X-ray source

The initial X-ray spectroscopy experiments were performed using hermetically sealed X-ray tubes [15]. Excillum, Inc. [16] has developed metal-jet X-ray tubes, which are conventional micro-focus tubes that utilize a liquid metal jet instead of a solid metal anode [15]. The metal jet supports a higher electron beam power, thereby generating a higher X-ray flux. Therefore, Excillum’s D2 was selected for laboratory XAFS measurements. The X-ray source was a high-purity liquid gallium jet anode. The size of the tube’s focal spot was 20 m × 80 m, which was transformed to a point source size of 20 m × 20 m at a take-off angle of 14.5°. The apex angle (opening angle) of the X-ray cone was approximately 8°, which covered a circular area 70 mm in diameter on the crystal. The maximum accelerating potential was 70 kV and the maximum current was 3.57 mA. The X-ray source window was a 50-μm thick beryllium window that transmitted 2-keV X-rays only half, and almost all (more than 82%) were above 3 keV.

Crystal

Johann-type spherically bent crystal analyzers (SBCAs) are used for monochromatizing and focusing polychromatic X-ray “bremsstrahlung” energy. To minimize the impact of strain fields resulting from spherical bending on the energy resolution, crystal wafers were sliced into strips 15 mm wide prior to being bonded with the glass concave substrate. These strips had an energy resolution similar to diced-bent crystals and also had a higher efficiency. All crystals were purchased from XRS LLC, Inc. [17]. The SBCAs had a bending radius of 500 mm and a surface diameter of 100 mm. To cover the widest possible working range within the 2.0–9.0 keV energy range and the spectrometer’s angular scanning range (55–80°), we are gradually expanding our collection of crystal analyzers. Currently, the available analyzer crystals include Si(111), Si(220), Si(311), Si (400), Si(331), Si(422), Si(533), Ge(110), and Ge(620). Table 1 provides an overview of the spectrometer crystals currently accessible, including their respective coverages within the first order of reflection across the working range and theoretical energy Darwin widths corresponding to those reflections. An energy gap exists between 2.4–3.2 keV because Si and Ge do not have crystal planes suitable for the intermediate energy range. Other crystalline materials (e.g., quartz) can cover this energy range; however, processing spherical quartz crystals remains challenging.

Detector

For transmission XAFS measurements, the spectrometer was equipped with an Amptek Fast SDD, which effectively captured the intensity of the diffraction signal throughout the data collection process. The Amptek Fast SDD is a vacuum-compatible detector with excellent energy resolution that suppresses higher-order harmonics in the diffraction signal while minimizing the background noise. It features a 500-μm thick silicon sensor layer bonded to the top of the electronic layer, providing an effective area of 50 mm². The SDD can achieve a counting rate of up to 1 × 10⁶ counts/s with a peak time constant of 0.2 μs. The energy resolution was 123 eV at 5.9 keV. Integrated within the SDD was a two-stage Peltier cooling system designed to prevent heat buildup and ensure efficient operation.

Motors and movement

The X-ray source was kept stationary outside the vacuum chamber. The crystal analyzer was mounted on a motorized module that provided vertical and pitch angular adjustments for alignment as well as horizontal and rotational adjustments for energy scanning. The goniometer held the detector and was positioned on two linear stages, D_x and D_y, as shown in Fig. 1(b). The energy-scanning mechanism of the spectrometer followed the Bragg angle equation and trigonometric formulas, which defined the geometric requirements of the Rowland circle.

The X-ray source, sample, and bent crystal analyzer were arranged on a Rowland circle with diameter R. The detector was rigidly connected to the sample, which was always directed at the center of the spherical curved crystal by a mechanical linkage. For the tests, the sample was placed in front of the detector. Whenever the Bragg angle θ or energy changed, the position of the detector on the Rowland circle needed to be shifted. Both the angle and distance of the bent crystal analyzer were simultaneously adjusted. Using the position of the X-ray source as the origin of the coordinate system, the x-axis was defined as the direction from the source to the crystal, and the y-axis was defined as the direction from the source to the detector. The bent crystal moved exclusively along the x-axis. The spectrometer’s motor positions (Ax, Dx, and Dy) for a specific photon energy E (measured in eV) are given by $E=\frac{nhc}{2d_{hkl}\text{sin}\theta }$ (1) $A_{x }=R\text{sin}\theta ,$ (2) $D_{x }= 2R\text{ sin }\theta {\text{ cos}}^2\theta ,$ (3) $D_y =2R{\text{ sin}}^2\theta \text{ cos }\theta ,$ (4) , where h is Planck’s constant, c is the speed of light, R is the radius of the spherical bent crystal analyzer (Rowland circle diameter), θ is the Bragg angle, d_hkl is the d-spacing for the given crystal type used, A_x is the x-coordinate of the bent crystal, and D_x and D_y are the x- and y-coordinates of the detector, respectively.

