Roll-to-roll fabricated perovskite solar cells under ambient room conditions | Hacker News Reader

Roll-to-roll fabricated perovskite solar cells under ambient room conditions

nature.com

153 pointsgnabgib107 comments

My two cents having formerly worked in perovskites trying to upscale the process:

Perovskites are exciting (or were exciting) because they have a high theoretical efficiency, are relatively simple to prepare, and the "worst" component in them is lead (an incredibly abundant material). The big problem with them is that they are famously horrifically unstable in ambient conditions.

Roll-to-roll processing means that you can fabricate them in mass scale. Ambient means that they claim to have solved issues like working in glovebox conditions.

Even if the price of solar panels has come down below labor, the fact that they are produced from rare earth minerals goes (in my opinion) underreported.

Consider the relationship between perovskites and multi-junction solar cells similar to the comparison between sodium and lithium ion batteries. Lithium will always have a higher capacity, but sodium is so abundant that for many applications it just doesn't matter anymore.

Do we not have lead free perovskites now?

Tin based perovskites have been studied for almost as long as the lead based ones but they have been less efficient and much less stable. Work continues to increase their efficiency and stability, e.g.:

"Efficient tin-based perovskite solar cells with trans-isomeric fulleropyrrolidine additives" (2024-01-29)

https://www.nature.com/articles/s41566-024-01381-7

Solar panels "produced from rare earth minerals" is "under-reported" because they are not made from rare earth minerals, and further: the minor metals they are dependent upon are byproducts of refining base metals, ie there isn't much additional impact from using them; we already make them.

I'm not really sure how someone who supposedly worked in solar panel research would think rare earth metals are used in solar panel construction.

Solar panels have decades-long lifespans (their rated lifespan is based on when they drop below 80% efficiency, not when they become useless), there's a growing recycling chain to sell complete aged panels to other markets (typically underdeveloped nations where daily equivalent hours of solar are very high and land is plentiful so efficiency doesn't matter), and the panels themselves are highly recyclable for the materials to make new panels.

Ever notice how the people 'concerned' about the environmental impact of mining rare earth minerals, which go into durable goods that are highly recyclable/recoverable, don't seem to have a problem with oil drilling, fracking, coal strip mining, etc - for something that is usable once, maybe twice?

This is true: i.e. they use rare metals not rare earth metals.

On HN, I hope we can share a correction like that respectfully: after all, they gave good info, except for a one-word slip of the tongue.

The critique seems to extend beyond correcting that error, becoming confrontational, questioning motivation and honesty with phrases like "supposedly worked in." and the long bit defending lifespan and enviromental impact against people who "don't seem to have a problem with oil drilling, fracking, coal strip mining, etc" - they didn't even touch on that subject.

Thank you for responding, I agree with your points, I did indeed make a mistake.

Which rare metals do they use? Silver for contact wires? If silver supplies were inadequate (they're not) these could be substituted for with copper, if a barrier layer was included between it and the silicon.

Maybe indium in ITO for those fancy transparent front contacts. Or tellurium in CdTe, supposedly still costeffective compared to “thick” Si cells. I would still give GGP a break it can be tricky to venture even small steps outside ones specialty these days

And, maybe in the future, gallium as a dopant in silicon cells, since it doesn't experience nearly as much light induced degradation as boron does. But dopants are used in very small amounts.

I think some power electronics uses europium silicide (or was that erbium?) as a gate material, so maybe in inverters? Again, the quantities would be small.

Which rare earth minerals go into solar cells?

No rare earth minerals are in solar cells. This is a famously bullshitty meme that was being propagated by Michael Shellenberger.

A simple search of the 'net will answer the question far better than this attack on Shellenberger. It will show that rare earth minerals can be used in PV panels as doping elements. Interestingly enough it is especially Perovskite PV cells which seem to benefit from the use of these additives [1,2]:

(1) Recently, use of rare earth (RE) ions doped nanomaterials in PSCs, has been identified as an effective means to address the aforementioned issues by expanding the range of absorption spectra minimizing the non-absorption loss of solar photons, enhancing light scattering and improving operational stability.

