 OK, so I think as most of you already recall, we start talking about fuel cells last class. And I'm going to continue with that today. And I suspect based on the level of questions that popped up last hour that it's likely that we'll just stay with that topic today and try to address some of these issues that we were just batting around before we started the class. But first, let me point out to you that I have given you a problem set. And it will be a do in class a week from today. And it's sitting on the website. So you can pull that down, PDF file, and you have one week to complete that. If there's any issues about accessing that problem set, I'm sure Tom will be delighted to hear about it. The other comment to make is sometime this afternoon or evening, the PowerPoint that I'm using will upload on to the website. That was the Royal Way. That means Tom has some more work to do. And so you'll have access to that. And then the final comment is that you can look high and low and barred and faulted to find things about fuel cells, and you won't. So you have the PowerPoint and whatever I might say today as your guide on that subject. What I didn't show you last time when we met was what a fuel cell looks like. So here's a fuel cell. Very unimpressive, a single cell. These are just cells that we've used in our laboratory. I'm going to have a few of them up here, so I'll pass them around. Let's see. I'll give you guys a couple here and shoot a few around here. So the circular material that you'll be seeing, that's naffion that we've been talking about. These fuel cells that you're looking at have been fairly well used. So they're looking a little on the sad side. The black material that you'll see the squares on there, that's the electrode material, the graphite cloth that we're using. What you can't see in all of that is in between the naffion and the electrode is a platinum catalyst bed. And the way we make these, by the way, is we simply take the electrode with the catalyst bed material on it and the naffion, and we hot press it. We iron it together, essentially. You get a structure like that. Now the impressive thing about that structure is the best of those cells, just that single cell, will generate 20 amps of current. And one that's not so good, or maybe more average, not bad, but more average, might generate 10 amps of current. So 20 amps of current, that's probably about the amount of electricity we're using in this room with all these wonderful lights in there. So we're generating enough current to power a room this size with a fair number of electrical utilities in it. What the cell won't do is it won't generate significant voltage. So that 20 amps is being generated at somewhere around a half a volt when we get that. And so to get the kind of voltage that one needs, you take a bunch of those, you stack them together. That's a fuel cell stack. And the voltage, of course, adds, because they're in series. And that's how we go. But all the data that I'm showing you, and this is what I showed you last time, is based on one of those single cell membrane electrode assemblies, MEAs. And as you recall from last time, what we are doing is we're interested in our lab in running these devices at higher temperature to get around some of the issues that I mentioned, namely CO poisoning, water management, and thermal management. And when one does that, a normal napheon-based MEA degrades, as I showed you last hour. But we do some material science. And we put a new membrane material in, which is the same napheon. But now we've added a metal oxide component, typically about 3% by mass. And as the orange curve shows you there, we can get a very good response, even at 130 degrees, that normally outperforms the regular cell response at 130. But outperforms the regular cell response at 80 degrees, the black curve, which is where the napheon cell likes to operate. So why is this happening? Why does this work? If you look in the literature around the time we were starting these experiments, mid-90s, then it was explained in the literature that what was really happening was when you added things to napheon, you were doing two things. One was you were increasing the number of protons available. Napheon remembers an acid. And more protons makes a better electrolyte. And the second was that you were making the material more water-retentive. That is, water could not evaporate out of the material as easily was the claim. And of course, since the protons that we're interested in are being transported by the aqueous phase, you need to have water in there. In fact, the degradation of that red curve I showed you a moment ago is apparently due to a lack of water. So that was the explanation. That didn't seem to make any sense to us. First of all, we were adding only, oh, three to six. Maybe in the worst case, it's 10% by weight of our metal oxide phase to the napheon. So how could that do that? Second of all, in terms of this acid idea, napheon is a super acid. It's a sulfonic acid with an essentially it's equivalent of a trifluorosulfonic acid because the membrane backbone, the polymer backbone is fluorinated as are the side chains. So it's a very strong acid. Metal oxides might be a little acidic, but they're not going to be as acidic as that. And so one might guess, correctly it turns out, that when you add a metal oxide, if you did anything, you would withdraw protons from the system. That is, the acid's going to protonate the metal oxide and now those protons aren't available for transport. And so that doesn't make a lot of sense. And even if there wasn't some other way, there wasn't enough material here to really explain that. And by the same argument, there's not enough material here really to change the water retention properties, per se, of the napheon. That is, even if you say, well, one of the materials, for example, you're adding is silica. And we know silica gel is great at retaining water. It's not enough silica gel to make a difference. It's only a small amount of silica gel. And second of all, even if that's what it was doing, that would be a big problem because the way metal oxides retain water is they chemisorb under the metal oxide surface and they get fixed there. They're stuck there. And we need mobile water that can do proton transport. We need those waters to be able to re-orient so the protons can hop from one to the next. So the kind of water that a metal oxide's going to retain isn't the kind we want anyway. Further, you can go and start doing some experiments. So we did some thermogravometric analysis of the water content of a regular piece of napheon versus these pieces that had the metal oxide in there. No difference. They both hold water to the same extent. So that doesn't seem to help. And the final measurement that we made is we measured the conductivity of these materials. Then we measure that conductivity outside of the cell. That is, we take a freestanding membrane without any electrodes on it. And we do an AC and penis measurement of the conductivity. And it doesn't change all that much. But there is a small change. And in fact, it turns out that outside the cell, the composite membrane is less conductive to a small extent than the material that's pure napheon. So all of this argues that the concept that we're doing something good for the water is not quite right in terms of a direct interaction of the water. So what are we doing? That is, we do see when we heat up a pure piece of napheon in a cell that the mature becomes more resistive. And that certainly is best explained by a lack of water in the napheon to transport the protons. So what we're going to do is first we're going to try and figure out when we add in the metal oxide what's happening at a molecular level, one of the sorts of interactions that we're introducing by doing that. And then we'll try and move from that molecular picture up to a bulk picture and ask, given that molecular interaction, what can we predict or explain is happening in the bulk material? So we started our study very simply. We said, we started with silica. That was the data that I showed you on the previous Tafel plot. And the next thing we did, we said to get some molecular information, let's just try a series of different metal oxides. If there's anything molecular going on here, different materials ought to give a different result. And so you're now looking at a series of current voltage curves all carried out at 130 degrees. The de Guza-Holes here simply indicates that that's the manufacturer that we're using for these particular materials. It's not so simple. It turns out to be a key result, actually. We'll get