WO2008017797A1 - Radiative post combustio - Google Patents

Abstract

Forced melt circulation of both the oxidic and metallic phases independently opens up the prospect of continuous processing up to 2200 °C using graphite lined reactors. Central to the new technology is oxygen-based combustion of CO with gas radiation from combustion zones to fully backmixed melt surfaces immediately below, from which CO is uniformly evolved, thus forming an upwards flowing laminar protective layer and then only mildly turbulent combustion zone upon subsequent addition of oxidant. Principal wall and roof areas are protected by reducing the already low values of the mass transfer coefficients by a factor of about 30, using CO transpiration under a total pressure gradient opposed to the direction of CO2 to provide service lives of 1-5 years. Post combustion offgases are quenched by entrainment in recycled CO2 using axisymmetric turbulent jets issuing from graphite-lined slots, readily accessible and easily replaceable on a routine periodic basis.

Description

RADIATIVE POST COMBUSTION

The present invention relates to metallurgical reactors in which reaction intensities are substantially lower than those associated with popular so-called "high-intensity" bath smelting reactors, such as characteristic of basic oxygen steelmaking and emerging smelting reduction technologies for iron and steelmaking from primary and secondary metaliferous feeds, such as lump ores, ore fines, partly reduced ore and metal-containing waste streams. The applicant has been promoting this alternative lower intensity technology for many years and, for example, has asserted at the Yazawa International Symposium (2003 TMS Annual Conference, San Diego U. S. A) that continuous steelmaking cannot sensibly be based on what Bessemer, himself, referred to as "a veritable volcano in state of active eruption."

The emphasis in the present invention is directed not at the smelting reactions themselves, but rather at the transfer of thermal energy from so-called "post combustion gases" into the molten phases involved, whether that be an oxidic melt such as slag, molten metals or combinations of both, in order to secure the lowest possible energy route practically achievable on a large scale.

For low energy consumption in ironmaking, attention must be directed at processes that produce carbon dioxide rather than mixtures of both CO and CO₂ gases and ideally, CO₂ should be formed within the ironmaking reactor itself without resorting to dilution of oxygen with nitrogen. This simple precept would appear to be violated in att the coal-based ironmaking processes currently under commercial evaluation.

Coal-based ironmaking using the smelting reduction approach was pioneered in Sweden and the "father" of smelting reduction is acknowledged to be Professor Sven Eketorp, of the Royal Institute of Technology, Stockholm. Eketorp presented a paper "Direct use of coal for production of molten iron" to an Institution of Mining and Metallurgy (IMM) conference in London in 1981. He cited four different processes attempted in Sweden. They all failed. His published response to a question that the applicant asked was:

"In reply to Professor N. A. Warner, there is very little hope of finding a process whereby we can deliver energy to the bath directly, i.e. by burning CO to CO₂." (Sven Eketorp, Report of discussion Extractive Metall. '81, D42). He went on to say: "If energy is produced by an oxidising reaction such as combustion of CO, the problem is to separate heat transfer and mass transfer.... The lining could not withstand heat pulsation and the FeO rich slag...."

In attempting to assess the likelihood of new ironmaking processes ever becoming serious threats to the iron blast furnace, it is necessary to keep Eketorp's comments in mind and realise that to date there has been a whole catalogue of not wholly successful attempts in various parts of the world. The frontrunners are Japan's DIOS process, Australia's HIsmelt process and the American Iron and Steel Institute AISI/US Department of Energy (DOE) process for coal- based ironmaking. Although each of these processes has its own set of extensive international patent cover, they are arguably all variants on a single theme. They are all incapable of 100 percent so-called "post combustion" of CO and CO₂ with technically pure oxygen within the ironmaking reactor itself without overheating and damaging the refractory lining and thus making the process inoperable. Quite simply, the accumulation of slag within the ironmaking reactor is the root cause of this problem in all these processes.

Slag accumulation results in the creation of a thermal barrier, which inhibits efficient heat transfer back to the site of the endothermic iron producing reaction in the liquid metal bath. The obvious solution is to float the iron oxide slag away as soon as it is formed. Clearly, generic melt circulation technology affords the mechanism for doing precisely this. Using established RH steel degassing technology, the melt can be pumped from a low hearth to a higher level and then allowed to overflow back to the lower hearth. Also, one side can be oxidising and one side maintained under neutral or reducing conditions. Consequently, if the fuel is burnt on one side, the combustion energy can be picked up by the melt and transferred to the other side. The melt acts as energy transfer medium. Slag formed on the higher side can be floated away as soon as it forms and allowed to overflow with the heavier melt into the lower hearth. These are exactly the requirements of an ideal smelting reduction reactor.

