            Tafel Region Extrapolation.

Direct measurement of the corrosion current is usually not possible. Corrosion current is the current between the anodic and cathodic sites and normally both the anodic metal dissolution process and the supporting cathodic process occur at microscopically adjacent points on the metal surface. Inserting an ammeter in this circuit can be difficult.

The potential-current response of a typical corroding electrode is as shown below. Note the semi-logarithmic plot. At any given potential the measured current is the sum of the anodic and cathodic reaction currents. At potentials below the free corrosion potential Ecorr the response is mainly that of the cathodic process. At potentials above Ecorr the reverse occurs. In an ideal situation both the contributions would be simple exponentials - straight lines on the plot. In practice one or both can be non-linear due to mass transfer limitations, formation of oxide film etc. Often, over a region either side of Ecorr, the current is proportional to the exponential of the potential - hence on a semi-logarithmic plot the response is linear. Extrapolation or curve fitting over this region can be used to determine Ecorr and Icorr, the free corrosion potential and the corrosion current.

In practice, the shape of the measured response depends on a number of factors, such as the speed of the measurement potential sweep etc. Also, in order to obtain the data, the test electrode has to be polarised appreciably away from its natural state. This in itself can cause irreversible changes to the electrode surface and hence this method is not recommended for on-line corrosion rate measurement.

The actual measurement can be done either galvanostatically (the cell current is set and the resulting potential measured) or, more commonly, potentiostatically (the test electrode potential is set and the resulting cell current is measured) using a potentiostat. One important source of error is that caused by the finite electrolyte resistance. This can cause the Tafel regions to appear distinctly non-linear (non-exponential ?) on the plot. The cure is to either use 'IR' compensation during the measurement or to fiddle the data afterwards. 