Bright X-Rays, AI, and Robotic Labs—A Roadmap for Better Batteries
Steven Cherry Hi, this is Steven Cherry for Radio Spectrum.
Batteries have come a long way. What used to power flashlights and toys, Timex watches and Sony Walkmans, are now found in everything from phones and laptops to cars and planes.
Batteries all work the same: Chemical energy is converted to electrical energy by creating a flow of electrons from one material to another; that flow generates an electrical current.
Yet batteries are also wildly different, both because the light bulb in a flashlight and the engine in a Tesla have different needs, and because battery technology keeps improving as researchers fiddle with every part of the system: the two chemistries that make up the anode and the cathode, and the electrolyte and how the ions pass through it from one to the other.
A Chinese proverb says, Give a man a fish, and you feed him for a day. Teach a man to fish, and you feed him for a lifetime." The Christian Bible says, follow me and I will make you fishers of men."
In other words, a more engineering-oriented proverb would say, let's create a lab and develop techniques for measuring the efficacy of different fishing rods, which will help us develop different rods for different bodies of water and different species of fish."
The Argonne National Laboratory is one such lab. There, under the leadership of Venkat Srinivasan, director of its Collaborative Center for Energy Storage Science, a team of scientists has developed a quiver of techniques for precisely measuring the velocity and behavior of ions and comparing it to mathematical models of battery designs.
Venkat Srinivasan [Ven-kat Sri-ni-va-san] is also deputy director of Argonne's Joint Center for Energy Storage Research, a national program that looks beyond the current generation of lithium-ion batteries. He was previously a staff scientist at Lawrence Berkeley National Laboratory, wrote a popular blog, This Week in Batteries," and is my guest today via Teams.
More Signal, Less Noise:The Radio Spectrum Podcast With Steven Cherry
Sign up to be alerted when
the next episode drops.
Venkat, welcome to the podcast.
Venkat Srinivasan Thank you so much. I appreciate the time. I always love talking about batteries, so it'd be great to have this conversation.
Steven Cherry I think I gave about as simplistic a description of batteries as one could give, maybe we could start with what are the main battery types today and why is one better than another for a given application?
Venkat Srinivasan So, Steve, there are two kinds of batteries that I think all of us use in our daily lives. One of them is a primary battery. The ones that you don't recharge. So a common one is something that you might be putting in your children's toys or something like that.
The second, which I think is the one that is sort of powering everything that we think of, things like electric cars and grid storage or rechargeable batteries. So these are the ones where we have to go back and charge them again. So let's talk a little bit more about rechargeable batteries that are a number of them that are sitting somewhere in the world. You have lead-acid batteries that are sitting in your car today. They've been sitting there for the last 30, 40 years where they used to stop the car for lighting the car up when the engine is not on. This is something that will continue to be in our cars for quite some time.
[n-dashes].
You're also seeing lithium-ion batteries that are not powering the car itself. Instead of having internal combustion engine and gasoline, you're seeing more pure chemicals coming out that have lithium-ion batteries. And then the third battery, which we sort of don't see, but we have some in different places are nickel-cadmium and metal-hydride batteries. These are kind of going away slowly. But the Toyota Prius is a great example of a nickel-metal hybrid. But many people still drive Priuses-I have one-that still has nickel-metal batteries in them. These are some of the classes of materials that are more common. But there are others, like flow batteries, that people haven't really probably thought about and haven't seen, which is being researched quite a bit, there are companies that are trying to install flow batteries for grid storage, which are also rechargeable batteries that are of a different type.
The most prevalent of these is lithium-ion; that's the chemistry that has completely changed electric vehicle transportation. It's changed the way we speak on our phones. The iPhone would not be possible if not for the lithium-ion battery. It's the battery that has pretty much revolutionized all of transportation. And it's the reason why the Nobel Prize two years ago went to the lithium-ion batteries for the discovery and ultimately the commercialization of the technology-it's because it had such a wide impact.
Steven Cherry I gather that remarkably, we've designed all these different batteries and can power a cell phone for a full day and power a car from New York to Boston without fully understanding the chemistry involved. I'm going to offer a comparison and I'd like you to say whether it's accurate or not.
We developed vaccines for smallpox beginning in 1798; we ended smallpox as a threat to humanity-all without understanding the actual mechanisms at the genetic level or even the cellular level by which the vaccine confers immunity. But the coronavirus vaccines we're now deploying were developed in record time because we were able to study the virus and how it interacts with human organs at those deeper levels. And the comparison here is that with these new techniques developed at Argonne and elsewhere, we can finally understand battery chemistry at the most fundamental level.
