Materials Science & Engineering A 720 (2018) 248–256

Contents lists available at ScienceDirect

Materials Science & Engineering A

journal homepage: www.elsevier.com/locate/msea

Kinetics and microstructural change of low-carbon bainite due to vanadium T

microalloying
⁎
Fateh Fazelia, , Babak Shalchi Amirkhiza, Colin Scotta, Muhammad Arafinb, Laurie Collinsc
a
CanmetMATERIALS, Natural Resources Canada, Hamilton,ON, Canada L8P 0A5
b
CanmetMATERIALS, Natural Resources Canada, Calgary, AB, Canada T2L 2A7
c
EVRAZ INC. NA, P.O. Box 1670, Regina, SK, Canada, S4P 3C7

A R T I C L E I N F O A B S T R A C T

Keywords: Noticeable strength increases, up to 120 MPa, have been obtained in hot strip and linepipe bainitic steels by
Linepipe steels microalloying with vanadium. However, there is no consensus on the mechanisms proposed in the literature to
Strengthening explain this effect. As such, this study has investigated the root cause and various strengthening contributions of
Phase transformations vanadium additions to bainitic linepipe steels. Two variants of a X100 linepipe steel with identical compositions
Vanadium
but different vanadium contents, i.e. a reference steel with residual amount of vanadium and a microalloyed
In-situ compression
variant with 0.06 wt% vanadium, have been subjected to dilatometric study to determine the kinetics of bainite
formation during continuous cooling and cooling-coiling scenarios. Further, a novel in-situ hot compression
experiment was developed to determine the strength of fresh bainite (prior to coiling) and its softening during
simulated coiling at 450 °C. Various microstructural features of the bainite, including the morphology, dis-
location density and lath size, as well as the presence of microalloyed carbonitrides have been quantified by TEM
characterization. It was found that the kinetics of the bainite transformation during cooling and the tempering of
fresh bainite upon coiling simulations were substantially influenced by vanadium microalloying. Vanadium in
solution shifted the bainite transformation to lower temperatures (by 30–40 °C) during cooling at 1–50 °C/s,
refining the lath structure and increasing the dislocation density of bainitic ferrite. The V-added steel demon-
strated a higher strength of fresh bainite (about 50 MPa) compared to the reference alloy. This superior strength
has been explained explicitly by accounting for the contributions from the increased dislocation density and the
refined lath structure of V-added variant. No evidence of fine VC or VN precipitation occurring during coiling at
450 °C was found. Further, the complex TiNbV(CN) particles that were observed were too coarse and too scarce
to make any strengthening contribution. The fresh bainite of both X100 variants tends to soften rapidly upon
coiling at 450 °C; about 110 MPa strength drop was noticed for the reference alloy. The implications of the
findings for alloy design and processing of advanced linepipe steels are discussed.

1. Introduction
The most advanced line-pipe grades are produced using low-carbon
Mn-Mo based compositions in conjunction with Ti-Nb-V microalloying
additions. Furthermore, tightly specified thermomechanical controlled
processing (TMCP) routes followed by on-line accelerated cooling
(ACC) have emerged as common practice in the past few decades [1].
The superior performance of modern linepipe steels (X100-X120) stems
from the replacement of the traditional ferrite-pearlite aggregates with
a bainitic microstructure which forms mainly during accelerated
cooling of hot bands and is further established during coiling. The
dominant strengthening mechanisms are attributed to the sub-micron
lath structure of bainite (lath thicknesses < 0.5 µm) together with high
dislocation densities of the order of 1014 m/m3 [2]. However, this
highly dislocated BCC structure with its large surface area of low-angle
lath boundaries is not stable at conventional coiling temperatures
(450–600 °C) and is susceptible to abrupt softening; this can give rise to
losses of up to ~150 MPa in yield strength.
The benefits of vanadium additions and the respective best pro-
cessing practices in ferritic pipes (< X65) and conventional HSLA steels
have been well characterised (e.g. [3]). Vanadium precipitation in
austenite and to a greater extent in ferrite have been exploited as a key
advantage to process various long and flat rolled products [3]. The
microalloying of modern bainitic steels with vanadium is not common
since the required fast cooling and low-temperature coiling of these
grades prevents significant precipitation of vanadium. There have been

⁎
Corresponding author.
E-mail address: Fateh.Fazeli@Canada.ca (F. Fazeli).

https://doi.org/10.1016/j.msea.2018.02.042
Received 20 November 2017; Received in revised form 9 February 2018; Accepted 10 February 2018
Available online 13 February 2018
0921-5093/ Crown Copyright © 2018 Published by Elsevier B.V. All rights reserved.
F. Fazeli et al. Materials Science & Engineering A 720 (2018) 248–256

