MT60146: ADVANCED ELECTRON

MICROSCOPY AND ANALYSIS

(3-1-0)

R. Mitra

Department of Metallurgical and Materials Engineering

IIT, Kharagpur
 Syllabus and Scope
• Design of Transmission Electron Microscope: Electron Sources, Lenses,
Aperture and Resolution, Aberrations, Electron Optics in bright and dark
field.
• Imaging in conventional TEM: Theory of amplitude contrast images,
extinction contours, imaging of line and surface defects, strain fields and
use of weak beam dark field imaging.
• Electron Diffraction: Concept of reciprocal space; indexing diffraction
patterns and Kikuchi patterns, Micro-diffraction; Convergent Beam
Electron Diffraction.
• Imaging and Diffraction in Scanning Transmission Electron Microscopy.
• High Resolution Transmission Electron Microscopy: Phase Contrast
Imaging, Contrast Transfer Function, Scherzer Defocus, Envelop damping
functions, Image simulation methods.
• Analysis: Energy Dispersive X-Ray Analyses – Principle
and operation, Resolution and Detector, Qualitative and
Quantitative Analyses, Corrections, ALCHEMI. Electron
Energy Loss Spectroscopy – Principle and operation of
PEELS, Concept of Zero Loss, Low and High Loss
Spectrum, Interpretation of EELS data.

Text Books:
1.David B. Williams and C. Barry Carter: Transmission Electron
Microscopy, Vols. I. Basics; II. Diffraction; III. Imaging; IV. Spectroscopy,
Plenum Press, New York, 1996.
2.Gareth Thomas and Michael J. Goringe: Transmission Electron
Microscopy of Materials, John Wiley and Sons, New York, 1979.
3.Sir P.B. Hirsch, A. Howie, Sir R. Nicholson, D.W. Pashley, and M.J.
Whelan: Electron Microscopy of Thin Crystals, Robert E. Krieger
Publishing Company, Malabar, Florida, 1977.
HRTEM LABORATORY,
CRF, IIT KHARAGPUR

JEM 2100 HRTEM – 0.19 nm point-to-point

and 0.14 nm lattice resolution.
 Brief History
• Louis de Broglie (1925) – electrons have wave-like
characteristics with  << visible light
• Davisson-Germer (1927), Thompson and Reid (1927) –
carried out electron diffraction experiments demonstrating
wave nature of electrons.
• Knoll and Ruska (1932) – Designed and used electron
lenses, as well as recorded electron images in Berlin–
Awarded Nobel Prize in 1932
• The Metropolitan Vickers EMI was the first commercial
TEM built in UK in 1936.
• Regular production started by Siemens and Halske in
Germany in 1939. Later – Phillips, JEOL, Hitachi, RCA
• Heindenrich (1949) – first thinned metal foils to electron
transparency.
• Hirsch & Coworkers in Cambridge – developed the theory
of electron contrast.
 Concept of Resolution
• The smallest distance between two points that
human eyes can resolve = 0.1-0.2 mm
• Any instrument revealing details finer than 0.1
mm can be referred to as microscope.
• Resolution of TEM means different for different
functions
• Image resolution in TEM in terms of classical
Rayleigh criterion for light microscopy.
 Raleigh Criterion

Two point sources at infinity have an angular

separation sufficient to place the maximum intensity of
the primary image peak of one source at the position of
the first minimum of the second.
 Raleigh Criterion
• Raleigh criterion implies that combined image of two sources will
show a small (13%) but detectable intensity minimum at the
centre.
• Smallest distance resolved =  = 0.61/μ sin 
where  is the wavelength of radiation; μ is the refractive index of
the medium; and  is the semi-angle of collection of the
magnifying glass.
• μ sin  is also known as the numerical aperture (N.A.).
• Considering N.A. = 1,   ½ 
• For light microscopy using green light ( = 550 nm),   300 nm
= 0.3 μm  1000 atom diameters = too big to identify features in
the microstructure.
• Electron microscope with better resolution as wavelength of
electron is much smaller (approx 0.0037 nm for 100 keV energy).
 e (nm) = 1.22/E, where E = energy in electron-volts (eV).
 Fundamental Properties of Electrons
• de Broglies’s equation considering wave-particle duality for
electrons:  = h/p;
where h = Planck’s constant and p = particle momentum.
• eV = (m0v2)/2, where m0 = rest mass of electron;
• P = m0v = (2m0eV)1/2
  (in nm) = h/(2m0eV)1/2 = 1.22/E (unit of E is electron-volt)

• At 100 keV, the electrons travel at half the speed of light (c)
 Relativistic effects cannot be ignored.
  = h/[2m0eV(1+eV/2m0c2)]1/2
 At E = 100 keV,   0.004 nm
It should be noted that  is much smaller than the diameter of the atom.
Wavelength – Energy Table
 Resolution - Information
• It is possible to obtain atomic resolution images using
Acceleration Voltage = 200-400 kV.
• It is not possible to achieve wavelength-limited resolution,
because it is not possible to fabricate perfect lenses.
• Lenses can be electrostatic or electromagnetic.
• Problems: (1) Spherical Aberration (Cs), (2) Chromatic
Aberration (Cc) and (3) Astigmatism.
• With Cs correction, it is possible to achieve resolution better
than 0.1 nm. (Small electron probes with high current)
• For thicker specimen, correction of Cc through filtering of
energy is useful. (images of electrons with range of specific
energies allowing band-gap imaging and chemical bond
imaging).
 Interactions of Electrons with
Specimen
• Electrons are a type of ionizing radiation, which is capable of
removing one of the tightly bound inner shell electrons from the
attractive field of nucleus – produces secondary signals from
specimens.
OPTIC AXIS
 Depth of Field and Depth of Focus
• Depth of Field: Measure of how much height or thickness
of the object remains in focus. It is also the distance along
the optic axis on both sides of object plane, within which
object can be moved without detectable loss in focus.
- In the TEM, the sample is usually in focus from top to
bottom surfaces of the sample at the same time, as long as
it is electron-transparent. e.g. dislocations or inclined
grain boundaries.

• Depth of Focus: Distance over which the image can move

relative to the object and still remain in focus. It is the
distance along the optic axis within which the image
appears focussed.
 Arrays of
inclined
dislocations.
Ray 1 originates at the highest point
where the object can appear to be in
focus, and this ray arrives at the
farthest point down the column,
where image appears to be in focus.
Ray 2 represents other extreme but
travels at same inclination to optic
axis.
If these rays appear to come from the
same point, the distances, dob and dim
correspond to the smallest distances,
which can be resolved in object or
image, respectively.
Depth of Field and Depth of Focus
im  tan im = (dim/2)/(Dim/2)
ob  tan ob = (dob/2)/(Dob/2)
The angular magnification is:
MA = im/ob.
The transverse magnification is:
MT = dim/dob.
Usually, MT = 1/MA;
The depth of focus,
Dim = (dob/ob).MT2 Greater depth of field and focus
The depth of field, are obtained by using small
apertures (small ).
 Depth of Field & Depth of Focus:
Alternative Treatment

Distance over which object remains in focus is

known as depth of field (d).
(/M)/d = tan 
  = Md tan ; or d =  /M tan 
Distance over which the image remains in focus is
known as depth of focus (D).

D = M cot  = M2d tan . Cot 

If  = , D = M2d
 Depth of Field and Focus
• In thin specimens, which scatter weakly, the
electrons emerge from the specimen as a
cone close to that defined by the angle, im
(10-4 rad.). This angle is controlled by the
condenser aperture (C2 lens).
• In thicker and more strongly scattering
specimen, the objective aperture defines the
angle, and 0 = 10-2 rad.
Sections of TEM
Ray Diagram of
TEM
Ray Diagram for Imaging and
Diffraction
 Vacuum Requirement
• Vacuum is required in the TEM because of
three reasons:
(1) To avoid scattering of high energy electrons
by residual gas in the microscope column.
(2) To ensure thermal and chemical stability of
the electron gun during microscope operation.
(3) To minimize or eliminate beam-induced
contamination of the sample during observation.
Electron Sources
 Thermionic Emission: Work Function

Thermionic Emission Current/Area,

Jth = AT2 exp (-Ew/kT)

(Richardson – Dushman Equation)

A = Constant, Ew = Work Function, T = Operating Temperature of

Filament.

 Jth larger for low Ew and high T

 Field Emission Gun

The first anode provides extraction voltage to pull electrons out of

the tip.
The second anode accelerates the electrons to 100 kV or more.
 Field Emission Gun (FEG)
• Cold FEG: A sharp tip of tunsten needle that emits
electrons is < 1 μm in diameter  Effective source
size is < 0.01 μm, and quite coherent.
• Hot FEG: Tungsten needle is heated to enhance
emission by electron tunneling – Shottkey
emission. These sources contain zirconium to
reduce work function. Larger energy spread and
larger source size than cold FEG, but more stable
and reliable for longer periods.
 Beam Current
The current density or maximum intensity of
electrons in the electron beam at crossover:

Jb = ib/(d0/2)2

Where ib = Emission Current, 100 – 200 A

d0= Diameter of the Crossover
(2 – 20 nm)
 Brightness
• Electron Beam Brightness, β is defined as the
Current density per unit solid angle.

β = current/[(area)(solid angle)]

= ib /[(0.25d02)(02)] A/m2sr

Where d0 = diameter of the beam at crossover at a gun

and 0 is the semi-angle of divergence at crossover. One
steradian is defined as the solid angle subtended at the
centre of a sphere of unit radius by a unit area on the
surface of the sphere.

Probe diameter, d0 = [ie/ 2]0.5

 Comparison of Filaments
Work Function Vacuum Crossover d0
(eV) (Torr) (nm)
Tungsten 4.5 10-4 – 10-5 50 X 103

Lanthanum 2.4 10-7- 10-8 10 X 103

Hexaboride

Field Emission 10-10 <10

Gun
It takes less than a minute for a monolayer to form at 10 -7 Torr; it takes
7 h at 10-10 Torr.
 Electron Sources
Beam Current Energy Spread 
Density (Am-2) (eV) (Am-2sr-1)
• Tungsten 5 X 104 3.0 109

• Lanthanum 1 X 106 1.5 1010

Hexaboride

• Field Emission 1 X 1010 0.3 1012-1013

Gun
Saturation of Filament
Undersaturated &
misaligned

Undersaturated Saturated
& aligned
 Electrostatic and Electromagnetic
Focusing
• Electrons are deflected by both electrostatic and electromagnetic
fields, and are brought to focus by engineering of the field
geometry.
• In the region of the electron gun, the electron beam is influenced
by electrostatic field created by anode and Wehnelt bias
cylinder.
• All subsequent focusing is electromagnetic, and is achieved by
using electromagnetic lenses equipped with soft iron (essentially
having zero magnetic hysterisis) pole pieces.
• Unlike optical lenses made from glass, the focal length of an
electromagnetic lens is variable, and can be controlled by
varying the lens current that flows in the coil surrounding the
pole pieces.
Wehnelt as Electrostatic Lens

Maximum brightness is
achieved at an intermediate
emission current and
intermediate bias.
 Path of electrons in TEM column
• An electron travelling in a magnetic
field is deflected in a direction at
right angles to the plane that
contains both the magnetic field
vector and the original direction of
travel of electron.
• In a uniform magnetic field, an
electron travelling off-axis follows a
helical path.
• Electrons of same energy and
travelling in a cone of directions
after originating from a point within
a uniform magnetic field is
converged to focus at a second point
after spiralling along optic axis.
 Path of electrons in TEM column
• The image of an object in a plane perpendicular to the axis of
electromagnetic lens is rotated about this axis.
• Focusing of the objective lens by adjusting lens current is
accompanied by rotation of the image about optic axis.
• Two images of the same object taken at different magnifications
will also be rotated with respect to one another.
• In light microscope, refractive index is constant. With
electromagnetic lens, deflection of electrons is continuous, and
the magnetic field varies continuously along the optic path
within the lens.
• Angle subtended by path of an electron with respect to optic axis
is < 1o  Optic path much longer than angular spread of
electrons.
• N.A. of objective in OM corresponds to  = 45 or 90o. N.A. for
electromagnetic lens is < 0.01.
Objective Lens Pole Piece
Link with Analytical Facilities
 Objective Lens Aberrations
Spherical Aberration

rs = Cs 3
Cs = 0.5 – 3 mm
Cs = 0.5 mm in JEM-2100 HRTEM (UHR)
 Objective Lens Aberrations
Chromatic Aberrations
Energy spread of e-beam, 3 eV
Energy Loss on passing through specimen, 5 – 50 eV

rc = Cc (E/E), Cc = 1 – 3 mm

Cc = 1.1 mm in JEM-2100 HRTEM (UHR)
 Objective Lens Aberrations
Astigmatism

-Exact alignment of lens components within the microscope column

is a critical factor in optimizing the performance of microscope.
-Variation in the focal length as the electrons spiral about the optic
axis. This leads to two principal focal positions on this axis that give
two line foci at right angles – called astigmatism.
-Astigmatism cannot be prevented because of: (i) inherent residual
asymmetry in lens and pole piece construction; (ii) extreme sensitivity
to minor misalignment, specimen geometry and contamination in
microscope.
 Objective Lens Aberrations
Astigmatism

rA = fA
 Astigmatism
• Objective lens is the most affected by misalignment. The
axial symmetry of this lens is especially sensitive to minor
disturbances associated with geometry, size, position and
dielectric properties of sample and carbonaceous
contamination.
• Complete two-fold astigmatism correction is achieved by
introducing sets of correction coils, whose magnetic fields
are at right angles to both the optic axis and the magnetic
field of the main lens coils.
• The correction coil currents can be periodically adjusted
during operation of the microscope to exactly balance
changes in magnetic assymetry due to build up of
contamination or displacement of sample.
 Effect of Diffraction
From Rayleigh criteria considering diffraction effects from nearest points appearing as discs,

Resolved Unresolved

Just Resolved

rd = 0.61/
Effect of semi-convergence angle on
spherical aberration & diffraction

At an optimum value of , the value of  is found to be the

minimum.
 Resolution
Increasing the aperture size increases the limit of resolution due
to spherical aberration and decreases the same because of
diffraction, optimum aperture size is:

 (or ) = (/Cs)1/4
Hence, rmin = 0.75 Cs0.25
 Polepieces and Coils
The distance between
the polepiece faces is
called the gap.
The bore-to-gap ratio
controls focusing action
of the lens.
Copper wire
surrounding polepiece-
As current is passed,
magnetic field is
created in bore.
The magnetic field is weakest on the axis,
Field is inhomogeneous
and increases in strength towards the side of
along length of the lens,
polepiece, so that the electrons are strongly
but axially symmetric.
deflected as they travel off-axis.
Objective lens is the most
important lens in TEM,
because it forms images and
diffraction patterns magnified
further by other lenses.
 Electron Ray Paths through
Magnetic Fields
• F = q(E + vXB) = -e(E + vXB)
• The Lorentz force, F is a vector
normal to v and B, which are
inclined to one another by angle,
.  is  90o for on-axis electron
entering lens.
• F = evBSin 90o = evB = mv2/r,
or r = mv/eB
• Considering relativistic
correction for v,
If V = 100 kV,
r = [2moE(1+E/2E0)]0.5/eB magnetic field,
Where mo and Eo are rest mass
and energy of electron. B = 1 Tesla;
r = 3.37X10-6[V(1+0.9778X10-6V]0.5/B r = 1 mm.
 Electron Ray Paths through
Magnetic Fields
The electron spirals through the field with a
helical trajectory.
The period of rotation is = cyclotron
frequency, w = 2/Tc = eB/m
With increasing B, pitch of helical path
becomes steeper.
Thus, the complete ray path through the
lens can be calculated.
Equations are called as “paraxial” (near-
axis) ray equtions, which describe r and .
The electron moves down in direction z,
rotating under the influence of
cylindrically symmetric field, B.
 Electron Ray Paths through
Magnetic Fields
Hawke’s equation states:
d2r/dz2 + 2B2r/2V1/2 = 0 (1)
d/dz = B/2V1/2 (2)
V = acceleration voltage of TEM.
= (e/2m0c2)1/2
Eq. (1)  rate of change of r is smaller
for more energetic electrons (larger
V), and larger for more intense field
strengths (larger B).

When energy of electron increases, stronger lenses can be used.

 The Lens Equation (Thin Lens)
• Distance from object plane to lens = u;
• Distance from the lens to image plane = v;
• The focal length = f
Considering the lens to be symmetric in strength on either side of
lens plane:
1/u + 1/v = 1/f

Magnification, M = v/u

M  semi-angle subtended at object () / semi-angle subtended at

image ()
 Strength of Lens – Focal Length

Above a certain magnification,

no further information is seen,
because other factors like
astigmatism limits resolution.

Strengthening of the lens shortens the focal length.

Weaker lens (f1) produces higher magnification (since M = v/u)
of the object.
Unlike light microscope, stronger lens magnify less and
demagnify more.
 Focussing

Overfocus: The stronger lens focuses before the image plane.

Underfocus: The weaker lens focuses after the image plane.
Focusing Conditions

Focused
Focusing Conditions

Under-focused
Focusing Conditions

Over-focused
Focusing Conditions

Astigmated Holey Carbon

 Axis Symmetric Tilt

Field of view remains constant during specimen tilts,

provided the area of interest is positioned at a point
defined by intersection of optic axis and tilt axis.
 Specimen Holder

Side-Entry Type Specimen Holders

Condenser Lens

Single Double
 Condenser Lens Optics

Parallel beam operation of TEM (conventional TEM work):  < 10-4 rad
Condenser Optics: Aperture

A smaller aperture creates a

more parallel beam at the
expense of total number of
electrons (smaller probe
current) reaching the specimen.

With smaller aperture,  is

reduced and the resulting
beam is more parallel.
 Condenser Lens Optics
Convergent Beam/Probe in TEM

Upper objective pole piece is used as

Focused C2 lens illuminating a 3rd condenser (C3) lens, and C2 is
small area of the specimen with switched off. Large do/di: high
convergent beam: large  demagnification
 Condenser Optics: Effect of C1 Lens
A strong C1 lens
causes greater
demagnification by
subsequent (C2 or
C3) lenses – causes
smaller probe size.
di of C1 is decreased,
while do of C2 is
increased.
Changing strength of
C1 also alters the
fraction of electrons
passing through the
condenser aperture.
Electrons emerging from exit
surface of specimen are
dispersed by objective lens to
form diffraction pattern in back
focal plane. These are
recombined to form image in
image plane.
Diffraction Mode:
Intermediate lens is adjusted to
select the back focal plane as its
object. The diffraction pattern is
projected on viewing screen.
Image Mode:
For looking at the image, the
intermediate lens is readjusted
so that its object plane is the
image plane of the objective Diffraction Mode Image Formation
lens.
 Selected Area Diffraction
• There are two ways for reducing the illuminated
area of the specimen:

– The beam can be made smaller: Use of C2 (and C3)

lens to converge the beam at the specimen.

– An aperture can be inserted above specimen so as to

permit electrons from a small area of specimen to
form DP: Use of selected area aperture in the image
plane of the objective lens  virtual aperture at the
plane of the specimen.
 Selected Area Diffraction (SAD) Aperture
• When the SAD aperture
inserted and objective
aperture is removed, electrons
hitting the specimen outside
the area defined by virtual
aperture will hit the
diaphragm, when it travels
into the image plane.
• Only electrons inside the
dimensions of virtual
aperture at entrance surface
of specimen are allowed to
contribute to SAD pattern.
 Selected Area Aperture: Spherical Aberration
Use of an SAD aperture to select an area in
the image plane has an added advantage: The
area is magnified 25x.
A 50 m diameter aperture selects a 2 m
diameter area on the specimen.
A smaller SAD aperture is not helpful,
because of the spherical aberration of the
objective lens.
The beams farther away from the optic axis
are bent more strongly.
For rays entering objective lens at angle  to
optic axis, the image formed is displaced by rm
= MCs3
The amount of displacement increases with
increasing  (Problem for dark field imaging)
The area selected by SAD aperture
corresponds to PP1 only for the direct beam.
 Ray Diagram for Imaging and
Diffraction
To get diffraction pattern:
Selected area aperture is
introduced, and objective
aperture is taken out.
The intermediate lens is first
focused on the objective
image plane to reveal the
specimen selected area.
Strength of the intermediate
lens is reduced until it is
focused on back focal plane
of the objective, to provide
the diffraction pattern.
Image and Selected Area Diffraction

Objective Aperture

Selected Area
Aperture
Diifraction Pattern from Single
Crystal of Aluminium

[011] Zone Axis

 Imaging
• The SADP contains bright
central spot and spots from
scattered electrons.
• To view an image formed
in the image plane of
objective lens, an aperture
(Objective aperture) is
inserted into the back focal
plane of the objective lens.
• Put aperture around direct
beam to form bright field
image.
Bright Field Dark Field
• Put aperture around
diffracted beam to form
dark field image.
 Bright Field and Dark Field Imaging

Size of objective aperture controls the collection angle, , which in

turn controls the aberrations.
BRIGHT AND DARK FIELD
Beam Tilt DF Poor man’s DF

Diffracted beam Diffracted beam inclined to

aligned with optic axis: All image points
optic axis will be elongated by
3Cs(g)2
Multiple dark field images Increase strength of
from diffraction pattern: intermediate lens to
Diffracted beams contributing to focus below the back
a specific dark field image can be focal plane (PQ) of
identified. objective lens 
each diffracted spot
is broadened to give
low mag image of
selected area.

Image - Not focused

Adjust objective
lens current, after
object plane of
intermediate lens
is fixed at RS.

Not
focused Focused Image is Focused
 Beam-Specimen Interaction
• The object can be considered as a geometrical plane,
perpendicular to optic axis – can be visualized as the
electron density distribution on the bottom side (the side
facing the objective lens) of a flat specimen.

• The electron density distribution occurs because the

incident electron beam of the initially homogeneous
distribution interacts with the specimen, and the
interaction varies from place to place within the specimen
depending on local structure.
 Distribution of Intensity
• When an incident electron beam of intensity I 0 falls on the
specimen,
(1) a fraction Ir is reflected from surface;
(2) Some electrons represented by Ia are captured by
specimen (absorption);
(3) Some electrons are scattered by specimen either
elastically (Ies) or inelastically (Iis);
(4) Rest of the electrons pass through the specimen (I t) along
the incident direction.
 I0 = Ir + Ia + Ies + Iis + It
 Scattering
• Electrons are scattered by the electrostatic field of the atoms
constituting a given material.
• Electrostatic field of an atom is generated either by the positive field of
atomic nuclei or by negative field of electrons surrounding the nuclei.
– Scattering by electrostatic field of nucleus – elastic
– Scattering by electrons surrounding nuclei – inelastic
• When the scattered electrons are stopped by objective aperture, a
contrast is produced in the image due to scattering – contrast of the
image is enhanced.
• Elastic scattering from regularly (periodically) arranged atoms  a
systematic interference occurs among the wavelets scattered from the
atoms. This phenomenon is known as “Diffraction”.
• It is generally believed that diffraction is caused by elastically scattered
electrons.
• Diffraction can also be caused by electrons scattered inelastically.
 Kinematical & Dynamical
Diffraction
• When the material scattering the electrons is crystalline, or
when the electrons are scattered by a regular array of
atoms, a discrete angular distribution of scattered electrons
occurs due to constructive interference resulting in
diffraction.
• Scattered electrons can often be re-scattered by atoms on
their path. Re-scattering is more frequent in the thicker
specimens.
• Single scattering events – Kinematical diffraction theory.
• Multiple scattering events – Dynamical effects.
 Kinematical Assumptions
• Only two waves, namely, the incident wave and the
scattered wave are considered, and no interaction between
the two waves exist. The incident wave is not altered either
on entering the specimen or thereafter.

