Process chain for the manufacture of hybrid bearing bushings

As already mentioned, the manufacture of hybrid solid semi-finished products by welding is already established, as it is the case with Tailored Blanks. Another option for joining different materials is pressure welding. This process is promising, because the resulting intermetallic phase seam is adequately narrow, and thus negative effects on properties (e.g. tensile strength) can be avoided [6]. Co-extrusion also belongs to the pressure welding processes, which allows the continuous production of hybrid semi-finished products made of at least two different materials [7]. According to the German standard DIN 8593-5, co-extrusion refers to joining by forming [8]. In principle, co-extrusion can be subdivided into two different main types of processes, in which the joining partners are formed either completely or locally. Each of these approaches utilize a different type of extrusion billet:

The LACE (Lateral Angular Co-Extrusion) process developed by Grittner et al. can be categorized as a co-extrusion process using conventional billets. Grittner et al. used this process to insert relatively stiff sheets and flat titanium profiles into the forming zone of the tool at an angle of 80°–90° to the extrusion direction. This procedure can be described as a quasi-continuous feeding of the reinforcement into the extrusion process [14]. In earlier studies by the authors, a LACE process in the laboratory scale for the jacketing of round steel bars with an aluminum alloy was developed. The resulting compound profiles consisting of 20MnCr5 case-hardened steel (1.7147—AISI 5120) and the aluminum alloy EN AW-6082 differed considerably from the intended coaxial arrangement of the two material components, due to specific limitations of the primary tool design [15]. Recent investigations by Thürer et al. employing a new tool design at near-industrial scale as well as 20MnCr5 steel tubes as reinforcing elements have already shown that sufficiently coaxial aluminum-steel hollow profiles featuring reinforcement contents of both 14 vol% and 34 vol% can be successfully achieved. In addition, shear strengths of up to 58 MPa ± 15 MPa were determined on sample segments that included the joining zone, which indicated the presence of a metallurgical bond by adhesion of aluminum to the steel [5]. For the planned subsequent die forging step within the Tailored Forming process chain (compare schematic process chain in Fig. 1), the desired coaxial arrangement of the two materials is indispensable. Furthermore, it is of major interest whether the developed LACE process can be transferred to reinforcements made of other materials and if a subsequent forming of the hybrid semi-finished product can be carried out successfully.

Depending on the initial state of the workpiece, multi-material forging processes are divided into compound forging and hybrid forging. The first method represents a combination of joining and forging operations in a single process step. Basic application studies of this concept showed that production of a uniform joint is challenging. The joining quality improves with increasing temperature [16]. However, increased formation of the oxide layer at elevated temperatures has a negative effect on the joining zone, and consequently on the joint quality [17]. For complex geometries with locally different strains and non-uniform material distribution, realizing a suitable joint becomes difficult. At this point, the second method (hybrid forging), which involves the utilization of previously joined workpieces, can be advantageous. The investigation of Förster et al. on forging of aluminum enclosed magnesium profiles showed that the joint produced by co-extrusion was maintained after the forming [18]. As demonstrated for bi-metal gas turbine discs by Klotz et al. [19], the quality of the joint after forming is largely influenced by workpiece’s initial condition. Domblesky et al. conducted forging tests on serially arranged friction welded workpieces that demonstrated good workability. The material flow depends on the load direction (axial and side compression) and material combination. In dissimilar compounds, deformation occurs preferentially in the lower strength material [20].

In the present study, the AlSi1MgMn aluminum wrought alloy (EN AW-6082) was employed as the matrix material in the co-extrusion experiments. The chemical composition of the aluminum alloy was measured using optical emission spectroscopy (Table 1). This aluminum alloy is one of the more easily formable wrought alloys [7], and thus was chosen for the LACE experiments. The billets were obtained from ST Extruded Products Germany GmbH (Vogt, Germany) and had an outside diameter of 142 mm. They were produced by continuous casting and then homogenized. The composition complies with the specifications in the standard DIN EN 573-3 [21].

