Introduction
X-ray fluorescence analysis has played a significant role in material elemental identification and charge state measurements during exploration of Mars, the Moon, and asteroids [1-3]. It also has important applications in advanced light sources [4-8] and scanning electron microscopy [9-10]. Depending on the practical scientific application, various excitation sources can be used, including X-ray sources [1, 12], electron excitation sources [13-15], proton excitation sources [16-18], and α particle excitation sources [2, 19, 20]. X-ray sources mainly include X-ray tubes, synchrotron radiation sources, and free-electron lasers, and they have the widest range of applications. Electron excitation sources include β-ray sources and electron guns, with electron guns being primarily used in electron-beam etching, scanning electron microscopy, and transmission electron microscopy. Proton sources are primarily small accelerators, whereas α particle excitation sources include radioactive sources and accelerators. Other excitation sources such as heavy ion and neutron sources are relatively complex and will not be discussed here.
The interaction mechanisms between the various excitation sources and materials are significantly different [17, 21-24]. The interactions of photons and electrons with atoms, for instance, exhibit distinct differences with regard to selection rules and energy-momentum conservation restrictions, etc. This results in varying processes and probabilities of the excitations for electrons located in different atomic shells when interacting with photons and electrons [24-26]. First, owing to the constraints of energy and momentum conservation, when the energy of a photon is absorbed by an outer-shell electron, triggering its ejection, the required momentum is supplied by the atomic nucleus. Therefore, the stronger the interaction between the electron and the atomic nucleus, the more easily the electron is excited. Hence, photons exhibit a higher propensity to excite inner-shell electrons during interactions with atoms. Second, because photons are electrically neutral, they mainly interact with one electron upon incidence. When a photon interacts with an outer-shell electron and is absorbed, the corresponding electron can only transition to certain energy levels and spin states. This indicates that specific selection rules must be complied with during this process.
In contrast, electrons, unlike photons, do not vanish after interacting with the outer-shell electrons and do not require momentum from the atomic nucleus. Additionally, the stronger the bond between the atomic nucleus and the outer-shell electrons, the more difficult it is for incident electrons to excite them. This is why electrons tend to excite outer-shell electrons more readily when they interact with atoms. Furthermore, when charged, electrons interact with multiple outer-shell electrons upon incidence. Finally, because electrons do not disappear before or after interactions with atoms, and can simultaneously interact with multiple outer-shell electrons, the excitation process does not need to conform to specific selection rules. Owing to the disparities in the electron shells and spin states excited by electrons and photons, differences may arise in the transition paths, as well as in the intensity of these transitions, reflecting the differing probabilities of each type of excitation event. Such differences among the excited electrons could lead to variations in the intensities of the characteristic spectral lines emitted by the same element during de-excitation. These variations in transition probabilities and their resultant spectral lines, in terms of both relative intensities and likelihood of occurrence, are critical for interpreting spectral data. The energy level transition paths pertaining to the electron shells and fluorescence spectral lines are depicted in Fig. 1.
The differences in the excitation processes between electrons and photons primarily affect the outer-shell electrons, which exhibit a more significant impact than the inner-shell electrons. This is because the outer-shell electrons are more susceptible to variations in excitation energy and interaction mechanisms. As a result, discrepancies in the spectral lines, particularly in the are more pronounced because these series involve transitions that are significantly influenced by the state of the outer-shell electrons. On the other hand, the α-line series, which results from transitions between the K and L shells, is characterized by more stable interactions, and thus shows less variation. A challenge arises when comparing the spectral lines of the β-line and γ-line series, which exhibit small energy separations (typically less than 100 eV, and in some cases, as minimal as 10 eV) for practical scientific applications. Such narrow energy gaps make it difficult to measure materials and identify elements using conventional detection devices such as silicon drift detectors (SDD). The energy resolution of an SDD is approximately 120 eV, which is insufficient for discriminating the narrowly separated individual spectral lines. To obtain more precise and detailed information about material structures, it is crucial to utilize a detector capable of high energy resolution across a wide energy range [27-30]. A detector using a superconducting transition-edge sensor (TES) can perfectly satisfy these requirements. By utilizing cryogenic X-ray spectrometers equipped with TES detectors, we can more effectively discern and accurately measure spectral features, particularly for the precise distinction of the spectral lines in the β-line and γ-line series, and relevant measurements can then be performed [31-36].
