We speak to the eminent physicist about cosmology research in the twenty-first century.
Joseph Silk is Professor of Physics at the Institut d’astrophysique de Paris (IAP) and Homewood Professor of Physics and Astronomy at Johns Hopkins University. Professor Silk discovered that a process called photon diffusion damping reduced the anisotropies of the early universe, a phenomenon named after him. A fellow of the Royal Society and recipient of multiple awards, including the Balzan Prize and the Gruber Prize in Cosmology, Professor Silk has numerous publications on cosmology, dark matter, and galaxy formation. Professor Silk (JS) spoke to Debdutta Paul (DP) during his visit to ICTS-TIFR for the Largest Cosmological Surveys and Big Data Science program.
The full text of the interview is reproduced below. The answers are lightly edited, and long paragraphs have been split for readability. The questions and initials are in bold, and DP’s additions are in square brackets. A shorter version of this interview first appeared in the ICTS Newsletter, volume IX, issue 1, 2023.
DP: Professor Silk, welcome to ICTS. How do you explain to a high school student interested in physics and mathematics about your field of research?
JS: The field of research I work in is cosmology, generally speaking, and that’s a study of the universe. More specifically, it starts off with looking at the universe, seeing that it’s full of galaxies and asking the question: Where do galaxies come from? How do they originate? What was there before the galaxies? And how far back can we go in time?
DP: Can you explain in detail any one of the questions that you’re studying?
JS: One of the main questions is the origin of the galaxies. Let me put that in perspective. The universe began in a gigantic primordial ball of fire, intense radiation, which cooled down since that explosion occurred. We call that explosion the Big Bang. Now, because it’s so hot, it’s very difficult for anything to condense out of that expansion. But as the radiation cools down, eventually condensation occurs. That’s partly because the radiation cools down rapidly enough so that it becomes as cold today as three degrees kelvin (3 K) — that’s what we measure in the fossil radiation that we view as the cosmic microwave background. But before then, the ordinary matter in the universe is more important. It has a stronger influence on gravity than the radiation does. The intensity of radiation has expanded away, and the matter then controls the gravity. And that means that there are areas which are slightly denser than the average, and they have a slight advantage: their gravity is stronger, they tend to collapse, and they eventually form the seeds of galaxies much later as time goes on. I figured out how to trace these seeds back in time and view them as slight fluctuations, hotspots if you like, in the cosmic microwave background radiation.
DP: We’d love to know about your journey in this field of research, starting from how you came into it to where you are working today.
JS: I began as an undergraduate in mathematics. I spent three years getting my degree at Cambridge University. Towards the end of those three years, I started wondering what I should apply my mathematics to. I experimented with a few things professionally, such as becoming an actuary, for example, I did an internship for that. I found it very boring. Then, by chance, I stumbled into a lecture which I wasn’t officially supposed to be attending. I sat at the back and heard a brilliant lecturer talk about Einstein and gravity and something called Mach’s principle, which is a mixture of philosophy and physics. This expedition captivated me, and I went away, decided to learn more, and eventually chose cosmology as my major interest in research.
DP: What are some of the open questions in cosmology?
JS: One of the questions that we’ve heard for many years now and have really got no nearer answering is… what is the nature of the dark matter? We know from observations that 90% of galaxies consist of something that we can’t see directly. It is mostly in the outer parts of galaxies and so must interact fairly weakly with ordinary matter. So we conjecture that it’s some form of weakly interacting particle. We search for these particles in particle colliders, for example. We do experiments that look for these particles, which can penetrate ordinary matter more easily because they’re weakly interacting, and search in laboratories deep underground. We haven’t found any evidence for these particles yet. That’s why this continues to be a very, very big puzzle — an outstanding question. Many searches are going on. We’re building bigger and better experiments. So far, there’s no indication [of] when we’ll be successful, if ever.
DP: In cosmology seminars, talks, and presentations, we keep hearing about the Hubble tension. What is your take on that?
JS: The Hubble constant is the rate of expansion of the universe, and we measure it by looking around us and measuring the rate of expansion of the distribution of galaxies, space if you like, from the redshift of the spectrum that gives us the expansion rate. Now, it’s complicated because you need some precise distance calibrator to get a precise velocity. Also, the average velocity of the galaxy itself is not good enough because galaxies have a random motion as well. What we’re trying to do is decipher the underlying overall flow, the expansion flow of the universe, the expansion of space. We do this by seeking out what we call distance calibrators. They’re like standard candles, yardsticks if you like, things that you can measure distance with. These, of course, are varieties of very bright stars; they may be variable stars called ‘Cepheid stars’. If you look a bit further away, the targets of choice are supernovae — exploding stars, which are also very good distance calibrators. We combine these together, pushing further out into the universe, and measure a certain rate of expansion where we think we’re more or less in some quiescent part of the expansion of the universe not perturbed too much by local objects, such as galaxy clusters.
