Process for the treatment of a silicate mineral

Abstract

A process for the treatment of a silicate mineral, includes: preparing a first composition including an alkali metal magnesium orthosilicate and optionally either (i) magnesium oxide or (ii) an alkali metal silicate, by reaction, at a temperature from 500 to 1200 C., of an alkali metal carbonate compound, which compound is an alkali metal carbonate, an alkali metal bicarbonate or a mixture thereof, with a magnesium silicate, the molar ratio of alkali metal carbonate compound, expressed as alkali metal oxide of the formula R.sub.2O, in which R represents an alkali metal, to magnesium silicate, expressed as silicon dioxide, of the formula SiO.sub.2, being from 4:1 to 1:4, and contacting the first composition with water to produce a second composition comprising an amorphous magnesium silicate hydrate (M-SH).

Claims

1. A process for the treatment of a silicate mineral, said process comprising: preparing a first composition comprising an alkali metal magnesium orthosilicate and optionally either (i) magnesium oxide or (ii) an alkali metal silicate, by reaction, at a temperature from 600 C. to 1000 C., of an alkali metal carbonate compound, which compound is an alkali metal carbonate, an alkali metal bicarbonate or a mixture thereof, with a magnesium silicate, the molar ratio of alkali metal carbonate compound, expressed as alkali metal oxide of the formula R.sub.2O, in which R represents an alkali metal, to magnesium silicate, expressed as silicon dioxide, of the formula SiO.sub.2, being from 4:1 to 1:4, wherein the reaction is obtained by contacting the alkali metal carbonate compound and the magnesium silicate to each other in a solid state, and contacting the first composition with water to produce a second composition comprising an amorphous magnesium silicate hydrate (M-SH).

2. A process according to claim 1, wherein M-SH is represented by an oxide formula in the form pMgO.SiO.sub.2.qH.sub.2O where p is from 0.5 to 2.0 and q is from 1 to 4.

3. A process according to claim 1, wherein the magnesium silicate is a magnesium silicate in which the molar ratio of magnesium oxide to silica is from 0.5 to 3.

4. A process according to claim 1, wherein the magnesium silicate comprises a magnesium silicate rock comprising a magnesium silicate of general composition:
m(MgO).t(QO).SiO.sub.2xH.sub.2O wherein m is from 0.5 to 3, t is less than or equal to 1, x is from zero to 2, and Q represents a metal or metals other than magnesium.

5. A process according to claim 4, wherein Q represents iron, nickel and/or chromium which process further comprises isolating an iron, nickel and/or chromium compound.

6. A process according to claim 1, further comprising isolating or producing magnesium oxide, magnesium hydroxide or an alkali metal silicate from the second composition.

7. A process according to claim 1, further comprising a step in which the second composition is carbonated to produce a third composition comprising a magnesium carbonate compound.

8. A process according to the claim 7, further comprising a step in which the magnesium carbonate from the third composition is calcined to produce a fourth composition comprising magnesium oxide.

9. A process according to claim 7, wherein the carbonation of the second composition is conducted in a second substep, after a first substep during which the second composition is formed by molding or pressing to form a shaped article, such that a carbonated shaped article is obtained.

10. A process according to claim 1, wherein the temperature is from 800 to 1000 C.

Description

EXAMPLE 1

(1) Ground pure olivine sand (Mg.sub.0.94Fe.sub.0.06).sub.2SiO.sub.4 was mixed with sodium carbonate hydrate powder in a 14:11 mass ratio. The molar ratio of this mixture is close to 1:1. 25 g of this powder was pressed into a pellet and placed into a lab furnace in a platinum crucible. It was heated up to 800 C. over 2 hours, maintained at 800 C. for 1 hour and cooled back to room temperature by natural cooling. In order to evaluate the reaction efficiency, separate samples of the sodium carbonate and olivine used underwent exactly the same heat treatment. The measured ignition losses are shown in the following Table.

