Resumo

ROJAS-VARGAS, Armando; RICARDO-RIVERON, Aymara; OJEDA-ARMAIGNAC, Elaine  e  SERRAT-GUASCH, Nurian. Uncertainty estimation for Cobalt and Iron determinations in paired samples of NiS and NiO by linear regression. RTQ [online]. 2024, vol.44, n.1, pp. 107-121.  Epub 31-Fev-2024. ISSN 2224-6185.

Interlaboratory comparison provides experimental evidence on the lab continuous performance to realize specific tests. In this work, uncertainty of measurements and the interlaboratory error for the chemical analysis of cobalt and iron was determined in paired samples of nickel sulfide (NiS) and sintered nickel oxide (NiO), using the least-squares linear regression. The analytical determinations were made by Atomic Absorption Spectrometry (AAS) with the participation of two laboratories of the corporative group for the production of nickel in Cuba. The assumptions of the linear regression model were verified using the Durbin-Watson statistic, homoscedasticity and normality test. The uncertainty in the slope (Sm), in the intercept (Sb) and the standard deviation of the measurement (Sy) were determined. As a result, the uncertainty (Sx0) associated with the expected value (x0) and error (ξ) in cobalt determination was Sx0 = 0,3 % and ξ = ( 0,60 % in Nickel Sulfide (NiS), and Sx0 = 0,03 % and ξ = ( 0,05 % in Nickel Oxide (NiO). As for iron determination, it was obtained Sx0 = 0,16 % and ξ = ( 0,32 % in NiS, and Sx0 = 0,05 % and ξ = ( 0,09 % in NiO. This method uses the historic database of interlaboratory comparison and provides information to strategic management decisions.

Palavras-chave : interlaboratory comparison; lineal regression; uncertainty.

        · resumo em Espanhol     · texto em Espanhol     · Espanhol ( pdf )