Schulich Faculty of Chemistry, Technion - Israel Institute of Technology, Haifa 32000, Israel.
Rapiscan Systems, Andover, Massachussetts 01810, United States.
J Phys Chem A. 2020 Nov 19;124(46):9656-9664. doi: 10.1021/acs.jpca.0c05909. Epub 2020 Nov 6.
Ion mobility spectrometry (IMS) is the method of choice to detect trace amounts of explosives in most airports and border crossing settings. For most explosives, the IMS detection limits are suitably low enough to meet security requirements. However, for some explosive families, the selectivity is not sufficient. One such family is nitrate-based explosives, where discrimination between various nitrate threats and ambient nitrates is challenging. Using a small database, machine learning methods were utilized to examine the extent of improvement in IMS selectivity for detection of nitrate-based explosives. Five classes were considered in this preliminary study: ammonium nitrate (AN), an ∼95:5 mixture of AN and fuel oil (ANFO), urea nitrate (UN), nitrate due to environmental pollution, and samples that did not contain any explosive (blanks). The preliminary results clearly show that the incorporation of machine learning methods can lead to a significant improvement in IMS selectivity.
离子迁移谱（IMS）是大多数机场和边境口岸检测痕量爆炸物的首选方法。对于大多数爆炸物，IMS 的检测限足够低，足以满足安全要求。然而，对于某些爆炸物家族来说，选择性还不够。硝酸盐类爆炸物就是这样一个家族，其中各种硝酸盐威胁与环境硝酸盐之间的区分具有挑战性。本研究使用一个小数据库，利用机器学习方法研究了 IMS 对硝酸盐类爆炸物检测的选择性提高的程度。在本初步研究中考虑了五个类别：硝酸铵（AN）、约 95:5 的 AN 和燃料油混合物（ANFO）、硝酸脲（UN）、环境污染引起的硝酸盐以及不含任何爆炸物的样品（空白）。初步结果清楚地表明，机器学习方法的结合可以显著提高 IMS 的选择性。
