Department of Chemistry, College of Idaho, Caldwell, ID 83605, USA.
Analyst. 2018 Feb 26;143(5):1188-1196. doi: 10.1039/c7an01761h.
To simplify and improve the safety of reprocessing used nuclear fuel, an initial assessment was made of Raman microscopy applied to microfluidic volumes with a view toward the on-line spectroscopic measurement of highly radioactive solutions. This study compares a microscopic Raman probe (excitation focal point diameter 70 μm) to a larger, well studied probe (excitation focal point diameter 125 μm) used in prior investigations. This was done by chemometrically modeling and predicting concentrations of HNO solutions (0 M to 8 M) as they flowed through microfluidic cells based upon spectra from each probe. Spectra recorded for each probe using the same static HNO solution set contained in 2 dram glass vials were used as training sets to produce models for the respective probes. Modeling required baseline, normalization and smoothing preprocessing to compensate for a reduced path length between the static glass vial training set (4 cm) and the reduced path length flow cell (1 cm), wide ranging solution concentrations, and the associated non-linear spectral changes, and abrupt and uneven concentration changes of flowing solutions. The micro-Raman probe is able to produce spectra that may be analyzed chemometrically to accurately predict the concentration of flowing HNO solutions down to microliter volumes. Based upon RMSECV and RMSEP modeling statistics concentration predictions of the micro-Raman probe are comparable to those obtained for a macro-Raman on identical samples.
为简化和提高再处理使用过的核燃料的安全性,初步评估了拉曼显微镜在微流体积中的应用,以期在线测量高放射性溶液的光谱。本研究通过化学计量建模和预测 HNO 溶液(0M 至 8M)的浓度,比较了微尺度拉曼探头(激发焦点直径 70μm)和先前研究中使用的较大、研究较好的探头(激发焦点直径 125μm)。这是通过从每个探头的光谱来完成的,这些溶液流经微流控单元。每个探头记录的光谱用于相同的静态 HNO 溶液集,该溶液集装在 2 dram 玻璃小瓶中,用作训练集,以分别为各个探头生成模型。建模需要基线、归一化和平滑预处理,以补偿静态玻璃小瓶训练集(4cm)和减小的流动池路径长度(1cm)之间的路径长度减小、广泛的溶液浓度以及相关的非线性光谱变化以及流动溶液的急剧和不均匀浓度变化。微拉曼探头能够产生可通过化学计量分析进行分析的光谱,以准确预测流动 HNO 溶液的浓度,直至微升体积。根据 RMSECV 和 RMSEP 建模统计数据,微拉曼探头的浓度预测与在相同样品上获得的宏观拉曼的预测相当。