Weathering of injection-moulded glassy polymers: changes in residual stress and fracture behaviour

Injection-moulded bars made from poly(methyl methacrylate) (PMMA) polystyrene (PS), polycarbonate (PC) and poly(vinyl chloride) (PVC), have been weathered outdoors in Jeddah, Saudi Arabia, for varying periods. The residual stresses in the bars have been found to vary considerably with exposure and in some cases the surface stress has reversed, becoming tensile. With PS and PVC, significant surface damage developed on exposure and the presence of surface flaws together with tensile residual stresses at the surface can be expected to reduce the resistance to fracture, especially in the case of PS which suffers considerable molecular weight reduction near the surface. The changes in residual stress in PMMA and PC were smaller, and the surface deterioration, examined by scanning electron microscopy, was much less than with PS and PVC. The performance of PMMA and PC in fracture tests generally declined less with weathering than did that of PS and PVC but the materials used were all commercial grades designed for use in Europe and the results obtained here should not be taken to indicate the relative weatherability of the different polymers. The significance of the results relate to the mechanism of failure involving the interaction of reversed residual stresses and surface flaws.     

On the properties of organic liquids affecting the crazing behaviour in glassy polymers

The properties of organic liquids affecting the environmental stress crazing in glassy poly(methyl methacrylate), polycarbonate and polyvinyl chloride were examined with reference to the diffusivity into the polymers. The isochronal bending tests for the sheet specimens were conducted by a traditional method using the elliptical bending device in environments of five alcohols, four n-alkanes and three good solvents at 30°C. The diffusivity of these liquids exclusive of n-pentane and the solvents was evaluated by measuring the weight gain of specimen under the accelerated soaking at 52°C. The absorption rate of each liquid, given by an initial weight gain per unit time, was shown to be correlated with a new factor consisting of the viscosity and molar volume of environmental liquid and of a ratio of dielectric constant of liquid to that of polymer. The critical crazing strains in alcohols and n-alkanes were also demonstrated to relate to this factor. In good solvents no crazing but dissolution of polymer surface was observed. It thus may be concluded that the crazing is strongly correlated with the diffusivity of environmental liquid, and in addition that the liquid with the above factor of greater value clearly tends to cause the environmental crazing in glassy polymers.     

The creep behaviour of isotropic polyethylene

Dead loading creep and constant strain rate yield experiments have been used to study the tensile creep behaviour of three grades of isotropic polyethylene. This has provided further evidence for the existence of two yield points in isotropic polyethylene. Two different models have been used to attempt to describe this behaviour. Although the results can be described by to both the two process model of Wilding and Ward and the co-operative jump model of Fotheringham and Cherry, it appears that the two process model provides a more convincing quantitative fit to the data. This revised version was published online in September 2006 with corrections to the Cover Date.     

The mechanism of creep in concrete

Summary An attempt has been made to investigate the fundamental physical mechanisms which cause creep in cement paste and in particular, to provide experimental evidence which will indicate which of the many theories of creep is most likely to be correct. The method used was to oven dry, load, and then resaturate cement paste and concrete specimens at different temperatures. Only one type of test clearly differentiated between proposed creep mechanisms. This test indicated that the shear or sliding of submicroscopic gel particles over one another, lubricated by water, must be a predominant factor. Existing theories regarding the size of the critical spaces in the cement gel which cause dimensional instability in concrete structures have been confirmed by the use of organic liquids of varying molecular size.

---

Résumé L’étude qu’on décrit ici est une tentative d’élucider les mécanismes physiques fondamentaux qui déterminent le fluage dans la pate de ciment et, en particulier, d’apporter une preuve expérimentale qui indique laquelle des nombreuses théories du fluage peut être considérée comme étant correcte. Il a été procédé par séchage à l’étuve, chargement, puis resaturation de la pate de ciment et des éprouvettes de béton à différentes températures. Un seul type d’essai a permis de reconna?tre clairement les mécanismes de fluage proposés. Cet essai indique que le cisaillement ou le glissement les unes sur les autres des particules de gèle sub-microscopiques lubrifiées par l’eau doit être un facteur prédominant. Les théories sur les dimensions des espaces critiques dans le gèle de ciment, qui provoquent une instabilité dimensionnelle dans les ouvrages en béton, ont été confirmées avec des liquides organiques aux molécules de diverses dimensions.

---

---

This work was carried out while the author was a Research Officer at the Central Electricity Research Laboratories, Leatherhead.     

Transitions in creep behaviour

Attempts to identify the mechanisms operating during creep are often made by examining plots which yield apparent activation energies, or the stress or grain size-dependences of creep-rate. The forms of such plots are here examined and the ambiguities which arise near transitions from one regime to another are noted. The ranges of temperature, stress and grain size commonly used are inadequate and serious errors in interpreting the results of creep tests will continue to be made until a better understanding of the interaction of the basic processes is developed, so as to enable the positions of transitions to be predicted.At Dept. of Metallurgy, University of British Columbia, Vancouver 8, BC, Canada until 31 March 1970.     

