Significant influence of heat input on microstructure evolution and mechanical properties of the simulated CGHAZ in a 1000 MPa grade ultra-high strength steel

ABSTRACT

The microstructure, impact properties, and crystallographic characteristics of the coarse-grained heat-affected zone (CGHAZ) in a new 1000 MPa grade ultra-high strength steel with varying heat inputs through were investigated by single-pass welding thermal simulations. As the heat input increased from 10 to 50 kJ/cm, the main microstructural constituents of the CGHAZ transitioned from lath martensite (LM) to lath bainite (LB) and subsequently to a combination of LB and granular bainite (GB). The CGHAZ exhibited a brittle fracture behaviour, and the impact energy initially increased and then declined with higher heat input. The lowest microhardness and the highest impact toughness were observed at a heat input of 20 kJ/cm. According to the results of EBSD, the proportions of high angle grain boundaries (HAGBs) and lower kernel average misorientation (KAM) values were the highest at a heat input of 20 kJ/cm. These factors collectively contributed to the enhancement of impact toughness.

KEYWORDS:

• Coarse-grained heat affected zone
• heat input
• ultra-high strength steels
• impact toughness

Introduction

Metal materials are widely used in transportation, construction, aerospace and many other industries. With the expansion of the application field of metal materials, the performance requirements for metal materials have gradually increased. The mechanical properties of metal materials are related to the microstructure. Therefore, many researchers regulate the microstructure of metal materials by various methods, so as to improve its mechanical properties, such as adding reinforcement phase [1], adjusting the alloy composition [2], controlling the preparation process [3–6], etc.

As the most widely used metal material, steel finds extensive use in hydropower stations, particularly high-grade hydropower steel employed in penstocks, spiral cases, and bifurcated pipes [7]. The hydropower industry currently holds the third position globally in electricity generation capacity. Given the swift economic development and the rising emphasis on environmental conservation, the progression of clean and renewable hydropower stands out as a significant trend. The scale of hydropower station construction is progressively expanding, accompanied by a corresponding increase in the diameter and hydraulic pressure of the penstock. Presently, the maximum diameter of penstocks has exceeded 13 m, with an HD value surpassing 5000 m². Currently, the predominant high-grade ultra-high strength steel (UHSS) is 800MPa grade. To meet the evolving requirements of hydropower units characterised by high head, high speed, high efficiency, and large capacity, the steel grade utilised in hydropower projects has advanced to 1000 MPa UHSSs, especially in specialised environments such as high-altitude hydropower stations with significant drops. The construction of large capacity pumped storage power stations further escalates the demand for 1000 MPa hydropower steel [1]. Consequently, an increasing amount of research is underway regarding the fabrication of ultra-high strength hydroelectric steel [7–10].

The welding process is inevitable when the application of ultra-high strength hydropower steel is considered. The yield strength of UHSS exceeds 550 MPa, while the tensile strength exceeds 700 MPa. To achieve optimal strength and toughness, commonly used techniques involve microalloying, controlled rolling and cooling processes, and a heat treatment procedure [11–13]. The microstructure of UHSSs is often characterised by a complex composition, including ferrite, pearlite, sorbite, residual austenite, martensite, bainite, and other microconstituents [14]. However, due to the complexity of thermomechanical strengthening mechanisms and microstructures of ultra-high strength steels, they are more susceptible to the adverse effects of welding on their properties compared to conventional carbon steels with lower strength ranges [14–16]. Typical problems related to the welding of UHSSs include cold cracking and the strain ageing embrittlement in the weld metal (WM) or heat affected zone (HAZ), HAZ softening, toughness deterioration, and lack of ductility [14,17].

Unreasonable heat input is among the contributing factors to these defects. The welding heat input alters the microstructure of UHSSs, thereby influencing the hardness, plasticity, and strength of the HAZ [18–20]. Consequently, the mechanical properties of welded joints are impacted. The HAZ can be divided into CGHAZ, fine-grained heat-affected zone (FGHAZ), inter-critical heat-affected zone (ICHAZ), and subcritical heat-affected zone (SCHAZ) as a result of distinct thermal cycles. The reduction in HAZ toughness can be primarily attributed to the phenomenon of grain growth, making the CGHAZ often the weakest part of the HAZ. The formation and development of microstructure as well as the final properties of CGHAZ are all influenced by variations in heat input during welding [21–23]. Therefore, it becomes necessary to characterise the microstructures and properties of CGHAZ of UHSSs produced during the welding at different heat inputs.

