A new experimental setup for time- and laterally-resolved X-ray absorption fine structure spectroscopy in a ‘single shot’

Abstract

In this work, a new setup for dispersive XAFS measurements is presented. This reproducible and scanning-free setup yields both time- and laterally-resolved XAFS experiments in a ‘single-shot’. It allows a straightforward adjustment for probing different elements covering many relevant applications in materials science. An incoming energetic broadband beam is diffracted by a Si (111) crystal after passing through the sample and collected by an area sensitive detector. Depending on the energy range of the incoming beam, XANES and/or EXAFS spectra can be recorded with a time resolution down to 1 s. The feasibility of this setup was demonstrated at the BAMline at BESSY II (Berlin, Germany) with reference Fe and Cu foils and the results are hereby presented and discussed. Additionally, an application where time resolution on the second scale is required is briefly evaluated. The presented example concerns studying early stages of zinc(II)2-methylimidazolate (ZIF-8) crystallization. This is particularly important for biomedical applications.

---

Introduction

X-ray absorption fine structure spectroscopy (XAFS) is a widely used method for material characterization, mainly owing to its element-specific character. This promotes XAFS to be one of the few methods that delivers chemical and structural information on amorphous phases, which is of crucial importance in many applications. This method entails both XANES (X-ray absorption near edge structure) and EXAFS (extended X-ray absorption fine structure) spectroscopy.¹ Additionally, XAFS has become very popular for in situ/in operando measurements in research fields where real-time monitoring is needed. One of these fields is the study of crystallization mechanisms of metal–organic framework (MOF) coordination compounds such as zinc(II)2-methylimidazolate (ZIF-8). The development of stable, easy to handle and non-expensive setups for use in different experiments and reaction environments is of great interest, both for academic and industrial research. In the last few years, the time resolution of XAFS measurements has significantly improved. One method that has replaced the step-by-step acquisition of every energy point by (quasi) continuous scanning of the whole spectrum, quick EXAFS (QEXAFS), is meanwhile well established.² With a fast-moving monochromator, whole spectra can be recorded in the millisecond range. QEXAFS can be found in some beamlines. For example, at the new beamline P64 (PETRA III, Hamburg), fully operating since April 2017, a time resolution in the 20–100 ms range is achieved.³ Another example is the SuperXAS beamline at the Swiss Light Source (SLS) at the Paul Scherrer Institute. XANES spectra can be measured in 10 ms and EXAFS spectra in 17 ms based on an optimized setup (new monochromator, ionization chambers and data acquisition system) installed in 2015.⁴ The beam is focused onto the sample with a width of typically 100 μm × 100 μm and a lateral precision of 50 μm due to the movement of the monochromator during the scan. Another example for QEXAFS is the BM23 beamline at ESRF where QEXAFS on the second timescale and a spot size of 4 μm × 4 μm is achieved.⁵

Another approach that has become a well-established method in the last few decades is dispersive XAS. The principle of this method is the diffraction of a polychromatic beam (with the desired energy bandwidth) by a bent crystal. This setup was first presented in 1981 ⁶ and since then numerous studies have been conducted to improve the performance and to determine the properties of dispersive EXAFS, as already published.^7–9 The different energies are reflected at different angles; they are focused on the sample and recorded after passing the sample by an area sensitive detector. Herein, CCD-based systems are often used. The advantages are as follows: on the one hand, it is a scanning-free measurement, leading to a high stability of the beam position; on the other hand, it allows fast recording of XAFS spectra since all energy points can be acquired in a single shot. Dispersive XAFS is in operation at the ID24 beamline from ESRF,^10,11 SOLEIL synchrotron, and in commissioning mode at the I20 beamline from Diamond Light Source. ESRF has developed a fast readout low noise (FReLoN) CCD camera, which enables 20 frames per second. Nowadays, a beam size of 5 μm × 5 μm with a time resolution of about 60 ms can be achieved.¹¹

The aim of QEXAFS and dispersive XAS is to achieve a higher temporal and spatial resolution. In QEXAFS, the time resolution is limited by the monochromator speed, and there are high requirements concerning the reproducibility of the monochromator positions and, subsequently, the beam stability. In dispersive XAFS systems, the readout time by area sensitive detectors is the limiting factor. Generally, spectroscopic CCD-based detectors have a readout time of tens of milliseconds. However, for many applications there is no need for such high temporal resolutions. In these cases, this setup is too elaborate and costly.

