February_EDFA_Digital

edfas.org 15 ELECTRONIC DEV ICE FA I LURE ANALYSIS | VOLUME 24 NO . 1 processes that cannowbe studiedby electronmicroscopy and othermethods. As the analysis so far has only focused on employing these AI methods to the analysis of single images with BPFA, there is potential to extend the resolu- tion limits for imaging even further in the future as more images of different samples are included in a training data set and using learning algorithms to improve reconstruc- tion quality. In addition to increasing the efficiency of the algorithms for the analysis of particular image/scattering processes in electron microscopy, an overall workflow can be developed that will improve imaging capabilities across multiple techniques (this workflow and the appli- cation software is currently being developed by Nuxutra). Figure 5 shows how the incorporation of different imaging methods into the workflow can bring multiscale, hyper- spectral measurements into the training datasets and these complete analyses can then be used to improve the resolution of wide-ranging expensive, difficult to use and over-subscribed scientific instrumentation that is the bedrock of the development of new advanced materials. ACKNOWLEDGMENTS Thisworkwas performed in theAlbert CreweCentre for Electron Microscopy, a shared research facility fully sup- portedby theUniversityof Liverpool. Aportionof thiswork was also supported by a Laboratory Directed Research and Development program at the Pacific Northwest National Laboratory (PNNL). PNNL is operated by Battelle Memorial Institute for theU.S. Department of Energyunder Contract No. DE-AC05-76RL01830. Aspects of this work were also supported in part by the UK Faraday Institution (EP/S003053/1) through the FIRG013 “characteriza- tion” project. This work was also funded by the EPSRC Centre for Doctoral Training in Distributed Algorithms (EP/S023445/1), Sivananthan Labs and Nuxutra. REFERENCES 1. S.J. Pennycook and P. D. Nellist: Scanning Transmission Electron Microscopy – Imaging and Analysis, Springer, 2011. 2. P.E. Batson, N. Dellby and O.L. Krivanek: “Sub-angstrom Resolution using AberrationCorrectedElectronOptics,” Nature, 2002, 418 (6898), p. 617-620. 3. R.F. Egerton: “Control of Radiation Damage in the TEM,” Ultra- microscopy, 2013, 127, p. 100. 4. A. Stevens, et al.: “The Potential for BayesianCompressive Sensing to Significantly Reduce ElectronDose inHigh-resolution STEM Images,” Microscopy, 2014, 63, p. 41-51. 5. L. Kovarik, et al.: “Implementing an Accurate and Rapid Sparse Sampling Approach for Low-Dose Atomic Resolution STEM Imaging of Electron Beam Sensitive Materials,” Applied Physics Letters, 2016, 109, p. 164102. 6. A. Stevens, et al.: “Sub-sampled Approaches for Extremely Low-Dose Scanning TransmissionElectronMicroscopy,” AppliedPhysics Letters, 2018, 112, p. 043104. 7. D. Nicholls, et al.: “Minimising Damage in High Resolution Scanning Transmission Electron Microscope Images of Nanoscale Structures and Processes,” Nanoscale, 2020, 12, p. 21248-21254. 8. D. Nicholls, et al.: “Sub-Sampled Imaging for STEM: Maximising Image Speed, Resolution andPrecision Through Reconstruction Parameter Refinement,” submitted to Ultramicroscopy. 9. R.G. Baraniuk: Compressive Sensing [Lecture Notes], IEEE Signal Processing Magazine, 2007, 24, p. 118-121. Fig. 5 Thephysics of beaminteractions canbe incorporated into sampling strategies formanyexperimentalmethods, optimizing the use of eachmethod individually and enhancing the scientific information obtained from themethods used together to solve materials challenges.