By using forced melt circulation of the molten metal phase, 100% post-combustion becomes immediately feasible without undue heat stress on the refractory lining of the reactor. This is because slag is not allowed to build up, but, as it is generated, it is removed continuously in a thin film floating on top of the flowing iron melt, overflowing with it into the lower hearth and thereby permitting high-intensity radiant heat transfer to what is effectively a clean molten iron surface. The applicant presented a discussion paper on melt circulation ironmaking encapsulating the aforementioned post combustion features to British Steel staff at the Teeside Laboratories in 1981. This is believed to be the same year that CRA disclosed that it had begun research on what is now the HIsmelt Process and it is also the same year that Professor Eketorp conceded defeat at the London IMM Conference for the concept of directly capturing post-combustion heat within a smelting reduction reactor.

The emphasis on post-combustion with technically pure oxygen is to facilitate carbon capture and storage. At the present time, there is a general consensus worldwide that there is sufficient evidence to link carbon dioxide with global warming, stressing the urgent need to take assertive action now to reduce global energy consumption. Climate change and global warming concerns are increasingly becoming major issues confronting the minerals industry. Accordingly, carbon capture and storage (CCS) must dominate any developing technology for smelting reduction in the general area of iron and steelmaking and related technology for lower energy production of stainless steel and ferroalloys. Most particularly, CCS must be taken into account from the outset for newly proposed technology for titanium metal production from ilmenite (FeTiOa) using fossil fuel both as the reductant and energy source. Enhanced post- combustion, totally reliant on radiant heat transfer between very high temperature post- combustion gases and melts associated with the proposed new technology, is the key to its implementation to secure potentially the lowest energy route to titanium metal available, as outlined in co-pending U.K. Patent Application No. GB0608080.8, which is entitled "Co- production of steel, titanium and high grade oxide". Therefore, the description of the present invention will be focused on this particular application, bearing in mind that by analogy the same principles can be applied by those skilled in the art to iron and steelmaking and other related fields.

The present invention reassesses the role of melt circulation in high temperature pyrometallurgy and concludes that there are metallurgical processes, which benefit from forced circulation of both slag and metal phases independent of each other and opens up the possibility that radiative post combustion to an oxidic melt or slag phase can be an attractive proposition in certain circumstances, rather than universally an impediment as implied in description so far of prior art, in which forced circulation at relatively high rates has never previously been suggested.

In terms of the two-film theory for interfacial heat and mass transfer between immiscible fluids (gas and slag), the present invention takes co-ordinated steps to reduce the thermal resistance of the slag "film" to convective heat transfer, such that the temperature drop across the slag film is normally no more than about 5O⁰C and preferably considerably less by substantially increasing the forced circulation rate of the slag above that perhaps originally envisaged, if it can be demonstrated that overall process benefits accrue. At a gas/liquid interphase receiving radiant energy from higher temperature gases representative of post combustion, the magnitude of convective heat transfer from gas to melt is normally relatively small compared to the radiant energy flux in what the applicant has termed "swimming pool reactors" with large gas free board to ensure a mean beam length for gas radiation, commensurate with the thermal load to be satisfied. In general terms, the emissivity of combustion gases in an isothermal enclosure depends on the temperature and (pco2L) and (pmoL), where L is the mean beam length and the first term of each product is the partial pressure of CO₂ and H₂O, respectively. Provision of adequate L, say in the region of 5 m for operation at 1 bar, is needed or proportionately less, if it is decided that operation above atmospheric pressure yields beneficial attributes commensurate with the additional cost.

Conducting pyrometallurgical operations in relatively low intensity reactors rather than the afore-mentioned high intensity reactors currently in vogue, presents opportunities just not available in compact reactors. Radiative post combustion is an important case in point. Admittedly, reactors of Olympic swimming pool dimensions are going to be needed for very large-scale operations. If these large reactors are lined with unmelted solid shells of the material being processed, the cost implications can be assessed in terms of the interest lost on the cash flow not realised, because of the hold-up of product within the process. This has to be balanced against the costs involved in conventional refractory lining of the reactors and the fact that unmelted shells of product material are indestructible as they can be replenished in situ during continued operations by controlled melting or freezing, employing electro-conductive heating in conjunction with steam raising or other heat removal means at high temperature.

In the proposed new titanium metal process, conventional refractories are not available for melt containment to satisfy the stringent demands of the technology. Similarly, there are no commercial refractories capable of withstanding the high temperatures in this particular application, where roof and wall temperatures are preferably 200⁰C or so greater than melt temperatures themselves. Suppliers of refractories are unlikely to guarantee their products, such as stabilised zirconia, which on temperature service grounds would appear suitable. This is because the new technology, although nominally continuous without cyclic temperature variations once steady-state conditions are reached, is still open to the possibility of adventitious thermal shock damage and the disastrous consequences thereof, effectively ruling out all known ceramic materials for the very high temperature (say 2100⁰C) service under the oxidising conditions associated with post combustion.

According to the present invention, transpiration of carbon monoxide gas across a porous carbon lining constituting the walls and ceiling of the reactor offers the required resistance to oxidation, whilst the exceptionally good mechanical and thermal shock properties of baked carbon or graphite at temperatures below about 2200⁰C and probably closer to the 2100⁰C level actually required for effective radiant post combustion to a melt with a surface temperature in the region of 1900⁰C or so in the preferred embodiment, ensure structural stability under the proposed very demanding conditions.

The applicant was a co-recipient of the Robert W. Hunt Award of the Metallurgical Society of AIME in New York during 1966 for a paper on the kinetics of the steelmaking reaction using the electromagnetic levitation technique. Central to this research was a definitive study of the rate-controlling step in the reaction of carbon dioxide with solid carbon. At temperatures around 1660⁰C₃ it was demonstrated conclusively that gas phase mass transfer exclusively controls the carbon oxidation reaction. In other words, a graphite sphere in the region of 1 cm diameter reacts with carbon dioxide at moderate gas flow rates at precisely the same rate as does a sphere of levitated molten carbon containing steel of the same size, provided the carbon level of the molten steel is above about 1 per cent. Furthermore, this reaction is virtually unaffected by total pressure and the rate is proportionally to In (1 + yco2) where yco2 is the mole fraction of CO₂ in the bulk of the gas phase. Interfacial chemical kinetics had absolutely no effect on the CCVcarbon reaction at the moderate gas flow rates used in the levitation studies.

Because of the reduced reaction intensities in the present invention as a direct consequence for opting for other than high intensity bath smelting, there can be no scope for chemical reaction kinetics involving dissociation of adsorbed carbon dioxide on carbon surfaces being used to slow down or inhibit the oxidation reaction. Thus gaseous molecular diffusion is exclusively the fundamental process involved in the interaction of the very hot post combustion gases and carbon-lined enclosure associated with radiative post combustion and thus the strategy for combating this must be directed towards drastically reducing the gaseous molecular diffusion rate.

The insight gained from the aforementioned experimental studies many years ago was a vital ingredient in coming forward with the present inventive step. Carbon monoxide transpiration decreases the mass phase transfer coefficient along lines quantitatively dealt with in texts, such as "Transport Phenomena" by R. B. Bird, W. E. Stewart and E. N. Lightfoot, John Wiley & Sons, Inc., New York, 1960, and this principle can be used with confidence to support configurational studies on various alternative options to identify a practical and effective radiative post combustion system in the present context. For the arrangement selected as the preferred embodiment, it can be shown that consumption of carbon in terms of advancement of the gas/carbon interface, because of the reaction with carbon dioxide in the post combustion gases can be reduced to mm/year levels with dense baked carbon or graphite materials, employing relatively moderate carbon monoxide transpiration flux levels. Rigid graphite insulation with low thermal mass and low thermal conductivity is consumed at a greater rate because of its lower bulk density, but it should not be ruled out as it may afford a lightweight alternative with replacement at say six monthly intervals instead of the 1-5 year time frame envisaged for graphite or baked-carbon dense materials.

Radiative post combustion is central to the technology recently proposed by the applicant (co- pending U.K. Patent Application No. GB0608080.8, which is entitled "Co-production of steel, titanium and high grade oxide"). This patent application purposely avoided claiming that one or other reductants for carbothermic reduction of ilmenite was fundamentally superior. The choice between carbon and natural gas, for example, would depend on local availability and costs. However, for the present invention the emphasis is directed at carbon with very low sulphur content such as that, which would become available, if decarbonisation of natural gas were ever introduced in the longer term to sustain a hydrogen economy. Wood charcoal, coal char, coal itself or perhaps petroleum coke, if comprehensively desulphurised, can all be identified immediately as alternative sources of carbon in the context of the present description.

The general arrangement of the overall titanium manufacturing process based on carbothermic reduction of ilmenite is shown in Fig. 1, so that the relationship of radiative post combustion with other in-line processing steps is made clear from the outset. The first of the in-line reactors is Smelting Reduction Reactor (SRRl), in which both the oxidic melt and metal phases are independently circulated to effect complete back-mixing of both so that unmelted shells of titanium oxycarbide at the solidus composition can be used for melt containment under virtually equilibrium conditions.

In the present preferred embodiment, SRRl may be optionally followed by a second melt circulation loop (SRR2), employed principally to increase the carbide content of the oxidic melt. There are various options for SRR2 including circulation of both the oxidic melt and metallic melt, circulation of the oxidic melt by itself, but in all cases the oxidic melt is force circulated to effect complete back-mixing of this phase, again to permit melt containment by unmelted shells of titanium oxycarbide at the solidus composition under virtually equilibrium conditions.

