SIGNAL FILTERING OF POTENTIOMETRIC STRIPPING ANALYSIS USING FOURIER TECHNIQUES

Four filtering functions, rectangular, triangular, exponential and dec ay exponential for Fourier filtering have been compared in this study of potentiometric stripping analysis (PSA) data. The filtering procedu re involves treating the potential coordinate as mathematically equiva lent to a time coordinate to transform data to the frequency domain. A fter Fourier transformation, each method yielded signal-to-noise enhan cement over the original data in the order, decay exponential>triangul ar>exponential>rectangular. Apart from the filtering functions, two di fferent filtering procedures, conventional and reflectance, have been compared in this study. Peak area measurement was unaffected by the re flectance procedure, while an undesirable increase in peak area was ob served in the conventional procedure. A linear peak area-concentration relationship was retained in all cases. The resolution of overlapping potential peaks, using copper and bismuth peaks as the example, was m ost improved using the rectangular filtering function and worst using the decay exponential filtering function. This is completely opposite to the results obtained in signal-to-noise enhancement study. Thus, a compromise between SIN ratio enhancement and peak distortion has to be made, as the latter is always a trade-off in signal processing. With the use of a personal computer (PC) and commercial data acquisition sy stems, the original saved data can be stored. Selection of a particula r type of filtering function to optimise the analytical signal can be achieved by applying various functions to the original signal, until a desirable output is obtained.