 Thanks very much Yep, I think I'm on so thanks for the invitation to come and come and talk about about plank and and the cosmology that we've been able to learn from it From its now from its full mission of data. Actually, it's satellite that flew for a few years And our isn't it not on Interesting it definitely looks like it's supposed to be on It's on super Let's try that better Good, okay So I want to show you some of our results our Main data that we're extracting cosmology from the main statistics We're using and what we see is kind of this of the key cosmological results So this is this is drawing on the work of a huge number of people and a huge number of groups a large collaboration of people And some of their logos are pictured here. This is definitely not single-person work. This is this is large large collaboration work Okay, so what I want to show you first is what is our main? What is our main data that we use to talk about things like primordial fluctuations testing inflation testing the contents of the universe and our main data to begin with really our Maps of the microwave background sky and I'm going to show you more or less three of them because it's three that we're currently using for cosmology This is a this is a this is a map that you've perhaps seen more Frequently than the other ones. I'll show you which is the Image of the temperature anisotropy in the microwave background so the temperature of the of the of light at 400,000 years of recombination when the universe transitions to being neutral We're capturing the fluctuations in that light that more or less trace the density fluctuations in space at that time at 400,000 years, and we've now mapped it out. This is an all-sky map unwrapped onto the screen and we've now sort of over the last 20 years we've kind of zoomed in on this on these features with higher and higher fidelity now at High resolution and high sensitivity measured from from plank and these temperature fluctuations the scale here is tends to hundreds of micro Calvin these are fluctuations about the mean temperature Small linear fluctuations that allow us to capture the physics of what's happening after 400,000 years and trace It back from inflation perhaps or whatever mechanism was that put in the initial fluctuations through to the time we capture them So this is the one that you've probably seen more we're now to do our cosmology. We've got new images of the sky And then the second one is the polarization anisotropy of the microwave background So what I'm showing you here. They're not as pretty pictures. There should be prettier pictures coming coming soonish from plank We can measure the temperature of the microwave background We can also measure the polarization of the light you can measure and so what we actually the physical thing We measure our q and u stokes vectors So polarization with this orientation in q and this orientation with u and we can map that over the sky as well as the temperature And that's shown here top and bottom for the q stokes vector and the u stokes vector And we're starting to see and I we're seeing these anisotropies that skirt that that gray the gray plot has got lumps and bumps in it And these anisotropies are tracing the physics of what's going on at recombination as An independent probe the temperature, but it's the same physics what we're what we're seeing The temperature of the C and B traces roughly the density fluctuations in the universe at 400,000 years And the polarization roughly traces the velocity So we have this tightly coupled photon barrier on fluid that evolves after primordial fluctuations are put in And we capture what that what that fluid is doing At 400,000 years both in both in density. It's density in its motion of the of over densities I'm just showing you a zoom in of a little patch of the sky about a hundred square degrees a small region sort of this big Measured from the ground-based telescope called at pole That complements plank by being able to zoom in on in a bit more detail on the polarization And here you can sort of see the fluctuations These are anisotropies in the polarization of these two q and u stokes vectors and It's not coincidence that your eyes will see a shape pattern like this in the top one and this in the bottom one that's the signature of Emote type pattern of polarization, which is the kind of it's a it's the pure if you take a polarization map in Of the sky and you look at it's the divergence part of the map and the curl part The divergence part is the thing that we think traces the motion of the fluid the photon barren fluid and that's what we're seeing in in In the signal here The other part that we could see is a b-mode a curl type polarization That that we haven't yet seen but I'll come back to that Okay, so so we have we get to trace these are We have temperature the cmb we have polarization of the cmb and now we also have this new one Which you'll see increasingly over the next decade Which is the anisotropy in the lensing potential of the cmb? so the micro background comes to us and gets gravitationally distorted by Cosmic structure that lies in between us and when the cmb set off Typically a photon will get deflected by a couple of arc minutes very small amounts of angle on the sky and So we can think of it as you know that the temperature of the cmb we see in some direction of the lens cmb is The unlanded cmb from some original direction Plus the gradient of some deflection field so a distortion a distortion angle It's given by the gradient of some lensing potential And That lensing potential the thing that bends the light is an integral of all the matter lying between us and last scattering And that's useful because being able to probe the integral of everything that's lying between us and last scattering Gives us a handle on things like the geometry of the universe It tells us more about the