 It's been really exciting over the last several years to see the solar industry finally take off. Here in the red curve, you see the price of solar cells as a function of time. Back in the 70s, it was about $100 per watt that you can get from a panel at noon under peak sunlight. And that dropped and then there was a stagnation where there was very little interest in renewable energy and so not a lot of research. And then more recently, the prices have dived and now it's down at around 50 or 60 cents per watt. And in a slide or two, I'll put that number into perspective for those of you who don't follow these numbers. But as, well, you can kind of get an idea that things get interesting at about a dollar per watt because when it got there, that's when the market really took off. And now about 65,000 megawatts are installed each year which is about the equivalent of 65 large coal-fired power plants. So we still have a long way to go but things are taking off and growing around 30 or 40% per year which is very large for something that requires manufacturing. And I'll probably don't need to say it this week but I think we're at a tipping point. People are taking climate change very seriously now all across the board which is good for the solar industry. Solar cells are primarily deployed in three sectors on homes and generating the power right where it's used. On large buildings in the commercial sector, you can reduce the installation costs as you go to these larger installations. And then you get the best cost when you go out to the utility scale and we're seeing a lot of this in California if you ever fly from Phoenix to San Francisco keep an eye out the airplane and you'll see a lot of this stuff out in the desert. And here you can see from the Department of Energy some of the numbers and on the residential scale. If you know the dollars per watt, you can calculate the cost of the electricity over the lifetime of the project to get what we call the levelized cost of electricity and without a subsidy in the residential scale right now that's around 20% but currently we do have a 30% subsidy and that brings it down to 14. DOE has set a goal that that needs to get to eight. That's when people would use solar pretty much everywhere but here in California, most people pay more than that and as a homeowner I get an electricity bill and around here we have tiers and up to a certain amount of electricity, you're in tier one and then you go to tier two. Tier four around here is 40 cents per kilowatt hour and if you ever use your air conditioner you get into that and so solar is very popular around here and what I think is gonna happen is electric cars are gonna become more and more popular. Eventually companies are not gonna allow people to charge up for free. I know of no employer that has a gas pump in the parking lot and allows people to have a free fill up whenever they want it so that's not gonna continue with electricity. People are gonna have absolutely ridiculous electricity bills at home when they have an electric vehicle and anyone who's smart enough to be able to afford a Tesla will be smart enough to switch to solar at that point and so I think it will really take off. Maybe skipping ahead to the utility scale, you see the prices come way down because it's just easier to do these large installations. Here the permit can be really expensive and it's not divided out over a whole lot of watts. Here it can get divided out of megawatts of power that are generated and you see that we still have maybe a factor of two or three to go in the cost reduction so the job is not done. At this point and you kinda get one, the difference between these two is not technology, it's the balance of systems and you're getting a hint that the installation is now a really significant part of the cost. It's probably about 75% here and maybe half there and so there certainly needs to be work in more efficiently installing the systems. It's not what I do, so it's not what I'm gonna talk about today but what people can do who make panels is they can just raise the efficiency and then you don't have to install as many panels and you effectively lower the balance of systems and also before I say more about that, you know you've arrived somewhat when you're finally covered in consumer reports and consumer reports now recommends that if you live in a white state, you should go ahead and buy solar, you will save money if you do that but in the blue states either there's not as much sunlight or electricity is so cheap there that right now you'd still be better off. So this drives some of the need to go to higher efficiency and here you can see the efficiency of some of the best technologies. I'll start with black because that's silicon and that's about 92% of the market. You can see it is one of the more efficient technologies but there's also hardly been any improvement in efficiency for about 18 or 19 years. There's been improvement in manufacturing to bring costs down but not improvement in efficiency. Gallium arsenide is better but it's about 40 times more expensive and I personally don't think that's gonna change but some people disagree with me. Down here are thin film technologies where you just have a very thin film of semiconductor on something like glass and it brings the costs down and traditionally it's been copper indium gallium selenide and cadmium telluride that were the leaders and then they get the other 8% of the market and then out of nowhere, perovskites arrived about four or five years ago. I was fortunate to be directing the center that made that discovery and so I knew about it right at the beginning and what's awesome about perovskites is we can print them. You could imagine taking an old film factory or an old newspaper factory tweaking the tools a little bit and printing solar cells on plastic and I've been working on that for about 16 years but our efficiencies were always way down here and now we have caught up and we can print 22% efficient solar cells. But I believe very high efficiency is where things need to go. We've done some calculations as have other people and we find that when you can raise the efficiency by one point you improve the value of the module by three cent per watt. So in other words, if we can go from 20% efficiency to 25% instead of paying 50 cent per watt you should be willing to go up to 65 cent per watt because of all the extra power. And then you can look in the market and there are products with a range of efficiencies and you