Below is the text version for the "Advanced Electrocatalysts through Crystallographic Enhancement" webinar held on July 31, 2019.
Eric Parker, Fuel Cell Technologies Office:
Good day, everyone, and welcome to the U.S. Department of Energy's Fuel Cell Technologies Office webinar series. This month we've got another great presentation from the Los Alamos National Laboratory on advanced electrocatalysts through crystallographic enhancement. My name is Eric Parker. I provide program support within the Fuel Cell Technologies Office, and I'm the organizer for today's event. We'll begin in just a second, but first, I have a few housekeeping items to tell you about. This WebEx call is being recorded, and the recording, along with the full slide deck, will be posted online in the near future. If you do not wish to be recorded, please disconnect now. If you speak during the call, you are presumed to consent to being recorded and to the use of your voice. That being said—if you could advance to the next slide, Jacob—all attendees will, however, be on mute throughout the webinar, so please submit your questions via the chat box as you see on the right-hand side of your WebEx panel. We will cover those questions during the Q&A at the end of the presentation. Again, the chat box should say Q&A. With that, I would like to introduce today's DOE webinar host, Nancy Garland, who is joining us at DOE headquarters. Hi, Nancy.
Nancy Garland, Fuel Cell Technologies Office:
I would like to give a brief introduction to Dr. Jacob Spendelow from Los Alamos Lab, today's speaker. Jacob received his bachelor's in science from Case Western Reserve University in 2002, and his Ph.D. in chemical engineering from the University of Illinois at Urbana-Champaign in 2006. He conducted research on electrocatalytic properties of model single-crystal surfaces there. He subsequently joined Los Alamos as the director's postdoctoral fellow in 2006, and he's been a staff member there ever since, since 2008. His research activities include work on catalysts, meso-structured electrodes, and diffusion media. Without further ado, Jacob.
Jacob Spendelow, Los Alamos National Laboratory:
Okay. Thank you, Nancy, for the introduction, and thanks to Eric for helping to set this up. Before I dive too much into the technical content, let me give you a brief introduction to Los Alamos National Lab, which is where I do my research. Los Alamos is located in northern New Mexico. It's part of the DOE national lab system. It's a very scenic area. The lab has been here since 1943. Originally, it was founded for the Manhattan Project. It's a fairly large lab, with a $2.7 billion annual budget and about 10,000 employees. Most of the lab budget has to do with national security work; however, we also have a large fuel cell program. It's actually the oldest fuel cell program in the national lab system. We've been doing research on fuel cells since the late 1970s.
Fuel cells are often seen as being a technology of the future. But at this point, they're really a technology of the present as well, because there are multiple fuel cell vehicles that are commercially available. You can see some examples of them here. Fuel cells have really transitioned in the past few years from being a laboratory technology to a real-world product. Now, that said, there are still technological improvements that could help to increase the level of deployment of fuel cells, and so that's largely what I'll be talking about today.
I'll give you a little overview of how fuel cells work—and in particular, polymer electrolyte fuel cells. That's the kind of fuel cell that we are working on at Los Alamos, and that's the main kind of fuel cell that the Fuel Cell Technologies Office does research on.
Polymer electrolyte fuel cells are defined by the presence of this polymer electrolyte membrane, which provides proton transport, and this membrane is surrounded by two electrodes. On one side, there is an anode, and on the other side, a cathode. On the anode side, we have oxidation of hydrogen occurring, while on the cathode, oxygen is reduced. These two reactions added together result in the overall hydrogen oxidation reaction to produce water and electricity. This electricity can be sent through an external circuit, where it can power a device such as a vehicle. Slow oxygen reduction on the cathode is the main barrier at this point to increased fuel cell performance, and it's one of the most important R&D areas for fuel cells. Due to the slow oxygen reduction kinetics, precious metal catalysts based on platinum and alloys of platinum are the state-of-the-art catalysts that are used to try to improve the oxygen reduction kinetics, but the cost of these catalysts is significant, and durability remains as a concern.
As I mentioned, cost is a big factor, which is limiting the rate of fuel cell deployment, and a lot of the fuel cell cost comes from these precious metal catalysts. The amount of cost due to the platinum varies with the amount of systems that you're producing. But, at high-volume manufacturing, it is projected that approaching half of the fuel cell stack cost would come from the expensive platinum catalysts that are used. In order to reduce this cost, there are a few strategies that you can employ. One is to design a better catalyst, which can have less platinum, and by using less platinum, you can reduce the cost. Another strategy is to try to get higher performance. If you can increase the amount of power that your fuel cell is producing from the same package, then you can make a smaller package. You can make a smaller fuel cell with less materials in it, and that will reduce the cost. You can see what a typical fuel cell performance polarization curve looks like here, with three primary sources of losses. The largest type of loss is activation losses, and these are basically kinetic losses that have to do with the slow oxygen reduction reaction on the cathode, and at higher currency also have increasing Ohmic losses and mass transfer losses. Reducing all of these losses is important to improve fuel cell performance and reduce cost.
