Below is the text version for the "ElectroCat Consortium: Advancing PGM-Free Electrocatalysts for Next-Generation Fuels" webinar, held on September 26, 2018.
Eric Parker, Fuel Cell Technologies Office
Good day, everyone, and welcome to the US Department of Energy’s Fuel Cell Technologies Office Webinar. We’ve got a great presentation this month from Deborah Myers and Piotr Zelenay on the Electrocatalysis Consortium, or ElectroCat. My name is Eric Parker. I provide program support within the Fuel Cell Technologies Office, and I’m an organizer for today’s meeting. We’ll begin in just a minute, but I have a few housekeeping items I want to tell you about. Today’s webinar is being recorded, and the recording along with the slides will eventually be posted online, and we’ll be sure to let you know. All attendees will be on mute throughout the webinar today. So please submit any questions you have regarding the content of the presentation for our hosts here in the WebEx chat option, and we’ll be sure to ask them at the Q&A portion at the end of the presentation.
And with that, I’d like to introduce today’s DOE webinar hosts, Dimitrios Papageorgopoulos, who is joining us here at DOE headquarters.
Dimitrios Papageorgopoulos, Fuel Cells Technologies Office
Hello everyone. And for those who don’t know me, I am the program manager for fuel cells within the Fuel Cell Technologies Office at the Department of Energy. So it’s my pleasure to introduce today’s speakers. First of all, Dr. Piotr Zelenay who is currently Los Alamos Laboratory Fellow, as well as project leader and team leader at Los Alamos National Laboratory, focusing primarily on fundamental and applied aspects of polymer electrolyte fuel cell science and technology, electrocatalysis and electrokinetics. Dr. Zelenay has published over 180 research articles, many in renowned scientific journals, including Nature Science, Chemical Reviews, and Accounts of Chemical Research, and has co-authored over 400 presentations.
He also has 21 patents. Our other speakers, because they’re presenting together, is Dr. Deborah Myers, who is senior chemist and leader of the hydrogen fuels and materials group in Argonne Chemical Sciences and Engineering Division. She has been working at Argonne for 29 years and has been the leader of the hydrogen fuels and materials group for 16 years. Her research focuses on materials development, characterization, and cell design, or polymer electrolyte fuel cells and electrolyzers, and she also plays a critical role in one of FCTO’s consortia that focuses on advancing fuel cell performance and durability, FCPAD. She and Dr. Zelenay are co-directors of the Electrocatalysis Consortium known as ElectroCat. And with that, I’d like to hand it to them, and take it away, fellows.
Piotr Zelenay, Los Alamos National Laboratory
All right, thank you very much, Eric and Dimitrios. It’s a pleasure to be here and to start this webinar with actually an acknowledgement that Electrocatalysis Consortium represents an early stage research and development effort funded by a DOE Fuel Cell Technologies Office in the Office of Energy Efficiency and Renewable Energy, EERE for short. Next please. All right, so DOE programmer has been primarily targeting development of fuel cell power systems for automotive transportation with the goal to afford these systems to be competitive with internally combustion engines.
Cost analysis for platinum group metal-based fuel cell stack is performed annually, assuming high volume production of fuel cell systems at a level of half a million per year. If we’re such a prediction is performed, one can – we realize that approximately 40 percent of the fuel cell costs start is due to the use of platinum group metal catalysts at the cathode and an anode of the fuel cell. Predominantly at the cathode. When we then remember that fuel cell stack represents approximately half of the cost of the fuel cell system, then a simple calculation reveals that between 20 and 25 percent are of the fuel cell system cost is due to the use of PGM-based catalysts in the fuel cell stack. As a result, when we then move to the spider diagram on the right inside of the view graph, we are noticing the bottom that the technology has been trailing the cost target with over $40.00 per kilowatt, with the ultimate target being $30.00 per kilowatt of generated power.
Similarly, there is a gap to cover here as far as durability over for 5,000 hours of stock operation is concerned. The other performance targets, such as peak energy efficiency by density specific power and start up from low temperature of minus 20 degrees C, and have there been largely met by now. But the durability and cost remain important challenges we need to address. And as part of addressing these challenges, electrocatalysis consortium or ElectroCat for short, was created in February 2016 as part of energy materials network, which comprises quite a few different materials related consortium, several of them relevant to fuel cells and hydrogen economy in general.
