Supermassive Black Hole Caught Red Handed in Stellar Homicide

all right good afternoon welcome to this week's CFA colloquium and it's a pleasure to introduce our speaker Suvi gazzarri Suvi did her master's degree got her master's degree at UCLA from UCLA and then moved to Columbia where she got a PhD in 2005 she then moved on to Caltech as a postdoc and spent a couple of years at Caltech and then was a Hubble fellow at Johns Hopkins before she moved to University of Maryland in 2012 as an assistant professor during her time at both Caltech and Johns Hopkins and now at Maryland Suvi has worked on a wide range of topics in time domain astronomy using primarily Galax which she will tell us some things that she will talk about today but over the last few years has also been a very important member of the Penn star's key project on variables and transients where she's done a lot of work with our group here at at the CFA so without further ado I'll turn things over to Silvia and she's gonna tell us today about yes so whenever you have a press release you need something dramatic and exciting and so this was our press release tagline and it did get a lot of attention supermassive black hole caught red-handed in a stellar homicide in fact I was interviewed by the Baltimore local news and when we were prepping for the interview I said you know could we do a transition like speaking of homicide but we did not he did not like that idea too sensitive a topic so yes it's wonderful to be here at Harvard CFA because of course I've been working with a lot of you some of you have already moved on from Harvard now but there are many people here that have been active members of this pan stars science consortium and so I have many collaborators who were are now or formerly at Harvard and of course I've done been doing a lot of work with the galaxy space mission so in my talk I'm going to tell you all the exciting things that I think we can potentially learn from tidal disruption events I'm going to show you the excellent progress we've made in the past decade in particular on detecting candidates across the electromagnetic spectrum in particular I'm going to focus on how we've searched under the lamppost by taking advantage of the wide field and high cadence of the Galax time domain survey in the UV and the pan-starrs medium deep survey in the optical to search for these events and then I'm going to focus on one excellent example in which we had enough evidence from this crime scene otherwise known as observations where we could determine the victim which in this case is a star and the perpetrator which is a hungry black hole and then I'll conclude with what really is the promising future for detecting large numbers of these tidal disruption events but some comments on how we can prepare ourselves and make this process more successful so usually when we talk about tidal disruption events we like to start with this schematic diagram from Reese's nature paper in 1988 where he shows you what was realized by theorists to be in an inevitable consequence of a supermassive black hole lurking in the center of a galaxy and that is that if a star approaches close enough namely within this tidal disruption radius the tidal forces of the black hole will overcome the self gravity of the star rip it apart and some fraction of the stellar gas will remain bound to the black hole and be available to accrete likely producing a luminous flare of radiation and then some fraction of the debris being ejected at high velocities so this was our picture in the late in 1980s however thanks to an rico ramirez ah sorry enrico thanks to an rico ramirez ruiz and his former student james willis Shawn who's now a fellow here at Harvard and unfortunately is not in the audience to defend himself they have really stepped up the quality of simulating exactly how the material in a tidal disruption event rains down on the black hole as a function of time and these are now detailed tighter dynamical simulations but interestingly enough even when you do all this beautiful detailed physics we get a pretty simple picture about what happens when you tidally disrupted our and basically as a consequence of the capillary and fall back of the material to the black hole as a function of time you get this simple characteristic T to the minus 5/3 power law decay another really interesting thing about tidal disruption events is that their timing is very sensitive to the mass of the central black hole and of course as astronomers we love weighing black holes especially potentially in distant galaxies and so this is a really unique probe of black holes in particular dormant black holes which are very difficult to study so in fact the rise time to peak or the time at which the fallback rate Peaks since disruption scales the black hole mass to the one-half power so as you increase the black hole mass the timescale increases as well as the peak mass accretion rate decreases but interestingly enough I also show here the Eddington limits corresponding to those mass accretion rates and you also expect different accretion behavior depending on the mass of the black hole namely that large black holes should be mostly sub Eddington accretion but the smaller black holes may in fact go through a large portion of their decay during a super Eddington regime so this is an interesting insight on accretion physics okay so what can we learn from TDs this is now my sophisticated diagram of a tidal disruption event and of course I just argued that we can potentially weigh black holes in the Centers of galaxies based on this characteristic timing of the fall back of material onto the black hole there's a neat consequence of the fact that the tidal disruption radius scales of the black hole mass to the 1/3 power whereas the Schwarzschild radius scales the black hole mass linearly such that as you get to bigger black hole masses you get to a point where the Schwarzschild radius actually becomes larger than your title disruption of radius so for above 10 to the 8th solar masses solar-type stars are actually swallowed whole before being disrupted however due to I we know from general relativity that if you have a maximally spinning black hole you can shrink your event horizon by quite you can shrink your event horizon and therefore actually increase the maximum mass that can disrupt a star so if you saw a tidal disruption event around a greater than 10 to the 8 solar mass black hole and you think it was a solar type start you could pretty clearly say that the black hole must have been spinning okay we can also probe the types of stars that are getting swallowed by these black holes and that's again because the fallback rate of material and in particular the tidal disruption radius at which a star is disrupted depends on the mass and radius of the star and it turns out that it also depends the fallback rate also depends on the internal structure of the star so there actually