Weld Power, Heat Generation and Microstructure in FSW and SFSW of 11Cr-1.6W-1.6Ni Martensitic Stainless Steel: The Impact of Tool Rotation Rate

Abstract

Friction stir welding (FSW) is a leading technique for joining high-strength steel. This study investigates the relationship between weld power, heat generation (HG), cooling medium, and parent austenite grain (PAG) size during both FSW and submerged FSW (SFSW) processes on 11Cr-1.6W-1.6Ni Martensitic Stainless Steel. Weld power and HG were determined by measuring plunge force and tool torque at various tool rotation rates (350–550 rpm). Additionally, the PAG size and microstructural phases in the base metal (BM), thermo-mechanically affected zone (TMAZ), and stir zone (SZ) were examined using scanning electron microscopy (SEM), electron backscattered diffraction (EBSD), and X-ray diffraction (XRD). The results indicated that the SFSW of martensitic steel required a plunge force twice that of the FSW process, along with greater weld power. The heat generated during SFSW was 130% higher than in FSW at 550 rpm. Despite this, the peak temperatures in the SZ were lower in SFSW as a result of the surrounding water’s high heat absorption. This difference in thermal behavior significantly affected the microstructure. While FSW resulted in a complete phase transformation to fine PAG, SFSW showed only minimal or partial transformation and a higher strain rate. Consequently, the SZ and TMAZ in SFSW exhibited a higher hardness than in FSW.

1. Introduction

1Cr11Ni2W2MoV steel is a martensitic heat-resistant stainless steel that was created to be utilized in the production of high-temperature bearing components for aero-engine components such as aerofoil blades, arbors, disks, and turbine blades because of its exceptional mechanical characteristics, which include high strength, toughness, creep resistance, and moderate resistance to corrosion [1]. However, joining this kind of steel with conventional fusion welding techniques resulted in numerous solidification defects and poor mechanical properties. Jia and Yue [2] and Zhang and Yang [3] studied the joining of 1Cr11Ni2W2MoV steel utilizing the tungsten inert gas (TIG) welding method. They revealed that 1Cr11Ni2W2MoV joints had a heat-affected zone (HAZ) width of 14 mm and a weld width of 4.8 mm. Furthermore, because of the dendritic microstructure in the nugget zone, the welded joints exhibited a significant decrease in hardness, fatigue limit, and tensile strength. The mechanical characteristics and microstructure of flash-welded 1Cr11Ni2W2MoV steel were examined by Yuan et al. [4]. They showed that delta ferrite was present in the weld microstructure, which reduced the hardness and creep resistance.

On the other hand, Zhang et al. [5] joined 1Cr11Ni2W2MoV steel by diffusion bonding. They investigated that in order to achieve good joint quality, the surface has to be polished using 2000# grit SiC paper. However, the joint size was constrained by the equipment available, and the procedure took a long time. Finding innovative welding methods for 1Cr11Ni2W2MoV heat-resistant steel is therefore a critical problem. Because it can provide high-strength, sound welds in materials that are challenging to weld, friction stir welding (FSW), a solid-state joining technology, has become widely used in industries like aerospace, automotive, the shipbuilding sector, and railroads [6,7,8]. FSW, which was created in 1991 by The Welding Institute (TWI), has a number of benefits over traditional fusion welding techniques, such as reduced residual stress, less distortion, and the removal of problems caused by solidification [9]. Although FSW was initially applied mainly to aluminum alloys, new developments have broadened its use to include copper alloys, steels, titanium, and magnesium.

Understanding the thermal cycle and how it affects microstructural evolution, specifically the behavior and transformation of the austenite phase, is crucial to the welding process when applied to ferrous metals, especially steels. Parent austenite grains’ (PAGs) development and evolution at high temperature during the FSW of steels are crucial because they determine the final microstructure and, in turn, the mechanical characteristics of the friction stir welds. The distribution, size, and shape of the PAG have an impact on the final microstructure after cooling [10].

