Background Information Secondary gypsum, or"Red Gypsum"as it is sometimes called, is typically comprised predominantly of calcium sulfate in varying states of hydration, along with oxides of iron in an amount varying from about 3-35%, and various trace elements. One method by which red gypsum is produced is as a by-product in the manufacture of titanium dioxide pigment via the well-known sulphate process, in which it is precipitated from acidic solution filtrates.

In a general sense, red gypsum is regarded by those skilled in the art as an industrial waste material which must be discarded, recycled, or otherwise disposed of.

Prior art methods of its disposal include burying the material in a landfill, or processing it in a refining operation to yield a purified gypsum product. However, further processing is generally undesirable from an economic standpoint, since the costs of refining are scarcely offset by the inherent value of the final refined product. Thus, red gypsum had an alternate use which did not require its further processing, such use would provide a vehicle for its disposal in a beneficial manner.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a plot of shear strength at various depths of OPC and a composition according to the invention ; FIG. 2 is a plot of strength vs. time for various mixtures according to the present invention; FIG. 3 is a plot of shear strength vs. pH of a mixture according to the present invention used in stabilizing soil; FIG. 4 is a plot of shear strength vs. depth of a mixture according to the present invention used in stabilizing soil; FIG. 5 is a plot of shear stress vs. strain at failure of a mixture according to the present invention; FIG. 6 is a plot of shear strength vs. depth of a mixture according to the present invention; FIG. 7 is a plot of shear strength vs. pH of a mixture according to the present invention;

FIG. 8 is a plot of shear strength vs. strain at failure of a mixture according to the present invention; FIG. 9 is a plot of shear strength against depth of a mixture according to the invention; FIG. 10 is a plot of shear strength vs. pH of a mixture according to the invention; FIG. 11 is a plot of shear strength vs. strain at failure of a mixture according to the present invention; FIG. 12 is a plot of shear strength against depth of a mixture according to the invention; FIG. 13 is a plot of shear strength vs. pH of a mixture according to the invention; FIG. 14 is a plot of shear strength vs. strain at failure of a mixture according to the present invention; FIG. 15 is a plot of shear strength against depth of a mixture according to the invention; FIG. 16 is a plot of shear strength vs. pH of a mixture according to the invention;

FIG. 17 is a plot of shear strength vs. strain at failure of a mixture according to the present invention; FIG. 18 is a plot of shear strength against depth for all boreholes tested with mixtures according to the invention; FIG. 19 is a plot of the average failure strains of mixtures used in various boreholes filled with mixtures according to the invention; FIG. 20 is a comparative plot of shear strength results from SCPT tests and laboratory tests; and FIG. 21 is a plot of shear strength vs. pH of samples from various boreholes in which mixtures according to the invention were evaluated.

Detailed Description A typical analysis of red gypsum suitable for use according to the present invention is as follows: Moisture 38. 5% pH 6.8 CaS04*2H20 43.0% Fe (OH) 3 16.9% Mn 0.41% Al 0.16% Sb <0. 1 ppm As 15 ppm Cd <0. 1 ppm Cr 71 ppm Co 4 ppm Cu 5 ppm Pb 6 ppm Mg 0.10% Hg <0. 1 ppm Ni 6 ppm Se <0. 1 ppm Na 217 ppm V 240 ppm Zn 111 ppm Till. 0.12% Si02 0.52% However, a wide variety of red gypsum products obtained as a by-product of titanium dioxide precipitation is suitable for use in a combination according to the invention.

Specification limits on red gypsum are: gypsum (CaSO4. 2H20 60-90%), iron (Ip and iron (E) oxide/hydroxides (expressed as Fe (OH) 3 3-35%) and main impurities Ti02 0.3-2%, Si02 0.1-4%, Na2O 0.1-0. 8%, CaC03 0.1-5%, MgO 0.1-0. 4%, K20 0. 1- 0.2%, P205 0.1-0. 2%.

