Type IMG data tree entries: 2-D Images

An IMG data tree entry is created whan a 2D diffraction image is imported into a GSAS-II project. Note that unlike every other type of entry in the GSAS-II data tree, images are not self-contained. The actual contents of the image is read from the original file whenever the image is needed, but all other settings utilized in image processing other than the image itself are stored inside the project. If the .gpx file is moved to another computer, if it contains images, the image files must also be moved.

A IMG data tree entry is always accompanied by four child tree entries, labeled Comments, Image Controls, Masks and Stress/Strain. There is also some information in the parent IMG entry.

When a 2D diffraction image (prefix IMG) or one of its children is selected from the data tree, the image is displayed as a color contoured picture. At a minimum, a small 'X' shows the currently defined location of the incident beam on the image plane (NB: it could be off to one side or the other depending on detector location, or even outside of the image). Optionally, the integration limits are shown as an interior arc, an outer arc and green (beginning $2\theta$ limit) and red (ending $2\theta$ limit) as well as “cake” section limits (dashed lines) connecting the two. These are set in Image Controls. There may be a frame mask (green) that encloses the area used for subsequent integration or other analysis. Red outlines and pixels indicated selected areas/pixels that are excluded (masked) from further use; the settings for these are found in Masks.

Parent IMG data tree entry

The data window for the main IMG data tree entry shows some controls, most of which are rarely modified (e.g., pixel dimensions) and are usually obtained via the image importer from either the image header or a metadata file associated with the image. It is the user's/instrument scientist's responsibility to ensure the accuracy of these values. When they are wrong, the image cannot be used.

What can I do here?

Apart from changing one of the listed values (caution: you do need to know what their true value is), you can determine the x-ray bean polarization and create a gain map for your detector. Both require an image with scattering from a purely isotropic amorphous sample (a glass slide mounted perpendicular to the incident beam is recommended) with the detector close to the sample so that the scattering angle at the edge of the detector is at least 35° $2\theta$; better is > 40° $2\theta$. A frame mask is recommended to remove detector edge effects. The image should be as free as possible (except for beam stop) from shadows and obstructions and normal to the incident beam. The detector orientation should have been previously calibrated with a known reference material (e.g., LaB₆ or Si).

Calibrate? (Polarization) - This begins the calibration procedure (Von Dreele & Xi, 2021) for the x-ray beam polarization, which integrates a 4° $2\theta$ wide ring sampling area with and without an arc mask positioned about 90° azimuth (top of image) with selected polarization values. The integrations match with the correct polarization value. You will be asked for a $2\theta$ position for the sampling mask; choose a value at least 2° $2\theta$ less than the maximum $2\theta$ seen for all edges inside the frame mask. The process takes about 5 min to complete, so be patient.
Make gain map - This uses the same image (from a glass slide) used for polarization analysis to determine a gain map for the detector. The process uses the result of an integration of the glass pattern to normalize the entire detector pixel array. The result (~1.0 for all pixels) is the scaled by 1000, converted to integers and stored as a GSAS-II image file (NB: this is a python pickle file and thus not usable by other programs) and entered in the GSAS-II data tree. You can view it to see what the map looks like (select its IMG entry). The gain map file can be imported into other projects using the same detector. If selected in Image Controls, the image is immediately corrected for the gain map.

Comments

This window shows whatever comment lines found in a “metadata” file when the image data file was read by GSAS-II. If you are lucky, there will be useful information here (e.g., sample name, date collected, wavelength used, etc.). If not, this window will be blank. The text is read-only.

Image Controls

This window displays calibration values needed to convert pixel locations to two-theta and azimuth. Also shown are controls that determine how integration is performed.

Three sets of menu commands are associated with the Image Controls tree item. The first, Calibration, provides commands to perform calibration from an image (where the calibration values are fitted from a diffraction pattern image taken with a calibrant). The Integration menu provides the ability to radially integrate an image to provide one or more powder diffraction patterns (requires that calibration parameters must be set first.)

