Sage Journals: Discover world-class research

Abstract

The combination of thin light metal sheets with fibre-reinforced thermoplastic layers in multi-layered fibre-metal-laminates advantageously combines the properties of both material classes. In this way, components can be developed which have both significantly increased specific properties (strength and stiffness with respect to density) and high energy absorption capacity compared with conventional design with mono materials. However, the structural behaviour of crash structures is decisively determined by material behaviour of the thermoplastic and metal constituents as well as the interface properties between both constituents and the corresponding delamination behaviour. To evaluate the structural response of multi-layered fibre-metal-laminates under highly dynamic loading conditions, Charpy tests were performed, where the test parameters, light metal material configuration, support length and laminate thickness, were varied. Moreover, the metal sheet surfaces were pre-treated by embossing to achieve different surface topologies. The influence of the different test parameters on the specific energy absorption capacity was characterised by the analysis of force–displacement curves.

Keywords

Lightweight structure fibre-metal-laminate interface topology Charpy impact test

Introduction

Lightweight structures are increasingly used in all fields of applications where masses are being moved to reduce energy consumption. Especially, the specific energy absorption is essential for impact- and crash-related applications. In this respect, laminated composites offer the possibility to design the property profile by varying sheet materials, thickness, stacking sequence, fibre orientation and process technologies.

Light metal alloys with aluminium, magnesium and titanium are characterised by a high formability, a non-directional high strength and stiffness and a comparably low density.^1,2 The crashworthiness of aluminium, carbon fibre-reinforced plastics (CFRP) as well as hybrid aluminium-CFRP hybrid tubes was experimentally proven under different loading angles.³ Such crash specific tubes are commonly designed for axial loadings.^4–6 In contrast, sandwich structures are used to resist loads normal to the surfaces.^7–9 Multi-layered fibre-metal-laminates (FML) as a combination of light metal semi-finished products with fibre-reinforced plastics in multi-layer design links the advantageous properties of both material classes and can be used to design an improved material behaviour.^10–15

A well-known example of a fibre-metal-laminate is GLARE, where glass fibre-reinforced epoxy layers and aluminium sheets are used.^16–20 In contrast, the present study is focussed on the usage of thermoplastic matrix materials, which offer significantly more energy absorption than thermosets, and thus are preferably used for impact-specific laminates.^21–23 In contrast, the adhesion of thermoplastic matrix materials to metals is often lower than those of thermosets. Moreover, the influence of residual stresses due to different coefficients of thermal expansion on the failure behaviour could be more relevant because of higher production temperatures with thermoplastics.

Also, faster production times and more efficient in-line production capabilities can be realised.^24,25 However, a complex deformation and failure behaviour^26–28 with a significant strain rate dependency of the material properties^29–31 has to be taken into account.

Besides in-plane material properties, interlaminar material properties play a key role for the design of crash loaded structures. Especially, the understanding of interface adhesion mechanisms under various process conditions and impact loads are essential since delamination is a dominant failure mode of highly dynamic loaded FML and affects the structural behaviour as well as the energy absorption capability. The delamination behaviour is significantly influenced by the loading conditions in the interface. Therefore, the structural response of FML under highly dynamic loading conditions is evaluated using Charpy tests.³² Specimens with both different support lengths and laminate thicknesses are tested and evaluated with respect to the bending and shear stress distribution ratios. Moreover, the influence of light metal surface topology and material configuration on the bending response is analysed.

Experimental set-up

Investigated material

For the metal layers, magnesium and aluminium alloys have been chosen based on their high strength-to-weight ratio and the subsequent applicability for lightweight designed vehicles. Twin-roll cast and hot-rolled magnesium alloy sheets (Mg-AZ31, thicknesses 1.0 and 0.6 mm) as well as aluminium alloy sheets (Al-5754, thicknesses 1.0 and 0.5 mm) are used. The sheets are in an initially annealed state, free of strain hardening with a homogeneous microstructure. The main alloying elements of Mg-AZ31 are 3% aluminium and 1% zinc, whereas the Al-5754 sheets are made from an aluminium alloy with 3% magnesium. The mechanical properties have been obtained parallel (0°) and perpendicular (90°) to the rolling direction with a universal tensile testing machine (AG-100 Shimadzu) with a maximum tensile capacity of 100 kN (see Figure 1).

Figure 1.

Mechanical properties obtained with a universal tensile testing machine of the Mg-AZ31 (a) and Al-5754 (b) sheets investigated.

The sheet surface roughness was analysed with a roughness-measuring machine, which determines the standard-compliant roughness values from the primary profile, waviness profile and roughness profile along a sampling length of 4.0 mm. The roughness of the investigated sheets, as mentioned in Table 1, has been determined before interface pre-treatment via embossing. These initial sheets serve as reference state and are hereinafter referred to as interface I0. The maximum profile height of 298 µm in Figure 2 is the distance from the lowest (centre of rough pattern) to the highest (elevation on edges) measuring point in that measuring range. Due to hydrostatic pressure during embossing, only the positive fraction of the tool surface profile is transferred to the sheet (as seen in Figure 3(b)). This sheet profile depth only amounts to local maxima of 150 µm.

Figure 2.

