Acicular Ferrite

Acicular ferrite is formed in the interior of the austenitic grains typically by direct
nucleation on the inclusions, resulting in randomly oriented short ferrite needles.

From: Treatise on Process Metallurgy: Industrial Processes, 2014

Related terms:

Austenite, Bainite, Ferrite, Microstructure, Titanium, Cementite, grain boundary,

Microstructures

View all Topics

Learn more about Acicular Ferrite

Acicular Ferrite
SirHarshad Bhadeshia Tata Steel Professor of Metallurgy, SirRobert Honeycombe
Emeritus Goldsmiths' Professor of Metallurgy, in Steels: Microstructure and Prop-
erties (Fourth Edition), 2017

7.3 Mechanism of transformation

Acicular ferrite and bainite are in many respects similar in their transformation
mechanisms. Their microstructures differ in detail because bainite sheaves grow as a
series of parallel platelets emanating from austenite grain surfaces, whereas acicular
ferrite platelets nucleate intragranularly at point sites so that parallel formations
of plates cannot develop. The nucleation site in the latter case is smaller than the
thickness of the plate, so that the inclusion is normally engulfed by the plate of ferrite
which it stimulates.

The growth of both bainite and acicular ferrite causes an invariant-plane strain
shape deformation with a large shear component (Fig. 7.4). Consequently, plates
of acicular ferrite cannot cross austenite grain boundaries, because the coordinated
movement of atoms implied by the shape change cannot in general be sustained
across grains in different crystallographic orientations. The lattice of the acicular
ferrite is therefore generated by a deformation of the austenite, so that the iron and
substitutional solutes are unable to diffuse during the course of transformation. If is
not therefore surprising that the concentrations of substitutional alloying elements
are unchanged during the growth of acicular ferrite.

Figure 7.4. Interference contrast micrograph showing the surface relief caused
when a metallographically polished sample of steel is transformed to acicular ferrite
(courtesy of Strangwood).

The deformation which changes the austenite into acicular ferrite occurs on particu-
lar planes and directions, so that the ferrite structure and orientation are intimately
related to that of the austenite. It follows that plates of acicular ferrite, like bainite,
must without exception have an orientation relationship with the austenite. This is
not necessarily the case when a transformation occurs by a diffusional mechanism,
because a grain of ferrite can easily grow into any adjacent grain of austenite with
which it happens to come into contact.

During isothermal transformation, the acicular ferrite reaction stops when the
carbon concentration of the remaining austenite makes it impossible to decompose
without diffusion. This implies that the plates of acicular ferrite grow supersaturated
with carbon, but the excess carbon is shortly afterwards rejected into the remaining
austenite. This of course is the incomplete reaction phenomenon described in
Chapter 6 for bainite, where the austenite never reaches its equilibrium composition
since the reaction stops at the curve of the phase diagram, Fig. 7.5 [8]. The obvious
conclusion is that acicular ferrite cannot form at temperatures above the bainite-start
temperature, and this is indeed found to be the case in practice.
Figure 7.5. Data from experiments in which the austenite is transformed isothermal-
ly to acicular ferrite, showing that the reaction stops when the carbon concentration
of the austenite reaches the curve (courtesy of Strangwood).

There are many other correlations which reveal the analogy between acicular ferrite
and bainite. For example, the removal of inclusions by vacuum arc melting, without
changing any other feature, causes an immediate change in the microstructure from
acicular ferrite to bainite [9]. The same effect can be obtained by increasing the
number density of austenite grain nucleation sites relative to intragranular sites. This
can be done by refining the austenite grains to obtain a transition from an acicular
ferrite microstructure to one which is predominantly bainitic (Fig. 7.6).
Figure 7.6. The transition from bainite to acicular ferrite because of the change in
the ratio of the number density of nucleation sites at austenite grain boundaries
to inclusions. (a)→(b) acicular ferrite promoted by introducing inclusions. (c)→(d)
acicular ferrite promoted by an increase in the austenite grain size. (e)→(f ) acicular
ferrite promoted by saturating austenite grain boundaries with allotriomorphic
ferrite. Figures courtesy of S.S. Babu.

