This invention relates to the production of refined liquid steel directly from iron ore materials, such as iron ore fines, using natural or manufactured gas. The chemical reduction of the solid oxide feed and its transformation into refined liquid steel within a single melt circulation reactor on a truly continuous basis is made possible by contacting a thin layer or sheet of molten oxide containing ferrous oxide floating on a carrier medium of molten iron, which is moving preferably countercurrent to a turbulent high velocity reducing gas stream containing hydrogen as its principal gaseous reductant either alone or in combination with carbon monoxide and other gases.

The aforementioned thin layer of oxide melt is formed almost immediately, when a preheated charge of iron ore materials, preferably but not necessarily iron ore fines which may previously have been benefϊciated physically to enrich its iron content, is added to the surface or alternatively injected subsurface into the molten iron carrier medium. When this occurs the immediate chemical reaction converts the charge material containing iron in whatever oxidation state it was initially into a melt in which the iron is present essentially in the ferrous state. The liquid ferrous oxide containing melt floats on the molten iron carrier medium and is transported away from the charge location along with the principal flow of molten iron along the reduction arm of a closed loop melt circulation reactor.

The initial principally liquid ferrous oxide diluted to a relatively minor extent with gangue oxide impurities is reduced progressively by the reducing gas as the oxide melt continues its passage along the length of the reactor with the iron oxide content decreasing until at the remote end from the charge location the now iron oxide deficient melt moves along with the carrier molten iron and overflows a weir to permit gravity separation of the two immiscible liquid phases, slag and molten iron. The molten iron continues on its way to be recirculated within the closed loop, while the slag accumulates downstream of the weir for intermittent tapping or perhaps it is removed continuously, possibly with the assistance of electromagnetic induction forces restraining the molten metal so that a clean separation of slag from the carrier molten iron is facilitated.

According to the dictionary (Chambers) "steel is iron containing a little carbon with or without other things". At present all commercial steelmaking is carbon based. The

dominant technology worldwide is the blast furnace (BfF) route to reduce iron ore materials with metallurgical coke to hot metal or pig iron containing 4 to 5% dissolved carbon along with other impurities. This is followed by batch refining of the hot metal in an oxygen blown steelmaking vessel to reduce the carbon level to typically around 0.04% before being subjected optionally to a further range of batch refining processes conducted in the transfer ladle in preparation for continuous casting.

Alternative iron sources (AI) such as direct reduced iron (DRI) pellets, hot briquetted iron (HBI) and sponge iron are gaining in popularity to supplement clean steel scrap in the electric arc furnace (EAF) route to steel, but the need still exists to decarburise the melt in the EAF process and this is seen to be advantageous in that the exothermic reaction of oxygen with dissolved carbon contributes a worthwhile amount of heat and thus reduces the electrical energy required.

It can be argued, however, that decarburisation in both the B/F and EAF routes to steel adds a further complication in the quest to go directly from iron ore material to liquid steel ready for continuous casting. This has long been the dream of steelmakers and it is generally referred to as "direct steelmaking", and by implication the technology ideally should be fully continuous.

This invention provides a means for achieving the direct continuous steelmaking objective in a single melt circulation loop of the type already proposed to satisfy a range of smelting and refining requirements in both ferrous and non-ferrous pyrometallurgy. The present invention has its foundations in coal-based smelting reduction for ironmaking originally proposed in EP 0266975 and US Pat. No. 4,701,217 and for integrated ironmaking with continuous steelmaking, again based on coal both as the source on energy and chemical reductant in PCT Patent Applications GB2003/003065 and GB2003/003069, published as WO 2004/00775 and WO 2004/007778 (US Pub. No. 2005/0269752 Al), respectively.

