User manual | University of Pretoria – Z Tang (2006)

University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 4 Microstructure and mechanical properties
CHAPTER 4 MICROSTRUCTURE AND MECHANICAL PROPERTIES
4.1 Acicular ferrite
4.1.1 Nucleation and growth of acicular ferrite
Microstructures with a significant proportion of acicular ferrite present an optimised
combination of mechanical properties if compared with mainly bainitic structures. It
is well documented that acicular ferrite formation is enhanced by the presence of
non-metallic inclusions in studies in weld pools[51-64], low carbon steels[65] and
medium carbon forging steels[66-69], and is characterized by elongated grains that are
“chaotically” arranged[4]. These second-phase particles act as point sites on which the
intragranular nucleation[70,71] of ferrite units develops. There may also be some M/A
islands present with a high dislocation density[11]. Acicular ferrite is a non-equiaxed
structure phase with an interior that contains a dense substructure of dislocations[4,72].
The carbon content in the M/A islands is higher than that in the surrounding matrix.
Accordingly, the M/A islands are carbon-enriched, whose formation may be attributed
to the partitioning of carbon during the transformation to acicular ferrite and the
post-transformation of carbon-enriched austenite. When the specimen is deformed in
the non-recrystallisation austenite region, high densities of substructure and
dislocations will be formed in the austenite, which increase the nucleation rate of
acicular ferrite, impedes the growth of the coherent and/or semi-coherent γ/α
interfaces and accelerates diffusion of carbon to these γ/α interfaces, which leads to
carbon-enriched austenite. During the accelerated cooling after finish rolling and
followed by the coiling process, part of the carbon-enriched austenite transforms to
martensite and the retained austenite coexists with the martensite[73]. The
transformation model is, therefore, a mix of diffusion and shear transformation[4,9,11,43].
The start temperature of the transformation to acicular ferrite is slightly higher than
that of an upper bainite[4,11].
Acicular ferrite always has an orientation relationship with the austenite grain, such
- 32 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 4 Microstructure and mechanical properties
that one of its close packed {110}AF planes is nearly parallel to the close-packed
{111}γ plane of the parent austenite. Within these parallel planes, a close-packed
−
−
<11 1 >AF direction of the acicular ferrite is found to be near to a close-packed < 101>γ
direction of the austenite[74]. This demonstrates that the growth of acicular ferrite
occurs by a displacive transformation and its growth, therefore, takes place without
carbon diffusion. The excess carbon in ferrite is probably rejected into the austenite
soon after nucleation. Acicular ferrite plates are never found to grow across austenite
grain boundaries and this is also consistent with the displacive transformation
mechanism, since the necessary co-ordinated movements cannot be sustained across
austenite grain boundaries. TEM work has revealed that the ferrite units belonging to
the same sheaf have the same crystallographic orientation in most cases[75].
As regards the carbon concentration of acicular ferrite structures during
transformation, experiments and thermodynamic theory have demonstrated that the
growth of acicular ferrite is diffusionless with the ferrite inheriting the chemical
composition of the parent austenite. The excess carbon in the acicular ferrite is
rejected into the retained austenite after transformation and can apparently occur
within a few seconds[74].
4.1.2 Two types of acicular ferrites: Upper and lower acicular ferrite
There are two different microstructural morphologies of acicular ferrite (AF) in
medium carbon micro-alloyed steels, depending on the isothermal treatment
temperature[75]. One is upper acicular ferrite, the typical acicular ferrite with an
interlocked microstructure (plate morphology) that is formed at high isothermal
treatment temperatures of typically 450 ºC. The secondary, is of acicular plates of
ferrite that had nucleated at the interface between the primary ones and the austenite
and are inclined at a high angle with respect to the substrate unit.
Another sheaf morphology of lower acicular ferrite, is composed of packets of plates
- 33 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 4 Microstructure and mechanical properties
following the same growth direction at lower isothermal treatment temperatures of
typically 400 ºC. A significant change in the morphology of the acicular ferrite is,
therefore, clearly apparent with a lowering of the formation temperature. It is
observed that the nucleation of the primary plates takes place intragranularly on the
same second-phase particles. These significant differences between the morphologies
of the two types of acicular ferrite can be distinguished in the early stages of the
transformation. At a high nucleation temperature of 450 ºC, single plates form at
second phase particles while at temperatures lower than 400 ºC, individual parallel
platelets are formed with residual phases in between them.
There may be two reasons for the formation of parallel AF units at low temperatures.
Firstly, the lower stability of the austenite close to the tip of the ferrite plate and
secondly, the strain field produced by the invariant plane-strain shape transformation,
both favour the formation of the same variant as that of the primary plate at these sites.
Further growth of these subunits seems to be possible parallel to the primary plate,
leaving a thin layer of carbon-enriched retained austenite between the different
subunits. Afterwards, these regions of austenite lead to the precipitation of interlath
cementite between the ferrite plates[75].
The autocatalytic formation of new plates of acicular ferrite is expected to depend
strongly on the carbon concentration profile of the parent austenite ahead of the
interface with the primary AF plates. This concentration profile will become more
pronounced as the transformation temperature decreases and the diffusion in the
austenite of the carbon rejected from the ferrite, becomes slower. Close to the acicular
ferrite tips, the carbon enrichment of the austenite could be low enough to allow the
transformation to proceed, leading to the formation of elongated sheaves. At high
transformation temperatures, the diffusion of carbon is more rapid and plate formation
on interfaces is more likely[75].
- 34 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 4 Microstructure and mechanical properties
4.1.3 Effect of the hot rolling process on acicular ferrite formation
Hot deformation will promote the subsequent formation of acicular ferrite. This is
because high densities of substructure and dislocations are formed in the austenite
during deformation of austenite in the non-recrystallisation region, which increases
the nucleation sites for acicular ferrite and promotes the acicular ferrite
transformation[4]. The temperature range for the nucleation of acicular ferrite moves
slightly towards higher temperatures and shorter times with hot deformation of
austenite if compared to an equivalent austenite without hot deformation. The growth
of acicular ferrite, however, is retarded in plastically deformed austenite[76]. With an
increase in the cooling rate after hot rolling, the fraction of polygonal ferrite decreases
and the fraction of acicular ferrite increases in volume and the grain size of the ferrite
becomes smaller[73].
4.2 Acicular ferrite and bainite
Bainite forms typically at temperatures between pearlite and martensite and the
transformation model is also a displacive mechanism. There are three kinds of
microstructure: upper bainite, lower bainite and granular bainite. Carbides precipitate
in-between the laths within upper bainite (which often results in a lower toughness),
while carbides are finely distributed within the bainite sheaths at a fixed orientation
within lower bainite together with some minor interlath formation of carbides also
here. Both forms of carbides have a specific orientation relationship between the
bainite lath and the carbides in lower bainite[11].