The heat generation, heat loss, and peak temperature are factors controlling the microstructure during and after the FSW process. The main sources of heat input during FSW are the friction between the workpiece surface and the rotating tool, as well as the significant plastic deformation that is caused in the material surrounding the tool pin and under the shoulder [11,12,13]. The weld macrostructure usually consists of three different zones: the Heat-Affected Zone (HAZ), the Thermo-Mechanically Affected Zone (TMAZ), and the Stir Zone (SZ). Submerged FSW (SFSW), which is performed underwater, was used to improve the quality of the welded joints [14]. Water quickly dissipates the heat produced due to its high thermal conductivity and heat capacity, making it an efficient cooling medium. Peak temperatures in the weld zone are lowered by this fast cooling, which may have an impact on material flow, microstructure, and the mechanical properties. But, particularly in thicker or harder alloys, the decreased heat may also make it more difficult to achieve complete material plasticization and consolidation. Consequently, a good combination of the FSW process parameters should be achieved to ensure adequate heat input.

The size and shape of the resulting PAGs are influenced by several factors during the FSW process, including the heating rate, peak temperature, time spent at high temperatures, and the initial microstructure. The degree of austenitization and following grain growth depends on the precise peak temperature and the residence time above the transition temperature [15]. PAG may become coarse if the temperature is too high or maintained for an extended period of time. The coarse PAGs tend to encourage the production of bainitic ferrite or coarse martensite laths, which may diminish mechanical characteristics like ductility and toughness. On the other hand, a more refined final microstructure, which is preferred in many technical applications, can result from finer PAGs. Moreover, the grain structure can be considerably refined by the dynamic recrystallization brought on by powerful plastic deformation in the SZ. Thermal effects, however, predominate in nearby zones like the TMAZ, where deformation is less severe [9].

The FSW process parameters that affect heat generation, peak temperature, and cooling rate include the tool material and design [16], tool rotation rate [17], tool tilt angle [18], welding speed [19,20], plunge depth [18], plunge force [21], and cooling medium [22]. The FSW tool’s shoulder diameter has been designed to enhance the workpiece’s deformation and frictional heating [16]. It was investigated that the axial force during FSW was decreased, and the heat generation was enhanced by the featured tools. Moreover, ridges on the tool shoulder provide superior mechanical qualities [23]. The thermal history and plastic strain in the stirred zone during the welding process are influenced by the tool rotation rate. The plastic strain and peak temperature were increased as the tool rotation rate increased [24]. On the other hand, welding speed has a dominant effect on the cooling rate [21]. It was reported that the peak temperature was increased and the cooling rate was decreased by decreasing the welding speed [19]. Furthermore, the tool tilt angle and plunge force have a positive proportional effect on the heat generation and peak temperature during the FSW process [18,25,26,27].

In fusion welding techniques such as Laser Beam Welding (LBW) and Gas Tungsten Arc Welding (GTAW), the relationship between heat generation, grain size of the PAGs, and mechanical properties has been thoroughly investigated [28,29]. However, because of the intricate thermo-mechanical nature of the process, this relationship is still developing in the context of FSW [22]. Furthermore, there are experimental difficulties in characterizing the parent austenite grains in FSW joints. Strong process–microstructure–property connections are urgently needed as interest in using FSW on high-strength steels (HSS) grows. Furthermore, the use of the SFSW process to enhance the quality of the HSS joints became an important research point [22,30].

The present work aims to investigate the correlation between weld power, heat generation, cooling rate, the PAG size, and hardness during FSW and SFSW of martensitic heat-resistant stainless steel. Different tool rotation rates (350, 450, and 550 rpm) were selected according to some experimental trials as well as the above literature. Moreover, the SFSW process was utilized to improve the microstructure stability and joint quality. Experimental data, including tool torque and plunge force, were utilized to determine weld power and the heat generation for each process at various tool rotation rates. Electron backscatter diffraction (EBSD) and X-ray diffraction (XRD) were employed to reconstruct the PAGs and to assess their impact on final phases after welding, dislocation density, and plastic strain in the processed zone.

