Below is the text version for the "Electrospinning Fuel Cell Electrodes to Improve Activity and Durability" webinar, held on March 28, 2018.

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. We've got a great presentation lined up this month from Peter Pintauro on electrospinning fuel cell membranes and electrodes to improve activity and durability. My name is Eric Parker. I provide program support within the Fuel Cell Technologies Office. And I'm the organizer for today's meeting.

We'll begin in just a second, but first I have a few housekeeping items to tell you about. Today's webinar is being recorded. And the full recording, along with the slides, will eventually be posted, and we'll be sure to let you all know. All attendees will be on mute throughout the webinar. But please submit any questions you may have via the chat box you can find on WebEx. We will cover these questions during the Q and A at the end of the presentation.

With that, I'd like to introduce our presenter for today, Peter Pintauro, who is joining us remotely. Peter is the principle investigator on this project and is the H. Eugene McBrayer Professor of Chemical Engineering in the Department of Chemical and Biomolecular Engineering at Vanderbilt University in Nashville, Tennessee. Prior academic appointments include Tulane University and Case Western Reserve University, where he was Chair of the Department of Chemical Engineering and Kent Hale Smith Professor of Engineering.

He is a fellow at the Electrochemical Society, a fellow of the American Institute of Chemical Engineers, and is house president of the North American Membrane Society. His research interests are in the areas of electrochemical engineering, membrane science, fuel cells, batteries, and nanofiber electrospinning. It's a pleasure to have you with us, Peter. Good afternoon.

Peter Pintauro, Vanderbilt University

Hello everybody, and welcome to my webinar today.

Eric Parker

All right, Peter. I've just made you presenter. Go ahead and share your webinar.

Peter Pintauro

Okay. So today I will be discussing past and current research work, where I'm using nanofiber electrospinning to prepare and evaluate membranes in electrodes for hydrogen/air fuel cells. The webinar outline is shown here. I'll first be talking about nanofiber composite fuel cell membranes. Talk about ideal membranes, the process of electrospinning polymer fibers, how we can make composite films. And then so some physical properties and durability test results with some of the membranes that we make here in my laboratory.

I'll then turn my attention to particle polymer nanofiber mats, which are used as electrodes that—in fuel cell MEAs, membrane electrode assemblies. Discuss a little bit about how we prepare the ink and the process of electrospinning, the physical structure of the fibers, and then talk about fiber performance in terms of power output and durability.

And I'll finish up with a brief discussion of future challenges and acknowledgments. At the last couple of slides in my webinar are just lists of papers that have been published dealing with both nanofiber membranes and particle polymer fiber mat electrodes.

So let's begin with a discussion of proton-exchange membranes for hydrogen/air fuel cells. So this is a major effort of my work. It has been funded by the Department of Energy. The ideal membrane, we want low sheet resistance. What that means is you want a membrane that is very thin but highly conductive to protons. In terms of durability, one needs a membrane that has low areal dimensional changes.

In a fuel cell operation, changes in membrane thickness during on/off, dry/wet conditions is acceptable. But significant changes in the lateral or areal dimensions of the membrane—those are the membrane swelling parallel to the electrode surfaces in MEA, which is shown schematically up here—any changes in this lateral areal dimensional swelling property of the membrane will decrease membrane durability during on/off fuel cell cycling. We need a membrane that's gonna be mechanically strong, with moderate volumetric swelling. And of course we need low fuel and oxygen crossover.

To achieve these properties, though, where it's somewhat contradictory in terms of what we need to do and the resulting effect on membrane—on the membrane itself. If we want a high conductivity, we put a lot of charge into our polymer ionomer, a lot of fixed charge. But this makes the polymer more hydrophilic. It swells greatly in water.

And that higher swelling will lower the effect of concentration of fixed charge sites. It also weakens the membrane mechanical properties and causes excessive swelling in the in-plane direction. And so it's been realized by many people that a single homopolymer, a single ion exchange, cation exchange polymer, really won't work in a fuel cell to achieve all of these targeted properties. And so one needs to consider blends, block copolymers, composites, perhaps crosslinking the polymer to control its swelling and mechanical properties.

Today I'm gonna talk about composite membranes. And this is the ideal membrane that we're trying to make. It's actually two types of membranes. This is what we call a nanofiber composite membrane, where we have ionomer nanofibers, proton conducting nanofibers, surrounded by an uncharged polymer matrix that provides mechanical strength and can—controls the swelling of our fibers.

Or we can do an inverse structure, where we have nanofibers of uncharged polymer, which reinforce an ionomer matrix. And so at least for option number two, this is not unlike, let's say, a GORE-SELECT membrane and other membranes that have been examined in the literature, where there's some kind of a inonomer impregnation into a preexisting fiber reinforcing mat.