Bainitic ferrite is very different from acicular ferrite in its shape. It possesses largely
parallel sheaves of ferrite with a lath-like structure with some granule-like or rod-like
cementite particles alongside or within the laths and along the prior austenite grain
boundary network that can be seen clearly. The ferrite in bainite nucleates at the
austenite grain boundaries[66,77-80], forming sheaves of parallel plates with the same
- 35 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 4 Microstructure and mechanical properties
crystallographic orientation with respect to the parent austenite. Interlath carbide
particles in upper bainite precipitate from the carbon-enriched retained austenite
trapped in-between the platelets, or in the lower bainite, from within the
supersaturated ferrite within the bainite lath[74].
Acicular ferrite micrographs, on the other hand, have been studied[81] in which the
acicular ferrite transformation starts through the nucleation of the primary plates at
non-metallic inclusion particles[33,52,54,56,82,84] and progresses by the formation of a
new generation of secondary plates of ferrite nucleated at the interfaces of the
austenite/AF primary plates[78]. Therefore, inclusions play an important role in the
formation of acicular ferrite in low alloy welded metals[85,86] because they provide
preferential sites for the nucleation[87-89] of the AF. The acicular ferrite matrix is
characterized by a fine non-equiaxed ferrite or interwoven nonparallel ferrite
laths[66,73,78,90], which have various sizes distributed in a random manner, very often
described as a “chaotic arrangement” of plates showing fine-grained interlocking
morphologies[78]. The prior austenite grain boundary network is completely eliminated
and some fine M/A island constituents are scattered throughout the matrix[4]. This is
due to the partitioning of carbon near the austenite/ferrite interface during the growth
of acicular ferrite. The carbon content in the austenite will be increased and
accordingly, the austenite’s stability will be increased. As a result, the partial austenite
that is carbon-enriched, remains and transforms to martensite during the subsequent
cooling process, resulting in the M/A island constituent.
4.3 Mechanical properties of line pipe steel
4.3.1 The ratio of yield strength to ultimate tensile strength (YS/UTS)
A low YS/UTS ratio a very important parameter in the API specifications for line pipe
steels as a high work hardening rate is required for this application. The lack of strain
hardening in high YS/UTS steels means that there is a reduced potential for strain
redistribution in thinned areas (thinned by corrosion or weld dressing) during service.
- 36 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 4 Microstructure and mechanical properties
A high strength is, of course, required for line pipe to transport oil or natural gas at
higher pressures. The American Petroleum Institute (API) specifies a YS/UTS ratio
not greater than 0.93 for an application involving pipelines. The 11 mm line pipe strip
steel currently produced by MITTAL Steel (South Africa) tends to have a slightly high
YS/UTS ratio of typically 0.93. This ratio is affected by the microstructure of the steel
and an optimised microstructure (such as acicular ferrite[32]) is, therefore, beneficial in
achieving a lower YS/UTS ratio by carefully controlling the hot rolling, cooling and
coiling schedules. This ratio is also an important issue in the development of higher
grades of line pipe steels. The specification of line pipe steels of the America
Petroleum Institute (API) is shown in table 4.1.
Table.4.1 Specification of line pipe steels of API[12]
Grades of
Minimum yield
Minimum tensile
Minimum
Maximum
steels
strength (MPa)
strength (MPa)
Elongation
YS/UTS
(%)
X65
448
530
20.5
0.93
X70
482
565
19
0.93
X80
551
620
17.5
0.93
X90
601
650
17.5
0.93
Some results showed that there is a slightly higher volume fraction of about 7% of
martensite/austenite (M/A) constituent with a higher finish rolling temperature and,
therefore, a more rounded stress-strain curve and a higher strain-hardening rate on this
curve[1]. It is beneficial to lower the YS/UTS ratio and that author found that when the
finish rolling temperature was 720 ºC the ratio was 0.69 and at 780 ºC, the YS/UTS
was 0.65 and the volume fractions of M/A constituents were 4.6 and 7.0%
respectively for a steel with composition 0.057% C, 0.27% Si, 2.04% Mn, 0.040% Nb,
0.112% Ti and 0.001% B[1].
- 37 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 4 Microstructure and mechanical properties
4.3.2 Toughness
The Charpy toughness specification for the control of fracture initiation normally does
not prove too onerous a requirement for higher grade line pipe. Toughness is also
important for line pipe steels. Higher toughness can be obtained by lowering the
carbon content and refining the ferrite grain size[38]. Niobium can improved Charpy
toughness[25,26] at lower finishing temperature below 980 ºC[91]. A lower vanadium
content is also useful to increase the toughness[38]. Acicular ferrite microstructures
resulted in an improvement of the Charpy toughness with no deterioration of the
strength[65,92] whereas bainite resulted poor toughness[66].
4.3.3 U-O pipe forming and Bauschinger effect
During U-O pipe forming, the plate materials are subjected to different cyclic strains,
depending on the location along the circumference and in the wall. Typical examples
of cyclic strain at locations 180˚ and 30˚ from the longitudinal seam are illustrated in
figure 4.1[93].
At the 180˚ location where mill tensile test specimens are usually taken, the outer
layer receives a tensile force during U-bending, a compressive force during
O-bending, a compressive force during shrinking, and a tensile force during
expansion. At the same time, the inner layer is subjected to a cyclic strain of
compression, tension, compression and tension.
The total process from U-bending to expansion is not a simple work-hardening
process but is actually very complex. On the other hand, during the flattening of a
curved pipe section for tensile specimens, the Bauschinger effect and work hardening
occur in the outer and inner layer respectively.
- 38 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 4 Microstructure and mechanical properties
Figure 4.1 Schematic stress-strain curves for the outer (top) and inner (bottom)
material during the U-O pipe forming process, with (left) at 180º and (right) at 30º
from the welding line[93].
The Bauschinger effect is a characteristic material behavior that is highly dependent
on testing conditions[94]. It results in the lowering of a proof-stress value after a
- 39 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 4 Microstructure and mechanical properties
previous single uniaxial initial loading in the opposite direction. When hot rolled strip
is converted into ERW line pipe, the pipe forming and sizing strains can significantly
modify the pipe yield strength by virtue of the Bauschinger phenomenon. Various
steels have different responses to the Bauschinger effect due to different stress-strain
curves. Steels with a yield plateau have a Bauschinger strain equal to that
corresponding to the strain equal to the yield elongation[95].
The Bauschinger effect is largely controlled by the carbon content and, to a
considerably smaller degree, by the manganese content (figure 4.2). Grain size
appears to have a minor influence, while the influence of residual-stress conditions is
strong (figure 4.3)[94].
Figure 4.2 The change of the Bauschinger Figure 4.3 The Bauschinger effect in
effect factor with carbon and manganese
micro-alloyed steel. The upper two curves
content[94].
are for steels with 0.2% C, 0.4% Mn,
unalloyed or alloyed respectively with Al,
V or Nb. The lower two curves are for
low-pearlite steels with less than 0.1% C,
2% Mn, and alloyed with Mo, Nb and
Ti[94].
- 40 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 4 Microstructure and mechanical properties
Work-hardening and the Bauschinger effect occur during the pipe-forming process
and during the flattening of the tensile test pieces before the tensile test. In the
pipe-forming process, work hardening by pipe expansion is more important than the
Bauschinger effect, while the reverse is true during sample flattening. The
ring-expansion test is used in measuring work hardening. The pipe after forming, has
a considerably higher yield strength than the plate, which indicates that work
hardening has taken place during pipe forming[96].
