Interferences and Errors Associated with the Excitation Process
Problems may arise from the different volatility of elements in the sample which will in turn be
modified by the matrix. Fractional distillation can occur and a serious error result unless the sample is
completely burned, because the emission intensity for each element will vary independently with time
(Figure 8.5(a)). On the other hand, this effect may well be exploited as in the case of volatile impurities
in a non-volatile matrix. By recording the spectrum only in the early part of a burn, they may be
detected with little interference from the matrix. The pre-spark routine discussed earlier (p. 290) may
also be used to improve precision.
Another aspect of the volatility problem concerns the sensitivity of the method – the greater the
volatility of a material the shorter its residence time in the discharge. It follows that volatility and
sensitivity will be related and that the lowest detection limits will be associated with low volatility.
Excessively rapid volatilization will introduce a serious error by self absorption. The high concentration
of the element in the cooler part of the discharge will absorb emitted radiation from the core, thus
depressing the intensity of the line. In exceptional cases, there may even be self-reversal of the line
giving a minimum in the intensity-wavelength profile (Figure 8.5(b)). Self-reversal is also observed
with intense lines from major constituents of a sample. Volatility differences present a problem as they
arise not only from variations in the properties of the analytes, but also from chemical reactions taking
place between analyte and matrix or the carbon of the electrodes. A further problem is the formation of
cyanogen radicals, CN, which provide interference from molecular spectral bands. In earlier years much
ingenuity was employed in minimizing complex interference effects but they remain a problem leading
to relatively poor precision and limits of detection. Dissatisfaction with arc/spark analysis on these
grounds is the primary reason for the displacement of arc/spark emission by more recently developed
techniques such as atomic absorption and ICP-AES,
Figure 8.5
(a) Variations of line intensity during burning of sample; element 1 is
more volatile than 2.
(b) Energy-wavelength profile and self-absorption.