Method and apparatus for producing an alkaline mixture comprising sodium silicate
By producing sodium silicate through the comminution, dehydration, and dehydroxylation of ores, the method addresses the energy and emissions issues of current production methods, enabling cost-effective and environmentally friendly production of alkaline activated and geopolymer cements.
Patent Information
- Authority / Receiving Office
- GB · GB
- Patent Type
- Applications
- Current Assignee / Owner
- CAVALIER MARCUS ALEXANDER MAWSON
- Filing Date
- 2024-11-20
- Publication Date
- 2026-06-17
Smart Images

Figure 00000000_0000_ABST
Abstract
Description
Field of the Invention The present invention concerns a method and apparatus for producing an alkaline mixture comprising sodium silicate. Sodium silicate is a generic name for a range of compounds with the general formula n(NajO) • SiOj, wherein n does not have to be an integer number, such as sodium metasilicate (NajSiOs) (when n = 1), sodium orthosilicate (Na4SiO4) (when n = 2), and sodium pyrosilicate (NagSijO?) (when n = 3 / 2). Background of the Invention Cement is the most widely consumed industrial product after clean water. Its main use is in the production of concrete for construction. Annual global production of cement now stands at about 4.4 Gt (i.e., 4.4 gigatonnes). Ordinary Portland cement (OPC), which currently accounts for over 95% of all cement manufactured, is made by a process comprising the thermal decomposition of limestone. Limestone is a rock or ore composed almost entirely of calcite, which is a mineral form of calcium carbonate (CaCOs). This thermal decomposition, called calcination, decomposes the calcite into calcium oxide, also known as quick lime (CaO), and carbon dioxide gas, according to the equation: CaCOs (s) CaO (S) + CO2 (gj [Eqn. 1] This process is highly endothermic, and therefore requires a large quantity of heat to proceed. At standard temperature and pressure (hereinafter, s.t.p.), the Gibbs free energy of the reaction of Eqn. 1 is highly positive, so that the limestone must be heated to a temperature above about 900 ’Celsius before the Gibbs free energy becomes negative and the reaction can proceed. This heat is usually supplied by burning a fossil fuel, such as natural gas, which is injected into the limestone until the limestone reaches the required temperature. Apart from the energy consumed, burning fossil fuel in this way generates its own very significant carbon dioxide emissions, which are additional to the carbon dioxide also liberated from the limestone according to Eqn. 1. Cement production is therefore responsible for at least 8% of all anthropogenic carbon dioxide emissions. In view of this and the present climate crisis, it is widely recognized that the carbon dioxide produced during cement manufacture should somehow be mitigated. Several proposals have therefore already been made for how to achieve this. One such proposal is to replace OPC with an alternative type of cement manufactured by a different process which produces lower carbon dioxide emissions. Amongst the most promising alternatives to OPC are alkaline activated and geopolymer cements. These have received widespread attention, firstly because they are as easy to use as OPC in the production of concrete, but also because concrete made using such alkaline activated and geopolymer cements often has superior performance to concrete made with OPC in several respects, such as in its compressive strength, durability and resistance to thermal variation and corrosion by acids and sulphates. Alkaline activated and geopolymer cements generally comprise at least two main constituents: a reactive aluminosilicate and an alkaline activator. Examples of reactive aluminosilicates include activated clays, blast furnace slag and fly ash. Whereas blast furnace slag and fly ash are both industrial waste products which generate no carbon dioxide emissions of their own during their production, supplies of blast furnace slag and fly ash are already diminishing as the industrial processes which produce them are replaced by more environmentally friendly alternatives. Moreover, since only limited amounts of activated clays occur naturally, they must otherwise be produced by the thermal decomposition of unreactive clays extracted from geological deposits. Global deposits of unreactive clays are so vast as to be essentially limitless, and the thermal decomposition of these clays releases water vapour from them instead of carbon dioxide, which are among the reasons why alkaline activated and geopolymer cements are considered attractive alternatives to OPC. However, the thermal decomposition of unreactive clays still requires heating them to a temperature of from about 500 to about 800 ’Celsius, and typically from about 650 to about 750 ’Celsius. At such temperatures, naturally occurring clays such as halloysite pSiOj-AhCk-AHzO) and kaolinite pSiOj-AhOa^HjO) thermally decompose into an activated clay, metakaolinite pSiOj-AkOa). Examples of alkaline activators include sodium hydroxide and sodium silicate. Whereas clays are widespread in nature, neither sodium hydroxide nor sodium silicate occurs naturally in geological deposits because both are water-soluble. They must therefore be manufactured, and are correspondingly more expensive than activated clays. This tends to make alkaline activated and geopolymer cements less economically attractive than OPC. Sodium hydroxide is a commodity chemical and its annual global production currently stands at about 62 Mt ( / .e., 62 megatonnes). However, the annual global production of sodium silicate is only about 10 Mt, which would only be sufficient to manufacture about 40 to 50 Mt of alkaline activated and geopolymer cements annually, even if the entire annual production of sodium silicate were dedicated just to the manufacture of alkaline activated and geopolymer cements. Current methods for producing sodium silicate are also highly energy intensive and generate their own carbon dioxide emissions. For example, sodium silicate can be produced by fusing ( / .e., melting) quartz sand ( / .e., silica, SiCh) and sodium carbonate (NajCOs) together at a temperature of between about 1350 and about 1450 ’Celsius, according to the equation: SiOz (sj + n NajCOs (s) n(Na2O) • SiO2 (s) + CO2 (gj [Eqn. 2] which therefore requires a large quantity of heat to fuse the starting materials and also releases carbon dioxide from the sodium carbonate. The United Nations Environment Programme Report on Eco-efficient Cements (Paris, 2017) therefore states (p. 22) that "Materials that could truly reduce CO2 emissions in the sector require the invention of a lower energy production process [for alkaline activators]". The present invention addresses this problem. As further background, ores of industrially important metals like iron and manganese also typically comprise significant proportions of one or more gangue mineral species, the most common of which are silicates. These silicate gangue minerals may include quartz and / or aluminosilicates, which may comprise one or more clays, for example. In the case of iron ore, wherein the iron is present as at least one of the mineral and mineraloid forms of iron oxide and iron oxyhydroxide (such as hematite, magnetite and limonite), quartz is the most common gangue mineral species found in the ore as well. Furthermore, ores of these metals may also comprise one or more siliceous minerals of the target metals (i.e., of iron and / or manganese) themselves. For example, if the target metal is iron, its ore may comprise one or more iron silicates, like minnesotaite (Fe2+3Si4Oio(OH)2) and stilpnomelane. Similarly, manganese ores formed by weathering of manganese carbonate minerals over geological time often also comprise siliceous manganese minerals, such as braunite (which has the overall chemical formula 3 Mn2O3 • MnSiOa) and rhodonite (which has the overall formula MnSiOa with substitutions of calcium cations for a minority of the manganese cations). However, the silicate content of any gangue mineral species and of siliceous minerals of the target metals has traditionally been separated out as a waste product when the target metal is extracted from its ore, either by prior beneficiation of the ore before one or more compounds of the target metal in the ore is reduced to obtain the corresponding metal in elemental form or by transferring this silicate content to a slag phase during reduction of the ore, or both. Other useful background information is contained in G. Pedro Smith et al.: "Reaction at High Temperatures between Air and Liquid Metal Solutions Containing Sodium. Effect of Solution Composition", J. Am. Chem. Soc., Vol. 77, No. 17, pp. 4533-4534 (1955) and Saburo Yuasa: "Spontaneous Ignition of Sodium in Dry and Moist Air Streams", Symposium (International) on Combustion, Vol. 20, Issue 1, pp. 1869-1876 (1985), the entire contents of both of which documents are incorporated herein by reference. Object of the Invention It is therefore an object of the invention to provide a method and apparatus for producing an alkaline mixture comprising sodium silicate. Description of the Invention Accordingly, in one aspect, the present invention provides a method of producing an alkaline mixture comprising sodium silicate. The method comprises comminuting an ore of a target metal into fines, wherein the ore comprises a siliceous mineral and at least one of iron and manganese as the target metal. The siliceous mineral may comprise the target metal itself and / or it may be silica and / or another silicate mineral present as gangue. The method also comprises dehydrating the ore and dehydroxylating hydroxylated compounds contained therein. The ore thus treated is then reacted in an inert atmosphere with an amount of liquid sodium in excess of the stoichiometric amount thereof required for a redox reaction between the liquid sodium and an oxide of the target metal from the ore, to precipitate out from the liquid sodium both the target metal in elemental form and other insoluble products comprising sodium oxide. The redox reaction is conducted at a temperature above about 320 ’Celsius to induce a reaction between the sodium oxide and silica derived from the siliceous mineral to produce sodium silicate. Depending on the molar ratio of the sodium oxide to the silica, the sodium silicate may comprise sodium orthosilicate and possibly also sodium metasilicate, for example. However, this reaction is also conducted below a temperature at which the liquid sodium can react with the oxide of the target metal to produce a ternary oxide thereof. At least some of the liquid sodium with sodium silicate dissolved therein is then phase-separated from the undissolved metal and the other insoluble products, before being reacted with at least one of oxygen and water to produce a mixture of sodium silicate and at least one of sodium oxide and sodium hydroxide. Thus the alkaline mixture comprising sodium silicate is produced from the silica content of siliceous minerals of the target metal in its ore and / or of siliceous gangue mineral species. Before being subjected to the above method, the ore may still undergo one or more processes of beneficiation, which may comprise known processes, to increase the proportion of the target metal in the ore relative to gangue. The method of the invention has at least the following advantages. The alkaline mixture comprising sodium silicate produced by this method is suitable for use as an alkaline activator for alkaline activated and geopolymer cements. However, most of the operations belonging to the method of the invention also form parts of respective methods for extracting iron and / or manganese from their respective ores. Thus the alkaline mixture comprising sodium silicate and these industrially important metals can both be produced simultaneously as co-products of each other. The energy consumed by both processes is therefore split between the production of these coproducts, which makes the method of the invention a low-energy method of producing an alkaline activator for alkaline activated and geopolymer cements. Because iron is the most widely used metal and manganese is the fourth most widely used metal after iron, aluminium and copper, co-production of the alkaline mixture comprising sodium silicate with either of these two metals allows the alkaline mixture to be produced at scale and in sufficient quantity for it to be economically viable or even economically advantageous to replace similarly large quantities of OPC by alkaline activated and / or geopolymer cements. Moreover, since no carbon or metal carbonate is used in the method of the invention, no carbon dioxide is produced as a result. This method is therefore at least carbon-neutral and makes no contribution to global greenhouse gas emissions. However, in some embodiments, the method of the invention may further comprise using at least some of the alkaline mixture to mineralize gaseous carbon dioxide by reacting the alkaline mixture with carbon dioxide to produce a composition comprising sodium carbonate and silica. For example, the carbon dioxide may be captured from atmospheric air and / or from one or more industrial processes, in order to mitigate greenhouse gas emissions. If so, the method of the invention can even have a negative carbon footprint overall. The liquid sodium may be produced by electrolysis, for example from sodium chloride, with the coproduction of chlorine gas. Electricity for the electrolysis may be provided by a source of renewable energy, such as wind or solar, or come from nuclear power, and therefore need not generate any greenhouse gas emissions. The method of the invention therefore need not produce any greenhouse gases at all. Because the alkaline mixture comprising sodium silicate is produced from the silica content of siliceous minerals of the target metal in its ore and / or of gangue mineral species, this silica content, which was previously treated as a waste product, is now consumed. The amount of waste produced during extraction of the target metal from its ore is thereby reduced in comparison to existing techniques, in which the silica content of the ore would instead have been diverted to mine tailings during beneficiation of the ore and / or to a slag phase during the ore's reduction to obtain the target metal. Reference is hereby made to the present applicant's co-pending UK patent application nos. 2417059.9 ("Carbon-Free Method and Apparatus for Producing Iron and Steel"; applicant's ref: NE-P-GB 001) and 2417063.1 ("Carbon-Free Method and Apparatus for Producing Manganese"; applicant's ref: NE-P-GB 008), the entire contents of both of which applications are incorporated herein by reference. Complete knowledge of the contents of these two co-pending patent applications is not essential, but may prove helpful, to understanding the present invention. The method of the invention will now be described in greater detail. Comminution Firstly, the ore is comminuted into fines to increase its surface area and to help separate mineral species in the ore from each other. Fines are generally considered to be less than about 6 or 7 mm across. Comminuting the ore may comprise a mechanical process, such as one or more of agitation, crushing, grinding, hammering, milling and rolling, as well as other similar processes. Alternatively or additionally, comminuting the ore may comprise a chemical process, such as one in which the ore is at least partially dissolved and / or particles bearing the target metal(s) are precipitated out from solution. Comminution produces fine ore particles with a range of different sizes. Smaller particles are more desirable because they have a greater surface area. This helps to accelerate dehydration and dehydroxylation of the ore particles, as well as their reaction with liquid sodium, described below. However, creating smaller particles with a greater surface area also requires more energy to be expended on comminuting the ore than would otherwise be expended on creating larger particles. The optimum size of particles will therefore be partly determined in any particular case by the most economic energy balance overall between comminuting the ore on the one hand and the next two stages in the method of the invention on the other. This energy balance will in turn depend in any particular case on the specific energy requirements of the comminution device or devices which are used, the prior chemical composition of the ore itself, and on the outcome of any beneficiation which the ore has undergone before comminution. If beneficiation has already resulted in finely comminuted ore particles (for example, having an average size of less than about 1 or 2 mm), then any further comminution will be unnecessary. Apart from increasing its surface area, comminuting the ore into fines can also change one or more of the physicochemical properties of the ore in a desirable manner. For example, if the ore is an iron ore, it may help to detach particles of iron oxide and / or oxyhydroxide from particles of gangue species also present in the ore. Comminution may also be effective not only in releasing water which is trapped in the ore, thereby helping to dehydrate it, but may also contribute to dehydroxylating hydroxylated compounds which the ore may contain, such as oxyhydroxides of the target metal(s) and / or clays present as gangue. The preferred range of particle sizes at the end of comminution can therefore be determined in any particular case from the best combination of overall energy expenditure with the desirable physicochemical properties of the resultant ore particles. Comminution may further comprise sieving, separating and recycling larger particles for further comminution into smaller particles. Dehydration and Dehydroxylation During and / or after its comminution into fines, the ore is dehydrated, and hydroxylated compounds contained therein are dehydroxylated, preferably completely, and at least to a high degree. This converts any oxyhydroxides of the target metal(s) in the ore into their oxides and avoids such other hydroxylated compounds as may also be present in the ore from reacting with the liquid sodium during the subsequent redox reaction between the liquid sodium and oxides of the target metal(s). It therefore minimises the production of gaseous hydrogen during this subsequent reaction, which might otherwise present an explosion risk. In some embodiments, dehydrating and dehydroxylating the ore may comprise heating the ore. If so, the ore is heated to a temperature of from 100 to 600 ’Celsius, inclusive. This is for several reasons, as follows. Firstly, it is above the boiling point of water. Water in the ore is therefore driven off as steam. Secondly, as the water vaporizes, if the water is trapped in pores within the ore, this can cause decrepitation ( / .e., fracturing) of the ore as the water expands into steam, thereby further increasing the surface area of the ore particles. Thirdly, since at atmospheric pressure, the boiling point of water is above the melting point of sodium, if the temperature of the ore is maintained above the melting point of sodium after dehydration and dehydroxylation, this also avoids the risk that the ore will quench the reaction of the oxide of the target metal from the ore with the liquid sodium by rapidly cooling the liquid sodium when the ore particles subsequently come into contact with the liquid sodium. Because the ore particles have a large surface area-to-volume ratio after comminution, they tend to heat up and cool down relatively quickly to any given temperature. However, if the ore is heated in order to dry and dehydroxylate it, the ore is also heated to a temperature of 600 ’Celsius or less. This has the advantage that it avoids undesirable and uncontrolled side reactions involving gangue mineral species, such as the thermal decomposition of any calcium carbonate present. In addition, heating the ore to too high a temperature also risks undesirably encouraging comminuted ore particles to sinter. For example, if the ore comprises at least one of the mineral and mineraloid forms of iron oxide and iron oxyhydroxide, 600 ’Celsius is below the Tammann temperature of any FejOa already present in the ore or formed during its drying and dehydroxylation, which is about 650 ’Celsius. It is also comfortably below the Tammann temperature of silica, which is calculated to be 759 ’Celsius and is found experimentally to be about 730 ’Celsius. Thus if silica is already present in the ore as gangue, for example, sintering the ore particles can be avoided. In order to reduce energy consumption, the ore is preferably heated to less than about 550 ’Celsius, more preferably less than about 500 ’Celsius and most preferably less than about 450 ’Celsius. Nonetheless, if desired, the heat-treated ore particles may be further comminuted back into fines after their heattreatment to reverse the effects of any sintering. On the other hand, 600 ’Celsius is an insufficiently high temperature for the thermal decomposition of the most prevalent silicate minerals of the target metal(s). For example, fayalite (FezSiO4) decomposes on heating in atmospheric air into iron (11,111) oxide and quartz only at temperatures of from about 800 to about 1000 ’Celsius. This is similar to the range of temperatures for the thermal decomposition of tephroite (Mn2SiO4), a manganese analogue of fayalite, which decomposes in atmospheric air into rhodonite, braunite and tridymite (a polymorph of SiOz) at temperatures of from about 800 to about 1100 ’Celsius. Iron silicates like stilpnomelane and minnesotaite, if any are present in the ore, tend to dehydrogenate before they dehydroxylate because dehydrogenation consumes OH groups which would otherwise be available for dehydroxylation. For example, in atmospheric air, minnesotaite dehydrogenates over a temperature range of about 300 to 700 ’Celsius. The evolved hydrogen reacts immediately with atmospheric oxygen to produce water vapour and leave an oxidised intermediate product called "oxy-minnesotaite", which retains the minnesotaite structure. Subsequent dehydroxylation of oxy-minnesotaite is complete by about 850 ’Celsius in atmospheric air and leads to its decomposition into iron oxides and silica. Similarly, the thermal decomposition of manganese silicates, such as braunite and rhodonite, only occurs at temperatures above 600 ’Celsius. For example, braunite thermally decomposes into hausmannite and silica at temperatures above about 700 ’Celsius. In the case of iron silicates, elemental iron can still be magnetically separated from diamagnetic gangue species after the subsequent reaction with liquid sodium because iron is strongly (and famously) ferromagnetic. However, in the case of manganese silicates, it is preferable that these should instead be converted into one or more trivalent manganese oxide(s) via an inorganic hydrometallurgical