Chemistry Reference
In-Depth Information
board so that the flow rate could be controlled. The distance between the menisci in the sample and buffer
vials was 1.5 cm, and the calculated flow rate was 0.88 nl s
-1
. The concept was proven using the cationic dye,
Methyl Violet and the neutral dye, Martius Yellow.
The potential of separation methods that counter hydrodynamic flow with electrophoretic migration seems
very promising for field analysis because it greatly simplifies the instrumentation and increases its overall
robustness. However, the type of samples that can be determined by this technique might be limited.
9.3
Possible ways of surmounting the disadvantages of CE
It has been mentioned above that some reluctance on the part of analysts to use CE is due to poor reproducibility
of peak migration times and areas. This is illustrated in Figure 9.3. In this example, an electropherogram of a
standard of phosphonic acids solution (Figure 9.3, a, i) is provided together with an electropherogram of an
extract of the same solution applied to sand (Figure 9.3, a, ii). It is evident that there is some resemblance
between these two electropherograms but the peaks of the standards can hardly be recognized among the
matrix peaks. The identification and quantification of pollutants is almost impossible. The reason is hardly
surprizing: the electropherograms were recorded by different analysts, using a portable custom-made
instrument without column temperature stabilization. Sampling was performed manually by means of a
syringe that forced a fixed volume of a sample to flow past the separation capillary inlet (see Figure 9.4).
Obviously, the results are not encouraging. Simplifying construction of the POC instrument to make it more
robust (by sacrificing a thermostat and making the sampling procedure more suitable for field operations)
decreases performance. Ways to improve it must be sought.
One way to improve the performance of portable CE is to look for more sophisticated and systematic
approaches to data processing that take advantage of the algorithms that modern software can provide. This
might appear to be making up data, but in fact it is not. By knowing the model of the signal, reliable corrections
can be made to the raw data using mathematical calculations. For example, base line correction is included in
most chromatographic software. Much more elaborate procedures can be developed. Moreover, using
sophisticated software to improve an analytical method without changing the instrumentation is obviously a
thoroughly green approach to analysis. This idea has existed since the first attempts to mathematically
separate chromatographic peaks were made in the 1970s. But it has not achieved the type of success needed
to impress a chromatographer who prefers to develop better columns. The appearance and widespread use of
diode array and mass-spectrometric detectors can significantly improve the prospects of using software to
improve analytical signals. Whether it is too optimistic to speak of 'mathematical' chromatography [57] is
difficult to judge at the moment - a column is still necessary - but it does appear that increasing the level of
sophistication of data processing enables the complexity of an instrument to be reduced, in clear conformity
with the goals of Green Analytical Chemistry.
Using software to improve the reproducibility of migration times has been of recent interest to several
researchers [58] who have studied the use of CE in metabolomics. The nonreproducibility of migration times
threatens the application of chemometrics to the sets (e.g. of samples of body fluids). Various algorithms have
been proposed [59]. The characteristic feature of these algorithms is their capability to process a large number
of electropherograms without human intervention. These elaborate algorithms require a representative
reference electropherogram and the rest of the set of recorded chromatograms must be very similar to the
reference electropherogram. This approach is not promising for field analyses where the composition of
the samples varies. One approach in this case is to mix the sample solution with known components of which
the migration time is outside the window of those of the analytes of interest and can therefore serve as
anchors. Assuming that the axis of deformation of the migration time is linear during the analysis process, the
axis can be corrected by software through simple linear interpolation. Contrary to metabolomic studies, the