In this study, the TES detectors demonstrated superior energy resolution and detection efficiency for X-rays within the energy range of 1.5 keV to 17.0 keV [33, 37-39]. Therefore, certain spectral lines within this range were selected for analysis. The K characteristic lines of copper (Cu) lie between 8 keV and 9 keV, whereas the L line series of tungsten (W) also falls within the sensitive range of our spectrometer. Therefore, both copper and tungsten are ideal candidates for our comparative study of the spectral differences observed between electron and X-ray excitation.
Although an electron gun linked to a high vacuum chamber is the optimal electron source for generating an electron-induced X-ray spectrum, implementing such an arrangement adds significant complexity to the experimental setup. To achieve a practical solution, we used commercially available X-ray tubes with copper and tungsten targets, which essentially replicate electron excitation by bombarding the anode material with high-energy electrons. These X-ray tubes provide a more accessible means of obtaining electron excitation spectra, theoretically akin to an electron gun setup. For the photon excitation spectra, we utilized the molybdenum (Mo) target X-ray tube generating X-ray radiation to irradiate high-purity copper and tungsten samples. By employing TES detectors to separately acquire and analyze the resultant energy spectra from the electron and photon excitations, we can discern and compare the K characteristic lines of copper and L characteristic lines of tungsten excited by either electrons or photons. The use of TES detectors was pivotal in this study, as they provided the acute energy resolution necessary to discern subtle differences and obtain definitive results.
Measurement and Analysis
Experimental Setup
Our chosen electron source for the experiment comprised X-ray tubes that utilize accelerated and focused electron beams to excite the metal samples and generate X-rays [25]. In the electron source, high-energy electron beams are injected into copper or tungsten targets. The principle underlying this X-ray generation is the interaction between the injected high-energy electron beam and the target atoms. When electrons in the beam encounter the atoms in the metal, they can either ionize them through collisional ionization or excite atomic electrons to higher energy levels. These excited electrons undergo transitions and return to lower energy levels, and in the process, they emit characteristic X-ray spectral lines indicative of their elemental origins.
In addition, we utilized a molybdenum target X-ray tube as the photon source to interact with high-purity copper and tungsten samples. In the X-ray tube, molybdenum atoms emit X-ray photons when an electric current is applied. These emitted X-rays possess the energy necessary to excite the high-purity copper and tungsten samples, thus exciting their inner-shell electrons. As these excited electrons return to their ground state, the released energy is emitted in the form of X-ray photons. These photons carry quantized energies that are characteristic of the element involved, thereby generating X-ray fluorescence (XRF) spectra with distinct energies and frequencies.
Theoretically, the adjustable parameters of the X-ray tube, such as the current and voltage settings, heavily influence the spectral output by determining the intensity and energy distribution of the emitted X-rays. In electron excitation experiments, X-ray tube parameter adjustment is crucial for controlling the spectral output of the electron excitation of copper and tungsten targets. The acceleration voltage (V) is meticulously set above the threshold excitation voltage (Vth) of the target material, which leads to the generation of a spectrum predominantly characterized by the inherent spectral lines of the material, supplemented by a continuous X-ray background, i.e., Bremsstrahlung. As V increases beyond Vth, an increase in the intensity of the characteristic lines (Ich) can be observed according to the empirical relationship.
