 Thanks a lot, Dina, for giving me a chance to show you guys what we're up to at the APS. So this beam line started to be developed when the the MBA upgrade to the APS was announced in about 2015. And actually most of the proposal was written by Ian McNulty and I when he was still at the APS. And this is sort of the extended list of contributors to the original science proposal that formed the basis for the beam line. These days actually I wanted to point this out too. So this is kind of a good example of what we're going to be able to do at a new beam line. So this is a data set measured at 34 IDC at the APS. And the nice thing about this is that the crystal size was very well matched to our X-ray spot size. And so we measured this data set in about 25 minutes at 34 IDC. And this is a little gold crystal in the beam. This corresponds to like a 15 second measurement at the upgrade beam line. So it's going to be a little bit of a horrifying flood of data that's going to come out of this thing. How it might come out of this thing, assuming it works. This is the people that are sort of involved in the day-to-day development of the beam line now. Once you organize are the people that come up with the dumb ideas. The engineering team is the group that kind of looks at us cross-eyed and says well maybe. And then the optics team is by far the most productive in recent time. And it's being led by Shamboshi. He's been developing the optical systems for all of the upgrade beam lines. And he's ridiculously productive. I got to admit that probably a third, maybe a half of what I'm going to show you guys today is a direct result of his work. So he's been extremely productive. So this is the kind of work we're doing today. And I think people are familiar with these types of experiments. We're doing sort of operando nanoscale imaging. We're doing it at 34 IDC. And the reason that we can sort of do these kinds of experiments is that we actually don't try very hard at the beam line. We have very modest optics. The advantage is we have a dedicated instrument. We very seldom take anything apart or change anything. And we can focus on sort of high Z materials where you get a good signal. We look at samples that are sort of compatible with our spot sizes, our spot to sort of a half micron on a good day. So 300 and 400 nanometer gold and platinum and battery particles fit pretty well in that beam. And we're also looking at dynamics that are very compatible with our measurement time. So like 10 to 20 minute measurements and things happening in crystals on the 10 to 20 nanominute or 10 to 20 minute time scale are the types of experiments you can really do well at our beam line. But we started forming a beam line proposal back in 2015 or 16 to start maybe pushing this a little bit further. And the idea was to start looking at things where the sort of nanoscale and sub nanoscale becomes the dominating sort of structural characteristics and trying to image those things at those resolutions. And our science case was built around these sort of topics. This was a contribution from David Tede in our chemistry division where he's looking at these or growing these synthetic leaf materials is what he calls them. So they're basically materials that will split water in the hydrogen and oxygen with just the input of sunlight. And these things are very efficient, but they don't really have a good grasp of the sort of structural characteristics of the active components in these materials. And at the time, there were some questions about what how these layered materials stacked up and sort of what sort of defects appeared in these layers. They actually did answer this question recently. There was a paper published last year where they finally were able to untangle the sort of structure of these these layered oxides that form these artificial leaves. We're looking at a lot of catalysis experiments at the beam line even today. And you could start zooming in on the sort of nanoscale and looking at actual active catalytic sites on even smaller objects and imaging the response of these catalysis crystals to the input reactants. You can start looking at structural materials and trying to correlate strain imaged in one way with defects on green boundaries all at the same time if you can start the image on the nanoscale. And then also metallic glasses were a topic that one of our contributors brought up. These things are extremely strong and when they fail, they fail catastrophically. And those failures are the sort of collective motions of groups of atoms. And if you could try to understand how these atoms start to move around before and after failure, you could start to have some impact on the actual application of these types of materials. So this was sort of the science case that we developed these types of ideas. And it led to the sort of guiding force of the beam line specification. My screen is moving extremely slowly. Let me see if I can solve that problem really quickly. And this is supposed to be up here. There we go. So we came up with this kind of specification for the beam line. This is our basic, our advertising slide when we first finished the proposal to try to get the beam line sort of built. We had this idea that we would work in the sort of 5 to 25 kilovolt range because that sort of hits a lot of capabilities as far as penetration and high scattering at lower energy. We live currently at a beam line with a canted sector and we have a Laue diffraction microscope as our partner at the beam line. So we've kind of had this thinking for the last, 20 years that being able to switch the beam from pink beam to mono beam and do Laue diffraction microscopy is a very clever way to get orientations of crystalline lattices. And so we actually started developing that with a partner user proposal at Los Alamos and Brigham Young and Carnegie Mellon University in the last few years to have a removable monochromator that we could pull out of the beam and do Laue diffraction on a small crystal and get an orientation matrix from it and use that to then go dial up Bragg peaks with a monochromatic beam and do coherent diffraction on them. Tasos Anastasios-Peteris publishes paper last year where we showed that we could do this in situ with this monochromator that now actually exists at 34 ID. So that's still one of our sort of driving goals in the development of the beam line is this ability to switch between pink and mono beams. We also wanted to include a zoomable focusing system into the beam line with this idea that we'd be able to change the spot size from sort of 50 nanometer scale to a couple of micrometer scale without moving the position of that focus and I'm going to talk a lot about this sort of technological development that we've been working on for the last few years at APS in the next few slides. And then this was our selling point that we're going to deliver extremely high resolution imaging sort of reaching the nanometer scale via coherent diffraction. We were also going to maintain this ability to do operando and in situ experiments with very high resolution imaging and our goal was to have a sort of five centimeter working distance in our focusing optics. And so that gives us space for doing in situ types of experiments. So these have been our sort of motivating goals in the development of the beam line and we're sort of sticking to these goals reasonably well. I wanted to show a little bit, I can't believe how slow my screen is going, show a little bit about some simulations I've been doing in the last couple of months to try to understand the type of data that we'd be able to measure, assuming we can put the photons on the sample and collect them. I've been working with some atomistic models that we can generate and relax with molecular dynamics. And then I can just do the lattice sum on these atomistic models and form diffraction patterns like this. And so I started getting asked questions by our engineering group, you know, if you have this huge detector that's say 4k by 4k, can you actually get a signal across that whole detector? And, well, that's an important question to answer. So I started running these simulations, again, just using PyNICS from Vincent at DSRF, summing these atomistic models and then just doing the sort of Thomson scattering corrections, nothing too sophisticated, just using E and M and H and C. And just rescaling to get a sense of what kind of felicon flux we would have in the diffraction patterns from such a little crystal. And it was actually surprising how full the detector was with signal. So I was simulating here from, so this is a crystal where the C-axis is pointing up and I was simulating here from L equals 1.5 to 4.5. And just doing, you