https://www.euro-online.org/enog/inoc2007/Papers/mac-slots.html https://www.euro-online.org/enog/inoc2007/Papers/m https://www.euro-online.org/enog/inoc2007/Papers/mac-slots.html

Solid as a composite polymer

Using a statistical technique known as multivariate image regression (MIR), Canadian chemical engineers have shown that NIR hyperspectral imaging can monitor both the chemical composition and physical properties of composite polymers.

Such composite polymers are becoming more common, as polymer producers increasingly replace virgin polymer resin with fillers such as natural fibres and recycled resin in order to reduce costs and make their products more environmentally friendly. But producing a polymer from a mixture of components is more difficult than producing it from a single type of resin. It requires ensuring that the different components are continuously mixed together in the right proportions and correctly dispersed throughout the polymer material, otherwise the resultant material won't have the desired physical properties.

Several groups have shown that NIR spectroscopy can be used to monitor the chemical composition of composite polymers, including filler concentration, as well as certain physical properties. But they have only done this for polymer melts during production, rather than the more challenging prospect of the finished, solid polymer product. Even though the cooling and solidification process can greatly alter the polymer's physical properties.

To extend NIR spectroscopy to finished polymer composites, Carl Duchesne and his colleagues at Université Laval in Québec turned to a combination of NIR hyperspectral imaging and MIR. In NIR hyperspectral imaging, a charged coupled device and data processing software are combined with NIR spectroscopy to produce a two-dimensional image in which each pixel contains spectral data.

Duchesne and his team used NIR hyperspectral imaging to analyse composites produced from a mixture of the polymer maleic anhydride grafted polyethylene (MAPE) and the natural fibre hemp, in which the hemp concentration varied between 0% and 60%. They then used MIR to try to relate the spectral data to the hemp concentrations and to physical variations that could affect properties such as tensile strength.

Rather handily, MIR uncovered two main latent variables in the data: one was related to the hemp concentration, while the other was related to physical variations caused by factors such as variations in the crystallinity of MAPE and the orientation of the hemp fibres. As they report in Industrial & Engineering Chemistry Research, this meant that by creating a model based purely on these two variables, Duchesne could accurately determine the hemp concentration and physical properties of the composites. Furthermore, because they used NIR hyperspectral imaging, they could track how these concentrations and properties varied across the composite.

Duchesne is now looking to apply this technique to the real-time monitoring of polymer composite production lines and to the design of new composites with desired physical properties.

Blog tags: