations, religions, or families remind themselves
of significant past events by noting centennials and
birthdays, by collective celebrating or mourning,
and sometimes even by erasing from the calendar
occasions of import. In the current fashion we note
anniversaries by writing books, publishing long
articles in newspapers, and pasting together old film
clips over a modulated narration for a television
program. We now celebrate our centennials, their
multiples or rational fractions, in private, with sub-
dued hoopla, with appeals to the intellect rather
than to the emotions. For example, last July 20th,
while the headlines were all turned on the moon,
the Germans quietly but insistently celebrated the
25th anniversary of the ill-starred attempt on
Hitler’s life by a group of army officers. From this
attack on Hitler the Germans derive a much-
needed belief that they were not all bad even in
those days, that there were a few people in Germany

FaLL 1969

who placed justice and humanity above their own
welfare. This gives them something to build on, an
assurance that the present democracy is founded on
real values, that it will not fall away like a mask at
the earliest opportunity, to reveal again the jack-
booted storm trooper. Such feelings require public
reminders, and they are on the whole useful to a
society justifiably ridden with guilt.

And how about us? Next year we too will have a
quarter century to celebrate: Alamogordo, Hiro-
shima, and Nagasaki. A quarter century of new
words: kiloton, atom bomb, radiation sickness,
mushroom-shaped cloud. The celebration has al-
ready begun. Paradoxically it is taking the form of
an attack on the peaceful aspects of the nuclear
world, perhaps because the huge arsenal of weapons
(how felicitous a euphemism, it makes one think of a
club or a spear rather than a super bomb) is difficult
to grasp, difficult to describe in human terms, and
somehow beyond the public reach. Still, some writers

31

Cover design for "Black Rain.”

are sure to try to celebrate Hiroshima in the most
direct form, and one of the most interesting har-
bingers of that flock is Masuji Ibuse, a well-known
Japanese writer. His book, “Black Rain,” is in the
form of a diary; the diarist, Shigematsu Shizuma, is
a real person, his diary exists, although it must have
been expanded and completely recast by Ibuse.

On the way to work in Hiroshima on the morning
of August 6, 1945, Shizuma was about to board a
train in Yokogawa Station. “At a point three meters
to the left of the waiting train, I saw a ball of blind-
ingly intense light, and simultaneously I was
plunged into total, unseeing darkness. The next
instant, the black veil in which I seemed to be envel-
veloped was pierced by cries and screams of pain,
shouts of ‘Get off!” and ‘Let me by, curses, and other
voices in indescribable confusion. The passengers
came pouring out of the car. I was squeezed off the
deck and flung onto the tracks on the opposite side
from the platform, landing on something soft that
seemed to be a woman’s body. Another body landed
heavily on top of me in turn. More bodies were piled
up on either side of me. A cry of pain and rage
escaped me, to be echoed in my ear by a similar cry,
in a heavy local accent, from a man whose head was
jammed against mine. With cries and groans rising
all about me, I shook myself free of those lying on
me and struggled to my feet. I pushed out with all
my strength, thrusting others out of the way, until
eventually I found myself being buffeted from behind
against something hard. Recognizing it as the edge
of the platform, I elbowed people out of the way and
clambered up onto it.” This is the private view of a
moment in history. A description of just what hap-
pened in Hiroshima on that morning follows. Refu-
gees by the thousands stream out of the city, not

32

knowing where to go, or what happened. Others
traverse ground zero going from one point in the city
to another in search of their families, their homes,
their children. The disaster is unimaginable.
Human beings are mutilated in unprecedented
ways; mostly by the heat flash which causes deep
painless burns on unprotected areas of flesh. Quickly
these begin to ooze. Hands reach up to find the cause
of discomfort, they come away slimy. How could it
be so bad, if there is no pain? They look at others to
comprehend the horror of themselves.

Still the question remains, what happened? Theo-
ries abound; the most popular is that some infinitely
poisonous gas has been dropped accompanied by
many bombs. But how is it possible from one B-29?
And what about the cloud? “. . . really shaped more
like a jellyfish than a mushroom. Yet it seemed to
have a more animal vitality than any jellyfish, with
its leg that quivered and its head that changed color
as it sprawled out slowly toward the southeast,
writhing and raging as though it might hurl itself
on our heads at any moment. It was an envoy of the
devil himself . . . who else in the whole wide uni-
verse would have presumed to summon forth such
a monstrosity?” This is one of Shizuma’s rare
outcries.

Really, the fire was the worst. Blast and heat flash
were confined to the center, but the fire storms swept
the entire city. Many were trapped in the fires, most
of the houses destroyed. Each house, each family,
every individual has a story to tell of that day. For
example: A father is trying to free his son who is
alive, but pinned by a collapsed building. As the
father works, frantically straining at the heavy
beams, the fire approaches, threatening to engulf
both of them momentarily. The ruined house blazes
up, the father looks at his son, turns, and runs away.
There is a crack and a rumble, the timbers in the
house shift, the son is freed, he jumps up and runs in
the opposite direction. Not all tragedies end in death.

Shizuma goes in and out of the city for several
days after the bomb was dropped. Horror follows
horror in his diary, until there is no pity left. Numb-
ness sets in. The reader begins to feel that the de-
stroyed city is, after all, a normal occurrence; that
violent death, radiation sickness, wandering
orphans, are simply man’s natural lot. Ibuse is
trying to evoke in his audience the same abrasion of
the senses that must have developed in the survivors
of Hiroshima. There is no anger. Shizuma, his
friends, or his family don’t seem to hate anyone. They
don’t rail at the Americans for dropping the bomb,

OaAxk RiDGE NATIONAL LLABORATORY Review

they show no resentment of the Japanese rulers for
having led their people to this fate. Many are not
even curious to discover exactly what caused that
blinding flash. At worst they seem perplexed that
anyone would go to such extremes, that anyone
would want to harm people to such an extent in
order to secure victory.

