' sy = A
Caleaa. N &
A-670 File L
Those Eligible
To Read the
Date Li=26-11, . Attached

Subject__ Notes cn Meeting of April 26, 194k

Copy #._5 Weinberg

= (Worary)

i W42

RBefore reading this decument, sizn and date below g&
S Ny
Name Date : 298 [,;1/“2&11' Date
A = UG
Y s W
arlmafs & :
7o TR
/J 'l—’ T // 7 .

wNTRAL RESEARCH LIBRARY"
OCUMENT COLLECTI
LIBRA 2 COPY

//’ B
DO NOT TRA}SFER TO0 ANOTHEQERSON

-l 3 S

If you wish someone else to see this document
send’in name with document and the library will
_arrange a loan.
CENTRAL RESEARCH LIBRAF.‘Y
DOCUMENT COLLECTION

8Nt connints -

Present: Fermi, Allison, Szilard, ‘,"‘T':gher, Yieinberg, Seitz, Morrisen,
Cooper, Vernon, Tolman, Wafson, Chlingsr

N (ee- V- t-29)

Fermi. His remarks

The first speaker in today's meeting was i

It was assumed for today's discussion that the aim of the chain
reaction would be the production of power.

The first type of pile assumed for this purpose was a permaneant
lzrge pile of about the Haaford size (but not the Hanford type necessarily)
for production of energy in the neighborhood of 10° kilowatts. The arrang
ment suggested was one in which one large mother plant would produce LG Zor
consumption in a series of smaller plants. In the mother plant, the ensrgzy
produced id be used to reduce the cost of the 49 produced. (Hr. Fermi
mentioned t he viewed the use of this power for the heating of cities
! ). There may be non~technical cbjections to this arrangement,
the shipment of 49 to the smaller consuming plants offers the

e s a
for example, I
serious hazard of its falling into the wrong hands, but these were to be

ted from this discussion.

rdamental aim in the mother plant would be toc gzt the mis—

eld, with full utilization of the metal as the zoal., If a
uchh a propesal can be found, then the schemes for isotope sep-

aration are not of great interest. If such a solution is not possible, ti

> schemes for isotape separation should undoubtedly be investigated further

ooon @

In the following discussion of full metal utilization, the
d 49 will be referred to as 8 and 9, respsciively. In the react
se thet cne fission of 9 and W fissicns of 8 take place in a
rengration. Then the production of neutrons will be U9 + ('é!
rons ave lost in the mederator, coclant, sic. I k
the numbsr used in producing 40-10. L

» absorption by 8 to produce 9 will be
iy e

(L -1)(Vg 3 Py

and the production of ¢ per cycle will be

The term 1 ¢ X represents the destruction of 9. Therefore, the ratio of pro-
duction of § to its destruetion, which we will call ", will be

ot

To ubilize a1l the metal, ¥ obviously must be greater thamn 1. If ~ is oaly
vary little vreauer then 1, the chain reaction would keep going with maxdirmm
economy of filssionable materials and would continue to g0 on until all the
metal were ussd, bub the value of such a pile would not be great and it would
only be good for, say, hardening materials (the VWigner effect) or possibly
{though less desirable) heating cities. The effective Vg is around 2.1 to 2.2.

Assume Pirst 2 Hanford type pile with an equivalent amount of 4%
substituted for the 25, i.e., in the early stages, 25 would be burmed to pro-
riuce 49 which would gradually improve its condition. The cez'lier estimate

of 1.9 for the ratio of the fission cross ssction of 43 to that of 25 has been
more recently estimated by ¥ as l.4. The ratic of absorption cross ssction
for 49 to that of 25 is around 1.5. TWith these conditions, L/ is about 10%
nigher than it was previously thought to be. (”’he actual valués of UV and
U effective are not really known so the discussion can only show ranges).
The situation then in a pile of Hanford design and latiice would be for a

/ effsctive (which will be referred to herzafter as, 4) of 2 to 2.2, vy will
from O 8 to 0.98. In the latter case, the pile is close to a balanced
sitvation but not quite there. To adjust such a pile without drastic changes
of design, large diameter slugs or more metal could be used te improve the
thermal wtilization and increase /(/. However, over-sized lumps increase the
difficulty of cooling since the annular type cooling is badly limited in
power producticn by the metal temperature. .

<

poCr
. W

The secordd type pille considered for the production of power was
he P-9 moderated pile. For a/Mof 2 to 2. 2y % woald be 0.93 to 1.,13.
hese values do not necessarily represent the optimum but are merely in-
dicative of whait can be done with P-9 piles and one with such a2 ¥ of 10 to
15% may or may not e an operable plant. The practical difference between
continuous and discontinuous P-9 plants is not large in this respect since
the loss by absorption for the coolant and its tubes practically compensates
for t’na less efficient reproduction in slurry piles. One might hope 1o im=-

2 situation by capturing the escaping neutrons in & refizctor bub the
ion in the pile containsr is an important problem.

