 Good morning from California. I'm Will Chu. I'm the faculty director of StorageX Initiative and a professor of material science and engineering at Stanford University. I'd like to welcome everyone to today's StorageX symposium. So today we have a focused discussion on a topic that received a lot of attention and is receiving a lot of attention, which is solid state batteries. And I'm very delighted to be hosting two of my colleagues from MIT to speak about the many aspect of material design, manufacturing, and also the availability of raw materials for solid state battery technologies. Our first speaker is Professor Jennifer Rupp, whom I will introduce just in a moment. And our second speaker is Professor Alsa Elevedi, as I mentioned, both from MIT. So I am extremely delighted to be joined by my longtime friend and colleague, Jennifer Rupp. Welcome, Jennifer. As you reminded me, we have known each other for a very long time, for nearly 20 years. And I just wanna say, Jen is a true delight to be with in all settings, professional and otherwise. And she is an innovator in the area of ceramics material. She has worked on a lot of different things. When I first met her, she was working on thin film fill cells. Then every few years she would reinvent herself. Then she was working on oxides, memory devices, memristors, and more recently, she has got it into energy storage. And the common theme is really ceramics, whether it's nano or macro, whether it's thin film or bulk, she's really a master of manipulating ceramics at all levels. And today we're really delighted to have her give an in-depth talk on solid state batteries with a focus on synthesis and manufacturing for something that will have to attack a problem of enormous scale. So Jennifer, I'm really delighted to be hosting today. Thank you for joining us. The stage is yours. Thank you so much for this very kind introduction, Will. And also I would like to thank you and stand forward for inviting me here today. And I'm really pleased to be virtually in sunny California today, thanks to Zoom. So here's a title of my talk. Today I will speak about design and manufacture of solid state batteries to its low cost. So this, when we think about the world, actually what motivates us and we think about the world actually, what motivates me most in thinking about new ceramics and what they can do actually in energy storage is to think about what is the best way to have cheap and accessible energy storage for a really wide section of the country. And of the countries and also actually in the world. So what I really like here is to show a map where we can actually see the climate change projection for 2017. And what we can see here on this map is that colored in red, you can see the regions that are actually most prone to be affected by climate change. So when we analyze this map really carefully, then we can see that actually especially in South America and in Africa and also in Asia, there will be a huge effect under 2070 of climate change. And in this little black dots, and I think that's quite impressive, you can see 67 cities that will have a new and unpresented climate like it's nowhere existing actually right now on Earth. And I think this should ring the bell really hard because very often these are places in the world where let's say the human development index, if we look for instance in South America and in some places also in Africa is lowered compared to the US. And therefore I think as scientists and actually as technologists, we should do everything we can and spend every minute that we have to think about what can I contribute to actually invent or find other solutions to have actually cheap accessible energy storage. And I think having for instance 67 cities with a climate like nowhere else on Earth right now in 2070 is really challenging because this has several implications on the human wellbeing. So there could be for instance race of new diseases. And also if we think about it, then there is actually when we look at the worldwide climate change projections, a very high affection by how many energy will be needed until 2050, 2070 due to that effect. For instance, when we look into India, we will have actually an uprise which is almost like a five times tall on the billions of kilowatt hours that will be needed actually to cool down the country due to the changes actually that are mitigated by climate change. So climate change is real. And I think we should really take care about that and invent actually new solutions. So let me go to the next slide. How do I jump actually from climate change to batteries? Well, that is quite easy. I think batteries are a very good way to store our 170,000 of Terrarat's solar income energy that we have for free into electricity and make it readily accessible and available in terms of infrastructure and mobility. So on the left side, I have a little diagram here where you can actually see a classic battery in the middle where you have actually the liquid electrolyte and a power separator that separates actually here the graphite anode and also the passage composite that you have over there. And in which you can actually store like in your conventional batteries of your electric vehicle or for instance of an electronic energy by shuffling lithium back and forth between the electrodes. Now, if we think about solid state batteries, what is meant here by solid is that we kick out actually this polymer separator and we bring in a ceramic material which encapsulates the lithium and it's just simply fast enough to transfer actually later the lithium between the cathode and an anode. Why this is actually so relevant is that this is the door opener to two main features for future battery technology. First, we can integrate pure lithium as an anode material because many of the solid state battery electrolytes are stable against lithium and this will increase actually later the storage density that you have and the overall energy density of the battery by a factor going up to 3,600 amps per in capacity. And also if we think about the cathodes and this is quite interesting because many of the solid electrolytes may be also a door opener to bring in much more high voltage cathode materials that conventionally may not all be stable in classic porous separator polymer electrolyte arrangements of the batteries. So by going to solid states, this may be the next batteries that will allow us to profit on higher volumetric and braver metric energies. So what does that mean for materials? So on the right side, you can see in a renews diagram where you have on the y-axis a lot of conductivity for the lithium and on the x-axis you have the restyprocal temperature. Now, if we look into that then there is two major classes that compete with today's polymer separators that you find in the classic lithium ion battery. There's the oxide classes where you can find materials like so-called lithium garnets and there is the sulphide classes where you can find materials like LGPS and others. They have actually quite high conductivity that I was in one order of magnitude comparable to today's liquid electrolyte porous separators. However, they are now in a solid state form. Now, there's a huge discussion right now in the field among TAG and also actually in science what will come first, oxide to sulphide. I just wanna declare here right away that I will not comment on that in my talk. However, I wanna highlight here some substantial differences. So in sulphide cells, if we look in a recent analysis by my colleague, Jürgen Yannick, then there is actually already a quite high advancement of cells that have been produced. There's a high activity actually in the field. And this has actually come already quite far along in having actually some first cell prototypes and different geometries tested. And the reason for that is quite simple. It is that actually sulphide-based electrolytes in terms of processing a low temperature processing that you can fabricate actually quite easily. These are razor soft materials. And that's actually quite advantages for that perk. What's a little bit challenging with sulphides is that you need very high pressure on the cells to operate them in later cell geometries. Now, when we look in the oxide-based materials, they're quite interesting because they can operate with pure lithium, which is a bit of a challenge for the sulphide-based cells. And also because they have been manufactured open open at high temperatures and high pressures, they need less pressure actually during operation. When we look at this analysis, overall oxide-based batteries, so to say batteries are overall, and that's one thing to consider, less advanced in various cell geometries and architectures that have been manufactured. So much more needs to happen, but I ascribe that to a strong ties that is existing actually to what is possible in terms of ceramic manufacturing as this operates in the manufactured higher temperature. Now, of course in Corona, we were a little bit bored with the team. So what we have recently done is we wrote, I think the longest paper we ever wrote so far comparing oxide and sulphide-based materials. And I just want to highlight here two, three key messages of this razor complicated plot, so we can compare transport electrochemical stability window and several mechanics. And also the thermal process window of different like lithium oxide compositions for electrolytes versus sulphide compositions. So why I personally think that lithium oxide materials are very interested and very unfair of unlike garnets is that this is, I think, a great way to bring in pure lithium into batteries. However, what we also see, and that is what my talk will be about, is that we have to bring down the process temperatures from traditionally like above 1000 degrees Celsius to much lower as this significantly scales later with costs. And this is actually where the motivation goes for the next slide. So when you do a little bit of the cost analysis, then we quickly realize that to meet the Department of Energy suggestion to have about 350 watt hour per kilogram of storage for a solar state battery system, this would translate to about $100 per kilowatt hour. Now in terms of ceramics, what it means is the following. So when one does a quick mass and one translates what it means in materials and processing, then one has about $30 per kilogram for processing and some materials of a solid state battery, which includes the anode electrolyte and cassoed. And if one looks, for instance, in other technologies that are available to do ceramic processing that are far more established like Soledoc said few says and assumes similar processes to make, for instance, solid state batteries in the future, like tape casting or other techniques, then this all breaks down to a cost bracket of about $8.5 to $12.5 per square meter on a cell pack level of total cost that the solid state battery would cost right now. However, this has not been optimized and that's very important to understand. So 75% of these costs are really related later to the manufacturer. So it's not the raw materials if that goes into massive production. It's mostly really the processing itself of the lithium oxide-based solid state battery materials. And this has to come significantly down to about $4 per square meter on a pack level to be competitive with current lithium ion batteries. If you would reach that, given the higher energy density because you can integrate, for instance, lithium, I think there's a very high likelihood that this is going to be the next generation solid state batteries. What I also will say is that not only a full solid state batteries of interest but also establishing the solid electrolytes for any hybrid models of batteries, where, for instance, a liquid like castor light that's just acting here as a separator and, for instance, then the dense castor nanode components there as well. So both is very possible. The take-home message that I want to give you from this cost analysis, and I think this is really important is that we have to reduce the costs and understand where they stem from in the manufacture of the ceramic solid state battery electrolyte park. So one thing that is amusing is that when the Department of Energy made the cost prediction of about $100 per kilowatt hour for solid state batteries, they assume a 20 micron stick electrolyte. However, I think up to this talk, there is not many opportunities or even manufacture options to make that. And I hope to convince you by the end of the talk that there is some new options to do so. Now, what I also want to show here is that in the battery design, as a smaller you can later go with the solid state battery electrolyte and the more room you give to the castor and the nanode part, the better it is for the battery because you will have an