Effect of Direct Rolling Process on Microstructure and Mechanical Properties of the Electron Beam Cold Hearth Melting Ti-6Al-4V Alloy

Abstract

Titanium alloy is a key supporting material in the field of engineering technology and high-tech, and also an extremely important national defense strategic metal material. However, the high cost limits its wider application. Direct rolling of titanium alloy melted by electron beam cold hearth melting (EBCHM) technology is considered to be an important low-cost plate production process. The rolling process has a considerable influence on the microstructure and mechanical properties of the alloy. Therefore, Ti-6Al-4V alloy melted by EBCHM technology was investigated in this study. The effects of different rolling processes on the microstructural evolution and mechanical properties of titanium alloy plates were evaluated. The results show that with the increase of deformation amount and rolling temperature, the more obvious the disintegration of lamella α and the higher the degree of equiaxation when the deformation is below the β transus. However, only dynamic recovery occurs when rolling temperature above the β transus, and lamella α does not undergo disintegration and equiaxation. With ultimate tensile strength (UTS) of 1076 MPa and elongation (EI) of 11%, the plate with 90% deformation at 950 °C has good strength and plasticity.

1. Introduction

Titanium alloys have many advantages, such as low density, high specific strength, corrosion resistance, good high and low temperature characteristics, etc. Titanium alloys are widely used in aerospace, military, marine engineering, chemical engineering, medical treatment and other fields. However, due to the high chemical activity, low thermal conductivity, low elastic modulus and other factors of titanium alloys, it is difficult to melt and process titanium alloys, and the high production cost limits its wider application [1,2]. Reducing the production cost of titanium alloy and its components has become one of the important issues that many scientific research institutions and production units have been tackling [3,4]. Using EBCHM technology to melt titanium alloy and directly roll it into products is considered to be an important low-cost technology route.

The typical method of preparing titanium alloy plate is to convert the round ingot, which is cast by multiple-running vacuum arc remelting (VAR) processes, into a flat ingot through a forging process, then mill to remove the oxide layer, and finally roll into finished plates through multiple rolling passes. Titanium alloy flat ingots can be melted and casted by EBCHM technology, which can then be used right away for rolling plates [5]. By streamlining the process flow and increasing material utilization, it can produce plates more affordably than the traditional VAR ingot process by skipping the hot working step that converts round ingots to flat ingots and the associated surface treatment step [6,7].

The plate is obtained by directly rolling EBCHM ingot, which saves the link of refining the original as-cast structure through forging. In the process of direct rolling into plates, in addition to achieving the target size, the fragmentation and equiaxation of as-cast lamellar structure become the key to ensure excellent mechanical properties. Weiss et al. [8] put forth a grain boundary separation model to explain how platelet structure fragments. Platelet α underwent deformation, forming an α/α sub-grain boundary and shear band. Diffusion embedded β phase into it, causing it to separate and become equiaxed. The α laths shear model was presented by Seshacharyulu et al. [9]. Near the shear line, deformation causes a significant number of dislocations with various signs. Dynamic recovery causes the dislocations with the same sign to form a new interface while the dislocations with different signs cancel each other out. The newly formed interface migrates through diffusion to lower the interface energy, eventually causing the platelet α to become equiaxed. The phase composition, morphology, distribution, dislocation, texture, and other microstructure characteristics of a titanium alloy largely determine its properties [10], and the degree of deformation and temperature have a significant impact on these microstructure traits. Kalinyuk et al. [11] investigated the microstructure and properties of plates under various hot working conditions with the aim of producing low-cost Ti-6Al-4V alloy plate by directly rolling EBCHM ingot. Wood [12] described the typical microstructure and characteristics of alloy plates that were directly rolled to a thickness between 25.4 mm and 50.8 mm. The effects of unidirectional and cross rolling, as well as heat treatment procedures, on the microstructure and properties of sheet metal were reported by Feng et al. [13]. The rolling process has a significant effect on the microstructure evolution of titanium alloy, and then affects the mechanical properties of the prepared plates. Therefore, it is very necessary to study the technology of direct rolling casting ingot.

In this paper, the Ti-6Al-4V alloy cast by EBCHM was directly rolled at different temperatures and deformations. The effects of different rolling processes on the microstructure and mechanical properties of alloy plates were studied. This will provide a theoretical basis and scheme for industrial production of Ti-6Al-4V alloy plates with excellent performance and low cost.

