Background

Fibre reinforced plastics (FRPs) offer many advantages over conventional structural materials. They have high strength and modulus-to-weight ratios, are fatigue and corrosion resistant, tailorable and require low maintenance. However, because of their unknown long term properties when exposed to a combination of in-service loads and environments, designers are still reluctant to use FRPs in primary load bearing structures. The effect of exposure to heat, moisture, hydrocarbons, fatigue and static loads etc, and more importantly a combination of these parameters may degrade the material’s stiffness and strength. The lack of long term data for FRPs and of an accelerated ageing methodology that will predict the effect such a degradation might have on the residual properties and future life of the structure are two of the major issues hindering their wider use.
Long Term and Accelerated Ageing

Ideally, composite materials and their structures that are intended for long term use should be tested in real time and with realistic in-service environments. Often this is not viable because the time involved would significantly delay product development. However, these long term data are invaluable when generated as in a programme currently running at Westland Helicopters. Test coupons and main rotor blade sections were naturally aged at a hot/wet site in Australia, in stressed and unstressed condition for periods up to ten years. Further specimens were exposed in environmental chambers at 45°C and 85%RH. Others were exposed for a five year period at a hot/dry site in Australia. The fatigue strength of the structures and coupons was unaffected by the 10 year wet natural exposure. Coupon static degradation was greatest in matrix dominated properties, although the fibre dominated properties of the glass fibre materials degraded more than that of the carbon. The coupons in the hot/dry conditions were unaffected after 5 years.

Such a long term approach is generally not viable, and accelerated ageing techniques are required. In polymers (thermosets, thermoplastics or elastomers) free space exists between molecular chains. This free space allows the polymer to absorb fluids to which they are exposed, especially those with similar solubility parameters. Such absorption physically weakens the polymer and may also chemically attack the polymer. The kinetics of these processes are governed by diffusion and chemical kinetics, both of which are governed by Arrhenius relationships with regard to the influence of temperature. Accelerated testing can therefore be performed at elevated temperatures, with the results being extrapolated back to service temperature for life prediction purposes. Diffusion characteristics can be measured by liquid mass uptake or gas permeation experiments. Chemical kinetics, classically involving concentrations of reactants and products, can employ the fact that for crosslinked polymers, the concentration of crosslinks is approximately proportionate to modulus or stiffness. Hence, measurements of changes in modulus from ageing can be plotted logarithmically against linear time (for 1st order reactions) at each temperature. From a series of such ageing plots at different temperatures, times to attain the same degree of modulus change can be used to develop the Arrhenius plot.
Thermal Ageing in Aerobic and Anaerobic Environments

For high mach number aircraft structures, aircraft engines, space satellites and other environments, composite materials are expected to be durable at high temperatures. Carbon bismaleimides (IM7/5260), polyimides (IM7/K3B) and amorphous thermoplastics (IM7/8320) may be considered for the next generation High Speed Civil Transport fuselage and wing structure. These materials may be exposed to temperatures in the region of 125°C and 175°C, representing Mach 2.0 and Mach 2.4 flight respectively. Weight loss, glass transition temperature (Tg), and tensile strength data were shown for ±45 and notched quasi-isotropic laminates exposed at these temperatures for up to 5000 hours. Significant reduction in properties, including Tg and strength, occurred for the bismaleimide after only 2000 hours at 175°C. There are differences in properties when test coupons are aged, as opposed to panels being aged and test coupons cut after the ageing, figure 1. This difference arises because oxidation occurs at the edges of the aged specimens, where damage also initiates in the post exposure test. The notched bismaleimide laminates showed extensive matrix cracking at the surfaces but the ultimate tensile strength was not significantly reduced because the 0° fibres in the load direction contribute to the majority of the laminate strength.

Figure1. Isothermal ageing of carbon/bismaleimide ±45 laminates.

In aircraft engines, composites may be used in structures ranging from outer nacelle to core bearing housing structures. The range of different applications causes a wide variation in operating environments. For civil engines, such as the Rolls Royce RB211, component life requirements may be in the order of 25 years. To evaluate woven carbon bismaleimide (T300/52502) thrust reversers, test specimens were aged in air circulating ovens at 230°C and 250°C for up to 2000 hours. Weight loss and the effects of ageing on the flexural and impact properties were measured, and thermal analysis and microscopy were used to investigate chemical and structural changes. Components fabricated from carbon/polyimide (PMR-15) for higher temperature parts were cycled between extremes of -50°C and +350°C. The mechanical stresses resulting from anisotropic thermal expansion of crossply laminates lead to microcracking degrading the PMR-15. Much of the damage occurred from the peak temperature rather than the cycling.