This latter technology employs three melt circulation loops to transform virgin iron ore materials to low carbon steel ready for continuous casting, with an in-line option for the provision of ultra-low carbon (ULC) steel if this is so required. Coupled with carbon abatement technology (CAT) in the longer term, this could lead to a near to zero gas emission future for integrated iron and steelmaking in response to growing concerns about

climate change and global wanning. Clearly, a similar approach can be applied to direct steelmaking from virgin ore with ore reductant and thermal functions being supplied by natural gas, substitute natural gas (SNG), synthesis gas or other manufactured gas. However, without the special provisions of the present invention, the opportunity for a real breakthrough in steelmaking technology is not realised and one ends up with just another direct steel process competing with what would appear at first sight to be less expensive options for steel production.

The present invention effectively does away with the need to incorporate a decarburisation step in the overall process and thus eliminates steelmaking, as presently constituted, altogether. Instead virgin iron ore materials are reduced to molten iron without carbon contamination in the very special way made available by the invention. The relatively small amount of dissolved carbon required to bring the molten iron up to the desired carbon level specification for a particular steel product range can be added later to the in-line refined molten iron once it has left the melt circulation loop. At the same time, other minor alloying elements can also be added, probably in association with continuous deoxidation preceding continuous casting. The technology involved here has been described in the publication "Towards coal based continuous steelmaking Part 2 - Low carbon to ultra low carbon steel", N. A. Warner, Ironmaking and Steelmaking, Vol. 30 (6), December 2003, pp. 435-440. It is to be appreciated that the molten iron withdrawn from the melt circulation loop may contain a relatively high level of dissolved oxygen, which can be dealt with using present state-of- the-art methods by controlled carbon addition followed by RH vacuum degassing, or alternatively on a truly continuous basis using the proposed Tower Refiner concepts outlined in the paper just referred to. In both cases, there is ample opportunity for controlled addition of carbon and minor alloying elements to meet desired specification levels dictated by the particular product range being processed at the time according to market requirements.

Natural gas is inherently a clean fuel and for power stations is supplied with low sulphur content. A similar quality natural gas comprised predominately of methane is available locally in many parts of the world or is available by pipeline transmission or as liquefied natural gas (LNG). However, indiscriminate addition or injection of iron ore fines and methane containing gas into a steelmaking reactor is not acceptable for direct steelmaking. The result would be an unacceptably high level of carburisation. Sekino, Nagasaka and Fruehan have published a detailed study of the kinetics of the reaction between methane

and liquid iron ("Kinetics of the Reaction OfCH ₄ Gas with Liquid Iron," Metallurgical and Materials Transactions B, Vol. 26B, April 1995. ρp.317-324.

It is also worth noting that processes for DRI and HBI production based on natural gas take steps to ensure that the natural gas is reformed to carbon monoxide and hydrogen or is at least partly oxidised with oxygen, in advance of the majority of the reducing gas actually coming into contact with freshly reduced solid iron at the operating temperature and pressure. Even so the DRI or HBI product normally has a carbon content above 1% and this can be as high as 5%.

If natural gas is steam reformed externally to carbon monoxide and hydrogen rather than being used directly in the steeimaking reactor, the immediate consequence of conducting the endothermic reforming reaction external to the steeimaking reactor is that an excessive amount of scrap melting capacity would have to be provided to satisfy the overall heat balance, if the reducing gas exiting the steeimaking reactor is not cooled and recirculated. This is a serious disadvantage if a straightforward direct steeimaking process with once- through gas flow for treating virgin iron ore materials is being sought. Steel scrap or AI is unlikely to be available at sufficiently low price to make such an overall steeimaking operation economically competitive with other emerging technologies, as for example coal- based direct continuous steeimaking of the type outlined in PCT/GB2003/003065, published as WO 2004/00775. Thus, at present for energy and cost reasons, there would appear at first sight to be no real alternative to using natural gas directly in the steeimaking reactor itself and accepting the consequential high degree of carburisation resulting from direct interaction of methane with the molten iron or steel.

However, the present invention overcomes the difficulty associated with carburisation and the implied requirement for decarburisation to reach carbon specification, by never letting methane and liquid iron come into direct contact with each other. The solution to the problem is provided by an adaptation of generic melt circulation technology.

