Almost half of the food thrown away in Germany ends up in the trash only because the best-before date has been exceeded. And yet these foods are usually still completely in order and not spoiled at all. The project DProFood wants to counteract this waste with modern technology: The project team around Professors Dr Hans-Jürgen Danneel and Dr Volker Lohweg is developing sensors that can assess the condition of food.

Is the product still fresh, or have processes already started that indicate spoilage? The aim of the project is to detect such processes before human senses can perceive them. Ideally, sensor monitoring should start with the raw materials at the beginning of the production chain and only end when the end consumer eats the food. “If we could always know exactly what condition the product is in at any given time, then we could drastically reduce the waste of food,” says Danneel.

Surprisingly little is known about spoilage processes

However, the DProFood team has a long way to go to reach this goal, because the problem is enormously complex. And the researchers are really starting at the very beginning: What actually happens when food spoils, and which substances can be used as spoilage indicators? It is quite astonishing that so little is known about this, says biochemist Danneel.

But his team has already gathered a wealth of experience from similar projects and knows where best to start searching. Interesting candidates for a reliable spoilage indicator are, for example, biogenic amines, which typically appear in spoiled fish and meat and cause a strong smell. However, by the time the human nose detects these substances, the food might already be inedible. Danneel says that a sensitive sensor can react to biogenic amines much earlier and warn the consumer in good time that this food should be consumed soon.

In the DProFood project, the research team will initially focus on food production and establish a monitoring chain for these production processes. This involves the use of various sensors, the measured values of which will then be merged into an overall result: from electronic noses and ultrasound to hyperspectral cameras that detect mould formation at a very early stage, practically anything is conceivable here.

The frozen pizza as an ideal test object

A frozen pizza is currently being used as a test object. Daniel Pauli, a research assistant in Danneel's team, explains this choice: “On the one hand, a pizza has many components, a wide range of raw materials is used - and each raw material spoils differently, so we have many starting points for different measurements here,” he says. “And on the other hand, a fresh or defrosted pizza spoils comparatively quickly. Therefore, we can observe the changes without having to wait weeks for something to happen every time we try it.” Later, the findings from the pizza experiments will be transferred to more durable product classes such as pudding or yogurt.

Hans-Jürgen Danneel is convinced that the automated monitoring technology will be a help to the employees in food production. “The people on the production lines have a great responsibility,” he says. “They have to keep an eye on the quality of the raw materials and yet produce according to economic specifications. A warning system that intervenes if necessary and perhaps says: Stop, at this point you have to heat up a little longer to make sure the product is safe - that would take a lot of pressure off the employees”. To this end, the system will be equipped with machine learning technologies so that at some point, it will be able to choose the best of many options autonomously.

In the long run, the DProFood team also envisages that food is monitored after production. Because off the production line, the external influences that can destroy a food product multiply: Now the goods are exposed to possible transport damage due to light, temperature or vibrations, can be stored incorrectly or thawed accidentally. And above all, it is no longer on a production line where it is transported past the required sensors but must carry its sensor with it - perhaps as a tiny component in the packaging.

Finding the right balance when using sensors

The DProFood team relies on so-called patches, which contain very small sensors and can, for example, react to external influences with colour changes. “Such patches already exist, but only for individual factors and almost always in the research phase,” explains Pauli. For example, the patches can react to an interruption in the cold chain and indicate this by changing colour. However, spoilage is such a complex process that it is not enough to detect individual factors - you would have to combine several sensors for this.

“At some point, however, the packaging may become too expensive if it is plastered with high-tech sensors,” Danneel concludes. It's a tightrope walk to find the right balance for a sensible use of sensor technology. And the scientists see another danger: What if monitoring the food only leads to the inferior raw materials being sorted out before production? That would shift waste from the consumer to the producer, “and that's obviously not what we want to achieve,” explains Danneel.

Nevertheless, the DProFood team believes that sensor technology can give new impetus to the often very conservative food industry and provide consumers with more security. And even if there is still a long way to go until complete freshness monitoring is achieved: The basic knowledge of how and when food spoils is in any case useful for everyone involved.