3 ¢b

Another type of pile to consider is one with very lititle or no
mederator (fast chain reacting type). From the nuclear point of view this is
very desirable and is simple in prmclnle but, practlca.l*", it involves serious
problsms in removing the heat. Ignoring the cool , and considering only the
nuclzar point of view, this type pile may be of either cne of two forms:

49 wiZ, throughout
mass

pure 28 or
tubealloy /

pure 28
reflector
13-

In Fig. A, a small spherical core of 49, say,10 cm in diameter, would be

surrounded by a sphere of 28 or normal tubealloy about 40 to 60 cm in dia-

meter. This arrangement is good from a v~ standpoint and one might expect

a Wof 1.3 to 1.4, because L can be made small since the fast neutrons from

the 49 get into the 28 readily. (Mr. Allison pointed out thet if 25 is not
considered for the surroundings here, thorium might be used). The pile shown

in Fig. A only requires a few kilograms of 49. To utilize more 49 it would be
possible to construct units like A with multipls 49 cores spherical or cylindrical -
in shape.

: Fig. B represents a homogeneous sphsre of 28 with 49 uniformly dis-
tribubed throughout the mass, the whole surrounded by a reflector of pure 28
vo catch the leakage neutrons. In this arrangement about 70% of the neutrons
get into 28 immediately to produce fast fissica. Assuming a mixturs of 49
end 28 in which X represents the percentage of 49, critical condiiions {i.e.,
where the chain reaction continues if the pile is of infinite size) would be
reacned with about 5% of 49 inm the mixture (X = 0.049). For values of /¥ of
2 %0 2.2, ¥ would be 1.37 to 1.57. As the pile size is decreased, the .
following results would be obtained. They are calculated without reflector.

IAPLR I
X
Critical ' 3 ' :
Radius of Sphere X {fraction of Q) . Az A= 22
100 cm 0.054, 1.23 1.43
70 cm ., 0.060 1.10 1.30
50 cm 0.05% .98 1.18

Adding a reflector would decrease the criticzl radiuz of the active sphere by
about 10 cm and improve very considerably the vaiue of Y since the reflsctor
would utilize the neutrons escaping from the active core. Taking the case of
the 70 cm sphere above, this represents about 1} &2 or say 30 toms of the mix=
ture. Therefors, 6% or about 2 tons of 49 would be required 4o keep this
machine rmmning. Thus a plant of this type rsquires a large quantity of 49
for operation although this is aot sufficient reason for discarding this type
of pils as a possibility. :

The serious objection to these fast chain piles is the removal of
the heat. Since practicaliy 211 ths heat is produced in ths 49 {about 70 to
80%), piles like those in Fig. A are harder to cool since it is mainly the tiny
cors which must be cooled while in Fig. B the whole mass iS5 to be cooled.

As anotier possibility, a compromise enriched pile might be designed
which would have enough moderator to reduce the percentage of enrichment re-
quirsd to keep the chain reaction going., But not as large an amount would be
required for-the conventional optimm conditions. . '

Hr. Ferml suggested that at a later meeting he would consider the
question cf how to use the 49.

Mr, Szilard was the second speaker and proposed approaching the
problem from a different viewpoint,-~that of assuming more optimistic valuss
of the comstants so as to indicate other potentialities. He pointed out that
the fast reaction is preferable to the slow chain reaction for producing 49
from tubsalloy and that this is probably more true if we assume more pess—
imistic velues for L/ or f. Eefore discussing these values of the constants,

hes of a possible design were distributed and deseribed briefly. These
are atvached hereto.

The sketches show two different arrangements. In sketch A, the
enriched tubealloy (enriched to where the chain reaction will go) and natural
tubzalloy would be distributed in the form of rods in a cylindrical pile, ia
which the snriched material would be in the center portion of the reds lying
within a circular area in the center of the pilse. Part of the rods, located
within three circular areas zround the center (as indicated in Fig., 1) vould
e arranged so the cylindrical bundles could sach be rotated about its axis.
In each of the rotating bundles, part of the rods would be natural tubealioy
znd the balance of natural tubealloy with the center section erriched.

In the beginning, the enriched material in the three bundles would
all face ths center of the pile and lie within a cylinder whose axis would
coincide with the axis of the pile and whose cylindrical surface would pass
through the three axes of the revolving bundles. By means of this arrange-

‘the muitipliication Factor increassd with the conbinued opsraition of
the envriched material could be rotated away from the canter of the
he natural tubsalloy brought towards the ceater where it in tum
enriched. In the center of the pile would be‘a single tube for
introducing mercury, liquid bismuth, or scme other absorbing or slowing msterial
for conbrolling the pile. The coolant for this type pile would be a bismuth-
lead alloy and would flow downward through the pile betwesn the static and
rotating rods. The possibility of using liquid scdium in place of bizmuth-
lead should also be looked inta. The volumetric heat can:
sodimm is zbout the same as that of tae bismuth-lead alloy but its densi
would e 10 times less, so that the pressure drop would be sbout 1/10 tF
he blsmuth-lead alloy or the velocity zbout 3 times larger for equal
by T Ia the scheme just dsscribed, the following approximate con-
1itions would obtain: {1) the bismuth-lead alloy would occupy aboub 1/3 of
hrough the pile at a velocity of about 15
h 2 om dizmeter rods raised to 700°C metal temper-
ature a2t the center of the central rod and with 150°C temperaturs increase
in the coclant, about 250,000 kw will be removed. The pumping powsr for the
coclant will consumz about 5% of the power produced. §

