1/8" Clear Green Edge Cast Acrylic Sheet (Plexiglass Size:48" x 96")

> Well there are a lot of confused arguments being made.

There certainly are! If you think more carefully and pay more attention to what I'm saying, you can stop making them.

Again, I'm well aware greenhouses can be economical; that's my whole point --- greenhouses are more economical than indoor cultivation! It's unfortunate that your annoyance with my tone is leading you to put so much effort into rebutting things I'm not saying in the first place.

The part where the economics aren't going to work out is where we switch from gathering sunlight with glass or plastic windows to gathering sunlight with solar panels and reproducing it through LEDs, losing 97% of it in the process. (That's assuming 20%-efficient LEDs.) Even at the 48.7% LED efficiency in the article you cite, you're still losing 92% of the sunlight in this process. The only way that can work out is if 13 square meters of solar panel come out to be cheaper than one square meter of glass or plastic skylight in the greenhouse, or if the artificial lighting is used as a supplementary source of light (as you say has been done for decades), or if the power for the LEDs comes from something even cheaper than low-cost solar panels. That last is definitely a possibility in places like Scotland, Norway, or Iceland, where sunlight is scarce and other power sources are available.

This is analogous to discussions about desalinated water. With modern reverse-osmosis plants, desalinated water is astoundingly cheap: US$0.58/kℓ as of 2016 at Israel's new Sorek plant, for example, so you can fulfill the drinking, cooking, and bathing needs of a household for about US$7 a year. But it's no longer cheap when you consider water use on an agricultural scale, where (in medieval units) that's US$715 per acre foot, two or three times the price farmers in the US commonly pay.

So, yes, energy is very cheap, and will get cheaper. But energy on an agricultural scale is still not cheap, and solar panels will never make energy on an agricultural scale cheap. You need a cheaper source of power.

As for the high-efficiency solar panels, yes, there have been 21%-efficient solar panels for a long time, since the 1970s in fact. The multijunction solar cells we used on our satellites were 30% efficient, if memory serves, and multijunction cells in the lab have exceeded 40%. As the article you linked explains, the 22%-efficient panels are monocrystalline solar panels marketed as "high efficiency"; you can see in https://www.solarserver.de/service-tools/photovoltaik-preisi... that they consistently cost about 60% more per watt, and therefore 120% more per square meter, than the 16%-efficient "low-cost" polycrystalline solar panels that are universally used in new solar power stations. So high-efficiency solar panels don't turn out to be a cheaper source of power than low-cost solar panels, any more than diesel fuel is, except, say, in Antarctica. High-efficiency solar panels are what you use when paying twice as much per square meter is okay as long as you get 60% more energy out of your very limited roof.

I understand how, since you lack that deeper knowledge about the context, it sounds to you like I'm being unreasonably assertive and dismissive. But in fact my assertivity and dismissal are based on understanding aspects of the situation you just weren't aware of.

The article on super-efficient white LEDs https://www.dial.de/en/blog/article/efficiency-of-ledsthe-hi... is very promising! I very much appreciate you bringing it to my attention. The calculation method they outline is solid. Unfortunately, their results are not reproducible or falsifiable, because they don't say which LEDs have these remarkably high efficiencies. (They do say that most illumination LEDs are less efficient, but of course that's no bar to LED-illuminated subterranean hothouses --- you could just not use the inefficient LEDs.) Lacking information on the particular brand and part numbers as well as testing conditions (temperature and current, mostly), it's impossible to say whether their astonishing results indicate that these near-50% efficiencies are reliably achievable, whether they are achievable but only under conditions that are otherwise unfavorable in some way (water-cooling the LEDs, for example), or whether they simply amount to some kind of measurement or calculation error. The results I've seen on manufacturer datasheets and other reproducible, falsifiable sources unfortunately do not reach this level of performance.

Nevertheless, even if those surprising claims about LEDs do pan out, it doesn't undermine my core claim, which is that a square meter of skylight is going to grow more vegetables than a square meter of solar panels, and that the solar panels are not, in the foreseeable future, going to be sufficiently cheaper than glass or clear plastic to make this an economic win. Low-cost solar panels currently cost about €30/m² wholesale, plastic wrap on Amazon costs about €0.001/m², and the solar panels provide one-sixth the energy, one-thirteenth after going through a 50%-efficient LED, or one-thirtieth after going through a 20%-efficient LED. If the LEDs were 100% efficient and there were no transmission and storage losses, the cost of energy would still need to fall by a factor of a hundred and eighty thousand for the costs to become equal. That's 17 to 18 doublings; expect it around 2050 to 2060.

Beating durable plexiglass should come sooner: https://www.amazon.com/Cast-FS-Plexi-Glass-size/dp/B01LYK5RD... is US$300 (€270) for 14.9 m² of 3.2-mm-thick acrylic, which is US$20/m² (€18/m²). That's only a factor of 10.4 cheaper than (wholesale) low-cost solar panels, or a factor of 50 cheaper than low-cost solar panels powering 20%-efficient LEDs; solar panels are on track to beat that in a decade or two. It'll be interesting to see what they find to protect the solar panels from hail, though. Maybe it'll work for greenhouses too.