- LIGO is a project aimed at measuring the bending of space and turning it into a measurable signal
- It uses laser interferometry to detect gravitational waves
- The waves are detected in the human audio band and can be converted into sound signals
- New mirrors are being developed to detect different wavelengths
- The measurements are challenging due to the distance and attenuation of the waves, but LIGO is able to look deeper into the universe than expected.

- The signal from the universe only decays by 1 over the distance, allowing us to observe deeper into the universe than expected.
- The Earth would only stretch by a small amount if hit by a gravitational wave.
- The Fabry-Perot interferometer is used in detectors, where the laser bounces back and forth, increasing sensitivity and signal power.
- Increasing laser power helps build up the signal, but is limited by quantum fluctuations and noise.
- The stability and precision of the laser is crucial for accurate measurements.
- Advanced LIGO benefited from the contribution of new, smart individuals who joined at the right time.

- Engineers faced numerous challenges when constructing and assembling the new detector for gravitational waves.
- One major problem was the interaction between the laser beam and the mirrors, which caused movement and interference.
- The team used a feedback control system and sophisticated techniques to try to minimize these issues.
- However, they are still looking for better optimization methods to avoid masking the gravitational wave signal.
- They are exploring linear subtraction and nonlinear regression techniques to improve data quality.
- The goal is to clean up the data and remove noise sources to enhance the detection of gravitational waves.

- The speaker discusses the history of black hole detections and improvements made to the detector.
- They believe there may be more signals that they have not yet identified.
- The speaker explains how black hole ringing frequency can be computed based on the size or mass of the black hole.
- They discuss the challenges of detecting signals from larger black holes due to technical noise and vibrations.
- Data science is being used to remove foreground noise and find deeper signals.
- The speaker explains that signals from black holes can be searched for by comparing them with known templates.
- They mention the possibility of hardware innovations to improve detection.
- The speaker discusses the resonance of mirrors and how they can be used to measure motion using lasers.
- They explain that the mirror resonates at specific tones and that thermodynamic vibrations can be heard in the LIGO data stream.
- The removal of specific frequencies is necessary to filter out non-gravitational wave signals.
- The speaker acknowledges the possibility of picking up signals that may not be gravitational waves.

- The speaker discusses the challenges of detecting gravitational waves and the need to differentiate them from other sources of interference.
- They mention various technical issues that have been encountered, such as misbehaving refrigerators and synchronous crystal frequencies.
- The speaker describes the process of finding and solving these problems, often through late-night discussions and experimentation.
- One specific problem mentioned is the interference caused by light bouncing off mirrors and picking up vibrations from unknown sources.
- The speaker explains plans to address this issue by using extremely black substances to block the light beams in the system.
- Different materials, such as stained glass and nanotubes, are considered for their absorptive properties.
- The speaker also mentions the possibility of building longer interferometers to improve detection capabilities.
- Overall, the focus is on the ongoing challenges and solutions in the field of detecting gravitational waves.

- Building a larger system to measure the universe would allow us to look back 10 billion years into the past and find signals from the first stars and the evolution of space-time.
- It would also help us understand if the universe started with different spatial dimensions and if gravity travels through three-dimensional space or through another dimension.
- The current limitations on Earth, such as vibrations and noise, make it necessary to build an interferometer in space.
- The LISA project aims to launch a triangular interferometer in space in about 16 years, which would greatly improve our ability to measure gravitational waves.
- Having a better gravitational wave detector would allow us to experience the richness and nuances of space phenomena, and help us understand where the laws of physics break down and new discoveries emerge.

- The video discusses the quest for knowledge in understanding the laws of physics, extra dimensions, and new particles.
- The speaker highlights the importance of studying gravity as a probe for understanding the universe.
- The video explores questions about the structure of the universe, the nature of empty space, and the reasons for the existence of galaxies.
- The speaker emphasizes that curiosity-driven research in science and technology has proven to be profitable and beneficial for society.
- The potential impact of a black hole merger closer to Earth is discussed, including the destruction of electronics but not major catastrophic events.

- LIGO is a project that uses laser interferometry to detect gravitational waves and convert them into sound signals.
- New mirrors are being developed to detect different wavelengths and allow for deeper observations into the universe.
- The stability and precision of the laser are crucial for accurate measurements.
- Advanced LIGO has faced challenges in constructing and assembling the new detector, particularly with the interaction between the laser beam and mirrors.
- Data science is being used to remove noise and enhance the detection of gravitational waves.
- The speaker discusses the challenges of detecting signals from black holes and the possibility of hardware innovations.
- The resonance of mirrors and the removal of specific frequencies are important for filtering out non-gravitational wave signals.
- The speaker acknowledges the possibility of picking up signals that may not be gravitational waves and the need to differentiate them from other sources of interference.
- Plans are being made to address interference issues by using extremely black substances and exploring longer interferometers.
- Building a larger system in space, such as the LISA project, would greatly improve our ability to measure gravitational waves and understand the universe.
- The video emphasizes the importance of studying gravity as a probe for understanding the laws of physics and the potential impact of a black hole merger closer to Earth.