The heat input for field welding exceeds 10 kJ/cm, therefore, this study conducted welding thermal simulation experiments with varying heat inputs ranging from 10 to 50 kJ/cm. The microstructure and mechanical properties of the CGHAZ of a 1000 MPa grade UHSS through metallographic characterisation, microhardness and impact testing. The focus was on determining the effect of welding heat input on microstructure, microhardness and impact toughness of CGHAZ. The findings of the present study carry significant importance for the design of the welding processes involving 1000 MPa grade UHSS.

Materials and methods

The test specimens were extracted from a 40 mm thick plate of 1000 MPa grade UHSS. This particular steel was a new product manufactured by a Chinese steel manufacturer. The average chemical composition of the steel is presented in Table 1, while Table 2 provides the tensile properties of the steel plate at room temperature.

Table 1. The average chemical composition of the steel (wt. %).

C	Si	Mn	P	S	Ni	Cr	Mo	V	Cu	Nb	Al	Ti	B
0.13	≤0.25	≤1.00	<0.02	<0.005	2.00~2.50	1.00~1.50	0.50~1.00	≤0.05	≤0.20	0.02	0.02	0.01	0.004

Table 2. The tensile properties of the steel.

Yield strength Rp0.2 (MPa)	Tensile strength Rm (MPa)	Elongation at break A (%)	Reduction in area Z (%)
948	986	18.4	70

Square rods measuring 10 × 10 × 55 mm³ were precision-machined from the central region of the steel plate. Welding thermal simulation experiments were conducted using the Gleeble-540 system (Gleeble, U.S.A.). Based on the actual process parameters of the on-site welding process, the heat inputs were set to 10, 15, 20, 30 and 50 kJ/cm. The welding thermal simulation samples were heated to a peak temperature of 1300°C at a heating rate of 130℃/s, followed by a 1 s hold at this temperature. Other welding thermal simulation parameters are presented in Table 3. Thermocouples were strategically welded in the middle of the sample surface, ensuring that the central portion of the sample reached the appropriate temperature. Figure 1 illustrates the sampling diagram of square rods taken from the steel plate, along with the corresponding welding thermal simulation diagram. The welding thermal simulation cycle curves, recorded by the thermocouples, at varying heat inputs are presented in Figure 2.

Figure 1. Sampling of square rods from the steel plate (a), and welding thermal simulation of square rods (b).

Figure 2. Thermal cycling curves with varying heat inputs.

Table 3. Experimental parameters of the welding thermal simulation process.

Heat Inputs (kJ/cm)	Peak Temperature (℃)	Heating Rate (℃/s)	Holding Time (s)	tH (s)		Cooling Time t8/5 (s)
				Above 900℃	Above 1100℃
10	1300	130	1	6.96	3.70	4.33
15				7.49	4.28	8.04
20				8.90	5.08	10.56
30				10.72	5.93	16.00
50				15.25	8.02	26.76

Metallographic samples were cut from a thermocouple monitoring zone and subsequently subjected to mechanical grinding, polishing, and chemical etching using 4 vol. % nital solution. The microstructural analysis of the simulated CGHAZ was conducted using Zeiss Axio Vert.A1 optical microscope (OM) (Zeiss, Germany) and TESCAN CLARA scanning electron microcopy (SEM) (TESCAN, Czech Republic). To delve deeper into the microstructure and precipitates of each CGHAZ, additional examination was carried out using 2100F transmission electron microscope (TEM) (JEOL, Japan). The thin film samples with a diameter of 3 mm were made by thinning and electrolytic polishing in a solution of 90 vol. % of alcohol and 10 vol. % of perchloric acid with the voltage of 30 V to 35 V, the current of 0.60A to 0.65A, and temperature was −20°C.

Vickers microhardness of BM and thermal simulation samples were measured by a HVS-1000z Vicker hardness tester (Veiyee, China) with a load of 1000 g and a holding time of 10 s. Microhardness values at five points were measured for each sample.

