Anomalous Small-Angle X-ray Scattering and Its Application in the Dynamic Reconstruction of Electrochemical CO₂ Reduction Catalysts

Abstract

The electrochemical CO₂ reduction reaction (CO₂RR) is a promising approach for mitigating the greenhouse effect arising from anthropogenic CO₂ emission. Nonetheless, poor product selectivity associated with electrochemical catalysts is the main technical problem for the application of CO₂RR technology. The catalytic performance of nano-catalysts is strongly dependent on their microstructural features. Anomalous small-angle X-ray scattering (ASAXS) is one of the most effective techniques for studying nanostructural change in an operando way, especially for complex systems and mixed-element catalyst situations. Furthermore, based on the research results of ASAXS, appropriate catalyst components and nanostructures can be designed to achieve stable catalytic performance of the catalyst, promote catalytic reaction rate, or improve catalytic reaction selectivity. In this paper, the basic concept, principle, and applications in different systems of ASAXS are reviewed thoroughly. Finally, the development prospect of ASAXS in the field of electrocatalysis is prospected. It is hoped that this review will further promote ASAXS technology to play a more far-reaching impact in the field of electrocatalytic CO₂RR.

1. Introduction

Climate change is causing unprecedented impacts on the whole planet, resulting in threats to the destruction of the global ecological balance [1]. Greenhouse gas concentrations directly affect the global average temperature, in which carbon dioxide (CO₂) is the most dominant one, accounting for about two-thirds of temperature change [2]. China made a solemn commitment to the world to strive for a carbon peak by 2030 and carbon neutrality by 2060 at the 75th United Nations General Assembly in 2020. Under this double carbon target, Chinese governments and scientific research departments at all levels have introduced and formulated a series of policies and plans, which inevitably make carbon capture, carbon sequestration, and carbon transformation developing important directions for the coming decades in our country [3,4,5,6,7,8,9].

The central carbon atom in one CO₂ molecule is covalently bonded to two oxygen atoms by double bonds. Its chemical structure is centrosymmetric and linear, so CO₂ molecules have no electric dipole with strong thermodynamic stability and chemical inertness. Electrocatalytic CO₂ reduction is an effective strategy for carbon capture, storage, and utilization. In recent years, progresses have been made in the field of electrically catalyzed CO₂ conversion with the development of renewable energy and the continuous cost decline of green electricity. It is becoming possible by using electrochemical methods to convert CO₂ into high-value-added chemical products and fuels. However, great efforts should be made for designing more effective and stable catalysts to finally commercialize electrocatalytic CO₂RR technology. Understanding the activation and maintenance mechanism of electrocatalyst activity, i.e., how to make the catalysts optimize the activity and selectivity, and how to achieve long-term catalytic performance under operating conditions, is one of the fundamental challenges in catalyst design [10]. Considering the symmetry, a catalyst design principle of CO₂ hydrogenation was put forward to construct symmetry-breaking sites to activate nonpolar CO₂. One of the strategies to break the symmetry of global structures was surface modification with phase reconstruction [11].

The structure and reaction mechanism of catalysts at the electrode–electrolyte interface is critical to understanding the structure–activity relationship of catalysts [12]. The catalyst is not static but undergoes a series of structural changes during the reaction, some of which are reversible [13] but some irreversible [14]. Reconstruction of a catalyst refers to the structural changes of the catalyst during the electrochemical reaction. Dynamic reconstruction of catalysts affects their catalytic activity, selectivity, and long-term performance. Understanding the reconstruction process is crucial for obtaining the complete structure–activity relationship, and is also an important basis for catalyst design.

Catalyst reconstruction is usually caused by the applied voltage bias and the change in environmental conditions during the catalytic process [15,16,17]. For example, when the applied potential is higher than the redox potential of the elements contained in the catalyst material, it is likely to lead to the valence state change of the surface atoms, which induces the surface structure reconstruction of the catalysts [15]. There are usually two types of reconstruction behaviors observed in the catalyst. One is that the electro reduction catalyst (usually metals, alloys, other compounds, etc.) is completely transformed into new phases [18]. The other is partial reconstruction in the electro reduction catalyst, such as defects (vacancy/step, distortion/disorder structure, cavity/pore structure, atom/ion doping), interface (heterointerface, grain boundary, surface adsorbate, etc.), and shape (crystal plane, size) changes [19].