dark energy about what what that could be and also tells us about the mass of neutrino particles This is kind of an exciting new thing to be able to see and so in the same way as the Cmb temperature anisotropy where you see blue spots The slightly less Lensing in that direction and red spots is like a hot spot. It's more lensing And so you should think of it where you see red The integral of the matter between here and where the cmb set off is higher than over there There's more matter in that direction than over there And this this kind of where we are right now with these lensing maps is Sort of you could think of it maybe like how Kobe was for the cmb temperature As in these these fluctuations you're seeing you're seeing them on large scales They're still a bit noisy But this this lensing map over the next decade should take the next steps should should really kind of Improve in the same way that we've seen that the temperature of the cmb improve But we're already learning things from it One one question I always get asked and then forget to explain so let me just say this now is It's how enough do you figure out what lends the cmb when you just get to measure light that's been lensed a Cartoon of what's what happens to this light is below This is a little snapshot of part of the cmb sky Before and after being distorted by intervening structure that's bending the light around it And it's a very small effect, but something that you can see slightly as though a coherent shifts in the background features Because the light is basically being bent around giant cosmic structures and what we do we actually use the fact that Different Fourier modes of the background signal are not coupled if there's no lensing But bending round a large object couples different Fourier modes together And so we construct an estimator of what did the lensing by looking at pairs of That's of the background temperature at some scale and some different scale length scale And we couple those together and estimate the lensing signal from that I'm not gonna say too much more here because it's maybe It's too You know you don't need to don't need that right now, but it's We appreciate it's not a trivial thing to do and one of the things that we work on a lot right now within plank We then have the collaborations is to is to get this better But this is important because this signal is the signal where you know, we hope when I come back It's in a decade We hope to say we have detected the mass of neutrinos from this and it will come from an improved measurement of this this map So now we'll come back to that Okay, so we have these that's I would say that's kind of our core data are these are these sort of Samples these maps of the temperature polarization and lensing of the of the CMB We then look at their statistics and the statistics we look at are Pretty much the two-point function of those maps because for Gaussian fluctuations, which they do seem to be That captures all of the information in those maps Although we do look for non-Gaussian features, too And so here is the the latest Power spectrum angular power spectrum from plank from this 2015 Data so up here is the is the variance in the map or the power spectrum of the map as a function of angular scale and these are You know tens tens of degrees or more Degree scale here and then sub-degree scale running along here to smaller scales And this is the the angular multiple moment of a spherical harmonic decomposition And so we're in this, you know, we plank is sort of Beautifully measured with very small errors this this the power spectrum now from large scales through to small scales here right down to these these a few arc minute scales and Shows incredible consistency with that red theoretical model That is our model. That's that's the lambda CDM cosmology So we do find a model that fits and it fits very well And you know, this is a flat universe with this handful of parameters that I'll discuss in the next the next slide Where we're seeing we think we're seeing, you know, just simple primordial fluctuations evolved through To recombination in a rather simple universe filled with neutrinos, baryons, photons, cold-up matter and a cosmological constant And the fit is really good I mean so so here down here shows the residuals of this lambda CDM model the red one compared to the data And the error bars are so small that you kind of have to see it on this residual plot and So anything that isn't this kind of simple lambda CDM model would show up as deviations of these data points around that red line Now one thing that your eye, you know goes to straight away is here This sort of dips this lack of power at the largest scales has been seen for a while is still there is a curiosity It's kind of a it's kind of deviant at like two to three sigma, but it's but not not a higher significance It's interesting At smaller scales there are some points that the deviate from the line, but not significantly And actually something that's happened in the last couple of years with improved plank analysis is that the kind of the residuals of this data With respect to the model have just really squeezed in So there's incredibly little wiggle room of anything that's not this model Compared to the data and let me just say just a couple a bit more specifics about what this model is Because what we do is we throw in six numbers Six variables and a handful of assumptions of assumptions about the universe when we're when we're predicting the theories that match the match to the data We assume a flat universe I'll come back to that We've seen flat universe and we can constrain three numbers that describe its contents and the expansion rate of space And we quantify them by a baryon density a cold-up matter density and a peak angle at which we see the CnB acoustic peaks and that maps on to What the contents of the universe are so