After the welding thermal cycle, the thermal simulation samples were processed into a standard Charpy V-notch impact test specimens. The Charpy V-notch was positioned at the centre of the test specimens. Subsequently, impact tests were performed using a NI-500C pendulum impact test machine (NCS, China), following the recommendations of the Charpy pendulum impact testing method for metallic materials GB/T 229–2020. The impact energies of specimens under various welding thermal cycle processes were determined at a temperature of −40°C. Three impact tests were carried out under the same heat input. Afterwards, the morphologies of fractured surfaces of Charpy impact test specimens were analysed using SEM.

EBSD analysis (NordlysMax2) was conducted to investigate the crystallographic textures of the thermal simulation specimens. Prior to analysis, the samples were electropolished in an electrolyte solution containing 90 vol. % of alcohol and 10 vol. % of perchloric acid at a voltage of 30 V for 30 s. The scanning area of the samples was 200 μm × 200 μm, with step size of 0.35 μm. The Channel 5 software was used for post-processing of data.

Results and discussion

Influence of heat input on microstructure

The microstructure of the base metal (BM) is shown in Figure 3. Combining this image with the preparation process of the steel plate, it was evident that the microstructure of BM was composed of tempered sorbite.

Figure 3. (a) OM and (b) SEM micrographs of BM.

Figure 4 shows OM micrographs of CGHAZ microstructures after thermal simulation tests with varying heat inputs. Notably, the grain size of the tested steel experienced significant growth after welding thermal cycling, and the microstructure of CGHAZ varied with heat input. Clear grain boundaries of prior austenite were observable, and parallel thin laths extended from the austenite grain boundaries into the grain interior. SEM micrographs in Figure 5 reveal that the CGHAZ microstructure comprised quasi-polygonal ferrite (QF), lath martensite (LM), lath bainite (LB), and granular bainite (GB). The laths with the same orientation aggregated into lath blocks, and several lath blocks were grouped together to form a lath packet. The boundaries of a packet are highlighted in Figure 5 with a white line. This lath block served as the minimum structural unit controlling both strength and toughness [24,25]. Additionally, the lath packets with different orientations divide the prior austenite into distinct regions. The EDS results in Figure 6 revealed the presence of carbide particles within and between the laths.

Figure 4. OM micrographs of CGHAZ produced with heat inputs of (a) 10, (b) 15, (c) 20, (d) 30, and (e) 50 kJ/cm.

Figure 5. SEM micrographs of CGHAZ produced with heat inputs of (a) 10, (b) 15, (c) 20, (d) 30, and (e) 50 kJ/cm.

Figure 6. EDS of CGHAZ produced with heat inputs of (a) 10, and (b) 15 kJ/cm.

During the thermal cycle, BM underwent rapid heating to achieve complete austenitization, followed by cooling to room temperature. Different heat inputs resulted in different cooling rates, leading to distinct microstructures [26]. The results presented in Table 3 highlight a positive correlation between heat input and t8/5, indicating an inverse relationship with cooling rate. Specifically, as the heat input increased from 5 to 50 kJ/cm, there was a notable increase in cooling time, signifying a substantial acceleration in the cooling rate. The CGHAZ exhibited mixed microstructures, predominantly consisting of LM when the heat input was 10 kJ/cm. The elongated and slender LM traversed the entire grain, forming one or two LM packets within the prior austenite grain. As illustrated in Figures 4(a) and 5(a), these LM exhibited narrow widths and a tightly packed arrangement. Figure 7 illustrates the TEM of CGHAZ test specimens with different heat inputs. Further TEM analysis in Figure 7(a) revealed that the LM had a width ranging approximately between 0.2 and 1 μm, with some dark M/A islands interspersed between LM along with precipitated carbide particles inside the LM. The orientation difference between adjacent laths was observed to be minimal. However, with an increase in heat input to 15 kJ/cm, both grain size and LM width noticeably increased. Some LM were more than 1 μm wide, and few M/A islands emerged amidst them. As the heat input was further increased to 20 kJ/cm, the cooling time was prolonged and the microstructure predominantly consisted of interwoven LB. The prior austenite grain could comprise one or multiple bainite packets exhibiting diverse orientations. The increase in heat input gradually led to the transformation of the microstructure into both LB and GB, as shown in Figure 7(c). The staggered GB distribution can be seen in Figure 7(d), that the high dislocation density was evident and some dark M/A islands interspersed between lath and packet boundary. As shown in Figure 7(e), there was an augmented width of the lath compared to the low heat input, and the size of carbide particle inside the laths increased. Furthermore, precipitation coarsening weakened the inhibitory effect of precipitates on grain boundaries, and the grain size was further increased.

