Francis O’Sullivan: From MIT, this is the Energy Initiative and I’m Francis O’Sullivan. Welcome to today’s podcast, one of a series we’re carrying out on game-changing energy technologies. We’re talking with colleagues from across MIT on the work they’re doing looking at defining the future of energy. Today we’re talking with three faculty members about the environmental and economical sustainability of materials. Professors Fikile Brushett, Elsa Olivetti, and Yogi Surendranath.
Elsa Olivetti: I’m Elsa Olivetti, the Atlantic Richfield Associate Professor of Energy Studies in the Department of Materials Science and Engineering.
Yogi Surendranath: I’m Yogi Surendranath. I’m the Paul M. Cook Career Development Associate Professor in Chemistry at MIT.
Fikile Brushett: Fikile Brushett, Cecil and Ida Green Associate Professor of Chemical Engineering.
FO: Elsa, Fikile, Yogi, thanks so much for joining us here today on the podcast. Let me start by saying, I’m really excited to have the three of you guys in the room for this particular episode. What I’d like to do to start us off is ask each one of you to reflect a little bit on challenges and opportunities you see today with respect to your work in the kind of context of materials science and how it’s relating to some of the bigger energy questions that you think are important.
EO: I like to think about the systems implications in how we use materials, particularly in energy development. In particular, as our energy technologies become more materials-dependent, so they’re sort of fundamentally driven by materials innovation and the complexity of materials we incorporate into our devices and technologies, that links us to this increasingly complex set of supply chains associated with garnering those materials and the devices that we’re using them in. I think one of the challenges, but also opportunities, is to be able to link our development of those technologies to our understanding of how those supply chains for those materials might evolve as a function of increased demand for them, different kinds of purities and grades of those materials, the nature of how we extract them from the ground, what else they’re extracted with when we extract them from the ground, how we manage the use of them at the end of whatever lifetime of that device. We often like to think about how the complexity of the materials and the constituents used in the materials has increased dramatically over recent decades, and we haven’t necessarily kept a pace with our ability to manage that at the end of life, or even in the manufacturing process associated with those materials.
FO: Right. Fik, how about you, in terms of your particular focus and how you see it being important in the bigger energy story?
FB: I agree with a lot of what Elsa just mentioned. My group plays a slightly different role in this ecosystem, but I do think we’re all connected. We’re particularly interested in trying to understand the connection between system level performance and goals of existent or new technologies and relating that to the property sets of the component materials that would go into it. I see our role as really trying to connect two different worlds. The world of people who will take a device and put it into, say, the grid or put it into some application, test it, see where it’s successful, where it fails, and then move on to the next type of device. Then there’s a group of people, a very large group and community, that focuses on materials development. These could be the chemists of the world, the polymer chemists, the materials scientists, many of whom have tremendous expertise in areas that are relevant to, say, energy storage and conversion and the development of new technologies. But for whatever reason are not as interested in the problem. It could be that they don’t fully understand where they would fit in. It could also be that these are very challenging problems with lots of moving parts and trying to understand how your particular skill set fits in is difficult. What we try to do is really provide folks like that with avenues or entry points into these types of systems to try to take a system-level goal or a target and break it back down into what component metrics are needed in order to hit that target and maybe provide some exemplar cases of things that might begin to walk in that direction or key challenges that exist. The idea is rather than saying “I need a better technology, I need a better battery, I need a better engine,” what specifically do you want to fix? If you can identify what that problem is and make it actionable, then there’s a whole wealth of knowledge at our disposal. It goes toward your earlier point of saying, if we need to do big things and we need to do them fast, we need all hands on deck.
FO: I think that’s a really… go ahead, Elsa.
EO: I was going to say, it links a little bit to what I said, of a way to fold in some of the more systematic issues. If you’re able to think about what specific problem are you trying to solve or a specific metric, I think is what you said. That would be a way to plug in metrics, like what Frank had said around scalability and systems implications. That’s a framing that I think helps also think about it in a holistic way, even though I think in part you were using it as a targeted way to go at what problem you’re trying to solve.