Venkat Srinivasan That is absolutely correct. If you go back in time and ask yourself, what about the batteries like the acid batteries and the nickel-cadmium batteries-did we invent them in some systematic fashion? Well, I guess not, right?
Certainly once the materials were discovered, there was a lot of innovation that went into it using what was state-of-the-art techniques at that time to make them better and better and better. But to a large extent, the story that you just said about the vaccines with the smallpox is probably very similar to the kinds of things that are happening in batteries, the older chemistries.
The world has changed now. If you look at the kinds of things we are doing today, like you said, that in a variety of techniques, both experimental but also mathematical, meaning, now computer simulations have come to our aid and now we're able to take a deeper understanding on how batteries behave and then use that to discover new materials-first, maybe on a computer, but certainly in the lab at some point. So this is something that is also happening in the battery world. The kinds of innovations you are seeing now with COVID vaccines are the kinds of things we are seeing happen in the battery world in terms of discovering the next big breakthrough.
Steven Cherry So I gather the main technology you're using now is ultraright X-rays and you're using it to come up with for the first time the electrical current, something known as the transport number. Let's let's start with the X-rays.
Venkat Srinivasan We used to cycle the battery up back. Things used to happen to them. We then had to open up the battery and see what happened on the inside. And as you can imagine, right when you open up a battery, you hope that nothing changes by the time you take it to your experimental technique of choice to look at what's happening on the inside. But oftentimes things change. So what you have inside the battery during its operation may not be the same as what you're probing when you open up the cell. So a trend that's been going on for some time now is to say, well, maybe we should be thinking about in situ to operando methods, meaning inside the party's environment during operation, trying to find more information in the cell.
Typically all battery people will do is they'll send a current into the battery and then measure the potential or vice versa. That's a common thing that's done. So what we are trying to do now is do one more thing on top of that: Can we probe something on the inside without opening up the cell? X-rays come into play because these are extremely powerful light, they can go through the battery casing, go into the cell, and you can actually start seeing things inside the battery itself during operando operation, meaning you can pass current keep the battery in the environment you want it to be and send the X-ray beam and see what's happening on the inside.
So this is a trend that we've been slowly exploring, going back a decade. And a decade ago, we probably did not have the resolution to be able to see things at a very minute scale. So we were seeing maybe a few tens of microns of what was happening in these batteries. Maybe we were measuring things once every minute or so, but we're slowly getting better and better; we're making the resolution tighter, meaning we can see smaller features and we are trying to get the time resolution such that we can see things at a faster and faster time. So that trend is something that is going to is helping us and we continue to help us make batteries better.
Steven Cherry So if I could push my comparison a little further, we developed the COVID vaccines in record time and with stunning efficiency. I mean, 95 percent effective right out of the gate. Will this new ability to look inside the battery while it's in operation, will this create new generations of better batteries in record time?
Venkat Srinivasan That will be the hope. And I do want to bring in two aspects that I think work complementarily with each other. One is the extreme techniques-and related techniques like X-ray, so we should not forget that there are non-X-ray techniques also that give us information that can be crucially important. But along with that, there has been this revolution in computing that has really come to the forefront in the last five to 10 years. What this computing revolution is that basically because computers are getting more and more powerful and computing resources are getting cheaper, we are able to now start to calculate on computers all sorts of things. For example, we can calculate how much lithium can a material hold-without actually having to go into the lab. And we can do this in a high-throughput fashion: screen a variety of materials and start to see which of these looks the most promising. Similarly, we can do it, same thing, to ask: Can we find iron conductors to find, say, solid-state battery materials using the same techniques?
Now, once you have these kinds of materials in play and you do them very, very fast using computers, you can start to think about how do you combine them with these X-ray techniques. So you could imagine that you're finding a material on the computer. You're trying to synthesize them and during the synthesis you try to watch and see, are you making the material you were predicting or did something happen during synthesis where you were not able to make the particular material?
And using this complementary way of looking at things, I think in the next five to 10 years you're going to see this amazing acceleration of material discovery between the computing and the X-ray sources and other techniques of experimental methods. They're going to see this incredible acceleration in terms of finding new things. You know, the big trick in materials-and this is certainly true for battery materials-if you can find those materials, maybe one of them looks interesting. So the job here is to cycle through those thousand as quickly as possible to find that one nugget that can be exciting. And so what we're seeing now with computing and with these X-rays is the ability to cycle through many materials very quickly so that we can start to pin down which of those which of the one among those thousand looks the most promising that we can spend a lot more resources and time on them.