technique, described below, before the subsequent reaction with liquid sodium. This is because such trivalent manganese oxide(s) are strongly paramagnetic and may therefore be magnetically separated from diamagnetic gangue species before the reaction with liquid sodium, whereas the elemental manganese produced thereby is only weakly paramagnetic and therefore cannot be magnetically separated from diamagnetic gangue species after this reaction. If the ore contains any clay-like minerals as gangue, dehydrating and dehydroxylating the ore by heating it as described decomposes these minerals into their constituent oxides, silica (SiOz) and / or alumina (AI2O3), and / or compounds thereof, releasing their water of crystallization, which escapes as water vapour. As mentioned above, halloysite and kaolinite decompose into metakaolinite pSiOa-AkOa). Gibbsite decomposes firstly into boehmite (AIOOH) and then into alumina (ALOa). Iron oxyhydroxides also decompose into their respective oxides, releasing water vapour. For example, goethite (Fe(OH)a or FeO(OH)) decomposes into ferric oxide (FejOa). Conversely, if the ore contains magnetite (FeaO4), heating the ore in atmospheric air in order to dehydrate and dehydroxylate it may also result in at least partial oxidation of the magnetite into hematite (FeaOa). If the ore is heated in order to dry and dehydroxylate it, the total quantity of heat required for dehydration and dehydroxylation, and therefore the temperature to which the ore particles should be heated, the dwell time of the particles at any given temperature, and the rate of heating them from their initial temperature, will all depend in any particular case on the original chemical composition of the ore (after any beneficiation), and on the effects of comminution on the ore, including its degree of comminution and the temperature of the ore particles after comminution. For example, if the ore contained less water initially and / or has been more finely comminuted, it will require less heat to dry. Moreover, comminuting the ore also tends to raise its temperature by converting mechanical energy into heat through friction. Thus a preferred temperature for the heat-treatment may depend at least in part on the initial chemical composition of the ore, as well as on its degree of comminution. In some embodiments, the ore may comprise red bauxite, which is primarily an aluminium ore also comprising a significant proportion of iron, and which may therefore be considered to be an iron ore as well. If so, the ore may initially contain a significant admixture of organic matter, such as tree roots and humus. This is because red bauxite is typically found in near-surface deposits which are extracted by open-cast mining. Preferably, therefore, if the ore comprises red bauxite, dehydrating and dehydroxylating the ore are carried out in an oxygenated atmosphere, such as in atmospheric air, and the ore is heated to at least about 250 ’Celsius, more preferably at least about 280 ’Celsius, and most preferably at least about 300 ’Celsius, which is above the auto-ignition temperature of any wood present. Any organic matter mixed in with the ore therefore combusts to produce carbon dioxide gas and water vapour, leaving ash. Comminution of the ore and therefore of any associated organic matter ensures that large pieces of organic matter are broken up, encouraging their complete combustion and preventing the formation of localised hotspots. Combustion reduces the mass of ash remaining to between only about 0.5% and 2% of the mass of organic matter originally present. Therefore, if, for example, the ore initially contained an admixture of 10% organic matter, this is reduced to only 0.05% to 0.2% ash remaining after combustion. The main components of the ash are calcium carbonate (typically about 25% to 45% of the ash), potassium carbonate (typically up to about 10%), alumina, silica and iron oxide, with trace amounts of other metal oxides. These are therefore incorporated into the other inorganic gangue species already present in the ore. In such a case, dehydration and dehydroxylation are preferably also carried out by enclosing the comminuted ore as it is heated so that it is contained in its own atmosphere and is not exposed to the surrounding environment. This has the advantages that the carbon dioxide thus produced, as well as any fly ash, are not released into the surrounding environment, and that the carbon dioxide may be captured. The captured carbon dioxide may be reacted, for example, with a portion of the sodium oxide in the other insoluble products which are produced by the subsequent reaction of the ore with the liquid sodium, as described further below. If the ore comprises a manganiferous mineral, it may be processed using at least one of two different techniques before being added to the liquid sodium. These are a reductant-free pyrometallurgical technique and an inorganic hydrometallurgical technique. Both techniques convert different manganiferous minerals into a trivalent manganese oxide, which may comprise at least one of manganese (III) oxide (MmCh) and manganese (11,111) oxide (Mn3O4). Since both these manganese oxides are strongly paramagnetic, they remain in the process stream if diamagnetic gangue mineral species are subsequently magnetically separated from them before they are added to the liquid sodium. In some embodiments, therefore, if the ore comprises a manganiferous mineral, it may be processed via the pyrometallurgical technique. This comprises heating the ore in atmospheric air to a temperature of less than about 600 ’Celsius without using a reductant (such as carbon or hydrogen). Depending on the initial chemical composition of the ore, this has several effects as follows. If the manganese ore comprises tetravalent manganese oxide, heat-treating the ore under the stated conditions reduces this to a trivalent manganese oxide. For example, if the manganese ore comprises MnOz in the form of pyrolusite and / or ramsdellite, heat-treating the ore as described thermally decomposes the MnOz into MnjOa and oxygen, according to the equation: 2 MnOz (sj Mn2O3 (s) + ½ O2 (gj [Eqn. 3a] However, if the manganese ore already comprises one or more trivalent manganese oxides, the stated temperature range is not sufficiently high to reduce the trivalent manganese oxide further into manganese (II) oxide if in the presence of atmospheric air without a reductant, such as carbon or hydrogen. Thus if the manganese ore initially comprises at least one of bixbyite and hausmannite, the trivalent manganese oxides in these minerals remain unreduced under the stated conditions. If the manganese ore comprises romanechite and therefore comprises water, heat-treating the ore in this manner causes water molecules trapped within tunnels of the romanechite to diffuse out and be driven off as water vapour. This heat-treatment also tends to convert the structure of the romanechite into hollandite as the water escapes. On the other hand, the thermal decomposition of hollandite and of other manganese oxides such as cryptomelane which are isostructural to hollandite, only starts above about 600 ’Celsius and is complete by about 700 ’Celsius, producing manganese (III) oxide and compounds such as baria (BaO) which contain the cations previously trapped within the tunnels. In some embodiments, the ore may comprise rhodochrosite as the manganiferous mineral. If so, the ore is heated to more than about 200 ’Celsius, because the conversion of rhodochrosite into a trivalent manganese oxide starts at about 200 ’Celsius. Heat-treating the ore in this manner causes the MnCOg to react with atmospheric oxygen and produce bixbyite (M^Og) with the release of carbon dioxide, according to the equation: 2 MnCOa (S) + ½ O2 (gj MnjOa (S) + 2 CO2 (gj [Eqn. 3b] Since manganese carbonate does not have high thermal stability, this reaction has a relatively low activation energy of only about 17.9 kJ mo^MnCOs)1 and is complete by about 300°Celsius. However, at temperatures above about 330 ’Celsius, the bixbyite starts to reabsorb the carbon dioxide. Therefore, if the ore comprises a significant proportion (for example, at least about 4% by weight, in some cases at least about 8% by weight, and in other cases at least about 12% by weight) of rhodochrosite, in some embodiments, the ore may preferably be heated to less than about 320 ’Celsius to avoid such reabsorption. If so, and the ore further comprises another manganiferous mineral apart from rhodochrosite which can be converted into a trivalent manganese oxide at a temperature of less than about 600 ’Celsius, this other manganiferous mineral may also be converted into a trivalent manganese oxide by subsequently using the hydrometallurgical technique as well. Some further reduction of the bixbyite to hausmannite may also start to occur under the stated conditions. If the manganese ore comprises any iron, calcium and / or magnesium carbonates as gangue species, the iron and magnesium carbonates also thermally decompose at temperatures of less than about 600 ’Celsius into Fe20a and MgO respectively, releasing carbon dioxide, but the more thermally stable calcium carbonate remains unchanged as CaCOg. Following the same pattern, kutnohorite thermally decomposes into hausmannite and calcite (CaCOs), releasing carbon dioxide gas. The thermal decomposition of kutnohorite in atmospheric air starts at about 500 ’Celsius. The thermal decomposition of siderite in atmospheric air starts at about 440 ’Celsius and is complete by about 540 ’Celsius. Thus the temperature of the heat-treatment can be selected to ensure the thermal decomposition of any of the aforementioned carbonate minerals, apart from calcium carbonate, based on prior analysis of the chemical composition of the ore. In order to prevent the carbon dioxide gas released by the rhodochrosite and by any other associated carbonates from escaping, and thereby contributing to global greenhouse gas emissions, in such cases, the ore is heated in atmospheric air closed off from its surrounding environment, and the carbon dioxide is captured. The captured carbon dioxide may then be reacted, for example, in a carbonation reaction with a portion of the sodium oxide in the other insoluble products, or with sodium hydroxide derived therefrom, to produce a mixture of products at least comprising sodium carbonate. In all the different embodiments which comprise heating the ore, the respective decomposition temperatures are dependent not only on the initial chemical composition of the ore and the crystal structure of each mineral component thereof, but also on the effects of comminuting the ore into fines. In other words, dehydration and dehydroxylation of the ore is a result of a combination of mechanical action (milling, etc.) and heating. This is because the enthalpy of dehydroxylation decreases with decreasing particle size. Therefore, it may not be necessary to heat the comminuted ore particles to a particularly high temperature, in order to achieve a high degree of or even complete dehydroxylation, if the particles have already been finely comminuted. For example, under laboratory conditions, clay-like minerals decompose into their constituent oxides over a temperature range of about 400 to 600 ’Celsius. However, dehydroxylation of kaolinite commences after just 2 hours of concentric disk milling at an ambient temperature of only 25 ’Celsius and is completed under the same conditions after 10 hours of such milling. Even without such extensive comminution, heating to a temperature of from 100 ’Celsius to about 500’Celsius is therefore usually sufficient to achieve nearly complete, or complete, dehydration and dehydroxylation of iron ore. With more extensive comminution, a lower peak temperature during dehydration and dehydroxylation of only about 300 or 400 ’Celsius may suffice. Moreover, heating the ore can also contribute to its comminution. For example, if the ore contains quartz and the ore is heated to above the quartz inversion temperature of 573 ’Celsius, a small percentage increase in volume which accompanies the phase transition from a to p quartz at this temperature can help to split the iron ore apart. Comminution, dehydration and dehydroxylation therefore interact synergistically with each other. Comminuting the ore into fines on the one hand and heating the ore on the other may also be at least partially combined, to discourage sintering of the ore particles as they are heated. In other words, some or all of the comminution of the ore may be carried at the same time as when the ore is being dehydrated and dehydroxylated. For example, a dryer of the ore may comprise one or more comminution devices, such as moving parts for separating the ore particles as they are heated. On the other hand, some or all of the comminution does not have to be carried out near to or in the same location as heating the ore. For example, prior beneficiation of the ore in one location may already result in finely comminuted ore particles, which may then be delivered to a different location for dehydration and dehydroxylation, before the subsequent reaction of the treated ore with the liquid sodium. As mentioned above, if the ore comprises a manganiferous mineral, before being added to the liquid sodium, the ore maybe processed using at least one of the reductant-free pyrometallurgical technique already described above and an inorganic hydrometallurgical technique, which will therefore now be described. Accordingly, in some embodiments, if the ore comprises a manganiferous mineral, it may be processed via this hydrometallurgical technique, which comprises adding at least some of the ore to hot, concentrated hydrochloric acid closed off from their surrounding environment to produce gaseous chlorine and an acidic aqueous solution comprising manganese (II) chloride. "Hot" in this context means that the initial temperature of the hydrochloric acid is at least about 15 ’Celsius and preferably more than about 23 ’Celsius above ambient temperature, in order to increase the rate at which the ore dissolves therein. The initial temperature of the hydrochloric acid is also preferably above about 50 ’Celsius, more preferably above about 60 ’Celsius and most preferably above about 70 ’Celsius. This has the additional advantage that the solubility in the acidic aqueous solution of the chlorine gas which is produced diminishes at higher temperatures. "Concentrated" in this context means that the initial concentration of the hydrochloric acid should preferably be at least about 2 M because at lower concentrations than this, the rate of dissolution of the ore in the hydrochloric acid may become unacceptably low. More preferably, the initial concentration of the hydrochloric acid should be at least about 3 M, more preferably still, at least about 4 M, and most preferably, at least about 5 M. For example, dissolving braunite in hot, concentrated hydrochloric acid proceeds according to the following equation: 1 / 7 [3 Mn2O3 • MnSiO3] (s) + 20 / 7 HCI (aq) -» MnCI2(aq) + 10 / 7 H2O M + 1 / 7 SiO2 (sj + 3 / 7 Cl2 (g) [Eqn. 4a] Dissolving the manganiferous mineral in the hydrochloric acid may be aided by stirring or otherwise mixing the ore particles into the hydrochloric acid. The rate of dissolution of the manganiferous mineral in the hydrochloric acid may be monitored and used to determine whether and / or when it has completely dissolved. For example, this may be done by measuring at least one of the flow rate of the chlorine gas thus produced and the pH and / or temperature of the resulting aqueous solution. The hydrometallurgical technique then comprises phase-separating from the acidic aqueous solution comprising manganese (II) chloride, a solid silica residue derived from the siliceous mineral. In the above example, this residue is represented by 1 / 7 SiOj (S). The phase-separation may be performed by at least one of settlement under gravity, filtration and centrifugation, for example. Sodium oxide, or sodium hydroxide derived from hydrating at least some of this sodium oxide, is then added to the acidic aqueous solution to produce an alkaline aqueous solution and a precipitate comprising manganese (II) hydroxide. In the case of sodium oxide, this reaction proceeds according to the following equation: Na?O (sj + MnCk (aq) + 10 / 7 H2O ( / ) Mn(OH)2 (S) + 2 NaCI (aq) + 3 / 7 H2O ( / ) [Eqn. 4b] If sodium hydroxide is added to the acidic aqueous solution instead, Eqn. 4b reduces to: 2 NaOH (S) + MnCl2(aq) -> Mn(OH)2 (S> + 2 NaCI (aq) [Eqn. 4c] At least some of the precipitate is then phase-separated from the alkaline aqueous solution. The phase-separation may be performed by at least one of settlement under gravity, filtration and centrifugation, for example. Potassium and barium cations derived from manganiferous minerals like hollandite and cryptomelane remain in the supernatant. At least some of the manganese (II) hydroxide is then dried and dehydroxylated in atmospheric air at a temperature of less than 100 ’Celsius to produce a trivalent manganese oxide according to the following equation: Mn(OH)2 (s) + V4 02(g) V2 Mn2O3(s) + H2O(g) [Eqn. 4d] wherein the oxygen is derived from the atmospheric air. Drying and dehydroxylating the manganese (II) hydroxide does not require the application of heat and can instead occur spontaneously in atmospheric air. The trivalent manganese oxide thus produced is y-MmOa, which does not occur naturally as a mineral and may only be produced artificially, as in the present case. The pyrometallurgical technique, in contrast, if applied to p-MnO2 (pyrolusite) produces a-M^Oa, which occurs naturally as the mineral bixbyite. In embodiments comprising the hydrometallurgical technique, the method further comprises neutralizing the solid silica residue with an aqueous solution of sodium hydroxide, washing the solid silica residue with deionized water and drying it, adding the washed and dried silica residue to the redox reaction with the excess amount of liquid sodium, as well as capturing at least some of the gaseous chlorine to prevent it from escaping into the external environment. As may be seen, therefore, the hydrometallurgical technique still comprises drying and dehydroxylating the ore. However, in this case, the manganiferous mineral in the ore is dehydroxylated by drying the manganese (II) hydroxide at a temperature of less than 100 ’Celsius and by drying the solid silica residue separately, before both are added to the redox reaction with the liquid sodium. The hydrometallurgical technique has the advantage that whereas the temperature of the pyrometallurgical technique is insufficiently high to thermally decompose manganese silicates, such manganese silicates may be converted into a trivalent manganese oxide via the hydrometallurgical technique instead. This allows diamagnetic gangue mineral species to be magnetically separated from the paramagnetic trivalent manganese oxide before it is added to the liquid sodium. Nonetheless, the hydrometallurgical technique still allows the silica content of these manganese silicates, as well as possibly of other siliceous minerals which may also be present in the ore as gangue, to be recovered as the silica residue which is added to the liquid sodium as well, even though silica is diamagnetic. The pyrometallurgical and hydrometallurgical techniques are therefore complementary to each other, and whether to use one or the other technique may be determined based on the initial chemical composition of the ore. Moreover, if the ore comprises a manganiferous mineral, such as rhodochrosite, which is better suited to the pyrometallurgical technique, and another manganiferous mineral, such as hollandite or braunite, which is better suited to the hydrometallurgical technique, then both techniques may be used to convert both types of manganiferous minerals in the ore into a trivalent manganese oxide. In such a case, however, the manganese ore is preferably subjected to the pyrometallurgical technique before the heat-treated ore is then subjected to the hydrometallurgical technique. This has the advantage of avoiding contamination of chlorine gas captured in the hydrometallurgical technique by carbon dioxide released from carbonate minerals in the ore. Moreover, if the manganese ore is subjected to the pyrometallurgical technique before being subjected to the hydrometallurgical technique, the heat-treated ore is preferably comminuted after heat-treatment and before being subjected to the hydrometallurgical technique. This has the advantage that if the reactivity of the manganese ore with the hydrochloric acid is reduced by reordering of its crystal structure as a result of its heat-treatment, then the reactivity of the ore with the hydrochloric acid may be increased again by comminution. Ideally, for the subsequent reaction of the treated ore with the liquid sodium, the ore would be completely dehydrated and dehydroxylated, since any water or hydroxylated compounds contained therein may otherwise react with the liquid sodium to produce a variety of different contaminants. However, complete dehydration and dehydroxylation may require more energy to be transferred to the ore than just a high degree of dehydration and dehydroxylation. The degree of dehydration and dehydroxylation to achieve can therefore be determined in any particular case from the overall balance between the energy required for dehydration and dehydroxylation on the one hand and the desired purities and percentage yields of the products from the subsequent reaction with liquid sodium on the other. The latter will in turn depend on the prior chemical composition of the ore itself, as well as on the outcome of dehydration and dehydroxylation. Thus a high degree of dehydration and dehydroxylation, rather than complete dehydration and dehydroxylation, may suffice in most cases. The degree of dehydration and dehydroxylation of the ore, or correspondingly of the precipitate if the hydrometallurgical technique is used, may be measured by repeatedly weighing the ore or precipitate, respectively, during dehydration and dehydroxylation, because they lose weight as they dry and as water vapour is driven off. A relatively rapid decrease in weight is therefore indicative of ongoing dehydration and dehydroxylation, whereas when the rate of weight-loss diminishes, this indicates that dehydration and dehydroxylation are nearly complete. The best conditions for dehydrating and dehydroxylating any particular batch of ore may be determined by sampling the ore before drying, and performing a loss-on-ignition (LOI) test on each sample thus collected. Alternatively or additionally, the degree of dehydration and dehydroxylation achieved may be measured by sampling the ore during and / or after drying, and performing a similar LOI test on each sample thus collected. If dehydrating and dehydroxylating the ore comprises heating the ore, the heat-treated ore particles may