The next in-line melt circulation loop is for continuous vacuum refining (CVR) of the oxidic melt. The refined oxidic melt then proceeds to electrochemical deoxygenation to produce titanium metal, optionally preceded by a degree of carbide removal by controlled oxidation using CCVoxygen. Alternatively, the oxidic melt containing titanium carbide, forming essentially an ideal solution in a thermodynamic sense, is chlorinated to produce a very high- grade titanium tetrachloride of about 99.95% purity, if pigment manufacturers prefer to capitalise on their vast experience in making pigments via the vapour-phase oxidation route. Under these latter conditions, environmental problems associated with disposal of waste chlorides would be reduced by at least one order of magnitude, because of the high purity of the oxidic melt initially chlorinated.

In the preferred embodiment, radiative post combustion is conducted on both arms of the melt circulation loop constituting SRRl. Segregation of post combustion to a separate arm is not necessary in the present invention. Combining both smelting reduction and post combustion advantageously gives rise to a uniform flux of carbon monoxide gas rising from the surface of the oxidic melt into the gas phase above throughout the entire melt circulation loop. Reactor dimensions are carefully selected such that the carbon monoxide evolved flows upwards under laminar flow conditions. Post combustion eventually takes place by admission of a preheated mixture of recycled carbon dioxide and technically pure oxygen into the turbulent well-mixed zone immediately above the laminar flowing carbon monoxide. The Reynolds number (Re) is less than say 2000-3000 in the laminar zone, while in the combustion zone turbulence is established by both increased mass flow rate as well as being beneficially promoted by the influence of the injected or otherwise admitted oxidant gas mixture and Re is typically above 5000.

By the aforementioned means, effectively an inert blanket immediately above the oxidic melt is established, extending vertically upwards at distance large enough to contain the relatively low intensity metal splashing, associated with a degree of sub-surface nucleation and growth of carbon monoxide bubbles, in parallel with direct evolution of CO at the gas/liquid interphase. The overall effect is a reduction in splashing by at least one order of magnitude compared with traditional bath smelting reduction reactors. But most importantly, the splash is not given the opportunity to react with oxidising gases. It merely returns to the oxidic melt surface having moved upwards and then downward through the protective blanket of carbon monoxide and any metal associated with the splash sinks back into the circulating metal phase underneath. Product loses due to oxidation are thus reduced to zero in the preferred embodiment.

There is a vast patent literature concerning what is called "transpiration cooling". Transpiration cooling is used to lower the temperature of the walls of combustors, internal components of advanced gas turbines, external surfaces of spacecraft and rockets on re-entry to the earth's atmosphere and the like, to enable available materials (metals or ceramics) to be used under extreme temperature conditions. The present invention does not use transpiration to effect any cooling whatsoever. On the contrary, cooling containment walls in radiative post combustion would be totally counterproductive and would encourage deposition of materials from the gas phase to form troublesome solid accretions. The preferred embodiment uses transpiration of carbon monoxide gas through porous carbon walls merely to alter the mass transfer coefficient by imposing a bulk flow component in the opposite direction to the net molecular diffusion flux involving transport of oxidant from the post combustion gases towards the carbon or graphite walls and ceiling, in which radiative post combustion is carried out. The fundamental processes involved are restricted to the gas boundary layer associated with the contacting of the solid with a bulk gas phase in motion. If any cooling does take place, this is co-incidental to the main objective.

The mass fluxes involved in the transpiration must be considerably lower than those associated with transpiration cooling, otherwise radiant heat transfer from the hot post combustion gases to the containment walls and ceiling would drastically reduce the thermal efficiency of radiative post combustion.

It is also worth pointing out that the first pre-requisite is to have relatively low intensity mass transfer conditions by employing planar surfaces and of course relatively low levels of turbulence. It is well known that small particles of carbon (soot) are oxidised by CO₂ at very high temperatures at rates controlled by chemical reaction kinetics and not by gaseous molecular diffusion. Similarly, academic studies, designed to eliminate the effects of gas diffusion, impinge high velocity gas jets onto solid carbon or melt surfaces to study interfacial chemical kinetics. It must be obvious that these situations and related phenomena have to be avoided in radiative post combustion. In the limit, high intensity eventually favours chemical rate control, but such conditions are rarely met in practical engineering applications and are therefore believed to be irrelevant to the present invention.

In the preferred embodiment, the oxidic melt, comprised principally of titanium oxycarbide and a whole range of titanium oxides TiO, Ti₂O₃, Ti₂O₅ and TiO₂ all in virtually thermodynamic equilibrium with each other, is force circulated at a high rate somewhere in the region of 1500 to 1 nominal melt circulation ratio based on equivalent TiO₂ in product melt, for reasons not immediately obvious, if radiative post combustion were being considered in isolation, which may pertain in other embodiments utilising the same generic principles, such as processes relating the continuous iron and steelmaking. The new titanium metal technology demands temperatures, which are considerably higher that say 1400-1600⁰C, more representative of smelting reduction for steelmaking, for example.

