Fast, high-quality infrared devices are expensive. That’s why they’ve been used mainly in applications such as space imaging and night vision for military helicopter pilots. But now MIT researchers are developing a method for making high-quality infrared devices for one-tenth of the cost, which could eventually lead to widespread use in civilian applications, such as cancer detection and night-vision displays in vehicles.
Relatively low-cost infrared devices are already available to consumers, but these devices are unreliable and tend to produce noisy, low-resolution images that refresh slowly, says Anu Agarwal, a research associate at MIT who manages the infrared project at the institute’s Microphotonics Center. High-end devices produce sharp images and video in real time, but need to be cooled with liquid nitrogen, she says, and they’re made with expensive materials and specialized tools.
The new method, which works at room temperature, uses materials that are much less expensive for converting infrared light into electrical signals for displays. Also, the detector can be made with tools similar to those used to make the electronics in the device, eliminating the need for specialized, costly equipment. Indeed, the sensor can be fabricated directly on silicon along with the electronics that read out the signal, which makes it possible to pack more pixels into a given area, increasing resolution, Agarwal says.
As a further benefit, each pixel can sense three or four specific wavelengths of light in either the visible or infrared range. With conventional technology, sensing multiple wavelengths requires using a pixel for each wavelength. Using one pixel for all the colors allows for significantly better resolution, Agarwal says.
Right now, the researchers are focusing on the development of devices that sense light at very specific frequencies. Within each pixel, multiple detector materials are also tuned to respond to specific wavelengths. Using specific wavelengths makes it possible to pinpoint, for example, the temperature of objects or certain substances. Cancerous tumors emit specific infrared wavelengths, for instance, and a detector set to a narrow frequency should be able to identify these against the background of the body’s heat, says George Kenney, associate director of the Microphotonics Center. Eventually, this feature could be used by firefighters who need to see light at the wavelengths emitted by a human body, without being distracted by light from fire or other sources, making rescue operations easier.
Firefighters could also make use of multiple frequencies at once, each color-coded on a display, says Agarwal. Hot spots detected at one frequency could be displayed as red; infrared light emitted by human bodies could be displayed as blue; and the critical temperature of flammable liquids could be yet another color. The areas on screen that are not at these specific frequencies would appear as mixtures of these colors – this is, in fact, precisely how digital cameras produce full-color images.
With future developments in materials it could be possible to sense wider ranges of frequencies. Agarwal says this might be especially useful for helicopter pilots. In one scenario, a detector material might be tuned to respond to all the visible frequencies. Its output would be displayed in a grayscale, creating a black-and-white image of the visible scene. On top of this background could be overlaid red images indicating the location of people, and other colors for identifying different types of vehicles.
By dramatically cutting the cost of such a system, the new technology could make high-quality infrared detection widely available for nonmilitary applications. These could include detecting bruised apples on an assembly line, surveying dark areas with security cameras, and monitoring industrial machines for overheating. Automaker BMW already offers cars with infrared systems that help drivers see at night, alerting them to animals or people beyond the beam of a car’s headlights. The new, inexpensive device might improve the quality of such a system and make it more widely available.
Agarwal declines to reveal the exact composition of the materials used, although she says they’re not new. The materials weren’t used in the past for this application because researchers believed they wouldn’t produce a clear signal. “People thought the signal-to-noise ratio would be horrible, so there’s no point in researching these materials,” she says.
While a prototype device for detecting narrow frequency ranges for early military applications could be ready within a year, Agarwal says other applications will require further materials developments to extend the frequency range of the detectors while maintaining a clear signal, and therefore could be several years away.
This new data poisoning tool lets artists fight back against generative AI
The tool, called Nightshade, messes up training data in ways that could cause serious damage to image-generating AI models.
The Biggest Questions: What is death?
New neuroscience is challenging our understanding of the dying process—bringing opportunities for the living.
Rogue superintelligence and merging with machines: Inside the mind of OpenAI’s chief scientist
An exclusive conversation with Ilya Sutskever on his fears for the future of AI and why they’ve made him change the focus of his life’s work.
How to fix the internet
If we want online discourse to improve, we need to move beyond the big platforms.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.