 It's my great pleasure to welcome everyone back to DistorageX International Symposium. My name is Will Chu. I'm a professor of material science engineering here at Stanford University and the core director of the StorageX Initiative. First, I just want to thank all of you for attending our symposium. Although maybe it's a little hard to see, we have enjoyed your attendance ranging from 1,500 to over 3,000 viewers at one time. In this difficult time of the pandemic, we're really excited to have sponsored a forum where everybody can get together, discuss science and think about the next steps. And to continue on that excellent set of talks we have heard over the past five weeks, it's my distinct pleasure to introduce our two speakers today. Speaking first will be Professor Yetming Chang from MIT's Department of Material Science Engineering. Yet is a true pioneer in the energy storage field. He has been working on energy storage since the 1990s. First approaching it from the perspective of electro ceramics. And then continuing on that to innovative chemistry, whether it's synthesis characterization properties and also understanding the many technical and automatic and business cases for it. Yet is also an extremely unique innovator in energy storage. Not only is he an innovator in science, fundamental science and understanding all the underlying material properties necessary for energy storage to function well. But also he is an innovator in commercializing and carrying out the very challenging task of tech transfer, taking laboratory curiosities to reality. And yet has started numerous companies, not all of them in the energy storage area I thought I'd just mentioned them. He started Americans Superconductor. A123, 24m, form energy and desktop metal. Yet I apologize if I missed a few. And it's extremely fun to see all the things that have been translated from his laboratory to the real world. And I'm excited to hear from yet today on understanding all the challenges, not only scientifically but also what it takes to take the technology and make it real. Yet is very accomplished. Scientist engineer. He has received many awards. He's a member of the National Academies of Engineering. And importantly, he has those trained many people. Some of whom are starting companies of our own and making contribution to the energy storage community in every possible way. Our second speaker, Professor Itui, my colleague and co-director of the initiative will be speaking second today. And he is similar to yet in that he's also an innovator, both scientifically and also on technology transfer. He is one of the most prolific scientists in material science. And has started three companies, Amprius, Forseier, EInnovate. And he too, like yet, is fully committed to seeing scientific innovations come from the laboratory to the real world. And I'm sure he will also allude to some of the many challenges and what it takes to get the technology to the real world. Yet, let's start with you first. I'm really delighted to hear from you today. I hope you'll share many of your exciting and perhaps crazy ideas on what the next steps is for energy storage. Yet the floor is yours. Thank you, Will. And thank you, and good morning, good afternoon, good evening, wherever you are listening in from. I thought I would talk about some things a little bit different from the previous speakers who have spoken in this symposium and whose talks I've listened to. So my title is energy storage from the macro to micro perspective. And so first I wanted to on this slide acknowledge the various programs and sponsors who've funded the research that I'll speak about. What I mean by the macro perspective here is that I'd like to begin in the next slide here. I'm talking about my personal take on, you know, why we do the research that we do. And specifically, you know, in this energy storage area. And, you know, if I look at the pie chart here of global GHG emissions, you know, the work that we do in storage touches on many those, and I'll just refer to the ones that I personally involved with. And so transportation, of course, we're all, all of you are aware that it relies on higher energy density, longer life, lower costs, safer batteries. And so, you know, my group has a number of projects underway in this area related to solve electrolytes, for example, but you already heard about that from your organic and Linda Nazar. I won't talk about that today. And we are, we have some work on electrolytes and in particular interfacial transport in electrolytes that just was published last week in Nature Energy but you know you've already heard about that so I'm not going to spend any time talking about that. And we have some projects on lithium metal interfacing interfacing with sound electrolytes. And I also won't talk about that. And that's supported by both DOE and now Aurora Flight Sciences because of the direction of, you know, electric aviation. And then we have good scale electricity. And I will talk a little bit about that and I'll spend some time trying to explain where the needs are and what we should keep in mind when we think about that particular problem. And then I have one other project that is really far out there, which is sponsored by Google, which is actually a fresh look at the problem or the potential of cold fusion. And so that really is the most speculative and far out problem that I work on. But the work that we do there is really focused on trying to create unusual new materials in the form of metal hydrides. And then I'm going to spend a little bit of time talking about electrochemistry applied to industrial processes and what I think is a slightly different way of looking at how we as battery scientists and electrochemists can contribute to that problem. Now, there's a big part of this pie chart where I haven't yet done anything professionally, which is the lower right hand quadrant, you know, agriculture, right. But I do practice personally some agriculture and so here's a view of our chickens, both the meat chickens and the egg chickens in our backyard. But I want to tell you that I'm completely convinced that electrochemistry can bring much to agriculture and we're trying to identify the right kinds of problems to solve there. So, this is the scope of the things I work on. What I'd like to talk about today, first is the problem of storage as it applies to enabling a renewable grid 100% renewable electric grid. And it's a macro perspective because I'm going to start from really the top down view and look what is necessary to make this happen. And the first thing to point out is that, you know, the way we generate electricity is actually rapidly evolving today. And in this narrow there are winners and losers. In the left hand chart here, what you see is that, you know, coal is declining as we all know, and it will continue to decline, but purely for economic reasons. You know, we might give it a push, but it's on its way out and we'll continue to go in that direction. And the main reason for that is the rise of natural gas. Nuclear is relatively flat. Conventional hydrogeneration is relatively flat. The only two that are increasing are renewables and natural gas, but you see that natural gas is about a factor of four higher than renewables. The reason for all this is apparent, I think to almost everyone, it's in the lower right hand chart here. It's the fact that, you know, in much of the world today, the lowest cost electricity we have available to us is renewables. And so one looks at this and you might come to the conclusion that natural gas is the natural winner here. And what could change that, though, is if we come up with a breakthrough in storage. And so what I'd like to now talk about is, you know, what would that breakthrough have to look like in order for for we to beat natural gas. And so that's a simple objective. Let's beat natural gas. And you get the impression from looking at the rise of grid storage based on lithium ion batteries that maybe lithium ion batteries are going to do all of the work for us, but I'll try to show that that is unlikely to be the case. And so over the last few years, the installations have been dominated by lithium ion, but at some point I really think that that is going to start to change. Okay, so, you know, lithium ion initially was put on the grid for for short term, short duration, high power applications. The photograph here is Laurel Mountain West Virginia and at the time that was the largest lithium ion farm in the world. Now it's eclipsed by quite a few others, several hundred megawatt hours. And the applications in that case were to help with wind ramping into the grid that, you know, when the wind energy came too fast, the grid couldn't handle it in that location. And you have similar applications with the duck curve that most of you are familiar with on the left, where the early, you know, evening demand is what creates a short term storage issue. And also recognize that in this whole scenario, the amount of storage we actually practice today is very small relative to total power capacity generation. So of the electricity we generate very, very little is stored is the little blue bubble in there, and then that tiny little circle is what we actually do with batteries. So, so grid storage today is a very small part of the electricity ecosystem. Okay. How much storage would we actually need to fully decarbonize the electricity system by I'll choose the date of 2050. And this is a problem that at the MIT energy initiative we have been looking into we're conducting a study called the future storage study. I'm sharing that study with Bob Armstrong, who's the director of the MIT energy initiative. And so it's still, you know, about a year from coming out in print. But I can tell you a couple of things that we're doing. One of these is to try to estimate how much good storage might be needed in order to fully decarbonize the electricity system. The estimates here come from a group, Patrick Brown and others. And there are a number of caveats here, but you see what the headline is. The headline is 100 terawatt hours of grid storage by 2050. Right. And where does the uncertainty in this estimate come from. Well, a lot of it comes from the fact that, well, we don't know how much new transmission will be built. We also don't know what the future cost of storage will be built, but that's the one that, you know, with better research of some influence on. And we also don't know how much the renewables will be overbuilt because the two kind of go hand in