So the primary goal of the consortium is to accelerate the deployment of fuel cell systems by eliminating the PGM catalysts, at least eliminating at the cathode. Next please. So in the next slide, we’re looking at technical targets for electrocatalysts for transportation applications. This table has been taken from multi-year research and development and early research, development and demonstration plan, published by DOE. It is slightly modified to show a comparison between the status of PGM free and PGM catalyst as well as 2020 performance targets.
In particular, when we look at the highlighted in blue line—referring to the performance of the fuel cell packed voltage of .8 volts—we notice that the PGM free catalyst generate performance which is approximately one-third of that flow generated by PGM catalyst, which in turn have largely met the performance target for 2020 performance target. Then they’re going down to next – to highlight the blue line. We’re looking at loss in performance at .8 per centimeter squared in millivolts. PGM free catalysts suffer from – still suffer from significant performance loss, and probably significant more than 50 millivolt at .8 per centimeter squared, whereas the corresponding number for PGM-based catalyst is only eight. Now we have PGM free catalyst activity target, which unlike for the PGM-based catalyst is expressed in terms of aerial – the current density, when for fuel cell operated on oxygen at [inaudible] performance measured at .9 volt after IR correction.
And this target – the ultimate target is 44 milliamps per centimeter squared, .9 volt, which actually corresponds exactly to the mass activity target established for PGM and base materials. And last year, the status – achieved status was 21 milliamps per 30 meters squared, approximately half of the 2020 target. In a week from now, we have another performance target of 25 milliamps per centimeter squared. Next slide, please.
Deborah Myers, Argonne National Laboratory
So the mission of ElectroCat, which was established in 2016, is to develop and implement PGM free catalysts, as Piotr mentioned. We’re focused primarily on the polymer electrolyte fuel cell and the cathode catalyst, but the broad mission of ElectroCat is to address all types of fuel cells in the FCTO portfolio that utilized PGMs that contributed to the cost of those systems, including phosphoric acid fuel cells, alkaline membrane fuel cells, and other types of fuel cells, in addition to the oxygen reduction reaction within our mission is also the hydrogen oxidation reaction and alkaline membrane fuel cells.
So that’s our broad mission, but within this first period of ElectroCat as Piotr discussed, we’re primarily focused on the cathode catalyst because that can make the most impact in terms of cost of the transportation fuel cell system. So in the beginning of ElectroCat in 2016, the ElectroCat steering committee comprised of Piotr, myself, and DOE managers solicited capability inputs from all the national laboratories, and those were down selected then to provide some core capabilities, and these capabilities are updated every fiscal year. So within those capabilities, we identified useful ones as the National Renewable Energy Laboratory and Oak Ridge National Laboratory, and these are primarily in characterization and implementation of the PGM-free catalyst in the fuel cell environment.
So the founding structure of ElectroCat and in view of all the energy material network consortia is to rely on atomic scale and mesoscale models to define the atomic level reaction kinetics, and also to look at, for example in the case of ElectroCat, the electrode structure and its impact on the fuel cell performance. Another founding principle of the [inaudible] is to utilize high-throughput techniques to synthesize and rapidly screen materials for their usefulness, and this is to accelerate the development and implementation of material in the electrodes and in the membrane electrode assemblies, and ultimately in the stack. And another founding principle is to aggregate all the data coming out of the consortium into an easily searchable and public database.
And this will allow the community and the consortium to develop and implement catalyst materials in the fuel cell environment. At the bottom of this slide, there’s a listing of the various capabilities at Los Alamos, Argonne, NREL, and Oak Ridge. Los Alamos is primarily responsible for PGM free catalyst development, electro chemical and fuel cell testing, and the atomic scale DFT type modeling, which you’ll see today. Argonne, their primary responsibility is high-throughput techniques in the synthesis and screening of the activity of materials and implementation in the fuel cell environment. Also a mesoscale model, so these are models of the electrode structure to define what is needed in the electrode structure to improve the performance. X-ray characterization studies using the advanced photon source here. And then also aqueous stability studies to look at the degradation mechanisms of the catalysts.