Natural gas thermally decomposes at high temperature to carbon and hydrogen according to Equation 1

CH4 _(g ) - (C) + 2H _2(g) (1)

This is a moderately endothermic reaction with ΔH° varying from 89.6 to 92.5 kJ/mol over the reaction temperature range relevant to this discussion of say 700°c to 1300 ⁰ C.

The chemistry of Equation 1 can be effected in a melt circulation reactor system of the type already proposed by the author for many pyrometallurgical systems in the past, but more recently at TMS 2003 for direct coal-based steelmaking, N.A. Warner, "New Reactor Concepts for Direct Coal-based Continuous Steelmaking," Proc. Yazawa Int. Symp. On Metallurgical and Materials Processing, 132 ^nd TMS Annual Meeting, San Diego, CA, March 2-6 2003, TMS Warrendale, PA, 2003, Vol. 1, pp. 881-900, and at AISTech 2004, N. A. Warner. "Continuous Steelmaking based on Natural Gas," AISTech 2004 Proceedings, " Vol. II, pp. 419-431, for scrap-based continuous steehnaking. If iron ore is added onto the surface of the circulating molten iron on the reducing side of a melt circulation loop, it would be converted almost immediately to liquid ferrous oxide and then together with the other dissolved oxide impurities, it would float away from the charge location and be carried downstream. The energy for the gaseous reduction of the liquid iron oxide layer and indeed its initial formation from the added preheated ore is provided by the sensible heat of the circulating molten iron, which in turn receives the appropriate amount of heat by thermal radiation to a lightly fluxed surface of the melt in the post combustion arm of the melt circulation loop. The flux referred to in the present case not only achieves the desired prevention of oxidation of the molten iron but also, at the same time, may perform the useful service of dephosphorisation, a necessary prerequisite for many virgin iron ore materials, if low phosphorous steel is the desired product.

Unlike the glass ribbon in the float glass process, which attains an equilibrium thickness on a liquid tin bath, the liquid oxide melt in this case will wet the molten iron carrier medium and will thus spread out and cover the entire melt surface on both sides of the melt circulation loop. In effect, this provides a protective layer against contamination. Accordingly, carbon from methane, soot or indeed reducing gases are not given access to the molten iron and thus the normal decarburisation step, which is a feature of all other steehnaking processes including those based on DRI or HBI employing natural gas, is no longer a requirement.

Now, if the natural gas is forced through a restricted space above the oxide melt layer, starting at the low iron oxide end of a countercurrent contact, the principal initial reaction

involved will be the thermal decomposition reaction of Equation 1. As a first approximation and a very considerable simplification of the real dynamic situation, the reducing arm can be split into two zones, namely, a methane decomposition zone in series with a liquid iron oxide reduction zone. Both zones are identical in terms of provision of the heat requirements from beneath by the circulating molten iron. There is, however, a major difference in the type of chemical reaction involved in each zone.

Decomposition of methane occurs homogeneously in the bulk of the gas phase via various gaseous species or soot formation precursors. With due attention to detail, a system can be designed such that minimum soot deposition occurs on the containment walls but rather individual soot particles are first formed in the bulk of the gas phase and then these grow therein on further retention. Accordingly, convective heat transfer from the melt surface into the bulk gas phase provides the thermal requirements for the methane decomposition zone. These are enumerated using the well-established empirical heat transfer correlation given by Equation 2, in which Nu, Re and Pr are the dimensionless groups well known to those skilled in the art.

Nu = 0.023 Re ⁰ ' ⁸ Pr ⁰ - ³³ (2)

In the present case, because of 100% post-combustion efficiency in the post-combustion arm, the off-gas from the steelmaking reactor can be made slightly oxidising and thus the inherent problems normally associated with stickiness of partially pre-reduced material do not arise. There is adequate heat contained in the off-gases to support a very high degree of charge solids preheating up to around 1300 ⁰ C. However, a high preheat of iron ore fines implies a mild steel refractory-lined spouted fluidised bed or circulating fluidised bed and its associated preheated solids feed stand-pipe and the gas off-take associated with the steelmaking reactor. This must be balanced against a perhaps cheaper option of a simple unlined fluid bed and stand-pipe fabricated from ferritic chromium containing steel, for example, and restriction in the temperature to the 600-650 ⁰ C level, already commercially well proven for steam raising applications in advanced power generation.

