ORNL-1463.txt

   
 
   
     
      

CENTRAL RESEARCH LIBRARY
DOCUMENT COLLECTION

= - OBNL 1463

Metallurgy m_t.l__,veralics
b e

R

METHODS OF FABRICATION OF CONTROL AND
SAFETY ELEMENT COMPONENTS FOR THE
AIRCRAFT AND HOMOGENEOUS
REACTOR EXPERIMENTS

OAK RIDGE NATIONAL LABORATORY

i] OFERATED BY

" CARBIDE AND CARBON CHEMICALS COMPANY
] / A DIVISION OF UNION CARBIDE AND CARBON CORPORATION

| e

POST OFFICE BOX P
OAK RIDGE. TENNESSEE

 

 
?
¥,

ORNL-1463

This document consists of 18§ pages.

 

Copy_5 of 159 copies. Series A,

Contract No. W-7405, eng-26

METALLURGY DIVISION

METHODS OF FABRICATION OF CONTROL AND SAFETY ELEMENT COMPONENTS

FOR THE AIRCRAFT AND HOMOGENEOUS REACTOR EXPERIMENTS
J. H. Coobs and E. S. Bomar

Powder Metallurgy by J. H. Coobs and E. S. Bomar
Welding and Brazing by G. M. Slaughter and P. Patriarca

DE b lASS IFIED

B ----AE[E T 557

———i s Mmook ma & Sy

 
 

FEB £ 6 1953

 

OAK RIDGE NATIONAL LABORATORY
Operated by
CARBIDE AND CARBON CHEMICALS COMPANY
A Division of Union Carbide and Carbon Corporation
Post Office Box P
O0ak Ridge, Tennessee

MARTIN MARIETTA ENERGY SYSTEMS LIBRARIES

R

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3 445k 03531485 4 |

 

 
ORNL-1463
Metallurgy and Ceramics

 

  
   

  
 
 
 
 
 
   
  
 
  
   
  
  

   
  
    
   
 

DISTRIBUTION
1. C. E. Center 27. A. Hollaender
2, Biology Library 28, M. T.
3. Health Physics Li 29, K. Z.
4, Metallurgy Libra 30. J. S.
9-6. Central Researc brary 31. A. S,
7. Reactor Experi al 32, C. S,
Engineering rary 33. C. E.
8-13. Central Fil 34. D. S,
14, C. E. Lars 35. D. W,
15. W. B. (K-25) 36. E. M,
16. L. B. t (Y-12) 37. A. J
17. A, M, nberg 38. D owen
18. E. H4aWaylor 39. P Reyling
19. E hipley 40. . Williams
20, F VonderLage 41. Coobs
21. . Briant 42 . S. Bomar
22, J. Fretague . M. Slaughter
23 A. Swartout . Patriarca
2 . C. Lind . . Levy (on loan from
A . L. Steahly Pratt and Whitney
. H. Snell Aircraft)

EXTERNAL DISTRIB

 
 
   
    
   
   

46-53. Aréonne National Labo

54-58. Atomic Energy Commi , Washington
59. Battelle Memoria titute
60-62. Brookhaven Natj Laboratory

63. Brush Beryllj
64. Bureau of

ompany

   

65-66. Californg esearch and Development Company
67-68. d Carbon Chemicals Company (K-25 Plant)
69-72. ) nd Carbon Chemicals Company (Y-12 Plant)
13. i o Patent Group
4. ow Chemical Company (Rocky Flats)

75-77. duPont Company

 
 

General Electric Company (ANPP) (1 copy ea. to W. Baxter,
H. C. Brassfield, J. A. McGurty, and W. J. Koshuba)
General Electric Company, Richland

Hanford Operations Office

aho Operations Office

a State College

s Atomic Power Laboratory

amos Scientific Laboratory

rodt Chemical Works

etts Institute of Technology (Kaufmann)

atory

isory Committee for Aeronautics, Cleveland
ory Committee for Aeronautics, Washington
of Standards

 
 
 
  
  
 
 
 
   
 
   
  
 
 
   
  
  
  
   
 
   
 

