Researchers developed a single photodiode sensor that detects full-color light using perovskite materials, impedance data, and machine learning algorithms.
(Nanowerk Spotlight) Conventional photodetectors are good at measuring how much light is present but fall short when it comes to identifying what kind of light they’re detecting. Extracting color or spectral information typically requires added hardware—filters, lens stacks, or multiple detectors arranged side by side or layered vertically. These systems tend to be bulky, expensive, and hard to scale, particularly when designing small, low-power devices. Even more recent strategies, like using heat or voltage to modulate a detector’s response, introduce other trade-offs, such as slower operation and higher instability.
At the core of this limitation is the nature of how standard detectors interact with light. Photons knock electrons free in the sensor material, generating an electrical signal. But those electrons don’t carry any identifying trace of the light’s color. Whether a photon is red or blue, the end result is the same kind of electrical charge. This process compresses a rich signal into something one-dimensional.
In a study published in Advanced Materials (“Full-Color Pixel with Only a Single Perovskite Photodiode”), researchers have created a color-sensing device that uses just a single photodiode—no filters, no moving parts, and no stacks of detectors. The key lies in using a perovskite material—a special kind of crystal that conducts both electrons and ions. These two types of charged particles respond differently when hit by light of different wavelengths. By tracking how these responses vary with frequency, and analyzing the resulting electrical patterns with machine learning, the researchers showed that a single detector could accurately identify full RGB color.
This approach is built on a fundamental difference in behavior between fast-moving electrons and slower-moving ions. When light strikes the perovskite, electrons begin conducting almost immediately. Ions, however, migrate more slowly and create delayed responses in the form of capacitance—temporary storage of charge. By applying a small oscillating voltage across the device and measuring how the current responds at various frequencies (a technique known as impedance spectroscopy), the team captured both the rapid and slow behaviors of these charge carriers.
A color imaging system built from a 10×10 array of single perovskite photodiodes. Each pixel detects color without filters by analyzing impedance signals. The system accurately reconstructs RGB values (b, c) and matches real object colors on a chromaticity map (c). Side-by-side comparison of the original (d) and reconstructed (e) images shows strong color fidelity and pixel uniformity above 95%. (Image: reprinted with permission by Wiley-VCH Verlag) (click on image to enlarge)
Light of different colors causes different ratios and movements of electrons and ions, and this leaves unique fingerprints in the frequency spectrum of the device’s impedance. However, these signals are not easily decoded by standard models. Instead of using simplified electrical circuit approximations, the researchers trained a neural network to learn the relationships between frequency-domain impedance data and the original light inputs. The result was a system that could reconstruct the red, green, and blue components of composite light with a relative error of less than 2 percent.
A central part of making this work was tuning the perovskite material so that ionic behavior became more useful for color discrimination. The researchers used small amounts of potassium and lithium to adjust how easily ions could move through the material. Potassium increased the energy barrier for migration, while lithium reduced it. By fine-tuning the composition with 1 percent lithium iodide, they achieved a balance where ionic contributions to the signal were strong enough to enhance spectral sensitivity without causing instability.
Color reconstruction remained accurate across a wide range of brightness—from 1 microwatt up to 200 milliwatts. At high light levels, many systems break down because the ions in the material become saturated. Once all the mobile ions are active, additional light no longer changes the signal in a measurable way. This limits the detector’s ability to distinguish color under bright conditions. To avoid this, the researchers used perovskite films thinner than 200 nanometers. Thin films prevent long-wavelength light from penetrating deeply and dominating the ion behavior, ensuring the response remains tied to photon energy rather than total light intensity.
To show that this approach could work in practical settings, the team built a small 10×10 pixel array. Each pixel contained one photodiode. When exposed to test images and color patterns, the array reconstructed accurate color information with over 95 percent uniformity across pixels. They also tested the sensor on real objects like fruit and leaves, confirming that the reconstructed color matched what a commercial spectrometer would record.
One challenge with impedance spectroscopy is speed. Scanning across a full frequency range can take several seconds, which is too slow for imaging. The researchers solved this by narrowing the frequency range and focusing only on the most informative segments. This brought the scan time down to under 4 milliseconds—fast enough for video frame rates.
This work shows that it is possible to extract rich spectral information from a single photodiode, using only the intrinsic properties of the material and a machine learning model. It avoids the need for optical filters, stacked structures, or mechanical modulation. The result is a simpler, smaller, and potentially cheaper device that could be used in robotic vision, wearable sensors, or compact analytical tools.
Get our Nanotechnology Spotlight updates to your inbox!
Thank you!
You have successfully joined our subscriber list.
Become a Spotlight guest author! Join our large and growing group of guest contributors. Have you just published a scientific paper or have other exciting developments to share with the nanotechnology community? Here is how to publish on nanowerk.com.
