US-MAP Webinar — Perovskites at the Edge of Tomorrow

Show video

>>Joel Jean: All right. Great. So yeah.  Thanks Harrison. Thanks Laura and Joe for  inviting me to speak today. Again I’m one  of the founders and CEO of Swift Solar which  is working on bringing perovskite tandem solar  cells to market. And I imagine that today  we have folks from all kinds of backgrounds  in the audience from students to perovskite  researchers to investors. So please bear with  me if I’m telling you things that you already  know. I’m sure I will. So anyway, our mission  at Swift is to build a world where all  energy is clean energy. And this has been  true since the very first meeting between the 

cofounders. We talked a lot about hopes and  dreams. And it was pretty clear that what we  all wanted was first to help fight climate  change and second to build clean energy  projects that serve the people who need  energy most. So that’s why we started Swift.  That’s how we built the company and the team  and that’s where I want to start today. As  a society we need to deploy lots of solar PV   because we need to do something about climate change. And we also think of climate change as  some far off problem for some island nation  sinking under the sea or for our grandchildren  to deal with. But it’s an existential risk 

today So almost every major problem we face  in the world today is either caused by or  amplified by climate change. Everything from  heat waves to power outages to pandemics  and refugee crises. They’re all  linked to climate in some way. So climate change got real personal for a lot  of us in 2020. This is a photo I took last  September from my front door. And the orange  sky, the raining ask are all caused by  these massive wildfires that we saw on the  entire west coast and even in Colorado. And  these wildfires are very closely linked to  about 150 million trees that have died in  California since 2013. That’s probably  because of drought which weakens the trees and 

makes them more vulnerable to things like  bark beetles which tend to hatch earlier and  reproduce more frequently in these warmer  winters that are caused by climate change.  We also see an increase in lightning  strikes which are a major cause of dangerous  wildfires. And we expect that with the  higher temperatures we’ll get even more  thunderstorms and lightning. Something like  12 percent per degree Celsius of warming in 

the US. And just overall we’ve seen that the  frequency of days in California, fall days in  California with extreme fire  weather has actually increased   by 2X since the 1980s and it’s only going to keep getting worse. This is also true for COVID actually. So  infectious diseases by definition involve  interactions between species, right, a host  and a pathogen, sometimes risk factors in  between. So infectious diseases rise when  humans come into closer contact with animals 

which happens when habitats are destroyed  by droughts or wildfires or flooding or  anything else from climate change. We also  see diseases rise when vectors like mosquitos  and ticks spread. And that includes when high  temperatures increase their metabolism or  when increased flooding gives  them more breeding sites.   And diseases also rise directly people are more likely to get sick which   is what we see with more extreme temperature swings and warmer winters which actually   reduces herd immunity in subsequent years. And the third thing is that these impacts aren’t  

actually isolated. So extreme weather can also increase infection rates and so on. So with all that said, what are we going  to actually do about climate? And the easy  answer is to stop digging, right? Stop drilling,  stop fracking, stock burning fossil fuels  and start using other energy sources that  don’t emit carbon. You all know this already.  And it turns out that solar PV is a pretty good  way to get zero carbon electricity. So here  we’re showing the life cycle carbon intensity  in terms of grams of CO2 equivalence per  kilowatt hour for all the different electricity  generation technologies we have. And 

obviously coal and gas are not so good. Gas  is generally better. But if you account for  methane leakage it’s probably not that much  better. And certainly with a whole array of  possible low carbon technologies including  PV which does pretty well. It’s about life  cycle right now about 50 grams of CO2 per  kilowatt hour. Thin film is a bit lower so 

that’s just silicon thin films are  maybe more like 20 to 30. These will   only go lower as we decarbonize the grid. Sunlight isn’t very energy dense so you need  lots of land. But it’s actually less than most  people think. So a lot of times when we  think of covering whole states with solar to  power the whole US but it is in fact – this  nice figure from my friend Patrick shows that  we only need this little  square. It’s pretty small,  

right? So this little square to satisfy all of the US electricity demand with PV. And that’s   with average efficiency. So if you actually go ahead and look in Arizona you need less.   So this is a good amount of land but it’s not that much especially if you’re looking at   comparing this to the amount of air we use for coal mining which is right about the same or   even combining missile testing ranges and golf courses you kind of get the same   amount of area as well. You only need about 50 percent of the area we use for corn ethanol.   So all these suggest that there are a lot of uses  of land that we could easily replace with  solar. And that’s doing pretty well on the 

climate side. Obviously we’re not looking  to go 100 percent solar. That’s not optimal  choice for the grid and the  land isn’t going to be limiting. And on top of that everyone has access to  sunlight globally. So obviously, right? So here  we’re showing the solar density instances,  megawatts per square kilowatt. This is just a  measure of density on the same scale as on  the x axis for uranium and oil. And these are  at large scales but on the same scale. So you  can see that solar varies by about a factor of 

four from the sunniest country which is _____  and the cloudiest which is Norway. Right?  So if you look at other energy resources like  Uranium for example you have a thousand x  variation between countries or  oil where you have a million x.   Those are way less equitably distributed. And probably most importantly solar PV  has already been used at large scale. It's  affordable compared to fossil fuels. We've  seen that PV is actually on track to surpass 

wind in the global installed capacity in 2021  with over 600 gigawatts already installed.  And about ten years ago I mean terawatt  scale solar sounded super far away. But now  we're almost there. So PV has gone from  super expensive to super cheap in the last 40  years dropping from about $8.00 a kilowatt  hour in 1975 equivalent which is something  like a utility bill of $7,00.00 a month for  a typical household in the US to less than 10  cents a kilowatt hour in 2020. Where it's now  the most affordable source of electricity for 

