Battery Revolution Clubhouse Recording - Batteries, Data & Solid-State

0:00Transcript

0:01Amazing. I see people coming in and we just got to, I think, have a constant, you know, in and out from people. And of course, people can also listen again to recording either on this app or on Spotify, Apple Podcasts and all these other platforms. With this, I will open today's session. We're really excited. When you look at the title, you can see there are 48 of these battery focused discussions we have. They're part of the Battery Insiders podcast and here we are on the Battery Revolution Clubhouse channel. And yeah, today we're really in for a treat. We have a really exciting, I think, you know, one of ours in front of us. Today, we're going to talk about all kinds of topics ranging from batteries to data to solid-state. And we've got a very qualified person with us because we got Tim Holme, Dr. Tim Holme from QuantumScape with us, who's the CTO at QuantumScape. And maybe just a few things before we start, maybe just a few ground rules here. You know, we really like to make these sessions interactive.

0:57I see lots of new people. So I'm just going to say a few things. If you're new on this app Clubhouse, you can raise your hand at any time to kind of join us on stage. And if you want to ask a question, I also want to highly encourage this chat function. I will just text in something in there now so you see how it looks like. On the left hand on the app, probably you can see that. We have for free to put any questions, any thoughts also into this chat. And they're also going to be preserved for afterwards. So that's a great resource as well. This is also really going to help us also from a moderation standpoint. If you want to ask a question, and you're in the audience, please put a question there. So we also know we want to bring you up on stage to ask it in person. Yeah, maybe one, two quick things on the stage, just because it might get crowded later on, because if many people want to join, you have this mute button. And you can use your mute button also to kind of show if you want to ask a question, or if you want to do some applause, because unfortunately, we don't see each other. So we have to go through this kind of little hex on the app. So maybe show you really quickly how you can raise your hand.

1:57Essentially by flickering your microphone very slowly. And you can also do applause by flickering your microphone quickly. Yeah, this is just a small things here on the app. And again, in the audience, and hopefully you have seen that. And also if you listen on Spotify and Clubhouse, you know, really, you're in for a treat. Because today, as I mentioned, we got Tim home with us, but we also not just have him. There's also a few other people here. One particularly here, I want to say my calls for today, Mariam, or let her chime in as well in a second. We're really happy to have her, one of our regulars here in these sessions, we're happy to have a course together. And there's also Raymond and Joachim you can already see. I'm sure there can be many more, want to join us in the discussion later on as well. Mariam, you want to say a few quick words from your side? Yes, I'm super happy to be here because this is such an important topic. Data and new technologies, and especially for batteries, with the scale, the costs, the timelines that are required for us to meet some of our goals. So it's a great topic, very happy to be here. And thanks, Simon.

3:06Thank you so much, Mariam. Brilliant. Yes. And also, some of you are going to know, I'm Babi, I got also one of our co-hosts, I'm sure we'll be on the next session again. Great, same as Catherine, actually, we have quite a few co-hosts now. Okay, with this, I think I don't want to take any more time away from kicking it off. Tim, really, really delighted to have you. I think you have a really interesting also past. And also, we saw maybe more research in Stanford, but also, of course, now, over a decade or so on QuantumScape. So yeah, I think you have a few words also in the beginning, maybe you can kick us off with off the topic of solid-state. Most people don't know too much about it. But with this, Tim, we'd so delighted to have you. Over to you. Great. Yeah, I'm really, really happy to be here. Thanks for inviting me. Sure. So, and solid-state has been a topic of battery research for decades now, but there still aren't really commercial solid-state batteries available. The EV that you drive, if you drive one, is not a solid-state battery. You know, your phone and laptops don't run on solid-state batteries. So that speaks to big challenges in developing practical solid-state batteries. So that's the problem that QuantumScape set out to solve. We really started off saying, let's build the world's best battery.

4:26It's still consistent with the laws of physics. And solid-state, it became apparent to us pretty quickly that solid-state was going to be an important avenue because our view was that the best battery would use a lithium metal anode. And that a lithium metal anode had been attempted many times before. And the classic problem with the lithium metal anode is dendrites. And our thesis was that the best way to approach that was with a solid-state separator that could sort of contain the lithium inside the lithium metal anode and make it behave nicely. So a lot of people are attracted to solid-state also because it might be a safer battery. I think that, you know, I want to be clear that there's, in my view, no such thing as an intrinsically safe battery, because if you're going to store enough energy in a dense enough fashion to drive a car, then there's going to be a lot of energy there. And if you're, if there are 100 million cars sold in the world every year, basically every corner case gets explored. So there'll be crashes and all kinds of things that allow the battery to release energy.

5:31However, I think a solid-state battery could be safer because you're replacing with a solid-state battery, you're replacing a lot of the liquid electrolytes with solid materials that are already in a lower energy or oxidized state. So there's not as much energy to release in a fire if a battery gets in a crash. So the liquid electrolyte in a lithium-ion battery has about as much chemical energy stored in the bonds of the liquid electrolyte as there is energy that the that's usable in the battery from lithium. So about 50% of the energy in the battery is, is just released in a safety event. And the solid-state battery hopes to reduce that. A few other advantages to solid-state as well. One is potentially better performance across a wider temperature range. So in a lithium-ion battery, if you go down to low temperatures, the liquid electrolyte will at some point freeze. And so near the freezing point, you get the viscosity going up and therefore conductivity going down. So low temperatures, there are issues and then at higher temperatures as well, some side reactions tend to kick in. And in solid-state battery, the electrolyte is already frozen. So it's already solid across the full operating temperature range. It'll just operate with the likely the same arneous behavior across the wider temperature range. So like I said, QuantumScape set out to try and make a battery like this commercial. We've spent about 10 years, more than 10 years doing that so far. The first five years roughly was just in pretty deep material science, exploring various materials and what would be possible to meet the requirements of the solid-state battery. And then since then, the last seven years or so, we've been really focused on scalability of the process to be able to make the required materials at higher volumes and in high enough quality that they actually work to satisfy the requirements. And then in the last couple years, we've started to release some of that data, which shows that first of all, in single layer cells, that just have one cathode, one anode, one separator that you can show pretty compelling performance.

7:53And then since then, in the last year, that can be scaled up to four and 10 layer cells without sacrificing the basic performance. And that's what we're still working on is scaling up. So this year, we're going to be making bigger cells and then going forward, moving into a pilot line and hoping to start actually running some cars off of these cells. So I hope that answers the question. Amazing. Thank you so much, Tom. I think this was a really, really helpful overview, especially for some people might not know so much about solid-state batteries. And I think, I mean, it's really interesting, right, kind of, I mean, I think there's a lot of buzz about this kind of different technologies. Some people wondering, you know, what kind of come next. And I think also, maybe to kind of also link it to another topic you want to talk about today, which is like the data topic, where you mentioned you release data, etc. You know, really trying to understand also kind of where can data help to kind of accelerate these developments. And I think, you know, of course, battery testing is a big topic, and maybe also you are here in Raymond, and some additional thoughts on this one as well later on is the battery death hackathon where you were so kind with QuantumScape also to provide some data for that. But I'm just kind of curious if you maybe can also maybe provide a bit of a ground of kind of what are the different aspects where data can be relevant and help, you know, in the battery from the research to production, testing, etc.

9:17Yeah, well, so, several years ago, Mark Andreessen published an article in the Wall Street Journal called Software is Eating the World. And the basic thesis there was software is becoming so powerful, so increasingly important, that it was going to affect all industries in a big way and really upset industries in a big way. And after reading that article, I got really fascinated by the idea of how could software and, and therefore data play a bigger role in battery development, which is often done in a highly empirical way. So we collect a whole lot of data in our lab. Actually, when we started the company, one of the first things we did is starting to build out a database to capture all of our experiments. We didn't want a bunch of researchers with their individual Excel files, not able to share data and not able to compare what, what they were doing with what had happened before or what other researchers were doing. So we built data into the infrastructure of the company from, from, I would say, just about day one. So our goal of our software team is to capture everything that happens in lab. Every time a material is transformed or goes through a process, we should be capturing the metadata about what happened to it. And then we also have, we also capture a lot of rich data of the components along their paths in the production process. So we're generating well over 100 gigabytes a day of data.

