Welcome everybody. This is a very special and urgent lecture episode featuring my friend and professor at my alma mater, Rick Gaetzkill. Joining us from Brown University, Rick is the spokesperson of the LZ Collaboration, which stands for Lux Zeppelin, which is located at the Sanford facility in America. Denny Sanford was a friend. He lived here in La Jolla. I met him many times. We have some conversations together. Really wonderful that you guys have done so much to celebrate this great contributor to philanthropy and also to our understanding of the cosmos.

So, Rick, I'm really excited. You're going to talk to us about one of the most exciting announcements in very recent history, and that has to do with an event that was announced just about a week ago. So this is really urgent and emergent, as I said. So, Rick, please take it away.

Brian, thanks so much. You know, it's marvelous to see you. So the Lux Zeppelin experiment is the sort of latest in a line of dark matter direct detection experiments that I've been involved in over the last 40 years. It always helps me to gain a little bit of perspective looking backwards as well as, as you say, what we've announced from the work of LUXEP just in the last week or so. It is always a little sobering that if you sort of just consider what's happened in the last 3 years, for instance. The universe has actually expanded measurably in the sense that it's about a 5th of a part per billion, which I always find quite remarkable. You know, just the universe continues relentlessly doing its thing. But from a dark matter perspective, we've been improving our sensitivity looking for so many or testing so many different models of dark matter.

And as you'll see, you know, in a moment, just how extensive that testing has been. And then my— one factor is my weight, which actually for once has fluctuated down and not up, which, you know, for 3 years is not bad. I am going to take the opportunity always to sort of remind people a little bit about why we're still engaged in looking for dark matter particles, even given that we've been looking for them for 40 years. And then, of course, as you've said, we are actually sitting on an event. An event has shown up in the LZ. Now, as anybody who sort of follows more closely rare event searches, you know, one has to recognize that a single event is, you know, simply a— at the start, if you like. And, you know, either statistically or in terms of our growing understanding of of the conditions that might have contributed to the event, that the interpretation is, you know, could either be that we're going to see subsequent events consistent with the dark matter hypothesis, or it could be we're going to see events consistent with some other, you know, more sort of, well, exotic but mundane at the same time background. And this, you know, as experimental physicists, is something we are absolutely focused on always in search The talk, or our discussion, there are so many abbreviations, buzzwords, acronyms these days.

I think I'm going to forego the usual jokes I make about some of these, other than to mention when people hear the word WIMP, for Weakly Interacting Massive Particle, which is after all these significant, you know, number of years that we've been trying to test such a hypothesis, You have to understand that physicists do have a little bit of a sense of humor, and the WIMP acronym actually came about at a time when dark matter could also be solved by MACHOs, which were massive compact halo objects. So there was a very deliberate, I think, sort of element of humor in the WIMP, MACHO. Now, MACHOs have actually been something that we've managed to test that particular hypothesis, and it is significantly— the amount of dark matter that could be satisfied using a Macho hypothesis is very much smaller and certainly would not solve the entire dark matter issue. Now the other thing I'm going to do a little bit of is I will end up mentioning supersymmetry, but again, I'm not going to get too heavily into the acronyms. And let's— anyway, let's move on.

I—

the other thing is, as you will see, these slides in fact do not have any AI used in their preparation. Certainly with respect to prettification.

Wow.

There is an interesting aspect of, of AI in rare event searches when it comes to the analysis chain. And while this is something that, you know, like many scientific, you know, experimental that we are, you know, doing a great deal of investigation in, we also have to subject any points in the analysis where AI has been used, you know, we have to be very rigorous about understanding how it's working, which is not always the way that people choose to use AI. We're a very constrained case, so we are excited to see what can be done in terms of making better and better use of data that we're taking. But equally, we, you know, as you might imagine, if we're talking about a single event, we don't want to be— we would not be in a situation where an analysis chain simply popped an event out of magic because of an AI component, and that that was what we stood behind. That it's simply, as you might imagine, you know, would not work like that. And that's not something that— The other thing is, I have to, you know, let's not talk about the AI. Let's talk about the natural intelligence that we have. This, like, you know, is a large scientific collaboration, relatively speaking.

There are over 250 authors on the latest paper that we've done. This is a photograph taken from a collaboration meeting here at Brown a year or two ago, but we have regular collaboration meetings, as you might imagine, and this work takes or combines the input of a broad range of people, both DOE institutions here in the U.S. and also universities and labs from other parts of the world, including the United Kingdom. And of course, we actually are based at the Sanford Underground Research Facility in South Dakota, which is a US deep underground lab that we work with and are supported by very closely to do this level of leading science. I always mention the other thing you have to bear in mind, that is, we're not just training people within our organ— within LZ to do astrophysics and cosmology. Obviously, that is something that we are trying to answer and break, you know, questions, answer questions there. But of course, much of the training that graduate students, postdocs receive extends into a broad range of other, you know, disciplines, including sort of analysis of massive datasets or— development of sophisticated simulations or indeed machine learning. And this is, it's an important part of our process.

Now, for me, I've actually, I'm about 40 years now trying to answer one single question, which is, can we fully identify what is dark matter? And on one hand, you may say 40 years, that's ridiculous to spend that long on a single question, but I— Much as we look back in history of science, it often seems like things are coming thick and fast. If you actually really break it down and look in specific sort of chimneys or specific areas of investigation, there's a much greater distance between the great— then of course 2 or 3 great discoveries come along at once, and then again you have this sort of extended, you know, period of time. If you dart around between different subdisciplines, then of course you can start bagging or seemingly get progress happening at a much higher rate. But, you know, look, science, you know, research is extremely demanding. And demanding because 95-plus percent of the time you're not going to get a negative result. And we held a sort of feedback session with graduate students and we had written feedback from a few students— this was a couple of years ago— who actually said they felt their supervisor was deliberately giving them work that did not give them— did not give answers. You know, the research was producing negative results. And we just realized we'd utterly failed our students in terms of really telling them about how research works, which is most of the time—

Yeah.

You know, you come up with a question that is well motivated, and that's, you know, how the, you know, using the community works. But it is so challenging you come up with well-motivated questions, but most of the time, if it's good research and you've got an imagination and, you know, a well-motivated imagination, but nonetheless a good imagination, of course it turns out that nature completely ignores any—

That's right.

Nature doesn't care. And, you know, I wish we had time to sort of discuss that, but, you know, even in the short 40 years that I've been addressing this, you know, so many Beautiful models have come and been ruled out for how we might solve dark matter. And you shouldn't— but what you shouldn't do is interpret that somehow that the entire question is broken. It's just simply that nature really doesn't care about beauty often. In a sense, certainly in the sense that when we construct a theory and try to say this is well motivated, Yes, it's well motivated. Yes, it's consistent with existing measurements, but nature doesn't need to pick it.

That's right.

And of course, it'll be— it will probably be very obvious. It always is when you get an answer and you look back and you go, ah, yes. But that again is something you have as a scientist to, you know, to recognize is not really. So I've worked in a number of labs. I often— one has to change one's clothing a bit. I was, you know, back in the '90s, for instance, I was, you know, northern Minnesota in the Soudan Mine, the CDMS2 experiment. We then moved in the noughties, the 2000s, to Gran Sasso. It requires a bit of a wardrobe change, of course.

One has to look more like an Italian physicist when one is in Italy. But we did a great deal of work on the early days of liquid xenon detectors. And then, as we mentioned, latterly we have been working on both the LUX and the subsequent LUX-Zeppelin. experiment in the Sanford Lab in South Dakota here in the US. And of course, one needs a bit of a wardrobe change if one's going to work in South Dakota. That's right.

The credit card state. The credit card state. Which is why Sanford was there, right?

Well, most— when I explain to people what it was like, the banking industry back in the late '60s, '70s, people don't really understand quite what an innovation credit cards were, and then equally how you had to restructure the legal environment in order for this clearing process to work. So it is fascinating.

That's right.

And only a limited number of states, I think, really understood this, and of course South Dakota, you know, did even at an early stage. I have to confess, you see, for 4 years I was actually an investment banker, you know, back in the '80s. Oh, that's right. So I actually, I think I understand a bit more about, you know, how finance—

You left it, you left that world for the high, high-pay world of experimental physics and professor and professing.

You know, back in the '90s, the BBC came into the lab in Oxford I was working at trying to do a story about how people were leaving academia for finance. And they were shooting a lot of B-roll. They were interviewing my head of department rather than me. I was standing in the background and I nudged one of the assistant you know, producers and said, I have to confess, I actually came out of finance back into academia. You know, is this going to wreck the story? And they just told me to shut up and just continue trying— just to continue twiddling whatever knob I was supposed to be twiddling to provide a backdrop to their B-roll, you know, for the story. But anyway—

Yeah.

So the dark matter itself is, you know, absolutely central in our attempts to build up this overarching model of our Milky Way. And I think everybody is probably to some degree familiar with it. What's fascinating is if we're talking about the overall composition of the universe, and we're now in a situation where around 25% of that total composition is dark matter. Now, that you might say, that sounds like a bit part, Rick. But no, if you want to understand bit parts, it's you and I. We're made from conventional atomic or baryonic matter, and that's less than 5% of the total composition of the universe. So you and I are the flotsam and jetsam of the conventional atoms, the protons, the neutrons, the electrons, flotsam and jetsam on a much more substantial matter component, which we know is there gravitationally. We have over the last— well, I think you can argue almost 100 years now of observations.

Where we have determined that the way that the galaxies and clusters of galaxies are behaving, that you clearly need to insert a great deal of non-luminous or dark matter, matter that doesn't show up directly at the telescopes but does end up— or some property that seems to actually affect the gravitational behavior and the gravitational evolution of our galaxies. But one of the things I've been lucky to sort of live through is also this tremendous change we had in that overall model, you know, in the '90s, and that we, you know, we have been refining the model of is that you also have to have— find room for about 70% of a thing called dark energy.

Right.

Which is a component— whereas the dark matter is helping us understand how gravitational formation at galaxy and cluster of galaxies are evolving and how they're holding together, dark energy is a very rarefied but finite term that appears to be pushing the entire universe at an ever-accelerating pace apart. The challenge is that right now we have these titles, dark energy, dark matter. We know dynamically how they influence the sort of evolution of our large scale in the Milky Way. The problem is that we don't know what either of these actually are yet. And that's what we're doing. We have this LZ experiment is very much targeted at trying to directly identify dark matter particles. Now, if the dark matter is due to particles, then they are— the abundance is large. We don't know the mass of the individual particles yet because we haven't directly measured them.

