In-situ Observation of Bainite Transformation in CGHAZ of 420 MPa Grade Offshore Engineering Steel with Different Mo Contents

Abstract

The 420 MPa grade offshore engineering steels with the different Mo contents were fabricated through Ca deoxidation. The effect of Mo content on the microstructure and the impact toughness of the coarse-grained heat-affected zone (CGHAZ) after the welding heat input of 100 kJ/cm were investigated by welding simulation and high-temperature laser scanning confocal microscopy (HTLSCM). As the Mo content is increased from 0.08 to 0.16 wt%, the impact toughness at −40°C is increased from 15 to 160 J, and the fracture is changed from cleavage to ductile and quasi-cleavage. Increasing Mo content leads to the increase in lath bainite (LB) and the decrease in granular bainite (GB) and acicular ferrite (AF) in CGHAZ. Electron backscatter diffraction (EBSD) results show that the fraction of high angle grain boundaries (HAGBs) is increased from 45% to 58%, while the effective grain size (EGS) is decreased from 10.4 to 6.3 µm in the HAZ of the steels. The growth rates of bainite ferrite range from 39.7 and 134.5 µm/s for Mo8 steel, and 51.2 to 165.6 µm/s for Mo16 steel. In-situ observation shows that increasing Mo content decreases the austenite grain size from 114 to 100 µm. As for Mo8 steel, AFs nucleate from the inclusions firstly. Then the different bainites nucleate, grow, stop growth and widen as they impinge the other bainites and grain boundaries. The secondary bainites finally nucleate from the prior bainite and AF. As for Mo16 steel, the AF cannot nucleate, and the bainites have a similar nucleating and growth behavior as the Mo8 steel.

1. Introduction

Offshore engineering steel is generally utilized for the petroleum and natural gas industries, such as marine platforms, marine energy equipment, submarine oil and gas pipelines, etc. Owing to its special service environment, the increasingly high strength and excellent toughness are required at low temperature. The varied kinds of offshore engineering steels are developed to cope with the increasing energy demand,¹⁾ whose yield strengths range from 420 to 690 MPa. To improve the welding efficiency, it is desirable to enhance the welding heat input in the construction of the offshore engineering steel. However, the heat-affected zone (HAZ) toughness is deteriorated due to the formation of the coarse austenite grain and brittle phases like martensite-austenite constituents (MAs), carbides and inclusions.²⁾ It is found that the large MAs are the direct factor that leads to the deteriorated impact toughness of the HAZ.³⁾

The oxide metallurgy by using strong deoxidizers like Ti, Mg and Ca⁴⁾ can be used to improve the HAZ toughness of the offshore engineering steel, in which large amounts of fine particles are generated in the HAZ. The austenite grain growth is inhibited by the nano-scale particles during the welding process. The intra-granular acicular ferrite (IAF) is promoted by the micro-scale inclusions,⁵⁾ whose inter-locked microstructures are beneficial to hinder crack propagation so as to improve the HAZ toughness. Yang et al.⁵⁾ found that the HAZ toughness of steel plate can be improved through Mg deoxidation with the heat input of 400 kJ/cm, and the size of inclusions decreases and the number density of inclusions increases with the increase of Mg content. Kato et al.⁶⁾ found that the Ca addition promotes the formation of fine TiN particles to obtain a fine-grained microstructure of HAZ. Zhang et al.⁷⁾ proved that the Ca addition increases the quantity of submicron-scale and nano-scale TiN particles, and demonstrated that submicron-scale TiN can efficiently refine the austenite grains.

Additionally, the HAZ toughness can also be improved by the microalloying elements. As is well known, Mo has an excellent effect on improving the hardenability of steel. Kong et al.⁸⁾ found that the temperatures of Bs and Bf are reduced and the size of the microstructure is decreased as 0.40 wt% Mo is added into the steel. Khare et al.⁹⁾ showed that Mo has limited influence on the rate of transformation of austenite into bainite, but a dramatic effect on the phase change temperature. Hu et al.¹⁰⁾ indicated that the addition of Mo promotes the bainite transformation and produces more bainite which leads to a higher strength than Nb and Mo+Nb steel. However, the effect of Mo on the microstructure and impact toughness of the HAZ has not been investigated thoroughly, particularly in the 420 MPa offshore engineering steel with Ca deoxidation.

Recently, high-temperature laser scanning confocal microscopy (HTLSCM) was widely utilized for in-situ observation.^11,12,13) Zhang et al.¹⁴⁾ studied the formation of IAF induced by the non-metallic inclusions by HTLSCM and found that the potent inclusions for nucleation of IAF are TiO, TiN and MnS. Terasaki et al.¹⁵⁾ found that the bainite belonging to the same Bain group is formed synchronously, even though they are at a relatively large distance from each other from the in-situ observation. Shen et al.¹³⁾ observed the lath martensite growth behavior by HTLSCM, and concluded that the preferentially nucleating sites of lath martensite are defects, grain boundaries, inclusions and grain interiors. These findings demonstrate that the HTLSCM is a practical and useful technique to study the phase transformation behavior.

The objective of the present work is to investigate the effect of Mo content on the microstructure and the impact toughness of the HAZ of offshore engineering steel. The HAZ specimens were simulated through Gleeble-3800 thermal-mechanical physical simulator with the heat input of 100 kJ/cm, and then the Charpy impact test was conducted at −40°C. The in-situ observation was carried out with HTLSCM, and the crystallographic microstructure and the precipitate were studied by electron backscatter diffraction (EBSD) and transmission electron microscope (TEM). The main points of this paper include: (1) the effect of Mo content on the HAZ toughness and microstructure; (2) the ferrite growth behavior through HTLSCM; (3) the effect of Mo content on ferrite growth rate with thermodynamic calculations.

2. Experimental

The experimental steels were melted in a vacuum induction furnace with a capacity of 50 kg, and the deoxidation treatment was carried out under Ar atmosphere. The deoxidants like Ca, Ti, Si and Al were added to reduce oxygen content and the target compositions were obtained with the CaO slag. The melts were then cast into ingots with the dimension of 120 mm × 180 mm × 240 mm, and then hot-rolled into thick steel plates. The rough rolling was conducted at 930°C with a reduction ratio greater than 30%, and the finish rolling was conducted at 800°C with a reduction ratio of over 30% as well. Eventually, the steel plates were cooled from 760 to 400°C at the rate of 10°C/s with a final thickness of 50 mm. A thermal camera for the surface temperature was used for the measurement of the temperature during hot rolling and cooling of steel plate. Four tensile specimens with the dimensions of Φ5 mm × 30 mm were cut from the position of 1/4 thickness and 1/4 width away from the edge of the steel plates for each tensile test, which is in longitudinal direction of the steel plate.

