The astrophysics of binary black holes after a decade of gravitational-wave.... - Sylvia Biscoveanu

追加: 2026-05-14

[00:00:15]Okay. And now Sylvia will tell us about the astrophysics of finding black holes.

[00:00:20]Hey Z. So thanks to the other organizers for inviting me. So I'm going to change gears a little bit um and tell you about the astrophysics that we can learn about following a decade of gravitational wave observations.

[00:00:34]So we've now observed gravitational waves across many decades in frequency.

[00:00:39]So here you can see the sensitivity of key gravitational wave observatories and facilities as a function of frequency.

[00:00:46]So starting all the way on the left hand side of this plot we have the nanohertz gravitational wave band where we now have evidence for a stochcastic background from we think the slow inspiral of super massive black hole binaries. And this was detected using a host of different pulsar timing arrays including um and the European um pulsar timing array and the Australian timing array, the Indian pulsar timing array.

[00:01:15]Um and going up in frequency we have Lisa in vanilla frequency band um which is a planned future space-based gravitational wave detector where the primary sources will be the mergers of these super massive black hole binaries but also um a whole zoo of galactic binaries including um combat object binaries with white dwarf components that we can't access with other bands.

[00:01:42]And this talk is going to be focused at highest frequencies shown on this plot which is in the audio band from about tens to thousands of hertz. Um where we have groundbased gravitational wave detectors like um the LIGO, Virgo and Cog detectors. Um and as we've already heard the primary sources of interest in this part of the band are mergers of stellar mass binary black holes.

[00:02:07]So how do these compact object binaries actually form? Well, there are two main theoretical formation scenarios that we think explain the stellar mass binary black hole mergers that we've seen um with the LIGO coctors. So, one possibility is the isolated binary evolution scenario or the field formation scenario. And in this formation channel, the compact objects are basically lifelong companions. So, they start out in a stellar binary. Um they're born together. They evolve um throughout many many stages their whole lives until both of the compact objects um end up as black holes in a tight enough binary that we can observe it merging within the age of the the universe. And there are many different astrophysical processes at play here.

[00:02:56]For example, we have tidal effects in the cellar progenitors and each of the objects in the binary needs to independently undergo a supernova um that leaves behind a black hole and that also keeps the binary bound rather than disrupted or a binary that results in a stellar merger. Now the other possibility is that these compact objects are more like social butterflies and this occurs in the dynamical formation scenario um which is mostly active in very dense stellar environments like globular clusters or nuclear star clusters where you have many compact objects and stars zipping around and they're constantly dynamically interacting with each other until some of them end up in binaries that are on tight enough orbits that they can merge within the age of the universe.

[00:03:44]And so these two formation scenarios, as you might imagine, predict very different um imprints on the properties of the binary black holes that we can observe with gravitational waves. So to highlight how we might learn about the formation channels of these systems using our gravitational wave observations, I'll walk you through the different signatures in terms of the spins of the black holes. So starting out for the isolated binary evolution scenario in terms of the spin magnitudes um we expect that the firstborn black hole in each binary is going to have small spin because of efficient angular momentum transport in the solar progenitor of this object. Basically what this means is that all the angular momentum is stored in the outer layers of the star. As the star evolves, those outer layers get blown off, leaving behind a slowly spinning core that collapses to a slowly spinning black hole. However, the second born black hole in an isolated binary can potentially be spun up due to tidal effects. So, this would occur at the stage where we have one evolved star and firstborn black hole. And at that point tides will work to synchronize the spin of that evolved star with the orbit which can spin up this object um which can then collapse to a rapidly spinning black hole and this is called tidal spin up.

[00:05:08]Now in terms of the spin angles um for binaries that form via isolated binary evolution we generally expect spins to be aligned to the orbital angular momentum and this is again due to tides.

[00:05:21]Basically, tidies are going to work to realign the spins of the stars in the binary with the orbital angular momentum after the individual supernovi that can kick the orbital plane.

