Skip to main content Skip to main content

Physics for the 21st Century

Dark Matter Interview with Featured Scientist Rick Gaitskell

Interviewer: What is an unanswered question in physics and how does your research address that question?

RICK: At the end of the 20th and the beginning of the 21st century, one of the most challenging questions that we’re facing is quite simply what is the composition of the universe? You would have thought that was a question of such simplicity that it had a simple answer and that we already knew it. Unfortunately, neither of those statements are correct.

It is a simple question but the answer, it turns out, is not simple. It’s probably one of the most technically challenging pieces of experimental cosmology we have ahead of us right now. However, we are absolutely convinced that we have the technology in order to discover what the majority of the matter in the universe is.

It is not you and I. The material that you and I are made of is not the dominant material in the universe. In fact, less than 1/20th of what the universe if made of, is made of conventional material. The rest of it is in the forms of substances, energy and matter that we currently know nothing about.

We’re working on the dark matter problem which is the title that we have for the reality that most of the matter in the universe is yet to be seen. It is non-luminous. It’s in the form of some new particle that we have yet to discover.

Interviewer: Can you give a summary of dark matter research?

RICK: The reason that we are aware of there being extra mass in the universe is because of the gravitational affect it has both on our own Milky Way galaxy, and in every observation we make of extragalactic phenomena.

We see the effect of the mass indirectly in the form of the way in which it causes one body, in this case a body being something as substantial as a galaxy, to move around another one. But yet when you actually try, for instance, to count the number of stars in that galaxy you find that is a very small fraction, a few percent, of the apparent mass of that galaxy.

The idea that galaxies weighed very much more than they appeared to just from counting the number of stars or, more technically, from looking at their luminosity, was something that was first experimentally determined nearly 80 years ago by Fritz Zwicky. He was making observations of the coma cluster. It was established at that point how bright the coma cluster was, therefore how many stars were in the various component parts.

But he was also crucially able to make a measurement of their velocity. If you measure the velocity of a particle, or in this case a galaxy, you are able to infer if it is gravitationally bound with other objects. You’re able to infer what the gravitational force is and if you are able to infer what the gravitational force on these objects is then you can infer what their parent gravitational masses are.

The result you get, and Zwicky found this, was that if the coma cluster really was held together by gravity, then the amount of mass that would be required to hold it together was over ten times greater than, at that time, the estimate of what the coma cluster must weigh, just due to counting the number of stars or the luminosity in that cluster.

That was when really I think the first suggestions that there was an awful lot of matter in the universe that was non-luminous first came to the fore. Since then, we’ve seen a succession of measurements, which have all basically led us to the same conclusion. For instance, when we look at how fast galaxies rotate, they appear to rotate substantially faster than you would infer simply from calculations that use the amount of luminous matter in those galaxies.

As you get to larger and larger radii this anomaly becomes more and more substantial. So you have to conclude that the majority of the mass, in not just our own galaxy, but all other galaxies, is dominated by some non-luminous dark matter.

Interviewer: How well do we understand dark matter?

RICK: The title “dark matter” tells us one of the key properties. It’s some form of matter, which does not interact with light. That leads us to immediately understand that it cannot be charged. If dark matter is charged we would have already seen it in conventional astronomy. So, that’s the first sort of criterion.

The next criterion—and it’s quite a technical point—is that when we’re making models of the universe, we need to be able to explain how it is that the structure that ultimately became our own Milky Way, or became the other galaxies formed. We do know that in the early universe and the Big Bang, initially that matter was distributed very homogeneously.

It was a very smooth distribution. There were small ripples in it but they were very small. They grew over time due to the fact that, as they say, gravity sucks. Therefore, if you get a small over-density of material in a smooth distribution it tends to continue to form. In order to really look in detail at the dynamics of how structure goes from this initial small ripple, to have the tremendous wealth of structure we see today (in the form of galaxies and galaxy clusters), the quantitative models that we have derived say that you need a thing which was very ready to sort of fall in and make structure, as gravity worked its process over billions of years.

That puts a constraint on this material, that it has to be, in our terminology, cold dark matter. What the term cold simply means is that the speed at which these particles were moving in the early universe was fairly low. Still fast by any reasonable terrestrial standards, but in terms of for instance the obvious benchmark of the speed of light, these particles were extremely slow. So, we also know that this dominant dark matter we’re looking for has this property of being cold.

