Elastomeric Seals and Materials at Cryogenic Temeratures
Elastomeric Seals and Materials at Cryogenic Temeratures
Elastomeric Seals and Materials at Cryogenic Temeratures
AD 214 1178
UNCLASSIFIED
NOTICE: When government or other drawings, speci-
fications or other data are used for any purpose
other than in connection with a definitely related
government procurement operation, the U. S.
Government thereby incurs no responsibility, nor any
obligation whatsoever; and the fact that the Govern-
ment may have fornalated, furnished, or in any way
supplied the said drawings, spec fications, or other
data is not to be regarded by iaplication or other-
wise as in any manner licensing the holder or any
other person or corporation, or conveying any rights
or permission to manufacture, use or sell any
patented invention that may in any way be related
thereto.
ASD TDR-6Z-31
TISIA b
NO Oi
AERONAUTICAL SYSTEMS DIVISION
NOTICES
When Government drawings, specifications, or other data are used for any purpose
other than in connection with a definitely related Government 1jrrcurement operation, the
United States Government thereby incurs no responsibility nor any obligation whatsoever;
and the fact that the Government may have f-s-aAW- i.: tiw i had. or in any way supplied
the said drawings, specifications, or other data, is not to be regarded by implication or
otherwise as in any manner licensing the holder or any other person or corporation, or
conveying any rights or permission to manufacture, -. ie, or sell any patented invention
that may in any way be related thereto.
Qualified requesters may obtain copies of this report from the Armed Services Tech-
nical Information Agency, (ASTIA). Arlington Hall Station, Arlington 12, Virginia.
Copies of ASD Technical Reports and Technical Notes should not be returned to the
Aeronautical Systems Division unless return is required by security considerations, con-
tractual obligations, or notice on a specific document.
ASD TDR-62-31
PUBLICATION REVIEW
M.elble, Chief
Elsorers and Coatings Branch
Non-Metallic Materials Laboratory
Directorate of Materials and Processes
iii
TABLE OF CONTENTS
Page
Introduction 1
1. Static Seals 1
1.1 O-Ring Seals 1
1.2 Heavy Plate Seal Tests 3
1.3 Functional O-Ring Seals 11
1.3.1 Extreme Lightweight Flange 13
1.3.2 Successful Flat Flanges 16
2. Moving Seals 21
3. Physical Properties Program 21
iv
Table of Contents (continued)
References 67
v
LIST OF FIGURES
Figure Page
vi
List of Figures (continued)
Figure Page
23 Loading and Unloading of Sample A at Room
Temperature 60
vii
LIST OF TABLES
Table Page
1 Flat Flange Seal Tests 4
2 Heavy Flange Test Results 7
3 Recipes and Properties of ASD Materials Used
for Heavy Flang Tests, Thermal Expansion,
and T 8
g
4 Seal Tests Using 2. 5" and 3" Diameter Flat
Flange s 19
5 Thermal Expansion and Tg 40
6 Torsion Tests at 76°K 46
7 Recipes and Properties of ASD Materials Used
for Torsion Tests 47
8 Composite Inorganic Test Specimens 55
viii
Introduction
It has also been shown by previous work [ 1 I that thin flat gaskets
of poly (ethylene terephthalate) or nylon make excellent cryogenic seals
with relatively low flange loads. The key to success in this case is the
method of applying sealing force to the gasket, and consists essentially
of setting up a narrow ring of highly concentrated stress to form the seal.
1. 1 O-Ring Seals
The elastomer o-ring concept has been explored in two additional
ways. These are the use of o-rings of small cross section in the tongue
and groove flange design and the use of o-rings between flat plates. 0-
rings of 1/16 inch cross section diameter have made satisfactory seals
in a tongue and groove flange design and required about half as much
bolt loading as 1/8 inch o-rings of the same material. Thus one of the
principal disadvantages of the original seals of this design can be sub-
stantially reduced.
The original tongue and groove flange design was dimensioned
to closely confine the o-ring at the end of the linear compression, and
then subject the confined elastomer to about 5% volume compression.
This required close machining tolerances in the flange construction, and
high bolt loads at the end of the compression. Although the resulting
seal is very reliable, and can be made with a large number of elastorner
compounds, it has now been found that some of the elastomers make
satisfactory seals without lateral or "side" confinement.
