 Hello, everyone, and welcome back to our regular webinars here at IBM Research, where we share with the world with you guys watching us right now all our latest milestones, breakthroughs, discoveries. I'm Katja Moskowicz, the editorial lead of IBM Research, and as usual, I'm here at IBM Research Lab in Zurich, Switzerland. The topic today is quite exciting, and also I would say quite important for many of us. It's energy. I don't think there are many people out there who don't care about energy, about energy storage, energy production, the cost of energy, and so on. I mean, think about it. If you're a kid, of course, you would care about your, I don't know, remotely controlled car to last longer. It's like you don't want to just keep running to your parents begging for new batteries. And if you're an adult, then you probably, to some, like many of us, care about sustainability, using renewable energy, which means better energy storage solutions, of course. And we all care about the cost of electricity, cost of energy. So energy for many of us is quite a crucial, crucial topic right now. And this is exactly where quantum computing, this super awesome next generation technology inspired by nature can really, really help. And of course, next to me, I have a model of a quantum computer. And I hope that at least some of you have watched our previous webinars on quantum computing, where we explain quite in detail what quantum computers actually do. If you haven't, then please do go back to our earlier webinars, especially the one I did in February with my colleague here in Zurich, our quantum lead, Hi Keryl. She's amazing and she explains really, really well and in great detail how this amazing machine actually works. So do go back and watch, but not now, of course. Do keep watching us right now for another hour. And we also going to be talking about AI, artificial intelligence. So how quantum and AI can help us build the next generation battery of the future. Before we go to our two presenters here, though, let me remind all our viewers to please send us your questions in the YouTube chat, because behind the scenes we've got an amazing team of scientists who will be answering your questions in real time. A lot of the questions we will also answer live, of course. They will be sent to me and I will be then asking our presenters. So without further ado, let me introduce our amazing presenters today, all the way from California, joining us Gavin Jones, a chemist who uses quantum computers to run molecular simulations. And we will talk about what exactly that means more in detail during the show. And Gavin will tell you all about it. And here in Zurich, joining us, computational chemist Teoleno, who has been looking into creating better batteries for electric cars for a few years now. Today, he's actually working on accelerating the discovery of new materials, including materials for energy applications. Welcome, guys. Thank you for joining us here today. Ciao, ciao, Raja, and good afternoon, good morning, good evening to everybody. Hello, everyone. Thanks for joining in. So the first question from me is for you, Gavin. And again, just to encourage our viewers here to start sending questions in the chat. But the first question is going to be from me. Please tell us what is actually this connection between quantum computing and batteries in a nutshell? Like if you can just explain what quantum specifically, how quantum specifically could help with battery applications? Sure. Well, I'll talk briefly about quantum computing since there may be viewers who are not familiar with quantum computing. Then I'll talk about quantum computing and chemistry, and then specifically batteries. So quantum computing is a new form of computing in which itself using bits comprising zero or one as a basic element of computation. We use qubits, which comprise superpositions of zero and one, meaning that it's a combination of zero and one at the same time. And the significance of this is that you can potentially represent more states of the quantum computer than you can with a classical computer. One other useful property of quantum computing besides superposition is entanglement in which qubits interact in a way such that the quantum state of each qubit cannot be described independently of the others, even when separated by a large distance. Which means that whatever is done to one qubit happens to be all the qubit. So this is one of the fundamental differences between a classical computer like your laptop or supercomputer, for example, and a quantum computer. We also use quantum logic gates in quantum computing are similar to classical logic gates like the and or not gates. Some examples are the poly-x and c-not gate, or represent the unitary operations of these quantum gates on qubits as rotations around a block sphere. And combinations of quantum gates can perform operations on qubits, which are composed into quantum circuits, similar to how we would compose a classical circuit using classical gates. Now moving on to quantum computing and chemistry. Chemistry is one of the near-term applications of quantum computing. One of the reasons for this is that it's better to use a quantum computer that has quantum properties to perform computations of a quantum system. And the reason why we perform simulations is to understand physical problems, of course, to create new theories, to explain these problems and to predict what will happen if similar problems are repeated, and to also explain unexpected results. Computational chemistry is a large field. For most of the talk, I'll be talking about quantum chemistry in particular, in which researchers are trained to model atomic interactions as accurately as possible using the