X-Ray Diffraction

Structural Analysis

PH 762 Experimental Techniques

Dr. Zulqurnain Ali
Waves:
A wave can be described as a disturbance that travels through a medium from
one location to another location.

Electromagnetic Waves
Do not need matter to transfer energy.
Are made by vibrating electric charges and can travel through space by
transferring energy between vibrating electric and magnetic fields.

Material Speed (km/s)

Vacuum 300,000

All matter contains charged particles that are always moving; therefore, all objects emit EM waves. Air <300,000
Water 226,000
Glass 200,000
Diamond 124,000
Electromagnetic Spectrum
Interference

One of the characteristics of a WAVE is the ability to undergo INTERFERENCE. There are TWO types.
Diffraction
This right triangle is
SIMILAR to the one made
P
by y & L.

q
q B.C. d
d

P.D.

P.D. = dsinq
P.D.

Notice the blue wave travels farther. The difference

in distance is the path difference.
Generation of X-rays

1665: Diffraction effects observed by Italian mathematician Francesco Maria Grimaldi

1868: X-rays Discovered by German Scientist Röntgen
1912: Discovery of X-ray Diffraction by Crystals: von Laue
1912: Bragg’s Discovery

Early X-Ray Tube (1900): This tube

was used for diagnostic
radiography and was made about
1900.
Characteristics of Common Anode Materials
Material At. # Ka1 (Å) Ka2 (Å) Char Opt Advantages (Disadvantages)
Min kV
(keV)
Cr 24 2.290 2.294 5.98 40 High resolution for large d-spacings,
particularly organics (High attenuation in
air)
Fe 26 1.936 1.940 7.10 40 Most useful for Fe-rich materials where
Fe fluorescence is a problem (Strongly
fluoresces Cr in specimens)

Co 27 1.789 1.793 7.71 40 Useful for Fe-rich materials where Fe

fluorescence is a problem

Cu 29 1.541 1.544 8.86 45 Best overall for most inorganic materials

(Fluoresces Fe and Co Ka and these
elements in specimens can be
problematic)
Mo 42 0.709 0.714 20.00 80 Short wavelength good for small unit
cells, particularly metal alloys (Poor
resolution of large d-spacings; optimal kV
exceeds capabilities of most HV power
supplies.)
Generation of X-rays
• X-rays are produced whenever matter is irradiated with a beam of high-energy charged particles or photons
• In an x-ray tube, the interactions are between the electrons and the target. Since energy must be conserved, the energy
loss from the interaction results in the release of x-ray photons

The photoelectric effect is responsible for generation of

characteristic x-rays.

An incoming high-energy photoelectron dislodges a

k-shell electron in the target, leaving a vacancy in
the k-shell
An outer shell electron then “jumps” to fill the
vacancy : L-shell to K-shell jump produces a K a x-ray.

M-shell to K-shell jump produces a Kb x-ray

The energy of the Kb transitions is higher that that of the Ka transitions, but
because they are much less frequent, intensity is lower
Continuous and Characteristic Spectrum Characteristic Wavelength values (in Å)
for Common Anode Materials

Anode Ka1 Ka2 Kb

Cu 1.54060 1.54439 1.39222
Cr 2.28970 2.29361 2.08487
Fe 1.93604 1.93998 1.75661
Co 1.78897 1.79285 1.62079
Mo 0.70930 0.71359 0.63229
Bragg‘s Law

In order to keep these beams in phase,

their path difference (SQ + QT) has to equal one or multiple X-ray wavelengths (nλ)
i.e SQ + QT = nλ or SQ + QT = 2PQ sin θ = 2d sin θ = nλ
Hence the path difference depends on the incident angle (θ) and spacing between the parallel crystal planes (d).
figure taken from: www.chem.ox.ac.uk
Particle or grain, Crystal, Crystallite and Domain size
Particle or grain, Crystal, Crystallite and Domain size
• A particle, or grain, may consist of one or more crystals which are fused together. The size of such a particle is generally not accessible with XRD, but with

light microcopy, for smaller sizes, scanning electron microscopy (SEM) is also an option. The Fusion of several crystals into a particle usually involves large

angle boundaries.

• The crystal size is also not accessible by XRD, unless the crystal is a single crystal (consisting of only a single domain). Possible size determination methods are

generally the same as for particle size determination.

• In the general case, a crystal may consist of several crystallites, which are fused by small angle boundaries (mosaic crystal). This means that, except for a

certain mismatch in the orientation, the crystallites show some degree of orientational relation to each other. The mismatch, however, is the reason why

crystallites diffract separately in XRD. Depending on the boundary conditions, crystallite size can be a quantity accessible with XRD.