also require cooling before they are introduced into the liquid sodium. The temperature to which the ore should be cooled after heating can be determined from the process conditions for the reaction of the ore particles with the liquid sodium, which are described below. The time taken to cool the ore particles down will be determined not only by their temperature immediately after being heated, but also by the temperature gradient to which they are exposed in order to reach the target temperature for the ore particles before they are introduced to the liquid sodium. Because of the large surface area-to-volume ratio of the ore particles after comminution, the ore particles can be cooled relatively quickly to this desired temperature, just by being exposed to the ambient temperature of the environment. This may be achieved without exposing the dried ore particles to the surrounding environment itself by including a counterflow system within a dryer of the ore, whereby dried ore particles leaving the dryer are brought into thermal, but not physical, contact with ore particles entering the dryer at or near to ambient temperature, such that heat is exchanged from the dehydrated and dehydroxylated ore to ore entering the dryer. An example of such a counterflow system is given below. Such a technique for heating the ore in order to dehydrate and dehydroxylate it may optionally be combined with one or more other known heating techniques. Reaction with Liquid Sodium After its comminution into fines, dehydration and dehydroxylation, the ore thus treated is reacted with an excess of liquid sodium. This may be done, for example, by introducing the comminuted, dehydrated and dehydroxylated ore into a bath of liquid sodium. Oxides of the target metal(s) in the ore thus treated are reduced by the liquid sodium to produce the target metal(s) in elemental form and the liquid sodium is oxidized to sodium oxide (NajO). In the case of trivalent oxides of the target metal(s), such as hematite and bixbyite, this redox reaction proceeds according to the equation: 6 Na ( / ) + MjOajs) 3 NajO (S) + 2 M (S) [Eqn. 5] wherein M represents the target metal(s), i.e., iron and / or manganese. In the case of other oxides of the target metal(s), such as magnetite and hausmannite, in which the target metal(s) are in a lower oxidation state, the reaction products are still the same as in the righthand side of Eqn. 5. In other words, the initial oxidation state of the oxides of the target metal(s) in the comminuted, dehydrated and dehydroxylated ore is immaterial to the eventual outcome of this reaction. Since the solubility of iron in liquid sodium is only a few ppm by weight and the solubility of manganese in liquid sodium is only about 0.01% by weight, any iron, as well as virtually all of any manganese, produced by the reaction of Eqn. 5 remains in solid phase in the liquid sodium. The solubility of sodium oxide in liquid sodium is low but increases with temperature. At temperatures below about 400 ’Celsius, however, the solubility of sodium oxide in liquid sodium is less than about 0.1% by weight of oxygen. Almost all the sodium oxide produced by this reaction therefore remains in solid phase as well, leaving only a trace amount dissolved in the excess liquid sodium. (The exact amount remaining depends on the temperature at which the reaction of Eqn. 5 is carried out and on how much liquid sodium is present.) The insolubility of all the reaction products of Eqn. 5 in liquid sodium ensures that this reaction is one-way. To ensure that all the oxides of the target metal(s) from the ore are consumed, the amount of liquid sodium present should be in excess of the stoichiometric amount thereof required for this redox reaction between the liquid sodium and each oxide of a target metal. This excess amount of liquid sodium also absorbs some of the heat generated by the reaction of Eqn. 5. The excess amount of liquid sodium to use can therefore be determined from the desired temperature profile of the reaction, described below. In general, the reduction of the oxides by the liquid sodium to produce the target metal(s) in elemental form is exothermic, but not violently so. For example, if the only oxide present were hematite, in which iron is in its most oxidized state (FejOa), this reaction produces about -73 kJ of heat per mol of liquid sodium consumed. For comparison, the reaction of elemental sodium with oxygen in atmospheric air produces about -208 kJ of heat per mol of sodium consumed. This reaction of sodium with atmospheric oxygen is therefore about 2.85 times more exothermic than the reduction of FejOa by liquid sodium. After the ore has been dehydrated and dehydroxylated, because the ore originally comprised a siliceous mineral, what is added to the liquid sodium may comprise any one or more of the following, in addition to an oxide of the target metal and possibly also alumina (AI2O3): silica (SiOz), dehydrated aluminosilicates like metakaolinite, iron silicates like fayalite, stilpnomelane and minnesotaite, which, depending on the temperature reached during the dehydration and dehydroxylation, may have been thermally altered into species like oxy-minnesotaite, and manganese silicates like tephroite, braunite and rhodonite. The liquid sodium attacks the crystal structure of these siliceous compounds, which are therefore reduced by it to produce the target metal(s) in elemental form, sodium oxide and silica (SiOz). The redox reaction of Eqn. 5 is conducted at a temperature of at least about 320 ’Celsius, more preferably at least about 340 ’Celsius, and most preferably at least about 360 ’Celsius, to induce a reaction between the sodium oxide and the silica to produce sodium silicate, as will be described shortly. However, the reaction of Eqn. 5 is also conducted at a temperature below that at which the liquid sodium can react with the oxide of the target metal to produce a ternary oxide thereof, as is also described further below. Cations of group I and II metals like potassium and magnesium, released, for example, by the reduction of stilpnomelane, behave as follows. The potassium cations are reduced by the liquid sodium to elemental potassium because the equilibrium oxide in the sodium-potassium system is sodium oxide (Na2O). The potassium metal thus produced is completely miscible with liquid sodium, so no potassium oxide is produced as a solid phase and the elemental potassium instead remains in solution with the liquid sodium as sodium-potassium alloy (NaK). In contrast, magnesium cations scavenge oxygen anions from the sodium oxide to produce magnesia (MgO), which, like sodium oxide, has very low solubility in liquid sodium, so remains in solid phase. At temperatures of from about 320 to about 350 ’Celsius, the silica reacts in a Lux-Flood acid-base neutralization reaction with the sodium oxide to produce sodium orthosilicate (Na4SiO4), as well as possibly also sodium metasilicate (Na2SiO3), according to the equations: 2 NajO (S) + SiOz (Sj -> Na4SiO4 (S> [Eqn. 6a] Na2O (sj + SiO2 (s) Na2SiO3 (s) [Eqn. 6b] Sodium orthosilicate is produced as the only product of these two reactions if the molar ratio of Na2O to SiO2 exceeds 2:1, whereas a mixture of sodium orthosilicate and sodium metasilicate is produced at lower molar ratios than this. In either case, sodium ortho- and metasilicate are both soluble in liquid sodium upto a saturation concentration of at least about 5%, which increases with temperature, and therefore readily dissolve in the liquid sodium. A disproportionation reaction between the liquid sodium and silica to produce sodium metasilicate and elemental silicon according to the equation: 4 Na ( / ) + 3 SiO2 (S) -> 2 Na2SiO3 (S) + Si (S> [Eqn. 7] can also occur at higher temperatures of from about 520 to about 550 ’Celsius. However, the activation energy for the reaction of Eqn. 7 is significantly higher than for the reactions of Eqns. 6a and 6b, and any elemental silicon thus produced will also tend to contribute to reducing the oxide(s) of the target metal(s) and therefore be oxidized back to SiO2, leaving sodium orthosilicate, as well as possibly sodium metasilicate, as the only remaining siliceous reaction products dissolved in the liquid sodium. If any dehydrated aluminosilicates like metakaolinite remain after dehydration and dehydroxylation, these are similarly attacked at temperatures above about 300 ’Celsius by sodium oxide produced by the reaction of Eqn. 5 to produce alumina, sodium orthosilicate, as well as possibly also sodium metasilicate as just described. The direct reduction by liquid sodium of either silica to produce elemental silicon or alumina to produce elemental aluminium in a manner similar to the reaction of Eqn. 5 is not thermodynamically favoured, since the Gibbs free energy of both such reactions, AG >0 across the range of available operating temperatures for the reaction of Eqn. 5. A disproportionation reaction similar to that of Eqn. 7, but instead between liquid sodium and alumina to produce sodium aluminate (NaAIO2) and elemental aluminium according to the equation: 3 Na w + 2 AI2O3 (s) -» 3 NaAIO2 + Al [Eqn. 8] has a Gibbs free energy, AG, which is only slightly negative across the range of available operating temperatures for the reaction of Eqn. 5 and has even slower reaction kinetics. Any effects of the reaction of Eqn. 8 may therefore effectively be ignored as negligible. Moreover, whereas any elemental aluminium which is able to form by this or any other mechanism could in principle dissolve in the liquid sodium, it is likely to be oxidized back to insoluble AI2O3 by reducing the oxide(s) of the target metal(s), SiO2 or Na2O. If any alumina is present in the reaction mixture, therefore, it too remains in solid phase. Subject to the following constraints imposed by the chemical composition of what is added to the liquid sodium, the reaction of Eqn. 5 is preferably carried out at or above the middle of the range of available operating temperatures, to increase the rate of reaction. The rate of reaction may also be increased by stirring or otherwise mixing what is added to the liquid sodium into the liquid sodium. If the ore comprises at least one of the mineral and mineraloid forms of iron oxide and iron oxyhydroxide, the redox reaction is conducted at a temperature of less than about 450 ’Celsius. Above about 720 K, reacting the liquid sodium with iron oxides from the iron ore risks producing ternary oxides like Na4FeO3, as the Na - Na?O - Fe system transitions to a Na - Na4FeO3 - Fe system. Such ternary oxides are generally soluble in liquid sodium, which both contaminates the liquid sodium and reduces the yield of elemental iron. In such embodiments, the temperature of the reaction should therefore preferably be controlled to remain below about 705 K, more preferably below about 690 K, and most preferably below about 675 K. In embodiments in which the ore comprises red bauxite, the ore may contain a small admixture of calcium carbonate and / or potassium carbonate as a result of the combustion of organic matter during dehydration and dehydroxylation of the ore. However, at temperatures below about 450 ’Celsius, these gangue species fail to react with the liquid sodium and also precipitate out. In the case of potassium carbonate, it has been established that no such reaction occurs until well above this temperature. In the case of calcium carbonate, the onset temperature for a reaction between liquid sodium and powdered calcium carbonate is 717 K, at the very top end of the range of temperatures for the redox reaction of Eqn. 5 when the ore comprises at least one of the mineral and mineraloid forms of iron oxide and iron oxyhydroxide, as is the case for red bauxite. In red bauxite, n(Fe?O3) = n(SiO?) present as gangue, where n denotes the number of mol of each respective component thereof. However, the stoichiometry of Eqn. 5 implies that 3 mol of Na?O are produced by this redox reaction for every mol of FejOs consumed. Typically, therefore, the molar ratio of Na?O: SiO? comfortably exceeds the minimum 2:1 ratio for all the sodium silicate produced by the reactions of Eqns. 6a and / or 6b to be in the form of sodium orthosilicate, leaving approximately 1 mol of unreacted Na?O remaining in the other insoluble products per mol of FejOs consumed, assuming that enough excess liquid sodium is also used to dissolve all the sodium silicate thus formed at the temperature of the liquid sodium. Nonetheless, if desired, in some embodiments, after analysing a sample taken from a particular batch of red bauxite, the molar ratio of FejOs: SiO? therein is found to be unusually low, this molar ratio may be increased by adding one or more iron oxide(s), for example in the form of mill scale, and / or another non-bauxitic iron ore, such as from a banded iron formation, to the red bauxite before the comminuted, dehydrated and dehydroxylated ore is reacted with the liquid sodium. This may be done, for example, to ensure that enough sodium oxide is produced by the redox reaction to react with all the silica in the ore, to alter the molar ratio of sodium orthosilicate to metasilicate produced by the reactions of Eqns. 6a and / or 6b, or to increase the amount of unreacted sodium oxide remaining in the other insoluble products. If the ore instead comprises a manganiferous mineral, the redox reaction is conducted at a temperature of less than about 600 ’Celsius. Above about 870 K, reacting the liquid sodium with a trivalent manganese oxide(s) risks producing ternary oxides like a-NaMnOj as the Na - NajO - Mn system transitions to a Na - NaMnOj - Mn system. Such ternary oxides are generally soluble in liquid sodium, which both contaminates the liquid sodium and reduces the yield of elemental manganese. However, even at temperatures below this upper limit, there is still a possibility of the highly reactive sodium oxide from at least the reaction of Eqn. 5 producing small amounts of other compounds by undesirable side reactions with contaminants. In such embodiments, therefore, the temperature of the reaction should preferably be controlled to remain below about 820 K, more preferably below about 770 K, and most preferably below about 720 K. In embodiments in which the ore comprises a manganiferous mineral, depending on the initial chemical composition of the ore and on which of the pyrometallurgical and / or hydrometallurgical techniques have been used to dry and dehydroxylate the ore, a trivalent manganese oxide which is added to the liquid sodium may be accompanied by lesser amounts of one or more other manganese minerals. For example, if only the pyrometallurgical technique has been used, what is added to the liquid sodium may comprise manganese minerals with crystal structures comprising tunnels, like hollandite and cryptomelane, in addition to siliceous manganese minerals like braunite. For manganese minerals with crystal structures comprising tunnels, like hollandite and cryptomelane, the liquid sodium attacks the MnOg octahedra, which are reduced to produce elemental manganese and NajO, in a similar fashion to the reaction of Eqn. 5. This process also releases the cations of group I and II metals, like potassium and barium, contained within the tunnels. For minerals like rhodonite, in which a minority of the manganese cations are substituted by calcium cations, this process also releases the calcium cations. Cations of group I and II metals like potassium and barium released by the reduction of manganese minerals isostructural to hollandite, and calcium cations released by the reduction of siliceous manganese minerals like rhodonite, behave as follows. The potassium cations are reduced by the liquid sodium to elemental potassium which, as described above, remains in solution with the liquid sodium as sodium-potassium alloy (NaK). In contrast, barium and calcium cations scavenge oxygen anions from the sodium oxide to produce baria (BaO) and calcia (CaO), respectively, which, like sodium oxide, both have very low solubilities in liquid sodium, so remain in solid phase. Thus after all these reactions, the reaction mixture comprises the excess amount of unreacted liquid sodium with sodium silicate (and possibly also a minor amount of potassium) dissolved therein, and a solid phase, which is insoluble in the liquid sodium. This solid phase may at least partially precipitate out from the liquid sodium and / or be at least partially suspended in it as a colloid. The solid phase comprises the target metal(s) in elemental form and other insoluble products at least comprising sodium oxide. Depending on the initial chemical composition of the ore and on the temperature at which the redox reaction is carried out, the other insoluble products may further comprise any one or more of magnesia, baria, calcia and alumina, calcium carbonate and / or potassium carbonate, as well as any undissolved silica, if the sodium silicate has reached saturation solubility in the liquid sodium. It also follows that if what is added to the liquid sodium comprises both iron and manganese compounds, for example if the ore originally comprised both at least one of the mineral and mineraloid forms of iron oxide and iron oxyhydroxide and a manganiferous mineral, then the temperature of the reaction with the liquid sodium should be controlled to remain below about 450 ’Celsius to avoid forming ternary oxides of iron with sodium oxide, and preferably below about 705 K, more preferably below about 690 K, and most preferably below about 675 K. If the weight ratio of manganese to iron is at least 2:1, the iron does not form a binary phase with the manganese, which remains as a-manganese from ambient temperatures up to about 707 ’Celsius. Carrying out the reaction with the liquid sodium in an inert atmosphere has the advantages of preventing the liquid sodium from reacting with atmospheric oxygen, and of inhibiting the formation of ternary oxides. For example, the inert atmosphere may consist of at least one of nitrogen and argon. Either nitrogen or argon, or both, may be produced on site by pressure swing adsorption (PSA) of atmospheric air. An on-site PSA generator of such inert gases may be powered, for example, using waste heat derived from cooling at least one of the comminuted, dehydrated and dehydroxylated ore, fresh liquid sodium before it is introduced into the reaction mixture, the liquid sodium with sodium silicate dissolved therein, and at least one of the undissolved metal and the other insoluble products. The inert atmosphere may be maintained at around atmospheric pressure or just above. A small positive pressurization of the inert atmosphere in which the reaction is carried out to above atmospheric pressure, for example by about 10 to 25%, is desirable to prevent ingress of air from the environment by leakage. The vapour pressure, p, in pascal of liquid sodium as a function of temperature, T, in kelvin from its melting point up to 700 ’Celsius is given by the following equation: logio p = 9.71- (5377 / T) [Eqn. 9] Table 1 below gives some representative examples of the values of this vapour pressure across the range of possible operating temperatures for the reaction of Eqn. 5. As this table shows, the vapour pressure of liquid sodium over the range of possible operating temperatures is always less than about 2.1 kPa, or less than about 2.1% of standard atmospheric pressure: T / “Celsius 320 370 420 470 520 570 p / Pa 4.39 22.3 89.3 297 850 2145 Table 1 Thus if the reaction of Eqn. 5 is carried out at around atmospheric pressure or just above, the loss of liquid sodium from the reaction mixture by vaporization is negligible and the effect of any such vaporization on the pressure of the inert atmosphere in which the reaction is carried out is negligible as well. If any small amount of residual water (for example, water of crystallization) remains in the ore after its dehydration and dehydroxylation, at temperatures above the melting point of sodium hydroxide (which at 323 ’Celsius is at the very bottom end of the range of operating temperatures for the reaction of Eqn. 5), this residual water is reduced by the liquid sodium to produce sodium oxide (NajO) and hydrogen, according to the following equation: H2O + 2 Na(<) NajO (S) + H2 (gj [Eqn. 10] The solubility of hydrogen in liquid sodium increases over the operating temperature range, but remains very low even at the top of the range. Hydrogen gas which is released into the inert atmosphere above the liquid sodium will therefore reach an equilibrium with the hydrogen thus dissolved in the liquid sodium, according to the temperature of the reaction mixture and the corresponding partial pressure of the hydrogen gas above the liquid sodium. Thus hydrogen released by the reaction of Eqn. 10 may cause a slight increase in the pressure of the inert atmosphere. For safety and in order to maintain the pressure of the inert atmosphere similar to or slightly more than that of the surrounding environment, the pressure of the inert atmosphere should preferably be monitored, for example by means of a pressure gauge. To prevent any undesirable build-up of pressure, any reaction vessel in which the reaction of Eqn. 5 is carried out may be fitted with a pressure-relief valve, through which pressurised gas may be vented from the inert atmosphere to the surroundings. If so, the vented gases should preferably be cooled to below the dew point of sodium vapour before being released, to recover any sodium and to ensure that the remaining gases are harmless. On the other hand, if the pressure of the inert atmosphere above the reaction mixture needs to be increased for any reason, extra inert gas can instead be introduced into the reaction vessel along with the comminuted, dehydrated and dehydroxylated ore. Separation of Liquid Sodium with Sodium Silicate Dissolved therein from Undissolved Solids Liquid sodium with sodium silicate dissolved therein may be phase-separated from the undissolved target metal and the other insoluble reaction products using at least one of several different techniques. For example, this phase separation may comprise at least one of settlement under gravity, filtration (i.e., trapping) and centrifugation. For example, the present applicant's co-pending UK patent application no. 2417052.4 ("Apparatus and Method for Separating a Contaminant from Liquid Metal"; applicant's ref: NE-P-GB 002), the entire contents of which is incorporated herein by reference, shows and describes an apparatus and method for separating undissolved contaminants, for example in the form of suspended or entrained particulates, from liquid metal, such as liquid sodium. In another example, a liquid sodium centrifuge is shown and described on pp. 29 to 33 of Summary of the APDA Sodium Technology Program by J.E. Meyers published under United States Atomic Energy Commission Contract No. AT (11-1)-865, Project Agreement No. 11 (June 1970), the entire contents of which is also incorporated herein