can be an imprint of the radial structure of the star on this fallback rate as a function of time so we can potentially measure or probe the mass and radii radial profiles and maybe even the chemical composition of these tidally disrupted stars what's both interesting and problematic is that in order to successfully model and interpret these events we have to understand the accretion physics going on and in particular what's exciting is the physics of a newly formed accretion disk so a lot of what Enrico's work Enrico's group is doing is trying to take these beautiful simon simulations which are showing you how this material is starting to build up around the black hole and maybe taking the next step in trying to figure out how does that material actually accrete and radiate a recent really exciting discovery from the Swift mission was that in fact not only do title disruption events produce these newly formed accretion disks from the stellar debris but as we know in astrophysics accretion and Jets often come hand-in-hand so it's not surprising that we now have a new class of tidal disruption events in which we're not seeing a flare of accretion onto the from on to the black hole but we're actually seeing a newly formed jet on access in the line of sight so tidal disruption of events might tell us some clues in general what it takes to form Jets in super in accreting supermassive black hole systems in addition of course these black holes are swallowing stars we have to get the stars into the tidal disruption radius around the black hole and that's due to scattering processes in the nucleus of the galaxy and so it's very sensitive to the orbital distribution of the stars in the center and of course as well as the the stellar population so the rates of tidal disruption events give us a window on the nuclear stellar populations and galaxies in addition to stars around black holes in the center of a galaxy there's also going to be gas and so we can potentially probe the circum nuclear environments of a black hole that was previously inactive by either photo ionizing this gas with the accretion flare from the tidal disruption event or as Ito Berger studied in detail looking at the interaction with a newly formed jet with circum nuclear material so what I really like about tidal disruption events is this idea that we can actually probe the demographics of supermassive black holes and even probe some more exotic scenarios which supermassive black hole aficionados enjoy discussing for example potentially detecting a recoiling supermassive black hole that happens to be tightly disrupting a star turns out that the rates of events are very sensitive to the presence of a binary supermassive black hole system you could potentially disrupt white dwarfs around intermedia mass black holes which would be a signature of such black small black holes and dwarf galaxies and again you could look for signatures of spinning supermassive black holes from looking for events in in more massive around more massive black holes so once we probe these black holes we can always then connect them to their host galaxies and what's really nice about tidal disruption events is that are particularly sensitive to this lower mass black hole range where these nice relationships between black hole mass and host galaxies properties like velocity dispersion are less well constrained so it's an interesting mass range to probe with a new method so tidal disruption events are amazing and let's go find some so first what are we looking for well based on some sort of simple assumptions of this fall back of material onto the black hole which then accretes you expect sort of a burst and power law decline of thermal radiation and you expect it to have if you just look at the characteristic temperature of a disc radiating at around the Schwarzschild radius up to the around the tidal disruption radius that corresponds to temperatures of about 10 to the fifth to 10 to the 6 Kelvin so it's basically the typical temperatures of the inner regions of an accretion disk and so you expect that to peak in the extreme UV and soft x-ray and in particular we would really like to find a tidal disruption event in an inactive galaxy so that we can really say that before this event the the the black hole was completely quiet and starved of fuel and suddenly we had this dumping of material due to the tidal disruption of a star because as we know ubiquitous property if active galactic nuclei is in fact variability so we were a little wary about being able to differentiate between large amplitude ajm variability and this more dramatic impulsive process so unfortunately the rate at which stars come close enough to be disrupted is quite low within a galaxy so maybe once every hunt to a hundred thousand years but we're now at the point where we have surveys at various wavelengths that are imaging galaxies millions of galaxies in multiple epochs and we now are surveying enough galaxies that we can be lucky enough to catch these events in the act so in fact in about in the last two decades and in particular in the last decade we've really gone now from a handful to over two dozen tidal disruption event candidates and they've actually been detected all across the electromagnetic spectrum I've been focusing mostly on the UV and optical which I will discuss in a second but shown here is sort of a timeline of tidal disruption event discoveries of the first when they were reported the first handful of candidates were from the rows at all Sky Survey and then when I started my postdoc at Caltech in 2005 I started searching for these events in the UV and optical and there have been several more candidates have emerged from archival studies in the x-rays but that there's been a lot of momentum now in the UV and optical in the last few years if you notice in this diagram which is a diagram of peak luminosity versus publication year there are two sources here which really stand out these are the swift hard x-ray selected title disruption event candidates and in this case they have super any super highly Eddington luminosities which and they do not look like a thermal spectrum and in this case these events look like they're actually do degenerate emission in your line of sight and so these are a completely different beast but very exciting for trying to understand jets in jet formation and tidal disruption events but I'm going to be focusing on this more classical thermal emission that we associate with the sudden accretion of the debris onto the black hole through this debris disc so as I said the first candidates were from the Rosado all-sky survey there have been several more from things like the xmm-newton Slough survey and archival searches in Chandra and these sources are very convincing in that they have high luminosities large amplitudes they have extremely soft x-ray spectra which are really hard to very different compared to an AGM spectrum and their rates are