The aluminate phase

In the early part of stage II, reactions of the aluminate phase will predominate. It is at this stage that the SO4... 

Although gibbsite and kaolinite are important in quantity in some soils and hydrothermal deposits, they have diminishing importance in argillaceous sediments and sedimentary rocks because of their peripheral chemical position. They form the limits of any chemical framework of a clay mineral assemblage and thus rarely become functionally involved in critical clay mineral reactions. This is especially true of systems where most chemical components are inert or extensive variables of the system. More important or characteristic relations will be observed in minerals with more chemical variability which respond readily to minor changes in the thermodynamic parameters of the system in which they are found. However, as the number of chemical components which are intensive variables (perfectly mobile components) increases the aluminous phases become more important because alumina is poorly soluble in aqueous solution, and becomes the inert component and the only extensive variable. 

The aluminate phase constitutes 5-10% of most normal Portland cement clinkers. It is tricalcium aluminate (Ca3Al206), substantially modified in... 

Destructive expansion from reaction with sulphates can occur not only if the latter are present in excessive proportion in the cement, but also from attack on concrete by sulphate solutions. The reaction involves the Al,0,-containing phases in the hardened cement, and in sulphate-resisting Portland cements, its effects are reduced by decreasing the proportion of the aluminate phase, sometimes to zero. This is achieved by decreasing the ratio of AljOj to Fe203 in the raw materials. In the USA, sulphate-resisting Portland cements are called Type V cements. 

Production clinkers have been found to contain cubic or orthorhombic forms of the aluminate phase, alone or in combination. The monoclinic modification has not been observed. The orthorhombic modification is also known as the prismatic, dark interstitial material, and is sometimes pseudo-tetragonal. It can arise only if sufficient alkali is available, but its formation appears to be favoured also by rapid cooling and by bulk compositions potentially able to yield a relatively high proportion of aluminate phase (M12). 

The aluminate phase in clinkers can also be characterized by its composition, determined by X-ray microanalysis this is discussed in the next section. 

Fig. 1.7 Portions of XRD powder patterns of clinkers containing (A) cubic, (B) orthorhombic and (C) pseudotetragonal modifications of the aluminate phase. Peaks marked A and F are due to aluminate and ferrite phases, respectively, and arc rc-indexed, where necessary, to correspond to axes in the text and Table 1.7, and to calculated intensities. After Regourd and Guinier (Rl).

Because of the close admixture with other phases, which is often on a scale of lOpm or less. X-ray microanalysis of the aluminate phase in clinkers is frequently difficult or unreliable. Data have been reported for cubic, orthorhombic, pseudotetragonal or unspecified forms of the aluminate phases in ordinary clinkers (R1,K1,B2,H3) and for aluminate phase (G3,G4,S1,B3) and glass (B3) in white cement clinkers. Tables 1.2 and 1.3 include, respectively, average compositions based on these somewhat scanty data, and suggested site occupancies based on them. The values in both tables take into account both the experimental data and the requirement of reasonable site occupancies. 

As with the aluminate phase, the average compositions take into account the requirement that these site occupancies should be reasonable from the standpoint of crystal chemistry. There is no basis for allocating cations to octahedral and tetrahedral sites separately as the preferences of some of the eations, especially Mg, in this structure are unknown, as is the temperature at which equilibrium is attained. This temperature probably varies between clinkers, and may be expected to affect the distribution. 

Table 2.3 lists some phases containing MgO that are in varying degrees relevant to cement chemistry. It is not a complete list of phases with essential MgO in the CaO-MgO-AljOj-SiOj system. As seen in Chapter 1, some MgO is also taken up by all four of the major clinker phases, typical contents being 0.5-2.0% for alite, 0.5% for belite, 1.4% for the aluminate phase, and 3.0% for the ferrite phase. Magnesium oxide (periclase), like calcium oxide, has the sodium chloride structure it is cubic, with a = 0.4213 nm, space group Fm3m, Z = 4, = 3581 kgm (S5) and refrac-... 

All the effects described above indicate that rapid cooling is desirable the aluminate phase reacts more slowly with water when finely grained and intimately mixed with ferrite, making it easier to control the setting rate (S24), decrease in alite content either from reactions involving the interstitial material or from decomposition is avoided, a higher MgO content can be tolerated, and the clinker is easier to grind. 

If the aluminate phase is orthorhombic, the appropriate composition (Table 1.2) should be substituted in the matrix A. If both cubic and orthorhombic forms are present, intermediate values may be used. 

The results will be less accurate for slowly cooled clinkers, as the compositions of the ferrite and possibly also the aluminate phases may differ significantly from those assumed here. At present, there are not enough data to deal with this problem. The method is not applicable without major modification to clinkers made under reducing conditions. It is doubtful whether the procedure is applicable to white cements, both for this reason and because they may contain glass. 

