COMPUTER MODELING OF MICROSTRUCTURE AND DEGRADATION

Over the past decade, the development of numerical models has provided new tools to investigate the influence of calcium leaching on the evolution of the properties of cement-based materials. A few years ago, Bentz and Garboczi¹⁴ used a cellular automaton-type digital-image-based model to study the influence of CH dissolution on the pore structure of hydrated C₃S pastes. They showed that leaching mechanisms can have a detrimental effect on the connectivity of the pore structure of the material.

The cement hydration and microstructure development model was later modified to account for the physical and mineralogical characteristics of cement grains on the properties of hydrated systems^{10, 11}. This new model, called CEMHYD3D, was used to investigate the effects of CH dissolution on the pore structure and diffusion properties of two series of hydrated cement pastes prepared at water-to-cement (w/c) ratios of 0.4 and 0.6, respectively. Hydrated microstructures were created by considering the characteristics of three commercial cements (A: CSA Type 10, B: CSA Type 50 and C: a Danish white cement). The chemical and mineralogical properties of these cements are given in Table (1).

 

Table 1: Chemical and mineralogical properties of the three cements

Oxides	Cement
	A: CSA Type 10	B: CSA Type 50	C: White
SiO2	19.78	21.45	24.29
Al2O3	4.39	3.58	1.71
TiO2	0.22	0.21	0.07
Fe2O3	3.00	4.38	0.32
CaO	62.04	63.93	68.60
SrO	0.26	0.07	0.13
MgO	2.84	1.81	0.54
Mn2O3	0.04	0.05	0.03
Na2O	0.32	0.24	0.14
K2O	0.91	0.70	0.03
SO3	3.20	2.28	2.11
LOI	2.41	0.86	1.13
Bogue
C3S	59	62	77
C2S	12	16	12
C3A	7	2	4
C4AF	9	13	1

 

The starting three dimensional microstructures were based on the measured particle size distributions of the cement powders and two-dimensional SEM/X-ray image sets in which the clinker phases had been individually identified¹⁵. The starting microstructures were then hydrated either for 2000 cycles or until achieving a degree of hydration, $\alpha$, that corresponded to the experimentally measured value (based on non-evaporable water measurements). The diffusivities of these "final" microstructures were then computed using the techniques described in the next section. These final microstructures were used as input microstructures for the leaching program. The CH in the microstructures was progressively leached as described previously¹⁴, and the diffusivity of the leached microstructures determined. In this way, the relative increase in diffusivity due to the leaching of CH from a hydrated microstructure could be assessed.