Surface topology of embossing tools measured by means of a 3D scanning microscope: fine roughness (a), coarse roughness (b), rough pattern (c) and fine pattern (d).

Figure 3.

Grey-scale pictures of the embossed sheets with sink eroded tool: interface topology I2 (a), coarse electron beam pattern: interface topology I1 (b) and photo shot of cut out samples, respectively (c).

Table 1.

Roughness values for the Mg-AZ31 and Al-5754 sheets (reference interface I0) investigated before interface pre-treatment via embossing.

		Mg-AZ31	Al-5754
Mean roughness value	R_a in µm	0.38	0.25
Mean roughness depth	R_z in µm	2.14	1.60
Maximum roughness depth	R_zmax in µm	4.92	2.07

Evaluating the influence of metal surface topology on the adhesion to the thermoplastic constituent, a pre-study regarding roughness and pattern has been performed. As roller embossing is a process used in bulk sheet production lines to fine-adjust final sheet surface properties, in the present work, direct embossing has been chosen as a pre-assessment method to evaluate different types of tool surface topologies in order to optimise mechanical adhesion in the fibre-metal-laminate. However, both sides of the metals sheets are pre-treated. As reviewed by Sinmazçelik et al.,¹⁰ most of the metallic surface pre-treatments are based on chemical processes such as anodising, etching or oxidation. The mechanical pre-treatment methods mostly serve as a precursor, as well as in this work, to give a good macro profile to further improve mechanical properties in the interface. Whilst chemical pre-treatment aims to enhance mechanisms of bonding such as electrostatic attraction, chemical bonding or molecular entanglement, mechanical bonds are achieved through specific micro- or macro-roughness.^33,34 On this basis, four different embossing structures have been chosen, resembling the most important characteristics of metallic surface structures. Two embossing tools with different roughness characteristics and two embossing tools with different patterns are investigated. The surface topologies of the embossing tools were measured by means of a three-dimensional scanning microscope and are shown in Figure 2. By blasting the first embossing tool with hard and sharp iron particles, a surface topology with sharp edges and high number of peaks per area has been created. In contrast, the second embossing tool surface was textured via a sink erosion method which causes the formation of rounded tips by briefly melting the surface. Compared to the blasted embossing tool, the parameters have been chosen to give a much coarser structure, which is distributed randomly and evenly, as well. The third and fourth embossing tools have been surface textured via electron beam. With an electron beam, it was possible to create customised patterns on the surface, including defined profile depth and dimensions. Here, two largely different dimensions where chosen as well, to examine the transferability of different tool topographies to the sheets via embossing.

The embossing tools have a width of 170 mm and length of 385 mm, chosen with respect to a sufficient size for further processing to multi-layered FML. Clamped in a specially designed device with plane-parallel tool surfaces, the sheets and embossing tool were pressed together in one step at room temperature. A hydraulic press with a maximum force of 10 MN (1000 t) was used to achieve the required embossing pressure. Preliminary tests have shown that due to hydrostatic pressure states, press forces above 500 t (125 MPa) do not increase the embossing depth significantly anymore. Thus, an equal force of 500 t was chosen to emboss the Mg-AZ31 and Al-5754 sheets with each tool. For the further investigations, one tool variant was selected from each of the stochastically structured variants (Figure 2(a) and (b)) and from the periodically structured variants (Figure 2(c) and (d)) to emboss the sheets. The sink erosion process causes the formation of rounded tips by short local melting of the tool surface. For further investigations, the sink eroded tool (Figure 2(b)) was selected, as it resulted in a much coarser and higher structured sheet than the sheet embossed with the blasted one (Figure 2(a)). This surface topology is named by the test parameter I2. With the aid of the electron beam process, it was possible to set tailor-made patterns on the tool surfaces with defined profile depths and dimensions (Figure 2(c) and (d)). In comparison to the sheets embossed with the tool of Figure 2(d), the selected ones embossed with the tool of Figure 2(c) are characterised by deeper impressions in the sheet surface, from which the authors derive more significant results. This surface topology is named by the test parameter I1. Figure 3 shows the chosen embossing texture used for the investigations. The pre-treated sheets with characteristic surface topologies I1, I2 as well as the reference interface I0 were cleaned with ethanol prior to the joining process of the hybrid laminates.

During sheet embossing with the sink erosion tool, a maximum roughness depth of 60 µm was achieved. The tool itself is characterised by profile peaks as high as 250 µm. Embossing with the electron beam textured tool has shown maximum roughness depth of up to 150 µm in the sheets. Whereas the tool exhibits profile peaks as high as 300 µm. The sheet thickness for both Mg-AZ31 and Al-5754 sheets has shown no measurable difference in embossing depth.

The CFRP layers consist of a unidirectional CFRP-6, namely, a CELSTRAN® CFR-TP PA6 CF60-01. It is a commonly used product in automotive and industrial industry. As many carbon fibre-reinforced thermoplastic materials, it combines advantages like high specific mechanical properties with high productive processes like pressing resulting in short process cycles. The material is used as a tape with 275 mm width and a fibre volume ratio of 60%. The associated product description³⁵ indicates a ply thickness of 0.13 mm and provides additional mechanical properties. The adhesion to the metal layers is supported by a 0.1 mm thick adhesive layer based on polyolefin: Cox 391 from Nolax. The hybrid laminates were produced using a hot press. The laminates were consolidated under vacuum conditions with a pressure of 5 bar and processing temperature of 250℃ for 20 min in order to ensure the melting temperature in the middle of all laminates. The test specimens were cut out by means of water jet cutting. The edges have been slightly grinded to assure a good edge quality.