for example if I tweak around the amount of dark energy the dark matter the cold up matter the Baryons that would change the peak angle. I see the CnB at We have three numbers that describe those And we have two numbers that describe their primordial fluctuations We think they're just Gaussian They're adiabatic So all the different fluids in the universe trace each other at early times they had they follow the same over densities And they are described by a power law With an amplitude and a spectral index and that spectral index differs from one It's not that's one of the big results That's that's consistent that's held up with the with the sort of improved analysis of Planck It's still there right the spectral index is significantly different from one Of that that describes that the fluctuations and those are the kind of probably the five parameters of main interest to this audience We also have an the CnB gets scat when the stars light up the universe the CnB light gets scattered and damped the signal And so the epoch when realization occurred affects our signal And is of great interest to many astrophysicists But perhaps of less interest to this audience That's one of our numbers And we have a bunch of assumptions. We assume the universe is flat We assume that the dark energy is just a cosmological constant. We assume there are three Species of neutrinos. We assume the primordial helium fraction is set by Big Bang Nucleosynthesis that it's 25% of the universe We assume that the total mass of the neutrinos in the universe is is 0.06 eV Which is the minimum mass scale that we get from? Neutrino oscillation experiments We assume there are no tents of fluctuations no cosmic strings and manic fields nothing right all those things are not are not in this this model and this is the model that really fits and And You know we looked for a lot of things that differ from it, and they're just not there And that's really held up even more by looking at these extra maps of the sky the polarization the lensing So here what I'm showing you is The this is now the power spectrum of the emode pattern that sits in the polarization maps If I take my maps of the CnB polarization, and I look for the pure divergence pattern in them And then I compute the two-point function of that Then that's shown here the the power spectrum of that emode signal as a function of Multi-pole again angular scale a large to small scales Here in blue are the data points from Planck and in red this this red curve is not fit to these data points That red curve is just lambda CDM The lambda CDM model that fits the temperature data plotted on the polarization data So the fact that this wasn't even fit to the data, you know says that this model is doing something right It provides a pretty good fit One of the residual worries of then Planck that's being worked on right now Is that if you zoom in and you look at sort of the the difference of the data with respect to the red Lambda CDM model there are there are some more Deviating points than you see in the temperature and We do think there are some residual systematic uncertainties left in the polarization data that kind of the one sigma or so level Probably from some of the temperature signal leaking into the polarization And that's why the main cosmology results are not don't use this polarization yet We sort of looked at it, but really the main results still come from temperature And that's something that you should you should see change in the next year But it's not you know, we're talking about small shifts. We're not talking about something significant We also see the correlation between the temperature and the polarization Down here again red curve not fit to the data It's just the prediction of lambda CDM of the cross correlation between the fluctuations at the temperature and the polarization of the signal Third then the third piece the third piece of that of the data is then the Two-point function the angular correlation the angular power spectrum of that lensing signal the lensing potential This then again shows the power spectrum the two-point function of the lensing signal as a function of multipole Again, these are degree scales in the sky Here the data from plank And again the black curve is just lambda CDM fit from the temperature data. It's not fit to this data And it provides and it fits very well You know, there's kind of this dip here, but it's it's not it's not really significant and this is useful because What this allows us to do so what yeah what this What this measurement allows us to do is do things like as coming back to neutrino mass But also at a more simple level measure the geometry of the universe that the amount The lensing essentially proposes the amount of clustering of cosmic structure at later times in the universe It typically picks up the size of cosmic structures kind of halfway through the universe's history Dominantly it's an integral but that integral peaks kind of halfway through the universe's history So we're picking up most of the signal kind of a kind of seven or eight billion years But it's but it's a broad kernel. It picks up signal from from earlier and from later And if we had a universe that kind of fit the background CMB data But was more geometrically was closed and had less dark energy than the under CDM It would actually produce more lensing because you'd have more clustering in the universe So if you have more clustering happening you boost the lensing signal and you would push that that whole curve up and that's given us these Strong constraints on the curvature of the universe So So here I'm showing you here the the fraction of the universe energy of density in dark energy or cosmological constant here Versus matter density here where the dash curve is a flat universe and anything off that flat curve is a In this direction is a closed universe And it used to be with