several recent studies showing that vanadium additions can strengthen

low to medium carbon steels with bainitic microstructures [4–8]; the
reported gains in tensile strength range between 50 and 150 MPa, de-
pending upon the processing parameters and the composition. How-
ever, the underlying mechanism is unclear; whether it is related to
vanadium precipitation, clustering or due to a solid solution effect on
phase transformation and tempering. For example, to suppress the
softening of fresh bainite during coiling the addition of V to bainitic hot
strips [4,8] and to X100 plate steels [5] has recently been proposed. It
was found in a laboratory Fe-0.04C-1.6Mn-1Cr-0.3Mo-0.024Al-0.08V-
0.012N grade that fresh bainite had enhanced resistance to tempering
in the temperature range of 400–550 °C, and exhibited up to 100 MPa
additional strength in the as-coiled state when compared to a V-free
reference alloy [8]. Whereas, the strength of the steels in the direct-
cooled state to room temperature was found to be similar in the V-free
and V-added alloys, implying that vanadium did not influence the Fig. 1. Different thermomechanical routes by dilatometer: (A) to develop CCT diagram,
nature of bainite transformation [8]. The beneficial effect of V has been (B) to simulate cooling and interrupted coiling and (C) in-situ compression to determine
attributed by some authors to the pinning of dislocations by V(C,N) softening of fresh bainite upon coiling.
precipitates resulting in sluggish recovery of dislocation tangles [4], but
BAHR 805 dilatometer was used to determine the kinetics of austenite
no direct evidence was provided. In another study, the addition of
decomposition during continuous cooling and cooling-coiling scenarios.
0.063 wt% vanadium was clearly found to enhance the yield and tensile
Three distinct thermomechanical routes were carried out as depicted in
strength (by up to 18%) in a laboratory X100 linepipe steel without any
Fig. 1. A common austenite reheating and conditioning practice was
compromise of toughness [5]. A detailed EBSD analysis confirmed a
used for all the three routes. The specimens were soaked for 5 min at
higher fraction of low angle boundaries and smaller grain structure of
1200 °C to simulate reheating of slabs prior to rolling and to ensure the
bainite in the V-added steel compared to the reference X100 plate,
dissolution of Nb and V carbonitrides, followed by sufficiently fast
however no explanations of the underlying mechanisms were provided
cooling to room temperature to suppress any subsequent precipitation.
[5]. In their supplementary TEM analysis, the authors quantified the
Thermo-Calc analysis using the TCFE7 database confirmed that Nb(CN)
increased volume fraction and refinement of complex TiNbV carboni-
are dissolved at 1090 °C, and V containing nitrides or carbides are not
tride precipitates due to V-microalloying and estimated that an addi-
stable above 920 °C in these alloys [9]. A second reheating to 1000 °C
tional 20 MPa contribution of precipitates to the overall strength would
followed by austenite conditioning at 850 °C, i.e. ε = 0.3 at 0.1 s−1, was
be plausible [9]. Obviously this is not sufficient to describe the observed
performed for refinement and pancaking of prior austenite grains.
~60–80 MPa of strength difference between the V-added and V-free
Route A entailed cooling the specimens at different rates to develop a
X100 steels and there must be other contributions that the authors have
CCT diagram, whereas route B included a 30 °C/s cooling segment to
not explored yet [9].
450 °C followed by isothermal holding for different times to simulate
It has to be noted that, during linepipe processing most V is in so-
complete and interrupted coiling. Route C was a novel in-situ com-
lution during hot rolling and fast cooling and there has never been any
pression test at 0.1 s-1 strain rate to measure the flow strength of the
reliable evidence of significant V precipitation occurring at coiling
material at various stages of coiling, namely the strength of fresh bai-
temperatures below 550 °C, even in studies using the latest atom probe
nite and its softening behaviour for different coiling times up to 10 h.
and neutron/X-ray scattering techniques [7,10]. Further, most pub-
The specimens processed by routes A and B were utilized for mi-
lished data is limited to as-coiled scenarios and little or no information
crostructural characterization. A Nova Nano-SEM equipped with field-
regarding the strength of fresh bainite for different vanadium contents
emission gun was used for SEM analysis of polished specimens lightly
has been reported in the literature. All the available data suggests that
etched with Nital 2%. A FEI Tecnai Osiris TEM operating at 200 kV with
our knowledge of the strengthening mechanisms of vanadium in bai-
conventional bright field-dark field imaging, along with Scanning
nitic microstructures is incomplete and open to debate; some authors
Transmission Electron Microscopy (STEM) mode using the high angle
claim that it is due to precipitation or clustering, while others believe
annular dark field (HAADF) detector, combined with Energy Dispersive
that vanadium suppresses dynamic recovery and thus prevents soft-
X-ray Spectroscopy (EDX) were used to study the microstructural fea-
ening.
tures of bainite and directly image the precipitates by elemental map-
This paper summarizes the findings of a systematic study on the
ping. TEM samples were prepared by electropolishing of thin foils using
formation and recovery of bainite in the presence of V in a reference
A8 electrolyte. The techniques described in [11] were adopted to
and a vanadium added X100 plate steel. The alloy compositions are the
measure dislocation densities using STEM images. HAADF images were
same as those reported on by Nafisi et al. [5,9]. The objective is to
taken at a low index zone axis (e.g. [011]) where all dislocations are
understand the mechanism of bainite refinement due to vanadium
visible. The foil thickness (t) was measured by Electron energy loss
microalloying and to quantify the ensuing strengthening contributions
spectroscopy (EELS) using the log ratio method [12]. The dislocation
for both fresh and aged bainite.
density (ρ) was calculated using a set of horizontal (Lx) and vertical (Ly)
reference lines superimposed on HAADF images of dislocations. These
2. Experiment
lines intersect the dislocations (ny times in the vertical and nx times in
the horizontal direction), such that the density is given by ρ =
The experimental alloys were variants of a X100 linepipe steel
(∑ n x / ∑ L x + ∑ n y / ∑ L y )/ t . The reported ρ is an average of at least five
containing 0.06C-1.85Mn-0.16Cr-0.03Nb-0.005Ti-0.006N (1.95Mn wt
measurements.
%) containing 0.004 wt% and 0.063 wt% of vanadium, respectively.
The STEM-BF micrographs were used to determine the microstructural
The steels were processed by pilot scale vacuum induction melting and
scale of bainite. Several fields capturing a high magnification of bainite
rolling facilities into 14 mm thick plate at CanmetMATERIALS. Details
packets were analyzed by the intercept method. Straight lines perpendicular
of the thermomechanical processing procedures can be found elsewhere
to the longitudinal direction of laths within a packet were plotted and the
[5]. Rod specimens with 10 mm length and 5 mm diameter were ma-
mean length of the intercepts by the lath boundaries were determined (Llath).
chined at quarter thickness of the plates for dilatometric studies and
The average thickness was estimated using λ = 2 Llath /π [13].
subsequent microstructure characterization. A quenching, deformation