Drop tower impact tests and ductility index

The drop tower impact test is a three-point bending experiment of either notched or unnotched beams. In the present study, an instrumented drop tower is used, namely, a Magnus 1000 from Coesfeld. Although the Charpy impactor follows a circular trajectory and the drop tower impactor falls vertically from top to bottom, the general test set-up is very similar (Figure 4(a)). The impactor hits the top of the specimen orthogonal and transfers its kinetic energy to the specimen. The used impactor and support geometry are accorded to DIN EN ISO 13802.³⁶ Additionally, a high-speed camera was used to record the structural behaviour of the specimens with a frame size of 768 pixel × 512 pixel and a frame rate of 10,000 Hz. The recording requires additional lighting that induces heat in the specimen. An active air ventilation was used to cool the test chamber resulting in a constant testing temperature of 29℃.

Figure 4.

(a) Drop tower experiment test set-up and (b) typical force–deflection curve.

The drop tower records impact force F and determines the specimen deflection u by solving the equation of motion according to DIN EN ISO 6603-2.³⁷ Integrated light barriers measure the impactor velocity, while the impactor force is measured by a load cell based on a piezoelectric effect. The measurement of force and deflection during the test enables an automated evaluation of the total absorbed energy E as the integral of the force–deflection curve. This total energy can be divided into two parts (Figure 4(b)). Energy losses due to bearing friction and air resistance can be neglected because of their small contribution to the energy balance.

The initial energy is defined by the integral of the force–deflection curve from u = 0 to u_i. Whereby u_i is the deflection where the force reaches its maximum. In contrast, the propagation energy is defined by the integral of the force–deflection curve from u_i to u_p that is the maximal deflection of the experiment $\begin{matrix} E = E_{i} + E_{p} = \int_{0}^{u_{i}} F (u) d u + \int_{u_{i}}^{u_{p}} F (u) 1 mu d u \end{matrix}$ (1)

In order to compare the absorbed energies of different material densities ρ and lay-ups, it is necessary to relate them to the structural mass.³⁸ Therefore, the measured energies are divided by the specimen mass m between the two supports. The mass is calculated by the product of the effective Volume V_eff that is the product of the support span, the specimen width, the laminate thickness t and the averaged density $\bar{ρ}$ . The averaged density of each lay-up is calculated by the sum over all layers n of the weighted density of each layer i $\begin{matrix} \bar{ρ} = \frac{1}{t} \sum_{i = 1}^{n} ρ_{i} t_{i} \end{matrix}$ (2)

As a result, the specific absorbed energy e is defined by $e = e_{i} + e_{p} = \frac{E_{i}}{\bar{ρ}; V_{ref}} + \frac{E_{p}}{\bar{ρ}; V_{ref}}$ (3)

The ductility index DI is defined as the ratio of propagation to initiation energy³⁹ $DI = \frac{E_{p}}{E_{i}}$ (4)

Additionally, a dimensionless characteristic parameter E^R to describe the energy dissipation behaviour is the ratio of impact energy E_imp to total absorbed energy E. It is calculated as follows $E^{R} = \frac{E_{imp}}{E_{i} + E_{p}}$ (5)

Experimental programme

Since the structural response is influenced by several parameters, a comprehensive test set-up programme was developed including 48 different configurations (Table 2). At first, the influence of loading condition on the structural response was analysed. Therefore, two different support lengths (24 and 60 mm) have been tested while the laminate thicknesses (nominal 3 and 6 mm) were not changed. The greater the thickness-support length-ratio, the higher the shear loading in the specimen compared to the bending loading. Moreover, the structural responses are influenced by the applied light metal material and hence two different light metals are investigated, namely, an Mg-AZ31 and an Al-5754. In order to identify the influence of the interlaminar properties, two mechanically pre-treated metal surface (I1 and I2) are tested against the reference interface (I0). Since it seems that the structural response is influenced by the laminate stacking sequence, five different lay-ups have been tested (L1–L5, see Figure 5). The scope of the present study is to investigate the influence of the metal-composite-interface and resulting delaminations on the failure behaviour of the FML. Hence, different amount of layers (from 4 to 16) were used. Moreover, two laminate thicknesses have been investigated since the loading by shear favours delamination initiation. In order to comply with these boundary conditions, the thickness ratio of the layers was left constant by almost all lay-ups (except L3). All laminates are symmetric due to different coefficient of thermal expansion. Furthermore, two different impact heights resulting in different loading velocities have been analysed to consider the strain rate-dependent material behaviour of both metal and CFRP material. Although some studies have shown that the influence of impact velocity on the structural response is not significant,⁴⁰ the scope of this investigation is to proof this statement for the investigated materials and interfaces. For each configuration, at least five experiments were carried out resulting in 266 experiments in total.

