Periodic Reporting for period 1 - DIGIPHASE (Development of Maximum Efficiency Phase Contrast Electron Microscopy)
Reporting period: 2015-07-01 to 2017-06-30
Microscopy has helped enhance our understanding in many fields of science, advancing societies capabilities from technology to medicine. The goal of this project was to investigate new techniques in electron microscopy and employ microscopy to study novel materials. Many materials can handle only limited doses before being significantly altered or destroyed, and in such cases the practical resolving power of the microscope can depend as much on the signal to noise obtained before the sample is damaged as on the imaging optics. This project investigated the use of ptychography to extract information as efficiently as possible.
In ptychography a pixelated detector is used to record the details of the electron scattering at low angles, where most of the transmitted electrons can be found, in a scanning transmission electron microscope (STEM). At every probe position an image of this scattering is recorded, creating a four-dimensional dataset. This dataset is then processed digitally to determine the phase and amplitude of the spatial frequencies transmitted by the specimen. It is then possible to construct a phase contrast image by interfering all frequencies and transforming back to real space. Such ptychographic phase contrast imaging was first used to overcome the limitation of spherical aberrations in the electron optics before the advent of aberration correction in hardware. This was achieved by using a small portion of the signal in Fourier space where the aberrations in the interfering disks cancels out. Now however aberration correctors are available that allow atomic resolution at relatively low accelerating voltages. Therefore it is possible to rely on the hardware aberration correction and make use of as much of the signal as possible in Fourier space.
In ptychography a pixelated detector is used to record the details of the electron scattering at low angles, where most of the transmitted electrons can be found, in a scanning transmission electron microscope (STEM). At every probe position an image of this scattering is recorded, creating a four-dimensional dataset. This dataset is then processed digitally to determine the phase and amplitude of the spatial frequencies transmitted by the specimen. It is then possible to construct a phase contrast image by interfering all frequencies and transforming back to real space. Such ptychographic phase contrast imaging was first used to overcome the limitation of spherical aberrations in the electron optics before the advent of aberration correction in hardware. This was achieved by using a small portion of the signal in Fourier space where the aberrations in the interfering disks cancels out. Now however aberration correctors are available that allow atomic resolution at relatively low accelerating voltages. Therefore it is possible to rely on the hardware aberration correction and make use of as much of the signal as possible in Fourier space.
The single side band (SSB) ptychographic method developed by the Marie Curie fellow intelligently determines where in Fourier space information is transferred for each spatial frequency, and extracts phase and amplitude information from just these regions. This maximizes the amount of signal gathered while simultaneously rejecting noise by excluding regions of Fourier space that do not contain signal. Because the details of the bright field disk are recorded at each probe position, it is also very simple to synthesize the other modes of bright field imaging commonly used in STEM from the same datasets. From such comparisons it is clear that the ptychographic method is significantly more efficient than the other bright field imaging modes in STEM. However the question of how efficient the ptychographic technique is compared to HRTEM remained.
Therefore the fellow performed simulations of STEM ptychography and various modes of HRTEM. In an electron microscope instrumental instabilities such as variations in the accelerating voltage cause the electron beam to loose coherence and the focus to fluctuate. Simulations allow one to test the limits of imaging modes without the constraints of particular instrumentation and complete control over all factors affecting imaging including partial coherence. Comparing simulations of perfectly tuned and partially coherent HRTEM and STEM ptychography with and without finite doses, it became clear that the HRTEM was much more susceptible to the effects of partial coherence. This can be understood from the fact that lines exist in the 4D ptychographic dataset along which the effects of partial coherence cancels out. Because of these so called achromatic lines, partial coherence limits the resolution of the finite dose HRTEM images much more severely than the ptychographic images. Examples with graphene are shown in the figure.