The steelmaking reactor itself is truly continuous, but if so required, liquid steel could be withdrawn batch-wise in accordance with downstream processing requirements. In the present case these may include final compositional adjustment and/or deoxidation before

passage to a continuous casting facility. There is more than adequate storage potentially available for this intermittent removal of liquid steel in the "swimming pool" sized melt circulation steelmaking reactor.

Because liquid oxide and melt containing significant amounts of dissolved iron oxide are notoriously difficult to process because of the extremely aggressive attack on all commercially available refractory materials, the new process adapts the teachings of PCT/GB 2003/003069 (published as WO 2004/007778 and US Pub. No. 2005/0269752 Al) in which un-melted shells of solid steel are used for melt containment, rather than conventional or traditional refractories, with the associated heat removal necessary to maintain a steady-state thickness of un-melted steel in a stable condition by steam generation from strategically located boiler tubes receiving thermal radiation from the cooler surface areas of the steel shells. The associated heat flow is not to be seen as a heat loss but rather and as an advantageous way of generating steam for advanced steam turbine power generation. The electricity generated is used principally to satisfy the plant requirements for air separation in order to produce oxygen, possibly compression of carbon dioxide to a supercritical fluid in the longer term in accordance with projected carbon abatement technology to combat climate change and global warming and for more general plant usage. Under these circumstances the synergy between power generation and the process requirements for combating problems associated with the aggressive nature of liquid iron oxide and molten iron containing appreciable amounts of dissolved oxygen must be addressed at the design stage and the features defined by this invention incorporated so that a true symbiosis is captured by the process designer with the associated benefits ultimately reaped by the process operator.

The frozen shell or unmelted solid steel approach for melt containment, besides its superiority over refractories for corrosion resistance for melts containing iron oxide, has other significant beneficial attributes, which are incorporated advantageously in the present invention. These will now be alluded to so that it can be appreciated that the combination of gaseous reductants, such as natural gas, with melt circulation technology opens up the prospects of a real break-through in ferrous extraction metallurgy.

Natural gas is commercially available at such purity that sulphur transfer so the molten iron produced on reduction, is no longer the issue as it is with all current ironmaking procedures

and therefore the need for desulphurisation is completely eliminated. Also silicon transfer from the gangue oxide materials associated with iron ore feed to the molten iron produced on reduction does not occur to such an extent that major de-siliconisation is a necessary step in producing a refined liquid steel, as it is with all other liquid steel producing current technology. This stems principally from the very low carbon activity in the molten iron carrier material and its relatively high level of dissolved oxygen.

The aforementioned benefits are only achievable by never permitting the molten iron and the carbon containing gas phase from coming into direct contact with each other. The thin layer of oxide melt wets the surface of the molten iron carrier material and covers it completely, thus supporting an effectively impermeable protective barrier. In regions in which this protection is negated, as for example when the floating layer of oxide melt overflows with the carrier molten iron into a downstream region for slag/metal separation, appropriate steps must be taken to shroud such regions with an inert purge gas, such as argon. Accordingly, it may be necessary to recover argon in a pressure swing adsorption unit or other appropriate gas treatment facility processing the final gas effluent after removal of water vapour and carbon dioxide, but this will depend on the financial viability of such an installation.

Provision of a protective layer of molten oxide material between the gaseous reduction and the molten iron has, however, the potential for seriously inhibiting the rates of heat and mass transfer required to sustain the reduction process within the bounds of an acceptable gas/liquid surface area, bearing in mind that the two must co-exist stably as separate continuous phases and never be allowed to be dispersed in one and other.

