Introduction

Hydraulic bonding via calcium aluminate cement (CAC) is the most applied mechanism due to both castable fluidity and green mechanical strength [1–5]. Nevertheless, several different alternative bonding systems for monolithic refractories have been developed in recent years mainly due to the necessity of the reduction in CaO content and thereby improve cold and hot strength, reduce porosity, increase density and corrosion resistance of castables [6–8]. At present, special emphasis is put on the preparation and application of reactive alumina able of forming hydrated phases in water, as an alternative hydraulic binder in unshaped refractory materials [9, 10]. Hydraulic alumina powders are widely prepared starting from Bayer gibbsite (α-aluminum hydroxide) via three steps; the reduction in the soda content, heat treatment and particle size reduction in an attritor ball mill [9]. Utility properties of reactive alumina are generally determined by the chemistry, especially by the soda content, primary crystal size, specific surface area (BET) and particle size distribution (mono-/bi-/multimodal). When it is in contact with water, it is gradually converted to bayerite polymorph of Al(OH)3 [10]. Incomplete reaction was found due to the differences in microstructure of hydratable alumina grains.

Besides, many chemical additives may be added to the refractory castables, e.g., to control the hydration process (accelerators/retarders) [11]. For example, Li+ salts and alkali or calcium hydroxides are considered to be strong accelerators [12]. Nevertheless, many accelerators have a negative effect on the properties of certain refractories that are applied in high-temperature and highly corrosive environments [13]. These disadvantages can be overcome by using a water-reactive substance chemically similar to the major compounds of CAC. In view of this state of affairs, the major thrust in development of superior cementitious systems is based on the combination of very fine particles of refractory oxides, e.g., reactive alumina micropowder, with other components tailored specifically for this purpose. In our previous works [14–17], we have investigated a hydratable compound, i.e., calcium zirconium aluminate with chemical formula of Ca7ZrAl6O18. It has a distinctly similar mechanism of hydration when compared to CAC, especially in terms of the C–A–H phases formed in hydrated cement paste. Calcium zirconium aluminate, Ca7ZrAl6O18, hydrates and hardens very quickly comparable to tricalcium aluminate, Ca3Al2O6 [18]. It liberates a large amount of heat almost immediately after mixing with water [14]. Hence, it may be used in high-temperature applications, especially as an additive in mass corundum castables, where it can help to shorten the setting time, or to accelerate the hardening process. An additional benefit of using Ca7ZrAl6O18 would be the improvement of physical and chemical properties of refractory castables due to the formation of calcium zirconate, CaZrO3, characterized by high melting point, low thermal expansion coefficient, high strength and excellent corrosion resistance against alkali oxides [14, 19]. In this paper, we propose and investigate the hydration processes of new binder based on the hydratable materials: calcium zirconium aluminate and reactive alumina as a new cementitious system for the technology of monolithic refractories.

Experimental

Synthesis procedure of calcium zirconium aluminate (Ca7ZrAl6O18)

The subject of the study was hardened Ca7ZrA6O18 and reactive alumina/Ca7ZrA6O18 pastes prepared with water/solid ratio of 0.5 and cured up to 32 days under different curing temperatures. Calcium zirconium aluminate, Ca7ZrAl6O18, was synthesized by high-temperature solid-state method. For a typical synthesis, stoichiometric amounts calcium carbonate (98.81 % CaCO3, Chempur), aluminum oxide (99.7 % Al2O3, Chempur) and zirconium dioxide (98.08 % ZrO2, Merck) were mixed for 2 h in a ball mill to obtain a homogeneous mixture. The mixture was then shaped in the form of cylindrical specimens (diameter 20 mm, height 20 mm) and calcined at 1200 °C for 10 h in an air atmosphere electric furnace. The heating was done at an average rate of 2 °C min−1. The material after the calcination process was crushed and ground in a ball mill to a fineness below 63 µm and again pressed under 80 MPa. All the pressed samples were sintered at 1500 °C and a soaking period of 30 h at the designed temperature. The as-obtained product was ball-milled to form powder. Additionally, industrial reactive alumina (PFR, Alteo) was also used as a raw material.