In my laboratory, what we do is we do not want to use this impregnation step. But what we do is we electrospin two fibers simultaneously, both the uncharged polymer and the charged ionomer. So we create this membrane, which I showed here. In this particular case, we have charged fibers surrounded by an uncharged polymer, which controls the swelling of these fibers. But we can create this kind of a structure by simultaneously spinning both uncharged and charged polymers.

What we do is we create a kind of forced assembly, a polymer mixture, which allows us to co-deposit these two different polymers, widely different properties, simultaneously on a collector surface, where we have intimate mixing of those two polymer components—the kind of mixing on a nano scale that one cannot achieve by simply putting these two polymers into the same solution and casting a film.

The resulting structure is—it decouples the mechanical and proton-conducting functions of the polymer membrane components. We can independently control both the size and the loading of both the proton-conducting phase and the uncharged polymers. And in this particular configuration we eliminate the need for a separate polymer impregnation step. And I'll show you how we do that in some of the next few slides.

First I want to talk a little bit about electrospinning. This method of producing fibers has been around for quite some time, well over 100 years. But it really took off and became much more important in the scientific world based on some of the pioneering work of Darrell Reneker at the University of Akron in the mid-90s. And so in a polymer electrospinning process, what we do is we take a polymer solution or polymer melt, and we force it through a needle spinneret under the influence of a very high electric field, where we—and we direct fibers toward a grounded collector surface.

If you look closely at the tip of the spinneret, the electric charge on the solution overcomes surface tension forces, and one has what's known as a Taylor cone, this necking down of the solution. Emerging from the Taylor cone is a fiber jet. That fiber jet then starts to move toward the collector surface under the influence of the electric field, at which time the solvent is evaporated. The fiber stretches, and there's a whipping action that occurs.

And the combination of all of these effects results in fibers that are sub-micron in diameter. I show you a YouTube video on the right side of this screen, where you can actually see these fibers being formed. And you can see the extreme whipping action here. So it's not uncommon to be able to generate fibers on the order of 100-nanometer diameter using electrospinning.

In order to electrospin a solution, there are a variety of process variables that one needs to consider—the concentration of polymer and solution, the applied voltage, the distance between the needle and the collector surface, which in combination with the applied voltage determines the electric field strength. We can talk about solution flow rate, humidity, and solvent type. And there's no real simple way of finding these correct conditions to make well-formed fibers. It's more of an empirical trial and error procedure.

Nevertheless, we can, in most cases, do this with an electric field and a needle. But there are also things such as needleless electrospinning, where one has a rotating cylinder that picks up a very thin solution, for example, a polymer, a rotating disk, spiral coils, rotating ball. And what happens is you create a whole series of node points at which fibers emerge, let's say, from this cylinder. And they are collected on some surface.

There's bifurcation of the fibers themselves as they leave the fiber generator. And so one can either use a needle or needleless electrospinning with an electric field. And there's also things that can be used where there is no electric field, such as a centrifugal or rotary jet electrospinning. You can see these in the literature.

Here we have a needle spinneret that rotates at very high speed, emits a fiber. And the centrifugal force causes that fiber to stretch and get deposited on the interior circumference of a cylindrical shell. And you end up with fibers without having to use an electric field. So there are various ways in which you can make electrospun—or spun fibers with diameters of sub-micron dimension.

In my case, what we do is we co-electrospin simultaneously an ionomer—let's say a perfluorosulfonic acid polymer for proton conduction—and an uncharged polymer, like polyphenylsulfone, which we will use as the reinforcing polymer. And we have a dual fiber mat. And we can control the flow rate of each of these particular components in order to adjust the relative amounts of each of the fibers in our fiber mat. They're well mixed, and they are co-deposited on the collector surface.

We then have two options as to how we want to process this very porous fiber mat. The fiber volume fraction here is maybe .1, .2—10 percent fibers only. We want to process this into a dense, defect-free membrane. And we have two choices. In one case, we can melt the uncharged polymer fibers and have them surround the ionomer fibers, in which case we'll end up with a interconnecting 3-D network of ionomer nanofibers that are surrounded by a uncharged polymer material.

Or we can choose the inverse structure, where we allow the ionomer fibers to soften and flow around the uncharged polymer nanofibers. And we retain these uncharged fibers in order to use them as a reinforcing mat. So in one case we have ionomer fiber surrounded by uncharged polymer. In the other case we have uncharged polymer fiber surrounded by the ionomer material. And we process these from the exact same mat. And we look at which one of these may be better for fuel cell applications.

So the general experimental procedures—you make a—prepare a polymer solutions for electrospinning, one with the ionomer, one with the uncharged polymer. We use perfluorosulfonic acid materials like DuPont Nafion, shown here, or lower equivalent weight, short side chain, PFSA materials such as those from 3M Company.

Unfortunately you cannot electrospin Nafion using typical alcohol water solvents—most solvents—because they don’t really form a true solution. The polymers do not unfold. And you need a polymer chain entanglement to create electrospun fibers. And so we have to add what we call a carrier polymer to our electrospinning solutions PFSA.