Getting more to the specifics about the research we've been doing on trying to attack these problems through development of better fuel cell catalysts—our objective has been primarily to design new catalysts, which are based on nanoparticle alloys of platinum with base metals, in which these alloys are ordered to form intermetallic structures, and then we're supporting these intermetallic alloy nanoparticles on graphitized nitrogen-doped carbon supports. We're primarily looking at binary alloys, but we've also done some work on ternary alloys. We're mostly looking at cobalt, but we've also done some work with nickel. We're avoiding Fenton-active metals, such as iron, due to the known durability issues that these metals can cause. A lot of the work that I'm going to show you has been done using commercial carbon supports, but we have been developing nitrogen-doped carbon supports as well, and we've had some very encouraging results with those in the past year that I'll be showing you later in the talk. We're also working on not just doing the fundamental science with these catalysts, but also demonstrating them in high-performance, durable MEAs and scaling up to 50-square-centimeter level.
This work is working to address several of the critical DOE targets, most notably the mass activity target, so we're trying to get high oxygen reduction mass activity, and we're also trying to do that in a durable way, so we're trying to make catalysts that can meet the DOE target of less than 40% loss in mass activity after doing durability testing in an MEA. Another critical target that we are addressing is the power density target of trying to get to higher than 1 watt per square centimeter.
Our approach has to do with designing ordered intermetallic catalysts, and our primary material set that we're looking at is the L10 (L-one-zero) intermetallic, and these are also known as face-centered tetragonal structures. These are alloys of platinum with a base metal, which is typically cobalt or nickel. Rather than being random alloys, these are ordered structures. In a conventional platinum cobalt alloy catalyst, such as what's used in a lot of the fuel cell products on the market today, the platinum and cobalt are randomly distributed, and they randomly fill the same types of lattice sites in the crystal structure. In contrast, these ordered intermetallic structures involve separating the platinum and the base metal, such as cobalt, and putting them on different types of lattice sites, so all the platinum can sit on one type of site, and the cobalt on another type of site. By forming this ordered structure, we can see some advantages in performance and durability that I'll be showing you later on. Most of our work has been done with binaries but we've also done a little bit of work with ternaries that I'll show you, and we've also done a little bit of work with some different ordered structures that I'll also show later in the talk.
Our approach to synthesis of these materials involves using atomic-level ordering, and we do that by preparing ordered cores through a heat treatment process that enables the platinum and the base metal to occupy these different lattice sites. By having this ordered structure, we can stabilize the base metal within that core. Then, we can further protect the core with a platinum skin, and we can use theory and computation to guide the design of these nanoparticle structures, and then we can support these nanoparticles on iron-free, nitrogen-doped graphitic carbon to enable them to be used in fuel cells.
In terms of characterization and testing, we're doing a lot of MEA testing in this work. It's a very MEA-focused project with not very much RDE testing. We're integrating the nanoparticles into MEAs, testing initial performance and testing durability in MEAs, and then performing a variety of MEA diagnostics. We routinely do things like impedance and limiting current methods to characterize what the loss mechanisms are and to help guide the design of the electrodes. We also do a lot of characterization, both at beginning of life and after durability testing, including X-ray diffraction, and then a variety of spectroscopic and microscopy techniques to better understand the structures we're making and how the structures evolve in the MEA environment. We're scaling up MEAs to larger sizes and also scaling up the catalyst synthesis up to gram-scale batches.
We have several partners we're working with on this project, including Brown University and University of Pennsylvania. Our university partners are doing a lot of the fundamental research on the synthetic methods and developing the new materials. At Los Alamos, really our role is to try to adapt these syntheses to make them scalable and to make them more amenable for manufacturing, and to try to recreate the same sorts of performance and durability in a more industrially relevant setting.
There is also a support development aspect of this work, because the interface between the catalyst and the support, of course, is critical to determining the performance and the durability. We're partnering with the University of Buffalo on designing new carbon-based supports that have several key attributes, including doping with nitrogen to improve the dispersion and stabilization of the nanoparticle catalysts on the support surface, and graphitization to improve the durability of the carbon support. Another key attribute is that we're avoiding iron, again, to avoid the Fenton degradation issues associated with iron. You can see an example of one of the high-surface-area carbon supports, which has high graphitic content and high nitrogen content, developed in this work.
