Hacker News Reader

Solar panels can be more efficient, but if your comparison is between “sunlight -> plants” and “sunlight -> solar panels -> electric lights -> plants”, you have to include the actual efficiency of e.g. photovoltaics and LEDs in your calculations.

Photovoltaics these days have something around 15-20% efficiency and LEDs have conversion efficiency around 50-60%. The magenta grow lamps are colored for more efficient use by plants, and you can pack more plants in a smaller space, but at that point you’re trying to offset energy losses on the order of 90%.

Yes, and I updated my comment just as you were posting this to mention losses due to reconversion. However, also keep in mind that even large efficiency losses can still lean in favor of photovoltaics/LEDs, because plants only use about 10% of the sun's energy in the first place.

The problem with talking about photosynthetic efficiency is that there are different endpoints you can talk about. The more efficient plants (C4 plants like sugarcane and maize) have something like 4% efficiency converting sunlight to biomass, but they actually absorb a 53% of the incoming light based on spectrum, and lose about 24% of the energy because photons with shorter wavelengths have excess energy which the plants cannot use. We’re not interested in the 4% figure, we’re interested in the 53% and 24% figure because they represent the part of the process that we can change.

Doing the math, that’s around 59% loss which you could mitigate by using LEDs that produce the correct spectrum—but solar panels and LEDs have 90% losses, so you’re noticeably worse off.

It’s worth remembering that the reason why plants only absorb certain parts of the spectrum is the same reason why photovoltaic panels only absorb certain parts of the spectrum—in both cases, you are using light to move electrons, and these processes only capture energy that corresponds to the underlying band gap. Light with shorter wavelengths has additional energy which is wasted, both for photovoltaics and for plants.

You can increase the efficiency by creating multijunction solar panels, which results in multiple band gaps. For most applications, these aren’t cost-effective. If I remember correctly, plants are also “multi-junction”, which explains why they are so efficient.

> It’s worth remembering that the reason why plants only absorb certain parts of the spectrum is the same reason why photovoltaic panels only absorb certain parts of the spectrum

Plants are green because they value light consistency instead of total energy. Green light has too many peaks and valleys and can overload the photosynthesis systems so they reflect a lot of it.

It's the "renewables without batteries" problem only in biology.

> Green light has too many peaks and valleys and can overload the photosynthesis systems so they reflect a lot of it.

This doesn’t make any sense to me. Why would peaks and valleys overload something? Why would green light have more peaks and valleys?

I was a bit sloppy with the way I phrased that—what I really meant was “the reason why plants use specific quanta of light is the same reason why photovoltaics absorb specific quanta of light” but I didn’t put much thought into how worded it.

Plants absorb light near two different spectral peaks. This is not entirely dissimilar to the idea of a multijunction photovoltaic cell. The color of light between the two peaks is green.

I don’t know if this is what the previous poster was getting at, some there is some theoretical evidence that the absorption wavelengths used in photosynthesis are not tuned for maximum power input but for stable power input: https://arxiv.org/abs/1912.12281

If this is the case it would be a motivation that is fundamentally very different than that of a multijunction solar cell.

I've read that paper, I didn't realize the comment was referring to that.

> If this is the case it would be a motivation that is fundamentally very different than that of a multijunction solar cell.

Solar cells are designed by humans with human motivations, and plants are "designed" by evolutionary processes which lack motivation entirely. Yet, in spite of this, there are astonishing similarities between the limitations of solar cells and the limitations of plants. Isn't that fascinating, that processes with such disparate origins have such similarities?

I don’t see any meaningful similarities beyond ‘they absorb light at more than one frequency’. It doesn’t seem particularly fascinating to me; it’s patently obvious that if you absorb at two frequencies you should absorb at two different frequencies.

> I don’t see any meaningful similarities beyond ‘they absorb light at more than one frequency’.

Ah, that's definitely not what I'm talking about. Look at the relationship between efficiency and wavelength for plants and photovoltaics... the similarity is clear.

Both photovoltaics and plants are capable of absorbing specific amounts of energy from incoming photons but not other amounts. Any excess energy becomes waste heat. The reason is because in both systems, the incoming light is used to move electrons from one state to another; the energy required to do this must usually come from one photon.

