But that seems unsatisfyingly wasteful, because most of the energy that the sun radiates is in the green part of the spectrum. When pressed to explain further, biologists have sometimes suggested that the green light might be too powerful for plants to use without harm, but the reason why hasn’t been clear. Even after decades of molecular research on the light-harvesting machinery in plants, scientists could not establish a detailed rationale for plants’ color.Recently, however, in the pages of Science, scientists finally provided a more complete answer. They built a model to explain why plants' photosynthetic machinery wastes green light. What they did not expect was that their model would also explain the colors of other photosynthetic forms of life too. Their findings point to an evolutionary principle governing light-harvesting organisms that might apply throughout the universe. They also offer a lesson that—at least sometimes—evolution cares less about making biological systems efficient than about keeping them stable.
The mystery of the color of plants is one that Nathaniel Gabor, a physicist at the University of California, Riverside, stumbled into years ago while completing his doctorate. Extrapolating from his work on light absorption by carbon nanotubes, he started thinking of what the ideal solar collector would look like, one that absorbed the peak energy from the solar spectrum. “You should have this narrow device getting the most power to green light,” he said. “And then it immediately occurred to me that plants are doing the opposite: They’re spitting out green light.”
In 2016, Gabor and his colleagues modeled the best conditions for a photoelectric cell that regulates energy flow. But to learn why plants reflect green light, Gabor and a team that included Richard Cogdell, a botanist at the University of Glasgow, looked more closely at what happens during photosynthesis as a problem in network theory.
The first step of photosynthesis happens in a light-harvesting complex, a mesh of proteins in which pigments are embedded, forming an antenna. The pigments—chlorophylls, in green plants—absorb light and transfer the energy to a reaction center, where the production of chemical energy for the cell’s use is initiated. The efficiency of this quantum mechanical first stage of photosynthesis is nearly perfect—almost all the absorbed light is converted into electrons the system can use.
But this antenna complex inside cells is constantly moving. “It’s like Jell-O,” Gabor said. “Those movements affect how the energy flows through the pigments” and bring noise and inefficiency into the system. Quick fluctuations in the intensity of light falling on plants—from changes in the amount of shade, for example—also make the input noisy. For the cell, a steady input of electrical energy coupled to a steady output of chemical energy is best: Too few electrons reaching the reaction center can cause an energy failure, while “too much energy will cause free radicals and all sorts of overcharging effects” that damage tissues, Gabor said.
And as the damage caused by climate change becomes more apparent, AI experts are increasingly troubled by those energy demands.“The concern is that machine-learning algorithms in general are consuming more and more energy, using more data, training for longer and longer,” says Sasha Luccioni, a postdoctoral researcher at Mila, an AI research institute in Canada.
Gabor and his team developed a model for the light-harvesting systems of plants and applied it to the solar spectrum measured below a canopy of leaves. Their work made it clear why what works for nanotube solar cells doesn’t work for plants: It might be highly efficient to specialize in collecting just the peak energy in green light, but that would be detrimental for plants, because when the sunlight flickered, the noise from the input signal would fluctuate too wildly for the complex to regulate the energy flow.