masonlw

Home \ Lauren Mason

About Lauren Mason

This author has not yet filled in any details.
So far Lauren Mason has created 247 blog entries.

Encoding information using optical imaging systems in the AI era

Abstract: The advent of artificial intelligence, particularly deep neural networks, is transforming the traditional criteria used to evaluate imaging systems. Increasingly, algorithms are used to process captured data, yielding outputs with little perceptual resemblance to its original form. This change broadens the scope of design possibilities while also creating a new challenge to define what constitutes a “good” imaging device when human interpretability is no longer a requirement. In this talk, we address this challenge using a probabilistic modeling approach, which can be used to quantify the information content of images captured under various conditions. We apply this framework to various imaging modalities, including label-free LED-array microscopy, lensless cameras, and single-shot 3D fluorescence microscopy, showing how it can directly evaluate performance without the need for labor-intensive post-processing algorithm development. This framework provides a new set of design principles uniquely tailored for the AI era.

Speaker Bio: I am a part-time postdoctoral researcher at UC Berkeley in the Computational Imaging Lab with Prof. Laura Waller in the Department of Electrical Engineering and Computer Sciences, where I received my PhD in Computational Biology and MS in Electrical Engineering and Computer Sciences. I am also the founder of Photomics, Inc., where I create open-source software for microscope control and label-free, computational microscopes. My research is focused on the design of hardware, software, and algorithms for data and information-driven design and control of optical microscopes. This work draws from many fields, including optical physics, machine learning, single-cell biology, immunology, software engineering, data science, computer vision, and information theory. More details can be found in the research section.

Imaging with multimode fiber endoscopes

Abstract:  In-vivo imaging through multimode fibers has been recently accomplished. Multimode fibers are attractive for endoscopic applications due to their thin cross-section, a large number of degrees of freedom, and flexibility. However modal dispersion and intermodal coupling preclude direct image transmission. The development of fast spatial phase control enables focus scanning and structured illumination for different novel imaging modalities. We discuss the implications of these techniques for ultrathin optical endoscopy.

Speaker Bio: Prof. Rafael Piestun received the Ingeniero Electricista degree from the Universidad de la República (Uruguay) and MSc. and Ph.D. degrees in Electrical Engineering from the Technion – Israel. From 1998 to 2000 he was a researcher at Stanford University. Since 2001 he has been at the University of Colorado Boulder where is a professor in the department of Electrical and Computer Engineering and in the Physics department. He is a fellow of the Optical Society of America, was a Fulbright scholar, an Eshkol fellow, received a Honda Initiation Grant award, a Minerva award, a Provost Achievement Award, and El-Op and Gutwirth prizes. He was associate editor of Optics and Photonics News and Applied Optics. He is founder of the company Double Helix Optics (SPIE Prism Award, First Place in the Luminate Competition) and the company Modendo Inc. His areas of interest include computational optical imaging, superresolution microscopy, volumetric photonic devices, scattering optics, and ultrafast optics.

‘Doughnut’ beams help physicists see incredibly small objects

In a new study, researchers at CU Boulder have used doughnut-shaped beams of light to take detailed images of objects too tiny to view with traditional microscopes.

The new technique could help scientists improve the inner workings of a range of “nanoelectronics,” including the miniature semiconductors in computer chips. The discovery was highlighted Dec. 1 in a special issue of Optics & Photonics News called Optics in 2023.

The research is the latest advance in the field of ptychography, a difficult to pronounce (the “p” is silent) but powerful technique for viewing very small things. Unlike traditional microscopes, ptychography tools don’t directly view small objects. Instead, they shine lasers at a target, then measure how the light scatters away—a bit like the microscopic equivalent of making shadow puppets on a wall.

So far, the approach has worked remarkably well, with one major exception, said study senior author and Distinguished Professor of physics Margaret Murnane.

“Until recently, it has completely failed for highly periodic samples, or objects with a regularly repeating pattern,” said Murnane, fellow at JILA, a joint research institute of CU Boulder and the National Institute of Standards and Technology (NIST). “It’s a problem because that includes a lot of nanoelectronics.”

4D Scanning Transmission Electron Microscopy for Multimodal and Multiscale Materials Characterization

Abstract: Physical properties of matter depend on structure across vastly disparate length scales, from well below the atomic to macroscopic.  In this talk, we’ll discuss scale-bridging scanning transmission electron microscopy (STEM) experiments and the algorithms used to quantify them, measuring quantities from picometer deformations of individual atomic columns in charge density wave materials under in-situ cryogenic cooling, to grain orientations of hundreds of crystallites in a single capture, to lattice parameter variations measured across the many micron lengths of LixFePO4 nanoplatelets in several stages of electrochemical cycling. Many of these datasets are large, and integrating computation and experiment is necessary in each case.  In atom tracking with high-angle annular dark-field (HAADF)-STEM, instabilities and bubbling from the cryogen can easily spoil in-situ measurements – by combining many fast-acquisition low-signal image captures with a registration algorithm tailored to nearly uniform lattices, measuring and visualizing ~pm lattice displacements in low-temperature CDW phases is possible.  In 4D-STEM, in which a 2D image of the diffracted electrons is collected at each position of the 2D beam raster, matching algorithms to experiment remains essential to make sense of the large and information rich datasets.  Examples will be selected to highlight a range of modalities, methodologies, and applications, and will include Bragg localization, amorphous/crystal classification, phase identification, automated crystal orientation mapping, and others.

Speaker Bio: Ben Savitzky is a postdoctoral scientist at the National Center for Electron Microscopy in Berkeley CA.  He created, maintains, and leads development of py4DSTEM, a Python software package for 4D-STEM data analysis.  He completed his PhD with Lena Kourkoutis and Cornell University in 2018.

