AD-A268 931 FINAL REPORT PROGRAMME, SUMMARY AND ABSTRACTS DISCUSSION MEETING. ., _ AGEING IN PROCESSED POLYMERS UNIVERSITY OF BIRMINGHAM DT% .- 5-7 MAY 1993 SEP07 19933 SSPONSORED BY THE UNITED STATES OFFICE OF NAVAL RESEARCH EUROPE Thi~ docmen h.:s ......: rved I ~ itsAND THE UNITED STATES AIR FORCE EUROPEAN OFFICE OF AEROSPACE RESEARCH AND DEVELOPMENT Conference Organisation: ,9 Dr M J Richardson Professor J H Magill National Physical Laboratory Office of Naval Research Teddington 223/231 Old Marylebone Road Middlesex TWI1 OLW London NW1 5TH Phone: 081 943 6785 Phone: 071 409 4131 Fax: 081 943 6755 Fax: V71 723 1937 THIS DOCUMENT IS BEST QUALITY AVAILABLE. THE COPY FURNISHED TO DTIC CONTAINED A SIGNIFICANT NUMBER OF PAGES WHICH DO NOT REPRODUCE LEGIBLY. TABLE OF CONTENTS 1 PROGRAMME 1 2 SUMMARY 3 3 ABSTRACTS: Enthalpic Ageing in Polymers: Structure/Property Relationships and Some Unresolved Questions 6 Ageing of Starch-Based Products below the Glass Temperature 8 Enthalpy Relaxation and Physical Ageing in Polymers 9 The Effect of Ageing on the Material Properties of Polymer Glasses 10 Structure and Relaxation in Glassy Polycarbonate 12 Thermal and Mechanical History-Dependent Properties of Polymers in the Glassy State 13 Crystallisation Phenomena in Glassy Polymers as Revealed by Real Time Dielectric Spectroscopy 14 Physical Ageing of Polymers Studied by DMTA and DETA Techniques 15 Investigation of Physical Ageing in Polymers by Positron Annihilation Spectroscopy 18 Coupling Model Approach to Linear and Non-Linear Relaxations in Polymers 20 Modelling of Creep and Physical Ageing in Thermoplastics 21 Prediction of Effects of Physical Ageing on Long-Term Mechanical Properties by Finite Element Methods 22 Several Aspects of Ageing in Glassy and Plastically Deformed Glassy Polymers 24 Sequential Ageing Theory- Comparison of the Model with Experiment 25 Interpretation of Mechanically-Induced Ageing and De-Ageing in Terms of Redistributing Internal Stress 26 93-20664 93 0 2 1 3 0 Iflh~llml - " -1- PROGRAMME Wednesday 5 May 13.55 Introduction and Welcome 14.00 Enthalpic Ageing in Polymers: Structure/Property Relationships and Some Unresolved Questions R Ferguson (Dept of Chemistry, Heriot-Watt) 14.45 Ageing of Starch-Based Products below the Glass Temperature S Livings (Cavendish Laboratory, Cambridge) 15.30 Tea 16.00 Enthalpy Relaxation and Physical Ageing in Polymers J M Hutchinson and U Kriesten (Dept of Engineering, Aberdeen) 16.45 The Effect of Ageing on the Material Properties of Polymer Glasses J N Hay (Dept of Chemistry, Birmingham) Thursday 6 May 09.00 Structure and Relaxation in Glassy Polycarbonate T Pakula (Max-Planck-Institut, Mainz) 09.45 Thermal and Mechanical History-Dependent Properties of Polymers in the Glassy State J Perez (INSA, Villeurbanne) 10.30 Coffee 11.00 Crystallisation Phenomena in Glassy Polymers as Revealed by Real Time Dielectric Spectroscopy T A Ezquerra and F J Balta Calleja (CSIC, Madrid) 11.45 Physical Ageing of Polymers Studied by DMTA and DETA Techniques R E Wetton (Polymer Laboratories, Loughborough) 12.30 Lnch -2- 14.00 Investigation of Physical Ageing in Polymers by Positron Annihilation Spectroscopy W J Davies and R A Pethrick (Pure & Applied Chemistry, Strathclyde) 14.45 Coupling Model Approach to Linear and Non-Linear Relaxations in Polymers K L Ngai (Naval Research Laboratory, Washington) 15.30 Tea 16.00 Modelling of Creep and Physical Ageing in Thermoplastics B E Read, G D Dean and P E Tomlins (NPL, Teddington) 16.45 Prediction of Effects of Physical Ageing on Long-Term Mechanical Properties by Finite Element Methods K C McEwan, T G F Gray and W M Banks (Mechanical Engineering, Strathclyde) Friday 7 May 09.00 Several Aspects of Ageing in Glassy and Plastically Deformed Glassy Polymers E F Oleinik (Institute of Chemical Physics, Moscow) 09.45 Sequential Ageing Theory: Comparison of the Model with Experiment N G McCrum (Engineering Science, Oxford) 10.30 Interpretation of Mechanically-Induced Ageing and De-Ageing in Terms of Redistributing Internal Stress C P Buckley*, P J Dooling+ and S Hinduja+ (*Engineering Science, Oxford and +Mechanical Engineering, UMIST) 11.00 Coffee and General Discussion and Evaluation .IiS CP.',_ 12.30 Lunch F--- "..