Particle physics, it turns out, is very ready to provide an answer for what or how it is we might be living in the universe that is composed for the majority of neutral, cold particles. This comes, for instance, from the field of higher energy particle physics where we are pushing the boundary towards super symmetry. The models of super symmetry very readily provide mechanisms for generating a large number of cold, neutral particles in the early universe, and yet ensuring that these particles are still around today, and that they are now formed into these halos.

The vast majority of this matter is in the form of this dark halo, but because it’s neutral we can’t see it and, therefore, we have to rely on new, innovative experimental techniques in order to provide direct evidence for the existence of the halos.

Interviewer: What is the motivation for your research?

RICK: This is an appalling admission because it means for the last 20 years I’ve been working in the same field. I had done all the reading and became aware of this question of missing matter in the universe, the dark matter problem, and went straight in as a graduate student developing a detector in order to look directly for dark matter. That, in fact, is where I find myself still 20 years later.

Interviewer: What is the leading candidate for dark matter?

RICK: The dark matter has to satisfy a number of properties. It has to be electrically neutral. It has to be dominantly cold which means that the velocity of distribution of these particles is very slow. It has to be produced in the early universe by some kind of mechanism, which usually stems from particle physics.

We find ourselves in a situation where there are many possible candidates for dark matter. I think it’s fair to say that the most favored candidate that we currently have, which seems to be extremely strongly supported by particle theorists, is what’s known as the WIMP, the weakly interacting massive particle. These WIMP particles would be a product of the early universe, the Big Bang.

Interviewer: Has a WIMP ever been detected?

RICK: Even though we have been looking for weakly interacting massive particles for at least 25 years I think it’s fair to say at the current time we’ve yet to have had a sort of confirmed observation of such a particle. Now, that is completely consistent with theory. It turns out that we still don’t know enough about weakly interactive massive particles to say precisely how likely it is for them to interact with conventional material.

Interviewer: How do you detect dark matter?

RICK: The way in which we’re trying to directly detect dark matter simply relies on taking a detector which is extremely sensitive and able to pick up the very occasional interaction between the dark matter particles that are pouring through us all of the time. If I hold my hand up right now the flux of these dark matter particles that are confined in the Milky Way in these sort of huge orbits, there are hundreds of millions of these particles going through my hand every second.

What we have been able to calculate is that there is a finite probability of one of these particles or some of these particles interacting with the conventional material we’re made of. Now, when I first started in the field 20 years ago we thought the rate was such that if you took a kilogram of material that the interaction rate would be at the level of say 100 or 1,000 events per day of one of these dark matter particles interacting with one of the atoms or, in fact, the nucleus of one of the atoms of a target. So we designed experiments to look for these very small amounts of energy being deposited.

Interviewer: Where is your experiment located?

RICK: The LUX (large Underground Xenon) experiment is being constructed at Homestake Mine in South Dakota. It relatively recently came into being. It was up until about 2002 actually the largest gold mine in the United States. In 2002, the operating company decided to shut it down. So, we have the opportunity to do an experiment right now at the 4,850 foot level, which for the type of experiment that I’m doing represents a significant increase in the depth. This allows us to have an experiment which performs even better, because every time we go that little bit deeper we are shielding out more cosmic rays and, therefore, keeping one of the many backgrounds that potentially stops our experiment from seeing dark matter. This potentially allows us to reduce that background below the signal that we’re looking for.

We intend to run the detector at a depth of 4,850 feet within the mine. Last year that level was still nearly 200 feet under water, so another component of the rehabilitation or the reopening of this new laboratory, the Sanford Laboratory, has been a tremendous program to pump over 150 million tons of water out of the mine in order to reduce it to a level that we can get back into the 4,850 level.

Interviewer: Why are dark matter detection experiments conducted underground?

RICK: In looking for a signal from the weakly interacting massive particles, we find it is an extremely rare signal. Our current theoretical estimates are that a detector the size of LUX—of which the target volume is in order of three-quarters of a meter by half a meter—that the event rate may be at the level of one event per day. It may be at the level of one event per month or per year from a weakly interactive massive particle.

