puddling furnace

[MattBower]

Grant, puddling starts with CI and passes through mild steel (basically) as carbon content falls. What makes you think it wouldn't be possible (not necessarily very practical, but he seems determined to try) in the middle of that process?

[Nakedanvil - Grant Sarver]

Melted cast iron is very fluid, making is relatively easy to "rabble" in the slag soup. As the carbon is reduced, it becomes "pasty" and forms into small "curds" that are gathered into a sponge ball. The process depends on maintaining a furnace temperature above the melting point of cast iron and below the melting point of iron.

Might be able to do it with steel scrap, but it's far better to attempt a known process and learn how that is done than to attempt a process that has no documentation and might be doomed before you start. I'd want to have successful experience puddling before I attempted something that experimental. I've read that often a percentage of scrap could be added, but I've never seen anything about using the process on a charge of just scrap.

[Nakedanvil - Grant Sarver]

Think about this Matt: if it was a smooth transition from cast iron to wrought iron and every stage in between, why wasn't the process used to make carbon steel? Like high carbon, medium carbon and low carbon. I may be all wet, but I think it remains pretty much a mixture of liquid cast iron, slag and lumps of wrought throughout the later part of the process until it's all converted.

[fciron]

Melted cast iron is very fluid, making is relatively easy to "rabble" in the slag soup. As the carbon is reduced, it becomes "pasty" and forms into small "curds" that are gathered into a sponge ball. The process depends on maintaining a furnace temperature above the melting point of cast iron and below the melting point of iron.

That was always my understanding of the puddling process; as the carbon cooks out, some of the iron condenses out of the solution. The Byers book describes it a bit differently, so I was hesitant to mouth off about it. (I poo-pooed the idea of powdered metal damascus years ago on keenjunk, so now I fact check before I post. )

[thingmaker3]

In the actual process, the stuff with the lowest melting point, the white cast iron, would stay liquid.

The stuff with the highest melting point, the WI, would precipitate. Any carbon would rapidly diffuse out into the atmosphere.

Mild steel has a much higher melting temp than CI - by hundreds of degrees F. Iron and mild steel are just a smidge apart on that scale. Your temperature controls going to be that tight?

-
also...


Tell me again how you plan to get the ferrosilicate into it? Ain't none in mild steel, you know. Plenty of it in white cast iron, though.

With no source of ferrosilicate, you'll only decarburise the mild steel into milder steel. Won't be "pure iron" because it will have manganese in it. Won't be WI because it won't have ferrosilicate.

[MattBower]

But I think they did sometimes make carbon steel -- or steely wrought, anyway -- that way, Grant. It sounds like a very tricky process that you'd immediately dump in favor of the Bessemer process, though, given the chance.

If in the operation of converting pig iron into malleable iron by puddling in a reverberatory furnace the process be stopped before the decarbonization is complete (the temperature of the furnace being a little lower, so that the partial solidification of the mass on decarbonization— "coming to nature" — takes place more easily), the resulting metal is a more or less carbonized iron, which, when prepared from pig free from any large quantity of sulphur and phosphorus, is susceptible of many of the applications to which steel is put. Considerable skill in manipulation is necessary in order to obtain anything at all approaching to a uniform product, the tendency being towards the production of a mass with lumps of soft wholly decarbonized iron, and sometimes of but little decarbonized pig irregularly distributed through it. This is best overcome by conducting the decarbonization more slowly and at a somewhat lower temperature than is usually done in ordinary puddling, and using less fettling and a less oxidizing atmosphere. A manganiferous pig is almost essential to the obtaining of a good product, first because the oxidation of the manganese gives a more fluid slag, and secondly because the small quantity retained by the product decreases the injurious effects of sulphur, phosphorous, &c., on the physical properties of the metal.

(1888.)

He had recently seen in Germany a process of producing steel by stopping the operation of puddling pig iron at a certain point, or intermediate state between cast and wrought iron, and hammering the mass at once into bars. The operation was one of much delicacy, and depended entirely upon the skill of the workman.