The opposite phenomenon, in which an inclusion-containing steel with bainite can

be induced to transform into an acicular ferrite microstructure is also observed. This
can be done by rendering the austenite grain surfaces ineffective as nucleation sites,
either by decorating the boundaries with a thin layer of inert allotriomorphic ferrite
(Fig. 7.7) or by adding a small amount of boron (30 ppm). The boron segregates
to the boundaries, thereby reducing the boundary energy and making them less
favourable sites for heterogeneous nucleation [10]. In general, any method which
increases the number density of intragranular nucleation sites relative to austenite
grain boundary sites will favour the acicular ferrite microstructure.
Figure 7.7. The change from a bainitic (a) to an acicular ferrite (b) microstructure
when the austenite grain boundaries are eliminated as nucleation sites by decoration
with inert layers of ferrite (courtesy of Babu).

> Read full chapter

Kinetics of bainite transformation in

steels
A. Borgenstam, M. Hillert, in Phase Transformations in Steels: Fundamentals and
Diffusion-Controlled Transformations, 2012

13.3.2 Lengthwise growth of plates

The lengthening rate of acicular ferrite illustrates the difference between the two
hypotheses. It was expected to be very much higher than observed experimentally
if it is controlled by the diffusionless mechanism. This discrepancy was explained
by Oblak and Hehemann (1967) through the suggestion that diffusionless growth
will build up strains that eventually will stop the growth and the result will be only
a subunit. Further growth will have to wait for a new nucleation event. The rate
controlling factor should then be the rate of nucleation of new subunits and their
final lengths, not the very high growth rate of each subunit. The subunits have been
identified with a substructure observed by electron microscopy of thin foils. Different
estimates have been made and it has, for instance, been mentioned that a typical size
of a subunit may be 0.2 × 10 × 10 μm3 (Bhadeshia and Waugh, 1982). Smaller sizes
have also been reported. However, alternative observations have also been made.
Figure 13.9 reproduces an in-situ observation of gradual growth of bainite made by
Nemoto (1974) using TEM. The data show that, if there had been a succession of
rapidly growing steps, their lengths must have been less than 0.1 μm which is much
shorter than expected. This may be taken as an indication that stepwise growth does
not occur.

13.9. Measured length of bainite at 380 °C as a function of time by Nemoto (1974)

showing that the growth of bainite is continuous in a steel with 0.51 mass%C and
9.1 mass% Ni.

Growth rate equations have been proposed for both hypotheses. For the diffusionless
hypothesis, expressions have been suggested that have two or three fitting para-
meters, one governing the rate of nucleation of a new subunit at the tip of the
preceding one and another representing the length of the subunits (e.g. Matsuda
and Bhadeshia, 2004). It can thus be used for representing experimental data after
fitting the parameter values to selected experimental information on growth rates.
From the diffusional hypothesis, expressions have been derived for the edgewise
growth rate of a plate based on the rate of diffusion of carbon in austenite and with
various degrees of sophistication (Zener, 1946; Hillert, 1957; Trivedi, 1970; Liu and
Ågren, 1989). These were recently compared (Hillert et al., 2003) and objections were
raised against the treatment by Trivedi. The comparison is illustrated in Fig. 13.10
and it is demonstrated that the sharpness of the edge, expressed as a radius of
curvature, , is not defined in any of the treatments. Hillert (1960) obtained an
equation of the following form by modifying Zener’s equation and applying Zener’s
criterion of maximum growth rate,

13.10. Lengthening rate at a supersaturation of Ω = 0.5 according to three different

treatments (Hillert et al., 2003).

[13.2]

where D is the diffusion coefficient for carbon in austenite, is the specific

interfacial energy, is molar volume of ferrite and x represents the various mole
fractions of carbon and, in particular, x o is the initial carbon content.

> Read full chapter

NON IRON POWDER NI-BEARING

ELECTRODES FOR LOW TEMPERA-
TURE ENVIRONMENTS
J. Koivula, J.H. Rogerson, in Welding in Energy-Related Projects, 1984

7016 CIL
Increasing welding energy decreases the volume fraction of acicular ferrite so that
the amount of M-A-C ferrite increases instead. That means a reduction in the
amount of high angle boundaries at the expense of low angle boundaries, less
energy is then needed for fracture. Transition temperature is also increased with
increasing welding energy. It can thus be concluded that increasing acicular ferrite
content in 7016 weld metal will lower the transition temperature.

