Chapter 3 Electron Scattering

1
Intruduction
Scattering is fundamental to all electron
microscopy. We cannot see anything in electron
microscope unless our specimen interacts and
scatters the electrons in some way. Thus any
nonscattering object is invisible. Electron Thin specimen
microscope is constructed to gather scattered
electrons and they will give us information we
seek about the structure and chemistry of the
specimen.
Without understanding phenomena of scattering
it is impossible to interpret the image, diffraction
pattern or analytical spectrum which microscope
produces.
Even after you are able to produce a very small
probe size (size of the beam on the sample) like Thin specimen
~ 1 nm and ask yourself question whether the
signal generated by the specimen coming only
from this area. The answer is: “NO” (especially
for SEM). Degraded spatial resolution is a direct
consequence of the electron scattering.

Fig. 3.1. (A) A uniform intensity of electrons, represented by the flat line, falls on a thin specimen. Scattering
within the specimen changes both the spatial and angular distribution of the emerging electrons. The spatial
distribution (intensity) is indicated by the wavy line. (B) The change in angular distribution is shown by an
incident beam of electrons being transformed into several forward-scattered beams.
2
3.1. The Characteristics of Electron Scattering
Electron scattering can be grouped in different ways. The most important are elastic and
inelastic. These terms are simply the description of scattering that results in no loss of
energy and some measurable (more than 0.1 eV) loss of energy, respectively. In this
case we consider the electrons as particles and
scattering involves some interaction like billiard
balls colliding (Figure).

However, we can also separate scattered

electrons into coherent and incoherent, which
refer to their wave nature. These two
distinctions are related, since elastic electrons
are usually coherent and inelastic ones are
usually incoherent.

The nature of scattering can result in different

Fig. 3.2. Schematic illustration of scattering angular distributions. Scattering can be either
processes that occur when an energetic
electron of energy E0 interacts with an atom:
forward scattering or back scattering. If
(a) Elastic scattering, in which the energy E1 electrons are scattered through the angle
after scattering equals E0. (b) Inelastic j < 90° then they are forward scattered, and if
scattering, in which E1 is less than E0. Note j > 90° they are backscattered.
that the elastic scattering angle je >> ji.

3
Different Kinds of Electron Scattering
A
1) Elastic scattering is usually coherent if
5
the specimen is thin and crystalline.
2) Elastic coherent scattering is usually
occurs at relatively low angles (1-10°) in
Thin specimen the forward direction
3) At higher angles (>10°) elastic scattering
3
becomes more incoherent.
1 4
4) Inelastic scattering is almost always
2
incoherent and relatively low angle (<1°)
forward scattering.
5) As specimen gets thicker, fewer
electrons are forward scattered and
more are backscattered until only
B
incoherent backscattering is detectable
in bulk.
5 Forward scattering causes most of the
signals used in TEM; backscattering in SEM
Bulk specimen

Fig. 3.3. Different kinds of electron scattering from (A) a thin specimen and (B) a bulk specimen: a
thin specimen permits electron to be scattered in both forward and back directions while a bulk 4
specimen only backscatters the incident beam electrons.
Probability of Scattering: Cross-Section

Electrons can be scattered more than once. The simplest scattering process is single
scattering. In TEM while the sample is very thin it is quite reasonable assumption that all
scattered electrons undergo one scattering event or not being scattered. If an electron
scattered more than once we use term plural scattering, and if it is scattered more than
20 times we call it multiple scattering. The characteristics of the scattering event are
controlled by factors such as the electron energy and the atomic number/weight of the
scattering atom. When we consider a specimen rather than single atom, factors such as
thickness, density and crystallinity of the specimen also become important.

The probability that electron will be scattered described by cross-section, s, or mean free
path, l.

A cross section is the area, which the scattering particle appears to present to the
electron: 3
N events / cm
s= = [cm2]
ni × nt ( )(
e / cm 2 × atom / cm 3 ) (3.1)

- N is number of events of a certain type per unit volume;

- ni is a number of the incident particles per unit area;
- nt is a number of target sites per unit volume.

