Properties of a Metal-Nonmetal Hybrid Joint With an Improved Shape of the Metal Insert



A vehicle's brake pedal transfers the force from a human being to the braking mechanism, and as such is considered as a safety component in the vehicle. In the past, vehicle weight-reduction initiatives resulted in a highly optimised design of steel brake pedal with an increased strength-to-weight ratio. However, any further reduction in the weight of the brake pedal is only possible by using combined, that is hybrid, materials. Hybrid technologies enable the design of lighter and cheaper brake pedals within the strict limits of the safety and technical requirements. This article presents an example of an innovative mechanical joint with a metal insert moulded into a polymer base structure that could be applied for a brake-pedal design. The metal insert contributes the required strength, while the polymer part provides the final form. However, the design of the metal insert not only provides the strength of the component, but also ensures the proper bonding of the metal-to-polymer joint. This article presents a comparison of the strength for simple and optimised forms of the metal insert within the hybrid joint. The strength comparisons were performed using numerical simulations, which were also experimentally validated.


Recently, the automotive industry has faced increasingly stringent environmental requirements that define the sophisticated technical solutions that manufacturers of vehicles are transferring to their suppliers that have an adequate development capacity.[1-12] Drivetrain components and fuel consumption are the focus of these initiatives, which as a result include reduced vehicle weights. However, this reduced weight affects the load-carrying capacity of the structure, and so sufficient strength, durability, and reliability must be ensured in the early phases of the product's development. The results of such vehicle weight-reduction initiatives are highly optimised designs of vital components with better strength-to-weight ratios. An example of such a component is the vehicle's brake pedal, which must be strong enough to perform its function, even in the case of impact, while keeping its weight as low as possible. Further weight reduction is possible only by using combined, that is hybrid, materials. An example of such a brake pedal is shown in Fig. 1.

Figure 1.

A hybrid brake pedal.

Recent hybrid technologies are implemented as alternatives or derivatives of injection over-moulded technology, a process that has been patented.[13] The first successful implementation of this innovation was in 1996, when Audi manufactured a front bumper from a metal-sheet–polymer hybrid.[14] However, hybrid technology was first used for a brake pedal in the 2006 Fiat Ducato. The core element is a thin-walled, U-profile, reinforced with PA 6-GF 30 polymer ribs. The polymer and metal components are joined mechanically, with polymer ribs and plugs, moulded into openings in the steel U-profile.[15] Some other solutions include ZF Friedrichshafen[16] and Trelleborg,[17] which are still in the development and validation phases.

After the bonding of two different materials, the resulting hybrid structure assumes the properties of the weakest component. If we wish to benefit from the hybrid structure (i.e., the high strength of the insert and the low weight of the base matrix), we have to ensure proper bonding of the joint. In this case, the hybrid structure demonstrates better properties. There are three basic bonding principles: material joint, geometric joint, and frictional joint. In the case of the material joint, the load is distributed between the bonded components by means of an additive material, which may be different (e.g., an adhesive) or the same as the base material (in the case of welding). In the case of the geometric joint, the load is distributed between the bonded components by means of a properly shaped joint; and in the case of a frictional joint, the friction is used to distribute the load between the bonded parts.

The most widely researched joints using metals and nonmetals were different versions of material joints. In almost all cases, the subjects of such research were adhesively bonded joints[18-28]; however, these are too expensive for mass production, since a sophisticated surface treatment or cleaning is required. Other limitations are the production technologies, the price of the adhesives, and the weight of such joints. In addition to the strength of the hybrid joint, cost effectiveness is also critical for serial production. The optimal solution for a hybrid joint would keep both parts tightly coupled, while leaving them intact. From this standpoint, a geometric or frictional joint would be a perfect solution. The problem is that the load capacity of such joints has not yet been researched in detail.

To improve the mechanical properties of the hybrid joint the metal insert must be shaped in such a way that the load is distributed between the insert and the polymer matrix by means of the geometric joint. This article will present some variations of the geometric joints between the metal insert and the polymer matrix that can be applied in hybrid structures. Tensile tests and numerical simulations were used to compare the load-carrying capacity of the different versions of such joints.

After the introduction, different versions of the geometric joints are introduced and their weaknesses and strengths are explained. The article continues with the experimental validation. This section describes the preparation of the specimens and the testing of the mechanical properties with a tensile test, where the results of the strength analysis for all versions of the hybrid joints are presented. Numerical Simulation of the Hybrid Joint and its Strength section shows numerical models used for the validation of the reference version of the joint's shape (Fig. 2) and a joint with geometric transmission of the load. In Comparisons With the Simulation Results and Discussion section we compare the measurements with numerical results and discuss the results. The main conclusions of the article are stated in the Conclusion section.

