Printed Solar Panels for Less Than $10 a Square Metre | Hacker News Reader

Printed Solar Panels for Less Than $10 a Square Metre

newcastle.edu.au

252 pointsdakna83 comments

Hey, I know the phd student who printed out these cells and set up the demonstration! It's part of the university's push for large scale organic solar demonstrations along with the smaller test cells.

From talking with him, the technology isn't really ready for prime time yet but it's getting pretty close. I think the key point is that efficiencies in small scale cells and larger scale manufacturing are still climbing (the same group has achieved greater than 5% in a cm^2 test cell iirc) and the printing is incredibly cheap and very amenable to fast scaling up.

It seems pretty obvious that you needed more efficiency for it to be a viable rooftop solution but the guy who set this up claimed that the fact he could just stick down some velcro and stick on the cells opened up some different use cases with cheap and lean installations supporting cheap cells.

All in all, if you look at how far the technology has come in the last 5 years alone, then it's a pretty exciting field to follow.

> It seems pretty obvious that you needed more efficiency for it to be a viable rooftop solution but the guy who set this up claimed that the fact he could just stick down some velcro and stick on the cells opened up some different use cases with cheap and lean installations supporting cheap cells.

I imagine one end-case could be using this as a cladding material, you could cover buildings, and the low efficiency would be balanced by the low cost.

The major problem is that these materials are unstable and degrade quickly, if you can get the lifetime up from 5-10 years now, to 20-30 years, then it really starts to look attractive.

I don't think commercial structures typically keep their cladding beyond 10 years, unless it's a warehouse, so I'd imagine if it's anything other than. A warehouse they'd be happy to just have it reliably last 10 years then whatever from that point forward.

Also depending on who's involved and the business vertical, plus tax and policy structure, there could be a cottage industry around skinning warehouses I'd imagine, even if it needed to be re-skinned ever 10 years, but I agree that anything less than 10 years isn't sturdy enough for many applications beyond super niche.

For many use-cases, the cost is far more important than the efficiency. e.g., I have a chicken coop that's a few hundred feet from the house and not worth the trouble of running wiring out there. In the past, I used battery-powered LED lighting. I'd much rather switch to solar. Same for the unfinished shed I started building years ago.

Those of us in rural areas often have to cope with wiring difficulties that solar would solve.

Looks like the conversion efficiency is between 2% and 3%, so, pretty low compared to silicon based PV
reference
https://cleantechnica.com/2017/05/17/researchers-australias-...
http://reneweconomy.com.au/uni-newcastle-team-tests-printed-...

edited to clarify 2/3%

This is what I was looking for. The 'Performance' section of the article was pretty lacking on any actual details. They mention that it's more consistent over 24 hour periods, but that could've just meant that output was low, regardless of the time of day.

To be clear to those reading, that's 2% to 3%, not two-thirds percent.

Thx, fixed

Last I heard silicon based was getting close to 25% efficiency. So for the same amount of energy of a square meter of a traditional panel, you'd need 10 square meters. That's a huge solar panel for only a regular amount of energy. I assume they'll continue to get more efficient though, maybe in a few years it'll be worth it.

That would put them at 20-30 watts-peak per square meter, or $0.50 to $0.33 per Wp if the cost is $10/m^2. At the lower efficiency, crystalline silicon modules are already cheaper. At the higher efficiency, it may be cheaper than crystalline silicon modules but just slightly. It would need to reach commercial-scale production very soon to avoid being overtaken by the relentless manufacturing optimization of standard c-Si technology.

Another problem is that "less than $10 a square meter" is almost certainly an estimate referring to scaled-up commercial production. Printing 100 m^2 of devices for that test site did not cost under $1000.

Finally, I'd wonder about lifetime and endurance. Crystalline silicon itself is extremely durable. With good supporting materials and thermomechanical design, c-Si modules can go 30 years before they degrade below 75% of original rated output. (There are at least a few solar modules installed in the early 1980s that are still working fine.)

Half the cost of solar panels is installation, because the price of panels has dropped so much. Low-efficiency panels are thus a lose; if they're less than half the efficiency of existing panels, they could cost zero to make and not be useful.