[00:05:34]And then in addition to this tilt angle with respect to the orbital angular momentum, there is potentially another interesting angle that we can use to learn about um binary information evolution. And this is this aimethal angle between the inplane projections of the spin vectors of the two component black holes. We call this 512.

[00:05:55]And in some cases for binaries that form um in this field formation scenario, it's possible that these systems will be caught close to certain resonant configurations where 512 will liate around specific values of aligned or anti-aligned. of y12 equals 0 or plus or minus pi rather than freely processing and this occurs really preferentially in these um isolated binaries. Now on the other hand for the dynamic information scenario um we expect that the spin directions will be isotropically distributed because these um black holes go through many many dynamical interactions that scramles the spin directions and there's no sort of preferred direction in the cluster and so you end up with this random distribution of spins. Now for the spin magnitudes um again because of the efficient angular momentum transport in the cell progenitors we expect low spin magnitudes except for in hierarchical mergers. So this occurs when the remnant of a previous merger is retained in the cluster and it actually pairs up and merges again with another black hole.

[00:07:06]And so in this case for um the merger remnant of two slowly spinning black holes we expect that it would have a spin of around.7. And so if we observe this um second generation object merging again, we would infer that it has a large spin.

[00:07:23]So I'm sure this will be familiar to many of you, but I just wanted to provide a little bit of context of how far we've come in terms of our observations of gravitational waves over the last 10 years. So our very first gravitational wave uh direct detection came over 10 years ago with GW5914 which was actually before the official start of the first observing run of the LIGO detectors. And we take this event for granted now because we've observed hundreds of other events like it. Um but I just wanted to take a little bit of time and really um impress upon you the significance of this first detection and the astrophysical impact that it had. Um so before GW514 um we didn't know whether or not binary black hole systems that merged within the age of the universe actually existed. Um and so the discovery of this event actually told us that there is some um binary astrophysics process that facilitates the um bringing together of black holes in binaries in tight enough orbits that we can observe their merger.

[00:08:31]Um this event was also unusual compared to the other stellar mass black holes that we knew about electromagnetically.

[00:08:38]Um and you can see them here in terms of their mass. Um and most of them are in binaries. So also their orbital period.

[00:08:46]Um and you'll notice that most of these electromagnetically detected black holes have masses less than about 20 solar masses. Well, GW50914 famously has masses of around 30 solar masses. And so this was already much more massive than the other stellar mass black holes that we knew about. And so this also indicates some potentially new astrophysical process that leads to the formation of such massive black holes.

[00:09:10]And then finally for this event, we didn't measure the spins particularly well. Um but the constraints that we did get indicated that it was likely very slowly spinning. And this is again in contrast to um the vast majority of these electromagnetically detected black holes which were consistent with potentially being even maximally spinning. And so already from this first event, we knew that binary black hole mergers exist and that potentially there are some different astrophysical properties at play in these black holes that we observe with gravitational waves compared to those that we already knew about electric.

[00:09:46]Okay. So, a few years after our first direct detection, um we had our first binary neutron star merger, which I won't talk about at all after this, uh slide given that we are focused on black holes. Um but this was a particularly exciting event because it was so far our first and only multi- messenger event um that was accompanied by a host of electromagnetic counterparts in addition to the gravitational wave emission. Um, and then a few years after that, we rounded out our portfolio of different classes of combat object mergers with the first detection of two mergers between one black hole and one star. And I should say that this plot just shows the cumulative number of detections as a function of observing time. Um so right now we have public data available through the end of the first part of the fourth live observing run and you can see that this includes 218 confident gravitational wave events. Now we are working very hard on analyzing the data from the rest of the fourth observing run. So the data from this middle chunk 04B should become public um in 2 weeks on May 26th. Um and then the O4C data um will be available at the end of this year. Um, but I do want to emphasize that you can see that there is a notable increase in the detection rate in ' 04 compared to 03. And this was facilitated by dramatic improvements to the detectors. And we now have close to 400 candidate gravitational wave events, which means that we're really entering the um population statistics era where we can begin to learn about the astrophysics not just from exciting individual events, but also the population as a whole.