Figure 2.

Basic version of the hybrid joint “Type I.”

Selection of the Design for the Hybrid Joint

The understanding of a hybrid joint's load-carrying capacity is a prerequisite for the development of hybrid products with increased strength requirements. Due to the limitations mentioned above, the objective was to select the proper form of the geometric or frictional joint between the metal insert and the polymer matrix, which would be simple enough for production while maintaining the required mechanical properties. The initial form of the joint is shown in Fig. 2. The basic dimensions were determined in accordance with the recommendations of the D638-08[26] standard.

The hybrid joint, as shown in Fig. 2, is made by the moulding of a thin, flat, S420MC steel-sheet insert into the matrix, made of Durethan DP BKV 60 H2.0 EF polyamide (polyamide 6 with 60% glass fibres, injection moulded, with improved flowability and heat-ageing stabilised). In order to facilitate the experimental determination of the joint strength, the insert is only partially moulded, which makes fixing the specimen into the tensile testing machine easier. The form of the specimen was determined in such a way that a rupture always occurs at the junction between the metal and the polymer.

In such a case, the strength of the joint is determined by the adhesion between the metal and the polymer (material joint) and the frictional force (frictional joint). The polymer matrix shrinks during the cooling phase and, therefore, it compresses the metal insert from all directions. The surface pressure is a prerequisite for transmitting a part of the load through the joint by friction. The use of the PA 6 polymer with 60% glass fibres limits the time-dependent relaxation of the polymer, so that the major part of the contact pressure on the metal insert is expected to be maintained.

To determine whether the shape of the joint adds bonding strength to the frictional joint a tensile load will be applied to the specimen, because a similar load type is applied to the brake pedal during regular operation. Different variations of the geometrical joint (Fig. 3) were designed by improving the existing solutions[15, 27-29] in order to establish the proper joint between the insert and the polymer matrix.

Figure 3.

Partially moulded inserts of different shapes.

The thickness of the steel insert (EN-ISO 1.0980) was determined on the basis of the steel-sheet thickness from existing brake pedals (Fig. 3–2d). The objective when designing the specimen (Fig. 3–3d}) was to achieve an uninterrupted flow of the melted polymer with the added glass fibres through the openings in the steel insert. The curved shape should also contribute to an improved bending strength of the specimen and the wedging of tabs in the load direction. Its weakness lies in the significant notch effect and the reduced cross section in the area of the curved tabs. Two additional specimens with a flat insert were prepared, made of carbon fibre (Fig. 3–4d}) and HSS steel sheet (Fig. 3–5d}). The strong joint between the polymer and the metal insert should be achieved with circular holes in the insert (Fig. 3–6d}), though the cross section is reduced. Through different investigations,[30, 31] it has been shown, that the shape of the holes does not influence the strength significantly, but increasing the size and the number of holes does. These results were considered in our design of specimen 6d. The specimen with a steel embossed insert (Fig. 3–7d}) enables a frictional and a geometric joint without any reduction in the cross section and a reduced notch effect. A short list of the tested inserts is below:

  • 2d: steel flat insert,
  • 3d: steel curved perforated insert,
  • 4d: carbon-fibre flat insert,
  • 5d: high-strength steel flat insert,
  • 6d: steel perforated insert,
  • 7d: steel embossed insert.

A detailed explanation of the specimen preparation and experimental testing is presented in the following section.

Experimental Determination of a Joint Strength

Preparation of the specimens

The specimens were made by first forming the inserts, then inserting them into the proper moulding tool, and finally, moulding them. All specimens were made by using standard technologies. First, the inserts were laser cut to proper dimensions and then for the cold forming of the metal inserts a special tool was used to form the openings (Fig. 3-3d) and the imprints (Fig. 3-7d). Although the forming was handmade, easy automation of the process is possible. A prototype mould tool was made for the production of the first series of the specimens. The pouring spot (gate) is the central longitudinal axis of the specimen. The metal inserts were positioned in the tool by means of a special holder, see Fig. 4, and heated to 70°C (temperature of the tool). A 60-t Negri Bossi injection-moulding machine was used with a heater temperature of 280°C, a pressure of 90 bars, a 20-mm dosing, and an injection speed of 30%.

Figure 4.

Specimen moulding tool and formed specimen.

