 Okay, good morning everyone and thank you for this opportunity to speak. So yeah, I'm going to talk here about the past, the present, and the future. I've been with GCEP right from the beginning in 2003 and there's really a lot that I could tell but only about 20 minutes. So the past will be very brief. I'll spend most of my time talking about results from just this year where some exciting things have happened and then a glimpse of the future. So the beginning for GCEP is very personal to me because I was already here and this was my research group and literally the day that I found out that I got one of the grants, I was having group meeting and telling some of these people that I thought they should probably look for a new advisor because my startup funds were gone, my NSF career funds were gone and I was basically going to have to lay off students. I just couldn't pay them and five of these seven were working on solar cells and after I gave that speech I go back to my office and the congratulatory email is there and I had a $900,000 grant, I ran back, they were still in the meeting room and there was a great celebration. I put five of those people on the grant, I was not supposed to do that, didn't ask for permission, over spent the grant by $90,000 but that was okay because other people hadn't done their hiring yet. So my over spending was offset by other people's under spending and anyway, everything worked out and so we got off to a start, we were working on organic solar cells and in the end I think we've written a couple hundred papers on that subject. We were later on able to get $25 million from Saudi Arabia to expand that out with about 15 other research groups and we learned a lot along the way and later on this is the research group and so in addition to saving us there in the beginning and allowing us to start, GSEP along with funds from other sources allowed us to have a large activity and in this photograph, every single one of them work on solar cells, after those original two who were not doing solar cells, everyone in my group did solar for well over a decade and that meant that we could do everything, we could make the materials, we could look at every layer in the solar cell, we could characterize it, the structure, we could look at the electrical, the optical properties, we could build models and then talking to people in industry at these meetings allowed us to learn about cost modeling, we were getting constant feedback on some of our early designs, we thought they'd be cheap and industry was just laughing at that and they're like, no, you're going to raise the costs by a factor of 10 using some of your really sophisticated nanotechnology designs and so we got a lot of feedback and even started getting into studying degradation and making cells durable which is not something you often see in academia and really allowed us to strengthen that whole thing and each generation of students passed on a lot of knowledge to the next. While all that was happening, the target that we needed to reach just got better and better and better and here's the story of the prices of silicon cells dropping and then in the last 10 years the deployment really rising dramatically and most of you probably know that story. I wish I could say I had a part of that, I didn't, maybe some of my students went to work for a few American companies and helped a little bit but mostly this is the story of Chinese silicon and now the prices have dived again this year due to another oversupply situation that the Chinese are expanding their factories again and now the prices is at 42 cents per watt for silicon which is really remarkable and yeah in places like California if you don't have solar on your roof you're just throwing money away picking a more expensive option for getting your power. Some people, it's interesting some people still tell me I'm wasting my time solar is never going to make it others are telling me I'm wasting my time because the problem's already solved. Fortunately some people think that we are at a good place. I think continued research is still needed. I think we need more efficient panels. The value of an extra efficiency point is 3 cent per watt meaning that you would rather pay 70 cent per watt to get 25 percent efficient panels than to buy the 16 percent efficient panels at around 42 cent per watt because more efficient panels you know they're giving you more power so you don't need to install as many of them and right now we're paying more to install the panels than people are paying for the panels themselves. So efficiency is really important and especially in a residential situation where your roof is not large enough to get the power that you'd like with today's cells and in that case there are situations where people it would make sense to pay even up over a dollar 50 per watt to get 25 percent efficient panels if they were available which today they're not you can't buy that. And then the other thing and Antonio Buena-Sisi at MIT makes this point really well it's really not clear that the solar industry is going to grow as needed to meet the climate targets and I'll just boil this down and make it really simple. The companies are not making enough profit to fund the next generation of factories in fact most of them right now are not making profit at all. Now the Chinese are building them anyway and I'm not sure I understand that that gets