How do you model dark matter?

Now, normally, when scientists want to get to grips with how something works, rather like a small child armed with a screwdriver and a toy, we try and take it to pieces. But the problem with dark matter is that, although we can see its effects, we can’t actually detect it or manipulate it. So how can we understand it? One way is to work out what it must be doing by using computer simulations. You tell the computer what dark matter does in some particular way and then ask it to predict how the Universe would look if that were true. And then you can come up with ways to test that theory in the real Universe. Colin DeGraf is a theoretician who does work like this at Cambridge University, and he spoke to Chris Smith...

Colin - Most of what I work with is what's called a cosmological simulation, so we don't try to represent single object in the simulation, instead we're simply modelling a huge volume of space that's representative of a very large amount of the universe. So we’re talking about boxes hundreds of millions of light years across that will include millions of galaxies and we evolved this starting from when the universe was only a few million years old, and we see how this evolves over the course of the entire age of the universe.

Chris - What ingredients do you put into your model in order to see that evolution?

Colin - There are entire different scales of ingredients you want to include. As you just heard, most of the matter in the universe is dark so we only need to include dark matter to be able to do a good job of modelling the large-scale evolution of the universe.

Chris - Your rationale is because this makes up the majority of the mass in the universe, if we just play with that to start with we're not leaving out much.

Colin - Exactly. And as long as you're only looking at the large-scale that's fine, but once you start including the formation of galaxies, the collapse of all this matter into what is comparatively small compared to the superclusters, you start to reach a point where you need to include baryonic matter -  the gas, dust and stars, and once you include that it affects how the dark matter collapses as well.

Chris - So you mean your simulations, while they were doing quite a good job, once you start to add the extra things and look slightly more in detail you start to see wrinkles that shouldn't be there, so you know you're missing something?

Colin - Yes, though in some ways it has even been the reverse where you start with dark matter only and then people realise “wait a minute, when you start looking at the dwarf galaxies we're missing some of the dwarf galaxies in the real universe, and the density of the material at the centre of these galaxies doesn't match” so we start having to add extra material into our models.

Chris - Does this begin to constrain ideas about how dark matter must be behaving and what its properties are, because when you make these models in order to get something which is representative of what we see when we gaze out there into deep space at the structure and fabric of the universe, the only way we can explain it is if there are additional properties to dark matter that we currently can't account for?

One of the problems is you can tweak the dark matter model and you can start to better explain some of the small scales at late times, but if you do that and it becomes hard to explain the very early universe. And if you tweak dark matter only to match the early universe then you can't explain what we see now.

Each time you tweak these models you end up changing what you predict to observe in the universe. So by continuously modifying both the dark matter and baryonic or normal matter, we can start to make predictions on how we would expect things to look like in the real universe. And so each one of these models gives us new predictions and then as we have new telescopes and new surveys taking more observations of the universe, we can actually rule some of these models out.

Chris - And what have you been able to exclude? What major predictions have now been able to be laid to rest or put to one side saying no, dark matter definitely doesn't do that, it has to be one of the above?

Colin - Truly ruling something out can be tricky with a simulation but what we can do is make a lot of predictions and so one of the things we look at is how dark matter behaves. The initial simulations we would just assume that the dark matter interacts gravitationally and that's it. Dark matter doesn't collide with itself, there is no other interaction there's just only gravity and that does very well at the large scales, at the small scales it doesn't and so some people will propose that you need warm dark matter instead of cold dark matter.

Chris - And what's the distinction between the two?

Colin - It's essentially just a matter of how massive is a particle of dark matter. Very massive dark matter would be very cold and less massive will be hot, which is the physicists way of saying basically what's the typical speed it's moving at in the early universe. Cold dark matter does a great job of the large scale, hot does not, and we know that the large scale very well and adding normal physics of baryonic material doesn't help. And at the small scales that's where you have a lot of very dense gas and star formation and that's where we really need to do a better job linking both the dark matter and the baryonic physics. And it looks like a lot of the small scale can actually be explained by the normal matter which means our current model strongly suggests that this dark matter particle would be relatively cold.

Chris - You've brought up the idea that dark matter's a particle, would it be reasonable to summarise then saying that the universe is dominated by this mysterious thing, it's some kind of particle, it has a lot of mass because it's very effective at exercising gravity and having a gravitational influence on things, but beyond that we don't actually know anything else about it?

Colin - For the most part, yes. But there have been a lot of different proposed explanations for what this particle could be, some of which you might not even classy as a particle. Initially, people proposed may be they're primordial black holes or they're very old stars, and we have been able to rule those out.

And as we rule these models out better and better, what we're left with is looking like this dark matter should really be explained by what we would consider to be this cold, primarily collisionless, dark matter particle.