by reference. If the phase separation comprises filtration, a filter substrate may be used which comprises, for example, a ceramic foam made, for example, of a titania- or zirconia-based ceramic, and / or a wire mesh or wool, made, for example, of one of the same materials as the inner surface of the first reaction vessel, such as grade 316 LN or 316 FR stainless steel or a titanium alloy having the composition described below. If any residual liquid sodium remains with the separated solid phase, this residual liquid sodium may be recovered and thus the degree of phase separation may be improved by flushing the separated solid phase with an inert gas which is hot enough for the residual liquid sodium to remain in liquid phase with the sodium silicate dissolved therein. Using an inert gas which is hot like this both avoids consuming the residual liquid sodium by forming more sodium oxide and prevents loss of the dissolved sodium silicate to the solid phase. Moreover, if the separated solid phase comprises undissolved metal, using such an inert gas also has the advantage of preventing reoxidation of the hot metal. Regardless of how this phase separation is carried out, phase separation of the same process stream may be repeated, in order to increase the degree of separation finally achieved. In some embodiments, the method may comprise, after phase-separating at least some of the liquid sodium with sodium silicate dissolved therein from the undissolved metal and the other insoluble products and before reacting at least some of it with at least one of oxygen and water, recycling at least some of it at least once to a reaction vessel in which the redox reaction is conducted, and introducing more of the comminuted, dehydrated and dehydroxylated ore into the reaction vessel when the liquid sodium with sodium silicate dissolved therein is recycled. This has several beneficial effects, as follows. Firstly, by returning the liquid sodium with sodium silicate dissolved therein back to the reaction vessel for reaction with more ore, the amount of sodium silicate dissolved in the liquid sodium can be increased, until it reaches a desired concentration. Secondly, introducing more of the comminuted, dehydrated and dehydroxylated ore into the reaction vessel replenishes the ore already consumed by the reaction. Thirdly, it also raises the temperature of the reaction mixture due to the exothermic nature of this reaction, whereby the temperature of the reaction can be maintained above its minimum temperature of 320 ’Celsius. If the liquid sodium with sodium silicate dissolved therein is recycled in this manner, the method may further comprise cooling the liquid sodium with sodium silicate dissolved therein after phaseseparating at least some of it from the undissolved metal and the other insoluble products and before recycling at least some of it to the reaction vessel. This has several advantages, as follows. Firstly, it allows the temperature of the redox reaction to be controlled to remain below a temperature at which the liquid sodium can react with a metal oxide from the ore to produce a ternary oxide thereof. Secondly, it also allows the liquid sodium with sodium silicate dissolved therein to be used as a heat transfer fluid for transferring heat produced by the redox reaction to another process. For example, the liquid sodium with sodium silicate dissolved therein may be circulated through pipework of a heat exchanger surrounding and / or contained within a dryer of the ore, to help the ore reach a desired temperature for its dehydration and dehydroxylation. If the liquid sodium with sodium silicate dissolved therein is cooled as just described, in some embodiments, the method may comprise cooling the liquid sodium with sodium silicate dissolved therein to below a temperature at which the sodium silicate reaches saturation solubility, thereby creating a supersaturated solution of sodium silicate in liquid sodium, seeding the supersaturated solution to produce sodium silicate in solid phase, and phase-separating at least some of the solid sodium silicate from the liquid sodium before recycling the liquid sodium to the reaction vessel with a reduced concentration of sodium silicate dissolved therein. This has several beneficial effects, as follows. Firstly, solid sodium silicate which is obtained in this manner may subsequently be added to the mixture comprising sodium silicate and at least one of sodium oxide and sodium hydroxide which is obtained by reacting liquid sodium having sodium silicate dissolved therein with at least one of oxygen and water. This allows the amount of sodium silicate in this mixture to be increased until the sodium silicate and at least one of sodium oxide and sodium hydroxide are present in the mixture in desired proportions. Moreover, since the liquid sodium which is recycled to the reaction vessel has a reduced concentration of sodium silicate dissolved therein, more sodium silicate is able to dissolve in the recycled liquid sodium before the concentration of the sodium silicate reaches saturation solubility. If the liquid sodium with sodium silicate dissolved therein is recycled back to the reaction vessel in which the redox reaction is conducted, in some embodiments, fresh liquid sodium may also be added to the reaction vessel when the liquid sodium with sodium silicate dissolved therein is recycled. This may be done even if liquid sodium is already present in excess of the stoichiometric amount thereof required for the redox reaction. This has several beneficial effects, as follows. Firstly, it replenishes the liquid sodium already consumed by the reaction. Secondly, it dilutes the concentration of the sodium silicate dissolved in the recycled liquid sodium, which allows more sodium silicate to dissolve in the excess liquid sodium remaining afterthe reaction. Thirdly, if the fresh liquid sodium has a higher temperature than the recycled liquid sodium, for example as a result of it having just been produced by electrolysis, this also helps to bring the temperature of the reaction mixture back up above its minimum temperature of 320 ’Celsius. In some embodiments, reacting at least some of the liquid sodium having sodium silicate dissolved therein with at least one of oxygen and water may comprise spraying at least some of the liquid sodium with sodium silicate dissolved therein into an oxygenated atmosphere to combust the liquid sodium and thereby produce a mixture comprising sodium oxide and sodium silicate. Liquid sodium is less viscous and has lower surface tension than water because it does not contain any hydrogen bonds. The liquid sodium with sodium silicate dissolved therein may therefore readily be sprayed to form a fine mist of droplets, which spontaneously ignites in the oxygenated atmosphere because of its large surface area and already elevated temperature. Preferably, the oxygenated atmosphere is scrubbed of carbon dioxide to prevent the hot sodium oxide formed by combustion from reacting with the carbon dioxide. Furthermore, the humidity of the oxygenated atmosphere is preferably also initially kept at a very low level (for example, by the oxygenated atmosphere first being dried). This prevents the droplets of liquid sodium from being coated with a surface layer of sodium hydroxide, which would otherwise inhibit their complete combustion. Nonetheless, after their oxidation, the humidity of the oxygenated atmosphere may still be increased to hydrate the sodium oxide and produce less reactive sodium hydroxide. If liquid sodium with sodium silicate dissolved therein is sprayed into an oxygenated atmosphere as just described, in such embodiments, the liquid sodium with sodium silicate dissolved therein may be sprayed into the oxygenated atmosphere inside a thermally conductive combustion chamber, and heat may also be extracted from within the combustion chamber via an exterior of the combustion chamber. Some examples of a suitable construction for the thermally conductive combustion chamber are described below. This allows heat generated by the exothermic oxidation and / or hydration of the liquid sodium with sodium silicate dissolved therein to be captured, rather than being dissipated into the environment, even though the liquid sodium is being dispersed by being sprayed. The heat thus extracted may be used for another purpose, rather than being wasted. For example, in some embodiments, at least some of the extracted heat may be used to contribute to the thermal decomposition of an unreactive clay, to produce an activated clay. This has at least the following advantages. Since combustion of the liquid sodium plateaus at a temperature above about 800 ’Celsius, whereas the thermal decomposition of unreactive clays requires them to be heated to a lower temperature of from about 500 to about 800 ’Celsius, by adjusting the rate of supply of the unreactive clay to the rate of production of the alkaline mixture comprising sodium silicate, most, if not all, of the energy required to activate the clay may be supplied by the production of the alkaline mixture comprising sodium silicate. Thus, subject to real-world losses and inefficiencies, the amount of energy consumed in producing the activated clay can be greatly reduced, which makes the production of an alkaline activated cement significantly more energy efficient overall. Moreover, both an activated clay and the alkaline mixture comprising sodium silicate, which may be used as the alkaline activator for the activated clay, are then produced in the same location as each other. In some embodiments, the method of the invention may comprise using at least some of the alkaline mixture comprising sodium silicate as an alkaline activator for an alkaline activated or geopolymer cement. Regardless of the source of other components of the cement apart from the alkaline activator, this has the significant advantage that the method of the invention then makes the production of an alkaline activated or geopolymer cement more economically attractive and therefore more competitive with OPC, with all the attendant environmental benefits that entails. If at least some of the alkaline mixture comprising sodium silicate is used as an alkaline activator for an alkaline activated or geopolymer cement, in embodiments in which liquid sodium with sodium silicate dissolved therein is sprayed into an oxygenated atmosphere inside a thermally conductive combustion chamber, using at least some of the alkaline mixture comprising sodium silicate as an alkaline activator may comprise adding the alkaline mixture to an activated clay which is produced by thermal decomposition of an unreactive clay using heat extracted from within the combustion chamber. Thus the method of the invention may produce an alkaline activated or geopolymer cement comprising both an activated clay and an alkaline activator, wherein the alkaline activator comprises sodium silicate and at least one of sodium oxide and sodium hydroxide, starting from an unreactive clay and a metal ore comprising a siliceous mineral as initial ingredients, whilst extracting a target metal from the ore to produce the target metal in elemental form as a co-product. Consequently, the total energy consumed is shared between producing the cement and extracting the target metal from its ore. In some embodiments, the method may comprise reacting at least some of the alkaline mixture comprising sodium silicate with gaseous carbon dioxide to produce a composition comprising sodium carbonate and silica. This reaction, which proceeds according to the equation: n(NajO) • SiOz (S) + CO2 (gj SiO2 (s) + n Na2COs (Sj [Eqn. 11] is the reverse of the reaction of Eqn. 2. Whereas the reaction of Eqn. 2 can only occur at high temperatures, the reaction of Eqn. 11 has a Gibbs free energy, AG <0 at temperatures below about 300 ’Celsius and can therefore occur at ambient temperatures. For example, the gaseous carbon dioxide may be captured from atmospheric air and / or from one or more industrial processes, in order to mitigate greenhouse gas emissions. In some embodiments, wherein dehydrating the ore and dehydroxylating hydroxylated compounds contained therein also releases carbon dioxide from the ore, the carbon dioxide released from the ore may be captured for mineralization by using the alkaline mixture comprising sodium silicate in this manner. In some embodiments, the composition comprising sodium carbonate and silica produced by the reaction of Eqn. 11 may be used as an ingredient in the manufacture of soda-lime glass, for which its composition is ideally suited. Traditionally, the manufacture of soda-lime glass has mostly used naturally occurring trona mined from geological deposits as a source of sodium carbonate. In comparison to using trona in glassmaking, using the composition produced by the reaction of Eqn. 11 instead therefore has the advantage that this composition contains carbon dioxide which has been captured rather than carbon dioxide which has been newly extracted from geological deposits. Even if the captured carbon dioxide was released from the ore during its drying and dehydroxylation, the ore has already been mined to produce the target metal in elemental form. Thus in such a case, the carbon dioxide successively passes through two industrial processes ( / .e., both production of the target metal and glassmaking), whereby the total amount of carbon dioxide extracted from geological deposits to produce both the target metal in elemental form and the glass is halved. The method of the invention may be carried out as a continuous, semi-batch or batch process. However, it is preferably carried out as a continuous process for reasons of economy and efficiency. Apparatus In a second aspect, the present invention also provides an apparatus for producing an alkaline mixture comprising sodium silicate from an ore of a target metal, wherein the ore comprises a siliceous mineral and the target metal is at least one of iron and manganese. The siliceous mineral may comprise the target metal itself and / or it may be silica and / or another silicate mineral present as gangue. The apparatus comprises a comminution device, a dryer, a gas-tight first reaction vessel, a first solid-liquid sodium phase separator and a second reaction vessel. The comminution device is for comminuting the ore into fines. The dryer is for dehydrating the ore and dehydroxylating hydroxylated compounds contained therein. The gas-tight first reaction vessel is for reacting the ore thus treated in an inert atmosphere with an amount of liquid sodium in excess of the stoichiometric amount thereof required for a redox reaction between the liquid sodium and an oxide of the target metal from the ore, to precipitate out from the liquid sodium a solid phase comprising both the target metal in elemental form and other insoluble products comprising sodium oxide. The redox reaction is conducted at a temperature of at least 320 ’Celsius to induce a reaction between the sodium oxide and silica derived from the siliceous mineral to produce sodium silicate, and below a temperature at which the liquid sodium can react with the oxide of the target metal to produce a ternary oxide thereof. The first reaction vessel comprises a first inlet for receiving the comminuted, dehydrated and dehydroxylated ore, a second inlet for the liquid sodium, and an outlet for the liquid sodium with the solid phase entrained therein. The first solid-liquid sodium phase separator is for separating the liquid sodium with sodium silicate dissolved therein from the solid phase, and comprises an inlet for receiving the liquid sodium with the solid phase entrained therein from the outlet of the first reaction vessel, a first outlet for liquid sodium with sodium silicate dissolved therein, and a second outlet for the solid phase. The second reaction vessel is for reacting the liquid sodium having sodium silicate dissolved therein with at least one of oxygen and water to produce a mixture comprising sodium silicate and at least one of sodium oxide and sodium hydroxide. It comprises a first inlet for the liquid sodium with sodium silicate dissolved therein, a second inlet for at least one of oxygen and water, and an outlet for the mixture thus produced. The comminution device may comprise one or more known types of comminution device, such as a rock crusher, grinder, hammer mill, ball, disc and / or rod mill, autogenous mill, roller mill, vibrator, and / or the like. The dryer may, for example, comprise a continuous tunnel dryer through which the comminuted ore is transferred, for example on a moving conveyor. Such a dryer may be compared and contrasted with a Dwight-Lloyd sintering machine, as follows. Like a Dwight-Lloyd sintering machine, the ore fines can be advanced through a hot zone of the dryer in order to dehydrate and dehydroxylate them, and the rate of advance may be similar to that in a Dwight-Lloyd sintering machine. However, unlike a Dwight-Lloyd sintering machine, the ore is not mixed with any other ingredients, other than possibly being a mixture of different ores (such as a manganese ore and an iron ore), and is not fired or exposed to a flame front, and is only heated instead. The dryer may have an interior which is closed off from the surrounding environment, for example by means of one or more doors or curtains, so that the ore fines can be enclosed within the dryer as they are heated, are not exposed to the surrounding environment and may instead be contained in their own atmosphere. Enclosing the ore fines as they are heated so that they are contained in their own atmosphere allows this atmosphere to be controlled, and also helps to prevent gases like carbon dioxide and sulphur dioxide which may be released from the ore if it is heated from escaping into the surrounding environment. Enclosing the ore in this way also helps to maintain the thermal efficiency of its dehydration and dehydroxylation by reducing the loss of heat to the environment. In some embodiments, the dryer may contain an atmosphere to which the ore is exposed during dehydration and dehydroxylation, and the dryer may comprise means for reducing a pressure of this atmosphere to less than that of atmospheric air outside the dryer. Thus when ore enters or leaves the dryer, this pressure differential causes atmospheric air from outside the dryer to enter the dryer along with the ore, thereby hindering or preventing the escape of gases like carbon dioxide and sulphur dioxide into the surrounding environment. The first reaction vessel may be made, for example, of steel of a type already used for containing and transporting liquid sodium, such as grades 316 LN or 316 FR stainless steel. In such a case, the inner surface of the reaction vessel may be provided with a sacrificial layer having the same chemical composition, for gradual corrosion by the highly reactive sodium oxide. Thus depending on the desired composition of other, minor elements mixed in with the target metal eventually produced, the chemical constituents of the sacrificial layer will either be the same as those already present in the products of the reaction between the ore and the liquid sodium ( / .e., Fe and / or Mn) or may be entirely acceptable as minor constituents of the reaction products ( / .e., Cr, Ni, Mo). Alternatively, the reaction may be carried out in a reaction vessel made from or lined with titanium or a titanium alloy, such as one having a composition by weight of 98.8% Ti, 0.8% Ni and 0.4% Mo. Titanium and titanium alloy are found to be highly resistant to corrosion by sodium oxide across the entire range of operating temperatures for the reaction between the ore and the liquid sodium, by forming a stable surface passivation layer of titanium dioxide. The first reaction vessel may, for example, be a stirred tank reactor. In some embodiments, the first inlet of the first reaction vessel may comprise an airlock to prevent atmospheric oxygen from being introduced into the first reaction vessel whenever the vessel is fed with ore. If so, the airlock may have an interior for holding the ore and may also comprise first and second gas-tight doors, an inlet to the interior of the airlock for inert gas, and an outlet from the interior of the airlock for atmospheric air. The first gas-tight door connects the interior of the airlock with a surrounding environment of the first reaction vessel, and the second gas-tight door connects the interior of the airlock with an inside of the first reaction vessel. The inlet for inert gas and the outlet for atmospheric air each comprises a respective valve for opening and closing a respective one of the inlet and the outlet. Thus a charge of ore may be introduced into the interior of the airlock via the first door, and atmospheric air may be purged from within the interior of the airlock by opening and closing the valves to replace the air inside the airlock with inert gas, before the ore is released into the first reaction vessel via the second door. If the first inlet of the first reaction vessel does comprise such an airlock, the apparatus may further comprise a heat exchanger for transferring heat from the first reaction vessel to the inert gas upstream of the inlet to the airlock. Thus the inert gas can be heated to a temperature similar to that inside the first reaction vessel before it enters the first reaction vessel, which helps to prevent a build-up of pressure within the first reaction vessel, and to purge atmospheric air from within the airlock. The first solid-liquid sodium phase separator may be of a type shown and described in co-pending UK patent application no. 2417052.4 ("Apparatus and Method for Separating a Contaminant from Liquid Metal"; applicant's ref: NE-P-GB 002), mentioned above. Like the first reaction vessel, it is preferably made of a material which is adapted to accommodate or withstand corrosion by sodium oxide, as also described above. The second reaction vessel is preferably also made of a material which is adapted to accommodate or withstand corrosion by sodium oxide, either as already described above or as further described below. Unlike the first reaction vessel, it does not need to be gas-tight. In some embodiments, the apparatus may further comprise a second solid-liquid sodium phase separator comprising an inlet downstream of the first outlet of the first solid-liquid sodium phase separator, a first outlet for solid sodium silicate and a second outlet for liquid sodium with sodium silicate dissolved therein, wherein the second solid-liquid sodium phase separator is maintained below a temperature at which the sodium silicate in the liquid sodium received from the first outlet of the first solid-liquid sodium phase separator reaches saturation solubility, and the second outlet of the second solid-liquid sodium phase separator is upstream of the second inlet of the first reaction vessel. The second solid-liquid sodium phase separator therefore