Appl. Phys. A 74 [Suppl.], S1224–S1226 (2002) / Digital Object Identifier (DOI) 10.1007/s003390101212
Applied Physics A Materials Science & Processing
In situ hydration of Portland cement monitored by neutron diffraction M. Castellote1,∗ , C. Alonso1 , C. Andrade1 , J. Campo2 , X. Turrillas1,∗∗ 1 Eduardo Torroja Institute for Construction Sciences, CSIC, Serrano Galvache S/N, 2 Institut Laue Langevin, 6, rue Jules Horowitz, BP 156, 38042 Grenoble, France
28033 Madrid, Spain
Received: 13 July 2001/Accepted: 24 October 2001 – Springer-Verlag 2002
Abstract. Ordinary Portland cement was mixed with deuterium oxide with a heavy water/cement ratio of 0.6 to monitor the in situ hydration, while acquiring diffraction patterns every 2.5 min. Two different experiments were carried out under different heating conditions. In one case, the temperature was uniformly raised from room temperature to 98 ◦ C at a heating rate of 20 ◦ C/h. In the second case, the heating was performed from 32 ◦ C to 100 ◦ C at a slower rate: 12 ◦ C/h. The disappearance and appearance of relevant anhydrous and hydrated phases were monitored and quantified by fitting isolated diffraction peaks to Gaussians. Ca3 SiO5 dissolved completely during the experiment and Ca2 SiO4 only partially disappeared. Ca(OD)2 precipitated and its growth rate exhibited a sigmoidal shape. Ettringite and hillebrandite formed but then dissolved before the end of the experiments. At the end only Ca(OD)2 and some Ca2 SiO4 remained as crystalline phases. PACS: 61.50.Ks; 61.12.Ld The industrial importance of cement is unquestionable. Cement and related materials are inextricably linked to our environment; the present lifestyle of humankind would not be possible without them. However, despite its relevance, many gaps in cement-chemistry knowledge still remain. Fundamental aspects such as its hydration – particularly during the early ages – remain poorly understood. The complexity of the physical phenomena during the hydration makes it difficult to choose the right experimental approach to isolate and study partial aspects of the process. Although the initial raw material contains a few well-defined crystal phases – calcium silicates mainly – the addition of water triggers a sequence of reactions that involve both crystalline and amorphous phases. Those phases are in dynamic equilibrium and evolve during times spanning up to several ∗ Corresponding
author. (Fax: +34-91/3020-700, E-mail:
[email protected]) ∗∗ Present address: ESRF, BP 220, 38043 Grenoble, France
months. Mechanical properties depend on the hydration evolution of the system during that period. Nevertheless, the fastest and crucial phenomena happen during the first few hours. Traditionally, the early-hours hydration has been studied by stopping the reaction and adding organic solvent mixtures with the hope of obtaining a ‘frozen’ picture of the system. This method although informative is clearly intrusive; noninterfering in situ methods are necessary to detect the phases that exist. The crystalline phases can be studied using diffraction techniques, preferably with an intense incident beam; such is the case for synchrotron X-ray or neutron sources. Both approaches are complementary. Synchrotron light can help to acquire, if necessary, diffraction patterns in a few milliseconds. However, the diffracting volume analysed is very small and it might happen that the observations are not statistically representative of the whole. Neutron-diffraction techniques overcome this problem. Typically, cylindrical volumes of a few cm3 are scanned. Over the years, the crystal structure of cement phases has been determined with both synchrotron and neutron sources, but only during the last ten years have dynamic in situ studies been done. In this respect, the works of Clark [1] and Barnes [2] can be cited. One feature of cement hydration is its exothermic nature. During the setting of large volumes of concrete, the temperature can reach 80 ◦ C. Furthermore, modern techniques of pre-cast concrete manufacture use steam-curing techniques; hence the interest of knowing the hydration as a function of temperature. Examples of these studies with synchrotron light have been published by Hall et al. [3]. Recently, small-angle neutron scattering (SANS) has been used to study the gel-like phases [4, 5]. This technique presents the advantage of allowing the study of the microstructure and the gel-like phases of cement. For the evolution of crystal phases, diffraction with a larger angular domain detection system can be performed. This has been done in the present work, where two examples of monitoring ordinary Portland-cement hydration at moderate temperatures are shown.