90-9
94.
95-98.
99-.101.
102. Mal
103. Massa
104-106. Mound
107. Nationa
108. National
109. National B
110. National Lea
111. Naval Research
112. New Brunswick L
113-117. New York Operati
118-120. North American Avy
121. Patent Branch, Wash
122. Rand Corporation
123. San Francisco Operatio
124. Savannah River Operatio
125. Savannah River Operations
126. Sylvania Electric Products,
127-130. University of California Ra
131. Walter Kidde Nuclear Laborato
132-135. Westinghouse Electric Corporati
136-144. Wright Air Development Center (1
Pratt and Whitney Aircraft)
145-159. Technical Information Service, Oak Ri

fice, Augusta
ice, Wilmington

n Laboratory
Inc.

to R. Strough,

 

111
i © i

 

 
METHODS OF FABRICATION OF CONTROL AND SAFETY ELEMENT COMPONENTS
FOR THE AIRCRAFT AND HOMOGENEOUS REACTOR EXPERIMENTS

J. H. Coobs and E, S. Bomar

SUMMARY

This report contains, in two parts,
the experimental work leading to the
fabrication of control and safety
components for the Aircraft and
Homogeneous Reactor Experiments.

The method evolved for preparation
of cylindrical inserts for the ARE
control and safety rods required the
hot-pressing of mixtures of boron
carbide with alumina and with 1iron.

A technique for fabrication of
control plates for the HRE was devised
to minimize defects found in earlier
sets of plates., The method involved
closely packing pressed-powder com-
pacts 1n a frame and then hot-rolling
to bond protective cladding sheets to
the core.

INTRODUCTION

The neutron-absorbing material
selected for control and safety
elements in both the ARE and HRE was
boron in the carbide form.

Cylinders with acceptable physical
properties and the requisite amount
of boron were made for the ARE safety
rods by consolidating mixed iron and
boron carbide powders. Consolidation
was effected by pressing atan elevated
temperature. Inserts for the control
elements were prepared in a similar
manner by using a dilute suspension of
boron carbide in alumina as a carrier.

The operating conditions to which
the HRE control and safety plates are
exposed require that the high cross-
section material be completely enclosed
in a protective envelope. Initially,
efforts were directed toward the
cladding of Boral with stainless steel
as a protective covering. Later,
pressed and sintered compacts of boron
carbide suspended in copper were sub-
stituted for the Boral to permit

rolling at a higher temperature for
bonding of cladding to core.

ARE CONTROL AND SAFETY ROD COMPONENTS

The design selected for the control
and safety rods for the ARE called for
the use of boron carbide as the high
cross-section material. Inserts for
the control rod were to be made in the
form of cylinders and composed of a
dilute suspension of boron carbide in
alumina. These were later prepared
with a minimum of difficulty.

The boron carbide bearing components
for the safety rods were to have es-
sentially the same shape, but were to
be made of pure boron carbide. The
reply to an inquiry to a supplier of
the special grade of boron carbide
required for the pure boron carbide
pressed cylinders indicated that the
cost for these items would be quite
high. Therefore experimental work was
done which revealed that pressed
cylinders containing an acceptable
amount of boron carbide bonded with
iron could be produced.

Experimental Work on Safety Rod
Inserts. Mixtures containing 80 vol %
boron carbide (Norton, metallurgical
grade, which analyzed 71 * 1% boron)
and 20 vol % copper, nickel, and iron
were hot-pressed to check the useful-
ness of these materials as cementing
agents. The powders used were all
-325 mesh, and were milled together,
using steel balls, for several hours.

The results of these runs are
tabulated in Table 1, The results of
the first runs indicated that nickel
did not promote especially desirable
properties in the pressed compact. A
liquid phase formed, but i1t did not
wet the boron carbide particles and
gave little cementing action,

 
 

 

TABLE 1. PROPERTIES OF HOT-PRESSED NICKEL, COPPER, AND
IRON PLUS BORON CARBIDE COMPACTS
PRESSING DENSITY AS PERCENTAGE
Cg?P?ffl;ION TEMPERATURE OF THEORETICAL REMARKS
e (°C) DENSITY

20 Ni-80 B,C 1200 74.0 We ak, much nickel loss

20 Ni-80 B,C 1200 70.4 Weak, soft

20 Cu-80 B,C 1100 74.5 Fairly strong, loss
of copper

20 Cu-80 B,C 1090 78.4 Fairly strong

20 Fe-80 B,C 1050 69.0 Soft, weak

20 Fe-80 B,C 1170 73.7 Soft, weak

20 Fe-80 B,C 1290 71.0 Soft, weak

 

 

 

 

Copper gave somewhat better cementing
effect than the nickel even though it
did not dissolve appreciable amounts
of boron or carbon. In every instance,
however, a portion of the copper was
lost by extrusion past the rams, and
the resulting compacts were still too
porous and weak to be useful.

