Choosing the suitable solvent is critical issue in order to obtain agglomerate-free say nanoparticles. First you need to use extremely little amount of your sample with different solvents. Then, send for ultrasonication for a period between 10-20 mins. In each case, you need to take a single drop from middle level of the solvent and carefully put it on the grid. With all these conditions in mind, hope you get better view of your sample without any agglomeration, noting that you need to view the grid carefully since it got approximately 400 meshes on which a trace of your sample is sticking on.
Your question reminded me of another question asked here in RG, in which the selection of the proper solvent for TEM powder specimen preparation was deeply discussed: https://www.researchgate.net/post/Which_solvent_is_the_best_for_diluting_powder_in_preparation_of_TEM_sample. I hope it is also useful to you.
You can disperse your powder sample in acetone, ethanol or other kinds of alcohols, but you should be careful about that your powder should not be solved. Only the powder should be dispersed.
We usually embedded powder in the g2 epoxy provided by gatan and afterwards cut thin slices using a microtme. Anyway, listen to Kenneth and tell us the aim of the study or at least praparation. And always take care of agglomerates!
you can prepare by suspending little amount of your powders on distilled water (considering that your powder is not soluble in water of course) in order to improve the suspension you can use an ultrasonic bath, then using a micro pipet (Pasteur pipet) put an small drop of your suspension on the center of carbon or parlodion filmed grid, 30 minutes at environmental temperature could be enough for dried completely, after that your sample is ready for TEM.