Heat treatments were utilized in 5Ni and 9Ni steel which resulted in the development of
tempered microstructures which contained either no measurable retained austenite (<0.5
pct) or approximately 4 to 5 pct retained austenite as determined by X-ray diffraction.
Microstructural observations coupled with the results of tensile testing indicated that the
formation of retained austenite correlated with a decrease in carbon content of the matrix.
Relative values of K t c at 77 K were estimated from slow bend precracked Charpy data using
both the COD and equivalent energy measurements. In addition, Charpy impact properties
at 77 K were determined. In the 9Ni alloy, optimum fracture toughness was achieved in
specimens which contained retained austenite. This was attributed to changes in yield and
work hardening behavior which accompanied the microstructural changes. In the 5Ni alloy,
fracture toughness equivalent to that observed in the 9Ni alloy was developed in grain
refined and tempered microstructures containing <0.5 pct retained austenite. A decrease in
fracture toughness was observed in grain refined 5Ni specimens containing 3.8 pct retained
austenite due to the premature onset of unstable cracking. This was attributed to the
transformation of retained austenite to brittle martensite during deformation. It was
concluded that the formation of thermally stable retained austenite is beneficial to the
fracture toughness of Ni steels at 77 K as a result of austenite gettering carbon from the
matrix during tempering. However, it was also concluded that the mechanical stability of the
retained austenite is critical in achieving a favorable enhancement of cryogenic fracture
toughness properties.
THE need for materials to contain liquefied gases has ductility, Charpy impact toughness, and the DBTT.
steadily increased in recent years. Low carbon ferritic Schwartz 8 has suggested that low temperature toughness
nickel alloy steels have been developed for use in increases as a function of retained austenite in 9Ni steel
cryogenic structural applications with costs being min- due to a TRIP mechanism operative at the crack front.
imized by selecting the appropriate nickel content for Sarno, et al ~ concluded that retained austenite con-
the intended service temperature range. In general, 9Ni tributes to toughness in 5Ni steel by depleting the
steel has received wide acceptance for liquefied natural matrix of interstitials and by simply increasing the
gas and nitrogen tankage due to its relative ease of amount of fcc phase which possesses higher resistance
production and good weldability. Recently, ~,25Ni steel to cleavage fracture.
has been proposed as a lower cost replacement for 9Ni Other results have been reported which are somewhat
steel in such applications and therefore is an alloy of contradictory. Pampillo and Paxton 9 reported a bene-
considerable interest. ficial effect of retained austenite on tensile ductility but
Austenite retention after heat treatment has been no effect on impact properties in maraging steels.
considered to play a key role in achieving low temper- Hwang, et al 7 reported that retained austenite improved
ature toughness in Fe-Ni alloys since the development tensile elongation and Charpy impact energy at cryo-
of low carbon 8 to 10 pct nickel steels for cryogenic genic temperatures in fine grained 12Ni titanium
service by Brophy and Miller? Marshall, et a[ 4 sug- gettered alloys but produced a decrease in K t c . Samples
gested that retained austenite increased low temperature containing retained austenite showed increased suscep-
toughness in 9Ni steel by scavenging carbon and other tibility to unstable crack propagation. Finally, Ooka, et
deleterious elements from the ferritic matrix. More allo found no quantitative relationship between retained
recently, Norstrom 5 found that cementite precipitation austenite and low temperature toughness.
was eliminated as a result of retained austenite for- It is evident from this brief review that the influence
mation in 9Ni steel and that this produced a cor- of retained austenite on the material toughness meas-
responding decrease in the DBTT. A mechanism other ured by various methods is complex and not well
than a scavenging effect seems to be supported by understood. In addition, in many previous studies the
recent reports 6,7 in low carbon, titanium gettered Fe-Ni effect of reduced strength due to tempering of the
alloys. The presence of retained austenite in such martensitic structure has been neglected in interpreting
systems was shown to have beneficial effects on tensile the role of retained austenite. A program was instituted
in our laboratory to define the microstructural mech-
J. R. STRIFE, formerly with Union Carbide Corporation, Tarry- anism responsible for high toughness in Ni steels by
town, NY, is now Research Scientist at United Technologies Research first considering quench and temper treatments in both
Center, East Hartford, CT 06108. D. E. PASSOJA is Senior Research
5Ni and 9Ni steel, and then by considering the more
Scientist, Union Carbide Corporation, Central Scientific Laboratory,
Tarrytown, NY 10591. complicated 3-step heat treatments in 5Ni steel. In
Manuscript submitted July 23, 1979. particular, an attempt was made to clarify the role of
ISSN 0360-2133/80/0811-1341500.75/0
METALLURGICAL TRANSACTIONS A 9 1980 AMERICAN SOCIETY FOR METALS AND VOLUME 11A, AUGUST 1980--1341
THE METALLURGICAL SOCIETY OF AIME
martensite tempering in conjunction with retained aus- log tempering time. The amount of retained austenite
tenite formation in both alloys. observed after heat treating exhibits a maximum as a
function of time. The time at which the amount of
retained austenite is maximized at a given tempering
EXPERIMENTAL PROCEDURE
temperature decreases as the tempering temperature
increases. This result can be understood by considering
Materials
that the initial austenite which forms will be rich in
Conventionally heat treated 5Ni and 9Ni plate ap- alloy content. As the time at the tempering temperature
proximately 13 mm in thickness was received from the increases, austenite continues to grow but the solute
A R M C O Steel Corporation. The chemical compositions redistributes itself over a larger volume of austenite.