overall higher energy storage. But I've done a little exercise with my team is to look into various solid state battery electrolyte chemistry. So there is Slypron, which is currently a material that is the one that is cycled up to 10,000 cycles and the longest in solid state battery. It's also one of the oldest known ones or established ones. There's many other chemistries. And once this plot actually shows is that to transfer a high temperature ceramic made above 1,000 degrees Celsius from a six millimeter type pallet, actually, into a synth film structure or anything that is looking like a film, it's very thin to give more room to the castor and the nanode. It takes, on average, 10 to 15 years. So it takes about 10 years for developing a new chemistry of a solid state battery electrolyte. And then you have to add, actually, another 10 to 15 to make it very small, thin, and ideally at a low temperature. So this is actually not so good. So I showed you in the beginning that we have to mitigate climate change. So we have to rag it up and become actually much faster. So we have to find actually ways to beat that and offer better manufacture. Now, what I want to show you actually here in this talk, and I have scheduled the talks along different challenges that I see. So I want to talk in the first challenge that I outlined here, is it possible to take, actually, some of these known chemistry is normally made at pallets or tapes and bring it down to synth films and keep, actually, the same lithium conduction properties for solid state battery electrolytes. Then in the second challenge, I would talk about how to make that at low cost with some new manufacture options and ceramics. And then I will talk about some other topics. So another challenge is how to make good interfaces. And I want to show you here some new material classes on polyamorphous garnets. And finally, I will talk about how to make very reduced resistances and interface, or specific resistances at the cathode electrolyte interface. So let's go to the first one. How do we make very thin solid state battery electrolytes to give more room to the electrodes in the overall battery design? So what you can see here is actually a very classic Arrhenius diagram like we have seen before. And there has been many publications that look in the Arrhenius diagram assessing different oxide-based lithium solid state battery electrolytes in their conductivity and over temperature. Now then the other diagram that is looked at is the electrochemical stability window. And among the different options of materials, we can see here that the composition of lithium gallium lanternum zirconium oxide, the so-called lithium garnets, actually quite OK-ish in terms of conductivity. They have about 1 millisiemens per centimeter at room temperature. But what is really cool about them is that they are actually stable up to very low voltages so they can directly operate with lithium without any use or need of a protective layer. Now one plus that you haven't seen so far in literature, unless you open up this publication and that we allowed ourselves with our team to postulate that it will become quite important, is to look at what is the potential to make now a synth film out of these chemistries and what is the thermal processing budget to do so. So what we can see on the left side, if you do a classic synthesis like taking powders of a material chemistry for the solid state battery electrolyte, you densify that by applying pressure and then go into high temperatures, you need actually above 1,000 degrees Celsius to get this cubic high conductive lithium garnet phase, which gives severe limitations in bringing cobalt-reduced cathodes in because simply by co-sintering in phase nature, many of them like LFP or NMC are stable. One really cool way to mitigate that is actually to reduce the thermal processing window of the solid state battery electrolyte and looking also into making this component smaller in sickness. So there's meanwhile several options to make synth films of these components. I just want to highlight here the blue, which is the lithium garnet. You can either make it as a cubic or amorphous phase and by doing so you can access the whole range below 900, which gives you no opportunities to bring in cobalt-reduced cathodes as well and also keeping a very small size factor. So I mentioned in the beginning that it takes about 10 to 15 years to transfer for a whole field on this chemistry from a pellet-taped plastic processing into something very small in sin. Now I want to show you here where the challenge is in doing that and why does it actually take so long. So when we have, for instance, a chemistry like silicium garnets, then one of the most challenging things that is very particular to the ceramics is to keep the lithium actually all in the structure. So when you do classic powder processing and then actually sintering, you can normally over-lisiate your powder. And then you hope that actually the whole lithium will stay in the structure. And then when you anneal and sinter it, if you have put in enough, it will be enough to face-stabilize by the increased lithium content the high cubic conducting phase. However, if you take, for instance, a pre-synthed pellet, even if you over-lisiate that, and you can only do that to a certain limit, and you ablate later with a laser, for instance, the pellet to make a synth film. So you can ablate that down by techniques like post-laser deposition. Then you will quickly see that the field was struggling for almost 10 years because they lost most of the lithium and they never made it to fully access the fast conducting cubic structure and chemistry of that film. So one of the key home message here is actually to go very thin, you have to be very, very vigilant in assessing that you have actually a very high lithium conductivity by keeping the lithium in check and in the structure. And you can't often do that by just over-lisiating the target because at high energy, you lose most of the material. So we thought about that and we came up with a little trick. That's a work going back to Richard Frenninger and Mihals Tulsig and my group. And what we thought about is, well, can we use two targets? We make a second one, which is lithium nitrite and which is very close to the wavelengths in the bank gap actually of our laser zeta blades. So it allows us a very fast trans-oxalicium nitrite on our substrate. And we could start to form a multi-layer where every certain nanometer, you have either lithium garnet, a bit unbalanciated or lithium nitrite, and we start to build that out like a consistent little skyscraper, but as a synth film. So if you do that and you later post-anilate, you can decompose the lithium nitrite of the sub-layers. But the lithium then sticks in the film and starts to elicitate and populates the lithium garnet phase that you have. And by doing so, you can suddenly access the cubic high-conducting phase. So this trick is described here on the side. So you can see here, lithium nitrite and the other target, we go here in a multi-layer deposition. And after annealing at only 650 degrees C, we get actually here a fully elicited garnet film. Now, if we look into that a bit further, so this is how that looks like, that's a substrate. And you can see here the cross-section and a scanning action microscope of such a multi-layer that we deposit. So the whole thing has only 500 nanometers. You have to imagine before things were manufactured on a pallet and tape level, which has surely be beyond 60 to 100 microns, more millimeter level. So this is a significant decrease of the electrolyte. And you can see here by contrast change that everywhere where the shape goes from duct to light contrast, you have actually here the different constituent phase of isolicium nitrin garnet. Now, going up to 650 in an anneal, the magic happens. And actually isolicium nitrite decomposes in a chemical reaction. And we can later see one phase contrast, which indicates that we have now the full phase of the lithium garnet. And in the analysis by Tau Simms, we can later see that we have a quite good decomposition. And we really have salivated lithium garnet constituents. Also everywhere we originally had lithium garnet monolayer. We can still see the aluminum dopant very well as a signature, so one can track how the film was really made here. Now, how does that affect actually later as a phase? So long story short, we have here Raman diagram. Raman spectroscopy is really good to probe, for instance, the near-order lithium oxygen vibrations, which are very light and normally hard to track in an X-ray diffraction. And the main signature I want to draw your attention to is actually this orange line at about 660 degrees C. So when I compare that to a reference, that's exactly the cubic lithium garnet phase. And what's very exciting is that by this new process, we brought down typical processing from 1,050 to 660 degrees C. And we brought down a structure from 100 micron to about 500 nanometer in thickness as a solid-state battery electrolyte. So what it demonstrates is it is very well possible to make these structures very thin and also at low temperature, but you have to control the lithium stoichiometry very well. Now, how does that affect the overall lithium conductivity? You can see here on the left side, the conductivity and the Rhenius diagram. And we can see here that the film's process through that route after decomposition of the multi-layer to a full film. We can see that they have actually quite high conductivities, about 10 over minus 5 Siemens per centimeter at room temperature. And if you were doing it in the traditional way, you would just lose most of the lithium, you're under-lisiated and you simply don't have a fast conducting film. So this just demonstrates a way to do it. Now, what I wanna show you is actually in a summary and comparison on this part of the work is actually where we stand. So you can see here the room temperature, lithium conductivity of this electrolyte over temperature. And the first really cool part is that it is close or actually even very competitive to like one. So it's another alternative sin film structure that can be manufactured, but is now a lithium garnet structure. And what's also exciting is that you can see here the temperature reduction in manufacture from above 1,000 to around 650, but being also in a sin film structure overall. And after our publication in 2019, I was thrilled to see that there has been several follow-up groups. So that's for instance, Jody from Emperor who for instance demonstrated also even using a similar approach in sputtering, also higher conduction. So this is really translatable to several labs all over the world right now as a technique. So after this work, we start with this is nice, but actually it's not good enough, right? So pulse laser deposition is a vacuum technique. So how is that gonna help the planet? Because it's just too expensive and manufacturing. So demonstrates we can go very thin, but it's not good enough to support I think overall to have actually cheap energy storage solutions. And that motivated actually a part of our next works that I'm very excited about to share. So how can we make now sin ceramic and robust electrolytes at low manufacturing costs? So I wanted to just highlight real quick here where industry is and where research is. So currently when I analyze the feed and what they are doing in ceramic manufacturing that we just made this plot with the team with my colleague Moran Balayish and many other colleagues in the team. So when we look for instance, first of all, the good news is that most of the chemistries you can process them already in either symptom structures, keeping a situation of check or actually with tapes or pallets as remix. However, what is a bit of a challenge is that 95% of the published research if I just analyze that for Lysim Garnet is mostly down on pallets. And that has to do often with the history of a research lab being very knowledgeable and trained in making ceramic pallets by densification of pallets and then going up into very high center temperatures. However, none of these pallets will ever make it into any battery product for cars or EV simply because that is not the way to manufacture something later on large scale and for cheap costs. What has been very exciting is the group of Eric Brexman and I think there has been also other great contributions by Sakamoto and many others demonstrating that one can also make tapes out of these materials, which is already much better because that's something ceramic industry knows how to process. And these tapes are normally like more on the scale of like 100 microns and above to be robust. However, they have very far away on where a classic polymer separator of my phone here is in the battery. So in