sem electron backscattered diffraction

Report
RMS Summer School 2010, Leeds
SEM ELECTRON BACKSCATTERED DIFFRACTION
Geoffrey E. Lloyd
School of Earth & Environment, Leeds University
http://www.see.leeds.ac.uk/research/igt/people/lloyd/rms/
Acknowledgements:
Niels-Henrik Schmidt, Austin Day, Pat Trimby (formerly HKL Software)
Obducat CamScan (Dick Paden)
Dave Prior (University of Liverpool)
Plan
•
•
•
•
•
•
•
•
•
Basic principals
EBSD pattern recognition
EBSD problems
Orientation contrast imaging
Automated EBSD analysis
Specimen requirements
EBSD applications
A few examples
Concluding remarks
Basic Principals
Electron:Sample Interaction
Single
nucleus
incident
electrons
Multiple scattering
emitted
electrons
‘focussed’
emitted
electrons
‘unfocussed’
‘lost’
electrons
‘lost’ electrons
Effects of Scattering
1. creates a ‘point’ source of
electrons, with all possible
trajectories, within the
material
2. crystalline materials:
electrons are diffracted by the
crystal lattice planes when the
Bragg condition is satisfied
note different
lattice spacing
Scattering from single lattice planes
Cone of
Intense Electrons
Electron Beam
Each lattice plane (hkl)
gives rise to 2 diffraction
‘cones’
Diffracting
Plane
Tilted
Specimen
Kikuchi
Lines
Phosphor
(after A. Day)
For SEM electron wavelengths, opening cone
angles are close to 180
Scattering from 3D lattice planes
• A 20keV electron beam
strikes a sample tilted at
65-75°
• The crystal structure at
the point of incidence
diffracts the electron
beam according to
Bragg’s law, nl = 2dsinq
• Each lattice plane (hkl)
gives rise to 2 diffraction
‘cones’
(after J. Hjeiling)
The image (EBSD pattern) is
unique for the crystal orientation
& composition at the point of
incidence on the
sample
(after
A. Day)
Typical SEM EBSD set-up
incident electron beam:
8-40kV, 0.01-50nA
EBSD detector - position
usually constrained by
chamber geometry
low-light sensitive (now
digital; originally analogue)
emitted
electrons
Specimen:
Surface normal
typically
inclined 60°-80°
to beam
EBSD detector distance
set to give ~90 °
angular range in EBSP
EBSD Pattern Recognition
EBSD pattern recognition
• EBSD patterns are unique for a specific crystal orientation
• The pattern is controlled by the crystal structure: space
group symmetry, lattice parameters, precise composition
• Within each pattern, specific ‘bands’ (i.e. pairs of ‘cones of
diffraction’) represent the spacing of specific lattice planes
(i.e. dhkl)
• EBSD pattern recognition compares the pattern of bands
with an ‘atlas’ of all possible patterns in order to index the
crystal orientation depicted
• This process WAS manual – it is NOW automated !
• Example - next slide
EBSD Patterns
diffraction from
specific lattice plane
pyrite
width = 1/d-spacing
FeS2
major crystal
‘pole’
HOLZ ring
1st order diffraction
• Unique for crystal orientation &
composition at the point of
beam incidence
• Can be >100° of total crystal
projection - easy to index as
symmetry decreases
• Spatial resolution (1m)
• Some pattern details:
2nd order diffraction
Example: pattern indexing
Original pattern
manual/auto-indexed bands
Computer indexed pattern
EBSD pattern ‘range’
Determined by crystal symmetry – defines the crystallographic
unit triangle that repeats the range of patterns over a sphere
Example: fcc (Cu)
Example: trigonal (quartz)
0001
001
rotate
101
unit
triangle
011
symmetrically
111
equivalent
unit triangle
-2110
11-20
-12-10
unit triangle
Indexing requirements
• SEM geometry:
beam energy, specimen & detector positions & orientations
usually fixed per SEM
• Crystallography:
sample lattice parameters & Laue/space group
input per phase (i.e. composition) as required
• Diffraction characteristics:
relative diffraction intensities from different (hkl) lattice planes
calculated per phase as required
Creating a crystal database
1. Select a crystal
e.g.
aluminium
2. Input lattice
parameters
e.g.
a = b = c = 0.405nm
a = b = c = 90
3. Input crystal
symmetry
e.g.
cubic
Laue group =
Space group =
or
m3m
225
Fm-3m
4. Input crystal
unit cell
e.g.