Two different steel grades were used for the reinforcing element: The unalloyed case hardening steel C15 (1.0401—AISI 1015) and the case hardening steel 20MnCr5 (1.7147—AISI 5120), which is alloyed with manganese and chromium. For both LACE experiments, the reinforcements were used in the form of tubes with a length of 660 mm, the outer diameter being 44.5 mm and the inner diameter 32 mm. Optical emission spectroscopy was applied to identify the actual composition of the steels used (Table 2). Their compositions comply with the respective standard specifications according to DIN EN 10084:2008-06 [22].

Figure 2a shows a schematic section through the relevant parts of the 10 MN extrusion press with the mounted modular LACE tool, which was developed specifically for the Tailored Forming process chain. Since the LACE process is used to encase a rigid reinforcement with a light metal, a mandrel component was designed, which is held in the tool by three support arms, arranged at an angle of 120° to each other. The procedure for this specific LACE process has been described recently by Thürer et al. [5] and Fig. 2b shows four steps of the simulated material flow of the aluminum. In a first step, the aluminum alloy meets the symmetrically designed inlet, where the material is split into two main metal streams. Then, the two metal streams each flow into pockets milled into the tool cavity (see second step). This prolonged flow into the pockets controls the overall flow of the aluminum alloy in such a way that the steel tube fed from the side is equally enveloped from all sides (see third step). This setup has been established in order to prevent the resulting compound profile from shifting or, even worse, from deforming. The aluminum is deflected and within the tool, the two aluminum streams then flow around the mandrel part, which results in a split into three parallel metal streams. These are then joined again in the welding chamber, while making physical contact with the steel tube (transition from the third to the fourth step). The tube is guided orthogonally to the direction of movement of the ram through the clamping cover and the mandrel part in the tool. In the mentioned fourth step, the compound profile exit the tool through the die. As the previous results by the authors showed [5], this specific tool and process design is capable of ensuring a mostly coaxial arrangement of the reinforcement inside the compound profile.

The LACE experiments required a modified tool holder for the extrusion press in order to enable an introduction of the reinforcing steel tube into the extrusion process. In this study, a non-heated tool holder was used, which had cavities that allowed for the lateral feeding of reinforcements. The co-extrusion was performed on a 10 MN extrusion press (SMS Meer GmbH, Düsseldorf, Germany). The extrusion ratio of the EN AW-6082 alloy that is used here depends on both the inner diameter of the container (d_container = 146 mm) and the opening diameter of the die (d_die = 62.7 mm). Additionally, the outer diameter of the steel tube (d_tube = 44.5 mm) must be included in the calculation (Eq. 1).

For these particular LACE experiments, the extrusion ratio for the aluminum equaled to 11:1. In addition, the ratio of the volumes of aluminum and steel in the hybrid profile corresponded to a reinforcement content of 34 vol%. To shorten the process chain, tubes were used instead of round bars for the LACE process. These already had the internal diameter of 32 mm required for the subsequent die forging process.

Prior to the LACE experiment, the steel tube was ground with 40 grit emery paper and then cleaned with ethanol. In ensured removal of any excess oxides that would curtail the metallurgical bond formation between aluminum and steel upon co-extrusion. Previous numerical investigations have shown that exerting relatively high temperatures as well as long contact times can obtain a metallurgical joint between the joining partners [23]. This transfers to high LACE process temperatures and low extrusion speeds. Hence, the billets were preheated to 530 °C for 4.5 h. The steel tube was at room temperature at the beginning of the test. The die was preheated to 490 °C and the container to 440 °C. At the beginning of the LACE process, a ram speed of 1.5 mm/s was chosen for upsetting of the billet inside the container. A stepwise decrease of the ram speed was then employed until the desired ram speed of 0.3 mm/s for the actual LACE process was achieved. This procedure was intended to counteract and reduce the rapid cooling of the matrix material billet during filling of the die to a minimum.