This isn't the first roll to roll process producing solar cells. Ovshinsky's Energy Conversion Devices made such things for years; in 2008, they were the largest producer of flexible solar cells. Their Uni-Solar unit is still active, but not a big player.[1]

(That was one of Stanford R. Ovshinsky's many inventions. He invented flexible solar cells, nickel-metal-hydride batteries, and much of thin-film electronic technology. But each time, some other technology pulled ahead.)

[1] http://www.uni-solar.com

I don't disagree, but that's really absurd. Installation is not particularly difficult. There are dangers, but It's not significantly challenging. The market requires it - because all those solar companies wouldn't exist if it wasn't profitable. Still... every time I see an installation quote I'm blown away by how much people pay. People don't seem to care about numbers once they have to get financed for a few years.

Installation is labor-intensive custom work. Panel manufacture is a high-volume production process.

And skilled labor at that. You don't want your roof leaking (if that's where it's installed), nor do you want an electrical fire.

I also presume the current installation prices might include some sort of maintenance guarantee. That adds to cost as well.

You should check out the planet money podcast on this topic. Turns out a big part of (consumer?) solar getting cheaper was the invention of a new tool, some kind of wrench I believe.

http://www.npr.org/sections/money/2015/04/10/398811199/episo...

> GOLDSTEIN: But again, just in the past few years, installation has actually gotten a lot cheaper. As a way it's gotten cheaper is it's gotten faster. You don't have to pay those guys to be out at your house for as long. And when I was out at John's, I actually saw some of the reasons it has gotten faster.

> RYAN BARNETT: This is the Zep Tool. This is the end-all be-all of tools.

> GOLDSTEIN: The Zep Tool? Led Zep Tool?

> BARNETT: Yeah, exactly. (Laughter). One tool to rule them all. Yep.

> GOLDSTEIN: That's a guy named Ryan Barnett (ph).

> What's a Zep Tool? I mean, it's a wrench. It's basically a wrench. It's got these little marks on it or whatever, but the key thing is it's part of this whole installation system they use.

https://www.youtube.com/watch?v=YtiKyp12ej0

I think it's much more accurate to call it an installation platform/system, as the wrench is fairly incidental. The hardware is the innovation. Pretty cool overall.

Since owning a house I am no longer surprised when seeing high costs for quotes of solar installs. Pretty much everything from a simple plumbing job to getting a window fixed has gotten very expensive.

I'm sure the costs for a contractor doing solar panels on a complete new build "cookie-cutter" subdivision could be quite reasonable but anytime it's a one off, with every house having its own unique quirks, it becomes expensive. It's also one of those things where I assume liability for the contractor can get quite high. Panel falling off the roof, or causing issues with standing ice or roof leaks, etc. Additionally it requires expertise in not only roofing but also electrical circuits (so a qualified electrician).

According to this [1], installation costs are way less than half of solar panel costs in 2016 US$. Looks like 10% - 15% depending on the installation case. See Figure ES-1.

[1] http://www.nrel.gov/docs/fy17osti/67474.pdf

Look at page viii on your link. "Hardware costs constitute about half the total price of our small-battery systems. The largest single hardware cost for these systems is the 6-kW battery-based inverter ($3,596), followed by the PV array ($3,584) and the lithium-ion battery ($3,000)."

PV only is the orange box at the bottom. The cheapest option is 3,584 for the panels and total cost of 15,581. The most expensive option is still just 3,584$ for panels and a total installed cost of 47,171$. Trying to drop that 3,584$ portion is not going to make much difference a this point. But, needing to install more panels would significantly increase other costs.

Agreed, and thanks. It's clear that dropping panel costs really wouldn't affect the installed bottom line much.

I just focused on installation costs, which in all cases seem far less than 50% of the overall system cost. This suggests that there isn't much economic leverage in reducing (or preventing growth of) installation cost, either. Only in the large battery cases is there a single component (the battery) that dominates other cost components.

Also notice that this is for solar with energy storage. If you want to tap into the grid then you need a permit and an electrician in addition to an emergency cut off for firemen. That ends up being $Ks which are related to installation (not panel/inverter/mounting) rather than a battery.

Check Project Sunroof for a quote: https://www.google.com/get/sunroof

It's not uncommon for installation costs to be over 30% (especially for smaller installations and depending on roof conditions), and when you only get the Fed subsidy when you pay taxes that number ends up looking bigger (even if it's the same % in the end).

Inverters are improving too. New semiconductors are being developed and improved.