When red gypsum is produced, the iron is in the ferrous state, but rapidly oxidizes (on exposed surfaces) to ferric when the pH is >6. When mixed with slag and lime, it rapidly oxidizes to ferric and so in a final composition according to one preferred form of the invention, the iron is present as ferric oxide/hydroxides. The iron particles are much smaller than the gypsum crystals, and tend to coat the gypsum crystals, which arrangement is suspected of being at least partially responsible for the unexpected behaviors we have observed. The properties of red gypsum are compared to those of ordinary white gypsum in table 1 below:

Property White Gypsum Red Gypsum pH 7. 4 7.4 Free Moisture (%) 10-17 10-50 Particle Density (mg/m) 3.05 2.71 Dry Density (mg/m) 1. 21 1. 21 Erodability (Dispersiveness) non-dispersive 2 dispersive 4 Liquid Limit 58 105 Plastic Limit (%) non-plastic non-plastic Californian Bearing Ratio (CBR % not obtainable 23.05 Optimum Moisture Content % 39 41 Consolidation Testing (50-400 kN/m) Coefficient of Volume Change (Mv m/MN) 3.643-0. 32 0.907-0. 179 Coefficient of Consolidation (Cv m/year) 1.19-0. 102 0.855-0. 232 Coefficient of Secondary Compression (Cα 0.05-0. 02 0.004-0. 002 m2/year) Compression Index Cc 0.218 0.242 Swell Index Cs 0.08 0.131 Permeability (kv X 10-m/s) 1. 043-6.92 3.314-6. 92 Undrained Shear Strength (Cu kN/m2) 95-104 194-355 Strain to Failure (%) 5 10 Table I-properties of white and red gypsum We have found that red gypsum has considerable strength, stiffness, and low permeability so as to enable it to be used as a civil engineering material. We have also found red gypsum to be useful when combined with other materials such as natural soils, clay minerals, and other materials such as ground granulated blast furnace slag (GGBS), BOS slag, or pulverized fuel ash (PFA or fly ash). We have found that addition of red gypsum to these materials improves their properties for use in civil engineering applications.

We have also found that red gypsum reduces the plasticity of clay soils, and that addition of red gypsum to clay soils increases the plastic limit, improving

"workability". The addition of red gypsum to clay soils also reduces their swell potential and reduces frost susceptibility.

We have also found that the strength of soils in general is maintained at much higher moisture content on addition of red gypsum, which we believe is responsible at least in part to the significant increase in strain to failure of soils and PFA which we have observed. Addition of Gypsum to PFA and GGBS also significantly improves undrained shear strength.

We have found red gypsum to be a relatively impermeable material.

Permeability of London Clay and Glacial Till is significantly reduced on addition of red gypsum. Addition of red gypsum also reduces the permeability of PFA and GGBS to values similar to that of clay soils. The addition of red gypsum to PFA and GGBS provides a decrease in permeability and erodability of materials.

Red gypsum was used in a deep soil mixing trial to strengthen the natural soil to allow it to be used for foundations for a railway line. The Trial Site was 2 valleys filled with weak peat and clay soils to a depth of 4-5m covering hard sand beds.

Formerly, the preferred method of treatment was to strengthen the weak soil with cement piles along length of railway. The cement was injected at 200kg/m3 of soil, to form 4-5m deep piles. A red gypsum/GGBS mix was used in place of cement. Trial with RG (red gypsum) powder pneumatically mixed with GGBS at 25: 75 and 75: 25 blends @ 200 and 250kg/m3 soil. 27 columns were installed on site to assess the strength gain of several mix ratios of GGBS/red gypsum. The strength of the columns was assessed after 7 and 56 days. Starting soil strength is 5-15 kPa, minimum requirement is 100kPa. After 56 days the columns were to be re-tested, however all

were found to be too hard to penetrate. Core samples were taken from selected columns and were tested for strength vs depth and mineralogy.

For purposes of this invention and the appended claims the word"soil"means any dirt composition, and includes without limitation kaolin, bentonite, mordenites, other clays, London Clay, Glacial Till, slurries, pulverized fuel ash, and ground granulated blast furnace slag.

Red gypsum may be present in a treated soil in accordance with the invention in any amount between about 2.00 % to 75.00 % by weight based upon the total weight of the final treated soil, including every hundredth percentage therebetween.

While the actual most preferred amount of red gypsum to be added to a particular soil will depend upon the makeup of the particular soil to be treated and its intended use, in general, in most cases it is preferable to have red gypsum present in any amount between about 18.00 % to 65.00 % by weight based upon the total weight of the final treated soil, including every hundredth percentage therebetween. The red gypsum so employed may contain calcium sulfate present as any of the various hydrated stages possible for calcium sulfate.