As I mentioned earlier, we're using theory and computation to help guide the design and guide our understanding of these intermetallic structures. One key aspect of these structures is trying to understand what makes them durable. To help guide our understanding, we've been doing DFT calculations with our partners at Brown University, and we've been using vacancy mediated diffusion as a model of the base metal diffusion and leaching in these intermetallic catalysts. What we found is that both diffusion barriers correlate strongly with the potential energy difference between just a few different states in the much larger diffusion process, and so that's a useful finding to help expedite these sorts of calculations. But in doing these calculations, we found that predicted diffusion rates for these intermetallic structures are orders of magnitude slower than for face-centered cubic platinum. If you look at the timescales on which this diffusion is predicted to occur, you would essentially conclude that on a device timescale, like for a vehicle, you wouldn't actually have significant diffusion rates. This actually might be part of the reason why the durability of these catalysts is as good as it is. However, we know that they're not perfectly durable, and we know that some leaching of base metals does exist, which suggests that alternative mechanisms such as oxygen place exchange might also be important in controlling the base metal leaching. Work is still ongoing in this area.
I'm going to move on now to talk about some of the experimental results, and I'll start with our work on fairly large particle L10 ordered platinum cobalt nanoparticles, which is what we started out working on in this project. You can see what these particles look like here. Again, they're fairly large as far as fuel cell catalysts go, with 9-nanometer particle size. Now later on, I'll show you some much smaller particles we developed more recently, but we started with these large particles in part because they're easier to study and understand. You can see what the local atomic-level structure looks like in these particles. This high-resolution dark-field STEM image is showing the contrast between the layers of platinum and cobalt stacked on top of each other within this nanoparticle catalyst. But if you look at the surface region, you can see that we have a shell of platinum coating the intermetallic platinum cobalt core. Now, when you look at XRD, you can see evidence of atomic-level ordering, which is present in these peaks that I've marked in green here. These are the super-lattice peaks, which show the existence of the L10 ordered structure. We've also done magnetic coercivity measurements that further attest to the ordering in these nanoparticles. Overall, the XRD, the coercivity measurements, and the TEM are all demonstrating that we have a high degree of ordering in these large particle catalysts, which essentially consist of platinum cobalt L10 cores surrounded by approximately 2-atom-thick platinum shells.
After developing these catalysts at Brown University, the group at Brown did RDE testing to study the performance for oxygen reduction and observed very high mass activity, and also very little loss in mass activity, during durability testing in an RDE environment. If you look at TEM images before and after the durability testing, you see no apparent changes in the overall structure. The particle size is still quite similar, and the composition also did not change very much after durability testing, so there was very little cobalt leaching that occurred.
These RDE results from Brown University were very encouraging. Brown sent their catalysts to Los Alamos for MEA testing, so I'll briefly go over the sorts of MEA testing that we do at Los Alamos. All the MEA testing that I'll be showing today uses fairly standard techniques, including ultrasonic spray to deposit catalyst layers on the membrane using conventional I/C ratios. We're using 29BC GDLs for all the testing, and Nafion 211 membranes for all the testing. Most of the testing I'll show is done in 5-square-centimeter differential cells that use high stoichs in order to have uniform conditions throughout the cell and further enable better understanding of what's going on in the MEA. Pretty much all of the work that I'm going to be showing uses an electrode loading of 0.1 milligrams of platinum per square centimeter on the cathode, and most of the testing was done at 150 kPa absolute pressure and 100% relative humidity, unless noted otherwise.
Here are a few of the tests that we standardly run on a catalyst as part of our MEA testing. Mass activity is measured by holding at 0.9 volts for 15 minutes in a hydrogen-oxygen environment and measuring the current at the end of that 15 minutes, so that's how we get mass activity. To do durability testing, there are two tests that we use. One is the catalyst accelerated stress test, and this has been defined by DOE to be a square wave between 0.6 and 0.95 volts with a half-second rise time, as shown here, so this test was done in hydrogen-nitrogen. We also sometimes do the support accelerated stress test, which involves a triangle wave between 1 and 1.5 volts, also in hydrogen-nitrogen.
We did some of this MEA testing using the catalyst developed at Brown, this large particle cobalt platinum catalyst. What we saw was very encouraging mass activity at the beginning of life, but, even better, we saw that after doing the accelerated stress test in the MEA, we only had about 20% loss in mass activity, which is much better than the DOE target of 40% loss. That said, we did see larger losses in the polarization curve at higher currents; however, we think this is mostly due to flooding of the electrodes with liquid water, because if you do the same test under drier conditions, you actually see very little change between the BOL and the EOL performance. That said, we are concerned about this loss in performance under wetter conditions, and we're also concerned about the low performance at high current densities. The power density is significantly below the DOE target, and we know some of the reasons for that. It has to do with how large these particles are, as well as how thick the electrode that we had to make for them was, but this was where we started, so our initial learning from this material was what guided the development of better catalysts, which I'll be showing you in the rest of this presentation.
