There is a different method of image formation in SEM and TEM, so don't expect them to look the same. You do not say what the material is that you are interested in (only what the matrix looks like) or say what your sample preparation is, or give any photos so this is all a rather guesswork. However, the general thing to bear in mind is that image formation is all about contrast. There must be some difference between objects that will result in electron contrast (not the contrast control on the panel! - that is different) 

In SEM there are two main factors at work in image formation; 1 topography, 2 atomic number. Without seeing any of your images or knowing more about your samples it is difficult to guess, but I suspect that the atomic number contrast is having a greater effect than the topography. Put very simply the lower the mean atomic number of a sample under a beam, the fewer electrons are generated/detected and the resultant pixel for that region will be dark, and with scanning it will result in a darker image. So a polymer, which is atomically light, will always appear dark IF it is next to something that is atomically heavier - your sample. This is most noticeably seen in backscatter electron images, especially of flat surfaces where the topographic effect is removed, but can also be seen in secondary electron images if the electron contrast is great enough. The topographic effect is - very simply - about the angle of the sample under the beam relative to the incident beam; the higher the slope the brighter the image, the flatter the sample, the darker it will appear.

In TEM the image formation is related to the scattering of electrons by materials of differing density, again related to atomic number, but this time it is in the opposite sense. The higher the density, the darker the TEM image, e.g. material stained with heavy metals, and the lower the density, the brighter the image e.g. carbon substrate appears nearly transparent. I would expect polymer matrix in this case to be lighter than your sample.

So in order to understand the effect in your case you need to know about the chemical makeup of your sample with respect to the matrix, you need to know what image formation system is dominant, and that will largely be determined by the sample preparation and image acquisition parameters. 

It is difficult to answer the question without knowing the sample and imaging details. In SEM, you can image the surface using secondary electrons or backscatter electrons. Generally secondary electron detection gives better understanding of the topography of the surface and backscatter electron detection provides distribution of different atomic densities on the surface. To get to the answer, I would start with

1. What is the composition of my sample? E.g. different phases, different materials etc. you don’t have to know exact details. Overall idea is fine too!!  

2. Was there any surface preparation involved? Like polishing with diamond or alumina slurry or chemical etching for phase contrast enhancement?   

3. As Ian Slipper pointed, was the surface that was imaged voided or tilted or was it smooth and flat.

4. Settings on the SEM (Acceleration voltage, working distance, beam current, image detection mode). There is plenty of material available in the open literature however I found this one useful since its brief and enough for any basic SEM use - http://it.stlawu.edu/~slumic/SEMTheory.pdf. Kindly go through the pdf. It’s very helpful.  

                I think this should point you to the answer. Further you can perform EDX analysis on the different phases to get more detailed composition of the material.

Hope this helps!  

http://it.stlawu.edu/~slumic/SEMTheory.pdf

Thanks for your valuable time.

Samples are thermoset segmented polyurethanes.

ok, thanks, so ALL of your material is atomically light. Maybe it would help if you posted your images (and how you obtained them) and say exactly what you don't understand about them. Then maybe the answers can become more specific.

@Ian: I am not very familiar with the image formation in TEM, but you perhaps mixed something: density and mean atomic number. As far as I know it is mainly the last one, i.e. there is no real correlation with the density. For SEM you mentioned this but for TEM you exchanged mean atomic number by density. therefore my question: is this really correct?

As comment: hard and soft do not have anything to do with the imaging contrast... You should formulate precisely... 

Thanks, Gert

Hi Gert, I guess what I am saying with this simply is the amount of matter that the beam hits, and is scattered by, in the TEM. You can have regions of the same mean atomic number giving a variation in image brightness if it varies in thickness or density. In SEM this doesn't apply as it is only the surface of that mean atomic number that is giving information to (backscatter) image formation; variations in thickness will be meaningless in this sense.

Well, I only wanted point out that for certain reasons the differentiation between mean atomic number and density seems to be important. As far as I know and understood this topic, mainly polynomials have been used to describe the change of backscattered coefficient with the atomic number. There is obviously no physical model to describe the backscattered coefficients of a material correctly.

Often it will be claimed it is density contrast what do we see. Therefore, I checked it with some combinations,i.e. Pb and Au, where Pb has a slightly higher atomic number but only half the density of Au...and Pb really appears brighter than Au in BSE contrast. I guess (but do not know) it is identical to TEM, despite the "problems" of thickness. I agree, this is another topic.  

I think thickness is also very important. TEM samples will be much thinner than SEM samples.

I hope the volume of electron beam interaction in SEM is much higher.

So TEM will be a better technic for analyzing narrow(nm) domains.

I don't understand your last comment:

I hope the volume of electron beam interaction in SEM is much higher depending on the sample preparation methods.

What does the sample preparation has to do with volume of the electron beam interaction? 

So TEM will be a better technic for precise analysis.

You will get what the technique and the operator is able to deliver. This must not automatically mean, the result is more precise. On the other hand, you are perhaps anyway more interested in more accurate measurements. Nice, if they are still precise. 

Yes. Gert Nolze Sir.

I have changed my words.

I have a question. I have SEM Images of cellulose nanofiber alone and lead acetate entrapped on the cellulose nanofiber surface. Can I conclude after entrapping lead acetate on the cellulose nanofiber surface, the surface looks brighter?