(2) TABLE-US-00001 Sample Loss on Ignition Sodium carbonate powder 14.97% Ground olivine sand 0.22% 14:11 mix of ground olivine sand + 16.69% sodium carbonate powder

(3) The ignition loss of the pure sodium carbonate was due entirely to loss of hydrate water and not to decomposition or evaporation of the carbonate. The pure olivine sample remained essentially unchanged during heat treatment, but the mixed sample reacted. From the measured LOI data, it is estimated that 60% of the CO.sub.2 from the carbonate was lost from the mixture. The phase constitution of the reacted sample was determined by XRD. The reaction products observed by this technique were Na.sub.2MgSiO.sub.4, MgO and Fe.sub.2O.sub.3 together with some unreacted olivine and sodium carbonate.

(4) 10 g of the reacted sample was put into 100 ml of de-ionised water under constant stirring at 40 C. for 1 h in order to evaluate its dissolution behaviour. The solution was filtered and the solid residue was analysed by XRD. The amounts of dissolved elements in the aqueous solutions were measured by ICP.

(5) The main solid phases detected in the filtered solid residue were Na.sub.2MgSiO.sub.4, MgO, Fe.sub.2O.sub.3 and olivine. All non-reacted sodium carbonate dissolved in the aqueous solution. Additionally, some dissolved SiO.sub.2 was detected with a concentration of 667 mg/I. From this value, one can estimate that 2 to 3% of the total silica from the mixture dissolved in the water under these conditions, (presumed to be as a sodium silicate, as the pH of the solution was measured to be about 12).

EXAMPLE 2

(6) The same ground pure olivine sand (Mg.sub.0.94Fe.sub.0.06).sub.2SiO.sub.4 and sodium carbonate hydrate powder as used in Example 1 were mixed in a 1400:1235 mass ratio. The molar ratio of this mixture is approximately 1:1. The powder was pressed into pellets and placed (in a platinum crucible) into a lab furnace which was kept at 900 C. The sample was air-quenched after 1 h of heat treatment. Separate samples of the sodium carbonate and olivine used underwent an identical heat treatment. It is important to mention that at this temperature, sodium carbonate is in the liquid state. (The melting temperature of this compound is 851 C.). The measured loss on ignition of all of the samples is presented in the following Table.

(7) TABLE-US-00002 Sample Loss on Ignition Sodium carbonate 14.96% Olivine 0.04% Olivine + Sodium carbonate 23.02%

(8) From the measured LOI data, a conversion rate (degree of decarbonation) close to 100% was obtained which means that the reaction was essentially complete. The reaction product was analysed by XRD and the phases detected were Na.sub.2MgSiO.sub.4, MgO, NaFeO.sub.2 together with some un-reacted olivine. No remaining sodium carbonate was detected, in agreement with the observed loss on ignition data. The MgO to Na.sub.2MgSiO.sub.4 mass ratio was estimated by Rietveld analysis of the XRD data to be 14:86, in reasonable agreement with mass balance calculations.

EXAMPLE 3

(9) 13 g of the powdered product of Example 2 was mixed with 260 ml of water to give a slurry with about 50 g/L solids concentration. This slurry was put in a closed pressure reactor with a total volume of 2 liters, maintained at 25 C. The reactor was first evacuated and then filled with pure gaseous CO.sub.2 up to one atmosphere pressure. The pressure of the gas and the pH of the aqueous solution were recorded as functions of time and are presented in FIG. 2. The rate of pressure drop indicates that the slurry captures gaseous CO.sub.2 rapidly, and, concurrently, the pH of the solution falls rapidly from an initial value of over 12 to a final value of about 8.5, consistent with the formation of a mixed sodium carbonate/bicarbonate solution. After this reaction the solids in the slurry were filtered and analysed by XRD. NaFeO.sub.2 was no longer detectable and the MgO:Na.sub.2MgSiO.sub.4 ratio had increased to an estimated value of 18:82, indicating that some sodium silicate had leached out of the Na.sub.2MgSiO.sub.4 phase into the solution and had then been carbonated.

