Hydration, carbonation and thermal stability of hydrates in Ca_7−xSr_xZrAl₆O₁₈ cement

Abstract

This paper studies hydration stages and phase transformation mechanism of the Sr²⁺-doped calcium zirconium aluminate cement at room temperature. Features of different hydration stages of this cement paste are identified by X-ray diffraction, scanning electron microscopy, differential thermal analysis, thermogravimetric analysis, evolved gas analysis and heat evolution test. It was found that the partial isovalent substitution of Ca²⁺ for Sr²⁺ such as Ca_7−xSr_xZrAl₆O₁₈, with x = 0.3, 0.6, 1.0, reduced the hydraulic reactivity of Ca₇ZrAl₆O₁₈ phase. The course of hydration of mixed oxides of the type 6CaO·SrO·3Al₂O₃·ZrO₂ documented using microcalorimetry was supported by investigations of the solid hydration products. Research showed that the hexagonal hydrates were stable at early and middle curing ages, an indication that including Sr in solid solution could effectively inhibit the conversion from both C₄AH_13–19 and C₂AH₈ to C₃AH₆ and AH₃. Calcium monocarboaluminate ${\text{C}}_{4} {\text{A}\bar{\text{C}}\text{H}}_{11}$ phases were also found in this system due to the carbonation process of C₄AH₁₃ phase. Substitution of Ca²⁺ ions by Sr²⁺ ions in hexagonal calcium aluminate hydrates causes structural disorder and then contributes to the broadening of lines in the powder X-ray diffraction patterns of layered structures of metastable hydrates. At later curing ages, the formation of two hydrogarnet phases, one Sr-rich (C,Sr)₃AH₆ and the other Ca-rich (C,Sr)₃AH₆, was proved.

Introduction

Calcium aluminates, i.e. calcium monoaluminate (CaAl₂O₄, CA), calcium dialuminate (CaAl₄O₇, CA₂) and dodeca-calcium hepta-aluminate or mayenite (Ca₁₂Al₁₄O₃₃, C₁₂A₇), are the most important constituents of hydraulic calcium aluminate cements (CACs) [1, 2]. Calcium zirconium aluminate (Ca₇ZrAl₆O₁₈, C₇A₃Z) is considered to be only a chemical compound containing zirconium (Zr) in CaO–Al₂O₃–ZrO₂ system that exhibits hydraulic reactivity similar to tricalcium aluminate (Ca₃Al₂O₆, C₃A) [3,4,5]. It is also well known that the hydraulic reactivity of these phases increases with the calcium content of the phase (CaO/Al₂O₃ molar ratio), and therefore, C₇A₃Z shows higher reactivity than mineral phases of CACs. The formation and dehydration of a number of C–A–H-like phases of hydrated Ca₇ZrAl₆O₁₈ were investigated in previous works [6,7,8,9,10,11,12].

The hydration process of the calcium aluminate cement produces similarly the various forms of hydrated calcium aluminate and gibbsite. Dissolution of the anhydrous phases causes a significant increase in concentration of Ca²⁺, Al(OH) ⁻₄ and OH⁻ ions [13]. The precipitation and growth of the hydrates takes place through the solution. When supersaturation is reached, calcium and aluminium ions can combine in variable ratios to precipitate different hydrates CAH₁₀, C₂AH₈, C₄AH₁₉, C₄AH₁₃, C₃AH₆ and AH₃ (C ≡ CaO, A ≡ Al₂O₃, H ≡ H₂O). These processes proceed with time and exhibit the temperature dependence [14]. The C₂AH₈ and C₄AH₁₉ belong to the broad AFm phases with the general formula [Ca₂(Al,Fe)(OH)₆]·X·xH₂O, where X is one formula unit of a singly charged anion (e.g. OH⁻) or half a formula unit of a doubly charged anion (e.g. CO ²⁻₃ ). These crystals have a layered structure consisting of positively charged [Ca₂Al(OH)₆]⁺ sheets with Al(OH) ⁻₄ or OH⁻, together with H₂O in the interlayer region. C₄AH₁₉ contains a complete additional layer of H₂O between the principle layers. C₄AH₁₉ can lose part of its interlayer water molecules to form C₄AH₁₃ at relative humidity below 88%. The initially formed metastable hexagonal hydrates are generally transformed into stable C₃AH₆ belonging to the regular system. The factors controlling transformation are: time, temperature, water content and CO₂ content in the system. In the presence of atmospheric CO₂, part of the interlayer OH⁻ can be replaced by CO ²⁻₃ via interlayer exchange to form two representative carbon-containing AFm phases: stable monocarboaluminate C₃A·CaCO₃·11H₂O (${\text{C}}_{4} {\text{A}\bar{\text{C}}\text{H}}_{11} ;\;{\bar{\text{C}}} = {\text{CO}}_{2}$), which contains only CO ²⁻₃ anions in the X position, and unstable hemicarboaluminate C₃A·0.5CaCO₃·0.5Ca(OH)₂·11.5H₂O (${\text{C}}_{4} {\text{A}}\bar{\text{C}}_{0.5} {\text{H}}_{12}$), which contains both OH⁻ and CO ²⁻₃ anions in the X position of hexagonal hydrates [15].

