Experimental design applied to the optimization of pyrolysis and atomization temperatures for As measurement in water samples by GFAAS

In general, research experimentation is often used mainly when new methodologies are being developed or existing ones are being improved. The characteristics of any method depend on its factors or components. The planning techniques and analysis of experiments are basically used to improve the analytical conditions of methods, to reduce experimental labour with the minimum of tests and to optimize the use of resources (reagents, time of analysis, availability of the equipment, operator time, etc). These techniques are applied by the identification of variables (control factors) of a process that have the most influence on the response of the parameters of interest, by attributing values to the influential variables of the process in order that the variability of response can be minimum, or the obtained value (quality parameter) be very close to the nominal value, and by attributing values to the influential variables of the process so that the effects of uncontrollable variables can be reduced. In this central composite design (CCD), four permanent modifiers (Pd, Ir, W and Rh) and one combined permanent modifier W + Ir were studied. The study selected two factors: pyrolysis and atomization temperatures at five different levels for all the possible combinations. The pyrolysis temperatures with different permanent modifiers varied from 600 degrees C to 1600 degrees C with hold times of 25 s, while atomization temperatures ranged between 1900 degrees C and 2280 degrees C. The characteristic masses for As were in the range of 31 pg to 81 pg. Assuming the best conditions obtained on CCD, it was possible to estimate the measurement uncertainty of As determination in water samples. The results showed that considering the main uncertainty sources such as the repetitivity of measurement inherent in the equipment, the calibration curve which evaluates the adjustment of the mathematical model to the results and the calibration standards concentrations, the values obtained were similar to international interlaboratorial comparison results.