 Good morning from Stanford University. My name is Will Chu. I'm the Faculty Co-Director of Storage X Initiative. Along with Professor Itwei, we're delighted to welcome you today to another Storage X seminar. So we have a very interesting topic today. One of the key beliefs at Stanford is that we need to span all the way from atoms to systems when it comes to cross-cutting energy storage. And that is exactly what we're going to talk about today. We're going to be examining the holistic picture for energy storage, not only for transportation, but also for grid and beyond. What do I mean by holistic? When it comes to technology, development, deployment, and optimization, one really has to consider how everything interacts at the smallest level, whether it is the materials, the battery cells. If we're talking about battery technology to the battery packs, and we're talking about grid-level storage, how the battery packs are configured, how they're deployed, and what applications they're being used. And it is an increasingly important topic because a lot of efficiency and economy can be gained from this level of optimization. And we are delighted to have two experts to speak to that. First, we have Professor Jessica Transe. Jessica is a professor at the Institute for Data Systems and Society at the Massachusetts Institute of Technology, and this is exactly her area of expertise focused on system-level analysis to inform policy and technology deployment. And her interest not only includes energy storage, but other forms of energy conversion as well, including hydrogen, low-carbon fuels, nuclear, and many others. And our second speaker will give the industrial perspective. Dr. Nicolo Campaniel from McKinsey & Company will be talking about the electrification of transportation. And he's going to tackle a rather specific system-level problem, which is comparing two chemistries, LFP, lithium-ion phosphate, and NMC, nickel-mechanics, cobalt oxide, for two very different type of battery technology at the materials level and how those differences will propagate to the system level and how that is influencing the skill-up, the technical omics, and how companies are making choices when it comes to selecting the technology. So I'm really delighted to be hosting both of them. So if I can ask Jessica to come to the stage, as I already mentioned, Jessica is a professor at MIT, and Jessica, we're really looking forward to hear your thoughts on system-level analysis and how it can inform policy and decision-making. Thank you so much, Will, and it's great to be here. Okay, so today I'm going to be talking about energy storage for deep decarbonization, and I want to start with some background on climate change to motivate this discussion of energy storage. As many of you know, climate change impacts such as wildfires, as the image shows here, and many others threaten human livelihoods and even human life, and addressing this risk requires completely decoupling greenhouse gas emissions and human economic activity. So the goal really is a return to a carbon cycle on earth that is in balance, where processes emitting carbon dioxide are balanced by processes absorbing carbon dioxide, and the emissions of other greenhouse gases such as methane and nitrous oxide are stemmed. So quantitatively, what is needed? Well, staying below a two degree Celsius global temperature rise in order to limit the damage of climate change requires a rapid reversal of rising greenhouse gas emissions trends. So I'm showing on this plot several of the trends that emissions, CO2 emissions could follow here, where depending on when the decline begins, the decline is either shallower or steeper. But what's needed is a rapid reversal of rising greenhouse gas emissions trends. And what has to happen is that emissions should fall faster than they've risen for over a century. This also likely means that wealthier economies should reach net zero emissions targets by mid century. So that's within three decades. Now, if we're talking about 1.5 degrees Celsius, the reversal would have to be even more dramatic. And either way, time is short and financial resources are never unlimited. So my work is all about informing innovation toward this future. And what we do in my group is to develop data informed models, both data driven and mechanistic models that can allow us to be more deliberate and effective in developing successful technological solutions. So the idea is to recognize always the uncertainty about the future, but to use insights that we can gain from past experiences in order to inform future technology development in order to increase the chances of success. So my research objective, this objective is shown here, it's to understand and shape technological innovation, especially in clean energy systems to accelerate equitable climate solutions. And today I'll be presenting some examples from my work on energy storage. Just very briefly, the methods that we use and develop, many of them we've developed in my group, involve developing data informed models that can be mechanistic or data based to measure technological innovation, understand the drivers and identify targets. And these models can be applied across scales from materials to devices to infrastructure, applied to hardware and also non hardware technology. And the impact of this work, the intended impact is to inform materials, technology and infrastructure decisions by engineers, private investors, policymakers and the public. And I should also add, well, I guess it's shown here, yeah, it's really about informing engineering as well as private investing, policymaking, and sometimes the decisions that consumers make. Okay, so I'm going to briefly touch on the role of energy storage and deep decarbonization. And then we'll get right into discussing some examples from my research on informing and ideally hopefully accelerating progress on energy storage. I also look forward to the Q&A session that's coming up. So what really is the role of energy storage and deep decarbonization? When we think about all of the different energy storage technologies that are out there, you know, ranging from batteries to pumped hydro storage to compressed our energy storage, and all of the realizations of those types of storage, I often say that the set of energy storage technologies is as diverse and likely more diverse than the set of possible energy conversion technologies. But what is the role of energy storage? What could it be? What role could it play in deep decarbonization? Well, it can certainly help enable the electrification of transportation. So there are two primary electric vehicles, hydrogen fuel cell vehicles and battery electric vehicles. Battery electric vehicles have been growing in popularity and in market adoption, and that has been largely enabled by the improvements we've seen in lithium ion battery technologies in recent decades. So energy storage can play a key role in electrifying and thereby decarbonizing transportation, especially if the electricity mix decarbonizes. And storage can also help there because it is one of the ways in which we can help support the integration of variable renewable energy, including solar and wind energy into the power grid in order to decarbonize electricity. And as you may see here that and I'll show sort of quantitatively how this plays out through some of the examples we'll go through. But as the economy electrifies the importance of energy storage grows. So one of the main ways to decarbonize the entire economy is to electrify as many energy services as possible. So services like transportation, various industrial energy services requiring, you know, processes requiring high temperature. You know, some may be a little bit more difficult to the highest temperature processes are more difficult to electrify but other aspects of industrial energy services can be electrified. Home heating can be electrified and so forth. Now all of this needs to be supported by a reliable decarbonized electricity supply mix and storage can play an important role here. But basically the point I want to make here is that as the economy electrifies and decarbonizes, the importance of energy storage is likely to grow. So it can be a really essential component to reaching the climate change mitigation targets that I started out with showing. The good news is that we've seen very significant progress in lithium ion battery technologies. And as I mentioned a minute ago, that has enabled, you know, the prospects and also the already the adoption of electric vehicles. As my group documented together with Micah Ziegler, my postdoc, we publish a paper in 2021 showing and estimating the cost decline in lithium ion batteries. These costs have fallen by 97% over three decades. That is a very rapid cost decline. And there have been improvements in other key technologies, namely two key technologies, that solar energy and wind energy. Because of these three trends that I'm showing here, because of the falling costs of solar energy, wind energy and lithium ion batteries and also because of increasing density, energy densities and lithium ion batteries, we're in a very different place to where we were just, you know, 10 years ago or even five years ago in talking about decarbonizing the economy. But there's more work to be done, because of course we need to actually see that downward trend in emissions that I started out showing. And in some areas, the technologies, you know, building on these three technologies and other technologies that are available. Some energy services are currently at least in terms of the technological availability. They're currently easier to decarbonize. That's a technical potential. They still have to be decarbonized. There are other energy services that are currently more difficult to decarbonize. So if we look at this pie chart, we can see that, you know, long distance road transportation is flagged. Certain industrial energy services are flagged for steel production, cement production, and then what's called load following electricity. So that's really when, you know, you get to that final push of decarbonizing electricity. As I'll show in a little bit, we need certain technological capabilities that we don't yet have. And storage can play a really critical role there. So, you know, the role for storage is really widespread across many of these energy services. So in decarbonizing home heating, as I mentioned, many industrial energy services, you know, light and medium duty transportation, you know, much of that can, you know, technologically speaking, be electrified. And, you know, certainly electricity. Now, for all of these energy services that are not labeled as currently more difficult to decarbonize, further development of storage will help to accelerate the process of electrification and decarbonization. So that's really important. But also further development of energy storage is really critical to this piece of the pie, the load following electricity, because it could allow for greater renewables, renewable energy adoption, and decarbonizing the decarbonize full decarbonization of electricity. And so let me now get into some research examples on, you know, how we can use models to inform and accelerate progress, you know, given that time is short, and the important role of energy storage and decarbonization, can we be a little bit more deliberate than we would otherwise be by examining what the targets are for storage performance, as well as what drives rapid progress. So in the case of lithium ion batteries, we saw that 97% cost decline in lithium ion batteries, solar energy costs for PV modules have fallen by 99%. That happened over, it's actually a bit more than that now, but that happened over four decades. These are very rapid cost declines, but what is underneath them and can we learn what what the reason for success was in order to inform future efforts. So that is what I'm going to be talking about. I'll give some examples of recent papers that we've published, and you'll get a sense for the methodologies that I've developed to address these these types of questions. Okay, so now let's talk about targets for collective innovation processes to work towards. So here the idea is, you know, not to stem creativity and research or anything like that, but it's to ask for certain technological components, and in this case, energy storage, what