Hydrated paste samples preparation and test methods

The hydraulic behavior of Ca7ZrA6O18 as well as of the mixtures of Ca7ZrA6O18 with reactive alumina was investigated by differential calorimetric analysis and other methods. The investigated mixtures were 75 mass% Ca7ZrA6O18/25 mass% Al2O3, 50 mass% Ca7ZrA6O18/50 mass% Al2O3 and 25 mass% Ca7ZrA6O18/75 mass% Al2O3. The single component samples were 100 mass% Ca7ZrA6O18 and 100 mass% Al2O3. To further simplify the notation, individual oxides are abbreviated as a single letter, i.e., C≡CaO, A≡Al2O3, Z≡ZrO2 and H≡H2O. Hydrated paste samples were prepared as follows: The following reagents in the given proportions were homogenized together and then mixed with water to produce pastes which were cured in a climatic chamber at 20 and 50 °C for 32 days at 90 % relative humidity (RH). The obtained samples were denoted as 100C7A3Z-H-T20, 25A-75C7A3Z-H-T20, 50A-50C7A3Z-H-T20 and 75A-25C7A3Z-H-T20 (cured at 20 °C) and 100C7A3Z-H-T50, 25A-75C7A3Z-H-T50, 50A-50C7A3Z-H-T50 and 75A-25C7A3Z-H-T50 (cured at 50 °C). For use in XRD, FT-IR and DTA–TG tests, the cores of broken pastes were ground and treated with cold acetone to stop the hydration process. X-ray diffraction (XRD), differential thermal/thermogravimetric analysis (DTA–TG) and Fourier transform infrared spectroscopy (FT-IR) were performed to identify the phase composition of the fine powder samples. X-ray powder diffraction (XRD) patterns were obtained on a X’Pert Pro PANalytical X-ray diffractometer. The XRD patterns of previously synthesized material, reactive alumina and acetone quenched hydrated powdered samples were collected by step scanning with step of 0.02 degree over the range 5°–90° at room temperature. The specific surface and grain size distribution of the synthesized Ca7ZrAl6O18 material were measured by a laser diffraction analyzer (Master Sizer 2000 version 5.60 apparatus of Malvern, UK).

DTA–TG was carried out using an NETZSCH STA 449 F3 Jupiter thermal analyzer. The temperature was programmed to rise at a constant heating rate of 10 °C min−1 up to 1000 °C. The test was performed under an air flow of 40 mL min−1. A Bruker Vertex 70v spectrometer was used for studying the curing of the Ca7ZrAl6O18 and reactive alumina/Ca7ZrAl6O18 pastes. The FT-IR spectrometer was operated in the mid-IR region (4000–400 cm−1) with a resolution of 4 cm−1. Each spectrum was the average of 128 scans. Samples were prepared by the standard KBr pellet method. The morphologies of the binder pastes were observed through using scanning electron microscope FEI Nova NanoSEM 200. For the analysis, freshly broken surfaces were prepared without hydration interruption before observation and coated with carbon.

Hydration heat of 100C7A3Z, 25A-75C7A3Z, 50A-50C7A3Z, 75A-25C7A3Z and 100A with water-to-solid ratio of 0.5 was measured by the differential non-isothermal–non-adiabatic calorimeter. The homogenized pastes of binders and water were prepared by an intensive mixing in a beaker for 1 min at 20 °C, and the obtained samples were immediately put into the calorimeter. The hydration reaction was carried out with the initial hydration temperature of 20 °C.

Results and discussion

Characterization of starting materials

High-purity commercial mono-modal reactive alumina powder (PFR, d 50 = 0.5 µm) from Alteo was used as raw material. XRD results showed that no peaks except α-alumina (corundum) were found in the XRD spectra. Particle size distribution of the synthesized calcium zirconium aluminate revealed the most of the powder is in micron level and around 50 vol% of the material was below 21 µm. The XRD study obtained from the synthesized sample revealed that Ca7ZrAl6O18 was identified as the major crystalline phase as well as very small peaks related to impurity of CaZrO3.

