No description available.

No description available.

No description available.

The stress corrosion cracking of model epoxy-glass 0°/90°/0° crossply and 0 unidirectional composites in aqueous sulphuric acid has been studied. Specimens, in the form of coupons cut from laminates were tested under conditions of uniaxial tension at constant load, whilst partially immersed in an acidic environment. Two different modes of failure, which resulted in fracture were observed. In the "so-called" Mode I failure, fracture occurred within that part of the composite immersed in the acid,whereas in Mode II, it occurred in the unimmersed part. Both failure modes were observed for crossply and unidirectional composites. In unstressed 0°/90°/0° specimens a third failure mode (Mode III) was observed, in which the damage took the form of transverse and longitudinal cracking of the unimmersed part. The failure mode depended upon the magnitude of the initial applied strain, the nature of the environment, and the type of environment cell. At initial applied strains of greater than about 0.15% only Mode I failure was observed. The similarity between the times-to-failure of laminates with those recorded for single E-glass fibres showed that the resin was not providing significant protection from the acidic environment. To account for this result it is postulated that the acid rapidly permeates these composites through environmental microcracks, which form parallel to the axis of the glass reinforcement. in the 0° plies. The formation of these microcracks is due to a reduction in the resin/glass interfacial strength in the presence of acids. Confirmation of this phenomenon was obtained from experiments on the transverse cracking behaviour of crossply laminates, immersed in aqueous acid and tensile tested at constant strain rate. At initial applied strains of less than about 0.15% and depending upon the experimental conditions, failure was by either Mode I or II. Mode II, which occurs in shorter times than Mode I was observed in acids (e. g. sulphuric acid) giving rise to relatively insoluble glass degradation products, providing evaporation of moisture from the unimmersed part of the specimen was possible. The Mode II failure mechanism has been identified with the transport of the aqueous acid along the glass resin interface from the immersed to the unimmersed part of the composite. Here precipitation of the less soluble glass degradation products causes a localised stress sufficient to initiate and propagate a stress corrosion crack. At these strains both Mode I and II failures occur at times significantly greater than those observed for glass fibres. Mode III failure is similarly attributed to the precipitation of glass degradation products within the composite. Microscopical examination of the stress corrosion fracture surfaces did not reveal any morphological differences between Mode I and II failure, apart from the appearance of more crystalline products in the Mode II fractures. Although glass fibre fracture is the ultimate failure mechanism in these composites, stress corrosion of the fibre matrix interface was found to be a necessary precursor.

The strength of carbon fibres as single filaments, bundles and in composite form has been investigated. Single filaments have been shown to have a strength distribution to which a two-parameter Weibull distribution may be fitted. The main objective of the work was then to reconcile the observed strength distributions of both simple (mono-fibre type) and hybrid composites to that of the single filaments. This has been done by preparing and testing a range of unidirectional bundles and composites of varying geometry. In all cases, the carbon fibre used was taken from the same batch of 1000 filament tow, thus ensuring a consistent statistical base for the comparison. The hybrid composites consisted of single bundles of carbon-fibre surrounded by glass-fibre in an epoxy matrix. These were compared with impregnated and dry bundles of the same dimensions. Multi-ligament hybrids were also prepared and tested to assess the effects of bundle dispersion. The strength distributions of both dry and impregnated bundles in relation to that of the single fibre have been shown to follow the trends predicted by Coleman and Harlow and Phoenix respectively. The hybrids by comparison show an enhanced strength which is only partially explained by thermal stress arguments. A hypothesis is advanced which proposes that the enhanced strength of a hybridised ligament is due to a larger critical group of individual fibre fractures being required to initiate catastrophic ligament failure than is the case for non-hybrid bundles. The significance of these effects with respect to practical composites is discussed.