Steven Cherry We've been relying on lithium-ion for quite a while. It was first developed in 1985 and first used commercially by Sony in 1991. These batteries are somewhat infamous for occasionally exploding in phones and laptops and living rooms and on airplanes and even in the airplanes themselves in the case of the Boeing 787. Do you think this research will lead to safer batteries?
Venkat Srinivasan Absolutely. The first thing I should clarify is that the lithium-ion from the 1990s is not the same lithium-ion we used today. There have been many generations of materials that have changed over time; they've gotten better; the energy density has actually gone up by a factor of three in those twenty-five years, and there's a chance that it's going to continue to go up by another factor of two in the next decade or so. The reality is that when we use the word lithium-ion, we're actually talking about a variety of material classes that go into the into the anodes, the cathode, and the electrolytes that make up the lithium-ion batteries. So the first thing to kind of notice is that these materials are changing continuously, what the new techniques are bringing is a way for us to push the boundaries of lithium-ion, meaning there is still a lot of room left for lithium-ion to get better, and these new techniques are allowing us to invent the next generation of cathode materials, anode materials, and electrolytes that could be used in the system to continue to push on things like energy density, fast-charge capability, cycle life. These are the kinds of big problems we're worried about. So these techniques are certainly going to allow us to get there.
There is another important thing to think about for lithium-ion, which is recyclability. I think it's been pretty clear that as the market for batteries starts to go up, they're going to have a lot of batteries that are going to reach end-of-life at some stage and we do not want to throw them away. We want to take out the precious metals in them, the ones that we think are going to be useful for the next generation of batteries. And we want to make sure we dispose of them in a very sort of a safe and efficient manner for the environment. So I think that is also an area of R&D that's going to be enabled by these kinds of techniques.
The last thing I'd say is that we're thinking hard about systems that go beyond lithium-ion, things like solid-state batteries, things like magnesium-based batteries ... And those kinds of chemistries, we really feel like taking these modern techniques and putting them in play is going to accelerate the development time frame. So you mentioned 1985 and 1991; lithium-ion battery research started in the 1950s and 60s, and it's taken as many decades before we could get to a stage where Sony could actually go and commercialize it. And we think we can accelerate the timeline pretty significantly for things like solid-state batteries or magnesium-based batteries because of all the modern techniques.
Steven Cherry Charging time is also a big area for potential improvement, especially in electric cars, which still only have a driving range that maybe gets to 400 kilometers, in practical terms. Will we be getting to the point where we can recharge in the time it takes to get a White Chocolate Gingerbread Frappuccino at Starbucks?
Venkat Srinivasan That's the that's the dream. So Argonne actually leads a project for the Department of Energy working with multiple other national labs on enabling 10-minute charging of batteries. I will say that in the last two or three years, there's been tremendous progress in this area. Instead of a forty-five-minute charge or a one-hour charge that was considered to be a fast charge. We now feel like there is a possibility of getting under 30 minutes of charging. They still have to be proven out. They have to be implemented at large scale. But more and more as we learn using these similar techniques that I can see a little bit more about, that there is a lot of work happening at the Advanced Photon Source looking at fast charging of batteries, trying to understand the phenomenon that is stopping us from charging very fast. These same techniques are allowing us to think about how to solve the problem.
And I'll take a bet in the next five years, we'll start to look at 10-minute charging as something that is going to be possible. Three or four years ago, I would not have said that. But in the next five years, I think they are going to start saying, hey, you know, I think there are ways in which you can start to get to this kind of charging time. Certainly it's a big challenge. It's not just a challenge in the battery side, it's a challenge in how are we going to get the electricity to reach the electric car? I mean, there's going to be a problem there. There's a lot of heat generation that happens in these systems. We've got to find a way to pull it out. So there's a lot of challenges that we have to solve. But I think these techniques are slowly giving us answers to, why is it a problem to begin with? And allowing us to start to test various hypotheses to find ways to solve the problem.
Steven Cherry The last area where I think people are looking for dramatic improvement is weight and bulk. It's important in our cell phones and it's also important in electric cars.
Venkat Srinivasan Yeah, absolutely. So frankly, it's not just in electric cars. At Argonne they're starting to think about light-duty vehicles, which is our passenger cars, but also heavy-duty vehicles. Right. I mean, what happens when you start to think about trucking across the country carrying a heavy payload? We are trying to think hard about aviation, about marine, and rail. As you start to get to these kinds of applications, the energy density requirement goes up dramatically.
