Chapter 12: Failure of polymer matrix composites (PMCs) in automotive and transportation applications – Failure Mechanisms in Polymer Matrix Composites


Failure of polymer matrix composites (PMCs) in automotive and transportation applications

P.K. Mallick,     University of Michigan-Dearborn, USA


With increasing need for fuel economy improvement, there is a great deal of interest in using polymer matrix composites to produce lightweight vehicles. This chapter provides an overview of the failure modes of two types of polymer matrix composites for automotive and road transportation applications. One of them is a random fiber sheet molding compound (SMC) composite which has many current applications in automobiles, light trucks and heavy trucks. The second type is the carbon and glass fiber reinforced polymer matrix composites that may find future applications in lightweight automotive structures to improve their crashworthiness.

Key words












12.1 Introduction

The main polymer matrix composites used today in the automotive industry are randomly oriented discontinuous E-glass fiber reinforced composites. E-glass fiber is used instead of carbon fibers because of its significantly lower cost. The polymer matrix can be either a thermoplastic, such as polypropylene, polycarbonate and polyamide 6,6, or a thermoset resin, such an unsaturated polyester and a vinyl ester. Epoxy, the most common thermoset polymer used as matrix in aerospace composites, is not commonly used in the automotive industry. The reason for not using epoxy is its longer cure time and higher cost than polyesters or vinyl esters. Thermoplastics are used mostly for functional, interior and semi-structural components. Thermosets are used for a variety of body panel, body structure and suspension applications. Injection molding, compression molding or injection-compression molding are the common manufacturing processes for thermoplastic matrix composites. compression molding, reaction injection molding and resin transfer molding are the common manufacturing processes used for thermoset matrix composites.

Carbon fiber reinforced polymer matrix composites have much higher modulus-to-density ratio and fatigue resistance than glass fiber reinforced polymer matrix composites and are more desirable for their weight saving potential in structural applications, such as door beams, front rails and springs, but they are significantly more expensive than glass fiber composites. Continuous carbon fiber reinforced epoxy laminates used in the aerospace industry are manufactured using hand or automated lay-up and autoclave curing. The processing time for lay-up followed by autoclave curing is much too long to be used for mass production of components in the automotive industry. For these two reasons, carbon fiber reinforced polymer matrix composites have so far found very limited application in the automotive industry, and that too, in low production volume, niche automobiles with relatively high price.

12.2 Polymer matrix composites (PMCs) used in automotive and road transportation applications

Polymer matrix composites (PMCs) in automotive and road transportation applications can be divided into three broad categories (Mallick, 2010; Brooks, 2000).

1. Functional, interior and under-the-hood applications, such as gears, brackets, interior door panels, instrument panels, intake manifolds, rocker covers. PMCs used in this category of applications include short or long fiber reinforced thermoplastics and bulk molding compounds (BMCs). The manufacturing processes include injection molding and compression molding.

2. Semi-structural body applications, such as hoods, fenders, seat backs, grille opening panels. PMCs used for these applications are sheet molding compounds (SMCs), long fiber reinforced thermoplastics, glass mat reinforced thermoplastics and chopped or continuous glass strand reinforced polyurethane. compression molding, injection molding and reaction injection molding are the principal manufacturing processes in this category.

3. Structural body, chassis and suspension applications, such as cross members, suspension arms, drive shafts, door intrusion beams, front rails and springs. PMCs considered for these applications are continuous fiber reinforced thermosets. Carbon fibers and epoxy are the major fiber and matrix of choice for these applications. The manufacturing processes for this category of applications are resin transfer molding, compression molding and filament winding.

12.3 Scope of the chapter

This chapter will consider the failure modes of two types of polymer matrix composites used in automotive and road transportation applications. The first type is the sheet molding compound composite (SMC), which contains randomly oriented discontinuous E-glass fibers in a thermoset matrix. The manufacturing technique used for making SMC parts is compression molding with a production cycle time in the range of 1–5 minutes. SMC is currently used in a variety of structural and semi-structural automotive applications, such as hoods, deck lids, radiator supports, grille opening panels, front and rear fenders, bumper beams, rocker covers and seat structures. They are also used in light and heavy trucks for a variety of components, such as pickup boxes, door panels, bumper beams, roof panels, air deflectors and spoilers. The reasons for such a wide range of applications for E-glass fiber reinforced SMC are:

• low cost

• parts consolidation

• low tooling cost and

• relatively fast manufacturing cycle time.

The second type of automotive composites considered in this chapter is the continuous carbon and glass fiber composites considered for crush resistant tubular structures, such as front rails. In recent years, there has been a significant amount of interest in such composite structures because of their weight saving potential and relatively high crash energy absorption per unit mass. The manufacturing methods for making crush resistant composite structures include filament winding, pultrusion and resin transfer molding.