know, I was using extra utilities then to use sort of a realistic detector geometry. And I was just simulating the Hkl values for each frame and looking at what the intensity was across the detector as a function of this scan from one point in reciprocal space to the other. And it was kind of surprising. So in the O02 break peak, which is the first one here, you get about 2 million photons per second. The O04 is coming out to about 0.5 million photons per second. And all the way out to the edge of the detector here, these brighter signals are sort of, you know, 10, 20, 30, 50 photons per second out on the edge of the detector. So it seems like, you know, this is just a 50 second simulation in the upgrade, assuming we get about 10 to the 12 photons per second on the sample. You know, it seems like this becomes a fairly reasonable kind of data set to expect. Of course, you know, it won't be exactly this because I'm not including everything, but at least it's an estimate of the type of signal we would see. So it's really looking encouraging that, you know, at least with a radiation hard sample that could handle this type of dose that we'll be able to measure sort of really extensive volumes of reciprocal space. And we've just started wondering how you face something like this. And then what that image even means is even a question. I wanted to put a little plug in for this website that's been developed by a graduate student at the University of Illinois in Chicago, Will Judge. He's been developing a really nice set of instructions for how to put together an atomistic model and how to use pinix to generate a coherent diffraction pattern from it. And he's even now starting to include documentation for using lamps, which is a molecular dynamics simulation package to actually relax this thing using an interatomic potential and get sort of a realistic sort of strain profile inside of the little crystal. So go to this documentation and take a look at it. Even I can follow the steps and do this, go from a little crystal to a coherent diffraction pattern. And I've been thinking about adding some instructions here for using extra utilities to simulate a sort of realistic diffractometer. And then maybe even include a module for doing the Thompson scattering and could even start to include other simulation modes. So a little bit back to the beamline now. So here's the layout. So originally we had hoped that we could have a green field where we would be alone and we could really do a beamline with no compromises and we were immediately told you're going to have to compromise. So we're going to be living again at sector 34 where we currently live. We're still going to be a canted sector. And so our sister instrument, the Laue diffraction microscope is being rebranded as 3DMN. And then we're going to be on the other branch and we're going to live down here at the end of the sector now in the atomic beamline. So our first compromise is we can only have one undulator. And so our undulator that we chose was a revolver. So we're going to have a, you know, finally the APS upgrades going to allow these revolvers that have been at the ESRF now for a very long time. The undulators we chose were a 2.5 centimeter period and a 2.1 centimeter period that will give us a sort of composite tuning curve that looks something like this across the entire bandwidth of the beamline. So we'll probably spend most of our time, you know, in this band between sort of 10 and 15 kilovolts. But, you know, you don't know where things are going to go. We might end up spending more time and much higher energy. So we're going to have a mirror here, a white beam mirror that will give separation between the beamlines. And we'll have a monochromature which is actually removable to allow the pink beam experiments where we've been doing a ton of work on the development of the Zoom optical system. And then we'll have our experimental instrumentation at the end. So I was going to go through these things a little bit and show you where our current state is. This is the floor plan. If you've ever been to 34 ID, this is what it currently looks like on the top here. We do our experiments in the C station here where our diffractometer lives. And then in the upgrade, we're going to build a new instrument enclosure at the end of the sector. And our KB mirror systems, the Zoom KB system will then now start in the C station. And the first mirror pair will be there. The second mirror pair will be down at the experiment. And then the C station is going to turn into an optical enclosure. There'll be both monochromators in here for each of the branches. And like I said, our first KB mirror pair. And then our white beam mirror will be at nominally three milliradians and live up in the A station here, the first enclosure. So one of the first things we had to look at was this idea of a removable monochromature. So this is the sort of schematic of the beam line. The white beam is reflected into a pink beam, comes out to the monochromator. We're going to have a very small offset of about a millimeter in this monochromature. And then we'll have our focusing optics at the end of the beam line. And then to do these Laue diffraction experiments, we have to move the monochromator out of the beam. But that also then shifts the, shifts the optical axis of the beam. So our solution to that was then just to have a specification for the white beam mirror that we would move it upstream and then move basically the pink beam axis into the zoom KB axis for the monochromatic beam. So this additional motion was included in our statement of work for our white beam mirror. And I think we're very close to awarding a contract for the white beam mirror system here. And it should be not so bad to add this additional motion. I think that was all I wanted to say about that. The way this is currently done, so even at 34 IDC where we now have a small offset monochromator, we actually heat the second crystal. And so when you switch from white beam or from monochromatic beam to pink beam mode, you're basically using a different source angle. And so by changing the angle of this first crystal by heating it, you actually kind of bring the monochromatic beam onto the same axis as the pink beam. And, and that's a trick that works when your monochromators running at room temperature. But our upgrade monos are looking like they're going to need liquid nitrogen cooling. And so when the first crystal is so cold, the temperature difference required between the two crystals is actually quite large. And so we had to start giving up on this idea of heating the second crystal to maintain this optical axis. And we switched to this idea of just moving the white beam mirror upstream to maintain the optical axis into the final focusing optics. So the ZoomKV system was something that was proposed in 2013 in this publication. And like I said, we started developing this sort of beamline concept in about 2015, 2016. And by 2016, they had actually demonstrated this at Osaka University with a bimorph mirror system. So the idea here is that you can use two pairs of mirrors, you can use two KB mirrors, and you can make them deformable. And you can move this first focus forward and backward with these deformable mirrors. And as a result, you can change the effective numerical aperture of the mirror system. So if you want a large spot size, you deform the first mirror, you move the first focus close to the second mirror, and you get a very small numerical aperture out of the second mirror. So you get a large beam. And then if you want to make a smaller beam, you just deform the first mirror, you pull that first focus further away from the second mirror, you now start to increase the numerical aperture of the second mirror. And you get a smaller spot. And if you tune this all up properly, you can do all this without moving the position of the final focal spot. And so this was a really attractive idea to us. And so we started including this in our beamline specification. And then this was actually demonstrated in this paper in 2016, where they used a pair of these deformable mirrors. And in one dimension, we're able to tune the spot size. So this gives you a lot of capabilities. It allows you to sort of tune the spot size to the sample size for a sort of Bragg CDI experiment. It allows you to tune the spot size for typography, sort of field of view experiments versus like imaging rate and resolution and damage. This was something I proposed early on that you could also tune the transmission efficiency of the KV mirrors. So in this high NA mode, this high numerical aperture mode, you're effectively overfilling the second mirror. And so when you're working at lower X-ray