this battery here we have actually a polymer separator that is 10 to 20 microns thick. So what can we do to actually come in that space? So if we look on the other hand on synth films that's far to send, it's not robust enough. It's like below one micron. And also in all the vacuum techniques that's simply too expensive for low cost manufacturing. So we have to be in this space here close to a polymer separator. And I do believe that I can say that I don't know any product right now that is in that space, but being a ceramic. So how can we replace a 20 micron polymer separator but with a ceramic? I think that's a really important question here. And it's important as this will contribute to reduce the overall costs of the system. Well, so I wanna show you another analysis that I just made very recently. And I wanna show you what's the EV world needs here in terms of manufacturing and what actually battery electronics need. So we have here the electrolyte sickness on the left side of the y-axis and you have here a manufacturable solid state or polymer electrolyte area for battery design. So when we think about prismatic batteries later going in production or also advanced systems like the blade battery cell to packs that I currently discussed, we need to be able later to quote about 30 to 10 square centimeters like full and dense and defect free with a ceramic. So this is really challenging. In case of going later into the space of electronics this is far more reduced. So the separator size here would only be three by five square centimeters. And these are values that are currently used in industry a check with some contacts. And I wanna demonstrate you now here what the ceramic industry can do. So I will paste that right now in the same plot. So if I paste it here, then we can take an example by thinking about what is the product that can be made as a ceramic separator by tapes in the industry that is the widest area size. And that would be alumina and alumina for instance you can buy right now in a size of 20 by 20 square centimeters. So it would be within the range of what electric vehicles later need for battery as a ceramic coating. And that has been about 40 years of development of an industry. However, when we look into tapes for less in garnets then the maximum size that is currently developed and that is crap free and dense that you can find by some ceramic manufacturers that are about to come up with products or just very recently is on the size five by five square centimeter. Scaling this up to the size needed here for EVI will be a challenge because unlike alumina this wiggles much more in making the tape. And also there is much more changes in much more challenges in establishing the face in it. So it's far more complicated and not an equivalent transfer just get from a tape of five by five square centimeter to 20 by 20 square centimeter. And I think there will be a lot of development by industry and researchers needed to make it there. I am very hopeful for that but I think it's gonna be challenging. On the other hand, if you look for instance in atomic layer deposition, a sin film technique you would drop in the sickness so you would need a support structure but you could think about something light. And that could be very interesting I think for the electronic space however it will be very challenging just by area codeable to just upscale that to an EV. So we thought about that very heavily with the team and thought about, okay how can we invent a new ceramic tax that will allow us to be there where a cell guard polymer separator is but with a ceramic that has about 10 to 20 micron and that's really important here has the option to later upscale far more easily. Going away from classic tapes and doing what actually the ceramicists are trained to do the best ditch the centering at all to avoid it to really reduce the manufacture cost. So here's actually a new techniques that I wanna show you that you can't find in a classic ceramic textbook which we call sequential deposition synthesis. The idea is here that we wanna make films that are now about 10 micron and sickness but we wanna do that without any sintering and we wanna do that with a complicated chemistry like listening brownies and find a protocol to avoid sintering so we wanna do the whole manufacture in a cheap low cost budget manufacturing way staying at temperatures below 600 degrees Celsius. So there are several techniques like pyrolysis known as ceramic or soldier or dip coating where you can have a precursor and how it normally goes is you have some metal source default in some high boiling point organic you bring it to a spray gun or any other way you wanna coat you create some droplets and you let it rain on a heated substrate. And if you hit it right by chemistry you can have a transfer of the metal salt into the metal oxide and can invent something that is like a four, five cation component oxide it has never been done for listening down it so it stayed better electrolyte best to my knowledge or not with high conductivities. Now the problem is that you can't roll out the boat here normally it's ceramic manufacture to be above one micron because the challenge is that some moments you go higher in sickness you will see actually that you have got drying cracks and it's very difficult to control the process because that deposition on the heated substrate when you transfer the metal salt to the metal oxide you have a lot of like cracks in the transformation because all the metal salts go directly in one step and at one temperature event in the chemical reaction into the metal oxide forming later a four, five caterine lithium gamma composition. So it's impossible with classic pyrolysis to make films that are above one micron also very hard or impossible with soil gel or other wet coating cheap techniques. So we thought about it and we came up with the following trick one thing that is really cool about lithium some of the lithium oxide based materials we work with like Garnettes is that they also existing in a non-lisiated state. So the lithium garnet if you take all the lithium out of the chemical structure then you can have a lot of vacancies in it and it can be fully stable by being actually existing in the pyrocloth phase that's normally what you try to avoid when you have lithium loss going to high temperatures and sintering but we use this trick here that this is possible to be a deliciated structure by itself to our advantage. So what we did is we invent some new sequential deposition synthesis so-called STS precursors that have a two-step chemical deposition reaction where we define that we let first decompose when we spray down actually the precursor on the heated substrate all the metal cations that are not lithium to form a pyrocloth phase but we have already the lithium salt in it and then in a second step going slightly higher in temperature we decompose the lithium salt bring that into oxide let it diffuse into the form pyrocloth backbone and thereby it allows us suddenly to control the drying much better so we spread that out over wider temperature range and that is a new trick how to make unique 10 micron sick films that I will show you now. So what we can see now in the next slide is actually here a little comparison just to previous like ceramic tacks so what the field is currently doing unless you've gone it's for tapes for instance in industry as they make first a powder then they have to densify that then they go to high temperature and symptom have grainbound and volume diffusion activated and then actually get a solid state battery electrolyte that's quite dense and sick and it made at very high temperatures as unwanted. Now in this technique here what's new also compared to classic solar gelers that we spread out the thermal processing window on purpose at a very low temperature and we define this precursor that we first form the pyrochlor, lanternum, zirconium, aluminum double structure but have a lithium salt in it already at deposition and then in a second step go in a slightly higher temperature and then we can actually see that the lithium salt decompose it lithium starts to diffuse into the preformed pyrochlor structure and then actually starts to form the lithium garnet structure at a quite low temperature. Now we can see actually in the slide how that looks like. So this is also a joint collaboration that we had together with Samsung so I want to give you an honorable mention to Lincoln Myra and Wonsuk Chang who have been working with us here and also to Yuntong Jiu and actually Zach Hood who have been carrying out the work. So what I'm really excited here to show you is I think the first films that have about two to about 10 micron sickness. So it's now the sickness of ceramic crack-free and dense with about 10 nanometer of grain size and you can see here the crack-free coating overall that are covering actually it's a whole substrate and this is now exactly the space you want to be for processing. So it's a low budget process and it's close to where polymer separator is. We can vary the sickness going anywhere from one to even up to 10 micron 15 roughly. And it's a really cool thing is that always made about 650 degrees Celsius and we have never really classically centered these structures. Now, if we look here on the next slide I want to disclose a little bit on how the basic fundamentals of this chemical reactions work. So I mentioned before this two step chemical reaction and we made it a little like flashlight diagram how to constitute and how to make such precursors to make such a deposition. So in the case of lithium garnets what you want to do is you always want to choose a higher decomposition temperature for the lithium salts that is premixed in your precursor together with the other metal salts of your pyroclonic and garnet phase. So that you first form the pyroclore phase and then actually have the lithium salt already like sit in it and then decompose control the drying by there and being able to go very sick. What's kind of interesting is so below here in the lower box you would see different options for different lanterns of zirconium and open salts you can pick for instance. And you want to make sure here that they all decompose in one temperature events. So these should be very close and color codes that we have. And then you can see your different lithium salt options that you can choose. You want to choose one that is dissolvable in a mix of high boiling point organics but it's slightly higher to go here with our flashlight diagram than the prefront pyroclore that you actually have here from the other components. Why I'm telling that here in this way is that this method is translatable to many other lithium compositions. So you could make this in cobalt oxide like that lithium titanate but also many other options you just have to re-adapt and build up a landscape for the selection of the right SDS precursor to build that up. But this is a door opener to make some very sick films being close to polymer separator. Now, what I want to also show you is a little bit more about the formation. So now a particular case, I go back to my favorite Raman spectroscopy because I love to look really into this near order on what's happening with the lithium and oxygen vibrational modes and also many others of the light element. So what's exciting is that in the beginning when we deposit at around 200, 300 degrees C so we form a razor like dense film. And in the film, we can show actually Raman spectroscopy that's a lithium salt which is a lithium nitride is even trackable in the beginning of the film. So it's an amorphous film of the other metal cations that form. And in it, you have actually a ready is a lithium salt that is trackable by the signal. So it's a composite that you first form. Then you can annealing. We actually start at some point when we go higher in temperature to melt up and decompose. So you can see is a melting and transmission electron microscopy of the lithium salt in the preformed film. And when that happened, you have actually coinciding the formation of an amorphous phase first for the other cations than a pyrochlor phase. And then once the lithium starts to diffuse and in those various techniques which you've reproved that, we can actually see that the lithium goes in the structure and then forms from this amorphous slash pyrochlor phase later over into the cubic lithium garnet phase. But the cool thing is that's now not made with any expensive vacuum methods. So it's a really low tech, cheap way to process that. And we get at a very low temperature these 10 micron films with the right lithium garnet cubic fast conducting phase. When it would go to very high temperatures then we see also in the transmission electron microscopy the change from this amorphous and then actually molten off lithium