AFM Artifacts

Post #1 – Equipment Error

As a researcher, I work with a lot of equipment on a regular basis. In my lab, I use lasers, spectrometers, cameras, cryostats, and several types of microscopes. But sometimes, my equipment doesn’t quite behave the way I’m expecting.1

This is an image of the cap layer of a quantum dot sample (the dots are buried underneath). You can see that our cap layer seems to be made up of ‘scales’ of material instead of being perfectly smooth. Note the scale bar – some of these bumps are so small, a normal optical microscope can’t see them! Instead, my lab uses an Atomic Force Microscope (AFM) to “see” what our samples look like.2

Here’s another AFM image of a similar sample. Notice how it looks kind of fuzzy compared to the first image? Now, it’s possible that the surface of this sample really is just shaped a little differently… but I don’t think so. I think the fuzziness is an artifact caused by some dust gumming up my microscope. How can I tell?

To understand why an AFM image might be wrong, you first need to understand how it’s generated. Here’s a diagram of how an AFM works:


Click image for source.

As the tip passes over the sample’s surface features, the cantilever will flex up and down, and the laser will hit a different spot on the photodiode. The computer interprets the photodiode signal, recording the tip height at each point on the sample.


Click image for source

Here’s a Scanning Electron Microscope image of an AFM tip. These tips are extremely sharp and pointy – the ones used in my lab have a radius of curvature of about 8nm. For comparison, a very sharp knife has a radius of curvature of about 4um – 500 times duller than an AFM tip!

Now, what do you think might happen if the tip wasn’t as pointy? What if something changes – like the tip slams into something hard and breaks, or picks up a piece of dust? The software can’t tell what state the tip is in. It’ll keep generating images assuming everything’s fine. Your AFM will be trying to take a height measurement every, say, nanometer – but the damaged tip might cause you to only get new information every few nanometers, or to think you’re looking at a gentle slope when it’s really a steep cliff.


Click image for source

Now that you know how an AFM creates an image, let’s look at a couple sample images and try to figure out whether we’re looking at an interesting feature, or an ugly artifact.


See how the first picture has triangles all over it? Also, notice how there’s sort of a minimum feature size, and you don’t see many things smaller than that? It’s possible that the sample just happens to be very uniform, but it’s very unlikely, especially when you consider the specific sample we’re looking at (InAs quantum dots on top of GaAs – crystal structure dictates that these features should be almost circular, definitely not triangular).

Now let’s look at the second picture. It may be hard to believe, but this is actually a picture of the same sample, just taken with a new AFM tip. The triangles are gone, and you can see hundreds of tiny, distinct dots with an underlying scale pattern. This picture verifies that the first picture was created with a broken tip. It’s a good thing we replaced it!

Now back to our original pair of pictures:

8 9

The logic for this one is a bit… fuzzier. If there is dust on the tip, it’s so small and soft compared to the surface features that it’s not leaving a signature shape all over the image. For dusty-tip images, I usually recognize them based on experience. But every so often, I see some incontrovertible proof:


In this case, the dust fell off the tip as the tip scraped across the sample (these AFM images were drawn from bottom to top). Yes, this does mean you can clean your AFM tip by just letting it run for a while (so long as the surface it’s scanning is clean!).

AFMs are used in a lot of fields, so even if you aren’t doing Quantum Optics, you might run into one. If you do, keep these artifacts in mind as you look at your images – you don’t want to have to tell your advisor that the awesome shape you fabricated was actually just a broken AFM tip!

(Are you curious about the ripples at the edge of each island? That’s another type of artifact, but not one caused by a broken tip. Stay tuned!)

To be continued…


GPSC_Bio_Pic_CroppedJasmine Sears – Graduate Student, University of Arizona

Jasmine Sears is pursuing her Ph.D. in Optical Sciences at the University of Arizona, having previously received her B.S. from Caltech. She works in the Quantum Nano-Optics of Semiconductors group under Professor Galina Khitrova, growing and characterizing quantum structures. She is also active in student government as the current Assembly Chair of the Graduate and Professional Student Council and an Outreach Co-chair of the Student Optics Chapter. Her interests include explaining cool science phenomena and making chocolates.

Comments are closed.