10:55Some of that is images of the components. Some of that is electrical data of the batteries as they're cycling and on test. And some of that is the metadata that I've discussed about what is happening to the components when and what tool, what recipes are being run. And we also capture time series data off of the the tools that are processing. So for example, what is what are the temperatures of a various tool over time so that we can then mine that and look for trends. So the goal of gathering all this data is to enable some more advanced analytics. So we do also have some data scientists and machine learning folks on staff who are embedded in each of the key teams and try and make use of all this rich data to help the scientists and researchers optimize their processes that they run. So a lot of researchers in our lab will run an A versus B comparison or a DOE that compares a number of results. And those comparisons, you know, should all be done statistically. But on top of that, we will mine further for data and say, you know, let's look across different shifts of operators and see if there are differences. Let's look across different batches of incoming material or different tools. These are all examples of things that we do on a very routine basis and have found really important trends that without this data set, we wouldn't have found. So just to take one example, years ago, we had an event where everything that we made across the lab started performing differently. And after this went on for a little while, we had to basically pull the whole team into a troubleshooting session and go through and say, what is the cause of this? And it was really only from having all the data that we were able to eventually root cause the issue. We had looked across all the batches of incoming material and found all the batches were responsible for producing good parts and bad parts. We looked across all the tools in our lab and said all the tools that have been responsible for producing good and bad parts. So that eventually allowed us to eliminate everything that we captured data on and narrow it down to a transfer in between two different processes.

13:17And we found that one of our suppliers that sold us the parts that we used for conveyance of parts between materials had an excursion in their process. And we went and talked to them and they said, oh, no, no, nothing changed. Nothing went wrong. So we went back and looked over our data again and did more experiments and became even more convinced. So we went back to the suppliers. They said, no, no, no, nothing went wrong. And we said, you know, really look again. We're very convinced it has to be. And eventually they found that there was something that they didn't document, but a technician made an unapproved, undocumented change to their process and it ended up affecting us. So that's just one example that happened years ago. But I have, you know, there are many more examples where having such a rich data set has really helped us to accelerate our production process to be able to find issues and also find opportunities to be able to say, hey, there's there's maybe a trend here that we weren't picking up on.

14:15So, yeah, that the data is kind of the the stuff that you work with to generate these insights. So data is enabling. It's not key by itself, but it does enable you to do very rich analysis. And Tim, for for a lot of the technologies that you're building, you mentioned that you spent quite a few years in fundamental materials research. Did you have to build your own standards for benchmarks on quality assurance of your materials on benchmarks of performance? Yeah, that's that's a good question. There aren't really standardized battery tests out there. Partly because there are so many different use cases of batteries, whether it's for consumer electronics or grid or automotive or, you know, there's also exotic batteries that have to work in outer space at super low temperatures or batteries that have to work on the head of a drill bit that does, you know, deep earth drilling and gets very hot. The medical devices, there's so many different use cases that we did basically have to develop our own set of standards for, okay, for the markets we're targeting, what do we think are the figures of merit and what are the best platforms to to use to measure those figures of merit? Platform development is a whole topic on its own that probably sounds super boring to a lot of people, but it's really important. How are you, what are the most reliable, consistent and high fidelity platforms that you can use to measure your battery materials and setting up those platforms, characterizing them? What is their noise? What's the so what's the signal to noise that you're going to be able to achieve from a given platform? That that's a boring but important piece of infrastructure that that a battery developer is going to need as well as the tests and the sort of standards or what metrics am I trying to achieve?

16:14And do you think that there should be an ecosystem where some of these standards and benchmarks are being shared amongst, say, players in the industry, amongst policymakers, organizations, etc? Well, yeah, I think that some work can be done along those lines. You know, when we talk to our potential customers, they would come up with a set of tests or a set of metrics to hit. But even there, there's not not a perfect consensus. You know, there's, there's a correlation across a lot of the metrics, but certainly it's not identical across the industry. There's, I think in computers or in semiconductors, there often are more standardized metrics, like just for example, flops, the floating point operations in a computer, that was often the way that you would measure performance of different computers. And then there were benchmarks that folks would develop. There's a little bit of that in batteries, like, for example, in the United States, there's the USABC, or battery testing, that has a set of standards they would like to hit, that they would like battery developers to hit.

17:26And then there's a little bit of that in the United States. And then there's a lot of that in the United States. There's one framework that we are using ourselves and encouraging others to adopt. We call it sort of the end test. So one of the things that makes battery development extremely challenging is it has a battery has to meet many metrics simultaneously. And if it, if it can't meet one of the metrics simultaneously from a very long list of metrics, then it's very possibly not a battery that's going to work in that application. So what we think of the end test is a battery that does demonstrate many things simultaneously. For example, cycle life at realistic use rate rates of power, charge and discharge, across realistic temperatures at realistic pressures, depth of discharge, etc. And doing all of that simultaneously in one test so there's the end test that we're promoting is doing cycling at 1C1C, meaning one hour charge and discharge. Because if you look at a supercharger today, it'll roughly charge the battery in an hour.

18:39And at room temperature with modest pressure, full depth of discharge, I think that's a good start. Of course, that's not a sufficient condition. That's, but we would argue it's a necessary condition, and a good starting place that you could use to compare different technologies. That's very interesting. So what you're saying is, is there an accelerated stress test that batteries have to go through essentially in order to emulate some of the real life conditions that it needs to be able to meet requirements for? Yes. And for example, that end test is, I think, one accelerated life test. You know, it cycles across full depth of discharge, 100% depth of discharge, which is more stressful than virtually every battery is used in a real world application. Right? If you think about your phone or your car, if you drive an electric car, you really don't tend to cycle it all the way down to zero and all the way up to 100 every time you use it. There are lots of shallow cycles. And so in this way, our end test is an accelerated life test. Also in the rates. So using it at 1C1C every cycle is quite aggressive. That would be equivalent to charging your car up at a supercharger every time you charge it and then discharging it on the Autobahn or I don't know, on a racetrack. Suppose you have a 500 kilometer range vehicle that you're discharging in an hour. That's very aggressive driving. It's not 500 kilometers an hour because things are not quite linear that way. But still, that would be extremely aggressive driving every cycle. So I think this test that that we use internally, and that we've shared a lot of our battery results on is quite an aggressive test.

20:32But again, like I said, that is perhaps a necessary test, but not a sufficient test. There are other accelerated life tests that battery developers should should use. So for example, this test that I've been describing tests cycle life more than calendar life. But batteries also need to have have long, what's called calendar life is how long the battery lasts if it's just sitting on a shelf. And for consumer electronics devices, that might be several years, say three, four or five years. But for a car, that should be 10, 12, or even more years. So there also should be accelerated tests that probe calendar life as well. And Tim, every time that you scale up to a new process, so for instance, when you're starting with your small cell or fundamental materials research, you're running some of these tests, you are maybe benchmarking and you're down selecting some of your materials, and then you'll go up to the next stage. You're scaling it up even to a larger cell, and then scaling that up to a pack.