sort of performance do we need, and we're going to be focusing now on cost performance, but we could look at other types of performance. What sort of performance might we need in order to support an overall decarbonization of the system? So you can start to think about maybe how you would go about setting up such a problem in a model, and I'll walk through an example of the method that we've developed in this area. Okay, so what I'm going to walk through is this first example, which is on cost targets for grid scale energy storage. So have a look at this plot. What I'm showing here is the our two capacity costs for energy storage. We have the power capacity cost on the y-axis, and the energy capacity cost on the x-axis. So this is the capacity cost for storage. We have different storage technologies shown in this main plot. We have pumped hydro storage, compressed our energy storage and lead acid. There's a lot of data uncertainty and also underlying variability in the cost. So that's why we have these wide ranges for different technologies. Now if I asked you, for example, if I point at one of these center points here in the lead acid cost data points, so I look at this point, and let's say I also look at this data point here, sort of up in the lower left here for compressed air energy storage, and I asked you, which technology costs less? Which one has lower costs? Well, it's a trick question because you can't answer it because there are two dimensions here, and neither one is dominating, neither one of these data points is dominating along both dimensions. So what do I need? Well, I need some sort of model to tell me what are the features of cost in this case, of this technology, that would allow it to be high performing in a particular context. And so the example that I'm showing here, and I'll just mention a walk through that very briefly, is the context of performing price arbitrage with energy storage that is storing solar energy in Texas. And that is where I get what are called these, what we call these ISO cost lines. So these are ISO performance lines. Once I have these lines, then I can see that this data point here, the compressed air energy storage data point, or really any storage technology that had this cost structure, so a lower energy capacity cost, and a higher power capacity cost would be advantageous as compared to that first data point. So this line here is an ISO performance line as we get toward the origin performance increases. So this is what's telling me that, you know, the line here is coming from a model, where we're looking at the context in which storage is being used. And that is really what's telling me which of these data points is advantageous at this point in time. In the case of storage, what the model really has to address is the resource management problem, because really that is the sort of raison d'etre, the reason for energy storage is to manage resources, right? You store energy at some point, you convert it to some useful form of energy at some later point. And so we need to model what is the exact resource management problem that our storage technologies are going to be addressing. And then from there, we can estimate desirable, we can gain insight about desirable technology features and maybe even estimate rough targets for performance, in this case for cost performance. So how do we manage the resource, how do we model rather the resource management problem? Well, at the heart of this, we have to consider fluctuations. We have to consider fluctuations in solar and wind energy. In this case, which is what I'm showing here, we also have to consider fluctuations in energy demands, and also the overall supply mix that we're using to provide electricity, as in this case. So if we look at price arbitrage, then what would happen to the solar and wind output, and that's where those ISO performance lines came from, is that it becomes more peaked, right? So you see output following these sorts of trends. We can also look at another resource management problem, which is one that we looked at and published a paper on in 2019, using that methodology that we had advanced for this paper in 2016. We applied that same methodology to a paper in 2019, but looked at a different resource management problem, which was to provide reliable electricity output to match the output shapes of different power plants, different types of power plants that are operating on the grid today. So we have baseload plants. We have intermediate plants. We have bipeaker plants, and we have peaker plants. And we looked at a handful of different locations and looked at how storage would optimally operate for wind and solar in order to meet the challenge presented by this resource management problem. And so what are the results of that? I'm going to get a little bit deeper into this problem, because this is really one of the key challenges that storage could potentially meet. And that is to be able to help provide that load following electricity in a deeply decarbonized system. But the question is, how well do current costs of energy storage compare to where costs would need to be in order to serve this role while also supporting low-cost electricity? So what we find is when we go through and we optimize the different sizes for different technological components in our system, what we find is that the optimal size of the renewables capacity, here I'm showing photovoltaics versus the storage capacity, that optimal size is going to depend on the costs of the different technological components, as we might expect. So what you can see here is that as we go from a lower energy storage capacity cost and a lower just storage cost overall to a higher one, what we see and focusing on the energy capacity, so that x-axis that I showed a minute ago, the energy capacity costs, what we see is that as the energy capacity costs increase, we see further oversizing of the photovoltaics power capacity, so that's what we see in the lower left-hand plot. And we also see an increase in the storage power capacity, again, as the cost of energy capacity of storage increases. We also see a drop in the storage duration and also the storage energy capacity as the cost of storage energy capacity increases. And so, you know, this is just to give you a sense that in the next plot, what I'm going to be showing are these costs, are the cost targets and the costs that I'm going to be showing in the next plot are based on this cost minimization process. So what do we learn about cost targets? Well, what I'm showing here is for a hypothetical scenario where we would be relying on 100% variable renewable energy, so optimal combinations, cost optimal combinations of solar and wind energy across these four different locations, Arizona, Iowa, Massachusetts, Texas. And again, I'm showing this plot using the same axes that I started out with, so we have storage energy capacity costs on the x-axis and power capacity costs on the y-axis. And what is shown sort of on the z-axis or through the heat map color here is the levelized cost of shaped electricity. So what is the cost of reliably providing electricity? Let's focus on Texas for a minute. With variable renewable energy and energy storage, and how does that cost of electricity depend on the different storage costs? And so that's what we're showing if we focus on that lower left-hand panel there. Now, the other thing I might want to ask is that I might want to ask what are cost competitive targets? If I look at other generation technologies over here on the right, we can see that we've estimated certain cost competitiveness targets. So at what point in this plot here do I reach cost competitiveness as defined by these other technologies? Well, what you can see is that we really want to be in that dark blue region. And so costs really have to fall to 10 to 20 or even in some cases lower dollars per kilowatt hour in terms of the energy capacity cost. We can survive with higher power capacity costs, and I'll show why that is in just a minute. But the idea, and there's no single cost target across locations, and this plot is intended to show different ranges of the cost of electricity as depends on the storage costs. But what we do see is that we have to see a significant drop in costs if we compare to battery technologies today, let's say lithium-ion battery technology costs. One would need to see 80 to 90% drop in costs in order to reach this dark blue region here. And so that gives us a sense of this cost target. So the cost target that we published is again a range. It's around 10 to 20 dollars per kilowatt hour in some places lower and some places higher. And this would be for a mix that's relying fully on variable renewable energy. So now we have our cost targets, but you might be asking, well, Jessica, what about if you consider other energy supply mixes? Maybe we don't want to rely entirely on variable renewable energy. There may be other supply technologies that can be used in order to help us get to deep decarbonization. So what does that do to the storage cost target? Well, this is what we can see here. And what I'm showing here on the X axis is the equivalent availability factor. And what this is essentially a proxy for is the percentage of electricity that's coming from renewables. So if we drop that equivalent availability factor by just 5%, we use just 5% of something else, we can see a significant drop in the cost of electricity. For example, for baseload wind, that cost can drop by 30%, or in some cases even more, just by using 5% of something else. And if we do that, when we look at, for example, a 95% equivalent availability factor, then the storage costs targets rise, they get closer to where battery costs are today. Of course, batteries aren't the only option for storage. One of the reasons why a lot of focus is placed on them, as you all know, is that they can be installed pretty much anywhere, whereas some of those other technologies like pumped hydro storage are limited to certain geographical regions. Now, I should point out that pumped hydro storage does have a favorable cost structure. But of course, we need to consider where it's available and also what impacts it would have along many dimensions other than cost. So there are arguments for really pushing forward with batteries, although in some locations other storage technologies may be available. But in any case, so what is the bottom line here? The bottom line is that we could use, if we diversify our electricity supply mix a little bit, there could still be a very important role for storage. In smoothing out the variability in solar and wind energy, electricity costs could drop. However, the challenge remains that that 5% needs to be met by some other source. And because of a certain nature of the fluctuations in the solar and wind energy resource, which I'll show you in a minute, that can present a challenge. So really one of the key results of this work that we found is the importance of the infrequent but large supply fluctuations. What you can see here is that if we look out over 20 years, there are a few events that we identified in this work that are large shortage events where you have larger than normal shortages extending over a few days, sometimes a week or more around that amount of time, where you have a below average availability of solar energy, that or wind energy and these shortage events happen only a handful of times over 20 years. Now, if you want to move to a system that's relying entirely on renewable energy, then a lot of whatever you install, if you use storage, if you install storage capacity, then that capacity would go unused for large periods of time. You can also rely on some of the other technologies that I mentioned, but you still would be having to meet these resource fluctuations and that can be costly really for any technology to operate in the kind of way that I'm showing here. So anyway, I think this work can help us think about what target should we be aiming for? And it's not a single target, but a target associated with a kind of deeply decarbonized energy system, electricity system in this case, and can sort of allow us to evaluate the efforts and the technologies that we're developing now. In addition to storage, demand side management, low carbon supplemental generation, also transmission expansion, these can all play a role. And we can be thinking about