Figure 5.

Investigated lay-ups for both Mg-AZ31 and Al-5754 hybrid laminates.

Table 2.

Test configurations and corresponding test parameter.

Config. nr.	Light metals	Inter- face	Lay-up	Support length in mm	Impact mass in kg	Impact height in mm	E_i in J	E_p in J	e_i in kJ/kg	e_p in kJ/kg	e in kJ/kg	DI	E^R
1	Mg	I0	L1	24	5.45	400	8.81	8.12	3.90	3.59	7.49	0.92	1.26
2					10	1000	6.73	6.34	2.98	2.81	5.78	0.94	7.51
3				60	10	400	8.63	3.19	1.53	0.57	2.09	0.37	3.32
4					10	1000	7.79	3.10	1.38	0.55	1.93	0.40	9.01
5		I0	L2	24	5.45	400	3.01	5.42	1.62	2.91	4.53	1.80	2.54
6					10	1000	2.97	5.55	1.60	2.98	4.58	1.87	11.52
7				60	10	400	6.58	6.05	1.41	1.30	2.71	0.92	3.11
8					10	1000	5.09	4.75	1.09	1.02	2.11	0.93	9.97
9		I0	L4	60	20	400	24.53	30.71	2.14	2.68	4.81	1.25	1.42
10					20	1000	26.55	24.96	2.31	2.17	4.49	0.94	3.81
11			L5	60	20	400	18.22	15.20	1.92	1.60	3.51	0.83	2.35
12					20	1000	18.57	17.90	1.95	1.88	3.83	0.96	5.38
13	Al	I0	L1	24	10	400	9.75	15.94	3.60	5.88	9.48	1.64	1.53
14					10	1000	6.45	22.59	2.38	8.33	10.71	3.50	3.38
15				60	10	400	3.31	10.28	0.49	1.52	2.01	3.10	2.89
16					10	1000	4.60	7.72	0.68	1.14	1.82	1.68	7.97
17		I0	L2	24	10	400	5.30	9.36	2.07	3.66	5.73	1.77	2.68
18					10	1000	6.31	9.11	2.46	3.56	6.02	1.44	6.36
19				60	10	400	5.70	11.42	0.90	1.81	2.72	2.00	2.29
20					10	1000	5.44	9.35	0.86	1.48	2.35	1.72	6.63
21		I0	L4	60	20	400	27.36	26.04	2.00	1.90	3.90	0.95	1.47
22					20	1000	28.72	27.31	2.10	1.99	4.09	0.95	3.50
23			L5	60	20	400	16.88	29.14	1.30	2.25	3.55	1.73	1.71
24					20	1000	20.43	27.92	1.57	2.15	3.73	1.37	4.06
25	Mg	I1	L2	24	10	400	2.84	3.97	1.53	2.14	3.66	1.40	5.76
26					10	1000	1.77	4.83	0.95	2.59	3.54	2.73	14.88
27				60	10	400	4.49	4.49	0.99	0.99	1.97	1.00	4.37
28					10	1000	4.92	2.33	1.08	0.51	1.59	0.47	13.53
29		I2	L2	24	10	400	2.84	5.36	1.49	2.82	4.32	1.89	4.79
30					10	1000	2.52	5.48	1.33	2.88	4.21	2.17	12.27
31				60	10	400	3.04	3.60	0.65	0.77	1.43	1.18	5.91
32					10	1000	2.78	3.98	0.60	0.85	1.45	1.43	14.50
33		I2	L3	24	10	400	2.85	4.33	1.53	2.33	3.86	1.52	5.47
34					10	1000	3.48	4.31	1.87	2.31	4.18	1.24	12.61
35				60	10	400	4.96	4.06	1.09	0.89	1.98	0.82	4.35
36					10	1000	4.60	2.79	1.01	0.61	1.63	0.61	13.26
37	Al	I1	L2	24	10	400	5.68	10.29	2.22	4.02	6.24	1.81	2.46
38					10	1000	6.53	9.53	2.55	3.72	6.27	1.46	6.11
39				60	10	400	5.96	9.97	0.95	1.58	2.53	1.67	2.46
40					10	1000	5.74	8.85	0.91	1.40	2.32	1.54	6.72
41		I2	L2	24	10	400	1.87	11.45	0.74	4.54	5.28	6.13	2.95
42					10	1000	4.75	8.51	1.88	3.38	5.26	1.79	7.40
43				60	10	400	2.93	6.46	0.51	1.13	1.65	2.20	4.18
44					10	1000	2.74	4.99	0.48	0.88	1.36	1.82	12.68
45		I2	L3	24	10	400	5.03	11.39	1.96	4.45	6.41	2.27	2.39
46					10	1000	6.81	9.13	2.66	3.57	6.23	1.34	6.16
47				60	10	400	4.86	8.06	0.77	1.28	2.05	1.66	3.04
48					10	1000	5.33	8.30	0.85	1.32	2.16	1.56	7.19

DI: ductility index.