Because the presence of even a film of oxide material can impede at least to some extent the heat and mass transfer processes necessary to effect the desired reduction process, it is imperative that the thickness of the melt layer is kept to an absolute minimum. The present invention ensures maintenance of such conditions by having melt circulation rates very much higher than has been considered acceptable in previous applications of generic melt circulation technology.

To date the various melt circulation loops associated with new proposals for ironmakdng and steelmaking as described in prior art, have all employed an adaptation of RH

(Ruhrstahl Hereaus) vacuum steel degassing to effect forced circulation of molten iron containing dissolved carbon around a closed loop. By association with RH technology, all these previous applications are based on refractory lined snorkels dipping into the molten ferrous material and thus by implication traditionally have accepted the need for periodic replacement of RH vessels with stand-by units, already at temperature so that thermal shock does not damage the refractories, when the stand-by unit is called into service.

For direct steelmaking based on virgin iron ore materials and natural gas, there is a need to increase the carrier molten iron velocity substantially above that previously associated with melt circulation in other iron and steelmaking situations, in order that the special requirement now identified as being crucial to the success of the invention, namely maintenance of the oxide melt layer being as thin as possible. The melt circulation ratio, defined as the mass rate of flow of the carrier molten iron material divided by the production rate of new iron, has typically in the past been less than 500 to 1 at the upper limit and normally somewhat less than this. The new process now being described will advantageously require a melt circulation ratio of 1500 to 1 or possibly higher for efficient operation.

Current state-of-the art for RH technology employs maximum melt circulation rates of about 200 tonnes per minute. Accordingly, provision of a number of RH type units operating in parallel would be required for melt circulation purposes in the present situation and this may not be financially viable. The present invention overcomes this difficulty by extending upwards both ends of the horizontal channels constituting the arms of the melt circulation loop an maintaining a differential pressure between the principal hearths or channels and a upper transfer launder system interconnecting both arms of the loop such that melt is forced into the vertical relatively short end channels and into the upper transfer launder system. In effect, so-called "barometric legs" are created by the vertical risers at the ends of both arms of the melt circulation loop. Admitting an inert gas into one of the vertical channels or end risers to create a two-phase flow region then effects melt circulation between the arms of the loop.

An embodiment of the present invention will now be described, by way of example only, with reference to the accompanying drawings in which:-

Fig. 1 is a schematic sectional plan view of a single melt circulation loop employing forced circulation of molten iron through a reduction arm and a post combustion arm interconnected at one end by a means for melt circulation such as a gas-lift device and an overflow or siphon at the other,

Fig. 2 is a schematic sectional elevation and plan view illustrating how a trapped gas space can be established initially as the basis for contacting liquid ferrous oxide melt with reducing gas in a simplified furnace hearth which uses force-cooled shells of unmelted solid steel for melt containment rather than conventional refractory material,

Fig. 3 is a schematic plan view of a pressurised melt circulation loop showing how two hearths, namely the reduction arm and the post-combustion arm, can be interconnected using a pair of upper transfer launders to constitute an overall circuit with pump action established in-situ by creating a two-phase flow region in one of the vertical channels connecting the principal hearths to the upper transfer launders,

Fig. 4 is a sectional elevation of the reduction arm shown schematically in Fig. 3,

Fig. 5 is a sectional elevation of the post-combustion arm shown schematically in Fig. 3.

Referring now to Fig. 1, the plant comprises a single melt circulation loop, a charge reduction arm 1 and a post-combustion arm 2 interconnected together by the gas-lift type pump 3 and a siphon 4 or other appropriate overflow device at the other end. Preheated iron ore fines 5 or other iron ore materials are added to the molten iron carrier medium 6 which flows around the closed loop melt circulation reactor. At the remote end of the charge reduction arm relative to the solid charge location, the molten iron carrier material and its associated now iron oxide-depleted melt layer overflow weir 7 into a phase disengagement region 8 from which molten slag 9 is either continuously or intermittently tapped. The preheated iron ore fines 5 are added continuously to the top surface of the molten iron carrier material via an appropriate distribution means 10 and are almost immediately chemically converted into liquid ferrous oxide containing normally a relatively minor amount of oxide gangue impurities. As the layer of iron oxide melt progresses along the reduction arm 1 floating on the molten iron carrier medium 6, it is progressively reduced and the molten iron thus formed joins the molten iron carrier