Experimental results

MTS 810.22 testing machine with the following specifications was used for the tensile testing (Fig. 5).

Figure 5.

Experimental equipment.

Testing site specifications:

  • — maximum force: −100 to +100 kN,
  • — maximum frequency: 20 Hz,
  • — air conditioned chamber: −130°C to +315-C,
  • — chamber dimensions: 356 mm, 432 mm, 559 mm,
  • — adjustable jaws: up to 100 kN,
  • — software: MTS 793.00-FlexTest Automation (PC) system software.

The specimens were fixed into the jaws with a clamping pressure of 7 MPa. The jaw-displacement velocity was 0.5 mm/min at room temperature and 50% relative humidity. To achieve statistical relevance of results, three specimens were tested for each variation. The test results were shown to be repeatable, which proved that the manufacturing process using the existing technology is controlled. A scatter range between the individual tests is 3% on the average.

Figure 6 shows the average values of the tensile test results for all hybrid specimens. The maximum fracture force for the basic specimen (2d) is 1360 N, while for the embossed specimen (7d) it is 4600 N. The measurements show that the specimen with the embossed joint carries more than three times more fracture force than the smooth flat specimen. Specimens 7d and 3d showed a pull-out failure mechanism which resulted from a fractured polymer matrix (Fig. 7). Specimen 6d failed at the metal insert resulting from a reduced cross section. Specimens with a flat insert 2d, 4d (carbon-fibre flat insert), and 5d (high-strength steel flat insert) experienced a pull-out failure mechanism after achieving the maximum fracture force, which can be seen by the “wave” effect for specimens 2d and 5d and a rapid decline for 4d. Results are shown for 2-mm displacement.

Figure 6.

Experimental results.

Figure 7.

Fracture mechanism of the geometrical joint 7d.

After testing of the specimens on the testing machine, the metal insert shape 7d was selected as the most appropriate from the standpoint of mechanical properties and manufacturing feasibility. This version will be used to validate the results of the numerical simulations together with the reference shape of the joint (2d). For the reference shape of the joint (2d), the bonding strength between the metal and nonmetal components only depends on the friction and the direct adhesion of the polymer to the metal surface. For the geometrical joint (7d), the bonding strength also depends on the shape of the polymer–metal joint.

In the next section, we present a numerical approach to simulating the selected hybrid joints and show the added strength of the geometrical joint in comparison to the experimental results.

Numerical Simulation of the Hybrid Joint and Its Strength

The 3D models of the specimens and the adjustable jaws from the testing machine were designed with Catia V5.[32] The finite-element simulations of the tensile test for the two selected hybrid joints were performed with Abaqus/CAE.[33]

Basic preliminary simulations were necessary to construct the foundation for our numerical model. Even though the results are not shown, it is important to give notice about the characteristics found through these tests. The material properties used have a proportional influence on the fracture force which means that if we halve the polymer material properties (σ, E), the fracture force divides by two as well. The element type, linear versus parabolic, also affects the results with a 25% force decrease with the parabolic elements. By not considering the friction coefficient between the materials, the shrinkage properties do not have a significant contribution to the fracture force, but on the other hand, using friction without the shrinkage also does not impact the fracture force. This shows the importance of using friction to model the behaviour of hybrid joints.

The elastoplastic material models with the true stress–strain curve were used for the steel insert and the polymer matrix (Fig. 8). The steel S420MC thermomechanical properties used: Young's modulus, E = 183 GPa (as measured), Poisson's ration, ν = 0.3, linear thermal expansion coefficient, α = 16 × 10−6. The following polymer thermomechanical properties were used (taken from CAMPUS plastics data bank): Young's modulus, E = 20 GPa (as measured), Poisson's ratio, ν = 0.35, linear thermal expansion coefficient, α = 4 × 10−5.

Figure 8.

Material properties of S420MC (above) and Durethan DP BKV 60 H2.0 EF (below). Tested on the MTS 810.22 tensile testing machine. Speed 0.5 mm/min at room temperature and 50% relative humidity.

As the goal was to achieve a comparison between the different joints and not an exact force match with the experiments, the chosen characteristic length of the finite elements was 3 mm. Smaller element size did not influence the maximum pull-out force significantly and with the larger elements, the computing time was reduced more than 10 times. In the case of the reference joint with the smooth insert (2d), C3D20R 20-node quadratic bricks were used (4309 elements), while in the case of the geometrical joint with the steel embossed insert (7d), the 10-node C3D10 quadratic tetrahedron type with 33989 elements were used due to the complicated geometry of the steel insert. The friction and adhesion in the contact between the steel insert and the polymer matrix were considered with the addition of friction and the shrinkage properties of the material. This method was chosen over the standard traction-separation method because of the complexity of the embossed specimen.

