Physics 441/2: Transmission Electron Microscope
Introduction
In this experiment we will explore the use of transmission electron microscopy (TEM) to
take us into the world of ultrasmall structures. This is the regime between 1000 ? and
atomic dimensions in which the continuing miniaturization of integrated electronics is
being pursued. It is also a very important region for structural biology. These length
scales are far below the limit where the resolution of conventional optical microscopy
becomes dominated by the wavelength of visible light (~5000 ?). The transmission
electron microscope (or TEM), first invented in the late 1930’s, has now developed into
the technique of choice for microstructral studies in a wide range of fields: materials
research, biophysics, polymer science, mineralogy, and health sciences, to name a
few.
In the first part of the experiment, you will get a feel for the capabilities and immense
resolving power of TEM by imaging some samples of DNA. We will measure the
diameter and pitch of DNA’s famous double-helix structure. In the second part of the
experiment, we will use TEM to study some “quantum well” structures made from
ultrathin layers of Silicon and alloyed Silicon-germanium. We will determine the point
spread function of the Philips 420 electron microscope and measure, at the highest
magnification, the width of a quantum well and the abruptness of its boundaries. These
are key quantities which determine the spectrum of electronic energy levels of the
quantum well. The Si/Si-Ge samples will also demonstrate the capability of TEM to
obtain selected-area diffraction patterns, from which detailed structural information can
be obtained at atomic dimensions.
Basic Principles
The uncertainty principle, Dp.Dx = h, sets a fundamental limit on the spatial resolution, D
x, that can be obtained by probing a system with a beam of particles having de Broglie
wavelength l (= h/Dp). Thus Dx @ O(l). In this way, the resolving power of an optical
microscope is limited by the wavelength of the light employed: (l > 3000?). In the
electron microscope, much shorter de Broglie wavelengths are possible by using highly
energetic electrons, thereby pushing the resolution limit down to the ? level.
Since they are charged, electron beams can be deflected by electrostatic or magnetic
fields. In virtually all commercial electron microscopes nowadays, magnetic lenses are
used to focus the electron beam carrying out the functions that glass
lenses serve in a conventional optical microscope. The Philips 420 was one of the first
high-resolution (Dx < 5?) microscopes on the market. Although this instrument was built
in the early eighties, the designs of the column and the magnetic lenses remain state-ofthe-
art.
The principle of operation is entirely analogous to the optical microscope. A collimated
beam of electrons, emitted from a hot LaB6 filament, is accelerated to ~120 keV. After
passing through a condenser lens the beam is incident on the sample. The size of the
beam can be varied, but is typically ~ 1 micron in diameter. Thus small selected areas
of the sample can be probed. The beam passes through the sample (which is thinned to
~ 200 ? to permit transmission of electrons in the 100 keV range) An objective lens,
situated immediately below the sample, then produces a magnified image of the sample.
Finally, this image is projected onto a fluorescent screen at the base of the column.
One of the powerful features of electron microscopes is their ability to display diffraction
patterns of the sample. In this case the wave-like nature of the electrons is utilized to
diffract the incident beam from the atomic structure within the sample. This can provide
information on the crystal structure of the sample and is particularly useful when the
atomic arrangement is regular and periodic, as in a crystal. It is very straightforward to
switch from a real image to the diffraction pattern: In diffraction mode the current to the
objective lens is turned off so that the diffraction pattern of the sample is simply projected
onto the fluorescent screen.
Imaging a “superlattice”
Introduction - The “superlattice” sample we’re going to investigate is an example of an
artificial crystal structure made by depositing alternating layers of two different (but
chemically related) materials on the flat surface of a semiconductor wafer. In this case
the two materials are Silicon and an alloy of Si and Ge, Si0.8Ge0.2. The confinement of
electrons within one of these layers is referred to as a “quantum well”. The deposition
process, carried out in an ultrahigh vacuum chamber (p < 10-10 Torr), is called
“Molecular Beam Epitaxy” (MBE). MBE is one of the most important and widely used
fabrication techniques in modern electronics technology; it is capable of producing
devices with precisely defined layer thicknesses of less than 10nm! For example, most
laser diodes used in DVD players are made by MBE. Another important example is the
sample used for the discovery of the fractional Quantum Hall Effect, which was
recognized by the 1998 Nobel Prize for physics.
TEM measurements - In this experiment you will use Transmission Electron Microscopy
(TEM) techniques to investigate the microscopic structure of a superlattice, including
measurements of key structural aspects such as:
· individual superlattice layer thicknesses
· sharpness of interfaces between the different layers
· crystallographic orientation of sample
· interatomic distances within each layer
Each of these parameters is very important to the physics of real electronic devices
based on superlattice materials. For example, in the example given above, energy
levels are established by forming a “quantum well” by means of the discontinuities in
electronic potential that exist at the interface between the two different materials:
Energy
Si SiGe Si
Fig. 1: Si -SiGe Quantum Well
I L I
Distance
Clearly, it is important to minimize roughness, and smearing of the potential discontinuity,
if the quantum well is to function as desired. Also, it is important that the width, L, is
precisely defined.
Procedure - The power of TEM lies in its ability to provide direct, high-resolution, images
of microscopic structures as well as “selected area” diffraction patterns of the same
region that is being imaged. In this way one can routinely obtain both a ‘picture’ of the
sample (which is really an electron absorption contrast map in a bright field image) as
well as information about the crystallographic arrangement of the atoms. All of this
information is obtained from the region probed by the electron beam (typically 1mm
across, in this microscope).
In this part of the experiment you will try to obtain as sharp an image as possible, limited
by the resolution of the microscope, and by the intrinsic chemical interdiffusion between
Si and SiGe layers. With the microscope set in the “M” mode (button on right of front
panel), set the magnification between 200,000x and 600,000x, and record (using the
CCD camera) several images of the superlattice at different microscope focus settings
to determine the optimal focus condition. Make a note of the magnification for each
picture.
Note that the superlattice sample is prepared as a thin cross-section (in order to permit
transmission of the electron beam, the thickness cannot be more than ~30nm) so that
the beam is parallel to the layers:
electron beam
Fig. 2: Cross-section
TEM sample
Si Si0.8 Ge0.2 Si