This question you should ask the operator who generated these images. I highly doubt that anybody will risk to give you any advice since who knows what these images display? Another question: Why did you make these images if you don't know what do they mean or tell you?

It is impossible to make quantification on your particles because they are too concentrated and aggregate

You should dilute them more and take images from fields with only individual particles.

Then you could use ImageJ for manual or automatic quantification.

I second the above recommendation for using ImageJ for either manual or automatic quantification. 

Do you want to quantify the particles individually, or quantify the area of the total particles? Individually will be more difficult, and less accurate, but you could adjust the threshold onto the greyscale of each individual particle as per your interpretation. Alternatively, you could install the erasure tool and use that to interpret the individual particles and than quantify each separately. 

If you want to quantify the images as a total, then you can create a binary mask image and use that particle analyse method - remember to ensure you are including all particles that touch the image edges otherwise you will get zero in each case. 

For either method, you need to set the scale prior to running the particle analysis. 

Let me know if you'd like any further information on those suggestions. 

Is it the same bioglass? Is it the same magnification for the 3 images?

For me, the response is no to my questions. In the first TEM image, the particles are big because they are black, stopping all electrons. The particles in the second TEM image are smaller (very small I could say) because they are transparent for electrons.

Is it different bioglasses, which you need to compare? What is your project/question?

I agree with Alexandra:

First, you have to deposit the same concentration of bioglass on the grid and take images at the same magnification.

Second, you have to take in account their aggregation state. Can you get individual particles or not? So particles analysis or surface area analysis.

Dear Shoaib,

You should use dark field TEM image for particle size distribution. Please find this article for further information.

D. Smolen, T. Chudoba, I. Malka, A. Kedzierska, W. Lojkowski, W. Swieszkowski, K.J. Kurzydlowski, M. Kolodziejczyk-Mierzynska, M. Lewandowska-Szumiel, Highly biocompatible, nanocrystalline hydroxyapatite synthesized in a solvothermal process driven by high energy density microwave radiation, Int. J. Nanomedicine. 8 (2013) 653–668. doi:10.2147/IJN.S39299.

I recommend the check the grid preparation for TEM or SEM, so,can you use image J

if possible you may took once again TEM otherwise FESEM it is good way to show your clear morphology 

The SEM and TEM analysis of your samples show that the particles are agglomerated and they irregular shape, some of the are a sphere, also the particles have spongy structure. Those dark color are considered as agglomarated and precipitated particles.