Files in this item



application/pdfMASHRAFI-DISSERTATION-2019.pdf (8MB)
(no description provided)PDF


Title:High-bandwidth high-precision robust x-ray microscopy - a control systems approach
Author(s):Mashrafi, Sheikh Tamjid
Director of Research:Salapaka, Srinivasa; Preissner, Curt
Doctoral Committee Chair(s):Salapaka, Srinivasa
Doctoral Committee Member(s):Hovakimyan, Naira; Voulgaris, Petros; Park, Hae-Won
Department / Program:Mechanical Sci & Engineering
Discipline:Mechanical Engineering
Degree Granting Institution:University of Illinois at Urbana-Champaign
Subject(s):Robust Control, Embedded Control, Optimal Control, Advanced Photon Source (APS), Argonne National Laboratory (ANL), Velociprobe, X-Ray Microscope, High Precision XYZ Scanning Stages, Nanopositioning Stage Dynamics, High Bandwidth, High Resolution, X-Ray Imaging Speed, X-Ray Image Spatial Resolution, Ptychography, Step Scan, Flyscan, Snake Scan, National Instruments, LabVIEW RT and FPGA, NI cRIO, Disturbance Rejection, Thermal and Sensor Drift Rejection
Abstract:In this thesis, a systematic framework for designing control for high-precision positioning stages of Velociprobe X-ray microscope at Advanced Photon Source (APS) at Argonne National Laboratory (ANL) is presented. In particular, our focus is on maintaining a precise position of the optics scanning stages in the XY lateral plane relative to the sample stages, which will ensure that the optics stages scans the focused X-ray spot on the sample along a predefined trajectory. We would also want to maintain a precise relative distance between the optics and sample in Z direction (direction of the X-ray beam) to make sure the X-ray spot size remains constant during a scan. Both precise positioning in lateral XY plane and constant relative displacement in beam direction would influence X-ray image spatial resolution and imaging bandwidth. Our framework facilitates control designs that achieve simultaneously specifications on tracking bandwidth and positioning resolution while guaranteeing robustness of the closed loop device to unmodeled uncertainties. To develop this framework, we used modern control techniques for modeling, quantifying design objectives and system-specific challenges, and designing the control laws. The control designs were implemented on a 3 degree of freedom piezo-actuated parallel kinematics stages dedicated for precision scanning of X-ray optics. Experimental results demonstrate significant improvements in positioning performance with H_∞ optimal controllers; for instance, improvements by over 134%, 149% and 132% in tracking bandwidths along X, Y, and Z stages, respectively, were demonstrated when compared to proportional-integral-derivative (PID) controller designs. Even with these high-bandwidth control designs, the positioning resolution of the order 1-2 nanometers were achieved, which is approximately the same as the PID controllers. Two different X-ray imaging technique, namely step scan and flyscan, were successfully carried out with the controllers. In the step scan technique, the optics stages tracked a typical raster scan pattern and successfully scanned the X-ray spot covering a 1 〖μm〗^2 area on the sample in 2.1 minutes, with NI control hardware and H_∞ control design. This resulted in 8 folds improvement in the imaging bandwidth compared to previously existing methods. In step scan technique, the X-ray spot is first positioned at point on the sample and corresponding diffraction pattern is recorded by the detector, then the X-ray spot is moved to new position in next step and imaging is continued. In contrast, we enabled the flyscan, where the optics stage continuously tracked a custom square snake scan pattern to scan the focused X-ray spot over a 1 〖μm〗^2 area of the sample in 0.01 secs while simultaneously recording the diffraction patterns at the area detector. Flyscan of 1 〖μm〗^2 area was done over 10^4 times faster than step scan with our control design and over 10^5 times faster than previous step scan performance at the APS beamline. In X-ray microscopy it is imperative that the relative position between the optics stage, that carries the X-ray focusing optics, and the sample stage follow a certain trajectory while either the optics or sample stage is being scanned. The state-of-the-art in X-ray microscopy at APS (as explained above) features an H_∞ control architecture applied to only the optics stage or both the optics and sample stage, achieving the objectives of large tracking bandwidth, good positioning resolution, rejection of environmental disturbance, attenuation of measurement noise, good X-ray diffraction image resolution and increased imaging bandwidth. However, the sensors and the fixtures that hold the sensors drift with time due to changing air temperature at the APS beamline. The drift of the sensor affects the lateral position of the zone plate focusing optics in the XY plane during scanning relative to the sample stage and the relative position between the optics and sample along Z direction. This results into imaging artifacts, image ambiguity and reduced image spatial resolution. Here, we identified this limiting factor and countered it by measuring the drift in real time and incorporated that in the optimal control architecture. We have shown that the effects of drift in the closed loop are practically removed. If our proposed method is adopted and applied to the X-ray microscope at APS beamline, it would significantly improve X-ray image spatial resolution and reduce imaging artifacts. We provide estimates of this improvements in this thesis.
Issue Date:2019-04-18
Rights Information:Copyright 2019 Sheikh Tamjid Mashrafi
Date Available in IDEALS:2019-08-23
Date Deposited:2019-05

This item appears in the following Collection(s)

Item Statistics