The 1nitial runs in which iron was
used as the bonding agent were also
being too soft and
There was,

unsatisfactory,
porous. however, no
evidence of loss of iron during the
pressing cycle; the iron may have
been converted to solid FeB or Fe,C.
This is confirmed by the work of
Nelson, Willmore, and Womeldorph,(l)
who pressed and sintered mixtures of
20 wt % 1iron (7.5 vol %) and 80 wt %
boron carbide at 1930°C for 30 minutes,
These sintered bodies were fairly
strong and underwent some shrinkage,
with no loss of material. Examination
by x-ray-diffraction techniques failed
to detect iron, but revealed the
presence of FeB, B,C, and graphite.
It thus seems that the bonding phase
is FeB formed by reaction of iron and

(1)J

Woneldo;ph,
(1951).

A. Nelson, T. A. Willmore, and R. C.
J. Electrochen, Soc. 98, No. 12, 465

boron carbide
cycle,

during the sintering

On the basis of these conclusions,

several runs were made at higher
temperatures; the results are shown in
Table 2. After several runs a hot-

pressing cycle of 5 min at 1525 to
1535°C was selected. This schedule
gave well-bonded slugs having adensity
close to 80% of theoretical with only
3 to 4% loss of material during press-
ing,

Production of Safety Rod Inserts.
Having selected boron carbide and iron
as the combination to be used, quanti-
ties of the powders were prepared to
give 80 vol % (56 wt %) boron carbide
and 20 vol % iron. Coarse boron
carbide powder was reduced to a fine
powder by grinding for 16 hr in a
steel ball mill with steel balls. The
sieve analysis of this material 1is
given in Table 3. The requisite amount
of -325 mesh iron powder was then
blended with the boron carbide by
grinding for an additional 8 hr in the
same mill,

The mixture was consolidated by
hot-pressing at 1520 to 1530°C and
2500-ps1 pressure in graphite dies.
TABLE 2. ADDITIONAL PROPERTIES OF HOT-PRESSED IRON PLUS BORON CARBIDE COMPACTS

 

 

 

PRESSIN , S PERCENTAGE
COTE??I;:DN TEJ?%;:TJ;E DFNiflffifiZOBETICflL fi%fi? REMARKS
(oG DENSITY

20 Fe-80 B,C 1520 77.6 Strong

20 Fe-80 B,C 1725 70.5 29 We ak

7.5 Fe-92.5 B,C 1740 68.2 6.0 | Fairly strong
20 Fe-80 B,C 1520 80.0 3.0 Strong

20 Fe-80 B,C 1520 77.8 4.2 | Strong

20 Fe-80 B,C 1525 80.3 5.5 | Strong

20 Fe-80 B,C 1530 77.8 6.5 | Strong

20 Fe-80 B,C 1525 79.0 6.5 | Strong

20 Fe-80 B,C 1520 78.3 8.0 Strong

 

 

 

 

 

Densities approximately 80% of theo-
retical for the mixture were obtained
under these conditions, One of the
assemblies used for the pressing
operation is shown in Fig. 1.

For assembly, the long punch and
mandrel were placed in the die with

TABLE 3. SIEVE
BORON CARBIDE POWDER

+100 mesh
-100 +200 mesh
-200 +325 mesh
-325 mesh

ANALYSIS OF

1. 2%
17. 2%
18.8%
62.8%

UNCLASSIFIED
¥-TOB9

 

M
dwfifié‘,{

Fig. 1. Graphite Die Assembly.