• Re-scattering of the scattered wave does not take place.

• No other interactions such as absorption, inelastic

scattering, etc. are considered.
Diffraction in Perfect Crystals
 Lattice Translation Vector
• Taking the origin at an appropriate lattice point,
the position of all atoms in the crystal can be
represented by a lattice vector defined as:
• rijk = ia + jb + kc
where i, j and k are positive or negative integers
including zero; a, b and c are the lattice vectors
defining the unit cell of the crystal structure.
 Basic Assumptions
• Electrostatic field of an atom is spread over a certain volume, it is
assumed to be represented by a point. (Replacing point scattering
centres by a true electrostatic field distribution around an atom is
merely to multiply the diffraction function by an additional slowly
varying function.)

• Material diffracting the electrons has a perfect crystal structure. (Point

scattering centres are situated at lattice points).

• The basic process in diffraction phenomenon is the interference

among the wavelets scattered in the same direction from scattering
centres.
Bragg’s Law

Scattering centres
2d(hkl)Sin  = n at O and L
Two beams are scattered from the
points, E and F lying on the same
plane.
As triangles EHF and FJE are
congruent, EJ = HF, and therefore
the path difference = 0.
The plane P3 lies half way Higher order reflections, ng
between P1 and P2
The planes, P1, P2 and P3
scatter in phase, if
2(d/2) sin  = 
The d-spacing between planes =
d/2
Coherent scattering occurs,
when g2 = 2/d, i.e., g2 = 2g
Scattering from plane, P3 occurs
no matter how atoms are
distributed in it. Even if no atom
is present.
Similarly, g3 = 3g & gn = ng
Dynamical Diffraction:
A beam which is oriented for Bragg
condition is re-scattered repeatedly.
The scattered beam is diffracted back
into the incident beam by same set of
planes, and is again diffracted.
 Path Difference
• The path difference = r = s.rijk – s0.rijk
where s0 and s are the unit vectors indicating the incident
and scattering directions.
 r = S.rijk
where S = s – s0
• Since s denotes the scattering direction, S is related to
direction of diffraction under consideration.
• Wavelength remains constant during the scattering
process. This implies that the phase difference between
two wavelets =  = (2/)S.rijk
  = K.rijk
where K = k – k0
and k = (2/)s, k0 = (2/)s0
 Interference
• A plane harmonic wave of k can be written as:
  = a exp ik.(r-ut)
• For the wave having phase difference of , the wave function
is:
  = a exp [ik.(r-ut) + i] = a exp [ik.(r-ut) + iK.rijk]
• Therefore, the resulting wave is the linear combination of the
two wave functions:

t =  +  = a exp ik.(r-ut)[1 + exp (iK.rijk)]

• Destructive interference occurs when exp (iK.rijk) = -1;  No

electron is scattered in the direction defined by K.
• Constructive interference occurs with maximum intensity,
when exp (iK.rijk) = 1;  Maximum density of electrons is
expected in the direction defined by K.
• A number of wavelets from all scattering centers should be
considered 

 = a[ [exp iK.r ijk ] exp ik.(r - ut)
ijk
n=n1-1 n=n2-1 n=n3-1


 = a[ [exp i nK.a] [exp j nK.b] [expk nK.c] exp ik.(r - ut)
n=0 n=0 n=0
• The origin is chosen such that the scattering centres are expressed
only by positive integers including zero.
• The summations inside the square brackets are geometrical
progressions having the common ratios of exp iK.a, exp iK.b and exp
iK.c.
 a0rm (from m = 0 to m=n-1) = a0(1-rn)/(1-r), where a0 and r are the
first term and common ratio, respectively.
  = a[(1-exp in1K.a)/(1-exp iK.a)] [(1-exp in2K.b)/(1-exp iK.b)] [(1-
exp in1K.c)/(1-exp iK.c)].exp ik.(r-ut)

• The density of electrons (or intensity of diffraction) I d is the product of

 and *

• Id = .* = a2[(1-cosn1K.a)/(1-cos K.a)]. [(1-cosn2K.b)/(1-cos K.b)].

[(1-cosn3K.c)/(1-cos K.c)].

• Id = a2[(Sin2 n1K.a/2)/(Sin2 K.a/2)]. [(Sin2 n2K.b/2)/(Sin2 K.b/2)]. [(Sin2

n3K.c/2)/(Sin2 K.c/2)].
• The intensity can be visualized in reciprocal space.
• Maximum intensity occurs for K.a = 2n
• The first minimum occurs at K.a = 2/n
Bragg Condition (Laue Equations)

• K.a = 2h
• K.b = 2k
• K.c = 2l
where h, k and l are integers
Reciprocal lattice vector, ghkl = ha* + kb* + lc*
where a*, b* and c* are unit reciprocal lattice
vectors.
 Reciprocal Lattice
a* = (b X c)/Vc; b* = (c X a)/Vc; c* = (a X b)/Vc
where Vc is the volume of the unit cell.
Vc = a.(bXc) = b.(cXa) = c.(aXb)

For orthogonal axes (e.g.

cubic, tetragonal,
orthorhombic),
directions of a and a*
are parallel.
 Reciprocal Lattice Vectors
• Since a reciprocal lattice consists of repeating points satisfying
the diffraction condition, it can be represented by the reciprocal
lattice vector as:
 ghkl = ha* + kb* + lc* where a*, b* and c* are unit reciprocal
lattice vectors.
• When the reciprocal unit vectors have the following relation to
real lattice vectors, a, b and c, the reciprocal lattice points
uniquely represent the diffraction condition. These relations are:
a.a* = b.b* = c.c* = 1
a.b* = a.c* = b.a* = b.c* = c.a* = c.b* = 0
• The above equations define the reciprocal lattice, where the
vectors (a*,b*,c*) have one-to-one correspondence with real
lattice vectors (a,b,c).
• ghkl.a = h; ghkl.b = k; ghkl.c = l;
• As K.a = 2h is the Bragg diffraction condition,
  K = 2ghkl is equivalent to diffraction condition.
 Reciprocal Lattice Vectors
• K = 2ghkl means the coincidence of K/2 with vector, g.
• Thus, k0/2 indicates the incident direction, and the length of the
arrow, k0/2 is proportional to 1/.
• Consider point, O as the origin of both real and reciprocal lattice,
and point O in the incident direction, such that OO = 1/
• The sphere having O as the centre and OO (=1/) as the radius
can be defined in reciprocal space, and this sphere is called the
“Ewald Sphere” or “Sphere of Reflection”.
• The vector from origin, O to any point, P on the Ewald Sphere is
equal to K/2 (=ghkl), and the corresponding direction of scattering
is along vector k/2.
Bragg’s Law: Derivation from Laue Conditions
A
The vector K can be considered
as perpendicular to the
hypothetical plane AB, which
appears to reflect the incident
wave, k0 to the direction, k.

Diffraction occurs when AB

coincides with the real plane
containing scattering centres.

K/2 = 2(k/2)sin  = 2sin /

When the diffraction condition
B is satisfied, K/2 = ghkl = 1/d
Combining the above relations,
the familiar Bragg condition
 Diffraction from a set of planes
separated by distance d

The vector, g is perpendicular to the diffracting plane.

Ewald Sphere intersecting array of
diffraction spots
 Reciprocal Lattice Vector
• The reciprocal lattice vector, ghkl is related to the real
crystal plane (hkl) as follows:
(1) ghkl is perpendicular to (hkl) plane.
(2) Abs (ghkl) = 1/dhkl
where dhkl is the spacing between adjacent (hkl)
planes.
 Reciprocal Lattice
Consider plane ABC with Miller
indices (hkl) and n as the
normal unit vector.
Reciprocal lattice vector, g(hkl)
is perpendicular to plane ABC
with Miller indices, (hkl)

dhkl = n.a/h = [(ha* + kb* +lc*)/g(hkl)].a/h = 1/ g(hkl)

 Real lattice vector = 1/Reciprocal lattice vector
Real and Reciprocal Lattice
FCC Unit Cell in Real and Reciprocal
Space along [001] zone axis

Real Lattice Reciprocal Lattice

(fcc) (bcc)
Ewald Sphere or Sphere of
Reflection, Radius = 1/
Bragg’s Law is satisfied on the
Ewald’s Sphere.
Ewald Sphere

Bragg’s Law is satisfied on

Ewald’s Sphere.
 Ewald Sphere
Sin  = (1/2d) / (1/)
Hence,  = 2dhkl sin
(1/)/L = (1/d)/OP’
(Since OOP is similar
to OOP)
OP  OP = R for small

 L = Rd = Camera
Constant (Equation used
for indexing diffraction
patterns)
If L and  are known and R is measured, d can be calculated.
 Diffraction Pattern from fcc
Polycrystalline Sample

Sequence: (000), (111), (200), (220), (311), (222),…

 Diffraction Patterns from
Au Polycrystals

(111)

(200)

(220)
(311)
Camera Length, L1 Camera Length, L2

L2 > L1, and d(hkl) = L1/R1, = L2/R2

Diffraction Pattern from
Amorphous Solid
Diifraction Pattern from Single
Crystal of Aluminium

[011] Zone Axis

 The Zone Law

Normal to a plane (hkl) is perpendicular to any direction [uvw]

contained in it. Hence [hkl].[uvw] = 0
Thus hu + kv + lw = 0 : Zero Order Laue Zone
 Atomic Scattering Factor
• Electrons are scattered by the electrostatic field of an atom,
which is extended over a volume surrounding the atom.
• Electrostatic potential is a function of position, and
efficiency of electron scattering should change with position
Thus, the weighing factor, V(ra) can be defined, where ra is
distance from centre of atom.
• Consider wavelet being scattered from centre of the atom
the wave function is:
 mijk = V(ra).dr. exp[ik.(r-ut) + i],
where  = K.(rijk + ra) and rijk is the lattice vector.
• mijk = V(ra).dr. [exp iK.ra].[exp iK.rijk].[exp ik.(r-ut)]
• The 3rd square bracket term is called “carrier wave” and
does not affect the intensity expression, and can be omitted
from the expression for diffraction amplitude.
 mijk = [V(ra).dr. exp iK.ra].[exp iK.rijk]
 mijk = [V(ra).dr. exp iK.ra].[exp iK.rijk]
• The first term is integration or summation over an
atomic volume, and hence it must be characteristic
of each type of atom. As it indicates efficiency of
atom for scattering of electrons, it is called atomic
scattering factor, fijk (K).
  = ijk = fijk exp iK.rijk
• If only one species of atom is involved, f ijk has a
unique value, f.
  = f.  exp iK.rijk
 Structure Factor
• Any lattice structure can be considered as consisting of a number
of primitive translational lattices interpenetrating each other.
• Atomic positions can be expressed by combination of (i) lattice
vector, rijk = ia +jb +kc; and the vector indicating translation of
one primitive translational lattice from another.
(ii) r0 = m1a + m2b + m3c, where m1, m2 and m3 are fractions of
unity.
• Sum of wavelets from a unit cell is:
ijk = f.exp i.K.rijk + f0 exp iK(rijk + r0)
• F(K) = f + f0 exp iK.r0
• When a number of primitive translational lattice interpenetrate
each other:
• F(K) = f + f0 exp iK.r0 + f1 exp iK1.r1+… = f + fp exp (iK.rp),
where rp = mp1a + mp2b + mp3c
• If lattice contains only one species of atoms, f = f0 = f1 = ..=fp
Total Amplitude of the Scattered Wave
• The total amplitude is the product of three
terms: the atomic scattering factor (which is
a function of scattering direction), the
structure factor and the diffraction function.
  = f(K)(1 +  exp iK.r )( exp iK.r )
p ijk
• Physical origin of diffraction function is
repetition of unit cells.
• One needs to know the atomic scattering
factor and structure factor only in the
neighbourhood of K = 2ghkl i.e.near
reciprocal lattice points.
 Structure Factor
• For bcc lattice:
• There are two interpenetrating primitive
translational lattices  F = f + f0 exp iK.r0
• r0 = 1/2a + 1/2b +1/2c
• In the vicinity of K = 2g,
• K.r0 = 2g.r0 =  (h + k + l)
• Thus, F = f + f0 exp i (h + k + l)
f + f0 for h + k + l = even number
f – f0 for h + k + l = odd number
• If lattice points are occupied by similar atomic
scattering factors, F = f [1 + exp i (h + k + l)]
= 2f for h + k + l = even number
= 0 for h + k + l = odd number
 Face-Centred Cubic
• (x,y,z) = (0,0,0), (1/2, 0, ½), (1/2,0,1/2) & (0,1/2,1/2)
• F = f0 + f0exp i(h+k) + f0exp i(k+l) + f0exp i(l+h)
• If the values of h, k and l are either odd or even, then all
the exponential terms are exp 2ni  the diffracted waves
are in phase.
– F = 4f0
• If one of h, k or l even and other two are odd, then two of
the terms will have odd multiples of , giving rise to two
terms which will have opposite (negative) sign with
respect to other two.
– F=0
 Structure Factor
• Simple cubic: all reflections present
• fcc (Al, Cu, etc.): reflections absent if h, k, l
are mixed .
• bcc (V, W, -Fe) : reflections absent if h +
k + l = odd.
• bct (martensite): reflections absent if h + k
+ l are odd.
Structure Factor (F) = amplitude of wave scattered by all atoms
of a unit cell/amplitude of wave scattered by an electron.
 Structure Factor for NiAl
• NiAl has B2 structure (Ordered BCC), such that the atom at
(000) is Ni and that at (½ ½ ½) is Al.
• Since the atomic scattering factors of Ni and Al are different,
the diffracted intensity is proportional to [fNi +
fAlcos{(h+k+l)}]2 +[fAlsin{(h+k+l)}]2
• Intensity is proportional to (fNi + fAl)2 when h+k+l is even, and
• Intensity is proportional to (fNi - fAl)2 when h+k+l is odd.
• {001} reflection will occur with an intensity proportial to the
difference in the scattering factors.
 Hexagonal Close-Packed
• Diffraction patterns are more difficult to index:
• Except for (0001), the patterns are different for every material, because of different c/a
ratios;
• While we use three-index notation to derive structure factor rules, four-index Miller-
Bravais notation is used to index lattice planes or diffraction patterns.
• HCP unit cell has two atoms located at (0,0,0) & (1/3, 2/3, ½).
• F = f0 + f0 exp 2i(h/3 + 2k/3 + l/2)
• Consider h/3 + 2k/3 + l/2 = X (X could be a fraction)
• F2 = f02.(1+exp 2X).(1+exp -2X) = f02(2+exp 2iX + exp -2iX)
• F2 = f02 (2 + 2cos 2X) = f02 (4cos2X)
– F2 = 0, if h + 2k = 3m and l is odd.
– F2 = 4f02, if h + 2k = 3m and l is even
– F2 = 3f02, if h + 2 k = 3m1 and l is odd
– F2 = f02, if h + 2k = 3m1 and l is even.
 11-20 and 11-26 reflections will be strong, while 11-23 will be weak. Moreover 0001 spot will be
missing.
 Hexagonal Close Packed
• If the direction [uvtw] lies in the (hkil) plane,
then uh + vk + ti + wl = 0
• The normal to the (hkil) plane is cartesian vector [hki(l/)] 
[uvtw] is [uvt(w)] in cartesian system, where 2 = (2/3)(c/a)2
• Using cartesian vector notation,
[uvt(w)].[hki(l/)] = 0
Therefore, the vector normal to (hkil) plane is the vector,
[hki{(3/2)0.5(a/c)l})
[11-20] is normal to (11-20) plane, because l = 0, but
[01-12] is not normal to (01-12) plane.

Angle between normals to planes (hkil) and (defg):

Cos  = [hd+ke+0.5(he+kd)+0.75lg(a/c)2] / [{h2+k2+hk+3/4l2(a/c)2}.
{d2+e2+de+3/4g2(a/c)2}]
 Size of Diffraction Patterns

• R = L/d where R is the distance from

central spot,  is the electron wavelength,
and L is the distance of the specimen from
screen or plate.

• If diameter, D of the ring pattern is

considered, D = 2L/d
 Plotting of Ring Patterns
• For randomly oriented aggregates of polycrystals, the reciprocal
lattice becomes a series of spheres concentric with the origin of
the reciprocal lattice.
• The radius of each sphere is proportional to 1/dhkl, and a number
of equivalent reciprocal lattice points contribute to each sphere.
• For cubic crystals, dhkl = a0/(h2+k2+l2)1/2 where a0 is the lattice
constant, h,k & l are the Miller indices.
• Rhkl = (L/a0)(h2+k2+l2)1/2
• As the value of (L/a0) is constant for a given diffraction
pattern, the values of Rhkl vary according to (h2+k2+l2)1/2 .
• The allowed reflections for fcc are either fully even or odd: The
order of rings in increasing radius are: {111}, {200}, {220},
{311}, {222}, {400}, {331}, {420}, {422},….and the radii are
proportional to 3, 4, 8, 11, 12, 16, 19, 20, 24,…..
 Diffraction Patterns from
Au Polycrystals

(111)

(200)

(220)
(311)
Camera Length, L1 Camera Length, L2

L2 > L1, and d(hkl) = L1/R1, = L2/R2

 Plotting of (uvw) Reciprocal Lattice Plane

-All points hkl lying in the uvw reciprocal lattice plane obey the
relation: hu+kv+lw = 0
-Choose two low index points h1k1l1 & h2k2l2, which satisfy this
condition.
-Magnitudes of reciprocal lattice vectors are proportional to
(h2+k2+l2)1/2
-Parallelogram ORPQ is completed to give P, Q & R as h1k1l1,
h2k2l2 & h1-h2,k1-k2,l1-l2, respectively, and this pattern is extended in
 Plotting of (321) Reciprocal Lattice Plane

• By inspection, (1 -1 -1) is a low index point in this plane.

• The perpendicular spot (h2 k2 l2) is determined by putting the
condition that h1h2+k1k2+l1l2 = 0, and adding the condition that
h2k2l2 lies in (321) plane.
• Thus, h2 - k2 - l2 = 0, &
3h2 + 2k2 + l2 = 0,
Solving these equations gives h2 k2 l2 = c(1 -4 5). However, (1 -4 5) is
 Indexing Unknown Diffraction Patterns
The substance giving the diffraction pattern is known, but its
orientation is unknown.
Information required: A table
of interplanar d-spacings –
ICDD/JCPDS

Consider three points, P1, P2

& P3 together with origin O.
Measure d values for these
Since these points lie in uvw points and determine the
reciprocal lattice plane, indices of each point. By trial
& error, a consistent set of
h1u+k1v+l1w = 0 indices is chosen such that
[h1k1l1] = [h2k2l2] + [h3k3l3]
h2u+k2v+l2w = 0
 Indexing of Spots
• The spot closest to the centre, (h2 k2 l2) is arbitrarily indexed for a given
combination of h, k and l.
• Consideration is then given to two other spots, (h 1 k1 l1) and (h3 k3 l3) such that: h2
= h1 + h3; k2 = k1 + k3; l2 = l1 + l3.
• The angles, 1 and 2 between lines from the centre spot to the two spots
measured, and the correct combinations of (h1 k1 l1), (h2 k2 l2) and (h3 k3 l3) may
be checked using the formula:
• Cos  = [h1h2+k1k2+l1l2]/{[h12+k12+l12]1/2[h12+k12+l12]1/2}

• Determination of Zone Axis, B: The zone axis [uvw] = B

• u = k1l2 – k2l1
• v = l1h2 – l2 h1
• w = h1k2 – h2k1
• Expressing the result in terms of reciprocal lattice vector, g corresponding to (h1
k1 l1 ), i.e. g1 and (h2 k2 l2), i.e. g2
• The zone axis, B = g1 X g2
Diffraction Pattern from fcc
Single Crystal

[uvw] = [001] Zone Axis

Diffraction Patterns - fcc
Diffraction Patterns - fcc
Diffraction Pattern - bcc
Diffraction Pattern - bcc
Diffraction Pattern - hcp
Diffraction Pattern - hcp
 Effect of Curvature of the Sphere of
Reflection - Laue Zones

The radius of sphere of reflection is 27 Angstroms-1 for

acceleration voltage of 100 kV
 Relaxation of Diffraction Condition
• The diffraction condition given by K.a = 2h, K.b = 2k &
K.c = 2l is the condition for maximum intensity.
• Appreciable intensity of diffraction occurs even if the above
condition is not strictly satisfied  diffraction condition is
relaxed.
• A small volume surrounding each reciprocal lattice point
represents the condition for appreciable diffraction intensities.
• The diffraction intensity depends upon distance the distance
from the reciprocal lattice point, and this can be visualized by
distribution of intensity with variation of K.a.
 Causes of Relaxed Diffraction
• Limited size of the specimen or small particles/precipitates cause
broadening of the diffraction peaks  Peak broadening (steaking
of diffraction spots, e.g. rel-rods or rel-plate) occurs along the
direction of specimen thickness or perpendicular to smallest
dimension of particle/precipitate.
• Thin foil specimen is often buckled  Orientation of specimen
changes from point to point.
• Lattice imperfections produce local disorder of atomic
arrangement, which in turn relaxes the diffraction condition.
• The incident beam is not perfectly collimated, resulting in the
angular spread of the incident beam.
• If the incident radiation is not strictly monochromatic, then the
sphere of reflection has a range of diameters corresponding to the
existing spectrum of wavelengths  Single Ewald sphere is
replaced by more than one spheres of refection.
Ewald Sphere or Sphere of
Reflection, Radius = 1/
Bragg’s Law is satisfied on the
Ewald’s Sphere.
 The width of reciprocal lattice
points in x, y and z directions
are 2/Nx, 2Ny & 2/Nz (Nx, Ny
& Nz are number of unit cells
along x, y and z directions,
respectively.)
The thickness along z-
direction is finite (~500 unit
cell), whereas it is infinite
along x & y directions.
Consequently, reciprocal In contrast, intensity
lattice points form streaks distribution along z-direction
parallel to z, the foil normal. is much wider than that in x &
y-directions..
Relaxation of Laue Condition along z-direction

Diffraction intensity will be obtained, even if the Bragg condition

is not exactly satisfied.
A vector s is defined describing the deviation from the exact
Bragg condition, when Ewald sphere cuts the streak.
As s increases, the diffracted intensity decreases, g = g + s
Diffraction Pattern Streaks
Streaking of Diffraction Spots
Diffraction from Mo-Si Multilayered Film
 Scattering of Electron Beam

Most of the incident beam is elastically scattered by the sample:

randomly in glassy or amorphous specimen; or coherently to
discrete angles (as in a crystalline phase).
Image may be formed from: (i) transmitted beam, (ii) diffracted
beam, or (iii) interference of diffracted beams.
Origin of Contrast in TEM Images

• Contrast arises from distinct image-forming

processes, which are termed as:
– Mass-thickness contrast
– Diffraction contrast
– Phase contrast
 Mass-Thickness Contrast
• The probability of an electron being elastically scattered
out of incident beam depends on atomic scattering factor
(which increases monotonically with atomic number) and
total number of atoms in the path of beam (total thickness
of specimen).
• Mass-thickness contrast reflects a combination of
variations in specimen thickness and density.
• Mass thickness contrast dominates the images formed for
biological specimens, amorphous specimens, replica, or
extracted phase.
• Intensity scattered out of the direct beam depends on the
energy of the electron beam, sample thickness & sample
density.
 Diffraction Contrast
• In case of perfect crystal, the contrast is associated with the
amplitude scattered from the incident beam into diffracted beam
by diffracting planes with a specific reciprocal lattice vector, g.
• This amplitude can be calculated by summing the diffracted
amplitudes from all unit cells that lie along the path of the
diffracted beam.
• The interference of wavelets can be shown schematically using a
“Phase-Amplitude Diagram”.
• Wave function can be re-written as:  = a exp i
= a (cos  + i sin ), where  = k.(r-ut) = 2g.(r-ut) = is the
phase difference in the amplitude scattered by a unit cell that is
at position r with respect to the origin in the column of material
responsible for contrast.
• When a number of wavelets are involved, the vectors
representing each wavelet are added on the diagram as:
  =  exp i =  cos  + i  sin 
aj j aj j aj j
 Phase-Amplitude Diagram

When diffraction condition is fully satisfied, the phase

amplitude diagram is a straight line, as there is no phase
difference between the wavelets.
 Deviation from Diffraction Condition
• When diffraction condition is nearly satisfied,
K/2 = g + k/2 (or s), where k denotes
the deviation from the diffraction condition.
 Deviation from Diffraction Condition
 ijk = F exp i(2g+k).rijk
= F exp i(2g.rijk).exp i(k.rijk)
where rijk = i1a + j2b + k3c and F is the scattering efficiency
of the unit cell.
According to the definition of reciprocal lattice vector, exp
i(2g.rijk) = 1
 ijk = F.exp i(k.rijk) = F.exp i

Since rijk is the multiple of unit lattice vectors,

j - j-1 = j-1 - j-2 = …..= 2 - 1 = j - 0 =  (This is one-
dimensional submission).