And in our case we use poly(ethylene oxide) shown down here. It is a water soluble polymer. So we spin fibers with Nafion with PEO, and then we can remove the PEO after electrospinning to achieve a neat Nafion fiber. So we have to identify electrospinning conditions for both—in our case polyphenylsulfone as the reinforcing fiber.

We've also used PVDF, shown here, as reinforcing fibers. And so we want to make fibers with no beads, no defects. Structures, we want well mixed system. And then we want to process this dual fiber mat into a dense and defect-free membrane. So that's what we're doing.

And today I'll talk mostly about DuPont Nafion with polyphenylsulfone. I'll show one example with this low equivalent weight 3M material. So here's the apparatus in our laboratory. It's very—rather simple. Custom made. We have a rotating drum collector that also oscillates in the lateral, horizontal direction so we get a nice, uniform coating of fibers on our collector surface.

We have a needle syringe filled with polymer solution and a syringe pump and a power supply. When we do dual fiber electrospinning, we have two of these syringe pumps, shown at the lower portion of this slide. Two needles coming in from opposite sides of our rotating drum. We can independently control the voltage gradient and the flow rates of each of these two solutions.

We end up with a fiber mat. It's highly porous. Typically, it's about 16 centimeters by 10 centimeters. And the thickness depends on how much polymer we want to deposit—typically 10 to about 120 microns. Of course, when you compress it and remove the porosity, that thickness decreases significantly.

So here are some typical examples of electrospun fibers. This is with 825 equivalent weight perfluorosulfonic acid polymer. On the left side we have a 10 weight percent solution of PFSA and PEO, where we change the PEO content from 25 percent down to one percent. You don't need much carrier in order to make fibers.

The more PEO you have in the fibers, you can see from the histograms here—this is the histogram of the fiber diameter distribution—high PEO content you make large fibers, maybe 600 nanometers or so in diameter. As you go down in PEO content, the fiber diameter decreases down to maybe an average of about 100 nanometers. We can dilute the material further in our electrospinning solution. There are some limits here.

But if we go down to 5 weight percent, we can also make fibers at 10 and 5 weight percent PEO. And you can see a very narrow fiber diameter distribution here. And we can make small fibers with more PEO if we dilute the solution. So viscosity of the electrospinning solution has some effect, for example, on the diameter of the final fiber.

And when we go too low in concentration, five percent, and too low in PEO content, one percent, we were unable to get fibers. And so we end up with spray droplets. And this is what we have. And in my research we want to avoid spray droplets and at all costs. We always want to make fibers.

So here's a typical fiber mat that we have. This is a combination of Nafion fibers and polyphenylsulfone fibers. They're indistinguishable from this top-down SEM of the electrospun mat. And then we want to process this mat in two ways. We want to soften the Nafion and keep the polysulfone fibers, or we want to soften and let the polysulfone flow while keeping the Nafion fibers.

And so what we do is we have a—just a simple experiment. These mats are left out in open atmosphere. They pick up a little water vapor from the air in our laboratory. And so when we hot press these mats under these conditions, 4 times at 10 seconds, there is sufficient water in the Nafion fibers and sufficient heat and pressure that the Nafion softens and flows and fills in all of the void space between the polysulfone fibers.

We then boil the membrane in sulfuric acid, make sure it's in the proton form. We boil it in water. These steps also are used to remove any PEO from our Nafion fibers. And we end up with a membrane where we have Nafion and reinforcing polysulfone fibers. A freeze-fracture cross-section is shown here.

You can see that there's nice filling of all of the void space. We do gas separation experiments, pressure experiments, to make sure that there are no defect voids in our membranes. These are very tight membranes. There are—there's complete filling of all the void space by the Nafion.

We can also take the same fiber mat, and instead of hot pressing, we can cold press, push the fibers closer to one another without melting or allowing the Nafion to flow. And then we expose the mat to chloroform vapor for just a few minutes, depending on the temperature—60 minutes at room temperature. And that chloroform vapor softens and plasticizes and allows the polysulfone fibers now to flow and fill in the void space between Nafion. And then we again anneal the Nafion.

We boil in acid and water, and we have a membrane now where we have polyphenylsulfone surrounding Nafion fiber network. If we do a freeze-fracture cross-section, we don't really see much here. But if we take the same membrane, and we immerse it into liquid chloroform, which allows all of the polysulfone to dissolve, what we end up is we can see here on the far right an SEM—top-down SEM of the Nafion network of fibers that exists in this particular membrane.

So now we go to looking at membrane properties. Here is proton conductivity as a function of Nafion volume fraction. This is conductivity measured in plane, in water at 25 degrees C. The dotted line here is a simple mixing rule based on the volume fraction of Nafion. And we see that conductivity essentially behaves as one might expect.

