 Good morning everyone. Let me first mention that Nick is actually one of my co-conspirator in this project that we started a little more than three years ago. So it's my great pleasure to talk about it. And I would like to highlight this is really the work of three very talented graduate students in postdoc, Xiaofei, leading in a madure, who have made this work possible. So before I start the work, let me just introduce myself briefly. In my group, we're looking at very basic electrochemical process such as ion insertion in batteries, fuel cells, and solar fuel process, and our approach is to establish design rules and rational engineering principle for these devices. So I don't think I need to tell this audience the importance of making carbon neutral energy available when and where it's needed. And this is just a picture showing you why this must be the case. And over the past few decades, there's really being a lot of progress made on how to use photo-electrochemical devices to enhance the utilization of solar energy, and particularly to convert it to chemical fuel by, for example, dissociating water into hydrogen and oxygen. And the biggest challenge in this process is really about using the entire solar spectrum. So if you look at the solar spectrum, you have ultraviolet, visible, and infrared, and often what happens is you're able to capture part of the spectrum, but not all. And what happens to the part of the spectrum that is not captured often ends up as heat. And in many of the current photo-electrochemical and solar cell processes, the heat is discarded and not utilized further. So the basic question we ask in this GSEP project is can we use the thermal energy to help us improve the performance of solar field process? And very briefly in a photo-electrochemical cell, you have a material that typically absorbs light, converts them into electron hole pairs, which then reacts with redox processes in liquid, typically in a water-based solution. So this is just a picture showing you what might be possible. Can we make earth abundant material like iron-based oxides and combine that with thermal energy to make the process more efficient? And one of the initial idea we had was to look at how transport properties of semiconductor depend on temperature. So as it turns out, many of these oxides, which are earth abundant, also have very low charge-carry mobility, which means it's hard to take the electrons or holes out of the material without them recombining and thereby converting them indeed. And this is just a plot of the mobility comparing silicon and iron oxide. And as you can see in silicon, the mobility is largely independent and temperature is very high, about a 100 centimeter square per volt per second. But for iron oxide, it's immeasurable at room temperature. And with increasing temperature, you slowly climb up, but you need to go to several hundred degrees Celsius. It turns out that this is a very general phenomena in metal oxides, because the d-orbitals in oxides, especially in the transition metal oxides, are typically very narrow. So this gives rise to quick, rather low, electron mobility, and this really limits the performance of this material. So the idea was, well, can we turn some of these oxides into better performing oxides by cranking up the temperature a little bit? And the temperature here would, in principle, come from the inefficient absorption of sunlight. And then, rather than throwing that away or by cooling the process, we simply turn that into activation energy to overcome some of the limiting process in the cells. So about three years ago, we started with iron oxide that is substituted with titanium. And we're trying to here perform the oxygen-evolution reaction. So this is the typical type of plots we have. We have current density as the y-axis and the potential as the x-axis. And what you want, it is a voltage curve which shifts to the left as you turn the light as much as possible. So this is just one curve showing you at 25 degrees Celsius what that looks like. And then on the dash line is what happens without the light. So the difference corresponds to the energy that is generated by absorbing light. So you can see in the dark, you have a quite high onset potential. So the current takes off at very high potential. When you turn on the light, the potential shifts by about 700 millivolts. And this is very nice. So what we want to happen as we turn on the heat to up the temperature, we want the curve to shift further to the left. But instead, what we see is as we crank up the temperature, the curve is actually shifting to the right, which means that the driving force is decreasing. So this was a bit disappointing. And moreover, if you look at what is called the saturation photocurrent, it is also not changing with temperature either. So this was rather disappointing. But there was one hope that if we look at the dark current, so this is how well the material behaved at without the light, you can see it is shifting in the right direction, meaning that potentially the catalysis is getting a little bit easier as you crank up the temperature. So we were thinking about why this was the case. So we went back to a simple electrochemistry and looked at how the carriers are being generated in the material. So it turns out that in the material that we're looking at, the activity at the surface, the efficiency for the material to extract, the carriers into the solution is actually quite slow. So all the benefit you get from enhancing with temperature may just get eaten up by recombination process at the interface. So to remove this effect, this was actually a suggestion from Nate a number of years ago, is to just look at a fast redox couple replacing water redox couple with just a sulfite couple. So this is a very quick redox couple, allow you to extract the carrier. It's as if you have a perfect catalyst on the surface without having to have one. And the results were very encouraging. Once we replaced the water oxidation reaction with the sulfite oxidation reaction, as you can see here in the top right, then the current began to increase with the temperature. So the blue