A near-infrared (NIR) spectroscopy platform can confirm the vitamin C content of a supplement or test how nutrient density varies according to the supply source, says Spanish startup Chemometric Brain.
As with many startups, Chemometric Brain was born out of a problem. In this case, food production services provider BlendHub needed a way of ensuring that active ingredients such as vitamins and minerals were evenly distributed throughout a blend, as well as verifying intra-batch homogeneity.
It also wanted to be able to authenticate the quality of the ingredients it was sourcing from different countries, to be sure that they were consistent in their quality and nutritional profile.
The company built a software platform that allowed it to use NIR spectroscopy to test ingredients and blends against these parameters. Branded Chemometric Brain, it was an internal, centralised solution that could be deployed by its customers and suppliers all over the world.
However, BlendHub soon realised that the platform offered potential beyond the ingredient authentication and blending control requirements within its own supply and value chains. In 2020, it was launched as a spin-out business and has since established itself as a SaaS (Software as a Service) platform for NIR spectroscopy solutions for the food and nutraceutical industry.
NIR spectroscopy uses the interaction between near-infrared light and matter to obtain information about a substance’s molecular composition and structure.
“Depending on the chemical composition of a product, it will absorb and reflect light at different wavelengths in different ways,” Jacob Kristensen Illán, founder and CEO of Chemometric Brain, told Vitafoods Insights.
“The beauty of this technology is that you can extract results in seconds and penetrate deeply into the sample, enabling homogeneous or heterogenous samples to be captured without the need for sample preparation.”
He said that while NIR spectroscopy offers vast potential outside the lab, currently it is only really being exploited in three applications: in commodity trading for analysing grain quality; by the animal feed industry for analysing feed composition; and by the dairy industry for measuring the fat, protein, lactose, and dry matter content of milk.
“The reason NIR spectroscopy is established in these sectors is that the hardware manufacturers have developed standardised models for these applications. Outside of these sectors, there is very little penetration and a lot of customisation,” Kristensen said.
Chemometric Brain offers a hardware-agnostic software platform for spectroscopy testing.
“Our strategy is akin to what Android is doing in phones,” explained Kristensen. “We don’t manufacture the hardware – there are over 70 manufacturers of instrumentation doing that already. We provide the operating system that makes it easier for model creators to develop apps for the food industry and for end users to access these applications.”
Users can use the platform to build quantitative models, which can then be used for measuring specific parameters, such as vitamin or protein content, or qualitative models, which might be deployed for authenticating ingredients or controlling homogeneity.
Qualitative models are built on the spectral “fingerprint” of the product, which is captured from a dataset.
“We train a model based on a group of samples – it’s a bit like facial recognition,” explained Kristensen. “This then allows you to compare new samples of that product to see whether they match. You can detect fraud, inconsistencies, and quality differences.”
The beauty of this solution, according to Chemometric Brain, is that it supports rapid in-house testing, allowing companies to move away from the cost-per-sample outsourced testing model that has historically dominated food quality testing.
“We’re betting on this technology because currently companies are very limited in terms of what they can do with analysis. Most testing is performed by external service providers – companies send out a sample and three or four days later, receive a result,” Kristensen said.
“This means it can’t be used as a decision-making tool, nor can it be applied exhaustively to production because of the cost. Once a business has invested in this solution, they can test as often as they like at no extra cost. The sky’s the limit.”
Instant test results also support data-driven decision making. For example, if a producer knows that an ingredient from a certain supply source has a different nutritional profile, it can adjust its process to account for that.
“We believe the future of quality in the food industry is rapid testing-based, and NIR spectroscopy allows for a broad spectrum of applications,” said Kristensen. “We are looking to deliver on this potential by creating a platform that will allow food companies of all sizes to benefit from faster and less expensive testing.”
