Cracking under pressure - failure identification from fractography of polymer composites

Materials World magazine
1 Jul 2012
Recovered tail from Airbus A300 accident

Groundbreaking research into the fractography of polymer composites is aiding composite failure identification, prediction and, ultimately, prevention. Dr Emile S Greenhalgh from Imperial College London offers an insight.

Download a PDF (1.12mb) of the original magazine article for full diagram information and more illustrations


The growing use of polymer composites in transport, sport and energy applications is leading to an increased need to solve in-service failures. Fractography, the study of fracture surface morphologies to glean information about the material microstructure and failure processes, is the principal method for achieving this.

In addition to being vital for post-mortem analysis of failed structures, fractography is used by researchers to provide an insight into composite failure mechanisms. For instance, in the 1970s scientists at the Royal Aerospace Establishment used fractography to identify failure processes under compression loading. These observations laid the foundations for modelling compression failure, which is perhaps one of the most critical areas in current composites research. Fractographic observations are now being used to underpin the development of physically based failure criteria, providing a vital link between predictive models and experimental observations.

Metallics vs composites

The underlying methodology for composite fractography is akin to that for traditional structural materials, such as metallics. There are, however, key differences. For metallics, the approach has been to collate an atlas of fractographs cataloguing all the different possible failure modes of a particular material. A researcher will deduce the failure of the component by correlating the fracture surfaces of interest with images in the atlas.

This approach has proven to be of little practical use for polymer composites, principally due to the huge range of different failure modes that develop in composites, and the interaction between them. With temperature and moisture also affecting fracture morphology, the approach pursued by composite fractographers has been to understand the mechanisms and relate these accordingly.

There are additional aspects in which composite fractography differs – a practical issue is the extensive fractured areas that develop. During metallic failures, particularly under cyclic loading, there is often a single crack that can be traced back to a site of initiation, and much of the excess energy released during failure is absorbed through plastic deformation. On the other hand, failure of polymer composites is usually violent and highly unstable, with the excess energy released through the formation of secondary failures – usually delaminations. These secondary damages complicate the analysis considerably, demanding additional time and resources. Therefore, composites fractography takes considerably longer to conduct than for an equivalent metallic component failure.

Gathering the evidence

Undertaking a fractographic analysis of a failed polymer composite component can be a challenging experience. The investigator is often presented with a complicated mixture of broken pieces, some of which amount to little more than piles of dust.

Failures in most structural (laminated) composites can be partitioned into three types:

  • Translaminar – through the thickness in which fibres have been broken (Figs 1-3)
  • Intralaminar – through the thickness in which only matrix, or fibre-matrix interfaces have been broken
  • Interlaminar – in the laminate plane, in which the layers have separated


Usually, these modes strongly interact. The approach followed is to characterise the mode and direction of all the fracture surfaces, and then to characterise the interactions between the different cracks, building up movement diagrams of the fracture surfaces. These are then collated to deduce the source and sequence of failure. Characterisation of the failures usually requires imaging at high magnifi cations, necessitating use of scanning electron microscopy.

The polymer matrix provides most of the information about crack-growth directions and is the main component associated with delaminations. One of the most important phenomena is the process by which multiple matrix fractures initiate along a crack front, begin to propagate on slightly di  erent planes and subsequently converge onto one plane. This leads to the formation of riverlines (Fig 4), which are one of the main means of deducing crack-growth directions on fracture surfaces where peel is the dominant loading mode. However, when interlaminar shear loading is present, the fracture morphology is very different and inclined platelet features (cusps) are apparent on the surfaces (Figs 5 and 6).

Predicting damage

Polymer composites is an exciting field, with novel composite incarnations under development, such as 3D-reinforced, nanoreinforced (hierarchical composites) and multifunctional materials.

Predictive modelling is now at an unprecedented level of maturity, with damage and failure modelling being used in component design. Furthermore, researchers are starting to explore new concepts such as crack arrest and self-healing of composites. Underpinning this development is a better understanding of how composites fail, and how to inhibit such failures.

As composites have developed, fractographic techniques have matured to the stage where researchers can confidently relate surface morphologies to the failure modes in a component. There is now a growing demand for materials engineers with composite fractographic skills. So, although being presented with a pile of broken bits and being asked to work out what has happened can be daunting, the rewards of gleaning the fundamental processes by which these materials fail are considerable.  

Sequencing cracks

Continuity of intersecting cracks is an important method for sequencing composite failures. For example, Figure 7 shows two intersecting cracks in the web of a CFRP I-beam that had failed in threepoint bending. Note that crack AA is continuous, while the second crack (BB) is discontinuous, with a step where it intersects AA. Clearly, AA must have formed first, since the presence of AA had subsequently modified the crack direction of BB. Similarly, Figure 8 shows a delamination surface, with a matrix crack extending vertically across it. Notice that the fracture morphology of the delamination is continuous across the crack, indicating the delamination had occurred first.

Defects, damage and in-service effects

Invariably, composites tend to fail from defect or damage and, therefore, an important aspect of fractographic analysis is identification of such defects, as well as knowledge of how they can influence strength and toughness. Perhaps the most common defect is fibre waviness (Fig 9), which can be associated with poor lamina drape at sharp changes in geometry or movement of the component during processing. Another important defect is voidage (Fig 10), which can develop from trapped volatiles during processing or, in the instance shown here, crimping induced by through-thickness stitches. Another defect that can be introduced during processing is contamination (Fig 11), which severely degrades the interfacial properties. Finally, although composites are extremely resistant to fatigue loading, failure of these materials under these conditions leads to unique features, such as matrix rollers (Fig 12).


Case studies of failures

Airbus A300 accident

Perhaps one of the most important in-service failures in the development of aerospace composites was that of the American Airlines A300 (AA587) in 2001 in New York. The accident investigation identified that the vertical tail and rudder had detached from the fuselage early in the accident, and there were concerns that there could have been a design problem with using composites in such a critical aerostructure. A detailed fractographic analysis of the retrieved tail component and attachment points (lugs) was undertaken at NASA. The culmination of this investigation was that the failure was shown not have been associated with impact damage, manufacturing defects or fatigue, but was due to highly excessive loading induced on the tail caused by pilot error.  

Formula One roll hoop

Composites have been extensively used in Formula One racing cars for several decades, and it is in this field that composites are perhaps pushed the hardest. Formula One requires fabrication of very complicated components that experience complex loading conditions, a good example being the roll hoop that protects the driver should the car turn upside down. This component is exposed to severe outof- plane loading, which can promote delamination formation, particularly at regions of high curvature. Unfortunately, such regions are also prone to manufacturing defects. In this particular investigation, a roll hoop was found to have failed prematurely. The subsequent fractographic investigation identified that the cause was poor consolidation and voidage at a change in thickness at one of the corners of the component. Such sites are recognised as problematic, and now many Formula One teams used technologies such as Z-pins (essentially, carbon-fibre nails) to resist failure initiation from such sites.