Selection and optimisation of furnace operating conditions

In principle there are no differences between calibration procedures for flame and ETA methods although the latter case will take longer, as has already been pointed out. Calibration standards should match the samples as nearly as possible with respect to major components otherwise standard additions must be used. Indeed the method of standard additions will have to be used at some time to check accuracy, so this procedure will usually be tried first. If the standard additions graph is parallel to the direct calibration graph then freedom from matrix interferences would be indicated. The standard additions principle as applied to ETA is now described. 

For more complex solutions and samples where it is not possible to remove the matrix during the ashing step it may be necessary to use the method of standard additions. Some workers advocate that this method should always be run initially to check for interference effects so that the best calibration procedure can be selected. If the standard additions graph and the direct calibration graph were parallel, freedom from interferences in the sample would be indicated, i.e. the element is in the same form in sample and standard immediately before atomisation, or the two forms give the same absorption response. 

It is advisable that all readings be obtained in duplicate or triplicate. There are again variations of the procedure. 

The validity of the method of standard additions depends on the forms of the analyte element in the sample and in the added standard responding in the same way during the atomisation step. This may not always be so in practice. For example, in determining lead in whole blood by this method it would have to be proved (or assumed) that lead added as a lead nitrate solution is atomised to the same extent as lead bound organically in the sample. This is most easily checked by running a certified standard material through the procedure. If no such reference standard exists, as with the above blood example, the sample should also be run after pre-treatment for removal of the matrix, e.g. wet or dry ashing, and the results compared. 

Two further limitations apply to the use of the method of standard additions. Simultaneous background correction must be employed because of possibly varying amounts of matrix material present in the tube during the several firings needed to make one determination. Furthermore, all readings must be within the linear portion of the calibration graph in order 

---

The selection and optimisation of furnace operating conditions will be dealt with in more detail in section VIIA of the next chapter (page 57). 

---