away from my material science expertise but it's not clear that you know that we're going to see it like a 10x expansion of factories due to a combination of low profits and the factories to make a silicon solar cell are rather expensive so if we could reduce the capital expenditure needed for a factory that would be extremely helpful. Here you see the efficiency of five kinds of materials advancing with time and I think the black line near the top is one of the more important ones that silicon and silicon is about 92 percent of the market right now and you know mostly that that has been flat around 25 percent for you know about 20 years although a world record of 26.3 was obtained about a month ago. Above that is Gallium arsenide and you can get 28 percent with that but it's about 40 times more expensive and I have a hard time seeing how Gallium arsenide becomes the primary photovoltaic material. The blue and purple that's the thin films cadmium telluride and copper indium gallium selenide and they're both at 22 and first solar is doing very well they're the most profitable solar cell company and they use thin films of cadmium telluride but the story I want to tell the rest of the day is the perovskites what just came out of nowhere and we went from a few percent and now they've caught sigs in cadtel they're also at 22 percent and once again GSEP helped me enormously I did not invent the perovskite materials but this as soon as I saw it I knew it was the material that I'd been looking for for over a decade and we wanted to get on it right away and GSEP came through with a large grant for Hema Krunadasa and I to get started on this right away. This is the crystal structure of the material the the original one that that generated the excitement with methyl ammonium lead iodide but I'll show you later there's other things that we can put into that crystal structure to tune the properties what's really exciting here is that you can dissolve this material and then just print it so picture like a newspaper printing press putting out solar cells and I wish I could tell you that you know we're just spray paint the stuff on the roof and it'll practically be free but that's not true there there are a number of layers in the cell it needs electrodes it needs to be properly packaged I don't have a cost model that I'm able to show publicly but the cost structure would be very similar to a thin film cadmium telluride cell and this is published data from Mike Woodhouse at the National Renewable Energy Lab and we don't we're not really going to change the three things on the right we would have similar electrodes we would have glass packaging we would have similar encapsulation and junction box what we can change is the buffer and absorber which is another word for the semiconductors and I think maybe we could cut that in half which which means that and you know right now cadmium tellurides at 40 cent per watt so we could maybe go down to 35 cent per watt but if you look at the top you'll see this is for a 16 percent module and I believe we can do better than that and I believe cadmium telluride will do better than that and and so you with higher efficiency while holding the cost per area constant you see a pathway to costs going down into 25 to 30 cents per watt here's an example of how we can tune the properties of the material on the top we're gradually replacing the iodine with bromine and you those are solar cells you can see metal electrodes on the top and you can see the color of the perovskite not changing on the left the band gap is low and the entire visible spectrum is being absorbed and the material looks black and as we go over to the right some of the red and orange photons are not being absorbed in the lower line the methyl ammonium is replaced by form of adenium and it has a slightly different size and later I'll show how replacing the methyl ammonium greatly improves the stability of these materials the reason I bring up tuning the band gap of the semiconductor is I believe the future of solar is going to be tandem's on the left you see a structure schematic that can give a 46% efficiency that's the world record right now but these are single crystal 3 5 semiconductors and I don't personally think those costs are gonna come down nearly enough there they're over $40,000 per meter squared compared to $90 per meter square for cadmium telluride so I think we have to to go away from the single crystals and and I'm excited to put things like perovskites on silicon silicon has a good band gap for the bottom cell in a tandem and perovskites can be the higher band gap and the idea is just that a higher gap cell is gonna harvest the higher energy photons and deliver a high voltage and then a lower band gap cell is gonna harvest the infrared photons and and still give us some power just with two semiconductors you can do better than then with one and the theoretical limit for a two terminal or two two semiconductor tandem is 43% and I I don't think in a cost effective manner we can do that but I absolutely think we can get 33% efficiency at something like $105 per meter squared and that would be an extremely powerful technology. There are couple ways you can imagine making this and on the left you see what we call a mechanically stacked four terminal tandem here the two solar cells have their own electrodes on the right it's a monolithic two terminal tandem when we go with