The big difference in transition temperature or toughness value at a given temper-

ature (Fig. 5) can be explained by the usual arguments of high dilution percentage
and severe strain ageing conditions at the root compared to subsurface passes.

Although Table 4 shows an increase in lamellar products in the microstructure with

increasing heat input it should be noted that no clear side plate structures can be
seen in any of the specimens under optical microscopy.

The best impact toughness and lowest transition temperature were obtained with the
lowest heat input. Similar microstructures were achieved when welding was carried
out at -50°C (Table 4) with Q = 3 kJ/mm. Very low heat input was unfortunately not
employed in low temperature welding.

> Read full chapter

Nucleation of Ferrous Solid–Solid

Phase Transformations at Inclusions
M. Enomoto, in Encyclopedia of Materials: Science and Technology, 2001

6 Influence of Particle Size

Barbaro et al. (1989) reported that acicular ferrite is formed on inclusions larger
than 0.4μm in size in the HAZ of low-carbon steels. It was observed that the critical
size of TiN particles for the nucleation of ferrite decreased with the amount of
undercooling in a nitrogen-added low-carbon steel (Fine Inclusions Forum 1998).
Similar observations had been made in weld metals (Abson 1989). A few explanations
of the size effects can be offered. In the first, the area of interphase boundaries, and
thus the number of nucleation sites, simply increases with the size of the particle.
Second, the activation energy for nucleation decreases, though weakly, with the
radius of the substrate particle (Lee et al. 1992). Furthermore, the volume of the
solute-depleted zone increases with the size of the particle. On the other hand, the
magnitude of the stress around inclusions due to mismatch in thermal contraction
does not depend primarily on the particle size.

Other transformations are also stimulated by inclusions. Manganese sulfide and

vanadium carbide have been observed to induce intragranular formation of pearlite,
which was ascribed to the formation of a manganese-depleted zone. Graphite was
reported to be nucleated on BN particles during the graphitization of a high-carbon
steel. The interfaces between graphite and BN were shown to be coherent (Iwamoto
et al. 1998). This may be an example in which epitaxy plays an important role in
the initiation of transformation at interphase boundaries. Microstructure control by
inclusions involves a number of fundamental issues concerning phase transforma-
tion in solids. The importance of understanding the operating mechanisms for the
development of this technique is emphasized.

> Read full chapter

Phase transformations in microalloyed

high strength low alloy (HSLA) steels
R.C. Cochrane, in Phase Transformations in Steels: Diffusionless Transformations
High Strength Steels Modelling and Advanced Analytical Techniques, 2012

6.7.1 Acicular ferrite steels

Given the attractive combination of mechanical properties provided by a fully aci-
cular ferrite microstructure, some weld metals can exhibit yield strengths of about
690 MPa coupled with 54 J impact transition temperature below −100°C and, conse-
quently, several attempts have been made to promote this mode of transformation
either directly to optimise mechanical properties in plate or strip steel compositions
or to modify the weld HAZ microstructure. In weld metals where such structures
can be readily obtained, there is little doubt as to the effectiveness of the inclusion
population which is some 10 times that of a commercial steel (weld metal oxygen
contents range from ~200 to 450 ppm compared to <10–50 ppm for steels). While
the effectiveness of the various inclusion substrates continues to be debated, some
features are well recognised. There needs to be at least one set of crystallographic
planes with a close match to ferrite; ‘TiO’ (Mills et al., 1987; Lee, 1994), galaxite (MnO
Al2O3) (Mills et al., 1987), TiN (Ilman, 2001; Ilman et al, 2011) or BN (Ilman, 2001)
have all been proposed. To form acicular ferrite in a steel, three conditions appear to
be necessary. Firstly, Ti deoxidation appears to be beneficial; the role of Ti appears
to be due to epitaxial ferrite nucleation on ‘TiO’-rich inclusions or TiN particles
or inclusions. Secondly, there needs to be a critical volume fraction and size of
inclusions for ferrite nucleation to be effective (Grong et al., 1995). Finally, displacive
transformation to ferrite must take place below or close to the Bs temperature
(Barbaro et al., 1989). Although weld metals are characterised by relatively coarse
austenite grain size and therefore the ratio of austenite grain boundary surface area
to total inclusion surface area approaches 1, this does not necessarily preclude the
formation of acicular ferrite in steels provided the intragranular nucleation rate is
enhanced by adding active substrates, such as ‘TiO’ (Shim et al., 1999), MnS (Tomita
et al., 1994) (Fig. 6.42), or rare earth oxides (Thewlis, 2006).

