Contents lists available at ScienceDirect

Composite Structures

journal homepage: www.elsevier.com/locate/compstruct

Direct injection molding and mechanical properties of high strength steel/composite hybrids

Yiben Zhanga,b, Lingyu Suna,b, Lijun Lia,b,⁎, Bincheng Huanga,b, Taikun Wangc, Yantao Wangc

a School of Transportation Science and Engineering, Beihang University, 37 Xueyuan Road, Haidian District, Beijing 100191, Chinab Lightweight Vehicle Innovation Center, Beihang University, 37 Xueyuan Road, Haidian District, Beijing 100191, Chinac Zhengzhou Electromechanical Engineering Research Institute, Zhengzhou 450015, Henan, China

A R T I C L E I N F O

Keywords:Metal polymer hybridMechanical propertiesFailureAnalytical modelFinite element methodInterface

A B S T R A C T

Metal polymer hybrid technology is a promising way of automotive weight reduction and safety improvement. Inthis study, high strength steel and glass-fiber reinforced PA66 composite were integrated by direct injectionmolding process. The mechanical properties and failure behavior of steel/composite hybrids under tension,bending, mode I and II fracture loads were investigated experimentally, theoretically and numerically. Theproperties of the interface between steel and composite were measured using double cantilever beam and end-notched flexure tests for the first time. The analytical model and finite element simulation considering theinterface mechanical behavior were proposed and validated. They can be utilized to predict the mechanicalproperties and progressive failure process of hybrid material cost-effectively and accurately. It is found that thecracks initiate at the interface, and then propagate through composite until the hybrid material loses load-carrying capacity completely. The steel dominates the load-bearing capacity, and the interface dominates thefailure process. The results obtained in this study offer a scientific guidance and theoretical basis for the en-gineering application of metal polymer hybrid materials.

1. Introduction

Due to excellent lightweight performance and high degree of in-tegration, metal polymer hybrids (MPHs) are being used with in-creasing frequency to replace traditional all-steel components such asautomotive bumper beams, instrument-panel cross beams and front-endmodules [1–3]. Especially in the automotive industry, high strengthsteel (HSS)/glass-fiber reinforced polymer (GFRP) hybrids have at-tracted a significant amount of attention for their low material cost andgood load-bearing capacity. However, fabricating MPHs is a challen-ging subject since metal and polymer have very different chemical andphysical properties. These differences cause problems such as poorbonding between metal and polymer.

Therefore, different methods have been developed to obtain MPHmaterials and parts. For example, metal and polymer are formed se-parately and then connected through adhesive bonding [4–6], plastic-rivets [7], or melting a pre-laid specific membrane on a metal surface inthe over molding [8,9] processes. However, adhesive bonding needsmore positioning and curing time. Plastic-rivets decreases the strengthof steel due to the holes required and melting a pre-laid specificmembrane increases the tooling cost and cycle time.

Direct injection molding (DIM) [10] is a new technology proposedfor almost one step manufacturing MPHs, which does not needpunching dozens of holes, processing over-molded edges or interlockingrivets. Nonetheless, there are still some problems that hinder its wideapplication. For example, a surface pretreatment prior to injectionmolding is often needed for improving adhesion between metal andpolymer. Nowadays, the pretreatment methods include primer coating[11], grit blasting [12], laser ablation [13], plasma treatment [14],chemical etching [15], chemical solution modification [16]. Combina-tions of these methods are also available and often utilized, e.g. am-monia aqueous etching and plasma treatment [17]. However, thesemethods have poor joining efficiency and compatibility with existingproduction lines because they need additional pre-process tools besidesstamping and injection molding. Therefore, it is imperative to develop anew DIM technology possessing low cost, high efficiency and goodcompatibility.

The limited knowledge on the mechanical properties of hybridsrestricts the further promotion of DIM technology. Hence, compre-hensive understanding the mechanical properties, damage propagationand final failure modes of MPHs is the premise of industry application.Moreover, most of published studies only focused on the overall

https://doi.org/10.1016/j.compstruct.2018.11.035Received 30 March 2018; Received in revised form 19 October 2018; Accepted 13 November 2018

⁎ Corresponding author.E-mail address: lilijun@buaa.edu.cn (L. Li).

Composite Structures 210 (2019) 70–81

Available online 14 November 20180263-8223/ © 2018 Elsevier Ltd. All rights reserved.