Due to a poor crystallinity of the hydration products formed at early age hydration and low water-to-cement ratio (w/c) or overlapping peaks due to hydrates with distinct but closely related crystallographic characteristics from a complex XRD profile, the methods of combined thermal analysis (DTA–TG–EGA) are promising ones regarding their qualitative and quantitative characterization. Previous works [16,17,18,19,20,21,22,23,24,25,26,27] indicated various decomposition temperatures of Ca–Al hydrates, obtained by thermal analysis methods. According to both Fleisher [16] and Hill [17], C₄AH_13–19 phase decomposes in the temperature range 200–280 °C or at a temperature of 250 °C, respectively. ${\text{C}}_{4} {\text{A}}\bar{\text{C}}{\text{H}}_{11}$ and C₂AH₈ do indeed give DTA peaks in the same temperature range 160–200 °C [18]. George [19] reports an endothermic peak due to dehydration of C₂AH₈ at 170 °C. Various authors [9, 20,21,22,23,24,25,26,27] have reported DTA curves including evidence for the presence stable cubic C₃AH₆. All report an endothermic peak due to dehydration somewhere in the range 300–330 °C. The products of hydration will also include alumina gel (AH₃-gel) and gibbsite Al(OH)₃ that decompose below 120 °C [18] and at a temperature of around 275 or 280 °C [19, 20], respectively.

It can be strongly argued that the importance of inorganic hydraulic binders is only exceeded by our lack of knowledge as to the factors influencing the synthesis, structures and properties of the calcium aluminates in CACs and calcium zirconium aluminate. One approach to developing this group of materials that can be taken is to examine the effect dopant cations, e.g. Sr²⁺, have on the properties and the hydraulic activity of Ca₇ZrAl₆O₁₈. Some series of oxides Ca_1−xSr_xAl₂O₄ and Ca_3−xSr_xAl₂O₆ were found in the literature review [28,29,30]. Following this approach in this paper, the hydration behaviour of a series of oxides of the type Ca_7−xSr_xZrAl₆O₁₈ using isothermal calorimetry at room temperature is presented. The hydration products were controlled by means of X-ray diffraction analysis, thermal analysis (DTA–TG–EGA) and scanning electron microscopy (SEM/EDS) of the cement stone.

Experimental

Synthesis

The synthesis of a series of oxides of the type Ca_7−xSr_xZrAl₆O₁₈, where x = 0.3, 0.6, 1.0, and the reference material with x = 0 as undoped Ca₇ZrAl₆O₁₈ was performed by a two-step firing procedure using two different sintering temperatures. In a first step, calcium carbonate (99.81% CaCO₃, Chempur), strontium carbonate (99.00% SrCO₃, Merck), α-Al₂O₃ (99.8% Al₂O₃, Across Organics) and zirconia (98.08% ZrO₂, Merck) powders were mixed with appropriate molar ratio of CaO, SrO, Al₂O₃ and ZrO₂ oxides, and then, the mixtures were homogenized for 2 h in a ball mill and pressed into cylinders having a diameter of 2 cm. All green pellets were calcined at 1300 °C for 10 h. In a second step, solid-state sintering of the pellets made from the calcined powder at 1420 °C for 15 h resulted in phases Sr²⁺-doped Ca₇ZrAl₆O₁₈. An intermediate grinding and mixing stage in order to improve homogeneity was necessary. As a final step, the reference Ca₇ZrAl₆O₁₈ and Sr²⁺-doped Ca₇ZrAl₆O₁₈ sinters were ground to a fine powder using a mortar and pestle, and then, they were homogenized for 1 h in a ball mill. The samples were characterized by X-ray powder diffraction using X’Pert ProPANalytical X-ray diffractometer. The XRD patterns of powdered samples were collected by step scanning with step of 0.02° over the range 5°–90° at room temperature. The phases developed during sintering were compared and confirmed using search–match reference ICDD database. The pellets were also polished, coated with carbon and then analysed using Nova NanoSEM 200 scanning electron microscope.

Cementitious paste preparation and methods of investigation

The course of hydration of a series of oxides of the type Ca_7−xSr_xZrAl₆O₁₈, where x = 0.3, 0.6, 1.0, and the reference material with x = 0 as undoped Ca₇ZrAl₆O₁₈ was investigated using isothermal microcalorimeter TAM Air (TA Instruments) by integrating the continuous heat flow signal during the 72-h hydration process. Calorimetric measurements of cementitious phases were taken using the in situ mixing treatments of powder with water with 1.0 of water/cement mass ratio (w/c) and temperature of 25 °C. All tests were performed by mixing the pastes inside the calorimeter to study the early hydration processes of cements. Two grams of dry powder was poured into a glass vial and water weighed into mounted syringes. Admix ampoule was introduced into the calorimeter prior mixing, thus equilibrating to isothermal environment. Mixing was done after equilibration inside the calorimeter to enable the quantitative access to the heat of the early hydration process. A second glass vial containing sand (providing the same heat capacity that the mass of binder paste) was used as an inert reference in the twin channel of calorimeter to ensure best stability of the baseline. Heat flow curves were normalized to the mass of total dry binder and reported in terms of power (mW g⁻¹ dry binder) versus time (hours).

