Process and mechanism of gradient structural pure copper sheets prepared by extrusion machining

Abstract

Due to the balanced strength-plasticity, gradient structural (GS) materials have received much attention in recent years. Extrusion machining (EM), a new severe plastic deformation, can prepare GS metal sheets in one step. However, studies on the forming mechanism and mechanical properties of GS plates prepared by EM are insufficient, and there is a gap in their microstructure evolution and thermal stability. Combining numerical simulation and experiments, the GS pure copper sheets prepared under different extrusion thicknesses t _ch were systematically investigated. The results showed that the prepared sheets exhibited significant gradient distributions in equivalent strain, grain size, and hardness, which could be adjusted by varying t _ch . As t _ch increased, the material strength decreased, and the elongation tended to increase. Moreover, the increased t _ch contributed to a higher lateral extrusion machining ratio. From the coarse-grained layer to fine-grained layer, the grain orientation tended to be randomized. When the annealing temperature exceeded 250°C, the gradient structure gradually disappeared, and the hardness greatly decreased. Upon annealing at 350°C, the tensile strength and elongation approximated the initial annealed samples. Accordingly, the GS sheets prepared by EM possessed better properties and provided relevant guidance to improve the theory and application of related processes.

Introduction

The severe plastic deformation (SPD) process, as one of the feasible approaches to improve the properties of the material, has attracted widespread attention. So far, researchers have proposed many SPD processes, such as equal channel angular pressing (ECAP),^1,2 high-pressure torsion (HPT),^3,4 accumulative roll bonding (ARB), ⁵ twist extrusion (TE),^6,7 and large strain extrusion machining (LSEM).^8–10 Based on experiments, finite element, and response surface methods, the optimal conditions and parameters for ECAP were analyzed and determined by Alateyah et al. ¹¹ Gunderov et al. ¹² found that HPT significantly increased the alloy’s microhardness, indicating an improvement in strength and yield stress. Boudekhani et al. ¹³ demonstrated that the implementation of ARB technology can reduce the average grain size of nickel tungsten alloys from 10 to 0.2 μm. Following three ARB processes, the microhardness of the resulting material increased by 86%, and the tensile strength exceeded 1200 MPa, but the ductility was compromised. Compared to the original samples, Attarilar et al. ⁷ explored that the strength and hardness of copper after TE increased but the ductility was greatly reduced. Wang et al. ¹⁴ revealed that the microstructure of the tool could effectively increase equivalent strain in the LSEM and reduced cutting force.

The most prominent advantage of SPD processes is transforming the initial coarse grains into ultrafine-grained or nanocrystals at room temperature, which significantly improves the material mechanical strength.^15,16 Unfortunately, the material ductility and structural stability are bound to decrease significantly when it acquires higher mechanical strength,^17–19 which limits its application in the industry. Therefore, it is urgent to find a new process that can enhance the material strength while maintaining reasonable ductility.

Many researchers have successfully prepared gradient structural (GS) metal materials by introducing strain gradients into ultrafine/nanocrystalline materials, thereby achieving high compatibility between material strength and plasticity. GS materials refer to metals that transition gradually from one phase, structure, or composition to another. It ensures a smooth evolution of mechanical properties with changing dimensions, enhancing coordination among different structural components within the material. Consequently, it greatly optimizes the material service and overall performance, achieving a balance between strength and plasticity. ²⁰

Currently, GS materials have captured wide attention for their outstanding properties and are expected to be extensively applied in industrial production soon. Accordingly, new SPD processes are required for the efficient and green fabrication of GS metal materials. Yang et al. ²¹ fabricated gradient structures on the surface of pure copper by surface mechanical attrition treatment. Lee et al. ²² prepared metal sheets with unique gradient microstructures by single-roll angular-rolling. By torsional extrusion of commercially pure copper, Prokof’eva et al. ²³ formed a gradient structure in the material. Based on finite element simulation, Vu et al.^24,25 prepared pure aluminum sheets by cutting and lateral extrusion machining. However, there are few comprehensive studies on GS sheets prepared by EM, including the forming mechanism and mechanical properties. Moreover, there are gaps in investigating the microstructural evolution and thermal stability of the gradient structure, which limits the expansion of related theories and the enhancement of their application.

In this study, GS sheets of different extrusion thicknesses (t _ch ) were successfully prepared from annealed T2 pure copper by EM process. Through numerical simulation and experiments, the effect of extrusion thickness t _ch on the equivalent strain, hardness, and tensile properties of the fabricated sheets was clarified. Furthermore, the microstructural evolution mechanism and the thermal stability were studied, providing insights into the performance improvement and wider application of GS metal materials.

The principles of the EM process

As shown in Figure 1, the height and width of the bulk workpiece are H and D (perpendicular to the paper surface) respectively, which were taken as 10 mm in this study. Through a pressure P, the bulk workpiece is pressed into the vertical channel of the die at a constant speed. After the extrusion cutting of the extrusion tool with an extrusion angle θ = 120°, the workpiece height is reduced to H₁ and its width remains as D. Therefore, the thickness of the cutting layer t _d  = H−H₁, which is uniformly taken as 2 mm. The die includes a lateral opening through which the material can also be squeezed out. The width of the lateral opening is taken as the extrusion thickness of the sheet as t _ch , and the length of the lateral opening (perpendicular to the paper surface) namely is the extrusion width of the sheet d = D = 10 mm. Meanwhile, a suitable back pressure P _bq can be applied to control the material flow and improve the forming proportion of the sheet, but it will also increase the design and manufacturing cost of the EM die and the performance requirements of the test equipment, which will cost more manpower and material resources. Therefore, the EM process was carried out without back pressure P _bq in this paper. Of course, the EM die and test equipment can be improved in the subsequent study to conduct a further study on the EM process with back pressure P _bq . The compression displacement of the bulk workpiece is recorded as L₀, and the extrusion length of the sheet is recorded as L.

Figure 1.

Schematic of the principles of the EM process.