combining these with energy storage to manage the variability in renewable energy. So one of the things that my group is working on now is to perform this analysis. We've developed the capability to do that really anywhere in the world, and really to understand how geography matters and sort of what the options are for deep decarbonization across locations around the globe. So very briefly, I want to take just a few minutes to go over a couple of more examples. And I'm going to give you one more example on research for targets that involves energy storage. Here I'm talking about targets for electric vehicle battery capacities, costs and charging infrastructure. And to identify targets for these technological and infrastructural components, as you can imagine, we may want to follow a similar process to what I mentioned before. And so I'm going to apply, and what we did in this work was to apply a very similar methodology to the one that we advanced for grid scale storage, which involves looking at the resource management problem. In this case, it's critical to look at energy consuming behaviors of people in vehicles and the diversity in energy consuming behaviors. We want to consider battery capacities, battery costs, charging locations and charging power. And the idea is to estimate targets for batteries and charging infrastructure that could really help with electric vehicle adoption. And so that's what we did in this research. And what we were able to do is to model the energy consuming patterns of people in vehicles around the U.S., looking at where they stop for how long, and then from there, thinking about and identifying strategic locations for charging stations, the power that these stations should have, and also then, importantly, quantify the effects on the technical adoption potential of electric vehicles. So what we found was that really layering on home charging, it can be of the level two variety and then workplace charging, highway, fast charging, as well as charging at certain overnight locations, this combination of these approaches can support electric vehicle adoption. And importantly, what this set of strategies allows you to do would be to mix and match different strategies depending on the particular situation in particular regions. And so that's something that is sort of helping with planning and thinking about where to invest in electric vehicle charging, considering people's behaviors. And essentially the goal here was to find, identify charging infrastructure locations that can offer convenience to people. Now, when it comes to batteries in this context, what's important to consider is that it's really the lower cost electric vehicles which have lower battery capacities that most people would be able to afford. So if we're talking about rapid adoption of electric vehicles, we need to think about planning the charging infrastructure around the battery capacities that people can afford. As battery capacities increase, energy densities improve and so forth, then the possibilities for charging infrastructure, it gives you a little bit more wiggle room. But at this point in time, considering the urgency of the problem, it's important to plan around capacities that are close to the vehicles that are available today at low cost and then also plan in flexibility for the future as battery technology continues to evolve. Okay. So just in a couple, I think two remaining minutes, I just want to touch on very briefly another kind of research. What I've talked about so far is about estimating targets for these important technological components, in this case, different types of energy storage that can help for rapid decarbonization. But what about the process of getting to where we are from at this point until the process of getting from where we are now to reaching those targets? That is going to take a certain amount of time. It's going to take a certain amount of investment. How can we make that process more deliberate? Well, one of the methodologies that we've advanced is to understand the mechanisms of technology innovation. So I was initially motivated to look at this problem by the technology of photovoltaic modules, which, as I mentioned, declined in cost by 99% over four decades. We've since then confirmed last year in 2021, we confirmed that lithium ion battery technologies have seen similar rates of improvement. So these are really standout technologies in terms of their rates of improvement. But the motivation originally for this was photovoltaics and asking, I would go around and talk to people and say, why do you think the cost fell? And everybody had a different answer. So that really begs the question, how can we study this quantitatively? And so that was the motivation for this model to really understand the mechanisms driving technological improvement. Technological improvement happens at many different levels. Here I've aggregated these mechanisms into three levels. So we can think about low-level mechanisms. Those are mechanisms happening at the level of devices or manufacturing processes that affect costs. And then high-level mechanisms are things like research and development, learning by doing economies of scale. And then policies are certain government policies can play a critical role in stimulating either activity in research and development directly. That's government-funded R&D or in incentivizing the growth in markets. And in the case of solar energy, both of these policies were really essential in driving down the cost of solar energy. And so really the quantitative innovation in this work here is in point two, what I'm showing here. And that's really in the basic ideas to relate engineering variables to cost change. And the challenge comes in in that you have many variables changing all at once and they're not all additive. You have many variables like something like efficiency is unitless. It's not an additive cost component, but it has an effect on many other cost components. So in this model, what we're able to do is tease out the contribution of these different variables that combine in different ways, not only additively, when many variables are changing all at once. And I'm going to jump forward to the results that we got when we looked at technological change in lithium-ion battery technology and specifically lithium-ion battery technology costs to see between two periods of time, as indicated on this plot here, what were the main drivers of the cost decline. And in the interest of time, I'm going to jump forward to the results. So here are the different variables that we looked at. And this is available in our paper. I'm not going to go through all of these detailed variables, but we have a cost model that relates these engineering variables to costs. And then we looked at high-level mechanisms. One of the things that you'll notice is that R&D played a really important role here. And that includes both public R&D and private R&D. The other thing that we were interested in is what is that R&D? And we were able to identify that chemistry and material science played a really essential role within that category of research and development. Some of the general insights that we arrived at were that research and development can be a dominant mechanism of cost decline even after the technology is well established in the market. And again, the idea here is to move toward quantitative insights, because of course we could say, yes, research and development is important, but the question is really, how important is it? How does it compare to economies of scale? And what might we learn from the underlying mechanisms in terms of what could be effective as an approach going forward? And one of the things we learned about those mechanisms is that one of the reasons why this technology may have succeeded is that it had access to diverse chemistries and materials that could be combined in a modular design. And so that's something that could also help other battery technologies. The model that we developed can also be applied prospectively to examine strategies for the future for different electrochemical batteries or really any technology. It's adapted to the particular technology, but that's something that can be done. And the idea is that we want to use these models to support advanced evaluation, prospective evaluation of development strategies for technology. So I will stop there. I hope that these examples have given you a sense for how we can potentially be more deliberate to support and maybe even accelerate progress in storage technologies and other technologies that are essential for society. And I look forward to your questions. Jessica, thank you so much for the wonderful presentation. We have quite a bit of questions, but we're a little bit out of time. So let me just pick on a few key ones. So in the first half of your presentation, in discussing grid-level energy storage, you showed a large number of plots on dependence to region, time, and so forth. And I noticed that many of those plots are fairly monotonic. So in terms of choosing an optimal solution, did your study reveal any sort of unexpected findings or unexpected optimal points in terms of what is the best configuration for a certain type of problem? Yeah, I think that the key finding of the study, which really I think applies not just to informing storage development, but also thinking about demand management, the role of supplemental generation for reaching a deeply decarbonized electricity grid, power grid, is that there are these large shortage events that happen infrequently, but really define the capacity of energy storage or something else that you would need in order to support this decarbonized power grid. And that was something that hadn't been uncovered before. So really for that study, that is the key result is really the nature of those larger supply shortages that wasn't known before. The reason it had been missed is that we had tended to focus as a community on one or two years of data. But once you start to extend out over 20 years of data or longer, now we're looking at 40 years for each location, then you start to see these low frequency, high amplitude fluctuations that really matter. So that was really the key, I guess, surprising results. And then of course, all of the quantitative results, what is the cost target? By how much does storage do storage costs need to drop? How important is the role of combining solar and wind together? Those were all results that we couldn't have estimated without a quantitative model of the kind that we developed. Great. Another interesting aspect you showed in the second half of your talk is the importance of spatial resolution when it comes to optimization. And I think you alluded to two very important aspect that this is extremely important for equitable energy access to look at the spatial dependence. And I think you were also hinting at some interesting coupling between the spatial and the temporal aspect. Could you maybe also point out a couple of interesting coupling that has both the spatial and the temporal component. I think you have one plot showing the time that they have charging and the location of charging. Yeah. So I mean, we see that generally speaking, that kind of coupling comes up in a number of areas of transportation when we're thinking about decarbonizing other modes of transportation, other types of vehicles, not just light duty vehicles as was the focus of our study. But yeah, in that particular study, it's, you know, so this is the thing, right? We want, we really need a bespoke model if we're evaluating these technologies with the goal of being, you know, a bit more deliberate about our technology development decisions. We usually need a bespoke model to address a particular energy service. And in the case of transportation, it's really important to think about that as you highlighted that temporal spatial component in order to provide convenience for people. So we really, and the diversity across the population, you know, so, you know, if you're in a car, like you alone are deciding where that car goes, and we need to understand and provide technological options for people that, you know, that take into account their behaviors and the diversity in their behaviors. Now there are certain factors that cause certain similarities across behavior. And that is like, for example, the diurnal cycle and the fact