hand. And the new removal overbuilding and storage are the tradeoffs on the same problem. Transmission is actually lower cost in storage today. But the US is not quote a copper play. Okay. And so there are significant buyers of transmission build out. And so one has to make some assumptions here. But essentially, with the early assumptions or early sorry early predictions here are estimations here say is that we might need on the order of 10 terawatt hours with us, multiply that by 10 for the world. It's 100 terawatt hours. And that's not counting the electric fleet. So to put 100 terawatt hours in perspective. That's equal to about a billion Tesla Model S cars. If by 2050. We have nine billion or so US living on the planet with really talking about one Tesla Model S size pack for every 10 people or so to kind of give you a scale of what that might mean. And so the question we might ask is, you know, so what do we need in order to meet that 100 terawatt hours. So the question of, you know, what lithium mine can do comes up over and over again. And I want to show you a couple of just two slides from some work that is ongoing, led by also also all of Eddie and Bob Jaffe colleagues mine here at MIT who are working on this feature of storage study. And the resource limitations are always in lines. And so lithium is one of them for sure. What this plot shows you is on the horizontal axis is on a long scale, how much storage you might need. And so you see our 100 terawatt hours there. And compared to that, you know, the resource limit is actually higher. So there is actually enough lithium available. But the bigger problem is that depending on where this lithium goes. So here we have 100% 50% 25% what that means is that the additional mining of lithium 100% of it goes towards, in this case, lithium ion batteries versus 50 or 25% of it. And then over on the left hand side, we have an axis which is compound annual growth rate of production, which means of mining. And so we would need to, if by 2050, we reach 100 terawatt hours, we use all of that lithium for lithium ion batteries, we need to increase production at a rate of 10% per year now through 2050. And if you look at the historical scaling production, which is this these yellow bars over here, we define this that it's not the upper limit. And so the basic message here is that it's a very heavy lift to get there. Not theoretically impossible, but a very heavy lift. If we look at cobalt and nickel and what would be required there. And so of course it depends on how much cobalt and how much nickel and that's why you have 811 and 622 shown here. We've left off 111. What you see is that the story is somewhat similar. The total resource would actually exceed the 100 terawatt hours you need, but the compound growth rate in production is large in all cases. And so it actually starts to tell us that, you know, maybe it's not the amount, but how do you scale mining at 10% worldwide a year now through the year 2050. So, as I said, very heavy lift. Okay, so let me then talk about how, you know, what kind of batteries might we need that might serve that function, which are not lithium ion. And describe a study that hence that was published, I think, in 29 late last year. And this is a case where in collaboration with Jessica transit, who those systems analysis here in MIT, we looked at the way that electricity is generated today and tried to predict how it could be done with renewable storage. And so the beginning part of this is just to tell you a little bit about how electricity is generated today. And they're basically three types of power plants in the US. We have no base low generation, which for example includes nuclear, but also others. And this is just 24 seven flat power. And so we have intermediate generation, which is about an eight hour block and then we have what we call peakers which about is about four hour block. And so talk about a study in which we greatly simplified the production, the output profiles of these into either just flat line, or simple square waves. The reality of these situations is that the output profile. So if one were doing this in a, you know, targeting specific applications and specific regions, you would have an output profile which much lumpier and more nuanced than when I'm showing you here in these simple square waves. But you know we're beginning with these simple square waves that have a rated power and a finite duration. Okay, we pick four locations in the US. And these are intended to choose a diversity of resources, and that big purple belt in the middle is the, you know, is the wind belt that's where the dust wall was. And so, and the north of the upper Midwest and then West Texas, Arizona, clearly a solar rich state, Massachusetts, poor solar poor way. So using those locations, I want to just show you in this chart, this is Iowa, where the wind is somewhat better than the solar. And what this shows you is at the end of the analysis, which what the analysis does is to calculate the combination of renewable generation and storage that gives you the lowest cost of the liver to electricity. The lowest LCOE. And, you know, we make reasonable assumptions here for how much wind and solar will cost you see them in the lower left here. 1500 a kilowatt for wind 1000 kilowatt for solar. And this is a calculation for essentially 100% availability so you can always, you never miss you always deliver the electricity that need that you need the study takes into consideration other cases such as what if you only put the need to be 95% reliability 5% of time people don't care what they want in terms of their electricity. And so results will vary. So there are nuances there. But the main thing here I want to point out is that the scale vertical scale is the state of charge of the battery 100. And so this goes all the way down to a fully depleted battery. And of course, that's the sizing of the battery there for the worst case scenario where you actually fully discharged the battery. And over the horizontal axis this year is 20 years. And this is based on backwards looking availability resource availability data for those for these for this location. And the main point here you see is that, you know, you have relatively infrequent, but multi day deep cycles deep discharges, and the purpose of the battery stays mostly charged. And you can certainly use that state of charge for other functions, but you know you need to cover these long durations of several days. Okay. So what does that mean in terms of the cost of the battery that you would need to do it to do this service function. And so this is one way of representing the data. The cost of power on the vertical axis in dollars per kilowatt and cost of energy on the horizontal axis dollars per kilowatt hour. And the main thing the color coding is to help with the, with the eye. And what we're trying to be here is CC GT in the blue. What this shows you is that in order to get down into the blue. First of all, the cost of storage on an energy basis has to be less than about $30 a kilowatt hour. The sensitivity to the cost of power is not so great. These are relatively vertical lines rather than horizontal lines. So you can tolerate some variation in cost of power, but cost of energy is super important. And then if you have kilowatts and kilowatt hours, of course, the slope is related to hours or duration. And what you see is here, you know, here's a case of 40 hours. And here's case 100 hours. And so you see the multi day aspect coming out. That's a four day storage cycle. And the other thing I'll point out is that, you know, lithium ion, I think is for for the grid is way off to the right about $250 a kilowatt hour today. So it's not even on this chart. Okay. So what we see is that we need something less than 30 kilowatt hours and out of four four day duration here. If you look at, if you take that kind of a plot and now just apply to the entire study that we did here, we have wind and we have solar, and there's also the combination of two, you can change the ratio to to optimize. And we have here in these, you know, these quadrant charts, wind for the four applications, you know, baseload intermediate load, peaker and there's something called a by peaker, which is just two cycles in a day. And then we have over here the storage capacity. Well, this is the four states that I talked about. And basically what you see here is that Texas is, you know, is good and wind and good and solar Arizona so many really only good and so on. What this leads us to is that in this study, we said, let's look specifically at a couple of different battery options. Let's consider one that's lithium ion like. And what lithium ion like means in this case is that it's a future looking perspective $150 a kilowatt kilowatt hour, you know, install. And the cost of power is relatively low $700 a kilowatt. And then a future low cost chemistry and I'll give you an example of that in just a bit, which is $20 a kilowatt hour so super low cost on an energy basis. And in this case because it's going to be a, you know, imagine a flow type batteries cost of power tends to be a little bit higher. And what the vertical axis here is is the dollars per kilowatt hour of delivering electricity. And each one of these small lower bars is the low energy cost option technology to the higher bar is the future lithium ion. And the color coding here is across the four different regions. And it tells you whether you choose to use mostly wind, which is the blue or mostly solar, which is direct. And so what you see is that the lower cost on energy basis battery always wins. Right. Again, argument for technologies other than lithium ion, right, if you can get there. So this is a complicated complex view of how to beat natural gas. But I want to give you a really simple way to do it. You might say, why don't you just do this at the beginning. Okay, natural gas today and natural gas power plant. It depends on what type, but it costs you about $1,000 a kilowatt delivered. This is the capital cost $1,000 a kilowatt to deliver electricity. And so if we had to run that with a battery over 100 hour period would require about $10 a kilowatt hour. Great electricity today is about $250 a kilowatt hour for delivery a couple of years from now. And that explains why it's suitable for the four hour duration. But so now the question of what can do the 10 $10 a kilowatt hour. All right. Quickly, very few battery chemistries actually clear this cost bar. Right. And the ones that are of interest have to be well below $10 a kilowatt hour just for the chemicals, which is what's planted here. Right. And what this shows you is that, well, sodium sulfur could be attractive. Right. It has a problem if it's high temperature because that has a lot of cost. You see iron iron on here you see zinc iron. So there are a few options that might make it. And we found one option here, which is essentially a room temperature version of sodium sulfur. And so I'll show you that one quickly. This we published a couple of years ago. It's a battery that uses a polysulfide on the left. So if you look down on the anode side here of this, it's a polysulfide lithium or sodium, but sodium is preferred where we cycle over a stable polysulfide speciation range and don't precipitate it and don't take it too far to where it's unstable. And then on the cathode side, we actually have a oxygen reaction. And what the sodium does is to change, as it crosses over, changes the pH and causes the oxygen breathing reaction. And so sulfur doesn't directly react with oxygen, but it's really a sulfur air battery. And this is one that has ultra low cost. And so this is the cost of the battery. And this is the duration over which it operates. And it has a curve because there's a power stack here that a short duration raises the cost. But the floor is the energy cost due to the low cost chemicals. So this would be one of these. Okay. One important point I want to make on these new battery chemistries is that there are going to be some important trade off. So a friend of mine likes to say, you know, batteries are just like people. Every one of them is deeply flawed in some way. It's just one of the flaws that you can live with. And if you look at our data for this battery that we publish, if you look at the discharge and charge curves, the round trip efficiency is only about 50%. And then in fact, you know, that was a big problem for our reviewers on this paper. And that's that. Is that really a battery? What kind of battery is this? Only 50% round trip efficiency, right? But on the right hand side, I will show, what I show you here is a plot which shows the levelized cost of electricity for a low cost flow battery versus the round trip efficiency for that battery. And on this plot, it looks pretty steep, but really, it amounts to only one cent a kilowatt hour difference in the levelized cost of electricity that you deliver. If you go from, for instance, 80% round trip efficiency all the way down to 55. So the message is that round trip efficiency is something that we can give up quite a bit of if we have very low cost chemistries here. Okay. So the kind of batteries we envision look something like this. This is a battery that has a power footprint like that natural gas, which you think is possible based on calculation, the calculations. One to two megawatts per acre, right, which is what a natural gas peaker does, and would occupy footprint like this relative to a the wind farm. And so we started a company, and this is a photo of five founders, and it's called Form Energy located in Somerville, Massachusetts. The CEO of the company is the fellow on the far left here, Matteo Haramio, who was formerly an executive at Tesla and in fact started the division called Tesla Energy. And so I want to just give you one recent headline. Long duration breakthrough. Form Energy's first project tries pushing storage to 150 hours. There's 150 megawatt hour of battery. And so the 150 hours here, you know, if you're a battery researcher, you might say, hey, 150 hours. Well, I can do 150 hours. My cell phone can do 150 hours. I just discharged it slowly. What I hope I showed you is that the point is not that you can discharge the battery over 150 hours. But but that is cost effective for that kind of an application because of the need to reach that no cost of electricity. Okay, so that's the first topic I wanted to talk about. I wanted to talk about another kind of kind of top down view on energy storage and electrolysis, which we all know and know quite a bit about as a means of chemical energy storage. And the image I am showing you here is a neutral water electrolyzer. So this is pH seven water dissolved in it is sodium nitrate. So what you're seeing here is a pH scale. And what it's showing you is that on the right hand side of producing a base on the left hand side of producing an acid. Of course, at the same time of producing hydrogen and oxygen gas. But rather than thinking of electrolysis is a way to create hydrogen, which we store as energy storage. Let's think about storing an acid in a base and not so much for electricity production, but for energy storage for other purposes. What is that other purpose, the one that we've been looking at is cement. So that today has the following problems. It is a giant CO2 emitter the largest industrial production process, the largest emitter ahead of steel. And that CO2 emission half of it comes from chemical CO2 due to the limestone and the other half comes from the thermal energy in the process. So what is limestone? Limestone is calcium carbonate. Here's limestone using architectural setting at MIT. It's just calcium carbonate and some impurities. And the cement that we talk about is the paste is the glue that holds to get to the gravel that we then call concrete. Okay, and so it's a four gigatons of CO2 per year produced from cement, one kilogram CO2 per kilogram of cement. And so what does the future look like here? If we look at the urbanization of the planet, and it's predicted that in 2060 they'll be double the number of buildings on earth that we have today. And what that means is that for the next 40 years, we will build out the equivalent in cement and concrete of one New York City every 30 days. So that's what it looks like. Where does the electrons is coming? Well, if we do what I showed you and produce an acid in the base in the electrolytic process, what we can then do is to take our lime and dissolve it in the acid, which is on the left here. Evolving CO2 along with the oxygen we produced at that electrode. And what this gives you is clean pure CO2, which then can be sequestered or used for actually more valuable functions than just being sequestered. The chemical process that you can carry out if you simultaneously have the base present is to take that calcium that's now dissolved and at pH 10 or 12 or so, we precipitate calcium hydroxide. And so this calcium hydroxide can now be captured on a flash forward here. And I'll just show you what the calcium hydroxide looks like here. This is an electrochemical reactor for making calcium hydroxide. This is cement, the active phase of cement called alite made with that calcium hydroxide. And this is what it does on the left is natural limestone. If you look through a reactor of this kind. First of all, it's selective, it purifies, gives us pure hydrated lime or calcium hydroxide in this case, and the precipitates are insoluble. And so what this leads us to is a scenario where now renewable electricity could be used in an electrolytic setting, producing hydrogen if it's a water electrolyzer, not only for energy storage, but for chemical storage, which we can then use in a number of reactions. Yeah, thank you very much for the wonderful talk connecting the macro to the micro. So we have time for some questions now. And maybe let me first begin by asking a high level question. In the systems analysis portion of your talk yet, you made very explicit comparison between electrochemical energy storage technologies and mechanical. So I see palm hydro feature very prominently on your plot. Can you also speak to a little bit about the comparison to thermal energy storage. Yeah. Yes, happy to so thermal storage. I think the first thing to be said about thermal storage is that thermal storage is a very, very low cost technology. If what you need is he, if you start as he, and you use it as he. And there are actually a number of reasons to do that. You know, those in the nuclear industry, for example, are now looking the much renewed interest in the heat output of nuclear reactors and how they can be used in industrial settings. And so that's, that's, that's a perfect, perfect application. And I think that thermal storage where you then have to convert back to electricity is an electricity that you want is that conversion cost that goes goes up quite a bit. And there are thermal. So, there are some technologies that are being looked at today. For for storing it as he and then re reacting back to, and then converting back to electricity. And I think, you know, my view on that is that that is potentially competitive with electrochemical storage. The energy density of electrochemical storage is probably going to be still a bit better. But it comes down to a system level cost. And I don't think we know the answer to what those are really going to be just yet. But I want to make it actually another analogy though the when we look at very low cost electrochemistry. We're doing something very similar to pump hydro compressed air and even thermal storage. Low cost electrochemistry is providing us with the equivalent of a very low cost working fluid. In the same way that water, air, heat of a very low cost working fluids. And so, when we have those very low cost chemistries. The cost starts to shift much more to the rest of the plan, how you journey the power or the balance of plan that you need, how you manage the fluids. So in that sense, we're, you know, we're in a similar technological realm, I would say. Thank you yet maybe just a quick follow up on that for me. So I think one of the major theme of your talk today in the systems analysis is that for long duration storage battery, the power requirement is very low you are charging and discharging. Especially you're discharging at fairly low rates. Do you think that could become an advantage for thermal storage if people start looking at not short but long storage. I understand one of the key challenges for thermal today is that charge and discharge power is a problem, especially if you want to convert it back to electricity. In the sense of power, it's a it's a potential advantage that helps thermal storage in terms of self discharge it's not. And so, you know, one of the issues with it depends on the scale of course, but in thermal storage you have, you know, heat leaks out, right. And that self discharge. And so the largest system and better off you are right. And so it's very system is going to be so much system size specific. And so that trade off the the thermal storage, the technology