The gradient structural sheets produced by the EM process exhibit a significant strain gradient along the thickness direction. To calculate the equivalent strain, the EM process of the sheet is segmented into three distinct stages. In the initial stage, the material within the boundaries defined by segments BF, FI, and the inner wall of the die undergoes a process equivalent to non-equal channel angular pressing (NECAP) with a 90° angle. ²⁵ The equivalent strain is calculated as follows:

$${\epsilon }_1=\frac{1}{\sqrt{3}}(\frac{t_{\mathrm{ch}}}{t_0}+\frac{t_0}{t_{\mathrm{ch}}})$$

(1)

Since $t_0=x{t}_d$ (0 < x < 1), and the ratio of the extrusion thickness t _ch to the cutting layer thickness t _d is defined as the extrusion cutting compression ratio, denoted by λ:

$$\lambda =\frac{t_{\mathrm{ch}}}{t_d}$$

(2)

Therefore, the equivalent strain of the material in the initial stage can be simplified to the following expression:

$${\epsilon }_1=\frac{1}{\sqrt{3}}(\frac{\lambda }{x}+\frac{x}{\lambda })$$

(3)

The second stage is the material flowing from the CDE area to BCE area after passing through segment CE. The material in this stage is equivalent to completing a NECAP process with an acute angle, ²⁶ and the equivalent strain is as follows:

$${\epsilon }_2=\frac{1}{\sqrt{3}}(\left|\cot \alpha \right|+\left|\cot \beta \right|)$$

(4)

The third stage is the material flowing from the BCE area to the П area after passing through segment BE. In this stage, the material undergoes a process tantamount to non-equal channel angular pressing (NECAP) with an obtuse angle. The equivalent strain is calculated by the following equation:

$${\epsilon }_3=\frac{1}{\sqrt{3}}(\left|\cot \gamma \right|+\left|\cot \delta \right|)$$

(5)

Therefore, the equivalent strain of material flowing into the area is as follows:

$$\epsilon ={\epsilon }_1=\frac{1}{\sqrt{3}}(\frac{\lambda }{x}+\frac{x}{\lambda })$$

(6)

However, the equivalent strain of material flowing into area II is as follows:

$$\begin{array}{c}\multicolumn{1}{c}{{\mathit{\epsilon }}_{\mathrm{\Pi }}={\mathit{\epsilon }}_1+{\mathit{\eta }}_1{\mathit{\epsilon }}_2+{\mathit{\eta }}_2{\mathit{\epsilon }}_3=\frac{1}{\sqrt{3}}(\frac{\mathit{\lambda }}{\mathit{x}}+\frac{\mathit{x}}{\mathit{\lambda }})}\\ +\frac{{\mathit{\eta }}_1}{\sqrt{3}}(\left|\cot \mathit{\alpha }\right|+\left|\cot \mathit{\beta }\right|)+\frac{{\mathit{\eta }}_2}{\sqrt{3}}(\left|\cot \mathit{\gamma }\right|+\left|\cot \mathit{\delta }\right|)\hfill \end{array}$$

(7)

where η₁ and η₂ represent coefficients associated with the thicker layers of the material, these coefficients are inversely related to the distance between the material location and the inner wall of side B. Thus, the gradient structural sheets produced by the EM process exhibit a distinct strain gradient along the thickness direction.

Numerical simulation process and experimental analysis

Numerical simulation process

Numerical simulation of the EM process was carried out by DEFORM software, which could deepen the understanding of material flow and variable distribution. The workpiece was set up as a plastic body, divided into 10,000 meshes, while the other parts (mold and punch) were rigid bodies with a mesh number of 3000. Considering that the high-deforming region at the lateral channel opening was the main focus of this study, it was further refined, as shown in Figure 2. Through the Mesh Window of the DEFORM software, the value of the mesh refinement ratio for the local refinement area could be directly set to 0.1. DEFORM-2D software comes with mesh reclassification technology, it can improve efficiency and stability while ensuring computational accuracy through strategies such as adaptive division and local reclassification.

Figure 2.

Finite element mesh model.

Considering that the workpiece material used in this study was pure copper after annealing and softening treatment, which was the same as the work of Asgari et al. in the high-pressure double twisting (HPDT) process study, the constitutive model expression was $\overline{\sigma }=330{\overline{\epsilon }}^{0.235}$ . ²⁷ The constitutive model effectively characterizes the stress-strain correlation of pure copper post-annealing and softening treatment, making it highly suitable for the finite element simulation in this study. In the primary deformation zone, the parts in contact between the workpiece and the tool and die were the main heat transfer boundaries with a coefficient of 40 N/s/mm/°C, while the heat transfer coefficient between the die and the environment was 0.02 N/s/mm/°C. Additionally, the environmental temperature was set to 20°C. ²⁸

Experimental procedures

The study utilized T2 pure copper as the bulk workpiece material (annealed at 600°C duration of 50 min, furnace cooling to enhance microstructural homogeneity), which was machined into rectangular samples with dimensions of 10 mm × 10 mm × 60 mm, and the chemical composition is listed in Table 1. After annealing and softening treatment, the hardness of the material changes from 105.03 to 55.34 HV.

Table 1.

The chemical component of high-purity copper (wt%).

Cu	Zn	P	Sb	Sn	Bi	Ni	S	O	Fe
≥99.95	≤0.005	≤0.002	≤0.002	≤0.002	≤0.002	≤0.002	≤0.0014	≤0.0015	≤0.0015

The EM experiment was conducted using a four-column hydraulic press equipped with a specially designed and manufactured die. The material of the die was Cr12MoV. As a cold working die steel, Cr12MoV offered superior strength and hardness compared to Cr12 alloy steel following quenching and tempering. It was commonly chosen as the raw material for manufacturing tools and dies that endured heavy operational loads and had intricate shapes and structures. Lubrication was provided by lithium-based molybdenum disulfide grease. A gradient structural pure copper sheet prepared by the EM process was shown in Figure 3. All experiments were carried out without back pressure P _bq , and the press velocity of the punch was 20 mm/s. In this study, the control variable method was employed to systematically vary the extrusion thickness t _ch at 0.8, 1.0, 1.2, 1.4, and 1.6 mm. This approach aimed to investigate the impact of extrusion thickness t _ch in the EM process. Since the cutting layer thickness t _d was equal to 2 mm in this paper, changing the value of the extrusion thickness t _ch alone was equivalent to changing the value of the extrusion cutting compression ratio λ. To minimize experimental error, each set of parameters was replicated at least three times.