Forming the largest diameter of the bearing bushing shown in Fig. 1 (right) is the main challenge in forging the hybrid semi-finished products manufactured by co-extrusion. In order to achieve sufficient formability, the individual materials need to be heated to their material-specific forging temperatures simultaneously. Therefore, an inhomogeneous temperature distribution within the bi-metal workpieces is necessary. This can be achieved by tailored application of induction heating, where the heat is generated by eddy currents. Due to the skin-effect, which dominates at middle and high operating frequencies, the highest current density is primarily concentrated on the workpiece surface, and thus is subject to the highest thermal input [24]. In the context of steel-aluminum workpieces, this phenomenon can be used to generate heat almost only in the steel component, while the aluminum heating is performed by heat dissipating from the steel. To avoid forming defects (e.g. cracks), the temperature of the steel part should be in the warm or hot forming temperature range (above 550 °C). However, the maximum possible temperature for steel is limited by the onset of melting of the aluminum alloy EN AW-6082 (solidus temperature of approx. 580 °C [25]).

Taking the present coaxial workpiece geometry into account that has steel on the inside, an induction heating concept with an internal induction coil was designed (Fig. 3). The heating tests were conducted using a middle-frequency generator TruHeat MF 3040 (TRUMPF GmbH & Co. KG, Ditzingen, Germany). In order to achieve a high temperature gradient, the maximum power output of 40 kW was used. The forging experiments were carried out using three heating strategies that differed in heating times. The temperatures were measured at the reference points located inside the center of the steel and aluminum components with thermocouples of type K, as shown in Fig. 3. The resulting temperatures before forging, including a transfer time of 2.5 s as well as the main generator parameters, are summarized in Table 3.

A 40 kJ screw press of the type Lasco SPR 500 (Lasco Umformtechnik GmbH, Coburg, Germany) was used to carry out single stage forging tests. The corresponding closed-die tool system that was installed in the forging cell is depicted in Fig. 4. The operating principle is described in detail in a previous publication [26]. Before the forming process, the lubricant Berulit 913 (Carl Bechem GmbH, Hagen, Germany) diluted with water in a ratio of 1:3 was applied to the semi-finished products. For this purpose, they were heated up to a temperature of about 150 °C in a furnace before being coated with the mixture. After the evaporation of the water, a thin layer of graphite and MoS₂ remains on the surface. Subsequently, the coated workpieces were heated up to forging temperature by the described induction heating strategy and then automatically transferred to the forging press with a robot arm. Thus, reproducible process conditions for all forging tests could be achieved.

An optical investigation by means of a metallographic characterization was performed on cross-sections from the front end (FE) as well as the back end (BE) of the compound profiles. Thus, deviations in the position of the reinforcement relative to ideal center positions were determined. It must be noted that while similar investigations of the material combination EN AW-6082 and steel 20MnCr5 have already been presented by the authors in the previous publication [5], the current results were obtained from the particular profiles that were used for subsequent die forging tests and thus, represent an extruded reference condition prior to forging. In addition, sections from extruded hybrid profiles made of EN AW-6082 and C15 steel have not been shown at all before, which enables a direct comparison of the two different material combinations used in this study.

Furthermore, the microstructures of the extruded matrix material was characterized. The location, where all LWS appeared to be macroscopically closed was defined as actual FE for all compound profiles. The position of the BE was reliant on the particular LACE experiment. In case the LACE experiment was stopped before the whole reinforcement was jacketed with aluminum, the part of the compound profile, which was closest to the tool, was examined in cross-section. In contrast, specimens taken from the forged bearing bushings were examined longitudinally in order to characterize the modification of the joint zone after forming in both the low and high deformation regions.

For all macro- and microstructural examinations of the bi-metal workpieces, the specimens were prepared metallographically. To contrast the secondary precipitates present within the aluminum alloy, the samples were treated with an etching solution consisting of HF and H₂SO₄. Both case-hardening steels were etched with Nital, to compare the microstructures before and after forming.