Calorimetric studies

The heat evolved values versus curing time are given in Table 1, the heat evolved Q(t) = f(t) is shown in Fig. 1, and the heat evolution curves dQ/dt = f(t) are shown in Fig. 2. The kinetics of hydration of the binders is shown in Fig. 2. The figure shows the rates of heat liberation (dQ/dt) during the first 12 h after mixing with water as a function of temperature. The calorimetric curves shown in Fig. 2 exhibit one major feature, i.e., an intense heat peak, which occurs immediately after mixing with water. This peak is referred to hydration of Ca7ZrAl6O18 when present alone (100C7A3Z) or in admixture with reactive Al2O3 micropowder (25A-75C7A3Z, 50A-50C7A3Z, 75A-25C7A3Z) and is associated with both wetting and reaction in which an anhydrous solid is converted to the hydration products [14]. The reason for this is because the calorimetric curve of alumina (100A) has only a mixing peak due to the initial wetting of the powder. The high heat evolution, accompanying the formation of hexagonal hydrates, raises the temperature of the hydrating binders appreciably. It would be reasonable to assume that the acceleration effect temperature on the conversion reaction of these hydrates to cubic hydrate (C3AH6) will occur. As shown in Fig. 1 and Table 1, with increase in reactive Al2O3 micropowder in the binder mass, the total amount of heat released during hydration decreases considerably. Nevertheless, the presence of reactive alumina in the binder mass can facilitate the heterogeneous nucleation of C–A–H phases.

Table 1 The total amount of heat released during hydration of binders versus curing time in J g−1
Full size table
Fig. 1
figure 1

Plots showing total heat evolution in J g−1 at 20 °C as a function of time for the first 12 h of hydration of reactive alumina, calcium zirconium aluminate and their binary mixtures

Full size image
Fig. 2
figure 2

Calorimetric curves plotting the rates of heat evolution in J g−1 h−1 by reactive alumina, calcium zirconium aluminate and their binary mixtures as a function of time when reacted with water for 12 h at 20 °C

Full size image

X-ray diffraction (XRD) analysis of hardened binder pastes

Figures 3 and 4 show the XRD patterns of the hardened binder pastes without (100C7A3Z-H) and with reactive alumina powder (25A-75C7A3Z-H, 50A-50C7A3Z-H, 75A-25C7A3Z-H) cured at 20 and 50 °C, respectively. It is observed that the products obtained by hydration of Ca7ZrAl6O18 at 20 °C are amorphous in nature and are characterized by the low-intensity XRD peaks; however, by XRD analysis were detected new peaks corresponding to tetracalcium aluminate hydrate, C4AH19 (Fig. 3). The quite broadened and low-intensity XRD peaks all suggest that C4AH19 is rather poorly crystalline. With increasing curing temperature at 50 °C, intense C3AH6 peaks were found, due to the high crystallinity of this phase after the thermal treatment. It is observed that Al(OH)3 is not amorphous in nature and characterized by the sharp peaks (Fig. 4). Different hydration degrees were seen for the hardened binder pastes cured at 20 and 50 °C (Figs. 3, 4). Absence of any major peak belonging to Ca7ZrAl6O18 suggests that transformation of this phase is complete in the hardened binder paste containing 50 mass% Al2O3 cured at 50 °C (50A-50C7A3Z-H-T50), while at the lower curing temperature remains present in the unhydrated form (50A-50C7A3Z-H-T20) (Figs. 3, 4). The unhydrated Ca7ZrAl6O18 residue could be explained by a surface coverage of the binder grains by early hydration, which impeded further Ca7ZrAl6O18 hydration. Nevertheless, no dormant period caused by a hydrated layer was occurred in the heat evolution curves (Fig. 2), as shown by previous calorimetry studies. Ca7ZrAl6O18 reacted with water, forming calcium aluminate hydrates and its accompanying hydration products, i.e., the different forms of aluminum hydroxide and CaZrO3 [14–17]. Free α-Al2O3 was also detected in the hardened binder pastes by XRD and becomes dominant with increasing the content of reactive alumina.

