2011_07_12_Spectrofluorimetric determination of chlorophylls and pheopigments using parallel factor analysis
Title and Authors
Title: Spectrofluorimetric determination of chlorophylls and pheopigments using parallel factor analysis
Authors: Ludvig Moberg1, George Robertsson2, and Bo Karlberg3,*
1Department of Analytical Chemistry, Stockholm Uni6ersity, SE-10691 Stockholm, Sweden
2Department of Analytical Chemistry, Stockholm Uni6ersity, SE-10691 Stockholm, Sweden
3,*Department of Analytical Chemistry, Stockholm Uni6ersity, SE-10691 Stockholm, Sweden
SUMMARY OF PAPER
INTRODUCTION
Chlorophyll a,b, and c (chl a, chl b, and chl c) is correlated with phytoplankton and so to primary productivity and biomass. So, it needs to be investigated in marine environments. Their degradation products pheophytin a, b, and pheoporphyrin c are also present with them. Chl b, c, and degradation products provide information about algae. Interference of excitation emission wavelengths is the major problem which contradicts the results from actual analysis, like the analysis of pure analytes obtained from different techniques like high performance liquid chromatography. Analysis of excitation emission plots and the results obtained show that some of the analytes are either overdetermined or underdetermined by the presence of other analytes. It means two or more analytes may have the same excitation emission spectra but with different intensities. So, the peak with lower concentration is suppressed by the presence of other peak with higher concentration.
Several methods have been used to find the number and amount of analytes in the samples. But, each method had its own flaw like suppression of some analyte or even absence of some analyte (which was actually present) etc. In this study, PARAFAC was used to decompose the analytes present in the samples.
EXPERIMENTAL
1 mg of chl a and 1 mg of chl b were dissolved in acetone water (9:1) and diluted to 100ml. phe a and phe b solutions after preparation were kept in freezer at -20oC. chl c and phe c were prepared and stored by a method different from those of chl a, b, phe a, and b. Stock solutions were prepared from these chls and phes. Concentrations in ?gL-1 for standards N1 ? N9, N10 ? N13, and V1 ? V5 are summarized in table 1.
Table 1
Standard    Analyte
    Chl a    Chl b    Chl c    Phe a    Phe b    Phe c
N1    10    10    10    75    10    4.8
N2    25    10    50    25    5    28.8
N3    25    50    10    10    30    28.8
N4    75    10    5    50    30    9.6
N5    25    5    30    50    10    28.8
N6    10    30    30    25    30    4.8
N7    50    30    10    50    5    64
N8    50    10    30    10    50    64
N9    50    50    5    25    10    64
                        
N10    70    0    0    10    0    0
N11    10    0    0    10    0    0
N12    70    0    0    70    0    0
N13    10    0    0    70    0    0
                        
V1    50    50    50    50    50    48
V2    100    10    10    100    10    9.6
V3    30    5    5    100    30    9.6
V4    40    40    0    20    20    0
V5    60    0    30    30    0    28.8

The samples were excited with visible light in the range of 360 to 500 nm with an interval of 2 nm whereas the emission wavelength range was 600 to 730 nm with a 2 nm interval. Chromatography was performed with an HPLC. PARAFAC was performed using MATLAB.


RESULTS AND DISCUSSION
Figure 1 below shows the peaks (sample V1) of different analytes except for phe c which is suppressed for these concentrations of analyetes.
 
Figure 1
PARAFAC model was fitted 10 times and the model calculations were based on nine standards. Core consistency was used to determine the number of components. 6 components were selected based on core consistency. The signal of the sixth analyte (phe c) was found very weak. Correlation coefficient between true and PARAFAC component was found higher than 0.97 whereas the coefficient for phe c was 0.79. The number of components and their concentrations were found to be matching with the true values.