12.4 Common in-service conditions causing failure

The in-service conditions in which automobiles operate can be very severe. The body structure and many body and chassis components of an automobile are subjected to a combination of bending and twisting loads. At the design stage, static bending stiffness, torsional stiffness and resonant frequencies of the body-in-white are determined either by finite element analysis or by laboratory tests. During vehicle operation, many body-in-white and chassis components are subjected to random fatigue loading due to road surface conditions, which include bumps, irregularities, undulations and pot holes. Fatigue failure is the most common form of failure that occurs in these components. Under impact conditions in a high speed collision, the vehicle body is subjected to a combination of compressive, shear and bending impulse loads at very high strain rates, which lead to global buckling, crumpling, tearing and bending failures. For body panels, the most important design parameter is their bending stiffness, which depends on the modulus of the material, panel thickness and their attachments. The common failure modes for steel body panels are dent formation and scratches due to low energy impact caused by hail storms, tool drops, stone chip impacts and accidental bumps. For a composite body panel, the failure under such low energy impacts may range from non-visible internal damage to complete fracture. The non-visible internal damage may include sub-surface delaminations, matrix cracking and fiber fracture, which will reduce the local stiffness and strength of the material in the damaged area.

The environment in which an automobile operates can also be severe. The operating temperature depends on the climatic condition in which the vehicle is operated, which may vary from 50 °c or higher in a desert to −20 °c or lower in arctic countries. If the component is near the engine or near the exhaust system, it may experience temperatures higher than 100 °c. The vehicle body may also be exposed to various levels of humidity and moisture. During winter months, it may be exposed to salty road surface. There may also be exposure to automotive fluids and chemicals, such as gasoline, motor oil, engine coolants, windshield washing fluids and various cleaning agents. In a severe high speed accident, the vehicle may catch fire, which can be a major concern for carbon fiber reinforced polymer composite applications due to the release of respirable carbon fibers in the atmosphere.

12.5 Sheet molding compound (SMC) composites

The SMC composites used in the automotive industry are randomly oriented discontinuous E-glass fiber composites in a thermoset matrix. The randomly oriented discontinuous fiber SMC, or simply random fiber SMC (Fig. 12.1(a)), is designated as SMC-Rxx, where ‘xx’ stands for the weight fraction of fibers in the material. For example, SMC-R18 is a random fiber SMC containing 18 wt% of fibers. SMC composites containing unidirectional continuous fibers or a combination of unidirectional continuous and randomly oriented discontinuous fibers (Fig. 12.1(b)) are also available, but their application has not been widespread. If the SMC contains unidirectional continuous fibers as well as randomly oriented discontinuous fibers, its designation is SMC-CxxRyy, where ‘xx’ stands for the weight fraction of continuous fibers and ‘yy’ the weight fraction of random fibers.

12.1 Sheet molding compound (SMC) composites: (a) SMC-R, containing randomly oriented discontinuous fibers and (b) SMC-CR, containing unidirectional continuous fibers and randomly oriented discontinuous fibers.

The discontinuous fibers in SMC-R are chopped E-glass fibers that are cut to the desired length from continuous fiber rovings and are randomly dispersed on a thin layer of resin formulation containing the matrix resin and several other additives (Mallick and Newman, 1990). Some of these additives act as processing aids and others are used to control shrinkage and surface quality. The matrix used in automotive SMC-R is either an unsaturated polyester resin or a vinyl ester resin. Polyester is more commonly used, but vinyl ester is preferred for parts which may be exposed to dynamic loading. In general, vinyl ester SMC composites show better fatigue performance than polyester SMC composites.

SMCs are produced in continuous ‘ready-to-mold’ sheet form and stored in an environmental chamber until they are ready to be processed to make parts. compression molding in a matched metal mold is the manufacturing process used for making SMC parts. The SMC sheet at room temperature is very pliable and can be easily cut into small pieces, stacked in layers and then placed in the mold. The average fiber length in an SMC is typically 25 mm. The fiber weight fraction may vary from 15 to 65% of E-glass fibers, depending on the modulus and strength required for a particular application. In terms of fiber volume fraction, this corresponds to approximately 7–45% of E-glass fibers. The higher the fiber content in SMC, the higher are its modulus and strength (Table 12.1). The SMC-R also contains low-cost fillers, typically calcium carbonate, mainly to reduce the material cost, but they also help control viscosity during processing, reduce molding shrinkage and increase the modulus of the matrix. At high fiber content, the amount of fillers is reduced; otherwise, it becomes difficult to mold the SMC because of increased viscosity. The details of sheet molding compounds and compression molding process can be found in several references (Mallick and Newman, 1990; Davis et al., 2003).

Table 12.1

Properties of various SMC composites

aSMC-C20R30 contains 20 wt% unidirectional continuous fibers and 30 wt% discontnuous random fibers. L and T are longitudinal and transverse directions, respectively.

12.5.1 Tensile characteristics

Figure 12.2 shows the tensile stress–strain diagrams of two random fiber SMCs, namely SMC-R25 containing 25 wt% E-glass fibers and SMC-R65 containing 65 wt% E-glass fibers. SMC-R25 also contains fillers, while SMC-R65 does not contain any filler. In both cases, the tensile stress–strain diagrams exhibit a break or knee at strain values in the range of 0.25–0.3%. The failure strain for both SMCs is less than 2%. For SMC-R25, the stress–strain diagram is slightly non-linear below the knee, but becomes linear at strain values higher than 0.5%. For SMC-R65, there is less non-linearity below the knee and the stress–strain diagram is also linear at strain values higher than 0.5%. Above the knee, the slopes of the stress–strain diagrams for both SMCs are lower than the initial slopes.

12.2 Tensile stress–strain diagrams of SMC-R25 and SMC-R65 composites, both containing randomly oriented 25 mm long E-glass fibers.