Figure 3.

A gradient structural pure copper sheet prepared by the EM process.

Due to the small size of the gradient structural pure copper sheet used in this study, standard sizing for the tensile property test sample was not feasible. Therefore, the sample was tailored to a non-standard size through wire cutting, as depicted in Figure 4. At the same time, to maintain the uniformity of the test surface and ensure the precision of the hardness test, this study employed wire cutting to slice the gradient structural pure copper sheet along its center line. The section obtained from the centerline was designated as the test surface. The hardness test samples were shown in Figure 4, and the hardness tests were carried out on different thickness layers of the gradient structural pure copper sheet by a Vickers hardness tester. All test loads were standardized at 1.96 N, with a dwell time of 15 s. The non-standard tensile property test samples were evaluated using an electronic universal testing machine at a tensile speed of 0.5 mm/min.

Figure 4.

Schematic of the hardness and the non-standard tensile property test sample.

The microstructure observation and grain size statistics of the gradient structural pure copper sheet were performed using an electron backscattered diffraction (EBSD) system and an intelligent inverted metallurgical microscope. The observed surface was consistent with the tested surface of the hardness test sample and the sample also needed to be coldly inset as well as the hardness test sample. For the EBSD sample, first of all, it was necessary to roughly grind, finely grind, roughly polish, and finely polish the sample until the mirror effect appeared. Secondly, the sample was subjected to alkaline washing – water washing – acid washing – water washing and quickly dried. Thirdly, the sample surface was electrolytically polished at room temperature with a volt of 1.8 V and a time of 7–10 min using the electrolyte (700 ml phosphoric acid + 300 ml water). Finally, the polished sample was placed on a field emission scanning electron microscope, and then the microstructure observation and grain size statistics were carried out by EBSD. For the metallurgic microscope sample, to begin with, it was also necessary to roughly grind, finely grind, roughly polish, and finely polish the sample until the mirror effect appeared. Secondly, the sample surface was metallically corroded at room temperature with a time of 25 s using the corrosion solution (50 ml hydrochloric acid + 100 ml water + 5 g ferric chloride). Finally, the corroded sample was placed on an intelligent inverted metallurgic microscope for microstructure observation.

Results and discussions

The forming process of the gradient structural sheet

Figure 5 illustrated the forming process of the gradient structural sheet. The images on the left displayed the results from finite element simulation, utilizing the Flow Net function of the DEFORM software. Different from the adaptive mesh repartition technology of DEFORM, the mesh generated by the Flow Net function only distorts and did not repartition, so it was always intact. Furthermore, the degree of mesh distortion was indicative of the extent of plastic deformation in the material. The figures on the right showed the microstructures of the gradient structural pure copper sheet observed by an intelligent inverted metallurgic microscope, and the figures on both sides corresponded to each other. The manufacturing process for the gradient structural pure copper sheet was separated into three distinct phases: the division of the workpiece, the initial stage of sheet formation, and the stabilization forming phase of the sheet.

Figure 5.

Forming process of the gradient structural pure copper sheet: (a, b) workpiece shunt stage, (c, d) initial forming stage of the sheet, and (e, f) stable forming stage of the sheet.

The workpiece shunt stage was shown in Figure 5(a) and (b). In this stage, as the punch was pressed down, the workpiece was shunted by the punch and the extrusion tool, and a little material flowed to the lateral opening, while most of the material flowed to the vertical channel below the lateral opening. The material flowing to the lateral opening began to be subjected to an “inverted branch” shearing action, which was shown in the left and right figures as mesh distortion and grain deformation respectively. At the same time, the “inverted branch” shearing action also validated the three stages of the EM process principle described previously.

The initial forming stage of the sheet was shown in Figure 5(c) and (d). In this stage, the material of the workpiece continued to flow into the vertical channel below the lateral opening and the lateral opening, in which the material flowing into the lateral opening was not only subjected to the “inverted branch” shearing action but also began to be subjected to the squeezing and rubbing action of the inner walls of both sides. The left and right figures respectively showed the further distortion of the mesh and the further deformation of grains, forming the initial sheet.

The stable forming stage of the sheet is shown in Figure 5(e) and (f). In this stage, with the flow of the material, the material flowing into the lateral opening was subjected to the double action including the “inverted branch” shearing and the squeezing and rubbing of the inner walls of both sides, and then flowed out of the lateral opening and formed the sheet steadily. The mesh of the material on side A that flowed out of the lateral opening was partially distorted and the meshes were elongated and refined, while the mesh of the material on side B that flowed out of the lateral opening was seriously distorted and the meshes were significantly elongated and refined. Therefore, the gradient structural pure copper sheets prepared by the EM process could form an obvious grain size gradient from side A to side B gradually.

Finite element simulation results and discussions of the EM process

Figure 6 showed the material flow velocities of the EM process under different extrusion thicknesses t _ch . Due to the impediment of segment BC of the extrusion tool, the low-speed triangle BCE area was formed, in which points B and C were the corner points of the die, and point E was the vertex of the low-speed triangle BCE area. The material flow velocity at point G of segment BC was almost 0, which could be defined as the “dead point.” The boundary between the material extruded through the lateral opening and the material outflowed through the vertical channel below the lateral opening was defined as the material shunt boundary. The material on the right of the material shunt boundary GHI was extruded through the lateral opening, while the material on the left was outflowed through the vertical channel. When the extrusion thickness t _ch increased from 0.8 to 1.6 mm, the material shunt boundary GHI shifted to the left gradually. Besides, when the extrusion thicknesses t _ch were 0.8, 1.0, 1.2, 1.4, and 1.6 mm, the material flow velocities at the lateral opening were 13.6, 10.6, 13.2, 11.3, 11.4 mm/s respectively.

