Amaury Micheli: Spontaneous quasiparticle creation in an analogue preheating experiment (TSVP Talk)

Yeah, that's that's about as good as it gets. Uh well, we can we can do the self introduction like this. Uh it is my pleasure to introduce uh Dr. Amari Miki. Apologies to both French and Italian speakers for any damage I I did to those words. Uh uh despite the Italian name, you're you're from France, I think. Uh and you were educated in France and indeed went to one of the very famous schools. Is that right? >> That's right. >> So you did your PhD in Paris two institute. One is a physics institute called Dubai >> and also work initute that changed name because my PhD is now called the laboratory >> they're actually cyclic permutations of names of French institutes. [laughter] >> Yeah. Um and so the topic of your PhD was on >> was okay half of my PhD was dedicated to the kind of thing I'm going to talk about today which is more um so a simpler version of this title would be some analog experiment an example of it and the other part is dedicated so early consideration with correlation >> so following your PhD you moved on to weekend that's a really weird filmation but yeah and I'm now working there weekend items in next Tokyo if you've never been come to this class we're doing a lot of things matter but also like biology >> so this is an interdicciplinary mathematical sciences institute in wakoshi >> yes >> next to Tokyo yes uh and uh who whom have you been working with there what kind of >> oh Okay. >> Working with other people in Japan doing some people in >> Okay. Well, that's a nice place. Okay. Well, uh it's my pleasure to welcome you here. My apologies for coming in late. Uh is analog gravity experiments an example the working title for? >> Yeah, the seminar. >> That's a good title. Thank you very much. >> I think so. Yeah. Um, first I'd like to say apologies. First, thanks for organizing last minute. Second is it's not really going to be a general introduction to and uh yeah and I've changed the title to this one uh for two reason. The first other one is very complicated and a lot of the to describing each of the words that go and also because there there was no gravity in the title as Nick Thomas said you didn't put gravity in the title you're going to be sell on people not realizing that this big stuff is actually has something to say about uh the talk is going to be divided into few parts. The first part is going to be really about explaining what is normal gravity and the second part is going to be about a precise example of another gravity experiment that I've worked on and let's say the reference for this work is this are these four papers here and especially this one where our latest publish okay so I will start by defining a concept called spontreation and I'll spend 10 minutes just talking about what is quantum creation why it's interesting and how does it appear in gravity what's include gravity and how can you try to simulate that in the lab that's actually then more specific example very quickly tell you about what I think are interesting perspective may be interesting um yeah that's we don't take questions. >> Um actually if you're willing to we can question my material was like >> is the room mic switched on? >> Sorry. >> Is the room mic switched on? >> Oh, this one. >> No, the the one up here. >> Okay, good. So, you you hopefully don't need a microphone unless you want to sing or something. Okay, so let's go quantification. So what is quantist particle creation? Simply put, I'll say it's the generation of particle in a region of space time. So at some point in space at some time which was first empty and then suddenly new to open that up. I will break this down in four ingredients for this part spontaneous creation. Four things that you need to know, four concept that you need to put together to understand. First thing that you need to know is our best current description of fundamental particles are that they're described by an object that is a quantum field. So when you think about having a particle at a point like it's an electron here really the description of that is more a quantum field. So an object that takes value everywhere in space has some configuration of that very scale. Okay. But this first fundamental particles are really some kind of configuration of an object called field that span all space. The second ingredient is that fields in general behave as collection velocities. So if you think about a field near Kai that has a value everywhere on the 2D sheets of the 3D space. So everywhere it has value and its value at every point can be understood as the oscillation or the movement of an oscillator at okay but that's the second the third thing that you need to do has to do with the quantum nature of that. I said this is general consideration for but this thing is a quantum. So what about one? I'll say that quantities typically have no well defined values. So for example and this the only example you really need to know about here the position of an oscillator. So let's take for example the classical case like here and imagine that you have a ball that is in a curved surface. This ball will have a definite can measure the bowl and you find it somewhere. That's a classical one. But now if you take the quantum oscillator, so it's a quantum bowl that is in curve surface here. If you do the same experiment exactly, repeated it perfectly and measure the position of this, you will not find it to be always at the same position because it doesn't have actually there's a fundamental uncertainty as to where this is. It's fundamental uncertainty here is it very tiny, but it's there. And typically here I there's a natural minimal level of uncertainty here that I will call vacuum fluctuation for the position of this oscillator. So it's you can try to reduce that but there was always something remaining meaning that okay you have some fluctuation naturally of the position. Okay so three ingredients. What is the last one? The last one has to do with a process by which you can amplify any type. So fluctuation can be amplified by classical or normal classical. And the process that I'm interested in here is that a parametric compation. So let me try to explain what it is. First the pitch is parametric complication is what happen when you're in the swing and you're trying to amplify the amplitude of your switch. That's a biometric amplification process. Let me try to explain that. Again, we have a bulb running on a curved surface. And at first, this surface is fixed. Fixed potential from the oscillator. It's going back. It's going up the right side of the wall, back down, up the left side, and just pass it. But now you can try to play around with that surface. You can try to open up that surface, close that surface while the ball is rolling. So, let's imagine that you start here. And what you're going to do is you're going to open up the surface. You're going to make this slope shallower. And you see that from the same starting point, of course, you're going to the ball is going to go further. And then it will reach a further light point. And