- 41 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 5 Background of current South Africa line pipe production
CHAPTER 5 BACKGROUND OF CURRENT SOUTH AFRICA LINE PIPE
PRODUCTION
5.1 Line pipe steel composition of Mittal Steel (South Africa)
The chemical composition of the current Mittal Steel (SA) line pipe steel is provided
in table 5.1.
Table 5.1 Typical chemical composition of the current 11 mm line pipe steel of Mittal
Steel, (wt%)
C
Si
Mn
P
S
Cr
Ni
Mo
Cu
Al
0.066
0.258
1.583
0.011
0.004
0.021
0.007
0.001
0.007
0.037
V
Nb
Ti
Sn
B
0.062
0.037
0.017
0.001
0.0002
- 42 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 5 Background of current South Africa line pipe production
5.2 Parameters of the hot rolling process at Mittal Steel (SA)
The parameters of hot rolling of line pipe steels at Mittal Steel are shown in table 5.2.
Reheating
tempe`rature, ºC)
Table 5.2 The parameters of the hot rolling process at Mittal Steel
Pass No
R1
R2
R3
R4
R5
R6
F4
F5
F6
1552
1150
1076
1053
1042
1008
938
915
896
1050
1008
969
915
896
879
Force (tons)
in
Temperature
(ºC)
1200
1150
out
Gauges
(mm)
F1
F2
1860
F3
in
240
195
160
120
85
60
40
28.87
21.02
15.76
13.51
out
195
160
120
85
60
40
28.87
21.02
15.76
13.51
11.70
0.21
0.20
0.29
0.34
0.35
0.40
0.32
0.32
0.29
0.15
0.14
16.1
22.6
20.9
26.2
1.8
2.7
3.3
3.9
Strain/pass, ε
Total strain
1.79
1.22
Strain rate
9.4
(s-1)
Roll speed,
~1.5
V (ms-1)
Inter-pass time, t, (s)
Inter-pass cooling
rate (ºCs-1)
~1.5
~1.5
~1.5
~1.5
10
10
10
10
10
3
3
3
3
3
2.7
1.8
1
1.5
40 ºCs-1 —for 6mm of the final thickness of strip
Cooling
finishing
rate
(ºCs-1)
after
20 ºCs-1 —for 11.5mm
50 ºCs-1 —for 5mm
~2.4%
t/D (Thickness/Diameter)
NB: F3-dummy for rolling
- 43 -
1.3
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 5 Background of current South Africa line pipe production
5.3 Typical microstructures and existing developments within Mittal Steel for line
pipe steel
The optical microstructure of the current 11 mm wall thickness line pipe steel for
Mittal Steel is shown in figure 5.1. This is the alloy that was used for a major part of
the pipe-line for large-scale gas transportation from Mozambique to Secunda in a 2.5
meters diameter pipe line. The microstructure is a mixture of polygonal ferrite,
acicular ferrite and pearlite.
Figure 5.1 The optical microstructure of cast #521031, Mittal Steel line pipe
Smaller thin walled (about 6 mm thickness) pipelines may be used in future within
South Africa for smaller scale gas distribution and reticulation to consumers. The
current steel produced by Mittal Steel tends to have a slightly high YS/UTS ratio of
0.93, which is on the maximum limit of the specification of the American Petroleum
Institute (API) of 0.93 for an application involving pipelines. The current line pipe
steel consists typically of about 0.06% carbon with micro-alloying elements of
titanium, niobium and vanadium (generally less than 0.05% for each) and is produced
either via the Electric Arc Furnace (EAF) or Basic Oxygen Furnace (BOF) route.
- 44 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 5 Background of current South Africa line pipe production
5.4 The hypothesis for this study
The objective of this study is to establish the relationship between micro-alloying
elements, the microstructure and deformation in austenite on the ratio of YS/UTS and
other mechanical properties. Therefore, a study is to be undertaken on the austenite to
acicular ferrite transformation with particular emphasis on the kinetics of the acicular
ferrite formation and as affected by the above process variables. The experimental
research would include a redesign of the chemical composition, dilatometer analyses,
simulation of the controlled rolling process on a Gleeble 1500 (initially the work was
started on a Gleeble 1500TM model but halfway through the study, the Gleeble was
upgraded to a Gleeble 1500D DSI), microstructural observations, tensile tests, SEM
and TEM investigations, etc.
5.4.1 Design of the chemical compositions of the investigated alloys
The V-content in the steel was decreased in this study because it only contributes to
dispersion hardening as V(C,N) in ferrite. Its dispersion hardening is less than that of
Nb in steels.
Niobium is a very important micro-alloying element in line pipe steels and formed
one of the main-alloying elements considered in this study. It contributes to dispersion
hardening (NbC in ferrite), promotes acicular ferrite formation (reduction of pearlite)
and raises the Tnr (“pancake” of austenite). The niobium concentration was increased
to about 0.045%wt (which is more than the 0.037% used currently by Mittal Steel).
Titanium is another micro-alloying element that retards austenite grain growth during
reheating. The reheating temperature may be as high as 1225 ºC due to undissolved
TiN. If titanium binds the free nitrogen in the steel, more niobium will be available in
the ferrite to precipitate as NbC and will increase the precipitation hardening.
Titanium can also control the shape of sulphide inclusions (TiS). Accordingly, the
titanium addition was kept at levels of 0.017 to 0.022% in this study.
- 45 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 5 Background of current South Africa line pipe production
Molybdenum contributes to phase transformation hardening and can be used instead
of V in the hardening of steels. It might benefit the increase in volume fraction of
acicular ferrite and M/A islands that are useful to lower the ratio of YS to UTS.
Molybdenum can diminish the Bauschinger effect during pipe forming. Thus,
molybdenum was added to the experimental alloys in this study. The molybdenum
additions considered here were 0.10%, 0.15% and 0.25%, respectively.
Carbon was slightly decreased to 0.05% C in this study for the purpose of improving
the weldability of steel.
Summarising the target analysis above, the chemical compositions of the experimental
alloys were designed as follows (Table 5.3):
Table 5.3 Design of chemical composition ranges of alloys that were investigated (in
wt%)
C
Si
Mn
S
P
V
Nb
Ti
0.05
0.2-0.25
1.0-1.2
<0.005
<0.01
0
0.045
0.022
N
Cu
Ni
Al
Cr
Ca
Mo
0.006
0.007
0.007
0.03
0.02
0.002
0.1-0.25
5.4.2 Design of the controlled hot rolling process
The austenite grain size of the current Mittal Steel alloy, reference alloy #6 (Mo-free)
was found to be 57 and 63 µm at 1225 ºC for 60 and 120 minutes respectively.
Therefore, the reheating temperature of 1225 ºC was chosen for this study to provide
almost complete dissolution of the niobium carbonitrides to achieve maximum
precipitation hardening, but little austenite grain size coarsening.
- 46 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 5 Background of current South Africa line pipe production
In this study, the finish rolling temperature was maintained at about 870 ºC, i.e. above
the Ar3 and, therefore, with no deformation in the (α+γ) two-phase region.