a CMB you could go all the way down that geometry down to zero dark energy And now you can't because the lensing tell it gives us a lot of new information the lensing prohibits, you know I've asked the closed universe and so Let's just concentrate on that on that blue Contour these are the 95% confidence limits for those two Contents of the universe and you see it's tightly Closed in around the flat the flat line and if I then Expressed it in a in omega k the the curvature of the universe Then just with the CMB I get a 2% is a 95% limits that the maximum deviation from Flatness is 2% just from the CMB Now if I then add in the positions If I look at then galaxies at later times that I look at where the galaxies are This is from the Sloan Digital Sky Survey Galaxies are typically separated by preferentially separated by a particular distance and that's seen here from the Sloan data They're typically separated at this this kind of 120 megaparsecs Where a megaparsec is a typical separation of a galaxy of pairs of galaxies There's this peak separation in that and that and that peak separation Gives us a handle on the expansion rate of the universe at later times And if we throw in that as well, then you get down to a 0.5% constraint on the curvature of the universe. That's that red curve here So really there's any any wiggle room from flatness is is is you know is vanishing If I look at now at the primordial fluctuations This shows our current constraints from both Planck and data including new data from a South Pole Telescopes the bicep to you and Keck array Down here is this is this power is this power the spectral index of fluctuations that I mentioned before where a value of one is scale invariant fluctuations and n less than one is Has less power at smaller scales and is what we would naturally expect from you know many simple models of inflation And this is as I said This is one of the main one of the main results from Planck is that if I just look at that red 95% contour I'll come back to this number in a minute It excludes one at high significance And it zooms in on kind of point nine six point nine seven that that range But now obviously the thing that the thing that's that's certainly of interest is do we see any gravitational waves? So if inflation happens, we would expect to see a background of tensor fluctuate We said expect tensor fluctuations that would source a background of gravitational waves And the strength of those the size that signal will depend on the energy scale of inflation And it's something that obviously we all would like to know If we can see it and what it is because that can that would tell us a huge amount about Both whether inflation happened and it's and its characteristics And so here is this we characterize it by a tensor to scalar ratio that the how big the tensor fluctuations are compared to the scalars and In a given slow roll inflationary model You know every every model will have a have a predicted space point on this NSR plane So here for example would be a Slow roll implant on potential with a Phi squared shape here will be a Phi shape And so here so with just with the plant data We limit this space in the NSR plane to here Where you can just about get this fiasco potential, but a five to the four term is it's gone off that space It's no longer doesn't no longer works and Those just come from meant looking at the temperature anisotropy But what we're really all looking for is is this this specific prediction that gravitational waves will put in a Polarization signal that has a curl part that might look that has this kind of pattern in the polarization and We have not seen that yet a primordial b-mode signal And that's you know We all heard last year that we thought we might have seen this from the bicep to experiment in the South Pole who did see a B-mode signal and would have put if the if the signal was just pure gravitational waves Would have put us up there somewhere a point that's actually inconsistent with blank But there's actually no evidence that it's gravitational waves and everything's pointing to the signal that was seen Just being a mission from our own Milky Way galaxy So that right now By including information from the Planck satellite that tells us more about the signal from the galaxy We've managed to sort of clean up the data from bicep to Clean it up and put actually a new limit on the tensor scalar ratio And that's what's shown here in green down here and That's kind of interesting and then if you and then the blue contour is if you add in the lensing data and external data But it's they're rather similar, but a bit squeezed in this direction but either of these two curves so so bring Bring this five-square potential into kind of then disfavored at more than the 95 percent level I mean they're not then it's not five sigma, but it's um, but it's certainly not preferred by the data and What many of us are doing to move beyond Planck are to try and make it a taxion Kind of in this in this kind of level if the signal is there to be seen or at least push down the limits You know down to here or below Because many you know that's obviously of great interest for constraining inflationary models But the current limit is RF less than point oh nine right now there's no evidence that anything is is That we have any fluctuations that don't just obey this power law Gaussian adiabatic description Okay, and that those constraints have really tightened with all the plant data We can look at the fraction of primordial fluctuations that could be non adiabatic So where the fluids at early times don't follow each other at early times as you might expect if they were imprinted by Inflaton fluctuations decaying to form all the fluids in the universe We can characterize that by some parameter alpha that shows the deviation away from adiabaticity and You know that is restricted to be you