The Structure of Faujasite

By Nancy Eaton

Dr. David C Palmer, founder of CrystalMaker Software, talks about the structure of faujasite.

What are we looking at here, and why is this material scientifically significant?

We’re looking at several different representations of parts of the crystal structure of faujasite: a type of mineral called a zeolite, found naturally in sediments such as at the bottom of the sea. Although they’re not that important geologically, zeolites have very useful properties, which make them important in industry, and they can now be created artificially.

Zeolites are microporous materials. They are crystalline, with ordered arrangements of atoms, which form a characteristic network of cages and channels that work like micro-sieves. Zeolite holes are tiny — perhaps a billionth of a meter, but chemists discovered that they could change the size of the hole artificially, by altering its chemical composition. So you suddenly have a whole class of material that can be manipulated to form useable products that can sieve at the molecular level or provide miniature reaction chambers to help speed up chemical reactions, and so act as catalysts in the chemical industry.

Why are computer image models such as these so beneficial in the understanding of zeolite crystals?

In order to understand exactly how the structure and the chemistry control the properties of these materials, we need to start thinking about what are the important parts of the structure, and that’s where the models come in. And there are, of course, many different ways in which materials can be visualized.

The arrangements of the atoms that make up these structures are very complicated, and what’s more, these are three-dimensional structures, which are difficult for many people to visualize. We needed a better way to see them, to de-focus and try to think about their coarse structure, and that’s one of the things that the CrystalMaker software can do.

Can you take us through a few examples?

Image 1 is a traditional ball-and-stick representation of part of the faujasite structure. We have red oxygen atoms and blue metal atoms (which could be silicon or aluminum) which are chemically bonded, as represented by the sticks. However, these models can be misleading, since the spheres aren’t actually the correct relative sizes of the atoms, and in any case, the model is too complicated to really see what’s going on; this is where the CrystalMaker program comes in.

For these types of structures, we’re not really interested in where all the red oxygen atoms are; instead, we want to focus on the metal atoms, since these are important for determining the zeolite’s properties. You can use CrystalMaker to hide all the oxygen atoms and just focus on, say, the relative positions of the metal atoms, which form a three dimensional framework, or mesh. We can then go one step further and say, well, part of this mesh consists of cavities, which are basically three-dimensional holes. So a better way to think about the cavity is as we show it in Image 2, where we show some of the smaller holes filled in, as solid objects, to create what I would call a polyhedral model. It’s a coarse overview of the different cavities. In Image 3, the atoms are actually plotted at their correct relative sizes. This is a space-filling model, where we can see the true relative sizes of some of these cavities in the structure, so we can determine what kind of molecules could be inserted into this structure.

How were these images used in research?

A few years ago I collaborated with Professor Martin Dove at the University of Cambridge Earth Sciences Department. He uses computer modeling to calculate the stability of structures, including minerals such as zeolites. For these calculations to be physically meaningful, he needs the computer to work with very large numbers of atoms: so many, in fact, that it gets difficult to see what’s going on. To overcome this problem we created a new feature in CrystalMaker 7.1 for Mac, called “depth-profiling.” It speeds up the computing process and allows scientists to look right inside a complex material in ways they could not do before, scanning thin slivers of material at different depths.

We used this faujasite model as an example to test out the depth-profiling tool. We can scan right through the model, with atoms fading in or out with changing depth. In our interface, we have a slider bar control, and you can smoothly move up and down to see the way the zeolite cavities change at different depths.

Are there other scientific uses for the depth-profiling tool?

Besides being used to look inside structures, we were surprised to learn it’s also being used by scientists interested in crystal surfaces. The way the atoms are arranged on the surface of a material can change according to the orientation of that material and the height at which the surface is created. Scientists can now use the depth-profiling tool to sample different orientations and heights.

There’s also interest in using computers to predict the structures of materials, whether they be crystals or glass or basically solid-state materials. The problem with this, as in Professor Dove’s situation, has always been that you had to ask the computer to work with very, very large numbers of atoms. If you plot millions of atoms on a screen, you just see a big blob, so how do you see into the blob? This depth profiling application lets them scan through a material very quickly for specific areas of interest just by clicking and dragging and moving up and down through the material.

Where do you see this kind of visualization evolving in the near future?

It would be instructive for scientists to be able to see calculations happening in real time, to visualize the different stages. Crystals aren’t really static because the atoms move all the time. So an area we’re interested in is the dynamics of crystal structures and how the atoms move. In the near future we may be able to look at a whole range of structures very quickly and see some of these changes in crystal lattices in real time.

It would also allow people to do “what if” experiments. For example, what if I’m trying to understand something about the catalytic properties of a microporous material such as this faujasite structure? It’s a material with large holes in it, and I want to know exactly where my molecule is going to fit. So I could click on the molecule with the mouse, move it and let go, and the computer could calculate where that molecule is most likely to fit. Those are the sort of calculations we can do now and view the final result, but what would be really great is if users could see it happen interactively.