21:40The behavior of the materials, I would imagine that would change as you're changing sizes, changing conditions, and so on. Do you have to repeat this process every single time? There is a lot of testing and optimization that has to happen on each level of scale. But one thing that's pretty interesting is the fundamental materials properties can be studied on a pretty small scale. And this is why a lot of battery research, if you look in academic papers, is done on relatively small cells, smaller than a commercial cell. There are things that do change are processing uniformity, edge conditions, things related to the battery cell packaging itself. Some of these get harder or more challenging as you scale up, some actually get easier. So just to take one example, things that scale with a perimeter to area scaling or surface area to volume scaling will change as you go up in size, often getting more favorable. Because if you were to think of an infinite plane, then that's going to behave ideally like your fundamental material. So the larger the material gets, the smaller the edge conditions are conditions that happen around the edge relative to what's happening in the center.

23:04Also another thing that will change as you scale up is more automation. And automation can remove a lot of the variability in a process. So there are some things that improve as you get to higher throughputs, higher scales. Also the processing tools can often be more uniform when they're processing big batches. And when you've got a tool that's sized to make a lot of material. So for example, you know, compare hypothetically, if you are an academic researcher doing a cathode casting, you might be using a drawdown table that's a tabletop machine that produces, you know, half a meter of cathode at a time. And there's, there's always head and tail effects when you're doing a casting, meaning the start and the end of that casting are typically different from the center. And so if that's happening over half a meter, you don't have much area that is uniform. But if you're running, you know, a large scale commercial coder, that's coding a kilometer or or many kilometers in a pass, and several factors wider also, that can be a much more uniform set of material, it's a lot more homogeneous. So like I said, there are some things that get better as you process at higher scales, scales, larger sizes, some things get more challenging.

24:24So you do have to re-qualify your processes and do a lot of testing at each level of scale through the scale up process. Fantastic. Thank you so much, Tim. I think lovely wisdom already there. And maybe just one thing before we kind of opening it also up to the floor and I already see some great questions by Jul and I think Milo's also with his hand on others. Maybe Raymond, are there any additional questions you would have at this time? You can unmute yourself. Yeah, yeah. No, this has been very interesting. Thank you. I'm curious, coming out from a more general data side, how do you, you mentioned quite a bit earlier how QuantumScape, you know, from the very start, you started basically database, you know, throwing everything into a database, all the research data. I'm wondering how you, how you keep, especially for like complicated root cause analysis, how you, how, how do you feel like you guys manage the context problem? I mean, specifically with the scale stuff, you know, certain suppliers are giving you information about something they did at, you know, for a test run. And then two years later, you have a problem with the material, but it was actually being manufactured at, you know, a pilot scale. So much, much larger production run. How do you, I feel like that must be a large problem to manage, like understanding this data, you know, these two data points that look very similar actually came from very different contexts.

25:49Um, is that a problem you feel like you've resolved or is still, you know, that's still a very manual, you know, you have to, you have to talk to the person who was on the floor at the time type of situation. Um, so there, there are some things for which you have to be on the floor to recognize all of the context. Um, but we do try and capture as much of that as possible in a format that is that then gets centralized and archived. So it does serve as the single, single point of truth. And that is our database. Um, so, you know, you're mentioning an example of raw materials. We log all the batches of raw materials as they come in, uh, any additional data that the supplier supplies, like a certificate of analysis will get logged with that batch. And then we record what batch is going into what downstream parts. So all of those links exist in the database and can be mined later. So then we have our power users or our data scientists going back. And if there is a root cause analysis problem, they will go and just jump into the data and start looking for all these links, um, sort of up, up and down throughout the process flow and figuring out using all that data to do the root cause analysis. Um, it's that approach doesn't give you the right answer every time. Sometimes it's really, you know, the, the process owner or even the technician who's running the equipment who notices the problem at some point. Um, but it certainly has solved the data approach has solved many of our root cause analysis problems in the past.

27:29Um, so that's one of the really important reasons I think that, that we log all of the metadata and keep it mindable and searchable and queryable in that database. Uh, the other reason, as I was saying earlier is to look for opportunities. So there are a lot of times where we notice, Hey, maybe here's a trend that, that I'm picking up a more subtle pattern in the data that nobody researched and maybe I should take a deeper look at that and run a focused experiment to look into that more. Yeah, that's cool. That actually, that really reminds me of this kind of the, the setup and philosophy that I've encountered at, uh, pharmaceutical companies with supply chain and manufacturing quality assurance. Yeah. And I think that big manufacturers would have what's called a, an MES manufacturing execution system that will track a lot of materials throughout the process. I'm not sure how granular, um, many of the big incumbents are in tracking the data. They might track lot to lot, but not sort of individual cells. I know that at least one of the major suppliers is, um, driving down from the lot level all the way down to individual cells. Um, but that's, that's something that we've built in really from day one, track all the individual components and as they go into the cell and ensure all that traceability at the most granular level. Very cool.

29:01Fantastic. I thank you. I think this is a really excellent point also with this traceability, which also reminds me, I think, you know, often we talk about traceability for supply chains, to ensure, you know, that the materials are coming from, you know, from, you know, low, low emitting sources, also ethical sources, et cetera. But I think, as you say, really in your production, even, you know, it's so crucial to kind of make sure the quality is there. And I found the story you brought earlier on the anecdote or like experience, um, you know, to find that supply, that something changed there, which really would influence your production sector really, really valuable. And I think, yeah, there's already like a quick takeaway from me here. You know, we really have to record everything we can, um, to, to this kind of fair prediction, but also between these. Um, otherwise also I think there's only some questions from the audience. So before we go there, as you know, we left an interactive discussion, maybe just one thing, just in the spirit of this podcast, the thing I want to highlight, nothing he has said as investment advice. I think that's just always really important to, to notice and please any questions here, keep it on the tech on the, you know, technology and policy, et cetera. Let's not talk about any individual stock prices, et cetera, what could be influenced. So I think that's just important to really open and nice discussion.

30:11And with this, I think, you know, one of the first question I saw in the chat, well, from Juhu. I'm also going to invite you now to the, to the stage in case you can speak. Yes, there you are. Juhu, welcome. Would you like to ask a question? Thank you. Thank you, Simon. And thank you, Tim, for the insights. I was wondering about the price of the cells when you actually get on the commercial level, that you have a commercial cell. Do you think that it will be like on competitive compared to like state of the art EV cells on like dollars per kilowatt basis? Because yeah, I think because at the beginning, I'm uploading your pick if you manage to do it from the beginning. What is your, like your own estimate? Yeah, I think for EVs in particular, cost is the most important metric. So the dollars or euros per kilowatt hour is probably the most important metric for an electric vehicle battery. So we, I think our market would be severely limited if, if our battery weren't cost competitive.

31:15So that's, that was one of the key metrics that we designed in and, or, or I should say designed for, from the beginning of the company, we said, how can you make a cheap battery? And one of our, one part of our thesis was make a battery with very high energy density, because you're going to pay a certain amount of dollars per kilogram of materials. But then what you're selling is sold for dollars per kilowatt hour. So if you're able to get very high kilowatt hours per kilogram, that should drive down the cost of the battery. So that is one of our approaches to get high energy density, then basically for each kilowatt hour, you have to buy and process fewer kilograms of material. Also, I think by, in, in our approach, we use a lithium metal anode, but it's lithium free as manufactured. So there's, there's no anode really, as the battery is manufactured, the anode creates itself in situ on the first time that the battery is charged. And that has the potential for, for great cost savings.

32:17If our separator is cost competitive with conventional separators, then what we've done basically is eliminate the anode manufacturing process. So you eliminate those, that set of materials, and the whole processing line. If you look in a battery factory, the main parts are coating of both electrodes, and then cell assembly, and then the formation and aging and test. So coating is, is like I said, one of those three big parts. In the coating areas, you have anode and cathode coders. If you're able to eliminate the anode coders or turn those anode coders into cathode coders, you get higher throughput out of your factory. That should mean lower cost. So lower cost is something that we have really designed for from the very beginning of the materials and technologies, techniques that we use. Jorge, thank you so much for that. And I think it was really fascinating though. Maybe also one thing kind of connected is kind of the overall lifetime, you know, costs. If you can last so much longer than I know in the pre chat, we had a bit about the, you know, the one million mile, the four million mile batteries, etc. out there and also kind of understanding with maybe, you know, what do you think could be possible to be, you know, and you know, if it's like a perfect solid-state battery kind of from a lifetime perspective, could you build like extreme robust system for that?

