...

Kannan-Valencia2001-HighStrengthSteelCastings-Imp+

by user

on
Category: Documents
2

views

Report

Comments

Transcript

Kannan-Valencia2001-HighStrengthSteelCastings-Imp+
䉷ASM International
JMEPEG (2001) 10:635–648
Evaluation of High-Strength Steel Castings Possessing
Improved Weldability
K. Kannan and J.J. Valencia
(Submitted 16 April 2001)
Naval components fabricated from HY-80 high-strength steels require an expensive preheat during welding
to avoid heat-affected zone (HAZ) cracking. Quenched-and-tempered low-C, high-Ni steels were evaluated
as potential alternatives to HY-80 steel castings with section sizes of 230 to 300 mm thickness. The
investigation examined the feasibility of obtaining mechanical properties equivalent to HY-80 by heat
treatment and evaluated weldability. The steel resulted in a crack-free casting, and preliminary tests
suggest that it could be welded without preheating. Optimized heat treatment provided reasonably good
yield strength (517 to 538 MPa) and Charpy impact toughness (63 to 80 J Charpy V-notch (CVN) energy
at ⴚ73 ⴗC). The former properties were just below HY-80 casting requirements of 550 MPa. Thus, while
this composition might not be a suitable replacement for HY-80, there are other potential casting applications. These include surface ship shaft struts and rudder inserts that have less stringent strength and
toughness requirements.
Keywords Charpy V-notch energy, continuous cooling transformation diagrams, double tempering, hardenability, heat treatment, high- strength steel castings,
HSLA-100, HY-80, Jominy test, low- carbon
steels, optical microscopy, SEM and TEM characterization, Tekken test, tensile properties, welding without preheat
1. Introduction
Castings and forgings of high-strength steels, such as HY80 and HY-100, are used on naval surface ships and submarines.
Specific cast components include hull inserts, rudders, struts,
stern tubes, foundations, and valve bodies. However, there are
several manufacturing and fabrication problems with the castings and forgings including weldability and sensitivity to
heat treatment.
The relatively high carbon and alloy content of these steels
dictates a minimum preheating temperature of 107 ⬚C for sections greater than 28.6 mm thickness.[1] This is to prevent
hydrogen-assisted cracking (HAC) during welding. Furthermore, HY-80/100 castings are sensitive to heat-treatment parameters. Improper heat treatment during manufacture can result
in untempered brittle martensite, which could lead to cracking
even after the final tempering treatment.[2]
Traditional high-strength steels, such as HY-80, derive their
strength and low- temperature toughness from a quenched-andtempered martensitic/bainitic microstructure. Lower carbon
HSLA-80 steels typically have a ferritic microstructure, achieving their strength by precipitation hardening without significant
martensitic phase transformation.[3,4] This facilitates ease of
welding, avoiding the need for a preheat. However, the HSLA80 steel has been certified for naval applications as a replacement for HY-80 only as wrought plates[3] and not yet as castings.
K. Kannan and J.J. Valencia, Concurrent Technologies Corporation,
Johnstown, PA 15904. Contact e-mail: [email protected]
Journal of Materials Engineering and Performance
Churchill et al. investigated a casting with the HSLA-80 plate
composition under a National Shipbuilding Research Program
SP-7 project.[4] They found that a 150 mm cubic-shaped casting
exhibited poor Charpy V-notch (CVN) energies at low temperature (9 J at ⫺73 ⬚C). The MIL-23008D specifications for HY80 call for a minimum yield strength of 551 MPa and CVN
energy of 95 J at ⫺18 ⬚C and 68 J at ⫺73 ⬚C.[5] Thus, the
cast HSLA-80 casting could not meet the specified minimum
properties for HY-80 castings. Other studies examined castings
with a composition similar to the HSLA-100 plate as a replacement for HY-80 castings with section thickness of 300 mm.[6]
However, the mechanical properties were not satisfactory (low
toughness of 57 J at ⫺73 ⬚C), though better than the cast HSLA80 composition. Furthermore, there were noticeable macrocracks and hairline cracks, both in the as-cast and heat-treated
conditions. Thus, the HSLA-100 plate composition was not
found to be practical as a casting alloy for the thick sections
involved.
Also in the SP-7 study, Churchill et al. investigated a lowC (0.04 wt.%), high-Ni (5.5%) steel with 1.5% Cr and 0.5%
Mo that relied on a quench-and-temper heat treatment to achieve
the combination of strength and toughness.[4] The relatively
high Ni content, in combination with Cr and Mo, was intended
to limit the formation of ferrite and bainite as well as to decrease
the bainite-start temperature, hence producing a finer carbide
distribution within the bainitic microstructure. Both effects, in
turn, were expected to improve toughness. Upon heat treatment,
the casting was found to exhibit good tensile and upper-shelf
CVN impact energies. However, the low-temperature (⫺73 ⬚C)
impact properties in 300 mm thick test blocks were marginal
or barely above the specified minimum for HY-80.
The present study involves investigating the potential for
further optimization of alloy composition and heat treatment
of alloys similar to those investigated in the preceding SP-7
study (referred to hereafter as alloys E and E-A). The objective
of this work is to evaluate if modified alloy chemistries and
heat treatment can lead to properties similar to the certified HY80 composition, while being weldable with little or no preheat.
