Igor Pikovski recently decided to tackle one of the holy grails of physics: discovering the quantum soul of gravity. “It’s the biggest mystery of physics,” says the Stevens Institute of Technology quantum information scientist. Quantum physics was formulated more than a century ago by the collective efforts of many scientists, including Albert Einstein, Max Planck, and Niels Bohr, to explain some puzzling experimental results, and it has been an active field of inquiry ever since.
We’ve come to understand that quantum physics can explain all kinds of strange effects in the universe. In fact, it can describe almost everything—except gravity. Many scientists have attempted to ferret out how gravity and quantum physics intersect, using string theory, for example, or theories of everything, says Pikovski. “It’s a very mathematical, speculative field where we don’t have answers yet,” says Pikovski.
READ MORE: What Physicists Have Been Missing
One way to connect them would be to observe a graviton, a hypothetical and unobserved elementary particle that acts as the force carrier for gravity in quantum field theories. Until recently, scientists thought that detecting a single graviton would be impossible. Pikovski says this thinking was rooted in faulty assumptions about the way the universe works. He is currently building a prototype graviton detector.
The hunt for the graviton has been overshadowed by the hunt for dark matter, says Pikovski, but he thinks a machine that could detect a graviton might be ready within a decade or two. “I would even wager that we’ll find a graviton before we find dark matter,” he says.
I recently spoke with Pikovski about his graviton detector and why AI is bad at creative physics.
What exactly is a graviton?
What we understand is that if quantum physics is really the underlying description of how the world works, it means that everything is made out of particles. Matter is made out of atoms. Atoms are made out of things like protons and electrons.
We can explain everything that interacts in a framework where there is a particle associated with it. This is the modern view of physics, which is known as the standard model. But then there is this one speculative particle, which is outside of the standard model, and that’s the particle associated with gravity, the graviton.
Now, we can write it down mathematically or postulate it, but the reason it was always assumed that it’s completely impossible to detect is because we had this intuition that gravity is something that really affects only the very, very big scales. Something like Earth or the sun or planets and beyond. That’s where gravity matters.
And then, at the other extreme, how do we see quantum effects? We have to go really small, to the level of atoms and electrons. We figured that you could maybe see a collection of gravitons, say 10 to 30 of them, but you could never see just one. Because at such small scales, gravity would play no role. That was the idea. But it turns out that’s not the full story.
People knew that atoms existed and maybe even other particles long before we could isolate them individually. It’s very similar with graviton detection. We figured out that the way people calculated it was not optimized. There’s a way one can extract a single graviton from a very, very large amount of them. That’s the essential idea.
Is that the idea behind your prototype detector?
Exactly. We published this work two years ago in theory. We said, “Look, it was an oversight. People thought you could never detect a graviton, but there is actually a realistic way.” And we outline how it’s done.
To give you a picture, gravitons are most likely around us all the time. It’s just we don’t feel them. It’s as if you’re on the shore and you’re hit by a giant tsunami, except this tsunami wave is what we call a gravitational wave made up of gravitons.
People detected gravitational waves 10 years ago, but the problem is that this tsunami is so weakly interacting that it essentially travels through Earth. It just shakes the detector a tiny bit. These particles are invisible, very much like neutrinos, for which the Nobel Prize was just awarded. Gravitons are even more transparent. You have many of them, but the goal is, can you find a way to fish out one graviton out of this tsunami?
That’s essentially what we found. And the prototype detector we’re building is based on that premise. It can be scaled up and start detecting these particles one by one.
Do gravitons always travel in waves?
This is a very, very deep question. It goes back, essentially, to the very early days of quantum physics, 100 years ago. One can think of gravitational waves in nearly exactly the same way as we think about light waves. One of the weird things of quantum physics is that a light wave sometimes behaves like a wave, and sometimes it behaves like a particle. And that’s exactly what it is. That was the puzzling thing in the early days of quantum physics where we see something that comes as a wave, but it turns out that actually this wave is made out of these tiny particles, which we call light photons.
Nowadays, we can detect photons. In 1905, Einstein figured out, “Well, we can explain everything if you realize that the light wave actually is made out of these tiny photons, tiny particles, but they’re a bit unusual.”
The short answer is yes, gravitons like photons always have different facets. If there are many photons, it looks like a wave overall, but actually the wave is made out of these tiny individual particles, and it’s the same with gravitons. They seem to have wave-particle duality. That’s the beauty of it, that we just don’t know. It’s all speculation. But once we can detect it, we can test it.
How does the mystery of the graviton relate to bigger questions about whether the standard model works or not right now?
It’s probably the biggest question about the standard model. It’s not even often asked. But the standard model has many issues, including the fact that there’s something that seems to be populating the universe, which is called dark matter.