when it has no energy anymore, it stops and it's going to go back in the other direction. And at this point, you're going to pull back your curved surface and it's going to give you kick so that it goes faster the other side. So compared to the previous situation, you're right back to the center, but this time you have more energy. So you're going to go more to the left. And you can play the same game of lowering the level of potential that has to go through to go to the left side. And this way is going to go further and further away. So by modulating this potential, you're amplifying the oscillation. Okay. So that's the process of parametric amplification and you can use that to amplify any kind of calculation that you okay so that's the four ingredients. Now let's do them together. So the first thing is do amplification can be amplified. So I told you that if you have a quantum oscillator it has some fluctuation on this position that you can't do anything about this. They are there there are tiny but you can amplify them. So let's say you have this ball here that is rolling about the bottom of the potential. It's not exactly at X. It can be a bit on the right or a bit on the left. And at first is just going around this potential with a certain you don't know exactly where it is, but it's about this central position with a small. But now what happens if you play the same game of opening and closing the potential this cursor as I was showing you? Well, in this case here, you don't know where exactly the particle is, but it can be slightly to the left or slightly to the right. It doesn't really matter. But when you're playing this game, if you're slightly more to the right, right, you'll go further to the right. And if you're slightly to the left, then you go much less to the left. And so the uncertainty, the space that there is in between these three ball here is going to increase. So the level of uncertainty that you have where the particle is when you play this game of opening and closing the potential right is good. So you exponentially amplify the uncertainty. Okay. So this way you can amplify this small uncertainty. Okay. Next let me go back to two entry. So two is the idea that the value of a field corresponds to having an oscillator at each space point. And let's say that here you take a classical field K and you say okay I want the field to have zero value that would correspond to all these vector being oppressed but if you take a quantum field denoted by K hat then you know that the position of this oscillator will fluctuate whatever you do so your quantum field will not be zero can picture that as okay wherever I measure my field I'm not going to find to be exactly zero maybe it's going to be more than zero maybe a bit less some degree of fluctuation so it's more noisy at representation. Okay, so that's all you need. Okay, so now let's put everything together. Let's assume that I start from a regional space with no particle. So this part I mean no particle guys means that I have a quantum field which is let's say zero on average like here but it still has some quantum fluctuation because it can never be exact. there's some fluctuation the value of the field is not exactly some fluctuation I'll call back so space value and now I can do the same kind of amplification process that I talked about I can change the environment of this field and by changing the environment of this field the space in which it leaves some I can amplify this calculation I will amplify this fluctuation by this process four and this amplified large quantum fluctuation here if you interpret them in terms of particle. What you will find is that now SL quantum fluctuation corresponds to a particle. It does the process of very schematically particle creation. Ites on the fact that particle are represented by field quantum field quantum fields always have function whatever you do. there's no particle in space and if you change space itself or the environment of this field you can amplify this partition to actually so that's the idea now I'm going to give you one example of this is famous I know most of the know about this example but it's the one on hawking radiation so Hawking radiation is the idea that black holes are not black and they radiate and that's the spontaneous particle creat radiation process. So what does it imagine that you have a space where you have some particle some matter that is very very dilute and this matter at first is very very dilute that's almost zero everywhere infinity is zero but this particle are getting closer and closer together towards the center they're collapsing and at some point they're going to form a black hole. So you go from a situation where all matter is very very dilute so there's no gravitational effect of this matter to a situation where actually matter is really really compact and then you form a black hole and then you have serious gravitational and during this process let's consider all other particle let's say at first I have no particle type for example here but again that means that I have a quant some vacuum fluctuation that lives in this space where the black hole is forming and when the black hole is forming This field of Kai will feel that something happened in this environment and that will foster amplification of this vacuum fluctuation here and actually if you do the map what you'll find is that this vacuum fluctuation are turned into particles that are emitting at the horizon that's really and this emission of particle is thermal and it has a certain temperature TH that characterize test. Okay, so that's one very very famous example of spontaneous partation maybe the most famous one but there are other there are kind of crucial inmology there is the generation of primities not going to talk about that it's important to describe the distribution of matter cosmological scale and the other one that I'm going to talk about is the preheating one where you generate all particles at some point in the universe through the preheating process it's a spontaneous creation process okay so creation crucial effect in cosmology ology why am I still talking about this because it seems to be very standard but there are two problems with this that we noticed a while back the first problem is observation we think it's very crucial in cosmology in gravity but actually we cannot observe it cannot observe it directly and the reason is that if you compute for example the temperature of this emission the level of this emission of the black hole it depends on the mass the black hole be forming and even if you take a very small black hole on astrophysical scale. Um then this and you take this mass to be as smaller than a reasonable mass black hole. So the mass of the sun you'll find that this temperature is 69. So 69 kelvin is interestingly very cold. But to give you um an estimate, even if the black hole with such emission was very close to us, we could see the we could know how the back is there. You could not see this radiation because the background radiation that we keep receiving from the cosmic microwave background that we cannot just