Fig. 3
figure 3

X-ray diffraction patterns of Ca7ZrAl6O18 without (100C7A3Z-H-T20) and with reactive alumina (25A-75C7A3Z-H-T20, 50A-50C7A3Z-H-T20, 75A-25C7A3Z-H-T20) and hydrated for 32 days at 20 °C

Full size image
Fig. 4
figure 4

X-ray diffraction patterns of Ca7ZrAl6O18 without (100C7A3Z-H-T50) and with reactive alumina (25A-75C7A3Z-H-T50, 50A-50C7A3Z-H-T50, 75A-25C7A3Z-H-T50) and hydrated for 32 days at 50 °C

Full size image

Spectroscopic analysis (FT-IR) of hardened binder pastes

IR analysis was applied to supplement XRD data and more fully characterize hydration products formed in the hardened binder pastes [20]. It was capable of differentiating poorly crystalline amorphous calcium aluminate hydrates as readily as crystalline phases which was a limitation of XRD method, beside C4AH19 and C3AH6 previously found in XRD measurements. The FT-IR spectra of the hydrated Ca7ZrAl6O18 without and with reactive alumina cured at 20 and 50 °C are shown in Figs. 5 and 6, respectively. The spectra of hydrated pastes prepared using different curing temperatures showing different absorption bands indicate not the same hydration products for these samples. FT-IR spectra of at ambient temperature metastable hydrates, CAH10, C2AH8 and C4AH19 showed similarities, as hexagonal plate-type structure, usually overlapped if more than one hexagonal hydrates coexist, but these were considerably different from the spectra of the hydrogarnet or cubic phase (C3AH6).

Fig. 5
figure 5

FT-IR spectra of the hydrated form of Ca7ZrAl6O18 and hydrated binary mixtures of Al2O3/Ca7ZrAl6O18 cured at 20 °C

Full size image
Fig. 6
figure 6

FT-IR spectra of the hydrated form of Ca7ZrAl6O18 and hydrated binary mixtures of Al2O3/Ca7ZrAl6O18 cured at 50 °C