2. Materials and Methods

The base metal (BM) used was a 4 mm-thick plate of 1Cr11Ni2W2MoV martensitic steel for the welding experiments. The forged steel plate BM was provided by BAOSHAN IRON and STEEL Co., Ltd., Jinan city, China and its chemical composition is given in

Table 1. Chemical composition of 1Cr11Ni2W2MoV martensitic steel.

%	C	Si	Mn	P	S	Ni	Cr	Mo	W	V	Fe
Min.	0.1	-	-	-	-	1.4	10.5	0.3	1.5	0.18	-
Max.	0.16	0.6	0.6	0.035	0.035	1.8	12	0.5	2	0.3	Bal.

. The welding process was carried out using a fully automatic machine model (EG-FSW-M1) [31]. The FSW machine is equipped with load cells to record machine data, including torque and tool plunge force, during the FSW and SFSW experiments. Figure 1 illustrates a schematic diagram of the SFSW process setup, including the features of the tool pin and shoulder. The FSW tool made from a W-25%Re alloy was employed for the process. The tool features a stepped tapered pin with a height of 3 mm, a tip diameter of 3 mm, and a root diameter of 7 mm. To enhance the stirring efficiency of the tool, a convex scrolled shoulder with a diameter of 15 mm was machined. The welding speed, dwell time, plunge depth of the tool shoulder, and tool tilt angle were fixed at 1.25 mm/s, 6 s, 0.3 mm, and 2.5°, respectively. Tool rotation rates of 350, 450, and 550 rpm were selected for the FSW and SFSW processes. A specialized clamping fixture within a steel tank was utilized for the SFSW process, as depicted in Figure 1. The tank is equipped with inlet and outlet valves to maintain a continuous flow of water over the workpiece’s top surface at a fixed water head of 15 mm. A water flow rate of 8.5 L/min was determined through preliminary trials to minimize temperature fluctuations in the water during the welding process. The water temperature before welding was measured as 25 °C. The fluctuations of the water temperature during the welding phase (after the plunge and dwell stages to the end of the weld length) were recorded as 27 ± 2 °C.

Samples for microstructure and EBSD examination were cut perpendicular to the welding direction. The samples were first ground with emery papers, progressing to a fine grit of #2500, and then mechanically polished with 0.5 µm diamond paste to achieve a final surface finish. For microstructure examination, the polished samples were subjected to a chemical solution (2 g picric acid dissolved in 50 mL ethanol) and investigated using both an Olympus PMG3 optical microscope (Olympus, Tokyo, Japan) and a QUANTA FEG250 SEM (FEI, Eindhoven, The Netherlands). The EBSD test required a final additional ion milling step on a Hitachi IM400 ion miller instrument (Hitachi, Tokyo, Japan). The EBSD test was accomplished using a GeminiSEM 500 SEM machine (ZEISS, Jena, Germany) with the Oxford instruments SYMMETRY S3 EBSD detector (Oxford Instruments, Abingdon, England). A 20 kV operating voltage, 14 mm working distance, and 0.5 µm step size were used for data acquisition. The AZtechCrystal 2.1 software was used for the post-processing of the EBSD data. Figure 2 shows the parent grain workspace in the AZtechCrystal 2.1 software. The PAG reconstruction starts with defining the parent phase, austenite (FCC), and child phase (martensite BCC), as shown in Figure 2. Then, choose the orientation relationship, for example, Kurdjumov–Sachs (KS), and set the window size (pix) as 24, which is equal to the number of crystallographic variants of martensite from one austenite grain. Thereafter, selecti a training region from one parent grain, which gives a lower refinement error and similar measured data for the child and parent grains in the pole figure. The PAG size distribution was measured utilizing ImageJ 1.44 software. Moreover, the final product phases after welding were studied using micro XRD. The XRD test was employed using a Bruker D8 DISCOVER XRD machine (Bruker AXS, Karlsruhe, Germany) with Co Kα radiation. Match 3 software was utilized for analyzing the XRD data. A Shimadzu HMV micro hardness tester (Shimadzu, Tokyo, Japan) was used to measure the Vickers hardness of the polished cross-section with a 500 g load and a 10 s dwell period. To determine the hardness value in each weld zone, the cross-section was etched following the hardness test.

