I have used electro-spray of nanoparticle suspensions onto aluminum foil targets with some success.

Agglomeration is always associated with wet synthesis techniques unless surfactants are used. Please check my paper "Microwave synthesis and characterization of nanocrystalline Mn-Zn ferrites". Full text available in my profile.

I have found similar problems in SEM with bulk synthesised magnetic nanoparticles, but have had some sucess with TEM imaging of magnetic nanoparticles, as they seem to disperse better on a hydrophyllic TEM grid (the magnetic nanoparticles I have made are synthesised in aqueous solution). Depending on what they might be coated in, you can use this to help disperse them on your surface. For example, a drop of particles in an organic solvent can be nicely dispersed on water, and when the solvent evaporates, a thin film of particles can be transferred onto a substrate for imaging. It can be a bit fiddly, but it might help you to get an single layer of particles which would be much easier to image and grainsize. I hope this helps

I agree with both Surender and Johanna. You need to use either surfactant for SEM or you may try TEM without any surfactant.

Thank you all, I will try using surfactants. If that doesn't work, I will try and look into TEM imaging.

You can prepare your specimens the same way as you would prepare them for TEM (on grid with carbon film) and then place them on carbon substrate (carbon sticky tape will work) and observe them with SEM. For me it is preferred way for nanoparticles observation with SEM. Minimal background for EDS and BSE, high contrast.

Actually, I do not recommend to use SEM nor TEM for quantitative particle size analysis. These are good for getting information about particle shape, and for getting an overview how broad the particle size distribution is, but finally for quantitative analysis, you should use dynamic light scattering or Laser Doppler method.

SEM and TEM can not show you a representative and statistically useful information, it is based on luck where you are looking at. DLS or LD deliver statistically reliable information.

However, DLS does not give you the particle size in a solution, but the hydrodynamic particle size, so can significantly over estimate particle zize, especially for small particles. It is best to use a variety of techniques and look for agreement between them so they can compliment each other. TEM/SEM, DLS and XRD can all provide information on the particle size and distribution, and if the values agree well, then they are more reliable than a size from just one technique.

I agree, one should always take several techniques in parallel and compare. However, I would like to underline again that statistically relevant and reliable figures can not be achieved by SEM / TEM.

Actually statistically reliable figures can be achieved by SEM/TEM. But it all depends on the nanoparticle formation and how much pain the operator can take. I have actually once made a good histogram plot of Ni nano particle distribution using SEM.

But yes, could do it only once. The sample somehow was really good.

Thanks so much, everyone. I sadly dont have the opportunity to get DLS for the sizes I am looking at. I did XRD and calculated the sizes. SO SEM is my next best option. How many would I have to count for a statistically significant number?

I have counted around 150. Got a nice histogram plot that 81 % particles is between 30-60 nm.

Michael, I don't know how big your nanoparticle amount is, but "150 particles" (as taken by Rajib) does not impress me and would not make me feel comfortable; you have billions of NPs in your sample, 150 is just a statistically irrelevant number.

Sure, you can get a nice histogram, but how reliable is that?

I assume, Munich University will have several particle size analyzers of whatever technique, maybe you ask in polymer department?

As we can learn from statistics sample size has nothing to do with number of manufactured particles (population size). For narrow and even distribution 150 particles (or even 15) could be good enough sample size. Sample size should be determined experimentally and be sufficient to provide needed statistical power. In measuring nanoparticles the main problem is not number of particles but right specimen collection and preparation.

yes, I agree, but how do you know in SEM that you look at a representative sample size and representative particles? the magnification is too big to be able to determine whether or not the selection one has to take is good enough, at least I think so. Therefore, I would always combine with light scattering or Laser Doppler method.

Thank you for all your input. I just want to give you a quick overview on my experience.

The SEM I had the opportunity to use did not have a high enough resolution for me to get any grain-size information.

I was able to get TEM images. I have suspended my particles in Isopropanol and put them into an ultrasonic bath for half an hour. The images are usable. However there is still agglomeration and the particle count is not as high as I wished it was. But I now also have information on the actual morphology of my samples which is great.

Lastly I was able to send my samples in for measurement on a Fritsch Analysette 22 nanotec plus. The resolution should be high enough to get a decent count. The samples were suspended in water and put into ultrasonic for 30min. However agglomeration was too strong to get any usable data.

You didn't try to use any surfactant (in water), like low MW PEG?

Consider electospray. With this technique you may be able to put electric charge on your nebulized nanoparticles sufficient to overcome their attractive forces that cause them to agglomerate. A syringe pump is used to pump a suspension through a capillary tube to a heated tip where a voltage potential is applied (~3000 to 4000 V). The dispersed particles are collected on metal targets a few inches away that have the opposite potential. A thin polymeric film of some type may be needed on the target to capture the nanoparticles and prevent them from re-agglomerating.

@Bernhard Wessling. I would have, but it wasn't me who did the measurement. Next time I will try PEG. Thanks

@Thomas Schmitt. That sounds like a great idea, I don't think we have such a machine in our institute though. I will ask around, though.

Even after agglomeration, if you could still recognize and resolve individual particles/grain, then some simple image processing software may be used to get the size. You could try using ImageJ. It has some plugins for particle analysis which you can use after calibrating the image to get pixel/size. This calibration can be carried out using the scale bar obtained from SEM/TEM image.

The particle identification then depends on the image thresholding you could set at the software.