Figure 7. TEM of CGHAZ produced with heat inputs of (a) 10, (b) 15, (c) 20, (d) 30, and (e) 50 kJ/cm.

Influence of heat input on mechanical properties

Figure 8 illustrates the microhardness of BM and CGHAZ test specimens with different heat inputs. The microhardness value of BM was 350.4 ± 2.6 HV1. Due to the formation of LB, when the heat input was 10 kJ/cm, the hardness is higher than that of the base material. The hardness of the bainite structure is lower compared to that of martensite, thus exhibiting a decrease in microhardness with an increase in heat input. As the heat input increased from 10 to 20 kJ/cm, the microhardness value of the CGHAZ progressively decreased from 364.7 ± 1.6 HV1 to 340.0 ± 16.4 HV1. With a further increased in heat input, the size of carbides increased, resulting in a slight elevation of microhardness of about 349.4 ± 10.6 HV1 and 346.9 ± 3.9 HV1. The occurrence of cold cracking was decreased when the hardness of HAZ less than 350 HV. Therefore, in terms of mitigating the susceptibility to cold cracking, it is not advisable to employ excessively high or low heat input levels for this 1000 MPa grade UHSS. Relevant preheating measures should be implemented prior to welding for optimal results.

Figure 8. Microhardness of CGHAZ test specimens produced using heat inputs of (b) 10, (c) 15, (d) 20, (e) 30, and (f) 50 kJ/cm.

Figure 9 displays the impact load-displacement curves of BM and CGHAZ test specimens at −40°C. These test specimens were obtained using various heat inputs. The entire fracture process could be analysed by examining the load-displacement curve, encompassing the stages such as crack initiation, stable growth, instability growth, and ultimate fracture. The entire area encompassed by the curve represented the amount of impact energy required for the test specimen to fracture. The left side of the peak value of the curve corresponded to the crack initiation stage, while the right side corresponded to the crack propagation stage. The regions encompassed by the curves on both sides corresponded to the crack initiation energy Wi and crack propagation energy Wp, respectively.

Figure 9. Dynamic load-displace curves of BM and CGHAZ test specimens at −40°C. As shown, test specimens were produced using different heat inputs.

As shown in Figure 9, significant differences existed in the load-displacement curves between the BM and CGHAZ samples. In the case of the BM, the load-displacement curve of the substrate exhibited a yield and work-hardening stage after the elastic section, followed by a stable expansion stage of ductile fracture after the curve reached the peak load. This resulted in a fibrous region of impact fracture. Subsequently, there was a rapid drop-off in loading due to brittle instability expansion, creating crack propagation regions of impact fracture. As cracks extended to sample edges, plane stress states emerged, leading to a gradual decrease in loading and forming the shear-lip regions. Conversely, for CGHAZ samples, load-displacement curves did not display an obvious stable expansion stage but instead declined rapidly after reaching peak loads until brittle instability expanded enough to cause fractures.

The relationship between impact toughness and heat input of both CGHAZ and BM test specimens at −40°C is illustrated in Figure 10. It is evident that the impact toughness of the CGHAZ test specimen was significantly lower compared to that of the BM. The impact energy of BM was 124.7 ± 26.4 J. As the heat input increased, there was an initial rise followed by a decrease in crack initiation energy, crack propagation energy, and total impact energy. Notably, when the heat input reached 20 kJ/cm, the impact energy had relatively high values. The crack propagation energy serves as an indicator of the sample’s resistance to crack propagation. It is evident that in the case of CGHAZ test specimens; there was a diminished level of crack propagation energy, signifying the occurrence of brittle fracture. As the heat input increased from 10 to 50 kJ/cm, the impact energy of the CGHAZ at −40°C initially increased up to 74.4 ± 9.6 J and then decreased. The initiation of cracks in the impact test was primarily associated with the hard phase. The presence of martensite at low heat input and the formation of coarse carbides at high heat input could be considered as potential crack initiation sites owing to their inherent hardness. Therefore, the Wi of CGHAZ lower than that of BM. The propagation of cracks was influenced by both the crystal structure and microstructure, such as lath type, grain size, etc. The grain boundaries increased with grain size decreasing, that lead to greater resistance to crack propagation [27]. The coarse grains in CGHAZ had a lower inhibitory effect on crack propagation, so the Wp of CGHAZ were significantly lower than that of BM.