3. Results and Discussion

3.1. Tool Plunge Force and Torque

The average plunge force and tool torque at the different tool rotation rates during the steady-state welding stage are illustrated in Figure 3. The plunge forces during the FSW process at 350, 450, and 550 rpm are 22, 19, and 17 kN, respectively. A higher tool rotation rate led to a reduction in plunge force. This is because the increased frictional and plastic heat generation at these speeds softened the workpiece material. The increased softening reduces the required plunge force. The plunge forces during the SFSW process exhibit a similar trend with respect to tool rotation rate as that observed in the FSW process. However, the plunge force during the SFSW process is significantly higher compared to the FSW process at the same welding parameters. During SFSW, the plunge forces at 350, 450, and 550 rpm are 45, 42, and 39 kN, respectively. Similar results were also reported by Kumar et al. [32] and Shi et al. [33].

The tool torque follows trends similar to those observed in the plunge force. During FSW, torque values were recorded at 85, 77, and 69 N.m for rotation speeds of 350, 450, and 550 rpm, respectively. In contrast, the SFSW process yielded higher torque values of 99, 89, and 80 N.m at the same rotation speeds. Similarly to the plunge force, the reduced torque at higher rotation rates can be explained by greater material softening, which decreases the resistance to tool movement around the pin and beneath the shoulder. Notably, both the plunge force and tool torque were significantly higher in SFSW (approximately 100%) compared to conventional FSW under identical parameters. This difference is primarily due to the rapid heat dissipation in the surrounding water during submerged welding. The water absorbs a substantial amount of heat, lowering the peak temperature in the SZ and reducing material plasticity. As a result, higher force and torque are required to achieve sufficient material deformation and complete the welding process operation [14].

3.2. Welding Power

The required weld power during FSW of high-melting-point alloys is essential to define a suitable FSW machine power. Choi et al. [34] conducted a detailed analysis of power consumption during FSW and reported that about 99% of the consumed power is delivered by the spindle motor. Their study also revealed that the weld power increased with the increase in the tool rotation rate speed due to higher frictional heating and plastic deformation at the workpiece–tool interface. This relationship underscores the importance of selecting a spindle motor with sufficient torque and speed capabilities.

Figure 4 compares the consumed power under various welding conditions. The welding power measured in this figure was determined from the following Equation (1) [14]:

$$\mathrm{P} \left(\mathrm{Watt}\right)={\displaystyle \frac{2\pi \mathrm{NT}}{60}}$$

(1)

where P, N, and T are the weld power in watts, tool rotation rate in rpm, and tool torque in N.m, respectively. The figure shows that an increase in the tool rotation rate led to a corresponding increase in power consumption. The consumed power during FSW at 350, 450, and 550 rpm was 3120, 3630, and 3970 watts, respectively. Furthermore, the weld power is higher during the SFSW process compared to that during the FSW process with the same parameters. At the same welding parameters, the weld power increased by about 16% during SFSW compared to the FSW process. This increase in the weld power is attributed to the need to overcome the effect of the heat absorption by water and the increase in the required tool torque during the SFSW process.

3.3. Heat Generation

Heat generation plays a crucial role in FSW, affecting microstructure evolution, joint strength and hardness, and defect formation [26,35]. Increasing the heat input results in dissolving the carbides and grain growth [35]. Figure 5 shows the heat generation during the FSW and SFSW processes at different tool rotation rates. The experimental heat generation from both the tool shoulder and pin was determined using the machine data and process parameters as mentioned in the following equations [36]:

$${\mathrm{Q}}_{\mathrm{p}}={\displaystyle \frac{{4\pi }^2}{3}}{\mathrm{\mu }\mathrm{P}\mathrm{\omega }\mathrm{R}}_{\mathrm{p}}^3$$

