Scanning probe microscopy
In this kind of microscopy a needle with a very sharp is moved across the surface in meanders (line by line) and a computer reconstructs an image from the interaction of the needle tip with the surface. Sometimes the tip quickly swings up and down while slowly meandering over the surface to take the image. Depending on the type of interaction used (electrical mechanical or other) scanning probe microscopy is further classifies in subcategories.
The two main types of SPM are:
- STM scanning tunnelling microscopy
- AFM atomic force microscopy
In both cases STM and AFM one would like to have only one single atom at the tip. But this is wishful thinking. In reality the average tip looks more like a rounded deformed mountain top. Images have been taken (with big transmission electron microscopes) that show that. See external links. Metallic STM tips (made from etched wire) are usually sharper than AFM tips etched from e.g. a silicon nitride wafer.
Contents
Electrical SPM – usually called: scanning tunneling microscopy STM
A very fortunate physical phenomenon makes it so that rather blunt tips are not too critical. When approaching the (electrically conductive) surface with a (conductive) tip that is connected to a different electric potential the current that arises makes a sharp (exponential) increase over a faction of an atomic diameter. This effectively hides all the atoms behind the ones at the very front. The reason for this behavior is that when the electrons (described as waves) seep into a space region where the potential energy is higher than their own energy (the vacuum between tip and surface) the wave becomes takes the shape of an exponentially decaying function. This is the infamous "tunneling effect". No magic here, just math describing physics usefully accurate.
Due to the direct physical contact between tip and surface (which is not present in electron microscopy) one can actually manipulate stuff like molecules and even atoms on that surface beside just merely imaging it.
The core principle of the electrical SPM (called scanning tunneling microscopy STM) is actually so low tech that that resolving of single atoms has been repeatedly archived by hobbyists in a domestic environment (living-room). No joke. This was limited to the easiest surfaces to image though. No shoving atoms around here. This is reserved for more professional systems.
Somewhat miraculous seems that even low tech damping systems can in fact suffice to prevent the tip from instantly crashing into the surface and destroying itself. Still one should refrain from coughing too loud when located in the in the same room as the SPM microscope.
Since images are taken line by line and the moving parts are so much larger than the imaged resolution taking an image can take considerable time (minutes to hours). At any time (even witout external vibrations) the tip can change its shape and thereby change or degrade the image. A very unnerving situation.
To take these images one needs to move a needle just fractions of the diameter of an atom in a decently controlled fashion. Again somewhat miraculous there really are actuators that make this quite easy: "piezo-actuators". In some cigarette-lighters one can find a pieces of quartz. When they are hit and quenched together a very tiny bit they produce a very high voltage (creating the spark for ignition). The piezoelectric effect. Operated in reverse when such a crystal is exposed to just a very tiny voltage it contracts/expands/shears a just a very very tiny bit.
To keep the tip from crashing into the surface in normal operation one uses a feedback loop (control engineering). When the current rises one pulls back a bit when it sinks one moves closer a bit. How fast and in which character one does that can be set by some control parameters. When following the height profile of the surface there is a fundamental speed limit though that can not be broken by any choice of control parameters. Problematically this speed limit is quite low since the moving mass of the piezo actuators is gigantic in comparison to the mass of atoms. Imagine moving a beard stubble (representing a molecule) with giant oil tanker (representing just the needle the much bigger piezos-actuarors are not even included).
- When moving to fast down an edge one makes a jump creating a "shadow" in the image.
- When moving to fast up an edge one crashes the tip into the edge.
Instead of keeping the current constant one can alternatively set a fixed height and then look at the variation of the current while meandering over the surface. At some level of current one should retract though to prevent a crash. Also vertical drift becomes an issue. Some magnetic types of SPM are done this way.
When steps on the surface are sharper than the tip one images the tip with the surface instead of the surface with the tip. This can be identifies by translation symmetric repeating patterns in the image. folding. This especially happens with blunt rough tips with two or more apexes. All the imaged structures are are doubled or multiplied in an other way. The image on the computer is a result of a mathematical folding operation between tip structure and surface structure. If the exact structure of the tip is known one can apply a reverse fold. This does not work for arbitrary blunt tips though. Noise! (TODO: lookup details of the limit here).
There have been attempts to automate the extraction of the tip structure (live?) from scanning over a surface with known structure and comparing the tip surface folding image with an ideal theoretical image of the surface.
Instead of removing the unpredictable shape-changes of tips by mathematical processing there have also been attempts to make tips that are more stable and or more defined. Very popular is picking up a carbon monoxide molecule that sticks down from the rather blunt tip like a needle with a very well known electron density distribution (DOS). Also attempted where attaching nanotubes or wrapping tips in graphene sheets. Modified tips can be very expensive. If not in cost when bought then costly in preparation time when self-made.
SPM microscopes themselves are quite small from human perspective (about the size of a bean can). This includes the tip/needle, the piezoelectric actuators and the suspension. Amplification electronics add a suitcase sized box.
But SPM microscopes that are featuring atomic resolution on a wide range on surfaces are enclosed in an ultra high vacuum (UHV) system and those usually fills a whole rooms. Optional cryogenic systems add to that.
SPM microscopes can be operated in air but there are barely any surfaces that can be imaged with atomic resolution. Notable exceptions are chemically very nonreactive (tech term: inert) surfaces like gold or single crystalline graphite (aka highly ordered pyrolytic graphite HOPG).
Mechanical SPM – usually called: atomic force microscope AFM
Many of the things mentined for STM's hold equally for AFM's.
AFM's usually swing up and down when moving over the surface. This can happen closer (tapping mode) or a little farther away (non contact mode). (Whether something contacts or not is a bit of a moot point in the soft sub nanoscale area. One can use various definitions for contact. There are VdW radii, covalent radii, and many other possibilities.)
Getting atomic resolution with AFM microscopes is harder than with STM microscopes. Since:
- the onset of the interaction is less sharp - (TODO: to very and quantify).
- tips are usually less sharp
To get atomic resolution (which initially was thought to be impossible) one needs to use a tuning fork with a higher stiffness.
Experimental difficulties
Samples inside of UHV system (e.g. Omicrom) are (often/always/sometimes?) purely mechanically tele-manipulated via crude grippers that are fed through stainless steel bellows. There also is rather bad visibility into UHV systems since there are only few view-ports in the system (fixed perspective). One can easily can crash the gripper that holds a sample into some internal obstacles loosing a valuable sample.
To retrieve stuff that fell down the whole system has to be opened. But this lets in water vapor. So after after and while pumping it down again the whole room filling system needs to be heated to oven temperatures over the course of several days!!
Miniaturization
The smaller scanning probe microscopes are built the faster they can work. (TODO: elaborate on that)
Related
External Links
- Wikipedia: scanning probe microscopy
- Wikipedia: scanning tunneling microscopy
- Wikipedia: atomic force microscopy
- At "Foresight Technical Conference 2013" Neil Sarkar gave a talk called "microscopic microscopes for the masses"
(TODO: Video recording was online (known!) - find it and add link to this vimeo video - was it removed?)
Now (2016) the microscope is available at a price point of $2,490 USD: (announcement)
icspicorp homepage
Images of tips
TEM-images of tungsten STM tips where atomic lattice features are resolved:
