R. L. Tobler, ~ R. P. Reed, 1 I. S. H w a n g , 2 M. M. Morra, 2 R. G.

Ballinger, 2
H. N a k a j i m a , 3 a n d S. S h i m a m o t o 3

Charpy Impact Tests Near Absolute Zero

REFERENCE: Tobler, R. L., Reed, R. P., Hwang, I. S., Morra, tation are necessary when the test temperature nears absolute
M. M., Baltinger, R. G., Nakajima, H., and Shimamoto, S., "Charpy zero.
Impact Tests Near Absolute Zero," Journal of Testing and Evalua-
The construction of large superconducting magnets for fusion
tion, JTEVA, Vol. 19, No. 1, Jan. 1991, pp. 34-40.
energy machines requires the evaluation of tough austenitic al-
ABSTRACT: We review Charpy impact testing at extreme cryogenic loys at 4 K. Computerized J-integral fracture tests of these alloys
temperatures, especially at liquid helium temperature (4 K), consid- at slow loading rates provide a reliable quantitative measure of
ering methods of testing and calibration, thermal behavior during the toughness at 4 K; simpler Charpy tests have been used for alloy
various stages of testing, and correlations between Charpy absorbed development and are sometimes proposed for quality control or
energy and quantitative toughness parameters. Because of the very
low specific heats of metals near absolute zero, any surface conden- acceptance tests. The absorbed energy Cv of a V-Notched bar,
sation of gases, convective or conductive heat transfer, or plastic and the lateral expansion t~ of the fractured specimen at the
deformation during a test will cause the specimen temperature to rise notch plane, are indicators of fracture resistance. As argued in
rapidly. Consequently, valid impact tests of alloys at 4 K can not be this paper, however, the meaning of impact measurements at
performed according to the procedure outlined in ASTM Methods
E 23-88. During Charpy tests, the temperature of austenitic steel temperatures near absolute zero must be questioned in view of
specimens, initially at or near 4 K, may in fact rise outside the cry- problems associated with specimen heating and temperature con-
ogenic regime. Fracture does not occur at the intended temperature, trol.
but at an uncontrolled temperature, since materials with different Although the existing standard does not convey it, to perform
work hardening rates heat differently. In view of the temperature an impact test at 4 K it is necessary to depart from the conven-
rise variabilityand scatter in measurements and property correlations,
we conclude that it is not possible to accurately estimate the 4 K tional test procedure. The specimen, after being cooled to 4 K,
fracture toughness of ductile steels, or rank them properly, using cannot be transferred through air to the test machine in the
Charpy tests. routine way. Instead, various methods of insulating the specimen
are used. Typically, the insulation is fractured along with the
KEY WORDS: adiabatic heat, austenitic alloys, cryogenic properties specimen, and a nonstandard calibration is required. Although
of materials, cryogenic tests, fracture toughness, impact tests, liquid
helium, mechanical behavior of materials, strain rate effects, test specimen heating during transfer may be eliminated using in-
standards sulation, severe internal heating of the specimen during impact
inevitably complicates any interpretation of results.

Simple methods of fracture toughness characterization are de-

sirable, especially at cryogenic temperatures where tests are ex- Procedures
pensive. Charpy impact testing has traditionally been used to
evaluate structural alloys at temperatures as tow as 77 K. For Most cryogenic laboratories are equipped with conventional
aerospace applications, the temperature of interest is 20 K; for pendulum-impact machines, but some use special machines and
fusion energy applications, it is still lower. Naturally there is apparatus for nonstandardized tests. As illustrated photograph-
some interest in performing Charpy tests at these temperatures, ically by Novikov [I] and schematically by Verkin [2], a pneu-
but special procedures and precautions regarding data interpre- matic machine is used in the Soviet Union for tests in liquid
helium. The device consists of a cylindrical chamber containing
a vertically aligned, pneumatically driven hammer. Elsewhere,
Manuscript received 11/13/89; accepted for publication 5/29/90. drop-weight [3] and instrumented impact tests [4] are conducted
~Ralph L. Tobler is Metallurgist, and Richard P. Reed is Cryogenic at 4 K. This paper considers only conventional pendulum-type
Materials Group Leader, National Institute of Science and Technology, tests of standard V-notched specimens, since little information
Materials Reliability Division, Boulder, CO 80303.
zI1 Soon Hwang is Researcher, and Martin M. Morra is Graduate on other methods is available.
Student, Department of Materials Science, Massachusetts Institute of
Technology, Cambridge, MA 02139; Ronald G. Ballinger is Associate
Professor, Department of Nuclear Engineering, Massachusetts Institute Conventional Testing
of Technology, Cambridge, MA 02139. The conventional procedure for conducting a Charpy impact
3Hideo Nakajima is Researcher, and Susumu Shimamoto is Division
Chief, Superconducting Magnet Laboratory, Division of Thermonuclear test is given by A S T M Methods for Notched Bar Impact Testing
Fusion Research, Naka Fusion Research Establishment, Japan Atomic of Metallic Materials (E 23-88). A 10 by 10 by 55 mm notched
Energy Research Institute, lbaraki, 311-02, Japan. bar is cooled to within 1 deg of a specified temperature, then
0090-3973/91/0001-0034500.00
34
 TOBLER ET AL. ON CHARPY IMPACT TESTS NEAR ABSOLUTE ZERO 35