In the alternative scheme B, comirol of the pile would be obtained
by means of a nest of tubes for the mercury or other controlling medium ar-
ranged as in Figs. 2A and 3B and 4A and 4B. The metal rods would all be
sbationary and vertical (nvs. 12, 13 and 14 in Fig, 34) and would be about
1/2 to 1 cm in diameter by about 2 meters long.

In both designs, the
ber by about the same

core weuld be ordinary tubealloy of the same rod size. The total diameter and
o

the height of the pile would be about 2 meters.

The objective of such a pile nmust be to produce as much extra 49 as
invested. It is assumed bhat the production will be double the original in-
ve For every atom of L9 disintsgrated, twe atoms of 49 would be pro-
ducsd. Part of these will be produced in the enriched core and part in the
surrounding natural tubealloy. Some of the producticn in the core will tend
to leak out into the natural tubsalloy and this lsakage must be kept within
cervain limits. Then k will increase over z period of time. As the chain
reaction goes on, the multiplicatiocn factor k will then increase so that ths
controls must provide for this as well as the normal operating control of the
pile.

IN—the slow chain reaction, 49 capturss neut x-ors in radiative nct
fission capture to produce a new element which we will call super plubcaium
or 40~10. It is assumed there is a 509 chance that this new element will be
Dissionablz, If 1t is not fissionable, it is assumed there is 50% chance that
it will be formed only in neg,l_gibl quantity in the capture of fast neutrens.
Thus, there is a 75% chance in a fast chain reaction that we may use L~ and
znot A din g:rt_no the pz'oduvhlon balance (/¥ = 2.2 nsutrons pger neutron abe

2.2 x 1.175 = 2.6 neutrons produced per neutron absorbed). As
he neutrons incrsases from thermal to f‘uu‘icu energles, it is

d decrease in /. The main argument in favor of the 1

T T,h"t if a fission neutroa is released in tubealiocy, it causes
n o preduce 1.2 neutrons (fast effect). IF all the neubtrons.
plured, the ovcra..l balance would be that for every atom of L9 dssiroyed,
oms of 49 would be produced. Cne goes back into the chain reactiocn, the
replaces the 49 destroyed, providing e net gain in L9.

in which a Ra - B neutron scurce was surrownded ov 28,
5.3% increase in the nuaber of nsubtrons
above the fission threshold. This means that the in-
r of neutrons for an infinite sghere would be 5.3 or
fis:xion cross section is taken at C.35 and the 1 - 0.63
cross section at 2.7 for a Vog of 2.2 to 2.6 £ will vary from 1.18

rring to the value above of U25 of 2.6, 1i' we were to use the
mere optimistic "e.-x._]_'bs reported by Y {that ¢ is 20% larger than Usg) then

Y/, o equals 3.1 nesutrons produced per neutron absorbed. If we are less optim-
e aad assue U9 effective=2.5 bubt use the 193% increase indicateq by the
iment meationed above, we have three neutrons produced in a mixture of 28
and 49 for cne atom of 49 destroyed.

It has been suggested that one of the subjects for one of the meet-
ings soon to be held would be a review of the availability of the metal pro-
ducing ores and other sources of tubealloy. This is to be given by Mr. P. Morrisan.

X
N
-\
1
1

g /;@,O ooeooo

Ziie deawing, teccther with all in‘flm‘d.fl?&-

uF

ard hE . s the'pres” SCHEME A
e the Unipesiiy el €l ;a,.apd_.qé‘*

use diselosura, of raproduetion of any part |

thereci may be made except’ by “written's-
authorization of the University of Chicago,.’
. \\\\\\\\\\\\\\\\\\\\\\\\\S T \\\A«\\\\
i //////V//////////////////////////A//M NG
_— =
IN = S /,
s AN
V| = Ei Eifi/ .;/ T
o AR ey
= NI
% == =N
N\ = = w.w/ N

T
wx\m\v LI

y

/.fl/_fl// NS

.

\\\\ \\\\\\\\m

Cerrrrrsa

N\

i iafisrmation

part

1 2f any

7

= Ba

=)

H” Am.u

o)

- 3 9

== 290

s O

i 25

i g

SRR
4 ¢ ) L

it a4
g {WEHOS

I,

et e Eamey

| o P !

B e

LS . N
ERRERE - N
_ __ 7

. biee s
o e
>
A

P / 7 7
LA

Z 2L
DANNNISE

|

SN rhIR M
DL,

Uy
S,
~ e
\\>>
\
3

Fire 36
Fection A-A

ritten

1e except bYW

mace

i the University of Chic:gp,"

iy

it

b
e

crvon A-4°