energies, in principle, if this was a fully tunable system, you could increase the angle of the second mirror, and you could retune the shape of it and increase the sort of transmission efficiency of the second mirror because you're not going to lose anything to the higher angle. This was shot down pretty quickly by Shambo. He said that the ability to reach the types of curvatures you would need to change the incidence angle of the second mirror set just wasn't going to work. But one of my other thoughts was we could also tune the working distance. If we've got this sort of fully tunable mirror system, I could move the mirror system further away from the center of the diffractometer for an experiment where I needed a larger working distance. And that's only going to be at the cost of the sort of smaller spot size you can get. And fortunately, that idea stuck around and we're not starting to specify a zoom KB system with a tunable working distance as well as a tunable spot size. So I'll show you a little bit about the current status of this design. What's really allowed us to keep this sort of tunable working distance idea is a very recent development by Osaka University. It's a hybrid bimorph mirror system. So this is a mirror that's now not only adjustable by piezo strips on the side of the mirror to tune the shape of it, it's also got an integrated bender system. So they can mechanically bend the system and then you can do a very fine tuning of the shape of the mirror with this integrated piezo system. So we're now starting to incorporate these types of mirrors into the zoom system. And we've got a bit of a collaboration going with Osaka and JTEC now to sort of characterize mirrors like this. Oh yeah, I probably should have said that that the point of these bimorph systems is that what they are is it's a piezo material that is bonded to the substrate, to the silicon substrate, and then you can apply a voltage to each of these electrodes and very finely tune the surface shape of the mirror along the length of the mirror. And we actually have one of these at APS now and we've been playing with it for a couple of years and getting used to how it works and I'll show you some of the results from that. So let's get a little bit into the specification of this mirror system. So this is the sort of consequence of going to a variable working distance mirror system and the consequence, the main driving difference here is that both of the final focusing mirrors have to be this hybrid mirror design. So it's this hybrid bender and bimorph mirror. But if we have a fixed working distance, the only one that really needs to be a hybrid mirror is the first one and the second one can actually just be a sort of regular bimorph. And the difference is that the regular bimorph is shorter. You don't need room for the bender so it can be a slightly shorter mirror. If you have to add a bender, the total mirror length grows by quite a lot. And so as a result, all of the d-mags change because this mirror is now longer and its distance from the first mirror is greater. And so that has an impact on the smallest spot sizes you can get. And I was actually quite surprised how much this changes as a function of x-ray energy. So it turns out now that with this sort of variable working distance mirror, we can get down to around 50 nanometers at 8 kilovolts or 9 kilovolts. But it's at the cost of a lot of flux. If we really want to get back flux, we actually have to pay for that in minimum spot size. So to get the sort of incident flux up to around 10 to the 12 photons per second, the spot size kind of grows to around 17 nanometers. And so we have this now the sort of concept of the high flux condition and the sort of low flux condition. And then at about 12 kilovolts, actually everything sort of levels out. We can now start to reach the sort of 50 nanometer spot size as a function of energy and maintain the sort of 10 to the 11, 10 to the 12 kind of photon fluxes onto the sample. So this was the trade-off between going with a fixed working distance or a variable working distance. And it's just something about this variable working distance that everybody thinks is really cool. So we've really decided to pursue this as the sort of target spec for the beam line. Here's some of the technology development we've been doing in recent years to sort of be able to do this and have some confidence that we're going to be able to do this. We got an Argon LDRD funded project in 2018 to develop the capability of doing a zoom optic. And at the same time that this LDRD was rewarded, this stands for Laboratory Directors R&D Project. At the same time that this LDRD was awarded, there was a DOE funded project across multiple institutions to develop wavefront preserving mirrors. And the Argon contribution to this was a wavefront sensor. So they're able to image a wavefront to very high accuracy. And so now you can kind of see that this is all coming together, that we can actually look at the wavefront coming off of a deformable mirror and tune that wavefront with a very high level of fidelity using these wavefront sensors. And they were developing both a transmission mode one and a diamond sort of beam splitter mode one. This diamond beam splitter one has actually been, I think, kind of abandoned because the phase structure introduced by the diamond was quite severe. But anyway, I'm going to show you a little bit about the development of this mirror development we've been doing under this LDRD. At the same time that this LDRD was awarded, Deming Xu, who's one of our sort of extremely capable engineers at APS was developing a bender system using his little constrained flexure benders here. So he had this idea that you could actually mount a mirror onto these constrained benders and then using two motors on the end, you would be able to get a very high fidelity ellipse if you design this whole thing correctly. And that's actually what he was finding in FEA simulation, that over 50% of the length of the mirror, the sort of middle portion here, you're getting a sort of 5 nanometer tolerance on the ellipse. And that was just in FEA. And then over the entire length of the mirror, there was sort of a 9 nanometer tolerance in a perfect ellipse if you design this thing very carefully. And so fortunately, he had already done this design and it was basically sitting in the can waiting for a lot of money to be thrown at it to build it. And our LDRD project got a surprisingly healthy funding. And so we developed a system here that would not only have his constrained flexure bending here, it would also have an array of capacitive sensors on the bottom of it that would actually be able to look at the metrology or look at the shape of the mirror in real time and allow us to then tune the shape of the mirror for a Zoom KB system. The idea was that we could use these types of mirrors as the first mirror pair, because you don't necessarily need a perfect first focus. You can deal with aberrations in this focus because you can correct it at the end with this very high fidelity bimorph system and the second mirror pair. And so we went about building this thing, and this is what it ended up looking like. So there's the benders on the end, there's the mirror flat on the top, and then here's this array of capacitive sensors. I think there's 23 of them along the bottom of the mirror. And then there's one on each end looking at where the bender sort of paddle is, the position of this bender. And actually this turned into a bit of an engineering project in its own right, because nobody had ever done a metal coating on both sides of a flat of a bender mirror. So the top is coated in platinum and the bottom is coated in gold to ground the capacitive sensors. And that even turned into a bit of an engineering project to develop the holders for this mirror so that you could coat both sides and then actually have leads on the end so you could attach the grounding wires. After we produced this thing, we took it to the the the Metrology Lab at APS and we used the Long Trace Profilimeter to now come up with a correlation between what the capacitive sensors on the back of the mirror were measuring and what the actual optical surface shape was. And this allowed us to develop a calibration curve between the the cap sensors and the optical surface. And so here's our here's our first measurement of this where we did the scans with the LTP and we read out the cap sensors. You can see a few of them were actually not working at the time. And so we were able to develop this sort of calibration curve knowing that if the cap sensors were reading out a given a given set, then we would have given an optical surface or given surface profile. It turns