T

performance of MPHs with specific geometry [18–21]. The interfacebetween metal and polymer is generally assumed perfect bonding intheir simulation. To eliminate the geometric or architectural effects, itis essential to obtain the fundamental properties at material couponlevel. Dlugosch et al. [4] fabricated the HSS/GFRP hybrid coupons byglue and then tested their mechanical properties. They found the hy-brids possessed different energy dissipative mechanisms and failuremodes from individual steel or composites. However, the interface inDIM hybrids is very different from the glue layer. Currently, the largeerrors in simulation of DIM hybrid structures [22] come from the per-fect bonding assumption due to the lack of experimental data and in-terface constitutive model. This new type of interface may possibly leadto different failure mechanisms, which needs another analytical modelto describe its fundamental mechanical behavior, and also requiresaccurate numerical simulation method based on the experimental data.

This paper aims to establish a comprehensive knowledge system forDIM MPH materials through a series of work. Firstly, a new DIMtechnology with the surface pretreatment method well compatible withstamping process was proposed, and four kinds of high strength steel/composite hybrid coupons were fabricated by this technology andtested (Section 2.1–2.3). Then, an analytical method was proposed topredict the mechanical properties of hybrids (Section 2.4). Next, thenumerical modelling and simulation procedures were introduced(Section 2.5). The progressive damage behavior of the interface be-tween metal and polymer was considered. Finally, the experimental,analytical and numerical results were compared and discussed (Section3).

2. Materials and methods

2.1. Materials

The metal used was dual-phase steel DP 590+Z (Benxi Iron andSteel Group Corporation, China) with 1mm thickness. The injectionmolding polymer used was Ultramid®A3HG7, a kind of 35 wt% shortglass-fiber reinforced PA66 (PA66-GF35) with 3mm length fibers,manufactured by BASF Application Chemical Company of China. Itsmelting temperature is 260 °C.

2.2. Fabrication process

In this study, a new direct injection molding process was developed,as shown in Fig. 1. Before injection molding, the steel surface was pre-processed by pressing with embossing die in order to obtain micro-mechanical interlocking structures. In the actual production, the mi-crostructures of steel stamping surface facing the injection part can bemanufactured by a punch with micro convexes. Simultaneously, the

steel surface pretreatment and stamping are completed. Hence, thisprocess is well compatible with the existing stamping line.

The direct injection molding process is comprised of three steps:Step 1. The steel plate was cut into specific shape and dimension.

Then, the steel surface facing the injection part was purified withethanol and embossed by a hydraulic indenter to form a diamond-shaped microstructure pattern.

Step 2. The embossed steel insert was heated to 370 °C in a heatertable, and then it was moved into the mould by hand rapidly. The steelinsert was positioned by the cylindrical magnets in the mould. Themold temperature was 80 °C. Before injection, the temperature of thesteel insert can be maintained at 200 °C based on multiple measure-ments, which was helpful to reduce the mismatched thermal expansionstresses between steel and polymer [23].

Step 3. The injection molding was carried out at the injection rate of20 cm/s under melt temperature of 270 °C. The maximum machineclamp force was 3800 kN. The injection pressure was 13.5 MPa, andhold pressure was 45MPa. The cycle time was approximately 60 s,wherein the hold time was 15 s and the cooling time was 30 s.

In order to improve the versatility and efficiency of the mould in thelab research, inspired by “Chinese movable type printing”, the mouldwas designed as a combination components that could replace thecavity core, as shown in Fig. 2. Four kinds of specimens were fabricatedand tested, i.e., tensile, bending, double cantilever beam (DCB) andend-notched flexure (ENF) specimens. For the DCB and ENF specimens,pre-crack was prepared by spraying release agent in the specified zoneof HSS surface before injection molding, rather than presetting poly-tetrafluoroethylene (PTFE) film due to the high temperature of injectionmolding would cause severe deformation of the PTFE film.

Fig. 3 shows the multi-scale images of pre-processed steel surface byhydraulic embossing technique. It is found that the periodical micro-structures (Fig. 3(a)–(c)) have appeared on the steel surface, whichimprove the surface roughness and contribute to form micro-mechan-ical interlocking (Fig. 3(d)). The length of the upper open rectangle is400–450 μm, and the depth is 150–200 μm. In injection molding pro-cess, the melt polymer flowed across concave objects and cooled down,and hence a series of micro-mechanical interlocking structures formedbetween steel and polymer.

2.3. Experimental procedures

In order to characterize the mechanical properties of MPH materialincluding modulus, strength and interface fracture toughness, the ten-sile, three-point bending, DCB and ENF tests were carried out. All ex-periments were conducted at room temperature environment on a MTSE45 machine with a load cell (range up to 50 kN). Loading velocity was2mm/min to assure a quasi-static condition. At least five valid

Fig. 1. Schematic illustration of fabrication process of MPH specimens.

Y. Zhang et al. Composite Structures 210 (2019) 70–81

71

experiments per variant were conducted to assure statistic robustness.