 
 

 

the mandrel centrally located along
the axis of the die., The punch extended
out of the bottom of the die. A
charge of 132 to 133 g of blended
powder was then loaded into the annular
space between the die and mandrel, the
top punch put in place, and a load of
2000 psi applied to seat the punches
prior to placing in the hot-pressing
rig. The set-up for hot-pressing is
shown 1in Fig. 2.

The lower punch rested on a graphite
block, which, in turn, was 1nsulated
from a Transite base by two refractory
bricks (Norton, RA 1190). A quartz
liner served as a retainer for the
bubble-alumina insulation poured 1in
around the graphite die prior to
heating. The quartz cylinder was, 1in
turn, externally supported by a
Transite sleeve that extended up to
the induction coil. Power was supplied
by an Ajax, 40-kw converter set at
30 kw for heat-up and at about 15 kw
to hold at temperature. An argon
atmosphere was introduced at the
bottom of the assembly to decrease the
rate of oxidation of the graphite dies.

It was found necessary to use
additional graphite spacers, as shown
in Fig. 2, to prevent excessive heating
of the press ram because of coupling
with the magnetic field of the induction
coil. A prism was mounted over the
sight tube of the die so that optical
temperature readings could be made in
a horizontal position. Also, a low-
pressure gage was added to the press
for this work, This gage, which 1is
not shown in Fig. 2, 1s connected to
the hydraulic line above the gage
shown,

It was found necessary to push the
mandrel out of the slug at high temper-
ature to avoid cracking during cooling
because of a marked difference in
expansion coefficients for the boron
carbide compacts and graphite. The
mandrel was pushed out of the compact
into the lower punch by using an
undersize mandrel,

The hot-pressed slugs were ground
to within the limits of tolerance for
length with a diamond wheel and brazed
into a finished azsembly, as shown 1in
Fig. 3. The brazing operation was
carried out in a hydrogen atmosphere,
and Nicrobraz alloy was used.

Experimental Work om Control Rod
Inserts. The first step 1n the work
on control rod inserts was toestablish
the compatibility of aluminum oxide
and boron carbide at the elevated
temperature used for fabrication. A
preliminary slug, less than full size,
was hot-pressed at 1750°C for 5 min
from a mixture of 90 vol % aluminum
oxide {(Norton, grade 38-500) and 10
vol % boron carbide (Norton, metal-
lurgical grade). The powder was con-
tained in a graphite die and con-
solidated at a pressure of 2500 psi.
Metallographic examination revealed
that the two materials were compatible.
A second slug containing the prescribed
amount of -325 mesh boron carbide had
a density 86% of theoretical and
possessed satisfactory physical
properties,

A full-size alumina-boron carbide
cylinder was pressed by using the
above conditions of temperature and
pressure. Considerable difficulty was
encountered in ejecting the slug from
the die. There was also a marked
tendency for the bubble-alumina in-
sulation to sinter at this temperature.
Subsequently, another full-size slug
was pressed, and a maximum temperature
of 1650°C was used. The compact was
removed without difficulty, and
sintering of the i1nsulating material
was minimized. The density of the
second compact dropped to 80% of
theoretical but the physical properties
were still acceptable. The boron
carbide investment was increased
proportionately to compensate for the
decreased density.

Production of Control Rod Inserts.
The hot-pressing equipment used for
making the boron carbide-iron parts
was also employed for producing inserts
UNCLASSIFIED
Y-7059

&

¥

 

 

 

 

 

Hot-Pressing Rig.

2.

Fig.
 

Fig. 3. Assembly for
for the control rods. New dies were
used, however, to avoid additional
boron pickup. Inserts for two rods
were to be made; one set with a boron
carbide content of 0,022 to0.024 g/cm?,
the other with a boron carbide content
of 0.005 to 0.006 g/cm?.

A special grade of alumina was used
(Norton, grade 38-500). Batches of
about 1 1b were prepared by mixing
appropriate quantities of the alumina
and =325 mesh boron carbide (Norton,
metallurgical grade). Fine boron
carbide powder was obtained by ball-
milling relatively coarse stock and
screening out the desired fraction,
Blending was carried out by tumbling
in a glass jar for approximately 8
hours. The boron carbide additions
were such as to give the correct boron
investment for a hot-pressed compact
having a density 80% of theoretical.