Volume 10(6) December 2001—635
Table 1 Chemical composition of alloys E and E-A (note
higher C, Mn and La, Ce in latter)
Element
C
Mn
Ni
Cr
Si
Mo
Ce
La
Cu
Al
V
Alloy E
(wt.%)
Alloy E-A
(wt.%)
Element
Alloy E
(wt.%)
Alloy E-A
(wt.%)
0.032
0.91
5.68
1.41
0.55
0.58
…
…
0.16
0.033
0.01
0.061
1.1
5.53
1.48
0.38
0.57
0.06
0.02
0.11
0.033
0.005
Zr
P
S
Nb
Ti
Sb
Sn
As
H
N
…
0.006
0.007
0.0001
0.002
0.001
0.001
0.003
0.002
1.3 ppm
36 ppm
…
0.01
0.006
⬍0.001
0.007
…
…
0.0078
…
…
43 ppm
…
2. Technical Approach
2.1 Casting Manufacturing
Castings of alloys E and E-A were procured in separate
heats from a commercial steel foundry. In both cases, the steels
were produced by the electric arc-furnace process and refined
in an argon oxygen decarburization vessel. Alloy E was cast
into a test block with dimensions 300 ⫻ 300 ⫻ 530 mm. Prior
to the removal of the risers and gates, the block was given a
proprietary homogenization and tempering treatment. Further
heat treatment (austenitization and tempering) was performed
in-house on thin subsections of the casting to identify optimal
heat-treatment conditions, as detailed in a later section.
Alloy E-A was melted and cast in a second heat and had a
slightly modified chemistry. This composition was designed to
have a somewhat higher C content (0.06%, up from 0.035%)
and Mn content (1.1%, up from 0.9%) than alloy E. The higher
C and Mn contents were intended to improve hardenability,
which, in turn, would potentially compensate for the loss in
strength upon tempering, while still ensuring weldablity. Alloy
E-A also had minor additions (0.06%) of rare earth metals
(REM) (La and Ce) to evaluate their potential benefits in refining the as-cast microstructure and controlling the morphology
of the sulfide inclusions.[7–9] Alloy E-A was cast into blocks
with dimensions of 300 ⫻ 300 ⫻ 530 mm as well as 230 ⫻
230 ⫻ 460 mm to investigate the effect of section size on the
mechanical properties. Unlike alloy E, these castings were fully
heat treated at the foundry, adopting commercial practices. The
austenitization and tempering treatments were chosen based on
the best mechanical property results obtained from thin slabs
of alloy E. The chemical compositions of the two alloys are
provided in Table 1.
2.2 Simulated Heat Treatment of Alloy E Using Thin Slabs
A novel experimental procedure was adopted, which enabled
a thin 19 mm slab to be used in each heat-treatment trial rather
than the entire casting. For each heat treat and water quench
operation, the cooling profile at the quarter-thickness (T/4)
location of the 300 ⫻ 300 ⫻ 530 mm casting was modeled by
finite element methods (FEM). (The T/4 location is the location
prescribed for evaluating mechanical properties in a test
636—Volume 10(6) December 2001
Fig. 1 Comparison of experimental cooling curves with those from
FEM computations. Experimental data points are overlaid on FEM
predictions (solid line)
block.[5]) The predicted cooling rate was duplicated on the thin
slab, which had been held at the desired temperature. The
desired cooling rate was accomplished by using a combination
of various slower cooling media, such as forced air convection,
cooling in still air, wrapping the slab in an insulating blanket,
etc. The thickness of the slab (19 mm) was decided based on
considerations of the Biot Modulus, which predicted that this
slab would cool uniformly with no thermal gradients for these
cooling media. Further details may be obtained from Ref 6
and 10.
Figure 1 shows the calculated cooling curves overlaid with
experimental results from cooling a 19 mm slab. The experimental technique affords a good match of the cooling at the
desired location, while enabling considerable time and material
savings and reducing experimental costs.
A potential criticism of this method is that the microstructure
and chemical composition of slabs from different locations in
the casting might not be representative of that at the T/4 location,
notwithstanding the homogenization treatment. While that is a
valid concern, the intent of these trials was merely to perform
a screening of various heat-treatment operations that could be
employed in heat treating a full-sized casting in a later trial.
Most of the slabs used in these trials were located as close to
the center of the casting as possible and between the two quarterplanes bounding the midplane. Furthermore, as will be seen in
Section 4.1, Jominy end-quench tests on alloy E reveal that the
hardenability behavior of samples taken from the center and
from the edge of the casting are similar. This observation,
along with the limited goals of this effort, precluded further
consideration of the effects of segregation and inhomogeneous
microstructure of the alloys studied.
2.3 Heat Treatment
Alloy E. Based on the finite element method (FEM) simulations, the heat-treatment evaluation of alloy E was conducted
on slabs of a 300 ⫻ 300 mm section and 19 mm thickness.
Table 2 shows the heat-treatment range employed. The austenitizing and tempering temperature ranges were chosen from
the literature,[4] and the hold time at temperature was chosen
following typical industrial practice for heavy-section steel castings and forgings (1 h for up to 25 mm in thickness, plus 30
min for each additional 25 mm).[11] Some sections of alloy E
Journal of Materials Engineering and Performance
received a double-tempering treatment, as in the SP-7 study,[4]
while others received only a single temper.
Alloy E-A. The knowledge gained regarding the effect of
various heat treatments on the mechanical properties was then
used to select the optimum heat treatment for the full-sized
castings of alloy E-A. This is shown in Table 3.
