WEBVTT 00:00:00.000 --> 00:00:00.000 Yeah, I can only request to record, I can't record. 00:00:00.000 --> 00:00:03.000 Request to host host. Okay. 00:00:03.000 --> 00:00:04.000 Perfect. 00:00:04.000 --> 00:00:05.000 Great, thank you. 00:00:05.000 --> 00:00:23.000 Okay. All right, well, thank you, everybody. I think in the interest of time, we should, um… we should get going. Uh, so this is the, as Maurice was mentioning, this is the inaugural meeting of the entire Meerkat Center. This is one of… 00:00:23.000 --> 00:00:31.000 The new microelectronics, uh, um. Science research centers funded by the DOE Office of Science. 00:00:31.000 --> 00:00:39.000 Uh, this one is, um… focused on the energy efficiency theme, so energy efficiency and microelectronics. 00:00:39.000 --> 00:00:51.000 And, um, it is composed of 8. Teams, um, funded for, um, a period of 4 years each, uh. 00:00:51.000 --> 00:01:03.000 Focusing on various aspects of of energy-efficient microelectronics spanning from sort of, uh, materials and device scale. 00:01:03.000 --> 00:01:12.000 Phenomena through, um. Circuits send, uh, architecture to, um. 00:01:12.000 --> 00:01:34.000 Applications. And, um, the, uh, the eight teams all have Uh, somewhat, uh, you know, quite distinct, actually, mission statements as to what they're trying to do, but they're obviously, uh, areas of commonality or cross-cutting interests, and Uh, we thought collectively the 8 PIs of these teams, that it made sense to have 00:01:34.000 --> 00:01:45.000 A series of meetings where we basically introduce each of the teams to, uh, each of the projects to the entire Meerkap Center Um, community. 00:01:45.000 --> 00:01:49.000 So, thank you for joining this first, uh, meeting, and, um. 00:01:49.000 --> 00:02:01.000 We will hear, um, uh, in our first presentation today, we'll have one presentation per week for the coming, I guess, this week and the next seven weeks. 00:02:01.000 --> 00:02:07.000 We'll hear from Maurice Garcia Severas from Lawrence Berkeley National Laboratory. 00:02:07.000 --> 00:02:13.000 Who's the, uh, lead PI. Uh, for one of the eight teams. This one is focused on nanohybrids. 00:02:13.000 --> 00:02:25.000 So I think… I assume that Maurice will start by explaining to us what nanohybrids are, and then we'll learn about some of the exciting plans for his research. 00:02:25.000 --> 00:02:26.000 Yeah, uh, thank you very much. Um, can you hear me okay? 00:02:26.000 --> 00:02:31.000 Project. 00:02:31.000 --> 00:02:44.000 Yeah, okay. So, uh. Yeah, this is the website of our, uh, project called Nanoscale Hybrids, a new paradigm for energy efficiency up to… energy efficient optoelectronics. 00:02:44.000 --> 00:02:49.000 And so I'll try to explain what… what that is, and what we're doing. 00:02:49.000 --> 00:02:58.000 Um, actually, I'm gonna give the first overview of the presentation, and then… Jackie Yao will talk a little bit… a little bit more about, uh. 00:02:58.000 --> 00:03:03.000 So, modeling and simulation, which I think might be of interest, uh, very broadly. 00:03:03.000 --> 00:03:15.000 So, that said, let me start… Let's see if I can make this full screen. 00:03:15.000 --> 00:03:29.000 Sure, sorry. Okay, uh… 00:03:29.000 --> 00:03:40.000 Okay, so that should now be full screen. Okay, so the first question is, this is an energy efficiency center, so energy efficient how, what we're doing. 00:03:40.000 --> 00:03:52.000 Um, just a bit of a cartoon. Uh, normally, or, well… more and more these days, we use sensors that have very high data rates. 00:03:52.000 --> 00:04:01.000 Uh, that we're going to collect a lot of information. Uh, and then, you know, we need a lot of computing to process that information and a lot of power. 00:04:01.000 --> 00:04:13.000 Uh, and what we would like to do is to have better sensors, in a way, that don't require… don't produce so much data, so produce information rather than data. 00:04:13.000 --> 00:04:24.000 I don't require as much computing. Um, so we're not actually dreaming in this… project of a new computer that can do the same job with less power. 00:04:24.000 --> 00:04:27.000 Actually, when I reduce the job that needs to be done. 00:04:27.000 --> 00:04:38.000 Uh, but we're not inserting edge computing here. To eliminate data transmission, which is… and storage, which also take up power, uh, but still, you know, doing the same computations. 00:04:38.000 --> 00:04:48.000 So, what we're talking about is a new kind of sensor, uh, where the raw output of the sensor is programmable, trainable information. 00:04:48.000 --> 00:04:53.000 Um, and the fetch computing is present, it's… comes after them. 00:04:53.000 --> 00:05:07.000 Um, so this is a… we're talking about, uh. Light sensor, so IR visible. Uh… Where, um… we can think of the sensor as a… just an element. 00:05:07.000 --> 00:05:19.000 Or, you know, the photon field. Uh, it's, uh, it's interacting with that element, and then information some pre-processed information comes out the end. 00:05:19.000 --> 00:05:24.000 Without any sort of. Electrical manipulation, um. 00:05:24.000 --> 00:05:36.000 Oh, that's through transistors, and so… Now, if you consider silicon, which is a… of the sensor, the photo sensor material. Uh, silicon sensor outputs electric charge, or current. 00:05:36.000 --> 00:05:42.000 Which is proportional to the illumination. This raw output has to be processed to extract information. 00:05:42.000 --> 00:05:47.000 Uh, you can't train silicon to respond differently to data than it did yesterday. 00:05:47.000 --> 00:06:02.000 It does what it does. Once it's built. And uh… and the response is, you know, given by the material properties, of course, you can dope it and so on, but the the… you can't really ask Silicon, for example, to absorb 00:06:02.000 --> 00:06:09.000 Ir. It won't. Uh, so Ananos… oops. 00:06:09.000 --> 00:06:24.000 Huh, sorry, uh, this is maybe the wrong… presentation, uh… Oops, uh… Let's see if this one's more up-to-date. 00:06:24.000 --> 00:06:32.000 Well, okay, so I don't know why this didn't change. This is supposed to be… artificial. 00:06:32.000 --> 00:06:38.000 Uh, I hope, I hope this is now, uh… 00:06:38.000 --> 00:06:53.000 Correct. So, a nanoscale hybrid is an artificial material, so to speak, where it's made… it's actually made up of separate elements smaller than the incoming wavelengths of light. 00:06:53.000 --> 00:07:01.000 But as far as the light is concerned, it's one material, just like silicon, because the light can't resolve these individual elements, the wavelength is too large. 00:07:01.000 --> 00:07:07.000 Um, so, uh… Unlike a natural material. 00:07:07.000 --> 00:07:23.000 With this kind of arrangement, we can independently control separate aspects of the photon matter interaction, so we can… we can, uh… manipulate the absorption separately from different suction, the amplification, we can introduce nonlinearities. 00:07:23.000 --> 00:07:28.000 So, in a single material like silicon, you get all these things. 00:07:28.000 --> 00:07:34.000 Together. You don't get to separate them out, and… adjust them separately. 