Finally, the Parms menu has commands allow the values on the window to be saved to a file, read from a file or copied to other images. The "Xfer Angles" menu command scales the current integration range for other images located at different detector distances.

Note that calibration parameters used in GSAS-II are closely related to those used by the pyFAI and Fit2D programs, but add 90 to the GSAS-II tilt plane rotation (labeled as "Rotation" in GSAS-II) to obtain the pyFAI value. The X and Y values determine the beam center location. In GSAS-II the values are in mm measured from the top left corner of the detector, while in pyFAI the values are measured in the same way, but in units of pixels:

\[X_{GSAS-II}/size_X = X_{pyFAI}$$ and $$ Y_{GSAS-II}/size_Y = Y_{pyFAI}\]

For Fit2D, the center is also in mm, but measured from the bottom left, so

\[X_{GSAS-II}/size_X = X_{Fit2D}$$ and $$ Y_{max} - Y_{GSAS-II}/size_Y = Y_{Fit2D}\]

where $Y_{max}$ is the detector size in pixels and $size_X$ and $size_Y$ are the pixel size in mm along the appropriate direction.

Masks

Image masks are used designate areas of an image with irregular in that should not be included in the integration, typically used due to detector irregularities, shadows of the beam stop, single-crystal peaks from a mounting, etc. Masks can be created with a menu command or with keyboard/mouse shortcuts. There are six types of masks:

Spot masks: exclude a circle with a selected center and diameter in image coordinates (mm).
Ring masks: excludes a specific Bragg reflection (a ring placed relative to the image center). The location and thickness of the ring are specified in degrees 2-theta.
Arc masks: exclude a section of a Bragg reflection, similar to a ring mask, except that in addition to the location and thickness of the ring, the mask has a starting and ending azimuthal angle.
Polygon masks: exclude an arbitrary region created by line segments joining a series of points specified in image coordinates (mm). Pixels inside the polygon mask are not used for integration.
Frame mask: excludes an arbitrary region created by line segments joining a series of points specified in image coordinates (mm). Typically, a point is placed near each corner of the image. Only pixels inside the frame mask are used for integration. Only one frame mask can be defined.
Pixel mask: excludes individual pixels that are found in a pixel mask search, which looks for pixels that have intensities that are significantly higher or lower than the median intensity for all pixels in a ring having the same two-theta value. Controls for the pixel mask search are found here and include the threshold for exclusion, which is expressed as a multiple of the standard deviation for the two-theta ring; the minimum and maximum two-theta value to be used in the search; and if the search should replace any previous pixel map search(es) or should be added to the previous pixel mask search. Finally, a button allows a previous pixel mask to be cleared.

What can I do here?

Masks of each type are created using the appropriate menu commands and then clicking as described in the section on "What can I do with the plot?" below, or by using keyboard shortcuts, also described in that section.

What is plotted here?

The image is shown, as described above. Note that the frame mask, if defined, is displayed in green, while the other types of masks are shown in red.

What can I do with the plot?

There are menu commands to create masks as well as keyboard shortcuts. If a menu command is used, then use left and right mouse clicks as described below. Note that on a Mac with a one-button mouse, some alternate actions must be used.

Spot masks: excludes inside a circular region of the image.
- Create Spot masks after a menu command by clicking on the location on the image that should be masked. There are also two ways to create spot masks with the keyboard:
  - Press the 's' key and then left-click successively on multiple locations for spot masks. Press the 's' key again or right-click to stop adding spot masks.
  - Alternately, move the mouse to the position for a new spot mask and press the 't' key. (Note that this can be used while the plot is in Zoom or Pan mode.)
  The default size for newly-created spot masks is determined by the Spot_mask_diameter configuration variable or 1.0 mm, if not specified.
- Edit Spot mask location by left-clicking inside or on the edge the of the mask and then drag the spot mask to a new location.
- Edit Spot mask radius by right-clicking inside the mask and then dragging to change the mask size.
Ring masks: excludes inside a ring of selected width that follows constant $2\theta$ as determined by the calibration (e.g. a Bragg diffraction ring).
- Create Ring masks with a menu command and then by left-clicking on the mask center; Or, by pressing the 'r' key and then left-clicking. (Right-click to cancel.)
  
