7 ms·

Once seen as fleeting, a new solar tech proves its lasting power

150 points by clouddrover 4y ago

ncmncm 4y ago
"Average temperature" doesn't mean much. Deterioration is concentrated in the periods of peak temperature, non-linearly. It makes it hard to express practical longevity with averages. You might be able to say something like, "it should last X years if it stays below T2, and spends no more than Y hours between T1 and T2." Then, actual lifetime at T2 provides an upper bound on that. They don't say in the article, but perovskites are amenable to stacking with different wavelength capture characteristics, so can be built to capture above 40% of incident solar energy. That, and being radically cheaper to produce than Si cells, means there is great interest in this result.
- smaddox 4y ago
  Deterioration tends to be roughly linear on an Arrhenius scale, which is basically the scale which is exponential on the inverse temperature. This is because deterioration tends to be roughly linear on the activation energy of defects.
  - ncmncm 4y ago
    What that means in practice is that hardly anything happens below a certain temperature, but it goes all to hell in a flash, above that. The "certain temperature" for any particular sample depends on what you consider "a flash". You would be pretty peeved if your solar panel went all to hell in a year, so you calibrate accordingly.
    - nicoburns 4y ago
      I guess these could be particularly well suited to colder / less sunny climates. Which could be a very good thing, given that in sunny climates existing technologies are already pretty much good enough.
      - ncmncm 4y ago
        "Floatovoltaics". The first place to put solar farms is floating on reservoirs and other calm water.
      - ncmncm 4y ago
        Deserts are just a really dumb place to put solar farms. But that is where ignorant investors want them, so that is where they are going, for now.
tomahunt 4y ago
https://www.science.org/doi/10.1126/science.abn5679 https://www.science.org/doi/10.1126/science.abn5679 link to the article in Science
dmix 4y ago
Wikipedia has a neat graph comparing the efficiency of different types of solar cells including perovskite, which shows how fast it's increasing compared to silicon/multi-junction. https://upload.wikimedia.org/wikipedia/commons/thumb/3/35/Best_Research-Cell_Efficiencies.png/1280px-Best_Research-Cell_Efficiencies.png https://upload.wikimedia.org/wikipedia/commons/thumb/3/35/Be...
- jopsen 4y ago
  Isn't it more cost than efficiency that matters?
  - kirrent 4y ago
    Not really. Materials and manufacturing costs for perovskite probably already beat silicon if anyone scaled up. It's efficiency and durability we're waiting on. If you look at pricing a large solar farm you'll be surprised that the cost of modules is probably only a third of the total cost but when you reflect on land, cabling, inverters, grid connection, and installation then it's not really surprising. Given these costs, you've got an obvious incentive to buy higher efficiency modules for a small premium for savings on installation, mounting, and land use and a disincentive to buy cheap and not yet high efficiency perovskites. This is why companies like OxfordPV are targeting tandem perovskite-silicon cells with efficiencies at 30%. Even if this costs 1.5 times as much as a 20% efficient silicon cell, it'd still bring down the price of your solar farm.
    - kippinitreal 4y ago
      Could you share any links that break down the cost structure you described for solar farms? I’d love to understand that more
      - conradev 4y ago
        NREL has great data on the US: https://www.nrel.gov/docs/fy22osti/80694.pdf https://www.nrel.gov/docs/fy22osti/80694.pdf See page 19 for all varieties of “about a third”
acomjean 4y ago
Seems like great advance but a lot of questions still. Like how efficient vs silicon panels at the start? If it makes manufacturing cheaper and see through panel potential is neat. We have had solar a number of years now. With the arrival of an electric car our seeming decent sized array isn’t enough. Im not sure the current market is going well, our panel manufacturer (LG) stopped making them..
walrus01 4y ago
The best monocrystalline Si cell manufacturers now have warranty of something like 83-84% of original STC rated output after 25 years, I don't think this tech is anywhere near it yet in terms of degradation over time.
- Robotbeat 4y ago
  Why not just read the article and find out? “The results showed a device that would perform above 80 percent of its peak efficiency under continuous illumination for at least five years at an average temperature of 95 degrees Fahrenheit. Using standard conversion metrics, Loo said that’s the lab equivalent of 30 years of outdoor operation in an area like Princeton, NJ.” Sounds pretty near your benchmark!
  - Ygg2 4y ago