People think that the standard model is not complete because it should include something like a dark matter particle. There are so many experiments but they’ve never found it yet. People look for it because certain observations of the galaxies suggest there is invisible matter somewhere. People don’t know what it is. But the main issue of the standard model is precisely that we don’t know how to put gravity in it.
One way to put gravity into the model is to have a particle called a graviton. But we don’t even know if these gravitons exist. So the standard model does not include it. We thought it would always just be a speculative question. But now it turns out it’s an experimental question.
How would your prototype work?
More than one hundred years ago, Einstein predicted gravity to be this strange phenomenon, which somehow bends space and time. This theory allows for many strange phenomena. It predicts black holes. Time can change. And all of that has been tested.
By now, it’s pretty well established that black holes exist. And there’s another prediction it makes, which are called gravitational waves. They’re very similar to light waves, except that in this case, they consist of gravity propagating and squashing space around it. And these waves, they also propagate at the speed of light. And they are produced all the time, but they’re very faint.
Gravitational waves were finally detected about 10 years ago when two black holes merged. There are these violent, cosmic events occurring all the time in distant galaxies. Black holes swallow each other or other stars. And in this process, they emit a lot of energy through these gravitational waves, which are typically invisible. But they propagate through the universe, through everything. They were detected about 11 years ago with LIGO, which was conceived in the 1970s, built in the 1990s and 2000s, and made its first detection in 2015.
Let me step back. When light passes by, what does light do? It shakes charges. For gravitational waves, something similar happens, but not exactly the same. It shakes space itself. As the gravitational wave passes through, everything is just kind of shaking a little bit. It’s so tiny that, of course, we don’t feel it, but these detectors like LIGO, they’re sensitive enough now to really measure it. They’ve been doing it for over 10 years now, and they measured a lot of passing gravitational waves. Every few days or so, it’s like, “Oh, there’s a, there was a black hole merger, and now we see the gravitational wave passing.” And there’s one second where this wave passes, and they measure it very precisely. That’s just standard wave detection.
We realized that if you just wanted to measure how this wave is made out of gravitons, it’s not much harder than measuring the wave itself. As the wave is passing, we essentially pick out a single graviton. What we want to do is not just simply absorb some energy from this passing wave, but we absorb energy which corresponds to exactly one particle.
Energy is deposited in discrete chunks, which are now called quanta, and these chunks are proportional to frequency, and this is what we now call a photon, and the exact same thing should be a graviton. So long story short, we simply want to have a detector that absorbs one quantum of energy from this passing gravitational wave.
The numbers work out exactly such that you have something like, 10 to the 30 gravitons, and your probability of absorbing is exactly one over 10 to the 30. It’s really a lottery. But with so many gravitons passing through, there is near certainty that you will absorb one. You just stand there with your detector, you wait for this gravitational wave to pass, and you measure how its energy is changing by exactly this discrete amount.
You said that LIGO took about 40 years to build, essentially. Do you think that it’ll take a similar amount of time or more to build a machine that can actually detect a graviton?
I use LIGO as a benchmark because somehow it took 40 years from conception to detection, but it was a success story. It worked. You start small. But I think for gravitons, it would go much faster. And the reason for that is because we have already LIGO. And our idea is very similar. Instead of looking at the shaking, we need to look at how energy is jumping in the system. We’re already halfway there. If I had to guess, I think it would take less than half the time, so maybe 20 or even 10 years.
I understand you sometimes use AI in your work but have found that on matters of physics it lacks creativity. A few years ago, you asked AI whether gravitons could be detected and it said, “Absolutely not. This is impossible.” You have called it a conventional wisdom machine.
To me, AI is very useful for certain things. But one has to always be mindful, at least at this stage, that it’s not a universal tool to generate entirely new ideas, at least at the moment. Just today, I was reading it solved some problem. But if you know what it’s about, usually these problems are complicated calculations. That’s usually where it currently is amazing, incredible. Calculations that humans have been thinking about it for 50 years, and AI can just do it in five hours.
The reason I asked AI about the graviton detector is that I already had this result. We had already even published it. But every time a new AI model comes out I like to test it to see how far it can think outside the box, because it didn’t yet have our paper in the training data. I would even guide it a little bit because I knew the answer. It would always push back, and that’s okay. Because it’s essentially a textbook.
This was also nice in a sense that, at least at the moment, it’s still humans who innovate something that didn’t exist before. To me as a physicist, at this stage, I’m happy because the most exciting thing about physics is really pondering the universe, and at the moment AI cannot replace that. These tedious calculations are not what’s fun about physics anyway.
Lead image: Roman Samborskyi / Shutterstock