I'll give you some numbers. If you look at today's lithium-ion batteries at the pack level, the energy density is approximately 180 watt-hours per kilogram, give or take. Depending on the company, That could be a little bit higher or lower, but approximately 180 Wh/kg. If we look at a 737 going across the country or a significant distance carrying a number of passengers, the kinds of energy density you would need is upwards of 800 Wh/kg. So just to give you a sense for that, right, we said it's 180 for today's lithium-ion. We're talking about four to five times the energy density of today's lithium-ion before we can start to think about electric aviation. So energy density would gravimetric and volumetric. It's going to be extremely important in the future. Much of the R&D that we are doing is trying to discover materials that allow us to increase energy density. The hope is that you will increase energy density. You will make the battery charge very fine. To get them to last very long, all simultaneously, that tends to be a big deal, but it is not all about compromising between these different competing metrics-cycle life, calendar life, cost, safety, performance, all of them tend to play against each other. But the big hope is that we are able to improve the energy density without compromising on these other metrics. That's kind of the big focus of the R&D that's going on worldwide, but certainly at Argonne.
Steven Cherry I gather there's also a new business model for conducting this research, a nonprofit organization that brings corporate and government, and academic research all under one aegis. Tell us about CalCharge.
Venkat Srinivasan Yeah, if you kind of think about the battery world and this is true for many of these hard technologies, the sort of the cleantech or greentech as people have come to call them. There is a lot of innovation that is needed, which means in our lab R&D, the kinds of techniques and models that we're talking about is crucially important. But it's also important for us to find a way to make them into a market, meaning you have to be able to take that lab innovation; you've got to be able to manufacture them; you've got to get them in the hands of, say, a car company that's going to test them and ultimately qualify them and then integrate them into the vehicle.
So this is a long road to go from lab to market. And the traditional way you've thought about this is you will want to throw it across the fence, right. So, say at Argonne National Lab, invent something and then we throw it across the fence to industry and then you hope that industry takes it from there and they run with it and they solve the problems. That tends to be an extremely inefficient process. That's because oftentimes that a national lab might stop is not enough for an industry to run with it-there are multiple paths that show up. And when you integrate these devices into the company's existing other components there are problems that show up when you get it up to manufacturing, when you start to get up to a larger scale; there are problems that show up and you make a pact with it. And oftentimes the solution to these problems goes back to the material. So the fundamental principle that me and many others have started thinking about is you do not want to keep R&D, the manufacturing, and the market separate. You have to find a way to connect them up.
And if you connect them up very closely, then the market starts to drive the R&D, the R&D innovation starts to get the people to the manufacturing world excited. And there is this close connection among all of these three things that makes things go faster and faster. We've seen this in other industries and it certainly will be true in the battery world. So we've been trying very, very hard to kind of enable these kinds of what I would call public-private[NDASH] partnerships, ways in which we, the public, meaning the national lab systems, can start to interact with the private companies and find ways to move this along. So this is a concept that I think of me and a few others have been sort of thinking about for quite some time. Before I moved to Argonne, I was at Lawrence Berkeley. And at Lawrence Berkeley-the Bay Area has a very rich ecosystem of battery companies, especially startup companies.
So I created this entity called CalCharge, which was a way to connect up the local ecosystem in the San Francisco Bay Area to the national labs in the area-Lawrence Berkeley, SLAC, and Sandia National Labs in Livermore. So those are the three that were connected. And the idea behind this is how do we take the sort of the national lab facilities, the people, and the kind of the amazing brains that they have and use them to start to solve some of the problems that is facing? And how do we take the IP that is sitting in the lab and how do we move them to market using these startups so that we can continuously work with each other, make sure that we don't have these valleys of death as we've come to call them, when we move from lab to market and try to accelerate that. I've been doing very similar things at Argonne in the last four years thinking hard about how do you do this, but on a national scale.
So we've been working closely with the Department of Energy, working with various entities both in the Chicagoland area, but also in the wider U.S. community, to start to think about enabling these kinds of ecosystems where national labs like ours and others across the country-there are 70 national labs, Department of Energy national labs-maybe a dozen of them have expertise that can be used for the free world. How do we connect them up? And the local universities that are the different parts of the country with amazing expertise, how do you connect them up to these startups, the big companies, the manufacturers, the car companies that are coming in, but also the material companies, companies that are providing lithium for a supply chain perspective? So my dream is that we would have this big ecosystem of everybody talking to each other, finding ways to leverage each other and ultimately making this technology something that can reach the market as quickly as possible.
