Powder XRD
Systematic Errors
Types of Systematic Error
• Flat Specimen Error
• Sample Displacement
• Axial Divergence
• Anomalous Peaks
– Kβ peaks
– Tungsten peaks
– Peaks from novel phases
• Kβ – Filter Asymmetry
Let’s Recall…
Remember Bragg’s Law and the geometry upon which it depends!
Diffraction Geometry
In a typical diffractometer configuration, the stage is stationary while the X-ray source
(tube) and the detector move while the sample stage remains stationary (a θ-θ
instrument). Some instruments, such as the AXRD Benchtop, are θ-2θ instruments, in
which the stage moves through a range of θ, the detector through 2θ, and the X-ray
source remains stationary.
Diffraction Geometry (cont’d)
Regardless of this geometry, there is
always a goniometer circle (which is a
fixed circle defined by the X-ray source
and the receiving slit) and a focusing
circle.
The focusing circle, like the goniometer,
is also defined by the source and the
receiving slit, but is additionally tangent
at all steps in 2θ to the surface of the
sample.
Therefore, as you scan through a range
of 2θ, the radius of the focusing circle
decreases.
Flat Specimen Error
In an ideal diffraction configuration, the
entirety of the sample is tangent to the
focusing circle, meaning that a perfect
diffraction experiment requires a curved
sample whose curvature changes with each
angular step as the radius of the focusing
circle changes. In practice, this is not possible
to do.
In a typical diffraction experiment, the sample
is mounted on a flat surface, meaning that
only one point of the sample is tangent to the
focusing circle. This results in small errors in
peak position related to divergence of the
beam (α) and the diffraction angle (θ):
Δ2θ = − 1ൗ 2 cot θ
6α
However, this error is relatively minor, and is
only significant at angles below 1˚.
Reference Surface
Sample Displacement
In many diffraction experiments,
the most significant error
Positive displacement (above affecting peak positions is
the reference surface)
caused by sample displacement,
i.e. the sample is too high or too
low, and so is not perfectly
tangent to the focusing circle.
No displacement
This error is often the result of
poor sample packing, but is also
an expected source of error in
non-pack mounts such as slurry
Negative displacement or top dusted mounts where the
(below the reference surface) sample is, by definition, above
the focusing circle.
Sample Displacement (cont’d)
This error is related to
the angular step (θ)
and diffractometer
radius (R):
Δ2θ = −2𝑠 cos θ /𝑅
Note that the effect
The peak shift shown above (as a of this error is
function of angle) is caused by a 50 greatest at low 2θ!
μm positive displacement.
Axial Divergence
Recall that axial divergence is caused
by the X-ray beam spreading out along
the diffractometer axis, on both the
incident and diffracted beam sides.
This causes errors in both peak
position and peak width.
Soller slits control this divergence by
blocking X-rays that are not parallel to
the series of openings in the soller slit
(grating), collimating the beam.
This error is related to the angular
aperture of the soller slits (Δ) and the
diffraction angle (θ):
The peak shift shown above (as a function of
Δ2θ = −1ൗ 2 cot 2θ
6Δ angle) is observed with a soller slit aperture of
Keep in mind, however, that the error 2.29˚. As the aperture angular width is decreased,
caused by axial divergence is complex,
and so the above equation is a close the error caused by axial divergence will decrease,
approximation. but intensity will also decrease.
Anomalous Peaks
Anomalous diffraction peaks are peaks that cannot be matched
to a known phase.
There are three main types of anomalous peaks:
–Kβ peaks
–Tungsten peaks
–Novel crystalline phase peaks
Identification of these peaks is not always easy, and requires
some familiarity with XRD patterns, the instrument being used,
and sample context.
Kβ Peaks
Recall that the production of X-rays results in characteristic radiation of multiple wavelengths. Typically, powder
diffractionists are only interested in one wavelength. Filters are used to remove the other wavelengths to
ensure that only one wavelength is being detected. If the mechanism for removal is faulty or forgotten, the
diffraction pattern may have peaks caused by Kβ radiation. These peaks can complicate an analysis by
appearing to be unidentifiable phases when in fact they are the result of diffraction of the wrong radiation.
In this example, a sample of hematite was analyzed with (blue) and without (red) a Kβ filter. In the red pattern,
Kβ peaks are clearly visible as they are the peaks that do not have corresponding peaks in the blue pattern.
Residual Kβ Peaks
As mentioned in previous slides, a β-filter is used to remove Kβ radiation when using a
Dectris Mythen detector. If your sample contains a highly-diffracting phase like quartz (as
seen below), the β-filter may not fully remove the Kβ radiation, resulting in small Kβ peaks.
The Ni filter may also cause asymmetry in very intense peaks as a result of incomplete
absorption of Bremsstrahlung radiation with a wavelength similar to that of Cu K α radiation.
Quartz (011), 2θα
β-filter effect
Quartz (011), 2θβ
Tungsten Peaks
Tungsten peaks are caused by tungsten atoms being vaporized off of the tungsten filament and deposited on the
anode. Recall that during an X-ray diffraction experiment, electrons given off by the cathode strike the atoms of the
anode, causing them to give off characteristic radiation, which causes diffraction and results in the peaks we observe. If
the anode is contaminated by tungsten from the cathode those tungsten atoms will also give off characteristic
(tungsten) radiation. This radiation can also cause diffraction, producing peaks that will not match known phases.
This is typically only a problem in older systems with older X-ray tubes, such as with the example below. This is a scan of
LaB6, a peak position standard, that shows anomalous tungsten peaks (black arrows).
Peaks from Novel Phases
If anomalous peaks cannot be explained as either Kβ peaks or tungsten peaks, it is possible that they may be explained
as representing a novel phase that has not previously been observed. Sometimes, observing novel phases is the goal of
the study!
The example below was obtained using a pure powder sample of the mineral blodite. ሷ The sample was exposed to sub-
freezing conditions (T ≤ -10˚ ) and a diffraction pattern collected. The diffraction pattern for Blodite
ሷ is well known, and
so the scientist was able to differentiate between blodite
ሷ and non- blodite
ሷ peaks. Those peaks therefore represent a
novel phase that had not yet been documented. But in order to say with confidence that these peaks represent a new
phase, the other potential causes of anomalous peaks must be ruled out.
In this figure,
the blue
pattern is
observed data,
the red is the
model. Black
lines below the
pattern
represent
blodite,
ሷ and
red lines
represent the
novel phase.
Final Review
There are a number of consistent factors that affect your data beyond the properties of the
material being analyzed. These factors, known as systematic errors, can cause apparent
changes in peak position, peak width, and peak symmetry, but they can also result in
“extra” peaks that, if improperly identified, can result in inaccurate phase identification.
The effects of some of these factors can be reduced with appropriate sample preparation,
such as sample displacement. Others can be mitigated by changing the instrument
configuration, such as filtering out Kβ radiation with a β–filter, controlling axial divergence
with soller slits, or by using a newer tube if your scans suggest tungsten contamination on
the anode.
Other potential issues, such as incomplete removal of Kβ radiation or the presence of β–
filter asymmetry, are difficult to eliminate completely. In these instances, you as the
scientist must be aware of these potential effects, and be on the lookout for them during
data analysis.