High spatial resolution might be a disadvantage when larger areas of a sample are to be investigated. With small beam sizes many points would have to be measured to scan the whole area. A 2D-mapping avoiding moving the sample can be performed with the use of a so-called letter-box beam as described elsewhere.¹² In this approach, stepwise energy-scanning was combined with a CCD camera to get a XANES spectrum for each pixel and hence a 2D-mapping of the sample.

The combination of time- and laterally-resolved XAFS analysis has been already previously reported by Katayama et al.¹³ In this case the beam is focused in one direction for energy resolution, while the other direction can be used for lateral resolution. This is achieved by using the traditional dispersive XAFS geometry, fulfilling the Rowland circle. This requires a precise placement of the sample to ensure that it is in the focus point of all energies. Furthermore, the dispersive element is placed before the sample. This yields additional unwanted scattering at the detector, which has to be prevented by the use of slits.

As already mentioned, we present a new DXAFS setup for time- and laterally-resolved XAFS in a single shot, for which the proof-of-principle is already published.¹⁴ The aim is to have a stable and reproducible setup for applications and case studies, in which certain dynamic processes occur on the second timescale. It represents, therefore, no concurrence with QEXAFS.

Some disadvantages which are present in other setups are overcome in our setup.

One big advantage is that the dispersive element is placed after the sample. This requires the detector to be placed at a 2θ angle, which prevents other effects from being detected, such as scattering. It is ensured that only the transmitted X-rays are reflected, and no slit systems are necessary.

Furthermore, this approach is stable, and it is easy to adjust owing to a bending mechanism allowing a simple and reproducible adjustment of the bending radius of the crystal. Like other dispersive XAFS setups, it is scanning-free as there are no moving components, thus leading to a high beam stability.

Summarizing, a broad polychromatic bandwidth beam is used to illuminate the sample and a convexly bent crystal is placed behind the sample as a dispersive element (Fig. 1). All energies are reflected at a different angle and collected by the CCD camera simultaneously. In this way, a whole XANES or EXAFS spectrum (depending on the bandwidth of the incoming beam) can be recorded in a single shot in the typical θ–2θ geometry. Simultaneously, the height of the incoming beam can be used to get a lateral resolution in the range of the beam size. This setup is versatile, because it allows quick and simple adaption to different energies without extensive geometrical considerations. In addition, there are no or only a few constraints from the side of the geometrical arrangement.

Fig. 1 Geometrical arrangement of the new DXAFS setup.

Experimental setup

Beamsource

All measurements were carried out at the BAMline at BESSY II, Helmholtz-Zentrum Berlin. The X-ray source is a superconducting 7 T wavelength shifter. It provides hard X-rays in the region of 4–100 keV.¹⁵ The beamline has two monochromators: a double-crystal monochromator (DCM) and a double-multilayer monochromator (DMM). Each of the two mirrors of the DMM consists of a Si substrate covered with W/Si coatings with an intrinsic bandwidth of ΔE/E = 1.7% full width at half maximum (FWHM). The DCM is an arrangement of two Si (111) crystals with an intrinsic energy resolution of 2 × 10⁻⁴ and was used for energy calibration. In this work, two options to generate the incoming broadband beam are used, both using the DMM. The first option combines the DMM in total reflection (‘mirror’) mode with a filter to obtain an energy bandpass, already described previously.¹⁴ This configuration allows a large usable energy range, which is necessary for EXFAS measurements. Depending on the DMM angle and the filter, different energy ranges can be obtained. In all cases, EXAFS measurements at K-lines of several transition metals as well as at L-lines of platinum group metals and rare earth elements are feasible. The second option is to use the intrinsic bandwidth of the DMM by setting it at the required energy. This covers in all cases the full XANES range.