back to that in a moment. But we have some titanium that we've put in here. These are 21 nanometer particles, some nanoparticles, some silica particles that we've put in there, some aluminum particles. And then the black triangles here would be a napheon that has no particles in it. We've just recast a film of napheon but didn't put any metal oxide in it. And so you can see that black curve. And we're a slightly different scaling here than we were on the other current voltage curve that I showed you. But that black curve is about the same as the pure napheon curve that I showed you previously. And you can see, if I add in silica or titanium, that I do better at 130 degrees than I do with the black curve. But I can pick certain materials, such as this alumina, and actually end up with a poorer result. So I can't just throw any piece of dirt into my napheon and get this sort of result. There's something happening here. Our breakthrough in terms of what was happening was actually had to do with the manufacturer. It turns out we purchased our titanium, which gave the best result, remember, on the last transparency, from two different manufacturers. One was the Degusa-Hulls, and the other was Alpha-Azar, AA over here. And when we made the Degusa-Hulls cell that I showed you on the last one, that's the best cell we've really seen. It's a fantastic material. We said, ah, titanium has some special properties that we like. And then we used the Alpha-Azar material and made a cell. And we sort of got lucky because that cell was worse than the worst cell we've ever seen. And if we had made that one first, we would have said, oh, titanium is bad material. And so we started asking ourselves, well, what's the difference between these two types of titanium? If you just look at the spec sheet, they're pretty much the same, slightly different particle sizes and things like that. But they should be about the same otherwise. And so we started thinking, well, maybe they made these materials two different ways. And potentially, our hypothesis was there was some organics involved in making these materials. And maybe one system had the organics left on the surface and the other one didn't. So we decided we would test out that concept. So what we did is we took some of this pore material. That's the red curve right there. Not a nice curve at all. Very high resistance. You'll notice from the slope right there. And we took it through a set of degreasing processes. We started with hexane and worked our way through methylene chloride and methanol, eventually to water and washed it. And then we took a curve. And we ended up with a curve that more or less looks like the purple or magenta curve that you're seeing there. A massive improvement. And we did the same thing now as shown on this transparency. But instead of carrying out a series of degreasing steps, we carried out a set of acid washes. Started with the pore material, took it through a series of mineral acids. Actually, we had acetic acid early on in this process, ending up with water. And again, you get a wonderful response. So it's beginning to look like there's something about the surface that isn't so good. Then we took the good material and we hit it with trichloro-methyl, excuse me. Trimethylchlorosilane to put an organic layer on the surface. There are surface TiOH groups on the titanium. And so we're reacting those with the chlorosilane bond and putting methyl, silo methyl trimethyl groups. Silo groups, let's try that again all over the place. And when you do that, you see a degradation from the nice magenta curve to the blue curve. It's starting to move down. And we found we could kind of go back and forth. You put organic stuff on the surface, gets bad. You take organic stuff off the surface by some kind of a washing procedure, it improves. So we conclude, ah, there must be a direct surface interaction between the napheon and the metal oxide, some sort of maybe bonding interaction, perhaps. And if you break that down, you don't get a good membrane material. So if you have this insulating organic material there, it's a problem. You need that direct interaction. OK. That story is a little bit more complicated than that. By the way, I should have pointed on the last transparency, but I will point it out on this one. All of these current voltage curves I'm showing you have activation or taffle regions that are identical. We're not making any change as we make these change in the rate of charge transfer at the interface. All the change is down here in the resistive region. Now, what I haven't told you so far is that the results that I've just shown you and showed you last hour only hold when you're below 100% humidity. In fact, if you really, really make sure your cell is fully humidified, then it really doesn't matter whether there's metal oxide there or not. You get a pretty good response. However, if you start to lower your humidity in the cell, this is where you see the sorts of things that I'm showing you. So here's a piece of napheon, no metal oxide in it, 100% humidity at 130 degrees, very nice response. And then even a modest drop-off to 88% relative humidity. There's a very small change in the partial pressure of water in the system. You can see it causes a major effect. 75% relative humidity, you're really down there. And below 75%, the cell just doesn't respond. Sort of the end of the line. So where we're seeing this good effect is in this region when we add the metal oxide in. So here's some metal oxide now at 75%. We could go lower than that. We've gone down to 20% or 30% actually relative humidity with the metal oxide. It works. But 75% relative humidity. And we're looking at a series of different materials, silica's from two manufacturers, titanium's from the two manufacturers. These are films that are 125 microns thick. That is this 5-mil material. We have pressurized our cell, again, because we're at a temperature where we have to be higher than the vapor pressure of water in order to get material into the cell. And this is a flow rate of the gases through our cell. And you can see that that Degusa-Holles TiO2 wins the day and that we have various other materials that give different responses. But all of them are better in this case than the control experiment, no metal oxide. It's 75% relative humidity. So this is where we're seeing the big effect. So what could be happening there? Well, there's got to be a specific interaction of some sort between the metal oxide and the polymer based on the data we have. So one guess might be that the polymer backbone somehow fizzes orbs onto that metal oxide surface, which I'm showing as a hydroxylated surface, because certainly under the conditions of humidity that we're working at and temperature, there are hydroxyl groups there. That's one possibility. Second possibility is that there's a hydrogen bonding interaction between the sulfonic acid group on the polymer and those hydroxyl groups on the surface. And the third possibility is that the sulfonic acid group on the surface coordinates to a titanium, if it's titanium oxide, for example, on the metal surface. That is a metal oxide surface. That is, there are some sites that have hydroxyl groups, but not all sites that have hydroxyl groups on a titanium or other metal oxide surface. And perhaps there's a straight interaction, a coordination interaction between the metal ion on the surface and this sulfonate group. Well, I'm just going to throw this one out without any data to start with, because it's chemically unreasonable. And remember my argument was, this is a strong acid. This is a much weaker acid. And so there's not going to be a hydrogen bond here. This might protonate that surface, but there's going to not be much of an interaction there. So it's either this one or this one that make the most sense. And by the way, if you actually do a titration of these surfaces, you find out that typically for these systems that work well, you lose about 3% of the protons from the naphthion protonating these surfaces. And that's why the conductivity drops off a little bit. And if you happen to have a metal oxide, such as alumina, where you lose a lot of protons, that is a strong acid-base interaction, then you get a rapport MEA when you make it. So this doesn't seem to make chemical sense at all. So we only have to consider the other two. And it has been very difficult to get any sort of direct data that tells us what's happening in the other two cases. But over the past year, we've come across an experiment that seems to answer the question. It's a thermogravimetric mass spec experiment. So normally, in a thermogravimetric