can find out what are the selling prices and of course the utility companies know what they're doing and it indeed is exactly three cents per watt is the extra value. And we find that in a place like Palo Alto or Menlo Park it turns out people would, it would make sense to pay like a dollar 50 per watt. That may seem crazy but yeah, if you have a Tesla you want to do anything you can possibly do to avoid paying the 40 cents per kilowatt hour price but your roof is not big enough to meet your electricity needs so you need to get the most efficient product you can get your hands on and it's worth it to pay well over a dollar and that's why I believe there's strong demand for a high efficiency product. Let me say a little bit about semiconductors for those in the audience who don't know much about them. They're interesting materials where there are a whole bunch of energy levels that can, electrons can go into and this is called the valence band and it's filled with electrons. Then there's a gap where there are no energy states and above it there's a conduction band with mostly empty levels. When light is illuminated on these materials you can excite an electron up into the conduction band. The lack of the negative charge creates what we call a hole with positive charge and until that electron falls back down the energy is stored and if we can somehow coax the electron to go to one side and the hole to go to another then we're able to extract that power and generate electricity. And when we do this we wanna get the highest current and voltage we can and to get a high voltage we want the band gap to be large because if we excite the electron high up into the band when it'll fall to the bottom of the band and we'll lose all that energy and the voltage is somewhat related to that energy gap and the larger it is the higher the voltage we can get. So that may imply that we want a large band gap. The problem with that is we can only absorb photons if they have a greater energy than the gap. So if we pick a large gap then a lot of the photons in the solar spectrum will go right through the solar cell unobserved. So when you look at the trade off you find the efficiency versus band gap looks like that and the ideal band gap is 1.4 electron volts and you can only get about 33% efficiency. But that's with one semiconductor. The way to do better is to have two semiconductors and use one of them to harvest the high energy photons that'll have a higher band gap and that cell can generate a higher voltage than silicon can do and then you can have also a lower gap cell to capture another part of the spectrum. And you can do even better with three or four or five however as you add layers the cost will go up and so I personally think two layers is ideal. And so I think that we can go from a practical limit of around 25 to something up in the range of 30 to 35 if we do this. And this definitely works. The world record solar cell has 46% efficiency. It has four solar cells stacked on top of each other. I personally think it is the most advanced semiconductor device that's ever been made. It costs well over $40,000 per meter squared compared to about $80 per meter squared for silicon. So there's certainly no way you can cover rooftops with this although they are great for powering satellites and space stations. But I think it just demonstrates the concept works but we cannot be using single crystals that are grown very slowly with a technique called molecular beam epitaxy. Instead we need an inexpensive material that works well even when it's defective. And that's where the perovskites I think come in. And the perovskites have the right band gap and it turns out for the bottom solar cell in a tandem silicon which is already the market leader has the right band gap. And so we can just put perovskites right on top of silicon. Here's what the perovskites are. This is the crystal structure. And the compound that generated a lot of excitement is methyl ammonium lead iodide. But it's all tunable. And here if we switch from iodide gradually to bromide you can see the color changing. That's because the band gap is increasing from 1.6 electron volts to 2.3. Over here all visible photons are absorbed so it looks black and over here the red photons are not being absorbed and that makes it look red. 1.8 EV is perfect for the tandem and so you see that right in there is the right compound for that. And you can also replace methyl ammonium with something called formamidinium and that will drop the band gap. Only have 20 minutes so I'll just skip right to it. We got our certified world record about two weeks ago. When you think you have a world record you send it to the National Renewable Energy Lab and they test it and they confirmed. And we were at 23.6% efficiency now and we needed to do one more thing to beat Silicon's record and I got the text on Saturday that that's done and 25.6 should fall at some point is in the next few weeks I think which is gonna be really exciting. If you join my group there's two ways to get one of those 26.2 stickers. The easy way is to run a marathon. The hard way is to make a solar cell of 26.2 and Kevin Bush is definitely pursuing the hard way and he's possessed and he's hunting it down and I hope we'll get to 30 in the next couple of years and I think we will. Another breakthrough and I think this is gonna be accepted in science this week we figured out how to do a low band gap perovskite and we sprinkled in a little bit of tin to replace some of the lead. We dropped the band gap down to 1.2 electron volts and here's a side structure of the first all perovskite tandem and we're only at 17% now. I can't even believe I'm saying only 17%. It took us 15 years. We started at 0.1% back in 2000 and now 17 is mundane. But we're at 17 and we have a list of problems that we've identified and we will fix these. The layer is just not thick enough. It's not absorbing the light. That's an easy one to solve and we don't quite have the right band gap here. You have to get the band gaps just right because you need the same current in both solar cells. If one puts out less current than the other it pulls the stack down. So we see some indicators that are a little too complicated to get into right now that it's a very good material with huge potential and so we think we're gonna clear 25% soon in an all perovskite structure meaning we can print the whole thing on plastic and have it be flexible. A year ago these devices were very unstable. They were only lasting minutes. We