Smartphones and digital cameras that we use every day have built-in image sensors that "see" the world, capture light and shadow, and transform images into digital images. The way these sensors recognize color is actually similar to that of our human eye: the human eye recognizes color through three photoreceptor cells (also called pyramidal cells) in the retina: red, green, and blue; Each pixel of the image sensor also captures red, green, and blue light, respectively, and then converts it into an electrical signal.
Currently, most image sensors are made of silicon. Silicon absorbs all colors in the visible spectrum, but in practice, a color filter is added to each pixel. For example, in order to capture red, blue and green light must be blocked out with filters. In this way, each pixel can only use about one-third of the light, and the rest is "wasted".
A new breakthrough in all-optical utilization: stacked pixels
To solve this problem, a research team from ETH Zurich and Empa has proposed a new image sensor solution. Their core material is perovskite, a new type of crystalline semiconductor that can be used to make solar cells and other optoelectronic devices. Compared to silicon, perovskites are not only easier to process, but they can also absorb light of a specific color simply by adjusting their chemical composition.
The researchers found that if more iodine is added to the perovskite, it absorbs red light; When bromine is added, it can absorb green light; When chlorine is added, it can absorb blue light. What's more, these perovskite materials are transparent to other colors of light – meaning that the red, green, and blue pixels can be stacked on top of each other like "sandwich cookies" instead of having to be arranged horizontally like traditional sensors. Each layer absorbs only the color it needs, and the rest of the light continues to travel downward.
This structure not only saves space, but also allows each photon to be fully utilized, theoretically delivering three times the light sensitivity and three times the image resolution.
Figure: Breaking the limitations of silicon, perovskite sensors have taken a key step towards commercialization
The prototype was initially successful and moved towards practical application
In the early years, the research team had already verified the feasibility of this design in experiments with millimeter-sized large-particle perovskite crystals. Now, for the first time, they have succeeded in manufacturing two prototype image sensors in the form of thin-film images, which are no longer just "proof of concept", but a big step closer to real-world application.
While the two prototype sensors are still large in size (about 0.5 to 1 mm), the team says that with a proven thin-film manufacturing process, it will be possible to continue to reduce the pixel size in the future, to be in line with existing commercial image sensors with pixels in the micron range, and even smaller than silicon. Just like the development of electronic components, the original transistor was a large piece of material, but now it can be scaled down to the nanometer level.
At present, the research team is also optimizing the circuit structure for pixel reading and connection, because the existing electronic reading system is designed for silicon and is not suitable for the material properties of perovskites. But they believe that these technical difficulties can be overcome gradually.
The application is promising, and it is not just about taking pictures
Perovskite image sensors are used for much more than just taking pictures with mobile phones and cameras. Since it can be chemically tuned to flexibly select the wavelength of light absorbed, it is particularly suitable for high-end applications such as "machine vision" and "hyperspectral imaging".
Unlike the three primary colors of the human eye (red, green, and blue), machines may need to perceive more or different wavelengths of light in specific tasks, such as agricultural monitoring, environmental detection, medical imaging, etc. The perovskite sensor can be customized with multiple "color channels", and the absorption band of each layer is clear and controllable, avoiding the need to rely on a large number of filters and complex algorithm processing like silicon, which is more efficient and the system is more concise.
The future is promising: high-sensitivity, high-resolution, and smarter "new eyes"
In several experiments, the researchers have verified the significant advantages of perovskite sensors in terms of light sensitivity, color reproduction, and resolution. In addition, because each pixel captures all color information in its entirety, it reduces common problems in digital imaging, such as mosaic effects and moiré fringes.
In the next step, the team will continue to promote pixel miniaturization and improve the electronic readout system, paving the way for perovskite image sensors to be commercialized. In the future, it may not only become the new standard configuration of smartphones and digital cameras, but also become a smarter eye for machines to "see the world" in the fields of medical care, environmental protection, and agriculture.