(2)

$${\mathrm{Q}}_{\mathrm{s}}={\displaystyle \frac{2\pi }{3}}{\mathrm{\mu }\mathrm{P}\mathrm{\omega }\mathrm{R}}_{\mathrm{s}}^3$$

(3)

where Q_p is the heat generation in watts by the pin, Q_s is the heat generation in watts by the shoulder, µ is the coefficient of friction (assumed as 0.35), P is the pressure in Pa (plunge force divided by shoulder area), ω is the angular rotation rate in rad/s, Rp is the tool pin radius in m, and Rs is the tool shoulder radius in m.

As is evident from the figure, the heat generation increases with the tool rotation rate due to the increased contact action between the tool and deformed materials, as well as the increase in plastic deformation. The heat generation during the FSW process at 350, 450, and 550 rpm was 2300, 2570, and 2800 J/s, respectively. During the SFSW process, the heat generation at 350, 450, and 550 rpm was 4730, 5670, and 6440 J/s, respectively. At the same welding parameters, the heat generation during the SFSW is higher than that during the FSW process. This effect can be explained by the higher friction between the tool surface and the less softened materials in the case of SFSW compared to that in the case of FSW.

The total heat input is the main factor governing the peak temperatures in the various metallurgical zones during FSW [37]. The total heat input is the difference between the heat generated and the heat losses. Thus, the surrounding medium has a significant effect on the final heat input and peak temperatures during the FSW process. Figure 6 illustrates the peak temperatures in the SZ at the different welding conditions, estimated numerically from our previous work [22]. The peak temperatures during FSW at 350, 450, and 550 rpm were 985, 1150, and 1260 °C, respectively. On the other hand, the peak temperatures during the SFSW process at 350, 450, and 550 rpm were 720, 845, and 920 °C, respectively. Although the heat generated during the SFSW process is 105 to 130% higher than that during the FSW process at the various rotation rates, the peak temperatures in the SFSW process are lower than those in the FSW one. This phenomenon clearly explains the significant difference between the heat loss to the air in FSW versus to the water in the case of SFSW. That effect will have a crucial impact on the PAGs at elevated temperature and consequently on the final microstructure after cooling.

3.4. Macrostructure and Microstructure Analyses

Figure 7a,b represents the macrostructure of the weld cross-section processed at 450 rpm using FSW and SFSW processes, respectively, as well as the areas of the different metallurgical zones, as shown in Figure 7c. The weld macrostructure consists of BM, HAZ, TMAZ, and SZ. It is obvious that the width and area of the metallurgical zones decrease during SFSW compared to FSW at the same welding parameters due to the low peak temperature and plastic strain. The SEM and EBSD tests were performed on the received BM as well as the TMAZ and SZ. The test locations in the TMAZ and SZ were at a depth of 1 mm beneath the workpiece’s top surface. The SEM and XRD patterns, shown in Figure 8a,b, show that the BM consists of full martensite laths (Figure 8a) and small amounts of Fe₃C and M₂₃C₆ carbides (Figure 8b). The PAGBs can be observed in the SEM micrograph. The EBSD-reconstructed PAG map and the PAG diameter distribution in the BM are shown in Figure 8c,d. The PAG diameter distribution in Figure 8d was measured using ImageJ software. The figures show that the PAG in the BM is coarse-grained with an average austenite grain diameter of 53 µm (M is the average grain size and SD is the standard deviation). Figure 8e shows the Kernel average misorientation (KAM) map in the BM, which will be compared to other zones later.

The SEM micrographs at low and high magnifications in the SZ at 450 rpm during the FSW and SFSW processes are shown in Figure 9. To reveal the microstructure in the SZ, the samples were immersed in the picral solution for 5 min. Figure 9 shows that the final microstructure in the SZ during both the FSW and SFSW processes was fully martensite laths. However, the PAGBs cannot be observed in the SEM micrographs even at high magnification. Thus, the EBSD in the TMAZ and SZ was used to reconstruct the PAGs and determine their size and distribution.