2. Materials and Methods

2.1. Materials

Sponge titanium with grade MHT-100, Al-55V alloy, and commercial-purity aluminum beans are the experimental materials. Given the loss of aluminum due to volatilization during the EBCHM process, the Al content is added in the range of 7.0 wt.% to 8.0 wt.%, and the V content is added in the nominal proportion of 4.0 wt.%. The components are combined, pressed, and joined. They are then processed in a big, 3150 kW electron beam furnace after being once melted by VAR process, alloyed, and homogenized. Figure 1 depicts the appearance and optical microstructure of the cast ingot, which, at the microscopic level, has a typical cast widmanstatten microstructure. The continuous heating metallographic method yielded a value of 1000 °C for the α + β/β transition temperature (T_β) of this alloy. The chemical composition of the titanium alloy ingot was measured by inductive coupled plasma emission spectrometer (ICP, Optima 8000), as shown in

Table 1. Chemical composition of Ti-6A-4V alloy (wt.%).

Ti	Al	V	Fe	C	N	O	H
Bal.	6.18–6.36	3.98–4.13	0.033–0.051	0.011–0.013	0.006–0.008	0.076–0.089	0.001–0.002

.

2.2. Direct Rolling Process

The rolling equipment used in this experiment is a two-roll mill (LG-300-6), and the working roll diameter is 170 mm. Splines with 120 mm × 30 mm × 8 mm were cut from the ingot for unidirectional rolling. The rolling temperatures were 850 °C, 950 °C and 1100 °C respectively, and the rolling deformation was 30%, 60% and 90%, respectively, as shown in Figure 2. The rolled plate was annealed at 700 °C for 2 h.

2.3. Testing

(1) Microstructure test. The microstructure of the metallographic specimen was observed by optical microscope (OM; OLYMPUS GX51) after it had been etched for 30 to 60 s by an etching solution with the volume ratio of HF:HNO₃:H₂O = 1:3:7.

(2) Tensile properties test. On the electronic universal testing machine UTM4304 at room temperature, a tensile test was performed. The tensile rate was 0.9 mm/min. The tensile specimens built in a horizontal plane for tensile property evaluation, as shown in Figure 3.

(3) Electron back-scatter diffraction (EBSD) test. Figure 4 illustrates how EBSD samples are cut from the rolled spline. The direction perpendicular to the RD-TD plane is designated as ND, while the transverse and rolling directions of the plate are designated as TD and RD, respectively. The Gemini SEM 300 field emission scanning electron microscope (SEM) was used for the EBSD experiment. It was outfitted with an Oxford Nordlys Nnos EBSD probe, and the central area of RD-ND surface was used as the detection surface. Electrolytic polishing was used to prepare EBSD samples with a volume ratio of 2:8 HClO₄:CH₃COOH. Voltage 40–60 V, current 0.6–2.5 A, temperature 10–15 °C, and corrosion time 50–90 s were the experimental parameters.

3. Results

3.1. Microstructural Characterization and Analyze

3.1.1. Metallographic Analysis

The OM microstructure in the middle area of RD-ND section of the alloy plate under various rolling conditions is shown in Figure 5. The original platelet α is partially twisted and some platelet α still exists in the form of a bundle after rolling at 850 °C when the deformation is 30%. More recrystallized α grains appear when the deformation reaches 60%, but platelet structure still predominates in the microstructure. The platelet α appears obviously equiaxed at a deformation of 90%, and the portion of the platelet α that has not yet completely disintegrated rotates, forming a layered distribution feature. The original platelet α is primarily twisted while rolling at 950 °C with a 30% deformation, but the equiaxed α grains have emerged. Platelet α clearly separates and disintegrates when the deformation reaches 60%, and the α grains take on a necklace-like shape. The platelet α fully breaks up, dynamically recrystallizes, and yields the equiaxed structure at a deformation of 90%. The β-transformed structure is always produced after rolling at 1100 °C under various deformations. The β grains flatten and lengthen as deformation increases, but the original structure is not disrupted.

3.1.2. Microstructural EBSD Analysis

Figure 6 illustrates the recrystallization of grain α following 90% deformation at various temperatures. The blue portion of the figure represents the recrystallized grain, the yellow portion the substructure, and the red portion the deformed grain. Following rolling deformation at 850 °C, it can be seen that the proportion of substructure is 43%, that of deformed grains is 35%, and that only 22% of them are recrystallized grains. This shows that dynamic recovery is the primary factor and dynamic recrystallization is the supporting factor during deformation. The percentage of recrystallized grains after rolling at 950 °C is 71%, while the percentage of deformed grains and substructure is low. This shows that significant dynamic recrystallization took place during rolling. After rolling at 1100 °C, the proportions of substructure, recrystallized grain, and deformed grain are 60%, 31%, and 9%, respectively.