I
x .
0'
1V
cuU)
___ - cn
Cz
There are several advantages to this modification: (1) the
flange loading is about half that required for a confined o-ring of the
same cross section; (2) the flanges require a minimum of machining
and there are no close machining tolerances; and (3) the surface
finish of the flanges is not critical - in fact, a somewhat rough
machine finish is advantageous since it helps prevent excessive flow
of the otherwise unconfined elastomer.
The top flange used for the tests shown in Table 1 was 3/8-inch
thick stainless steel, and compression of the 1-inch diameter o-ring
was by means of six 3/8-inch steel studs on a 2 1/4-inch bolt circle.
This construction is quite rigid, considering the size of the o-rings
used, but flange loading was in most cases high enough to cause
significant strains in the metal parts. This distortion of flanges and
studs "spring loads" the o-ring so that differential shrinkage during
cooldown is less likely to separate the brittle elastomer from the
confining flanges. A more severe test of the o-ring seals would be
one in which this spring loading effect of bolts and flanges has been
eliminated.
3
-0 0 u. r0 0l 00 0- 0n 0n
-4 Cd -
844)4
to UU) U
0 0L0 0 0A
0 04 0
za LA LA
>4 4 0 z
$400 0 0 0'. 0 0
4))
z z z Z~ > z
(44) ) 4) (0
En V u
0n 10 knl)C 0l
0. 4
4.~0 L t Ln tn 0 Ln LA A
00. Ln 00 N~
4) 0 0 0 0 0 04
.0( 0 0
0 9.0.
>.. 0 :
0. NC -' N 0 V
0 0 ' '3 N - to
ot 0 0
M U 0 0 t
0 t0 : , 0 0
>. 11 z
.00 (4..
-- 4 - ~4 U ~ ~ ~ 1.
.
-
0 0
.0 .. 0
0 004 Z ~ u i.. W 0 04 4 ) 0
044) 4) $ 04
4) (44 0 0
0. 04 4.8 0>
.00 ( 0 0
84 0 0
TO LEAK DETECTOR
V" THREADED
STUDS
MILLED :1 H I,,
DEPRESS ION "0" RING
HIGH PRESSURE
SOFT HELIUM
SOLDER
]> 5
0
c'J
The amount of squeeze (percent compression) of the o-ring was in-
creased by placing thin metal discs in the bottom of the depression,
or decreased by placing a thin metal insert between the flanges at the
bolt circle. All surfaces were given a normal machine finish.
6
v AE ' 1.1 II 4 li 'A A l I 14I 4.
4) * E
0 0
*0
4) 1
4)>. 0
4) n n4 .0I
0 0 0 0 000000o
00 000 00 00
o> 1. 1. -vQ
7- 0 0
4)> vIIfz N z I 4 zz v w Z Z
(n) .;
N:
F-.
0VwoI ~ ~~ ~ ~
nInct ~ V4)Da no o ~ ~ oI ~n 444
cI
1 oor6 -a oa oa r DO -0 4))4 -6
a. o*
0 0
v C s.
0 c
13 Iv o 4
<VI 0
E) u 00 00 Ir uri o wrgp~ u0 00 f u u 0
o r 0
0 o 0 0 to a. a. 0
u 0 0 0 au u
0 0 00 00 .~000 0 -7
'0
U)
0.0
00
14
LLA
0 n LA 0 LA) 0f -n0t
m - * 8 .0o
4) . -(d. o~r
00 4H.9 H
o U* 4 X~.~- 0
x0j* ~ o 0 (
$4 N 0 W
UZ0~~
>,~N Z 04 d 44:~
u~3 (04 4 oX e
14 34
0.F
.00
34$ a.,~0
0. a
00
.0 C00
> 0.
0 >0.
z 00
0< F( e-
0 0 U.
N co *oL
L)8
m 00
1 4
ka 0o 00ao t 00
0 0 0 0o 0
ON N 0 LI-0L 0 -ni ne
:::i1
044
m$. - 4 00 0
.