The PAG size in the processed zone at elevated temperature was affected by the peak temperature and plastic deformation during the FSW process. Furthermore, the plastic deformation and temperature during FSW were increased with the tool rotation rate [9]. The reconstructed PAG maps and the PAG diameter distribution in the TMAZ at the different tool rotation rates during FSW and SFSW processes are shown in Figure 10 and Figure 11, respectively. The PAG diameter distribution figure showed the mean PAG diameter (M) and the standard deviation (SD). The PAG maps show that the PAG diameter in the TMAZ during FSW was decreased with the increase in the tool rotation rate. The PAG diameters in the TMAZ at 350, 450, and 550 rpm during the FSW process were 31, 26, and 24 µm, respectively. At 350 rpm, the temperature and deformation in the TMAZ were lower than those at higher tool rotation rates. Consequently, the martensite in BM was partially transformed into austenite grains at elevated temperatures, resulting in fine PAG in local regions compared to the BM. Figure 10b shows the presence of very fine PAG with grain diameters below 10 µm. With an increase in tool rotation rate, the peak temperature and plastic deformation increase. This resulted in complete dynamic recrystallization and fine PAG in the TMAZ at 450 rpm and 550 rpm.

During the SFSW process, the peak temperature in the TMAZ is lower than that during the FSW one. In the SFSW process, the PAG diameters in the TMAZ at 350, 450, and 550 rpm are 49, 35, and 27 µm, respectively. Figure 11a,b shows that there is no obvious fine PAGs in the TMAZ at the 350 rpm-SFSW condition and the PAG diameter and grain distribution were slightly similar to that in BM (Figure 8c,d). This means that there is no recrystallization occurring in the TMAZ under that processing condition. By increasing the tool rotation rate (≥450 rpm), i. e., increasing the heat input, fine recrystallized PAG appeared in the TMAZ. The degree of dynamic recrystallization in the TMAZ was increased with the tool rotation rate and the amount of fine PAG was increased, as shown in Figure 11c–f. However, the current work demonstrated that dynamic recrystallization within TMAZ was less pronounced in SFSW than in conventional FSW. This suppression is directly attributable to the lower peak temperatures and reduced levels of plastic deformation generated by the SFSW process across all rotation rates investigated.

It is well known that the SZ bears a higher temperature and plastic deformation than the TMAZ. Thus, the PAGs in the SZ exhibit a higher degree of dynamic recrystallization at elevated temperatures and plastic strain compared to the TMAZ. Figure 12 and Figure 13 illustrated the PAG maps and grain diameter distribution in the SZs at various rotation rates during the FSW and SFSW processes, respectively. The PAG diameters in the SZ during the FSW process at 350, 450, and 550 rpm were 19, 22, and 23 µm, respectively. The higher heat input and peak temperature (in the single-phase region) in the SZ during FSW cause the complete transformation of the parent microstructure into fine austenite grains during the welding stage. However, the increase in the tool rotation rate resulted in larger PAG diameters, which can be attributed to the higher peak temperatures achieved. These elevated temperatures promote grain growth by providing the necessary thermal energy for grain boundary migration and coalescence. This can be observed in the reduction in the amount of the fine PAG diameters below 15 µm.

