ColloidTek was recently featured in Tekniikka&Talous (Technology & Business), a Finnish news magazine that is focused on technology and innovations.
The article “Tampere based company ColloidTek analyzes liquids even through containers – Fingerprint of the liquid is measured by self-learning data-analysis” (in Finnish) introduces the company and Collo Online Liquids Analyzer. CEO Matti Järveläinen and founding member Ilkka Sillanpää tell about the features of Collo®, and share examples on how it’s used in process industry.
The following summarizes the article in English.
ColloidTek helps with the challenge of measuring colloidal suspensions and liquids containing fine solids in real time. This is especially difficult, if also chemical variables need to be analyzed at the same time. Collo® is an innovative radio-frequency analyzer that detects electrical parameters from the analyzed liquid.
Collo® system probes liquids with a MHz-range radio wave pulse, and the gained spectrum, that ColloidTek calls “fingerprint of a liquid“, is transmitted to a computer that analyses the data by comparing it to a database.
The measurement can be done either with an immersible sensor through a plastic or glass container, or with a flange connected to a metallic pipe. The measured information can be sent straight to cloud if desired.
Data analysis is based on machine learning
Järveläinen stresses that even more important than the physical measuring principle, is the data analysis that is based in machine learning: ”Computation is not standardized, AI follows a few dozen parameters from the spectrum and the needed variables are taught to it based on known examples case by case”. In comparison, spectrometers and selective electrodes measure a theoretically well-known and mathematically reduced quantity such as absorbance or voltage.
The method developed by ColloidTek is particularly suitable for colloidal suspensions, because of the electrical double layer surrounding the colloidal particles. Radio waves spot the changes in those particles well. ”Double layer prohibits the particles to stick to each other. Should the layer disappear, a sedimentation is formed. This can be either good or bad, depending on the process. Typically, this phenomena needs to be carefully controlled”, explains Järveläinen, who has made his PhD-thesis on colloidal chemistry.
Järveläinen and Sillanpää tell that the company has gained Finnish customers for example from chemical, mining and water industries. One customer has optimized the amount of added chemicals in their mineral refinement process, whereas another monitors properties of a ceramic slurry. Earlier this customer used a method based on mechanical vibration, and Collo is a remarkable improvement to that. Light or infrared sensors are not able to help, since the slurry is very thick.
ColloidTek was founded 2017 as a spin-off from TUT, from a joint-project by materials science and automation and hydraulics laboratories.