Microscopy Conference 2023

Session chairs

Prof. Dr.

Characterisation of a timepix detector for use in SEM acceleration voltage range

IM 3: SEM and FIB developments

IM 6: Phase-related techniques & 4D STEM

Abstract text (incl. figure legends and references) Introduction Hybrid pixel direct electron detectors are getting more popular in electron microscopy due to their excellent properties. These cameras can be affordable which makes them ideal tools for experimentation especially in combination with an SEM setup. To support this, a full detector characterisation of an Advacam Minipix detector system in the 15-30keV range was made. At low beam currents, detector layer behaviour should be dominant, which could make the findings equally applicable to Medipix and TImepix3 sensors provided the same detector layer is used. Objectives Determine modulation transfer function (MTF) and detective quantum efficiency (DQE) of the Advacam Minipix detector for acceleration voltages of 15 – 30 keV and different detector thresholds. Materials &amp; methods Advacam Minipix (TimePix) with 300 μm thick Si sensor 256x256 pixels of 55 μm size; Tescan Mira FEG SEM; 16 μm FIB aperture. The MTF is obtained via Fourier transform of the Point Spread Function (PSF) acquired with single spot illumination [1], mechanically scanning over individual pixels with a 16 μm FIB fabricated mask mounted on a piezo stage. The DQE is calculated from separate homogeneous illumination measurements of DQE(0) corrected with cluster size estimation [2], Noise Power Spectrum and MTF with same conditions [3]. Results Fig.1 and Fig.2 represent MTF and DQE obtained for beam energies of 15 – 30 keV and detector threshold settings 5 – 25 keV. It is seen that the MTF is quite close to ideal. MTF gets further from ideal with an increase of HT at constant threshold. Higher thresholds improves MTF but at the cost of reducing the number of detected electrons and thus a worse DQE. Figure 1. MTF of TimePix detector depending on Acceleration voltage (HT) and camera threshold. Figure 2. DQE of TimePix detector depending on Acceleration voltage (HT) and camera threshold. Conclusion We provide a full MTF and DQE characterisation of an Advacam Minipix detector for beam energies in the SEM range. We observe a near ideal MTF with only minor effects from clustering. In this sense the detector is very close to its theoretical limit. The DQE on the other hand remains always under 65% which we speculate to be due to a combination of backscattering and absorption in the front metal contact that is deposited on the detector layer. References [1] Tore Niermann et al. "A new linear transfer theory and characterization method for image detectors. Part II: Experiment", https://doi.org/10.1016/j.ultramic.2012.01.011 [2] D. Jannis et al. "Event driven 4D STEM acquisition with a Timepix3 detector: Microsecond dwell time and faster scans for high precision and low dose applications" https://doi.org/10.1016/j.ultramic.2021.113423 [3] G.McMullan et al. "Detective quantum efficiency of electron area detectors in electron microscopy", https://doi.org/10.1016/j.ultramic.2009.04.002 [4] The authors acknowledge the financial support of the Research Foundation Flanders (FWO, Belgium) project SBO S000121N. [5] The authors are grateful to Dr. Ivan Lobato for productive discussion of methods