NREL is primarily involved in advanced fuel cell characterization or operando characterization of the materials and electrodes, and high throughput electrode fabrication and testing. And Oak Ridge in advanced electron microscopy, atomic level characterization using those techniques and x-ray photoelectron spectroscopy studies. Next slide, please. So at the beginning of ElectroCat in 2016, we established a website to sort of collate and advertise the capabilities of the four national laboratory partners, and these were organized in terms of their functionality, so we have three broad categories of synthesis, processing, and manufacturing. Characterization and testing and computation modeling and data management. Within each one of these three broad categories, we have sub-categories of how these techniques are applied.
For example, in synthesis, we have high surface area catalysts, model system synthesis, and fabrication of electrodes, and membrane electrode assemblies from the PGM free catalysts and so forth. So if you’re interested in utilizing these capabilities, I encourage you to go to the website, which is listed at the top of this slide and look through these capabilities, and it is possible then for you to utilize these capabilities through various agreements with the national laboratory. Next slide.
Piotr Zelenay
All right, we would like to start detecting a part of this presentation with highlighting two paths for catalyst development – PGM free catalyst for oxygen reduction, and starting with a catalyst that uses cyanamide, PANI, iron, carbon, zinc. Acronym for like we’ve been for some time pursuing the high temperature path and catalyst development, and this the first example is actually in that particular category. In addition to the transition map precursor, the main one of iron, tetrachloride in this case, we are using zinc chloride as a performer and two nitrogen [inaudible] cyanamide, then polyaniline with polyaniline being the main precursor for the active site together with iron, and cyanamide being another performer in the synthesis.
So we are using high temperature treatment in typically the range of 950 to 1,100 degree C in an inert atmosphere to generate catalysts where at [inaudible] they are tested. Both in [inaudible] electrolyte via rotating [inaudible] and then fuel cells. This has been a best performing catalyst in terms of activity for us, and this is the one that allowed us to meet the activity target last year of – activity target was 20 milliamps per centimeter squared .9 volt. We talked about it briefly a few slides back, and the same catalyst can demonstrate 105 milliamps per centimeter squared at .8 volt when the fuel cell is operated in there. And that actually represents an improvement of close to 30 percent of a performance demonstrated in June 2017. Next slide please.
So through our research, we’ve been able to correlate performance of PGN-free catalyst with the presence of iron sites or nitrogen coordinated iron sites within the carbon matrix. This is not a direct evidence that of FeN4 or FeNX in general to be an active site, but it’s a strong indication that we can reach high performance only when such sites are present in the catalyst phase. Through that research, we’ve realized that by increasing this version of iron to the atomic level, we are able to improve the performance, and that was the rationale behind the developing catalyst, so called atomically disbursed catalysts, such as the one shown in the bottom half of the slide, which is labeled as AD in parentheses, iron, nitrogen, carbon catalyst. This is the metal organic framework, the right catalyst from the precursor we call ZIF-F, F standing for fiber. That is after the predominant morphology, [inaudible] morphology observed in the precursor, and actually to a larger degree in the catalyst after the heat treatment. The initial morphology has been preserved in this particular catalyst after the heat treatment, and in most of these catalysts, the dominating component is carbon of 90 atomic percent content.
Then we have sub-one percent content of iron, and single percentages of nitrogen and oxygen with virtually no thing detected at least in the samples shown. What is important is we are not seeing iron rich nanoparticles, so which we do see for the catalyst, they’re shown in a previous slide, which indicates very good or more accurately atomic dispersion of the catalyst. Next one please. All right, the TM characterization performed at Oak Ridge does indeed reveal high level of atomic dispersion, and also like in the other catalyst shown two slides ago, good correlation between the presence of iron and where the presence of nitrogen. Collocation of these two elements. The challenge with atomically disbursed catalyst has been to reproduce high activity measured to rotating [inaudible] study in fuel cell environment.
And we’ve been gradually improving the performance of fat catalysts in fuel cells with progress between 2017 AMR and April 2018 highlighted in the graph on the right-hand side. Now through linear combination state of [inaudible] results at Argonne, we’ve been able to correlate activities measured via halfway potential in rotating [inaudible] collectors testing with the presence of FeN4 in the catalyst phase, that’s a small graph in the lower left corner of the [inaudible] graph. Next one, please.
Deborah Myers
As mentioned earlier, one of the core activities within ElectroCat for Argonne is mesoscale modeling of the electrode structure, and those mesoscale models rely on nanocomputed, X-ray computed tomography data taken at the advanced photon source as shown on this slide, and once acquiring that data and it’s segmented using kind of pained various properties of the electrode – for example, the distribution of solids within the electrode structure, the distribution of port, and from those distributions then calculations can be made on transport properties of those electrodes, such as shown in the upper right plot, the shortest path to tortuosity in the electrode structure in the three different directions of that electrode structure.