The technique for hot-pressing the
alumina-boron carbide powder es-
sentially duplicated that used on the

UNCLASSIFIED
Y-7080

Boron Carbide-Iron Slug.

boron carbide-iron powder, except for
an increase in maximum temperature to
1650°C, Components for an assembly
and a completed unit are shown in
Fig. 4. The brazing operation was
again done by using Nicrobraz alloy in
a hydrogen atmosphere.

HRE CONTROL ELEMENTS

The design of the control elements
for the HRE calls for plates bent into
semicircular shapes. The plates will
be exposed during operation of the HRE
to high-pressure water and gas. The
plates are to be made of a high cross-
section material contained in a stain-
less steel jacket., For the first set
of elements built, sheet stock of
boron carbide dispersed in aluminum
was canned in type 347 stainless steel
by covering the boron carbide stock
with the steel and edge-welding.
Several of these plates were found to
distend following return to atmospheric

 
UNCLASSIFIED
Y=-7T0814

DUTER TUBE &

INNER TURE ™

 

Fig. 4.

pressure after exposure to operating
pressures, Failure was accredited to
an accumulation, as aresult of leakage
through porous welds, of high-pressure
gas 1n the space between the cladding
and insert. The porosity may have been
caused by melting and alloying of the
aluminum with the weld metal.

The purpose of this investigation
was to determine a procedure of fabri-
cation that would minimize the defects
found in the first set of plates.
Hot-rolling was employed to eliminate
the void by bonding the cladding to
the core. In an effort to use ma-
terials readily available, bonding of
the boron carbide-aluminum sheet with
stainless steel was attempted first.
Work was later shifted to the cladding
of pressed-powder mixtures of boron
carbide suspended in copper, since
this method appeared to offer a better
chance for success in the limited time
available for the work.

Reproducibility of the quality of
the edge weld was assured by using a
straight-line arc-welding train.
Conditions were determined for operation
of the arc welder to obtain sound
welds without melting of the aluminum
core.

 

Assembly for Boron

Carbide—Aluminum Oxide Slugs.

Experimental Work on Stainless
Steel—Boral. A preliminary check on
the feasibility of bonding stainless
steel to boron carbide-aluminum sheet
stock was carried out on material com-
posed of 37 vol % boron carbide and
63 vol % aluminum (35 wt % boron
carbide and 65 wt % aluminum). This
material is called Boral and is of the
same composition as the stock used 1in
the preparation of the original set of
control plates.,

Cores measuring 3/4 in. in diameter
were prepared for small samples by
pressing =100 +200 mesh boron carbide
and =100 mesh atomized-aluminum
powders, in the volume ratio given
above, at a pressure of 50 tons per
square inch, These compacts were
placed in frames blanked from type 347
stainless steel,

Cladding plates for the cores were
prepared as indicated in the following
groups:

Group 1. Thin layer of type 302 stain-
less steel powder sintered
at 1150°C to the inside sur-
face of cladding.

Group 2. Thin layer of copper powder

sintered at 1020°C to inside
surface of cladding,
Plain cladding, bright
annealed in a dry hydrogen
atmosphere,

The cladding and frames were spot
welded to hold them together, heliarc
welded around the edges, and given the
treatments listed in Table 4,

Group 3.

TABLE 4. TEMPERATURES AND PERCENTAGE
REDUCTIONS FOR EXPERIMENTAL
ROLL-CLADDING OF BORAL
WITH STAINLESS STEEL

 

 

 

TOTAL
CLADDING ROLLING REDUCTION
GROUP TEMPERATURE (%)
1 600°C, 3 passes 32.5
1 600°C, 1 pass; 31
500°C, 2 passes
2 600°C, 3 passes 35
2 600°C, 1 pass; 34
500°C, 2 passes
3 600°C, 3 passes 33
3 600°C, 1 pass; 37
500°C, 2 passes

 

 

 

Group 3 claddings were prepared to
check straight metallurgical bonding
of stainless steel to aluminum. The
copper layer of Group 2 was introduced
as an ald to bonding, for possible use
in case a bond was not obtained between
stainless steel and aluminum. The
sintered layer of stainless steel
powder 1n Group 1 was 1introduced to
offer a rough surface for mechanical
keying in the event metallurgical
bonding could not be obtained at the
rolling temperatures employed.

