染料與顏料之拉曼光譜資料庫的建立與應用

Abstract

本研究使用了兩種基於不同原理撰寫的光譜資料庫，分別為均方根誤差和交互相關係數原理，作為評估準確性和相關性的衡量指標。由於每次測量的訊號強度和位移間距都各不相同，因此通過對光譜進行正規化處理，成功解決了光譜不一致的問題，可以檢索和比對多種常見的有機染料和無機顏料。為了應對不同狀態的顏料樣品，使用了三種不同的前處理方式。這些樣品通常呈現不同的形式，例如包裝好的水彩顏料、上色後的畫作以及化學合成的染料等，因此針對這些樣品採用了相應的檢測方法。透過使用不同的前處理方法，結合表面增強拉曼散射技術可以對一些無法直接使用拉曼光譜儀測量到的樣品進行分析。除了驗證了這些方法的可行性之外，同時結合資料庫的使用，比較兩個資料庫的準確度，從而實現了更快速的鑑定結果，為顏料的鑑定提供了更多的選擇性。透過在實際樣品上的成功應用，證明了該資料庫在不同顏料和染料比對上的有效性，同時也展示了在表面增強拉曼技術方面的適用性。

Two spectral databases built on different principles, namely the root mean square error and the cross-correlation coefficient principles, were used as measures to evaluate accuracy and correlation. Since the signal intensity and displacement spacing of each measurement are different, the problem of spectral inconsistency was successfully solved by normalizing the spectra, allowing for the retrieval and comparison of a variety of common organic dyes and inorganic pigments.In order to deal with pigment samples in different states, three different pre-treatmentmethods are provided. Considering that the pigment samples that need to be identified usually come in different forms, such as packaged watercolor paints, colored paintings, and chemically synthesized dyes, we provide corresponding detection methods for these samples. By using different pre-treatment methods, combined with surface-enhanced Raman scattering technology, some samples that cannot be measured directly using a Raman spectrometer can be analyzed.In addition to verifying the feasibility of these methods, it also combines the use of databases to compare the accuracy of the two databases, thereby achieving faster identification results and providing more selectivity for the identification of artworks. The successful application on actual samples demonstrates the effectiveness of the database in comparing different pigments and dyes, and also demonstrates its applicability in surface-enhanced Raman technology.