Department of Chemistry, Kookmin University, 77, Jeongneung-ro, Seongbuk-gu, Seoul, 02707, Republic of Korea.
Sensor System Research Center, Korea Institute of Science and Technology (KIST), 5 Hwarang-ro 14-gil, Seongbuk-gu, Seoul, 02792, Republic of Korea.
Anal Methods. 2020 Jun 18;12(23):3032-3037. doi: 10.1039/d0ay00612b.
In situ real-time and nondestructive identification of packaged chemicals is essential for applications such as homeland security and terrorism prevention. Although various Raman spectroscopic methods such as spatially offset Raman spectroscopy (SORS) and time-resolved Raman spectroscopy have been investigated for real-time detection, the background interference originating from packaging materials limits the accuracy of the analysis. In principle, the Raman background from the packaging cannot be removed completely. To overcome this limitation, we developed a SORS-based dual-offset optical probe (DOOP) system that offers real-time prediction of 20 chemicals concealed in various containers by completely removing the background signal. The DOOP system selectively acquires the Raman photons generated from both the outer packaging and the inner contents, whose intensities are dependent on the penetration depth of the laser. The Raman spectra obtained at two remote offsets are automatically subtracted after normalization. We demonstrate that the DOOP method provides the pure component spectra by completely removing background interference from three plastic containers for a total of 20 samples in three different containers. In addition, an artificial neural network (ANN) was applied to evaluate the accuracy of the real-time chemical identification system; our system led to drastic improvements of the ANN prediction accuracy.
原位实时、非破坏性地识别包装化学品对于国土安全和反恐预防等应用至关重要。尽管已经研究了各种拉曼光谱方法，如空间偏移拉曼光谱（SORS）和时间分辨拉曼光谱，用于实时检测，但源于包装材料的背景干扰限制了分析的准确性。原则上，包装的拉曼背景无法完全去除。为了克服这一限制，我们开发了一种基于 SORS 的双偏移光学探头（DOOP）系统，通过完全去除背景信号，实时预测隐藏在各种容器中的 20 种化学品。DOOP 系统选择性地获取来自外包装和内部内容的拉曼光子，其强度取决于激光的穿透深度。在归一化后自动减去在两个远程偏移处获得的拉曼光谱。我们证明 DOOP 方法通过从三个塑料容器中完全去除背景干扰，为总共三个不同容器中的 20 个样本提供了纯组分光谱。此外，还应用了人工神经网络（ANN）来评估实时化学识别系统的准确性；我们的系统导致 ANN 预测精度的大幅提高。
