 Okay, hi. Good morning and good afternoon and good evening around the world. This is E-Tray. Together with my co-director of Storage X, Professor Will Che, we would like to welcome you to our Storage X symposium again. And this time, the X equals to heat. In the past six months, you have been seeing we have outstanding speakers and lithium-ion batteries and new chemistry, high energy, and also in the larger scale production industry panel as well. Our Storage X, the mean of X, is expanding. So in today's symposium, we have X equals to heat with two experts to join us, Professor Bob Laughlin at Stanford University and Andrew Ponack, the CEO of Antora Energy Company. I will start by introducing Bob Laughlin first and my co-director, Professor Will Che, will introduce Andrew later. It's absolutely a great honor to introduce Bob. I joined in the Stanford faculty in 2005. Luckily, my office was in the same floor as Bob. Pretty soon after I joined in, after Bob came back from Stanford, to Stanford from Christ, as the Christ President, I have the fortune to the opportunity to interact with Bob. He is absolutely an intellectual thinker. He is curious about many, many research topics. Certainly, he has achieved a lot. He is a Nobel Prize winner in 1998 for his explanation of fractional quantum Hall effect. He has an understanding of many research topics. I remember my group meeting on the third floor of my color building, Bob. Maybe you still remember. Oftentimes, I have a group meeting, Bob walking in the hallway, picking through the window and seeing, wow, this topic is interesting. Bob will come into my group meeting and have a very active discussion with my whole research group. That was just fantastic to have your intellectual horsepower to participate in my group. Over the years, of course, Bob cares about what's going on in the world, particularly climate change. He also put his intellectual horsepower behind this problem and tried to come up solution for it. That is the reason we invite Bob today to speak about the topic he has been working on, the Britain batteries for the largest scale energy storage. With that, I would like to invite Bob onto the stage to start his presentation. Wow, what an introduction. Thank you. It's all true of course. Good morning, everybody. Wherever you are, thank you for tuning in to listen to me talk for a little bit. As Professor Trey said, the topic today is this, this image that I'm showing you here. It's a storage technology. There's a story behind it that I want to tell you. I like to start seminars from the answer. The answer is this. What this is a picture of is a closed cycle brake engine. See, there's a working fluid going around and around in it. This is a compressor turbine pair. These round things are counter flow heat exchangers. The tanks contain heat storage fluid. In the case of the hot one, it's molten nitrate salt. In the case of the cold one, the blue one, it's something else that we argue about, that for the time being is not important. The technology is electromechanical. That is to say, there is no connection from the grid to any electrochemistry. It's entirely mechanical. That's what I'm going to talk about today. Before I jump into my slides, let me tell everybody that I came to this particular approach the long way. My background is not in mechanical engineering. My background is in semiconductors and also actually in batteries. I have quite a lot of electrochemistry background from my time at Bell Labs. It was thinking about the limitations of semiconductors and electrochemistry generally that led me to this particular approach. The reasons, therefore, for looking at a machine like this as opposed to some other are important and require discussing. But I want to make clear it's not because I don't understand electrochemistry. It's because I do understand it. This is a formal publication that occurred in 2017 in Bloomberg. I'm showing this chiefly to emphasize that there are a lot of aspects of the technology that I can't talk about today. The flip side of trying to do something important is that other people's money gets involved. They then call the shots. I have NDA restrictions on what I can talk to you about today. But everything I am talking to you about today is clear. Let's start with the problem. This is a double graph. The bottom of this graph shows the total energy consumption of the world as a function of time. You see it steadily growing. You see that the coal, oil, and natural gas components are also steadily growing. This little yellow sliver here is renewable. We like to think, well, the bulk of that is wind. The wind, of course, is growing like crazy. But this is not so. One third of that yellow sliver is corn ethanol. Corn ethanol may or may not be renewable, depending on how you do the calculation with the diesel used in the agriculture. There is a market penetration problem with renewables that we all know about. There it is. At the top, you see the keeling curve, which is the CO2 concentration measured at the top of Mount Aloha as a function of time. What you see is that the steady increase in CO2 concentration has continued over the course of time, in particular, while this yellow sliver is growing. This is a problem that is out of control and that is not solved. Now, we're talking about electricity, not cars. We're talking about electric grid. This is a distillation of what the electricity problem is. This is a plot of the power used by the California ISO. All of California is a function of time for a particular day in 2009. It doesn't really matter what the day was. What you see is two curves here at the bottom. The gray one is a theory made two days before about what the load is going to be. And then the red one is experiment. That's the amount of power that the load actually demands. And you see it does pretty well, but the variation is enormous. It's a 50% variation. The green line at the top is standby power. So the green line is contracted to stay online but not turned on and to be ready to switch into the grid if there's a shortfall with the red line. Electricity companies like to say that they manufacture electricity when it's needed. And the reason they do that is there's no way to store it. The good news about electricity is it travels at the speed of light. The bad news is they're traveling at the speed of light and therefore you can't store it anywhere without creating explosion danger. Actually, you can't store it at all. The technologies for doing this have a cost problem. So this is the problem. There are other grid storage problems that occur on longer time scales. But the first one, the first one you have to deal with is this 24-hour surge. As a result of this problem, the price of electricity produced by the wind becomes negative in many parts of the world. Certainly happens here. It definitely happens in Northern China. It happens in Northern Europe. So the wind blows and it produces electricity but the electricity is not needed. So the wind producers are willing to sell it to the distributor for negative prices. What does that mean? It means they have to pay the grid to take the power and they still make profit because they have a subsidy. But it's obviously very bad business-wise and it's telling you there's a problem. In the case of China, which is a particularly important market, we know that the wind people have been ordered to stop making windmills because of this problem. A way to stash very large amounts of energy that works and that's known. You pump water uphill. There's a number of problems with this but the point is that in this country, the United States, the amount of so-called pump storage capacity for storing energy is about one quarter of what one needs for the whole country. So in principle, one could build the rest of it. This is very important because it's the first indication that the core of this problem is not technology is money. What's holding up the building of these facilities is a money problem. There is also danger. So when we're thinking about storing really large amounts of energy for cities, that amount of energy is dangerous. So it's a very different proposition from a car. It will burn up. It's so bad you lose the car. But if a big facility fails, it can destroy lives, a great number of them. So there's a danger problem and it's big. So a big metropolitan area, like I picked LA Metro, storing the power consumption of LA Metro for eight hours is the energy equivalent of a pretty large nuclear weapon. It's not the biggest but it's not the smallest either. So it's a largest nuclear weapon which is to say if you store this energy in a place where it can come out quickly, you have created an explosion danger big enough to destroy cities. So this amount of energy is really dangerous and way at the top of one's engineering considerations has to be safety. The failure of pumped hydro was facilitated in part by the development of this. This is an industrial gas turbine. It comes to us from the airplane industry. It is a fabulous technology and right now in the world this is the technology that levels the load. In those cases where you don't have hydropower to turn off and on, instead you turn off and on these extremely cheap gas turbines. They work wonderfully and they're reliable. So the power producers like them. So when we're talking about storage technology, it's important to keep in mind that this is the technology to beat. The batteries would be great but right now it's all happening with the what amount of jet engines. Now this is also critical to the problem. This is a plot of total, in California, total power versus time of day for a particular day. Actually it was April 2019 last year. Now let me call your attention to this great big yellow blub bubble here. This is the energy supply from the solar field to the Mojave Desert. You'll see over here that the big peak in demand is much later in the day and at present it's handled by some hydro chiefly imported from the Pacific Northwest but mainly turning on these natural gas turbines. Now so there's a time shift problem between when the sun shines and when the load is that could in principle be solved by a storage of technology but isn't. Now centrally important for battery people is this picture up here. This is also from the California ISO showing the battery flows in and out of the grid as a function of time over at the same day and you'll notice that the units are the same. So it's in multiples of gigawatts and it goes from minus point one to point one so this curve I can't plot it on the bottom here because it wouldn't show it would be too small. So what we're talking about here is kind of a hundred megawatts in and out maximum for periods of maybe 20 minutes. Now this is chiefly lithium ion and there's a message here which is this there is a business model to use batteries for stabilizing the phase of the grid that's what these 20 minute excursions are but there is not a business model to level that big solar load. So this curve tells you that batteries can interact with the grid very nicely but they have a cost problem that makes it not possible right now for them to make solar energy really work. So that means it all depends on cost and now when we're talking about cost this is not science folks this is science fiction and I apologize for this but now I'm showing you this picture chiefly to illustrate. Now first of all this red line is my technology why is it the best well it's my graph okay so I get to make I get to make the graph any way I want so I show the technology I prefer to be the best. That's not the important thing right now the important thing is that there are two metrics not one for for how it functions one one is the intercept down here this is what I call the cost per engine watt this is the amount that it costs you like for example a pump storage facility it's the cost of the dam and me and the and the generators and but if there's no lake you don't you don't store any power so the cost per watt as a function of of storage hours it's what plotted here one of the numbers is the intercept the other number is the slope and the slope of this line is denominated in cost per joule so it's a cost per watt of the intercept and a cost per joule now what you see is that