Schrodinger equation. Now the energy sector is one of the important sectors of chemistry that we believe will be impacted by quantum computing because the world demands for energies growing at a dramatic pace. More research is needed to discover alternative energy sources, including batteries. And to cut the story short, we believe that quantum computing could play a pivotal role in this space because it may be possible for quantum computing to enhance the types of simulations that we can perform on some of the processes that occur in batteries because they are really complex, complicated processes that classical computers cannot really deal with at the moment. So we believe that quantum computing could improve our ability to make more accurate predictions. All right. And if I understand correctly, it's mostly about, or at least in part, about better assimilating molecules. Because we, obviously, to create a new material, we first need to create a new molecule. And to create a new molecule, you have to have this configuration of atoms arranged in a specific way. And is that saying, if you can explain from, as you're obviously the chemistry expert, how exactly it works, how can you find this perfect molecule for your new material and why quantum computers are better in that task than classical ones? Well, I mentioned quantum chemistry, and I will grossly understate things by saying that it's very complicated. There are many types of simulations that we can perform. I mean, even normal chemistry is quite complicated, right? Yes, it's very complicated as well. Yeah, quantum chemistry is just another order of computation about that. So there are many types of simulations that we can perform, and we can obtain answers that will allow us to make assumptions based on the results that we generate, can form hypotheses, we can innovate based on these hypotheses. But quantum chemistry as practiced today on the classical computers also struggles with many problems. And this has mostly to do with the accuracy of the quantum chemistry methods and how easily we can actually carry out these simulations. For example, one of the most widely used methods that we use is density functional theory or DFT. But we have to carefully evaluate the accuracy of DFT methods by benchmarking. And this is time consuming. There's no really reliable way to improve your results. And in addition, some methods maybe may not accurately simulate chemical species with exotic electronic structures, such as radicals, baradicals, and so on with unpaired electrons. And these are normally generated in batteries. DFT methods are not universally accurate. In order to characterize the electronic structures of these species, we like to use more accurate methods such as wave function methods to study these types of systems, but they can be very expensive in terms of the memory required, the amount of space required, the amount of time required for the computation. We believe that quantum computing could help us by allowing us to perform such types of electronic search calculations or wave function calculations more efficiently. Yeah, that's a great explanation. Thank you for that, Gavin. And of course, as you mentioned, all these simulations are super important for new materials and for our discussion today, for new batteries. And with that, let me go to Theo now and tell, what's your view? Why do we need new batteries? I mean, Gavin touched upon that already, obviously, but I would really love to hear what you think. Why is it that you guys are actually working so hard to create new batteries for the world? I mean, we already have batteries, right? Sure. But the reality is that we need really for different reasons. And one of those, one of the, I think the most important one, that may not necessarily be of relevance for electrical mobility is the fact that we need the battery technology that is also capable of absorbing the energy that is produced with the sustainable renewable sources. These are very fluctuating sources of energy. The sun one day shines the other day, it's cloudy, you have strong winds one day and no winds for a week. And we need to be able to absorb that energy and store it for when the society really needs it. On the other hand, we need to, if we now look more at the consumer market, like electrical mobility, it's crucially important to push the technology to, let's say, advance more and more the experience of driving an electric car. We know that, I mean, there are different reasons for that, of course. I mean, one is also when I started working on the battery fields, and Katya, this goes back a little bit of times ago, like 12 years ago, we started, we initiated a project in IBM that was called Battery 500, and we really put lots of effort in exploring new technologies, mostly metal air batteries that may have been actually very good candidates for creating that gap that would have really provided more value and a better experience for people driving electrical mobility. And of course, you do that for different reasons. I mean, one is the primary experience of a person driving a car and not having the anxiety of thinking that the battery is going to discharge quickly and you will be in the middle of nowhere. So basically elongate the range, but we also need to do that for other reasons, because there are so many other sectors. As I said, electrical mobility is a very important one. It's one that I do care personally, so I really am a strong supporter of electrical mobility. But if we think to other types of technologies, like even semiconductor technologies, there have been very fascinating projects even within IBM where we were trying to understand whether the replacement of specific wiring inside the semiconductor may