0
I. Cr
Z
o0
D
01
554
0 0
Z: U u
to-
U- . -. -
U u u x u9
In . 070-inch width the latter two materials showed little or no material
failure at 88% true compression, and made reliable high vacuum seals
at temperatures from room temperature to between 76 and 93°K, and
at pressures to 1250 psig. Satisfactory seals were also obtained with
these compounds in . 140-inch width o-rings, but only after com-
pression reached the point of severe material failure.
10
THIS
PAGE
IS
MISSING
IN
ORIGINAL
DOCUME NT
Helium
Pr essu r e
To Leak Detector
Flange Seal
Assembly
I Liquid
N2 or H2
12
cycling was accomplished by means of an off-on switch which operated
the three-port solenoid valve shown in Figure 3.
13
92LBORE, 8 HOLES EQUALLY SPACED ON 4.000 B.C.
LOCATE WITHIN 0.010 OF TRUE POSITION
3.0850D
.R9
.0 9 0
3L
.00 -. 09
1.5
15
1. 3. 2 Successful Flat Flanges
16
7--9/32 BORE, 10 HOLES EQUALLY SPACED ON 3.345 B.C.
I 7 0.200A
1 5/32
1/2
f J-
j~ 2 .5 10 D
17
9/32 BORE, 8 HOLES EQUALLY SPACED ON 4.000 B.C.
- 3 .0 8 5 D
.090
R0.640
SI I A
3.005 D 0.200
18
Fitinge. (-Riitg Si ite Tlo,1 oil Bolts Colillivits and Reisults
li g. 7 ' 1.H
. D). I 2" i-Il, Al. lnot anodlized. (Jsed( retainintg
steel to Al. 070''0W sl-eve. No ltak at 1 50Opsig, roomn
Fi g. 7 (07' li I1ii
I. 0 itll 11) 1I hu
itwate(-r to .N 2 (yclIe s 1I WE'
st vI to ,tt-l 0. Y"' Wir-v ove rnight bake. 410 cycles, warm
i-.'4 Ring. all(d (1l, 1 'JO psig to I attn. No
I iatil leak at 250O p5 ig rootii t(-ttip. Nfo
le,-ak at 1 5 0 ps ig 7 f, "K.
ditto 3. 30' 1. 1). 100 in-lb (Flange OK warm, held 1 atm at 7t 'K
.070"W edges touch at 50 Small leak at 1 50 psig, 7i 'K
in- lb)
,
Fig. 7 2. 84 I. I). 1_ S I-lb Al. not anodized. U.,ed .hvev
steel to Al . 070 W OK at I 0 p. , warm. l,eaked
during t'ooldtwn.
Fig. 7 3. 239" 1. I). 150 in-lb Al. not anodized. renip cycled 5
steel to Al . 070' W times, pressure cycled 10 times.
No leak at 7t,'K, 110 psig.
Table 4 (Continued) SFA[L TESTS USING 2. 9" and 3" DIAMETER FLAT FLANGES
"0
into a fairly uniform cone shape (although some of the seals deformed
into a flower pattern, showing some bowing between bolts). When
the flexing was relatively uniform it served to reduce the initial
compression required to maintain the seal during cooldown. Measure-
ments indicated that good seals were being made with 65 to 70%
compression of the o-ring, whereas 80 to 90% is required if spring
loading through distortion of flange parts is eliminated.
2. Moving Seals
An apparatus for study of packings for slow-turning shafts has
gone through several modifications and is shown in its present form
in Figure 8. It was necessary to go to the double seal arrangement
shown in order to eliminate end thrust on the shaft when pressure
was applied to the seal. Attempts to oppose this thrust through a
ball bearing were partially successful, but measurements of the
torque required to turn the shaft were more difficult to interpret
because the resistance of this bearing varied with pressure and
temperature. Another important improvement in the present
apparatus is lining the packing sleeves with molybdenum-filled and
reinforced polytetrafluoroethylene.
