When you take a "high carbon steel" (let's use O1 as an example here), to a "yellow" heat past critical temp (colors are so subjective....but in reality even that is irrelevant here)...and allow it to air cool, your resulting structure is simply pearlite. It is neither hard nor brittle. But if you were take said high carbon steel to a "yellow heat" and cool it SLOWER than still air (I am talking about the vermiculite, wood ash, or similar overnight cooling treatments), and then AFTER you try to harden that structure with your austenitizing heat and quench, then yes, the resulting martensite matrix will have carbides (mostly cementite in the lower alloy carbon steels like O1 etc) that are in the grain boundaries, which does create a more "brittle" structure than if the initial cooling process was allowed in still air. That is why metallurgists suggest to always allow a still air cool following normalizing/cycling, and not a wood ash/vermiculite treatment on hypereutectoid steels.
If your steel has a eutectiod or hypoeutectoid count (roughly .77%, 1080 and below), then the wood ash/vermiculite/slow cooling process will not cause this problem. If your steel is hypereutectoid (roughly 0.77% and above) like 1095, O1, 52100, White, Blue, etc... then it is NOT advisable to do a slow cooling process during normalizing and/or cycling.
You may notice many of the smiths practicing traditional techniques will indeed do the wood ash/overnight cooling process with very high carbon steels like the Shiro, Ao, and AoSuper steels. These methods are traditions of smiths of centuries old. I cannot comment on if this is the BEST method to prepare these steels for hardening or not. The reason I say this is because these steels, especially the Hitachi White and Blue steels are used mostly in kitchen knives, where there *may* be a benefit to the said added *strain* with that brittle structure (strictly slicing and not chopping/hard impact). When you inquire these things to contemporary and well known/established metallurgists/knifemakers/swordmakers (I stress METALLURGISTS....plural), they'll tell you that any sort of cooling that is slower than still air on hypereutectoid steels like 1095, O1, White, Blue, etc, this treatment will place carbides in grain boundaries instead of within the aus grain, and thus a more brittle microstructure once hardened/quenched/tempered will occur, and are best with still air cooling during normalizing/cycling.
My point here being, there are old vs new methods/thoughts. Some may be better than others, and I do NOT have the answers, but hope that this can help your question *and* (ha ha!) stir up a debate. I am not a metallurgist by any means, and I am constantly learning new things, and I REALLY like the idea of sharing them. Right or wrong.....I'm thrilled to be corrected in these matters, and we all can learn from it.
