Ok, some details need to be provided in order to aid understanding. To start, I've seen zero legitimate research that indicates pre-collapse eutectic corrosion occurred
(no, Steven Jones work doesn't count. For one, he doesn't claim sulfidation corrosion, he claims thermite melting. For two, his work isn't legit). And two papers
... point at the opposite - corrosion having occurred after the collapse - as being more likely.
Sisson and Biederman merely noted that
"...preliminary experiments[5] indicated that the reaction was not fast and dissolved little metal in 24 h."
Their point is that the sulfidation attacks are not rapid chemical reactions. Both towers collapsed within 2 hours. A 2 hour sulfidation attack isn't outside the realm of possibility given high enough temperatures, but compared to the days to weeks exposure in the rubble pile fires, it's simply more likelly to have occurred there rather than in the standing towers. Especially when the experiments indicate the need for many hours to corrode significant amounts of steel.
Banovic and Foecke made several observations related to banding and microstructure formation (I'm still studying in order to understand it), but their one observation regarding the pattern of corrosion is pretty straightforward: The corrosion on a specific exterior column indicated that it was "prone" at the time of the sulfidation attack. Given that it's a column (albeit a perimeter one), and given that its size indicates it was no higher than the 51st or 52nd floor of either tower (therefore not in the impact zones of either building and not punched horizontal by the jets), it's obvious that it was vertical prior to collapse. If it was vertical prior to the buildings collapses yet prone during the sulfidation attack, it doesn't take a genius to understand that the corrosion occured in the rubble pile. After the collapses.
There is more there than that - read the Worchester Polytechnic team's works (plural), as well as Banovic/Foecke - but I've yet to see one work that demonstrates pre-collapse eutectic erosion. In contrast, I've found at least these two that point at post-collapse sulfidation being the likely case.
------
Regarding the contining moronic, disproven proposals of thermite: No. Biederman's, Sisson's, Barnett's, and Sullivan's work fully demonstrate the top-end temperatures the eroded pieces of steel could have possibly experienced, and they do so by demonstrating the formation of certain steel phases and iron sulfide species that would be destroyed above that temperature. The corroded steel could not have exceeded 1000
oC. It's definitely possible they were lower; Banovich's and Foecke's work indicates lower temperatures, albeit still above 850
oC. Regardless, that is well,
WELL below the temperatures that thermite reach. The characteristics of the steel's microstructure surrounding the corrosions, as well as the sulfidation layers, would simply not exist if thermite were used, because they would have been destroyed. The eroded steel negates the possibility of thermite use. That idiocy can be dropped forthwith.