5
Probability of Scattering: Mean Free Path
Mean free path, l, is an average distance electron travels in the specimen between the
events:
A
l= (3.2)
N A × r ×s
- A is atomic weight (g/mol);
- NA=6.02×1023 atoms/mol is Avogadro number;
- r is density g/cm3.
The bigger l means that probability for the event of scattering is less. Typical values of l
for scattering at TEM voltages are of the order of tens of nm. The total mean free pass of
the several scattering events is calculated:
1 1 1 1
= + + + ... (3.3)
l l1 l2 l3
In single or plural scattering the probability of single electron being scattered n times
while traveling distance x is given by Poisson equation

P(n ) = (1 / n!) × (x / l ) exp(- x / l )

n
(3.4)

6
Multiple Scattering: Monte Carlo Simulation
For the multiple scattering all primary electrons can be assumed as scattered many times by several
different mechanisms. In this case Monte Carlo simulation (random numbers) is applied to describe the
process. Our knowledge of the values of s and l is very sketchy (for particular scattering events may
be known within a factor of two). By Monte Carlo simulation we can predict the electron paths as the
beam is scattered.
The term “Monte Carlo method” was coined in the 1940s by physicists Stanislav Ulam and John von
Neumann working on nuclear weapon projects in the Los Alamos National Laboratory to calculate
path of neutrons through deuterium and tritium. Monte Carlo simulation methods are especially useful
for modeling phenomena with significant uncertainty in inputs and in studying the systems with a large
number of coupled degrees of freedom.

A B

http://www.gel.usherbrooke.ca/casino/index.html
Fig. 3.4. Monte Carlo electron-trajectory simulation of the beam interaction in iron, E0 = 20 keV, tilt = 0°.
(A) Plot of five trajectories, showing random variations. (B) Plot of 100 trajectories projected on a plane
7
perpendicular to the surface, giving interaction volume.
3.2. Elastic Scattering. Scattering by Isolated
Single Atom
Elastic scattering is a process, which might change the direction of the primary electron without
changing its energy detectably. It is known as Rutherford scattering. This type of scattering is a result of
Coulomb interactions (involving electrostatic charges). It’s convenient to divide the elastic scattering
mechanism into two principal forms: electron scattering from isolated single atoms and collective
scattering from many atoms together within the specimen when the atoms are ordered.
In case of interaction of a single electron with isolated atom
elastic scattering can occur in one of two ways. 1) The electron
from the beam may interact with the electron cloud of the atom,
resulting in a small angular deviation. 2) Alternatively, if an
electron penetrates the electron cloud and approaches the
nucleus, it will be strongly attracted and may be scattered through
a larger angle, which can approach 180°. You should be aware
that either of these two interactions may not be truly elastic, so
our separation of scattering into elastic and inelastic is a bit of
simplification.
Probability of electron being scattered elastically is given by
equation:
1 E0 is the energy of the electron;
( )
Pq ~ 2 (3.5) q is the angle of scattering.
E 0 × sin q
4

Probability of the electron scattering is higher for: 1) the lower energy electrons; 2) and electrons
scattered with the angles close to 0 (forward) and 180 deg (backscattered electrons).
Fig. 3.5. Two mechanism by which high-energy electron is scattered by an isolated atom. Coulombic interaction within
the electron cloud results in low-angle (q) scatter while Coulombic attraction by the nucleus causes high q scatter or
complete backscatter. The potential within the electron cloud is always positive. 8
 Interaction of two waves The distance through
which a wave change its
2.5 2.5
p 2p p 2p phase on 2π is called
1.5 1.5 wave length. A rule of
0.5 0.5 wave theory states that
waves reinforce one
-0.5 -0.5
another when they are in
-1.5 -1.5 phase, i.e., when they
-2.5 -2.5 are coherent. Two waves
2.5 2.5 will be coherent when
1.5 1.5 they are in phase or
0.5
when their phase
0.5
difference is 0×p, 2×p,
-0.5 -0.5 4×p... etc. When we have
-1.5 -1.5 a set of coherent waves
-2.5 -2.5
they form a plane wave
2.5 2.5
for which all waves from
the set are in the same
1.5 1.5
phase perpendicular to
0.5 0.5 the direction of the wave
-0.5 -0.5 propagates. And if we
have set of waves which
-1.5 -1.5
are not coherent we don’t
-2.5 -2.5 form a plane wave.
constructive interference deconstructive interference
 Diffraction of Coherent Laser Light on the slit

