process (Ea). This predicts material change with time when subjected to accelerated ageing: ‘life’ under service conditions can be extrapolated. The validity of this approach has been verified for nitrile rubbers by comparing long term degradation at temperatures just above service with short term, high temperature testing. In this instance, the results were comparable, giving confidence in the accelerating principle. This technique is more difficult to apply to other polymers let alone composites; some may present limited changes in mechanical properties during exposure, making it impossible to use Arrhenius to extrapolate service life. An example is a carbon fibre filled thermoplastic subjected to exposure in a sour fluid containing a small amount of hydrogen sulphide and carbon dioxide in methane, oil and water at 380 bar (around 5500 psi). Samples were immersed in the oil phase of the test fluid at 180, 190 and 200 C (356, 374 and 392 F) and retrieved at four different sampling points per temperature. Tensile (ASTM D3039M) and three point bend (ASTM D790M) testing was performed on unaged as well as exposed specimens. Tensile testing proved unreliable as the presence of an oily film on the surface of the samples meant that gripping was difficult and often led to slippage. The properties measured during the flexure tests were of much better quality; the data is summarised in Fig. 37.2. After an initial decrease in flexural modulus from time 0 at all temperatures, due to the absorption of the test fluid effectively plasticising the material, the property appears to remain stable over the test duration, even at the high test temperatures ranging from 180 to 200 C(356–392 F). The same can be said for the maximum strength and strain. They appear to both increase and then “stabilise” within one standard deviation from the mean. If a polymer appears to remain unaffected by the exposure fluid even after long times at elevated temperatures, it is not possible to establish a service life. In other words, for more chemically resistant materials test conditions may have to be so severe to initiate any material changes that this causes effects unlikely to be seen in service and so invalidates the approach of thermal acceleration. In addition, fluid absorption in materials such as composites is even more complex than in elastomers, thermoplastics and thermosets alone. Fluid diffuses into the composite matrix, with the extent of the absorption depending on the chemistry and morphology of the polymer as well as the volume fraction and configuration of the fibres, whether bubbles and voids are present in the matrix and wicking occurs at the interface between fibres and resin. The combination of fluid absorption and elevated temperature will cause plasticisation of the matrix, which allows relaxation of the polymer chains. It also affects the residual stresses present within the composite and may allow micro-crack formation which could ultimately lead to increased water absorption and/or weakening of the fibre/matrix interface; physical processes not accounted for by an Arrhenius approach to ageing. The test fluid may also chemically attack fibres and/or resin, compromising the fibre/resin bond or even dissolving fibres and/or resin. The effects of fluid absorption tend to become more apparent the higher the exposure temperature. This is due to the fact that the higher the temperature the faster the rate of diffusion and ultimately the higher the level of absorption over a fixed time period. It is clear from the above, that the tools to define service life for composite materials have not yet been fully developed. To help identify what mechanisms are involved in ageing, Element Materials Technology Hitchin is investigating different ln( /time to % change) /temperature service life 0 2 4 6 8 10 12 0 10203040506070 property level ageing time 50% Change in property level with exposure time and temperature: service temperature (black) with increasing temperature (blue to red) Schematic Arrhenius plot of rate of property change against reciprocal temperature using 50 % decrease in property level as failure criterion Fig. 37.1 Arrhenius principal 322 S. Munch et al.
RkJQdWJsaXNoZXIy MTMzNzEzMQ==