2.3.1. Tensile testThe dimension of tensile specimens (Table 1) was determined ac-

cording to ASTM standard D638. The MPH sample was made of 1-mm-thick steel sheet and 1-mm-thick polymer using the direct injectionmolding process. The GFRP sample was 2-mm thick fabricated by theinjection molding process. The steel sample was machined from a 1-mm-thick steel plate.

The tensile tests were conducted until total failure and full separa-tion of both clamped ends. The clamping bar was held constant for the

entire duration of the experiments. Additional cap strips were not re-quired since no slipping occurred and failure was rarely initiated withinthe clamps. In the stress-strain curve, the slope in the linear elasticregime was interpreted as the modulus, and the maximum stress wasdefined as the tensile strength of hybrids.

For a better understanding of the failure mechanisms of hybridsunder tensile loading, tensile experimental processes are recorded bythe high-speed photography. SEM images and optical micrographs areused to analyze the observed mechanical response to the micro-structure. The digital image correlation (DIC) method is used to mea-sure the full-field deformation and obtain the failure strain of MPH

Fig. 2. Mould used in fabrication and four kinds of samples.

Fig. 3. Morphologies of embossed HSS from (a) electron microscopy, (b) laser microscope and (c) three-dimensional imaging; (d) cross-sectional SEM image of aMPH sample with interfacial structures.

Y. Zhang et al. Composite Structures 210 (2019) 70–81

72

specimen.

2.3.2. Bending testBending specimens were determined according to ASTM D790, as

listed in Table 2. MPH sample consists of 1-mm-thick steel and 1-mm-thick GFRP. GFRP sample is 2-mm thick, and steel sample is 1-mmthick. The 3-point bending tests were conducted using a setup with twosupports and one crosshead with a radius of 5mm. The supports had aspan distance of 32mm.

For bending tests, the modulus Eb is assessed according to the re-spective standard for FRPs DIN EN ISO 14125 with Eq. (1) as

=El F

sbtΔ

4Δb03

3 (1)

where l0 is the free length between the supports. FΔ and sΔ are thechange of force and the change of displacement, respectively.

Bending strength σmax is defined as

=σ F lbt

32

·maxmax 0

2 (2)

where Fmax is the maximum force.

Table 1Dimensions of the constituent and hybrid material samples for tensile tests.

Materials Geometry Thickness (t)

MPH 2mm (1mm steel+ 1mmGFRP)

GFRP 2mmSteel 1 mm

Table 2Dimension of the constituent and hybrid material samples for bending tests.

Materials Geometry Thickness (t)

MPH 2mm (1mm steel+ 1mmGFRP)

GFRP 2mmSteel 1 mm

Fig. 4. Dimension of the DCB sample of steel/composite hybrids.

Fig. 5. Dimension of the ENF sample of steel/composite hybrids.

Y. Zhang et al. Composite Structures 210 (2019) 70–81

73

2.3.3. Interface fracture testsMode I and II cracks are the most common interface failure in

structures. DCB test is the representative method for characterization of

the mode I fracture toughness. Specimens for DCB tests were designedaccording to China aerospace standard HB 7402, consisting of 2-mm-thick GFRP phase and 1-mm-thick metal phase (Fig. 4). ENF test wasused to characterize mode II fracture toughness. Specimens for ENFtests were designed according to China aerospace standard HB 7403,consisting of 2-mm-thick GFRP phase and 1-mm-thick metal phase(Fig. 5).

In DCB and ENF tests, the matt white paints were sprayed on thespecimens to assist the observation of crack tip propagation.Additionally, a paper ruler was glued at 5mm internals under the bondline to facilitate the crack length measurement. A digital camera wasused to record the crack lengths per 5mm intervals during testing sothat the crack length parameters could be generated. Interface char-acterization tests were conducted until the crack propagated by 40mm.

Mode I fracture toughness (GIC) was calculated via Eq. (3):

= ×G Fδba2

10IC3

(3)

where a is the delamination length counting from the load/displace-ment application points; b is the specimen width; F is the critical load at

Fig. 6. FE models of (a) tensile specimen; (b) bending specimen; (c) DCB specimen; (d) ENF specimen.

Fig. 7. Bi-linear constitutive relationship in cohesive zone model.

Fig. 8. The tensile (a) force-displacement and (b) stress-strain curves of the hybrid and constitute materials.

Y. Zhang et al. Composite Structures 210 (2019) 70–81

74

the point of deviation from the linearity of the load-displacement curveand δ is the corresponding displacement.

Mode II fracture toughness (GIIC) is calculated using Eq. (4):

=+

×G Fδab L a

92 (2 3 )

10IIC2

3 33

(4)

where L2 is the span length.