- ; -3- SUMMARY The manufacture of almost any plastic article involves some form of quenching as occurs, for example, in extrusion or injection moulding. The primary intention is to extract the product from, say, a mould as quickly as possible in order to maximise productivity. This, unfortunately, cannot be equated with the structural stability of the workpiece and so the properties of the formed component may subsequently change (ie age) as it slowly relaxes to a more stable state. Such a process is purely physical in origin and may be reversed on heating: it must be distinguished from the more familiar chemical effects which are generally irreversible and are associated with decreases in molar mass. "Physical ageing" was not a serious problem when polymers were used only for low grade applications (although even here it contributed to the "cheap and nasty" image that is still widespread for plastics -a simple example is the warping of an article that was well formed initially). Now that plastic components are increasingly designed for use in demanding situations it becomes essential for the manufacturer to have the ability to forecast long-term changes (particularly in the dimensional, mechanical and thermal properties of the processed component). This, in turn, implies the existence of simple and rapid tests that are able to define the condition of an aged component at some time after its fabrication. This Discussion Meeting was intended (i) to survey our present knowledge, (ii) to identify common features in the various manifestations of physical ageing and (iii) to indicate outstanding problems - and plausible routes to their solution. Lectures considered: 1 specific techniques that are used to characterise ageing 2 the relationships of these to ultimate properties 3 empirical and fundamental models 4 the predictive validity of such models. One of the commonest techniques used to study ageing in polymers is differential scanning calorimetry (DSO) because of the ease of sample pieparation and the simple experimental -4- procedure. An extensive body of DSC "relaxation" information is now available for polymers and blends and it is becoming feasible to discuss "structure (repeat units) - property" relationships in a meaningful manner, although we are still some way away from real predictive capability in this respect. One definite advance for binary blends with widely separated Tg's is that ageing is dominated by the component with lower Tg. There is still controversy over the equilibrium enthalpy limit to which a glass can relax -is it given by a simple extrapolation of the supercooled liquid or is the non-linearity of the enthalpy- temperature curve greater then is implied by a linear C.p-T curve? The latter is suggested by the results when this quantity is considered as another unknown but this approach does not seem to be validated for the few cases when the extrapolation can be made using data obtained for oligomers. At the other experimental extreme from DSC as a technique is positron annihilation spectroscopy (PAS) which gives information on the free volume (Vf) -both the average value and its distribution -in a system. Vf is a key parameter in many theories of ageing and the complexity of the PAS procedure will only be fully justified if Vf can be unambiguously defined as a function of mechanical and thermal history. Although DSC and specific volume (density) measurements undoubtedly reflect the effects of physical ageing, it is not clear how the observed changes are related to those parameter measurements of more immediate practical value - the several mechanical moduli or yield stresses, for example. There was general agreement at the Workshop that the changes in thermodynamic properties correlate with the increases in the elastic modulus and the tensile, flexural and compression yield stresses (which, in turn, imply decreased impact strengths, fracture toughness and ultimate elongation -ie increased brittleness). By contrast, creep data correlate only with changes in specific volume and not with DSC results. In this instance creep and specific volume are found to relax much more rapidly than enthalpy. The point is emphasised by data obtained for polycarbonate at room temperature; this polymer is stable with respect to enthalpy (on time scales of up to 15 years) whereas the creep compliance and density change for several days after quenching the sample. At the meeting there was a widespread (but mixed) feeling that samples are "deaged" by stress (due to the regeneration of free volume) so that at some yield limit much of the sample history may be eliminated and consequently correlations between thermodynamic and mechanical behaviour may be somewhat fortuitous. It seems highly likely that direct 1:1 relationships between properties are the exception rather than the rule. Different experimental techniques are likely to sample different regions of the overall relaxation spectrum that are influenced by the summation of -5- imposed restraints experienced during the history of a sample. Even the thermodynamic quantities, density