The temperature range for rough rolling was chosen to be from about 1190 to 1000 ºC
in this study, which is above the non-recrystallisation temperature Tnr. The total or
cumulative true strain in the rough rolling stage was chosen to be about 1.4 with
individual pass strains of more than 0.2.
The final temperature for the finish rolling stage was between 840 to 870 ºC. The
finish rolling temperature in this stage was maintained above the Ar3 temperature. The
deformation for this stage was, therefore, carried out in the austenite
non-recrystallisation region. The total strain in the finish rolling stage was about 0.54
in order to accumulate enough rolling strain within the austenite grains for the
subsequent ferrite transformation, leading to a pass strain of more than 0.2 as well.
The initial and final thicknesses of the ingot and plate for laboratory hot rolling were
planned to be 43 and 6 mm respectively, with a total heavy reduction of 86%.
The cooling rate after finish rolling has a significant effect on the subsequent
microstructure of the line pipe steel. Rapid cooling is useful to increase the volume
fraction of acicular ferrite and this contributes to good mechanical properties. It
results in a continuous stress-strain curve, decreases the Bauschinger effect during
pipe-forming and leads to a low ratio of YS/UTS. In this study, various cooling rates
and with/without prior deformation in the austenite on the Gleeble were used to
establish the effect of these parameters on the ratio of YS/UTS. For this purpose, the
experimental challenge was how to measure the volume fraction of acicular ferrite?
The coiling temperature is also very important to the degree of precipitation hardening
of line pipe steels. If the coiling temperature is too high, the precipitates become too
- 47 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 5 Background of current South Africa line pipe production
coarse. At low temperatures acicular ferrite may form, but the temperature cannot be
too low because the precipitation requires diffusion, and in this study two coiling
temperatures of 600 and 575 ºC were chosen. These are also the previous and the
current coiling temperatures respectively used by Mittal Steel in their 11 mm strip
steel for line pipe.
In this study a hypothesis that the acicular ferrite or an optimised mixture of acicular
ferrite and polygonal ferrite, is the most suitable microstructure for decreasing the
ratio of YS/UTS of steels, was, therefore, tested.
A series of tests on the Gleeble were also carried out to study the effect of cooling rate,
coiling temperatures and deformation prior to transformation on the ratio of yield
strength to ultimate tensile strength.
- 48 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
CHAPTER 6 EXPERIMENTAL PROCEDURES
This chapter describes the experimental procedures for the investigation, including the
alloy design, the melting and casting of the ingots, the hot-rolling process, the testing
of the austenite grain size and the presence and identification of undissolved particles,
the determination of the non-recrystallisation temperature, the determination of the
strain-free and the strain affected CCT diagrams and the determination of the
mechanical properties, etc. Distinguishing between the acicular ferrite and polygonal
ferrite in the microstructures of the samples was done by TEM through shadowed
carbon extraction replicas and thin foils. The chemical compositions of the
experimental alloys were typical of commercial line pipe steels with Nb-V-Ti
micro-alloying elements with the current line pipe steel from Mittal Steel as the
reference steel.
6.1 Alloy design
As indicated earlier and briefly summarised here again, the design of the experimental
alloys was based on the considerations set out below.
1. The chemical compositions of HSLA line pipe steels are normally low in carbon
and contain some micro-alloying elements that may be only one, or a combination
of any two or three of the micro-additions (vanadium, niobium and titanium). A
low carbon level was selected for improving the weldability and toughness, to
provide less pearlite and more effective dissolution of niobium during reheating
that will increase the precipitation hardening of these steels.
2. Niobium has strong dispersion hardening characteristics due to the formation of
NbC in ferrite. It promotes the transformation to acicular ferrite that can be
beneficial to lower the ratio of yield strength to ultimate tensile strength of these
steels. Niobium also causes refinement of the austenite grains during the rolling
process by raising the non-recrystallisation temperature (Tnr). Accordingly, a little
more niobium was considered than in the current Mittal Steel alloy while
- 49 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
vanadium was reduced or left out, which decreases the non-recrystallisation
temperature.
3. Titanium can retard the austenite grain growth during reheating of slabs due to the
presence of TiN particles. As described above, considering the stoichiometric ratio
of Ti to N (3.4 /1) and preventing MnS stringer inclusions, (which requires Ti/S in
TiS of 1.5/1), the ideal titanium addition was calculated from equation (2.1).
4. The addition of molybdenum in Nb-containing steels can improve transformation
hardening (increased volume fraction of acicular ferrite and M/A islands), can
provide grain refinement and precipitation hardening. It also greatly suppresses or
delays the formation of polygonal ferrite and pearlite[4]. Additions of molybdenum
were, therefore, considered instead of the usual vanadium for enhancing the
strength of the steels. In steels with molybdenum, the stress-strain curve of the
as-rolled plate is usually continuous, without an upper yield points[33]. This may
provide control of the Bauschinger effect and contribute to an increase in yield
strength from plate to pipe.
Thus, the newly designed alloys that were made up, all had the same low carbon,
niobium and titanium levels but were made with and without vanadium, and also had
varying amounts of molybdenum.
6.2 The melting of the experimental alloys
Sections from the currently produced line pipe steel from Mittal Steel were used as
feed stock material for the melting of the new alloys in a 50 kg vacuum induction
melting furnace at Mintek. The liquid steel was cast into ingots of 43 × 66 × 235 mm.
The chemical compositions of the five new alloys are listed in table 6.1. The Mittal
Steel line pipe steel is included in the table as a reference.
- 50 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
Table 6.1 Chemical compositions of the experimental alloys, in wt.%
Alloy #
1
2
3
4
5
6(Mittal Steel)
Cu
0.02
0.03
0.03
0.02
0.02
0.007
Al
0.004
0.055
0.054
0.055
0.065
0.037
C
0.05
0.06
0.05
0.05
0.05
0.066
Co
0.006
0.009
0.008
0.009
0.01
--
Si
0.29
0.25
0.23
0.24
0.25
0.26
Mn
1.21
1.29
1.05
1.31
1.14
1.583
V
0.06
<0.005
<0.005
<0.005
<0.005
0.062
P
0.014
0.018
0.019
0.019
0.019
0.011
Nb
0.055
0.05
0.051
0.052
0.055
0.037
S
0.011
0.01
0.011
0.011
0.011
0.004
Ti
0.017
0.019
0.019
0.019
0.021
0.017
Cr
0.07
0.05
0.04
0.05
0.05
0.021
Sn
-0.003
0.003
0.003
0.003
0.001
Ni
0.07
0.04
0.04
0.04
0.04
0.007
Mo
0.01
0.09
0.09
0.12
0.22
0.001
B
Free-N
0.0001 0.0068
0.0003 0.0035
0.0002 0.0032
0.0003 0.0032
0.0003 0.0027
0.0002
--
6.3 The effect of reheating temperature and soaking time on the austenite grain
sizes
The samples from the cast ingots were machined into cubes of about 10 × 10 × 10 mm.