that human beings need to sleep, most of us, right? So I don't know over there in Silicon Valley, maybe people don't have to sleep, but, you know, most people have to sleep and maybe for ideally seven, eight hours a night. So if you can put chargers there, then, you know, they can be of a lower power of variety. Wherever people park when they're at home, that can be really important. So these are the kinds of rules of thumb that we can learn about. And then for each location, you really need a different solution that matches that location. But it's, you know, it's, yeah, I mean, I think that for some of these decarbonization challenges, it's really important to examine behavior in this, in this very fine grained way. Thank you, Jessica. Oh, please go ahead. Hi Jessica, very nice talk. I want to ask you a question. I think it's probably embedded in your model, right? So this energy storage application for solar and wind for the grid. We often talk about long duration. This all type of timescale right there is probably building in your model in the use case, like 10 hours, 72 hours or longer. So one thing has not really showing up as much maybe even going to seasonal, right? So, so how is this influence in your model? And that's the first question related to that is also, well, all these stories technology has a lifetime. Palm hydra can go nearly forever, like half a century long, battery one, and that as it may be five years, maybe shorter and lithium iron, I don't know, seven, eight years, 10 years, like how is the lifetime of this technology like building to the model? Yeah, yeah, thanks. Yeah, so the way that we constructed the model developed the model really the to answer your first question, the results about the duration of storage, that is an output of the model. So what we do is, you know, we simulate this power system in order to understand what the cost optimal mix of renewable energy, other energy supply and storage would be in order to provide a reliable electricity supply. And then what we do, and this was really I think the key innovation in terms of relating this, these sorts of simulations to technology features and really understanding how we can use models to inform the prioritization of technology features. In this case, we're looking at those two dimensions of cost but also estimate cost targets. So this was really the advance that we made in that 2016 paper with a former student of mine, Will Braff and also Joshua Mueller. And that was to, you know, essentially consider a full range of different technology costs and features. And then from there for each cost of technology, we estimate an optimal size. And so the duration to answer your question, you know, the storage duration that is associated with a cost optimal energy mix is an output of the model. And so what we see for that particular location, you know, if you're performing price arbitrage, the duration is a couple of hours, that's the optimal duration of storage, right? And of course, that's the energy capacity divided by the power capacity of storage. If we're looking at this deeply decarbonized energy mix, which was the example I went into in more detail, then the duration of storage that you're going for goes up to, you know, above 100 hours. And the reason is, and this is if you're using 100% variable renewable energy, it comes down if you're using other supplemental generation. But that's really essentially where you can see that is in the slopes of those ISO cost or ISO performance lines. Those slopes are the hours of duration. So that's really where we can see that result. But yeah, the result, so this was one of the, you know, the insights really allowed us to then say that, yes, long duration storage is important for deep decarbonization via this pathway of using a lot of variable renewable energy. So I know that's a kind of a detailed answer, but hopefully, hopefully clear. And yeah, to get to your second question about lifetime. So once we have these results, so we have these cost competitive targets, then we can ask, how do technologies compare real technologies that we have compared to those targets? And the lifetime is a critical determinant of the overall cost of a storage technology in terms of the service it provides, the cost per unit service that it provides, because if you build all of this capacity and it only lasts for, you know, seven years versus 40 years, then, you know, you're basically dividing your costs by, I mean, it's an approximation we have to take into account various other parameters, but you're basically, you know, amortizing those costs over either 40 years or seven years. And so it has a huge impact on costs. So the way that lifetime comes into this is in estimating the cost of whatever the storage technology is that should be taken into account when we're looking at the cost estimates for these different technologies. But yeah, thanks for the questions. Yeah, thank you, Jessica. Back to you, Will. All right. I think we are a bit out of time, so we can have maybe the rest of the questions for the panel discussion. So thank you so much, Jessica. So, Yi, would you like to continue? Yeah. Yeah, thank you, Will. Nicolo, Nicolo, let me bring you to the stage. Let me give introduction to our second speaker, Nicolo Compagno. He is the solution manager at Mackenzie. Certainly, I think this is the first time we have Mackenzie background, a person to give a presentation. Nicolo is, well, you know, we all know Mackenzie is one of the leading business consulting firm globally. Nicolo and Mackenzie co-founded and now co-manages the battery's insight, which is a stop-up within Mackenzie, developing new suite of product to support clients in the battery industry. So Nicolo has the training with electrochemistry background, you know, with PhD in there and he got PhD in 2016 in Belgium and doing part of his research at MIT. I think with this scientific background, now with going a lot of business insight, he will be able to share us with very interesting insight, particularly the topic today is very interesting, MMC versus LFP. Nicolo, let me get you to share the screen and share your insights. Thank you very much. Can you hear me all right? Yeah. Fantastic. So thank you very much for the introduction. So, hello everybody. Again, this is Nicolo Campagnol, manager of battery insights in Mackenzie. And today I'm going to talk about this phase between these two technologies, these two chemistries, MMC and LFP and how it's not just about chemistry, but it's also about how the real world, the technology, the markets, quality policies actually influence the choices of these chemistries right there. So just a couple of words about Mackenzie. As you said, I think Mackenzie is relatively known as a company. We are a business consulting firm, but I think one few things that people don't know is that we actually have, for example, few sub companies or startups within Mackenzie. They are still part, obviously, of the modern firm, but we act somehow independently within and we focus on a specific aspect of a certain industry. So one of them is mine spans, for example, focusing on a mind-by-mind calculation, the cost, emissions and similar energy insights that gives an idea of all the source of energy and how the energy transition is going to look like and the sources are going to look like. Then there are hydrogen insights and battery insights, battery insights where I belong, that we focus only on the battery industries from mining, all the way through recycling. And our important thing that a lot of people might don't know, even new clients, sometimes they were very surprised, is that we do not hire only smart people from finance and from economics. We do, obviously, and we're very proud of all those, but we also have plenty of colleagues, including me, with a PhD in technical subjects. In my case, it's electrochemistry and batteries, but we also have plenty of colleagues who came from industry. So from automotive OEMs, from cell players, from active material players and so on who joined the firm and actually built the knowledge and bring the knowledge to our projects. And lastly, and this is something that I was particularly interested when I joined the firm and I was so happy to be able to do, is that for many aspects, we actually have our own, we produce our own data. So in the case of automotive here, it's me in a teardown lab of our batteries and automotive. So we tore down cars in the recent past, including the battery cells. And I was responsible for the battery cells teardown. So this is a great occasion for us to actually learn how the industry produces these battery cells and how they look like internally. And it gives really like the knowledge, it gives us the knowledge to deliver a better job, basically the deliver better insights to our clients, because we are, they are based not only on knowledge that are, they are predating McKinsey, so carry case of people with industry experience or due only to academic work, but actual real data from the real world. So in my talk today, and it's about again LFP and NMC, I will talk a bit about the basics. So the science behind it, not going through deep in details, I know there are plenty of smart people with PhDs and whatnot in electrochemistry, so I'm not claiming to teach you this kind of things obviously. But an interesting jump that I will do, at least for me, when I was at the university especially, is how this science actually translate in real life. So what's going on and why in, in the real world out there and then how the choices that we are taking today actually influence the future trends. So without further ado, easy. What's, what's in a battery, right? So a battery normally is formed by form modules and then cells, cells read what we normally call manuals called batteries, but that's, that's the technical name and every cell, whatever chemistry you are working with is composed by an anode and electrolyte with or without a generator and a cathode. And if we have to take six parameters, six performance parameters, we say like, these are the important things that I want to look into for my battery, right? Although all three components play a role in all, you might try, you might group them and say that the anode is actually mostly responsible for our, for the charging speed and the number of cycles. The electrolyte is mostly connected to the safety of the battery and the battery cell in particular. And the cathode is responsible both for the sustainability, as well as for energy density, which gives the range in the case of a car. And finally, for the cost is the most expensive component of the cell. So it's driving the cost of it. If we take the most common chemistry, cathodic material chemistry today, these are lithium iron phosphate. So it's somehow a fertilizer with a bit of lithium inside and a layer structure, nickel manganese cobalt oxide that is intercalated with lithium. Okay, it's called NMC. Okay, these are the two beasts that we consider normally when we talk about lithium ion batteries today. There are not, these are absolutely not the only ones, absolutely not the case, but they are far by far the most common. So if we compare the two of them in the same scale of values, if you want, it's a bit qualitative obviously, but it gives an idea that we were comparing the cathode and the electrolyte before. We see that at first glance, LFP is quite well performed. So it's better in safety, potentially better in lifespan is obviously up for debate depends on the condition and similar. For sure, it's better in cost and in specific power for the way it's actually produced, because normally you make smaller particles for LFP than you would normally do for NMC, which is as most of us might know about kinetics drive what you can do in terms of power. On sustainability, the fact that LFP is actually made for some so common materials, normally should drive the better sustainability overall in theory, at least versus NMC, which is based on the transition metals, nickel, cobalt in particular, they are more difficult to mine. And therefore, I have more energy that you have to put in and therefore more CO2. In all these, we see that there is one parameter which is where NMC is winning, which is specific energy. And specific energy, it's actually the most important parameters for one market, which is mobility. The mobility drives basically