Full size image

The FT-IR spectra of the hydrated pure Ca7ZrAl6O18 paste and cured at 20 °C (100C7A3Z-H-T20) allowed identification of the hexagonal hydrates CAH10, C2AH8 and C4AH19. Broad band in the absorption range of 4000–3000 cm−1 was due to the O–H groups in calcium aluminate hydrates and alumina gel. The IR spectra of CAH10 have a very broad and intense band due to hydroxyl vibration in the 3550–3400 cm−1 region, with maxima located near to 3500 cm−1. The peaks located at 3622, 3528, 3472, 1645, 584, 534 and 420 cm−1 were probably due to C2AH8 [21], but some of these absorption bands supported the assumption of the presence of C4AH19 hydrate. Aruja [22] discussed the solid solution formation between tetracalcium aluminate hydrate and di-calcium aluminate hydrate. Hence, it is difficult to specify the characteristic vibration bands of C4AH19. A very weak band at 1645 cm−1 corresponding to H2O deformation mode supported the evidence for the presence of hexagonal hydrates. This is the main difference between cubic phase and the hexagonal ones. The IR spectra also present OH-free band at 3676 cm−1 due to the formation of stable C3AH6 phase. Others fundamental bands due to the stretching and bending vibrations of the Al–O in the AlO6 groups appearing at about 802, 525 and 412 cm−1 identified in the Ref. [23, 24] could overlap with the hexagonal phase absorption bands presented in our experimental spectrum. Nevertheless, the strong band at about 534 cm−1 appearing in the spectrum Fig. 5 is an essentially “pure” or “isolated” AlO6 octahedra vibration [25]. According to FT-IR analysis results, it was confirmed that both the hexagonal and cubic hydration products coexist under these conditions. It has been clearly shown that these results are in good agreement with the microcalorimetric results. The heat generated by the Ca7ZrAl6O18 hydration has raised the temperature of pastes and accelerated conversion reaction causing coexistence of both unstable calcium aluminate hydrates and thermodynamically more stable C3AH6. Moreover, the absorption peaks appearing at 3622, 3528 and 3472 cm−1 due to O–H stretching vibration in gibbsite supported the evidence for the presence this compound. Another gibbsite bands, i.e., 1020 cm−1 coming from O–H bending vibration and at 534 cm−1 due to AlO6 vibration, are in very good agreement with that found in the literature [23, 25]. It was found that both hexagonal and cubic hydration products are susceptible to carbonation and can react with atmospheric CO2. The stretching vibrations of CO 2−3 in carbonates were observed at 1416 cm−1 (Fig. 5). The bands at 878 and 862 cm−1 are also calcium carbonate bands. The presence of a broad band centered at 733 cm−1 confirmed that still some unhydrated Ca7ZrAl6O18 was also present. In mixtures reactive alumina/Ca7ZrAl6O18 cured at 20 °C (25A-75C7A3Z-H-T20, 50A-50C7A3Z-H-T20, 75A-25C7A3Z-H-T20), the modes corresponding to the α-Al2O3 coexisting with calcium aluminate hydrates can be identified. According to Tarte [25], the most characteristic feature of the IR spectrum of corundum being the strong absorption bands near 650 and 600 cm−1. The IR spectrums of systems with reactive alumina contain absorption band at about 453, 492, 594 and 640 cm−1 caused by α-Al2O3 (Fig. 5). The absorption bands observed in the ranges 4000–3000 and 1750–1580 cm−1 considered as the characteristic peaks of hexagonal calcium aluminate hydrates gradually disappeared with increasing reactive alumina content (25A-75C7A3Z-H-T20, 50A-50C7A3Z-H-T20, 75A-25C7A3Z-H-T20—Fig. 5). Also, the hydrated Al2O3/Ca7ZrAl6O18 mixtures were less susceptible to the carbonation process due to decreasing the amount of C–A–H phases.

The shapes of FT-IR spectra for the hydrated Ca7ZrAl6O18 without and with reactive alumina cured at 50 °C (Fig. 6) are noticeably different. They clearly show that at high temperature, noticeable sharpening of the bands occurs. The strong sharp bands appearing in the 3670–3400 cm−1 are due to valence OH vibrations of the water molecules in the cubic C3AH6 hydrate and aluminum hydroxide, Al(OH)3. The strong peaks at 3661 and additionally at 519 cm−1 correspond to C3AH6, which was formed when calcium zirconium aluminate was subjected to hydration reaction at elevated temperature (50 °C). The 3665 cm−1 band is distinctive for this phase while the strong 519 cm−1 can be assigned to the vibration of AlO6 group in the hydrogarnet structure [25]. Considering the data from spectra collected during the present investigation, it was possible to indicate the absorption bands of aluminum hydroxide, gibbsite at 3626, 2526, 3460, 1022 and 961 cm−1. The broad peak located at 793 cm−1 supported the presence of unhydrated Ca7ZrAl6O18 in the hydrated pure paste. In the spectra for the hydrated pure Ca7ZrAl6O18 (100C7A3Z-H-T50) and the Al2O3/Ca7ZrAl6O18 mixture containing 25 mass% reactive alumina (25A-75C7A3Z-H-T50) (Fig. 6), there is a peak at 1637 cm−1 caused by the H–O–H deformation band and provides spectroscopic evidence of the presence of secondary hexagonal hydrates. Through IR spectroscopy, it can be seen that hydrated samples cured at 50 °C were also strongly affected by atmospheric CO2.