When the plane wave interacts with the point shape obstacle it forms a spherical wavelet.
We can argue that all the waves from the initial set will stay in phase on the surface which
is spherical in shape. 10
Interaction of Plane Wave with two Atoms
0th order

1st order

11
3.2. Elastic Scattering. Diffraction
Let’s consider the electron wave interaction with the specimen as a whole. Each atom in
the specimen acting as a source of “secondary” spherical wavelets. A rule of wave theory
is that waves reinforce one another when they are in phase, i.e., when they are coherent.
Waves cancel one another when they are out of phase. Diffracted waves are in phase with
one another only in certain directions. Zero-order wave which proceeds in the same
direction as the incident wave we’ll refer in TEM as the direct beam of electrons. There
are also higher-order waves which propagate in directions that are at some fixed angle to
the incident wave and we’ll call these the diffracted beams.

Fig. 3.6. A plane coherent electron wave

generates secondary wavelets from a row of
scattering centers (e.g. ordered atoms in the
specimen) The secondary wavelets interfere,
resulting in strong direct (zero-order) beam
and several orders of coherent beams
scattered (diffracted) at specific angles. Thus
the low-angle elastic scattering distribution is
modified by the crystal structure of the
specimen, and intense diffracted beams
emerge at specific angles.

To get the diffracted beam we need to satisfy two conditions: 1) we need to have our electron waves 12
to be coherent and we need to have 2) the structure of atoms to be ordered (crystal lattice).
 Diffraction: Von Laue Approach
Von Laue used light-optics approach to argue that the diffracted waves (x-rays) are in phase
if the path difference between waves scattered by adjacent scattering centers is a whole
number of wavelength, hl, and h is an integer. Thus, if scattering centers are spaced
distance a apart and the incident beam makes
an angle q1 with the line connecting the
scattering centers, and is diffracted at an
angle q2, then the path difference (AB-CD) is
a(cosq1 - cosq2 ) = h × l (3.5)

b(cosq3 - cosq4 ) = k × l

c(cosq5 - cosq6 ) = l × l
If all three Laue equations are satisfied
simultaneously then a diffracted beam is produced.
The letters hkl are the indices of the diffracted beam
and are equivalent to the Miller indices (hkl) of a
crystal plane.

Fig. 3.7. The approach used by von Laue to calculate the path difference for the x-ray waves valid
for any wave including the electron wave (wavelength l). In this one-dimensional figure the wave is
incident at angle q1 and scattered at angle q2 from two atoms (B and C) spaced distance a apart.
13
The path difference between scattered waves is AB-CD.
 Diffraction: Bragg’s Law
Laue’s approach was simplified by Bragg who argued that the waves behave as if they
“reflected” off atomic planes. Bragg argued that waves “reflected” of adjacent scattering
centers must have a path difference equal to an integer number of wavelength if they are to
remain in phase. The path difference between electron waves “reflected” from the upper and
lower planes in Figure 3.8 is (AB + BC). Thus if “reflecting” hkl planes are spaced a distance
d apart and the wave is incident and “reflected” at an angle qB, both AB and BC are equal to
d·sinqB and the total path difference is 2 d∙sinqB. Then the Bragg’s law:

n × l = 2d × sin q B (3.6)

Elastic scattering is important

because it is major mechanism
by which electrons are deflected
and because they make main
contribution to the diffraction
pattern (used in TEM) and
backscattered electrons signal
(used in SEM).

Fig. 3.8. The Bragg description of diffraction in terms of the “reflection” of a plane wave (wavelength l)
incident at an angle q to atomic planes of spacing d. 14
3.3. Inelastic Scattering
Inelastic scattering is a general term, which refers to any process causing the primary
electrons to lose a detectable amount of their energy (it has to be more than 0.1 eV to be
detected). When the high energy electron encounters the atom, it first penetrates the
outer, loosely bound electrons of the cloud, then it passes the inner, more tightly bound
core shell electrons, and finally it may encounter the nucleus.