This Thesis traces the development of high flyash content concrete. This concrete contains a high proportion of flyash in the cementitious content, usually between 50 and 80% by volume. The concrete was initially developed as a roller-compacted hearting (or interior) concrete for dams. Concrete with a low workability, a low thermal (and non-thermal) movement, and relatively high tensile strain capacity was required. In 1976, the Author had proposed that this could be obtained by using a concrete with a high paste content and that the high paste content should be obtained by using large quantities of flyash and a low cement content. This concept was studied in depth during a CIRIA project, which is described. For roller-compacted concrete, the optimum proportion of flyash in the cementitious content was found to be between 70 and 80%. The properties of such a concrete were found to be suitable for use in dams. High flyash content concrete has now been proposed for two dams, one in America and the other in the U. K. In the Thesis, a trial mix programme is described which extended the concept of a high flyash content in concrete into the workability range of immersion vibration. During this study, it was postulated that there is a relationship between the flyash contribution to compressive strength and the water/cementitious ratio, in the same fashion as that proposed for cement by Abrams in 1912. It is shown that the contribution of flyash to strength is more sensitive to the water content than the contribution of the cement, and that the conventional methods of mix design for flyash in concrete may not be using flyash to its best advantage. The Thesis concludes with a description of a placement of high flyash content concrete in a road and storage area. The sub-base of the road was placed through a paver-finisher without roller compaction and was a concrete in which flyash made up 75 to 80% of the cementitious content. The pavement-quality concrete was compacted by immersion vibration and flyash made up 60 to 65% of the cementitious content. Both concretes were extensively tested and have performed very well. It is concluded that high flyash content concrete is a new material, and that it should have a use in many forms of Civil Engineering construction, in particular dams and roads. Usually, there are substantial economic advantages in such a use, and the long-term in-situ properties are generally better than those of similar conventional concretes.

A laboratory scale drum winder has been designed and built for the production of pre-preg. Cross-ply hybrid matrix laminates were made from the pre-preg with glass fibres/epoxy resin in the longitudinal plies and glass fibres/epoxy resin-urethane elastomer in the transverse ply. The addition of urethane to the matrix in the transverse plies alone increased the applied strains necessary for the initiation and development of transverse cracking during the extension of cross-ply laminates. This resulted in a smaller reduction in laminate stiffness (due to damage) at a prescribed level of strain. Damage resistance was similar to that in cross-ply laminates with urethane additions to the matrix in both the transverse and longitudinal plies (a uniform matrix laminate). It was found that urethane additions lead to improved damage resistance in cross-ply laminates because they lower the transverse ply modulus and increase the matrix toughness in the transverse ply. During the extension of cross-ply laminates, stable (constrained) transverse cracking was observed in thin transverse plies and unstable (brittle) transverse cracking in thick transverse plies. The effects of urethane additions on the development of constrained transverse cracking and brittle transverse cracking were modelled with a shear lag stress analysis combined with an energy balance and a statistical expression for the transverse ply strength respectively. The ultimate properties of hybrid matrix laminates, having improved damage resistance, were expected to be better than uniform matrix laminates with a similar urethane content in the matrix. However, the tensile strength of circular centre-notched (0,90)s hybrid matrix laminates was lower than uniform matrix counterparts and the compressive strength of (02,902)2s hybrid matrix laminates was similar to uniform matrix counterparts.

Throughout the materials community, there exists extensive literature on carbon fibre chemical, physical and mechanical properties, and the effect of these on composite performance. To date, however, the atomistic chemical nature of the changing carbon fibre surface when surface treatments are applied has been a very difficult area to study using experimental measurements alone. This study shows the potential of molecular simulation methods to be applied to real engineering materials, and the chemical insight that can be gained in using them. Inverse gas chromatography experiments show how the dispersive and acid-base nature of the carbon fibre surface changes as a function of surface treatment. In combination with these experimental measurements, the SORPTIOISFTM molecular modelling algorithm is used to develop models that provide a representation of the surfaces of these materials in terms of specific chemical functional groups. These models are compared with separate experimental data, and used to predict and describe aspects of the adhesion in the interfacial region of a carbon fibre-thermoset matrix resin composite.