development efforts that I know of today are still mostly focused around less than a day of storage maybe a little maybe longer than four hours, but not yet multi day storage. Thank you yet so sounds like it may be better in terms of power relax and power requirement but the self discharging issue will become more significant. But maybe just have to make it really big. Right. So the second question on the systems level has to do with your perspective on transportation batteries. So you discuss a lot of the challenges and opportunity with grid level storage, but certainly we are nowhere near the desired cost level for transportation battery technologies. And, you know, in recent years, we have really sort of seeing a renewed interest in less expensive but lower energy density options for transportation batteries for example, in the iron base chemistry like lithium iron phosphate with, you know, several major implications of the return of LFP to transportation. I was wondering if you can talk a little bit about sort of your perspective and potential pathways to seeing, say something of an iron or make any space chemistry that is potentially much lower in costs than chemical costs, but then could have the possibility to compete or or or approaching to compete on other performance metrics what does the pathway look like there. Yeah, as you know, as you know, lithium iron phosphate is very near and dear to me. I think also not only for transportation, but lithium ion for the grid. You know, if you look at the low, what are the lowest cost lithium ion options for the grid. I think it's going to, you know, evolve more and more towards LFP. A number of things are happening with LFP. So first of all, you're one big, of course, being on base doesn't rely on the cobalt or nickel but that's on the lithium. So the one straightforward option, I think is attractive would be LFP coupled with lithium metal, for example. We're going to solve the lithium metal problem. Then, you know, whether it's with a solid electrolyte or with a liquor electrolyte, we can take advantage of that energy density boost and still, you know, get to the get and still be competitive. You know, the best performance cars, perhaps, but for many of them, if we use LFP as the cathode. So that's one that's certainly on my mind. Thank you yet. Can you comment in particular if there could be a pathway to increasing the energy density of iron or making space chemistries on the cathode side of lithium ion batteries. You know, certainly, we know the voltage increase from iron to manganese phosphate. So those those mixtures are known and have been fairly well developed. I think that those those are possibilities. Other compounds, I think is really what you're alluding to are there other compounds. And so, don't have a a favorite or a particularly high potential prospect there. But maybe others do. Thank you yet. So maybe we can also get another question on more of the technical detail side. So, on the second part of your time on the, the software based chemistry, one side of the electrode is for electron transfer process and other stuff also proposed to electron transfer process involving other gases. So this is a bit counterintuitive to use multi electron transfer because they're inherently more sluggish. Can you talk a little bit about sort of the trade offs by going to multi electron transfer, both in terms of power and energy that you're playing with here in the technology or a commercializing a form. Yes, let's see. I can tell you what some of the challenges are, you know, we publish these, we publish papers that are that are not, you know, of course, talk that describe directions in in the technology. And the problems are not fully solved when we publish these papers. And so I would say that, you know, the, the biggest challenges. I really pointed out the round trip efficiency challenge. Right. The big part of that is actually the membrane. Right. It's not so much the kinetics in on either side. So the kinetics on the air breathing side. But I should just back up one second and say that in general, if we look at the, the really the bottom of the scale in battery chemistries from a cost perspective. My only belief is that the lowest cost ones are going to be some very low cost in organic with an air electrode. And so if you accept that, then the challenge is the air electrode in terms of kinetics. Right. But really, I think much more so than the than the other side than the other side. One of the reasons that in that everything a crossover case, our efficiencies the maximum we ever measured was around 70%. That was limited by the air cathode. Right. And so at the lowest possible current densities you talk about for for the reasons that many of you are aware of, you know, we can't really get much higher than about 70%. But I also pointed out why that shouldn't really be a factor. The, the reaction kinetics in the polysulfide are not limited. But the membrane is and so the membrane in this case is the biggest challenge is that it has to be both high conductivity and blocking, but blocking to both acid and base. Right. And that's a particular technical challenge. I think that membranes that can separate strong acids and strong bases is an area of research that could be very, very enabling of a lot of new chemistries, if we can come up with membranes of that kind. So I hope that helps. So the kinetics really the the the limitations on the air cathode should I think what you were looting to, but not so much the polysulfide, but the membranes really have to spend a lot of time worrying about. Thank you. Yeah, that's good to have another challenge for us material scientists. So I think now is a good time to finish up with yet. Thank you very much and yet stay with us. We still have a panel discussion afterward. And now we'll search over to my colleague each way. Okay, thank you. Well, well, yet very, very nice talk. And when we was given introduction, I, I look at what I have been doing and say, wow, you know, I think what do your entrepreneurship has been inspirational to me. You know, I'm trying to follow your footsteps. So let me now switch topic quite a bit, you know, yet is giving you the macro to micro perspective, you know, the particular grayscale for, you know, also for the larger scale using electrochemistry CO2 capture and so on. So let's go down to the nano scale for the materials and the interface, the understanding and design. Well, I want to now come back to this and after lithium ion batteries, right, the Nobel's was giving very successful commercial activity right there. So we asked these questions. A wheelchair and I, you know, are leading storage acts initiative at Stanford. So we asked these questions. Let me give you some of those grand challenges. We are talking about right here. So, and what are the big problems we want to work on for storage. Now, by the way, is can we keep increasing the energy density, while per kilo or while per liter, right, it's going from 250 while per kilo to 500. For the audience, you all know about battery 500 consortium that is really trying to solve this challenge right there and for the transportation. If you go to 500 while per kilo with a reasonable size battery pack we can go to 500 miles driving range. So, what can extend the battery cycle life even longer, where 10,000 cycle 30 years together, you know, certainly you could see some of the battery probably reaching 5000 might even 10,000 but you need to think about the constraint as well can I keep the same energy density. So, certainly not LFP LTO combination can give you very long cycle life by your cut energy density by half also remember that. Can we, you know, maintain the energy density having very long cycle life right then there's many materials the chemistry issue start to jump out. But charging, can we do it fast enough, you know, 10 minutes or less. I mean this can change the whole equation, even for the grease gas storage right so when you have, you know, certain dollars per kilowatt hour of electricity if you could boost up the power and the power cost is reduced as well. Also, also the safety we need to talk about safety somehow the safety is not emphasized enough you know we keep seeing the accidents coming out in the grease gas storage you see the burning in Arizona you see the burning and and South Korea, many of those large scale energy storage for certainly for cars you know every day this this happened. The cost down below $50 a kilowatt hour I mean from the long duration when it needs even more and yes presentation you have seen it. And how do we sense and know the health condition of the batteries, I will say we this is the part we probably have. We have done it poorly as as the whole community and battery so far it's only two terminal device right there we rely on the wattage the current the impedance and so on to tell us some information. What the local strategies for battery use and recycle and the grease scale and seasonal even seasonal storage right there I think any progress we may and any of this bullet point will be super exciting. Let me give you a summary of over the last 15 years my group have been doing try to address some of the challenges. And the next about 2530 minutes also of course no no no way to go through all of this but this is all this challenges higher on my to do list. So the nano scale perspective understanding and in design or materials interface. So I'm showing you right here is a micron particle right when I use the nano scale. I mean to all of them think about this what does this mean to you nano scale. Does it mean to you just simply go down to the particle size going to the nano particle some sort of nano scale. That's the case in your mind. I want to open up your mind a lot more. This is only part of it. If you only see going down to smaller size. I think that's quite narrow. And then you are going to see many problems high surface area to high surface area too much chemical reactivity you say no nano would not work right so that's that's not that's very narrow definition right here I want you to open your mind a lot more. What we're talking about a nano can include as well as nano scale coating how do you do the best coating to stabilize interface for example. It could also means how do you build those particles together forming secondary particle with nano scale control nano scale porosity and think about a mass and electron transport. And the phenomenon wise electron and iron transport needs to be considered and down to nano scale. You know some of the