EBSD.com is an educational site by Oxford Instruments

Because the luminous intensity of the diffraction pattern on the phosphor screen is low, successive frames may need to be integrated in the CCD camera to improve the signal to noise ratio in the image.

Using too short an integration time on the CCD camera may give a poor image and in this case, integrating for longer on the CCD will improve the visibility of the diffraction pattern (Figure 1). The camera may be cooled to reduce electronic noise in the CCD when used in this way. The yield of backscattered electrons increases with atomic number, so low atomic number materials will require a longer integration time than higher atomic numbers.

Figure 1 Effect of integration time and probe current on diffraction pattern from Austenitic steel.

Effect of integration time and probe current on diffraction pattern. 36 ms, 2 nA (Probe currents are approximate)

(a) Effect of integration time and probe current on diffraction pattern. 36 ms, 2 nA (Probe currents are approximate).

Effect of integration time and probe current on diffraction pattern. 36 ms, 200 pA (Probe currents are approximate)

(b) Effect of integration time and probe current on diffraction pattern. 36 ms, 200 pA (Probe currents are approximate).

Effect of integration time and probe current on diffraction pattern. 360 ms, 200 pA (Probe currents are approximate)

(c) Effect of integration time and probe current on diffraction pattern. 360 ms, 200 pA (Probe currents are approximate).

The CCD camera resolution can also effect the integration time required to collect a diffraction pattern. Current CCD cameras can collect 12 bit images at a resolution of 1344 x 1024 pixels. Using “pixel binning”, neighbouring pixels in the CCD can be added together to form a single pixel in the image. When pixels are binned in this way, less integration time is required to achieve a given signal in the image pixel because the detecting pixel area is larger. For example, if a satisfactory diffraction pattern is obtained in 12 ms at 1344 x1024 resolution, a comparable pattern can be obtained in 3 ms at 672 x 512 resolution.

In addition, the transform works well on noisy images (Figure 2) so some experimentation is necessary to determine the optimum integration time required when collecting maps.

Figure 2 Diffraction patterns can be automatically solved in the presence of noise.

Diffraction pattern from Austenitic steel

(a) Diffraction pattern from Austenitic steel.

Indexed diffraction pattern

(b) Indexed diffraction pattern.

Noisy diffraction pattern.</p>
		</div>

(c) Noisy diffraction pattern.

Indexed noisy diffraction pattern

(d) Indexed noisy diffraction pattern.

Electrons of all energies scattered from the sample form a background to the diffraction pattern, which reduces the contrast of the Kikuchi bands.

The background intensity can be removed to improve the visibility of the Kikuchi bands. The background can be measured by scanning the beam over many grains in the sample to average out the diffraction information.

Figure 3 Background removal from a diffraction pattern of Austenitic steel.

Background

(a) Background.

Original pattern

(b) Original pattern.

Pattern after background subtraction

(c) Pattern after background subtraction.

The EBSD sample is usually tilted at approximately 70° relative to normal incidence of the electron beam to optimise both the contrast in the diffraction pattern and the fraction of electrons scattered from the sample.

For smaller tilt angles the contrast in the diffraction pattern decreases (Figure 8).

Figure 8 Effect of tilt on silicon diffraction pattern.

Sample tilt 70 degrees

(a) Sample tilt 70 degrees.

Sample tilt 60 degrees

(b) Sample tilt 60 degrees.

Sample tilt 50 degrees

(c) Sample tilt 50 degrees.

It is very important to understand the effect of varying the microscope operating conditions on the diffraction pattern.

Probe current

Increasing the probe current will increase the number of electrons contributing to the diffraction pattern and so allow the camera integration time to be reduced (Figure 1). However, this must be balanced with the spatial resolution required, because increasing the probe current will also increase the electron beam size.

Accelerating voltage

Increasing the accelerating voltage reduces the electron wavelength and hence reduces the width of the Kikuchi bands in the diffraction pattern (see equation 2). Also, because more energy is being deposited on the phosphor screen, this will result in a brighter pattern which requires a shorter integration time (Figure 4). Changing the accelerating voltage may require adjustment to the Hough transform filter size to ensure the Kikuchi bands are detected correctly. Higher accelerating voltages may be required to penetrate conducting layers, and lower accelerating voltages for restraining the beam to thin layers, or for charging samples.

Figure 4 Effect of changing accelerating voltage on diffraction patterns from Austenitic steel. Note that there is an effect on the bandwidth, sharpness and contrast.

Accelerating voltage 10 kV

(a) Accelerating voltage 10 kV.

Accelerating voltage 20 kV

(b) Accelerating voltage 20 kV.

Accelerating voltage 30 kV

(c) Accelerating voltage 30 kV.

Working distance and magnification

Because the sample is tilted, the SEM working distance will change as the beam position moves up or down the sample, and the image will go out of focus (Figure 5). The image will also be foreshortened because of the tilt and at low magnifications much of the field of view could be out of focus. Some EBSD systems can compensate for the image foreshortening by using different horizontal and vertical image beam steps and can adjust the SEM focus automatically as the beam is moved over the sample (Figure 5).

Figure 5 Tilt correction and focus maintenance.

Image without tilt or dynamic focus compensation

(a) Image without tilt or dynamic focus compensation.

Image with tilt compensation and no dynamic focus compensation

(b) Image with tilt compensation and no dynamic focus compensation.

Image with tilt and dynamic focus compensation

(c) Image with tilt and dynamic focus compensation.

In addition, movements of the beam will alter the pattern centre position on the phosphor screen and this can affect the EBSD system calibration (Figure 6) . EBSD systems can compensate automatically for shifts in the pattern centre by calibrating at two working distances and interpolating for intermediate working distance values. It is important to know the range of working distances for which the EBSD system will remain accurately calibrated.

Figure 6 The effect of changing working distance on pattern centre position.

Left: With a tilted sample, the pattern centre position will depend on the sample working distance. Middle: The top and bottom of the field of view may have a different working distance and hence pattern centre positions. Right: If the sample is moved, the working distance and hence pattern centre position will change

(a) Left: With a tilted sample, the pattern centre position will depend on the sample working distance. Middle: The top and bottom of the field of view may have a different working distance and hence pattern centre positions. Right: If the sample is moved, the working distance and hence pattern centre position will change.

The green cross shows the patteren centre with working distance 14.4mm

(b) The green cross shows the patteren centre with working distance 14.4mm.

The pattern centre moves down the screen as the working disance increases to 18.9mm

(c) The pattern centre moves down the screen as the working disance increases to 18.9mm.

The pattern centre moves down the screen as the working disance increases to 22.5mm

(d) The pattern centre moves down the screen as the working disance increases to 22.5mm.

Pressure

Diffraction patterns can also be collected from samples at low vacuum in environmental SEMs (Figure 7). This can be useful with specimens which may otherwise charge, such as ceramic or geological materials.

Figure 7 Effect of SEM vacuum on diffraction pattern from Silicon sample.

Pressure 0Pa

(a) Pressure 0Pa.

Pressure 10Pa

(b) Pressure 10Pa.

Pressure 70Pa

(c) Pressure 70Pa.

Pressure 130Pa

(d) Pressure 130Pa.

Go To Top