In this manuscript, three examples are shown for testing the new experimental setup: the Fe-K edge (7.112 keV), Cu-K edge (8.979 eV) and Zn-K edge (9.669 keV). Table 1 shows the experimental parameters. The settings of the DMM for the bandpass were adopted from the standard XAFS-measurements, where the DMM is set in the total reflection mode at the required energy level to reject the higher harmonics. For every bandpass in this experiment the lower energy limit is due to the used filters (e.g. 60 μm Al filter) and the upper one is due to the W–L lines from the material of the DMM.

Table 1 Summary of the experimental parameters for each measuring mode

Mode	Element	Parameters	Energy range
EXAFS (bandpass)	Fe	DMM@38 keV with a 10 μm Be filter	3–10 keV
	Cu	DMM@50 keV with a 60 μm Al filter	7–10 keV
XANES (DMM@Eedge)	Fe	DMM@7.112 keV	150 eV around the edge
	Cu	DMM@8.979 keV	190 eV around the edge
	Zn	DMM@9.7 keV	200 eV around the edge

---

Dispersive element: Si (111) crystal

The used crystal is a commercially available single side polished Si (111) wafer from Siegert Wafer GmbH (Aachen, Germany) with a thickness of 525 μm and a diameter of 100 mm. It is convexly bent by means of the so-called wafer bender, which was developed in house (Fig. 2). The crystal is placed onto a vertical polished rod. It is held by two steel rods on the left and on the right side with slits to host the crystal. They are joined by an aluminium bar that is fixed on a linear stage with a maximum travel distance of 10 mm. The setup is equipped with a stepper motor and a position recorder to allow symmetrical and reproducible bending of the crystal.

Fig. 2 Details of the wafer bender with a (a) linear stage for the bending, (b) rotating motor for setting the required angles, and (c) linear stage for positioning with respect to the incoming beam.

In order to estimate the shape of the bending curve, a simulation with Abaqus software from SIMULIA¹⁶ has been carried out. This was performed for different applied forces, meaning different bending positions of the stepper motor. For this purpose, it was assumed that the middle of the crystal is fixed where it is placed on the vertical polished rod. Then a displacement on the two sides of the crystal perpendicular to its surface corresponding to the motor position was taken as the boundary condition. For defined points on the crystal their displacement in x-, y- and z-directions was simulated. The results of these simulations show that the bending curve in the horizontal middle of the crystal is nearly a circle line and that the bending is directly proportional to the displacement on the sides of the crystal, i.e. the bending position. From this proportionality, it follows that the bending process for every position of the bending motor can be calculated from the simulated bending position of 1 mm, which is denoted as h_s1 shown in Fig. 3.

Fig. 3 Simulation of the bending of the crystal, with a displacement of 1 mm perpendicular to the crystal surface (bending position h_s1) on both sides of the crystal. The inlay shows the displacement of every point of the crystal surface perpendicular to the surface (in the y-direction) from 0 mm (dark blue) to 1 mm (red). The black line corresponds to the plotted black crosses.

Furthermore, the angles on every point of the crystal surface can be calculated from the simulated bending line. This information is necessary to determine the bending position for a given energy range, which should be calculated prior to the experiment. Let us consider the following example: XANES measurements at the Fe-K edge with an energy range from E₁ = 7.050 keV to E₂ = 7.200 keV. The corresponding angles are θ₁ = 16.286° and θ₂ = 15.938°. These angles should represent the limits of the illuminated area on the crystal. For the bending position h_s1 (bending position 1 mm) the angles at points of the crystal surface near to the edges of the illuminated area are known from the simulation, and they are denoted as θ_1s1 and θ_2s1. Since the applied force and bending of the crystal are proportional, the required bending position for the desired angles is:

Assuming that the crystal is symmetrically bent and the difference between the average angle and θ₁ and θ₂ is the same, there is no difference in determining h by taking the maximum or the minimum angle.