In order, therefore, to be able to see such small event rates we need to reduce all other possible sources of events. We have taken the LUX detector deep underground, in order to try to remove cosmic rays, which on the surface, are streaming through my hand at the level of about one or three events per second. When we take ourselves deep underground, the rate drops to a little bit more than about one event per year through my hand. That is an extremely important aspect of being able to see the WIMP signal.

In designing the LUX experiment, one of the things we realized is that if you take a large volume of xenon in order to look for very rare events, one of the great strengths this kind of detector has is that the outer layers of xenon shield or stop particles in the conventional background from getting into the central region. So, we are now constructing a detector which is nearly a third of a ton in mass and by doing that it means we can, for the first time, really get a level of radioactivity or absence of radioactivity in the core of the xenon that will simply be unparalleled in previous experiments that have taken place.

What we have is a very quiet, if you like, inner volume of xenon protected by all these outer layers of xenon. And what we hope will happen is that while all the conventional radioactivity that is trying to get in from outside gets ranged down in all these layers of xenon, that dark matter particles, which are highly penetrating and will pass through the earth without even noticing they’re there, then become an extremely clear signal as a residual event rate in the core of this detector.

Interviewer: How does a dark matter particle reach your detector?

RICK: The dark matter particles, the WIMP particles, are in orbit in the galaxy. They are all each individually probably on these orbits that take hundreds of millions of years for them to make a single lap—in the same way that the earth or the solar system is on an orbit that takes nearly a quarter of a billion years to make its way around the Milky Way.

These particles are moving at a few hundred kilometers per second. But let’s, for instance, think about one special particle that happens to be on such an orbit around the Milky Way that it passes not just through us here but actually down through the earth. Again, it’s so weakly interacting that it barely notices the existence of the earth. But it just happens by chance that this particle passes through one of the detectors that we have set up, the LUX detector, in the Sanford lab, Homestake Mine laboratory.

It is moving at a few hundred kilometers a second, which is obviously very fast. But it’s relatively light. In fact, it weighs about the equivalent of a typical atom. It interacts with the nucleus of our one of the target nuclei in our detector and causes a nucleus to recoil. The amount of energy that is imparted during this recoil is a small amount of energy.

It’s a relatively modest amount of energy but because it’s imparting directly to the nucleus of one of these atoms it actually knocks the nucleus sideways by a distance of say a few hundred nanometers (a nanometer being a billionth of a meter). That’s in the order of a thousandth width of a hair. So, it’s a very localized, very small scale effect. But we have specifically designed our detector so that this modest amount of energy being imparted just into a single nucleus and causing the nucleus to recoil, that nucleus goes on to interact with other atoms in our target generating a small amount of light.

That light is picked up by an array of photomultiplier tubes, photodetectors that we have around the periphery of this target volume, so we are then able to see a small flash of light. Now, that’s not the only thing that the recoiling nucleus does. It generates, by interacting with other atoms around it, a small amount of light but it also generates charge.

It actually causes some of the electrons associated with the atoms to become disconnected from their core nucleus, and those electrons will then be drifted in our detector and could be measured as well. So, we have designed this detector to both detect the initial light from the interaction, and also to effectively detect the charge liberated as well. By having more than one detection method, it turns out we can tell a lot more about the nature of the original interaction. It’s incredible when you consider how little energy is being deposited, but you can actually tell precisely where in the detector, by the combination of this light pulse and the measurement of the ionization, the interaction occurs.

Interviewer: What is the target material made of?

RICK: The target itself is made of liquid xenon, which scintillates when small interactions occur in it. We then take this detector inside a large nearly eight-yard diameter tank of water. We then put that tank of water about a mile underground, keeping away the cosmic rays. The result is that we have a central volume of xenon, which is extremely quiet, with very few interactions taking place in it, even on a time scale of months.

In the middle of this detector there are very, very few interactions occurring. So what we’re able to do is to watch that target volume of xenon and try to see whether there is some additional signal in there. We can go look for these events at a frequency that may be as low as one or two events per year in the experiment. That’s the technical challenge.

Interviewer: What is noise in your experiment?