Two methods were used to process these samples in order to measure the austenite
grain size. The first was to reheat the samples at temperatures of 1150, 1200, 1225 and
1250 °C, respectively, and then quench them into water. The samples were etched
using many different etchants (see table 6.2), but the results revealed that these
etchants were not suitable to reveal the prior austenite grain boundaries for the alloys
studied. The carbon content of the alloys was probably too low for this. The more
successful technique was a modified McQuaid-Ehn method by carburising the
samples after reheating in argon, in-situ within the austenite region in a dry carbon
monoxide gas atmosphere at 927 ºC for up to 5 hours directly after reheating at the
above four austenitisation temperatures, i.e. without going through the ferrite
transformation. Pro-eutectoid cementite formed on the prior austenite grain
boundaries during very slow cooling from the carburisation temperature at 927 ºC
down to 690 ºC at which temperature the sample was removed from the furnace. Thus,
- 51 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
the cementite layers indicated the sites of the original austenite grain boundaries,
which then became easy to measure. The carburising process is illustrated in figure
6.1. After carburisation, the samples were polished and etched in a 2% Nital solution
and then the original austenite grain sizes were measured by the mean linear intercept
method[97].
Table 6.2 The composition of the etchant solutions
Solution Number
#1
#2
#3
#4
#5
#6
Chemical
Quantity
Picric acid
1g
Hydrochloric acid
5 ml
Ethanol alcohol
95%
FeCl3
1g
H2O
100 ml
FeCl3
1g
Hydrochloric acid
5 drops
H2O
100 ml
Sodium bisulphite
34 g
H2O
100 ml
Hydrochloric acid
40 ml
Sulphuric acid
10 ml
H2O
50 ml
10% aq. Oxalic acid
28 ml
H2O2(30%)
4 ml
H2O
80 ml
- 52 -
University of Pretoria etd – Tang, Z (2007)
Temperature, °C
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
Figure 6.1 Schematic of the modified McQuaid-Ehn carburising process of the
samples directly after reheating.
6.4 Measuring the presence and composition of undissolved particles
During reheating, any undissolved small particles will retard the austenite grain
growth. Before entering the first rough rolling pass, a fine austenite grain size is
beneficial to the later strength and toughness of the steels. The quantity and the size of
the undissolved particles are related to the reheating temperature. Austenitisation
temperatures ranging from 1150 to 1250 °C were employed, together with soaking
times of 15 to 120 minutes for this part of the investigation. The samples were
quenched into water or were fast cooled in helium gas. The details of these treatments
are listed in table 6.3 below. The samples were mechanically ground and polished
after the treatment, then were lightly etched in a 2% Nital solution without an
apparent visible optical microstructure of the matrix, and vacuum coated with carbon.
The shadowed carbon extraction replicas were similarly made, also after light etching
in 2% Nital, then the vacuum application of Au-Pd shadowing at an angle between 20
to 40º before vertical coating of the carbon. The size of the particles was measured on
the micrographs obtained from the transmission electron microscope (TEM) and their
volume fraction calculated using the following equation[98]:
f =
π
6
N s ( x + σ Al )
2
2
Al
- 53 -
(6.1)
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
σ
where
x
2
2
Al
=
n ∑ x Al − (∑ x Al ) 2
(6.2)
n(n − 1)
is the planar arithmetic mean of the particle diameter;
Al
f is the volume fraction of particles;
Ns is the total number of particles intersecting a unit area;
σ
x
2
Al
Al
is the standard deviation of the particle size distribution;
is the diameter of a particle;
n is the total number of particles measured.
Table 6.3 Temperatures and soaking time of the treatment
for undissolved particles
Temperature (ºC)
Time (min)
1150
15
60
120
1200
15
60
120
1225
15
60
120
1250
15
60
120
6.5 Non-recrystallisation temperature (Tnr)
The finishing temperature of rough rolling is associated with the non-recrystallisation
temperature (Tnr) and heavy reductions must take place within this recrystallisation
region at temperatures higher than the Tnr in order to obtain a fine recrystallised
austenite grain size at the start of finish rolling below the Tnr. This is beneficial to high
strength and good toughness of the steels. Accordingly, the non-recrystallisation
temperature is an important parameter that should be considered when the hot rolling
process is designed. Many researchers have studied the recrystallisation of
austenite[99-107]
and
some
mechanisms
- 54 -
have
been
proposed[108-110].
The
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
non-recrystallisation temperature may be a function of the parameters[99,111-113] of the
rolling process, such as pass strain[114,115], strain rate and inter-pass time[48], and may
also depend on the content of micro-alloying elements[116], etc. The hot torsion test
has been widely used to simulate industrial hot rolling processes[117-121]. In this study,
the recrystallisation behaviour of the steel was investigated during multi-pass
compression deformation on a Gleeble hot working facility on cylindrical samples of
8 mm in diameter and 15 mm in length that were machined from the as-cast ingots.
6.5.1 Testing schedule for the determination of the Tnr
Reheating temperatures should be high enough to dissolve all the precipitates (mainly
the Nb precipitates, but except the TiN) because the micro-alloying elements affect
the non-recrystallisation temperature. The reheating temperature for Nb precipitates
can be determined from the following equation[122,123]:
12 ⎞
6770
⎛
log[ Nb]⎜ C + N ⎟ = 2.26 −
14 ⎠
T
⎝
(6.3)
The calculated solution temperatures for Nb(C,N) in the experimental alloys are listed
in table 6.4.
Table 6.4 Calculation equilibrium Nb carbonitride solution temperature
Alloy number
Solution temperature of Nb(C,N) (ºC)
1
1145
2
1149
3
1179
4
1130
5
1137
- 55 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
Laasraoui[124] reported that the niobium carbonitrides remained undissolved at
reheating temperature of 1100 ºC for an 0.04% Nb steel. Thus, a maximum reheating
temperature of 1225 ºC was selected here.
Samples of alloy #6 (the current Mo-free Mittal Steel alloy) were heated to 1225 °C at
a heating rate of 100 °Cmin-1 and held at this temperature for 15 minutes. The
multi-pass compression tests were carried out using the test parameters shown in
tables 6.5 and 6.6. Pass strains, ranging from 0.15 to 0.32, and inter-pass times
ranging from 4 to 50 seconds were employed, at a constant strain rate of 1 s-1. In two
particular tests, the inter-pass time and pass strain were held constant. Another test of
strain rate ranging from 0.1 to 2.22 s-1, was also employed at a constant pass strain of
0.2 and constant inter-pass time of 8 seconds (see table 6.7). Such a multi-pass
compression testing schedule is illustrated schematically in figure 6.2 below.
Table 6.5 Testing parameters for Tnr at strain ranging from 0.15 to 0.32
Inter-pass time (s)
8
8
8
8
8
8
Strain rate (s-1)
1
1
1
1
1
1
0.15
0.2
0.24
0.28
0.30
0.32
Strain/pass
Table 6.6 Testing parameters for Tnr at inter-pass times ranging from 4 to 50 seconds
Inter-pass time (s)
4
6
8
15
20
30
35
40
50
Strain rate (s-1)
1
1
1
1
1
1
1
1
1
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
Strain/pass
Table 6.7 Testing parameters for Tnr at strain rate ranging from 0.1 to 2.22 s-1
8
8
8
8
8
8
8
8
8
Pass strain, ε
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
Strain rate (s-1)
0.1
0.47
0.9
1.22
1.38
1.67
1.80
2.0
2.22
Inter-pass time (s)
- 56 -
University of Pretoria etd – Tang, Z (2007)
Temperature, °C
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
100 ºCmin-1
Time, min
Figure 6.2 Schematic schedule employed in the multi-pass compression tests for the
Tnr.