experiment, we simply ramp the temperature on a sample. This is just a pure piece of membrane, piece of naphthion, in this case, nothing in it. And one observes the mass of the membrane as you change the temperature. And you can see there's a mass loss here. And you'll notice I put arrows here putting out all kinds of interesting events. But if you remove those arrows and just looked at that red curve, I think you would be really hard pressed to say, anything is happening there. There's nothing very distinctive about that curve. But what we are doing is, at the same time, in this case that we're getting that mass loss, we're looking in the gas phase using a mass spec at what's coming off. So if we set our mass spec to mass 18, while we carry out this temperature transition, we observe two fairly ill-defined peaks right there. One of them starts at about 100 degrees and continues above that temperature. Now we dehydrate to some extent, but not really well, the naphthion before we put it into the TG mass spec. Naphthion just sitting out in the air will rehydrate. It's a strong acid. It's rather hydrophilic. So water, we sucked into it. And so what this is, obviously mass 18 is water, is there is a certain amount of water that we have just left in our membrane. You get up to 100 degrees and it boils out. We get the second feature that starts. You'll notice around 300 degrees if you have really good eyesight. Higher than 300 degrees at the bottom there. That's a little unusual for water. Obviously it's mass 18, but 300 degrees and water don't seem to go together very well. Remember, there's no metal oxide in here. This is just naphthion, so this is just the acid. But you will notice that concomitant, same temperatures as this is coming off, you see this mass 64 material coming off, which has a mass, which is SO2. So if we take those two peaks together, that's sulfonic acid. So we see sulfonic acid falls apart in the gas phase and we get these two signatures. So a mass 18 and a mass 64 that are right on top of each other means that you have sulfonic acid. So we're cleaving the sulfonic acid groups off the membrane at that point. And then at higher temperatures out here, we see other masses that are associated. This one's associated, this kind of purpley one, with the breakdown of the side chains on the naphthion and this one with the green one there with the backbone. So you can see what's happening is the first thing that happens is we lose some water that's around. The next thing that happens is these sulfonic acid groups fall off the side chains. That's followed by the side chains themselves decomposing at much higher temperature. You'll notice well over 400 degrees and then eventually the backbone going in a similar temperature range. Now this decomposition is not bad news for the fuel cell itself because remember we're not going to run our fuel cells above, say, 150 degrees and all this exciting stuff's happening above 300 degrees. So this is just the materials analysis. It doesn't have any impact. So given this mass spectral information, now I can put these arrows on here and show you that these little wiggles in this curve really do correspond to chemical events. Now do exactly the same experiment and keep your eye, by the way, on this peak right here. But do it with a membrane that has this 3% by weight titanium in it. Same water loss peak above 100 degrees. Look what happens to that sulfonate loss peak. Very sharp at lower temperature and you'll notice the same thing happens over here. So you notice the sulfonic acid that's doing it. So that is a beautiful signature for a catalytic event. There was no catalysis beforehand when there was just naffion there. I put the metal oxide in and the metal oxide catalyzes the decomposition of the sulfonic acid. The rest of these things haven't changed really. You do see also now in the actual TGA this step where that catalysis occurs. So in other words, we're saying we get up to some magic temperature and then instantaneously all the sulfonic acids fall off. There's only one way that can happen if the metal oxide is responsible. And that is there must be a surface interaction between the metal oxide and the titanium, excuse me, and the sulfonate group or the polymer. So it must be that that third picture I showed you is the correct picture. That is that we have a metal oxide sulfonic acid interaction and it would be a covalent coordination interaction between a titanium surface site and a sulfonate group. That must explain all of this. So now we know what's happening. So what? By the way, you can run through a whole series of this really quickly. We have different metal oxides. And the punchline is if the metal oxide gives you a better fuel cell electrode, then you see the sharp catalytic peaks in the TGA. And if it doesn't help you, you see these broad things. You don't see any catalysis. So only when the sulfonic acid is coordinated to the metal oxide do you get this positive effect. No. First of all, if it did, the membrane would not work. The question is, is there enough metal oxide in there so that you have surface sites that can tie up all of the sulfonic acids? And remember, those sulfonic acids are critical to proton transport. So if you tie them all up, you'd have a very poor material. Second of all, there simply aren't enough based on the masses that are available. What we believe is happening at 300 degrees or a little higher than that is there's enough mobility in the polymer that we're getting a lot of sulfonic acids sampling the titanium surface in a relatively short time. But our picture would be something more like this. Where there's a few sulfonic acids, there's your metal oxide particle, there's your three chains of the polymer, a few sulfonic acids that are interacting via this coordination interaction, tying this whole structure together like that. So this might be a fancy way, if you will, of doing some crosslinking of your polymer chemist, but not a strong type of crosslinking. These are very weak interactions. We can't see them easily spectroscopically. The only way we're seeing them is this indirect technique that I've described to you. So what's that going to do for you? What are we going to predict for that? Well, the first thing is one would predict that we're going to cut down on polymer mobility if we do this sort of thing. So we should change the mechanical properties of the polymer. Make it a more rigid structure than if we're not doing this. Second of all, any phase transitions, and there happens to be a glassy phase transition in aphion, we would expect would change in temperature. Because again, this is a more rigid structure. And to do a phase transition, we're going to have to get these polymers to move around. They have to change positions. And it's going to be harder to do that. And finally, if we have this sort of rigidity, we expect better water retention under a stress load. What am I talking about? It sounds like I'm almost turning to an engineer here. You'll recall I told you last time the only way to do this science is by partnering with an engineering group. So let me tell you what I'm talking about. First of all, let's start with chemistry here. We have measured this phase transition to this glassy state. Remember, a naffion itself is a highly organized, self-assembled material. And so if it's going to become a more fluid state, it's going to lose that organization. And what we find is we've done a mechanical analysis dynamic, mechanical analysis here, where we're stressing the material and we're seeing how it responds to a function of temperature. And we get a transition point for the materials this way. And you can see here is pure naffion. And it really doesn't matter what we add in terms of these metal oxides. They're all good improvements. We see that transition point increases in temperature. And it turns out, in fact, that the ones that increase it the most give us the best MEAs. Now, these actual numbers over here aren't actually too important because this measurement is done in this DMA machine where the membrane is freestanding and being twisted. But you take a fuel cell, you take those MEAs that I've passed around, and you bolt them between two rather thick pieces of carbon. And that's all held in place by two big brass plates. So it is a mechanically constrained environment. And that happens to affect this transition temperature. So all I can tell you is that the transition temperature increases. The actual value in the fuel cell, I don't know from this sort of measurement. We've done another study where we have looked at the grazing angle x-ray scattering of this