表面聲波有機氣體感測器之研製與應用

Abstract

表面聲波有機氣體感測器之研製與應用 摘要 製備不同的表面聲波one-port 315MHz感測器以偵測不同有機揮發性氣體。在本研究中可分為三個部分:首先,發展銀(I)/大環胺醚-22的表面聲波氣體感測器偵測不同的烯油類,如:烯、烴。接著,一個多頻道的表面聲波感測器結合主成份分析和線性判別分析發展出偵測不同的極性與非極性的有機分子。最後,倒傳遞的神經網路與多變量的線性迴歸分析結合多頻道的表面聲波感測器可以發展偵測不同有機氣體的定量與定性分析。 銀/大環胺醚-22的表面聲波感測器可以應用發展成為氣相層析儀的一種偵測器,以偵測不同有機氣體,如:烷、烯、炔等。此銀/大環胺醚-22的表面聲波氣相層析偵測器對1-己烯(1-hexene) 和1-己炔(1-hexyne)在不同濃度對應訊號值時呈現良好的線性關係值。比較表面聲波(SAW)和石英微量天平(QCM)兩者間的靈敏度,發現表面聲波的效果遠勝於石英微量天平。對同6個碳數的烴類(烷、烯和炔)則偵測結果為1-己炔>1-己烯>>1-己烷,而感測七個碳數的順、反異構物,則可得順-2-庚烯>反-2-庚烯和順-3-庚烯>反-3-庚烯。對烯、炔莫耳質量的探討時也觀察出表面聲波偵測器對烴類有極好的再現性和偵測極限,例如:順-3-庚烯的偵測極限為0.6 mg/L。並比較表面聲波偵測器和傳統氣相層析儀熱導電偵測器兩種感測的效能,表面聲波的感測效果優於熱導電偵測器。另外,還探討氣體流速和溫度對此表面聲波偵測器的影響。 在本研究中製備多頻道表面聲波感測器偵測不同類的有機氣體,藉著SAS軟體來操作主成份分析並選出合適的塗佈物塗佈於表面聲波感測器上偵測有機氣體,氣體如:己烷、1-己烯、1-己炔、1-丙醇、丙醛、丙酸、1-丙胺。在多頻道的表面聲波感測系統中以19種不同的塗佈物偵測上述七種有機氣體,經SAS軟體操作後選出前六個主成份分數,利用主成份分數散佈圖可以分辨七種氣體。除此之外,利用SPSS軟體對上述的這些氣體做線性判別分析,也可達到100%的定性辨識效果。將選出的六種塗佈物18C6、Cr3+/cryptand-22、stearic acid (SA)、polyvinyl pyrrolidone (PVP)、triphenyl phosphine (TPP)、polyethylene glycol (PEG)對七種有機氣體做雷達圖分析,而雷達圖也可以清晰辨識這些氣體。不同有機氣體濃度的效應對應多頻道表面聲波偵測系統的訊號也在本實驗中被觀察和探討。 倒傳遞類神經網路也可以用來確認上述七種氣體,它不但可以辨識不同的單一氣體還可以區分混合氣體,有關倒傳遞神經網路系統的學習速率和隱藏層單元也被觀察研究。最後,多變量的線性迴歸分析結合多頻道表面聲波感測器可應用於這些氣體的定量分析且可得小於10%的誤差,多變量線性迴歸分析還可以決定混合氣體1-己烯、1-己炔和丙醛它們各別的含量。
Preparation and Application of Surface Acoustic Wave Organic Gas Sensors Abstract Various one-port 315 MHz surface acoustic wave (SAW) sensors were prepared for various organic vapors. There were three parts in this study. Firstly, the Ag (I)/cryptand-22 SAW gas sensor was developed for various olefins, e.g. alkene and alkynes. Secondly, a multichannel SAW gas sensor with the principal component analysis (PCA) and linear discrimination analysis (LDA) was developed for various polar and nonpolar organic molecules. Finally, the back propagation neural network (BPN) and multivariate linear regression analysis (MLR) with the multichannel SAW sensor were developed for the detection of various organic vapors qualitatively and quantitatively. The Ag(I)/cryptand-22-coated sensor was applied as a GC detector to detect various organic vapors, e.g. alkanes, alkenes and alkynes. The Ag (I)/cryptand-22-coated GC-SAW detector exhibited linear response to 1-hexene and 1-hexyne. The comparison between SAW and quartz crystal microbalance (QCM) sensors demonstrated that the SAW sensor exhibited much better response than the QCM sensor. The frequency responses of the Ag(I)/cryptand-22-SAW sensor to normal C6 hydrocarbons were in the order of 1-hexyne >1-hexene >>hexane, while responses to various heptene isomers were in the orders of cis-2-heptene > trans-2-heptene and cis-3-heptene > trans-3-heptene. The molar mass effects examined for alkynes and alkenes indicated that the SAW detector was quite sensitive to the hydrocarbons with good reproducibility and quite low detection limits, e.g. 0.60 mg/L for cis-3-heptene. The performance of the Ag (I)/cryptand-22 SAW GC detector was well comparable to the commercial thermal conductivity detector(TCD). The effects of flow rate and temperature on the response of the detector were also investigated and discussed. The prepared multi-channel SAW detection system was employed to detect various organic molecules in this study. The principal component analysis (PCA) method with SAS software was applied to select the appropriate coating materials onto the SAW crystals for organic vapors, e.g. hexane, 1-hexene, 1-hexyne, 1-propanol, propionaldehyde, propionic acid, 1-propylamine. A dataset for multi-channel sensor with 19 SAW crystals for 7 analyses was collected after comparing the correlation between the 19 coating materials and the first six principal component (PC) factor. The principal component analysis scores map could distinguish 7 gases. Furthermore, linear discriminate analysis (LDA) with SPSS software was also used to detect these organic vapors qualitatively which could reach 100% discrimination for these gases. The profile discrimination maps were applied for the discrimination of these organic vapors. These organic molecules could be clearly distinguished by a six-multichannel surface acoustic wave (SAW) detection system with coating materials, e.g. polyethylene glycol, 18 crown 6 (18C6), Cr3+/cryptand22, stearic acid, polyvinylpyrrolidene and triphenyl phosphine. The effect of concentration for various organic vapors the responses of the multichannel SAW detection system was also investigated and discussed. The artificial back propagation neural (BPN) network was also used to recognize various organic gases. It showed not only the distinction of unity organic vapors but also mixture gases. The effects of learning rate and the hidden unit of neural network system for BPN analysis were investigated. Finally, the multivariate linear regression analysis (MLR) with the multichannel SAW sensor was also applied to quantitatively detect these organic vapors with <10% error. The MLR analysis was also applied to determine the concentration of each component in a mixture of 1-hexene, 1-hexyne and propionaldehyde.

Description

Keywords

表面聲波, 有機氣體, 主成份分析, 倒傳遞類神經網路

Citation

Collections