Chemistry Reference
In-Depth Information
maintained even if the fluorescence intensity changes. This is possible if a portion of the total fluorescence
spectrum along a trajectory connecting points of identical intensity (isopotential trajectory) is selected from
an initial point to final excitation and emission wavelengths. In order to ensure a sensitivity level similar to
that of a direct determination in absence of background fluorescence, trajectories must pass by the fluorescence
intensity peaks of both analytes. The performance of this technique can be improved by using derivative
methods (DMISF). It was found that the simultaneous detection of spectrally closely overlapping compounds
is possible using MISF method combined with derivative techniques. The DMISF method has high sensitivity
and selectivity, but at higher noise levels the sensitivity of detection decreases more rapidly. DMISF is capable
to discriminate fluorophores of similar quantum yield being in not more than 1
concentration compared to
interfering high concentration components at low noise level. In practically noiseless circumstances, MISF is
capable to discriminate compounds differing in their concentrations in more than two orders of magnitude
[77]. This method has been applied to the simultaneous direct fluorimetric determination of diflunisal and
salicylic acid in serum [78]. The simultaneous determination is accomplished by using a single isopotential
trajectory that passes by the maxima in the fluorescence bands of diflunisal and salicylic acid in order to
ensure a sensitivity level similar to that of a direct determination in absence of background fluorescence.
Although the MSIF technique suppressed the fluorescence background of serum, the overlap between the
diflunisal and salicylic acid MISF spectra precluded the simultaneous determination of the two compounds.
Satisfactory results could be obtained by measuring the analytical signals at the zero-crossing points in the
first derivative of matrix isopotential synchronous scan. The main advantages of this method include: aqueous
solutions are used over organic solvents required in others methods; it can be directly applied to untreated
serum samples; and limits of detection are similar or lower than obtained in the other described methods [78].
%
13.3.1.4
Combination of multivariate statistical methods with derivative SFS
Multivariate statistical methods, such as principal component regression (PCR) and partial least-squares
(PLS), have been applied in the analysis of mixtures, using mainly infrared or UV-Visible absorption
spectroscopy. Only a small number of methods have been described in which fluorescence data are used for
multicomponent analysis, by multivariate calibration methods [79]. Jones
et al
. [80] reported the use of
excitation, emission and synchronous spectra for the determination of acyclovir and guanine, applying
several multivariate calibration methods. In addition, they investigated the utilization of second-derivative
synchronous spectra in the multivariate calibration methods. They found that the use of synchronous spectral
data was the best choice for the determination of acyclovir, whereas the use of second-derivative synchronous
spectra was the best choice for the determination of guanine. Aghamohammadi
et al
. [81] evaluated the
potential of combining normal, synchronous and derivative fluorimetry with multivariate methods to
determine aflatoxin B
1
(AFB1) in complex matrix of pistachio samples. Normal fluorescence spectra and
their derivatives and synchronous spectra and its first derivative were used separately as analytical signal to
make the calibration models (multiple linear regression (MLR), PCR and PLS1). They concluded that AFB1
determination by synchronous fluorimetry in combination with MLR or PLS calibration showed a better
performance when compared with derivative spectral and normal fluorescence spectra.
13.4
Use of derivative signal techniques in liquid chromatography
One of the most serious problems with chromatography workers are confronted is the occurrence of only
partially resolved peaks arising from coelution of solutes in the samples or from similarities between their
retention times. Traditionally, this type of problem was addressed by modifying the experimental conditions
by trial and error until the aforesaid errors were minimized. Thus, different mobile or stationary phases
