Novel idea – arrange solar panels like Nature designed it

From Slashdot, something so obvious you wonder why nobody tried it before: arrange solar panels like tree leaves for better efficiency. See the story and graphs I’ve provided below.

The experiment: linear solar array and tree array

13-Year-Old Uses Fibonacci Sequence For Solar Power Breakthrough

An anonymous reader tips news of 7th grader Aidan Dwyer, who used phyllotaxis — the way leaves are arranged on plant stems in nature — as inspiration to arrange an array of solar panels in a way that generates 20-50% more energy than a uniform, flat panel array. Aidan wrote:

“I designed and built my own test model, copying the Fibonacci pattern of an oak tree. I studied my results with the compass tool and figured out the branch angles. The pattern was about 137 degrees and the Fibonacci sequence was 2/5. Then I built a model using this pattern from PVC tubing. In place of leaves, I used PV solar panels hooked up in series that produced up to 1/2 volt, so the peak output of the model was 5 volts. The entire design copied the pattern of an oak tree as closely as possible. … The Fibonacci tree design performed better than the flat-panel model. The tree design made 20% more electricity and collected 2 1/2 more hours of sunlight during the day. But the most interesting results were in December, when the Sun was at its lowest point in the sky. The tree design made 50% more electricity, and the collection time of sunlight was up to 50% longer!”

His work earned him a Young Naturalist Award from the American Museum of Natural History and a provisional patent on the design.


Good for him! Here’s the output graphs:


newest oldest most voted
Notify of

Now that sounds like a job and research well done. Guess it doesn’t have to be done only in Ivory towers!


That is really cool.


Nice idea, but i think it will go the way of Solyndra – complicated structures are more expensive than simple ones, so the simple one wins in the marketplace (yes i know the solar market is distorted by subsidies, but within those conditions, technologies still compete against each other).

Pamela Gray

Now that’s cool.


It’s an interesting result, but I’m not sure the data presented support the conclusion of “20-50% more energy.” He measures voltage, not energy. Solar panels have to be loaded to their maximum power point (MPP) to get the most energy. The voltage x current is maximum at MPP, but voltage is well below the peak (open circuit) voltage.
With his design, each cell has a different angle to the sun. There may not be a well-defined MPP.
The point of PV arrays is to create energy. The measurements are for voltage. He’s demonstrated that a range of angles generates voltage for longer than a flat panel array with a single angle. He has not demonstrated any increase in energy harvest. There is reason to believe that the energy harvest from his array will be lower.


It is a good idea and great study.
In practice, I expect the cost of the structure would outweigh the increased output especially as panel costs continue to fall. And, it could be argued that some solar panel systems already mimic nature by tracking the sun – just like some plants.

Young, bright, minds unfettered by agendas and paradigms?


Excellent and congratulations to such a young and brilliant mind. Hope to see more good things come from our younger generations.


Oops, should have said that “The point of PV arrays is to harvest energy.” Energy can’t be created or destroyed.

richard verney

Good for him.
I had always understood that to maximise the input energy, it was necessary for solar panels to be steered much the same way as a movable satelitte dish. Indeed, I believe that there are already mechanisms on the market which allow panels to be steered so as to track the movement of the sun.


Clever and cute, but the siting of the tree seems to be influenced by the house.
Also I count 10 panels on the house and at least sixteen on the tree. Sixty percent more panels might actually produce 20% more power.


“The pattern was about 137° and the Fibbonacci sequence was 2/5” is gibberish without context. Couldn’t you maximize generation even further with panels that track the direction of light and use low input, high efficiency motors to orient themselves (like sunflowers)?

Anyone who likes Fibonacci numbers has a promising future.
Very clever idea. Needs bigger panels that can work while partially shaded – trees do a good job of intercepting nearly all the sunlight.
BTW, a great place to watch a partial solar eclipse is from a tree. The leaves make a lot of pinhole cameras, and you can control the number and size by climbing up and down.

Dan Santo

Not to be all hard-assed on a really AWESOME bit of work, but that ‘typical’ solar array in the picture really ought to be at the same height as the ‘tree’ solar array. Otherwise it will be hit by shadows significantly earlier than the ‘tree’ solar array.
OTOH, that picture may very well be only for display purposes and doesn’t reflect the actual experiment.