    35C outdoors doesn't sound like much. Illuminated areas can trivially reach +10-15C of the temperature in the shade. Edit: Article goes in more details, but they did fast age it in 100C, so it should be durable. That said, I'm still doubtful it would perform as well in real world, especially since global warming would raise air temperature. Did the 30 year models account for rise in temperature?
    - parineum 4y ago
      >Did the 30 year models account for rise in temperature? Do the Si estimates?
  - hulitu 4y ago
    Constant temperature testing ? Good luck. Where i live the difference in temperature between night and day can reach 20 C. A lot of things perform well at constant temperature.
  - walrus01 4y ago
    Needs to last through several years of accelerated aging in a test chamber to simulate 20-25 years real world before we know the actual degradation.
ZeroGravitas 4y ago
Similar to batteries, millions of minor process improvements are the real story but make poor headlines compared with radical new tech. Having said that, once/if they crack this, it opens up whole new frontiers for solar, which is already on track to utterly dominate electricity production without this factored in. The three main new areas are low cost, low weight, and as a layer on existing cells to improve efficiency. They'll expand the frontiers of where solar makes sense in several directions. Another interesting battery (and software dev) parallel. They're not just excited about the current record, but excited that they have a way to evaluate long term performance of new methods. This is very similar to the work done by Jeff Dahn at Dalhousie (in partnership with Tesla). Once they had the method to measure results they could throw things at the wall and see what would still stick after a full service life without waiting 30 years. His latest work is about how you build batteries to last 100 years. https://cleantechnica.com/2022/05/26/jeff-dahn-the-100-year-battery https://cleantechnica.com/2022/05/26/jeff-dahn-the-100-year-...
- toomuchtodo 4y ago
  90% of new generation capacity in the US annually is solar and wind. That number will only continue to climb; we're really constrained by how fast you can kick out PV panels from the fab and polysilicon supply. Adoption curve goes vertical before 2030 imho, with natural gas slowly being edged out as storage and more generation rolls out. Coal is dead, nuclear would be without subsidies (which we should provide as long as the reactors can run safely until an extended EOL date).
  - throwaway894345 4y ago
    This is exciting stuff, but I’m still confused about how we solve for base load generation with solar and wind? What do you do when an entire region is sunless and windless for a week? My impression is that you can’t simply bring enough power from neighboring regions to meet demand. So do we have really big storage systems and/or massive overprovisioning? And if so, how does solar and wind compete with those costs factored in?
    - 4y ago
asymmetric 4y ago
> the Theodora D. ’78 and William H. Walton III ’74 Professor in Engineering These prefixes (for money) are getting out of hand
- zirror 4y ago
  Could you elaborate? I didn't understand what this even means. Did someone named "Theodora D." in 1978 donate alot of money and now the chairs are named after them?
  - wrycoder 4y ago
    Theodora’s last name is Walton and she graduated in 1978. She is married to the other Walton. The “Walton Professor of Engineering” would have been sufficient. Lots of money apparently buys lots of wordsmithing.
bawolff 4y ago
Yeah, but have they fixed the lead/toxic material problem
- jonnycomputer 4y ago
  ? could you expand on this comment?
  - pfdietz 4y ago
    The perovskite materials contain some metal ions. In some varieties, it's lead ions (others use cesium or tin ions.)
    - zbrozek 4y ago
      Though aren't these extremely thin films, such that the amount of material used is very small?
tigerlily 4y ago
> …researchers returned to their labs to tend to their experiments in carefully coordinated shifts, Zhao noticed something odd in the data. One set of the devices still seemed to be operating near its peak efficiency. Ahh, this method of discovery never gets old <3
NickRandom 4y ago
Although I can't find the link to an article I read about the early development of PV cells using a purple glass filter I did manage to dig out a research paper that provides more or less the same regarding colored filters https://web.archive.org/web/20220709132444/https://reader.elsevier.com/reader/sd/pii/S2214157X20303439?token=D712E898BFAB825FEF279C6C0AAFA3E60C4F6E537E003477BE578C2F76C138BCB8D65EE02F66AD7535CE4C79C2150B9F&originRegion=eu-west-1&originCreation=20220709132338 https://web.archive.org/web/20220709132444/https://reader.el... https://reader.elsevier.com/reader/sd/pii/S2214157X20303439?token=D712E898BFAB825FEF279C6C0AAFA3E60C4F6E537E003477BE578C2F76C138BCB8D65EE02F66AD7535CE4C79C2150B9F&originRegion=eu-west-1&originCreation=20220709132338 https://reader.elsevier.com/reader/sd/pii/S2214157X20303439?...