Figure 6.

Material flow velocities under different extrusion thicknesses t _ch : (a) 0.8 mm, (b) 1.0 mm, (c) 1.2 mm, (d) 1.4 mm, and (e) 1.6 mm.

In this paper, the Measure tool of DEFORM was used to measure the extrusion lengths of the sheet under different extrusion thicknesses t _ch . When the extrusion thicknesses t _ch were 0.8, 1.0, 1.2, 1.4, and 1.6 mm, the extrusion lengths of the sheet were 24.9461, 27.1265, 29.853, 28.2527, and 28.2559 mm respectively (Figure 7).

Figure 7.

Extrusion lengths of the sheet under different extrusion thicknesses t _ch : (a) 0.8 mm, (b) 1.0 mm, (c) 1.2 mm, (d) 1.4 mm, and (e) 1.6 mm.

To characterize the forming effect of the sheet, the ratio between the volume of the sheet extruded through the lateral opening and the volume of the bulk workpiece pressed down through the vertical channel was defined as the lateral extrusion cutting ratio ω. As shown in Figure 1, since the extrusion width of the sheet d was equal to the workpiece width D, the lateral extrusion cutting ratio ω in this paper was equal to the ratio between the cross-section of the sheet extruded through the lateral opening and the cross-section of the bulk workpiece pressed down through the vertical channel. Namely, the formula of lateral extrusion cutting ratio ω could be simplified as:

$$\omega =\frac{L\times t_{\mathrm{ch}}}{L_0\times H}$$

(8)

where L was the extrusion length of the sheet, L₀ was the compression displacement of the bulk workpiece, which was 48 mm, t _ch is the extrusion thickness of the sheet, and H is the height of the bulk workpiece, which was 10 mm. Therefore, the formula of lateral extrusion cutting ratio ω could be ultimately simplified as:

$$\omega =\frac{L\times t_{\mathrm{ch}}}{48\times 10}=\frac{L\times t_{\mathrm{ch}}}{480}$$

(9)

The lateral extrusion cutting ratio ω under different extrusion thicknesses t _ch was calculated by equation (9), and the results were shown in Figure 8. When the extrusion thickness t _ch increased from 0.8 to 1.6 mm, the lateral extrusion cutting ratio increased from about 0.042 to about 0.094, and the forming proportion of the sheet increased gradually. As described above, when the extrusion thickness t _ch gradually increased from 0.8 to 1.6 mm, the material shunt boundary GHI gradually shifted to the left, so that more material was extruded through the lateral opening, and less material was outflowed through the vertical channel. Therefore, with increasing t _ch , the lateral extrusion cutting ratio ω increased gradually.

Figure 8.

Lateral extrusion cutting ratio ω under simulation for different extrusion thicknesses t _ch .

The equivalent strain of the EM process under different extrusion thicknesses t _ch was shown in Figure 9. The sheet extruded through the lateral opening was subjected to the double-action including the “inverted branch” shearing and the squeezing and rubbing of the inner walls of both sides, and then appeared an obvious equivalent strain gradient, which might indicate an obvious grain size gradient. At the same time, due to the impediment of segment BC of the extrusion tool, a higher equivalent strain area was formed near segment BC, which was similar in shape and size to the low-speed triangle BCE area described previously. This indicated that the material flowing through the low-speed triangle BCE area could produce severe plastic deformation and accumulate more strain, which was an important reason why the equivalent strain of the material of side B was significantly higher than that of side A. Besides, with the increase of extrusion thickness t _ch , the proportion of the high equivalent strain layers decreased gradually, while that of the low equivalent strain layers increased gradually.

Figure 9.

Equivalent strain under different extrusion thicknesses t _ch : (a) 0.8 mm, (b) 1.0 mm, (c) 1.2 mm, (d) 1.4 mm, and (e) 1.6 mm.

To illustrate the equivalent strain gradient across the sheet, segment PQ was established to the right of the lateral opening. There were 20 measuring points uniformly arranged on segment PQ, and the equivalent strain of these points was measured to draw the equivalent strain curve. The equivalent strain curve was shown in Figure 10. The analysis revealed that the largest equivalent strain typically occurs at point Q, while the minimum strain tended to be near point P. Material at point P was affected by the squeezing and rubbing action of the inner wall of side A, so its equivalent strain had been improved. The gradient in equivalent strain from point P to Q was visible. With increasing extrusion thicknesses t _ch of 0.8, 1.0, 1.2, 1.4, and 1.6 mm, the equivalent strains at point P were measured as 2.079, 2.076, 1.872, 1.869, and 1.862 respectively. Correspondingly, at point Q, the strains were 4.871, 5.114, 5.302, 5.479, and 5.311 respectively. As the extrusion thickness t _ch increased, the equivalent strain of the material at side A generally decreased, while that of the material at side B initially increased before decreasing. Notably, at an extrusion thickness of 1.4 mm, the strain difference between points P and Q was at its maximum, highlighting the significant impact of t _ch on the strain gradient.

Figure 10.

Equivalent strain from point P to point Q under different extrusion thicknesses t _ch .

To describe the equivalent strain change as it passed through points A and B under different t _ch , this study utilized the point-tracking function of DEFORM software. The paths A₀–A₁ and B₀–B₁ in Figure 9(c) were followed to record the equivalent strain change of the material, with the results depicted in Figure 11. The dashed lines represented the instances when the material passed through points A and B, respectively. Analyzing the figure reveals that the equivalent strain change of the side A material in the gradient structure pure copper sheet remained relatively constant before nearing point A, but gradually increased as it approaches. Particularly notable was the sharp rise in equivalent strain change upon reaching point A, followed by a tendency toward stability. This behavior was attributed to the intense shear forces acting on the material near point A. Conversely, the equivalent strain of the side B material began to increment gradually before reaching point B. Upon passing through point B, a similar sharp rise in equivalent strain was observed, followed by a stabilization trend. This response was influenced by the substantial shear forces experienced by the material upon entering the low-speed triangular region BCE.

