Thermogravimetry (TG) or thermogravimetric analysis (TGA) measures the change in sample weight as a function of temperature. The sample is usually placed in a sample pan connected to a balance and lowered into a furnace purged with an inert gas. This gas flow is to establish a quick exchange of material and will transport any evolved volatile component from the sample. Usually, gases such as dry nitrogen or dry helium are used as purge gases. In special cases, normal air or gas with a specific water contents can be used.

Thermogravimetry is less interesting when dealing with polymorphs, as polymorphic transitions or melting events do not involve a change in chemical composition. TG can, however, reveal decomposition events, as they are normally accompanied by the evolution of volatile decomposition product such as water or carbon dioxide. Due to the same reasons, TG is especially interesting in the characterisation of solvated/hydrated crystal forms.


TGA plots
Figure 1 Example of the TG thermograms of (a) a non-stoichiometric and (b) a stoichiometric hydrate

By plotting the measured weight vs. temperature, information about the stoichiometry of the investigated solvate/hydrate can easily be extracted from the weight loss per known amount of initial sample. Thus, this technique is possibly the easiest and quickest way to establish the molecular relationship of host to guest in a solvate/hydrated crystal structure, without the necessity to grow single crystals and determine the crystal structure by diffraction methods. In fact, valuable information can be gained from TG measurements, which then help the crystallographer to solve the crystal structure of a given solvate/hydrate for example from powder data.

In addition to the stoichiometry of a solvate/hydrate, TG measurements give information about the thermal stability of the crystal form. Depending on the onset of the mass loss, the solvate/hydrate can be classified as more or less stable. Less stable forms,

Figure 2 Isothermal thermograms of the same crystal form at different temperatures.

e.g., will start desolvation immediately, which will be partly caused by the dry atmosphere the sample is placed in. More stable solvates/hydrates will show a stable weight before releasing the incorporated solvent/water. The release can be a single, well-defined step or can cascade through several steps. Thus it is possible to identify intermediate, more or less stable solvates/hydrates with lower stoichiometry than the original crystal forms.

TG measurements can be run isothermally, i.e. measuring the sample weight as a function of time at a certain temperature. These measurements are generally used to investigate the reaction kinectics of the solvent release. For this purpose, several isotherms are measured and investigated by different methods. The model fitting approach fits empirical reaction models to the measured data, in which the model is a theoretical and/or mathematical description of what is observed experimentally and can be converted into a mathematical expression (rate expression). This approach cannot be used for desolvations/dehydrations following a two-step mechanism, as only one model can be fitted to the whole reaction. Model free approaches normally only give the activation energy Ea of the desolvation/dehydration, without modelistic assumptions. Semi-model-free approaches such as the local Avrami method fit a model to every point of the desolvation/dehydration thus allowing for a change in model and an easy recognition of such changes.



A. Khawam and D. R. Flanagan, J. Pharm. Sci., 2006, 95, 472-498.