- Next, we have Steve Howell. Hi, Steve.

- [Steve] Hi, how are you?

- Good .

- [Steve] Okay, great. Thanks, everyone, thanks ExoPAG for giving me some time to talk to you about a program we've been running. We can go to the next slide, Jennifer. I wanna start by sort of motivating this, it's a program of high resolution imaging we've been doing for over a decade now, starting with Kepler and continuing now into the TESS Extended Mission and beyond. But here's a good example just to let you know why we need something like this, this is the simplest case and why everybody thinks it's important, but we'll see in a minute there are many reasons. So this is a star from the Kepler Mission, Kepler-1002, a planet. And on the top row of this figure, you can see this field of view image by Kepler in the center. If TESS would image this field of view, which it finally didn't, Kepler field image by TESS, on the left, NAC limited image, that is just a sort of standard ground-based CCD image of the star. And Kepler-1002 is the star in the middle of the two frames on the right and sort of all blended as a mass in the TESS image on the left. On the bottom row are much different scaled images of Kepler-1002. The two on the left are images done with speckle interferometry. The image on the bottom right is an infrared AO image. And you notice that in these images, you see that Kepler-1002 is actually a pair of stars that are quite close together, something like 0.4 arc seconds apart. So the top row doesn't know this, the bottom row of course does know this. And the reason it's important in this simple example is that if this was a single star, which is assumed in the Kepler planet catalog, this planet has a radius of 1.4 earth radii, which is kind of an interesting radius. It's at the boundary of sort of rocky and non rocky and falls near the Fulton gap. But of course the star really is a binary star here, and so if the planet orbits the brighter or the primary of the stars, its radius is really 1.8 earth radii. And if the planet happens to orbit the fainter or the secondary star here, its radius is three and a half earth radii. And in fact, it's not always so simple to tell, except in very unusual cases, which star the planet actually does orbit. We mostly take it for belief that it orbits the primary because if the secondary star is much fainter, then it's harder to see a transit there. And so we just assume it's the primary but there's no guarantee that it is. There are a few cases Kepler-13, for example, where we've done a lot of work to show which star it orbits. Next slide, please, Jennifer. So this is the sort of initial reason why you think about doing this. And that is that if you have a planet transit observation and the planet orbits a star, then the depth of the transit gives you the ratio of the size of the planet to the star. But if there's another star there, and this could be a true bound companion, or this could be a line of sight star that just happens to be very close in line of sight, then the true depth of the transit is not seen. You see a shallower observed depth. And there's a nice paper that David Ciardi led, and there was this relationship here where if you know about the two stars or three stars, in case of some triple systems, you can adjust the observed depth to the true depth and if there is another star, the planet always becomes larger because the true depth is always deeper. So this was sort of the idea of starting this way back in time. But if we go on to the next slide, we'll see that there are many other reasons why we need to care about this. The largest one might be that something like half of F, G, and K stars are binary or multiple star systems. So this means that something like half the planet you know, might not really be the size you think they are, unless you fully understand the scene around that star. And I don't mean stars that are four arc seconds away, I mean stars that are very, very close, well within half an arc second, I don't wanna read all these, these slides will be posted, but in all of these, there's something that we need to do to mitigate these risks. And they all involve some high resolution image knowing about the scene of that star or not, and provide characterization for both the exoplanet properties, but also for the star properties that it orbits. Next slide, please. So if you don't know what speckle interferometry is, here's a little visual primer. We take images such as those on the left. For example, we take thousands and thousands of 60 millisecond images, and each of those looks like the image on the left filled with a bunch of little focused point sources called speckles. And the image is messed up because of the atmosphere, of course, which is why we're doing this. And in each of those images, you actually get a frozen piece of the atmosphere because your frame is so short, then you can see that atmosphere distortions to your wavefront in each of those speckles. If you take all those images and just co add them, which is the equivalent of an integrated image, you would see the image in the middle, which gives you a single blob that say something like one arc second across. But that of course isn't really the true image here, the true image is what you get as represented here on the right in this case, that you can then remove the effects of the atmosphere mathematically using 48 transform techniques. And that allows you to get this reconstructed image on the right side here, showing a binary star, in this case also about 0.4 arc seconds apart. On the bottom, I show you what you actually would see if you co-ad these in Fourier space. If there are two stars, you actually see a fringe pattern. That's why it's an interferon metric technique. And that fringe pattern allows you to get all the information on the stars, separation, their position angle, and the difference in magnitude between the two stars. Next slide, please, Jennifer. I just wanted to show you briefly the instruments that we've built over the last three years I guess now. These were funded by NASA through the Exoplanet Program office and NASA headquarters. The instruments in this case unlike infrared AO images are quite simple. Infrared AO systems involve sometimes lasers, maybe always laser, they involve deformable mirrors, wavefront sensors, all kinds of hardware, and they make the correction in hardware on-the-fly. Speckle interferometry basically takes very high speed images and makes the corrections after the fact mathematically in Fourier space. Each of our instruments are essentially identical in the hardware they use and the software interface. There's one at the WIYN 3.5 meter telescope at Kitt Peak, and one each at the two Gemini eight-meter telescopes in Hawaii and in Chile. They both obtain simultaneous images in two colors, each with a filter wheel in front of them, using and or EMCCD cameras. And we have sort of a red and a blue channel and some filters for you to choose from in each of those wavelengths. Next slide, please. Thanks, so the reason that I'm talking here at the ExoPAG is if you don't know this over the past, about 10 years or more, this is a community service program. We observe about 100 nights a year on these three telescopes and all of our targets come from the community. They come from the community through the Kepler K2, and now TESS program offices as a list of POIs that are prioritized by community members and by the TESS project. We get objects given to us by community members that need one observation or need more than one observation of targets to be done. Maybe they're writing a paper on an exoplanet, maybe they're interested in old stars that have hot Jupiters, and things like that. And we provide all of this data or reduced data as well as the raw data, the fully reduced data, at the NExScI-EXOFOP without any proprietary period. So there are three ways one can get data. You can go to the NExScI archive and get any of the data we've taken, the raw data, if you wish, I don't know why you'd want it, but if you're crazy, you can get that. Or you can get the fully reduced data for any objects that we've done with the Kepler, K2, or TESS. You can also send requests, then I have a call in the NExScI, exoplanet archives has a call before every observing run, they ask the community for targets. And you can also write your own proposals to use these instruments. They are not PI instruments, they are completely open to the community and anybody can write a proposal to use them through the NOIRLab "open skies" proposals to either Gemini or the WIYN telescope. Next slide, please.