Results and discussion

Figure 6 illustrates the results from Table 2 in a reordered sequence. First, a distinction is made between the both support lengths of 24 mm and 60 mm. Second, it is sorted by the light metal: Mg-AZ31 and Al-5754. Then, the data are sorted by the interface and by the lay-up. For each configuration, two columns are drawn representing two different loading energy levels. Besides, the standard deviation for each type of energy is represented by the grey error bars. Due to measurement errors during the tests, only one valid experiment for configuration 30 could be analysed. Hence, no error bars are drawn for this configuration.

Figure 6.

Results of drop tower experiments: total absorbed specific energy e divided into initial specific energy e_i (light colour) and specific propagation energy e_p (dark colour).

Provided energy

The two neighboured columns represent the same test set-up but different impact energies. There is no clear tendency for a correlation of impact energy and absorbed energy. Moreover, the difference lies mostly within the scatter of the results indicating that the provided energy is high enough to break all specimens.

Support length

In general, the absorbed energy level of the 24 mm support length experiments is significantly higher than the 60 mm ones. However, a direct comparison of the configurations is difficult. The thick laminates cannot be tested at the short support length due to clamping between impactor, specimen and support. Comparing the thin laminates (L1 to L3) of the short support length with the long support length, the averaged absorbed specific energy drops from 4.62 to 1.89 kJ/kg for Mg-AZ31 and from 6.76 to 2.10 kJ/kg for Al-5754.

The smaller the support length, the more the specimen is loaded by interlaminar shear stress in relation to the bending stress. Since interlaminar strengths are significantly lower than the in-plane ones, the shear stress distribution results in delaminations between the metal and the CFRP layers. An initial delamination leads to a drop in the force signal. The propagation of delaminations increases the absorbed propagation energy. Consequently, the averaged ductility index of the thin lay-ups drops from 1.98 for the short support length to 1.35 for the long support length. Also, the deformation behaviour significantly differs from the shear loaded specimen (Figure 7). Figure 7(a) illustrates the deformation and failure behaviour of an Al-5754 specimen with lay-up L1 with short support length. Because of relatively high interlaminar shear stresses over the full length of the specimen, the metallic layers deform plastically near the support and the load fin. Moreover, the interlaminar shear stress induces energy absorbing delaminations. The bending moment causes a compressive load in the upper layers of the specimens. Therefore, these layers move upwards after delamination due to buckling failure and form an s-shape.

Figure 7.

Governing parameter support length: (a) Delamination and plastic deformation at short support length (config. nr. 14) and (b) almost elastic deformation at long support length (config. nr. 16).

In contrast, the major plastic deformation of long support length experiments (Figure 7(b)) is concentrated near the loading fin. This deformation zone is relatively small compared to the effective volume of the specimens. Additionally, a localised in-plane failure near the load fin occurs due to compressive loads in the top layer resulting in a bulging failure. This failure initiation does not lead to more propagating failure mechanisms. Because of these localised process zones compared to the effective volume of the specimens, the specific (mass related) absorbed energy is smaller with higher support lengths.

Light metals

The comparison between Mg-AZ31 and Al-5754 hybrid laminates shows more significant trends, since every configuration is tested with both light metals. Averaging the values of the specific energy for one material and support length combination does not represent an engineering characteristic. However, the comparison allows a qualitative evaluation of the materials’ energy absorption capacities under equal testing conditions. Each of the testing configurations (interface I and lay-up L) is represented equally in both material combinations allowing to derive a generalised tendency. With the support length of 24 mm, the Al-5754 hybrid laminates absorb approximately 6.76 kJ/kg, while the Mg-AZ31 hybrid laminates absorb approximately 4.62 kJ/kg. Moreover, the ductility index is significantly higher for Al-5754 at 2.31 than for Mg-AZ31 at 1.65. In general, the Al-5754 hybrid laminates allow more deflection until final failure because of their high forming capacity. This results in more delaminations and plastic deformation which increases especially the absorbed specific energy. The effect becomes smaller with increasing support span.

For the tested support length of 60 mm, the absorbed specific energies are almost the same (Mg-AZ31: 2.54 and Al-5754: 2.59 kJ/kg). Bending loading conditions dominate the material response and the failure phenomena are completely different to the tests with the short support span. There are almost no delaminations observed for both metal material types due to low interlaminar shear stresses. In contrast, the failure behaviour near the loading fin differs significantly. The Mg-AZ31 shows less deformation capacity compared to Al-5754, which exhibits a higher plastic deformation. Therefore, the failure mechanism for Mg-AZ31 is tensile fracture in the bottom metal layers. The density-related tensile strength is higher for the used Mg-AZ31, which forces these specimens to absorb relatively high specific initial energy compared to the Al-5754 specimen (1.34 to 1.03 kJ/kg). In contrast, the failure behaviour of the magnesium alloys is much more brittle resulting in less specific propagation energy (1.17 to 1.56 kJ/kg; Figure 8). Even though the overall specific energy is nearly the same for both metal hybrid laminates, both the deformation and failure behaviour is very different resulting in a significantly higher averaged ductility index for Al-5754 at 1.71 than for Mg-AZ31 at 0.87.

Figure 8.

Governing parameter light metal material: (a) Brittle tensile ply failure in Mg-AZ31 layers (config. nr. 4) and (b) more plastic deformation in Al-5754 layers (config. nr. 16).