The following is a list of several uses for red gypsum in civil engineering applications: soil conditioning for temporary site works; reduction of swell potential of soil under pipes and footings ; treatment of waste slurries to ease handling problems; treatment of waste slurries to permit their use as an engineered fill; treatment of a'pumpable'fill for use in conjunction with geosynthetics and tunnel linings; modification of indigenous soils for use as landfill liners; modification of soils for use as structural fill; modification of soils for use as an impermeable barrier;

soil stabilization for surface applications such as pavement subgrade improvement ; soil stabilization utilising deep mixing techniques; and slope stabilization using'piles' (augured holes filled with additive and capped).

According to another aspect of the present invention, we have discovered a novel binder which comprises varying ratios of red gypsum ("RG") and ground granulated blast furnace slag ("GGBS"). A deep soil mixing field trial was conducted which yielded excellent results in that two of the RG/GGBS mixes were comparable (mix C) or stronger (mix D) than the cement mix (mix E). The results also illustrate that the pH is very important for strength development, and that the pH should be greater than about 10.5, which may be reasonably interpreted to mean that a pozzolanic reaction is taking place. The strength of mixes A and B (75% RG/25% GGBS binder ratio) were poor in comparison to mixes C and D (25% RG/75% GGBS binder ratio), probably because the pH was lower. The pH would be lower in mixes A and B as RG has a lower pH than GGBS, and mixing with the soil would reduce the pH even lower. This is likely because earlier lab work combining 60% RG/40% GGBS on their own also shows a large strength increase. Mix Strength (N/mm) Days 0 2 7 28 33 20% RG Powder/80% GGBS 0 0 2 39 42 40% RG Powder/60% GGBS 0 0 0 22 29 20% RG Cake/80% GGBS 0 0 3 38 43 30% RG Cake/80% GGBS 0 0 0 35 38 50% OPC/50% GGBS 0 15 28 55 100% OPC 0 31 46 59 90% GGBS/10% lime 0 0 2 12 Table II-strengths of various mixtures, some according to the invention

Five rotary boreholes ("BH") were drilled from a platform layer 1.5 meters above the top of the columns to a depth of 7. 5m (2m below column base). The cores were 1.5m long and had a diameter of 100mm. Once drilled, all samples were placed in sealed plastic tubing before being delivered to the testing facility. A list of the identification and composition of the samples taken is shown below.

BH 83 200 kg/m3 concentration, 25% Slag, 75% Gypsum (Mix A) BH 79 250 kg/m3 concentration, 25% Slag, 75% Gypsum (Mix B) BH 75 200 kg/m3 concentration, 75% Slag, 25% Gypsum (Mix C) BH 71 250 kg/m3 concentration, 75% Slag, 25% Gypsum (Mix D) BH 69 200 kg/m3 concentration, 100% cement (Mix E) Drilling records show that the percentage recovery was quite low for all boreholes, particularly below 3m. This can be attributed to sandy material being present at lower depths and the columns being 4m in length.

The three main testing methods used in this investigation were the quick undrained tri-axial compression test, pH test, and X-ray diffraction test. The quick undrained triaxial compression test is fully described in the British Standard, BS 1377: Part 7: 1990. The samples tested had dimensions of 100mm in diameter and 200mm in length. This size was chosen because the diameter of the cores was approximately 100mm. Reducing the diameter of the core could have caused damage to its structure thus influencing measured strengths and strains. Also using larger samples in more representative of the soil mass. The sample is compressed at a constant rate (in this case 1. 5mm/min) whilst under an all round confining pressure. The confining pressure used in these tests was 100kPa. As the axial load is applied, the sample suffers continuous deformation. Throughout the tests, until the sample fails, readings of stress and strain are made at regular intervals. If the sample continues to deform without failing, then the value of stress at 20% strain is recorded as the failure stress.

The stress at failure can be converted into a value peak undrained shear strength (Cu). This is measured in kilo Newtons per meter squared or kPa. All samples tested were then submitted to pH and x-ray diffraction testing.

The pH of the samples was determined by the method outlined in BS 1377 Part 3: 1990. The method employed was the electrometric method, which gives a direct reading of the pH value of a soil suspension in water. The equipment used was a Jenway pH meter (model 3150). Samples were prepared by drying material then mixing 30g with 75ml of distilled water and left to mix for at least 8 hours.

X-ray diffraction is a technique for the study of crystal structure. The basic principal behind x-ray diffraction is that when a beam of X-rays meets a crystalline solid, the x-rays interact with the solid and the beams are scattered. From the pattern of scattering one can infer the pattern of distribution of electronic charge in the crystal and hence the nature of the crystal structure. The essential feature of the diffraction of waves of any wavelength is that the distance between scattering centres be about the same as the wavelength of the waves being scattered. The dimensions of x-rays and the spacing between atoms in crystals meet these conditions (Moore and Reynolds, 1997).