2.4. Analytical prediction method

2.4.1. Tensile performance of hybrid materialThe metal polymer hybrid coupons can be regarded as a sandwich

structure composed of metal-interface-polymer three phases, thereforeits effective modulus can be derived based on the rule of mixture.Moreover, as reported in Ref. [24], the tensile curves of metal polymerhybrid materials show a three-stage feature: elastic deformation stagebefore steel yielding (Stage I), the stage from the yielding of the steel tothe fracture of composite (Stage II) and the stage from the fracture ofcomposite to the fracture of steel (Stage III). Based on these char-acteristics, an analytical model is proposed.

Based on rule of mixture, effective tensile modulus of MPH is cal-culated by Eq. (5):

= + +E E A E A E A A( · · · )/t MPH t HSS HSS t Interface Interface t GFRP GFRP MPH, , , , (5)

Fig. 9. Failure process of hybrid in tensile test: (a) initial loading; (b) interface failure; (c) composite failure; (d) steel failure.

Fig. 10. Fracture morphology of the steel: (a) longitudinal section, (b) transverse section and its local views at magnifications of (c) 300 and (d) 800.

Y. Zhang et al. Composite Structures 210 (2019) 70–81

75

where Et is the effective tensile modulus. A is the corresponding crosssection ratio, whose subscript indicates the type of material or phase.The composite with short fibers is assumed to be isotropic.

Assume that all materials have identical tensile strain. The equa-tions for tensile stress σMPH are different according to the deformationstage:

Before the steel yields (εy), average stress σMPH can be calculated byEq. (6):

= + + <σ ε E A E A E A A ε ε·( · · · )/ ,MPH t HSS HSS t Interface Interface t GFRP GFRP MPH y, , ,

(6)

For the strain from the yielding of the steel to the fracture of GFRP(ε f1), σMPH can be calculated as Eq. (7):

= + ⩽ <σ σ A ε E A K A ε ε ε( · · · · )/ ,MPH y HSS t GFRP GFRP MPH y f, 1 (7)

where σy is the yield stress of steel, and K is the strain hardeningcoefficient of composite.

For the strain from the fracture of composite (ε f1) to the fracture ofthe steel (ε f2), σMPH can be calculated as Eq. (8)

= ⩽ <σ σ A A ε ε ε( · )/ ,MPH y HSS MPH f f1 2 (8)

Therefore, the equations for tensile force Ft MPH, are as follows:

Fig. 11. Fracture morphology of the composite: (a) longitudinal section, (b) transverse section and its local views at magnifications of (c) 300 and (d) 800.

Fig. 12. The bending curves of hybrid and the constitute materials.

Fig. 13. The force-displacement curves of (a) DCB and (b) ENF specimens of hybrids.

Y. Zhang et al. Composite Structures 210 (2019) 70–81

76

=

⎧

⎨⎪

⎩⎪

<

+ ⩽ <

⩽ <

F

E S S

σ A E A K S S S

σ A S S S

·

· · · ·

·

t MPH

t MPHSl y

y HSSSl t GFRP GFRP y f

y HSS f f

,

,

,

0

0 1

1 2 (9)

where l0 is the gauge length; S is the displacement, the subscript y f, 1and f2 indicate the deformation stage as described in Eqs. (6)–(8).

2.4.2. Bending performance of hybrid materialThe effective bending modulus is expressed by Eq. (10):

= + +E E A E A E A A( · · · )/b MPH b HSS HSS b Interface Interface b GFRP GFRP MPH, , , , (10)

where Eb is the effective bending modulus. A is the corresponding cross

section ratio, and the subscript indicates the type of material or phase.Combined with Eq. (1), the equations for bending force Fb MPH, are

calculated as Eq. (11):

=

⎧

⎨

⎪⎪

⎩⎪⎪

<

+ ⩽ <

⩽ <

F

E S S S

E S E S K S S S

E S S S S

· ·

· · · · ·

· ·

b MPH

b t

l b MPH y

b t

l b HSS yb t

l b GFRP y f

b t

l b HSS y f f

,

4· ·,

4· ·,

4· ·,

4· ·,

MPH

HSS GFRP

HSS

3

3

3

3

3

3 13

3 1 2 (11)

where S is the displacement, and the subscript indicates the deforma-tion stage as described in Eqs. (6)–(8).

2.5. Numerical simulation

2.5.1. Finite element modellingIn order to evaluate the mechanical behavior of steel/composite

hybrids numerically, four kinds of FE models were established ac-cording to the experimental conditions in Abaqus/Explicit software, asshown in Fig. 6. The metal phase and composite phase were modelledusing C3D8R elements with 8-node linear brick, reduced integration

Fig. 14. Comparison of force-displacement curves between experiments and analytical results: (a) tension; (b) bending.