material. And there, we're doing it. Here's an analysis of the data as a function of the amount of water in the material. So we're looking at the spacing between diffraction events and the amount of water. And you can see that if you just have a pure piece of naphthion that you'd get this curve right here. But if you put in these various metal oxides, only at the point where you have almost no water around, do you see the same distribution of water pools in this material? We're looking at essentially a crystallinity associated with the water pools in the material. And that once water is there, you see a very different spacing for the crystallinity of this material. That is, the water pools are larger and closer together when I have the metal oxide around. And one would argue, based on the other piece of data, that if as I raise the temperature I'm going through this glassy transition, then I'm going to move from a crystalline state to a non-crystalline state, and that the metal oxide base in this data is going to help me retain that crystalline state, at least for some amount of temperature. So just to put that into a nice pictorial form, I start off with my piece of naphthion that looks something like this cartoon over here, where I have the spaghetti, which would be the polymer material. And I have these pools of salt water or acidified water, which are the green things. And the pools are connected by these channels, like I showed you before. And what I'm trying to show you here is I have a fair amount of order to those pools. And that's why I get x-ray diffraction. So there's actually a periodicity to these pools. Self-assembled, it's crystalline. I heat it up, and I evaporate some of the water. And doing that, but more importantly, I break down that order. So I no longer have a nice array of pools of water that are evenly spaced. I no longer have channels that are interconnecting them. And so at that point, clearly, I no longer have a material that is going to be a good proton conductor, moving protons from one side to the other side. So to the extent that I can delay this transition, I have a higher temperature membrane material. Now there's a second effect going on here. And it's really unfortunate, because it's a pure mechanical effect. It really isn't in the realm of chemistry, but it's very important. And that is, I just argued that this piece of napheon is squeezed between two plates. In fact, the whole thing is bolted together. It's 8 quarter 20 bolts that are holding that little square you saw together. So it is well bolted together. And so what has to happen? Well, as the cell starts to run, it makes water. Some of that water ends up in the napheon. Where does it have to end up? It has to end up in these pools. Have to expand. Well, the only way the pools can expand is that the material expands or swells. It's pushing against those plates. So if it has to do that, it's got to do a lot of work to accommodate that water that goes in. On the other hand, if I have a metal oxide phase in there, then the metal oxide phase takes up some space. It makes void spaces in the polymer membrane. It doesn't pack together as well as it would in the absence of the metal oxide. And that means when I go to draw watering, because I'm making it in my fuel cell, I have plenty of room in there to put the water in without having to expand out against these electrodes that really are fairly immovable. And so I can keep water in my membrane more effectively because of this structure, this mechanical effect. So I keep water in A because I don't do the phase transition, which is going to tend to remove water from the system, and B because of this mechanical effect. So it doesn't have to do with evaporation. It has to do with the naphthion. And if you're a big fan of stress strain curves, you can see this effect pretty nicely right here. If I have any chemical engineers in the class, I don't know, right? If I take a pure piece of naphthion and I carry out a stress strain curve, this is what I get. I get an elastic region here, and then I get an irreversible deformation above this level. I do the same thing with 3% titanium in there. And first of all, you notice my elastic region is about twice as large as before. And now I don't deform. I just stagnate when I get in this region. So in other words, I've changed the mechanical properties of the naphthion. When I have the metal oxide there, I don't have to be concerned about an irreversible deformation that destroys the structure that I have there. There's another really interesting effect. I wish I had discovered this, but my colleague, Jay Benzinger, discovered this in their cell. They were playing around. As an electrochemist, making an electrode-electrolyte interface is the easiest thing in the world, because your electrolyte is usually a solution. And so you take your electrode, you throw it in a beaker of the solution, and you say, I've done this very impressive thing. I've made an electrode-electrolyte interface, and you don't worry too much about it after that. And so as an electrochemist, it never dawned on me that perhaps how much you tighten the bolts on your fuel cell would have anything to do with the quality of your interface. But Professor Benzinger is an electrochemist, so he went and turned the bolts. And he made a rather interesting discovery. So we're looking at the cell resistance right here. We're a little messed up on this axis, but it's cell resistance taken off the linear part of those current voltage curves. And he's doing that as a function of the number of times he turns the bolts on this thing. And at first, you'll notice the resistance goes down. That's always a good sign as you turn the bolts. And then it hits a minimum. And then it starts going back up again. And at the same time, he finds in this region where there's a minimum that is below the red line here, when he looks at the current response or the voltage response that it starts to oscillate. That's always a bad sign in electrochemistry. But this oscillation is a little strange. So here's the current oscillation. Here's the voltage oscillation. And you'll notice that it always goes down to the same level here and up to the same level over here. It's a periodic oscillation. It's not kind of a random sort of noise type of thing. And in fact, he finds out that he can change the period reversibly by pulling around with the tension of the bolts in this region down here. He can dial it in. So what does he do? First thing he does is he is giving a talk at Ballard, big fuel cell company. So he's talking to them. He says, by the way, have you ever noticed anything like this? And they say, yeah, it happens on our stacks every once in a while. And he said, well, what's it due to? What do you do about it? And they say, well, if we see this behavior when we build a stack, we say that's a bad stack and we throw it away. So he started thinking about, well, what is exactly causing this? And the argument says, follows. We have our two big chunks of electrode material with the bipolar plates, the big graphite sheets in there. And it's all bolted together with these huge bolts. And we have our poor little MEA squashed in there. Now, as we start up that cell and we start to generate water, if the bolts aren't too tight, the material will swell and things will get tighter. And that's a good thing to happen because you get an improved membrane electrode interface by doing that. And you can enhance that process by starting to tighten down those bolts so you don't need as much swelling to get good contact between the membrane and the electrode. However, if you overdo it, then you get into this region over here where you've tightened this down so much that there is no room for the water in there. The membrane can't swell when the water is generated. And as a result, the water just flows out of the membrane. And so it's like a sponge. You're ringing it out. And you don't have enough water in the membrane to support the proton current. And so you get this decay. Now, right here on this critical point, what's happening? You've got just enough of a gap here and spacing so that the water can start to be generated. The material will swell. It first optimizes. You get a nice current flowing as the membrane and electrode interface optimizes. And as it optimizes, you have more current. And as a result, you have more water. And now you're pushing too hard against the electrodes. And the water is coming out because there's not enough space there. And so the structure collapses on you when you go into these oscillations. And so these are actually steady states at the top and at the