Table 1 shows the chemical compositions of the two experimental steels with the different Mo contents, which were designated as Mo8 and Mo16, respectively. Table 2 gives the mechanical properties of the two experimental steels, and these newly designed steels have a similar tensile strength level to the EH420 offshore engineering steel plate.

Table 1. Chemical compositions of the two experimental steels (wt%).

Steels	C	Si	Mn	S	Nb	Ti	Cu	Cr	Mo	V	Ca	Al	O	N
Mo8	0.067	0.140	1.040	0.003	0.002	0.004	0.240	0.015	0.081	0.005	0.0023	0.0120	0.0031	0.0025
Mo16	0.063	0.110	1.010	0.003	0.002	0.004	0.220	0.020	0.160	0.005	0.0021	0.0130	0.0034	0.0024

Table 2. Mechanical properties of the two experimental steels.

Steels	Thickness (mm)	Tensile strength (MPa)	Yield strength (MPa)	Elongation (%)	Charpy impact at −40°C (J)
Mo8	50	527–535	381–428	≥ 34	336, 330, 333
Mo16	50	539–542	420–449	≥ 34	319, 318, 321

To evaluate the HAZ toughness, the steel samples with a dimension of 11 mm × 11 mm × 71 mm were cut down from the position of 1/4 thickness and 1/4 width away from the edge of the steel plates with 50 mm thickness. The one-pass electro-gas arc welding was simulated by the thermal simulator (Gleeble3800, Dynamic Systems Inc.). The t_8/5 is set to be 24 s, which corresponds to the welding heat input of 100 kJ/cm. Figure 1 presents the heating and cooling curves of the welding simulation of 100 kJ/cm and HTLSCM. The peak welding temperature was set to be 1350°C and held for 1 second, then cooled to 800°C at the average cooling rate of 55°C/s, followed by the average cooling rate of 12.5°C/s to 500°C and finally cooled to room temperature in the atmosphere. After welding simulation, the specimens were cut into 10 × 10 × 55 mm³ with a notch tip in the middle part then subjected to the Charpy impact tests at −40°C according to ASTM E23 standard.

Fig. 1.

Heating and cooling curves of the welding simulation of 100 kJ/cm and HTLSCM. (Online version in color.)

The HAZ samples were cut from the weld simulated specimens and then were mechanically ground, polished, and etched in 4% nital (4 ml HNO₃ + 96 ml C₂H₅OH) for optical microscopy observation (OM, DM2700M, Leica Microsystems) and scanning electron microscope observation (SEM, EVO 18, Carl Zeiss) together with energy dispersive spectroscopy (EDS, Oxford, X-max). The average size of two-dimensional prior austenite grains was measured with the intercept method by counting more than 200 grains within 20 mm². The microhardness was measured by using 0.5 kgf (kilogram-force) for 5 seconds load in a Vickers hardness tester (HX-1000, Shanghai Optical instrument), and each specimen was measured for 10 points. To investigate the nano-scaled particles in the HAZ, the extraction replicas method was implied. The replicas were submerged in 10% nital to remove the carbon film from the sample surface. Then the carbon film was picked up by using the Cu grid and was observed by TEM (JEM-200CX, JEOL) operating at 200 kV.

The HTCSLM (Laser-tech, Japan) equipment enables the in-situ observation of the nucleation and growth of bainite ferrite in the grains during the cooling process. The cooling rate was set to be the same as the welding simulation to obtain the similar microstructure. The HTCSLM samples were polished and etched in 4% nital for the subsequent observation. The α orientation maps of HAZ specimens and HTCSLM specimens were measured by EBSD (SYMMETRY, Oxford Instruments) in the scanning electron microscopy (SEM, EVO18, Carl Zeiss) system at an accelerating voltage of 20 kV, tilt angle of 70° and step sizes of 0.4 μm. The post-process was carried out by AZtecCrystal software. To estimate the effect of Mo on the phase transition, the austenitizing temperature was obtained by dilatation test (DIL 402C, NETZSCH) with the specimens heated and cooled under Ar atmosphere.

3. Results

3.1. HAZ Microstructure

Figure 2 presents the SEM and OM micrograph of microstructure in the CGHAZ for Mo8 and Mo16 steels after the heat input of 100 kJ/cm. The microstructures mainly consist of lath bainite (LB) with the bainite growing in the same direction in the grain, and granular bainite (GB) with the polygonal shape. As shown in Figs. 2(a) and 2(b), the prior austenite grain boundaries (PAGBs) and plenty of white particles in the sub-grain boundaries of LB and GB can be easily identified. Additionally, the black particles are identified as carbides or MAs in Figs. 2(c) and 2(d).

Fig. 2.

(a) and (b) SEM, (c) and (d) OM micrograph of microstructure in the CGHAZ for (a) and (c) Mo8, (b) and (d) Mo16 steels after the heat input of 100 kJ/cm. (Online version in color.)

Figure 3 illustrates the inverse pole figure (IPF) maps, image quality map with HAGBs (≥15°), and kernel average misorientation (KAM) maps of Mo8 and (d–f) Mo16 CGHAZ. It is clear that the crystallographic morphology consists of LB and GB, and the AF with a larger aspect ratio was also found as shown in Fig. 3(d). Furthermore, the CGHAZ of Mo8 steel contains less LB and more GB than that of Mo16 steel, indicating the inhomogeneity of Mo8 steel. The high angle grain boundaries (HAGBs) are defined as the misorientation angle larger than 15°, while the low angle grain boundaries (LAGBs) are defined as the misorientation angle smaller than 15°. The white line represents the HAGBs as shown in Figs. 3(b) and 3(e). The HAGBs are widely distributed along the grain boundaries and the different packet boundaries. It implies that the distribution of HAGBs is uneven, and a large number of HAGBs are found in the PAGB for Mo8 steel. Regarding Mo16 steel, more HAGBs are found in LB than in the GB as pointed out in zone A and B from Fig. 3(e). Moreover, the fraction of HAGBs in Mo16 steel is significantly higher than that in Mo8 steel. The KAM map represents the local strain or crystal deformation level, which means that the local strain or crystal deformation level in Mo16 steel is larger than that in Mo8 steel.