The creep behaviour of Synroc and alumina

The creep of Synroc C and alumina in four-point bending in argon was investigated in terms of the relaxed symmetric stress and the reference asymmetric stress; the alumina being used as a reference material. The creep tests were undertaken in the temperature range from 850°C to 1300°C. The rupture behaviour of Synroc at 950°C indicated a high stress exponent, and that the creep ductility was unusual in that the strain increased with increasing test time. A scanning electron miscroscopy examination of Synroc after creep revealed the development of defect-free oxidised surface layers. For Synroc, neither prior exposure to pre-heating in air, nor prior indentation affected the creep rate behaviour. This was attributed to the formation of the oxidised surface layers and the associated healing effects of the damage produced by the indentations.     

The creep behaviour of ultra-high modulus polypropylene

The creep and recovery behaviour of highly drawn polypropylene monofilaments has been studied over the temperature range 20 to 50° C. A range of samples was examined to identify the influence of draw ratio and molecular weight. It is concluded that the permanent flow creep arises from the presence of two thermally activated processes, one of which relates to the -relaxation process and is associated with the crystalline regions of the polymer, and the second with the molecular network.     

The competition between shear deformation and crazing in glassy polymers

Whereas thin films of some polymers such as polystyrene readily form crazes when strained in tension, thin films of other polymers such as polycarbonate rarely exhibit crazing under the same testing conditions; the polymers that rarely craze tend to form regions of shear deformation instead. Polymers such as polystyrene-acrylonitrile which lie between these two extremes of behaviour may exhibit both modes of deformation. Thin films suitable for optical and transmission electron microscopy (TEM) of six such co-polymers and polymer blends have been prepared. After straining, the nature of the competition between shear deformation and crazing is examined by TEM. It is found that in these polymers many crazes have tips which are blunted by shear deformation. This process leads to stress relaxation at the craze tip, preventing further tip advance. In this way short, but broad, cigar-shaped crazes are formed. Examination of the deformation at crack tips in the same polymers shows more complex structures, the initial high stress levels lead to chain scission and fibrillation but as the stress drops, shear becomes the dominant mechanism of deformation and the stress is relieved further. Finally, at long times under stress, chain disentanglement may become important leading to fibrillation and craze formation again. The nature of the competition is thus seen to be both stress and time dependent. Physical ageing of these polymers, via annealing below T _g, suppresses shear leading to the generation of more simple craze structures.     

The effect of molecular size on non-Fickian sorption in glassy polymers

Rutherford backscattering spectrometry has been used to determine the concentration against depth profiles of n-iodoalkanes diffusing into a polymer glass photoresist. All the iodoalkanes smaller than iodohexane show strongly non-Fickian, or Case II, diffusion. After an induction time a sharp front forms, with almost no concentration gradient behind the front. Ahead of the front the concentration decreases exponentially with depth, a form predicted for Fickian diffusion ahead of a moving boundary. Values of the diffusion coefficientD extracted from this Fickian precursor decrease strongly withn, the number of carbon atoms in the iodoalkane. A similar decrease is observed for the front velocity, the magnitude of which is in qualitative agreement with that predicted by the Thomas and Windle model of Case II diffusion. For the larger values ofn, D decreases asn ^–2, prompting speculation that these longer chains diffuse into the glass by a reptation-like mechanism.     

The fracture behaviour of ABS polymers

The fracture behaviour of three different grades of ABS has been investigated over a range of temperature and strain-rates. A two-stage fracture process has been observed consisting of a slow crack propagation region followed by catastrophic failure. The surface features in these two regions have been examined in detail using scanning electron microscopy. In all cases, the fracture initiation site was located at either a surface defect or at an internal inclusion much larger than the average rubber particle size.     

The macroscopic yield behaviour of polymers

A yield criterion, not previously compared with the actual macroscopic behaviour of polymers, is herein compared with the pressure-modified octahedral shear stress criterion earlier suggested by others. This new relation, which is a version of the von Mises criterion, accommodates differences in tensile and compressive yield strengths and accounts for any dependence of yielding on the hydrostatic component of the applied stress state.With the use of thin-wall tubes accounting for the majority of experimental points, the yield behaviour of polycarbonate and polyvinylchloride was investigated. Besides these findings, results previously reported by others have also been utilized in this paper. Since these various studies employed quite different polymers, the excellent overall correlation of experiment with prediction should merit the serious attention of persons interested in the macroscopic yield behaviour of polymers.Comparisons between this new criterion and the modified octahedral shear stress are also made in regard to the effect of pressure on subsequent yield behaviour. Although not fully verified, it is suggested that the predictions which result using this new criterion, seem a little more reasonable.     

Deformation kinetics by crazing in glassy polymers

In order to confirm the proposed theory of the plastic deformation caused by crazing, the theoretical deformation curves derived from the dislocation analogue were calculated using the data of craze behaviour measured under static tension and compared with the experimental results of creep and constant strain-rate tension in PMMA and PVC plates wetted by kerosene as a crazing agent. Favourable agreement between the theoretical and experimental results may provide evidence that the proposed method is valuable for estimating the amount of the plastic strain caused by crazing.     