249
F. Fazeli et al.
Fig. 2. Dilatometric findings for the base plate (red) and the V microalloyed variant (blue): (a) Measured CCT diagram, where the solid and dashed lines represent 5% and 50%
transformation respectively, and (b) bainite formation during cooling at 30 °C/s and upon subsequent isothermal coiling at 450 °C.
3. Results
The measured transformation kinetics during continuous cooling
cycles were complied into CCT diagrams and are illustrated in Fig. 2a.
The transformation start temperature of ferrite (5% transformed) ap-
peared to be similar for the reference and V-added steels, however for
cooling rates spanning 1 °C/s to ~50 °C/s, where bainite dominates the
microstructure, noticeably lower transformation temperatures can be
seen in the vanadium alloy. The shift for the 5% transformation start
temperature was estimated to be between 30 and 40 °C with similar
values for 50% transformation. The transformation stasis temperature,
where the bainite reaction can no longer proceed, was less influenced
by vanadium addition, i.e. during 30 °C/s cooling bainite formation
ceased at 448 °C and 430 °C for the reference alloy and the vanadium
variant, respectively. The coiling temperature selected by industry was
450 °C, which is just above these bainite end temperatures. It has to be
noted that the Ac3 temperature, i.e. the completion of austenite for-
mation during very slow heating, was identical for the two steels and
uninfluenced by the difference of vanadium contents.
Fig. 2b shows the dilatometric signal during cooling and coiling
(Route B). The 30 °C/s cooling rate suppressed any decomposition of
austenite to ferrite, indicated by initial linear segment, until the bainite
start transformation temperature, Bs was reached. It is obvious that
bainite started at a higher temperature (580 °C) for the base steel (red
line) whereas in the V-added alloy bainite appeared around 550 °C. The
reference alloy maintains this 30 °C higher temperature of transforma-
tion during the entire cooling segment between Bs and the coiling
temperature, i.e. 450 °C. Most of the bainite reaction occurred during
the cooling segment, spanning less than 3–4 s, and it reached its com-
pletion within the 20 s of coiling (not shown here). The overall fraction
of bainite at the end of coiling was similar for the both alloys ap-
proaching close to 100% transformation. No martensite reaction was
detected by the dilatometer signal during cooling from 450 °C to room
temperature. The rapid completion of the bainite reaction implied that
the fresh bainite would undergo tempering during the prolonged time
of coiling.
The SEM micrographs of fresh bainite from specimens cooled at
30 °C/s and held for 1 min at a coiling temperature of 450 °C (route B),
are illustrated in Fig. 3a and b, for the reference and V-added steels,
respectively. The microstructure is composed of two bainite morphol-
ogies; (a) packets containing several elongated bainite laths and (b)
randomly distributed bainite platelets with relatively equiaxed shape
and irregular boundaries (see annotations). The platelets appear to have
different elevations; some raised above others depending on their re-
action tendency with the etching solution Nital 2% [14]. Further, the
250
Materials Science & Engineering A 720 (2018) 248–256
lath and platelet boundaries are outlined by the etching products de-
spite a very short immersion time. It was found that the V-added variant
had a more refined bainitic structure and a more dominant lath mor-
phology compared to the reference steel, as can be seen qualitatively in
the SEM micrographs of Fig. 3a and b.
Additional TEM characterization, as seen in Figs. 4 and 5 for the
reference and V-added steels, confirmed that the microstructure of
specimens from coiling simulation (route B) were almost fully bainitic
with a slight presence of high carbon residual constituents. The bainite
was composed of noticeably fine, highly dislocated bainitic ferrite laths
and platelets. This is evident in Fig. 4a showing a few adjacent laths of
bainite in the reference steel. The laths featured 300–500 nm thickness
and the interlath boundaries were often free form carbide or retained
austenite. The bainitic ferrite platelets are also highly dislocated and
relatively equiaxed ranging from 500 to 900 nm in dimension, as shown
for the V-added alloy in Fig. 5a. Evidence of slight cementite pre-
cipitation (darker constituents) at platelet boundaries is seen in Fig. 5a.
These cementite particles are enriched with Mn as indicated by the EDX
map of Fig. 5f.
Random distributions of undissolved particles were occasionally
found by EDX analysis (see Figs. 4b-e and 5b-e). As indicated by the
EDX map, these coarse carbonitride precipitates (20–50 nm size) ap-
peared to contain Ti and Nb in the reference alloy. Complex TiNbV(CN)
particles were also found in the V-microalloyed steel (Fig. 5c-e). The Ti
signal was absent in some of the NbV containing particles (see the
bottom right of Fig. 5c-e). No signature of fine NbC, VC or VN particles
was detected. Under optimum imaging conditions a high density of
dislocations could readily be revealed inside the bainitic ferrite laths
and platelets, an example of which is illustrated in Fig. 4f.
More insight pertaining to the nature of these complex TiNbV par-
ticles was found by EDX scans at higher magnifications. Fig. 6a shows
two adjacent laths and TiNb particles separated by a lath boundary in
the reference alloy coiled for one minute. It is clear that cuboid TiN
particles provide a suitable substrate for subsequent precipitation of
NbN particles (Fig. 6b-d). The carbon signal from these NbN pre-
cipitates was very faint. Similar features were observed in V-added
specimens (Fig. 6e-i). Fig. 6e shows a TiNbV(CN) particle in the vicinity
of a lath boundary for a fresh bainite microstructure. Two hemi-
spherical NbCN particles have precipitated on the facets of a TiN cu-
boid. It appears that the TiN contains a substantial amount of vanadium
atoms whereas the neighbouring Nb(CN) precipitate shows very faint
signal of vanadium; only subtle segregation near the interface with the
TiN substrate was detected. The EDX map of C in Fig. 6j also confirmed
the presence of cementite particles at the lath boundaries.
The initial strength of fresh bainite and its softening during coiling at
F. Fazeli et al. Materials Science & Engineering A 720 (2018) 248–256