[00:11:37]So let's take a look at the properties of these um roughly 150 uh confident gravitational wave events through the last catalog. We've already seen a version of this slide from Alison's talk, but it wouldn't be an LBK talk without masses in the study here as well. Um and so already looking at the masses of all these events, we can begin to notice some interesting trends. So um we can see that the gravitational wave detected black holes which are shown in blue tend to be systematically more massive than the electromagnetically detected uh black holes which are shown in red um including some really unexpectedly massive events um that we've already heard a little bit about today. And then looking at the low mass end of this plot we can see that there are some mystery objects that are colored um both orange and blue. This is because we don't really know whether they are a black hole or a neutron star.

[00:12:34]And so let's take a look at some of these particularly interesting events in a little bit more detail. Starting with one of these mystery objects. So GW230529 um was actually the first gravitational wave event that we detected at the start of the fourth observing run. Um and it is most likely the merger between a neutron star and a black hole. Um but if this is its actual interpretation, this is the most symmetric neutron star black hole merger event that we've ever observed. Um consisting of a approximately 1.4 solar mass neutron star merging with a 3.6 solar mass black hole. And the existence of this object is potentially interesting and surprising because based on uh the electromagnetic observations of neutron stars and black holes in this low mass um part of the compact object mass distribution, there was evidence for a gap between the most massive neutron stars and least massive black holes. And we called this the lower mass gap. Um well this object lies staunchly in that lower mass gap between around 3 to five solar masses. And so the existence of this 3.6 6 solar mass black hole that we've observed in gravitational waves tells us that one these kinds of objects can exist and that there's some supernova process that leads to their formation and potentially there is some um astrophysical selection effect that disfavors the presence of these kinds of objects in systems where they can be detected electromagnetically.

[00:14:07]Next we have GW23123 which we already heard about a little this morning. Um but this event is remarkable because it because it is likely both the most massive and the most rapidly spinning binary black hole merger. Um here you can see the posteriors on the component masses and the component spins um with different waveform models and we've already heard about the um waveform systematics for this event. Um but astrophysically this event is a little bit of a puzzle um because it is actually so massive that we think that black holes with this mass should not be able to form from direct stellar collapse. Um and if it was some kind of hierarchal merger, it's actually so rapidly spinning that it's not easy to explain this particular set of parameters from a hierarchical merger scenario. And so this event doesn't have a very clean astrophysical interpretation. Um maybe it could have formed in an AGN disc or requires some other sort of interpretation beyond the kind of two standard possibilities.

[00:15:14]Then we also heard a little bit about this pair of events GW241011 and GW241110 um the palend drones. Um, and what I want to highlight about these events is they're very interesting spin measurements. So, um, they're both consistent with having primary spins of around.7, which is, um, this smoking bone signature of hierarchal formation.

[00:15:38]Um, and this, um, first event from October has a spin that is slightly um, misaligned but um, less than 90 degrees with respect to the oral angular momentum. And then the second event from November has a spin that is um nearly antilim to the orbital angular momentum.

[00:15:58]And so these spin misalignments are also an indication that these two systems could have formed in a dynamical environment via a hierarchical merger like the globular cluster. Now you might be thinking you know these hierarchical merger events should be preferentially more massive because they include uh the remnant of a previous merger and these events have quite low masses um in the teens. Um and it turns out that the existence of these low mass hierarchal mergers is consistent with um predictions from high metallicity globular cluster simulations. So those um where the environment in the cluster has around solar metality. So this is not necessarily an anomaly even though we think these are hierarchical mergers that have quite low mounts.

[00:16:48]Then we also heard a bit about GW250114 which is the twin of our first detection nearly 10 years later and we can draw a nice parallel between the improvements made um between the Hubble deep field and the JWST deep field. Um, we saw our original friend GW5914 with an SNR of about 25. And then 10 years later, we have GW uh 25114 with an SNR of around 80. They have remarkably similar masses.