Fig. 3.

EBSD results of (a) and (d) inverse pole figure (IPF) maps, (b) and (e) LAGBs (2°–15°) and HAGBs (≥15°), and (c) and (f) KAM maps for (a–c) Mo8 and (d–f) Mo16 CGHAZ. (Online version in color.)

Figure 4 depicts the misorientation angle distribution and the effective grain size (EGS) in the CGHAZ of Mo8 and Mo16. It is worthwhile mentioning that the EGS is defined as the size of the sub-grain with the HAGBs larger than 15°, which plays a significant role in its properties and impact toughness.¹⁶⁾ It can be seen that increasing Mo content increases the area fraction of HAGBs from 45 to 58%, and decreases the average value of EGS from 10.4 to 6.3 μm.

Fig. 4.

Misorientation angle distribution and EGS in the CGHAZ of Mo8 and Mo16. (Online version in color.)

Figure 5 is the TEM photographs of (a) and (d) the morphologies of nano-scaled particles on the carbon film, (b) and (e) selected area electron diffraction (SAED), and (c) and (f) EDS results for the HAZ specimens of (a–c) Mo8 and (d–f) Mo16 steels. It can be seen that the TiN particles have the rectangle shape while the MoC carbides have the amorphous shape. Those nano-scaled particles were mainly TiN with the size of 18 nm in Mo8 HAZ. Regarding Mo16 steel, plenty of particles are identified as MoC in addition of TiN. It indicates that increasing Mo content promotes the formation of the MoC carbides in the HAZ.

Fig. 5.

TEM photographs of (a) and (d) the morphologies of nano-scaled particles on the carbon film, (b) and (e) SAED, and (c) and (f) EDS results of the HAZ specimens for (a–c) Mo8 and (d–f) Mo16 steels. (Online version in color.)

3.2. HAZ Toughness and Microhardness

Table 3 lists the measured Charpy impact energy at −40°C and the micro-hardness of the HAZ and matrix. As the Mo content is increased, the HAZ toughness is significantly improved from 15 J to approximately 160 J at −40°C. Both microhardness of the matrix and HAZ in Mo16 steel are also larger than those in Mo8 steel.

Table 3. Measured Charpy impact energy at −40°C and microhardness of the matrix and HAZ.

	Charpy impact energy at −40°C (J)	Hardness of matrix (HV0.5)	Hardness of CGHAZ (HV0.5)
Mo8	15, 15	220 ± 4	269 ± 12
Mo16	159, 162	235 ± 4	297 ± 15

Figure 6 exhibits the fracture surface and the magnified morphologies in the HAZ of Mo8 and Mo16 steels. It can be seen that the fracture is brittle for Mo8 steel. The cleavage facet size is measured to be about 12.7 μm, which can be determined from the surrounding tear ridge as referenced from Shibanuma et al.¹⁷⁾ The tear ridge with small dimples can absorb partial impact energy and is beneficial to the toughness. As for Mo16 steel, the fracture surface is the combination of ductile and brittle zones. The dimple with small and deep holes has the size of about 9.2 μm in the ductile zone, indicating that the ductile zone is formed due to the growth and coalescence of voids. In addition, the fracture is found to be quasi-cleavage in the brittle zone which is between ductile and cleavage.

Fig. 6.

SEM micrographs of fracture surface and the magnified morphologies in the HAZ of (a–c) Mo8 and (d–f) Mo16 steels. (Online version in color.)

3.3. HTLSCM Observation

From in-situ observation by HTLSCM, it is convenient to study the growth rate and growth behavior of the ferrite during the continuous cooling process. The cooling curves are set to be the same as the welding simulation to obtain a similar microstructure, as shown in Fig. 1.

3.3.1. Austenite Grain and Room-temperature Microstructure

Figure 7 shows the microstructures and grain size distributions at the room temperature of the HTLSCM specimens for (a) and (c) Mo8, (b) and (d) Mo16 steels. The microstructures of both steels are LB and GB which are consistent with those of the HAZ specimens. The sizes of the prior austenite grains are significantly decreased as the Mo content is increased from 0.08 to 0.16 wt%. The average sizes are 114 and 100 μm for Mo8 and Mo16 steel, respectively.

Fig. 7.

Microstructures and grain size distributions at the room temperature of the HTLSCM specimens for (a) and (c) Mo8, (b) and (d) Mo16 steels. (Online version in color.)

Figure 8 gives the IPF maps, image quality map with misorientation distribution, KAM maps and pole figure maps of HTLSCM for Mo8 and Mo16 steels. It can be seen that the results are similar to those of HAZ specimens. The crystallographic microstructures contain LB and GB, and there are more LB and less GB in the Mo16 steel than those in the Mo8 steel. The KAM value of Mo8 steel is evidently smaller than that of Mo16 steel, implying that increasing Mo content increases the local strain and the crystal deformation level. The poly figure of zone A and B from Mo8 and Mo16 grains are plotted with the calculated results are present in Figs. 8(d) and 8(h). It indicates that the relationship of ferrite with prior austenite keeps K-S.

Fig. 8.

EBSD results of (a) and (e) the IPF maps, (b) and (f) the image quality maps with misorientation distribution, (c) and (g) the KAM maps, and (d) and (h) the pole figure maps of HTLSCM for (a–d) Mo8 and (e–h) Mo16 steels. (Online version in color.)

Figure 9 illustrates the misorientation angle distribution and EGS for Mo8 and Mo16 HTLSCM specimens. As the Mo content is increased, the area fraction of HAGBs is increased while the EGS is decreased, which has the same tendency as those of CGHAZ specimens.

Fig. 9.

(a) Misorientation angle distribution and (b) EGS for Mo8 and Mo16 steels. (Online version in color.)