No standard laboratory tests were available to measure the durability of these samples so a new test was developed for the purpose. The philosophy behind the test was to simulate a realistic but worst-case scenario of conditions that may occur in the field, in this case columns being soaked in a large ground water supply. Samples measuring 50mm length by 100mm diameter were cut from the sample cores and completely immersed in 1.5 litres of distilled water. The pH and conductivity of the water was then measured at intervals, the physical condition of the samples was monitored and photographs taken. It was also intended to measure the strength of the samples using a pocket penetrometer but the samples proved to be too brittle.

It is hoped that measuring the pH and conductivity of the samples over time will indicate what level of interaction is occurring between the sample and the water, with the thought being that excessively high pH or conductivity should be indicative of sample degradation.

Results The results of triaxial and pH testing are presented here. Each borehole is considered individually with attention focussed on shear strength-depth, shear strength-pH and shear strength-strain at failure relationships.

Figure 1 shows a graphical comparison of the shear strength of OPC and a mixture according to the invention and illustrates similar behaviour of these materials in the test. Figure 2 is a plot of strength vs. time for various mixtures according to the present invention which shows that mixtures of the present invention eventually reach strengths very near those of ordinary cement, which makes them in many cases useful as a replacement for cement.

Borehole ("BH") 83 was filled with Mix A, which is a mixture of 25% slag, 75% red gypsum at a concentration of 200kg/m3. Figure 3, shows the Shear strength versus pH graph of BH 83. In Figure 4 is shown a graphical representation of shear strength versus depth of BH 83. Figure 5 shows graphically the shear stress versus strain at failure for BH 83. The shear strength against pH graph (Figure 3) shows no distinct relationship between shear strength and pH. This is surprising as it is recognised that the acidity of the soil is important to bring alkali and silica into solution, which is necessary for cementitious reactions to take place. The shear strength against depth graph (Figure

4) shows an initial Cu of 75 kPa dropping to 30 kPa at 2. 1m, coincident with a weak zone in the peat layer encountered in borehole SA6453; this then increases to 225kPa at 4.4m. The shear strength against strain at failure graph (Figure 5) shows samples with higher strengths failed at lower strains. All failure strains fall with the range 2. 5-14%.

Borehole 79 was filled with Mix B, which is a mixture of 75% gypsum, 25% slag at 250kg/m3 concentration. Figure 6 shows the shear strength versus depth graph of BH 79. In Figure 7 is shown a graphical representation of shear strength versus pH of BH 79. Figure 8 shows graphically the shear stress versus strain at failure for BH 79.

The shear strength against depth graph (Figure 6) shows the shear strength at the top of the column is quite low, 23 kPa. It then ranges between 70 and 94 kPa from 1. Om to 4. 2m.

Again a distinct reduction in strength occurred at-2. 0m depth, coincident with a weak zone detected by exploratory boreholes. The shear strength against pH graph (Figure 7) shows no clear relationship between Cu and pH. All samples have a pH of greater than 9.3. Initially there does not appear to be a relationship between shear strength and strain at failure (see Figure 8). However, if the test conducted at 0.2m is discounted then a pattern of reducing strains at failure with increasing strength can be seen. The sample taken at 0.2m is likely to have been topsoil whereas the rest of the samples were taken from the peat. Failure strains vary between 5 and 15%.

Borehole 75 was filled with Mix C, which is a mixture of 75% gypsum, 25% slag at 200kg/m3 concentration. Figure 9 shows the shear strength versus depth graph of BH 75. In Figure 10 is shown a graphical representation of shear strength versus pH of BH 75. Figure 11 shows graphically the shear stress versus strain at failure for BH 75. Figure 9 shows the shear strength at lm is 175 kPa but the sample taken 1.4m exhibits a much lower shear strength of 30 kPa, which rose to 830kPa at a depth of 2. 0m. Samples

tested from depths 2.8 and 3.9m exhibit shear strengths of over 1060 kPa. At 4.3m the Cu reduced to just 7 kPa but this may have this may been below the base of the column. There appears to be a positive relationship pH and shear strength. However, all pH values lie between 9.2 and 11 (see Figure 10). The shear strength against failure strain graph (Figure 11) shows samples with a higher shear strength exhibited lower strains at failure. The three strongest samples have failure strains at 1.5% or below.

