I have various bimetallic nanoparticles which have formed alloys. I then used NTA and TEM to characterize the size of my nanoparticles. The TEM shows that the size of nanoparticles is considerably less than the NTA. Can anybody help explain the difference in size?
You don't mention how homogeneous your solutions are or how much "considerably less" is. As summarized in this paper, "Under ideal conditions, particle size analysis by DLS and NTA both can be accurate to 2%, but heterogeneous samples can have errors an order of magnitude larger." (See references in that paper for details).
It's also important to keep in mind that you are comparing techniques which measure samples in very different ways and under very different conditions. TEM measures individual particles in a dried (and often treated) state in a vacuum, while NTA measures an ensemble of particles dissolved in solution. In each case, the size distribution is often just as important as the so-called "size". If the size-distribution of your measurements overlap, perhaps there's no conflict after all.
Also, if your particles have some sort of coating which Au nanoparticles, for example, often do, then you might expect differences in the apparent size measured in these two different states.
Using controls is a good way to get around some (but not all) of these problems, but then the same questions need to be raised about the control particles and any difference in their behavior from the particles under study.
So, as is usually the case in well-done science: "It's complicated..."