Fig. 15. Comparison between experiments and analytical results: (a) tensile modulus; (b) tensile strength at GFRP rupture; (c) bending modulus; (d) bending strengthat GFRP rupture.

Table 3Constitute parameters of PA66-GF35 composite and steel.

Materials n A (MPa)

PA66-GF35 0.96 6200DP 590+Z steel 0.28 912

Y. Zhang et al. Composite Structures 210 (2019) 70–81

77

and enhanced hourglass control. The interface between steel andcomposite was modelled using 8-node three-dimensional cohesiveCOH3D8 elements. Wherein, the supports and crosshead in the three-point bending and ENF models were simplified as rigid bodies. Themesh in the deformed zone was refined, and the minimum size of ele-ments was 1mm.

In the tensile and DCB models, the fixed surfaces were applied a fullconstraint at all degrees of freedom (DoFs). The moving surfaces wererealized by displacement boundary conditions. In the bending model,the rigid impactor was assigned a mass and an initial z-velocity whileconstrained at the remaining DoFs. Based on literature data [25], ageneral friction coefficient of 0.3 was applied to all surface pairings inthe contact domain.

2.5.2. Constitutive modelsFor steel and PA66-GF35, the segment of true stress-strain curves

after necking onset has to be modified for a better insight into thematerial local response of the specimen which is important in validatesimulation. Two essential steps similar as Ref. [26] are used in thispaper.

Step 1. Conduct tension test and process the strain hardening curveup to the necking point.

Step 2. Extrapolate the curve according to a hardening law. In thisstudy, Ludwik Law [27] is used:

= +σ σ A ε¯ ¯ ·(¯ )p n0 (12)

where, A and n are the parameters. With two constraints keepingcontinuity of the curve and the identical slope, only one of the twoparameters is independent.

Cohesive zone model is introduced to simulate the interface beha-vior, in which stress is expressed as a bi-linear function of the dis-placement [28,29], shown in Fig. 7. Fracture toughness is considered asthe most important factor in the damage condition analysis [30].Hence, in this paper, fracture toughnesses (GIC, GIIC) are identified fromDCB and ENF experiments. The initial stiffnesses (EI , EII) and maximumstresses (TI ,max, TII ,max) are identified from simulation-based inverseidentification methods.

3. Results and discussion

3.1. Experimental results

3.1.1. Tensile propertiesFig. 8(a) shows the force-displacement curves of high strength steel/

composite hybrid, pure steel and pure composite under tensile loading.Since the cross-section areas are different from individual specimengeometries, the force level of the composite has been scaled in order toprovide comparability. The force of hybrid material is carried by steel

Fig. 16. Comparison between experiment and simulation under tensile loading: (a) PA66-GF35 composite; (b) steel.

Table 4Cohesive zone properties of the interface in hybrid material.

Parameters EI(GPa)

EII(GPa)

EIII(GPa)

GIC(kJ/m2)

GIIC(kJ/m2)

GIIIC(kJ/m2)

TI,max(MPa)

TII,max(MPa)

TIII,max(MPa)

Values 1.5 0.75 0.75 0.17 0.45 0.45 11.4 10.51 10.51

Fig. 17. Influence of the cohesive zone parameters on (a) DCB and (b) ENF numerical results.

Y. Zhang et al. Composite Structures 210 (2019) 70–81

78

and composite simultaneously. Fig. 8(b) shows the stress-strain curvesof hybrid material and its constitute materials. PA66-GF35 compositespecimen shows a typical brittle behavior, i.e., stress increasing nearlymonotonously, and then, followed by a brittle fracture. Steel specimenhas a power relationship between stress and strain after yielding. Thehybrid material exhibits a complex piece-wise increase of stress withincreasing strain. At the initial stage, the load was shared by steel andcomposite, and stress increases nearly proportionally with the in-creasing strain until the steel yield. At the second stage, stress increasesstably with increasing strain until the fracture of composite.

Subsequently, the load was carried solely by the steel, so the curve ofhybrid material is consistent with that of steel.

Fig. 9 shows the failure process of hybrid material in the tensileexperiment at lateral direction. It is found that the initial crack occursin the interface (Fig. 9(b)), rising to the loading direction in parallel,followed by composite phase failure (Fig. 9(c)) and steel phase failure(Fig. 9(d)), which are consistent with the force response in Fig. 8.