EXAMPLE 4

(10) 5 g of the powdered product of Example 2 was mixed with 200 ml of water to give a slurry of about 25 g/L solids concentration. This slurry was then boiled gently for one hour, after which it was filtered and the solid residue analyzed by XRD. The relative concentration of the Na.sub.2MgSiO.sub.4 phase in the residue was clearly greatly reduced compared to the original untreated residue, and the relative concentration of MgO greatly increased. The NaFeO.sub.2 phase had also completely disappeared but peaks for a layered double hydroxide phase probably having a formula close to 4MgO.Fe.sub.2O.sub.3.CO.sub.2.10H.sub.2O were seen clearly, together with weaker peaks for Fe.sub.2O.sub.3 (haematite) and unreacted olivine.

(11) The liquid filtrate was also analyzed by ICP and the results (see the Table below) showed a high concentration (6.92 g/L as Na.sub.2O) of sodium in solution, as well as 0.47 g/L of SiO.sub.2. This confirms that most of the sodium had leached out of the sample, and that about 15% of it was probably in the form of a sodium metasilicate solution, the rest presumed to be a mixture of sodium hydroxide and sodium carbonate.

(12) TABLE-US-00003 Chemical analysis of the aqueous solution Chemical species Concentration [mg/l] SiO.sub.2 466 Al.sub.2O.sub.3 6.65 Fe.sub.2O.sub.3 1.78 CaO 0.51 MgO 1.51 K.sub.2O 2.40 Na.sub.2O 6920 SO.sub.3 2.53 P 0.58

EXAMPLE 5

(13) The powdered product of Example 2 was mixed with water to give slurries (suspensions) with various solids concentrations, in some cases with the addition of various soluble salts to the initial aqueous solution. A sample of slurry was put in a closed reactor with a total volume of 1.65 liters, maintained at 35 C. and agitated with a mechanical stirrer operating at 500 rpm. The reactor was first evacuated and then filled with pure gaseous CO.sub.2 up to one atmosphere pressure. The pressure of the gas, which decreased with time due to its absorption by the slurry, was recorded continuously. Whenever the pressure reached a relatively constant value, further CO.sub.2 was added again to bring it back to one atmosphere. By following the change of pressure with time between refills, it was possible to estimate the total amount of CO.sub.2 consumed by reaction with the slurry. Results for a series of such experiments are summarized in the following Table.

(14) TABLE-US-00004 CO2 Total captured, CO2 Slurry Slurry Total as % of Experiment captured volume conc. solids. theoretical Slurry Temp. Duration Solid products n (mol) (mL) (g/L) (g) maximum additives ( C.) (h) detected 1 0.0641 260 50 13.0 35 Na2CO3: 35 24 Hydromagne 21.15 g/L site 2 0.0724 260 50 13.0 40 Na2CO3: 35 24 Hydromagne 53.85 g/L site 3 0.1363 260 50 13.0 75 NaHCO3: 35 30 Hydromagne 53.85 g/L site 4 0.0352 260 11.5 2.99 84 35 26 Artinite

(15) The theoretical maximum CO.sub.2 capture was calculated on the assumption that all of the carbonatable solids, expressed in terms of MgO and Na.sub.2O in the solids, would carbonate to give MgCO.sub.3 and Na.sub.2CO.sub.3, respectively, irrespective of any slurry additives present. However, it was observed that the main solid products were usually hydromagnesite and, in one case, artinite, which would imply a slightly lower amount of CO.sub.2 capture than the maximum theoretical value. In experiment 4, during which no slurry additives were used (i.e. pure water was used to make the slurry) the amount of CO.sub.2 captured, at 84% of theoretical, is actually slightly more than would be expected if artinite were the main magnesium carbonate formed, so it is likely that other carbonates were also formed but not detected. In any case, the result of experiment 4 shows that it is possible to essentially fully carbonate an aqueous suspension of the reaction product of Example 2 in about one day at atmospheric pressure. (Note also that the duration of these experiments was probably longer than necessary because they had to be left overnight unattended, during which time no additional CO.sub.2 could be added to bring the pressure back up. If one atmosphere pressure of CO.sub.2 had been maintained continuously, the reaction times would probably have been significantly shorter).