00:07:34.000 --> 00:07:44.000 Uh, so, okay, this is what I said, we can optimize sensing process. The sensing process, for example, to extract the maximum possible information from the photon field. There's a later slide. 00:07:44.000 --> 00:07:50.000 On this. Um… But in this project, what we aim… we aim to do is go a bit further than that. 00:07:50.000 --> 00:07:56.000 And to use the nanoscale hybrid also to process information, ideally in a programmable way. 00:07:56.000 --> 00:08:07.000 So basically, make an artificial material that can learn. Okay, so now, just to make… put that a little bit on the real axis. 00:08:07.000 --> 00:08:16.000 Here are some examples of things, um… Along the way of what we're trying to make here, so… So this is a pretty smart sensor. 00:08:16.000 --> 00:08:33.000 Uh, but there's, you know, it doesn't actually use nanoscale hybrids. What it… what it is is a… as the title says, a two nimble bipolar photodiode, so by a suitable arrangement of different materials, um, and shown in the cartoon. 00:08:33.000 --> 00:08:54.000 You can manipulate the band gaps, and so, um, in such a way that photons bigger than a certain frequency will produce negative charge carriers, and And, um, below that frequency, they'll produce positive charge targets, so you basically get either positive or negative charges, that pumps that… 00:08:54.000 --> 00:09:01.000 Hence bipolar. Uh, depending on, um… Depending on what wavelength is coming in. 00:09:01.000 --> 00:09:17.000 Of course, you only get two choices, positive or negative, but with those Two choices already, you can, uh, you know, you can voltage control the threshold at which you distinguish between positive and negative charge, depending on… so the threshold wavelength, basically. 00:09:17.000 --> 00:09:38.000 That is controllable. By an applied voltage, and you can then program, uh, or… or produce some sort of a learning device, which can recognize, uh, spectral features this way, or can distinguish material. So this is… I need to add the reference for this, sorry about that. I have some missing references, so if we upload these slides somewhere. 00:09:38.000 --> 00:09:43.000 I will, um… I will, um, put the references in there. 00:09:43.000 --> 00:09:55.000 Um, so here's another example of an in-sensor learning. Again, without nanoscale hybrids. So in this case, uh, what we have is, um. 00:09:55.000 --> 00:10:02.000 An electro-optical modulator, uh, which is, in this case, just, uh, you know, a liquid crystal. 00:10:02.000 --> 00:10:06.000 Uh, which can, of course, we never rotate the polarization and so on. 00:10:06.000 --> 00:10:18.000 Um, but we control it with a diode, with a reading of a photodiode that's in the same pixel, let's say Uh, and again, with some, uh, with some electoral feedback loop. 00:10:18.000 --> 00:10:26.000 And so… so in that case. Um, what you can do is engineer an optical nonlinearity that's low power consumption. 00:10:26.000 --> 00:10:42.000 Compatible with incoherent light, because the photodiode doesn't care about wavelength in this case, and it's CMOS compatible, so you can build a chip out of it, and here's a chip with half a million of these pixels, and in real time, you can change, for example, the 00:10:42.000 --> 00:10:58.000 Do a image contrast enhancement with using this device. Uh, by adjusting You know, which, uh, the pixels that receive more light reflect less on the pickles that receive Let's light reflect more. 00:10:58.000 --> 00:11:03.000 Just on that in real time, automatically. Of course, this involves an electrical feedback loop. 00:11:03.000 --> 00:11:14.000 And what we aim to do in this project is implement such feedback and couplings between nanoscale elements, so before any transduction into electrical cities. 00:11:14.000 --> 00:11:21.000 So, longer term. Uh, so what I've shown so far is kind of single pixel stuff. Longer term. 00:11:21.000 --> 00:11:41.000 Uh, we are trying to develop concepts to make networks this way. So, uh, the except… you know, the inspiration… preparation for this is, of course, how vision works, so in the vision, we… we don't record images, right? We have a visual cortex actually outputs information. 00:11:41.000 --> 00:11:45.000 That goes… sorry, the retina, actually, I was information that goes to the visual cortex. 00:11:45.000 --> 00:11:49.000 The visual cortex that's further processing, all with very low power. 00:11:49.000 --> 00:11:57.000 Um, so for this. Kind of arrangement, uh, you need some sort of, um, some sort of, um. 00:11:57.000 --> 00:12:11.000 Math or manipulation of the… of the input, uh, to, um, to produce the information at the output, and so that's… in order to do that without some sort of a circuit. 00:12:11.000 --> 00:12:23.000 Uh, what we need is interactions between multiple of these nanoscale hybrids, so we just make a device that's a lot of nanoscale hybrids, but they can actually interact with each other. 00:12:23.000 --> 00:12:32.000 Um, so one possib… oops. So, um, so here's then an example of a network, again, without nanoscale hybrids. 00:12:32.000 --> 00:12:38.000 Um, this is, again, using circuits, but the point here is to have optoelectronic neurons. 00:12:38.000 --> 00:12:49.000 Uh, so I get to essentially make an artificial retina. Um, in a way that you, you know, use his sensors and then circuits. 00:12:49.000 --> 00:13:01.000 To process the information from the sensors, and then… but then do that optically, so there's also lasers, so that information comes in, uh, optically, and also is shared optically among the network. 00:13:01.000 --> 00:13:13.000 Um, but again, this… this is… this is, uh… you know, not doing… not using directly interactions between nanoscale hybrids, so it doesn't actually have nanoscale hybrids. 00:13:13.000 --> 00:13:19.000 And so one possibility to make some sort of coupling between nanoscale hybrids is using plasmons. 00:13:19.000 --> 00:13:28.000 So this, this, uh, org shows, um how, you know, um… My prior work using, you know, using, um. 00:13:28.000 --> 00:13:39.000 Arrays of materials placed in a particular geometric arrangement. To produce, uh, different frequency absorptions, uh. 00:13:39.000 --> 00:13:47.000 And be able to adjust the frequency absorption by the… by, uh, some sort of small. 00:13:47.000 --> 00:13:56.000 Dielectric properties between these plasmas can be used, in this case, to sense molecules on the surface. 00:13:56.000 --> 00:14:01.000 And depending on the concentration of a of the target molecule that's being sensed. 00:14:01.000 --> 00:14:07.000 Uh, you can… you get more or less wavelength shift. 00:14:07.000 --> 00:14:14.000 Uh, of the… of the resonant frequencies in this, uh, plasmonic structure. 00:14:14.000 --> 00:14:21.000 So that could be a way in which we can couple Uh, nanoscale, uh, hybrids, uh, together. 00:14:21.000 --> 00:14:34.000 So… so now, um… So that was kind of an introduction to, you know, prior art and what would, you know, from the people who are in the proposal, you always see the PI at the top. 00:14:34.000 --> 00:14:41.000 And, um, and, uh, what we're aiming to do, and now I'm sort of switching a little bit to what we're actually doing. 