as determined in our laboratory are listed in Table I. Thus, the alloy content of the austenite at the tempering
temperature decreases causing the stability of the aus-
tenite to decrease as the time at temperature increases.
Microstructural Characterization
At some point, the austenite stability is reduced to a
Specimens to be investigated optically were mounted level which, when cooled to low temperature, will
in epoxy and polished using standard metallographic readily transform to martensite, resulting in a decrease
techniques. Specimens to be used for observation by in the volume of austenite retained at low temperature.
transmission electron microscopy were prepared from As the tempering temperature increases, austenite
sections cut parallel to the fracture surface of Charpy growth increases concomitantly, and the austenite sta-
specimens. Disks of 3 mm diam were spark machined bility is reduced for shorter aging times. Tempering
from the sections and chemically thinned in a dual temperatures of 863 and 813 K were selected for the 9Ni
submerged jet polisher using a chromic-acetic electro- alloy from these data to produce tempered martensitic
lyte. The thin foils were examined in a JEOL electron microstructures with and without retained austenite.
microscope operated at 200 KV. Time at temperature was limited to the linear portions
Quantitative measurement of the amount of retained of these curves to prevent formation of untempered
austenite was determined by X-ray diffraction methods. martensite on quenching.
The peak pair technique of Durnin and RidaP ~was The heat treatment for 5Ni steel as developed by
adopted and a standard procedure using a copper X-ray A R M C O L2 is a relatively complex three-step treatment
tube and graphite monochromator was implemented. which involves !) austenitizing and water quenching,
Samples for X-ray measurement were polished metallo- 2) reheating into the A c,-Ac, range to form a lamellar
graphically and deeply etched to remove the disturbed structure and water quenching, 3) reheating to just
metal resulting from polishing. In general, three deter- above the A c, and air cooling or quenching. The final
minations of retained austenite content were made for microstructure is composed of ferrite, tempered marten-
each heat treatment. The relative integrated areas under
the austenite peaks indicated that any texture effects
were minimal. The accuracy of this technique was
estimated as + 0.5 pct.
10
9
Heat Treatment uJ 893K
I---
8
To establish heat treating parameters, the critical I,g
I-r 7
temperatures were determined from dilatometric meas-
/ / 1 3 K
urements on specimens which had been austenitized at rgJ 6
1173 K and water quenched. The A c, and A c, temper- z_ 5
atures were defined as the first deviations from linear <
I-
thermal expansion behavior. These were determined as 4
923 and 1073 K for the 5Ni steel, and 863 and 1013 K
for the 9Ni steel.
In the 9Ni alloy, austenitizing was performed at
1116 K to remove the effects of prior heat treatment
3
2 /
and to produce a martensite packet size of approxi- o I | I I
o lO 10 2 10 3 10 4 10 5
mately 10 ttm. A series of tempering treatments was TEMPERING TIME [MINUTES]
implemented in order to alter the formation of retained Fig. 1--Pet retained aus,!.enite v s tempering time for 9 pet Ni steel
austenite. The data obtained from these treatments are quenched from 1116 K and tempered at the indicated temperatures
plotted in Fig. 1 as percent retained austenite at 77 K v s followed by quenching to 77 K.
