Table of Contents ICAC'98
Static Loading and Low-Velocity Impact Characterization of Graphite/PEEK LaminatesF. Aymerich, P. Priolo and D. Vacca
Dipartimento di Ingegneria Meccanica
Università di Cagliari, Piazza d'Armi - 09123 Cagliari (ITALY)
|TABLE OF CONTENTS|
Impact damage is widely recognised as one of the most detrimental damage forms in composite laminates, which dissipate the incident energy by a combination of matrix damage, fibre fracture and fibre-matrix debonding. While in ballistic impacts damage is clearly visible by visual inspection, low-velocity impacts are usually responsible for internal damage that is difficult to detect but which is likely to reduce the load-carrying capability of the laminate. Static deflection tests are often used to predict the response of thermosetting graphite-reinforced composites to low velocity impacts; for high-toughness thermoplastic composites, however, the rate-dependence of the resin properties and the lack of relevant data mean that results obtained by static loading need to be confirmed under dynamic conditions. In this study, both static indentation-flexure and low-velocity impact tests were carried out on graphite/PEEK laminates. The internal damage, assessed by means of a range of techniques (X-radiography, ultrasonics, deply, optical microscopy), was related to the absorbed energy as obtained by instrumented testing. Test results are compared and discussed to characterize the onset of individual damage modes and to clarify the role of the loading rate in the failure process.
Fibre-reinforced composite materials have been increasingly used in load bearing structures since they have a number of advantages over more conventional materials: high specific strength and stiffness, good fatigue performance and corrosion resistance. A serious obstacle to more widespread use is the sensitivity of composite laminates to impact and static loads in the thickness direction [1, 2]. The energy absorbed during impact is mainly dissipated by a combination of matrix damage, fibre fracture and fibre-matrix debonding and this may lead to significant reductions in the load carrying capability of the laminate. While in ballistic impacts the damage is very localized and clearly visible by external inspection, low velocity impacts involve long contact time between impactor and target, and so produce global structure deformation with internal damage at points far from the contact region.
For this reason, low-velocity impact tests are often simulated by simple static indentation-flexure tests, neglecting the influence of dynamic effects on the structural response [3, 4]. In epoxy based laminates, static and dynamic tests are observed to produce the same damage modes and comparable force-deflection behaviour . However, impact and interlaminar toughness tests conducted on composites based on matrices such as PEEK point to a strong rate-sensitivity [6-8], therefore indicating possible differences in response between impact and quasi-static loading. In this study. both indentation-flexure and low-velocity impact tests were carried out on quasi-isotropic graphite/PEEK laminates. The internal damage, assessed by a range of techniques, was related to the absorbed energy as obtained by instrumented testing. Test results are compared and discussed to characterize the onset of individual damage modes and to clarify the role of the loading rate in the failure process.
The material studied was a quasi-isotropic graphite/PEEK laminate with 63 % by volume of continuous AS4 graphite fibres. The lamination sequence was [45/-45/0/902/0/-45/45]s with a thickness of 2.2 mm. The specimens were square plates 90 x 90 mm2 cut from 420 x 420 mm2 panels supplied by Fiberite Europe.
Impact tests were carried out on a purpose-built drop weight impact testing machine; the testing rig consists of a drop tower with two vertical rod guides, an impactor which runs on the guides through two linear roller bearings and an anvil with a specimen support fixture. A pneumatic catch mechanism captures the impactor after the rebound in order to prevent multiple impacts on the specimen, which was clamped between two rings of 70 mm internal diameter. The impactor was instrumented with a semiconductor strain-gauge full bridge bonded to the tup, provided with a hemispherical nose of 12.5 mm diameter. Impact and rebound velocities were measured by an infra-red sensor which illuminates a flag attached to the falling mass. The indentor displacement was evaluated by double integration of the contact force signal, while the kinetic energy lost by the impactor was measured from the impact and rebound velocities.
Different impact energies and velocities can be obtained by changing the falling mass and the drop height. In this study the impactor mass varied between 0.8 and 5 kg while the drop height was changed over the range 0.09-1.15 m so as to obtain impact velocities ranging between 1.3 and 4.8 m/s.
Static tests were carried out utilizing a servohydraulic testing machine at constant impactor velocity of 0.025 mm/s, as obtained by an LVDT. The same indentor and support fixture were used for both static and impact tests
Impact damage in composite laminates, unlike metallic materials, consists of a complex three-dimensional pattern of different fracture modes. Several evaluation techniques are then necessary to assess accurately the nature and extent of the damaged zone.
In this investigation the material studied was non-destructively examined by means of penetrant enhanced X-radiography, ultrasonic inspection and optical microscopy. X-ray radiography utilizes a radio opaque dye (a zinc iodide solution was used in this study) to infiltrate the damaged areas before irradiation of the specimen. This method offers a very detailed picture of fracture modes, but internal damage not connected to the surface cannot be impregnated with the zinc solution and so remains undetected. The method can provide through-the-thickness details if proper stereoscopic techniques are adopted, but the three dimensional reconstructed picture is not clearly decipherable and the information obtained reliable only for quite low levels of damage.