The optimal distance between the crystal and the detector, L₀, can be calculated from the beam width, bw, the area of the detector screen, A_CCD, and the difference between the average angle and the minimum or maximum angle, δθ = θ₁ − θ_av = θ_av − θ₂ (Fig. 4):

Fig. 4 Reflection of the beam on the crystal and the illuminated area on the CCD screen. bw is the beam width, θ_av is the average angle and the angle of the tangent on the bent crystal in the middle of the illuminated area (K), δθ is the difference of the angles at the edges of the illuminated area and the average angle. L₀, L₁ and L₂ are the distances between the crystal (middle and edges) and the CCD screen.

For this example, the bending position is h = 0.1874 mm for a beam width of 10 mm and a detector area of 20 mm.

Simulation of the beam path

As previously stated, this setup is based on a simple θ–2θ geometry. The key feature is the simultaneously polychromatic and spatially broad incoming X-ray beam. In order to determine how the energy range is distributed over the CCD and based on the simulated bending curve, the beam path was simulated by an own-developed routine in the IDL 6.1® program (Harris Geospatial Solutions, Broomfield, USA). For a given bending position h, average angle θ, detector angle θ_CCD, the distance between the detector and the crystal, L₀, and the beam width, bw, the expected position of the beam on the detector screen can be simulated. The simulations have been performed to check the calculated values and to prepare the experiments. Keeping with the same example (XANES@Fe-K edge) the result of the simulation is shown in Fig. 5. The given values are: h = 0.1874 mm, L₀ = 822 mm, θ_av = 16.1117°, θ_CCD = 32.2234°, and bw = 10 mm. These values have been calculated based on the reflections described above. Hence, a 20 mm length of illumination of the CCD screen (black line in Fig. 5) was considered.

Fig. 5 Simulation of the reflected polychromatic X-ray beam onto the CCD for XANES measurements at the Fe-K edge.

Detector

The used detector consists of a CCD camera, pco.4000 manufactured by PCO AG (Kelheim, Germany). It has a resolution of 4008 × 2672 pixels with a pixel size of 9 μm × 9 μm. The minimum exposure time is 5 μs. In combination with a Rodenstock objective (Qioptiq Photonic GmbH & Co. KG, Feldkirchen, Germany) with a focal distance of 100 mm and a Nikon objective (Nikon Instruments Europe B.V., Amsterdam, Netherlands) with a focal distance of 180 mm the pixel size of the recorded images is 4.8 μm. This allows a usable sensitive area of 19.46 mm × 12.83 mm. To convert the X-ray radiation into visible light, to which the camera is sensitive, a fluorescent screen (P43) from ProxiVision GmbH (Bensheim, Germany) is implemented in front of the first objective. The bit depth n is 14 for the pco.4000 meaning that the camera records values between 0 and 2¹⁴ (= 16 348) for the intensity pattern. This intensity pattern, corrected by a flatfield, is a direct measure of the absorption profile of the sample. For camera control, image acquisition and archiving of images, the software Camware is used. Images are saved in .tif format and processed with ImageJ software or with a self-written IDL® program.

Geometrical setup

The wafer bender is placed on a rotating table DMT65-DM4-HSM and a linear table MTM60-10-HiSM, both from OWIS GmbH (Staufen, Germany), to adjust the angle between the incoming beam and the crystal and the position in the beam. The adjustment is done remotely during operation.

All this, depicted in Fig. 6, is mounted on a frame, together with the detector at an adjustable distance, which in turn is fixed on another rotating table (Art312XM-HMC) from Aerotech GmbH (Fürth, Germany) allowing the θ–2θ geometry.

Fig. 6 The new DXAFS setup. The polychromatic beam passes through the sample before it is dispersed by the bent crystal (a) and collected by the CCD (b). The wafer bender is mounted on a rotating table and a linear stage. The distance between the crystal and the camera is adjustable.

Results

Energy calibration

After the measurements, energy calibration is required. This is performed with the same experimental adjustments as the measurements. For this purpose, the DMM is removed and the DCM is placed in the optical part of the beamline. Pictures are taken every 10 eV scanned by the DCM. In the case of the XANES@Fe-K edge, the recorded lines are shown in Fig. 7(a). The inclination is due to the divergence of the incoming beam. The differences between the upper and lower end of the line are between 53 and 77 pixels or 3.7 eV and 4.8 eV (calculated after the calibration from the pixel difference). Taking the middle of every line, it is possible to assign a pixel to every energy and plot a calibration curve. This curve can be compared with a calibration curve obtained from the simulation (Fig. 5) by calculating the intersections of the reflected beam of each energy with the screen of the detector. Both energy calibrations, the experimental and the one obtained from the simulation, are shown in Fig. 7(b). Simulation and experiment are in good agreement.