6.5.2 The determination of the non-recrystallisation (Tnr)
A typical set of curves of the true flow stress versus true strain from a multi-pass
compression test is illustrated in figure 6.3.
160
866℃
interpass time:4s pass strain:0.20
strain rate:1/s
140
895℃
s tre s s ,M P a
120
928℃
100
80
1108℃
1064℃
1018℃
1153℃
60
40
20
0
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
strain
Figure 6.3 The curves of flow stress versus strain in a multi-pass compression test on
alloy #6.
The mean flow stress of each pass was calculated from the following equation and the
flow stress-strain curve (see figure 6.3):
- 57 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
−
σ=
1 ε
∫ε σdε
ε b −ε a
b
(6.4)
a
The non-recrystallisation temperature was determined from the relationship between
the mean flow stress (MFS in MPa) of each pass and the inverse temperature (1/T in
K-1) of the compression deformation, as illustrated in figure 6.4. This typical curve is
divided into two stages by a slope change in the two straight line sections. In the
lower slope stage (which corresponds to a high temperature deformation), full
recrystallisation takes place during the pass because there is no strain accumulation
and the increase in the mean flow stress is solely due to the decrease in temperature of
the inherent strength of a well annealed microstructure. However, in the higher slope
stage (deformation below the Tnr), there is only partial dynamic recrystallisation, or no
recrystallisation at all, indicating that the strain is accumulated from pass to pass, and
the mean flow stress increases more rapidly with decreasing temperature[125]. The
intersection of two straight lines provides the Tnr.
140
M e an flo w stre s s, M P a
120
i n t e r p a s s t i m e: 4 s p a s s s t r a i n : 0.2 0
s t r a i n r a t e : 1/s
100
80
60
40
20
0
0.65
0.7
0.75
0.8
1000/T
0.85
0.9
(K-1)
Figure 6.4 Determining Tnr from the mean flow stress in MPa versus the inverse pass
temperature in K, during a multi-pass compression test on alloy #6.
- 58 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
6.6 CCT-diagram
The alloys used for the CCT diagrams, were the as-cast alloy #5 (with 0.22% Mo) and
alloy #6 (Mo-free) (Mitall Steel reference alloy). The chemical compositions of alloys
#5 and #6 were listed in table 6.1. Samples of with a size of a 15 to 16 mm long
cylinder with a diameter of 7 mm were used for the strain affected CCT tests, and a 8
to 9 mm long cylinder with a diameter of 7 mm for the strain-free CCT test were used.
The temperatures of phase transformations were measured by the C-strain facility on
the Gleeble which measures the change in diameter of the sample during cooling for
the strain affected CCT tests.
6.6.1 The Ac1 and Ac3 test
The Ac1 and Ac3 are the important critical equilibrium temperatures for starting and
completion of the austenitisation transformation during phase transformation of low
carbon hypoeutectoid steels (less than 0.77% C). The determination of the Ac1 and Ac3
temperatures were made on a single-LVDT THETA dilatometer facility. Phase
transformations are normally associated with a non linear volume change in the
temperature range of the transformation. The linear thermal expansion or contraction
of the samples, therefore, takes place in a manner that allows the subtraction of the
linear relationship between dilatometry and temperature from the non-linear
transformation portions. A schematic sketch of the dilation with temperature is
represented in figure 6.5[126]. Figure 6.6 shows the schematic determination of the Ac1
and Ac3 temperatures of steels on the heating curve of dilation versus the testing
temperature. In order to completely dissolve all the Nb-alloyed precipitates in this
study, e.g. the Nb(C,N), the reheating temperature was chosen as 1225 ºC for 15
minutes with a heating or cooling rate to and from this temperature of 3 ºCmin-1 for
equilibrium conditions. Samples for the THETA dilatometer were cylinders with a 7
mm diameter and a 10 mm length. The chamber containing the samples was kept at a
vacuum of 10-4 torr, to prevent any significant oxidation of the samples.
- 59 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
Figure 6.5[126] Schematic dilation as a
Figure 6.6[126] Schematic determination
function of testing temperature.
of the Ac1 and Ac3 temperatures on the
heating curve.
6.6.2 CCT diagram without prior deformation
The continuous cooling transformation (CCT diagram) without prior deformation was
also made on the THETA Dilatometer. The reheating rate was taken as 100 ºCmin-1 up
to 1225 ºC and held for 15 minutes at this temperature for the purpose of complete
dissolution of the Nb-precipitates. The samples were subsequently cooled down to
980 ºC at a rate of 5 ºCs-1 and held for 5 minutes at this temperature before finally,
cooling down to 25 ºC at the various almost linear cooling rates of 0.1, 0.2, 0.5, 1, 2, 5,
8, 10 and 20 ºCs-1, respectively. The sample chamber was evacuated before cooling
after 980 ºC and the cooling media was either flowing argon or helium gas, depending
on the required cooling rate. The chamber of the THETA dilatometer and the
schematic test schedule are illustrated in figures 6.7 and 6.8, respectively. After
cooling, samples were polished and etched in 2% Nital and the microstructures
examined with an Olympus PMG3 optical microscope. A combination of the optical
micrographs and the cooling curves of dilation with the test temperature, were used to
determine the phase transformation temperatures. The CCT diagram was then
constructed from the various curves of temperature on a linear scale with the test time
on a log scale and the phase transformation temperatures indicated on these.
- 60 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
Thermocouple
Sensor
sample
Figure 6.7 Dilatometer chamber
Figure 6.8 Schematic schedule of the test for the CCT diagram on the THETA
Dilatometer.
6.6.3 Strain affected CCT diagram
The stain affected CCT diagram with prior deformation, was carried out on Gleeble
1500D DSI hot simulator. The reheating rate was also 100 ºCmin-1 up to 1225 ºC and
held for 15 minutes at this temperature for the complete dissolution of precipitates in
an argon atmosphere. The samples were then cooled down to 860 ºC at 10 ºCs-1, and
held for 5 minutes at this temperature. The samples were then compression deformed
with a strain of 0.6 (45% reduction below the Tnr) at 860 ºC at a strain rate of 0.5 s-1.
Samples were finally cooled down to room temperature after deformation, at the
various linear cooling rates of 0.1, 0.2, 0.5, 1, 2, 5, 8, 10, 20 and 40 ºCs-1, respectively.
The cooling media was also argon (for a slow cooling rate) or helium gas (for a rapid
- 61 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
cooling rate). The steps for preparing the micrographs were the same as for the
strain-free CCT diagram above. The chamber of the Gleeble hot simulator and the
schematic test schedule are illustrated in figures 6.9 and 6.10, respectively. The
C-gauge facility of the Gleeble was used to measure the dilation of the sample during
cooling after compression deformation.