And using those data, then we can determine then which properties need to be changed within that electrode structure to improve their hydrogen air performance, and that’s what is shown in the three plots on the upper left to lower right where we look at the effect of the electrode fitness on the hydrogen air performance, the electrode tortuosity, and the micropore size and volume fraction of those micropores. And as you can see from those plots that there is a profound influence of electrode tortuosity on the hydrogen air performance. And so that says if we can then decrease this tortuosity, we can then improve the performance of the fuel cell, especially in the high current density region.
Other properties that can then be tweaked, and this is for the atomically dispersed catalyst, is to look at changing the micropore sites. So decreasing the micropore sites then improves the transport properties in high current density region. As may be intuitive, if you decrease the electrode thickness, you improve the transport properties of that electrode, but with the caveat that you have to maintain the same volumetric active site density, which is a challenge with these catalysts. Next slide, please. So also one of the main activities that Argonne is to define routes, which can enhance the active site density of the materials. And we do this primarily by using the combinatorial setup, but also using institute x-ray characterization. So what is shown on this slide is the institute x-ray characterization of the precursors for the atomically dispersed catalyst as a function of heat treatment temperatures. So this is – these are measurements taken while we’re heat treating the samples, and a parallel study was done at Oak Ridge National Laboratory with Institute TEM to look at the morphological change of the catalyst during the heat treatment process.
And so what is summarized on this slide is the Iron K edge absorption data as a function of temperature, and if you look in the upper right plot, you can see that from the absorption in this energy range, you can determine how much zinc is in the sample. From the edge step, you can then determine how much iron is in the sample, and this is all as a function of temperature. So highlights from these x-ray absorption data are that in the low temperature region, in the precursors, in this particular precursor, we saw that the iron head and Fe203-like coordination structure, which then as we started increasing the temperature of the precursor, the iron precursor was reduced to something more akin to Fe2 plus.
There was a formation of particles as shown in the TEM data, and there was coordination of the Iron 2 sulfide to form an iron sulfide like structure in the 500 to 550 degree C range. As we started increasing the temperature above that range, we saw the formation of a carbide phase, and increase in the material density as shown by the absorption at 7.6 KEV. And then as we transition to the higher temperatures, this iron carbide like phase transitioned to what we believe is the active phase, which is the FeN4 like coordination. And if we look at the upper right plot, concurrent with the formation of the FeN4 coordination, what was loss of zinc.
If we look at then the TEM data, we see that the zinc is forming particles, and those particles are evaporating in that temperature range. At the higher temperatures, we see then if we hold at, for example, 1,000 degrees C, we can see the conversion of the FeN4 active sites back to the carbide and iron metal. So the bottom line from these studies was if you want to enhance the active site density, you can go to pyrolysis temperatures as low as 900 degrees C, which is also sufficient to remove the zinc, and if you have higher iron content than the precursors, you want to avoid these very long holds at high temperatures because this then starts to destroy your active site structure to form these inactive species of carbides and metal.
Next slide, please. So another one of the activities that Argonne is high-throughput synthesis and characterization of materials. For our pilot study of with the robotic system, we decided to look at the atomically disbursed class of materials because it does have a very high activity as shown in the rotating disc electrode measurements, and really a small fraction of the different parameters were studied by Los Alamos. So using a robotic system that many of you may be familiar with from the pharmaceutical industry, we combine the precursors to the atomically disbursed catalysts in various ratios, looking at, for example, the iron to zinc ratio, going from no iron to 7.5 atomic percent iron in the precursors. Looking at different salts. So Los Alamos has been focusing in their bulk synthesis on using iron sulfate as the precursor, so we also explored nitrate and acetate as precursors. Then we also looked at three different heat treatment temperatures, 900, 1,000, and 1,100 degrees C.
So we utilized the robotic system to mix the precursor solutions, drive off the solvent, and then do a heat treatment of multiple samples at the same time at the same temperature, and then characterize those samples in a high-throughput fashion using something akin to a rotating disc electrode measurement, but it’s called a multi-channel slow double electro cell that allows us to screen the oxygen reduction activity of four materials simultaneously, and this gives us information identical to what you’d obtain from a rotating disc electrode measurement.