consistent with what you'd expect from for tidal disruption events and if you place these these are all the rosette flares if you place all their points together in general they're consistent with this chi to the minus 5/3 parallel decline but these are very sparse light curves basically at best you have sampling of a few years in these events so my goal was even though the emission from these events should be peaking in the soft x-rays unfortunately we don't have an all-sky soft x-ray survey to work with right now so why don't we search under the lamppost and take advantage of these beautiful especially its optical supernova surveys which have daily cadence and high sensitivity in wide areas so that we can actually get detailed like curves of titled destruction events and try to do all the things that I motivated in my other slides namely try to get enough information that we could actually probe the black hole mass the mass and radius of the star it's polytropic exponent and even the spin of the black hole so let's look under the lamppost and also with using these wide field time domain surveys enable prompt multi-wavelength follow-up so that maybe we can capture that high energy emission in follow-up observations so the first proof of concept for this with searches that we did with Galax that happened to overlap with the CF h GLS supernova legacy survey as well as searches in purely optical data with the Sloane stripe 82 survey and in these cases we detected we sort of went out looking for nuclear flares with the galaxy data it wasn't necessarily nuclear because the spatial resolution isn't as good but we're just looking for large amplitude flares and in the what's really interesting is the optical sources are shown here the optically selected sources these two sources here were selected based on their UV properties but yet their optical light curves that we extracted after the fact from the overlapping optical data look very similar to the optically selected candidates and basically these look like blue transients that follow this T to the minus 5/3 power law decay so this was sort of proof of concept that we could study these high energy sources at longer wavelengths note here that we don't have any constraint for these sources on the rise time of the events which is one of the most critical Diagnostics for black hole mass so there was more work to be done so while we're starting to get some nice light curves we really want to catch these things on the rise to use them to weigh black holes so again I wanted to use the gallic satellite because if it's wide field of view and it's UV sensitivity but I wanted instead of studying galaxy evolution and the star formation rate history of the universe I wanted to use this telescope to study high-energy transients in the time domain and to do this we also wanted to team up with pan-starrs which is a beautiful wide field optical time-domain Survey in particular the pan-starrs medium/deep Survey which has about ten seven square degree fields that it visits every night as it cycles through four bands GRI Z and then Y bands during bright time so we took these two telescopes and they're actually very well matched in terms of their per epoch sensitivity as well as their field of view and we match the cadence we designed the cadence of Galax to be complementary to pan-starrs and so we get this nice coat I guess we call it now Co observing between galaxies and pan-starrs and so any source could then be studied any transient source could be studied both in the UV and optical without worrying about your decisions on follow-up so shown here just the pan-starrs fields that we monitored with Galax Galax are the blue circles and pan-starrs is the green circle and these are famous extra galactic legacy fields so the UV is actually a very exciting place to look in the time domain because it's relatively unexplored compared to the optical x-ray and gamma-ray and there are lots of science drivers for why you should care about the UV transient universe I talked at lunch about this exciting potential to detect core-collapse supernova very early after explosion we think that titled we are we know the type of disruption events should be intrinsically very bright in the UV for months to years but it also turns out that active galactic nuclei are very UV bright as well as have larger variability at UV wavelengths and variable stars also show a larger variability at UV wavelengths so here is now just the temporal overlap between the two surveys and when we analyze all of this multi epoch near UV imaging we find that the UV universe is mostly made up of AGMs and quasars here I'm just differentiating which look like point sources in the optical in which I have an extended galaxy in the optical but we're really dominated by accreting variable supermassive black holes we also detect some variable stars including M dwarfs are a liar II and some C B's but what's exciting is we have this gray slice here extra galactic transients which don't fall into necessarily the category of AGMs or supernovae and so these are where we are kind of looking for these potentially large amplitude nuclear flares from quiet and galaxies associated with the tidal disruption event so one of our most extreme or the most extreme extra galactic transients was in fact this one here which showed over four magnitudes of variability in the near UV and this is compared to all the quasars and AGMs in our survey so they also have large amplitudes of variability but on a different level than this source so this plot is really indicating that this looks like it's a true transient source in the UV it's not just a fluctuation of ongoing behavior it was also detected very clearly an optical difference imaging by the pan-starrs medium/deep Survey and so when we took a spectrum of this with the MMT telescope thanks to a collaboration with IDO Berger and Ryan Charnock here at Harvard it was interesting because some experienced spectroscopy looked at this and said oh it's just a quasar that's just magnesium to a trench of 0.96 so now you know in retrospect we should have said well even if it's a quasar that kind of behavior is unlike any quasar we've ever seen so who cares if it's a quasar it's still interesting turns out it was not a quasar because we actually had to wait until late time spectra so luckily Ryan Charnock knew that it would and we we had detected with galaxies we knew it was interesting but we were smart enough at least to keep following it even though we thought maybe was a quasar and over time the continuum emission from the transient faded away which initially was very blue and eventually got to see the calcium HK absorption features from the underlying galaxy indicating that this was in fact a galaxy at a redshift of 0.1 7 early type galaxies and that that broad isolated feature was in fact helium two four six eight six which is actually a high in ization line that had been originally