(2) Rare earth ion doped nanomaterials can be used in perovskite solar cell to expand the range of absorption spectra and improve the stability due to its up conversion and down conversion effect. This article reviews recently progress in using rare earth ion doped nanomaterials in mesoporous electrodes, perovskite active layers, and as an external function layer of perovskite solar cell.

[1] https://www.sciencedirect.com/science/article/abs/pii/S10020...

[2] https://www.sciencedirect.com/science/article/abs/pii/S13877...

That's just in the lab. If you buy PV modules right now the cells will not be doped with rare earth elements. And almost everything demonstrated in the lab doesn't progress beyond there (which is fine; that's how technology R&D works.)

I think the closest one could come to making the "REE in solar" claim make sense would be decoloring agents for the glass. Cerium could be used for this, but I think manganese is cheaper.

Cerium is common in solar coverglass used in space, but I'm not sure I've heard of it being used for terrestrial applications.

Here is the graph of exponential installed solar capacity [1]. Like Moore’s law, continuous technological innovation and investment will be required to keep the pace. We just hit 1 terawatt— and doubling time seems to be about every 3 years. So… 1,2,4,8,16,32,64,128,512,1028, 2056 terawatts in 30 years?

With 20% capacity, that’s equivalent to >300,000 Million Tons of Oil (MToE) per year. Current global energy consumption is 14,000 MToE [2].

[1] https://ourworldindata.org/grapher/installed-solar-pv-capaci...

[2] https://en.wikipedia.org/wiki/World_energy_supply_and_consum...

That chart ends at 2022, but 2023 was an even bigger year than you would expect from that curve, with over 500GW of new solar installed. https://www.iea.org/reports/renewables-2023/executive-summar...

Also: “ In 2023, an estimated 96% of newly installed, utility-scale solar PV and onshore wind capacity had lower generation costs than new coal and natural gas plants. In addition, three-quarters of new wind and solar PV plants offered cheaper power than existing fossil fuel facilities.”

Ahh, the classic case of seeing the bottom half of an S curve and projecting it out to infinite exponential growth.

The number of times things have experienced infinite exponential growth in all of history starting from the Big Bang: 0.

/jk just wait long enough and that zero will go to Infinity

Isn't the universe expanding exponentially since the Big Bang?

The expansion rate slowed down dramatically after the big bang and then sped up again, from Wikipedia:

> Cosmic expansion subsequently decelerated to much slower rates, until at around 9.8 billion years after the Big Bang (4 billion years ago) it began to gradually expand more quickly, and is still doing so.

Sure, but until we see the inflection point we can't know how much longer the bottom half of the S curve lasts— it might be 2, 5, 10 years, or it might have already passed; either way we'll only know in retrospect.

Those different options make a big difference on how much PV is part of the long term global energy picture.

Ah the classic case of rounding a two digit number to infinity to make a strawman point?

Nobody said "infinite".

The upper asymptote of an S-curve is often called its "carrying capacity". We expect an inflection point about halfway toward this point. What do you think the maximum capacity of global solar energy is? The total amount of solar energy hitting Earth is about 4.4 * 10^16 watts -- 44,000 Terawatts. If we covered 1% of the Earth in solar panels at a meager 10% efficiency, that's 44 Terawatts -- this is a reasonable low estimate for the "carrying capacity" from total solar irradiance. We're at about 1 Terawatt right now. A high estimate (remember, this is the absolute maximum) might be 10% of the Earth at 20% efficiency -- 880 Terawatts. Of course, if we run out of space on Earth, there's always more space in ... well, space.

Another "carrying capacity" could be the materials needed for production. As TFA illustrates, we have enough different ways of producing solar panels that we are not anywhere near maxing this out either.

So I think there's pretty good justification to think we're still at the very early part of this S-curve.

> 44,000 TW

Check your arithmetic; it's considerably more than that.

173,000 terawatts [1]

So, 1% at 20% capacity is 346 terawatts. That seems like a reasonable upper limit for earth systems.

[1] https://sos.noaa.gov/catalog/live-programs/energy-on-a-spher...

There's a considerable difference between the amount that hits the atmosphere and the amount that hits the surface of the Earth. My number includes average cloud cover. It's this order of magnitude, and a higher number only strengthens my point.