00:14:41.000 --> 00:14:45.000 Uh, so, um, so in terms of the… in terms of how to build. 00:14:45.000 --> 00:14:53.000 Or what… what a nanoscale hybrid should be, or what, what, uh, um… in practice, there's first some theoretical work. 00:14:53.000 --> 00:15:11.000 That, um, um… aims to understand Uh, what, uh, what kind of properties at nanoscale hybrid should have to make you know, these artificial materials, so… so what should be… how, you know, in terms of a… quantum, uh, quantum states. 00:15:11.000 --> 00:15:14.000 You know, what are the quantum states that carry out absorption? 00:15:14.000 --> 00:15:23.000 And what are the… and how does that, uh. Correspond to, then, uh, transaction into electrical signals. 00:15:23.000 --> 00:15:32.000 And, you know, in this work, it shows that it's possible, for example, to make a a detector that has, uh. 00:15:32.000 --> 00:15:44.000 Uh, spectral… spectrally resolved response with a… with a, you know, essentially 100% quantum efficiency within some uh, whether it's a frequency band. 00:15:44.000 --> 00:15:52.000 Um, by having, uh, you know, a suitable combination of absorbers, and then… and then states that carry out transduction. 00:15:52.000 --> 00:15:58.000 Um, so this is all, uh. Quantum mechanics and calculations. 00:15:58.000 --> 00:16:08.000 There are experiments that then suggest that, uh. Nanoscale hybrids made with 1D and 2D materials can actually achieve these properties. 00:16:08.000 --> 00:16:20.000 So this is a past work that's showing that, that essentially informs how we're doing things in this project. Uh, but basically, all this work shown here uses carbon nanotubes. 00:16:20.000 --> 00:16:28.000 That's the, uh, device that carries out the transaction to electrical signals, and then different devices to absorb light. 00:16:28.000 --> 00:16:39.000 And coupled to the carbon nanotes. Um, so then, in addition to theory, then there's modeling to actually, uh. 00:16:39.000 --> 00:16:45.000 To actually predict how devices will actually work once we make them, and how should we interface them to circuits. 00:16:45.000 --> 00:17:04.000 Uh, so this is the later talk by Jackie. Um… So the modeling tools that are being used and developed for this project are, I think, possibly of either quite, quite a, quite broad interest in meerkats, so… so that's why we chose to have Jackie. 00:17:04.000 --> 00:17:10.000 Focus on this presentation later, because of course, we can't do… we can't dive into every aspect of the project. 00:17:10.000 --> 00:17:14.000 So then, um… Sorry, I need something on there. 00:17:14.000 --> 00:17:26.000 So, um… So then let me talk a little bit about the ingredients and the assembly methods that we're developing. 00:17:26.000 --> 00:17:32.000 To actually make some prototype nanoscale hybrids to carry out these functions. 00:17:32.000 --> 00:17:43.000 Uh, so in terms of 1D and 2D materials. You know, we have work on, uh, lithographically defined TMD, transition metal dichot Cogginide. 00:17:43.000 --> 00:17:53.000 Alloids and heterostructures. Um, this shows a picture of a structure that has layers of of TMDs. 00:17:53.000 --> 00:18:02.000 Uh, we're using a carbon nanotubes. Um, as, uh, as devices that then we have to process further, and I'll talk about later. 00:18:02.000 --> 00:18:08.000 Uh, and in this case, trying to place them on HBM Flicks as a dielectric, as a gate dielectric. 00:18:08.000 --> 00:18:19.000 To have a very low noise, for single photon. Response, and also tellurium nanowires instead of carbon nanotubules, which can be lithographically defined. 00:18:19.000 --> 00:18:25.000 And also have, you know, similar properties, uh, for, uh, transfunction. 00:18:25.000 --> 00:18:33.000 So, there's also, for one, the N2D materials, it's… not only being able to produce them, but characterization is critical. 00:18:33.000 --> 00:18:39.000 And so we have an extensive, uh, extensive capabilities for… and expertise for characterization. 00:18:39.000 --> 00:18:45.000 And, you know, like I said, they're extensive, you can look here, Foundry Tools, just in one area. 00:18:45.000 --> 00:18:57.000 Uh, but let me just give one example. Um, that was done for… specifically for characterizing TMDs, which is the fabrication and design of of, uh, silicon nitride membranes. 00:18:57.000 --> 00:19:05.000 Uh, you know, where you can produce these, uh, TMDs through this process called lateral conversion. 00:19:05.000 --> 00:19:12.000 Synthesis, which is what we're using. But then also, uh, analyze them with transmission electron microscopy. 00:19:12.000 --> 00:19:30.000 Uh, thanks to the thin membrane. And so there's a lot… this… this kind of characterization allows one to see what happens you know, with the precursors, and then after the After the, uh… the conversion, uh… and understand, really. 00:19:30.000 --> 00:19:48.000 Uh, things like effect of temperature. Okay, so then we also need CMOS substrates, because everything we're… when we actually make these nanoscale hybrids, we have to read them out somehow, and… to be able to talk to them, and this is all gonna be done with a CMOS backend. So this is a CMOS plus X. 00:19:48.000 --> 00:19:58.000 Development, let's say, so we need… Uh, you know, our own CMOS substrates, so we have, so far, produced Um, in the… in the previous project. 00:19:58.000 --> 00:20:08.000 Uh, which was called, uh, nanoscale hybrids, and… nanoscale sensors on CMOS, Uh, we, uh, produced wafers. 00:20:08.000 --> 00:20:15.000 On, uh, TSMC 130 nanometers. And on Skywater, 130 nanometers, two different paths. 00:20:15.000 --> 00:20:33.000 Um, and, uh, managed to get, in both cases, the required polarity of the surface for integration of nanomaterials, which is of order 1 nanometer Um, and uh… and uh, we can now use these devices in this project, these, uh, these back-end CMOS chips. 00:20:33.000 --> 00:20:53.000 And we're additionally starting to design Uh, for 40 nanometer TSMC through the DoD Microelectronics Commons project that will also, uh, help us produce a… Yeah, another way for more… features that we can use for these, uh, for these, uh, for this new project where we wanted 00:20:53.000 --> 00:21:01.000 Try processing as well as develop processing, or in hybrid processing as well as sensing. 00:21:01.000 --> 00:21:07.000 So then, of course, we need parenting and imaging. Again, we have lots of capabilities there. Here are some examples. 00:21:07.000 --> 00:21:15.000 Uh, of, uh, planning, uh, trenches, these horizontal lines are trenches, there's a… a close-up view. 00:21:15.000 --> 00:21:25.000 And I'll… again, and then on the… on crossing the trenches at a right angle are carbon nanotubes, and here you can see the individual tubes. 00:21:25.000 --> 00:21:37.000 They're very straight and parallel, in this case. Um, and uh… And then this… you know, this is a prototype using just a blank substrate. 00:21:37.000 --> 00:21:50.000 Uh, but then we want to do this on a CMOS chip, and there's a picture, or an AFM of a… of a trench on a surface of a SEMA ship where it doesn't look as good yet, but we hope it will look better in this 40 nanometers process. 