The presence of non-linearity before the knee is an indication of damage formation even at low strain levels (Watanabe and Yasuda, 1982). The damage is in the form of debonding at the fiber–matrix interface and as the knee is approached, cracks start to form in the matrix; however, these cracks are prevented from growing by the fibers, and in the case of SMC-R25, also by the fillers. In other words, small matrix cracks or debonds extend up to the fiber or filler surfaces and are interrupted. As the strain increases with continued loading, multiple matrix cracks are formed at random locations throughout the material. This may also be accompanied by fiber fracture, fiber–matrix interface debonding, fiber pullout and fiber bridging. Because of multiple damage formation, the modulus of the material is decreased above the knee. The amount of damage continues to increase until complete fracture as several microcracks join together to form a major crack which propagates through the width of the specimen.

It should be noted that random fiber SMC composites show planar isotropic behavior when they are molded under controlled flow conditions that produce two-dimensional random fiber orientation in the molded plate. However, under production molding conditions in which there is considerable amount of material flow in the mold, there may be preferential orientation fibers in the flow direction, leading to lower modulus and strength in the transverse-to-flow direction relative to the flow direction. Because of such direction-dependent properties, it can no longer be treated as a planar isotropic material. Preferential orientation of fibers also occurs at the knit line of a molded SMC-R plate, which is formed by the joining of two or more flow fronts during the molding operation (Fig. 12.3). The tensile strength of the material is lower in a direction transverse to the knit line, and therefore, failure may initiate at the knit line, particularly when there is a significant stress component in the transverse direction.

12.3 Knit line formed at the joining of two flow fronts in a compression molded SMC-R plate.

If the SMC contains a combination of unidirectional continuous fibers and randomly oriented discontinuous fibers, its strength and modulus are significantly enhanced in the longitudinal (L) direction of the continuous fibers. The SMC-C20R30 composite shown in Table 12.1 has the same total fiber content as the SMC-R50 composite, but 20 wt% of fibers in SMC-C20R30 are unidirectional continuous fibers and 30 wt% are 25 mm long randomly oriented discontinuous fibers. Its tensile modulus and strength in the continuous fiber direction (L-direction) are 21.4 GPa and 289 MPa, respectively. Both are significantly higher than the tensile modulus and strength of SMC-R50. However, modulus and strength of SMC-c20R30 transverse (T) to the direction of continuous fibers are much lower than the corresponding values in the L-direction. Thus, the material is non-isotropic. SMC-C20R30 and similar SMCs are useful for beam-type applications, such as cross members and bumper beams in which there is a predominant stress direction along their lengths.

12.5.2 Tensile strength variation

Under normal molding conditions, SMC properties can show significant variation in tensile strength. This is because of numerous internal defects that may be present in a molded SMC part. Among them are dry fibers, non-uniform fiber and filler distributions, resin-rich areas, voids, fiber length variation, fiber bundle separation, etc. The tensile strength variation is modeled using a two-parameter Weibull distribution (Shirrell, 1983) given by the following equation:


where F(σ) = distribution of tensile strength σ, = location parameter, and α = shape parameter. The location and shape parameters of several SMC-R composites are given in Table 12.2. Note that a low value of the shape parameter α is an indication of greater scatter in the tensile strength data.

Table 12.2

Weibull parameters for strengths of SMC-R28 and SMC-R50 composites

12.5.3 Effect of stress concentration

The effect of stress concentration on the tensile strength of SMC-R composites has been studied by several investigators using a drilled round hole (Fig. 12.4) as the source of stress concentration. In general, SMC-R has a relatively low sensitivity to the presence of holes. This is because microcracks formed at the naturally occurring flaws in the material, such as fiber ends and voids, at low tensile loads mitigate the adverse effect of stress concentration. Mallick (1988) and Shirrell and Onachuk (1986) have shown that the Whitney–Nuismer point-stress failure criterion (Whitney and Nuismer, 1974) can be used to predict the tensile strength of SMC-R. For a planar isotropic SMC-R composite plate containing a round hole at its center, the failure strength can be written as:

12.4 Damage forming at the edges of a hole in an SMC-R plate under tensile loading.


where Snet = net tensile strength of an SMC-R plate in presence of a central round hole, So = tensile strength of the SMC-R without any stress concentration, λ = R/(R + Ro), R = hole radius, which is assumed to be much smaller than the width of the plate, and Ro = a characteristic distance from the hole edge (which is determined experimentally) whose value depends on the fiber content of SMC-R.

12.5.4 Fracture

The fracture process in an SMC-R in the presence of a finite length crack starts with the physical appearance of matrix microcracks at several locations in the material. Some of these microcracks may appear away from the crack ends. The growth of these microcracks is controlled by the discontinuous fibers and fillers dispersed in the matrix. Fiber matrix interface debonding and fiber failure also occur. Eventually many of the microcracks join to form a long crack, ultimately leading to fracture by separation. However, the crack propagation takes place in a zigzag fashion as it propagates around or through the randomly oriented fibers. Fracture toughness values of three SMC-R composites as determined by Sun and Sierakowski (1980) are given in Table 12.3. It demonstrates that fracture toughness of SMC-R composites increases with increasing fiber content.