The diary ends on August 15, with the surrender
of Japan. This news is more profoundly shocking to
the Japanese than the bomb, but they accept it too,
fully expecting that the American occupation will
bring a greater disaster to Japan than anything that
preceded it.

Shizuma survives, he has told his story, he draws
no moral from it, he exercises no judgment. He sur-
vives, his wife survives, and his niece who has
radiation sickness because she was caught in the
fallout, the muddy black rain, survives and may
even get married. In Hiroshima, at least, all life
was not extinguished.

S— e

It is doubly disconcerting to follow a book like
“Black Rain” with “The Game of Science.” Cer-
tainly no one, since Hiroshima, has thought of
science as an innocent game, “simply and purely a
search for knowledge,” as the authors say. With a
large measure of callousness one might call war a
game. Trading in stocks and bonds or running large
corporations may be a game. Dealing in art objects,
the popular entertainments, even politics may be
likened to a game. But the results of science are
terribly serious and permanent; they have the power
to change man’s view of himself, his world, and his
place in it. The question is not how the game of
science is played: what are the amusing rules, who
the inveterate players, who the champions, who the
villains; the problem is how mankind can absorb
new information, and profit by it without doing harm
to itself. To call science a game, to reduce it to its
elements which indeed bear some resemblance to
play, is to deny its surpassing contributions to man'’s
intellectual and material progress and to shun the
responsibility for the awesome perils that can be
brought about by scientific work. Games have a
beginning and an end; science is forever.

Right now we need books, as we never have before,
to explain what science is all about, to convey-to the
non-scientist the methods of science, the relation-
ship between applied and basic research, the diffi-
culties involved in finding anything new, and the

FarLL 1969

rewards that come with understanding science,
quite apart from participating in it. It is probably
correct to say that discovery in any branch of science
enriches every scientist. With the aid of popular
books like this it ought to enrich all men as well. It
is here that McCain and Segal really fail.

The authors describe what science is, how it is
done, and who does it. They have a gift for writing,
the book never flags, the tempo is sustained at the
pitch of excited expostulation. It is what they have
to say that is boring.

To explain the structure of science the hierarchy
is set up once again with basic research at the pin-
nacle. Again the whole body of science is revealed to
contain only the noblest human motivations, to be
pure even beyond the fantasies of a detergent copy
writer. Is this a realistic fact to present to the lay
reader? Ought not one try for a balance: science has
done marvelously well for us; it has also created great
troubles. But fortunately science can find a way to
fix many problems, if it is given the job. Is this not
preferable to McCain and Segal’s attitude that
science is a great blessing and that the evils associ-
ated with it are somebody else’s baby?

In many cases the book is just plain wrong, as
for instance, “The level of abstraction varies
greatly. . . , with the scientists’ theories often far
removed from ordinary sense data, whereas the
practitioner deals with the here-and-now.” And why
is it that every explanation of inductive reasoning
eventually rests on the sun rising tomorrow? This
is far from a universal phenomenon, and it would
clearly surprise an Eskimo to learn that induction,
so fundamental to scientific thought, rests on the
belief that the sun will rise tomorrow. Think up
something original, can’t you, McCain and Segal?

The book has its merits. Both authors are psychol-
ogists, a fact carefully concealed in the text, and
when they leave science and turn to the scientists
themselves, things pick up. Here the reader unfa-
miliar with the scientists is likely to come away with
a pretty clear picture. Unlike science, the scientists
are not lily white. They are motivated by the same
sort of feelings that are common to us all. They are
believable human beings, and the only general
requirement set out is that they have an IQ greater
than 120, and at least a tolerance for mathematics.
While in his private life, a scientist can be “an
alcoholic, a braggart, a lecher, a thief-or even a
member of a civic club, and still be forgiven; but one
instance of his deliberately falsifying data brings
eternal condemnation by his peers.” This is quite

33

true, and must sound odd to a used-car salesman.
There is also a gossipy section which classifies
scientists into “players,” “operators,” and “by-
standers” and gives some characteristics of each
subspecies.

McCain and Segal talk at length about the train-
ing of a scientist, about his formal as well as his
informal education. They advocate a broad, unspe-
cialized curriculum for the aspiring candidate. “The
plea is for a liberal education at the undergraduate
level — attuned to the broad reaches of current knowl-
edge and thought.” And specialization is not only in
science: . ..in a typical business-administration
program the student gets approximately twice as
many specialized hours as does a physics major.
Perhaps business is more complex than physics; or
perhaps the object of a business school is to produce
the equivalent of a trained seal. Among their other
excellent qualities trained seals perform on com-
mand, present a uniform, neat appearance, and slide
smoothly through the water, creating no waves.”

34

Now here is a statement that is sure to gladden
the heart of a physicist or a seal with aspirations
businesswise. It may even attract an unsuspecting
youth to a career in physics, where he will soon
find out that his education is very specialized and
that he will be attuned to the broad reaches of cur-
rent knowledge only at the price of great personal
commitment.

McCain and Segal display an astonishing breadth
of interest in the closing sections of the book.
Chapters are entitled “Science in the World,”
“Ethics, Free Will and the Scientific Study of Man”
(14 pages for all this), and there is even a section
describing how a scientist should write a research
proposal to the funding agencies. This is all pretty
shallow stuff. The impression one gets is that the
authors have emptied their heads of all material
about science or in any way related to it, and strewn
it, like the results of a weekly shopping trip to the
A & P, on the kitchen table. Santayana described
it: to “spread a feast of what everybody knows.”

Jacket design on "Game of Science.”