the demand for batteries today, as well as probably in the short future. So although LFP looks better a bit across the board, if you want, you actually have NMC, which is best for the main applications of it. And according to one of our scenarios, it's going to be the same also in the future, keeping the same anode and same electrolyte. So according to the science, according to what you would drive through a real drawdown, okay, this is better for this, this is better for us, you would expect NMC to be leading almost 100% in mobility applications, while LFP taking the lead in grid storage. And this is where we want to move away from the size, away from the lab and have a look actually what's happening in the real world. Well, in the real world, especially in the recent past, that was often exactly the opposite. So you have this in China in particular, which is the biggest market for batteries. There was a particular attention on LFP. And you have plenty of vehicles which were running with LFP batteries. At the same time, in the western world and far east, so Korea and Japan, you would see plenty of NMC powered grid storage among the biggest one, even the biggest one actually that comes in mind, lithium ion ESS in fields like you know, plants you want to call them like that, they're actually based on NMC still today. And now we are moving towards LFP, even for those applications. So why is it the case that it's often, not always, but it was often exactly the opposite. Well, there are several reasons for that, several aspects and several drivers. One of them is governments, which play an important role. And Jessica was actually referring to that as well. And so it was quite interesting. So one of the effects of the policies that was adopted by the Chinese government for safety, for buses, basically was to, and they didn't formalize it, but basically they didn't put an MC producers, foreign MC producer, because most of MC producers were foreigners, in the list of the players could get a tax rebate or anyway, a subsidy for batteries for buses. And that drove all the industry to move towards LFP in China. So virtually all buses that are produced in China use LFP. And this is actually the golden standard also outside China by now. And at the same time, the same government for other reasons decided that they wanted to push their electric vehicle passenger cars industry towards a product with higher performance in terms of range. So they were giving subsidies only to cars or graduated in a higher for cars with longer range than cars with shorter range. What happened then is that the Chinese automotive player decided to change, not only make batteries bigger, but also change the chemistry from what was adopted before LFP to an MC. So very interesting how, even indirectly, I don't think you don't have to see one of these government acting in the interest of one specific chemistry. Obviously they don't care, but they always try to do their best for what the people want. But in pushing this direction, actually they push exactly the different, like the opposite direction, two different types of interest. So it's quite interesting the results of the policies. At the same time, there is another very important driver, which is technology. So again, from our lab experience, we know that OK, NMC should be more expensive and has normally higher energy density, but LFP is on the opposite floor. So lower energy density but cheaper. Well, if we take two cells that we open from our tear down in China, we observe that LFP cell that we open and that NMC811 that cell that we open actually had a rheumatial cost, it was almost a long pair. So the drivers were several. One reason was that the cathode, the cathode thickness was much smaller in the case of the LFP versus the MC. The cell itself was smaller. Therefore, there was more packaging than back. But it is interesting to see that in the end, a product that you can price much lower than the NMC811 was made with a rheumatial cost that was almost the same. Therefore, pushing the price premium over rheumatial to a much smaller amount in comparison. So again, price costs not necessarily better for LFP if you're designing in a certain way. On the other side, and this is an interesting result from a Chinese producer, BYD, in their BYD Han, they were able to, with a new invention, this blade battery, so skipping the module and using directly as a component of the cell, as a component of the pack in a way, also structural component for the car. They are able to use an LFP chemistry guaranteeing 600 kilometers of range. To give an idea today, we have around 400, like my car has around 400 kilometers of range, and sports and NMC battery. So they used a cheap technology, a cheap cell, LFP, and they achieved a very high energy density, which is higher than what NMC on average can propose. So very interesting to see how these two levers can actually basically flip the results upside down. Another important, very important driver, especially today in the market is the price of raw materials, right? The raw materials in this moment are going a bit crazy. And we could go before, we could go all the way to 2020, then the trend will be very similar. We see the lithium today is around five times what used to be at the beginning of last year, and all the other raw materials are also up. So you have nickel and cobalt up 20%, you have cobalt up at 88, and aluminum as well. What this entail? Well, this entail that the whole industry, whether it's batteries or not, actually, saw an increase in the raw materials, therefore they had to increase the price. And this is the first time in, I don't remember how long actually, Professor Chanix showed a bit of how the price went historically. It's probably the first time that we see bump up in the price, at least in the cost, but also in the price from what we hear from our customers in the price of batteries. And in a moment where we buy policies everywhere in the world, we are pushing to get more electric vehicles in the market. So that's what happens. I mean, lithium is in both lithium mine batteries, both in LFP and NMC, but it's LFP is still cheaper than that. So what's happening is that many players, this is only based on announcements, so we're not even checking if they are following suit with what they are saying or not. This is just based on announcement. You see that while all Western players, these are only Western and Far East Asian players, produce cars mostly with NMC or NCA, a lot of them in 2021 announced like, okay, we are at least exploring LFP for our entry level version. So we see that the pressure in raw material prices pushed the OEMs to have a product, an entry level product, which is interesting for customers, push them basically to explore this new chemistry, this that they were not using it before, which is for our client, this is a bit of a double-edged sword. On one side, it's good for them because they can expand their portfolio and bring more product to the market, which is great because you have more choices as a customer. On the other hand, having a more complex portfolio brings also complexity. So it's a very important balance that you have to strike as an OEM, considering the amount of cars you produce, considering the amount of demand that you have between how many chemistry and how many offerings you want to give versus how many products that you want to, in order to approach more clients. Now, this is all good for today, right? Already 20, this was all the announcement for 2021, but let's see what all this story of LFP and NMC will, how will this story impact the future, right? So here, it's a very easy schematic of what you normally do at the end of life of this battery. So LCO, lithium cobalt oxide, as well as high cobalt NMCs, normally would go for recycling. The value of the material is so high that it makes a lot of sense, especially with the price of raw materials today, to send it to recycle. At the same time, one of the reasons why LFP is so cheap is that the raw materials are cheap. So recycling doesn't really make a lot of sense. There's a lot of cost involved in recycling, and there's not much to get out of it, right? Because it's iron phosphate, it's not much there, right? So reuse could be the best choice. In the middle, you have NCA, NMC811, and so on and so forth. All these, it's a bit of a question. Should I reuse? Should I recycle? And what's best? Well, the increase in LFP should, in theory, push up recycle. While if we stick more to NMC, it's a bit on the higher side, you would see more, sorry, reuse. Well, if you see more NMC, even today, in 10 years, you would expect more recycle to happen. So that's a market that might be very interesting tomorrow, right? But this is not, again, always the case, right? Unfortunately, so we did a calculation also for a player who was interested to see what to do in the end with their cars, with their batteries, on the cars. And what we observed, that for sure, for an LFP, you want to go for reuse, there is no much value on recycle, so you rather reuse and then recycle. But for NMC, except if you expect the raw material prices to be very high now and very, very low tomorrow, you can always recycle at the end of life, of the second life. So the business case of second life or reuse, it's very compelling if you're able to fix their economics of it. So if you're able to have enough, if you have enough volume and if you are efficient enough in the dismantling and reassembling, which is not too big of an effort, because normally this battery pack are extracted from the car and put on racks. You don't disassemble the modules, you don't disassemble the cells. You test, obviously, the pack, but then you put it in a rack as is. And the whole intelligence is to make the whole product in the end is to make all the batteries work together, because you can immediately have an instant lift from 2016. You have a Tesla 3 that went in an accident two years ago, so everything has to play together. That's where the, where's the difficult part, not really in the disassembly and reassembly. So that's interesting because even in this case, you would see that you would postpone a lot of material that will come from recycling because of the use of second life. So the business case for recycling, the business case for second life, they need to be taking into account all these trends, but also what is the economics of each of the two. If we see the other raw materials, a bit of all the raw materials, because that's also quite important to understand, okay, how the market will evolve tomorrow and why would that evolve in that direction. We see that cobalt and lithium are absolutely depending on batteries. So a swing in the battery demand up or down or a swing in the supply basically would impact each other. So it will impact prices heavily. Okay, so then you see that especially for in 2030, lithium would be mostly driven by by the use of batteries while cobalt because of the transition towards away from cobalt not as much. But there is another element that is very much under spotlight these days, which is nickel. Nickel is divided in nickel class one, nickel class two, nickel class one being of high purity, nickel class two low purity is still nickel because the element is nickel, but the products are different because there is a lot of effort is needed to switch from one to the other. From one to not much, but from two to high purity 99 it is very difficult. And you see that nickel class one actually is under a lot of pressure on batteries. We're expecting by 2030 the majority of nickel class one actually to be demanded by this industry. So what happens is that OEMs see this time the producers of cells, producers of cars, they see these trends. And so for the first time is they are actually moving upwards in the value chain a lot. So for the first time actually investing in mines. So in the past we never heard about big OEM investing on an iron ore mine. That's something so odd because it's not really their business. Well now we have plenty of agreements in terms of take or pay, in terms of even direct investments of the low part of the value chain with the high part of the value chain with the mines to ensure that there is going to be enough investments to build these mines to guarantee there is going to be the supply of these raw materials. And I spoke a bit of nickel and just