Paper to the Polymer Processing Society European Meeting, Strasbourg, 29th-31st August, paper 5-3.

S F Bush

Introduction

To view the introduction, please click on the link: ModellingLace

Paper (6-2-8) to the 10th International Annual Meeting of the Polymer Processing Society, Akron, USA, 5th-8th April 1994.

S F Bush with D R Blackburn, O K Ademosu, F B Yilmaz and P F Zhang

Introduction

Earlier papers^{(1, 2, 3, 4)} have described the factors affecting fiber-matrix cntacting, the organization of long fibers into coherent lace-like or mat-like structures, and the dependence of tensile strength on these two classes of variable. Equations for the mean number of touches¹ in the fiber structure and tensile strength² have been proposed which allow for the main variables present in moldings and extrusions, including differing matrix properties. Fiber management technology developed under the generic acronym SAFIRE⁴ (Self Assembling Fiber Reinforcement) has recently been applied commercially to practical moldings such as hard hats and pallets, in both of which examples impact strength is a key property.

Accordingly, a wide series of experiments has been carried out using the Izod method to determine the variation of impact strength as a function of the fiber length, fiber concentration, the fiber-matrix interface, fiber reinforcement structure and matrix properties. Both commercially available long-fiber granules and laboratory-compounded types have been used with different mold configurations and molding conditions for a variety of polypropylene and polyethylene matrices.

References

[1] S F Bush, Control of Fiber Structures in Melt Extrusion, 36 Ann Mtg Can Soc Chem Eng, Sarnia, Canada (1986) paper 32d.

[2] S F Bush, O K Ademosu, D R Blackburn, F B Yilmaz, Factors Affecting the Strength of Long-Fibre Reinforced Injection Moldings, Poly Proc Soc Eur Mtg, Prague (1992) paper 6-06.

[3] D R Blackburn and O K Ademosu, Factors Affecting Fiber-Matrix Contacting in Fiber-filled Granules, Poly Proc Soc, 9th Ann Mtg, Manchester 1993, Paper 6-14.

[4] S F Bush, Self Assembling Fibre Reinforcement (SAFIRE) processes, “Textile-reinforced Composites for Engineering”, Bolton Institute, Bolton, England, 12-14 January, 1994.

Paper (5) to the Institute of Materials 6th International Conference on Fibre Reinforced Composites, Newcastle, England, 29-31 March 1994.

Published in Plastics, Rubber and Composites Processing and Applications, Vol 24, No 3 (1995).
S F Bush with F B Yilmaz and P F Zhang

Abstract

A wide series of experiments has been undertaken to measure impact strength as a function of fibre length and concentration, the fibre/matrix interface, and induced fibre-mat structure and matrix properties. Both commercially-available long-fibre polypropylene granules and in-house polypropylene and polyethylene glass-fibre compounds have been used where the interface conditions are known and can be varied. For the fibre-mat structures achieved, notched impact strengths rise with fibre lengths and with fibre concentration, giving in all cases an improvement on the virgin polypropylenes – for some conditions a five-fold improvement at 25% w/w concentration.

Paper to the Polymer Processing Society 9th Annual Meeting, UMIST, 5th-8th April 1993, 02-35.

F B Yilmaz with S F Bush

Introduction

There are a number of different techniques for the production of long glass-fibre reinforced polypropylene (LGFRPP) granules for injection moulding. Dispersion and wetting of glass-fibre in the granule will be different according to the type of polypropylene grade and the compounding methods. Most of the materials used in this study were produced by the melt coating method and contain two bundles of filaments. The granules which are used in this study are given in Table 1.

* They have coupling agent.

Injection moulding is a two-step operation which involves first the plasticization of solid materials in a screw extrusion unit, followed by the high pressure pumping of molten material into a mould cavity. Most fibre breakage occurs in the screw extrusion unit with injection moulding of fully dispersed LGFRPP granules[1]. There is in fact an advantage in using melt coated granules. When glass-fibre bundles are travelling in the screw unit, they protect each other against severe fibre breakage. On the other hand to disperse the fibre bundles a separator is necessary. In tis study a proprietary dispersion device was used between the mould and the screw unit.

References

[1] R S Bailey and H Kraft, Int. Polymer Process. 22, p94-101 (1987)

[2] S F Bush, O K Ademosu, D R Blackburn, F B Yilmaz, “Factors Affecting the Strength of Long-fibre Reinforced Injection Moldings”, Polymer Processing Society 1992 European Meeting, Prague, 21-24 September, paper 06-06.

Paper to the Polymer Processing Society European Regional Meeting, Palermo, Sicily, 15th-18th September 1991.

S F Bush with C A Benson

Introduction

The design of reaction injection moulding processes requires knowledge of the system chemistry, an evaluation of the possible mixing processes which can be used, and characterisation of the subsequent in-mould behaviour of the reacting mixture. Mixing times are required to be a few milliseconds, moulding filling times typically a few seconds, and cycle times from one to three minutes.

The major problems in handling the phenolics chemistry for process design are how to characterise the variety of resins used and how to reduce the complexity of the chemical mechanisms to manageable proportions.

Phenolic reinforced reaction injection depends on rapid mixing of two or more unequal reactant streams with flow rate ratios typically somewhere in the range from 1 to 20 to 1 to 3. The major stream may contain reinforcing fibres of lengths up to 1.5 mm. The work reported here has concentrated on phenolic foams of densities up to about 500 kgm^-3. Fibre reinforcement at about 5% of final part weight can increase strength and modulus by about 50% for foams in the range 300 to 400 kgm^-3. On the basis of the chemical and mixing models an RRIM machine and mixhead have been designed and built for maximum shot volumes of 3 litres and an injection rate of 1 litre s^-1. This gives for instance a 1 m x 0.5 m x 19 mm foam moulding of 400 kgm^-3 density.