- [Woman] You have about 10 minutes remaining, including questions.

- [Steve] Thank you. We heard earlier today about the exoplanet science gaps. So again, I won't read these, but just to show here that we do directly address five of the science gaps that are in the Exoplanet Program office science gap, and we also enable through addressing these five gaps, we enable four other gaps. So I think there's only two that we missed. So it's a pretty good program for covering lots of bases to help out with a lot of exoplanet research, both from the space and from the ground. Next slide, please. To show you our progress on TESS TOIs so far. On the left shows a histogram of the number of TESS TOIs we've observed for the prime mission. So the red histogram are planet radius TESS TOIs, and the green bars are how many of those we've observed. And on the right is our initial start at the extended mission. I think this chart is maybe two months old now or something. And so we have more, in fact, we were observing last night, we have more observing time in June and July as well this year. So we're making progress. We haven't done everyone yet, but give us a few more years and maybe we'll catch up on all of these. Next slide, please. Just to show you an example here of what kinds of contrast and resolutions we can get. This is from a recent paper which made observations about 100 TESS objects using the Gemini telescopes. And this shows our observations made at 832 nanometers, so the red part of the optical spectrum. The vertical dashed line is the diffraction limit of the eight-meter telescope. In this case, it's about 20 some milli arc seconds. And the red line with the kind of shaded band is our mean contrast curve covering these 100 objects. We just put them all together and made a contrast curve. If the seeing is bad or it's cloudy, just like with every other observation, if the sky background is bright, you can change your contrast curve a little bit up and down within that shaded area. And the blue dots here show for this particular set of objects, companions that were detected around TESS TOIs. So you see a couple of things here. There are companions at sort of everywhere here within 1.2 arc seconds, all the way down to the diffraction limit. And also there are companions here that look funny because they're better than our contrast limit, but of course as I mentioned, that's our main contrast limit. So if you have a brighter star or exquisite scene conditions, which you often get at Gemini, you can have observations that really push that elbow much further to the left. So just an example of what sort of space we can observe within. Next slide, please. I wanna just review a few of the findings that have come out of this project over the last few years. All of these are available at the very last slide of this, which I won't cover, but it will be in the presentation, gives a reference list where all this material is from. Mostly published papers. So we've found that the TESS TOIs, on the left here you see a mass ratio diagram, they tend to have mass ratios that agree with those so-called field binaries. And of course, field binaries is a funny term because we know if binary stars here are TOIs, we believe they have a planet, but if we don't see a planet, it doesn't of course mean the star doesn't have a planet, but it means we don't necessarily see that planet. So those would be the field binaries. And as such, we have binary stars that mostly have roughly equal mass ratio. And then they're kind of a constant mass ratio, a number at maybe 40 or 50% of that value out through all the mass ratios. And by the time we get to a mass ratio in this diagram of about 0.4 or 0.3, that fall off is due almost certainly to sensitivity in seeing much fainter companions. For example, if you have a G-Star and an M-8 companion, that would be a target that infrared AO might pick up, but that's very difficult for an optical imager to see because the Delta magnitude there is very large. On the right we look at the period distribution, the orbital period distributions of TESS TOIs that are binaries. And the purple histogram on there is again, the field stars where the two colored histograms are from our studies at the WIYN telescope and the Gemini telescope. And they show as this paper discusses in detail that the average separation of binary stars that have planets is slightly larger at about 100 AUs than the average separation of stars that are not known to have planets, which is about 40 astronomical units. Next slide, please, Jennifer. Another thing that we've been doing over decades for a couple of will take us maybe three or four years with TESS, is do astrometry of the orbital pairs. And this is an example from Kepler target KLI 270. I forget what the Kepler planet is here. But on the last, we show the relative motion of the primary star on the secondary star through space over time from 2011 to 2017. And you can see that the stars track together through space quite nicely. On the right hand side, we show relative astrometry of putting the primary star all stacked up in the center, and then looking at the motion of the secondary star. So this is starting to see the orbital motion of the two stars in space. And in this particular case, it's very interesting because it looks like the secondary star is gonna diagonally cut through