Portal de Periódicos da CAPES

Artigo Acesso aberto Produção Nacional Revisado por pares

BIBTEX RIS

Accurate automatic titration procedure for low sharpness and dichroism in end point detection using digital movies as detection technique

2017; Elsevier BV; Volume: 133; Linguagem: Inglês

10.1016/j.microc.2017.04.041

1095-9149

Lucas A. Siqueira, Iara S. Nunes, Pedro Lemos de Almeida, Wellington da Silva Lyra, Renato Allan Navarro Andrade, Mário César Ugulino de Araújo, Luciano F. Almeida, Ricardo A. C. Lima,

Microfluidic and Capillary Electrophoresis Applications

This study addresses the challenge of performing fast and accurate titrations with end points that are difficult to identify. In this sense, the combination of digital movies (DMs) with a flow-batch analyzer (FBA) has shown to be a promising analytical tool. In a digital movie-based automatic titrator (DMB-AT), a DM is recorded by a webcam as the titrant is added in the mixing chamber (MC). The DM is then decompiled into frames that are ordered sequentially. In the first frame, the region of interest (ROI) of 257 × 255 pixels is defined. The red (R) component is used to calculate the Pearson's correlation coefficient (r) between the first frame and each subsequent one. The titration curve is obtained in real time by plotting r values and the opening time of the titrant valve. The experimental titration curve is fitted in a sigmoidal regression, and the new titration curve is used to estimate the end point by means of the second derivative method. A software program written in Adobe AIR for a desktop platform programmed in ActionScript 3.0 controlled the opening time of the solenoid valves, gathered the analytical signal and carried out the analytical procedure in real time. The feasibility of the method was confirmed by determining the total hardness and the total alkalinity in mineral waters. The results were compared with classical batch titration and did not exhibit statistically significant differences when the paired t-test at the 95% confidence level was applied. The proposed method is able to process approximately 67 and 86 samples h− 1 for total hardness and total alkalinity, respectively. Its precision was confirmed by overall relative standard deviation (RSD) values of less than 1.0% in both applications.