  The default thickness for newly-created ring masks is determined by the Ring_mask_thickness configuration variable or 0.1 degrees ($2\theta$) if not specified.
- Edit Ring mask location by left-clicking on either the inner or outer circle and drag the circle to the new radius.
- Edit Ring mask thickness by right-clicking either on the inner or outer circle and drag the the circle change spacing between the inner and outer circle.
Arc masks: excludes inside an arc of constant $2\theta$, similar to a ring mask, except that in addition to the location and thickness of the ring, the mask has a starting and ending azimuthal angle.
- Create Arc masks with a menu command and then by left-clicking on at the mask center; Or, by pressing the 'a' key and then left-clicking. (Right-click to cancel.)
  
  The default size for newly-created ring masks is determined by configuration variables:
  - thickness: Ring_mask_thickness (0.1 degrees $2\theta$ if not specified) and
  - azimuthal range: Arc_mask_azimuth (10.0 degrees if not specified.)
- Edit Arc mask location by left-clicking on either the inner or outer circle and drag the circle to the new radius. Alternately, left-click on the upper or lower arc limit (the straight lines) and drag them to rotate the center of the arc azimuthal range to a new position.
- Edit Arc mask thickness or range by right-clicking either on the inner or outer circle and drag the the circle change spacing between the inner and outer circle. Alternately, right-click on the upper or lower arc limit (the straight lines) and drag them to change the arc azimuthal range.
Polygon masks: excludes inside an arbitrary region created by line segments joining a series of points specified in image coordinates (mm). Pixels inside the polygon mask are not used for integration.
- Create Polygon masks with a menu command and then by left-clicking successively on the vertices of the polygon shape surrounding pixels to be excluded. After the last point is defined, right-click anywhere to close the mask. Alternately, press the 'p' key and then left-click, as before, to define the mask and right-click anywhere to close the mask.
- Edit Polygon mask by left-clicking on any point at a vertex in the polygon mask and drag that point to a new position. If the vertex is dragged to the same position as any other vertex in the mask the dragged point is deleted.
- Add a point to Polygon mask by right-clicking on any vertex and dragging. A new point is added to the mask immediately after the selected point at the position where the mouse is released.
Frame mask: excludes outside an arbitrary region created by line segments joining a series of points specified in image coordinates (mm). Typically, a point is placed near each corner of the image. Only pixels inside the frame mask are used for integration. Only one frame mask can be defined.
- Create a Frame mask with a menu command and then by left-clicking successively on the vertices of a polygon. After the last point is defined, right-click anywhere to close the frame mask. Alternately, press the 'f' key and then left-click, as before, to define the mask and right-click anywhere to close the mask. Note that if a Frame mask already exists, using the 'f' key or the "Create Frame" menu item causes the existing frame mask to be deleted.
- Edit the Frame mask by left-clicking on any point at a vertex in the frame mask and drag that point to a new position. If the vertex is dragged to the same position as any other vertex in the mask the dragged point is deleted.
- Add a point to the Frame mask by right-clicking on any vertex and dragging. A new point is added to the mask immediately after the selected point at the position where the mouse is released.
Pixel mask: excludes individual pixels that are found in a pixel mask search, which looks for pixels that have intensities that are higher or lower by a threshhold than the median intensity for all pixels in a ring with the same two-theta value. By default this mask contains no pixels, but pixels are added using a search.
- Search to add to the pixel mask using either the Operations->"Pixel mask search" or the Operations->"Multi-IMG pixel mask search" menu commands. This searches for pixels that have intensities that are significantly higher or lower than the median intensity for all pixels in a ring having the same two-theta value. The threshold for exclusion is determined by the standard deviation for the pixels in the ring multiplied by a user-supplied value.

