Automated Centering and Tomography of Crystalline Samples

Crystal centring algorithms developed within BIOXHIT, XREC and C3D, are now implemented at many Partner sites. In addition to work on the individual optical centring algorithms the developers have been collaborating on testing joint procedures. Further advancements occurred on the hardware side, including better optical illumination in the visible range and crystal detection via UV illumination and fluorescence.

One example is a joint combination of XREC and C3D employed by Partner 3 (SRS) which was successfully transferred to beamlines at Partner 17 (Diamond). The implementation is shown in Figure 3.5.a. Background image is stored within the GUI at the start of the experiment and is subsequently subtracted form the crystal image. The automatic crystal centring and the movement of x, y and z is carried out based on weighted average of C3D and XREC output.

Figure 3.5.a. Implementation of crystal centring in Generic Data Acquisition

Figure 3.5.b. Crystal centring by X-rays using entropy

Figure 3.5.c. Crystal shape reconstruction by X-ray tomography

In some cases where optical crystal centring fails (particularly for small crystals) it may be advantageous to use X-rays for centring using entropy values. A direct approach was used from ADSC detectors directly with modified C programs from ADSC. In this case total number of spots were used to plot against movement of x, y and z of the goniometer, Figure 3.5.b.

We have also developed a proof-of-principle of crystal shape determination by both via optical and X-ray tomography, Figure 3.5.c. A detailed study on performing tomographic experiments on traditional MX station and the feasibility is assessed of the 3D reconstruction of the separated objects of the crystalline sample, the solvent around and the sample holder has been submitted for publication. The data gathered on this way allows optimisation during centring to avoid cracks and other crystal defects, as well as optimising the orientation of the absorbing volume using multi axis goniometers. Further possibilities for more sophisticated processing of diffraction data with accurate absorption correction is also highlighted.