transferred to the anvil of a test machine and fractured within 5 as low as 77 K, were forced to abandon Charpy testing at 4 K
s. This procedure is satisfactory at temperatures down to 77 K, because "the problems of temperature control and scatter were
and it is often used to identify the fracture transition tempera- unacceptable" [10].
tures and lower shelf energies of ferritic steels. However, for Preliminary thermal analysis [6] predicts a rapid temperature
tough austenitic alloys below 77 K, it proves inadequate. Two rise above 4 K during direct transfer unless precautions are taken.
major problems are encountered: the specimen temperature rises To confirm this expectation, Dobson and Johnson [7] recorded
during transfer to the anvil, and the specimen heats severely as the thermal histories of AISI 304 test specimens with embedded
it fractures at high strain rates. The large data scatter of replicate thermocouples. Using precooled tongs, they transferred bare
tests at cryogenic temperatures is widely attributed to these two specimens from a helium Dewar through ambient air as quickly
factors. as possible, but the specimen temperatures increased from 4 to
67 K or more before the specimens could be fractured. Yoshi-
mura et al. [11] did better by covering their specimens during
Cryogenic Testing
transfer; still, their specimens heated from 4 to about 15 K before
The most common cryogenic test temperatures are the fixed the pendulum's strike.
bath temperatures of the liquid cryogens: 77 K (nitrogen), 20 K The precooling temperature must be accurate to within 1 K
(hydrogen), and 4 K (helium). One way to verify the adequacy according to A S T M E 23. Therefore the following efforts have
of impact tests at these temperatures is to attach thermocouples been made to insulate the specimen (and sometimes the entire test
to calibration specimens and record the temperature history dur- machine) to achieve a fixed initial test temperature near 4 K.
ing each stage of testing.
Thermocouples may be located at the specimen surface or
implanted in the center of the thickness by means of drilled holes.
Boating, Boxing, and Encapsulation
Thermocouples of Chromel-versus-Gold + 0.07% iron are pre- A paper boat method was developed at Ohio State University
ferred for maximum sensitivity below 20 K. Results for specimens for tests at 20 K [6]. As described by Zambro and Fontana [5],
initially cooled to 77 K show that there is no appreciable tem- a standard specimen is glued to the bottom of a paper boat which
perature rise within the 5 s allotted to transfer and break the has a perforated cover. After it cools in a Dewar of liquid hy-
specimen per A S T M E 23 [5]. On the other hand, for specimens drogen, the boat containing the specimen is lifted by threads and
initially cooled to 20 or 4 K, the temperature rises quite rapidly transferred through air to the anvil. Once the boat docks on the
immediately after removal from the Dewar flask [6, 7]. Conse- anvil, the threads are cut and the pendulum unleashed. Thermo-
quently, tests at 77 K or above can be conducted as specified by couple responses show that this method is successful in that the
ASTM E 23, but tests at 20 K or below require other techniques paper container is able to retain sufficient liquid hydrogen to
which tend to compromise the simplicity of the test to varying forestall heating during specimen transfer.
degrees. A variation of boating also has been used for tests at 5 or 6
To avoid the complications of testing at more extreme tem- K, as described by Long [12]. According to this technique, the
peratures, it is tempting to use data acquired at 77 K to make specimen in a paper boat is dipped first in liquid hydrogen, then
decisions about 4 K applications. Unfortunately, that approach in liquid helium to freeze the hydrogen around the specimen.
is unreliable. Mazandarany et al. [8], for an AISI 316LN base The solid-hydrogen-encased specimen is then transferred and
metal and its welds, compared C, measurements at 77 K with J~ tested. The procedure is timed using thermocouples to identify
measurements at 4 K but did not find a good correlation. Thus the proper moment for pendulum release to obtain a 5 to 6 K
the problems of fracture characterization at 4 K cannot be re- initial test temperature.
solved simplisticly by Charpy tests at higher temperatures. This Although boating permits tests at 20 K or lower, hydrogen is
should be expected, because the deformation mechanism hazardous and special precautions are necessary for safety.
changes between 77 and 4 K. Typically, at 4 K, the tensile stress- Therefore Jin et al. [13] developed an alternative procedure using
strain curves at normal strain rates are serrated from discontin- liquid helium and a plastic box for insulating specimens tested
uous yielding (unstable plastic deformation), and they are strain- below 6 K. Open-celled, grooved polystyrene foam is taped
rate sensitive. A t 77 K and higher, the stress-strain curves for around a slightly undersized Charpy test specimen which is then
the same materials are smooth and continuous; there is no evi- encased in a thin-walled acrylic plastic box. Plastic inlet and
dence of discontinuous yielding, but the conventional process of outlet pipes are cemented to the box, and liquid helium from a
slow-stable, nearly isothermal deformation prevails instead [9]. storage dewar is continuously pumped through the system. A
In the quest for a fully satisfactory Charpy test procedure at temperature of 5 to 6 K is achieved in 60 to 80 s, and, while
4 K, three general approaches have been tried: direct transfer liquid helium is still flowing through the box, both the box and
of the specimen from the cryogen to the anvil of the test machine, the specimen are impacted.
encapsulation of the specimen, and encapsulation of entire test Ogata et al. [14,15] sought simpler methods of encapsulation.
machines. These are briefly reviewed in the following sections. They wrap their specimen in a 3 by 46 by 200 mm piece of grooved
polystyrene foam to create a pipe-like casing with inlet and outlet
ports at the ends for liquid helium. Longitudinal grooves on the
Direct Transfer
inside surface of the foam facilitate the flow of helium and its
Researchers have tried to apply ASTM E 23 at temperatures contact with the specimen. The casing is sealed with glue and
down to 4 K while making no concessions at all for the extreme wrapped in cellophane tape to prevent it from tearing as a helium
temperature. They conclude that the conventional procedure transfer line is inserted at one end. With the transfer line inserted
fails at 4 K. For example, Lee and Dew-Hughes, after obtaining and taped to the casing, the assembly is placed on the anvil and
Charpy data for stainless steel castings at several temperatures helium is continuously transferred. Typically, the specimen tern-
36 JOURNAL OF TESTING AND EVALUATION

