A thin section of eucalyptus lies under the microscope objective, and white light, passing through two polarizing filters, turns the wood into a mosaic of purple, yellow, and blue patches. These colors are not random: like letters on a map, they show the direction of the invisible cellulose threads inside the cell walls. Previously, such a map could not be read without destroying the sample, because an electron microscope cannot distinguish cellulose from neighboring substances, and a conventional polarization microscope confused two different orientations of the fibrils.
The method relies on a simple property: cellulose refracts light differently depending on the direction of the wave, so under crossed polarizers it begins to glow with a specific color. But this glow has a blind spot: fibrils tilted at 45° and 135° gave the same signal, and the microscope could not distinguish them. The researchers added a compensator to the optical setup: a plate that shifts one part of the light wave relative to another by 530 nanometers. This shift, like a translator, separates the two previously indistinguishable directions and turns them into different colors.
To verify that each color indeed corresponds to a specific arrangement, the scientists built a computer model of the cell wall as a hollow cylinder made of many layers. In each layer, they set the tilt angle of the fibrils and calculated what color the microscope would see. It turned out that five types of architecture: line, helix, ring, crossed helix, and twisted helix: have unique color profiles. On a cross-section, the line is always purple, while on a longitudinal section, the helix and the ring differ by which color appears when the compensator is rotated.
Experiments on real cells confirmed the model's predictions. In the trunk, rings made up 6%, helices 27%, and twisted helices 67%. In the branch bark, out of ten cells, two turned out to be crossed helices, which are indistinguishable from linear ones under a conventional microscope but give a purple color instead of blue under polarized light. Three different identification methods gave the same result on the same cells, confirming the reliability of the color labels.
Now the color map of wood becomes a tool that can be read automatically. The color profiles are so distinct that they are ideal for machine vision and neural network algorithms. Already, the method allows simultaneous analysis of dozens of cells, counting their types, and noting the location of clusters. The next step is to train a neural network on simulated and experimental images for fully automated classification. This will open the possibility of rapid screening of large arrays of wood and other plant tissues, and will also help in creating biomimetic materials with improved structural integrity.
Cellulose microfibrils in a tree's cell wall are thinner than the wavelength of visible light, but their collective arrangement creates colors visible to the eye. In the eucalyptus trunk, 67% of cells have a twisted spiral structure that resembles the cholesteric liquid crystal phase.
The idea of reading the 'memory of a tree' through color patterns echoes Ursula K. Le Guin's novel 'The Word for World Is Forest', where the forest stores information in its structure. Here, polarized light acts as a translator from the language of cellulose spirals to the language of color.