Table 12.3

Mode I fracture toughness of SMC-R composites

(from Sun and Sierakowski, 1980)

12.5.5 Fatigue

Most of the fatigue data on SMC-R are based on tension–tension stress-controlled fatigue tests. Figure 12.5 shows the S-N diagrams of SMC-R25 and SMC-R65 composites at three different temperatures. No fatigue endurance limit can be observed in these diagrams; instead, the number of cycles to failure is increased as the maximum stress amplitude is decreased. Table 12.4 gives the fatigue strength of three SMC-R composites at 106 cycles. As the data in Table 12.4 shows, the fatigue strength increases with increasing fiber content, but the ratio of fatigue strength and static tensile strength decreases with increasing fiber content.

Table 12.4

Fatigue strength at 106 cycles of SMC-R composites (at 23 °C, R = 0.05)

Material Fatigue strength (MPa) at 106 cycles Ratio of fatigue strength and tensile strength
SMC-R25 40 0.49
SMC-R50 63 0.38
SMC-R65 70 0.31
SMC-C20R30 130 (L), 44 (T) 0.45 (L), 0.52 (T)

12.5 Fatigue S-N diagrams of SMC-R25 and SMC-R65 composites (frequency of cycling = 5 Hz, R = 0.5).

Fatigue damage in SMC-R composites appears in the form of matrix microcracks and fiber–matrix interface debonding at a very low number of cycles and the number and size of these microcracks grow as the cycling is continued. Several microcracks eventually join to form numerous visible cracks; however, failure by separation may not occur for a long period of cycling after the appearance of the visible cracks. As a consequence of microcrack development and coalescence, the modulus of the material decreases with increasing number of cycles.

Mallick (1981) has shown that crack formation in SMC-R depends on the matrix type. For the same number of cycles and at the same maximum stress level, the visible surface crack density in the SMC-R with a polyester matrix is much higher than that with a vinyl ester matrix. As a result, the polyester SMC-R shows a much greater reduction in modulus than the vinyl ester SMC-R. Furthermore, the post-fatigue residual strength of the polyester SMC-R is lower than that of the vinyl ester SMC-R.

Fatigue damage in SMC-R appears in the forms of matrix microcracks, fiber-matrix interfacial debonds and fiber fracture (Wang et al., 1986). Wang et al. (1983a) proposed the following fatigue damage growth model for SMC-R composites:


where D is the cumulative damage after N fatigue cycles, B is a constant and A is a function of the damage state that depends on the loading history, stress amplitude, mode of loading, temperature, etc. The damage parameter D is expressed in terms of the modulus values as:


where Eo is the initial modulus and E is the modulus after N cycles of fatigue cycling. Thus, (Eo − E) in Equation [12.5] represents the reduction in modulus due to fatigue damage accumulation in N cycles. The modulus ratio E/Eo of an SMC-R as a function of number of load cycles in tension–tension fatigue tests is shown in Fig. 12.6.

12.6 Modulus degradation during fatigue cycling of an SMC-R at four different maximum fatigue stress levels (1 represents the lowest maximum fatigue stress and 4 represents the highest maximum fatigue stress).

Wang et al. (1983b) also noted that fatigue crack growth rate in SMC-R under Mode I fatigue loading can be represented by the Paris power law equation which is commonly used for metals:


where, a = instantaneous crack length, da/dN = rate of steady crack growth with increasing number of cycles, KI = Mode I stress intensity factor, and C, m = material constants.

Since crack growth in SMC-R may not take place in a self-similar manner, the crack length a in Equation [12.5] includes the size of the damage zone in front of an embedded crack. For an SMC-R50, Wang et al. (1983b) determined the material constants C and m to be 7.58 × 10−12 mm/cycle and 9.65, respectively. It is noteworthy that the value of m is considerably higher than that for metals, which is between 2 and 4. This means that the damage accumulation in SMC-R occurs at a higher rate than that in metals. The high value of m for SMC-R50 was attributed to the high density of crack initiation sites both at the crack tip as well as away from the crack tip.

12.5.6 Impact

Low energy impact on SMC-R panels may not cause a complete structural failure or penetration of the impactor through the panel thickness, but, as shown in Table 12.5, there may be extensive surface and internal damage that can cause reduction in modulus and strength of the material (Chaturvedi and Sierakowski, 1983). Low energy impact in in-service conditions may be caused by accidental bump with another vehicle, stone chip impact, hail storm, etc.

Table 12.5

Post-impact residual tensile properties of SMC-R composites

(from Chaturvedi and Sierkowski, 1983)

Tests conducted with 45.4 g, 22.2 mm diameter steel balls fired at SMC-R35 plates at four different impact speeds ranging from 48 to 96 km/h showed local indentation or denting and a ring of disconnected microcracks surrounding the dent on the impacted surface (Khetan and Chang, 1983). A damage zone consisting of a cluster of dense microcracks was visible on the back surface (Fig. 12.7). Both front and back surface damage increased with increasing impact speed, but they were greater in the back surface than in the front surface at all impact speeds. Matrix cracking and fiber–matrix interface debonding were the principal micro-failure modes observed in these damage zones.

12.7 Damage on the back side of an impacted SMC-R panel (Kau, 1990).

When perforation occurred, the front and back surfaces of the panel also had different failure appearances. Examination of SMC-R25 panels after puncture tests at a constant speed of 2.5 m/s (Kau, 1990) revealed that the damage zone on the back side of the panel was larger than that on the front side. There was evidence of delamination on both front and back surfaces surrounding the impacted area. There were also numerous radial cracks emanating from the damage zones, but these cracks were longer and deeper on the back surface than on the front surface. Matrix cracking, fiber–matrix interface debonding, fiber fracture and fiber pull-out were the principal micro-failure modes observed in the impact damage zones.

