arXiv subjects

Joshua L. Vincent

Publications and source records attributed to Joshua L. Vincent.

8 recordsLinked to original sources

arXiv2021

The role of surface spin polarization on ceria-supported Pt nanoparticles

In this work, we employ first-principles simulations to investigate the spin polarization of CeO$_2$-(111) surface and its impact on interactions between a ceria support and Pt nanoparticles. For the first time, we report that the CeO$_2$-(111) surface exhibits a robust surface spin polarization due to the internal charge transfer between atomic Ce and O layers. In turn, it can lower the surface oxygen vacancy formation energy and enhance the oxide reducibility. We show that the inclusion of spin polarization can therefore significantly reduce the major activation barrier in the proposed reaction pathway of CO oxidation on ceria-supported Pt nanoparticles. For metal-support interactions, surface spin polarization enhances the bonding between Pt nanoparticle and ceria surface oxygen, while CO adsorption on Pt nanoparticles weakens the interfacial interaction regardless of spin polarization.

cond-mat.mtrl-sci↗

arXiv2021

Unsupervised Deep Video Denoising

Deep convolutional neural networks (CNNs) for video denoising are typically trained with supervision, assuming the availability of clean videos. However, in many applications, such as microscopy, noiseless videos are not available. To address this, we propose an Unsupervised Deep Video Denoiser (UDVD), a CNN architecture designed to be trained exclusively with noisy data. The performance of UDVD is comparable to the supervised state-of-the-art, even when trained only on a single short noisy video. We demonstrate the promise of our approach in real-world imaging applications by denoising raw video, fluorescence-microscopy and electron-microscopy data. In contrast to many current approaches to video denoising, UDVD does not require explicit motion compensation. This is advantageous because motion compensation is computationally expensive, and can be unreliable when the input data are noisy. A gradient-based analysis reveals that UDVD automatically adapts to local motion in the input noisy videos. Thus, the network learns to perform implicit motion compensation, even though it is only trained for denoising.

eess.IV↗

arXiv2021

Adaptive Denoising via GainTuning

Deep convolutional neural networks (CNNs) for image denoising are usually trained on large datasets. These models achieve the current state of the art, but they have difficulties generalizing when applied to data that deviate from the training distribution. Recent work has shown that it is possible to train denoisers on a single noisy image. These models adapt to the features of the test image, but their performance is limited by the small amount of information used to train them. Here we propose "GainTuning", in which CNN models pre-trained on large datasets are adaptively and selectively adjusted for individual test images. To avoid overfitting, GainTuning optimizes a single multiplicative scaling parameter (the "Gain") of each channel in the convolutional layers of the CNN. We show that GainTuning improves state-of-the-art CNNs on standard image-denoising benchmarks, boosting their denoising performance on nearly every image in a held-out test set. These adaptive improvements are even more substantial for test images differing systematically from the training data, either in noise level or image type. We illustrate the potential of adaptive denoising in a scientific application, in which a CNN is trained on synthetic data, and tested on real transmission-electron-microscope images. In contrast to the existing methodology, GainTuning is able to faithfully reconstruct the structure of catalytic nanoparticles from these data at extremely low signal-to-noise ratios.

cs.CV↗

arXiv2021

Deep Denoising For Scientific Discovery: A Case Study In Electron Microscopy

Denoising is a fundamental challenge in scientific imaging. Deep convolutional neural networks (CNNs) provide the current state of the art in denoising natural images, where they produce impressive results. However, their potential has barely been explored in the context of scientific imaging. Denoising CNNs are typically trained on real natural images artificially corrupted with simulated noise. In contrast, in scientific applications, noiseless ground-truth images are usually not available. To address this issue, we propose a simulation-based denoising (SBD) framework, in which CNNs are trained on simulated images. We test the framework on data obtained from transmission electron microscopy (TEM), an imaging technique with widespread applications in material science, biology, and medicine. SBD outperforms existing techniques by a wide margin on a simulated benchmark dataset, as well as on real data. Apart from the denoised images, SBD generates likelihood maps to visualize the agreement between the structure of the denoised image and the observed data. Our results reveal shortcomings of state-of-the-art denoising architectures, such as their small field-of-view: substantially increasing the field-of-view of the CNNs allows them to exploit non-local periodic patterns in the data, which is crucial at high noise levels. In addition, we analyze the generalization capability of SBD, demonstrating that the trained networks are robust to variations of imaging parameters and of the underlying signal structure. Finally, we release the first publicly available benchmark dataset of TEM images, containing 18,000 examples.

cs.CV↗

arXiv2021

An $\textit{In Situ}$ Light Illumination System for an Aberration-Corrected Environmental Transmission Electron Microscope

