Pros and cons of the technique of processing

In isotope ratio mass spectrometry (IRMS), any sample (S) measurement is performed as a relative-difference (δi) from a working-lab-reference (W), but the result is evaluated relative to a recommended-standard (D): δi. It is thus assumed that different source specific results (δi, δi ...) would represent their sources (S1, S2 ...), and be accurately intercomparable. However, the assumption has never been checked. In this manuscript we carry out this task by considering a system as -IRMS. As “δi → δi” conversion is a post-measurementtheoretical task, we here examine the designs of typical δ-scale-conversion processes, and clarify how accuracy of δi can be ascertained on measured-data (δi ...). We present a model for a priori predicting output-uncertainty leading to the selection of a proper evaluation-method. Our study shows that scale-conversion, even with the aid of auxiliary-reference-standard(s) Ai(s), cannot make “δi” free from W used for measurements; and the “δi δi” conversion-formula normally used in the literature is invalid. Besides, the latter-relation has been worked out, which leads to e.g., fJ([ ± p%],[ ± p%],[ ± p%]) = ( ± 4.5p%); whereas FJ([ ± p%],[ ± p%]) = ( ± 1.2p%). That is, contrary to the general belief (Nature 1978, 271, 534; Anal. Chem. 2006, 78, 2439; RCM. 2007, 21, 3006), the scale-conversion by employing one than two Ai-standards should ensure “δi” to be more accurate. However, a more valuable finding is that the transformation of any δ-estimate into its absolute value helps improve accuracy (viz.: [ ± 1.2p%] → [ ± 0.014p%]), or any