Wiener filtering of electroanalytical data by means of fast Fourier transform

Stripping electroanalytical techniques, in spite of their high sensitivity, usually suffer from poor signal-to-noise ratios. Both the numerical values of the ratio and the distribution of the noise across the recorded frequency spectrum must be considered carefully in order conveniently to separate the noise from the analytical signal. Traditionally, this separation has been achieved by various filtering techniques that are time-consuming, require a visual and therefore subjective interpretation of the signal spectrum and have in general limited scope. In this paper, a filtering technique that overcomes these limitations based on the Wiener filtering method in the frequency domain has been developed. According to this scheme, the detection and separation of the noise components in the voltammetric signal are automatically achieved in an optimal way. The principle of the method is that the noise in the signal can be estimated from the blank signal and subsequently eliminated. The feasibility of the proposed method for filtering in stripping analysis was first assessed with simulated voltammograms and then the technique was applied to real voltammetric data.