Figure 9.

Governing parameter surface topology: Similar deformation and failure phenomena of specimen with different pre-treatment technologies (a) reference (config. nr. 8) and (b) cross texture (config. nr. 28) and point texture (config. nr. 32).

Metal surface topology

The developed pre-treatment process influencing the interface between the metal and the CFRP layers is currently only applicable for one side of the metal layers. Hence, only sandwich laminates were manufactured with metal cover layers and pretreated inner surfaces (lay-ups L2 and L3). Delaminations do not occur for sandwich laminates. Therefore, no influence of the surface topology on the energy absorption can be seen by high-speed camera images (Figure 9). In detail, a difference of failure mechanisms is not observable. All specimens exhibit comparable deformation and failure behaviour. There is a very localised zone near the loading fin where the specimens deform plastically in all configurations. The test specimens are bent to almost 90° until the lower metallic cover layer fails due to high tensile stresses. At this stage, some fibre tensile fracture occurs. Because there are no differences in deformation and failure behaviour, energy absorption is also comparable. In detail, the absorbed energy of Mg-AZ31 hybrid laminates with modified surface topology decreases up to 23% for 24 mm support length (config. nr. 26 vs. 6) and 47% for 60 mm, respectively (config. nr. 31 vs. 7). In contrast, no clear tendency for Al-5754 FML can be seen. While the absorbed energy is increased for short support length up to 9% (config. nr. 37 vs. 17), it is decreased for long support length up to 42% (config. nr. 44 vs. 20).

Since no delaminations occur, it is assumed that the modified metal surface topology induced stress concentrations in the metal layer resulting in less bending strength and in less energy absorption consequently.

Lay-ups

The lay-ups have a significant influence on the energy absorption capacity. All laminates tested with the short support length have a nominal thickness of 3 mm. The specific tensile stiffness and strength of CFRP is higher than the magnesium alloy ones. Lay-up L1 consists of more outer lying CFRP layers resulting in higher bending stiffness and strengths and that leads to higher initial energies compared to lay-ups L2 and L3 (Mg-AZ31: 3.44 compared to 1.49 kJ/kg and Al-5754: 2.99 compared to 2.07 kJ/kg). Moreover, the lay-up L3 absorbs more energy because the layers of CFRP with high strength and stiffness are outwardly positioned resulting in a higher structural stiffness and strength and consequently a higher absorbed energy compared to lay-up L2. Again, the reason for a higher propagation energy is the delamination occurrence due to lower interlaminar properties between metal and CFRP layers (Mg-AZ31: 3.20 compared to 2.62 kJ/kg and Al-5754: 7.11 compared to 3.86 kJ/kg; Figure 10). The same logic applies for lay-ups L4 and L5 with 60 mm support length, where the total absorbed energy is significantly higher than the other lay-ups (Mg-AZ31: 4.16 compared to 1.89 kJ/kg and Al-5754: 3.82 compared to 2.10 kJ/kg). Figure 10 illustrates the deformation and failure behaviour of lay-up L1 (a) and lay-up L4 (b). While the failure behaviour of the lay-up L1 is mainly characterised by tensile failure of the lower metallic layers, the lay-up L4 shows a more complex failure behaviour. Here, in addition to the very local tensile failure of the lower metallic layers with the bulging of the upper metallic layer, shear-induced delaminations and high plastic deformations over the entire specimen length can be observed. Since these failure mechanisms have a large-volume effect, they also have a significant influence on the energy absorption capacity. Consequently, the energy absorption capacity increases with increasing laminate thickness. Since the lay-up L5 exhibit lower energy levels (initiation and propagation), it can be assumed that a higher number of layers leads to higher energy absorption at the same thickness. This observation can also be explained by the more complex failure mode interaction in lay-up L4.

Figure 10.

Governing parameter lay-up: (a) Very local failure phenomena just as observed with tensile ply failure of lay-up L1 (config. nr. 4) and (b) delamination and plastic deformation of lay-up L4 (config. nr. 10).

As another result it can be stated, that the lay-up has the most influence on the structural performance of hybrid laminates. Since laminates with several layers invariably absorb much more energy by delaminations, the potential of multi-layer hybrid laminates has been demonstrated.

Conclusions

Lightweight materials are of increasing interest for the application in crash- and safety-relevant structures. Numerous possibilities exist to design the structural response of laminated composites under highly dynamic loading condition. In the present study, fibre-metal-laminated composites are investigated within low-velocity three-point-bending experiments to analyse the complex failure and energy absorption behaviour and to derive a fundamental understanding for the design of crash loaded structures.

The hot pressing process is generally used for the manufacturing of multi-layered hybrid laminates. Here, up to 16 single layers made of Mg-AZ31 or Al-5754 in combination CFRP have been consolidated. The influence of different stacking sequences, impact energies and mechanical interface topologies on the specific energy absorption and failure patterns was explored. In accordance to the study by Múgica et al.⁴¹ and Pärnänen et al.,⁴² it was found that the failure behaviour between Mg-AZ31 and Al-5754 hybrid laminates differs significantly. The deformation capacity of Mg-AZ31 under highly dynamic loadings condition is limited and causes a more brittle failure behaviour than the Al-5754 variants. This results in an averaged 40% higher ductility index for the Al-5754 hybrid laminates which means that much more energy is absorbed by propagating failure. The specific absorbed energy by propagation is higher for the Al-5754 variants (from −29% up to 197%). Although the increase in the total energy absorption capacities varies from −19% up to 85%, a generalised tendency of a higher total energy absorption capacity of the aluminium compared to the magnesium alloy hybrid laminates can be derived.