(1843.)

I think the idea is a fool's errand -- but I still think it might be possible to make it work.

[MattBower]

Tell me again how you plan to get the ferrosilicate into it? Ain't none in mild steel, you know.

Wouldn't melting in some sand take care of that?

[thingmaker3]

Wouldn't melting in some sand take care of that?

Sand is silicon dioxide, not ferrosillicon.

You wouldn't put chlorine gas and sodium metal on your eggs, would you?

[connor bachmann]

I suppose I could start with cast iron (fining it first, of course) then add mild steel halfway through the process in order to bulk up the mix.

[MattBower]

Sand is silicon dioxide, not ferrosillicon.

You wouldn't put chlorine gas and sodium metal on your eggs, would you?

Whatever the slag in wrought iron consists of by the time the wrought has become, well, wrought, it starts life as silica impurities in the ore -- no?

[thingmaker3]

Whatever the slag in wrought iron consists of by the time the wrought has become, well, wrought, it starts life as silica impurities in the ore -- no?

You'll need a blast furnace for that.

[MattBower]

Why? I'm not saying you're wrong. I'd just genuinely like to know why. I'm not a chemist, but would've guessed that SiO2 + Fe + heat = Fe2SiO4, and all those things would be present in a puddling furnace charged with mild steel -- if you added sand.

[thingmaker3]

I'm not a chemist either. Closest I got was to marry a chemist's daughter.

I also hope I'm not coming off as discouraging - I'd really like to see someone pull off a hobby scale puddling furnace. But they'll only succeed if they play by Mother Nature's rules.

The reaction for ferrosilicon starts with reducing silicate to silicon: SiO2 + C > Si + CO2
There's some calcium involved, too.

Then the iron descends through the silicon, dragging it toward the bottom of the blast furnace or bloomery furnace. (This is why no stringers occur when molten glass sits on top of molten crucible steel.)

When mills make H-13 or some of the S series steels or 300 series stainless, they don't add sand. They add ferrosilicon.

As previously noted, molten steel is at a higher temperature than our average puddling furnace.

[MattBower]

Then the iron descends through the silicon, dragging it toward the bottom of the blast furnace or bloomery furnace. (This is why no stringers occur when molten glass sits on top of molten crucible steel.)

When mills make H-13 or some of the S series steels or 300 series stainless, they don't add sand. They add ferrosilicon.

As previously noted, molten steel is at a higher temperature than our average puddling furnace.

No stringers in pig iron from a blast furnace either, though; the slag floats to the top of the fully molten, denser iron. The slag is tapped off before the pigs are cast, and skimmers remove the rest during the pour. (I think the slag would segregate out in a regular puddling furnace, too, except for all the mechanical mixing that occurs as the iron starts to solidify.) There will be some silicon impurities in the pig iron, but not in the form of slag stringers. To get wrought iron from pig in a puddling furnace, you have to add slag (with one possible caveat -- see below). The question is the form in which it's added. Which leads me to a question that's been bothering me since you raised the ferrosilicon issue (and ties into your comment on H13 and the S- steels): are you sure sure silicon as an alloying element is the same as silicon as a slag constituent? I'm really not sure of that at all. Here are a few points of reference:

The hearth [of a puddling furnace] is made of furnace slag and iron ore. This slag is composed mainly of iron oxide and silica and is called cinder....(1) The door is closed tightly, and the heat is so regulated that the iron and the cinder become pasty and melt down together. This requires about 30 minutes, and is called the melting-down stage.

(2) After the charge has been melted and the iron and cinder well mixed, the clearing stage follows. The puddler's helper uses an iron bar with a bent end to stir the whole charge thoroughly, working through the hole in the door. The stirring brings the impurities of the iron into contact with the oxides of the hearth and of the charge, and these, assisted by any oxygen coming from the air which enters through the fire box, oxidize the remaining silicon, manganese, and a further amount of the phosphorus. During this stage, the furnace is kept very hot. A slag is formed, containing the oxidized impurities and much iron oxide.