OAK RIDGE NATIONAL LABORATORY Review

FaL1 1969

Herb McCoy first worked at ORNL as an
undergraduate codp student out of Vir-
ginia Polytech in 1953. In 1955 he married
Ann Thompson, a Laboratory Personnel
Division employee, and transferred to UT,
where he received his M.S. in metallurgy
in 1958, the year he joined the ORNL staff.
His work at the Laboratory has been with
Aircraft Nuclear Propulsion, the Gas-
Cooled Reactor, and currently the Molten
Salt Reactor Experiment. McCoy heads up
a group that is engaged in fundamental
studies on the mechanical properties of
metals in the Metals & Ceramics Division.
The story he tells here of the development
of the particular alloy that would stand up
under the exacting demands peculiar to
the MSR is one of innovation and ingenuity.

The INOR-8 Story

By H. E. McCoy

HE PHYSICISTS’ NOTEBOOKS are filled with

attractive reactors that probably will never be
built because of inherent materials problems. The
factors usually involved are excessively high op-
erating temperatures and corrosiveness of the fluids
involved. Other reactors have been constructed
without full consideration of the component materi-
als, and their operation has been plagued with prob-
lems. The Molten Salt Reactor Experiment at ORNL
might well have been one of these problem cases,
had not the talents of several individuals and the
facilities of the Laboratory and several commercial
metal producers been applied to developing a struc-
tural material for the metallic parts of the reactor.

35

4‘:’"»1
’v‘t
‘-
-
S P
o, . &
)
-
".

o9 *
L
..'

z s 2 A3 kg e e A
S eyl ey R Y S
;.r t '..:.,‘: .'}".-'.‘l"g "’u‘) : : " ’:‘ ":.“‘\. 'l"““
e y i v = - ) - - o
.‘ '-." f.: “.="'. ‘0.:'-’ M m'\. ’.5‘}5.9.?" alfa.lgt""“\.‘
et WA A r':‘;\'\f‘-—-{\{“r-'.. !‘ a2, .,
vl T i re g AN e T gy et $:
e A e MPYT B “o \'\.fi.‘lfl » o ,)- A e )
;_n‘« .I";f. i-_.p.w—-. v g;‘\u t & .:‘,‘3 % \c- E fll‘;-r'-‘n.
weest CpAsLec 0wt sk ter  F LozL. W
- . o i - T : R s
S A AR ST g
.b"" AT 1Y ‘f » r‘-‘.‘& ..‘-‘ t.'fl:‘:~ [ ] % ™ s
Y o 23 . i’ ‘.\- ."-’ Q“. 13 ‘-i_ (
.t » - ' ‘ g. 2o e ,.o!.‘ n 8
o AN L O N o N
A ! i & b Pye “ s
0 oo d S _.
’ .;\‘ * ‘.’% ~ LRy i
v .‘h \.\‘ » A .
-
¢ -

This material had to be chemically compatible with
molten fluoride salts and capable of being fabricated
into complicated shapes such as characterize the
containment vessel, piping, and the heat exchangers.

Such a material was developed; it was called
INOR-8 at the Laboratory, but commercially it is
known as Hastelloy N and Allvac N. This alloy
has performed admirably in the MSRE, but the
prospects of molten salt power reactors with plant
lives of 30 years means that we must develop a
modification of this alloy that is more resistant to
embrittlement by long-term neutron irradiation.
Let us look briefly at the development of INOR-8
and then analyze the progress toward developing
an improved modification of this alloy for future
molten salt power reactors.

During the 1950’s when the United States was
attempting to develop a nuclear-powered aircraft,
one of the concepts considered was a reactor fueled
with UF, contained in a mixture of other fluoride
salts. A relatively complicated reactor called the
Aircraft Reactor Experiment (ARE) was constructed
and operated to demonstrate this concept. The fuel, a
mixture of NaF, ZrF,, and UF,, was pumped through

36

Figure 1. Photomicrograph of a sample of INCONEL 600
(20%Ni-15%Cr-5%Fe) after exposure to fluoride salt

in a pumped loop for 15,000 hours at 1300°F.

Voids near the surface are formed

as chromium is removed

selectively by the salt.

small Inconel 600 tubes at temperatures up to
1620°F. Inconel 600 is a nickel alloy containing
chromium and iron. The tubes were surrounded by
blocks of beryllium oxide that served as moderator
and reflector. The reactor achieved criticality in
November 1954 and operated 221 hours before
being dismantled for examination.

The ensuing period between the successful
operation of the ARE in 1954 and criticality of the
MSRE in June 1965 saw a change from thinking
of molten salt reactors in terms of military applica-
tions to the concept of their use as civilian power
plants. The work on molten salt chemistry and
materials development that took place during this
period resulted in a vastly simpler core design than
that of ARE.

Much of the complication of the ARE core came
from the fact that the Inconel 600 tubes were neces-
sary to separate the molten salt from the beryllium
oxide moderator-reflector, which was not compatible
with the salt. Graphite is also a good moderator,
and the discovery that graphite and salt are com-
patible allowed the core to become a block of graphite
with small holes for fuel passages. Assuming that
the graphite is available, the big materials ques-
tion becomes, What material can be used for the
containment vessel and the other metallic parts of
the system?

And this is where INOR-8 was used.

Requirements of a Structural
Material

One of the questions that we usually ask in choos-
ing a material is whether it will corrode. This ques-
tion applies equally well to many objects that we
touch each day as well as to nuclear reactors. Rusted
tin cans, pitted automobile trim, and tarnished
silverware all represent various types of corrosion.
The severity of corrosion can vary from the very
thin oxide film that gives the penny its tarnished
appearance to the complete dissolution that we
observe when a penny is dropped in nitric acid. The
former type of corrosion is ignored; the latter type
is intolerable.

QOaK RipDGE NATIONAL LABORATORY Review

Fluorine is very reactive with metals, so many
jump to the conclusion that the fluoride salts are
equally reactive. The free energy of formation is
commonly used to measure the stability of a com-
pound; a compound is stable if the free energy is
negative and the stability is greater the more neg-
ative the free energy. Table 1 shows the relative
stabilities of several fluorides of interest, including
the potential fuel salt constituents as well as the
structural metals under consideration.