perature drops to 5 K after 60 s and to 4.6 K after 170 s [15]. As indicated above, the energy to break a specimen-capsule
This method is perhaps the simplest and least expensive yet combination is greater than the sum of the energies required to
proposed. During continuous operation, each test consumes only break an empty capsule and a bare specimen separately. There-
1 L of helium. Little energy is expended to fracture the foam, fore tests of encapsulated or insulated specimens cannot be cal-
so this method can be used for both high and low toughness ibrated simply by subtracting the energy required to break an
materials. empty capsule from the total absorbed energy. Instead, a cor-
The glass Dewar is another technique used recently in Japan rection factor is deduced by comparing the results of tests per-
[16-18]. A specimen is placed in a small double-walled vacuum- formed on empty capsules with the results for capsules containing
insulated glass tube capable of retaining liquid helium for several specimens. The comparison is made at 273 K [16] or 77 K [17],
minutes. The tube containing the specimen and helium is then and the resulting correction factor is then adopted at 4 K, with
placed on the anvil of the impact machine and fractured. This an assumption that the temperature effect on the calibration
method ensures an initial test temperature of 4.2 K, but the C,, factor is negligible.
readings must be corrected for the extra energy absorbed in A second calibration technique is described by Mori and Ku-
breaking the specimen-container composite. Little energy is re- roda [18]. In Charpy testing, it is customary to measure both Cv
quired to fracture glass at 4 K, but a significant energy term and ~, and the two are approximately linearly related for certain
arises because there is a vacuum space between the specimen conditions. Unlike C~, t~ is not affected by the presence of in-
and the anvil. During fracture, glass shards are haphazardly in- sulators, and this provides the rationale for the second calibration
terposed between the specimen and the anvil, causing friction procedure: if it is assumed that d(Cv)/dt~ at 4 K equals d(C~)/dd~
[18]. The correction factor proves to be rather large and some- at 77 K, then the value of ~ at 4 K gives, by correlation of 77
what variable (20 to 35 J). Consequently this method has limi- K data, the value of Cv at 4 K [18].
tations and cannot be recommended for alloys with lower tough- Table 1 lists typical calibration factors for the various test
ness where the magnitude of Cv approaches the magnitude of methods. The correction factor is lowest for the paper boat tech-
the calibration factor. nique and highest for the glass Dewar technique. For this reason,
In older work, the impact test machines as well as specimens the glass Dewar method is not recommended for low toughness
were encapsulated. DeSisto [19] built a sheet-metal structure to alloys.
house his impact machine, a specimen storage drum, a cold box, Table 2 compares, in summary form, the various approaches
and a cooling mechanism. With the chamber evacuated, a feed to extreme cryogenic testing that are referenced in this review.
mechanism would automatically insert cold specimens to the test The minimum achievable test temperatures are listed, along with
machine anvil. Initial test temperatures were as low as 8 K. Kiefer the principal advantages and disadvantages that can be cited for
et al. [6] also enclosed their machine in a chamber complete with each technique. The cryogenic apparatus and techniques are de-
a glove-box and viewing ports. After evacuating the air, they scribed and illustrated in the literature, but none is specified or
filled the chamber with helium gas to prevent moisture conden- referenced in national or international standards. Each method
sation during transfer. Within 2 s, Charpy specimens could be has specific advantages and disadvantages, but none predomi-
manually transferred from an open-mouth hydrogen Dewar nates in current practice.
within the enclosure and tested at temperatures as low as 25 K.
Both of these approaches achieve freedom from atmospheric
Adiabatic Heating
condensation, a major factor that raises the specimen temper-
ature during transfer. This advantage is outweighed by the com- Whereas specimen heating during transfer can be eliminated
plexity of the apparatus required, and, to our knowledge, these by boating, boxing, or encapsulation, the heat generated within
techniques have not been used by anyone other than the origi- the specimen during high strain-rate deformation and fracture
nators. cannot be avoided. Adiabatic heating is affected by material and
test variables, including the total quantity of plastic work con-
Calibration sumed in fracturing the specimen, the rate of straining, the spe-
cific heat of the test material, and the cooling efficiency of the
Any test involving boated, boxed, or encapsulated specimens medium. Austenitic steels are particularly susceptible to adi-
requires a correction factor to account for the presence of in- abatic heating because typically they are very tough and have
sulation and its effect on the energy absorbed in fracture. There low specific heats. In fact, austenitic steels heat perceptively
are two methods of calibration, as discussed in the next section. during conventional tensile tests at room temperature [20]. The
In both cases, assumptions are involved in arriving at the cor- problem is much worse, however, at impact strain rates, and still
rection factor. worse at cryogenic test temperatures where the specific heats of