Vacuum chamber

The tender X-ray monochromator employed a custom vacuum chamber with a pressure of 10^-6 mbar, which was achieved using a 700-L/s vacuum pump. To prevent mechanical deformation due to pressure differences, all the mechanical components were mounted on a separate sturdy steel plate within the vacuum chamber, ensuring that the diffraction plane was aligned. Owing to the volume limitations of individual components, achieving proximity during the operations proved challenging, resulting in vacuum chamber dimensions of 1 m × 0.8 m × 0.6 m and Bragg angles ranging from 55° to 80°.

Monochromatic Flux

The monochromatic flux was tested with Ge (620) crystals in the energy range of 7040–7050 eV (Fig. 2(a)). When the voltage was held constant and the current was decreased, the counting rate decreased. The counting rates corresponding to the same power were similar. When the current was held constant and the voltage was increased, the counting rate increased, reaching a maximum at 60 kV, 3.5 mA, and 210 W (approximately 5 × 10⁵ counts/s).

Fig. 2Full-screen View

(Color online) (a) Monochromatic flux tested with Ge (620) crystals in the range of 7040–7050 keV in different powders. (b) Fluorescence peaks of gallium Kα₂ and Kα₁ obtained by using the characteristic peaks of the X-ray source

Energy Resolution

Directly evaluating the energy resolution of laboratory source systems is challenging. Two factors contribute to the total energy resolution: energy broadening corresponding to the core-hole lifetime and the intrinsic energy resolution of the spectrometer. To characterize the energy resolution of the spectrometer, we used the characteristic peaks of the X-ray source to obtain a strong signal. The X-ray source was operated at 120 W (40 kV, 3 mA) using a Si553 crystal monochromator. Figure 2(b) shows the test results: peaks A and B were fluorescence peaks for gallium Kα₂ (9223.8 eV, Bragg angle 71.905°) and Kα₁ (9250.6 eV, Bragg angle 71.403°), respectively. The full width at half maximum (FWHM) of peaks A and B were 2.96 eV and 2.89 eV, respectively. The energy broadening corresponding to the core-hole lifetime of the Ga Kα₂ and Kα₁ peaks were 2.66 eV and 2.59 eV, respectively [21]. The final energy resolution of the instrument (△E) was 1.298 eV at 9223.8 eV and 1.28 eV at 9250.6 eV.

Additionally, we theoretically derived the instrument resolution for different energy ranges (Eq. 5). First, we obtained the value of △E₁ for the crystal based on the tested value of △θ and the crystal energy at 71.905º. The value of △E₁ is the difference between the energy resolution of the spectrometer and the energy Darwinian width (ΔE_D) of the crystal. In a laboratory spectrometer, different energy ranges correspond to different crystal diffraction planes. Therefore, Darwin width correction was applied to each diffraction plane to obtain the final instrument resolution (the Darwin width of each crystal at a Bragg angle of 71.905° was obtained from the XAS data for the elements). As shown in Table 3, the energy resolution ranged from 0.36 to 1.30 eV at 2.0 to 9.0 keV, covering the instrument’s energy range. The energy resolution is expressed as $\frac{\text{Δ}E}{E}=\text{Δ}\theta \text{ cot }\theta ,$ (5) , where the θ is the Bragg angle, △E is the intrinsic energy resolution, and E is the energy.

XANES

In Figs. 3(a) and (b), the normalized K-edge X-ray absorption near the edge structures (XANES) of the titanium foil and anatase TiO₂ are shown alongside the synchrotron data from the BL14W1 beamline of the Shanghai Synchrotron Radiation Facility (SSRF). All XAFS data were background removed and normalized using the software package Athena [22]. The Bragg angle, denoted as θ = 66.865°, was determined based on the first peak observed in the derivative spectrum of the Ti foil and assigned a value of 4966 eV. The consistency of the spectra shown in Fig. 3(a) suggests that the energy resolution of the laboratory monochromator was similar to that (1.0 eV) achieved using a double-crystal Si(111) monochromator for a synchrotron. Nevertheless, recognizing the inherent challenges associated with directly assessing the energy resolution in laboratory source systems is crucial, as it encompasses both the measurement of the FWHM and comprehensive characterization of the entire energy spectrum. The FWHM of the pre-edge peak of the Ti foil was approximately 1.2 eV. All features could be reproduced, demonstrating that the laboratory device was suitable for a wide range of applications.