12.5.7 Environmental effects

Since SMC-R composites in automotive applications may be exposed to elevated temperatures, moisture, automotive fluids and various types of chemicals, the effects of these environmental conditions on their strengths have been of interest to several researchers. Both tensile strength and modulus of SMC-R decrease with increasing temperature (Denton, 1979), which is mainly due to the softening and weakening of the matrix with increasing temperature. Similar effects were also observed for the fatigue properties of SMC-R.

Springer et al. (1980) studied the effects of several environmental conditions, such as humid air at 50 and 100% relative humidity, saturated salt water, No. 2 diesel fuel, lubricating oil, anti-freeze and gasoline, on the tensile strengths of SMC-R25 and SMC-R50 composites. For the SMC-R25, the matrix material was a polyester, but for the SMC-R50, both polyester and vinyl ester resins were used as the matrix materials. The test temperatures were 23 and 93 °c and the time of exposure was up to 6 months. The tensile strength of all SMC-R composites decreased in varying degrees depending on the type of liquid. In most cases, the SMC-R50 with polyester matrix showed a higher reduction in tensile strength than the SMC-R50 with vinyl ester matrix. The reduction in tensile strength was higher at 93 °c than at 23 °c. The extent of tensile strength reduction is shown in Table 12.6.

Table 12.6

Environmental effects on the tensile strength of SMC-R (from Springer et al., 1980) composites after continuous exposure for 6 months

Environment Effect
Humid air • Virtually no effect at 23 °C, 50% RH
• 20–25% reduction at 23 °C, 100% RH
• 35–45% reduction at 93 °C, 100% RH
Salt water • 20–35% reduction at 23 °C
• 47–55% reduction at 93 °C
Motor oil • 0–20% reduction at 23 and 93 °C
Anti-freeze • Slight reduction at 23 °C
• 55–75% reduction at 93 °C
Gasoline • Slight reduction at 23 °C
• 15–30% reduction at 93 °C

12.6 Polymer matrix composites (PMCs) for crashworthy structures

Increasing concerns and awareness about occupant safety have led to the concept of crashworthiness in designing automobiles. Both front and back rails of automobiles are designed for crashworthiness, which means they should be capable of absorbing high speed collision energy through progressive crushing over a length of 300 to 500 mm while maintaining a force level commensurate with tolerable deceleration for the survival of the occupants. These rails are usually tubular steel structures and are designed to crush by local plastic buckling that creates multiple folds. Carbon and glass fiber reinforced polymer matrix composites are attractive alternatives to steel and aluminum alloys for such structures since, when properly designed, they have not only shown progressive crushing, but also higher energy absorption per unit mass than steel and aluminum alloy tubes (Thornton and Jeryan, 1988).

Many studies on the quasi-static and dynamic crush behavior of thin-walled polymer matrix composite tubes have been published in the literature. Reviews of these studies are also available (Jacob et al., 2002; Mamalis et al., 1998; Ramakrishna, 1997). A variety of tube configurations, laminate constructions and fiber architectures, fiber and matrix combinations, processing methods and test conditions were employed in these studies. Various triggering mechanisms are used not only to induce a progressive crush, but also to control the maximum load at which the crush initiation occurs.

Table 12.7 gives a list of tube design and material parameters considered in the quasi-static and dynamic crush experiments with polymer matrix composite tubes. Reported specific energy absorption1 values were between 14 and 226 kJ/kg (Jacob et al., 2002). In comparison, specific energy absorption values for steel and aluminum tubes are 30–60 kJ/kg and 70–85 kJ/kg, respectively.

Table 12.7

Parameters considered in the crashworthiness studies of composite tubes

Tube configuration Circular cylinder
Square and rectangular cylinder
Tube dimensions Outer diameter (D)
Wall thickness (t)
D/t ratio
Fiber architecture  ± ? with ? = 15, 30, 45, 75 and 90
0/ ± ? with ? = 30, 45
0/90/ ± ?
Knitted fabric
Chopped fiber mat
Fiber volume fraction 15–60%
Fiber/matrix Carbon/epoxy, Carbon/PEEK, Carbon/PPS
E-glass/epoxy, E-glass/polyester, E-glass/vinyl ester
Kevlar 49/epoxy
Processing Filament winding
Roll wrapping
Resin transfer molding
Hot press molding
Triggering mechanism Beveled edge
Tulip edge
Drilled hole pattern
End plug with radius
Test Quasi-static
Dynamic (drop tower)
Environment Temperature

12.6.1 Failure modes of polymer matrix composite tubes

Depending on the fiber and matrix types, fiber architecture and other parameters listed in Table 12.7, the following major failure modes (Fig. 12.8) are observed in polymer matrix composite tubes under quasi-static compressive loading:

12.8 Failure modes observed in axial crushing of composite tubes.

1. Fragmentation mode The fragmentation mode is characterized by multiple short interlaminar and longitudinal cracks, less than the laminate thickness in length, creating numerous fragments of lamina bundles. The edges of the lamina bundle progressively shear to form a conical crush zone. This mode of failure is also called the transverse shearing mode.