After the specimens were rolled,
samples were taken from each of the
three groups for metallographic exami-
nation, thermal cycling, and autoclave
tests. Metallographic examination
indicated that bonding had occurred at
the Boral-to-cladding interface in
each group. No cladding-to-core

reaction zone was evident in those
samples rolled partially at 500°C,
Samples reduced entirely at 600°C did

show a region of interaction between
the Boral and the cladding stock,
Bonding of cladding to the picture
frame occurred in the Group 2 samples
that had the sintered-copper layer.
Figure 5 shows the interface region
for a sample from each group rolled at
600°C.

Neither cycling 25 times between
room temperature and 500°F nor exposure
to helium for 18 hr at 1500 psi
following sudden pressure release was
found to affect the samples.

Several samples, measuring 2 by
8 i1in., with Boral cores and copper-
coated stainless steel cladding were
next prepared. Rolling was carried out
at 600°C to ensure bonding. At this
temperature, the Boral was reduced
preferentially and presumably built up
sufficient pressure ahead of the
rolls to burst an end from each of the
laminates after four or five passes,
Altering the per cent reduction per
pass seemed 1lneffective as long as the
total reduction exceeded 30%. Subse-
quently, two laminates similar to the
above were processed by using three
passes at 600°C to obtain 20% reduction
and two additional passes at 500°C to
obtain a total reduction of 30%. These
plates rolled very satisfactorily, and
there was no evidence of the end
rupture experienced with earlier

samples,
The edge-welding problem was 1in-

vestigated by using a 6-in, length
from one of the 2- by 8-in. plates
prepared as outlined above. The

excess stock along the length of the
sample was sheared off at an angle to
the core so that the core-to-edge
distance varied from 3/8 to 1/16 inch.
The plate was sealed along this sheared
edge by using the straight-line arc-
welding train, The arc was observed
for any irregularity in appearance, and
irregularity was detected for edge-
stock thicknesses of the order of 1/8
inch., Examination of the edge of the
plate under the microscope after the
welding operation showed defects at

 

 
 

——— - UNCLASSIFIED
¥=-5573

—=— TYPE 347 STAINLESS STEEL
CLADDING

-s— SINTERED LAYER OF TYPE
302 STAINLESS STEEL

       

Rl b e aghe U e R
(a}-'-“":v':.;fitlt‘:"{{é &:‘;"3 Tl NS y s H“u"*#%fi
i %"P-‘.h-- VT MG S Y Lt A ™ .:'}'*".‘:T__fi_*a‘..fifim
UNCLASSIFIED
¥-5574
- ® <
- . & -— TYPE 347 STAINLESS STEEL
ok ; CLADDING

-%— SINTERED LAYER OF COPPER

 

-=— BORAL
UNCLASSIFIED
’ Y-55T72
. -%— TYPE 347 STAINLESS STEEL
CLADDING

==— STAINLESS STEEL-ALUMINUM
REACTION ZONE

   

- = . - g - . i .- ‘l "~ ¢ . . I
. A b * Ly ."

LA PR el . )
- . ¢ . &

Fig. 5. (a) Interface of Stainless Steel Cladding with Sintered Stainless
Steel Powder Against Boral Core. 500X (b) Interface of Stainless Steel
Cladding with Sintered-Copper Powder Against Boral Core. 500X (c¢) Interface
of Stainless Steel Cladding Against Boral Core. 750X

 
thicknesses up to 3/16 inch. At best,
the 1/8-in.-wide frame requested for
the stainless steel clad Boral 1s a
border-line requirement for sound edge
welds.

The results of the experimental
vork on Boral-stainless steel laminates
indicate that full-size plates can
probably be prepared by using the
procedure outlined above,

Experimental Work on Stainless
Steel-Borcu. Stainless steel laminates
rolled at 600°C work-hardened quite
rapidly, and the resulting plate was
very rigid. Since the plates were to
be formed into semicircular shapes
after rolling, with a radius of about
7 3/4 in., for the smallest-diameter
pair, it would be helpful if the
rolling process could be carried out
at somewhat above 600°C. Obviously,
this would necessitate replacing the
aluminum with some other carrier for
the boron carbide.