6.42. Examples of ferrite plates nucleating at ‘TiO’ inclusions in a simulated heavy

steel section. After Al-Hajeri et al. (2006).

Several attempts have been made to produce AF microstructures in steels or in the

weld HAZ by doping with suitable inclusions of an appropriate size; between 0.25
and 2.5 μm appears to be optimal from studies on welds and/or by lowering the
transformation temperature by alloying (Al-Hajeri et al., 2006). Ensuring sufficient
oxide particles are present means that the oxygen content must be above some
minimum value, estimated to be 50–80 ppm. At this level, Al deoxidation is not
desirable and to some degree, this also mitigates against using Nb since Nb2O5 is
formed and would reduce O below that desired at steelmaking temperatures. Hence,
V additions seem to be more appropriate. Whilst some success in promoting AF has
been achieved on a laboratory scale for Nb steels, prospects for scaling to commercial
production appear limited by the need to control, firstly, the state of deoxidation as
this requires a departure from conventional steelmaking; secondly, the cooling rate
needed to form inclusions of an appropriate size is unobtainable with conventional
continuous casting, except by the thin slab process route. Although some commer-
cial steels have been promoted as having improved HAZ properties as a result of
promoting acicular ferrite formation (Hart and Mitchell, 1995), only V treated steels
appear to give rise to a fully AF microstructure (He and Edmonds, 2002; Jun et al.,
2006) under HAZ conditions. However, a recent study by Garcia and co-workers
(Garcia-Mateo et al., 2008a, 2008b) strongly implicates VN formed in austenite as
a nucleant for AF (Fig. 6.43). However, given that little or no precipitation would
be expected during a typical weld HAZ thermal cycle, such an interpretation seems
contrary to the work of He and Edmonds (2002) and the balance of studies on welded
structures. It seems probable that all these observations can be reconciled in terms
of a sequence of segregation, clustering followed by precipitation with the dominant
mechanism being dependent on steel composition and/or processing.
6.43. Acceleration of acicular ferrite formation in a V-N steel after a deliberate heat
treatment to precipitate VN in austenite after holding at 900°C for 2500 s (b), com-
pared to simply cooling from the re-heating temperature of 1125°C for 120 s. The
solid lines correspond to the start of transformation. Note the broadening of the AF
range and depression of martensite start temperature, Ms, on the CCT diagram (a),
compared to that in (b). Steel composition, 0.08%C-1.46%Mn-0.25%V-0.018%N.

Adapted from Garcia-Mateo et al. (2008a, 2008b).

> Read full chapter

INFLUENCE OF NITROGEN CON-

TENT ON DEPOSITED WELD METAL
NOTCH TOUGHNESS
C. Bonnet, A. Gaspard Angeli, in Welding for Challenging Environments, 1986

CONCLUSION
The different experiments carried out for this study show that, in a multipass deposit
where the structure of the as-welded areas is mainly acicular ferrite, the regions with
the lowest notch toughness are those reheated to a temperature below Ac1. On the
contrary, the regions reheated to a temperature slightly above Ac3 show the highest
notch toughness. The difference between the transition curves for these respective
regions is around 30°C.

An increase in the nitrogen content of the weld deposit leads to a large decrease in
the notch toughness properties of all the regions of a multipass weld. This decrease
is greater when small quantities of Niobium and Vanadium are present in the weld
deposit.

> Read full chapter

Bainite
W.T. ReynoldsJr., in Encyclopedia of Materials: Science and Technology, 2001

3 Diffusional Mechanism
Following a systematic investigation of microstructures in a series of alloy steels,
Hultgren proposed a diffusional hypothesis for bainite formation. He suggested
bainite evolved with the nucleation of probainitic (acicular) ferrite, then nucleation
of cementite particles on the ferrite/austenite boundaries, followed by the cementite
being surrounding by growing ferrite, and eventually nucleation of new cementite
particles. In contrast to bainite, pearlite was suggested to develop by nucleation
of a cementite plate, followed by an adjacent ferrite plate, and then coordinated
lengthening of the plates.