bottom here associated with the cell. And in fact, Jay has now identified, I think, five different steady states depending on exactly how you assemble your cell and run it. Well, independent of the amount of current, you don't want it to oscillate, obviously. That's bad for power conditioning. But the pretty long frequencies, by the way, you probably couldn't have seen the time curve on there. But those oscillations can be over a period of an hour type of thing. So there is an optimal tightness, where you're not oscillating, but you also have optimized your membrane-electrode interactions. And it's kind of interesting. You can't just sit down and build a fuel cell. That is, it's a pretty simple device. But if you just go and buy the parts from your local fuel cell supplier and assemble them, I guarantee you it won't work. Because there's all kinds of folklore that you need to know. And for example, there are magic numbers that you tighten your bolts so you don't just don't take a wrench. You take a torque wrench and you very carefully torque yourself just the right amount. So people have known about these effects, but they never bothered to ask themselves, why is this happening at a chemical level? So you'd have to know, I want to torque at this amount and things like that. The argument is that we start off here and we have some contact between our catalyst bed and our membrane, but it's not great. And as we start to either torque it down or let the membrane swell, we improve the contact. But at some point, we overdo it. We don't get no more contact here, but we start squeezing the water out of the membrane because we've deformed the membrane too much. It turns out there's another positive effect. You don't see this effect when you're working with 125 micron thick membranes. But I think I mentioned to you, everybody wants to go to thinner membranes. And in fact, the state of the art material today is a 40 micron thick membrane. One of the issues you have to deal with your membrane material, including napheon, is a so-called crossover problem, which I haven't really talked about very much. But it's very simple. Napheon is not a permanent barrier, not an ideal barrier for the gases staying on the appropriate sides of the cell. And in particular, you can imagine hydrogen is a very mobile, small species. It can get through. And so you have a hydrogen crossover problem where the hydrogen ends up at the oxygen electrode. And the symptom of that is that your open circuit voltage falls off when that happens. There's other reasons why your open circuit might fall off, but that's a great one. And so you can see here, if I start off with 125 micron thick membrane, I get a nice open circuit here. I drop down to 40 microns. And it falls off because I'm getting hydrogen crossover, which can be measured separately, by the way. And then I go to a 40 micron composite, and it comes back up again. So that's the open circuit changing. And what I'm showing you with the bar graph is how that hydrogen, which is directly measured on the wrong side of the cell, is changing. So I have very little crossover to start with. I get a lot when I go to 40 microns, and it drops back down when I go to the composite. So I make a more mechanically rigid material, and hydrogen can't permeate through it as well. When you get to these very thin structures, that becomes a critical parameter. So just to summarize a bit, we see that there's an increase in the glass transition temperature of these materials that allows us to keep water in the membrane. We don't lose our self-assembled structure as easily when we have a metal oxide around. We see we have this improved mechanical rigidity, which gives us this ability to be stable when we're applying this mechanical stress to the electrodes to the membrane. And at the same time, we maintain good catalyst contact while eliminating this water loss due to over swelling by having the metal oxide around. Well, if you can remember, all the way back to last class, I had argued the reason for doing this is that you do not want yourself to be poisoned by carbon monoxide. I haven't said a word about that, so does it work. So some more current voltage curves. I call your attention, first of all, to the solid red squares right there. That is a cell operating under pure hydrogen, standard NAF beyond 80 degrees, right? Where it likes to operate. That's our control experiment. We now take that cell and we bleed in some CO into the cell, 100 parts per million in this case, not a tremendously large amount. And after a while, that's our current voltage curve. That is a dead cell. That is a CO poison cell, and that will not recover because the CO is irreversibly bound to the platinum at that point. We take our cell and now take the temperature up to 130 degrees and we have added in some titanium into our membrane nails so that we can do that. And we get the solid orange curve right there. And you'll notice that not only does it obviously outperform the CO poison curve, it actually is still outperforming the normal temperature NAF young curve in the absence of any CO. We're doing better. We add in some more CO, we go up to 500 parts per million and there's a slight decay in the performance you'll notice there, but still we're outperforming there. And you'll notice, by the way, for the first time that not all the data up here in the taffle region falls on top of each other. That is, we are playing now with the charge transfer kinetics by CO poisoning our surface. We have actually gone up now to 1,000 parts per million of CO and at that point the curve is about the same as this red curve. So we can tolerate up to 1,000 parts per million of CO at 130 degrees and still have a very respectable fuel cell. So yeah, it works. How long? Five minutes and 29 seconds. No, that's a very important question. First of all, it's kind of interesting to see how poisoning at these kinds of low levels in particular CO does not happen instantaneously. It takes several hours to just see the effect. So if you do a really quick look you say, ah, no problem. We have run these cells only for about 10 hours. These curves were all taken after 10 hours so that it's a fair comparison. And we see that. We have not run for the thousands of hours that one would like to do some technical limitations that we have. Others, though, have run for at least many hundreds of hours at elevated temperatures and have seen a similar result that's been reported in the literature. No, no, it's very steady. This is so, I'm showing you at least after 10 hours a very stable curve. And as I said, there's other reports in the literature that have gone up to several hundred with a stable curve. Okay, so I'm not going to repeat this again but I've thrown that in there just in case you decide to look at this off the website and want to remember what the key results are. There they are, okay. Now we get to the bonus material. We have a choice here. I think people were asking questions at least last hour and before class today about, well if you really wanted to use this fuel cell in a car, not only do you need a fuel cell that works, but you need to be able to carry hydrogen around. And that's not an electrochemical question so I really have no right to be talking about this class but it's a very interesting and important question and one that interests me. So I thought it would be okay if we took a little time anyway and ran through the issues here. So let's talk about how are we going to store hydrogen and what are the hopes and what are the realities of that? What are the issues that we have to deal with? Okay, so you can sit down and just have some fun and make up a list of the sorts of issues that come about if you want to stick some hydrogen on board your car or what have you. And the first one that seems to jump into everybody's mind is safety, am I going to detonate Caltech by building my car and running it around campus? And of course that whole safety issue really comes from New Jersey. It's not a California issue at all. It comes from this Hindenburg event that happened sometime back in New Jersey, South Jersey. This is in fact the Department of Energy has been promoting of course the use of hydrogen and fuel cells and one of their big concerns is that they're not being accidents because that would really put a damper on things. And they've, this whole safety issue they talk about as the Hindenburg effect. Okay, nobody was concerned about hydrogen obviously until the Hindenburg decided to sort of burn up. And there's this horrible picture I'm sure you've all seen it, front page of the newspaper with Hindenburg on fire and