Analysis of thermal dehydration of hardened binder pastes by DTA and TG

DTA–TG is considered one of the most powerful tools in the investigation of the hydration of CAC pastes [16, 26]. Curves have yielded useful results on the calcium zirconium aluminate hydrated for various lengths of time under different conditions [14–16]. Amorphous or poorly crystallized compounds not easily identifiable by XRD give characteristic thermal peaks. Curves of the Ca7ZrAl6O18 and the mixtures Ca7ZrAl6O18/reactive Al2O3 for 32 days at 20 °C and even at elevated temperatures (50 °C) obtained in this study are shown in Figs. 7 and 8. DTA curves of Ca7ZrAl6O18 without and with reactive alumina cured at 20 °C (Fig. 7) show three endothermic effects corresponding to the stepwise dehydration of pastes. There are some differences regarding the type of calcium aluminate hydrates that forms initially in Ca7ZrAl6O18 paste, as it has been discussed in Ref. [14]. The results do not seem to be consistent, and much more systematic work is needed to establish the products in the all stages of hydration of Ca7ZrAl6O18 at 20 °C. The CAH10 decomposed at 88 °C was reported to be the preferred product. The 32-day hydrated compound (100C7A3Z-H-T20) gives three peaks at 123, 188 and 293 °C due to the stepwise dehydration of both amorphous and crystalline hydration products AH3-gel, CAH10, C2AH8, C4AH19, Al(OH)3 and C3AH6. Ukrainczyk et al. [21] found three endothermic effects corresponding to the stepwise dehydration of C2AH8 at about 110, 175 and 300 °C. The thermal effect at about 120 °C in hydrated pastes in of great interest in the interpretation also dehydration reaction of CAH10. The endothermic peak below this temperature may be partially due to the presence of the amorphous AH3-gel. The two endothermic effects at 188 and 293 °C (Fig. 7) occur very close to the hexagonal phase studied by Ramachandran and Feldman [27] showing two typical endothermal peaks at 195 and 280 °C owing to dehydration of C2AH8 and C4AH13, respectively. Considerations on the stability in the system CaO–Al2O3–H2O conducted by Lothenbach et al. [28] indicate that at ≥88 %, relative humidity C4AH x is stable as C4AH19 but dehydrates to C4AH13 at ≤81 % relative humidity. C4AH19 is also considered as phase structurally derived from C4AH13 hydrate by the addition of an extra layer of H2O molecules. It should be noted that the structural changes of crystalline C4AH19 (Fig. 3) probably occurred with increasing temperature during the DTA–TG measurement and the C4AH19 phase dehydrates to the C4AH13 phase during this drying procedure. An endothermic peak at about 293 °C represented the presence of the remainder C4AH13. This DTA peak may be also due to the formation of Al(OH)3. The mixtures reactive Al2O3/Ca7ZrAl6O18 which had been hydrated at 20 °C indicated endothermic effects for the expulsion of water at the same temperatures found for the hydrated Ca7ZrAl6O18 (Fig. 7). Only, the shift in the first peak temperature may be due to the presence of AH3-gel that forms initially in the reaction of fine-particle alumina and water. Recrystallization of dehydrated calcium aluminates was manifested by sharp exothermic reaction in the range 920–950 °C in the mixtures of Al2O3/Ca7ZrAl6O18 (Fig. 7). It was concluded that in pastes of Al2O3/Ca7ZrAl6O18, reactive alumina favor the nucleation and crystallization of calcium aluminates.

Fig. 7
figure 7

DTA curves of Ca7ZrAl6O18 without (100C7A3Z-H-T20) and with reactive alumina (25A-75C7A3Z-H-T20, 50A-50C7A3Z-H-T20, 75A-25C7A3Z-H-T20) and hydrated for 32 days at 20 °C

Full size image
Fig. 8
figure 8

DTA curves of Ca7ZrAl6O18 without (100C7A3Z-H-T50) and with reactive alumina (25A-75C7A3Z-H-T50, 50A-50C7A3Z-H-T50, 75A-25C7A3Z-H-T50) and hydrated for 32 days at 50 °C

Full size image

The pastes cured for 32 days at 50 °C exhibited an endothermic effect at about 320–340 °C (Fig. 8) which was attributed to dehydration of the binder phase C3AH6. The formation of stable cubic phase (hydrogarnet) by the interaction of Ca7ZrAl6O18 with water at 60 °C has been studied in the literature [15]. Nevertheless, the DTA curves presented in this paper showed that the endothermic peaks shift toward higher temperature. The mixes different in reactive alumina content formed a more ordered cubic phase C3AH6. The possibility of influence of alumina on the nucleation process of the hydrates has received some attention in recent years [29, 30]. This is particularly evident for the paste with 75 mass% Al2O3 (Fig. 8) where an additional endothermic peak at 286 °C representing the presence of Al(OH)3 is observed.