Atom
Al
Al
Al
Al
z
0
0.5
0.5
0
symmetry indicates
4 atomic positions
x
0
0
0.5
0.5
y
0
0.5
0
0.5
Occ
1
1
1
1
Create a diffraction database
• To index EBSD patterns, we must know the relative intensities of the
(Kikuchi) bands (reflectors) in the patterns
• Most approaches use the kinematic electron diffraction model
• This model calculates the structure factor (intensity) for each (hkl)
reflecting plane:
intensity of
(hkl) plane
number
of atoms
  F (hkl )
lattice planes


F ( hkl )   f g  exp  2ih  xg  k  y g  i  z g 
structure factor
for (hkl) plane
N
g 1
atomic
scattering
factor
atomic position
of atom g
Diffraction database
Conventional software packages automatically calculate
the diffraction (reflector) database of (relative) intensities
e.g. aluminium:
Reflectors
No.
{111}
4
2.338
100.0
{200}
3
2.025
69.4
{202}
6
1.432
27.6
{113}
12
1.221
18.2
{222}
4
1.169
16.2
etc.
d-spacing Å Intensity %
EBSD Problems
Pseudo-symmetry
Spatial resolution
Angular resolution
Specimen preparation - see later
Pseudo-symmetry
• Occurs where 2 orientations cannot easily be distinguished
due to an apparent n-fold rotation axis
• Especially common in lower symmetry crystal structures:
e.g. the orthorhombic structure with a  b can appear to be tetragonal
when viewed down the c-axis
• Specific examples (minerals):
quartz
trigonal but can appear pseudo-hexagonal
olivine
orthorhombic but can appear pseudo-hexagonal
plagioclase triclinic but can appear pseudo-monoclinic &/or pseudohexagonal
Example
Quartz:
60° rotation about
the c-axis
Often results
in similar
EBSD patterns
Recognition
depends on the
identification of
‘minor’ bands
(black) - often
not selected
automatically
negative
rhomb
distinguishes
positive
rhomb
Spatial resolution
• Depends on the penetration &
deviation of electrons into a sample
(plus beam diameter)
• Typically ranges from ‘few’ m for
W-filament SEM to a few 100nm for
FEG SEM
Example:
P
R
R
• Penetration depends on:
sample atomic number
accelerating voltage
beam current
(plus, coating depth &
surface damage - see
later)
P
note ‘down slope’ effect of tilting – scan ‘uphill’
Several Monte Carlo based
simulation packages are
available via the Web
(e.g.
http://www.gel.usherbrooke.ca/casino/index.html)
Angular Resolution
• Angular resolution of an individual EBSD pattern is typically ~1
• Also important when determining the misorientation between two
(adjacent) crystal lattices (e.g. grains) – ‘misorientation analysis’ is
becoming a popular application of EBSD as it provides information
on sample properties & behaviour
• But, calculations of misorientation axes from 2 individual measurements
with misorientation of <15° contain increasingly large errors
• Angular resolution depends typically on basic EBSD set-up configuration,
EBSD pattern quality & hence indexing software (parameters,
composition, pseudo-symmetry, etc.)
• Examples (next slides)
Angular resolution 1:
sample-detector considerations
Small detector distance
Large detector distance
poor for indexing but good angular
resolution
good for indexing but
poor angular resolution
important for constraining
misorientation axes.
Angular resolution 2:
effect of angle imaged
large angular spread:
low angular resolution
low angular spread:
good angular resolution
Changes in high resolution EBSD patterns can be used to define better
rotation angles & more accurate misorientations
Orientation Contrast
Imaging
Control of crystal orientation
on emission signal
polycrystalline sample
note variation in image grey-scale level - depends on penetration &
emission, which depend on crystal orientation
EBSD microstructural images
• Electron beam is scanned over an area of a tilted sample, rather than
positioning the beam on a point for EBSD patterns
• Forescattered
electrons (FSE) with
intensities determined
by penetration (i.e.
crystal orientation) are
emitted towards the
EBSD detector
quartzite
• FSE signal detected
by silicon devices
attached to EBSD
detector
• FSE Orientation Contrast image of variation in crystal orientation contrast variations only qualitative (next slide)
Warnings
2 grains with the same
orientation can have
different OC signals
beam
2 grains with different
orientations can have the
same OC signal
beam
due to different
‘incident angles’
due to effectively
the same
‘incident angles’
OC images not quantitative !