a lot of the world. So LCOE obviously isn’t the whole story  because sunlight is variable and intermittent.  But the cost of storing that energy, that  excess energy from when the sun is out with  lithium ion batteries is also  coming down at an insane rate.   So because of this huge progress in solar PV in the last ten years   that a lot of people will now say solar is solved. Right? Solar has won. We should inspect our   work on agriculture or cement or steel or electric aviation or fusion or something   else. And obviously everyone wants to work on and invest in the more unsolved problems. So that's the dominant narrative. Solar is  solved. And honestly I think it's true for the 

most part at least on the surface. This wasn't  obviously true ten years ago or even five  years ago. But now there is good evidence  that we can likely eliminate about 90 percent  or more of power sector emissions without a  whole lot of change in the average cost of  electricity or using just existing technologies,  things like PV, wind, hydro, lithium ion  and new transmission. And there's some nice  work by _____ and others that shows that 

you can scale up renewables and electrify  everything using EVs and heat pumps and  batteries. And that alone can actually  eliminate something like 70 to 80 percent of US  emissions by 2035. So that's  with no new technologies. So today I actually believe that silicon is  probably good enough. We don't need to wait  for perovskite or other new technologies, PV  technologies to solve most of that climate  problem and that's actually really good for  the world. But personally I also think it would  be a huge mistake to stop innovating in  PV technology. Just about every single low 

carbon pathway, every integrated assessment  model, every power system model that I've  seen relies heavily on PV to lead the way  in decarbonizing the power sector. So that  means that even a small change in LCOE can  make a huge difference. Right? The 10 to  30 percent savings you might get from making  PV more efficient and easier to install  could translate to on the order of $100  billion of value per year in the US alone. But 

today we're systematically under  investing in new solar technology. So here's some data on most of the high  profile US solar investments over the last 20  years or so. And a bunch of companies raised  hundreds of millions of dollars from 2006  to 2012 to build the next big thing. Right?  And if you haven't seen a bubble freeze this is  what it looks like. After Solyndra blew up in  2011 after over a billion dollars raised US  investments in solar PV basically froze for  the rest of the decade. So I've seen this first 

hand with doing fundraising where I have  had VCs say that their LPs won't let them  invest in solar after Solyndra. So this is  clearly a barrier. This kind of rise and fall has  been a barrier to continuing  to fund new solar technologies. So in the last decade solar startups in the  US weren't really getting funded. But in the  background obviously technology was still  improving steadily. So perovskites this new 

technology has now surpassed every other thin  film in small area cell efficiency. But a lot  of - there's been very little private funding  that's gone to the US PV companies during the  same period. So over the last three years it's  only been something like $20 million in total  that's gone into US startups building new  and emerging PV technologies. And just for  comparison in this same time period we've  seen over $400 million invested in nuclear  fusion. The same amount invested  in waste disposal startups.   So anyway my goal here is not to - obviously not to downplay   the importance of zero carbon baseload power or sustainable waste treatment. But if we are going  

to rely on solar to solve 20 to 50 percent of the climate problem and if we're spending   trillions of dollars deploying solar worldwide in the next 30 years and   indefinitely after that it seems worth it to keep making it better. So silicon is the only PV   technology we ever have it's not the end of the world for climate. But we are missing out on a   huge opportunity to build a cleaner and more equitable world. So as we look forward to a potential future for  perovskite I think it's useful to look at why  previous solar startups struggled and what we  can learn from them. I think there are a few  different factors worth considering. And some  of these include just bad timing, to enter 

the market when China was scaling up silicon  PV massively. The choice to compete with  silicon head on on price which forced them to  scale the factories too early. So you spend  hundreds of millions of dollars on factories  that are running not completely proven  processes which ultimately led to running out  of money before finding product market fit  and ultimately selling the company to a questions  acquirers. And a lot of these things are  difficult to control. Right? A lot of them  were issues of circumstance or macro-economic 

realities. But anyway, it's not a pretty picture. And we have to look at these failure learning  opportunities and try to chart a different  path. So to kind of walk through these,  some of the learnings you might get. I mean  fundamentally the world in 2020 is different  from the world of 2010 for a lot of reasons. 

We importantly now have some real public  awareness and momentum on climate change.  We know where China stands. Governments  do seem to be more aware of the  strategic importance of energy and domestic  manufacturing. And with things like stacks 

and other capital sources for better or for  worse we're starting to see new pathways to  both capital and assets for clean tech which  was nonexistent in the first clean tech wave. We also know not to underestimate China in  silicon where the targets for cost floors and  efficiency ceilings are always moving. We  need to outperform rather than undercut  silicon and also find customers who will  pay for good products and not just cheap  products. And of course another option is to  couple perovskites with silicon in a tandem  cell. Another important learning is to  develop technology and manufacturing processes  and find that product market fit before you  go and build and scale up your factories. And 

maybe the ultimate learning, a very important  learning at least is that product market fit  and unit economics matter. So new PV  technologies need to have products that have  good unit economics. They need to have a  financing plan that matches their market  strategy which could include mission aligned  private investors and government support.  And I think it's really important to recognize  the value of government support here which  really should be sustained even in the face  of high profile failures. One of the shining 