[00:17:21]Again, both approximately 30 solar mass and equal masses. Um, but as we heard about in a lot more detail this morning, the highr of this event allows for some really exciting precision tests of general relativity. Okay. And next up, I'm going to show you uh some so far unreleased uh events. So, they're coming out in five hours. So, I will ask you not to take any photos or tweet about this for the next five hours. Um but we have our last two um exciting individual events from 04. So between the data that we get at the photo diode of the detector and the strain that we release to the public there's a calibration step in between and normally we can perform this calibration of the data using INC2 measurements of the state of the detector. So we have a a variety of different ways that we can calibrate the detector um using measurements that we can make within the detector itself. And this calibration is typically parameterized in terms of some amplitude and phase deviations um as a function of frequency. Well um in addition to measuring this calibration using the detector itself, it turns out that if you trust GR, you can actually measure this calibration using the astrophysical signal. um because we can compare the data in our detector that includes any potential miscalibration to the prediction from general relativity. And so far we've done this astrophysical calibration um measurement for every single event that we've ever analyzed, but it's not been informative um until these two events. So this is GW24925 and GW25207.

[00:19:12]And you can see that for the event in September, we were able to measure the detector calibration both using the NC2 measurement shown in blue, which has some uncertainty, but it's generally quite precise. Um, and using this astrophysical calibration method shown in pink and the two measurements agree with each other, which means that we're doing something right. um which was a good validation because for the second event um it occurred at a time when the Hanford detector was still stabilizing and we were not able to get a confident measurement of the NC2 detector calibration which would have meant that we would have to throw away the Hanford data entirely and not analyze it because it was poorly calibrated. Um however we were able to use this astrophysical calibration method to infer the calibration state of the detector at that time and we get an informative measurement. Now you might be wondering you know why would I want to be able to measure the calibration state of the detector that's not really telling me anything about astrophysics. Um well as I mentioned if we don't know the calibration state of the detector then we have to throw away that data. Um, and one area that could potentially be affected is the sky localization of the events at these time periods of poor calibration. And so you can see here for these two events, the initial um, sky map that was estimated when we throw away the poorly calibrated data from Hanford in green. Um and you can see that these are uh an order of magnitude larger than the updated uh sky maps that we get when we incorporate this astrophysical calibration measurement shown um in the pink um with two different assumptions and um in the blue here. So this is just one example of the extra science that we can access using this astrophysical calibration measurement. Um, and I'll also just add that actually the um, parameterized tests of GR for this um, second event from February actually um, provide the tightest upper limits on the 2pn and higher deviations um, across all events including better than 25114. And so that measurement would not be possible without this astrophysical calibration.

[00:21:31]And this is because actually both of these events are quite high SNR. So this event had an SNR of around 32 and the February event had an SNR of around 70.

[00:21:42]Okay. So now that we've completed our survey of interesting individual events, um we want to go from individual event inference to population level inference.

[00:21:51]So here is just a quick schematic of how this process works. You can see here the posteriors on the primary and secondary spins of all of the events in our last catalog GWGC4.

[00:22:07]And what we want to uh determine is what is the overall population level spin distribution. So here in the dashed white lines we have some different possibilities. Maybe it's uniform across all spins. Maybe it peaks at small spins. Maybe it peaks somewhere in the middle. And just by looking at the individual event posteriors, you can't really tell if there's a particular trend because they're quite poorly constrained individually. And so this hierarchical basian inference method allows us to one account for the statistical uncertainty in our individual event measurements. And then it also allows us to account for selection biases. So because we have a deterministic uh prediction for what the signal from these binary black hole mergers should look like, we can quantify exactly how sensitive our detector network is to a signal with a given set of parameters. And we can account for the fact that some signals are easier to detect than others. For example, those that are more massive or closer to us. And so the results that I'm about to show you for the population um as a whole represent the distributions of the true underlying astrophysical um population rather than the population that we observe which is um a biased uh representation of this underlying astrophysical population.

