以全管柱偵測法探討吸附型管柱中之樣品衝破

Abstract

吸附層析法(adsorption chromatography)，利用固定相與不同物質分子吸附能力的差異而達到混合物的分離效果，其過程為移動相分子與物質分子競爭固定相吸附中心，通常用於預濃縮樣品與製備型分離樣品。吸附型層析是科學史上第一種層析技術，直到今日，此技術仍被大量使用，但一般研究所注重的卻僅限於濃縮與分離的結果，至於管內樣品的吸附機制與濃度分布則無深入研究。本研究以全管柱偵測 (whole-column detection, WCD) 系統，搭配可透光之玻璃管柱，來觀察樣品亞甲基藍與甲基橙之吸附，並和Pai所提出的郵包模型(parcel model)理論互相比較。 研究中發現，若把管柱全長分為n段，隨著低濃度樣品持續注入管柱，樣品雖是如肉眼所見的由第一段開始載入至第n段，但以WCD偵測時，樣品並不會先載滿第一段後才開始載入第二段，而是會受到樣品的動態吸附平衡常數、樣品濃度、流速、溶劑種類之影響。 研究也發現，理論圖形和實驗圖形之誤差和n值有關。當n值設定為10時，經由郵包模型所模擬出的理論圖形會與實驗圖形有約23.01%之誤差，當n值為20時，則理論圖形與實驗圖形將有42.58%之誤差，若n值提高到100，理論圖形甚至完全背離實驗圖形。 除了流速提高會造成單位體積樣品的載入速度較快外，管內吸附圖形亦會隨著動態吸附平衡常數之不同而產生大幅改變；若樣品濃度越高，則因有效管柱容量增加及聚合體的形成，而使得單位莫耳數樣品載入速度較慢；且由於水相對於乙醇具有較低的溶劑強度，會造成樣品載入速度較慢，因此，當溶劑中的乙醇比例改變時，樣品於管內空間的訊號將產生劇烈變化。

Adsorption chromatography enables separation of mixtures by utilizing difference of adsorption affinity on different compounds. Analytes compete with the mobile phase of the adsorptive active sites on the stationary phase. This technique has been mainly applied in sample enrichment or preparative separations. Being the first chromatographic technique in the history, it has still been heavily applied up-to-date. However, the researches concerning separation mechanism and the on-column analyte distribution profile have been largely ignored. In this study, we analyzed adsorption of methylene blue and methyl orange on the octadecylsilane packed in a glass column using whole-column detection (WCD). The results were compared with the parcel model proposed recently by Pai. The column was artificially divided into n sections. While the sample of low concentration was injected into the column, it appeared that the sample would fully occupy the first section before move to the second section, observed by naked-eyes. However, sample molecules, monitored by WCD, were found to penetrate the second section even the adsorption had not reached saturation in section one. The mechanism was dependent on the dynamic adsorption constant, sample concentration, and the solvent type. The deviation between the simulated data using the parcel model and the experimental data was n-dependent. A 23.01% deviation was found for n = 10, 42.58% for n = 20. Simulated with n = 100 it deviated completely from the empirical data. Higher flow rates would affect the adsorption profiles when the sample concentration remained the same. The on-column adsorption profiles might change drastically with the changing dynamic adsorption constants of the samples. Larger effective column capacity and aggregate (such as dimers and trimers) formation of the sample with higher concentration slowed down the sample loading onto the stationary phase. The water, with a relatively lower solvent strength, would also decrease the sample loading rate. Thus, the on-column adsorption profile appeared significantly different from that of the sample prepared in water with ethanol .