and enthalpy, which normally change in parallel with each other under simple conditions, show divergent behaviour when stress is applied to the specimen. Modelling of ageing behaviour was discussed from two very different points of view (i) the engineering and (ii) the molecular dynamics approaches. Ideally, the two should be comparable but this desirable situation is still not experienced. Creep models for engineering calculations, which incorporate ageing behaviour, are reasonably successful. They may use empirical functions that are not accepted universally but show the general form that is implied in a more basic analysis. In the latter category, several phenomenological models have had success in predicting/explaining such features as the fine structure of DSC curves in the vicinity of the T. region. Unfortunately, the anticipated "structural parameters" that may be extracted from the measurements seem to have little basic meaning. The need for realistic microscopic models remains high and the promising approaches used by Ngai, Pakula, and Perez were discussed. These models must use independent molecular parameters in order to simulate the reasonable success of the empirical models. Only when this occurs will the problems of physical ageing be resolved meaningfully. -6- Enthalpic ageing in polymer glasses: a) Structure property relationships, b) Some unresolved questions Roderick Ferguson Department of Chemistry, Heriot Watt University, Edinburgh, Scotland Differential scanning calorimetry has proved to be a simple and reliable technique for monitoring physical ageing processes in polymer glasses. Methods for analysing the Cp data obtained from such experiments fall into two general categories:- a) Cp curve shape analysis (eg the phenomenological multiparameter model of Hodge) or b) Enthalpy change as a function of ageing time and temperature (eg the empirical Cowie Ferguson model). We have concentrated on the latter methodology because:- (i) Thermal lag effects have more influence on the shapes of Cp curves than for enthalpy changes. (ii) The ageing parameters obtained from the phenomenological model vary both as a function of ageing time and temperature and hence have limited use in a predictive sense. (iii) Although the phenomenological models can fit Cp curves quite well, the corresponding enthalpy changes due to ageing are found to be at least 2-3 times bigger than those found experimentally. (iv) Enthalpy changes due to ageing can be correlated with volume relaxation experiments. Enthalpic ageing data have now been obtained for a range of homopolymers, some copolymers and several blend systems, so that one can start to consider the dependence of the enthalpic ageing parameters on the chemical structure of the polymer repeat unit(s) ie structure property relationships. Two useful parameters that can be obtained from the Cowie/Ferguson model parameters (log tc and 0) are the time to reach 99.9% of thermodynamic equilibrium, te and the average segmental activation energy, <Ea>. This latter quantity is related to the relaxation time distribution function, p('r), which in turn can be obtained from the relaxation function, 0(ta). Values of <E,> obtained at an ageing temperature of Ta = Tg-10 K are of the same order of magnitude as the activation energy requirements for processes such as the mechanical P relaxation in PMMA. Furthermore, there is a strong linear correlation between <Ea> and the enthalpic Tg. Values of Log(tc) obtained at Ta = Tg-10K also exhibit a reasonable linear correlation with Tg. For a range of SAN copolymers, the CF model ageing parameters were found to vary with copolymer composition. For blends where the Tg's of the two components are not too far apart, the ageing parameters were found to exhibit a weak dependence on the copolymer composition. However, in blends where the two component Tg's are far apart (such as PVME and PS), the physical ageing is found to be dominated by the more mobile lower Tg component. -7- There still remain several unresolved questions in the field of enthalpic ageing in polymer glasses, namely:- 1a) Is it valid to obtain the equilibrium enthalpy change for an infinitely aged sample, AT.U.), from a simple extrapolation of the liquid enthalpy data ? (as has been assumed by many workers in this field). 1b) Does the presence of topological constraints in the polymer glass mean that AH_(T.) will always be smaller than expected ? 2) How can the phenomenological multiparameter model(s) be improved so that a) the parameters obtained from them do not vary with ageing time or temperature and b) these models successfully predict the observed enthalpy changes ?