Measurement and Results
To obtain the intrinsic spectra of copper and tungsten using the electron source, it is essential to appropriately adjust the experimental conditions. For copper, with its K-edge at 8.980 keV, we selected voltage settings of 11 kV, 12 kV, 13 kV, and 14 kV to ensure that the beam's energy surpassed this threshold, facilitating the excitation of the characteristic lines of copper. Similarly, because the L1, L2, and L3 edges of tungsten are 12.102 keV, 11.540 keV, and 10.200 keV, the voltage settings of the tungsten target were 15 kV, 16 kV, 18 kV, and 20 kV, respectively. For both targets, the currents were carefully selected from the milliampere (mA) range to achieve the required energy levels. The electron beam was accelerated and focused to generate the desired energy levels necessary for these experiments, with each run lasting approximately 12 h to ensure sufficient data of adequate quality and statistical significance.
Regarding the use of the X-ray source, for the molybdenum target with the K-edge of molybdenum at 20.000 keV, the most intense Kα1 line energy produced was 17.479 keV, which provided sufficient energy to excite the inner-shell electrons in the studied samples. Consequently, the voltage settings for the molybdenum target to illuminate the pure metal were chosen as 30 kV, 40 kV, 45 kV, and 50 kV. The corresponding current levels were chosen in the appropriate ampere (A) range to effectively optimize the sample excitation and X-ray emission within the desired energy range. Subsequently, the XRF mode was employed to obtain the energy spectrum. Given the negligible differences in spectra across various experimental setups, except for count rate variations, the results at the 50 keV setting were adopted as a benchmark for further analyses, ensuring a standard reference point for comparing X-ray emission efficiencies and sample excitations across the experiments.
Figure 3 illustrates a comparison of the intrinsic spectrum of copper acquired from the electron source and the corresponding XRF energy spectrum derived from the photon source. For the electron-source-derived spectra, we applied voltages ranging from 11 kV to 14 kV to the copper target for a comprehensive analysis. In contrast, for the photon source, the spectrum resulting from the 50 kV configuration using a molybdenum target is presented because it consistently provided representative data. All data were processed after calibration and normalization procedures before analysis, and compared with the spectrum obtained under similar experimental conditions using the SDD detector. Our analysis shows that for the copper K-edge energy range, the characteristic Kα and Kβ lines exhibited remarkable consistency across the two different excitation sources.
The normalization approach applied in Fig. 3 for copper leverages the Kα peaks, comprising Kα1, and Kα2. These peaks correlate with the energy level transitions of KL3 and KL2 in copper, respectively. From our analysis, it is notable that across varying voltages within the intrinsic and XRF energy spectra, the intensities of the left Kβ lines exhibited no significant variations under either the electron or photon excitation sources. This result aligns with the theoretical understanding that the transition paths of all K line series are associated with the K shell, which is a singular energy level without subshell divisions. Upon further scrutiny of the spectral behavior, we confirmed that the total counts and spectral line probability densities for the copper K line series were substantial, which guaranteed the detection result.
Similarly, Fig. 4 compares the intrinsic spectrum of tungsten elicited by the electron source and the XRF energy spectrum of tungsten obtained using the photon source. The operational voltages of the tungsten target for the electron source were 15 kV, 16 kV, 18 kV, and 20 kV. For the photon source analysis, a 50 kV configuration using a molybdenum target was selected as a representative example. All data were processed after calibration and normalization procedures and compared with the spectrum obtained under similar conditions using the SDD detector. A comparative analysis revealed that within the energy range of the characteristic lines of the L-edge of tungsten, the Lα lines of tungsten showed no noticeable variations under different excitation sources. However, distinct differences were observed within the Lβ line series. For a focused examination of these discrepancies, the energy spectra within the Lβ line energy range were enlarged and compared, selecting the intrinsic electron source spectrum at representative voltages of 15 kV and 20 kV, along with the XRF energy spectrum of tungsten obtained from the photon source, as shown in Fig. 5a. In particular, for the intrinsic electron source spectrum at a representative voltage of 20 kV, each peak within the Lβ line energy range was analyzed and marked with its energy level transition path. To complement this, Fig. 5b provides a schematic of the energy level transition paths for the corresponding illustration.