at short time scales batteries win and that will just be forever so when we're doing things like load leveling of the phase the technology is already here there is no need to make new technology is out there deployed it's fine where are we getting the trouble is out around four hours and there's a crossing now the green line which are lithium you'll notice they're they're sloping upwards why is this well the way conventional batteries work is that the engine and the storage medium are the same thing so you can't they're the basically the electrode and so you cannot buy more storage without buying more engine and so that's why the slope is so steep you're basically buying more electrode when you make a bigger battery and that slope is eventually what kills you for a long long duration pump hydro by contrast the storage medium and engine are separate and and so once you've got the engine built then the more storage cost you nothing once just the slope there is so small it doesn't matter so these are kind of the two paradigms and right at the moment historically where the crossing point is is about four hours now i've got a flow battery line up here i don't have time to talk about this right now but let's let's move on so this is the basic idea that to make to get the price down is necessary for long duration is necessary to separate the the engine from the storage medium now that holding that thought i want you to travel with me to southern spain and you look at that solar field in the background there and you'll see those two little tanks those two little tanks are filled with molten nitrate salt this is a technology that's deployed already chiefly in concentrating solar plants and it's deployed a lot what happens is the solar heat heats molten salt there's these two tanks have molten salt in them at two different temperatures and you pump from the cold side to the hot side and heat with the solar field and then at night you do the other way around you go from the hot tank to the cold one and run a steam turbine now the important thing for us right now is that this technology is widely deployed so the first deployments were andesol one two and three this happened in 2008 they're 50 megawatts because the was a limit placed on the spanish by the spanish government but now you see the massive replication of the technology including many facilities in china now what this tells you is that the investors think think they can get their money back okay so this is the the first deployment is the technology now the technology exists the investors are willing to put their money down because they they're they're persuaded the technology works now when i was researching for all this i i got very depressed you know why why can't you do this for the wind because wind is a lot more wind deployed in the world than than solar energy and um then it hit me that you could and uh the reason is because energy storage is not about energy it's about entropy so your your task uh is is is to take the jewels and put them somewhere in a way that minimizes the amount of entropy that you create um now it turns out that modern jet engine technology is so good that it actually makes relatively little entropy so the the engines have tremendous power and they're quite small this is true also for steam technology by the way um so so actually these are pretty good in the entropy department so that means uh you can run it backwards so let's do a very quick um review of how a jet engine works there's a compressor here's a turbine the air goes in one side there's some fire at this point i usually ask the physics audience you know what what's the difference between the compressor and the turbine and they say they don't know because they don't it's very subtle actually the pressure in this in this uh chamber here is uniform but it's hotter at one end than the other and the work done is pdv so the the it's the volume expansion associated with the heat that makes the turbine different from the compressor otherwise they're just a they're just a couple of fans so um you can imagine now getting rid of the fire and putting in place of it uh what amounts to a locomotive boiler with some hot fluid let's say molten salt going one way and the uh working fluid of the turbine going the other and if you counter flow them as i've shown here then the temperature drop between uh the two sides one going one way one going the other is very small and this can be made as small as you want just by making the heat exchanger as big as you want so with proper amount of steel deployed here you can get rid of the entropy creation in the heat exchange uh now this leads to the picture that i showed you which has just exactly that the heat heat exchanger is replacing the fire and then we have a mirror image of that on the low pressure side uh basically does the same thing with cryogens there is a whole bunch of mechanical engineering cycle stuff uh that we could talk about offline um this is all published in the referee literature this cycle involves heat going in and heat going out this is actually a charge cycle all you do to switch from charge to discharge mode is that you run the movie backwards if it's not making any entropy then you can run the movie backwards and what used to be an engine is now a storage machine now um there are some other things here that matter this uh making these technologies with heat involves materials and it's sort of a quintessential material problem i'm running out of time here so i'll i'll skip this those of you who are interested please uh have a look at the the materials considerations particularly the considerations of the steels it turns out that the molten salts in question have a wonderful coincidence with what's called the creep cliff of modern steels and it's right where they where they begin not having a strength so this this is where supercritical steam plants work and this is exactly where the molten salt plant works also um let's skip this here those of you who are interested here are the equations for the turbines there's there are two numbers here called the adiabatic efficiencies per stage or so-called polytropic efficiencies the there's a little formula for what the round trip efficiency is for the entropy you're making which is easy to calculate and it turns out to be 0.7 actually 0.73 so the asymptotic efficiency of this machine is somewhere in the low 70s and uh that is enough uh to make money especially since the heat you're losing isn't lost it's actually discarded at a uh temperature that's useful for other things the engine in question is not new this uh was the first closed cycle engine uh gas turbine ever built it was built just before the outbreak of world war two in switzerland there are there's a history of plants that have been built chiefly for coal burning so the point important point is that the technology side the the turbo machinery side of the of the technology is exists and it exists uh with the corroboration of all the key numbers now when you're doing stuff that matters um you have to do this this is a whole bunch of patents and the uh you'll see i'm first author on all of them and sole author in the first one much of what's happened uh in this subject has been dictated by the history of the of these patents and uh this is a another subject to talk about but i don't want to get on to right now this patent stuff has to do with secrecy it's exactly the opposite of what we're supposed to be doing in university but in any way there you are um moving forward to do something about this problem you have to do get involved with people's uh other people's money so i meant at the end now the technology i put on the table is actually not different from underground case storing the observation is simply that when you compress air compress a gas the energy is stored in its heat not in the pressure the storing energy and pressure makes no sense so actually underground pressure storage has always has a heat transfer section at the at the top and what the technology is we're talking about here is you just get rid of the cavern and instead of doing one compression and do a compression followed by an expansion so there's no need for the cavern anymore no explosion danger because everything is in thermal equilibrium the stuff is hot but it can't ever explode so the nuclear explosion danger is gone no land use issues so deploying pomtigro has the problem that it uses land which people don't like and out here in western united states where we live the water is more valuable than the power uh i'm talking to battery people so let's just say i'm not a great fan of flow batteries because of this little ion exchange membrane problem in the middle but we're running out of time now so i can't talk about that the semiconductor issue is something i hope that will come up today uh the grid operators turn out hate semiconductors because the way the grid works is actually mechanical and they much prefer to have what they call rotating assets on the on their grid and i'm done thank you everybody for your patience and your forbearance and i think maybe i ran over five minutes and if i did i apologize well i think you are fine you are just uh about on time you're not running over so that's good okay good yeah uh so great well thank you for your great talk um let me start by asking you questions um looking at the brayton baff using uh the molten salt right what will still be the biggest one or two challenges you know in order to make this to work and deploy in larger scale what what are the critical things what are the barriers you still need to overcome um well since this is all happening right now yeah i know the i know the answer to this question and of course the the simple thing is money that's the answer is money but uh let's be more precise there is no technical impediment at all there's no scientific impediment there's no discovery you have to make the principles are all known versions of the technology are deployed okay it turns out it's just straight engineering that you have to pay for so this technology is very much like building a rocket it you need a whole bunch of things to work correctly and as you know very well since you've run a lab getting things to work is so hard but getting things everything to work at the same time is very hard and it requires teams of people who are talking to each other we're finding that the turbo machinery community is very risk averse and so for example steam turbine people who who have the technology for generation are very reluctant to run their machines backwards to make compressors even though engineers in the next building over are doing just that they're making compressors for gas turbines so that means some a lot of working with teams is necessary and a lot and some experiments and these experiments are are pricey and so um i guess the answer is to get the team working together costs money and so that money required to get the team working together is the key problem and there aren't any others yeah so bob let let me let me go down to a little bit detail right so uh you show a table with uh now many uh i was an experimental demonstration of using these type of technology similar molten salt right base um is there a lifetime consideration due to corrosion from the flowing um these uh molten liquid uh in and out of your system uh we all know molten liquid these ionic what it's called molten salt molten ionic liquid now is corrosive uh what the system lasts for certain number of years uh is it already well documented and demonstrated it's already documented and demonstrated the basic research for this uh technology was done actually in this country in the 90s uh the technology was first deployed here but it was deployed commercially first in spain so all of those problems had to be discharged before the investors would come in so the fact that it's deployed in the field tells you that those are non problems for a plant you need um you need 40 years and so there is some corrosion but it's not it's okay in 40 years now a little bit of physics here uh molten salts are actually less corrosive than you might think if there's no water so they're they're they attack metals like crazy if there's water but if they're hot they actually don't uh the major