salt and then elicited structure into actually forming a grain-grain boundary. So that's what you see here by the shade differentiation where you can see the progressing grain rows into the films. And just to show you and I wanna highlight here also some work of Jesse Hinricher in my group and also where Andrea Murano is also a big part of is actually that this can not only be used later to be coated on then substrate but you can also use for instance, para-strap straight. So I have here for instance like a little glass fiber that I just hold here in the camera and you can use it actually here to even coat glass fibers, then some crack free more or less in really gaining like some first coatings because the capillary forces of the droplets coming down will actually later like lead to a densification of the film. So I think this is some very interesting new ways to process that go beyond the classic textbook of ceramics. The films do have connectivity that are close to other films made on the lower end like 500 nanometers like the past laser deposition film but we ditched here really the vacuum methods and the expensive processing. Now what I will show here is just a little nutshell of what that means. So compared to traditional processing we show here that we are close to a solid separator but for ceramic. The way to do that is to increase the drying budget and stretch that out at lower temperature. That's where you wanna be with your chemistry and defining the chemical reaction of your SDS precursor. Why is that relevant? Well, if we can process now at 60650 you can bring in LFP or other like structures like NMCO and NCA as castles that are normally not stable in classic sintering when you have to go to very high temperatures. So if you're here in that space it's very hard to bring that in but now we have an opportunity to bring that actually together. And I think this is very relevant to look into cobalt reduced battery overall designs with a sin low budget manufactured electrolyte. Now what I wanna show you in the next slide is just a summary of that. So where's the fields of foreign terms of design? So if you look at the full solid state battery space there's been tapes made, right? And these use very high temperatures. So we bring that down to 650 and you can see a package of about 20 micron sickness dense and crack free of such a lithium garnet electrolyte. And I think this is the door opener to go to other new cell designs in the future not only for full solid state but also for hybrid battery device architectures where you could either think about having for instance also a liquid component in it and just using this ceramic separator to bring in lithium and possibly cobalt reduced castles. Now what I wanna show here just in a nutshell why I also think this is important for production later. So we see here again the tape. So we are currently here we can code and that's what you can buy a five by five square centimeter as that old tape. That's what you can find. So as that old refers back to the same garnet and the upscaling potential if it would be the same case like alumina but there's far more challenges by the face nature here for the same garnets and pure alumina you will be able to go probably in the range where you make cells for instance for electronics it will be very tough to access the room and the range for battery cells and EV. What I think is quite exciting is that with this technique of spraying and using the sequential deposition synthesis that I showed you there's an upscaling potential. So spraying is something that the industry knows very well how to do that in various products and just if you think about color coding now the chemistry is of course not transferable like the color coding but I think there could be opportunity over the next years to go for some upscaling and potentially also access the range. So I think this is now to be seen over time how that develops but I'm quite excited that there is some options actually ahead. Yeah and with that I just realized that I'm almost at 1046 so will I I'm not sure whether I have time to go in the next topic can you can you highlight that to me? Yeah maybe we can have about five more minutes will that be fine? Yeah all right, yes I will try to wrap it up. So I want to show you here one example which is talking more about polyamorphous sort of act like materials. So when I showed you before were actually the different structures you could actually see that there is the option to make crystalline cubic lysium garnet structures. However one could also see in the Raman spectrum other indicators that it's possible to also make amorphous faces. And fundamentally this is really interesting because imagine you could have a sort of state battery electrolytes that has no grain boundaries. So in terms of lysium dendrites this could be very attractive for somebody positions. And we were among the first to publish the existence of these amorphous faces for this lysium garnet material that you can find when you do classic symptom processing. We also found that the conductivity is very variable depending on what type of amorphous face you have and is so far undescribed in literature. Now why is this so interesting? And I want to do a quick analogy to Lypon. So Lypon is one of some materials which is known since 2030 years which was like very well developed in Oak Ridge National Labs. And which shows you about 10,000 cycles of the soil state battery life. And it's in part because it's a symptom structure it's amorphous and has no grain boundary. So it's a very interesting material to work with. So we were the finders of these described the first poorly amorphous lysium garnet structures where you can have this chemistry but you can have various amorphous states. Now what's very different to Lypon is that Lypon is a classic Saharan glass. So we have all these different tetrahedra so that's the little triangles you see here are the phosphor oxygen nitrogen and some can form dimers so you just like click here with another triangle and then you can have the lysium actually migrate in this material and when them work. So what really fascinates me is that has been challenging to describe for 20 years the best arrangement of these triangles or tetrahedra actually here in the amorphous state to find the best lysium diffusion and also like describe how to process that. So it has been quite challenging but it has come an enormous way and shows us really high cycle numbers. So that's great. But if that has been a long way then the question is well, lysium garnets do not have one building units they have even up to four. So they have the lysium on an octahedral and tetrahedrocytes. The zirconium on an octahedrocyte and you have the lantinum oxygen also in a dodecahedrocyte in the structure. So it was really fascinated. It's going to a good zoo and to watch this animal that becomes more complicated like how to describe the lysium diffusion in such an amorphous structure. And what excites me a lot is there has been recent demonstrations that after our publication of the existence of the amorphous lysium garnet structures that these were integrated actually in first cells and they show very high cycle stability in life. So there is a first demonstrations of 500 cycles at a C rate of 10. So these amorphous lysium garnet structures could be interesting protective layers later for various structures ranging for sulphide or oxide-based cells because they're also made at quite low temperatures. So there could be good protective coatings to lysium. Now, when we looked into that and we tried to understand the structures and I want to highlight here is a work of Yun-Tung Ju and Heyman Pike and many others actually in the group and also a good collaboration with Igor Lubomirski at the Weizmann, Anatoly Frankl and Stony Brooks and Claire Gray in Cambridge. We were interested in trying to understand how we describe cesium amorphous phases. So as mentioned, you can make them those packs also as being amorphous by either SDS or PLD. And when we look for instance, the lysium NMR what's really exciting is that depending on the post and kneeling temperature so where I show it here was my cursor we are everywhere amorphous still we can find very different lysium hopping dynamics and only if we crystallize to the cubic phase we can actually see here that we have a classic like corner bonding of the bonding units. However, for the various amorphous phase we see it's a non-saharias in class because it can bond actually over the phases and ashes even for the building units. And it forms depending on which amorphous phase you are in a different condensation of the overall structure. So it's in some analogy very different to lipon and it's another type of a non-saharias in class. Now what I wanna show you here is that the impact on conductivity is striking. So you can see an orange different amorphous tailored phases that are existing for the lysium garnet in the arenas diagram. And depending on what temperature you use to freeze the amorphous structure in the manufacturer you can see here very different overall conductivity compared to the cubic one. So this is kind of very exciting because that correlates back to huge structure changes that we observe here. So when we look for instance on the conductivity over the post-saharying temperatures then we can see that for the various amorphous states depending on how actually the building blocks are connecting we can see here actually that we have different conductivities which has a maximum around 4,500. And I wanna highlight here work from the Weizmann from David Ehren, Igor Lobomirsky where they showed also that amorphous perovskites such phenomena exists. However, we have four local bonding block units compared to a perovskite that would only have technically one. So it's far more complicated. What we can say right now is that the lysium and the zirconium and the exupson NMR evidence is acting as a network former with its building block units where the lantern acts like in a classic glass as a network modifier. And we can see here that with increasing annealing of the amorphous phase, we see an improved lysium local orderliness. However, it has a peak. And we can actually see here that it depends on how much edge and phase sharing you have of the local bonding block units for the oxygen to a breach and how the lysium is later diffusing through. So in a nutshell, what is interesting here is that this could be alternatives to lipon that have actually new tailored structures. And because I made it quite low temperatures so you can bring them down to quite low temperatures, they could either be constituents of future source state battery electrolytes and or also be like protective coatings, for instance, for oxide or sulfite based battery designs. Okay. And with that, I would like to jump to the end. Sorry, I'm doing a little jump here. I want to thank very much actually my team and my collaborators. Is there something you enjoyed in this talk? This is to all those people and I show you here actually some pictures of the team pre-COVID because they are very joyful and happy, I think. And I want to thank all of them so much for support. I particularly wish to thank all our sponsors, especially also here, Samsung on that one for all the fantastic collaboration. I want to thank you for your attendance and I'm open for your question in a minute. And I want to just highlight for a quarter minute that if you're also like looking for opportunities, I recently founded the Lila Materials Science Mentoring Program supporting minorities, females and LGBTQ in solid state ionics. So if you're interested in having me as a mentor, you don't need to be at MIT. You can be at a company, you can be anywhere out in the world, please apply and I'm very happy to do my share and write a paper less a year, but also support here as a career. Thank you so much. Yeah. Jen, thank you so much for that very exciting talk. I think we have now time for a few questions before we go to Elsa. So Jen, let's talk more about your new approach for the sequential spray deposition for the solid electrolytes. Can you tell us a little bit more about the difference of the property of this film that you make in comparison to a traditionally processed thick film of the same dimension? Could you help us understand the difference between? So you did 10 micron traditional process versus 10 micron process your way. You already highlighted the difference in the processing conditions and the difference in the conductivity. But how about other properties like mechanical property or cell behavior, dendritic resistance, be great to get some additional details. Yes, so I will try to disclose as much as I can. So first of all, there is, so the best of my knowledge, no alternative processing available to make such a 10, 20 