For three dimensions, k.rijk = n1 a.k + n2 b.k + n3 c.k

= n11 + n22 + n33
 Phase-Amplitude Diagram
  = F j from j=0 to j=n-1
  = F  exp i(j ) from j=0 to
j=n-1
  = F.[(1 – exp in)/(1 – exp i
)]
Due to conditions of constant F
and , the phase amplitude
diagram becomes a regular
polygon.
 Phase-Amplitude Diagram
Taking P at the center of the
polygon, the vector
representing each wavelet and
P forms an isosceles triangle.
OP and PQ also form an
isosceles triangle.
Sin /2 = F/2R
Sin n/2 = Abs ()/2R
where R is the radius of
polygon.

Eliminating R, one obtains Abs () = F [(Sin n/2)/(Sin /2)]

Total amplitude, Abs () has its maximum as  approaches zero.
 Phase-Amplitude Diagram
• On the phase-amplitude diagram, the polygon increases its
radius and becomes a straight line as the diffraction condition
is exactly satisfied.
• When F is relatively small compared to radius of polygon, the
polygon can be approximated by a circle, & Sin /2 = /2
Abs () = F [(Sin n/2)/(Sin /2)]
becomes: Abs () = F.2(sin n/2)/
• Sin /2 = F/2R becomes  = F/R
- The radius of a circular phase-amplitude diagram is inversely
proportional to the deviation from diffraction condition,  or K.a.
-Therefore, radius of the circle becomes infinite as  approaches
zero – Kinematical Assumption
-When polygon is approximated by circle, summation becomes
integral.
Kinematical Theory of Diffraction Condition
• It = I0 – (Ir + Ia + Ies + Iis)
• Ies and Iis cause contrast only if electrons are
prevented from reaching the screen by objective
aperture.
• If only elastic scattering is considered, I es = Id =
intensity of diffracted beam
• I t = I0 – I d
• Bright field image, if objective aperture is kept to
pass the transmitted beam;
• Dark field image, if objective aperture is kept to pass
the diffracted beam.
• Bright field contrast is complimentary to dark field,
if a single diffraction causes the image contrast.
Transmitted and Diffracted Beams

Bright Field Dark Field

BRIGHT AND DARK FIELD
Beam Tilt DF Poor man’s DF

Diffracted beam Diffracted beam inclined to

aligned with optic axis: All image points
optic axis will be elongated by
3Cs(g)2
 Column Approximation
• A foil specimen is divided into
columns of suitable size, which are
approximately parallel to the
incident direction.
• Diameter of each column is several
unit cell distances.
• The column diameter is >> t sin
2, where t is the thickness and  is
the Bragg angle.
• Interference is considered only
among the wavelets scattered from
scattering centers within a column.
• Atomic arrangement is more or
less uniform from one to another
end of the column.
 Contrast in Perfect Crystal

• A perfect crystal is defined as a crystal,

which does not contain lattice defects such
as point defects, dislocations, stacking
faults, or grain boundaries.
• Contrast arises due to: (i) Buckling; (ii)
Change in thickness.
 Deviation from Diffraction Condition
• When diffraction condition is nearly satisfied,
K/2 = g + k/2 (or s), where k denotes
the deviation from the diffraction condition.
Deviation from Bragg Condition
g = g + s

s = 0, Exact Bragg Condition

s < 0, Deviation from Bragg

condition

s > 0, Deviation from Bragg

condition
 Thickness Contours

Bright Field

Schematic Representation Dark Field

 Phase-Amplitude Diagram
• A polygonal phase-amplitude diagram can be
approximated by a circular phase-amplitude diagram,
when amplitude of a component wavelet is negligibly
small with respect to that of polygon.
 ijk = F exp i k.rijk can be re-written as:
 j = F exp i k zj, where k is the component along z axis
taken as along the column, and zj is a discrete distance
along the z-axis.
• The z-axis taken along the diffraction direction, k. Since
specimen foil is about perpendicular to incident beam, the
z-axis may also be considered along the incident beam.
• Summation of wavelets along the column:
  = F  exp i k z can be shown as integral along the
j
column.
 Summation by Integration
• Taking origin, z = 0 at the middle of column,
integration can be carried out (along the length of
column) from z = -t/2 to z= t/2
  = F [exp ikz / ik] from z = -t/2 to z = t/2
  = F {2sin (k t/2)} / k
• The maximum amplitude occurs when k.t = 
• Therefore max = 2F/k
• Considring that max is the diameter of the circular
phase amplitude diagram, R = F/k
 Thickness Contours
• When a crystal of uniform thickness is perfectly flat,
every column has exactly same intensity of
diffraction  No contrast results.
• When there is only a slight deviation from
diffraction condition, phase amplitude diagram is
circular.
• Depending on thickness of the specimen, the vector
representing diffraction amplitude can be decreased
from the diameter of the circle (maximum
diffraction amplitude) to zero.
 Thickness Contour: Phase Amplitude Diagram
• The oscillation of diffraction amplitude is
shown by vector OP as P changes its position
around the circle.
• In bright field microscopy, the dark bands
correspond to maximum intensity of
diffraction, whose amplitude is the diameter of
the circular phase-amplitude diagram.

-The equation, k.t =  defines the unit thickness in which the

diffraction amplitude oscillates periodically.
-Thickness of a specimen can be determined from a set of thickness
contours, if the value of k is known.
 Thickness Contour: Phase Amplitude
Diagram

• The equation k.t =  defines the unit thickness in which the

diffraction amplitude oscillates periodically;
• This relation is particularly important when the dynamical effect
is considered. Accordingly, the phase-amplitude diagram
converges to a circle of definite diameter as the exact diffraction
condition is approached.
• When the diffraction condition is satisfied, the unit thickness
is characteristic only of the material and the diffraction
condition (i.e. g vector) which is operating.
• This unit thickness is known as the “extinction distance.”
 Thickness Contours
• For a wedge-shaped specimen, the dark bands which
appear approximately parallel to the edge of the
specimen are called as “thickness contours”.
Buckling - Phase Amplitude Diagram
for Bent Crystal

• When a specimen is buckled, the radius of phase

amplitude diagram changes from column to
column, according to R = F/k as k changes.
• Diffraction amplitude changes along the specimen
plane.
• The portion of the specimen, which diffracts
strongly appears dark in bright field is known as:
BEND CONTOUR.
Bent Crystal
Beam Intensity along
Bent Crystal
Images of Bend Contours
 Bend Contours
(Bright and Dark Field Images)
 Bend Contour
• The dark regions appear as bands, which
contain a fine structure consisting of parallel
fringes.
• The fine structure is clearly seen in dark field,
where the contrast is reversed.
 Bend Contours

Dark Field Image showing Bend Contour and Stacking Fault

Contrast of Imperfect Crystals
• Dislocations
• Stacking Faults
• Antiphase Boundaries
• Grain Boundaries
 Dislocations

Bright Field Image showing dislocations, dislocation jogs,

dipoles, grain boundary ledges and thickness contours.
 Contrast of Imperfect Crystals
• A crystal containing a lattice imperfection is called as
imperfect crystal.
• Because of the lattice imperfection, the scattering centres
deviate from the regular lattice sites by a vector R(r).
• The vector, R(r) is a fraction of unit lattice vectors, and is a
function of position relative to imperfections.
• Due to irregularity of scattering centres, the equation for
perfect crystal, ijk = F.exp i(2g + k).rijk is modified as:
ijk = F.exp i(2g + k).(rijk + R)
= F. exp i (2g.rijk +k.rijk + 2g.R + k.R)
= F. exp i (k.rijk + 2g.R), since the first term is a
multiple of 2 and the fourth term is negligibly
small with respect to the second and third terms.
The total wave amplitude diffracted from a column is:
 = F. Integral of exp i(k.z + 2g.R)dz within column, where R
depends on the position of the column, and varies along the column.
 Contrast of Imperfect Crystals
• Comparison of the integral equation,
 = F. Integral of exp i(k.z + 2g.R)dz within column with that
for perfect crystal ( = F. integral of exp i k.z ) indicates that
the contrast effect due to imperfection is due to the term, 2g.R.
• As R is characteristic of the imperfection, the resulting contrast
also depends on the type of imperfection.
• Since 2g.R is a dot product of the displacement vector of
imperfection and the reciprocal lattice vector, only the
component of displacement along the reciprocal lattice vector
contributes to contrast effect.
• The reciprocal lattice vector satisfying diffraction condition is
nearly perpendicular to the incident beam, k0  Any
displacement along the incident direction can not be seen by
diffraction contrast.

• Condition for extinction of contrast is g.R = 0

• This principle can be used to determine direction of the
 Phase-Amplitude Diagram
(1) When k.z and g.R have same
sign, the displacement term
increases the phase difference
among the wavelets  curvature of
phase-amplitude diagram is
increased.  Diffraction amplitude
is reduced  No significant
contrast effect.
(2) When k.z and g.R have
opposite sign, the displacement
term reduces the curvature of the •Incident direction is taken as
original circle  diffraction is positive direction of z.
enhanced by imperfection 
k is defined as positive, when
imperfection is seen as dark
contrast in bright field. reciprocal lattice point is inside
the Ewald’s sphere.
Deviation from Bragg Condition
g = g + s;
s = k/2
s = 0, Exact Bragg Condition

s < 0, Deviation from Bragg

condition

s > 0, Deviation from Bragg

condition
 Stacking Fault
• A stacking fault is a plane imperfection, which divides a crystal
into two portions.
• Each portion of the crystal is perfect, and has exactly same
crystallographic orientation  displacement occurs only on the
stacking fault plane.
• Displacement is constant throughout the stacking fault plane, and
is confined to the plane.
• In a fcc structure, the stacking fault vector (Shockley partial
dislocation) is a/6<112> and the fault lies on {111} plane.
Stacking Fault in FCC

String of vacated B-sites along the

line O – Part of the Shockley Partial
dislocation  b = a0/6[112]
 Stacking Fault
• In a fcc structure, Shockley
partial dislocation has b =
a/6<112> and the fault lies on
{111} plane.
• A stacking fault is inclined to
the specimen plane, and
intersects the column at z = z1.

The phase factor

determining the contrast
The crystal below the fault is identical to that changes from zero above
above in spacing and orientation, but is the fault to 2g.R below
translated with respect to it by vector R. the fault.
 Contrast Due to Stacking Fault
• Since both sides of the fault are perfect crystals of the same
orientation, the portions of the phase-amplitude diagram are circles of
equal radii.
• However, a sudden phase shift occurs across the fault plane.
• b = a/6<112> implies that R = (1/6)a + (1/6)b + (1/3)c
 the phase shift is  = 2g.R = 2(h/6 + k/6 + l/3) where (hkl) is
the index for the operating reciprocal lattice vector (h, k & l must
either be all odd or all even).
• The resultant amplitude,  = F{Integral from –t/2 to z1 (exp i kz dz)
+ Integral from z1 to t2 (exp.i[kz + /3(h + k + 2l)]dz}, where the
column intersects with the stacking fault at z = z1.
• Substituting the possible combinations of h, k & l, the phase shift 
is 2n, (2/3 + 2n) or (4/3 + 2n), where n is 0 or an integer.
Phase Amplitude Diagram for Stacking Fault
Consider Integral of exp.i[kz + /3(h + k + 2l)]dz
• For hkl = 11-1; h+k+2l = 0;  = 2n,
• For hkl = 1-11; h+k+2l = 2/3;  = (2/3 + 2n)
• For hkl = 111; h+k+2l = 4/3;  = (4/3 + 2n) where
n is 0 or an integer.
• When  = 2n, the existence of stacking fault does
not affect diffraction  the stacking fault can not be
seen. Fault will be also invisible for g.R = 0. e.g. R =
1/3[111] will be invisible for g = 2 -2 0, 3 -1 1.
• For other values of , the corresponding phase-
amplitude diagram is two circles of identical diameter
rotated with respect to each other by 2/3 or 4/3 (See
next slide).
 Phase Amplitude Diagram
When a column diffracting
electrons hits a stacking fault
plane, the corresponding phase
amplitude diagram shifts from
original circle to one of the two
possible circles depending on the
magnitude of g.R.

For values of  other than 2n, the corresponding phase-amplitude

diagram is two circles of identical diameter rotated with respect to
each other by 2/3 or 4/3.
Phase Amplitude Diagram for Stacking Fault
- A column intersecting a The intersection of two circles, Q
stacking fault produces a defines scattering from stacking
diffraction contrast different fault plane.
from that of a perfect column.

-Depending on the position,

where the stacking fault
intersects the column, the
intersection of the two circles,
Q, moves along the original
circle in a different manner.
- Whenever Q coincides with
the origin O, no contrast effect
is produced by stacking fault.
Phase Amplitude Diagram
Stacking Fault - Contrast
Stacking Faults
 Contrast Due to Stacking Fault
• For the usual
specimen thickness, Q
rotates a few times
around the original
circle.
• The result is fringe
contrast, and number
of fringes indicates
number of depth
oscillations.
 Contrast Due to Anti-Phase
Domain Boundary
• An anti-phase domain boundary can be considered as similar to a stacking fault
plane with respect to super-lattice sites.
• At the domain boundary, the atomic position of the component changes from one
sub-lattice to another sub-lattice.
• The effect of phase shift due to this translation can be treated similar to that for
stacking faults.
• Consider that complete ordering has taken place in a specimen with stoichiometric
composition.
• Displacement vector, R = a/2<110> for fcc; R = a/2<111> for bcc.
• For R = (½)a + (½)b,  = 2g.R = (h + k)
• Here, the phase shift is n, where n = 0, 1, 2.
• The corresponding phase-amplitude diagram consists of two circles of same
diameter contacting each other at point Q, which corresponds to antiphase domain
boundary.
• Contrast is due to the Superlattice diffraction from Anti-phase domain boundary.
Phase Amplitude Diagram-Anti-
Phase Boundary
Anti-Phase Boundary - Image
 Dislocation

Consider a dislocation line along the y-axis. The dislocation is

parallel to the specimen plane. A screw dislocation produces one-
dimensional displacement in the direction of screw dislocation.
R = b/2 = (b/2). tan-1 (z/x)
Screw Dislocation
Screw Dislocation
 Contrast Effect Due to Screw Dislocation

• R = b/2 = (b/2). tan-1 z/x (displacement along y-direction)

• Consider a dislocation parallel to the specimen plane.

• The amplitude of diffraction from a column located at a
distance x from the dislocation line is =  = F{Integral from z1
to z2[exp i(k.z + g.b tan-1 z/x)dz
• For a perfect dislocation (such as b = a/2<110> for fcc), g.b = 0
or an integer.
• The second term, (g.b tan-1 z/x) dz distorts the phase-amplitude
diagram from the original circle.
• The magnitude of distortion depends on the absolute value of
g.b., and distance of the column from dislocation.
• For large x, the 2nd term is negligible. Thus diffraction contrast
exists only very near the dislocation line.
 Contrast Effect Due to Screw Dislocation
• Invisibility Criteria: The extinction of
diffraction contrast is given by the
condition, g.b = 0  If b lies on the
specimen plane, the dislocation contrast
disappears when a certain diffraction
operates.
• The value of b, which is perpendicular to
the operating g (two-beam condition) can
be found out  Burgers vector of the
dislocation.
Invisibility Criteria
Dislocations (Invisibility Criterion)

Dislocations A,
B and C are
visible in (a);
Dislocation A
vanishes for g =
-3 11 in (b)
Screw
dislocation with
b = ½[0 -1 1]
Invisibility Criteria (Partial
Dislocations)
 Contrast Effect Due to Screw Dislocation
• The diffraction amplitude for a screw dislocation as a function of position, x, and for various combinations
of k and g.b is given by numerical evaluation of:
  = F{Integral from z1 to z2[exp i(k.z + g.b tan-1 z/x)dz
• The apparent width of a dislocation depends on the numerical value of g.b and the value of s ( = k/2).
• Close to the Bragg position, the contrast will be a maximum and the width of the observed contrast
maxima is typically of the order of 10 nm.
 Contrast Effect Due to Screw Dislocation
• The sign of the strain-field associated with a dislocation is
reversed, when the column of diffracting crystal is moved across
the projected position of the dislocation line in the image, so that
the sign of the phase shift is also reversed.
• The above-mentioned phenomena depends on whether g.b has
the same sign as k.x.

• On one side of the dislocation, the phase-amplitude diagram is

expanded, enhancing the contrast.
• On the other side of the dislocation, the phase-amplitude
diagram collapses, reducing the observed contrast.
 n= g.b; 2sx = k.x

Unwinding side of dislocation Winding side of dislocation

The contrast from the dislocation is found only on one side of the
dislocation line, where phase-amplitude diagram is unwinding.
Thickness Contours, Dislocations, Low
Angle Boundaries
 Contrast Effect Due to Screw Dislocation

• If k changes its sign due to tilting or buckling

of the specimen, the position of contrast changes
side.

• Appreciable contrast is observed only when the

perfect diffraction condition is nearly satisfied.
Therefore, dislocation contrast is observed most
near a bend contour.
•When s (or k) and R are of the same sign, the phase amplitude
diagram is expanded and strong contrast is expected.

•The region of maximum contrast lies to one side of dislocation

line and moves to the other side, if the dislocation crosses a bend
contour.
 Contrast Effect Due to Screw Dislocation
• When a screw dislocation is inclined to the specimen plane by
angle, , the displacement within a column at a distance, x
from the dislocation is different from that due to dislocation
parallel to the specimen.
• This dislocation line is assumed to be along new y-axis (y = y
cos ) and z is perpendicular to direction, y’. Hence, the
displacement at the column is:
• R = (b/2).tan-1(z/x); z = z cos 
 R = (b/2).tan-1(z cos  /x);
The inclined dislocation has the same effect on the contrast at
the distance, x as the parallel dislocations at the distance, x/cos
.
• Contrast of the inclined dislocation is narrower by a factor,
cos 
Edge Dislocation
Edge Dislocation
 Contrast effect due to edge
dislocation
• Consider an edge dislocation lying along the y-axis, and
its Burgers vector to be along x-axis.
Diffraction by Edge Dislocation

Planes near the core of edge

dislocation bent for diffraction
 Arrays of
inclined
dislocations.
Displacements around edge dislocation
 Edge Dislocation – Displacement Field
• The displacement has two components,
• Rx = (b/2) [tan-1(z/x) + {1/2(1-)}{xz/(x2+z2)}]
• Ry = 0
• Rz = -(b/2)[{(1-2)/4(1-)}ln(x2 + z2) + {1/4(1-)}{(x2-
z2)/(x2+z2)}
• Considering the expression for Rx, an edge and a screw
dislocation have the same influence on the contrast at
different distances, xe and xs, respectively.
• For contrast to be same from both dislocations,
tan-1 (z/xs) = tan-1 (z/xe) + {1/2(1-)}{xez/(xe2+z2)}]
• In order to satisfy this condition, Abs (x e) > Abs (xs)
• Therefore, the edge dislocation appears wider than the
screw dislocation.
Contrast effect due to edge dislocation

• If the Burgers vector of an edge dislocation is along the z-

axis, the component Rz produces the contrast effect.
• Since the displacement perpendicular to the Burgers vector
is smaller than that parallel to Burgers vector, the contrast
effect is less significant.
• Because of the displacement component perpendicular to
the Burgers vector, the g.b condition does not lead to
complete extinction for edge dislocations.
 Dislocations

Tangles Cell-Walls
 Dislocation Dipoles

The sense of black and white oscillations at the ends of

dislocations a and b is opposite, indicating that the
dislocations are of opposite sign.
 Superdislocations (in Ni3Mn)

The sense of black and white oscillations at the ends of adjacent

dislocations a, b is the same, indicating that dislocations are of
same sign.
 Dislocations

Bright Field Image showing dislocations, dislocation jogs,

dipoles, grain boundary ledges and thickness contours.
 Dislocations

Dark Field Image showing dislocations, dislocation jogs,

dipoles, grain boundary ledges and thickness contours.
 Displacement due to precipitate/Inclusion
• Consider a spherical inclusion in an
isotropic matrix.
• The displacements due to the
inclusion are radial and are given by:
R = r03/r2 r  r0
= r r  r0
where r0 is the radius of inclusion,  is
the elastic strain due to the misfitting
precipitate.
• The bending of lattice planes gives
rise to contrast.
• There should be a line of zero contrast
through the precipitates normal to g,
where g.R = 0
Coherent Precipitates: Line of Zero Contrast
perpendicular to g.
Displacement Due to Precipitate
 Contrast caused by Coherency Strains

• Coherence strain distorts lattice plane  it can become a

source of diffraction contrast.
• If the displacement vector, R is known, it can be
substituted in:  = F. Integral of exp i(k.z + 2g.R)dz
• The coherency strain field around a precipitate appears
similar to that around a dislocation.
• Since the direction of displacement is reversed along a
column near the precipitate, the winding part changes to
unwinding part or vice-versa from one side of precipitate
to another.
• Resulting amplitude is exactly the same at both sides of
precipitates.
Displacement due to precipitate/inclusion

• For the inclusion, R/z changes sign for points in

the column above and below the
precipitate/inclusion.
• Contrast effects from precipitate are expected to
be relatively weak. Strong contrast is observed
only if the crystal is in strongly reflecting
condition.
• Maximum contrast is expected for a column close
to the surface of the particle.
 Contrast due to precipitate/inclusion
• The phase angle  = 2g.R = 2gr04/(r02 + z2)3/2 where r0 is the
radius of the particle, and z is measured from the point where the
column touches the particle.
• The amplitude-phase diagram for perfect crystal of the upper part
of column is a circle with radius 1/2s or (1/k).
• Considering r0 as small compared to circumference of circle,  has
a maximum value of 2gr0 for z  r0. The circle corresponding to
lower part of the column is displaced by r0 = 2gr02.
• Change in amplitude is given by the ratio of separation of two
circles to diameter: s/s = 2gr02/(s)-1 = 22gr02s
• Fractional change in intensity = I/I  42gr02s  If s is large,
contrast is more.
Phase Amplitude Diagram

Amplitude-Phase Diagram from

column AB near inclusion. The centers
of initial & final circles are displaced
by r0.
Lattice Image of Precipitate:
Rotational Moire Fringes

(100)Al//(100)Al3Ti (L12)
[100]Al//[100]Al3Ti (L12)
 Moire Patterns
• Moire patterns occur when overlap occurs between two crystals
with almost equal or multiple lattice parameters.
• If a beam diffracted by both crystals is allowed to recombine
with the transmitted beam to form final image.
• Example: A thin foil containing a plate-shaped semi-coherent
precipitate particle, such that the same reflection is excited in
both matrix and precipitate.
• The reciprocal lattice vector changes by g on passing through
the interface.
 Moire Patterns
• If a vector, rn describes a matrix atom position in absence of the precipitate, a
vector rn+R must be used to take into account the actual atom position in the
matrix because of strain in the matrix.
• Phase change =  = 2g.Rn where Rn will depend on the displacement vector rs in
the interface.
• As g.Rn is an integer and g is parallel to the interface,  = 2g.Rn = 2g.rs
  is a linear function of position in the interface, and the image contains fringes
whose direction is given by g.rs = constant.
• This is the equation of straight line perpendicular to g  the fringes run
perpendicular to g.
• In formation of Moire fringe, the transmitted and diffracted beams are
recombined in the image  The objective aperture admits the beams T and D3.
 Parallel Moire Fringes
• Both crystals have parallel reflecting
planes, but different lattice parameters.
• Using the analogue of two line
gratings, a structure of larger
periodicity is formed.
• Fringes are formed with spacing, D =
d1d2/(d1-d2), where d1 and d2 are
spacings of each grating.
• The Moire patterns with spacings 5-
500 Å are produced upon the
combination of d-spacings for the
operative reflections.
• The Moire fringes are parallel to the
reflecting planes, and perpendicular to
g as pointed out above.
 Rotational Moire Fringes

• The two gratings have

identical spacings, but a
slight angular rotation, .
• The lines in the Moire
pattern are approximately
perpendicular to the lines
of the gratings
(perpendicular to g.
• The Moire fringe
spacing,
D = d/2sin (/2)  d/
 Moire’ Fringe Images

Rotational Orientation Parallel Orientation

 Mixed Moire Fringe
• Overlapping gratings that are rotated relative to one
another, and also have different spacings d1 and d2.
• D = d1d2/{(d1-d2)2 + d1d2}1/2
Annealing Twins in Nanocrystalline Cu
Diffraction Pattern from Twins
 Diffraction Pattern from Twin
• Extra spots arise in the SADP, because the twin in the thin foil is
oriented differently from the matrix relative to incoming electron
beam.
• Twin is mirror image of the crystal structure of the matrix across a
certain crystallographic plane, Twinning plane (K1).
• The twinned material above the twin plane has been distorted by
shear, 1 relative to matrix below it.
Diffraction Patterns from Twins

Twin patterns from fcc Ordered L10 structure

crystal structure, B [110] B[201] twinned on (112)
twinned on (1 -1 1)
 Diffraction Pattern from Twin
• The twinning plane, K1 contains B: e.g. Zone axis,
B[1 -1 0] lies in the plane (111).
• The electron diffraction plane of the matrix and twin
are mirror reflections across the (111) plane (that is
180o about the [111] direction).

