|
|
Low Energy Electron Microscopy
The need to observe nanometer scale processes on surfaces in real-time lead to the development
of a LEEM instrument (low energy electron microscope) in 1985 by Ernst Bauer
and Wolfgang Telieps, more than 20 years after its invention by E. Bauer in
1962. While the first Instruments were installed at the University
of Clausthal in Germany, Dr. Ruud Tromp and Dr. Marc Reuter initiated the
development of a new LEEM at IBM. This instrument started operation in 1991 and
was later on sold to several places worldwide. With
new theoretical concepts for the correction of imaging errors of magnetic
lenses, magnetic deflectors, and analyzers, new LEEM designs became desirable.
These include the energy filtered LEEM (SPE-LEEM)
for chemical analysis, the spin polarized LEEM (SP-LEEM)
for magnetic imaging, the Baby LEEM approach and the SMART
project. In 1998 IBM built the first LEEM including a 90° Beam deflector
worldwide. This LEEM instrument is currently installed at the T.J. Watson
Research Center in Yorktown Heights, NY and is used for surface science research.
 |
 |
Figure 1: Basic setup of LEEM II |
Figure 2: A photo of LEEM II |
The basic setup of the IBM LEEM II is shown in Image 1. Electrons leave the
electron gun at the top of the image with an Energy of 15000eV. These high
energy ("fast") electrons fly downwards, and pass a set of
lenses and steering coils that take care of electron focus and beam position
before they reach the prism array. The prism array is the magnetic 90°
deflector mentioned before. Here the electrons are deflected by 90°: they
leave the prism flying to the right, into the sample chamber - still having an
energy of 15000eV. When the electrons pass the objective lens at the entrance of
the sample chamber, they are "slowed down" and decelerated to an
energy of only a few eV, since the sample, which is located in front of the
objective lens, is held at a potential of nearly 15000V. The low energy electrons
are scattered at the sample surface, analog to a LEED
experiment and reflected back into the objective lens. Due to the potential
difference between sample surface and objective lens, the electrons are
again accelerated to 15000eV on this path. The prism deflects the electrons
downwards into the imaging column, where the projector lenses and the screen are
located.
Even though details of the lens setup and its purposes would be hard to
describe here, comparison of the setup with a simple optical microscope can help
to understand the basic idea. The electron gun and the condenser lens play the
role of the illuminating lamp of the microscope. While in the optical case the
illumination optics is mounted below the sample and the observation is performed
in transmission, there is no need for a prism - this is different in the case of
the LEEM. However, the objective lens is equivalent to the objective in the
optical microscope and the projector is equivalent to the optical microscope
projector.
 |
 |
Figure 3: a) Si(001) atomic arrangement
b) the corresponding diffraction pattern |
|
|
When a parallel and coherent beam of low energy electrons hits a single
crystal surface composed of atoms arranged on a periodical lattice, the
electrons are diffracted at the surface and the result is a diffraction pattern.
This effect is used in the LEED instrument (low energy electron diffraction),
but also plays an important role in the LEEM case. In Fig.3 a) the atomic
configuration of a Si(001) surface is shown, with the surface atoms relaxed into
their minimum energy configuration (the "(2x1)" reconstruction)
and minimizing the total number of unsaturated ("dangling")
bonds. Each two adjacent surface atoms move together slightly and form a double
bond with their neighbor (pairs of red and green atoms in case of Fig.3a).
However, due to the properties of the Si crystal lattice, the bond geometry is
rotated by 90° on each atomic step on the surface. Fig.3a shows two terraces,
an upper terrace on the left, and a lower terrace on the right.
What
would a diffraction pattern of a surface like this look like ?
First of all: diffraction reflects periodicities on the surface. We
would expect
to find the underlying square lattic of the atoms in the diffraction
pattern, in
addition to spots that are due to the reconstruction. The experimental
result is
shown in Fig.3b). The spot in the center is the directly reflected
beam, the
green and red ones are caused by the reconstruction. Concentrating on
the
green spots first, which correspond to the left terrace, it is
obvious that the
periodicity in the vertical direction in (a) and (b) is not changed
compared to
the underlying square lattice. In the vertical direction, the
periodicity in the
real-space configuration is doubled, which gives rise to additional
spots in the
diffraction pattern at 1/2 positions. The same argument explains the
red spots, which are rotated by 90° relative to the green ones and
are generated by
the lower terrace in Fig.3a.