Figure 11.

Equivalent strain change of material passing through two points A and B under different extrusion thickness t _ch .

The equivalent strain rate served as an indicator of the rate at which the material equivalent strain changes. A higher equivalent strain rate implied a faster change in strain and more intense plastic deformation. Figure 12 presented the equivalent strain rates observed during the EM process across various extrusion thicknesses t _ch . This graph illustrated how variations in extrusion thickness influenced the rate at which the material strain changes, thereby affecting the severity of its plastic deformation. When the extrusion thicknesses t _ch were 0.8, 1.0, 1.2, 1.4, and 1.6 mm, an “inverted branch” high equivalent strain rate area was formed on the left of the lateral opening, and its values were all greater than 13s. The “inverted branch” high equivalent strain rate area was the same as the “inverted branch” shearing action area described previously. Besides, since the material was subjected to the “inverted branch” shearing action on the left of the lateral opening the material had a severe plastic deformation and its strain changed quickly, so that “inverted branch” high equivalent strain rate area was formed.

Figure 12.

Equivalent strain rate under different extrusion thicknesses t _ch : (a) 0.8 mm, (b) 1.0 mm, (c) 1.2 mm, (d) 1.4 mm, and (e) 1.6 mm.

Figure 13 illustrated the extrusion cutting temperatures during the EM process for different extrusion thicknesses t _ch . When the extrusion thicknesses t _ch were 0.8, 1.0, 1.2, 1.4, and 1.6 mm, the maximum extrusion cutting temperatures were 80.5°C, 78.7°C, 77.9°C, 75.1°C, and 73.7°C respectively. With increasing t _ch , the maximum extrusion cutting temperature decreased gradually. At the same time, the highest extrusion cutting temperature generally appeared at point B or near point B, because it was usually the area where the plastic deformation of the material was most severe. Besides, a high extrusion cutting temperature area with an extrusion cutting temperature greater than 72.5°C was formed near the lateral opening. With increasing t _ch , the area decreased gradually, and the trend was consistent with the changing trend of the maximum extrusion cutting temperature. The recrystallization temperature of T2 pure copper was generally above 200°C. The overall extrusion cutting temperature was far lower than the recrystallization temperature of pure copper, so it could be considered that the extrusion cutting temperature had little effect on the microstructure of pure copper during the EM process.

Figure 13.

Extrusion cutting temperature under different extrusion thicknesses t _ch : (a) 0.8 mm, (b) 1.0 mm, (c) 1.2 mm, (d) 1.4 mm, and (e) 1.6 mm.

Experimental results and discussions of the EM process

The macro morphology of the gradient structural pure copper sheets prepared by EM process under different extrusion thicknesses was depicted in Figure 14. The samples exhibited a regular structure and good formation. To reduce the impact of accidental errors, each extrusion length measurement of the sheet should be repeated three times. When the extrusion thicknesses t _ch were 0.8, 1.0, 1.2, 1.4, and 1.6 mm, the average extrusion lengths measured by the vernier caliper were 17.3, 26.5, 31.7, 37.9, and 37.6 mm respectively.

Figure 14.

Macro morphology of the sheet under different extrusion thicknesses t _ch : (a) 0.8 mm, (b) 1.0 mm, (c) 1.2 mm, (d) 1.4 mm, and (e) 1.6 mm.

The compression displacement of the punch in the experiment and the finite element simulation were the same, both of which were 48 mm. Equation (9) was used to calculate the lateral extrusion cutting ratio ω under different extrusion thicknesses t _ch , and the results are shown in Figure 15. When the extrusion thickness t _ch increased from 0.8 to 1.6 mm, the lateral extrusion cutting ratio ω increased from about 0.029 to about 0.125, and the forming proportion of the sheet increased gradually. The results were consistent with the previous finite element simulation results.

Figure 15.

Lateral extrusion cutting ratio ω under different extrusion thicknesses t _ch .

The bulk workpiece material used in this paper was T2 pure copper with a face-centered cubic structure. After the annealing treatment, the internal stress and the work hardening of the bulk workpiece material were eliminated. Its hardness was about 55.34 HV, and it had low hardness and good plasticity. For the original annealed sample, it was necessary to grind and polish the sample until the mirror effect appeared. Secondly, the sample surface was metallically corroded at room temperature for a time of 25 s using the corrosion solution. Finally, Figure 16 showed the microstructure observed on an intelligent inverted metallurgic microscope. The figure illustrated that the grains of the raw annealed samples exhibited an equiaxial morphology with an average size of approximately 100 μm, and the size and distribution of these grains are uniform and homogeneous.

Figure 16.

Microstructure of the original annealed sample.

As illustrated in Figure 17, when the extrusion thickness t _ch was 1.4 mm, the material microstructure of side A, the intermediate transition material, and the material of side B of the GS pure copper sheets prepared by the EM process observed by EBSD. This graph clearly showed the grain of the material via the EM process appeared to evident refinement. As shown in Figure 17(a), after the shearing, squeezing, and rubbing action of the inner wall of side A, most of the grains of the material of side A were elongated grains with a size range of 3–7 μm and the average grain size was 4.83 μm. Figure 17(c) showed that the grains in the material of side B, which underwent more severe shearing and squeezing actions, were mostly fine elongated grains or even finer grains post-crushing, ranging from 0.6 to 1 μm with an average size of 0.87 μm. The intermediate transition material, as depicted in Figure 17(b), endured a degree of plastic deformation that lied between that of side A and side B, reflected in its grain size which ranges from 1 to 3 μm, averaging at 2.14 μm. Therefore, a clear gradient in grain size was evident from the material of side A, through the intermediate transition material, to the material of side B. This gradation confirmed the results predicted by finite element simulations, demonstrating the effectiveness of the EM process in refining grain structures across different regions of the copper sheets. At the same time, both A-side materials, intermediate transition materials, and B-side materials contained <001>, <101>, and <111> grain orientations, and there was no obvious grain orientation segregation.