The inverse pole figure (IPF) maps show different preferred grain orientation for alloy plate α phase deformed by 90% at various temperatures, shown in Figure 7. When rolling at 850 °C, the majority of the grains are blue-calibrated, indicating that they are primarily distributed to <01$\overline{1}$0>. Among these, the recrystallized grains are primarily calibrated by the transition colour, which is obviously different from subgrains and deformed grains, indicating that the grains rotate during recrystallization. The majority of the grains are red or nearly red when rolling at 950 °C, and among them, the orientation of subgrains and deformed grains rotates noticeably and differs noticeably from that of rolling at 850 °C. The majority of the grains are also concentrated in <0001>. Platelet α’s orientation is loose at 1100 °C when the deformation temperature is reached, but no concentrated orientation is created.

Generally speaking, 2° ≤ θ (misorientation angle) ≤ 15° is defined as low-angle grain boundary (LAGBs), and θ > 15° is called high-angle grain boundary (HAGBs) [14]. There may be 12 α-phase variants in a β grain, and there are 5 types of α/α grain boundaries among the 12 α-phase variants [15,16], according to Burgers orientation relationship between α-phase and β-phase of titanium alloy and crystallographic symmetry. The distribution and frequency diagram of α-phase misorientation angle after 90% deformation at various temperatures is shown in Figure 8. As can be seen from Figure 8a, when rolling at 850 °C, most misorientation angles are distributed at 0°–10°, among which <2° accounts for a high proportion, and there are only very low frequency-misorientation peaks located around 60° and 90°, which indicates that the dynamic recrystallization degree is low, the α phase has obvious variant selection, and there are a large number of deformed grains. HAGBs does not show continuous distribution and its proportion is very low, which is consistent with the low recrystallization degree of α grains. When rolled at 950 °C, the percentage of HAGBs increased to 44.8% and distributed continuously. There were also no frequency-misorientation peaks around 10°, 60°, or 90°, indicating that there was clear variant selection and a high degree of dynamic recrystallization. When the temperature reaches 1100 °C, there are clear peaks in the orientation difference around 10°, 60°, and 90°, which only match the particular orientation relationship, and the HAGBs does not exhibit continuous distribution. This is because the original structure is not broken after rolling, resulting in β-phase transform into α-phase followed Burgers’ correlation.

The pole figures of the plates obtained after 90% deformation at various temperatures is shown in Figure 9. When rolled at 850 °C, a typical T-shaped texture [17] with a maximum polar density of 60.25 and strong texture strength is produced, which is obviously concentrated orientation in both ends of TD on {0001} pole figure. According to the distribution of pole density in the three pole figures, it can be known that the texture is divided into {10$\overline{1}$0}<11$\overline{2}$0>, which is the result of prismatic slip [18]. After rolling deformation at 950 °C, a B-shaped basal texture is formed, but the c-axis of the grain deviates from ND to TD by about 60°, the maximum pole density is 9.69, the texture type changed and the strength weakened. After rolling deformation at 1100 °C, six different α-phase orientations can be seen from the {0001} pole figures, while the number of the pole density points on the {10$\overline{1}$0} pole figure is obviously higher than that on the {11$\overline{2}$0} pole figure, which indicates that the other six α-phase variants are rotating around the <11$\overline{2}$0> axis with the first six variants.

3.2. Mechanical Properties

Figure 10 displays the tensile characteristics at various rolling temperatures after 90% deformation. The tensile property is best after rolling at 90 °C; its utimate tensile strength (UTS) is 78 MPa higher than that at 850 °C, and its elongation (EI) is 2% higher. The tensile property is the worst after rolling at 1100 °C; its UTS is 197 MPa lower than that at 950 °C, and its EI is 8% lower than that at 950 °C, and EI is significantly deteriorated. Figure 11 depicts the 90% deformed tensile fracture morphology at various rolling temperatures. While the fracture of the specimen rolled at 950 °C has micropores, deep dimples, and tearing edges in some places, the fracture of the specimen rolled at 850 °C has small and shallow dimples, local dimple bands and long tear edges in others. Both of them exhibit mixed toughness and quasi-dissociation fracture characteristics. The specimen’s fracture rolled at 1100 °C, which has dense and tiny dimples, may have resulted from the fracture of platelet α in the β crystal. The dissociation plane and secondary crack can both be seen simultaneously, displaying the properties of quasi-dissociation fracture.

