Interval Estimates of the Total Content of Antioxidants for Different Analytical Assays

Abstract

Determination of the total content (с_Σ) of similar analytes recalculated to a standard substance X_st is a commonly used but a metrologically incorrect measuring procedure, leading to significant systematic errors (B-type uncertainty). Interval estimate с_Σ (Vershinin et al., 2016) is an alternative method that does not require recalculation to X_st and, in contrast to calculating total indices, slightly depends on the nature and ratio of the sought-for group components present in the sample. Such estimates are used for evaluating the total content of antioxidants (AOs) of foodstuffs. However, the dependence of these estimates on the choice of the group reagent and the signal measuring procedure is uncertain. To address this question, model mixtures of AOs with known с_Σ values, ranging from 10^–5 to 10^–4 mol/L, were prepared and tested. Generalized signals were measured by spectrophotometry using Folin–Ciocâlteu (F–C) and FRAP assays, and traditional and interval estimates of с_Σ were calculated and compared. The true values of с_Σ were within the respective intervals for all model mixtures; in the case of the F–C assay, the intervals were wider and shifted towards higher с_Σ values. The FRAP assay was modified using AO concentrations in mol-eq/L, reducing reaction time, and substituting the auxiliary reagent. This modification successfully brought the sensitivity coefficients of individual AOs closer to one another and made the relative widths of intervals three times narrower. Modified FRAP assay was used for the group analysis of wines, black tea infusions, and juices. Corresponding interval estimates of с_Σ were obtained and compared. The unresolved issues and future research directions related to interval estimates are discussed.