Borehole 71 was filled with Mix D, which is a mixture of 25% gypsum, 75% slag at 250kg/m3 concentration. Figure 12 shows the shear strength versus depth graph of BH 71 (note only the sample taken from 1. lm failed during test, the remaining 4 tests may be stronger than the graph suggests). In Figure 13 is shown a graphical representation of shear strength versus pH of BH 71. Figure 14 shows graphically the shear stress versus strain at failure for BH 71. As can be seen on Figure 12 all but one of the samples tested from BH 71 was too strong for the testing equipment.

The one sample that did fail had a shear strength of 1275kPa. The four samples that did not fail have shear strengths in excess of 1340kPa. There is no clear relationship between pH and shear strength shown on the graph (Figure 13), although all samples had pH values in excess of 10.75. The only sample to failed so at a strain of 1.3%. The other samples were at less than 1% strain when the tests were terminated (see Figure 14). Low strains to failure indicate that the sample is very brittle, brittle samples may have initial high strengths but when a shear plane forms the shear strength reduces to a much lower residual value.

Borehole 69 was filled with Mix E, which is pure portland cement at 200kg/m3 concentration. Figure 15 shows the shear strength versus depth graph of BH 69 (the sample taken at 4. 1m did not fail, the shear strength may therefore be higher). In Figure 16 is shown a graphical representation of shear strength versus pH

of BH 69. Figure 17 shows graphically the shear stress versus strain at failure for BH 69 (the sample taken at 4. 1m with a shear strength of 1343kPa did not fail, its shear strength and strain at failure may be higher than represented on the graph). The 3 samples between 0. 5m and 2. 1m had shear strength between 400 and 830 kPa, the sample at 3. 1m did not fail during the test and had a shear strength in excess of 1300 kPa (Figure 15). The pH against depth graph shows there is a positive relationship between shear strength and pH. All pH values were between 11.1 and 12.1 (Figure 16). The samples that failed during the tests show that with increased shear strengths the strain at failure reduces. The samples that did not fail were under 0.5% strain when the tests were terminated (Figure 17).

Figure 18 shows a combined graph of shear strength against depth, for all boreholes tested. Figure 19, shows average failure strain for different boreholes tested.

Laboratory work was conducted to examine the strength and durability of peat when mixed with different mixture and quantities of binder material.

It can be seen from figure 18 that mix A and B (75% Red Gypsum/25 % slag @ 200 kg/m3 and 250 kg/n3 concentrations) exhibited significantly lower shear strengths than the other samples tested. Mixes C and E (75% slag/25% red gypsum at 200 and 250 kg/m3 respectively) show considerable variation but have approximately the same average shear strengths. Mix D (75 % slag/25% red gypsum at 250 kg/m3 concentration) showed the highest shear strength. All mixes show troughs in shear strength between 0.9 and 2. 1m, corresponding with results from in-situ testing results and is possibly due to natural zones of weakness within the peat layer.

Figure 19 shows that mix A and B had higher average strains at failure with averages of 10.6% and 7.4% respectively. Mixes C and E averaged 5.8 and 4.0 % respectively. The average failure strain for mix D was less than 1.0%, but it should be remembered that only one of these samples reached failure before the test was terminated, and these failure strains may be considerably higher.

As can be seen in Figure 18, which is a comparison of shear strength results from SCPT tests and Laboratory tests, the shear strength of mixes C, D, and E are higher than those found during the in-situ testing (increases of 350%, 820% and 280% respectively), it would appear therefore, that strength continued to develop in the period between SCPT testing and recovery of the samples. In the other boreholes the shear strength remained the same, the average shear strength from BH 79 actually reduced by 50%.

All samples exhibited pH values between 8.5 and 12.1. Samples from mix A showed the lowest values of pH. The overall pH results do not indicate that samples with higher pH values have higher shear strengths but that high shear strengths are not reached unless the pH is above 10.5 to 11. This is not surprising as a pH above 10.5 is required to bring alkali and silica into solution in order to produce cementitious compounds (Simpson, 2001).

Boreholes 83 and 79 (mix A and B) had pH values below this range, they were also the weakest samples to be tested. The pH values recorded in this study confirm the conclusions drawn by Simpson (2001) that high shear strengths were not reached in the laboratory due to low initial pH values. Typical shear strengths of the laboratory mixed samples being between 20 and 80 kPa, with typical pH values being 8.5.

