< back to main site

Publications

Using principal component analysis in the evaluation of SERS-active substrates.

Tantra, R; Brown, R J C; Milton, M J T (2007) Using principal component analysis in the evaluation of SERS-active substrates. In: Abstract Book of Colloquium Spectroscopicum Internationale XXXV, 2007, Xiamen, China.

Full text not available from this repository.

Abstract

The drive towards improved SERS reproducibility has resulted in a wide variety of currently available enhancing substrates, ranging from metallic colloids to solid metal supports. Substrate performance can be affected by the method of preparation, adding complexity to the issue of reproducibility. Clearly, there is a need to have a common platform to graphically display and assess the relative performance of various SERS substrates to each other; such a platform can potentially be used to establish control limits for fit-for-purpose SERS substrates. Principal component analysis (PCA), a ubiquitous technique for data analysis (with the capacity to examine data objectively, rapidly and precisely) has considerable potential to be used as a tool to summarise classification performance of various SERS substrates. In this study, PCA was used to visualise the relations between different sources of SERS gold colloids; classification was handled through the use of identifying clustering patterns between data replicates. PCA was able to distinguish successfully between colloids that were made `in-house¿ at NPL and those that were commercially purchased (by comparing their SERS response to Rhodamine 6G). Colloids made in-house were shown to give best reproducibility. Within the context of PCA analysis, the term reproducibility refers to sources of data variations identified by individual principal components (PCs). The associated loadings results for the PCs suggest that the main variations in the dataset are dominated by certain spectral peaks. Data variations surrounding these bands may be due to variations in the electromagnetic SERS enhancement experienced by the adsorbed analyte, which in turn may be influenced by the size and shape of metal clusters formed during colloid aggregation events.

Item Type: Conference or Workshop Item (UNSPECIFIED)
Subjects: Analytical Science
Analytical Science > Trace Analysis and Electrochemistry
Last Modified: 23 Jul 2018 13:02
URI: http://eprintspublications.npl.co.uk/id/eprint/4042

Actions (login required)

View Item View Item