33:42Yeah, yeah. Great question. So when it comes to life, obviously that's going to be important on its own, but then it could really open up the grid market as well. So what's important for the grid or stationary energy storage market is cost over the full lifetime of the battery. And so that's basically initial cost and cycle life. If you get a super long cycle life, the battery can operate for 20 plus years, then it's probably going to be cost competitive on the grid market. So low cost and long cycle life would open up whole new markets that, you know, as big or bigger than the whole EV market. So batteries, I think our type of solid-state battery has the potential to have longer lifetime than lithium-ion for one key reason here, which is the SEI formation growth on the anode. So this is getting down a little bit more technical. But one of the main reasons that your phone or your laptop lasts less long today than it did when you bought it is the fade. The fade of the cell is largely due to what's happening on the anode side. So on the graphite particles, they on the during formation, what's forming is the SEI solid electrolyte interface on the graphite surface. And then as the battery is cycled and charged and discharged, this SEI can tend to thicken and then start to crack and have to reform which consumes more lithium, more electrolyte out of the battery. And that's one of them, one of the larger drivers of capacity fade. And in a system like ours, we don't have that SEI formation. So by reducing one of the failure modes or one of the fade mechanisms, at least, we hope to have longer lifetime as well. Do you think I find the difference between the full solid-state battery and the the electrolyte battery quite interesting? And that do you think there was a difference between having like less things to optimize? I mean, so when you're manufacturing, you know, a liquid electrolyte lithium-ion battery, you have to optimize, you know, the burning phase, you have to optimize electrolyte formulation with that SEI formation in mind.

35:56Is that, do you think that's an advantage behind the fundamental choice of solid-state, even though initially it was, you know, probably quite risky? Can we make this at scale? But once you're getting there, it seems like there's, there's just a lot less knobs you need to turn. Well, to talk about formation in particular, yeah, I think we're certainly hoping that the formation process can be much simpler and much reduced. And again, that's one of the biggest cost areas. So imagine you're operating a factory, once you have made the battery, then it needs to go into a formation process where it can spend weeks or a month before you can even ship the battery. You've tied up all of your costs, all of your inventory into this finished product, but you can't ship it and then therefore even start the revenue recognition process until it spends quite a long time in the formation process. And so we are hoping that because we don't have the SEI to form, that our formation process can be simpler, shorter and therefore cheaper.

37:01Fantastic. Thank you. Also, I want to put another question on me, the formation, I know it's a big one. I think also you, you had another question. Hi, Simon. Hi, Tim. Thanks for all the opportunity to ask a question and congrats for what you achieved so far. I'd like to hear about if you could comment on the major challenges on the technical and non-technical side that you're facing. And if, and I like also to hear about if the supply chain crunch is affecting, or having a short term or long term impact on your business. Thanks. Okay, good question. So on the supply chain, the entire battery industry is booming right now. I think there's exponential growth going on, probably literally in the battery industry. So all of the traditional battery suppliers and battery materials are under a lot of stress. And that on top of COVID, on top of the war in Russia and Ukraine is really stretching the supply chains quite thin. So that does, that does pose some challenges in the near term. And then probably as we're scaling up, it will be challenging to work through. We do have a planning and supply chain group that does form pretty deep partnerships with our, with all of our suppliers. And we try and get on top of that.

38:38Most big equipment in the battery space has lead times that not measured in weeks, not measured in months, but measured in quarters or year plus timeframes. So if the supply chain issues add a few months to that, that does pose a delay. But as a percentage of, you know, a year, if you add a month to a project, it's actually not delayed by that much. So the other part of your question was, what are the technical and non-technical challenges? So right now that the biggest technical challenges are in the scale up, we have to install and qualify a much higher volume equipment. And to qualify the equipment means running your process on it, making sure that you can run and get the same quality at the output that you did with other tools in a smaller scale process. So that requires a lot of process engineering, a lot of careful, again, statistical and data analysis, but also a lot of experienced folks walking the line who are able to, to, you know, see things with their eyes and touch them with their hands and see is, is this different?

39:46Is that, does it look like it's supposed to look? And the non-technical challenges, you know, probably every business, the biggest non-technical challenge is, is working with people. So we've got a lot of great people. I think it's really a world-class team. And the best part about working at QuantumScape is the team we have and the folks that I get to work with and learn from every day. It's really a fantastic set of folks, but also presents some challenges because you'll have a lot of people with a lot of different experience. That's good. The diversity of opinion is what you want in a team. If everybody were the same and thought the same way, you, you wouldn't get nearly as far and as fast as you can with a diverse set of opinions and backgrounds. But then working with that, especially under some pressure when people are feeling they have to deliver on a pretty aggressive set of timelines and milestones, then, you know, if you have different opinions and, and pressure, that can also create some challenges to work through. And so that's probably the biggest area of technical challenge, of non-technical challenge right now. Thanks very much.

40:55I like to try a follow-up question. Like if you had a lamp and you would have wishes, you could, what would be the wishes? Boy, that's a, that's a big one. I'm a genius, you know. You know, as, as much as I pour my heart and soul into batteries, I'm not sure my wishes would be related to batteries. It might be, you know, world peace or... I'm glad I made you laugh. So something like free and ubiquitous, uh, solar power fusion and batteries. Um, if you get, if, if your genie is not quite so powerful, then, and I had to have, go for a more local wish. Um, um, um... It's a hard one. Yeah. You know, really it's, I'm, I'm just struggling to, to grasp the scale at which I should be wishing, right? Should I wish for the company to succeed or do I need to get more granular than that? Should I wish for the industry to succeed and, and scale up and replace fossil fuels?

42:12There's, there's a lot of things to wish for. Um, sorry, I'm not sure I'm giving you the answer you're looking for. Um, like broad question, broad answer. Well done. Thanks. Fantastic. Thank you. Yeah. I liked it. We were talking about wishes. There's definitely something out there. Yeah. Thank you. I think now also Milo's has raised his hand. Would you like to question? Otherwise I'm also going to get, um, um, for our gone here, maybe also one quick thing in the meantime, because we've now been half come in and maybe also gives you a few seconds to image. You want to get a glass of water or something. I know we have been asking lots of questions. I can be the interval. Um, maybe if we take a rest, uh, maybe just here as a reset of the room, we now about 45 minutes into one and a half hours. We really delight today as part of the battery insiders, a battery revolution podcast, to have Tim home with us, Dr. Tim home, CTO of QuantumScape. And I think it has been a really fascinating discussion. We're going to continue this for about another 45 minutes.

43:14Just really quick thing. If you listen to this on Spotify, our podcast, all these other platforms, and be free to also join us on life. And then you can also ask your own question. And personally do this session once a month, always on the first of the month. And yeah, we have a really exciting lineup and also had a really interesting, um, already lineup of conversations, starters and speakers over the last 48 sessions. And if you want to listen to them, you can go on, um, battery insiders on Spotify, Apple podcast, et cetera, you can find the recordings on there. Maybe this just as a bit of a crease quickly said, and yeah, please also keep the questions coming. I think there's about 18 or so already in the chat. So you'll try to address them there. Or this maybe Milo's would you like to, to, to ask your question? Yes, Tim. Tim, I have the question about commercialization of your technology. When do you realistically expect your technology to be commercialized?