Since the fuel compounds are more stable than the
structural materials, the salts and the metals should
be satisfied with their respective roles and the cor-
rosion rate should be low. However, the relative
ranking of the metals indicates that nickel and
molybdenum are least likely to form fiuorides and
that chromium is most likely. Thus, a reasonable
container material for these salts would be a nickel
or molybdenum base alloy with minimum quantities
of chromium and iron. The question arises, Why
not use pure molybdenum or nickel? Pure molyb-
denum is difficult to fabricate and is quickly re-
legated to the role of alloying addition to nickel
to give strength. Pure nickel is quite weak at high
temperatures and acutely sensitive to impurities
such as sulphur and phosphorus and would not be
suitable for use in an engineering system. Both

FaLL 1969

nickel and molybdenum oxidize rapidly at high
temperatures and would have to be protected from
exposure to air; alloying with chromium is normally
used to alleviate this problem in nickel base alloys.
We have already mentioned the Inconel 600 that
was used in the ARE. However, this alloy corroded
to depths of 10 mils in 1,000 hours when exposed
to circulating salt at a maximum temperature of
1500°F and would not be suitable for a long-term
application. The nature of the attack is shown in
Figure 1.

The list of available commercial alloys was
searched further. Two, Hastelloys B and W, satis-
fied the corrosion requirements, but “aged,” i.e.
became very brittle, after being heated for long
periods of time.

Development of INOR-8

This completed the screening tests on the avail-
able commercial alloys and all attention was turned
toward developing a new alloy in late 1956. The
application was still the ANP program where the
service temperature was 1500°F. The requirements
for this alloy were 1) good corrosion resistance to
fluoride salts, 2) good oxidation resistance, 3) moder-
ate strength and ductility at high temperatures,

37

Figure 2. Variation of the rate of oxidation of an 80%Ni-20%Mo alloy
with chromium content. Note the very sharp decrease in rate at chromium

levels of 5 to 7%.

4) stable properties at the service temperature, 5)
ability to be melted and fabricated into complex
shapes, and 6) ability to be joined by welding and
brazing. As you may have already gathered, such
a program involved several experimental disciplines.
We have already mentioned corrosion, mechanical
properties, fabrication, joining and physical metal-
lurgy. Fabrication in this case involved shaping 5-
ton ingots into useful items like plate, pipes, rods
and sheets, using such techniques as forging, rolling,
swaging, extrusion, and drawing. Physical metal-
lurgy is a term that covers about everything that
is not included in the other terms, but its principal
aspect here was a study of the stability of the vari-
ous alloys by aging at temperatures of 1200°F to
1500°F.

38

This program was directed by W. D. Manly of
Union Carbide; ORNL’s H. Inouye served as the
program’s technical conscience in addition to being
involved in much of the experimental work.

The facts accumulated at this point indicated that
the alloy base should be nickel, with molybdenum
(less than 20%) for strengthening, and some chro-
mium for oxidation resistance (less than the 15%
in Inconel 600). The melting practice used at that
time introduced small amounts of manganese, sili-
con, boron, oxygen, nitrogen, and some iron since
some of the chromium was charged as ferrochrome.
The metals used for melting stock also contain
small amounts of sulphur, phosphorus, carbon, and
myriad other impurities since scrap from previous
melts is quite often used. And so a commercial melt

OAxK RIDGE NATIONAL LARORATORY Review

hecomes a mmulticomponent system, and the so-called
scale-up from small laboratory to large commeraial
meits is an important and unforesecable experience.

The largest portion of the development program
spanned the years of 1958 to 1958, Small melts up
to 100 ib could be made at ORNL and larger ones
of §,000 to 10,000 Ib were made by Battelle Memorial
Institute, International Nickel Company, Westing-
house, and Haynes Stellite. The small ingots were
used as 8 screening step and larger melts were
made of only the compositions that appeared at-
tractive. (The screening step was taken because
these allova cost about $6 per Ib.) The large batches
of metal, or “heats,” were fabricated by the vendors
with OENL personnel on site in most cases. The
fabrication data derived from this were extremely
valuable to both parties involved, since this was an
anknown alloy, of which a 10,000-11 masa had to be
gotten to a useful shape. Fabrication was in the
form of tubing for corrosion studies, sheet for me-
chanical property measurements, and thick plate
for joining studies. The New England Testing Lab-
oratory near Boston assisted in the mechanical
property studies, and welding assistance was avail-
able from Rensselaer Polytech.

Hundreds of small melts were made by metallur-
gists at ORNL and large melis were made by the
vendors of the alloys shown in Table 2. All of these
alloys fabricated well except INOR-4. This alloy
contained 1.58% titanium and 2% aluminum, These
elements in this concentration form a very brittle
intermetallic compound with nickel called “gamma
prime.” It forme the basic hardener in a series of
nickel base alloys developed in recent vears. Such
alloys require special fabrication procedures not
practicable for our purposes.

My casual mention that most of the alloys fabri-
cated well does not pay adequate tribute to the team
Unouye, T. ¥. Roche, 1. E. Rosson and . Golston}
who worked many hours at ORNL and in the fab-
rication shops of the vendors. At this time most of
the tubing available was made by forming fiat
strips into a cylindrical shape and welding the eon-
tacting pieces together. This was called “welded
tubing” and had a terrible reputation for being
badly flawed. What we really wanted was a seam-
lgss tubing. The normal procedure for getiing this
is to deform an ingot some by forging Chammering),
extrude this forged pisce to obtain 8 tube shell (a
crude pipe with a thick wall), and draw, i.e., stretch,
this tube to obtain high guality, thinwalled tubing.
The extrusion step was the difficult part. Extruston

Fary 18648

involves simply pushing the metal through a die
te obtain a particular shape. This is what you do to
footh paste, but metals take g good deal more push,
To minimize the energy reguired, the metal is usu-
ally heated nearly to melting point, transferred
guickly to the extrusion press, and shoved through
the die in a matter of a few seconds. This procedurs,
nowever, consistently gave serap instead of a tube
shell. One of the team, in a2 momend of inspivation,
reasoned that the energy of extrusion alone could
be enough to heat it up and might even exceed the
melting point. The solution? Slow down the rate of
extrusion. We ended up with good tube shells that
were later drawn into miles of tubing by Superior
Tubing Company.