This gives me hope! Thank you for sharing!

> So… 1,2,4,8,16,32,64,128,512,1028, 2056 terawatts in 30 years?

What did 256 terawatts ever do to you?

> So… 1,2,4,8,16,32,64,128,512,1028, 2056 terawatts

Ah, a power series.

> So… 1,2,4,8,16,32,64,128,512,1028, 2056 terawatt

Multiplying by 2 is hard. Let's go shopping!

* That Wikipedia link gives an average global energy consumption of 4.810^12 watts (assuming 11.63 TWh per MToE). The solar insolation of Earth is about 210^17 watts The total power output of the Sun is about 410^26 watts. The total solar power output of the Milky Way galaxy is about 410^37 watts
Assuming exponential growth and assuming 20% utilization, that gives us

A fully solar economy in ~13 years * Kardashev Type 1 in ~59 years * Kardashev Type 2 in ~272 years * Kardashev Type 3 in ~381 years

Wondering if anyone could help shore up my understanding or point out a resource to get me a better handle on this.

Using the solar maps from here[0], you can find the kWh/day/m2 for the US. If I am in a say 5.7 kWh/day/m2 region and I have 1 m2 of a 20% solar efficiency panel, does that mean I would get 1.14 kWh usable out the other end? Or is it 20% * X% horribly lossy conversion factor?

If I want to math out 11kWh/day in the 5.7 region, back calculating would put me at requiring 9.6 m2 of panels (11 / (5.7 * 0.2). Again, if there is a horrible lossy conversion factor, that would just go into my denominator, correct?

Or am I missing something entirely? I tried to use this calculator[1], but I could not recapitulate the numbers they were generating.

[0] https://www.nrel.gov/gis/solar-resource-maps.html

[1] https://pvwatts.nrel.gov/

The map shows solar resource, irrespective of the technology you're using. For example, you could be using heat collectors that would capture closer to 100%. To keep things simple, for back of the envelope calculations, you can imagine 1kW per square meter. Subtract cloud coverage, night hours and then multiply with 0.2 for PV panels efficiency.

If you want 11kWh/day, you need: (5.7 * 0.2) * Y = 11, so Y = 10 square meters. You can double check this: 10 sqm should have about 10KW of solar potential energy, but with PV efficiency you're getting about 2KW, so to reach 11kWh, you need 5 good hours of sunshine on average.

You'd be able to generate 1.14 kWh at the panel level if you kept the panel pointed directly at the sun throughout the day [1]. This is called "2 axis tracking" and it was sometimes used for solar farms when solar panels were much more expensive. Now that panels are much cheaper, 2 axis tracking has practically vanished from the market. The added expense and mechanical complexity isn't worth it. Single axis tracking, where the panels just rotate to track the sun from east to west, is still popular in large solar farms. It captures more sun than leaving the panels stationary but has less complexity than 2 axis tracking.

For a rooftop solar panel, you're not going to have any sort of sun tracking. The lack of tracking will reduce your output at the panel level. You will also lose more output if dust, debris, and bird droppings don't get cleaned away regularly.

You also lose some energy when the direct current electricity from your panels gets converted to alternating current in the inverter. How much loss depends on the inverter and how heavily loaded it is.

The NREL tool you linked says it's designed for "homeowners, small building owners, installers and manufacturers", which implies that it's for rooftop systems. It includes estimates for those loss factors I mentioned above, which is why I expect that it falls short of the number you calculated.

[1] EDIT: I forgot another significant factor: temperature coefficient of performance. A panel gets its efficiency measured at "standard test conditions" which include a moderate (near room temperature) panel temperature. Panels lose some efficiency as they heat up, which means that they don't perform as well as you might naively expect in the middle of the summer. The loss varies by panel technology. The very best conditions for panel output -- where they actually surpass reported efficiency -- is "bright sun but cold air," like noon on a freezing cold day with clear skies.

So, the bottom line is that the simple kWh/day/m2 * panel efficiency * m2 of panels should be within the theoretical ballpark of generation, but the real world is a harsh mistress and will undercut you.