stars of clean tech 1.0 was Tesla which some  people may know funded its first factory  with DOE vehicle manufacturing loans and  actually have been supported heavily by EV  subsidies and emissions credits. So a lot of possible learnings we can take  from the last generation. So to kind of look  forward I just want to start to introduce  some of the new technologies including  perovskites and lay out the landscape a bit.  So there's many ways you can try to classify  and organize the PV technologies. One way  is to use material complexity as sort of a  guiding metric. And this is very hand wavy  but you can kind of look at a spectrum from 

things like _____ silicon which are wafer  based, have relatively simple materials, single  atoms in a structure in a basis. But then  you have very complex processing, very high  temperatures, requires crystallization,  doping, things like that. On the other side of the spectrum you  have things like quantum dots and organic  molecules which are very complex if you look  at the building blocks of these materials.  But they are processed in very simple ways  at low temperatures. And these materials on  the right side, these emerging materials  including perovskites tend to be more abundant. 

They tend to be more tolerant to impure  precursors and disorder. So you can actually  manufacture them more cheaply in many cases. And those lower process temperatures,  lower manufacturing temperatures are a big  advantage. So perovskites and other emerging  technologies can be manufactured at below 

200 Celsius, 200 degrees Celsius compared  to 500 plus for silicon and commercial thin  films like CdTe and CIGS. And these lower  temperatures reduce your energy use. They  make it easier to design equipment. For  example compared to CIGS and CdTe and they  don't melt a cheap and flexible  plastic substrates which is helpful.   So in that direction if you're looking to make a very high   powered weight or very lightweight solar panel you really need to go to a very tin substrates. And  

these are important for mobile or portable solar applications. And this is showing the specific power in  watts per gram against the thickness of  different substrates. So these are on a _____  scale. So obviously these efficiencies are a  bit dated but changes don't really matter to  _____ scale. In any case at the thick substrates 

you obviously have the substrate is dominating  the weight of the full module or the full  cell. And in contrast as you go to the left to  a regime where you have very thin substrates  you get to a point in the limit where active  layers dominate the weight. And in reality  you're going to somewhere in between these.  But it's only when you get to this left side  where the active layers start to dominate  that having these lightweight thin flexible  active layers really matters. So perovskites  start to shine when you go to thinner  substrates alongside other  emerging PV technologies.

So at MIT we did a silly thing and we put a  solar cell on a soap bubble. This is obviously  pretty useless. But we did this to show sort  of what you could do in the limit of having a  very thin substrate, very lightweight solar  cell. And we were able to achieve over 6,000 

watts per kilogram. So that's like a whole  rooftop installation in a kilogram. Right? So  quite a lot of power in a  small lightweight package.   This is actually an organic solar cell, not a perovskite. One more thing is that because these emerging  technologies use abundant materials that  are produced in high volumes they can often  reach terawatt scale deployment without  requiring a lot of growth in production of  the raw element. So you compare silicon which  needs about six years of silicon production  to satisfy all world electricity demand if you  were to use 100 percent silicon PV. telirium  you need something like 1400 years. And 

this has been kind of a dead horse that's  been beaten a lot for CdTe. And I do want to  emphasize that you don't need to satisfy all  electricity demand to be a useful technology  so don't want to downplay this. But certainly  when you look at the emerging thin films on  the right side you need something like less than  three years for all of these technologies to  generate all the materials, all the raw  materials you need to produce terawatts of  electricity, something like 25 terawatts with  all this single technology. So that's actually  really important from a macro perspective.  It lets you - if you can avoid having to 

actually increase the production volume of  these materials. Then you're not limited by  mining. You're not limited by biproduction  dynamics of how these raw limits are  produced. And all of those things make you  a less risky technology in the end when  you're looking at large scale. So the point of all of this is that emerging  solar technologies do have some fundamental  advantages compared to silicon in terms of  weight, flexibility, potentially efficiency,  abundance, cost, manufacturing. A lot of  potential advantages. And anyway we're all here  because we want to see these advantages realized  in the market and perovskite is the most  promising. So let's look at a few ways that  perovskites might actually be useful in real 

applications. So natural use case for perovskite  in orbit in space where the SpaceX and  OneWeb and others are looking to deploy  thousands of satellites to provide internet  connectivity around the world. So this could  be tens of megawatts each year of solar and  market size of hundreds of millions for the  solar alone. And perovskites here have a 

natural advantage because they can  be lighter weight, more efficient   and radiation hard it turns out than silicon cells. Another application is powering unmanned  aerial vehicles, UAVs and air ships for  months at a time using only sunlight. So  you can do a lot of things with UAVs but one  exciting application to me is you can bring  internet and communication services also to  remote areas or improve them. So if you have  a group of people spread out in a rural area  without access to a cell tower you can  actually introduce the UAV or network of UAV  flying ahead in the stratosphere and that  can act like a flying cell tower and create a  network that acts as a go between  to link people to base stations. So you can think of these UAVs as a  complimentary form of satellite. Easier to launch 

and actually recoverable. And they're fully  dependent on efficient lightweight solar. So  leading aerospace companies, a lot of them  are actually working at building these  platforms. So it's still early days for the  applications that I just describes but there's  actually a lot of weight behind them. And  we can help unlock this market by making  lightweight and efficient solar cells in  a more affordable way than conventional   space cells.

One application that's actually really interesting  to me is looking at solar water pumping  and other productive loads in India. This  is obviously coming from, coming out of the  MIT Gridex Solar Program which was focused  on these kinds of applications. But if you  look at India's electricity about 20 percent of  it is used for agriculture today. And a lot of  that is for irrigation. And irrigation is  super important. But a lot of the small farmers  actually don't have consistent access to  grid power. So solar water pumps have become  an interesting way to bridge that gap.  

