In a Cambridge laboratory, a team of researchers and their start-up GlitterinTech have shrunk a tool that light has long been asked to reveal its secrets through, the spectrometer, into a device a few centimeters across that costs about ten dollars. Such instruments have always forced a trade-off, where making them small meant making them less precise, and this design turns that bargain by folding the mathematics of convolution into the light itself. The promise reaches beyond the bench: a sensor cheap and small enough to live inside everyday wearables might one day listen to the body's chemis
Cambridge's $10 Optical Chip Could Bring Lab-Grade Sensing to Wearables
Lab-grade sensing in a device that costs about $10
Mark
What's the core idea here? Why does a $10 chip matter?
Mimi
Small spectrometers have always traded away performance. This one does a math operation on the light itself, so it stays small and cheap without losing much precision.
Mark
And the test results were strong?
Mimi
It classified plastics, medicines, coffee, flour and tea with 100% success, and measured solution concentrations to about 0.01%.
Luke
Worth noting that's 100% in the reported tests. The source doesn't say how many samples or how varied they were, so I wouldn't read it as a general guarantee.
Mark
What about the health side? Glucose is the big prize.
Mimi
They measured skin moisture, blood alcohol, lactate and glucose, and tracked glucose over time without piercing the skin.
Luke
In one participant. That's a demonstration, not validation. The source itself says further testing is needed before it's a medical product.
Mark
The temperature range seems relevant for wearables.
Mimi
Yes, it kept working from -20°C to 80°C, which suits devices worn outdoors or used in factories.
Luke
The claim that it beats benchtop instruments is also limited to those experiments. And the bandwidth and resolution scaling is described as theoretical.
Mark
So what's the next marker?
Mimi
Clinical validation and whether GlitterinTech can bring it to market.
Der Puls
- A ten-dollar spectrometer has matched laboratory precision in tests, challenging a long-standing assumption that small meant imprecise.
- The design dispenses with conventional light-splitting and heavy reconstruction, pushing the mathematics directly into the optical hardware.
- Early trials picked out plastics, medicines and foods with perfect accuracy and tracked glucose in one participant without needles.
- The team is now racing toward clinical validation, while the commercial path runs through GlitterinTech and its eventual wearable and factory deployments.
- Nothing is ready for patients yet, and the decisive test is whether the noninvasive glucose signal holds up across many bodies and many months.
In a Cambridge laboratory, a team of researchers and their start-up GlitterinTech have shrunk a tool that light has long been asked to reveal its secrets through, the spectrometer, into a device a few centimeters across that costs about ten dollars. Such instruments have always forced a trade-off, where making them small meant making them less precise, and this design turns that bargain by folding the mathematics of convolution into the light itself. The promise reaches beyond the bench: a sensor cheap and small enough to live inside everyday wearables might one day listen to the body's chemistry without breaking the skin. For now it remains a promising beginning, with the long human work of clinical proof still ahead.
Researchers at the University of Cambridge, working alongside GlitterinTech, a start-up born from their own group, have built an optical spectrometer no bigger than a few centimeters that costs about $10. They claim it rivals laboratory instruments many times its size, and that it is small enough to slip into phones, portable electronics and wearable health devices.
Spectrometers read light to learn what it has touched, and they serve medicine, factories, chemistry and environmental science. Making them smaller has always meant giving something up, usually range, resolution or accuracy. This design takes a different path. Instead of splitting light apart or reconstructing it with heavy computation, it applies a mathematical operation called convolution directly to the light, using a sequence of optical components tuned in proportion to one another. The approach runs on a silicon nitride chip with onboard electronics, measures across near-infrared wavelengths, and returns a result in under a second.
In trials, the device identified plastics, medicines, coffee, flour and tea without a single error, and it measured solution concentrations finer than some commercial benchtop machines. It worked across a temperature range unusual for an instrument of its size. The health experiments drew the most interest: under conditions meant to mimic real physiology, it sensed skin moisture, blood alcohol, lactate and glucose, and in one participant it followed glucose noninvasively over time.