In the space environment, temperature extremes are typically -150°C to +120°C with up to 30,000 cycles in a geostationary orbit, and -90°C to +90°C in low earth orbit. The principal concern is the dimensional stability of components such as communications antenna dishes with effectively zero overall coefficient of thermal expansion. The resultant microcracking from thermal cycling can be reduced by using lower cure temperatures and new toughened epoxy and cyanate ester resins.
Ageing in Liquids

For carbon/toughened epoxy (T800/924C) laminates, maximum moisture content attained in unidirectional specimens was about 1.4% (by weight), reached in 36 days in boiling water. The moisture uptake mode was Fickian. For multidirectional laminates, maximum moisture content was reached in 70 days. This difference in diffusivity is a function of laminate stacking sequence. Similarly, in carbon/PEEK (APC2) and carbon bismaleimides (5245C), the diffusivity rates are different for different lay-ups and thicknesses when exposed to liquids such as water, jet fuels and other aviation fluids. In addition, the diffusivity rate changes when the laminate is loaded mechanically, changing the internal stresses. Thermal spikes on different configurations of the 5245C material in the temperature range of 100-200°C caused enhanced moisture absorption by an increase in the free volume of the matrix at the spiking temperatures.

Liquid diffusion may also alter the strength and failure mode of composites. The compressive strength of the saturated unidirectional T800/924C laminates was reduced by 50% over the dry laminates. The failure mode also changed from in-plane fibre microbuckling, figure 2, to out-of-plane microbuckling for the saturated case.

Figure 2. In-plane fibre microbuckling in a compression loaded unidirectional laminate.
Physical Testing

Physical tests such as dynamic mechanical thermal analysis (DMA) and differential scanning calorimetry (DSC) were sensitive to changes in plasticisation of two woven carbon toughened epoxies (T300/924C-833 and T300/914C-833). These materials were exposed to various combinations of temperatures up to 70°C for up to 90 days in a variety of aerospace fluids including engine oil, hydraulic, anti-icing and cleaning fluids. The DMA results exhibited good correlation with high temperature interlaminar shear strengths. In a damage tolerance study of carbon and glass epoxy (F913C and F913G), laminates were impacted after immersion in water. The impact damage area was more wide spread and the subsequent compression after impact strength was lower than in non-exposed laminates.

Raman spectroscopy may be used to monitor swelling in matrix materials and determine diffusion coefficients. By monitoring the peak position of strain-sensitive Raman bands, the axial deformation of the fibre may be determined and used to define the states of localised stress and strain in the composite matrix. The dielectric technique may be used as a non-destructive examination method. The frequency domain can provide data on the extent of water ingress into a structure and the conversion of oxide to hydroxide at the interface. Time domain measurements can be used to identify regions of ingress and disbonding within the structure.

Modification of the resin system may enhance the composite's hydrophobic performance. The synthesis of halogen-substituted tetraglycidyl methylenedianiline (TGDDM) resins can reduce the water uptake of TGDDM resins by as much as 40%, while showing relatively minor changes in Tg.
Conclusions

General conclusions on technology needs for the long term exposure of composites may be drawn as follows.

Ageing occurs from the surface inwards and requires time to penetrate the FRP laminate fully. In addition, the fluid diffusion rate and the subsequent chemical ageing rate can be anisotropic and is dependent on the applied stress of the laminate. Thus, it is important to determine how stresses are altered during ageing on a ply-by-ply basis. Local changes in matrix properties around edges, holes cut-outs and other structural geometry features should be modelled.

Representing the true service history of the material or structure is vital, but this is often costly to do for screening purposes. A simpler but still representative screening procedure should be established. The tests for screening and exposure tests should be matrix dominated tests that will reflect matrix degradation while still providing data for design codes. These tests could include the transverse flexure or tension test, an interlaminar and in-plane shear test and interlaminar fracture tests including the double cantilever beam.

At present, no consistent method is available to accelerate effects of ageing from thermal or moisture exposure. For chemical properties an Arrhenius plot may be used. However, this approach is less well studied for changes in mechanical properties in FRPs and may not apply to realistic service exposures. Clearly an accelerated approach that accounts for a spectrum of loads, temperatures and liquid exposures is required.

Much of the ageing work being performed is either strictly application dependent or idealised, such as applying isothermal conditions or immersions in single liquids. Much must still being done to increase the confidence in the use of FRPs for long term applications. In an attempt to address the above technology needs, MERL has initiated a multi-sponsor programme that all industries may join which will examine all factors involved both experimentally and theoretically where possible, with the intent of developing a software tool to determine end-of-life factors for FRP structures.