Jeff Carlson

great work … for static arrays that may be the way to go … a solar tracker is really the ideal but this is some great work from a 13 year old …


Clever but what happens if the flat panel can be moved and trace the sun? This is only good for static installations?


I can see at least 16 panels on the “tree” and 10 panels on “control”… no wonder.

Simply brilliant. Well done Aidan.

Eric Anderson

Very interesting. When I bought panels on my house a couple of years ago it was like pulling teeth to get the contractors to do anything other than all panels on the same roof, at the same angle, facing the same direction. I wanted to add a couple more panels on perpindicular roof slopes, but they all insisted that they couldn’t do it because of limitations of how the inverter dealt with incoming load (couldn’t handle wide variance between the two circuits). Pretty frustrating and poor design, if true, although I never could ascertain whether it was a true design constraint or if they just had their “warranty” installation instructions they couldn’t deviate from.
One question about the teen’s experiment above: Is it just me, or does it look like the tree version has way more panels than the flat version? If so, was that taken into account?

John F. Hultquist

The backing wall of the building has an off-white surface with a strong albedo. The “oak-leaf-like” solar panels appear as though they might benefit from this. The “linear” (flat) solar array appears not to benefit from the reflected light.
As an aside, the WSJ recently ran an article on “bottle trees” and if you search on that phrase you can find many hits. Blue is a preferred color – keeps evil spirits away. It was noted that some folks find such structures unattractive.


What a brilliant young man. It makes sense that nature is using its own type of fractal antenna since it is collecting electromagnetic radiation.
“A fractal antenna is an antenna that uses a fractal, self-similar design to maximize the length, or increase the perimeter (on inside sections or the outer structure), of material that can receive or transmit electromagnetic radiation within a given total surface area or volume.”


While it does make sense to me in principal, I wonder what effect (if any) the sunlight reflecting off the white paint in the picture above would have had on the experiment. The tree-arranged panels are closer to it than the traditionally arranged.


This really doesn’t seem like it should make that big of a difference, but it is hard to argue with data (as it should be).

Robert L

A cute idea and well done for a 13 year old. Uses panels in a 3 dimensional tree-like rather than normal planar arrangement. But produces LESS power per panel as they are occasionally shaded by other panels and structure. So totally useless given that panal costs are the primary cost in PV installations and economics (the land area is almost free in comparison).

Gary Krause

Nature spends millions of years evolving efficiencies. We just need to get past our selves.

Well done Adian, I don’t know the lads selection criteria, oak seems reasonable, but what of other plants that are known to specifically seek the maximum solar input…..say a sunflower for example?

Richard S Courtney

A clever lad! well done.
[sarc on] Now, all we need to do is make the Sun shine at night and solar electric power would be useful. [sarc off]


In the picture I see the tree cells next to a white wall, so some reflected light may benefit the design, especially when the sun is off of the optimum angle. Where as the conventional arrangement , tilted 45 degrees will not benefit from the wall reflections.
There must be more to the experiment than this.


Great idea.
I’m sure it’s been tried, but wouldn’t panels that rotate with the sun and self-correct their angles work even better? I understand it would consume energy, and I’m sure it’s been thought of before.
Just another non-scientist thinking out loud.


I’m going to contest the idea that any design that mimics nature is automatically superior. I’m not prepared to argue in depth, but it seems to me to put limits on the imagination.

Leon Brozyna

This is what is known as thinking outside the box.

Massimo PORZIO

In middle 90s I designed a sun tracking solar panel system for the camper of a friend of mine. It worked great for getting the most energy from the panel even if the panel was partially shaded by trees. I remember that I get the best results by moving quickly the panel every 30-40 minutes and putting the system to low power for the rest of time. That because the panel had a quasi-co-sinusoidal response as function of the angle from its normal (it was probably a lambertian emission law pattern).
For that reason I really wonder seeing the discontinuities at the top of the “Standard” (I guess it’s the flat) panel response.
Does he tested the system in the place of the photo?
If he did it, that’s not really the right place to get the maximum from your solar panels, I can see many shades from the leaves and branches of the trees in front of the house.
One more, the efficiency should be measured as function of the produced power along the time, not as the voltage at the solar panel terminals.
If that measured voltages were taken without any load attached to the solar panels, those graph are meaningless, because many solar cells produce the nominal voltages with a little light on them but as you connect a small load the voltage drops quickly.