When the low energy electrons used for the illumination of the sample in the
LEEM hit the crystal surface, a diffraction pattern is the result. For real
space imaging the experimentalist chooses one of the diffraction spots with a
contrast aperture located in the imaging column. Depending on what spot was
chosen, the resulting image can be completely different. If the specular (center)
beam is used, the resulting image is called a "bright field" image,
images taken with any other beam are called "dark field" images. An
example for dark field imaging is shown in Fig.4 using
the Si(001) (2x1) surface from above.
Fig 4.
 |
|
 |
|
 |
| a) LEED Image of Si(001) |
|
b) Dark-Field image of the surface in a) using one of the green spots for imaging |
|
c) Same as b; using a red spot |
 |
| Fig. 5: Step phase contrast on Si(111) |
While the lateral resolution of the IBM LEEM II is "only" 5
nanometers, the vertical resolution can be much higher due to diffraction
effects. An example for this was already given in the above paragraph, where the
dark field imaging was used to resolve terraces on the surface with a height
difference of only one atomic step (=1.36 Å for the case of Si(001) ). Another
possibility to resolve single steps, is to make use of the wave nature of the
incoming electron beam. When the wavelength of the beam is chosen in a way, that
destructive interference occurs between adjacent terraces, all terraces
will appear with the same gray level in the microscopic image, but the steps
separating the terraces will appear as dark lines in the image. An example is
given in Fig.5
 |
PEEM: Photoelectron emission microscopy. The electron gun of the LEEM
is switched off, Electrons are excited with a UV light source making use
of the photo effect. The resolution is not as good as in LEEM (The
intensity of the used HG is not high enough for 5 nm resolution), but at lower
magnification differences in the work function of different materials are easily
visible. There is no LEEM image visible in the intermediate image plane, but
the PEAD - the Photoelectron angular distribution.
|
| |
 |
MEM: Mirror Electron Microscopy: The electron energy is reduced to the
limit, when the electrons return in the retarding field, before they hit the sample
surface. It is hard to understand, how the contrast in these images is formed.
The basic mechanism is, that all height variations on the sample surface, such
as steps, grains, ... change the local properties of the retarding field and
therefore take influence on the reflected electron beam. The intensity of the
reflected beam is very high, and there is no LEED image visible: since no
scattering process took place, all reflected electrons are in the specular beam.
|
| |
 |
Dark field imaging: Usage of one LEED spot in the intermediate plane
for imaging All areas on the surface that contribute to the existence of this
spot appear bright in the image, all other areas appear dark. See above chapter
for detailed explanation.
|
| |
 |
phase contrast: Usage of the wave nature of the incident electron beam
to generate a vertical diffraction contrast, e.g. to make steps visible on the
surface. See above chapter for detailed explanation.
|
| |
 |
reflectivity contrast: Different areas on the surface might show a
difference in electron reflectivity, depending on the surface material, or even
depending on the surface structure. Since the reflectivity coefficient depends
on the incident electron energy, the contrast can be optimized. The most famous
example is the difference between the (7x7) reconstruction and the (1x1)
structure on the Si(111) surface at ~850°C. At an electron energy of about 10eV the (7x7)
areas appear much brighter than the remaining surface.
|
| |
 |
LEED: since a diffraction pattern is formed in the backfocal plane of
the objective lens, it is possible to image this pattern on the screen (LEED).
See above chapter for detailed explanation.
|
| |
 |
Microdiffraction: By restricting the electron beam only to a very
small area on the surface (fraction of a µm), it is possible to determine the
LEED pattern of small areas on the surface, like the LEED pattern of single islands
or terraces in order to
determine their crystal structure and orientation.
|
|
|