Screening of Potential Core Metallics.

Iron, nickel, types 302 and 410 stain-
less steel, and chromium powders were
considered as substitutes for aluminum.
Coarse evaluation tests were made by
pressing compacts containing 37 vol %

boron carbide with each of these metals.
The compacts were sintered at 1150°C
for 30 min with the results listed 1in
Table 5. These tests indicated that
in all cases the boron carbide reacted
with the matrix metal to formabrittle
intermetallic or low-melting eutectic
phase, with accompanying detrimental
effects to the physical properties of
the compacts.

Laminates containing cores of the
above compositions gave uniformly poor
results after hot-rolling. Several
laminates containing iron or type 302
stainless steel inthe cores wererolled
at 1225°C; at this temperature the
cladding was attacked by the core
material and most of the center portion
melted away. Several other laminates
containing iron, type 410 stainless
steel, and chromium matrices were
rolled at 1050 to 1125°C; all these
laminates blistered badly because of
ruptures within the cores,

Copper was considered next for the
metallic of the core. Information from
the literature indicated that copper
does not react with or dissolve
appreciable amounts of boron or carbon
at temperatures far above i1ts melting
point,

TABLE 5. PROPERTIES OF SINTERED COMPACTS OF NICKEL, IRON,

TYPE 302 STAINLESS STEEL,

TYPE 410 STAINLESS STEEL,

AND CHROMIUM PLUS BORON CARBIDE

 

 

 

 

DENSITY AS PERCENTAGE
COMPOSITION OF THEORETICAL DENSITY PROPERTIES
Green Sintered
Ni-B,C Ni-B eutectoid formed and
flowed out of compact
Fe—B4C 76.5 70.0 Brittle, fairly strong,
liquid phase formed
Type 302 stainless 74.0 59.5 Very brittle, liquid
steel-B,C phase formed
Type 410 stainless
steel-B,C Very brittle, weak
Cr-B,C Brittle, poorly sintered

 

 

 

 

10
A compact of copper and coarse boron
carbide was prepared by sintering a
mixture of the powders and then cold-
pressing, resintering, and cold-
rolling the mixture. The compact was
successfully clad by hot-rolling at
1000°C, with a 50% reduction in seven
passes., Figure 6 shows the interface
between type 347 stainless steel
cladding and the boron carbide-copper
core. The finished laminate could be
bent to a l-in. radius without failure.
When bent further, failure occurred
through the core and not at the bond
surface.

A revision in the scheduled time for
completion of the full-size plates for
the HRE forced an early decision as to
which fabrication method would be used.
The copper-bearing core was chosen as

 

STAINLESS o 0 &

e,

.-1‘ -eas . T & ) ! 4

Fig. 6.
Core. 250X

 

the material most likely to fabricate
satisfactorily.

Preparation of Boron Carbide-Copper
Cores. The boron carbide bearing
section for each plate was assembled
by using multiple inserts. A press
with 150-tons capacity was available
with which to consolidate the powders.

Since a considerable amount of wear
caused by the hard boron carbide
particles was anticipated, it was

decided to form the compacts in a die
at 25 tsi, sinter, and coin the sintered
compacts between hardened-steel plates
at 50 tsi, in an effort to minimize die
wear. A hardened-steel die was there-
fore made to form 1- by 3-in. compacts
of the required thickness.

Stock for cores was prepared as a
mixture of 41 vol % (16.3 wt %) boron
carbide (Norton, metallurgical grade,

Rl 4 - NyncLassiFien)

Y-5692

e - - N
o . wr
».