The dynamic reconstruction of an electrocatalyst is a multi-dimensional, multi-temporal, and spatial evolution process. Advanced characterization techniques, such as X-ray absorption spectroscopy, X-ray diffraction, X-ray scattering, X-ray photoelectron spectroscopy, infrared spectroscopy, Raman spectroscopy, transmission electron microscope, scanning electron microscope, scanning tunnel microscope, and atomic force microscope, have been widely used to distinguish the electrochemical reconstruction behavior of catalysts [20,21,22,23,24,25,26]. Synchrotron radiation techniques have unique significance in revealing the dynamic reconstruction process of electrocatalysts, especially in third- and fourth-generation synchrotron radiation sources. The higher intensity provided by the beam source makes the time resolution of detection higher and higher, and the state of the sample can be detected as increasingly complex, approaching the real operando. With the continuous progress of in situ synchrotron radiation technology, this will continue to promote the establishment of a complete structure–property relationship within the atomic, nano, and micron size hierarchy of electrocatalysts under working conditions [25]. The electrode–electrolyte interface reaction under electrocatalytic conditions can be monitored and studied in real-time by using the deeply penetrating and high flux of high-energy X-rays. Synchrotron radiation small angle X-ray scattering (SAXS) [27,28,29] can obtain qualitative information such as the uniformity of the electron density of the system, the monodisperse or polydisperse of particles, the clarity of the two-phase interface, the fluctuation of electron density in particles, and the fractal characteristics of the catalyst. Quantitative information can also be given, such as the particle shape, size and distribution, correlation distance, average wall thickness, particle volume fraction, specific surface area, average interfacial layer thickness, fractal dimension, molecular weight, porosity, and nanoscale periodic structure (small-angle diffraction).

The in situ electrocatalytic reaction device is highly complex; the optical path of the beam goes through the whole device and contains at least: a front window, working electrode (catalyst), product, electrolyte, diaphragm, electrolyte, contrast electrode (catalyst), back window, etc. For the scattering experiments in terms of particle size and other structural information, the accurate deduction of the scattering background becomes a big challenge. In addition, a variety of catalysts, such as multi-elemental catalysts, alloys, or composite catalysts [30,31,32], further complicate the scattering information composition, making it more difficult to obtain information about the structural changes of the target particle or target element. The anomalous small-angle scattering technique (ASAXS) is an important branch of the SAXS technique. SAXS patterns were collected at several energies near and away from the absorption edge of the target element by changing the incident beam energy [33]. Through specific calculations, the scattering information of the target element could be separated from the rest. This is suitable for in situ/operando characterization of electrochemical catalyst dynamic reconstruction [34,35,36]. It can give the distribution of different elements in the same catalytic particle, or the particle information composed of different elements, respectively, or separate the catalyst scattering information from the complex background scattering [37]. The in situ/operando ASAXS technique is essential for identifying short-lived intermediates and monitoring the surface structure and electron state of specific elements in catalysts during the electrolytic process, thus providing a comprehensive understanding of the essence of catalytic reactions. This paper briefly introduces the principle of ASAXS, reviews its application in a catalyst quantitative measurement, and finally gives prospects for the development direction of this technology.

2. Anomalous Small-Angle X-ray Scattering (ASAXS): Theory

When there are nanoscale electron density fluctuations in the nanomaterials at a range of 1~100 nm, the X-ray scattering information will locate at a small angle extent around the incident beam (typically within 5°), which is the small angle X-ray scattering (SAXS) [28]. The total SAXS intensity of an object is the sum of scattering intensities from all electrons irradiated using the incident X-ray. For a monodisperse system, after averaging all orientations of the particles, the SAXS intensity can be expressed as:

$$I(q)=I_e{\displaystyle \underset{V}{\int }{\displaystyle \underset{V}{\int }\rho \left(r_1\right)\rho \left(r_2\right)\frac{\sin qr}{qr}}}d{v}_{1}d{v}_2$$

(1)

Here, $I_e$ is the scattering intensity of an electron. $\rho \left(r_i\right)$ are the electron densities of the scattering particle at $r_i$ and $\stackrel{\rightharpoonup }{r}={\stackrel{\rightharpoonup }{r}}_1-{\stackrel{\rightharpoonup }{r}}_2$. $q$ is the modulus of the scattering vector and $q=4\pi \cdot \sin \theta /\lambda$. In a homogeneous matrix, the randomly oriented particles with similar shapes and sizes parameterized by $R$ are dispersed, and the scattered intensity is rewritten using [38]:

$$I(q)=\frac{N_p}{V_s}\Delta {\rho }^2{\displaystyle {\int }_0^{\infty }P\left(R\right)V{\left(R\right)}^{2}F}{\left(q,R\right)}^{2}S\left(q,R\right)dR$$