[00:23:27]Okay. So we can start off by looking at the black hole masses. And this is our um population level distribution on the mass of the more massive black hole in each binary. Um the different colors show uh two different population models.

[00:23:43]In green is one that makes fewer assumptions about the shape of this population model and um really just allows the data to tell us what it holds. Um whereas the blue model folds in some more astrophysically motivated assumptions for what the shape of this distribution might be. Um and we can see some interesting features. We see that there are two peaks. One a global peak at around 10 solar masses, another one at around 35 solar masses. Um and also the evolution of this distribution is different at low masses compared to high masses. And so we can begin to look for certain signatures of these various astrophysical processes that we talked about. So looking at the high mass end um we see that there are um high mass black holes with masses greater than about 50 solar masses which we think cannot form from direct cell collapse because of a particular kind of supernova called a parent instability supernova where stars that are initially between about 150 and 250 solar masses um they just get completely disrupted leaving behind no black hole remnant and so we would expect there are no black holes um with masses greater than about 50 solar masses that form from direct cellar collapse. And so the existence of black holes in this mass range potentially hints to us that these objects are forming hierarchically or through some other channel potentially in an Asian disc. Now we have this feature at around 35 solar masses. Um and this feature is of ambiguous astrophysical origin. There's no single astrophysical process that predicts a clear peak at 35 solar masses. It's not really expected um due to a pileup of black holes um at this particular mass from this instability supernova process. Um and so potentially this uh feature is due to a confluence of different formation channels being active at different masses.

[00:25:46]Okay. Now let's take a look at the spins. So here are our population level measurements of the black hole spin magnitude. Again using two different population models and we can see that in general black hole spin magnitudes are small. We have a lot more support at um small spin magnitudes compared to large spin magnitudes. And this is consistent with this picture of efficient angular momentum transfer in the massive solar progenitors of these black holes.

[00:26:14]However, um the spin distribution that we infer also seems to indicate that the spins are not negligibly small across the whole population. And um we can approach this question from um one particular method that I developed a few years ago called spin sorting where usually we tend to sort the components of the binary um into the primary and the secondary based on which one is more or less massive. Instead, we can sort the components of the binary based on their spin. So, we call Kai A the black hole that is more rapidly spinning and Kai B the black hole that is more slowly spinning. And here you can see the um population level constraints on these distributions of these spin sorted parameters from GWTC4.

[00:27:00]And you'll notice that Kai A is consistent with peaking away from zero at around point 4 while Kai B is consistent with peaking at zero. And so in recent work with a Princeton undergraduate Lily Shenai, we showed using simulations that this measurement of the population level distributions of Kai A and Kai B indicates that at most 80% of sources are non-spinning using the spin sorting method. And so this is perhaps indicative of an astrophysical process like tidal spin-up where in isolated um binary evolution you end up with uh spin preferentially on one of the two black holes in each object.

[00:27:43]Okay. Now what about the spin tilts?

[00:27:46]Here you can see our population level constraints on the coine of the tilt angle where a cosine of one means the black hole spins are aligned to the orbital linear momentum. zero means they lie in the orbital plane and minus one means they are anti-aligned. Um and we find that there is significant support for some fraction of the population having significantly misaligned spin tilts. And again as a reminder it's difficult to get such large um misalignments in the isolated binary evolution scenario. And so this tells us that some binaries are likely forming dynamically. Now you might also notice this feature at around um cos theta of zero. So in plane spins um and in recent work led by Jacob Steman we showed that this potential excess of systems with inplane spin tilts could be explained um by a significant fraction of the population actually forming in stellar trickles. And so these would be mergers that are facilitated by coiloff oscillations where um we have an inner binary that merges um due to the presence of an outer um object on a wider orbit. And in this case this channel sort of uniquely predicts that there would be an excess of spins um in the lower plane. And so potentially we have some binaries that are forming dynamically with very large misaligned tilts and some that are forming in solar triples that contribute to this excess space in the plane.