So the Indian government has actually been able to deploy something like 1.75 solar pumps in the next   few years which is almost a gigawatt market and potentially higher. And the   importance here is that if you want to be able to share panels a lot of the farmers actually   share the economics and avoid theft, you need portable lightweight probably rollable panels   that can be moved around and set up by a single person.

And there's a lot of companies working on  these kind of solutions already in India. There  are many of them based on crystalline silicon  today. Claro Energy, Atom, for example.  These are photos taken by our Grid Edge  teammates who basically saw these use cases. 

And you can see that they're on wheels which  is really interesting. They're often put on  carts and weight naturally is a challenge  here. These are human powered solar, mobile  solar. So you can actually look at powering  all kinds of other productive loads as well,  right? Things like sugar cane or fruit juicing  machines for street vendors. On the right we  have electric rickshaws for  short distance transport,   wheat threshing, weaving, all kinds of agricultural productive loads as well. And I think learning from ______ over at  MIT in particular and the Good Edge team in  general I think the e-rickshaw market on the  right is actually really interesting. It's a very 

large market and these things are often used  as a last mile connection for a few tens of  kilometers per day. and you can actually retrofit  them to do a lot of things. And they have  a hard time getting access to power since they're  often just standing in the road during the  day and they're off grid. So lightweight  solar can make a lot of sense there too. But going back to the US or globally we  can also help open up a new class of self-  charging EVs, solar powered EVs, cars and  trucks and buses that charge themselves with  sunlight. So a lot of people think that  sunlight is not energy dense enough to power  vehicles. But if you actually go through the  calculations you find that if you have solar  that's efficient enough you can actually  drive something like 5,000 miles a year on  sunlight or almost half your annual miles. So  with a relatively small upfront investment,  right, a couple thousand dollars, maybe a  thousand dollars you can actually get a lot of  convenience of plugging in your car a lot  less often, the piece of mind of having that  extra source of power. Even  cost in CO2 savings for not  

charging with grid power so I think that's pretty cool. If only because perovskite technology is  flexible we can start to make solar integrated  roofing products that roofers can install at  the same way and same time that they install  your new roof which could actually drive down  the cost of residential solar dramatically  largely by reducing soft costs and help  make it more affordable for almost every  household. So obviously this has been tried  before a few times with CIGS products and  silicon products. But to my knowledge it's  never been done with tech that actually can be  more efficient and more affordable than  silicon. So perovskite has really opened up I  think a new opportunity in  this range, in this domain.

And especially for this market, sorry, for this  audience I think it's important to recognize  that these are all viable markets for  perovskites made in the USA. Right? So these, the  economics work out such that you can  manufacture in the USA and serve all these  different markets. There's a lot of strategic  importance for the country I think in being  able to open up these applications  and deploy perovskites. Anyway, I know some of you wanted to hear,  you probably showed up here to hear what  we're up to at Swift. so let me quickly  introduce the team and the company. So these are 

the founders of Swift. We have six cofounders  coming from five different countries. And  we've worked together on this technology for  a lot of the past decade. So several of us  have been, had a lot of intertwining paths.  Thomas and I were actually classmates at  Stanford during undergrad and through PhDs  and post docs at MIT, Oxford and Stanford.  And ultimately we came together  at NREL in 2018 to launch Swift. And I'll just say that my cofounders are all  world leaders in the field of perovskites. 

They're pioneers of perovskite tandems.  And on the team we have experts across many  aspects of solar technology and techno  economics. And we've been lucky to add team  members and advisors who have really helped  push the industry forward over the past 20  years or so. So we started with 6 people. We  have 18 including a couple interns and part 

time folks. One thing I like to point out  is that we have people from ten different  countries out of 18 people which maybe   just goes to show how international the community of perovskite and solar research is   and hopefully how globally we can deploy solar as well. So quick rundown of the numbers. We've  raised about $16 million in equity financing. 

Only about $1.4 million from traditional VCs  for a lot of the reasons that I named earlier.  We again have 18 team members and 9 PhDs for  better or for worse, a lot of PhDs sitting  around. And we have four federal grants.  And we've been able to use our funding to  actually build out a facility with 10,000 square  feet in Can Carlos, California. It’s about  half lab and half office. And it's completely  tailored for perovskite elements. So the  design of our lab pulls from our team's  experience at a lot of the leading labs, NREL,  Stanford, Oxford and elsewhere. And we  have equipment for making cells and many 

modules for testing cell performance  and testing product reliability. And the technology that we're focusing on that  we believe is the future for perovskites is  all perovskite tandems. So again the core  advantages of perovskites are that they can be  formed from these blended and low cost  precursors at low temperatures but they still  have this very, very high quality. right? To  create high quality semiconductors at these 

low temperatures which means you can make a  very efficient solar cell with a thin film.  And a particularly interesting property is  that you can actually tune the colors by  tweaking them through a composition. So we stack two perovskites up on top of each  other. So one absorbs the UV in more blue  light, transmits the red and the infrared  which gets absorbed in the second cell. And  because you're optimizing each cell for a  narrower part oft he spectrum you can make a  tandem that's much more efficient at  converting sunlight into electricity   than a typical cell.