line is at 7 degrees Celsius and the red line at 72 degrees Celsius. You can see the current is slowly climbing. The voltage is still shifting a little bit, but it is much better than the previous case. And on the bottom, I'm showing you how much the current is enhanced over room temperature. So without the sulfite fast redox couple to scavenge the hole that is being generated, you have a very, very flat line. That's what's in black. But when you put sulfite, then the curve goes up quite a bit and the coefficient we're getting is about 1% per Kelvin. So every degree you turn up, the photo current increases by 1%. This was good, but not great. We were predicting something much larger on the order about 5%. So we asked the question, what is going on here? And we come up with a simple picture, which I will come back to in a little bit. There are two regions in the material in which you can perform the separation of the electrons and holes as the light is absorbed by the semiconductor. One region is assisted by the electric field. This is known as a space charge. So this is the region here denoted by W naught. And there's the second region in which the charge separation is facilitated by the diffusion of the carrier. So this is driven by a concentration gradient of the charge carrier. And this is known as the minority carrier diffusion region. So it turns out for a material like iron oxide, the two layers are comparable in thickness, a few nanometer for the space charge region and a few nanometer for the minority carrier diffusion length. And it turns out that the space charge region is not strongly activated by temperature because the mobility really governs how thick the minority carrier diffusion length and it only changes the space charge region negligibly. So the second idea we had was, well, can we increase the minority carrier diffusion length so that we can absorb more light in the region in which the temperature enhancement will be more noticeable? So we switched to a different material. This is bismuth venidate that is doped with malibnum. And this is the type of nanostructure that we have been able to create by a combination of electro deposition and annealing. And let me just point out the minority carrier diffusion length here is much larger than iron oxide. So it's 2 to 4 nanometer for iron oxide and about 70 to 100 nanometer in bismuth venidate. The space charge region as comparably thick in both is on the order of a few to 10 nanometer. This is resulting from the high doping level that we have to do in order to get enough majority carrier mobility. And the results were even more encouraging. So these are showing the IV curves with and without light. So without light is the black dash curve with light are blue going to red as we increase the temperature just by 32 degrees Celsius. We're able to increase the saturation photo current density from about 2 milliamp per centimeter square to almost 3.5. So if I plot this and I show you the coefficients, now we're up to about 4 percent per Kelvin and the shift in the voltage, it's only about 2 millivolts per Kelvin. So this is a really impressive performance because the loss in the voltage is relatively small compared to the gain in the current. So on the net basis, you're getting more photo power out of this system. And if I compare this to the iron oxide example I have on the previous slide, we have almost a flat line showing that 1 percent behavior and iron oxide. But in the red line I showed the bismuth venidate is now up to 4 percent. So here we're really seeing a very dramatic enhancement by temperature simply by having more of the light being absorbed in the minority carrier diffusion length rather than in the space charge region. So we took advantage of this and we created a bismuth venidate cell that is coupled with a silicon tin oxide junction. So this is essentially embedding a photovoltaic into the system. On the left, you can see a cross-sectional view with bismuth venidate on the top. We have an omic contact layer tin oxide in between, which is also transparent, and then we have n-type silicon as a substrate. The reason why we added the silicon is if you notice on the previous slide, we were losing voltage about 2 millivolts per Kelvin, so it would be good to offset that. So by adding a silicon solar cell embedded in our photoelectrochemical electrode, then we're able to push the voltage further to the left, which is what you desire. We're able to add about 300 millivolts additionally to the system, and by comparing against other type of bismuth venidate cell, we're able to achieve a very low onset potential and also a very high current density. And this is by combining both managing the voltage and also enhancing the current with temperature. So this was a very nice demonstration showing that temperature and thermal energy can actually be a viable tool for tuning the performance of devices and elevating the temperature will give you a net improvement performance rather than a decrease, which is commonly observed in solar cells. So the next experiment we thought West will understand what's going on and can we validate the model I proposed in the competition of light absorption between the space charge region and the minority carry diffusion region. What we did now was to look at a third material, titanium oxide, and this is a very well established model system for studying photoelectric chemistry. The performance is typically not very good because of the large bang gap, but nevertheless we're able to see what is going on. So what we have done here, it is to make nano wires of titanium dioxide so we can control the diameters and the distances and the relative transport length more accurately. This is just one of those nano wires you can see in the cross-sectional image there. We're showing you again the improvement in the photo current as a function of temperature. The increases is okay, it's on the order about 1% per Kelvin, so typical comparable to the iron oxide case. And what we have done here it is to use the controllability of the synthesis by varying the diameter of the wire. So as you increase the diameter of the