[00:29:21]Now so far all the distributions that we've looked at have assumed that the mass and spin parameters of these binary black holes are uncorrelated on a population level. while in reality we expect that these different formation channels should encode correlations between these parameters. And so um in a model that I developed a few years ago we wanted to look explicitly for this correlation. And so to look for this correlation instead of modeling the component spin magnitudes and tilts we look at this K effective parameter the effect of line spin which is the mass weighted spin projected onto the orbital angular momentum. So this is a lossy parameter. For example, there are multiple ways to get a KI effective zero. Either the black holes are non- spinning or they both have spins directly um pointing into the orbital plane, but it is actually the spin parameter that we measure best with gravitational waves. And so we can get more information about it on the population level. And so in this correlated model, we allow the mean and the width of the K effective distribution to evolve with red shift.

[00:30:25]And here you can see um the high effective distribution that we inferred using our previous um catalog at different slices in red shift starting at the smallest red shift and this narrowest distribution in green going out all the way to the largest red shift of about 1.8 in the um widest brown uh band there. And so the fact that this distribution broadens as you go to larger and larger red shifts tells us that black holes were likely born with larger spin in the cosmic bounds. Now when you update uh the inference here including the latest events from 04 um the evidence for this broadening which had been previously found at around 98% credibility um using just the events from 03 increases to greater than 99% credibility in 04. Um, so we are actually becoming more confident in this correlation as we included more data.

[00:31:21]Now you might be thinking, well maybe the evidence for this correlation is just driven by the particular model for the correlation that you built into your population analysis. And so in work led by MIT Brown student Jack Heisel, we wanted to check for exactly this. So um we changed our correlation model from this more prescriptive linear model for the correlation to a more flexible spline model. And you can see here our results. So this is the evolution of the width of this KI effective distribution as a function of red shift again using data from the last LBK catalog. Um and the linear results are shown in orange and the more flexible spline results are shown in blue. And so you can see that even with this more flexible model um that really lets the data speak for itself and we still find evidence for some kind of evolution. So you can't draw a flat line through the blue either.

[00:32:17]Okay. And then the last thing that I want to say about spins is we alluded to the astrophysical information that we can potentially gain from this as a mutal angle between the two um in projections of the black hole spins by one two. And so in recent work I um calculated the population level distribution for this parameter using um the GWTC3. So the previous catalog um unfortunately if we don't measure the spin tilts very well we measure the asth spin angles even worse for individual events. Um but we as I mentioned are now getting to the size of population where we can potentially begin to tease out features on the population level. And so we see hints of excesses in the 512 distribution at 512= plus or minus pi.

[00:33:05]Um and this is consistent with some potential contribution from these binaries that are in these special um nearly resonant configurations where instead of the asamucle really circulating and processing it is locked um around these particular values. Um and this is a signature of the mass transfer history of the binary in the isolated evolution scenario. So the fact that we find excesses at plus and minus pi rather than at zero actually tells us that more of the black holes that um end up as the more massive black holes have started out as being the more massive star rather than the other way around which we would call mass ratio reversal.

[00:33:45]So this is obviously a very weak measurement. Um but as our catalog doubles hopefully we can begin to get better population level constraints on this parameter as well to learn about mass transfer in these biners.

[00:33:59]Then finally we can look at the red shift distribution. So here you can see the population level um distribution on the binary black hole merger red shift um again using two different models. And so the interesting feature here is that when you compare it to um how quickly we expect the um cosmic star formation history to evolve which is shown in white with arbitrary normalization. Um it seems that our binary black hole um red distribution actually evolves more quickly which means that it would peak at um smaller red shifts compared to the overall star formation history. And so there are two competing effects here.