Figure 10. The relationship between impact energy at −40°C and heat input in the case of CGHAZ and BM test specimens.

The macro-morphologies of surfaces fractured at −40°C are illustrated in Figure 11, revealing distinct features for both the BM and the CGHAZ test specimens. The fibre regions of fractured surfaces were demarcated by white lines and labelled as F, the crack propagation regions (radiation regions) were delineated by yellow lines and labelled as R, and the shear-lip regions were outlined by orange lines and labelled as S. On the fractured surface of the BM sample, pronounced plastic deformation was evident, with prominent fibre and shear-lip regions. Conversely, the fractured surfaces of CGHAZ test specimens exhibited brittle fracture appearance characterised by minimal fibre and shear-lip regions, except for the sample produced with the heat input of 20 kJ/cm. These observations aligned with the corresponding impact load-displacement curves and impact toughness results.

Figure 11. Macro-morphologies of the surfaces fractured at −40°C in the case of (a) BM and (b-f) CGHAZ test specimens produced using heat inputs of (b) 10, (c) 15, (d) 20, (e), and (f) 50 kJ/cm.

Figure 12 displays micro-morphologies of the radiation regions in both BM and CGHAZ test specimens after the impact fracture. Notably, the presence of quasi-cleavage fractures was evident, comprising cleavage facet, dimples, and tear edges. The river pattern on these river-like cleavage facets appeared relatively short while converging towards adjacent tear edges composed of dimples. The regions with cleavage facets in the case of the BM test specimens were evidently limited in their presence, followed by the CGHAZ test specimens produced with heat input of 20 kJ/cm.

Figure 12. Micro-morphologies of crack propagation regions derived after fracture at −40°C in the case of (a) BM and (b-f) CGHAZ test specimens produced using heat inputs of (b) 10, (c) 15, (d) 20, (e) 30, and (f) 50 kJ/cm.

The lath bundle boundaries, and the lath block boundaries exhibited a significant inhibitory effect on crack propagation. Previous studies have demonstrated that ferritic structures (BCC), such as ferrite, bainite, and martensite, are susceptible to cleavage fracture along the {100} crystal planes [28,29]. Moreover, the increased orientation difference between the {100} crystal planes enhances the barrier effect of these boundaries against crack propagation. When there is minimal orientation difference between adjacent laths, no apparent crystallographic discontinuity exists. Under such circumstances, dislocation cracks tend to propagate across multiple laths along the {100} crystal planes with limited resistance from these boundaries [28].

The LM within the grains exhibited greater parallelism. For this reason, it was unable to resist crack propagation when the heat input was 10 and 15 kJ/cm. The multi-oriented bainite packets within the prior austenite grains of samples produced using the heat input of 20 kJ/cm had more block and packet boundaries to resist crack propagation, thereby resulting in improved impact toughness. However, excessive heat input led to lath and grain coarsening, i.e. to a decrease in prior austenite grain boundaries, lath block boundaries, and lath packet boundaries, consequently diminishing crack resistant capabilities.