So obviously the leading technology today is  silicon. It's about 95 percent of the market.  It's been in the lab since the 1950s and it's  creeped up in efficiency, right? Still going up  at the cell level up to a theoretical limit  of about 29.5 percent. In perovskite single  junctions we know came around around 2010.  And our team members have been at the  forefront of this really helping advance  the technology to over 25 percent today. But 

these single junction cells are always going  to hit a ceiling around 30 percent as well. So  what we find exciting about tandems but all  perovskite and perovskite silicon tandems is  that they've improved just as quickly and they  can break through that ceiling. So they can  ultimately reach efficiencies as high as   around 32 or so percent in the near term and well over 40 percent eventually in theory. So this is the exact same strategy  that's been used successfully with other  semiconductors, specifically with three to  five materials like _____. And they've reached  efficiencies as high as 39  percent with fixed junctions   but they also cost like $10,000.00 per square meter so 10x more than silicon.   And what the perovskites offer is the same kind of performance, maybe almost as good of   performance at a cost that's actually lower than silicon and that's a game changer.

So if you want to go ahead and make an all  perovskite tandem you have to think about  the entire stack and the entire module  structure. And how do you design this? Right? So  there's a lot of pieces. On the low gap cell  you need to make a low gap perovskite which  usually introduced tin to the material which  tends to produce stability challenges. So you  need to be able to make a stable and efficient  low gap perovskite and you also need to  make a stable and efficient wide gap  perovskite which introduces bromine which can  cause phase segregation and other stability  challenges as well. You need a rear electrode  that actually physically blocks ions from  reaching the metal contact or simply just doesn't  react with ions in the perovskite and other  layers. And these problems usually show up 

over time over the longer times.  So we need to be aware of these. Maybe the most important layer in a tandem  is the recombination layer which goes  between the lower light gap perovskites. You  have to electrically connect the sub cells  with this layer by efficiently recombining  holes from one cell and electrons from the  other. We've had a couple of design  considerations. You do need to make sure that this  recombination layer isn't too conductive  laterally so you can avoid coupling the shunts.  You also need to block ion diffusion between  the sub cells and also damage to the bottom  cell during processing of the top cell. So  there's a lot of competing requirements, 

optimization variables here for the recombination  layer and that's why we spent a lot of  time working on making this a very  high performing recombination layer. And finally the optics are really important  here. So if you change the short circuit current  a small amount in a tandem it actually has  a much bigger impact for tandems than for  single junctions. So we need to optimize the  layer thicknesses in optics to reduce optical  losses and match the sub cell currents. That's  obviously true for two terminal primarily  for the current matching at least. So obviously  this whole stack needs to be made using 

low cost materials and manufacture that at  high throughput. So here's an example of a  real all perovskite tandem stack worked on  by _____ at NREL and two of my cofounders  Charles and Thomas and many of other folks,  many of our collaborators at NREL. This  was made into a very high  efficiency flexible cell. And really what these tandem perovskite  tandems open up is the possibility of making  really nice looking solar products that are  as lightweight and flexible as wrapping paper  with better performance than state of the art.  So it's this combination again of this high 

efficiency, of the lightweight, flexibility  and affordability that really opens up the new  opportunities for solar that  didn't really exist before. So we've been heading in that direction,  designing very lightweight PV products based on  these perovskites. These pictures are for  single junctions but certainly you could do the  same thing for tandems. And a key piece of  this is making a lightweight package. So you 

need to again because of the substrate and  the packaging dominates the weight, you need  to have very lightweight packages with good  barrier properties to make these very high  less over kilogram products. You can see  they're obviously very flexible as well.   So that's great. If we want to put perovskites a bit more into  context we can compare them with other PV  technologies. So starting with silicon, right,  efficient, very long lived, very low cost but  not lightweight at all. CdTe  is actually very similar.  

CIGS is efficient, could be lightweight, long lived but generally has   had trouble getting to very low costs. Organic solar cells efficiency is a downfall there   but they can be very lightweight and have other properties like being transparent or colorful   and stability or long lifetimes are still under development. III-V as I mentioned very efficient,   very lightweight, generally quite stable as well but cost has been the Achilles heel there. So we look at perovskites single junction  have pretty solid efficiency and can be very  lightweight and low cost as well. Stability  is a challenge. I think tandems are what we 

find exciting in that they can kind of check  all the boxes doubly in this case. So you can  reach very high efficiencies, over 25 percent  per module, possibly over 30 in the long  run, lightweight so you can get to over two  kilowatts a kilogram if you do the right  packaging and use perovskite tandems, all  perovskite tandems. They can be quite long  lived we believe in the future. Still under  development though. And ultimately they can  help reduce the cost of solar  on a per kilowatt hour basis. But the big question for perovskites is always  stability. Right? Will they last 25 years?  And the short answer is I don't know. I think  perovskites have a good chance of achieving 

the 25 year lifetimes. And one data point is  the experience of silicon, CdTe and CIGS. A  lot of PV technologies that have started out  unstable and ultimately been able to reach 20  plus year lifetimes. So their history seems  to suggest that once technology moves from  commercial production into deployment  stability problems usually will get solved over  time. Maybe that's because of more focused  engineering or optimization efforts or more  field experience. Obviously this is true only  if stability problems aren't caused by some 

fundamental unavoidable degradation mechanism   which doesn't seem to be the case necessarily for perovskites. So I think   our best bet there is to understand where all the holes are and start filling in systematically. And none of this really means that stability  ever stops being a concern. But for a lot of  other technologies it seems to have become  just one of many variables to consider in  product design and testing. Whereas for  perovskites stability is that one big cold shower  that we take in every commercial discussion.  So the stability issues aren't clearly 

fundamental but they're far from trivial to  solve. So I do want to come at this stability  question from a few different directions  starting with a bit more history. So in the US we've had three major programs  for solar RD&D, research development and  demonstration before Sun Shot. And this really  started in the oil crises years in the '70s. 