So also in addition to screening the oxygen reduction activity in aqueous environment, we’ve developed or further developed an apparatus that is commercially purchased from Nouvant Systems to look at the membrane electrode assembly performance of these various materials in the fuel cell environment, and this is for 25 electrodes simultaneously using a segmented flow field in this hardware. We also use high-throughput characterization of the materials, a high-throughput XRD in the high-throughput research lab, and as high throughput as possible x-ray absorption, spectroscopy characterization at the advanced photon source. Next slide, please.
So looking at these various parameters and utilizing the robotic system to synthesize the materials, we were then able to synthesize 40 unique compositions of the catalyst in this atomically dispersed class, and what I’ve shown here in the bar chart is the oxygen reduction mass activity of these materials at .75 volts as measured using the multi-channel flow double electrode cell. And as you see at the bottom there, there’s a legend as to what all these catalyst codes refer to. So the bottom line from these studies is that we found the highest oxygen reduction activities for materials synthesized from the iron nitrate and iron sulfate precursors, and that we achieved the highest oxygen reduction activity at the lower pyrolysis temperatures of 900 to 1,000 degrees C.
We were able to correlate that oxygen reduction activity with the content of the FeN4 side by characterizing all these materials using x-ray absorption spectroscopy, and then linear combination fitting of the [inaudible] region. So out of these studies, we’ve found materials that potentially have five times the oxygen reduction activity of the baseline materials. The baseline materials explored by Los Alamos are shown with the stars on the bars. So the next step with this then is to reproduce this synthesis, implement it in the fuel cell, and actually hopefully realize that activity in the fuel cell environment. Next slide.
Piotr Zelenay
All right, so now there is a general agreement in the PGM-free electrocatalysts community that further development of these materials with concurrent increase in the activity will largely depend on but the enhancing, the understanding of the origins of our activity. And then the regaining higher – more control over the design of catalysts. The biggest challenge may be the identification of the active site or in catalyst obtained via high temperature treatment with relatively little control over the process.
And in the past few months, we’ve been able to demonstrate that at iron sites are indeed present on the surface of PGN-free catalysts. In this case, they’re atomically dispersed iron, nitrogen, carbon catalysts we touched upon a little earlier. The technique that we’ve adopted from biological studies is relatively new as related to most of our spectroscopy and called nuclear resonant vibrational spectroscopy, NRVS, or Nervous for short. When we combine this technique with the use of molecular probes, such as nitric oxide, NO, which expected to bond strongly to iron, and that had been demonstrated in biological studies I already mentioned.
And we have – we are able to look for iron as sites on the surface of catalysts, and indeed, the results shown in there in this slide indicate that once NO is added to the system, there is – there are additional vibrational features in the nervous spectrum indicating, improving that iron is present on the surface. The experiments shown in this slide was carried out with isotopically enriched iron. With the isotope here is the most power – the isotope of iron 57 in order to increase the signal in the nuclear resonant vibrational spectroscopy experiment at Argonne. While we cannot state for sure that iron is the active site, we have been able to demonstrate maybe perhaps for the first time that iron can survive in acidic environment, on the surface of catalyst once coordinated by nitrogen. Next slide, please.
We already mentioned the molecular probes in nitric oxide is one. Another one is actually nitrite and iron. We’ve used both in Los Alamos not only for demonstrating the presence of transition metals, iron is specifically on the surface of catalysts, but actually for the purpose of counting the number – counting the active sites present on the catalyst surface. Again, the results shown here are for atomically dispersed iron, nitrogen, carbon catalyst. On the left upper corner, we are showing the effect of adding nitrite probe to aqueous electrolyte, and its impact on oxygen reduction reaction polarization plot obtained at a – using [inaudible]. We can contaminate the activity pretty much completely. These are the plots, they’re the movement from green to red plot in the view graph, and then we can reductively remove the poison from the catalyst surface and fully regain the original performance.
By assuming certain reduction process from NO2- to Ammonium cat-ion, and during reduction, we can determine the third stripping charge, and also the surface concentration of active sites on the catalyst surface. And then I have by referring the most current density generated at any given potential, we can estimate the turnover frequency, which in this particular case is 1.2 electrons per site per second at .8 volt, which is significantly lower than the turnover for your frequency generated or measured for where PGN or platinum-based catalysts. We can also estimate the number of active sites per atom of a specific type. Carbon, nitrogen, iron. Those numbers are shown on the right-hand side in the table.
