A glass act

4 min read

Bristol university researchers employ aerodynamic levitation technique in an attempt to manufacture containerless glass at speed. Stuart Nathan reports.

A process for making glass without a container could lead to new insights into how glasses are formed as well as improved mass-production techniques, according to a team from

Bristol University


The process, which can make samples of glass in minutes, could also mean the development of new types of high-performance glass.

The team, led by Adrian Barnes, is studying the properties of glasses formed at very high temperatures. 'If you heat things up, you expect to put your sample materials into a crucible or container of some sort, and put it into a radiative furnace,' he said. 'But the problem is, the higher the temperature gets, the more difficult it is to find a suitable furnace. You have the additional problem of what to put the sample in.'

There are two main problems with containers, explained Barnes. First, the team is making glasses which need temperatures above 3000K, and there are a limited number of materials that can withstand these temperatures. 'Even with very good containers, you get impurities coming in, and they can contaminate the sample,' said Barnes.

The second is in the nature of glassmaking itself. Glasses are amorphous solids, with the molecules arranged in a disordered state somewhat similar to liquids. To make a glass successfully, the ingredients have to be melted together, then cooled or quenched in such a way that the sample solidifies without crystallising. Once a crystal forms, it tends to encourage the formation of even more crystals; and one of the factors that encourages them to form is surface features in the container that holds the molten sample.

This is known as nucleation, and it's a particular problem with crucibles, said Barnes. 'Crucibles tend to be ceramics, which have a lot of crystalline areas on their surface where nucleation can take place,' he explained, 'and that can stop you forming a glass at all.'

The answer is to go to containerless production, where the sample literally floats. Previous attempts at this have used electrostatic fields to support the sample, but this only works if the sample is very small and metallic; it's also very difficult to control. Barnes's group is instead using aerodynamic levitation - a technique similar to the effect of blowing a stream of air beneath a table tennis ball, although the physics is a little different.

'We sit the sample on a conical nozzle and pass a stream of pure argon gas underneath it. That lifts the sample a few tens of millimetres off the surface - not very far at all, but enough for us. The gas effectively acts as the container.' The equipment can handle samples up to 2-3mm in diameter, he said.

The heat for the system is supplied by a 240W carbon dioxide laser, shining vertically downwards on to the sample; the amount of energy can be controlled by changing the diameter of the beam. Temperature measurement is a problem, however. 'Obviously, you can't put anything directly on to the sample, but once we're above 1,000


C it starts glowing. We have a pyrometer which determines the temperature from the emitted light.'

One of the team's goals is to make glasses which haven't been possible before. 'We've made some aluminates which previously have nucleated too fast and crystallised completely. A traditional glassmaker would take the sample out of the furnace and pour it on to a cold block to quench it, so there would be a risk of nucleation on the block.

'With us, if we want to get a maximum quench-rate we just turn the laser off.' The heat then radiates away from the sample rapidly, close to the condition known as black-body cooling, a theoretical model of perfect radiation of infra-red. The team can achieve cooling rates as fast as 500K/sec, said Barnes.

Another goal is to investigate the processes that occur as glasses form, which could have a considerable influence on the development of new glassmaking techniques.

Monitoring the temperature of the samples carefully during both heating and cooling is providing considerable insights, said Barnes. 'If you know the power going into the sample and the absorption of the energy is constant, then you should be able to work out the temperature.' he said.

'As we do this, we can see things like phase transitions taking place in the sample. For example, if you take a bead of aluminium oxide and cool it to 300-400


C below the melting temperature, it suddenly flashes light at you, re-radiating all the latent heat of fusion as it suddenly solidifies.

'That means that it's crystallised, so if you can ensure that it doesn't do that, it means you've made a glass. But it also means that you can look at glass transition in real time. Nobody has done that before.'

The team is also trying to measure the electronic properties of their glasses, which could help in the development of electrical insulators for the electronics industry. Again, new technology has been developed by the researchers to allow them to take precise measurements.

The team is measuring the conductivity of the samples, using a coil carrying a current. When a conductor is placed inside the coil, it changes the way that electricity flows - 'it's like a glorified metal detector,' said Barnes. But the high temperatures make it difficult to take measurements, because the coil warms up along with the sample and this affects the electricity flow.

'That swamps the system,' explained Barnes, 'so we've developed one that oscillates the coil up and down. If we oscillate it fast enough, the effects of the drift in temperature are smaller than the timescale of the measurement we want to make.'

The system itself is only suitable as a production process for very small quantities of glass, said Barnes, but it makes a highly-effective test bed. 'If we find a sample in the lab with optimal properties for a particular application, we look at how we made it and see if it can be made by conventional means.'

'And one major advantage we have for developing new glasses is the speed. Conventional methods using a furnace can take days. We can make a sample in five minutes.'