have been possible through the use of specific liquids. At the time, we were talking about electric plug. An electric bulb is nothing more than a liquid that is carrying out electricity and then using a concept which is very similar to the concept of a battery extracting the energy out of the fluid for power in the electronics. So I think, I mean, there are so many levels in our society from consumer goods to infrastructure works, thinking to a stabilization of breathe and storage of renewable sources, but even all the way to miniaturization and the semiconductor where the further development for battery materials is something that we cannot ignore anymore. Yeah, I mean, I totally agree. We can't ignore it for sure. And also for IBM specifically, it's such a huge topic, right? A huge field of research that I think last year we even had it on our like a 5-in-5 list of predictions. So prediction that we think is going to come true within the next five years. Is that right, yeah? Absolutely, absolutely. The energy was one of the five predictions and people may not be aware what is 5-in-5 and maybe I may also ask the colleagues if they want to face the link of the 5-in-5 of 2020. It's a bet that we do as IBM research on technology on fields of technology where in the next five years there will be a revolution. And last year it was about materials. Materials they may have had an impact on technologies that are relevant for sustainable goals. I also, I loved a lot 2020 with respect to the pandemic and all the situations that happened worldwide because it was the year for the UN sustainable goals. So we had really a very strong drive even in analyzing the patterns and identifying the five main things where the development of novel materials may dramatically change the technology, the situation of the technology and the energy was there as a development of materials for batteries. Cool, yeah, absolutely. Thanks for that. And we now have questions coming in actually from the audience and there is one for Gavin. So somebody here is asking looks like if any one of us wanted to actually simulate chemical reactions in batteries using the IBM quantum tools right now what are the current IBM quantum computers actually capable of? Gavin? Sure, so as I mentioned most of the efforts in quantum simulations for chemistry has been on using quantum chemistry. We have had some degree of success but there's still a lot of work to do. So we've developed several error mitigation techniques such as those designed to deal with device readout error, those to perform noise extrapolations and so on. We've also developed strategies to reduce the number of qubits required to simulate larger molecules. For example, exploiting symmetry, use of frozen cores, use of active spaces for reactions. We have created new onsets that can be used to actually do these simulations on quantum simulators as well as on quantum hardware. We've created new, we've used near-term variational algorithms such as VQE or variational quantum egg and solver to perform hybrid simulations part of done in classical hardware and part of done in quantum hardware or with quantum simulators. But there are many challenges and these mainly have to do with the development of better quantum algorithms, developments of better hardware. Some of the main challenges are that we need to develop and use better quantum algorithms as I mentioned. At present, because we have a limited number of noise qubits, we're dependent on quantum algorithms like VQE in which you run a quantum circuit, hundreds if not thousands of times in order to perform statistics and obtain a final energy. There are other algorithms that have been developed for chemistry like phase estimation but they require too many qubits or they require the ability to operate on deep quantum circuits which is not something that we're capable of doing today. We have a limited number of noise qubits but because of these limitations, we try to limit the number of qubits that we are actually using as much as possible because their error increases while increasing the number of qubits that you're using. So at present, we can only perform calculations of small chemical systems or we study systems in which we limit the number of orbitals being used in order to map onto fewer qubits. We also limit the size of the basis set which describes the atomic orbitals to a limited set in order to reduce the number of qubits and gates needed. So in general, the development of better quantum hardware and error mitigation techniques or fully fault tolerant qubits would help. Right now, we interface with quantum computers in the cloud. In the future, we hope that quantum computing would work right alongside classical computing for chemistry problems so that it would be even easier for us to actually use the quantum computers that we have. By the way, I can't help but comment. You're talking about all this super cool futuristic stuff and you're joining us from your bedroom which is quite amazing. It's such a norm right right now, right? People are just working from home and even if you're working on quantum applications and really amazing stuff as you mentioned through the cloud, of course. So you can access your quantum computers through the cloud. You don't have to actually be anywhere else right now, right? Right. But what you were just talking about right now, actually you kind of answered the question that's been sent to me now for you from IU Zhang who is asking, the current quantum computer is still NISC? How to apply this kind of device with limited size and error to real applications? So maybe you could just summarize, well, first of all, what NISC is but you kind of answered that already but maybe just to simplify a little