Electron microscopy observations of bainite and pearlite have led to a mechanistic

basis for this approach. The resulting, more detailed, model articulates the circum-
stances under which bainite and pearlite form during a general eutectoid reaction
(Lee et al. 1988). This model relates the internal morphology of a eutectoid product
to two factors: (i) the relative growth rates of the eutectoid phases, and (ii) the
rate at which the slower-growing phase renucleates on the boundary between the
faster-growing phase and the parent (Fig. 1). An attractive feature of the model is its
ability to explain a variety of bainite morphologies (Fig. 2) in terms of simple factors
that, at least in principle, can be measured or estimated. For example, the model
has been applied to interpret transitions in internal morphology and in aggregate
shape of evolving eutectoid structures (the external morphology) as a function of
transformation temperature and carbon concentration in a series of Fe–C–2 wt.%
Mn alloys (Spanos et al. 1990a, b). By considering how the nucleation and growth
rates of ferrite and cementite change with temperature and carbon concentration, it
is possible to explain the relatively sharp transition between upper and lower bainite
and the more gradual changes from upper and lower bainite to nodular bainite.

Figure 1. Variation in eutectoid microstructures with ratio of the product growth

rates, G /G , and nucleation rate of the slower growing product, – J* , when the
volume fraction of the slower growing phase is small (afterLee et al. 1988).

Figure 2. Schematic illustrations of various bainite morphologies. The white con-

stituent represents the majority eutectoid phase (e.g., ferrite) and the dark con-
stituent represents the minority phase (e.g., cementite). (a) Nodular bainite, (b)
columnar bainite along a prior matrix grain boundary, (c) a sheaf of upper bainite
laths, (d) lower bainite, (e) grain boundary allotriomorphic bainite, and (f ) inverse
bainite (after Spanos et al. 1990b).
Hultgren’s original model of bainite formation also proposed an explanation for
transformation stasis (Hultgren 1947). Two types of ferrite were envisioned: ortho-
ferrite, which differs from austenite in carbon concentration and in the concentra-
tions of substitutional alloying elements, and paraferrite (paraequilibrium ferrite),
which differs from austenite in carbon concentration but not in alloy content. It
was suggested that the Bs temperature is the temperature below which paraferrite
appears, and transformation stasis just below the Bs temperature is caused by a slow
reaction rate associated with the upper portion of the “C” curve of paraferrite on
the TTT diagram. This explanation for the Bs temperature and transformation stasis
does not explain subsequent experimental results. Partitioning data and measured
growth kinetics from a variety of Fe–C–X alloys indicate the transition to paraferrite
occurs at temperatures well above the range of observed Bs temperatures (Aaronson
et al. 1988). Also, the amount of ferrite found during transformation stasis in
Fe–C–Cr and Fe–C–Mo alloys is well below the metastable fraction expected to form
under paraequilibrium (Enomoto and Tsubakino 1991).

A promising framework for understanding transformation stasis was built on the

effects of alloying elements on ferrite growth. Measured transformation kinetics in
ternary alloys (e.g., Reynolds et al. 1990a, b) have shown transformation stasis is not a
general property of bainite, but rather a characteristic of particular alloying elements
and concentrations. In general, alloys that exhibit transformation stasis also tend
to be those in which ferrite growth is significantly slower than predicted under
paraequilibrium (Shiflet and Aaronson 1990). The ability of specific solute species to
slow ferrite growth is widely attributed to a solute drag effect, or solute drag-like
effect, depending on how one describes the origin of the drag force (Reynolds et
al. 1990a, Purdy and Brechet 1995, Enomoto 1999). Efforts to ascertain the degree
of solute adsorption at ferrite/austenite boundaries have yielded somewhat mixed
results, although the presence of molybdenum accumulations at these boundaries
in Fe–C–Mo alloys now appears certain (Aaronson et al. 1990, Aaronson et al. in
press). There is, however, a large body of microstructural and kinetic evidence for
the solute drag effect in steels, and investigations into interactions between solutes
and growth interfaces are likely to be a productive area for research.