- The Earth's acoustic modes could be excited if a binary black hole pair passed through the solar system at a specific frequency, potentially causing significant damage.
- The current method for detecting gravitational waves is not the best idea and there are other possibilities being explored, such as acoustic detectors and space detectors.
- There is a need to improve the transduction of space-time curvature into an electrical signal and light may not be the best option.
- Coherent quantum feedback is an idea being considered to optimize the detection of gravitational waves, turning the system into an unstable feedback system.
- There is progress being made in solving the problem and a scale model of the new system is being built to test its sensitivity.

The Earth's acoustic modes could be excited by a passing binary black hole pair, posing a risk of damage. Alternative methods for detecting gravitational waves are being explored, including acoustic and space detectors. Improving the transduction of space-time curvature into an electrical signal is necessary, and coherent quantum feedback is being considered. Progress is being made, with a scale model being built to test sensitivity.

Rana Adhikari is a Professor of Physics at Caltech and a member of the LIGO team, who were the first to measure gravitational waves - https://en.wikipedia.org/wiki/First_observation_of_gravitational_waves

Their detection papers are available here - https://www.ligo.caltech.edu/page/detection-companion-papers

Rana on Twitter - https://twitter.com/RanaXAdhikari

YC's podcast is hosted by Craig Cannon - http://twitter.com/craigcannon

So maybe, yeah, maybe we should just start out explaining, like, what is LIGO?

LIGO is a huge project aimed at being able to take the bending of space that we think is happening all the time and turn it into some kind of signal that we can use and measure.

It's like having a voltage meter or a microphone, but it's like a microphone for space.

Can you do the beginner's explanation of what the device actually is?

Yeah, so we think, or we now know, based upon Einstein's relativity theory from 100 years ago, that there isn't a real force of gravity, like there is a force between magnets or between charges or things like that, but instead the way that gravity works is that it curves space.

And so it's a lot like imagining what happens when you're jumping up and down on your bed and somebody else is jumping on their bed, I guess, because if this happens to you all the time, then this is something you're familiar with.

But whoever's heaviest makes a big impression in the bed and then whoever is littler has to account, what kind of an analogy is this?

Whoever's littler has to account for the depression in the middle of the bed, and just are jumping accordingly, and you tend to slide into the biggest dimple in the bed.

So those who do trampolines or jumping up and down on their bed, understand how, you know, well-trained, and understand how gravity works.

So to detect this on the Earth is incredibly hard, and people have long ago measured the curvature of a space due to the Earth, and due to the Moon, and other planets, and that sort of thing.

And so we well understood that in fact space is curved, but we had no evidence to support the idea that the curvature of space could travel through space as waves.

The detection of gravitational waves, for decades, had been debated in the scientific community.

That it was waves of thought, people called it, because they thought it was just waves of mathematics.

It was just some equation he wrote down, but it didn't make any sense.

And some decades ago, people realized it was real.

And then some crazy people said, Hey, let's try to measure this, even though it was millions of times not, I mean, like factors of millions, not possible.

But luckily, through this is just something I'm not capable of, but through a combination of optimism and courage and not knowing the right answers to several equations, they were able to start up the field and start to look for these things.

I think if they had known how tough it would be or that it was gonna take 50s, 5 years to have success, probably no 1 would have started.

And so now here in the modern times, the way we do it is we use the tool of laser interferometry, which is, for those of you who are interferometer aficionados, it is a Michelson type interferometer with a lot of extra stuff added onto it.

For those who are not, the concept is simple.

So you take a laser, like a laser pointer, but much more expensive and therefore much more stable.

Is it a billion dollars now into the project?

You can do, probably you could do the whole thing with a hundred thousand dollar laser.

You split it in 2 and you send it in 2 separate directions.

And then when the waves come back, they interfere with each other.

And you look at differences in that interference to tell you the difference in how long it took for 1 beam to go 1 way and the other beam to go the other way.

And so this, the, the way I said it was really careful there because, there's a lot of confusion about the idea of there, these are waves and space is bending and everything's shrinking and how come the light's not shrinking and so on.

There's, there's no real difference, between the ideas of space and time warping.

It could be space warping or time warping, but the only thing that we really know is what we measure.