FB: That’s exactly right.
FO: I like concept a lot. I guess that’s one of the overriding philosophies here at MIT, is to really understand a problem and go try and solve that, rather than just doing something for the sake of it. It’s rather simplistic, but certainly that’s important. I think it’s increasingly important for some of these energy challenges. That leads me to some of the work you’ve been working on, Yogi. If we look at the energy transition and the bigger challenge for managing carbon and so on, a lot of the focus has been on electricity. I’m sure we’ll come back to it. Fik’s going to tell us what a better battery looks like later. There is a whole host of other needs, for molecules, for materials, which require us to alter profoundly in many instances how we create those molecules because we’re still going to use or we’re still going to want to require those materials going forward. I’d like your reflections on that particular aspect of the challenge.
YS: When we look at energy problems and a global thinking on what’s really needed, we think a lot about what it takes to be able to efficiently interconvert energy from one form into another. Whether that’s interconverting it from light energy to electrical energy, which we can do now with really impressive efficiencies in commercial solar cell technologies. Things we haven’t conquered are how we convert electrical energy into chemical energy, how we efficiently back-convert those, how we convert one form of chemical energy into another one. Those energy interconversion processes are ultimately related to materials that allow for those interconversions to occur very efficiently, with the minimal loss of energy wasted to heat or other processes that don’t end up storing it or efficiently interconverting, or selectively interconverting it. We really see the problem as a materials problem related to the development of efficient catalysts for energy conversion processes, which is there really across the landscape of technologies that you’d want to deploy in a future low-carbon sustainable energy future. Our focus has really been on, how do we understand at an elementary level what about a material allows it to perform an energy conversion reaction or process efficiently versus inefficiently? Our emphasis has always been that, in many cases for energy conversion processes, they rely heavily on the interfaces between materials. They rely on the surfaces of catalysts, the interfaces between an anode in a battery and the electrolyte. The list can go on and on, but a lot of energy challenges relate to our inability to control material surfaces effectively at an atomic and molecular level and be able to dynamically influence those processes when we need to for an energy conversion challenge. There’s really a vast space that needs to be recast and understood in materials development.
FO: With that said, and with what each of you guys have laid out there, many of the challenges that we’re facing, that the energy system in the economy more broadly is facing, really require not just that innovation to be successful, but that innovation to be translated. I know you’ve been thinking deeply about trying to support that. Do you think we’re actually making progress? Do you think industry broadly–and this isn’t just the energy industry, but it’s all material-centric industries–are they getting better at more rapidly making that transition? Or are we still stuck with the same kind of development and deployment paradigms that we’ve had for decades?
EO: I would probably have to say I think we’re more stuck than improving. [Laughter] But you can find, there are exceptions. I think that there are ways in which all sorts of–just the fact that communication moves more quickly now, that probably has helped that to some extent, in all the ways that it has influenced our lives. There is improvement in learning cycles based on the fact that that information flows more quickly. I don’t know that’s fundamentally changed the way we do the transition process as much as just sort of increase the velocity, but–
YS: I’d just like to jump in there. I think that in thinking about the technology development trajectory, one of the biggest, I think, predictors of how effectively you can do that is how much you understand about the materials you’re working with. You think about how we’re on a systematic plummeting price curve on solar cells. It’s because the vast majority of solar cell technologies occurs on silicon technologies for PN junction-type solar cells. Where the understanding of the underlying materials, the understanding of its defect properties, its transport properties, its interface properties, really were developed over many decades, well before what we now herald as a solar revolution. As we now want to tackle, and need to tackle, additional challenges in energy storage, interconversion, et cetera, we need to really ask ourselves, do we know enough about the materials that we’re working with to be able to predict down the line where the roadblocks may be? It’s not to say that we can’t get great new technologies without a full understanding. But I think that that development cycle and getting on the right learning curve is very much predicated on knowing the nature of the underlying properties and materials you’re working with.