people falling down and whatnot and you see all the flames and the smoke and whatnot and it's just bad news and maybe you've also heard the news reel and nothing good here. What's interesting is that wasn't actually a hydrogen fire but the Hindenburg of course had to be able to hold the hydrogen in its canvas cloth that it was made out of and so they had to put something on that surface so the hydrogen wouldn't come out. And they chose a mixture of metal particles and metal oxide particles. It was actually, I think it was an aluminum chromate that they actually put in aluminum with a chromate. When I teach freshman chemistry, go back to Princeton and teach freshman chemistry, I do this demo which is called the thermite reaction. Has anybody not seen that reaction? Have you heard of the thermite reaction? Don't know the thermite reaction. Thermite reaction is a reaction between a metal particle of powder of metal and a metal oxide. So classically it's done with some rust, some iron oxide and some aluminum. That's how we kind of demonstrate in class. And you mix those together and of course nothing happens and then you apply some heat and you get a spectacular exothermic reaction. Fireworks, it's so hot that the iron that's formed from that reaction from the iron oxide comes out in a liquid state. Well it turns out that metal works for aluminum and iron oxide. It works for just about any metal and metal oxide. Exactly what they had on the sick end of this thing. So there was a small hydrogen fire. It wouldn't have gone anywhere if that's all had been. It would have burnt a little bit but the balloon was sectioned off so it wouldn't have been an issue. But what it did was it supplied the activation energy needed to ignite a thermite reaction in the skin. And that whole picture you see when you see the Hindenburg going down is the thermite reaction. But everybody says hydrogen, it's bad. And the whole idea was we used helium in our lighter than air machines. The Germans used hydrogen. We had an advantage. So we have safety. That picture is in everybody's mind and we don't want to do that again. The second thing is that we demand a certain range and a certain amount of power for automobiles. The automobile manufacturers have said if you can't build a hydrogen powered car that has about the same performance characteristics of a car today, don't bother to do it and nobody will buy it. I don't know if that's true or not but that is the assumption. So you need to have a certain amount of power. Nothing more or less than about 50 kilowatts and maybe up to 100 kilowatt engine. The small Toyota fleet that's running around California, that is a fuel cell SUV fleet, they have 90 kilowatt fuel cell power plants in there. And they have performance that's comparable to a gasoline SUV. The other thing that's very important lesson learned with the full electric battery vehicles is you need range. If you've got to fill up your tank every 50 miles or so, that's not good. You need the magic numbers apparently 300 miles. We expect to go about 300 miles between fillups, otherwise unacceptable. So if you factor all that in, you need to carry somewhere between 5 and 10 kilograms of hydrogen on board to get the kind of performance and range that you're interested in. So we need to be able to do that. We need to be able to do that without adding excessive weight to the vehicle, obviously. We have a little problem with the density of hydrogen. Of course, at STP, we're not going to get anywhere in terms of these numbers up here. And we have to work to get beyond STP. And the question is, what do we gain for how much PV work we do? How much should we compress this stuff to get it more dense? Should we make liquid or solid hydrogen and use that as our fuel, for example? We gain a lot by doing that. We have a volume requirement. One of the big failures of the battery car was the lack of a backseat, because it was filled with lead acid batteries and lack of a trunk and things like that. There was not a lot of room in the passenger compartment. So if you're going to fill up your backseat with hydrogen tanks, or even your trunk maybe with hydrogen tanks, that's going to be an issue. So there's a volume requirement and also a geometry requirement. That is, if tanks have to be a certain size and shape, then maybe you can't put them just anywhere you want. Imagine you know the tanks that you have in your lab today. You know what they look like. If our tanks have to look like that, they're very long, right, and things, where do you stick them in your car? You can't just put them anywhere. So there's some geometry to consider. Another issue is, obviously, you have to refill your car. So how are you going to do that? First of all, where are you going to get hydrogen from? Try buying it today, right? And the second of all, how long is it going to take to refill your car? Is it like the two minutes, like you expect to spend in the gas station right now? Two hours. And then there's always cost in the background. What's the cost? Good news about that is if gasoline keeps going up like it is, we won't be talking about that one anymore. My colleagues over in our engineering and environmental department say that once the cost of gasoline hits $5 a gallon, things start to look good for hydrogen. So we're about halfway there. You disagree with that one? I won't agree or disagree. That's just what they tell me. You disagree. Well, the assumption they're making is you're going to make your hydrogen from coal in that case. And the reason, quite simply, is if the cost of gasoline hits $5 a gallon, then the cost of oil is way up there. It's twice what it is today, right? And the cost of natural gas is also up there. And right now, if you go and buy your hydrogen from Arab products or wherever, you're getting it from steam reforming of natural gas. So it doesn't help you. So you need an alternate source of hydrogen. And they've looked at coal gasification as a potential alternate source. That's where the $5 comes from. You could also think, though, about electrolysis of water and issues like that at that point. They haven't done that analysis. OK, so what about hydrogen? How much should we work? If we're going to store hydrogen in some kind of pure form, how much do we work to get this thing on board? And what we're looking at here is the three phases of hydrogen. And if we have hydrogen gas at whatever pressure you want, then we have to have it in a steel container. Let's say that's the assumption right here. And so you're not going to do better than 1% of the mass being hydrogen in that case. Now, of course, if you go to a liquid or solid, it can be pure hydrogen. You don't need your heavy steel container. Not totally a true statement, because obviously, to keep us a liquid or a solid, we need some very, very good insulation. And there's some mass associated with that, but not a steel container. So it's not really 100%, but it's not 1%. The important point to look at is this is hydrogen gas stored at 200 bars at room temperature. This is liquid hydrogen, 20 degrees Kelvin. And there's solid hydrogen, 4 degrees Kelvin. Now, if you look at the density, you're not gaining that much. That is, the first thing you'll notice, there would be really absolutely no reason to go from liquid to solid. The gain here is trivially small. So mostly you want to go to liquid. But going from high pressure hydrogen to liquid is only a factor about 5. And there's a big, big energy difference in going from high pressure hydrogen to liquid. It's a lot more than a factor of 5. So if you're going to use hydrogen in a pure form, it's going to be high pressure hydrogen. Energy-wise, it just doesn't make sense to do it otherwise. Now, by the way, remember those numbers for density, about 0.01 ballpark, 0.1 to 0.01. So what are our options? Steel tanks. We have steel tanks. No problem today that I'll do 2,000 to 5,000 PSI. Lab tank is usually somewhere around 2,02500 PSI. 5,000 is not a problem. They've been built, they've been around forever. So no one technology, great safety record if you ignore a few minor mishaps that weren't hydrogen but other high-pressure gases. There is, for example, the steel tank that is embedded in the ceiling of the chemistry laboratories at Harvard. Because some graduate student many years ago foolishly decided to transport a tank without taking the regulator off and putting the cap on. And it dropped. And the regulator broke off. And there was a missile-type effect, which fortunately missed everybody, but the tank ended up embedded in the ceiling. And they decided they would leave that tank there as a reminder to