The development of internal stresses in planar random Saffil fibre reinforced aluminium with a range of fibre volume fractions has been studied theoretically and in monotonic and cyclic deformation (Bauschinger) experiments at room temperature and 77K. The conventional method of analysing Bauschinger experiments is extended to allow for a separation of the mean and thermal stresses. This analysis is applied to experimental results enabling the mean stress hardening rate and the magnitude of the thermal stress to be measured. The experimental results are compared with predictions of the mean field model, which is based on the Eshelby method of determining internal stresses. For that purpose the Eshelby S tensor for a planar random array of fibres is calculated. Because the aluminium/Saffil composites are not isotropic in the transverse directions, the plastic strain used in the calculations has to be determined experimentally. A method for quantifying the anisotropic plastic flow of the aluminium/Saffil composites is proposed and the results are used in calculations of the mean stress hardening rate. A comparison of predictions for the mean stress hardening rate with results of the experimental analysis proposed here shows that good agreement is obtained for low fibre volume fractions at 77 K. The results also show that relaxation of the mean stress increases with fibre volume fraction and that at 77 K the mean stress hardening rate is about a factor of two larger than at room temperature. The measurements of the thermal stresses obtained in the Bauschinger experiments are in quantitative agreement with results obtained in monotonic tests. The magnitude of the thermal stress at room temperature or 77 K is independent of fibre volume fraction and a comparison with predictions shows that relaxation of the thermal stress increases with fibre volume fraction. Cycling in the Bauschinger experiments reduces the thermal stress and hence the separation of the mean and thermal stresses is essential for a reliable measurement of the mean stress hardening rate. Matrix hardening contributes considerably to the overall hardening of the composite, both at room temperature and 77 K. The modified Orowan-Wilson model, which enables the plastic friction coefficient to be measured in copper-tungsten composites, has been applied to the aluminium\Saffil composites. The model requires both the mean stress and the peak stress curves obtained in Bauschinger experiments to be linear in plastic strain. Most of the peak stress curves for the aluminium/Saffil composites are non-linear but for the curves which are linear the predictions of the model are not in quantitative accord with experimental results. This may be because relaxation reduces the mean stress and the source shortening stress in different proportions. The diameter of the Saffil fibres is also close to the lower end of applicability of the model. The temperature dependence of the mean stress hardening rate suggests that relaxation is thermally activated. A model for relaxation of the mean stress is proposed. An equation is derived for the number of Orowan loops per fibre and it is assumed that the rate controlling mechanism of relaxation is cross slip of screw dislocations. The estimated activation energy is independent of fibre volume fraction but the activation volume decreases with increasing fibre volume fraction. The magnitudes of activation energy and activation volume support the assumptions of the model. A preliminary study on the early stages of fatigue shows that persistent slip bands form in the matrix of the aluminium/Saffil composites.

The incorporation of a ductile second phase into a monolithic ceramic is a well-established toughening technique. This study is concerned with the toughening of alpha-alumina/20 vol% iron ceramic matrix composites. The main emphasis has been to investigate the effect of the morphology of the iron on the toughness of the composite material. The fabrication process employed was that of hot pressing. Manipulation of the starting powders has resulted in the formation of hot pressed materials with morphologically different microstructures (a difference in the distribution of the iron particles throughout the alumina matrix). The fracture toughness of the materials has been measured using indentation, double torsion and double cantilever beam methods. The indentation technique was found to be inappropriate for the finite measurement of fracture toughness due to the non-ideal crack patterns produced with each indentation event. All three tests, however, found the composite materials to be tougher than the parent matrix and comparatively accurate fracture toughness values were produced by the double torsion (DT) and double cantilever beam (DCB) testing techniques. The toughest composite was found to be the one which possessed a combination of long-crack (a network distribution of iron particles) and short-crack (discrete iron particles) toughening mechanisms. The DCB test was developed in this study to enable the in situ observation of crack/particle interactions at the microscopic level. Observation of the interactions was made possible by performing the tests on a straining stage which was mounted inside a scanning electron microscope. The observed toughening mechanisms for each composite were isolated and incorporated into a numerical analysis. This analysis allowed specific energy values to be allocated to each type of toughening mechanism. It was found that the energies for the intrinsic toughness of the matrix, debonding and plastic deformation were in the correct proportion to each other and the energy attributed to plastic deformation compared favourably with a theoretically derived value.