Keywords

1 Introduction

A new interest is currently growing in the use of nano-aluminosilicates as a novel type of additive in cement-based materials [1]. These nano-materials can be classified under non-calcinated nano-clays and amorphous nano-aluminosilicates based on the crystalline structure. Moderate quantities of clay minerals at nano-particle size range in cement-based materials can nucleate the formation of calcium silicate hydrate (C-S-H) [2], improve certain fresh state properties such as green strength [3], and decrease gas and water permeability of the hardened specimens [4].

Amorphous nano-aluminosilicates are potentially seen as a new generation of nano-particle additives that can be used in cementitious materials. The interest in these materials resides in their amorphous character, which may offer the possibility of nucleating alternative hydration products to C-S-H, such as calcium aluminate silicate hydrate (C-A-S-H), especially in systems where cement is a minor component, such as high volume fly ash mixtures.

The research work reported in this manuscript analyzes the effect of a group of amorphous nano-aluminosilicates on the early hydration of Portland cement paste. The data aids in identifying the mechanism undergone by these amorphous nano-aluminosilicates in the presence of cement pore solution and the effects on the early hydration reaction.

2 Material and Methodology

The following reagent grade chemicals were used in the preparation of the sols: tetra-ethyl orthosilicate (TEOS), aluminum-tri-sec-butoxide (ATSB), aluminum ethoxide (AE), sec-butyl alcohol (C4H10O), ethanol (C2H6O), nitric acid (HNO3), sulfuric acid (H2SO4), hydrochloric acid (HCl), and deionized water (18.2 MΩ cm). The aluminosilicates sols were synthesized by mixing 1 mole of TEOS and 1.3 moles of the corresponding aluminum alkoxide, at a water to alkoxide molar ratio (r) of 194 and at mole of acid to mole of alkoxides ratios (A) of 0.22 or 0.13 depending of the nature of the aluminum alkoxide. The principal variables in the synthesis, summarized in Table 1.

Table 1 Variables of the sol–gel synthesis process
Full size table

The synthesis protocol used was a variation from the method described in Pozarnsky and McCormick, [5]. Following the synthesis, all sols were dialyzed [6]. A reactor was created to study the interaction between nano-particles and ions in the cement pore solution. A total of 167 g of cement and the corresponding 0.75 % wt. replacement of nano-aluminosilicate were immersed in the reactor in two separated Spectrum Spectra/Por 3 RC dialysis membranes at a water to cementitious material ratio (w/c) of 3, purged with N2 and stirred for 72 h. In order to simulate realistic cement pore solution conditions in the reactor, the membrane containing the nano-particle sol was immersed into the vessel once the reactor reached equivalent calcium concentration values as those in real cement paste samples [7]. After the 72 h period, the remaining nanoparticles were collected from the dialysis membrane and dried for analysis with the FTIR spectrometer.

An Empyrean Series 2 X-ray Diffraction System manufactured by Panalytical was used to examine the mineralogy of the nano-particles. The size range was determined with a Zeta Potential Analyzer ZETA PLUS manufactured by Brookhaven Instruments. Xerogels were analyzed using a Digilab Excalibur FTS 3000 Series Fourier transform infrared spectrometer. The spectra of KBr pellets with 0.3 % sample concentration were collected at 4 cm−1 resolution and co-adding 32 scans per spectrum. Isothermal calorimetry experiments were performed at 25 °C using a TAMair model TA instrument. Samples of cement paste containing 0.75 wt. % of nano-particles were prepared at water to cement ratio of 0.45.

3 Results and Discussion

3.1 Nanoparticle Characterization

The amorphous nature of the synthesized nano-aluminosilicates was revealed by the absence of sharp peaks and presence of humps in the XRD diffractograms. All the nano-aluminosilicates had the same diffractogram independent of the type of alkoxide and acid used during the synthesis, with humps at 7.4°, 26.5°, 39.8°, 65.2° and 68.1° 2θ. With the exception of the hump at 7.4° 2θ, which was associated with zeolite type mineral, the observed humps were linked to sillimanite mineral (Al2SiO5). The 26.5° 2θ is also a common hump in amorphous silicates [8].

