Julia Schmidt

4D-STEM Imaging With the pnCCD (S)TEM-Camera

4D-STEM imaging describes a powerful imaging technique where a two-dimensional image is recorded for each probe position of a two-dimensional STEM image. For typical STEM images of 256x256 probe positions, a total of 65,536 2D images needs to be recorded. This amount of data can be recorded with the pnCCD (S)TEM camera in a practical timeframe. This camera uses a direct detecting, radiation hard pnCCD with a minimum readout speed of 1000 full frames per second (fps) [1]. With the pnCCD (S)TEM camera a 4D data cube consisting of 256x256 probe positions with a 264x264 pixel detector image for each probe position can be recorded in less than 70 s. Several measurements have been performed to prove the capability of the camera for 4D-STEM imaging, among them strain analysis, magnetic domain mapping and most recently electron ptychography.

Analysis of Polymorphs Using Simultaneous X-ray Fluorescence and Diffraction with an Imaging Spectrometer

X-ray fluorescence (XRF) is a widely used method for materials characterization. However, for polymorphs such as Rutile and Anatase, both of which are TiO2, the fluorescence spectra will be nearly identical. Differentiating between these polymorphs requires a second measurement with an X-ray diffractometer. Using two instruments increases the cost of the measurement, and it introduces issues with image registration and sample preparation. Previously, simultaneous XRD-XRF has been done by attaching a silicon drift detector (SDD) to an X-ray diffraction (XRD) system. However, due to the typical geometry and source monochromation of an XRD system, the count rate in the SDD is typically low, making the measurement process slow [1]. Overcoming these limitations requires a detector that is both position sensitive and energy dispersive.

A New Approach: Quantitative Imaging Spectrometry for Iron Oxide Analysis

Solid state X-ray detectors have undergone major improvements and advances in the last 50 years. Such advances include, but are not limited to, improvements in energy resolution, count rate and solid angle. One significant improvement is the development of detectors capable of recording both the position and energy of every X-ray event that hits the detector. Such detectors are known as imaging spectrometers, since they combine the spectroscopic performance of the silicon drift detector with the position resolution of an imaging detector.

Advanced 4D STEM Imaging with the pnCCD (S)TEM Camera

Inspired by recent advancements in detector technology, there has been increasing interest in the technique of four dimensional scanning transmission electron microscopy imaging (4D STEM). Typically, a focused beam of electrons is applied to a sample in a two dimensional (2D) raster pattern. At each point, a 2D image is captured which intrinsically contains bright field (BF), dark field (DF), and high angle annular dark field (HAADF) signals. In total, a 4D dataset is recorded allowing a comprehensive analysis and enabling a number of techniques such as strain analysis, magnetic domain mapping, scanning electron diffraction, TEM tomography, and electron ptychography. All these 4D STEM imaging techniques are possible with the pnCCD (S)TEM camera, which meets the following key requirements: (i) Fast acquisition of data to minimize effects of beam and sample drift, and beam induced damage to the sample; (ii) a large number of detector pixels to precisely discriminate between diffraction spots or to determine the position of a bright field disk with high accuracy; and (iii) a sufficiently radiation hard detector compatible with high electron rates as used in scanning electron diffraction experiments.

Pushing the Limits of Fast Acquisition in TEM Tomography and 4D-STEM

In both TEM and STEM many experiments profit from r ecording two-dimensional camera images at very high readout speeds. This is especially true f or tomography in TEM and ptychography in STEM. The pnCCD (S)TEM camera is routinely running at 1 0 00 frames per second (fps) in full frame mode [1]. This camera uses a direct detecting, radiation hard pnCCD with 264x264 pixels and features binnin g and windowing modes, which allow to further increas the frame rate substantially. For example, 4-fold binning in one direction, i.e. 66x264 pixels, yield s a readout speed of 4000 fps. In windowing modes, up to 20 000 fps are possible. Benefitting application s range from imaging on the microand millisecond timescale to strain analysis or electric field mapp ing.