We have approximately one active site in 1,000 carbon atoms, 55 nitrogen atoms, and approximately half of iron atoms are involved in the active site base on this approach. Next one, please. In addition to carrying out experiments involving the probes, our atomic-level modeling has targeted among other targets the identification of where possible alternative probes to the nitric oxide, nitrite, we’re using binding energies as a guiding magnitude for active development.
In the tables shown, we are comparing binding energies calculated for different assumed active sites for iron-based catalysts, starting with FeN4 in the middle of the graphing sheet, which is a model system for the calculation, and then moving to FeN4 with attached to OH, which is a very likely configuration, again, based on DFT calculations. And then for the active sites, located at the edge of the graphing sheet. Of the systems shown here, carbon monoxide, nitric oxide, chlorine, sulphur, and OH, the O does indeed show the highest strength of binding strength to the active site. OH would have been an excellent active site probe, however, they cannot be used in an aqueous environment as such. Next slide, please.
Deborah Myers
Well we discussed previously one of the big challenges in this class of catalyst is identification of the active site and utilization of those active sites throughout the thickness of the electrode on which are typically in range of 80 to 100 microns thick. And also their utilization as a function of electrode potential or cell voltage. So this slide represents a lot of work by all four national laboratories in the synthesis of the atomically dispersed catalyst by Los Alamos and fabrication of the electrodes, and then testing of numerous cells at National Renewable Energy Laboratory as a function of oxygen partial pressure, and then analysis of those data and deconstruction of the contribution from various processes to the losses in the cell to extract the losses due to the kinetics of the oxygen reduction reaction.
And then correlation of that with the cell potential. And so what we’ve seen throughout all of these tests with multiple cells and different oxygen partial pressures is that there is a consistent dependence of the active site availability on cell voltage. And this has then been correlated – as has been previously by the Mukherjee group with the oxidation state of the iron center. Corresponding with that then, we did Operando and in situ x-ray absorption and spectroscopy to determine the oxidation state of the iron as a function or potential, and that is shown in the upper right plot. Whereas we decreased the cell voltage or the electro potential. We start to reduce the Fe3+ to Fe2+ in the range of 800 to 600 millivolts, which corresponds very well with the background [inaudible] shown in the lower right. So using these XANES, we can determine the number of electrons transferred as we go through this redox potential, and that correlates very nicely with the modeling results shown in the lower left plot.
Where we have an electron transfer of approximately .8 electrons – I’m sorry, .7 electrons as we go from partial utilization of the catalyst at the high potentials to full utilization at the low potentials in the Fe2+ form. With that redox potential being or the nurse potential for that redox being approximately .79 volts. Also corresponding very nicely with the background [inaudible]. So it’s this type of study then points to the Fe2+ being the active center for the oxygen reduction reaction, and that is the cell performance of course is improved as you start to decrease the cell voltage because you’re converting the Fe3+ to Fe2+. Next slide, please.
Piotr Zelenay
We mentioned early on in this presentation that catalysts had durability – presents a serious challenge for PGM-free materials. And how serious the challenge is after they were presented in the two graphs shown in this slide with the performance under steady-state conditions decreasing, it quickly in both oxygen and air environment in the fuel cell. We tried to model the kinetics of performance degradation using two relatively simple models. The first one is first order degradation, and the second is a two-step auto-catalytic degradation, which would indicate the performance loss due to a chemical attack of one of the products of oxygen reduction reaction during testing. And actually, the latter model is fitting the experimental data much better than first order degradation model, indicating it is – the performance loss is due to the attack of hydrogen peroxide or hydroperoxyl radical derived from hydrogen peroxide, causing degradation of the active site. Next slide, please.
There are several possible degradation mechanisms happening at the atomic-scale level, and also at macro mesoscale levels, and they are listed in the left-hand side of this slide with possible degradation process involving demetalation at an atomic-scale level, local N/C corrosion, active site poisoning. On a macro or mesoscale, a carbon corrosion is a likely mechanist. Also, ionomer membrane degradation loss leading to the loss of ionic connectivity within the catalyst layer.