44:11And especially when, uh, can you describe a little bit that we can understand and we can expect your technology to be, uh, consumer electronics, or maybe in the cars? Some timeline, realistic timeline. Yeah. Um, so the targets that we are shooting for are to to move in and really implement our pilot line this year, to be using the pilot line next year to make, um, cars at, you know, at least one, hopefully more than one cars operating with our batteries. And then to start to, to be in parallel ramping up higher volume production so that in, uh, 2025 timeframe, you could start to see commercial cars running with our technology in it. So those are our timeframes. I think it's quite aggressive. Um, but I think we have to be aggressive as a company and as an industry, um, to, to be able to win business and to be able to displace fossil fuels as quickly as we can. Um, you also asked about consumer electronics.

45:18So we haven't, uh, announced any partners in that area yet. It's, it's an interesting market because everybody interacts with their consumer electronics devices, uh, dozens of times, hundreds of times a day, maybe. So it's something that's really front and center for, for most people in terms of climate change. It's obviously a much, much smaller, uh, or if, if any impact on climate change. So that's one reason that we're really focused on cars as well as the much, much bigger market that automotive represents. However, um, we have announced some results that I think make our batteries applicable to consumer electronics, which is, um, pressure. So we haven't talked much about this today, but most batteries, even lithium-ion batteries operate under a little bit of pressure to keep the layers in, in good contact with each other. Most solid-state batteries require elevated pressure. And if you think about the form factor of most consumer electronics devices, they have to be super thin. Your, you know, your phone or your tablet or a laptop has to be very thin. So there's just not much space for any fixture to apply pressure to that battery. But we did announce in December, um, that our batteries could operate under no applied pressure. So that does open up the consumer electronics market. I'm not sure about the timeframe there. It's possible because consumer devices are smaller that they could, you know, come in in a shorter timeframe, but as a company, we really haven't made that our sole area of focus. So I don't have a timeline for you that I could share.

46:52Okay. And what's the most, uh, important underlying technology. I mean, if you can simply describe that when I understand, so it's that separator are using gels or, uh, what can you describe, uh, simply the technology? Yeah, sure. Um, we almost zoomed past, past that. Um, so our, our technology is, we use a cathode that's, um, fairly similar to advanced lithium-ion batteries today. The cathode has particles of a nickel rich NMC, a very high nickel NMC. It has an organic catholite. So the catholite is the electrolyte inside the cathode. That's what conducts lithium in the long range between the particles and to all the way to the separator. We have a solid-state ceramic separator. And that is, as you were intimating, that's the real key to make the lithium metal anode work to, uh, resist the tendency of lithium deformed dendrites. And also to improve Coulombic efficiency, we have a solid-state separator that allows us to use lithium metal anode with no lithium as manufactured. Um, and then the, the key technology here is, um, I would say two parts. The first is the separator. So what materials and what process do you use to make that separator that makes it have the properties that are required, especially for, um, being in stable contact with lithium over 10 plus years when lithium is the most reducing element there is. And then the second key area of technology is once you have that separator, how do you build a cell around it that works? How can you, um, assemble the cell, deal with the lithium expansion and contraction, for example, and, and, um, how do you deal with edge conditions over the entire lifetime of the battery? So lithium, as it plates out in a, in a lithium metal anode will expand, lithium is packed more densely in the cathode than it packs as lithium metal. So any lithium metal battery would have some expansion, particularly if you have a zero lithium as manufactured cell, then the expansion is larger in terms of percentage of the overall cell thickness.

49:18So our cell will expand and contract by about 15% throughout the cycle. So we have to come up with a cell form factor that is capable of accommodating that expansion and contraction. And by the way, this is another reason that operating at low pressure is important because if you have to maintain a high and uniform pressure on a material that's expanding and contracting, that becomes pretty difficult to do from an engineering perspective in a way, if, if you're also trying to maintain low cost and high energy density, if energy density, weren't an issue, then you could maybe design some big bulky fixture that that is able to do it. But if you need to apply high and uniform pressure by using a small volume in mass to do so, that's hard. So I think that's, that's one reason that we've really focused on driving down the pressure that our cells operate at. So anyway, the two key areas of technology are our separator material and production process and then the cell and cell assembly process. So keep your eyes out for our, some announcement in the future. Sorry, my daughter is here.

50:28Thank you so much, Tim. I appreciate it. It's very early morning for Tim. So we're really, really thankful for him to spend his time with us today. Hopefully this was helpful, Mylos. Maybe that was helpful. It was helpful. Yes, absolutely. Fantastic. Thank you, Mylos. And I think then we also have Basim here. Would you, would you like to ask a question? I have some comments. Yes, I have like three questions. I will take like five minutes and then I'll be more to do it. Can I do that? Absolutely, no problem. And we still have time. So yeah, maybe just as a quick reminder, we got about 35 minutes or so. Also from experience, the last half an hour tends to be the most busy might also, you know, yeah, you know, I think the most questions come then also more time zones are waking up. So yeah, if you have your questions, make sure to bring them up now. Also invitation for all the women as well listening, please also join us, raise your hands, really trying to, you know, have a, you know, diverse participation in these podcasts and questions. So please, please step up and step in. We would love to have you on here as well on the panel. Brilliant. Is there any other question kind of from this room? Otherwise, I can also maybe go to Mark. Would you ask, like to ask a question Mark in the chat?

51:59Okay. There you are. Hey Mark, good to see you. Good morning, Dr. Holm. I have a question for you. I was reading that QuantumScape has some alternative cathodes and Florida cathodes in its IP collection. I'm just curious how this could be used in other applications, whether it's cars, planes, or what form, because I know you have more than just the lithium cathode in your IP collection. Yeah, there's a couple of alternative cathodes that I can talk about. So first of all, we are researching alternative catholites. And even if you keep the same active material, there are catholites that might enable higher energy density and or higher power density if they have higher, greater conductivity or greater stability properties. But I think you were probably asking more about the cathode active material. So when the company started, we were looking also at, at very high potential energy cathodes like, um, uh, conversion, and conversion cathodes. So to give an example, um, iron trifluoride, FEF3 is one of the more commonly discussed conversion cathodes. This was researched more, call it 15, 20 years ago than it is today. It's fallen out of favor a little bit today, uh, because there are a lot of challenges in making these cathodes work. The advantage is, um, it's got super, very high theoretical energy density as well as, um, low cost because it just uses iron as the, as the metal, which is, uh, much lower cost and higher availability than nickel or cobalt. And that's, that's used in today's lithium-ion. So FEF3 is, is a technology we looked at a little bit, um, quite a long time ago. I think we developed some new IP, found out some, some really interesting ways to solve some of the challenges with conversion, conversion cathodes, but ultimately decided that biting off everything, you know, reinventing a cathode in and separator in an anode all at once would, was a bit too much to tackle all at once. So decided to, to go with more conventional cathodes. There are still, um, lots of challenges with conversion chemistry cathodes.

54:31Um, there are others who are looking at formation cathodes like a lithium air or lithium sulfur. I think those are still pretty far out. There's quite a bit of technology development required before they, they can be commercialized. So when we were really looking at how do we make the best battery, it was pretty clear to us that the single biggest change you could make is on the anode side, going from graphite to lithium metal anode would have quite, quite a large improvement. And then after that, the cathode becomes the limitation again. So if, if we were to, you know, go 10 years down the road and, uh, QuantumScape is successful and is scaling up, then we probably have our research team turn to look more at the cathode side. Um, and, and and clearly there's a lot of activity in academia looking at the, at different alternative cathodes today. So hopefully that technology matures so that it, it could be commercialized at some point in the future.

55:32Thank you, Dr. Holm. Fantastic. Thank you so much. And also maybe now one thing I would like to ask everybody on the call, we have about half an hour left. If at the bottom, you can go on the little share icon, there's a little, there's a little arrow up in the box going up. And if you want to go share on clubhouse or you can of course also share on social media, but if you just want, if you enjoy, I mean, I may probably should say if you enjoy this conversation, please do that. If not, then don't. But if you enjoy this conversation and I find it absolutely fascinating, um, please, please share this room here on clubhouse and we're going to create a bit of a more buzz because also no, no more people in North America are just waking up and to make sure this goes straight to their newsfeed and they're all going to be, you know, have a chance to kind of join us for another half an hour. So if everyone who likes this conversation can do that, it would be absolutely amazing. Paseem, are you back inside?