The hydration products and the thermal stability of hydrates were investigated by differential thermal analysis–thermogravimetric analysis–evolved gas analysis (DTA–TG–EGA), X-ray diffraction analysis and scanning electron microscopy–energy-dispersive spectroscopy (SEM–EDS) (mentioned above). For this purpose, the neat cement paste composed of Ca_7−xSr_xZrAl₆O₁₈, where x = 1.0, and water with w/c = 1.0 was homogenized by hand, in a glass baker. Sample in paste form was placed in sealed polyethylene bag and cured up to 21 days in a climatic chamber with the relative humidity maintained at 95% and temperature of 25 °C. As a next step, samples were dried by acetone quenching at specified intervals of time, i.e. 15 min, 0.5 h, 1 h, 24 h, 3 days, 7 days, 14 days and 21 days. Dry powders were then analysed by XRD and thermal analysis. A simultaneous DTA–TG–EGA method (NETZSCH STA 449 F5 Jupiter coupled to QMS 403 D Aëolos) at a heating rate of 10 °C min⁻¹ under a flow of Ar (50 mL min⁻¹) was applied. The containers for samples and reference were corundum crucibles. The reference corundum crucible was kept empty during all DTA–TG–EGA tests. The initial sample mass was 25 mg.

Characterization of microstructure evolution of cement paste by SEM was performed on the freshly broken surfaces of the hardened cement pastes after 24 h, 3 days and 11 days of curing duration. The samples were coated with carbon in order to remove any charge.

Results and discussion

Synthesis of the Sr²⁺-doped calcium zirconium aluminate cement

The samples with different concentrations of strontia sintered at 1420 °C for 15 h were ground, and the phases were identified by XRD analysis. The X-ray diffraction results present no diffraction peaks of SrO and other Sr-bearing phases admixtures, e.g. strontium aluminates. Moreover, the peaks of Ca₇ZrAl₆O₁₈ shifted gradually with increasing the content of SrO, thus indicating that Sr²⁺ ions have been doped into Ca₇ZrAl₆O₁₈ lattice to form the mixed oxides of the type Ca_7−xSr_xZrAl₆O₁₈ solid solutions, where x = 0.3, 0.6, 1.0. The XRD pattern for concentration of strontia x = 1.0 unhydrated composite will be presented later together with the powder XRD patterns for the Ca_7−xSr_xZrAl₆O₁₆ (x = 1.0)-based cement paste. Effect of Sr²⁺ ions doping on Ca₇ZrAl₆O₁₈ lattice was confirmed by comparing the values of the observed peaks with their standard values as per JCPDS No. 98-018-2622 [31]. It is found that with the increase in the SrO content, a shift in the lines corresponds to a somewhat larger unit cell for the “pure” undoped Ca₇ZrAl₆O₁₈ in comparison with Sr-doped Ca₇ZrAl₆O₁₈ solid solutions since the substituting Sr²⁺ ion is larger than Ca²⁺ (1.13 vs. 0.99 Å) [28]. Figure 1a, b shows the results of the SEM and EDS observations of Ca_7−xSr_xZrAl₆O₁₆ (x = 1.0) sinter heat treated at 1420 °C for 15 h. Almost pure crystal phase of Sr²⁺-doped calcium zirconium aluminate with some amount of accessory (Ca,Sr)ZrO₃ was found in the SEM microstructure. The EDS spectrum was taken on the region where strontium formed a solid solution within the matrix, which is shown as grey region in SEM image (Fig. 1b). The mass percentage of elements determined by the quantitative EDS microanalysis in SEM is 26.73 mass% Ca, 12.82 mass% Sr, 12.96 mass% Zr, 20.36 mass% Al and 27.14 mass% O. This results closely relate to the concentration of elements in the stoichiometric composition, designed as (Ca₆Sr)ZrAl₆O₁₈.

Fig. 1

a SEM image of the microstructure of as-synthesized Sr²⁺-doped Ca₇ZrAl₆O₁₈ sample heat treated at 1420 °C for 15 h. b EDS spectrum and quantitative composition analysis of grey region in (a)

Hydration and crystalline hydration products of the Sr²⁺-doped calcium zirconium aluminate cement