across the primary star, which would be in the same orbital plane as the planet in this system. We assume the planet here orbits the primary. We don't know if this is true for every star. Like I said, the Kepler stars are farther away. Their separations that we see are much larger than for TESS. And so for Kepler, it'll be decades and decades before we get pretty good orbits. For TESS we can do it in a few years. Next slide, please. And finally, this recent result, this just as is in press from a postdoc at NASA Ames, Katie Lester, where we have plotted on the left, the planet radius versus the planet period for planets that are seen in single stars, as in the blue dots, and planets that we find what we know, the host stars are binaries. And the little vertical errors are where we make corrections to the planet radii because of the third light contamination. And you can see, in some cases it moved the planet radii quite a bit, in other cases, it didn't move it very much at all. But what you notice dramatically from this diagram is that very small planets, things that are about two earth radii or smaller are not seen in binary star systems. And this is no doubt due to the contamination of the transit light, and the transit becomes too shallow to be measured in that case. So these systems, while they no doubt exist, show that we have a large bias here for finding small planets. We won't find them in binary star systems, which again is something like half of all the stars. So this is a big effect that has to be taken into account if you care about occurrence rates, for example. Next slide, Jennifer. And now I just wanna show two quick slides to highlight some other things our instruments are highly useful for. We're starting to do tests on looking at what we call widefield speckle interferometry, where we move away from point sources and actually look at things that have two dimensional sizes to them. These are some initial tests with the planets Jupiter and Saturn, because they're both pretty and cool and they're also scientifically of high interest. On the right is the asteroid face on which we observed, I think last December it was, or maybe the December before that, I guess 2017. Wow, quite a while ago. Time just flies when you observe 100 nights a year. And you can see that the asteroid here is quite resolved into an elongated shape. And Jennifer, the next slide, please. Just some other things we've been doing with our instruments. We've been looking at microlens targets to do followup of microlensing events, where both the lens and the source can be separated, and each of the individual stars can have their identities known in terms of color and type. Our cameras because they're high-speed LiDAR cameras and the electron multiplying properties of them, we can do very faint objects at very high speed, fractions of seconds, to many minutes, depending on what speed you need. And we do two colors simultaneously. This is an example of a two-color light curve of an eclipsing pair of white dwarfs. And you can see quite well here that the eclipse in the blue is much deeper than the eclipse in the red. And you can also see the secondary eclipses there at the end of the plot. And on the right here, we just show a nova shell that we imaged just a little while ago for Nova Carina V906 Car. And you see the nova shell is quite resolved there with a bright star in the middle. And I think my last slide is, next, Jennifer. Just the summary which I'll leave up and take any questions if you like. As I mentioned, the slide after this, which we don't need to look at, but it will be in with the slide deck here, has all the references to the material that I talked about. Thanks.

- Okay, great. Thank you, Steve. I think we're a little over. There is one question in the chat though, if you wanna comment briefly on the relevant performance between speckle imaging versus AO.

- [Steve] Yeah, sure. The performance, it's not a real fair comparison to try to say one is better than the other. AO gives you wider fields of view. Our field of view is 1.2 arc seconds for point source speckle. We go a little wider in our new sort of widefield mode. But once you get beyond about one arc second, you lose the spectral coherence between objects, and things start becoming de-correlated. So you really can't do a super-great speckle reconstruction and hit the diffraction limit if you go out past about 1.2 arc seconds. But as I mentioned earlier, the big advantage of infrared AO is that faint red stars, we won't be able to see. Things like brown dwarfs we can't detect. They're just too faintly optical. They have Delta magnitudes of 10 or 12 or 15 or something like that. But if you move out to the K band with the infrared AO, then very late M stars become quite easy to see in the AO. We have about twice the resolution in the optical because the wavelengths are three or four times shorter. So it's sort of, they both are really lovely to have together. You get a little different resolution, a little different discovery space.

- Awesome. Thank you so much. And yeah, we have references. So Michael Meyer asked about the Lester et al, it looks like there's a press release coming, so that's the thing.

- [Steve] There will be on Monday, so look forward to that, yeah.

- Perfect. Awesome, thank you so much for this great service.

- [Steve] Great, thanks.