SEM/EDS observations of hardened binder pastes

Addition of reactive alumina influences both the hydration behavior of Ca7ZrAl6O18 in binder systems, as previously highlighted in XRD, FT-IR, DTA–TG discussion, and morphology of the formed hydrated products of hardened binder pastes (Figs. 9, 10). Figures 9a and 10a present SEM micrographs of neat Ca7ZrAl6O18 pastes hydrated at 14 days at 20 and 50 °C, respectively. Figures 9b and 10b showed SEM micrographs of hardened Ca7ZrAl6O18 pastes containing 25 mass% reactive Al2O3 hydrated for 14 days at 20 and 50 °C, respectively. Figure 9a showed that the micrograph is composed of dense structure of amorphous gel and poorly crystalline hexagonal calcium aluminate hydrates. The micrograph of Ca7ZrAl6O18 paste containing of fine Al2O3 particles cured under sealed conditions at 20 °C (Fig. 9b) indicates that a large amount of more crystalline form of C–A–H phase was formed. In Fig. 10a, one can observe a very dense structure in the hardened matrix. The main hydration products of the Ca7ZrAl6O18 matrix formed at 50 °C are fully crystalline hydrates C3AH6 and Al(OH)3 having close structure. Figure 10b shows a mixture of well-crystalline hydration products and calcium hydroaluminate binding phase formed in the hardened Ca7ZrAl6O18 paste containing 25 mass% reactive alumina. SEM results show that considerable amounts of amorphous products were produced, especially during the hydration of both the Ca7ZrAl6O18 phase at 20 °C and the Ca7ZrAl6O18/reactive Al2O3 mixtures cured at 50 °C.

Fig. 9
figure 9

Morphologies of hydrated Ca7ZrAl6O18-based materials cured at 20 °C: a the reference, b Ca7ZrAl6O18 containing 25 mass% reactive alumina

Full size image
Fig. 10
figure 10

Morphologies of hydrated Ca7ZrAl6O18-based materials cured at 50 °C: a the reference, b Ca7ZrAl6O18 containing 25 mass% reactive alumina

Full size image

Conclusions

  1. 1.

    The hydration heat of the binder containing reactive alumina is smaller than that of the mono-phase hydraulic binder based on calcium zirconium aluminate (Ca7ZrAl6O18) while binding properties of the two-component binder systems were maintained.

  2. 2.

    Curing temperature affected the hydration products formed in the hydration products of both Ca7ZrAl6O18 and Ca7ZrAl6O18/Al2O3 blends greatly. At the curing temperature 20 °C, X-ray amorphous hydrates were mainly formed during the Ca7ZrAl6O18/Al2O3 binder’s hydration process and the water-rich calcium aluminate hydrate, i.e., C4AH19 was the only detected crystalline hydration product. At the elevated curing temperature 50 °C, the formation of predominant crystalline Al(OH)3 and the C3AH6 phases occurred.

  3. 3.

    Aluminum oxide micropowder influences the hydration behavior of Ca7ZrAl6O18 facilitating the nucleation and growth of hydration products and lead also to differences in the microstructure of the hardened pastes. It may be concluded based on DTA measurement that reactive alumina favored also recrystallization of dehydrated calcium aluminates.

  4. 4.

    Accompanying hydration products, i.e., refractory binary compound—calcium zirconate, CaZrO3, was detected at both 20 °C and elevated curing temperatures.

  5. 5.

    Hence, due to this unique combination of properties, Ca7ZrAl6O18/Al2O3 binders can be used in many types of new refractories including unshaped mixtures.