FO: It does strike me clearly that today there are a broader set of tools being developed–machine learning, artificial intelligence, and so on–you see them popping up everywhere, and the question is, okay, well, is this a new set of kind of tools for the materials science space that we can leverage to support some of this accelerated work? In your experience, have you guys seen successful applications of some of these new types of technologies? At least in the more energy-centric technology development? Or has it been relatively limited?
YS: It’s still early days. I would say the successes that we’ve seen so far–and there have been successes–have been small steps in the right direction. It could be something like a quantitative structure property relation derived from taking a training set of materials and using that to develop a computer algorithm that might be able to predict what the best possible material might be. There have been successes there. These are, of course, very small steps. One needs to continue to grow that to be able to say, can you take all the genomic calculations, and have you really captured the underlying physics? As we mentioned, it’s a complex environment. A lot of times you can learn something off of a computer, put it into a practical system, and it tells you something completely different.
FB: I mean, I would add to that. I think that the efforts that have been made, obviously the advances in computational approaches to materials discovery and design have been very, very impressive, and there are very good demonstrations of that. I think that it’s, again, one of these questions of, the computer will not tell you an answer if you don’t provide the computer with the information that’s germane to the device or application or technology or catalytic reaction you’re doing. These tools are extraordinarily powerful but are down constrained on some level by the information we can feed into them. Whether that be about the properties of the material, the degradation pathways of a material. To even train those calculations requires a greater level of understanding at an experimental level of how these materials behave under these extremely unusual operating environments that we put them in when they are performing their work in electrochemical technologies and particularly energy technologies.
EO: I’ll just add another flavor to that point. When you think about artificial intelligence and machine learning and the ways in which it’s discussed and all the attention it gets, the order of magnitude of data that the tools are being developed around, when you think of the Facebooks of the world, the number of cat pictures on the Internet relative to the amount of fundamental data on materials, it’s a totally different ball game. You just have to remember that when you think about the reasons it’s early days. There’s obviously a lot of opportunity and a lot of excitement, but we still have a long way to go to have that underlying, not just physical understanding, of course, but also just the baselines and benchmarks to be able to make sure we’re building the right model. To a certain extent, that points to other opportunities in the algorithm development, too. You have this other set of problems. There’s materials science-specific challenges that need to be overcome, and so maybe that points to algorithm development, but it also echos the need for data and need for more fundamental understanding.
FB: I think this goes towards a topic that we keep coming back to, though in different guises, which is interfaces. To have the folks who are developing the computer algorithms need to be interfacing with the experimentalists and the subject matter experts to make sure that whatever you’re putting into the computer is germane and leads to an outcome that can actually be validated experimentally to see whether, hey, does this work or does it not?
FO: Tell me then, like, looking to the future a little bit, each of you have some really interesting work ongoing at the moment. From a longer term or larger goal perspective, what would you guys like to be able to achieve with your own personal research program? What is that question that you want to spend the next five, ten years really having impact on?
YS: I’d be happy to start in this area. We approach this related to the problem of the chemistry of interfaces, the chemistry of catalysis. In particular, the chemistry of how you use electrical energy to do chemical work, and how you rearrange chemical bonds with electrons. That’s the large, big picture goal. We really, at the very fundamental side of that, would like to write the rulebook on how one designs a material to achieve a particular chemical transformation in a directed way, in a selective way, in an efficient way. That rulebook simply does not exist. It does not exist because we don’t have enough of a fine-grained atomistic level understanding of how complex interfaces at catalysts’ surfaces really work. It’s our grand goal, which we’re going to chip away at over the next few years, and have been over the last five years, in writing that fundamental playbook that really needed to be built before we were able to make high-efficiency solar cells or do any of the other things that we now can do really quite well. We don’t have that rulebook for interfacial catalysis.
FO: Fik, how about you?