4. Discussion

4.1. Influence of Deformation and Temperature on Microstructure

Figure 5 shows how the OM microstructure of the Ti-6Al-4V as-cast alloy is significantly influenced by the amount and temperature of deformation. Ti-6Al-4V alloy is a representative α + β titanium alloy, which has α- and β-phases. The percentage of β-phase in the alloy increases as temperature rises. The alloy fully enters the β single-phase region when the deformation temperature exceeds T_β. It is simple to deform β-Ti because it has a bcc structure and numerous slip systems that can be started. Since β-Ti has a high level of stacking fault energy, it benefits from dislocation cross-slip and climbing during deformation, which causes dislocations of different signs to cancel out, dislocation density to decrease, corresponding distortion energy to be low, and insufficient driving force for recrystallization [19]. The β grain primarily deforms and experiences dynamic recovery when the alloy is rolled at 1100 °C due to the combined effects of these two factors. Therefore, even if the deformation is increased to 90% at 1100 °C, the as-cast microstructure cannot be significantly refined and equiaxed.

Platelet α in the original tissue twists, breaks up, and equiaxes when the deformation temperature is lower than T_β. Platelet α’s morphological change and orientation are closely related. “Soft orientation” occurs when the c axis tends to be perpendicular to the direction of the stress. In this region, layer α is preferentially deformed and disintegrated, whereas “hard orientation” occurs when the c axis tends to be parallel to the stress direction [20,21]. With this orientation, platelet α is difficult to deform and transitions to “soft orientation” as deformation increases. The local platelet α is twisted and disintegrated, but the local platelet structure is still present when rolling at 850 °C and 950 °C, leading to an uneven microstructure. The initial hard-oriented platelet α rotates to form long strips α, and the microstructure is layered, while the initial soft-oriented platelet α completely disintegrates and becomes equiaxed when the deformation is increased to 90%. In addition, unidirectional rolling is used in this study, which is easy to cause α grains twist in the same direction, which is not conducive to the uniform refinement of the structure.

In the model of platelet α transformation to equiaxed developed by Weiss [8], Seshacharyulu [9], and others, the process benefits from an increase in deformation degree and deformation temperature. More α/α sub-interfaces will form as the deformation rises from 30% to 90%, and at the same time, the total distortion energy and recrystallization driving force will increase, which will hasten the disintegration and equiaxed platelet α. Comparing the microstructures deformed differently at 850 °C and 950 °C reveals that the deformation is the same and that a higher proportion of the equiaxed α-phase was produced after deformation at the higher temperature. This is due to the fact that an increase in deformation temperature promotes diffusion and speeds up the migration of grain boundaries and embedding of the β-phase into the α/α interface. The percentage of deformed grains and substructure is 75% after 90% deformation at 850 °C, and the percentage of LAGBs is 84.6%, indicating that dynamic recovery is the primary factor during deformation, and no significant dynamic recrystallization occurs. After 90% deformation at 950 °C, the proportion of recrystallized grains reached 71.4%, and the proportion of HAGBs reached 44.8%, forming a continuous HAGBs frequency-misorientation distribution, indicating that significant dynamic recrystallization occurred during deformation.

According to Weiss, the degree of β-phase embedding has an impact on how difficult it is to separate α/α sub-grain boundaries. The α/α interface separation is simpler when the depth of β-phase embedding is greater than half the thickness of platelet α. The transition point between α-Ti and β-Ti is 882 °C. α→β phase transformation takes place when the deformation temperature reaches 950 °C. As a result of the phase transformation, platelet α’s thickness is inevitably reduced, which facilitates the separation of the α/α interface. Additionally, as the β-phase ratio rises, the structure’s ability to coordinate deformation increases, making it simpler for platelet in the hard orientation to rotate to the soft orientation and disintegrate. Based on the above analysis, it is considered that rolling deformation in α + β phase region is more conducive to refinement and equiaxation of as-cast microstructure.

4.2. Effect of Deformation Temperature on Micro-texture

When the alloy is rolled deformed at 850 °C, the texture takes on a typical T shape (Figure 9a). The findings demonstrate that the normal compressive stress and the tensile stress along the rolling direction can be roughly equivalent to the plane strain state during the rolling deformation process [22]. As a result, the slip surface and slip direction are parallel to the plate’s surface and rolling direction, respectively. Since the long axis ratio c/a value of α-Ti (1.587) is slightly smaller than that of ideal hcp structure (1.633), cylinder {10$\overline{1}$0}<11$\overline{2}$0> slip is the main slip system [18,22]. Therefore, in the rolling process, the slip surface {10$\overline{1}$0} in this slip system tends to be parallel to the rolling surface of the plate, while the slip direction <11$\overline{2}$0> tends to be parallel to the RD. At this time, the {0001} crystal plane tends to the TD of the plate, forming a T-shaped texture. Comparing Figure 6a and Figure 7a, it can be seen that the deformed grains and subgrains are mainly distributed in the direction of <01$\overline{1}$0>, and the cylinder {10$\overline{1}$0} is just parallel to the rolling surface, so the texture at 850 °C is mainly caused by deformation.

