Page 509 - Handbook of Thermal Analysis of Construction Materials

Section 9.0 - Industrial Applications 481

higher temperature. Satisfactory resolution has been obtained using 5–8 mg
samples in hermetically sealed aluminium pans using a 15°C/min heating
rate. The resolution of the TGA mass loss profiles is similar to those for the
DSC results. The successive mass losses are not adequately resolved (even
using the derivative trace) using an open pan system to permit quantifica-
tion. With the use of a sealed system (pan with a pinhole), the calcium
sulfate hydrate peaks can be resolved (Fig. 26). A typical calculation
involving the separated peaks in the DSC trace is as follows:
The area associated with the low temperature DSC endo-
therm is measured. A calibration curve is used to deter-
mine the amount of gypsum in the cement. Using the peak
area ratio for the dehydration of pure gypsum (found to be
3.3), the amount of heat in the second stage of the cement
dehydration associated with gypsum is determined. The
percentage of hemihydrate is then determined using a
calibration curve for pure hemihydrate and the difference
in peak area of the second peak, i.e., the area not associ-
ated with the original amount of gypsum in the material.
A similar approach for the quantitative estimation of the
calcium sulfate hydrates can be used with TGA analysis.
The pinhole arrangement in the sample pan allows the
initial water of dehydration to escape, but at a rate that
retards the second hydration step. Using suitable refer-
ence standards, a calibration curve similar to that gener-
ated for DSC can be developed and subsequently used to
determine the percent of gypsum at levels below 1% with
an accuracy of 3 to 5%. Results of this quality requires a
TG with a high sensitivity and low thermal drift because
the presence of, for example, 0.5 % hemihydrate in a 2 mg
sample, results in only a 160 µg mass loss.

Fujii and co-workers have reported the important factors governing
the quantitative analysis of gypsum by DTA. [20] These included the follow-
ing recommendations:
• Covering the sample with inert substances and bottom
packing to improve the ability of reproduction of the DTA
curves.
• Use of the fact that the peak area associated with the
dihydrate is independent of the rate of temperature change.