materials electron cannot move so fast that to insulating iron cannot move so fast. Some face transformation is deep in the nano scale we need to consider nano morphology and the wires tubes and you know different type of shapes allow you to you know maximize the electron transport. You should also consider it in the actual scale you're taught you all to design all those thinking you need to come down to the nano scale so it's very very rich phenomenon is it's not just nano particles. This why this is so important. Let me show you this plot first we want to increase energy density for example. This is the plot I'm making my postdoc a student, a previous student yeah you are now a postdoc in MIT. The physical assets is relative volume chains horizontal assets is the amount of lithium your store versus the host atom number, right we are really sitting on the left hand side right here in the commercial space this is the number we are using is roughly one to six. Once you increase the amount of lithium your store, it's going to go up more and more and more like this ratio just go higher, you know all this new material start to show up the volume expansion is going to change a lot more. This is the type of material we need to deal with to increase the amount of energy storage we have so the material design thinking needs to change because of this dramatic change. So I only want to focus on lithium metal on in this talk and due to limited time. lithium metal of course is well known as problem like it's a plating stripping problem. You know you need three million mile per centimeter square of capacity or higher so you are talking about 15 micron of lithium deposition and strip away 15 micron. And this is a challenge we have never been able to deal with successfully, you know, for the batteries during this plating process. And because of lithium we have this electrolyte forming this SCI and this plating cause the volume change is going to break the SCI somewhere because the plating will not be layer by layer uniformly. Again, a nuclear the hotspot and grow out this a dendritic structure this filamentous structure during stripping it's not uniform with strip where you cause the dead lithium formation. And before long and your battery will lose its efficiency and die fast and plus the safety problem we all understand this. So I want you to keep in mind is this a three million hour. One square is 15 micron you have to deal with how do we handle that I mean this is in the single electric scale if you have multi electric stacking up your batteries the whole battery cell is going to go crazy, swelling, you know shrinkage swelling and react reaction with SCI or the electrolyte forming SCI. So in this perspective and reveal papers, we highlight the center part of that. Lithium metal has its problem high chemical reactivity, as well as relative infinite volume change like this volume change going from empty states during stripping and to a field state during a deposition. The relative changes infinite unit overcome the rest of surrounding other problem you observe. So we really need to overcome the is fundamental. I also the root causes in the center right there. So that has been guiding my whole research group I believe in the many scientists and the research field will agree with me. This is a very important rule cause we need to overcome. So we have a build a team right here particularly with a collaboration of the collaboration with Steve Chiu and Jenan Bao. You know, one area we are building is, you know, learning from how graphite hosting lithium iron. And we need a host host lithium matter as well. Otherwise, this volume fluctuation will be too big. You cannot handle imagine that 15 micron up and down that change. I don't know how we are going to handle that without stable host. The next will be how to build a stable interface understand and building this interface is very important. So, speaking of that, if we think about lithium plating and stripping, do we really understand I mean the first thing you need to face is the new creation, particularly if you stuff is something like copper foil as a current collector. And what's the behavior of that Alan pay my graduate student back in 2017 together we publish this paper. What this is a very classical heterogeneous nucleation problem. And you that positive lithium onto copper they have different crystal structure different lattice spacing, and they don't necessarily like each other. And lithium would require a new creation barrier right there. If you increase over potential as drawing right here of this for new creation, you know classical new creation theory, what will tell you increased over potential you drive a harder you have a smaller nuclear by the number density is increased a lot. The scaling load is showing on the top left in a critical new radius, nuclear radius, right will go down with over potential the relationship is one over over potential. And the number density will go out as a cubic function of over potential. So we actually measure that top right is the nuclear size. Versus the current density to hire the current density to hire the over potential when you plot that and it fits into the classical heterogeneous nucleation very nicely. That picture right here is showing you this different current density from left to right top to bottom. You clearly see the, you know, the radius of the lithium nuclear I becomes a lot smaller with the current density, the number density is increased tremendously with current density, fitting nicely into the classical radius new creation. So with this observation let's go one step further. Let's look at this a wattage curve lithium deposition onto copper showing on on the lab right here this this new creation barrier highlighted by this wattage deep by 40 milliwatts and the very low current density 10 micro ampere centimeter also, but you can find other substrate such as gold. Here is a lithium plating highlighted right here is no wattage deep. So this clear difference between different materials right there. Kaya was my postdoc now working in Apple and and discover this is a well let's try to understand this. This is very important for us to control lithium matter deposition down the road. So, so we come to explanation, looking at the face diagram between copper and lithium on the top gold and lithium at the bottom right here. And you look at the most right hand side, lithium and copper the copper has no solubility and negligible right solubility and lithium, but gold would have some solubility you look at this domain. This, what does this mean, it means less lithium coming in doing deposition before lithium matter deposition have a lithium coming in, you know, arrow is gold. And it has the ability to dissolve gold away because gold has solubility and lithium. So lithium has this ability to make gold more and more look like lithium. And after that lithium deposition happened happening this gradual change really remove the nucleation barrier, but lithium on copper is different story. So that very different and lattice constant and increase the structure so lithium deposition has nucleation barriers. And I think initial discovery actually we screen a number difference of substrate from gold, silver, zinc magnesium where you see this this we ship the watches curve. You look at this you say well for all the substrate, they don't have new creation barriers right there. You go look at the face diagram between this materials this matter and lithium it fits the explanation really fits. And then you look at copper nickel and carbon it always has some new creation barriers in this what is deep right there. So having this now allow us for the first time and we design this experiment to have spatial control where lithium deposition will take place. This is an example of a gold pattern onto copper. You see lithium go on to go but not copper because lithium has a new creation barrier on copper they don't really like copper you can you know do this do this experiment going up to reasonable current density. Of course, if you drive it too hard with high current, then the overall potential is too much you're going to have deposition everywhere. So with that, we design our first host concept of hollow carbon with gold seed in there. Hollow carbon has new creation barriers. This is a kind of amorphous slightly graphitic carbon having some conductivity lithium can still penetrate through these carbon. And we suspect when you do deposition within certain current density, we can only like to nucleate inside this hollow carbon that is hollow carbon will be protecting lithium from the outside electrolyte. Certainly the assumption will be this hollow carbon cannot have too big of a pause otherwise electrolyte will wet will go in and then that destroy the purpose. So we made this structure and with this gold nanoparticles inside this hollow carbon sphere. And I want to show you an aceto video to indicate right when you do lithium deposition inside this hollow carbon with gold as seed. Once lithium coming in, you see this gold is dissolved away. So very powerful and lithium sit inside. Why should strip lithium out. You are going to see the gold nanoparticles coming back. So really proving our idea of the gold seed is having ability to absorb lithium and lithium can dissolve it away. That's the way to promote nucleation. So this lithium with gold seed and this carbon the bottom this hollow carbon that we see after deposition certain capacity you don't see this lithium filament going out. But the top hollow carbon case without gold seeds you see the lithium matter deposition outside very clear difference if you you look at the battery column deficiency. The top is very low the bottom is a lot better. This is back in 2016. Now, then we asked the question, you know, what type of carbon will be the best. We have this is amorphous carbon with slightly graphitic domain right there. We found out those are mechanically is not strong enough. That's why understanding the materials properly need is super important. If we could grow this highly graphitic carbon right using, you know, the some of these catalysts for example, having a nickel coating right there. This would, you know, grow these graphitic carbon and later dissolve away the inside is nickel, and you can, you know, produce this crafting cage, a gold seed in there is the nucleation location. Indeed, improve the performance quite a bit and the electric becomes a lot more resilient mechanically, if you put mechanical force on it. So we saw the