destruction of the metals comes from oxidation so nitrate salts have a lot of oxygen in them and they rust they rust the iron now it's stainless steel so they don't rust very fast um but that's chiefly it if you're dumb enough to put two different metals in contact with this stuff of course that's that's the end because they're electrolytes so you have to make sure them to be careful in materials but if you are then electrochemistry doesn't happen and there's no there's no uh debilitating problem yeah uh next question from uh from my audience um can you describe a little bit about you know now you have the heat right exactly the process of heat converting to electricity uh what's the process like is is there a thermoelectric thermoelectrical conversion system needed and i i think i think no no i rushed through that very fast yeah let me just answer in words um it's the same as a steam plant okay so uh in a steam plant this is why it's hard to talk to battery people so there's a different culture where you work with steam so most of the world's electricity is still generated by steam and what you do is you make high pressure gas which is steam and then you expand it through a turbine and it gets colder and you get work out that then spins mechanically spins a generator this mechanical spinning then connects to a highly efficient mechanical generator which then connects to the grid this uh so the conversion is mechanical now so that's the answer is it is it is mechanical now as far as the grid operators are concerned it's very important the mechanical nature is good because the grid itself is mechanical it's it's electric fields going back and forth at 60 hertz and the generators and loads are all connected with ropes that are made of light the tram the trap the forces travel speed of light so it's mechanical so when you have um dc uh storage media such as any electrochemistry you need some semiconductors to fake this mechanical this mechanical motion and they don't do it very well so the uh technology this particular technology goes back leverages the fact that the grid is mechanical and generates electricity mechanically it also takes electricity off the grid mechanically so the storage process is literally you take a movie of the thing generating and then you run the movie backwards yeah so bob um what's the uh experimental uh well i think this is for the benefit of audience like what's the conversion efficiency during this process the energy conversion efficiency by using okay uh the from heat to electricity is many many people ask this question and it's just a it's it's it's we have to get a blackboard that so one answer is 100 so no energy is lost now the better answer is that the um efficiency concept doesn't apply when you're using the same engine to go backwards so a typical thermal engine has some efficiency going forward well if you're if it doesn't make any entropy uh and you run it backwards the storage efficiency is one over the uh uh original generation efficiency so the round trip efficiency is one uh in other words the the traditional loss the loss issues actually cancel out when you're doing storage and generation uh with the same cycle now in reality the the cycle does generate entropy and so you have to slough off some of the energy is heat the answer is asymptotically it's 30 so asymptotically you lose you lose a third the actual amount you lose is greater than that because you have to make engineering compromises so it's probably in the in the high 60s now i say probably because the amount of this extra loss can be beaten down with money so if you if you add more capital cost you can get the round trip efficiency up so you have to make a business decision about what's the right what's the right uh round trip efficiency point and the current models say it's somewhere in the in the high 60s yeah okay uh but next question um uh from both audience for me so looking at the economics um what's the uh you know the upper limit the size of the system and down limit of system you know i think down limit the smallest size right system starting to make make sense economically both and thinking about what the power as well as thinking about the energy what hour uh and uh you know you show the table some started from 50 megawatt i remember right some like hundred in the table so and starting from the lower limit what started to make sense economically and thinking about uh the size of the system yeah it it turns out it's actually backwards yeah the big ones are better yes okay and so it's different from batteries the big ones the bigger the bigger you make them the more efficient the present target right now is 100 megawatt 100 megawatt in 100 megawatt out and it's because the turbo machinery really begins to be efficient when you get that size so you what you want is something that looks a lot like a steam turbine for for a commercial power plant it's just designed so it so it you know so there's a compressor on it which there normally isn't it turns out that making small ones is harder there's a simple reason why um for best efficiency you need for the blade tip to be moving at about macho 0.7 so that means if it's smaller it has to spin faster so if it's working at 60 Hertz which is the sort of nominal in this country the radius has to be about half a meter okay for it can't be any smaller than that because uh because uh it has to spin too fast and then you have frequency conversion and as you know there's no way to do that it's a physics problem so you can work it you can spin it slower than that as you do an esteem plant but not faster so the the small ones are really tough to do and that's interestingly one of the impediments because it's difficult to make a small prototype uh 100 megawatt though no problem uh and 100 megawatt look at steam so a hundred to a thousand so a hundred to a thousand megawatts is kind of the window the power scales up easily and then the storage time the marginal cost of the jewel is entirely the salt and it turns out you look at the numbers the price numbers the salt is so cheap that there is no advantage to substituting some other storage medium for timescales less than 20 hours uh there that's a that's a corner you don't have to cut so the the the salt actually you just make a bigger tank uh you want you want uh you move it up to a day fine just make a bigger tank yeah so Paul let me ask you one last question before we move on to next talk uh if you can provide uh you know a concise answer you know um so looking at lithium ion batteries or other battery technology let's just pick lithium ion the cost curve is coming down fast um and uh and the system level costs including the cell right the BMS and and and so on very likely uh will go down to well this is from the EV data certain the grid will be different uh down to about let's see a hundred dollars per kilowatt hour uh in that range right uh would that change the whole economics you're showing this plot right there you know uh that's by the assumption the kilowatt hour of the lithium ion battery I don't know what's the number you use right there yeah okay now this is a conversation I have this is of course the key question because lithium is marching because of other forces chiefly cars and it's the big that's by far the biggest market so so if uh lithium comes down enough it will solve this problem so that's that's basically the question is lithium going to come down enough I think it won't now you ask why not well it's just a guess you know it's uh I'm looking at how lithium ion batteries work and how the price has been squeezed out and I I just uh and I don't believe uh these are the optimistic uh statements about how the battery costs will come down what we know right now as I showed you a plot we know that prices today are not adequate the I had a second line on my graph which is the aggressive projections of that cost coming coming down and and you see it's it's it's you know it's getting close so if the lithium batteries in particular came down a factor two uh that changes the landscape just it just does so anyway that's that's the answer it's it's very close and it's not enough right now and that if uh every amazing thing you could project for lithium came true then it would barely make it into the kind of four to five hour range that you need uh but I have to guess about whether it's going to make it or not and my guess is no they're just parenthetically there's a there's another issue with electrochemistry which is maintenance and um the these batteries uh that do this if they're doing surge leveling they're working hard okay they're discharging and charging every day all the way and that's tough on the electrodes and so there's a downstream cost issue with with electrode failure that uh in this application that you really have to worry about you know Bob I think I agree with you it's lithium iron needs to work hard to this maintenance issue and different temperature climbing environment charging back and forth the cycle life degrades very fast particularly in the hot environment oftentimes where you have solar is so it's very hot right there so lithium iron well you know you know you don't run the batteries hot uh I don't think it's it's a thermal issue the batteries it's a matter of cycle just just just working the electrodes very hard at normal temperatures in cars as you know there's a weight penalty so when you're designing car batteries or actually for electronics too it's a paramount to get the get the weight uh minimized in a stationary application there there isn't I think what's going to happen historically is that the is that most engineers will be working on that weight problem because that's where the money is and therefore the the batteries will what that will get are will be optimized for weight as opposed to cycle rigidity and that's another reason I'm suspicious that the market just isn't going to is it going to provide exactly what you need for stationary storage okay well thank you Bob thank you so much for sharing with us uh with that if you can still stay online with us you know later we'll come back to a short panel discussion right uh I'll pass this to uh professor we'll try to introduce the next speaker thank you E and Bob thank you for a very insightful lecture um Andrew if I can also have you come to the stage please terrific um well two weeks ago um we had the great pleasure of hosting one of Stanford's distinguished alums uh JP Straubel to speak about scaling up challenges for lithium-ion batteries and today we have another outstanding alum uh Andrew Honak who is the founder and CEO of Antora Energy so just like yeah we'll tell a little story uh Stanford is a great place as we of course say and many of these encounters happen in the hallways of a building or sometime between buildings and uh I have the great pleasure of seeing some of the early development of Andrew's technology uh when we were walking between our thermodynamics class in um Encina Hall past the Hoover Tower and then back to the home of material science um in the Durand and McCullum buildings I think this was three years ago and we'll talk about all sort of things and you know mostly crazy idea so you know years later I was uh not surprised that the crazy idea turned out to be not so crazy after all and then Andrew and his colleagues and Torah brought it to life so it was an amazing journey to to watch uh Andrew and I should also add this is not Andrew's first stint uh he I think we almost lost him to an earlier stint uh dragonfly uh systems where he was developing uh power electronics for solar cell which is um um being an incredible success I think that was maybe six years ago Andrew right and fortunately we somehow was able to get him back to Stanford uh where he was just barely finishing and then found his uh co-conspirator for a Torah so in that context we welcome Andrew uh another uh terrific innovator and and and I'm truly a serial and entrepreneur and uh Andrew was so delighted to hear about your company and your new technology today wonderful okay well first of all well thank you for the very kind introduction and uh I must say I'm really honored to be here um you know it has been a really fun last three years since we were you know chatting about all these