Fig. 3Full-screen View

(Color online) Normalized K-edge XANES spectrum of (a) titanium foil and (b) anatase TiO₂ compared with synchrotron data obtained using the BL14W1 beamline at the SSRF. (c) normalized K-edge XANES spectrum of rutile TiO₂ and anatase TiO₂ based on the laboratory spectrometer. (d) Ti K-edge XANES spectra of the pristine Ti₃AlC₂ before and after MS-etching and Ti K-edge XANES patterns of Ti₃AlC₂ (a type of MXene) before and after MS-etching, alongside reference spectra of Ti foil and TiO₂ (reprinted with permission from Ref. [23], copyright 2021 Wiley Online Library). Normalized M5 edge XANES spectrum of (e) ThF₄ and (f) UO₂ compared with synchrotron data obtained using the XAFCA beamline at the SSLS. (g) U-M4 edge conventional XANES data of U-doped La₂Zr₂O₇ -MS samples (reprinted with permission from Ref. [24], copyright 2022 Royal Society of Chemistry). (h) U-M5 XANES spectra of UCoO₄ (black) and UO₂ (orange) (reprinted with permission from Ref. [25], copyright 2021 ACS Publications)

The XANES spectra of the TiO₂ in anatase and rutile forms is displayed in Fig. 3(c). The spectrum consisted of pre-edge components A1–A3, a distinctive shoulder B, and multiple peaks C1–C3. Other studies [26–28] have described the origins of these features. In the rutile TiO₂ sample, features A1 and A2 appeared at lower energies compared to the anatase sample, whereas features B and C appeared at the same energies as those of the anatase.

In our previous study, we employed a tender X-ray monochromator to analyze MXene materials (a type of 2D material [29]), as shown in Fig. 3(d). We primarily focused on performing near-edge structural analysis of Ti₃AlC₂ and Ti₃C₂T_x materials (a type of MXene), particularly those pertinent to supercapacitor applications [23]. Building on these findings, we thoroughly investigated the structural attributes and valence changes.

Unlike the actinide L-edge, the M-edge of actinide elements exhibits less pronounced core energy level broadening and greater sensitivity to valence states. However, the M-edges of Th (3.3 keV) and U (3.5 keV) are outside the range of hard X-rays (> 5 keV) and have not been effectively tested using laboratory X-ray absorption spectrometers. However, these values are within the energy range suitable for tender X-ray monochromators. Accordingly, the Th and U M-edge absorption spectra were experimentally examined using a laboratory light source, as shown in Figs. 3(e) and (f).

Figure 3(f) shows the normalized XANES spectra of UO₂ and UO₂(NO₃)₂ along with the synchrotron data obtained from the XAFCA beamline at the Singapore Synchrotron Light Source (SSLS). XAFS data were normalized using the Athena software package [22]. Compared to the XAFCA and SSLS data, the UO₂(NO₃)₂ and UO₂ spectra collected by the laboratory spectrometer showed a high degree of agreement, indicating that the energy resolution of the spectrometer was comparable to that of synchrotron radiation near 3.5 keV. All the observed features were reproducible, and this level of replication was adequate for most applications.

To support the nuclear energy sector’s efforts to immobilize actinide waste efficiently (Fig. 3(g)), we employed conventional U-M4 edge XANES spectra to test the valence state of uranium ions and their coordination environment in samples of U-doped La₂Zr₂O₇ -MS [24]. In the regime of efficient catalysts (Fig. 3(h)), the XANES analysis revealed an upward shift in the U-M5 XANES spectra obtained from the transmission mode measurements of the UCoO₄ catalysts [25], confirming the presence of U⁶⁺ and providing critical evidence of electronic structure modulation among polymetallic sites.

EXAFS

The Bragg angles ranged from 55° to 80°, allowing the extended X-ray absorption fine structure (EXAFS) measurements to extend up to or beyond 1000 eV above the absorption edge. Ni and Co foils were selected as samples for the different energy ranges. The XAFS data for these samples are shown in Fig. 4. Figures 4c and d show the k-space spectra of the two samples, and Figs. 4e and f display the R-space spectra. The EXAFS spectra obtained from the laboratory spectrometer and 14W1 beamline at the SSRF were comparable in their k-space oscillations (2 – 12 Å-1), as shown in Figs. 4(a-f). In the regime of electrocatalytic CO₂ reduction (Figs. 4g–i), the transition metal EXAFS of MPc (M=Fe, Co, Ni) catalysts supported on carbon nanotubes was analyzed using a tender X-ray monochromator [30]. The coordination number and bond length were ascertained via the fitting results, significantly contributing to the elucidation of the catalyst’s structure.

Fig. 4Full-screen View

(Color online) (a) Ni K-edge XAFS spectrum. (b) Co K-edge XAFS spectrum. (c) Ni K-edge XANES spectrum. (d) Co K-edge EXAFS spectrum. (e) Ni K-edge FT-EXAFS data. (f) Co K-edge FT-EXAFS data. EXAFS spectra of three MPc-CNT samples: (g) FePc- CNT at Fe K-edge, (h) NiPc-CNT at Ni K-edge, and (i) CoPc-CNT at Co K-edge. (Reprinted with permission from Ref. [30], copyright 2023 Wiley Online Library)