all future graduate students that when you move a tank, you better not have a regulator on it. So good safety record, assuming you follow the rules. That is the weak leak in any tank as the regulator. Some have said, said, that metal tanks are subject to hydrogen embrittlement. And this would be an issue. I'm not sure this is really an issue in that in the lab we've been using hydrogen tanks for a long time, and we don't have them failing because of hydrogen embrittlement, but certainly some inspection routine would be needed. So there's a little something going on there, but I'm not sure that's a real problem. If that tank fails for any reason, whether it's the regulator or the tank, then you have pieces of steel flying all over the place. So you have shrapnel. And that's a big problem. That is clearly a safety problem. So in other words, if you have a steel tank, then you somehow have to armor the container that it's in. So if it does fail, the shrapnel doesn't go and wipe out the 405 at rush hour type of thing. I just mentioned that. It was totally backed up today. And everybody was sitting on it. I was imagining a hydrogen car detonating in the middle of all this traffic. But anyhow, it wasn't a pretty sight. Now, there's another problem which you're probably not familiar with because you get to use the hydrogen coming out of the tank. You don't have to fill the tank. But it's not that easy to fill a hydrogen tank. You do know that they're heavy also. If you've had a lug one around, that's an issue. But you also know that most of that weight is not due to the hydrogen. It's due to the steel. So you're paying a big weight cost just to hold the pressure in. You have poor volatil metric storage in part because hydrogen, when you get up to the sorts of pressures that you're dealing with, is very non-ideal. That is, ideal gas law is not a good approximation. You have pretty big numbers for your A and B coefficients in the Van der Waal equation. So you're losing something like 20%. If you took the ideal gas law and said, oh, I'll get this much hydrogen under these conditions of pressure and volume, you'd find you're off by 20% in a real life hydrogen tank at those pressures. That's a big loss. The other problem is that hydrogen heats up when you pressurize it quite a bit. And so if you think you're going to pressurize a tank in two minutes, you better have a really good way of dumping the heat because you're going to have a very, very hot tank, which probably is not a good idea around hydrogen. So those are issues, but you can see they're engineering issues. One could get around them. Another solution, the solution at the Department of Energy by far favors at the moment is they say, tank technology is a good technology. Steel is bad, so let's go to these composite tanks. The composite tanks today are rated for 7,000 PSI. It's a carbon composite. It's a wound carbon filament here. The inside of the tank has a polyethylene sleeve, which is necessary to keep the hydrogen in there. It'll leak out of the carbon windings otherwise. It is believed that they will be rated for 10,000 PSI in the not too distant future. Already they're storing something like 7% hydrogen by weight. That is, we have a lot of pressure and a low mass on the tank material, so it goes way up there. So we have high storage capacity. We have lightweight. This tank, unlike a steel tank, will not fragment upon failure. If the tank fails, it's very clever. The polyethylene insert powders, which is really good news on two fronts. First of all, it takes a lot of energy to break all the polyethylene down into a powder, so the whole shockwave doesn't get out of the tank. And the second is, you're not going to have some fragment shooting through the passenger compartment of the car next to you and taking out the lever. They're expensive. These are tanks that were developed by NASA and they're pretty expensive. There's another issue in that there's a tank, but right there's the hole. And you have to put a regulator in there. And you have the same limitations on regulators that you have on steel tanks. That's the weak link. So you have to armor that area, at least. What about other options? The first option up here has been more or less ruled out. And that is, we know how to put gasoline in a car, so let's put gasoline in a car. Instead of burning it in an engine, we'll put it through a reformer that's right in the car. We'll turn it into hydrogen and CO2. We'll feed the hydrogen into our fuel cell and away we go. And Honda was in favor of this approach for a while, but they have abandoned it also. You have a nice thing. You have the energy density of gasoline. It's a liquid. It's got a lot of hydrogen in it. It has a lot of energy in it. Everything's good there. You have great access to gasoline, obviously. You don't have to worry about where does the hydrogen come from? System integration is poor. What is that? Reforming takes place at about 07 to 800 degrees C, because it's an endothermic reaction of room temperature. Keep it up to make it free energy negative. And the fuel cell that we're talking about runs at, let's say, 150 degrees on a good day. So you're going to take this 900 degree box, and you want to bolt it to this 150 degree box. And so there's some engineering tricks there about how you do that and keep them all happy. A separate problem is you're going to push your foot on your accelerator, your gas pedal. And instead of injecting gasoline into an engine, you're going to be turning on a chemical factory that takes gasoline, runs it through all these reforming steps to get to hydrogen and whatnot, and then takes it into a fuel cell. And so that kind of 0 to 60 business in reasonable time, usually chemical factories don't work that fast. So you have issues. And then finally, one of the big things that have been touted for a fuel cell car is that, A, it's going to be good in terms of not spewing out environmentally negative gases. By that, we mean the kinds of things that tend to make smog in this area. But B, it's also now supposed to save us from the greenhouse effect. That is, we don't want to spew out CO2. And of course, if you're using gasoline through a reformer, nothing's changed. In fact, you may, if you do it wrong, spew out more CO2 than you spew out of a car burning the gasoline. I think that would be true also of a maintenance. Well, no, the idea is you make hydrogen from coal in some factory somewhere, and you trap all the CO2 and do whatever you're going to do with it. And so if you have to trap the CO2 coming out of tailpipes, you can't even think about winning. At least you can think about, we'll ship it to a neighboring state or something like that and make it a factory. There's possibilities. One of the favored possibilities now is using a metal hydride, a pure solid state system. Everybody agrees it's safe, big plus. That hydrogen is not going to go anywhere. You don't need a big steel tank now, but there are metal hydrides. There are metal alloys, so they are intrinsically heavy. You've got a big chunk of metal in your car. So you have a weight issue, very expensive materials. These aren't your run-of-the-mill metals that we're talking about. That's an issue. The thermodynamics and the kinetics is not working well for you if you want to think about getting the hydrogen in and out of the metal hydride. That is, the safety issue is based on the fact that when the hydrogen is in the metal hydride, it's sitting in a nice, deep thermodynamic well. That is, delta G is very positive for getting it out. But of course, you want to get it out when you run your car. So you have to heat up your system quite a bit to get the hydrogen out. So there's a big waste of energy in getting the hydrogen out. On the other hand, you need to put hydrogen in when you're refueling your car, and that's going to have a major exotherm associated with it. And so you have to be able to take care of that. Again, you can't fill it up too quickly, or you'll generate too much heat, and you'll have issues. So the in and out here is somewhat of an issue. Now if you look at sort of the metal hydrides that people are talking about, and right now the favored one seems to be the magnesium hydride, if you look at the weight percent, you'll notice it's very favorable compared to pure hydrogen or tank hydrogen, assuming it's a steel tank. Now the magnesium hydride can do as good as those composite tanks. So you're doing pretty good there. The other interesting thing is you'll notice the actual density of hydrogen in the materials higher than a pure phase of hydrogen. It's comparable to solid hydrogen, a little better