Thermocouple
C-gauge
Piston
sample
Figure 6.9 Chamber of the Gleeble 1500D DSI
Figure 6.10 Schematic schedule of the test for the strain affected CCT diagrams on the
Gleeble.
- 62 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
6.7 The thermo-mechanical process
6.7.1 Cooling unit
A specially constructed cooling unit was used to cool the laboratory melted ingots
after hot rolling in a controlled manner. The cooling unit consists of nozzles, a water
pump, control valves, and a water box etc. The experimental arrangement is shown in
figure 6.11. The coolant is fed with compressed air and sprayed onto the cooling
samples through nozzles. The linear distance between the nozzles was about 70 mm,
so that uniform cooling of samples of about 100 × 300 mm could be attained. The
cooling rate, for instance, was 21 °Cs-1 using fresh water at 24 °C as the coolant and
with compressed air spraying of the water with an air pressure of 580 MPa. A higher
figure of 47 °Cs-1 was achieved with a 10% NaCl aqueous solution instead of water.
Cooling Box
Air gauge
Air switch and
Water Switch
Water gauge
Water out
Water container
Pump Switch
(a)
Water pump
Water in
(b)
Water valve
Water in
Water in
(c)
(d)
Figure 6.11 Experimental arrangement of the cooling unit for controlled cooling: (a)
overall view, (b) controller for mixing of gas and water, (c) valves for the nozzles and,
(d) cooling spray in the chamber from the spraying jets.
- 63 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
6.7.2 Hot rolling process of the laboratory ingots
The schedule for the laboratory hot rolling of the ingots, i.e. the pass strain, the total
reduction and inter-pass times etc., were controlled as far as possible to optimise the
process for austenite grain refinement[114,115]. The experimental casts were hot rolled
on a two-stand laboratory hot rolling mill equipped with 10 inch rolls. The ingots of a
machined 43 × 90 × 66 mm size for the alloys #1 to #6 were reheated and hot rolled to
6 × 100 × 300 mm plates with the final thickness of 6 mm chosen to simulate a
possible future reduced gauge for Mittal Steel from their current 11 mm strip.
6.7.2.1 Reheating before laboratory hot rolling
From the results of the study of the effects of temperature and time on the austenite
grain size, a reheating temperature of 1225 °C and a time of 60 minutes were taken.
This is quite safe as it has been reported that austenite grains will not coarsen unduly
before an austenitisation temperature of about 1250 °C is reached in V-Nb-Ti micro
alloyed steels[35].
6.7.2.2 Rough rolling of the laboratory hot rolling
As indicated before, the metallurgical function of roughing is to refine the coarse
austenite grains after the reheating and soaking to achieve the finest possible
dynamically recrystallised grains before entering the finish rolling below the Tnr. Pass
strains should, therefore, be at least 0.2 or higher to promote Dynamic
Recrystallisation (DRX) during rough rolling to produce finer recrystallised grains.
The starting temperature for rough rolling of the laboratory cast ingots was between
1148 and 1190 °C and five passes in this roughing stage were undertaken, with a total
roughing strain of 1.43.
- 64 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
6.7.2.3 Finish rolling of the laboratory hot rolling
As before, the objective of this stage is to accumulate sufficient rolling strain within
the austenite grains to promote a finer ferrite transformation after rolling. Ferrite
nucleation sites are, therefore, greatly multiplied in number and a very fine ferrite
grain size can be generated during the subsequent controlled cooling[6]. In this study
three passes were used with a total strain of 0.54 and the finish rolling temperature
was maintained between 857 and 865 °C.
6.7.2.4 Cooling rate after laboratory finish rolling
The cooling rate (Vc) and the finishing temperature (Tc) of the accelerated cooling
after finish rolling, are important parameters of the thermo-mechanically controlled
processing for the experimental alloys to achieve their optimum strength and
controlling the Tc and Vc may lead to the control of the precipitation of carbonitrides
during the accelerated cooling[9]. A rapid cooling rate helps to promote finer
precipitation of Nb(C,N), ferrite grain refinement and acicular ferrite formation. The
latter is preferred for a low ratio of yield strength to ultimate tensile strength[4], and it
also avoids the development of any pearlite in the microstructure. A rapid cooling rate
of 47 °Cs-1could be achieved in the cooling box by using an aqueous solution of 10%
NaCl for alloys #1 to #5 while a rate of 39 °Cs-1could be obtained for the Mo-free
alloy #6.
6.7.2.5 Simulation of coiling after laboratory hot rolling
The coiling temperature will influence the effectiveness of Nb-carbonitride but
especially V(C,N) precipitation in the ferrite, thus controlling precipitation
strengthening of these steels. If the coiling temperature is high, the precipitates will
become coarser during coiling. A coiling temperature of 600 °C was selected for the
study of these alloys as this was also the temperature used in the past by Mittal Steel
before they recently lowered it to 575 ºC. After reaching 600 °C in the cooling box,
the small plates were placed in a furnace at 600 °C until the temperature became
- 65 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
uniform throughout the plate. Thereafter they were well insulated for 24 hours using a
generous covering of vermiculite for a slow cool simulating the actual coiling process
in the plant.
6.7.2.6 Hot-rolling process curve
The hot-rolling process schedule is illustrated schematically in figure 6.12. The
Temperature, °C
symbols R and F in the figure refer to the rough and finishing passes, respectively,
Figure 6.12 Schematic schedule of the hot rolling process on the experimental alloys.
6.8 The identification of acicular ferrite
The optimum microstructure of line pipe steels appears to be one that contains
acicular ferrite and some polygonal ferrite. Acicular ferrite is very different from
polygonal ferrite and bainitic ferrite. Polygonal ferrite, as the name implies, has
polygonal boundaries with a carbon concentration in the ferrite that is almost uniform
together with a lower dislocation density. Acicular ferrite, on the other hand, is
characterised by fine non-equiaxed or interwoven nonparallel ferrite laths, which have
various grain sizes and are arranged in a random manner[78]. Some M/A
(martensite/austenite) islands and a high density of dislocations in-between the AF are
common[11]. The carbon content in the M/A islands is higher than that in the
surrounding
matrix
and
accordingly,
- 66 -
these
islands
are
carbon-enriched
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
martensite/austenite islands whose formation may be attributed to the partitioning of
carbon during the transformation to acicular ferrite and the post-transformation of
carbon-enriched austenite. During accelerated cooling followed by the coiling process,
part of the carbon-enriched austenite transforms to martensite and the remaining
austenite will coexist with the martensite[73]. The accepted transformation model of
acicular ferrite is a mix of diffusion and shear transformation[ 4,9,11,43].
Acicular ferrite contributes to a low yield strength and a higher tensile strength due to
the lower carbon in the matrix and the M/A islands having a high density of
dislocations and this, therefore, leads to a lower YS/UTS ratio. It was found initially
that it was easy to confuse acicular ferrite with polygonal ferrite under an optical
microscope when samples were etched in 2% Nital because the grain boundaries
between them do not become clearly etched. Therefore, one of the key aspects of the
experimental techniques in this study was how to distinguish between acicular ferrite
and polygonal ferrite in low carbon, Nb-Ti micro-alloyed steels. Various techniques
were initially attempted to identify the acicular ferrite, including the use of optical
microscopy,
Scanning
Electron
Microscopy
(SEM),
Transmission
Electron
Microscopy (TEM) using carbon extraction replicas with and without shadowing and
finally, thin foil TEM samples.