(Link.)

The material operated on in puddling is iron containing from 2½ to 4 per cent. of carbon. During the first stage of the process this iron is melted down to a fluid bath in the bottom of a reverberatory furnace. Then the oxidation of the carbon contained in the iron commences, and at the same time a fluid, basic cinder, or slag, is produced, which covers a portion of the surface of the metal bath, and prevents too hasty oxidation. This slag results from the union of oxides of iron with the sand adhering to the pigs, and the silica resulting from the oxidation of the silicon contained in the iron.

This cinder now plays a very important part in the process. It takes up the oxides of iron formed by the contact of the oxidizing flame with the exposed portion of the metal bath, and at the same time the carbon of the iron, coming in contact with the under surface of the cinder covering, where it is protected from oxidizing influences, reduces these oxides from the cinder and restores them to the bath in metallic form. This alternate oxidation of exposed metal, and its reduction by the carbon of the cast iron, continues till the carbon is nearly exhausted, when the iron assumes a pasty condition, or "comes to nature," as the puddlers call this change.

Link.

The essential chemical constituents of the slag produced in the puddling furnace and retained in part by the iron are iron oxides, both ferric (Fe2O3) and ferrous (FeO), oxide of manganese (MnO), silica (SiO2), and phosphoric acid (P2O5). Of these the oxides of iron and manganese are basic in their chemical affinity while silica and phosphoric acid are acid. These bases and acids combine with each other to form neutral compounds: silicates and phosphates of iron and manganese.

[Link.]

Shifting gears, here's something interesting that I came across while reading on this subject. More here, too. This "busheling" process sounds very much like what Connor proposes to do -- and apparently it was quite common.

By busheling scrap is meant the charging of the puddling-furnace with baskets of small steel scrap, bringing them to incipient fusion, balling the mass, and squeezing out and rolling as in the true puddling process. This produces a bar or sheet composed of particles of steel compactly welded together, with intermingled slag, closely resembling puddled iron, but containing much more manganese (and generally less phosphorus). The structure of this material is extremely hard to distinguish from puddled iron on the one hand and charcoal-hearth iron on the other, because the distinct lines of separation between iron and steel as formed in fagoted material are missing.
On the other hand, if the steel scrap is actually melted in the puddling-furnace, there is nothing in etching to distinguish the structure of the finished product from wrought-iron.

[Link.]

[thingmaker3]

are you sure sure silicon as an alloying element is the same as silicon as a slag constituent?

I'm sure it is NOT. Slag is not silicon, it is ferrosilicon. Sand is also not silicon, it is silica. Just to be thourough, sand is not slag.

Thank you for the links to reading material! I shall enjoy going through them!

[MattBower]

I'm sure it is NOT. Slag is not silicon, it is ferrosilicon. Sand is also not silicon, it is silica. Just to be thourough, sand is not slag.

Thank you for the links to reading material! I shall enjoy going through them!

You're misreading my question. My question was whether you're sure silicon exists in the same form in both slag and ferrosilicon. More directly, my question is whether you're sure that slag is ferrosilicon. The chemical formula for ferrosilicon seems to be FeSi. I haven't yet found anything to suggest that FeSi is an important component of slag in a bloomery or puddling furnace, both of which seem contain a lot of iron silicates (e.g., fayalite, Fe2SiO4). Basically, I'm skeptical that slag is FeSi -- although it certainly seems reasonable to think that FeSi could become slag. But I'm not yet convinced that you need to make FeSi in order to make Fe2SiO4 from various forms of SiO2, Fe and iron oxides.

[Nakedanvil - Grant Sarver]

I've done a fair amount of research on puddling, but I'm certainly no expert. I've not seen any reference to ferrosilicone being used in puddling. The slag in puddling is usually identified as silica sand and iron oxide.