Although we knew there were some issues with the high-current performance, we were still very interested in further studying this catalyst to better understand what makes it so active and so durable, so we did XRD testing at Los Alamos. At the beginning of life, we saw, again, these super-lattice peaks, which demonstrate the presence of the ordered structure, but we still saw the same peaks present after doing 30,000 cycles of durability testing in an MEA. That was a big result, showing that the ordered structure can exist even after extensive durability testing. We also observed that the peak positions did not shift very much after the durability testing, indicating that the lattice established by the ordered core isn't really changing very much, and we did see some change in composition with some cobalt leaching, but we were still keeping most of the cobalt in the nanoparticles even after the durability testing.
We sent these materials to our collaborators at Oak Ridge National Lab to do some advanced microscopy characterization. What we saw from that dark-field STEM and STEM EDS imaging is that we had roughly 1-nanometer platinum shells surrounding intermetallic platinum cobalt cores after doing the durability testing. Again, this is demonstrating that the core can survive and stay ordered, even after durability testing, with really the only change being that you have a thicker platinum shell forming around the core, and this is protecting the particle interior from further leaching. Now, this platinum shell is too thick for the cobalt in the core to be providing a significant ligand enhancement after the durability testing; however, the kinetic enhancements due to the strain are remaining, even after 30,000 cycles. This is the most likely explanation at this point for why the mass activity is retained so well by these catalysts. These results, I should mention, were published last year in Joule.
Summing up performance for this catalyst, it did very well in terms of the mass activity and the mass activity loss was very good; however, there were still some issues, especially with the performance at higher current densities, so the power density wasn't where it needed to be. We think that has to do with the very low ECSAs, due to the large particles used, as well as the thick catalyst layers used.
Then we moved on, and this is work that was done at Los Alamos. At Los Alamos, we moved on to start synthesizing smaller particles that are still ordered, still L10 ordered platinum cobalt.
Here is one of the initial results from synthesis at Los Alamos. You can see what the nanoparticles look like here on a Vulcan XC-72 support. One thing you'll notice is that the particle size distribution is wider than it was for the catalyst developed at Brown, and this is partly because the techniques used at Los Alamos provide less control of the particle structure; however, they are more manufacturable, which is a significant advantage. Despite a wider spread than we would have liked in the particle size distribution, we saw very good performance and very good durability, which really encouraged us. You can see that even though the particles are much smaller here, with most of the particles being 4 nanometers or smaller in diameter—despite the small particles, we can still see evidence of ordered particles in the XRD patterns.
We were interested in understanding this catalyst better, so, again, we sent samples to Oak Ridge for electron microscopy, and you can see some more images here showing a better view of what these particles look like. Again, we have mostly smaller particles. There are a few that are perhaps bigger than 5 nanometers but the majority of them are 4 nanometers or smaller, even down to around 2 nanometers, so these catalysts are more suitable for fuel cell applications due to the smaller particle size and the resulting higher electrochemical surface area.
We also did high-resolution dark-field STEM imaging, which enabled us to directly visualize some of these ordered structures. Again, you can see the ordering in this sort of striped pattern, where you have the alternating planes of platinum and cobalt within the nanoparticle. On the left side here, we have the catalyst as synthesized, and then on the right side, the same catalyst after we put it in an MEA and did 30,000 AST cycles on it. What you can see is that the original structure didn't change very much. We still were able to find these ordered particles after all the durability testing. In terms of STEM imaging, you can always hunt around and try to find a particle that looks the way you want it to, so you have to always take these results with a grain of salt. However, we supplement this by doing X-ray diffraction as well, which is a bulk-averaged technique. You can't just find the particle you like with this technique. It's really showing you what all the particles look like. When we do XRD, we can see that these super-lattice peaks showing the ordered structure are still present, confirming that this ordered structure can survive the durability testing, even in very small particles of four nanometers or smaller. We also did EDS mapping to look at the distribution of platinum and cobalt within the nanoparticles before and after the durability testing, and we saw very little difference before and after the test. Really, the only difference is that the platinum shell got somewhat thicker, but that platinum cobalt core, which was an ordered structure, is still intact after the AST.
This slide summarizes some of the MEA results from this catalyst, the small L10 platinum cobalt catalyst developed at Los Alamos. This catalyst does very well with respect to the DOE targets. It's meeting pretty much all of them, including the mass activity and the durability targets. The most difficult target for us at this point is the power density target. We're very close to meeting that, even at low pressures, but we still have a little bit to go there. However, when we increase the pressure to more real-world relevant conditions, like 250 kPa, we can get more than 1 watt per square centimeter with these catalysts. Part of the reason for the high performance of these catalysts has to do with the small particles and the resulting high ECSA. At beginning of life, the ECSA measured by CO stripping is more than 60 meters squared per gram. This is quite competitive with conventional catalysts, so this is definitely a high enough ECSA to provide the sort of surface area that you need for good MEA performance.