shield. It's just there. It's orders of magnitude. So whatever radiation black hole is having is swamped by this. This is very very bad. So okay black hole evaporate black emits radiation but very we signal you cannot see directly and you have similar problems with these other spontaneous creation instances another problem is conceptual and I'm going to be very quick on this one but the idea is that although now I'm telling you okay the black holes actually emitted very very little radiation if you trace back in time where does radiation come from you look at how the frequency of that radiation changed uh when it when it was generated by vacuum fluctuation. The vacuum fluctuation that it corresponds to corresponds to fluctuation at very very high frequency of your field the original and you don't expect to be able to correctly treat this large frequency without quantity effect that I did not discuss here. So it seems like there's a self-consistency issue not inversion. Okay. Okay, so there are two problems here and that's where I'm driving and it was more or less born with this paper in 1981 by Uno which was called experimental black question mark. So it's actually yes um and the idea was okay this is a really interesting phenomena but the way we are computing things maybe have some problem and also observation we cannot see it. Is there a way to transpose that phenomenon into a lab setting where we can actually see it and check that first it's real and yeah check that is real and check that our computation that's the ID of no gravity and let me explain how you can so four ingredients for spontaneous particle creation and what what you will notice if you look at all the words here very quickly you will notice that nowhere is the word value. Okay? Because in fact, gravity only plays the role of the classical source. So you need some classical source to amplify your fluctuation and that's what the generation of the black hole does. But it doesn't need to be gravity. You can amplify this fluctuation with some other source. That's what's happening for example in the shrinker effect where let's say you have empty region of space but you have a very strong electricity. This very strong electric field will generate pairs of electrons and positrons in the space. But that's also spontaneous particle creation fostered by an electric field, not a gravitational field. That's very legitimate. And you tell me, okay, well, you could just do that then. You don't need to use gravity. You don't need to go in space. You can just do that. Uh the problem is that even in this case, you need to reach very very large electric field to be able to see significant creation and you cannot at this stage do it with these numbers. So there's an experiment that is trying to do it this way. You can there are people who have done other version of this recently and I'm not too aware how that works to you probably managed to decrease the threshold that you need to reach to see something but at first this number is too large for something. So you did not solve the problem by saying okay I'm not going to use gravity I'm going to use what can you do? So in another gravity what we say is okay so first we are not going to use gravity to source the amplifications too hard. We're going to use another source but in addition what we are going to remove is the idea that we need to study a quantum field with excitation of [clears throat] sorry we are going to study other generic quantum fields but their excitation do not correspond to so thea the case that we study as that of conate so very very cold atoms like quantum field and you're going to study the motion this is what I'm going to do and then you have a quantum field it's excitation are not fundamental particles but they are easier to create and other examples for example polarons and there has been many many other so in one box analog gravity experiments give you a repeatable and comfortable way to [snorts] observe spontaneous creation in the lab you can see it and second you can test the fact that spontaneous creation works even in a non ideal situation it's not pen and paper you have atoms in a lab there's noise there is environmental things that are happening and still if you still see spontaneous creation there it mean it's a good test of how it's affected by all these things okay I not going to talk about this you can ask me later so this was more of a very conceptual introduction to analogy in school but you can what you can actually do is give a more mathematical framework to the analogy where you write down according to the dynamics of a fluid and you show that in a certain limits the dynamics of this fluid the equation describing it are exactly the equation that corresponds to the dynamic of a field in a space you might know you might conver so I'm not going to show this because it's not necessary for the rest but just know that this is possible and you get that velocity potential the perturbation of the velocity potential of your field will follow kind of equation matrix here determined by type that's mathematical. Okay. Is are there any question um or or or it's already but in in any case. So uh next I want to introduce the specific analog experiment that I worked on and tell you about the model of this experiment before telling you about the result and I need to acknowledge the fact that I'm just one member of a much larger team and first there are the people really work hard the experimental team at the institute of so let's say this is south Paris basically and the P the PIs of the team are Don and Chris Westbrook And among all these people there is one person particular Victor who was done most of the precise experiment on the day-to-day basis for this project they are doing many different project but for this project was done most of the work that's why his name is on top of first and there's also a few team with Scott here who was my PhD supervisor and he's now moved to in France small smaller town in France and someone here that was a very important figure in another gravity called Ronaldo Parani and unfortunately he passed away before the completion of this work but we owe a lot [clears throat] okay so let's go so I want to tell you about another preheating experiment so first I want to tell you what is preheating so for this let me describe the history of the universe times run from the left to the right and I'm going to describe that as a two-stage process first you have a period very very The first period that we have some clue might happen is called inflation in which the universe expanded in a quasi exponential very very and that is nice because actually gives you a homogeneous universe I'll explain why but it also means that let's say you have some particles in the universe with a certain density suddenly all the particle the volume in which they are contained is going to grow exponential suddenly the density of this is directed down and you have a universe that is near to the whichever was there before inflation just stop. So you have an almost empty universe. So again I'm going to look at some some species of particle Kai