2.4 Mechanical Properties and Microstructure Evaluation
Tensile and CVN impact properties were evaluated
according to ASTM Standards E-8 and E-23, respectively.[12,13]
These were evaluated on the heat-treated slabs of alloy E and
at the T/4 location for alloy E-A. The tensile results reported
are the average of two tests on standard 13 mm gauge diameter
samples conducted at 25 ⬚C, while the CVN results represent
the average of 5 tests on standard 10 mm square cross-section
samples, conducted at ⫺18 and ⫺73 ⬚C. Limited dynamic
tear (DT) testing[14] was also performed to determine the DT
energies at ⫺40 ⬚C. Metallographic samples taken from the
tensile and CVN specimens representing various heat-treatment
conditions were evaluated by optical microscopy. Limited transmission electron microscopy (TEM) was used to examine double-tempered structures from alloy E, and x-ray diffraction was
employed to determine retained austenite. Scanning electron
microscopy (SEM) was used to inspect the tensile and CVN
fracture surfaces.
2.5 Weldability
Weldability of alloy E was evaluated by the Tekken test that
has been standardized by the American Welding Society (AWS
B4.0-95).[15] Here, a single test weld is deposited on the sample
of the appropriate “Y-groove” configuration. Weldments are
then inspected for signs of HAC, either in the weld metal (WM)
or HAZ. Welding was done under controlled conditions of 16
⬚C (i.e., no preheat) and 82% relative humidity to simulate
shipyard welding on a cold, humid day. Welding was performed
under the following conditions:
Table 2 Temperature range used in heat treatment of
alloy E
Austenitization
First temper
899 to 954 ⬚C
Water quenched(a)
649 to 732 ⬚C
Water quenched(a)
Second temper
● None
● 566 to 621 ⬚C
water quenched(a)
(a) Cooling rate equivalent to water quench for a 300 mm thick casting,
calculated from the FEM model
•
•
shielded metal arc welding (SMAW) using E-10018M electrode, 1.1 kJ/mm energy input, and
gas metal arc welding (GMAW) using MIL-100S electrode
and an experimental lower carbon LC-100 electrode with
1.46 kJ/mm energy input.
Alloy E was compared to various traditional steels (an HY80 equivalent A757-EQ21 casting and an HY-100 plate).
3. Results
3.1 Alloy E Casting
Tensile and CVN Impact Properties. The tensile and CVN
impact properties of alloy E were evaluated over a wide range
of austenitization, tempering, and double tempering temperatures. In general, it was found that the austenitizing temperature
did not have a significant effect on the tensile properties over
the temperature range of 899 to 954 ⬚C. For example, in the
single-tempered condition (677 ⬚C), the same yield strength of
607 MPa was obtained in samples austenitized at 899 ⬚C as
well as 954 ⬚C. The Charpy values also appeared comparable
within the range of scatter (51 ⫾ 25 J and 48 ⫾ 13 J at
⫺73 ⬚C after austenitizing at temperatures of 899 and 954 ⬚C,
respectively, and tempering at 677 ⬚C). It was also found that
a single subcritical temper (i.e., below Ac1 ⫽ 665 ⬚C[16]) was
not sufficient to produce the desired combination of yield
strength and low-temperature toughness. It appeared that a double-tempering operation consisting of an intercritical temper
followed by a subcritical temper was required to improve low
temperature toughness.
Table 4 shows the best tensile and CVN properties of alloy
E resulting from single intercritical-tempering treatments and
those resulting from a double-tempering treatment (intercritical
followed by subcritical) at the temperatures indicated. The CVN
energy values are shown along with the 95% confidence limits.
Following austenitization and a single temper in the 677 to
705 ⬚C range, alloy E exhibited adequate tensile properties and
upper-shelf CVN energies at ⫺18 ⬚C, sufficient to meet MILS-23008D requirements for HY-80. However, the CVN properties at ⫺73 ⬚C were not achieved. The double-tempering treatment at 621 ⬚C increased CVN impact properties at both
temperatures as well as the DT energy, but the yield strength
values were reduced to below the 551 MPa minimum specified
for HY-80. Thus, though this heat treatment leads to satisfactory
impact properties, it did not meet MIL-S-23008D strength
requirements. The best combination of strength and toughness
was obtained upon an intercritical temper in the 677 to 705 ⬚C
range, followed by a subcritical temper at 621 ⬚C.
Table 3 Heat treatment of full-sized blocks of alloy E-A
Composition/
Block ID
Dimensions (mm)
Preliminary HT
E-A-1
300 ⫻ 300 ⫻ 530
E-A-2
230 ⫻ 230 ⫻ 460
Proprietary homogenization and appropriate
temper for removal of gates and risers
Journal of Materials Engineering and Performance
Austenitization
927 ⬚C/6.5 h water
quench
927 ⬚C/5 h water
quench
Temper
691 ⬚C/6.5 h water
quench
691 ⬚C/5 h water
quench
Second temper
621 ⬚C/2 h water
quench
621 ⬚C/2 h water
quench
Volume 10(6) December 2001—637
Table 4 Optimum mechanical properties of alloy E
Tensile properties (room temperature)
Austenitization (ⴗC)
927
First
temper
(ⴗC)
Second
temper
(ⴗC)
Yield
strength
(MPa)
Tensile
strength
(MPa)
Elong. %
51 mm
gauge
Reduction in
area (RA),%
677
None
621 (2 h)
None
621 (2 h)
607
538
690
531
869
724
959
745
24
27
18
26
71
72
65
71
551
…
20
…
705
HY-80 MIL-S-23008D
specifications
(minimum)
Microstructure and Fracture Evaluation. Microstructural examination of the homogenized cast material revealed
the presence of a very coarse microstructure with a prior-austenite grain size ranging from 250 to 300 ␮m. The coarse microstructure may be expected due to the large size of the casting.