00:21:50.000 --> 00:21:56.000 And, um, and then once we have carbon nanotubes on here, we can connect them. 00:21:56.000 --> 00:22:01.000 To the chip below, and we can also make here… or we can add lithographically. 00:22:01.000 --> 00:22:10.000 The lunar nanowires is a… picture that shows this, this skinny line here is a nanowire, and then there's an electrode and a readout. 00:22:10.000 --> 00:22:15.000 Uh, or let's say, interconnect pad to connect it to the chip. 00:22:15.000 --> 00:22:28.000 Uh, so… so this is all work towards making some demonstrator that is effectively will behave like this, um… photodiode, uh, that distinguishes frequencies, as I mentioned before. 00:22:28.000 --> 00:22:44.000 So in this case, we can have some carbon nanotubes sensitive to one set of frequencies on another uh, set to another frequency, so we could… we could be, um, you know, more… Uh, we could have more bins than just plus or minus, as in the case of this… of this photodiode. 00:22:44.000 --> 00:22:49.000 And of course, in a much… at a much smaller scale. 00:22:49.000 --> 00:22:57.000 Okay, so then… so then, uh, we have to do some… something more, so in this case, if we just put in carbon nanotubes. 00:22:57.000 --> 00:23:04.000 Is not enough, the carbon had to do with it. Device that does the transaction, it's not in itself a nanoscale hybrid. 00:23:04.000 --> 00:23:10.000 We have to add stuff to it, so we need to add in particular things that absorb light of the required wavelength. 00:23:10.000 --> 00:23:16.000 And then for that, we can use quantum dots. And so the question is, how do you put quantum dots on a carbon nanotube? 00:23:16.000 --> 00:23:24.000 And, um, and that's where the trench comes in. So this… the fact that we have the carbon nanotube is a bridge crossing a trench. 00:23:24.000 --> 00:23:28.000 Allows access from… all around the carbon nanotube. 00:23:28.000 --> 00:23:33.000 And that allows to wrap DNA around it. And then, uh, in this process of, uh. 00:23:33.000 --> 00:23:46.000 The decoration with DNA, self-assembly, we have DNA wrapped around the carbon nanot And then quantum dots that have been functionalize with DNA tails. 00:23:46.000 --> 00:23:53.000 Uh, and then they can… those quantum dots will then attach selectively, because we use the right DNA sequences. 00:23:53.000 --> 00:24:08.000 Uh, to the… to the carbon nanotube. Where the trench is, so we go from this… a bunch of quantum dots and a carbon nanotube to a carbon nanotube with the quantum dots on it, and there's a… picture of what that would look like. 00:24:08.000 --> 00:24:14.000 Um, I don't know what that actually looks like, and this is an actual carbon nanotube with the DNA on it. 00:24:14.000 --> 00:24:32.000 Um, so… so then, you know, that's another thing we can do with self-assembly, is then place things on a CMOS chip that are, you know, very small, so instead of building Um, uh, nanoscale hybrids directly on that chip, we can build them in solution, let's say. 00:24:32.000 --> 00:24:44.000 Uh, and then place them on the chip with this directed self-assembly, which uses this technique of peptide brushes, which can be patterned on the surface lithographically. 00:24:44.000 --> 00:25:01.000 But then, wherever you pattern them. Uh, you can then… you can select, you know, you can, again, use the right sequences so that a particular DNA Um, uh, sequence will attach to a particular peptide And then you can… we can attach this DNA origami, so-called DNA origami. 00:25:01.000 --> 00:25:21.000 These are breadboards made of DNA, or yummy, so… so there's a lot of DNA strands in here, but it's basically a flat little patch of DNA, which has… which has selective attachment to some of these Peptides. On top of it, it can carry a circuit, carry a nanoscale hybrid. 00:25:21.000 --> 00:25:29.000 Uh, with carbon nanotubes and quantum dots, etc, and then can attach, you know, like, the green ones could attach here, and the blue ones could attach somewhere else, for example. 00:25:29.000 --> 00:25:36.000 So, so that's… this is a demonstration of, uh, of a pattern of these, uh, of these patches. 00:25:36.000 --> 00:25:46.000 Attached in specific locations of a substrate. So, so that, that's it for the… very quick introduction on what is it we're working on. 00:25:46.000 --> 00:25:56.000 What we would like to do, and so now I'd like to… send it over to Jackie to explain more about the modeling and simulation tools. 00:25:56.000 --> 00:25:58.000 Sure, thanks, Maurice. Let me sh… okay. 00:25:58.000 --> 00:26:04.000 I'll stop sharing, yeah. I can figure out how to… okay, stop. 00:26:04.000 --> 00:26:05.000 There we go. 00:26:05.000 --> 00:26:11.000 Okay, great. Uh, desktop, I'm sharing my entire desktop, but I want to show my slides. 00:26:11.000 --> 00:26:16.000 Presenting mode. Is that visible? Great. Okay. 00:26:16.000 --> 00:26:17.000 Yes, yes. 00:26:17.000 --> 00:26:25.000 Alright, um, I'm hoping to finish within 10 minutes so that we can have some time for questions and discussions. 00:26:25.000 --> 00:26:39.000 But, um, so yeah, thanks, Maurice, for highlighting the modeling effort. So basically, I would like to give an overview of the microelectronics simulation packages that we have been built over the past years. 00:26:39.000 --> 00:26:47.000 And we're hoping that these packages can serve as a tool that is available to all of the projects within MSRC. 00:26:47.000 --> 00:26:59.000 And a lot of work are, um, in collaboration with the nanohybrid team and other microelectronics team. And you can get access to the open source packages through this GitHub link. 00:26:59.000 --> 00:27:05.000 I will copy and paste this link at the end of my presentation in the chat. 00:27:05.000 --> 00:27:24.000 Okay, so here is an overview of what physical mechanisms we're including as of now. We have several packages named Artemis Electronics, Ferrox, and Magnex. I'll introduce them one by one in the following slides. 00:27:24.000 --> 00:27:32.000 Basically, we're taking care of electromagnetics, whether in the static or dynamic fashion. The ferromagnetic materials. 00:27:32.000 --> 00:27:39.000 So that gives us the switch of the magnetization, the spin wave propagation, and the quantized magnon interactions. 00:27:39.000 --> 00:27:54.000 Ferroelectric material that enables non-volatile memory functionality, quantum transport that include electron transport in the carbon nanotube, and we're working toward enabling it for other materials and also phonon transport. 00:27:54.000 --> 00:28:01.000 And finally, the superconducting, the classical superconducting theory represented by London equation. 00:28:01.000 --> 00:28:07.000 That allows us to simulate the electric signal flowing in the superconducting material and circuits. 00:28:07.000 --> 00:28:12.000 The key features of our modeling tool is that they are flexible. 