the two terminal tandem we have to design it such that the two semiconductors produce the same current density if one of them has a lower current density than the other it limits the current of the whole stack and that can certainly be done although when the when when clouds come along it changes the spectrum and it and it's hard not just hard it's impossible to always have that current matched so on the left gives us a little more flexibility and what we really like about the architecture on the left we don't have to change the silicon at all we just take it however any company wants to make it and we put our stuff on top and by the way you could use SIGS as well copper indium gallium selenite also has an ideal band gap we made the first of both of these prototypes but right now EPFL has the world record for the left architecture at about 25.2 and I'll show you soon our world record of 23.6 for the structure on the right a lot of people initially criticized this idea of a mechanically stacked tandem they said you don't want to have four wires coming out you don't want to have two separate inverters I completely agree and that's why we would do something like you see here we would just adjust the areas of the cells as needed and then put the bottom string in series with the top string there's actually a lot of things you can do various ways of current matching or voltage matching but it's only necessary to have two wires coming out and if the current match changes as the spectrum changes you could imagine a little circuit in there that would would make adjustments and fix that for you so that you would always extract the maximal amount of power I really have to fly through how we got that world record we and you know we and and I have to acknowledge there are thousand literally a thousand or more researchers out there in the world who are who are hard at work in the labs as I speak developing all these new compounds and and here we catalog some of them and we we don't yet have our ideal band gap of 1.80 v but we do have the cesium formidinium lead bromide iodide compound that works really well has a has a band gap of 1.6 it's very stable and we we were able to put it on top of a heterojunction silicon cell from Zach Holman at ASU and and I guess the point I'll just make here in the spirit of this session is that this is the culmination of of what we've been doing for 17 years there's a lot of layers in this stack and and you know we know how to deposit them all we know how to select them we know how to get the energy levels right and we have programs that allow us to calculate how thick all these layers should be and we got this working remarkably quickly did not have to make hundreds of of tandems to get this NREL certified 23.6 world record and I will say we're in the last couple months we've developed five improvements we usually isolate the improvement we don't on a daily basis make the entire stack and when we do the the next generation it's going to be about 26 or 27 percent efficient so really excited about that and at that point we truly will have upgraded the world record silicon solar cell I should also point out this is a key number that's a 21 percent efficient silicon cell on the bottom so we've upgraded a 21 percent cell up to 23.6 using a perovskite that was only 15.5 and now we have better perovskites that'll allow us to get 26 or 27 yeah so two weeks ago a science paper came out Giles Eperon is a graduate student or was a graduate student at Oxford with Henry Snafe and he figured out how to get a lower band gap with tin and Henry's a good friend and they knew we were good at making tandems so they sent the material over here for us to make a tandem and we further optimized the material here you just see how the we can adjust the band gap by adjusting the tin to lead ratio and this is an scm of of an all perovskite tandem and this first one was at 17 percent and we we will be able to do way better when we just figure out how to make this new material thicker it has very low voltage loss and we just haven't made it thick enough to absorb all of the light so i i'm very optimistic that that could go up to 25 percent fairly soon we've got 20 percent when we do the four terminal tandem where the 1.2 and the 1.80 materials are or 1.6 have separate contacts and so in a way i like this business opportunity of upgrading silicon i think it has an easy path to adoption but i now also see how we could do it all with perovskites and and we could print this on plastic and and we could have flexible cells in the reasonably near term at 25 percent and and i think 30 percent is ultimately doable degradation is is one of the main concerns here and i have to say i was i was trying to be very cautiously optimistic about a year ago back then ourselves were only lasting a few minutes we have improved this by well over four orders of magnitude in the last year the the key is you got to replace metal electrodes with something like indium tin oxide because the halogens and the perovskites react with most metals you've got to get the methyl ammonium out of there and then you have to package it so here here's a picture of a cell that's been on a hot plate at 100 degrees the the the reddish square that's where we have ito yellow is where we don't have ito when methyl ammonium leaves you're left with lead iodide and and the lead iodide is yellow so you see that the ito holds in