The corresion experiments on these alloys showed
that the corrosion susceptibility increased in this
order: iron, niobium, uranium, chromium, tungsten,
and aluminum. It is very close to that predicted
by the relative stahilities of the fluorides, i.e., alumg-
nunt would like to be a Suoride mwore than any of
the slements listed and will be attached prefer-
entially by a fluoride salt. Tungsten is the only
element that fell appreciably out of place. The cor-
rosion rates of allovs containing only nickel, molyh-
denurn, iron and chromium were found to be guite
acceptable. Jack Devan and Bob Evans showed by
two very good fracer experiments that the corrosion
rate of such alloys in a flueride salt containing UF,
was controlled by the diffusion rate of chromium
in the metal. The corrosion reaction involved chro-
mium in the alloy reacting with UF, in the salt to
produce UF, and CrF,, both of which are soluble in
the salt. This reaction can be written:

Cridissolved in + UF; == CrFydissolved + UF,

the metal) in salt)

Compesition, % by weight {(Base: Nij

Alloy e Lr Fe Ti 0 A&l M W
MOR-1 | 20
INOR-2 | 16 5
NOR-3 | 16 |1 T
OR-4 | 18 B5 2
INGR-5 0 15 P .
HORE 1 16 5 BE 3
INOR-7 | 16 & E i
HMOR-8 | 18 & 5
INOR-$ | 17 5 3

Table 2. Several promising nicke! base slloys
melted in the wwurse of developing INOR-8,

38

The reaction is temperature-dependent and the net
effect is that chromium is removed from the hotter
parts of the system and deposited in the cooler parts.
Thus, the chromium level should be as low as pos-
sible to obtain the lowest corrosion rate in fluoride
salts.

However, Inouye’s studies of oxidation in air
revealed a need for chromium. As Figure 2 shows,
the oxidation rate decreased markedly as the chro-
mium level was increased. In fact, as much as 6%
chromium was required to obtain an oxide that was
adherent and kept the metal from oxidizing too rap-
idly. And the need for 6% chromium was corrobo-
rated by the observation that of the alloys shown
in Table 2 only INOR-7 and INOR-8 had acceptable
oxidation resistance at 1500°F.

All of the alloys had adequate strength, because
of the 15 to 20% molybdenum present. Titanium and
aluminum increased the strength even further,
but the gains in strength were not worth the fabri-
cation problems mentioned earlier. The precise
levels of chromium and iron did not affect the
strength appreciably.

All of these alloys could be welded by the tung-
sten electrode with inert cover gas (TIG) process
as long as the boron, sulphur and phosphorus levels
were each below 0.01%. These are impurity ele-

40

ments and they can be kept at this level with rea-
sonable care.

All of these factors led ultimately to the alloy
composition specified in Table 3. The final composi-
tion is nearest to that of the experimental alloy
INOR-8, which is how the alloy has always been
known at ORNL. A patent application was filed
on March 3, 1958 and patent No. 2,921,850 was
issued on January 19, 1960 to H. Inouye, W. D.
Manly, and T. K. Roche.

Molten Salt Reactors for
Civilian Power

The ANP program was discontinued in 1960 and
the mission of the program was shifted to civilian
power reactors and became known as the Molten
Salt Reactor Program. A demonstration reactor
was needed and construction on the Molten Salt
Reactor Experiment (MSRE) began in 1961.

Very little additional development was done.
Long-term corrosion experiments were conducted
and the mechanical properties were measured on
several heats procured for the MSRE to insure that
the properties were equivalent to those measured
for the development heats of INOR-8. One problem

OAk RIDGE NATIONAL LABORATORY Review

worthy of mention developed in the form of cracks
that occurred during welding. The cause of this was
never determined, but as a consequence the heats
of INOR-8 for the MSRE were purchased from
Haynes Stellite (now Materials System Division of
Union Carbide) on the specified condition that their
weldability be first demonstrated by the vendor.
The actual number of heats rejected as a result of
this clause is not known, but we got the general im-
pression that the markedly improved quality re-
sulted from some revisions in the melting practice.

The fabrication of the metal components of the
MSRE proceeded without major incident and the
reactor went critical on June 1, 1965. The heat ex-
changer shown in Figure 3 gives an idea of the com-
plexity of the structure. In all, some 100,000 lbs
of INOR-8 went into fabricating the system. The
performance of the MSRE has been extremely satis-

FarvL 1969

Figure 3. Completed tube bundle for the MSRE
primary heat exchanger. The bundle consisis

of 163 half-inch OD tubes of INOR-8

Jjoined to a 1%” thick header.

fying to all concerned and the role of INOR-8 in this
success story is a vital one. We have removed
several pieces of INOR-8 for examination and have
not been able to detect any corrosion. In spite of
this, however, the salt in the MSRE continues to
pick up chromium from the metal, and the amount
after about 30,000 hours at 1200°F is equivalent
to removing the chromium from a surface to a depth
of about 0.7 mil (a human hair is one to three mils
thick).

Need for Improvement

Since the MSRE was fabricated, we have found
that INOR-8, like most iron and nickel base alloys,
is embrittled by neutron irradiation. Metallurgists
and solid state physicists have known for many

41

DEFORMATION (PERCENT)

0 200 400 600 800 1000 1200
TIME (hr)

years that metals are hardened when they are ir-
radiated at relatively low temperatures. This hard-
ening is referred to as “displacement damage,” since
it is caused by atoms being displaced from their
normal positions by neutrons. These displaced atoms
interfere with the normal mechanisms of deforma-
tion and make the metals stronger. If the neutron
irradiation occurs at higher temperatures, thermal
motion enables the displaced atoms to return to
their normal positions and no hardening is observed.
This self-annealing temperature is about 1000°F
for alloys such as INOR-8 and the general conclusion
until about 1966 was that these materials could be
used in a neutron environment at higher tempera-
tures without significant change in properties.