The microstructure also consisted of lath ferrite, surrounding
small regions of angular/blocky bcc phase, with small amounts
of retained austenite (1 to 3%).
The prior-austenite grain size of the austenitized, single- and
double-tempered material was found to have a considerable
scatter. For example, samples from single tempered material at
677 ⬚C/1 h after austenitization at 927 ⬚C/1 h varied from 27
to 65 ␮m. Double-tempered samples at 677 ⬚C/6.5 h and 593
⬚C/6.5 h after austenitizing at 927 ⬚C/6.5 h showed a grain-size
range of 21 to 60 ␮m. Thus, a direct correlation between the
heat treatment and grain size was not possible. However, the
heat treatments used substantially refined the microstructure of
the initial homogenized cast material.
Figure 2 shows the light optical microstructures (LOMs) of
alloy E in the single- and double-tempered conditions, respectively. Note that single- and double-tempered materials did not
reveal clear differences in their microstructure, which primarily
consisted of bainite, some ferrite, and possibly some martensite,
as shown in Fig. 2(b) and (d), respectively. Additionally, independent of the heat treatment employed, the “dendrite ghosts”
of the as-cast microstructure are still reminiscent (Fig. 2(a) and
(c)). Further TEM inspection of the double-tempered specimens
by Fonda et al.[17] indicated that the microstructure primarily
consisted of ferrite laths, a highly dislocated lenticularlike bcc
phase, and very small amounts of retained austenite (Fig. 3).
X-ray diffraction of the double-tempered steel indicated retained
austenite levels of approximately 0.3%.
As a note of reference, the microstructures following similar
double-tempering treatments were examined earlier.[4,18] It has
been suggested that the improvement in toughness following
a double temper is due to the following. During the first (intercritical) temper, some of the existing microstructure transforms
to austenite while the rest undergoes overtempering (softening).
Upon water quenching, the austenite transforms to martensite.
In the second (subcritical) tempering operation, the martensite
is tempered, thus leading to overall improved toughness.
In general, inspection of the CVN fracture surfaces revealed
that all samples tested at ⫺18 ⬚C exhibited a fully ductile
behavior (Fig. 4a). Conversely, both single- and double-tempered samples tested at ⫺73 ⬚C had a mixture of ductile and
638—Volume 10(6) December 2001
CVN energy (J) and % shear
ⴚ18 ⴗC
133
195
121
193
⫾
⫾
⫾
⫾
24
11
8
11
95
100%
100%
100%
100%
ⴚ73 ⴗC
61
92
49
91
⫾
⫾
⫾
⫾
9
24
9
30
68
28%
44%
47%
49%
DT energy (J)
and % shear
ⴚ40 ⴗC
805 67%
1205 100%
418 43%
1159 100%
…
brittle fracture (Fig. 4b and 5a and b), with the double-tempered
samples exhibiting more ductile behavior than the single-tempered samples (Fig. 4b versus 5b). It was also observed that
samples with a smaller and narrower prior-austenite grain-size
range, in both single- and double-tempered conditions, had a
higher CVN energy than those with a larger size distribution.
For example, a CVN energy of 75 J was obtained from a
double-tempered material (677 ⬚C/6.5 h and 593 ⬚C/6.5 h after
austenitizing at 927 ⬚C/6.5 h) with a prior-austenite grain size
of 24 to 34 ␮m. Specimens with the same tempering conditions
and a prior-austenite grain size of 21 to 52 ␮m had a CVN
energy of only 48 J. An additional feature of many fracture
surfaces was the presence of relatively large inclusions (5 to
15 ␮m). Energy dispersion spectroscopy (EDS) analysis indicated that the inclusions were complex oxides containing primarily Al, Ca, Fe, Ni, and Zr (Fig. 5c). Their presence may
have adversely affected the impact energies and other mechanical properties.
Weldability Evaluation of Alloy E. Tekken tests using
SMAW of alloy E and HY-100 using E10018-M revealed significant differences (Fig. 6a and b). HY-100 exhibited poor
weldability with HAZ and WM cracking. This is as expected
for a higher carbon steel welded without preheat. On the other
hand, alloy E exhibited no HAZ cracks, indicative of better
weldability. However, it had WM cracks, indicating the need for
a more “hydrogen-free” process such as GMAW with different
consumables, for instance, MIL-100S or the lower carbon LC100. While the MIL-100S electrode is sometimes used for
welding thin sections without preheat, in the present case, it
resulted in WM cracks and some HAZ cracks (not shown)
and, thus, was not found to be suitable. Figure 6(c) shows the
weldment of alloy E obtained by GMAW using the experimental
electrode, LC-100, which was formulated to be used without
preheat. This exhibits good weldability, as seen by the absence
of cracks either in the WM or HAZ. However, due to material
limitations, only one valid test was performed for alloy E under
this condition, while the AWS-4.0 test procedures call for a
minimum of three. Thus, the tests were not conclusive, but
preliminary results were encouraging.