crazy ideas uh on the walks as you mentioned and and really uh just again an honor for me to be here with uh you know on a lineup that includes at some of the other talks people like JB Strable and Mateo Harmeo which have definitely been inspirations for me uh and our team as we uh pursue this so um yeah and Torah Energy was founded a couple years ago um you know about a year after some of those conversations uh Will and I were having uh after class um and uh really excited to share what we've been up to all right so um I'm gonna divide the the talk into three sections the first two will go relatively quickly because Professor Laughlin covered them pretty well you know why do you need uh long duration storage a little bit about why uh thermal energy storage maybe uniquely suited uh to this problem and then and Torah's approach and what are the advantages and disadvantages of the way we're trying to do thermal energy storage versus other approaches including Professor Laughlin's uh and and a number of others out there all right so uh again the need for long duration storage uh you know for me it really came out of my time at Sunpower which was the company that acquired my first company Dragon Dragonfly Systems and um you know one of the things is as a young engineer working at a big company like this you know we were very focused on driving down the cost um you know this was an example of a plant that was supposed to be sort of a slam dunk uh you know great solar resource very close to transmission had sort of everything going for it um but not only did Sunpower see everything was going for this site but a lot of other companies did too everybody installed solar at the same area pretty soon there was lots of curtailment negative electricity prices during the day um so it became very clear that storage was going to be necessary going forward and this wasn't an original idea that I had but it was certainly very personal for me you know working on the cost side but then seeing even the cheapest plants not be able to be competitive in the market uh without without storage and there are a number of ways to look at the storage problem uh Professor Laughlin uh you know showed one way and and there's a number of great academic analyses that go into why you need long duration storage this is just a really quick visual showing you know if you have um the uh uh you know solar plus wind in California uh over the last uh yeah I think this was in 2017 you can see it's kind of all over the place if you start filtering that even with about eight hours of storage you still have something that kind of can't be relied upon for uh the grid you know every day of the year if we're trying to go to deep decarbonization and sort of a hundred percent renewables um and it's only once you get about a hundred hours or more of storage that it's actually stable enough that you can rely on it you know sort of every day of the year so so this is this is the problem now a lot of people have been saying for a few years you know the long duration storage is needed it's been certainly talked about in in academic environments um you know it's kind of fun for us to start seeing the first uh you know big commercial procurements uh California Community Choice Aggregators put out a call recently for 500 megawatts of longer duration storage this is eight plus hours so we're starting to see kind of what was uh seen a little bit ahead as a necessary thing for the grid is start to show up in in procurements but very exciting for us obviously as a business uh we need to not just think that our stuff is needed but actually see the customers requesting that so um you know long duration storage is is um you know it's it's a a good word in some ways but in in other ways it's not so good when people talk about long duration storage there's nothing inherently good about a battery that can discharge slowly you can definitely take a lithium ion battery for instance and discharge it really slowly um and it'll be you know long duration um so really when when people talk about long duration storage what what they really mean is a cheap battery they mean storage that is so cheap that even if you're only cycling it a few times per year you can still pay back the capital cost of that battery and so that that's sort of reflected in this chart here if you look at the Department of Energy they put out this gray target zone that shows that the longer the duration of the battery the fewer cycles the battery goes through over the course of a year and the cheaper the battery has to be to be competitive and this is sort of the uh an analogous plot to the one professor Laughlin showed this is in dollars per kilowatt hour rather than dollars per kilowatt um so this is why rather than a slope and intercept you get kind of these asymptotes down um but it's it's the same idea and I would say you know uh we're maybe a little bit more bullish on on lithium ion batteries we you know lithium ion batteries are quite expensive now we we do think that lithium ion battery systems will get to the hundred dollar per kilowatt hour point which means that by the way the cells and packs are substantially below that in the future um and so we actually think lithium ion is it is it going to be a great solution for shorter durations maybe even up to about 10 hours but if you look at some of these extremely long durations like you know 50 to 100 hours um lithium ion even in the future we think is about an order of magnitude off from where it needs to be uh to be competitive and so this is really the the need that we've identified and I will just make a caveat though you know I've put this big green critical need up there it covers 20 to 100 hours these are very very different markets you you have a very very different use case for a 20 hour battery versus a 100 hour battery 100 hour battery may get used largely for resiliency to provide a few days of of cloudy weather or a few days of resiliency for a public safety power shut off which is a big market here in California due to the fires recently but the 20 hour battery might be cycled you know a lot more maybe not fully on a daily basis but but more often so you know for especially these longer durations and you might notice that I'm talking about a little bit longer durations even than Professor Laughlin was one thing that that's pretty clear is that you need extremely low cost for just the energy component of the system in fact we think you know for a lot of these longer durations you really need to be under $10 per kilowatt hour of incremental energy stored in the system now this is a really really tough thing to do and to give one example of why this is so tough is just to look at the the cost of the storage and this is a great plot from ARPA-E this is just showing just the cost of the container so you know let's just put up you know some target that you want your your stuff to be $10 per kilowatt hour or less and you know even if you came if you know if you're a researcher you come and you say hey I came up with the best ever thermal energy storage or flow battery or whatever and it uses a free material absolutely free it costs nothing but you say the only problem is it does need to be stored in in stainless steel if you just look at the cost of the stainless steel tank for your free material you have to if you're less than about 150 watt hours per liter which is a pretty decent energy density for a lot of longer duration storage you've used up your entire cost budget just on the tank so this is something that I think is you know again lots of reasons that it's hard to get to something like 10 but but this is one example of how you can't ignore other aspects of the system it can't just be that your you know electrolyte or your thermal energy storage medium is cheap you also have to have other things like a high energy density so the rest of your storage isn't too expensive and sorry the rest of your system isn't too expensive all right so thermal energy storage you know there's a lot of different types you know we think thermal energy storage is pretty uniquely suited to this for a few reasons and the first of which Professor Laughlin covered really well which is safety you know thermal energy is relatively difficult to discharge all at once and so when you are starting storing massive amounts of energy it's very helpful to have it be stored in in thermal energy rather than chemical or mechanical you know another thing that's really important about thermal storage is that you have very cheap raw materials you know a third thing is you can get very high energy densities and there's a range you know different technologies will have different energy densities but typically you can get up into those few hundred watt hours per liter which is really where you need to be and then you know one other thing that again Professor Laughlin mentioned is if if you're working with the right technologies you can have pretty minimal cycle-based degradation so you can get these long plant lifetimes and you really do see for these utility assets that you want to have decades of life and so it is helpful to have something that again theoretically there's a lot of ways this can go wrong and we can chat about them but you can have very minimal degradation. I also wanted to just reference a little bit Professor Laughlin had a very similar picture but this is one of these great salt storage tanks I think this one was about a gigawatt hour of storage so it's it's definitely possible to store huge amounts of energy with something like this. One area where we differ a little bit when we were starting out our journey thinking about energy storage and thermal energy storage when we talked to a lot of the folks in the concentrated solar world we actually heard a lot about the challenges of working with molten salt about the corrosion about the pumps and pipes and heat exchangers and you know the ability of this salt to freeze and cause problems and so we were actually pretty strongly advised that if you can help it try to stay away from you know a molten storage medium like that because it can can cause a lot of headaches. I think one thing that you've seen with that while there have been pretty substantial deployments of of thermal energy storage systems using salt it's actually trailed off quite a bit recently and there's a lot less interest in this area than there was a few years ago and simultaneously you see lithium ion batteries just taking off and lithium ion it either is already or soon will have surpassed all the energy storage in all of these molten salt tanks around the world so in the past this was definitely a much bigger thing than electrochemistry but lithium ion is really taking off and we expect that to continue. So if you want to store energy that you know they're in in the form of heat there are a number of different things that you can use again here's molten salt you can see relatively cheap under that sort of ten dollars per kilowatt hour for the salt itself does have a pretty low energy density so most of the time the salt plus the containment ends up at over ten dollars per kilowatt hour I should mention by the way these are dollars per kilowatt hour thermal whereas the the previous plot I used was electric so all of these get sort of twice as you know two x worse on cost if you have say a 50 percent efficient system or fresher lafflin was talking about high 60s so not quite two x worse so some of the other options by the way these are are liquids here in the blue solids here in the orange and you know concrete very cheap material can store a lot of energy there's a really interesting cluster of you know cheap materials carbon iron alumina that you can have pretty high volumetric energy densities and very low cost these are the ones that we focus on in particular carbon also kind of fun to note out note you know anyone looking to start with say oh these ones have really high energy density what are the what are these silicon and boron if you actually go through the phase change like freezing liquid silicon you can store a ton of energy in a very small space the problem with these is typically that the containment is actually extremely difficult the same freezing issues you know corrosion issues get a lot