The bonding behaviour of the used adhesive layer was found to superimpose the influence of the different surface topologies and significant differences could not be identified. Therefore, either the pre-treatment technology may be adjusted or an adhesive layer with lower mechanical properties should be applied in future studies. The largest impact on the structural behaviour was found in the stacking sequence. Generally, it can be stated that thinner and more layers lead to more specific absorbed energy (up to 78%). Especially, more delamination can be observed with an increasing number of layers. Also, a higher shear-to-bending-ratio leads to more delaminations. This results in a higher energy absorption capacity based on higher failure propagation energies In accordance with Múgica,⁴⁰ no significant influence of impact velocity could be derived.

It has been shown that FML could have a potential to be used in crash-loaded lightweight structures. In general, laminates with more layers are beneficial. Especially, the delamination behaviour has a significant influence on the structural response as well as on the energy absorption capacity.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research,authorship,and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research,authorship,and/or publication of this article: The authors gratefully acknowledge the financial support of this research within the project “hybCrash” by the European Union (European Regional Development Fund) and the Free State of Saxony under the grant agreement no. 100285086.

Supplemental material

Supplemental material for this article is available online.

References

Dursun

Soutis

. Recent developments in advanced aircraft aluminium alloys. Mater Design 2014; 56: 862–871.

Bloeck

Aluminium sheet for automotive applications. In: Rowe

(ed). Advanced materials in automotive engineering, Cambridge: Woodhead Publishing, 2012, pp. 85–108.

Zhu

Sun

et al.

Energy absorption of metal, composite and metal/composite hybrid structures under oblique crushing loading. Int J Mech Sci 2018; 135: 458–483.

Dlugosch

Fritsch

Lukaszewicz

et al.

Experimental investigation and evaluation of numerical modeling approaches for hybrid-frp-steel sections under impact loading for the application in automotive crash-structures. Compos Struct 2017; 174: 338–347.

Reuter

Sauerland

K-H

Tröster

. Experimental and numerical crushing analysis of circular cfrp tubes under axial impact loading. Compos Struct 2017; 174: 33–44.

Liu

et al.

Lateral crushing and bending responses of cfrp square tube filled with aluminum honeycomb. Compos B Eng 2017; 118: 104–115.

Liu

Zhang

. Impact responses of sandwich panels with fibre metal laminate skins and aluminium foam core. Compos Struct 2017; 182: 183–190.

Zhang

Wang

et al.

Energy absorption and low velocity impact response of polyurethane foam filled pyramidal lattice core sandwich panels. Compos Struct 2014; 108: 304–310.

Bergmann

Heimbs

Progressive bearing failure of composites for crash energy absorption. In: Lopresto

Langella

Abrate

(eds). Dynamic response and failure of composite materials and structures, Woodhead Publishing, 2017, pp. 299–334.

10.

Sinmazçelik

Avcu

Özgür Bora

et al.

A review: fibre metal laminates, background, bonding types and applied test methods. Mater Des 2011; 32: 3671–3685.

11.

Carrillo

Cantwell

. Mechanical properties of a novel fiber-metal laminate based on a polypropylene composite. Mech Mater 2009; 41: 828–838.

12.

Sadighi

Alderliesten

Benedictus

. Impact resistance of fiber-metal laminates: a review. Int J Impact Eng 2012; 49: 77–90.

13.

Nguyen

Zatar

Mutsuyoshi

Mechanical properties of hybrid polymer composite. In: Thakur

Thakur

Pappu

(eds). Hybrid polymer composite materials, Cambridge: Woodhead Publishing, 2017, pp. 83–113.

14.

Zopp C, Nestler D, Tröltzsch J, et al. Catpual – an innovative and high-performance hybrid laminate with carbon fibre-reinforced thermoplastic polyurethane. In: 21st Symposium on composites (vol. 742 of Key Engineering Materials), pp. 294–301. Zürich: Trans Tech Publications, 2017.

15.

Nestler

Trautmann

Nendel

et al.

Innovative hybrid laminates of aluminium alloy foils and fibre-reinforced thermoplastic layers. Materialwissenschaft und Werkstofftechnik 2016; 47: 1121–1131.

16.

Abouhamzeh

Sinke

Jansen

KMB

et al.

Thermo-viscoelastic analysis of glare. Compos B Eng 2016; 99: 1–8.

17.

Hua

et al.

Bending failure mechanism and flexural properties of glare laminates with different stacking sequences. Compos Struct 2018; 187: 354–363.

18.

Zarei

Brugo

Belcari

et al.

Low velocity impact damage assessment of glare fiber-metal laminates interleaved by nylon 6,6 nanofiber mats. Compos Struct 2017; 167: 123–131.

19.