EXAMPLE 6

(16) A crushed sample of serpentine from Horsmanaho, Finland (ground in a ball mill to a powder with 43% passing a 75 micrometer sieve, and containing, by mass, 37.9% silicon expressed as SiO.sub.2, 38.7% magnesium expressed as MgO, 7.4% iron expressed as Fe.sub.2O.sub.3, and with an ignition loss of 14.9% at 950 C.), was mixed with anhydrous sodium carbonate powder in a 5355:4645 mass ratio (molar ratio approximately 1:1.3). About 5 kg of the mixed powder was pressed into a steel crucible and calcined in a lab furnace at 950 C. for 4.5 hours. The mass loss during calcination was 27.9%. From the measured mass loss it can be estimated that the decarbonation reaction was complete. The reaction product was analysed by XRF spectrometry and shown to contain 27.8% silicon expressed as SiO.sub.2, 30.4% magnesium expressed as MgO, 5.7% iron expressed as Fe.sub.2O.sub.3, and 33.4% sodium expressed as Na.sub.2O. An XRD analysis showed the major phases present in the product to be Na.sub.2MgSiO.sub.4 and periclase (MgO).

EXAMPLE 7

(17) The reaction products of Example 2 (referred to hereinafter as product X) and of Example 6 (referred to hereinafter as product XS) were ground to powders, and the finenesses of the resulting powders were measured using the Blaine Specific Surface Area (BSS) method. For each sample of product X or XS, 75 g of the solid were added to 1.5 liters of deionized water in a glass reactor equipped with a stirrer with a helicoidal Teflon paddle operating at 500 rpm. Pure CO.sub.2 gas was bubbled continuously through the agitated suspension (via a porous glass frit at the bottom) at a flow rate of 12 normal liters per hour at close to one atmosphere absolute pressure. The reactor contents were maintained at 70 C. by a jacket heated by circulating hot water. After various periods of time, samples of the suspension were taken to assess the progress of the carbonation reaction. The samples were filtered and the liquid filtrates were analysed for dissolved elements by ICP. The solid filter-cakes were dried at 110 C. and then analysed by XRF spectrometry for elemental composition, by XRD for qualitative phase composition, and by thermal analysis coupled with evolved gas analysis for the quantitative detection of combined CO.sub.2 and water. In order to calculate the amount of magnesium that had reacted, it was assumed that all of the CO.sub.2 in the dried filter-cake was present in the form of hydromagnesite.

(18) The results for three different product samples are given in the following Table

(19) TABLE-US-00005 Blaine Estimated Mg in Estimated specific Main Total Mg Mg in aqueous degree of surface crystalline content of CO2 hydro- phase reaction of Experiment area of phases dried filter- content magnesite expressed MgO in N; anhydrous detected in cake, of dried expressed as % MgO anhydrous anhydrous Duration product, dried filter- expressed filter- as % MgO relative to product, product used (h) m.sup.2/kg cake as MgO, % cake, % in filter-cake filter-cake % 5 3.5 420 Hydro- 31.2 15.9 18.2 0.6 59 Product X magnesite 6.sup.1 4.0 600 Hydro- 30.3 16.4 18.8 0.5 63 Product X magnesite 7 4.0 205 Hydro- 27.7 17.2 19.7 0.4 71 Product magnesite XS .sup.1In experiment 6, a high-power agitation system was used instead of the normal stirrer

(20) It can be seen from the above results that it is possible to carbonate products X and XS at only one atmosphere pressure in aqueous suspension and obtain conversion yields of the order of 60-70% of the total magnesium in the starting material in about 3.5 to 4 hours. Based on XRF analyses of the solid phases coupled with ICP analyses of the liquid phase, it is estimated that roughly 90% of the initial Na and 50% of the initial Cr in the product leached out during the experiment.

EXAMPLE 8

Formation of Eitelite

(21) The procedure of Example 7 was repeated using two slurries (prepared from products X and XS from Examples 2 and 6, respectively) at a concentration of 150 g/L, i.e. three times the concentration used in Example 7.

(22) During the 3 first hours the same trends were observed as those in Example 7. But after 4 hours, XRD revealed the presence of eitelite as well as hydromagnesite. The results are given in the following Table.

