A long-oil alkyd paint and an oil paint coated on the zinc surface of zinc rich paint and galvanized steel cause the failure of adhesion with exposure in short period. This paper is to evaluate the factors which may influence the loss of adhesion of these paints. Galvanized steel was coated with the long-oil alkyd paints, an oil paint, an epoxide paint, a vinyl paint and a wash primer, and exposed out-door. After exposure, adhesion was measured by torque method and by cross hatch adhesion test and materials produced at the inter-surface of paint films that were peeled off by means of amalgamating galvanized steel were analyzed by ATR method of infrared absorption, and viscoelastic measurement was carried out for long-oil alkyd paint films. Some papers state that calcium plumbate paints are excellent for direct use over zinc surface, so calcium plumbate paints were treated by the same method. The results are as follows : (1) Long-oil alkyd paints and the oil paint react to zinc, and produce the zinc soap between zinc and paint films. Adhesion decreases with these products. (2) From the viscoelastic measurement, it is found that the films which produced zinc soap on galvanized steel became harder and more brittle than the films without metal soap on tin plate. This result suggests that the inter-surface of paint films on zinc was very brittle. (3) Calcium plumbate inhibits to produce zinc soap and so, calcium plumbate paints keep good adhesion.
The electric capacitance method, which was originally proposed by Brasher and Kingsbury to determine the amount of absorbed water in a paint film, involves a few problems in its basic assumption. As an example, the water distribution in the paint film was assumed to be always uniform even when the stage of water absorption was transient. Therefore, the authors investigated the accuracy of the B-K method by the simulation with the digital computer and also by the immersion test. Firstly, the distribution of water through the paint film was calculated on the basis of the onedimentional diffusion theory, and then the overall amount of water uptake W1 was estimated by the integration as a function of time duration. Furthermore, the overall electric capacitance of the painted specimen was estimated by integrating the local electric capacitance which varied in accordance with the amount of absorbed water. The amount of absorbed water W2 was again estimated from this ovarall electric capacitance by the ordinal B-K method. The comparison of the values of W1 with that of W2 revealed the fact that the error of the B-K method must be in practical quite small except the initial period of water absorption. In the immersion tests of the painted specimens in the sodium chloride solution, however, a rather great discrepancy was found between the directly obtained water uptake by the weight method and the estimated by the electric capacitance method. Several reasons were speculated for explaining the observed discrepancy. In the most case, the plots of the weight and the electric capacitance against the time nearly followed the analytical equations derived in this investigation as a rough approxi-mation. A method of estimation of the diffusion constant of water in the paint film was devised by using these analytical equations. By this method, the values ranging from 10-9 to 10-11 cm-2/sec was obtained as diffusion constants through various protective paint films from the observed electric capacitancetime relations. Faculty