• The twinning plane, K1 randomly oriented relative to

B: The existence of twins is not immediately
recognizable from the symmetry of electron diffraction
pattern. A few extra spots may occur, and centred dark
field images can be used to check if spots arise from
twins.
Transmitted and Diffracted Beams

Bright Field Dark Field

BRIGHT AND DARK FIELD
Beam Tilt DF Poor man’s DF

Diffracted beam Diffracted beam inclined to

aligned with optic axis, &  is the
optic axis divergence angle: All image
points will be elongated by
3Cs(g)2
Conventional Dark Field & Weak
Beam Dark Field Imaging
Conventional DF Weak Beam DF
Weak Beam Dark Field
 Dark Field Imaging
• To take a beam tilt DF of an area oriented for two-beam
(000 & g beams), the incident beam is tilted in such a way
that the diffraction condition remains unchanged, when the
reflected beam is along optic axis.
• To ensure this, transmitted beam is moved to bring the
weakly excited –g reflection towards the centre of the
viewing screen.
• As this beam approaches the optic axis, it becomes
strongly excited and this is the reflection to use for DF
image.
• In the reverse situation, when the transmitted beam is tilted
so as to bring the strongly excited spot g towards the
centre of the screen, the original two beam condition is
disturbed, i.e., the reflection 3g becomes strongly excited.
Weak Beam Imaging
Interface Dislocations
Interface Dislocations
 Weak Beam Imaging
• Weak beam images are formed in beam centred
dark field using a faint diffraction spot.
• It is essential to avoid any bright reflection in the
diffraction pattern.
• Resolution increases with increasing deviation
from the Bragg condition.
• The maximum usable specimen thickness depends
on the amount of inelastic scattering which
destroys contrast. Amount of inealastic scattering
depends on atomic number and type of reflection.
 Weak Beam (WB) Approach
• Weak Beam technique is a convenient and important method of
producing high-resolution diffraction contrast images of crystal
defects.
• Diffraction contrast from the core of the defect only is
responsible for the image.
• Consequently, systematic analysis of a given crystal defect is
possible at a resolution approaching that of a microscope.
• Major advantage of WB approach over the two-beam
technique lies in its improved resolution combined with the
high contrast in the image.
• Since beam recombination is not required, lens aberrations are
not so important.
• The image has a very high level of contrast.
 Kikuchi Line and Experimental
Determination of k
• A “Kikuchi line” is a pair of dark and bright lines on a
diffraction pattern.
• The pair of dark and bright lines is formed by the diffraction of
electrons, which have been inelastically scattered once.
• Its formation is usually enhanced, when a thicker specimen foil
of lighter element is used.
• A fraction of the incident electrons is inelastically scattered as
they enter the specimen.
• If these electrons are not rescattered, a general background
intensity will result from these electrons. The intensity falls off
for angles further away from the incident direction.
• A pair of bright and dark lines will result when the electrons
are diffracted by a set of crystallographic planes.
If the specimen is thick
enough, it generates
electrons scattered over a
large angular range along
the forward direction.
 Kikuchi Pattern Formation
Inelastic scattering event at O.
Loss in energy of about 50 eV.
1st scattering
Ray OB reaching atom B along
plane, XY is diffracted
2nd scattering strongly along BC, causing less
intensity in direction BE.

The events of diffraction at A

and along AD are rare.
Excess Low
intensity line intensity line
 Kikuchi Pattern Formation
The cones (Kossel cones) are
constructed by considering all
the vectors oriented at angle
B (and -B) to the (hkl) plane.
There is a pair of Kossel cones
for g, and another pair for
2g.
Since the screen is flat, the
Kossel cones appear as
parabolas.
In regions close to optic axis,
the parabolas appear as
parallel lines.
 Kikuchi Pattern Formation
• Although electrons involved in Kikuchi line formation are
inelastically/incoherently scattered, the energy loss is
small (15-25 eV).

• The diffraction effect due to these electrons can be treated

as that of monochromatic electrons.

• The paths of the electrons diffracted from a set of

crystallographic planes form two cones having their axes
normal to the diffracting plane.

• The intersection of these cones with the screen produces

Kikuchi lines.
Kikuchi Pattern
Cones of high intensity (Excess)
and lower intensity (defect) are
formed.

The cones intersect the Ewald

sphere forming Kikuchi lines.

Kikuchi lines are parallel to the

projection or trace of the (hkl)
plane involved.
 Kikuchi Lines
Since Kikuchi lines are the
result of diffraction from a set
of planes, one should be able to
index the line by indexing the
diffracting plane.
Kikuchi line spacing in the
diffraction pattern is directly
related to the spacing of the
operative crystal planes by the
equation: pd = L, where p is
the spacing of the Kikuchi line
pair.
Sin  = (p/2)/L = p/2L Considering  as small, p = 2.L
There is no error arising because of deviation from the exact
Bragg condition comparable with that in spot patterns, as
Bragg law is always satisfied for the Kikuchi Lines.
 Characteristics of Kikuchi Lines
• If the specimen is tilted through a small angle, the Kikuchi
lines move whereas the intensities of diffraction spots hardly
change. Kikuchi lines are more sensitive to beam specimen
tilts than SADPs.
• When the g Kikuchi line passes through reflection G, sg = 0
(that is, Bragg condition is exactly satisfied), and the –g
Kikuchi line passes through the central spot.
• If the direct beam is exactly parallel to the plane hkl, the +g
and –g Kikuchi lines are symmetrically displaced about the
central spot, O with the g Kikuchi line passing through g/2
and –g Kikuchi line passing through –g/2.
Indexing Kikuchi Patterns
(Beam along [001])
 Kikuchi Map Construction
For fcc crystal:
• The lines are drawn for the case
where [001] pole is exactly along
the optic axis.
• These lines are perpendicular
bisectors of g vectors. The distance
between lines = g = 1/d.
• Next, keeping the common 020-g
vector pointing in the same
direction, map for the [101] pole is
constructed.
• It is noted that 020 and 0-20
Kikuchi lines are common to both
[001] and [101] pole.
 Kikuchi Map
Construction
• Next [112] pole is added.
• [112] pattern shares:
2-20 and -220 reflections/Kikuchi
lines with [001] pole; and
-1 -1 1 and 11 -1 reflections with
[101] pole.
• Next [111] and [011] poles can be
drawn
Kikuchi line pairs are not straight
lines connecting the poles.
These lines are curved, as these are
parabolic over large angles.
Kikuchi Map
 Kikuchi Pattern
(Beam along [011])
 Kikuchi Lines

Kikuchi lines are traces of corresponding hkl planes. Hence, Kikuchi

lines move as if rigidly fixed to the crystals, so that their directions and
magnitudes of movement reveal changes in orientation of the crystal.
The orientation of the specimen can be uniquely determined from
two pairs of independent Kikuchi lines. When the directions of two
indexed crystallographic planes are known, the orientation of the
crystal with respect to a reference direction can be specified.
 Kikuchi Lines: Crystal Orientation
• Using Kikuchi patterns, it is possible to estimate
orientation of the beam relative to crystal with accuracy of
1o.
• The beam direction passes along O, whereas the Kikuchi
lines represent the planes with indices (h1k1l1), (h2k2l2) and
(h3k3l3).
 Kikuchi Lines:
Crystal Orientation
• A, B & C are major poles, i.e.
zone axes.
• The zone axis, A represents
intercept of (h1k1l1) & (h2k2l2)
planes.
• The zone axis, B represents
intercept of (h2k2l2) & (h3k3l3)
planes.
• The zone axis, C represents
intercept of (h3k3l3) & (h1k1l1)
planes.
 ,  and  are semi-angles
between pairs of plane normals.
 Kikuchi Lines: Crystal Orientation
The angles between the
Kikuchi lines from different
{hkl} planes correspond to the
angles between planes,
because the Kikuchi lines are
parallel to the trace of planes.

- If distances OA, OB and OC are measured, and if the camera length

is also known, the angles between the optic axis and lines of
intersection between pairs of planes, i.e. 1, 2 and 3 can be
determined.
- It is then possible to determine [uvw] using 3 simultaneous
equations based on the dot product with plane normals, [hkl].
Deviation from Bragg Condition
g = g + s
s = 0 (or k = 0), Exact
Bragg Condition

s < 0 (or k < 0), Negative

deviation from Bragg condition -
The bright Kikuchi line shifts
towards the centre

s > 0 (or k < 0), Positive

deviation from Bragg condition –
The bright Kikuchi line shifts
away from the centre
 Deviation from Bragg Condition
• O = Origin of reciprocal lattice plane, i.e.
transmitted beam spot (000)
• L = Camera length
• x = Shift of bright Kikuchi line due to deviation
from Bragg condition with respect to
diffraction spot (G).
  = x/L = x/Rd, where d = 1/g and the
distances, x and R are measured on the
photographic plate.
• The angle  = s/g
• Taking  = , s = g = (x/L)g = x/Ld
• For small angles, R = 2BL  R/L = 2B = /d
• s = x/Ld = (x/d).(/Rd) = (x/R).(/d2) = (x/R).
(g2)
 Kikuchi Lines
• As the incident direction of electrons approaches the exact
Bragg condition, the bright Kikuchi line approaches the
corresponding diffraction spot, while the dark Kikuchi line
approaches the central spot.

• The bright Kikuchi line shifts towards the centre of the

diffraction pattern when there is a negative deviation from
the Bragg condition (k is negative).

• The bright Kikuchi line shifts away from the centre of the
diffraction pattern when there is a positive deviation from
the Bragg condition (k is positive).
 Stereographic Projections
Orientation of grains and orientation
relations can be represented using
stereographic projections.

Orientation of any plane is represented

by inclination of normal to that plane
Angle between the planes = relative to normal of reference plane.
angle between their great
Plane normals intersect the sphere at
circles. It can be measured
points called as poles.
along great circle connecting
P1 & P2.
Stereographic Projections
Stereographic Projections
Stereographic Projections
Stereographic Projections
Stereographic Projection
Foil with normal along
[001] direction.
Normals to the planes
(100), (111), (011) and
(-111) lie on the
circumference of the
same circle around the
sphere of projection.
If an image has to be
formed from the 0-22
reflection, the specimen
has to be tilted so that
0-11 pole rotates until it
is on primitive circle Tilt about a pole, that is 90o to 0-22, i.e.
(90o to beam direction). [100], [111] or [311] zone axis.
Orientation Relationship
Orientation
Relationship:
(001)bcc//(0-11) fcc
Orientation Relationship
 Dynamical Treatment: Wave Equation
A wave function for the electrons of total energy, E in a potential
field, V(r)/e must satisfy the time-independent Schrödinger wave
equation:
 2(r) + (82m/h2).[E – V(r)].(r) = 0 where = 2/x2+2/y2+2/z2
 [(-h2/82m) 2 + V(r)].(r) = E.(r)
- The first term (in 2 ) represents the kinetic energy, and the second
term represents the potential energy.
- In the TEM, the electrons are subjected to acceleration voltage and
crystal potential.
- Charge on the electron is negative and electric field associated with
acceleration voltage points towards the gun.
- Electrons have a kinetic energy due to acceleration given to the gun
 Total initial energy of the electron.
- When the electron passes through the crystal, it will have a potential
energy due to periodic potential associated with atoms in the crystal.
• The signs:
• The charge on electron, q is a negative number, -e, where e is
a +ve number.
• The acceleration voltage, -V (between 100 kV and 1 MV), is
–ve for positive charge leaving the gun. –V is the electric
field potential.
• The initial energy given to electron is a +ve number, E (in
eV). It is charge times the acceleration voltage.
• The potential inside the crystal, V(r) is a +ve number reaching
a local maximum at the positively charged nucleus of an
atom.
• The potential energy,  of the electron outside the crystal is
zero. It decreases when the electron is inside the crystal, i.e., 
= q.V(r) = -e.V(r), and therefore is always a –ve number.
 In Vacuum
 2(r) + (82m/h2).[E – V(r)].(r) = 0
• When the electron moves in vacuum (outside the crystal)
between the source and specimen, V(r) = 0 
(r) = exp (2i.r), provided the magnitude  of the wave
vector  is so chosen that h22/2m = total kinetic energy of
the electron.

.A modified version of the equation is:

2(r) + (82me/h2).[EAV + V(r)].(r) = 0
where EAV is the acceleration voltage
 Potential inside crystal
Beam electron
experiences a
small positive
attraction as it
enters the crystal.
Its kinetic energy
increases.

A. The local charge sensed by the beam electrons as it passes

through a metal, represented as a row of ion cores in a sea of
electrons. The local charge is very large and positive in the vicinity
of the ion, and approaches zero between the ions.
B. V(r) is the potential of the electrons. Potential energy is negative.
 The Crystal
• The basic property of a crystal is that it is periodic.
• V(r) = V (r + R)
where R represents any lattice vector of the crystal and r
represents any real-space vector.
• This equation is the fundamental definition of a perfect
crystal: the nature & environment at point r is identical to
that at r+R.
• The electron, which shoots through the crystal always sees
a positive potential, as the nucleus is positively charged.

• V(r) = = Vgexp (2ig.r) = (h2/2me).Ugexp (2ig.r)

This summation extends over all the reciprocal lattice
vectors, g of the crystal. Vg is defined as the g component
of V in the Fourier series, and Vg = (h2/2me).Ug
Wave Mechanical Treatment of Diffraction
Schrodinger Eqn: 2(r) + (82m/h2).[E – V(r)].(r) = 0

• The electron beam can be described by its total wave

function, tot, which must always be a solution of the
Schrödinger equation.
• A solution to the Schrödinger wave equation can be
written as:
  = aj exp i(kj.r) = ajUj,
where Uj = exp i(kj.r)
• The equation for (r) is called a “Bloch Function” when
the potential V(r) is periodic as it is in a crystal.
 Bloch Waves
• Since the electron is in a periodic potential, its wave
function must have the symmetry of the crystal.
• The solutions to the Schrodinger equation having the
required translational property are known as Bloch
waves.
• Each Bloch wave is associated with just one kj but it
is a continuously varying function of r.
• Each Bloch wave is a sum over all the points in
reciprocal space.
• Thus, each Bloch wave depends on every g,
conversely each g beam depends on every Bloch
wave.
Wave Mechanical Treatment of Diffraction

• Schrodinger Equation:
[(-h2/82m) 2 + V(r)].(r) = E.(r)
• Upon substituting Uj = exp i(kj.r) into the Schrodinger
equation for , one has:
- aj kj2 Uj + 82m/h2[E – V(r)] aj Uj = 0
• Multiplying the above equation by (–h 2/82m) gives:
 aj.(h2kj2/82m).Uj – E ajUj + V(r) ajUj = 0
• The term, h2kj2/82m is the kinetic energy of the electron
wave (or Bloch wave), whose wave vector is kj.
• Writing the kinetic energy as Ej, one obtains
 ajEjUj - E ajUj + V(r) ajUj = 0
• Or [V(r)Uj + (Ej – E)Uj].aj = 0
• Multiplying the above equation by Um* = exp (-i km.r) and
integrating over an atomic volume (),
 [Um*V(r)Ujd + (Ej – E)UjUm*d]aj = 0
• It may be noted that UjUk*d = jk
 jk is the Kronecker delta
 jk = 1 when j = k; jk = 0 when j  k
Therefore, -(Em – E)am = [Um*V(r)Ujd]aj
 [(Ej – E)Um*Ujd]aj = (Em – E)am
Considering Vmj = Um*V(r)Ujd,
-(Em – E)am = Vmjaj
 Dispersion Relations
-(Em – E)am = Vmjaj
• Since m is an arbitrary integer, the above equation
is really a set of equations:
 V0jaj + (E0 – E)a0 = 0
 V1jaj + (E1 – E)a1 = 0
 V2jaj + (E2 – E)a2 = 0
 Vmjaj + (Em – E)am = 0
• For light, dispersion means separation of
electromagnetic radiation into constituents of
different wavelengths.
• For electron optics, it is separation of
electromagnetic radiation into different k vectors
or energy.
 Zeroth Order Approximation
• The approximation is:
• Vmj = Um*V(r)Ujd = 0
which indicates that (Em – E)am = 0
• Since am is the amplitude coefficient of a wave function, am 0 
Em = E
• Thus, Em can only be total initial energy.
• Thus, the zeroth order approximation is restricted to monochromatic
waves.
• Single value of Em = E0 and the corresponding wave vector is k0.
• An allowable wave vector for a diffracted wave is: k g = k0 + 2g
• When there is only a single diffraction excited, two waves k 0 and kg
exist in the material, and the corresponding wave function can be
written as:  = a0 exp ik0.r + ag exp ikg.r
 Zeroth Order Approximation
  = a0 exp ik0.r + ag exp ikg.r, where a0, ag 0
• Since the intensity of diffracted wave must be zero at the
entrance surface of the specimen, this solution is not
compatible with the boundary condition for a normal case
of diffraction.
• The approximation is reasonable if a0 >> ag
• Thus, the kinematical theory applies better where the
atomic scattering factor involved is relatively small, i.e. in
X-ray diffraction (as compared to electron diffraction), or
where the diffraction condition is not completely satisfied
(when amplitude of diffraction is small).
 First Order Approximation
• This approximation is also called the “Two Beam Dynamical
Treatment”
• Assumptions: All Vij’s = 0 except for the cases i= 0, j = g and i
= g, j = 0; that is V0g and Vg0 are the only non-zero coefficients.
v0g and vg0 are complex conjugates of each other, and
correspond to atomic scattering factor. This assumption
provides a single atomic scattering factor.
• Two waves are related in the same manner as incident and
diffracted waves.
• The set of dispersion equations reduces to two equations:
• V0gag + (E0 – E)a0 = 0 (i)
• Vg0a0 + (Eg – E)ag = 0 (ii)
• These simultaneous equations can be solved to determine the
amplitudes, a0 and ag, only if the determinant of the
coefficient matrix vanishes.
V0g (E0 - E)
=0 (iii)
 First Order Approximation
• (E0 – E)(Eg – E) = V0g.Vg0 = Vg2 (iv)
• Where Vg2 = U0*V(r)Ugd Ug*V(r)U0d
• Vg = V0g = Vg0
 ag = - [(E0 – E)/V0g] a0 = -[Vg0/(Eg – E)]a0
• Hence, rewriting the Equation (iv), we get
• E = 1/2{(E0 + Eg)  [(E0 – Eg)2 + 4Vg2]0.5} (v)
• Since the total energy E (kinetic energy of incident
electron) is fixed, there must be two sets of E0 and Eg to
satisfy both plus and minus signs, namely, E00 Eg0 and E01,
Eg1.
• For each kinetic energy, the corresponding wave vector
exists. E00  k00; E01  k01; Eg0  kg0; Eg1  kg1
 First Order Approximation

• E = 1/2{(E00 + Eg0) - [(E00 – Eg0)2 + 4Vg2]0.5}

• E = 1/2{(E01 + Eg1) + [(E01 – Eg1)2 + 4Vg2]0.5}

• For each kinetic energy, the corresponding

wave vector exists
 E00  k00; E01  k01; Eg0  kg0; Eg1  kg1
 First Order Approximation
• Relation between incident and diffracted waves: kg0 = k00 +
2g; kg1 = k01 + 2g
• It should be noted that the magnitude of the diffracted wave
vector is not equal to that of the incident wave vector unless
the diffraction condition is not exactly obeyed. The
corresponding dispersion point is not on a Brillouin zone
boundary.
• A single value of  was assumed for all existing waves in the
kinematical treatment.
• Thus, existence of two sets of wave vectors seems strange.
• Above analysis is suggestive of multi-valued wavelength,
since four energy states appear to exist when monochromatic
electrons enter a material.
 Bloch Wave Model For Dynamical Diffraction
-Electron energy is transmitted
through the crystal as two pairs
of waves with a small
wavelength difference within
each pair.
- The two wave pairs have their
amplitude maxima at (A) and
between (B) the atomic/ionic
positions: A. Stronger
interactions; B. Weaker
interactions.
- Amplitude is transferred from
one pair of Bloch waves to the
Two types of Bloch wave in a other as electrons propagate
crystal aligned at Bragg condition. through the crystal.
 Dispersion Surface:
Pictorial representation between k and energy.
kI = k0/2; kD = kg/2 For an electron moving in vacuum, dispersion
surface is a sphere of radius, (2mE)1/2/h =1/