At the same time, we use the cosmic microwave background as our anchor in the distant universe. From studying the fluctuations in that background radiation, we can infer a distance scale because those fluctuations are predicted to have a certain scale and will be too large or too small depending on how far away they were. By correcting for this uncertainty in [the] distance, we can infer the scale of the universe — what it should be from the microwave background. The problem is that when we look at the microwave background, we work on our parameters, and we figure out what the missing link is in terms of the expansion of the universe — that’s one parameter we get basically from the sky. We compare that with the local value from the variable stars and the supernovae, and it’s not the same. That’s the Hubble tension. The difference is only a few per cent, but it’s very persistent. The precision of our measurements is such that we think it’s a very real discrepancy.
DP: During the program, you talked about precise measurements carried out by proposed experiments on the far side of the Moon. How and why will these measurements yield new information about cosmology?
JS: The basic goal in the cosmology I want to do next is to penetrate the Dark Ages, that’s long before there were any galaxies or stars very far away in the early universe. That’s the equivalent of a redshift of 50. That is, the wavelengths are stretched out by a factor of 50 due to the expansion. Now, the only things to detect in the dark ages are hydrogen clouds because they’re the building blocks of the later galaxies. One way to look for these hydrogen clouds is to measure them by using the 21-centimetre line of atomic hydrogen, which is produced when the electron spin flips in a hydrogen atom. When the spin of the electron is antiparallel relative to the proton, the energy is very slightly lower. So the spin-flip is excited by absorbing background radio waves at a very precise wavelength, and we see this as absorption against the diffuse radio background. So you can see these cooler clouds as shadows against the background radiation. However, because they’re so far away in the early universe, the absorbing radiation is highly redshifted. From 21 centimetres, it stretches out all the way to 10 metres. That corresponds to a very low frequency, a frequency so low, or a wavelength so long that it’s almost impossible to do these experiments on Earth. Because the Earth’s ionosphere deflects these waves, it stops us from seeing them from so far away and gives us all sorts of extra noise. It turns out that the far side of the Moon is the best place to do this experiment in the nearby universe. In fact, it’s said to be the most radio-quiet region in the entire inner solar system. That’s not only because there is no ionosphere around the Moon to give you radio noise but also because the far side is shielded completely from the Earth, which means there is no radio interference from the Earth from our cell phones or TV or whatever. That’s why it’s such a perfect place to do low-frequency radio astronomy.
DP: What will you learn from these measurements about the early universe, the hypothesis of inflation, and about dark matter and dark energy?
JS: The goal is to learn something about inflation. That’s the major goal of these experiments on the far side of the Moon. The way we’ll do that is the following. Inflation occurred in the first instance of the universe, 10 to the power of minus 36 seconds (10-36 seconds) after the Big Bang. It produced a dramatically huge expansion of the universe and then rapidly settled down to a more normal expansion rate. But during that brief instant of very accelerated early expansion and the settling down to the usual expansion rate, a field of gravitational waves was generated. Those waves, of course, redshifted to a very low frequency, but they leave a tiny imprint on the microwave background. This is because gravitational waves shear matter as they pass through it by a very tiny amount. They leave this shear signal on the microwave background, which is unique to the passage of gravitational waves. We can measure this as tiny twists in the fluctuations in the background radiation. It’s a non-compressive mode of polarisation, a shearing mode of polarisation, that’s the signal of gravitational waves.
With the cosmic microwave background, we’ve been trying very hard to look for this signal. It turns out that it’s really difficult to find. It’s very, very weak. But above all, the inflation models don’t give you any definite prediction. They tell you what is possible, but they don’t tell you what is guaranteed. So we’ve had to think of an alternative way to get a definitive result on testing inflation. That definitive result comes about because inflation generates the fluctuations from which all the structures are made today — it’s one of the great successes of the theory. But in so generating those fluctuations, it leaves slight twists and turns. We call this non-Gaussianity, non-randomness in the pattern of the fluctuations in the sky. We will use the enormous amount of information we have in the low-frequency radio waves from all these early clouds to try to detect this deviation from Gaussianity in these primordial fluctuations. We can see this in very low-frequency radio, effectively. That’s how we’ll test inflation because inflation is guaranteed to produce these deviations from Gaussianity.
DP: We recently heard about the proposal to have gravitational wave detectors on the Moon, the LGWA. How important do you think is LIGO-India in the LIGO collaboration? And will it be useful if it comes up in, say, the next ten years?
JS: LIGO-India is very timely. Right now, we have three functioning experiments: two in the US, one in Italy, and a fourth one in Japan is about to come on the air. Adding a fifth detector in India will be a great improvement. That’s because, with these experiments in different parts of the globe, you can greatly improve the localisation on the sky of the gravitational wave sources. That India will bring into the game. Its great distance from the other observatories means we’ll be able to pinpoint much more accurately where the sources are coming from. That’s really important. Even if it will take us ten years to get there, that’s still fine; there’ll be no competition. And that will be a wonderful contribution to the field.