However, more recent work has shown that this
assumption was wrong and that the ductility of al-
loys such as INOR-8 at 1200°F is reduced drastically
by neutron irradiation. This embrittlement has been
attributed to the helium that is produced by the
transmutation of boron-10. A thermal neutron re-
action, it can be written:

B +n — *He + "Li

Cyclotron injection experiments have shown that
the damage is due to the helium rather than the
lithium. Boron is present in these alloys as an im-
purity that comes primarily from the refractories
used during melting ('°B constitutes only 18.2%
of natural boron). The cross section for this trans-
mutation is quite high and all of the *B will be
transmuted to helium in a molten salt reactor after
a neutron fluence of about 1 X 10?' neutrons/cm?.

The exact mechanism by which helium embrittles
these alloys is not understood, but ease of crack

42

Figure 4. Results of creep tests on irradiated (curve B)
and unirradiated (curve A) samples of INOR-8. The samples
were stressed to 30,000 psi at 1200°F.

propagation seems to be the most dramatic effect.
This is best measured in the response of a metal
sample to a sustained load in the irradiated and
unirradiated conditions. Such tests are normally
referred to as “creep tests,” and simulate the type
of loading experienced by a reactor vessel under
steady-state operation. The load (force) or stress
(force per unit area) is applied and the resulting
deformation is observed. This deformation or creep
will cause small cracks to form in the metal that
eventually link together until the sample fractures.
A well-behaved engineering material such as
INOR-8 will respond as shown in curve A in Fig-
ure 4. When this same material is stressed while
being irradiated, curve B is obtained. A lot of metal-
lurgy is tied up in the shape of these curves, but the
important fact for our analysis is that the curves
are identical for the irradiated and unirradiated
samples for the first 100 hours, or up to where the
irradiated sample fractured. Thus, the processes
that cause the stressed metal to deform (creep) are
not influenced by irradiation, but the irradiated
sample fractures in a shorter time owing to the
easier propagation of cracks. This is illustrated
further by the photomicrographs shown in Figure 5.

Tests were run to determine quantitatively the
effects of this embrittlement on the materials used
in the MSRE structure. These tests showed that the
fracture ductility was reduced as the quantity of
helium in the metal was increased by going to higher
thermal neutron fluences. Another important ob-
servation was that the fracture ductility depended
heavily upon how fast the metal was deformed. The
lowest fracture ductilities occur at a deformation
rate of about 0.1% per hour. Such high deformation
rates would not be encountered normally in a re-
actor system but may occur during transients in-
volving rapid change in temperature and pressure.
The ductility then improves slightly as the deforma-
tion rate is reduced further to a more realistic rate
of 0.001% per hour. (The MSRE was designed to
deform at a rate no higher than 0.00001% per hour.)

The properties of INOR-8 are adequate for the
MSRE where the anticipated thermal neutron flu-
ence is less than 5 X 10" neutrons/cm?. However,
the design lifetime of a power reactor such as the
Molten Salt Breeder (MSBR) is 30 years and the
fluence on the vessel will be at least 1 X 102! neu-

QAx RIDGE NATIONAL LABORATORY Review

trons/cm?. The need for an alloy with better resist-
ance to embrittlement, therefore, is evident.

Reduce the Boron Content?

You might think the obvious solution to a problem
that results from the presence of boron would be to
get rid of the boron. And although at low deformation
rates the fracture ductility is reduced from 30 to
3% by the presence of only 1 ppm helium, we pur-
sued this avenue for a short time. The main source
of boron is from the refractories (ceramic vessels)
used in melting the alloys. The regular melting
practice used in making the metal for the MSRE
produced material with 20 to 50 ppm boron. By
employing vacuum melting practice and using
very pure components the content was reduced to 1
to 2 ppm, probably the lowest attainable by com-
mercial vendors. We have even made small labora-
tory melts containing .02 ppm boron, but we found
that even these alloys were very brittle after ir-
radiation. Thus we have been forced to conclude

that the levels of helium required to embrittle this
alloy are so low that reducing the boron level will
not offer much improvement.

Modification of the Chemical
Composition

A closer look at the mechanism of embrittlement
by helium reveals several other important charac-
teristics. Metals are made up of grains, or small
crystals. At low temperatures they usually frac-
ture across the grains (intragranular) and at ele-
vated temperatures they fracture along the grain
boundaries (intergranular). Helium is embrittling
to these alloys only in the temperature range where
the fracture is intergranular. Since helium is nearly
insoluble in metals and moves very slowly, it fol-
lows that only the helium that is generated near
the grain boundaries can have an effect on the frac-
ture process. Since the range of the transmuted
helium atom is only about two microns in iron and
nickel, the '’B atoms must be located very near the
grain boundaries. With the boron present in parts-

Figure 5. Photomicrographs of the samples plotted in Figure 4. The irra-
diated sample on the right shows relatively few cracks that occurred before
they linked together to cause fracture, in contrast with the unirradiated
sample ( left ) which deformed 13% before fracturing but showed
numerous cracks

ot ey,

vl
- d
B e 1 9 K .2;
: . =k | §-3
. a 5 - - “' & L i “
A B S R R A
i 3 Y LY i = & g
B a I / : \_,.é 5 b 5L =
R SR e WA o o ol
: A AEE i ; 4 TR Ll ]
jia #1)
4 3 " -

FaLL 1969

AL
Al 2

'_1

43

per-million concentrations the shattering effect
it has is surprising. However, the boron atom is
smaller than that of iron and nickel and is not ac-
commodated easily in the crystal structure. Thus
boron gravitates toward the grain boundaries where
the arrangement of atoms is interrupted by the in-
tersection of grains of different orientations. Hence,
the laws of nature provide a mechanism for concen-
trating boron in a region where it is most detri-
mental.