Fig. 3. SEM micrograph of the specimen cooled at 30 °C/s to 450 °C and held for one minute showing a mixture of lath and platelet morphologies for: (a) reference steel and (b) V-added
variant.

450 °C are illustrated in Fig. 7, showing the response of the reference steel
(red) and the vanadium alloy (blue). The yield strength was determined at
0.5% offset strain from flow curves obtained from in-situ compression tests at
450 °C, as shown in Fig. 7a. This novel in-situ compression strategy permits
the assessment of strength for a target microstructure at any stage of ther-
momechanical processing without artefacts associated with cooling to room
temperature, such as unwanted transformation, precipitation or strain aging.
The superior strength of fresh bainite for the vanadium microalloyed alloy
compared to the reference is readily seen (about 40 MPa), and this higher
strength was maintained during different coiling times. Further, the fresh
bainitic microstructure tended to undergo a substantial softening during
holding at 450 °C; about 80 and 90 MPa strength loss were noticed for the V-
added and reference steels, respectively (cf. Fig. 7b), implying a slightly
higher resistance of the V-added alloy to softening.

Fig. 4. TEM micrograph of the reference steel cooled at 30 °C/s to 450 °C and held for 3 h showing: (a) STEM-BF, (b) EDX scan of Nb superimposed on BF image, (c) Nb distribution (d) Ti
map, (e) Mn map, and (f) a higher magnification STEM-BF of dislocation tangles.