: R

Interface of Stainless Steel Cladding Against Boromn Carbide-Copper

11

 
which analyzed 71 * 1% boron) in
copper powder (U. S. Metals Refining
Company, type C). Two per cent of
stearic acid dissolved 1n acetone was
added to minimize classification of
the boron carbide during handling of
the mixture. The acetone was driven
off by gentle heating prior to com-
pacting. The compacts were pressed at
25 tsi, heated in air at 400°C to
remove the binder, and sintered at
950°C for 2 hours. The compacts were
then repressed at 50 tsi between
hardened-steel plates and resintered
for 15 min at 950°C to promote densi-
fication. Preparation prior to
cladding was then completed by cold-
rolling to the desired thickness,
loading the laminates into the frames,
edge welding, and evacuating. Finally,
the composite plates were hot-rolled
at 1000°C, as described above. It was
discovered during the preheating period
for rolling the large plates that an
outward distortion of the cladding

occurred. The cause could not be
determined with certainty. The edge
welds on all plates were checked for
soundness prior to rolling by a test of
ability of the laminate to hold a
vacuum. The possibility of a volatile
residue from the binder after firing
at 950°C does not seem likely. By
delaying sealing of the exhaust tube
for the laminates and applying vacuum
to the core during the preheating
cycle, the difficulty with swelling
was eliminated. Subsequently, camphor
was substituted for stearic acid as a
binder to help minimize residual con-
tamination following sintering.
Components for preparation of a
laminate are shown in Fig. 7.

The rolled laminates were x rayed
to locate the cores and the excess
stainless steel stock was sheared from
the edges to leave aminimum of 1/8-in.
border, which was sealed by using the
automatic argon-arc welder. In this

UNCLASSIFIED
¥Y-T284

Hot Rolled Composite

Pressed 8 Sintered Cu B4C

 

Fig. 7.

12

347 Picture Frame

347 Cladding

Components for HRE Control Plate.

 

 
operation, the boron carbide-copper
mixture possesses an additional
advantage over Boral in that its higher
melting point reduces the chances of
melting of the core and the associated
contamination of the weld.

The actual core area for each plate,
except plate 8, was determined from
the x-ray pictures; plate 8 was
delivered for a bending experiment
before an x~ray picture was taken. The

edges of plate 8 were trimmed by using
the outline of the core that was
visible on the cladding surface as a
guide,

The level of the boron carbide
content of these plates was slightly
below the 100-mg minimum requested,
except for plate 16. The deficiency
may be accredited to a lack of famili-
arity with the properties of the
powder mixture.

13

 
 

APPENDIX

A. HRE CONTROL PLATE DATA

 

 

PLATE NUMBER

 

 

6 10 5 9 1 14 12 16
Over-all length, in. | 16 1/8 16 1/8 17 1/8 17 3/16 18 1/8 19 11/16 | 19 3/16 13 13/16
Over-all width, 1in. 1 1 1/32 15/16 15/16 19/16 6 1/14 6 1/16 9 5/16
Core length, in. 15 3/4 15 3/4 [16 5/8 16 11/16 | 17 3/4 19 5/6 18 9/16 13 5/16
Core width, in. 11/16 21/32 15/16 15/16 11/8 53/4 513/16 | 9 1/32
Core area, in.? 10.8 10.3 15.6 15.65 20.0 111.0 107.9 120.3
B,C-Cu investment, g | 56.3 53.7 82.8 82.9 108.3 596.3 599.3 744
B,C investment, g 9.17 9.08 13.5 13.5 17.565 ar1.2 97.7 122
B investment, g 6.46 6.46 9.62 9.62 12.56 69.2 69.5 87.6
B per cm? of core, g |92.8 97.3 95.7 95.4 97.5 97 100 113

 

 

 

 

 

 

 

 

 

 

*Plate deliveared for bending experiment before x-ray picture could be taken for determination of aize.

Equipment:

Material:

Welding Conditions:

14

APPENDIX B.

WELDING CONDITIONS HRE CONTROL PLATES

G-E, WD4200, d-c welder

Linde, CM 37, machine carriage

Airco, machine heliweld holder

G-E, hi-thoria, 0.040-1in.-dia tungsten electrode

Nine stainless steel clad boron carbide and copper sheets,
1/8 in. thick, in varying widths of 1 to 9 1/4 in.;
accumul ated edge welding required was approximately
27 ft for the nine plates and varied in length from 13
to 19 inches.

Shielding gas,
Gas flow, 25 ft®/hr
Generator open-circuit voltage, 70 volts
Arc voltage, 10 volts

Arc current, 46 amp

Welding speed, 10 in./min

argon