One solution to this problem that seems reason-
able is to add an element that will react with the
boron to form a stable compound. Ideally, by suit-
able annealing, the compound could be deposited
as fine precipitates located within the grains rather
than at the grain boundaries. However, the location
of the precipitates along the grain boundaries may
be acceptable if the transmuted helium remains
associated with the precipitate and does not in-
fluence the fracture process. Very fine precipitates
along the boundaries can even retard the propaga-
tion of cracks and might prove beneficial from this
standpoint. Hence, the key item in our thinking
was to add small amounts of elements such as tita-
nium, zirconium and hafnium that formed com-
pounds with boron.

One thing to understand here is the extreme im-
portance of the electron optic equipment that has
only become available in recent years. An ingenious
device, the microprobe analyzer, is capable of de-
termining the chemical composition of an area about
two microns in diameter, permitting analysis of
both the precipitates and the surrounding areas
(matrix). These precipitates can be extracted and
collected by a process which involves dissolving
away the matrix material. They can then be identi-
fied by x-ray diffraction, analyzed by normal wet
chemistry techniques, or placed in the electron mi-
croscope for electron diffraction and for qualitative
analysis by the microprobe attachment for this
instrument. The metal can also be corroded with
acids until it can be seen through by the electron
microscope to determine its structure and the role
that the precipitates play in deformation and frac-
ture. Moreover, these same techniques can be ap-
plied to irradiated samples since very small amounts
of material are involved. This work has been per-
formed by R. E. Gehlback, J. O. Stiegler, R. S.
Crouse, and H. V. Mateer and has been the key
item in the program to modify INOR-8.

We also examined the other elements in INOR-8
for the purpose of simplifying the alloy as much as

possible. The microstructure shown in Figure 6 is
characteristic of the standard Hastelloy N. The
large precipitates are carbides of the MgC (six metal
atoms to one of carbon) type where M is 50% Mo,
40% Ni, 5% Cr, 1% Fe, and up to 4% Si. These par-
ticles are located in “stringers” in the metal, causing
the grain size to be much smaller near the stringers.
The particles themselves are very brittle, and frac-
ture as the material is deformed. When the material
is irradiated and is also relatively brittle these
cracks can propagate to cause failure. The particles
are too large to improve the strength, so we sought
to eliminate these from the microstructure.

Our analysis of the matrix (excluding the preci-
pitates) revealed that only about 12% of the Mo was
actually dissolved in the alloy. The bulk of it went
to form the massive MyC carbides. We made several
small melts with varying Mo content and found
that alloys with 10% and 12% Mo were free of the
massive precipitates and those with 14% and greater

Oak RipGE NATIONAL LABORATORY Review

Mo concentrations contained the precipitates. The
strength penalty for decreasing the Mo content from
16% to 12% was modest enough and we decided to
make this change in the alloy composition.

As mentioned previously the amount of chromium
in the alloy was a compromise between resistance
to oxidation and to salt corrosion. Hence, we did
not alter the concentration of this element. The ad-
dition of chromium as ferrochrome also adds about
4% iron. We did not think the iron was harmful,
but its elimination gave us one less variable.

Because of the very reactive elements (titanium,
zirconium and hafnium) that we were adding, we
chose to use the vacuum melting practice, in which
the molten alloy is not exposed to air. This melting
technique has become very common over the past
few years and is preferred by most vendors for melt-
ing even standard INOR-8. The vacuum melting
practice also reduces the amount of so-called “re-
sidual” elements introduced such as manganese,

FaLL 1969

I-—-

TR

Figure 6. Polished cross section of a sample of INOR-8
deformed at 1200°F. The large carbide particles
have fractured, and near the upper edge the

3
e fractures have resulted in a crack
E, w Propagating to the surface.
=19
~
8 (3
ol
5
.l§_
Yz

silicon, boron, nitrogen and oxygen. Nickel-base
alloys are notorious for being embrittled by small
amounts of sulphur, so we retained some manganese
n our specification to tie the sulphur up as a stable
Mn-S compound. The other residual elements were
kept as low as possible.

Recent work by C. E. Sessions has shown that the
strength of INOR-8 depends very strongly on the
carbon content. As the C is decreased below the
normal (0.04% to 0.08%), the strength decreases
dramatically; higher concentrations improve the
strength gradually. Inouye had observed earlier
that they could not fabricate tubing from material
containing more than 0.1% C. These factors in-
dicated that we should leave the carbon specification
at 0.04% to 0.08%.

Our resulting modified alloy then contained 12%
molybdenum, 7% chromium, 0.2% manganese, 0.06%
carbon, and small additions of titanium, zirconium
or hafnium,

45

Figure 7. Transmission electron
micrograph of a Ti-modified sample of
INOR-8 irradiated at 1175°F and
creep-tested at 1200°F.,

The dislocations resulted from

the creep test in which the sample
deformed 5% before it fractured. The most
important feature is the grain boundary
lined with small MC-type carbide
precipitates. These help to

disperse boron and

inhibit crack propagation

along the boundary.

Properties of the Modified Alloys

Several small melts were made to investigate the
effects of Ti, Zr and Hf on the mechanical properties
after irradiation. (The behavior shown in Figure 7
is characteristic of the Ti addition.) These samples
were irradiated at 1200°F and then tested in the
hot cell to determine the strength and ductility
after irradiation. Both parameters were improved
dramatically as the titanium level exceeded 0.3%
to 0.4%. Zirconium and hafnium brought about
similar improvements.

We sought to keep the titanium, zirconium and
hafnium additions as low as possible for two rea-
sons. First, these elements depress the melting
point of nickel-base alloys and can cause cracking
in the early stages of a melt or during welding. A
second reason is that these elements form brittle
compounds with nickel. The exact levels required
to achieve undesirable results were not known, so

46

we started with very low concentrations.