251
F. Fazeli et al.
Fig. 5. TEM micrograph of the V-added steel cooled at 30 °C/s to 450 °C and held for 3 h showing: (a) STEM-BF, (b) EDX scan of V and C superimposed on BF image, (c) Nb distribution (d)
Ti map, (e) V map, and (f) Mn map.
4. Discussion
The effect of vanadium alloying, particularly in solid solution, on
the transformation kinetics of bainite during continuous cooling cycles
has rarely been studied. The shift of the bainite reaction to lower
temperatures has been reported by Sourmail et al. for medium carbon
steels (0.24 wt%) containing 0.2 and 0.4 wt% vanadium [7] where
18 °C shift in Bs was found between the reference alloy and a 0.2 wt% V-
containing steel cooled at 2 °C/s. The authors did not discuss the me-
chanism for this delay nor its implications on mechanical properties.
However, Sourmail et al. studied the isothermal kinetics of bainite
formation and found that vanadium additions have no tangible effect
on the reaction in the 375–450 °C temperature range but slows down
the rate of bainite formation above 450 °C. They attributed this effect of
vanadium to its segregation at prior austenite grain boundaries thus
reducing the potency of these interfaces for bainite nucleation, as
suggested by Enomoto [15]. However, no direct APT evidence for this
was presented. Similar delays in the isothermal kinetics of bainite for-
mation have also been reported for 0.03 wt% V and 0.07 wt% V addi-
tions in a 0.05C-1.5Mn-0.3Si steel examined between 500 and 600 °C
[6]. The underlying mechanism responsible for the retarding effect of
vanadium on bainite kinetics is apparently unknown and requires fur-
ther studies outside the scope of the current paper. Our dilatometric
findings have clearly demonstrated a shift in the bainite transformation
to lower temperatures over a wide range of cooling rates due to vana-
dium microalloying in low carbon steels. As such, the implications for
microstructural change and in particular, to understand the frequently
reported strengthening of bainite will be analyzed in the following
section.
The key morphological features of bainite, namely the lath thickness
252
Materials Science & Engineering A 720 (2018) 248–256
and dislocation density scale with the transformation temperature.
Finer scale bainitic ferrite laths containing denser dislocation tangles
result in stronger bainite [2]. The specimens processed through route B
representing fresh bainite microstructures, were analyzed by TEM to
determine the dislocation density and lath thickness of bainitic ferrite
for both reference and V-added alloys. The characterization of different
foils confirmed that the microstructure was fully bainitic consisting of
mainly slender laths with a smaller fraction of equiaxed platelets. A
microstructural refinement (about 12%), in terms of lath thickness and
platelet size, was found for bainite that formed at a lower temperature
due to vanadium microalloying. Table 1 summarizes the measured
quantities for lath thickness and equivalent area diameter (EQAD) of
bainitic ferrite platelets. The scale of the bainite microstructure varies
with its formation temperature and alloy composition. Singh and Bha-
deshia determined experimentally the lath thickness of bainitic ferrite
for several alloys treated isothermally at various temperatures [16].
Their measured data for low carbon steels, i.e. 0.1 wt%C with ~2 wt%
Mn, which are close to the investigated linepipe steels are reproduced in
Fig. 8a. The measured lath thicknesses for the reference steel and V-
added variant are also included in the graph. Although our data is re-
lated to bainite formation at higher temperatures (between 580 and
450 °C), it closely follows the linear trend of Singh and Bhadeshia.
Another impact of transformations occurring at lower temperatures
where dilatational or shear strains are associated with a phase transi-
tion is manifested in an increased dislocation density. In the case of
austenite decomposition to ferrite, the dilatational or shear components
of transformation strains can be accommodated plastically in both
phases with more participation from the softer phase, i.e. ferrite in low
alloy steels [2]. The laths or platelets of bainitic ferrite in this study
demonstrated very high densities of dislocations. The measured values
F. Fazeli et al. Materials Science & Engineering A 720 (2018) 248–256

Fig. 6. High magnification EDX scans of fresh bainite showing the elemental constituents of complex precipitates in: (a-d) the reference alloy, and (e-j) the V-added steel. Coiled for one
minute at 450 °C.

for the mean dislocation density, along with the respective standard microalloying is obvious, showing close to 75% increase of dislocation
deviation (SD), are given in Table 1 for the fresh bainite formed below density in bainitic ferrite as the result of the transformation occurring at
Bs during cooling after 30 s holding at 450 °C. The benefit of vanadium ~30 °C lower temperatures. The measurement of dislocation number

Fig. 7. Findings of in-situ compression tests at 450°C showing, (a) hot compression flow curves for different coiling times; 1 min (fresh bainite) and 10 h (aged bainite), and (b) strength of
fresh bainite and its softening upon coiling. (For interpretation of the references to color in this figure, the reader is referred to the web version of this article.)