bring this example to close to because it's all a loop. So we saw that nickel prices actually influence heavily which chemistry are going to use, whether you're going to use LFP or whether you're going to use NMC in the future. And at the same time the amount of NMC and LFP also influence how much nickel demand there is, which influence the price. So it's all really, really a loop and a circle. So this gives an idea just a small deep dive on what we are expecting in terms of plus one, plus two demand and you see that the supply by 2030 could make it in terms of units, but it's completely decoupled like it's not perfectly proportioned in terms of which class it's needed. So there are some, while the units are there, so there is enough nickel mined or supplied by a different resource, there's going to be a shortage, potentially it's going to be a shortage of high purity nickel products and this will probably potentially have impacts in terms of decisions for OEMs and as well as mines. And we see on the right side a bit of a numerical tweak of what could be in different scenarios for materials and you see that for example comparing cobalt and nickel in an NMC. You see that an increase of 200 percent of cobalt would not have the same impact of an increase of 200 percent of the nickel price. This is about it from my side. Thank you very much for the attention and yeah, I'm looking forward to your questions. So thank you so much, Nicolo. This is a very interesting comparison from so many different angles. I myself certainly have witnessed this whole change of LFP and MMC just within the last let's say five, seven years also. I mean just very, very interesting story right there. So looking at these two comparisons, let me open up the comparison a little bit more. I mean in the audience is also people asking this question. Certainly the highest value is on the cathode and if I recall back in five, seven years ago it looked at the battery different components core. So I do this also separator, this electrolyte but that part just keep get the suppress is percentage wise. Nicolo, do you also compare what's going on there as well? So what happened there to make the cause? What did they do to get so much lower? Yeah, so there are three main, globally there are three main drivers to bring down the cost or up. So one side you have the technology. So how much you are able to squeeze out of this cathode and we saw that the cathode even by the increase of nickel but not necessarily you basically were able to get more million per hour per gram in each cathode and that helps the whole system. Or you were able to do for example you went from what was I eight microns to now six, five microns in terms of thickness for copper foil and that obviously reduces the amount of copper, increase the energy density and potentially increase the price. Then there is another important parameters, parameter which is the scale. So what we do in McKinsey we have this cost model and we simulate the cost of each single plant in the world. So we take the global average and the global average is pushed down by the fact that there are more giga factories than there used to be in the past. So while a single factory doesn't see the scale effect necessarily, the whole industry sees this push down because of this. Then there is another parameter which is not necessarily going down and that's the raw material prices and that's what we were discussing today and until a couple of years ago actually raw materials were getting under control, we were getting out of 2008 which sort of material prices going up and down was going down. Now prices are to the roof and that's what we saw like LFP above 100 dollars per cell per kilowatt hour per cell. So it's impressive the prices are spot prices today. So I'll have a couple more questions we'll go to Charmin any time. So I also look at the, well let's see, the cathode LFP and MMC anode graphite now silicon is also coming out of course graphite's cost is low, silicon coming in increase the energy density of the cell. Would the silicon story coming in change how LFP and MMC's you know competition? The reason to say so right Nicola you are aware of this is the silicon coming in can boost up the energy density of MMC more than LFP due to the voltage reason LFP is low and the silicon you have some voltage loss the boost of MMC is higher is more right and then that will also is further because a lot more energy this will drive down the help drive down the cost assuming silicon size cost is can maintain low. So what's your what's the thinking you know in Nicola in your guys analysis did you see this potential things can happen or what's your prediction? Yeah we are monitoring quite quite closely both silicon and all the type of silicon because there are so many different options to get silicon done right high silicon done as well as lithium metal whether with solid state or not and again the math you know science says like do an MMC right with with silicon or even with lithium metal you want to use an MMC well not really especially with lithium metal we see more LFP for a start rather than MMC why because you can keep the voltage a bit more under control so it's easier to do because there it's much easier to make LFP because LFP it's you know you don't really have the same type of purity of the air that you need or controlling in terms of humidity and so it's easier to make LFP so what we saw for example is one of the start of modern one actually was that they were they were preparing something which is super advanced say solid state silicon silicon whatever not but they were using LFP as a as a cathode rather than an MMC although as you rightly say you would expect the biggest boost through the future what we could see is that there is going to be a be forkation between high performing and low performing vehicles or products and in that case you're going to see two different drivers and mostly it's going to be cost so if silicon also brings down cost which is what we expect in the long term you might see silicon on both sides both in the high end products or in the low end in the low end product while if it doesn't you will see it only in the high end high end products which are normally based on MMC that's what we see and what we can predict at the moment yeah I will Nicola outstanding presentation I I think it really your work highlights the importance of a system level analysis LFP actually it's at the cell level per kilowatt hour it's the same cost as an MC although most people say LPs less expensive but it's also lower energy density but it's really amazing to see that the improvement comes out at the pack level and not at the cell level in terms of cost due to higher volume and weight utilization so I really just want to point out the importance of the system level analysis very quickly you you show this really interesting slide which is this opposite trend between China and say the western world and where China was pushing you know through its 2015 policy which basically eliminated LFP from the incentive program but yet that's how the industry developed in China and in the US where we're pushing for grid level storage they're going for MMC instead of LPs so everything is not making a lot of sense here can you explain a little bit how the market developed in this way where were the what were the market forces that led to this outcome I just thought it was really interesting this everything is out of balance yeah so it's um it's a bit of that's the way it is you know like not much and it started like that so China started with LFP which was easier to produce and that's probably why they focused on that they didn't have the same protection in terms of IP although for what we understand this is not a showstopper this is not there's not been the showstopper the fact that LFP was protected by IP because you can always buy and pay it's not it's but the Korean players and Japanese players developed first MMC mostly for consumer electronics where cost was not so important and diverted into into ESS grid storage as well as automotive in the case of the first Tesla right um and and that's how the west as well as far east got the MMC propensity like the MMC focus uh first and at the same time in China because they started with LFP also to protect their own industry although the product was less performing in terms of energy density they they used used the different levers and uh I mean safety is for sure a good a good lever because safety it's it's true left is more safe than so it makes out of sense to have for example a bus that is full of people with LFP so they they use this um this um lever also uh to um somehow yeah these levers basically made the the the local producers of LFP thrive in China and then becoming became also MMC producers in the future in the in the like in the recent history so Nicole just to confirm so what you're saying is that the regulatory aspect of safety drove the scaling up of LFP whereas the incentive continued to help MMC supply to develop in China it's not only that but there is a component to that yes there is a component to that and as I said there is plenty also of IP that was the same of historic propensity of some players to do another one product rather than another so yeah there is there are several aspects to it and and that's how the market developed but the moment is a bit rebalancing it uh itself there are there are a number of battery historians out there I think we should have more battery historians to look back back to you E. Yeah well I want to jump in a little bit Nicole if you don't mind this question we are asked there also seems to be a consideration if you look at Chinese policy and 2015 like they started to you know only giving incentive to the government incentive to the high energy chemistry MMC but not earlier it's they were in 2015 one reason maybe it's a rumor right is because the MMC technology in China the development at the time before 2015 wasn't ready to go to the global market to compete so not until 2015 the local supplier there you know the technology gets there then they started to giving incentive right to to MMC way until they become a lot more competitive otherwise the foreign supplier coming in will wipe out the whole Chinese market in your analysis did you see this correlation I mean it's hard to to to comment on on policies that the Chinese government or any other government took so I'm not I'm not going to to comment on that it is possible that the local players were not prepared but I wouldn't I wouldn't be able to say that it was a decision of a policy targeting a certain player or a certain industry to do a certain result yeah it's it's hard to say maybe there is a correlation but I don't know if there is any causation yeah okay maybe I'll just ask one last question then we should bring Jessica also to the stage there's a question from the audience this is interesting one so I thought about similar thing is there a reason you know investment and nickel purification is lagging you mentioned you know class one class two right high purity one is in high demand for the batteries industry you know we have a lot of nickel in the world and what about purification wouldn't that make sense to just to invest a lot more in the purification process and get a class one up there and also low cost as well yeah so the question is the cost that's that's the big issue right so the without going into the details of of sapro light and laterite type type of war right so there are basically nickel likes very much iron right they are really they really get along and so when they are together within the ore it's difficult to separate them and if you have to make stainless steel you don't have to separate them so that brings a product in the market which is the class two which has a mix of nickel and an iron which is perfect for the for the stainless steel industry and brings down the price of nickel in the in this mix uh a total obviously taking this out brings a lot of cost right and so that's why it's it's very complex to do and even not always from the same ore you can do economically uh you can make uh nickel class one so I'm not saying that are or that are class two and or that are class one that's not correct but there are or that are more prone to be used for class two and or there are more prone to be used as class one then price can always