2. Splaying mode The splaying mode starts with the formation of long interlaminar, intralaminar and parallel-to-fiber cracks in the tube wall, creating several long fiber bundles. As the loading is continued, the fiber bundles bend and form fronds that spread both outwards and inwards, but do not fracture.

3. Folding mode The folding mode occurs with the formation of small folds due to local buckling. This is very similar to that observed in steel and aluminum crush tubes.

4. Brittle fracture mode Like the fragmentation and splaying modes, the brittle fracture mode also starts with the formation of interlaminar cracks and fiber bundles; but instead of transversely shearing or bending, these fiber bundles fracture at the ends of the interlaminar cracks. The lengths of the interlaminar cracks in this case are reported to be between one to ten times the laminate thickness.

Depending on the fiber architecture and fiber/matrix properties, the fragmentation, splaying and brittle fracture modes are observed with many carbon and glass fiber reinforced epoxy tubes, whereas the folding mode is usually observed with Kevlar fiber reinforced epoxy tubes. In general, both splaying and folding modes produce high energy absorption and the brittle fracture mode produces the lowest energy absorption.

The load–displacement diagrams for the splaying and folding modes of failure are similar to those observed with steel and aluminum tubes. Initially, the load increases almost linearly with increasing deformation until the maximum load is reached (Fig. 12.9) and the load drops abruptly to a lower value. crushing of the tube begins at or, in some cases, before the maximum load is reached, due to high stresses in and around the crush initiator. After the first load drop occurs, stable crushing continues with the load fluctuating within a small range of values for a considerable crush deformation (Fig. 12.8). The mean load value and the area under the load-displacement diagram after the stable crushing is initiated are indicators of the amount of energy absorption in crush tests.

12.9 Experimental and predicted load–displacement diagrams in axial impact loading for (a) 2-ply and (b) 4-ply carbon fiber/vinyl ester [0/ ± 45] braided tubes of hollow square cross-section. An external plug and 45 ° chamfer were used for stable crush initiation (McGregor et al., 2010).

12.6.2 Energy dissipation mechanisms

Energy dissipation during crushing experiments of polymer matrix composite tubes takes place through several micro-failure mechanisms that include transverse matrix cracking, transverse matrix crushing, fiber breakage, fiber buckling and matrix crushing in the fiber direction, delamination and fiber–matrix interface debonding. Some of these micro-failure modes create debris that accumulates between the delaminated layers and becomes an additional source of energy dissipation through friction. The events that take place during crushing experiments in which splaying mode of failure is observed include the following:

1. formation of debris wedge due to crushing of the crush initiator (beveled edge, for example)

2. propagation of a Mode I opening crack at the apex of the debris wedge

3. extensive delamination in the fronds in the region of small radius of curvature at the wedge

4. flexural damage due to multiple transverse cracking through the individual plies

5. propagation of axial splitting between fronds

6. multiple longitudinal cracking through the individual plies of the fronds

7. frictional resistance due to the penetration of the debris wedge between the external and internal fronds

8. friction between adjacent plies on passing through the deflection arc of the crush zone

9. friction between the fronds and the loading platen.

It is estimated that nearly half the energy dissipation takes place due to friction between the fronds and the loading device. A significant amount of energy is also dissipated due to the bending of fronds (Mamalis et al., 1997). Friction between the delaminated layers is also a source of energy dissipation during crushing.

12.6.3 Energy absorptions in quasi-static and dynamic tests

Energy absorption of composite tubes is also influenced by the rate of load application. It is reported that energy absorption in high speed impact tests of polymer matrix composite tubes can be either higher or lower than that in quasi-static tests. Over a crush rate range of 10−4 to 10 m/s, Thornton (1979) reported a 20% change in the specific energy absorption for [0/90] woven fabric glass fiber/epoxy tubes, but only a 2% change for [0/90] woven carbon fiber/epoxy tubes. Farley (1991) reported that the energy absorption by [0/ ± θ]2 carbon fiber/epoxy tubes was not sensitive to crushing speed, while the energy absorption by [ ± θ]3 carbon fiber/epoxy tubes increased by up to 35% over the crushing speed range of 0.01 to 12 m/s. Kevlar fiber/epoxy tubes with [0/ ± θ]2 and [ ± θ]3 lamination orientations exhibited between 20 and 45% increase in energy absorption over the same crushing speed range. The difference in energy absorptions was attributed to fiber-dominated failures in the case of [0/ ± θ]2 tubes vs. matrix-dominated failures in [ ± θ]3 tubes. Boeman and Caliskan (2002), on the other hand, observed a large decrease in energy absorption capability of a variety of glass fiber reinforced composite tubes when tested in quasi-isotropic and dynamic tests. For [0/ ± 30] braided tubes, the energy absorption decreased by almost 55% between the crushing speeds of 0.083 mm/s and 2 m/s, but remained nearly constant between 2 and 5 m/s.