3.2 Alloy E-A Castings
Tensile and CVN Impact Properties. The tensile and CVN
impact properties of castings E-A-1 and E-A-2, evaluated at
the T/4 location, are listed in Table 5.
Journal of Materials Engineering and Performance
Fig. 2 Optical microstructures of alloy E austenitized at 927 ⬚C/6.5 h: (a) and (b) single tempered at 677 ⬚C/1 h, and (c) and (d ) double tempered
at 677 ⬚C/6.5 h ⫹ 593 ⬚C/6.5 h. Etchant: modified Winsteads and 2% Nital
Journal of Materials Engineering and Performance
Volume 10(6) December 2001—639
Fig. 3 The TEM microstructure of double-tempered E-steel showing
mostly lath ferrite, some bcc-ferrite, and possibly small pockets of
austenite
Note that yield-strength values are in the range of 517 to
538 MPa and, thus, fall short of HY-80 requirements. These
may be compared to the results in Table 4, based on subscale
heat treatment of alloy E. The austenitization and tempering
treatments in both cases (alloy E subscale and alloy E-A full
scale) were quite similar. It may be seen that the yield-strength
values are in the same range (531 to 538 MPa for E and 517
to 538 MPa for E-A). At first glance, it would appear that the
alloy modifications (such as increased C and Mn) did not lead
to any significant improvement in hardenability, at least for the
section sizes under consideration.
The elongation at failure and reduction in area are quite low,
indicative of hydrogen embrittlement (HE) problems. Interestingly, HE problems were not encountered in the trials done on
alloy E castings, where heat treatment was performed on 19
mm slabs cut from the castings. This appears to be a potential
problem in heat treatment of thick sections. It is likely that there
is considerable microshrinkage and porosity at the midplane of
such a thick section, where hydrogen would tend to accumulate.
If this is not diffused out or dispersed by a hydrogen-baking
treatment, HE is a strong possibility. An experimental baking
treatment of 232 ⬚C/24 h was tried on tensile samples machined
from regions adjacent to the T/4 location. This increased the
elongation to failure to 25%, while the yield strength was unaffected. Thus, this seemed to be an appropriate treatment for
eliminating HE, at least for the thin sections employed (13
to 19 mm thick). Thicker sections might call for increased
temperature and/or holding time.
Analysis of the yield-strength values in conjunction with
640—Volume 10(6) December 2001
Fig. 4 The SEM image of fractured charpy samples of alloy E (double
tempered): (a) tested at ⫺18 ⬚C and (b) tested at ⫺73 ⬚C
the CVN energies indicates that the desired combination of these
properties has not been achieved. Satisfactory CVN energies, as
seen in the case of E-A-2, appear to be combined with relatively
low yield strengths. Another feature of the impact testing was
the nature of fracture surfaces observed. Several samples exhibited an anomalous appearance, marked by the presence of “flat
facets,” in addition to the usual ductile dimples or trans- and
intergranular cleavage. This has been referred to in the past as
“rock candy fracture” (RCF).[19] The reasons behind RCF are
not quite clear; it has been previously speculated that aluminum
nitride (AlN) embrittlement might be a probable cause.[19] While
RCF can lead to lower CVN energies, another effect of RCF
is to increase the scatter in CVN energy data.
While the mechanical properties obtained from these fullsized castings fall short of HY-80 specifications, they represent
an improvement over results from prior investigations.[4,6] These
castings possess improved low-temperature (⫺73 ⬚C) impact
properties than cast HSLA-80 and HSLA-100 blocks (Table
6). These castings could also be cast and heat treated without
any cracking problems, as encountered with the HSLA-100
composition.
Journal of Materials Engineering and Performance
(c)
Fig. 5 The SEM image of fractured Charpy samples of alloy E (single tempered) (a) tested at ⫺73 ⬚C, (b) tested at ⫺73 ⬚C, and (c) EDS spectra
of inclusion shown by the arrow in (a)
Microstructure Evaluation. Optical microscopic examination of E-A steel castings showed that the microsegregation of
the as-cast material persisted even after homogenization and
full heat treatment. Figure 7 shows the macrostructures found
in CVN specimens taken from the 230 mm thick casting.
In general, these castings had a greater extent of microsegregation and a higher porosity level than the castings of alloy
E. The dendrite arm spacing, which is an indicative parameter
of microsegregation, was found to be ⬃400 ␮m in E-A compared to 170 to 240 ␮m in alloy E. The reasons for this are
unclear; the modified alloy chemistry or a change in the melting
and pouring process may have contributed to some degree.
Journal of Materials Engineering and Performance
Most of the porosity was found at the dendrite ghosts of the
interdendritic regions of the castings, as shown in Fig. 7. In
addition to shrinkage porosity, other regions of the specimens
showed porosity with rounded features, which may be indicative
of gas porosity. An example is shown in Fig. 8.
The SEM examination (Fig. 9) also revealed the presence
of relatively large clusters of REM inclusions (⬎10 ␮m). The
EDS spectrum (Fig. 9b) of the REM oxide inclusions in alloy
E-A also indicates the presence of sulfur and some aluminum.
The REM inclusions tended to be associated with porosity.