and here you have no okay so that's the first piece. The second piece of the history is given by the standard model of so-called lambda where you can explain the formation of galaxy the structures of the galaxy that you see today in the sky and posate that all these things started by all this form of matter being in a hot dense plasma and that's where the nucleus that we see form and such that's the second part and you can see that there seems to be a gap here where You go from an almost empty universe to a hot and heating is the bridge between it's you go from the end of inflation to the standard model and this is a period that is not very well known actually it's hard to to know exactly what happened there but I'm going to describe the general sequence of events that we think happened. So going back to inflation, inflation was the exponential expansion of the universe. And in the simplest instance, you imagine that this universe was filled with just one type of particle, the inflat. And that there was as many particles in everywhere in the universe. It was a very homogeneous universe. Everywhere this inflatable has the same has the same value everywhere and this value is slowly decreasing and this loss of energy of the inflat the value of the platon here is going down slowly is what's powering this quy exponential expansion. The is quy exponential because the inflat is losing energy expansion. That's what's causing. But at some point it's going to stop. The inflat will start rolling quicker on this potential with quicker energy and will oscillate at the bottom of this potential. Then inflation stops. So here again you're in a situation where you've emptied out your particle. So all particles are described by your field. There are some just vacuum fluctuation. The minimal level of fluctuation field has. But it lives in a universe where you have this field that has a value that is oscillating everywhere. So you somehow have an oscillating source everywhere in your universe that is coupled to this field. And you're going to have the same process where this field is going to feel that the environment is shaking some. And this is going to excel this vacuum fluctuation to actually create particles from the back. More precisely, you have a parametric resonance process for all the matter still here. And what you create is that you create pairs of particle that are going in opposite direction with an exponential rate. And that's so that's and that's the first stage of reheating. Preheating is the first stage of preheating that's there there's there's several reason why preheating but I said preheating is before the preheating the part and then after that these particle that you created are going to scatter and they're going to termalize and that's what we don't need to understand the second part of this process here of the sequence of element is very nonlinear so this means two things it means that whatever happened there and any kind of quantum correlation that might be there is completely blown in the final state it's very hard to observe because looks like just an odd dense super particle. So you can't really probe by looking at the last the final stage of reheating you cannot really probe this quantum part of the process first thing and second thing even if you want to describe the second part of the process where everything collides it's actually really hard really nonlinear system and even on the latis means even when you do simulation that's actually not easy to simulate even this process so for these two reason it might be nice to have a system that follows a similar sequence of when you start from vacuum fluctuation you find you produce particles quite particles they scatter around and they these analytic experiment you don't have to make computation you just do the experiment and it does the sequence of event okay and you can try to stop it at different time and see okay that's where it's happening here that's what's happening there that's the motivation behind having an analog of preheating now at this stage really what I'm going to talk about is preheating I'm not going to talk about the second part. This is harder and we've not any question at this stage. Okay. So now let me explain how you do that. So that was like very cosmology. But now how you do that? So the way we do it is we use a gas of cold helium atoms that we trap in a magnet. So you imagine that you have one laser actually we have two lasers you have two lasers that traps your atom in a very tight region here and you have one laser that is stronger than the other which means that your atoms are trapped very uh strongly radially but there are the trap is a bit looser vertically so you have almost 1D gas we have this almost one chain of atoms there and this gas because it's a very cold atom gas is well described by the quantum density field row and the quantum density field. So you have quantum field described the gas and we are in a situation actually where the gas is in a quasi condensed state which at the technical level means that this good approximation to treat two quantum fields by saying okay there's a part of this field that corresponds to condense atoms and that's going to be described by two classical field row 0 and 0 to the condense part of the atom. So you can imagine that row 0 just a background density. So everywhere the same in the gas on top of which you have some pointation data around. And here you can see that you already have an ingredient that we need for particle creation process which is that we have row 0 that is going to play the role of a classical source then it's going to act on some quantities delta okay but you need of course these two objects to talk to each other and actually you don't need to do anything for that because the dynamics of this row and pt if you write is nonlinear so the equations if you write down the equation will naturally mix the green objects, the classical objects with the blue object. Here I want to emphasize that this is generally the way people design analytic experiment. Take a system that is generally quantum. So you know it's really quantum you have quantum fields then you identify a situation where a part of the system can be classy say okay that's my classical background and I'm going to study all this classical background acts on my quantum and can excite this backation this is what we are doing now what is the exact process by which we make this classical part act on this quantum field explain that the way we do it is that we simply modulate a bit so You have your gas and you're going to modulate. So you're going to compress it. Compress it expand. What it does is that it's going to modulate the density go. So you have an oscillating density background and so your quantum your perturbation here on top will feel this oscillating density background and will react and actually what you expect is that this compleation are going to create opposite momenta waves on top of your so you have waves of atoms with a resonant frequency that are created exponentially on. So that's the basic setup and now I'm going to move on to something