Half the cost of solar panels was not installation for my system. Care to provide figures?

A simplest comparison would be it's 10x less efficient and 20x cheaper. For residential might not be useful as you'd might want to maximise output.

For solar park or industrial rooftops (especially when building new) that could be good price advantage.

Sure in the First World that makes sense. But if you're looking ahead and pondering how to bring power to the millions (if not billions) who are not First Worldites then this is fascinating.

If it's easy to manufacture local to need that's also an additional perk.

Would you clarify for me? Do you mean 1/10th efficiency and 1/20th the cost?

Yes. I am not sure what's the problem, sir.

I asked for clarification, because 20 times cheaper than $10 is $200 less or $-190. One 20th of $10 is $.50.

What is the price of silicon based panels?

Industry average 300w panel (17% efficient) is about 1.6m² and comes in at about $0.40/W wholesale. At scale you can go cheaper. Manufacturing cost is likely lower. Shipping costs can be 15-20% on top of that.

~$120/m².

What would be helpful to know:

How much of every material is required/m²?

How much does 1m² weigh?

What is the 75% output lifetime?

Efficiency graph at varying solar iridescence?

Manufacturing yield rates?

For a very rough comparison, according to this, conventional solar panels cost about $10-$12 per square foot, or very roughly $100 per square meter.

https://www.quora.com/What-is-the-cost-per-Sq-ft-for-solar-p...

Of course, what we really want is a comparison in terms of cost per watt.

Maybe equally important to the cost of these panels is the ease and cost of installing them. These new printed panels are very flexible/lightweight and can be deployed easily and even temporarily.

Yea, can't make a meaningful comparison without (a) cost per watt or (b) cost per total present-value of lifetime power delivered

Yep, but the article states "production costs":

>The technology delivers unprecedented affordability at a production cost of less than $10 a square metre.

the cost you are citing are "sale prices", they are not the same thing.

Given that the Quora post cited is accurate (and "current" as the price of conventional panels is dropping constantly, and a two years old post sounds like outdated) the ratio is much smaller than 1/10, more probably around 4/10 or 5/10.

Still a "big thing", however, provided that it works, that the cost estimation is accurate, that the efficiency is comparable, etc.

Does anyone have any information on how efficient these are how long they last and that kind of thing? All I got for the article was a lot of hype and $10 a meter.

Woild like to know normal panel's efficiancy and costs for comparison

You can buy solar cells off aliexpress at ~$0.33 / watt (this is the lowest price atm). The most efficient cells are at 23-25% efficiency, but these are a bit more expensive (closer to $0.5).

Here's the first example I could find (https://www.aliexpress.com/item/10-Pcs-17-6-125-x-125MM-Mono...).

In the real world you cannot expect to need less than around 5 square meters for 1kW of peak power (using the most efficient panels), the standard "rule of the thumb" being 7 sq meters per 1 kW, some lower efficiency panels (not used anymore) will stretch that up to nearly 9 square meters per 1kW.

This puts the efficiency in a range between 1/9 and 1/5, i.e. between 11 and 20%.

So if this new panel has a 1/10-1/8 of the best above, it plainly means that you will need in your typical single home installation more space than your roof has.

Let's take a 6 kW installation, with best current panels it will be 30 square meters or however at the most 45, with these printed thingies more like 240-300 sqm or more.

And 2-3 years of expected life!

Maybe the principle is fine but it must be modified/bettered to raise the efficiency and to increase the durability.

> "In terms of efficiency, the solar cells are currently showing around a 2-2.5 per cent power conversion rate" [1]

Since you can probably get 250-350 watts of sunlight (avg) per square meter, it means 1 sq m generates 5 to 8.75 watts. Which is 0.5 to 0.875 watts per dollar (or $1.14-$2 per watt).

I don't think that's cheaper than the current offerings, especially if you don't consider these printed panels take 10X more space. Also the lifespan isn't known.

But consider that the price of these new panels will likely fall, if mass-produced.

[1]: https://cleantechnica.com/2017/05/17/researchers-australias-...

EDIT:

Other people say it's $0.40 per watt. [2]

[2]: http://www.thefifthestate.com.au/energy-lead/energy/cheap-an...

I'm not sure where you are getting the 250-300W/m2 of irradiance. The standard value used in all solar industry calulations is 1000W/m2, but in practice, you should expect to see 800W/m2 at sea level and moderate lattitude.