150 points by clouddrover 4y ago

"Average temperature" doesn't mean much. Deterioration is concentrated in the periods of peak temperature, non-linearly. It makes it hard to express practical longevity with averages. You might be able to say something like, "it should last X years if it stays below T2, and spends no more than Y hours between T1 and T2." Then, actual lifetime at T2 provides an upper bound on that. They don't say in the article, but perovskites are amenable to stacking with different wavelength capture characteristics, so can be built to capture above 40% of incident solar energy. That, and being radically cheaper to produce than Si cells, means there is great interest in this result.

smaddox 4y ago
Deterioration tends to be roughly linear on an Arrhenius scale, which is basically the scale which is exponential on the inverse temperature. This is because deterioration tends to be roughly linear on the activation energy of defects.
- ncmncm 4y ago
  What that means in practice is that hardly anything happens below a certain temperature, but it goes all to hell in a flash, above that. The "certain temperature" for any particular sample depends on what you consider "a flash". You would be pretty peeved if your solar panel went all to hell in a year, so you calibrate accordingly.
  - nicoburns 4y ago
    I guess these could be particularly well suited to colder / less sunny climates. Which could be a very good thing, given that in sunny climates existing technologies are already pretty much good enough.
    - ncmncm 4y ago
      "Floatovoltaics". The first place to put solar farms is floating on reservoirs and other calm water.
    - ncmncm 4y ago
      Deserts are just a really dumb place to put solar farms. But that is where ignorant investors want them, so that is where they are going, for now.

tomahunt 4y ago

https://www.science.org/doi/10.1126/science.abn5679 https://www.science.org/doi/10.1126/science.abn5679 link to the article in Science

dmix 4y ago

Wikipedia has a neat graph comparing the efficiency of different types of solar cells including perovskite, which shows how fast it's increasing compared to silicon/multi-junction. https://upload.wikimedia.org/wikipedia/commons/thumb/3/35/Best_Research-Cell_Efficiencies.png/1280px-Best_Research-Cell_Efficiencies.png https://upload.wikimedia.org/wikipedia/commons/thumb/3/35/Be...

jopsen 4y ago
Isn't it more cost than efficiency that matters?
- kirrent 4y ago
  Not really. Materials and manufacturing costs for perovskite probably already beat silicon if anyone scaled up. It's efficiency and durability we're waiting on. If you look at pricing a large solar farm you'll be surprised that the cost of modules is probably only a third of the total cost but when you reflect on land, cabling, inverters, grid connection, and installation then it's not really surprising. Given these costs, you've got an obvious incentive to buy higher efficiency modules for a small premium for savings on installation, mounting, and land use and a disincentive to buy cheap and not yet high efficiency perovskites. This is why companies like OxfordPV are targeting tandem perovskite-silicon cells with efficiencies at 30%. Even if this costs 1.5 times as much as a 20% efficient silicon cell, it'd still bring down the price of your solar farm.
  - kippinitreal 4y ago
    Could you share any links that break down the cost structure you described for solar farms? I’d love to understand that more
    - conradev 4y ago
      NREL has great data on the US: https://www.nrel.gov/docs/fy22osti/80694.pdf https://www.nrel.gov/docs/fy22osti/80694.pdf See page 19 for all varieties of “about a third”

acomjean 4y ago

Seems like great advance but a lot of questions still. Like how efficient vs silicon panels at the start? If it makes manufacturing cheaper and see through panel potential is neat. We have had solar a number of years now. With the arrival of an electric car our seeming decent sized array isn’t enough. Im not sure the current market is going well, our panel manufacturer (LG) stopped making them..

walrus01 4y ago

The best monocrystalline Si cell manufacturers now have warranty of something like 83-84% of original STC rated output after 25 years, I don't think this tech is anywhere near it yet in terms of degradation over time.