Figure 10 gives the micrographs of the typical AF and LB nucleated from the inclusions and their EDS results for (a) and (b) Mo8, (c) and (d) Mo16 steels. For Mo8 steel, the AFs with the aspect ratio of 1.6–2.2 were nucleated from the inclusion in Fig. 10(a). EDS results show that the inclusion is a CaO–Al₂O₃–MnS–TiN complex compound. For Mo16 steel, the AF with an aspect ratio of 3.3 and LB both nucleated from the inclusion. Generally, the LB with a parallel shape has a large aspect ratio than that of AF, and is also separated by the low-misorientation boundaries, carbides and M-A constituent films,¹⁸⁾ which can be distinguished with the AF. The EDS results show that the inclusion is a CaO–Al₂O₃–MnS–TiN complex compound as well.

Fig. 10.

SEM images of HTLSCM of the typical (a) AF and (b) LB nucleated from inclusion, and the corresponding (d) and (d) EDS results for (a) and (b) Mo8, (c) and (d) Mo16 steels. (Online version in color.)

3.3.2. Bainite Transformation

Figure 11 reveals the HTLSCM snapshots of Mo8 steel at the different temperatures of the nucleation and growth of different bainite ferrites, which are designated as B1 (B1-1, B1-2...), B2 (B2-1, B2-2...), B3 (B3-1, B3-2...) and B4 (B4-1, B4-2...), respectively. B1, B2, B3 and B4 represent the bainite nucleated from boundaries, inclusions, intra-grain and prior-ferrites, respectively. In addition, B1, B2 and B3 grow up first which could be classified as the prior bainite. B4 grows up from the prior bainite, which could be classified as the secondary bainite. The red arrows represent the growth direction of the bainite. All photos are taken from the video which is shot during observation. In the same shooting field, the AFs at different times are captured, and the lengths of different AFs and bainites are measured in different times, then the growth rates of AFs and bainites are obtained in Figs. 12(f) and 14. The identification of different phases is from their transformation features based on the previous works.^14,19,20) The LB and AF can be easily recognized from in-situ observation. The LB has a larger growth rate and length, while the AF has a small growth rate and length, which is often found to be nucleated from inclusions. However, the GB is identified as the remaining phase which is based on its morphologies, which is usually difficult to be distinguished from the HTLSCM snapshots.

Fig. 11.

HTLSCM snapshots of Mo8 steel at the different temperatures for (a–h) the nucleation and growth of different bainites. (Online version in color.)

Fig. 12.

Enlarged zone from Fig. 11(a) of Mo8 steel at the different temperatures for (a–e) AF nucleating at inclusions and the growth rate (f). (Online version in color.)

Fig. 14.

Growth rates of the different bainite ferrites observed with HTLSCM of (a) Mo8 and (b) Mo16 steels. (Online version in color.)

When the temperature was 586.2°C, the AF was nucleated at the inclusion as shown in Fig. 11(a). Then B1-1 and B2-1 were nucleated from the grain boundary, and their growth rates were almost the same. Meanwhile, the bainite B3-1 was nucleated and grew up from the intra-grain, as well as the B3-2. As the temperature was cooling down, more bainite ferrites were nucleated from the boundaries. The bainite ferrite is the ferrite in bainite, especially in the upper bainite in which the ferrite and the caribes are distributed individually. At about 529.6°C, B2-1 and B2-2 were nucleated from the inclusions. It should be noted that this kind of ferrite was significantly different from the AF. Since it had a large aspect ratio, it was recognized as LB. Subsequently, the bainites of B1, B2 and B3 ceased growing as they impinged with the other ferrites, then became widened and coarsened. Finally, the secondary ferrite of B4 started to be nucleated at the prior ferrite, grew up rapidly and occupied the remained space. From the observation of the video of HTLSCM, it can be determined that the start temperature of AF is about 586.2°C, while the bainite start temperature of B_s and end temperature of B_f are 555.8°C and 476.6°C, respectively. Additionally, the quantities of AF and B3 are far less than those of the other bainites, which leads to that the bainites of B1, B2 and B4 dominate the microstructures.

It is confirmed that the AF nucleates earlier than the bainite from the in-situ observation. Figure 12 is the enlarged zone from Fig. 11(a) of Mo8 steel at the different temperatures for AF nucleating at inclusions and the growth rate. The morphologies of AFs are the black lines started from the black inclusion dots in their initial formation periods. The red arrows indicate the AF and their nucleation site. As the temperature was cooling down, some ferrites were nucleated at the inclusions with the radial shape and grew up much more slowly. Those ferrites were significantly different from B3, and could be determined as AF based on their small aspect ratio. When the temperature was about 583.6°C, several AFs were found and nucleated at the inclusions in Fig. 12(a), then many AFs were nucleated at the inclusions and grew up slowly in Figs. 12(b) and 12(c), which were indicated by the red arrows. At about 527.4°C, some AFs impinged the other AFs, ceased growing and started to coarsen as shown in Fig. 12(d). Figure 12(f) shows the average growth rate of AF1, AF2 and AF3, which ranges from 6.5°C/s and 9.1°C/s.

Regarding Mo16 steel, the HTLSCM snapshots of Mo16 steel at the different temperatures of (a–h) the nucleation and growth of the different bainites are shown in Fig. 13. When the temperature was 551.0°C, the bainite of B1-1 was first nucleated from the grain boundary. Then at 544.1°C, the bainite of B3-1 was nucleated from the intra-grain, which can also be deemed as the defect. At 531.8°C, the bainites of B2-1, B2-2 and B2-3 were nucleated simultaneously from different inclusions, and their growth rates were nearly the same in the process of cooling as shown in Fig. 13(c). Subsequently, these prior bainite ferrites stopped their growth in front of the other bainites or the grain boundary. Then these prior bainite ferrites became widened and the secondary bainites were nucleated from the prior bainites. For example, the bainites of B4-1, B4-2 and B4-3 were nucleated from the prior bainite, grew subsequently, and were restrained by the prior bainites and the grain boundaries. From the observation, the bainite start time of B_s and ending time of B_f are 551.0°C and 453.0°C, respectively. In addition, the start time of bainite from the boundary is earlier than those from defects and inclusions.

Fig. 13.

HTLSCM snapshots of Mo16 steel at the different temperatures of (a–h) the nucleation and growth of the different bainites. (Online version in color.)