medium and, at the same time, the equivalent amount of molten iron is normally overflown or otherwise removed continuously from the melt circulation loop at any appropriate location, in this case via a product siphon 11. Preheated natural gas 12 or other manufactured gas, probably based on coal, petroleum coke or biomass material is added to the reduction arm 6 at the remote end relative to the solid charge location and is forced to flow at high velocity through a confined space having a gap of only a few centimetres between the top surface of the oxide melt and the roof of the reduction arm. This gap would normally measure 5 - 6 cm in height but perhaps somewhat larger depending on gas phase pressure drop considerations or even smaller if appropriate control means were developed and integrated into the gas flow circuit. The reducing gas not consumed along with products of reaction are passed through a crossover gas duct 13 into the post- combustion arm 2 either at a single location or more probably at various points along the post-combustion arm where preheated oxygen 14 is added again via a number of ports, such as 15, distributed along the post-combustion arm. Finally, the fully combusted off-gas is ducted at 16 to the solid charge preheater.

Referring now to Fig. 2, this demonstrates in simplified terms how a closed loop is constituted, omitting at this stage that separation of slag or spent flux from the liquid steel will be necessary. Melt circulation is induced along the length of the principal hearth 17 by admitting an inert gas to the bottom of a vertical channel 18 that connects the bottom principal hearth 17 to an upper level transfer launder 19 or return launder for liquid steel flow. At the other end of this return launder is a second vertical channel 20, which connects again with the principal hearth 17. This simple arrangement is shown as a sectional elevation in Fig. 2a. In Fig. 2a liquid steel fills both the principal hearth 17 and the upper return launder 19 and its associated up-flow 18 and downflow channels 20. Also shown in 2a are two vertical ducts 21, 22 behind the melt transfer launder 19, representing on the left-hand side one for admission of methane gas 21 and further downstream a vertical duct 22 for removal of product gases. The liquid steel rises to the same level in both these gas ducts corresponding to that in the vertical liquid flow channels and transfer launder system, assuming they are all open to the atmosphere or provided with gas at substantially atmospheric pressure. Figure 2c shows the arrangement of 2a in plan view. For purposes of illustration, the gas flow ducts are shown as pairs in this plan view, one arranged on each side of the return launder for liquid steel so that one of each is still visible behind the return launder in the sectional elevation shown in 2a. If the gas ducts 21, 22 are

now interconnected and a pressure above atmospheric pressure is admitted, the liquid steel is forced out of the gas ducts with the equivalent volume of liquid steel overflowing from the outflows 23 of the top launder system. By continuing to increase the applied pressure, the melt level recedes in the principal lower hearth so that a trapped gas space 24 is formed in the region between the top of the melt surface and the upper containment enclosure or roof. Now, provided the gas pressure is kept constant, a stable space for gas flow is established and gas admitted via the left-hand duct(s) will flow across the melt surface and discharge vertically upwards into the right-hand duct(s). Consider now that a liquid iron oxide melt is floating on the liquid steel and that preheated methane is admitted on the left and product gases flow out through the right-hand duct(s). With injected inert gas 25 admitted to the up-flow channel via a vertical lance 26 (not shown in Fig. 2), the liquid steel and the main reducing gas flow will move countercurrent to each other. Furthermore, if iron oxide is fed onto surface of the melt at the right of the principal hearth, it would quickly melt and chemical reduction would proceed along the principal hearth, the volume of liquid steel would increase and the corresponding amount of liquid steel product would overflow continuously from the outflow ports of the top return launder system.

