Error propagation in normalization of stable isotope data: a Monte Carlo analysis

Abstract

A higher analytical precision of a stable isotope ratio mass spectrometer does not automatically guarantee accurate determination of the true isotope composition (delta-value) of samples, since estimates of true delta-values are obtained from the normalization of raw isotope data. We performed both Monte Carlo simulations and laboratory experiments to investigate aspects of error propagation during the normalization of carbon stable isotope data. We found that increasing both the number of different reference standards and the number of repetitions of each of these standards reduces the normalization error. A 50% reduction in the normalization error can be achieved over the two-point normalization by either analyzing two standards four times each, or four standards two times each. If the true delta-value of a sample is approximately known a priori, the normalization error may then be reduced through a targeted choice of locally optimal standards. However, the difference in improvement is minimal and, therefore, a more practical strategy is to use two or more standards covering the whole stable isotope scale. The selection of different sets of standards by different laboratories or for different batches of samples in the same laboratory may lead to significant differences in the normalized delta-values of the same samples, leading to inconsistent results. Hence, the same set of standards should always be used for a particular element and a particular stable isotope analytical technique.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 78 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.