Fracture surfaces for steel and composite observed by SEM areshown in Figs. 10 and 11. Fracture surface of HSS is about 45° withrespect to the loading direction (Fig. 10). Elongation at fracture is ap-proximately 5.8 mm and proportion in reduction in area is approxi-mately 47%. The higher magnifications in Fig. 10(c) and (d) reveal thatsome macroscopic voids are intermingled with fine microscopic voidson the fracture surface. Therefore, steel exhibits a typical ductilefailure.

Fracture surface of PA66-GF35 composite is perpendicular to theloading direction (Fig. 11). Elongation at fracture is approximately1.9 mm and proportion in reduction in area is approximately 8.5%,both of which are much lower than steel. Fibers stretched out uni-directional on the fracture surface in the higher magnification picturesas shown in Fig. 11(c) and (d). The composite exhibits brittle matrixfracture and fiber pull-out failure.

3.1.2. Bending propertiesFig. 12 depicts the force-displacement curves of hybrid material

compared to pure steel and pure composite under bending loads. Sincecomposite specimens and steel specimens have different cross sections,the force level of composite has been scaled in order to provide com-parability. The force carried by hybrid material is contributed by steeland composite simultaneously PA66-GF35 composite specimen shows atypical brittle behavior, and steel specimen exhibits ductile metal be-havior. Hybrid material exhibits a complex piece-wise increase of stresswith increasing strain.

3.1.3. Interface propertiesThe force-displacement curves of DCB and ENF specimens are

shown in Fig. 13(a) and (b). The curves overlap each other well, in-dicating the test results are reliable. For DCB experiments, force-dis-placement curves show obvious bi-linear response. Force increasesproportionally with the increasing displacement to the maximum load,after that force decreases linearly with some fluctuation. These fluc-tuations are caused by the fracture resistance [31,32]. According to Eq.(3) and the relationship in Fig. 7, Mode I fracture toughness (GIC) is0.17 kJ/m2. For ENF experiments, the force-displacement curves alsoexhibit bi-linear response and mode II fracture toughness is 0.45 kJ/m2.

3.2. Analytical prediction

Fig. 14(a) compares the experimental tensile force-displacementcurve with the analytically predicted result. Fig. 14(b) compares theexperimental bending force-displacement curve with the analyticallypredicted result. Furthermore, Fig. 15 compares the results of themodulus and strength at GFRP rupture. It is found that the analyticalresults agree well with the experiments.

The tensile strength and bending strength are affected significantlyby the properties of steel phase. The calculated values of bendingmodulus of hybrids are higher than the experimental results(Fig. 15(c)), and the error is approximately 22.22%. The errors mayresult from the cross section differences of pure steel specimen and purecomposite specimen. Hence, a further improvement of analysis modelsare needed to characterize the variation of corresponding cross sectionratio for steel, interface and composite.

Fig. 18. Comparison between simulation, experiment and DIC results of hybridmaterial under tensile loading: (a) force-displacement curves; (b) strain dis-tribution of simulation; (c) DIC strain contour of experiment.

Y. Zhang et al. Composite Structures 210 (2019) 70–81

79

3.3. Numerical simulation

3.3.1. Material model validationThe constitutive relations of steel and composite in hybrid material

are obtained by Eq. (12). The parameters n and A are identified andlisted in Table 3.

Fig. 16 compares the experimental tensile stress-strain curves withthe simulation results of PA66-GF35 composite and steel, respectively.The agreement is very good. Therefore, the constitute models are va-lidated.

Based on the experiments in Section 2.3.3, the cohesive zoneproperties of the interface are obtained, as listed in Table 4. The in-fluence of each parameter on simulation results is assessed by DCB andENF finite element models. For DCB case (Fig. 17(a)), either stiffnessvariation in the range of 0.5–2 times or maximum stress increased by20% would have relatively small influence on the force-displacementcurve response. Nevertheless, the fracture toughness increased by 20%,which affects the maximum force more significantly. Similar results areobserved in ENF case (Fig. 17(b)) in the aforementioned variationrange.. Therefore, it is concluded that the numerical results are notsensitive to changes of stiffness and maximum stress. Similar resultswere reported by Zou et al. [29].

3.3.2. Tensile simulationFig. 18 shows the comparison of tensile mechanical response and

failure modes between simulation, experiment and DIC results for steel/composite hybrid material. The tensile modulus and peak force valuesof simulation show good agreement with experimental results(Fig. 18(a)). Furthermore, the strain distributions in three re-presentative stages of simulation (Fig. 18(b)) agree well with DIC re-sults (Fig. 18(c)): When t= t1, the steel yields, and the strain is ap-proximately 0.002; When t= t2, the PA66-GF35 composite fractures,and the strain is approximately 0.03; When t= t3, the steel fractures,and the fracture strain is approximately 0.3. The hybrid material showsa typical progressive failure process. The simulation method developedin this paper can be used to predict the mechanical behavior of hybridmaterial cost-effectively and accurately.

