多頻道石英壓電晶體感測器的研製與應用

Preparation and Application of Multi-channel Piezoelectric Crystal Gas Sensor

張平(Ping Chang)

指導教授 ： 施正雄

國立臺灣師範大學/理學院/化學系/博士(1999年)

摘要

石英壓電晶體基於它的共振頻率對於對於晶體表面的質量改變具有非常好的靈敏度，因此將特定的化學物質塗佈在其電極表面作為吸附劑來偵測特定的化學物種，可設計成相當好的化學感測器。
在本論文中，Ru(III)/cryptand 首先被置備來塗佈於晶體表面上，用以偵測非極性的有機氣體。由於非極性的有機氣體因為若僅靠凡得瓦力作為吸附的作用力，將使吸附型的感測器靈敏度不佳且選擇性太差，若使用Ru(III)/cryptand作為塗佈物質則可增進感測靈敏度，且對單鍵、雙鍵、三鍵的的非極性氣體具有選擇性，其訊號強度依序為炔類＞烯類＞烷類。
以石英壓電晶體做為換能器最大的優點在於可隨意的更換塗佈物種而達到偵測各種分析物之目的。多頻道的化學感測器，對於選擇性不佳的情形能夠有相當程度的改善。本研究中利用主成份分析來幫助研究者在眾多塗佈物種中選擇最適合偵測目的的物質來塗佈。十四種塗佈物質被考慮用以偵測常見的有機氣體，其中 polystyrene、polyvinyl alcohol、stearic acid、fullerene、polyethylene adipate和polyvinyl pyrrolidone等塗佈物質在主成份分析中具有代表性被選擇，利用主成份分析法中的成份分數，可將變異量集中在少數的維度中，藉以辨別不同的分析物，例如有機酸、胺類化合物、醇類化合物及芳香族化合物。
類神經網路發展應用也是本研究的重點之一。利用倒傳遞類神經網路，可以幫助來辨識不同氣體分子，而且不經過人為判斷更顯得客觀而易於應用於警報系統。利用訓練組數據，經過網路的訓練學習過程，能夠找到網路參數來作辨識工作。利用多頻道石英壓電感測系統配合類神經網路除了能夠分辨不同的分析氣體之外，能更進一步的得到一個混合氣體的樣品中，可能含有的氣體成份，例如有機酸、胺類化合物、醇類化合物及芳香族化合物等。
經由類神經網路分辨氣體物種後，複回歸分析被應用在定量上，很顯然的，簡單回歸是無法計算混合有機氣體中個別的氣體濃度的，利用多頻道石英壓電偵測系統的感測訊號，配合類神經網路的辨別，最後加上複回歸做濃度的分析，可以直接偵測到混合氣體中目標氣體的濃度，誤差約在5∼20﹪。
整個的實驗系統包含記錄多頻道頻率的即時多頻道系統都是在實驗室自製，可由個人電腦直接控制，直接監測頻率的變化。控制程式及倒傳遞類神經網路的程式則自行以Qbasic程式語言撰寫。主成份分析及複回歸分析則以商用統計軟體SAS計算。

關鍵字

感測器 ； 石英壓電晶體 ； 主成份分析 ； 類神經網路 ； HASH(0x8ab041c)

並列摘要

In this study, piezoelectric quartz crystal detection system with a home-made computer interface was prepared and applied as a gas chromatographic detector. The oscillating frequency of quartz crystal decreases due to the adsorption of organic molecules on the coating. The home-made computer interface includes the digital counting system Altera and Intell-8255 data processing system. Software was written to control the interface and data acquisition.
The Cryptand/Ru(III) coated piezoelectric quartz crystal detection system demonstrated good selectivity and high sensitivity for olefins. The frequency shifts were in the order: 3-hexyne > 1-hexyne > cis-2-hexene > trans-2-hexene > hexane. It means those alkynes seem greater adsorption on Cryptand/Ru than alkenes and alkanes. Effects of temperature and interference of water were also investigated and discussed.
A multi-channel piezoelectric quartz crystal detection system was prepared and employed in the present study to detect various organic molecules from a flow system. By application of principal component analysis (PCA) to a data set containing piezoelectric quartz crystal frequency shifts measured in responses to detect adsorbed on special coating materials, a reduced set of orthogonal factors had been defined to provide maximum discriminating information for all analytes of interest. The data set obtained from 14 piezoelectric crystal sensors for 30 analytes and the first four factors of the reduced set explained almost 90.3% of the variation. Six interested coating materials were selected after compared the correlation between the 14 coating materials and the four factors. Propylamine, Formic acid, Propanol and Toluene can be distinguished clearly by the six-channel piezoelectric detection system without considering the retention times of the analytes.
Back propagation neural network (BPN) was used to distinguish the species in the organic mixture and multivariate linear regression analysis (MLR) was used to compute the concentration of the species. A six-channel piezoelectric sensor detecting organic molecules in static system was investigated and discussed. Amine, carboxylic acid, alcohol and aromatic molecules can easily distinguished by this system with back propagation neural network. Furthermore, the concentrations of the organic compounds were computed with about 5-20% error by multivariate linear regression analysis (MLR). Organic Mixture with amine, carboxylic acid, alcohol and aromatic molecules detecting by this way also had good qualitative and quantitative results. To have better distinguishability, changing the fault-tolerance in back propagation neural network was also investigated and discussed in this study.

並列關鍵字

sensor ； QCM ； PCA ； ANN ； BPN

參考文獻

42. M. T. Gomes, T. A. Rocha, Armando C. Duarte and Joao P. Oliveira, Anal. Chem., 1996, 68, 1561

1. Y. Okahata, and T. Tsuruta, Langmuir, 1988, 4, 1373-1375

2. K. Sode, and Y. Kameda, Anal. Chem. Acta, 1980, 220, 251-255

3. M. R. Deakin, and O. R. Melroy, Electrochem. Soc., 1989, 136(2), 349-352

5. H. Muramatsu, Anal. Chem. Acta, 1989, 225, 339-480