Steel color change - how come?

I don't think there is any interaction. I think the brass just adds thermal mass to avoid temperature swings.

chuck

Whaaa?! You do realize ferrous oxide has a *melting point* in excess of

2,000°F, don't you?

Maybe under heavy reduction, where the hydrogen and, more importantly, carbon in the flame are able to actively reduce the oxide from FeO to Fe (such happens readily with copper metal), but not under any other circumstances.

Tim

-- Deep Fryer: a very philosophical monk. Website:

Here's the deal. Take a piece of iron. It has a perfectly clean surface, I mean the oxygen and nitrogen molecules of the air are bouncing directly off the pure iron (and occasionally iron carbide and others) surface. Now at room temperature, a few of these do bust up on the iron surface and oxidize it. Like aluminum, this layer is invisible, but because it's somewhat indifferent, it doesn't change the chemistry of the iron and you don't notice it. This layer is only as thick as it is because the oxygen molecules can't get past it at this temperature.

Ok, so let's raise the temperature. Radiation heats the air molecules near the steel to the same kinetic energy, that is to say, hot air's molecules move faster. So they hit the iron surface with more energy, and occasionally one will pass through the oxide layer and oxidize more iron. (It probably passes by diffusion, where the surface oxidizes to magnetite - Fe3O4 - or rust - Fe2O3 - which is then passed backwards to the metal, which reduces Fe3O4 and Fe2O3 back to FeO, at the price of more Fe metal being burned.) Just as carburization can diffuse hardening carbon (or nitrogen in some cases) only so far, likewise the oxygen only goes so far through the oxide. It's always passing through, even at room temperature, so the response of oxide growth is probably logarithmic - it tapers off quite quickly as thickness rises, but never stops completely. It's just that thermal response is exponential, so it'll take about two million years to eat a tin can, while at orange heat, your tin can will hold molten aluminum for only about fifteen minutes!

A temperature of 350°F for a few minutes produces a nice light straw color (hard to spot because the oxide takes time to grow, and you can't anticipate it because this is the first interference layer, around 80 nanometers thick?), but the same temperature extended to an hour gives a purple coloration.

Actually, they come back several times, but each time harder to see because the light has to travel through more oxide thickness. If you heat a shiny bar from one end in plain air, you'll see the first layers, yellow, purple, blue; then a darker run of yellow, purple and blue, and so on for maybe three or four total modes. What's happening is light is spending 1/2, 1

1/2, 2 1/2, ... wavelengths inside the thickness of oxide (1/4, 3/4, 1 1/4, ... wavelength thick layer, for the wavelength of light *passing in the medium* (light slows down per the index of refraction), of the *canceled frequency*). Thicker layers attenuate more, so it quickly (500nm?) starts looking black. Fe3O4 is a wonderful electromagnetic-absorbent material, after all. No wonder the military uses it...

Tim

-- Deep Fryer: a very philosophical monk. Website:

Just using a propane torch the blue will disappear as the I continue to hold it to the steel piece I'm heating. It doesn't come back as it cools.

Peter

Yup. What Tim said.

Steve

Tim Williams wrote:

See what else Tim said. The color disappearing doesn't mean the oxide does.

Steve

Peter Grey wrote:

"Tim Williams" wrote: Here's the deal. Take a piece of iron. It has a perfectly clean surface, I

^^^^^^^^^^^^^^^ I followed your explanation with great interest. You seem to take me back to some the beliefs I held originally, and abandoned briefly as this thread unfolded. I take it that you do not agree that a certain color on the surface is correlated with a certain temperature in the metal. As I thought originally, the color DOES correlate with film thickness, which, in turn, is dependent on the time/temperature history.

Since "drawing the temper" of steel, as done by blacksmiths, using the surface colors, is a way of raising the steel to the desired temperature, why does it work? Wouldn't a longer time at a lower temperature produce the same interference colors as a shorter time at a higher temperature? Could it be that the time/temperature history produces the same effect on color that it does on hardness? Or am I completely off the track here?

Thanks.

It can't be, since I've torch tempered and oven tempered metal myself. The oven pieces come out considerably darker, in the purple range as I mentioned.

Because it works in the short term. As you expand time exponentially, you get more "out" of it, but after a while, to get even a small amount out, takes a very long time. The endpoint is arbitrary; by eye, over a few minutes, you'll probably spot between 300 and maybe 500°F (SWAG). In the oven, you get a bigger change out.

Why does it work as far as the metal? Two reasons. For one thing, it's just simple carbon steel, you can't really go wrong with it (short of overheating before quench, which makes it crunchy no matter what!). Number two, the reactions in the metal, where bainite and whatnot break down to more stable phases during tempering, is the same kind of time-temp governed reaction as the oxidation is. It might not proceed at the same rate (it would be interesting to compare this!), but who knows.

Exactly! So it may be that my purple blades at 350°F for an hour are overtempered, while the yellow-for-a-few-minutes torch tempered jobs are undertempered. This is where some imperical evidence comes in handy. The last blade I tempered was the brass handled knife here:

I've sharpened this good enough to shave with (not real comfortable, but it cuts the hair smoothly anyway..), and in the process I don't notice much of a burr turned up so it must be pretty hard. That's fine with me since it's so thick it'll "never" break.

I tempered that to 350°F for an hour and it came out purple (splotchy mind you, fingerprints for instance are prime spots to prevent oxidation). It seems to be simple 1080-1090 (used to be a chisel), nice yellow-white bursts on grinding, same as a file.

Tim

-- Deep Fryer: a very philosophical monk. Website:

As Tim mentions, the longer you leave steel in a tempering oven at a

*constant* temperature, the color keeps changing.

Tempering by colors works if you use a consistent timing. If you change from warm forge exhaust to direct torch heat, I don't think the same color indicates the same temper. Not that I've ever tried this, but it seems to make sense. So in the end, the temper colors are a guide that you have to experiment with to find what works for you.

Steve