In this work, an optic fiber based $\textit{in situ}$ illumination system integrated into an aberration-corrected environmental transmission electron microscope (ETEM) is designed, built, characterized and applied. With this illumination system, the dynamic responses of photoactive materials to photons can be directly observed at the atomic level, and other stimuli including heating and various gases can also be applied simultaneously. Either a broadband light source or a high power laser source aiming to expedite photoreactions can be utilized, fitting different application needs. The optic fiber enters the ETEM through the objective aperture port, with a carefully designed curvature and a 30° cut at the tip to orient the emitted light upwards onto the TEM specimen. The intensity distributions striking the sample from the broadband and laser sources are both measured, and due to the non-uniform distributions, an alignment procedure has been developed to align the bright spot with the electron optical axis of the TEM. The imaging and spectroscopy performances of the ETEM are proved to be maintained after incorporating this illumination system. Furthermore, Langmuir evaporation is observed when in situ laser light is applied to GaAs, demonstrating the phenomenon of optical heating on suitable semiconductor materials.

cond-mat.mtrl-sci↗

arXiv2021

Fluxional Behavior at the Atomic Level and its Impact on Activity: CO Oxidation over CeO$_{2}$-supported Pt Catalysts

Reducible oxides are widely used catalyst supports that can increase oxidation reaction rates by transferring their lattice oxygen at the metal-support interface. The interfacial oxidation process is typically described in terms of a Mars-van Krevelen mechanism. However, many outstanding questions remain unanswered regarding the atomic-scale structure and dynamic meta-stability (i.e., fluxional behavior) of the interface $\textit{during catalysis}$. Here, we employ aberration-corrected $\textit{operando}$ electron microscopy to visualize the structural dynamics occurring at and near Pt/CeO$_{2}$ interfaces during CO oxidation. Finite element simulations are performed to develop a reaction rate analysis wherein the atomic-level structural observations are directly correlated with the catalyst's turnover frequency for CO oxidation. We show that the increasing frequency of catalytic turnover correlates with dynamic fluxional behavior that (a) destablizes the supported Pt particle, (b) marks an enhanced rate of oxygen vacancy creation and annihilation, and (c) leads to increased strain and reduction in the surface of the CeO$_{2}$ support. Overall, the results implicate the interfacial Pt-O-Ce bonds anchoring the Pt to the support as being involved also in the catalytically-driven oxygen transfer process, and they suggest that oxygen reduction takes place on the highly reduced nearby CeO2 surface before migrating to the interfacial perimeter for reaction with CO. The $\textit{operando}$ electron microscopy approach described here should be applicable to a large number of nanoparticle catalysts. This technique will enable the identification of catalytically functional surface structures and strengthen our ability to establish (dynamic) structure-activity relationships.

cond-mat.mtrl-sci↗

arXiv2021

Developing and Evaluating Deep Neural Network-based Denoising for Nanoparticle TEM Images with Ultra-low Signal-to-Noise

A deep convolutional neural network has been developed to denoise atomic-resolution TEM image datasets of nanoparticles acquired using direct electron counting detectors, for applications where the image signal is severely limited by shot noise. The network was applied to a model system of CeO2-supported Pt nanoparticles. We leverage multislice image simulations to generate a large and flexible dataset for training and testing the network. The proposed network outperforms state-of-the-art denoising methods by a significant margin both on simulated and experimental test data. Factors contributing to the performance are identified, including most importantly (a) the geometry of the images used during training and (b) the size of the network's receptive field. Through a gradient-based analysis, we investigate the mechanisms learned by the network to denoise experimental images. This shows that the network exploits global and local information in the noisy measurements, for example, by adapting its filtering approach when it encounters atomic-level defects at the nanoparticle surface. Extensive analysis has been done to characterize the network's ability to correctly predict the exact atomic structure at the nanoparticle surface. Finally, we develop an approach based on the log-likelihood ratio test that provides a quantitative measure of the agreement between the noisy observation and the atomic-level structure in the network-denoised image.

cond-mat.mtrl-sci↗

arXiv2020

Chemical Kinetics for Operando Electron Microscopy of Catalysts: 3D Modeling of Gas and Temperature Distributions During Catalytic Reactions

$\textit{In situ}$ environmental transmission electron microscopy (ETEM) is a powerful tool for observing structural modifications taking place in heterogeneous catalysts under reaction conditions. However, to strengthen the link between catalyst structure and functionality, an $\textit{operando}$ measurement must be performed in which reaction kinetics and catalyst structure are simultaneously determined. To determine chemical kinetics for gas-phase catalysis, it is necessary to develop a reliable chemical engineering model to describe catalysis as well as heat and mass transport processes within the ETEM cell. Here, we establish a finite element model to determine the gas and temperature profiles during catalysis in an open-cell operando ETEM experiment. The model is applied to a $SiO_2$-supported Ru catalyst performing CO oxidation. Good agreement is achieved between simulated compositions and those measured experimentally across a temperature range of 25 - 350 °C. In general, for lower conversions, the simulations show that the temperature and gas are relatively homogeneous within the hot zone of the TEM holder where the catalyst is located. The uniformity of gas and temperature indicates that the ETEM reactor system behavior approximates that of a continuously stirred tank reactor. The large degree of gas-phase uniformity also allows one to estimate the catalytic conversion of reactants in the cell to within 10 percent using electron energy-loss spectroscopy. Moreover, the findings indicate that for reactant conversions below 30 percent, one can reliably evaluate the steady-state reaction rate of catalyst nanoparticles that are imaged on the TEM grid.

cond-mat.mtrl-sci↗