The central assumption needed to explain transformation stasis below Bs is that the
solute drag effect must be strong enough to slow markedly or stop ferrite growth.
This is feasible when sufficient concentrations of carbon and of an effective solute
are present, the reaction temperature is relatively low, and ferrite boundaries have
migrated far enough to pick up the dragging solute. Under these circumstances,
renucleation of ferrite on immobilized ferrite/austenite boundaries (sympathetic
nucleation) becomes feasible if sufficient undercooling is available. The tem-
perature below which copious sympathetic nucleation begins corresponds to the
kinetic Bs temperature. The degenerate ferrite morphology found below the bay
temperature results from a change from relatively smooth, ledgewise migration of
ferrite boundaries to irregular “start–stop” migration during which ferrite nucleates,
grows until stopped by solute drag, and is then forced to renucleate again (Reynolds
et al. 1990a).

> Read full chapter

Weld Microstructures
SirHarshad Bhadeshia Tata Steel Professor of Metallurgy, SirRobert Honeycombe
Emeritus Goldsmiths' Professor of Metallurgy, in Steels: Microstructure and Prop-
erties (Fourth Edition), 2017

13.2.2 As-deposited microstructure

The microstructure obtained as the weld cools from the liquid phase to ambient
temperature is called the as-deposited or primary microstructure. Its major compo-
nents include allotriomorphic ferrite, Widmanstätten ferrite, and acicular ferrite (Fig.
13.3). There may also be some martensite, retained austenite or degenerate pearlite.
These latter phases occur in very small fractions, and are known by the collective
term microphases. Bainite, consisting of sheaves of parallel platelets, is not generally
found in well-designed welding alloys. Instead, acicular ferrite is induced to nucleate
heterogeneously on non-metallic inclusions [7].

Figure 13.3. (a) Schematic illustration of the essential constituents of the primary
microstructure in the columnar austenite grains of a steel weld deposit. (b) Scanning
electron micrograph of the primary microstructure of a steel weld (courtesy of Rees).
The terms , w and a refer to allotriomorphic ferrite, Widmanstätten ferrite and
acicular ferrite, respectively.

In practice, the gap between the components to be joined often has to, for thick
steels, be filled by a sequence of several weld deposits. These multirun welds have
a complicated microstructure (Fig. 13.4). The deposition of each successive layer
heat treats the underlying microstructure. Some of the regions of original primary
microstructure are reheated to temperatures high enough to cause the reformation
of austenite, which during the cooling part of the thermal cycle transforms into a
different microstructure. Other regions may simply be tempered by the deposition
of subsequent runs. The microstructure of the reheated regions is called the reheated
or secondary microstructure.

Figure 13.4. The macrostructure of a multirun weld, made by sequentially depositing

a number of beads in each of the 7 layers (courtesy of Kamellia Dalaei, ESAB AB).

> Read full chapter

HIGH-STRENGTH, HIGH-FRAC-
TURE TOUGHNESS SUB-
MERGED-ARC WELD FOR ARCTIC
LINE PIPE
F. Hamad, ... S. Bhole, in Welding for Challenging Environments, 1986

WELD DESIGN
Producing weld metal with high strength is not a challenge. However, the attempt
to combine high toughness with high strength is not an easy task.
Thus, the toughness of the weld metal can be optimized by minimizing the vol-
ume fraction of inclusions, such as sulphides and silicates, and by developing a
fine-grained, cleavage resistant microstructure, such as acicular ferrite.

Consumables
The flux was selected with the objective of minimizing the effect of welding para-
meters, such as voltage and current on the mechanical properties and chemistry
of the weld. A fused, semi-neutral flux was chosen. Manganese and silicon oxides
were below levels where a significant transfer of manganese and silicon to the
weld can take place. Those levels were 5% MnO and 18% SiO2 (ASM - Metalworking
and Manufacturing Processes Series, B 002, 212). The flux also contained high
percentage of CaF2 (15%) and CaO (9.5) to ensure a high desulphurization potential.
The other major constituent of the flux TiO2 (20%) which delivers Ti and TiO2 to
the weld. Both are known to be instrumental in nucleation of the fine-grained,
interlocking microstructure (Ito, Nakanishi, 1976). The flux had a grain size of 32
× 200 mesh (according to Tyler Screen sieves) in order to stabilize the arc and enable
high current applications.