For large
energy or
small ,
surface of
sphere is flat.
Cross section through two spheres of radii kI and kD centered
on O and G, respectively. The spheres represent surfaces of
constant energy, and the dotted line is the trace of diffracting
plane (equivalent to Brillouin Zone Boundary).
• As the splitting of a degenerate energy state, E into two
states, E00 and E01. The energy shifts are E00 and E01,
respectively.
 E00 = E – E00 (1)
 E01 = E – E01 (2)
• E = E0 -1/2(E0-Eg)  {[1/2(E0 – Eg)]2 + Vg2}1/2
= E0 - E  [E2 + Vg2]1/2 (3)
where E = ½(E0 – Eg) = ½(E01 – Eg1) = ½(E00 – Eg0) (4)
• For the two sets of energies,
 E00 = E – E00 = -E – [E2 + Vg2]1/2 (3a)
 E01 = E – E01 = -E + [E2 + Vg2]1/2 (3b)
• The magnitude of E depends on the deviation from
the diffraction condition. At the exact diffraction
condition, E = 0 as the magnitudes of incident and
diffracted wave vectors are equal.
• The energy difference between two incident or diffracted waves,
E01 = E00 – E01 (5)
Since E00 = E – E00; E01 = E – E01
• E00 – E01 = E01 – E00 = 2[E2 + Vg2]1/2
Or E01 = 2[E2 + Vg2]1/2 (6)
• An energy state has its corresponding wave vector. Difference in
wave vectors corresponds to difference in two energy states.
• E = h2k2/82m  E = (h2k/42m).k
 k = [42m/h2k].E = (k/2E).E (7)
 k01 = (k/2E).E01 (8)
 k01 = (k/E).[E2 + Vg2]1/2 (9)
  = 0a0U00 + 1a0U01 + 0agUg0 + 1agUg1
= 0a0exp ik00.r + 1a0exp ik01.r + 0agexp ikg0.r + 1agexp ikg1.r
(10)
• The amplitude coefficients, a’s are related by two sets of
equations equivalent to:
 0ag = - [(E00 – E)/Vg].0a0 = -[Vg/(Eg0 – E)].0a0 (11a)
 1ag = - [(E01 – E)/Vg].1a0 = -[Vg/(Eg1 – E)].1a0 (11b)
• The boundary condition states that the diffracted amplitude
is zero at the entrance surface of the specimens. Incident
amplitude is at its maximum value.
 0a0 + 1a0 = 1 (Unity is full strength) (12)
 0ag + 1ag = 0 (13)
 0ag + 1ag = -[(E00 – E)/Vg]0a0 - [(E01 – E)/Vg]1a0 = 0
• (E00 – E).0a0 + (E01 – E).1a0 = 0
• Substituting of 0a0 + 1a0 = 1 yields:
• (E00 – E).0a0 + (E01 – E).(1 - 0a0) = 0
 a00 = (-E01 + E)/(E00 – E01) = E01/E01 (14)
 a01 = 1 - a00 = -E00/E01 (15)

Since ag = - [(E0 – E)/V0g] a0 = -[Vg0/(Eg – E)]a0

 a01 = - [(Eg1 – E)/Vg]ag1
 a00 = - [(Eg0 – E)/Vg]ag0
• Adding these two relations:
 a01 + a00 = 1 = - [(Eg1 – E).ag1 + (Eg0 – E).ag0]/Vg
• Vg = (ag0 + ag1)E – Eg1ag1 – Eg0ag0
= - Eg1ag1 + Eg0ag1 = (Eg0 – Eg1)ag1

ag1 = Vg/(Eg0 – Eg1) = Vg/(E00 – E01) = Vg/E01 (16)

ag0 = - ag1 = - Vg/E01 (17)
 t = 0a0exp ik00.r + 1a0exp ik01.r

= (1/E01){E01 - E00 exp ik01.r}exp ik00.r (18)

 g = 0agexp ikg0.r + 1agexp ikg1.r

= -(Vg/E01){1 - exp ik01.r}exp ikg0.r (19)

The boundary condition requires that k01 be normal to the

entrance surface. The z-axis is chosen perpendicular to the
surface.
Therefore, k01.r = k01z (20)
The intensities of the transmitted and the diffracted waves are:
It = t2 = t.t*
= [1/(E01)2]{(E01)2 - 2E00E01cosk01.r +(E00)2}
It = [1/(E01)2]{(E01)2 - 2E00E01cosk01.r +(E00)2}

Substituting Eqs. (3a), (6) and (20), one obtains

It = 1/{4(E2 + Vg2)}.{4E2 + 2Vg2 + 2Vg2 cos k01 z}
= {E2/(E2 + Vg2)} + {Vg2cos2 (k01z/2)/(E2 + Vg2)} (21)

Id = g2 = Vg2/(E01)2 {2 – 2 cos k01.r)

= [Vg2/(E2 + Vg2)] sin2(k01 z/2)
(22)
 Significance of Two Beam
Dynamical Approach Results
(1) It + Id = 1 = I0  The transmitted and the diffracted
intensities are complimentary, as the energy is conserved.
(2) E indicates a deviation from the exact diffraction
condition. When the diffraction condition is perfectly
satisfied, E becomes zero. This implies that:
It = cos2(1/2)k01z
Id = sin2(1/2)k01z

These equations indicate that the maximum depth at which one of the
waves (either transmitted or diffracted) reaches its peak intensity,
corresponds to the minimum intensity of the other wave or vice-
versa. Maximum intensities of both waves are equal & minimum
intensities are zero.
 Significance of Two Beam Dynamical Results
It = cos2(1/2)k01 z
Id = sin2(1/2)k01 z
(3) The above equations predict that the intensity of diffraction oscillates
along the depth (i.e. along a diffracting column) of a specimen even
when the exact diffraction condition is satisfied. The kinematical
theory predicts that there is no oscillation at the diffraction condition
and the corresponding phase amplitude diagram is a straight line,
implying that intensity of diffraction increases linearly along column.
The periodicity of the intensity oscillation can be found from the
above equations: (1/2) k01 z0 = (1/2) k01 g = , where g is the
extinction distance.
- The periodicity depth, g depends on the magnitude of deviation from
the diffraction condition  g = 2/k01 = h2/[2m.(E2 + Vg2)1/2]

- At the diffraction condition, E = 0  g = h2/2mVg = E/Vg

 It = cos2(1/2)k01 z; Id = sin2(1/2)k01 z
 g = 2/k01. or k01 = 2/g
• For Z = sample thickness (t),
• Id = sin2 (t/g )
Variation in intensity of
a dark field (Id) and
bright field (It) image as
a function of specimen
thickness and extinction
distance.
 Significance of Two Beam Dynamical
Approach Results
• The kinematical theory assumes that t2 >> g2

• Therefore, according to the kinematical theory, the

intensity loss from the transmitted wave due to the
diffraction is so small that there should not be any
significant contrast effect observed in bright field.

• In the dark field, there would be a contrast effect.

However, the illumination would be so small that it
would be difficult to observe the effect. That the
diffraction contrast is observed in a routine manner
can only be explained using the dynamical theory.
 Atomic Scattering Factor
• Vg corresponds to the atomic scattering factor.
• Vg = U0*V(r)Ugd = Ug*V(r)U0d
• Where U0* = exp –ik0.r; Ug = exp ikg.r
• Therefore, Vg = V(r).exp i(kg - k0).r d = V(r).exp iK.r
d, where kg - k0 = K
• When atomic scattering coefficient is small (Vg  0), It
= 1 + [Vg2cos2(k01z/2)]/E2  1
• Ig  (Vg2/E2).sin2 (k01z/2)  0
• Thus, this case approaches the kinematical case.
 Significance of Two Beam
Dynamical
ApproachResults

• Incident beam is scattered continuously.

• A part of diffracted beam is also scattered in the direction of
incident beam and vice-versa. This takes place continuously
changing amplitude in those directions.
• Assumptions in Wave-Mechanical Concept:
• Zeroth order: Efficiency of scattering from one direction to
another ~ 0  Weak scattering (Kinematical Theory).
• First order (Two-Beam Case): Incident beam is scattered
strongly to diffraction direction and vice-versa.
 Significance of Two Beam Dynamical
Approach Results
Both the wave amplitudes, 0 and g
vary with z, the depth of the crystal.
We consider the changes d0 and dg
in 0 and g, respectively, when the
electron passes through a small slab
dz in the column of the crystal.
The atoms in the slab dz scatter the There are two directions,
incoming waves 0 and g in all for which scattered
directions. In most directions, the waves reinforce each
scattered waves interfere other and give rise to a
destructively and the effect is large effect.
ignored.
 Two Beam Dynamical Approach Results
Each plane parallel to surface makes a contribution as:
(i/ g).exp(-2k.z)  g = (i/ g). exp(-2k.z)

For Perfect Crystal:

 (r) = 0(z).exp (ik0.r) + g(z).exp (ikg.r)
• d0/dz = (i/0)0 + (i/g)g exp (ik.z)
• dg/dz = (i/0)g + (i/g)0 exp (-ik.z)
  = Fexp ikz dz

For Imperfect Crystal:

Replace k.z by k.z + 2g.R
• d0/dz = (i/0)0 + (i/g)g exp i(k.z + 2g.R)
• dg/dz = (i/0)g + (i/g)0 exp –i(k.z + 2g.R)

 Two Beam Dynamical Approach Results
Rewrite the equations:
 0 = 0exp iz/0
 g = gexp (-ikz + iz/g)
The wave optical equation:
• d0/dz = (i/g).g
• dg/dz = (i/g).0 + ikg
Note that: 02 = (0 exp iz/0)2
g2 = [g exp (-ikz + iz/0)]2
 Two Beam Dynamical Approach Results
 g = (g/i)(d0/dz)
• Eliminating g,
• d20/dz2 - ik d0/dz + (/g)20 = 0
• Two independent solutions:
 0 = C0(1)exp 2i1z; C0(2)exp 2i2z
 1 and 2 are solutions of:
• (2i)2 - ik (2i) + (/g)2 = 0
 1 = {s – [s2 + (1/g)2]}/2
 2 = {s + [s2 + (1/g)2]}/2

 g = Cg(1)exp 2i1z; Cg(2)exp 2i2z

 Energy of Electron Inside Crystal
• When a wave enters a crystal, it is refracted, because of
sudden potential change. The magnitude of the wave vector
is proportional to momentum: P2/P1 = 1/2 = k2/k1 = n0 =
refractive index.
• The magnitude of wave vector of electron inside the
specimen, k0 (or ) = n0. = n0.[(2mE)/h], where n0 =
refractive index, where  is the wave vector (1/v) in
vacuum.
• Since the potentials outside and inside the crystals are
different by 10-30 V, n0  1.
• k02 = 2 = 2mE/h2 + U0
where U0 is the scaled mean inner potential of the crystal. As
U0 is a positive number, kinetic energy or velocity of
electron in crystal is greater than that in vacuum.
Due to refraction,
the wave with
wave vector k00
(or ) changes to
that with k0 on
entering a
crystal.
• Due to strong diffractions, the dispersion surface deviates
from a true spherical shape near the Brillouin zone
boundary (the diffraction condition).
• A strong diffraction occurs when the dispersion point
approaches the intersection of the dispersion surface and
Brillouin zone boundary.
• There is a strong exchange of electrons between the
incident wave and diffraction directions when the
diffraction condition is approached. The diffracted wave
serves as the primary beam and is diffracted back again to
incident direction.
Dispersion Surface for 2-Beam Dynamical Diffraction

Dispersion surface consists of two

branches labelled as (1) and (2).
Each point D on either of these
branches is associated with wave
vectors, k and k + g drawn to the points
O and G in the reciprocal space.
When k = k + g, the condition of exact
Bragg reflection is satisfied, and the
points D for which this is true lie on
the plane bisecting OG, which is the
Brillouin zone boundary.
All points D on the dispersion surface
describe Bloch waves b(k,r) of
energy, E
 Dispersion Surface for Strong Diffraction
Dispersion surface is pictorial representation between k and energy.
• Near the Brillouin zone boundary, the
dispersion surface deviates from the sphere
so that the dispersion point, L splits into L0
and L1, as indicated by the quadratic
equation (v):
E = 1/2{(E0 + Eg)  [(E0 – Eg)2 + 4Vg2]0.5}
• Because of the strong diffraction near the
Brillouin zone boundary, the refractive
index becomes anomalous and the
dispersion surface deviates from the sphere
around this region.
• The incident beam is nearly normal to the
surface. Only when the BZB is nearly
normal to surface or contains incident
direction, strong diffraction occurs.
  k01 = L1L0. Further, k01 = kVg/E
The vector, O0O1 is k01
k01 = k00 – k01
For usual TEM, k00 is nearly
normal to specimen surface.
Only when the BZB is nearly
normal to the surface or
nearly parallel to incident
beam, strong diffraction will
occur.
k01 = L1L0
For points along the BZB,
substituting E = 0 into
k01 = (k/E)[E2 + Vg2]1/2
kd01 = k.Vg/E = Vg/E = 1/g
Dispersion Surface Near Diffraction Condition
When the electron is inside the crystal, there are two values of k
& Ug 0

Near the Brillouin

zone boundary, the
two branches of the
surface constitute a
hyperbola, the
asymptotes being
spheres with centres at
O and G.
At the BZB, the
separation k = (k(2) –
k(1))/cos B
 Deviation from
Bragg Condition
by sg
Dispersion surfaces
 Scanning Transmission Electron Microscopy
• While conventional TEM uses a nearly
parallel electron beam approximately
normal to the plane of sample, it is possible
to focus the electron beam at a much larger
convergence angle into a focal spot on the
sample.
• Transmitted signal can be collected as a
function of the beam location, when it is
rastered across the sample – Mode of
operation is called STEM.
Scanning Transmission Electron Microscopy
Convergent probe is used.
Scanning coils between C2 lens
and C3 lens or upper objective
pole piece.
Scan coils pivot the beam at front
focal plane of objective lens.
The C3/upper objective pole
piece ensures that all electrons
emerging from pivot point are
parallel to optic axis.
Image of C1 lens crossover is
formed on specimen plane.
The beam must scan parallel to
Scanning Transmission Electron Microscopy

If objective lens is symmetrical,

and lower objective pole piece
is strong, then a stationary
diffraction pattern is formed in
back focal plane.
If the beam scanning is stopped,
a CBED pattern is formed in
back focal plane.
Same scan coils control
the beam scan on the
specimen and computer
display screen.
Scanning Transmission Electron Microscopy
Z-contrast: Capability of imaging
heavy atomic mass element-rich
area on low Z substrate.
Single atoms scatter incoherently
and the image intensity is sum of
individual atomic scattering
contributions.
Normal ADF detector accepts
Bragg scattered electrons.
High angle annular dark field
(HAADF) detector is used for Z-
contrast images.
HAADF detector gathers electrons
scattered at >50 mrad (3o).
Scanning Transmission Electron Microscopy

The contrast is related to

cross-section for elastic
scatter by matrix (A) and by
alloying/dopant element (B)
C = (A/B – FB)cB
cB is atomic concentration of
alloying element.
FB = fraction of alloying
element that substitutes for
matrix atoms.
Principle of Reciprocity: TEM-STEM
Convergence Angle
 Number of detectors are
employed for STEM
imaging.
The BF detector counts the
no. of electrons per unit time
as a function of the position
of the focused beam on the
sample. Small scattering
angle, 1 < 10 mrad.
By adding annular detector,
forward-scattered electrons
can be collected over a large
scattering angle, 2 > 10-50
mrad.

Electrons scattered to larger angles, 3 > 50 mrad can be detected

using a high angle annular dark field (HAADF) detector.
 Scanning Transmission Electron
Microscopy
• None of the elastically diffracted electrons reach
the HAADF detector, and the contrast is due to the
inelastically scattered electrons.
• These images are termed as Z-contrast images,
since there is correlation between local contrast
and local mass-thickness, which depends on the
value of atomic number, Z.
• Regions enriched in elements with higher value of
Z appear brighter.
• The mass concentration may even be estimated by
correlating the local concentration using a suitable
expression for inelastic scattering.
 Convergent Beam Electron
Diffraction (CBED)
• CBED is the oldest TEM diffraction technique.
• It was developed by Kossel and Möllenstedt in 1939.
• SAD has the following disadvantages:
• (1) Smallest area selected by SAD is 0.5 m (with
intermediate voltage and very low Cs, the smallest area of
SAD is 0.1 m).
• Many crystal defects and precipitates are smaller in size.
• The limitation is overcome by using CBED.
• This region is is a function of beam size and beam-specimen
interaction volume, which increases with specimen thickness,
but a lot smaller than SAD.
• Several microdiffraction techniques have been developed to
overcome the spatial resolution limitations of SAD in TEM.
Selection of Alpha ()
Nano-Probe
 Convergent Beam Electron Diffraction
The big advantage of
CBED over all other
diffraction techniques is
that most of the
information is generated
from miniscule regions
beyond the reach of other
diffraction methods.

SAD Pattern: (i) Electron

CBED Pattern:
Beam incident on the sample
(i) E-Beam is convergent; is parallel;
(ii) beam is 10-100 nm in diameter (ii) beam is 1 – 10 mm in
and  = 1-10 mrads. diameter
Diffraction Patterns of Si [111]

CBED Pattern SAD Pattern

CBED Pattern and Ewald Sphere

As beam converges over an angular range of 2, the Ewald

sphere is effectively rotated 2 about the origin.
The Ewald sphere has an effective thickness of 2 because of
beam convergence, and so intercepts a range of HOLZ reciprocal
lattice points.
CBED Patterns
Convergent Beam Electron Diffraction

A strong upper
objective pole piece is
required to create a
convergent beam, so
that diffraction
patterns are created.
Before development
of STEM, CBED was
possible by addition of
a mini-lens below C2
lens of conventional
TEM.
 STEM Microdiffraction and
Annular Bright and Dark Field
STEM Bright Field & Dark
Field
 Obtaining CBED Patterns by TEM
CBED patterns can be formed in any TEM that is capable
of creating a small (<< 1 m) beam with a convergence
semi-angle () > 10 mrads.

Four microscope variables to be controlled for

obtaining the CBED pattern are:
• The beam convergence semi-angle, .
• The camera Length (L)
• The focus of the pattern
• The size of the beam.
 Obtaining CBED Patterns In TEM
Mode
• Start with your specimen in the eucentric plane, and
form a focused image on the TEM screen with the area
to be examined.
• Select a large C2 aperture about 100-200 m in
diameter, and centre it. Adjust the C2 lens to form a
focused spot on the area of interest.
• Keep C1 weakly excited to give a relatively large spot,
about 100-200 nm containing sufficient current to give
high intensity pattern.
• Select a small camera length < 500 mm to give a wide-
angle view of the pattern.
• To observe CBED pattern, switch to diffraction mode,
making sure that both objective and SAD apertures are
retracted.
Choice of C2 Aperture for CBED Pattern (TEM Mode)

The convergence semi-

angle,  is set by
changing the C2 aperture.

The pattern of non-

overlapping discs is a
Kossel-Möllenstedt
(K-M) pattern. To get
this pattern, C2
aperture should be
selected such that 2
< 2B for a particular
specimen and
Rings of intensity represent electrons orientation  C2
scattered at high angles ~ ±10o aperture of 10-50 m.
Choice of Camera Length (TEM mode)

• The choice of camera length (L) depends on the

nature of information required.
• A large L gives a high magnification view, but
spans only a small angular range.
• L > 1500-6000 mm to observe detail in (000) BF
disc.
• L < 500 mm to observe low magnification or
whole pattern containing electrons scattered at
high angles.
CBED Patterns for Different Camera Lengths

A. At high L, CBED pattern shows the details of (000) spot.

B. At intermediate L, CBED pattern shows details of all
diffraction discs.
C. At lowest L, higher order Laue zone is faintly visible.
 Focusing of Pattern (TEM)
• The beam has to be focused on specimen,
rather than underfocused or overfocused by
adjusting strength of either objective or C2
lens.
• The CBED pattern has to be focused on
back focal plane of the objective lens.
• Use intermediate lens fine focus to sharpen
the image of C2 aperture.
 Effect of Focusing

Underfocussed
Focussed

Overfocussed
 Choice of Beam Size (TEM)
• The beam size should be reasonably large to
provide reasonably good intensity.
• The volume sampled by beam controls
spatial resolution.
• For thin specimens, spatial resolution is close
to beam size.
• In thicker specimen, inelastic scattering will
spread the beam and degrade spatial
resolution.
 Effect of Specimen Thickness
• For very thin specimens, kinematical
diffraction conditions are prevalent, i.e.
diffraction discs are uniformly bright.
• Moving to thicker region provides
dynamical contrast in diffraction patterns.
• For CBED, specimens should be thicker
than one extinction distance.
 CBED Patterns for Different
Thicknesses

For Thinner Specimen For Thicker Specimen

 Energy Filtering
Inelastically scattered electrons hide useful information.