Not all solar power systems work this way.

Just for comparison's sake, aren't plants/photosynthesis something like 5% efficient?

I think that's correct, but that difference seems to be within the error margins for this calculation, and some other loss (e.g., transmission, storage, etc) could eat any hypothetical advantage that the vertical farming position might've enjoyed.

Blurple and magenta LEDs seem to be on their way out last time I looked into growing lights. It is white (or white appearing) LEDs now, which are getting 190+ lumens per watt. The light spectrum is still tuned for plant growth efficiency, but it is a broader spectrum per LED rather than more specific spectrums from specific diodes.

It is worth noting too that these newer white LEDs are spectrum efficient enough that people growing with them have had to turn up the overall room temperatures to compensate for the drop in leaf temperatures created by "waste" wavelengths that the sun and old sodium bulbs use to provide.

190+ lm/W is still less than 50% efficiency. I don't understand how a white lamp can be tuned for plant growth efficiency, could you elaborate?

50% efficiency is pretty massive for a light or most anything really. And you can get higher if you underdrive them.

Just because a light appears white doesn't mean the spectrum is evenly distributed though, generally if you graph the color spectrum of these lights green is in a huge dip. If you go to page 13 of this data sheet they show show spectrum graphs. https://cdn.samsung.com/led/file/resource/2020/09/Data_Sheet... All of which appear white in real life. They may be more cool blue white or warmer reddish white, but white nonetheless. Especially with the intensity that they shine.

> Note how most plants are green, which is to say they're content with reflecting the most energy-dense range of the visible spectrum.

Photosynthesis is very complex, and there are reasons why plants shed about 10% of green light: https://science.sciencemag.org/content/368/6498/1490 . It's not that evolution was too dumb to discover the simple fact of green light being more energy dense, it's that there's other more important constraints happening at the molecular level in photosynthesis.

Plants desire more stability in energy output from photosynthesis (e.g., clouds moving overhead) at a lower level, rather than attaining maximum output. Trying to always maximize output means you'll have high fluctuations in energy output.

There is this recent-ish paper (DOI: 10.1126/science.aba6630) on why it might be, that plants forego the peak of the solar spectrum. Essentially it boils down to being able to regulate the photochemistry of photosynthesis. If it were centered on the peak of the spectrum there's not a lot of regulation possible by means of shifting the reaction energy levels around.

By placing the light absorbing parts of photosynthesis on the slopes of the spectrum, by mere adjustment of the energy levels the reaction undergoes it can shift its activity to parts of the spectrum with more or less light intensity.

Even more to the point- solar panels live quite easily on roofs, deserts, highway medians. Quite a bit tougher to put functioning agriculture in these places!

Green roofs are a thing, though. https://en.wikipedia.org/wiki/Green_roof

A green roof is pretty. It has plenty of upside. It isnt a productive farm which is the appropriate comparison here

>It isnt a productive farm which is the appropriate comparison here

Why would 'vertical farming' be more productive then?

Because productive farms require lots of expensive maintenance/equipment. A vertical farm lets you optimize for that use case since its the only use of the infrastructure and you can align everything towards maximum use of those scarce resources.

Random roofs don't optimize for that usage pattern, and so you need to use them in an application where the primary cost is the install cost rather than an upkeep cost. Solar panels are a perfect fit. Farms are a terrible fit.

Depending on the geography and building design, you can carry light to the plants via optic fiber. They act as a light pipe/optical waveguide. No need for any photovoltaic solar panels. However this design necessitates a skyscraper in the middle of the desert with nothing else blocking line of sight. Great if you are in the Middle East. Not so great for New York or Seattle. For the latter cities, a permanent barge on the Hudson/Puget with a fiber optic connection could be a solution (the losses would be great and it may not be much cheaper than using electricity, geography and land costs will have to be carefully accounted for).

Here are some simplified designs for home use: https://www.lowes.com/pl/Tubular-skylights-Skylights-accesso...

Some commercial suppliers: http://www.huvco.com/

The technology is very cut and dry. If you are a well funded startup, it may be more economical to acquire an optic fiber skylight manufacturer instead of ordering OEM.