Poster

A novel ion source for Lithium-ion microscopy and 3D tomography

Abstract text (incl. figure legends and references) Liquid Metal Alloy Ion Source (LMAIS) is an emerging Focused Ion Beam (FIB) source technology. It has been established for nanofabrication over the last years [1]. Most recently, ion imaging with Lithium ions from GaBiLi eutectic alloy has become feasible [2]. Multiple ion species such as Lithium and Bismuth are emitted simultaneously from a single source and are separated in a subsequent Wien filter. Different ion species, light or heavy and fast or slow, can be selected quickly without any mechanical sample or column adjustment [3]. Lithium is the lightest ion available from LMAIS. It features the smallest virtual source diameter and lowest energy spread [4]. Hence Lithium offers the smallest probe size of all ions emitted from eutectic alloys, resulting in sub 2nm image resolution Besides Lithium, also Gallium and Bismuth ions as well as Bi clusters are available from the same source. Bismuth ions and in particular Bi clusters have an excellent depth resolution for sample delayering. The combination of Bi and provides outstanding ion beam imaging capabilities. Switching between Li and Bi can be accomplished within a few seconds and enables access to different ion beams within the same FIB column. This setup allows stable imaging of 3D structures and sample reconstruction. With the FIB system presented (here), Ion beams are always perpendicular to the sample and no stage movement or sample tilt is needed resulting in a stable and reliable process for tomography and 3D sample reconstruction. In this contribution we describe the working principle and capabilities of LMAIS with a focus on GaBiLi. Results of 2D Lithium-ion microscopy and workflows for stable 3D tomography by Bismuth milling and Li imaging will be presented. Figure 1: Li ion image of Sn/C Figure 2: Li ion image of graphite Figure 3: Region of interest (RIO) for tomography. Box created by 80 Bi+ milling steps and in between Li+ ion imaging Figure 4: 3D reconstruction of semiconductor chip Stack of Li images from RIO (Figures 3) [1] J. Gierak et al, Journal of Vacuum Science &amp; Technology B 36, 06J101 (2018) [2] N. Klinger et al, Beilstein J. Nanotechnology, 11, 1742–1749 (2020) [3] W. Pilz et al, Journal of Vacuum Science &amp; Technology B 37, 021802 (2019) [4] L. Bischoff et al, Appl. Phys. Rev. 3, 021101 (2016)

Dr.-Ing.

Multi-voltage FIB-tomography of the hierarchical nanoporous gold

MS 2: Metals and alloys

Abstract text (incl. figure legends and references) Hierarchical nanoporous metals have attracted scientific interest due to their enhanced mechanical properties, e.g., strength and stiffness, in comparison to the continuous metals [1]. Hierarchical nanoporous gold (HNPG) represents a uniform network of nanoscale pores and solid metal ligaments with a size in a range of 15-110 nm. FIB-tomography where a slice-by-slice FIB milling of the HNPG sample is followed by a subsequent imaging of the obtained block faces by using high-resolution SEM, allows to reconstruct the investigated HNPG volume with a high precission [2]. However, in order to do so, the impact of the so-called shine-through-effect should be considered during the reconstruction process [3]. This effect though, might not only seen as an artifact but contribute additional information of the ligaments&quot; shapes and positions when simultaneously imaged different accelerating voltages. Thus, the images obtained at the higher voltage will contain more information from the depth of the sample while the low-voltage images would give the information about the shallow level as presented in Fig.1. However, it is challenging to interpret this information accurately as there is no unique mapping between material and intensities. In such situations, more complex reconstruction methods such as machine learning-based segmentation methods over classical thresholding methods should be preferred [4]. Additionally, our multi-voltage images, together with a different machine learning method, can further improve segmentation performance and reconstruct the nanostructures more accurately. Figure 1. SEM images of the HNPG structure obtained at 1kV (A), 2kV (B) and 4 kV(C) and the mean intensity of the specified area (red dashed square) as a function of the accelerating voltage (D). [1] Shi et al., Science 371 1026 (2021) [2] Winter et al., J Micros 233(3):372–383 (2009) [3] Moroni, Thiele, Ultramicroscopy 219 113090 (2020) [4] Sardhara et al., Frontiers in Materials 9 2296-8016 (2022)

Dr. habil.

Implementation of elastic and inelastic scattering processes for electrons with energies below 30 keV in Geant4 for the simulation of SEM