The FTIR analysis of the samples displayed in Figure 1 provides additional information regarding the structure of the nano-particles. The tentative assignments of the principal bands in the 400–1,300 cm−1 region are summarized in Table 2. The FTIR spectrum showed typical characteristics of aluminosilicates. The Si-O-Si asymmetric stretching band in the samples was shifted toward a lower wavenumber (964 cm−1) with respect to silica (1,082 cm−1). This shift was caused by the incorporation of Al3+ into the Si4+ sites [9]. Additionally, the spectra showed a strong and broad vibration band between 493 and 690 cm−1. This region was associated with the presence of bending and stretching bands of Si-O-Al and stretching of Al-O for condensed AlO6 octahedral [9–11]. The FTIR spectrum did not show significant differences despite the type of alkoxide. Only the vibrational bands associated with the amorphous aluminosilicate backbone (1,010–1,250 cm−1) showed differences due to the type of acid used in the synthesis. This is the case for the band at 1,064 cm−1 associated with vibrations of tetrahedral SiO4 groups. Both samples, AT-57 and AE-50, synthesized in the presence of H2SO4, had a broader and more intense 1,064 cm−1 band than the other nano-aluminosilicates. These differences in the band can be attributed to the incorporation of sulfate ions by the nano-aluminosilicate [12].

Fig. 1
figure 1

FTIR spectra of the two nano-aluminosilicates synthesized series: ATSB (a) and AT (b)

Full size image
Table 2 Characteristic bands and assignments of the nano-aluminosilicates before and after 3-days exposure to simulated pore solution
Full size table

3.2 Isothermal Calorimetry Data

In all the cases, the hydration of not only the tricalcium silicate (C3S) peak but also the alumina-bearing phases was accelerated with respect to the control. The enhancement in the hydration of the alumina-bearing phases was caused by the strong capacity of this type of nano-particle to incorporate sulfates from the cement pore solution. Table 3 summarizes the magnitude of the acceleration effect in both C3S and alumina-bearing phases. The comparison based on alkoxide type revealed that the AE-based nano-aluminosilicates had less capacity to accelerate the hydration reaction then ATSB-based nano-aluminosilicates. Finally, a comparison based on the type of acid, indicated that AT-57 and AE-50 nano-particles synthesized in the presence of H2SO4 had less capacity to accelerate the hydration of both phases than the other two acids tested. One explanation for this behavior is the fact that these two nano-particles had the coarsest particle size distribution among all synthesized nano-aluminosilicates (see size distribution in Table 1). In addition, it is expected that the pre-existence of sulfate ions in the structure of these samples may mitigate the affinity for additional sulfate ion uptake from the pore solutions and therefore contribute less to the acceleration of the alumina-bearing phase.

Table 3 Summary of the effect of nano-aluminosilicates on the early hydration of cement
Full size table

3.3 Effect of Simulated Cement Paste Pore solution

As is summarized in Table 2, the broad band in the original samples located between 493 and 690 cm−1 evolved towards a more defined group of bands at 532, 617, 673, and 707 cm−1. At the same time, the Si-O-T asymmetric stretching band shifted toward higher wavelength numbers and became better resolved. These changes indicated an increased polymerization and structured order of the nano-aluminosilicate caused by the incorporation of calcium from the simulated cement paste pore solution. This re-organization of the alumina-silica network was different depending on the nature of the nano-aluminosilicates. This difference is illustrated in Figure 2, where the spectrum before and after exposure are compared for two nano-particles, AT-58 and AE-52. Those nano-particles synthesized with AE alkoxide, such as AE-52 in Figure 2, developed an intense band at 1,110 cm−1. This band was assigned to Si-O asymmetric stretching caused by a highly polymerized sheet and/or a framework structure [10], which indicated a higher propensity of AE-based nano-aluminosilicates to polymerize under cement pore solution conditions.

Fig. 2
figure 2

Comparison of FTIR spectrum of nano-aluminosilicates before and after 3-days exposure to simulated pore solution. Sample AT-58 (a) and AE-52 (b)

Full size image

4 Conclusion

Two series of nano-aluminosilicates were synthesized with different types of aluminum precursor in the presence of three acids. Both series were capable of accelerating the hydration of C3S and alumina-bearing phases in Portland cement. However, those nano-particles synthesized using the ATSB showed higher acceleration capacity, most likely due to the higher content of aluminum in tetrahedral coordination. The higher degree of agglomeration and presence of sulfate ions in the nano-particle framework were identified as potential mechanisms to justify the lower acceleration in the hydration of cement in samples AT-57 and AE-50. The exposure to simulated pore solution triggered the incorporation of calcium into the alumina-silica framework causing an increase in the polymerization in all samples and more significantly in the AE-based nano-aluminosilicate.