If you know nothing about the morphology of your samples, it is probably a good idea to first try SEM, partly because it has a much greater dynamic range in scan size than AFM, as mentioned by Tapas.  Although SEM images map the intensity of scattered electrons, not the real height of the surface, they do in most cases give you a good perception of surface morphology.  Once you have some SEM results, you will have a better grasp of the morphology of your samples.  If you need to know the real 3D image of or to estimate the roughness of your samples, you have to use AFM to get there.

I haven use SEM yet. But with AFM you can get up to nm resolution. AFM is much more difficult to use and time consuming than SEM. 

Luis

It depends on the area you wish to investigate. AFM has higher resolution, and gives you a quantification of surface topography. With SEM you obtain just 2D images, with qualitative information about surface appearance.

Totally agree with Lucia Ferrari. But additionally I would like to add that with AFM you can get a 3D image. With SEM you can only get 2D.

Luis

Depend on what information you want.  To get  nanomaterials particle size and shape you can use FE-SEM (Field emission SEM).  Using normal SEM we can see particles minimum size of  300 nm.  I do not know much about AFM. 

Always better to start with SEM. Much more informative, wide possibilities with different signals, fast, multiple magnifications, wide field possible, 3D software can be used.

AFM - for specific tasks, not overall morphology. Speaking about topography AFM excels in height measurement (while possible with SEM, AFM takes a more direct approach). 

Just to see the surface topography of the sample is better SEM or if you need more resolution, you can use the FEG-SEM, but if you want to know some properties of your sample, like roughness, adhesiveness or any other mechanical properties, the AFM is a better choice, cause with this technique of microscopy has a probe that stay or have contact with your sample.

So you need analise what do you want to know and which information you want obtain.

High Resolution SEM gives you lot of information about surface morphology. But it totally depends on the kind of information you want to get. If you want to get roughness AFM is more direct one. If possible it will be better if you use both

Depends on your sample and what you look for. Suppose if your sample is smooth, SEM wonot be able to show surface morphology, it just shows like plain surface. On the other hand, AFM can show the real surface morphology even if your sample is so smooth. The z resolution in AFM is extremely high. Also, SEM is not a pure surface analysis tool (it probes sub-surface also) but AFM topography (without artifacts) is a surface analysis tool.

Do not forget it depends on the nature of the sample ; you can only visualize conductive sample with SEM whereas  the AFM can also be used for insulating material. The biggest advantage of the AFM is that you can image also under liquid and  some biological sample usually do not like vacuum.

Hello my friend .

AFM is much better . if i want tell truly, SEM is not acceptable for surface topography. Also SEM is utilized for polymer grafting on poros media . but topography shows the roughness and SEM resolution hasnt  enough accuracy for nano metric roughness detection . therefore i suggest strongly to use AFM analyses. 

In SEM one detects the scattered secondary electrons emitted from the first few monolayers (typically about ~5 nm) from the surface to visualize the surface topography with a large scan area from 1mm to few nm. Meaning, you can vary the magnification from lowest (about 20x) to highest (about 200k or more). Using a modern field emission (FE) scanning electron microscope (SEM) at first glance you can have a good idea of shape and size of  the particle with a 2D image with a resolution of better than 3 nm. Moreover, by varying the tilt angle of the stage on which you mount your sample you can visualize the three dimensional look of your particle. But the quantity "roughness" (average or rms value whatever you want ) which is defined as fluctuations of height distributions with respect to a reference plane, meaning how much bumpy or zagged the sample surface is depends on the variations of the features in z direction (vertical direction). For such quantitative roughness estimation of course AFM is much superior to SEM. In AFM a sharp tip scans over your sample just like a stylus and one measures the variation of current flowing in between the tip and sample surface keeping a constant height of the tip. If your sample surface has some bumps the tip will be much closer to the surface allowing more current to flow through the tip and if there is a valley or hole /depressed part then the tip will be away from the surface and one notices a decrease in current. Thus variation of the current flowing from sample to tip which  is mapped as height variation measures the roughness. So, in principle one can test the smoothness of the sample, i.e., zig zag/ bumpy nature or roughness of the sample almost in atomic resolution (in 1 or 2 Angstrom scale for AFM operating in ultra high vacuum condition with no ambient disturbances) and with much better resolution than SEM for AFM generally operating in atmoshpheric condition. BUT ONE IMPORTANT ISSUE IS THAT MUCH OF SUCH FINE MEASUREMENT IS DEPENDENT ON THE FABRICATION OF ATOMICALLY SHARP TIP WHICH IS MAJOR CONSTRAINT FOR AFM AND ALSO YOU CAN SCAN THE SAMPLE WITH MICRON AREA. IN SEM THIS IS ADVANTAGEOUS BECAUSE THE SCATTERING OF ELECTRONs FROM THE SAMPLE GIVES YOU THE CLOSE TO ACTUAL MORPHOLOGICAL INFORMATION with much larger scan area than AFM. In principle any kind of sample that does not degrade in vacuum condition can be observed within a SEM. Of course insulating and organic samples should be coated with an ultrathin gold for observation in SEM. In a nutshell on first hand you see your sample in SEM, know the shape size and how it looks like. Then if you need to know how smooth your sample surface is quantitatively use an AFM.