Physical and thermodynamic aspects of the glassy state,and intrinsic non-linear behaviour of creep and stress relaxation

Rheology in the highly viscous liquid and the glassy state is reviewed and discussed. A distribution relaxation time due to the co-operative molecular motion is exhibited both in the highly viscous liquid and in the glassy states. However, only in the glassy state has the structure been frozen-in at some particular internal state resulting from the incomplete establishment of a thermodynamic equilibrium state. Therefore, the intrinsic non-linear rheological behaviour of the glassy state is explained from the physical and thermodynamic aspects in the glass transformation region. The volume relaxation of soda-lime-silica glass and the thermal history of glass during forming process have been studied. Finally, the role of rheology in thermal stress and fracture mechanics is also mentioned, which subsequently will allow us to re-evaluate the mechanisms of toughening and weakening of composite materials.     

The compression creep behaviour of silicon nitride ceramics

A comparison has been made of the compression creep characteristics of samples of reaction-bonded and hot-pressed silicon nitride, a sialon and silicon carbide. In addition, the effects of factors such as oxide additions and fabrication variables on the creep resistance of reaction-bonded material and the influence of dispersions of SiC particles on the creep properties of hot-pressed silicon nitride have been considered. For the entire range of materials examined, the creep behaviour appears to be determined primarily by the rate at which the development of grain boundary microcracks allows relative movement of the crystals to take place. Now with the BNF Metals Technology Centre, Wantage.     

Yield criteria for amorphous glassy polymers

Three amorphous polymers, polymethyl methacrylate, polystyrene and polycarbonate were tested in uniaxial tension, uniaxial compression, plane strain compression and simple shear, over a range of temperatures. In each test, the yield point was precisely determined via residual strain measurements after unloading. With the yield stresses determined for these four different stress states, two pressure dependent shear yield criteria, i.e, the modified Von Mises and the modified Tresca criteria, were checked and compared. It is shown that (i) in each case (material, temperature, initial ageing state), the yield locus is satisfactorily described by either one or the other of the two criteria, and (ii) each criterion can be associated with a specific deformation mode (either homogeneous or localized in shear bands). As for the temperature dependence of the yield stress sensitivity to the hydrostatic pressure, it appears to be related to the glass transition temperature (Tg) and more precisely to the α and β relaxations. Finally, the pressure dependence of the yield stress can be possibly explained as being due to two effects: (i) the influence of pressure on molecular motions leading to yielding and (ii) the influence of pressure on the microstructural state. This revised version was published online in November 2006 with corrections to the Cover Date.     

Non-linear creep behaviour of PET

The non-linear creep behaviour has been studied on PET films at room temperature. A particular value of the stress, _c, was used to characterize the change between the linear to the non-linear domain. The variations of the elastic modulus, the relaxed modulus and _c revealed great sensitivity to the morphology of the crystallization. A molecular model of non elastic deformation, assuming (i) hierarchical correlated molecular motion, and (ii) nucleation and expansion of sheared-microdomains, was used to analyse the role of stress on anelasticity. To take into account the two-phase structure of semicrystalline films, a phenomenological series/parallel model was applied to express the mechanical coupling between amorphous and crystalline phases. Quantitative agreement was found between theoretical predictions and experimental data for low and high stresses. However, there was a discrepancy in the rate of recovery because the model predicts a strain recovery slower than the experimental behaviour. Consequently, it is proposed to develop further the molecular model mentioned above by specifying the energy profile of a sheared-microdomain and its stress dependence. Then, the difference between creep and recovery strain rate could be explained.Nomenclature A Anelastic equilibrium compliance - A Parameter proportional to the relaxation strength - b Shear vector - Correlation parameter - _i Particular value in the distribution - _e Average value in the material ( = 0.27) - Correlation parameter characterizing the ability of chain orientation - d _a Amorphous density - d _c Crystalline density - Parameter of the mechanical mixing law - E Tensile modulus - E _c Crystalline Young's modulus - g _i Statistical weight in the distribution - G ₀ Shear modulus at 0 K - J _max Creep compliance at the end of the creep time - J _max(0) Value ofJ _max for low stresses - J _u Unrelaxed compliance - J _i( _i;A _i) Calculated compliance for a couple ( _i;A _i) - J _exp Experimental creep compliance - J _a Compliance of the amorphous part - J _c Compliance of the crystalline part - J _sc Compliance of the semi-crystalline material - k Boltzmann's constant - Parameter of the mechanical coupling law - R Radius of a shear micro domain - ₀ Stress necessary to cross the energy barrier only by mechanical activation - T _c Crystallization temperature - t _c Creep time - _mol Time for a translational motion of a structural unit over a distance comparable to its size - Particular value of _mol in the time distribution - Characteristic time for the secondary relaxation - ₀ Time proportional to the Debye time - t ₀ Scaling time parameter determined by the experimental value of _mol - U Activation energy for an elementary molecular motion - X _c Crystallinity ratio - V _a Volume fraction of the amorphous part - V _c Volume fraction of the crystalline part