harder at very high temperatures so again we're focused on carbon and not only is it pretty good in energy density and in the in cost it has a lot of advantages as far as being very very temperature stable you can take it up to stupid high temperatures and it doesn't really do much and it's also very thermally conductive which for a lot of reasons is is useful all right so i'll just run into a bit of you know how entourage going about thermal energy storage how it differs from other approaches and and why we think it's the right way and but what are some of the reasons why it may not be so looking at the system we take electricity in from the grid we use it to resistively heat we just dump the electricity directly through carbon so really cheap storage medium that gets it hot we store that hot material inside an insulated shell to try to keep the heat leakage down until we're ready to use that stored thermal energy we actually get the the carbon hot enough that it's emitting light it's glowing mostly in the infrared but actually hot enough that you can also see it it's glowing in the visible as well so we have thermal radiation coming off this thermal energy storage medium and we actually use photovoltaic cells very simple conventional solar cells to convert that thermal radiation directly back into electricity so that's the part that's maybe a bit of a twist very surprising is to use photovoltaics rather than a conventional heat engine like a turbo machinery like a steam or a brain cycle turbine so it's still electricity and electricity out you know but at the end of the day the the energy stored is heat so why do we want why do we think this is the right way to do it the first thing is again we can use a very very cheap thermal energy storage medium carbon the feedstock here comes from petroleum coke which is made in massive massive quantities and is pretty nearly free also relatively high energy density which is important and then the other thing is the photovoltaics you know we can piggyback on a lot of the learning and the supply chain that's built up in various parts of the photovoltaics field in the last few decades rather than building all of that from the ground up so you know photovoltaics are are really transforming energy generation right now we're hopeful that we can do the same thing with storage okay so this was a very schematic we're going to go one one level deeper into at least a cartoon i'm just going to skip this cartoon of the system where you have these carbon blocks with slots between them so the carbon blocks are restoring the energy you're resistively running current through them to resistively heat them when you want to store the energy again it's within an insulated steel shell and then when you want that that back out just zooming in here you put something it looks very similar to a solar panel into one of these slots and that can be soaking up all the light that's coming off of the sides of these hot walls of carbon and turning that into electricity so uh zooming one more in on that process you have a hot carbon hot piece of carbon this thing is sitting up at uh you know over a thousand degrees celsius so it's it's glowing hot again when you want that energy back out you you insert this photovoltaic cell now a couple things happen that a hot object is emitting some high energy photons it's emitting a black body spectrum or gray body and emitting some high energy photons photovoltaics are very very good at converting those into electrons that's what they do all day under the sun that's what they like to do the problem is for that black body spectrum there's a lot of light that's coming in the mid or even far infrared so these are really low energy photons these photons are below the band gap of the photovoltaic cell and so they don't get absorbed so they can't be turned into electrons they don't get absorbed there it's a it's a lost mechanism one thing that we can do in this application though is put a very good infrared mirror behind the cell so this reflects all of that infrared light that's that's going through the photovoltaic would have been lost reflect it right back to that hot emitter so we're recycling all of this energy that would have otherwise been lost and this allows us to get to much higher efficiencies than standard photovoltaics i'll talk about that in a second but like any heat engine it's got a hot side and cold side you have to reject heat on the cold side so we use water cooling the power densities aren't so high that this is really problematic but it does probably require water cooling in some cases we may be able to do an air cold system all right so talking about the the efficiency which i just sort of teased might be higher than in solar here's kind of the the you know spectral way of looking at this this is where you know you have some spectrum of light coming from the sun if you have a single junction cell you choose one band gap you know the light that's above the band gap gets converted but you know if that's too much energy that energy kind of goes to waste it gets thermalized within the semiconductor and then you have another loss mechanism which is all of the light that's below the band gap and that's wasted and so if we look at the analogous picture for thermo photovoltaics using this method you still have you know light above the band gap that gets if as too much energy gets thermalized although in this case because it's only a little bit above the band gap usually because the whole spectrum is is sort of redshifted you actually get less thermalization loss than in a regular solar PV you do convert some of the light there into electricity and usually that huge loss mechanism which would have been much bigger if you were kind of doing it in the naive way and not trying to recycle any of it if you do recycle it you can actually get most of that energy back there will probably be some sliver here at the bottom that it's not a perfect reflector but overall that leads you to higher efficiency so a single junction solar cell there's never been one made over over 30 percent in fact the the fundamental limit is is 33 percent approximately the shocky quaisal limit you know whereas for a single junction thermo photovoltaic you can actually get over 50 percent and the theoretical maximum efficiencies are far higher than that so this is sort of the fundamental difference between solar and and thermal photovoltaics that we take advantage of to get a system with with very high efficiency so here's an example of what one of these cells looks like this cell was made about a year ago and we first started making these thermo photovoltaic cells this cell was made you know the semiconductor here was grown on basically a hand-built reactor at the national renewable energy laboratory you know very very non-commercial we were very excited because we were able to get pretty good performance but one of the things that has been very important to us throughout this process is making sure that any technologies we're putting into place are not ones that can only be done in sort of a lab setting and that can't be scaled up and commercialized and so we've actually taken the same basic design made some tweaks and but now we're actually producing these cells commercially with a with a commercial partner on standard equipment and we're actually able to get even higher efficiencies than we get at the at enrol so very exciting to have taken that first step toward commercialization now another thing you might notice about this picture is the cell is very small this is only about one centimeter on a side and this is another aspect of thermal photovoltaics that's that's pretty interesting this cell in our application can do about five to ten watts so that's far far higher than what it could do under the sun and the reason is that we're much closer to our thermal energy storage medium this carbon that's emitting the light we're much closer to that than a solar cell is to the sun so the amount of light that you're getting per area is far higher and you know the the approximate ratio there is is a few hundred to one so this is just a little visual example if you wanted a kilowatt of photovoltaics with standard solar PV have a few large modules basically you know one cell-sized one regular solar cell-sized object does about a kilowatt in our application this is also something that's very important for the economics of the system because it means that we can afford to put more money into the the the cell per area to get high performance and to have it be a little different than a solar PV cell while still having a very good cost per watt of the of the PV so this is just some some fun recent data and Torres demonstrated the highest efficiency of any thermal photovoltaic cell but actually beyond that it's the highest efficiency of any solid state heat engine that isn't using the the sun as its as its heat source so this is more efficient than thermal electrics or thermionics as well which are two other common solid state heat engines so you know we're pretty proud of this just for sort of the the science-based part of it the there's a peak efficiency you can see and it rolls off on both sides the the left side of that roll-off at lower temperatures is because there aren't enough photons above the band gap so the ratio of how many are useful photons versus bad photons gets worse the farther the lower temperature you go and on the right side you're actually getting so much power so much power per area that you're starting to run into resistive losses you're trying to push so much current through the cell that that you're you're losing some to heat just in in ice where at our losses so you know kind of to compare our current efficiencies are sitting somewhere between sort of a standard car internal combustion engine and the you know diesel engine say in a truck you know we do have some near-term improvements we think we can make that'll get us up to about 45 percent and that's putting us a little closer to very good heat engines like you know Toyota Prius which is a lot better than most cars a big steam turbine or even close to sort of the the best simple cycle single cycle heat engines which are big marine diesels all right um you know we're talking about lots of little things or you know these cells you know there's also the the big industrial side of this which is you know big blocks of carbon getting really hot and all that we're working on that as well i'm not going to go into too much of the details but it's been really fun to kind of move to that larger scale make sure that we're not missing anything and the rest of the system and then beyond that we're preparing for a much larger system which is actually a five megawatt hour customer-cited pilot that we're doing near near Fresno that's supported by the California Energy Commission okay so finishing up here just wanted to be very transparent about the pros and cons of this system so cons first because they're less fun tpv definitely has a higher technical risk than standard heat engine so there's a lot of r&d we need to do you know we have to take that on we can't go to a commercial supplier and say hey give me a really efficient tpv cell you know we have to do that work ourselves and that that takes time another disadvantage we're doing this this way is we have a lower efficiency just sort of fundamentally than a system that it uses the heat engine as a heat pump like professor loughlin mentioned um so this means that you know we're always going to be limited by the efficiency of the pv cell there aren't too many other loss mechanisms so we can get pretty close to the pv cell efficiency but even when we project very far into the future with what we think are are you know some of the best cells we can make we won't see much more than about 50 round trip efficiency on these systems so definitely lower than the sort of 60 to 70 that you can get with a heat pump system we haven't seen really strong evidence that that the market uh would prefer say a 60 or 70 percent solution to a cheaper 50 percent solution but it but it's certainly a disadvantage and that's a hypothesis that needs to be tested um and then also by storing the heat at a higher temperature you do have more heat leakage so you can