more. Oh yeah, sorry something that I forgot to mention but I already emphasized that a lot that this creation here is generally a spontaneous creation because you can really have quantum and I can really target the back talk about now I'm going to give you a more uh formal version of this. So here is going to be a more technical question. So this is Newtonian describe the gas where I've assumed basically that okay my gas is really 1D. So I can really do whatever I want on the radial side and what I'm doing can control it very well and I only care about the dynamics of the gas very longitude. So I have a one dimension here and the thing that I want to emphasize is that what I can do on the radio part this compression of the gas this modulation of the trap frequency is packaged in this G of T here that controls the interaction. So here this G of T is related to how I compress that. Now this is the general but I'm really interested in what's happening to my perturuation on top of the background. So I expand this participation theory. I expand the delta delta theta and I collect in order of thetas. I have my background dynamics. Not going to care about this. I have this second order amonian. This second order amon is what's going to define what I call quite particles and this contains the physics of spontaneous creation. But then you have actually infinitely many orders and this order encode interaction of this quite here the quite particle are free but there you have interaction of the quasi particle and I'm also going to use so next I will spend some time describing spontaneous creation in this setting does it happen in this setting and then after I'm going to say okay but what happens when you start to turn on interaction how does it affect this spontaneous creation Okay, so that's the second order in H2 written plainly. And what you can do here, so this is telling you about the coupling between the density partation on top of the gas and the face pertubation on top. Yeah, what you can do is reccast it as a sum of collective motion. So mixed motion of density and phase and you can write it as sum of or quasi particle defined by this annulation operator P of K here which is a mixture of density and face. So here instead of saying okay I have waves on top of my gas of density and wave that communicate I'm going to say okay I have a certain collective motion with a certain wavelength K and with a certain frequency magnet and I have 1 2 3 this motion so I'm going to describe my state using a certain two types of numbers the first number is the number of excitation of each wavelength that I have n here and the second number is not If you have collective motions of your gas, different type of collective motion, but this type of motion can be correlated. This one correlation I'm going to care about is the correlation between corrective motion in one direction K and in the opposite direction. This is given by this CK per correlation. You have other types of correlation but for what matters here I'm just going to consider two. These two numbers are really the main players of the rest. Okay, so that's the description and now in this terms what is the expert really the expert what the experiment is doing starting there. So again G this this coupling is related to the dropping of the gas. So you start with a fixed trap then you modulate for a bit. So you make G oscillate and then you start modulating and you look at what so if that's to you you can think about in and out state that are connected by this volume of transformation or you can just say okay I'm going to evolve this quite particle operator and find what happens. What you'll find is that what this quasi particle operator is initially is mapped to a mixture of quasi particle operator EK and the the creation operator max and that correspond also to two that's another okay so that's the general dynamics but now you can also compute what happens to a specific situation and what I care about is really the case where I start from a gas where I do have some motion already. You would have no motion of your gas if your gas was at zero temperature. But because you're not at zero temperature, you have some excitation to start with. So you have a thermal amount of excitation here and pH, but they are incorrect. So the motion in the K and the minus K direction are not correlated. This is zero. Now you're doing this process where you're exciting your gas and you look at the excitation that you have after that. And what you'll find is that you have amplified the excitation for more excitation to be clean and also that the excitation between of the collective motion with a wavelength K and the wave vector K and the wave vector minus K are not. This is the result of the excitation process and instead of just showing algebra here going to show you a video that show that shows what happens when you do this numeric. So I'm going to solve the dynamics given by the sin numeric and I'm going to say okay I modulate this g of t cenosidally at a certain amplitude a and with a certain frequency omega m what I expect is that I'm creating pairs of quasi particles exponentially at a rate given by a * omega m. So the more the stronger I modulate the faster it grows and I'm only going to create pairs of particle at some resonant frequency here. So collective motion at frequency omega k that's rationed. So in this plot here you'll see as a function of K and K. So the number of particle in red and the correlation between particle with momentum K and momentum minus K in blue and I'm starting from a situation here where I have some excitation because I have a gas that is not at zero temperature but they're not correlated. And what you're going to see is that as the number of oscillation increases, you're going to have a growth of both the correlation and the number of particles around this blue line. That is the resonant. Good. See that here you have the green line that rolls around two points. And if you're careful, you see that the red line below also grows. So you're creating particles both in the momentum in the mode K and minus K in pairs and because CK grows as well these things are correct. Okay. So it seems that it's working if the experiment if I can do that experimentally I would create pairs of of quite particles of my gas right but there's a catch which is that if you just look at this what you can tell is okay I've generated quite particle but there were some to start. So what can I tell that really what I've done is a really genuine quantum process and I just didn't take some classical motion of the gas and converted to another classical that really I took the absence of motion in this mode and I generated something out of that. So it's a bit subtle but the way to do that is to use entanglement. So entangle moment I want to check whether the mode can minus can angle which mean that they are correlated in a way that is not a lot of possible and in this case to check entrelation ck is larger than the number of excitation in the so when this delta k is negative the state is entangled and you can check that this delta k using the formula