bit for the audience who are maybe not so, don't have so much expertise in quantum computing terminology. Yes. So NISC stands for, this refers to these near-term computers that we're using today. So these are quantum computers that are limited in number of qubits and are also noisy. And as Kenny mentioned, yes, I did answer a lot of this already. So we've tried to implement a lot of techniques to deal with the noise that we have. We tried to limit the number of qubits. We try to develop error mitigation techniques and use them while we're doing all these simulations in order to more accurately simulate the systems that we're interested in. Great. And as we mentioned earlier, and you keep repeating it, so we are using these quantum computers for molecular simulations and we will be using them more in the future when they will be working better than classical ones, specifically for molecular simulations to create better materials and hopefully better batteries in our case. But at the moment, we are not quite there yet. So today we have to rely on today's technology and now with that, I'd like to go back to Theo because that's exactly where your work comes in, right? Because you work on AI and how can AI help us with creating a new battery, better battery right now today? Absolutely. Before I jump into the AI topic, there is one thing that I wanted to comment on what you mentioned by Gavin. It's a very exciting thing because to a certain extent we are experimenting exactly what quantum chemistry was experimenting 30 years ago when we still didn't have the computational power that we have today. So we are really trying to adopt algorithms and grow and adopt the algorithms with the growth of the corresponding hardware. I think it couldn't be a more exciting time for people that are developing software, for people that are thinking how to apply the physical schemes on algorithms. So it's really a very fascinating time. And now, following on your question, AI, of course, I mean, we start from a different perspective, the perspective of the data. And I think, I mean, I'm actually here in the lab and I'm apologizing for all the people that are listening. If there is a little bit of background noise, it's truly live with operations in the autonomous lab. But the AI is really, in the last years, really got a very important role in the space of material design. We have been using AI for collecting information. We have been using AI for building models that were capable of predicting new materials' hypotheses. That, at the end, is one of the main advantages of machine learning and data-driven models. Reasoning on data, extracting part of data, identifying the correlation, and using those signals to support the main expert in producing new type of materials for battery application. AI is also doing something else, though. And that's really very important. It helps us in synthesizing those materials. And that's the reason why I'm here today. The space where I'm broadcasting from is actually the first autonomous lab connected in the cloud, reachable through Internet. It's really an embodiment, a chemical laboratory embodied in the cloud. So, having said that, we really talked a little bit about the use of AI in the technology stack. And AI is really helping to accelerate a different stage of the research. And this is what we actually, in IBM, call accelerated discovery. We are taking a big advantage of AI, but also of quantum and the cloud, different type of technologies to accelerate the research and be able to design new materials with 10x factors on time and cost. So, this is really the objective of the accelerated discovery. Reduce as much as possible the trial and error to be able to design new materials, vector materials faster and with a more precise specification on possible packaging. There is another space that I believe a little bit more disconnected from AI. It's a little bit more general. And I think that that will be another sector where we will be observing a lot the use of AI. I think in mostly all modern electric vehicles, there are systems of artificial intelligence for managing the battery packs. So, that's really very important. I mean, real-time data analysis of all the sensors that are in that battery packs that really allow to calibrate the use of the pack and again improve the experience for the people. So, AI is having a humongous role for what is basically most relevant for me, for us that are talking about researching the design of new materials. But it will have really incredible implication on how complex battery infrastructure are going to be easily handled in the future. It's very similar if I can make a last parallelism to the fly-by-wire concept. Why did people introduce the fly-by-wire? Because there is a lot of complexity in handling unstable design of aircraft. And you really need to rely completely on computer. And here in this space, whenever you start putting together different battery packs, so we are talking about technology, but we are talking also about all the package that is taking care of the cooling of the battery, how much power you can drain out of the battery, and how much power you can even inject in the battery, the system starts to be incredibly complex. And of course, it's beyond the human capabilities to handle the complexity and it's there where AI can come and help. Right, right. Well, so you kind of mentioned that you're in your AI lab right now, but I think what you haven't done yet is actually named the machine that's doing all this amazing work, right? You're next to it, aren't you? So maybe you could introduce us to your robotic helper and what it does exactly? Absolutely. It's really a pleasure. So the environment that you see behind me is