But the one I want to highlight here is the  flat plate solar array program. That was the  first major program here in the from '75 to  '85. This was very relevant for stability  development for silicon. So if you look at this  project, this program the goals were really 

to build a reliable and standardized  silicon module design with relatively high  performance for the time, long lifetimes  and reasonable cost. And the way this worked  was that DOE funded industry silicon R&D  and then bought modules. They bought the  modules and then ______ researchers at  JPO and elsewhere tested them and the test  results were used to improve the manufacturing  processes so there's a lot of exchange  between industry academics, government, nonprofits   to really develop this technology and develop manufacturing. And I see a lot of benefits from block buys.  It really encouraged the PV industry to use  the latest technology. It did evaluate modules  using a uniform test or series of tests. And  importantly actually created these close  collaborations between the manufacturers and the  evaluators which led to really fast  learning. So in this ten year period we saw 

module efficiencies, warrantees and costs  improving dramatically. We saw a lot of  foundational technologies and manufacturing  processes developed for silicon. And I  think a lot of industry experts - I wasn't around  then - but certainly a lot of people say that  FSA was foundational to the  terrestrial PV industry. So if you look at where FSA started these  are typical solar modules in 1975 silicon  modules. And they're actually maybe not  that different from the kinds of perovskite  modules that you're seeing today. But over  that ten year period we saw across these five  sort of block buys a quick evolution to  the commercial format that we see today for  silicon. So the cell packaging factory  increased. We saw module efficiency, power and 

costs all improve quite dramatically. And  probably most importantly it helped make  silicon PV reliable. So if you look across  this program it helps reduce module failure  rates from over 20 percent to 0.1  percent in the last block buy.

So there's a few possible take aways for  perovskites. Field testing is obviously critical.  We need to move from passing the standard  qualification tests to understanding how  perovskite specifically fail in the longer  term and then design the qualification tests to  evaluate those failure modes and design  solutions to mitigate them. Obviously  perovskites can benefit from a lot of these  reliable designs that were developed through  FSA and the thin film programs, things like  substrates, interconnects and capsulants. And  maybe other kind of learning is that government  coordination if it's done right can do a  lot to push PV technology forward. So  perovskite companies and researchers need to  work together to advance perovskite reliability.  I think it's going to take government and 

national lab support to  coordinate this kind of program. Maybe a slightly different perspective. This  is some nice data from a really great analysis  by Dirk Jordan and Sarah Kurtz over at NREL.  And this shows that historical, these are  mostly silicon module warrantees have often  been way out ahead, about five or more  years longer than the longest published field  tests at the time. So that's interesting. Yeah.  To be clear I'm not saying anyone should go  and warrantee perovskites to their heart's  content today. That would be really bad for  the industry. But there's some precedence for 

selling PV without fully proving out the  warrantee of the lifetime in the field. As long as  the science is reasonably well understood. So if we look at how perovskite solar cells  might break to try and understand that science  there's a lot of different mechanisms so  intrinsic and extrinsic mechanisms that we need  to understand and solve. So a lot of intrinsic  ones which happen even in inert conditions.  There's structural instabilities in the  perovskite which could lead to decomposition.  There's thermal stresses which can cause the  organic cations and the perovskites to leave  to sublimate and also cause delamination  between any of the layers in the stack. Ion  diffusion can happen leading to these contact  and halide reactions which tend to cause  shunts and defects in the  cell and reduced performance.  

It can also have light induced instabilities anything from halide segregation   to dealloying of the perovskite to defect formation. So these are all possible   under light which is a problem if you have a solar cell. And certainly there's extrinsic mechanisms   like moisture and oxygen ingress which can lead to decomposition as well. So for the intrinsic things you basically  just want to turn your perovskite  compositions carefully. You want to choose  your contact layers carefully and design 

internal barrier layers in those cases to try  and block that ion motion which is unique to  the perovskite. In terms of the extrinsic  mechanisms the probably the simplest design  strategy is just to choose good packaging  solutions. Certainly the better we can do in  terms of intrinsically stable perovskites  the better off we are from an extrinsic   perspective as well. So there's a lot of key stability tests that  need to be done. I think light and heat, high 

temperatures, high in testing lights is  very important, particularly important of  perovskites. But certainly things like the  MP, UV, thermocycling and combined stresses  are all important tests. And it's important  to know that you actually need different tests  for different applications. So space product  has to be radiation hard for example. So at Swift we've been able to pass - just to  give you an example we've been able to pass  damp heat tests with flexible packaging. So  this is just over 1,000 hours the efficiency  change for tandems in wide gaps and  low gaps. And obviously going for these 

flexible applications makes stability challenges  more challenging. You can't just use glass  as a barrier. But we are  making very good progress here. So what's the outlook on stability? I think  personally I think it will take a few more years  to prove out this ten plus year field  performance in reliability. And then you have  preventability as well. So it's not going to be  a fast process either way. So shorter lifetime  applications like these space or UAV, things  like that are already in reach today. But if 

you really want to go after the longer term  applications soon I want to show one possible  strategy that you could consider to  get perovskites into the market sooner. So the context here is that DOE and others  often say that we need 50 year life modules to  hit the Sun Shot targets, to hit our climate  goals. But we've seen that PV technology just  keeps getting better and better consistently  and rapidly. Costs are coming down. 