Figure 17.

Microstructure of the gradient structural pure copper sheet observed under EBSD: (a) the material of side A, (b) intermediate transition material, and (c) the material of side B.

As presented in Figure 18, when the extrusion thickness t _ch was 1.4 mm, the microstructure of the materials of both sides of the gradient structural pure copper sheets prepared by the EM process were observed by an intelligent inverted metallurgic microscope. As shown in Figure 18(a), after the shearing, squeezing, and rubbing action of the inner wall of side A, most of the grains of the material of side A were elongated grains with a size range of 3–7 μm. As depicted in Figure 18(b), the material of side B, which experienced more severe shearing and squeezing actions, contained mostly fine elongated grains, and even finer grains post-crushing, with grain sizes between 0.6 and 1 μm. This was consistent with the results observed by EBSD, which also verified the grain size gradient in the gradient structural pure copper sheets prepared by the EM process.

Figure 18.

Microstructure of the gradient structural pure copper sheet observed under metallographic microscope: (a) the material of side A and (b) the material of side B.

During EM process, the materials separated from the surface layer of the workpiece underwent a series of actions including “inverted branch” shear and extrusion friction by the inner walls on both sides A and B. This process resulted in forming a GS pure copper sheet. The grain of the sheet suffered a complex evolution process, which could be divided into three stages, as illustrated in Figure 19. Following annealing and softening treatment, the internal stress and work hardening of the workpiece material was significantly reduced, allowing for a notable recovery of plasticity to its initial state. Before contact with the extrusion block, the grains of the workpiece consist of coarse grains similar to the original annealed sample. Upon contact with the extrusion block, material shunting occurred, leading to the initiation of “inverted branch” shear action on the material flowing toward the side opening. This process induced dislocations with varied orientations in the grain, causing the density of dislocations to gradually increase and form dislocation cells. These cells elongated in the direction of shear action, resulting in the breakage of grains and the formation of smaller grains. As the punch continued to press, the material flowing into the side opening experiences both “inverted branch” shear action and extrusion friction on the surfaces of inner walls A and B. This led to the reorganization of dislocations within the grain, transforming them into oriented sub-grains boundaries, and eventually elongated sub-grains, thereby achieving grain refinement. Due to the more intense shear and extrusion friction on side B compared to side A, the grain elongation, breakage, and refinement on side B were more thorough, resulting in a noticeable gradient in grain size as the material transitions from side A to side B.

Figure 19.

Grain evolution process of the gradient structural pure copper sheet.

The microstructures of the formed sheets were depicted in Figure 20. A notable refinement of the grain structure was evident, exhibiting a gradient pattern that varied in depth. The microstructure of the sheet could be divided into two distinct regions based on the degree of grain refinement: a low-strain zone (LSZ) and a high-strain zone (HSZ). The HSZ comprised a transitional layer (TL) and an ultra-fine grain layer (UGL). Compared to the elongated grains found in the DGL, the grain size was further reduced within the HSZ. The grains located near side B underwent even greater refinement due to the combined effects of shear, extrusion, and friction, predominantly aligning horizontally within the UGL layer. The DGL, TL, and UGL regions corresponded to the three stages in Figure 19, respectively.

Figure 20.

The microstructures of the formed sheets.

The EM process markedly increased the hardness of gradient structural pure copper sheets than that of the original annealed sample. The Vickers hardness of the gradient structural pure copper sheet from side A to side B under different t _ch , as illustrated in Figure 21. These sheets exhibited a hardness exceeding 122 HV, which was significantly higher than initial samples. The maximum hardness was generally near the surface of side B, while the minimum hardness was generally at a depth layer near the surface of side A. It was because the surface material of side A was subjected to the squeezing and rubbing action of the inner wall of side A, so the work hardening was improved and the grain was further refined, which improved the hardness.

Figure 21.

Vickers hardness from the surface of side A to the surface of side B under different extrusion thicknesses t _ch .

As the gradient structural pure copper sheet gradually transitioned from the surface of side A to side B, there was an obvious hardness gradient on the whole, and the variation trend was consistent with the variation trend of equivalent strain gradient and grain size gradient described previously. Due to the intense shearing and intense extrusion friction action, grains were elongated, broken, and refined. As the accumulated equivalent strain in the material increased, the material experienced stronger plastic deformation, resulting in finer grains. According to the mechanism of fine-grain strengthening, the finer the grain was, the greater the hardness was. Therefore, from the surface of side A to side B, the progression in hardness generally corresponded to the gradients observed in equivalent strain and grain size. Besides, when the extrusion thickness t _ch was 1.4 mm, the hardness difference of the gradient structural pure copper sheet was the maximum, reaching about 16.5 HV.

To study the tensile properties of the gradient structural pure copper sheets prepared by EM process under different t _ch , non-standard tensile workpieces were manufactured in Figure 4. The non-standard tensile samples after testing were registered in Figure 22. From left to right, they are the original annealed sample, the 0.8, 1.0, 1.2, 1.4, and 1.6 mm gradient structural samples. All the samples showed an obvious necking phenomenon after the tensile property tests.

Figure 22.

Non-standard tensile samples under different extrusion thicknesses t _ch : (a) original annealed sample, (b) 0.8 mm, (c) 1.0 mm, (d) 1.2 mm, (e) 1.4 mm, and (f) 1.6 mm.

