Anodic dissolution of aluminum in hot chloride solutions produces a high density of fine etch tunnels that extend along [100] directions. Tunnels evolve from cubic etch pits when all but one of the pit wall surfaces become passivated; dissolution then occurs at the one active face at a rate that may initially be as high as 20 A/cm2. Tunnels have square cross sections with sides ∼1 μm and aspect ratios as high as 100:1. Tunnel growth may be considered a unique form of pitting corrosion in which dissolution and passivation occur simultaneously with a sustained balance between the two processes.

Crystallographic pits were grown on Al͑100͒ at temperatures from 30 to 90°C, in solutions of HCl ϩ H 2 SO 4 . Pits grew during a galvanostatic anodic pulse for 5-100 ms that was preceded by a galvanostatic cathodic pulse of 50 ms. The anodic potential has a small peak for 1-2 ms and then remains constant. The cathodic pulse causes rapid pit nucleation so most pits nucleate within 5 ms, and pit passivation and pit growth are the dominant processes at longer times. It was determined that a substantial fraction of pits passivates during the pulse, so growth rates were calculated from the increase in largest pit size with pulse duration. The growth rate is constant at each temperature and follows an Arrhenius temperature dependence with an activation energy of 7.2 kcal mol Ϫ1 . There are significant differences between the growth of pits and the growth of etch tunnels. The activation energy for pit growth is one-half that for tunnel growth, and pit growth rates are greater than tunnel growth rates. It is proposed that the chemisorbed chloride complex postulated to be an intermediary for Al dissolution changes from one structure during pit growth to another, more stable, structure during tunnel growth.

There are few published results of etch tunnel growth in aluminum at temperatures below 60°C because a high tunnel density is not achieved at these temperatures. We find that etching high purity capacitor foil with its as-received surface oxide produces a low but sufficient density of tunnels so steady state tunnel growth rates can be measured at temperatures down to 28°C . These rates follow the same Arrhenius temperature dependence as rates measured at high temperatures. It is concluded that the mechanism of steady state tunnel growth is the same from 28 to 97°C . It is found at lower temperatures that there is a transition period of a few seconds between crystallographic pit growth and steady state tunnel growth during which tunnels grow at a rate that is slower than for pit growth but faster than steady state tunnel growth. The activation energy for this rapid tunnel growth is 7.2kcal∕mol , the same activation energy as for pit growth but one-half that for steady state tunnel growth. This transition period grows shorter with increasing temperature and is not seen at 70°C or above. Tunnel segments grown during this period have a different appearance than steady state tunnels, indicative of a different growth mechanism.

Stresses in porous anodic alumina coatings have been measured for specimens stabilized in air at different temperatures and humidities. In ambient atmosphere the stress is tensile after anodic oxidation and is compressive after sealing. Exposure to dry atmosphere causes the stress to change to strongly tensile, up to 110 MPa. The stress increase is proportional to the loss of water from the coating. These changes are reversible with changes in humidity. Similar reversible effects occur upon moderate temperature changes. The biaxial modulus of the coating is about 100 GPa.