The boundary condition for the simulation was a 2-mm longitudinal displacement of the movable jaw (the insert side), while the jaw at the polymer side was fixed. Both parts of the specimen, that is the polymer matrix and the steel insert on the right, are rigidly connected with jaws.

The simulations were performed in three steps. In the first step, the jaws were clamped. In the next phase, the shrinkage of the polymer matrix was determined by cooling it to the specified temperature difference. In the last step, the movable jaw was shifted in the y direction. The simulation failure criterion is met when the tensile force drops below the maximum value.

For the described finite-element models, a parameter study was carried out (Table 1). In this study, the influence of the friction coefficient in the contact was studied. The friction coefficient value was selected on the basis of known values from the literature,[25] while the temperature variation in the cooling phase ΔT was determined on the basis of the polymer glass transition temperature. The results show a large difference between the values achieved with and without the friction coefficient. We also have to note, that the difference for the geometrical joint is double the amount of the difference for the reference joint. We need to use friction to simulate the reference joint, but comparing to experimental results (1360 N for 2d and 4600 N for 7d) we find, that simulations without considering friction for the geometrical joint are more applicable.

Table 1. Numerical analysis results
 Frictional coefficientΔT(°C)Force (N)
Reference joint (2d)0.0−100  0
 0.1−100 716
Geometrical joint (7d)0.0−1004600

The comparison of the simulated and experimentally determined deformation behaviour is shown in Fig. 9.

Figure 9 shows the relation between the tensile force in the specimen and the relative displacement of both jaws, for the reference specimen shape (2d) and the geometric joint (7d), between simulated and experimental results. Considering friction and shrinkage, we get a maximum simulated tensile force of 716 N for 2d and 6400 N for 7d. The maximum tensile force achieved with the experiment is 1360 N for 2d and 4600 N for 7d joint.

Figure 9.

Comparison of the fracture force derived from experimental tensile tests and numerical simulation (using 0.1 friction coefficient).

The Von Mises stress values for the embossed specimen (7d) at the maximum achieved force during the test are shown in Fig. 10, for the purpose of studying the deformation behaviour. The failure mechanism is similar to the failure mechanism in Fig. 7 from the experiment.

Figure 10.

Von Mises comparative tension during the fracture of a geometric joint.

Comparisons With the Simulation Results and Discussion

The experimental results showed that the fracture force is significantly increased by changing the basic shape of the metal insert joint into the geometric shape of the joint, which was then confirmed by numerical simulations. Despite some differences between the experimental and numerical results, the numerical analysis data substantiated the findings of the experiment. The reason for the discrepancies between the experimental and the simulated results originate from the unknown material behaviour properties, including the friction coefficient and the material shrinkage as well as unknown production process mechanisms that influence the joint, and the element size and type used in the simulations. We can explain the lower measured geometrical joint strength in comparison to the simulated one (with friction) as a consequence of bad over-moulding of the embossed insert. But the lower strength for the reference joint can be designated to unknown real friction and adhesion parameters. An experimental test is suggested, in which specimens should be designed so that the friction (and adhesive) portion of the joint would be reduced by adding a layer of PTFE on the contact surface between the steel insert and the polymer matrix. In this way we would be able to estimate the sole contribution of the shape of the insert.


This article describes the selection of the shape and a comparison of the hybrid joint strengths between simple and optimised forms of the metal insert.

By using the embossed inserts, the bonding strength between the two materials is significantly higher, without any additional treatment of the materials that are being used.

The experimental and numerical analyses of the hybrid specimen were performed in order to estimate the improvement in the bonding strength of the joint between the metal and the polymer in the case of the hybrid brake pedal.

Although the test results are promising in terms of the integration into the structural components and indicate that the technology is appropriate, a further improvement of the numerical simulations is required, which would better reflect the real development within a hybrid joint. These findings will be used in an application in which tests for reliability, manufacturing capability, and recycling feasibility will be executed.


The results presented here are a part of the doctoral research: “Yield of bended beams, made of metal-nonmetal hybrids”, which was co-funded by the Slovenian Technology Agency - TIA, P-MR-09/47. Operation part financed by the European Union, European Social Fund. The authors of this article would like to thank all who participated in the research and testing of the hybrid specimens.