Modification of Philibert–Tixier ZAF correction for geological samples

Two steps are usually used in electron microprobe correction: (1) development of a theoretical mode based on physical phenomena within pure or binary targets; and (2) simplification of certain complexities due to multi‐element samples by using average coefficients in the theoretical equations. It is proposed to use a more detailed method of estimating the general correction coefficients concerning A/Z relationships in the absorption and ionization terms of the theoretical formulation. In the absorption formulation the factor h is taken as (1.2 ∑ C i A i /Z i 2 )/ (1 + K a ). The value of a is given as ‐1. In the ionization relationship M = (∑ C i Z i /A i )/(1 + K b ) with b = E 0 /100, where E 0 is the accelerating voltage in kV. K , deduced from experimental results, is given as K = ∑ C i ( Z i ‐ Z¯ ), where Z¯ is the average atomic number of the target. These values are derived empirically through direct measurements on samples whose correction properties induce normally strong analytical errors: oxides with 11 ⩽ Z ⩽ 14 analysed within a heavy matrix.