out we actually we actually did this wrong. And it took us a while to realize that we didn't have the the amplifier set up correctly on the cap sensors. And so we didn't have the maximum sensitivity. And it turned turned into quite a quite a severe error in the calibration of these cap sensors. And we didn't know it at the time. And we just kept moving forward with testing this thing. And we were going to go back and recalibrate these things right when the pandemic started. And suddenly we weren't allowed to go to the lab. So we took it to 1pm to start learning how to use the thing. So here we have the the bender mirror and a helium box. And then downstream of this we have this grading interferometer system to actually do the wavefront measurement. And here's some measurements that came out of this beam time. So this wavefront is actually this this wavefront measurement here is actually able to give us the sort of response function of the of the mirror. So we're able to see if we move the if we move motor one the the one in blue here if we move motor one a certain distance and bend the mirror a certain amount. This is the sort of change we get in the in the response of the wavefront to a single micron movement of that mirror. And so you can build up a whole series of these calibration curves for for different positions of these two pico motors that bend the mirror and and develop this entire sort of table of response functions that can allow you to then tune in any sort of shape that you want. So we said about trying to tune one and and here's the sort of result of this. So so we started out with a wavefront measurement like this the black curve. And then we adjusted the two two motors to to try to get to a target wavefront measured in the in the wavefront sensor. And what we found was that we measured this the sort of blue curve here or the red curve. And and here's the prediction error. So this is actually quite severe. This you know you can't deal with sort of 20 nanometer or 10 nanometer wavefront errors. And and we weren't searching what was what was causing this until we started to realize that our our cap sensors were actually not well not well calibrated. And that ended up being the cause of these types of errors which are sort of endemic in all of our our work now on this system. But like I said we think we can fix this because we can do a better a better job of the calibration of the of the cap sensors. Masrafi was a postdoc working with us on this project for for about a year. And he said about playing with some machine learning for for being able to tune this because again if you have a whole set of these calibration curves for the wavefront to motor position, you now need to be able to to predict what you need for for a given a given wavefront. And so he worked on a set here now where where he did a bunch of measurements. So he read out the cap sensors for a whole bunch of different values of the force motors on the ends of the motor on the ends of the mirror. Built this whole sort of tuning tuning set that he could then use to train a machine learning algorithm or a neural network so you could feed in the cap sensor values or the mirror profile that you want and out would come the two values you need to set the the force motors to to to get a specific or a specified mirror shape. And he showed that this was actually producing a result that was again very similar to a sort of a search algorithm or an interpolation algorithm from the from the sort of composite tuning curve set that we had. So he was able to show that his neural network was producing very similar results. But again this is very bad error because of our our cap sensor calibration. So so we anticipate that once we improve the the cap sensor calibration both of these will be much more accurate. So for it we'll be able to say we need this shape of mirror and we'll be able to pull out you know what you have to set the motors to to get that shape of mirror with a fairly high fidelity. So that's the current state of of the sort of bender system that we've been using and developing as our first mirror. We've also spent a lot of time with this this bimorph. This put bought under this again this sort of multi-institution wavefront preserving mirror project. So it's actually been getting quite a lot of use in these two different projects. This is what the mirror looks like. There's the piezo with the electrodes on the edges there. It sits in this box that was delivered by JTEC. And then here's the electrodes that apply the voltages to the the individual individual electrodes. And then there's the mirror surface. So our plan for for a mirror like this is is we also want to be able to measure the in situ sort of metrology. We want to be able to measure the shape of the mirror and and ensure that we're tuning up the the shape that we need for for a given spot size. And our plan here is to integrate a set of interferometers across the top of the mirror here to actually look at the optical surface of the of the mirror. And and Deming had actually worked out an entire mounting system for I think I think it was like nine or 10 adto cube interferometers that would be arrayed across the top surface to read out the profile. And all of that stuff was was delivered right when the pandemic started. So we haven't even managed to do that mounting and test the system yet. But anyway, that's our plan for these for these bimorphs is to look at the top surface then with with interferometers so that we can you know do real sort of real time metrology and and maintain the shape that we need for a given a given focus. So we did eventually get around to putting the two mirrors together and do a sort of 1D zoom demonstration. This was how we worked. We first put the the bender into the beam and we use the interferor the wavefront sensor the grading and detector system to tune up a set of shapes in the first mirror system. So that's what you see here. We were optimizing the the shape of the mirror to get as close to the ideal shape as we needed for a given spot size. And then and then we would put the second mirror in this bimorph and then tune up the whole system. So we would then look at the the focus or the beam coming off of both mirrors and and correct for a given a given system and so or a given spot size. And so this was actually kind of an iterative process because we haven't really gotten the whole the whole sort of system working yet with you know being able to dial up a given shape on here due to our our cap sensor calibration area. And we also don't have interferometry yet on the bimorph. So we have to kind of you know play with the voltages on each of these piezos to get this the shape that you want. This is a response function sort of set for the the bimorph. So this is essentially the exact same thing we did where we looked at the wavefront as a function of the bender motors on the on the bender motor. But this is now as a function of the 18 different channels that you can apply voltages to on the bimorph. So you apply a voltage of 100 volts to one electrode and you get the purple curve and the other electrode get the orange curve. And so from this sort of set of response functions you can then form any any mirror shape or any wavefront shape that you want. And so we went through the sort of iterative process to to correct the the the wavefront coming into the second mirror and get a relatively optimal focus. So that's actually kind of the neat thing about these bimorphs is that you you really don't need a very good first focus because the bimorph can correct everything that's wrong with the wavefront coming into it. So here's a little movie of the the bimorph actually responding. So so we we start with a you know relatively flat mirror and then we put all the voltages onto all of the electrodes and you see this thing settle over the course of you know sort of a minute or something like that to a relatively flat focus or a relatively decent focus. And so we did this for for three different three different configurations. We got a relatively small spot of about three microns. We were able to tune up a five micron focus and then about a 10 micron focus. And and you know this is the sort of agreement between what the the focus spot should have been and what it was actually measured to be. So so we were able to actually do this and and actually dial up desired spot sizes based on a on an input from a model to to tune the two mirror systems. So it seems like it's all coming together and I'm actually pretty optimistic that I keep teasing Chambot that I want to slide her on the beam line control where I can slide it from 50 nanometers to two microns and and and it'll just work and it seems like we're getting close to being able to to do something like that. Here's the specification for the for the benders for the actual beam line design the sort of shapes we're going to need to be able to achieve with with Demian's bender system. Our current test mirror is 200 millimeters long and we can almost bend it this much. So we're pretty confident that with a longer mirror we're going to actually be able to push the bending a little bit further and get to the shapes we need. We've actually ordered a second set of components to build a second mirror so we don't have to be so worried about breaking it and and we're going to try to push it to to the furthest possible extent and get a sense of where it'll break. This is the specification for the for the bimorph mirrors the hybrid bimorphs that'll be the second kb mirror pair. So this is the sort of like greatest curvatures we're going to need to achieve for the smallest spot sizes and and you can see here this is for the red and blue curve here are for the different working distances so the red I think is for the the short working distance mode and the the black here is for the longer working distance mode and what's pretty crazy about this is that the bender and bimorph system combined probably can't achieve these these curvatures. So what we're going to have to do is actually polish a prefigure into the mirror. So so even when the mirror is flat it's going to be bent to or it's going to have a figure that that sort of matches this blue curve here. So so even with zero voltage and zero bending we'll we'll have some focusing and I forget what this what this shape corresponds to I think it corresponds to about a 300 nanometer focus or something like that and then the the bender and bimorph will will move us between the the red and the black line. So this is the spec for the horizontal mirror this is the spec for the vertical focusing mirror again for the for the two working distances. And and you know we're we're currently collaborating with Osaka University in JTEC to to really prove that something like this can come into existence. So that was kind of what I wanted to say about the optics and I guess I got a few minutes left. The least developed component of the beam line is actually the end station here so again we're going to be building a new hutch down on the end of the sector here. So I thought I'd spend a little bit of time with the the thing we know the least about and and I'm actually looking for a lot of advice here because because we're starting to get asked I'm starting to get asked a lot of questions by our engineering team with regard to to what we need to do to to do the experiments. So this is the sort of conceptual layout that we we've had for for a long time it's been getting slowly more and more refined. One thing we actually asked to do is to make a hutch that could extend out into the aisle a little bit to give us more space. So from where our diffractometer will be to where the wall of the hutch will be is is limited by the the aisle of the APS so people walk in this space here and we have this mezzanine it's sort of a mechanical mezzanine up above the walking space and I asked for special permission to extend our hutch out into this aisle a little bit to give us a little bit longer detector arm. We were allowed to go to just the edge of the posts that hold up the hold up the roof of the APS and so we got basically an extra meter of space here. The plan is to have a control room on just upstream of the hutch here where we can sit and work. The control rooms are the first things that are being descoped in all the beam lines so I hope we get to keep it but anyway we'll have a small a small door here for people to go in and out and we'll have a larger door on the inboard side here for moving equipment in and out. Another sort of feature of this is that we're going to have a sort of a tube on the end of the hutch here. Our original design had actually filled the whole sector with enclosure and I started to realize that you know some space outside of the hutch is as important as the space inside the hutch but I wanted to keep this ability to, wow that's kind of neat I don't know how I did that, I wanted to keep this ability to do you know sort of small angle scattering experiments which are much with a much longer detector distance and Gary who was an engineer at the time said well why don't we just stick a tube um off the end of the hutch that can be shielded and you can have a track in there and you can just stick a detector in there if you if you want to use it so we're still kind of keeping that design feature of the enclosure. I wanted to point out that this is kind of a scary big space so this is where the the hutch is going to go at the end of sector 34 and you know this is the ceiling this is about six meters where these ventilation ducts are so we're basically filling this whole space between this post and the end of the Laue diffraction microscope with enclosure and I'm really just trying to decide if I if I really feel like we need this sort of bump out idea here and we really need all of this height but the sort of five meter detector arm distance was motivated by our sort of largest spot sizes sort of two micron beams need about a four meter detector distance with a sort of reasonable pixel size and so that was sort of the motivation for this sort of bump out here. We haven't done final design on this enclosure yet so there's still time to to sort of rethink all of this. I keep in sort of regular touch with Garth Williams at Brookhaven who's designing the or leading the Coherent diffraction imaging beam line there and we try to maintain a sort of overlap between what's going to be the capabilities of the two beam lines and and really they're going to be you know solidly working on this and the sort of bigger detector distance capability and and so this is all overlap and so this idea of going with a detector to sort of four or five meters isn't really necessary from a sort of capabilities point of view because Brookhaven is certainly capable of doing this or will be capable of doing those kinds of experiments so that's a place where I'm still trying to decide if I really want to keep pushing the the size of the detector arm mostly because it's also an engineering challenge to put a very large detector so far away. This is sort of our block diagram of the of the instrument so we're going to have the the KB system here this goniometer we're hoping can be a very high spec goniometer that will manipulate the sample I'll say a little bit about that at the end and then there's the sort of slits and shutter and a wave front sensor for tuning up the KB mirror system this can be removable we can move it in and out of the beam you know maybe attenuators so this is the sort of sort of straw man that we're starting to sort of assemble into a more realistic engineering sort of layout so Scott Izzo is the engineer that I've been working with on the sort of layout here and this is the sort of model he's been slowly developing so we we're still kind of sticking to this idea of a sort of two circle detector arm holding a big honking detector that can move from you know sort of a half meter out to five meter distance the sort of KB mirror drawing here is an actual engineering component that's being developed to hold the hold the bimorph the hybrid bimorph system so that's that's relatively realistic it still needs to be refined a bit you can see there's no vacuum chamber in the drawing yet and we have a slit and we have some sort of boxes here now that represent a wave front sensor perhaps a BPM a high speed shutter and when I say high speed I mean sort of millisecond and then an attenuator of some type and these things are slowly being you know turned into more realistic more realistic elements and then the the the variable working distance will be accomplished by essentially moving the whole KB mirror system so a lot of people are probably familiar with this this new instrument at the APS called the velocicrobe and this was something that was developed by by Kurt Preissner who's now our our lead engineer at the atomic beam line or he was actually making stages out of granite and so this thing is an air bearing stage that you can turn on the air the granite floats on top of the other piece and you can motor it left and right and then there's a wedge system here to motor it up and down there's a gantry system here that you can also motor the optic forward and backward so so these things are just starting to get picked up and used by the engineering staff at APS in fact the the monochromator that was built for our our allowing diffraction partner user proposal with Los Alamos was