See also the section on Development of New Products and Processes.

Extracts from three Prosyma Research Ltd reports to the Everite Group, 6th April 1990 to 19th November 1990.

S F Bush

(The focus of the SAFIRE work moved to Santar situated in South Africa.)

SAFIRE A – Pipe extrusion (6th April 1990)

The work in the period under review (the S-runs) was taken up with extending the work of the R-runs (reported in Report 7) to a wide range of SAFIRE materials.

Altogether we have demonstrated pipe making capability in the SAFIRE polypropylene granules typical of the Santar materials as well as the Santar specified HDPE Hostalen GM5010. Note that this HDPE grade is black and this adds a dimension of potential difficulty because of the tendency for carbon black to cling to upstream discontinuities.

SAFIRE A – Pipe extrusion (1st May 1990)

The last report highlighted the likely effects of temperature on the presence or absence of weld lines. Consequently the work in the period under review (T-runs) has concentrated on repeating, where possible, the S-runs but at higher temperatures.

The results so far show that a significant decrease in polymer MFI requires a substantial increase in temperature in the mixing zone. This is particularly evident with Hostalen 5010, the HDPE grade proposed by Everite. Its MFI is about 0.3 at 230 ^oC compared with 0.4 for the PP Profax used up to now. Runs at 240 ^oC insread of 220 ^oC used in the S-runs show an appreciable increase in burst strength which may tentatively be ascribed to more complete disappearance of the weld line.

Injection Moulding (19th November 1990)

Recent results at UMIST have shown that SAFIRE granules may be injection moulded with separation and average fibre length retentions comparable with the values found in extrusion. This is achieved with a minor adaptation of a standard injection moulding machine and a simple version of the Fibre Separating Unit.

Paper to the Polymer Processing Society Regional Meeting, Brunel University, UK, 18th-19th September 1988.

S F Bush

Summary

As is well known, a particular feature of the injection moulding of thermoplastic polymers is the very wide range of strain rates developed at different parts of the flow field. Furthermore, the effects of high strain rates at the gate for instance are felt at points in the moulding remote from their point of origin. These effects show up principally in two ways, (a) in the fluid state and (b) in the subsequent solid state.

In the fluid state, the wide range of strain rates across the flow field means that the mean chain shape at a particular point is a function not only of the local strain rate, but also of the chains convected to the point from other parts of the flow. The chain shape determines the number of entanglements which in turn is a major determinant of the resistance to flow. Representing the flow resistance by means of a local viscosity, even if introduced as a function of local strain rate, will thus potentially give rise to significant error.

In the solid state, the incidence of high strain rates and the rapid cooling necessary for economic production give rise to substantial anisotropy and potentially damaging distortion as strained chains gradually relax.

To allow for these effects in a complex moulding simulation, chain shape has been set up as a primary dependent variable like the velocities and the temperature. The net rate of change of shape at any point and instant is thus the sum of convection terms and a local rate of change. To embed this treatment within a general two or three dimensional moulding or extrusion simulation entails a considerably simpler approach to the description of the chain shape than that adopted by other workers whose objective has been to derive point or equilibrium viscosity strain rate relations.

Using the present approach, attention is focussed on the processes of entanglement and disentanglement. The kinetic or non-equilibrium treatment adopted is suitable for a finite cell modelling system and also allows estimates to be made of the viscosity under special-case steady flow conditions. Furthermore, because of the strong dependence of crystal growth rates on local chain shape or orientation, the approach allows a prediction of the space variation of crystallinity developed in a solid moulding.

See also the section on Polymer Morphology & Fibre Reinforcement Mechanisms.

Paper to the 3rd International Conference of the Polymer Processing Society, Stuttgart, paper 05-7, 11th-15th April 1987.

S F Bush

Summary

Arguably the main problems in injection moulding reside in post-moulding distortion, manifested as shrinkage, warping and sinking. A major contribution to these problems is the existence of non-equilibrium conformations or shapes of the polymer chains in the post-mould state. As is well-known, the degree of this non-equilibrium condition varies throughout the moulding, usually being greatest in the wall regions closest to the gate. This variation arises essentially from the many large differences in strain rate experienced by a chain flowing into and within a mould, coupled with the fact that production rates of cooling do not, in general, allow chains time to fully relax while still in the fluid state.

During the injection process, the conformation of the chains, particularly their degree of entanglement, determines, with the temperature, the local viscosity. Both viscosity and conformation change with time and from point to point. In fact chain shape (and therefore viscosity) is the product of local conditions inherited from upstream. Mathematical models of injection moulding, however, customarily treat viscosity, where it is not taken as constant, as an empirical function of the local variables usually a principal strain rate and temperature.

By contrast the approach adopted by this paper is to employ a formulation which treats chain conformation and viscosity as dependent variables to be calculated within the computation along with the usual microscopic variables, namely bulk velocities and temperature. The chain shape is characterised by variables whose local rates of change are functions of the shape variables themselves, the velocity gradients and temperature. The chain shape in a finite region of the mould at any instant is then obtained as the resultant of the local rate of change and values of the chain shape convected into the region from upstream. The local viscosity and elasticity are then obtained. In this way we both obtain a prediction of molecular orientation and account directly for the viscosity and elasticity memory.

The results of this formulation are presented as simulations of three-dimensional injection mouldings of a variety of basic shapes. The simulation model is part of a design system which includes a graphical interpreter of mould cavity detail. Velocities, temperatures, pressure and molecular orientation are displayed by the post-processor as colour-graphic contours at various stages of mould-filling.

See also the section on Polymer Morphology & Fibre Reinforcement Mechanisms.