predicted to be possibly associated with tidal disruption events because it requires soft x-ray photons so now we had the redshift and from the host galaxy luminosity so here is the before it's not detected in the UV here's the after here you see the host galaxy and the blue transient on the galaxy based on the luminosity of the galaxy the black hole mass in the center of this galaxy should be around a few million solar masses which is safely in the range of a black hole that can disrupt a solar type star and has low star formation rate and now here is the beautiful pan stars and galaxies like curve so the pan-starrs data with its beautiful three-day cadence really captures now the rise and the rise and decay of the light curve and this is a logarithmic scale in time and you see the rise to the peak and you see that the what's really striking is that the UV emission remains very bright over a year after the peak and so if any of you study supernovae or you'll know that sure we can detect supernovae in the UV at early times but they always expand and cool and the UV emission drops extremely rapidly on the timescale of days so this long-lived UV emission and this beautiful rise in decay was really striking so we can take this beautiful like her and fit which we're at the time where the best models numerical bottles for the fallback rate in a tidal disruption event and I haven't done any translation here between accretion rate luminosity have simply said that the luminosity in each band just follows the accretion rate and you get a remarkable agreement between the shape of the light curve in detail and this simple calculation of what the fallback rate the capillary and fall back rate as a function of time is for a type of disruption event so we can you can use them the rise time of this event to actually weigh the perpetrator here the black hole and determine that the black hole mass is 2 million solar masses but you still have an uncertainty here related to the mass and radius of the star disrupted what's really neat is recently or a last cycle we applied for HST follow up with the idea that while the source has faded away is probably faded away to below the limits of certainly ground-based imaging in the UV in principle we should still be able to detect this event now over four years after the detection and so baked just by extrapolating just a chi to the minus 5/3 para law here very simple you can predict you know given the range of models what you'd expect to see with HST and so we just observed it in June and in fact it's continuing to decline in this simple power law fashion and what's really fun is that while in the optical which is the gray and black contour is the grayscale and black contours while in the obstacle it looks like a galaxy in the UV it's a point source so the magenta is the UV image so it's you know right now in the UV all you're seeing is this fading tidal disruption event and it's still shining and we can use the beautifulest Rama tree or the spatial resolution of HST to really pinpoint and confirm that this source is in fact a nuclear we're going to start approaching some of the systematic errors in determining what is the nucleus of a galaxy but we can certainly continue to strengthen this idea that this really is a nuclear accretion event on to the black hole so we it took us a while to identify these helium two lines and there's a reason for that which is that you don't normally see a spectrum with only helium two emission lines you would also expect to see bomber line emission and so we can measure the velocity of this line and it's pretty broad about 9,000 kilometers per second and given the lack of H alpha detected we get a ratio of about greater than five helium two to H alpha so we if you use standard emission line Diagnostics that would indicate that your hydrogen mass fraction in the gas is very low okay so our idea was that okay this these lines are telling us about this disrupted star and in fact we're seeing high-velocity debris from the star that was disrupted and that this star that was disrupted was hydrogen poor it was actually helium rich so potentially we're seeing here the victim in the crime scene we're seeing seeing this high-velocity he lionized helium gas indicating the type of star was disrupted so that's when we could really put our black hole behind bars because if we assume that this was in fact a helium rich star based on the lack of star formation in the host galaxy we really and the timing of the event we rule out an evolved star like a massive wolf or a star which is lost its hydrogen due to winds we think that in fact this could be the core of a red giant star that was stripped by the same black hole so is stripped of its hydrogen at earlier times and now we're seeing the disruption of the helium rich core and the we can measure the luminosity of the vent the total energy how much mass a lower limit to how much mass was accreted and based on what we think the structure of a stripped red giant is we can plug in the mass and radius of the core of a stripped red giant and then pinpoint the black hole mass even further to three million times the mass of the Sun so the idea is that in this event not only did we have this beautiful light curve to get the timing of the event but we actually knew about the structure and chemical composition of the star disrupted so we could actually model the black hole mass so I'll say that it's you know it's hard to put error bars on a mass like this because it really depends on your assumptions for the star disrupted and it's also model dependent in that so far I've been using these mass accretion rates curves to predict the timing of a flare but of course what we observe is the radiated luminosity so we LeSean at all tried to take their mass accretion rate curves but then look at the structure and emissivity of the disk producing the emission and try to actually model the radiated flux and then see how that evolved with the function of time and so what's really what's really problematic matic and this source is that given its low temperature and the fact that with a declining luminosity you would expect a declining temperature you would expect band effects in that in the UV an optical band you would expect a different power law decline if you're below metric luminosity was tracing the mass accretion rate so they have a really hard time fitting this light curve when they use sort of a realistic disc like spectrum for what they think the creating debris disc should look like and so in order to get the model to fit the data they basically have to make it such that the specific luminosity in the UV acts optical band basically follows the mass accretion rate and they do that using a reprocessing layer and I'll talk about that in a second there's also possibilities another way to explain the low effective temperature of the submission so we fit a black body temperature of thirty thousand degrees Kelvin