Maybe that's the 24-hour average?

Average capacity factor for solar cell is about 25% (night, clouds, seasons, etc).

1000W/m2 * 0.25 => ~250W/m2.

1000W/m^2 is average daily peak irradiance. That number is useful for designing system capacity, but doesn't reflect output throughout the day.

1000W/m2 is at noon at the equator without clouds/haze/dust.

1000W m^2 is actualy closer to high noon at Kansas on a fully clear day (this is equivalent to 2,000 umol/cm^2/s-1) Equatorial solar irradiance is actually higher (closer to 1,050W/m^2) because the equator is just a bit closer to the sun.

>Equatorial solar irradiance is actually higher (closer to 1,050W/m^2) because the equator is just a bit closer to the sun.

No, no, no. Equatorial is higher because the sun is more directly overhead on average than in Kansas.

That's not how the inverse square law works for photon flux density.

I'm a researcher in this field. Printable, flexible solar cells have been researched for over a decade and improvements on their efficiency plateaued at about 10% almost five years ago. That's at lab scale being tested in a nitrogen glove box, using the most exotic designer molecules and printing techniques that don't scale. Many companies have tried and failed to commercialize this technology in the recent past, Konarka being one example. I believe Heliatek in Germany is the latest iteration, no doubt propped up by their massive solar subsidies.

The main problem is that the raw organic semiconducting materials are currently not produced at scale, and no chemical company will risk ramping production when the end product is only a 5% efficient cell that will be outcompeted by c-Si in almost every application except niche cases like the chicken coops described by others here.

That being said, my best guess is that printable, flexible electronics will be seen in flexible displays/wearables within 5 years, RFID tags on product packaging in the same time frame, biomedical and soft robotics applications within 10, and flexible large-area lighting (think wallpaper with programmable lighting zones) within 10 as well.

But in the energy game, efficiency and raw material costs determine winners, and printed OPV is losing on both.

> On our lab-scale printer we can easily produce hundreds of metres of material per day, on a commercial-scale printer this would increase to kilometres. If you had just ten of these printers operating around the clock we could print enough material to deliver power to 1000 homes per day,” said Professor Dastoor.

That being said, may-be 10Km worth of these can power 1000 homes.

It costs $10 per sq.meter.

The costs of the inverters, installation, and connection already dominate the cost of solar.

Before incentives in California, you are looking at $5k/kw and the solar cells are less than 20% of that.

Is that true? I was looking at $4/kw (all in) 5 years ago. Not in CA, but that should come close to a 25% difference. It was mid-tier panels, but a high quality inverter.

Yes, but did that include federal and state subsidies? With Fed tax incentives (as most people quote) on a 5kw installation in CA it's $15-20k depending on who and where. So that would be in the $3-4 range... a lot of that is labor (almost half) so other parts of the country are lower. Rural places tend to be higher though.

My real point was that labor price variations can easily be larger than the total actual panel cost.

Curious, what is the need for a 5kW array? Do they typically only produce at 5-10% of peak? Or do the relatively low panel prices mean that if you are going to build 1kW, you might as well build 5?

I just went outside and checked my utility company smart meter... the whole house isnt even using 100W right now (though thats on the low side, I'd think).

That will change as soon as your fridge/washer-drier/TV much less hairdrier, AC, or stove turn on (it's peaky so net neetering really helps). Typical 3 person SFH uses ~1.5kw average or 1000kWh/mo ($110/mo at CA rates).

Well installed solar at US lattitudes in a very sunny area generally gives you something like 25-30% of peak as an average (it's not up half the day and averages less than 45deg over the day ignoring clouds).

Lighting/laptops/phones take a tiny amount of energy. For example, a phone battery nowadays (e.g. iPhone 6) is about 10 watthours with say 5 hours of screen-on-time. Disregarding it's on doing background stuff all the other hours, it approximately uses just 2 watt on average.

But the very first microwave (oven) I looked up on Bestbuy used 1500 watt, and if you've got something cooking for 40 minutes in there you're easily using 1 kWh, or about three months of iPhone use.

> Curious, what is the need for a 5kW array? Do they typically only produce at 5-10% of peak?