Our theories, in fact, span a very wide range of possible masses. But for the kind of particles that we're looking for with LZ, you know, you're talking about of the order of 1 to— or 0.1 to maybe 100 of them per liter, or per, you know, about this sort of volume. And they are moving sufficiently rapidly that through your body, you've probably got about 100 million of them moving through your body. Now, they are not interacting with your body, except in the general sense of providing gravity that holds the galaxy together.

Right.

But their actual rate of interaction is so weak at this stage that we now One way to imagine it, you're gonna see sort of cross-section numbers, but one way to imagine it is if we fired a single dark matter particle, hypothetical dark matter particle, through lead, we could actually pile that lead all the way out to the closest star beyond the sun, you know, sort of, you know, Proxima Centauri or even Alpha Centauri, and actually go about twice, more than twice that distance, so 10 light years. And the— even though this dark matter particle was traveling through lead, less than 50/50 chance it would have interacted at that point. So that's a very weak interaction, but because as you can see the fluxes are large, 100 million per second through your body, and if we use detectors that are massive enough, then the probability of getting an individual interaction over a matter of weeks or months starts becoming finite.

Yeah.

And that's what we've been doing. So that's how we're trying to study these dark matter particles. Now, the WIMP, idea really is nearly 50 years old now. There were a number of papers that were seminal. I can't list them all, but I think Lee Weinberg— I'm just pointing out that it was given— if we go back to that period in the late '70s, early '80s, we were just understanding how important and experimentally verifying the W particle and how weak interaction physics and electroweak unification, and in fact, of course, really bedrock of the standard model of particle physics, that it seemed very natural for weak-scale physics to provide a dark matter particle. Why? Because all particles are equal in the early universe. They're all being created and destroyed. As the universe cools, the masses of these particles becomes more relevant.

And what was fascinating to realize is that a particle with an interaction strength and a mass that was of the order of the weak scale and the weak scale physics that we understood, that would very naturally provide a significant component of the matter of the universe as long as it was stable.

Yeah.

So you have to plug in by hand a mechanism that makes this exotic weakly interacting massive particle stable. You do that, you formulate a model. Now it turns out you can actually do that over a huge parameter space. You can vary the masses of these particles, you can vary the interaction strengths which depend themselves on the exchange of other particles, sort of so-called gauge particles or coupling particles. And that we have been, you know, over the last 40 years, we've been looking directly for the interaction of those dark matter particles with nuclei. It turns out that nuclei are of the similar order of mass to these WIMPs. To these things. So that means that you get a sort of— you get momentum exchange taking place.

Now, often when we're trying to understand what the likely rate is, we make various simplifying assumptions. And one of the things, you know, we're about to go on to is that it turns out that while simplifying assumptions of the nature of the interaction are very useful, it has rather limited the parameter space over which we have been specifically searching. For dark matter. And I mean, you know, theorists have been— were addressing this and, you know, we go back sort of 10, 15 years. They developed frameworks that actually said, look, don't just constrain yourself to this very limited, simple interactions. It's quite possible that dark matter interactions may be happening with nuclei which have a slight— Now, why you might be concerned about this, and I— let me just go straight into the calculation. It's really, if you're, you know, if you are a graduate student in physics or, you know, you're comfortable with keV, kiloelectron volt units, really when you're trying to understand what the sort of energy that could be transferred to a nucleus is, if you've paid attention in previous dark matter, many previous dark matter results and what have you, you'll realize that we often talk about tens of keV. A keV is a sort of energy associated with X-rays.

You know, so for instance, if you're being exposed to an X-ray in a hospital, that's a few hundred, you know, to 800 keV of energy. We've been looking for dark matter with sort of tens of keV energy, so down at the sort of bottom range of that. But if you actually calculate how much kinetic energy a dark matter particle in the Milky Way as it's traveling through you is carrying, you realize that because its velocity is order a few 10 to the minus 3, so a few thousandths of the velocity of the speed of light, then actually calculating half mv squared, if you're used to sort of a unit switch, you can just bring out the— or put in factors of c or c squared, and you realize that something with a mass of 100 GeV that's moving a few times 10 to the minus 3 the speed of light is actually carrying an energy that's as high as, say, 400 or around 400 keV. And the reason we usually don't talk about transferring as much as 400 keV is because under the most simple types of interactions, the so-called spin-independent interaction, you get a simultaneous scattering from all the nuclei simultaneously in a nucleus, which gives you a significant coherence effect because it turns out that the amplitude from each scattering, if the amount of momentum that's exchanged or the amount of energy that's being exchanged is relatively modest, you can maintain coherence across them. What's fascinating is if you do— if you sort of start messing around with the numbers, you realize that if you want to exchange more than a few tens of keV, actually you're going to lose coherence across the individual nucleons and you have to go to a much more— a more sophisticated, not terribly more sophisticated, but just a little bit more sophisticated calculation of the cross-section. And at that point, what's even more entertaining is if, as you play around with the potential parameters, it turns out you can actually get a situation where the dark matter wants to interact preferentially at higher recoil scatterings rather than the lowest one. It is heavily suppressed, but as you see, as you'll see in a moment, the LZ experiment is many, well, 7 tons of active volume. It's a large detector and we're well placed to look for relatively, well, relatively very weak—

Rick, if we could just go back, go back to the kinetic energy. So this is, I want to point out, this is like freshman physics. It's very cool. It's just like, if you want to see a detectable outcome of a collision, you don't take like a cannonball and shoot it into a ping pong ball. You really will see much more when they're roughly matched in mass range. And that, I think, is why you choose that. The thing that always kind of elides my discussions of it, because I'm not as much of an expert as you, when I teach cosmology and I talk about WIMPs and I talk about dark matter, It's always like, well, these things are scattering off of— so it's about, as you say, the WIMP is bouncing off the nucleus or vice versa, depending on your reference frame, right? But when we say bouncing, let's be precise. There's some gauge boson being exchanged, right? So all forces, including collisional forces, have to be mediated by some force-carrying mediator.

In most cases, an electron, electron scattering, Bhabha scattering, or whatever, it's a photon. What is being exchanged? If these things only interact weakly, it must be the W or Z, right?

That's correct, and it's because at the moment we don't have a specific model that we've identified. We have to be as broad as possible about considering the types of particles that can be exchanged. So as you'll see for this LZ event, You can— That to see it at the rate that we're seeing would lead you to estimate the sort of mass scale of the particle itself and then also, you know, a particular species of or type of gauge particle that's doing it. But in general terms, I think actually in this— so on this slide here, and this is one I— you remember I used in 2016. I've only— I've cut all the other previous slides that ran into this out. But the idea was that I was trying to give you a sense. In this case, we're just plotting the mass of the dark matter particle. But we are— what we then do is these specific models that are labeled here are often characterized by narrowing the gauge particle or the specific particle that's being exchanged in order for that.

And as you're— With the case of this LZ event, you can actually do it, for instance, with suggesting that the dark matter is— and you can see it actually on this plot— that, well, there's— it's a Higgsino-like particle but it's actually a doublet. And you can move between the 2 states as the scattering is taking place, and that itself introduces another twist in the allowable range of momentum or energy that's exchanged and will heavily suppress, in fact make it impossible for you to scatter depositing low energies. You have— there's a finite amount of energy that you must exchange in order to include the process which requires you to move move from the 1 Higgs, you know, so Higgsino, sorry, Higgsino state to the higher Higgsino state. There is in fact a subsequent decay associated with that. Unfortunately, it happens way outside of our detector, so it's not something we're in a position, you know, to study. So yeah, the way— and this is, you mentioned propagators. So one of the reasons— so we look at this plot. I usually say to a student, if they put up a plot that has what's this, 9 orders of magnitude on the vertical.

I'll usually say to them, look, that's ridiculous. Physics doesn't work. 9 orders of magnitude under most circumstances is a ridiculous amount of dynamic range to put in it. But this is a historic plot versus year of the sensitivity in terms of just normal— sorry, particle cross-section for dark matter on the particle— on the—

On the—

targets that were terrestri— the subterrestrial, the underground, you know, event searches that we've been doing. And we have actually covered, in terms of results, nearly 8 orders of magnitude over the last 40 years. Why is it 8 orders of magnitude? Because that propagator you mentioned, actually, if it's— if the particle that's being exchanged actually has mass, then in the limit of the low momentum exchanges, you're basically— you're often dealing with a 1 over mass to the 4th 4th term for that propagator. So one order of magnitude uncertainty or increase in propagator can give you 4 orders of magnitude in terms of a cross-section or interaction probability or rate. So it turns out nature can deliver models really without trying very hard that span this enormous range of potential interaction sensitivities or cross-sections. So So, you know, you don't automatically just say because you didn't find it in the first 10 years of searching that that's ruled out dark matter because this is a problem. In fact, the way I would say it is we had to flip it around. We had to try to make damn sure that for those of us who are involved in this kind of search, that the rate at which we improve the performance of the detectors, which is the sort of the size of the detectors and also their sense sensitivity and their ability to eliminate other potential background sources, that we have to keep improving that at a rate that is fast enough that we really can meet this challenge we've got here.

And, you know, one way I've put it in the past is that we are actually beating Moore's Law, you know, because whereas I think— let's see, Moore's Law is what, 3 and a bit orders of magnitude every— no, sorry, it's an order of magnitude every 10 years. I think we've been delivering an order of magnitude in sort of 2/3 that time.

Mm-hmm.

So, you know, 6 years or so. So we've been going faster than— now, you know, obviously I'm— you know, this is dark matter sensitivity. But just to give you a sense of how rapidly we've been able to— and we've done this through changing the type of technology that we use and just being absolutely laser-focused on improving our ability to test new models, which is something we've done.

Actually, yeah, if you go back a slide.

Yes.

I mean, I just wanna point out a couple things. Your density of fascinating information per second exceeds any cross-sectional flux that we'd get from any particle.

My apologies.

Except maybe the neutrino. But we're gonna have a podcast after this, and I will refer people to it. But I wanted to get this video out first. But I can't resist putting on my physicist podcast physicist hat And there's a couple things I would say. You know, I joined Brown University in the basement where you are now, I believe, and when that curve says 1993. So I was pretty early with Charlie Elbaum. And they were working on helium-4, proton—

Bob Lenoux.

Yeah, Lenoux.

Jim Scheidel, absolutely. Humphrey Maris.

Humphrey Maris, yeah, of course. How could I forget Humphrey? And I love those guys, but I went to Charlie Elbaum, the late, great— I love Charlie. He's one of the reasons I went to Brown. And he wanted me to work in his lab. And he said, why don't you go and talk to my student? I said, okay, great. So I went down to the dungeon, the basement there, you know, 144, below 144 Barrison-Holly. And I asked the graduate student and he was very enthusiastic and we spoke. And at the very end I said, oh, just one more thing, you know, just in terms of like, you know, career and so forth.