3. 1. 1 Compression Dilatorneter
When ani aln orphous rubber-like material composed of long
chain trolecul'(- rdnird,,1y distributed is compressed, the chains tend
21
ROTATING
SHAFT
HELIUM" GA T
LEAK DETECTOR
OR GAS METER
DISC
REINFORCEDPAKN
IOINTS
HELIUM
PR ES SURE
22
! Warm Seal and
Drive Pulley Leak "a out
at top - -
--Hollow Drive
Shaf t
Flexible
'I Coupling
Pressure and
'Condensate Line
B
Seal Housing
and Torque Tube
Rotating Face
Bellows-
Type '-S
]i
Sleeve Seal s
*-Seal"b
. :. Leak "b"out
Liquid N2
or
Liquid H2
CODE
S Cryogenic 0-ring Seal
B Cryogenic Bearing
T Torque Tube Bearing
24
DIAL INDICATOR
OR MICROMETER STAINLESS
STEEL
LEVER
THERMOCOUPLE
IsTA STAINLES
VAR
STAINESSIN
BELLOWS
BE LLOW S I )( BE LLOW S
PURGE i ( PRESSURE
LINE -,LINE
26
z 0
w2
-,cr 0
U)-
0 crw0
C/)~
ww
0
-0 -0- a
_ _ _ _ _ _ _ _
C__ _
0.,
CL.
wE
0
0
w0
00
Of-
0
0
N 0
0* 0 o0
SS40U!' HIOLN3-1
27
C
0' V
00
0C
0
U 0
o ho
if I
2 0
w
C~C
0
LLJ
0(
00
0 0
U-
280
temperatures than if the pre-stressing were not performed. Alternately,
a smaller squeeze (and smaller bolt torque) would effect a good seal at
some given temperature. This result was observed in a few of the
functional seal tests, but has not been studied with specific reference to
creep effects.
3. 1. 2 Sealing Theory
We suggest now a method for analytically treating the system
where an o-ring is compressed between two flat plates and cooled
for sealing service in the cryogenic temperature range.
UT - = a E(T I - T2 ) (1)
T T ?1
0
T - 12
TT aEdT (2)
1 I
1
T-= u(T, L)
which,upon differentiation.gives
29
(8T )L
(dT + a
8"L 2 T dL (3)
a- La (8L -
L L T
By definition:
L -iE,
aOLIT
and - (lia
L KT /T
- dL -- aEdT. (6)
1 1
30
3. 2 Force Evaluation Experiment
Four thin layers of mica insulation are placed on the top and
bottom of the bellows caps to provide insulation for the bellows
assembly. The floating plate further isolates the o-ring from the
liquid nitrogen in the bellows and transmits the force exerted by the
o-ring.
31
TO WHEATSTONE - VENT
BRIDGE CIRCUIT- -TO LEAK
DETECTOR
LIQUID N 2 /
STAINLESS STEEL
TUBES
VACUUM
- SPACE
- LOADED FORCE
WASHER
INSULATION
LIQUIDLIQUI "
N2 .- TEMPERATURE
LOADING COMPENSATING
SLEEVES- 0 91FORCE WASHER
THERMO- FLOATING
COUPLE PLATE
LEADS
HIGH PRESSURE
HELIUM
O-RINGI
33
0
0
0
CP
.S
-4--
0'
CP o
0o 0
- 0
IA~ -
00
0 0 VO
0~ In0 ,-
X* '3niL3~
0n
34L
With this arrangement there will be an initial contraction of the
invar sleeves, force washers, and bellows caps. This will take
place immediately and the "initial" force on the o-ring will be taken
at this point. After this initial contraction, the only further con-
traction affecting the percent of compression will be due to the invar
pillars contracting; this should be about . 0004" instead of . 008" as
with the previous arrangement.
3. 3 Thermal Expansion and Second Order Transitions of
Polymers in the Temperature Range 300--76°K
3. 3. 1 Introduction
In the reporting year considerable effort has been directed
toward the experimental measurement and theoretical understanding
of the second order transition point (T ) and the thermal expansion of
polymer materials classed as elastomirs. The objectives of this
part of the program are three-fold:
1. To experimentally measure the low temperature thermal
expansion of elastomers, most of which have not been reported
comprehensively in the literature. Thermal expansion can be re-
lated to o-ring sealability, and is important in other possible low
temperature uses of elastomers.
2. To relate the thermal expansion to molecular structure
and compounding procedures, with the idea of directing the synthesis
of materials with lower thermal expansions.