As for peak production, well yes-ish. The average peak solar hours in the US is about 4 per day, Europe is probably at 3, in Texas it's probably 5. That means a 5kW array will produce about 25 kWh a day, or 750 a month in Texas, saving about $85 a month / 1k a year in electricity purchase at 11.5c a kWh. So yeah, on average a panel in the US produces at roughly 15% or so of peak. (±4 peak hours per day).

Wow,Solar prices in the USA are pretty high. I live in India where solar is under 1.2USD/watt installed. This is for a good module and very good inverter.

I have seen quotations for under a dollar a watt using cheaper components

If labor was free, and you didn't need a permit or certified equipment to attach to the roof and utility lines, and there were no taxes, then that's close to what you would pay. $0.60 cells and $0.40 string inverter $0.20 cheap brackets.

For very large commercial/utility installations it falls close to that ($1.40 including land and transformers).

https://pv-magazine-usa.com/2016/09/29/nrel-u-s-utility-scal...

That was with all costs included and no subsidies. With subsidies it was about $1-2/kW. That was 5kW before derating.

Wow! Sounds like a good deal. You could do it for components without installation costs, but even utilities can barely get that today in California on 100MW builds.

2% efficiency would give you ballpark 5W per square meter (peak) or about .25W for a sheet of paper (less indoors or at night)

So, if this can be combined with a paper-thin e-ink display (and, if needed, a flat sheet capacitor for power storage), would that be enough to make true paper-thin displays at reasonable price?

I wonder if these could be used to produce panels on a remote planet/planetoid. I.e. could you use this to create a bunch of panels to place on the moon, even with the low efficiency you'd still save a lot by not having to ship them.

You would still have to ship the raw material, as well as the machines to assemble/print them.

and figure out a way of transferring power through space..

An interesting addition to the mix. The design space around solar power systems focuses on either cost (like in this example where efficiency is low) or efficiency gains regardless of cost[1].

Presumably if we get to a point where you can cheaply print 25+% efficient cells then we're "done" as it were on improving solar cells :-)

[1] https://arstechnica.com/science/2017/03/japanese-company-dev...

> “By reinventing the delivery model we remove the need for initial lump sum outlays, overcoming the key barrier to community uptake and ensuring that the science actually ends up on our rooftops,” said Professor Dastoor.

But can't banks just solve this, by financing panels upfront? There's quite some money to be made there, I'd guess. And the risk is limited.

A lot of cost in installations is now other factors such as casing, peripheral electronicls like invertors, labor, financing, etc.

A French company has been doing solar panel printing for while now. What's so new with this one ?

Shingles are about $8 per sq meter, uninstalled. Plywood is somewhat more. How durable is this stuff?

If the efficiency could just increase (to ~8% maybe) and inverters got a bit cheaper still, I think you'll start to really open the "just for fun" part of the market. I wouldn't hesitate to put them on my shed.

Plenty of cheap inverters - design your system to run at higher voltages. Run your panels in series to get to ~120V first, then get a simple switch inverter since you're already matching (in North America where I am) your mains voltage, those inverters are cheaper and usually more efficient. Then you feed that into your MPPT, which should have the step-down for 12V battery charging and cleaning the input 120VAC power..

Efficiency is the bottleneck of that's project, perhaps if it became an open source project creative people can create fantastic uses for it.

Could I wear it and generate power while I walk around? Could I sell the power I generate when I reached my destination?

How those films hold under direct sun?

What's the cost per Watt?

What is the price per watt?

Yeah but do they mine bitcoin

> If you had just ten of these printers operating around the clock we could print enough material to deliver power to 1000 homes per day,” said Professor Dastoor.

The US has 100M homes. That would require 100,000 days, or 300 years...

One could consider having more than just ten printers.

You missed the "just ten".

That also means if these things decay at the same rate as normal solar panels (producing an equivalent of 30 years of like-new power through about 35 years of lifetime), only 100 presses will be enough to power the entire U.S., or about 300-400 presses if you include all power needs, not just home. Which doesn't sound like a lot, one press per million people.

But where are you going to put these things? They are quite inefficient compared to conventional solar, so it’s not like you can just cover roofs in them and be done.

Yes i already realized that... Essentially that is half-scam, with that low efficiency they are pretty much useless.

How about a calculation with the number of new homes per day? In other words how many printers do we need to satisfy the solar energy demand of all future homes?

Reply on news.ycombinator.com