improvement like that. But what we eventually really want to do is to embed lithium into, you know, a host materials, not by electrochemical deposition. We want to do it by certain way you can pre may already pre store lithium in there so this lithium metal holds composite can be used to pair with other type of electrode they don't have lithium to start with. So Ding Chang and Yah Yuan, you know, really pioneered this area and my group. And I have shown this before, you know, if you melt lithium right we want to develop a modern lithium process. And onto different materials, lithium metal does not really like to wet or many materials until one day and Ding Chang Yah Yuan try reduce the graphene oxide it melts it goes in. So our, you know, lithium has this wetting property, and we're similar to water hydrophobic hydrophilic now we call it lethal philic or lethal phobic. And this with this graphene oxide right is the way is very small amount is 8% only you can embed lithium now between the graphene oxide layer this function as the host to stabilize lithium. When you strip lithium away, it's whole volume will not collapse to completely empty. Indeed, if you don't strip all the lithium completed away, you can hold the volume. All these holes right there, fix is a thickness. This helps you to increase the stability of lithium metal. So I will jump through this. And over the years, you know, not only my group, certainly the whole research field already develop a number of holes that look very promising to the, you know, promote the state stable cycling of lithium metal. Now we ask one more question with all these holes right there, you know, fundamentally thinking about lithium transport. And what else we need to take care of. Let me use graphene oxide as an example about the torch all steel lithium metal foil with your metal holes. The torch all steel has been used right for your cathode transition metal oxide cathode or graphite and no touch all these. It is important. So to to leave a metal maybe it's even more important. The reason is, you know, lithium metal is through deposition stripping mechanism, imagine the top row is a graphene oxide horizontally aligned. You can deposit a lithium metal in there. Why this torture, you know, horizontal aligned graphene will have really high touch honesty. And, and this will only promote lithium metal deposition on top part of the graphene oxide, not to the bottom. The bottom row is if you have vertically aligned graphene, you know this ionic path is more uniform compared to the height tortures path or horizontally aligned. You can deposit a lot more uniform lithium metal to the vertically aligned that will promote stability. So we have this hypothesis and how I started to work on this and develop a method using the cooling the ice crystal formation like this is graphene oxide. Dispers and water and then you cool down from the bottom pre produce the ice equestrized and then it's going to propagate from bottom to the top. Turn out to be you can align graphene oxide vertically this way. So this produce a low touch honesty. You can also cool it down from the side and align graphene horizontally. Now you have a high touch honesty. We could also do the cooling, you know, surrounding cooling. Now you have a random distribution. So this is a three different torch honesty comparison. And it's very clear and the vertical align one has low touch honesty give you a much stable cycling. And look at the figure a right here is vertical aligned graphene. That's the most stable and could make efficiency that that position is more uniform, followed by the random graphene and then followed by horizontally aligned graphene. It's very clear effect right here. And this touch honesty difference. It's, you know, horizontal a lot horizontally align is about 4.5 render is about one point a and vertically align is about one point two one two or three also. And you have to focus on the bottom. SEM this is called session real is after you know 40 cycle 40 cycle for the horizontal and randomly align you see this is highly those cut look like more silly information, but vertically aligned one giving you even after 100 cycle this is very, very solid, very dense field with the metal deposition go inside very nicely. So a very clear difference. And the, in the behavior. So this is all the material design we know interfaces very important right to understand the interface, we really need a new tools to know that and help us to guide us the design of the interface. And with you, if you look at in a fragile bathroom material such as lithium metal on the team. I mean, this is a long temperature. And, and this lithium matter is going crazy, you know, you just start to zoom in. And you're going to destroy this lithium metal then drive right away. There's no way you can do, you know, image it. And understand, you know, these materials understand this interface. So, a few years ago, roughly three years ago, my two of my students, you know, we discussed about this problem. And say, quite a genetic electron microscopy microscopy might be able to solve this problem. Cryo EM has been developed by biology community structural biologists, right in 2017 winning the Nobel Prize. And we actually is a one, one of the earliest group to adopt this technique, develop the lithium matter deposition and doing the pond free without exposed to the air and the liquid nitrogen environment and transfer into the T and stabilize this lithium metal. We know lithium metal react with nitrogen. However, lithium metal does not react with liquid nitrogen, we can stabilize these lithium metal start to do imaging on this materials. We were able to see atomic scale resolution or metallic lithium. I think this is for the first time in the past, even though some people show some TM image. And we doubted that's really the metal, because lithium metal get destroyed so fast by the e beam. Now in the liquid nitrogen environment. We were able to resolve this. What's really important is this offer very powerful tool has come to the interface. We know the SCI. And also in the past. And there's a proposal by a door on our bar and by play about, you know, what could be the SCI structure. So I'm taking this question. We could now imaging this lithium metal and resolve this SCI layer. This white line right here is the interface between lithium metal on the top, the bottom layer is really the SCI about 20 nanometer thick. And this is using ECDC electrolyte. We could observe the inorganic particles such as lithium oxide and lithium carbonate dispersed into this amorphous matrix sounds like a mosaic in model. That's what the play proposed, you know, a few decades ago. And in the same time, when we change the SCI, sorry, the electrolyte, you know, component slightly by adding in fluorinated carbonate FEC into the electrolyte. This is the whole SCI structure change completely. Now what you are seeing is this amorphous layer in the bottom and the top layer is beautiful inorganic coating, right? FEC is oftentimes now you add FEC, clumbic efficiencies improve. So, and this SCI change, the structure change completely explain this. And with this inorganic coatings uniform on the top, these really stabilize the lithium metal a lot more clumbic efficiencies gets higher. So this really resonate with Doron's layer by layer model SCI. This is some modification is actually not inorganic coating on the top. Indeed, the cryoem is so powerful. Over the past three years also we have been using the cryoem for the battery research for other fragile materials such as the metal organic framework. And we work on electro catalysis, you know, looking at the catalyst, perovskite solar cells, and it opened up, you know, really exciting opportunity for I think material science community. I mean, to renew the inches of using cryoem, cryoem was invented in material science community. You know, the impact wasn't big enough by structural biologists who took this tour develop it further for solving protein crystal structure. Now, with all this development, it can have great impact now coming back to the material science community. You know, in the last three years, these 10 papers we published, this many things are the first time we can resolve so I will not have time to discuss this. And also, please see the work by Shirley Mung and UCSD and Lena's work in Cornell and they have been doing also nice work as well in the cryoem. Now I want to go deeper into the interface issue. We keep talking about SCI, for example for lithium metal, you know, for silicon for graphite. Now, can we really understand how the transfer happening at the interface, before you form the SCI. For example, right here we have a lithium metal, right, this you have a lithium that's solvated by these schematic drawing the, you know, salvation shell. And to deposit lithium, lithium I need to pick up an electron, electron needs to transfer to lithium I and deposit become atoms. And if without the SCI, how does it look. The reason we are interested in this question without SCI is this type of study allow us to really understanding the lithium salvation shell, its impact to this transfer process free of SCI. If I see I come and another scenario will start to to to appear. So in order to understand that we develop a tools a few years back, using this a microelectro showing on the top left right here. This is about radius about 12 micron. Also, why the exposure very small areas. And then when you are doing the CV measurements scan your voltage measure your color. You can scan very fast. This is microelectro cross section is so small, and the lithium flux going in. So this is kind of point sources, right point source, and it's less limited by the mass diffusion. So through this what is called the CV scan. You can see, you know, Jay that's current density versus the water, you can scan very fast, you know, up to 2030 watt per second. So we are you're finishing this measurement one scan less than a second. So now you look at the zoom in and look at this a range of the small wattage window that it's independent of scan rate. This is the regime right now is electron transfer control regime you're not controlled by the mass transfer right here the mass transfer. This microelectro really give the advantage to study the chest transfer, a free of, you know, mass transfer limit. So this is not to be quite exciting for us to go into a detail to analyze this curve within the electron transfer limit. So and David Boy, my graduate student right there did a great job and here analyzing for example