IM 5: Quantitative image and diffraction data analysis

Abstract text (incl. figure legends and references) A Monte Carlo tool called Geant4SEM is presented for the simulation of secondary electron (SE) emission, back scattered electron (BSE) trajectories and transmission signal in a scanning electron microscope (SEM). The tool is based on the powerful open source platform Geant4 [1]. Although initially it was developed for high energy physics simulations, the toolkit is well suited for analyzing interactions of particles with matter in many areas and has been used for similar purposes before [2]. Within Geant4, physics models can easily be implemented and tested individually. Thus far, four physics models have been implemented into the framework of Geant4. Mott cross-sections based on ELSEPA [3] have been used for the elastic scattering model. A dielectric function theory approach for cross sections of inelastic scattering and generation of SEs coming with JMONSEL [4] according to Shinotsuka et al. [5] has been adopted. Additionally, an inelastic scattering model for electron-phonon interaction based on Ganachaud and Mokrani [6] has been implemented as well as a model for boundary crossing of electrons. The detection of electrons can be set to occur right after exiting the sample or at a distance and can be filtered by electron properties such as kinetic energy and emission angle. Aforementioned models have been validated against experimental results. Figure 1 shows a comparison of simulated SE yields of gold at multiple primary electron energies with other simulations and experimental results of Walker et. al. [7]. Fig 2 displays signal profiles of an SEM image of a nominally 100 nm high and a 250 nm wide silicon step on a silicon wafer at a beam energy of 1.3 keV and an equivalent simulation. To achieve a quantitative comparison between the two, the exponential decay of the SEM signal has been fit to the simulated data and to the over multiple lines averaged measured data from the left edge towards the center of the silicon step sample. The exponential factor of our simulation lies within the margin of error of the experimental data's fit. It is demonstrated that Geant4SEM is capable of quantitative predictions of experimental SEM images. It deepens our understanding of the SEM imaging process and enables accurate measurements of nanosize structures such as semiconductor structures, fibers and particles of varying sizes, shapes and materials. [1] S. Agostinelli et al., Nucl. Instrum. Methods Phys. Res. Sect. A 506, 250-303 (2003) [2] E. Kieft, E. Bosch, J. Phys. D: Appl. Phys. 41, 215310 (2008) [3] F. Salvat et al., omput. Phys. Commun. 165, 157-190 (2005) [4] J. Villarubia et al., Ultramicroscopy 154, 15-28 (2015) [5] Shinotsuka, Tanuma, Powell, Surf. Interface Anal. 51, 427-457 (2019) [6] J.P. Ganachaud, A. Mokrani, Surface Science 334, 329-341 (1995) [7] Walker et al. Scanning. 30. 365-80. (2008)

Prof.

Video frame interpolation neural network for FIB-SEM tomography

LS 3: Imaging of large volumes and plastic section tomography

Abstract text (incl. figure legends and references) Introduction Focused-Ion-Beam Scanning-Electron-Microscope (FIB-SEM) tomography is a destructive imaging technique useful to investigate the 3D internal structure, and therefore the properties, of various materials composites. This technology is often employed for the study of nanomaterials, such as graphene nanosheet networks, whose mechanical and electrical properties are highly dependent on the material morphology. Through serial milling and imaging hundreds of network cross-sections, a high-fidelity 3D reconstruction of a nanostructured network can be produced and analysed [1]. The typical slice thickness is about 10-20 nm, while the resolution in the cross section is about 5 nm. It would be advisable to have the same resolution along the milling direction, this will result in a 3D reconstruction made of cubic voxels, which facilitate the studies of certain morphological properties. However, most imaging systems do not allow to produce thinner slices of the material and the process would be more expensive. In addition thinner slices have the disadvantage that the damage produced during one cut can propagate down to the next one. Objectives The solution proposed in this work [2] is to use a neural network developed for video frame interpolation, namely RIFE [3], to generate the missing slices between the collected images, and therefore obtain cubic voxels. Materials and methods The first application of the model was performed on a dataset made of graphene nanosheet networks, with pixel size of 5 nm and slice thickness of 17 nm. Along with computer vision metrics, such as mean squared error (MSE) and structural similarity index (SSIM), the evaluation of the network porosity in each frame was used to quantitatively compare the results to the ground truth and a simple linear interpolation method. Results The results show that the proposed model successfully generates the missing frames and outperforms the linear interpolation method on the graphene dataset. The application was extended to datasets of materials of a different scale generated using various imaging techniques, such as micro CT scans and serial block face SEM. Conclusion This work demonstrates that the resolution along the cut direction of FIB-SEM data, or other imaging techniques, can be improved by employing a neural network developed for video frame interpolation. Thanks to this model, it is possible to obtain cubic voxels in the 3D reconstruction despite the acquisition of thick slices, which facilitates the imaging process. [1] Gabbett, C. Electrical, Mechanical &amp; Morphological Characterisation of Nanosheet Networks. Diss. Trinity College Dublin, 2021. [2] Gambini, L., Gabbett, C., Coleman, J. N., Jones, L., Sanvito S. Video frame interpolation Neural Network for FIB-SEM tomography. In preparation. [3] Huang Z., Zhang T., Heng W., Shi B., Zhou S. Real-time intermediate flow estimation for video frame interpolation. In: Proceedings of the European Conference on Computer Vision (ECCV), 2022.

