Joint venture – polymer development
A 10-year academic and industry collaboration has changed the way polymers are developed for commercial production, says Professor Thomas McLeish of Durham University, UK.
The way people think about polymer resin design and manufacture is undergoing a quiet revolution. A major part of this has been initiated by a UK project linking industry research groups to eight university teams with expertise in physics, chemistry, engineering, mathematics and computer science.
The Microscale Polymer Processing initiative was conceived at a workshop in York, UK, a decade ago and run by the UK Polymer IRC, and the Polymer Engineering Group. It is a mark of the close connectivity of the field that senior research staff from four UK-based polymer producers were able to indentify the need for radical change in the science base, and had the academic teams to achieve it. The frustration stemmed from the empirical nature of polymer materials and processing development, despite the progress in the molecular chemistry and physics of polymer processing.
The aim of the project was to design new plastics at the molecular level, with all the subsequent processing and product properties targetted from the design of the polymerisation itself. Rather than explore a polymerisation technology and then look for markets, market knowledge motivated intelligent molecular design of new materials. The late 1990s saw rapid development of new catalysts such as the metallocenes which promised greater control of molecular weight distribution and branching of polyolefins. With these new molecular design tools available at the industrial scale, it became even more important to know how to use them.
As a result of the York workshop it was decided to link up the best UK academic groups in a downstream structure, novel for the academic sector, that mirrored the pattern long-used in industry. Polymer chemistry at Durham and Sheffield was linked to physics in Leeds, chemical engineering at Bradford and Cambridge, and mechanical engineering at Oxford. Parallel to the experimental structure, a modelling stream fed into it at each level, with molecular physics leading into computational fluid dynamics.
An academic research scaffold was created that self-connected vertically, as well as horizontally, to industrial teams in BP Chemicals, BASF, DSM, Dow Chemical, Du Pont Teijin and Lucite International. The final ingredient was advanced experimentation. If the modelling tools to be developed were truly molecular, then the test materials would need to be probed with experiments sensitive to molecular structures and configurations, not just the standard rheological and macroscopic flow behaviour. X-ray and neutron scattering were therefore added to the optically-probed melt flow rigs at Cambridge and Bradford Universities.
In parallel, theories used to calculate stress-fields and flow of melts, were also developed to make predictions for scattering experiments. The project idea fitted hand-in-glove into the EPSRC managed programme, Materials Processing for Engineering Applications, announced the same year. With £2m from the programme and a further £500,000 from industry, the project launched in 1999.
The structure gave the project several advantages. Conceptual steering received a great deal of input from industry throughout its run. This allowed the grant to be exploited, steering resources to fundamental problems. A new understanding of why molten plastic streams swell rather than contract when emerging from a die – a delicate interplay of molecular structure and process conditions – became a major study. Also, the connectivity of the research groups enabled the construction of multiscale experimental and theoretical research that no single group could attempt. Finally, experiments on well-defined polymers with known molecular weights and architectures could be used to test and hone the theoretical models before being applied to more complex and less well characterised industrial materials.
One such experiment was ambitious – could an entire science team be put together in a way that would allow molecular-level understanding of a polymer melt flow to be checked at every point in the process? Together, the Bradford and Durham groups built a recirculating melt flow rig that could maintain a precious quantity of deuterium labelled polystyrene passing continuously through a contraction and expansion component. The entire rig was transported to the Neutron source at Grenoble, France, and the beam scanned to follow the entangled chains as they stretched out (relaxed). A mirror experiment on a computer predicted scattering patterns at the measured points in the flow and compared the two. Although a science-based experiment, this breakthrough paid industrial dividends by identifying hidden anisotropy in polymers that appear perfectly optically isotropic yet the oriented neutron signal correlates with direction-dependent crack propagation in thesolidified product.
The second phase of the project, running from 2004-2009, has more international and UK-based partners, including groups in the Technical University of Eindhoven, the Netherlands, University College London, and the University of Reading, both UK. The industrial consortium was joined by Mitsubishi Chemical, Tokyo, Japan, Ineos, Lyndhurst, UK, and Basell-Lyondell, Roterdam, the Netherlands, to tackle quantitative models for fully industrial polymers, flow-induced crystallisation and the flow of filled polymers.
All streams used the micro-scale polymer processing methodology of working from clean experimental systems through multiscale modelling to industrial application. Metallocene catalysts that incorporate long side branches into a growing polyethylene chain created a family of complex-topology melts. It proved possible to calculate the expected distribution of linear, star-shaped, combs, trees and more complex architectures within such a resin, and also to build mathematical models of its process flows. The image top right is typical of the degree of detail now computable in the stress field of a real processing industrial melt, and its direct comparison with observation of crystal morphology and density, which are essential for final product properties.
The group now knows the basic rules that connect molecular structure to process performance in polymer melts, and how to tailor the rules for any chemistry of polymer chain. Nature is kind – it is possible to apply the generic science separately to styrenics, polyolefins and acrylates.