Even after durability testing, we still retained very good ECSA, attesting to the durability of these catalysts and explaining why the performance, even at high current density, doesn't suffer much after the AST. If you want to really dig in and figure out why are these catalysts providing the durability benefit, I think the most important thing to look at is the cobalt retention. That's really the difference that the ordered structure provides. In this ordered platinum cobalt, you leach a lot of the cobalt out of the catalyst during MEA testing, or during operation in the real world. So that cobalt leaves your catalyst, it's not helping your catalyst anymore, and now it's actually sitting in other places in your fuel cell where you don't want it to be, and it's hurting your performance elsewhere. Really, the advantage of these ordered catalysts is that they can retain high cobalt content and they don't lose much of their cobalt content during MEA testing. Even after 30,000 cycles in an MEA, they have much higher cobalt content compared both to a disordered platinum cobalt that we made at Los Alamos, as well as compared to commercial platinum cobalt catalysts.
The results with these platinum cobalt catalysts were certainly encouraging. We were happy with the performance and durability levels that we achieved, but we're never satisfied and we always want to get better. One of the ways that we've been trying to do that is by starting to incorporate additional elements into these ordered nanoparticles, so moving beyond binary catalysts to ternary catalysts. Again, our collaborators at Brown University did a lot of the initial work in this area, starting with some DFT calculations, which predicted that adding a third component, such as nickel, to the ordered platinum cobalt could provide improvements in the binding energy of some of the oxygen reduction intermediates. On the experimental side, the Brown group synthesized a variety of nickel platinum and cobalt platinum, as well as ternary cobalt nickel platinum catalysts, and tried to form ordered structures by annealing. What they observed was that for the pure nickel platinum, it was very difficult to form any ordered structures. You don't see any of those super-lattice peaks here in the XRD patterns. But, as you add more cobalt to the material, the ordering starts to become apparent, and so this 50-50 cobalt nickel catalyst was able to have good ordering evident in the XRD patterns. Based on the DFT and XRD results, this ternary cobalt nickel platinum catalyst looked quite promising.
Also, based on RDE results at Brown, it looked promising, so the oxygen reduction halfway potential measured in RDE was the same or even a little bit higher than the binary cobalt platinum catalyst. We were excited about this catalyst, and we did testing at Los Alamos in MEAs. However, at this point, the MEA performance really isn't measuring up to what we would expect based on the RDE performance. This is a common problem that I think most people who are active in this field are aware of, the challenge of trying to reproduce good RDE results in the MEA environment. Despite the good RDE performance, the mass activity in the fuel cell wasn't meeting the target, and the performance at high currents wasn't very good either. Now, we're certainly far from giving up on these materials based on our theoretical understanding, as well as the RDE tests—we think there is a lot of promise here—but there is a lot of work to be done in trying to translate that still to the MEA environment.
Most of the work that we've been doing has been on the L10 ordered catalysts, but we've also done some work on L12 catalysts, and this has mostly been done by our collaborators at University of Buffalo. You can see one of the catalysts that they synthesized here, which is actually based off of their PGM-free cobalt ZIF-based catalyst. They made their cobalt ZIF catalyst, but then they added platinum salts and did heat treatment to form platinum cobalt nanoparticles, and then further heat treatment to form ordered platinum cobalt on the nitrogen-doped carbon surface. You can see some really fantastic dark-field STEM images here showing this ordered structure, which was very beautiful and we're very excited about it, so the Buffalo group sent us some of these catalysts to do MEA testing at Los Alamos. We did see encouraging performance with this L12 similar to what we saw earlier with the L10 structure. Results from this study were published last year in Nano Letters.
Another type of support that our collaborators at Buffalo have been developing is this hydrogel-based support, which is made by taking a polymer hydrogel, which is typically based on polyaniline, and adding manganese salts and then doing a heat treatment. This manganese can help to catalyze the formation of graphitic structures in the resulting carbon support. You can see here what these carbon supports look like. They're essentially stacked planes of nitrogen-doped graphene, which are folded tightly on themselves forming this porous, highly graphitic and nitrogen-doped structure.