And that's the mantra of the true empirical person, I guess.

We send out the light and the light comes back and interferes and the pattern changes and that tells us something about effectively the delay that the light saw.

And it could be that the space time curved so that the light took longer to get there.

But you could also imagine that there was a change in the time in 1 path as opposed to the other instead of the space.

But it's a mixture of space and time, so it sort of depends on your viewpoint.

But this warping of space-time is what's measured.

And we turn it into a real signal by putting the interference of the 2 beams onto a usual photo detector that's like a solar cell that you would use, and so it turns light into electricity, and that's the whole thing.

And then you measure it by looking at waves.

Well, the whole measurement is right there in the electrical signal out of the photo cell.

Okay, and so then how do you go about converting it to, I mean I've seen a bunch of these like sound, sounds of things you've measured.

Yeah, it's the same as like an electrical instrument.

If you have an electric guitar, when you play whatever you're playing on there, it generates an electrical current in the pickup coils of the guitar, and that signal comes out through a little cable and goes into an amplifier, and then that directly makes sound.

So the photo detector detects a signal which turns into electricity, and then we take that electricity and we drive a speaker, and then it makes sound.

So there's no, I can understand what you're asking.

How is it that the wave from outer space can directly get turned into a signal in a speaker?

It happens to be, And then I think this is a whole other topic to discuss.

It happens to be that the waves that we're detecting and the waves which are easiest to detect are exactly in the human audio band.

So the waves that you and I can hear with our ears.

That's the whole frequency range for gravitational waves that we can detect.

Gravitational waves happen at all frequencies, but they're really loud right in this band.

And so our detectors are aimed for this band because we expected the audio band would be a good place.

But our technology also happens to be only capable of detecting things in this band.

For a lot of technical reasons, which I can tell you about, that happens to be the case.

And so it is a little weird, but the signals directly make sound.

So what I wanted to talk about then, I saw in 1 of your talks, I guess it was from last year, you were talking about creating new kinds of mirrors to focus, rather, the mirrors that you had created focused the vibrations to certain parts of mirrors, but you're working on creating new mirrors to detect other things.

When I first heard about it, it was, I was going to say, when I first heard about it, I thought maybe it was unbelievable, but I have to say, I don't think I ever, when I first heard about this project, I don't think I understood enough to even understand that it was impossible.

So I never went through this disbelief belief period.

But a lot of people with good reason say, it's impossible to make these measurements.

And the reason is, if you imagine zooming into where the black holes are in outer space, they're like a billion light years ago, a few billion light years ago, and the universe is only 13 billion light years, 13 billion years old.

So it's a good fraction of the size of the known universe away.

And if you get close to these black holes, the amount, as they're merging and eating each other, the amount that the space is warping is enough so that, like this water glass, it would be shattered, but it would be stretched, if it was stretchable, to, you know, like this.

But the wave as it propagates to us, like waves do, they get attenuated.

So imagine this stretching has a lot of energy in it, but as it spreads out, the amount of energy has to be conserved as it propagates.

And so the amplitude of the wave as it comes to us gets reduced like 1 divided by the distance.

And so the energy, or the energy in the wave goes like 1 divided by the distance squared, like any other kind of radiation.

We measure the amplitude instead of the power.

We measure directly the stretch rather than the heat or something like that.

So we're able to look a lot deeper into the universe than you would naively expect because our signal only decays like 1 over the distance instead of the distance squared.

So the wave comes to us, and by the time it gets to us because it's billions of light years, the squeezing and stretching is much less than 500% it's more like a part in 10 to the 21 or 22.

So that means if you have...the whole Earth for example is about 10,000 kilometers in size.

And so that, the whole Earth will be only stretching by about 1 hundredth of a micron.

A micron is like the wavelength of light.

Or my hair is the diameter of my, what little hair I have here, is about 100 microns.

So it's 10,000 times smaller than the width of this hair.

Is how much the entire earth would stretch when it was hit with a gravitational wave that you initially measured?

We've not seen anything of that size since then, since that first 1.

Our detectors are big, they're 4 kilometers, so they're not 10,000, unfortunately.

If I was in charge, I would drill through the center of the Earth, and there would be mirrors.

I would put a big L in the center of the earth and then there would be mirrors on both ends.

That would be ideal and that's what we would use.

But then, okay, so to divert a little bit, can you explain how the Fabry-Perot interferometer works, because it's not, it's 4 kilometers, but it's also kind of like bounce back and forth at the same time, right?

It's the Fabry-Perot cavity means rather than just send the laser beam down, we send it into a thing which has 2 mirrors into it.