FB: That’s a really good question. The way I think about it is maybe a little level higher than Yogi and a little bit further down the innovation pipeline. What I’m very interested in is, as we transition towards a low-carbon economy and as solar and wind and other renewable technologies may begin to make up a larger fraction of the energy that we use, electrochemical systems are going to become increasingly relevant. As Yogi said, the science and engineering behind electrochemical systems is somewhat limited right now. We do certain things relatively well, but there are other transformations that we just don’t really have a good handle on. What I’m very interested in is trying to identify those opportunity spaces for electrochemical technologies to come in and to maybe allow a process that wasn’t possible beforehand, or to enable a more sustainable route to a particular chemical. I think there’s tremendous opportunity once you look beyond what we’re originally looking at right now, which is grid energy storage, in terms of industrial electrochemistry, being electrochemistry into the chemical plant. I think those are very interesting questions for a range of different disciplines, ranging from the chemist who’s interested in the fundamental transformations, to the engineer who wants to know, how would I build the reactor and how would I integrate it? To the economist who’s saying, okay, what’s the value-add at the end of the day? Those are areas where I think we’d like to be impactful over the next, I’ll be aggressive and say 10 to 20 years.
EO: It’s funny the way that Yogi said the rulebook on designing material because sometimes I say, part of what the group is working towards is a guidebook or a rulebook towards scaling the use of materials and the systems implications of doing that. I think that’s another analogy–maybe you did that on purpose, getting us in the room together. But this idea of trying to be able to anticipate the unanticipated consequences of how we develop these technologies, because I think we keep saying when we transition to a carbon economy. Obviously, the urgency of that is really great and we need to make sure we’re not doing that in a way that’s at the expense of other kinds of decisions and other kinds of problems we would link ourselves to that would either prohibit that happening at the scale that it needs to happen, or lead to other consequences that would put us in otherwise dire positions. That manifests in a few different ways, some of which I mentioned earlier. The supply chains that we link ourselves to and the materials that we’re using, the broader lifecycle implications of that, but also just linking the manufacturing piece to the materials development as early as we can in that development process so we identify dead-ends or areas needing further inquiry before we’re 10 years into the technology development.
FO: Of course, it was deliberate to bring you three into the room, but not necessarily for that particular point. I think when we were chatting about this particular podcast, there was a desire to have a group come together that could reflect a little bit about the holistic aspect and role of materials science. That spans everything, as you said, Yogi, from the molecular interface, right the way through to the broader macro integration, into processes, and then all the way through to the very macro kind of global impacts of all of this. And loops back around to say, we therefore need this kind of integrative approach. We need the rulebooks. We need all of those rulebooks. That perhaps is the game-changing opportunity. In this series of podcasts, we’ve had folks speaking about a number of quote/unquote “game-changers.” That’s been very interesting. That’s been innovations on photovoltaics and their deployment partnering, that’s been innovations on fusion energy. But actually, the more we thought about that, the more we reflected back on the more fundamental role that all of these innovations are linked to materials science and us really embracing materials science from end to end. I think that is what the real game-changer is. Now we’re really just stepping into a challenge, actually. Given the timelines we have, the timelines that we have to achieve success over, it’s really important that we get that coherence together to the greatest extent possible. I’m very fortunate to be able to work and learn from you guys and see the great work you’re doing. I think it gives me a lot of confidence when I’m speaking to others that I think we here at MIT at least will have a good stab at addressing this problem. With that, I’d like to thank Yogi, Fik, and Elsa for joining us today. It’s been absolutely fabulous, guys. Maybe we’ll talk again soon and see how we’re making progress on our guidebooks.
YS: Let’s do it again some time.
FB: Thank you.
FO: Show notes and links to this and other episodes are available at energy.mit.edu/podcast. Please tweet us @mitenergy with your questions and comments, and of course, show ideas. And do also please subscribe and review us where you get your podcasts. From the MIT Energy Initiative, I’m Francis O’Sullivan and thank you for listening today.