After rolling deformation at 950 °C, the grains are mainly distributed in the direction of <0001>, and the texture type becomes basal texture (Figure 9b). It can be seen from Figure 7a that during rolling deformation at 850 °C, the grains with dynamic recrystallization have obviously rotated, showing a transition trend to <0001> direction. It can be seen that dynamic recrystallization promotes the grain rotation to <0001>. In the process of deformation, the slip between the basal plane and the prismatic does not change the orientation of the grains in the microscopic arrangement, but only makes the grains rotate around the c axis [23]. However, the slip of the pyramidal will tilt the grain, which will make the grain rotate.

Schmid factor and the critical cutting stress are crucial variables in determining whether the slip system initiates. The slip system has a greater starting probability and is more readily initiated the higher the Schmid factor [24]. The frequency chart of the Schmid factor of a rolling surface deformed by 90% at 850 °C and 950 °C is shown in Figure 12. It can be seen that the Schmid factor of pyramidal <a> slip deformed at 850 °C has a high frequency in the range of 0.33 to 0.43 and that of 950 °C has a high frequency in the range of 0.37 to 0.50. Because of this, as the temperature increases, the pyramidal <a> slip becomes simple, encouraging the grain to rotate in the direction of <0001> and resulting in the formation of the basal texture.

Burgers orientation relationship governs the transition from the β-phase to the α-phase after rolling in the β single-phase region, which is {0001}_α∥{110}_β and <110>_α∥<111>_β; no obvious variant selection occurs during rolling.

4.3. Effect of Deformation Temperature on Mechanical Properties

The mechanical properties are significantly impacted by rolling temperature. The original lamellar structure is broken and equiaxed after rolling in the two-phase region, and the area of the α/β interface increases along with the resistance to dislocation slip, resulting in a high strength [25]. When the sample experiences tensile deformation, the slip initiates in the primary α grain with the largest individual orientation factor. If there are many primary α grains, the deformation can be quickly dispersed to them, preventing stress concentration and grain cracking in individual grains [26]. Because of this, a microstructure with more α particles is much more flexible. After 90% deformation at 950 °C, there are more primary α grains, thus showing better plasticity.

The alloy structure is β transformation structure after rolling at 1100 °C, and the semi-coherent Burgers phase relationship is satisfied at the α/β phase interface of this structure [27]. Dislocation movement readily crosses the α/β interface, which has little effect on preventing dislocation slip [28]. In addition, after rolling, the α colony are still large and intertwined to a low degree, and dislocations can easily pass through, resulting in a low level of alloy strength. Although the β grain boundary can prevent dislocation slip, dislocations frequently congregate there, concentrating stress. Microcracks will form at grain boundaries and spread along grain boundaries, resulting in a decrease in alloy plasticity, when stress concentration brought on by dislocation accumulation exceeds the material’s fracture strength.

5. Conclusions

Ti-6Al-4V alloy ingot melted by EBCHM technology was directly rolled at different temperatures and deformation to prepare plate. The following conclusions can be drawn:

(1) The rolling temperature is 850 °C, and the dynamic recovery is dominant during deformation. The recrystallized grains account for only 22% during 90% deformation. At 950 °C, when the rolling deformation is 90%, the original as-cast lamellar structure is broken and equiaxed to the highest extent, with 71% of recrystallized grains and 44.8% of HAGBs. When rolling at 1100 °C, the alloy obtained by rolling with different deformation was always a β-transformed structure and failed to effectively break the as-cast structure.

(2) On the premise that the deformation amount is 90%, when the rolling temperature is 850 °C, the micro texture is a T-type texture, and when the rolling temperature is 950 °C, the B-type basal texture is obtained. The increase of temperature promotes the rotation of recrystallized grains and the start of pyramidal <a> slip, which changes the texture type.

(3) Alloy plate has best strength and plasticity, 1076 MPa and 11%, respectively, deforming 90% at 950 °C. At 1100 °C, there is a 90% reduction in deformation, the lowest strength, the worst plasticity, and quasi-dissociation fracture characteristics in the fracture morphology.