Elsevier

Cement and Concrete Research

Volume 33, Issue 12, December 2003, Pages 2037-2047
Cement and Concrete Research

Effect of hydration temperature on the solubility behavior of Ca-, S-, Al-, and Si-bearing solid phases in Portland cement pastes

https://doi.org/10.1016/S0008-8846(03)00224-2Get rights and content

Abstract

The concentrations of Ca, S, Al, Si, Na, and K in the pore solutions of ordinary Portland cement and white Portland cement pastes were measured during the first 28 d of curing at temperatures ranging from 5–50 °C. Saturation indices with respect to solid phases known to form in cement paste were calculated from a thermodynamic analysis of the elemental concentrations. Calculated saturation levels in the two types of paste were similar. The solubility behavior of Portlandite and gypsum at all curing temperatures was in agreement with previously reported behavior near room temperature. Saturation levels of both ettringite and monosulfate decreased with increasing curing temperature. The saturation level of ettringite was greater than that of monosulfate at lower curing temperatures, but at higher temperatures there was effectively no difference. The solubility behavior of C-S-H gel was investigated by applying an appropriate ion activity product (IAP) to the data. The IAPCSH decreased gradually with hydration time, and at a given hydration time the IAPCSH was lower at higher curing temperatures.

Introduction

Temperature is a key variable affecting the curing of cement-based materials because it influences both the early hydration kinetics and the properties of the hardened cement paste or concrete. Although concrete initially gains strength more rapidly when cured at elevated temperatures, the final strength is lower and the permeability is higher. Elevated temperatures also reduce the tendency for irreversible creep and shrinkage. Most of these effects can be related to the increased rate of silicate polymerization at elevated temperatures [1], [2], [3], which densifies and stiffens the C-S-H as it forms. SEM studies [4], [5], [6] have shown that at elevated temperatures the outer-product C-S-H gel is denser and does not fill the capillary pore space as effectively, and thus the microstructure is more heterogeneous. Elevated temperatures may also alter the equilibrium assemblage of solid phases in cement paste. A particularly important example is the relative stability of ettringite and calcium monsulfoaluminate (monsulfate). During the early hydration period when sulfur concentrations in the pore fluid are high, ettringite is more stable than monosulfate at ambient temperatures. As the temperature increases, monosulfate eventually becomes the favored phase due to its smaller enthalpy of reaction. Lack of formation of ettringite during initial curing can lead to an expansive and damaging condition known as delayed ettringite formation (DEF) [7].

This study reports measurements of the concentration of major elements in pore solutions from two different types of cement paste cured at constant temperatures ranging from 5 to 50 °C over the first 28 days of curing. Of particular interest were any changes in the solubility behavior of C-S-H gel with temperature, and the changes in the calculated relative stability of ettringite and monosulfate as a function of temperature.

The hydration of cement produces solid phases in intimate contact with a pore system, and the composition of the pore solution can thus provide information about the hydration products. Comparing the ion activity product for a solid phase with its equilibrium solubility product gives the degree of over or undersaturation, and tracking this saturation level over time, or comparing the levels in different types of paste, can be quite useful. Dilute suspensions of cement powder are not representative of normal cement paste and concrete, and thus hardened cement pastes should be used. Pore fluid can be extracted from cement paste after set by applying pressure to crushed paste with a steel die [8], [9]. A number of studies of this type have been conducted (e.g., [10], [11], [12], [13], [14], [15]), and the basic solubility behavior in the Na, K, Ca, S, and hydroxyl system has been established. Solutions become modestly supersaturated with respect to portlandite a few hours after mixing, and then saturation is approached slowly over a period of several days. Gypsum, which is present in the starting cement, is saturated during the first few hours of hydration and then undersaturated once it is consumed through reaction with C3A to form less soluble calcium sulfoaluminate phases. The details of this process depend on the presence of free lime, the time to set, the presence of mineral admixtures, and other factors.

Few studies extending past the first few hours have reported concentrations of aluminum and silicon, which occur at such low concentrations that they are difficult to measure. Their neglect, while it does not significantly affect calculations for portlandite and calcium sulfate, prevents saturation levels of the calcium sulfoaluminate phases and C-S-H from being calculated. Exceptions include studies by Lawrence [13], Xue et al. [15], and recent work by the present authors [16], all conducted near room temperature. The latter study [16] reported elemental concentrations of Ca, S, Na, K, Al, and Si during hydration of two types of Portland cement at 20 °C, along with the saturation levels with respect to ettringite (AFt) and calcium monosulfate (AFm), as well as a few different ion activity products for C-S-H. The present study extends this work to different temperatures. In addition, data from the three studies noted above are reanalyzed to try to obtain a more complete picture of the solubility behavior of C-S-H gel in hydrating cement paste.

Section snippets

Experimental

Type I ordinary Portland cement (OPC) pastes were mixed at a water-to-cement ratio (w/c) of 0.4 by weight and hydrated under sealed isothermal conditions at 20, 30, 40 and 50 °C. White Portland cement (WPC) pastes were mixed at w/c=0.5 and hydrated at 5, 20, 37, and 50 °C. The composition of the cements are listed in Table 1. The WPC has significantly lower alkali content than the OPC, and this, along with the higher w/c used to hydrate the WPC, led to significantly lower alkali contents and

Thermodynamic analysis

Calculating saturation levels requires measuring the elemental concentrations in the pore fluid and then performing a thermodynamic analysis to obtain ionic activities. In electrolytes such as cement pore solutions, each element exists as more than one ionic species. These species can be simple charged ions such as Ca+2, or ion pairs such as CaOH+ and CaSO4°. Calculating the equilibrium ionic activities from the elemental concentrations is an iterative process that is facilitated by the use of

Elemental concentrations

Fig. 1 shows the total alkali (Na+K) concentrations in the OPC and WPC pastes at each temperature. The alkali concentrations in the OPC paste are significantly higher than in the WPC paste, as expected. In both pastes, alkali concentrations increase with time as the hydration reactions consume liquid water and concentrate the pore solutions. This process occurs more quickly at higher temperatures, leading to differences in the alkali concentration at earlier times. At later times the alkali

Summary

The elemental concentrations of Ca, S, Al, Si, Na, and K in the pore solutions of normal w/c OPC and WPC pastes hydrated at temperatures in the range of 5–50 °C were reported for the first time. Changes in the saturation levels during the early period of hydration occurred more rapidly at higher curing temperatures, as expected. At all temperatures, portlandite exhibited small and decreasing levels of supersaturation, and gypsum was saturated during the first few hours and greatly

Acknowledgements

This work was supported by the National Science Foundation under contract CMS-007-0922.

References (30)

Cited by (111)

View all citing articles on Scopus
View full text