On that calculator resource, they provide a monthly and hourly spreadsheet, but even with the more detailed numbers, I was still failing to corroborate their presumably much more sophisticated modelling which accounts for other losses.

Thanks. Just spit balling numbers and trying to see what things look like.

> which include a moderate (near room temperature) panel temperature. Panels lose some efficiency as they heat up, which means that they don't perform as well as you might naively expect in the middle of the summer.

This is why vertical solar panels are becoming a thing, the additional cooling benefits increase output up to or beyond the optimal angle to the sun, and the better cooling also prolongs their life.

And they work better in winter if there's snow.

The "System Losses" breakdown shows the various additional factors they are derating the system by. The figures seem reasonable enough, and give an additional loss of 14%.

I put in my own address, which is in the 4.0-4.5 kWh/day region, and set the DC system size to 1 kW, which corresponds to 6m2 of panels (courtesy of their rooftop calculator). The NERL website estimated that such a system would yield between 2.45 and 6.48 kWh/day, with an annual mean of 4.71 kWh/day.

That works out pretty close to what the map indicates for my region: 4.5 kWh/m2day * 0.2 conversion factor * 0.86 losses factor * 6 m2 of panels = 4.64 kWh/day

Super appreciate your reverse calculation. The downstream 14% loss still is an easy enough fudge factor to use for a plausible output number.

The market need for cheaper solar cells seems to have evaporated, since the vast majority of the cost of solar projects these days is always in labour/land/wiring/inverters/grid connection/maintenance contracts.

That means saving a bit of money on the panels in return for lower efficiency is never a good deal.

Some are trying to combine perovskite with silicon solar cells because they specialize at capturing different wavelengths. So-called tandem solar cells.

I wonder what happened to singlet fission cells and other things trying to get around the SQ limit

Still being worked on.

https://interestingengineering.com/energy/paderborns-new-sol...

Thanks for the link - I used to be peripherally involved in this field and last I heard MIT (Baldo et al.) had resorted to some hafnium oxynitride layer, which is really not gonna drive costs down at all.

If the materials are cheap enough, we might be able to build them into other stuff that was going to use labor anyway (shingles, asphalt, siding, etc). No idea what the economics of this look like though, and electricians (a pretty expensive form of labor) will need to be involved no matter what, but at least theoretically cheaper cells can also deal with labor costs.

Like solar fences: https://next2sun.com/en/solar-fence/

Solar panels are already cheaper than wood, which is amazing.

Where are PV panels cheaper than wood? A 4x8 sheet of plywood runs about 40$ in north America, and roughly double that in Europe. I don't see PV panels anywhere near that price.

I can't recall ever seeing a plywood fence in North America.

You do see them from time to time in especially rough rural areas. I can remember one that was apparently painted from the "oops" paint section of the local hardware store surrounding a strip club next to a junkyard.

Sure. I don't recall seeing one, but the near-inevitability of this is why I didn't assert they don't exist.

Rarity speaks to how poorly suited the material is for building durable fences and the ~irrelevance of the cost of plywood in this subthread.

You don't want a plywood fence though. PV panels are cheaper than the decorative wood fences are made of. Wood fences are easily 100€/m around here.

It’s not cheaper than wood. It’s cheaper than certain specific kinds of wood used in specific applications.

A 2x4 stud is like $3, for example. Decorative cedar is quite a bit more expensive.

I don't think that you actually thought that the person you're replying to said that solar panels are now cheaper than all wood always.

Ya, that's my take too. Continuing R&D on alternatives like perovskite might open up new use cases. Like using transparent solar cells as windows. It's worth investigating.

Bonus, it keeps scientists employed, maintaining our capacity.

>> other stuff that was going to use labor anyway (shingles, asphalt, siding, etc)

No. None of that ever works. Everyone has the "good idea" of cramming PV into some other product thinking that doing so will somehow reduce labor. It never does. Solar shingles are typical. They sound great but in reality require hundreds or thousands of electrical connections all spread over the moving flexible surface that is a wooden roof. You will be chasing electrical gremlins the moment the temperature shifts. And fixing any of those gremlins will involve penetrating the waterproofing, the core function of any roof. It is far easier to build and maintain a normal roof and then mount dedicated panels atop. The same too with siding. Want solar walls? Build normal walls and hang solar panels on them.