Yes, there may be some minor issues with the design — I did a biology experiment in 10th grade that was way lame compared to how I would design it today. That’s not a reason to knock it down but to ask questions and design a better follow-up experiment. Two specific comments:
1. Yes, it looks like there are more panels, but if that were the case the peak volts should go higher than 5. So maybe this is an illustration.
2. Tracking arrays are more efficient than static arrays. The problem is not the energy cost but the maintenance costs. Remember those wind farms Anthony posted about where 25% of the windmills are out of service at any give time due to maintenance needs? And I’ve seen the same thing.
3. Flat panels will probably be cheaper to install than trees, and land is cheap (for the sake of argument) but you’re forgetting about maintenance again. Didn’t Anthony post photos from a German solar plant where the ground was overgrown with weeds? Yes, here it is,
I can definitely think of times and places where an efficient small footprint “tree” would be better in the long run than a large flat static or tracking array.

Doug Proctor

If we want cheapness and efficiency, how about growing trees and then burning them in wood stoves: free fuel, good exercise and a zero-sum CO2 game. Forget solar panels entirely. Go Green. Nature is way ahead of us.
Should be a business plan here, perfect for Africa, if not New York City.
Congrats to the kid! He followed through with an idea. We all have ideas, and the world would be a better place if we put many of them into practice, getting better with time. Must be satisfying to him and his family. If Gore/Suzuki/Hansen followed through with their low-impact lifestyle they say is so necessary (for others) we’d be impressed also. Let’s let them take a lesson from this story: follow-through and see what happens.


richard verney says:
August 19, 2011 at 7:51 am
“I had always understood that to maximise the input energy, it was necessary for solar panels to be steered much the same way as a movable satelitte dish. Indeed, I believe that there are already mechanisms on the market which allow panels to be steered so as to track the movement of the sun.”
Yes indeed, but no more used in many new installations – panels have become so cheap that the cost of the mechanics for the tracker becomes uneconomic. Just add more panels instead of buying an expensive tracker.


its usually the obvious simple that works unfortunately as an afterthought LOL well done lad


I give young Mr. Dwyer much credit for an innovative experiment and for putting much thought into this. However, his results do not add up for me.
1. From the pictures, it appears that there are 10 solar cells in the standard array, but a larger number in the tree. It’s possible that there are more solar cells in the standard array which are on the “far side” of the gabled roof in the picture, but if so they would be shaded and fairly useless as far as collecting solar energy.
2. The standard array appears to be angled at 45 degrees to the horizon. If the location is at 40 degrees N latitude — just a guess — then it makes sense that it performs very poorly in December. The array should be significantly more vertical.
The combination of these 2 factors can explain the fact that the peak voltage from the standard array never gets as large as that from the tree arrangement. It makes sense that the tree arrangement can collect energy for a longer period of time, due to its diversity of orientations. And this shows in the comparison of voltage traces. But a properly situated linear array should generate a larger voltage when all of its panels are at an optimum orientation relative to the sun, and this is not the case in his results.
Again, congratulations to Mr. Dwyer. It is indeed a valuable insight to look at how Nature has evolved its creations, and to learn what works best in a practical, non-ideal environment. In this case, however, I don’t think that a tree design will improve on a (properly oriented) linear array as far as average power is concerned. However, the idea of multiple orientations can be used to extend the duration of large energy collection. [At a cost in either average power or number of cells.]

Luther Wu

fine Job, young man!!!

Robert F

Like others have said, no load voltage readings are meaningless. My 5,000 watt PV system tries to start every morning before sunrise, for the panel voltage rises over 250V as soon as there is some light in the sky. It quickly shuts down and tries again in 10 minutes. This may happen 3 or 4 times before the system has enough light to produce at least 75 watts.
In addition, shaded panels drag down the performance of the others in the string. That would mean that all of the “leaves” on the solar tree would have to be independent, which would be a wiring nightmare.
I give the young man a lot of credit for being creative, however.