00:28:12.000 --> 00:28:32.000 Portable, open source, and massively parallel. So I will introduce the details one by one. One, um… key point I would like to emphasize is that our device package modeling tools sit in between material physics and circuits and architecture applications. 00:28:32.000 --> 00:28:47.000 In another word, we… look into the junction physics and device physics. We solve continuum form of the PDEs to predict the physical interaction in the system. So we do not work on first principle material. 00:28:47.000 --> 00:29:04.000 We get the output of the DFT simulations, uh, which is that they predict the material properties and hand it over to us. We use the material properties at the input into our model, and the output of our model is the circuit response. 00:29:04.000 --> 00:29:12.000 Uh, for example, the device's IV curve and the transistor's transconductance. So this is where we sit. 00:29:12.000 --> 00:29:18.000 I just want to make it, uh, clear in case you get interested in using some of our models. 00:29:18.000 --> 00:29:47.000 All right, so I mentioned that our code packages are portable, meaning that Um, because we're fortunate to leverage the access scale computing product, AMRX, our code packages are functional on different GPU and CPU platforms. If you switch from ProMuter to Frontier, um, to Aurora, you don't have to rewrite the code, uh, like some of other packages do. You can just 00:29:47.000 --> 00:30:12.000 Use the model and ran the simulations. Um, it's accessible, meaning that it's open source, everyone can download And use the package, and if you have questions, feel free to reach out to us. It's also… got flexibility in the algorithm, meaning that even though we only include this, um, packages and the associate physical mechanisms. 00:30:12.000 --> 00:30:24.000 Now, if there are interests to implement new physical mechanism, and the additional coupling, then we're happy to do that. 00:30:24.000 --> 00:30:35.000 And we use time domain algorithms that are suitable for nonlinear problems, because nowadays, a lot of these problems are nonlinear and highly coupled. 00:30:35.000 --> 00:30:45.000 We've got almost ideal GPU and CPU scalability, um, because we're fortunate to leverage the MRX ECP product. 00:30:45.000 --> 00:30:52.000 So, this is an example of the weak skilling efficiency of one of our packages. 00:30:52.000 --> 00:31:02.000 As you can see that up to a thousand of GPUs, we're seeing a flat efficiency, meaning that there's no overhead runtime from the GPU or CPU communication. 00:31:02.000 --> 00:31:09.000 So, this package is… it's very suitable for, um, especially for larger problem simulations. 00:31:09.000 --> 00:31:28.000 Alright, so at this point, I've overviewed the features and the basics of the package. Uh, now let me go through them one by one, so that you know, beyond just the fancy words and you can understand what's going on behind the scene, and what equations we're solving. 00:31:28.000 --> 00:31:38.000 The first package I'd like to introduce is called Artemis Adaptive Mesh Refinement Electromagnetic Solver. We solve Maxwell's equations in the fully dynamic form. 00:31:38.000 --> 00:31:52.000 Um, and or with the coupling to the micromagnetic LOG equation listed here, um, the superconducting classical London equation with the extra superconducting flow included. 00:31:52.000 --> 00:32:04.000 Or the ferroelectric Ginsburg-Landau equation. We have implemented the Ginsburg-Landau in electrodynamics. We have to validate it before we, um, uh. 00:32:04.000 --> 00:32:21.000 Use it for more applications. But this is the backbone of Artemis. And here I'd like to show two application examples. First, on the left is a hybrid quantum circuit. We basically have a co-planar waveguide resonator. 00:32:21.000 --> 00:32:39.000 And on top of that, we have a ferromagnetic materials. Which is the cobalt iron material. So if we excite excitation of both the CPW resonator, which is photon And the ferromagnetic resonance, which is the magnum. 00:32:39.000 --> 00:32:51.000 And we tuned them so that they're in the strong coupling region. We're able to get this anti-crossing spectrum feature. And this is from Artemis Maxwell plus LRG simulation. 00:32:51.000 --> 00:32:58.000 Then, uh, we run the Maxwell plus London simulation. And simulate this quantum chip. 00:32:58.000 --> 00:33:05.000 This is a simplified view, because the complete view is still proprietary. 00:33:05.000 --> 00:33:13.000 Um, but this is a nice showcase of Artemis' capability of simulating a larger-scale quantum circuit. 00:33:13.000 --> 00:33:36.000 Next, um, electro… electronics. Um, this is for non-equilibrium quantum transport, and this is with, uh, this is… achieved, uh, with collaboration, uh, with Francois, mostly. Um, so what this model does is it solves the electrostatic Poisson equation, which is simplified from the four dynamical Maxwell's equation. 00:33:36.000 --> 00:33:43.000 And we solve the quantum transport represented by the non-equilibrium Green's function method, the NEJF method. 00:33:43.000 --> 00:34:08.000 And of course, we have to solve this two equation systems in a self-consistent session. Um, here is one example of the model application. We have carbon nanotubes, either in the fully aligned structure or misaligned structure. And then we run, um, the NEGF simulation and post-process computed 00:34:08.000 --> 00:34:30.000 The transconductance here. Um, to our best knowledge, this is the one-of-a-kind non-equilibrium transport model that can accurately simulate the misaligned carbon nanotubes. So as you can see, the misalignment has a significant impact in the transconductors. 00:34:30.000 --> 00:34:38.000 It causes a 12% um, difference, uh, if we have different kind of alignments. 00:34:38.000 --> 00:34:48.000 Different kinds of, uh, different level of misalignments. Um, yeah, so this is, uh, the electronics code package for non-equilibrium. 00:34:48.000 --> 00:35:02.000 Quantum transport. Currently, we have carbon nanotube, but we're actively working toward generalizing it to more materials and more kind of particles, like phonons. 00:35:02.000 --> 00:35:08.000 Next, I would like to introduce Ferrox, that's a face-filled module for ferroelectric devices. 00:35:08.000 --> 00:35:32.000 Here shows the mathematical model of Farrow X. Uh, we solved the three equation systems in a self-consistent way. First is on the left, we have the Ginsburg-Landau equation in the time domain that is describing the evolving of polarization in the ferroelectric layer. Then we have the electrostatic Poisson equation, which is pretty similar to electronics. 00:35:32.000 --> 00:35:58.000 Um, and then we have the, um, the, um, carrier calculate… the carrier density calculation in the semiconductors. Uh, we have both the Fermi direct distribution and the Maxwell-Bosman distribution. We also have the drift-diffusion career transport implemented in this, in FaroX package. 00:35:58.000 --> 00:36:24.000 Um, I would like to show the, uh, negative capacitance field effect transistor applications to showcase the capability of FarrowX. So the first figure here shows the polarization evolving at a function of time. Then the lower panel in the middle here shows The Faro X simulated capacitance enhancement effect, which is one of the signatures of negative capacitance. 