3.3.3. Bainite Growth Rate

To further investigate the influence of Mo content on the phase transformation, Fig. 14 shows the growth rates of the different bainite ferrites observed with HTLSCM of (a) Mo8 and (b) Mo16 steels. The growth rates are in the increasing order of B3<B1<B2<B4 for Mo8 steel, and B1<B3<B2<B4 for Mo16 steel. In general, it can be concluded that the growth rate of bainite nucleated from the grain boundary is lower than that from the inclusion and the prior bainite. Increasing Mo content also increases the bainite growth rate which may be contributed to the larger undercooling level, because the addition of Mo can decrease the banite transformation temperature of Bs.⁸⁾

3.4. Dilatation and Phase Transition

Dilatation is an efficient way to investigate the phase transition in steels. Figure 15 is the dilatation of Mo8 and Mo16 steels, schematic diagram of the L(T), L_γ(T) and L_α(T), and the volume fraction of the α phase in the two steels. The heating and cooling rates of Mo8 and Mo16 steels are set to be 5°C/s between the room temperature and 900°C/s. In the heating and cooling processes, the austenite starting temperatures of Ac1 and Ar1 and the austenite ending temperatures of Ac3 and Ar3 are measured. As the Mo content is increased from 0.08 to 0.16 wt%, the dilatation shows little differences between the two steels.

Fig. 15.

Dilatation of (a) Mo8 and (b) Mo16 steels, (c) schematic diagram of the L(T), L_γ(T) and L_α(T), and (d) the volume faction of α phase in the two steels. (Online version in color.)

As for the continuous cooling, the volume fraction of the α phase can be calculated with Eq. (1).²¹⁾   

$$\frac{f(T)}{f_{max}}=\frac{L(T)-L_{\gamma }(T)}{L_{\gamma }(T)-L_{\alpha }(T)}$$(1)

where f(T) is the volume fraction of the ferrite and f_max is the final volume fraction of ferrite which is assumed to be 1. L(T), L_γ(T), and L_α(T) are the relative dilatation (mm) and can be measured as illustrated in Fig. 15(c).²¹⁾ Thereby, the volume fraction of the α phase is calculated as shown in Fig. 15(d). The results show that the fraction of α phase and the rate of phase transformation change little as the Mo content is increased.

4. Discussion

As shown in Figs. 2 and 7, the microstructures in the CGHAZ and HTLSCM mainly consist of LB, GB, AF, M–A constituents and carbides in the present work. These microstructures are basically identified by their morphologies based on the previous work.²⁾ The GB phase is with the islands of M–A constituents and granular ferrite matrix, while the LB is with the islands of M–A constituents and lath ferrite matrix. The AF is the common microstructure in the HAZ, and usually nucleated from the inclusions, which has needle-shaped morphology. Wu et al.²²⁾ found that the morphology of acicular ferrite varied from lath to plate, with the length, width and thickness being normally less than 36, 6 and 3 μm, respectively.

4.1. Effect of Mo Content and Austenite Grain Size on Phase Transition

It is well known that Mo as well as Mn and Cr, can enhance the hardenability of the steel, and the microstructures are profoundly influenced by those microalloying elements., The Ac1,²³⁾ Ac3,²³⁾ bainite start temperature (Bs)²⁾ and martensite start temperature (Ms)²⁴⁾ can be evaluated by the following equations: 3   

$$\mathrm{A}\mathrm{c}3=910-230w_{\mathrm{C}}{}^{0.5}-15.2w_{\mathrm{N}\mathrm{i}}+44.7w_{\mathrm{S}\mathrm{i}}+104w_{\mathrm{V}}+31.5w_{\mathrm{M}\mathrm{o}}$$(2)

  

$$\text{Ac}1=723-107w_{\text{Mn}}-3.9w_{\text{Ni}}+29w_{\text{Si}}+16.7w_{\text{Cr}}+290w_{\text{Al}}$$(3)

  

$$\mathrm{B}\mathrm{s}=830-270w_{\mathrm{C}}-90w_{\mathrm{M}\mathrm{n}}-37w_{\mathrm{N}\mathrm{i}}-70w_{\mathrm{C}\mathrm{r}}-88w_{\mathrm{M}\mathrm{o}}$$(4)

  

$$\begin{array}{l}\text{Ms}=545-330w_{\text{C}}+2w_{\text{Al}}-14w_{\text{Cr}}-13w_{\text{Cu}}-23w_{\text{Mn}}-5w_{\text{Mo}}\\ \hspace{1em}\hspace{0.5em}-4w_{\text{Nb}}-13w_{\text{Ni}}+3w_{\text{Ti}}+4w_{\text{V}}\end{array}$$(5)

Table 4 shows the calculated and measured temperatures of Ac3, Ac1, Bs and Ms for the two experimental steels. As the Mo content is increased, the values of Ac3 and Ac1 keep consistent. The Bs is slightly decreased, while the Ms is slightly increased. Moreover, the Bs observed from HTLSCM for Mo8 and Mo16 steels are between the calculated values of Bs and Ms, indicating that the observed phase with the HTLSCM during the cooling process is bainite rather than martensite. However, HTLSCM can only observe surface of the specimen and there should be a surface effect on transformation temperature, which need to be in-depth studied and be discussed in our future work.

Table 4. Calculated and measured temperatures of Ac3, Ac1, Bs and Ms for the experiment steels.

		Ac3	Ac1	Bs	Ms
Mo8	Calculated	834.4	709.7	648.3	475.7
	Measured	827 (Dilatation)	700 (Dilatation)	555.8 (HTLSCM)	/
Mo16	Calculated	838	709.4	646.3	478
	Measured	824 (Dilatation)	690 (Dilatation)	551 (HTLSCM)	/

From the results of Figs. 2 and 7, the prior austenite grain size in the HTLSCM specimen is far larger than that of the CGHAZ sample. Owing to the long heating time of HTLSCM, the austenite grain size in HTLSCM specimens is far larger than that of the CGHAZ specimens. Many researchers clarified that the austenite grain brings the change of phase transformation in the steel. Lee et al.²⁵⁾ investigated the effect of austenite grain on the transformation kinetics of upper and lower bainite in a low-alloy steel, and indicated that the transformation rate of upper bainite is increased with increasing the grain size, the transition temperature between upper and lower bainite is irrespective with the austenite grain size. Recently, Ueji et al.²⁶⁾ proved that the austenite grain refinement could accelerate the bainite transformation and reduce the whole transformation time. Considering the austenite grain size, it can be inferred that increasing the Mo content in steel lead to smaller austenite grain size, which also affect the transformation kinetics of bainite.