know less than a Percent or so on even less so with temperature and even less if we include the polarization data which is It's still Early days for that and there's no evidence for non gas entity the constraints on this FNL parameter That described the deviation of a particular kind of deviation from non gas entity Where zero says their Gaussian is now tightly constrained to be Two and a half or different different types of it for different shapes of non gas entity two and a half plus or minus six and All these other two numbers here that show also a consistent with zero so no evidence of that from the CMB and No evidence that the primordial fluctuations deviate from a power law So any any kind of scale dependence of the spectral index is constrained at the kind of 1% level I can't I can't vary more than that and there's no significant kind of features in the spectrum that could say you have some significant Oscillation or some significant boost at some scale that could be indicative of some more unusual model. We're just not seeing it We've learned a lot about neutrinos With plank and with ground-based experiments. We're able to measure the effective number of species Which for us in cosmology could be neutrinos sorry the effective number of relativistic species that could be neutrinos It could mean other things it could be It could be axions it could be anything that produces a relativistic background We would count as being this these relativistic species and So what we do is we assume it's all neutrinos and we constrain the number of them Okay, so that's it and we allow that to be a continuously variable number. We don't just assume it to be integer values And so I'm showing you here the constraints Well, let me just that we just give you the numbers that we have which is the number of neutrinos species or the effective number of species It's now pretty tightly constrained at three point one plus or minus point three or point two at a one sigma level Depending whether or not we include this is plank data, and this includes galaxy position data And this is this is this is really taken a big and this comes from the fact that if you increase neutrinos You basically sort of imagine you're switching radiation from From photons that are coupled to barion's to neutrinos that are not coupled to barion's The sound waves at the early times propagate differently in those two fluids And so neutrinos additionally damp the power spectrum and slightly shift it could the peak positions So getting a measure of the damping tail from plank And then increasingly with polarization you can squeeze you can really limit how much non photon Radiation you have in the universe and this number used you used to be able to have like ten in the CNB Just like five years ago ten was fine, right? So this the fact that it's you know, even four now looks pretty Disfavored so this this shows that number as a function of barion density And this little if I look at any of these combinations of data for is just just does is really on on the edge Now again, you could have some different Different relativistic Species that could contribute a fraction of an N It could be point two point one ten to the minus five and that's still in there But this is that we are squeezing down on the possibilities there And in the next few years that should Be halved or better that uncertainty from from measuring the CNB even better And we can also look at The primordial helium fraction and that looks that's also now much better constrained with like if I look at neutrino mass There are many effects of neutrinos But just the one I want you to sort of have Carry dominantly at the moment because it's the one where we I hope to make best progress is if you switch coal dark matter If you take some of it and make it to be turn it into neutrinos instead Then the universe that has more massive neutrinos has more suppression of clustering of cosmic structure Because those neutrinos even if they're massive they started off relativistic in the early universe They basically behave like radiation because the universe was so hot so a massive neutrino behave like radiation And then like and then like cold up matter later And the fact that it had this epoch of behaving like radiation it free streamed out and it Doesn't cluster like cold up matter. And so you get less clumping of stuff in the universe and less lensing So you tune up the neutrino mass and you push down the lensing power spectrum And there are other threats too, but this is the the main one to think about and that's given us this measurement of This constraint on the neutrino mass The sum of the neutrino masses that's less than point. This is too many decimal places. I'm sorry I did not choose these number of decimal places point six eight right is a All point seven I take carry point seven right that it should be less than point seven EV at 95 percent confidence and if you add in the Galaxy information, then it's less than point two three or point two EV And this is kind of interesting because this is getting into the regime of you know We're getting we think that there should be a signal at point. Oh six EV or greater We have to be careful because we're doing cosmology and this is all indirect measurements. We're not seeing neutrinos We're just measuring their effects cosmologically So we have to be really careful about whether it could be mimicked by other things that we don't understand yet like dark energy And there are some degeneracies with those but they're not really strong And so this is gonna this is this is where in the next ten years. We're you know, we're hoping cosmologically to actually detect neutrino mass Again with with these caveats of being an indirect detection But I think that's we're in we're in an exciting regime Let me just let me just raise a couple of problems or clues and then I'll then I'll stop I'm