couldn't even ship them to NREL for validation because they were dead by the time they got there. This was troubling me greatly and then we have had tremendous progress. We did three things. We replaced the methyl ammonium which was leaving the film with a combination of cesium and formidinium and then it turned out we were having corrosion with our metal electrodes. We replaced that with a metal oxide and then we took advantage of being in Silicon Valley and we just went to a nearby solar cell company and they showed us how to do a proper package with a rubber edge seal and with those three things we're now able to pass the industry standard tests. There's a building full of torture chambers at a company called D2 Solar. Things like you put it in an oven at 85 degrees Celsius, 85% humidity and leave it there for six weeks and usually if you pass that your cells will last for 25 years even in a place like Miami and we passed it on the first try. Another one, we have an oven on top of a freezer with an elevator that keeps taking the solar cells up and down and so they shrink and expand, shrink and expand and we made it through the 200 cycles which is the tests that you have to pass to be able to sell a product. So we're really excited about that. A year ago I would have said the probability of passing these tests by now was less than 1% but somehow we managed to do it. And so I think the outlook is great. 25% is inevitable for the tandems. I think 30% is gonna be achievable and I certainly don't wanna say we're done on stability. We have a ways to go and we haven't done field testing but stability is looking really good and one of the next challenges will be can we actually print it on huge area? Right now these are only one square centimeter so we need to show that we can do it at big scale and get the uniformity that is needed but things are looking great. And if you wanna learn more about solar cells you have a couple of good options. If you want maybe sort of the small dose then you could take Bruce Clemens course and a third of it is solar cells, a third is fuel cells and a third is batteries or if you wanna go all the way you can take my quarter long class on solar cells and Fritz Prins does a whole quarter on fuel cells and Will Chew does a whole quarter on batteries and then you'll really get into the details. Well, I thank you all for listening and I'll try my best to answer your questions. You know something about solar cells I can tell. So UV, yeah we haven't done that yet. We are setting up all the lamps in our lab to do some of that kind of testing and so we'll hopefully find out soon and yeah we're not, we really haven't done any of those tests that you mentioned at this point. Do you know of anybody? I would say I mean my group and the company Oxford PV are far away ahead of everyone else on doing this. No one else is packaging solar cells and doing this sort of thing and I'm really, really glad that my little team is beating their 25 million pound funded 30 person company and I know we have higher efficiency and higher stability so proud of what the students are doing. So if we choose to just make perovskite panels not as tandems we think we know what they're gonna cost. We partnered with Mike Woodhouse at the National Renewable Energy Lab who runs a team that does cost modeling and he thinks that it's gonna be, he's projecting silicon is gonna go down to about 40 cent per watt and he thinks we could get to 34 cent per watt so a little cheaper, not massively cheaper, just a little bit but what's really interesting the tandem will also be about 34 cent per watt so about the same as the perovskite but the efficiency would be up at 25 or 26 percent and not effectively lowers the balance of systems and so that's where the really big win is. It's essentially just holding the cost in place and raising the efficiency and these are just projections at this time there is no perovskite factory anywhere in the world. You know silicon right now at 50 cent per watt you'd be getting 16 percent efficiency at a dollar per watt now you could be up at 21 or 22 and we project all those efficiencies to come up by about two points over time and we expect the costs to drop from like 50 to 40 cent per watt, something like that. What do we do about the lattice mismatch? Well, yeah when you're alloying them it just changes the lattice constant and it mixes. And when we make our tandem solar cells they're not single crystals, there is no, it's not like the three five tandems that you might be familiar with where it's important for one crystal to grow off the other, there is no epitaxial growth for us. Well, you know as academics I personally I have my idea of what I think is the most promising path but I'm also open-minded that and I'm aware that things are unpredictable and here at the university I like to push multiple paths in parallel of course and one of the reasons you do that is for every student to have their own unique project that they can push forward. I mean the all-Perovskite, we didn't have that two months ago, that is really, really new. So originally the plan was to build on top of silicon so that we're upgrading an existing product which is an easier market entry but you know if we make a 25% perovskite tandem things have to be re-evaluated and a problem that I didn't touch on the solar industry, I'll make an aside comment. A lot of people tell me that they like what I'm doing but a lot of people tell me that I'm completely wasting my time because the problems are already solved and a lot of people tell me that I'm wasting my time because it's obvious that solar will never make it which I find very interesting that it's obvious to them on both ends there and you know in the time I had I showed a lot of positive things but what I didn't tell you is that most solar companies are really struggling. They have a lot of debt and first solar is the only one that has positive cash on hand and that's because the tools are so expensive and they're not really in a position to grow right now because the profits if they have them are not sufficient to build the next generation of factory and you don't, that gets lost when you just look at that dollar per watt figure. We do, our factories could potentially be a lot cheaper and so that is why some would be very excited by an all perovskite tandem or even just a single junction perovskite. Well, I see the zero minute sign so I guess I'm done and hope to get to know some of you during your time here at Stanford.