High-Fidelity Ptychography of Highly Periodic Structures

Lensless imaging based on ptychographic coherent diffractive imaging enables diffraction-limited microscopy at short wavelengths, overcoming the limits of imperfect optics.1,2 Ptychographic imaging of highly periodic structures has been challenging, however, due to the lack of diversity in the recorded diffraction patterns, which leads to poor convergence of the reconstructed sample images. Although techniques (such as modulus enforced probe and total variation regularization) have been explored to address this challenge, they suffer from slow convergence, heavy reliance on constraints on the samples, or both. This significantly limits ptychography’s application to a wide variety of periodic structures in photonics, nanoelectronics and extreme ultraviolet (EUV) photomasks.

Ab initio structures from nanocrystal molecular lattices

Electron diffraction has dramatically increased in popularity amongst chemists given its renewed application for ab initio structure determination from molecular nanocrystals. In one implementation, popularly referred to as 3D ED or MicroED, crystals nanocrystals orders of magnitude too small for conventional X-ray analysis are interrogated by an electron beam to determine atomic structures. However, these approaches are thwarted by disordered, overlapping, or otherwise poorly diffracting domains.

Spatially resolved diffraction mapping techniques can overcome some of these limitations, and have seen limited application in X-ray diffraction. In electron microscopy, such approaches, including 4D scanning electron microscopy (4D-STEM), have grown popular. We demonstrated that 4D-STEM can be used to determine ab initio structures of molecules by direct methods, from small ordered nanodomains of single microcrystals. In our approach 4D-STEM is used to generate diffraction scans that enable ex post facto reconstruction of digitally defined virtual apertures. The synthetic patterns derived from these scans are suitable for direct methods phasing of molecular structures.

In addition, this approach unveils that coherently diffracting zones (CDZs) in molecular crystals form unpredictably distributed striations. The observation of these zones and our ability to determine structures from these regions of nanocrystals empowers us to explore their atomic substructure and their response to radiolytic damage.

Operando Spectral Imaging of the Li-ion Battery’s Solid-Electrolyte Interphase

Considering the scale of the lithium ion battery (LIB)  industry, it is surprising how poorly the function of LIBs is understood at the molecular level. While much is certainly known, this knowledge has been gained via inference and expensive trial-and-error because it is difficult to look inside a functioning LIB to “see” what is going on. The battery is a bulk device with a liquid, air-sensitive organic electrolyte. With use, there forms on the LIB electrodes an almost magical solid-electrolyte interphase (SEI) that is an insulator for electrons but a conductor for Li+ ions. The main mysteries of LIB function involve the chemical composition and structure of this layer. We present the first images of the LIB SEI acquired under room-temperature operando conditions with high spatial and spectroscopic resolution. This combination gives us an unprecedented view of the SEI’s development, where we can make chemical identifications localized to nanometer precision while the electrode is in the very act of intercalating. We image the bulk SEI, not just its surface, by contriving electrochemical fluid cells that are only 50 nm thick. With these thin cells we can map the Li itself by its unique spectroscopic fingerprint, an achievement described as “practically impossible” just a few years ago.

Accurate quantification of lattice temperature dynamics from ultrafast electron diffraction of single-crystal films using dynamical scattering simulations

In ultrafast electron diffraction (UED) experiments, accurate retrieval of time-resolved structural parameters, such as atomic coordinates and thermal displacement parameters, requires an accurate scattering model. In this article, we demonstrated dynamical scattering models that are suitable for matching ultrafast electron diffraction (UED) signals from single-crystal films and retrieving the lattice temperature dynamics. We first described the computational approaches used, including both a multislice and a Bloch wave method, and introduced adaptations to account for key physical parameters. We then illustrated the role of dynamical scattering in UED of single-crystal films by comparing static and temperature-dependent diffraction signals calculated using kinematical and dynamical models for gold films of varying thicknesses and rippling as well as varying electron probe energy. Lastly, we applied these models to analyze relativistic UED measurements of single-crystal gold films recorded at the High Repetition-rate Electron Scattering (HiRES) beamline of Lawrence Berkeley National Laboratory. Our results showed the importance of a dynamical scattering theory for quantitative analysis of UED and demonstrated models that can be practically applied to single-crystal materials and heterostructures.

Deep-Learning Electron Diffractive Imaging

Coherent diffractive imaging (CDI) is revolutionizing the physical and biological science fields by first measuring the diffraction patterns of nano-crystals or non-crystalline samples and then inverting them to high-resolution images. The well-known phase problem is solved by the combination of coherent illumination and iterative computational algorithms. In particular, ptychography – a powerful scanning CDI method – has found wide applications with synchrotron radiation, high harmonic generation, electron and optical microscopy. However, iterative algorithms are not only computationally expensive, but also require practitioners to get algorithmic training to optimize the parameters and obtain satisfactory results. These difficulties  have thus far prevented CDI from being accessible to an even broader user community. Here we demonstrated deep learning CDI using convolutional neural networks (CNNs) trained only by simulated data. The CNNs are subsequently used to directly retrieve the phase images of monolayer graphene, twisted hexagonal boron nitride and a Au nanoparticle from experimental electron diffraction patterns without any iteration. Quantitative analysis shows that the phase images recovered by the CNNs have comparable quality to those reconstructed by a conventional iterative method and the resolution of the phase images by the CNNs is in the range of 0.71-0.53 Å. Looking forward, we expect that deep learning CDI could become an important tool for real-time, atomic-scale imaging of a wide range of samples across different disciplines.    

Go to Top