Traceable size measurement of polysterene particles with sizes up to 500 nm using STEM-in-SEM together with simulation-based image analysis

MS 3: Low-dimensional and quantum materials

Abstract text (incl. figure legends and references) Nano and micro particles possess size-dependent, desirable properties, but also pose risks to health and the environment. For this reason, exhaust emission standards such as Euro6 stipulate limits for the number of emitted particles, which are checked in exhaust emission tests (AU). Such official measurements must be traceable to SI units. Hence, a traceable sizing method for polysterene particles with sizes between 200 nm and 500 nm has been developed. For this purpose, an already existing procedure for smaller particles [1] has been adapted to the new task. The experimental procedure has already been described in Ref. [1]. In short, the polysterene particles are deposited on a TEM grid and imaged by STEM-in-SEM (or TSEM), using the bright field detector, see Fig. 1. The traceability to the SI unit meter is ensured by means of a two-dimensional grating standard. For size determination, the image pixels belonging to the particle must be separated from the background pixels. To achieve highly accurate measurements, the signal profiles are linearly interpolated and an individual threshold dependent on the particle size is used [1]. For relatively large polysterene particles, the threshold is located in the outer, non-linear section of the signal profile. Thus, a two-step approach has been chosen as depicted in Fig. 2: First, a preliminary value for the size is determined at a fixed threshold of 60%, which lies in the linear region of the signal profile. To obtain the final size, a tabulated value is added that has been determined by Monte Carlo simulations in advance. The simulation of the image formation process is performed using Geant4SEM. For this purpose, we have added our own physics classes [2] for single scattering processes of electrons to the powerful open source Monte Carlo framework Geant4 [3]. The elastic scattering cross sections have been calculated using ELSEPA [4] with muffin-tin radii. For the inelastic scattering cross sections, we have implemented a phenomenological model approximating the dielectric function. We measured polysterene particles with three different nominal sizes: 200 nm (PSL200), 300 nm (PSL300), and 500 nm (PSL500). Table 1 lists our results in comparison to the reference values of two international intercomparisons [5, 6] and the manufacturer's data sheet for PSL500, respectively. In all three cases, a very good agreement is shown. In addition, preliminary studies reveal that the uncertainty of our measurements is generally smaller than the pixel size. References [1] T. Klein et al., Meas. Sci. Technol. 22, 094002 (2011) [2] See also the MC2023 contribution of Martin Hermann et al. [3] J. Allison et al., Nucl. Instrum. Methods Phys. Res. 835, 186 (2016) [4] F. Salvat et al., Comput. Phys. Commun. 261, 107704 (2021) [5] F. Meli et al., Meas. Sci. Technol. 23, 125005 (2012) [6] H.-L. Lin et al., Metrologia 56, 04004 (2019)

JEOL`s JSM-IT800 series: high performance SEM for all users

Abstract text (incl. figure legends and references) Looking for a modern, high-resolution SEM allrounder that fits to your individual needs? JEOL&quot;s popular FE-SEM JSM-IT800 will serve you for fast and crisp high-resolution image acquisition, high throughput while maintaining maximum ease of use. The microscope now comes in 3 &quot;flavors&quot; in form of 3 objective lenses &quot;HL&quot; &quot;SHL&quot; and the new &quot;i&quot; – a model line-up, exactly designed for your individual purposes. Understand how tailored lens-technologies will enhance your results and how easy and automated it is to operate JEOL&quot;s flagship SEM featuring the fully integrated JEOL EDS system.

Advanced focused ion beam milling techniques for TEM specimen preparation of solid bulk samples

Abstract text (incl. figure legends and references) Introduction: In contrast to time-consuming conventional TEM-specimen preparation methods, the focused ion beam (FIB) sputtering in a dual beam SEM system allows to produce high-quality specimens for various TEM studies automatically and quickly. Different TEM based techniques require specific considerations during the specimen preparation. Objectives: In this work, we discuss three special FIB preparation methods: cutting of the electron beam shaping device from a bulk sample, preparation of the Ga ion beam sensitive nanotube cross-section and transport of the SiN-Membrane with magnetic structure on the surface. Materials and methods: For the preparation of specimens, we used focused Ga ion beams in FEI Helios dual-beam scanning electron microscopes. Results: Figure 1 shows the beam-shaping device for off-axis measurement of the phase shift around the vertical current. The U-shape cross-section TEM lamella of a Ga ion beam sensitive nanotube can be seen in the figure 2. Figure 3 illustrates the cross-section TEM lamella of the magnetic structure on the surface of a SiN membrane. Focused Ga ion beam sputtering is used to prepare high quality electron transparent specimens for different TEM experiments as well as to prepare some non-conventional specimens considering desired geometry, topology and material beam sensitivity.