The deeper electron penetrates into the atom, the

higher energy it may loose. It is possible (but very
rare) for the electron to lose all it’s energy in a
single nuclear interaction. There are many
interaction processes, which can cause loss of the
energy. It is important to realize that inelastic
scattering is eventually responsible for stopping the
electrons by the solid. Almost all electron energy
will end up as a heat in the specimen. A small
portion of energy will escape as secondary
electrons or as X-rays.

Fig. 3.9. Cross section for the various inelastic scattering processes in Al (aluminum) as a function of
the incident electron energy, assuming a small angle of scatter (q ~0°); plasmon (P), K and L-shell
ionization (K, L), fast and slow electron generation (FSE, SE). For comparison purposes the elastic
15
cross-section (E) is also included. The values are relatively insensitive to the given range of the beam
energy.
Three Major Components of Inelastic
Scattering

We separate the inelastic processes into three components:

• Process that generates X-rays.

• Process that generates other (secondary) electrons.

• Process that is a result of collective interactions with many atoms.

16
Generation of Characteristic X-ray Signal

Fig. 3.10. The ionization process. An inner (K) shell electron is ejected from the atom by high-energy
electron. When the hole in the K shell is filed by an electron from the L shell, characteristic Ka X-ray
emission occurs. The beam electron loses energy but continues on through the specimen. 17
 Characteristic X-ray Signal: K, L, and M Signals
The difference in the two shell
energies equals the energy of
characteristic X-ray. If K shell hole
was filled with the L shell electron
we get Ka X-ray, but if it was filled
from the M shell we get a Kb X-ray.
If the hole was in the L shell and
we filled it with the electron from
the M shell, we get an La X-ray,
and if we fill it from the N shell we
get an Lb X-ray. The notation is
more complex because we
differentiate the a X-ray in terms of
a1 and a2 depending on which
subshell the electron falls to fill the
hole (Figure).

Full list of energies can be found on: http://xdb.lbl.gov/

Fig. 3.11. The complete range of possible electron transitions that give rise to K, L and M
characteristic X-rays. Not all these X-ray are detectable by EDS in the electron microscope.
Allowed transitions are determined by the selection rule 18
Quantum Numbers and Selection Rules
The principal quantum number, n, describes the shell. The azimuthal quantum number, l,
describes the subshell. In chemistry and spectroscopy, ℓ = 0 is called s orbital, ℓ = 1, p
orbital, ℓ = 2, d orbital, and ℓ = 3, f orbital. ℓ = 0, 1, 2,..., n − 1. The magnetic quantum
number, ml, describes the specific orbital (or "cloud") within that subshell. The values of
mℓ range from −ℓ to ℓ, with integer intervals. The spin quantum number, ms, (not shown in
the schematics) describes the intrinsic spin of the electron within each orbital. The labels
in the image show the quantum numbers nlml with l encoded as letter.

The Laporte rule is a selection

rule formally stated as follows:
In a centrosymmetric
environment, transitions
between like atomic orbitals
such as s-s, p-p, d-d, or f-f,
transitions are forbidden.
The orbital quantum number
change: Δl =±1;
The magnetic quantum
number change : Δml =0;±1;

Selection Rules: https://www.youtube.com/watch?v=hWo2b-i6UiE 19

Characteristic X-ray Signal: Critical Ionization
Energy
The electron beam has to transfer an amount of energy greater than a critical value to
inner-shell electron to ionize the atom. This energy is called critical ionization energy (Ec),
which is equal to the binding energy. The beam energy E0 should be greater than Ec if
we are going to generate characteristic X-ray. The value of Ec increases as the electrons
are more tightly bound to the nucleus, so K shell has highest Ec. Atoms with higher Z
(atomic number) have higher Ec. Complete list of ionization values is available as
tabulated data. Ionization cross section area is given by Bethe equation:
where ns is the number of electrons in the
æ p × e 4 bs ns ö æ cs E 0 ö (3.7)
s =ç ÷ logç ÷ ionized subshell, and bs and cs are constants
è E0 Ec ø è Ec ø for the shell.