micron film. So there is nothing out there. So in terms of what chemical space, there's not really another method that can do it. The best I can compare it to is, let's say, a 60 to 100 micron thick tape, right, that you can process. And the lower end would be than a synth film made by Erdo is similar. So what's really interesting in terms of other properties to mention, as you asked, is that the aspect ratio of grains to the total thickness is very high. That's quite unique about these films. So you have to imagine you nucleate directly in a solid body and because of that, you have an average 10 to five nanometer grain size. So you have a lot of grain boundary area, but your thickness can be 10 microns. So to shoot an example, if I would make a similar film, but I'm restricted to 500 nanometer or a micron in ALD or PLD, you have also 10 nanometer grains, but you have only maybe 500 nanometer on film. So you can't stack so many grains over the thickness. If we look in tapes, you have maybe like 50 microns of average grain size, but you have a tape of maybe, I don't know, like 100, 300 microns, or maybe you have like 20 microns of grain size. But still your aspect ratio is very bad. That means that you don't have many grains going over the film thickness, which probably in terms of the mechanics is at a disadvantage for a sintered products. And which I think here is very interesting because you're kind of like here in the sweeter spot. Now I can't disclose anything right now on the cell level testing and other performances because that is in the making. However, what I will say is that there are cells that are very promising in recycling. All right, Yi, I think you have a question. Please go ahead. Oh yeah. Hi, Jennifer, it's a great talk here. This processing is amazing, making a very thin film right now. A couple of questions actually all related, but I think I'm fascinated by this new processing you develop. The first question will be, so when you do a kneeling is on a substrate, then you will need to release that, right? Do you need to release that become freestanding? Kind of one to 10 micron film. And then how do you do that? Because they're kind of bonded to the substrate and the kneeling temperature go up to 500, 600 degrees Celsius. Second question is certainly, you know, ionic conductivity is still a little bit further away from what you are looking for, right? So you probably want to look for 10 to minus three, at least multiple times 10 to minus four. And this probably is a balance and they are kind of conflicted is the answer to have a high ionic conductivity you want to have in the temperature or say, maybe even going to close to center in temperature. So what will be the balance of that to get to the high ionic conductivity? Basically two questions, yeah, yeah. Yeah, so those are great questions. Thank you so much Yi. So come to number one, right? Which was about how to peel it off. So I think for the best battery architecture later, you don't, you just take and choose a power support structure, which could either be an electrode or something else, but it's good in terms of weight later, reducing that, that's not all for ceramics, but you could have a power support structure and then actually fab it on. So I showed you one example where we fab on something Paris. So the idea is here not to peel it off, just have it as an integration of a part of the battery later. So you would just like leave it on as a coating just coat it on. To come to number two, you're correct, right? So the lithium conductivity is a little bit lower right now. And I ascribe said that there is still room for optimization. Yeah, yeah, yeah, there's still room for optimization. However, I think it was first for us to demonstrate some new tech. I don't think you have to center to high temperatures. I don't believe it. It's more fixing maybe the total listening concentration in the precursor. And it could also be that because you have a lot of grain boundary area, that's a question to be answered by science. That's a connectivity is a bit lower. But even then, I'm not sure I really care right now from the processing or battery standpoint because it's not so much and you're just so much thinner in terms of form factor. I would like to do everything to avoid to go here to very high temperatures and avoid and ditch the century. So Jennifer, for the case, you don't have to go to free standing, right? So you have NL, you have electrolyte, cathode. What would be the way to have compatibility? Because you are not going to build a multi-layer. You still need annealing. Then lithium metal can rotate annealing temperature, right? So yeah, it's what I'm thinking about. Maybe the reverse architecture is you make cathode first, you may electrolyte, then you deposit lithium, then that's it. And then you just, you know, roll it up. Yeah, Yi, you know, you actually answered my question already, which is fantastic. So that is one way to do it. So it could, of course, be 10 other ways to do it that I'm very happy to talk to you about offline. OK, sounds good. Yeah. Thank you, Yi. I think in the interest, we should probably move on. But, Jenna, I just want to make one observation. I think this idea of sequential lithiation is really a great one and it really mirrors what is done on the cathode side for lithium-ion battery. You can always, you know, think about when the lithium goes in as controlled by the kinetics. So I think this is something that is makes a lot of sense to me. So thank you for sharing that. But we'll have time to talk to Jennifer more after Elsa's talk. So thank you very much, Jennifer. So our next speaker, Elsa Elevedi, is here. Elsa is also a professor of material science and engineering from MIT. So we are having an MIT MSC day today. So very pleased to be hosting you both. As I mentioned earlier, Elsa has spent her career looking at the supply chain aspect, materials availability, circular economy and technical economics related to many different processes. And she has made a lot of outstanding contributions to plastics, to catalysis, to energy storage and other fields. And applying this really comprehensive frame of analysis to have the holistic picture in the area of batteries. She has also looked at many components. For example, she has worked on the availability of cobalt and its impact has looked at the processing of solid state batteries. And she has also looked at many other aspect of batteries. And these are going to be essential topics for translating new technology to market. So, Elsa, we're so pleased to have you. Thank you very much for joining us. The stage is yours. So thank you so much for this opportunity. I am grateful to add our for the opportunity to add our perspectives on challenges in solid state battery manufacturing, scaling in particular. So I'm just going to share a couple of stories of work we've done, nominally, two things. And then one little idea on at the back end here, looking at materials availability and manufacturing scalability. And the first piece of this will be focused on solid state electrolytes, but I'll broaden it to lithium ion systems more generally. And the space that this fits in within the work of my group that we'll just give a great overview of is looking at scalability and research in particular and support of electrification towards decarbonization. And in that domain, there are basically four spaces that we like to think about within the group, how supply chains, particularly material supply chains, need to evolve, how we think about materials, properties and what we're aiming for in terms of technical performance and how does that contribute to scaling more generally related to process cost. And then finally, the overall ecosystem in which these materials operate and that gets to kind of market dynamics. And so I'll touch on supply chains, but a little bit about kind of process and materials, properties particularly around scaling. So that's the space that we'll cover today in the time I have with you all. And the premise of the work around scalability of what we're trying to do in the group, and just beginning these efforts and always are grateful for input on this, is as researchers we're trained and we train our students to develop new technologies with really the aim of optimizing this narrow set of performance metrics over here on the left around using particular materials and how that relates to technical performance based on the set of lab-based process and equipment and those sorts of transport environments. But I feel that in order to scale at the pace that we need to, the sort of unprecedented and overwhelming at times pace, that we have to be able to think about manufacturing at scale as early on as we can. So try and avoid unforeseen manufacturing, assembly and integration challenges that frequently arise. And so this is our small attempt at trying to contribute to that conversation, but always very interested in expanding that. And so when it comes to solid state batteries, as I'm sure has been looked at throughout the day and what Jen said as well, one of the challenges is thinking about interfaces in particular and how we might manage interfaces better in terms of successful integration of solid state electrolytes into fully all solid state batteries. And so in that, and that's just this first case that I'm going to talk about where we've been using some methods both on the economic side, but also a little bit of data extraction to try to understand what are strategies that have been pursued to deal with solid electrolyte interface layers and inter diffusion at solid electrolyte cathode interfaces in trying to improve technical performance. But then what does that how does that represent in a manufacturing context and what that looks like to us from this kind of very schematic perspective is, you know, we're thinking about the kinds of things we think about at lab scale, you know, making things in coin cell forms, impressed. And then what does that mean in terms of manufacturing at scale, where we would have different cell configurations, different methods that are used and how does that manifest in yield and throughput and in particular what the kind of my punchline on this first piece is just that we as much as we can think about the cost in manufacturing implications of strategies that we propose to deal with interfaces as an example, you know, the obvious statement is that technical performance needs to outweigh the associated costs or the associated implications in terms of materials that we're using, et cetera. So that's kind of my punchline, my takeaway from this first part here and in what we've tried to think about in the methods that we've used to try to do that. So to that end, from a methodology perspective, what we do is basically twofold. One is looking at the cost implications of this, so developing technoeconomic models of what scale production as a function of a particular technology means and then how do we think about ways in which technology might be improved or performance might be improved in the materials and cost implications of that. And so for this work, which was recently published in Juul, we looked at oxide and sulfide-based solid-state batteries and you just see the example process steps on the left where we've just looked at what the manufacturing context would look like, what the steps were required and the associated materials with that. Again, just looking at an oxide and sulfide-based chemistry as examples and we have a little bit more resolution among the oxide and sulfide families, which I'm just illustrating on the right-hand side here. We end up with this baseline cost and the total cost numbers are not the point here, but just to give you a sense of, we do have those numbers but really what we're interested in is how those change as a function of these different strategies that might be pursued to address interfacial issues and just to explain these plots on the right-hand side a little bit in a little bit more detail, we're just looking at total cost in dollars per kilowatt hour as a function of different oxide and sulfide-based chemistry that we're looking at and then just broken down by cost component where the materials cost is dominant, as one might expect. And on the right-hand side for the oxide-based system, we just looked a little bit comparing if they're trying to normalize by density differences between the oxide and sulfide-based systems and capacity-based differences. So those details are a little bit beyond the scope of what we were, of the point that I want to make with you all this morning here, but just to explain what's going on there. But this being driven by materials cost is not unexpected but also drives the ways in which we pursue the analysis beyond this point. So that's the baseline for the economic analysis that