The results of isothermal calorimetric experiment (Figs. 2, 3), X-ray diffraction analysis (Figs. 4, 5), simultaneous thermal analysis (Figs. 6, 7a, b) and scanning electron microscopy (Figs. 8–10) were used to study the evolution of phase composition during the process of hydration of Sr²⁺-doped calcium zirconium aluminate cement paste with w/c of 1.0 at room temperature. Figures 2 and 3 exhibit the influence of chemical and structural modification of Ca₇ZrAl₆O₁₈ phase induced by strontium doping upon the calorimetric curves at constant value of water/cement ratio. Figure 2b demonstrates the rate of heat evolution for the blended cements over a 72-h period. As the Ca₇ZrAl₆O₁₈ phase was doped with Sr²⁺ ions (Ca_7−xSr_xZrAl₆O₁₈; x = 0 (as undoped, reference), 0.3, 0.6, 1.0), the intensity of the peak heat rate decreased gradually with increasing strontium content (Fig. 2a). This figure presents not typical heat evolution rate with only one peak that corresponds to initial reactions which follow the mixing of the cement with water. There is also not any occurrence of induction period observed; hence, the hydration reaction starts immediately after mixing of these cementitious phases with water at room temperature. The first peak corresponds to initial reactions which follow the mixing of the cement with water. It records the heat of wetting, dissolution, hydration of Sr²⁺-doped calcium zirconium aluminate and the early formation of hydration products. There is no occurrence of the dormant period or the induction period in the heat flow curve (Fig. 2). After about 2 h, the hydration reaches a steady state and future hydration is diffusion controlled. Thus, the mass transport of water and dissolved ions through the hydration barrier controls hydration process. No other exothermic peaks were also registered during 72 h of experiment. Figure 3 shows an integration of the heat evolution over a period of 72 h. It is particularly easy to make the very clear conclusions that for given strontium-modified cements all hydration produces less heat as compared to pure Ca₇ZrAl₆O₁₈. According to the initial structural model given by Fukuda et al. [5], the highly disordered crystal structure of Ca₇ZrAl₆O₁₈ with the five types of positions of Ca atoms and four types of AlO₄ tetrahedra was found. Ca atoms are located between [Al₆O₁₈] rings created by interconnected six AlO₄ tetrahedra in this structure. The incorporation of Sr (instead of Ca) causes both chemical and structural modifications of Ca₇ZrAl₆O₁₈, thus influencing its hydration kinetics.

Fig. 2

Heat flow curves as a function of hydration time (a the first 2 h; b the period of 72 h) of the undoped reference Ca₇ZrAl₆O₁₈ and the Ca_7−xSr_xZrAl₆O₁₆ (x = 0.3, 0.6 and 1.0) cement pastes prepared with 1.0 of w/c (water-to-cement) ratio and hydrated at room temperature, normalized to mass of binder paste (with in situ mixing condition)

Fig. 3

Cumulative heat curves as a function of hydration time of the undoped reference Ca₇ZrAl₆O₁₈ and the Ca_7−xSr_xZrAl₆O₁₆ (x = 0.3, 0.6 and 1.0) cement pastes prepared with 1.0 of w/c (water-to-cement) ratio and hydrated at room temperature for the period of 72 h

Fig. 4

Powder XRD patterns for the Ca_7−xSr_xZrAl₆O₁₆ (x = 1.0)-based cement paste prepared with 1.0 of w/c ratio and hydrated at room temperature for the period of 0.5 h (b), 24 h (c), 7, 14 and 21 days (d–f). a Unhydrated Ca_7−xSr_xZrAl₆O₁₆ (x = 1.0) phase

Fig. 5

Powder XRD patterns for the Ca_7−xSr_xZrAl₆O₁₆ (x = 1.0)-based cement pastes collected at 2θ from 27° to 35°

Fig. 6

Simultaneous TG (in grey)/DTA (in black) curves of cementitious binder samples measured in argon at flow rate 50 mL min⁻¹ (heating rate 10 °C min⁻¹, initial mass 25 mg)

Fig. 7

Gas evolution curves for representative mass spectroscopic ion fragments of H₂O (a) and CO₂ (b) vapours during the thermal decomposition of the cementitious binder samples in flowing argon, measured in situ by online coupled TG/DTA–EGA–MS system (Ar flow 50 mL min⁻¹, heating rate 10 °C min⁻¹)

Fig. 8

SEM images of the fracture surface of 24-h hydrated Sr²⁺-doped calcium zirconium aluminate cement paste (room temperature)

The course of heat evolution during hydration of cementitious phases described by means of calorimetric curves plotting the rate of heat evolution in mW g⁻¹ can be pertained to changes in the phase composition of the hardened cement pastes controlled by means of X-ray diffraction (XRD).

The mineralogical composition of the acetone quenched hydrated cementitious pastes determined by XRD is presented in Figs. 4 and 5. The original sample is an unhydrated compound (Fig. 4a) that undergoes a phase transformation when it reacts with water at room temperature. The hydration process occurs in the Sr²⁺-doped calcium zirconium aluminate cement at various time intervals (0.5, 24 h; 7, 14 and 21 days). As indicated in Eq. (1), the overall reaction proceeds as chemical changes take place leading to the formation of two separate products (C,Sr)₄AH₁₉ and (C,Sr)₂AH₈ at the early stage of hydration. Since strontium ions can substitute the calcium in the hexagonal “pure” C–A–H and AFm phases [30, 32, 33], broadening of the diffraction peaks caused by structural disorder of layered metastable hydrates [34] was observed.