These hydrogel-based supports were used—so we deposited platinum on the supports and did RDE as well as MEA testing. I'm showing here some of the MEA testing results from Los Alamos. In addition to just looking at the performance and durability, we also did some characterization by looking at the carbon corrosion rate during accelerated stress testing. This test was the support AST, which involves rapid cycling at high potentials between 1 and 1.5 volts, and we measured the carbon corrosion by looking at the amount of carbon dioxide coming out of the cell using NDIR. What we saw is that if you compare to a conventional platinum on Vulcan catalyst, the platinum on this hydrogel catalyst has much, much lower carbon corrosion rates, and we think that's due to the higher graphitic content of this hydrogel-based structure. As a result of the lower carbon corrosion rates, the MEA durability was much better than the Vulcan support, so the Vulcan support lost most of its performance following 5,000 cycles, whereas the hydrogel support stayed very durable. We're encouraged that this hydrogel support can enable us to meet the DOE support durability targets; however, we're certainly not meeting the mass activity or the high current density power targets with this support. Now, that said, we wouldn't expect to from this test because this was just a platinum catalyst. There was no cobalt or anything else to improve the performance of this catalyst, but we wanted to look just at platinum on this catalyst before we moved on to platinum cobalt. I'll show you platinum cobalt results in the next few slides, but I did want to mention that actually just earlier this month, our manuscript on this study was accepted for publication in Energy and Environmental Science.
After seeing the great results with the platinum on the hydrogel support, we moved on to start adding cobalt to this material. Again, we did the seed-mediated approach that we've developed at Los Alamos, so by adding cobalt salts and then doing a high-temperature heat treatment, we were able to form these platinum cobalt nanoparticles. You can see here in the STEM image that the particle size distribution is now smaller and narrower than what I showed you earlier for the XC-72 support, so we can make a better dispersion of platinum cobalt nanoparticles by using this new support developed by our collaborators at Buffalo. We also zoomed in to do some high-resolution STEM imaging. Again, we were able to see clear evidence of ordered particles in this catalyst. You can, again, see these stacked planes of platinum and cobalt on top of each other, even for some really small particles, even particles down to 2 nanometers, we can still see ordered cores within these particles.
We did MEA testing on these platinum cobalt on hydrogel support catalysts and saw encouraging results. We saw very high mass activity. This is actually the highest mass activity we've measured up to this point, and that's probably due to the smaller particles and the better dispersion that the support provided. We also saw that the durability was still quite good, meeting or basically meeting all of the DOE durability targets for catalysts in MEAs. The high-current performance was a little bit lower than what we saw for the XC-72. It was still, I would say, pretty good but not quite as good, and, clearly, some more work needs to be done still to get to this 1 watt per square centimeter target. But we haven't done any MEA optimization on these catalysts yet, so I think there is still a lot of opportunities for getting better results, even without any tweaks to the catalysts. Also, the ECSA for these catalysts is very high, and this is, again, due to the smaller particles that were enabled by this new support. Overall, the combination of the Los Alamos platinum cobalt intermetallic technology, and the Buffalo support technology is providing really good mass activity and really good durability, and so we're excited about continuing the study of these new catalysts.
Most of the results I showed you up to this point were from fairly small synthesis batches, typically around 100 to 200 milligram batch sizes. We started doing gram-scale synthesis early this year, and the initial gram-scale synthesis results have shown similar performance to what we get for the smaller batches. The performance is slightly lower, but it's fairly close, and we're further optimizing the synthesis and the scale-up of the synthesis to try and improve that and better understand how to translate these catalysts to larger batch sizes. I also want to note that the intermetallic L10 cobalt platinum that we've been developing in this project is compatible with a lot of different carbon supports, and you can see that in the polarization performance here. We actually have three different kinds of carbon supports here, and the performance is really quite similar for all three: the solid Vulcan carbon, for porous carbon, as well as for that hydrogel carbon that I showed you previously. The fact that our catalyst technology can be integrated with all these different supports shows how flexible and how robust this approach to synthesis is.
Summing up our status versus some of the critical DOE targets, for the two most promising catalysts that we've developed so far, both are doing quite well versus the DOE targets. The best one at this point is still the L10 platinum cobalt on the Vulcan support. However, we have been able to get even higher levels of mass activity using the novel supports that we've developed in this project, but we still have some work to do on these to try to improve the power density.
All the results that I've shown you came from a three-year FCTO-funded project. That project is coming to an end this September, so all the future work here is somewhat tentative at this point. But if we are able to find additional funding to continue this work, there are several promising areas that could provide significant further improvements. One of those areas that we're interested to explore is to increase the high-current performance and durability by improving the dispersion of these L10 platinum cobalt nanoparticles on the support. We can do this by implementing some further improvements to our synthetic process, as well as improving the support structure. By tailoring the nitrogen doping—we've already demonstrated that we can improve the dispersion by adding nitrogen doping, and by further tailoring it, we think that we can get even further improvements in dispersion, which will provide better mass activity and better durability, as well as better performance at high current density. Most of our results so far have come from binary platinum cobalt catalysts, but we see a lot of promise in moving to ternary systems, especially by adding other elements such as nickel. Preliminary results have been promising in this area, but there is a lot of work that needs to be done to translate the preliminary RDE studies to the MEA environment, so that's an area we're interested in further exploring.
