The weird world of water at the nanoscale

We all know what happens to water when it freezes - it becomes ice. But it does some weird stuff as well, and there are still lots of unknowns about how this sort of transition...
20 June 2010

Interview with 

Angelos Michaelides, London Centre for Nanotechnology


High-power shot of a hexagonal snow crystal.


Chris - Now we all know what happens to water when it freezes.  It becomes ice, but it does some weird stuff as well because the ice is less dense than the water and it floats, and there are still lots of unknowns about how this sort of transition from water to ice actually happens, especially on the nano scale.  So understanding this could help with a wide range of different industry applications, as well as our predictions of what will happen with climate change.  So this week, Meera Senthilingam went along to the London Centre for Nanotechnology to meet Angelos Michaelides to find out how his team are looking at how water turns into ice.

Angelos -   There's really many, many interesting things about water.  It's an everyday substance - we're made of water.  It's all around us.  Our Earth is covered in water, but we still don't understand many of the properties of water at the molecular scale, and one Hexagonal Ice Crystalof the processes we're trying to understand is just how water freezes; how individual water molecules come together to form little ice particles.

Meera -   Well why does ice form in the first place?  So why does water become ice?

Angelos -   When we're below zero degrees, ice is simply more stable than water.  Most of the times when water freezes, it happens at the surface of some impurity particle, and so, it's the actual presence of these little impurity particles, maybe minerals or clays or dust particles that actually allow you to nucleate ice.  The water will stick to the little particles, form a little ice particle, so this is a nucleation process, and then once you have a small ice particle, just like in crystal growth, it's very easy for more water molecules to stick to the little crystal and to grow, and to expand.  So we really want to know how exactly, looking really at the level of individual molecules how this ice forms.

Meera -   More specifically, you look at this formation of ice on particular materials or solids such as metals rather than just how ice forms in the first place.

Angelos -   Yes, so water really is one of the most incredible substances that we have.  It occurs in our environment in all three phases, solid, liquid, and gas, and in the solid phase, there are 15 different phases of ice alone.  And so, you can imagine when you introduce some other material like a surface, then you get all of this beautiful array of complex and interesting behaviour.

Meera -   How do you set about looking at this?

Angelos -   One of the techniques that has really transformed our understanding of water at interfaces at the nanoscale, is the scanning tunnelling microscope, and this is an incredible technique that allows you to see individual molecules and atoms at surfaces.  And then we will take this experimental insight that comes, and we will try and build models to really understand at the molecular scale how these structures form, what structures form, and why.

Meera -   And so, what have you manage to find out so far then?

Angelos -   Well one recent example that was interesting because it went sort of against common belief was where we were looking at water on a copper surface, and we noticed that water forms chains.  And when we did our models to understand what is the structure of these chains, we noticed that the water is actually forming pentagonal arrangements.  So it's actually forming chains built out of water pentagons.  This is interesting because we know from daily life that snowflakes come as hexagons and have this hexagonal motif, and out of all of the 15 different ice phases, none of them are actually built out of pentagons.  If we go to another surface for example, if we go to a palladium surface, we don't form pentagons, but we don't form traditional ice-like structures either.  We form something that looks like really beautiful rosettes, and a lacy structure. As the structure of the humble snowflake attests to ice crystals normally come in hexagons. However, our simulations (along with experiments from Andrew Hodgson and co-workers at the University of Liverpool) show that ice can come in pentagons too. The movie shows an ab initio molecular dynamics simulation of a chain of water pentagons on a Cu surface. To learn more see or Carrasco et al, Nature Mater. 8, 427 (2009).

Meera -   So that's really quite fascinating that it just varies so much on different solids or different metals in this case.  Now you've also been looking at the formation of ice on clay particles which is important for understanding our atmosphere.

Angelos -   Yes, so when we're starting the formation of clouds, this typically starts off with a super cooled water droplet.  So a small micron-sized water droplet that is already below zero degrees, but it won't freeze at super cooled.  This freezing process is helped by a foreign substance which in the atmosphere, quite often is a clay particle.  Little nano-size clay particles typically float about in the atmosphere.  They get there from dust storms in desserts.  So we're interested in understanding these clay particles and how they make the rain, and make the snow because without these agents that would nucleate the ice, it simply wouldn't rain.  Outside the tropics, we require ice nucleation to have rain.

Meera -   I mean, it's clearly interesting to understand this simply to know why it does rain as you've just mentioned, but what are the applications of potentially of knowing this about the formation of ice on so many levels or on so many materials?

Angelos -   In the atmosphere, we want to know how and why ice forms if we want to understand issues like global warming to understand the impact of ice and clouds, and water in the environment, and how it impacts upon - say for example, the temperature of the globe.  Aside from this sort of environmental interest, there's all sorts of technologies who are interested in ice nucleation. 

One example would be the airline industry.  Their planes, as they're flying through these super cooled water droplets at minus 30 degrees, it's incredibly useful that ice doesn't form on the wings, on the fuel lines, all over the planes.  Like that plane for example that crash landed a couple of years ago at Heathrow.  It did so because ice formed on one of the fuel lines which cut off one of the engines.  So there's an industry who are very worried about ice nucleation.  Ice cream manufactures are also interested in this.  Ice cream doesn't taste very good if it's covered in little icy particles.  And going back to the atmospheric connection, it's incredibly important for these people who want to control rainfall, who want to so-called 'seed clouds' to make it rain.  They're interested in knowing what is an efficient material at making ice so as to make rain, and one of the most widely used materials is silver iodide because it has a structure which resembles very closely the structure of ice.

Meera -   So that's quite a wide range of applications.  So, what's the next step in this research then, in order to help this wide array of industries?

Angelos -   At the moment, we can identify what structures form and we can understand why, but we're not at the stage that we can predict.  So we would ideally like to be able to predict and to say, "Okay, you give me a material and I'll tell you in advance how it will behave", to try and identify the basic principles and the basic trends that control ice nucleation at interfaces.

Chris -   Who would've thought the most common molecule on the earth within - well, a few words of magnitude, could be so complicated.  That was Angelos Michaelides.  He's from London Centre for Nanotechnology which is a collaborative institute between UCL, University College London and Imperial College London.  He was talking to Meera Senthilingam.


Add a comment