• Grey-levels cannot be
inverted to give orientation
• Grains in different
orientations could have the
same grey shade
• One image may show only
60% of boundaries – so,
move the image slightly
Percentage of boundaries imaged
• Grains in the same
orientation but different
positions have different grey
shades
120
100
80
60
40
20
0
0
2
4
6
No of Images Used
(after D. J. Prior)
8
10
Automated EBSD Analysis
Automated EBSD analysis
• Computer controlled movement
of the electron beam across a
sample
• EBSD pattern ‘captured’ at each
point
• Indexing of EBSD patterns is via
pattern recognition software
• Software writes the crystal orientation
(3 Euler angles), & phase information
per pattern to a data-base for later
analysis
• BUT – important to run a manual visual
check of solutions before the
automated analysis!
Automated EBSD
analyses
crystal orientation variation
provides a variety of
information
orientation contrast
pattern quality - ‘strain’
P
T
crystal orientation ‘pole’ figures
many other parameters:
e.g. ‘misorientation’
(becoming very important in
microstructural analysis)
Specimen Requirements
Specimen Preparation
• Polished blocks, thin-sections, natural fractured or grown
surfaces
• Surface damage (m-mm) created by mechanical polishing
must be removed:
• chemical-mechanical (‘syton’) polish
• etching
• electro-polishing
• ion beam milling
• Insulating samples may require very thin carbon
coat, but uncoated samples may perform OK - next
Charging Problems
• Reduce charging by coating but only at expense of image
detail &/or resolution
Uncoated
3-5nm
C coat
K-feldspar. 20keV ~15nA (after D.J. Prior)
• Note: specimen damage can occur in absence of charging
Effect of coating on OC images
uncoated
~4nm C
200m
(after D.J. Prior)
~8nm C
Interim Summary
• Orientation contrast images:
variations in crystallographic orientation & sample microstructure
• EBSD patterns:
full crystallographic orientation of any point in OC image
• Spatial resolution:
~100nm (FEG, metals) to ~1m (W, rocks)
• Angular resolution:
~1 - 2° (misorientation >5 °)
• Materials:
most metals & ceramics; many minerals - depends on composition
• Automated analysis:
100’s of EBSD patterns/second (record ~800/sec via stage scanning)
but indexing accuracy may suffer (use of fast or sensitive EBSD
detectors increasing depending on requirements)
EBSD Applications
What can EBSD be used for?
• Measuring absolute (mis)orientation of known materials most popular/obvious usage
• Phase identification of known polymorphs - becoming
popular
• Calculating lattice parameters of unknown materials difficult, only possible for relatively simple structures?
• Measuring elastic strain
• Estimating plastic strain on the scale of the electron
beam activation volume
• Estimate aggregate elastic stiffness matrix (Cij) from grain
crystal texture – used to predict material properties (e.g.
thermal, electric, acoustic, magnetic, etc.)
Recommended applications
• Tremendous significance for many types of materials
research, including:
- deformation & recrystallisation
- understanding processing histories
- effects of pre-heating & heat treatments
- identifying phases in multi-component systems
- microstructural characterisation & calibration
(including boundary geometry, etc.)
- modelling microstructural processes
- constraining micro-chemical data
- estimating physical properties (e.g. elastic, thermal, sonic, electric,
magnetic)
- etc.