TABLE 1--Typical calibration correctionfactors for nonstandard 4 K tests using various

specimen insulating containers.
Calibration Correction
Container J (ft . lbs) Reference

Paper boat 0.27 (0.2) 5

Foam wrapper 1.5 (1.1) 14
Plastic box 4.9 (3.6) 13
Glass Dewar 20 to 35 (14.7 to 25.8) 18
 TOBLER ET AL. ON GHARPY IMPACT TESTS NEAR ABSOLUTE ZERO 37

TABLE 2--Comparison of Charpy test procedures used to evaluate material behavior at cryogenic temperatures.

Minimum Initial
Method Test Temperature,
[Reference] Cryogen K Advantage Disadvantage

Pneumatic machines He 4 Precise initial temperature Not a pendulum test

[1,2]
LN, direct transfer N 77 Low cost, standard procedure Unreliable for predicting alloy
[8] performance at 4 K
LH, direct transfer H >20 Hazardous fluid; lack of temperature
I5] control
LHe, direct transfer He 20-60 Non-hazardous fluid Lack of temperature control
[7,10,11]
LH boat H 20 Simple paper insulation Hazardous fluid; H facilities are required
[51
Frozen H boat H and He 4 Simple paper insulation Hazardous fluid; H facilities are required
[121
Plastic box He 5 Low cost; steady temperature; no Preparation time; significant calibration
[13] temperature gradient correction
Foam wrapping He 5 Low He consumption; negligible Preparation time
[14,15] calibration correction factor
Glass Dewar He 4.2 Precise initial temperature Preparation time; large calibration
[16,17,18] correction
Glove box H 25 Specimen transfer in an inert Hazardous fluid; elaborate apparatus
[6] environment
Encapsulated machine He 8 Transfer in vacuum, free from Elaborate apparatus; costs; low rate of
[191 condensation; automatic feed; variable testing
test temperature >- 8 K

these materials decrease by several orders of magnitude for a shown in Fig. 1. From Mori and Kuroda [18], for austenitic
temperature change between 298 and 4 K, and approach zero at stainless steels, we have
zero K.
Heating during fracture can be studied analytically or exper- +(mm) = 0.0167 C,(J) (2)
imentally. Both approaches indicate a huge departure from the
initial'test temperature for specimens that are precooled to 4 K.
For AISI 304 steel upon impact, Dobson and Johnson [7] The volume of the plastic zone for a specimen with a 10 by 10
recorded temperatures as high as 150 K, and this temperature mm cross section and a 2 mm deep notch is
is defined by some as the upper limit of the cryogenic regime.
Their observations are supported in this paper by the following Vp(mm3) = a 2 ~ (10 mm) (8 mm) = 2.67 a C,.(J) (3)
thermal analysis based on the conversion of plastic work to heat.
An energy balance between the mechanically generated heat where a proportionality constant, et, is introduced for experi-
and the specimen temperature rise can be written: mental calibration. (If the plastic zone width is equal to 2~, then
et = 1.) The thermal diffusion distance during the impact is small
9 enough to be neglected, so the energy balance reduces to
0.9 C~ = pV~ CaT (1)
r~
0.337 (mm3) f9
where pet = r, CedT (4)
Cv = Charpy absorbed energy,
p = material density, Finally, for austenitic stainless steel, the specific heat [21] for
Vp = volume of plastic zone, temperatures up to 400 K is given by
C~ = specific heat,
T, = temperature before impact, and
Cp[J/(kg • K)] = - 3.9056 + 0.15424T + 5.3141
Tr = maximum temperature after impact.
In Eq I we assume that 90% of the absorbed energy is converted × 10-2T2 - 3.8490 × 10 4T~ + 1.0382 x 10-6T4 - 9.7572
into heat. The volume of the plastically deformed zone can be
estimated from the measured value of lateral expansion, ~, as × 1 0 ' ° T 5 (5)
38 JOURNALOF TESTING AND EVALUATION

Notched Bar Impact Specimen

(
400
(55 x 10 x 10 mm)

300

"\ Nctch, 2 mm deep

I.
~D
e.,
200
E Stainless Steel
[.-
e~
• Dobson& Johnson
lOt) ........ alpha=l
l>J.~ Specimen
/ ~ Halve~~ .... alpha=0.5
t~
- - - alpha=2
Notch -~ ! • I

2 mm deep !00 200 300

initial Temperature (K)

FIG. 2--Predicted temperature rise during a Charpy test as a function

of initial temperature for austenitic stainless steel.

4OO

FIG, 1--Schematic diagram of fractured Charpy specimen showing 300 .~...-: ...-"
lateral expansion at the fracture plane. We assume the lateral expansion s

]
t~ is proportional to the volume of the plastically deformed zone Vp.
200 ,..."