Fig. 7 (a) All images of the energy calibration taken together in one image. The number of pixels is denoted in red, and the corresponding energy in blue. (b) Calibration curves from the experiment and from the simulation for the Fe XANES measurements.

Fe reference foil

The first XANES measurements were performed on Fe reference foil with a thickness of 10 μm from Goodfellow Cambridge Limited (Huntingdon, England). The crystal bender linear stage was at a position of 0.12 mm; the distance between the detector screen and the crystal was approximately 200 mm; the average angle (θ_av) between the crystal and the beam was 15.77°, and the average angle (2θ_av) between the detector and the beam was 31.54°. In this case the DMM was set at an energy of 7.15 keV, with a beam size of 5 mm × 15 mm (height × width). The measured flatfield had the form of a Gaussian, corresponding to the intrinsic bandwidth of the DMM. The full width at half maximum is calculated to be approx. which corresponds to ΔE ≈ 121 eV for the adjusted energy. The energy range with intensity higher than 35% of the maximum intensity is approximately 150 eV. This range is reasonable for XANES measurements. Fig. 8 depicts the results of the measurement of the Fe foil. Herein, the flatfield has been chosen corresponding to its maximum being approximately at the position of the Fe-K edge. This leads to a good visibility of the post-edge features as the energy intensity of the flatfield is sufficient for this region.

Fig. 8 XANES measurement of 10 μm thick Fe foil with an exposure time of 10 s for the whole spectrum (DXAFS) and a total of 20 minute acquisition time for standard XANES. The flatfield is the limiting factor for the detectable energy range.

Cu reference foil

XANES measurements at the Cu K-edge were carried out with 12.5 μm thick foil from Goodfellow Cambridge Limited (Huntingdon, England). The DMM was set at 9.039 keV, with a usable energy range of approx. 190 eV and a beam size of 3 mm × 20 mm (height × width). The crystal bender linear stage was at a position of 0.5 mm; the distance between the detector screen and the crystal was approximately 640 mm; the average angle (θ_av) between the crystal and the beam was 13.26° (angle values displayed by the motor control of the rotating tables). The results of the flatfield and the corresponding XANES spectrum with an exposure time of 1 s are shown in Fig. 9, compared with a conventionally measured XANES spectrum of the same foil, recorded within 20 minutes.

Fig. 9 XANES measurement of 12.5 μm Cu foil with an exposure time of 1 s for the whole spectrum.

For EXAFS measurements on Cu foil, the DMM is employed in mirror mode at an energy of 50 keV in combination with a 60 μm thick Al filter (see bandpass mode in Table 1). The beam size was 3 mm × 6 mm (height × width). The crystal bender linear stage was at a position of 1.25 mm. This leads to a broader reflected energy range than needed for the measurements. Therefore, there were more possibilities to optimize the image on the CCD screen by means of the two adjustable angles. The angle (θ_av) between the crystal and the beam was 12.07°, and the best image on the CCD screen was obtained at a camera angle of 24.8° at a distance between the camera and the crystal of about 520 mm. In Fig. 10(a) the EXAFS spectrum obtained in 6 s by DXAFS is shown, together with a conventionally measured EXAFS spectrum of the same foil (40 minutes for the whole spectrum). The wavefunctions in the k space and the Fourier-transformed data in the R space are depicted in Fig. 10(b) and (c).

Fig. 10 EXAFS measurement of 12.5 μm Cu foil with an exposure time of 6 s by DXAFS. (a) The spectrum is compared to a conventionally obtained one in 40 minutes. (b) Wavefunctions in the k space and (c) Fourier-transformed data in the R space.