the methyl ammonium and and and you get a 10 000 x improvement right there but you you want to have multiple layers of defense and it's also best to just replace the methyl ammonium with cesium and formimidinium uh why do we use both cesium is too small to fit in the lattice formidinium is too large and um on if you mix them then on average they're the right size and it gives you a more robust crystal structure and that um that also that just gives us the the temperature stability that we need and then with the combination of those two and of course that by the way makes it a lot easier for us to put the indium tin oxide down on top and then those combinations allow us to properly package the cell. DuPont was extremely helpful to us and i guess it's in their best interest to teach us how to use their materials to to package cells and and and they did that and uh we didn't even try to invent anything here we just use industry standard materials and and then we went out to pass the industry standard tests the so-called IEC tests one of them is a thousand hours at 85 degrees C 85 percent humidity and two of our three packages we passed on the first run in fact if you look closely at the data the efficiency actually went up um previously we we couldn't anneal ourselves because they they were so thermally unstable and now we're able to anneal them and and even improve the efficiencies uh so we're really excited about that um we um we've passed a temperature cycling test 200 times between 85 degrees C and minus 40 where the cells are constantly expanding and contracting and they did not delaminate we've passed a test under intense ultraviolet light and we're now doing a lot of testing under under one sun or and sometimes more than one sun conditions so outlook is looking just vastly better than it did a year ago uh and so yeah on efficiency um the single junction right now is at 22 percent and a pretty good slope versus time and and I think there's a good chance single junction cells will be at 25 we now have the band gaps we want for both single junction perovskites and uh tandems the tandems um the four terminal is already over 25 percent um and and I think 30 percent in the in the near term um is is certainly doable uh there are now at least five companies commercializing uh uh perovskites um and uh one of them is as uh founded by uh Colin Bailey and I that's Colin there he um he I might not be here today he texted me to say that he's sick uh but um perhaps he'll he'll um be able to come by in the in the afternoon uh so we're gonna try to do the mechanically stacked tandems and uh Oxford PV uh is gonna try to do the monolithic uh two terminal tandems I think it's great that Henry founded his company and I founded mine but we're still able to work together and and uh pop out that science paper uh this month um but also in terms of of of GSEP impact um Richard never likes to take credit for these uh he only likes to take credit for one of these 11 companies uh that that have all come from uh these were all founded by students um out of my research group and and you know you can read off the the um the technologies and and a lot of times the students um did not spin out a company based exactly on their PhD research uh but they learned from this ecosystem and while I was tackling some really hard long term problems um and and I never felt that the time was right to found an organic solar cell company so they went a lot of times with uh with shorter term ideas and and and um uh most of these companies have done very well some of them have already been uh bought like for example D fly was bought by uh Sunpower um and and uh so GSEP really created this environment and I literally I I I was I told you at the beginning I was at a point where I was gonna have to give up on solar um uh and and I may have even left academia all together and and gone and gotten a job at a company so I'm not sure any of that would have happened uh with without GSEP and yeah so finally there's almost too many people uh to uh to thank um uh Kevin Kevin Bush made this world record 23.6 percent cell and Thomas lightens uh took the lead on the perovskite perovskite tandem uh but uh really a lot a lot of people um uh from Stanford and other places as well who have contributed thank you all for your attention thanks very much Mike um but really ready for lunch but maybe if someone has one pressing question uh we can have one question and then we'll we'll go on to lunch Lawrence I don't have a question but in a way in demon I listened uh was fascination with Michael and Thomas presentation and thinking of what I said before it would be so important that CGP can communicate this state of the innovations and the problem you are facing what has we discussed yesterday on the funding of new companies I think the communication of where do we stand in innovation is absolutely central again to have this positive loop so I don't know what you can do like a sort of a reporter on results or telling the stories I don't know but of would you have to go from of course this very sort of substantive presentation to something like government and others just make have a sense of where this pointing to thank you we appreciate your comments okay well I think uh Mike I think we have to go to lunch so you'll be around lunchtime and I'm sure you can ask Mike lots of questions so let's thank Mike one more time and thank you Mike