another way that this has been accounted for is another paper which thought that maybe it was a very low viscosity disc and that produced a lower temperature and then there's also this other factor and that some of the luminosity is coming from the viscous evolution of the disc and some is actually coming from the accretion process so this is all going to change how you relate your model your light curve fits to the actual black hole mass so what is the nature of the continued emission in these tidal disruption events that were detecting so all of our events are shares similar characteristics and they seem to show this hot thermal emission and they do show this nice power law decline of behavior but their temperatures are actually a lot lower than what you'd naively expect from an accreting debris disc so shown in this diagram are the bull the bull met people metric luminosities of the events versus their black body temperatures and you also see here a selection effect in that the optical events that we detect are biased to lower temperatures and the x-ray events that we detect are bias to higher higher temperatures if these things were really black bodies emitting at a fixed radius then they should follow these diagonal lines on the luminosity temperature plane and shown in the dashed red lines are the Swart shield right radius of a 10 to the 6 to attend the 7 solar mass black hole and the gold lines are the radii of the titled disruption radius of the 10 to the 7 solar mass black hole 10 times the tidal disruption radius and a hundred times the titled disruption radius so basically from theory you might expect that you're gonna see the secreting debris disk with emission coming from may be as small as the Schwarzschild radius and may be as far as maybe twice the tidal disruption radius and that's gonna give you temperatures that should scale with luminosity in this rectangle here so we certainly do not see that what we do see is that the UV opt and optical candidates seem to be telling us that the radiation is coming from larger radii we also appear to have a constant temperature we both observe fixed colors with time but we also don't see this shallower decline that you would expect if the temperature of the emission we're changing lumenocity were declining we still see this T to the minus 5/3 parallel decline the x-ray candidates have the opposite problem they appear to have uncomfortably small radii some even smaller than the Schwarzschild radius and so perhaps they're probing more more something like a 'no creating hotspot in the inner regions of the debris disc so there is some old work by avi Loeb and Omer in 1997 thinking if there were some sort of envelope of debris that could reprocess this high-energy emission from a tidal disruption event that would be really fortunate because the temperatures and luminosities would be much better for optical surveys to detect so in fact that seems to be the case we seem to be finding this type of emission in our candidates so the question is is that what's going on is there some sort of reprocessing layer the only issue with the reprocessing layer is that it has to have a sort of a complex behavior and that it has to regulate itself such that it has to shrink in just the right way for the temperature to remain constant so basically if the luminosity is declining to see the minus 5/3 you need the radius to be climbing as T to the minus 5/6 so that you end up getting a constant temperature but there might be physical ways in which the radius of the photosphere could be changing due to opacity and I think James and Enrico we're starting to investigate that in their paper so is our object PS 1:10 JH the archetype archetype so it turns out that the Palomar transient Factory went back into their archives of all of their transient detections and they ended up finding several or three nuclear flip flares one of which in particular had a striking resemblance to PS 110 JH both in its light curve so the open circles are our object and the solid circles are from the Palomar transient factory object and they have I don't think there's any scaling here I think they just literally pasted them on top of each other but you can see that the rise the shape of this initial peak as well as even the late time decline is very similar to our events but not only are they similar photo metrically but they are also similar spectroscopically in that this object which has the nicest light curve also appeared to have helium broad helium emission and no corresponding H alpha emission and they have two other objects they don't have as nice like curved sampling but they do have spectra for their other two objects there was also another type of disruption event candidate reported in a nearby galaxy which had a very nice late time like curve sampling and what they found is that there seems to be a spectrum of spectroscopic signatures from these tidal disruption event candidates so all of these were selected from optical surveys but it looks like they show either they show only broad helium emission or some bro show both hydrogen and helium emission and so this is exciting because this is potentially we're seeing gas we're actually tracing now the kinematics an ionization state and chemical composition of the gas in these tidal disruption events so there's a lot of information here so the question is what is the origin of these broad lines originally in the literature people focused a lot on line emission that could be come from the unbound debris intercepting the accretion flare but I think this line emission was expected to be quite weak and Enrico and James I think emphasized in this paper that in fact the the cross-section of the stream is very is very narrow and so you actually would have very weak or unlikely catch in the line of sight emission from that unbound debris so they proposed in fact that the the line emission is coming from this newly formed accreting debris disc that somehow circular eise's and starts feeding the black hole and then there there was an so then the question is you know are we seeing so let's say these broad lines are coming from this creating debris why are some high why do some only show healing emission where is the hydrogen emission line hydrogen emission in these objects so either it's because the disrupted star itself is hydrogen poor or the alternative explanation is that it's due to the ionization state of the gas and you're sort of looking at the inner regions of you could think of as like a truncated normal a GN disc and if these conditions would just suppress the hydrogen and emissions so that you would get an enhanced helium to hydrogen ratio so there was a recent paper by Gaskell running cloudy and using sort of a truncated broad line reaching cloud and trying to get as high a helium to H alpha ratio as possible and what they found was that you could get a ratio of about up to four but only at a certain critical density of about 