• A crystallite may be subdivided into two or more domains by two-dimensional defects, like twin boundaries or stacking faults. The nature of such defects

causes the domains within a crystallite to have a rather well defined relation to each other concerning their crystal lattice. Whether or not domains within a

crystallite diffract coherently with each other depends on the nature of the boundary and its relation to the diffracting crystal planes. Within these

limitations, domain size is thus accessible with XRD.

• Crystallite and domain structure, and the defects separating domains within a crystallite, can only be directly investigated with high resolution transmission

electron microscopy (HRTEM).

Diffraction pattren
 Thermo Scientific ARL EQUINOX 3000 Series
X-Ray Diffractometer Powder Diffraction

X-ray source
Generator: 3500 W (60kV / 60mA - Option 30kV limitation)
Standard sealed X-ray tube : Cu radiation
Detector
Curved Position Sensitive X-ray Detector, CPS120
Curvature radius: 250mm
Acquisition in real time over 120° 2Theta

Goniometer
No moving parts, Acquisition in asymmetric mode,
Sample holders
Fixed, with or without sample rotation, Thin Film Assembly
ARL EQUINOX 3000 Goniometer

Asymmetric Acquisition Mode:

• That Means for fixed theta incidence of sample, you detect all the
diffraction peaks on the 2theta detector.
• For a fixed sample theta incidence, you detect all diffraction peaks
simultaneously on the detector within 2theta.
• So no, detector, sample or X-ray source movement is required
ARL EQUINOX 3000 Sample Holder
You can use XRD to determine

• Unit cell lattice parameters and Crystal Structure

Index peak positions
Lattice parameters can vary as a function of, and therefore give you information about, alloying,
doping, solid, strains, etc.
By Rietveld refinement of the entire diffraction pattern
• Crystallite Size and Microstrain
Indicated by peak broadening
Other defects (stacking faults, etc.) can be measured by analysis of peak shapes and peak width
• Phase Composition of a Sample
Quantitative Phase Analysis: determine the relative amounts of phases in a mixture by referencing the
relative peak intensities
Information of an Idealized Diffraction Pattern
Many factors may contribute to the observed peak profile

• Instrumental Peak Profile

• Crystallite Size
• Microstrain
• Non-uniform Lattice Distortions
• Faulting
• Dislocations
• Antiphase Domain Boundaries
• Grain Surface Relaxation
• Solid Solution Inhomogeneity
• Temperature Factors

• The peak profile is a convolution of the profiles from all of these contributions
A single crystal specimen would produce only one family of peaks in the diffraction pattern.

2q

At 20.6 °2q, Bragg’s law
fulfilled for the (100) planes,
producing a diffraction peak.
The (110) planes would diffract at 29.3 The (200) planes are parallel to the (100) planes.
°2q; however, they are not properly Therefore, they also diffract for this crystal. Since
aligned to produce a diffraction peak (the d200 is ½ d100, they appear at 42 °2q.
perpendicular to those planes does not
bisect the incident and diffracted beams).
Only background is observed.
 A polycrystalline sample should contain thousands of crystallites. Therefore, all possible
diffraction peaks should be observed.

2q 2q 2q

• For every set of planes, there will be a small percentage of crystallites that are properly oriented to diffract (the plane
perpendicular bisects the incident and diffracted beams).
• Basic assumptions of powder diffraction are that for every set of planes there is an equal number of crystallites that will
diffract and that there is a statistically relevant number of crystallites, not just one or two.
Peak Shape Analysis
The net broadening is the “sum” of all sources of broadening B ( FWHM ) = Bi Å Bc Å Bs Å BSF Å ...

XRD Line Broadening

• Unresolved a1 , a2 peaks.
Instrumental Bi • Non-monochromaticity of the source
(finite width of a peak).
• Imperfect focusing, etc.
Crystallite size BC

Strain BS • ‘Residual Strain’ arising from dislocations,

coherent precipitates etc. leading to broadening
Stacking fault BSF
§ In principle every extended defect
contributes to some broadening.
Other defects § Localized defects (e.g. isolated
point defects) cause diffuse
scattering.
 Counts

Debye Scherrer Formula Width=0.007 rad

XS ~ 19 nm
1000
Width=0.002 rad
XS> 90 nm
"#
Size =
$∗&'( ) 500

0
23 24 25 26
Position [°2Theta] (Copper (Cu))

• Apart from the shape of the crystallite and the crystallographic direction, the shape of the diffraction peak profile,

the definition of the profile width and the size distribution of the crystallites all influence the Scherrer constant. For

example, Langford & Wilson tabulated four different shape factors per lattice plane, corresponding to four different

definitions of peak width, for cubic, tetrahedral and octahedral crystallite shapes

• Due to this complexity, the observation that the Scherrer constant is usually close to unity, and the fact that in

most cases the actual crystallite shape is either unknown or ill defined, it has become common practice to use an

"overall average" shape factor, which is typically around 0.9.