Ptychography was also investigated as a tool for charge transfer sensitivity. The experimental ptychographic phase contrast images of single layer boron nitride collected by the fellow show the contrast expected for successful charge transfer sensitivity, and a publication is in preparation on the subject. Another development in which the Fellow participated during the project was the measurement correction of residual aberrations with ptychography. Residual aberrations affect the contrast in phase images, and when looking for such subtle changes in phase as those caused by charge transfer it can be important to rule these aberrations out or remove them. From these studies it became apparent that although ptychography can be performed with very few pixels in the detector, there is a tradeoff between detector speedup and the ability to measure and correct aberrations in the 4D data. Fortunately, detector technology is advancing rapidly and with the technology emerging today, it is no longer necessary to greatly reduce the resolution of the detector to achieve sufficiently rapid speeds.
In addition to ptychography, a method of imaging was developed that is extremely sensitive to the local stacking of atoms by measuring the center of mass of the medium to higher angle scattering with a pixelated detector. By digitally mapping the magnitude and phase of the center of mass it was possible to determine the structure of Graphene/hBN van der Waals Heterostructure in 3D. Another example of a novel detection sensitivity the fellow contributed to was the demonstration of atomic mass sensitivity in the electron microscope. Finally, the fellow's expertise in electron microscope contributed to the successful publication of 12 other publications, and more are set to follow.
Therefore the fellow performed simulations of STEM ptychography and various modes of HRTEM. In an electron microscope instrumental instabilities such as variations in the accelerating voltage cause the electron beam to loose coherence and the focus to fluctuate. Simulations allow one to test the limits of imaging modes without the constraints of particular instrumentation and complete control over all factors affecting imaging including partial coherence. Comparing simulations of perfectly tuned and partially coherent HRTEM and STEM ptychography with and without finite doses, it became clear that the HRTEM was much more susceptible to the effects of partial coherence. This can be understood from the fact that lines exist in the 4D ptychographic dataset along which the effects of partial coherence cancels out. Because of these so called achromatic lines, partial coherence limits the resolution of the finite dose HRTEM images much more severely than the ptychographic images. Examples with graphene are shown in the figure.
Ptychography was also investigated as a tool for charge transfer sensitivity. The experimental ptychographic phase contrast images of single layer boron nitride collected by the fellow show the contrast expected for successful charge transfer sensitivity, and a publication is in preparation on the subject. Another development in which the Fellow participated during the project was the measurement correction of residual aberrations with ptychography. Residual aberrations affect the contrast in phase images, and when looking for such subtle changes in phase as those caused by charge transfer it can be important to rule these aberrations out or remove them. From these studies it became apparent that although ptychography can be performed with very few pixels in the detector, there is a tradeoff between detector speedup and the ability to measure and correct aberrations in the 4D data. Fortunately, detector technology is advancing rapidly and with the technology emerging today, it is no longer necessary to greatly reduce the resolution of the detector to achieve sufficiently rapid speeds.
In addition to ptychography, a method of imaging was developed that is extremely sensitive to the local stacking of atoms by measuring the center of mass of the medium to higher angle scattering with a pixelated detector. By digitally mapping the magnitude and phase of the center of mass it was possible to determine the structure of Graphene/hBN van der Waals Heterostructure in 3D. Another example of a novel detection sensitivity the fellow contributed to was the demonstration of atomic mass sensitivity in the electron microscope. Finally, the fellow's expertise in electron microscope contributed to the successful publication of 12 other publications, and more are set to follow.
This combination of interpretability and dose-efficiency therefore offers a new route to imaging radiation sensitive specimens that promises to overcome present barriers to revealing the atomic structures of fragile beam sensitive materials and molecules. A paper has been written on these findings and is being submitted to Physical Review Letters. The low-dose imaging community is a large and diverse group working in many fields. In molecular biology low dose imaging is used for instance to enhance our understanding of the workings of the body and infections. Another example of low dose imaging is to better understand energy materials such as Li based batteries. Enhanced imaging of beam sensitive materials will enhance knowledge in all these fields, from medicine to our ability to produce and efficiently make use of energy.