S1225
1 Experimental A Spanish commercial Portland cement was used. Its chemical composition is expressed in weight per cent as SiO2 : 19.37, Al2 O3 : 6.12, Fe2 O3 : 3.13, CaO: 62.86, MgO: 1.78, SO3 : 3.23 and free CaO: 1.28. Portions of approximately 5 g of cement were mixed with deuterated water in a water/ cement ratio of 0.6 by weight and, after manual stirring, introduced into amorphous silica tubes of 8-mm internal diameter. The instrument D1B of Institut Max von Laue–Paul Langevin was used to perform the diffraction studies. This is a high-flux two-axes diffractometer coupled to a multidetector with 400 cells encompassing 80 ◦ C in 2θ. The wavelength used was 2.5253 Å. The sample holder was introduced into a furnace with control thermocouples of K -type. Another thermocouple of the same type was introduced inside the sample holder
to monitor the actual temperature of the specimen. The temperature was regulated within an accuracy of ±1 ◦ C. Two experiments with different heating rates were done. In the first, cement and deuterated water were mixed one hour before starting the heating. From 22 ◦ C, a maximum temperature of 95 ◦ C was reached in approximately 4 h, i.e. a heating ramp of 20 ◦ C/h. For the second experiment, the mix was prepared 11 h before. During this time the tube was kept at room temperature. For the neutron-diffraction experiment, the temperature was increased from 30 ◦ C to 100 ◦ C in approximately 6 h; that implies a heating rate of ∼ 12 ◦ C/h. For a more detailed account of the heating profile see Figs. 1 and 2, where the right-hand axes display temperatures. In both cases diffraction data were collected continuously, storing the detector counts every 150 s. The 2θ angular domain explored was 10◦ to 90◦ .
Fig. 1. Thermodiffractometry contour map of the first experiment. The more prominent diffraction peaks are labeled. Cement chemistry shorthand notation is used. C2 S = Ca2 SiO4 , C3 S = Ca3 SiO5 , CH = Ca(OD)2 , CSH stands for hillebrandite and Ett for ettringite
Fig. 2. Same kind of plot for the second experiment. The mix of cement with D2 O was kept at room temperature for 11 h and then analysed by neutron thermodiffractometry
S1226
Data are displayed as the projection – in the 2θ–time plane – of a pseudo-three-dimensional map obtained after sequentially connecting the diffraction patterns. The resulting plot is a contour map which allows us to consider major features of the whole experiment, i.e. the phases, the existence of domains and their growth and decay.