actually what we label RoboRXN. RoboRXN itself is not something physical. It's really a blending of three technologies. We have AI, we have cloud, and we have commercial automation hardware. Behind me, you can see the third component, the commercial automation hardware, whether with analytical instruments here on my lab. So RoboRXN was really the extension of the work that we did and that we named in 2018 IBM RXN for chemistry. The primary goal of that activity is the extension, the use of natural language processing technologies to domain specific languages. And in our case, chemistry. So we have been really demonstrating like you can use the same architecture that we routinely daily use for translating between English and Spanish or Italian or Japanese and use exactly the same technology, train those architecture with chemical data to support the chemist in design synthesis, in synthesizing new materials. And in this lab, we are strongly specialized on small organic molecules. And while most of the battery application is actually solid state, so we are not really talking about organic molecules, here instead, we can design and synthesize molecules that can really make a difference for energy storage device like electrochemical flow cells that are very relevant and very important for large scale application. So it's possible to synthesize materials that molecules that have specific redox potential and then optimize the difference in redox potential in such a way to maximize the output voltage of the entire technology. And here I think we are straining some of the images out of a sample of a chemical synthesis that we have been doing for one of these small molecules in the context of semiconductors, but the images are pretty simple. So very interesting technology. Again, even if highly tailored for small organic molecules are very relevant even for optimizing large scale energy storage devices that may have a very relevant application for the day. I remember not so long ago, you even created a molecule life in front of a live audience at the shore of the Zurich lake. So if people want to watch Theo do that, so Google something like Theo IBM Research Molecule Life and I'm sure you'll come across the YouTube video and it was quite a show, it was quite amazing. Absolutely, it was the time we were streaming by the lake of Zurich. This time I decided to stream from the lab. It's really the final concept of providing even to chemists an home office opportunity. You can do chemistry from the sofa and the comfort of your apartment. Yeah, absolutely. Before we go to the questions from the audience, there are quite a few that are waiting now. I'd like to ask you Gavin, actually, what your thoughts are on RoboRXN and specifically on AI more generally, actually. Do you think in the future quantum computing can help or actually not help but work alongside AI to together these two technologies kind of accelerate our research of new materials for batteries even more? What do you think? Yeah, sure. So first of all, I think RoboRXN is very cool. I'm a trained chemist and I actually think that a lot of that I see in AI and machine learning, especially for chemistry, is almost like magic. I know a lot of my experimental chemistry colleagues at Almedin Rare work were very excited to see the RoboRXN demonstration and what it could mean for their research. And it was exciting for me as well. I think a lot of my experimental chemistry colleagues were excited about the fact that they could actually, as Teo said, actually do experiments from home, which is perfect for the current times that we're in, right? So we're talking about the combination of AI and quantum computing. Actually, we've already started to make progress on uniting quantum computing and machine learning for investigations of materials. In particular, we deployed quantum machine learning to assist classical DFT methods in a study of disordered crystalline materials used on lithium-ion batteries. We've also used neural networks to improve the precision of observables for chemical systems to arrive at more accurate results. So these examples highlight the pathways that one could take when leveraging AI and quantum computing for chemistry. So using machine learning to improve quantum algorithms and or using quantum machine learning to improve on classical results. Right. Well, here's a question for you, Gavin, from the audience here. I don't have the name this time. I'm sure you all are not the only ones researching the use of quantum and AI for applications on batteries. Really? Well, I hope you're not. Anyway, do you have any stories of how you learn, you collaborate with other researchers on this topic? That's a good question. Certainly. So we actually work with partners. So we're not on this path alone. We work with a number of partners. And typically what we depend on from these partners is that they come in with the subject matter expertise. So they're the ones who are coming in with projects and saying, hey, how can we actually use quantum computing to study this type of material or this type of system? And so we've partnered with companies like Daimler, Mercedes-Benz, as well as K University, JSR, well, K University, as a university, of course, JSR, Mitsubishi Chemicals, to actually study these types of systems. And typically they come in with a use case, and then we try to help them tailor what they're interested in to the quantum computer. And can you share any of the latest results maybe with any of these companies? I mean, how close are we to applications, really? Oh, sure. Yeah. So we have, as I mentioned, we worked with Daimler, Mercedes-Benz actually on the project that I mentioned earlier using QML for study of lithium ion crystalline