Steven Cherry And right now, who is waiting on whom? Is there enough new research that it's up to the corporations to do something with it? Or are they looking for specific improvements that that they need to wait for you to make?
Venkat Srinivasan All of the above. That is probably quite a bit of R&D that's going on that industry is not aware of, and that tends to be a big problem-there's a visibility problem when it comes to the kinds of things that are going on in the national labs and the academic world. There are things where we are not aware of the problems that industry is facing. And I think these kinds of disconnects where sometimes the lack of awareness keeps things from happening fast is what we need to solve. And the more connections we have, the more interactions we have, the more conversations we have with each other, the exposure increases. And when the exposure increases, we have a better chance of being able to solve these kinds of problems where the lack of information stops us from getting the kinds of innovation that we could get.
Steven Cherry And at your end, at the research end, I gather one immediate improvement you're looking to make is the brightness of the X-rays. Is there anything else that we should look forward to?
Venkat Srinivasan Yeah, there are a couple of things that I think are very important. The first one is the brightness of the X-rays. There's an upgrade that is coming up for the advanced photon source that's going to change the time resolution in which we can start to see these batteries. So, for example, when you're charging the batteries very fast, you can get data very quickly. So that's going to be super important. The second one is you can also start to think about seeing features that are even smaller than the kinds of features we see today. So that's the first big thing.
The second thing that is connected to that is artificial intelligence and machine learning is becoming something that is permeating through all forms of research, including battery research, we use AI and ML for all sorts of things. But one thing we've been thinking about is how do we connect up AI and ML to the kinds of X-ray techniques we've been using. So, for example, instead of looking all over the battery to see if there is a problem, can we use signatures but of where the problems could be occurring? So that these machine learning tools can quickly go in and identify the spot where things could be going wrong so that you can spend all your time and energy taking data at that particular spot. So that, again, we're being very efficient with the time that we have to ensure that we're catching the problems we have to catch. So I think the next big thing that is going on is this whole artificial intelligence and machine learning that is going to be integral for us in the battery discovery world.
The last thing which is an emerging trend is what is called automated labs or self-driving labs. The idea behind this is that instead of a human being going in and sort of synthesizing a material starting in the morning and finishing the evening and then characterizing it the next day and finding out what happened to it and then going back and trying the next material, could we start to do this using robotics? This is something that's been a trend for a while now. But where things are heading is that more and more robots can start to do things that a human being could do. So you could imagine robots going in and synthesizing electrolyte molecules, mixing them up, testing for the conductivity and trying to see if the conductivity is higher than the one that you had before. If it's not going back and iterating on finding a new molecule based on the previous results so that you can efficiently try to find the answer for a higher conductive electrolyte than one that you have is your baseline. Robots work 24/7. So it kind of makes it very, very useful for us to think about these ways of innovating. Robots generate a lot of data, which we now know how to handle because of all the machine learning tools we'll be developing in the last three, four, five years. So all of a sudden, the synergy, the intersection between machine learning, the ability to analyze a lot of data, and robotics are starting to come into play. And I think we're going to see that that's going to open up new ways to discover materials in a rapid fashion.
Steven Cherry Well, Venkat, if you will forgive a rather obvious pun, the future of battery technology seems bright. And I wish you and your colleagues at Argonne and CalCharge every success. Thank you for your role in this research and for being here today.
Thank you so much. I appreciate the time you've taken to ask me this questions.
We've been speaking with Venkat Srivinasan of Argonne National Lab about a newfound ability to study batteries at the molecular level and about improvements that might result from it.
Radio Spectrum is brought to you by IEEE Spectrum, the member magazine of the Institute of Electrical and Electronic Engineers, a professional organization dedicated to advancing technology for the benefit of humanity.
This interview was recorded January 6, 2021 using Adobe Audition and edited in Audacity. Our theme music is by Chad Crouch.
You can subscribe to Radio Spectrum on the Spectrum website, where you can also sign up for alerts, or on Spotify, Apple, Google-wherever you get your podcasts. We welcome your feedback on the web or in social media.
For Radio Spectrum, I'm Steven Cherry.
Note: Transcripts are created for the convenience of our readers and listeners. The authoritative record of IEEE Spectrum's audio programming is the audio version.
We welcome your comments on Twitter (@RadioSpectrum1 and @IEEESpectrum) and Facebook.
See Also:
Battery of tests: Scientists figure out how to track what happens inside batteries
Concentration and velocity profiles in a polymeric lithium-ion battery electrolyte