The reason that the Moon will provide an important addition to gravitational wave telescopes that will probably occur on roughly the same timescale, maybe a few years later, is the following. These gravitational waves from far-away merging black holes pass by us. They shake the interferometer, the telescope, one that eventually will be LIGO-India; currently, it is LIGO and Virgo. We measure the resulting signals at a certain frequency corresponding to the speed at which the waves (at the speed of light if you like because gravity travels at the speed of light) can traverse a few kilometres — that’s the length of the arm of the interferometer, the beam as it were. That’s your measuring rod, and the vibrations in it give you the signal that you measure in these gravitational wave detectors. Now the Moon is some 8000 kilometres across. So when a gravitational wave passes there, it shakes the Moon very, very slightly. That gives you a vibration on a scale that’s not four kilometres or 40 kilometres (which will be the new generation of ground-based telescopes after LIGO-India, but that’s projected), but something much, much longer, and therefore a frequency that’s much, much lower. And that is really, really nice because if we had gravitational wave telescopes on the Moon, we could then measure a frequency range that corresponds to black holes coming together. They move slowly at first, gradually speeding up, so we can measure their approach as they are produced as the black holes begin to merge together. It’s a very important missing link in our understanding of how black holes merge. We would see their approach. We can do that very simply on the Moon because all we need to do is to put seismometers — very, very precise seismometers — on the Moon. Now, Apollo did that 50 years ago, and they measured the first lunar quakes. These new seismometers, much more precise, will measure the tiny, tiny vibrations of the Moon from passing gravitational waves. A few of those installed on the Moon will be a wonderful new telescope, a futuristic one, but one that we’ll be able to build in perhaps 20 years’ time and complement all the other gravitational wave telescopes we have.
DP: You have studied the possibility that dark matter is made up of a large number of tiny asteroid-mass black holes. What attracts you to this scenario over other possible explanations of dark matter?
JS: The problem with dark matter is that we haven’t found it yet. We’ve been looking desperately for weakly interacting particles, and if they existed, they should be produced in collisions, high energy collisions of known particles, [where] you’ll see events with missing energy or missing momentum. But we haven’t seen those yet. So we’re being forced to think of different possibilities. One of those is a black hole. Because we know they exist, we’ve measured black holes, and we’ve even imaged very massive black holes. Black holes are dark, so they’re ideal for dark matter. The problem is, if the black holes are, say, produced by dying stars a few times the mass of the Sun, then we can set very strong limits on how many of those there are from basically their merger rate and the gravitational waves they produce. There simply are not enough. We also have other types of experiments, looking for dark things passing in front of nearby stars. All of these say that most black holes are one per cent of the dark matter. Also, dark matter cannot consist of very massive black holes, [because] they wouldn’t be dark: they’d be glowing as they’d inevitably be accreting ambient gas. So you have to say, well, maybe it’s not black holes produced astrophysically like the mass of the Sun, but they could be much smaller.
In principle, primordial black holes could be really, really small because the universe was very dense very early on. If regions collapsed today, they would make enormous black holes, but early on, they would make microscopic black holes. So the question then is… what masses of black holes could you imagine that could be the dark matter? They can’t be too small. Because if you made the black holes less than roughly the mass of a small kilometre size asteroid, actually, about a billion tons or 10 to the power of 15 grams to give you a number, then they would undergo a process discovered by Stephen Hawking, it’s called evaporation. So, they would disappear and couldn’t be the dark matter. If they were about a hundred million times more massive than that, they would deflect the light too much when they are in between us and nearby stars. We would see light deflections, and the process that we call lensing, or microlensing, would allow them to be detected. So there’s a narrow range, not so narrow really, it’s between asteroid mass and lunar mass basically, a respectable range of mass where we could hide the black holes. They wouldn’t deflect the light from stars, they wouldn’t Hawking-evaporate, and they’d be stable. If they were produced in the early universe, they would be the dark matter today.
The reason why we think this is an interesting option is that if you go back early enough in the universe when these were made, they could be very rare events early on. You could make very tiny fractions of these at that time, as a fraction of the energy density of the universe. What happens is the radiation all expands away, but the black holes are left behind. So, the tiniest numbers at the beginning, amounting to fractions of a billion compared to the density of energy then, could be the dark matter today. That’s the attractive part of the hypothesis. Rare events can make them — events that just involve gravity, so it’s not a great mystery. We’re not inventing new particles. It is a little bit unusual in the sense that our theory didn’t predict these, but we can tweak the theory to make them, and they could be the dark matter.
DP: But how do we test these primordial black holes?
JS: Our best hope is that they would actually collect where the dark matter is. In the centre of a galaxy where there often lurks a supermassive black hole, we know that there’ll be lots of these tiny black holes around it as the black hole itself grew from smaller beginnings. Early on, far away in the past, those tiny black holes, the dark matter, would cluster around the massive central black hole. Many of them would fall in and give out some gravitational waves. And that will result in something detectable. Although you probably couldn’t see individual gravitational wave events from the [process of them] falling into the black hole, you would imagine sort of a stochastic background, as ripples in the gravitational waves in the background sea of gravitational waves. That would be a vital clue.
DP: Thank you so much, Professor Silk, for speaking to us.
The author thanks Spenta Wadia and Ananya Dasgupta for inputs.