(2)

where $N_p$ and $V_s$ are the number of particles and the sample volume irradiated by the incident beam, respectively; $\Delta \rho$ is the electron density contrast; $P\left(R\right)$ is the size distribution function; and V(R) is the volume fraction of particles. $F\left(q,R\right)$ and $S\left(q,R\right)$ are the form factor amplitude and the structure factor, respectively. The structure factor is dependent on concentration and scatterer sizes and has to be considered carefully when scatterer sizes and distances between them are similar. SAXS cannot discriminate the different element structures (e.g., spatial distribution) in nano-catalysts containing multiple compositions when the incidence X-ray is a fixed energy.

The remarkable anomalous SAXS (i.e., ASAXS) technique stems from the X-ray energy dependence of the atomic scattering factors and gives selective access to the SAXS contributions of target elements in nano-size phases. The technique utilizes the energy dependence of the X-ray scattering factor, $f\left(E\right)$, of a target element when the incident X-ray energy is changed near the absorption edge of the element [39,40]:

$$f\left(E\right)=f_0+f^{\prime }\left(E\right)+i{f}^{″}\left(E\right)$$

(3)

where $f_0$ is the number of electrons of the resonant atoms. $f^{\prime }\left(E\right)$ and $f^{″}\left(E\right)$ are the anomalous dispersion corrections. The $f^{\prime }\left(E\right)$ can be obtained from the $f^{″}\left(E\right)$ value by using the Kramers–Krönig relation when it is very close and within the absorption edge region. There are two ways to extract the SAXS signal of the target element in the catalytic particles from the ASAXS data; one method is subtraction, and the other is decomposition.

2.1. Subtraction Method

The scatters in the sample can be divided into two parts, which are, respectively, target elements and remain matrix. The SAXS signals of the target elements change obviously when the energy is near the absorption edge of them, but the SAXS signals of the matrix have little changes with energy. Thus, the scattering of the target elements in scatterers can be obtained via subtraction of the SAXS performed at the two energies $E_1$ and $E_2$ [41],

$$\begin{array}{l}I\left(q,E_1\right)-I\left(q,E_2\right)=\left[\Delta {\rho }^2\left(E_1\right)-\Delta {\rho }^2\left(E_2\right)\right]\frac{N_p}{V_s}\\ \times {\displaystyle {\int }_0^{\infty }P\left(R\right)V{\left(R\right)}^{2}F}{\left(q,R\right)}^{2}S\left(q,R\right)dR\end{array}$$

(4)

where $\Delta \rho \left(E\right)=\left|n_p{f}_p\left(E\right)-n_m{f}_m\left(E\right)\right|$; $n_p$ and $n_m$ are the average atomic densities of the resonant particles and of the matrix, respectively. $f_p\left(E\right)$ and $f_m\left(E\right)$ are the atomic scattering factors of the resonant atoms and of the matrix, respectively.

For the simple case of a uniform spherical particle in a homogeneous matrix in a monodisperse system, the interference scattering can be neglected, where $S\left(q,R\right)\approx 1$ can be assumed [42,43,44]. The SAXS can be rewritten as [41]:

$$I\left(q,E_1\right)-I\left(q,E_2\right)=K{\displaystyle {\int }_0^{\infty }P\left(R\right)V{\left(R\right)}^{2}F}{\left(q,R\right)}^{2}dR$$

(5)

where $K=\left[\Delta {\rho }^2\left(E_1\right)-\Delta {\rho }^2\left(E_2\right)\right]N_p/V_s$ is the scale factor.

Generally, matrix scattering can strongly appear in the small-angle scattering region. The interference between target elements in the catalyst particles and the matrix is also attributed to the total SAXS intensity. The interference is generally omitted in the ASAXS data handling in the subtraction method. This interference could have a significant impact on the SAXS curve, implying that the qualitative and quantitative findings can be strongly degraded by the subtraction methods of ASAXS analysis [45]. When it takes into account the interference terms, the full SAXS can be written as:

$$\begin{array}{l}I\left(q,E\right)={\left|f_p\left(E\right)\right|}^2{n}_p^2{S}_{pp}\left(q\right)+{\left|f_m\left(E\right)\right|}^2{n}_m^2{S}_{mm}\left(q\right)\\ +2\mathrm{Re}\left[f_p\left(E\right)f_m^{\ast }\left(E\right)\right]n_p{n}_m{S}_{pm}\left(q\right)\end{array}$$