It is like building a computer into a desk. It seems like a great idea that will save space and keep your office tidy. There are lots of youtube videos about such builds. In reality, it is expensive on day one and extremely inconvenient to maintain in the long run. Nobody ever does it twice.

Part of why it doesn't work, though, is that PV is too expensive.

If it's cheap enough, you can tolerate failures and poor illumination of the panels for things like fence panels or whatever.

I do agree you need big panels to not have excessive labor from connections.

>> If it's cheap enough, you can tolerate failures

But you just can't. When you are using lots of tiny things all connected through each other then you have less tolerance for faults, not more. One bad connector can mean that an entire run of shingles is dark. So even a 1% fault rate, if you have a few hundred connections in each run of shingles, means that basically nothing is connected. Or think of a long fence. One broken bit can mean the entire fence after that break is no longer connected. You're just setting yourself up for a long day of checking connectivity only to have the fence shift again.

Most of what you say was anticipated by the comment you replied to:

> > I do agree you need big panels to not have excessive labor from connections.

> You're just setting yourself up for a long day of checking connectivity only to have the fence shift again.

If only we had ways to make long runs of wiring relatively reliable.

My point is: there's second order effects: expensive panels need to have as high of a capacity factor as possible; high capacity factor constrains installations and increases other costs. If you cut 2/3rds of the cost of the panel away, other costs decrease, too, and more types of installation become reasonable.

There are also non-linearities. Obviously there are some regions in in the cheapness/efficiency/durability space that vastly increase the practical ability to deploy these things. If we had 99% efficient panels that cost pennies per square meter, and last for years, then lots more applications could potentially open up. A 50% cheaper panel may not unlock that now, but it brings us closer.

Even if we never get to any of these thresholds, its worth a shot. Cleaning up the energy sector needs to be all-hands-on-deck and people researching this stuff doesn't preclude policy changes (subsidies, federal job guarantee/new CCC, etc.) to address the labor angle.

PEV in metal roofing seems more workable.

Attaching to the roof requires screwing fasteners through the metal in fairly arbitrary positions based on the underlying framing. It’s not going to be easy to have electric connections.

Normally folks only put holes through the lifted up/corrugated sections (so as to minimize the likelihood of leaking) so all the area in-between (the larger flat sections) are where the electronics/solar arrays would go.

I think future designs of panels might be designed in such a way an electrician isn't required. All foolproof plug'n'play connectors and designed in such a way you cannot plug them in in an unsafe way.

You don't call an electrician every time you plug in a hairdryer, and a hairdryer is typically higher voltages and currents than a single panel.

Higher voltage than a single panel, but a string of panels easily hits hundreds of volts. Even worse they can be hard to make safe, since as long as the sun is shining they are generating energy and roof installers don't like working at night.

You can avoid this by using microinverters, but they're a pretty substantial premium on each panel and an added point of failure.

There is lot of tech around solar panels that is being effectively obsoleted by the plummeting costs of the panels themselves. Why bother trying to squeeze out the last few percentage from each panel when it's so much cheaper to just install a couple more panels to make up the difference? This is the big difference between countries like the US where solar installs are still expensive at $3-$6/watt and countries like Australia where home solar installs are under $1/watt.

It's even worse, perovskite degrades faster than silicon cells so you are getting lower efficiency short life cells. Tons of money is being invested in trying to fix the lifespan problem.

Ground/roof solar is still >60% panels around here

Labor costs are less when the weight is 10x less.

A little, but the design of the support structure is dominated by wind loads. Depending on where you are, it can need to survive wind gusts of 90 - 160 mph. At the low end, 90 mph corresponds to 1000 N / m^2, which is more than the weight of any kind of solar panel.

I've been waiting for consumer level panels to get cheaper forever. You'd think by now that you could get a 200W panel for $50. But they have been the same $200 for what seems like a decade now (I suppose they didn't go up with inflation, but still)

You can pull up at MO Wind and Solar off of MO 60 and grab 600W panels for $195/ea today: https://windandsolar.com/risen-595-watt-mono-solar-panel-bif...