materials. And we've also worked with them on lithium sulfur batteries. We've also collaborated with Mitsubishi Chemical and K University, Lithium in Japan, on the research involving lithium O2 batteries where there we were looking at a reaction that's proposed to occur in lithium air batteries. Of course, we're also working with other partners on other topics in chemistry and physics. For example, we worked with Mitsubishi Chemical, K University, and JSR on simulations of materials for OLED applications. So we believe that partnerships like these are very valuable, as I mentioned, because the partner comes in with the subject matter expertise and challenges us to actually be able to study these systems on quantum computers. And so you mentioned lithium. So basically, I guess my question is, are you mostly trying to kind of improve the current lithium ion batteries, or are you trying to move away from lithium ion completely and create something completely new? We're trying to do everything basically. That's a goal of research, right, is to be able to improve the things that we have now, as well as to actually look at some new areas that could potentially offer us a new pathway or new technologies that we can actually iterate on and can actually improve the things that we're working on right now. Right. And Teo, what do you think about lithium ion? What's wrong with current lithium ion batteries right now anyway? Let me also follow up first on the question that was coming from the audience. Beyond the very relevant, very important collaboration with Industrial Partner, we also have stories within governmental funded projects. So an example here in Switzerland for people that may be aware of that is the Marvel Center for Search, basically, where we have been really using in the last seven years AI resources and technology to design new cut of materials, new unknown materials within a specific type of technology. So I think it's a very interesting, I'm sure, I mean beyond even the one that I'm mentioning, if we go across the entire research unit we will find very amazing stories of collaboration in the space of factories. As I said, it's really a 10 years plus belief that we have in novel materials that can change the landscape of storage devices. And why do we need to do that? Let me connect to your question about lithium technology. There is nothing really bad about lithium. In reality we have a very solid network of recycling for lithium. However, the current trend, the current extrapolation, if we really extrapolate to an entire humanity relying on electrical energy storage devices, one of the main concern is really availability of lithium. And so it's really important to explore different type of technologies. There will never be one battery that fits all type of problem and all type of application. This is something that we will have really to realize and live with that. We were discussing earlier about the battery application for grid application. These type of batteries are completely different from the batteries that are running in electric vehicles. In electric vehicles we want to have re-usability, we want to have rechargeability and we want to have even a larger number of cycles. Now to a certain extent, lithium technology made it really possible. It made it possible in terms of gravimetric energy density. And this basically means that with the same weight we are carrying an amount of energy that even if less than the one that a liter of gasoline may be producing is still pretty much comparable in the overall economy of the energy conversion in in a piston engine. But there are other technologies. I mean we have been looking at sodium. We have been looking at other metals providing the really the source, the main energetic drive. I think it will be important to have a specific differentiation where we are really identifying different technologies for the specific use. And different technology and different type of materials for the specific use. And there the importance of using AI, the importance of using quantum for calibrating, optimizing the material design quite often can make a big difference in terms of several equivalents of megawatt hours of energy that are stored in a device. So yeah I'd like to just interject here. So yeah first of all, yes I agree with Matteo. So some of the other some important reasons as some of them Matteo mentioned is of course abundance of lithium as a natural resource. There's a finite amount of lithium available. It's worthwhile to look for an alternative battery material. For example sodium as Tia mentioned is widely abundant. It's a lot cheaper. Moreover in some cases the technologies for lithium batteries rely on other natural resources. For example cobalt which are obtained in problematic ways. So they're in conflict zone for example. And if we can create batteries that don't rely on such materials then it would be better for us all I believe. There's also a possibility of creating safer batteries. So safety is also another important concern here as well. Yeah for sure I totally agree. And actually important question here from the audience which we probably should have addressed a little bit earlier but you know better late than never. H. Lander is asking what actually describes the battery of the future and is it just an accumulator so reloadable or just a battery that is like gone after draining. So what's your view? I'll pass it over to Tia. I want to remark what I said before. There is not going to be the battery of the future. There are going to be batteries of the future. What we need to do and this is the responsibility of science and research is really push all the technology to the level where we are really achieving a very high level of