actually built on one of these granite stages so you know we're getting quite a bit of experience with them so we're kind of picturing something like this we'll have a granite stage here that essentially moves the KB mirrors forward and backward by 100 millimeters to to change the the working distance of the system with our monochromator we've gotten a pretty good sense of the sort of angular stability of these granite stages and it's actually surprisingly good we can move the the mono in and out of the beam and and it's still you know well within the rocking curve of of the silicon 111 crystals and you know just a little bit of tune up and you get back on the peak so these things are are pretty good and they can probably be even better with a little bit of careful thought that goes into the sort of angular reproducibility there's a wavefront sensor idea something that's been actually developing quite a lot at APS and and we're starting to to get to the point where we think this thing will work in in pretty near real time like sort of one hertz or maybe half hertz type frame rates where we'll be able to analyze the get a fully analyzed wavefront out of the instrument and it sort of you know hurts kind of rates which will make tuning up the the mirror systems pretty pretty effortless I think if you can get that data directly into a control system that then feeds back to the shape of the mirrors we should be able to tune up the the zoom KB system pretty well we're gonna need two of these we're gonna need one between the mirrors and then one at the end downstream of the mirrors to tune up the the second pair and and so we're kind of still developing that that sort of idea these things are sort of a half meter long one thing I've been also looking at recently is is the space upstream of the KB mirrors you know fitting all of these bits into the to the space between the wall and where the KB mirror goes has been a little bit of a concern we originally had about 1.4 meters between the wall and the center of the diffractometer and so I was starting to to be a little bit concerned that as we started to move the center of the diffractometer downstream to get more space how would that impact our our sort of detector distances that we could achieve given the you know the distance to the hutch walls and so I just recently did this little little simulation where I moved the diffractometer downstream and computed the distance to all of the walls and and you can see that as we move downstream from sort of the initial position that we expect to about two meters downstream we only lose about 700 millimeters of sort of detector distance and then there's an envelope here you know you can get the detector further away up here than you you can over here this is really cool I keep getting little lines um so anyway I wanted to do this just to uh to get a sense of what the impact was of moving this diffractometer downstream and and it works out to not be so bad so I think the model at scott is currently developing has the uh the sort of diffractometer moved downstream by a half meter or 750 millimeters or something like that so it's like the second or third frame in this this movie here um so the guineometer is uh something that we we've made very slow progress on and a lot of it is because we uh we suddenly started to worry about vibrations and and the interaction between the detector arm and the sample and and it sort of put a freeze on the development of the the guineometer I've been showing this slide for a number of years now where I had this this sort of picture in my head of a sort of inverted joystick where you'd have this highly polished cradle that you could tip and tilt by moving the the joystick on the bottom here and uh and you could get a third axis by rotating about this sort of center axis of the joystick and and then as a sort of a point of uh of a merit I wanted to point out that the Hubble space telescope has a figure of 30 nanometers so in principle if you could polish this cradle to as perfect the sphere as the Hubble space telescope mirror you could have a 30 nanometer sphere of confusion um three axis guineometer um again I say that with a smile on my face because I I know that that's uh that's a little bit overly simplified but um you know we're starting to explore now with some uh companies you know how far could you push the ability to uh to uh you know get a very high spec guineometer with a sphere of confusion that's approaching our x-ray spot sizes if I want to do a a scan of hkl from you know 1.5 to 4.5 I'm rotating my sample you know 20 30 40 degrees and I'm trying to do that in a 15 nanometer x-ray beam you sort of need a very high spec guineometer to do that so we're trying to explore how far we can push this and uh and and trying to get some uh some engineering companies involved in in perhaps uh you know seeing seeing what we can do um and when we started talking about this guineometer we also started talking about vibration if the uh the like if the the whole system was going to be vibrating on the scale of 100 nanometers it didn't make any sense to try to make a guineometer that had a 15 nanometer sphere of confusion so so Kurt Preistner started looking at you know just how how good is even the floor at sector 34 um so so that was one of the first measurements he did um I guess this wasn't back in uh in January started looking at how good is the floor in the east station at 34 id so this is where the allowing microscope is so it's just upstream of where our instrument's going to live um shambo has a spec on the kb mirror system of 10 nanoradians and and so we started to say well can the floor even do 10 nanoradians and it looks like it barely can so in the sort of five to uh sort of five to 15 hertz range here you know it's just making that 10 nanoradians spec at least in two axes um he has to redo the measurement to get the third axis um out of the floor um he also went to 34 idc and he started doing measurements um on the floor and on our diffractometer and on our optical table and on our sample stage and even on our detector arm as a function of moving our detector arms we're kind of curious how much does moving the detector at 34 idc impact the sort of vibrations felt at the sample and he actually found it wasn't zero so as he moved the detector arm up and down and inboard and outboard he started getting a sort of floor vibrations that were you know sort of you know 50 nanometer type vibrations on the floor that was amplified onto our um onto our sample stack by you know the order of microns as the the detector arm was moving um it's well known that our optical table at our beam line is the worst optical table at the aps it's it's it's normally vibrating a micrometer um so so to see it go up by a factor of two wasn't wasn't too surprising um and i thought was really neat was he actually measured on the detector arm as well so as he was moving the detector arm up and down he was looking at vibrations on the detector arm just just due to the motion and and he was finding you know quite large vibrations like 13 micron level vibrations and 60 micron level vibrations on our detector arm as as it was moving so that that could actually impact this ability to sort of fly scan the detector if the the detector is vibrating on the scale of the size of a pixel of the detector um that that might have an impact on our ability to do fly scans um at the beam line so so it was an interesting sort of exercise but at least we're we're confident we can well we're relatively confident we're going to be able to meet the spec of the sort of 10 nano reading stability of the the kb system and then the rest of the instrument's going to have to be designed to try to you know really preserve that um as we go forward um i've just been asked by the project to start working out the the detector specification uh the straw man detector that i told the engineers to to use was an igar 16m mostly because it's really big and it's really heavy it's 55 kilograms it's 400 millimeters in size so it gave them a sort of straw man to say okay if i'm going to try to move this thing from half a meter to five meters and i'm going to try to sweep a a diffraction signal from you know 10 degrees to 40 degrees you know how am i going to be able to move something that big so it gives them something to at least uh at least think about um but we don't really have a detailed spec yet for the detector um i've got a slide here that that shows in a few seconds the kind of fluxes we're expecting and uh and i'm a little bit worried about that so i did a simulation similar to my little little gold crystal that i showed in the first few slides but i simulated a sort of micron scale um silicon