Ptychographic X-ray computed tomography at a high-brilliance X-ray source

First X-ray spectral ptychography and resonant ptychographic computed tomography experiments at the SWING beamline from Synchrotron SOLEIL

X-ray ptychography and ptychographic computed tomography have seen a rapid rise since the advent of fourth-generation synchrotrons with a high degree of coherent radiation. In addition to quantitative multiscale structural analysis, ptychography with spectral capabilities has been developed, allowing for spatial-localized multiscale structural and spectral information of samples. The SWING beamline of Synchrotron SOLEIL has recently developed a nanoprobe setup where the endstation's first spectral and resonant ptychographic measurements have been successfully conducted. A metallic nickel wire sample was measured using 2D spectral ptychography in XANES mode and resonant ptychographic tomography. From the 2D spectral ptychography measurements, the spectra of the components of the sample's complex-valued refractive index, δ and β, were extracted, integrated along the sample thickness. By performing resonance ptychographic tomography at two photon energies, 3D maps of the refractive index decrement, δ, were obtained at the Ni K-edge energy and another energy above the edge. These maps allowed the detection of impurities in the Ni wire. The significance of accounting for the atomic scattering factor is demonstrated in the calculation of electron density near a resonance through the use of the δ values. These results indicate that at the SWING beamline it is possible to conduct state-of-the-art spectral and resonant ptychography experiments using the nanoprobe setup.

Integration of Fourier ptychography with machine learning: an alternative scheme

As the core task of the reconstruction in conventional ptychography (CP) and Fourier ptychographic microscopy (FPM), the meticulous design of ptychographical iterative engine (PIE) largely affects the performance of reconstruction algorithms. Compared to traditional PIE algorithms, the paradigm of combining with machine learning to cross a local optimum has recently achieved significant progress. Nevertheless, existing designed engines still suffer drawbacks such as excessive hyper-parameters, heavy tuning work and lack of compatibility, which greatly limit their practical applications. In this work, we present a complete set of alternative schemes comprised of a kind of new perspective, a uniform design template, and a fusion framework, to naturally integrate Fourier ptychography (FP) with machine learning concepts. The new perspective, Dynamic Physics, is taken as the preferred tool to analyze a path (algorithm) at the physical level; the uniform design template, T-FP, clarifies the physical significance and optimization part in a path; the fusion framework follows two workable guidelines that are specially designed to keep convergence and make later localized modification for a new path, and further establishes a link between FP iterations and the gradient update in machine learning. Our scheme is compatible with both traditional FP paths and machine learning concepts. By combining ideas in both fields, we offer two design examples, MaFP and AdamFP. Results for both simulations and experiments show that designed algorithms following our scheme obtain better, faster (converge at the early stage after a few iterations) and more stable recovery with only minimal tuning hyper-parameters, demonstrating the effectiveness and superiority of our scheme.

Enhancing Multi-Distance Phase Retrieval via Unequal Interval Measurements

In the conventional methods of multi-distance phase retrieval, the diffraction intensity patterns are recorded at equal intervals, which can induce slow convergence or stagnation in the subsequent reconstruction process. To solve this problem, a measurement method with unequal intervals is proposed in this paper. The interval spacings between adjacent measurement planes are decreased gradually. A large gap accelerates retrieval progress, and a short span helps to recover detailed information. The proposed approach makes full use of the available measured dataset and simultaneously generates variations in diversity amplitude, which is a crucial issue for the techniques of multi-image phase retrieval. Both computational simulations and experiments are performed. The results demonstrate that this method can improve the convergence speed by 2 to 3 times and enhance the quality of reconstruction results in comparison to that of the conventional methods.

First ptychographic X-ray computed tomography experiment on the NanoMAX beamline

Documentation is presented for the first ptychographic X-ray computed tomography experiment on the NanoMAX beamline, along with a quantitative analysis of the reconstruction quality and a discussion of possibilities for future improvements.

Ptychographic X-ray computed tomography is a quantitative three-dimensional imaging technique offered to users of multiple synchrotron radiation sources. Its dependence on the coherent fraction of the available X-ray beam makes it perfectly suited to diffraction-limited storage rings. Although MAX IV is the first, and so far only, operating fourth-generation synchrotron light source, none of its experimental stations is currently set up to offer this technique to its users. The first ptychographic X-ray computed tomography experiment has therefore been performed on the NanoMAX beamline. From the results, information was gained about the current limitations of the experimental setup and where attention should be focused for improvement. The extracted parameters in terms of scanning speed, size of the imaged volume and achieved resolutions should provide a baseline for future users designing nano-tomography experiments on the NanoMAX beamline.

Spectroscopic imaging with single acquisition ptychography and a hyperspectral detector

We present a new method of single acquisition spectroscopic imaging with high spatial resolution. The technique is based on the combination of polychromatic synchrotron radiation and ptychographic imaging with a recently developed energy discriminating detector. We demonstrate the feasibility with a Ni-Cu test sample recorded at I13-1 of the Diamond Light Source, UK. The two elements can be clearly distinguished and the Ni absorption edge is identified. The results prove the feasibility of obtaining high-resolution structural and chemical images within a single acquisition using a polychromatic X-ray beam. The capability of resolving the absorption edge applies to a wide range of research areas, such as magnetic domains imaging and element specific investigations in biological, materials, and earth sciences. The method utilises the full available radiation spectrum and is therefore well suited for broadband radiation sources.

Accelerated single-beam multiple-intensity reconstruction using unordered propagations

In conventional multiple-plane phase retrieval method, the wave propagations that proceed in the same ordered sequence could slow down convergence or lead to stagnation. In this Letter, a novel, to the best of our knowledge, algorithmic technique to accelerate phase retrieval using an unordered sequence of propagations is demonstrated experimentally. The main advantage of the technique is the significant increase in the change in amplitude, a key and essential element for a successful iterative phase retrieval. For N planes, the number of possible ways to implement the sequence of propagations is increased by a factor of N!(N-1)!, providing diversity for the change in the amplitude. Compared to the conventional algorithm, the technique performed 2 times faster and reduced the number of needed intensity patterns for the test object wave used. The technique may be adopted in the other multiple intensity-based phase retrieval methods.