Page 131 - Handbook of Thermal Analysis of Construction Materials

114 Chapter 3 - Formation and Hydration

In cement pastes, the calcium silicate hydrate phase contains
adsorbed, as well as interlayer, water. In order to differentiate and estimate
these two types of water, Feldman and Ramachandran [58] carried out DTA/
TG of cement pastes under controlled humidity conditions. Two endother-
mal peaks emerged in samples exposed to different humidities, one due to
the adsorbed water at 90–110°C and the other at 120–150°C caused by the
interlayer water.
The mechanism of setting of cement pastes is of considerable
interest. It is generally thought that the ettringite formation is a prerequisite
for the setting of the paste. Some studies have suggested that some
hydration of C S should take place before setting can occur. The evidence
3
is based on the existence of an endothermal peak due to the dehydration of
calcium hydroxide present at the time of setting. This was observed for
several cement pastes containing various types of admixtures. [59]
The TG method has also been found to be suitable for following the
hydration of portland cement. [17] The TG of a mature paste indicates a loss
at 425–550°C primarily due to the CH decomposition, and the AFm phase
also shows a step in the same temperature range. The loss at 550°C is partly
caused by CO and partly due to the final stage of dehydration of C-S-H and
2
hydrated aluminate phase. For a typical portland cement cured for 3–12
months, the CH amount determined by thermal or XRD method is in the
range of 15–25% on the ignited basis. Thermogravimetric method (static)
was adopted by Feldman and Ramachandran [60] to construct adsorption-
desorption isotherms in cement pastes. Water content was measured at
different humidities by heating the material in a vacuum for 3 hrs at 100°C
in a TG apparatus. This was equivalent to exposure to d-dry condition. In
the d-drying method the sample is evacuated, using a trap with solid carbon
dioxide. In Fig. 32, the isotherms of cement paste exposed to different
humidity conditions are presented. Preparation A was a paste exposed to
11% RH for 6 months. Preparation B was d-dried and exposed to 11% RH.
Preparation C was obtained by heating A at 70°C for 2 hrs. The sample
contained 3% more water than preparation B. The samples were exposed to
an equilibrium RH values of 11, 32, 44, 58, 66, 76, 84, and 100% and
reconditioned from 100% to 11% RH. Scanning curves were also con-
structed at different humidities. In comparing the two isotherms for B and
C it can be observed that the sorption boundary for curve B (d-dried) has a
greater slope. A large hysteresis loop was found for preparation B. Since
preparation C was not d-dried it did not allow complete regain of water
when exposed at low or intermediate humidities. Interlayer water re-enters