that I give you earlier composites in two terms there's one term here that depends on how much classical excitation we have to start with is positive and there's a second term here that can make this delta k go negative which means you reach entanglement which if you do the algebra you'll see comes from commutator so this is really the bit that talks about quantum vacum fluctuation and that is not classical if I do this computation class this would be zero and delta is so if I want this to go negative this has to be large and so this can only go negative negative if I really pro back my goal is to make the experiment and measure a delta G that's negative and that's a generic return analog experiment you want to create excitation and you want to check that the excitation that you created are because else you're just knowing that okay I can create excitation but did you create them from back you cannot you want them to be intact and if you check that plot here what you will find is that the green line for CK K is above the red line for NK. So actually if the experiment really does that then you generate problem is that the inter the experiment is harder than this and in particular there are some interactions in the experiment that will model the picture a bit and maybe prevent you from seeing. Um so I only have I can use how much can I use? I I'm at 14 minutes from I think >> keep keep going. We started slightly late. [laughter] >> Okay. Okay. Then I'm gonna explain the effect. It's fine. Okay. [clears throat] It's going to be very brief anyway. So I remind you that all the process that I was talking about here was at the second order of partation theory or first order. Now I'm going to the next order and I'm looking at the interaction between So what we've done is that we've identified what are the dominant process for this gas and the dominant process are these two processes here the life process where if you have a quasi particle with a certain wave vector K it can split into two quasi particle one with a K minus Q and one with a momentum Q or the collective motion with wavelength K and combine with one with wavelength Q and give you one So this means that even if I I've created some excitation with momentum K, they're not going to stay there. It's they're going to leak out to other. And we computed what is the effect exactly on the number and correlation of the excitation of these two processes. And what we find is that the number of excitation and the correlation of this excitation decay exponentially at a rate given by this gravity which is proportional to the temperature of the gas inverse. So if you put the excitation in your mode they will just go. Okay. The next step is to look at what happens when you both complexation. So you do your process of spontaneous creation but at the same time the excitation you put there go away. Okay. So what is the sweet spot between these two process to be able to say okay that's the right regime of parameter where actually even if there's some interaction in the system I'll still be able to see particles to see excitation sorry and to see that this excitation entangle not generated. So that's the second part of the work with Thomas Scott and uh what we found is that if so it's a very simple relation it's very simple inequality which is telling you that if GK which is the rate at which you pro excitation which is related to a hard pump basically is smaller than this dissipation rate here which the particle go away time the amount of classical excitation you start with if you satisfied if you if you're in this regime then you will never for which means that you generated quite particles but you cannot check your generated vacuum so your experiment fails so what it's telling you is which is kind of obvious is that you have to look at low temperature but also you have to work at high density and actually these two things don't go together so it's difficult and that you have to modulate strongly to have a large decay to overcome this decay rate in this population but also you have to modulate for a short amount of time that not from the inequality but that's from other things that we can talk about later. Don't worry about the so with that in mind you can try to do the experiment and and see. So now I'm going to move to the experiment. So what is the measurement proc? So I remind you the idea is that you have a gas that is trapped by two laser in a very small region. What we do first is that we are going to modulate this trap by changing the bar of the laser. Basically by the bar of the laser the trap is more or less stiff and so the gas is going to os. So here it's real data here you see the oscillation of the size of the gas. So you have an oscillating profile for the size of the gas and at some point you stop that and then you open the trump. So you release You release the laser. You turn off the laser and then the gas will expand and it will fall. It will fall on this detector. We let the we let the atoms fall on the detector and we'll count the atoms. So what's happening here during the fall is very interesting because imagine that we had atoms. We created atoms. We created we created motion of the atoms and in particular we created pairs of motion some going up some going down. Okay. So when you open the trap, some of the atoms have a upward velocity, some have a downward velocity and most of them basically are static. So the one that were going that were going up when you release the gas go up and then they fall. The one that were going down will go straight down and the one in the middle are just going three with no initial. So you have now three clouds. You have the top clouds here corresponding to the one that went up. Middle clouds corresponding to the one you type and the lower correspondence. So by measuring the time at which your atom arrived on your detector, you can map that to their momentum. So you can either okay that's the K mode that's the minus K mode and this detector here it's called microch plates that allows you to detect single atom you can count atoms to single one atom that very briefly some experimental data here to show what I mean so this is the density arriving on your detector as a function of time and you see that you have the first peak [clears throat] big peak corresponding to the one in the next eye and a second almost the same I as the first one which makes sense we expect now it means that really the observable we have access to this setup is number of you're just counting that's what you measure saying that many atoms with this momentum that many at that momentum and I can look at the correlation of this number of atom and I can repeat to compute this kind of So you want to do physics. First thing is that although you detect atoms, what you really want to know about is their motion in the trap. Okay, you know the atoms are there but you want to know okay what were their motions when they were trapped and when I excited. So there's something called polar evaporation where if you open the trap slowly enough basically the you have a map between the quasi particle excitation and the