wire, the region which is occupied by the minority carry diffusion length is actually getting smaller and smaller. As you increase the temperature for the smaller wires, eventually you'll run out of enhancement because you would have eventually activated entire wire by increasing the minority carry diffusion length. But for the larger wire you will tend to have still a very much inactive core. So simply by using geometry, varying temperature and the wire diameter at the same time, we can try to test our theory that it is the light absorption competition between the space charge region and in the minority carry diffusion region that's controlling the performance. And this is just to show you that we have fabricated three different diameter nano wire. And to convince you that the surface of these nano wires are the same, we carried out high resolution imaging on the electron microscope just to confirm that the termination layer where the titanium intersects with the liquid are in fact the same. So the difference between the three wires are only the diameter. And what I'm about to show you is a data set that looks at the photo current as a function of temperature for each diameter. And what we have done here, it is to develop a simple geometric model to fit all the data points for all the sample using one equation. So you can see when the wire is very, very thick, such as the green line over there, it's very much increasing linearly with temperature. So that means as you increase the temperature, you're getting more and more current because you're not running out of wire to absorb the light. But when you have a very small wire, eventually you are activating the entire wire through thermal activation of the minority carry diffusion length. That's why the blue line tends to plateau. And our model perfectly captured this transition from a linear behavior to a more plateau behavior. And from this data set, which took three sample and about a day to measure, we're actually able to back out the minority carry diffusion length, the activation energy of minority carrier transport, and also the space charger layer width by using the equation I have down there, just by considering the geometry, the diameter of the wire, the thickness of the space charge region, and the thickness of the minority carry diffusion. There's a lot of assumption, but we are very happy that such a simple model can explain the data accurately. And this is a plot showing you how temperature is activating the minority carry diffusion length from about 6 nanometer at 10 degrees Celsius to 12 nanometer at 70 degrees Celsius. And we're also able to extract a space charge layer width of 9 nanometers, which is very comparable to what's expected for this substitution level in Titania. And we're also able to back out an activation energy of 0.1 electron volt, which is consistent with the type of electron trapping that you see in Titania. So it turns out that the fraction of light collected within the minority carry diffusion length is what's dictating the thermal enhancement. And here we provide a very unified view of temperature enhancement. This plot shows the performance or the extent of temperature enhancement as a function of the space charge layer width and also the minority carry diffusion length. And you can see why iron oxide didn't work very well to begin with. The dash line shows where it is equal, so this is same diffusion length as the space charge layer. Iron oxide sits pretty much on the line, so that gives you a fairly low enhancement. Titania sits a little bit further to the right, which is good. That means more light is absorbed in the region that can be activated by heat. But Bismuth-Venadate sits far to the right because the minority carry diffusion length is much longer than the space charge layer width. And this is consistent with our experimental observation that we have minimum enhancement for iron oxide, a little bit better enhancement for Titania, and much better enhancement for Bismuth-Venadate. And for the last part of my talk, I want to discuss how to go above 100 degrees Celsius. So if you don't pressurize the water, basically they will boil at about 100 degrees Celsius. And we see that there is a benefit to temperature enhancement, and the next question we ask is, can we get more out of it by going above 100C? So we needed to replace the water with a solid-state electrolyte, and what we've developed here is an all-oxide approach and replacing water with an oxide-based solid electrolyte. And this is the type of cell that we have prepared. My student, Madur, has been working very hard over the past few years to make this happen. We're using, as a demonstration system, Itria stabilized zirconia as an oxygen ion conductor. We fabricated a very thin membrane on the order of a few hundred nanometer, which is supported on a silicon substrate, but this is not photoactive in this particular setting. The scanning electron microscope images and the chemical map below shows you the cell from the top. What we have done here is we needed to add one more ionic conductor. So if you just put something like bismuth-venidate on zirconia, then only the intersection between the two will be active. That's where you can extract the photo carrier, but it will be good to surround the bismuth-venidate with something like water, something that can also help you transport the ions. And that's what we have done here. We added, in addition to the material that absorbs light, a material that also transports ion as well. And this, we also use bismuth-based material for compatibility. So we're using here bismuth-copper-vanadium oxide, which is an excellent ionic conductor. We mix that in with the bismuth-venidate you can see here in the color images. The red images here shows you the distribution of copper, which is not present in bismuth-venidate. And by mixing the two at the boundary of these interfaces, you have the electron-hole pair coming from the bismuth-venidate. You have the oxygen ion as your mobile carrier in the bismuth-copper-vanadium oxide, and you have the whole thing on a solid-state electrolyte. So we examine this system as a