this dotted line is our dotted dot is our experimental curve to fit this curve to not to be the classical but a warmer model will not be able to fit it. It can only fit within very narrow of plus minus 15 milliwatt of wattage range. It's going over that doesn't fit. And we need to go to Marcus model. These organization energy lambda equals to about upon three watt. If it's very nice nicely Marcus model really, you know, take into account this a voltage dependence. Now, when you have charge transfer, the solution share the solution structure, you know, so we need to reorganize itself to accommodate this change so consider that this energy costs right there. Another theory allow us to fit this very nicely. And, and with this in mind is more than in mind we now can look into different electrolyte, you know, for example, and this plot I'm highlighting, and the same ecdc solvent with different this an iron arsenic fluoride phosphorus fluoride and the PCORI they have different binding strength with lithium and its solution structure is different. And this actually fits very nicely. And this arsenic fluoride one the current density is very high and your day not exchange current current is much higher than a PCORI and phosphorus hexafluoride. And the reason is, this has an iron is much bigger advance weaker with lithium and your solution. Energy right there is smaller. And also, a screen different solvent electrolyte we also establish the correlation solution energy affecting the transfer as well as viscosity. So we start to be able to establish very nice correlation of electron transfer across this interface, because this scan is so fast less less than a second this is really in a situation free of SCI SCI doesn't grow that fast. So I want to emphasize the interface, you know, we really need to modify eventually this interface make it stable to prevent further chemical reaction between very reactive annual with electrolyte. You know, we can see the hollow carbon sphere in the past, but on nitrogen graphene to be layer materials and electrolyte additive right continue to be important. And, you know, new type of material such as nano diamond that very strong chemically very stable to suppress the lithium metal danger formation. And, you know, we look into lithium fluoride. How do you form very dense lithium fluoride and in collaboration with Brutstein and UCLA, we look at that, you know, developing a gas species, you know, a full, a full process to form lithium fluoride and new type of lithium nitrate formation and high temperature, very dense. So we explore further with the number explore cell healing polymer dynamic polymer that can sell here. So these are all of a new type of idea for the interface. Due to limited time, let me only highlight one example. Certainly, you know, electrolyte continue to be important. Jason Zhang in PNNL, also Kang Xu, and Chen Sun Wang, they have been doing really great job and there. Let me report back to you. And recently we discovered new molecule function as solvent highlighted right here called FDMB. This is solvent based on the rational design. We know DME is actually can form good SEI, right? This is either. But DME is not stable against the high wattage. Below four watt versus lithium is started decompose quite a bit. So we want to extend this chain further now becomes a more hydrocarbon chain. This increase the stability and still maintain the either the two oxygen, you know, to forming good SEI. However, this is not, you know, stable enough the DME. We need to add a problem to problem right here. Turn out to be this molecule didn't exist before. This is first time making this molecule. It has amazing effect and as a solvent. You know, these are three different type of molecule as solvently compare right and these FDMB with fluoride it can stabilize go up to higher wattage during this current versus wattage measurement. And these half cell, the lithium copper half cell and cycling to Columbia efficiency within five cycles go up to quite high, you know, your deposition you see right away, more than 99% right away. Using our method to do the cycling that is you deposit certain capacity and you cycle shallow that and then you strip all the lithium away to judge the Columbia efficiency. That should get up to 99.5%. So in this nature energy we actually have a lot more data to show you to pair with the real cathode is very, you know, stable all for quite amazing performance. So this is a some of the picture the top bill. This is DME. This is DMB. This C is FDMB in the the grain size become a lot bigger is actually a lot more stable. We also discover the SCI using the choir EM imaging for for the FDMB. This SCI is all amorphous, but very uniform. It's really thin only six nanometers. Previously easy DC electrolyte easily is a 20 nanometer SCI. The DME is 10 nanometer is less uniform. It has some, you know, other, you know, in homogeneity right there, turn out to be FDMB is very uniform. So the further study indicate through the simulation if you look at FDMB after the color is brownish color in other electrolyte will be transparent. Turn out to be is lithium coordination can not only coordinate with oxygen and the FDMB, but also the foreign, you know, through these molecular dynamics simulation we try to now figuring out already to some degree of how coordination in this electrolyte. So it's actually quite exciting and culture YouTube to read about is our new work just coming out a couple weeks ago, and this discovery or new type of solvent to form electrolyte. And having, I mean, one of the I was one of the best performance discovered so far, you know, for lithium matter in the whole better community. So I think my time is up. I won't, you know, going to detail summary of that so the material design and the nano scale as well as the interface so important. And, you know, doing research to address the grand challenges we identify continue to be the theme of the research group. So let me end my talk by thanking my whole research group, and also the funding support, particularly from DOE. Thank you for your attention. I will be happy to answer any questions you have and also going to the panel discussion this is yet later. Thank you very much for the very comprehensive talk. And I took the liberty to organize the question since we're a little bit tight on time. So the first group of question has to do with the effect of nucleation on the subsequent morphology. So in many of the experiments that's done in the laboratory there are complete cycling so every single cycle you have to strip all the way down and nuclear again. But in the real life operation often as you mentioned the cycling is very shallow so you don't undergo nucleation every time. We reported several ways to control the nucleation morphology, which has an effect on the subsequent morphology. So the question has to do with how well does the growth morphology remember the nucleation morphology in these realistic cycling conditions. Yeah, I mean that's that's a question we asked ourselves all over the time and so after first new creation, right, I mean it's very important the first new creation will determine the second in the first cycle, how it will deposit imagine you have a lot of nuclear versus fuel and nuclear you already, you know, grow this lithium down morphology will be very different. And then with more and more cycle going on. You know, this is the question we have not studied carefully yet with more and more cycle right there. This memory can last for how long. However, I bet the damage if there's any damages to start with you already have it so you want to do it even it's the first field cycle memory. It's already very important you can lose a lot of lithium this way. And that will be the answer but later I mean in the future I think it's memory effect we could look into more. Thank you and related to this point on there's been a very significant body of work, many from you on the effect of deposition of lithium. How about the stripping on the morphology. What do we know about the effect of stripping. Well, this is an excellent one so I rarely talk about stripping even though we have a few study and one of our PNAS paper. My previous post of Feifei did a study on the stripping is actually very important as well because stripping is the case you can form voids stripping can cause significant uniformity right there and your SCI won't be as you know uniform and then the stripping coming in cause the void formation which we tend to see the pits formation easily. So stripping I will say equally important as as that position so we shouldn't overlook. Thank you. The next question has to do with the effect of SCI on the electron transfer kinetic so you show through the ultra microelectro experiment it's possible to learn the native electron transfer on lithium metal before the SCI forms. The question has to do with after you let the SCI form. How does the exchange current density how does the reorganization energy. Have you seen a big dependence and evolution of those quantities upon SCI formation. The SCI formation certainly number one will change the exchange current density, reduce it by a lot for example by about 100 times we could see that change. And if we don't SCI you know easily we see 10 milli amp per centimeter square type of exchange current density but once as your formation it dropped a lot. And then once it drops a lot then electron transfer of course as it is also blocking the electrons right and electron transfer becomes it's not a limited step anymore. So that would be you know how the lithium be solvated go through the SCI I think that's the limiting a step so. Great. Thank you. And then just the final question. So you discuss very briefly the importance of safety. And you have shown a lot of success in controlling the morphology in controlling SCI on lithium metal. So suppose one day all these problems are fully solved. So how about safety. What are some of your directions on managing the safety of a lithium metal battery in the charge state. Yes, even with all those problems. Well, we will be very nice because all those problems you mentioned, even with those problems saw we have not solved the safety problem yet so safety right here. Not only means the shorting of due to lithium down drive but also in the charging state. And this carries so much energy. If you're going to the nail penetration test like you heat up the batteries are carrying so much energy lithium metal has high surface area and the safety issue still exists. And some of the strategy