Contrary to conventional wisdom, what you see is not always what you get. There are hard limits to what the human eye is capable of perceiving. For example, in most cases, we cannot tell the difference between normal and cancerous tissue by eye. Similarly, we cannot see the pollutants that are floating around in the air, or identify specific nutrient deficiencies in crops. For jobs such as these, we need tools that can give us a more detailed look at the world.
Spectral sensing is one such tool. Using this technology, materials, biomarkers, abnormal tissue, and so on can be identified by analyzing the complex way that they interact with light. Given what spectral sensing is capable of, it is an extremely useful tool. However, existing systems are highly complex and very expensive, which means they are impractical for most use cases, and rarely ever see the light of day outside of a research lab or industrial setting.
A schematic of the optical sensor (📷: X. Cui et al.)
Now, researchers at Aalto University in Finland are changing that narrative. By combining miniaturized hardware with intelligent algorithms, they have developed a spectral sensing device that is small enough to fit into a smartphone or wearable device.
Just how small is this sensor? Very. It measures just 5 micrometers by 5 micrometers — an area 200 times smaller than the cross-section of a human hair. Yet this tiny sensor is capable of incredible precision, distinguishing between thousands of colors with a wavelength identification accuracy of approximately 0.2 nanometers. This level of accuracy makes it capable of identifying materials such as organic dyes, metals, semiconductors, and dielectrics
Unlike traditional spectral sensors that rely on large optical components like prisms or gratings, this new sensor uses its electrical responses to light to perform spectral differentiation. This approach not only reduces the size and cost of the device, but also makes it easier to integrate into everyday devices.
Organic dyes that were identified by the system (📷: X. Cui et al.)
The device works by exposing the sensor to various colors of light during its training phase. It then generates unique electrical fingerprints for each type of light, which are decoded by advanced algorithms. This novel method allows the sensor to analyze light in a way that eliminates the need for bulky optical components, making it ideal for applications requiring portability and scalability.
In lab tests, the sensor demonstrated its capability to identify unknown wavelengths with a high level of precision. It achieved a wavelength identification accuracy of 0.19 nanometers in free-space applications. Moreover, when integrated with a silicon waveguide, the device’s accuracy remained competitive, coming in at 2.45 nanometers. In further tests, the system was demonstrated to be able to classify the unique spectral fingerprints from a variety of substances, such as organic dyes, metals, and semiconductors, with a high degree of certainty
Despite these promising results, the researchers acknowledge that challenges remain. Noise in electrical signals, device miniaturization, and material variability are some factors that can negatively affect the system’s performance. Addressing these issues will be important as the technology moves from the research lab to commercial applications, where “good enough” will no longer be good enough.