We’ve developed a system for parallel measurements of carbon dioxide transition methyl and fluoride emissions during fuel cell operation, which is shown here, and we’ve been carrying out those experiments in addition to modeling experiments in the next slide. Next slide, please. So in this slide, we are showing an approach which involves a descriptor for activity loss called knock-on displacement threshold energy, KODTE, which is given in with the value of which is given in terms of electron energy to minimum energy of an electron required to knock off one of the atoms in the active site. Examples of such active sites are shown on the top of the view graph.
The calculations have been done by that [inaudible] at Los Alamos. Several assumed the structures involving different transition metals, such as iron, manganese, and cobalt in particular, graphing itself, and armchair and zig-zag edges of graphing sheets. From that calculation, we observe the highest accessibility of nitrogen to be removed from the active site. We are not seeing significant differences between different transition atoms, and we’re also not seeing to a significant effect of OH in unstability of the active sites. Next slide, please. So when we – XPS experiments, we are noticing actual a decrease in the nitrogen content in the catalyst, especially the content of pyridinic nitrogen, often associated with [inaudible] activity of these materials. And in – which is in agreement with the calculations just presented. We can also use limiting potential calculated using the DFT as an activity descriptor for PGM-free catalysts. And if we assume nitrogen is indeed being further removed first from the active site, we can predict the change – corresponding change in activity, and that change for one assumed active site of FeN40H at the zig zag edge of the graphing sheet is from .8 to .64 for the vault [inaudible] versus computational hydrogen electric, which is actually represents in more than one order of activity loss in the performance of these catalysts. Next slide, please.
Deborah Myers
The remaining challenges to implementation of PGM-free catalysts and polymer electrolyte fuel cells are outlined here. As we’ve discussed, the number one challenge is durability of these materials, so for the metal organic framework-based catalyst powders and electrodes, there is degradation, not just of the operating electrode, but of the catalyst powders as they’re sitting on the bench. So there’s an aging issue with the powders, so that needs to be addressed, and these indeed are the highest activity materials, those derived from metal organic framework material. So there’s also limited stability of the electrodes under steady state and load cycling conditions in the fuel cell environment.
Inadequate understanding at this point of the mechanism, there is some indication that peroxide is involved in the degradation mechanisms as just mentioned by Piotr. So the oxygen reduction activity of course is in need of improvement as Piotr discussed. The active site density is much lower than that of platinum, and in addition, the turnover frequency of those active sites is approximately an order of magnitude or more lower than that of platinum. So that indeed is in need of improvements such that you can decrease the thickness of the electrodes and improve the transport properties, and then hopefully improve the durability at the same time. So there’s challenges in determining how many of those active sites are on the surface of the catalyst, and in electrode design and catalyst ionomer integration to improve mass transport through the thickness of electrodes. Then also the ion is catalyzing the formation of hydroperoxide radical, which is detrimental to the ionomer and electrode layer, and also to the membrane, so ultimately you’d want to replace that iron with another transition metal, such as cobalt.
At this point, the cobalt-based catalysts have lower oxygen reduction activities than the iron-based materials, those are in need of improvement. Then of course the ultimate challenge is once these challenges are met is to integrate these materials into automotive fuel cell stack systems, and implementation in vehicles. Next slide, please. So back in FY17, there was a federal opportunity announcement, and four projects were selected out of that solicitation, led by Carnegie Mellon University, Giner, Greenway, and Pacific Northwest National Laboratory. Since we’re kind of running short on time, I won’t go into detail on each one of these present projects, but there are presentations on the DOE website from the last annual merit review that details all the approaches of these projects. Next slide, please.
There was also a FOA this fiscal year and an announcement made recently on the projects selected out of that FOA. These are led by Northeastern, Indiana University, Purdue University, Vanderbilt, Pajarito Powder, and United Technologies Research Center, and you can see the project titles on this slide. And we look forward to learning more about these projects as they get underway in FY19. Next slide, please. So this slide shows the co-authors. Of course, the consortium has very large teams at the various national laboratories, and these are the people who have been hard at work obtaining the data that I showed today. Next slide, please.
Eric Parker
Okay, so thank you, Debbie and Piotr for that awesome presentation. Just as a reminder, if you do have any questions about anything you just saw that was presented, please submit them now, and we’ll start Q&A in a minute here, and I’ll turn it over to my colleague, Dimitrios, to begin that.