[00:34:38]one is the actual um red shift dependence of the star formation history from which these black holes are born and then the other component is the delay time between the black hole formation and the eventual merger. And so the fact that this might peak at smaller red shifts is potentially indicative that the either the stellar progenitors of these black holes which are tend to be more massive are um being generated from a star formation history that also peaks at smaller um red ships or that there's some long delay time between when these um stellar objects are born and when we actually see the black holes. merch. So, you might notice that this plot cuts off at around a red shift of 1.5. That's where 99% of our detectable black holes are um detected.

[00:35:35]And so, again, hopefully as we um add more events and as our detectors improve, we'll be able to probe this out to larger red shifts.

[00:35:43]Now we might even be able to learn about the most distant binary black hole mergers um before we detect them individually using the stochcastic background. So in the band of groundbased gravitational wave detectors we expect that there should be a stcastic background of all the unresolved very quiet very distant um stellar mass compact object mergers that we can't individually detect. Um and they should overlap to produce this background. It's not actually a Gaussian background because um the duration of these signals in the LIGO band is much shorter than the time between them. And so it's actually a pan or a popcorn background. Um, and so in recent work with an undergraduate student at Northwestern, Nico Bears, we wanted to see if we look for this non-gaussian astrophysical binary black hole background, can we use it to learn about um the population properties of high red shift binary black holes. And so what this amounts to is a population analysis without thresholds. Usually all the previous plots that I've shown you have been obtained using a high confidence subset of all of the individual wave events that we detect that we are actually confident are real events. And so in this method instead we analyze all the data agnostic as to whether or not there actually is an event in the particular search of data that we're analyzing and then we probabilistically combine all the data um accounting for our uncertainty. um for whether or not an event is failing. And so in this sense, there's no distinction between foreground and background. And so using a simulation um we find that we can actually potentially constrain the shape of the binary black hole merger rate um which I'm plotting here out to red shifts beyond the peak of star formation using our detectors at current sensitivity when we search for this um unresolved population. And you can see our constraints using the combination of the foreground and the background in orange compared to what we would get just using the foreground in blue. And so you can see that when we include these unresolved mergers, we actually can learn something about the shape of the red shift distribution at large red shifts through around three.

[00:38:03]Okay. And then I just wanted to highlight that in hopefully in about 10 to 15 years uh we will have no astrophysical background at all. Um because we will have some uh new next generation detectors that will be so sensitive that we'll be able to detect every single stellar mass binary black hole merger of stellar origin in the universe. Um so these are uh cosmic explorer um which is the concept being developed out of the US and Einstein telescope which is being um developed in Europe which would instead have a triangular design. So cosmic explorer would be like LIGO L-shaped um but the arms would be 10 times longer so 40 km while one current proposal for Einstein telescope is this triangle with 10 km long arms that would be built underground. So to give you a sense of the observing power of these detectors um you can see here the reach of various detectors um for binary black holes shown on the right in white and binary neutron stars on the left in yellow. And so you can see that at the end of the last observing run 03, we were detecting a good number of binary black holes and basically just scratching the surface of the binary neutron star population. Even with the most sensitive that we could potentially make a gravitational wave detector in the current LIGO facilities, which is this orange line, um we're barely um detecting, you know, the beginning of the binary neutron star population. Whereas with these next generation detectors shown in pink and green um we can detect all the binary black holes and the majority of the binary neutron stars. Um so this would be a gamecher. All right. So just to summarize in terms of what we've learned um so far and what we hope to learn in the future. Um for supernova physics we've seen that there's no evidence for a lower mass gap between neutron stars and black holes detected in gravitational waves. And this tells us something about the supernova engine that leads to the formation of these compact objects. And we also saw that the primary mass distribution extends into the upper mass gap both for a few individual events that we've observed with masses in this range and for the whole population. And so this tells us something either about um the formation channel through which those very massive binaries are forming or that we got something wrong about this um pulsational parent stability supernova process. Then in terms of the information channels, we think we are seeing contributions from both dynamical assembly and isolated binary and potentially triple evolution um based on the spin tilt distribution and with uh the detection of the stochastic background in the band of groundbased detectors. This will reveal the population of high red shift binary black holes before we can detect a single one of these mergers individually. And this would allow us to determine whether those high red shift sources have distinct properties from the ones that we are observing in the local universe. Um and then finally um with our next generation detectors we'll be answer and we'll be able to answer all of these questions and more. And I'm happy to take any questions. Thanks.