That's ballpark where you say, about $65/200W.

They also have remanufactured panels, like the 410W trina ones, even cheaper. Good outfit, nice and helpful, bought several lots of panels off of them.

It seems like the size & cost of the panel has stayed the same and the power output has gone up.

I just put another 5x550W panels up 2 months ago, and they are now 101eur per panel including tax! 30e less than I just paid. wish I had space for more!

It's frustrating. Panels around $0.25/W exist, but it's really difficult to get your hands on them in small quantities as an individual. You can either string together a bunch of tiny eBay specials or drive halfway across the country to find a distributor of the panel you want who's willing to sell to consumers.

In Germany, 405 Watt panels new can be had for 65 euros. The law allows up to 800W to be connected to homes with relatively little bureaucracy, as a balcony solar power plant which renters can install without modifying the building. This seems to have pushed the price down, so there are many 800W kits including panels, an inverter and cables available for under 400 euros.

Mind sharing a link to buy one? Not in Germany but am in Portugal looking to buy

Sure thing, here is the solar panel: https://shop.green-cluster.de/products/405-watt-full-black-p... and here is the kit: https://www.netto-online.de/balkonkraftwerk-830-w-600-w-phot...

There are always lots of new deals, if you search for "photovoltaik" on mydealz.de you can find the latest offers in Germany

Cost of labor might be significant in some countries, but in a lot of countries, cost of materials is still the significant cost barrier to installation of solar. If the cost of materials were to come down significantly, even if it is at the expense of some efficiency, and also make the installation much more flexible, then definitely more installation will happen. Capital invested will be recovered in a shorter duration making the investment a lot less risky. This makes lending programs much more accessible, making the whole thing a more self-reinforcing cycle.

It feels like the problem here is less about the solar industry and more about the construction and skilled trades generally. Everything involving an electrician / electrical contractor has gone way up in the last 10-15 years, along with other things. So whatever savings are being saved on materials are just being eaten up by rising labour costs.

That and government incentive programs for home energy efficiency seems to have just inflated prices and stimulated demand to make the installation costs worse. Quotes I've gotten on heat pumps for example have been ridiculous, and solar much the same.

Hate to say it, but a recession might be what "fixes" this. Not that I want to deprive trades people of a good livelihood, but it feels like the end consumer is getting screwed right now.

True but the cost of labor goes down if they get cheap enough.

For example: buy panels so cheap you just leave them on the ground. If they get damaged who cares. Some people are already using them as material for fences. Not a great angle, they’re cheap who cares.

Perovskite solar cells are the best candidate for solving the problem that you pointed out. The best perovskite/silicon tandem cells in the laboratory have 33% efficiency, and the theoretical limit for this type of cell is 43%.

Do you know if there is any (even theoretical) work to solve the lifetime issue? Perovskites degrade in sunlight (order of months) making it seem unlikely they would ever be useful outside the lab.

The are plenty of places with really low labor costs that would love cheaper panels

> That means saving a bit of money on the panels in return for lower efficiency is never a good deal.

This does not at all logically follow from your proceeding statement. Cheaper solar panels mean they can be used in different ways with different labour/land/wiring/inverters/grid connection/maintenance requirements.

if the substrate really can remain even a little bit flexible however, this opens up entirely new deployment opportunities.

(In the US)

In my part of EU the cost of getting 10kW of solar installed has gone from around €7000 in 2021 to €2000 or even less today. That is after government incentives, but the incentives have not changed during that time - it's a fixed amount per kW. The price reduction is due to the cost of panels and equipment going down.

As an outsider, what's the appeal of PeSC? The paper says they are less efficient and harder to manufacture, but also that they are the "next generation".

I would have thought the next generation would be more efficient or easier to manufacture or both.

As I understand it, silicon cells have largely been optimized to near their peak potential. There's not much room left for improvements at this point.

Organic and perovskite cells have a higher potential efficiency. Just like was the case with silicon, it will take years of development and incremental improvements to see higher efficiencies in these technologies. Silicon cells were also not very good at the start of their development.