function of temperature, just like the bismuth-venidate in water. So here it's a plot on the left of the photovoltaic as a function of temperature. The blue shows you the liquid system. So you can see there's a slight slope. The photovoltaic decreases. So we extrapolated that all the way down to about 400 C, and this is where the photovoltaic would go to zero. And this is just from the expectation that the material is becoming more intrinsic, so you're losing your junction at the interface. The red curve shows the cell. So it close up from almost zero voltage at low temperature, because the ions are not mobile, but as it increases the temperature, the volta voltage actually goes up. So this is opposite that of the liquid system, and this is one demonstration that you can actually use a ionic carrier in a solid as a way to facilitate charge separation. We've also done experiment where we took out various component of our electrode. It's only when you have both bismuth-venidate and the bismuth-copper-venidate then you will get an excellent photovoltaic. On the right here I'm showing the current density with and without light, and you can see that we're getting quite an appreciable photocurrent. We're now able to reach almost 100 milliampere square centimeter, and our goal, which I keep increasing, is now 1,000 milliampere square centimeter. So this system has the advantage of high current, but low voltage. So this is something we're trying to understand how we can make use of it, certainly increasing the junctions in the system, or by combining with other processes or external bias, this can become a useful approach. So to conclude, let me just show you the following schematics on how we think about thermally enhancing the generation of solar fields. In the conventional world, we're looking at solar cells and PEC. We're taking light, and we're shining on it, and we're getting solar fuels or electricity out of it. If we're able to concentrate the light, as it is sometimes done in concentrated photovoltaics, we often have to cool the system in order to prevent the degradation of the performance. But as shown in this work, it may be helpful to consider how the heat can actually benefit carrier transport and catalysis in certain systems, and then we can propose something maybe without the cooling. And then we can take advantage of the heat that is being generated. And with that, I would like to thank you very much for your attention. Thank you. So I meant to ask, many of the structural modifications that you do will significantly influence the optical absorption properties as well. So have you tried to exploit that possibilities for design of your system? Yeah, Nick and I have been thinking that from day one, but it has taken us longer than we thought to get to this point, so that we'll have to raise more resources to tackle that challenge. Great. Nate, thanks. It was really good. The voltage in solar cells goes down 2 millivolts a degree. Like we predicted it was going to do. So you're getting more current. Eventually you'll get a lot of current, but these high temperature cells, fundamentally. But one can learn what the recombination mechanism actually is by looking at that temperature dependence in that high current limit when you keep constant current and then don't have the confounding effects of more current giving you more photo voltage when you should be getting less. So have you deconvoluted all that and does it make sense that you want to run these at high temperature while just inherently having more recombination and losing voltage or do you really want to cool them and use the heat for something else? Well actually, Nate, so from the photo voltage dependence on temperature, our current understanding is that actually it is dominated more by the loss of Fermi-level splitting as opposed to the dependence of recombination on temperature. And I absolutely agree with you that for the high temperature case, the photo voltage is falling too much. So basically intersection point is not optimal right now. So for the liquid cells, that's actually much more advantageous, because in that region, there is actually a peak photo power that lies above ambient. But for the solid state approach right now, because the photo voltage is falling too quickly, so where the photo current is picking up, it's already beneath the optimal photo power output. So the two approaches we're considering right now is one, can we envision this as more as an light enhancement of electrolysis cell rather than just having a standalone PC? Two, then can we think about ways to try to improve the photo voltage so that moves the voltage up? But you're absolutely right that I think for the liquid case, there is a much better case for increasing the temperature slightly above ambient to get peak photo power. But for the solid state, right now at 300 degrees Celsius, we're above the peak power. So we can find an ionic conductor that can go down to 150 or 200 degrees Celsius with a suitable photo current, then we're talking. Thank you, Nate. Other questions? Well, maybe I can ask one on our own work is always an interesting thing to do. So one of the things that comes up often is where is the heat going to come from, right? So if we do want to go to these higher temperatures, are you going to have to have a photo cell plus heat? Do you have enough heat from the solar spectrum to do these kind of processes? Well, for the high temperature, very uncertainly, some degree of concentration will be needed and of course spectral selectivity will help with that. The low temperature I'm not too concerned to get up to 80 degrees Celsius is fairly straightforward simply by absorbing some of the IR passing through. So I think in the former cases, we're talking about almost a costless improvement. To perform as in the second case, we have to rigorously consider all the costs associated with increasing the concentration to get the temperature you want. But even just with a few, say, 10-sun concentration, you should be able to get to 300C with good spectral selectivity for the cells. Great. Any other questions? If not, let's thank Will again. Thank you.