You lose a certain fraction of the conductivity of Nafion depending on the volume fraction of polyphenylsulfone in the membrane. And it really doesn't matter if you have polysulfone fibers and Nafion matrix or Nafion fibers and polysulfone matrix. So both configurations essentially follow a linear mixing rule. If you have 50 percent of the—of membrane is composed of polysulfone, you only end up with 50 percent of the conductivity that you would expect if you used—looked at a neat Nafion film.

On the right here we show volumetric water swelling at 100 degrees as a function of the Nafion volume fraction. And here we see that the swelling falls below a simple mixing rule based on the swelling of neat Nafion—oops, I'm sorry—the swelling of neat Nafion at 100—at—without any reinforcing fibers and the swelling of polyphenylsulfone in water, which is zero because it is a highly hydrophobic material. And again, what we see here is both membrane structures have essentially the same volumetric water swelling.

For a fuel cell membrane, it's in-plane swelling which is very important. This has something to do—this affects the durability of the membrane during hot, wet, dry, cold, on/off fuel cell conditions. And so we also looked at the in-plane swelling as a function of Nafion volume fraction. And here we see a fundamental difference between our two membrane structures.

If we have Nafion fibers surrounded by polyphenylsulfone, that polyphenylsulfone has a 3-D interconnecting network. And so what it does is it creates isotropic swelling in our membrane. In other words, the thickness swelling and the in-plane swelling are almost the same. And what happens is we see high in-plane swelling for our particular membrane.

If we look at the opposite structure, where we're using polysulfone as a reinforcing matrix, we don't have this 3-D interconnectivity between the polysulfone fibers. These polysulfone fibers are reinforcing Nafion, but when water enters into the Nafion phase, those fibers—the Nafion can swell, and the fibers can move apart. And if they can move apart, they increase the swelling and the in-plane—I mean in the thickness direction.

And the result is that there's less swelling in the in-plane direction. The Department of Energy has a targeted in-plane swelling of about five percent. And you can see that we can achieve this five percent in-plane swelling at a much higher Nafion volume fraction when we use the polysulfone as reinforcing fiber as opposed to using the polysulfone to surround Nafion fibers.

And so this is the preferred—this blue data is in fact the preferred structure for how to reinforce a ionomer membrane for hydrogen/air fuel cell operations. Of course, we wanted to see if this worked in a fuel cell. So we adjusted the membrane thickness to compensate for the loss in conductivity. Again, we have a simple volume fraction mixing rule.

So if we have 40 weight percent—if we have 40 percent polysulfone to our membrane, we're gonna have to make our membrane 40 percent thinner in order to get the same areal—or the same resistance in a fuel cell. And so that's what we have done here. The red data is a typical Nafion membrane, with 51 microns in thickness. The blue is a voltage current density plot for a nanofiber composite membrane.

Since it's only 60 volume percent Nafion, we lose conductivity. But we make that up by making the membrane thinner. And then it behaves the same. So you can adjust the thickness and conductivity of these membranes, get the same sheet resistance, and then those should work equally well as a homopolymer material in a fuel cell MEA.

We also look at durability here. And we did a wet/dry cycling durability test, where we look at open-circuit voltage. So we make a membrane electrode assembly, adding cathode and anode catalyst layers to our membrane. This is the same membrane I showed in the previous slide.

And we look at open circuit voltage. It's above .9 volts. And then we begin 2 minutes of 100 percent RH hydrogen air and 2 minutes of 0 percent RH hydrogen air. So we add wet and dry hydrogen and air, cycling every 2 minutes at 80 degrees C. And we look for changes in the open circuit voltage in our MEA.

As we expand and contract our membrane, if this in-plane swelling is too high within a fuel cell test fixture, it's well known that cracks, defects, pinholes will develop in the membrane. And when they do form, there is going to be mixing of hydrogen and air and a decrease in the open circuit voltage. And we see that for our Nafion 212 membrane, as shown here, using a arbitrary voltage of .8 volts as a cut-off as to when the membrane fails due to gas crossover and pinhole formation.

We see that the Nafion failed after 546 hours of this humidity cycling. When we looked at the nanofiber membrane, which swelled much less in the in-plane direction, we improved the lifetime by 54 percent, up to 800 hours. So we can see that this lower in-plane swelling improves membrane durability.

We next go—turn our attention to electrospinning fuel cells—electrodes. And so now what we want to do is take an ionomer, like PFSA. We want a carrier polymer, and we now want to add catalyst particles and create an ink with a solvent. And now we want to extrude this ink through a needle spinneret under the influence of an electric field and deposit it onto a collector surface.

What we're trying to do here is create a new electrode structure. It's one thing to have new catalysts and new high activity catalysts, whether they're alloy catalysts, high surface area materials. But there's also the need to arrange and organize those catalysts into an electrode structure that improves the overall performance of the membrane electrode assembly. And so our focus was how to organize catalysts using nanofiber electrospinning into electrode mat structures that improve the performance of the fuel cell.