The lead author, Dr. Chunhui Yao, credited a return to the underlying mathematics, while Prof. Qixiang Cheng, who led the project, praised the work for being robust and fully packaged rather than merely effective in a lab. The team pictures cheap sensors continuously checking product quality in factories, judging food freshness, and monitoring hydration, intoxication and blood sugar in wearables. Much more testing is required before any of this reaches patients, and the path forward will be defined by clinical validation and whatever commercial rollout GlitterinTech attempts next.
Researchers at the University of Cambridge, working with GlitterinTech, a start-up founded by the same research group, have built an optical spectrometer that measures only centimeters across and costs about $10. Their claim is that it matches the precision of laboratory instruments many times its size, and that it is small enough to sit inside portable electronics and wearable health devices.
Spectrometers analyze light to learn about the substances it has interacted with, and they are used across medicine, manufacturing, chemical testing and environmental research. Shrinking them has usually cost something. Compact versions tend to have a narrower measurement range, lower resolution or reduced accuracy, which leaves them better at identifying a substance than at measuring it precisely.
The Cambridge design, called a convolutional spectrometer, takes a different route. Conventional devices either split incoming light into its wavelengths or use heavy computation to reconstruct a spectrum. This one applies a mathematical operation, convolution, directly to the light. The team arranged optical components with repeating spectral responses in a sequence, such as unbalanced Mach-Zehnder interferometers or micro-ring resonators, and tuned them in proportion to one another so the response shifts linearly across wavelengths. Fast Fourier transforms then recover the spectrum. The researchers say this outperforms dispersive, Fourier-transform and computational-reconstruction designs in several key areas.
The device is built on a silicon nitride photonic platform with onboard electronics. It works across near-infrared wavelengths from 1200 to 1700 nm and collects and processes a measurement in under a second. Because the optical patterns repeat, the bandwidth can in theory be extended almost without limit, and adding components in sequence can raise resolution exponentially. Because the measurements follow a linear relationship, recovering the spectrum takes little computing power and is less vulnerable to noise than methods that depend on elaborate calibration.
In tests, the sensor classified plastics, pharmaceuticals, coffee, flour and tea with a 100% success rate. It measured concentrations in water-based and organic solutions to roughly 0.01%, which in those experiments beat commercial benchtop spectrometers. It also kept working between -20°C and 80°C, an unusually wide range for a miniature instrument.
The health results drew the most attention. Under conditions designed to reflect real physiological environments, it measured skin moisture, blood alcohol, blood lactate and blood glucose. The team also tracked glucose noninvasively over an extended period in one participant. The source stresses that much more testing is needed before this could become a medical product, but the result hints at a way to monitor blood sugar without repeatedly piercing the skin.
Dr. Chunhui Yao, a lead author, said the key was returning to the mathematics and using the convolution theorem directly in the optical domain. Prof. Qixiang Cheng, who led the project, credited Yao with turning a mathematical concept into a robust, fully packaged system, and said the work stands out for technical readiness as well as lab performance. Prof. Richard Penty, who contributed to the photonic design and integration, called the architecture simple, scalable and manufacturable.
The team imagines factories fitted with cheap spectrometers that continuously check material composition and product quality, along with sensors for food freshness and environmental analysis. In healthcare, wearables might track hydration, signs of alcohol intoxication, fitness biomarkers and glucose for people with diabetes. Cheng said the aim is to make spectrometry as ubiquitous as temperature or motion sensing. What to watch next is clinical validation of the health measurements and any commercial rollout through GlitterinTech.
Bemerkenswerte Zitate
Our vision is to make spectrometry as ubiquitous as temperature or motion sensing.— Prof. Qixiang Cheng, project lead
The architecture is simple, scalable, and manufacturable, yet it delivers precision that rivals or exceeds much larger instruments.— Prof. Richard Penty
Möchten Sie die ganze Geschichte?
Das Original lesen bei Science Daily ↗
Technology