Also, looking more at the synthesis and the manufacturing of these materials could be critical, because there have been questions about how compatible is the synthetic approach with high-volume manufacturing due to the high-temperature treatment; how long do you need to do that treatment; what will that do to the manufacturing throughput? In our lab work, we've been doing fairly long annealing times, which would not be suitable for manufacturing, but we've seen evidence that we can substantially reduce the annealing time, and so I think there is a lot of interesting, fundamental work to be done on understanding the ordering process that could lead to substantial acceleration of the ordering, which would further reduce the manufacturing cost of these materials.
Lastly, in terms of future opportunities, everything that I showed you was sort of developed with an eye toward light-duty vehicle applications; however, recently there has been burgeoning interest in heavy-duty vehicle applications of fuel cells. Arguably, the intermetallic platinum cobalt that we've been developing in this project is even more applicable and even more relevant to heavy-duty. That's because in heavy-duty applications, durability is even more important than for light-duty, and also efficiency is even more important. Those are the two areas that our intermetallic platinum cobalt really shine in, because conventional disordered platinum cobalt suffers from high amounts of cobalt leaching. This cobalt—it leaves the catalyst, so it's not providing a performance enhancement anymore, and it's even hurting you because it goes other places in the fuel cell where you don't want it to be. By keeping more of the cobalt within the catalyst, we can improve the durability and, by retaining high mass activity, we can improve the efficiency. So, we're very interested in heavy-duty applications for these catalysts in the future.
In summary, our intermetallic platinum cobalt catalysts are providing high activity and high durability in MEAs. We've found that the ordered catalysts, which are surrounded by platinum skins, can retain high cobalt content even after extensive durability testing in MEAs, and L10 ordering is still apparent in these catalysts, even after doing 30,000 voltage cycles in an MEA. The best catalyst that we've developed can meet the DOE performance and durability targets, although there is still room for further improvement in that area. Overall, I also want to emphasize how important MEA testing is to evaluate oxygen reduction catalysts. Los Alamos National Lab is a world leader in MEA fabrication and testing. There may be people on this webinar who are developing catalysts themselves. Maybe you don't have access to MEA testing capabilities. We're always happy to collaborate, so if you have a promising catalyst, you're welcome to contact me by email and we can talk about possibilities for testing at Los Alamos.
I would like to acknowledge our funding agency, so support from the U.S. Department of Energy through the Fuel Cell Technologies Office is what has enabled us to do all this work, and I especially want to thank Dr. Nancy Garland, who has been the manager for this project over the past three years, as well as Dr. Dimitrios Papageorgopoulos, who is the fuel cells team leader. I also want to acknowledge and thank all of my collaborators at Los Alamos, as well as at our partner institutions who have done the work that I showed you today.
With that, I would be happy to take any questions. If you submit those through the WebEx, I'll do my best to answer those. Thank you very much for your attention. If you want to stay in touch, here is some contact information. Thanks again.
Eric Parker:
Thanks, Jacob. With that, we do have some questions in already, and feel free to submit more as you think of them. I'll let Nancy lead our Q&A session. Thanks, Nancy.
Nancy Garland:
Okay. Thank you, Eric and Jacob. First question, per Slide 15 for the L10 sample, the question is, "Is the platinum shell formed from the acid leach or from some other method process?"
Jacob Spendelow:
Oh, okay. Yeah, it's formed by acid leaching. After the initial synthesis and the initial heat treatment to form an alloy structure, we then do an etching in a mild acid solution. Typically, it's something like 0.1 molar perchloric acid. After a few hours, we wash and dry the catalyst, and then we do a further heat treatment to heal the surface and form a smooth platinum skin surrounding the platinum cobalt core.
Nancy Garland:
Okay. Thank you. Second question, they're asking, "What deliquescent materials are used inside MEAs to reduce moisture levels, which could be harmful in MEAs?" The second question, "Does using such material damage MEAs in any way?" They're asking about deliquescent materials.
Jacob Spendelow:
Yeah, so that's getting a little bit outside my area of expertise. We're really focusing on the catalyst development. But the question about moisture content in MEAs is a good one because water in MEAs is critical both to providing good performance but it also plays a role in the durability. Without water, your MEA actually doesn't have any performance at all, because you need water for proton conduction in the polymer electrolyte. However, if you have too much water, multiple problems can result. One is you can flood your electrodes, and I showed you an example of where that happened for the large particle platinum cobalt earlier in the slides. That's one problem. Another issue is that higher water content can accelerate catalyst degradation because some of the degradation mechanisms involve essentially dissolving the metal ions and having them be solvated by water. Reducing water content is one approach to try to improve durability, and there has been work on that by doing things like incorporating hydrophobic agents within the MEA. Frankly, I'm not going to be able to say a lot about that because that's work that other people have done, but there is literature about that if people are interested in further exploring that.