10 to the 11 per cubic centimeter so we can pretty stringently measure the amount of H alpha both at peak in our spectrum as well as we now have some nice a late time spectra the late time spectra are nice because now we're just seeing the host galaxy light so we can improve our subtractions to isolate the emission from the transient and we're pretty sure that our helium 2 to H alpha ratio during peak is greater than 5 and we still don't see any H alpha popping up in this late time spectrum however we always have to be close to our observers because it turns out that rest frame H alpha happens to land in a telluric line so we're trying to be very careful ryan Charnock is being very careful in his reduction of this spectrum so one of the predictions of the ghoulish on paper is that as this debris disc ice expands with time so shown here is their model with the this is ionizing photons flux versus density and as this as this disc expands with time the densities fall and you start to get into a regime a more normal regime seen and other agents where you would have a stronger H alpha d helium ratio so the red line is showing where our September 2012 spectrum lands in their models and at that point just according to this plot this isn't a model this is a set of models from an Ag on paper that they used but based on where that red line lies you would expect an H alpha to helium ratio of greater than 3 so you would actually expect H alpha to be stronger than helium 2 and we can I think that would actually be detectable in our spectrum if that wasn't in fact the case so it seems that we don't see the H alpha line popping up as you would expect in this explanation also we still have to understand why these PTF tidal disruption events which also had spectra at early times why why do they show hydrogen when you'd expect it early times for the hydrogen to be weak so this is a very exciting area and there's a lot to be learned from the spectroscopy of title disruption events in real time so I think we're entering the era of precision tidal disruption event observations where and precision I mean we're going from light curves with the time sampling of years to a time sampling of days and the next great LSS the nice great a title substrate event factory is really going to be LSST because of its major increase in survey volume and by scaling the detection rates in our study with the Sloan survey you would expect that a survey like pan-starrs would maybe get a couple events per year which isn't maybe a factor of a few off of what we observed but something like LSST is going to detect thousands of tidal disruption events per year but if we're going to find thousands of tidal disruption events we're going to find millions of supernovae which is great for some of you in the audience but not for me and so the question is how in the world are we going to sift through all of these transients to find our really exciting tidal disruption events so that we can then go take our spectrographs and go take our other follow-up instruments to study them in detail and study their central black holes so with the Sloane study we found a parameter space where these tidal disruption events did stick out from supernovae in particular and that was in the color evolution with time versus their color so basically tidal disruption events have constant colors with time and are intrinsically very blue so they are hot and unlike a supernova which expands and cools tidal disruption events seem to have relatively constant temperature so you can use this parameter space to really rule out most supernovae you can also look at the host galaxies colors to look at where title assumption event host appeared to be redder than of course the star forming like galaxies of hosts to core-collapse supernovae so we've tried started trying to do this with pan-starrs so Sloane had the you band and LSST will have the you band in pan stars we did not have the you band but we did have these beautiful deep light curves of many optical transients and my graduate students ed Kumar went through and wanted to classify all of the extra galactic transients so any transient with a galaxy host and went through and did some very intensive light curve fitting to decide we basically just made the hypothesis that the extra galactic universe is made up of two types of transients AGN and super novae so it turns out there are other exciting things in there like tidal disruption events but that was our main hypothesis that either you have bursting transient sources or you have stochastically varying sources and so we did he did like curve fitting to determine is something so classically variable or is it burst like then with the next step we said was okay what is it's offset with the host galaxy we made the other basic assumption is that anything that's nuclear and stochastically variable is probably an AGM anything off nuclear that is a bursting like curve is also probably a supernova it's probably associated with an exploding star then but then there's some other off diagonal terms for example a nuclear source that is a bursting light curve certainly it could be a supernova because supernovae can go off near the Centers of galaxies but it could also be a tidal disruption event which is what we're looking for so we use this combination of like general light curve class and offset and we can do some fun things like look at all of our photo metrically identified a GN and supernovae and look at some of their basic properties for example host magnitude or the difference between the transient magnitude and its host and you get this neat relationship that the Aegean and supernova kind of lie in different parts of these parameter spaces and so instead of spending the hours of crunching through all of these light curves at the telescope with just a couple measurements for example with an LS s T transient alert if all you had was the host magnitude in the eye band and it's transient detection in the eye band you could tell you could give us some probability of whether it wasn't a GN or supernova so if the transient was very bright compared to its host you're pretty sure it's got to be a supernova whereas if it's in this regime where it's a close in magnitude or fainter than magnitude than as host then it's harder to differentiate between a supernova and a GM so we can take all of these photometric classifications and kind of look at this space again of color decay versus color and this is very busy this is what it's like for me every day looking through looking for something interesting this is what you have to deal with so you stare at this and you stare it like curves and in here are amazingly exciting objects but they're also boring ones too and what's neat is that in fact the supernovae of course do stick out in this quadrant here they're red and they evolve quickly with time they cool quickly with time you'll notice that there are purple circles here these are