Fluorescence

 

Problem

This data is made from known analytes. They can be used for second order calibration tests, and for other concepts of interest to test on a fluorescence data set. It can easily be divided into smaller units if the user so desires. The samples were generated and measured by Åsmund Rinnan (KVL, DK) and Jordi Riu (Universitat Rovira i Virgili, Spain).

 

Get the data (28 MB)

 

The data are available in the data set format defined by the Dataset Object freely available at Eigenvector Research and also part of the PLS_Toolbox, and was made in Matlab version 6.5. If you do not have the Dataset Object or PLS toolbox, the data will simply be in a structure format. There are two structures. One with the EEMs, the other with the concentration profiles. Download the data and write “load fluordata" in MATLAB. If you use the data we would appreciate that you report the results to us as a courtesy of the work involved in producing and preparing the data. Also you may want to refer to the data by referring to

  1. Rinnan, Å.: Application of PARAFAC on Spectral Data, 2004, PhD thesis, Royal Veterinary and Agricultural University (DK)
  2. Bro, R, Rinnan, Å, Faber, K.: Standard Error of Prediction for Multilinear PLS. 2. Practical Implementation in Fluorescence Spectroscopy, Chemometrics and Intelligent Laboratory Systems, 2004, Accepted

 

Data

There are a total of six different fluorophores in the dataset: catechol, hydroquinone, indole, resorcinol, tryptohpane and tyrosine. These were chosen on the basis of closeness to the 1st order Rayleigh scatter line and their overlap in both emission and excitation spectra. In total 405 samples are recorded.

Figure 1: The pure excitation and emission spectra for the six analytes.

Out of these six analytes, 12 datasets where measured (see Table 1). All of the datasets are more or less equal, only with small differences among them, described in the column to the right. An ‘x’ in columns 2-7 indicates if the fluorophore is present or not in the dataset.

 

Table 1: A short description of the 12 datasets.

Dataset

Fluorophores

Description

Catechol

Hydroquinone

Indole

Resorcinol

Tryptophane

Tyrosine

0

x

x

x

x

x

x

Pure standards

1

 

 

 

x

x

x

Tyrosine kept constant, while factorial design on Resorcinol and Tryptophane

2

x

x

x

x

x

x

All samples have 2 or 3 fluorophores

3

x

x

x

 

x

 

Catechol and Hydroquinone vary equally

4

x

x

x

 

x

x

All samples have 2 or 3 fluorophores

5

x

 

 

x

 

 

Test of Catechol and Resorcinol

6

x

x

x

 

x

x

All samples have 3 or 4 fluorophores

7

x

x

x

 

x

x

Low concentrations

8

x

x

x

 

 

x

 

9

x

x

x

 

 

x

Catechol concentration slightly lower than previous max (possible non-linear on the previous maximum concentration level)

10

x

x

x

 

x

x

All samples have 2 or 3 fluorophores

11

x

x

x

 

 

x

Replicate of dataset 8 and 9

 

Instrumentation

The instrument used for the analysis was a Varian Eclipse Fluorescence Spectrometer. The instrument setup was optimized to give satisfying resolution within a limited time. It took roughly 12 minutes per sample (including all the sub-sampling). In Table 2below, the instrument setup is shown. Sub-samples state that the sample was left in the instrument and scanned five consecutive times. Every measurement consists of 136 emission wavelengths × 19 excitation wavelengths.

 

Table 2: Instrument setup

Instrument settings

Value

Slit width

5 nm (both Emission and Excitation)

Excitation

230-320 (nm) - Every 5 nm, Highest wavelength first

Emission

230-500 (nm) - Every 2 nm

Scan rate

1920 nm/min

PMT Detector voltage

600V

Sub-samples

5

 

General note

In seems that Catechol contains some impurities, and thus gives rise to an extra component in the data sets where it is present.