56:23Would you like to ask a question now? Otherwise I go to someone else in between? No, I can ask now, uh, just a bit of intro. So I have been following Consk for some time now myself, and I come from automotive, different types, buses, trucks, cars, and now I'm in electric planes, which is a totally different problem for us. So, um, maybe different questions. The first question related to, are you looking into some different variants of cells where it can, for example, for, as you mentioned about customer, uh, consumer electronics, when we talk about calendar life, but for example, for, uh, electric planes, we look more into energy density, safety, and, uh, then we talk about cycle life because yeah, it's, it's a different story and also about cost. Are you looking at different variants here to try to satisfy different markets or you mainly focus now on electric vehicles and then the cost and different aspects are more important for you? Yeah. So for electric aviation, I think specific energy, what hours per kg should be the most important metric, like you were saying, and then safety, I think you had a really good ranking, um, specific energy and safety after that life and cost, but, um, probably power. So specific energy, safety, power, then life and cost. Um, we could potentially, um, work, work with electric aviation markets. It is a much smaller market today and for the near future. So again, that's, that's, we haven't really focused there, but there are, I know a lot of companies working at making these flying cars or flying taxis services. And because again, we eliminate the anode, we do have the potential for higher specific energy. That is what hours per kg than lithium-ion. Um, and I think the higher you go, you can unlock longer and longer range flight. So lithium-ion today is of course, good for drones, but to get longer range, you would need to get higher specific energy batteries.

58:35And then maybe the second point related to the form factor of the cells, maybe I missed it. I didn't join from the beginning of beginning, but, uh, when we look, for example, yes, you are saving on the energy density on a cell, but for example, when you are building cylindrical versus big cells, then you are saving on energy density on a cell level or maybe cylindrical, but then in the end on a, on a bigger scale, on a battery pack, when I want to build a big battery for a car or a plane, then I lose on energy density on the higher, on a, um, on a pack level due to a lot of things like form, form factor. And also that I need to put a lot of, uh, fire retardants, et cetera, et cetera, et cetera. How do you look about the form factor of your cell? Or that is something you didn't look at yet, but which form factors are you going to go? Yeah, good point. So it's my opinion that a lithium metal cell won't use a cylindrical form factor because of the volume expansion issue that I mentioned earlier. So every time lithium plates and strips and the cell changes state of charge, the cell expands and contracts ours as we expect to expand and contract about 15% across the operating range. And so if you imagine a cylinder where the inside is trying to expand by that much, it, uh, hard to imagine that working unless you create internal pores for the lithium to expand into, but then those internal pores have their own space, their own volume, which reduces your volumetric energy density. So I think that lithium metal anodes will be, uh, uh, stacked rather than wound, then some sort of a prismatic form factor. If you're thinking about electric aviation, then you probably want a pouch rather than a can just to save a little bit of weight.

1:00:20And then another way to save weight is making larger cells so that the packaging becomes a smaller fraction of the overall volume, uh, or mass of the cell. The problem I see now with suppliers, for example, sorry to interrupt you because it's a very good point you made now, but the funny point is that most of the suppliers working with that type of technology, they are looking at cylindrical, unfortunately. And it's, yeah. That's interesting. Yeah. I mean, well, from a volumetric perspective, the cylinders don't pack as densely as bricks would. Um, maybe some of the area between the cylinders you could use for cooling or to run some electronics. Um, but for electric aviation, it's, it's more about mass. So I'm, I'm a little bit surprised you're not hearing a lot about pouch cells. Yeah. They are just trying to use a known, yeah, they are trying to use a known form factor, basically 21, 700 or something. That's their main point. But I agree with you. Yeah. Unfortunately, but thank you so much for the answer. Thank you.

1:01:24Sure. Yeah. And another thing you brought up was the, the overhead at the pack level and how you, um, arrange cells. So we're, we're going to make a prismatic form factor that I, I think stacks more densely at a pack, but also another reason that people are interested in solid-state batteries is they think that maybe if they're, if they're safer or more thermally tolerant, there would be less overhead at the pack level to build in safety, safety features, safety mechanisms, and thermal cooling mechanism. So that's another potential benefit of solid-state technologies. I agree. Thank you so much. Thank you, Basim. And then we have quite a few more additions to the panels. You can see Tim lots of interesting questions, which is fantastic. Esteban, I think you're the next one, would you like to go next? Simon, thanks a lot for organizing this. Thanks, uh, Dr. Tim for, for being here. Um, a question coming back to the materials aspect, uh, is, uh, your ceramic separator technology stable, uh, at high voltage window, for example, could it handle, um, you know, lithium, micken, manganese oxide, high voltage pinel or such, or, or such families of materials?

1:02:38Yeah, that's a good question. So that the separator is only part of the stability question. Um, a lot of it actually is the catholite material. So you need to develop a catholite that's going to be stable. It's going to be in a high surface area contact to voltage. The separator has one interface, uh, one planar interface in contact with high voltage materials, but the catholite has, you know, throughout the volume of the cathode is in contact with a much higher area. So the catholite, I think, is the, is actually the area to focus. Um, now in terms of stability of catholites, I can't talk much about our catholite, but from what is known, if you use carbonates or ionic liquids, they will have a voltage stability window that's dictated by their set of materials and could be studied independently of the separator. If you look at solid-state catholites, the most commonly discussed ones are sulfides that have really a, an even more limited range of voltage stability than the traditional carbonates or liquids do. So if you're trying to enable high voltage cathodes, my advice would be to look pretty carefully at what catholite would you use that's stable.

1:03:59Got it. Thank you. Thank you. Fantastic. Thank you. Maybe also one quick thing, actually mentioned Esteban also was part of the battery DAF competition, great team there. And also, you know, I was playing around with some of your data. Thanks for participating. I hope it was fun. I hope you learned something. We struggled and suffer a bit, but it was very nice. Fantastic. Thank you. Juan, would you like to go next? Are you there, everyone? Otherwise, maybe in the meantime, we can go to Mohammed. Hello, Arne. Thank you for having me. Thank you for sharing the good information on the one of the great work and the great work. My question is actually about the rail earth materials or rail materials that are used. Ceramic type, how did the state, electric lights, for example, the ground cannon, the common material that is used in this type of material, in this type of lights. Yes. The question is, how do you think about the potential shortage of such materials in the future?

1:05:38For example, we have problems with robot. Do you think such materials could have similar problems in the future? Yeah. So the question was about the availability of materials. So one of the objectives we had in the company was to make a battery that would enable mass market cars. That was really our key objective from when we started. And so we knew that if you wanted to go after the mass market of automobiles, you had to talk about extremely high volumes of material. So we constrained ourselves to only using materials where there was enough abundance in the world that it could supply the large markets. So we use materials that are available at high volumes and we have a supply chain team that's in contact with suppliers to ensure that the supply will be there when we need it. However, as you point out, even things like nickel and cobalt, which are sort of global commodities, nickel more than cobalt. But, you know, we could even talk about copper and graphite. The battery materials is scaling up so quickly that people are wondering if there's going to be enough graphite production capacity around enough nickel production capacity around. So the, the EV industry is, is almost incomprehensibly large. So at some level of scale, I'm sure that we will be one of the, the players out there saying, you know, let's, let's ramp up some mines. Let's, let's go faster.

