Chemistry

S. D. Okorokov, S. L. Golynko-Vol’fson, T. N. Yarkina

On the Possibility of Directed Change in the Course of Mineral Formation in the System CaO—Al₂O₃—SiO₂

(Presented by Academician P. A. Rebinder on 30 X 1962)

To obtain Portland cement with high strength indices, it is very important to ensure, during firing, an increased content of highly active minerals and, correspondingly, a reduced content of weakly active minerals. Of the two calcium aluminates whose formation is possible in Portland-cement clinker—$C_3A$ and $C_5A_3$—the latter is more desirable, since, in comparison with $C_3A$, it gives considerably higher strength $({}^{1})$. Likewise, of the two calcium silicates—$C_2S$ and $C_3S$—the second is more desirable, because it surpasses $C_2S$ in absolute strength and in rate of hardening.

It is known, however, that when mixtures consisting of $CaCO_3$, $Al_2O_3$, and $SiO_2$ are fired, in the process of binding calcium oxide with acidic oxides, as the temperature and duration of heating increase, $CA$ is first formed from $CaO$ and $Al_2O_3$, which then passes into $C_5A_3$ and subsequently into $C_3A$; analogously, from $CaO$ and $SiO_2$, $C_2S$ is formed at first, which, upon reaching high temperatures, passes into $C_3S$. It is important to note that $C_3A$ is formed at a lower temperature than $C_3S$; thus, normally, saturation of silica with lime to $C_3S$ is possible only after saturation of alumina to its most basic compound—$C_3A$.

We set ourselves the goal of determining whether it is possible to change the course of mineral formation in the system $CaO—Al_2O_3—SiO_2$ in a directed manner such that, after the formation of $C_5A_3$, calcium oxide would proceed not to the formation of the less desirable $C_3A$, but to the saturation of $C_2S$ in order to obtain highly active tricalcium silicate. To solve this problem we made use of the known observations on the instability of $C_3A$ when it is heated with fluorides $({}^{2–6})$, as well as data on the considerable intensification of the formation of $C_3S$ when fluorine-containing mineralizers are introduced into the raw mix $({}^{7})$.

Below are presented the results of experiments demonstrating the possibility of directed mineral formation in the system under consideration, namely the obtaining, from one and the same raw mix, instead of the normal composition—$C_3A + C_2S$—of the more desirable composition—$C_5A_3 + C_3S$.

In the first series of experiments, two mixtures of the following compositions (in moles) were prepared from previously synthesized pure $C_3A$ and pure $C_2S$: 1) $3C_3A + 4C_2S$ and 2) $3C_3A + 8C_2S$.

Each of these mixtures was divided into three equal parts. To the first of them no fluorides were added; to the second and third, $CaF_2$ and $Na_2SiF_6$ were added in an amount corresponding to 2% fluorine by weight of the mineral mixture. Mixtures without mineralizers were fired at 1400°, and those with mineralizers at 1300°.

The resulting clinkers were ground to the same specific surface area, approximately equal to 3000 cm²/g. In the powders the content of free $CaO$ was determined, and their strength on hardening was established by testing small specimens. X-ray patterns were also taken for the clinkers of composition $3C_3A + 4C_2S$.

Chemical analysis showed that the clinkers contained no free $CaO$. X-ray examination showed that, when the two-mineral mixture of composition $3C_3A + 4C_2S$ was fired without addition of fluorides, no changes occurred in its mineralogical composition: on the X-ray patterns—

Fig. 1. X-ray diffraction patterns of mixtures of minerals \(3\,C_3A + 4\,C_2S\), fired without an additive (a) and with an addition of \(CaF_2\) (2% F) (b), with an addition of \(Na_2SiF_6\) (2% F) (c)

... corresponding to this clinker (see the lower X-ray diffraction pattern in Fig. 1), only diffraction maxima of interplanar spacings characteristic of \(C_3A\) and \(C_2S\) are visible. We see a completely different picture in the X-ray diffraction patterns taken from clinkers fired with an addition of fluorides. In this

case (see the two upper curves in Fig. 1), only diffraction maxima of interplanar spacings characteristic of $C_5A_3$ and $C_3S$ are observed on the X-ray diffraction patterns. The lines of $C_3A$ and $C_2S$ are completely absent.

These data, as well as the absence of free lime in the sinters, convincingly show that, during firing of a mixture of two minerals—$C_3A + C_2S$—in the presence of fluorides, a redistribution of lime between alumina and silica takes place. The addition of fluorine-containing compounds leads to decomposition of tricalcium aluminate, and as a result a more low-basic aluminate, $C_5A_3$, is obtained, while the liberated calcium oxide combines with $C_2S$ to form an increased amount of $C_3S$. The change in mineralogical composition occurs in accordance with the following equations (the percentage contents of the initial and newly formed minerals are given in parentheses):

Table 1

Strength during hardening of two-mineral sinters obtained by firing without an additive and with fluoride additives

\[ 3C_3A\,(54.05) + 4C_2S\,(45.95) = C_5A_3\,(39.10) + 4C_3S\,(60.90), \]

\[ 3C_3A\,(37.04) + 8C_2S\,(62.96) = C_5A_3\,(26.79) + 4C_3S\,(41.73) + 4C_2S\,(31.48). \]

Such a substantial change in phase composition, caused by the fluoride additives, leads to a very great increase in the strength of the products obtained during their hardening. This is convincingly demonstrated by Table 1.

