Silicate Bond Characteristics in Calcium–Silicate–Hydrates Determined by High Pressure Raman Spectroscopy

Abstract

The mechanical and thermal properties of the gigatonnes of concrete produced annually are strongly affected by the anharmonicity of the chemical bonds in its main binding phase, nanocrystalline calcium−(alumino−)silicate−hydrate (C−(A−)S−H). Improvements in C−(A−)S−H design increasingly depend on simulations utilizing a set of effective interatomic forces known as "CSH-FF", yet these assumptions have never been directly examined at the chemical bond level, and there is no guidance for their improvement. In this work, we use high-pressure Raman spectroscopy to directly measure bond anharmonicity in a representative series of C−(A−)S−H samples with varying composition and two natural model minerals, 14 Å tobermorite and xonotlite. We find that structural water molecules effectively scatter thermal energy, providing a heuristic for improving the thermal resistance of concrete. A comparison of experimental and calculated bond anharmonicities shows that a stiffer Si−O interaction would improve the transferability of CSH-FF to the thermal properties of C−(A−)S−H. High-pressure Raman spectroscopy is suggested to improve the calculations of C−S−H and to characterize other complex, nanocrystalline materials.

Plastic deformation mechanism of calcium-silicate hydrates determined by deviatoric-stress Raman spectroscopy

“…Typical standard error of the fit peak center is 1 cm -1 . Incorporation of Al in C-A-S-H does not lead to meaningful increases in peak width compared to the C-S-H sample [28,29].…”

“…Raman spectra were acquired by loading the sample into a BX90 diamond anvil cell by a standard procedure with a pressure medium, either a 4:1 volume/volume mixture of methanol to ethanol, or silicone oil [34]. It was necessary to use multiple pressure media so that the sample peaks could be observed without interference from the pressure media peaks [29]. Methanolethanol mixtures have interfering peaks around 1000 cm -1 , while silicone oil has an interfering peak around 650-700 cm -1 .…”

“…The standard ruby lines could not be used for hydrostatic pressure measurement when measuring the minerals in this study because the high laser power needed for the calcium silicate hydrate samples activates the ruby fluorescence, resulting in unacceptable convolution with the samples' weak Raman lines. Instead, the pressure in the diamond anvil cell was obtained from the Raman frequency of the top diamond culet, as described elsewhere [29].…”

“…When combining the separately obtained Q 2 bending and stretching peaks, as in the hydrostatic points of Figure 5, a linear relationship between wavenumber and pressure was assumed for one mode -in the C-S-H sample, the Q 2 bending peak, and in the C-A-S-H sample, the Q 2 symmetric stretch peak [29]. The frequency of that mode was estimated at the pressure of the other mode.…”

“…Different modes were used as the basis for the plot because the linearity between peak center and pressure was better for different modes between the two samples. These plots are available from our previous work [29].…”