I was only talking about carbonates and their affect on Al corrosion. The link I found pointed out that dissolved halides were much worse at removing the protective oxide layer than dissolved carbonates.
Carbonates in a boiler are bad news for several reasons and you don't need
900F to suffer. When you heat water the solubility for CaCO3, Mg2CO3 and others drops off and they come out of solution. This is what forms scale on heat transfer surfaces (even in home water heaters, and LP boilers). The really bad part is that once the CaCO3 plates out, it will not re-dissolve into the boiler water when you shutdown the boiler. You can't get the scale back off with out 'mechanical' cleaning or 'chemical' cleaning. One way to combat this in the old days was to add tri- and di-sodium phosphates. The Na would combine with the CO3 to form a scale that was easier to remove. The first trade name of such boiler-chemistry control was 'CalGon' (same as the retail soap) which stood for 'calcium-gone'. But the large amount of phosphates wasn't so good for the environment so phosphates have been regulated.In steam plants it is now common to use 'polishers' (ion-exchange resin beds) to continuously demineralize the condensate/feed-water. Keeping total conductivity down well below 0.1 micromhos/cm is quite common. This slows the buildup of TDS in the boiler water.
Dissolved oxygen in boiler is also bad for Cl-stress corrosion. But modern low-carbon steels have a problem with too-low an O2 content as well. Yes, O2 levels are controlled with several techniques (deareating feed tanks, contact feed-water heaters and such). We actually have to keep O2 levels between 10 and 50 ppb for best results.
daestrom