A Fully Automated Small Molecule Single Crystal X-Ray Diffraction System Mark E Light * & Michael B Hursthouse, Department of Chemistry, University of Southampton, UK Diffraction Unit Cell Success Strategy Data Collection Data Process System Y PreScans Yes No BruNo Mount BruNo Unmount Setup via GUI Sample Tray No SampleTray GUI BruNo Decision Making Programming BruNo is a Mitsubishi Movemaster RV-1A 6 axis industrial robot with a CR1 controller. It possesses an onboard program detailing the positions of the samples and the location of the goniometer head. At present it enables the automatic collection of 24 samples from loading of pre-mounted crystals to full structure solution and refinement (see poster 0791). The system can be run in either unit cell or full data collection mode to carry out general sample screening, rapid data collection, polymorph screening and day to day hands free data collection. The heart of the system is it’s automated decision making process. This relies on the analysis of a number of prescan images to determine the diffraction properties of the sample. These are used to set appropriate parameters for the unit cell and data collection procedures or to reject samples in cases of poor diffraction. The advice of Rob Hooft (Bruker Nonius) has been invaluable during this project. He has made possible the full use of the collect python modules and provided help with many other programming issues. The system is programmed in python to allow the essential incorporation of the collect 1 python modules. This makes access possible to the goniometer positioning functions enabling synchronisation of the robot and diffractometer during sample mounting. In addition, through the unique nature of the collect 1 suite, manipulation of the core data collection modules is possible. This has greatly facilitated the building of a fully automated data collection program. Thus it has been possible to replace all user inputs with intelligently calculated automatic responses. In addition all graphical output from collect 1 has been suppressed. Frank Jin (Bruker Nonius) developed the initial BruNo control program and has helped with it’s subsequent development. He also provided advice on the robots basic operations. The experiment is driven from our in-house SampleTray GUI. This enables sample selection and input of metadata including sample name and formula etc. It also allows configuration of the experiment type and operational parameters. 1 "Collect" data collection software, Nonius B.V., The fine tuning of the decision making parameters is an on going process which will improve with operational experience. However, it is possible to select parameter ranges within which the decisions are made. This not only aids system improvement but enables some manual control over the way the experiment behaves. We have plans to introduce a scheduling system to intelligently queue samples in order of diffraction quality. This would maximise throughput within a limited time frame. At present the sample tray holds 24 samples but it would be a simple task to increase this number with the introduction of a basic rack system to make operation almost continuous! CCECCE Comb-e-chem The system has just reached the first stages of operational testing. The reliability of BruNo has been explored over a period of 24 hours continual sample changing, and the automatic data collection procedure has proved to be reliable over a range of crystal qualities. The system as a whole has collected a handful of samples from start to finish and is about to enter an extended period of full testing. We have also developed a direct interface to BruNo to enable a range of simple procedures such as sample mounting and basic diffractometer positioning. Through serial and network connections the GUI manages control flow between the robot, diffractometer and the automated data collection program. Data reduction is spawned as a separate process to reduce dead time in the system and the data is placed in a central store for access by System-Y (see poster 0791). The robot is marketed with proprietary software 2 enabling access to the on board program functions and details of the robot axes positions. It’s use is necessary during recalibration of the robot target positions. 2 Melfa Personal Computer Software, Mitsubishi, 1999.