3.3.3. Bending simulationFig. 19 shows the comparison of bending mechanical response and

failure modes between the simulation and the experiment. The force-displacement curves agree well. Moreover, the FE model can predict thebending modulus, maximum force, delamination and fracture very ac-curately. As shown in Fig. 19(b), although the polymer is broken intotwo sections, the broken composite and the deformed steel are stillintegrated. It indicates that a strong bonding between the compositeand the steel by injection molding process has been formed.

4. Conclusions

In this study, high strength steel and PA66-GF35 composite wereintegrated by direct injection molding process. The mechanical prop-erties of steel/composite hybrid material were investigated using ten-sile tests and three-point bending experiments. The properties of theinterface between steel and composite were measured for the first timeusing double cantilever beam (DCB) and end-notched flexure (ENF)tests. Analytical models based on the rule of mixture were developed toexplain and support the experimental results. Moreover, a finite ele-ment simulation method to predict the mechanical behavior was pro-posed and validated. The following major conclusions can be drawnfrom the present study:

(1) A new direct injection molding process combining the embossedtechnique and direct injection molding was validated by fabricationdifferent kinds of steel/composite hybrid specimens. This processpossesses the advantages of low cost, high efficiency and goodcompatibility with existing automotive production line.

(2) The mechanical properties of hybrid material resembles a sum ofthe steel and composite. The steel dominates the load-bearing ca-pacity, and the interface between steel and composite dominatesthe failure process. Moreover, the contribution of each constituentphase is affected by loading condition.

(3) The rule of mixture can be used to predict the effective modulus andstrength of hybrid materials composed of metal and polymer. Thenumerical simulation approach with cohesive zone model of theinterface can predict the mechanical behavior of hybrid materialscost-effectively and accurately.

(4) A comprehensive utilization of the analytical models and FEmethod proposed in this study can contribute to the performanceevaluation and structural design of hybrid materials efficiently andaccurately, which may serve as a foundation for the engineeringapplication of metal-polymer load-bearing components.

Acknowledgments

This work was supported by the National Natural ScienceFoundation of China (No. 51575023 and No. U1664250), the NationalKey Research and Development Program of China (No.2016YFB0101606) and the Open Foundation of Henan Key Laboratoryof Underwater Intelligent Equipment (No. KL03B1805).

References

[1] Dlugosch M, Volk M, Lukaszewicz D, et al. Suitability assessments for advancedcomposite-metal hybrid material systems in automotive crash structural applica-tions. Int J Automot Compos 2017;3(1):14–28.

[2] Grujicic M. Injection over molding of polymer-metal hybrid structures. Am J SciTechnol 2014;1(4):168–81.

Fig. 19. Comparison between experiment and simulation results of hybrid material under bending loading: (a) force-displacement curves; (b) failure modes.

Y. Zhang et al. Composite Structures 210 (2019) 70–81

80

[3] Amancio-Filho ST, Dos Santos JF. Joining of polymers and polymer-metal hybridstructures: recent developments and trends. Polym Eng Sci 2009;49(8):1461–76.

[4] Dlugosch M, Lukaszewicz D, Fritsch J, et al. Experimental investigation of hybridmaterial systems consisting of advanced composites and sheet metal. Compos Struct2016;152:840–9.

[5] Giampaoli M, Terlizzi V, Rossi M, et al. Mechanical performances of GFRP-steelspecimens bonded with different epoxy adhesives, before and after the agingtreatments. Compos Struct 2017;171:145–57.

[6] Ulger T, Okeil AM. Analysis of thin-walled steel beams retrofitted by bonding GFRPstiffeners: Numerical model and investigation of design parameters. Eng Struct2017;153:166–79.

[7] Zoellner OJ, Evans JA. Plastic-metal hybrid-A new development in the injectionmolding technology. In: ANTEC 2002 Annual Technical Conference, Sanfrancisco,CA, 2002;3:2966-2969.

[8] Fortunato A, Cuccolini G, Ascari A, et al. Hybrid metal-plastic joining by means oflaser. Int J Mater Form 2010;3(1):1131–4.

[9] Roesner A, Scheik S, Olowinsky A, Gillner A, Reisgen U, Schleser M. Laser assistedjoining of plastic metal hybrids. Physics Procedia 2011;12(1):370–7.

[10] Grujicic M, Arakere G, Pisu P, et al. Application of topology, size and shape opti-mization methods in polymer metal hybrid structural lightweight engineering.Multidiscip Model Mater Struct 2008;4(4):305–30.