you know nanowire so sort of two microns long by you know 500 nanometers by 200 nanometers because that was a lot of atoms and it took a lot long time to simulate the data set i uh i generated that sort of simulated data set and i was uh getting like 30 million counts per second in the 004 bregg peak um in that simulation so so we're talking about sort of 10 to the 8 photons per second or more um from you know maybe realistic samples um in the sort of realistic estimate of the flux of the beam line and uh there's not a lot of detectors in the world that can swallow 10 to the 8 photons per second continuously you know there's these aegyp technologies there's the young frow technology that's that's coming into existence that that has higher dynamic range but it's um a little bit scary that you know something like an igar doesn't really behave linearly above 10 to the 6 photons per second um so specification detector is going to be hard um this was a slide i showed years ago where there's the mm pad that was developed at cornell that can show you know sort of sustained 10 to the 8 photon per second um measurements um chambeau gave me this number originally that in our 15 nanometer spot if we got the direct beam on the detector for something like a small angle typography measurement you could expect something like 10 to the 11 photons per uh per second in the brightest pixels um so so those are you know pretty and you know it's probably not going to be that high in reality but it's still it's still going to be a very large number and and spec in a detector i think is going to be quite difficult um but fortunately the project is going to let us do that last and so you know hopefully we'll just buy the best thing we can buy in you know 2021 or 2022 or 2023 when the APS comes into existence um another question i was asked by the engineers was what about the detector pose and so we started thinking you know if this detector that we have at 34 idc if this sort of two-circle goniometer is going to impact the vibrations that in the optics and sample what are some other models we could use for for moving the detector arm around and and you know everybody's going to be pretty familiar now with the robot at nanomax where you can you know basically position the detector anywhere you want but um what is the impact of the accuracy with which the detector is pointing at the sample or what we've been calling the detector pose how perpendicular it is to the incoming diffraction signal um what is the impact of that and what specification do you need to be able to you know first of all measure it and second of all deal with it if you go away from something like a two-circle goniometer that we're all very familiar with um so there's some other ideas you know like the robot arm or perhaps uh one circle like this with a stage stack that moves up and down and then a tip circle that you know rotates the detector to point at the sample or or even something like this which is sort of similar with what they're developing at what garth is doing at the cdi beam line at brookhaven where you have a sort of large arc where there's there's no motors near the sample you you move the detector arm on this arc you have a stack on the on this arm that can move forward and backward and up and down and so you can you know basically move your detector around through through two circles in a scheme like this where you know all the motors are very far away from the sample i've been slowly starting to work on this but it's actually quite a hard problem to to work on i wasn't really expecting it to be to be so hard but at least it's been hard for me um so so i can use extra utilities to to say okay i'm putting a tip into my detector a tilt into the detector about one axis and i can simulate the coherent diffraction pattern i would see as a function of that tilt angle but then coming up with how that impacts the uh the image that you retrieve from the phase retrieval and then coming up with a spec for how accurately you need to know the the sort of pose error in the detector um and and the impact that that that that error will have on the on the image has been a little bit difficult mostly because when i simulate data sets like this you know something like a 4k by 4k detector i'm really rapidly getting up to like six and 10 gigabyte coherent diffraction patterns and even doing sort of phase retrieval on a 10 gigabyte data set um in 3d is is has been difficult so i haven't gotten very far with with coming up with a specification for this so i'm kind of interested to to hear what what people have had um you know and their experience at other instruments and and and you know Kurt's really bugging me about this because he really wants to know um sort of what we have to specify uh for our detector manipulator and i guess that was about all i had planned to say and just wanted to say thank you and um figure because argon paid so much for the slide template i should use it it's a little bit funny um and uh hopefully there's some some good questions and discussion um thank you so much ross this was very interesting and i i've already a couple of people asking to make questions so the first was jen robinson please can you unmute yourself jen and ask your first question that that's um an incredible amount of of work that's gone into all that and um i i wonder you know five years from now how much of it's all going to be true but uh that's uh you won't you won't it won't happen if you don't try um the i had a couple of questions about the optical design of the uh the zoom system uh firstly a comment that you you said it yourself but the wavefront sensor is is uh indisputable you you absolutely have to have that and we've talked about this also for garth that there's no way to to align anything like this unless you can put in something in between the optics and and actually align each piece separately so so i think you said that but uh that does need to be emphasized because it's a much more complicated optical system so my i've got two related questions one is why you've constrained the intermediate focus to be the same for the horizontal and the vertical because that is a degree of freedom that would give you a lot more options i would i would think um i so so you basically mean this prefigure this blue curve no i mean you've decided i think i think without justifying it that the the vertical focus and the horizontal focus of the intermediate one the one that moves um are both us in the same place uh i still don't think i understand what you mean the one that's basically at the kb a zoom kb system is is a separate system in the horizontal and the vertical so it's a horizontal pair and a vertical pair um and that intermediate focus could be different for the two in a different position for the two uh oh you mean you mean you mean for the for the first focus okay yeah so the the secondary source basically so this thing yeah yeah certainly there's there's there's there's no reason that you can't do them completely separately so you could have a a two micron by 50 nanometer beam if you are and and certainly that that's all that's all baked into the design yeah so you are you are actually allowing those folks those two foci to move separately because you can't put the vertical and the horizontal mirrors at the same place you have to have them split so it doesn't seem to make any sense to have that focus at the same place yeah yeah exactly yeah yeah there's there's there's no reason you can't do any combination so you can have an any any spot size in any direction and uh yeah yeah exactly yeah okay that's certainly planned or baked into the system okay good yeah and then the the second half of the question is you you never mentioned the word coherence in any of this and I'm wondering uh where your coherence aperture is in in this design yeah that's a good point I never mentioned the word coherence so so here's the basic uh beam line layout and uh and currently we have a slit um after the monochromator before kb1 okay so so that'll be sort of our our coherence defining aperture and all of the flux specifications that shambo gives me is for a single coherent mode by his definition of a coherent mode and so you can always give up on that a little bit and increase the uh increase the flux we also have a slit before kb2 um sort of a cleanup slit or a collimating slit and uh we haven't really thought too much about how these two slits are going to play against each other because if you remember olig gang um no olig jubar here who's been advising garth says that it's it's bad to use slits for that because you um you know you get diffraction from them and it shows up in your in your focus and so forth but I guess