The second series of experiments was carried out by us in order to show that directed mineral formation can be achieved not only when firing mixtures of ready-made minerals, but also when firing raw mixes of the corresponding composition. For the experiments, two raw mixes were prepared from chemically pure calcium carbonate, silica gel, and aluminum hydroxide, calculated so that after firing they would give (under conditions of normal mineral formation) clinkers of the same two compositions as those adopted in the first series of experiments: 1) $3C_3A + 4C_2S$ and 2) $3C_3A + 8C_2S$.

The raw mixes were fired without additive and with an addition of fluorides according to the regime adopted in the first series. After preparation, the clinkers were ground to a specific surface corresponding to 3000 cm²/g. No free CaO was found in the clinkers. X-ray diffraction examination showed that the process of mineral formation during firing of raw mixes without an additive and with a fluoride additive proceeds quite differently, as a result of which the phase composition of the fired products is different. The clinker of calculated composition $3C_3A + 4C_2S$, when obtained from a raw mix without fluorides, consisted only of $C_3A$ and $C_2S$. Conversely, clinkers of the same initial calculated composition, when fired with fluoride additives, contained no $C_3A$ or $C_2S$; on the X-ray diffraction patterns of these clinkers, only diffraction maxima of interplanar spacings characteristic of $C_3S$ and $C_5A_3$ were visible. The data of Table 2 show that the strength of clinkers obtained with fluorides very greatly exceeds the strength of the corresponding clinkers fired without fluorides. This is also, although indirect, nevertheless weighty evidence of the directed course of mineral formation during firing.

From the foregoing it is clear that the principal distinctive feature of directed mineral formation is the circumstance that, in the presence of fluorides, calcium oxide combines (after formation during firing—

…material \(C_5A_3\) and \(C_2S\)) not with pentacalcium trialuminate, but with dicalcium silicate. For an additional verification of this, we decided to carry out a third series of experiments. A mixture of the following composition (in moles) was prepared from previously synthesized \(C_5A_3\), \(C_2S\) (in the \(\gamma\)-modification), and \(CaCO_3\): \(C_5A_3 + 4C_2S + 4CaCO_3\).

The amount of \(CaCO_3\) in the indicated mixture was deliberately taken such that it would be sufficient either for the complete conversion of all \(C_5A_3\) into \(C_3A\), or for the complete conversion of \(C_2S\) into \(C_3S\). The experiments were to show with which mineral—\(C_5A_3\) or \(C_2S\)—\(CaO\) would combine during firing of the indicated mixture without fluorides and with the addition of fluorides.

Table 2

Strength during hardening of clinkers fired without additive and with an addition of fluorides

After preparation, the mixture was divided into three parts, to two of which \(CaF_2\) and \(Na_3SiF_6\) were added (calculated as 2% fluorine relative to the expected weight of the fired product).

The firings were carried out according to the regime adopted in the first series. Free \(CaO\) was not found in the clinkers. In an X-ray examination of the radiograph of the clinker obtained without the addition of fluorine-containing compounds, only diffraction maxima of interplanar spacings characteristic of \(C_3A\) and \(C_2S\) were detected. This indicates that normally (i.e., in the absence of fluorides) \(CaO\) combines preferentially with \(C_5A_3\), and not with \(C_2S\). On the radiographs taken from clinkers with additions of \(CaF_2\) and \(Na_2SiF_6\), only diffraction maxima of interplanar spacings characteristic of \(C_3S\) and \(C_5A_3\) were visible; maxima characteristic of \(C_2S\), \(C_3A\), and \(CaO\) were absent. Thus, during firing of the charge with an addition of fluorides, \(CaO\) combines preferentially with \(C_2S\), and not with \(C_5A_3\). Mechanical tests confirmed the above-described features of mineral formation during firing of a charge of composition \(C_5A_3 + 4C_2S + 4CaCO_3\) without additive and with an addition of fluorides: in the latter case the strength proved to be considerably higher than in the former.

The results of the third series of experiments once again clearly show that the presence of fluorine-containing mineralizers in the fired charge, as it were, imposes a prohibition on the formation of the highly basic calcium aluminate of composition \(C_3A\), and, conversely, promotes the more rapid formation, at lower temperatures, of tricalcium silicate.

The general conclusion on the basis of the three series of experiments carried out is that introduction into the raw mix of fluorine-containing additives in the proper amount (of the order of 1–2%, calculated as fluorine) makes it possible to effect a directed change in the course of mineral formation during firing, leading to the production of high-strength, rapidly hardening minerals \(C_5A_3\) and \(C_3S\), instead of the low-strength and slowly hardening \(C_3A\) and \(C_2S\).

Leningrad Technological Institute
named after Lensovet

Received
16 X 1962

Cited Literature

1. S. D. Okorokov, Interaction of Portland-cement clinker minerals in the process of cement hardening, 1945.
2. W. Eitel, Zement, Nos. 2, 3 (1941).
3. N. A. Toropov, B. V. Volkonskii, V. I. Sadkov, Cement, No. 4 (1955).
4. B. V. Volkonskii, Proceedings of the Conference on the Chemistry of Cements, 1956.
5. V. I. Satarin, Cement, No. 3 (1957).
6. S. D. Okorokov, B. V. Volkonskii, T. N. Yarkina, Cement, No. 4 (1962).
7. S. D. Okorokov, S. L. Golynko-Vol’fson, R. I. Sadykova, Proceedings of the Leningrad Technological Institute named after Lensovet, issue 52 (1961).