How long have you been a grad student? And he said, 9 and a half years. And I just couldn't believe it, Rick. And I felt like that took him, you know, to the '80s, you know, predating you in this field. And now we're in, you know, you're projecting out to 2040. I wanna push back with love and respect. And that is to say the following. Moore's Law is, you know, probably a slacker compared to the progress here. But another dimension, you know, it's created incredible technology that's, not only aided physicists and philosophers alike, but now it's taking over the world with transistors and software and computers and GPUs.

And we'll get into that in the podcast. I want people to watch our discussion about AI in the classroom and beyond with one of the world's great educators, Rick Gaetzkill. But Rick, tell me, I mean, how do we keep justifying it? I mean, the CMB world, I'm used to this. I've been looking for B-mode polarization, as you know, since the year 2000. And we haven't detected primordial B-mode polarization, but we've made a lot of progress. We've detected things. This, if this holds up, this is the first detection, right? This is a huge breakthrough. But on the other hand, it's sort of depressing as well because the length of the curve on the x-axis is so long.

How do you, how do you, you know, maintain through, through the rain, through the fog, etc.? We'll get into the neutrino fog, but, but how, how do you keep your spirits up, you know, and how do you maintain this? And what if, God forbid, you know, or I wanna say God forbid, I'll just say nature might not care about what you want as a student or as a postdoc or as a professor. So what happens if this is just another, you know, upper limit?

I know, I mean, the critical thing about when you're conducting scientific research is, I think, to make sure that it is well motivated and relevant. It's, of course, there have been notable examples where I think people, because they were total mavericks, went off in a direction and that actually yielded something. Of course, we also know that the mavericks we never talk about are the ones who went off, did their own thing, and completely busted. So, you know, you always have to— you've got to be painfully aware of this post facto selection effect that occurs, you know, when you're doing it. But I think, you know, when it comes to science, what we're trying to do as a community is to, by making presentations, by producing results, and by talking about, you know, discussing it, it's the question always is, is continuing a particular direction well motivated? Now, I, I think, you know, one of the, one of the areas, you know, saying, um, uh, you know, how many people does it actually take to do this type of work and how much resource does it take? In other words, I, I guess I, I know again you can say it's relative, but honestly, this, this type work is relatively cheap. It's— so it's not— but it does give you this incredible leverage in terms of sweeping through or being able to test simultaneously a very large number of potential models in— of particle dark matter.

And—

Yes. By staying sort of in touch with the rest of the community and looking at how the other results are going, I mean, one of the things is we both know, is extraordinary is the cold dark matter model, which, you know, for decades was sort of vying with a number of other potential models. It has survived and in fact has continued to make predictions and agree with new measurements at a level that I think few theories survive that long.

Yeah.

It is remarkable. And as a consequence, that sort of continues to steal one— make it so critical that if there is cold dark matter, and by that, by just cold, the cold means non-relativistic dark matter. If there is this dominant term that we call cold dark matter there, trying to figure out what it is is critical because that's the only way we're going to actually understand. And this is where I'd love to say we're going to understand the fundamental laws of physics, but of course it's only one damn universe. We we get to play with. So what we actually end up doing is— or getting a direct explanation of our universe, since it is quite clear that, you know, you can imagine situations where nature itself is giving birth to a larger number of universes, and we only are lucky enough as experimentalists to get to do one in one. But nonetheless, what we're hoping is that The— getting some of the properties of this cold dark matter, this dominant matter in the universe, will allow us to understand better how our universe is put together. Because at the moment, both the cold dark— the dark matter placeholder, if you like, and the dark energy placeholder, 95— that's 95% of the composition of the universe.

We do not know what the actual mechanism is what the actual composition is. And we've— we're trying to answer this with the direct detection strategy and there are many experiments doing this, trying to look in other specific parameter spaces. But at the same time, we're also trying to make dark matter in the Large Hadron Collider.

Uh-huh.

And we're also making astrophysical measurements looking for decay products from the dark matter. So we're trying to test this hypothesis. And honestly, 40 years— go back and look at history— 40 years is not a lot of time to be working on a problem. And I know you measured in human lifespan or what have you, you may say that's rather dramatic, Rick, but it's—

It's only 4 of Charlie's grad students' lifetimes.

Well, indeed. You know, that— and this is the way scientific research has to be conducted. If you— what you don't do is you don't stay on a question necessarily in a completely bloody-minded way because then that does lead to a sort of terrible sort of slowdown in rate of progress. You're always questioning, are we actually— given the other data that's coming in from other experiments, given what's happening in terms of our evolving understanding of how the universe is put together, is looking for particle dark matter still well motivated? And I would say right now, given everything else we're seeing. It's extremely well motivated. And the fact that we've tested 8 orders of magnitude of models, as I say, unfortunately, because of the way the physics works, you know, in terms of the parameter space that nature could have picked, you know, we don't get a sort of complete sweep. We're still testing right today, and we will continue to test models that are extremely well motivated. And in a sense, because they are a little bit more exotic, a little bit more removed, when we get the answer of what the particle is, of course, that's likely to really blow open a whole new— not new universe, but a whole new area of physics that right now is just one of a whole slew of possible models.

But when we know which particular model nature has chosen, then— And, you know, one does sincerely, uh, look forward to it then leading to answering a whole slew of other questions. Yeah, that's right.

Yeah.

So, um, so you mentioned being a graduate student. Now, I, I, I did— I was a graduate student in the UK where they used to cut your funding off after 3 years. So, so, I mean, literally they did bang, 3 years, and I was living on bread and water for about the last 14 So not 40, 4, sorry, last 4 months of my studentship. Oh, sorry, end of studentship. I was— but this was a dark matter detector, which is only 10 grams or so. That was the scale, you know, back in the— We have essentially the same sensitivity in 7-ton detector as we did in a— or you call it 10-ton detector as we did in 10 grams. So that's over a factor of a million. change.

And it's really down to extraordinary sort of ingenuity of so many colleagues trying to understand how you can measure individual electrons and individual photons, which is what we do in the xenon detectors now.

Right.

And we— and in this case, we are trying to use those individual measurements, those single photon and single electron measurements and combining them together to say that yes, we have evidence of some kind of interaction happening in the middle of our detector that really couldn't be put there by conventional backgrounds or conventional sort of radioactivity, but are instead due to the occasional interaction of a dark matter particle. So I should— actually, I'm going to skip this. So let me talk about SURF and our detector. So we're operating at an old— what was previously a gold mine. you know, up until just the turn of the millennium, you know, 2000— just at the beginning of the 2000s. And we repurposed a significant amount of the infrastructure for science here in the U.S., you know, and through the support of Denny Sanford and the state of South Dakota and Mike Rounds, who was then the governor, we've been able to— and all the subsequent support, and of course the Department of Energy, who've been absolutely critical to to, you know, making this laboratory, you know, function so well that we're operating this detector, the LUX-ZEPLIN detector, which for scale is about, you know, your height. It's about my height. You know, it's sort of one— just over 1.5 meters in terms of the active central volume of xenon.

It's about 7 tons of xenon. Xenon is— Xenon is actually the rarest of the gases in the air you're breathing in right now.

Mm-hmm.

Every 1 out of 10 million atoms that you're breathing in right now, which means a very large number of them are in fact xenon. And xenon has a number of sort of useful properties, but the main one that we're exploiting is in fact that it is very— can be made very pure and is very low in intrinsic radioactivity.

Yes.

It's also quite dense when it's cooled to about -100 degrees centigrade, so it becomes a liquid. It's about 3 grams per cubic centimeter, so 3 times that of water. And it can be purified so that light has a 10-meter-plus mean free path traveling through the xenon, and also electrons that are liberated in the middle of the xenon do not immediately combine. We can simply drift them using fields, electric fields, and pull them to the surface. And that makes for an ideal detector. The actual process where a particle interaction happens and scintillates, you scintillate. You know, everybody's fairly scintillating.

It's—

the trouble is you can't get the light or the ionization out of you because you're far too dirty and you're far too opaque.

Hey, speak for yourself.

Well, I'm definitely extremely opaque, as my students will tell you. But, but, so whereas the xenon has this fascinating condensed matter physics that a number of the noble elements have, where it emits light after excitation, it emits light at a wavelength that is actually— has a very low interaction probability with the material itself, so you can get the signal, the signal out. So that's what we exploit. We put photomultiplier tubes around the outside and we look and count individual photons. And we put a field on and we count individual electrons jumping out of the liquid surface. And the combination of that allows us to tell the position of where an interaction is and how much energy is deposited. And as you might imagine, in the middle of this very large amount of xenon, it is very difficult for regular particles to get in because their mean free paths are typically measured in a centimeters. So trying to get into something which is 150 centimeters across, all the conventional interactions tend to be confined around the edge.

Now, the only slight exception to that is there are one or two radioactive isotopes that will dissolve or can potentially— do, sorry, do dissolve in xenon. So, you know, one of the areas where our teams have spent a lot of time on is trying to make sure that as little of, for instance, the trace gas krypton, which can have or does have a radioactive isotope associated with, or radon, which is one I think everybody's familiar with because you get work done to survey radon in your basement. It turns out that the same radon daughters that you worry about with of excess radon in your basement are also capable of producing events in the middle of our detector. So we work very hard to drive— also to drive down the dissolved radon and krypton levels amongst others. I mention those 2 because those are the ones that end up being the most challenging to remove, partly because their chemistry, of course, is noble. Those are examples of other noble gases and therefore they—

They're inert.

They are more challenging chemically to remove than many other forms of contamination. So I mentioned I should pick up the speed a little bit, but just get to the event. But let me just show you. So this is just time measured in microseconds. So from here to here is around 1,000 microseconds or a millisecond. And this is just— this is the lowest— these are This example here is the lowest of all events that we see. And for conventional dark matter, this is often the regime in which you're going. And we actually see neutrinos.

This type of event I'm showing you here could be very—

You say lowest, you mean lowest energy or lowest—

Lowest energy. Thank you. Sorry. Absolutely right. Lowest energy. So we've actually, for instance, seen boron-8 neutrinos scattering, coming out of the sun, scattering in our detector, depositing very small amounts of energy. this S1, which is a primary light scintillation labeled. That is just a few photons that are caught by the photomultiplier tubes associated with the initial interaction.