Figure 23 displayed the engineering stress-strain curve derived from the tensile property assessment of the sheet under different t _ch , and the tensile properties were illustrated in Table 2. Analysis of Table 2 revealed that the original annealed sample exhibited the largest tensile strength of 212.57 MPa and an elongation rate of 52.62%, with a hardness of approximately 55.34 HV, consistent with the expected properties of annealed pure copper, which were low strength and hardness, and high plasticity. After EM process, the largest tensile strength of the gradient structural pure copper sheet increased obviously, while the elongation rate decreased. As t _ch raised from 0.8 to 1.6 mm, the largest tensile strength dropped from 432.84 to 399.34 MPa, while the elongation rate increased from 16.25% to 25.13%. In other words, this trend indicated that increased t _ch resulted in a gradual decrease in tensile strength and a increase in elongation rate. As described above, as t _ch increased, the proportion of the high equivalent strain layers decreased, while that of the low equivalent strain layers increased gradually. Higher accumulated equivalent stress led to finer grains. Thus, as t _ch increased, the ratio of fine grain layers gradually decreased, while the ratio of coarse grain layers gradually increased. Therefore, the tensile strength of the GS pure copper sheet decreased, while the elongation increased, that is, the plasticity increased.

Figure 23.

Engineering stress-engineering strain curve of the sheet under different extrusion thicknesses t _ch .

Table 2.

Tensile strength and elongation of the sheet under different extrusion thicknesses t _ch .

SPD process	Extrusion thickness t ch and pass	Yield strength (MPa)	Largest tensile strength (MPa)	Elongation rate (%)	Reference
EM	Original annealed sample	49.6	212.57	52.62	This study
	0.8 mm	334.61	432.84	16.25	This study
	1.0 mm	278.72	423.94	17.99	This study
	1.2 mm	213.05	416.04	19.18	This study
	1.4 mm	146.48	408.31	23.57	This study
	1.6 mm	81.53	399.34	25.13	This study
CR	1.5 mm	354.80	358.6	1.9	[29]
RTR	1.5 mm	342.1	347.0	2.1	[29]
ECAP	0 pass		144	51.6	[30]
	1 pass		319.9	12.4	[30]
	4 passes		373	9.4	[30]
	8 passes		400	8.2	[30]
	12 passes		408	11.6	[30]
ARB	0 pass		220.67	43	[31]
	1 pass		363	5.5	[31]
	5 passes		445	5.7	[31]
	9 passes		480	5.2	[31]

When the largest tensile strength was equal, the elongation of the GS sheets prepared by the EM process under different t _ch was significantly higher than that of the ultrafine-grained pure copper produced through the conventional SPD methods, as shown in Table 2. The improved balance of strength and plasticity in pure copper sheet was attributed to the synergistic effect between the fine-grained layers with higher strength and the coarse-grained layers with lower strength. Under uniaxial tension, the coarser-grained layer, the weaker coarse-crystalline layer yielded first and then progressed to the stronger fine-grained layer. This process effectively inhibited the occurrence of yielding, necking, and fracture.

The thermal stability of the gradient structural pure copper sheet

Figure 24 illustrated the annealing process for GS sheets. Firstly, the temperature was raised from room temperature to 140°C, 190°C, 240°C, 290°C, and 340°C with a heating rate of 10°C/min, and then kept at this temperature for 5 min. Moreover, the temperature was further elevated to 150°C, 200°C, 250°C, 300°C, and 350°C, this time advancing at a slower rate of 2°C/min, and each stage was sustained for 60 min. Finally, the process concluded with the sheet being removed from the furnace and permitted to cool back to ambient temperature.

Figure 24.

Annealing process route of the gradient structural pure copper sheet.

The microstructure of the GS sheets at various annealing temperatures was depicted in Figure 25. Comparison of the Figure 25(a) and (b) with Figures 17 and 18 revealed that there was no significant change in the grains on both sides of sheets at 150°C, which was similar to the grains in the sheets before annealing, proving a better thermal stability at that temperature. Comparing Figure 25(c) and (d) with Figures 16 to 18, it could be found that the grains on both sides of the sheets changed significantly at 350°C. The elongated grain which occupied a larger proportion before annealing was gradually transformed into the equiaxed grain with a similar shape and size to the original annealed sample, indicating reduced thermal stability at this higher temperature.

Figure 25.

Microstructure of the gradient structural pure copper sheet under different annealing temperatures: (a) the material of side A annealed at 150°C, (b) the material of side B annealed at 150°C, (c) the material of side A annealed at 350°C, and (d) the material of side B annealed at 350°C.

The different microstructures of the GS pure copper sheets under different annealing temperatures were closely related to the microstructure evolution. When polycrystalline metals underwent plastic deformation below the melting point, their microstructure and properties changed, and part of the increased energy in this process would be stored as strain energy. The changed microstructure and properties were returned to the state before plastic deformation by heat treatment (or annealing treatment). This process mainly included three stages: recovery, recrystallization, and grain growth. In the first stage, part of the strain energy stored by plastic deformation was released, which promoted the atoms diffusion at high temperatures, and reduced the number and density of dislocations in the polycrystalline metal. After the recovery stage, the grain was still in a state of high-strain energy. In the second stage, the grain state with high strain energy was transformed to the equiaxed grain state with no strain energy and low dislocation density, which was driven by the internal energy difference between the grains with distortion and those without distortion. The grain boundary moved away from the grain nucleus and the grains gradually formed the new equiaxed grain until it completely replaced the original grain. After the recrystallization stage, the polycrystalline metal became softer and weaker, but its plasticity and toughness were improved. Whether the recrystallization stage could occur usually depended on the recrystallization temperature of a specific material. It was generally observed that the greater the plastic deformation underwent by a material, or the higher its purity, the lower its recrystallization temperature would be. However, when the degree of plastic deformation reached a certain value, the recrystallization temperature tended to a constant, which was called the minimum recrystallization temperature. The process of recrystallization was initiated when the temperature exceeded the minimum recrystallization temperature, and it did not occur when the temperature was below this minimum. After the recrystallization stage, if the polycrystalline metal continued to be in a state of high temperature, the equiaxed grain without strain energy and with low dislocation density would continue to grow. This process was the grain growth stage. As the grain grew up, the grain boundary area decreased gradually and the grain boundary surface energy decreased gradually, which was the driving force of the grain growth. In the recrystallization stage, not all grains grew up and the larger grains absorbed the smaller ones. For polycrystalline metals, the average grain size generally followed the following formula:

$$d^n-d_0^n=\mathrm{Kt}$$

(10)

where d and d₀ were the average grain size at time t = n and t = 0 respectively, t was time, K and n were constants, and the value of n was usually greater than or equal to 2. According to equation (10), as time went by, the average grain size of the material gradually increased and reached a range after a certain moment. After the occurrence of the grain growth stage, the polycrystalline metal became softer and weaker, and the plasticity was further improved.