if you want to view the topo, I would suggest you to find a cold FEG SEM like Hitachi s 4800 for your sample assuming it's solid. AFM can measure the surface roughness if you need.

At the first, what do you mean 'view'? What kind of structure You expect? Another question is what kind of material? Eg. AFM may measure (not preview or taking photo) the surface of ceramics and mostly grains will be visible. SEM - due to specification of electron beam - will visualize the same ceramics as pyramids. I faced this situation measuring grains of limescale. AFM is 3D measurement unit. SEM is not measurement device and sometimes grayscale doesnt represent depth. Anyway in my opinion, mentioned techniques are different and sometimes not comparative.

Found pictures - not the same/similar magnification, but will give you simple 'overview'. Anyway, as more engineer than scientist, I prefer 'mechanical' microscope - AFM.

Dear Friend,

AFM has many parameters like Amplitude, Statistical and Spatial parameters to describe the surface topography. Even by line profile, 3D and 2D image profile and Histogram analysis, you can explain sizes and counting rate of the domains of the materials and surface morphology. SEM, FESEM or In-lens FESEM also describe the morphology of the samples but gives qualitative analysis.

I must say that both have own importance when you want to know the top most surface morphology and need not  to compare for the application of materials.   

As already pointed out in the majority of the comments, it is not the question either AFM or SEM. They are practically complementary and do not give in general the same information because of different interaction physics. Both techniques have their specific advantages and disadvantages which does not mean that they can substitute each other. Again: it depends which interest do you have. It is similar to a question like: What is better for measurement of distance: a folding ruler or a micrometer caliper. If you want measure the size of your cabinet, certainly the ruler despite the worse precision and accuracy in comparison to the micrometer caliper.

here I explain this question very detailed with examples :

Wear resistance of nanocomposite coatings

http://www.sciencedirect.com/science/article/pii/B978085709211300008X

@Janusz Rebis.

How come you are trying to discuss abilities of SEM and attach picture of extreme poor quality? Sure you'll never get good results from SEM with so bad operator.

Dear colleagues, please abstain from giving advises on things you do not know. This thread is ridiculous. Please read reply of Gert Nolze  and be whenever possible quiet.

@ Vladimir Dusevich

Quality is enough for purpose mentioned: 'overview'. SEM is only preview tool for me.

If you know nothing about the morphology of your samples, it is probably a good idea to first try SEM, partly because it has a much greater dynamic range in scan size than AFM, as mentioned by Tapas.  Although SEM images map the intensity of scattered electrons, not the real height of the surface, they do in most cases give you a good perception of surface morphology.  Once you have some SEM results, you will have a better grasp of the morphology of your samples.  If you need to know the real 3D image of or to estimate the roughness of your samples, you have to use AFM to get there.

Thanks to all contributors! I'm learning a whole lot! I'm yet to assimilate all the information shared by Tapas :-) This is an excellent discussion for an expose of pros and cons of various methods. Specifically AFM & SEM. I'd need at least these two methods to calibrate and establish  NDE XRD as the preferred method for "topography" of mono-crystalline materials. The current disadvantage of XRD is the significantly lower spatial resolution of 2D XRD detectors and source strength. At best today, 2D detector spatial resolution is in the 1-5um range. But the ability to measure reciprocal space volume at such small spatial VOXEL sizes is a distinct advantage over SEM & AFM. If you mean "surface topology" instead of topography, then I'd agree AFM & SEM would be the choice. The sample prep and associated alterations to Nano structure is my primary concern :-)

The ability of making in situ measurements with little or no sample prep using XRD Topography is a huge BOON! It is also very convenient to observe topography of various (hkl) planes in situ besides the ability to monitor various surface depths non-destructively. XRR can give the average rms surface roughness with tremendous precision besides any film thickness if present. Please review the following discussion about XRD Topography and the NIST 2000 SRM when interested:

https://www.researchgate.net/post/What_is_the_best_way_to_interpret_REAL_TIME_3D_X_Y_Omega_2Theta_reciprocal_space_relative_intensity_data_from_HRXRD_rocking_curve_analyses

best  and really method  for  surface topography was SEM