Mapping the Polymer Universe

First published in British Plastics & Rubber, August 2002
Article reproduced by permission from Clive Maier, Econology Ltd.
Copyright Clive A Maier 2002

Optimum Selection of Polymers | Polymer MaterialUniverse | CAMPUS and CHEMRES |
Philosophy | Browsing the database | Selection | Summary


Selecting an appropriate plastics material for a project can be done in a number of ways, none of them ideal. But Clive Maier has found a materials selector that just might be.

If you work in plastics, sooner or later you will find yourself specifying plastics. Sooner, if you are a product designer; eventually, if you are a processor. That is however much you may feel it is someone else who should say what they want, and just leave you to make the thing.

If this is going to happen anyway, you may as well decide to like it, and if you can become good at it there should be a profit. Buyers are looking for full service providers, suppliers who will take responsibility for getting a product into being at the right time, the right price and the right quality. Fragmented supply chains are a problem for them and the problem is that responsibility can disappear into the cracks between the supply chain partners.

So specifying the right plastics material should be a key skill for companies in the plastics community but how do you go about it? After all, there are tens of thousands to choose from. You certainly can't do it off the top of your head. So do you use generic data to find the right plastics family, then ask suppliers for a grade that fits? Or do you start with grade data and whittle it down to a shortlist by specifying key properties? Or do you use whatever you used last time, without ever knowing whether it is the optimum?

The optimum choice is what concerns Granta Design, a company backed by the University of Cambridge and ASM International, and noted for its Cambridge Engineering Selector (CES) software. The company has applied its selection methodology to the world of plastics and has come up with a new software tool, the Polymer Selector. This aims for the best of both worlds by combining generic and grade-specific data sources with very powerful graphical selection methods devised by company co-founder Professor Mike Ashby.

The Polymer Selector has been designed to deal with problems in our current selection methods. Most generic data consists of point values rather than the range of values that would more realistically represent the behaviour of a polymer family. And generic data is hard to come by for some materials, particularly the newer polymers and blends.

There is a different problem for grade-specific data. Grade coverage is good, even very good, and grade-to-grade comparisons are now practical thanks to a massive industry-wide effort of standardisation inspired by CAMPUS. The problem is that so much data is missing. If you look at any datasheet you are likely to find blanks against quite a few properties. When you make a selection search, you run the risk of screening out useful materials, simply because the data that would establish their suitability is missing.

The Granta solution is a completely new generic database for plastics. Its Polymer MaterialUniverse data module is believed to be the broadest such database, embracing thermosets, elastomers and composites as well as thermoplastics, and running to some 570 distinct types and sub-types.

More than 50 property values are given for each one, and these are ranges not spot values. Crucially, there are no blanks. Every generic datasheet includes every value.

When you start selection, there is no chance of a candidate material being overlooked because the property is missing. Of course, defining those value ranges was quite a problem for the developers at Granta. Most values have been taken from published sources, while others have been estimated from known physical values by methods generally accepted as reliable. All estimated values are starred for recognition and the data sources are declared. No values were calculated by computational chemistry because Granta sees the technology as too dependent for this context on the molecular make-up of the material. The generic values include the all-important price information that plays a key part in materials selection.

Two other data sources are particularly important for the Polymer Selector. Grade-specific data is supplied by the CAMPUS module. This contains the familiar point and multi-point data supplied by all plastics producers subscribing to the CAMPUS consortium, and currently stands at more than 5,500 commercially available grades. The other data module is ChemRes . This is derived from data on the chemical resistance of plastics collected over many years by Rapra Technology. ChemRes ranks the materials in the Polymer MaterialUniverse by their resistance to more than 190 different chemicals. You can work with all three data sources - Polymer MaterialUniverse, CAMPUS, and ChemRes - within the common POLYMER SELECTOR interface.


  Click here for a full size picture

Working with CAMPUS


This brings us to the three-fold philosophy behind the POLYMER SELECTOR. First in selection, you move from the general to the particular; in other words, from generics to commercial grades. Second, material performance is rarely governed by single datasheet properties but by combinations of them. This can be expressed by performance indices derived by mathematical analysis of the engineering problem. Another way is to map how materials perform when you compare key properties in pairs. You do this in the POLYMER SELECTOR with an Ashby chart and that brings us to the third point which is, see the whole picture. In conventional selection schemes you can see your hits but it is hard to know how near to misses they were, or how close to success the misses came. If this method is a page torn from the AtoZ, the Ashby chart is a sheet map of the whole city.

Let's see how it works in practice. In the POLYMER SELECTOR you can browse, select or search. You do this on datasets or tables that are arranged in tree-like hierarchies. We have already talked about Polymer MaterialUniverse and CAMPUS, and there are similar structures for suppliers, references and for the ProcessUniverse where all relevant plastics processes are set out.