solve that by adding more insulation but uh you know you're always going to have a little bit harder challenge there we typically design for about one percent per day heat leakage so if you store for a week you're maybe losing five percent or a little bit more of your of your power all right the pros um this one is really important which is that that by taking a solid state storage medium the solid carbon and a solid state heat engine um it vastly reduces the cost and complexity of these systems and this is really really important uh for for large-scale deployments and it's something we've heard sort of over and over again um you know from our customers that it's really uh you know important to keep things simple uh you know a good example of this would be turbo machinery based heat engines like regular power plants often have uh large numbers of people that are there for operations and and maintenance a lot of costs they are hopefully with a simpler system we can get down to to very minimal uh o and m comparatively another advantage is that we have very high energy density you know by by using carbon which is one of the best um materials for for energy density um and using a relatively wide temperature window because we're at these very high temperatures we can get energy densities per volume that are getting comparable to lithium ion at least at the sort of pack level um not quite at the at the cell level and then uh you know this is this is one that we've really uh I think uh come to appreciate more and more is that you you have faster design cycles so we're able to uh run a new um you know a new tpv cell very very quickly make improvements improve the efficiency improve other aspects of its performance whereas you know if you're trying to build a big you know steam turbine or gas turbine or something like that um it gets very challenging it often takes a very long time to go through a cycle of learning um one other thing uh about the system that we really like is fast charging so as I mentioned at the beginning there's nothing inherently good about a battery being slow in fact you'd rather have it be really cheap and fast um in this case we're only kind of halfway there we can charge it really fast because the resistive heating really isn't limited to anything other than how big a transformer and an interconnection you have so even though we discharge a little slower um and that's because you know you don't want to pay for too many of these photovoltaic cells you can charge it on the other hand very very fast dump electricity from you know otherwise curtailed wind and solar into the device and then uh the final thing here is that it's it's very scalable and modular and this also is a little bit related to those faster design cycles um but the modular aspect means we can test on small scales and have a very strong idea of how that performance is going to be on a larger scale because it's just the same unit repeated over and over there isn't the same uh scaling factors for example that professor Laughlin was talking about with turbo machinery and then also on the scalable side you know uh we've seen exactly the same thing that professor Laughlin mentioned about you know you need these turbo machinery units to be quite big you know 100 megawatts in many cases you know you might be able to get down into the tens of megawatts but it's not going to work at a much smaller scale and we've seen that their first there are some aspects of the market that uh we can address by going down to say the megawatt level that a turbo machinery solution couldn't but then finally this is actually just a money question it's a lot easier to sell your first units when they cost a million dollars versus a hundred million dollars and so it really reduces the finance risk and sort of the roll out of this technology to have you know bite sized pieces for the system okay so those are uh pros and pros and cons of the system um you know all done here uh I just really want to thank again the the organizers here for inviting me and also a shameless plug uh we're always hiring and uh for those who are might be Stanford students we do have an internship position open now we have other positions that will open in the future uh so uh regardless of your background please send us a note at hiring it into our dot energy love to hear from you okay Andrew thank you very much for the talk both the technical side and business side we really appreciate it so time for questions um let me maybe start off with one of my own question and this is I think just due to time limitation you focus the talk mostly on your power module the the thermal PV um can you talk about the challenges on the energy side um so you mentioned the heat leak it's a problem another thing I could think about is the high temperature storage would also effectively increase the containment cost as well um what are some of the challenges there what are you working on yeah yeah great question um so a couple of things there I think the first thing is um one thing that we like about the the system side is that there are a number of high temperature industrial processes that look relatively similar to this use similar temperatures um there's even a technology or a manufacturing method for for graphite we have these big graphitization furnaces looks very very similar to what we have you run current directly through the graphite you pack it with a bunch of insulation you operated it even higher temperatures than we're operating at so there's some sort of industrial analogs that gives us a lot of confidence uh that that uh this is not going to be the the hardest part of the system one thing that I didn't mention is we do um you know that steel shell that I mentioned has the insulation inside of it so the steel shell sort of the containment doesn't go to a very high temperature that's something that's really nice uh when you have a solid storage medium you know it doesn't sort of fill all the nooks and crannies so it's really easy to just you know have the solid storage pack insulation around it pack a shell around it that shell can stay cool that really helps the containment cost we don't need any fancy materials there but we do use an inert gas so uh you know carbon at these temperatures will oxidize it actually won't burn they use graphite electrodes in the steel industry in air they slowly corrode in the oxygen but they don't burn but we do have to keep an inert atmosphere so that we have the the long life there so that's another aspect of the system that's you know needs to be thought about thank you um just to follow up a little bit more on that point so in your in your cost analysis plot of the container you know you show even stainless steel is quite prohibitive but am I correct to understand that where you could win is on the x-axis you're moving high energy density so you're still using um somewhat expensive containment but just much less of it yeah we both think that we can use we don't have to use stainless steel for instance because it's a cooler temperature there's no corrosive stuff in there and then also yeah higher energy density those two combined can bring it down from you know the kind of the random place I put which was you know $10 per kilowatt hour down to half of that or less for the containment which is really important for the economics terrific um let me move into more of a material science aspect of the tpv um so you know here at stanford and many other places um spectral selectivity has really taken off as um not only for heating but also for cooling uh you know he is an expert in this area um I know that this is not something that the tpv has really focused on and I think you're here you're using a mirror basic at the back iron mirror to help with it are there any opportunities for spectral selectivity on the front side of the cell to really help balance the emission versus absorption um any opportunity there yes absolutely and and this has been looked at in in the rather small tpv community for a while there are really three places people look at at selectivity one is make a special emitter the the hot object that emits only the photons you want as you mentioned put a filter between the emitter and the cell could be on the front of the cell that reflects back whatever photons you don't want or as we do put the mirror on the back let them come through the the cell and reflect them all the way back out um just on the the first one the the selective emitter what we found is that um the selective emitters we've seen don't perform all that well compared to the other methods and there's also a degradation and cost problem basically you know in our view you want to keep whatever's hot in the system absolutely as simple as possible so for us big dumb carbon blocks that never move and just sit there their whole life that's about the only thing you want to be hot in the system um you know the the filters we think have more promise than the the selective emitters um but still have have some challenges and what we found is that we can actually create a better reflector using the the semiconductor itself as sort of a filter where it absorbs the above band gap light and lets the rest through and then sort of a non-selective mirror behind that the commonity combination of those two uh can perform better than the the front mirror you can by the way also combine these different methods of selectivity it's possible we'll do that in the future but we don't think we need to right now terrific um Andrew so um let me ask um I get a question from the audience uh so Jeff McConaughey from Stanford asks um I think this is inspired by your plot on the efficiency and comparing that to various uh internal combustion engines for cars um is there an opportunity to use TBV as a way to convert the thermal energy from combusting fuel to electricity say for mobile applications or stationary so now for storage so just one way to go from chemical to heat to electricity uh rather than say chemical to heat to mechanical for propelling a car any opportunities there because I think the efficiency comparison was quite striking uh even yeah yeah yeah absolutely it's a great question um so we we have looked at that in a couple of applications this isn't our main focus um you know a good example of where this might uh work so in cars we think lithium-ion batteries are are going to work there's there's no reason and we don't really want people to be burning fuels in in cars anyway so that's not something we're so interested in pursuing but absolutely the general idea of of combustion and using the TBV for that is interesting a couple of places we have looked at one is something like biogas so um you know biogas coming off of you know wastewater treatment plants anaerobic digesters taking like cow manure and these sorts of things it's often really nasty gas and you have to do a lot of work to clean it up if you want to put it into a pipeline as renewable natural gas or if you want to burn it in a turbine engine and and uh also you typically have these really small sources of this gas and so a good a good place that we would be interested at some point in the future perhaps in looking at is you know can you have uh a small source of renewable natural gas combust that use the tpv to convert that into energy because it's sort of an external combustion engine uh it can be very tolerant to whatever impurities or low btu content of the gas all of those other things so absolutely there's some really interesting applications of this outside of of just storage right um so related to this question um is there an optimal scale for the tpv in terms of size um you know how big does it have to be and how does the scale um as you make it bigger yeah yeah so the tpv itself um scales uh pretty or it's pretty independent of scale so that the performance should be about the same you know the tpv cells themselves we need to keep relatively small because they produce so much power per area it's just gets unwieldy if you have you know a kilowatt or 10 kilowatts coming off of a single cell so we keep the cells pretty small but once you tile them into a module it doesn't matter if that module's you know that big or that big or you know really really big um in order to do that on the system