(23) TABLE-US-00006 Blaine Mg in Estimated specific Main Total Mg Estimated aqueous degree of surface area crystalline content of CO2 Mg in hydro- phase reaction of Experiment of phases dried filter- content magnesite expressed MgO in N; anhydrous detected cake, of dried expressed as % MgO anhydrous anhydrous Duration product, in dried expressed filter- as % MgO in relative to product, product used (h) m.sup.2/kg filter-cake as MgO, % cake, % filter-cake filter-cake % 8 4.0 600 Hydro- 28.06 16.45 18.83 0.5 67.12 Product X magnesite & eitelite 9 4.0 600 Hydro- 25.74 15.82 18.11 0.8 70.37 Product magnesite XS & eitelite

EXAMPLE 9

Effectiveness of Water Leaching

(24) 50 g of product X was washed in 1 L deionised water 6 times in series (15 minutes stirring in between at room temperature and atmospheric pressure, i.e. 25 C. and 1 bar). Each time a small sample was taken to analyse the solids and the liquids. The XRF results for product X coupled with ICP solution analyses lead to the results given in the following Table, which the percentages of each element leached from the product are given as a function of the number of washing steps.

(25) TABLE-US-00007 % leached Si Ca Mg K Na S Cr Washing 1 9.85 0.46 0.00 19.76 36.44 13.77 46.12 Washing 2 3.20 0.09 0.00 1.94 8.69 0.60 1.30 Washing 3 1.41 0.02 0.00 0.25 3.47 0.04 0.21 Washing 4 1.22 0.02 0.00 0.03 3.53 0.13 0.32 Washing 5 0.75 0.02 0.00 0.02 1.87 0.14 0.12 Washing 6 0.51 0.05 0.00 0.00 1.30 0.05 0.08 TOTAL 16.94 0.66 0.01 22.00 55.31 14.73 48.15

(26) The first washing step was clearly the most efficient. Therefore 500 g of product X were washed in the same conditions, the solid was then filtered and dried overnight at 110 C. Finally the same carbonation experiment as described in Example 8 was performed on this solid product. This time XRD analysis revealed only hydromagnesite; no eitelite was detected. This shows that washing can be used to enhance Na recycling in the process.

(27) The analysis of the solution obtained after the first washing (table below) showed it to have a high pH and to contain about 4000 mg/L of Na and 637 mg/L of Si, the other elements being present in much smaller amounts. On this basis, the solution was estimated to contain about 23 milimoles/liter of sodium metasilicate (Na.sub.2SiO.sub.3) and 130 millimoles/liter of sodium hydroxide (NaOH), possibly also including some carbonate ions.

(28) TABLE-US-00008 In mg/l (elements) Si Ca Mg K Na S Cr Washing 1 637.5 0.299 0.267 4.1 4000 1.38 28.4

EXAMPLE 10

Hydromagnesite Formation at 60 C.

(29) 1 L of a 50 g/L slurry of product X was prepared and poured into a 2 L autoclave. The system was closed and 1 L of pure CO.sub.2 at 10 bars was added without purging the residual air, after which the slurry was stirred and heated up to 60 C. (which roughly corresponds to the dew point of exhaust gases in a cement plant). After 2 hours the pressure dropped to close to atmospheric and heating of the autoclave was stopped. The next day, the slurry was filtered and dried overnight at 110 C. XRD revealed hydromagnesite as the main crystalline product.

EXAMPLE 11

Formation of Magnesite as Main Product at 120 C.

(30) The procedure of Example 10 was repeated but at 120 C. (a typical temperature for exhaust gases from a cement plant) using the same autoclave, and adding CO.sub.2 each time the pressure dropped close to two bars (the equilibrium water vapour pressure at 120 C.). Three additions of CO.sub.2 up to 10 bars were made in the space of one day. XRD on the dried solid revealed magnesite (MgCO.sub.3) to be the main product, but also showed some traces of magnetite (Fe.sub.3O.sub.4). ICP analysis of the aqueous phase showed 90% leaching of Na. The combined CO.sub.2 content in the solids was analysed by means of a high-frequency induction furnace coupled to a Horiba EMIA-820V gas analyser and showed that the amount of MgCO.sub.3 present accounted for about 47% of the original Mg in Product X.