00:36:24.000 --> 00:36:35.000 So, as you can see that, um, with different portion of different, uh, different ratio of the mixed face. 00:36:35.000 --> 00:36:42.000 Here, we can have different level of capacitance enhancement. Uh, if I… if I reiterate. 00:36:42.000 --> 00:36:56.000 Here, in the middle, we have the polarization color map, the domain pattern of the polarization, only within the orthromic phase. And surrounding the orthorhombic phase, we have the trigonal phase, which is nonpolar. 00:36:56.000 --> 00:37:07.000 So, with the depolarization effect coming from the O-Face and T phase mixed up, we can get different levels of capacitance enhancement Depending on the grain size. 00:37:07.000 --> 00:37:21.000 And here shows the GPU scaling up to, um, this is permuter node numbers, if you multiply by 4, we get the actual GPU numbers. So up to about 500 GPUs, we get pretty perfect weak scaling performance. 00:37:21.000 --> 00:37:29.000 Meaning that we get no overhead between, uh, no overhead coming from the GPU communication. 00:37:29.000 --> 00:37:43.000 Last but not least, the Magnex micromagnetic module for magnetic devices. We saw the allergy equation, the Ginsburg-Landau, uh, sorry, the lift… the Landau lift-Shit-Stilbert equation. 00:37:43.000 --> 00:37:54.000 And we have all of the effective magnetic field included. Um, and those could either be fully dynamical field or quasi-static field. 00:37:54.000 --> 00:38:10.000 I'm going to just very briefly introduce the scaling, uh, studies, the cutting-edge spatial temporal numerical discretization. Uh, we also validated against the very standard MIUMAC problems. 00:38:10.000 --> 00:38:26.000 Then, uh, this is one demonstration, uh, how we use MagnaX to validate experimental data. So here we have a spin cycloid texture in one kind of material, which is the bismos ferrite, BFO material. 00:38:26.000 --> 00:38:42.000 We have the texture set up, and then we deposit spin wave from one end of the material, and that let the spin wave propagate through either along the cycloid direction or parallel to the uniform direction of the cycloid. 00:38:42.000 --> 00:38:51.000 Then we post-process and get the spectrum performance of the Magnum. So here shows the spin moment. 00:38:51.000 --> 00:39:15.000 Cone angle as a function of the frequency, and we get the resonance at different frequencies. Then, we calculate the, uh… the transport efficiency of the magnolon along those two different directions, and you can see that our experimental result falls in the range of the numerical simulation. 00:39:15.000 --> 00:39:22.000 Um, and this is, uh, agreeing with the BFO, the bismuth ferrite material properties. 00:39:22.000 --> 00:39:38.000 So we're… we're upgrading the next functionality into more advanced materials, like anti-ferromagnetic material, with more mechanical coupling so that we can model magnetoelastic material. 00:39:38.000 --> 00:39:59.000 Okay, with that, I would like to thank you for your attention, and again, if you're interested, you're welcome to check the package, and if you have questions or comments, feel free to reach out to Andy or myself Thanks. 00:39:59.000 --> 00:40:08.000 Okay, so that concludes our presentation, and hopefully we have some time for discussion or questions. 00:40:08.000 --> 00:40:09.000 Yeah, I'm happy to share other questions for Maurice and Jackie? 00:40:09.000 --> 00:40:13.000 I don't know if… Paul, if you want to share the… 00:40:13.000 --> 00:40:15.000 Or for other members of our team. 00:40:15.000 --> 00:40:38.000 Or other members of the team, yeah. First of all, thanks for the presentation. I think it was very thought-provoking. I have a question, um… uh, related to the, you know, the kind of design principles for a nano-hybrid… a nanoscale hybrid, um. 00:40:38.000 --> 00:40:54.000 Uh, structure. You've shown, Maurice, that you're… through the… the team of PIs you have, you've got access to a variety of different nanoscale 1D and 2D semiconductors, there are… you know, various kinds of transduction and memory. 00:40:54.000 --> 00:41:10.000 Possibilities, uh, is there a particular… Um, is there a particular kind of, uh, nano-scale hybrid that the team feels most interested in starting with? I mean, you've got all these interesting building blocks. 00:41:10.000 --> 00:41:18.000 Uh, is there a way that you're thinking about prioritizing which ones you put together Uh, and, um, and for what? 00:41:18.000 --> 00:41:20.000 For what purpose? 00:41:20.000 --> 00:41:28.000 Yeah, well, the more advanced work is in integrating carbon nanotubes with quantum dots. 00:41:28.000 --> 00:41:33.000 Uh, and the idea there is, uh, basically to have different. 00:41:33.000 --> 00:41:39.000 So these are single-carbon nanotube is the goal, so I have a single carbon nanotube device. 00:41:39.000 --> 00:41:48.000 That has quantum dots on it, and then the quant… the… that would be read out like a transistor, so would I get a drain source on the carbon nanotube. 00:41:48.000 --> 00:41:53.000 So that… so that we're… we're, you know, we're moving along, uh, trying to get that. 00:41:53.000 --> 00:42:05.000 Or making prototypes of that that… you know, we're pretty close to having you know, a single… a prototype just for a single color of quantum dots, but the goal is to have, like. 00:42:05.000 --> 00:42:18.000 Two or three colors on the same pixel, so different… same carbon nanotube, different quantum dots on each one, right? So 3… let's say 3 tubes with… One with green dots, one with red dots, one. So that's… that's kind of our… 00:42:18.000 --> 00:42:34.000 Currently, most… the thing we've done the most work on. Uh, but, you know, the… The issue with carbon nanotes is that they're hard to deal with, so they're basically, especially individual tubes, so they're… so we need lots of, uh. 00:42:34.000 --> 00:42:39.000 Gymnastics to place them, connect them, decorate them, right? Sorry, someone. 00:42:39.000 --> 00:42:44.000 Yes. Gregoris has a question, I think, or a comment. 00:42:44.000 --> 00:42:45.000 You know, I think he'll comment. 00:42:45.000 --> 00:42:52.000 No, I… why should I be commenting? Notice, Daki, thank you very much for the presentation. 00:42:52.000 --> 00:43:19.000 I have many questions regarding the package simulation packet, and maybe proportion between your goal is for… to fabricate and to simulate. So is… Is your goal to still develop the simulation package, to treat this as deliverable in your in your project, or as a tool to simulate and fabricate. So maybe this is the kind of follow-up question on what Paul requested, and because it's going to be heading towards 00:43:19.000 --> 00:43:29.000 You know, the preferred structures, preferred methodologies. Or maybe preferred even technologies, yeah? 00:43:29.000 --> 00:43:32.000 Okay, so I guess Jackie should take that. 00:43:32.000 --> 00:43:37.000 Um, I… let me… let me clarify if I understand the question. So the question is. 00:43:37.000 --> 00:43:50.000 Do we plan to continue developing these code packages? Right. We definitely plan to continue developing this code package, right? Within Nanohybrid. 