I need to study surface of nanoparticles if form of powder. Can I use to it AFM. How can I fix them on SiO2 support for example ? It's very damadge for cantilever.
You can try suspending the particles in a solvent and either spin casting or drop casting onto the substrate. To be honest it very much depends on what the particles are even rather inert things like C60 can be very hard to image by AFM because getting lone particles is very rare.
Make sure that you know that particles are not conglomerating this can be done by choosing a clever solvent. Also you may want to check purity of the resulting sample by XPS or other technique if possible.
If they have any affinity to carbon, you may drop-cast a dispersion of your nanoparticles onto an HOPG. Then, you will find them stuck along terrace edges of HOPG. They will rather be well-separated.
If your nanoparticles are hydrophobic, you may prepare a Langmuir-Blodgett film on any atomically-flat hydrophobic substrate and imaging them with any technique available for the scale of your nanoparticles.
The only way I have ever destroyed a cantilever is by snapping it off either by dropping it or moving the tip down when I thought it was going up. If the cantilever gets damaged but is still intact you should be able to tune the tip again and recover it without problem. It is far, far more common to damage and dull the tip. This is generally done by normal usage, very large features on the surface, scanning very quickly over a rough surface, or getting junk stuck to the tip.
There are a lot of examples out there for images caused by 'bad' tip. The general things to look for are dull images, where features look blurred this caused by a tip getting something stuck to it or by being rounded to a wider point. The other common artifact is repeating shapes typically triangles or squares, this caused by the tip being clipped and the 'tip' then looks like a pyramid without a top.
Generally the method to fix this is to replace the tip. When working with soft substrates like gold or HOPG it is possible to recover a tip by intentionally crashing it and therein making a new tip that is sharp. I have never been able to get as good of a tip by crashing as compared to a new tip it but for non-publication quality images it does work fine.