Case study: early stages of ZIF-8 crystallization

As already stated, DXAFS allows us to follow in situ reactions that take place on the second timescale. First tests of time resolved measurements were performed to observe changes in the coordination of metal ions in solution as it is experienced during typical ZIF-8 synthesis.^17,18 Therefore, Zn²⁺ in aqueous solution was mixed with aqueous 2-methylimidazole solution, leading to subsequent coordination of 2-methylimidazole to zinc ions, which occurs within seconds. During the whole reaction, a picture was taken every two seconds. The DMM was operated at an energy of 9.7 keV, The beam size was 4 mm × 12 mm (height × width); the crystal bender linear stage was at a position of 0.12 mm; the angle (θ_av) between the crystal and the beam was 11.18°; the camera was at an angle of 23.7° and at a distance of approximately 520 mm.

The obtained spectra are shown in Fig. 11. Zinc chloride shows a typical single peak spectrum that can be observed in this case in the beginning of the experiment. This has been associated with zinc ions that are octahedrally coordinated by 6 water molecules within the first coordination sphere and two chloride anions within the second coordination sphere.¹⁹

Fig. 11 XANES measurement of aqueous ZnCl₂ solution that is mixed with aqueous 2-methylimidazole solution. Pictures are taken every 2 seconds.

Within 6 seconds, the spectrum changes and bears a resemblance to the spectrum of tetrakis(1-methylimidazole)zinc²⁺ ions. The spectrum in the presented kinetic measurement further converts to tetrahedrally coordinated zinc(II)2-methylimidazolate (ZIF-8).

The observed formation of an intermediate compound that resembles the spectra of tetrakis(1-methylimidazole)zinc²⁺ ions and likely is tetrakis(2-methylimidazole)²⁺ is in line with recently published pair distribution function (PDF) based results.²⁰ More importantly, the presented measurements indicate that tetrakis(1-methylimidazole)zinc²⁺ is a stable and chemically very similar surrogate to tetrakis(2-methylimidazole)²⁺ which only forms transiently during early stages of ZIF-8 crystallization.

Conclusions

A new setup for reproducible time-resolved XANES and EXAFS measurements has been tested successfully. With the newly developed bent wafer setup it is possible to record a whole XANES or EXAFS spectrum in a single shot. The beam path has been calculated theoretically and has been simulated for different sets of parameters. The experimentally found energy calibration is in good agreement with the simulations. First tests with metal foils have been performed.

The Si crystal as a dispersive element enables a completely scanning-free measurement of the spectra which are recorded by a CCD camera. The wafer bender mechanism and the rotating tables allow a simple and reproducible adjustment of the bending radius and the reflecting angles for different energy ranges. The parameters for different probing elements can be optimized and adapted to address questions in materials science such as in studies that address early crystallisation stages of ZIF-8. Here, the combination of DXAFS and classical XAFS has revealed the chemical similarity between stable tetrakis(1-methylimidazole)zinc²⁺ and tetrakis(2-methylimidazole)zinc²⁺ that only forms as an elusive intermediate during ZIF-8 synthesis. This finding provides further opportunities for studying the coordination chemistry of ZIF-8 during its formation. The letter-box beam can be used for scanning the sample in one direction. Simulations of crystal bending have shown that the bending curve is directly proportional to the applied force. This allows us to theoretically calculate the bending curve of the crystal for arbitrary applied forces from a single simulation. One aim for the future is to get the energy calibration curve directly from the simulation instead of measuring it for every single adjustment. The first comparisons between simulation and experiment show promising results. Future applications will be XANES and/or EXAFS measurements during chemical reactions for in situ monitoring of structural changes as demonstrated here for prototypical MOF ZIF-8. This also includes the use of lateral resolution using the height of the beam, e.g. to track gradients of oxidation states e.g. in catalyst materials.

Conflicts of interest

There are no conflicts of interest to declare.

Acknowledgements

The authors want to thank HZB for the allocation of synchrotron radiation beamtime, Monika Klinger and Bettina Röder (BAM) for the technical drawings of the setup and Cetin Haftaoglu (BAM) for the simulations of the bending properties of the crystal. S. Beyer appreciates the postdoctoral Adolf Martens Fellowship.