When we think about X-ray as electromagnetic radiation which travels with the speed of
light, c, we consider it like a wave with a specific length, so the energy of X-ray related to
the wave length through relationship:

hc 1.24
E = hn = (3.8) l= (3.9)
l E
For example if a K shell electron has been knocked out of molybdenum atom and replaced by the
jumping in of an L electron the energy difference DE, is 17400 eV (Table 3.1), which is emitted as Ka
X-ray of Mo. The wavelength of this X-ray calculated by this formula is 0.071 nm. 20
X-Ray Critical-Absorption and Emission Energies
for some Elements K and L Series (in keV)
Table 3.1. X-Ray Critical-Absorption and Emission Energies in keV
Table 3.1

21
Bremsstrahlung X-Ray

If the electrons in the beam penetrate completely through the electron shell of the
atom they can interact inelastically with the nucleus. If the electron is decelerated
by the charge of the nucleus, it emits an X-ray. Since the electron can suffer any
amount of deceleration, then these X-rays can have any energy up to the beam
energy. Such X-rays are known as “bremsstrahlung” or “breaking radiation”. It
appears on all spectra as a background. (No use for the analytical techniques).

22
Secondary Electron Emission (SE)
Secondary electron emission (SE) represents electrons of the specimen that are ejected
by the electrons of the beam. They can be discussed as three distinct groups:
• If the electrons are in the conduction or valence bands then it doesn’t take much
energy to eject them, and they are called “slow SEs” with energies typically below
50 eV.
• If electrons are strongly bound inner-shell electrons they are less readily ejected, but
when ejection happened they can have a significant fraction (up to 50%) of the
beam energy, and they called “fast secondary electrons” or FSEs.
• If the electrons are ejected by the action of a characteristic X-ray passing its energy
to an outer-shell electron, then these secondary electrons are called Auger
electrons.
Slow SEs are ejected from the conduction or valence bands of the atoms in the specimen. Usually
slow SEs are assumed to be free electrons, i.e. they are not associated with a specific atom and
contain no specific elemental information. But due to their weakness slow SEs can escape only if
they are near the specimen surface. So we use them in the SEM for forming images of the specimen
surface (topography).

Fast secondary electrons (FSEs) are high-energy electrons generated in the specimen. They carry
large fraction of the beam energy. Due to this fact they travel significant distances in the specimen
and are capable to escape from deeper areas within specimen compared to slow SEs. As a result
FSEs degrade spatial resolution of the imaging and microanalysis. So, they are unavoidable and
undesirable. We don’t use them in SEM. In electron microscopy term “secondary electrons” is usually 23
applied to slow (low) energy electrons.
 Auger Electrons Emission of Auger electrons is an alternative to
X-ray emission when an ionized atom returns to
the ground state. The ejected Auger electron has
an energy which is calculated based on the
energies of the localized electrons involved. For
instance:

EA = EK – EL2 – EL3

There are few different possibilities for electron

jumps, since there are more than one localized
electron on each shell/subshell. So there are
going to be few Auger peaks on the spectrum.
Hence, Auger electron has characteristic energy
and it depends on the electronic structure of
ionized atom. The Auger process is favored in
atoms with a small binding energy, i.e. lighter
elements. Auger electrons (having energy of a
few kV) are able to escape from very close to the
surface. Auger electrons form a signal, which is
in big practical use in Auger spectroscopy for
analysis of the light elements.
Fig. 3.12. The process of inner (K) shell ionization and subsequent Auger electron emission. Note that
L1 ® K transition is not allowed in the atom since it involves s-cloud electron of L-shell jumping to the s-24
cloud of the K-shell. K- shell has only s-cloud. (s ® s transition forbidden by the Laporte rule)
Relationship Between Probability of X-Ray
and Auger Electrons Emission

Auger electron emission and characteristic X-ray are alternative processes

by which relaxation of the excited atom might occur. They do not have the
same probability. Fraction of the atoms, which emit an X-ray, known as the
fluorescence yield, w, varies strongly with atomic number, Z:

Z4
w= 4 (3.11)
(
Z +c )

c has a value of about 106 for atoms whose K shells have been excited and is
higher for L and M shells. The yield of Auger electrons is 1-w.