that's that's all been thought through yeah not really by us mostly because we we have some constraints so there is a paper from osaka where they uh they explored what happens if you add an appetizing aperture in between the two mirror sets so so mirror set one you put an appetizing slit um at the first focus and then and then mirror two produces the focus and you can see in in simulation they got extremely clean um beams you know here's the slit diffraction that's demagnified and then and then they cleaned up all of that slit diffraction and then they actually showed that it improved contrast and coherence fraction so they actually did a coherent diffraction measurement from a a little uh grating or something like that and showed that the the sort of uh visibility of the fringes was improved by using this appetizing slit um you know because we have uh we have the 3d m and experiment between our two kb mirrors we don't have a lot of freedom to do stuff between our kb mirror sets and uh and you know those guys they fight for every millimeter they've you know they're not happy about having a beam pipe through their hodge so so it's been a bit of a challenge to uh you know you know convince them that that they're going to be fine with like you know 500 millimeters of space at their end station here but at least we won't have to have pumps or anything in there i guess right we thought that one helped before yeah the other way around good thank you thank you uh there was a question from Richard sunberg about the maximal sample to detector distance i think you talked about it yeah yeah we discussed that at the end does he have any other comments about yes richard do you have any other maybe you can unmute yourself or write if you want a talk and and i would just like to invite everybody else to um either raise your hand in the chat or unmute yourself and ask this is a very informal talk so please profit of this uh situation and ask ross all all the questions you you would like i don't see any urgent question now actually a question about the kb ross and it's always um uh you know it's something that i also have to face at a certain point um so the um bendable kb versus a fixed curvature kb um i mean there is a there is a big risk of stability isn't it um so the the fixed curvature i've had this amazing quality of being stable over a long time now i think nanomax has proved this quite nicely uh while on the on the bendable uh especially if you want to add this active feedback which may induce this you know hurts or sub hurts instabilities uh can you comment on this well that that was our motivation for adding the the cap sensor array behind the mirror so so we can basically constantly watch the shape of the mirror and uh and and look for drifts in the shape of the mirror um so so so that's our motivation for that so we basically in the whole zoom system we're including an in-situ metrology so so for the benders we have cap sensors on the back and then for the for the bimorph we're going to have this uh interferometry array on the top surface that we'll be looking at um looking at the optical surface of the of the mirror um because we know that we're going to have to to monitor the shapes of the mirrors and and probably maintain them at at a relatively high rate um well we we don't know how what kind of rate we're going to need to correct at yet but uh we're going to have the ability to to you know monitor the shape and and maintain it um and hopefully what's ever needed okay there is a question from dina schiefer about uh energy scans do you consider using energy scans for bcdi yeah certainly there's no reason not to um we're we're slowly starting to do it more at 34 idc so this the sort of small offset monochromator idea um should be very capable of doing energy scanning um we've already specified a slew rate for our our undulator gap um to uh to maintain the uh the there's a synchronization between the the undulator spectrum and the monochromator and it's going to be pretty fast like the full range in something like five seconds or 10 seconds so so we've uh we've certainly been been uh maintaining this idea that we want to be able to do to do energy scanning uh coherent diffraction as well dina is commenting again maybe um i we also thought of the possibility of using automatic attenuator um between the sample and the detector just in case you know during this uh we're far more concerned about the sample than we are about the detector so so the attenuation is probably going to be to stop the sample from melting when you're putting 10 to the 12 photons on it um then then then we are worried about the detector at this point um actually i started looking at at at the attenuators a little bit because you know don't currently we don't really use the attenuators in any way that's calibrated we don't really care you know about very specific attenuation we just want to make the beam dimmer because our sample is getting destroyed or because the detector is saturating um so i've been looking a little bit at different ideas for attenuators and and i was thinking of you know maybe exploring like a piece of silicon that's wedged and uh the thin bit will be less attenuation and the thick bit will be more attenuation and we'll just motor that wedge of silicon through the beam uh to adjust the attenuation uh but you know again it's all tied to this sort of space we have upstream of our diffractometer um i'm starting to you know worry about that and worry about you know what goes into this space here and you know this is a this attenuator thing here this box is actually an idt attenuator block and it's like i think eight eight foils or something like that and it's half a meter long you know it really eats into the space we have upstream of our our kb mirrors so so i'm starting to think about what's going into these gray boxes um quite a bit now okay uh i see that we have gone now 20 minutes after let's say the the hour which is good we are still 30 people um is there any other urgent questions from the audience because i think the discussion can go on a long time okay um i have a question go ahead steven i already asked one oh thanks uh so we're talking about 15 seconds for a data set right uh possibly what are you well you're probably doing that now right uh not quite um but i'll tell you in a couple of weeks um the uh the thing that is a bit concerning is the ability to scan the sample right have you thought about this at all yeah yeah the ability to scan that fast right and yeah and uh yeah you know so i was reading actually your fly scan paper um yesterday and uh and starting to come up with ways to actually measure how much time we're currently spending on motor movement and detector on triggering um so yeah i mean that that's all that's all kind of baked into this sort of specification of the sample goniometer and uh and and i'm really getting concerned that we're not making any progress on this um we we made some progress late last year with developing a sort of request for information document to send out to companies and say here's what we'd like to do you know how close do you guys think you could come and we were starting to set up meetings for uh you know discussing this with different companies and uh and it all sort of came to a crashing halt in the last uh sort of you know nine months or something like that so so i really want to uh get back to this and uh and i think now that we have a better handle on things like you know the vibrations of the floor and the beam line and and the sort of spec of the detector arm that we're uh we're going to be looking at we can start you know approaching the sort of specification of a goniometer with a little bit more sort of confidence and uh and start talking to companies about this but you're right i mean scanning a sample very fast and then perhaps fly scanning and continuously measuring and uh it's all it's all stuff that's you know slowly being developed you know you guys are making progress on it um and and we're going to start you know working on it at a certainly you know different scale than you guys but uh but it's all it's all got to be solved and and i think we're making progress on it okay thank you well before going ahead maybe with a more informal discussion i would like just to uh close officially this session uh thank ross enormously for this contribution and and remind that actually we are now opening for more beam lines so you uh in a in a week time in a couple weeks time i think it's um steven who is going to give a talk on uh i d1 or the srf that there is also on the line nsls2 and so lay and i'm in contact with more scientists so please keep keep tuned and the program is going to be updated we