So there's a need for high power at low platinum loading. We also want highly durable catalysts where we minimize carbon corrosion and platinum dissolution and the effects of these corrosion and dissolution processes on the power output. And then we also want to be able to create an electrode where we have some system operational control in order to—in particular—get oxygen and air to a catalyst sites and get water out of the fuel cell, especially at high current density conditions.

So here we are electrospinning a fiber mat. We have to add a carrier. In this case we're adding polyacrylic acid, shown down here. We can also—I'll show you—add PVDF as a carrier and also something like polyvinylpyrrolidone. What—people have put particles into fibers for many, many years, not necessarily for fuel cell applications.

But what distinguishes some of our work for—as compared to others' is that we're putting lots of catalysts into our fibers, greater than 50 weight percent catalyst in our fibers. So we get fibers that have—are primarily catalysts. We use this perfluorosulfonic acid as a binder. And we're able to electrospin.

So here are some of the experimental conditions that you have to prepare an ink. There's certain steps. I won't go through all of them, but you can look at the steps here or look at our papers. You can see the voltage conditions, the flow rate, the distance collector to needle. We control the humidity of the air in which we electrospin.

And what we can do is we can make fibers. And so here are some high resolution SEMs of electrospun fiber electrodes. This data was collected for us by Karren More at Oak Ridge National Laboratory. This is a Johnson-Matthey platinum on carbon catalyst. Typically our fibers contain 65, 70, 72 weight percent catalyst.

You can see that there's a range of Nafion and also some polyacrylic acid. Again, that's a water soluble polymer that we believe is removed from the fiber mat. So what you can see here from these SEMs is that we have fibers with a highly roughened surface. This high resolution shows that we have very good distribution of ionomer around individual—oops—individual catalyst particles.

And we also see that the fibers themselves have porosity within the fiber. And that's due to some of the humidity that we use in the air during electrospinning. Water vapor enters into the fibers and causes phase separation because the water and Nafion are incompatible with one another. So we have very good mixing of catalyst and binder at the spinneret tip.

We have this stretching of the fibers. We have very rapid evaporation of solvent during fiber formation in the electrospinning apparatus. And as a consequence, we have this kind of unique structure. We don't see much agglomeration of particles. Very thin coating of Nafion on the catalyst and porosity within the fibers.

We wanted to see how robust our fiber mats were, so we subjected them to different hot pressing conditions. We don't want to ruin the fibers during hot pressing. And we've made these fiber mats by either depositing them onto a current collector surface, like a carbon paper. We can deposit them onto a foil and then transfer them onto carbon paper during MEA fabrication.

Or we can even electrospin them right onto the membrane to make a catalyst-coated membrane. But we wanted to see whether or not the catalyst fiber mat were—had sufficiently robust mechanical properties that they would withstand the conditions for hot pressing without melting. And so we looked at hot pressing at 140 degrees at 10, 16, and 80 megapascal. And you can see that the fiber structure is retained.

And you can also see that the—there is lots of porosity between fibers. So we have fiber mats that have porosity, interfiber porosity and intrafiber porosity. And it's the combination of two that helps with the performance. So here are some typical fuel cell performance data. Johnson Matthey platinum on carbon with small, 5-square center MEA.

This particular catalyst fiber mat for our cathode had 65 weight percent platinum, 23 percent Nafion, 12 percent PAA. We used a painted anode with no PAA, just Nafion. We had a .1 milligram per square centimeter platinum loading. And we looked at different platinum loadings of our fiber mat. And we compared it to a .1 painted cathode electrode MEA.

The dotted lines here are the performance—the fuel cell performance, cell voltage versus current density. This is at 80 degrees, ambient pressure, 100 percent humidity. And that's the performance of our typical Nafion-based slurry painted electrode MEA. And you can see at the same fiber content, same platinum loading in a fiber electrode, we see a substantial increase in the power, the product of voltage and current density, at any given voltage.

And so for example at .65 volts, we see about a 33 percent increase in power at the same platinum loading going from a slurry to a nanofiber. We also see that we can decrease the catalyst loading down to .065 milligrams per centimeter squared and still see an improvement in the power density as compared to a .1 milligram per square centimeter. If we go very low, down to .029, we're seeing somewhat of a drop.

This is partially due to the fact that we don't have good fiber distribution on our fiber mat at these very, very low platinum loadings. But what you can see is that you get higher power in a nanofiber mat electrode, and we attributed this to inter and intrafiber porosity, which leads to a higher electrochemical surface area for our catalyst, higher oxygen reduction mass activity. And we get very good expulsion of water.

If we look at some alloy catalysts now and compare them to platinum carbon, so here we have—these are all nanofiber electrode cathode MEAs. We also used a nanofiber anode here. Both the cathode and the anode are .1 milligram per square centimeter. We looked at platinum cobalt and platinum nickel from TKK. And we looked at platinum carbon from Johnson Matthey.

