3 ms·

Setting aside the veracity of the article for a moment ... ... a key factor in industrial-scale silicon production (microchips, PV, or otherwise) is the purity

Setting aside the veracity of the article for a moment ... ... a key factor in industrial-scale silicon production (microchips, PV, or otherwise) is the purity of the raw material. It may well be that high grade silica is fundamentally scarce, or that access to it may be restricted or limited by other factors. Sand theft is an issue that's emerged as a concern, largely as it's a huge industrial input, particularly in concrete, and the most suitable sands are found in riverbeds, rather than either deserts or beaches, both of which are far more abundant. Discussed five years ago on HN here: <https://news.ycombinator.com/item?id=21758301 https://news.ycombinator.com/item?id=21758301> There are abundant materials whose useful stock or rates of production are limited by other factors. Nitrogen fertiliser is a canonical example: the raw element is abundant in air (70% of the atmosphere), but its fixation is extremely energy intensive, relying most keenly on natural gas supplies and prices. Ammonia and ammonia-based product costs (fertilisers, cleaning supplies, explosives) tend to fluctuate strongly with natural gas prices, despite the underlying abundance of nitrogen itself.

h0l0cube 2y ago
I don't refute this point, but the alternative materials would have to be more available than high-grade silicon for it to matter.
- dredmorbius 2y ago
  Pretty much agreed on that, and I haven't followed the perovskites discussion closely enough to know where it falls on that basis. I have been aware of the technology vaguely for a decade or so. The raw material itself is CaTiO₃,[1] which are three fairly abundant elements, though how that compares with sufficiently pure silica isn't clear to me. My general feeling (from what I've read and experience with other strongly-promoted energy alternatives) is that applications are more likely to be niche, where conditions specifically favour perovskites' specific advantages. Those seem to be greater incident sunlight conversion efficiencies (~30% vs. 15--20% more typical of silicon PV, which is often less significant than might at first be apparent), and potentially lower manufacturing costs. As is noted in the discussion, panel cost itself is increasingly dwarfed by less-fungible labour and infrastructure costs (e.g., physical support, see <https://news.ycombinator.com/item?id=43135371 https://news.ycombinator.com/item?id=43135371>). Worth keeping a finger on the pulse, but not something I'm getting particularly excited about. Pace of development has lagged hype/expectations over the decade or so I've been hearing about the tech. ________________________________ Notes: 1. See: <https://en.wikipedia.org/wiki/Perovskite_(structure) https://en.wikipedia.org/wiki/Perovskite_(structure)>
  - h0l0cube 2y ago
    Yeah, I think the major problem with them is lifetime. Oxford PV's tandem panels are supposed to meet or exceed typical panel lifetimes, but being tandem, they also use silicon.
    - dredmorbius 2y ago
      That would be another factor. PV tech to date has a presumed lifespan of about 20 years, at which point it's both typically degraded to 80% of nameplate functionality and cheaper to replace with newer, more functional, components. That said, extending life by a factor of 50--100% could be a game-changer, particularly where PV is a major (or majority) factor of electricity generation or all energy inputs. At that point, needing to replace 5% of all installed capacity every year becomes its own daunting task. Reducing that to 2.5% would be a tremendous win if that could be achieved at a competitive cost. I've done some reading on both how and why panels fail and durability/lifespan testing (much of this comes out of NREL in the US), and as it's due to multiple degradation pathways achieving greater lifespans isn't a trivial task, and the 20-year benchmark is itself quite dependent on specific experiences. E.g., a heavy hailstorm won't much care how old your panels are, but on average the expected incidence of such events is factored into the 20-year life expectancy (along with other similarly variable factors).

h0l0cube 2y ago

I don't refute this point, but the alternative materials would have to be more available than high-grade silicon for it to matter.

dredmorbius 2y ago
Pretty much agreed on that, and I haven't followed the perovskites discussion closely enough to know where it falls on that basis. I have been aware of the technology vaguely for a decade or so. The raw material itself is CaTiO₃,[1] which are three fairly abundant elements, though how that compares with sufficiently pure silica isn't clear to me. My general feeling (from what I've read and experience with other strongly-promoted energy alternatives) is that applications are more likely to be niche, where conditions specifically favour perovskites' specific advantages. Those seem to be greater incident sunlight conversion efficiencies (~30% vs. 15--20% more typical of silicon PV, which is often less significant than might at first be apparent), and potentially lower manufacturing costs. As is noted in the discussion, panel cost itself is increasingly dwarfed by less-fungible labour and infrastructure costs (e.g., physical support, see <https://news.ycombinator.com/item?id=43135371 https://news.ycombinator.com/item?id=43135371>). Worth keeping a finger on the pulse, but not something I'm getting particularly excited about. Pace of development has lagged hype/expectations over the decade or so I've been hearing about the tech. ________________________________ Notes: 1. See: <https://en.wikipedia.org/wiki/Perovskite_(structure) https://en.wikipedia.org/wiki/Perovskite_(structure)>
- h0l0cube 2y ago
  Yeah, I think the major problem with them is lifetime. Oxford PV's tandem panels are supposed to meet or exceed typical panel lifetimes, but being tandem, they also use silicon.
  - dredmorbius 2y ago
    That would be another factor. PV tech to date has a presumed lifespan of about 20 years, at which point it's both typically degraded to 80% of nameplate functionality and cheaper to replace with newer, more functional, components. That said, extending life by a factor of 50--100% could be a game-changer, particularly where PV is a major (or majority) factor of electricity generation or all energy inputs. At that point, needing to replace 5% of all installed capacity every year becomes its own daunting task. Reducing that to 2.5% would be a tremendous win if that could be achieved at a competitive cost. I've done some reading on both how and why panels fail and durability/lifespan testing (much of this comes out of NREL in the US), and as it's due to multiple degradation pathways achieving greater lifespans isn't a trivial task, and the 20-year benchmark is itself quite dependent on specific experiences. E.g., a heavy hailstorm won't much care how old your panels are, but on average the expected incidence of such events is factored into the 20-year life expectancy (along with other similarly variable factors).