RICK: In the LUX experiment we’re looking for the extremely rare particle events from these WIMPs, so we have to work very hard to eliminate any possible signals in our detector that could be misinterpreted as being WIMPs. Now, there are an enormous number of possible sources of, if you like, fake events. It could be simply related to cosmic rays. In order to remove those we go deep underground. It could be due to radioactivity coming from the rock or people in the laboratory. In order to remove that effect we use an eight-meter diameter water tank and place the detector in the middle.

Interviewer: Why do you use a water shield?

RICK: It’s almost like a large swimming tank, except that the water itself is of extremely high purity. It is millions of times more pure than the water that comes out of your tap, or sea water. It’s very low in anything other than water itself.

Now, water is not radioactive. By removing all the impurities we end up with a very, very clean environment into which we insert or lower the detector itself. Through this multi-layered strategy we can end up with an extremely low background event rate in our detector.

Interviewer: Describe the LUX detector.

RICK: The LUX detector is nearly a third of a ton of liquid xenon surrounded by photomultiplier tubes and a cryostat. The previous experiments that we’ve done were considerably smaller—about 15 kilograms of material. We’ve now grown the detector up by over a factor of 20.

What this allows us to do is to use not only the fact that the detector, being that much larger, is able to see events more readily. But that it also provides, like an onion, layers of shielding. We can use the outer layers of the xenon in order to shield the inner layers, where we get the most dramatic contrast between signal and any potential sources of noise.

Interviewer: What is the temperature of the liquid xenon?

RICK: This requires us to be at about -100 degrees Centigrade. At that temperature the xenon forms a liquid. If you cooled it just a few degrees further, the xenon would then turn into a solid. So, one of the challenges of our experiment is in fact to hold the xenon at a temperature that is just cold enough that the gas has turned into a liquid but not turned into a solid. In water, for instance, there are 100 degrees Centigrade between freezing and boiling. For xenon the difference is much tighter and therefore it’s more of a challenge to hold the liquid precisely.

Interviewer: What happens when a WIMP interacts with your detector?

RICK: The LUX detector is designed to look for the very occasional interaction of one of these WIMP particles that is in orbit in our galaxy. Most of these WIMPs pass through the detector and do absolutely nothing. But very occasionally one of these WIMP particles will interact with one of the nuclei in our target, letting off a flash of light which we then detect using the photodetectors in the LUX experiment.

Interviewer: How do you interpret the data from photodetctors?

RICK: The experiment is designed that the third of a ton of liquid xenon is being watched by a 120 photomultiplier tubes arranged both above and below the liquid. When a WIMP particle comes in it interacts with a nucleus, causes a flash of light to be emitted, which is then picked up by those photomultiplier tubes.

The photomultiplier tubes are basically like a television set only running in reverse. So rather than in television where we have an electronic signal that is then delivered to a cathode ray tube, or at least it was in the old days, which then caused light to be emitted from the front face, which you saw as television, a photomultiplier works in the reverse direction. Light is emitted on a photocathode on the face and it’s converted back into an electronic signal.

What we do is we take that electronic signal; we boost it; we multiply it by over a million times using an arrangement of electrodes inside the photomultiplier tube. That signal then is routed back along with the signal from all the other photodetectors to a bank of electronics which convert the analog signal back into a digital signal. It is then displayed directly in all its detail, all 120 channels—basically some small spikes showing when electrons or electronic signals are coming from the photo tubes. And we know that electronic signal from the photo tubes corresponded to light hitting the front face.

We can measure the position of the interaction and we can measure the amount of energy, and we will compare those predictions with the theoretical predictions we have for the rate of which we are expecting to see dark matter particles and critically the energy distribution that we expect from such events.

Interviewer: How do you know you’re not measuring a neutrino or something else?

RICK: The signature that we expect to be associated with a WIMP is energy of the order of a few tens of kiloelectron volts. The electron volt is a measure of energy where it implies that we’re at a level comparable to sort of energy you see associated with the binding energy of interelectrons in atoms.

We’re often asked how we’re going to know that the signal that we’re getting within this LUX detector is due to dark matter and not some other source, for instance, neutrinos. They’re neutral particles. They are weakly interacting so that they wouldn’t leave a signature anywhere else in our detector as they came in. The reason why we can rule out neutrinos is because they are relatively light. I say relatively but they’re a billionth times lighter than one of these WIMPs.

