TEM image: why it's not always what it seems

[Jie Xiang]

Dear class:

Head down the link below for an example of how even high-level researchers, journal editors and reviewers can be fooled by TEM images:

http://www.sciencemag.org/content/296/5568/611.full

[samontoy]

Hello everyone, 

For anyone who does any type of microscopy (SEM, TEM, AFM, MFM, …), I'm quite certain you know by now that it's not so obvious what your imaging. It certainly takes some time to learn how to operate any of these tools efficiently, and yet even more time to obtain meaningful data. 

From my experience, I know that each of the above require plenty of hours of operation before one feels comfortable working with the machines alone. After all, the last thing you want to do is break something. Unfortunately, breaking an AFM or MFM tip is bound to happen… whether by mistake, not understanding the V-I readouts of the tip bias, or simply because the tip has simply reached it's shelf life. However, I believe the most difficult thing with microscopy techniques is  the interpretation of what your seeing. It helps to know what it is your attempting to see. Say you want to know how layers in a multilayer structure are separated, it is helpful to know what the materials are involved so one can discern between the contrast of each layer (Figure 1).

Figure 1. Si\Au multilayer. (Fullerton)

On other cases, imaging is simpler because you know what you should see. For instance, if you have a magnetic film with perpendicular magnetic anisotropy and it is stripe domain configuration then you should expect to see an MFM image as:

Figure 2. FeGd in Stripe Domain Configuration at RT.

On other cases, you really don't know what you are going to see…. but you expect to see something based on the model of your experiment. A few months ago, I was fooled by MFM images I was obtaining for a physics project where we attempted to see quanta flux pinning in a superconductor\ferromagnet. My initial scan (Figure 3), demonstrated I was observing something but not what I was expecting. I would expect to see some dots.

Instead I got a wave-like pattern occurring where pinning points could be in the superconductor/ferromagnet sample. When I performed a more detailed scan of the area (smaller resolution) I noticed this wave-like pattern had shrunken equally over the entire scan area. (Figure 4). This was a clear sign that something was not right… 

Figure 3. 

Figure 4.

After a few more attempts, I observed what I had been looking for (Figure 5). By the way, this took about 6 whole days to measure. 

Personally, I suggest anyone in the class to attempt to learn at least one of these imaging techniques. They certainly are very useful in research and it's very rewarding to be able to image the things you work on.

Cheers, 

Sergio