And so here is a platinum carbon nanofiber MEA. We are at one atmosphere back pressure, 80 degrees, 100 percent humidity. You can see the GDL carbon paper we're using. You get a substantial improvement in the nanofibers when you go to a more active catalyst. These platinum alloy catalysts are known to be more active than platinum.

And we see that improvement in activity when we go to nanofibers. And so what this means is as new catalysts are being developed and as new higher activity catalysts are being developed, if you can put those into a nanofiber cathode, for example, structure, you're going to see that improvement in the nanofibers, that the nanofiber works better than a slurry, whether it—no matter what catalyst is being used. Every catalyst that we see, we—as we get better and more active in a slurry system, we get more active in a nanofiber system.

And you can see here, this is the mass activity that we measure for these various catalysts. Here's the power densities at .65 volts. And so we see very nice increase in power when we go to nanofiber mats. We also wanted to look at the durability of our cathodes in a nanofiber configuration.

We looked at two different tests. These are accelerated stress tests, ASTs, that have been developed by the Department of Energy and National Laboratories. There's a—and also auto companies, like Nissan—there's a triangular wave test between one and 1.5 volts that's supposed to simulate the long-term fuel cell start/stop operation. There's also a platinum dissolution test, where there's square wave cycling between .6 and .95 volts that's supposed to simulate load cycling, acceleration, and deceleration.

So we performed these accelerated tests. We run thousands of voltage cycles, either triangular wave, 1 to 1.5 volts, or .6 to .95 volts and determine how well our membrane electrode assemblies work. So here are some data where we look at an electrospun cathode MEA and a sprayed cathode MEA. This is data that was collected at Nissan Technical Center.

This is one of their non-optimized sprayed cathode MEAs. And so on the right is the performance at 100 percent RH; on the left 40 percent RH. These are power density—or current—voltage current density curves. The product of those, the current density and voltage, is power. So we want to push our performance up to the upper right hand quadrant of these plots.

Anything down in this lower quadrant down here means that the MEA is not performing well. What we see here, if you look at this solid triangle data, is that's the beginning of life, BOL, data for our nanofiber cathode MEA. This is with platinum on carbon, Johnson Matthey, at .1 milligram per square centimeter. This is a 25-square centimeter MEA operated at a slightly higher back pressure, 1 atmosphere of back pressure.

And if we look at the—here's the initial performance. And again, we see better performance initially than we see with the spray electrode. That's what I've shown in previous slides. And so we see the benefit of the nanofibers in terms of the beginning of life performance due to this structure that we have.

We then did 1,000 voltage cycles, triangular wave, to do our carbon corrosion test. And when we did that, we saw a substantial drop in the performance of the spray electrode at Nissan. And that's due to loss of carbon, loss of some platinum. But it's also due to flooding of the electrode, of our cathode, that during the process of carbon oxidation and loss of carbon, one also produces a certain carbon oxygen species on the catalyst surface, which makes the catalyst more hydrophilic.

It holds on to water, and the catalyst floods. Water coats the catalyst surface, and oxygen can't access catalyst sites. And you see a decrease in the performance of the MEA. And that's what you see here, a substantial drop in the performance for the spray electrode.

When we run our nanofiber electrode through the 1,000 voltage cycles, we also see a drop in performance, but not nearly as dramatic as the spray. And we see some loss in surface area of the catalyst. We see a loss in carbon. But we do not see as much of a loss in power.

And that is because our nanofiber electrodes are very good at removing water. The porosity between the fibers enhances water removal. And so what we see here is that the performance of our nanofiber electrode after this corrosion test examined durability, is about the same as the starting performance of the spray cathode MEA.

If we look at 40 percent RH, we get a very different and unusual effect that's going on. If we look at 40 percent RH, at beginning of life the spray electrode at Nissan worked much better than the nanofiber electrode. And that's because our nanofiber electrodes don't hold water very well. So at low RH gas operation, they're dehydrating. We're losing water, losing conductivity. And we see this drop-off in fuel cell performance.

If we look at the end of life for the spray and the fibers, what we see for the sprayed system, at end of life there's substantial drop in performance. There's a loss in carbon. There's a loss in platinum. There's—the electrodes become slightly more hydrophilic, so they're able to hold onto water better. But the loss of carbon and the loss of platinum overwhelm this hydrophilic effect, and you get an overall loss in power. And so you see the open circles here.

With the nanofiber electrode, after carbon corrosion, there's actually an improvement, an increase in the power. And the power is beginning to approach the beginning of life performance of the spray electrode, as you can see from the open triangles that's starting to approach the solid circles here. What happens during the carbon corrosion is that we do lose some carbon. We lose some platinum.