No energy filtering With energy filtering

 CBED Patterns: Kossel Pattern from
Si with beam along <001> zone

CBED pattern at exact focus Overfocussed condition: focus in

a plane before back focal plane
of objective lens.
Large Angle CBED
Specimen raised by h
from the eucentric (object)
plane in objective lens.
Specific electrons in the
incident cone along CE-DE
and AE-BE are at exact
Bragg condition from both
sides of the hkl plane.
A spot pattern is created in
eucentric plane.
LACBED pattern of lines
is formed in back-focal
plane.
All electrons diverge
through the spots and are re-
focussed by the objective
lens to form spots K, L, E.
In back-focal plane of
objective lens, there is
LACBED pattern consisting
of a disk of intensity crossed
by deficient lines like
Kikuchi lines.
But these lines arise from
elastic scattering, and are
called Bragg lines.
The deficient lines come
from lines in specimen along
which strong diffraction
occur from hkl plane.
 Use of SAD for LACBED
• If SAD diaphragm is inserted into image
plane so as to permit only direct beam to
pass through the aperture – BF LACBED
pattern is obtained. This contains hkl
deficient lines of intensity
• If SAD aperture is kept around ghkl
reflection in the image plane, such a pattern
consists of hkl excess lines of intensity.
• Contrast can be improved by using a
smaller SAD aperture.
 Large Angle CBED

Without SAD
aperture

With SAD
aperture
Montage of Si[111] BF LACBED Pattern
surrounded by six {2-20} DF LACBED

Symmetry information can

be obtained to find out
point group of the crystal
structure.
 CBED In STEM Mode
• First the beam scanning is stopped.
• The spot on the STEM screen is positioned
on the region of interest.
• In TEM, the area of specimen can not be
seen without spreading the convergent
beam.
• In STEM, the convergent beam is scanned
to see the image.
 ZOLZ Patterns
• On increasing L above ~800 mm, only first
few diffraction maxima are observed.
• CBED pattern consists of discs similar to
array of spots in SAD pattern, i.e., discrete
diffraction maxima surrounding the central
(000) disc.
• This is ZOLZ pattern as hu + kv + lw = 0.
 HOLZ Patterns
• The central bright portion of CBED pattern is due to relatively
intense low angle scattering.
• At higher angles, the ZOLZ intensity drops because the atomic
scattering amplitude f() is decreased.
• Intensity increases when the Ewald sphere intercepts the
HOLZ planes in the reciprocal lattice, and a ring of diffraction
intensity is observed.
• The HOLZ intensity arises from weak high angle diffraction
from crystal planes that are not parallel to the beam.
• Radial distance from (000) in a diffraction pattern is related to
angle of scatter. Use a smaller camera length (L) to see spots
formed by higher angle scattering.
 HOLZ Patterns
• Consider the intersection of Ewald sphere
with reciprocal lattice.
• The HOLZ planes in the reciprocal lattice
cross the sphere, (unlike the zero-order
layer, which is tangential to it), and
intersection of a sphere with a plane creates
a ring.
Laue Zones
hu + kv +lw = N,
where N is an integer

The first ring is called the FOLZ,

because hu + kv + lw = 1
CBED Pattern and Ewald Sphere

As beam converges over an angular range of 2, the Ewald

sphere is effectively rotated 2 about the origin.
The Ewald sphere has an effective thickness of 2 because of
beam convergence, and so intercepts a range of HOLZ reciprocal
lattice points.
 CBED Patterns showing HOLZ

There is 3D crystallographic information in the CBED pattern,

whenever significant HOLZ diffraction intensity is present.
The HOLZ-ring radius depends on: (i) interplanar spacing in the
crystal, (ii) the electron wavelength (or acceleration voltage), (iii)
camera length (L) and (iv) any misalignment of beam with optic axis.
 CBED Patterns showing HOLZ
• There is 3D crystallographic information in the CBED pattern,
whenever significant HOLZ diffraction intensity is present.
• To observe HOLZ rings in addition to the ZOLZ pattern, very
small L (<500 mm) is chosen, so that full angular range of the
back focal plane is permitted by imaging system.
• Since the scattering into HOLZ spots is weak, the exposure
time to reveal HOLZ maxima is unusually long so as to ensure
that ZOLZ is overexposed. Two exposures are required to
record both ZOLZ and HOLZ patterns.
• One should tilt to a low-symmetry zone axis (e.g. <114>), it is
easier to observe a FOLZ than a high symmetry zone axis
(<001>).
 Kikuchi Lines In CBED Patterns
Kikuchi lines in Kikuchi lines in
SAD pattern CBED pattern

• In CBED patterns, one sees sharp Kikuchi lines, while in SAD

patterns, the Kikuchi lines are diffuse.
• The difference arises because the convergent beam samples a much
smaller region of specimen than that selected by the SAD aperture.
• In the volume of specimen contributing to the CBED pattern, there
is usually little or no strain, either elastic (due to specimen bending)
or plastic (due to lattice defects).
• In zone axis patterns, ZOLZ Kikuchi lines appear as bright bands.
 Kikuchi Lines In CBED Pattern

• Generation of Kikuchi lines in CBED is more complex

than that in SAD pattern.
• The incident electrons span an angular range and
therefore some electrons in the beam may already be at
the Bragg angle to a ZOLZ plane.
• Hence, there will be an elastic scattering contribution to
the Kikuchi lines, where they cross the disks in CBED
patterns.
 Kikuchi Lines In CBED Pattern
• If Kossel conditions (i.e. 2 > 2B) are chosen, there will
always be electrons in the beam with the correct trajectory
for exact Bragg diffraction from the planes in the UVW
zone, and so there is elastic contribution to Kikuchi lines.
 Relationship Between HOLZ
Lines and Kikuchi Lines
• Kikuchi lines arise from inelastic scattering by HOLZ planes
and HOLZ Kikuchi lines exist in CBED patterns.
• The HOLZ Kikuchi lines are in principle more useful than
ZOLZ Kikuchi lines, because those come from planes with
much larger Bragg angles (and g-vectors). Hence, those lines
are more sensitive to changes in lattice parameter than the
ZOLZ lines.
• Since, g = 1/d; g = -d/d2
• Thus, for smaller d, the value of g is larger at the same d.
• Advantage is taken by using the HOLZ lines, which are the
elastic part of HOLZ Kikuchi lines, which lies within the
diffraction discs.
• Lines arise when the incident beam at the correct Bragg angle
for diffraction by a HOLZ plane are scattered to high angles,
creating a bright line through the HOLZ disc and leaving a
dark line in (000) disc.
 ZOLZ & HOLZ Kikuchi Lines in
CBED Patterns
The HOLZ lines come in pairs,
like Kikuchi lines with bright
(excess) lines within HOLZ discs
and dark (deficient) lines within
the (000) disc.
As HOLZ lines contain 3D
information, they show the true
fcc threefold {111} symmetry.
However, the ZOLZ Kikuchi
lines and spots [three principal
pairs of (2 -2 0) lines show six- Schematic illustration of
fold two dimensional {111} {111} CBED pattern.
symmetry.
A C

B C. Deficient HOLZ lines are

visible in the central (000)
disc of a low symmetry (114)
zone axis parttern.

A. Low L and large  CBED pattern showing a wide area of

reciprocal space, away from zone axis pattern conditions.
B. When the specimen is tilted to low index zone axis pattern, a
small C2 aperture is inserted, a ring of excess HOLZ lines
appear associated with HOLZ discs.
 Indexing HOLZ Lines
Record another pattern with a
small L and large  to show
HOLZ discs.
Index HOLZ hkl maxima and
observe which maxima show the
clearest HOLZ excess lines.
Each HOLZ line pair will be
perpendicular to the g-vector
from 000 to HOLZ maximum.
Information from CBED Patterns
• Specimen Thickness
• Unit Cell and Precise Lattice Parameters
• Crystal system and true 3D crystal symmetry
• Primary Advantage: Information generated from
small regions beyond the reach of other diffraction
methods.
• Problem: Localized contamination and beam
damage or heating
 Thickness Determination
In zone axis patterns recorded under
2s < 2B, concentric diffuse
fringes (Kossel-Möllensedt fringes)
are visible.
On moving the specimen (if it is not
bent) under the beam, the number of
fringes changes.
Number of fringes increases by one
every time, the thickness increases The K-M fringes
by extinction distance, g contain information on
thickness.
If specimen thickness is less than g,
no fringes are visible.
 Thickness Determination
We tilt to two-beam condition with only one strongly excited hkl
reflection: CBED discs show parallel intensity oscillations.

As in elastically bent crystal, the convergent beam provides a range

of incident angles to the diffracting hkl planes.
Oscillations are symmetric in hkl disc and asymmetric in 000 disc.
 Thickness Determination

The central bright fringe in hkl disc is at exact Bragg condition, s=

0
The fringe spacings correspond to angles i.which are related to
deviation si for the ith fringe.
si =  (i/2Bd2)
where B is the Bragg angle for the diffracting hkl plane, d is the hkl
 Thickness Determination
If the extinction distance, g is
known, the foil thickness t is
determined from the following
equation:
si2/nk2 + 1/g2nk2 = 1/t2
Assign n = 1 to the first fringe,
which corresponds to excitation
error s1.
Assign n = 2 to second fringe, s2.
Plot (si/nk)2 vs 1/nk2. If a straight
line is obtained, arbitrary
The intercept is t2 and the
assignment is considered good.
slope is -1/g2 If a straight line is not obtained,
re-assign n=2 to the first fringe.
CBED Patterns from a Carbide Particle

Kossel pattern recorded with C2 Same pattern taken with C2

aperture having size of 150 m aperture having size of 20 m.
- Consider a specimen with
known crystal structure.

- The projected radius (G) of

the HOLZ ring provides the
lattice-repeat vector of the
crystal parallel to the beam
direction.

- The spacing of the

reciprocal lattice parallel to
the electron beam = H

- The value of G can be

measured more easily from
the Kossel pattern.
 Interplanar spacings parallel to
the electron beam
Since the radius of the Ewald
sphere increases with
decreasing , the value of G
for a given orientation
increases as the acceleration
voltage is raised.
HOLZ rings are difficult to
see on increasing the
acceleration voltage.

Assuming that the terms with H2 are negligible,

- FOLZ ring radius, G1 = (2H/)1/2
- SOLZ ring radius, G2 = 2(H/)1/2
Both G and H are in reciprocal lattice units (nm-1)
Interplanar spacings parallel to the electron beam
• 1/H = 2/G12
• The value of H-1 can be expressed in real space units
(nm) through the measured radius, r (mm) and the
camera constant, L (nm.mm):
• 1/H = (2/).(L/r)2  by measuring the values of r
(= radius of HOLZ rings), one can determine the
value of H.
• Low symmetry zone axis with a small H will give
rise to a small HOLZ ring of diameter, r on the
diffraction pattern, which will be easier to observe at
a chosen L.
• Hence, the crystal repeat spacing parallel to the
electron beam can be determined, if the Miller
indices for the direction of H are known.
• 1/H is directly related to the real space direction
vector as:1/H = [u v w]
 Interplanar spacings parallel to the
electron beam

For fcc crystal, 1/H = a0(u2 + v2 + w2)1/2/p

where a0 is the lattice parameter,
p = 1 when (u + v + w) is odd, and p = 2 when (u + v
+ w) is even.
For bcc crystal, p = 2, if u, v and w are all odd, and p
= 1, if u, v and w are all even.
These conditions for p take into account structure
factor effects, which causes systematic absence of
reflections or whole rings in some cases.
 Interplanar spacings parallel to the
electron beam
If there are no absences of the HOLZ rings:
• In a crystal system with orthogonal axes
(cubic, tetragonal or orthorhombic),
• 1/H = (a2u2 + b2v2 + c2w2)1/2; where a, b and c
are the lattice constants. For cubic structure, a
=b=c
• For hexagonal or rhombohedral systems,
For three index system
• 1/H = [a2(u2 + v2 – uv) + c2w2]0.5
 Lattice Parameter Determination
• It is possible to get reasonably accurate
measurement of the lattice constants by indexing
reflections in ZOLZ and/or by measurement of the
HOLZ ring diameter.
• Best method is to use the positions of HOLZ lines
in the (000) disc. As these lines arise from high
angle reflections, their positions are sensitive to
changes in lattice parameter.
• Results with more than 0.2% precision can be
obtained by computer simulation of the position of
HOLZ lines using different lattice constants..
Lattice Parameter Determination

HOLZ lines in the

(000) disc in [111]
CBED pattern.

Results with precision

of ~ 0.2% can be
generated by computer
simulation.

Changes in the pattern

symmetry within the
encircled region is
easily distinguished.
 Lattice
Parameter
Determination
For determination of precise
lattice parameters, the
experimental HOLZ line
pattern is compared with
computer-simulated version.

Changes in lattice
parameter can occur
due to composition or
strain.
 Point Groups
• Crystallographic point groups are the sets of
crystallographically permissible
symmetries, which are formed when sets of
axes intersect in a common point.
• The axes correspond to rotation, inversion
and mirror symmetry elements.
Point Group Symmetry
A. Symmetry 2 refers to a Symmetry in CBED Patterns
two-fold (diad) rotation
axis, i.e. pattern has
symmetry when rotated
by 180o.
B. 2mm is a diad
symmetry with two
independent mirror
planes parallel to diad.
C. 3m indicates three fold
rotation (triad)
symmetry with one
mirror plane, i.e.
rotation symmetry
every 120o with one D. 4mm indicates a fourfold rotation
mirror plane present at symmetry (tetrad) with two independent
each 120o mirror planes parallel to the tetrad.
 Obtaining 3D Crystal Symmetry Information
Whole Pattern Symmetry
• Use large C2 aperture and
small camera length to get
Kossel pattern, such that
HOLZ rings are continuous.
• HOLZ line symmetry contains
the true symmetry of the
crystal.
ZOLZ pattern (inset) is a square array
corresponding to [001] – could be cubic,
tetragonal or distorted othorhombic
structure with a  b
Measurement of FOLZ ring
Kossel pattern has symmetry of 2mm
diameter confirms that c-axis
with the mirrors M1 and M2 indicated,
repeat distance is 0.75 nm,
which is suggestive of orthorhombic which is larger than a or b
crystal. measured from ZOLZ.
 Whole Pattern Symmetry
CBED Pattern from Cu:
The ZOLZ Kikuchi line
pairs show six-fold
symmetry.
Ring of excess HOLZ
line can be seen.
Deficient HOLZ lines in
the (000) disc show only
threefold symmetry.
 Obtaining 3D Crystal Symmetry Information
Bright Field Symmetry: HOLZ
lines present in (000) disc.

ZOLZ Kikuchi pattern from

<111> fcc shows six-fold
symmetry in the form of a star.

HOLZ lines in a wide angle CBED

pattern show three-fold symmetry
Phase Contrast Imaging (HRTEM)
The most significant difference with
the conventional BF or DF imaging
is the number of beams collected by
the objective aperture or detector.
BF or DF image requires the
collection of a single beam
(transmitted or diffracted) for image
formation.
Phase contrast image requires the
selection of more than one beam.
More beams collected improve the
resolution of the image.
Contrast in TEM images can occur due to
the differences in phases of the electron
waves scattered through a thin specimen.
Multi-Beam Images
 Phase Contrast Images
• The appearance of the image varies with small
changes in: (i) thickness, (ii) orientation, (iii)
scattering factor of the specimen, (iv) focusing
conditions, and (v) astigmatism of the objective
lens.
• This sensitivity is the reason why phase contrast
can be exploited to image the atomic structure of
thin specimens.
• The requirements are: (i) TEM with sufficient
resolution is required to detect contrast variations
of atomic dimensions; (ii) Proper control of
instrument parameters that affect the phases of the
electrons passing through the specimen and lenses.
 Illumination System
• The electrons start from electron source or gun.
• Depending on the actual details of source, electrons are emitted with
a spread of 1-2 eV.
• These electrons are accelerated by acceleration voltage of electrons,
say 200 kV. In reality, the energy imparted due to this is 200,
where  is due to the ripple in high voltage source.
• As a rule, the electrons from the source are emitted incoherently
with a range of directions.
• These electrons pass through a number of apertures, which limit the
angular range of the electrons reaching the specimen; e.g. condenser
lens cuts off some of the directions.
 Effect of Illumination System
• The range of angles used to form the image is considered as the
convergence of illumination.
• The range of angles (convergence) is characterized by either a
Gaussian distribution of directions (when the illumination is
defocused) or a cone of directions characterized by half angle of the
cone (when illumination is fully focused).
• Because of chromatic aberration of the condenser lens, the electrons
having different energies due to ripple in accelerating voltage and
natural energy spread of accelerating electrons are focused at different
positions.
• Condenser aperture limits the net energy spread of the electrons.
• The two main parameters carried forward from the illumination are (i)
energy spread of electrons, and (ii) convergence of illumination.
 Effect of Specimen
• A wave with a unique wave vector before
the specimen is split into a wave with a
number of different wave vectors by the
specimen diffraction (elastic scattering).
• These different wave vectors are all
coherent.
• The specimen is not truly stationary, but it
drifts or vibrates.
 Post-specimen Imaging
• It is conventional to consider all remaining lenses as just one lens,
i.e. the objective lens.
• With this lens, the wave exiting the specimen is focused on to the
phosphor screen/photographic film, or CCD camera.
• Image may be kept slightly out of focus to compensate for the
spherical aberration of the objective lens, which brings waves
traveling in different directions into focus at different positions.
• An ideal objective lens focuses the specimen exactly, while in
reality there are some instabilities in the current of the objective
lens. This leads to distribution of the actual focus of the final
image, and is termed as the focal spread.
• The energy spread of the electron source has the same effect as the
focal spread at the objective lens. Together, these two parameters
are termed as the focal spread of the microscope.
 Additional Effects
• Astigmatism of objective lens.
• Orientation of incident illumination with
respect to optic axis of the objective lens
(post-specimen lenses), is called the tilt of
illumination or “Beam Tilt.”
• The last effect is the objective aperture.
• In modern HRTEM, very large objective
aperture is used.
 Abbe Interpretation of Imaging

An object with periodic structure produces two diffracted beams in

directions given by the Bragg’s law, together with the unscattered
beam. These are focused to points in the back-focal plane. The image
is formed from interference between spherical waves from each of
the three sources in back focal plane.
 Wave Optics
• A wave in real space of the form (r) can be decomposed into components
(k), where k is the wave-vector of the wave, by the following Fourier
integral:
 (r) = (k).exp (2ik.r).dk
• The inverse of the above-mentioned Fourier transform is:
 (k) = (r).exp (-2ik.r).dr
• The wave function (k) has a spatial frequency of k.
• For electron diffraction, we refer to both the position vector (r) in real space
and the wave vector (k) in three dimensions.
• For imaging theory, this can be simplified to two dimensions normal to the
incident beam direction: the vector, u (not k) can be used to describe the
spatial frequency in two dimensions.
 Phase Contrast Images – Lattice Fringes
• Because of aberrations in the microscope, the spatial frequency,
u is changed as: (u).exp (2iu.r)  (u).exp (2iu.r - i)
• Two beams are allowed to interfere by selection of objective
aperture.
 (r) = A.exp 2i(u.r) + B.exp 2i(u.r-)
where A and B are real numbers.
The intensity is then:
I = ABS [(r)]2 = A2 + B2 + 2A.B. cos (2[u-u].r + )
• If  has a fixed value, term on the right of the equation
represents fringes in the image with spatial frequencies of u-u
caused by the interference of two waves.
• If  has a distribution of values, on averaging over different
values of the cosine term, this term becomes zero  No
interference between the waves.
 Coherent and Incoherent Aberrations
• If the phase change,  is fixed for each value of u, we refer
to the aberration as coherent.
• If the phase change,  is not fixed for each value of u and
has a random distribution, we refer to the aberration as
incoherent.
• Coherent waves interfere, and coherent aberrations affect
the interference between the waves.
• Incoherent waves do not interfere, and incoherent
aberrations can be considered by summing of the
intensities.
• Coherent Aberrations: Defocus, astigmatism, tilt and
spherical aberration.
• Incoherent Aberrations: Focal spread, Drift, Convergence
 Factors affecting Resolution
Coherent Aberrations
• Defocus (To compensate for spherical aberration and other
specimen induced effects)
• Astigmatism
• Beam Tilt (Orientation of beam with respect to optic axis)
• Spherical Aberration
Incoherent Aberrations (Random character)
• Convergence (Angular spread of electrons)
• Focal Spread (Energy spread of electron source and Instability
in Objective lens current)
• Drift
 Coherent Aberrations
• All coherent aberrations can be expressed rigorously by
expanding the phase shift,  as Taylor series:
 (u) = A + B.u + C.u2 + (D.u)2 + ….
• Let us assume that the electron wave is passing exactly through
the centre of the lens along the optic axis. Then, by symmetry,
all the odd terms must vanish. We are left with the series:
 (u) = A + Cu2 + (Du)2 + Eu4 + …
• By comparison with Fresnel propagator, C = z
where z is the objective lens defocus.
• The astigmatism of the lens is the term (D.u)2, which is written
in the form:
• (Du)2 = ..(.u)2
where  is the astigmatism in Angstroms and  is the unit
vector in the direction of the astigmatism.
• E = (/2).3.Cs
 The Role of Optical System
• The microscope transforms
each point on the specimen
into an extended region in the
film image.
• Since each point on the
specimen is different, the
specimen is described by
function, f(x,y).
• The extended region in the
image, which corresponds to
point (x,y) in the specimen, is
described as g(x,y).
• Both f and g are functions of
x and y.
If two nearby points, A and B are
considered, two overlapping images, gA
The Role of
and gB are formed. Optical System
Each point in the image has contributions
from many points in the specimen.
This result is mathematically expressed
as:
g(r) = f(r)h(r-r)dr = f(r)  h(r-r)
where h(r-r) is a weighting term telling us
how much each point in the specimen
contributes to each point in the image.
Since h(r) describes how a point spreads
into a disk, it is known as the point-spread
function or smearing function.
g(r) is convolution of f(r) with h(r), i.e. f
 The radio analogy
• In audio system, it is necessary to hear different frequencies, such
as drum, other instruments and music.
• High frequency in audio is related to 1/t, while frequencies in
lattice images are related to 1/x.
• High spatial frequencies correspond to small distances in real
lattice, but large distances from optic axis in diffraction pattern.
• The rays passing through lens at large distances are bent through
larger angles by objective lens. These rays are not focused at the
same point by objective lens, because of spherical aberration.
• Spreading of point in the image happens. Discs are observed.
Magnification is increased, but confusion is observed for fine
details.
• One can write g(x,y) as g(r), considering the discs to be of
uniform intensity.
• g(x,y) = g(r) = G(u).exp (2iu.r)
• Here, u is a reciprocal lattice vector, the spatial frequency for a
particular direction.
• G(u) is the Fourier transform of g(r).
• G(u) is defined as product of two other Fourier transforms,
i.e. G(u) = H(u).F(u)
• h(r) shows how information in real space is transferred
from specimen to image.
• H(u) is the Fourier transform of h(r).
• H(u) tells how information or contrast in reciprocal space
is transferred to the image. It is called the contrast transfer
function.
• F(u) is the Fourier transform of f(r).
• So convolution in real space [g(r) = f(r)h(r-r)dr = f(r)
 h(r-r)] gives the multiplication in reciprocal space (as
shown by equation: G(u) = H(u).F(u).
• The factors contributing to H(u) include:
– Aperture  A(u)
– Attenuation of the wave  E(u)
– Aberration of lens  The aberration function, B(u)
• H(u) = A(u).B(u).E(u).
• The aperture function, A(u) indicates that the objective
aperture cuts off all values of u (spatial frequencies)
higher than some selected value governed by the
radius of aperture.
• The envelope function, E(u) has the same effect, but is
a property of the lens itself.
• B(u) is expressed as B(u) = exp (i(u)).
 (u) = zu2 + (1/2)Cs3u4
where z is the value of defocus.
 Derivation of (u)
• Due to the combined effect of spherical aberration and defocus of
the objective lens, a point at the specimen appears as a disc with
radius () = Cs3 + z.
• The rays which pass through objective lens at angle  are not
focused to the same point due to spherical aberration and finite
defocus, z.
• For a range of angles (values of ), averaging can be done as: D()
= ()d (with limits = 0 and ) = Cs4/4 + z2/2
• Bragg’s law states that: n = 2d sin B
• For small B (Sin B  B) 2B  g = scattering angle (as g = 1/d).
• Hence, in the equation for D(u),  is replaced by u, where u is the
reciprocal lattice vector (u = g).
 (u) = Phase = (2/).D(u) = (2/).[Cs4u4/4 + z2u2/2]
• Thus, we have (u) = zu2 + (1/2)Cs 3u4
• Thus, Sin (u) depends on the values of Cs (the lens quality), 
(acceleration voltage), z (defocus value) and u (spatial frequency)
 Coherent Aberrations
• Phase shift for the HRTEM in the absence of beam tilt is =
(u)
 (u) = /(z2u2 + 1/2Cs4u4)
• In the presence of small beam tilt, the origin of the phase
shift term is shifted from u = 0 to u = w.
 (u) = /[z2(u-w)2 + 1/2Cs4(u-w)4]
• The phase shift term, (u) is central to understanding the
imaging process.
• 1/2Cs4(u-w)4 = ….. + 2Csu2w2 + 2Cs4(u-w)2
• The first of these terms is equivalent to a defocus, whereas
the second is similar to astigmatism. Hence, the tilt can be
partially cancelled out by astigmatism.
 Transfer Theory
• For a given value of the electron energy, defocus and the wave
leaving the specimen is modified by the phase term (u), which
depends upon the vector u, the spatial frequency of each
Fourier component of this exit wave.
• This modified wave forms an image after the microscope
lenses.
• If (r) is the wave exiting the specimen, the imaging process is:
 (r) --(Fourier Transform) (u) –(Phase Change) (u) ---
(Fourier Transform) (r) –(Intensity) [(r)]2 ---
(Incoherent average) I(r), Final Image.
• The incoherent effects are included by averaging images at
different energies by means of changes in the lens focus to
obtain the final image.
• Most of these involve approximations about the form of the
wave leaving the specimen.
 The Specimen
• The specimen function, f(r) is called the specimen
transmission function.
• f(r) = f(x,y) = A(x,y).exp (-i(x,y))
where A(x,y) is the amplitude (not aperture function) and
(x,y) is the phase which depends on thickness of the
specimen.
• For HRTEM applications, A(x,y) = 1.
• Phase change depends only on the potential, V(x,y,z),
which the electron sees while passing through the
specimen.
• Let us assume a thin specimen with thickness, t, such that
its two-dimensional projected potential is = Vt(x,y)
• Vt(x,y) = V(x,y,z)dz from z = 0 (lower limit) to z = t
(upper limit)
• It is an attempt to create a two-dimensional projection of
the crystal structure.
 The Specimen
  = h/(2meE), where E is the acceleration voltage in Volts.
• When the electrons are in a crystal,  is changed to , such
that  = h/[2me(E + V(x,y,z)]0.5
• While passing through a material of thickness, dz, the electrons
experience a phase change given as:
• d = 2(dz/) - 2(dz/)
 d =2(dz/)({[E+V(x,y,z)]0.5/E0.5}-1)
 d = 2(dz/)({1 + [V(x,y,z)/E]}0.5 – 1)
 d = 2(dz/)(1/2){V(x,y,z)/E}
 d  (/E).V(x,y,z)dz
 d  V(x,y,z)dz, where  = /E.  is the interaction
constant, and is an expression related to elastic interaction. It
remains constant, as V increases, because  and E compensate
each other.
 The total phase shift is dependent only on V(x,y,z), and is =
V(x,y,z) dz (integrated from z = 0 to z = t) = .Vt(x,y)
 The Specimen
• Taking into account absorption,
• f(x,y) = exp [-iVt(x,y) - (x,y)]
• This model shows that the specimen is a “phase object”, if
absorption is neglected.
• In other words, it is “Phase Object Approximation (POA)”.
• For very thin specimens (which is necessary for HRTEM),
Vt(x,y) <<1, and (x,y)  0.
• Thus, expanding the exponential function and neglecting  and
higher order terms,
• f(x,y) = 1 - iVt(x,y)
• This is “Weak Phase Object Approximation”, which assumes
that for a very thin specimen, the amplitude of the transmitted
wave function is linearly related to the projected potential of the
specimen (scattering is weak).
• The projected potential takes into account the variations in z-
direction, and is different for electrons passing at different
angles to the optic axis.
 Charge Density Approximation
• One of the simplest approximations is the charge density
approximation, which gives an idea what an image at a
relatively large defocus will look like.
• For a very thin specimen, the exit wave can be approximated
by a Kinematical expression as: (r) = 1 - itV(r), where t is
the crystal thickness,  is a constant depending on electron
voltage, which is = (2me/h2k), and V(r) is the crystal
potential. Fourier transforming leads to: (u) = (u) - itV(u),
for the decomposition of the wave into spatial frequencies.
• Multiplying by the phase shift term, the modified wave after
the objective lens imaging is: (u) = (u).exp[-i(u)].
• Assuming no beam tilt or astigmatism, the spherical aberration
can be neglected for small values of u, and the exponential
term is expanded as: exp [-i(u)] = 1 - i(u) = 1 - iu2z.
So that, (u) = (u) + u2ztV(u) - itV(u)
Charge Density Approximation
 (u) = (u) + u2ztV(u) - itV(u)
• Back Fourier transforming the modified wave in the image
plane is:
 (r) = 1 - itV(r) + (zt/4)2V(r)
• Taking the modulus squared, the image is:
• I(r) = 1 + (zt/2)2V(r) + ….
• Neglecting all terms containing t2 as small, the crystal
potential and charge density (r) are related by the
equation: 2V(r) = -(4/0).(r)
• I(r) = 1 – (2zt/0).(r).
• For small spatial frequencies and for very thin specimens,
image is proportional to the projected charge density of the
specimen.
 (r) = 1 - itV(r)  The effect of the
specimen is to change the phase of the part
of the wave.
• For a completely perfect microscope, the
image intensity is: I(r) = 1 + [tV(r)]2.
• Inclusion of defocus and spherical
aberration leads to additional phase change
to scattered wave.
Weak Phase Object Approximation
 (u) = (u).exp[-i(u)].
 (r) = 1 - it V(u).exp[-2iu.r-i(u)]du.
Taking the modulus squared to generate image,
and neglecting terms in t2 as small,
• I(r) = 1 - it V(u).exp[-2iu.r-i(u)]du + it
V*(u).exp[2iu.r+i(u)]du
• If specimen is thin and absorption can be
neglected, V(u) is conjugate symmetric:
• V(-u) = V*(u)
• If the crystal has centre of symmetry, then V(u) =
V(-u)
• If there is no beam tilt, (u) = (-u)
 Contrast Transfer Function
• I(r) = 1 - it V(u).exp[-2iu.r-i(u)]du + it
V*(u).exp[2iu.r+i(u)]du leads to:
• I(r) = 1 - it/2 V(u){exp [-2iu.r-i(u)] - exp
[2iu.r+i(u)] - exp [-2iu.r+i(u)] + exp [2iu.r-
i(u)]}.du
= 1 - t V(u).{sin [2iu.r+i(u)] - sin [2iu.r-i(u)]}du
= 1 + t V(u).cos(2u.r){-2sin((u)}.du
{-2sin((u)} is the weak phase object contrast transfer term.
The sign and amplitude of the contrast for the fringe structure
in the image corresponding to the spatial frequency, u is
governed by contrast transfer function.
The electron microscope is a filter, changing the sign and
contrast of the spacings in the image, filtering out some of them.
 Contrast Transfer Function
• Which frequencies are
passed through depends
on the exact value of
objective lens defocus.
 Contrast Transfer Function
• T(u) becomes zero at u = u1.
• One will like the value of u1 to be as large
as possible.
• If T(u) crosses the u-axis, the sign of
transfer function reverses.
• The value of u1 defines the limit at which
the image may be directly interpreted.
It is useful for the graphs to use
reduced coordinates:
u* = (Cs3)1/4.u;
Z* = -(Cs.)-1/2Z