go to the roof and you know everything is is uh economically uh interesting at that point right so but uh except in that case normally to to to stay conscious you rather go for for example um sulphidic ores tend to be better to make um to make uh uh nickel while uh you know mix already with iron uh like in the in the nickel belt so that would be between uh fuba in tenizia and all those tend to be better uh to make uh uh nickel pig iron and so on yeah okay well I mean with this now we have also Jessica uh back um and let's have uh you know some more discussion uh where if you don't might let me kick off the first meeting and and this uh you know panel discussion I'm looking at both of your talks fantastic I mean giving uh I think all the I'm giving myself you know many perspective you know from your angle you analyze the problem so perhaps first question I want to ask uh both for Jessica and Nicola um your um and your research analysis um particularly you know this lithium iron you share some of the analysis uh on the lithium iron did you see what are the technology gaps very clear to you to say hi we need to do more R&D on those uh and I'm sure in our audience right here there's a lot of people doing R&D so they will be probably wondering whether you know your analysis can point to the direction or confirm what what they are working on is uh it's worth doing uh who wants to take this first maybe Jessica do you want to take it first we'll give Nicola a you know 30 second break yeah sure um yeah so I mean I think it's a great question um you know as everybody most people on this call and listening to this probably know within the category of lithium iron battery technologies there's quite a diversity of um you know an answer we can broaden that even further I mean there's a great diversity of different options and directions for R&D um I think that is one of the reasons why I think it's can be useful to use the kind of model that I presented prospectively to understand okay if we you know improve these cell components let's say we I mean there's there's really like you know two if we're talking about costs and costs isn't the only performance metric of interest we also um and as Nicola mentioned we care about the specific energy and the energy density for certain applications but if we're talking about cost there's really um this question of you know various efficiency parameters um and both both in the actual cell operation but also in the manufacturing of the technology and then there's the question of the upfront um material and um you know labor investments that go into manufacturing any of um one of these technologies and so unfortunately I'm not going to be able to give you like the silver bullet here in my answer but what I would say is that there's always there's often a trade-off between those two um at least in the um cell operation in that you know we can often save on materials and you know low-cost manufacturing processes and give up a little bit in terms of the operational efficiencies but I think it's not always straightforward ahead of time to know what that trade-off is and that's where I think these more detailed cost models can be used to really understand what those trade-offs are now going forward as you know um I mean every design decision within the cell affects other design decisions which is why it's it's important to really model not just the variables that I mentioned but their interactions and that's actually something that we're doing so we're hoping to publish a paper in the near term taking a couple of dominant um lithium ion battery technology designs and applying the model that we've developed prospectively to investigate some of these these trade-offs and try to get some quantitative insight. Nicola, I had the time to take notes so that Jessica didn't so I will actually go for my top three that doesn't mean that anything else is not important quite the opposite but I give you my top three and you tell me what what you think yet so for me we really should focus on chemistry leap so we need to go for the next gen this is the way that we can drop both cost as well as improving the energy density sustained like substantially and that would be mostly involved in the anode choice that would be lithium metal or silicon I I I want to keep my my mind open any of the two is a great solution but this is really like the first objective on the short term that we can do to decrease cost improve CO2 emission and and as well as energy mass the second one would be to go away from uh this uh if possible from the cathode that's we have today okay so try to make something which is uh for example we have sulfur that might be a bit too far although there are some people who can do it today but anyway something that is relatively cheap and relatively well performing it's a key to answer the the requirements of the market today so it's anode cathode and if you have to change the electrolyte you have to change the electrolyte obviously but not necessarily and the last point if you if you see how much does it cost if you see like that in the process uh to produce a lithium ion battery what actually takes the energy and the capex in the factory you want to tackle the drying part okay so that's what a lot of people are now proposing as dry coating this could be one of the options not the only one but this is really like the point the pain point that we can that you want to with the highest impact you can improve everything obviously but if you want to have a high impact 50 percent uh drop in that would have the highest impact the globally versus 50 percent of another part of the process yeah well thank you so much uh nico the really good point i mean these are three points right i think really is uh uh smiling on this you know i i can appreciate that more let me be specific we're fearful to try any time i'm just asking specific questions still regarding to my first question hey giving the lithium cause nicole you analyze right the percentage you will it will take what about sodium yes we have sodium ion battery guys that is actually commercial and it's going to solve a lot of problems but if you end up using a lot of nickel and similar yeah i'm not sure how we can yeah yeah yeah appreciate that then what about solid state right well jessica please yeah please yeah no i mean i was just going to say i was kind of um i had to skip over these results in the presentation because we were short on time but um you know it's interesting also to map on some of the results that we see going back in time um looking at what drove lithium ion battery technology costs down uh compared to what nico said so you know if we look at for example we grouped a set of variables that we looked at in terms of cell performance changes so that includes you know changes to the cell charge density charge capacity utilization cell voltage cathode cathode um foil area was another one and what we see is that the contribution of cell charge density is by far the dominant one um among that category so much larger than for example cell voltage um which was much smaller now that isn't to say that is what will happen in the future but i think it's really interesting and and there are some arguments for really emphasizing that and then the other major contributor is cathode materials prices um looking back and so that's also interesting um you know and then plant size i mean i did mention that research and development had a dominant effect and that's both from the private sector and public sector on cost decline but also you know economies of scale um and processes that can rapidly be scaled up that is also um something to consider yeah thank you will do you want to ask absolutely such an interesting discussion Jessica and Nicola i really appreciate you looking back right because we have now so much data for the development of battery field Jessica you have also compared to the solar field so i want to ask a little bit on on this aspect right so a lot of people are saying okay maybe you know this is it lithium-ion battery has won it's the cost learning curve it's impossible to compete but if we look back historically and we look at lead acid the transition to nickel metal hydride the transition to lithium-ion what lesson can we learn here and and and the specific question i like to pose is in my mind there is a certain learning curve i just got you mentioned the solar cell and lithium-ion battery were very similar on the on the cost learning curve so that's probably kind of the best case scenario for any new technology so what comes to mind is it's the gain in the improvement in the properties and the performance of the technology that really has to be high in order to justify um even operating at the same cost learning curve and then the the related question i have is do you see an opportunity for accelerating the cost learning curve for the next generation technology so that you can learn everything we have learned from lithium-ion battery and then make the next cost learning faster so i know this is a lot of question packed into one but yeah really want to build off what you have presented on the past performance at the past data looking at the evolution of the industry jessica yeah sure um yeah no i mean i think i definitely do feel that we can be more deliberate in technology innovation it's not to say we're always going to make the perfect choices but if you think about how a lot of the choices are made today or you know sort of maybe it's changing a little bit now but it's it's a little bit of a random process in a sense and sort of intuition guided but i think that we can for example um you know both and like i was presenting on both and estimating targets you know so for example you you know and and looking at what specific design decisions might allow for both more rapid improvement and then also the greatest quantitative improvement in cost or some other performance metric so um you were taking the example of lithium-ion battery technologies and asking you know and i think one of the participants also asked a question i saw in the q&a you know is this technology going to um i mean you didn't ask the question this way will but you know other people will ask like well do we need anything else right should we just push on this right um and i think that putting some numbers on the performance that would be needed for certain applications for storage you know so for example estimating that we still need an additional 80 to 90 percent cost decline in our battery technologies if they are to work in the role that i studied for providing reliable deeply decarbonized electricity where the electricity is mostly coming from solar and wind you know that is kind of giving us a sense of how we should like how much we should diversify across different chemistries battery designs etc um and i think the answer is we still need that diversification because it's not clear also based on you know the analyses that we've done and others have done that there really is going to it's really going to be possible for battery for lithium-ion battery technologies within that set to drop in cost by that much so yeah and jessica this is a great point and if you look at solar it's kind of an interesting discussion right as solar is about 15 years ahead of batteries so that was one in which the technology diversification never quite happened there was many attempts at displacing silicon and many discussions of new technology that could be lower in price better in performance um but at least i think over the past 15 years this really hasn't taken place so i wonder if you have any thoughts on the learnings from the solar industry yeah i mean i would say we can't compare you know energy storage with solar energy so as i mentioned you know this set of energy storage technologies it's like such a large set as you well know and you know most people on this call know there's so many options even within electrochemical batteries um and they can serve different purposes so you know in the case of solar energy um there was a certain dominant design based on silicon silicon isn't the ideal band gap material but it has certain properties that make it easy to work with which is also why cost i think there's a lesson there for battery technology you know i'm thinking which which materials lend themselves to scalability and um you know sort of easy to scale