
114
CHAPTER6/SCANNING PROBE MICROSCOPY
nanometer scale, analogous to photolithography. It has been demon-
strated that the STM can produce patterns with line widths on the order
of 10 nm.
6.7.4 Current Fluctuations
Electron traps may exist on surfaces (e.g., electronic states located in
the bandgap of semiconductors or poorly conducting species). When
a tunneling electron impinges on the surface and is captured by these
charge traps, the local potential becomes more negative and suppresses
further tunneling. This results in a drop of the tunneling current. Some
time later, the electron is released from the trap, resulting in resumption
of normal tunneling. The net result is that the tunneling current fluctu-
ates with time at these trap sites. Therefore, this provides a method for
imaging these electron traps. Further, by measuring how the rate of
tunneling current fluctuation as a function of tunneling bias, it is possible
to determine the energy location of these electron traps. For further
information, see the articles by M. E. Welland and R. H. Koch, Appl.
Phys. Lett. 48, 724 (1986), and R. H. Koch and R. J. Hamers, Surf.
Sci. 181, 333 (1987).
6.8 LIMITATIONS OF STM AND SOLUTIONS
There are two major limitations of the STM. First, the specimen surface
must be reasonably conducting. Under typical operation conditions,
the resistance of the gap separating the tip and the specimen is on the
order of 10 megaohms (e.g., tunneling at 1 nA under a bias of 10 mV).
‘‘Reasonably conducting’’ means that the resistance of the electrical
path from the specimen to the return circuit should be small compared
with 10 megaohms. This rules out many ceramic and polymer materials
from consideration. One solution is to put a conduction coating (e.g.,
gold) on such surfaces, assuming that the coating faithfully reproduces
the surface topography of the substrate. Another solution is to use AC
tunneling, i.e., the bias is allowed to change sign rapidly. The basic
idea is that in the forward cycle, electrons are injected from the tip
onto the surface. The behavior of the tunneling current with respect to
tip–surface spacing is as predicted by Eq. (6.3) in this portion of the
cycle. Therefore, feedback control can be ‘‘locked’’ to the tunneling
current in the forward cycle. In the next half cycle, the polarity is