There is some corrosion effect here in our fibers. But the fibers become more hydrophilic due to the oxidation of the carbon surface of the catalyst. And that means the fibers hold on to more water. We see an increase in—with more water there's an increase in the activity of the catalyst toward oxygen reduction. There's a higher conductivity and—of the ionomer. And so we actually see an improvement in performance with our fibers. Very unusual effect here.

We also looked at a little bit on using PVDF as a—not only as a carrier for Nafion in our electrospun fibers—electrodes but also to adjust the hydrophobicity of the binder in order to control carbon corrosion. So during this on/off cycling, carbon corrodes because it reacts with water, forming CO2. And so we thought by adjusting the amount of PVDF in our fibers, which we need to spin the Nafion, we could also adjust the hydrophobicity.

And so we can make fiber electrodes ranging from pure PVDF up to a Nafion/PVDF ratio of about four to one. So here we have a highly hydrophilic binder. Here we have a very hydrophobic binder. In all of these cases, the catalyst content of our fibers is about 70 weight percent. So here the PVDF stays in our fiber mat, and then we look at performance.

So here is some beginning of life and end of life performance of these different Nafion/PVDF binder fiber cathode MEAs. This is all at .1 milligram per square centimeter. The dotted line here is a conventional slurry MEA. What we see in the data here is that when our binders have mostly Nafion, 80/20 or 67/30, but the major component is Nafion, we have a fuel cell performance that outperforms the slurry MEA.

So here we're seeing high conductivity in our binder. And we're utilizing our fibers quite nicely, so we get good oxygen access to catalyst. And we're getting good expulsion of water. When we have fiber mats that are mostly PVDF, so we only have 33, 20 percent or no Nafion whatsoever, at beginning of life we see that the performance is not as good as that of Nafion.

And that again is due to the fact that we have a hydrophobic material. We're diluting the Nafion. Conductivity goes down. There's not much water on the catalyst surface. Water helps to catalyze the oxygen reduction reaction.

But even if we don't have—even when we don't have any Nafion, we still get power. It's coming out of our fuel cell. There's enough water that enters into even a pure PVDF catalyst fiber that we can get some conductivity. And we get some power.

After we go through 1,000 voltage cycles for our carbon corrosion test, what happens is these highly hydrophilic binders that have mostly Nafion, they tend to lose power a little bit. The fibers that are mostly PVDF and are hydrophobic, they become more hydrophilic, and they're—they gain power after the carbon corrosion tests. And our conventional GDE, well, it just totally fails and loses lots of power.

So after carbon corrosion what we see is that almost all of the nanofiber electrodes that have a mixture of Nafion and PVDF, no matter what—how much Nafion we have, all the way down to 20 percent Nafion, we see essentially the same polarization curves, the same fuel cell performance. And what we see is that a pure PVDF fiber with no proton conducting Nafion, after carbon corrosion it performs the same as a conventional slurry electrode with just neat Nafion binder.

So if you're looking at beginning of life performance, 80/20 Nafion/PVDF is best. If you want to look at end of life performance, you use a more hydrophobic binder. It's gonna give a lower initial power, but after carbon corrosion, it's gonna produce the best. So that's what we see here.

Here are some results where we're looking at load cycling. Here we're looking at platinum cobalt catalyst, where we are looking at a square weight potential between .6 and .95 volts. Here we're simulating acceleration/deceleration. Here we're looking at platinum cobalt catalyst in the nanofiber and in the conventional slurry electrode.

So the solid symbols here, red circles, that's platinum cobalt nanofibers. Initially very high power, much higher than the black circles, solid circles. That's a conventional GDE. These are five square centimeter MEAs. You can see platinum cobalt, Nafion, with PAA. There's the composition.

We—slurry electrode, we didn't add PAA. We don't have to add this carrier. Our anode is a mixture of Nafion and catalyst. And here are the conditions under which we run these tests. One atmosphere of back pressure at 80 degrees. When we perform our 30,000 voltage cycles, square wave, to simulate load cycling, acceleration/deceleration, we see a substantial drop in the performance of the conventional slurry MEA with platinum cobalt.

Not nearly as much of a drop in power with the nanofibers, which are shown by the open red circles. And if we look at the power losses at .65 volts or the maximum power loss, what we see for nanofibers is we lose between 20 and 30 percent of the power after 30,000 voltage cycles in this accelerated stress test, whereas in the conventional MEA we're losing between 60 and 70 percent of the power.

What we're seeing here is that, again, not only is carbon corrosion having a less detrimental effect on the performance of nanofiber MEAs, but we're seeing that metal dissolution is also having less of an effect. And here we're not exactly sure what's happening. We're seeing less metal dissolution actually from the catalyst surface in the nanofibers.

We're not exactly sure why. That's what's producing this higher power retention. This is something that we are still looking at. But we see this both for platinum catalysts as well as platinum alloy catalysts. So what we see then is we have nanofiber electrodes, high initial power, very good durability. We can go to low platinum loadings, produce high power.