253
F. Fazeli et al. Materials Science & Engineering A 720 (2018) 248–256

Table 1 The neighbouring laths within a packet of bainitic ferrite are sepa-
Microstructural features and their strengthening contributions for fresh bainite (cooled at rated by low angle boundaries [2]. EBSD studies of the reference and V-
30 °C/s below Bs and held 30 s at 450 °C).
added alloys showed that the frequency of these low angle (2–10°)
Reference steel 0.06 wt% V Δσ, MPa boundaries was slightly higher in the V-microalloyed specimens im-
microalloyed steel plying a more refined microstructure. These lath boundaries could
impede the passage of dislocations and thus act as effective obstacles
Bs,°C 580 552
with a strengthening contribution proportional to the inverse thickness
Lath thickness (SD), μm 0.41 0.36
Platelet size as EQAD, μm 0.78 0.68 of the laths, i.e. σlath =k/λ [18,19]. The coefficient k is related to the
Dislocation density (SD), m−2 3.34×1014 5.8×1014 dislocation line tension and the Taylor factor. Different values for k are
(1.26 ×1014) (0.41 ×1014) proposed depending on the assumed dislocation-obstacle interaction
Lath strengthening, MPa 280.5 314.2 33.7 and the boundary misorientation [19], however the value of
Platelet strengthening, MPa 147.4 169.1 21.7
115 MPa μm was adopted in this work. The contribution from lath
Dislocation strengthening, 134.1 176.7 42.6
MPa strengthening in fresh bainite for the two alloys is included in Table 1
Linear combined 336.1 398.0 61.9 showing a 33 MPa strength gain due to refinement of the lath structure.
strengthening corrected Correcting for the temperature at which the compression tests were
for 450 °C
carried out, i.e. 450 °C, a 0.81 reduction factor should be considered
Quadratic combined 252.1 292.3 40.2
strengthening corrected resulting in 27.3 MPa strengthening at 450°. The factor 0.81 is the ratio
for 450 °C of dislocation line tension at 450 °C compared to its value at 25 °C,
correcting for the temperature dependency of the shear modulus as
suggested by Frost and Ashby [20]. The strengthening contribution of
density was carried out in both lath and platelet regions and the de- platelets can be estimated in a similar way, considering their submicron
termined mean values include both morphologies. It is noteworthy to scale. Using k = 115 MPa μm/λ, the strength gain from the refinement
compare these measured quantities with those reported in the litera- of bainite platelets was estimated to be 21.7 MPa at room temperature
ture. Takahashi has analyzed bainitic ferrite for a range of transfor- (17.5 MPa at 450 °C).
mation temperatures and proposed an empirical equation showing the The strengthening contribution from dislocations can be estimated by
dependency of its dislocation density with temperature [17]. Fig. 8b the well known Taylor relationship, σdislocation= αMμbρ1/2, where α is a
compares Takahashi's equation with data points from this study. The factor of the order of 0.38 [18], M is the Taylor factor, μ is the shear
measurement of defect densities is associated with significant un- modulus and b is the magnitude of the Burgers vector. The values de-
certainties, nevertheless the determined values from our measurements termined for the dislocation density were used to estimate the ensuing
are comparable with Takahashi's proposed trend. dislocation strengthening in the fresh bainite for the reference steel and the
As illustrated in Fig. 7a, the fresh bainite of V-microalloyed steel has vanadium microalloyed alloy. The results indicating that about 42 MPa
a superior strength (~40 MPa) compared to that of the reference plate. strength difference is developed between the two steels; corresponding to
Further, this strength advantage is maintained during coiling simula- 35 MPa at 450 °C (Table 1). The combined contribution of lath strength-
tions for up to 10 h at 450 °C. The additional strengthening of bainitic ening and dislocation strengthening is believed to be the prevailing factor
strip steels [4] and linepipe plates [5] as a result of vanadium additions for the superior strength of fresh bainite in the vanadium microalloyed steel.
has been reported by other studies. However, the reported strength In order to add these two contributions a superposition law was adopted.
values were associated with specimens at the end of coiling, where However, to combine the contribution of different obstacles with various
various microstructural changes such as softening or precipitation strengths and populations a suitable addition exponent should be selected
prevented a direct interpretation of the observed strength gains due to [21]. A linear summation of strengths (upper bound: n = 1) is valid where a
vanadium alloying. Indeed, experimental data for the strength of fresh high density of weak obstacles and a small number of strong obstacles are
bainite and its evolution with coiling are lacking in the literature. In our present. On the other hand, the quadratic addition (lower bound: n = 2) is
study, the unique in-situ compression tests (route C) and the inter- more suitable for systems containing obstacles with comparable strengths,
rupted coiling simulations (route B) permit to understand the and thus the overall strength can be assessed by a linear summation of
strengthening contribution of vanadium in fresh bainite. obstacle densities. For fresh bainite with a high dislocation density and fine

Fig. 8. The measured lath thickness (a) and dislocation density (b) of bainitic ferrite along with data from the literature.