Efficiencies are going up. And degradation  rates are going down as well. So we're at a  point now where the module hardware and  the module installation are actually only a  small fraction of the total cost of solar and  that's especially true in the US. So if this  balance of system, these non-module costs  dominate then you might think that a long  module lifetime maybe isn't as critical because  you can take these shorter lived modules  and actually replace them one or more times  during the system life given that most of that  infrastructure is already  in place at least in theory.   So you can upgrade these modules periodically with better and better modules. 

And obviously this is particularly important  if you start out with a shorter lifetime. So we did some basic LCOE calculations  with NREL benchmark pricing data for US PV  systems. This is on the left here for a standard,  sorry, standard operating strategy for a PV  system. So you incur a bunch of costs up  front as you do with a PV system. You'll spend  some at the end to decommission. This is for  100 megawatt utility scale system. You can 

see in red here the efficiency of the modules  degrades over time. And this is within  warrantee but it does degrade over time. The  degradation rate here is constant because  you have the same modules percent per  year. The capacity degrades alongside the  modules. The EC capacity is - sorry, the DC  is usually over size compared to inverters.  So that actually stays constant and  the generation falls off slightly.

So if you go ahead and include module  replacements you see if you imagine one  replacement event at year 15 here in the  middle you incur module and inverter costs. You  incur some installation costs at that time  but it's relatively small compared to upfront  costs especially with DOS. So at that 15  year mark you get to jump up in efficiency to  where you assume modules are at that point.  Degradation rate is a lot lower or can be a  lot lower. And in this case you jump up your  capacity as well both in DC terms and as 

you upgrade your inverter or when you replace  your inverter you also upgrade so the AC  capacity jumps as well. And that's when  you get to a step up in your generation. So if we go ahead and compare a few different  options for PV technologies so you can  start out with commercial technology so this  might be a silicon or a thin film, commercial  thin film module. You can compare that with  a lost cost hypothetical module in green  here which has the lower price but  also slightly lower efficiency and high  degradation rate. And then a third case  where you have a high efficiency emerging  technology which starts at a higher  price, can go to lower, has a very high  efficiency going even higher and starts out  with a high degradation rate, in this case a ten  year lifetime to start and improving  over the course of these 30 years. So if you look at how these perform without  module replacement you see that in general  the emerging PV LCOE tends to be higher   by about 15 or 16 percent than the standard commercial PV LCOE for all the system types.  

But if you go ahead and replace the modules every 15 years the LCOE for   these emerging PV, this levelized cost of electricity it actually becomes highly   competitive, right. So its the blue one here compared to the black here. And importantly   as you go longer term this grey one assumes that you don't replace but have the ultimate   performance for these technologies. You can do better with these emerging technologies than   even commercial technologies that are extrapolated out in performance. So the core idea here is to think of modules,  PV modules as an upgradable technology  and not just this one time infrastructure  investment. So by doing that you can make  modules that are short lived start out  competitive actually with long lived modules. So 

again this emerging PV with even a ten year  life to start as long as it's improving quickly  can have an LCOE that's competitive. And  obviously there are a lot of caveats to this.  The module replacement concept isn't always  economically favorable. So you need to do  detailed financial modeling for specific systems.   As a general rule you need an efficiency gain over your installed modules to be   large enough to justify that extra cost of replacement. And you need lifetime to be   improving as well. And we'll probably need module recycling to reduce the environmental  

impacts of doubling the number of modules you're using for example. Backward compatibility is going to be an  issue where you need to be compatible with the  structural or electrical BOS. Otherwise you're  not leveraging that existing infrastructure.  You might need new interconnection  agreements, transmission upgrades to  accommodate the higher generation capacity.  There's the risk that you get a higher cost of 

capital because of using a new technology  which might actually negate the LCOE  benefit. And certainly there’s not a huge  amount of pure economic incentive to do this.  You can see the benefit is still not that great  compared to just continuing to use silicon.  So until you reach that long term end game  it's actually not a huge advantage in terms of  LCOE. So I think government support for  this strategy might be one approach to  derisking and scaling up new  technologies at a reasonable cost.  

But ultimately this concept of module replacement can let   PV technologies like perovskite tandems or perovskites in general which could have very   high performance but are unproven into the market without that 25 year life. So clearly there are a lot of challenges ahead.  I just want to leave you with a different  perspective on perovskites and the future of  solar. So when I'm feeling particularly nerdy  I do like to think of Swift as a nuclear  fusion company. They've also raised a lot of  money a lot more money than solar so maybe  that's a good way to frame it for our  investors too. But we like to call this modular  remote fusion. Right? And this is actually a 

fact. The fun is a free fusion reactor. We  do wireless power transfer across 100 million  miles which is pretty remote and our modular  technology is going to be the lowest cost  selector of solar energy. So in short  we're making the cheapest fusion   reactor in the universe. So I'll actually lean into this a bit. What I  think is actually very cool abut light, starlight  specifically is that it's the energy carrier of  choice in space. Right? That's it. There's no  real other option. So if you're going to  use this I mean PV is really the best way to 