Stress/Strain

This allows one to evaluate strain typically induced by a pure axial load (e.g. no shear) on a polycrystalline sample (e.g. a steel bar). This strain will distort the Bragg diffraction rings seen by the 2D detector. This follows the method of He & Smith (Baoping Bob He & Kingsley Smith, (1997). Adv. in X-Ray Anal. 41, 501.) to determine the 3 unique terms of the axial strain tensor. One can examine the results as a series of diffraction line d-spacings vs azimuth angle; if no strain, these are straight, otherwise they will show a single sinusoidal variation with maxima at the maximum strain direction (90° & 270°) for a tension load. The signs are reversed for a compression load. One can also examine the local intensity variation as multiples of a random distribution (MRD) due to texture. Before embarking on this analysis be sure that your detector is carefully calibrated for orientation and position; you are looking for very slight variations in ring shape which may be biased by inadequate detector calibration. Commonly, the calibrant (typically $\rm CeO_2$) is painted on one sample surface (be sure to note if in front or back of sample!) and the sample ½ thickness is used in the Sample delta-z box (significant only for residual stress analysis).

What can I do here?

Menu Operations -

Append d-zero – this adds a d-spacing value to the stress strain ring list; this also can be done by picking a ring from the plot.
Fit stress/strain – this fits the three unique axial strain tensor elements (ε11, ε22 & ε12) for each ring to the local ring maxima at 1 mm intervals about each ring. Results display on console and table. The fitted d-zero is calculated from the mean d-spacings found for the ring & given as d-zero ave and can be compared to the d-zero value for any residual stress/strain.
Plot intensity distribution – this makes an MRD plot with azimuth for each ring.
Save intensity distribution – this saves the intensity distribution curves as a simple text file
Update d-zero – this updates the d-zero values with the d-zero ave ones thus removing any effect of residual stress/strain.
All image fit – this performs the Fit stress/strain operation for a selected sequence of images. NB: the stress/strain data should be copied to all other images before doing this. Results are reported in Sequential strain fit results.
Copy stress/strain – this copies the stress/strain data from the current image to other selected images in preparation for doing an All image fit.
Save stress/strain – this saves the current stress/strain data to a file with .strsta extension.
Load stress/strain – this loads previously saved stress/strain data from a file with .strsta extension.
Load sample data – this opens a metadata file containing sample load data.

What is plotted here?

The Strain plot shows the variation of d-spacing for each selected ring with azimuth angle. If there is no strain the points will scatter about a straight line. If there is strain, the points will describe a negative “cosine” curve if the sample is under tension or a positive “cosine” curve if the sample is under compression. If the fit has been done, a calculated curve will be shown along with a dashed black line for the fitted average d-spacing of the calculated curve. This average is either the mean (“Poisson mean” = False) or the Poisson mean which is ¼ or ¾ of the interval from d-min to d-max depending on the comparison between ε11 & ε22. Detector calibration errors will distort these curves.

The Ring intensities plot shows the local intensities as MRD taken at 1mm intervals about the circumference of each ring.

What can I do with the plot?

The Strain plot is best examined line by line by zooming in on each. The deviations are quite small and can not be discerned over the full d-spacing range. You should examine the calibration lines to ensure they are straight.

The Ring intensities plot will respond to the following key strokes:

Key 'l' – this progressively shifts the RMD lines to the left
Key 'r' – this progressively shifts the RMD lines to the right
Key 'u' – this progressively shifts the RMD lines up
Key 'd' – this progressively shifts the RMD lines down
Key 'o' – this resets the shifts to zero
Key 'g' – this toggles display of a grid on the plot
Key 's' – this saves all the plot data as a CSV file.