Your typical neutrino detector often will rely on neutrinos scattering from electrons rather than it interacting directly with a nucleus. So what that means is if we were likely to be detecting neutrino signals in our detector, it would be in the form of events that were interacting with the electrons, not with the nucleus.

So the LUX detector is designed to tell the difference between a nuclear recoil and electron recoil. Most background events are in the form of interactions with atomic electrons. The signal we’re looking for very specifically is a WIMP recoiling from a nucleus.

So, by distinguishing between the two types of events it allows us to reject or ignore the vast majority of background events while still maintaining a sort of search sensitivity to the WIMP events.

Interviewer: Why would neutrinos be so unlikely to interact with the nuclei?

RICK: Actually, as it turns out, neutrinos will interact with the nucleus. But because the neutrino is a billion times lighter than a WIMP, a WIMP can sort of collide with a nucleus and give a quite a kick. When a neutrino comes in and tries to interact with a nucleus it will do so, but the amount of recoil that you get is very, very much smaller. This means that although there are going to be neutrino signals in our detector, the nuclear recoil ones will be of a very much lower energy than the WIMP signals.

Interviewer: How many events do you need in your data before you can confirm dark matter has detected?

RICK: One of the things that really tells you whether the result from discovering this case of WIMPs in one experiment is correct will be the subsequent observation of a similar signal in a second experiment. That is an enormously powerful technique and one that’s used in science a great deal—which is the reproducibility, if you like.

Clearly, the experiment that sees the signal first, as it were, gets the satisfaction of having been there first. But in a very real sense, subsequent experiments are every bit as important. Because, using a different set of processes, if they can demonstrate an observation of a signal which is theoretically consistent with the first one, that will be an enormously important confirmation that we do have WIMPs.

Interviewer: How long does it take you to get your data?

RICK: We are intending to run for at least a period of one year. During that time period, I think in order to convince ourselves that we actually have a signal that is due to WIMPs, we would like to see at least three to five events.

To be absolutely positive you would like to do statistical tests on the data, so the more events you can get the better. I think if in the space of the initial first year of running we were to see a handful of events, then firstly we would almost certainly continue the data run.

The other thing we might well do is say, well look, if using a detector which is a third of a ton we’ve seen a handful of events in a year, then if we were to build a detector that say was 10 times bigger then we would have not just three or four events but actually 30 events. That would be another quite logical thing to do in the event of seeing events.

Interviewer: What if you detect no events?

RICK: I’ve been working in the dark matter field for 20 years and I’ve worked on five different detector technologies. We’ve done a number of search experiments for these weakly interacting massive particles, and every experiment to date we have come to the conclusion that we have seen nothing.

Now, one might begin to question, why is it one does experimental physics, if one is repeatedly doing experiments that see nothing? Sometimes it’s difficult to understand but often seeing nothing from a theoretical point of view can be as significant as seeing something. Because, in a sense, when you see nothing you are learning something new. You are eliminating theories that previously we thought were just as good as the next theory.

If we run an experiment with an entirely new level of sensitivity, we are probing new models. And when we see nothing we’ve eliminated those models. So, we are ensuring the science and our knowledge is marking forward. Now, what we would like to do at some point is transition from ruling out models to start actually seeing particle events, and therefore allowing us to say “Ah, this is now a signal and it’s consistent with a particular set of models.” Then we can start to refine those models in order to see whether we can understand the much deeper nature of both cosmology and particle physics.

Interviewer: What are your predictions for the LUX experiment?

RICK: I think it will be fair to say that everybody on the program knows that we are building what will be the most sensitive dark matter detector in the world and, therefore, there is every prospect of us, if there is a dark matter signal out there, seeing it. If you like, we have the necessary condition of being the most sensitive detector in the world. That will allow us, if there is a signal, to be the first to see it, which I think would excite people enormously.

Even if after the duration of a year of running that detector we don’t see anything, we will have equally ruled out a whole set of dark matter models and, I think, uncowed we will simply move on again to build an ever larger detector in order to look again.

Series Directory

Physics for the 21st Century


Harvard Smithsonian Center for Astrophysics in association with the Harvard University Center for the Environment. 2007.
  • Closed Captioning
  • ISBN: 1-57680-883-1