TL;DR
This paper introduces a flexible, open-source framework for accurately fitting the surface shapes of X-ray mirrors, crucial for high-performance synchrotron and free-electron laser facilities. By providing a standardized approach to defining mirror geometries and optimizing their fit to measurement data, this framework aims to improve X-ray optics characterization and fabrication.
What Problem Does This Paper Address?
X-ray mirrors are essential components in synchrotron radiation (SR) and free-electron laser (FEL) facilities, precisely focusing and directing intense X-ray beams for a wide range of scientific experiments. The accuracy of these mirrors' surface shapes directly impacts the quality of the X-ray beam, influencing its size, flux, and background noise. Consequently, precise characterization of mirror surfaces is critical during fabrication and before installation to ensure optimal performance and avoid wavefront aberrations.
Existing methods for characterizing these surfaces often rely on specialized, facility-specific software that may not be openly available or easily adaptable to different mirror geometries and metrology instruments. This creates a need for a universal, extensible, and user-friendly framework that can accurately fit measured surface data (slope or height) to theoretical mirror shapes, accommodating various geometries and measurement conditions.
Methods
The authors propose a four-layer framework for fitting X-ray mirror surface shapes:
- Standard Shape Definitions: This layer defines mathematical expressions for common grazing-incidence X-ray mirror shapes, including elliptic cylinders, hyperbolic cylinders, ellipsoids, hyperboloids, and a newly introduced sagittal collimating diaboloid. These expressions are provided in two conventions for flexibility.
- Surface Generation with Pose Parameters: This layer generates theoretical mirror surfaces by incorporating the mirror's "pose" – its position and orientation relative to the metrology instrument. This is achieved using six degrees of freedom: three translation parameters (chief ray intersection) and three rotation angles (roll, pitch, yaw).
- Selective Least Squares Optimization: This is the core fitting engine. It uses a nonlinear least squares method to minimize the difference between the measured data and the generated theoretical surface. Crucially, users can select which parameters (shape parameters and/or pose parameters) are optimized and which are held constant, allowing for tailored analysis.
- Function Wrappers: This top layer provides user-friendly functions that abstract the complexity of the underlying layers, allowing users to easily fit data to specific mirror shapes (e.g., a concave elliptic cylinder for Kirkpatrick-Baez mirrors).
The fitting process is iterative. It starts with an initial guess for the mirror's pose and shape parameters, generates a theoretical surface, calculates the error between the theoretical and measured data, and then updates the parameters to minimize this error until a convergence criterion is met.
Datasets
The paper demonstrates the framework's effectiveness using several representative real measurement datasets of various X-ray mirror types, including:
- 1D slope data of a concave elliptic cylinder.
- 1D height data of a concave elliptic cylinder.
- 2D height map of a concave hyperbolic cylinder.
- 2D height map of a concave ellipsoid.
- 2D height map of a concave hyperboloid.
- 2D height map of a simulated tangential collimating diaboloid.
These datasets cover common mirror geometries and measurement types (slope and height, 1D and 2D), showcasing the framework's versatility.
Key Insights and Ideas
- Unified Framework: The four-layer structure provides a systematic and modular approach to mirror surface fitting, making it adaptable to new mirror shapes and metrology techniques.
- Pose Parameter Importance: Accurately accounting for the mirror's pose (position and orientation) is critical for precise surface fitting, especially when dealing with real-world measurement data that may not perfectly align with the ideal coordinate system.
- Parameter Selection Flexibility: The ability to choose which parameters to optimize is a significant advantage. This allows users to focus on specific aspects of the mirror's shape or to account for known constraints, leading to more meaningful and practical fitting results. For instance, optimizing only certain pose parameters can reveal subtle "twist-like" errors.
- Open-Source Contribution: Making the framework open-source (MATLAB and Python) is a key contribution, fostering collaboration, transparency, and further development within the X-ray optics community.
- Diaboloid Surface: The introduction of the sagittal collimating diaboloid as a standard shape expands the repertoire of available theoretical models for X-ray optics.
Key Results
The paper demonstrates the framework's effectiveness through several fitting examples. In each case, the framework successfully fits the measured data to the theoretical mirror shapes, achieving low residual errors (e.g., 0.143 µrad RMS for slope fitting, 0.37 nm RMS for height fitting). The examples highlight how different parameter optimization strategies can lead to improved fitting results and reveal subtle surface errors. For instance, including the grazing incidence angle in the optimization for a concave elliptic cylinder minimized coma-like residuals. The framework also successfully handles complex 2D shapes and simulated misalignments, showing its robustness.
Limitations and Future Work
- Weighting: The current framework does not incorporate weighting schemes that account for the non-uniform distribution of the beam footprint on the mirror surface. This could be important for mirror alignment and wavefront control with adaptive X-ray mirrors, as it could reduce the impact of figure errors at edges.
- Toroidal Mirrors: Toroidal mirrors, another common type of X-ray mirror, are not yet implemented in the current framework.
- Ex Situ Metrology: The framework is designed for ex situ mirror metrology data. Future work could extend it to in situ measurements or incorporate more advanced fitting techniques.
The authors suggest that the framework's modular design makes it readily extensible to more complex shapes and that future developments could address the noted limitations.
Summary updated