Figure 2 shows the predicted final temperature from Eq 4 as [-, alpha=l

a function of initial temperature. (The data of Dobson and John-
100
"
S.S.
son [71 are best predicted with c~ = 1 at low temperature, as e-

shown in this figure.) Clearly, no significant difference in final

temperature is predicted for initial temperatures of 4 or 77 K. AI
Hence we expect that the absorbed energy for a given steel must 0-1 • i ! •

be very similar despite the two different initial test temperatures. 0 100 2O0 3OO
Matsumoto et al. [16], who tested austenitic stainless steel welds Initial Temperature (K)
at 273, 77, and 4.2 K, report data trends that appear to support
this expectation. FIG. 3--Predicted temperature rise during Charpy tests of copper and
We also made predictions for copper and aluminum (highly aluminum compared with that of austenitic stainless steel.
conductive metals) using the appropriate experimentally derived
specific heat correlations [22] in Eq 4. As in the case of stainless temperature, whereas data for face-centered-cubic (fcc) metals
steel, we expect that the Co data for initial test temperatures of show a gradual increase or decrease of impact energy with test
4 or 77 K should be similar, since the final temperature is nearly temperature. Ferritic steels are representative of the bcc metals,
independent of the initial temperature, as shown by the results
whereas austenitic steels are representative of the fcc metals.
in Fig. 3. Verification of this model is necessary and can be Usually, the gradual trends for austenitic steels extend into the
achieved by thermocouple-instrumented Charpy tests at select
cryogenic range. For example, high Mn austenitic steel [23]
temperatures. However, both our thermal analysis and the ex-
showed a gradual decrease of Cv at five temperatures between
perimental data available so far support the present view that, 295 and 4 K. As stated in ASTM E 23, the Charpy test is practical
because of adiabatic heating, the 4 K Charpy test has little sig-
for ferritic steels because it enables the ductile-to-brittle transi-
nificance.
tion temperature to be determined, and it is less appropriate for
austenitic alloys or other fcc metals that do not exhibit transi-
Discussion tional behavior.

Data Trends Data Scatter

Typically, Charpy impact energies for body-centered-cubic Data scatter has long been a problem in Charpy tests, and this
(bcc) metals over a range of test temperatures show an abrupt problem is magnified when the tests are performed at temper-
decrease of energy at the ductile-to-brittle fracture transition atures approaching absolute zero. There are several reasons.
 TOBLER ET AL. ON CHARPY IMPACT TESTS NEAR ABSOLUTE ZERO 39

First, a bare specimen warms continuously as it is being trans- or experience in view, and as emphasized in this paper, it is
ferred, so the temperature at the time of testing varies with the actually impossible to perform valid Charpy tests at 4 K according
transfer time. Second, the scatter at extreme cryogenic temper- to the procedure presently outlined in A S T M E 23 and as re-
atures increases relative to the level observed at room temper- quired by regulatory agencies.
ature because the effects of metallurgical and mechanical test Based on the facts available at this time, it is appropriate to
variables are stronger at cryogenic temperatures. Third, the scat- revise A S T M E 23 in regard to cryogenic testing. The revision
ter in Charpy measurements at 4 K will be increased by any should restrict the temperature for cryogenic testing to a lower
uncertainty associated with the required calibration correction limit of 77 K or else specify some appropriate standard tech-
factors. nique(s) for testing at temperatures below 77 K.

