Other Procedures for Error Detection

There may well, on occasions, be people reading this book who wish to make sure that their data is reliable, but who cannot afford either the time delay or the cash outlay needed to acquire a selection of appropriate certified reference materials. Fortunately, there are other useful approaches worthy of consideration which allow detection of spectral, chemical, or transport interferences in analytical flame spectrometry. 

The easiest way to check for the absence of a spectral interference is to reanalyse the sample using one or more different wavelengths for the determination. It is highly improbable that the same wrong answer will be obtained at two different wavelengths, and even less likely that the same erroneous answer will be obtained at three different wavelengths. 

The extent of chemical interferences in flame spectrometry varies with flame conditions and analyte concentration. Thus it is most unlikely that the same wrong answer will be obtained at two different heights in the flame or at two different fuel-to-oxidant ratios. Indeed it has been suggested that the former of these two options may provide automated detection of chemical interferences.2 The burner was moved up and down using a microprocessor-controlled stepper motor. Alternatively, results in air-acetylene and nitrous oxide-acetylene flames may be compared. Similarly, it is unlikely that the same wrong answer will be obtained at two different dilutions. Thus if it is thought that there might be a risk of chemical interference, the determinations on a selection of samples should be repeated under diverse conditions, either on the same or different instruments. 

If a collimated light source is fitted inside the spray chamber, and a photodiode detector is used to measure the amount of light scattered by aerosol droplets inside the spray chamber, it is possible to detect directly changes in aerosol 

The above suggestions do not constitute an exhaustive list, but should provide ideas for sufficient easily applied tests to allow the analyst to have greater confidence in the accuracy of his data. However, if time allows, it is still better, in addition, to analyse authentic or simulated reference materials. 

---