3. To determine the direction to be taken if one tried to
synthesize a gum elastomer of appreciably lower T than thoseg! avail-
able at present.
A dilatometer has been developed to measure thermal expansions
at low temperatures, and some measurements have been made. The
results, together with a brief discussion of thermal expansion and
some possible interpretations of the measurements, will be reported.
3. 3. 2 Brief Theory
Thermal expansion of polymers seems to be due to changes in
the configuration of the chains. The expansion is due not so much to
the motion of one chain past another, but to the internal motion of the
units within the molecules[ 6]. This change depends on the energy
changes of the long chain molecules and the units which make up the
molecules, and on the various forces present. These forces include
surface tension, primary atomic forces, and secondary molecular
forces. The forces can be varied by many method ': 1) inserting
fillers, 2) plasticizing, 3) varying molecular weight (length of chain),
polarity, symmetry, and orientation.
35
The two commonly designated points which signify the stiffening
of an elastomer when it is cooled are Tg and the brittle point. The T
of high polymers is usually very close to the brittle point. The Tg is
the temperature at which the rotation of the units in the molecules of
polymers stops or begins, depending on whether one is following the
temperature up or down. The brittle point is the temperature at
which a substance shatters under specified standard test conditions.
This depends on how fast the molecules can absorb the applied force.
If the molecules do not react fast enough, the material shatters.
36
He
GAS
L
INVA R
STYROFOAM RODS
INSULATION
$4*9
*.He
- GAS 2-
0 37. *-
The polystyrene is enclosed by a flanged metal can which is immersed
in a bath of liquid nitrogen. Thus the major part of the cooldown is
by slow conduction through the N2 gas and expanded polystyrene
surrounding the sample. The final part of the cooldown is achieved
by introducing liquid nitrogen directly into the space around the
sample. Average cooldown rate is maintained at about 40 Kelvin
degrees per hour.
38
.02a
.020 _ __
.018_ _
.0 16 2
.12E
.0 14 -1 D __
.00 2 - - 12
0 15 12820 25030
.350
40
F
I
F
I
and monochlorotrifluoroethylene C = C If it is assumed
I
C1 F I
that the chlorine and fluorine atoms in monochlorotrifluoroethylene
are similar or more similar in size than the fluorine and hydrogen
atoms are in vinylidene fluoride, then it can be seen that the con-
figuration of I2D with a monomer ratio of 50/50 will be rougher and
have available more "hooks" on its chains than that of 12E with a
monomer ratio of 70/30. This may be one of the causes for the higher
T g of
. 1ZD, which results in lower total expansion even though the ex-
pansion coefficients of the two compounds are identical. Besides
inhibiting the lengthening of the chains, the "hooks" may serve to
capture other branches and hold them in close contact, enabling the
molecules to act on one another with greater force. This may be a
mechanism by which the units in these molecules are hindered in
rotating, thereby resulting in a higher T in 12D than in 12E.
The comparison of 8D, 1ZB and 12C, all of which are copolymers
of vinylidene fluoride and perfluoropropylene with monomer ratios of
70/',, and having the same compounding recipes, should be instructive.
According to ASD, the molecular weight of 1ZB is greater than that of
I2C, which in turn is greater than 8D, and all are between 60, 000
and 100, 000. It can be seen that the expansivities of 8D and 12B are
very similar, with that of 12C larger in the temperature interval
above T Also, the T 's of 8D and 12B are similar, with that of 12C
slightly'igher. From this information we see that molecular weight
is not an important factor in influencing the T or thermal expansion,
providing, at least in the case of T Is 18] the molecular weight is
lar ge. g
41
CF3 F
difference between the form of perfluoropropylene C I = CI
F= F
Cl F
and monochlorotrifluoroethylene
I I
C C The former
I =
I
F F
42
G ~M t
MG (1)
kI O(Za) 3 (Zb)
During a test the top plate rests on a dewar, and the test area
is immersed in the coolant. Weights are placed on the pans and the
resulting twist angle is measured. Hysteresis effects are reduced by
cycling the load several times before the readings are taken. At least
two tests were performed on each sample, and each reported test
result is an average of several load applications. Twist angles are of
the order of 2 to 10 degrees. The weight involved is around 1/2 pound,
which applies about 1. 5 pound-inches of torque to the sample. Sample
dimensions are of the order of 0. 511 x 0. 08" x 2"1 free length.