The LOM analysis indicated that the microstructure of the
steels in the fully heat- treated condition consisted primarily
Volume 10(6) December 2001—641
Fig. 6 Low magnification pictures of Tekken test weldments: (a) HY-100, SMAW (MIL-10018); (b) alloy E, SMAW (MIL-10018); and (c) alloy
E, GMAW (LC-100). All tests performed at 16 ⬚C, 82% relative humidity, 1.1 to 1.46 kJ/mm energy input
Table 5 Mechanical properties of E-A-1 and E-A-2 castings at T/4 location
Tensile properties (room temperature)
Section size/
block ID
YS
(MPa)
TS
(MPa)
Elong. %
51 mm
gauge
RA, %
ⴚ18 ⴗC
ⴚ73 ⴗC
Comments
300 mm (E-A-1)
230 mm (E-A-2)
538
517
662
717
7
11
20
22
99 ⫾ 34 35%
123 ⫾ 22 50%
79 ⫾ 31 35%
80 ⫾ 26 36%
HE in all tensiles(a)
HE in all tensiles
551
…
20
…
95
68
…
Alloy
E-A
MIL-S-23008D specifications (minimum)
CVN energy (J) and % shear
(a) HE ⫽ hydrogen embrittlement
of bainite with some ferrite and possibly very small pockets of
martensite (Fig. 10).
Also, no differences in microstructure (at the LOM level)
were found between the interdendritic and dendrite areas in
this steel, except that the prior-austenite grain size at the interdendritic regions appeared to be 30 to 40% smaller than at the
642—Volume 10(6) December 2001
dendrite regions. However, the variations observed in prioraustenite grain size may be attributed to the large variations in
cooling rates from the surface to the center of the casting from
solidification to temperatures below Ac1.
Table 7 shows the prior-austenite grain-size ranges for
samples from heat-treated alloys E-A and alloy E. Note that
Journal of Materials Engineering and Performance
Fig. 7 Low magnification optical microstructures of homogenized
and heat-treated (Table 3) 230 mm section thickness steel E-A
Table 6 Mechanical properties of HSLA-80 and HSLA100 castings from literature
Casting
composition
[4]
HSLA-80
HSLA-100[6]
Section
size
(mm)
150
300
Tensile properties
(room temperature)
CVN energy (J)
YS
(MPa)
TS
(MPa)
ⴚ18 ⴗC
ⴚ73 ⴗC
546
558
625
703
130
138
9
57
the grain size for commercial HY-80 steel is given for
comparison.
While there is a large variation in the prior-austenite grain
size within and among the various castings, it would appear
that E-A castings have a narrower and relatively smaller prioraustenite grain-size range than alloy E. However, both alloy
E and E-A steels have a coarser prior-austenitic structure than
the HY-80 steel, despite the latter having a larger cross section
(note that different vendors manufactured HY-80 and Esteels).
Large variations in prior-austenite grain size may be expected in castings with large cross sections. These variations
may be attributed to the chemical composition, manufacturing
practices that vary from vendor to vendor (melting, deoxidation, alloying, pouring parameters, casting size, mold materials, design, etc.), and homogenization treatment. The
difference in parameters, such as dendrite arm spacing and
porosity, between the two batches of castings also indicates
that the casting process was not standardized (especially for
these experimental alloys). More work is needed to determine
the effects of casting manufacturing parameters and solidification phenomena (microsegregation and solidification rates
within the casting) on the parent-austenite grain size. This
would help develop techniques to control the grain size for
the optimization of microstructure and mechanical properties.
Journal of Materials Engineering and Performance
Fig. 8 Porosity in homogenized and heat-treated (Table 3) 230 mm
section thickness casting of alloy E-A: (a) light optical view and (b)
SEM view
4. Discussion
The quenched-and-tempered low-C, high-Ni steel compositions (alloy E and its modification E-A) resulted in crack-free
castings. Preliminary tests suggest that alloy E can be welded
without preheat. However, neither E nor E-A achieved the
required mechanical properties to meet MIL-S-23008D for HY80 castings. This is perhaps indicative of a lack of hardenability
or a nonoptimal microstructure of the alloys studied. To address
these issues, first, the hardenability of these steels was evaluated
using the Jominy test,[20] in conjunction with relevant continuous cooling transformation (CCT) diagrams. Second, the casting
results were compared with mechanical property data from
forgings, which were made from a casting of the same alloy
E.[21]
4.1 Hardenability
A potential factor in the relatively low mechanical properties
of these steel castings is their hardenability. This can be seen
Volume 10(6) December 2001—643
(b)
Fig. 9 The SEM views and EDS spectra of REM inclusions in alloy E-A
from Fig. 11, which compares the hardenability for alloy E and
the HY-80 equivalent steel (A757 E2Q1 steel cast as a 350 ⫻
350 ⫻ 1050 mm block). Note that the data presented in Fig.
11 are from samples taken from the center as well as closer to
the surface of the castings, in order to account for potential
effects of segregation. The HY-80 steel exhibits a hardness of
about 41 RC (Rockwell C) at the quenched end, similar to what
might be expected of a martensitic structure for that carbon
content.[22] With increasing distance from the quenched end
(i.e., with decreasing cooling rate), there is a gradual decrease
644—Volume 10(6) December 2001
in hardness to about 39 RC. On the other hand, alloy E exhibits
a considerably lower hardness than HY-80, and the hardness
throughout the sample remains practically constant (32 to 33
RC). The lack of a significant hardness variation with distance
indicates that alloy E is insensitive to cooling rates. It is also
interesting to note that the hardness of the samples taken from
the center and those taken closer to the surface of the casting
are quite similar. Thus, segregation, if present, does not appear
to have a significant effect on the hardenability in both steels.