gain in terms of energy stored and very little losses. So I was thinking when when Katia you were reading this question I don't know why it came to my mind the thermonuclear batteries that we use in satellites. This type of batteries or in rover that are exploring other planets on behalf of humans these batteries are going to be there and because they offer a certain number of reliability and independence that different type of technologies are not really capable of offering. But this is really only one example. We will have batteries that will be specifically optimized for electrical mobility. We will have batteries that will be optimized for grid application where actually there is not too much pressure and stress on the number of cycles that we can achieve because normally when when you talk about batteries for mitigating the grid or absorbing a surplus of energy produced by renewable sources you are really you should really picture like huge containers, thousands and thousands of leaders of solutions. These are the electrochemical flow cells where you are actually storing the excess energy and whenever needed this is actually released but you are not doing the same type of use that you do with batteries in electrical mobility where maybe you are charging today is charging in three hours and then recharging again in four hours. So there will not be one battery but there will be many batteries and I think our role is really the one of being sure and that the technology like AI and quantum are really used for improving the design of the material storage technology. Yeah for sure and of course as you mentioned I mean storage is super important for renewables right because obviously when we've got solar panels or wind turbines you're not gonna like we do have to store that energy for longer even during the night or when there's no wind right so we shouldn't be dependent on the weather conditions and we are right now so that's also important I totally agree. Here is a question from Akhil Kara for Gavin I guess and Akhil is saying are there any new tools under development especially with respect to quantum computing that will help simplify and speed up quantum chemistry and computational chemistry research? Yes that's a good question. So at IBM we're always trying to make make the tools that we're using and the tools that we're offering to the wider public as to make it as easy for our users as possible and yes so we're also we're always trying to make improvements to the main software that we use which is called Qiskit so that a lot of it that users are exposed to like application researchers are exposed to would be abstracted away from the user. Also if you are interested in say not only working on applications and quantum computers say to improve better research or so on then you could also look at things like if you're interested in say making better qubits they could look at Qiskit metal so we have a lot of tools that we're trying to actually implement that will make things easier for users to actually use the tools that we have. Great yeah absolutely thanks for that Gavin and here is a question about recycling from the audience here which is actually a really important aspect as well that we haven't really talked about much yet. So Charles Goh is asking is there any fundamental change of battery technology? What about recycling? How to save the weight in order to use battery in aviation industry? So I guess that's several questions in one but tell maybe you could talk talk to us about recycling specifically for now and address aviation. This question is just touching me on my most interest which is not recycling but aviation actually. Of course because you fly right you're a pilot. Absolutely. So let me let's first start addressing a fundamental change of battery technologies. There have been in the last 10 years a very I would say important attention by many companies including us to metal air technologies. I would really say that these technologies at least on paper it's actually the one that can deliver an incredibly competitive gravity metric and I'm always referring to weight for a reason. I'm arriving there compared to the fuel that we are burning in piston engine. There are some difficulties however and this is more or less where I personally from my perspective I really left the field of metal air batteries. One of the side products in this process is really the formation of substances that are highly oxidizing. We are talking about peroxide superoxides and the reason is that one of the other one element is the metal that you are carrying with you but the other element is the oxygen that is in the air and that's the beauty of this concept. You don't need to care the anode and the cathode actually or the anode and the cathode material but one of the two the oxygen that is in the air is directly available and you don't doesn't really contribute to the weight that you are carrying. There are aware as I said important challenges and one of those that we have been looking for several years in a period where AI was not yet there unfortunately was the stability of the electrolytes. You always need a medium for diffusing the ion and closing the cell. Most of the electrolytes are organic molecules and one of the main complexity was really to identify stable electrolytes that may have undergo to a specific number of cycles. Generally for user application we are talking about several hundreds if not a thousand cycles during the lifetime of the battery without really degrading and this unfortunately didn't really happen and I think there is always a lot of research the interest for metal oxygen batteries of course shifting towards solid state concept so you are completely getting rid of the