The Phase Shift term is then:

(u) = (u*4 - Z*u*2)
Large pass bands, where
contrast for a range of
frequencies is almost same.
These occur for small
negative values of defocus
(positive z*), where the
transfer function has a value
close to -1 for a broad range
of the values of u*.
 Scherzer Defocus
• Scherzer Defocus corresponds to Z* = 1
• Balance the effect of spherical aberration
against a particular negative value of
defocus.
• All the beams will have nearly constant
phase out to the first cross-over of the zero
axis.
Plots of T(u) (i.e, CTF) or 2sin
(u) with u
 sin  starts at 0 and decreases.

When u is small, the effect of Z

on CTF is large.

The nature of this curve does not

depend on the specimen.
 Contrast Transfer
Function
Plots of T(u) (i.e, CTF) or 2sin
(u) with u 

Smaller Cs values give the larger

values of u1 (i.e. value of u at
which T(u) = 2sin (u) = 0) 
Smaller Cs leads to higher
resolution.

High spatial frequencies  large

diffraction angles  larger effect
of objective lens C
Images of Amorphous Ge film
and their corresponding
diffractograms. Z (Defocus)
has the following values: (A) 1
Sch, (B) -1.87 Sch, (C) -2.35
Sch, (D) -3.87 Sch
 Effect of Aberrations

Well aligned Drift and No

Astigmatism

Some
Astigmatism More drift

More Well
Astigmatism aligned
 Linear Imaging Theory
• Weak phase object approximation suffers from the
defect It does not include the effect of incoherent
imaging phenomena.
• Based on weak phase object analysis, it is possible to
infer that resolution is limited by the extent of broad
pass band. Objective aperture can be used to remove
spatial frequencies beyond this.
• The next step is to consider the average effect of
incoherent aberrations. It can be done either by
approximate analytical methods or full numerical
integration approach.
 Envelope Term
• We use weakly scattering object for the analysis:
• I(r) = 1 + t.V(u).cos(2u.r).{-2sin((u))}du
• It is noted that, (u) = (2/).[z2u2/2 + Cs4u4/4]
(u,f,w) = (/)[(z+f)2(u-w)2 + ½ Cs4(u-w)4]
= (u,0,0) + w.(u,0,0) + fu2 + ….
• Thus, I(r) = 1 + t.V(u).cos(2u.r).{-2sin((u) + w.(u) +
fu2)F(f)S(w)dfdw}du
• This expression is modified as:
• I(r) = 1 + t.V(u).cos(2u.r).{-2sin((u))}E(u).du
where E(u) is the linear envelop term (involving focal spread and
convergence/beam tilt).
E(u) = cos(w.(u)).S(w).dw.cos(fu2).F(f)d(f)
where S and F are the cosine Fourier transforms of the convergence
(w) and focal spread (f) distributions, respectively;
 Envelope Term
• E(u) = [S((u)/2)].[F(u2/2)]
• Gaussian distributions for convergence and focal
spread imply that:
  = 2/4<2>,  = 1/[2<(f)2>]
where < 2 > and <(f)2 > are root mean square
(r.m.s.) values of convergence and focal spread,
respectively..
• S(w) = (/)exp(-w2);
• F(f) = (/).exp(-.f2)
• E(u) = exp(-[(u)]2/4).exp(-22u4/4)
 Physical Significance of
Envelope Function
• The envelop function, E(u) is a damping function,
which limits in a very fundamental way the spatial
frequencies obtained in the image.
• If Sin (u) = 0 for a particular spacing and defocus
value so that this spacing does not appear in the
image, different defocus can be used to increase
contrast.
• If envelop is small, the information can not be
seen in the image. This represents fundamental
limit to attainable resolution.
 Physical Significance of
Envelope Function
• Convergence Contribution: When the beam
direction is changed, the effective value of spatial
frequency to be used in the phase shift term (u) is
changed. Convergence vanishes when:
(u) = 0 = 2u(z + Cs2u2)
i.e. z = -Cs2u2
• This particular defocus value is called the overlap
defocus for the spacing, u.
• At more negative defoci, the convergence
contribution is reduced for larger values of u,
thereby the resolution is effectively improved.
 Physical Significance of
Envelope Function

• For the focal spread, there is no defocus

contribution, and nothing is gained by
changing defocus.
• The envelope term is like a soft aperture,
which fundamentally limits the resolution.
Low Angle Grain Boundary
 Analysis of Diffractograms
One can use the values of defocus (z) and
spherical aberration (Cs) by measuring the radii
of bright and dark rings in the diffractogram.
Bright rings correspond to sin ((u)) = 1; dark
rings correspond to sin ((u)) = 0
sin ((u)) = 1, when (u) = n/2 and n is odd;
sin ((u)) = 0, when (u) = n/2 and n is even
At least two rings are required to find two
unknowns, z and Cs.
(u) = zu2 + (1/2)Cs 3u4
Substituting (u) = n/2 implies that
n/u2= Cs3u2 + 2z
 Plot n/u2 against u2, and fit to a straight line
with slope = Cs3; intercept on u2 axis = 2z.
 Image Simulation
• The HRTEM image needs simulation,
because of the loss of phase information
during formation of an experimental
intensity map.
• This implies that the image does not
represent the structure.
• The experimental image is compared with
simulated images.
Formation of the image is sensitive to the
following factors:
(1) Precise alignment of the beam with respect
to specimen & optic axis,
(2) The thickness of the specimen,
(3) The defocus of objective lens,
(4) Chromatic aberration, which becomes
more important as thickness increases,
(5) Coherence of the beam,
(6) Intrinsic vibration in the material.
 The Multislice Method
• Multislice algorithm is used to calculate the wave
at the bottom of the crystal (neglecting
absorption), and then simulate the image formed
by this wave by using convergence and Cs of the
instrument as input parameters.
• The basic approach is to section the specimen into
many slices, which are normal to the incident
beam.
• It is a fast numerical method of calculating the
diffraction from an arbitrary crystal using
numerical integration approach by exploiting Fast
Fourier Transform (FFT).
The basic physical idea is to
divide the crystal into slices
along the beam direction. Multislice Method
The potential within a slice
is projected onto the first
projection plane: this is the
phase grating.
The amplitudes and phases
generated by interacting
with this plane are
calculated for all the beams.
These diffracted beams
propagate in free space to
the next projection plane,
and this process is repeated.
.
 Multislice Method

The electron wave travelling downwards is scattered by each slice 

Multiplication by a Phase Grating, transformation to reciprocal
space, propagation to next slice by multiplying by a propagator,
back Fourier transformation
 Multislice Method
(Reciprocal Space Approach)
• The multislice method considers three components:
• (i)  describes the electron wave,
• (ii) P is the propagator of the electron wave in free
space: the microscope, and
• (iii) Q is the phase grating: the specimen
• This process can be described by the equation:
 n+1(k) = [n(k).Pn+1(k)Qn+1](k),
where n+1(k) is the wave function in reciprocal space
at the exit of the (n+1)th slice and the symbol 
denotes a convolution; Pn+1(k) is the propagator for
the (n+1)th slice; and Qn+1(k) is the phase grating
function (transmission function) for the (n+1)th slice.
• All functions are in reciprocal space.
 Multislice Method
(The Fast Fourier Transform Approach)

 n+1(k) = F[{F-1[n(k).Pn+1(k)]}.qn+1(r)]
where qn+1(r) is the real space form of Qn+1(k)
• Steps:
(1) Multiply n(k) by Pn+1(k)
(2) Take the inverse Fourier transform of the product
(3) Multiply the result by qn+1(r)
(4) Fourier transform of the final result
 Multislice Method
- The first point is to decide about crystal orientation, and how
to cut it into slices.
- For this, the first step is to set up a two-dimensional unit cell
normal to the beam direction.
- Due to the use of FFTs, the programme takes two-
dimensional unit cell and extends it to produce an infinite
lattice.
 Multislice Method
• After determination of 2D unit cell, the repeat distance along the
beam direction normal to the unit cell is considered.
• This repeat distance, together with 2D unit cell vectors defines
the unit cell for calculation.
• The next step is to represent the 2D unit cell numerically, and size
of his array determines the accuracy of calculation.
• For a right angled unit cell, the largest diffracted beam included
in the calculation along each axis is 2/3p (Å)-1
where p is the distance of two adjacent elements of the array
along each axis.
• The thickness of each slice is chosen to be about 2 Angstroms or
smaller.
• Within each slice, information is provided about: (i) how many of
each type of atoms are present, (ii) positions of atoms as fractions
of the unit cell parameters, (iii) their names and atomic number,
(iv) type of scattering factors to be used, (v) the occupancy of
each atom. Other inputs: slice thickness and acceleration voltage.
 Imaging
• Multislice provides the wave leaving the specimen
• Imaging transfers this information to the final image.
• The wave-function leaving the specimen is altered by both
coherent aberrations of microscope (Cs and z) and
incoherent effects (beam convergence and focal spread).
• Procedure involves shifting the phase of the wave in the
reciprocal space depending upon the values of coherent
and incoherent aberrations.
• To calculate the image, operating conditions of the
microscope should be known fully. The basic parameters
required are: (i) Cs in mm, (ii) defocus in Angstroms, (iii)
Convergence in terms of the radius of a filled aperture in
mRad, (iv) Root mean square standard deviation of focal
spread due to high voltage instability and energy spread of
electrons starting from the filament.
Simulated Images of TiC [011]
Al-TiC Interface in Composite
Experimental and Simulated
HRTEM Images

Symmetric Tilt Boundary ( = 5) in Nb

Atomic Resolution Image

Au [001]
HRTEM Image of Precipitate

(100)Al//(100)Al3Ti (L12)
[100]Al//[100]Al3Ti (L12)
HRTEM Image of Al-Ti
Multilayered Film
 Electron Specimen Interaction

When high voltage electrons traverse a thin foil specimen one

of the primary inelastic interactions is that of inner-shell or
core-shell ionizations.
Core-Shell Transitions
 Characteristic X-ray Production
• Fluorescence yield (): Fraction of ionizations that lead to X-ray
emission. The value of  decreases with decreasing atomic number (Z).
 Characteristic X-ray Production
• Ionization cross-section (Q): Measure of how easily a particular atom
is ionized by a particular electron (units of
ionizations/electron/(atom/cm2).
• Q = 6.51X10-20.[nsbs/UEc2].ln (cs.U)
where ns = number of electrons in the shell
Ec = ionization energy for shell in keV
U = overvoltage = E0/Ec where E0 = incident electron energy
bs and cs are constants.
Q is proportional to (ln E0/Ec)/E0Ec
For a particular X-ray photon (Ec), doubling the incident beam kV
(e.g. from 100 kV to 200 kV) only results in a small decrease in Q.
So, X-ray production is slightly less efficient at high acceleration
voltage.
 Bremsstrahlung X-ray Production

• Bremsstrahlung or Contiuum X-rays: These X-rays

constitute a background intensity across the full energy
range of the X-ray spectrum.
• Bremsstrahlung arises as a result of deceleration of the
high energy electrons in the field surrounding the
nucleus and inner shell electrons.
• Characteristic lines are superimposed on background.
• The intensity at a given energy, E is a function of
atomic number, Z and beam current, i.
X-ray Collection Angle
 Energy Dispersive Spectrometer

Schematic diagram showing the energy dispersive spectrometer.

 Energy Dispersive X-Ray Detector
Si(Li) Detector is used.
• Voltage pulses produced by detector are proportional in size to incoming
X-Ray photon energy.
• The Si(Li) crystal has a band structure, in which conduction band states
are empty, while valence band states are filled.
• When an energetic X-Ray photon is captured, electrons are promoted
into conduction band, leaving holes in valence band.
• Under an applied bias, electrons and holes are swept apart and collected
on the electrodes on faces of the crystal. The ideal number of charges per
incident photon of energy (E in eV) is n = E/ ( = 3.8 eV).
• Detector is cooled for noise reduction. Signal is amplified electronically.
EDS Detector
 EDS Detector
• Si(Li) spectrometer has a protective window of Be (7.6 mm
thick), a front surface electrode of Au (20 nm at front face and
200 nm at back) and inactive layer of Si (20-200 nm). X-rays
pass through those to reach active Si detector.
• The whole crystal is p-i-n device, with relatively shallow
junctions 200 nm deep on either side (dead region) of the
central 3 mm thick intrinsic region (active region).
• Electrons and holes are separated by the internal reverse bias of
a very narrow p-n junction.
• As X-rays penetrate matter more easily than electrons, a thicker
region is required for X-rays to generate electron-hole pairs and
lose all their energy.
• Absorption in the Be window eliminates nearly all X-Rays
below 600 eV. Above 2 keV, all X-Rays are transmitted.
Between 600 eV and 2 keV, absorption decreases with
increasing energy. At 1.5 keV, 70% is transmitted, while at 1.0
keV, 45% is transmitted.
 Energy Dispersive Spectra

EDS spot analysis from Al EDS spot analysis from TiC

Illumination System Artefacts
• Ideally the electron probe should be the sole
source of ionizing radiation incident upon
the specimen.

• In practice, stray electrons and X-rays

generated in the illumination system can
produce X-rays from areas away from the
point of interest
 Sources of
spurious X-rays.
(1) Stray electrons that are
either poorly collimated or
travel around the final C2
aperture holder can generate
specimen characteristic X-
rays.
(2) Hard X-rays that
penetrate the C2 aperture can
fluoresce the specimen.
In both cases, the specimen
X-rays are not localized at
the point of interest, and are
therefore a spurious
contribution to the detected
spectrum.
 Illumination System Artefacts
• To detect any stray radiation, the probe should be
positioned down in a hole in the specimen, and see if the
specimen-characteristic spectrum is still detectable (hole-
count)
• If hole-count is non-zero, either stray X-rays exist, or some
electrons are not confined to the region of the main probe.
• To minimize stray radiation from illumination system:
(i) if stray electrons are present because they pass around
C2 aperture, this area is plugged by a stray aperture;
(ii) if hard X-rays generated in the C2 aperture are
present, they can be reduced by using ultra-thick top-hat
shaped Pt apertures with a tapered bore to restrict further
electron scattering.
Artefacts due to Microscope Stage Region
• Problems arise because the specimen is electron
transparent, and therefore high energy electrons
are transmitted and scattered in the forward
direction.
• About 30% of electrons hitting the pole piece or
stage travel back through the column to hit the
specimen.
• Backscttered electrons are responsible Cu peak
obtained from grid or specimen holder.
• On tilting, backscattered electrons may intercept
the specimen off-axis. These problems increase
with specimen thickness and tilt angle.
 Artifacts in EDS Analysis
• Peak Broadening
Natural width: 2 eV at FWHM
Statistical distribution in number of charge carriers
created by capturing X-Ray photons of a single energy, and
uncertainty created by thermal noise.
• Peak Distortion
Lack of symmetry due to absorption and incomplete
charge collection.
• Si Escape Peaks
A small peak at an energy of 1.7 kV (Si K) less than than
the original peak.
 Artifacts in EDAX Analysis
• Absorption edges of Si and Au
Above the absorption edges of Si and Au, mass
absorption coefficient rises sharply, resulting in decreased
continuum X-radiation..
• Si Internal Fluorescence
Photoelectric absorption of X-rays by Si dead layer causes
emission of Si X-Rays. Small peak at 1.7 keV is observed.
• Stray Signals
Fe, Cr and Ni peaks can come as stray signal from the
wall of the detector. Depends on the tilt of the
specimen.
Background
 Minimum Detection Limit

Atomic concentration of interest = x (<1)

Counting time = t sec.
Total no. of counts at the peak = (xP + B)t
RMS noise or fluctuation = [(xP + B)t]0.5
For signal to be detected, xPt > 5 X noise, (Bt)0.5 (for small x)
x > 5/(P2t/B)0.5  1/(it)0.5 is the minimum detection limit.
 Optimization of AEM for X-ray
Microanalysis
• It is essential to maximize the count rate of the
characteristic X-rays.
• For small probe size and long count rates, contamination
builds up due to hydrocarbons.
• Time to acquire X-rays should be short enough, so that
specimen drift or contamination does not degrade the
spatial resolution.
• Spatial resolutions are optimized by use of small probe
size and thin specimens. These conditions are opposite of
those required for high count rates.
• A compromise is required to have ~±5% accuracy and 10-
30 nm spatial resolution.
 Optimization of AEM for X-ray
Microanalysis
• Choice of acceleration voltage:
It is possible to predict that peak to background ratio
increases with acceleration voltage. Spatial resolution is
also maximized.

• Choice of Electron Gun:

Highest brightness source is desirable when counting
statistics is a limitation. FEG is ideal if high spatial
resolution is required for imaging and microanalysis.

• Choice of Probe Parameters:

One can select the probe size (C1 lens strength) and probe
convergence angle 2s (C2 aperture size) Increasing the
value of 2s increases the probe current.
Optimization of AEM for X-ray Microanalysis
• EDS Variables:
- The EDS detector has fixed geometry with respect to active area (30
mm2) and take-off angle (0o, 20o or 70o).
- The solid angle is maximized to maximize number of detected X-ray
photons by placing it close to specimen.
• Diffraction Conditions During Analysis:
When a crystalline specimen is close to strong Bragg condition,
anomalously high X-ray generation occurs due to electron channeling
effects/Bloch wave interactions – “Borrman effect”.
Can be avoided by operating away from the Bragg condition or with
large 2 (as in STEM)
Advantage - Used for determination of atomic site occupancy through
ALCHEMI (Atom Location by Channeling Enhanced Microanalysis)
• Measurement of X-ray Peak and Background Intensities:
Can be done by the computer.
 Quantitative EDS
• Concentration, C = k(Ispecimen/Istandard)
where k is a sensitivity factor that takes into
account the difference between the generated and
measured X-ray intensities for both standard and
unknown specimen.

