The relationship between wetting amount and corrosion rate in initial stage of atmospheric corrosion was investigated to determine whether the peak corrosion rate observed under solution film also occurs under dropwise wetting. Zinc ER sensors were deposited with NaCl and tested for 360 hours at constant temperature and humidity. The corrosion rate was compared with wetting amount, which could be calculated from the amount of deposited salt and the relative humidity. Peak corrosion rates did not occur in this experiment, and corrosion rates increased with increasing wetting amount.

Various environmental factors （temperature, humidity, light/salt exposure, etc.） act in complex ways to degrade the corrosion protection performance of coating films. However, it is difficult to evaluate the individual and combined effects of all these factors. On the other hand, corrosion protection performance has a strong correlation with the impedance of coating films, which corresponds to the resistance to the flow of electrical charges in the coating. Therefore, the effects of light irradiation on the penetration behavior of moisture and salt into the coating films were investigated. As a result, it was found that material changes due to light irradiation accelerated the penetration of water and/or salt , thus decreasing the impedance. When water and/or salt penetrated into the coating films, a peak in the DRT （Distribution of Relaxation Times） spectrum was observed at relaxation times less than 10^－1 s.

In this study, we propose a method to efficiently estimate equilibrium constants for multiple reactions using titration curves of metal ion complexes. First, titration curves were measured, and candidate complex reactions were listed. An efficient numerical simulation method was then developed to compute titration curves parameterized by equilibrium constants. To estimate these constants, we formulated an optimization problem maximizing the likelihood between simulations and experiments, applying Bayesian estimation to reduce instability. Additionally, model selection based on Akaike’s information criterion （AIC） identified key complex reactions influencing pH changes. The method’s effectiveness was validated through two tests. First, simulated titration curves with artificial errors demonstrated accurate equilibrium constant estimation. Second, applying the method to aluminum hydroxide and aluminum fluoride titration curves, we identified eight major reactions from 21 candidates.