And then we wait of the order of 800 microseconds and there's this sharp— this taller S2 event, which is again light, but it's light being generated from an electron. Electrons are fascinating. They— electrons, in xenon, when they drift through liquid, they scatter but non-radiatively. And critically, they don't recombine because we've made the xenon very pure. So there's a very low cross-section for being trapped, if you like, or scattered. The electron gets to the liquid surface, we have enough field to encourage the electron to jump out of the liquid. In gas, it is still drifting because of applied fields that we have, but now it's radiative. The scattering of the electron is now radiative, so the entire trace of the electron lights up, and we measure the photons associated with that track of the electron for the last sort of, you know, 20 millimeters or so that it propagates in the gas.

And this combination of an initial pulse of light and the subsequent pulse of light from the electron arriving at the liquid surface and entering the gas makes for an absolutely fantastic— we call it a time projection chamber, which sounds very exotic.

Yeah, it does.

It's just saying that the drift time of the electron and the location of where the electron hits the surface and lights up like a Christmas tree, that's the projection bit, that we can actually infer what the original XYZ or Z location of the interaction was. And you can do this in 7 tons. It's bloody, you know, it's quite fantastical. You know, we have, you know, the physical scale is 1.5 meters, both laterally diameter and drift. And we're able to do it at energy— we're able to measure energy depositions that are at the keV level and then counting, you know, in this sort of quantum world we live in, we are counting individual electrons, counting individual photons. And, you know, so the students, postdocs, you know, worked incredibly hard to basically characterize exactly how these signals are generated and how they depend on the nature of the original particle interaction. So I must pick up. So this slide on the top right, what it was just designed to show is along the bottom is that S1, the primary light.

On the vertical is the is the secondary light, the S2, due to the electron. So effectively, the ratio or the amount of S2 light versus S1 light, it turns out not only does it give you the position, the TPC part of it, but actually the amount of ionization you have relative to the primary scintillation light actually tells you about the nature of the original interaction. And this is enormously important because dark matter, we are looking for the most part, although we do have side searches. I think, but for the most part, we're looking for nuclear recoils from the dark matter coming in and interacting with the nuclear— You mentioned the business of the mass inequity, that the non-equal masses make— so an electron could, in principle, scatter from a dark matter, but you get so little that the masses are so different in many scenarios we're looking at that it's a very poor exchange. So primarily it's It's the dark matter scattering from the nucleus. And the ratio, the amount of S2 signal versus S1, which is the orange— we're using neutrons as a proxy for dark matter in this plot— that the ratio of the S2 over S1 is also telling us about the nature of the original interaction. So since most mundane backgrounds tend to be of an electron recoil type, which is characterized by the tritium betas on this particular plot, their ratio is some— of S2 over S1 is somewhat different from the ratio of S2 over S1 that you get for neutrons. And this holds, as you'll see, not just at the lowest energies, but even up to really very substantial energies, hundreds of keV, and that's of course where we've ended up looking for—

So just to summarize, Rick, the S1 signal, the nature of that is what exactly? It's coming from the dark matter, putative dark matter particle, or any particle that will interact weakly, and then it's it's from the recoil, right? It's a recoil. They're both recoil, and one is nuclear, one is electron.

That's right. So both of these processes, we're down at low energies, we're down at sort of 10 keV or less in this particular plot, which is— which for anybody who's worked with detectors will tell you is a very small amount of energy. And typically these recoils are happening over distances that are just measured in, you know, sort of— well, for the nuclear recoil, 10 nanometers, 50 nanometers. It all happens in this very short range. For the electrons at these sort of energies, you're still talking about submicron recall. So you're really measuring recalls in both cases in the nanometer scale. And as a consequence, it's— what's fascinating is that the— and to do dark matter, as my thesis had a lot of, for instance, condensed matter in it, I was trying to use superconductors at that time to detect dark matter. Here we have xenon.

But the condensed matter physics is— you get to study it in this exquisite detail in these energy regimes that often we're the first to sort of really go in and look.

Right.

And what we're seeing and we have to characterize it at highly detailed way. We have to characterize what would happen if 10 keV or say a 5 keV nuclear recoil or a 5 keV electron recoil was to go into xenon and deposit its energy along these very short tracks. How many ionization electrons are going to be liberated or how many internal excitons are forming in the xenon itself which then de-excite in a rather exotic way, in fact, forming these sort of double dimers that then emit photons. And all this is happening— I should— I didn't mention this before, but all this is happening in the VU— what's known as VUV, vacuum ultraviolet, about 175 nanometers. It's so named because those ultraviolet that is that short doesn't travel through air. You have to pull a vacuum in order to allow it. It turns out you can actually get 175-nanometer radiation or photons through xenon itself. Xenon doesn't want to interact with those even though it generated those photons.

It turns out it's very low probability of reabsorbing them, which is critical when you're building these massive detectors. You don't want them eating their own signal, basically.

So, yeah.

We, we, so Brown, like many other groups, we, we actually built the sort of, uh, these 2, uh, 250 PMT arrays that, that, you know, span about 1.5 meters. And this, this is going back some time now, but back in 2019, a whole slew of students, undergraduates, graduate students, uh, postdocs, all worked to assemble under very clean conditions. And then you— we shipped everything to the— to, to, to the Sanford Lab. And then by about 2021, so sort of coming out of COVID we were commissioning the detector. And what I'm talking about here is one of multiple data analyses we've done. We've done more conventional dark matter searches, but we've also looked very specifically in a dataset we took between sort of mid-'23 and early '24, which spans about 250 20 live days of data looking for dark matter events at, as you'll see, a higher energy regime than is usual for the searches. Now, it turns out that there are some challenges associated with doing that and that's why it's taken us a sort of couple of years to get to this point of actually saying we can estimate with some certainty, as you must, not only what the efficiency of seeing dark matter events are, which which that's, relatively speaking, relatively easy. The bit that you've got to quantify is what's the chance of more conventional interactions faking a dark matter signal in that area.

And that's something we've spent a lot of time not just trying to understand, but also you have to quantify it. Because the degree of certainty or the statistical significance you associate with a result is often based on what you are able to estimate is the— how unlikely is it that a background event could have faked your signal? And that's— people often talk about sigma confidence levels and that's— it's from that process or that likelihood of, is this a— is this dark matter signal versus likelihood of a more mundane— I don't know why I call them mundane because I can tell you the background has to be bloody exotic to fake this stuff, but it's an exotic more conventional particle somehow getting into your detector and interacting in a way that looks like a, you know, the dark matter signal. So we— again, you know, as you— I'm sure you would understand, we had to do an immense amount of internal analysis but also discussion and presenting and re-presenting and reassessing what we felt best categorized our understanding of the behavior of backgrounds in our detector to convince ourselves that if we see events, and we have, we saw one event, that is in the region where we expect dark matter, but what's the chance of it just being due to background? And I think again, I'm concerned about the time, so let me just It's okay.

I think we'll finish this today and then we'll go on. We'll have the podcast. I'm coming out to Brown. I'm on Stefan Alexander's Theoretical Physics Center. I'm on their board of directors. So, I'm supposed to come out there. So, we'll do an in-person review and we'll also do in the lab. So, take as much time as you want now.

I've got another 20 minutes if you do.

Okay.

Okay.

Well, right. So, What we're looking at right now is a conventional WIMP search from LZ. I'm picking the one in '24 because it— we have had subsequent results from there. But I guess the key thing I'm, you know, we need to look at is on the horizontal scale in the gray is energies, and this says 0.8 keV EE, 5 keV nuclear recoil. So that— what that is doing is telling you that for this particular search, that we are looking at extremely low energies, events that are extremely low energies, a few— well, less than 1 keV in electron equivalent around 5—

it's—

the reason you have to talk about nuclear recoil and electron recoil is because they— the amount of signal, the amount of S1 light and S2 light that is generated is somewhat different for a given particle of a given energy. But This was an early example. We actually pushed our threshold down to about, yeah, a couple of keV nuclear recoil in later search. But as I say, this particular plot is sort of, I think, easier for somebody to understand. So these events up here are conventional electron recoil events primarily being produced by residual amounts of radioactivity decaying in the middle of the detector and that's actually shown here. This is your radius of the detector and this is just the Z or the Z height, the drift thing. And you can see this sort of smattering of events. And this is taken over the same 220 days that we're going to talk about for the, you know, for the high-energy analysis.

But you can see these events are fairly uniformly spread out. What that's telling you is actually they're far less likely to be due to radiations coming in from the outside. These sort of gray regions actually where you see a much larger number of points, those are the typical radiation coming in from the inside that all gets stopped near the edge of the detector, but we have residual amounts of radioactivity dissolved. But because they're electron recoils, not nuclear recoils, they have a higher S2 over S1. It's actually this magenta— it's not magenta, purple. This purple region here is where we expect dark matter to appear. Now, you can actually see one or two events sort of getting into this region here. And in fact, that's because the separation between the electron recoil band up here and the nuclear recoil band down here is— while the separation is pretty good, still at about the sort of 1% or fraction of a percent level, events are sort of getting down into this region, and we have to— we take account of that.

So when we're looking for candidate nuclear recoil events in this region, which is where you expect the WIMP nuclear recoils to appear, we do have to say there's a competition between more conventional backgrounds just leaking down into this region and WIMPs. So in this particular result, there is nothing that stands out as an exceptional number of nuclear recoil events. nuclear recoil events that would be consistent with WIMPs. So what we did is we ended up just eliminating models that would have put more than a certain number of events into the— and as the joke goes, we were— we still are, I think, world-leading. Yeah, we are world-leading for seeing nothing in this sort of 9, and in fact down to about 5 GeV mass range because of the scale of LZ and because of the how long we've run it and because it's working very well, we're able to look for, you know, models in this low-energy regime and do so in a way that, you know, is very sensitive. But as I say, we're not seeing a buildup of low-energy nuclear recoil events. So what we did is we're using the same data to look for an interaction between the nucleus and you know, chi here, which represents the dark matter particle. And this is, if you like, sort of the conventional type of interaction where you get coherent scattering across all the nuclei.

But that really is the sort of vanilla version. You're summing up all these nuclei and you're assuming because it's very low energy interaction that the interaction itself is coherent. Meaning that the phase, for those people following quantum mechanics, when you're attempting to do scattering amplitudes, Of course, as you may remember, that your momentum exchange as you go in, interact with one particular nuclei, one point sort of exchange interaction, and you come out, there's a certain P associated with that. There's a certain— and that momentum exchange, if you like, has an amplitude and also a phase associated with that. And if you move location, if you go to different nuclei, If the amount of momentum you're exchanging is small and therefore the wavelength associated with that is, you know, and certainly principle, one way of picturing that in your head. But, you know, if the characteristic inverse of the momentum, which is a distance, if the distance is large, that means that the phases across all of those, all of that nuclei scattering are all very similar. So when you add them in order to get the overall, you know, amplitude squared, the sort of matrix element or the cross-section calculation, coherence. Phases are all very similar, they all add together.