Data Correlations
References
In a large superconducting machinery development program
sponsored by the Japan Atomic Energy Research Institute [1] Novikov, N. V., "Mechanical Property Measurement Techniques
of Structural Materials at Cryogenic Temperatures," Advances in
(JAERI), C, data were used to screen new cryogenic steels [24- Cryogenic Engineering, Vol. 22, 1977, pp. 113-1t8.
28]. Ninety alloy compositions were Charpy tested at 4 K, using [2] Verkin, B. I., "The Development of Cryogenic Materials Science
the glass Dewar method. Alloys having C~ greater than 80 J were in the USSR," Advances in Cryogenic Engineering, Vol. 26, 1980,
then produced in greater quantities and evaluated using J~ tests pp. 37-47.
[3] Nishijima, S., Takeno, M., and Okada, T., "Impact Tests of Rein-
at 4 K. When the J~e measurements became available, the C~ forced Plastics at Low Temperatures," Advances in Cryogenic En-
data and K~¢ estimates from J~ were compared. In general, C~ gineering, Vol. 28, 1981, pp. 261-270.
increased with J~ or the K~ estimates, but the data for individual [4] Yoshida, H., Kozuka, T., Miyata, K., and Kodaka, H., "Instru-
alloys did not show a perfect 1-to-1 correlation [27,28]. mented Charpy Impact Tests at Low Temperatures," in Austenitic
The J A E R I correlation was limited to certain high toughness Steels" at Low Temperatures, R. P. Reed and T. Horiuchi, Eds.,
Plenum, New York, 1983, pp. 349-354.
austenitic steels because a 4 K toughness (Kjc) of 200 MPa. m ~ [5] Zambrow, J. L. and Fontana, M. G., "Mechanical Properties, In-
had been specifically targeted for alloy development. In future cluding Fatigue, of Aircraft Alloys at Very Low Temperatures,"
work we plan to extend the correlation to low toughness alloys Transactions ofASM, Vol. 41, 1949, pp. 480-518.
as well. There are good reasons to expect that C, should not [6] Kiefer, T. F., Keys, R. D., and Schwartzberg, F. R., "Charpy.
Impact Testing at 20 K," Advances in Cryogenic Engineering, Vol.
correlate 1-to-1 with J~ or K,~ at 4 K. First, the two tests do not I0, 1965, pp. 56-62.
measure exactly the same material properties: the Charpy test [7] Dobson, W. G. and Johnson, D. L., "Effect of Strain Rate on
involves fracture initiation from a notch as well as propagation Measured Mechanical Properties of Stainless Steel at 4 K," Ad-
from a notch, whereas the K~, and J,~ parameters pertain to ini- vances in Cryogenic Engineering, Vot. 30, 1984, pp. 185-192.
tiation only. Second, the two tests are conducted at different [8] Mazandarany, F. N., Parker, D. M., Koenig, R. F., and Read, D.
T., "A Nitrogen-Strengthened Austenitic Stainless Steel for Cry-
strain rates, and the discontinuous yielding process at 4 K is ogenic Magnet Structures," Advances in Cryogenic Engineering,
strain-rate sensitive. Especially at strain rates higher than 10 -3 Vot. 26, 1980, pp. 158-169.
which are encountered in impact tests, the discontinuous yielding [9] Reed, R. P. and Walsh, R. P., "Tensile Strain-Rate Effects in