X-ray diffraction testing has not shown any evidence of thaumasite or ettringite in any samples.

Table II-VI below shows stress test results on samples having the indicated compositions using test method EN 197. Composition pH H2O Stress (N/m2) added % RG %GGBS 7 days 28 days 33 days 20 (pwd.) 80 10 245 g 2.3 38.7 41.5 40 (pwd.) 60 9.8 285 g 0. 0 22. 3 28.5 20 cake 80 10 196 2. 5 38.1 42.6 30 cake 70 9. 9 170g 0. 0 34. 8 37.9 Table II

The samples in table II show the strength at various times of mixtures comprising 20% and 40 % by weight of red gypsum powder (obtained by drying and grinding red gypsum filter cake produced in titanium dioxide manufacture via the sulfate process) in admixture with ground granulated blastfurnace slag (GGBS). The samples in table II also show the strengths at various times of mixtures comprising 20% and 30 % by weight of red gypsum filter cake in combination with GGBS as well. The measured pH of the mixtures is shown, as well as the amount of water added in each case. Composition pH H20 Stress (N/m2) added % RG (pwd.) % GGBS 7 days 28 days 56 days 20 80 12.3 275 21. 7 39.9 51.4 40 60 12.2 315 8. 8 34.2 39.9 60 40 12.1 350 2. 8 20.6 18.6 80 20 12.0 390 g 0 2.8 4.1 Table III-

The samples in Table III show the measured strengths of materials having the indicated compositions and containing the indicated amounts of water. However, as contrasted to those samples set forth in table II, the samples in table III contained 1.0% lime by weight, to adjust the pH. Composition pH H2O Stress (N/m2) added % RG (cake) % GGBS 7 days 28 days 56 days 20 80 12. 3 252 g 18. 8 38. 8 49.8 40 60 12.2 283 g 10. 7 33.8 38. 3 60 40 12.1 322 3. 5 18 20.0 80 20 12.0 364 g 1.6 4. 3 4.5 Table IV- Composition H2O added Stress (N/m2 7 days 28 days 56 days 50:50 GGBS:PC 225 g 29.6 54.6 - 100% PC 225 g 47.0 58.6 - 90: 10 GGBS: Lime 225 g 9.4 18.8 20.3 Table V- Composition Total H20 Slump PD Stress (N/m2) 2 days 7 days 28 days 40% Cake 5170 70 2245 0.5 5.7 26. 3 80% Cake 5785 55 2130 0.0 too soft 1. 1 Table VI-

The examples in table IV are the same as those in table III, except that the red gypsum used in the examples in table IV were made using red gypsum filter cake that was produced as a by-product from manufacture of titanium dioxide using the sulfate process, whereas those in table III were made using crushed and dried filter cake material.

From the data in the above tables, it can be seen that the reaction between the various components is probably pozzolanic (best strength occurs when the pH of the mix is >10.5 by adding 1% lime). These data show that the strength gain is slower on a time scale than that seen in ordinary Portland cement ("OPC"). However, surprisingly we have discovered that after 56 days, strengths similar to OPC can be obtained with a RG: GGBS: lime composition. The results also show that GGBS + lime alone does not reach the same strengths as RG: GGBS: lime. However some applications, e. g. decorative block paving, do not require high strength and so higher ratios of red gypsum are beneficial, and especially when a red color is required, which is the majority of products in this type of market.

Optical and electron microscopy shows a general fusing together of the individual RG, GGBS grains over time which will be one of the main contributors to strength gain. We have also determined that ettringite crystals (CasAl2 (S04) 3 (OH) 12 * 26H20, Hydrated Calcium Aluminium Sulphate Hydroxide) are formed in the mixture and these can be detected when the samples have been cured 28-56 days. It is well known that ettringite can provide strength when formed during the initial cure. In the mixtures of our invention we find it interesting, curious, and wholly unexpected that the ettringite is not found in the initial cure, but develops slowly over time, as opposed to the prior art. The only other phase detected is gypsum.

Consideration must be given to the fact that although this invention has been described and disclosed in relation to certain preferred embodiments, obvious equivalent modifications and alterations thereof will become apparent to one of ordinary skill in this art upon reading and understanding this specification and the claims appended hereto.

Accordingly, the presently disclosed invention is intended to cover all such modifications and alterations, and is limited only by the scope of the claims which follow. All parts and percentages in this specification and claims are expressed in weight percent, unless otherwise noted.