Figure 5
Ability to index the diffraction pattern as described in §3.1 , given an input beam position at various coordinates on the detector. The portion of the detector shown is the same as in Fig. 3 , with the true beam position shown as a black dot. The sample is the same as in Figs. 2 and 3 , with the low-angle spot-to-spot separation from the longest cell dimension being L = 1.5 mm. Red pixels are input beam positions that lead to correct refined beam parameters, unit-cell dimensions and Bravais symmetry. Positions where otherwise correct indexing gives monoclinic symmetry (instead of orthorhombic) are given in yellow. Incorrect indexing is shown in lavender, and white pixels indicate that no indexing was possible. The first two panels show the control where the input beam position is used directly for indexing, with Bragg reflections from either one (a) or two (b) images. Note that in the single image used for panel (a), the c axis is parallel to the incident beam. Consequently, two lattice angles are poorly determined, accounting for the preponderance of input beam positions giving monoclinic rather than orthorhombic symmetry. When two images are used (b) this ambiguity disappears. The lattice-like arrangement of lavender patches in panel (a) corresponds to the lattice of probable (but incorrect) beam positions in Fig. 3(c) . In the last two panels, the true beam position is predetermined by a grid search, based on one (c) or two (d) images. |