Image by Jill Wellington from Pixabay
Author profile picture

MantiSpectra, an Eindhoven University of Technology (TU/e) spin-off company, has developed a sensor that can be used to quality check unpasteurized milk and check if cows have health problems. Milk (as well as other food products and materials) reflect infrared radiation that can give valuable information to its checmical make-up and quality. Unfortunately, this light is invisible to humans – meaning that we have to use expensive methods to conduct quality checks.

With the help of a machine-learning algorithm, the MantiSpectra chip-sensor can be used by farmers to detect infrared light and use it to quickly determine if milk has any defects.

“The problem is, this is usually measured during the processing of the milk,” explains Maurangelo Petruzzella, CEO at MantiSpectra. “If you are able to measure this directly at the farmers level, then you know also the health status of the cow.”

The device is already in use by early adopters of the tech – the most well-known is its use in determining when a crop, such as a tomato, is ripe for picking. Farmers generally rely on their experience to know when tomatoes are ready to harvest but this is not always a fool-proof method. Picking a tomato too early can dull its flavor – while tomatoes picked past their prime may not make it to the supermarket. The infrared lights reflected by crops is one measurable indicator for when they are exactly ready for picking.

The sensor can also be used to sort plastics in recycling, determine whether a bottle of pills has been labelled correctly, and even scan cheese to adjust the blending process.

“For the food and pharmeucutical industries, it allows you to have a level of control at many different stages of the product’s development,” says Petruzzella. “This is important to happen before it reaches the customer or next processing step.”

Nano spectrometer

The MantiSpectra chip is a spectrometer – a device that gathers light, isolates the infrared spectrum of the light, and uses it to gather information about the chemical compound of the material. Petruzzella and his team initially developed their sensor in 2017 as a nano-spectrometer that could fit in a smart-phone. After undergoing a multi-year development phase in the TU/e Nanolab, the chip-sensors are ready and are being tested as a prototype targeted for volume production.


Tests are currently underway for the applications in dairy farms but the idea would be that farmers would have access to quick data on their milk (and their cows) – without having to go to a lab. The hope is to scale-down the price of the sensors so that it can be used by these first-stage manufacturers (as well as other industries) as an extra control in their production process.

Read about how robots make agriculture more sustainable.