Unpicking the fundamental reactions behind gypsum, the infinitely-recyclable mineral.
Gypsum is one of the most widely-used building materials in the world. It is applied in several interior building uses, including in wallboard. Normally, calcium sulphate dihydrate (CaSO4•2H2O) is used as a raw material and a subhydrate such as hemihydrate (CaSO4•0.5H2O) or anhydrite (CaSO4) is produced. The subhydrate is the binding material, which reacts with water back to form gypsum.
There are three main diagrams that can help us understand this cycle in detail. The first describes the process of water leaving the structure of gypsum to form hemihydrate or anhydrite. The second depicts the temperature-dependent solubility of the calcium sulphates in water. The third illustrates the hydration energy of hemihydrate as it reacts with water.
Diagram 1 - Dehydration
The first process is the formation of hemihydrate or anhydrite from dihydrate. To obtain more information about this step, gypsum crystals were grown from sodium sulphate and calcium chloride via a diffusion process. A large test tube, approximately 200mL in volume, was filled with 3g of sodium sulphate (Na2SO4). A smaller test tube, around 25mL in volume, was filled with 3g of calcium chloride (CaCl2). The small test tube was immersed in the larger one and the complete system was filled with deionised water and sealed. The tubes were stored for around 10 weeks, until gypsum crystals had grown. The crystals were collected and used for thermal analysis. Natural gypsum crystals were also cut into 15mm-long pieces along different axes to measure thermal expansion using dilatometry.
Figure 1a shows the test tube with crystals, Figure 1b shows the natural crystal and Figures 1c and 1d show the directions along which dilatometry was measured.
The dehydration of gypsum can be observed using thermo-gravimetric measurement and differential scanning calorimetry. Depending on the parameters of measurement, a two-step process can be observed. The formation of beta-hemihydrate starts at around 85°C and is shifted into the formation of anhydrite III. Due to this, it is tricky to measure the mass loss of wet gypsum. Physically-bound water evaporates from the starting point until around 120°C, depending on the heating rate.
If physically-bound water and chemically-bound water are to be measured individually, a drying phase with dry gas at around 40-60°C is recommended. Under such conditions, most of the physically-bound water will evaporate. After this, the chemically-bound water can be measured accurately.
The surfaces of the crystals experience some cracking because of the changes in crystal structure and the water leaving the structure, although no changes are visible until after ~5.5% of the overall mass is lost. A volume change can be observed depending on the axis of the gypsum crystal. The surface of the formed beta-hemihydrate is larger, because the surface of the crystal cracks and a fishbone structure appears, as shown in Figures 2c and 2d.1
The formation of anhydrite III takes place in a second step, which is also not easy to observe as a ‘perfect peak’ in thermal analysis. At around 400°C a change in crystal structure of anhydrite can be measured due to the formation of anhydrite II. At around 700°C a sintering of the crystals takes place. This process is shown in Figures 2e and 2f. This sintering is responsible for the decrease in solubility and the formation of insoluble anhydrite. At around 1250°C, anhydrite I will be formed. At higher termperatures it decomposes into CaO and SO3. These last two processes will not be discussed in this article.
Diagram 2 - Solubility of different phases
The solubilities of different calcium sulphate (hydrate) phases in water are also key to understanding the system, especially the hydration of hemihydrate. Figure 3a is presented in almost every building materials textbook.2 - 4 The results presented are roughly correct, but some points have to be taken into account. For hemihydrate below ~100°C it is not really a ‘solubility’ being plotted, but the maximum ion concentration during the solving of hemihydrate. This has to be mentioned, because solubility is defined in equilibrium and because of the formation of dihydrate, which has a lower solubility, an equilibrium cannot be reached. The ‘solubility’ or solving rate for hemihydrate decreases with higher temperature.
Crystals with such behaviour normally have a positive solving energy. Solving hemihydrate is exothermic. For gypsum, a nearly constant solubility can be observed, so the solving energy of gypsum is around zero. Solving energy and crystallisation energy in addition should also amount to zero, so the crystallisation energy of gypsum is also close to zero and for sure much lower than the solving energy of hemihydrate. The heat of hydration of hemihydrate is mainly influenced by the solving energy of hemihydrate, besides a small solvation energy at the beginning of the reaction, when the hemihydrate comes into contact with water. For anhydrite, a lower solving energy should be expected.
Two different types of hemihydrate - alpha and beta - are discussed in the diagram. Beta hemihydrate is formed when gypsum is heated up to around 100°C in a normal open kettle or rotary kiln. Alpha hemihydrate is formed in an autoclave above 100°C in a water (suspension autoclave) or in a steam autoclave.5 - 7 This is possible as, above 100°C, hemihydrate becomes less soluble than dihydrate. The solubility of dihydrate above 100°C is not shown in the diagram, but it can be expected to increase. In suspension autoclaves, large crystals without branches form, while in steam autoclaves large crystals with more surface area, including branches, are formed.