(6)

the partial structure factors of phases $i,j\in \left\{p,m\right\}$,

$$\begin{array}{l}{S}_{ij}\left(q\right)=e^{-iq\cdot \left(r_i-r_j\right)}{\displaystyle {\int }_V{p}_i\left(r\right)e^{-iq\cdot r}dr{\displaystyle {\int }_V{p}_j\left(r^{\prime }\right)e^{iq\cdot r^{\prime }}d{r}^{\prime }}}\\ =e^{-iq\cdot \left(r_i-r_j\right)}{A}_i\left(q\right)A_j^{\ast }\left(q\right)\end{array}$$

(7)

where $p_i\left(r\right)$ is the spatial distribution function of phase $i$, and $A_i\left(q\right)$ is the Fourier transform thereof. The catalyst–particle phase was assumed to consist of only one single spherical particle of radius $R$ [45]. When it takes the interference term into account, the true ASAXS subtracted signal is:

$$\begin{array}{l}\frac{I\left(q,E_1\right)-I\left(q,E_2\right)}{\left[{\left|f_p\left(E_1\right)\right|}^2-{\left|f_p\left(E_2\right)\right|}^2\right]n_p^2}\approx S_{pp}\left(q\right)+\frac{n_m{f}_m}{n_p{\overline{f}}_p}{S}_{pm}\left(q\right)\\ =S_{pp}\left(q\right)+\alpha S_{pm}\left(q\right)\end{array}$$

(8)

where $\alpha$ is the support/catalyst scattering ratio. For some specific systems, this scattering ratio could be larger than 1, such as for light-metal nanoparticles on a high-Z metal-oxide support. Therefore, the particle–support interference cannot be intentionally ignored in the ASAXS subtracted signal.

2.2. Decomposition Method

The decomposition method decomposes the total scattered intensity into three components: the independent normal SAXS, a cross-term related to the interaction between the target elements in catalytic particles and the matrix, and the pure resonant scattering term only originated from the target elements in catalytic particles [38]. When the change energy $E_i$ is close to but below the absorption edge of the resonant atoms, the SAXS intensities $I\left(q,E_i\right)$ can be written into three scattering functions:

$$I\left(q,E_i\right)=I_0\left(q\right)+2f^{\prime }\left(E_i\right)I_{0R}\left(q\right)+\left[{f^{\prime }}^2\left(E_i\right)+{f^{″}}^2\left(E_i\right)\right]I_R\left(q\right)$$

(9)

where $I\left(q,E_i\right)$ is the measured and corrected scattering intensity with $E_i$ energy. The function $I_0\left(q\right)$ is measured at energies far from any absorption edge of the elements in catalytic particles. The $I_R\left(q\right)$ contains scattering information on the spatial arrangement of the resonant scattering element alone. The $I_{0R}\left(q\right)$ is the cross-term between $I_0\left(q\right)$ and $I_R\left(q\right)$.

Formula (9) could be solved in many examples from the literature by using the minimum number of energies (three). It was found that the reliability is better if Equation (9) is overdetermined. The computing result can give sufficient reliability by using five energies below the adsorption edge [46]. The ASAXS analysis is presented in a catalyst containing Pt nanoparticles supported on mesoporous MCM-41 silica [47]. Decomposition and subtraction methods were compared, and it was found that the purely resonant contribution is in good approximation for higher q values (larger than 0.1 nm⁻¹), but may lead to erroneous results at smaller q values in the simple subtraction at two energies. The complete ASAXS analysis should be decomposed into three partial intensities with the scattering data measured at different energies (best five), which will give the scattering function of the resonant particles without any prior assumption or condition.

3. ASAXS Applications in Catalysts

Owing to the dependence of catalyst performance on the nanostructures of target elements, it is crucial to identify their nanostructural evolution or reconstruction during the catalytic CO₂RR. In this section, the reconstruction examples of electrocatalysts are extensively investigated for CO₂RR via SAXS and ASAXS techniques. Then, similar ASAXS studies of other candidates, such as Pt, Ni, Au, Cu, etc., are introduced, highlighting the universal nanostructural changes and available impacts on catalysis. As shown in

Table 1. Summary of different elements in catalysts investigated via ASAXS.