effectively behaves like a cold trap for trapping and removing dissolved sodium silicate from the liquid sodium received from the first outlet of the first solid-liquid sodium phase separator, before liquid sodium with a reduced concentration of sodium silicate dissolved therein is returned to the first reaction vessel. This allows liquid sodium to be circulated through the first reaction vessel and for solid sodium silicate to be continuously extracted from the reaction mixture. However, unlike a conventional cold trap, the second solid-liquid sodium phase separator has the advantage that it is also able to provide a constant supply of solid sodium silicate. Like the first solid-liquid sodium phase separator, the second solid-liquid sodium phase separator may be of a type shown and described in co-pending UK patent application no. 2417052.4 ("Apparatus and Method for Separating a Contaminant from Liquid Metal"; applicant's ref: NE-P-GB 002) mentioned previously, and is preferably also made of a material which is adapted to accommodate or withstand corrosion by sodium oxide. If the apparatus does comprise a second solid-liquid sodium phase separator as just described, the first outlet of the second solid-liquid sodium phase separator may be arranged to add the solid sodium silicate it produces to the mixture comprising sodium silicate and at least one of sodium oxide and sodium hydroxide produced in the second reaction vessel. This has the following advantage. Since the concentration of sodium silicate dissolved in the liquid sodium which enters the second reaction vessel is limited by its saturation solubility at the temperature of the liquid sodium, the mixture produced in the second reaction vessel comprises a majority of at least one of sodium oxide and sodium hydroxide, but only a minority of sodium silicate. However, by adding solid sodium silicate from the first outlet of the second solid-liquid sodium phase separator thereto, an alkaline mixture comprising sodium silicate and at least one of sodium oxide and sodium hydroxide in desired proportions may thus be obtained. In some embodiments, the second reaction vessel may comprise a thermally conductive combustion chamber, the first inlet of the second reaction vessel may comprise a nozzle for spraying liquid sodium with sodium silicate dissolved therein into the combustion chamber, and the second inlet of the second reaction vessel may also be adapted to admit at least one of gaseous oxygen and water vapour into the combustion chamber. For example, the combustion chamber may be made of a high thermal conductivity material such as copper or a copper alloy and have a corrosion-resistant lining to protect the copper from the alkaline mixture produced by the reaction of the liquid sodium with the oxygen and / or water vapour, as well as from the high temperature of this strongly exothermic reaction. For example, the lining may be made of a nickel-chromium alloy, such as Inconel® 625 or Hastelloy® X, both of which are highly resistant to corrosion in caustic environments, but which also have coefficients of thermal expansion which are either very similar to or the same as those of copper and several copper alloys at or around the combustion temperature. Moreover, since such nickelchromium alloys also have relatively good thermal conductivity, this allows heat to be rapidly transferred from within the combustion chamber to an exterior thereof, which simultaneously prevents the interior of the combustion chamber from overheating and also allows the extracted heat to be used for another purpose. If the second reaction vessel does comprise a thermally conductive combustion chamber as just described, then the apparatus may further comprise a clay transport pathway, itself comprising an inlet for an unreactive clay and an outlet for an activated clay, wherein the clay transport pathway is adapted to bring the unreactive clay into thermal contact with an exterior of the combustion chamber. This has the advantage that it allows heat from the exothermic reaction inside the combustion chamber to be used to contribute towards the thermal decomposition of an unreactive clay, to produce an activated clay. If the apparatus does comprise such a clay transport pathway, the outlet of the clay transport pathway may be arranged to add the activated clay to the mixture comprising sodium silicate and at least one of sodium oxide and sodium hydroxide which comes from the outlet of the second reaction vessel. If so, the apparatus may therefore be used to produce an alkaline activated or geopolymer cement comprising both an activated clay and an alkaline activator, wherein the alkaline activator comprises sodium silicate and at least one of sodium oxide and sodium hydroxide, starting from an unreactive clay and an ore of a target metal as initial ingredients, wherein the target metal is at least one of iron and manganese and the ore comprises a siliceous mineral, whilst also extracting the target metal from its ore in elemental form. Brief Description of the Drawings Further features and advantages of the present invention will become apparent from the following detailed description, which is given by way of example and in association with the accompanying drawings, in which: Fig. 1 is a flow diagram of a first embodiment of a method of producing an alkaline mixture comprising sodium silicate; Fig. 2 is a flow diagram of part of a second embodiment of such a method; Fig. 3 is a flow diagram of a third embodiment of a method of producing an alkaline mixture comprising sodium silicate; Fig. 4 is a flow diagram of a fourth embodiment of such a method; Fig. 5 is a flow diagram of a fifth embodiment of such a method; Fig. 6 is a flow diagram of a sixth embodiment of such a method; Fig. 7 is a flow diagram of part of a seventh embodiment of such a method; Fig. 8 is a flow diagram of an eighth embodiment of such a method; Fig. 9 is a flow diagram of a cycle in a ninth embodiment of a method of producing an alkaline mixture comprising sodium silicate; Fig. 10 is a graph schematically representing the temperatures of successive stages in the cycle of Fig. 9; Fig. 11 is a flow diagram of part of a tenth embodiment of a method of producing an alkaline mixture comprising sodium silicate; Fig. 12 is a flow diagram of part of an eleventh embodiment of such a method; Fig. 13 is a schematic diagram of a first embodiment of an apparatus for producing an alkaline mixture comprising sodium silicate; Fig. 14 is a schematic longitudinal section through an embodiment of a dryer; Fig. 15 is a schematic diagram of an embodiment of an airlock; Fig. 16 is a schematic longitudinal section through an embodiment of a reaction vessel for reacting liquid sodium having sodium silicate dissolved therein with at least one of oxygen and water; Fig. 17 is a schematic diagram of a second embodiment of an apparatus for producing an alkaline mixture comprising sodium silicate; and Fig. 18 is a schematic diagram of part of a third embodiment of an apparatus for producing an alkaline mixture comprising sodium silicate. In the accompanying drawings and the associated description, the subscript "(Na)" attached to a chemical species as a state symbol denotes that the species is dissolved in liquid sodium. Detailed Description Fig. 1 shows a first embodiment of a method 600a of producing an alkaline mixture comprising sodium silicate from an ore of a target metal, wherein the ore comprises a siliceous mineral and the target metal is at least one of iron and manganese. The method 600a comprises comminuting 301 the ore into fines, and drying 302 the ore to dehydrate it and so that hydroxylated compounds contained in the ore are dehydroxylated. Comminution 301 of the ore and its dehydration and dehydroxylation 302 interact with each other as described above. The comminuted, dehydrated and dehydroxylated ore particles are then reacted 601 in an inert atmosphere with an amount of liquid sodium in excess of the stoichiometric amount thereof required for a redox reaction between the liquid sodium and an oxide of the target metal from the ore. This reaction precipitates out from the liquid sodium both the target metal in elemental form and other insoluble products comprising sodium oxide. The reaction 601 is conducted at a temperature of at least 320 ’Celsius to induce a reaction between the sodium oxide thus formed and silica derived from the siliceous mineral to produce sodium silicate. However, the reaction 601 is also conducted below a temperature at which the liquid sodium can react with the oxide of the target metal to produce a ternary oxide thereof, such as Na4FeO3 if the target metal is iron or a-NaMnO? if the target metal is manganese. At least some of the liquid sodium with sodium silicate dissolved therein is then phase-separated 305 from the undissolved metal and the other insoluble products. Finally, at least some of the phase-separated liquid sodium having sodium silicate dissolved therein is reacted 602 with at least one of oxygen and water to produce the alkaline mixture comprising sodium silicate and at least one of sodium oxide and sodium hydroxide. Fig. 2 shows part of a second embodiment of a method 600b of producing an alkaline mixture comprising sodium silicate as just described in relation to Fig. 1. Like the method 600a, the method 600b again comprises comminuting 301 the ore into fines, but in this case, dehydration and dehydroxylation 302 of the ore comprises heating 302c the ore to a temperature of more than 100 ’Celsius. Being above the boiling point of water at atmospheric pressure, this helps to dehydrate and dehydroxylate the ore. On the other hand, the ore is also heated 302c to less than 600 ’Celsius, which is comfortably below the Tammann temperature of silica. Thus if the ore already comprises silica as the siliceous mineral and / or silica is formed by thermal decomposition of the siliceous mineral during dehydration and dehydroxylation 302 of the ore, undesirable sintering of the silica can be avoided. Moreover, 600 ’Celsius is also below the respective thermal decomposition temperatures of such siliceous minerals of the target metal as braunite, rhodonite, tephroite and fayalite. This therefore enables these minerals, all of which are paramagnetic, to remain in the ore if diamagnetic gangue mineral species are subsequently magnetically separated from the ore before the ore is added to liquid sodium. Thereafter, the method 600b proceeds in the same manner as the method 600a described above in relation to Fig. 1, in other words by reacting 601 the comminuted, dehydrated and dehydroxylated ore particles with liquid sodium, performing the phase separation 305 and reacting 602 the phase-separated liquid sodium having sodium silicate dissolved therein with at least one of oxygen and water. Fig. 3 shows a third embodiment of a method 600c of producing an alkaline mixture comprising sodium silicate from an ore of a target metal, wherein the ore comprises a siliceous mineral, the target metal is iron, and in this case, for the sake of example, the iron is present in the ore as at least one of the mineral and mineraloid forms of iron oxide and iron oxyhydroxide. Like the method 600b of Fig. 2, dehydrating and dehydroxylating the iron ore in the method 600c also comprises heating 302c the ore to a temperature of from 100 ’Celsius to 600 ’Celsius, inclusive. Not only is this comfortably below the Tammann temperature of silica, it is also below the Tammann temperature of FejOa, which therefore avoids undesirably sintering iron (III) oxide either already present in the ore or formed during its dehydration and dehydroxylation. Moreover, in the method 600c, when the comminuted, dehydrated and dehydroxylated ore particles are then reacted with liquid sodium, the iron ore thus treated is reacted 303 with an excess amount of liquid sodium in an inert atmosphere and at a temperature of less than 450 ’Celsius, which is below the temperature at which the liquid sodium can react with the iron oxide to produce Na4FeO3 as a ternary oxide thereof. As in the method 600a of Fig. 1, the method 600c then comprises phase-separating 305 at least some of the liquid sodium with sodium silicate dissolved therein from the iron and the other insoluble products formed in the reaction 303. In addition, however, after this phase separation 305, iron is separated 304a from the other insoluble products to yield the iron as an end-product. The iron may be magnetically separated from the other insoluble products because iron is strongly (and famously) ferromagnetic at temperatures below its Curie temperature, whereas the other insoluble products are diamagnetic. However, iron is also denser than all the other main species present in the reaction 303, and is nearly 3.5 times denser than sodium oxide. Thus particles of iron produced by the reaction 303 are less buoyant in a fluid like liquid sodium than equally sized particles of any of the other insoluble products. Therefore, as well as or instead of magnetic separation, separating 304a the iron from the other insoluble products may comprise separating them based on their different densities. In the method 600c, the temperature of the reaction 303 is controlled 313 as follows. At least some of the liquid sodium from which the iron and the other insoluble products have now been separated 305, but which still has sodium silicate dissolved therein, is cooled 309c, and then returned 314 to the reaction 303 in a loop or circuit. Before it is returned 314 to the reaction 303, the liquid sodium is cooled 309c by being circulated through pipework of a heat exchanger surrounding and / or contained within a dryer of the ore. Heat from the liquid sodium is used to heat 308 the ore in the dryer, which helps the ore to reach the desired temperature for its dehydration and dehydroxylation. Adjusting 315 a flow rate of the liquid sodium in this cooling circuit also allows the temperature of the reaction 303 to be controlled 313 to remain below 450 ’Celsius. The flow rate of the liquid sodium may be adjusted 315 by such things as an initial choice of how much liquid sodium in excess of the stoichiometric amount thereof required for the reaction 303 is used and / or an initial choice of the diameter of the pipework. Alternatively or additionally, a pumping rate for the liquid sodium around the cooling circuit may be varied. When the liquid sodium with sodium silicate dissolved therein is returned 314 to a reaction vessel in which the reaction 303 is conducted, more of the comminuted, dehydrated and dehydroxylated ore is introduced into the reaction vessel as well. This both replenishes the ore consumed by the reaction 303 and also raises the temperature of the reaction mixture again due to the exothermic nature of the reaction 303. Thus the temperature of the reaction 303 can be maintained within its desired temperature range of between 320 and 450 ’Celsius. Moreover, as a result of returning 314 the liquid sodium with sodium silicate dissolved therein back to the reaction 303, the amount of sodium silicate dissolved in the liquid sodium can also be increased, until it reaches a desired concentration. Thereafter, the recirculation 314 stops and the method 600c proceeds in a similar manner to the method 600a of Fig. 1, by performing a final phase separation 305 of the iron and the other insoluble products from the liquid sodium having sodium silicate dissolved therein, and then reacting 602 the phase-separated liquid sodium having the sodium silicate dissolved therein with at least one of oxygen and water. Fig. 4 shows a fourth embodiment of a method 600d of producing an alkaline mixture comprising sodium silicate from an ore comprising a siliceous mineral, where in this case, for the sake of example, the ore is red bauxite. Red bauxite is an aluminium ore also comprising a significant proportion of iron, which may therefore be considered as a type of iron ore within the context of the present invention. Since red bauxite is typically found in near-surface deposits which are extracted by opencast mining, it may initially contain a significant admixture of organic matter, such as tree roots and humus. In the present embodiment, therefore, dehydrating and dehydroxylating 302b the ore are conducted in an oxygenated atmosphere, such as in atmospheric air, and the ore is heated to at least 250 ’Celsius, which is above the auto-ignition temperature of any wood present. Any organic matter mixed in with the ore is therefore combusted 307 to produce carbon dioxide gas and water vapour, leaving ash, which is incorporated into the other inorganic gangue species already present in the ore. Meanwhile, comminution 301 of the ore and hence of any associated organic matter ensures that large pieces of organic matter are broken up, encouraging their complete combustion and preventing the formation of localised hotspots. The dehydration and dehydroxylation 302b are also conducted by enclosing the comminuted ore as it is heated, so that it is contained in its own atmosphere and is not exposed to its surrounding environment, and the carbon dioxide released by the combustion 307 is captured 306. The method 600d then comprises reacting 303 the comminuted, dehydrated and dehydroxylated ore with liquid sodium in the same temperature range as described above in relation to Fig. 3, for the same reasons. The method 600d then proceeds as in the method 600c by phase-separating 305 at least some of the liquid sodium with sodium silicate dissolved therein from the iron and the other insoluble products formed in the reaction 303, reacting 602 the phase-separated liquid sodium having the sodium silicate dissolved therein with at least one of oxygen and water, and separating 304a the iron from the other insoluble products to yield the iron as an end-product. In the present embodiment, however, after the other insoluble products have been separated 304a from the iron, these other insoluble products, a major portion of which is sodium oxide and which therefore has a high affinity for carbon dioxide, is passed 311 through the atmosphere in which the ore is contained during its dehydration and dehydroxylation 302b. The sodium oxide in the other insoluble products, which are still hot from the reaction 303 with the liquid sodium, now reacts 312 with the carbon dioxide gas that was captured 306 to produce at least sodium carbonate. Thus no carbon dioxide is released into the environment and a useful co-product is produced instead. As noted above, it is typically the case that in red bauxite, nfFejOa) = n(SiO?) present as gangue, where n denotes the number of mol of each component, which leaves approximately 1 mol of unreacted NajO remaining in the other insoluble products per mol of FejOa consumed, to react 312 with the carbon dioxide gas. This is more than sufficient sodium oxide left in the other insoluble products to react 312 with all the carbon dioxide, since typically, n(CO?) « nfFejOa), where n(CO?) denotes the number of mol of carbon dioxide released by the combustion 307. Nonetheless, in another possible embodiment, which is a variant of that shown in Fig. 4, if after analysing a sample taken from a particular batch of red bauxite, the molar ratio of FejOa: SiO? therein is found to be unusually low, the method 600d may also comprise increasing this molar ratio by adding an iron oxide and / or iron ore from another, non-bauxitic source, to the batch of red bauxite, before the comminuted, dehydrated and dehydroxylated ore is reacted 303 with the liquid sodium. Fig. 5 shows a fifth embodiment of a method 600e of producing an alkaline mixture comprising sodium silicate from an ore comprising a siliceous mineral, where in this case, for the sake of example, the ore comprises a manganiferous mineral. The method 600e comprises comminuting 501 the manganese ore into fines and heat-treating 502 the ore in atmospheric air to a temperature of from 100 to 600 ’Celsius, inclusive. Comminuting 501 and heat-treating 502 the ore interact with each other as described previously. This not only dehydrates the ore and dehydroxylates any hydroxylated compounds contained therein, it also reduces any tetravalent manganese oxide, such as pyrolusite, in the ore into a trivalent manganese oxide. However, the heat-treatment 502 is performed without using a reductant, such as carbon or hydrogen. The trivalent manganese oxide may comprise at least one of manganese (III) oxide (MnjOa) and manganese (11,111) oxide (Mn3O4). The heat-treatment 502 also thermally decomposes any rhodochrosite and / or manganite present in the ore into a trivalent manganese oxide, respectively releasing either carbon dioxide or water vapour. On the other hand, the temperature of the heat-treatment 502 is not sufficient to thermally decompose siliceous manganese minerals, such as braunite and rhodonite, into hausmannite and silica. If the manganese ore comprises romanechite, heat-treating the ore in this manner causes water molecules trapped within the romanechite to diffuse out and be driven off as water vapour. The heat-treatment also tends to convert the structure of the romanechite into hollandite as the water escapes. However, the temperature of the heat-treatment 502 is not sufficient to cause the thermal decomposition of hollandite itself or of other manganiferous minerals such as cryptomelane which are isostructural to hollandite. The method 600e then comprises cooling 504a the heat-treated ore back down to less than 100 ’Celsius and magnetically separating 504b paramagnetic components of the ore at least comprising the trivalent manganese oxide from diamagnetic components of the ore. At least some of the heat released by cooling 504a the ore back down may be recovered and used to help in heat-treating 502 the ore. The magnetic separation 504b will tend to separate any diamagnetic silica and / or silicate originally present in the ore as the siliceous mineral and / or silica formed by thermal decomposition of a siliceous mineral during the dehydration and dehydroxylation 502 of the ore from the paramagnetic components of the ore. However, since siliceous manganese minerals, such as braunite and rhodonite, are paramagnetic, as are hollandite and other manganiferous minerals such as cryptomelane which are isostructural to hollandite, all these manganiferous minerals remain in the paramagnetic components of the ore, along with the trivalent manganese oxide. The paramagnetic components of the ore are then added to an amount of liquid sodium in excess of the stoichiometric amount thereof required for a redox reaction 505b between the liquid sodium and the trivalent manganese oxide. This reaction precipitates out from the liquid sodium both elemental manganese and other insoluble products at least comprising sodium oxide. As before, the reaction 505b is conducted at a temperature of at least 320 ’Celsius to induce a reaction between the sodium oxide thus formed and silica