1:07:17But at the same time, we should talk about recycling. The, one of the great advantages to batteries is you, because you're not burning something and just using it one time, you could enable a circular economy. So we have made contact with some players in the recycling space and are pretty excited about the ability to recycle our materials. If you're talking about a ceramic separator, again, those elements, you know, the grams of materials that you're putting into the separator don't go anywhere throughout the life of the cell. So they could potentially be entirely reclaimed and reused. And so hopefully that becomes a big part of, of the picture when we talk about converting the whole industry, the whole world over to electric vehicles that we really do enable the circular economy. Fantastic. Thank you so much. Thank you. Thank you. And I think the next one, Bo, would you like to be next? Yeah, two questions. One was related to the, what you just mentioned about recyclability.

1:08:25Is there anything specific you're doing now in the construction of the, the packs or the cells to enable easier recyclability? Um, we're not specifically designing for recyclability, but I do think that a solid-state battery could potentially be more recyclable. If it's, if it's easier to take apart, then maybe the recycling process could be cheaper or more effective. Um, you know, what, what's publicly known is that JB Straubel is on our board of directors. He was the founder of one of the founders of Tesla and is now running Redwood materials, which is a battery recycling company. Um, so he's got quite a, quite a bit of insight into both battery, into batteries and electric vehicles and recycling. And, um, he has said that he's pretty excited about the potential of our technology to be recycled. But like I said, if, if the battery comes apart more easily into its components, if you're able to easily separate it out the anode cathode and separator, and then direct them into more specific, um, purification technologies that could make for a lot easier recycling than what's often done now, which is to literally shred the battery, basically shred the whole thing, and then try and sort it out using a pyrometallurgical or hydrometallurgical process.

1:09:48Excellent. Thank you. And another question was related to, uh, how the solid-state, uh, battery technology compared to others like zinc air for stationary storage in cost, um, and feasibility. Mm-hmm. Yeah. Um, okay. So I think we should, we should as an industry be thinking large. Gigafactories are large, but still only satisfy a fraction of a percent of the overall vehicle market. So to satisfy the whole vehicle market, you have to be thinking about terra factories. Uh, and then when you get to grid plus automotive market, it's, um, tens, hundreds of terawatt hours. So probably all battery technologies will be necessary when you talk about that sort of scale. So there's probably going to be a space for zinc, air, and iron, everything else that you could think of. That said, I do think that one anecdote has been instructive for me, which is the anecdote of Aquion. So I think, you know, I know Jay Whitaker and I think he's a great researcher and I was rooting for Aquion, but Aquion was a company that was trying to make very low cost aqueous based batteries for grid applications and, you know, got well past proof of concepts to where they were really making large batteries. But eventually the price of lithium-ion fell so quickly that it, um, forced them out of business. So from looking at that, from looking at the solar industry and how a lot of advanced technologies were undercut in price by sort of older, but, but cheap and scalable technologies. Uh, one of the things that I take away from those anecdotes is in battery specifically, um, learning curves can bring down costs by quite a bit. And, um, the watt hours per kilogram, that you can get out of a battery is a really important factor. If you think about an equation where dollars per kilowatt hour is very important for batteries, dollars per kilowatt hour is, uh, you could decompose that into dollars per kilogram times kilograms per kilowatt hour. So you want to have low dollars per kilogram of material, but you also want to have high energy density because high energy density will multiply, um, to, to get you what is the cost. So you take dollars per kilogram divided by energy density, you get dollars per kilowatt hour. So high energy density is a, just as important a factor as the cost of the raw materials is in the equation of what is the cost of the overall battery. So this is a long way of saying that I think lithium-ion, um, or solid-state lithium metal could be even lower costs than, um, and a dollars per kilowatt hour basis than something that has a lower dollars per kilogram like zinc air. So I guess I gave you both sides of the argument there. Maybe the battery market is so large that there's room for all kinds of technologies. But on the other hand, uh, it could just be that lithium-ion or solids, especially solid-state lithium metal battery would just come down the learning curve and be so cheap that it could apply for most of the large markets. Great. Thank you.

1:13:16That's awesome. Learning curves can bring down costs. Uh, Sharath, do you want to go next? Yes. Do you have a question? Yes. I have a couple of questions. Um, first one is regarding the LFP prototypes. I know there was a blog and back in December that quantum skip supports LFP, um, lithium phosphate, uh, but I didn't see much, uh, uh, in the news about the prototypes related to LFP apart from the partnership with Fluence for state of stationary storage, I assume is LFP, but is there any progress on LFP prototype in automotive segment, uh, with Volkswagen along with lithium-ion? Uh, just wanted to know that. And second question is about the catholite. I know, uh, Tim mentioned that, uh, we, uh, QuantumScape has a solid-state to ceramic separator, but some sort of a gel, uh, catholite. Uh, but in future, like near future, next five years, 10 years, are you looking at replacing that also with the, uh, solid electrolyte or is it not worth, uh, pursuing further even for next 10 years? These are the two questions. Thank you.

1:14:23Yeah. Great. Thanks for the questions. So regarding LFP, I think what we did was basically a proof of concept to show that we thought our battery should be more or less agnostic to the cathode that really the key was in the separator that enables a lithium metal anode and that we could use it with different cathodes. So all we did was very simple proof of concept to demonstrate that, um, yeah, in fact, it looks like our batteries could work with LFP. And so then what we intend to do is, you know, more or less offer our customers the choice. Do you want LFP, which could have the advantage of low cost, um, or do you want the, uh, higher cost, but higher range NMC based? And we'll kind of let our customers drive that question. Um, you know, LFP has some natural advantages in that iron is cheap and the supply chains are large and you don't have to worry about cobalt. So lots of advantages to LFP. I think what's, what's pretty interesting though, is combining LFP with lithium metal. So LFP, since it's lower voltage benefits even more from a low voltage, high capacity anode. So you can get, um, you know, theoretically with LFP and lithium metal, you can get to a pretty attractive energy density, especially if there are any further savings on the pack side. So you can bring the energy density of LFP on to close to par with today's lithium-ion nickel rich cathodes, but much lower cost. So, um, what we're hoping is that our technology basically enables you to shift the frontier outwards. So if today's frontier with lithium-ion is nickel and LFP being sort of two points on the cost versus range curve, that we could shift both of those points to lower cost and higher range by using our technology.

1:16:15So that's sort of our take on LFP. Then you also asked about catholites. Yeah, we are researching advanced catholites. There are a lot of advantages to using advanced catholites. Um, we are using an organic catholite today. And the key reason is that it's, it's much easier, offers a shorter path to market. And, you know, a lot of the people who listen to this podcast might care about what catholite is in the car that they drive. But imagine your neighbor, for example, when they go out to buy their next car, do you think that they're going to be asking, well, what's the catholite inside the battery? They're probably going to care about what's its range, what's its charge time, what's its price, you know, the other features in the car. They're probably not going to be asking what's the catholite inside. So we're choosing for a catholite that meets the key customer requirements. And that offers us the most straightforward path to market. At the same time, yes, we do have research programs on alternative catholites.

1:17:17Thank you, Tim. That's wonderful. Juho, you had a question. Yes, it's a continuous role also about this catholite. It's a bit hard to handle. And in your design, basically, you have no solid, solid, constant solid, solid interface because the lithium metal and the solid electrite interface will be constantly reformed, right? And then the second question is, how thin can you make your ceramic separator at the moment? And do you still aim to like shave your microns away from it? Yeah, so I caught the second question. I missed a little bit of the first. So let me answer the second one. We're making our separator in the low tens of microns. Obviously, it's got to be pretty competitive with where the polymer separators are today to still get high watt hours per liter. The first question, you were breaking up at least for me a little bit, so I'm not sure I caught it. Yeah, sorry. So the solid, solid interfaces are known to be really hard to handle, like if you have a constant solid, solid interface. But in your design currently, you basically don't have any solid, solid interfaces between the materials, right? Except the lithium metal and the solid aeroclite, but that is constantly reformed anyways.