Hydration

$$\begin{aligned} & 2\left( {{\text{Ca}},{\text{Sr}}} \right)_{7} {\text{ZrAl}}_{6} {\text{O}}_{18} + \, 59{\text{H}}_{2} {\text{O}} \to 2\left( {\left[ {\left( {{\text{Ca}},{\text{Sr}}} \right)_{2} {\text{Al}}\left( {\text{OH}} \right)_{6} } \right] \cdot {\text{OH}} \cdot 6{\text{H}}_{2} {\text{O}}} \right) + 5\left( {\left[ {\left( {{\text{Ca}},{\text{Sr}}} \right)_{2} {\text{Al}}\left( {\text{OH}} \right)_{6} } \right]\left[ {{\text{Al}}\left( {\text{OH}} \right)_{4} } \right] \cdot 3{\text{H}}_{2} {\text{O}}} \right) + 2{\text{ZrO}}_{2} \\ & 2\left( {{\text{C}},{\text{Sr}}} \right)_{7} {\text{A}}_{3} {\text{Z}} + 59{\text{H}} \to \left( {{\text{C}},{\text{Sr}}} \right)_{4} {\text{AH}}_{19} + 5\left( {{\text{C}},{\text{Sr}}} \right)_{2} {\text{AH}}_{8} + 2{\text{Z}} \\ \end{aligned}$$

(1)

The typical crystalline phases of first hydration products were hexagonal tetracalcium aluminate 19-hydrate ((C,Sr)₄AH₁₉) and dicalcium aluminate hydrate ((C,Sr)₂AH₈) that were observed after about 0.5 h of curing in cementitious paste as evidenced from the characteristic diffraction peaks centred at specific angles 2θ = 30.868° (d = 2.89448 Å) and 2θ = 8.2787° (d = 10.67150 Å). This is in good agreement with JCPDS cards No. 00-014-628 and No. 00-045-0564 for C₄AH₁₉ and C₂AH₈, respectively. Nevertheless, the X-ray diffraction patterns of (C,Sr)₄AH₁₉ and (C,Sr)₂AH₈ are very close making no further clear distinction between these two hydration products.

As time progresses, (C,Sr)₄AH₁₉ losses part of its interlayer water to form (C,Sr)₄AH₁₃, according to Eq. (2):

Partial dehydration

$$\begin{aligned} & 2\left( {\left[ {\left( {{\text{Ca}},{\text{Sr}}} \right)_{2} {\text{Al}}\left( {\text{OH}} \right)_{6} } \right] \cdot {\text{OH}} \cdot 6{\text{H}}_{2} {\text{O}}} \right) \to 2\left( {\left[ {\left( {{\text{Ca}},{\text{Sr}}} \right)_{2} {\text{Al}}\left( {\text{OH}} \right)_{6} } \right] \cdot {\text{OH}} \cdot 3{\text{H}}_{2} {\text{O}}} \right) + 2\left( {3{\text{H}}_{2} {\text{O}}} \right) \\ & \left( {{\text{C}},{\text{Sr}}} \right)_{4} {\text{AH}}_{19} \to \left( {{\text{C}},{\text{Sr}}} \right)_{4} {\text{AH}}_{13} + 6{\text{H}} \\ \end{aligned}$$

(2)

The formation of tetracalcium aluminium 13-hydrate (C,Sr)₄AH₁₃) between 24 h and 14 days in the Ca_7−xSr_xZrAl₆O₁₆ (x = 1.0)-based hydrated cement paste was identified by X-ray diffraction analysis, as shown in Fig. 4c–e. The diffraction patterns show the most detectable peak (the “100% peak”) of C₄AH₁₃ at about 2θ = 11.1638° (d = 7.91968 Å) that is in good agreement with reference data (JCPDS card No. 00-011-0203).

The interlayer exchange of OH⁻ by CO₃ ²⁻ proceeds via a stepwise manner such as shown in Eq. 3. This exchange is accompanied by a shift in the position of the diffraction maxima from 2θ = 11.164° to 2θ = 11.6293° (d = 7.60964 Å). The latest one belongs to the calcium carboaluminate hydrate C₃A·CaCO₃·11H₂O (JCPDS cards No. 00-036-0377) which has been identified in 21-day cured cementitious paste.

Carbonation

$$\begin{aligned}&2\left( \left[ \left( {{\text{Ca}},{\text{Sr}}} \right)_{2} {\text{Al}}\left( {\text{OH}} \right)_{6} \right] \cdot {\text{OH}} \cdot 3{\text{H}}_{2} {\text{O}} \right) + {\text{CO}}_{2} \to 2\left( \left[ \left( {{\text{Ca}},{\text{Sr}}} \right)_{2} {\text{Al}}\left( {\text{OH}} \right)_{6} \right] \cdot 0.5{\text{CO}}_{3} \cdot 2.5{\text{H}}_{2} {\text{O}} \right) + 2{\text{H}}_{2} {\text{O}} \\ & \left( {{\text{C}},{\text{Sr}}} \right)_{4} {\text{AH}}_{13} + {\text{CO}}_{2} \to \left( {{\text{C}},{\text{Sr}}} \right)_{4} {\text{A}\bar{\text{C}}\text{H}}_{11} + 2{\text{H}} \\ & \left( {{\text{C}},{\text{Sr}}} \right)_{4} {\text{AH}}_{13} + {\text{CO}}_{2} \to \left( {{\text{C}},{\text{Sr}}} \right)_{3} {\text{A}} \cdot {\text{CaCo}}_{3} \cdot 1 1 {\text{H}}_{2} {\text{O}} + 2{\text{H}} \end{aligned}$$

(3)

In the conversion reaction, compounds (C,Sr)₄AH₁₉ (or (C,Sr)₄AH₁₃) and (C,Sr)₂AH₈ redissolved or interact to give (C,Sr)₃AH₆, hydrogarnet, along with additional AH₃, which over time tends increasingly to approximate to gibbsite (Eqs. 4a–c).