43
/360 DIVISION
,, SCALE
PULLEY1
STYROFOAMI
BALL BEARINGS
t DEWA R
TEST
SAMPLE
The results indicate that 21D and ZlE are the softest
compounds at 76°K with 8A, 21B and 21A also having shear moduli be-
low 3 x 104 psi. However, the moduli of all materials tested are
within a factor of 2. Assuming a Poisson's ratio of 0. 25, the relation
between G and Young's modulus E is:
E = Z. 5G (Z)
45
ASD Monomer G E
No. Polymer Ratio (psi) (psi)
8A Natural Rubber -- 2.47 x 10 4 6. 19 x 1
1 8D Isobutylene (A)
and 97. 5/2. 5 3.51 8.77
Is oprene
18D Isobutylene
aa-d 97. 5/2. 5 3. 7Z 9. 31
Isoprene
18C Ditto 99/1 3.63 9.07
21D Butadiene
and 77/23 2.23 5.59
Styrene
21E Ditto 57/43 2. 26 5.65
18B Chlorinated
Isobutylene 1/97/2 3.42 8. 57
and
Isoprene
21 F Polyurethane Unknown 3. 57 8.92
21G 1, 1 Dihydroper- -- 2 .9 4 (B) 73 B
fluorobutyl
Ac rylate
21 B Butadiene
and 70/30 2.74 6.84
Ac rylonitrile
21A Butadiene
and 80/18 2.50 6.25
Ac rylonitr ile
46
U
wA 0A at 0 IA
C4)C
.0
04= 0 0 0 0 0
r In In In 0 a
4)0 00 !; 0 0A
M n In Ns
In o ' IA
IA AI IA 14 InI
'A In 0) t) u e 0n
0 Ri 1 0
- *0 0.- N 0 0
11 01 O~4
c 040
0 Go
C) C)
4100,
0 0 - I-
S. 0IA
V-
0 0 0'r
r 04
LL41
4) 4-
U U) En
14
P4
0
O 0 0
.0 00 co. ZD w
00 - (L'
N N N N
4) - e
In4
OcO 0 NO -:I0 -N
O4 V) 0 ) n a. u U
4) -
0-
0 0 ~ 0 0 0
00
z~
uu CI~
48)
properties but has no phase changes in the temperature range of
interest. The sample and the reference are cooled or warmed at the
same rate, and any abrupt changes in the differential between the two
indicates a first or second order transition in the sample. A some-
what simpler approach to DTA has been reported recently[ 14]. This
more recent method, which eliminates the reference material by
measuring the thermal lag in the sample, has been incorporated into
the elastomer program and will be continued through the coming year.
At this time the apparatus will be described and preliminary data
obtained with a natural rubber sample supplied by ASD will be dis-
cussed.
49
EXPERIMENTAL SET-UP INSTRUMENTATION
THERMOCOUPLE TO INDICATOR
THERMOCOUPLE
COUPLINGBUCKING
POTENT IOM ET ER
DEWAR
SAMLEAMLIFIEl
SUPPORT
VOLT
7 DIFF. T.C.
00.5
51
oo '3nlV83,
0* 0 3fL~d~ I 0
o0 0 0 0
T T
____ ___T___
IC
PE N >
a:I
z E
0.
ul CM
3.Jzia
SIIOOJ:IW 'IVILN3
-ldn3OW 3H0
52U
190
16 - _ _ _ _
42 201 K
0 14
E
ACTUAL-S
-J DIFFERENTIA
:512
zw
w
a.
S10
LU
6 (25-3§L~y) ________
4 5 6 7 8 9 10
RUNNING TIME, inches
(inch
19.1 mni n.\
I
54
The promising results obtained from these tests prompted
preliminary investigations of the mechanical and viscous behaviors
of the composites, at room temperature and at 76K. It is hoped
that comparisons between results at various test temperatures will
aid in understanding the composite structure and its possible use for
low temperature seals.
