That was excellent, Joules.

I'm not casting steel, of course, just iron (lower temp, higher carbon) but it was very interesting anyway. More information than I have found on my own.
My guess is that pure silicon is pretty ephemeral in a melt, and more so than ferro-silicon. Generally in home foundry work, the effect of ferro-silicon addition is said to last no more than 5 minutes after addition, so you don't add it until the very end, before pouring the mold.
I do use ferro-silicon to prevent chill (hardening -- carbide formation) in melts for thinner castings. Or if a scrap type seems to chill easily, by experience.
I originally bought the pure silicon in hopes of adding it to aluminum to try to make aluminum extrusion scrap a better sand-casting material. But shortly after receiving it (EBay purchase) I read that others had tried to add it to an aluminum melt, and hadn't succeeded because of its high melting temperature (1414C). So then I thought about using it for iron, and wondered why ferro-silicon was normally used, and I'd seen no mention of pure silicon.
I've since heard of one person who said they were successful in adding pure silicon to aluminum, but no explanation (so far) of how that was achieved.
I have personally dissolved copper in aluminum at much lower than copper's melting temperature, so I don't discount that it may be possible with silicon. That's dissolution, not melting. I would like to know about that if it's been done.
I've also seen one video on YouTube where a very amateur melter, with poor stock (beverage cans?) did seem to combine silicon and aluminum in a covered electrical muffle furnace at very high (for aluminum) melt temperature. Maybe he got it to silicon melt temp? If so, and somehow he kept oxygen out, I guess it's possible. But in a normal aluminum melt, gross overheating is asking for a big loss in quality, and wholesale production of dross. Considering what he started with the metal quality of the final castings looked poor, with the Si/Al example slightly better than without, to my eye.
Anyway, thanks Joules, and nope not a fab lab, just EBay, and at a price for a small quantity not much different that the ferro-silicon I purchased in the past.