nuclear transients so we didn't we couldn't decide if they were in a GN a we couldn't decide if they were a supernova or a tidal disruption event so we just call them nuclear transients but probably if they're in this box they are supernovae the yellow highlighted are supernovae that have been followed up mostly from the harvard supernovae pan-starrs program so you see that what we say our supernovae are in fact supernovae when you observe them spectroscopically but then you have this see here of blue or objects with less color evolution and are bluer and the two stars are the two spectroscopically confirmed tidal disruption events from pan-starrs and so while they are in there and maybe there aren't that many nuclear transients around them so you know maybe there's a manageable number it's still a little bit daunting how much potential contamination there are from supernova and AGMs and so what if we then took out put on our any near UV goggles and we could figure out which of these sea of transients actually had bright UV emission we're also bright UV variable sources and ah suddenly things get a little quieter which I guess I mean that's sad and that you have less sources but you know that these sources a lot of them are really interesting so the UV preferentially picks up mostly the AG ends and quasars the quasars of the blue dots AG ends are the green circles the UV does pick up a few supernovae which is always exciting because supernovae are only bright in the UV at early times and so at lunch at the lunch talk I talked about at least one of these super novae where we actually caught the rise of the supernova in the UV and so it is very exciting when you see a supernova in the UV but there very few of them that actually get detected because they only last for a short period of time and it also depends on the type of star exploding but what's exciting is here you have our - tidal disruption events and now the confusion from other sort of nuclear transients has really quieted down and there are only a few sources left in this color space that could potentially stick out as tidal disruption event candidates and a couple of them that I don't know if you can tell are in yellow we have spectra of and it looks like they're in fact a GN and we could also look at another view like their host colors and you can see that the title disruption event hosts are much redder than a lot of the other hosts and so that could be another important discriminator so my main point of this is that while the optical transient universe is exciting and active it's also very confusing and it really helps to have another perspective on transients especially if you're trying to find high-energy events like tidal disruption events it really helps to have that high energy information so what is the future I'm sure there are definitely people in this room who want a space mission at higher energies to complement LSST and shown here are all European missions this is just a proposal a UV proposal I Rosita is gonna launch in 2016 but for that to actually complement LSST you'd have to extend its baseline mission and then there are some ideas for a hard x-ray all-sky telescope which would be very sensitive to these relativistic tidal disruption events these jetted events where the emission is non thermal but really we need these are all European emissions we really need a NASA mission so that we're ready for for LSST to be able to capture the most exciting energetic events so to summarize tidal disruption events are potentially a relatively clean laboratory to study black hole mass and the nuclear stellar populations and galaxies and potentially black hole spin there are very important tests of accretion physics meaning that we have to understand the accretion physics for us to be able to model them properly and they can also tell us about jet formation and black hole environment I showed you one example where we think we actually caught a black hole red-handed in a stellar homicide whatever that means and the future really is promising for detecting more of these events but we have to be smart in how we find them and having a high-energy emission high-energy mission like Galax or roast that would be extremely powerful Thanks absolutely yeah it's remarkable that we see events that show this simple tweet of -5 there's behavior I mean it doesn't make any sense that it would follow so cleanly because like you said that's the fallback rate that is not the actual accretion rate on to the black hole because you're assuming it's the biggest time is very short and somehow the materials are creating out that rate but it's it's shocking and it really does look that simple so but certainly yeah if there was some sort of time delay for example between the disruption fall back and formation of the disk somehow for some reason accretion rate on the disk at that same parallel behaviour observational II we can't we can't probe that time delay unless something like Enrico's worked on the possibility that at the time the star is disrupted you have some compression they could form a shock they could produce a burst of x-ray radiation so potentially you could actually time the time of disruption but that's very faint and difficult to detect so use methods to weigh the black hole and you said that if you find some that are greater than ten of the eight solar masses you can infer that spinning but to the fact that it's spinning make your methods for way in the black hole any less reliable or is that not an issue so there have been calculations for the fallback rate so taking these curves and predicting them in the case of a spinning black hole and there are some changes in the timing of the event but the actual shape doesn't change so I think there could be some subtle changes in the in the timing of the curve but it would probably be on a smaller order than the fact that the timescale you measure is so long that it must be a very big black hole I could imagine that early in the universe all of the stars that were easily going to be eaten to eat and many other situation like the Milky Way we had a lot of sparse basically orbit around the black hole and they're going to be perturbed perhaps only a very little bit by the slight granularity in gravitational potential and so I'm wondering perhaps almost all of the impact parameters might actually just be marginal to give you only a little bit of title of destruction and the star might even survive another couple orbits and be tidally disrupted increasingly as time went on but perhaps a little bit most of these events are with impact parameters that are very very marginal for getting actual disruption of yeah Enriquez group explored that parameter space and found that there was a large range of impact parameters in which the core of the star remains intact and there also have been some papers thinking about like he said multiple passages and some sort of periodic signature from the accretion