Bikakis

GSE

. Simulation of the dynamic response of glare plates subjected to low velocity impact using a linearized spring-mass model. Aerosp Sci Technol 2017; 64: 24–30.

20.

Botelho

Silva

Pardini

et al.

A review on the development and properties of continuous fiber/epoxy/aluminum hybrid composites for aircraft structures. Mater Res 2006; 9: 09–09.

21.

Mullins

Liu

Sue

H-J

Chapter 2 – mechanical properties of thermosets. In: Guo

(ed). Thermosets, 2nd ed. Amsterdam: Elsevier, 2018, pp. 35–68.

22.

Peters

Engineering thermoplastics—materials, properties, trends. In: Kutz

(ed). Applied plastics engineering handbook, 2nd ed. (Plastics Design Library). Norwich: William Andrew Publishing, 2017, pp. 3–26.

23.

Larson

An overview of thermoplastic materials. In: Larson

(ed). Thermoplastic material selection, Norwich: William Andrew Publishing, 2015, pp. 97–143.

24.

Gerstenberger

Osiecki

Kroll

et al.

Processing and characterization of cathodic dip coated metal/composite-laminates. Arch Civil Mech Eng 2016; 16: 467–472.

25.

Nestler

Trautmann

Zopp

et al.

Continuous film stacking and thermoforming process for hybrid cfrp/aluminum laminates. Procedia CIRP 2017; 66: 107–112.

26.

Hufenbach

Gude

Böhm

et al.

The effect of temperature on mechanical properties and failure behaviour of hybrid yarn textile-reinforced thermoplastics. Mater Des 2011; 32: 4278–4288.

27.

Böhm

Gude

Hufenbach

. A phenomenologically based damage model for 2D and 3D-textile composites with non-crimp reinforcement. Mater Des 2011; 32: 2532–2544.

28.

Böhm

Gude

Hufenbach

. A phenomenologically based damage model for textile composites with crimped reinforcement. Compos Sci Technol 2010; 70: 81–87.

29.

Hufenbach

Langkamp

Gude

et al.

Characterisation of strain rate dependent material properties of textile reinforced thermoplastics for crash and impact analysis. Procedia Mater Sci 2013; 2: 204–211.

30.

Hufenbach

Hornig

Zhou

et al.

Determination of strain rate dependent through-thickness tensile properties of textile reinforced thermoplastic composites using l-shaped beam specimens. Compos Sci Technol 2011; 71: 1110–1116.

31.

Hufenbach

Gude

Ebert

et al.

Strain rate dependent low velocity impact response of layerwise 3D-reinforced composite structures. Int J Impact Eng 2011; 38: 358–368.

32.

Hufenbach

Marques Ibraim

Langkamp

et al.

Charpy impact tests on composite structures: an experimental and numerical investigation. Compos Sci Technol 2008; 68: 2391–2400.

33.

Baldan

. Adhesively-bonded joints and repairs in metallic alloys, polymers and composite materials: adhesives, adhesion theories and surface pretreatment. J Mater Sci 2004; 39: 1–49.

34.

Rider

Olsson-Jacques

Arnott

. Influence of adherend surface preparation on bond durability. Surface Interface Anal 1999; 27: 1055–1063.

35.

Material Data Sheet. Standard product values. Celstran cfr-tp pa6 cf60-01, http://tools.celanese.com/ (2014).

36.

DIN EN ISO 13802: 2016. Verification of pendulum impact-testing machines – charpy, izod and tensile impact-testing. Corrected version 2016-04-01.

37.

DIN EN ISO 6603-2. Plastics – determination of puncture impact behaviour of rigid plastics – part 2: instrumented puncture test. Corrected version: 2002-04-00.

38.

Sun

Wang

et al.

Energy absorption mechanics and design optimization of cfrp/aluminium hybrid structures for transverse loading. Int J Mech Sci 2019; 150: 767–783.

39.

Beaumont

PWR

Riewald

Zweben

Methods for improving the impact resistance of composite materials. In: Greszczuk

(ed). Foreign object impact damage to composites, West Conshohocken: ASTM, 1975, pp. 134–158.

40.

Múgica

Aretxabaleta

Ulacia

et al.

Rate-dependent phenomenological model for self-reinforced polymers. Compos A Appl Sci Manuf 2016; 84: 96–102.

41.

Múgica

Aretxabaleta

Ulacia

et al.

Impact characterization of thermoformable fibre metal laminates of 2024-t3 aluminium and az31b-h24 magnesium based on self-reinforced polypropylene. Compos A Appl Sci Manuf 2014; 61: 67–75.

42.

Pärnänen

Alderliesten

Rans

et al.

Applicability of az31b-h24 magnesium in fibre metal laminates – an experimental impact research. Compos A Appl Sci Manuf 2012; 43: 1578–1586.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.04 MB

An experimental study on the bending response of multi-layered fibre-metal-laminates

Abstract

Keywords

Introduction

Experimental set-up

Investigated material

Drop tower impact tests and ductility index

Experimental programme

Results and discussion

Provided energy

Support length

Light metals

Metal surface topology

Lay-ups

Conclusions

Footnotes

Declaration of Conflicting Interests

Funding

Supplemental material

References

Supplementary Material