Nancy Garland:
Okay. Thanks. Okay, third question. "How does the hydrogel support surface area compare to Vulcan support surface area?" There's a follow-on question, "Does CO2 production correlate with carbon surface areas?" The first one comparing surface area for hydrogel support, and then with a Vulcan support.
Jacob Spendelow:
Yeah, those are both really good questions. The hydrogel support has significantly higher surface area than Vulcan carbon. If I remember right, Vulcan carbon is typically, what is it? It's a little over 200 meters squared per gram, and the hydrogel is something on the order of 600 meters squared per gram, if I remember right. Those details are in the EES paper, which was just published earlier this month, but they're pretty close to that, to those numbers. Now, in terms of how that surface area translates to the CO2 evolution, despite having higher surface area by a factor of three or so for the hydrogel, we actually see significantly lower CO2 evolution rates. In this case, they're not correlated, and that's because the hydrogel support has higher graphitization than the Vulcan support.
Nancy Garland:
Interesting. Okay, next question. The first comment, "Very impressive presentation. One quick question on Slide 20. The blue curve, Slide 20, what is the first peak 2-theta 15—approximately 15?"
Jacob Spendelow:
Oh, so actually there are peaks here, which we're not always sure what all of these peaks are, but there are peaks that are associated with the ionomer, and peaks that are associated with the carbon support. The super-lattice peaks are the ones that I've labeled in green here. There is actually one super-lattice peak also here that's overlapped by other peaks, so I didn't put an arrow there. But, yeah, these peaks don't actually come from the platinum cobalt.
Nancy Garland:
Okay, and then the first questioner says, "Thank you," and then the next question, "You didn't talk about industrial partners. Is it too soon? I guess you did also talk about applications with heavy-duty vehicles. Is it too soon, or what do you see there?"
Jacob Spendelow:
Yeah, that's also a great question. We have been collaborating with one industrial partner, and I didn't really talk about that much here, but we have been working with IRD Fuel Cells. We recently sent them some of our material for testing. We haven't received those results yet, so I can't show them at this point, but they are working with us on this project. In terms of other partners, we've had interest. Of course, I'm not going to name names, but we've had several developers who have contacted us about wanting to test these materials. Part of the problem up to this point has been the batch sizes, that we were mostly making very small batch sizes and they weren't sufficient to give material to these potential partners. Now we've made progress on that recently with our gram-scale synthesis, so that's one of the areas that we really want to do more, is start putting this material in the hands of others and start spreading it around outside of Los Alamos.
Nancy Garland:
Sure. Okay, and then the last question has to do with cost. "Have you done any analysis for the cost?"
Jacob Spendelow:
Yeah, so that's a great question. The short answer is we haven't done any formal cost analysis but it's something that we would like to do, and that's something that—Los Alamos isn't going to do that on its own, but, through the DOE program, we have been talking about the possibility of having somebody like Strategic Analysis help with doing some techno-economic analysis on the ordering process, and figuring out how much the thermal treatments we're using, how much those would contribute to manufacturing costs, and then figure out—that could really help to sort of establish what the envelope is that we should be looking in, like what is a reasonable annealing time to not have a significant impact on the material costs? That's something that we would love to do in future work.
Nancy Garland:
Okay. Great. I think we've got one more question. "Can private or public commercial organizations test MEAs at Los Alamos, or is it protected by government security clearance?"
Jacob Spendelow:
Oh no, we're happy to collaborate with people. I mean, usually we do testing that's more along the lines of people send us catalysts and we put it in an MEA ourselves. Now, if somebody has an MEA that they've made on their own and they want to test it, we don't usually do it that way but there's no reason we couldn't. Yeah, I mean anybody that wants to explore collaboration is certainly welcome to email me, and I would be happy to talk about what our capabilities are and how we could work together.
Nancy Garland:
That's great. Okay, that's all the questions. Anyway, thank you very much.
Eric Parker:
Yeah, thanks, Nancy, and thanks for everyone who asked questions. Like I said before, if you think of any later, please feel free to email them directly to Jacob or myself. If you want to advance to the last slide again and leave that up for everyone while I close out today's presentation.
In conclusion, a big thank you to everyone who joined today, and our presenter for that awesome presentation, but that does conclude our webinar for today. If we didn't get to your question, please feel free to email the FCTO webinar mailbox, like I said. I would also like to remind everyone that the full recording, along with this slide deck, will be posted online at the FCTO webinar web page in the coming weeks. I also encourage you to sign up for our newsletter, where you will hear about upcoming webinars like these and other really interesting topics. With that, I would like to wish everyone a great rest of your week, and goodbye.