There is wide agreement that substitution of alkali metal ions retards the early reaction of the aluminate phase, which is thus less for the orthorhombic than for the cubic polymorphs (S35,B43,RI3). The effect has been attributed to structural differences, but the early reaction of pure C, A is also retarded by adding NaOH to the solution, and the OH ion concentration in the solution may be the determining factor (S35). The reaction of C,A is also retarded by iron substitution and by close admixture with ferrite phase formation of a surface layer of reaction products may be a determining factor, at least in later stages of reaction, and the retarding effect of such a layer may be greater if it contains Fe (B44). 

The reaction of the aluminate phase in a cement paste is influenced by the presence of the silicate phases and of alkali (Section 7.6). Ghorab and El Fetouh (G62) studied the effect of the latter on reactions in the pure system. 

Stubby rods of AFt phase are also seen (D25 D27). They are typically some 250 nm long and 100 nm thick. Studies using wet cells show them to occur both on the surfaces of the grains, and at some distance away (S41,S68) (Fig. 7.6b). They are probably more abundant near to the surfaces of the aluminate phase, and appear to nucleate in the solution and on the outer surface of a layer of gel. On drying, this layer shrinks, and the AFt crystals fall back onto the surfaces of the cement grains. The early products thus differ in morphology and composition from the exfoliating foils or honeycombs of C-S-H that have been observed in CjS pastes. 

The concentration of S04 must drop rapidly inside the shells as the aluminate phase reacts, and AFm phase often forms within the shells, any AFt phase formed there initially being replaced by AFm as a result of continued reaction of the aluminate phase (S41,S68). A single specimen may show AFm phase within the shells and AFt phase outside them. The XRD evidence (Section 7,1.1) shows that significant quantities of AFt phase may persist, apparently indefinitely this is presumably material that has been precipitated outside the shells. If the 864 concentration in the solution outside the shells drops before these have sufficiently isolated the anhydrous grains, relatively large ( 10 pm) crystals of AFm phase can form throughout the paste (S68). 

Some workers have concluded from QXDA evidence that the rate of consumption of the aluminate phase is not significantly affected by the presence of gypsum (L33,L36,018), and setting has been attributed largely to... 

The effects of the limestone are partly physical and partly chemical. As with many other finely divided admixtures, including pfa, the hydration of the alite and aluminate phases is accelerated. Because of its fineness the material also acts as a filler between the grains of clinker, though it is unlikely to be as effective in this respect as microsilica. Chemically, it reacts with the aluminate phase, producing C ACHjj, thus competing with the gypsum. 

Jennings et al. (J33) found that with C3S, the retarding effect of sucrose was greatest if it was dissolved in the mixing water. This contrasts with the situation for cement. They concluded that the retarder was incorporated into the initial product and impeded its transformation into a second product. This explanation is compatible with that discussed above if incorporation is taken to mean adsorption on nuclei or growing crystals of the second product. Although there may well be a phase transformation, this evidence does not seem to demand one. Even without one, the retarder would be more effective the sooner it was added, because a smaller area would need to be poisoned. The situation with cement differs due to the competing effect of the aluminate phase. 

The fact tliat both conventional water reducers and superplasticizers are more efTective if added some time after mixing provides a strong indication that adsorption probably occurs at least in part on the hydrated phases, as the anhydrous surfaces have by that time become covered with hydration products. Chiocehio cl al. (C57) found that the optimum time for addition was at the start of the induction period. More of the admixture seems to be taken up by the early hydration products, especially of the aluminate phase, if it is added before the early reaction has subsided. 

Figure 19. Cement hydration process. Calcium silicate hydrates to form C-S-H, a quasi-amorphous gel of composition close to C3S2H3. The excessive rate of hydration of the aluminate phase is controlled by gypsum through the formation of a calcium trisidfoaluminate hydrate, ettringite. (Reproduced with permission from reference 5. Copyright 1991 Elsevier.)...

The hydration of the aluminate phase is very fast and it is essentially over within the first few hours. The contribution of the final product to the mechanical strength of the hardened cement paste is fairly low. It is also susceptible to attack by sulfate ions, which leads to expansion and weakening of the final product. 

In the hydration reaction alite absorbs about 40% by weight of water, of which 24% is chemically bound, and releases 500 J/g. For belite, 21% by weight of water is chemically absorbed, only 250 J of heat per gram are released, and less than half the amount of slaked lime is formed compared with the reaction of alite with water. Hydration of the aluminate phase is the reaction, which consumes most water, up to twice its own weight of water can absorbed in the final product, and releases most heat, 900J/gram. 

In the production of Portland clinker it appears possible to lower the energy consumption by adding by-products containing barium to the raw meal, and by simultaneously lowering its lime saturation factor. As a consequence of these measures the amount of dicalcium silicate in the clinker is increased at the expense of tricalcium silicate. Owing to the presence of Ba ions in its ciystalline lattice this phase becomes converted into its O modification, and its reactivity is increased. At the same time Ba also becomes incorporated into the aluminate phase and modifies its properties (Rajczyk and Mocun-Wczelik, 1992). 

Together with alite and belite, the aluminate phase may somewhat increase the early strength of the hardening cement (this effect being due to the considerable heat of hydration that this compound evolves). Its own hydraulic properties are slight, however. 

---