The relevant LACE process parameters such as ram force as well as ram speed were recorded. Figure 5 shows typical force–time curves of two experiments with an extrusion ratio of 11:1 using EN AW-6082 as matrix material and tubes consisting of C15 (a) or 20MnCr5 (b) as reinforcing elements. A process-related cooling of the tool caused by the use of a non-heated holder during the LACE experiment was determined via the thermocouple positioned inside the die, near the bearing surface. To counteract this cooling, significantly faster ram speeds were chosen at the beginning of the process. Initially, a ram speed of 1.5 mm/s was used in order to upset the billet and fill the tool relatively quickly. As shown in Fig. 5a, b, the ram force rapidly increased until at plateau at around 2.6 MN was reached. After this plateau of the first step, the ram speed was reduced to 0.5 mm/s. As the filling process progressed (second and third step), the ram force continued to increase. The target ram speed of 0.3 mm/s for the main LACE experiment was set after reaching 6 MN (C15) and 7 MN (20MnCr) respectively, reflected by a short drop in force (Fig. 5). Subsequently, the ram force increased continuously to a maximum value of ≈ 8.2 MN (C15) or ≈ 8.0 MN (20MnCr5), at which point (fourth step) the hybrid profile exited the LACE extrusion die. At this point, the ram force did not drop, but increased to ≈ 9.0 MN (C15) and ≈ 8.8 MN (20MnCr5).

Figure 6 shows exemplarily a longitudinal section through the hybrid hollow profile made of EN AW-6082 and 20MnCr5. Semi-finished products for the die forging process were later produced by sawing and turning the compound profiles. Under macroscopic visual inspection, there seems to be a largely coaxial arrangement of the reinforcing element and the aluminum jacket. The positions of the cross-sections that were extracted for closer examination FE and BE are marked with blue dotted lines. During the separation of the compound profiles from the tool, cooling lubricant partially leaked between the joining partners in the case of the profiles with EN AW-6082 and C15. In the case of the profiles made of EN AW-6082 and 20MnCr5 this could not be detected.

The cross-sections taken from FE and BE of the compound profile (having a total length of 215 mm) of the material combination EN AW-6082 and C15 are shown in Fig. 7a, b. In comparison, Fig. 7c, d show the cross-sections of the FE and BE of the compound profile with the material combination of EN AW-6082 and 20MnCr5 (total length of 480 mm). All cross-sections had an almost circular geometry. Deviations from the intended diameter of the aluminum jacket were found for both profile variants. While the deviation was the same over the entire profile length for 20MnCr5, a change in values was determined for the profile with C15 (see Table 4).

These deviations in diameter indicate that the material flow of the aluminum in the tool has not yet been fully adjusted by the milled pockets. As a result, a higher wall thickness of the aluminum jacket can be observed on the side with LWS 3. Thus, the coaxial arrangement of the steel tube is not ideal. In the material combination with C15, the desired position was approximately achieved at the beginning (see Fig. 7a). However, this changed over the profile length (see BE in Fig. 7b), hence an influence of the aluminum flow on the reinforcing element in the x-direction must be assumed. For the profile with 20MnCr5, the offset of the tube from FE (see Fig. 7c) towards the BE (see Fig. 7d) decreased. Further investigations on the influence of the steel grade on the material flow of the matrix material are still pending. The fact that this is not yet ideal is evident from the angle between the LWS. The position of the LWS did not quite correspond to the expected position in all specimens. In particular, LWS no. 3 was tilted away from the recipient in the z-direction. In addition, the angle between LWS no. 4 and no. 2 was typically lower than the expected 120°, with all samples having an angle of about 105° between LWS nos. 2 and 4 instead.

Deviations of the reinforcement from the ideal coaxial position can be explained in both cases (C15 and 20MnCr5) by the initial position of the tube inside the tool. The reinforcing element is guided through the mandrel. In order to prevent the tube from getting jammed in the mandrel, the outer diameter of the tube was chosen slightly smaller than the opening of the mandrel, after which the steel tube enters the welding chamber and comes into contact with the matrix material. In contrast to the work of Grittner et al. [14], where no offset was possible because a titanium sheet was used as external reinforcement, in the LACE process described here an offset of up to 1.0% is possible due to the diameter tolerances between tube and mandrel part. The coaxial arrangement of the two joining partners was achieved by turning, which is indicated in Fig. 7 by the white dashed line in the aluminum. For this purpose, the semi-finished products were clamped inside the bore of the steel and the aluminum was machined to achieve fully coaxial cylindrical surfaces on the outer diameters of both the steel tube and the aluminum jacket. In the case of the compound profiles with C15, a maximum of 1.5 mm of aluminum was removed in diameter. In the case of 20MnCr5, up to 2 mm of aluminum were turned.