Transformation

$$\begin{aligned} & 2\left( {\left[ {\left( {{\text{Ca}},{\text{Sr}}} \right)_{2} {\text{Al}}\left( {\text{OH}} \right)_{6} } \right] \cdot {\text{OH}} \cdot 6{\text{H}}_{2} {\text{O}}} \right) + \left[ {\left( {{\text{Ca}},{\text{Sr}}} \right)_{2} {\text{Al}}\left( {\text{OH}} \right)_{6} } \right]\left[ {{\text{Al}}\left( {\text{OH}} \right)_{4} } \right] \cdot 3{\text{H}}_{2} {\text{O}} \to 2\left( {{\text{Ca}},{\text{Sr}}} \right)_{3} {\text{Al}}_{2} \left( {{\text{O}}_{4} {\text{H}}_{4} } \right)_{3} + 15{\text{H}}_{2} {\text{O}} \\ & \left( {{\text{C}},{\text{Sr}}} \right)_{4} {\text{AH}}_{19} + \left( {{\text{C}},{\text{Sr}}} \right)_{2} {\text{AH}}_{8} \to 2\left( {{\text{C}},{\text{Sr}}} \right)_{3} {\text{AH}}_{6} + 15{\text{H}} \\ \end{aligned}$$

(4a)

$$\begin{aligned} & 2\left( {\left[ {\left( {{\text{Ca}},{\text{Sr}}} \right)_{2} {\text{Al}}\left( {\text{OH}} \right)_{6} } \right] \cdot {\text{OH}} \cdot 3{\text{H}}_{2} {\text{O}}} \right) + \left[ {\left( {{\text{Ca}},{\text{Sr}}} \right)_{2} {\text{Al}}\left( {\text{OH}} \right)_{6} } \right]\left[ {{\text{Al}}\left( {\text{OH}} \right)_{4} } \right] \cdot 3{\text{H}}_{2} {\text{O}} \to 2\left( {{\text{Ca}},{\text{Sr}}} \right)_{3} {\text{Al}}_{2} \left( {{\text{O}}_{4} {\text{H}}_{4} } \right)_{3} + 9{\text{H}}_{2} {\text{O}} \\ & \left( {{\text{C}},{\text{Sr}}} \right)_{4} {\text{AH}}_{13} + \left( {{\text{C}},{\text{Sr}}} \right)_{2} {\text{AH}}_{8} \to 2\left( {{\text{C}},{\text{Sr}}} \right)_{3} {\text{AH}}_{6} + 9{\text{H}} \\ \end{aligned}$$

(4b)

$$\begin{aligned} & 3\left( {\left[ {\left( {{\text{Ca}},{\text{Sr}}} \right)_{2} {\text{Al}}\left( {\text{OH}} \right)_{6} } \right]\left[ {{\text{Al}}\left( {\text{OH}} \right)_{4} } \right] \cdot 3{\text{H}}_{2} {\text{O}}} \right) \to 2\left( {{\text{Ca}},{\text{Sr}}} \right)_{3} {\text{Al}}_{2} \left( {{\text{O}}_{4} {\text{H}}_{4} } \right)_{3} + 2{\text{Al}}\left( {\text{OH}} \right)_{3} + 9{\text{H}}_{2} {\text{O}} \\ & 3\left( {{\text{C}},{\text{Sr}}} \right)_{2} {\text{AH}}_{8} \to 2\left( {{\text{C}},{\text{Sr}}} \right)_{3} {\text{AH}}_{6} + {\text{AH}}_{3} + 9{\text{H}} \\ \end{aligned}$$

(4c)