saying, you know, I giving maybe the a slightly We see nothing in the plank data and another CNB data that says that we need anything It's not just this simple lambda CDM model. It's kind of a bit freakishly good fit That if I if I were to raise any Problems or clues of where we might see something new something some new physics right now if we look at how Galaxies are lensed around large-scale structure called cosmic shear The current measurements of the size of clustering When this this number kind of quantifies size of clustering Okay, this this green contour doesn't quite overlap with this black one and this black one is the plank one and this green one is kind of the one from Lensing of galaxies. They don't quite overlap I think it's more likely that the gap that the measurements of the lens in galaxies is still in early days and might have systematic uncertainties So I don't say that this is I don't think that this is a huge problem, but it's potentially interesting The the peaks of the CNB are a little bit too smeared out if we introduce this strange parameter that I call a Lensing AL and a lensing should be one if the peaks are all completely as I expect and Actually, if I this is a bit too busy a plot But this blue curve here is the preference from the plank temperature data and it prefers a slightly higher value And so it's that the peaks are slightly more smeared out than I'd expect in just lambda CDM But it's not a physical parameter This is not a physical model but you know if one has a physical model that actually does predict a bit more peak smearing then Then there's some room in the data for that and it's large scales The fact this is dip and power at the largest scales is very interesting and I think we have to look in the polarization data and Just keep it in mind that it's there even though maybe it's not hugely significant Okay, so right now the data from plank the CNB data They demand lambda CDM a bit more and more confidence and so It's got a look like little lambda CDM, even if it's not and in particular if it's not inflation It's got a look a lot like inflation If that was the thing that put in the primordial fluctuations And we're busy searching for gravitational waves And I think the neutrino sector is an area that that holds a lot of promise for the next decade in cosmology And I hope we have a lot of interesting further things to say with the next generation of experiments Thank you great overview A quick question Would you like to comment something about the neutrino mass hierarchy? Yeah, I would yeah, so here so so really we're Realistically for a long time now to come we can only measure the total mass So we just sensitive we're slightly sensitive to the to the hierarchy But mainly we sensitive to total mass and so if it but if the higher if it's inverted Then it will be at least point one two EV will be the total mass sum and we'll get there sooner. Yeah So we could detect an inverted hierarchy I think in five years cosmologically with with the caveats, right? So that that will be interesting and you also mentioned degeneracies with other What is the most dominant source? It's a great question So right now me and my PhD student is actually working on a paper to really push, you know explore kind of as many as possible the If you just want to use the cleanest probes so CMB lensing and the Barry and the galaxy positions then they're most degenerate with with curvature and With the dark energy equation of state So if you allow the dark energy equation of states to vary with kind of a two parameter model doubly not w a then then it then it could it could as much as double the error on the neutrino mass and And And those are those the main ones and but there were kind of then there's this then other data should help break this Generacies again, but those are the two main ones because it's like basically Neutrinos suppress clustering But you can change the expansion rate to look a bit like that that's that's that's why those those are degenerates But they're not completely the same When you say that the quadratic model is disfavored that refers only to single-term potentials, that's right. Yeah, sure. Sure, definitely Yeah, okay You're talking about the time frame. You said that in the next decade, we will know the message and neutrino to another money to Using which experiment? Oh, yeah, it's a good question. So so There's two prongs one is improving the CNB lensing and so for CNB lensing we've got what we call Or an DOE speak stage three CNB experiments These are experiments from the ground that will map half the sky from the ground And that's the advanced act telescope in Chile And the South Pole telescope third generation experiment So both of the ones that have already been doing small-scale CNB They are putting on more detectors and covering more sky and that will get us kind of part the way Then the CNB stage four that is plans, but not fully funded yet Which is the natural extension of all these ground-based experiments, which we're saying we're not going to get a new satellite anytime soon what we can do is use or I say that there are there are Strong proposals for large for smaller satellites to measure the larger scales But to map out the whole sky to get neutrino mass you want to get as much the sky as possible at higher sensitivity And so from a mixture of these ground-based telescopes in Chile the South Pole potentially even Greenland will do that at the same time the measurements of galaxy positions will improve with the Daisy Galaxy clustering survey that's starting in a couple of years that moves on from the stone-digital sky survey So it will be the combination of I think the thing that we can actually make a detection from will be what we call CNB stage four plus The Daisy galaxy clustering survey so, yeah Okay, I'm afraid we have to move on and you can talk to speak it later. Yeah