EXAMPLE 12

Selective Separation of Chromium by Reduction in Solution

(31) Powdered product X was stirred with deionised water for 15 minutes at a 2:1 Water:solids mass ratio. The yellow-coloured aqueous phase was filtered and a sample taken for ICP analysis. Excess ferrous sulfate (FeSO.sub.4), a reducing agent, was then added in powder form to the solution, after which a chromium-containing precipitate formed. The liquid was again filtered, giving a colourless solution which was again analysed by ICP. The chromium content of the yellow-coloured aqueous phase was 772 mg/L. The colourless solution contained only 29 mg/L of chromium.

(32) These results, when compared with the analysis of the raw materials, indicated that about 50% of the total chromium in the original raw material (olivine) was converted into a readily soluble (chromate) form by the process used to make product X, and 96% of the Cr leached into the solution was precipitated by addition of ferrous sulfate.

EXAMPLE 13

Concentration of Hydromagnesite from the Solid Residues by Flotation

(33) Separation tests were performed by flotation on the carbonated products produced from the application of the process to olivine and serpentine, similarly to the products shown, respectively, for experiments 5 and 7 in the table of Example 7. Combinations of sodium oleate, carboxymethylcellulose (CMC) and methyl isobutyl carbinol (MIBC) were used in the aqueous phase sequentially to disperse the solids in the form of a slurry at a solids concentration of 90 g/L.

(34) First sodium oleate was added to the slurry and stirred for 5 minutes in a beaker in order to render carbonated particles hydrophobic. Then CMC was added to depress silicate hydrophobicity, followed by a further 5 minutes of stirring; and finally MIBC was added in order to stabilise the foam formed by air bubbling.

(35) Air was bubbled through the treated slurry in a miniature flotation cell, and the solids carried over by the foam were collected as concentrate. The residual solids were collected as tailings. Results of four such experiments are presented in the table below. The CO.sub.2 content was analysed by means of a high-frequency induction furnace coupled to a Horiba EMIA-820V gas analyser. The CO.sub.2 contents of the concentrates were typically 3-4 times greater than those of the tailings, showing that hydromagnesite can be effectively separated by flotation in this manner.

(36) The XRD results confirm the separation. They are expressed in a qualitative way by different symbols expressing the probability of a phase presence: (o) not present, (*) possibly present, (X) definitely present, (X+) present in abundance.

(37) TABLE-US-00009 Chemical additive (dosage, ppm) Test 1 Test 2 Test 3 Test 4 Sodium oleate 1500 1000 2000 1000 CMC 200 200 200 200 MIBC 20 20 20 20 Product tested by flotation: Carbonated product X Carbonated product X Carbonated product X Carbonated product XS Results of flotation tests Tailings Concentrate Tailings Concentrate Tailings Concentrate Tailings Concentrate Components SiO2 % 34.1 17.6 33.7 18.5 33.8 22.0 33.3 16.9 measured by Fe2O3 % 6.4 3.4 6.4 3.6 6.4 4.3 6.9 3.0 chemical MgO % 34.6 37.2 34.5 37.2 33.9 36.9 29.2 35.9 analysis Na2O % 2.3 1.3 2.3 1.4 2.6 1.3 2.3 1.2 LOI % 22.3 40.0 22.6 38.9 22.7 35.1 26.2 41.6 CO2 % 5.7 22.2 5.8 20.2 5.7 17.3 9.4 23.0 Phases Sodium 0 0 0 0 0 0 0 0 detected by magnesium XRD silicate Periclase X * X * X * * * Olivine X * X * X * 0 0 Hydromagnesite X X + X X + X X + X X +