The powder X-ray patterns of 24-h–14-day cured cementitious paste and 21-day cured cementitious paste indicate that at least two different kinds of components of hydrogarnet were present in these samples (Fig. 4c–f). In essence, the diffraction patterns of a sample of Ca_7−xSr_xZrAl₆O₁₆ (x = 1.0), treated with water within the curing time range between 24 h and 14 days (Fig. 4c–e) and hydrated up to 21 days (Fig. 4f), show the presence of two series of similar lines that are located very close together, indicating that two hydration products having a similar crystal structure are formed. Going into details, two isostructural hydrates were formed like Sr-rich and Ca-rich cubic phase (C,Sr)₃AH₆. The position of characteristic line of Sr-rich cubic phase (C,Sr)₃AH₆ is very close to the location of Ca-rich one, but it was found at lower values of 2theta. All diffraction lines in powder patterns belonging generally to cubic phase, as compared with standard X-ray diffraction pattern of C₃AH₆ (JCPDS card No. 01-071-0735), are shifted to lower values of 2theta. There are two phenomena in the presented results. Firstly, the three most intense peaks at 16.7921°, 38.1904° and 31.0501° are due to the Ca-rich cubic phase (C,Sr)₃AH₆ in the XRD pattern of Ca_7−xSr_xZrAl₆O₁₆ (x = 1.0) cement paste prepared with 1.0 of w/c ratio and hydrated at room temperature for the period of 21 days (Fig. 4f). Secondly, Fig. 4c–e clearly emphasizes the formation of a second hydrogarnet phase, Sr-rich cubic (C,Sr)₃AH₆, as evidenced from the diffraction peaks positioned at 16.5731°, 37.9245° and 30.8052° of 2θ for 14-day cured sample, given as an example. The significant 2θ peak shifts to lower angles when Sr²⁺ was used, with respect to peaks from “pure” C₃AH₆, and formation of both Sr-rich (C,Sr)₃AH₆ and Ca-rich (C,Sr)₃AH₆-type solid solutions could be attributed to induced crystal-field effects due to the different ionic size of Ca²⁺ (0.99 Å) compared with Sr²⁺ (1.13 Å) [28]. Hence, Sr-rich (C,Sr)₃AH₆ was stable within a reaction period from 24 h to 14 days (Fig. 4c–e). The consecutive formation of two separate solid solution members, supported by Ref. [28], one close to the strontium (Sr-rich (C,Sr)₃AH₆) and the other to the calcium (Ca-rich (C,Sr)₃AH₆) end, leads to the conclusion that the strontium reacted first, by reason of its greater solubility, leaving a solution richer in lime to form later. Shifting to the left from the standard XRD peaks of C₃AH₆ due to replacing Ca²⁺ with Sr²⁺ is evident from both Fig. 5c–f and Table 1. Moreover, residues of unhydrated (Ca,Sr)₇ZrAl₆O₁₈ were observed up to 14 days of curing, and this phase continues to hydrate slowly over the next 7 days (Fig. 5e–f).

Table 1 Reference data for “pure” C₃AH₆ (JCPDS card No. 01-071-0735) and the related 2θ [°] positions and d [Å] values obtained in this work

Thermal stability of hydration products of the Sr²⁺-doped calcium zirconium aluminate cement paste

The simultaneous TG/DTA measurements show several decomposition stages of the Sr²⁺-doped calcium zirconium aluminate cement paste after curing time period of 15 min to 21 days (Fig. 6). The progress of cement hydration consumes water and produces hydrates; thus, the highest loss of mass due to the evaporation of the water chemically bonded to C–Sr–A–H phases was achieved after 21 days of moist curing. The total mass loss due to decomposition of the cement paste between about 25 and 1000 °C is 9.89, 10.23, 12.98, 28.71, 32.05, 35.53, 35.59 and 38.00% of the initial mass, corresponding to the length of curing time according to thermogravimetric analysis results displayed as TG curves (in grey) for samples cured as long as 15 min, 0.5 h, 1 h, 24 h, 3 days, 7 days, 14 days and 21 days, respectively. These curves can be compared to the in situ analysis of the evolved gaseous species by the online coupled EGA–MS system. Main evolution curves of H₂O (m/z = 18) and CO₂ (m/z = 44) are also shown as ion current versus temperature in Fig. 7a, b. Evolution of H₂O shows parallel run to the DTA curves having local maximums at the same temperature values. Nevertheless, the endothermic dehydration peaks are more sensitively detected by the MS than DTA curves and the characteristic peaks of evolution of H₂O are clearly visible in all the samples. In the temperature range of 25–1000 °C, the peaks attributed to H₂O are apparently strengthened, thus confirming the water consumed in the formation of the hydration products with hydration time. The signals for H₂O change from one peak at 259 °C to strong peaks between 90 and 310 °C. The EGA peak attributed to H₂O at 259, 258 or 266 °C was due to the decomposition reaction of the crystalline (C,Sr)₄AH₁₉, which was formed preferentially during the first 15 min and exists up to ca. 1 h of hydration (Fig. 7a). Next, fraction of bound water of (C,Sr)₄AH₁₉ was lost, and within a time range between 24 h and 7 days, the (C,Sr)₄AH₁₉ lost part of its interlayer water to form (C,Sr)₄AH₁₃. At the temperature of 273 or 274 °C, the peaks attributed to H₂O due to the decomposition reaction of (C,Sr)₄AH₁₃ were observed in the mass spectroscopic gas-evolution curves of the cementitious pastes hydrated between 24 h and 7 days (Fig. 7a). Starting from 14 days of hydration, this peak was not observed clearly in the m/z 18 curve that proves the partial disappearance of (C,Sr)₄AH₁₃ phase including its carbonation and formation of the crystalline $\left( {\text{C,Sr}} \right)_{4} {\text{A}}\bar{C}{\text{H}}_{11}$ phase within a time range between 14 and 21 days. The peak of tetracalcium aluminium carbonate hydrate coincides with the peak (C,Sr)₂AH₈ at about 178 °C. The EGA peak attributed to H₂O evolution at 158, 166, 173, 179 or 178 °C was due to the decomposition reaction of the crystalline (C,Sr)₂AH₈, which was detected within a time range between 0.5 h and 21 days (Fig. 7a). The gradual shift of these peaks to higher temperature as curing time progresses may be caused by an increase in the degree of crystallinity this hydrate.