00:43:50.000 --> 00:44:19.000 Two of the packages are of special interest. One is Artemis, which is the electromagnetic Um, solver, because within nanohybrids, we're interested in looking into this optical computing and photonic effects. The second package is the electronics package. That is the non-equilibrium quantum transport, because we have continued interest in carbon nanotube and interactions between electrons and photons and possibly phonons. 00:44:19.000 --> 00:44:37.000 So, yes, we're definitely interested in keep developing those, uh. Packages. But I think, um, it… it does not conflict with the interest of fabrication, right? So our goal is Um, at the end of the project, we can deliver a… 00:44:37.000 --> 00:44:49.000 Self-consistent modeling plus fabrication tool packages. Um, and this is actually, uh, continuing the initial interest of co-design. 00:44:49.000 --> 00:44:58.000 So, I think, yes, there's continued interest, but that does not mean that we will not… we're not interested in fabricating the device. 00:44:58.000 --> 00:44:59.000 Is that what you're asking? Okay. 00:44:59.000 --> 00:45:02.000 Effectively, yes, you answered my question. 00:45:02.000 --> 00:45:19.000 So, since I know Gregor's, maybe I can add something, is that one of the aims eventual aims is to actually… is to use the… one of these pipes… so there's not… not directly integrate, but use one of these packages together with spice simulation, or… so basically… 00:45:19.000 --> 00:45:24.000 To model the entire circuit of the nanoscale hybrid on a CMOS chip. 00:45:24.000 --> 00:45:33.000 And, uh, you know, and for example. To make amplifiers that have a carbon nanotube as the input transistor. 00:45:33.000 --> 00:45:46.000 Right? And so then… then you would need the results of the of the simulation tools to tell you what the transitor is doing in order to build a model for the Or to use in spies, basically. 00:45:46.000 --> 00:45:51.000 Mm-hmm, mm-hmm. Thanks a lot. 00:45:51.000 --> 00:46:01.000 Any further questions? 00:46:01.000 --> 00:46:16.000 I guess I have one question that's maybe a little bit more on the mechanics of how your team works together. What is the… sort of nature of the meetings that you have, and uh… and, you know, how are you… how are you kind of 00:46:16.000 --> 00:46:18.000 Structured, I guess. 00:46:18.000 --> 00:46:38.000 Yeah, we're basically working as one group, we don't really have… we have a bit of a… you know, a bit of side meetings for a specific specific things, like technical details of, you know, placing, um, quantum dots on something, or technical details of lithography, and so on. 00:46:38.000 --> 00:46:49.000 But mostly we have a weekly general meeting that we all attend, and we kind of… percent, um, to each other the progress of the different areas. Of course. 00:46:49.000 --> 00:46:57.000 Sometimes a single meaning isn't enough, but we just move to the next, you know, move the material to the next week, or sometimes we don't have enough. 00:46:57.000 --> 00:47:04.000 But basically, yeah, we just have a… One, we work as one group with one meeting, as far as syncing up every week. 00:47:04.000 --> 00:47:11.000 And then in between those weekly meetings, there's kind of side meetings of a few people doing the specific work. 00:47:11.000 --> 00:47:15.000 But we always come together once a week to sort of sync up. 00:47:15.000 --> 00:47:16.000 Yeah, thank you. 00:47:16.000 --> 00:47:19.000 We also have the steering meeting, Maurice, between the 00:47:19.000 --> 00:47:24.000 Yeah, right. So then we, yeah, we have a steering group. 00:47:24.000 --> 00:47:29.000 A monthly steering group meeting. For the, kind of, more, you know. 00:47:29.000 --> 00:47:40.000 Discuss, sort of, strategic strategy, and are we meeting our goals, and… Well, you know, on reporting and… hiring, and so on, but… 00:47:40.000 --> 00:47:43.000 Okay, very good. Pops. 00:47:43.000 --> 00:47:49.000 Alright, thanks. I really enjoyed that presentation. Um, I was curious about two things. 00:47:49.000 --> 00:48:01.000 Um, there have been a few people who've been working on aligning… aligning the depositions of the carbon nanotubes. Is your team leveraging Some of that expertise, or at least some of that, you know, know-how, such as it is. 00:48:01.000 --> 00:48:09.000 And the second question is, when you were functionalizing the DNA and getting the three colors that you were describing, you know, the green. 00:48:09.000 --> 00:48:24.000 Blue, and so on, uh, it seemed to me that the number of dots that land might be a bit stochastic. Are you planning on flooding the carbon nanotube so that, you know, such variations Turned out to be, uh… control, you know, not… not material to… 00:48:24.000 --> 00:48:34.000 Yeah, let me first ask… answer the first part, and then Greg, who's connected, can talk about all the Headaches he has with quantum carbon densities. 00:48:34.000 --> 00:48:50.000 But for the aligned tubes, we actually used a vendor. That had this… developed this proprietary process to… to place carbon nanotubes, and they just transferred them from From a fabric… from a substrate where they fabricate them to, uh, to the chip of interest. 00:48:50.000 --> 00:48:56.000 As a… as a big patch of aligned tubes. Uh, so they did that for us. 00:48:56.000 --> 00:49:02.000 So, we hope to continue, you know, making use of that kind of process, outside process. 00:49:02.000 --> 00:49:03.000 Thank you. 00:49:03.000 --> 00:49:04.000 Maybe, uh, maybe I can… 00:49:04.000 --> 00:49:12.000 And I can answer the second part of your question, um, the number of quantum dots we get per carbon nanotips is definitely stochastic. 00:49:12.000 --> 00:49:23.000 As you acutely noticed, um, and… Well, according to Francois simulation, you know. 00:49:23.000 --> 00:49:30.000 It doesn't really affect the efficiency too much, as long as there are some… a few quantum dots on the carbon nanotube. 00:49:30.000 --> 00:49:44.000 Uh, but yeah, we would like to have them more uniform, but, you know, this is the nature of attachment of DNA to carbon nanotubes. It's not… covalent, it's, uh, based on the PyPi stacking. 00:49:44.000 --> 00:49:50.000 Uh, with the carbon nanotube, and it's very hard to control. 00:49:50.000 --> 00:49:51.000 But, you know… Yeah. 00:49:51.000 --> 00:49:55.000 But there's problems first. No, but thank you, obviously. It's a good first step, thank you. 00:49:55.000 --> 00:49:59.000 Thank you. 00:49:59.000 --> 00:50:07.000 Great. Further questions, or Observations? 00:50:07.000 --> 00:50:12.000 Okay. Nope. Doris, I guess, another… Question? 00:50:12.000 --> 00:50:20.000 Very quick question about manufacturing. Are you planning to use in-house capabilities, or to develop capabilities. 00:50:20.000 --> 00:50:31.000 Or, you know, use some external capabilities, commercial-grade foundries. Can you maybe comment on this? Because, I mean, we know that manufacturing is coming at high price, generally speaking, and this is… 00:50:31.000 --> 00:50:38.000 Yeah. Yeah, so the answer is both. So… so, for example, of course, for the CMOS wafers. 