12.6.4 Modeling of progressive crushing

Modeling of progressive crushing of polymer matrix composite tubes is performed using explicit finite element codes, such as LS-DYNA, and using a phenomenological material model available in these codes. Two material models based on continuum damage mechanics are the MLT model, first proposed by Matzenmiller et al. (1995), and the CODAM model, developed at the University of British columbia by Williams et al. (2003). The MLT model uses strain-based damage parameters, whereas the CODAM model uses a strain-based damage potential. Recently a comparison was made between the MLT and CODAM models to simulate progressive crush behavior of [0/ ± 30]n tri-axial carbon fiber/epoxy braided tubes (Xiao et al., 2009). The specific energy absorption predicted by the CODAM model was in better agreement with the experimental results than the MLT model. This was attributed to the better predictability of the unloading segment of the load–displacement diagram by CODAM. Local unloading occurs when the material moving out of the crush front becomes part of the continuously growing crush fronds.

Another material model used for crush modeling is based on lamina stiffness degradation after the lamina fails due to one of the following failure modes:

• fiber rupture,

• fiber buckling,

• matrix cracking due to transverse tension and/or in-plane shear, and

• matrix failure due to transverse compression and/or in-plane shear (chang and Chang, 1987).

This material model has been applied by Huang and Wang (2009) to simulate the crush characteristics of carbon fiber/BMI tubes with [ ± 45/90/0/0/90/0]s lamination configuration. Even though the specific energy absorption predicted by the finite element analysis using the Chang–Chang model was within 5% of the experimental value, the simulated load–displacement diagram was not in very good agreement with the experimental load–displacement diagram.

12.7 Implications of preventing failure

Failure of polymer matrix composites used in automotive and road transport applications is not much documented in the published literature. Structural damage due to either fatigue or high energy impact are not reported; however, implications of structural damage, when they occur, can be very serious. This can include warranty issues, recalls, severe injuries, loss of life and litigation, all of which the automotive companies would like to avoid. There are no standards in the automotive industry for determining the assessment of structural damage on subsequent use or reuse of the composite structure. The repair issues for the damaged structure have also not been addressed. Replacement of damaged structure is not a good option, since it may be expensive to replace composite parts, especially if they contain carbon fibers.

Another concern for composite applications in the automotive industry is that once the vehicles are in customers’ hands, there is very little routine inspection of the vehicle or its components. Non-destructive inspection at regular intervals during the service operation of vehicles is not an easy task. Therefore, if there is damage to underbody structural composite components due to fatigue or low energy impact, it will remain undetected until structural failure occurs.

SMC-R composites described in this chapter are used mostly for semi-structural body applications, such as body panels and bumper beams. The principal design issue for body panels is the bending stiffness, which depends on the modulus of the material and the panel thickness. Since SMC-R has much lower modulus than steel, the bending stiffness in SMC-R body panels is achieved by using a higher thickness than steel body panels. Surface appearance, paint adhesion and joining are some of the other issues considered in SMC-R applications. To reduce the possibility of occurrence of visible damage due to low-energy impacts, the design stress can be limited to the knee stress level (shown in Fig. 12.1); but since this is a relatively low value, most body panel designs are based on the tensile strength of the material and a moderate to high factor of safety in the range of three to five. Any dents or visible damage that occur in service are repaired using adhesive patches.

SMC-R has also been used in bumper beams which are tested using pendulum impact tests at specified impact speeds (which, according to standards set by the National Highway Traffic Safety Administration, a US Government regulatory body, are 4 km/h for the frontal mid-width impact and 2.4 km/h for corner impact). In pendulum impact tests, the bumper beam is required to prevent damage to the structure behind it, but it is allowed to sustain damage and even structural failure. The principal material properties considered in the design of the SMC-R bumper beam are the strength (which determines the maximum impact load it can tolerate) and the modulus (which determines the maximum deflection it will undergo). The damage tolerance of the SMC-R is not usually considered at the design stage. In an actual accident, if the bumper beam has extensive damage or structural failure, it is replaced with a new bumper beam, which can become very expensive.

12.8 Future trends

Although at present carbon fibers are not used in any high-volume production vehicles, there is an industry-wide consensus that in order to increase the fuel economy of future vehicles, carbon fibers must be adopted replacing glass fibers in polymer matrix composites. The advantages of carbon fibers are their lower density, higher modulus-to-density ratio, higher strength-to-density ratio and higher fatigue resistance. They provide a much larger weight saving potential than glass fibers and will probably find increasing applications in body structure and chassis applications. However, there are many limitations of carbon fibers that must be resolved before they can be adopted on a larger scale. currently, the major limitation is their high price and lack of availability in large quantities. The manufacturing processes used for continuous carbon fiber reinforced epoxy are too slow and labor intensive for the automotive industry where the volume of production ranges from 50000 to 500000 per year; therefore, faster manufacturing processes need to be developed for automotive application of carbon fiber reinforced composites. The failure of carbon fiber reinforced composites in in-service conditions and its consequences must also be considered. For example, carbon fibers have low strain-to-failure and the damage tolerance of carbon fiber composites to low-energy impact may not be as high as that of glass fiber composites. If carbon fiber reinforced polymer composites are used in future automobiles for larger weight reduction, good understanding of their long-term behavior in the automotive environment, including damage development and their implications, needs to be developed.

12.9 References

Boeman, R.G., Caliskan, A.G. A novel capability for testing crash energy management structures at intermediate rates, 2002. [SAE Technical Paper No. 2002-01-1954, Warrendale, PA, Society of Automotive Engineers].

Brooks, R. Composites in automotive applications. In: Kelly A., Zweben C., eds. Comprehensive Composite Materials. Oxford: Elsevier Science; 2000:341–363. [Vol. 6].