What has just been described is effectively a single hearth furnace arrangement for reducing a pure iron oxide material with energy input not from post-combustion of the reducing gases leaving the reduction arm but rather from some other unspecified source such as electrical conductive heating not only merely to bring the reactor up to steady state operation but on a continual basis. The direct steel process actually being proposed would normally would use electrical conductive heating only for start-up and for holding at temperature for short periods to effect essential maintenance or after prolong stoppage once the steel melt has frozen in situ, preferably with the gas pressures maintained until a complete freeze up has been attained.

Referring now to Fig. 3, for a real natural gas based direct steel process, the return launder system 19 described in Figure 2, is replaced by a second principal hearth 2, probably but not necessarily identical to that described for the reduction arm 1, except that a much greater gas freeboard is needed for efficient radiative heat transfer to supply all the process energy requirements. This post-combustion (PC) hearth 2 would be at substantially the same level as the reduction arm. If there is an appreciable gas phase pressure drop across the reduction arm, where the gas space clearance may well be down to 5 - 6 cm to achieve

the desired intensity of reduction based on heat and mass transfer requirements, the reduction arm would be tilted probably no more than a degree or so in order to preserve uniform flow of the liquid phase against the total pressure gradient. This aspect is readily quantified by applying a mechanical energy balance to the flow system. Also the upper level launder system would be replaced with two shorter transverse launders interconnecting the flow between the reduction arm 1 and the PC arm 2. If the reduction and PC arms were operating close to atmospheric pressure, the upper transverse launders would be evacuated and the up-flow liquid channel would in flow terms resemble an RH upleg and the downflow liquid channel would be equivalent to the melt siphon of former arrangements.

For various reasons, including those relating to gasification of soot particles, an increased pressure of operation at 2 - 3 bar has many attractions. Not the least of these is associated with the physical provision of the unmelted steel shell arrangement for the up-flow and downflow channels and the interconnecting lateral transfer launders. Also, the two-phase pumping of the melt around the loop is facilitated by having available a barometric upleg of greater length than that sustained by a vacuum as in the RH system. Accordingly, a minor pressure (2 - 3 bar) operation is recommended with the upper launder system perhaps virtually at atmospheric pressure. This latter arrangement facilitates continuous overflow of liquid steel to continuous casting or perhaps to optional compositional adjustment including deoxidation and ultra low carbon (ULC) steel production ahead of continuous casting.

Fig. 4 is a sectional elevation of the reduction arm 1. It is worth noting that the inert gas flushing action afforded by the lance(s) 26 injecting argon or other inert gas into the molten iron will ensure that any hydrogen pick-up from the liquid iron, when first reduced from iron oxide, will be efficiently scrubbed from the circulating molten metal and thus removed from the system. Also the same comments apply to adventitious pick-up of nitrogen via air entering the circuit. Therefore, one can conclude that the hydrogen and nitrogen contents of the molten iron leaving the melt circulation loop at 11 will comply with the strictest specification limits for both of these elements. This is rather ironical, because at first sight it may have been concluded that hydrogen pick-up in particular would be a serious problem in what after all is effectively a predominately hydrogen based steelmaking process.

Now referring to Fig. 5, there are very new features in the diagram except that Fig. 5 is dominated by the need for a relatively large gas freeboard 29 above the melt surface to facilitate gas phase radiative heat transfer from the very hot post-combustion gases to the circulating molten iron 6, which is covered by a very thin layer of flux (not shown in Fig. 5) in order to increase the emissivity as well as providing protection from melt oxidation by direct gaseous interaction with the molten iron surface. Also, the very useful feature of inline dephosphorisation is accomplished by provision of an extended surface layer of liquid flux, which floats on the molten iron carrier medium with the spent flux and molten iron overflowing together across a weir 7 into the phase disengagement zone 8.

Industrial scale thermal decomposition of natural gas is not new technology. It has been used for many years in a semi-continuous mode, one reactor on-line whilst the other is being heated by combustion with hydrogen at temperatures up to 1400 ^υ C. Currently there is renewed interest in this type of reaction as a promising means for direct decarbonisation of fossil fuels in order to combat global warming and climate change by eliminating carbon emissions to the atmosphere and sequestering or storing solid carbon rather than CO ₂ . Because methane decomposition is only moderately endothermic, the decomposition reaction requires considerably less heat that the steam reforming process for manufacturing hydrogen from methane and it is claimed that less than 10% of the heat of combustion of methane is needed to drive the process. Carbon is an inert material under ambient conditions so it can be safely stored with nainimal ecological uncertainties.