ZAF Correction
• Z: Atomic Number
• A: Absorption of X-rays within specimen
• F: Fuorescence of X-rays within specimen
 Background Subtraction
• X-rays are generated by the “bremsstrahlung” process, as
the beam electrons interact with the coulomb field of the
nuclei in the specimen.
• The intensity distribution of the “bremsstrahlung” X-rays
decreases continuously as the X-ray energy increases,
reaching zero at the beam energy.
• The number (NE) of bremsstrahlung photons of energy E
produced in a given time by a given electron beam is
given by Kramer’s law:
NE = KZ(E0-E)/E
Where Z is the average atomic number of the specimen,
and E is the X-ray energy in kV.
 Determination of k Factors
• The k-factor is not a constant.
• It is a sensitivity factor depending on X-ray
detector, microscope, microanalysis
conditions, background substraction and
peak-integration methods.
• The k factors can be determined:
(i) experimentally using standards, and
(ii) by calculation from first principles.
 Experimental Determination of kAB
• A thin specimen of known composition, CA, CB can be
examined, and the values of IA and IB can be obtained
from EDS spectrum.
• Since CA and CB are known, kAB can be found.
• Precautions required:
(i) The standard must be a well-characterized specimen.
(ii) The standard must be capable of being thinned to
electron transparency.
(iii) The thinning process should not introduce chemical
change.
(iv) Thin regions should be characteristic of the bulk.
(v) The thin foil should be stable under the chosen
acceleration voltage.
 Calculation of kAB
• The k-factors are not standard, but are
sensitivity factors.
• k factors obtained on different AEMs are
identical, if the same standard is used at the
same accelerating voltage, same detector
configuration, same peak integration and
background-subtraction routines.
• The k-factors can be calculated from first
principles by involving the determination of
each of the factors influencing X-ray
production.
 Calculation of kAB (First Principles)
• IA/IB = CA/CB.(ZAF) where Z, A and F are atomic number,
absorption and fluorescence corrections, respectively.
• Assuming that A and F are negligible in a thin foil, CA/CB =
1/Z.(IA/IB)

• kAB = 1/Z = [QKKaK/A)A] / [QKKaK/A)B]

= (Qa)AAB / (Qa)BAA
where Z is atomic number correction factor for bulk specimens, Q K is the
ionization cross section for K lines, K is the fluorescence yield for K
lines, a is the ratio of intensities of K/(K + K), and A is the atomic
weight.

• Q is strongly affected by the acceleration voltage.

• Atomic number affects , A and a
• The choice of peak integration method will also affect a.
• Thus, kAB = 1/Z = (Qwa)AAB / (Qwa)BAA
• This equation assumes that equal fractions of the X-
rays generated by the elements A and B are collected
and processed by detector  This assumption is true if
X-rays are neither strongly absorbed nor pass
completely through the detector.
• To take into account the detector effects, the above
equation is multiplied by B/A, where  represents the
detector efficiency in terms of absorption of weak X-
rays. The term, A is a function of mass absorption
coefficient for element A, densities and thicknesses of
Be window, Au contact layer, Si dead layer, and
thickness of active Si layer.
• kAB = 1/Z = [(Qwa)AAB] / [(Qwa)BAA].(A/B)
 Absorption Effect
• Preferential absorption of X-rays from one of the elements  Detected X-ray
intensity will be less than the generated intensity. In such a case, C A is not
proportional to IA.
• The k-factor is modified to account for reduction in I A.
• This is a problem, if specimen is too thick or one or more elements have
lower characteristic energies (< 1 keV).
• If kAB is the true sensitivity factor when the specimen thickness t = 0, then the
effective sensitivity factor for a specimen in which absorption occurs is given
by kAB*
• kAB* = kAB (ACF)
• ACF = {B(t).exp [(-/)Bspec t cosec ]}d(t)/{A(t).exp [(-/)Aspec t
cosec ]}d(t)

 = (n/V)(A/N) where n = no. of atoms of

average atomic weight A in a unit cell of
volume V, & N is the Avogadro’s number.
Absorption path length, t = t cosec 
 Absorption Effect
• The value of (/)Aspec =  (Ci/)Ai
where Ci is the fractional concentration of element
i in the specimen such that:
Ci = 1
The absorption of X-rays from element A by all
elements i in the specimen is summed, including
the self-absorption by element A itself, as well as
by other elements present.
Absorption Effect
 Fluorescence Correction
• X-ray absorption and fluorescence are related, because the
primary cause of X-ray absorption is fluorescence of another
X-ray.
• Strong absorption effects occur when there is a small amount
of one element, whose X-rays are being absorbed by the
presence of a relatively large amount of another element.
• Example: Absorption of Al K X-rays by Ni in Ni3Al, which
leads to significant decrease in the X-ray count for Al and
small increase in Ni X-ray count, because Ni is the dominant
element.
• However, Ni L X-rays are fluoresced by absorption of Al
K,. Ni L X-rays are not considered for microanalysis.
Here, fluorescence is a relatively minor effect and usually
happens for X-rays of less interest.
Fluorescence Effect
EDS Mapping
 Beam-Specimen Interaction

- Minimum detection and spatial resolution are related. High

spatial resolution is obtained by reducing the analyzed volume
and therefore decreasing the signal intensity.
- Then, the minimum detection limit is increased. It is required to
be as low as possible.
 ALCHEMI
• Atom Location by Channeling Enhanced Microanalysis
• Close to two-beam conditions, the Bloch waves interact strongly
with crystal planes  Characteristic X-ray emission is enhanced.
• ALCHEMI is a quantitative technique for identifying
crystallographic sites, distribution and types of substitutional
impurities in crystals.
• In this method, spectrum is acquired under strong channeling
conditions, such that Bloch wave interacts very strongly with a
particular row of atoms.
• The channeling orientation is chosen such that the planes
interacting strongly with the electron beam also contain the
candidate impurity sites.
• Prior idea about the position of substitutional atoms is desirable.
A. Borrman effect: The variation of
characteristic X-ray emission close to
two-beam condition as the beam is
rocked across {400} planes of GaAlAs. B. ALCHEMI allows
The X-rays from Al, which occupies determination of site occupancy
Ga sites follow the Ga X-ray emission, of atom X in columns of atoms A
while As follows complementary and B. By tilting to s> 0 and s< 0,
pattern. BSE is inversely proportional Bloch waves strongly interact
to amount of electron channeling. with planes of A and B atoms.
 ALCHEMI
• When a Bloch wave is maximized on a particular plane of
atoms, the X-ray intensity from the atoms in that plane will be
the highest.
• Orientations are found for which maximum channeling is
observed for the selected atoms, say A and B.
• Suppose there is a substitutional element, X.
• X-ray intensities are measured from each element in
orientations (1) and (2).
• Then, a non-channeling orientation (3) is found, where
electron intensity is uniform for both planes.
• For the non-channeling orientation, the ratio k = IB/IA
where IB and IA are X-ray counts from elements B and A,
respectively for non-channeling orientation.
 ALCHEMI
• For channeling orientations (1) and (2), two
parameters,  and  are defined.
  = IB(1)/k.IA(1)
  = IB(2)/k.IA(2)
• If it is assumed that the element X sits on specific
sites, i.e. it substitutes for atom B, an intensity
ratio term R is defined.
• R = [IA(1).IX(2)] / [IX(1).IA(2)]
• The fraction of atom X on B sites is given by:
• CX = (R-1)/(R-1+-R)
• Fraction of X atoms on A sites is 1 – Cx.
 ALCHEMI
• ALCHEMI gives a direct measure of the
occupation of substituional sites.
• Difficulties may arise because intensity difference
for different orientations are often small.
• Good X-ray statistics is required.
• Acquiring such statistics may become difficult, if
good spatial resolution is also required.
 ELECTRON ENERGY LOSS
SPECTROSCOPY
• When the electron beam traverses a thin specimen,
it loses energy by a variety of processes.
• EELS is carried out to separate the inelastically
scattered electrons and quantify the information.
• The spectrum also contains electrons, which have
not lost any energy.
 Why do EELS?

• It can detect and quantify all elements in periodic

table.
• It is excellent for light elements.
• Better spatial resolution than EDS
• More analytical sensitivity than EDS
• Provides additional information about weakly
bound valence and conduction band electron
distribution
• Provides information about chemical state of each
element.
 Scattering
• Single scattering occurs when each electron undergoes
at most one scattering event.
• Plural scattering (>1 scattering event) imply that the
electron has undergone a combination of events.
• If multiple scattering occurs, the specimen is too thick
for EELS.
• The symbol  refers to scattering semi-angle.
• Characteristic scattering semi-angle, E for an energy
loss, E is give as:
 E = E/2E0  E/(m02) where v = velocity of electron,
m0 = rest mass of electron,  = (1-v2/c2)-0.5
The intensity of scatter through an
angle, .
The incident beam (i) and
elastically scattered beams (ii) are
limited by thermal energy of the
beam (source dependent).
The energy loss electrons (iii), (iv)
and (v) show different angular
spreads depending on the type of
inelastic interaction responsible for
energy loss
Ionization and plasmon loss
electrons will like phonon scattered
electrons, be picked around the
elastic diffraction maxima, but will
not enter the spectrometer..
 The angular spread of
energy loss electrons
relative to the diffraction
pattern in the TEM.
If an objective aperture is inserted, a
large fraction of the ionization loss
and plasmon loss electrons will
contribute to the image.
 Energy Loss Processes
• Four most common energy loss processes:
(i) Excitation of phonons,
(ii) Excitation of electrons from one energy
level within a shell to another, or from one shell
to another (intra- and interband transitions)
(iv) Inner shell ionization,
(v) Collective, quantized oscillations of
conduction band electrons.
 Energy Loss Processes
• Phonon excitations result in energy losses of 0.02
eV  detection is below the resolution limit of
spectrometer.
• Single electron interactions occur in the range of 1-
150 eV, but their excitation probability is small in
metals.
• For materials science, two excitation processes are of
major interest: (a) inner shell ionization, and (b)
collective, quantized oscillations of conduction band
electrons (plasmons).
 Energy Loss Spectrum
• Three principal regions of the EELS spectrum contain:
• (i) The zero-loss peak, which consists primarily of elastic
forwarded-scattered electrons, but also contains electrons
that have suffered minor energy losses.
(ii) The low loss region up to energy loss of 50 eV
contains electrons, which have interacted with weakly
bound outer shell electrons of atoms. (predominantly
plasmon)
(iii) Electrons in the high-loss region have interacted with
more tightly bound inner shell or core electrons.
(ionization loss)
 Zero Loss Peak
Spectrum from the low
loss region of a thin
specimen of pure Al.
The small peak
adjacent to the zero loss
is the first plasmon loss
peak at 15 eV.
 Low Loss Region of the Spectrum (5-50 eV)
Spectrum from low loss region
of thicker sample of pure Al:
Due to free electrons.
Spectrum from low loss region
of stainless steel: Broad
spectrum when free electrons
are not abundant, - interactions
Plasmon loss in materials with free with tightly bound valence
electron structures. electrons in
amorphous/crystalline
materials.
Ep = (h/2)p =
(h/2)(ne2/0m)0.5
n = free electron density,
0 = permittivity in
vacuum, E  5-25 eV
Low Loss Region of the Spectrum (5-50 eV)

The low-loss spectrum from

specimens of Al and Al-
containing compounds,
showing differences in
intensity that arise from
differences between bonding
in different materials.
 High Energy Region of Spectrum

• Detection of the beam electron that ionized

the atom is independent of whether the
atom emits an X-ray or Auger electron.
• EELS is not affected by fluoresence-yield
limitation that restricts light element X-ray
analysis.
High Energy Region of Spectrum
• The edge falls in intensity, never reaching zero, until the
beam energy is reached.
• To ionize an atom, an electron loses just the critical energy
necessary for ionization (Ec). Ec = EK for K-shell
ionization, Ec = EL for L-shell ionization.
• Losing more energy transmits more energy to the ejected
K-shell electron, and has no effect on subsequent de-
ionization and X-ray or Auger electron generation.
• The chances of losing more energy than Ec during the
interaction drop off rapidly until E = E0.
• Background intensity after ionization edge consists of the
tail of edge.
• If there are many elements in the specimen, then the
background intensity increases in proportion to the number
of edges in the spectrum.
 High Energy Region of Spectrum
• The K-shell electron is in the 1s state and gives rise to a
single K-edge.
• In the L-shell, the electrons are in either 2s or 2p orbitals.
If a 2s electron is ejected, we get L1 edge, and ejection of a
2p electron causes either L2 or L3 edge.
• Ionization cross-sections are small, and the mean free
paths are relatively large.
• It is possible to ionize an atom by transfer of E>Ec
• Chances of ionization become less with increasing energy
above Ec, because the value of cross-section decreases with
increasing energy.
• Ionization-loss electrons have a distribution showing a
sharp rise to a maximum at E = Ec, followed by slowly
decreasing intensity towards background. This triangular
shape is called an “edge”.
High Energy Region of Spectrum

The full range of

possible edges due to
inner-shell ionization,
and their associated
nomenclature.
 High Energy Region of Spectrum
• Above 50 eV , the spectrum consists of rapidly
falling background upon which are superimposed
peaks or edges.

Background consists of electrons that have suffered multiple or

random losses, until the first edge is reached. Increased thickness
will increase the background intensity.
 EELS Spectra

Typical triangular shape Hydrogenic edge

spectrum from isolated hydrogen superimposed on
atom: Hydrogenic ionization background from plural
edge. elastic scattering.
Energy loss near-edge structure (ELNES) within 50 eV of Ec due
to chemical bonding
Extended energy loss fine structure (EXELFS) – Oscillations
detected at >50 eV from E caused by diffraction by ionized atom.
 High Energy Region of Spectrum

Ionization edges obtained from a specimen of BN supported on

a thin carbon film showing large variation in edge heights and
rapidly decreasing background.
 High Energy Region of Spectrum
Correspondence
between the
energy levels of
electrons
surrounding
adjacent Ni & O
atoms.
The plasmon peak
corresponds to
valence electron
band below Fermi
level.

Ni edge in NiO appears different from that in pure Ni. Ejected L

shell electrons may end up in unfilled d-band states.
 EELS Spectrometer

• The spectrometer consists of a magnetic prism: (A) dispersion

and focusing of the electrons in the plane of the spectrometer;
(B) lens focusing action normal to spectrometer.
EELS Spectrometer
PEELS comprises a YAG
scintillator coupled via fibre
optics to a thermoelectrically
cooled semiconductor
photodiode array in the
dispersion plane of the
spectrometer.
The resultant spectrum across
the whole energy range
simultaneously.
There is an integration time for
the PEELS spectrum.
EELS Spectrometer
 Spectrometer Collection Angle

• The spectrometer collection angle is a crucial factor

in EELS.
• Large collection angles give high intensity in the
spectrum.
• In image mode without an objective aperture, energy
resolution is good.
• In diffraction mode, the entrance aperture controls .
• Large aperture and large  worsens the energy
resolution.
Spectrometer Collection Angle
 is the semi-angle subtended at
the specimen by the entrance
aperture to the spectrometer.
For dedicated STEM:

Depending on diameter (d) of the

spectrometer entrance aperture and
the distance from the specimen to
aperture (h),  (in radians)  d/2h.
Since h is constant, the range of 
is controlled by number and size of
available apertures.
If h  100 mm, for 1 mm size
aperture,  = 5 mrads.
Spectrometer Collection Angle
The angular distribution of
electrons entering the
spectrometer aperture
below the centre of the
specimen is independent
of the entrance aperture
size.
The angular distribution of
electrons contributing to
TEM image is controlled
by the size of objective
aperture in the back focal
plane of objective lens.
 Ionization Loss Spectrometry
• Choice of operating parameters
• Accumulation of the spectrum
• Calibration of the spectrum
• Spurious effects in the EELS spectrum
• Quantification
• Limits to microanalysis using ionization
edges
• Choice of operating parameters
• In STEM mode, it is essential to optimize the gun
alignment, filament saturation as well as largest
C2 aperture.
• In STEM mode, the region analyzed is selected by
positioning the probe on the point of interest in the
STEM image.
• TEM mode can be operated using either imaging
or diffraction.
• For high energy resolution, the angular spread of
electrons entering the spectrometer is reduced:
STEM or TEM diffraction mode. In TEM imaging
mode, objective aperture is used.
Qualitative Microanalysis

Ionization edge is the

energy loss at which
there is a discrete
increase in the slope of
the spectrum.
This value is the edge
onset, Ec, i.e. the
critical ionization
energy.
Qualitative Microanalysis: Stainless Steel Sample

TiC precipitate TiN precipitate

 Quantitative Microanalysis
• It is assumed that the electrons contributing to the edge have
undergone single ionization event.
• Let us assume that the K edge is being quantified.
• The K-shell intensity above background, IK is related to the
probability of ionization, PK and total transmitted intensity, IT 
IK = PKIT
• In a thin specimen, incident intensity  IT, neglecting backscatter
and absorption effects.
• PK = NK exp (t/K) where N is the No. of atoms per unit area of
specimen with thickness, t, K is the ionization cross-section, and
K is the mean free path for ionization losses.
• As the exponential term is close to unity, IK = NKIT.
• Therefore, N = IK/(K.IT).
• It is possible to measure the absolute number of atoms per unit
area of the specimen, by measuring the intensity above
background in the K edge and dividing by total intensity in
spectrum and ionization cross-section.
 Quantitative Microanalysis
• In a spectrum containing two edges from elements, A and
B:
• NA/NB = IKAKB/ IKBKA
• Similar expressions also apply to L, M edges, etc.
• We could accurately subtract the background under the
ionization edge.
• One can not gather the whole of the energy-loss spectrum
out to the beam energy, E0, because above 1-2 keV, the
intensity drops to a level close to the system noise.
• While ionization-loss electrons can have any energy
between EC and E0, the intensity in the edge falls to
background level within about 100 eV  it is imperative to
restrict the integration of spectral intensities to a window,
 of 20-100 eV.
• IK() = NK()IT().
• IT() is the intensity of zero-loss electrons combined with
low loss electrons.
 Quantitative Microanalysis
• It is not possible to collect spectrum over solid angle = 4
steradians.
• Considering collection semi-angle = ,
• IK() = NK()IT().
• The factor, K() is termed the “partial ionization cross-
section.”
• The absolute quantification for N is given by: N =
IK()/[K()IT()].
• For ratio of two elements, A and B, the low loss intensity drops
out as:
• NA/NB = [IKA().KB()] / [IKB() KA()].
• Two essential steps are required:
(i) The background subtraction to obtain, IE;
(ii) The determination of partial ionization cross section, K().
 Background Subtraction
• Curve fitting

• The graphical method

• Using difference spectra

 Curve Fitting

A window,  is chosen in the background before the edge onset

and fit a curve to the channels.
Then, the curve is extrapolated over the energy window  under
the edge, and subtracted to get edge intensity, IK.
 The graphical method

The curve-fit region before the unprocessed edge is extrapolated to

give estimated background, which is then removed, leaving
background subtracted edge. Curve fit equation: I = AE-r, where I is
the intensity in the channel of energy loss E, A & r are constants.
 Difference Spectra

It involves taking two spectra, offset in energy by a few

kV, and subtracting one from another.
Partial Ionization Cross-section
• Can be done using standards with known
composition.
• The standard and unknown have same
thickness and bonding characteristics.
• Spectra should be gathered under identical
conditions, such that , , E0 and t are
same.
 Spatial Resolution
• Unlike EDS, beam spreading is not a major factor in
EELS.
• The factor affecting resolution in STEM mode or in
probe-forming mode on TEM is the size of the probe
with sizes < 10 nm.
• In TEM mode, the spatial resolution is a function of
selecting aperture, i.e. spectrometer entrance aperture
and its effective size at the plane of specimen.
Spatial Resolution
 Detectability Limits
• The factors influencing the detectability
limits are:
• (i) The edge intensity,
• (ii) The signal-to-background ratio (jump
ratio)
• (iii) The efficiency of signal detection,
• (iv) The time of analysis.
 Specimen Preparation

Sectioning of disc from bulk component.

Mechanical Thinning
 Mechanical Polishing
• A polished, parallel-sided disc is prepared and then thinned
from one or both sides using a rigid jig to maintain planar
geometry. The sample thickness should be 100 m.

• Once the thickness of the specimen is reduced to 100 m or

less and the surfaces have been polished to metallographic
finish, the disc is secured on an optically flat baseplate, and
dimpled.

• In dimpling, a fine grinding medium removes material from

the central area of the disc sample, as it is rotated in contact
with a polishing wheel. This is, essentially, a lapping process
in which particles of the grinding media are displaced in the
region of contact shear and continuously fractured to expose
new cutting edges.
Cross-section Specimen Preparation
(a) Rectangular sections are cut
from the wafer.
(b) The sections are glued
together to form a block
greater than 3 mm in
thickness, from which a 2.8
mm diameter rod is trepanned.
(c) The rod is inserted and glued
into a 3 mm outer diameter
metal tube.
(d) Thin sections are then cut
from this assembly.
(e) - (g) Finally, these sections are
mechanically thinned.
Jet Polishing
In jet polishing, a current is
passed through a stream of
the polishing solution as it
impinges on the disc sample.

Thinning is accomplished by
electropolishing at a high
current density.
 Ion Milling

Precision ion milling permits thin film multilayer assemblies to be

thinned perpendicular to the plane of assembly.
Two ion guns are normally used in order to sputter from both sides
of sample. The sample is rotated about the axis perpendicular to the
plane of assembly.
 Extraction Replica

Instead of preparing a thin slice from a component for direct

examination, it is possible to take a replica from the surface.
 Reasons:
Replica Methods
• Non-destructive examination may be necessary.
• Selecting one component of a complex sample for investigation.
• Extracting specific phases from a polyphase material. Suitable
chemical etchants can be used to isolate a selected phase.
• Correlation of microstructures using alternative imaging methods.

A negative replica of the surface is obtained on a flexible, soluble plastic. The

plastic may be cast in place and allowed to harden, or it may be a plastic sheet that
has been softened with a suitable solvent and then pressed onto the surface before
allowing the solvent to evaporate. Once the plastic replica is hardened, it can be
peeled away from surface and then shadowed with gold palladium.
After shadowing, a carbon film 100-200 nm in thickness is deposited on the plastic
replica and the plastic is then dissolved in organic solvent.
The carbon film retains particles removed by plastic from replicated surface, as well
as heavy metal shadow which reflects surface topology.
The carbon replica is then rinsed and collected on a fine mesh copper grid.
 Books for Further Reading
• Sir P.B. Hirsch, A. Howie, Sir R. Nicholson, D.W. Pashley, and M.J.
Whelan, Electron Microscopy of Thin Crystals, Robert E. Krieger
Publishing Company, Malabar, Florida, 1977 (ISBN 0-88275-376-2)
• Gareth Thomas and Michael J. Goringe, Transmission Electron
Microscopy of Materials, John Wiley & Sons, New York, 1979 (ISBN 0-
471-12244-0)
• J.W. Edington, Practical Electron Microscopy in Materials Science, Vols.
1-5, Philips, Eindhoven, 1974.
• David B. Williams and C. Barry Carter, Transmission Electron
Microscopy, Vols. I. Basics; II. Diffraction; III. Imaging; IV. Spectroscopy,
Plenum Press, New York, 1996, ISBN 0-306-45247-2
• David B. Williams, Practical Analytical Electron Microscopy in Materials
Science, Philips Electronic Instruments, Inc., Electron Optics Publishing
Group, NJ, USA, 1987.
• John C.H. Spence, Experimental High-Resolution Electron Microscopy,
Clarendon Press, Oxford, 1981 (ISBN 0-19-851365-8)
Thank You !!