The normalization process is shown in Fig. 4 was performed based on the peak intensities of the Lα lines of tungsten, including Lα1 and Lα2, corresponding to the energy level transition paths of L3M5 and L3M4, respectively. Additionally, the enlarged Fig. 5a shows that the peak intensities of Lβ2 (L3N5) and Lβ6 (L3N1) were consistent with the Lα lines across various representative voltages within the intrinsic and XRF energy spectra. This observation suggests that the spectral lines resulting from the de-excitation process of excited electrons related to the same outermost energy level, the L3 subshell, exhibit similar peak intensities. For the other characteristic spectral lines, the peak intensities of Lβ4 (L1M2) and Lβ3 (L1M3), associated with the innermost energy level of the L-edge, the L1 subshell, were noticeably weaker in the intrinsic energy spectrum than in the XRF energy spectrum. This reflects the lower probability of exciting electrons in the inner subshell compared to the outer subshell during electron excitation. Similarly, the peak intensity of Lβ1 (L2M4), related to the subouter energy level, the L2 subshell, was weaker in the intrinsic spectrum than in the XRF spectrum. Moreover, we detected a unique peak located at the high-energy end near Lβ1 in the intrinsic energy spectrum, which was absent in the XRF spectrum. Accurately targeting the energy value of this peak center allows it to be numerically matched with the energy difference between the L2 and M5 levels. However, L2M5 is an energy level that contravenes the selection rules; thus, it is not represented as a characteristic energy level in the XRF energy spectrum. This finding indirectly verifies the initial assertion that specific selection rules are obligatory in the photon excitation process, whereas electron excitation is not bound by these rules.
Drawing on the foundational concepts introduced in the Introduction and reinforced by our experimental data, we can establish that photons and electrons engage with atomic structures under profoundly different constraints of energy-momentum conservation and selection rules, dictating the likelihood and nature of electron excitation across various atomic shells. Building upon the meticulous analytical observations delineated for both copper in Fig. 3 and tungsten in Fig. 4, our discussion further elucidates the intricate dynamics of electron excitation and photon interaction within these elements, as evidenced by their spectral behaviors. For copper, the consistent behavior of the K line series underlines the singular nature of the K shell, which is a foundational aspect that underpins our understanding of the electron structure and transition probabilities within this metal. The uniform response across varying excitation modes not only underscores the theoretical frameworks guiding our interpretations but also highlights the robustness of the detection methodologies employed, ensuring the reliability of the observed spectral features. As for tungsten, the contrast in the peak intensities between the Lα and Lβ lines clarifies the nuanced differences attributed to the subshell configurations within the L-edge. The variations observed in the Lβ line series capture the inherent variations in the excitation probabilities between the inner and outer shells during electron interactions. This asymmetry underscores the role of charged electrons in readily exciting outer-shell electrons, as opposed to their neutral photon counterparts, which abide by selection rules and are more likely to interact with inner-shell electrons. Especially noteworthy is the revelation provided by the unique peak near the Lβ1 line, an anomaly that poses a challenge to conventional selection rules and paves the way for a deeper investigation into the mechanics of electron excitation versus photon excitation. This anomaly enriches our understanding of tungsten's spectral lines under diverse excitation scenarios, highlighting an excitation pathway that is available only to electron excitation.
Conclusion and Outlook
In conclusion, our study provides insight into the spectral responses of copper and tungsten when subjected to various excitation sources. For copper, the experimental data revealed that the intensities of the Kα and Kβ lines remained constant across the electron and photon excitation methods. This observation aligns with theoretical insights into the structure of the K shell, suggesting that the lack of subshell divisions results in uniform transitions that are less sensitive to the type of excitation source used. In contrast, the spectral analysis of tungsten demonstrated distinct behaviors, especially for Lβ lines under varying excitation sources. Although the Lα lines were consistent, the Lβ lines displayed notable variations, which can be attributed to the complex subshell configuration and its influence on the excitation and de-excitation processes.