So to finish things up, let me just mention something about future challenges in terms of membranes and electrodes. Future challenges are related to cost, performance, and durability. One needs to make membranes that are very thin, low—to reduce the cost of ionomers. Use inexpensive ionomers. The membranes have to be very thin with very good mechanical properties, low gas crossover. Right now people are looking in the 15 to 20 micron range.

We need to improve the performance, which means high conductivity over a whole range of different humidity conditions. And create membranes that can hold on to water at both high temperature and low relative humidity. And of course we want to minimize the chemical degradation of the fuel cell membrane during operation and also minimize in-plane swelling effects to maintain high durability.

And for electrodes, again, cost is an issue. Lower the platinum loading without losing power. We want to create membrane electrode assemblies using roll-to-roll manufacturing processes. I want to say that electrospinning is a commercial manufacturing process right now.

It can make materials—fiber mat materials and roll-to-roll process schemes at very, very high throughput rates. So electrospinning is compatible with high throughput roll-to-roll manufacturing. We want electrodes that generate higher power. We need to optimize the porosity, the binder properties.

We want to decrease the binder thickness. We want to improve power at low humidity. We want the binder to hold on to water if we can. It has—the binder also must have good compatibility with the membrane.

And finally we need to make sure that we have good durability. So we want to adjust the composition and structure of our electrode to minimize power losses when there is metal dissolution, carbon corrosion, things like freeze/thaw cycling.

So with that I want to thank everyone for attending this webinar. I want to acknowledge the people and funding that made this work possible. A whole series of postdocs and graduate students, some prior people who were in my lab, some people who are currently here. They're listed.

We've done some work with Nissan Technical Center, shown here. I want to thank Karren More, who has been very helpful in doing SEM analysis of our fibers. The membrane work was funded by the U.S. DOE, EERE, FCTO office. The fuel cell work was also funded and is currently funded by Fuel Cell Technologies Office of EERE.

I've had past monies from Nissan, Merck, the National Science Foundation. And I want to thank TKK for supplying some of the platinum-alloy catalysts. So with that I'd like to thank everyone. Sorry I took a little too—a little longer than I expected. I'd be happy to answer questions.

You have my email address. You can always send me an email or call me if you have questions that we can't get to right now. Thank you.

Eric Parker

Thanks, Peter. I will address some of the questions that were sent. We had a lot of great ones, and we'll see if we can fit a few in right now. Off the bat, we had a question come in. How much electric voltage is on the spinneret?

Peter Pintauro

We always ground our spin—our spinneret tip has somewhere around 10 to 12 kilovolts. And typically that's what we operate. There are electrospinning processes that can go much—that go much higher than that. One can do melt polymer electrospinning.

Sometimes that requires a much higher voltage. But right now for both the membrane fibers and the electrodes we're somewhere in the vicinity of, I'd say, 8 to 12 kilovolts. And we ground our collector.

Eric Parker

Another one, real quick. Is the solvent for the catalyst ink aqueous or non-aqueous?

Peter Pintauro

We—when we're using Nafion with PAA, we're using a alcohol/water mixture. Not unlike the kind of dispersion that one has with—when you buy a Nafion dispersion. When we use a PVDF as a carrier, there we have to use a [coughs]—excuse me—an—a different solvent. Okay. We use an organic solvent, like DMAC acetone mixtures when we have PVDF as the carrier and we're making Nafion/PVDF fibers.

Eric Parker

How long does it take to make the fiber sheet with 16 by 10 size?

Peter Pintauro

Right. We have a single needle. So it takes us an hour or two to make the mat. It depends on how thick the mat is. So that would depend on what platinum loading we wanted for our catalyst or how thick we want our final membrane to be.

But in an industrial process you would have hundreds of these needles. And it's not uncommon to have linear fiber mat speeds of 10, 20 feet per minute making fiber mats. So there are electrospinning manufacturing processes and equipment and companies that we are going to start working with that can make a 3-meter or 2-meter wide fiber mat at a linear mat speed of maybe 10 feet a minute.

Eric Parker

Great. And –

Peter Pintauro

In our lab it takes a couple hours. In an industrial process it's very, very fast because they use multiple needle spinnerets.

Eric Parker

Awesome. Well, I want to be respectful of everyone's time. We have a bunch more other submitted questions, but we are at the end of the hour. I will make sure these all get forwarded to Peter, and he can answer you privately.

But in the meantime I want to thank everyone for joining. And that concludes our webinar for today. Also, please feel free to submit future questions to the DOE webinar email inbox you see there. I'd like to thank everyone for joining and remind you that we will make the full recording of the webinar and the slides available online shortly.

Please be sure to check out the presentation when it does. Also, I encourage everyone to sign up for the monthly newsletter, which includes information and registration for future webinars. And have a great rest of your week, everyone. And goodbye.

Peter Pintauro

Thank you, everyone.

[End of Audio]