This relative insensitivity of hardness to cooling rate was
Journal of Materials Engineering and Performance
Fig. 10 Optical and SEM microstructures of fully heat-treated alloy E-A casting
Fig. 11 Jominy hardenability for alloy E and HY-80 equivalent (A757 E2Q1) cast steels austenitized at 927 ⬚C and 899 to 921 ⬚C for 1 h per
squared inch, respectively
Table 7 Prior-austenite grain size in alloys E-A, alloy E, and HY-80 steel
Alloy E-A-2
Casting section, mm
Grain size range, ␮m
ASTM
230 ⫻ 230 ⫻ 460
27–35
7.2–6.7
Journal of Materials Engineering and Performance
300 ⫻ 300 ⫻ 530
32–45
7–6
Alloy E
HY-80 casting
300 ⫻ 300 ⫻ 530
21–60
8–4.7
990 ⫻ 990 ⫻ 330
⬃16
9
Volume 10(6) December 2001—645
Fig. 12 Optical microstructures and hardness values of samples of alloy E-A that were controlled cooled from approximately 1000 ⬚C: (a) cooled
at 0.5 ⬚C/min, 332 HV and (b) cooled at 132 ⬚C/min, 343HV. Etchant: modified Winsteads and 2% Nital
observed in alloy E-A castings as well. This was evident from
a microstructural and hardness analysis of samples subjected
to controlled fast and slow cooling cycles, as shown in Fig.
12. The microstructure of the sample cooled at 132 ⬚C/min was
found to be predominantly martensitic with a hardness of 343
HV (⬃35 RC), while that of the slow-cooled sample (0.5 ⬚C/
min) was mostly bainitic with a hardness of 332 HV (⬃33 RC).
Thus, the hardness of martensite is fairly similar to that of
bainite, perhaps due to the low carbon content of the martensite
produced and with the resultant body-centered tetragonal (bct)
structure having a low c/a ratio. The insensitivity to cooling
rate is also in keeping with the properties of the 300 and 230
mm thick castings of Table 5. This analysis shows that the
maximum hardness values of alloy E (from the Jominy test)
and alloy E-A (from the controlled cooled sample) are quite
similar, indicating that the slight increases in C and Mn in the
latter had no effect on promoting hardening.
The lack of hardenability may be further corroborated by
an inspection of a partial CCT diagram developed for alloy
E-A[6] and overlapped with that of a similar alloy (1118) from
the SP-7 study.[4 ]This is shown in Fig. 13, along with the
predicted cooling rate at the T/4 location for a 125 mm and a
300 mm thick section subjected to water quench. The 300
mm thick section results in a nonmartensitic structure upon
austentization and quench from 927 ⬚C. This correlates well
with the bainitic microstructure observed in this alloy (Fig. 10).
Quantitative estimates of hardenability, such as the ideal
critical diameter (DI), have been established mostly for mediumcarbon steels.[22] The DI is a measure of the thickest round bar
of a given composition that will result in at least 50% martensite
at its center, given a perfect quench. Such data are not available
for lower carbon alloys, such as alloy E. To a first approximation, a section size of 125 mm or less might produce a significant
amount of martensite, as shown in Fig. 13. However, even if a
646—Volume 10(6) December 2001
predominantly martensitic structure were obtained, the hardness
would not be comparable to that of HY-80. Note that the E-A
sample cooled at 132 ⬚C/min gave a hardness of 35 RC (343
HV), which is lower than the hardness of 41 RC at the waterquenched end of HY-80. This is primarily due to the low carbon
content in alloy E.
4.2 Comparison of Casting and Forging Properties
Analysis of tensile and CVN impact properties of alloy E
in the forged condition[21] showed that the specified properties
in alloy E could be achieved by refining the microstructure and
closing porosity by forging. In this case, the forged and fully
heat-treated material had a prior-austenite grain size of approximately 18 ␮m (8.4 ASTM) with a primarily bainitic microstructure. The results are compared in Table 8.
This analysis indicates that it would be very difficult for alloy
E to meet the MIL-S-23008D specified mechanical properties in
the cast form. However, microstructural refinement and elimination of casting defects, such as porosity, may result in better and
more consistent properties than those observed in the castings of
this investigation.
5. Conclusions
Quenched-and-tempered low-C, high-Ni steels were evaluated as potential alternatives to HY-80 castings. The focus
of the study was on optimizing the heat treatment to obtain
mechanical properties equivalent to HY-80 and evaluating their
weldability. The results of this study indicate that the compositions studied were not completely able to meet the tensile and
CVN impact properties specified by MIL-S-23008D for HY80 cast steels. However, significant improvements in the heat
Journal of Materials Engineering and Performance
Fig. 13 The CCT diagrams for alloy E-A and alloy-1118 from Ref 4
Table 8 YS and CVN properties for alloy E forged at 1093 ⴗC
Normalizing/
austenitizing (ⴗC)
1010/927
1010/927
Tempering
(ⴗC)
YS(0.2%) L/T
(MPa)
CVN at ⴚ18 ⴗC/ⴚ84
ⴗC (J)
Reduction ratio
649
649
594/598
566/571
206/133
197/119
3:1
5:1
treatment of the low-C, high-Ni steels were achieved. A doubletempering treatment led to reasonable yield strength (517 to
538 MPa) and CVN energy (80 J at ⫺73 ⬚C) in castings with
230 to 300 mm thick sections. This represents a considerable
improvement over the cast HSLA-80 and cast HSLA-100 compositions, especially with regard to low-temperature CVN
impact properties. The castings of the alloys studied were reasonably sound and free of cracks, and preliminary tests suggest
that castings of this material could be welded without preheat.