side though I will mention you know uh you know the power conversion side to scale independent the system side you do get lower heat leakage the bigger you make this system um that's just surface area to volume ratio so you probably don't want to go below tens of megawatt hours of storage so let's say it's a 50 hour system one megawatt 50 megawatt hours that would be sort of a small system for us um you know very small compared to say a turbo machinery based solution but still not so much like a household scale or something like that terrific maybe I can make one observation maybe this could be a seed for the panel discussion in a minute or two um I think injury your technology really strikes me as um really fulfilling the energy to power ratio conundrum for long duration storage so your storage um has to have scale right the energy has to have scale so it has to be large to minimize heat loss but your power stack actually works very well as small scale just like electrochemistry yes so I think in that sense is somewhat similar to a flow battery where you have a very nice decoupling but also I think um the the size dependence is also different so allowing e2p scale really nicely uh and I think maybe this is a question that Bob can also talk about is sort of the e2p aspect um for using the breakdown cycle for example as the power module which I think will be has to be larger in size so it seems like there could be nice complementarity between the two technologies in terms of the size of deployment so with that Andrew thank you so much for sharing your thoughts again and if I could ask everybody to rejoin I think we have about 15-20 minutes for hopefully another spirited discussion E and Bob will come back well Andrew great talk yeah I have to do all these zoomy things here okay I'm I'm I'm I'm I'm resumed so to speak so so uh maybe well we can do it in this way Bob did you hear uh Will's questions about this e2p uh ratio and uh and any uh I did yeah you want to are we are we paneling now is this is a panel discussion thing that's right a spirited panel discussion you know um once we start talking about this and things like it uh we're we're this is business model issues okay and um so I have to have dollars in the sentence for me to pay attention the um that energy to power ratio is an engineering concept what I need to know is what's the dollars per store jewel and the dollars per watt of the engine and um the the marketplace at least that I see is is going to be just vicious on those things right now it's for the battery dominated parts of the market it's vicious and cost per watt okay uh and that's why batteries are are are doing well they have a low they have a low cost per watt but they have this this energy to power problem that gets worse and worse and worse as you get the longer duration um so uh so nothing so well I hope you could refocus on is not what that ratio is as an engineering concept but what the dollar ratio is as a business concept um uh well and you do you have some comment on that uh on this question I do have question following question related to this I want to give you the chance to say it first oh yeah I was just going to mention I think that's that's exactly the right way to think about it you know you have to think about dollars per watt or dollars per kilowatt and then the dollars per kilowatt hour the dollars per energy and that that's one area you know we're often looking at at longer durations because we think we have a bigger advantage in those areas because we have a very very low per energy cost you know lower materials and containment costs than molten salt and so that kind of pushes us toward uh you know having a bigger delta at longer durations um I you know from a business perspective also we see there we think there's going to be a lot of competition again maybe I'm a little bit more bullish than you professor Laughlin about the lithium ion world we think there's going to be a lot of competition anywhere in the kind of 10 hours or less category uh from lithium ion batteries and what we really wanted to make sure we didn't do was end up with a solution where our selling point was hey we're going to be you know 20% 30% cheaper than a lithium ion battery and there's no way that a reasonable uh financier or utility is going to want uh to go for a really untested technology over something that now is quite tried and true like lithium ion installations if there's a small cost delta so that has also pushed us to say you know what are the areas kind of far away from lithium ion batteries where there's a big hole in the marketplace yeah can I weigh in on this please that's correct um the uh the lithium situation is that is the elephant in the room and uh how one approaches the problem depends critically on how one measures that now um and you know you can hurry their way for for a business perspective you want to find a market that you can get and that isn't going to be taken by somebody else so I'm that's totally right the difficulty we face is the first market that matters is the 8 to 10 hour one that's the first market that matters that hasn't been nipped off by by lithium so that you know you can go to longer times which I'm interested in as well uh even seasonal but the problem is the immediate cash is in this shorter time scale and so what you're betting you know one way is whether the lithium people are going to nip off that piece or not uh I'm aware of the um uh you're right about the business problem that's one of the difficulties I faced uh and okay it's thought about this now so the the assessment that the lithium people are going to have trouble is personal and like all personal assessments you know you put your money on the line and see what happens I uh so we'll see yeah so I have a following question kind of related but slightly different um so looking at e to p ratio and also dollar coming in right I I let's look into more slightly more complex situation um so when you talk about a long duration when you think about more than 10 hours maybe five days right so that situation this is the only thing about one situation but in reality uh the when you install something in the grid scale storage in that station a really big one you actually use that for multipurpose and long duration at the same time day to day this operation power in and out you know hour to hour so and then you look at the characteristic of the technology characteristic uh both of you work on a different one you know sharing some common rarity right I every time I think about lithium you know lithium can have very low power cost if lithium you are running it doing fastest charge let's say one hour or you know 30 minutes actually the power cost goes down but once you store more energy you need to have the energy right there you destroy a long duration that's the power rating the cost goes uh will increase however under the situation sometimes you you do need to you know do that hour 30 minutes charging this charging and then sometimes you can do you know five to ten hours under this complex situation now what's your thought about uh the comparison like what you're working on with lithium and then thinking about your technology characteristics how that changed the whole cost argument the dollar who's first yeah you go first Andrew sure I'll I'll take a crack at this so um yeah I think this is a really interesting uh area you know there's there's right now value stacking you know trying to do you know different things with the same battery as we're talking about is is pretty much necessary for most of these energy storage installations and and of course there's a there's a whole spectrum of durations of energy storage you know from balancing really quick fluctuations on the grid to you know longer you know 15 minute one hour you know and then there's a long duration you know we um when we look at the modeling we really think that there's shorter there's going to be enough lithium ion on the grid that all of those shorter duration things that you would do trying to balance you know with you know every 15 minutes or something like that there's going to be enough lithium ion that it's going to soak up all the potential in that area so in our models we typically don't imagine that we're going to be value stacking those short duration services on top of the longer duration services when when we look at um you know the different areas that we might be addressing it would be more like we're doing some sort of diurnal cycling you know near the top of our our thermal SOC and then every once in a while you know maybe doing a few days worth of discharge because there's low solar and then you know back up again and then and then that whereas we're not going to be I think dealing with the really really small fluctuations because lithium ion does a good job of that. Bob? Yeah I concur completely uh uh the graph that I showed demonstrates that this battle is over that lithium has already won it and I say lithium as opposed to other electric chemistry it's so far ahead in the in lowering the production costs um that I don't think anybody else can catch up to it and so that's that's a done deal the uh the phase-leveling timescales of 20 minutes to half hour in there that's going to be lithium I think forever and this by the way includes all sorts of quick response issues like for example you're running a company and the grid goes out okay and you need your backup to come in real fast so that to not wreck all the memory of your computers that's a job for batteries and so I can foresee uh the batteries taking over that and then if you have a problem longer longer than 30 minutes or an hour then you bring online something else that won't respond quite so quickly uh but lasts what lasts longer and is cheaper this by the way changes also the engineering mandate for alternatives so when you're if there's no lithium you have to design a machine that will respond fast on the other hand in a world where lithium has taken care of that problem uh your design problem uh is easier because the machine you have to has has to be steady has to have timescales that are typical of a modern gas turbine stay 20 minutes okay uh and the quick response issues are now not a non problem yeah thanks Bob will pass back to you thank you um so Bob and Inju I since we're talking about dollar I thought I would go there even a little bit further um so this is a point very obvious at a both of you but just for all of our audience listeners as well the the value of long duration or longer duration storage really also depends on the level of decarbonization we desire in the electrical grid and obviously this is highly tied to policy the cost of carbon and so forth um at the current deployment right very low um a sort of modest penetration of variable generation on the grid um what is the can you help us understand what is the market size the potential revenue from say you know 100 hour storage level versus sort of the uh 30 minute frequency regulation to something more intermediate do you have a good understanding of how much money there is um today at the current level or put another way what levels do you think that your technology will will require in terms of um variable generation penetration on the grid in order for longer duration to really kick in and create the market um or is it already here today again I'll defer to you Andrew first sure sure yeah so um it's a good question and as a business you know we're obviously very uh aware of this you know we don't you know it's very difficult if we're creating a technology that isn't going to be used for 10 years that would be very tough position to be in fortunately a couple of things that we've seen one is that even while you're at a certain penetration level on the grid as a whole um you get much more uh extreme localized effects and Professor Laughlin mentioned this in a few locations around the world where you know locally on the grid you know in parts of Texas or parts of China you have so much renewables that you really have hit that nearly 100 percent renewables land where where you really need this long duration storage and so it's sort of a slow roll on from you know kind of isolated geographies where you have specific conditions toward the grid