I do find references to ferrosilicone used in steel production as an alloying material.

Ferrosilicon is used to remove oxygen from the steel and as alloying element to improve the final quality of the steel.

LINK

It is a de-oxidizer and is used to inhibit de-carburization. In puddling the whole point is to de-carburize the charge, right? Can you cite some references?

[thingmaker3]

You're misreading my question. My question was whether you're sure silicon exists in the same form in both slag and ferrosilicon. More directly, my question is whether you're sure that slag is ferrosilicon.

I am sure (very sure) that I will be happy to see the results of your and Connor's experiments.

I am sure I have read of "ferrosilicon stringers" or "ferrosilicate stringers" or some such in WI (as the defining quality of WI) and will check my source when I have time to spend on such endeavors. In the interrum, please do not let my concerns or questions slow or delay your experiments. :)

[Nakedanvil - Grant Sarver]

Ferrosilicates would make sense. I researched it because I keep toying with the idea of trying a small puddling furnace. As I alluded to before, I think I would make it gas fired, maybe a 50-75 pound charge. Be really cool to do a Bessemer converter too. Could feed a converter with an old-fashion cupola to melt the cast iron.

[MattBower]

Ferrosilicate -- yes.

The more I read and think about this, the more intrigued I am. If it were me, I think I'd build a shallow tray of high alumina, 3000 degree dense firebricks, mortarted together. Line that with hammer scale and sand. Put a big propane ribbon burner on top of it, pointing straight down into the tray, similar to the flat-top modular forges I've seen around here, running a strongly oxidizing atmosphere. (I'm not sure a reverb is absolutely necessary, as long as you're running a clean fuel and an oxidizing atmosphere.) Cook the sand and scale until they start to get soupy, then add small pieces of mild steel (I'm thinking of the scrap bin full of punch press slugs at the structural steel place up the road), stir it all together, get it all up to a good welding heat, consolidate it, and start working it. I might also consider adding a little cast iron scrap to the mix to help get things nice and soupy, then throwing in some additional hammer scale, if necessary, to decarburize the CI and start fusing the whole mess.

If the slag wasn't satisfactory this way, I'd try running a reducing atmosphere for a while, and possibly adding a little lime to the mix, before adding the steel and switching to an oxidizing atmosphere. (As far as I can see, CaO is the one thing that goes into blast furnace slag that wouldn't be present in the sand-scale mix.)

I think you could also run this on just scrap cast iron.

Of course, like many things, there's no way I have time to actually do this.

[MattBower]

By the way, I know the "strongly oxidizing atmosphere" part sounds wrong, considering that we're talking about forge welding. And maybe it is wrong. But it seems necessary for the decarb we want, and I think the saving grace might be the slag bath. From what I can tell, the slag acts as a flux and dissolves iron oxides. I think that's why most wrought iron welds nicely even without flux, and I think it's why you might be able to weld with the wrong atmosphere in a puddling furnace.

[Nakedanvil - Grant Sarver]

Got Larry (monstermetal) interested now. He's got some parts and pieces, I've got some parts and pieces....................

[MattBower]

Got Larry (monstermetal) interested now. He's got some parts and pieces, I've got some parts and pieces....................

Cool! Except for the probably prohibitive cost of shipping all the way across the country, I'd send you some unused cast iron plates from an old weight set...

[MattBower]

By the way, I guess the firebrick (or refractory) bottom isn't really necessary. A lot of the old puddling hearths were made of cast iron. Seems pretty ballsy to me, but apparently they made it work.

[Nakedanvil - Grant Sarver]

My temptation is to do true reverbatory furnace, but we might just use your suggestion with a ribbon burner in the roof. Larry has a large ribbon burner and I have some large blowers. Funny, often these projects get more complicated is they reach reality, but this seems simpler now than it ever did. Might cast a basin using a castable refractory. Same with the roof. Kind of a turtle shell construction. Fifty pounds of scale, fifty-sixty pounds of cast iron and we'll be ready to go.