4.2. Effect of Mo Content on Ferrite Growth Rate

With the assumption that the ferrite growth is controlled by carbon diffusion, the ferrite growth rate can be explained by the Zener-Hillert Eq. (6).²⁷⁾   

$$V=2\mathrm{\hspace{0.17em} }500\mathrm{\hspace{0.17em} }000\cdot \frac{\overline{D}T}{x^0}{\left(x^{\gamma (\alpha +\gamma )}-x^0\right)}^2$$(6)

where V is the growth rate (m/s). $\overline{D}$ is the diffusion coefficient of carbon in the austenite (m²/s). T is the temperature (K). x⁰ is the mole percent of the carbon in the steel which is 0.28% in the present work. x^γ/(α+γ) is the carbon concentration in the γ/(α + γ) interface (mol).

Trivedi et al.²⁸⁾ proved that the weighted-average coefficient of carbon gives an exact growth rate of ferrite. Thereby, the diffusion coefficient of $\overline{D}$ can be described as:   

$$\overline{D}=\frac{1}{x^{\gamma (\alpha +\gamma )}-x^0}\int D_{\mathrm{c}}^{\gamma }(x_{\gamma },T)d{x}_{\gamma }$$(7)

  

$$\begin{array}{l}{D}_{\mathrm{c}}^{\gamma }=0.5\times {10}^{-4}\times exp(-30x_{\gamma })\\ \hspace{1em}\hspace{0.5em}\times exp\left(-4.18\times \frac{38\mathrm{\hspace{0.17em} }300-19\mathrm{\hspace{0.17em} }000x_{\gamma }+5.5\times {10}^5{x}_{\gamma }{}^2}{RT}\right)\end{array}$$(8)

where x_γ is the carbon concentration in the austenite (mol), and $D_{\mathrm{c}}^{\gamma }$ is the diffusion coefficient of carbon in the austenite at different temperatures (m²/s); R is the mole gas constant (8.314 J/(mol∙K)).

The interfacial concentration of elements can be calculated according to the different models. In a certain solid system such as Fe–C–X ternary system, as the cooling rate is slow enough, the chemical potential of the elements is equal in the two phases zone of α + γ. The full equilibrium is so-called ortho-equilibrium (OE) which was first proposed by Hultgren.²⁹⁾ When the austenitized steel at a high temperature cool down rapidly, some interstitial elements like C and N with the large diffusion coefficients can reach the equilibrium, while some substitutional elements of X like Mn and Mo having the smaller diffusion coefficients thereby remain immobile. This kind of model can be determined as para-equilibrium (PE).³⁰⁾ T0 is the temperature at which the free energies of α and γ are the same in the α + γ zone. Bhadrshia et al.³¹⁾ proposed that the bainite reaction ceases as the carbon in the γ reaches the critical value which is near the T0 line. The T0’ line shows that the stored energy is about 400 J/mol in the bainite ferrite due to the shape change in invariant-plane strain.³²⁾

Figure 16 shows the phase diagram of ortho-equilibrium Ae3 line, para-equilibrium Ae3 line, T0 and T0’ lines with the driving force of 400 kJ/mol calculated by Thermal-Calc software. The Fe-1.0Mn-0.08Mo-0.06C phase diagram in the mass percent is calculated with the database of TCFE9, which contains only α, γ and θ phases. It can be seen that the PE-Ae3 and PE-Acm lines under PE conditions are lower than that of the OE condition. T0 line is located between the PE-Ae3 and α zone, and the T0’ line is below the T0 line. It illustrates that the temperature of ferrite transition in the T0 line is lower than that in the T0’ line if the bainite transition follows the T0 line. Figure 16(b) shows the mole fractions of carbon with PE-Ae3, T0 and T0’ lines. There is no significant difference as the Mo content is increased from 0.08 to 0.16 wt%. To further explore the effect of Mo content on the PE-Ae3, T0 and T0’ lines, the diagram of the assumed steel with the Mo content of 1 wt% is also calculated which is named Mo100. As the Mo content is increased from 0.08 to 1 wt%, the PE-Ae3, T0 and T0’ lines are slightly shifted to the upper positions.

Fig. 16.

Phase diagram of ortho-equilibrium Ae3 line, para-equilibrium Ae3 line, T0 and T0’ lines with the driving force of 400 kJ/mol calculated by Thermal-Calc software. (Online version in color.)

Figure 17 shows the comparison of the calculated ferrite growth rates along with PE, T0 and T0’ line with the measured values in Mo8, Mo16 and Mo100 steel. As the Mo content is increased from 0.08 to 0.16 wt%, the growth rates of ferrite have little difference. When the Mo content is further increased to 1 wt%, the growth rate is significantly increased. It means that increasing Mo content can increase the ferrite growth rate. In addition, the measured values from HTLSCM are plotted near the T0 zone and far from PE-Ae3 and T0’ lines. It indicates the bainite growth is under T0 condition. Additionally, the growth rate of AF is much lower than those of the bainite growth rates. Bhadershia¹⁸⁾ indicated that the bainite sheaves grow as a series-parallel platelet from the γ grain face, while the AF is nucleated from a point so that the parallel platelet cannot develop. Moreover, the lattice of AF is generated by the deformation of γ, and the iron and substitutional atom cannot diffuse during the phase transition, which leads to a slower growth rate.

Fig. 17.

Comparison of the calculated ferrite growth rates along with PE, T0 and T0’ line with the measured values in Mo8, Mo16 and Mo100 steel. (Online version in color.)