Browsing is easy. You navigate through trees using the familiar Windows paradigm to expand or collapse branches. When you arrive at the end of a branch, for example at polyamide type 69 unfilled, a double click brings up the datasheet, and information on suppliers and data sources is just another click away. Searching is similarly straightforward; just a matter of entering a text string in a search window.


  Click here for a full size picture

A bubble chart selection in progress


As the name implies, selection is what the Polymer Selector is all about and it is here that the power of the system is concentrated. You begin selection by working with Polymer MaterialUniverse, the generic database. When you have isolated a few generic hits, you then select these in the CAMPUS module and refine the selection process to locate a few acceptable grades. Of course, CAMPUS is strong on thermoplastics but less so on thermosets, and of little or no use with foams or composites. If your generic hits are all misses in CAMPUS, the next stop should be the suppliers listed on the generic datasheets.

There are three ways to select in the POLYMER SELECTOR: by graph, limit or tree. You proceed by stages; as many as you like or as many as it takes. Each selection stage finds the hits, if any, against your selection criteria in that stage. The next stage works on the hit list of survivors from the previous stage. Tree selection is simple; you just navigate to a branch and click-select on that. It is an in-or-out selection so you would probably use it to make a broad cut, for example, to pick out all the thermoplastics that are suitable for injection moulding.

Selection by limit involves setting a value for a key requirement. This is either a matter of clicking a check box or typing in maximum and minimum values for a chosen property. No guide values are given, for example the range available in the current hit list, so you do have to know your stuff to select by limits. The hard fact is though, that if you can't define your service requirement well enough to enter limits, you can have no assurance that your design will work as planned. If defining those limits seems tiresome, now perhaps is the time to consider the cost of product failure.

Now we come to selection by graph, and this is where the real power lies. You can plot any pairing of numerical, discrete, logical or link attributes on selection charts. Depending on the combination you pick, you will finish up with either a table-like count chart, or a bar chart, or a bubble chart. The bubble chart is the classical Ashby selection chart. This is what you get when you compare two numerical values, like price and fracture toughness. Each material in the Polymer MaterialUniverse has a range for each of these values. The bubble that represents each material is the ellipse that circumscribes those values when plotted on the X and Y axes.

To begin with, the bubble chart shows the entire population of the Polymer MaterialUniverse. Now you make a selection with the mouse by dragging a selection box over the portion of the chart that interests you. In our example, you would probably want low price and high fracture toughness. As you drag the box, the materials bubbles that fall within the box are highlighted and the hit list updates in real time. Click on any bubble and its identity comes up in the style of a tool-tip. You can make the chart easier to read by zooming in on any region, and by hiding all the 'failed' material bubbles.


The Polymer MaterialUniverse tree structure


Now, because you are looking at the sheet map and not the AtoZ, you can see just how your hits qualify and you can check up on the near misses. Even better, you can drag the selection box about and resize it at will. As the hit list updates in real time you can see exactly how materials pass in and out of contention. For instance, you can see just what compromise must be made to bring a favoured material into the running.

You can do a lot more with these charts. You don't have to accept the built-in datasheet attributes for the X or Y axes: you can define your own. In our example, we used the datasheet price which is given in £/kg. Instead, we could use the expression price times density to give the rather more revealing cost per unit volume. If that now seems like a good idea, it is easy to put it into action. You can go back at any time to any of the selection stages and edit them to modify the criteria.

The selection box is not the only way to work with bubble charts. You can use a gradient line or a point line instead. You draw a gradient line by specifying its slope and clicking on a point that the line must pass through. A point line is drawn by clicking on two points that the line must pass through. In either case, bubbles on one side of the line pass and those on the other side fail. You choose which is the pass side. Once again, you can drag these lines about with the mouse to alter the selection, and the hit list will update in real time.

You build up these selection stages until you have defined the key requirements of the design. You will either have no hits at all, in which case a design rethink is required, or you will have a manageable short list of suitable generic candidates. Now you switch to the CAMPUS module and select just those generics as a starting point. This is likely to give you a fair number of grades. You can now work on those using exactly the same selection tools and techniques until you finish up with a shortlist of grades, preferably from more than one supplier. The CAMPUS dataset properties do not correspond exactly to those of the Polymer MaterialUniverse , so some adjustment to the selection criteria may be necessary. The other difference you will see is in the bubble charts. CAMPUS has point values so you will see materials represented by small uniform dots instead of ellipses in all sizes.

There are other polymer selection tools on the market. Rapra's Plascams has been the long-time leader in generics. On the grade-specific front, MCBase, the CAMPUS database consolidator from M-Base is prominent, while for comprehensive coverage the Prospector products from IDES take the prize. But POLYMER SELECTOR is the only tool that combines both generics and grade-specifics, and it is the only plastics selector to use the powerful and instructive Ashby bubble charts. These are well established in other branches of materials engineering and they seem destined to make their mark in plastics too.