The theory of metallic corrosion in the light of quantitative measurements.—Part III

Abstract

The present paper describes an investigation into the degree of reproducibility obtainable by the oxygen absorption method of measuring corrosion, and gives comparative results for this and the loss of weight method. The work described in Part II which dealt with the corrosion of zinc in dilute potassium chloride has been extended up to 4N solution, and to a series of potassium sulphate solutions from N/10,000 to N. Slight vibration of the apparatus has been found to have been the source of error in three of the experiments described in Part II, namely those performed in N/200 and N/10 KCl. The corrosion-time curves for these solutions show too steep a slope owing to convection of oxygen set up by the vibration. Curves more nearly representing corrosion in truly stagnant conditions are given in fig. 4 of the present paper. The effect of vibration was only important in those experiments which were controlled by the rate of oxygen supply, whereas most of the experiments of Part II were controlled by the concentration of chlorine ions.