2 Results and discussion In Figs. 1 and 2, contour maps for the diffraction data are sketched. On the right are represented the internal temperature of the samples. The main cement components, Ca2 SiO4 and Ca3 SiO5 , are the only ones detected in the beginning. The diffraction peaks of Ca2 SiO4 overlap with those of Ca3 SiO5 . However, it is possible to distinguish that, while Ca3 SiO5 dissolves after approximately 6 h in both experiments, Ca2 SiO4 remains until the end of the experiments, although some of it has been dissolved. If we focus our attention on the first experiment, which was performed one hour after mixing, we can appreciate that calcium hydroxide does not start to grow noticeably before one hour. This is by far the more conspicuous phase in both experiments. In the second, calcium hydroxide is already present, since the cement was mixed with deuterated water 11 h earlier. A comparison of calcium hydroxide growth can be seen in Fig. 3. In the first experiment, the growth shape is clearly sigmoidal and the error associated with every point is small enough to envisage quantitative kinetics studies. Another interesting crystal phase is ettringite. Its presence is equally notorious. One reflection isolated from the rest (116) permits us to follow its fate. The other reflections overlap with the Ca2 SiO4 ones. There are differences in both experiments regarding the formation of ettringite. In the first, ettringite does not appear until a temperature of ∼ 44 ◦ C is attained, i.e. there is a latent period of 80 min. In the second experiment the ettringite is already formed. Another striking difference is their unequal thermal stability. In the first experiment, ettringite exists up to 95 ◦ C, then abruptly decomposes (see Fig. 3). The ettringite formed at room temperature
is more labile; it decomposes at 70 ◦ C following a similar decaying trend. This is probably related to different compositions, ettringite being a crystal structure very tolerant to variations of constituent ions. This is also confirmed by the quite different spacing associated with the (2 1 6) reflection along the domain of existence; approximately, the d spacing is 0.5% larger in the ettringite formed during the first experiment. It is well known that dissolution of Ca3 SiO5 and to a lesser extent Ca2 SiO4 results in calcium silicate hydrates that are more or less crystalline, which often are designated in the literature as CSH gels. Most of those phases do not show diffraction peaks because their crystallinity is very poor. However, it sometimes can be spotted. In the present experiment, one of these phases known as hillebrandite (Ca2 SiO4 ·H2 O) was noticed. This phase is often found in experiments of calcium silicates with water under mild hydrothermal conditions [6, 7]. There is an isolated reflection (0 6 2) whose intensity can be measured. In Fig. 3, the profile of precipitation and subsequent solution can be seen for both experiments. The remarkable fact is its coincidence in time for both hydrations, despite being conducted under different conditions. CSH is formed at ∼ 50 ◦ C and decomposes at 94 ◦ C in the first experiment, while in the second it precipitates and dissolves at practically the same temperature 95–92 ◦ C. That probably means that the temperature is not the factor responsible for its presence, but rather the complex system of a solid–liquid phase with a variable concentration of ions. 3 Conclusions With these experiments there have been shown some of the possibilities that neutron diffraction can give to study in situ complex and rapidly changing chemical systems involving solution and precipitation of crystalline phases. It has been demonstrated how the solution of calcium silicates takes place with time and temperature. Also, it has been shown that the precipitation of calcium hydroxide could help to quantify and monitor the curing process of cement even under hydrothermal conditions, which is of relevance to the cement industry and more precisely in the context of concrete steam curing. Acknowledgements. We are grateful to all Instrument D1B staff for their help and dedication. The experiments were done thanks to beamtime (Exp 5-25-39) granted by the Institut Max von Laue–Paul Langevin.
References
Fig. 3. Evolution of crystal phases. Normalised integrated intensities vs time. Open symbols refer to first experiment, solid ones to second. Squares represent ettringite (2 1 6) reflection, triangles hillebrandite (0 6 2) reflection and circles represent Ca(OD)2 (1 0 1) reflection growth
1. S.M. Clark, P. Barnes: Cement Concrete Res. 25, 639 (1995) 2. P. Barnes, X. Turrillas, A.C. Jupe, S.I. Colston, D. O’Connor, R.J. Cernik, P. Livesey, C. Hall, D. Bates, R. Dennis: J. Chem. Soc. Faraday Trans. 92, 2187 (1996) 3. C. Hall, P. Barnes, A.C. Jupe, X. Turrillas: J. Chem. Soc. Faraday Trans. 92, 2125 (1996) 4. R.A. Livingston, D.A. Neumann, A. Allen, J.J. Rush: Mat. Res. Soc. Symp. Proc. 376, 459 (1995) 5. B.J. Heuser, M. Popovici, R. Berliner: Mat. Res. Soc. Symp. Proc. 376, 95 (1995) 6. H.F.W. Taylor: Cement Chemistry (Academic Press, London 1990) p. 369 7. Y.S. Dai, J.E. Post: Am. Mineralogist 80, 841 (1995)