Robotbeat 4y ago
Why not just read the article and find out? “The results showed a device that would perform above 80 percent of its peak efficiency under continuous illumination for at least five years at an average temperature of 95 degrees Fahrenheit. Using standard conversion metrics, Loo said that’s the lab equivalent of 30 years of outdoor operation in an area like Princeton, NJ.” Sounds pretty near your benchmark!
- Ygg2 4y ago
  35C outdoors doesn't sound like much. Illuminated areas can trivially reach +10-15C of the temperature in the shade. Edit: Article goes in more details, but they did fast age it in 100C, so it should be durable. That said, I'm still doubtful it would perform as well in real world, especially since global warming would raise air temperature. Did the 30 year models account for rise in temperature?
  - parineum 4y ago
    >Did the 30 year models account for rise in temperature? Do the Si estimates?
- hulitu 4y ago
  Constant temperature testing ? Good luck. Where i live the difference in temperature between night and day can reach 20 C. A lot of things perform well at constant temperature.
- walrus01 4y ago
  Needs to last through several years of accelerated aging in a test chamber to simulate 20-25 years real world before we know the actual degradation.

ZeroGravitas 4y ago

Similar to batteries, millions of minor process improvements are the real story but make poor headlines compared with radical new tech. Having said that, once/if they crack this, it opens up whole new frontiers for solar, which is already on track to utterly dominate electricity production without this factored in. The three main new areas are low cost, low weight, and as a layer on existing cells to improve efficiency. They'll expand the frontiers of where solar makes sense in several directions. Another interesting battery (and software dev) parallel. They're not just excited about the current record, but excited that they have a way to evaluate long term performance of new methods. This is very similar to the work done by Jeff Dahn at Dalhousie (in partnership with Tesla). Once they had the method to measure results they could throw things at the wall and see what would still stick after a full service life without waiting 30 years. His latest work is about how you build batteries to last 100 years. https://cleantechnica.com/2022/05/26/jeff-dahn-the-100-year-battery https://cleantechnica.com/2022/05/26/jeff-dahn-the-100-year-...

toomuchtodo 4y ago
90% of new generation capacity in the US annually is solar and wind. That number will only continue to climb; we're really constrained by how fast you can kick out PV panels from the fab and polysilicon supply. Adoption curve goes vertical before 2030 imho, with natural gas slowly being edged out as storage and more generation rolls out. Coal is dead, nuclear would be without subsidies (which we should provide as long as the reactors can run safely until an extended EOL date).
- throwaway894345 4y ago
  This is exciting stuff, but I’m still confused about how we solve for base load generation with solar and wind? What do you do when an entire region is sunless and windless for a week? My impression is that you can’t simply bring enough power from neighboring regions to meet demand. So do we have really big storage systems and/or massive overprovisioning? And if so, how does solar and wind compete with those costs factored in?
  - 4y ago

asymmetric 4y ago

> the Theodora D. ’78 and William H. Walton III ’74 Professor in Engineering These prefixes (for money) are getting out of hand

zirror 4y ago
Could you elaborate? I didn't understand what this even means. Did someone named "Theodora D." in 1978 donate alot of money and now the chairs are named after them?
- wrycoder 4y ago
  Theodora’s last name is Walton and she graduated in 1978. She is married to the other Walton. The “Walton Professor of Engineering” would have been sufficient. Lots of money apparently buys lots of wordsmithing.

bawolff 4y ago

Yeah, but have they fixed the lead/toxic material problem

jonnycomputer 4y ago
? could you expand on this comment?
- pfdietz 4y ago
  The perovskite materials contain some metal ions. In some varieties, it's lead ions (others use cesium or tin ions.)
  - zbrozek 4y ago
    Though aren't these extremely thin films, such that the amount of material used is very small?

tigerlily 4y ago

> …researchers returned to their labs to tend to their experiments in carefully coordinated shifts, Zhao noticed something odd in the data. One set of the devices still seemed to be operating near its peak efficiency. Ahh, this method of discovery never gets old <3

NickRandom 4y ago

Although I can't find the link to an article I read about the early development of PV cells using a purple glass filter I did manage to dig out a research paper that provides more or less the same regarding colored filters https://web.archive.org/web/20220709132444/https://reader.elsevier.com/reader/sd/pii/S2214157X20303439?token=D712E898BFAB825FEF279C6C0AAFA3E60C4F6E537E003477BE578C2F76C138BCB8D65EE02F66AD7535CE4C79C2150B9F&originRegion=eu-west-1&originCreation=20220709132338 https://web.archive.org/web/20220709132444/https://reader.el... https://reader.elsevier.com/reader/sd/pii/S2214157X20303439?token=D712E898BFAB825FEF279C6C0AAFA3E60C4F6E537E003477BE578C2F76C138BCB8D65EE02F66AD7535CE4C79C2150B9F&originRegion=eu-west-1&originCreation=20220709132338 https://reader.elsevier.com/reader/sd/pii/S2214157X20303439?...