actually. So that all looks pretty good if you can get around the other issues. So following the same line of thought, what about some other chemical hydrides instead of metal hydrides? Yes. If you remember, because why is it, how could hydrogen be more dense in a metal hydride than in solid hydrogen? Because in solid hydrogen, it's H2, right? And there's a certain metal hydrogen-hydrogen distance in H2 that we can't get beyond that old 612 well thing. But in the metal hydride, it's sort of hydrogen atoms. So it actually can be more dense. Not much more dense, but yeah. OK, so a chemical hydride, such as sodium borohydride or there are aluminum hydrides that have been suggested. I have to give you a disclaimer here. I am picking sodium borohydride and the hydrogen-on-demand system. That is a trademark name for a system made by Millennium Cell. And I'm affiliated with that company. So I'm going to tell you about it, because I happen to know more about that one than the others. But I don't mean to suggest it is the solution. Might be the solution, but you now know where I stand on this. So what we're talking about is if you take sodium borohydride action, you just put it into water. You start to bubble out hydrogen. But if you put it in strongly basic aqueous solution, say 2 to 10 molar depending on sodium hydroxide, it is indefinitely stable. It'll sit there on the shelf forever as sodium borohydride. And if you then have a catalyst that you can pass the solution over that will strip up the hydrogen, you have a nice way of storing hydrogen. And so the idea would be that you have a little, this is just a teeny little system, not a car-sized system. But you have two reservoirs. One reservoir that has the aqueous, basic sodium borohydride solution in it. A second reservoir that's going to be empty, you have a little catalyst system down here and the ability to take the hydrogen out and you bubble the hydrogen out. And then you have this remaining solution of sodium borate aqueous that you end up pumping into the second canister over there. That's your waste from the system. It's not too bad a waste. You'll notice, and that's borox. That's cleaning solution. So it's not too, too bad. On the other hand, if we went to this system, the bad news is that what are we going to do? We're going to have mountains of broxo lying around on our corner stations because we have to empty this stuff out. That doesn't seem to make much sense. Now, we need a way of taking that material and reforming it back into sodium borohydride. And it has to be energy efficient and cost efficient to do that. And although this part of the cycle, the part I'm showing you works just fine, it's wonderful. The recycle part has not been worked out well enough yet to be practical. But assuming you can do that, and that is a big but, you have the nice advantage of your system is not going to burn up. It's water. So you don't have to worry about a fire from the system. You get a very high effective pressure of hydrogen, 7,000 psi, because you can make very concentrated solutions of sodium borohydride, and B, because you can store them just in polyethylene containers. So I showed you this Daimler Chrysler unit test car before. That uses just a plastic tank in place of the gasoline tank, 20 gallons, and it has a 300-mile range. Non-flammable, yes. Maybe that's a bad word that it doesn't burn. No fire. Let up front. I didn't switch it over. Yeah, yeah, you're absolutely right. That's the wonderful thing about PowerPoints. Tom, you're going to fix that before you put that up, right? It's a low-volume system. It's a simple system, just a solution and a catalyst. Red is chemical safety. Although millennium cell touts that this is a safe system, it doesn't burn, they don't have to mention what happens if you decide to take a bath in 10 molar sodium hydroxide. So there are some issues there. If your tank were to break and it was to spill on somebody something, then there obviously are some corrosion issues. That one would be concerned about sort of a trade-off with gasoline. No, but the battery is a small little thing and totally sealed, and this is something that you have to be adding to and probably 20 gallons and not totally sealed. And you could imagine maybe in an accident or something like that, the tank might crack. You could see how this could happen. Or somebody fooling around the sodium borohydride station, I was going to call it a gas station, whatever it is, decides to have a squirt you with whatever. It would be an issue, I guess. So you have to deal with that. We talked about the recycle issue. Can't wait. It won't burn, though. Can you make sodium borohydride at reasonable cost? And then right now, this is just too new a system. So any guess as to what it would cost to actually do this is a wild guess. So whether it's ridiculously expensive or reasonable is very unclear. But it does work, and it does scale nicely from this little thing up to the car size. The story is sodium borohydride today is used for two things. The primary use is in the solid fuel boosters of the shuttle. It's the reducing agent. And then another use actually is in as a whitening agent in the paper industry. It's a fairly new use for it. As a result, you don't have sodium borohydride factories all over the country. In fact, you have basically one sodium borohydride factory. It's passed hands from different companies. Today it's owned by Roman Haas. The technology in that factory, although the most recent factory was built, I think, perhaps in the 70s it was rebuilt, is 1950s technology. That's when the original factory was set up, and they're doing it the same way. And basically what it is is sodium metal to sodium hydride, sodium hydride reacting with the borate, which you just dig out of the ground to make sodium borohydride. Now you're going way uphill in energy to make the sodium metal, and then just dropping it off as you get back down to the sodium borohydride. So this is a very expensive approach, both in terms of energy and cost. But of course, if your major client is NASA, that's OK. So there are a number of research groups that are very interested in some more reasonable approaches to that recycle. One approach that is being considered is an electrochemical approach, where you would take sodium borate and electrochemically convert to sodium borohydride. We've started actually working on that in our laboratories. You need a good electrocatalyst for that, which is what we're interested in in appropriate electrolytes. You're obviously not going to do this in water for the obvious reasons. So it's a non-aquia system. Another approach is just finding a good chemical reducing agent where the activation barrier isn't very large. That will let you do this. And that reducing agent is recyclable. And so far, although there are reducing agents, there aren't reducing agents that you can turn over and get back to their reduced state. So it's an open question at this point. Hydrogen itself won't do it. That is, nobody knows what this is a multi-step process, so this catalyst is not a reversible catalyst. Whoops, and what just happened there? Let's get back in there. Let's get back to where we want to be. OK, we're there. So those are, and again, you can do this now with aluminates, aluminum hydrides, with various organic substituents on it. And those systems do tend to be more reversible. So there is some understanding of how you would go back here with the aluminum-based systems. That's work that's primarily been done at the University of Hawaii. But on the other hand, the sodium borohydride does a better job of storing hydrogen. So sodium borohydride is better right now in this direction. The aluminum systems, perhaps, are better in the other direction. And nobody really knows where this is all going to come out. So those are your options. Throw it open to you for questions. OK, no questions. OK, so let me tell you where we're going to move on Thursday. So essentially, everything I've set up to now in this class is more or less background information. That is, if a random student had walked up to me in the hall and said, I decided I want to do an electrochemistry experiment or my professor's decided I want to do an electrochemistry experiment and I want to know what to do and how to do it and what technique I should use, you haven't learned a thing so far. You've learned all the important background information. So where we will start on Thursday is with the pragmatic details of a potential-controlled electrochemistry experiment. Remember, there's a homework problem set on the website. Do next Tuesday, and we'll see you on Thursday.