Normally, the surface area of suspension autoclave alpha-hemihydrate is smaller compared to the steam autoclave alpha-hemihydrate, which is in turn lower than that of beta-hemihydrate. This is visible in Figures 3b and 3c, which shows the flat and smooth surface of the crystals in comparison to the rough surface of the beta-hemihydrate. Therefore, a minimum of three lines for the ion concentration of hemihydrate should be given, one for suspension autoclave alpha-, one for steam autoclave alpha- and another one for beta-hemihydrate. Indeed, different autoclaves and different raw gypsum sources will result in a different hemihydrate surface. The authors suggest showing the hemihydrate solubility as a ‘field’ that is roughly in the ranges given by literature, rather than as clearly-defined lines.
The maximum ion concentration reached depends on the surface of the hemihydrate: The higher the surface area, the higher the concentration. Another factor is the defect density in the crystals. The higher the ion concentration in the solution, the more gypsum seed crystals are formed - and the more rapidly they are formed. Besides, a higher ion concentration leads to more branches on the gypsum crystals.
Technical properties like strength and porosity also depend on the crystal morphology.8 At higher temperatures, larger crystals with fewer branches can be expected. When the temperature increases and a very low hemihydrate solubility is reached, a seed crystal formation is very slow, comparable to anhydrite. Super-saturation now influences the speed of seed crystal formation.
At higher temperatures the super-saturation is lower and the expected reaction speed is lower. However, as the ions can move faster at higher temperature, the speed of the hydration reaction is increased. With a higher temperature fewer seeds are formed and the gypsum crystals can grow larger (Figure 3d - f). At a lower super-saturation not only are larger crystals formed, but they normally have even fewer branches. At temperatures around 90°C the ion concentration required for seed crystal formation is only reached very slowly, if at all. So, as in the case of anhydrite at around 30°C, no reaction can be observed for some hours.
From Figure 3 and the information above it is also clear why anhydrite needs an accelerator, as the super-saturation is too low for fast seeding. At lower temperatures, the hydration speed of anhydrite is much higher and a slow hydration of pure anhydrite can be observed.
Diagram 3 - Heat of hydration of hemihydrate
The third type of diagram that needs to be understood is the heat of hydration of hemihydrate. The heat of hydration is measured with heat flow calorimetry. As explained before, energy of hydration is mainly the solving heat of the hemihydrate. Figure 4 shows the heat of hydration for a typical alpha-hemihyrate, with and without seed crystals, and for a beta-hemihydrate.
The first peak shows a small solvation energy due to the first solvation of hemihydrate. The finer the hemihydrate is, the more pronounced the first peak. This is ascribed to an increased solvation energy due to more rapid dissolution of the hemihydrate. Thus, a higher supersaturation can be reached.
The next step is the retardation period, which is shorter for a higher supersaturation in the reaction solution, because the seeding is faster. If that is the case, the retardation period shortens, or even becomes completely absent, as additional seeding crystals are used as accelerators (See Figure 4). For beta-hemihydrate, which has a very high surface area, the maximum ion concentration is very high and is reached very rapidly. Seeding is so fast that only a very short retardation period will be observed, if at all. The more seeds that are formed, the smaller the gypsum crystals will be. In theory, small hemihydrate crystals mainly influence the seeding process during the hydration and larger crystals influence the growth of the gypsum crystals.
After the retardation period, the next peak indicates the start of the main reaction, which can be divided into two parts: Before and after the peak. Up until the peak of heat evolution, the crystal growth of gypsum is the main influencing factor because there is enough hemihydrate left to deliver the needed ions. On the right side of the energy peak maximum, most of the hemihydrate is dissolved, and only parts of the largest crystals are left. The dissolution of hemihydrate becomes the limiting factor until all hemihydrate is consumed.
After this step, no energy can be registered and the reaction stops. The solution is still a saturated gypsum solution. By drying such samples, all the dissolved ions will form gypsum. If the drying process is slow, the gypsum crystals that had already formed simply grow larger. However, in the case of rapid drying, a second generation of smaller crystals is formed. The size of the second generation of crystals depends on the drying speed: The faster the drying, the smaller the crystals. The second generation of crystals can be very fine and can consequently have a high impact on the crystal surface area. The effect can, for example, be analysed using gas adsorption. If the reaction is stopped, for example by adding ethanol, small crystals will form, as a result of the change in the dielectrical constant of the liquid phase.