EXAMPLE 14

(38) Ground natural talc from Luzenac (France) with nominal composition Mg.sub.3Si.sub.4O.sub.10(OH).sub.2 and containing some minor impurities (1.1% Al.sub.2O.sub.3, 0.9% Fe.sub.2O.sub.3 and 0.9% CaO by mass) was mixed with anhydrous sodium carbonate (Na.sub.2CO.sub.3) at a 4863:5136 mass ratio (molar ratio approximately 1:1). The composition was chosen in order to obtain an atomic ratio of 2:1 Na:Si in the final sample. 20.5 g of this powder was pressed into a pellet and placed into a lab furnace in a platinum crucible. The sample was heated at 900 C. for 1 hour, followed by cooling in air. It was weighed before and after treatment and the measured loss on ignition of 23.3% was consistent with evaporation of carbon dioxide from the sodium carbonate plus bound water from the talc; it represents about 95% of the theoretical value for complete reaction of 24.6%. The phase constitution of the reacted sample was determined by X-ray diffraction: the main products detected were Na.sub.2MgSiO.sub.4 and Na.sub.2SiO.sub.3.

EXAMPLE 15

(39) A sample of Product X was leached in water following a procedure similar to that given in example 9. After drying at 105 C., the powdered material (which had an ignition loss of 8.9%), was analyzed by X-ray fluorescence for its major elements, and found to contain 28.1% SiO.sub.2, 39.1% MgO and 17.9% Na.sub.2O. An X-ray diffraction analysis showed that the main crystalline compounds present in the powder were periclase (MgO), sodium magnesium orthosilicate (Na.sub.2MgSiO.sub.4) and forsterite olivine (Mg.sub.2SiO.sub.4); but it was known also to contain amorphous magnesium silicate hydrates (M-SH). 5 parts of this material were mixed manually with 1 part of deionised water in a rubber bowl, using spatula. 7 g aliquots of the resulting paste were compressed in a cylindrical mould at a load of 3 tonnes to give cylindrical pellets 19 mm in diameter and 10 mm in height. These pellets were subject to curing in a flow of pure CO.sub.2 gas at atmospheric pressure in a chamber at 202 C. Two different humidity conditions were tested: dry (i.e. no water added to the gas stream); and wet (in which the gas stream was bubbled through water at the bottom of the curing chamber before passing over the pellets). The uptake of CO.sub.2 and/or water by the pellets was followed by taking them out quickly and weighing them once a day. The experiment was stopped after one week, as the weight increases had begun to level off. At this point, the pellets carbonated under dry conditions had gained 3.0% and all four of the pellets tested gave essentially identical weight changes. On the other hand, the pellets carbonated under wet conditions (close to 100% relative humidity) showed a wider pellet-to pellet variation in weight increase, with a mean of 7.8% and a standard deviation of about 1%.

(40) It was notable that all of the pellets that had been carbonated under one atmosphere of CO.sub.2 became superficially much harder than companion pellets that had simply been stored in air. The pellets carbonated under humid conditions also showed a considerable amount of efflorescence. A sample of this efflorescence was scraped off and analyzed by X-ray diffraction. It was found to contain nesquehonite (MgCO.sub.3.3H.sub.2O), nahcolite (NaHCO.sub.3), trona (Na.sub.3H(CO.sub.3).sub.2.2H.sub.2O), and sodium carbonate mono-hydrate (Na.sub.2CO.sub.3.H.sub.2O). The wet-carbonated pellets themselves, analyzed by the same technique, showed the presence of primarily of periclase (MgO), sodium magnesium orthosilicate (Na.sub.2MgSiO.sub.4), olivine (Mg.sub.2SiO.sub.4) and nesquehonite (MgCO.sub.3.3H.sub.2O). It thus appears that nesquehonite was the main binder phase and that it was probably produced to a significant extent by carbonation of M-SH.

(41) Pairs of treated and untreated pellets were compressed to failure in a compression machine. The results are summarized in the table below:

(42) TABLE-US-00010 Compressive loads at failure, Curing regime applied to pressed pellets kN, (for 2 pellets) Stored in air 3.0; 3.2 Carbonated under humid conditions 11.5; 18.5 Carbonated under dry conditions 36.0; 36.7

(43) It is clear that the atmospheric-pressure carbonation process greatly increased the strength of the pellets, and that carbonation under dry conditions was preferable to carbonation under humid conditions.