we're doing. And then that we couple with the text mining analysis that we did in collaboration with Jen, looking at extracting not only the process conditions within the literature, so what you're seeing on the left here as a function of the chemistries that have been pursued in the literature, a different sulfide and the LZO chemistries as a function of the temperatures used in the synthesis. That we have extracted automatically using that pipeline that we've developed within the group to do so through natural language processing to try to summarize what are the recipes that are used that then feeds into that economic model that I just showed you, right? If we know the steps and the temperatures at which they were occurring, then we can begin to estimate the scaling of those as a function of things like production volume and how that contributes to yield as well as the interfacial strategies that have been pursued. So this is just giving you a quick example of a basic set of extractions that we've done here. And what you're seeing is as a function of chemistry, the highest frequency temperature, reported temperature for a heating step within this. And just again as an example, we can break down the LZO process flows by the temperature and heating step that that temperature was applied. So we're showing here on the right hand side, the LZO broken down by a sintering step, annealing, calcination and drying. And you can just see the evolution of that. And then also what's interesting is to dig into the range of temperatures that are used and go into unsure some of the things that Jen talked about. I apologize for missing the details there. So the baseline economic analysis coupled with details on the literature, details within the literature on the processing conditions, we can pull out very quickly, couple that with other chemistry approaches that are adopted in order to try to get at how were temperature reductions achieved without sacrificing performance in processing temperature. And I'm sure again, Jen talked about this as well. So here on the left, we're looking at processing temperature as a function of dopant essentially. So where that's modeled as a variation from LZO molecular weight, right? We would expect a relationship between variation and molecular weight based on the degree of doping and how that would be pursued and then try to correlate that with processing temperature. So here just trying to give you a little bit of example of how we leverage what's in the literature to extract out ranges in processing temperature, ranges in chemistry that are pursued. And then also on top of that, different approaches that are used to address interfaces such as adding interfacial layers, changing slurry chemistries, et cetera. And so all that information feeds together in a set of analyses that I'm just gonna show you two examples of around use of text mining to learn processing conditions to then translate that to economics, to try to inform some questions around potential scaling implications as early on as we can in the pursuit of these strategies within the literature and within the research community. And here, as I said already, my punchline is that we would want to quantify as early as possible where performance gains are able to outweigh costs of implementation. And so I just picked two examples here combining the techno-economic work I said with the data mining to look at what are the potential scaled implications of strategies that have been pursued. And I'm comparing two different strategies here that have been pursued to address interfacial issues and the particular recipes associated with those strategies. So in the first example on the left-hand side, there's a sputtered tin layer that to help with interfacial resistance performance that we see the baseline cost in blue there where we don't have the added tin with performance increase of about 60 million powers per gram as reported. But in this case, the cost of the deposit layer is sort of has equivalent cost in that outweigh essentially the performance gains. So that's the left-hand side where we're sort of our green that's shown as the improvements in performance lowering the dollars per kilowatt hour but then as a trade-off associated with increased costs associated with sputtering the tin. And obviously the performance at scale, we don't have the direct measurement of and we have modeled what the tin sputtering would cost at scale. But just this idea that there may be a challenge in having those performance gains outweigh the cost of implementation is a flag for us to think more thoroughly about in the design of that particular research or pursuit of that particular direction. And that is comparison on the right-hand side where this is a different chemistry obviously but the example here, what the strategy was was an optimization of the slurry management of the slurry chemistry and in addition of more binder that saw a significant increase in performance as shown by the green directionality there at not so much added cost, system cost. And we think about that in the context of the model. So in this case, we would see that potentially that strategy, the optimization of slurry chemistry as an example, if we can see these sorts of performance gains may be a more cost-effective strategy. So again, not to say that these are the costs or that this is the sort of a strategy to be endorsed but just this ability, this sort of demonstration of these tools to try to inform manufacturing costs at scale was what I wanted to share with you all here. And then so the second story I wanted to share has to do with going back to my initial framework that I talked about in terms of considering scalability and research. What I just talked about sort of overlaps this materials and process cost idea. But the other thing that we like to think a lot about is supply chains, is sort of, excuse me, materials availability related to rapid scaling effectively deployment in terms of supportive electrification in the scale at which that needs to occur. So I'll just use the second part of my time here to tell you a little bit of that story. And in that case, my punchline again to scoop myself is that it's not about running out. I'm sure folks are very familiar with the sort of cycles of hype in the popular press about, oh, we don't have enough X, we don't have enough Y. For us, it's about trying to identify what is it about a particular materials market or how it's used in a manufacturing or towards the science perspective, what's needed in terms of the purity or nature of a material. How does that contribute to potential supply chain disruptions? And that's important for rapid deployment as folks could intuit because if the supply chains are disrupted or even perceived disruption in supply, that can impact the ability to deploy, impact people's interest, market adoption, companies risk effectively. The risk being deemed too high by a decision maker in order to pursue that strategy. And that can be a function of many things, the market imperfections, how substitutable a particular material is, the nature of how it's mined, et cetera. And just to drive the point home a little bit further, the reason why we particularly like to think about this issue in addition to the supply chain disruptions is that the implications that that has on price, on cost can again discourage technology adoption at the pace at which it needs to occur. And so in the right-hand side here, this is some work done by Yeming Cheng and Bill Green at MIT looking at learning curves associated with battery manufacturing capabilities. As those decrease, as shown in the top line here, we start to asymptote towards the materials costs. And so we become more subject or more sensitive to volatility in the price of materials. And so we've done a lot of work to understand what are the factors that drive price increase and therefore could influence the cost of a technology and its ability to scale. And so the point that I wanna make here is kind of a couple fold, is that we look at a set of kind of screening metrics associated with materials availability. Oh, I skipped a slide, that's the problem. And as folks know, we have to get specific then about what we mean by a particular chemistry. And as I said, my first example was focused on solid-state electrolytes, when I talked about the manufacturing scalability, here I'm broadening to materials generally within lithium ion chemistries and in particular looking at some lithium cobalt nickel and then I'll show some metrics around manganese, et cetera, as we go through. And as we know, we need to think about what chemistries we mean because as a function of what chemistry is gonna mean different things in terms of materials intensity, this audience knows that in spades. And so for each of these elements, I'm not implying, obviously it's used in its metallic form, but just to, it's easier to talk about the data in that way. In the case of lithium across the just even NMC chemistry is lots of uniform, or the total quantity of materials intensity is similar versus when we start to think about cobalt nickel and there's much broader range even within the NMC family. And so in terms of our methodology, in terms of the research that we pursue, the idea is to look at some basic screening metrics to get a sense of which materials might we need to dig into in more detail. And so I'll just make a couple of points just again to illustrate our approach, but then also try to draw out some conclusions based on that. So what I'm showing on this slide is in two plots the same information on the x-axis is fraction in top country, meaning how concentrated a particular supply chain is geographically versus the what's called static depletion index on the y-axis, which is the reserves, the economically extractable quantity of a particular material divided by the annual production. So this idea of how many years that current economically extractable content would we have material? And the point is about the contrast between the two. On the left-hand side, you see a much narrower axis in both the x and the y for things that are mature elements, nickel and manganese dominated by their use in steel, not a lot of movement in these metrics. Excuse me, I forgot to tell you one critical thing, which is that each of the points that I'm showing here are five year increments from 2005, 2010, at 15 and 20, obviously these numbers change over time, both numbers in the numerator and the denominator of the x-axis, y-axis change over time. And so for nickel and manganese, there's not a lot of movement if you account for the more narrow scale compared to on the right-hand side, looking at lithium and cobalt, again as contrasting examples, where over the course of my five year increments, lots of movement in these dots, where we're finding more, where the nature of extraction is changing, and the point of concern in particular, as folks know, because it's been often talked about in the literature and in the popular presses, cobalt, that has both become, over the past, since 2005, has both become much more concentrated in geographically, but also that we are not finding material at the rate at which we are at an increased rate relative to the rate we are extracting, such that the number on the y-axis is also going down. So what these provide for us is a screening for where there may be materials of more concern to understand in more detail. And another way to think about this visually is just to think about the geographic distribution of these materials where I'm then now comparing just cobalt on the left in terms of trade flow in million US dollars, where export is shown in red, import in green, on the left-hand side for cobalt and on the right-hand side for lithium and the contrast is very visually apparent, the degree of concentration for cobalt geographically versus lithium and cobalt being concentrated not only in mining, but also in refining, whereas the lithium supply chain much more diversified and as folks know, also more diversified in terms of how we mine it from both brine and rock. So just again, screening. The other concern I'm sure as folks know is for cobalt is the byproduct nature of the extraction that cobalt is mined as a byproduct of other carrier metals, copper and nickel principally. And so that's another flag for us in terms of where there may be concerns, supply chain concerns that could lead to more volatility in price. And the right-hand side, I'm just showing this framework that we've used to kind of evaluate this byproduct challenge in a more comprehensive way across energy materials and I'm happy to take that and talk about this more in questions which I'm hoping to leave a couple of minutes for here at the end, but I won't focus on it more now. So those are some metrics on the supply side. The other