One major problem with living underground is you lose all of that wonderful energy provided by the sun. This energy is useful for many things including growing plants for food. A window is effectively a very short pipe for sunlight to enter a room in a normal house. Windows clearly allow enough of the right kind of light to enter to allow plants to grow. Could someone in an underground bunker use a bunch of fiber optic cables to pipe enough light into a room to allow plants to grow? Assuming that's possible would it be possible to use the same technique to power a solar oven?

The basic setup would be 144 strand fiber cables ranging from 100 to 300ft long attached to the bunker ceiling to provide overhead lighting. The cables would poke out from the ground to collect sunlight. Some cables would run up trees others would peer out from crevices in rocks, and some would mixed in with low vegetation like grass. The distribution of the cables would be randomized and not point directly back to the bunker in a straight line to reduce the chances of the bunker being found. The cables would have a fish eye lens attached to each end like a borescope. The idea wouldn't be to form a cohesive image but just to collect and pipe the light.

The implementation details described above can be adapted if needed. Some of my concerns relate to the general feasibility of using fiber optic cables to pipe light. Are readily available commercial fiber optic cables tuned to a certain frequency range of light and as a result severely attenuate or filter out the wavelengths I need? Would there be too much loss in the transmission to get the light output I'd need? Some quick googling showed a fiber is 50 microns, and a 144 fiber bundle is a bit under 3/4 of an inch in diameter. That means my "window" is going to have to be roughly 2.5 times bigger than a real window in normal house. This seems workable but I don't know if the losses are going to be so large that the ratio gets significantly worse. That being said I don't know how many watts typical food crops really require out of the roughly 1000 watts per square meter the sun provides (in ideal conditions). I also don't know if I'm missing something obvious and fundamental....

So to reiterate my question could commercially available fiber optic cables pipe enough light under ground to allow food crops to grow? If they can pipe enough light to grow food could they power a solar oven?

  • 1
    $\begingroup$ Depends on how much you want to pay, and how much area you want to dedicate to light capture. The sunlight falling on one square meter of land can be captured and directed underground to illuminate one square meter of substrate. So if you want one hectare of underground agricultural substrate you must capture light from one hectare of land -- this would be easily noticeable. (And you don't want to use modern single-mode fibers, which are tuned to a narrow frequency band, you want old-style multimode fibers or dirt-cheap plastic light conductors.) $\endgroup$
    – AlexP
    Nov 19, 2018 at 19:10
  • $\begingroup$ For example look here. Google for "light conductors" illumination. Your mental image is correct, but those are the lowest of the low end of light conductors. $\endgroup$
    – AlexP
    Nov 19, 2018 at 19:48
  • $\begingroup$ First thing that springs to my mind: The topside ends of the cables would need constant cleaning, otherwise they wouldn't be useful for long. $\endgroup$
    – Erik
    Nov 20, 2018 at 9:51

2 Answers 2



Light Pipes are an established architectural thing. Breaking it up into optical fibres (effectively many small light pipes) would probably be expensive and finicky, but presumably cost is no object.

Relevant text from the wikipedia article:

Optical fibers can also be used for daylighting. A solar lighting system based on plastic optical fibers was in development at Oak Ridge National Laboratory in 2004. The system was installed at the American Museum of Science and Energy, Tennessee, USA, in 2005, and brought to market the same year by the company Sunlight Direct.

Edit: To address issues brought up in the comments, mixed- or multi-mode fibers, while thicker than telecommunications optical fibre, are capable of transmitting many different wavelengths, which would serve to transmit at least a pseudo-daylight to the bunker.

Alternatively, light conductors are less specialized still, and are employed in buildings today

(Relevant text from the above:

The Fiber Optic Skylight by HUVCO Daylighting Solutions comprises an exterior mounted panel containing 64 computer-controlled lenses that focus sunlight into optical fibers. The fibers transport the natural light up to 60 feet (18 m) to deliver it where not previously possible.


  • 1
    $\begingroup$ Hmmmm sunlight is a very impure spectrum light, can fiber optics really convey all wavelengths? $\endgroup$
    – anon
    Nov 19, 2018 at 19:02
  • 1
    $\begingroup$ @Erik - as I said, finicky - the fisheye lenses suggested by the OP would be necessary, and the area of light acquisition would have to be substantially greater than an equivalent window, as indicated by the linked article. $\endgroup$
    – jdunlop
    Nov 19, 2018 at 19:03
  • $\begingroup$ @Erik: Not really fiber optics, but a rather more coarse form usually called "light conductors". They are readily available and much cheaper than real fiber optics. $\endgroup$
    – AlexP
    Nov 19, 2018 at 19:04
  • 1
    $\begingroup$ @jdunlop Your answer relies too heavily on the link provided. You should outline the contents of any link if they are necessary for the answer to be useful. This way if a link dies, the answer remains relevant and helpful. If this answer remains as is it may be deleted as "link only" $\endgroup$
    – James
    Nov 19, 2018 at 20:36
  • 1
    $\begingroup$ @jdunlop Answers should answer the question AND provide support/evidence of why that answer is correct. Never assume another user has the same knowledge base you have. Thanks for editing. $\endgroup$
    – James
    Nov 19, 2018 at 21:55


Crops need light. If they do not get enough light they whither and die. On the other hand there is a sturation point, so not every Watthour of light energy is reliably translated into sugars. The light needs to stay between the compensation point and the saturation point, to be effectively growing the plant.

How much light? The light saturation point, which we will try to stay near to, depends on the plant species and on their current developental phase. 2% - 35% of noon sunlight are values i found. So i'll round and say we need 10% of 40°N-summer-noon power. Noon, that is. Power from the sun will be diminished before and after noon. So whatever (sub-one) efficiency your fibers have, they will neccessarily curb the duration of the growth on each day. (Outside, the plants would still get, e.g., juuuust enough sunlight at 18:00, but with the light transmitted in fibers, another 10% is lost, and the plant ceases to effectivly photosynthesize; even worse, plants need to eat too, and the longer the stretch of no-photosynthesis, the higher the percentage of the sugars-produced during high-light-phases that they will snack on during low-light-phases)

If you want to grow a m² of crops fiber-lighted, they will either grow more slowly, or you need to bring in light from more than one m². If you go the other way, and take the light from one m² and distribute it on, say, two m², you may be able to still get some growth for a few minutes around high noon... So for every x m² of crops you need at least x *(1+lossratio) m² of fibre optics - those might get hard to camouflage topside...

Let's be generous and say that your fibre itself looses nothing, because the distance is short enough, and the sunlight for some reason needs no specific setup to couple into the fibre in all the different sun-angles, and during overcast weather (absolutely impossible) - this still leaves the ends, which introduce 0.3dB each - meaning you would only loose about 7% of light. Thus, you would not suffer that much of a hit in terms of growth duration over the day and may be able to get away with 1:1 fibre-input:crop-area; this would mean we need about 1000m² of fibre input to feed one person... If the sourrounds are not greyish grasslands of fiber-y quality, this will be absolutely impossible to camouflage.

Problems with sun-angle and no real throughput during overcast conditions means that indeed you would need x*10 or more, imo.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .