喫煙室の排気に含まれる金属濃度のリアルタイム分析と階層ベイズモデルによる各銘柄によるニッケル排出寄与の推定：島根大学における事例

抄録

A hyphenated analytical system with a gas exchange device (GED) and an inductively coupled plasma mass spectrometer (ICP-MS) enabled us to measure the concentration of metals in atmospheric particulate matter in real time. In this study, with a special focus on passive smoking whereby non-smokers inhale cigarette smoke in the environment, the concentration of metals contained in smoking booth exhaust was directly measured by GED-ICP-MS. Furthermore, we used a generalized linear mixture model (GLMM) and hierarchical Bayesian model (HBM) to explain the analytical results. To compute the statistical models, the number of smokers of each brand was used as an explanatory variable, and the emission contribution of each brand to the metal concentration in the exhaust was estimated. GLMM and HBM analyses were carried out based on random effects such as observation errors and errors among smokers. It became clear that the number of smokers of brand C contributes to the increase in the concentration of Mn, Fe, and Ni. As for Ni, where we also introduced a hierarchical Bayesian model to compare brands, the probability of the regression coefficient of brand C being larger than 0 was 1.0. Moreover, the probability that the regression coefficient of brand C is larger than those of other brands was 0.97 or more. Furthermore, in the scenario where four smokers smoke brand C near the exhaust port at the same time, the Ni concentration in the smoke exhaust was estimated to be increased by 0.05 ng/m³, and if it was assumed that this exhaust is exposed for lifetime, the carcinogenic risk was estimated to increase by a maximum of 1.2×10⁻⁸.