interesting point that we've looked at on the demand side, gets to this degree of deployment issue in particular around scaling in those sort of how scale do we need to get in order for or how much production needs to change as a function of the pace of deployment and the potential for resource constraints even where we wouldn't expect it because of that growth trajectory. And here I'm comparing lithium, cobalt and nickel and as folks know, other chemistry has become of more interest in looking at manganese or iron phosphates and I'm happy to talk about that as well. But what I'm showing here for lithium cobalt and nickel is the historic compound annual growth rate. Sort of how these materials have grown in the past in terms of fraction relative to certain amounts of deployment, so five terawatt hours in 2030 versus a hundred terawatt hours in 2050 for each of these elements in comparison to the historic. And what I'll just particularly highlight is the significant compound annual growth rate required for nickel and this is for a set of chemistries, NMC based chemistries effectively through 2050 for a hundred terawatt hours. So this is thinking about both vehicle and grid based deployment potentially that this 15% growth rate required is far above the historic cagger. And then the other nuance that folks may be aware of associated with nickel is that the vast majority of nickel is currently extracted is not relevant for the EV market because it's extracted for steel application and a ferro nickel. And so being dominated and even growth in nickel has been dominated by expansion of nickel pig iron for stainless steel or ferro nickel with fewer or discoveries that would lead to battery grade nickel. And so I think that the implications there are interesting to consider in the nuance around that we don't mean running out, we mean the sort of change in the nature of the supply of that material and what that means for the supply chains that is interesting. And so nickel is particularly of concern there versus something like lithium. And what I'm showing on the bottom here is over time lithiums, the dynamics associated with the end use for lithium use in batteries. And this blue line here is showing that end use increasing to significant fraction such that the mining supply chains have become more used to that market as an end use for extraction actually. I think there was just an article in the New York Times this morning maybe or yesterday about lithium growth domestically, which was interesting as well. So I'm headed towards the end of my time here and I just wanted to bring this, oh, how funny, there we go. Bring this back a little bit actually to my interfacial point around solid state electrolytes. And we did a similar analysis looking at a couple of strategies that had been pursued for interfacial resistance mitigation in solid state electrolytes, looking particularly at germanium, tandem and lanthanum as kind of going through that screening analysis that I just gave you some metrics around the elements that popped out for further consideration related to solid state electrolytes more broadly. And so the point being here is that the tandem raised some concerns for us in terms of the, you can see the blue dotted line here, sorry, I didn't explain the plot. So the x-axis has to do with deployment in terms of terawatt hours in 2030 for solid state battery production versus that compound annual growth rate that I talked about before. And the blue dashed vertical line here corresponds to reserves for tantalum in 2018. And then these little stripes here are the caggers, the historic caggers for germanium and tantalum in particular. And so tantalum is a flag for us in terms of scaling along two dimensions, one the sort of resource, the reserve limit and the degree of change from historic deployment. As well as on the right hand side, tantalum is also extracted as a byproduct material that's similar to the way cobalt is. In tantalum's case it's from niobium and tin. And so what I'm plotting over time here is tantalum production relative to the potential for added tantalum capacity relative to extraction of niobium and tin. And we see that this is sort of like the buffer, the production buffer associated with tantalum relative to these other elements. And we see that as you compared to germanium, which is log scale here, that there were very much below the buffer for current germanium production. It's not so true for tantalum. And so again, just the sorts of screening metrics that give us pause and mean we wanna dig into these and more, dig into this in more detail to understand how cost implications might arise. So I will end there to reinforce the point I made before about this gains in technical performance must outweigh intended materials and manufacturing costs. And as early as we can understand that as possible might help to refine and inform the research directions that we pursue. And then I just have a couple of points here on these materials availability. Lithium as is often reported, has been able to be responsive to these increases in demands. Whereas cobalt and nickel, there's some nuances in those supply chains that mean we wanna be thinking about other chemistries that will help to diversify that challenge. So with that I'll say thanks for this opportunity and I'm happy to take a question or two before I have to log off. Thanks so much. Also, thank you so much. I know you have a busy day today. So thanks for making the time. Maybe I will just take a very quick question here. So on the slide that you showed the cost distribution for conventional solstice batteries, I think you have the material cost at about 90%. This is quite extraordinary. Is this pointing to the low cost of the processing or the high cost of the raw materials? In this case it's more the, I guess high cost of materials or just that the materials cost is what dominates kind of any gains in that we're able to make in terms of production materials or scaled material production effectively. I mean that's a lot of times what happens in these newer technologies or evolving chemistries is that the ability to make those materials in bulk at scale is, there's a learning curve associated with that. And so that's actually something that we're digging into now is the, it's linked to my rapid deployment point, right? Depending on the trajectory of what we're, it's good to have these time horizons in mind as we're developing these technologies because that would point to that. So it's more driven by the materials cost, but I think that they're very much linked, right? The ways in which we improve, this is the point, the ways in which we are able to improve yield through improved production, to the points that I'm sure Jen was making earlier on, then that just automatically drives cost down. Sure. And I think it also assumes quite a thick membrane for example, for the cost you're projecting. So certainly making a thinner would decrease the cost. Of course. That's very nice to hear given the two talks today. And then I know your time is very limited. So I wanna thank you again for joining. And now we're gonna have a discussion with Jennifer. It was awesome to see the two rock star faculty from MIT present back to back on a similar topic. So this really highlights, I think, the depth of the bench at MIT and the stars of the player. So I really was very happy to see that. So now we have a few moments to talk more broadly. And I thought I will start, Jen, by asking you this question about scaling up. Can you tell us a bit about where we are in scaling up solid state batteries, specifically the solid electrolyte, and where do you think a big gain can be made in terms of the scaling up itself? Not about the chemistry specifically, but just about where we are and where do we need to go for scaling up? Right. So that's a great question. Well, so my understanding is from listening to some of my contacts in car manufacturing industry for EV in particular, right? If we wanna bring it to a cell with 10 by 30 square centimeter, that's a lot of coking for a solid state battery electrolyte. And if we think about it, and I apologize for that, will to be for a moment technical, if we think about what it means in terms of homogeneity and in controlling shrinkage and in controlling how well we can manufacture things for a powder, I think for a tape, it's, I do see perspective, but it needs the best ceramicists, right? And how working with ceramic industry to go from something that is, I think manufacturer like five by five square centimeter of a tape to 10 by 30. And my example is that for alumina, it took them decades of development to get it to 20 by 20. And alumina by phase is far more easily made. So let's say you shrink it, it's far less than compared to lithium garnet, just as an example. So this is the challenge I think we face in manufacturing. And then as Elsa also mentioned to be compatible and price, my personal suggestion is to really look into what are alternatives to not manufacture from powders for scale up, but raise it directly from liquid chemistry into solidification, which saves you one process step is good on costs. And also I think in coding it on could be a game changer in terms of scaling this up and move completely away from any vacuum tag. I don't think it's a production volume later is so high as anticipated. It's gonna be challenging I think, yeah. But it's easy, it's different routes, yeah. Yeah, absolutely good. I think for especially battery electric vehicle applications, I think the vacuum route is going to be cost controlling, I think at the end. But perhaps for specialty applications that will be reasonable. But let me ask you a further question. So you highlighted this problem with the area, right? That really making a large area. Am I correct to understand that's the path toward lowering the cost production is to scale up the production area? Or are you thinking more about the size of the cell that could be made in terms of the requirements on area? I think we have to differentiate that. I think area is later going to define whether it's going to be an application like razor electronics or drone because then you need like significantly lower area, right? You cannot work with other cost factors or whether it's going to be EV. And I think for EV, it's going to be large. And so it can be that some ceramic tech will make it easily into the electronic space but it's more challenging to scale up and will outrule some methods for EV. Now, the other part of the question I think is that you had was on costs. And what I do believe is I think in terms of costs, if you can, I think if you can find any method where you don't have to go to a manufacturer, get the powder, densify, go to very high temperature, also in terms of making the cassoed on, right? So you have to co-center that. You have to get a mechanical good bond. And I think the only way and a stick to that to bring in cobalt reduced cassoeds is to walk away from classic co-centering. So it's not like a fuser. So battery is very different here. And in order to do that trick, I think it could be that run route is to just not traditionally process from powders and then classic centering. Yeah. Now, absolutely agree. That will affect the costs, I think, very much. So I think it's two separate things on area and cost brackets. No, I think they're highly related. And one thing that really occurred to me is although you didn't show any pictures of your photos of your setup, I can imagine it's very inexpensive to have the sprayed deposition setup compared to a very precise temperature-controlled kiln. Do I have the correct understanding there that the capex for production is also gonna be crucial as well in driving the cost? Yes, so currently it's not. So in my analysis that we published, it's 75% of the total costs will be given by which production route you take. And I think it's very clear, right? If you have to go to a manufacturer like BSF or someone and get the powder, then apply pressure densification, center very high up. There's many processing steps that you have to execute that are quite expensive. Then the problem is like many of the best-less in conductors are those structures that a small change in any stoichiometry. Locally, just let's think about a tape, right? Can lead to challenges in controlling shrinkage warping. Warping is a big challenge if you scale up. So these factors don't scale linearly. And I think one idea here is by taking the center element out and taking a powder to densification route out, but coming from nucleation and decomposition reactions that it would allow you later to, for instance, coat wider areas, you're more independent of the substrate. Also maybe in terms of weights, that's interesting because isn't it better to have a right power structure in which you coat maybe 10 microns instead of having a thick dense warpable tape, right? So that's also good for the battery, I think overall. And being here, I think at a lower temperature budget is very significant. And I think this should be able to upscale but we haven't done it yet. It's early days. We just disclosed this new opportunity in processing and I think this is still to be shown. But I see that as perspective, yeah. Yeah, I fully agree. And I think in addition to the absolute temperature, right? Which also sets the thermal budget for your process. I think the requirement for temperature uniformity if I'm not mistaken, it's probably gonna be more crucial at the high temperature. So I think my impression is that for high temperature centering and processing of films, of thick films, it's gonna be very demanding on the equipment themselves, which I think will also dictate the cost if I'm not mistaken, would you agree with that? I agree with that fully, yes. And I think again, there's a big difference whether, so you could argue, well, there's maybe like production volume of 10,000 fuel cells or year or something, but we're going here in a billion or something for battery. So it's a very different production volume, but even if you anticipate that, you can change face, nature and chemistry. And many of the seam conductors are having significant much challenges to face stabilize over wider areas. So it's a different game, I would say. Absolutely, absolutely. And maybe let me ask one more sort of detail question before we zoom out a little bit. So on the scale up, have you considered throughput and how important is throughput in this case in determining the cost? Do you see an advantage for a process that could, for example, maybe Elsa was making this point, maybe the performance or quality is not good, but the throughput is very high. So thereby giving you the proper trade-offs, I think certainly as scientists and engineers ourselves would like to make the best possible material, but economically, there's always a trade-off. So maybe you can talk a little bit about what you see as the potential trade-offs between cost and performance and processing, and where with the sweet spot why. Yeah, that's a great question. So I think in terms of throughput, I, if it's about me, so I'm not sure whether it's so significant, whether you are exactly at 10 over minus four, 10 over minus 15 minutes per centimeter, as long as you're very slim. And you have, for instance, the other side. In terms of throughput, it could mean that local variations at low temperature process in, you know, slight structure changes are not as relevant as at high center temperatures, but may allow you to deposit more quickly. So we have to consider, even though in terms of throughput, right? If you think about it like in traditional ceramic processing, you have to densify. So that's one step you can't take away per piece. So your throughput is limited by that to do that. And then later you can maybe like, and that's great work from Liangbing Hu and others from the Maryland for sure you can in some seconds actually bring it on very high temperature. But then still in terms of throughput, you have to core center a cassoed. So if I can take away as a core centering and some of the densification up there, I think in throughputs that may be more competitive to design other process lines later to take it. And I think later in throughput, one could imagine to have like a roll where it just goes through and you have like various sprays arranged, right? That could be quicker to count than to densify from powder. And there's different like products where that is done, right? So there's a likelihood. But now we have to also be realistic that coating a car color with some spray is a different dimension and then coating, for instance, as ceramic, yeah. And it's unclear for now, for instance, how much do sickness variations play a role? Is there other like chemical compatibilities with other components? I think there's like still it's very early days, yeah. Thank you, Jen. So maybe zooming out a little bit, I always like to draw analogy between different fields. And I know you are a very well positioned to do. So having done quite a number of things tied together by ceramics, in this area of thin electrolyte for solace the battery, where do you see an analogous field which has achieved the necessary success in the scale? Or is it something completely new that it is really there's no example to look to elsewhere in ceramics? That's a fantastic question. Looking at the specs, right? So it's orders of magnitude away from, for instance, fuel cells or supercabs or other tech where I think there is a feed up in rising in ceramics is centers. So there's like more and more need for centers, for instance, where you would also use more similar tech, I think, to this, right? But still it will never be at the numbers if we think about 2030 and having 30% electrified cars driving, right? The production volume is huge. Now, the question is like how competitive will sold state batteries be by then or hybrids? And how much will be replaced in that share, right? So I think it's an unprecedented case for ceramic industry. And also it's, and that's my point, right? In terms of material and phase, it's unprecedented because many of the other ceramics have lower numbers, but then maybe complex chemistry. I mean, alumina, for instance, right? So, that would be one of the most produced ceramics that are around, but they are not, in most cases, electrochemically functions ceramics in the use case. So this is really new. And that is why I think it is very relevant for the fear to not sleep and do great fundamental research, but collaborate with industry to find a way to make a high volume of these functional electrochemicals remix. I think it's unprecedented. So I love your question. Well, unprecedented requires innovation. So this is great, great problem deep. So, Jen, normally we don't take any more Q and A's at this point. However, two of our esteemed colleagues have questions for you. So I thought I would just take the final few minutes to ask them on their behalf. So Stan Winterham has a question. So do you think that the electrolyte can be used at slightly elevated temperatures, say about 100 degrees C, to improve performance? Any challenges in raising the cell temperature? That's a great question, Stan. So thank you so much also for attending the talk. So I do believe that you can operate some at higher temperatures. So once we manufacture that at about 400, then dissociates the lithium salt, go to crystallize at 650, then the structure is stable. So it should not be a problem to be about 400, 500 degrees lower in operation. And as a rule of thumb in ceramics, so you normally free, you kind of like stabilize the structure at about 1200 degrees higher. So having a gap here of 400, 500 should not be a problem on long-term stability because diffusion is not as activated anymore at this point. Right, and his second question is, can the interface between the cathode and the electrolyte when used at this processing route, is it possible to eliminate the use of a liquid as an interfacial layer? That is a great question. So it is possible to directly spray bilayers of a cathode electrolyte. So we have one example, but we have to do more in the electrochemical testing to say the truth. And I think potentially it could be because in microstructure it's like really these 5, 10 nanometer grains over 10 microns, it's razor dense. But now I think later for production, we're gonna have to run helium leakage tests through some standard tests to also confirm for so and so many 100 cycles running with the liquid will still be stable. But I think there's option to engineer that without a liquid by playing now here very carefully with interface engineering and the design of the cathode. But I think this is, to be honest, another two to three years of research of my life to look at the cathodes here, but it is on my to-do list. And I am planning so to develop actually these chemistries with the SDS technique here in the next years. Up and coming, great question. Excellent, Jen. Thank you, Stan, for the great question. And I also have another question from Linda Nazar who would like to ask about the advantages of amorphous LOZO. So she highlights, of course, it has been shown that dendro issue can be a problem even in single crystal. So she would like to understand the advantage of amorphous LOZO beyond the lowering the temperature of processing. Yeah, thank you so much for the great question, Linda. So that's a good one. Right, so what was shown in single crystal experiments by the colleagues where that wherever you have a defect you're prone to formalism dendrite. However, now we have to think about what's a realistic scenario in a cell and there you won't have a single crystal. You know, very bad manufacturer you would have something with grain boundary and grain microstructure. So where the lithium dendrite's way to go first and predominance in many of the lithium oxide based materials is along the grain boundary in forming the dendrite. So by having now an amorphous structure that's probably why Lypon works so well that it's also stable towards lithium. This could be interesting to just, you know code a little bit as a buffer layer and one thing that I will highlight here is the amorphous films we can make them at three 200 degree seam. So this is suddenly close to where you make sulfites. So technically it could become an interesting protective layer to bring on sulfite cells for instance for lithium to avoid mitigation or it could be a small layer you just bring in by nature and process after having fat for instance the soil state battery electrolyte core of the lithium oxide based materials. So, but there we have the first TM data where we see it stable with lithium for the amorphous phases but now we have to do more electrochemical investigations and also highlight and it's a three four year project with Claire and some other colleagues where we try since three, four years to write a paper on what are these amorphous structures of lithium garnet. So give me another half year and I hope we have finished them and really understood and can publish what they are actually. So that's to be resolved first. Well, thank you, Jen. We certainly hope to have you back very soon to speak and Linda, thank you very much for your question. So I think this brings us to the conclusion of the symposium but Jen, I was wondering if you have any words of wisdom for our viewers especially students in postdoc who are just beginning their training and their professional career. Oh, that's a great one. Yeah, so first of all what I wanna encourage you you can do more than vote green. You have all those amazing brain power. So don't waste any minute and think about the toughest problems we have in the field and do those. So I would love you to encourage your supervisor and really challenge them and asking yourself every day when you walk to the lab or do computer science is that really one of the toughest problems that is right now existence? The risk for failure is high but be courageous, be brave. You're allowed to do so. And I think this is what we need as a field to strive, right? And really make it forward and we have here an option to exchange things for society. So by doing good chemistry. So I hope this motivates you and I look personally forward to see all successes of the next generation coming for sure. Thank you, Jen. It certainly motivates me. So this is great. And Jen, thank you very much again and I also like to add my thanks for Elsa for joining us. Was truly wonderful to spend two hours talking about scaling up and supply chain for solid state batteries. So I'm also pleased to announce the rest of the schedule for this spring quarter. So in two weeks, we're gonna have a wonderful panel discussion on the connect the intersection of energy storage and building energy efficiency. And we're gonna have a really outstanding scholar and practitioner to speak to us on May 21st. And on June 4th, we're going to have Frank Blum who is the head of the battery business at Volkswagen who will discuss the latest strategies there. And then finally on June 18th, we will have a deep dive on battery cathode chemistry with Professor Yankuk Sun at Hanyang University and Hooper Gastiagar at the Technical University of Munich. So just a reminder, we do a lot of our information sharing on our LinkedIn page. So if you're interested, please join and follow us on LinkedIn. And I also wanna mention for those of you who are interested in learning about energy technology and the energy transition, Stanford offers numerous online courses that you can find by going to online.stanford.edu. And with that, I'd like to thank everyone again for joining and have a wonderful weekend.