Thus, while this alloy cannot be a direct replacement for cast
HY-80, there might be other potential applications for which
this could be considered. These include surface ship shaft struts
and rudder inserts with less stringent strength and toughness
requirements.
Acknowledgments
This work was conducted by the National Center for Excellence in Metalworking Technology, operated by Concurrent
Technologies Corporation under Contract No. N00014-00-C0544 to the U.S. Navy as part of the U.S. Navy Manufacturing
Technology Program. The authors acknowledge the technical
contributions and advice of Paul Konkol and Gerard Mercier,
Concurrent Technologies Corporation, and Dr. Richard Fonda,
Journal of Materials Engineering and Performance
Naval Research Laboratory, in performing TEM characterization of the alloys.
References
1. Anon.: “Fabrication, Welding and Inspection of HY-80/100 Submarine
Applications,” Military Specification MIL-S-1688A, Naval Sea Systems Command, Arlington, VA, 1990.
2. M.J. Kleinosky, C.L. Trybus, D.L. Winterscheidt, and D.W. Yuan:
Advanced HY-80 Castings Heat Treatment Analysis, NCEMT TR No.
00-36, Concurrent Technologies Corporation, Johnstown, PA,
2000.
3. E.J. Czyryca: Development of Low-Carbon, Copper-Strengthened
HSLA Steel Plate for Naval Ship Construction, DTRC-SME-90/21,
David Taylor Research Center, Carderock, MD, 1990.
4. R.K. Churchill, J.H. Devletian, and D. Singh: High Yield Strength
Cast Steel with Improved Weldability, National Shipbuilding Research
Program, NSRP No. 0326, Advanced Technology Institute, North
Charleston, SC, 1991.
5. Anon.: “Steel Castings, Alloy, High Yield Strength (HY-80 and HY100),” Military Specification MIL-S-23008D (SH), Naval Sea Systems
Command, Arlington, VA, 1990.
6. K. Kannan, P.J. Konkol, J.R. Martinez, K. Ray, B.P. Tipton, and
J.J. Valencia: Evaluation of Candidate Casting Compositions (Interim
Report): Enhanced Processing of High Strength Steel Castings and
Forgings for Naval Components, NCEMT TR No. 99-47, Concurrent
Technologies Corporation, Johnstown, PA, 1999.
Volume 10(6) December 2001—647
7. M.P. Seah, P.J. Spencer, and E.D. Hondros: Met. Sci., 1979, vol. 13
(5), pp. 307-314.
8. W.G. Wilson, L.J. Heaslip, and I.D. Sommerville: J. Met., 1985, vol.
37 (9), pp. 36-41.
9. C.I. Garcia, G.A. Ratz, M.G. Burke, and A.J. DeArdo: J. Met., 1985,
vol. 37 (9), pp. 22-28.
10. K. Kannan, K. Ray, J.R. Martinez, and P. J. Konkol: Int. Symp. Steel
for Fabricated Structures, R. Asfahani and R. Bodnar, eds., ASM
International, Materials Park, OH, 1999.
11. P.J. Konkol: Lenape Forge, West Chester, PA, private communication, 1998.
12. Anon.: “Standard Test Methods for Tension Testing of Metallic Materials,” ASTM E8-98, ASTM, West Conshohocken, PA, 1998.
13. Anon.: “Standard Test Methods for Notched Bar Impact Testing of
Metallic Materials,” ASTM E23-96, ASTM, West Conshohocken,
PA, 1996.
14. Anon.: “Standard Test Method for Dynamic Tear Testing of Metallic
Materials,” ASTM E604-83, ASTM, West Conshohocken, PA, 1994.
15. Anon.: “Oblique Y-Groove Test,” AWS B4.0-95, American Welding
Society, Miami, FL, 1995.
648—Volume 10(6) December 2001
16. C.A. Papesch and J.J. Valencia: Thermophysical Properties Determination of HSLA Steels, NCEMT Laboratory Report TR No. 98-07, Concurrent Technologies Corporation, Johnstown, PA, 1999.
17. R.W. Fonda, J. Feng, and G. Spanos: “High Strength Steel Castings and
Forgings for Naval Applications,” TMS 1999 Fall Meeting, Cincinnati,
OH, 1999.
18. S. Bechet and K. Rohrig: Weldable High Strength Steel Cast Steel,
Climax Molybdenum Company Report, Greenwich, CT, 1982.
19. N.H. Croft, A.R. Entwisle, and G.J. Davies: Advances in Physical
Metallurgy and Applications of Steels, The Metals Society, London,
1982, pp. 286-95.
20. Anon.: “Steel-Bars, Forgings, Bearing, Chain, Springs,” ASTM Standard A 255, ASTM, West Conshohocken, PA, 2000, pp. 43-63.
21. D.W. Yuan, P.J. Konkol, K. Ray, B.P. Tipton, and J.J. Valencia: Evaluation of Alternate High Strength Steel Forgings: Enhanced Processing of
High Strength Steel castings and Forgings for the Naval Components,
NCEMT Report TR No. 99-32, Concurrent Technologies, Johnstown,
PA, 1999.
22. G. Krauss: Steels: Heat Treatment and Processing Principles, ASM
International, Materials Park, OH, 1990, p. 147.
Journal of Materials Engineering and Performance
Fly UP