derived from the siliceous mineral to produce sodium silicate. However, the reaction 505b is also conducted in an inert atmosphere and at a temperature of less than 600 ’Celsius, which is below the temperature at which the liquid sodium can react with a manganese oxide to produce a ternary oxide thereof, such as a-NaMnOj. After the redox reaction 505b, the liquid sodium with the sodium silicate dissolved therein is phase-separated 506 from the undissolved manganese and the other insoluble products. If, for example, the paramagnetic components of the ore comprised cryptomelane, the liquid sodium may also comprise some dissolved potassium. This is converted into a small admixture of potassium hydroxide during the subsequent reaction 602 of the liquid sodium with at least one of oxygen and water. If, however, the paramagnetic components of the ore comprised romanechite and / or hollandite, for example, the solid phase from the phase separation 506 may comprise some baria. Similarly, if the paramagnetic components of the ore comprised kutnohorite, the solid phase may comprise some calcia. In any case, the solid phase is then added to liquid water to hydrate 507 the sodium oxide therein and produce an aqueous solution of sodium hydroxide, in which the elemental manganese remains undissolved. The manganese is then phase-separated 508 from the aqueous solution of sodium hydroxide. Any small admixture of baria and / or calcia in the solid phase after the phase separation 506 is converted into barium hydroxide and / or calcium hydroxide during hydration 507 of the sodium oxide, which dissolve into the aqueous phase and are phase-separated 508 along with the sodium hydroxide to leave just the elemental manganese behind. The aqueous solution of sodium hydroxide may subsequently be recycled to produce more liquid sodium by electrolysis. Fig. 6 shows a sixth embodiment of a method 600f of producing an alkaline mixture comprising sodium silicate from an ore comprising a siliceous mineral, where in this case, for the sake of example, the ore comprises rhodochrosite as a manganiferous mineral. This sixth embodiment is a special case of the fifth embodiment descried above in relation to Fig. 5. The method 600f therefore differs from the method 600e in that the manganiferous ore is heat-treated 502a in atmospheric air to at least 200 ’Celsius to ensure the thermal decomposition of the rhodochrosite and that the heat-treatment 502a is conducted with the ore closed off from its surrounding environment to allow carbon dioxide released by the rhodochrosite to be captured 510. If desired, the heat-treatment 502a may also be conducted at less than 320 ’Celsius to discourage the carbon dioxide from being reabsorbed by the trivalent manganese oxide thus formed. The method 600f then proceeds as described above in relation to the method 600e of Fig. 5 until after phase-separating 508 the aqueous solution of sodium hydroxide from the elemental manganese in solid phase, when at least some of this aqueous solution of sodium hydroxide is then reacted 511 with the captured carbon dioxide to produce sodium carbonate as an additional co-product. Fig. 7 shows a seventh embodiment of a method 600g of producing an alkaline mixture comprising sodium silicate from an ore comprising a siliceous mineral, where the ore again comprises a manganiferous mineral. In this embodiment, however, before the ore is reacted 505b with the excess amount of liquid sodium, the manganiferous mineral in the ore is converted into a trivalent manganese oxide using a hydrometallurgical technique 503. This hydrometallurgical technique 503 firstly comprises adding 503a at least some of the ore to hot, concentrated hydrochloric acid closed off from their surrounding environment. The ore dissolves in the acid to produce gaseous chlorine and an acidic aqueous solution comprising manganese (II) chloride, leaving behind a solid silica residue derived from the siliceous mineral. The solid silica residue is phase-separated 503b from the acidic aqueous solution. Sodium oxide, or sodium hydroxide derived from hydrating at least some of this sodium oxide, is then added 503c to the acidic aqueous solution to produce an alkaline aqueous solution and a precipitate comprising manganese (II) hydroxide. At least some of the precipitate is then phase-separated 503d from the alkaline aqueous solution, and the manganese (II) hydroxide is dried and dehydroxylated 503e in atmospheric air at less than 100 ’Celsius to produce the trivalent manganese oxide. Dehydration and dehydroxylation 503d of the manganese (II) hydroxide in atmospheric air does not require the application of any heat to produce y-phase MnjOa as the trivalent manganese oxide. The solid silica residue is finely divided, so in principle, it could be reacted directly with sodium hydroxide to produce an alkaline mixture comprising sodium silicate. However, this reaction proceeds only slowly at ambient temperatures and the solid silica residue is also likely to be contaminated with chloride ions, which would make the resulting alkaline mixture unsuitable for use as an alkaline activator for alkaline activated and geopolymer cements. Therefore, the solid silica residue is instead treated with just enough of an aqueous solution of sodium hydroxide to neutralize 522 any of the acidic aqueous solution remaining thereon. Only a relatively small molar amount of aqueous sodium hydroxide is required to neutralize 522 the relatively small molar amount of remaining acidic aqueous solution, and conventional means may be used to determine when a neutral pH is achieved. The neutral pH ensures that the silica residue does not react any further with the aqueous solution of sodium hydroxide to produce an alkaline mixture comprising sodium silicate. The neutralized silica residue is then washed 523 in deionized water to remove the aqueous solution at least comprising sodium chloride which results from the neutralization 522, and dried. Once again, the washed and dried 523 silica residue, which has now been decontaminated, could in principle be reacted directly with sodium hydroxide to produce an alkaline mixture comprising sodium silicate. However, since this reaction proceeds only slowly at ambient temperatures, the washed and dried 523 silica residue is instead added to the reaction 505b of the trivalent manganese oxide with the liquid sodium. Since the silica residue and the trivalent manganese oxide are added to the liquid sodium as separate process streams, the relative rates of addition of each can be varied to control the relative molar amounts of each which are present in the reaction mixture. For example, the rate of addition of silica can be matched to the rate of production of sodium oxide. After the reaction 505b, the method 600g then proceeds as described above in relation to the method 600e of Fig. 5, namely by performing the phase separation 506 and reacting 602 the phase-separated liquid sodium having sodium silicate dissolved therein with at least one of oxygen and water to produce the alkaline mixture comprising sodium silicate. In addition, the method 600g also comprises capturing 525 chlorine gas evolved from the hot, concentrated hydrochloric acid, and using 526 at least some of this captured chlorine gas to make hydrochloric acid. The captured chlorine may be used to make hydrochloric acid in at least one of several different ways, some of which are described in the applicant's co-pending UK patent application no. 2417063.1 ("Carbon-Free Method and Apparatus for Producing Manganese"; applicant's ref: NE-P-GB 008) mentioned above. Regardless of how the hydrochloric acid is made 526, the method 600g then comprises recycling the hydrochloric acid thus produced for dissolution 503a of the ore, as shown in Fig. 7. Fig. 8 shows an eighth embodiment of a method 600h of producing an alkaline mixture comprising sodium silicate from an ore comprising a siliceous mineral, wherein both iron and manganese are target metals. The method 600h is therefore suitable for producing the alkaline mixture comprising sodium silicate as a co-product of the production of ferromanganese. The method 600h consumes both an iron ore, in which the iron is present as at least one of the mineral and mineraloid forms of iron oxide and iron oxyhydroxide, and a manganese ore comprising a manganiferous mineral. Processing of the iron ore initially proceeds as described above in relation to Fig. 3, which is represented on the left-hand side of Fig. 8. In other words, the iron ore is comminuted 301 into fines, dried and dehydroxylated 302c and then added 303 to an excess amount of liquid sodium to react therewith in a redox reaction, the temperature of which is controlled to remain within a range of temperatures of from 320 to 450 ’Celsius. Comminution 301 of the iron ore and its dehydration and dehydroxylation 302c interact with each other as described previously. Meanwhile, processing of the manganese ore initially proceeds as described above in relation to Fig. 5, as represented on the right-hand side of Fig. 8. In other words, the manganese ore is also comminuted 501 into fines and heat-treated 502 in atmospheric air to a temperature of from 100 to 600 ’Celsius, inclusive. Comminution 501 and heat-treatment 502 of the manganese ore interact with each other as described previously. Thereafter, the heat-treated ore is cooled 504a back down to less than 100 ’Celsius and paramagnetic components of the ore are magnetically separated 504b from diamagnetic components of the ore, before the paramagnetic components of the ore are then added 505b to an excess amount of liquid sodium to react therewith. In this case, the temperature of the reaction with the liquid sodium is controlled to remain within a range of temperatures of from 320 to 600 ’Celsius. The method 600h then proceeds as follows. After the reaction 303 of the iron ore with the liquid sodium, the iron is separated 304b from the other insoluble products using a "wet" separation technique, in which both are still entrained in the liquid sodium. This may be done, for example, by applying a magnetic field gradient to the liquid sodium with the iron and the other insoluble products entrained therein to separate the ferromagnetic iron from the diamagnetic other insoluble products and / or by using a density-based separation technique. In the latter case, a stream of the liquid sodium with the iron and the other insoluble products suspended or entrained therein may, for example, be supplied to an inlet of a device functioning like a hydrocyclone (which may therefore be called a "natrocyclone"), in which the denser iron particles are directed to an underflow outlet thereof and particles of the less dense other insoluble products are instead directed to an overflow outlet thereof. In any event, the separation 304b produces two process streams, one of which is liquid sodium with sodium silicate dissolved therein and an entrained solid phase comprising sodium oxide as well as any diamagnetic gangue species, and the other of which is liquid sodium with sodium silicate dissolved therein and elemental iron entrained therein in solid phase. This second process stream is mixed 612a with the products of the reaction 505b between the paramagnetic components of the manganese ore and liquid sodium, which therefore at least comprise liquid sodium with elemental manganese and sodium oxide entrained therein in solid phase, but which may also comprise sodium silicate dissolved therein if the manganiferous mineral initially present in the manganese ore was a siliceous manganese mineral, such as braunite and / or rhodonite. Next, the first process stream from the separation 304b is phase-separated 305b into a solid phase comprising sodium oxide and diamagnetic gangue species, and a liquid phase comprising liquid sodium with sodium silicate dissolved therein. The result of mixing 612a the second process stream from the separation 304b with the products of the reaction 505b is also phase-separated 506 into a solid phase and a liquid phase. In this case, the solid phase comprises a mixture of elemental iron, elemental manganese and sodium oxide, and the liquid phase comprises liquid sodium with sodium silicate dissolved therein. Both solid-liquid phase separations 305b, 506 may be conducted using one or more of the techniques described previously. The two liquid phases from these two phase separations 305b, 506 are then combined by mixing them 612b, and the resulting mixture is reacted 602 with at least one of oxygen and water to produce the alkaline mixture comprising sodium silicate with at least one of sodium oxide and sodium hydroxide. Meanwhile, the solid phase from the phase separation 506 is added to liquid water to hydrate 507 the sodium oxide therein and produce an aqueous solution of sodium hydroxide, in which the elemental iron and manganese do not dissolve. These undissolved target metals are then phase-separated 508 from the aqueous solution of sodium hydroxide to leave a mixture of iron and manganese suitable for the production of ferromanganese. Similarly, the solid phase from the phase separation 305b is also added to liquid water to hydrate the sodium oxide therein and produce an aqueous solution of sodium hydroxide, in which the gangue species remain undissolved. This undissolved gangue may then be phase-separated from the aqueous solution of sodium hydroxide to recover the sodium hydroxide, which may also be combined with the aqueous solution of sodium hydroxide from the phase separation 508. Either or both streams of recovered sodium hydroxide may then be recycled to produce more liquid sodium by electrolysis, for example. Fig. 9 shows a ninth embodiment of a method 600i of producing an alkaline mixture comprising sodium silicate from an ore comprising a siliceous mineral, wherein liquid sodium with sodium silicate dissolved therein is recycled. Thus the method 600i comprises reacting 601 comminuted, dehydrated and dehydroxylated ore particles in an inert atmosphere with an amount of liquid sodium in excess of the stoichiometric amount thereof required for a redox reaction between the liquid sodium and an oxide of the metal from the ore. For clarity, the prior comminution, drying and dehydroxylation of the ore have been omitted from Fig. 9, but are still present in the method 600i nonetheless. As described above in relation to Fig. 1, the reaction 601 is conducted at a temperature of at least 320 ’Celsius to induce a reaction between the sodium oxide thus formed and silica derived from the siliceous mineral to produce sodium silicate, but below a temperature at which the liquid sodium can react with the metal oxide to produce a ternary oxide thereof. As also described before, the reaction 601 precipitates out from the liquid sodium both the target metal in elemental form and other insoluble products comprising sodium oxide. In this embodiment, the amount of liquid sodium to react with the ore particles is selected based on the silica content of the comminuted, dehydrated and dehydroxylated ore, so that the concentration of sodium silicate dissolved in the excess liquid sodium which remains after the reaction 601 is relatively high, and the other insoluble products may even comprise a residual amount of undissolved silica. At least some of the liquid sodium with this relatively high concentration of sodium silicate dissolved therein is then phase-separated 305 from the undissolved metal and the other insoluble products and is cooled 309 until the dissolved sodium silicate reaches saturation solubility in the liquid sodium. The supersaturated solution of sodium silicate in liquid sodium this produces is then seeded 603 with nucleation sites, such as with crystals of sodium silicate obtained previously, to precipitate out from this supersaturated solution, sodium silicate in solid phase. At least some of this solid sodium silicate is then phase-separated 604 from the liquid sodium before the liquid sodium, which now therefore has a relatively lower concentration of sodium silicate dissolved therein, is recycled 314 back to a reaction vessel in which the reaction 601 is conducted. At this point, apart from adding more comminuted, dehydrated and dehydroxylated ore to the reaction mixture, fresh liquid sodium is also added to the reaction vessel. This therefore replenishes the liquid sodium which has been consumed by the reaction 601 previously, dilutes the concentration of the sodium silicate dissolved in the recycled liquid sodium to allow more sodium silicate to dissolve in the excess liquid sodium which remains after the reaction 601, and, if the fresh liquid sodium has a higher temperature as a result of having just been produced by electrolysis, also helps to bring the temperature of the reaction mixture back up above its minimum temperature of 320 ’Celsius. Meanwhile, the amount of solid sodium silicate which has been phase-separated 604 from the liquid sodium continues to accumulate as the cycle is repeated. Fig. 10 is a graph plotting temperature in the cycle of Fig. 9 on the y-axis or ordinate against the progress of a sequence of successive stages in that embodiment on the x-axis or abscissa, where for the sake of example, the target metal is iron. Whereas the ordinate of this graph is marked by a linear temperature scale, the abscissa does not represent time but instead represents an ordered sequence of events, on which for illustrative purposes only, each stage of the cycle of Fig. 9 has been assigned an equal portion of the abscissa. In practice, however, each of the stages may last a different length of time from each other. In Fig. 10, portions of the abscissa which are labelled "Reaction and phase separation" each comprise both the reaction 601 of the comminuted, dried and dehydroxylated ore with liquid sodium and the subsequent phase separation 305 of the products of that reaction into a solid phase and a liquid phase. At point A on the graph, the comminuted, dried and dehydroxylated ore particles are first introduced to the liquid sodium. The initial temperature of this reaction mixture at point A is chosen to be above a minimum temperature of 350 ’Celsius, which is indicated on the graph by the dashed line labelled Tmin (reaction). This temperature is itself well above the minimum temperature of 320 ’Celsius to induce a reaction between the sodium oxide formed by the redox reaction and silica derived from the siliceous mineral in the ore to produce sodium silicate. As a result of the reaction 601, which is exothermic, the temperature of the reaction mixture rises until it reaches point B on the graph. However, enough heat is extracted from the reaction mixture to ensure that point B remains below the temperature at which ternary oxides of sodium and the target metal can start to form, which is indicated on the graph by the dashed line labelled T (ternary oxides form). Since in this example, the target metal is iron, the line labelled T (ternary oxides form) is therefore at 447 ’Celsius on the graph, which is the temperature at which the ternary oxide Na4FeO3 can start to form. The phase separation 305 of the products of the reaction 601 into a solid phase and a liquid phase is also conducted whilst ensuring that the temperature of both the solid phase and the liquid phase does not rise above the line labelled T (ternary oxides form). Between points B and C on the graph, the liquid sodium with sodium silicate dissolved therein is then cooled 309 to below a temperature at which the sodium silicate reaches saturation solubility in the liquid sodium, which is indicated on the graph by the dashed line labelled T (supersaturation). This temperature will depend on the concentration of the sodium silicate in the liquid sodium after the phase separation 305, but for the sake of example, it is shown as being at 265 ’Celsius. On the other hand, the liquid sodium with sodium silicate dissolved therein is only cooled 309 enough to remain well above the melting point of sodium, which is indicated on the graph by the dashed line labelled Tm (Na). Between points C and D on the graph, which are approximately isothermal, the supersaturated solution of sodium silicate in liquid sodium is then seeded 603 to precipitate out sodium silicate in solid phase from the supersaturated solution, and at least some of this solid sodium silicate is then phase-separated 604 from the liquid sodium. Between points D and A' on the graph, the liquid sodium, which therefore now has a relatively lower concentration of sodium silicate dissolved therein than before it was cooled 309, is combined with fresh liquid sodium of a higher temperature. This brings the temperature of the combined liquid sodium at point A' back up above the dashed line labelled Tmin (reaction) again. The combined liquid sodium is then returned 314 to the reaction vessel in which the reaction 601 is conducted, and more comminuted, dehydrated and dehydroxylated ore is added to the reaction vessel as well. The cycle then repeats in the same manner from point A' to points B' and C' on the graph (and D', and so on) as just described above in relation to points A, B, C and D, and the amount of solid sodium silicate produced continues to accumulate as the cycle is repeated. Thus it may be seen that the temperatures of the reaction 601 and of the phase separation 305 always remain between the lines labelled Tmin (reaction) and T (ternary oxides form), whereas the temperatures of the seeding 603 and of the phase separation 604 always remain between the lines labelled T (supersaturation) and Tm (Na). Fig. 11 shows part of a tenth embodiment of a method 600j of producing an alkaline mixture comprising sodium silicate. Initially, the method 600j proceeds as described above in relation to Fig. 1 until it produces liquid sodium with sodium silicate dissolved therein. In the present embodiment, however, reacting 602 at least some of this liquid sodium having sodium silicate dissolved therein with at least one of oxygen and water comprises spraying 605 at least some of the liquid sodium with sodium silicate dissolved therein into an oxygenated atmosphere to combust the liquid sodium. The liquid sodium is therefore oxidised to sodium oxide in a strongly exothermic reaction, which produces an alkaline mixture comprising both sodium oxide and sodium silicate. In this embodiment, the liquid sodium with sodium silicate dissolved therein is sprayed 605 into the oxygenated atmosphere inside a copper combustion chamber lined with a nickel-chromium alloy. This lining protects the copper combustion chamber from the alkaline mixture thus produced, as well as from the high temperature of this exothermic reaction, which plateaus above about 800 ’Celsius. However, because both the lining and the copper of the combustion chamber have good thermal conductivity, this also allows heat to be extracted 606 from within the combustion chamber. At least some of the heat thus extracted 606 is