We obtained a few 100-1b ingots from commercial
vendors to have some plates for investigating the
weldability. We found that the zirconium reduced
weldability when only 0.05% was present. We did
not investigate these alloys further. Alloys con-
taining up to 1% Ti and 1% Hf had good weldability.

We could not afford to carry parallel development
programs on alloys containing titanium and haf-
nium additions. Titanium is present in several iron
and nickel base alloys, so we felt more comfortable
with this addition and proceeded further with its
development. Our development schedule for the near
future included several 100-lb melts to study the
effects of variations in titanium and carbon levels
and one 5,000-1b melt. (The original development
program for INOR-8 involved about 30 melts of
this size, but today’s austerity does not allow this
approach.)

Our initial irradiation tests were carried out at

Oak RipGE NATIONAL LABORATORY Review

"
1200°F, operating temperature of the MSRE. The
" strength and fracture strain of the titanium-modified
alloys were excellent after irradiation, and welding
studies showed that fhey generally had good weld-
ability. The single 5,000-lb melt fabricated well
and about 50% was converted into useful products.
This large heat had one idiosyncrasy: it cracked
when welded with filler metal from the same heat.
However, sound welds were made with eight other
heats of filler wire and we were not concerned about
the basic weldability of the titanium-modified alloys.

All seemed well with our new alloy until, anti-
cipating that molten salt breeder reactors will
operate at 1300°F, we irradiated some samples at
higher temperatures. Much to our dismay, we found
that the samples irradiated at 1300°F and higher
reverted to about the same limitations as the stand-
ard alloy.

We examined thin sections of some of the irra-
diated samples in the electron microscope and found

FALL 1969

Figure 8. Another sample of the same
composition irradiated at 1300°F before
creep-testing at 1200°F. This one deformed
only 0.4% before it fractured, so fewer
dislocations occurred. Much of the

grain boundary lengths are free

of precipitates and contain occasional
helium bubbles. What few precipitates are
present are coarse and of the M,C type.

that the microstructure varied dramatically with
irradiation temperature. The microstructure shown
in Figure 7 was noted after irradiation at 1175°F
and creep testing at 1200°F. The most important
feature is the grain boundary lined with very small
particles. If you tried to shear this boundary, you
would probably find that it is very strong and that
these particles would make crack propagation dif-
ficult. These particles were found to be carbides of
the MC type where M consists of molybdenum and
the modifying elements added such as titanium,
niobium, zirconium, or hafnium. A sample of the
material irradiated at 1300°F and tested at 1200°F
has the microstructure shown in Figure 8. The
grain boundaries in this sample are free of preci-
pitates for large distances and the precipitates
present are relatively coarse. Occasional helium
bubbles are visible along the grain boundaries.
The precipitates in this sample are of the M,C type
with M being 90% molybdenum and 10% chromium.

47

Hot swaging
in ORNL’s
metals laboratory

Recall that the proposed mechanism of embrittle-
ment of this alloy during irradiation involves the
accumulation of helium (from transmuted boron-10)
at the grain boundaries. The fine precipitates
formed at 1175°F would disperse the °B and also
provide a boundary that resists crack propagation.
Both of these factors probably contribute to the
good properties of the alloys containing the fine
MC precipitate. The coarse M,C type would not be
effective in either regard.

Thus a critical part of our improved alloy seemed
to be the fine MC precipitate and we looked at
further chemical modifications that would make this
carbide stable at higher temperatures. From our
own work and from the literature, we compiled a
list of those carbides which were favored by various
alloying elements of interest. This list (Table 4)
only indicates the trend, and the alloys have to be
made to determine how much of each element is
required to form the desired MC carbide. For ex-

Elements Type of Carbide Favored
Ti, Zr, Hf, Nb MC
Cr M23C6
h. 3
Mo, W M.C (high Si present)

M.C (low Si present)

Table 4. Carbide formation in Hastelloy N.

48

ample, adding titanium favors the formation of the
MC type, but we found in an alloy containing 0.5%
Ti that MC was formed at 1200°F and that M,C was
formed after long periods of heating at 1300°F.
This type of information could only have been
gathered by making the experimental alloy.

We have irradiated several alloys containing var-
ious concentrations of the elements listed in Table
4. To date alloys having chemical compositions that
cause the formation of the MC type precipitate
have very good mechanical properties after irradia-
tion. The properties of several of the more attractive
alloys and those of standard Hastelloy N are com-
pared in Figure 9. These alloys were all irradiated
at 1400°F and retained their good properties. The
properties of the alloy containing 1% Ti with 1% Hf
are very good, and those of the alloy containing 0.5%
Ti with 2% Nb are acceptable.

Further work will be required to decide the exact
composition of the alloy to be used in future reac-
tors. The information obtained from our small melts
will be used to write specifications for several large
5,000-1b melts. These alloys will be tested to deter-
mine their strength and ductility after irradiation.
Large sections will also be welded to insure that the
alloys can be joined. These studies will lead to a final
chemical specification that will be used for procur-
ing materials for future molten salt reactors.

Oak RinGe NATIONAL LABORATORY Review

. UNIRRADIATED

. IRRADIATED

N

L+ JH L 'Q3131I00W

IH 1 a313100W

__ ACTV QHVANVLS

Qo o (@) (@] (] O O 0 o b
o (@] (o] (®) 0 N -
s} » ] =
(INIOH3d) NiVHLS IHNLOVHS
¥
11} + JH'Q313a0W
-E---I..I---I.ll
SRS B 55 T 1 ] e
1 R g 6 v
f & £ 0 BAaT & i
s i z -

(44} 34NLOVYHY OL 3NIL

Figure 9. A comparison of the creep properties of standard INOR-8 and
several modified alloys in the unirradiated and irradiated conditions.

Test condition was a stress of 35,000 psi at 1200°F.