But if you ramp up the attempted momentum exchange, that corresponds to decreasing the wavelength of the— associated with that wavelength, or what it means is that the phase is now varying much more rapidly across the size of the nucleus. the, you know, it's Fermi level, 10 to the minus 15 meter, you know, sort of Fermi level scales at these nuclei. And that change in phase means that when you start co-adding the terms, they are now not adding as cleanly together. They're actually starting to interfere with one another. And it produces a very rapid suppression, which is shown here, this is just interaction rate versus— this is actually recoil energy, not momentum, but it's— this is non-relativistic stuff, so you just say p squared over 2m effectively, where your m has to be the reduced mass. But effectively, you're just— this is just larger momentum as you go up in recoil energy. And what we see is while the rates are very large when you have full coherence for small momentum exchanges or small energy of recoils, you rapidly suppress as you get to 30, 40 keV. You've killed your signal and that's— it's not because the WIMPs carry enough energy to give the new— to give the xenon a kick.

It's that the process of attempting to exchange the momentum, that the Q squared as we say or the the momentum exchange, the wavelength associated with that is now getting smaller than the size of the WIMP— sorry, than the xenon nucleus, and it's becoming— the phase is changing rapidly across the nucleus and the terms are now interfering with one another and you've suppressed the interaction rate. But the reason is we've failed to, if you like, apply any imagination because we are taking what is sometimes referred to as the absolute vanilla of all possible interactions, which is just this scalar-like interaction where it's just, you know, WIMP in, WIMP out, point-like effectively interaction with a nucleon. You add them all up and you get a total sort of interaction strength. And we do sometimes talk about a slightly more exotic version of that, which is called spin-dependent, but it turns out even that is Although, introduction to gamma— sorry, we haven't got time to talk about gamma matrices. You have to develop— okay, so if you're— what we're going to do here is we're going to start trying to generalize the way in which the dark matter particle and the nucleon, these individual nucleons. And one of the ways you're going to do that within a relativistic theory is you want to start including the fact that these particles And for anybody who's slog— and it is fascinating. I shouldn't call it a slog. I always found it quite entertaining.

But if you're going to develop a framework to try to understand what the coupling strength is, you now have to have your particles represented by something that actually— the spin are in the jargon, but something that actually gives a little extra degree of freedom which is associated with the spin. And we're going to do that for the nucleon, and we're going to do that for the dark matter particle, and then you have to figure out the way in which the particle interaction strength between the two, how that is going to— Now, I guess what I should emphasize is, relatively speaking, we're not— we're still trying to look at this in a very general sense. And you're going to hear this phrase effective field theory. We're not trying to take this apart under a specific Bose, a specific gauge particle, a specific gauge. We're just trying to generalize and say, Let's, you know, what's a very simple interaction that includes the spin terms but doesn't have any other additional sort of complications? And the fact that we call this L15 may begin to tell you that actually we've just skipped over a whole load of other possible interactions and that's what we can do. And I think I should— let me put this slide up in this form. So it is perfectly natural when you're talking about particle interactions to say that this interaction strength the effective field theory that's describing them could include a momentum exchange term. And this is now the strength of the interaction.

So I'm not trying to do— I'm not thinking about how phase coherence is occurring across the nucleus.

Right.

What instead I'm doing now is saying that the point-like, you know, if you like, interaction between the dark matter particle and the nucleon, that it turns out it actually cares about what the value of the momentum is exchanged. And if we give that a q squared or a q to the 4th, which are both acceptable, they're Lorentz invariant, they don't violate— you can put it in without violating sort of any reasonable relativistic particle theory. But having put in terms like that, If it's proportional to what's going like q squared or q to the 4th, of course it's getting stronger as the momentum exchange is larger. So we're in a regime now where we've lost coherence across the entire nucleus, so that's suppressed it, but the actual interaction strength between the dark matter particle and one of the nucleons is actually being increased like q to the 4th or q squared or q to the 4th. And because of that— So you're saying that the interaction strength is actually increasing? That would mean that when we go to look for dark matter signature, if, you know, ultimately what you then have to explain is why your specific dark matter model would favor, you know, this L10 as shown on this graph, the L10 type interaction against an L1 interaction, which is much more simple, but you'd have to come up with a reason, a specific reason for why that's being suppressed. if the interaction was dependent primarily on L10 effective field interactions, that those would manifest themselves at higher Q squared or Q to the 4th, and that's higher recoil energies. So that's— and this, you know, this work was proposed both by my colleague Gigi Fan and Matt Reese, you know, back in, I think, 2010 or something like that, and also Wick I think Haxton and collaborators have studied this also in great detail, and I think they really laid out this sort of effective— all the effective interactions, Lagrangians in the language, of what you could have. But it is important to test all of those.

Is that to rule out, like, look elsewhere in other Well— Or is that on the theory side, calibrating the theory side?

It's because we don't know what the dark matter particle is, we don't actually know what the preferred Lagrangian or preferred interaction is. So while it is fair to say that an L1, or, you know, the basic scalar-style interaction would be natural because it can include coherence, which often makes it dominate. It is also fair to say that because our detectors now are, what is it, 8 orders of magnitude more sensitive than when we originally started, they are more than comfortable probing for these more exotic Lagrangians. Now, the fact that we haven't seen anything in the past 40 years means that nature somehow has decided to suppress the conventional spin-independent because we would have, you know, we might well have seen interactions much earlier on because of the significant coherence enhancement. But also what's now happening is because we have these massive detectors that with some perfect— with some reasonably natural assumptions about particle masses and the gauge particle masses, you could— and you could— we could well be in a situation where dark matter is choosing to not— or is actively suppressed Interacting through the sort of L1 Lagrangian, but is, say, favoring an L10 interaction. And these plots here— let's see, I should focus on the L10 here, but— sorry, that's L15 actually. Sorry, L10's here. But sorry, this is just the mass— sorry, this is the recoil energy range.

And I guess these plots are way too complicated for a talk like this, but one's putting them up just really to show how you don't just automatically get this major enhancement at low energies, which is the more conventional plot. What you actually see is a dominant preference for high-energy interact— higher-energy interactions. And as I say, because of the scale of the— Now there is a— and this is different. You can also talk about inelastic scattering where in this case the dark matter comes in 2 states that are just— are not quite degenerate. They're separated by some energy delta and that this does— this also is a mechanism whereby as a dark matter particle comes in, if its primary interaction is actually to be excited from one state to a nearby If the energy separation there is measured in hundreds of keV, it turns out that that would have a very direct effect on the rate of particle interactions that we'd see in these dark matter direct detection events. And there are, again, supersymmetry models, the one with a sort of Higgs-xeno doublet, with separations where the overall mass of the particles are of the order of 1, 1.1 TeV. So, sort of range, and that the separation is measured between the 2 Higgs Zeno masses is a few hundred keV. And for instance, that could be consistent with the type of higher energy preference.

And that's the doublet aspect of it. I feel like we buried the lead maybe a little bit. I mean, the mass of this is quite large, right?

Well, that is of course one of the things that, you know, for instance, all the LHC searching that we've been doing over the last What is it, 20 years now? You know, we've been sweeping up there, you know, in terms of production in mass. So in many cases what we've been doing is saying that as natural as we might have originally thought it was to produce supersymmetry down at lower mass scales, having searched for production of particles in there in many models that in order to, you know, to be consistent with the non-observation of many of the searches we've done, that would suggest higher mass particles. Now, it turns out, of course, that actually if you go as high as sort of 1.1 TeV in this sort of double Higgsino style model, that's actually very challenging reaching those kinds of center of mass— oh, sorry, reaching that kind of production particle. But so this is one that— this is a model that— model So you're saying that the Higgs is not the only particle that could be that produced? would be consistent with the direct detection thing, but is very much more challenging to actually see in— even with the high luminosity LHC. But this is only, you know, this is just one model that happens to be consistent with the one particle event we've seen. And of course, it's so early days that, you know, you don't— you know, you do this in order to make— help you begin to understand how this might work into various models. You're not saying that this is the definitive model. That would be the wrong way to ever look at a single event search.

So, yeah, I should— this particular plot, I was just trying to show you the contrast between at low energies where we usually look for dark matter and now— sorry, this orange and this purple thing, those are nuclear recoil sources, neutron sources that we put directly on the LHC. the detector. And what you're seeing is a band which interestingly, the red band is actually moving away from the electron recoil band which is shown here. This is an S1 versus S2 plot again. And so we have, you know, nominally what is very clear separation at these higher energies between nuclear recoil response and electron recoil response. And of course, from the point of view of a single event, that means that having a single nuclear recoil appear in the middle of a detector, which, you know, we're very confident is not due to a neutron because to get a neutron that deep into our detector, given the cross-sections for regular neutrons, is high enough to mean that it really can't get that deep in, you know, makes it so powerful when we are looking for, you know, events in the center of the detector. So, yeah, I should— I was— we designed LZ— what I was trying to do with these slides, but we designed LZ to be very good at doing this, to look for very occasional events and to make damn sure that this wasn't being produced by a more conventional detector. So for instance, we have an outer detector which is loaded with gadolinium and is scintillating, which is very specific.

Fine.

designed to catch neutrons either on the way into our detector or on the way out if they say they were generated by a piece of material inside the construction material of LZ. That gadolinium is very opaque to neutrons and lights and produces 8 MeV, a very high-energy burst of light or energy when that capture takes place. So we're able So we do see neutrons trying to get into our detector, but firstly, the absolute rate is extremely low, and secondly, we're able to very cleanly characterize them by using the multi-layers of our detector in order to sort of tag them. And this is something people have worked very hard on to make work so effectively, and is in fact— we use this to convince ourselves that we're seeing. So this is actually the science result. So this is the nuclear recoil band in the red and the blue is the electron recoil band and this is the S1 primary light signal and the S2 signal. And we're actually now up to around 250 keV nuclear recoil. So the original WIMP search was all happening down here.