then used 607 to contribute to the thermal decomposition of an unreactive clay, to produce an activated clay. In general, any of the alkaline mixture comprising sodium silicate produced by a method according to the present invention may be used as an alkaline activator for an alkaline activated or geopolymer cement. This may be done, for example, by combining the alkaline mixture with an activated clay in desired proportions, and then subsequently adding water to them to initiate their reaction. However, in the particular embodiment of Fig. 11, at least some of the alkaline mixture from the combustion chamber is used 608a as an alkaline activator for at least some of the activated clay which is produced using heat extracted 606 from within the combustion chamber, in order to produce a dry cement, ready for activation. Nevertheless, in other possible embodiments, at least some of the activated clay may come from another source. Whereas in the embodiment of Fig. 11, at least some of the alkaline mixture produced by the present invention is used as an alkaline activator for an alkaline activated or geopolymer cement, the alkaline mixture of the invention has many other possible uses as well. For example, it may be used to mineralise captured carbon dioxide. Fig. 12, therefore, shows part of an eleventh embodiment of a method 600k of producing an alkaline mixture comprising sodium silicate, wherein the method 600k comprises reacting 610 at least some of this alkaline mixture with gaseous carbon dioxide to produce a composition comprising sodium carbonate and silica. In the embodiment of Fig. 12, the carbon dioxide may be captured 609 from atmospheric air and / or from one or more industrial processes. For example, the carbon dioxide may be captured 306 from the combustion of organic matter mixed in with bauxitic ore, as in the embodiment of Fig. 4, and / or it may be captured from the thermal decomposition of a carbonate mineral, such as the carbon dioxide which is captured 510 from the thermal decomposition of a manganese ore comprising rhodochrosite, as in the embodiment of Fig. 6. However, regardless of the source of the carbon dioxide which is captured 609, in the present embodiment, the method 600k further comprises using 611 the composition comprising sodium carbonate and silica thus produced as an ingredient in the manufacture of soda-lime glass. Fig. 13 schematically shows a first embodiment of an apparatus 6a for producing an alkaline mixture comprising sodium silicate from an ore of a target metal. The apparatus 6a comprises a comminution device 2, a dryer 4, a gas-tight first reaction vessel 130, a first solid-liquid sodium phase separator 190 and a second reaction vessel 620. The comminution device 2 is for comminuting the ore into fines and the dryer 4 is for dehydrating the ore and dehydroxylating hydroxylated compounds contained therein. The ore is fed firstly to the comminution device 2 and from there to the dryer 4. Water vapour and other gases like sulphur dioxide are liberated from the ore in the dryer 4. The sulphur dioxide may be captured using a known technique for scrubbing flue gases. A dryer 4 suitable for use in the apparatus 6a is described below in relation to Fig. 14. The first reaction vessel 130 comprises a first inlet 131 for receiving the comminuted, dehydrated and dehydroxylated ore and a second inlet 132 for liquid sodium. The first reaction vessel 130 is for reacting the ore thus treated in an inert atmosphere with an amount of liquid sodium in excess of the stoichiometric amount thereof required for a redox reaction between the liquid sodium and an oxide of the target metal from the ore, to precipitate out from the liquid sodium a solid phase comprising both the target metal in elemental form and other insoluble products comprising sodium oxide. The redox reaction is conducted at a temperature of at least 320 ’Celsius to induce a reaction between the sodium oxide and silica from the ore to produce sodium silicate, and below a temperature at which the liquid sodium can react with the oxide of the target metal to produce a ternary oxide thereof. For example, the first reaction vessel 130 may contain a bath of liquid sodium to which the ore is added. Liquid sodium consumed in the reaction is replenished by fresh liquid sodium introduced via the second inlet 132, in order to maintain the level of liquid sodium in the first reaction vessel 130 roughly constant. The first reaction vessel 130 also comprises an outlet 134 for the liquid sodium with the target metal and other insoluble products comprising sodium oxide produced by the reaction entrained therein. The first reaction vessel 130 is sealed in a gas-tight manner from its surrounding environment to prevent the liquid sodium it contains from reacting with oxygen from atmospheric air. A head space in the reaction vessel 130 above the bath of liquid sodium is filled with an atmosphere consisting of one or more inert gases, such as nitrogen and / or argon. Either nitrogen or argon, or both, may be produced on site by pressure swing adsorption (PSA) of atmospheric air. An on-site PSA generator of inert gas may be powered using waste heat from the first reaction vessel 130 itself or from another source. To prevent atmospheric oxygen from being introduced into the first reaction vessel 130 whenever the vessel is fed with ore, in this embodiment, the first inlet 131 comprises an airlock 20, which is described below in relation to Fig. 15. The first reaction vessel 130, which is a stirred tank reactor, is made of steel of a type normally used for containing and transporting liquid sodium. An interior surface of the first reaction vessel 130 comprises a sacrificial layer to accommodate corrosion caused by sodium oxide produced by the reaction. Liquid sodium introduced into the reaction vessel 130 establishes a gradient of concentration of dissolved species in the sodium bath, from the purest liquid sodium at the second inlet 132 to the highest concentration of dissolved species at the outlet 134. For safety, the first reaction vessel 130 may also comprise an overflow outlet for liquid sodium, as well as a pressure relief valve. This is preferably located at or near the top of the head space in the reaction vessel 130, not only to relieve excess pressure, but also to vent any light gases, such as hydrogen, which would tend to collect there. The outlet 134 of the first reaction vessel 130 is connected to an inlet 191 of the first solid-liquid sodium phase separator 190, which is for separating the liquid sodium with sodium silicate dissolved therein from the solid phase comprising both the target metal in elemental form and other insoluble products comprising sodium oxide. The first solid-liquid sodium phase separator 190 therefore comprises, in addition to the inlet 191 for receiving the liquid sodium with the solid phase entrained therein from the outlet 134 of the first reaction vessel 130, a first outlet 194 for liquid sodium with sodium silicate dissolved therein, and a second outlet 195 for the solid phase. The first outlet 194 of the first solid-liquid sodium phase separator 190 is connected to a first inlet 161 of a second reaction vessel 620, which is for reacting the liquid sodium having sodium silicate dissolved therein with at least one of oxygen and water to produce a mixture comprising sodium silicate and at least one of sodium oxide and sodium hydroxide. The second reaction vessel 620 therefore comprises, in addition to the first inlet 621 for the liquid sodium with sodium silicate dissolved therein, a second inlet 622 for at least one of oxygen and water, and an outlet 624 for the mixture thus produced. A suitable second reaction vessel 620 is described below in relation to Fig. 16. Fig. 14 schematically shows an embodiment of a dryer 4 suitable for use in an apparatus of the invention, such as in the apparatus 6a. The dryer 4 comprises an entry 6 and an exit 7 located in proximity to each other at one end of the dryer 4. The entry 6 and exit 7 may each be provided with spring-loaded doors and / or curtains to help maintain and control an atmosphere within the dryer 4. The atmosphere within the dryer 4 may also be held at a pressure below that of atmospheric air outside the dryer 4, to draw air in through the entry 6 and / or exit 7 and thereby hinder or prevent the release of gases from within the dryer 4 via the entry 6 and / or exit 7. Ore 1 of the target metal at or close to ambient temperature enters the dryer 4 via the entry 6 on a first conveyor 8a and travels towards a hot zone 9 located at an opposite end of the dryer 4 from the entry 6 and exit 7. The hot zone 9 comprises means for heating the ore 1, such as pipework containing liquid sodium as a heat transfer fluid. This establishes a temperature gradient from the entry 6 and exit 7, increasing along the dryer 4 towards the hot zone 9 at the opposite end thereof. A fan 10 mounted in a duct 11 extracts water vapour and other gases, such as sulphur dioxide, driven off from the heated ore through a vent 12 also located in the hot zone 9. This reduces the partial pressure of these gases within the dryer 4, which promotes dehydration, dehydroxylation and desulphurization of the ore 1. The fan 10 can also be used to maintain the total pressure within the dryer 4 below that of the air outside the dryer 4 by overcompensating for the increased pressure of the atmosphere within the dryer 4 caused by its expansion on heating. The dryer 4 also comprises an inlet 13, whereby, for example, dry air or inert gas may be introduced into the dryer 4, in order to adjust the level of oxygen within the dryer 4. Thus, for example, if the iron ore 1 comprises red bauxite having organic matter mixed in with the ore, dry air may be introduced into the dryer 4 via the inlet 13 to replace oxygen consumed by combustion of the organic matter within the hot zone 9. When the ore 1 reaches the hot zone 9, it falls under gravity onto a second conveyor 8b contained within the dryer 4 and is transported by this in the opposite direction back towards the exit 7. As the ore travels from the hot zone 9 back towards the exit 7, it radiates heat towards the ore entering the dryer 4 on the first conveyor 8a. The dryer 4 may have a reflective lining to increase a rate of this radiative heat transfer. The radiative heat transfer may also be enhanced by convective heat transfer, by circulating the atmosphere within the dryer 4 from the second conveyor 8b towards the first conveyor 8a using appropriate ducting and / or one or more additional fans. Thus heat is transferred from the dehydrated and dehydroxylated ore leaving the dryer to fresh ore entering the dryer in a counterflow system within the dryer 4. When the ore 1 reaches the end of the second conveyor 8b, it falls under gravity through the exit 7 onto a third conveyor 8c and is transported by this towards the first inlet 131 of the first reaction vessel 130. Whereas in Fig. 14, the dryer 4 is schematically represented as being only about twice as long as it is high, this is for illustrative and explanatory purposes only. In practice, the dryer 4 may be several tens of metres long in comparison to a height of only about one or two metres, so that the entry 6 and exit 7 are remotely located from the hot zone 9, and a significant temperature gradient can be established and maintained between the entry 6 and exit 7 at one end of the dryer 4 and the hot zone 9 at the other end. If the means for heating the ore in the hot zone 9 comprises pipework 5 which contains liquid sodium intended for the first reaction vessel 130, the counterflow system described above will increase the temperature at which the ore arriving in the hot zone 9 reaches thermal equilibrium with the liquid sodium, but will also decrease the temperature at which the ore leaves the dryer 4. Thus liquid sodium leaving the pipework 5 will enter the first reaction vessel 130 at a higher temperature, but the ore leaving the dryer 4 will enter the first reaction vessel 130 at a lower temperature, than if no such counterflow system were present. However, the total quantity of heat transferred to the first reaction vessel 130 by the liquid sodium and the ore together remains the same as if no such counterflow system were present. If on the other hand, the means for heating the ore in the hot zone 9 supplies heat to the ore from another source, the total quantity of heat transferred to the first reaction vessel 130 will be different. Fig. 15 schematically shows an embodiment of an airlock 20 suitable for use as part of the first inlet 131 of the first reaction vessel 130. The airlock 20 comprises two gas-tight doors 21, 22. The first such door 21 connects an interior 25 of the airlock 20 with the surrounding environment E, whereas the second such door 22 connects the interior 25 of the airlock 20 with an inside 135 of the first reaction vessel 130. The airlock 20 also comprises an inlet 23 to the interior 25 of the airlock 20 for inert gas and an outlet 24 from the interior 25 of the airlock 20 for atmospheric air. The inlet 23 and the outlet 24 each comprises a respective valve 23V, 24V for opening and closing a respective one of the inlet 23 and the outlet 24. To introduce a fresh charge of ore via the airlock 20 into the first reaction vessel 130, the first gas-tight door 21 is opened and the ore particles are introduced into the interior 25 of the airlock 20, whilst the second gas-tight door 22 is kept closed. The first door 21 is then closed again and the interior 25 of the airlock 20 is flushed with inert gas by opening the valve 23V, and then also opening the valve 24V to purge atmospheric air from the airlock 20 via outlet 24. Both valves 23V, 24V are then closed again, before the second door 22 is opened to introduce the ore into the first reaction vessel 130, whilst the first gas-tight door 21 is kept closed. The second gas-tight door 22 is then closed again, and the airlock 20 is ready to be used again, to introduce another charge of ore into the reaction vessel 130. To prevent the first reaction vessel 130 from being pressurized by inert gas which enters the reaction vessel 130 from the interior 25 of the airlock 20 together with the ore, by reason of the inert gas expanding as it heats up from the temperature of the surrounding environment E to the temperature inside 135 the first reaction vessel 130, the inert gas may be preheated before it is injected via inlet 23 into the airlock 20 to the same temperature as the temperature of the atmosphere of inert gas inside 135 the reaction vessel 130. This may be done using waste heat from the first reaction vessel 130 itself or from another source. For example, as shown in Fig. 15, the inert gas may be heated by liquid sodium from the first reaction vessel 130, via a heat exchanger HE6. Preheating the inert gas also has the advantage that it pressurizes the inert gas before it is injected into the airlock 20, thereby avoiding the need for the inert gas to be pumped into the airlock 20 and helping to purge the atmospheric air from the interior 25 of the airlock 20. Fig. 16 schematically shows an embodiment of a second reaction vessel 620 for reacting liquid sodium having sodium silicate dissolved therein with at least one of oxygen and water, which is suitable for use in an apparatus of the invention, such as in the apparatus 6a. The reaction vessel 620 comprises a combustion chamber 625 which has a first inlet 621 for liquid sodium with sodium silicate dissolved therein, a second inlet 622 for at least one of oxygen and water, and an outlet 624 for a mixture comprising sodium silicate and at least one of sodium oxide and sodium hydroxide. The first inlet 621 comprises a nozzle 623 for spraying the liquid sodium with sodium silicate dissolved therein into the combustion chamber 625 and the second inlet 622 is adapted to admit at least one of gaseous oxygen and water vapour into the combustion chamber 625 as well. Liquid sodium with sodium silicate dissolved therein which is sprayed through the nozzle 623 forms a fine mist of droplets inside the combustion chamber 625 and spontaneously ignites in the presence of oxygen because of its large surface area and already elevated temperature. At atmospheric pressure, the temperature of the liquid sodium rises as it combusts until it reaches a plateau above about 800 ’Celsius. The combustion chamber 625 is made of copper and has a lining 626 made of a nickel-chromium alloy. The lining 626 is resistant to corrosion by the highly caustic sodium oxide formed by combustion, but also has good thermal conductivity, allowing heat from this intensely exothermic reaction to be transmitted through the combustion chamber 625 to an exterior 627 thereof. Since this combustion reaction consumes gas, a negative pressure differential is created from the exterior 627 to inside the combustion chamber 625, which draws gaseous oxygen and / or water vapour in through the second inlet 622 without any need for them to be pumped. Oxygen and / or water vapour may be supplied from ordinary atmospheric air, the other main constituents of which ( / .e., nitrogen and argon) act as inert carrier gases. The air may be scrubbed of carbon dioxide before being introduced into the combustion chamber 625 to prevent it from reacting with the sodium oxide formed by combustion. The humidity of the air should initially be kept at a very low level (for example, by firstly drying the air) because too humid air results in the droplets of liquid sodium being coated in a surface layer of sodium hydroxide, which inhibits their complete combustion. Combustion may therefore be conducted in two stages, firstly by admitting dry air into the combustion chamber 625 via the second inlet 622 to produce a mixture of sodium silicate and sodium oxide, followed by water vapour or humid air to hydrate the sodium oxide thus formed and thereby produce a mixture comprising sodium silicate and at least one of sodium oxide and sodium hydroxide in a desired ratio. The resulting mixture comprising sodium silicate and at least one of sodium oxide and sodium hydroxide forms a powder, which settles to the bottom of the combustion chamber 625, from where it may be removed via the outlet 624. The reaction vessel 620 may further comprise a stirring mechanism (not shown) to ensure complete combustion and proper mixing of the reaction products. The reaction vessel 620 also comprises the following safety features to ensure that this combustion can be controlled and is able to be extinguished if desired. The walls of the combustion chamber 625 are sufficiently thick for it to withstand complete evacuation. The outlet 624 is fitted with a door 628 which remains closed during normal operation and is only opened to remove the accumulated mixture of powdered combustion products. The second inlet 622 is fitted with a filter 622a, comprising a substrate made, for example, of sintered titanium fibre felt, which can capture the powdered reaction products and which also prevents flames from escaping from within the combustion chamber 625. The filter 622a also protects the inside of the combustion chamber 625 from contamination. If the filter 622a becomes blocked, the reaction vessel 620 fails safe because the combustion reaction dies due to lack of oxygen. However, during normal operation, the filter 622a is unlikely to come into contact with any reaction products because of the negative pressure differential which draws the gaseous reagents into the combustion chamber 625. The first inlet 621 is fitted with both a non-return valve and an on / off valve upstream of the nozzle 623. The reaction vessel 620 further comprises a plurality of valves 629a, 629b for injecting the combustion chamber 625 either with an inert gas, such as nitrogen and / or argon, or with liquid nitrogen. The valves 629a, 629b are multiple in number to provide redundancy in case any one of them fails. Thus, the reaction rate may be controlled not only by varying the rate of supply of liquid sodium through the first inlet 621, but also by injecting inert gas through the valves 629a, 629b, which reduces the relative amount of the gaseous reagents present inside the reaction vessel 620. Inert gas may also be injected via the valves 629a, 629b to help the powdered reaction products settle to the bottom of the reaction vessel 620. If it is desired to extinguish the combustion reaction, the supply of liquid sodium may be stopped by shutting off the on / off valve upstream of the nozzle 623 and liquid nitrogen may be rapidly injected into the combustion chamber 625 via the valves 629a, 629b instead. This is effective at rapidly cooling both the reaction products and any remaining liquid sodium without contaminating them by vaporizing the liquid nitrogen. Any remaining oxygen is expelled thereby back out of the second inlet 622, but the filter 622a inhibits the escape of reaction products and / or flames. The reaction vessel 620 may then be cooled back down to ambient temperature before the door 628 can be opened. After such an event, the filter 622a may be removed and replaced with a new one. Fig. 17 schematically shows a second embodiment of an apparatus 6b for producing an alkaline mixture comprising sodium silicate. Like the apparatus 6a described above in relation to Fig. 13, the apparatus 6b also comprises a comminution device 2, a dryer 4, a gas-tight first reaction vessel 130, a first solid-liquid sodium phase separator 190 and a second reaction vessel 620. The comminution device 2, the dryer 4 and the first reaction vessel 130 in the apparatus 6b are arranged and function as described previously. In the present embodiment, however, the first outlet 194 of the first solidliquid sodium phase separator 190 is connected to a two-way valve VI, which is itself connected both to the first inlet 621 of the second reaction vessel 620 and to a first inlet 631 of a precipitation tank 630. The two-way valve VI can therefore be operated to direct liquid sodium with sodium silicate dissolved therein from the first solid-liquid sodium phase separator 190 either to the second reaction vessel 620 or to the precipitation tank 630. In addition to the first inlet 631 thereof, the precipitation tank 630 also comprises a second inlet 632 for receiving crystals of solid sodium silicate and an outlet 634 for liquid sodium having both sodium silicate dissolved therein and solid sodium silicate entrained therein. Moreover, the apparatus 6b further comprises a second solid-liquid sodium phase separator 640. This second solid-liquid sodium phase separator 640 comprises an inlet 641 for receiving liquid sodium with both sodium silicate dissolved therein and solid sodium silicate entrained therein from the outlet 634 of the precipitation tank 630, a first outlet 644 for solid sodium silicate and a second outlet 645 for liquid sodium with sodium silicate dissolved therein. The inlet 641 is connected to the outlet 634 of