It has been pointed out that sequestering CO ₂ in aquifers or in the ocean may become environmentally unacceptable. For ocean disposal OfCO ₂ , the surrounding pH could have significant harmful effects on marine life. Pressurising CO ₂ in underground wells conceivably could cause underground structural changes and allow catastrophic release of CO ₂ , which would cause asphyxiation to air breathing animals and humans. These are all matters of very considerably concern, which clearly will generate much debate hi the years ahead.

If direct decarbonisation turns out to be the preferred option, then gas-based steelmaking presumably will use hydrogen in the really longer term, but for the foreseeable future using natural gas in direct steelmaking, even without CO ₂ sequestration, would reduce greenhouse gas emissions down to a level of at least one third of those associated with

primary steelmaking based on the blast furnace route. Most environmental experts would probably agree that this goal in itself is worth pursuing. However, if it can be demonstrated that in parallel with the environmental benefits, significant relief in carbon emission charges is assured without adversely affecting costs to the extent that steel becomes noncompetitive against alternative materials, then an even greater incentive exists.

One mole of natural gas is a better proposition than two moles of hydrogen as a reducing agent in the present case. The solid carbon produced in the form of soot on thermal decomposition of methane, which is entrained in the gas flow, does a worthwhile service in reducing H ₂ O back to H ₂ and CO ₂ back to CO all the way along the path of the reduction zone and since both these secondary reactions are themselves endothermic, a greater temperature gradient is generated across the liquid oxide layer than would otherwise be the case with hydrogen alone. This in turn has a significant positive influence in promoting natural convection and ensures that liquid phase mass transfer does not impede the overall rate of gaseous reduction at the oxide melt/gas phase interface.

hi laboratory studies on gaseous reduction of iron oxides, the gas flow is normally set at a level above that to avoid reagent starvation at the lower end of the scale and by equipment limitations at the upper level. Under these conditions the gases emerging from an experiment are normally well removed from thermodynamic equilibrium.

A somewhat different situation would exist in a real direct steel operation. To avoid excessive natural gas consumption, a balance has to be struck between maintaining a large driving force for gas phase mass transfer throughout the reactor length and energy conservation. There is an obvious trade-off here, which needs to be optimised on a cost basis as well as on other grounds. Midrex™and HYL™ DRI processes both use extensive gas recirculation with natural gas or reformed gas make-up to maintain an optimum driving force for gaseous reduction. Because of the very high temperature involved in terms of differential between operating temperature and heat resistant alloy upper temperature limits, gas recycling is not considered a realistic option in the present case, so for the direct steel process being discussed the gas is once through plug flow. On the other hand, the liquid metal is completely back-mixed and circulates at a rate many times the actual steel production. Accordingly, the gas emerging from the reduction arm will be closer to equilibrium than normally the case for laboratory experimentation and therefore both the

forward and reverse chemical reaction rates need to be taken into account. The net forward rate must be well in excess of that to satisfy the demands of the gaseous diffusion processes occurring between the gas/liquid melt interface and the bulk gas phase. This time the mass transfer analogue of Equation 2 is all-important, i.e.

Sh = 0.023 Re ⁰ - ⁸ Sc ⁰ - ³³ (3)

in which the dimensionless group, Sh, Re and Sc are well known to those skilled in the art.

The process and plant according to the present invention are not solely applicable to steelmaking from iron oxide ores but are also applicable to stainless steelmaking and other alloy steelmaking, in general, from mixed oxide feed materials.

From the foregoing discussion, one skilled in the art can readily ascertain the essential features of this invention and without departing from the spirit and scope thereof, can make various changes and modifications.