and adiabatic heating behaviors for austenitic steel change Liquid Helium," Advances in Cryogenic Engineering, Vol. 34, 1988,
pp. 199-208.
abruptly [9]. Third, there could be a specimen size effect since [10] Lee, K. S. and Dew-Hughes, D., "The Effect of ~-ferrite Upon the
the Charpy specimen has a smaller ligament (8 by 10 ram) com- Low Temperature Mechanical Properties of Centrifugally Cast
pared to the typical compact specimen (15 by 25 mm) used in Stainless Steels," in Austenitic Steels at Low Temperatures, R. P.
Kk or J,~ tests. Reed and T. Horiuchi, Eds., Plenum, New York, 1983, pp. 221-
Impact tests are of questionable value at extreme cryogenic 242.
[11] Yoshimura, H., Shimizu, T., Yada, H., and Kitajima, K., Tetsu-
temperatures. Soon, an inexpensive alternative screening test to-Hagane, Vol. 65, 1979, pp. 681-686 (in Japanese).
may be in demand. The above considerations suggest that, for [12] Long, H. M., comment in Advances in Cryogenic Engineering, Vol.
austenitic alloys at 4 K, slow-bend tests of fatigue-cracked Charpy 19, 1974, p. 378.
specimens might offer better prospects for correlation with J~ [13] Jin, S., Horwood, W. A., Morris, J. W., Jr., and Zackay, V. F.,
"A Simplified Method for Charpy Impact Testing Below 6 K,"
values. Witzke and Stephens [29] used slow bend tests for an Advances in Cryogenic Engineering, Vol. 19, 1974, pp. 373-378.
alloy development program at 77 K, but as yet there is no ex- [14] Ogata, T., Hiraga, K., Nagai, K., and Ishikawa, K., "A Simplified
perience with slow bend tests at 4 K. Method for Charpy Impact Testing Near Liquid Helium Temper-
ature," Cryogenics, Vol. 22, 1982, pp. 481-482.
[151 Ogata, T., Hiraga, K., Nagai, K., and Ishikawa, K., "A Simple
Conclusion Method for Charpy Impact Test at Liquid Helium Temperature,"
Transactions of the National Research Institute for Metals, Vol. 26,
Until now, the limitations of Charpy testing at liquid helium 1984, pp. 238-242.
temperature have not been fully recognized beyond a relatively [16] Takahashi, Y., Yoshida, K., Shimada, M., Tada, E., Miura, R.,
small community of cryogenic researchers. For example, regu- and Shimamom, S., "Mechanical Evaluation of Nitrogen Strength-
ened Stainless Steels at 4 K," Advances in Cryogenic Engineering,
latory agencies today still require that liquid helium storage tanks Vol. 28, 1982, pp. 73-81.
be designed according to the A S M E Boiler Code which requires [17] Matsumoto, T., Satoh, H., Wadayama, Y., and Hataya, F., "Me-
Charpy tests of materials of construction at 4 K. Although the chanical Properties of Fully Austenitic Weld Deposits for Cryogenic
meaning of Charpy data at liquid helium temperature is ques- Structures," Welding Research Supplement, Vol. 66, 1987, pp. 120-
126s.
tionable, the tests are nevertheless conducted because the code
[18] Mori, T. and Kuroda, T., "PredictionofEnergyAbsorbedin Impact
mandates them. However, neither the code nor the notched bar for Austenitic Weld Metals at 4.2 K," Cryogenics, Vol. 25, 1985,
impact test standard was originally written with 4 K applications pp. 243-248.
40 JOURNALOF TESTING AND EVALUATION