1:18:35Yeah, yeah. Good point. So that's a lot of why we opted for the organic catholite and in early products is it does solve some of the problems of maintaining good interfaces, even at low pressure across the battery life. And the solid, solid interfaces, as you were describing, yes, they are a challenge, no doubt. Juho, did you have a follow up question? No, I got my answer. Thank you. Perfect. So Ron, do you have a question? Yes. Can you hear me okay? Yes, thanks. Okay, great. The question was, are you still using the batch process to make your ceramic separators or have you been able to successfully transition to a continuous roll-to-roll process where you're actually getting a strip or a ribbon of the ceramic separator? Well, we have designed our process to use continuous flow technologies. So the two key steps in making our separator are a coating process and a heat treatment process. And we've designed both of those to be continuous processes. So we knew, again, we need to get to very high volumes and low costs. So anything that's a batch process or a high vacuum process was not going to be able to scale to the volumes and costs that we needed. So we designed those out from very early on.

1:20:05That's strip or roll-to-roll now. That's wonderful. Ron, do you have a follow-up question? Yes. Are you able to do that now? Have you successfully transitioned to the continuous roll-to-roll or strip process currently? Well, I have to stick with what's been publicly disclosed and what I've answered is what is publicly disclosed. Thank you. Joachim, you've got a great question about experimental data. Yeah, sure. So, yeah, just coming back to the data topic as well. Very generally speaking, you can use data for experimental design and also for understanding performance or aging as well as manufacturing processes. However, I would say they're all interlinked, but in a way they're also separate. And also in particular during the last year, there were a few startups formed that claim to use AI for experimental design and also for material discovery. And I was wondering how much of this learning from data for making decisions on experimental design is incorporated in the quantum scale process. And what are you doing or what have you been doing the last couple of years in that regard?

1:21:33Yeah. So we do a lot of data mining and data science on the results of our experiments to try and understand what works or find root cause. The field of AI for materials discovery, I would say, is still pretty early stages. And one of the challenges there is the availability of big data sets. So if you're a consumer internet company, you have billions or hundreds of billions or more of records in your database and the ability to learn and do online learning and everything, I think the tools of machine learning and AI today are very well suited for your process. So you can throw big data at a huge neural net and really achieve some pretty amazing things. But if you are in the field of materials, you know, and you're trying to look at, say, 40 data points and try and use that to discover new materials. Well, even by building physics into your model, that's still a pretty difficult challenge. So that's one availability of good data. Consistent data is one of the challenges. And then another challenge is even in a single material system, the processing conditions can really change the materials properties. So, you know, what are, to take a trivial example, if you take bulk gold and measure its melting point, and then you take an nanoparticle of gold, exact same material and measure its melting point, you can get a number that differs by hundreds of degrees. So the same material processed in different ways can have very different properties.

1:23:16And that's something that's very hard to capture, I think, in a robust enough way to do good data science or AI on top of. So I know there are companies that are trying to do this and best of luck to them. I think it's still early days, so maybe they'll be successful in full amount of time. I think one area that's pretty exciting though, is combining AI with experiments. So if you're able to do a lot of experiments in a high throughput fashion, generate a lot of data, and then feed that into, for example, a Bayesian optimization system, that's able to tell you what's the next experiment to run, and then you go get that data and feed it back into the optimizer, and it'll continue to optimize and tell you the next experiments to run. I think that's a pretty interesting area. So I think anybody doing AI in the space will have a greater probability of success if they also build a high throughput experimentation system. And that's really the core of the QuantumScape approach is, how do we create very rapid cycles of learning, generate and capture a lot of data, so then we can learn from all that data and the rapid cycles to drive faster progress. So even from when the company started, our first tools that we got, we ran 24-7, and our lab still today runs 24-7, so that we're able to crank through a lot of experiments and shorten the learning cycle time.

1:24:46Yeah, thank you so much for sharing, Tim. I very much agree with a lot of what you said, and in particular using Bayesian optimization, I think it's a very interesting field, lots of things moving there, but then with this tool you also have the course of dimensionality, and if you want to go to very high dimensions, then this might not be the most effective way of doing it. And then in addition, you have all the challenges with material discovery you just outlined. So yeah, I think I'm quite excited to what is about to happen in this field, and also what startups come up with. And yeah, thanks so much for sharing. It's a great point about the course of dimensionality. Yeah, if you build a Bayesian optimizer and say, okay, what temperature and pressure should I run my reactor at? You can go through that space pretty efficiently, but what you usually find out is that there are totally other dimensions that you didn't think of at first that are really key, maybe the key or at least important factors to consider. So then you need to start building those into your model. And that becomes still, I think, very challenging to do purely computationally. And that's why so many experiments are still driven by the intuition of of the humans in the loop.

1:26:09Yeah, absolutely. No, no, thank you so much for sharing. And in particular, I mean, one point, maybe to add on top of this is, I think it's very interesting to see that lots of research that would have been traditionally carried out at universities is now moving to startups. I mean, on the one hand side, I think it's good that capital is flowing into that research. On the other hand, I feel like sharing is an important issue because academia usually was always very open with their results. And this was very good for the ecosystem that everyone was sharing their results, whereas now lots of research is carried out in startups. And then sometimes it's very difficult to share data, right? And you've been encountering that as well. How do you see open data in the space? And what could be things to do to have more data sharing and more sharing of results in the battery space? Yeah, so actually, one of our collaborators, Venkat Viswanathan and Professor Carnegie Mellon, is one of the people trying to bring together a battery data genome to address exactly this issue.

1:27:24So their proposal is that the community should come together and open up their data sets and contribute all the metadata to. So what's very difficult to capture is, like I was saying earlier, you can capture the temperature and pressure that you run your system at. But those are only a few of the easiest variables to measure. But you know, what are the impurities in your process? And you know, this course of dimensionality that we mentioned earlier means that there are dozens, hundreds, thousands of things that may be important to the experiment and to get everybody to document those all in a systematic way so that it all could come together, for example, in a relational database that with a nice table structure, tabular structure can, I think we're not there yet, even in academia to let alone companies. So open sourcing big data sets was a real key that drove forward, for example, convolutional neural nets in the field of image processing. I think the ImageNet database of a million labeled images was one of the key things that enabled the convolutional neural net to take off. That doesn't exist in batteries. And I think it's challenging to assemble, but that doesn't mean that we shouldn't do it. That just means that the community should try hard to do it because it could potentially unlock a lot of value. Thank you so much Joakim for that question. And we're getting a lot more questions. So it just goes to show how relevant this topic is. But we're coming on time here. So I wanted to close out the room by saying thank you, Tim, for being here. This was a really relevant topic. And Simon, do you have anything, any other comments?

1:29:15Thank you, Miriam. And I think really, really big thanks to Tim, especially appreciating that it's a really early morning for you on a Saturday. And I think it's not a given that someone would spend your time. And I think it's highly appreciated by all of us. You can tell there's a little buzz and I know we probably could talk easily another few hours. But also, I think we have to be mindful of everyone's time. But I just want to say a massive, massive thanks to Tim. Also, you can see he's new on this platform, the little party popper sign there. So if you want to follow him there, I don't know, maybe you're going to stick around in some future sessions. And also anybody else you enjoy, please, please follow each other. Another thing I want to say, you know, you can listen again to this entire recording on our Battery Insiders podcast, will be up probably in a week or so on Spotify, other podcasts and any other platform where you listen to your podcasts. And we also have these sessions regularly. So we have a next one in one month's time, always on the first Saturday of the month.

1:30:11So next one going to be the 7th of May. So please also join us then. And I'm sure it's going to be another exciting discussion. But yeah, for today, I just want to say a massive thanks to you, Tim, and anybody else who joined and brought the questions and thoughts. And anybody for listening. It was a really, really wonderful occasion today. Yeah, the time flew by. Thanks a lot for the questions and the interest. So, you know, you you you you you