We have used here TiO2 powder for the study of TGA analysis under the air . After the certain temperature level, we observed weight loss and further that it will gain rapidly. Please let me know the details.
It looks like some thing is not alright. But , may be if you give the exact details of your experiment, one may try to figure out. I cannot think of any process, which results in weight gain, unless you are going to lower temperature and due to water vapor condensation leading to weight gain.
Your result in terms of weight gain is very strange; anyway it could be related either to a misfunction of the machinary or to some residual into the environement of the crucible,such as a condensing vapor. Moreover,it could be helpful to better understand the situation if you could quantify the weight gain in terms of percentage.
We need to look at the data to get a better insight. With our TGA we also observed a "mass gain" as a strong function of time. We could proof through different tests (including no sample and zeroing the balance) that our balance has an unacceptable drift during temperature hold measurements (conditions: hold set temperature and hold for min. 30 minutes). Fortunately, we have two different TGAs, so we could also verify this malfunction which affects only one machine, the second does not show any problems.
In general terms weigh gain in TGA are due mainly to oxidation processes. One thing you could check in your experiment can be to do the same temperature program but using pure nitrogen. I bet you will not obtain weigh gain under this conditions.
The titania powder may possibly contain some unbound humidity; while being also partially hydroxilated, so that thermal dehydroxylation may occur upon heating: ≡Ti─OH + HO─Ti≡ → ≡Ti─O─Ti≡ + H2O. Substoichiometric TiO2-x may oxidate slightly by heating in air. Also to be noticed is that the TGA instrument should be calibrated at analogous experimental conditions, including selected gas and its flow rate. If air is used, its humidity content should also be kept constant.