Chang, F.K., Chang, K.Y. A progressive damage model for laminated composites containing stress concentrations. J. Composite Matls.. 1987; 21:834–855.

Chaturvedi, S.K., Sierakowski, R.L. Residual strength assessment of SMC composites subjected to dynamic impact. Composite Structures. 1983; 1:137–161.

Davis, B., Gramann, P., Osswald, T., Rios, A.Compression Molding. Munich: Hanser Publishers, 2003.

Denton, D.L.Mechanical properties characterization of an SMC-R50 composite, SAE Technical Paper No. 790671. Warrendale, PA: Society of Automotive Engineers, 1979.

Farley, G.L. The effects of crushing speed on the energy-absorption capability of composite tube. J. Composite Matls.. 1991; 25:1314–1329.

Huang, J., Wang, X. Numerical and experimental investigations on the axial crushing response of composite tubes. Composite Structures. 2009; 91:222–228.

Jacob, G.C., Fellers, J.F., Simunovic, S., Starbuck, J.M. Energy absorption in polymer composites for automotive crashworthiness. J. Composite Matls.. 2002; 36:813–850.

Kau, H.T. A study on the impact behavior of chopped fiber reinforced composites. Polymer Composites. 1990; 11:253–264.

Khetan, R.K., chang, D.C. Surface damage of sheet molding compound panels subject to a point impact loading. J. Composite Matls.. 1983; 17:182–194.

Mallick, P.K. Fatigue characteristics of high glass content sheet molding compound (SMC) materials. Polymer Composites. 1981; 2:18–21.

Mallick, P.K. Effects of hole stress concentration and its mitigation on the tensile strength of sheet molding compound (SMC-R50) composite. Composites. 1988; 19:283–287.

Mallick, P.K.Materials, Manufacturing and Design for Lightweight Vehicles. Cambridge: Woodhead Publishing, 2010.

Mallick, P.K., Newman, S.Composite Materials Technology: Processes and Properties. Munich: Hanser Publishers, 1990.

Mamalis, A.G., Robinson, M., Manolakos, D.E., Demosthenous, G.A., Ioannidis, M.B., Carruthers, J. Crashworthy capability of composite material structures. Composite Structures. 1997; 37:109–134.

Mamalis, A.G., Manolakos, D.E., Demosthenous, G.A., Ioannidis, M.B.Crashworthiness of Composite Thin-Walled Structural Components. Boca Raton, FL: CRC Press, 1998.

Matzenmiller, A., Lubliner, J., Taylor, R.I. A constitutive model for anisotropic damage in fiber-composites. Mechanics of Materials. 1995; 20:125–152.

McGregor, C., Vaziri, R., Xiao, X. Finite element modelling of the progressive crushing of braided composite tubes under axial impact. Int. J. Impact Engng.. 2010; 36:662–672.

Ramakrishna, S. Microstructural design of composite materials for crashworthy structural applications. Materials and Design. 1997; 18:163–173.

Shirrell, C.D. Variability in static strengths of sheet molding compounds (SMC). Polymer Composites. 1983; 4:172–179.

Shirrell, C.D., Onachuk, M.G., Influence of mold coverage upon the notch tensile strength of R25 sheet molding compoundHahn H.T., ed. Composite Materials: Fatigue and Fracture, 1986:32–50. [ASTM STP 907].

Springer, G., Sanders, B.A., Tung, R.W. Environmental effects on glass fiber reinforced polyester and vinylester composites. J. Composite Matls.. 1980; 14:213–232.

Sun, C.T., Sierakowski, R.L. Fracture characterization of chopped fiber glass composites (SMC). SAMPE Quarterly. 1980; 11:15–21.

Thornton, P.H. Energy absorption in composite structures. J. Composite Matls.. 1979; 13:247–262.

Thornton, P.H., Jeryan, R.A. Crash energy management in composite automotive structures. Int. J. Impact Engng. 1988; 7:167–180.

Wang, S.S., Chim, E.S.-M., Zahlan, N.M. Fatigue damage and degradation in random short-fiber SMC composite. J. Composite Matls.. 1983; 17:114–133.

Wang, S.S., Chim, E.S.-M., Zahlan, N.M. Fatigue crack propagation in random short-fiber SMC composite. J. Composite Matls.. 1983; 17:250–266.

Wang, S.S., Chim, E.S.-M., Suemasu, H. Mechanics of fatigue damage and degradation in random short-fiber composites, Part I – Damage evolution and accumulation. J. Applied Mechanics. 1986; 53:339–346.

Watanabe, T., Yasuda, M. Fracture behavior of sheet molding compounds Part 1. Under tensile load. Composites. 1982; 13:54–58.

Whitney, J.M., Nuismer, R.J. Stress fracture criteria for laminated composites containing stress concentrations. J. Composite Matls.. 1974; 8:253–265.

Williams, K.V., Vaziri, R., Poursartip, A. A physically based continuum damage mechanics model for thin laminated composite structures. Int. J. Solids & Structures. 2003; 40:2267–2300.

Xiao, X., McGregor, C., Vaziri, R., Poursartip, A. Progress in braided composite tube crush simulation. Int. J. Impact Engng.. 2009; 36:711–719.

1Specific energy absorption (SEA) is the energy absorption (equal to the area under the load-displacement diagram) divided by the mass of the crushed length of the tube.