of the star but yes certainly the rates depend on how often stars are scattered into the Moscow I would suppress when you showed those colorblindness tests yeah yes yeah that you didn't use the time history it seems to me the title of disruption only happens once whereas an IGN player that's how I made these shapes so I defined so nuclear transient means that you had a bursting shape but only once only once exactly well anything's stochastically variable that was in a galaxy nucleus I called nay GN a stochastically variable source off-center from it galaxy nucleus would show up here probably as a gray circle and that could be a recoiling supermassive black hole but before you get excited it could just be fattest rama tree so this is this is take the all of these classifications are photometric and they take into account the light curve shape and the offset from the host galaxy what about the long term history you have five years of pain starts today's young'uns come back exactly so anything we produce I call an AGM is specifically modeled with a damn random walk stochastic so it isn't as bad as you make it you know I know what these are ends still a headache I know that all of these are bursting transience but they could be super novae or they just ignore the green boxes yeah yeah I'm going to read out green tea right it's not that bad take your means the purple is what we're after t-shirts I think it's quite important so I think the point that we're trying to make is these are AGN that are made from the inner parts of the outer parts you know so the disk basically start from the inner part and expand and if you take a broad line region of an Ag n you know because of reverberation mapping and also looking at the width of the lines you know that basically the higher ionization lines are very deep in and as you go outside that are the lower annotation lines so the prediction is that initially you know everything will be highly ionized you want to see it and then you go from the high initiation to the - agent stay now you're absolutely right that we really have to understand in detail the structure of the reprocessing layer and assuming maybe that it's a scaling loss of the problem and Megan are very similar you may be incorrect because here you have this elliptical disk that is not you know perfectly spherical expanding but then if you really want to go into the other regime which is okay I'm gonna assume that the star is a helium star particularly those masses just a Rado you know pure healing stars it's tiny and even if you have a red supergiant if you want to treat the hydrogen you know what we show is that you can also you know you can not strict as much as you know hydrogen as that remaining in the core because you know these red giants tend to rather than expand by losing mass and start contracting it's very very difficult to do it so I feel like I don't believe that you know we know the Iron Man relatively well and we don't you know and now we have like forty percent of all the tile disruptions being in helium stars I think father will remain there eight other main sequence and what we're seeing is you know they're changing in this view process unless you I mean the environment that these stars are in our extreme I mean now we have this g2 cloudstock any measures it's nothing like the the life story of the red time so you will have to go on a straight example and then changes periastron you know it's very centered distant typically to disrupt the corner and the probability of that is miniscule so that is for the there is the wrestle about I there's a weeper yeah okay that's right but there any more subdued or there's a paper there's a paper that goes through the dynamics book the Nova chat all and they say that you know as long as you you have to create all that hydrogen gas that got stripped but they do have some orbital solutions where you could get the core sub was greatly disrupted but my argument against the the hydrogen to helium ratio is that it's just so hard to suppress the hydrogen enough and Gaskell barely did it with sort of standard cloudy calculations so even if the hydrogen is ionized you're still going to get recombination and then it just it's very I think it's just really hard to get such an extreme ratio so until I see a calculation of does it will show you okay I think it's not so then why are these guys at early times showing hydrogen barely mentioned yeah so they're useful analogy there you weights in between - I mean this is her - hot off the presses that I don't have not on the group's working on it but now the idea is that g2 isn't just a simple gas cloud but g2 might be a it does have an interior star that's still intact but it has gas that's being stripped off the layers but we still have to explain this bizarre dusty envelope around the star so in the guesses groups paper they suggested it could be a stellar merger so that's fun and it still emerges are rare but it could be that you know in the environment around supermassive black hole that's not crazy so back in the day Kobayashi at all 2004 predicted that you should see a population of helium stars and you know stripped red giants in the Centers of galaxies so I don't I mean I'm not I'm fine with either explanation I wouldn't say just because they were rare I think it's still interesting because these are extreme environments but I also just know from you know the spectra of AGMs over time and even supernovae it's just very hard to suppress the hydrogen emission so if you could do it that we have to understand you know how you do it and and then how you get the different but we have more observations the better more overtime IOC you know pointed out to us is that you do see this trend that things that peak at very early times which means that that this has been developed as much so no hydrogen and helium things that pika slightly later show you know basically the helium and they're these things that basically shows on hydrogen that in fact if you look at this standard scaling laws of a GN they're very far off than hydrogen and you know a mission lighter or argument there is maybe you have a receipt drill disk you know before the time disruption and you have enough gas at low intensity you okay I think it's a it's interesting that we've seen a broad line reach and form in real time absolutely and yeah I think to try to understand what other properties of those definitely and and actually modeling your line ratios as a function of time is extremely exciting and we as observers we need to give you more spectra tomorrow great so we all have

Channel: CfA Colloquium

Views: 5,305

Rating: 4.8333335 out of 5

Keywords: Black Hole (Celestial Object Category), Astronomy (Field Of Study), Astrophysics (Field Of Study), Supermassive Black Hole, Harvard–Smithsonian Center For Astrophysics (Organization), Physics (Field Of Study)

Id: L_WiYE6ahRk

Length: 70min 54sec (4254 seconds)

Published: Thu Oct 23 2014