The wire chosen was of a low carbon, manganese-molybdenum chemistry as shown

in TABLE 1.

TABLE 1. Wire Composition

Weight Percent
C Mn S P Si Mo
.09 1.8 .015 .006 .69 .45

The selection of such wire-flux combination was aimed at producing a low carbon,
high hardenability, titanium containing weld.

Welding Conditions
The submerged arc welding was done on a spiral pipe mill using DC-AC tandem
setup under the condition shown in TABLE 2.

TABLE 2. Welding Parameters

1st Pass (0D Bead) 2nd Pass (ID Bead)

Current (Amp) 950 500 900 500
Voltage 32 34 31 34

The welding speed was 27.5 mm/s (65 in./min.).

Base Metal
The pipe was X-70 arctic grade with 1067 mm O.D. x 14 mm W.T. (42 in. O.D. x .55
in. W.T.). The base metal was calcium-treated control-rolled microalloyed steel with
typical chemistry as shown in TABLE 3.

TABLE 3. Base Metal Chemistry

Weight Percent
C Mn S P Si Cb Mo Al N O
.09 1.7 .006 .01 .20 .06 .20 .03 .01 .01

Microstructure and Toughness

It is known that weldments composed of ferrite-bainite microstructures exhibit an
inverse relationship between strength and toughness as shown schematically in Fig.
1.

Fig. 1. Diagram showing inverse relationship between toughness and strength.

There are a number of factors which influence the fracture toughness of a steel.
At temperatures above transition, fracture involves microvoid formation and so
the amount, size and distribution of inclusions are the major factors. At lower
temperatures, however, the toughness is strongly influenced by the grain size and
the type of microstructure, as shown in Fig. 2.
Fig. 2. Main factors which influence toughness.

> Read full chapter

Joint integrity
Norman Bailey BMet. CEng. FIM. FWeldI, in Weldability of Ferritic Steels, 1994

Weld metal toughness

As with other systems, ferritic steel weld metals normally have their toughness
reduced as their strength is increased. However, there is a major exception to this
rule. If conditions (as discussed in Chapter 1) are suitable for the development of
acicular ferrite, the toughness of as-deposited ferritic steel weld metal will improve
if the amount of alloying elements is increased until coarse primary ferrite has been
replaced by fine-grained acicular ferrite. Within this range, strength and toughness
both increase. However, with further alloying, toughness begins to fall owing to
the introduction of bainite or martensite when the microstructure is completely, or
almost completely, one of acicular ferrite.

The benefits of having a fine acicular ferrite microstructure survive PWHT, but not
normalising, either as a separate heat treatment or by succeeding weld passes.
However, the fine ferrite produced by normalising by succeeding weld passes (Fig.
1.11f) ensures the retention of good toughness.

When it is necessary for a weld metal to have a yield strength in excess of about
650 N/mm2, it is not possible to maintain an acicular ferrite microstructure, because
the degree of alloying required is such that bainitic or martensitic microstructures
result. Except in the Cr:Mo weld metals (particularly 2.25Cr: 1Mo) which are bainitic,
 it is advantageous to aim for a martensitic microstructure of low carbon content. In
a fully martensitic weld metal, nickel is the only element known to be capable of
improving toughness. However, other factors, such as a high degree of refinement
in multipass welds and low inclusion contents are also known to be helpful.

The only way to increase the strength of fully martensitic weld metals in the as-weld-
ed condition is to increase the carbon content – and this poses severe problems in
regard to avoiding cracking, particularly hydrogen cracking. If heat treatment is to
be carried out, then weld metal strength can be increased in some low carbon weld
metals by ageing treatments, such as are applied to maraging steel welds.

> Read full chapter

ScienceDirect is Elsevier’s leading information solution for researchers.

Copyright © 2018 Elsevier B.V. or its licensors or contributors. ScienceDirect ® is a registered trademark of Elsevier B.V. Terms and conditions apply.