Chemical Forums

Specialty Chemistry Forums => Materials and Nanochemistry forum => Topic started by: Albedo on June 18, 2019, 04:22:55 AM

Title: Iron Oxide Reduction with Hydrogen
Post by: Albedo on June 18, 2019, 04:22:55 AM
Greetings Chemists, I am struggling a bit with how to solve for reaction time and the input/output state of the elements/compounds involved. Just a heads up, I am a Mechanical Engineer and it has been a over 5 years since I had last touched Chemistry.

Ok, onto the detailed explanation; Iron Oxide can be reduced in the following manners: 3Fe2O3 + H2  ::equil::  2Fe3O4 + H2O
then further 2Fe3O4 + H2  ::equil:: 6FeO + 2H2O
and lastly 6FeO + 2H2  ::equil:: 6Fe + 6H2O

Now, to perform all three reactions, a temperature of over 570°C is needed, which of course is a good thing since this would take place inside a blast furnace. The first problem I am struggling with is what states the input and outputs will be. For instance, the Iron Oxides will obviously have to be a solid since that is the Iron Ore. But what about H2? Does it have to be a gas or liquid, or both could work? What about the outputs then? The pure Fe needs to be in a liquid state. And the H2O would be preferably be a gas. But alas, I don't know how to figure that out, my chemistry knowledge doesn't delve that deeply.

The second problem is the reaction times, which is critical to blast furnace function. Only so much Ore and Limestone (in case you don't know metallurgy too well, Limestone is used to absorb impurities found in the Iron Ore and forms a slag on top of the molten liquid) can be added at a time so good quality ingots can be made. My chemistry knowledge is nowhere near this level, so I expect a lot of help will be required for this problem.

I have tried reading up on the PDf's out there for this topic, but there isn't too much since it is experimental right now. If needed, I can link all the ones I have read.

Thanks for any help provided,
Title: Re: Iron Oxide Reduction with Hydrogen
Post by: Borek on June 18, 2019, 06:58:40 AM
Some of your questions are rather easy to answer:

Check boiling points of hydrogen and water - can any of them be liquid at 570 °C?

Check melting point if iron - will it be liquid at 570 °C?

Reaction time is quite a different beast and actually the best approach is to check things experimentally.
Title: Re: Iron Oxide Reduction with Hydrogen
Post by: Enthalpy on June 18, 2019, 07:12:53 AM
Welcome, Albedo! (An astronomer maybe?)

At 570°C, hydrogen and water exist only as gas, whatever the pressure. Iron is a solid, it needs some 1500°C to melt. In a blast furnace, this difficult temperature is provided by the combustion of (refined) coal, which is very much helped by recycling heat from the exhaust flue into the blast air or oxygen.

A blast furnace uses to input (refined) coal rather than hydrogen to reduce the iron oxides. Whether it happens over the Fe3O4 and FeO route is a detail, only the initial and final iron compounds determine the amounts of reactants. While coal makes dangerous carbon monoxide, it's by far not as bad as hydrogen, and it's cheaper.

Limestone binds impurities like phosphorus and sulphur that are highly undesired in steel. Their amount depends fundamentally on each ore.

The speed doesn't result from simple properties of the materials. It's much an experimental observation that you can try to explain afterwards with subtle models.

In which context do you study this? An economic setup would need strong reasons to choose hydrogen. For instance if willing to produce iron but emit no CO2, some processes make H2 from sunheat, but they pass over intermediates (Zn and ZnO? I could check if needed) which might react more directly with the iron ore to reduce iron, instead of going through H2. Or if electricity shall make H2, then one might instead electrolyse molten iron ore as is done for aluminium. Or inject methane in the furnace rather than converting at part of it into H2.
Title: Re: Iron Oxide Reduction with Hydrogen
Post by: Albedo on June 18, 2019, 06:19:48 PM
Thanks for the quick responses!

Borek, thank you for explaining how that works, I didn't realize it was that simple. As to trying to do an experiment with this, perhaps when I was attending the university I could've got permission to set one up, but right now I do not have the money or contacts to try and set up an experiment. I am trying to do all this with proof of concepts on paper in order to even gain any traction with this. You and Enthalpy both pointed out how I neglected to check the melting point of Iron, thank you both for catching that mistake.

