Groves W A, Zellers E T
Department of Environmental Health Sciences, University of Michigan, Ann Arbor, MI 4801-2029, USA.
Ann Occup Hyg. 2001 Nov;45(8):609-23.
This article describes the development and evaluation of a small prototype instrument employing an array of four polymer-coated surface acoustic wave (SAW) sensors for rapid analysis of organic solvent vapors in exhaled breath and ambient air. A thermally desorbed adsorbent preconcentrator within the instrument is used to increase sensitivity and compensate for background water vapor. Calibrations were performed for breath and dry nitrogen samples in Tedlar bags spiked with 16 individual solvents and selected binary mixtures. Responses were linear over the 50- to 400-fold concentration ranges examined and mixture responses were additive. The resulting library of vapor calibration response patterns was used with extended disjoint principal components regression and a probabilistic artificial neural network to develop vapor-recognition algorithms. In a subsequent analysis of an independent data set all individual vapors and most binary mixture components were correctly identified and were quantified to within 25% of their actual concentrations. Limits of detection for a 0.25 l. sample collected over a 2.5-min period were <0.3xTLV for 14 of the 16 vapors based on the criterion that all four sensors show a detectable response. Results demonstrate the feasibility of this technology for workplace analysis of breath and ambient air.
本文介绍了一种小型原型仪器的开发与评估，该仪器采用由四个聚合物涂层表面声波（SAW）传感器组成的阵列，用于快速分析呼出气体和环境空气中的有机溶剂蒸气。仪器内的热解吸吸附剂预浓缩器用于提高灵敏度并补偿背景水蒸气。对置于Tedlar袋中、添加了16种单一溶剂和选定二元混合物的呼气和干燥氮气样品进行了校准。在所研究的50至400倍浓度范围内，响应呈线性，且混合物响应具有加和性。所得的蒸气校准响应模式库与扩展不相交主成分回归和概率人工神经网络一起用于开发蒸气识别算法。在随后对一个独立数据集的分析中，所有单一蒸气和大多数二元混合物成分均被正确识别，且定量结果在其实际浓度的25%以内。根据四个传感器均显示可检测响应的标准，对于在2.5分钟内采集的0.25升样品，16种蒸气中有14种的检测限<0.3倍阈限值。结果证明了该技术用于工作场所呼出气体和环境空气分析的可行性。