Full volume ultrasonic analysis is capable of revealing internal damage and of providing information about its extension and depth. The ultrasonic inspection was performed with a system consisting of a three-axis computer-controlled scanning bridge (0.025 mm resolution), a 150 MHz Krautkramer HIS2 ultrasonic pulser-receiver and a 500 MHz digital oscilloscope. The specimens, immersed in water, were scanned in pulse-echo mode by a 50 MHz transducer with 12.5 mm focal length. The complete ultrasonic waveforms, including the front and back face reflection, were acquired at each scanning point, stored on the computer hard disk, and then processed by selecting the appropriate gate width and location in order to obtain images of the delamination at the desired depth. Both time of flight C-scans (showing the trough-thickness position of damage) or amplitude C-scans (presenting a slice of the material at the selected gate position) could be reconstructed from the acquired ultrasonic data. All the specimens were examined from the two sides and the information obtained was recombined to a single volumetric image.
The extent and distribution of fibre fracture were observed after deply  of damaged specimens. In order to separate individual layers of different orientations, the specimens were heated in an oven at 600° C for 1 minute. In this way, by partially pyrolizing the interfacial matrix-rich zones, individual layers could easily be separated from one another, while sufficient matrix material remained intact to hold the fibres in the original position. The fibre fracture paths were observed on the two sides of each deplied layer with a light microscope and their length, measured perpendicularly to the fibre orientation, was determined with the aid of a micrometric translation stage.
One of the key factors controlling the composite's response to dynamic loading is the rate sensitivity of the constituent materials' mechanical properties. This effect has often been neglected in the past, in that numerous studies on common graphite/epoxy laminates indicated only a minor strain-rate influence on impact phenomena.
A first way to examine the influence of velocity is to consider the absorbed energy as a function of impact energy. In fig. 1 the energy absorption values of graphite/PEEK laminates are plotted versus imparted energy for both static and impact tests; for comparison a similar graph for quasi-isotropic [±45/0/90]2s graphite/epoxy laminates is shown in fig. 2. The contact duration ranged between 2 and 8 ms, depending on the impactor mass, for impact tests and between 100 and 150 s for static tests.
|Fig 1: Energy absorption versus impact energy||Fig 2: Energy absorption versus impact energy for quasi-isotropic graphite/epoxy laminates|
The total delamination area -as obtained by summing the delamination areas at single interfaces- and the fibre fracture length -evaluated by summing the length of broken fibres paths at each deplied layer- are plotted in figs. 3-5 versus absorbed energy. While, as expected, no difference exists between the dynamic and static response of graphite/epoxy composites (fig.3), we notice that in graphite/PEEK samples higher loading rates are associated with larger delamination areas and with smaller amounts of broken fibres, thus indicating a distinct rate-dependence of graphite/PEEK laminates, probably as a consequence of viscoelastic effects dominating at low rates.
|Fig 3: Total delamination area vs. absorbed energy for quasi-isotropic graphite/epoxy laminates||Fig 4: Total delamination area vs. absorbed energy|
A more precise description of the dominant damage mechanisms induced by transverse loads can be obtained by combining instrumented testing data with the information deduced through the different damage evaluation techniques adopted.
|Fig 5: Fibre fracture length vs. absorbed energy||Fig 6: Force versus indentor displacement curves for impact and static loads (impact energy = 4.3 J)|
|Fig 7: X-radiographs of damaged graphite/PEEK laminates (impact energy = 4.3 J). (static load, left - impact, right -same length scale)|
Fig. 6 also shows a static force-displacement curve (energy = 4.3 J), which is similar to that for dynamic tests with the same energy, with an abrupt load decrease again indicating the onset of major fibre fracture. The pattern of statically induced damage, consisting of a network of matrix cracks, delaminations and fibre fracture -at both front and rear face-, is comparable to that of impact-induced damage, but is characterized by a smaller projected and total delamination area (fig. 7) and by a larger amount of broken fibres. A typical through the thickness distribution of fibre fracture lengths is shown in fig. 8 for both static and impact loads.
|Fig 8: Through-thickness fibre fracture length (impact energy = 4.2 J)|
For higher energy (8.2 J), the force-deflection curve (fig. 9) exhibits several decreases in load, corresponding to the attainment of the load-carrying capability of the laminate, with progressive propagation of fibre fracture from the back surface through all the laminae. Again, static and impact curves are similar, but damage differs in matrix damage extension (figs. 10,11) and fibre fracture amount, with static tests associated with smaller delamination areas and longer fibre fracture zones.
Fig 9: Force versus indentor displacement curves for impact and static loads (impact energy = 8.2 J)
Fig 10: X-radiographs of damaged graphite/PEEK laminates (impact energy = 8.2 J) (static load, left - impact, right)
|Fig 11: Amplitude ultrasonic C-scans of impacted graphite/PEEK specimen (impact energy = 8.2 J)||Fig 12: X-radiograph of damaged graphite/epoxy laminates (impact energy = 8.2 J)||Fig 14: Fibre fracture path in impacted graphite/PEEK sample|
|Fig 13: Through-thickness fibre fracture length (impact energy = 8.2 J)|
Static and low-velocity impact tests were carried out on quasi-isotropic graphite/PEEK laminates in order to study the influence of load-rate on the main damage mechanisms produced by transverse loads.
Static and dynamic-loaded samples exhibited comparable energy absorptions, but impacted laminates showed larger extensions of delaminations and, at the same time, smaller amounts of fibre fracture. The variation in the dominant damage modes with impact velocity reveals a clear rate-sensitivity of the material,a probable consequence of the viscoelastic nature of the matrix. For this reason, simulating a low-velocity impact phenomenon with a static test of the same energy can be hazardous since this may lead to underestimation of the delamination area. Therefore, predictions of graphite/PEEK impact resistance have to be based on results obtained under conditions reproducing the real deformation rate of the structure.