Enthalpy, you mention that the coke used in traditional blast furnaces is not as bad as hydrogen, could you explain a bit further what you mean? I am aware hydrogen is flammable and when used to form H2O at elevated temperatures, a steam explosion must be considered. But other than those concerns I see it as a significantly more desired reduction agent than coke, which heavily pollutes and is a limited resource unlike hydrogen which is abundant and can be separated using electrolysis. As for using electrolysis for Iron separation like other metals, like Aluminum as you mentioned, I have only read that it is in the experimental stages as well and that a chromium-iron anode would have to be used along with a whole lot of other mixtures. I can link the article I checked out if you want, but I do not see electrolysis as being the best way forwards with iron reducing. And lastly, the context I study this in is that of an engineer seeing an opportunity: coke is great and all by providing heat and the reducing agent, but the pollution is just too great and it cannot be renewed, in addition to those, often times there is too much carbon in the wrought iron ingots produced and they must go through a process to lower the carbon content. Then to top that, some really important steel alloys require no carbon in there mixture, making carbon usage very undesirable. Hydrogen on the other hand can be renewed from water electrolysis and it is very abundant. If a furnace were to use this instead, then the prices would go down making it more viable. Then when we include the savings a company will make from not having to worry about the pollution taxes and regulatory laws, this furnace is a viable alternative, despite its' problems.

Thanks again both of you for your fast replies and I look forward to discussing this more with you!
Title: Re: Iron Oxide Reduction with Hydrogen
Post by: Enthalpy on June 19, 2019, 06:15:36 AM
Hydrogen explosions are badly dangerous. If hydrogen gets mixed with air and catches fire (you have hot parts everywhere), almost always it detonates at huge speed, instead of making a slowes flame like for instance natural gas uses to do. This detonation is much more destructive. With hydrogen it happens over a very wide mix ratio with air.

Carbon monoxide is a much smaller risk of explosion, and usually it doesn't detonate. But it's a bad poison, sure.


hydrogen can reduce iron oxides, but I suppose less efficiently than carbon monoxide does. Reacting with one mole of oxygen atoms produces 275kJ for Fe, 242kJ for H2 (making vapour), 283kJ for CO. All data at 298K.

Maybe someone can estimate the proportion of unreacted hydrogen around 1600°C. I suppose only a fraction makes H2O, the rest must be separated and recycled, but without wasting the heat. Not trivial.


Coal is extremely cheap, even more so than natural gas. Hydrogen from electrolysis has no single chance to compete with it. By the way, hydrogen is produced from methane presently because this is much cheaper. I doubt that hydrogen can replace coal.

Electrolysis of iron ore was my attempt to linger this cost problem, because the path over hydrogen means important additional losses. But even that way, Fe would need as much electricity per mole as aluminium, perhaps with a lower voltage, and with moles twice as heavy. So if aluminium costs 5€/kg to construction companies, much if it resulting from electricity cost, steel obtained by electrolysis would cost 2.5€/kg instead of 1€/kg. Bad. Hydrogen would be worse.


I fully agree that steel production too should go emission-free, just like lime production, transports, electricity production... But it's badly difficult!

Perhaps the cheapest route is to sequester the CO2 produced by the steel plant.

The best hope I have would reduce iron ore using sunheat, in some indirect process involving other metals.


Carbon is removed from pig iron to make any steel. This is such an old process that it has become really cheap.

And yes, some steel demands very low carbon content, but the existing processes are cheap enough that any alternative would have a hard time to compete. They can be as simple as: heat in an electric oven, blast oxygen through the melt.
Title: Re: Iron Oxide Reduction with Hydrogen
Post by: Enthalpy on June 21, 2019, 11:23:00 AM
[...] some processes make H2 from sunheat, but they pass over intermediates (Zn and ZnO? I could check if needed) which might react more directly with the iron ore to reduce iron, instead of going through H2 [...]

One such process is the iron oxide cycle
that heats iron (III) oxide, or rather ferrite, to 1400°C with sunheat to reduce it to iron (II) oxide, or rather its mix with an other metal. The second step that makes hydrogen and ferrite would drop away.

It lets me hope that sunheat can do at least a third of the job towards metallic iron, at a higher temperature if needed - a blast furnace is hotter than that. Saved 1/3 coal and CO2 emissions.

Whether, when and how the ore's other components should be removed, I dunno.


An other process is the zinc–zinc oxide cycle–zinc_oxide_cycle (–zinc_oxide_cycle)
where sunheat decomposes ZnO at 1800°C to metallic zinc and oxygen. The second step that makes hydrogen and ZnO would drop away. Gaseous zinc would reduce iron (II) oxides to metallic iron.

Whether Zn/ZnO is the right cycle to reduce FeO, I have no opinion. Other cycles exist, and a blast furnace is hotter than that. Sunheat would do all the job towards metallic iron.


To produce 10 000 t Fe, a set of furnaces may need 1.5×1014J in 8 hours daylight, or 5GW sunheat collected over 10km2, better as smaller units. The biggest Solar powerplants have similar areas and investments. Many iron producing plants are as big, and not in the desert.


Extracting so much iron ore under reliable sunlight, Australia could export FeO or Fe rather than ore. Emit no CO2, transport no coal, only iron.

Marc Schaefer, aka Enthalpy