We're now looking at a much broader energy range. But because we're saying potentially, you know, the Lagrangian or the operator, you know, the nature of the coupling could be allowing much more of the energy of the WIMP to be transferred to the nucleus. And this is the physics result after 220 days and we unblinded the data, although it turns out we're not claiming this as a blinded search just simply because post facto we went back and looked at the way we blinded and we— we're trying to be conservative. We're arguing that a really smart researcher in our group could have effectively probably told that— statistically figured out the difference between injected salt, as it's called, that we use for blinding, and possible event. And so, as such, we've decided not to call it a blind analysis, although I, you know, I'll emphasize that a lot of the cuts that we use were effectively fixed very early on in the analysis process in a relatively simple So we're— the sort of things that salting and blinding is designed to avoid, which is biases in terms of cuts, we've been able to, because of the way the detector works, really try to stay away from making any marginal cuts in the data. And so when we finally opened up and unblinded, you know, the data, although as I say, we're not calling it a blind analysis, we found we got left with one actual event. in the data, and it lies close to the nuclear recoil band, which is statistically where you expect nuclear recoils to occur, but it is way up here. There is an event down here that is actually expected.

That's due to accidental coincidences, and it turns out that at low energies, in many of our previous papers we discussed this, one of the backgrounds that we have to fight against at low energies is just accidental coincidences of a single S1 and a single S2 light. It turns out that for the higher energy region, that's not really— it's not a dominant background source. As we've gone through the analysis, it's the most significant contributor to backgrounds are actually high-energy gamma rays. that are potentially multiply scattering in the detector. And that's something we spent a lot of resource making sure that we could understand it, model it as well as we could, and produce some kind of statistical estimates— or not some kind of— produce well-developed statistical estimates of— And this is the weird thing. This is 220 days, but effectively with our simulations, we end up having to run for the equivalent of sort of 220,000 days. So what's that?

So can you translate the x and y axes? I mean, I'm sure they mean a lot to you, but to the audience, what—

Yeah, no, no, sorry. So this is our standard S1, the primary light plot. And this is the secondary light, or the light that comes from the ionization, the S2. So each one of these dots is a single event that occurred in the science data that we ran for 220 days, 220 live days. And down here, these events tend to be typically dominated by intrinsic electron recoil events occurring from beta, from low-energy contaminants. producing electron recoils. We also see evidence of specific gamma energies. Again, it turns out that when we're calibrating the detector for— that there are short-lived radioactive excitations that occur from the neutrons.

For this type of work, they— the energy deposition remains sort of well contained in this sort of S1, S2 plane. So we're comfortable that those types of events are not leaking or not producing events that are further down. But what we do in the paper spend quite some significant amount of time talking about is what happens if you have a high-energy gamma ray generated near the walls of your detector from residual contamination and that that gamma ray tries to get into your detector. Now, since that's going to scatter from electrons. Typically, that would produce events in the band here. But I think as I should show you here is— so that red line corresponds to a gamma ray trying to get in, doing a single interaction and then leaving. That would be a single interaction. Now, what if it scatters twice? What happens is first interaction here, second interaction here.

You get 2 lots of S1 light. The particle is so quick to propagate between those 2 vertices that effectively you can barely see any difference in the time of the S1. It's 10 nanoseconds when we're typically reading our S1 events over 100 nanoseconds or more. But the S2s, because the The delay time here between the S1 and S2 is related to how long it takes the electrons to drift and because they're actually only moving millimeters per microsecond, the electrons, that you actually get physical separation between the 2 S2s. But because you can see the 2 S2 signals arriving when the electron— small number of electrons actually reaches the top, you can clearly see 2— that there must have been 2 vertices. So nobody's going to— you're not going to confuse that with the dark matter.

No.

Particle. So, but, you know, only the paranoid survive in this game. So you have to start thinking about, yes, but what could possibly remove one of those S2 vertices? I— there's still 2 vertices, but something eats the, the, uh, these ionization signals. So that's what we call an MSSI, which is just simply multi-scatter single ionization. So what, what What removes? Well, you know, in some senses, we have a significant challenge because in order to apply a field between— to get electrons to drift upwards, that's effectively a positive field, if you like, pointing downwards. And we achieve that with a, you know, positive potential on the gate relative to a negative potential here down on the cathode. Now that cathode is actually running, you know, at say 100 kilovolts, minus 100 kilovolts. We've now got it, we've now got to get rid of that high voltage before we get down to the PMTs which are running much closer to the ground.

So we have what we call a reverse field region here and that means the electrons actually drift downwards and because they drift downwards you can't detect them. They just get lost.

So this is all, I like to say, you know, with Feynman's permission, you know, the first rule of physics is not to fool yourself, and the second rule is you're the easiest person to fool. So these are all ways that you're guarding against deceiving yourself, right, Rick?

That's exactly right. Now, you know, you might look at this and say, well, right, so how does this lead to the problem? And the problem is sort of shown here. Because you have 2 lots of S1, then you basically say that— say the first event is sort of here, and then the second event adds this amount of S1. light gets you to here. But under normal circumstances, the 2 lots of S2 here would also just boost you up and keep you, you know, essentially inside this sort of electron recoil region. Also, you'd see the fact that there are 2 bangs in the S2 and that would clearly tell you you had multi-site. But imagine one of them goes away, then what you're seeing is just 2 lots of S1 but only 1 lot of S2, which is getting you— brings you down into this region. So what you have to do is to make sure that you understand what the rate at which such events, multi-site single ionization, or, you know, an S2 loss, is going to occur.

But in order to understand this, you have to be simulating or thinking about events that are not occurring necessarily at the 1 in 220-day level, because we're actually trying to suppress this. We are thinking about whether these events would happen in a time period of sort of quarter of a million days, 220,000 days. Why? Because, you know, going with the Feynman theory, you know, you have to realize that even if something is, you know, if something's going to creep into your data, it can, because there are so many ways that things can possibly creep into your data, you have to be prepared to allow for the idea that some very rare mistake, mistaken identification might have occurred. So you have a process which, while being incredibly rare at the level of about 1 in 220,000 days, which is, you know, way beyond the amount of time we're running this detector, but it's the kind of timescale we have to simulate the detector, that one of these events randomly fluctuated, you know, happened to— bad luck— fluctuated into the detector. Detector. And it's doing that kind of work and understanding the sort of details of the response that it was necessary for us to do in order to make any kind of quantitative claim concerning what the likelihood, if you like, you know, of this event being due to a misidentification. And we— and it turns out that about the level of 1 in 200, so that's 1 in, what, 500 million— half a million days. But at the level of 1.5 million days, there is— we believe there is a rate of these MSSI, this misidentification of a thing.

It's, you know, 1 in 200 level of such an event sort of fluctuating in. I'm actually being sort of quite conservative because you mentioned earlier that there's a thing called look away— or sorry, look elsewhere effect.

Yeah.

So for you and I to be discussing this event You know, neither of us ahead of this result said we were going to see an event here. So actually what you have to say statistically is that an event could have cropped up over a much larger range of possible events. And you have to, when you're to make sense of the statistics of how significant an event is within that sort of signal, broad signal band, given that we have a large number of potential physics models that could generate dark matter recoils over that range.

Mm-hmm.

And because of the way that the background physics, the fluctuations occur, we actually, as, you know, people are familiar with the statistics, we started out a local significance that was in excess of 3 sigma, but by the time you include look elsewhere effects, and, you know, that sort of conservative couching that we have to do, that you end up with about a 1 in— sorry, 2.6 sigma, which I think is 1 in a couple hundred chance of you and I having a discussion over what is effectively a background event having fluctuated.

You quoted it as a range of confidence intervals, which is a little bit unusual. Can you explain that, Rick? Why is there a range between 2.6 and 3.4? What determines that range, and what would have to happen before the collaboration wave function collapses in regards to the effect.

So physicists, we're very— what's the word— very rigorous, very honest with ourselves. The only real language, as I always have to remind students, adjectives don't cut it. You have to associate quantitative numbers with things. So what you would do in this case is you've got one event. And what— so locally, if you like, locally in that energy region, what you would start off by doing is looking at every possible mechanism you can come up with that might deposit an event in that location you're seeing. And we looked at a lot of possible mechanisms, and most of them are— it's vanishingly small that they could, you know, they could in any way accidentally or randomly create an event in that region. The MSSI effect was the one that ended up sort of leading in terms of the probability. Still a very small probability.

So because that probability was at the sort of level of, I guess it's about 1 in 1,000 or something like that, that is associated with a 3.4 sigma. Talking about a sort of 3— Actually, it's less than 1,000. It's probably a few thousand, 1 in a few thousand. But you then, as I say, you have to go back and say, well, yes, but in order for this event to— an event to be interesting, it could have been occurring over quite a broad range of possible recoil energies. So, and then you end up forming a sort of global likelihood. And here is where it can get a little— not— I don't think subjective is the right word, but—

Subjective.

the word, but there are different ways of presenting such an analysis. And one of the reasons we published is because we're looking forward to getting input from people as to whether they feel our global significance is conducted in a way that they think is most natural for this. And there is— it is definitely one of those areas where there's no absolute, you know, this is the only way to do it, people. You have to, you know, you have to end up deciding exactly how you're going to statistically combine all these possible models for signal and all the possible contributions from the background, although that bit's a little easier. But it's still— that still gets— no, I should never say it's easy. It is a very exacting process. And I'm— you know, we spent a lot of time discussing this point.

Yeah.

Anyways, so what happens is that whereas it's a 3.4 sigma local effect in terms of how unusual it would be. In terms of us, you and I just sitting down and discussing, or indeed the experiment producing a result that had any event in this red nuclear recoil region, the chances of that happening but being, as it turns out, caused by a random fluctuation of the MSSI background into this band that global significance drops us to about, I think, 2.6 sigma, which is this 1 in 200 level. It's— Now, you know, the right thing to do is firstly, when we're talking about one event, is to keep emphasizing this could be a random fluctuation of a background event. It could also just be that we have misunderstood some aspect of the running of of our detector because we're trying to understand the detector at a level that is, you know, uh, well past the limit.

Wouldn't that have to be, I mean, how far into the process was this event? If you, once you, you know, unblinded it, you're then allowed to go back and see when did this event occur, right?

Yes. So I think, I think I have a, yeah. Yeah, there it is. So this slide actually, so the, this particular run, uh, was hot off the presses when we when, you know, we started doing analyses of this kind, you know, this was the data run from March '23 to April '24. But we were focused mainly on the low dark matter, low energy recoils, you know, as a sort of flagship analysis. But we also started working on higher energy regime. But a couple of challenges there. Firstly, we had to— do much higher calibration statistics and nuclear recoil statistics to really make sure we understood where the nuclear recoil band was.