119] DeSisto, T. S., "Automatic Impact Testing to 8 K," Technical Re- for Cryogenic Use," Advances in Cryogenic Engineering, Vol. 30,
port 112/93, Watertown Arsenal Laboratories, Waterton, Mass., 1984, pp. 169-176.
1958. [26] Suemune, K., Sushino, K., Masumoto, H., Nakajima, H., and Shi-
[20] Ferron, G., "Influence of Heat Generation and Conduction on mamoto, S., "Improvement of Toughness of a High-Strength, High-
Plastic Stability under Uniaxial Tension," Materials Science and Manganese Stainless Steel for Cryogenic Use," Advances in Cry-
Engineering, Vol. 49, 1981, pp. 241-248. ogenic Engineering, Vol. 32, I986, pp. 51-56.
[21] Mann, D., Ed., LNG Materials & Fluids Handbook, National Bu- [27] Tone, S., Ogawa, M., Yamaga, M., Kaji, H., Horiuchi, T., Ka-
reau of Standards, Boulder, Colo., 1977. samatsu, Y., Nakajima, H., Yoshida, K., Takahashi, Y., and Shi-
[22] Hands, B. A., Ed., Cryogenic Engineering, Academic Press, New mada, M., "Preliminary Study on Structural Material Selection for
York, t986, p. 95. Large Superconducting Magnets," in Austenitic Steels at Low Tem-
[23] Miura, R., Ohnishi, K., Nakajima, I-I.,Takahashi, Y., and Yoshida, peratures, R. P. Reed and T. Horiuchi, Eds., Plenum, New York,
K., "Mechanical Properties of 18Mn-5Cr Austenitic Steel at Cry- 1983, pp. 263-275.
ogenic Temperatures," in Austenitic Steels at Low Temperatures, [28] Nakajima, H., Yoshida, K., Okuno, K., Oshikiri, M., Tada, E.,
R. P. Reed and T. Horiuchi, Eds., Plenum, New York, 1983, pp. Shimamoto, S., Miura, R., Shimada, M., Tone, S., Suemune, K.,
287-293. Sakamoto, T., and Nohara, K., "Fracture Toughness of Newly
[24] Tone, S., Shimada, M., Horiuchi, T., Kasamatsu, Y., Nakajima, Developed Structural Materials for Superconducting Coils of Fusion
H., and Shimamoto, S., "The Development of a Nitrogen-Strength- Experimental Reactor," Advances in Cryogenic Engineering, Vol.
ened High-Manganese Austenitic Stainless Steel for a Large Su- 32, 1986, pp. 347-354.
perconducting Magnet," Advances in Cryogenic Engineering, Vol. [29] Witzke, W. R. and Stephens, J. R., "Development of Strong and
30, 1984, pp. 145-152. Tough Cryogenic Fe-12Ni Alloys Containing Reactive Metal Ad-
[25] Masumoto, H., Suemune, K., Nakajima, H., and Shimamoto, S., ditions," Cryogenics, Vol. 17, 1977, pp. 681-688.
"Development of a High-Strength High-Manganese Stainless Steel