On the other hand, the peak temperature and residence time in the SZ during the SFSW process were not enough for the complete phase transformation and dynamic recrystallization. Figure 13a,b showed that the reconstructed PAG map has tiny grains, with about 65% of the grains below 10 µm in size. However, the peak temperature and deformation rate in that zone were lower compared to those in the case of the FSW process. In addition, the reconstruction results showed that about 16% of child martensite grains did not reconstruct into PAGs. This result means that most of the tiny grains at 350 rpm-SFSW were not dynamically recrystallized PAGs, and most of the parent microstructures may be deformed or crushed by the stirring action of the welding tool [20]. This finding agrees well with that detected by Li et al. [38] during the application of severe plastic deformation for the duplex stainless steel. In other words, the microstructure in the SZ at the 350 rpm-SFSW condition was deformed without or with a little amount of phase transformation or dynamic recrystallization. In fact, increasing the tool rotation rate elevated both the plastic deformation and the peak temperature through increased frictional and mechanical energy. The PAG maps and the grain distribution showed that the SZs have finer PAGs compared to BM. Compared to BM, the fraction of PAG diameters below 30 µm was increased. This effect was due to partial dynamic recrystallization that occurred in the SZ at 450 and 550 rpm during the SFSW process. Despite the higher heat generation during the SFSW process, the water’s high heat absorption lowered the peak temperature and plastic deformation. One would expect the conditions in conventional FSW, characterized by high peak temperatures, extended residence time, and a slow cooling rate may cause considerable grain growth and larger PAG diameters. However, the opposite was observed: PAG diameters at 450 and 550 rpm were larger in SFSW. This phenomenon may be attributed to the greater plastic deformation in the FSW process, which increased the number of nucleation sites for new PAGs at elevated temperatures, ultimately leading to a finer grain structure than in SFSW.

Table 2. PAG size in µm at the different welding conditions.

	350 rpm		450 rpm		550 rpm
	FSW	SFSW	FSW	SFSW	FSW	SFSW
TMAZ	31	49	26	35	24	27
SZ	19	9	22	30	23	26

summarized the PAG diameters in the TMAZs and SZs.

The Kernel average misorientation (KAM) indicated local strain due to the accumulation of geometrically necessary dislocations [39]. Figure 14 shows the KAM maps in BM, TMAZ, and SZ at 450 rpm during FSW and SFSW process.

Table 3. KAM_av values in the SZs during FSW and SFSW processes at different tool rotation rates.

Condition	BM	350 rpm	450 rpm	550 rpm
FSW	1.61	1.787	1.487	1.488
SFSW		1.943	1.659	1.584

lists the KAMav in the different zones in Figure 11, which was calculated using the following equation [40]:

$${\mathrm{KAM}}_{\mathrm{av}}={\sum }_{\mathrm{i}=1}^5\left({\mathrm{A}}_{\mathrm{f}}(\mathrm{i}-1,\mathrm{i})(2\mathrm{i}-1)/2\right)$$

(4)

where Af (i − 1, i) is the KAM area fraction between i − 1 and i degrees. The as-received forged BM shows high strain magnitude and dislocation density. The SZ at 350 rpm during the FSW process has higher KAMav than that at 450 rpm and 550 rpm. The KAMav during the FSW at 350, 450, and 550 were 1.787, 1.487, and 1.488, respectively. The KAMav during the SFSW process had the same trend. The KAMav during the SFSW process at 350, 450, and 550 rpm were 1.943, 1.659, and 1.584, respectively. This may be attributed to the complete recrystallization occurring at a high tool rotation rate, which in turn produced a dislocation-free grain. However, the SZs during SFSW have higher KAMav compared to those during FSW at the same welding parameters due to the low number of recrystallized grains in the case of SFSW. Moreover, the SZ at 350 rpm during the SFSW process had the highest KAMav, which emphasized that the tiny grains in that zone were not recrystallized grains but deformed.

3.5. XRD Analysis

Figure 15 showed the micro XRD patterns in BM, TMAZ, and SZ at 450 rpm during both the FSW and SFSW processes. It was obvious from the figure that the BM consists of full martensite and a small amount of Fe₃C and M₂₃C₆ carbides. The final microstructure in the TMAZ and SZ during the FSW and SFSW process was a full martensite phase. The carbides on the SZ and TMAZ were dissolved at a high temperature during the welding process. The other main difference in Figure 15a was the decrease in the peak’s intensity from the BM toward the SZ. This effect may be attributed to the dissolution of the carbides into the matrix, which in turn increased the carbon content of the martensite matrix. This increase in the carbon content decreased the path difference between the X-rays, which showed destructive interference. In addition, Figure 15b–f showed that the full width half maximum (FWHM) values of the peaks were increased in SZ compared to BM. This is due to the increase in the carbon atoms in the octahedral sites in the SZs. This effect increased the tetragonality of the martensite, which means higher strain and dislocation density in the SZ compared to the BM [41]. Moreover, the FWHM in SZ during the SFSW process was higher than that in the SZ during the FSW process, which means higher dislocation density. This also agrees well with the KAM maps. The higher cooling effect of the water during the SFSW compared to the FSW decreased residence time and increased strain and dislocation density.