organic electrolytes to have instead the depressants of a solid state electrolyte with huge benefits in terms of chemical and electrochemical stability so these are some of the I would say forefront technologies nothing that unfortunately we will be able to consume tomorrow maybe in few years it's always the big question mark are we going to make that technological discovery that is going suddenly to make the entire technology usable. In the lab metal air batteries are amazing and are one of the still well thought and always considered ideas to promote mobility in aviation so electrical aviation and there we are going to the second part of the question what do we really need well we need light batteries we need really batteries that are competitive in terms of gravimetric and volumetric energy density in an airplane it's important also the volumetric components because the wings which is normally where we are storing fuel the wings of the airplane have a fixed volume so you want to maximize also the volumetric the volumetric density normally with the design of the cars the volumetric density is important but you can always play a little bit increasing the volume of the car making a car a little bit bigger and the stronger focus is really on gravimetric density in the case of airplanes instead the two are playing both of them a very important role one because the volume is normally pretty well defined and the second because the weight is going to have a very important component if you are too heavy you are not taking off from a runway in all that and I will close it here recycling it's an important component of every battery technology we we are talking a lot about sustainable goals but when we talk about sustainable goals it also means that all these technologies that are not necessarily uh how to say uh renewable in principle because we are we are taking uh resources land resources we're taking we are taking other minerals for building the batteries we have to act in a very responsible way and the amazing things uh for lithium ion batteries is that after roughly 30 years of consumer uh consumption by the people we have in place I don't have really the last digits of my availability but we have in place a mechanism a chain of recycling that can actually recycle more than 95 97 percent of the batteries that's that's really where we have to aim and ideally even push the number the number higher are we are we working on any of that uh somebody is actually asking whether you guys are doing any research that could lead to uh more kind of easily recyclable batteries so regarding recyclability one of the main components uh beyond the material the material itself is really defining the performance the retrochemical performance but the recyclability and how easy is to recycle is really a packaging problem normally is is an engineering problem and it's there that AI comes to our hands we have uh some projects that we are incubating inside our research lab where we are using AI to drive the design of uh packaging and and user products not only the the design of the materials the chemical composition or the formulation but really how you should assemble uh your uh your consumer goods in order to maximize specific properties this is a very interesting concept for AI aided design of uh uh of products in semiconductor industry but also in the uh energy storage yeah fantastic thank you for that Theo and just a very last question here uh probably very short but important for the audience somebody is asking where uh they can actually go and experiment with quantum computers today uh and uh with RoboRXN so if you guys can point our our viewers here to to the right direction to you know play around with Qiskit and I don't know if it's possible to do anything with RoboRXN but let us know Gavin sure yeah so um as I mentioned uh the IBM computers are accessible through the cloud um if you go to um quantum computing dot IBM dot com today you can sign up for an account you can start playing around with um IBM quantum computers you can start learning more about um quantum computing um and so this is uh this is some this is a resource that uh that is open for for everyone and everyone can use it today fantastic and I I want to stress the appeal for for Qiskit also an appeal to all the quantum chemistry or computational chemistry that are in the audience uh take the initiative and start really uh experimenting with the uh early technology early quantum computing technology is going to be an incredible investment also for uh career prospects I mean we should learn by the history and as I said at the beginning this period is resembling a lot uh what I experienced only partially at the beginning of the 90s uh so fantastic here regarding RoboRXN and ArxN all the models we make available all the models so if there are people that are interested in chemistry synthetic chemistry how can you predict the outcome of a chemical reaction are you preparing yourself for chemistry classes uh reach out IBM ArxN for chemistry uh just google and you will land on the website and you can use the resources the train model for uh prepping your chemistry class exam thank you fantastic thank you so much uh Theo and thank you Gavin that was a fantastic session I hope our viewers enjoyed it as as much as I did and thank you for um thank you to everyone for joining us today of course and to our stellar team of scientists behind the scenes that were answering questions in the chat and if you guys have any comments ideas questions you can always find me for example on on twitter and send me a question or whatever you like DM me and please don't forget to subscribe to our channel and join us again next month thank you and goodbye thank you all right