6.8.1 Observation with optical microscopy and by SEM
Samples taken after hot rolling that were etched in a 2% Nital solution, were
examined with an Olympus PMG3 optical microscope, both JEOL 5800LV and 6000F
high resolution SEM, respectively. Acicular ferrite could not be distinguished from
polygonal ferrite on the micrographs after etching with a light (5 seconds) or a deep
etch (up to 120 seconds) under the optical microscope and SEM.
- 67 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
6.8.2 Observation of replicas by TEM
6.8.2.1 Preparing replicas without shadowing
Firstly, the polished samples were deeply etched from 30 to 60 seconds in a 2% Nital
solution and thoroughly washed to remove all loose etching debris from the surface.
Carbon coating was done under vacuum on the etched surface of the samples and the
coating separated from the sample’s surface in a mixture of 7 ml nitric acid and 75 ml
ethanol before floating-off in distilled water.
6.8.2.2 Preparing shadowed carbon extraction replicas
The etched surface of the samples was first shadowed through vacuum evaporation
with a gold-palladium alloy before the vertical deposition of carbon. Deep etching
from 30 to 60 seconds was employed before the shadowing and the shadowing angle
with respect to the surface was varied from 20° to 40°. The carbon was then coated
vertically onto the shadowed layer. The technique used for separating the carbon film
from the sample was the same as described above for the replicas without shadowing.
6.8.3 Thin foil TEM samples
Thin foil samples were used to further distinguish between acicular ferrite and
polygonal ferrite and to also validate the results of the shadowed replicas in the TEM.
Thin slices of material were cut by electro-discharge in a wire-cutting machine in
order to reduce any adverse effect of deformation on the dislocation density in the
samples that could have formed from mechanical cutting. The disc size was 3 mm in
diameter and 0.6 mm in thickness and the procedure for preparing the thin foil
samples was as follows:
•
The original cylinder of material with dimensions of 3 mm diameter and 15
mm length was machined in a lathe;
•
Five discs of 3 mm diameter and 0.6 mm thickness were cut from the cylinder
with the electro-discharge wire-machine;
•
The disc samples were carefully and lightly polished to between 50 to 80 μm
- 68 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
in thickness on fine grinding paper;
•
The final thinning was done by electro-polishing with a solution of 1.25 l
acetic acid, 0.08 l perchloric acid and 0.7 g chromium oxide at 25 °C.
All samples for replicas and thin foils were examined in a Philips CM200 TEM,
operated at 60 or 200 KV, respectively.
6.9 Test of subsize samples on the Gleeble with various cooling rates, coiling
temperatures and prior deformation
In order to study how some parameters of the controlled hot rolling process influence
the YS/UTS ratio of steels, a series of experiments were planned. These parameters
include cooling rates, coiling temperatures and deformation in the austenite.
The cooling rate after hot rolling has a strong effect on the fraction of acicular ferrite
in Nb-Ti micro-alloyed steels and particularly, accelerated cooling after finish rolling
helps to increase the volume fraction of acicular ferrite which may, in turn, influence
the YS/UTS ratio. The coiling process controls the precipitation of particles in the
matrix of these steels which affects dispersion hardening. The reduction during hot
rolling also has an effect on the CCT diagram of these steels as well. A series of tests
were designed for this purpose.
6.9.1 Hot rolling plates for Gleeble samples
Alloys #3 and #6 were selected for these tests. Casts of these alloys, firstly, were hot
rolled to 6 mm thickness (the parameters of the hot rolling for plates are shown in
Appendix H). Two types of preliminary samples were machined: the first type A to a
rectangular size of 6 ×10 × 100 mm was used to study the effect of cooling rates and
coiling temperature without deformation (figure 6.13-(a)), while the other type B
shown schematically in figure 6.13-(b), was used to study the effect of prior
compression in austenite, from there the shorter gauge length. At this stage, the gauge
- 69 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
lengths had not been machined into the rectangular samples, hence their being called
“preliminary”.
Figure 6.13 Preliminary samples on the Gleeble of (a) type A and (b) type B
6.9.2 Tests on the Gleeble
The first two groups of type A samples were tested at different cooling rates after
austenitisation in which the cooling rates ranged from 1 to 51 ºCs-1 for the Mo-free
alloy #6 (sample #A124) and 1 to 54 ºCs-1 for the 0.09% Mo alloy #3 (sample #AF3F).
The process graph is shown schematically in figure 6.14.
- 70 -
University of Pretoria etd – Tang, Z (2007)
Temperature, ºC
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
Time
Figure 6.14 Graph of the heating and cooling process on the Gleeble for samples
#A124 (the Mo-free alloy #6) and sample #AF3F (the 0.09% Mo alloy #3).
The second two groups of type A samples were subjected to a 60 minutes coiling
simulation at 575 (sample #B113) and 600 ºC (sample #A113) for the Mo-free alloy
#6, respectively, after cooling in Gleeble. The process graphs are illustrated
Temperature, ºC
Temperature, ºC
schematically in figures 6.15.
(b)
(a)
Time
Time
Figure 6.15 Graphs of the heating and cooling cycles in the Gleeble on the Mo-free
alloy #6 for samples (a) #A113 and (b) #B113.
The last group, which consisted of type B samples (sample #TEN06 from the Mo-free
alloy #6) was tested on the Gleeble with a prior 45% reduction in the austenite (only
33% reduction below the Tnr) before cooling and coiling simulation at 575 ºC for 60
minutes. Figure 6.16 shows the test process.
- 71 -
University of Pretoria etd – Tang, Z (2007)
Temperature, ºC
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
Time
Figure 6.16 Graph of heating, cooling and deformation process in the Gleeble for
sample #TEN06 (the Mo-free alloy #6).
6.9.3 Tensile tests
All of these rectangular and preliminary samples that were subjected to the above tests
on the Gleeble, were thereafter machined with their gauge lengths to subsize tensile
samples of type A (see figure 6.17-(a)) for the tensile test which was done on an
INSTRON-8500 Digital Control tensile testing machine for the first four groups of
samples and the type B (see figure 6.17-(b)) on a smaller instrumented Hounsfield
tensile testing machine.
- 72 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
(a)
(b)
Figure 6.17 Tensile test samples of (a) type A and (b) type B. T is original thickness of
the plates and 6 mm for the as-rolled alloy and the Gleeble samples, respectively.
6.10 Test of mechanical properties on the as-hot rolled alloys
The specimens for the tensile tests from the hot rolled plates were cut from the middle
of the rolled plates in the longitudinal and transverse directions and were machined to
the subsize tensile of type A (see figure 6.17-(a)). The tensile tests were carried out at
room temperature on an INSTRON-8500 Digital Control tensile testing machine with
an initial cross-head speed of 0.25 mm min-1 until the elongation of 0.5 mm was
reached and then a second cross-head speed of 2 mm min-1 thereafter.
- 73 -

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project