3.6. Hardness Measurements

The hardness measurements were carried out to examine the effect of heat generation and cooling medium on the mechanical properties of the welded joints. Figure 16 shows hardness profiles of the produced joints at different tool rotation rates in the FSW (Figure 16a), the SFSW (Figure 16b), and the average hardness values in the TMAZ and SZ compared to BM (Figure 16c). The hardness profiles have similar trends at all the welding conditions. The hardness values decreased from the BM toward the HAZ and then increased again from the TMAZ until they reached their maximum values in the SZ. The hardness of the BM was 548 HV. The TMAZ and SZ have higher hardness values than the BM. The increase in the TMAZ’s hardness at the different welding parameters was attributed to the dissolution of the carbides during the welding process, which increased the carbon content in the martensite matrix and/or the decrease in the grain size due to recrystallization. The hardness values in SZ during both processes, i.e., FSW and SFSW, were decreased with increasing the tool rotation rate. This is due to the increase in heat generation, which in turn increased recrystallization and grain growth and decreased the dislocation density at high tool rotation rates. However, the SZs during SFSW have higher hardness values compared to the SZs during FSW. The highest hardness value can be found in the SZ at 350 rpm during SFSW due to the relatively low heat generation and the cooling effect of the water, which resulted in a deformed microstructure with a high magnitude of strain and dislocation density. With the increase in the tool rotation rate to 450 and 550 rpm, the increase in the heat generation and recrystallization resulted in relatively low hardness values.

4. Conclusions

The present study was conducted to explore the relationship between tool rotation rate, weld power, heat generation, cooling medium, and parent austenite grain size during the FSW and SFSW of martensitic ultra-high-strength steel. The following findings have been drawn:

• The plunge force of the tool decreased from 22 kN at 350 rpm to 17 kN at 550 rpm during the FSW process. A similar trend was observed in the plunge force during the SFSW process. However, SFSW requires nearly double the plunge force compared to FSW at the same welding parameters.
• The tool torque decreased, while the weld power increased with the increase in tool rotation rate during both SFSW and FSW processes. However, the required weld power increased by approximately 16% at SFSW compared to FSW to counteract the heat absorption effect of the cooling water.
• Heat generation increased with increasing the tool rotation rate. The heat generated during the SFSW was higher than that produced during the FSW process. At a rotation rate of 550 rpm, the heat generation at SFSW was 130% greater than that at FSW. However, the peak temperatures in the SZs during SFSW were lower than those observed during FSW.
• The parent microstructure in the TMAZ during FSW at elevated temperatures is partially transformed into fine PAGs at rotation rates of 350 and 450 rpm and is fully transformed into fine PAGs at 550 rpm. The phase transformation in the SZ during the FSW process occurred at all tool rotation rates studied. During SFSW, the parent microstructure in both the TMAZ is only deformed without undergoing dynamic recrystallization at 350 rpm and is partially transformed at 450 and 550 rpm. In the SZ at 350 rpm, the parent microstructure is deformed with very little degree of dynamic recrystallization, while partial and full dynamic recrystallization occurred at 450 and 550 rpm, respectively.
• The carbides in BM were dissolved in the SZ resulting in a high carbon content in the martensite matrix. Additionally, the SZ at SFSW exhibited a higher dislocation density and strain rate compared to both the SZ at FSW and BM.
• The average hardness values in the TMAZ and SZ were higher than the BM. Moreover, the hardness values in the SZ during SFSW were higher than those during FSW by about 6 to 7%.