In the present oxidic melt, the mole fraction of titanium carbide close to the liquidus temperature is typically between 0.1 and 0.15 at 1900⁰C, an operating temperature typical of SRRl in the preferred embodiment. Titanium carbide has an exceptionally high thermal conductivity (k) of about 43 W/mK at 1900⁰C, so its presence even at the aforementioned level in the oxidic melt is bound to increase k above normal values for current industrial titania slags, whose effective k in the liquid state has been taken as 5 W/mK by Pistorius et al. (J. H. Zietsman and P. C. Pistorius: Minerals Engineering, 19, Issue 3, March 2006, 262-279) in modelling and ilmenite-smelting DC arc furnace process. Extrapolation of available data indicates that k for pure TiO₂ is about 3 W/mK at 1900⁰C, so it seems reasonable to expect k in the present embodiment to be about 6W/mK. On this basis, the dimensionless Prandtl Number (Pr) is also about 6.

By comparison, for slag melts containing ferrous oxide associated with smelting reduction, Pr is typically around 16 and k values typically about 1.2 W/mK. The corresponding values for molten steel are Pr about 0.17 and k about 32 W/mK. This rather simplistic analysis illustrates that the thermal properties of the oxidic melt are superior to those of normal slags encountered in smelting reduction, but still far removed from the corresponding values for steel melts. Notwithstanding this, detailed assessment of the thermal resistance of the liquid film at the oxidic melt/gas interface indicates that convective transfer of heat through the liquid film is not going to be a serious impediment to efficient energy transfer in radiative post combustion. It is the combination of the relatively low intensity inherent with swimming pool reactors, coupled with the enhanced thermal properties of the oxidic melt in the preferred embodiment, which together give rise to this very advantageous position for effective utilisation of radiative post combustion in the proposed titanium technology.

The titanium process embodiments of the present invention will now be described, by way of example only, with reference to the following drawings in which: -

Fig. 1 shows schematically the arrangement of two melt circulation reactors (SRRl and CVR) in which radiative post combustion is used to transmit thermal energy released in post combustion of smelting reduction gases. A second melt circulation reactor (SRR2) for optionally increasing the carbide content of the melt in order to produce a higher-grade product in CVR is not shown.

Fig. 2 is a schematic sectional plan view of the first of the smelting reduction reactors (SRRl) showing the extensive flow area (rectangle abed) through which carbon monoxide produced by carbothermic reduction rises upwards.

Fig 3 is a schematic external plan view of the melt circulation loop reactor (SRRl) shown in Hg. 2.

Fig. 4 is a half-sectional elevation of the plan views in Fig. 2 and Fig. 3 of the melt circulation reactor (SRRl). Fig. 5 is a sectional elevation taken across the width of the melt circulation loop (SRRl).

Referring now to Fig. 1, the major portion of the thermal energy demand of the processes conducted in SRRl₃ SRR2 and CVR is satisfied by radiative post combustion. It can be shown quantitatively that for a plant producing 500,000 tonnes per annum of TiO₂ equivalent, SRRl requires two arms of a melt circulation loop, each arm being say 10m in width by about 50 m in length. For an oxidic melt depth of 0.5 m and an average velocity of 1.25 m/s, the oxidic melt flow is characterised by a Froude Number of less than unity, so the flow is sub-critical turbulent. With a radiative post combustion heat input on both arms of sufficient magnitude to satisfy the total thermal demand and a melt circulation ratio of 1500 to 1, the temperature drop across the oxidic melt "liquid film" associated with convective heat transfer into the bulk of the circulating oxidic melt is evaluated to be less than 10⁰C.

For the preferred embodiment, the very low temperature drop of 10⁰C clearly demonstrates that the thermal resistance for convective transfer of heat into the bulk oxidic melt from the melt surface, receiving radiant energy input from post combustion, is negligible. However, due to its critical importance in the present context, a brief statement concerning its evaluation is appropriate.