Sample A B C D E
Skeleton 430 ss*',c 430 ss 304 ss Molybdenum 430 ss
Skeleton bonded? Yes Yes Yes Yes No
Skeleton density 23% 23% 22% 31% 23%
Impregnate indium indium indium silver silicone
resin
slurry
'. Stainless Steel
55
2" DIA. LOADING
MEMBER
HIGH
VACUUM
CRYOSTAT
-5/6" STAINLESS
STEEL ROD
-77-
56
3. 6. 2 Testing Procedure
Two samples of the 430 ss + indium composite and one each of
the other three composites were available for testing; therefore more
tests were performed on the 430 ss + indium than on the others. The
tests were performed and reported by T. F. Durham of the Mechanical
Properties section of the N. B. S. Cryogenic Engineering Laboratory.
The deformations refer to total deflection. Thus "the sample was
deformed 5%" means that the total deformation of the sample and the
loading rods was 0. 05 x 0. 5 = 0. 025 inches. Following is Mr.
Durham's report of test procedure on each sample:
57
and unloading curves were obtained for all tests. After warm-
up the specimen was found to have sustained only about 0. 001"
permanent deformation.
3. 6. 3 Discussion of Results
Since the deformation of the loading members played an im-
portant part in the total deflection recorded on the original graphs, the
data were translated to working curves and these corrected curves
plotted in Figures 22 thru 27. In all cases the stre -s was calculated
on the basis of the initial cross section area of the sample.
58
0 0
o 0L
o 0.
0
CP
0 .2
b-
0 cm
It-
isdS3N2
(/-A
590
raft
LL
z0
C
0- 0
0 0
0- 0
cm -
i (V/A 3M
600
U) 0
Wo 0
)-i w
U() 0
z L0.
ONJ
0
("D
0z
CP
0 8 8 0
000 0 0
0 0
61
o.
0
00
0 - a
00
00
0J S
a;
o U)
62
E
4-
E
0
q~. cr
0
H
0
0
Q0
C-4-
I' CL
0
"w E
C)
* 0
Cx
E
o 0
n-
w cm 0
0
o 0 0) 0 0 0
i$d (V/_-) SS381S oM
63 u.
OD
0
.0
Cl)
CL
E
0
V
0
4-
00
0 00 0
0 w 00
V cm to
isd S380 (/j)
64.
than that for other tests, and some analysis of the recovery was
possible. Note that stress was calculated on the basis of the initial
cross section of the sample, not a good approximation for this test
since the area increased significantly during the loading period. Note
also the yield for both tests, and the non-recover,,.ble plastic flow
after yield for the second test.
65
tures. It is well known, for example, that polymers have high ex-
pansivities and become glass-like below characteristic transition
temperatures. These properties might lead one to expect that
elastomers clamped between metal flanges cannot retain a seal at
cryogenic temperatures. The fact that o-ring seals have neverth .less
remained strong and tight far below their brittle point emphasizes the
need for a better understanding of changes in physical properties
throughout the low temperature range.
The work to date has shown that there are no simple correla-
tions between seal performance and expansivity, brittle point temperature,
or compression and shear modulus at 76°K. The measurement of
these properties, and others, from above the brittle point to cryogenic
temperatures is nevertheless very valuable and will eventually enable
us to explain and predict the performance of elastomers as seals.
Thermal expansivity measurements will continue to receive important
emphasis and an attempt will be made to determine shear modulus
and compression modulus at all temperatures between the brittle
point and 76°K. The force evaluation experiment will yield diagrams
of the forces exerted by an o-ring during cooldown and while function-
ing as a seal. Other information, continuous with temperature, will
be obtained from DTA measurements, as described above, and a
rebound experiment which is in the design stage. It is hoped that
most of these properties can eventually be correlated with seal per-
formance, with polymer variation, and with compounding parameters.
66
REFERENCES
67
References (continued)
68
GPO 854396
04
V
Z u 0.
0
0 4)
~ ~Id
.
>
0O 4
0.'
14
0 0 *.O0
04)
u 14* (L~ (
4)4)0
kE
0.I~
0 ~,
o)44
00 40
'o
x~o~
0~~~
Io4~ l
- 4 oI
4)* o' I
<c 04 4) .0
<z