The water release from Sr-rich (C,Sr)₃AH₆ and Ca-rich (C,Sr)₃AH₆ begins as soon as the conversion of metastable hydrates to the stable higher-density phases starts after 24 h of hydration and it is visible as EGA peak at ca. 300 °C in the evolution curves of H₂O of the Sr²⁺-doped calcium zirconium aluminate cement paste hydrated within a time range between 24 h and 21 days (Fig. 7a). Despite the fact that the cubic phases Sr-rich (C,Sr)₃AH₆ and Ca-rich (C,Sr)₃AH₆ were clearly distinguished from the XRD patterns of both 14-day and 21-day hydrated cement pastes as shown in Figs. 4, 5, there is no clear difference in the EGA peaks at 309 and 308 °C (Fig. 7a), respectively.

According to the DTA, TG and EGA curves, the loss of mass due to the decomposition of AH₃ gel takes place below 120 °C with the broad endothermic effect in all blended cement pastes but is clearly visible as a strong effect in samples cured for at least 24 h. The decomposition step due to the dehydration of crystalline gibbsite Al(OH)₃ is observed in Fig. 7a as the EGA peaks at 273 °C that precede the peak of Ca-rich (C,Sr)₃AH₆ under long-term curing age up to 21 days. The evolution of CO₂ gas from the carbonated Sr²⁺-doped calcium zirconium aluminate cement consisted of different stages with broad EGA peaks (Fig. 7b).

Microstructural investigations of the Sr²⁺-doped calcium zirconium aluminate cement paste

The microstructure of hydration products of the Sr²⁺-doped calcium zirconium aluminate binder paste was examined by SEM/EDS (Figs. 8–10a). The SEM micrographs in Fig. 8 show that the flake-like crystals of C–Sr–A–H and amorphous gels are the major hexagonal hydration products and they were formed at age of 24 h at room temperature. As hydration proceeds, the unhydrated grain core is continuously consumed to form hydration products and the hydrated product of the Sr²⁺-doped calcium zirconium aluminate compound in a grain of cement adheres firmly to the unhydrated core of the grains of cement as shown in Figs. 9 and 10a. The characteristics of the hydrates were examined by EDS analysis after the SEM examination, as illustrated in Fig. 10b and Table 2. These results show that the hydrated matrix contains calcium, strontium and aluminium, which can be related to the calcium strontium aluminium hydrates (C–Sr–A–H). There was also a clear segregation of zirconium within the reaction zone generated from unhydrated grain core of (Ca,Sr)₇ZrAl₆O₁₈ to the hydrated matrix C–Sr–A–H (points 1–2–3).

Fig. 9

SEM images of the fracture surface of 3-day hydrated Sr²⁺-doped calcium zirconium aluminate cement paste (room temperature)

Fig. 10

SEM images of the fracture surface of 11-day hydrated Sr²⁺-doped calcium zirconium aluminate cement paste (room temperature). Spot 1–3 EDS analysis. a EDS spectrum of sample in the micro area 3

Table 2 EDS analyses of the points in Fig. 10a (unit: atom%)

Conclusions

• 1. Since no diffraction peaks of SrO and other Sr-bearing phases, e.g. strontium aluminates, can be detected in samples and the peaks of Ca₇ZrAl₆O₁₈ shift gradually with the composition, it indicates that SrO has been doped into Ca₇ZrAl₆O₁₈ lattice to form uniform solid solution.

• 2. Intensity of the first peak of heat evolution is gradually reduced as the x value in the Ca_7−xSr_xZrAl₆O₁₈ (x = 0.3, 0.6, 1.0) formula increases.

• 3. The hydration heat first peak of the mixed oxides of the type 6CaO·SrO·3Al₂O₃·ZrO₂ is generated simultaneously with both the wetting of cement grains and the formation of Sr-doped hexagonal C₄AH_13–19 and C₂AH₈ hydration products.

• 4. Including strontium ions in the solid solutions of unstable hexagonal hydrates could effectively inhibit the conversion from both C₄AH_13–19 and C₂AH₈ to stable C₃AH₆ and AH₃ products.

• 5. The new phase C₃A·CaCO₃·11H₂O (${\text{C}}_{4} {\text{A}}\bar{C}{\text{H}}_{11}$) was found in this system due to the activity of carbon dioxide; the formation of C₄AH_13–19 and C₂AH₈ is prohibited and the generation of stable cubic phase is delayed in the early hydration process.

• 6. Two hydrogarnet phases, one Sr-rich (C,Sr)₃AH₆ and the other Ca-rich (C,Sr)₃AH₆, were formed due to the inevitable and irreversible conversion of metastable hydrates.