A comprehensive experimental and theoretical study of turbulence structure and transport mechanism at the free surface in an open channel flow was made Komori et al. (Int. J. Heat Mass Transfer, 1982, Vol. 25, (4), 513-520). Also, more recently Wang et al. (Computers & Fluids, Vol. 34, 2005, 23-47) in a theoretical evaluation of turbulent open channel flow with heat transfer by large eddy simulation have validated the use of Pr^{" I/2} in the evaluation of mean turbulent heat transfer coefficient near the free surface, as implicit in standard surface renewal theory. The Komori et al. equation can be shown to reliably predict data obtained on a full-size commercial plant employing vacuum distillation of the circulating lead of a zinc blast-furnace (J. G. Herbertson and N. A. Warner, Trans. Inst. Mn. Metall, C, Vol. 82, 1973, 16-20), which is the only known published work on high temperature mass transfer at a free liquid surface. The Komori et al. equation in conjunction with the analogy between heat and mass transfer using the inverse square root dependence validated by Wang et al. is the basis of the evaluation of the temperature drop under discussion, on which the viability of radiative post combustion is so dependent. Referring now to Fig. 2, the carbon monoxide rises upward from the melt surface 1 in a stable condition of laminar flow. The turbulent flowing post combustion gas zone begins in a horizontal plane above and is thus not visible in the drawing. For a gas freeboard of 5 m above the melt surface within the turbulent combustion zone, the radiative flux to the oxidic melt surface is estimated to be about 128 kW/m² for an oxidic melt at 1900⁰C and a combustion gas temperature of 2100⁰C. In the absence of CO transpiration, the carbon walls 2 and ceiling, enclosing the radiative post combustion zone above the oxidic melt, would be oxidised at a rate of 3.6 x 10^"6 kmol/s m² by combustion gases containing 85% by volume CO₂ and the balance principally CO. The combustion gases, including recycled CO₂ as well as recycled but not transpired CO in this hypothetical case, flow upwards on each arm at an average velocity of about 0.24 m/s. Under these conditions, the bulk gas phase Re is about 12000, indicating that the flow upwards has progressed from initial laminar flow through transition to well- established turbulence.

Assuming that pressurised clean CO is available for transpiration and that for every kmol TiO₂ equivalent some 1.70 k/mol of this CO is transpired through porous carbon walls and ceilings, the mean transpiration flux is about 1.6 x 10^"4 kmol/s m². With transpiration at this level, the mass transfer coefficient for transport of CO₂ to the carbon surfaces is reduced drastically from about 1.14 x 10^"3 to 4.22 x 10^"5 m/s. For a graphite wall this corresponds to 25 mm depletion of graphite wall thickness per year.

The CO used for transpiration in this particular embodiment comprises CO released in electrochemical deoxygenation for titanium metal production, optionally CO released in association with lowering the carbide content of the oxidic melt in advance of deoxygenation and CO recovered from the vacuum system associated with CVR These streams of CO represent significant thermal energy, so steps are taken to recover this as the streams are cooled, filtered and then compressed in preparation for preheating ahead of admission to the various faces of the walls 2 and ceiling all constructed from porous baked carbon or graphite. By these means, effective amelioration of the oxidation of the lining above the melt is ensured by drastically reducing the gas phase mass transfer coefficient.

Referring now to Fig. 3, shown in section is a single baked carbon or graphite lined gas offtake 3, through which post combustion gases are ejected under the influence of a total pressure gradient by means of slot jets into the quench zone immediately above (not shown). The hot gas off-takes 3 are designed 4 to be readily accessible and therefore easily replaceable employing standby units with consumable baked carbon or graphite linings 5 in place ready for immediate installation on a routine periodic basis. The objective is to achieve partial quenching of the post combustion gases from a temperature in the region of 2100⁰C down to a lower temperature level, say 1300⁰C to 1200⁰C immediately they leave the upper region of the post combustion zone. On flowing through a slot 4 an axisymmetric jet of hot gas is formed, which is capable of entraining recycled cooled combustion gas, principally CO₂ at around 125⁰C, to effect cooling of the post combustion gases containing various elemental vapours and compounds without ever allowing the gases to contact solid surfaces and form accretions.

By the aforementioned means, stable condensed dispersoids are formed in the bulk of the gas phase and quenched to a temperature level such that they are non-sticky and are entrained in the bulk gas, enabling the downstream use of conventional refractory lined gas off-take ducts. Energy recovery from the off-gases now at about 1300⁰C is achieved by heat exchange using a moving packed bed regenerative preheater or similar device in order to raise the temperature of recycled CO₂ and technically pure oxygen to about 1200⁰C, before injection into the radiative post combustion zone. This injected or otherwise added mixture of CO2 and O₂ promotes turbulence and supplies the oxidant for combustion of the evolved smelting reduction gases and CO associated with transpiration through the porous carbon walls and ceiling.

After regenerative heat exchange, the partially quenched combustion gases, still containing some CO, then become the carrier gas for fluid bed preheating of dried mineral concentrates, such as ilmenite under controlled oxygen potential conditions to say 1350⁰C to ensure removal of sulphur to very low levels in advance of feeding the preheated concentrate into SRRl .

Also shown in Fig. 3 are the two rows of lances 6 and 7 for establishing the gas/liquid two- phase regions associated with independent forced circulation of both the oxidic melt and liquid metal phases, respectively. It is important to appreciate that these lances, which are submerged in the appropriate liquids, are withdrawn upwards in the event of a prolonged stoppage and melt freeze over. Normally, electrical conductive heating is employed to maintain the oxidic melt close to the liquidus temperature for shorter periods of times, such as involved in replacement of the consumable baked carbon or graphite-lined gas off-take from the post combustion zone. Referring now to Fig. 4 and Fig. 5 showing in order starting from the bottom, the unmelted shell of titanium oxycarbide 8 of solidus composition for melt containment, the liquid steel phase 9, the oxidic melt 10, the laminar flowing carbon monoxide effusion 11 above the melt due to smelting reduction reactions, the turbulent radiative combustion zone 12 extending to the porous carbon walls 2 and porous carbon ceiling 13 through which transpiration of preheated CO occurs under a total pressure gradient.

In the central region of Fig. 4, the gas off-take from the combustion zone is shown in half- section. For a nominal production of 500,000 tonnes per annum of titanium dioxide equivalent, say the initial internal dimensions of the off-take slot were 1.2 m wide x 10 m length, at the end of the tenth day, gas phase mass transfer controlled oxidation at 2100⁰C with 85% CO₂ at 15% CO by volume would enlarge the slot 4 to about 2.7 m wide x 11.5 m length, for example. The carbon depletion rate is non-linear with respect to time and as the slot enlarges the carbon mass depletion rate progressively decreases. Initially, the gas phase Re is about 125,000 and by the end of the tenth day this has decreased to 98,000 so the flow remains very turbulent over this period. However, the carbon depletion cannot be allowed to continue indefinitely, because of the need to quench the post combustion gases 14 to a required lower temperature level by entraining recycled cooled combustion gas 15, principally CO₂ at around 125⁰C, within a specified distance before the slot jet impinges on solid surfaces 16 and forms accretions. For an axisymmetric turbulent jet the extent of cooler gas entrainment was studied in the now classic work of Ricou and Spalding (F. D. Ricou and D. B. Spalding, J. Fluid Meek, 11, 1961, 21-32). The relationship proposed by these authors was used to identify the approximate time frame for renewal of the off-take slot, using the hydraulic diameter for a slot in place of the diameter for a single round nozzle. If, for example, it was deemed necessary to quench the off-gases to a specified extent within a distance of say 10 m in the present case, the slot would need replacing on the eleventh day. The actual amount of carbon consumed annually is minimal at about 810 tonnes for the scale of production under discussion. By comparison, the projected annual consumption of carbon by the smelting reduction reactions works out at 191,500 tonnes.

For purposes of illustration only, the end results of computer evaluation of the preferred embodiment of radiative post combustion involving co-production of titanium metal and steel from preheated ilmenite concentrates using continuous carbothermic smelting in-line with electrochemical deoxygenation for titanium metal, will now be presented. The mass percentage chemical analysis of a typical ilmenite concentrate is>

TiO₂ 50.4; FeO 34.1; Fe₂O₃ 12.1; Al₂O₃ 0.55; SiO₂ 0.76; MgO 0.76; MnO 0.61;

P₂O₅ 0.01; CaO 0.04; Nb₂O₅ OAl; V₂O₅ 0.28; Cr₂O₃ 0.09.

The concentrates are fluid bed preheated and then fed at 1350⁰C into SRRl, which is operated nominally at one bar pressure and 1900⁰C. Melt is over flown or otherwise removed from SRRl to SRR2, which operates at nominally one bar pressure at 2000⁰C, where the carbide concentration is increased to facilitate vacuum refining in CVR at 1920⁰C and at operating total pressure of 1 mbar. The principal thermal demands are in SRRl, where a heat input of 612 MJ per kmol of TiO₂ equivalent is needed and CVR, where some 155 MJ per kmol TiO₂ equivalent is required.

SRR2 has been designed to have a relatively smaller thermal load of 55 MJ per kmol TiO₂ equivalent, because operation at 2000⁰C makes it more demanding to supply energy by radiative post combustion, if a maximum operating temperature of 2100⁰C or thereabouts is imposed so that structural graphite can be kept well below 2200⁰C₅ a conservative recommended upper temperature limit for engineering applications.

Carbon monoxide is exhausted from the CVR melt circulation loop under vacuum using a steam jet ejector system based on established steelmaking practise for oxygen top-blown circulating flow vacuum degassing of liquid steel. An overall energy balance on the entire process indicates that the steam consumed by the ejector pumping system accounts for 60-70 per cent of the steam generated in waste heat recovery, leaving after allowing for other minor steam requirements, about 15-20 per cent available for on-site power generation, if so desired.