4.3. Effect of Mo Content on Microstructure and Toughness

With regard to the 420 MPa offshore engineering steel, the steels with two different Mo contents were manufactured. The results in Table 3 show that with increasing Mo content from 0.08% to 0.16%, the impact toughness of HAZ at −40°C is increased from 15 J to approximately 160 J. From the observation of microstructures by OM and SEM as shown in Fig. 2, it is hard to distinguish the difference between the two steels. Hence, it is important to investigate the crystallographic microstructure from EBSD. Both HAZ and HTLSCM specimens show the similar tendency in Figs. 4 and 7. The higher Mo content increases the area fraction of HAGBs and decreases the EGS. Especially, EGS is an important index to estimate the mechanical property of steel. The yield strength of the steel can be estimated by Eq. (5):³³⁾   

$${\sigma }_{\mathrm{y}}={\sigma }_0+{\sigma }_{\mathrm{P}}+{\sigma }_{\mathrm{S}}+{\sigma }_{\rho }+k{d}_{\mathrm{e}\mathrm{f}\mathrm{f}}^{-\frac{1}{2}}$$(9)

where σ_y is the yield strength (MPa). σ₀ is the friction stress for pure iron (MPa). σ_p is the precipitation hardening (MPa). σ_s is the solid solution hardening (MPa). σ_ρ the hardening of dislocations (MPa). $k{d}_{\mathrm{e}\mathrm{f}\mathrm{f}}^{-\frac{1}{2}}$ is the Hall-Petch grain strengthening where d_eff is the effective grain size (MPa).

The nano-scale carbides of MoC are more precipitated in the higher Mo steel as can be seen in Fig. 5. Consequently, increasing Mo element content can increase the yield strength by the precipitation hardening, solid solution hardening and grain strengthening.

In addition, to estimate the yield strength, the empirical formula of Eq. (10) is used considering that the microstructure is mainly bainite ferrite in the HAZ.³⁴⁾   

$$Ys=-84.8+2.646H_V$$(10)

Then the yield strengths of Mo8 and Mo16 steel are calculated to be 624 and 701 MPa, respectively.

The Griffith equation is as follows:³⁵⁾   

$${\sigma }_{\mathrm{F}}=\left(\frac{4\gamma E}{\pi }\right)\cdot d_{\mathrm{e}\mathrm{f}\mathrm{f}}^{-\frac{1}{2}}$$(11)

where σ_F is the critical fracture stress of the ferrite (MPa). γ is the effective surface energy that can be taken as 14 J/m². E is Young’s modulus with a value of 210 GPa. d_eff is the crack size of the ferrite (m) and can be replaced by the EGS. Then the values of σ_F are calculated to be 742 and 833 MPa for Mo8 and Mo16 steel, respectively, which are higher than their yield strengths of 624 and 701 MPa. Therefore, the Mo16 steel should have higher HAZ toughness than that of Mo 8 steel.

Figure 18 shows the schematic diagrams about the effect of the Mo content on the bainite growth behavior for Mo8 and Mo16 steels during the welding simulation. Based on the in-situ observation and the discussion results, the phase transformation behavior can be concluded as follows: As for low-Mo steel of Mo8, the AFs nucleate from the inclusions firstly. Then the different bainites nucleate from inclusions, grain boundaries and intra-grain. Subsequently, those bainites grow, stop growth and widen as they impinge the barriers like the other bainites and grain boundaries. And finally, the secondary bainites nucleate from the prior bainite and AF, and dominate the remained space. As for high-Mo steel of Mo16, the difference is that the AF cannot nucleate, and increasing Mo content decreases the Bs temperature and increases the bainite growth rate, leading to that the microstructures consist of LB mainly and GB a little. The bainites have a similar nucleating and growth behavior as the low-Mo steel. In addition, increasing Mo content also decreases the austenite grain size and the EGS of the sub-grains, so that the Charpy impact toughness of the HAZ zone is improved.

Fig. 18.

Schematic diagrams about the effect of the Mo content on the bainite growth behavior for Mo8 and Mo16 steels during the welding simulation. (Online version in color.)

5. Conclusions

The effect of Mo content on the phase transformation and the low-temperature impact toughness of 420 MPa grade offshore engineering steel is studied. The conclusions are addressed as follows:

(1) As the Mo content is increased from 0.08 to 0.16 wt%, the impact toughness at −40°C is increased from 15 to 160 J, and the hardness is increased from 269 to 297 HV_0.5. The fracture is changed from cleavage of Mo8 steel to ductile and quasi-cleavage of Mo16 steel.

(2) Increasing Mo content leads to the increase in LB and the decrease in GB and AF in the CGHAZ. In addition, EBSD results show that increasing Mo content from 0.08% to 0.16% increases the fraction of HAGBs from 45% to 58%, while the EGS is decreased from 10.4 to 6.3 μm for the HAZ of the steels. The tendency is consistent with the results of the HTLSCM specimens.

(3) The growth rates of bainite ferrite range from 39.7 and 134.5 μm/s for Mo8 steel, and 51.2 to 165.6 μm/s for Mo16 steel. The ferrite growth is controlled by carbon diffusion and can be evaluated under the T0 condition. With increasing Mo content, the growth rate of ferrite is increased. However, there is no evident difference in the growth rates between the two steels due to the small difference of Mo contents between 0.08 and 0.16 wt%. In addition, the growth rate of AF in Mo8 steel ranges from 6.5 and 9.1 μm/s which is much smaller than that of bainite ferrite.

(4) In-situ observation shows that increasing Mo content decreases the austenite grain size from 114 to 100 μm. As for Mo8 steel, the AFs nucleate from the inclusions firstly. Then the different bainites nucleate, grow, stop growth and widen as they impinge the other bainites and grain boundaries. And finally, the secondary bainites nucleate from the prior bainite and AF. As for Mo16 steel, the AF cannot nucleate, and the bainites have a similar nucleating and growth behavior as the Mo8 steel.

Acknowledgments

This work is financially supported by the National Natural Science Foundation of China (Grant No. U1960202), the Science and Technology Commission of Shanghai Municipality (No. 19DZ2270200) and the Opening Project from Shanghai Engineering Research Center of Hot Manufacturing (Grant No. 18DZ2253400).

