[M-IS12] 結晶成⻑、溶解における界⾯・ナノ現象

[MIS12-03] 機械学習を用いたLC-TEMその場観察における核生成検出と結晶成長速度測定

*勝野 弘康¹、木村 勇気¹、山﨑 智也¹、瀧川 一学^2,3 (1.北海道大学低温科学研究所、2.理化学研究所革新知能統合研究センター、3.北海道大学化学反応創成研究拠点)

キーワード：透過電子顕微鏡、溶液成長、機械学習、粒子検出、成長速度

The nucleation of crystals is the first stage of crystallization and the origin of all materials. There are numerous studies of the crystallization of materials. The general understanding of crystallization is summarized in the classical nucleation theory, which is an extension of a macroscopic picture to a microscopic picture. The validity of the nucleation theory on an atomic scale has been studied in vaper growth. Recently, nonclassical nucleation phenomena have been reported in solution growth. Although the development of LC-TEM has enabled us in-situ observation in solution, it is difficult to capture the stochastic nucleation phenomena. To support the detection, recording, and analysis of nucleation events during in situ observations, we developed an early detection system by machine learning for nucleation events observed using a liquid-cell transmission electron microscope. Detectability was achieved using the machine learning equivalent of detection by humans watching a video numerous times. Our developed detection system was applied to the nucleation of sodium chloride crystals from a saturated acetone solution of sodium chlorate by electron-beam radiolysis. Nanoparticles with a radius of greater than 150 nm were detected in a viewing area of 12 um x 12 um by the detection system. The use of machine learning enabled the detection of numerous nanometer sized nuclei. The nucleation rate estimated from the machine-learning-based detection was of the same order as that estimated from the detection using manual procedures. The analysis of the change in the size of the growing particles as a function of time suggested that the crystal phase of the particles with a radius smaller than 400 nm differed from that of the crystals larger than 400 nm.