Multicomponent analysis using an array of piezoelectric crystal sensors
Multicomponent analysis using an array of piezoelectric crystal sensors
复制标题
使用压电晶体传感器阵列进行多组分分析
DOI：
--
复制
发表时间：
1987
期刊：
影响因子：
--
通讯作者：
B. Kowalski
中科院分区：
文献类型：
--
作者：
W. Carey;Kenneth R. Beebe;B. Kowalski
一个阵列的9个压电石英晶体，每个涂有不同的部分选择性涂层材料，被构造用于有机蒸气的多组分分析。该阵列的有用性进行了评估，通过定量已知的混合物样品在两个和三个组件的情况下，使用两种校准技术，多元线性回归（MLR）和偏最小二乘法（PLS）。在使用微传感器（例如涂覆的压电晶体）的情况下，传感器之间可能存在高度的共线性，这对回归结果有影响。在两个组件的情况下，PLS方法产生了4- 6倍的改进，预测能力超过MLR。虽然单个传感器在响应中产生3-6%的相对误差，但是当使用传感器阵列和PLS方法时获得的平均相对预测误差在两个组分的情况下为4.6%，而在三个组分的情况下，共线性的影响降低了预测能力。
An array of nine piezoelectric quartz crystals, each coated with a different partially selective coating material, was constructed for multicomponent analysis of organic vapors. The usefulness of this array was evaluated by quantitating known mixture samples in both two and three component cases using two calibration techniques, multiple linear regression (MLR) and partial least squares (PLS). With the use of microsensors, such as coated piezoelectric crystals, a high degree of collinearity between the sensors may exist, which has an effect on the regression results. In the two component cases, the PLS method yielded a 4- to 6-fold improvement in prediction capability over MLR. Although the individual sensors produce a 3-6% relative error in response, the average relative prediction error obtained when using an array of sensors and the PLS method was 4.6% in the two component cases, while in the three component case the effect of collinearity decreases prediction capability.