25
Electron-Hole Pairs and Cathodoluminescence
(CL)
If your material is a semiconductor, incoming
electron might give some energy to the electron
of the material, which is in the valence band
(Figure). This amount of energy is enough for
the electron of material to jump to the
conduction band leaving the hole in the valence
band. The hole is filled by a conduction band
electron falling back to the valence band. Upon
e
recombination a photon of light is emitted with
the frequency determined by the band gap
width.

Fig. 3.13 Schematic illustration of CL. (A) Initial state

before a beam electron interacts with valence-band
electrons. (B) A valence-band electron is exited across the
gap into the conduction band, leaving a hole in the valence
band. (C) The hole is filled by a conduction-band electron
falling back into the valence band hole. Upon
recombination a photon of light is emitted.

26
Plasmons
Plasmons are collective oscillations of
free electrons that occur when the
electron beam passes through the
“sea” of electrons. It might be
considered as an oscillation, which
creates regions of varying electron
density. Plasmon process has largest
cross section and it is the most
common inelastic interaction
occurring in materials, predominantly
in metals. In plasmon exciting
process the primary beam loses 5-30
eV

27
Phonons

Phonons are oscillations where all the

atoms in the crystal lattice vibrate
collectively. Such vibrations of the atoms
are equivalent to specimen heating. The
amount of lost energy is usually less than
1 eV and the mean free pass is large, of
the order of µm. This doesn’t mean that
phonon scattering is not important. All
electrons, which remain in the solid, are
likely to excite phonons eventually and
this is how the solids are heated by the
electron beam. You can reduce the
number of phonons by cooling the
specimen.

28
3.4. Interaction Volume
It is clear that more than one inelastic process can take place. In a solid specimen many
such events will occur until the electron is stopped inside the specimen or leaves it
surface. For the transmission electron microscopy, when the sample is thin (usually about
1000 A) electrons also have a possibility to pass through the sample. The trajectories of a
few typical electrons, calculated by Monte Carlo simulation are shown on figure. It can be
seen that the majority of electrons are brought to a halt within the solid, but a few are
backscattered and leave the specimen. The volume within which 95% of the primary
electrons are brought to rest is called interaction volume

Fig. Monte Carlo electron-trajectory

simulation of the interaction volume in iron
at E0=20 keV for the 0° tilt of the specimen.

29
Interaction Volume. Kanaya-Okayama Range
Interaction volume depends on electron beam energy, E0, atomic number of the element of
the sample, Z, specimen surface tilt angle, j (Fig. 3.14), and density of the specimen, r
and atomic weight, A. Radius of the interaction volume can be estimated by Kanaya-
Okayama equation:
0.0276 × A × E 01.67
RKO = (3.12)
Z 0.89 r

and dependence on tilt angle by:

RKO (j ) = RKO ( 0) × cosj (3.13)

A B C φ

Fig. 3.14. Monte Carlo electron-trajectory simulation of the interaction volume in iron at
E0=20 keV for various surface tilts: (A) 0°, (B) 45°, (C) 60°. 30
Sampling Volume
Even though radiation is generated within interaction volume, it will not be detected unless
it escapes the specimen. Possibility for the secondary signal to escape from the specimen
depends on how deep inside the specimen interaction takes place, what type of secondary
effect occurs and how high the energy is. The volume of material contributing to the
specific signal is called sampling volume. Figure 3.15 schematically illustrates the size of
the sampling volume for the
different types of the secondary
signals. This example
demonstrates why we usually get
signal in electron microscope
from the area on the sample,
which is considerably larger than
the size of our primary electron
beam.

Fig. 3.15. Schematic illustration of the electron-beam interaction in a Cu, showing the
electron interaction volume, the backscattered-electron sampling depth, the secondary-
electron sampling depth and the x-ray generation range for Cu Ka, Cu La; E0=20 keV. 31