will be we absolutely need a very good fire retardants put it into the batteries but fire retardant going in change the risk also make it risk so very hard to put a whole lot in. How do you kind of encapsulate the fire retardants and put it into the batteries. I mean, as one example, I will say to help the safety. The chemistry materials level all the way up to, I was a system level of design, I think can be in place to adjust the safety. Thank you there are many more questions that we cannot get to today so I encourage the audience to reach out directly to Professor Tui or Professor Chang for further discussion. So now we have about 10 or 15 minutes left I'd like to invite yet to come back. There you are yet. Welcome back. And now I have had the have the pleasure of having a conversation with the both of you. Which is a daunting task I would say. So I thought I would go back to how I introduced the both of you being innovators both in the academic setting but also in the tech transfer setting. And one thing I have noticed, and this is well recognized in the startup world is that there's always a competition between the degree of disruption one hopes to achieve. And the time and the resources that it takes and often we see folks trying to make incremental disruption but you can do it very quickly. So there are people trying to make the huge disruptions that takes 10 or 20 years lithium on battery being one of those that required extensive investments of resources. I was wondering if we can have a discussion around this to think about how this is being done today, maybe also ideas on how this could be changed in the future or how to affect the investment in the tech transfer environment, so that we can make the necessary big breakthrough that's needed. So yeah maybe we can start with you and share some thoughts on this topic. I was thinking that, as you were speaking there that, you know, sometimes the investors take care of that for you. If it's incremental they won't invest. But it's incremental but impactful incremental is not necessarily a bad word. If it's impactful, you know, industry is often. If it's a drop in, right, if it's a drop in, there are many avenues to have impact. And you can have big impact even if it's incremental. But, you know, for startups, if it's incremental it's not interesting. If it's a 20 year project is not interesting. But what's a really great sign is that there are so many, you know, climate and clean tech mission oriented investors out there now who, number one, their mission is in the area of this kind of hard technology that we're all interested in. And the second is that they have a, you know, kind of a 10 year view on things. And, you know, if you, of course they always want to do it faster. But if you think in terms of 10 years and not three, you can do a lot. Thank you yet. I could we also have your thoughts on this as well. Absolutely. So, I agree with what yet is saying. I also have the observation and say, well, sometimes, you know, you have a great idea. You think this can change the world can be big impact. And, and your judgment, my judgment oftentimes in the early days is not completely right. How long can you take because I know less about industry. That's why talking to industry super important you calibrate yourself I think yet has a lot of experience already and know how long it will take. And so, then you say, well, when is the right time. I stand for MIT. We have a lot of students go from the stock up. And that's wonderful. You know, it's really brave and going in. And certainly invested were coming and say I agree with yet maybe seven years, maybe 10 years, maybe a little bit shorter, I need to see something. Then you ask the question, when are we going to take up this technology to the real world to, you know, doing commercialization. I always have something one one foot in my mind as if I can raise funding and university continue I see so many problems I need to work on. I just continue doing that for a little bit longer, but it's not always entirely possible to raise funding to do so, then you can consider saying well I want to transfer the industry. A little bit earlier. This is a this is a zone right there is not the, you know, half deadline but this is so you can do a little bit early a little bit later, highly depend on the real situation if you can raise funding in university why not you know, university has a lot of resources to further understand this. If not, I mean going into the industry, you can see a pathway within seven years you could hit somewhere then go for it. Maybe will if I could come back in with just a couple of more comments on that you know I think that it's a big, it's a big surprise when you see a really top notch industry lab and realize that it can do 100 times the experiments that you can do in your university lab. And so, if the way I think about it there are problems where scaling the rate at which you can do the work really is the key to success. Then you really time to get it out there and just let that happen. Right. But if you have a stage where it's an idea and, you know, more money doesn't help solve the problem. And it's really you need a better idea for part of what you're trying to develop. That's when you know that's what it should still that's something you should focus on your university lab. And more money isn't necessarily going to solve problem it may give you more shots on goal in some sense. But when it comes down to you know if it's being if doing experiments to 10 times the pace will just get you to the answer that much faster that is probably time to start let that happen. Thank you yet I think that's extremely great and simple description of the delineation between academic research and and beyond. So both of you, Ian yet highlighted that developing technologies take a lot of time. And if we do a sort of a simple estimation. So many reports really argue that we need to have solutions starting to roll out on a massive scale. I think yet you mentioned 2050 as one number. So that's exactly 30 years from today. And if we look at the technology development history of today's lithium-ion battery. You know one can argue exactly how long it took but it probably took somewhere from 20 to 30 years to get it to a point where it is near to what it is today. So, do you both have an indication. How much time do we have to innovate at the very early stage. By considering how much time is needed to transfer it to have it come to market or some massive adoption by 2050. How much time do we have and how hard do we need to work. Yeah, do you want to take that first. You know if you look at the history of lithium ion. It required a designer material for every single part of the battery required a designer cathode designer electrolyte designer and designer membranes and all these attitudes. So, you know, I'm the design materials, I think, you know, I'm as big a fan of them as anyone, but I think the skill quickly. We have to start looking at creative ways. The idea is often not the bottleneck. Right. It's the process of making a designer material work. The creative ways of using these also low cost materials. The way that one way sometimes scribe it is that if you're, if you're a chef, it's actually it's much more creativity to make a great dish with really low cost ingredients, then with expensive ingredients. Right. And so you have to I think we have to apply that kind of their creativity to looking at these ultra abundant materials and find new ways to make them work. Okay, I didn't answer how long. That's, that's clearly the approach for taking a form, you know, trying to get some, you know, low cost, very abundant materials into market in scale in very much shorter than the usual time. So, so well, if you look at, you know, about 2030 years, or you mentioned 2050, the CO2 level, we need to go below certain level with two degrees C's within control. You know, hit to the scale. I will say any, probably any new ideas, new chemistry, we better implement this within the, within 10 years to make an impact, you know, right there. And then there's another 20, 30 years, 10, 20 years you need to scale up to the level. So probably about 10 years or so we need to get, get this all, all in. Yeah, we really don't have much time. And then how do we speed up that working harder is one solution. But also, I see, I mean, I watched the solar industry and the growing and people think solar was mature when you're back to 10 years ago. No, not not yet. And there's a lot of collective innovation together the whole supply chain, all need to come together somehow this communication speed needs to go faster to all do it and all the way from mining, you know, to to the top. I think the whole supply chain just need to communicate so frequently to push forward. Thank you. So, before we finish here I just want to make a quick comment is truly a pleasure to have both of you present today. Site by side. Between the two of you you have catalyzed the investments of billions of dollars into battery industry, some of your own startups. I think I can join the rest of the community by thank you for your many contributions now. Normally I would close but I just received a question from from Sting winter hand so I think I cannot ignore that question. I will read Stan your question verbatim. So this is question is for yet. So Stan asked, do you see a resurgence of all of them for grid storage and even EVs as demand for nickel and cobalt goes up. So Stan gets to have the last question here and yet you can have to the last word for today. Thank you. Absolutely. I think so I think that low cost lithium ion is for the grid. And even for you is going to be defined by nine resource constrained materials and I, you know, LP has so many things going for it as a grid storage chemistry. I find it hard to believe that we will not revert to that more, especially as the cobalt nickel availability and prices become under pressure. Thank you. So we will continue next week at our usual time. Our next two speakers are Professor Martin winter at the University of Munster, who's leading many of the large efforts in the European Union. So he will be giving us the European perspective on energy storage. And we'll also be joined by Professor Shirley mung from the University of California, San Diego, who will also discuss many of the materials challenges and next generation battery technologies. And with that, I'd like to thank you on the behalf of each way of myself for joining us symposium. I hope to see you all next week.