And just like, I don't know if everybody does this, but when I was a kid I wondered, here's this bathroom mirror and what if, you know, I can see my reflection and then if I go into the fun house like during Halloween, I can see multiple reflections and what if you put 2 mirrors together and you set a flashlight and you took it out, would it bounce around infinity times and explode the mirrors?

And you know, I've always tried to, you know, I wondered about it and it turns out no, sadly.

You cannot destroy the universe by facing 2 mirrors together.

And the reason is because the at each mirror surface, a little less than 100% of the light gets reflected back.

Some of it gets turned into heat or things like that.

And so the Fabry Pro optical resonator is just 2 mirrors facing each other.

So it's set up so that let's say 1% of the light comes out and 99% goes in.

Or, no, 99% gets reflected and 1% goes through.

And so when you set it up at first, you have these 2 mirrors and you put in the beam from this side, let's say.

And so that little bit gets in and starts bouncing around.

But by the time it comes back, you're already putting in more laser light, you're constantly putting in more laser light.

And that builds up constructively with the waves, electromagnetic waves you're sending in.

And slowly the power builds up in the system just through this little leakage.

And it builds up until the point where the amount coming out is about the same as the amount going in.

And at that point you have, let's say, a few hundred times more laser power in this system than if you had just sent in a single beam.

And it gives you basically a factor of 200 extra sensitivity than you would get.

So the extra power of the laser generates more sensitivity?

It's just like you said it though, it's effectively like the laser bounces a bunch of time.

And so you can imagine, here's this space time which is curved.

So now the laser has to travel through this curved path and so it's a little bit longer of a distance.

And when it goes down and comes back it picks up a little bit of extra phase shift.

And if you do 200 round trips, you get 200 times the phase shift.

But does that net out 200 times the noise as well?

So, the more times you go around, the more you pick up the signal from the mirror motion.

And so the signal-to-noise, trying to get below the fluctuations of the mirror, don't actually get better by building up more power in the system.

And in fact, if, it can even go the wrong way.

At some level, at some point, the quantum mechanical fluctuations in the, the number of photons that are in your system, becomes so large that the mirrors are shaking.

And so the more power you have, it just gets worse.

And so there's a limit to how much power you want to put in there before you get into trouble.

But so you might reasonably ask, why do you do it then, if it doesn't help?

It helps just in building up the signal, and if you're limited by the fluctuations of the mirror motion due to the environment or something like that, then more power is not any good.

But most of the time, in most cases when people are doing precise measurements with lasers, they're limited not by the mirror motion, but by the noise due to the fact that they have a finite number of laser photons.

So if you just take a laser and put it onto a photocell and you listen to it with your speakers, which you can do, it sounds kind of like a hiss.

And that hiss is because quantum mechanically, the energy of the light is sort of, is in the discrete packets, which we call photons.

And so if you have a 1 watt laser or something like that, you have some many billions and billions of photons, but you end up with a hiss level of noise due to the, basically the quantum nature of the light that you can't get beyond.

And increasing laser power builds up your signal.

So if you double the laser power, you double your signal, but your noise, due to these random photon fluctuations, only goes up like the square root of that power.

So you win a little bit for that particular noise.

And why not just have a super powerful laser?

Okay, but you need it to be even stronger.

It is tough to build a super powerful laser.

trying to think of how to put this in context.

So the people who have worked with these lasers will understand why we do it this way.

If you have these days, like if you have $50,000 in your pocket, you can go onto the internet and buy yourself a scientific research grade laser that's a couple of watts and will work fine for you.

And there's no such...you can't buy anything like that these days.

You can buy lasers that have that much power, but their frequency is not very stable.

And our laser is used as the meter stick for doing the measurement.

It's like each wave of the laser is 1 tick on a meter stick.

And if you have an unstable laser, that means these little waves, instead of being very precise, are jiggling all over the place.

And so your meter stick, it would be like having a meter stick where the tick marks are kind of dancing around.

Right, and it's so precise, because like, just so I don't mistake it, the first, the 2 black hole measurement, how much did it move?

About 10 to the minus 18 meters, which means 1 billionth of about the size of an atom.

So you have to, there's no laser in the world which is good enough to measure this.

And so we take the best laser in the world that we can find, and then we stabilize it and make it about 10 million times more stable than what you can buy, and then it's kind of just barely good enough.

And we're going to have to do better if we want to do better.

What changed between LIGO and Advanced LIGO?

We got a bunch of new younger people who are really smart.

And because they were, yeah, like miraculously, they showed up just at the right time, you know, in this, They showed up for graduate school in like 2010 or 2011, right when we happened to need them.

You know, without knowing about the timing, they just showed up and they finished some classes and then they decided to ship out and live at these remote sites in Louisiana and Washington.

