Bower-Barff Rustless Iron

Write the two down before you light anything:

Fe₂O₃ — red rust. Porous, flaky, larger in volume than the iron it replaced. Falls off and exposes fresh metal.

Fe₃O₄ — magnetite. Black, dense, hard, tightly adherent. Seals the surface.

Both are iron plus oxygen. The whole process is about steering the reaction to the second one — which means hot, and starved of free oxygen.

Wire-brush the iron part to bare metal and degrease. Place it in a closed retort or muffle — the chamber must seal well enough to exclude free air.

Air in the chamber is the failure mode: free oxygen at red heat gives red scale, which is exactly what you are trying not to make.

Heat the sealed retort to a dull to bright red, roughly 500–650 °C.

Below about 500 °C the reaction is impractically slow; much above 650 °C and the coating grows thick, brittle and prone to spalling when it cools.

Pass superheated steam through the hot chamber for 20–40 minutes.

The reaction is 3 Fe + 4 H₂O → Fe₃O₄ + 4 H₂. Iron takes oxygen from the water and releases hydrogen.

Two things follow from that equation, and both matter. The oxygen is delivered bound in water, never as free O₂ — which is what selects magnetite over red rust. And the hydrogen given off fills the chamber and keeps air out, so the reaction protects its own conditions. Vent it safely; hydrogen burns.

Cut the steam and let the retort cool closed. Opening it hot lets air onto hot iron and puts red scale over your black oxide at the last moment.

Expect a uniform matt to satin blue-black surface, hard, and firmly attached — it will not rub off on a cloth.

Oil it. As with phosphate, the magnetite layer is thin and slightly porous; it resists rust far better than bare iron and better still with oil in it.

Put the treated part and an untreated control somewhere damp for two weeks.

Expect the control to be orange within days and the treated part to stay black.

Then scratch the treated part through to bare metal and keep going. Rust will start at the scratch and creep under the coating. That is the honest result and the process's real limitation — magnetite is a barrier, not a sacrificial coating. Zinc protects a scratch because zinc corrodes instead of the iron; black oxide has no such mechanism and simply stops where it is broken.

Who did what. Frederick Settle Barff, a chemist, patented the superheated-steam method in 1876. George Bower, an ironfounder at St Neots, added a stage using producer gas — alternating reducing and oxidising atmospheres — which made the process faster and commercially viable around 1880. The hyphenated name reflects a genuine two-stage collaboration, not a courtesy.

On the missing numbers. We could not verify either British patent number from a primary source. The process is well documented otherwise — trade pamphlets from the Bower-Barff Rustless Iron Company, a Popular Science Monthly account of 1886, and surviving architectural ironwork. Rather than repeat a number we cannot check, this blueprint states the dates and names and leaves the numbers open. An unverified citation is worse than an acknowledged gap.

Where you have already seen it. The same magnetite chemistry, at different scales: gun bluing, the black finish on Victorian railings and lamp standards, and the seasoned surface of a cast iron pan — which is magnetite plus polymerised oil, exactly the coating-plus-oil combination step 5 describes.

Why it faded. It needs a sealed retort at red heat, so the part must survive that temperature and fit in the chamber — no good for anything hardened, thin, or large. Galvanizing protects a scratch and Bower-Barff does not. It survives today mainly as black oxide finishing on tools and fasteners, done in hot alkaline salt baths rather than steam retorts, and valued as much for appearance and lubricant retention as for corrosion resistance.