254
F. Fazeli et al.
lath structure where the obstacle strengths (i.e. kb versus αMμb2) are
comparable, the quadratic superposition is a better description for the
combined contribution of laths and dislocations. The calculated combined
2 2
strengthening of lath and dislocation, i.e. ∆σ = σlath + σdislocation , for fresh
bainite is shown in Table 1. The values were corrected for the temperature
effect. Included also in the Table is the combined strength assuming a linear
superposition law. It can clearly be seen that the estimated additional
strength gain due to vanadium microalloying, i.e. 40 MPa using quadratic
addition law, describes very accurately the experimental strength advantage
measured by in-situ compression tests (cf. Fig. 8a). It has to be noted that,
for these calculations, only the lath refinement was considered. As men-
tioned earlier, the bainitic microstructure of these alloys is composed of lath
and platelet morphologies with lath dominating; and a quantitative analysis
to determine the actual fractions of different morphologies is not trivial to
carry out. Accounting for the platelet morphology (e.g. 40% platelets and
60% laths mixture) would only result in a few MPa lower combined
strength, 38 MPa.
Our analysis indicated that the refinement in lath structure in combi-
nation with the increased dislocation density contributed to an additional
40 MPa strength at 450 °C (~ 50 MPa at room temperature), sufficient to
describe the observed strengthening difference of fresh bainite by in-situ
compression tests. The yield strength difference for the actual steel plates,
controlled-rolled and coiled at 450 °C, was reported to be about 60 MPa in
a previous study [9]. In that study a comprehensive characterization of the
microstructure revealed that vanadium microalloying promoted a higher
number density of complex TiNb(CN) particles and a smaller size dis-
tribution of these particles in hot-rolled plates (see Table 1, Ref [9].),
finding also that no fine-scale vanadium precipitates had formed during
cooling or coiling [9]. From Ashby-Orowan analysis this provides an es-
timated additional precipitation strengthening of 20 MPa in the V-added
steel compared to the reference X100 plate, which is not sufficient to
explain the overall strengthening difference between the two alloys and
there must be other contributions to describe the strength benefit of V-
added variant [9]. Obviously, the findings of this paper confirms that the
contribution of lath refinement and increased dislocation density must be
explicitly accounted for. It is noteworthy to emphasize that the thermo-
mechanical processes in this study (routes A, B and C) were different from
the actual plate rolling schedule (cf. Ref [5].) resulting in a much lower
number density of complex carbonitride particles during high temperature
processing. Further, the fast cooling and low temperature coiling pre-
vented the precipitation of vanadium carbides or nitrides. As confirmed by
our TEM characterization and the micrographs in Figs. 4–6, only a few
undissolved coarse TiN in the reference steel and (TiV)N particles in the V-
added alloy, along with attached NbN precipitates, were present in the
specimens from coiling simulations (route B). These particles were coarse
and had very low number densities, such that their strengthening con-
tribution was insignificant. Thus the estimated 20 MPa precipitation
strengthening difference in actual hot rolled plates reported previously [9]
is not relevant to the specimens studied here.
As pointed out earlier, most literature studies analyzed the strength
of as-coiled specimens [4,5,8,9], as such the final strength difference
due to vanadium alloying must have been influenced by several mi-
crostructural changes occurring during coiling. For example, the
strength gain reported by Siwecki et al. [4] depended on the coiling
temperature; a difference of 70 MPa in yield strength was found for
coiling at 400 °C whereas coiling at 500 °C led to 110 MPa additional
strength in a low carbon Mn-Mo-Cr hot strip steel as a result of 0.08 wt
% vanadium microalloying. As illustrated in Fig. 7, fresh bainite tends
to soften rapidly at coiling temperatures; about 90 MPa drop was noted
for the reference plate whereas the V-added alloy demonstrated a
higher resistance to softening. A similar positive effect of V on the
temper resistance of martensite has long been known [22]. Upon tem-
pering of fresh bainite, several microstructural changes occur [2].
Dislocation tangles within bainitic ferrite are rearranged into more
stable configurations and neighbouring laths with small misorientations
255
Materials Science & Engineering A 720 (2018) 248–256
tend to merge forming bigger monolithic plates; these two events
combine to significantly lower the overall strength of bainite. It has to
be noted that the tempering response of bainite depends on its mor-
phology and the tempering temperature. Lath morphologies tends to
undergo more rapid softening compared to granular morphologies [23].
Sluggish tempering of granular bainite tempered at high temperatures
600–650 °C has been reported [23], however the observed softening
trend could be due to simultaneous VC precipitation and dislocation
recovery; the former could offset the softening effect of dislocation
recovery. The low coiling temperature examined in this study, i.e. 450°
prevents VC or cementite precipitation and thus the measured softening
was merely due to recovery of the bainite substructure. At coiling
temperatures above 500 °C, potential clustering or precipitation of va-
nadium could interact with the rearrangement of dislocations resulting
in sluggish recovery of bainite [4]. However, there has never been any
reliable evidence of significant V precipitation occurring at these low
temperatures, even in studies using the latest atom probe and neutron/
X-ray scattering techniques [7,10]. Thus more detailed experiments are
required to determine the exact mechanisms of bainite softening in the
presence of V. These aspects of vanadium microalloying, i.e. cluster
formation and precipitation of vanadium during coiling and its