00:50:38.000 --> 00:50:55.000 You know, fabrication, that's all done at foundries. Uh, as you know, but we've been lucky to take advantage of some initiatives, like the than this nano… nanotechnology accelerator that you… that produced the… wafers at, uh, Skywater, you know, that was free to us. 00:50:55.000 --> 00:51:04.000 Uh, because… and it was a pilot program, and then the, um… Uh, and then the one that is upcoming through the, uh. 00:51:04.000 --> 00:51:09.000 Michael Johnny's comments, TOD, Michael Electronic Commons. Also, we just pay for the waiver, but not for the masks. 00:51:09.000 --> 00:51:15.000 So, um, because it's a multi-project thing, uh, sponsored by that. 00:51:15.000 --> 00:51:31.000 Effort. Um… So, uh, so… so then for the other parts, for the… post-processing, let's say, for example, if we need to planarize and do chemical mechanic collagen or wafer coring, you know, we send all that stuff out. 00:51:31.000 --> 00:51:42.000 Um, the e-beam lithography will do in-house. Uh, so the, you know, regular lithography, also we can do in-house, so depositing palladium. 00:51:42.000 --> 00:51:50.000 For example, contacts for carbon nanotubes, and then And then platinum electrodes on top of the palladium, all that's done in-house. 00:51:50.000 --> 00:52:06.000 The tellurium deposition for making tellurium nanowires also in-house. As all the TMD, the TMD fabrication is all at the family. 00:52:06.000 --> 00:52:08.000 All right, Angelo. 00:52:08.000 --> 00:52:31.000 Yeah, maybe… maybe it's a little tricky question to see on the first one, or presenting, but I was wondering… Um, you know, when you wrote your, uh, proposal, did you imagine additional capabilities that you would like to have in this project, and that you could not include, and that you may think maybe at this stage that some other of the teams within 00:52:31.000 --> 00:52:38.000 The center could potentially have, and that we… Could that be a project? Is there…? 00:52:38.000 --> 00:52:44.000 Yeah, certainly there's… there's a lot of things that are… you know. 00:52:44.000 --> 00:52:57.000 We don't, you know, haven't done before, and we figure out how to do it, and it takes a while, and… If someone already has a process for that, it works, but they're typically detailed things. I don't think I have a list handy, but… 00:52:57.000 --> 00:53:03.000 Yes, absolutely. I mean, just like the carbon nanotube, a line carbon nanotube, like I said, we went to a vendor. 00:53:03.000 --> 00:53:09.000 But if someone in the project is doing something like that, we would definitely be very interested. 00:53:09.000 --> 00:53:19.000 Uh… you know, yeah. 00:53:19.000 --> 00:53:22.000 All right, we have a few minutes left. Any, uh, any last questions? 00:53:22.000 --> 00:53:38.000 Maybe… maybe someone else wants to comment on that, I don't know, Archana, or… Or, um, Francois, you guys are connected, so… so again, capabilities that we would like to make use of, maybe it'd be good to have a list of that. 00:53:38.000 --> 00:53:43.000 I think, uh, maybe there was… Recently, we talked about, um. 00:53:43.000 --> 00:53:57.000 Uh… sputtering of, uh… Polymers, what was it? Uh… sort of, uh, spiral deposition of, uh, from methane gas, or something like that. 00:53:57.000 --> 00:53:58.000 Was that Archana, do you remember? 00:53:58.000 --> 00:54:12.000 Yeah, yeah, actually, um… just maybe I'll, uh… So one of the things that we've been trying to do, uh, is prepare a real monolayer of quantum dots on, um. 00:54:12.000 --> 00:54:28.000 Different substrates, including carbon nanotubes and 2D materials. Greg's approach, uh, is one way to do it, but for… We've also been trying to implement some kind of substrate functionalization. 00:54:28.000 --> 00:54:34.000 Um, we've had success in the past with polymerizing methane on, um. 00:54:34.000 --> 00:54:49.000 Silicon oxide. Uh, but they don't let us do that in our… In the… in the Foundry tool anymore, so in case, uh… you know, anyone has a tool where we can, um… I'll, you know, stick a wafer in and 00:54:49.000 --> 00:54:56.000 Polymerize… to create, like, you know, few nanometers of polymer methane on top. 00:54:56.000 --> 00:55:07.000 We found that to be a very good substrate to create a real monolayer of quantum dots. Usually, you can create, like, an ordered array of, like, you know, multi-layers of quantum dots. 00:55:07.000 --> 00:55:12.000 Or you can create, like, these individual dots with extremely dilute solutions. 00:55:12.000 --> 00:55:25.000 And the reason we want to create this true monolayer is because we're seeing, um, really exciting results with, um, in terms of… excitation transport being enhanced at that limit. 00:55:25.000 --> 00:55:35.000 Uh, which is… which is to say, it's one of the mechanisms in Francois' theoretical framework. It's one of the tuning knobs. You know, how do you get the information from the light? 00:55:35.000 --> 00:55:41.000 To your nano-hybrid Center for Dissociation, the Center of Interest. 00:55:41.000 --> 00:55:58.000 And controlling that. So this is just, um, yeah, I can just drop the paper that we've been following that Traditionally, we could do it at the foundry, but… Um, they turned out to be quite a dirty process that was interfering with other, uh, more delicate processes that are 00:55:58.000 --> 00:56:10.000 Going on, so if you guys have a janky… Plasma… plasma, where we can run some methane safely. 00:56:10.000 --> 00:56:11.000 Yeah, a joint list of capabilities. Yeah. 00:56:11.000 --> 00:56:19.000 I mean, I think it would be very nice to have a… some sort of a… room or something, or even… but some way that someone in the center can say, hey. 00:56:19.000 --> 00:56:20.000 Yeah. Should we start, like, a center Slack channel or something? 00:56:20.000 --> 00:56:25.000 I need this, does anyone have it? And someone can reply, so… Or something, yeah, we should have. 00:56:25.000 --> 00:56:35.000 Yeah, I don't know… I think that would be really good, so that, ultimately, I feel like We should be able to post something in the postdocs, or students, I think that's where the real… lubrication needs to happen, so we've been, like, reaching out to different clean rooms, and… 00:56:35.000 --> 00:56:41.000 Yeah. 00:56:41.000 --> 00:56:51.000 I think we found something in kernel based on Francois' postdoc suggestion, but it's unclear how all of the logistics… ideally, if we could keep it within the center, that would be nice 00:56:51.000 --> 00:57:13.000 Mm-hmm. That's a great suggestion, thank you. Well, we've reached the hour, and we should be respectful of everyone's time. Thanks to Maurice and Jackie and the whole team, the Nanoscale Hybrids team, for being our lead-off presenters, and uh… we'll pick this up again next week, and uh… see everyone, hopefully. 00:57:13.000 --> 00:57:15.000 Everyone on the call, maybe a few more people then. Have a good one. Bye. 00:57:15.000 --> 00:57:21.000 Thank youThanks