In addition to the variation of the shape of the supplied reinforcements (e.g. rods, tubes), the successful variation of the material spectrum is also of interest for a wide range of applications of the LACE process in order to be able to adapt the material to the load case in the future. In this study, C15 and 20MnCr5 reinforcements were successfully jacketed with the aluminum alloy EN AW-6082. An influence of the steel grade on the parameters of the LACE process and the positioning accuracy of the reinforcement in the compound profile was not found. Upon removing the profile with C15 from the tool, cooling lubricant leaked between the joining partners. It can therefore be assumed that no complete metallurgical bond was present here. Such a behavior was not detected in any compound profile with 20MnCr5.

The hybrid bearing bushings forged with different temperature profiles resulting from the parameters given in Table 3 are shown in Fig. 8. Other parameters of the forging process were the same for all investigated cases. With heating strategy A (8 s of heating), a vertical crack along the whole length of the component occurred as a result of an insufficient formability of the steel part, whose actual temperature was well below usual forming temperature in this test. The most critical temperature is not present in the center of the steel part but at the joining zone, where steel and aluminum have an almost equal temperature under 450 °C. Thus, the steel part had insufficient formability and failed during the forming process. Blue discoloration of the fractured surface relates to the blue-brittleness effect caused by dynamic strain aging at low forming temperatures [27]. This behavior is accompanied by maximum ultimate strength and minimum ductility, which can lead to premature material failure even at low strain. Different from strategy A, the critical temperature threshold for steel forming was exceeded with strategy B (9 s of heating) even at the joining zone. Consequently, the bearing bushing could be successfully formed without any macroscopic defects. It should be mentioned that changing the temperature field by applying a different heating strategy could also have an impact on e. g. the local tribological conditions, strain and strain rates even if the stroke kinematics were kept unchanged. Damage initiation in hot forming can be affected by the prevailing strain rate as discussed in [28]. Thus, the success of the process setup is determined not only by the temperature fields of both workpieces, but represents a result of the entirety of the process conditions of heating strategy B. In order to the determine process stability limits increased temperatures of the steel component, strategy C employing a heating time of 10 s was investigated. However, the aluminum component reached a temperature above 500 °C and thus, failed on the outside diameter due to hot cracking.

Figure 9 represents the material distribution of a bearing bushing successfully forged using heating strategy B. There are three main characteristic areas: the largest diameter, the inclination and the smallest diameter. Metallographic examinations were carried out on one part consisting of EN AW-6082/20MnCr5 and on two parts made of EN AW-6082/C15. Representative micrographs recorded at characteristic points along the joining zone are shown in Fig. 10 and marked from 1 to 5.

All micrographs from the inclination area show a sufficient bond. A separation of the joining partners can be detected at both ends of the specimen with EN AW-6082/20MnCr5 (Fig. 10a). The gap size is up to 60 µm at position 1 and about 20 µm at position 5. In the lower part with the smaller diameter of the hybrid bushing, the separation took place particularly within the aluminum alloy, therefore, some aluminum remains adhered to the steel. This indicates that the bond that was established during extrusion did not fail, but crack formation occurred inside the softer aluminum alloy during forging.

Similarly, the bond in the lower part of both specimens consisting of EN AW-6082 and C15 did not failed during the deformation (Fig. 10b, c). In the upper part of specimen 2, the joining zone obtained a wavy morphology after forming (Fig. 10b). The gap size at position 1 is between 30 and 50 µm. There are some particles of steel visible on the aluminum side of the gap. In contrast to specimen 2, specimen 3 shows a complete joint along the whole interface zone except for a couple of air inclusions as can be seen at position 1 (Fig. 10c). All micrographs of the samples having the material combination EN AW-6082/C15 demonstrate an extremely deformed pearlite phase on the steel side close to the joining zone. This effect appears after forming and can be explained by the influence of the inhomogeneous heating strategy and specific material flow in the joining zone.