References

• 1)   Y.  Kitagawa and  H.  Kawasaki: Kobe Steel Eng. Rep., 63 (2013), 16.
• 2)   S. G.  Lee,  D. H.  Lee,  S. S.  Sohn,  W. G.  Kim,  K. K.  Um,  K. S.  Kim and  S.  Lee: Mater. Sci. Eng. A, 697 (2017), 55. https://doi.org/10.1016/j.msea.2017.04.115
• 3)   P.  Mohseni,  J. K.  Solberg,  M.  Karlsen,  O. M.  Akselsen and  E.  Østby: Metall. Mater. Trans. A, 45 (2014), 384. https://doi.org/10.1007/s11661-013-2110-3
• 4)   J.  Yang,  K.  Zhu,  R.  Wang and  J.  Shen: Steel Res. Int., 82 (2011), 552. https://doi.org/10.1002/srin.201100046
• 5)   J.  Yang,  L.  Xu,  K.  Zhu,  R.  Wang,  L.  Zhou and  W.  Wang: Steel Res. Int., 86 (2015), 619. https://doi.org/10.1002/srin.201400313
• 6)   T.  Kato,  S.  Sato,  H.  Ohta and  T.  Shiwaku: Kobe Steel Eng. Rep., 61 (2011), 32.
• 7)   Y.  Zhang,  J.  Yang,  D.  Liu,  X.  Pan and  L.  Xu: Metall. Mater. Trans. A, 52 (2021), 668. https://doi.org/10.1007/s11661-020-06105-4
• 8)   J.  Kong and  C.  Xie: Mater. Des., 27 (2006), 1169. https://doi.org/10.1016/j.matdes.2005.02.006
• 9)   S.  Khare,  K.  Lee and  H. K. D. H.  Bhadeshia: Int. J. Mater. Res., 100 (2009), 1513. https://doi.org/10.3139/146.110222
• 10)   H. J.  Hu,  G.  Xu,  L.  Wang,  Z. L.  Xue,  Y. L.  Zhang and  G. H.  Liu: Mater. Des., 84 (2015), 95. https://doi.org/10.1016/j.matdes.2015.06.133
• 11)   W.  Mu,  P.  Hedström,  H.  Shibata,  P. G.  Jönsson and  K.  Nakajima: JOM, 70 (2018), 2283. https://doi.org/10.1007/s11837-018-2921-1
• 12)   X.  Wan,  K.  Wu,  L.  Cheng and  R.  Wei: ISIJ Int., 55 (2015), 679. https://doi.org/10.2355/isijinternational.55.679
• 13)   Y.  Shen and  C.  Wang: Metall. Mater. Trans. A, 50 (2019), 4955. https://doi.org/10.1007/s11661-019-05420-9
• 14)   D.  Zhang,  H.  Terasaki and  Y.-i.  Komizo: Acta Mater., 58 (2010), 1369. https://doi.org/10.1016/j.actamat.2009.10.043
• 15)   H.  Terasaki and  Y.  Komizo: Metall. Mater. Trans. A, 44 (2013), 2683. https://doi.org/10.1007/s11661-013-1610-5
• 16)   H.  Tervo,  A.  Kaijalainen,  S.  Pallaspuro,  S.  Anttila,  S.  Mehtonen,  D.  Porter and  J.  Kömi: Mater. Sci. Eng. A, 773 (2020), 138719. https://doi.org/10.1016/j.msea.2019.138719
• 17)   K.  Shibanuma and  S.  Aihara: Tetsu-to-Hagané, 98 (2012), 184 (in Japanese). https://doi.org/10.2355/tetsutohagane.98.184
• 18)   H. K. D. H.  Bhadeshia and  J. W.  Christian: Metall. Trans. A, 21 (1990), 767. https://doi.org/10.1007/BF02656561
• 19)   X.  Zou,  J.  Sun,  H.  Matsuura and  C.  Wang: Metall. Mater. Trans. B, 49 (2018), 2168. https://doi.org/10.1007/s11663-018-1326-4
• 20)   X. D.  Zou,  J. C.  Sun,  H.  Matsuura and  C.  Wang: Metall. Mater. Trans. A, 51 (2020), 1044. https://doi.org/10.1007/s11661-019-05604-3
• 21)   D.  Quidort and  Y. J. M.  Brechet: ISIJ Int., 42 (2002), 1010. https://doi.org/10.2355/isijinternational.42.1010
• 22)   K. M.  Wu: Scr. Mater., 54 (2006), 569. https://doi.org/10.1016/j.scriptamat.2005.10.054
• 23)   N.  Huda,  A. R. H.  Midawi,  J.  Gianetto,  R.  Lazor and  A. P.  Gerlich: Mater. Sci. Eng. A, 662 (2016), 481. https://doi.org/10.1016/j.msea.2016.03.095
• 24)   K.  Ishida: J. Alloy. Compd., 220 (1995), 126. https://doi.org/10.1016/0925-8388(94)06002-9
• 25)   S.-J.  Lee,  J.-S.  Park and  Y.-K.  Lee: Scr. Mater., 59 (2008), 87. https://doi.org/10.1016/j.scriptamat.2008.02.036
• 26)   R.  Ueji,  Y.  Kimura,  K.  Ushioda,  T.  Ohmura and  T.  Inoue: ISIJ Int., 61 (2021), 582. https://doi.org/10.2355/isijinternational.ISIJINT-2020-389
• 27)   M.  Hillert: Metall. Mater. Trans. A, 25 (1994), 1957. https://doi.org/10.1007/BF02649044
• 28)   R.  Trivedi and  G. M.  Pound: J. Appl. Phys., 38 (1967), 3569. https://doi.org/10.1063/1.1710174
• 29)   M.  Hillert and  J.  Ågren: Scr. Mater., 50 (2004), 697. https://doi.org/10.1016/j.scriptamat.2003.11.020
• 30)   Y.  Xia,  G.  Miyamoto,  Z.-G.  Yang,  C.  Zhang and  T.  Furuhara: Metall. Mater. Trans. A, 46 (2015), 2347. https://doi.org/10.1007/s11661-015-2866-8
• 31)   H. K. D. H.  Bhadeshia and  D. V.  Edmonds: Acta Metall., 28 (1980), 1265. https://doi.org/10.1016/0001-6160(80)90082-6
• 32)   H. K. D. H.  Bhadeshia: Acta Metall., 29 (1981), 1117. https://doi.org/10.1016/0001-6160(81)90063-8
• 33)   S.  Morito,  H.  Yoshida,  T.  Maki and  X.  Huang: Mater. Sci. Eng. A, 438–440 (2006), 237. https://doi.org/10.1016/j.msea.2005.12.048
• 34)   E. J.  Pavlina and  C. J.  Van Tyne: J. Mater. Eng. Perform., 17 (2008), 888. https://doi.org/10.1007/s11665-008-9225-5
• 35)   T.  Hanamura,  F.  Yin and  K.  Nagai: ISIJ Int., 44 (2004), 610. https://doi.org/10.2355/isijinternational.44.610