[11] Grujicic M, Pandurangan B, Bell WC, et al. A computational analysis and suitabilityassessment of cold-gas dynamic spraying of glass-fiber-reinforced poly-amide 6 foruse in direct-adhesion polymer metal hybrid components. Appl Surf Sci2008;254(7):2136–45.

[12] Li X, Gong N, Yang C, et al. Aluminum/polypropylene composites produced throughinjection molding. J Mater Process Technol 2018;255:635–43.

[13] Huang BC, Sun LY, Li LJ, et al. Experimental investigation of the strength ofpolymer-steel direct adhesion (PSDA) joints with micro-structures ablated by laser.J Mater Process Technol 2017;249:407–14.

[14] Sperandio C, Bardon J, Laachachi A, et al. Influence of plasma surface treatment onbond strength behaviour of an adhesively bonded aluminium-epoxy system. Int JAdhes Adhes 2010;30(8):720–8.

[15] Kimura F, Kadoya S, Kajihara Y. Effects of molding conditions on injection moldeddirect joining using a metal with nano-structured surface. Precis Eng2016;45:203–8.

[16] Izadi O, Mosaddegh P, Silani M, et al. An experimental study on mechanicalproperties of a novel hybrid metal-polymer joining technology based on a reactionbetween isocyanate and hydroxyl groups. J Manuf Processes 2017;30:217–25.

[17] Yeh RY, Hsu RQ. Improving the adhesion of plastic/metal direct bonding by

injection moulding using surface modifications. Adv Mater Process Technol2016;2(1):21–30.

[18] Costas M, Díaz J, Romera LE, et al. Static and dynamic axial crushing analysis of carfrontal impact hybrid absorbers. Int J Impact Eng 2013;62:166–81.

[19] Kim HC, Shin DK, Lee JJ, et al. Crashworthiness of aluminum/CFRP square hollowsection beam under axial impact loading for crash box application. Compos Struct2014;112:1–10.

[20] Shin DK, Kim HC, Lee JJ. Numerical analysis of the damage behavior of an alu-minum/CFRP hybrid beam under three point bending. Compos B 2014;56:397–407.

[21] Dlugosch M, Fritsch J, Lukaszewicz D, et al. Experimental investigation and eva-luation of numerical modeling approaches for hybrid-FRP-steel sections under im-pact loading for the application in automotive crash-structures. Compos Struct2017;174:338–47.

[22] Hopmann C, Schöngart M, Weber M, et al. Crash simulation of hybrid structuresconsidering the stress and strain rate dependent material behavior of thermoplasticmaterials. AIP conference proceedings. AIP Publishing; 2015. p. 050007.

[23] Grujicic M, Sellappan V, Omar MA, et al. An overview of the polymer-to-metaldirect-adhesion hybrid technologies for load-bearing automotive components. JMater Process Technol 2008;197(1–3):363–73.

[24] Wu G, Wu ZS, Luo YB, et al. Mechanical properties of steel-FRP composite bar underuniaxial and cyclic tensile loads. J Mater Civ Eng 2010;22(10):1056–66.

[25] Reuter C, Tröster T. Crashworthiness and numerical simulation of hybrid alumi-nium-CFRP tubes under axial impact. Thin Walled Struct 2017;117:1–9.

[26] Dunand M, Mohr D. Hybrid experimental-numerical analysis of basic ductile frac-ture experiments for sheet metals. Int J Solids Struct 2010;47(9):1130–43.

[27] Ludwik P. Elemente der technologischen Mechanik. Berlin: Springer; 1909.[28] Gupta P, Pal S, Yedla N. Molecular dynamics based cohesive zone modeling of Al

(metal)–Cu50Zr50 (metallic glass) interfacial mechanical behavior and investiga-tion of dissipative mechanisms. Mater Des 2016;105:41–50.

[29] Zou P, Li Y, Zhang K, et al. Mode I delamination mechanism analysis on CFRPinterference-fit during the installation process. Mater Des 2017;116:268–77.

[30] Jousset P, Rachik M. Comparison and evaluation of two types of cohesive zonemodels for the finite element analysis of fracture propagation in industrial bondedstructures. Eng Fract Mech 2014;132:48–69.

[31] Sarrado C, Turon A, Renart J, et al. An experimental data reduction method for theMixed Mode Bending test based on the J-integral approach. Compos Sci Technol2015;117:85–91.

[32] Matthews T, Ali M, Paris AJ. Finite element analysis for large displacement J-in-tegral test method for Mode I interlaminar fracture in composite materials. FiniteElem Anal Des 2014;83:43–8.

Y. Zhang et al. Composite Structures 210 (2019) 70–81

81