Normally, seeding takes place only once in the system, because only one peak for supersaturation can be measured. If two or more generations of gypsum crystals can be found in a system, seed crystals have been used or effects explained above have taken place. The seed crystals are normally larger than gypsum crystals formed during hydration. To prevent extra seed crystal formation from the supersaturated solution, a large number of seed crystals are needed; normally, a mixture of grown seed crystals and crystals formed from the supersaturated solution can be found.
According to Lavoisier and Le Chatelier,9 the reaction is a process of dissolving hemihydrate to create a super saturated solution of gypsum, from which gypsum is formed. This is possible due to the lower solubility of gypsum compared to hemihydrate. This theory is proven by Figure 4a - 4i and it shows why gypsum can act as an accelerator. If seeds are added, the reaction can start directly, either with or without or besides extra seed crystals. In case of low concentrations of gypsum crystals, extra seeds will be formed and, in the hydrated product, two or more generations of gypsum crystals can be found. Cavazzi and Traube10 described a gel-like interphase, which is stepwise forming gypsum crystals. Fiedler11 described an inner hydration of the hemihydrate. Periderij12 and Eipeltauer13 combined Fiedler’s and Le Chatelier’s and Lavoisier’s theory by describing an inner hydration beside the gypsum formation out of a super saturated solution.
Gypsum is a great material for use inside buildings. However, outside use is tricky, as strength is lost in the presence of moisture. In this article we discuss the hydration of hemihydrates. A video of the hydration of alpha hemihydrate can be downloaded at: https://www.chemie-biologie.uni-siegen.de/bwc/personen/anlagen/alpha_mit_wasser_240_min.avi?m=e .
The hydration of hemihydrate and anhydrite is a question of dissolution and recrystallisation. To understand the process, these two steps must be understood in detail. With the three diagrams presented here, a basic knowledge of gypsum can be gained.
References
1. Pritzel, C., Kowald, T., Sakalli, Y., Trettin, R.; ‘Binding materials based on calcium sulphates,’ Cementitious Materials: Composition, Properties, Application, Editor H. Pöllmann; S. 285 – 309; ISBN: 978-3110473735; De Gruyter, 2017.
2. Foerst, K.W.; ‘Ullmanns Encyklopädie der technischen Chemie,’ Aufl., Bd. 8, p 97-132, Munich; Urban & Schwarzenberg, 1957.
3. Freyer, D.; ‘Untersuchungen zur Phasenbildung und -stabilität im System Na2SO4 – CaSO4 – H2O,’ Dissertation, Technische Universität Bergakademie Freiberg, 2000.
4. Förthner, S.; ‘Untersuchungen zur Kristallisation und zur Kristallmorphologiebeeinflussung von Gipsen zur Steuerung der makro-physikalischen Kenngrößen’, Dissertation, Friedrich-Alexander-Universität Erlangen-Nürnberg, 2011.
5. Fischer, K.W., Uschmann, W.; ‘Die Erzeugung von Alpha-Halbhydratgips in technischen Salzlösungen; Silikattechnik15 Nr.11, 1964.
6. Oetzel, M., Koslowski, T., Wiesgickl, G., Wosnitza, F.; ‘Patentschrift Verfahren zur Herstellung von Calciumsulfat-Alpha-Halbhydrat,’ Aktenzeichen 19737778, 1997.
7. Powell, D.; ‘The α- and β-Forms of Calcium Sulphate Hemihydrate,’ Nature 185, 375–376, 1960. https://doi.org/10.1038/185375a0.
8. Pritzel, C., Trettin R.; ‘Influencing the morphology of gypsum; Proceedings of the 10th International Congress for Applied Mineralogy (ICAM);’ Editor A.T.M. Broekmans (Editor); S. 541 – 548; ISBN-Nr. 978-3-642-27681-1; e-ISBN-Nr. 978-3-642-27682-8; Springer-Verlag; Berlin, 2012.
9. Le Chatelier, M.H.; ‘Crystalloids against colloids in the theory of cements,’ Trans. Faraday Soc. 14, 1919.
10. Cavazzi, A.; ‘Das gelatinöse Calciumsulfat und das Abbinden des Gipses,’ Kolloid-Zeitschr. 11, 1912.
11. Becherer, G., Fiedler, H.; ‘Über röntgenografische Untersuchungen des Abbindevorgangs bei Gips,’ Silikattechnik Heft 7, 6.Jahrgang 1955.
12. Perederij, I. A.; ‘Theorie der Bildung, Erhärtung und Festigkeit von normalem Gips und hochfestem Gips GP,’ Chem. Techn. 8, 1956.
13. Eipeltauer E.; ‘Erzeugung von krichfesten Hartgipsen,’ Zem.-Kalk-Gips 6, 1960.