Crystallite size analysis: Summary

• There are dozens of methods for crystallite size determination by XRD, which will all yield different results.
Even when using the same method on the same material, different choice of calibration or pre-assumptions
may yield different results.
• Comparison of results from different groups or people is impossible in most cases.
• Results should be published with a detailed description of the determination procedure if the absolute
crystallite size values are meant to be used for external comparison.
• Use a constant procedure for a series of samples and rely on the relative trends seen in the results rather than
believe in absolute values!
Strain Broadening

Strain effect: dislocations, vacancies, interstitials, substitutionals, and similar defects are imperfections which generate
a distribution in the d-spacings of lattice planes and thus result in peak broadening.
Strain Analysis
• Lattice strain or microstrain is a local deviation of d-spacings from the average value, caused by local defects.
• The origin may be point defects like interstitial or missing atoms, or substitution with different atoms.
• Since such defects have a size differing from that of the regular atoms in the lattice, the d-spacings are contracted or
expanded accordingly around the defect.
• One-dimensional defects, like screw dislocations, or two-dimensional defects like twin boundaries may also cause
microstrain,
e.g. if the atoms at the boundary need to be slightly displaced in order to match the lattices of both domains.
• Typically, the forces of compression and expansion will be balanced, and thus result in diffraction peak broadening
(variation of d-spacings translates into variation of diffraction angle), instead of a peak shift.
• This is in contrast to macrostrain, or stress, which is directed an thus causes peak shifts, depending on the relation between
the diffraction vector and the strain vector.
• While both size and strain broadening affect the width of diffraction peaks, they show a different angular dependence
of the peak width.
• The integral breadth of the strain-related diffraction peak broadening is proportional to tanθ.
Strain Analysis Methods
Williamson-Hall Plot
1. Fourier methods like Warren-Averbach analysis
deliver mean-square strain values.
2. A relatively simple and thus still widely used
method is the Williamson-Hall Plot. [G. K.
Williamson, W. H. Hall, "X-ray line broadening from
filed aluminiumand wolfram", Acta Metall. 1 (1953)
22-31]
3. A more sophisticated method which can be directly
integrated into Rietveld software is the Double-
Voigt Approach.[D. Balzar, "Voigt-function model in
diffraction line-broadening analysis" in: Defect and
Microstructure Analysis by Diffraction, edited by R.
Snyder, J. Fiala , H. J. Bunge, (1999) 94-126.
IUCr/Oxford University Press]
Phase Composition of a Sample
• Databases such as the Powder Diffraction File (PDF) contain dI lists for thousands of crystalline phases.
The PDF contains over 200,000 diffraction patterns.
• Computer programs can help you determine what phases are present in your sample by quickly
comparing your diffraction data to all of the patterns in the database.
• The PDF card for an entry contains a lot of useful information, including literature references.
• With high quality data, you can determine how much of each phase is present
• The ratio of peak intensities varies linearly as a function of weight fractions for any two phases in a
mixture
• The reference patterns are represented by sticks
• The position and intensity of the reference sticks should
match the data
• A small amount of mismatch in peak position
and intensity is acceptable experimental error

The X-ray diffraction pattern is a fingerprint that lets you figure out what is in your sample.
Phase identification (qualitative analysis)
There are various ways to perform a phase identification. Usually, the diffraction pattern is matched against reference
patterns from a database: diffraction pattern: either the measured data (may require background subtraction) or a peak list
(positions and intensities) derived thereof peak list generation:
peak search (position and intensity at maximum height of the peak) or peak fitting (position and intensity more accurate)
database: commercial (PDF = powder diffraction file, maintained by the ICDD) manual or automatic

How many peaks must match between a reference PDF pattern and a measured
diffractogram?
Generally, ALL peaks found in a PDF pattern must also be seen in in the diffractogram,
otherwise it is not a valid match.
Possible exceptions:
a) Small peaks may be not detectable if the noise level is too high.
b) Missing peaks may be the result of a very strong preferred orientation effect.
®If this is the case, the relative intensities in general are changed and show a systematic
dependence from hkl.
c) “Missing” peaks may be the result of anisotropic disorder.
®If this is the case, the FWHM of the peaks should show a systematic dependence from hkl
(some reflections become so broad and low that they are not recognized anymore).
Phase Composition of a Sample