Notes and references

1. G. Bunker, Introduction to XAFS: A Practical Guide to X-ray Absorption Fine Structure Spectroscopy, Cambridge University Press, 2010, p. 270.
2. R. Frahm, J. Stützel, and D. Lützenkirchen-Hecht, State of the art Quick-EXAFS: Applications in Catalysis, in Users Meeting 2009 and Associated Workshops, ESRF Grenoble, 2009.
3. O. Müller, D. Lützenkirchen-Hecht, and R. Frahm, Instrumentation for QEXAFS Spectroscopy at P64, 2013.
4. O. Muller, et al., Quick-EXAFS setup at the SuperXAS beamline for in situ X-ray absorption spectroscopy with 10 ms time resolution, J. Synchrotron Radiat., 2016, 23(1), 260–266.
5. O. Mathon, et al., The time-resolved and extreme conditions XAS (TEXAS) facility at the European Synchrotron Radiation Facility: the general-purpose EXAFS bending-magnet beamline BM23, J. Synchrotron Radiat., 2015, 22(6), 1548–1554.
6. T. Matsushita and R. P. Phizackerley, A Fast X-Ray Absorption Spectrometer for Use with Synchrotron Radiation, Jpn. J. Appl. Phys., 1981, 20(11), 2223–2228.
7. A. M. Flank, et al., EXAFS in dispersive mode, Nucl. Instrum. Methods Phys. Res., 1983, 208(1), 651–654.
8. E. Dartyge, et al., X-ray absorption in dispersive mode: A new spectrometer and a data acquisition system for fast kinetics, Nucl. Instrum. Methods Phys. Res., Sect. A, 1986, 246(1), 452–460.
9. K. Ukyo, M. Tadashi and K. Kazutake, A Dispersive Method of Measuring Extended X-Ray Absorption Fine Structure, Jpn. J. Appl. Phys., 1981, 20(5), L355.
10. S. Pascarelli, T. Neisius and S. De Panfilis, Turbo-XAS: dispersive XAFS using sequential acquisition, J. Synchrotron Radiat., 1999, 6, 1044–1050.
11. S. Pascarelli, et al., Dispersive XAS on a High Brilliance Source: Highlights and Future Opportunities, in AIP Conference Proceedings, IOP Institute of Physics Publishing Ltd, 2007.
12. J.-D. Grunwaldt, et al., 2D-Mapping of the Catalyst Structure Inside a Catalytic Microreactor at Work: Partial Oxidation of Methane over Rh/Al₂O₃, J. Phys. Chem., 2006, 110, 8674–8680.
13. M. Katayama, et al., Development of dispersive XAFS system for analysis of time-resolved spatial distribution of electrode reaction, J. Synchrotron Radiat., 2015, 22(5), 1227–1232.
14. A. G. Buzanich, et al., Time- and spatial-resolved XAFS spectroscopy in a single shot: new analytical possibilities for in situ material characterization, J. Synchrotron Radiat., 2016, 23(pt 3), 769–776.
15. H. Riesemeier, et al., Layout and first XRF applications of the BAMlineat BESSY II, X-Ray Spectrom., 2005, 34(2), 160–163.
16. Dassault Systèmes, https://www.3ds.com/products-services/simulia/, 2002–2017, [cited 2017 2017-05-17].
17. S. Beyer, et al., Maintaining Stable Zeolitic Imidazolate Framework (ZIF) Templates during Polyelectrolyte Multilayer Coating, Colloid Interface Sci. Commun., 2018, 22, 14–17.
18. S. Beyer, et al., Ultra-Sonication of ZIF-67 Crystals Results in ZIF-67 Nano-Flakes, ChemistrySelect, 2016, 1(18), 5905–5908.
19. P. D'Angelo, et al., Structural characterization of zinc(II) chloride in aqueous solution and in the protic ionic liquid ethyl ammonium nitrate by x-ray absorption spectroscopy, J. Chem. Phys., 2011, 135(15), 154509.
20. M. W. Terban, et al., Early stage structural development of prototypical zeolitic imidazolate framework (ZIF) in solution, Nanoscale, 2018, 10(9), 4291–4300.

---