In that sense perovskite has the potential to be the next generation of solar cell. New developments, such as the one demonstrated in the linked paper, are just a step towards that ultimate goal of more efficient solar.

I'm not an expert in this field, so please feel free to correct any mistakes I've made.

Which ones don't have heavy metals?

"Harder to manufacture" is relative. The hope is that they would be easier, because they're not monocrystalline and don't require the high energy https://en.wikipedia.org/wiki/Czochralski_method to produce a semiconductor substrate. The paper (and the general "pitch" of perovskites) plans to use roll-to-roll printing on flexible substrates.

a-Si has been at or beyond this level of efficiency since the 90's and also does not require a Cz or FZ process.

It's encouraging to see progress but where perovskite thin films show potential is in an integrated mechanical stack application with silicon, where they can supplement each other's barely-double-digit efficiencies focusing on different parts of the spectrum to combine to reach something on par with traditional crystalline silicon, but thinner and with lower production costs.

Seeing thin film beat crystalline silicon is like seeing nuclear fusion become cost-effective. It's perpetually 10 years away, and has been since the 70's.

The benefit of PeSC is that it can get into efficiencies higher than 20% (higher than traditional industrial cells which typically top out around 18-19%) but the problem with them to date has been the cost of manufacturing. This article is all about solving the "cost of manufacturing" problem.

If we can get the cost of manufacturing PeSC cells down to the same levels of traditional crystalline silicon PV cells then the old style will become obsolete. It's a simple evolution of photovoltaic technology with PeSC cells being the next generation. Not so different from any tech where it's expensive when first introduced but as mass adoption and manufacturing improvements take place the cost comes down.

PV cells on thin flexible substrates are interesting for space applications, where they could have very high power/mass.

Roll to roll amorphous silicon solar panels go back to the mid-2000s.[1] Those were commercial products, and, at the time, not much more expensive than rigid cells.

[1] https://en.wikipedia.org/wiki/Energy_Conversion_Devices

What is "roll-to-roll", in this context?

https://www.ronaldindia.com/all-you-need-to-know-roll-to-rol...

It means, to print long sheets on rollers like a newspaper might.

Let’s make sure they use little to no lead first before we deploy them to homes all around the world.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9860350/

Ovonics was doing roll-to-roll solar cells in the 1980s, were they not? This is exciting if the technology scales better or something, but I can't help feeling that we've stagnated a bit.

What material though? Nanosolar was also doing CIGS roll-to-roll cells, and that didn't work out too well for them.

costs of solar panels of all sorts have dropped 100 fold in that time. I'm not sure I'd call that stagnation.

70 USD for 100 W seems pretty bad? Especially given the low efficiency.

I would expect the price to decrease over time given perovskite solar cells are currently not yet mass produced and still actively being researched.

Or, the degradation problems keep them from ever taking off. If they do, they may only do so as part of tandem cells with a silicon PV bottom layer.

A quick look around gives me a current figure of ~1 USD/W in Australia (dunno if that evolved a lot in a short time with the current inflation and currency rates).

Money quote:

"The cost for [production] Seq[ence]. B is likely to be lower than 1 USD W−1, and Seq. C could be lower than 0.5 USD W−1. These represent a significant reduction to the cost estimate from previous works of around 1.5 USD W−147. This results from a similar or lower cost in $ m−2, and a higher recorded efficiency. However, the technology is still not able to compete with mass-produced silicon solar cells, for which module spot prices have been lower than 0.30 USD W−148. Despite this, opportunities may exist in niche markets that value the lightweight and flexible nature of these modules, as discussed in our previous work47. The next step for the technology would be exploring high-value PV markets at the predicted manufacturing costs while addressing the remaining high-cost components to sustainably advance the technology towards commercialisation. Supplementary Fig. 12, with about 5 USD m−2 module cost (excluding encapsulation), shows the potential for the further cost reduction by eliminating the remaining high-cost components."

The cost model is based on high labor cost assumptions of $25-36 USD/hr. See page 14 below. I wonder how much lower China's industrial capacity could reduce the costs.

https://static-content.springer.com/esm/art%3A10.1038%2Fs414...

Reply on news.ycombinator.com