Example applications
Crystal orientation data from
SEM/EBSD
• Individual orientation measurements related to
microstructure:
crystal lattice preferred orientations/texture analysis (i.e. inverse/pole
figures, orientation distribution functions
misorientation data (similar types of plots)
• Non destructive
data be collected from representative samples
• Automated
statistically large/viable data sets acquired
• BUT! Samples must be oriented:
Materials - RD, ND, TD
Rocks – X, Y, Z or NSEW
CPO/texture analysis
• SEM/EBSD:
faster than optical & gives full crystallographic information
faster than X-ray goniometry & applicable to all crystals
faster & simpler than neutron diffraction & applicable to all crystals
does not suffer from problems associated with ODF calculations
• BUT! only surficial analysis:
restricted to electron beam penetration depths
not 3D (unless incorporate serial &/or three orthogonal sections)
Example: quartz
colour-coded
pole figures
misorientation
angle profile
subgrains
Dauphine
twins
Automated phase identification
100m
Crystal
orientation
(Euler) image
100m
Phase 1:
quartz
100m
Phase 2:
feldspar
Phase 3:
mica
In Situ Experiments
“Crystal-probe” HT (1200C) FEGSEM
(D.J. Prior, University of Liverpool)
Note: column tilted to 70° - allows horizontal sample movement
& greater access to sample chamber (i.e. various other electronspecimen signal detectors)
Publication rate - EBSD papers
projected 2010
(~3500)
3500
3000
April
2010
No. publications
2500
2000
1500
1000
500
0
1990
y = 0.6714x3 - 4013.4x2 + 8E+06x - 5E+09
R² = 0.9992
1992
1994
1996
1998
2000
Date (yrs)
2002
2004
2006
2008
2010
Selected bibliography
•
As of April 2010, searching for EBSD in Web of Knowledge results in ~3000 hits! So, here are
some ‘early’ papers.
•
Lloyd, G.E. 1987. Atomic number and crystallographic contrast images with the SEM: a review of
backscattered electron techniques. Mineralogical Magazine 51, 3-19.
•
Schmidt, N.-H. & Olesen, N.O. 1989. Computer-aided determination of crystal lattice orientation from
electron channelling patterns in the SEM. Canadian Mineralogist 27, 15-22.
•
Randle, V. 1992.Microtexture Determination and its Application. The Institute of Materials, London
174pp.
•
Randle, V. 1993. The Measurement of Grain Boundary Geometry. Institute of Physics Publishing,
Bristol, 169pp.
•
Field, D.P. 1997. Recent advances in the application of orientation imaging. Ultramicroscopy 67, 1-9.
•
Wilkinson, A.J. & Hirsch, P.B. 1997. Electron diffraction based techniques in scanning electron
microscopy of bulk materials. Micron 28, 279-308.
•
Humphreys, F.J. 1999. Quantitative metallography by electron backscattered diffraction. Journal of
Microscopy 195, 170-185.
•
Prior, D.J. 1999. Problems in determining the orientation of crystal misorientation axes for small
angular misorientations, using electron backscatter diffraction in the SEM. Journal of Microscopy 195,
217-225.
•
Prior, D.J. et al. 1999. The application of electron backscatter diffraction and orientation contrast
imaging in the SEM to textural problems in rocks. American Mineralogist 84, 1741-1759.
•
Trimby, P.W. and Prior, D.J. 1999. Microstructural imaging techniques: a comparison between light
and scanning electron microscopy. Tectonophysics 303, 71-81.
•
Wilkinson, A.J. 1999. Measurement of small misorientations using electron back scatter diffraction.
Electron Microscopy and Analysis, Institute of Physics Conference Series, 161, 115-118.
Crystallographic data sources
• Pearson’s Handbook, Desk Edition, Crystallographic data for
intermetallic Phases, ASM international, 1997. ISBN 0-87170603-2.
• American Mineralogist - http://www.geo.arizon.edu/xtal-cgi/test
• International Tables for Crystallography Volume A: Space Group
Symmetry. Edited by T. Hahn, revised edition, 1996. ISBN 07923-2950-3; see also:
http://ylp.icpet.nrc.ca/SGHT/1983/
http://www.cryst.ehu.es/cryst/
• Altwyk site: http://ylp.icpet.nrc.ca/altwyk/
Laboratory demonstration
• Aims to show:
basic SEM/EBSD set-up
orientation contrast imaging
EBSD pattern capture & indexing (including crystal & diffraction
database construction)
automation
questions & answers (?) plus general advice

similar documents