as a whole over time and if it weren't for that I would honestly say it'd be very very challenging to commercialize uh any long duration storage if we had to wait for you know 10 years when everything was going to flip and then suddenly we need it so so it rolls on in different places at different times there's also a number of interesting um you know aspects like I mentioned uh the public safety power shafts things that are requiring long duration storage for reasons other than uh just a a you know renewables penetration getting so high and those are also stepping stone markets that I think you know and Torah but but also many of the companies in this area are going to be looking at uh in the meantime these are pretty big markets you know hundreds of millions or billion dollar markets for resiliency but it's not the you know hundreds of billions of dollar market of the grid as a whole um okay my turn uh a little more bullish um but you know along again sort of along the same line but a little more bullish first of all I'll point out to everybody that in this country it's a non-starter while without regulation because no one can beat natural gas in a fair fight so in places where natural gas is extremely plentiful as is the case here in north america isn't it's very political this question so in california at the moment for example there is a great um desire to decarbonize completely whether that's going to happen or not we'll see but uh if the laws weren't there it would definitely not happen because this is a natural gas state and natural gas is cheaper and better than these other things now we can refocus the question though places like germany or most of china okay uh that are not natural gas rich and in fact germans are importing a stupendous amount of natural gas right now through the north spring pipeline and about to double it with north spring too uh so this is good for decarbonizing germany but it's a national security issue because you're importing all your energy this is also the case uh in china so the the there's there are political reasons international political reasons to cut back on coal might be hard to do and the natural gas assets aren't there at the moment so natural gas is imported from russia now in those those markets are very different from the ones uh that we have here uh in europe uh i think uh the the need is strong enough that the market would make itself in other words uh the utilities right now have nothing to buy so so they're stuck you know they have to use water and import natural gas but if you gave them something to buy they would do a business calculus and think you know do i need to import all that gas and if your answer is we can we can stop it um then they i think they would so i think the northern europe market is very strong it's a little harder to call in china because of coal politics in china which is you know like coal politics everywhere it's just more because there's more coal used in china i think the answer is that it's like europe i think the answer is there's enough pressure and also internal desire to cut down on coal um that the technology uh would would make its own market all right anyways so it's position dependent the answer yeah well you have something you want to follow on yeah if i i don't mind but if you know my e um bob let me you you brought up a very important point which is um it leads to this issue of co2 footprint as well something that um we didn't discuss today uh and this is really where lithium-ion battery on a lifecycle basis the co2 footprint isn't small it's it's it's pretty large um you know andrew i couldn't help to notice that you're using a lot of carbon so maybe this is can be viewed as carbon negative in in some sense um but of course it takes carbon to manufacture all the parts a co2 to make all the parts too um have you thought far enough in terms of what the co2 footprint might be could the thermal technology be more competitive on the co2 footprint than lithium-ion battery that could have also other impact on sustainability as well yeah okay go ahead just just really briefly um certainly i would i would say we you know we're excited about the fact that we think we have a lower co2 footprint in manufacture but we don't think that'll actually be a big competitive aspect of the technology i don't our personal view is that the the co2 footprint of current lithium-ion batteries is not so high that that is going to prevent deployment we think the the economics are going to drive that you know much faster than any hesitation about the the co2 um you know as you mentioned i don't know if it's it's carbon negative but uh you know by certain definitions but we certainly do like the fact most of this pet coke right now that we'd be using to make our our batteries gets burned in coal-fired power plants um and so it's it's pretty nice to be diverting some of that stream that would have otherwise gone into the atmosphere um one fun thing that i know i've spoken with a few people recently about is um you know there are other technologies like methane pyrolysis that might produce vast quantities of carbon that we don't really know what to do with yet um and so we're also excited to provide a sink for for some of those things okay uh i think i think i'm on board with this one too um the carbon footprint of manufacture is something that concerns the precourt institute a lot and i think that concern is incredibly misplaced uh this may be something you worry about two centuries from now but right now it's not the problem uh there's a series of more urgent problems that the wind cannot penetrate more than 30 without a new invention and that's the problem to solve right now and the uh the later the more cosmic things will sort of solve later um so um in physics you know we have the from book on poly you know there's these three levels of significance you know there's right wrong and not even wrong this carbon footprint thing is not even wrong very very interesting answer i i i want to switch gear a little bit and by asking both of you this question um when i compare looking at the lithium ion batteries history right and uh when you know the battery was developed in 1970 starting that lithium matter and so on it was about energy crisis it turned out to be you know 40 years later close to 50 years later right so uh it's uh it's also for fighting this purpose but in the you're doing this time period about 50 years um lithium ion happened through a history of power and consumer electronics first in the market and then these about 15 years you know of history on that and then going to electric car and the cost going down performance cycle life is going better you know things gets better and better safety gets better right the electric car now is dominating by lithium ion now lithium ion coming into the grid so it has this market sector one by one just keep coming up you know with the technology growing the cost and making sense just this sequence right there my question now come to both of you for the thermal base storage would you be able to imagine some other market application sectors that's not entirely stationally connected with solar with wind yet allows you to explore first and before you go to the scale of doing solar wind right away and whether that makes sense or not i i don't know so just brainstorming you know just other applications you could also leverage to build up the manufacturing of your technology to reduce the cost to get everything you know behind you to support these uh these these type of technologies you're on and all right perfect this is this is a very good question this is something we think a lot about because i think you see this in almost every technology adoption you know you have these these stepping stones of smaller markets maybe that are less price sensitive to larger ones that are a little bit or better yeah that are more price sensitive to really large ones that you have to be very very low cost and there aren't a lot of examples of technologies that leapfrog all of those and get right to the end you know massive scale and very low cost you know i certainly i saw that also in solar exactly the same thing as batteries where you have you know you put them on buoys and then you put them on you know rich people's houses and then you start installing them in the desert you know big changes in scale um and uh really uh you know we a couple of thoughts here first of all we we hope we can piggyback a little bit again on the photovoltaics journey you know this is something that has been developed for decades you know most we're trying to combine you know photovoltaics technology that is more developed and the sort of more standard industrial high temperature storage like i mentioned you know that a graphitization furnace so that hopefully we can leapfrog some of those early ones by uh taking advantage of other industries but even within that uh you know we absolutely are always looking for markets that are those smaller earlier markets might have something special about them that allows us to operate in them first before we get to the really large eventual markets i mentioned uh psps markets and i think this is actually something also that professor laughlin alluded to if i'm not mistaken was you know what is the waste heat coming out of these these systems you know that's another thing that we're looking at one of our potential early customers is very interested in this for combining power applications so getting the electricity out of the the storage but then also using the heat for agricultural processing in in in this case um and so yeah absolutely we're looking for these areas that are stepping stones to the larger just like lithium ion batteries had more consumer electronics okay yeah um that's right right that the history shows you need little steps now uh man i thought about this endlessly you know endless hours thinking about this and the conclusion i came to was that this problem is why the technology doesn't exist yet in other words the blockages to making little steps are fundamental and that is why they haven't happened and the energy industry as it exists today kind of counts on that the the uh the entry barrier to actually nailing the problem uh is very high one of the things i hope that came out in my own lectures that if you wanted to solve it right now and they and the feds were to decide they want to do it boom six months okay because all the parts of how to do it are actually necessary so the constraints all come from costs and in particular this this issue of doing it incrementally uh so you know it's going to work out the way it works out it might turn out that you can do it incrementally or it might turn out that this is more like building a rocket that you just have to build one and you need a big enough money source to actually do it to get over the hump now um so history is going to now work out and we'll see we'll see what happens um one thing for sure this this question is fundamental it's a fundamental question is this problem like solar cells or is it this problem like rockets thank you for sharing yeah yeah what a what an inspirational note on which we can end our session i was looking forward to a spirited discussion i think we had one thank you both again for taking your early morning today to talk to us e and i and also the rest joining from across the world as being a very much a learning experience best of luck to the both of you for scaling up the technologies and very excited to see where it goes next um so if i could have the closing slide please justin thank you very much um so as he mentioned we are on this session of uh this several sessions on understanding x beyond just batteries but we'll actually take a a brief return to chemistry of batteries and so in two weeks we will have our very own professors in imbal and eric coaxman from the university of maryland to talk about chemistry and then three weeks later right after thinsgiving we will have our final talk for the year in which we will come back to the topic of long-duration storage uh so with that i would like to thank you all for participating and andrew and bob thanks again for taking the time it's been a great thank you thank you paul and andrew have a good morning thank you