Scanning electron diffraction using the pnCCD (S)TEM Camera

Scanning electron diffraction (SED), performed in a STEM, is a powerful technique combining information in reciprocal space and real space to achieve nanoscale crystal cartography of materials structure. SED involves scanning a focused electron beam across a specimen and recording an electron diffraction pattern at each position to yield a 4D dataset comprising a 2D diffraction pattern at every position in the 2D scan region. Obtaining high quality data depends on fast acquisition, large dynamic range, and accurate recording of the location and intensity of diffraction spots. Here, we present SED measurements using the pnCCD (S)TEM camera taking a Ti-Fe-Mo alloy for demonstration. The pnCCD (S)TEM camera provides fast acquisition of 2D camera images using a direct detecting, radiation hard pnCCD with 264x264 pixels [1]. Routinely, the readout speed is 1000 frames per second and can be further increased by binning and windowing. The large number of pixels and high readout speed of this camera enables the recording of high quality diffraction patterns in a short acquisition time. Further, the camera properties can be changed by modifying the voltages applied to the pnCCD providing several camera operation modes [2]. Considering scanning electron diffraction experiments, which are performed at high electron beam intensities, the combination of data recorded in two different camera operation modes allows a comprehensive diffraction pattern analysis with quantitative and spatial information. The 4D datasets acquired with the pnCCD (S)TEM camera can be analysed in a number of ways [3], most simply by plotting the intensity of a subset of pixels as a function of probe position in flexible post-experiment schemes to obtain ‘virtual diffraction images’ or to perform differential phase contrast analysis. Using virtual diffraction images, the two phases existing in an ultra-fine lamellar microstructure [4] in this Ti-Fe-Mo alloy can be clearly discriminated. [1] H. Ryll et al, Journal of Instrumentation 11 (2016), p.04006 [2] J. Schmidt et al, Journal of Instrumentation 11 (2016), p. P01012 [3] P. Moeck et al, Crystal Research and Technology 46 (2011), p.589-606 [4] A.J. Knowles et al (2016), Proceedings of the 13th World Conference on Titanium DNJ, RKL & PAM acknowledge: ERC grant 291522-3DIMAGE and EU grant 312483-ESTEEM2. PA11

Setup and Practical Applications of a pnCCD Based XRF System

Micro-focused X-ray fluorescence (μXRF) analysis has proven very useful as a complementary X-ray imaging technique [1]. The technique requires a focused X-ray beam, a stage capable of moving in a raster pattern, and an X-ray spectrometer (normally a silicon drift detector, SDD). Unfortunately, images are built pixel-by-pixel, using the stages to bring the sample under the stationary beam. This means that the time required to create an image is limited by the speed of the stage, not the detector. High speed stages tend to have poor position reproducibility, and accurate stages tend to be slower and more expensive. To eliminate this limitation, a new method of X-ray imaging must be established. In optical experiments, the incoming light is focused with an optic onto an imaging device. This device records all of the information simultaneously, greatly improving the speed with which an image is produced. It preserves both the position and the energy information of each incoming photon, making it an ideal imaging device. Unfortunately, these imagers are not normally sensitive to X-rays. However, a new type of CCD, known as the pnCCD is capable of detecting both the position and energy of each X-ray event.

Characterization and Comparison of Detector Systems for Large Area X-ray Imaging

In the last 15 years, there has been a major shift in the type of detector used in X-ray microanalysis [1]. The silicon drift detector (SDD) is now the predominant energy dispersive X-ray spectrometer, having replaced the lithium drifted silicon detector. With high spectral resolution at input count rates over 100 000 counts/s, an SDD reduces the total time required to record a high quality X-ray image. In 2005, a multi-element SDD was also introduced, which combined multiple SDDs into a single detector [2]. These detectors combine large solid angles (for more efficient X-ray collection) with fast spectral processing through multiple pulse processors. Yet its primary drawback is that the SDD does not record the position of each X-ray event in addition to the energy. An array of SDDs could do this, but the position resolution would be poor. Otherwise, an X-ray scintillator combined with an ordinary charge coupled device (CCD) can record the location of the X-ray event, but the energy resolution is usually poor. An ideal imaging spectrometer would be one that can record both position and energy.

Dynamic range of fully depleted pnCCDs: modeling and experimental confirmation

pnCCDs are a special type of charge coupled device (CCD) which were originally developed for applications in X-ray astronomy. At X-ray Free Electron Lasers (XFEL) pnCCDs are used as imaging X-ray spectrometers due to their outstanding characteristics like high readout speed, high and homogenous quantum efficiency, low readout noise, radiation hardness and a high pixel charge handling capacity. With pnCCDs it is possible to separate one photon from no photon and two photons as well as being able to measure up to up to 104 photons per pixel per frame. However, extremely high photon intensities can result in pixel saturation and charge spilling into neighboring pixels. Because of this charge blooming effect, spatial information is reduced. Due to the deep understanding of the internal potential distribution we can enhance the pixel full well capacity even more and improve the quality of the image. This paper describes the influence of the operation voltages and space charge distribution of the pnCCD on the electric potential profile by using 2D numerical device simulations. Experimental results with signal injection from an optical laser confirm the simulation models.

High Speed, High Throughput Two Dimensional Direct Electron Detector Based on the Concept of pnCCDs

Two dimensional direct detection electron imaging solid state sensor systems, subject to intense radiation, suffer from charge overflow to neighbouring pixels. This not only affects the dynamic range and linearity, but also the position precision of e.g. the electron impinging on the detector. Various techniques have been proposed to overcome the problem of charge overflowing the pixels. Several sensors are able to cope with up to a few million signal charges per pixel. In many scientific applications, e.g. in free electron laser science and electron detection in TEMs, this is by far not sufficient. We have developed a technique to handle 1 billion signal charges per pixel without charge spilling over to neighbouring pixel within one readout frame of typically 1 ms. The quality of the point spread function remains unchanged but amplitude information may be degraded. We call this mode of operation: controlled charge extraction (CE) mode.