Elements	Catalysts	Types	Structure Information	Ref.
Pt (L3-edge: 11,564 eV)	Pt25Cu75	Fuel cell	Pt-Cu core/Pt shell structure; Cu dissolution.	[34]
	Pt/carbon Pt3Co Pt/carbon black	Polymer electrolyte fuel cell	Dissolved smaller particles promote the growth of larger particles; Co size distribution.	[35] [36] [48]
	Pt/carbon	Fuel cell	Pt particle size distribution; Pt core/oxide shell structure.	[43]
	Pt/ordered mesoporous silica Pt/LZ-M-5	Catalytic application	The ASAXS subtraction method.	[47] [49]
	Pt/ATO (Sb-SnO2 nanopowder)	Oxygen reduction reaction (ORR)	Pt size distribution; Particle coarsening and Ostwald ripening.	[50] [51]
	Pt (Ni)/TiO2 Pt (Ni)/TiO2-C	Direct methanol fuel cell	Small Pt particle size; Ni particle size.	[52]
	PtNix (x = 3, 6)	Polymer electrolyte membrane fuel cell	Particle size evolution; Surface Ni dissolution; PtNi core/Pt shell structure.	[53]
Ni (K-edge: 8333 eV)	5NiCu-17	CO2 reduction	NiCu core–shell–shell structure.	[54]
	Ni/SiO2	Catalytic application	Ni particle size distribution.	[55]
	Raney-type Ni	Industrial fields	Ni particle size distribution.	[56]
	Ni/MgAl2O4	Synthesis gas	The ASAXS decomposition method; Ni core/NiO shell structure.	[57]
Au (L3-edge: 11,919 eV)	Au/C	Fuel cell	Au particle sizes.	[30]
	Au/C		The ASAXS subtraction method.	[41]
	AgAuNP@tin-rich ITO	Glucose oxidation	Au spatial distribution inside core.	[58]
Cu (K-edge: 8979 eV)	Cu-microcrystalline cellulose	Generation of hydrogen	Cu2+ particle volume distribution.	[59]
	CuO(Cl0.1)/SiO2	Photocatalytic hydrogen generation	The ASAXS decomposition method; Cu size distribution.	[60]
Co (K-edge: 7709 eV)	CoMoP/Al2O3	Hydrotreatment	Two analytic methods for ASAXS; shape parameter.	[38]
	Co, Co3O4	Fischer–Tropsch Synthesis	Size distribution of two kinds of Co particle sizes; the ASAXS subtraction method.	[61]
Ru (K-edge: 22,117 eV) Se (K-dege: 12,658 eV)	RuSex/C	ORR	Ru and Se mean sizes; their size distributions; the ASAXS decomposition method.	[46] [62]
Ce (M-edge: 884 eV)	CeO2, CeO2/Au	Oxidation reaction	Bimodal distribution of Ce particle size.	[63]

, the recent development of the ASAXS technique by probing the unique structure information obtained from different elements in catalysts is summarized.

3.1. CO₂ Reduction

Recently, numerous electrocatalysts have been developed for the CO₂RR to convert CO₂ into high-value-added chemical products or fuels. Generally, an effective method is size engineering, which can boost the CO₂RR performance. The surface-to-volume ratio can result in a greater surface area and higher density of active sites with a decrease in particle size. In addition, the catalytic activity can also be significantly affected by the geometric and size effects. Therefore, it is important to understand the size–activity relationship for the development of electrocatalysts in the CO₂RR [64]. SAXS is a very key technique for investigating the microstructures of catalysts and electrolytes in CO₂RR, which is the foundation of the ASAXS technique. Here, we start with the introduction of SAXS technique applications.