Dimitrios Papageorgopoulos
Well thank you, Debbie and Piotr. I have a couple questions that have come in. The first one actually is a clarifying question if you can go to Slide 10. MEA performance states one bar air partial pressure. Is that supposed to be bad pressure?
Piotr Zelenay
No, this is actually partial pressure. So we are using one bar partial pressure of hydrogen at the anode and one bar partial pressure of air which means some of partial pressures of oxygen and nitrogen at the cathode.
Dimitrios Papageorgopoulos
Thank you, Piotr. Another one is a different type of question. At the start of the presentation, you mentioned the activity target the DOE has for PGM-free catalysts. Is there one that could be used for PGM and PGM-free catalysts?
Piotr Zelenay
In combination, right? Yeah.
Dimitrios Papageorgopoulos
In combination, yeah. The second leg to this question is what is – if the activity is you’re targeting is 44 milliamps per square centimeter, what is the status for PGM catalysts at this stage?
Piotr Zelenay
Well, we have in the table – we are showing right now, we are showing N/A, not applicable, because indeed it is PGM with the foremost target is given in terms of mass activity, .44. And I am – to this mass activity target as shown in the calculation at the bottom, it isn’t quite relevant to .44 milliamps per centimeter squared. I don’t remember exact numbers. I don’t know if Debbie, you remember these numbers, but I think PGM catalysts have come very close to – or even exceeded for some cases this target of .44 amps per milligram of PGM.
Deborah Myers
That is in the table in the column that says 2018 PGM status. So the mass activity status as when this table was established is .53 amps per milligram PGM, and then given the loading of .1 milligram PGM per square centimeter, that translates to .053 amps per square centimeter.
Piotr Zelenay
Yeah, 53 milliamps.
Dimitrios Papageorgopoulos
Another one has come in. Is there evidence that cobalt will produce less peroxide than iron, and therefore be more stable?
Piotr Zelenay
I don’t know that off the top of my head. I don’t think we’ve been testing the peroxide yields, which we can do actually only rotating this collector [inaudible], then fuel cell itself. We’ve done some very preliminary studies years ago of – and cobalt – durability of cobalt-based catalyst. I don’t recall seeing significant differences in the rates of performance loss between cobalt and iron, but that was at a time when activity of all our PGM-free catalysts was relatively low.
Deborah Myers
Yeah, and I think the point really then is to eliminate the iron such that you’re not generating the hydroperoxide radical from the peroxide that can be generated as part of the oxygen reduction reaction.
Dimitrios Papageorgopoulos
Thank you. I’ve got one last one here. The question is the following. Regarding catalyst active site durability, are there any indications on the composition of the active sites remaining after degradation approaches a minimum?
Piotr Zelenay
Well this is exactly the kind of experiments we’ve been planning, including this week. We are going to look at our catalyst, [inaudible] advanced photon source especially, and also using this [inaudible], and in combination with active sites approach to see which – tried to see which catalyst site can be correlated to the performance loss. At this point, we do not have such strong indication, which form of iron involving sites disappears. Debbie?
Deborah Myers
Well, hot off the press data that I was looking at last night, it looks like the FeN4 site is remaining intact, and what’s happening at least for the atomic re-dispersed catalyst and this select sample is that you’re removing the spectator species over time with conditioning of the electrode and operation at .8 volts. So early indication, and we certainly have to verify this, is it’s not destruction of an FeN4 site that is causing the loss and performance of the atomically dispersed catalyst-based electrodes.
Piotr Zelenay
In other words, there is maybe a possibility that it is the carbon next to the FeN4 that is affected over the operating time. So this result is preliminary results do not necessarily undermine the concept of the active site being FeN4, but it can prove show that it is the carbon matrix which is being affected over the time of operation of the catalyst.
Eric Parker
Okay, that’s going to do it for questions today. So big thank you to our presenters for informative presentation on ElectroCat, and that concludes today’s webinar. If we didn’t get to your question or you think of one later, please feel free to e-mail Deborah and Piotr’s e-mail listed here on the slide, or us at DOE in the webinar inbox. And I’d like to thank everyone for joining in today, and to check back soon for a posting of this full recording and the webinar slides on the DOE website, and I encourage everyone to sign up for our monthly newsletter, which includes information and registration for upcoming webinar topics. With that, I’d like to wish everyone a great rest of their week, and goodbye.
[End of Audio]