The whole idea is completely stupid – even for a kid. When building complex structures for the mere purpose of holding some solar panels into the sunlight is an option, you do it right and let the panels track the sun. You don’t build an artificial tree and let half your panels face north (not even if there is a white wall right behind you). Leafs don’t cost much for the tree – and they work to some extend even when they’re lit from behind. Unless solar panels drop to a few cent/squarefoot and become semi-transparent and/or unless the structure of a rooftop isn’t already there and paid for anyway, this won’t fly.

Mike Spilligan

I really can’t undertstand how he managed to get started without a huge grant. 😉

Honest ABE

Sorry but I don’t buy it.
Trees aren’t designed just to catch sunlight – they have to compete with other plants to get at light, produce visible flowers for reproduction, conserve water and absorb CO2 – and they can’t do it from consistently flat surfaces like roofs.
A structure designed purely for collecting sunlight would be far more efficient that something like this.


A thought. Photosynthesis is efficient enough that most plants don’t do much to track the sun (with light hitting the leaf surface at less than perpendicular angles. They get enough light by remaining stationary and taking in both the direct rays of the moving sun and the light reflected from the atmosphere. That is why leaves face in all directions around the plant. Notice that a tree growing out in a field alone in the sun doesn’t favor the south (in the Northern Hemisphere), it grows symmetrically. Solar panels just aren’t efficient enough to get away with the same system.

This “work earned him a Young Naturalist Award from the American Museum of Natural History,” did it?
An engineering approach like this has anything to do with being a “Naturalist“?


Well done, the lad did good!
But I would comment that a trees leaves are not just designed to capture the sun, but also to restrict surface area and reduce evapo-transpiration etc – so, not exactly a direct ‘design’ comparison IMHO. And of course, the average tree has a fairly dense canopy arrangement to capture as much light as possible. As others have commented the design also requires more intricasies and indeed strength to provide the support to the panels, etc. – so clearly design as a function of cost and productivity is a major part of the equation!

S Matthews

Robert F;
‘In addition, shaded panels drag down the performance of the others in the string. That would mean that all of the “leaves” on the solar tree would have to be independent, which would be a wiring nightmare.’
Maybe not that bad with micro-inverters. See :


Wouldn’t a “tree” installation have a higher profile, air resistance-wise, than a conventional array? One good storm might bring the whole setup crashing down. Scaling it up to standard PV array sizes would also be problematic; the support structure would have to be more elaborate to handle the weight.


By definition, in a FIbonacci sequence, a particular term is the sum of the two preceding terms, so what does he mean when he says that “the Fibonacci sequence is 2/5”?


There are 16 panels on the tree, 10 on the front plane surface and 6 on the back of the plane.
This to me is not proper. All 16 should have been on the same, sun facing plane.
Everyone knows panels work best when the sun is at a normal vector to the panel. You can do the trig to find out the solar exposed area. Given my understanding of trig, the only way he could generate more in the tree was to put the 6 panels on the shady side.


I count 20 PV cells on the tree as compared to 10 on the flat panel. However, in the original article it is stated:

I made a second model that was based on how man-made solar panel arrays are designed. The second model was a flat-panel array that was mounted at 45 degrees. It had the same type and number of PV solar panels as the tree design, and the same peak voltage. My idea was to track how much sunlight each model collected under the same conditions by watching how much voltage each model made.

From this I must assume that the photograph of the experiment isn’t an accurate representation of the full experiment.
While some point out the light colored wall as a confounding variable in the experiment, I don’t see it that way. The traditional angled array of solar cells doesn’t benefit from the backscattered light from the wall whereas the tree array does. The fact that the angled array is lower than the tree and sees shorter exposure to the sun in comparison to the tree is actually a fairly accurate model. Many traditional solar arrays are on the ground below what would be the tree line.
What I get out of this is the solar array arranged in a Fibonacci sequence is able to capture and utilize more indirect light than a flat array and is less orientation dependent. Placing solar panels on a structure that raises the PV cells higher off the ground will result in exposure to the sun and backscattered light for a longer period of the day. I also imagine that using a tree style structure with a Fibonacci design allows for a greater density of PV cells for a given area of land.
Anyway, my conclusion is take some grant money away from the “Team” and give it to this kid. It would be money better spent as this kid is thinking out of the box – something we need more of in science today.