Materials scientists in Japan have developed a machine learning method to predict the mechanical properties of new polymers. Doctors Ryo Tamura, Kenji Nagata, and Takashi Nakanishi from the National Institute for Materials Science in Tsukuba have applied their technique to homo-polypropylene polymers. By analysing X-ray diffraction patterns under various conditions, they provide a non-destructive and cost-effective alternative to traditional manual testing methods.
Their approach, detailed in Science and Technology of Advanced Materials, leverages machine learning to predict material properties based on descriptors from X-ray diffraction data. This method captures detailed polymer structures, including changes during moulding. The researchers successfully correlated X-ray diffraction features with properties like stiffness, elasticity, and flexibility. This study offers a promising, non-destructive option for polymer analysis and suggests potential applications for other material data types, including X-ray photoelectron spectroscopy.
Read the full story here.