the precipitation tank 630 via an open / shut valve V2. Thus the inlet 641 is also downstream of the first outlet 194 of the first solid-liquid sodium phase separator 190, being connected thereto via the open / shut valve V2, the precipitation tank 630 and the two-way valve VI. The first outlet 644 is arranged to add solid sodium silicate to the mixture comprising sodium silicate and at least one of sodium oxide and sodium hydroxide which is produced in the second reaction vessel 620 and which comes out of the outlet 624 thereof. The first outlet 644 of the second solidliquid sodium phase separator 640 is also connected to the second inlet 632 of the precipitation tank 630. This allows a small portion of the solid sodium silicate from the first outlet 644 to be abstracted from a major portion thereof which is added to the alkaline mixture produced in the second reaction vessel 620, to provide nucleation sites for precipitation of solid sodium silicate in the precipitation tank 630. The second outlet 645 of the second solid-liquid sodium phase separator 640 is upstream of and connected to the second inlet 132 of the first reaction vessel 130. During operation, both the precipitation tank 630 and the second solid-liquid sodium phase separator 640 are maintained below a temperature at which sodium silicate in the liquid sodium received from the first outlet 194 of the first solid-liquid sodium phase separator 190 reaches saturation solubility. Occasionally, however, as part of routine maintenance, the temperature of the precipitation tank 630 and of the second solid-liquid sodium phase separator 640 may be raised above the temperature at which sodium silicate in the liquid sodium received from the first outlet 194 of the first solid-liquid sodium phase separator 190 reaches saturation solubility. This allows the tank 630 and the separator 640 to be flushed through with liquid sodium and for any residual solid sodium silicate remaining in the tank 630 and / or in the separator 640 to redissolve in the liquid sodium, which prevents the tank 630 and the separator 640 from plugging. Like the first reaction vessel 130, and the first and second solid-liquid sodium phase separators 190, 640, the precipitation tank 630 is made of steel of a type normally used for containing and transporting liquid sodium. The apparatus 6b allows liquid sodium with sodium silicate dissolved therein to be circulated in a loop through the first reaction vessel 130. The apparatus 6b therefore functions as follows. When valve VI is positioned to direct liquid sodium with sodium silicate dissolved therein from the first solid-liquid sodium phase separator 190 to the precipitation tank 630, valve V2 can be closed to allow the precipitation tank 630 to fill up, and crystals of solid sodium silicate are introduced into the precipitation tank 630 via the second inlet 632 thereof. Since the precipitation tank 630 is maintained below a temperature at which sodium silicate in the liquid sodium reaches saturation solubility, these crystals of solid sodium silicate act as nucleation sites for solid sodium silicate to precipitate out in the tank 630. Valve V2 is then opened again, the precipitation tank 630 drains into the second solid-liquid sodium phase separator 640, and the precipitated solid sodium silicate is separated from the liquid sodium with sodium silicate dissolved therein, but now at a lower concentration, by the second solidliquid sodium phase separator 640. This liquid sodium with a relatively lower concentration of sodium silicate dissolved therein leaves the second solid-liquid sodium phase separator 640 via the second outlet 645 thereof and is returned to the first reaction vessel 130 via the second inlet 132 thereof, where it is diluted down but also heated up by being mixed with fresh liquid sodium produced by electrolysis, and where more comminuted, dehydrated and dehydroxylated ore is introduced to the first reaction vessel 130 as well, via the first inlet 131 thereof. Meanwhile, solid sodium silicate continues to accumulate at the first outlet 644 of the second solid-liquid sodium phase separator 640. This process may be repeated until enough solid sodium silicate has accumulated at the first outlet 644 of the second solid-liquid sodium phase separator 640, whereupon the valve VI is repositioned to direct liquid sodium with sodium silicate dissolved therein from the first solid-liquid sodium phase separator 190 to the second reaction vessel 620 instead. The recirculation of liquid sodium with sodium silicate dissolved therein from the first outlet 194 of the first solid-liquid sodium phase separator 190 back to the first reaction vessel 130 therefore ceases. Instead, the liquid sodium with sodium silicate dissolved therein now reacts with at least one of oxygen and water inside the second reaction vessel 620 to produce a mixture comprising sodium silicate and at least one of sodium oxide and sodium hydroxide. Since the concentration of sodium silicate dissolved in the liquid sodium is limited by its saturation solubility at the temperature of the liquid sodium, this mixture comprises a majority of at least one of sodium oxide and sodium hydroxide and only a minority of sodium silicate. However, by adding the solid sodium silicate which has accumulated at the first outlet 644 of the second solid-liquid sodium phase separator 640 thereto, an alkaline mixture comprising sodium silicate and at least one of sodium oxide and sodium hydroxide in desired proportions can be obtained. Whereas in Fig. 17, the first outlet 644 of the second solid-liquid sodium phase separator 640 is arranged to add solid sodium silicate to the mixture comprising sodium silicate and at least one of sodium oxide and sodium hydroxide which comes out of the outlet 624 of the second reaction vessel 620, in other possible embodiments, solid sodium silicate from the outlet 644 of the second solidliquid sodium phase separator 640 may alternatively or additionally be directed inside the second reaction vessel 620. Fig. 18 schematically shows part of a third embodiment of an apparatus 6c for producing an alkaline mixture comprising sodium silicate. Like the apparatuses 6a and 6b described above, the apparatus 6c comprises a second reaction vessel 620 for reacting liquid sodium having sodium silicate dissolved therein with at least one of oxygen and water to produce a mixture comprising sodium silicate and at least one of sodium oxide and sodium hydroxide. As before, the second reaction vessel 620 comprises a first inlet 621 for liquid sodium with sodium silicate dissolved therein, a second inlet 622 for at least one of oxygen and water, and an outlet 624 for the alkaline mixture thus produced. The second reaction vessel 620 comprises a copper combustion chamber 625 having a lining made of a nickelchromium alloy, as described above in relation to Fig. 16. In the present embodiment, however, the apparatus 6c further comprises a clay transport pathway 650, which itself comprises an inlet 651 for an unreactive clay and an outlet 654 for an activated clay. As schematically represented in Fig. 18, the clay transport pathway 650 is adapted to bring the unreactive clay into thermal contact with an exterior 627 of the combustion chamber 625. Thus heat can be transferred from within the second reaction vessel 620 to unreactive clay travelling along the clay transport pathway 650 and can contribute to converting the unreactive clay into an activated clay. As also shown in Fig. 18, in this embodiment, the outlet 654 of the clay transport pathway 650 is arranged to add activated clay from the outlet 654 to the mixture comprising sodium silicate and at least one of sodium oxide and sodium hydroxide which comes out of the outlet 624 of the second reaction vessel 620. An arrangement as schematically shown in Fig. 18 may be incorporated into the apparatus 6a of Fig. 13 or into the apparatus 6b of Fig. 17, or indeed into any other apparatus according to the invention, in order to provide further embodiments thereof. In summary, therefore, the present invention provides a method and apparatus for producing an alkaline mixture comprising sodium silicate, which mixture is suitable for use as an alkaline activator for alkaline activated and geopolymer cements, but which may also be used, for example, to mineralize captured carbon dioxide. Since this alkaline mixture is produced whilst extracting a target metal in elemental form from its ore, the energy consumed is split between the production of these two co-products. This makes the method of the invention a low-energy method of producing an alkaline activator for alkaline activated and geopolymer cements, which in turn makes such types of cement more economically viable as replacements for OPC. Furthermore, the invention does not consume any carbon, so is at least carbon-neutral. On the other hand, it consumes the silica content of gangue mineral species and / or of siliceous minerals of the target metal in its ore, which have previously been treated as waste products. The amount of waste material produced when extracting the target metal from its ore is thereby reduced as well. Whereas the present invention has been described above by reference to particular examples and embodiments, the scope of the invention should not be taken to be limited thereby and is instead defined by the appended claims.
Claims
1. A method (600a - 600k) of producing an alkaline mixture comprising sodium silicate, wherein the method comprises:comminuting (301) an ore of a target metal into fines, wherein the ore comprises a siliceous mineral and at least one of iron and manganese as the target metal;dehydrating (302, 302b, 302c, 502, 502a, 503) the ore and dehydroxylating hydroxylated compounds contained therein;reacting (303, 505b, 601) the ore thus treated in an inert atmosphere with an amount of liquid sodium in excess of the stoichiometric amount thereof required for a redox reaction between the liquid sodium and an oxide of the target metal from the ore, to precipitate out from the liquid sodium both the target metal in elemental form and other insoluble products comprising sodium oxide, wherein the redox reaction is conducted at a temperature of at least 320 ’Celsius to induce a reaction between the sodium oxide and silica derived from the siliceous mineral to produce sodium silicate, and below a temperature at which the liquid sodium can react with the oxide of the target metal to produce a ternary oxide thereof;phase-separating (305, 305b, 506) at least some of the liquid sodium with sodium silicate dissolved therein from the undissolved metal and the other insoluble products; andreacting (602) at least some of the phase-separated liquid sodium having sodium silicate dissolved therein with at least one of oxygen and water to produce a mixture comprising sodium silicate and at least one of sodium oxide and sodium hydroxide.
2. A method (600b) according to claim 1, wherein dehydrating and dehydroxylating (302) the ore comprises heating (302c, 502) the ore to a temperature of from 100 to 600 ’Celsius, inclusive.
3. A method (600c) according to claim 2, wherein the ore comprises at least one of the mineral and mineraloid forms of iron oxide and iron oxyhydroxide, and the redox reaction (303) is conducted at a temperature of less than 450 ’Celsius.
4. A method (600d) according to claim 3, wherein:the ore comprises red bauxite;dehydrating and dehydroxylating (302) the ore comprises heating (302b) the ore in an oxygenated atmosphere to at least 250 ’Celsius closed off from its surrounding environment; andthe method further comprises capturing (306) carbon dioxide gas produced by combustion (307) of organic matter mixed in with the ore.
5. A method (600d) according to claim 4, further comprising increasing a molar ratio of FejOa: SiOz in the red bauxite by adding at least one of an iron oxide and a non-bauxitic iron ore to the red bauxite before reacting (303) the comminuted, dehydrated and dehydroxylated ore with the excess amount of liquid sodium.
6. A method (600e, 600h) according to claim 2 or claim 3, wherein the ore comprises a manganiferous mineral, the ore is heated (502) in atmospheric air without a reductant, and the redox reaction (505b) is conducted at a temperature of less than 600 ’Celsius.
7. A method (600f) according to claim 6, wherein:the ore comprises rhodochrosite;the ore is heated (502a) to at least 200 ’Celsius closed off from its surrounding environment; andthe method further comprises capturing (510) carbon dioxide gas released from the rhodochrosite by its thermal decomposition during the dehydration and dehydroxylation (302) of the ore.
8. A method (600g) according to any one of the preceding claims, wherein the ore comprises a manganiferous mineral, the redox reaction (505b) is conducted at a temperature of less than 600 ’Celsius, and the method comprises, before reacting (303, 505b, 601) the ore with the excess amount of liquid sodium, converting the manganiferous mineral in the ore into a trivalent manganese oxide using a hydrometallurgical technique (503) comprising:adding (503a) at least some of the ore to hot, concentrated hydrochloric acid closed off from their surrounding environment to produce gaseous chlorine and an acidic aqueous solution comprising manganese (II) chloride;phase-separating (503b) from this acidic aqueous solution, a solid silica residue derived from the siliceous mineral;adding (503c) sodium oxide, or sodium hydroxide derived from hydrating at least some of this sodium oxide, to the acidic aqueous solution to produce an alkaline aqueous solution and a precipitate comprising manganese (II) hydroxide;phase-separating (503d) at least some of the precipitate from the alkaline aqueous solution;anddrying and dehydroxylating (503e) at least some of the manganese (II) hydroxide in atmospheric air at a temperature of less than 100 ’Celsius to produce the trivalent manganese oxide;and wherein the method further comprises:neutralizing (522) the solid silica residue with an aqueous solution of sodium hydroxide; washing (523) the solid silica residue with deionized water and drying it;adding (524) the washed and dried silica residue to the redox reaction (505b); and capturing (525) at least some of the gaseous chlorine.
9. A method (600c, 600i) according to any one of the preceding claims, comprising, after phaseseparating (305) at least some of the liquid sodium with sodium silicate dissolved therein from the undissolved metal and the other insoluble products and before reacting (602) at least some of it with at least one of oxygen and water:recycling (314) at least some of it at least once to a reaction vessel (130) in which the redox reaction (303) is conducted; andintroducing (A') more of the comminuted, dehydrated and dehydroxylated ore into the reaction vessel (130) when the liquid sodium with sodium silicate dissolved therein is recycled (314).
10. A method (600c, 600i) according to claim 9, further comprising cooling (309) the liquid sodium with sodium silicate dissolved therein after phase-separating (305) at least some of it from the undissolved metal and the other insoluble products and before recycling (314) at least some of it to the reaction vessel (130).
11. A method (600i) according to claim 10, comprising:cooling (309) the liquid sodium with sodium silicate dissolved therein to below a temperature at which the sodium silicate reaches saturation solubility, thereby creating a supersaturated solution of sodium silicate in liquid sodium;seeding (603) the supersaturated solution to produce sodium silicate in solid phase; and phase-separating (604) at least some of the solid sodium silicate from the liquid sodium before recycling (314) the liquid sodium to the reaction vessel (130) with a reduced concentration of sodium silicate dissolved therein.
12. A method (600i) according to any one of claims 9 to 11, further comprising adding (D) fresh liquid sodium to the reaction vessel (130) when the liquid sodium with sodium silicate dissolved therein is recycled (314).
13. A method (600j) according to any one of the preceding claims, wherein:reacting (602) at least some of the liquid sodium having sodium silicate dissolved therein with at least one of oxygen and water comprises spraying (605) at least some of the liquid sodium with sodium silicate dissolved therein into an oxygenated atmosphere to combust the liquid sodium and thereby produce a mixture comprising sodium oxide and sodium silicate;the liquid sodium with sodium silicate dissolved therein is sprayed (605) into the oxygenated atmosphere inside a thermally conductive combustion chamber (625); andthe method further comprises extracting (606) heat from within the combustion chamber via an exterior (627) of the combustion chamber (625).
14. A method (600j) according to claim 13, further comprising using (607) at least some of the heat extracted (606) from within the combustion chamber (625) to contribute to thermal decomposition of an unreactive clay.
15. A method (600j) according to any one of the preceding claims, comprising using (608) at least some of the alkaline mixture comprising sodium silicate as an alkaline activator for an alkaline activated or geopolymer cement.
16. A method (600j) according to claim 15 when dependent on claim 13, wherein using (608) at least some of the alkaline mixture comprising sodium silicate as an alkaline activator comprises adding (608a) the alkaline mixture to an activated clay produced by thermal decomposition of an unreactive clay using (607) heat extracted (606) from within the combustion chamber (625).
17. A method (600k) according to any one of the preceding claims, comprising reacting (610) at least some of the alkaline mixture comprising sodium silicate with gaseous carbon dioxide to produce a composition comprising sodium carbonate and silica.
18. A method (600k) according to claim 17, further comprising using (611) the composition comprising sodium carbonate and silica as an ingredient in the manufacture of soda-lime glass.
19. An apparatus (6a, 6b, 6c) for producing an alkaline mixture comprising sodium silicate from an ore of a target metal, wherein the ore comprises a siliceous mineral and the target metal is at least one of iron and manganese, the apparatus comprising:a comminution device (2) for comminuting the ore into fines;a dryer (4) for dehydrating the ore and dehydroxylating hydroxylated compounds contained therein;a gas-tight first reaction vessel (130) for reacting the ore thus treated in an inert atmosphere with an amount of liquid sodium in excess of the stoichiometric amount thereof required for a redox reaction between the liquid sodium and an oxide of the target metal from the ore, to precipitate out from the liquid sodium a solid phase comprising both the target metal in elemental form and other insoluble products comprising sodium oxide, wherein the redox reaction is conducted at a temperature of at least 320 ’Celsius to induce a reaction between the sodium oxide and silica derived from the siliceous mineral to produce sodium silicate, and below a temperature at which the liquid sodium can react with the oxide of the target metal to produce a ternary oxide thereof, the first reaction vessel (130) comprising a first inlet (131) for receiving the comminuted, dehydrated and dehydroxylated ore, a second inlet (132) for the liquid sodium, and an outlet (134) for the liquid sodium with the solid phase entrained therein;a first solid-liquid sodium phase separator (190) for separating the liquid sodium with sodium silicate dissolved therein from the solid phase, wherein the first solid-liquid sodium phase separator (190) comprises an inlet (191) for receiving the liquid sodium with the solid phase entrained therein from the outlet (134) of the first reaction vessel (130), a first outlet (194) for liquid sodium with sodium silicate dissolved therein, and a second outlet (195) for the solid phase; anda second reaction vessel (620) for reacting the liquid sodium having sodium silicate dissolved therein with at least one of oxygen and water to produce a mixture comprising sodium silicate and at least one of sodium oxide and sodium hydroxide, wherein the second reaction vessel (620) comprises a first inlet (621) for the liquid sodium with sodium silicate dissolved therein, a second inlet (622) for at least one of oxygen and water, and an outlet (624) for the mixture thus produced.
20. An apparatus (6a, 6b, 6c) according to claim 19, wherein the dryer (4) contains an atmosphere to which the ore is exposed during dehydration and dehydroxylation, and the dryer (4) comprises means (10, 11, 12) for reducing a pressure of the atmosphere within the dryer (4) to less than that of atmospheric air outside the dryer (4).
21. An apparatus (6b) according to claim 19 or claim 20, further comprising a second solid-liquid sodium phase separator (640) comprising an inlet (641) downstream of the first outlet (194) of the first solid-liquid sodium phase separator (190), a first outlet (644) for solid sodium silicate and a second outlet (645) for liquid sodium with sodium silicate dissolved therein upstream of the second inlet (132) of the first reaction vessel (130), wherein the second solid-liquid sodium phase separator (640) is maintained below a temperature at which the sodium silicate in the liquid sodium received from the first outlet (194) of the first solid-liquid sodium phase separator (190) reaches saturation solubility.
22. An apparatus (6b) according to claim 21, wherein the first outlet (644) of the second solidliquid sodium phase separator (640) is arranged to add the solid sodium silicate to the mixture comprising sodium silicate and at least one of sodium oxide and sodium hydroxide produced in the second reaction vessel (620).
23. An apparatus (6a, 6b, 6c) according to any one of claims 19 to 22, wherein:the second reaction vessel (620) comprises a thermally conductive combustion chamber (625);the first inlet (621) of the second reaction vessel (620) comprises a nozzle (623) for spraying the liquid sodium with sodium silicate dissolved therein into the combustion chamber (625); andthe second inlet (622) of the second reaction vessel (620) is adapted to admit at least one of gaseous oxygen and water vapour into the combustion chamber (625).
24. An apparatus (6c) according to claim 23, further comprising a clay transport pathway (650) comprising an inlet (651) for an unreactive clay and an outlet (654) for an activated clay, wherein the clay transport pathway (650) is adapted to bring the unreactive clay into thermal contact with an exterior (627) of the thermally conductive combustion chamber (625).
25. An apparatus (6c) according to claim 24, wherein the outlet (654) of the clay transport pathway (650) is arranged to add the activated clay to the mixture comprising sodium silicate and at least one of sodium oxide and sodium hydroxide from the outlet (624) of the second reaction vessel (620).