It depends on the equilibrium water content and on the thermal stability of the polymer. If the water content is 1-2% by weight, you can safely measure it using vacuum oven at temperatures where the evaporation rate is high enough but lower than the decomposition temperature of the polymer. (This is acceptable for most plastics or when checking drying conditions before injection molding). If you have very small water contents you may consider Karl Fischer titration on finely ground samples. Sometimes broad line NMR can also be used to determine mobile water fraction but it should be calibrated. The same is true for IR or NIR taken on films. If the water content is high, or very high, you should do the drying (followed by weighing) in several successive steps. It may happen that a complete removal of water may lead to the degradation of the polymers (e.g. polysaccharides).
TGA may not be the specific technique for low content moisture analysis but if you only have this technique available I would suggest you perform a dynamic experiment from ambient temperature to 110ºC and the hold the temperature at 110ºC for 20-30 min.
TGA can overestimate the amount of water if you have some other volatiles coming out of the polymer at this temperature. In order to have more accurate results you could couple your TGA to a mass spectrometer and try to isolate water from other compunds. Also, as suggested by Gyorgy Banhegyi you can also use a specific technique such as Karl-Fischer titration.
Generally, moisture affects the processability of some plastics , especially in case of PBAT , Nylon, .. etc. For these materials, defects will occur if they are processed with a moisture content outside of the recommended range. So, the moisture content for plastics can be determined by Karl Fischer (see ASTM D 6869) or by loss in weight .
Employing Aqua-track is another means of measuring moisture content and is primarily employed in plastics-processing industry. , for measuring moisture content of granular solids. Measuring principle entails reaction between water and calcium hydride to produce hydrogen. The reaction takes place in the sealed reaction vessel of the instrument, which is evacuated via a built in vacuum pump, prior to measurement. The container is then heated up to the required measuring temperature (80—200°C). The evaporating water reacts with the reagent calcium hydride to generate hydrogen; the change in pressure due to generated gas is proportional to the water content in the sample and is monitored by means of a piezoelectric transducer. Since the principle of operation is an absolute chemical method, no calibration is required for different substances to be tested. With this equipment it is possible to detect even the most minute traces of moisture (less than 0.01% H2O) in large, representative testing samples (up to 100g). In the case of lower-weight samples, the measuring range can be increased to detect very high degrees of moisture.
Moisture content can be measured by taking a certain amount of polymer then after keeping it in the oven for a few hours (for better results try to keep it in the oven for 24 hours) then after keeping the polymer immediately in a desiccator for 20-30 mins. Then weigh the polymer to find the moisture percent in the polymer by using a simple formula. This is the general procedure to find the moisture content in the polymer.