[00:41:17]>> Thank you Sylvia. Are there any questions?

[00:41:21]I have two and I made a mental note of the slides. One is slide 21 when you were talking about spin sorting.

[00:41:28]>> Yes.

[00:41:29]>> I just want to make sure I understand what you said. I think you said that yeah at most 80% of sources are non spinning. What do you mean by sources?

[00:41:37]Do you mean the progenitor spins?

[00:41:40]>> No, sorry. I just mean of the binary buck holes that we observe.

[00:41:45]>> But so do you mean the pre merger?

[00:41:48]>> Yes. Yeah. Yeah. The components pre merger. Yes.

[00:41:52]>> And and the second question is about the new event slide 18.

[00:41:56]>> So when I look at the um sky map on the right, >> yes, >> I'm supposed to be comparing the green uncalibrated with the blue or red, I guess. No.

[00:42:13]>> Yes.

[00:42:13]>> And and so those sky locations are completely different.

[00:42:17]>> Yes.

[00:42:17]>> Can you explain what's going on?

[00:42:19]>> Yeah. So actually here we are plotting only the 90% area. If I were to have plotted the 99% area it would extend over into this region. It just so happens that for this event um we happened to you know be in a weird part of the sky where the you know two detector sky ring and the three detector sky ring overlap a small probability.

[00:42:43]Um, and so this is kind of just, you know, it the fact that these don't overlap shouldn't happen often. Um, but it's not anomalous that they do. It just happened to be in a weird part of the sky. Yeah.

[00:43:02]>> The other side.

[00:43:04]>> Okay.

[00:43:05]>> So when you say at most 80% spinning, so I I miss at least two numbers to understand the statement. So you mean at most 80% have spin less than something and then the overall statement has some significance.

[00:43:18]>> So at most 80% of the binary black holes in the underlying population have two component objects that have both exactly zero spin.

[00:43:33]Uh this was what we compared against for the simulation. Right? So there have been other um methods used to probe this question where you can define some very narrow spike of spins around zero or you can compute base factors for spins that are you know take on any value versus those are that are exactly zero for our simulations we used exactly zero I don't think there would be any practical difference between a black hole that has a spin of 0.05 or 0.01 01 compared to a spin of zero, but we had to pick something for the purposes of the simulation. Um, and this is a qualitative statement. So, um, in the paper, you know, you can look and see the exact method that we used to do this, but we basically compared against, um, different simulated populations. And so, uh, I actually don't have error bars on this number because it's kind of a handwavy statement, but it was, um, just to demonstrate that rather than, you know, recomputing all of the parameter estimation for all of the events using a non- spinning hypothesis, you actually get a very similar constraint on this overall non-spinning fraction using this, you know, more qualitative method that's a lot cheaper.

[00:44:48]>> When you showed the black hole sources. There are very few between Voyager and then the VG detectors. Is there a reason?

[00:45:00]>> Uh yeah, because there are no stars. Red shift 10 um that will collapse that will have already collapsed to form black holes. And so this is just for one cosmic star formation history. It's possible that with some variations you might get a few more, but in general we don't expect stellar mass, binary black holes of stellar origin at red shift greater than 10. And if there are black holes of stellar mass at redist greater than 10, people generally say that's a smoking gun of primordial black hole formation.

[00:45:32]>> Okay, then no further questions. So there will be afternoon tea in Rubenstein now and then there will be dinner at 6:30 [snorts] in Simon's Hall. And with that, let's thank Sylvia once more.