• The pattern shown above contains equal amounts of TiO2 and Al2O3
• The TiO2 pattern is more intense because TiO2 diffracts X-rays more efficiently

With proper calibration, you can calculate the amount of each phase present in the sample
The weight fraction of each phase can be calculated if the calibration
constant is known
Counts

TiO2, Rutile 49.4 %

Fe2O3, Hematite 28.7 %
3600 TiO2, Anatase 21.9 %

1600

400

0
25 30 35 40 45
Position [°2Theta] (Copper (Cu))

The calibration constants can be determined:

By empirical measurements from known standards
By calculating them from published reference intensity ratio (RIR) values
Example Iron Oxide
Common Questions
Q1. All peaks of a PDF pattern are in the measured data, but the measurement contains additional
peaks. What does this mean?

Answer: The identification is probably correct, but the measured pattern represents a phase
mixture. Keep the reference pattern you found, then continue searching for references to explain
the additional peaks. Proceed until all peaks are explained.

Q2. All peaks in my measurement are explained with a PDF reference, except for some very small
ones. I tried to identify them as an impurity, but failed. What could this mean?

A: Two possible reasons are:

a) The peaks are artifacts resulting from spectral impurities (other wavelengths, e.g. Kb, W L). Thus,
they are weak “duplicates” of very strong peaks.
b) The peaks are real, but they belong to the reference compound, not an impurity. It may be that your
diffraction pattern is “better” in terms of signal/noise ratio than the (possibly old) PDF pattern. After
all, the diffractometers have improved with time.
®Use an hkl or Rietveld fit to check if the positions of the additional peaks matches the unit cell
parameters of the reference compound.
Common Questions
Q3: All peaks of my diffraction pattern are explained with the reference PDF card of my target compound. The peak positions
and intensities agree very well. Does this mean that my sample is pure?
Answer: No. Your sample may or may not be pure. There could be impurities which are invisible to XRD, either because the
amount is below the detection limit, or because they are “XRD amorphous”.

Q4: How many peaks must match between a reference PDF pattern and a measured
diffractogram?
Answer: Generally, ALL peaks found in a PDF pattern must also be seen in in the diffractogram, otherwise it is not a
valid match.
Possible exceptions:
a) Small peaks may be not detectable if the noise level is too high.
b) Missing peaks may be the result of a very strong preferred orientation effect.
®If this is the case, the relative intensities in general are changed and show a systematic dependence from hkl.
c) “Missing” peaks may be the result of anisotropic disorder.
®If this is the case, the FWHM of the peaks should show a systematic dependence from hkl (some reflections become
so broad and low that they are not recognized anymore).
Sources of Error in XRD Data
Background
Background can be produced from a number of sources. These include:
1. Fluorescent radiation emitted by the specimen
2. Diffraction of a continuous spectrum of wavelengths
3. Diffraction scattering from materials other than the specimen including soller slits, specimen binder, sample mount, and air.
4. Diffuse scattering from the specimen itself, including:
a. Incoherent (Compton) scattering, which increases when light elements are present.
b. Coherent scattering: Diffuse scattering due to various crystallite imperfections.
c. Extremely short-range ordered material, like glass, cause intense diffuse scatter.
d. Low intensity maxima contribute to background when the number of unit repeats normal to the diffracting plane is small.
e. Temperature (diffuse scattering is general small unless soft materials are involved)

• Penetration of the beam into the sample. (Sample transparency). The more intense the incident radiation, the farther the beam
penetrates into the sample. This creates a situation similar to flat specimen error as well as penetration into the substrate holding the
sample. In the latter case, diffraction from the substrate (off the focusing circle) will contribute to the pattern.
• Zero alignment: The source, the sample, and the detector must all be perfectly in line at 0° θ.
• X-ray source - Gaussian distribution for Kα1 radiation. Control choices for the operator (you) are kV and mA settings of the
generator.
Other sources of error

Flat specimen error

The entire surface of a flat specimen cannot lie on the focusing circle
Creates asymmetric broadening toward low 2theta angles
Reduced by small divergence slits; eliminated by parallel-beam optics
Poor counting statistics
The sample is not made up of thousands of randomly oriented crystallites, as assumed by most analysis techniques
The sample might be textured or have preferred orientation
Creates a systematic error in peak intensities
Some peaks might be entirely absent
The sample might have large grain sizes Produces ‘random’ peak intensities and/or spotty diffraction peaks
Softwares
PANalytical HighScore Plus
Match

Free Softwares
GSAS- Rietveld refinement of crystal structures
FullProf- Rietveld refinement of crystal structures
Rietan- Rietveld refinement of crystal structures

PowderCell- crystal visualization and simulated diffraction

patterns
JCryst- stereograms