Beyond optical microscopy  (Page 4/5)

The very high degree of focus of a STM allows it to be used as a spectroscopic tool as well as a larger scale image producer. Properties of a single point on a sample surface can be analyzed through focused examination of the electronic structure.

Complications and Caveats

The integral use of the tunneling current in STM requires that both the probe and the sample be conductive, so the electrons can move between them. Non-conductive samples, therefore, must be coated in a metal, which obscures details as well as masks the actual properties of the sample. Furthermore, like with the SEM, oxidation and other contamination of the sample surface can be a problem, depending on the material(s) being studied. To avoid this, STM work is often carried out in an ultra-high vacuum (UHV) environment rather than in air. Some samples, however, are fairly well-suited to study in ambient conditions; one can strip away successive levels of a layered sample material in order to “clean” the surface as the study is being conducted.

Another seemingly simple problem involved in STM is control of vibration. Since the distances between probe and sample are so minute, the tiniest shake can result in data errors or cause the tip to collide with the surface, damaging the sample and possibly ruining the tip of the probe. A variety of systems have been implemented to control vibration, often involving frames with springs, or a sling in which the microscope is hung.

STM is plagued by artifacts, systematic errors in the observed data due to the mechanistic details of the microscope. For example, repetition of a particular shape in the same orientation throughout the image may be a case of tip artifacts, where a feature on the sample was sharper than the tip itself, resulting in the tip’s shape being recorded rather than that of the sample feature. Lack of optimization of the microscope’s feedback loop can produce large amounts of noise in the data, or, alternatively, cause a surface to be much smoother than it is. Finally, while sophisticated image processing software lends much-needed clarity to STM data, it can be misused such that meaning is created where there is none. Image filters used must be carefully evaluated against more “raw” image data to affirm their utility.

Counterbalancing the technique’s obvious usefulness is the general difficulty of STM as a process. Whereas a scanning electron microscope can be operated successfully by a researcher with minimum skills as a technician, STMs are notoriously finicky and require expertise, time, and patience to produce a decent image. They are therefore not particularly popular research tools, though improvements in design and artifact control have been and are being made, making STM increasingly more practical.

Questions for review

• In what types of situations would constant current microscopy be preferred over constant height? And vice-versa?
• What are potential problems of the large amount of data filtering and processing involved in STM?
• What errors are likely to be present in data from a particularly jagged, sharp-featured sample, and why?