so if you have atoms if you have like 10 quasi particle with a certain wave vector you will find 10 atoms with the same wave vector when you count all that's really nice which means that really what we have access to is under this assumption correlation of quasi particle numbers which is really what we want to And then we need to count this positive particle numbers and check that we have the creation of one. So the first resulted here is shown here and it shows you the number of atoms which correspond to the number of quite particle in the mode K minus K as a function of time. So this time correspond to let's say how much excitation you have for what you see here is that this grows and this is a lot it grows exponentially that's what we expect and in addition if you check the rate of this growth you'll find that it grows at the expected rate before related to the so okay first stage result we creating particle the way we expect them to create them we can also play around with some parameters but that's not enough because you want to know that you're creating pairs of particle that are entangled. So how do you do that from these numbers? You're going to make some we are going to make some assumption on the state of this excitation. So the center gian homogeneous state for the technicality and then under this assumption you can compute that this number here which is you first look at for every run you say okay here at 10 excitation with mod k 10 excitation with mod minus k and then you repeat this correlation count for every experiment that gives you this number and you normalize by the average number of excitation k average number of excitation minus k and if you compute this you and relate them to these two points I mentioned before this CK the correlation between K and minus K and this NK I remind you that your state is entangled when this CK is larger than NK so basically here your state is entangled when this correlation function that you can measure is larger than two okay what you want to do is count get some statistic measure this and show that not only did you produce particle but this number is larger and that's what we've shown in this video. So here you see this color vector G2 shown the the blue dots are the experimental data as a function of the mean detected atom number. So basically this number will grow if you produce more if you produce more excitation. So it's it's not so different from the time I was showing before. What you see here is that okay first when you have a low detected number of atoms you not pump really hard it's it's really hard to see right it's not very clear that G2 is above two you've produced something but to as as this number has a large uncertainty but after a while this number clearly sits above two which means that you do have G2 that is larger than two and you did create an intended state okay and so that is a signature that we successfully obtain have spont Continuous creation of course I okay >> question >> yes >> sure can we can we cross check this assumption because the this this estimate is based on assumption the state is center >> yeah okay very good so um it's difficult to check and we did not check it be clear I so it's not completely true in the sense that I did check it using simulations where it seems that you do Um but experimentally we did not check it because it's really hard to check nonverity. You would have to check correlation of like three or it's really hard actually this is a simple version of the assumption but Victor the author here the experiment actually written a theory paper that also went to where he has shown that you can relax this threshold. You can relax the assumption and get a larger threshold uh which is shown in red. So the real threshold that we use is this one where you still need to assume but there's some assumption and you see that it's not always two here it's it's it's a bit harder but the true answer is this was not computation no so it's an assumption okay and so under this assumption and okay there's also the fun operation that we did cannot check we demonstrated spontaneous creation Let me wrap up here. Um, some perspective on the on the experiment itself. Um, you could try to push the analysis at later time. Here I've just shown you the creation, but you can care about what happens after how particles interact and how you get to a thermal distribution because you do kind of get to a thermal distribution. That's actually what they've done in the first version of experiment which fail and they just pump too hard and go to a thermal state. >> [clears throat] >> But what I really want to talk about is more perspective of the field itself and generally ask what can you learn from this kind of gravity experiment if you care about you care about not just atomic experiment. So there's the first case a where you see what you expect and it's already interesting is for example this vacuum amplification phenomena spontaneous creation phenomena has only been seen in a handful of experiments meaning four five total. Okay, there's something that is so relevant in physics. So it's already nice that you get that. In addition, there were some confirmation of other key effects of quantum theory like to pull vacuum decay where they were able to design a situation. So not an expert on this, but they were able to design a situation where you could make the QFD computation and check to an extended QF computation against actual observation of the system. So that's nice. Uh what would be nicer that's to be that never happened so far is to see things you don't expect. You have a model for your gas. It somewhat matches to to an extent to a point of the curse bas and you thing that you don't expect in your experiment and if you're lucky enough maybe this is something that is not just specific to your system but also exist in the model that you could import in the gra that would be really interesting that has not happened. And I want to end with a question for the audience for the gravity people in the particular which is do you know of any problem in gravity where okay it's conceptually interesting you get a number you get a prediction but you cannot actually test gravity this prediction and you could maybe use analy to try to test that the concept exist outside of gravity okay I close here >> so thank you for very clear and interesting in questions. >> I think Johan went up first. >> I have that but let's say like if one goes like a step back and consider like the derivation of the analog metric in the most connected base like usually the derivation of the metric is done without taking into consideration the back reaction of the quantum field on the classical part. But for example in your work do you have to consider that in order to get your results or no or it is not necessary. >> So uh yeah okay I'm not so yeah so you're absolutely correct. I just want to get back to the metric. So this results under many assumptions which are not necessary satisfied and in fact we don't have a number. So it's there's no formal analogy to this level in our case and not so much because of backation because