They just, day in and day out, they just figured out every problem and solved it.

And so, yeah, what are some concrete examples?

I mean, there were some engineering changes in going into this new detector, which was 10 times more powerful laser, and we isolated the mirrors a lot better from the environment, and the mirrors are much higher quality.

They're super beautiful and much heavier and really good.

So there's a bunch of these technical things which were changed.

But, and each 1 of them on their own worked really well because the engineers who were constructing them did a great job.

The problem was when trying to put it all together and These things just never reveal themselves when you're sitting in a little room and designing your widget.

But you put it in a suitcase and carry it on the airplane and try to bolt it onto the 4 kilometer, billion dollar machine, and then it's just like, Total disaster.

Well, because when I first heard you talk, you were talking about deer getting close to the tube and all these things.

Here in our labs at Caltech, you measure 1 thing or 2 things, but the full problem of putting it all together and what's the problem that only shows up when you have, you know, hundreds of kilowatts of laser power and giant mirrors and it's all running together.

For example, I would say, I mean, it's still 1 of the toughest problems and it's not completely solved, is it has to do with the interaction between the laser beam and the mirrors.

And normally when you think about these things, you say, well, the laser beam goes out there and it bounces off that thing and it comes back and that's all there is to it.

And maybe the mirror is shaking around so that's a problem.

But in fact there's so much laser power that when you, it's weird to think about, but there's so much laser power when we hit the mirror, it moves the mirror.

And the mirror, to imagine it, is about this big.

And, in the previous LIGO 20 years ago, I would just pick up a mirror and you could carry it in and put it in, But there's no longer any of this like, yeah, I'll pick this thing up and carry it around.

Yeah, it's way too expensive and way too heavy.

But the super heavy thing, we're hanging from handmade glass fibers, which are super thin, and it's sitting there and swaying around.

And when the laser power hits it, it just moves.

And even more annoyingly, when the mirror moves a little bit like this, the laser beam at the other end, 4 kilometers away, it moves.

And so then that mirror twists a little bit like this.

And now the reflected beam moves a little bit.

And these 2 things are talking to each other through this, the pressure from the radiation, the laser light.

And that's super annoying and it doesn't, it's not a thing that you can test if you're in a tabletop.

You have to put the whole thing together.

You can simulate it and calculate it as we did.

But when you put it together, it's a lot more trouble than expected.

It kind of works, but when we start increasing the laser power, a lot of these interactions happen, which are really troublesome.

And luckily we have a fresh stream of new people coming into grad school.

And if they, we, we, we hope they, they, they remain as, as good as the people we've had so far.

If that doesn't work out, we're in trouble.

Yeah, so what do you suspect will be the changes that suspend the mirrors in a way that the laser doesn't move them?

I don't think we have any way of doing it.

What we're doing right now, we have just a sophisticated feedback control system that we measure the light beams which leak out of the system at a bunch of places.

We detect it and then we, we have a, a system of, something like 20 feedback loops, which, push puts forces on the mirrors to try to keep it aligned and keep this from being such a problem.

The trouble is, well, you know, we're trying to detect gravitational waves, which are tiny, and so when you do something, you imagine like this, let's imagine like this is the mirror, and this little, there's this little ting, which is about how long the gravitational wave lasts, like this.

And now we're trying to control the pointing of this thing because it's getting steered around by the beam, and so I have my feedback control like this trying to hold it on.

it just screws up the whole thing if you're applying too much feedback because you mask the signal that you're trying to detect.

And so we're in a place where we would like to figure out how to better optimize our feedback controls so that they don't mask the gravitational wave signal so much.

And luckily there is a community which thinks about how to optimize control systems, and they've been a great help to us.

But we're now at the limit of what I understand.

And so I'm looking for someone who knows more than me to help us improve this situation better.

And I hope some of our modern learning techniques and signal processing techniques can be used for this.

Yeah, I mean, I want to talk to you about that as well.

How are you, what techniques are you applying right now to the data and what do you hope to apply in the future?

We basically, everything that we can find, we just read a lot of things on the internet.

And everything that sounds clever, we want to use it.

So since 10 years ago or more, we have been trying every single kind of linear subtraction.

So we have, I don't know, let's say tens of thousands of sensors which are measuring the environment and the motion and these feedback controls and all kinds of things like this.

And we take each 1 of those and we compute the optimal Wiener filter, which is the, yeah, that's the optimal filter that you can apply to the sensor signal and subtract it from the data.

See all Y Combinator transcripts on Youtube

The Technical Challenges of Measuring Gravitational Waves - Rana Adhikari of LIGO