TEM sample preparation with high-energy Xenon ions and focused low-energy Argon by plasma FIB

LS 7: Advances in sample preparation

Abstract text (incl. figure legends and references) Optimizing ion species for specific (S)TEM sample preparation tasks is not only becoming common with multi-ion species plasma FIB, but rather, essential as our understanding from experience and simulation data grows. Xenon ions provide a high sputter yield and are excellent for large 3D-volume-data-acquisition and correlated microscopy applications. Reactive oxygen ion beam processing has proven to dramatically improve milling quality for carbon-carbon bonding materials by smoothing the cut-face very effectively during sputtering. In this presentation, recent progress of low-energy-focused-argon ion beam cleaning is explored from trials with a Helios 5 Hydra Plasma FIB for TEM specimen preparation. Argon ion beam can be used down to 500 eV for specimen preparation, which is producing lower damage in finished samples. Ions species selection is critical for near "damage-free" TEM specimen preparation. Xenon plasma FIB first opened an opportunity for "gallium-free" TEM specimen preparation and had a benefit of lower amorphous layer thickness than gallium. However, Xenon ions induce higher vacancy-density in target materials. With conventional gallium FIB, there is a challenge to reduce or eliminate gallium in TEM specimen. Unfortunately, gallium implantation commonly leads to reactions and migration with many materials (e.g., steel, aluminum alloy, GaN/AlGN etc). Experimentation with the Helios Hydra for making (S)TEM samples has confirmed that the focused argon ion beam is suitable for near damage-free TEM specimen preparation guided by the investigation of ion-solid interaction SRIM data [1]. With Plasma FIB argon ion energy at 500 eV, it is becoming routine to produce near "damage-free" TEM specimens of these reactive materials which opens new opportunities for better material structure-property-function investigation. [1] C. Jiao, J. Graham, Xu Xu, T. Burnett, and B. van Leer. Low Energy 500 eV Focused Argon Ion Beam Provided by Multi-Ions Species Plasma FIB for Material Science Sample Preparations. Microsc.Microanal. 27 (Spupl 1). 2021

Using machine learning and topographic SEM imaging for software assisted fractography

Abstract text (incl. figure legends and references) The aim of a fractographic investigation is the evaluation of macroscopic and microscopic fracture surface characteristics and, as a result, the determination of the fracture mechanism of a component from a failure case. The basis for such evaluations of fracture characteristics comes from actual comparative mechanical testing and from the literature. A fractographic analysis can be very complex and, in any case, requires considerable experience. In the IGF project "iFrakto", software is being developed that quantitatively determines fracture characteristics and fracture mechanisms utilizing digitized expert knowledge, machine learning, and standard 2D and topographical data from SEM imaging. Topographical data are obtained from 4QBSE detector using shape-from-shading technology (Fig. 1). The developed software tool provides knowledge-based suggestions for the evaluation of fracture surfaces (Fig. 2). As a basis for this, round robins were carried out among fractographers to create a knowledge base, to query the practice-relevant requirements for such tools and to carry out first practical tests. The approach described can be used with all scanning electron microscopes. The results presented here were generated with a Tescan Vega3 and the BSE detector based topography measurement system DISS6 from point electronic GmbH. The investigations were mainly carried out on various metallic fracture surfaces. The results show that the use of machine learning software together with SEM image and topography data makes an important contribution to the quantification of fracture surfaces and effectively supports the operator in the interpretation of fracture surfaces.

PS 10
Poster session IM 3: SEM and FIB developments

Termin

Session chairs

Programm

PS 10Poster session IM 3: SEM and FIB developments

Termin

Session chairs

Programm

PS 10
Poster session IM 3: SEM and FIB developments