Yeah, no doubt. For this high— and it turned out actually our initial estimates of where that was were slightly wrong, which is one of the reasons why the salting, the blinding— spot. They didn't— they weren't quite what we subsequently showed through higher statistics calibrations was going on. So 2 years elapsed, as it were, in the analysis and the deep— trying to really understand at a deep level what the chances of multiple scattering and other background were of producing, you know, fake events in this region. And we decided this year that that analysis had matured enough that we were indeed ready to unblind, which we did, and also to go ahead and publish, you know, the results from that unblinding. But of course, since April '24, we've been continuing to run the detector.

Right.

So we're actually in this interesting situation of now having, what, over 3 times as much exposure closure at this point. And, you know, so that, of course, in itself, if this is physics— and again, it's statistics— but if you say that one event every 220 live days is the rate at which we would like to accumulate physics, it means we have enough data statistically to go ahead and sort of really test whether we are accumulating physics.

That's true also— yeah, you're absolutely right. But that's also true of the background. And it's true of the experiment, right? So you should have accumulated, you know, if it's background, I mean, then the, there's obviously there's a lot of people that are critical of a single event, right? I mean, there's a thousand papers. Yeah. Yeah.

Yeah. We understand that.

We're not trying to say— I think, I think, you know, your, your H-index was very high, but I, I think it's gone up, you know, by a couple orders of magnitude even, or the citation count. I mean, that happened with BICEP too as well. I mean, I was getting emails from my former Brown professor, you know, Robert Brandenberger, you know, 4 days before the result, and they were already writing papers. But tell me, Rick, I mean, this is exciting in all different ways, right? You've already made the case, you know, in the worst possible case, right, for whatever— I don't even know what metric that refers to, because the truth is God's alone, or Mother Nature's if you're an atheist, right? But let's say this turns out not to be a dark matter candidate. Well, you've learned about the experiment, you've learned about the tool, you've learned about condensed matter physics, which, you know, you've helped me appreciate. My colleague Kaishuan Ni here, a very good friend of mine and been a guest on the podcast.

Yeah, no, I've worked with Kaishuan.

And yeah, and Elena April, a past guest on the podcast from the liquid, from Xenon-100 and whatever they're at now, 1,000. They also have, and I think this is a, you know, I want to chastise you in your field because you guys do so much for understanding practical nuclear physics and the nuclear condensed matter physics that nobody ever talks about and how productive and useful that is. So that's the bear case. That's the, you know, going back to your investment days, right? That's the bear case. The bull case is that either, you know, it's new physics or some new background, which would also be incredibly interesting. So I wanna, you know, 'cause we're coming up on almost 2 hours and I love talking. Yeah, yeah, yeah. And we'll talk more in public, but I know you gotta teach and I gotta teach.

But tell me, Rick, so where do we go from here? Are there other slides that you must show right now? 'Cause I have a lot of questions I wanna put to rest before we wrap up. So let's go on to— yeah.

Yeah, I mean, as we've talked about, obviously one interpretation is how can we adjust effective field theory parameters and say that this event, the reason that we're seeing it up at 250 keV and we're not seeing an accumulation of events at lower energies, you could play around with that idea and that suggests specific Lagrangian you have to say that whatever mechanism is causing the interaction must be actively suppressing low-energy events and— but, you know, benefiting— now, we're not suggesting from one event. This is where you need events plural. This is where you need many events because now, as you get an energy distribution, 2 things are going to happen. Firstly, are the events actually lying in the nuclear recoil band, which is a necessary condition for it to be WIMP. If those additional events are not lying in the nuclear recoil band, then that's a very good indication that we actually have a systematic background that we didn't understand, but that is coming in at a rate of the order of 1 every 220 days. Then the last hypothesis, of course, is that this was just a random fluctuation. And like, you know, plenty of experiments before, you every so often you get that unlucky. Why? Because we do so many experiments—

Yeah.

you have to get unlucky in terms of a background fluctuation, which is why, you know, we often talk about 5 sigma as being a necessary requirement if you're, if you're really going to say, you know, a signal is robust enough to, you know, to claim direct discovery.

That's right.

So yeah, so that's, that's really, you know, where we're going. And there are specific, you know, models, you know, inelastic Higgsino double doublet submodels. There are also ones where specific choice of Lagrangian that they, of course, have already produced people saying, well, have you looked in another place in your data? And, and it's been very exciting because in some cases people actually said, if you look in your appendix of your paper you've published, we've actually done an analysis already of your data and making the following assumptions, we can rule out or we could model which is lovely. I mean, obviously everybody understands that you haven't done the efficiency explicitly and you haven't done the thing, but it's still showing us, you know, very directly on how you can test a whole range of models using the data. I think all I would say is people have to bear with us because it is enormously exacting trying to estimate not just the response of the detector, but also the leakage, the signal leakage. In order for us to, as scientists, to be able to look at an event and say, how much weight should we put on this event, or we hope subsequent events, you have to have formed a very quantitative model of the background. So, you know, that's what we're doing. That's what we're doing.

So anyway, I guess I should say, you know, We're trying to make the quietest place in the universe inside our— or quietest known watched place in our universe. And I think we're doing a pretty damn good job of it. But obviously, you know, we always hope that it's not too quiet in the sense that ultimately one of these, one or more of these dark matter particles does actually start depositing energy in our—

If you want to find a quiet place, Rick, just find out where the string theorists are hanging out. Just kidding. I'm just kidding out there. I love string theorists. Some of my best friends are string theorists. I just, you know, I don't know if I'd want them to marry my daughter. Rick, this has been awesome. I just have a couple of questions because it's so rare we get to, you know, hang out and chat.

But, um, but I guess, you know, fundamentally, you know, the next, the next generation, if we, we're going to extend, you know, if you had made that, uh, I assume you made the Gordon Moore, you know, kind of law for dark matter detectors, you know, before this event occurred. But This might throw things in, throw a wrench into it, might make another knee on that plot to descend even farther faster. What are some of the ultimate limits? What's Mother Nature's veil that she, as Feynman said, refuses to let you pick up? Is it the neutrino background? Is it some other exotic phenomenon?

Well, so actually December of last year, we announced a result where we have accumulated around 20 20 neutrino events in the LZ detector, and those are events from the 8 boron component of the solar neutrinos. Those are neutrinos that are just energetic enough that when they coherently scatter off xenon, we can see the recoils. Now, that's a 15 MeV neutrino. That flux is, well, in our terms, quite high.

Yeah. 20 events.

So that means when we're doing very— when we're looking for signals at the lowest end of our signal now, we have to allow for the fact that the dark matter events, if the dark matter events are down at those low energies, they're going to be mixed in with neutrino, low-energy 8 boron neutrino events. Now, as you go to higher energies, actually solar neutrinos, you know, they cut off at 15 MeV, so we're not going to do it. But, you know, atmospheric cosmic neutrinos. So that's, you know, cosmic ray interactions producing neutrinos in the atmosphere. Those energies go up much higher, those neutrinos. And it's not going to occur in LZ. But if we were to, you know, go, as is quite natural, we build not a 10-ton detector, but a 100-ton order detector, and we run it for 10 years or something like that. What starts happening is, is we do expect to start seeing nuclear recoil signal, and that comes about from atmospheric neutrinos.

It actually also somewhat amazingly is we could start seeing the occasional event also from neutrinos from the diffuse supernova background, which is—

Right.

That's basically all the supernovas across the entire universe over time going off bang, and it turns out that the GeV, those GeV neutrinos could also be depositing energy through scattering off nuclei in a xenon detector. So that is a sort of fog in the sense that you start seeing individual events occurring and now you have to say that that could be associated with the neutrinos, it could be associated with dark matter, and statistically what we have to do is determine the neutrino signals as well as we possibly can so that we can see any anomalous or increase in the rate above that which would indicate dark matter. So that's what, that's what we mean by entering the fog, is we're now statistically having to allow for, you know, 2 possible hypotheses for them. So that's, it's, that's some way away from this kind of work we're doing at the moment. We have gotta do a lot more pushing to get into the fog.

Another question I have, I can't resist. I ask it whenever I talk to your friends Alina and Kaishuan and anyone else who's working in this field. But, but what are your thoughts about the current state of DAMA? Um, I did a video about them last year in connection to, you know, the, the most persistent signal that's believed by nobody that has the best, uh— well, it's not believed by nobody, but, um, and I'm gonna ask, it's not just to dish dirt, but, but, uh, tell me, Rick, first, because I think, you know, I'm going to ask you a follow-up which is basically what I've wanted them to do for a long time, which is to build a DAMA Southern Hemisphere. But tell me, Rick, what are your thoughts as one of the world's foremost experts on experimental physics in this field? What are your opinions about DAMA?

Well, so '90s, I was around in '97 when they came out with the first annual modulation. Yeah, 29 years ago. it— obviously, we were enormously excited.

Yeah.

It's the mechanism that dark matter uses to modulate in that way, you know, is a very, you know, very natural, well understood. And it implied that there would be a very high rate of dark matter in not just the DAMA experiment but actually a number of the other experiments we were running at the same time. So it really motivated, you know, us looking for corresponding events in similar detectors. Now, as the decade went on, so, you know, through the noughties, if you like, we managed to build these other detectors that were more and more sensitive. And unfortunately, we did just did not see any individual events. So DAMA was seeing, if you like, a statistical process, which was this few percent modulation of a large number of events which— where the hypothesis was that a small— sorry, that some fraction of those events were conventional background and some other fraction were dark matter and the modulation was in the dark matter component and the observation of modulation over the year, higher in June, slightly a few percent lower in December and up, down, up, down like that, that that was evidence for the dark matter. But we were running other experiments that on a— literally on a single event by a single event basis could tell the difference between a nuclear recoil and more conventional background. And we just weren't seeing any of the— any events in our other detectors that would be consistent with the dark matter.

Now, there were ways to modify the theories which actually includes inelastic dark matter. In fact, that was a time when inelastic dark matter became interesting because that was potentially a way to explain the difference in results. Although one of the things we did with our xenon detector in the mid-noughties, 2006, around then, is we made a very sensitive, just 10-kilogram detector. But the fact that it did not see events from dark matter because the xenon's heavier than the iodine in the DAMA experiment, you could no longer use this inelastic get-out clause to sort of effectively explain why why nobody else was seeing the DAMA events. So just from a pure scientific process point of view, the fact that we've been unable to replicate the DAMA results either in sodium iodide experiments, which have now reached a level of sensitivity that are comparable, you know, to DAMA. And DAMA was an exquisitely designed experiment.

Yeah.

So it took a long time for other people to replicate everything that they got right in the DAMA experiment. But also at the same time we had all these other competing— well, not competing, but, you know, other dark matter, direct detection dark matter experiments that were not seeing events. That's a problem for a— for one result if you can't seem to get confirmation with other experiments. So, you know, the net result is it's a very clear modulation. It's gone on for— they've stopped the experiment now, although I think they're rebuilding it with an independent group.

