The Oxford group has also been developing a system to scan a collimated X-ray source to provide reference trajectories through multiple layers of silicon sensors that would in turn allow sensors in different layers to be surveyed/aligned with respect to one another after they had been assembled into the final tracker but before installation of the tracker inside the ATLAS experiment.
William showed a few diagrams of how such a system could be configured to provide an array of precisely located x-ray beams in 3 dimensions and briefly described what he understood of the system. More information on the Oxford plans can be found at their website . In particular William showed the first two figures that can be found on this page which show how the system designed to survey the SCT could be deployed to survey the pixel detector.
Discussion of the benefits of being able survey the pixel detector with such an x-ray system versus the costs it would incur followed. Clearly we should learn more about the precision that could be expected from such a system and whether it could be used to set overall radial scale of the pixel detector. This should be compared to the planned construction accuracy of the pixel system including optical survey of fiducials and touch-probe coordinate measurements that are expected of the different sub-assemblies of the pixel detector.
If such studies showed that significant cross-checks could be made and additional information could be provided for an initial alignment database then the focus should turn to the compatibility of the x-ray sources being studied by Oxford with the pixel detector readout. These studies could be started with single chip pixel assemblies and a simplified x-ray source.
It was agreed that William would get more techinical information about the capabilities of the proposed Oxford x-ray system and that we would try to convene a conference/video call in mid-June with the LBL pixel detector assembly engineers to compare x-ray capabilities with anticipated construction tolerances.
It was stated that even if plans to do a full 3-d survey of the pixel detector turned out to be prohibitively costly (in terms of time or extra services to operate a significant fraction of the pixel detector outside the ATLAS cavern) it might still make sense to have a simple 2-d system to survey modules on the front/back faces of the disk sectors that are going to be assembled in North America.
For more information on the status of the use of an x-ray alignment system for the ATLAS pixel detector contact william@physics.utoronto.ca.