Looking ahead, the findings of our current study can serve as a springboard for advancing our understanding of atomic excitation dynamics and spectroscopic analysis. With the nuances of spectral line behaviors now in sharper focus, our upcoming research initiatives will aim to refine the methods we use to analyze and interpret data from an expanded spectrum of materials. An electron gun spanning a 1 keV to 30 keV energy range has been procured for the laboratory, and the development of a matching high-vacuum chamber is already in progress. These tools will allow us to extend our investigations to metals beyond copper and tungsten and delve into the intricacies of the Lβ line series across different elements. A TES detector equipped with a sufficiently high energy resolution and wide-range capability promises a more meticulous deconstruction of individual spectral lines. This development is intended to reveal finer details of material structures, offering a richer dataset for energy-dispersive spectroscopy applications, particularly within the realm of electron microscopy.
Furthermore, experiments are planned using Am-241 as an α particle source to juxtapose and contrast the energy spectra evoked by electron versus α particle interactions. These research endeavors will ultimately establish a comprehensive database. This reservoir of knowledge is envisioned to function in sync with sophisticated space science instruments such as APXS, enhancing both terrestrial and extraterrestrial material analyses. Through these ambitious projects, we aim not only to bolster the foundation of spectroscopy, but also to contribute a significant corpus of data and analytical expertise to the expansive fields of material science, chemistry, and astrophysics.
Ground-based verification and data processing of Yutu rover Active Particle-induced X-ray Spectrometer
. Chin. Phys. C 39,Elemental composition of manganese- and phosphorus-rich nodules in the Knockfarril Hill member, Gale crater
. Mars. Icarus 392,First Operation of TES Microcalorimeters in Space with the Micro-X Sounding Rocket
. J. Low Temp. Phys. 199, 1062–1071 (2020). https://doi.org/10.1007/s10909-019-02293-5Hard X-ray micro-focusing beamline at SSRF
. Nucl. Sci. Tech. 26,High-resolution X-ray emission spectroscopy with transition-edge sensors: present performance and future potential
. J. Synchrotron Radiat. 22, 766–775 (2015). https://doi.org/10.1107/s1600577515004312A practical superconducting-microcalorimeter X-ray spectrometer for beamline and laboratory science
. Rev. Sci. Instrum. 88,L-edge spectroscopy of dilute, radiation-sensitive systems using a transition-edge-sensor array
. J. Chem. Phys. 147,Soft X-ray spectroscopy with transition-edge sensors at Stanford Synchrotron Radiation Lightsource beamline 10-1
. Rev. Sci. Instrum. 90,3D characterization of porosity and minerals of low-permeability uranium-bearing sandstone based on multi-resolution image fusion
. Nucl. Sci. Tech. 31, 105 (2020). https://doi.org/10.1007/s41365-020-00810-wEvaluation of energy-dispersive X-ray spectrometer with transition-edge sensor microcalorimeter on scanning transmission electron microscope for detection of low-concentration phosphorus in Fe–P alloy
. Appl. Phys. Express 13Hyperspectral X-ray Imaging with TES Detectors for Nanoscale Chemical Speciation Mapping
. J. Low Temp. Phys. 200, 437–444 (2020). https://doi.org/10.1007/s10909-020-02456-9Method and portable facility for measurement of trace element concentration in prostate fluid samples using radionuclide-induced energy-dispersive X-ray fluorescent analysis