Nominal Ni C N Mn Mo Si AI S P Fe
5Ni 5.05 0.085 0.017 0.52 0.28 0.21 0.046 0.009 0.022 bal
9Ni 9.10 0.084 0.035 0.54 0.065 0,19 0.038 0.011 0.026 bal
microscopy results suggest that retained austenite acted tion. Tempering treatments at 923 K which produced
as a sink for carbon in the 9Ni alloy. retained austenite were also investigated. It was dis-
In the 5Ni steel, tempering at 873 K (<0.5 pct covered, however, that the austenite which formed at
retained austenite) produced a microstructure which the tempering temperature was relatively unstable, so
was indistinguishable from the 9Ni alloy tempered at that a maximum of 2 pct was retained on quenching in
813 K. An optical micrograph of a polished and etched liquid nitrogen. An optical micrograph for a sample
specimen, Fig. 4(a), shows extensive carbide precipi- tempered 6 h at 923 K, Fig. 4(b), illustrates the fact that
tation both at the packet boundaries and in grain austenite present at the tempering temperature effec-
interiors. Observations in TEM showed that the as- tively gettered carbon from the ferritic matrix. Note
quenched martensitic structure was identical to that in that no carbide precipitation is observed in the grain
9Ni and that tempering at 873 K resulted in partial interiors. The packet boundaries are covered with
dislocation recovery and extensive carbide precipita- patches of untempered martensite transformed on
quenching from the tempering temperature. treatments both produced a very refined microstructure
As was previously discussed in the experimental which appeared partially lamellar and partially
procedure section, a heat treatment of 1 h at 1158 K, equiaxed. The average grain size determined from TEM
water quench, followed by 1 h at 1008 K, water quench, micrographs of these microstructures, Fig. 6, was less
produced a refined lamellar grain structure in the 5Ni than 1/~m. This represents an order of magnitude
alloy. After this treatment, the structure consisted of reduction in grain size when compared to the quenched
approximately 50 pct tempered ferrite and 50 pct fresh and tempered microstructure, Fig. 4(a), since the effec-
martensite. As a third step in heat treatment, specimens tive grain size in the quenched and tempered micro-
were reheated for 1 h at 873 K to produce a tempered structure is the martensite packet size which was
structure, and 1 h at 923 K for development of retained determined to be 8/~m.
austenite in combination with tempering. X-ray meas- The microstructural features associated with the three
urement showed that these treatments produced %0.5 step treatments were also investigated using TEM.
pct and 3.8 pct retained austenite respectively. Obser- Shown in Fig. 6(a) is a representative micrograph for a
vations in light microscopy, Fig. 5, showed that these specimen tempered 1 h at 873 K as the third step.
Fig. ~-Transmission electron micrographs of the structure observed Fig. 7--Transmission electron micrographs of the 5 pct Ni steel
in 5 pct Ni steel for three step heat treatments 9 (a) 1 h at 873 K for the showing retained austenite resulting from a three step heat treatment
third step. Carbide precipitation designated by arrows. Grain bound- utilizing 1 h at 923 K for the third step. (a) Bright field micrograph.
ary precipitate consisting of elongated carbide and possibly retained (b) Dark field micrograph using austenite diffraction spot as the
austenite at A. (b) 1 h at 923 K for the third step. Both elongated and operating reflection. Retained austenite is located at the ferrite
equiaxed grains are seen with little evidence of carbide precipitation9 boundaries9
Extensive carbide precipitation was observed as des- described previously. In the specimens tempered 1 h at
ignated by arrows on the micrograph. In addition, a 923 K, Fig. 6(b), very little carbide precipitation was
laminar precipitate was observed at some grain bound- observed, but retained austenite was readily docu-
ary locations as shown at position A on the micrograph. mented using darkfield techniques. Figure 7(a) is a
The morphology and contrast suggest that these are two brightfield micrograph and Fig. 7(b) the corresponding
phase precipitates consisting of elongated carbides darkfield micrograph using an austenite diffraction spot
coupled with films of retained austenite. Although as the operating reflection9 It can be seen that patches
confirming electron diffraction evidence was not ob- of austenite decorate the ferrite grain boundaries. Thus,
tained, the possibility that these are retained austenite it is apparent from a comparison of the microstructures
films must be considered since austenite distributed in associated with the three step treatments that retained
this manner at low volume fractions (~0.5 pct) would austenite formation has effectively gettered the carbon
go undetected using the X-ray diffraction techniques in the 5Ni alloy.
~ 1 HR. 863K. wq
retained austenite (Fig. 9). The net result was that the
9Ni specimens containing retained austenite exhibited
considerably more energy absorption as measured by
area under the load-displacement curve. These data are 10
listed in Table IV. The energies measured were cor- z