Yeah, LIBRA, aren't they? Is that right?

Yeah, well, actually, LIBRA was part of DAMA and they went on. There's now— I've forgot what they're calling it, but they're basically taking the same setup and they're going to do it again to try to see whether they can see how this— how the annual modulation might be getting in somehow to the experiment.

Yeah, the final thing I wanted to ask you about, you know, because how often do I get a chance to talk to not only a friend, a valued and treasured colleague, but a professor at my alma mater, involves the work of a former— well, I should say, involves the work of someone who's at a university that's also my alma mater, namely my undergraduate alma mater, Case Western Reserve University. And that's where I was an undergraduate before I came to Brown in '93. I cannot believe it's been 33 years since I started there. But nevertheless, it is. and you weren't there, unfortunately. I would have loved to have you as a professor, colleague, friend, you know, uncle in the lab, so to speak. You're not that much older than me. But tell me, Rick, the work that Stacey McGaugh and others have done on MOND.

You mentioned—

Yes, yeah.

You mentioned work by my former colleague, Kim Greist. I mean, he's my emeritus colleague. He's still a friend. Kim Greist here on MACHOs, and that was really big here, you know, in the '90s and 2000s, but that's largely gone away. But recently, a paper, you know, by my friend Alessandro Melchiorri, I just made a video about on my channel this week called Why the Dip? And that has to do with Gaia data from 3 years ago, also around the same time as your event, Rick, by the way. Kind of interesting. Gaia came out with a map of the Milky Way's stars and velocities and kinematics, and they show that there's actually a semi-Keplerian drop, which is highly unexpected, both in the dark matter paradigm that you and many others are more or less committed to in one way or another. But it also rules out MOND because MOND was invented to do what dark matter does, but without requiring those fanciful Higgs-ino doublets or whatever you may have detected, or that one.

I mean, it reminds me of the Valentine's Day event. What do you feel? What is your take on MOND?

Well, so Stuart, firstly, cold dark matter needed You know, if you're gonna really develop your understanding of dark matter and cold dark matter specifically, you needed competing theories. So of course you will remember when we used to also be very serious about hot dark matter, relativistic particles, and it was enormously important to just compare those 2 models, both sort of dark matter models. But— and it really drove us to understand rates of structure formation and the way in which data and obviously increasingly precise data that we've managed to get in terms of, you know, the structure, evolution, history of our, of our, you know, universe. And the MOND in a different sense was enormously important. Again, you know, there have to be competing theories in order for you to understand any model. It's always extremely instructive to have other competing models. So it would be, you know, in a sense, it's not good if there is only the one theory. Now, of course, cold dark matter has been so damn successful.

It is you have to be quite brave coming up with theory. MOND has done very well, except that you always have to remember that MOND was kind of structured to solve a specific problem, which, as you know, is to do with the individual galaxy rotation curves. So when students come up to me and say, Rick, you do realize that, you know, rotation curves in galaxies is a solved issue, it's like, hang on, you've got to be a little careful about your logic there, because realize that MOND was designed to solve solved that issue. So let's try to look at how MOND fares in other regimes. And that, of course, is, you know, that's how any good theory gets tested. And cold dark matter has done phenomenally well over a huge number of length scales and obviously time, you know, evolution. And, you know, MOND has had some challenges in that respect. You know, and then if we end up having to say, well, okay, you know, it's MOND plus dark matter.

You know, in a sense—

Yeah.

Of course, nature has proven that nature is quite willing to give us mixed models. But from a, you know, from a theory point of view, obviously we'd like to, if we can, define whatever the dominant solution is for dark matter or the dominant solution is for dark energy. And we have to hope a little bit that nature decides to make one of the models dominantly the solution is. If that's not the case, if it turns out nature's used 5 separate, you know, components all to contribute—

You know, I always say, you know, if you said I'm looking for matter, ordinary matter, then, you know, I'd say, well, what kind? There's 116 of them, you know, on the periodic table. But, you know, most people don't agree with me. And lastly, you know, I often often say, you know, people use dark matter as a canard, as a polemic against physicists, saying, you know, you also alluded to it in a way of being humble against hubris, but they'll say things not as charitable. They'll say, you physicists are idiots, you don't know what 95% of the world, the universe is made of, and yet you claim to be able to tell us all sorts of things about, you know, the future and so forth. I like to point out that dark matter has already been detected, and you guys, even if this particles correct, it won't be the first, because we've known since 1956 that neutrinos exist, and we've known since 2008 or before, 2001 maybe, that neutrinos have mass. So they fit every definition of dark matter. They're massive, they're weakly interacting in this case, and they don't interact with light. So people that claim that there must be some alternative like MOND, I feel like that's the strongest objection.

In other words, Yes, it is true, it doesn't make up the complete density of the universe, the missing density in terms of matter, non-baryonic matter, but neither does xenon itself, or neither does iron. Iron doesn't make up much of the universe as well. Is it important to us? Absolutely. If the only thing we knew about was iron, we'd be, you know, we'd have some knowledge, but we couldn't surely say that there has to be some other alternative form of matter or whatnot, right? So why do people reject the dark matter WIMP paradigm. So I'm thinking of Sabine Hossenfelder, who's a friend and been on the show many— she thinks this is all a big waste. I mean, she's hopeful. She said, you guys are in a gold mine hoping to get gold in the form of a Nobel Prize. And I hope that you do if you're right.

But the point is, people mock dark matter, but we know dark matter exists. And I see this from Elon Musk. I see this from everybody that it's all a scam. And we haven't done anything in physics since before string theory. So where do you take us? Take us out on this final question. The existence of the neutrino, doesn't that truly substantiate that dark matter is particulate in any case? May not be only from neutrinos, but doesn't it give us a big boost?

My, you know, those early dark matter detectors were indeed testing the hypothesis that Dirac neutrinos, the kind of conventional sort of neutrinos you're talking about, if they were the dominant mass in the universe and if they were massive enough to be cold, or a Dirac neutrino was massive enough to be cold, we were able to directly test that hypothesis. And we, you know, that was enormously instructive that we managed to rule out, you know, that there were massive Dirac neutrinos making up the dominant part of the dark matter. I mean, it's always interesting. Again, it comes back to this sort of— you work on a question for 40 years. For some people, just the fact that you had to work on it for 5 makes it a pointless— they immediately assume that—

Instant gratification, right?

And it's, gosh, you're really not— you're failing to understand how science works. Now, the fact that we are in a society or a structure that that is prepared to, you know, back us and to allow us to do work over these long periods. I mean, if, you know, if I'd been an ancient astronomer, you know, and coming up with dark matter would have been the difference between whatever my ruler was, the difference between them maintaining power or getting usurped because clearly they're not—

Not a brown, not a brown.

Well, I mean, astronomers always have this sort of leg up with respect to eclipses. But of course, if you got one eclipse wrong, you know, that probably terminated your career in a very unfortunate manner.

Maybe your life.

Yeah, well, no, indeed, that's what I'm getting at. So, you know, we— but, you know, in the lifetime of a researcher or the, you know, all this, all the attention span, if you like, of a researcher, some research, that's not the useful metric here. You really have to look at the problems on this larger scale. And also keep looking at all of the other data that we're getting to see how consistent the hypotheses are with what we're getting. You know, if we manage to construct an experiment that demonstrates signal or a propensity to develop or to generate dark matter through some other mechanism, and we do have— we're doing tests of that type in many other different channels. If we start seeing a significant sort of positive indicator there, then obviously that would suggest certainly that looking for a WIMP particle, you know, is less well motivated because it's going to be a sub-dominant or very small component.

It's—

trying to rule it out completely is always going to be, you know, a very significant challenge.

Would that mean that the nightmare scenario is that dark matter exists, it's similar to inflation right? Inflation could have happened, but could be so undetectable, so low in energy, that we can never prove it or rule out and falsify alternatives. Is that the nightmare scenario for you, that the neutrino fog, you know, will be the ultimate limit? And are there any other, you know, proposals to, you know, clean that background? I thought it was hard to clean the galaxy of its polarized dust B-mode emission, but scrubbing the universe of neutrinos seems impossible by comparison. So— Well, That's your nightmare scenario.

I mean, the neutrinos and the dark matter, of course, have 2 very different signatures when you start considering the typical direction in which they're hitting your detector. So it is fair to say that if you want to associate a signal of nuclear recoils with a galactically interesting astrophysical source, the fact that the sun is in motion, you know, 230 kilometers a second around the Milky Way and therefore that the Earth is being carried with it. That, you know, that Cygnus, you point towards Cygnus, that's where the sun's heading. That does skew the recoils you get from dark matter away from Cygnus. So that is a signature. Now, in xenon, we've tried to look for possible ways of getting directionality in liquid xenon. That hasn't happened, but other researchers have demonstrated that using alternative targets, usually with gas, although there has been work on some solid, you know, stilbene for instance, and— but that it will be possible to see the recoil direction. And of course, if you're able to do an experiment that's measuring the recoil direction of your— The recoil direction of the beam.

your nuclei, then the neutrino fog actually is no longer a fog because it's statistically, it's, you know, isotropic or, you know, it doesn't have that correlated singleness.

That's right.

So there are ways through it. Of course, it requires a lot of research to make a— I often talk about Sisyphean index, technology that wants to work. If an experiment or a particular technology wants to work, I call that a low Sisyphean index.

Right.

If it's going to roll down the hill on you every time you turn your back or you try to go to sleep one night and the technology comes back down the hill and you've got to push it up again, high Sisyphean index, not so good. So we're always, when we're doing dark matter experiments, we're always rather dependent on finding low Sisyphean index materials. Xenon, I think, has a very low Sisyphean index. It really wants to work. If we're going to build a 100-ton detector made of gas, we're filling cathedrals. We want to make sure that we're doing it with a technology that really wants to work. It's quite doable on the scheme of systems that we've built in the past.

Well, Rick, this has been fascinating. I want to cut it off before we hit the 2-hour mark because this is just too exciting, too delicious. And hopefully by the time I'm visiting you again, I will be able to do it in person, and then we'll have a little bit more clarity what it is, what it was, and we'll refer back to this, you know, watershed epochal event. So, Rick Gaetzkill, you know, tremendous, tremendous amount of gratitude. You've been unbelievably kind and generous with your time during this incredibly busy time for you traveling around faster than a neutrino. Rick, this has been great. Thank you so much for communicating with me and my audience. Thank you.

Brian, thank you so much. It's been a great pleasure talking with you again.

Thank you, my friend.