convert energy from this form that the  universe likes, electromagnetic waves to a form  that humans like, electricity. So what's cool  about that is that any technology that we  build today to capture the starlight around  Earth, perovskite or otherwise. if we make it  good enough and efficient enough this technology  that we build today can really benefit  humanity for all time and that's no  matter where in the universe we go. So we've actually done the synchrotron test  with our colleagues at Cambridge and found  that perovskites are in many ways actually  the best technology for Mars and for space. In  a lot of early tests they can survive proton  eradiation something like five times better than  today's leading space technology. They can be  a lot lighter weight and they're an order of  magnitude lower in cost. So that means if  Matt Damon here had perovskite instead of 

solar satellite III-V technology or these  super heavy panels made out of weird reflective  silicon cells, if he had perovskite I  think Matt Damon would have had a way   better time on Mars. So with that, enough of Matt Damon. Let's  wrap up. Swift Solar has a world class  technical team, technology in the market that  we expect will let us scale up sustainably  and profitably. But more importantly I  think for everyone out there working on  perovskites for other emerging technologies  I do think we're on the frontier of an  incredible opportunity to build  technologies that can serve humanity   for a really long time no matter where we end up.  

So yeah. Thank you again to the Swift team and our supporters, to all of our collaborators at   NREL and Cambridge and elsewhere, to the US- MAP team and to all of the _____ members and to   all of you for listening in today. I'm happy to take any questions. >>Laura Schelhas:   Thank you Joel. This is really  interesting. I definitely appreciate  the Matt Damon references at the end. We don’t  have a lot of time for questions. But one  that has come up in a couple of different forms  is related to recyclability and end of life.  And I think that this is a really important  question given that you're recommending a  module replacement strategy. so what  are your thoughts on opportunities for 

recycling perovskite cells?  And do you see a path forward   for kind of what the end of life looks like for these? >>Joel Jean: That’s a great question.  I think it's going to become  particularly important as we deploy a lot of  it obviously. So in the early years it's not a  huge deal. I think in terms of recyclability  we haven't done much work on this. I think  there's a lot of interest in it. So there are  a couple of points I guess. The actual raw  material use in here is relatively small. So  that's one small comfort. But assuming as we  go and deploy a lot of this stuff it is going  to be a concern. So I think there's some early 

work that shows that you can pull out a lot   of the constituent materials, the lead, the halite, a lot of the raw materials that come   out of these modules so you can break it down. So I guess the way we look at it is if you,   if there is that demand for it which we expect there will be eventually we can develop recycling   programs with partners to actually reclaim modules and break them down. So I think in terms of the environmental  impacts of this kind of module replacement, we  did a bit of an analysis here looking at life  cycle impacts on the environment. So with  silicon I mean you certainly have some major  impacts and in all cases actually for metal  depletion because these different PV technologies  all use critical materials, metals. But in 

many other categories including carbon  emissions and kind of toxicity and all kinds of  things from an environmental perspective  adding module replacement actually is not so  bad. For silicon obviously it increase  dramatically but for CdTe and CIGS it has a  relatively low life cycle impact and that's  going to be similar for perovskites. They are  manufactured in a similar way, use very  similar module structure. So I think from that  perspective replacement isn't  going to make a huge difference.   Recycling is going to be important either way.

>>Laura Schelhas: Great. Another question here   that I have interest so I’m going to be biased and ask it. How much outdoor exposure   testing have you conducted so far on your samples or modules? >>Joel Jean: Great question. We have a small  test going on I think in Charles living  room. But otherwise we have not done much  outdoor testing. It’s mostly it’s a very high 

priority for us and I think that’s something  that NREL and US-MAP can help with a lot. >>Laura Schelhas: Great. I like that answer.  And then I think a nice question to kind  of wrap up as we approach the hour. In what  timeframe do you see solar and renewables  completely replacing nonrenewable sources  on the grid? What’s your prediction? >>Joel Jean: What’s my prediction? Wow.   I think it’s going to be somewhere between 2035 and 2050. Not sure if that’s a good answer.

>>Laura Schelhas: It’s a fun estimate. I think  I asked Joe the same question. I think he  was a little bit more optimistic. So we can  decide on the conditions later. But good to  hear. Harrison we’ve just hit the  hour. Is there anything you want to  

mention? Joe I think you’re coming in. >>Harrison Dreves: Well, no. I was just going  to say I would totally defer to Joel since  he’s got the business experience and acumen. >>Joel Jean: I would defer to a lot of  other people who know a lot more than me. So  appreciate that though. >>Laura Schelhas: Going out on a limb.  We’ve got people in the chat saying 

anywhere between 2030 and 2050. We should  take a survey of the community at some  point and see who is right. I  think it’s fun to think about. >>Harrison Dreves: I think we can  make that happen on the webpage maybe. >>Laura Schelhas: Yeah. I think we should do it. >>Harrison Dreves: We could  set up a poll if you guys want.

>>Laura Schelhas: Cool. Let’s  do it. Guys, anything to add   as we approach the end here? >>Harrison Dreves: Just a reminder we will  get the recording posted next week on the  US-MAP website. So to news and events and  then click the link to the webinars page to  find that. And stay tuned for our  March webinar. We’ll also post that   on that page once we get it prepared. Male: Excellent and do check out the  US-MAP webpage. We’ll be also posting about 

for potential companies interest in staffing  needs, the NREL post docs and graduate  students are having events that we’ll feature  there as well. So I just want to thank Joel  again for spending some time with us. I think  it was a really, really great presentation. I  especially like the power __ I must confess. Yeah. >>Joel Jean: Sounds good. Thank you all.

>>Harrison Dreves: All right. We’ll  end it there. Thanks everyone.

2021-02-22

Show video