strengthening contribution are the subject of our current study in sev-
eral experimental alloys with additions of V, Nb or V-Nb. These findings
provided insight for alloy design and process optimization of V-bearing
line-pipe steels with higher as-coiled strength.
5. Conclusions
The strengthening mechanism of vanadium microalloying in a low-
C bainitic microstructure was studied in a reference X100 linepipe steel
and a vanadium microalloyed variant. The main findings are summar-
ized in the following:
• It was found that, over a wide range of cooling rates (1–50 °C/s),
vanadium atoms in solution delay the bainite reaction to lower
transformation temperatures (by 30–40 °C).
• The shift in bainite transformation to lower temperatures resulted in
refinement of the lath structure and an increase of the dislocation
density of fresh bainite for the V-added steel compared to the re-
ference alloy.
• No signature of vanadium precipitation during fast cooling and
subsequent low temperature coiling (450°) was found from TEM
characterization.
• The presence of undissolved TiN in the reference alloy, and (TiV)N
in the V-added variant was confirmed by TEM characterization.
Hemispherical caps of NbN were invariably associated with these
undissolved Ti containing nitrides.
• The complex TiNbV(CN) were coarse and few in number density, as
such no tangible strengthening contribution could be made.
• The additional strengthening contribution from the finer laths and a
higher dislocation density was estimated to be around 40 MPa, which
is consistent with the experimental strength gain measured by in-situ
compression tests of fresh bainite at 450 °C.
• Further, interrupted coiling simulations in conjunction with in-situ
compression tests confirmed that the fresh bainite softened rapidly
upon isothermal holding at 450 °C, i.e. 90 MPa strength drop during
10 h for the reference steel.
• In the presence of solute vanadium the fresh bainite demonstrated
more resistance to softening during coiling.
Acknowledgments
The work was supported by the Natural Resources Canada through
A-base funding. Further, the materials were supplied by Evraz Inc. NA.
F. Fazeli et al.
References
[1] S. Vervynckt, Control of the Non-recrystallisation Temperature in High Strength
Low Alloy Steels, Ph.D. Thesis Ghent University, 2010.
[2] H.K.D.H. Bhadeshia, Bainite in Steels, Institute of Materials, London, 2001
(2nd ed.).
[3] R. Lagnebrog, B. Hutchinson, T. Siwecki, S. Zajac, The Role of Vanadium in
Microalloyed Steels, Vanitec, London, 2014.
[4] T. Siwecki, J. Eliasson, R. Lagneborg, B. Hutchinson, Vanadium microalloyed bai-
nitic hot strip steels, ISIJ Int. 50 (2010) 760–767.
[5] S. Nafisi, M.A. Arafin, R. Glodowski, L. Collins, J. Szpunar, Impact of vanadium
addition on API X100 steel, ISIJ Int. 10 (2014) 2404–2410.
[6] C.-Å. Däcker, O. Karlsson, C. Luo, K. Zhu, J.-L. Collet, M. Green, P. Morris,
R. Kuziak, N. Kwiaton, C. Stallybrass, Z.I. Olano, Bainitic Hardenability – Effective
Use of Expensive and Strategically Sensitive Alloying Elements in High Strength
Steels (BainHard), Final report of RFSR-CT-2007-00023, Publications Office of the
European Union, Luxembourg, 2012, http://dx.doi.org/10.2777/22306.
[7] T. Sourmail, C. Garcia-Mateo, F.G. Caballero, S. Cazottes, T. Epicier, F. Danoix,
D. Milbourn, The influence of vanadium on ferrite and bainite formation in a
medium carbon steel, Metall. Mater. Trans. 48A (2017) 3985–3996.
[8] B. Hutchinson, T. Siwecki, J. Komenda, J. Hagström, R. Lagneborg, J.-E. Hedin,
M. Gladh, New vanadium-microalloyed bainitic 700MPa strip steel product, Iron
Steelmak. 41 (2014) 1–6.
[9] S. Nafisi, B. Shalchi Amirkhiz, F. Fazeli, M. Arafin, R. Glodowski, L. Collins, Effect of
vanadium addition on the strength of API X100 linepipe steel, ISIJ Int. 56 (2016)
154–160.
[10] Y. Oba, S. Koppoju, M. Ohnuma, T. Murakami, H. Hatano, K. Sasakawa, A. Kitahara,
J.-ichi Suzuki, Quantitative analysis of precipitate in vanadium-microalloyed
medium carbon steels using small-angle X-ray and neutron scattering methods, ISIJ
Int. 51 (2011) 1852–1858.
[11] D. Rojas, J. Garcia, O. Prat, L. Agudo, C. Carrasco, G. Sauthoff, a.R. Kaysser-Pyzalla,
256
Materials Science & Engineering A 720 (2018) 248–256
Effect of processing parameters on the evolution of dislocation density and sub-
grain size of a 12%Cr heat resistant steel during creep at 650°C, Mater. Sci. Eng. A
528 (2011) 1372–1381.
[12] T. Malis, S.C. Cheng, R.F. Egerton, EELS log-ratio technique for specimen-thickness
measurement in the TEM, J. Electron. Microsc. Tech. 8 (1988) 193–200.
[13] L.C. Chang, H.K.D.H. Bhadeshia, Austenite films in bainitic microstructures, Mater.
Sci. Technol. 11 (1995) 874–882.
[14] E.V. Pereloma, J.D. Boyd, On the nature of raised ferrite in a low-carbon micro-
alloyed steel, Scr. Mater. 34 (1996) 703–706.
[15] M. Enomoto, N. Nojiri, Y. Sato, Effect of vanadium and niobium on the nucleation
kinetics of proeutrctoid ferrite at austenite grain boundaries, Mater. Trans., JIM 35
(1994) 859–867.
[16] S.B. Singh, H.K.D.H. Bhadeshia, Estimation of bainite plate thickness in low-alloy
steels, Mater. Sci. Eng. A245 (1998) 72–79.
[17] M. Takahashi, H.K.D.H. Bhadeshia, Model for transition from upper to lower bai-
nite, Mater. Sci. Technol. 6 (1990) 592–603.
[18] C.H. Young, H.K.D. Bhadeshia, Strength of mixtures of bainite and martensite,
Mater. Sci. Technol. 10 (1994) 209–214.
[19] K. Zhua, O. Bouaziza, C. Oberbilliga, M. Huang, An approach to define the effective
lath size controlling yield strength of bainite, Mater. Sci. Eng. A 527 (2010)
6614–6619.
[20] H.J. Frost, M.F. Ashby, Deformation Mechanism Maps, Pergamon, Oxford, UK,
1982.
[21] U.F. Kocks, P. Hassen, V. Gerold, G. Kostorz (Eds.), Proc. 5th Int. Conf. On Strength
of Materials and Alloys, Pergamon Press, Oxford, UK, 1979, p. 1661.
[22] R.A. Grange, C.R. Hribal, Porter, Hardness of tempered martensite in carbon and
low-alloy steels, Metall. Trans. A 8 (1977) 1775–1785.
[23] O.V. Sych, A.A. Kruglova, V.M. Schastlivtsev, T.I. Tabatchikova, I.L. Yakovleva,
Effect of vanadium strengthening upon tempering of a high-strength pipe steel with
different initial microstructure, Phys. Metal. Metallogr. 117 (2016) 1270–1280.