How do I join AFMS

AFMs tailored to your application
AFM basics
DME Channel on YouTube
AFM configuratorNew
  • Publications on thermal microscopy
    In the last few months a few publications about thermal microscopy in SEM-AFM have appeared, which can be found under SEM-AFM applications.

  • The new controller C32 is available
    The C32 controller offers many new possibilities.
Post-processing of AFM images

The atomic force microscope is an instrument that can precisely measure changes in altitude on an atomic scale. Often, however, AFM images should not only allow certain details to be measured, but should also look good. In most cases, this requires post-processing. The following is a brief description of how simple processing steps can improve the appearance of an AFM image without any loss of information.

Unprocessed AFM image

The image on the right shows an unprocessed AFM image of a gallium nitride (GaN) surface. You can just see the course of individual atomic layers. GaN is a semiconductor material that is used, for example, in the manufacture of blue and white light-emitting diodes. The picture was taken with our small DS 45-40 scanner in the clean room of the Institute for Applied Physics at the TU Braunschweig. The scanner is mounted in the "Igloo" platform, which stands on a normal metal table. Since the scanner is not covered, the permanent air flow and running pumps in a clean room generate a small, low-frequency noise that is noticeable in horizontal lines in the image. Furthermore, the sample is tilted by approx. 0.4 ° in relation to the scan plane. Such tilting can hardly be avoided and is particularly evident in the case of flat samples, as in this case.

Level correction

A significant improvement is achieved by approximating the sample surface through a plane and subtracting this plane from the data. As a result, the tilt disappears and the interesting structures become more clearly visible. Now, however, the low-frequency interference that can be eliminated by line averaging is also more pronounced.

Line mean

The image on the right is the result of level correction and line average. You can now see the course of the atomic layers very clearly. The "background noise" present within the atomic layer surfaces cannot have been caused by regular GaN atomic layers, since the noise is smaller than the distance between the atomic layers. It could be half layers or adsorbed atoms, which cannot be further distinguished in this way.

Line mean and plane correction are standard methods of scanning probe microscopy. They are used very often and usually bring a gain in quality. The data is often processed in this way during the scanning process, so that the quality of the result can be assessed during the scanning process. In addition, our scan software always saves the unmanipulated original data, so that you can go back to the original even after using certain image processing techniques. From the pictures shown here you can also see that you can easily achieve atomic height resolutions with the AFM even on a normal table under less than ideal conditions.

If a sample is scanned with surface structures that are not randomly distributed, as in this case, the line mean can also produce undesirable artifacts, since this function works best when all lines have roughly the same "height mean".

Influence of "lighting"

In order to bring out surface structures better, it is often worth adding a kind of lighting effect (even with two-dimensional images). Essentially, you add or subtract the change in height in a certain direction from the original image. The following table shows a comparison of "illuminated" and "non-illuminated" images and the real 3D representation, all structured and unstructured gallium nitride surfaces:

The lighting effects can give a strong three-dimensional impression, even with a two-dimensional representation. Often times you can see a lot more details and the pictures look more interesting.

With the last strip-like structure, the representation is not yet ideal. You can see the details on the surface of the stripes, but you cannot see exactly what it looks like in the deep "valley" between the stripes. Here it is sometimes advisable to simply switch off the color palette and only use the lighting effect (or swap colors in the color palette). Since the real height information is then lost, you can also switch to the real 3D display:

In the last two pictures you can see the surface structure within the valley very clearly. In addition to the nice appearance, the real 3D display has the disadvantage that you can no longer measure the proportions exactly, which is actually the point of the AFM or SPM. In this respect, you should normally avoid this display method.