Are there any systems that can actually capture close to 100% of sunlight falling on the roof and carry it somewhere else? All the systems you link are only for replacing illumination for humans. They don't provide anywhere near light flux sufficient for plants.

Say a plant needs 1/10th of actual full sun during the day, could you theoretically remove the electricity step entirely but "just" (putting "just" in quotes because I'm sure it wouldn't be easy) having some sort of fancy mirror setup on a 10 storey building to send 1/10th of the sunlight hitting the roof to each floor?

Mirrors take up space too.

Might be significantly higher efficiency overall though vs. putting solar panels on the same roof?

I guess the devil is in the details but I can't see how mirrors will be positioned to support vertical farming. It's be interesting to see some viable ideas (if those exist).

Thinking aloud here; perhaps we could somehow use optical fibers to bend the light from rooftop to vertical farm?

It's beyond my pay grade. I have no idea if 'optical fibers' are an option here. I suspect 'not' only because it hasn't been done already and feels like it if worked for vertical farming, it would also work to supplement office lighting so there is some monetary incentive to get such a system implemented in some way.

The other problem is that you still have the same issue. That is, for every square cm of sunlight you collect the best you could do is direct that light to a square cm of a plant. That is, you can't just increase surface area of collected light so to provide sunlight for vertically stacked plants, you need the equivalent horizontal surface area to support those.

But I have no idea.

Close, it's 1/10th of the spectrum not 1/10th of the intensity.

The 'fancy mirrors' required would simply be a prism. You'd separate out the red and blue light and aim it at the plants, and send the green and infrared light to solar cells, to be used for running red and blue LEDs. (Or you could use flourescence if you had something that glowed red or blue when exposed to green or IR light).

Check out this spectral plot for chlorophyll-a and -b:

http://hyperphysics.phy-astr.gsu.edu/hbase/Biology/ligabs.ht...

and compare it to this plot of the solar irradiance at Earth's surface:

http://hyperphysics.phy-astr.gsu.edu/hbase/vision/solirrad.h...

There are peaks in the clorophyl absorption spectrum at 400-500 and 600-700 nm (blue to UV and red), but the sunlight provides energy in an atmosphere-attenuated blackbody curve everywhere from 300-1000nm.

In theory, by providing illumination with just 10% the energy of the sunlight's full spectrum in a narrow band from 420-430nm where photosynthesis is most efficient, you could have plants receive the same amount of energy.

Unfortunately, solar cells have the exact same problem: Just as the proteins in chlorophyll only use certain spectral energies, so too the semiconductors only make efficient use of certain energies. Multi-bandgap solar cells can help:

https://en.wikipedia.org/wiki/Multi-junction_solar_cell#/med...

but in both chemical and electrical solar energy extraction, you're working with a particular photon energy and there's going to be waste.

Do the plants actually need the full strength of that red and blue light for optimal growth though? I thought the comment above mine was implying that they don't.

So my proposition was that if we started with, say, midday sun, maybe we could divide that up (in brightness, not just by spectrum) to cover a larger area (multiple floors in this case).

Are there materials that fluoresce at a shorter wavelength than what they absorb? I was under the impression that it was a one-way thing (absorb short wavelength and emit longer wavelength).

Yes. Some materials absorb multiple photons for each one they emit and thus have enough energy to emit at shorter wavelengths. Cesium vapor is one (though probably not a good choice for this application!).

maybe if the plants are fixed on a very tall pole, and spaced some meters from each other, because of the sunrays inclination, all the plants on the pole can get sunlight. this will not save space in a big farmland setup, because then the poles need to be spaced from each other in proportion to their height, or they will cover the sun from the neighbour pole, bit this can make small isolated empty spaces useful for growing a lot of plants.

Plants use green light: https://www.canr.msu.edu/news/green_light_is_it_important_fo...

In fact the idea of light frequency restriction is dying a death in LED grow lights, sun-like full spectrum with deep red and UV ranges have proven to be superior to specific ranges (blurple).

Light being a bottleneck isn't very meaningful as plant growth is adjusted by a bunch of levers, e.g. increase respiration and nutrients and more light energy can be utilized - obviously there's limiting factors in plant biology.

"Light, its what plants crave!"