manufacturing processes um but it wasn't the ideal material um however it i mean it wasn't the ideal material from a band gap perspective but it had other advantages and it has really come to dominate the market there are other photovoltaic technologies that could still grow and meet certain specific applications you know like um you know for you know for residential applications and and so forth so anyway but solar energy was is a certain conversion and photovoltaics rather is a certain conversion technology when we talk about energy storage electrochemical batteries even lithium ion battery technologies there's such a diversity of different you know chemistries and designs that we're thinking about and so i think it's really not a one-to-one comparison the other thing is that and this was also mentioned by Nicolo is that when we when we look at how like like what markets drove adoption and lithium ion battery technologies and then where the cost in our work where the cost decline came from what you can you know and and also what we did in one of the 2021 papers that i presented was to look at if we adjust the cost decline for increases in battery energy density and specific energy you know what would the rates of cost improvement look like and just to kind of keep this answer somewhat relatively short i mean i know it's been already a long answer but you're also asking hard questions so um that require long answers so i can't take i don't have to take full responsibility for that but i don't think but anyway um good questions but yeah so you know what we see is that it you know there's a way in which you can estimate what the cost improvement might have been if energy density hadn't been a concern right but for the application that this technology was being developed for it was very much a concern both specific energy and energy density now if we relax those requirements you could potentially have seen faster cost improvement again it's unclear what's going to happen going in the future but when you're talking about grid scale storage you are relaxing a number of constraints and so you know that is something that makes me think that the evolution going forward isn't i'm not saying it's it's we're not going you know we can't predict the future in this sense there's not much data to go by here even if we wanted to make a data-driven prediction but um there is the possibility certainly of much more diversification uh among the energy storage technologies thank you jessica sorry sorry for catalyzing uh an extensive answer nicole uh i think the main uh learning for me that's the most interesting learning from uh from jessica's talk was actually this transition towards a 10 hours plus uh requirement in terms of energy density and this is not in terms sorry in terms of hours for for storage and this is not lithium ion right lithium ion is relatively good uh has a relatively cheap power in terms versus uh energy okay while if you want a lot of hours you want something which is relatively cheap in terms of storage is terms of energy but potentially high in terms of cost so that's that tells me that basically for grid storage lithium ion in the future might not be the right right choice so we need we will see as we were saying like relaxing or changing slightly or the requirements we will probably see other technology popping up that are going to be more convenient for uh for that use case so today uh ESS is around four hours and that's lithium ion it's quite good but if you want to go for 10 hours plus that probably not lithium ion at the same time if you see mobility lithium ion is actually in the sweet spot but it depends on this where you strike the balance between cost and uh fast charging as well as range so if we do all for example swapping fast charging is not an issue anymore so you can definitely you might see uh movement away from very thick graphite uh annals for example because we don't need them anymore right so you can increase energy density decrease cost or if you go for something which is like if the price of lithium skyrocket then it's the moment that you would see again let's say a swapping of sodium ion batteries and you just would go a bit more often to the swapping station but but that's it or the charging station obviously that's it so um I agree that it's we should see the curve not as the silicon uh PV curve but it's actually slightly different in terms of energy storage because we have so many different options that we can take and we have just to strike a different balance between the requirements of each market to decide what's the best technology at that specific moment I want to add in something sorry to come back to the same similar topic again very interesting discussion I want to mention this and see what you're thinking Jessica and Nikola it's this leveraging effect from another industry if you look at silicon solar industry you know it actually leveraged the whole semiconductor industry computer chips right and the microprocessor right the whole semiconductor all the knowledge built processing supply chain and this learning is free to solar industry and certainly silicon has also its own benefit why is dominion semiconductor is stability silicon silicon dioxide passivation amazing like processability of silicon oxide all these things coming in solar industry get this for free basically right if I look at now lithium ion for transportation uh for e-mobility and this whole industry get it for free from consumer electronics consumer electronics already explored that for nearly what 20 years right something years before people started series explore for transportation so I think it's still a lessons learned you know you these uh industry mutually leveraging and uh is that the important for us to think further right Jessica you mentioned about 10 dollars perhaps per kilowatt hour of capital costs for you know long duration stories how could we learn from a maybe a completely different industry to leverage that yeah so I do have some thoughts on that and I think you bring up a really good point um again I think the solar case is a bit different from the lithium ion battery case in the sense that you know for solar panels it's absolutely correct that the industry learned from the semiconductor industry when solar panels you know our solar solar panels from the 1970s were providing electricity at 100 times the cost of competitive electricity at that point their their advancement from there was largely supported by government policy and there were two types of instruments market creation and rnd funding but there was a really important role of government policy in supporting the development of solar so I think for many of the remaining challenges including stationary energy storage we need to be um really recognizing the importance of government policy now the thing that we see in solar is that it was government policy but it stimulated an even greater effort in the private sector so these market expansion policies led to a lot of private sector investment and we estimate that 60 percent of the cost decline from um that 99 percent cost decline from the 1970s came from these market expansion policies but it was really private industry that kicked in and and actually did the innovating in that case so um I think in the case of stationary storage what we can learn from that actually we can take a lesson from solar which is that if you get these policies right you can potentially really accelerate development and um you know there's so much more I think you guys are asking great questions there's so much more I could say on this um I don't know if we have time and I don't want to take up Nicolo's a couple of minutes here so um Nicolo um I I think I think yes grid storage is already piggy banking on on automotive because it came a bit later it's just that we always focus on the battery cell because that's the most uh come on we are electrochemist it's the most amazing part of the story but even when we are going to change all the battery cells to another technology let's say sodium ion the grid storage industry already is going to use the same inverters the same battery management system etc that they developed using lithium ion which was developed by the automotive industry which was coming from from the consumer electronic industry so I think it's if we don't focus only on the cells it's already like something that they are using for free from another industry the grid storage industry and it's actually automotive one thank you Nicolo real back to you well we're having so much fun I'm afraid our time is almost up now ordinarily at the end of this seminar uh E or I will ask you to sort of give advice to students and upcoming scientists working in the area but I thought I would make a twist um you know over the next 10 years trillions will be invested in this area and uh you know some of those decision makers are in this seminar so I wonder if you each have a one minute um advice to give to the decision makers who are making big billion dollar bets into the future especially when it comes to this holistic level thinking maybe I should give a pause so you can form your answers yeah I mean I can go I can go first um I have let's say two um pieces of advice and um you know that would be to um you know one is it is well they're both related to being more deliberate about investment so I think oftentimes we follow intuition more than we follow the data and more than we follow careful mechanistic studies of technology development but as we've all experienced over the last two years there and and I would say the field of public health epidemiology and vaccine and and medical intervention development does this better than other areas of technology but you know there can be a very deliberate effort in rapid technological development but that means looking really under the hood of technologies looking at the mechanisms of improvement to also understanding how a single technology will impact a larger system and I think the one point I wanted to make more than anything else at my talk today is that we can be more deliberate in this area as well in energy storage in fact energy storage is a key candidate for this and I think this also applies to students so that's also a little message for students um and you know my other point is really about thinking about the right balance between diversification and concentration in your investment portfolios and I'll stop there um and move to Nicolo so he has some time as well thank you Jessica so for me there are two points so I would like to you know decision makers should should look not only in how cool is this a technology or what technology it's very important to keep an open mind of the next generation batteries etc there are so many different ideas that we are used to be a bit dispersed if we don't focus also on the economics so not start with okay I have this cool product what can I do with it it's often enough but am I answering an actual problem of the world with this technology and therefore is it worth worth exploring or at least keep up on a balance and open mind on those and the second one I would like to underline and I would like us all to focus on something which is also sustainable because it's great that you have a new technology that can bring down the cost and in the short term might be a solution but the old story what we are all actually here for in electrochemistry especially is to get to a decarbonization of our industry so if we have this new technology that is based on elements that are super rare and super difficult to to get out of the ground or they might actually be harmful maybe this is not the right answer you know even if it's not too expensive that we should also think about how much does it emits to to get this material to transform them and to make it into the batteries that we're going to use every day. Jessica Nicola thank you very much for this spirited discussion and your deep presentations I really very much appreciate it so this brings us to the end of the seminar and Evan if I can have the closing slides please we're going to have two more seminars this this quarter this winter quarter so please stay tuned for more information and register for the new events and with that I'd like to thanks everyone again for attending and joining us this morning at Stanford thank you very much and have a great weekend