Radiative post combustion for CVR is not possible directly within the melt circulation loop at total pressure of 1 mbar, so the circulating oxidic melt is returned back to atmospheric pressure using a barometric leg into a hearth of the required surface area to satisfy the requirements of radiative heat transfer from combustion gases derived from recycled and preheated carbon monoxide combusted with a mixture of recycled carbon dioxide and technically pure oxygen, also regeneratively preheated to a temperature of say 1350⁰C. The oxidic melt then flows upward back to the evacuated CVR melt circulation loop through another barometric leg. By these means, the total thermal demand for continuous vacuum refining of the oxidic melt can be supplied by radiative post combustion. If the oxidic melt flow were fiictionless, there would be no net potential energy consumption involved in this diversion of oxidic melt from vacuum to atmospheric pressure and back again to vacuum using barometric legs for the purpose.

In the preferred embodiment, all the energy requirements associated with the smelting reduction and vacuum refining thermal loads are provided by radiative post combustion. Using this approach, the door is open to titanium metal production, directly from ilmenite concentrates at an overall energy consumption about one third of the current best available technology employing the Kroll process.

Claims

1. A method for capturing energy released by complete combustion of CO within combustion zones of melt circulation reactors by direct radiative heat transfer to melts at temperatures up to 2200⁰C located immediately below the particular combustion zone and thus receiving direct gas radiation therefrom, which comprises the following essential technical features:

(i) provision of fully-backmixed melt circulation with melt surface dimensions and melt circulation rate such that the rate per unit bath area or flux of CO evolution is uniform throughout the length of the reactor thus providing an upwards flowing laminar layer of CO in which the Reynolds Number is less than 2000 so that a protective gas blanket typically in the range of 0.5 to 1.5 m in thickness is formed between the melt surface and the combustion zone in which gaseous oxidant (oxygen plus recycled CO₂) is admitted to the upwards laminar flowing CO with the effect that the Reynolds Number is increased to about 5000 and thus still only a relatively low level of turbulence established,

(ii) protection of the graphite walls and roof elements in contact with the oxidising post combustion gases, recognising that at the very high gas temperatures involved the oxidation of carbon is controlled exclusively by gaseous diffusion from the bulk gas through the boundary layer to the solid carbon surface and the gas phase mass transfer coefficient itself is dependent on the level of turbulence in the upwards flowing post combustion gases, which at the Reynolds Number of 5000 referred to in (i), is already restricted but realistically the already low mass transfer coefficient can be reduced still further by a factor typically around 30 by modest rates of transpiration of CO such as 1.6 x 10^"4 kmol/sm² through the graphite surfaces, opposed to the direction of CO₂ diffusion, under the influence of a total pressure gradient dictated by the permeability of the particular grade of graphite selected.

(iϋ) maintenance of high partial pressures of CO₂ in the post combustion gases by avoiding dilution with nitrogen using tonnage oxygen rather than air so that gas radiation is enhanced by the associated increase in emissivity and thus the area of melt surface required for a given endothermic reaction requirement is reduced, whilst at the same time because of the high concentration of CO₂ in the effluent gas, carbon capture and storage (CCS) is simplified.

2. A process plant employing the radiative post combustion method according to claim 1, in which an oxidic melt passes in sequence through a smelting reduction melt circulation reactor (SRRl), optionally through a second smelting reduction reactor (SRR2) and then to a continuous vacuum refining melt circulation loop (CVR), in which gas radiative heat transfer is not a viable option at a total pressure in the region of 1 mbar, so the circulating oxidic melt is returned back to atmospheric pressure using a barometric leg into the a hearth of the required surface area to satisfy the demands of radiative heat transfer from combustion gases derived from recycled and preheated CO combusted with a mixture of recycled CO₂ and technically pure oxygen with the oxidic melt then routed back upwards to the evacuated CVR melt circulation loop via another barometric leg, ideally an overall zero net potential energy consumption scenario, if the oxidic melt flow were frictionless, but in reality energy has to be added to the oxidic melt to combat friction losses.

3. A process plant according to claims 1 and 2, in which all the energy associated with the smelting reduction and vacuum refining thermal loads is provided by radiative post combustion.

4. A process plant according to claim 1, in which post combustion gases are ejected into a quench zone to cool the gases down to 1200-1300⁰C under a total pressure gradient by entrainment of recycled cooled combustion gases at around 125⁰C by means of slot jets immediately above the active combustion zone, which because of the increased gas velocities and associated turbulence can no longer be adequately protected against oxidation in the longer term to provide a service life of 1 to 5 years before replacement, as applicable to the graphite walls and roof elements, and therefore these hot gas offtakes are installed so they are easily accessible and readily replaceable employing standby units with consumable baked carbon or graphite lining in place ready for immediate installation on a routine periodic basis.

5. A process plant employing the method of radiative post combustion according to claim 1, in which the mineral being smelted is ilmenite or titaniferous magnetite and the two immiscible liquid phases are liquid un-refined steel as the denser phase in contact with an upper oxidic melt bulk phase composed principally of titanium oxides, titanium oxycarbide and various impurity oxides, all in solution in the liquid state.