Atomic Force Microscopes (AFM) allow scientists and researchers to peer on matter surface looking at molecules. Molecules are too tiny to be visible by observing them with light: the light wavelength is longer (broader) than single molecules and the light wave just move around the molecule without being reflected like a sea wave moves around a tiny rock. In an AFM scientists use a sort of stick (a very tiny one) moving it over the surface of the material. In this movement the tip is subject to forces caused by the bumpiness of the material surface, that is by the molecules, and these forces displace the tip of the stick (a cantilever). These displacement are recorded by a laser beam (notice that a tiny displacement of the tipi of the stick, the one interacting with the molecule, leads to a much bigger displacement of the whole stick that can now be detected by the light beam).
The problem with AFM is that they are bulky and very expensive (a very cheap one starts at 30,000$, a good one is over 500,000$). It seems like a general rule: the more you are interested in something small, the bigger the tools you have to use. Think about the gigantic Super Hadron Collider that runs for kilometres underground to detect atomic particles.
Now researchers at the University of Dallas, Texas, have managed to squeeze an AFM into a chip, thus dramatically saving space and, more important, cost. Their result is presented in an article on IEEE Xplore.
Chips are not "cheap" but their overall cost does not change much with volume, meaning that if you produce many chips the unitary cost goes down rapidly. In fact, researchers at Dallas claim that the cost for a single chip may go down to a few dollars, making the overall cost for an AFM system (it takes more than a chip, you need a computer to analyse the data coming from the chip, software and tools to place and control the specimen you need to analyse) in the order of a few thousand dollars.
Now the question is who may want to buy such systems. Probably not you, nor me. According to the researchers there is a market for affordable AFM. The production of silicon chips (as an example for IoT), as an example, may benefit from the availability of AFM since it would be possible to check for faulty structures before packaging, thus saving cost. An AFM may then enter into the production lines of many chips, even those produced in small scale by small companies, focussing of niche -low volume- markets.
For sure it is another example of the progress we are making in manipulating matter at atomic level, something that will be a trademark of the next decade.