Figure 26 illustrated the Vickers hardness across the GS sheets from side A to side B at various annealing temperatures. At 150°C, the hardness exhibited a slight decrease yet largely retained its pre-annealing gradient, demonstrating good thermal stability at this temperature. At the temperature of 200°C, there was a noticeable decrease in hardness, although the hardness gradient was still preserved. Upon reaching 250°C, the hardness exhibited a significant decline, yet a gradient in hardness remained discernible. As temperatures ascended to 300°C or higher, the hardness aligned with that of the original annealed sample at approximately 55.34 HV, and the gradient dissipated, indicating a return to the microstructural and mechanical properties of the original annealed state. Consequently, it is recommended that annealing be conducted at or below 150°C to maintain the thermal stability.

Figure 26.

Vickers hardness from the surface of side A to the surface of side B under different annealing temperatures.

Figure 27 illustrated the non-standard tensile samples obtained after testing the tensile properties under different annealing temperatures. The figure revealed that all samples exhibited significant necking following the tensile tests. Besides, different from the gradient structural sample before annealing, the elongation of the annealed sample after the isochronous annealing treatment under different temperatures had a great difference, in which the elongation of the 150°C annealed sample was the smallest, while the elongation of the 350°C annealed sample was the largest. The elongation of the samples progressively increased with the rise in annealing temperature. This pattern underscored the significant impact of annealing temperature on the mechanical properties of the GS pure copper sheet.

Figure 27.

Non-standard tensile sample under different annealing temperatures: (a) 150°C, (b) 200°C, (c) 250°C, (d) 300°C, and (e) 350°C.

Figure 28 showed the engineering stress-engineering strain curve for a 1.4 mm thick pure copper sheet at different annealing temperatures, with the tensile properties detailed in Table 3. According to Table 3, the largest tensile strength of the GS sheet with a thickness of 1.4 mm extruded before annealing was 408.31 MPa, and the elongation was 23.57%. When annealed at 150°C, the largest tensile strength of the annealed sample was 366.46 MPa, a decrease of 10.25% compared to the sample before annealing, while the elongation rate was 26.49%, an increase of 12.39% compared to the sample before annealing. When annealed at 200°C, the largest tensile strength of the annealed sample was 300.03 MPa, a decrease of 26.52% compared to the sample before annealing, while the elongation rate was 39.31%, an increase of 66.78% compared to the sample before annealing. When annealed at 250°C, the largest tensile strength of the annealed sample was 250.05 MPa, a decrease of 38.76% compared to the sample before annealing, while the elongation rate was 46.03%, an increase of 95.29% compared to the sample before annealing. When annealed at 300°C, the largest tensile strength of the annealed sample was 229.93 MPa, a decrease of 43.69% compared to the sample before annealing, while the elongation rate was 49.01%, an increase of 107.93% compared to the sample before annealing. When annealed at 350°C, the largest tensile strength of the annealed sample was 218.56 MPa, a decrease of 46.47% compared to the sample before annealing, while the elongation was 51.25%, an increase of 117.44% compared to the sample before annealing. At the same time, the largest tensile strength and elongation were close to the original annealed sample’s largest tensile strength of 212.57 MPa and elongation of 52.62%. The pattern indicated a gradual decline in tensile strength accompanied by a corresponding increase in ductility, as evidenced by the elevated annealing temperatures. Initially, at 150°C, the copper sheet exhibited optimal thermal stability. However, as the annealing temperature continued to rise, the thermal stability of the sheet progressively deteriorated, aligning with previous observations.

Figure 28.

Engineering stress-engineering strain curve of the sheet under different annealing temperatures.

Table 3.

Tensile strength and elongation of the sheet under different annealing temperatures.

Annealing temperature (°C)	Yield strength (MPa)	Largest tensile strength (MPa)	Elongation rate (%)
Original annealed sample	49.6	212.57	52.62
Untreated sheets	146.5	408.31	23.57
150°C	134.8	366.46	26.49
200°C	80.2	300.03	39.31
250°C	63.9	250.05	46.03
300°C	56.3	229.93	49.01
350°C	50.1	218.56	51.25

Conclusions

• (1) The gradient-structured pure copper sheet, produced with varying t _ch , transitioned gradually from material on side A to material on side B, displaying a consistent gradient. With increasing t _ch , the material flow boundary shifted to the left, the size and proportion of the high equivalent effect region decreased, and the lateral extrusion cutting ratio increased.

• (2) The tensile strength of GS sheets decreased with increasing t _ch , while the elongation increased. Compared with the ultrafine-grained pure copper prepared by the traditional SPD process, the GS pure copper prepared by EM process possessed higher strength-plasticity matching.

• (3) Following annealing at 150°C, the sheet displayed good thermal stability at this temperature. However, when the annealing temperature exceeded 250°C, the gradient structure of the sheet gradually disappeared, and the sheet was least thermally stable when the temperature reached 350°C.

• (4) As the annealing temperature raised, there was a gradual decline in the hardness, with a noticeable reduction occurring when the temperature reached 200°C or higher. Concurrently, the tensile strength of the material diminished, while the elongation progressively increased. Upon annealing at 350°C, the tensile strength and elongation of the sheet approximated the values of the initially annealed sample.