Influence of heat input on crystallographic texture

The IPF, grain boundary, and KAM diagrams are presented in Figure 13 for heat inputs of 10, 15, 20, and 30 kJ/cm, respectively. In the grain boundary diagram, low angle grain boundaries (LAGBs) with misorientation angles ranging from 2 to 15◦ are indicated by grey lines, HAGBs with misorientation angles from 15 to 45° are marked by red lines, and HAGBs with misorientation angles larger than 45° are designated by black lines. The grain boundary statistics is illustrated in Figure 13(e). It can be seen that the misorientation angles of the grain boundaries were primarily distributed in the range of 0–10° and 50–60°. At a heat input of 10 kJ/cm, the lath blocks primarily consisted of LAGBs, with HAGBs observed between austenite grain boundaries and lath packets; however, there were relatively few lath packets within the grains. Among these angles, HAGBs larger than 15° accounted for approximately 43.5%, while those larger than 45° accounted for about 41.8%. At a heat input of 10 kJ/cm, HAGBs larger than 15° accounted for approximately 44.8%, while those larger than 45° accounted for about 41.0%. The grain boundary statistics were very close at the heat inputs of 10 and 15 kJ/cm. With an increase in heat input to 20 kJ/cm, there was an increase in different orientations within the grains, along with a rise in the proportion of HAGBs larger than 15° to approximately 56.7%. Additionally, the proportion of HAGBs larger than 45° increased to about 52.7%. Further increasing the heat input to 30 kJ/cm led to a decrease in the proportion of HAGBs boundaries to 49.2% owing to the reduction in lath packet boundaries, and the proportion of HAGBs larger than 45° was 40.2%. HAGBs, especially those with a misorientation angle larger than 45°, enhanced the impact toughness because they could arrest propagation of cracks and redirect it [30–33]. Comparing the heat inputs of 10, 15, 20, and 30 kJ/cm, the impact toughness of the CGHAZ test specimens at −40°C was the best when they were produced using heat input of 20 kJ/cm, according to the grain boundary statistics.

Figure 13. Crystallographic characteristics (GB, IPF, KAM) of CGHAZ produced using heat inputs of (a) 10, (b) 15, (c) 20, and (d) 30 kJ/cm; (e) grain boundary statistics; (f) KAM statistics.

Comparing the KAM diagram and KAM statistical data diagram in Figure 13 for different heat inputs, the KAM values provided insights into dislocation density and strain distribution. From the KAM diagram, it is evident that the majority of KAM values fell within the range of 0 to 2°, with higher values observed between the laths where the M/A component’s KAM values could reach up to 4–5°. As the heat input increased to 30 kJ/cm, there was a significant rise in KAM values within the range of 1–4°. Higher geometric dislocation densities were associated with elevated strain levels, leading to grain fracture and deformation, as well as an increase in residual stress within the metal. The initiation of crack was more likely to occur at higher KAM values. Consequently, this impaired the plastic deformation ability during the impact testing and reduced the impact toughness. According to EBSD crystallographic characteristics of CGHAZ samples, the optimal toughness of these samples was achieved when the heat input was 20 kJ/cm, which aligned with the outcomes of Charpy impact testing.

Conclusions

The microstructure, impact properties, and crystallographic characteristics of the CGHAZ of a new 1000 MPa grade UHSS produced with different heat inputs were analysed in the present study through single-pass welding thermal simulations.

1. The microstructure of the BM was composed of tempered sorbite. The primary microconstituents of the CGHAZ transitioned from LM to LB and subsequently to a combination of LB and GB, with an increase in heat input from 10 to 50 kJ/cm. The width of the lath gradually increased with the rise in heat input.

2. The microhardness of the CGHAZ with heat input of 10 kJ/cm was larger than BM, and the microhardness of test specimens decreased and then increased with higher heat input. The impact toughness tests demonstrated that the CGHAZ exhibited a pronounced brittle fracture behaviour, with the highest impact toughness observed in CGHAZ test specimens produced using the heat input of 20 kJ/cm. Macro-morphologies of the fractured surfaces revealed an absence of discernible fibre regions within the CGHAZ, with the shear lip regions being most prominent in CGHAZ test specimens produced with the heat input of 20 kJ/cm. The micro-morphologies of the fractured surfaces displayed characteristics typical for quasi-cleavage fracture.

3. At a heat input of 20 kJ/cm, the proportions of HAGBs and lower KAM values were the highest, which effectively inhibited crack propagation. These findings indicated that CGHAZ produced with the heat input of 20 kJ/cm exhibited excellent toughness. This was consistent with the results of Charpy impact testing. Therefore, it has been proved that heat input of 20 kJ/cm may be optimal for single-pass welding of UHSSs investigated in the present study.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was funded by the Research and Development Fund of Xi’an Thermal Power Research Institute Co., Ltd. [TQ-23-TYK13], the Opening Fund of Shaanxi Key Laboratory of Friction Welding Technology [20200101], and the Natural Science Basic Research Program of Shaanxi [2018JM5076].