I was trying sonicating the TiO2 in isopropanol for several hours. But the size was still more than 100nm. After the coating the size was even bigger, up to micron meters.
You need to use some surface modification of the TiO2 NPs. In the case of organic solvents (i.e. non-polar solvent) I suggest to use something like Hexamethyldisilazane (HMDS). It will be covalently bounded to the surface (the rest of HMDS is evaporatred) and protect NPs against the agglomeration. First you need to drop few drops of HMDS on NP powder and leave to react in closed bottle for few hours (a day?). Afterwards, leave open to evaporate the rest (or little heat up if necessary). Mix with organic solvent you need (with polymer if required), use for spin-coating, and enjoy you nanoparticle film.
I completely agree with Brian, you need a surfactant to make TiO2 nanoparticles soluble in organic solvent. HMDS is a good example, but any carboxilic acid with long alkyl chains will be also effective.
Thanks for the answers. I've considered adding some surfactant. Can I heat the film at lower temp. (ex: 100C) to evaporate the surfactant because I don't want residue in my film?
If you use HMDS, the liquid is evaporating fast even at room temperature without additional heating (it is siccative). Additional heating make evaporation of residues faster, of course. Usually is used 100-150 degree for few minutes to ensure that HMDS is totally evaporated. The heating is removing only the surfactant which is not covalently bounded to the NPs, but the organic envelope does not do any barrier and the thickness is less than 1 nm so you can neglect it. Only thing which is changed are the wetting conditions (i.e. the contact angle) ane the NP layer is hydrophobic. To remove the covalently bounded HMDS (or other surfactant) I strongly recommend to use the UV/ozone cleaning which will remove all additional organic residues and will not influence TiO2 NPs. This is better than heating because of the film re-ordering during the high temperature heating (required to remove covalently bounded organics).
I did not try to modify ZnO before, but it should work since it is via oxygen (covalent bond). In addition, there are some reports on ZnO modification by HMDS, such as
Really appreciate your reference. I tried your solution yesterday and it worked. Previously the ZnO NP solution in EthOH was cloudy, as soon as I added some HMDS into it and shaking vigorously, the solution became clear. I will test its electrical properties soon.