. Nucl. Sci. Tech. 27, 136 (2016). https://doi.org/10.1007/s41365-016-0133-3Calibrating the linearity between grayscale and element content for X-ray KES imaging of alloys
. Nucl. Sci. Tech. 33, 1 (2022). https://doi.org/10.1007/s41365-022-00986-3Advanced β-ray-induced X-ray spectrometry for non-destructive measurement of tritium retained in fusion related materials
. Fusion Eng. Des. 109–111, 1569–1573 (2016). https://doi.org/10.1016/j.fusengdes.2015.11.020Tritium distribution analysis of Be limiter tiles from JET-ITER like wall campaigns using imaging plate technique and β-ray induced X-ray spectrometry
. Fusion Eng. Des. 160,Broadband ultrahigh-resolution spectroscopy of particle-induced X rays: extending the limits of nondestructive analysis
. Phys. Rev. Appl. 6,Applications of laser-induced breakdown spectroscopy for cultural heritage: A comparison with X-ray Fluorescence and Particle Induced X-ray Emission techniques
. Spectrochim. Acta B 149, 1–14 (2018). https://doi.org/10.1016/j.sab.2018.07.012PIXE and Its Applications to Elemental Analysis
. Quantum Beam Sci. 3, 12 (2019). https://doi.org/10.3390/qubs3020012Statistical characterization of PIXL trace element detection limits
. Acta Astronaut. 212, 534–540 (2023). https://doi.org/10.1016/j.actaastro.2023.08.032Mars Science Laboratory Alpha Particle X-ray spectrometer trace elements: Situational sensitivity to Co, Ni, Cu, Zn, Ga, Ge, and Br
. Acta Astronaut. 165, 32–42 (2019). https://doi.org/10.1016/j.actaastro.2019.08.026Comparison of particle and photon excited X-ray fluorescence applied to trace element measurements of environmental samples
. Nucl. Instrum. Methods 106, 525–538 (1973). https://doi.org/10.1016/0029-554X(73)90318-2PIXE and XRF analysis of particulate matter samples: an inter-laboratory comparison
. Nucl. Instrum. Methods Phys. Res. B 266, 2401–2404 (2008). https://doi.org/10.1016/j.nimb.2008.03.056X-Ray Interactions: Photoabsorption, Scattering, Transmission, and Reflection at E = 50-30,000 eV, Z = 1-92
. At. Data Nucl. Data Tables 54, 181 (1993). https://doi.org/10.1006/adnd.1993.1013Thermal detectors as X-ray spectrometers
. J. Appl. Phys. 56, 1257 (1984). https://doi.org/10.1063/1.334129Optical response of Al/Ti bilayer transition edge sensors
. Chin. Phys. B 23,Fabrication and characterization of Al-Mn superconducting films for applications in TES bolometers
. Chin. Phys. B 30,Development of series SQUID array with on-chip filter for TES detector
. Chin. Phys. B 31, 28504 (2022). https://doi.org/10.1088/1674-1056/ac2b91Transition edge sensor-based detector: from X-ray to γ-ray
. Nucl. Sci. Tech. 33, 84 (2022). https://doi.org/10.1007/s41365-022-01071-5Elemental composition x-ray fluorescence analysis with a TES-based high-resolution x-ray spectrometer
. Chin. Phys. B 32,Data Acquisition and Analysis for a TES-Based X-Ray Spectrometer
. J. Low Temp. Phys. 209 1017–1023 (2022). https://doi.org/10.1007/s10909-022-02744-6A calibration system adopting 405 nm laser for soft x-ray TESs
. Proc. SPIE 12169 369–374 (2022). https://doi.org/10.1117/12.2622281Study on data processing for x-ray spectrometer based on microcalorimeter
. Proc. SPIE 12169 2051–2057 (2022). https://doi.org/10.1117/12.2625121TES-based x-ray spectrometer developed for SHINE project
. Proc. SPIE 12169, 381–385 (2022). https://doi.org/10.1117/12.2622288Development of basic theory and application of cryogenic X-ray spectrometer in light sources and X-ray satellite
. Acta Phys. Sin. 70,Zhi Liu is an editorial board member for Nuclear Science and Techniques and was not involved in the editorial review, or the decision to publish this article. All authors declare that there are no competing interests.