we're working on where this construction you can define an inverse metric it's not an inverted metric this way so you can have this construction but although you can define an another metric what you can do you're still doing backification genuine so at the conceptual level Now this construction is about free field on the fixed curve space time and so you don't have back reaction and you don't have the dynamics of space limit there are some attempts try to treat back reaction in an analog system and see how you expect back reaction gravity case I am not familiar with this attempt and but in some and I think it's indeed an interesting direction to come back to. >> Thank you. I think you also had a question. >> Um, so first of all, I wanted to hear a bit about uh if we learn anything about reheating through this. >> Yeah. Well, not necessarily from this exper experiment, but in general from analog systems if there's uh because as you said, breathing is kind of a nightmare to >> Yeah. So, so okay. So I think the So I said that I said okay this is really hard to handle. So what if >> so I think there's there's that's why I pitched it. So do I believe [laughter] >> and the thing is actually so there's there's a bit there's a bit of a problem here. So the the I think the state-of-the-art in terms of experiments is people were they were able to produce um excitation here I showed you that excitation in two bands but actually you have infinitely many resonance band and they were able to see many resonance band and see the interaction between this resonance band um and so that's that's nice now the mapping to reading is a bit complicated the sense that okay you have a parametrication process but in reading We don't exactly know what is there's many possible scenario heating but it's expected that the expansion of space plays a role in reheating as well the preheating part also which means that don't really address a specific the resonant B shifts so we don't get the exact same excitation spectrum so it's not so much that many people will be able to quantum simulator field theory and this experiment that I'm telling you with several of them they were able to map what they were seeing to precise field theory and check the vertex of the entire this field theory become the simulator the question would be more like is this field theory that I'm getting mapping one to one extre this one second yeah is it is it is it an exotic one or that also to get the expansion space is tricky >> there was Another question or not? >> You have another question. [laughter] >> Okay, go on. >> Um, so you mentioned also at the beginning about um, you know, one of the reasons why people started looking models was the transunction problem. >> Yes. >> And do you think one can answer like people consider it's kind of sold. So what's happening here? So I I should mention that. Yeah. So I it's not on site anymore but there's a there's a paper by mention which is what have you learned from studying platform that talks about this and so one of the thing is that here uh I'm I'm telling you that you can describe this both gas using quantity but that's only to an extent because at some point you have atoms at some point you have a cut you have a cut in your j and yet you're spread access creation So you have a natural caloric theory and you have no transplantation problem. It's still missing quantification. So that's that's one level. In the second level you can do the theory. You can say okay I have the one breaking the UV what what does it change to radiation spectrum and the answer is that it changes the radiation spectrum but it does not kill the paper by that. I was asking because it feels like what you're mentioning now is more about you know we have a cut off and we can see that despite the cut off everything seems to work >> but the way I usually think about the transparent problem is really about you know accessing a new regime where you have to ungravity regime and in this situation this is not really a gravitational system right so I don't I wanted to hear. >> Okay. Are you saying that you're pulling things that are in the full column gravity regime to a non gravity regime? This gives you a mapping to access this colum gravity. Yeah. So in this case be like oh actually I'm starting to see the particle like behavior fact that it's not so it could be a different type of so when you you're looking to some UV physics at the cut off physics which in this case is like mass physics in the other case more physics so I wouldn't know if I would trust the fact that here I would define as a >> okay so absolutely no so there's what What what analog system can tell you is like under the assumption that the model or gravity does match the model in both case I can run the same mathematical procedure this same mathematics you can write the same mathematical procedure there's one case for in the sense that I make a prediction so all of this runs through bearing the fact that you need to assume that the model that we have for a black hole is a good model that that's you cannot do by looking at that's something >> okay Yoshi >> yes can I ask >> sure >> yeah more question you mentioned that the original idea has a conceptual problem related to quantum gravity is there of designing or changing the experiment towards that limit >> sorry I think we could changed a lot the way we're doing relative was proposed originally by >> so originally in the original frame kind of idea if you mentioned that the the idea has a conception problem related to cont gravity right >> yes so that's the trans problem >> okay so in a con physics what would be the the the asytoic kind of manipulation bring the system towards that kind of limit >> um so Okay. So that's that's that's related to what Luka was talking about. I would say let me put it this way. So the conceptual problem here is that you assume that the system that you know perfectly well the gravitational system that the field will behave the way it wants uh in the UV and and and and that's not going to change. That's the for example the spectrum of excitation that's an assumption we don't know and in the cond matter case so in the benste case you actually know oh it's going to be you know here and you know it's not true you know that actually if I draw my my excitation spectrum first it's linear at some point it's not linear anymore quadra and uh and so here we know that it's not true and Despite the fact that it's not true, it still work. You still see that. So some here the conceptual problem is not a question. It's a fact like the field the quantum field is not a good description of the system arbitrarily large energy and still the phenomen. I'm not sure answer the question. >> Maybe I don't I'm not addressing my question but I >> No, no, no. It's okay. It's okay. get. >> So, I have a question. Um, how much longer are you here? And how should people find you if they want to talk to you? >> So, I'm here until the 29th. If you want to find me, uh, find TS offices. >> That's 03. >> On floor EF on floor EF. >> Yeah. Okay. So, then I suggest we thank Amori again for this job.