The SAXS technique was performed to investigate the microstructures of the synthetic Zn-MOF material for highly efficient electrochemical conversion of CO₂ to CH₄ [65]. It displayed that the morphology varied from rod-like to sheet-like and to spherical, with increasing ZnCl₂ mass fractions in the solution. From the SEM image (Figure 1a), it can be found that the Zn-MOF synthesized at mass fractions of ZnCl₂ = 0.38 had a smooth surface with a length of about 10 μm. The pair distance distribution function (Figure 1b) resolved from the SAXS data gave the shape and size information based on the SEM image. The conductivity of the electrolyte and double-layer capacitance at the interface of the electrode/ionic liquids should be interrelated to the microstructure of the electrolyte. SAXS studied the ternary electrolytes of electrocatalytic CO₂RR, consisting of ionic liquid, organic solvent, and H₂O [66]. According to the scattering intensity curve which migrated to a high q range and Braggs law $d=2\pi /q$, it was illustrated that the size of the ionic liquid aggregates decreased with increasing H₂O content (H₂O < 5 wt%). The weaker cation–anion interactions resulted in the higher solubility of the CO₂ in the electrolyte because of the small size of the ionic liquids aggregates. The SAXS intensity moved slightly to a lower q range at a higher H₂O content (H₂O > 5 wt%), which suggested that the aggregate size increased with increasing H₂O content. Using the liquid crystal cubic phase formed with monoolein and Myverol as a matrix, a catalytic complex of nickel (II) and 1-hexadecyl-1,4,8,11-tetraazacyclotetradecane was supported for CO₂RR. The structures of the monoolein and Myverol cubic phases were determined using SAXS [67]. Based on the ratio of repetition distance, Ia3d and Pn3m structures were distinguished, enabling us to monitor the CO₂ concentration in the aqueous solution corresponding to the first- and second-order diffraction rays. The first-order reflection of the Ia3d structure corresponded to the d [211] and d [220] diffraction planes, while the first-order reflection of the Pn3m structure corresponded to the d [110] and d [111] diffraction planes.

Typically, the SAXS patterns exhibit three principles: (i) Porod, (ii) mass fractal, and (iii) the Guinier regime. The Porod curve (high q range) is described by the q⁻⁴ decay of the scattering intensity from the surface of the primary particles. The fractal regime (at middle q-values) is characterized by q^−Df decay of the scattered intensity. When 0 < D_f < 3, it corresponds with mass fractal; additionally, 3 < D_f < 4 is for a surface fractal. The Guinier curve at low q depicts the mass fractal agglomerates of aggregates. That the value of the fractal dimension is lower means the catalytic nanostructure has a more open structure, which is important to enhance the performance for the catalytic application. A series of nickel–cerium-promoted mesoporous silica (SBA-15) catalysts modified with different loadings of yttrium were tested in a CO₂ methanation reaction and characterized via SAXS [68]. The Porod slope of the SBA-15 was significantly lower than that of the loaded samples, which supported a significant change in the catalyst morphology from rough surfaces to more fractal properties (Figure 1c). Three strong diffraction peaks, 100, 110, and 200, appeared, with the strongest peak being at q = 0.06 Å⁻¹. These peaks were attributed to the highly ordered 2D hexagonal structure (p6mm) of SBA-15. The SAXS intensity increased after SBA-15 was modified with Ni-Ce and Ni-Ce-Y. The scattering increase could be explained by either the partial pore destruction of 15Ni10Ce/SBA-15 and 15Ni10Ce10Y/SBA-15 catalysts, or an increased contrast between the filling pore and the matrix. The fractality of Au-Bi₂O₃ was measured via SAXS and the D_f was 1.80 (Figure 1d) [69]. The Au-Bi₂O₃ material exhibited exceptional selectivity towards formate (97%) in CO₂RR.

Catalytic materials are often composed of different elements, and the performance of different elements in the catalytic process is different. Normal SAXS is difficult to distinguish the contributions of different components. The ASAXS technique can be used to strengthen the contribution of one or more components to SAXS so that the interested component can be separated from the composites. In general, it is not easy to subtract the background scattering of the porous substrate due to the complex phase systems (metal, substrate, and void) of metal catalysts. To better understand the effect of the microstructure on the catalytic performance and prove the potential of ASAXS technology in the multi-scale organization of catalysts, Franziska Emmerling et al. investigated a NiCu core–shell–shell structure of nanoparticles (NPs) and their chemical composition for catalytic reduction of CO₂ to CO [54].

The final NP dispersion was analyzed via SAXS using the Weibull distribution, and NP dispersion with a diameter of 16.8 nm and a distribution of 14% was obtained (Figure 2a). The SAXS data were accurately fitted with a core–shell–shell structure like the model in the inset. The SAXS data of NiCu NPs prepared with different amounts of OAm were shown in Figure 2b, and the data were fitted with a Schulz–Zimm fit (blue line). According to the SAXS data, the NP size increased from 6 nm to 15 nm when the OAm/metal ratio increased from 5 to 90 (Figure 2c). Moreover, the NiCu SAXS data of different TOP contents were fitted with a Schulz–Zimm function [54], and the size of polydispersity NPs decreased proportionally from 30 nm to 6 nm. Three ASAXS intensities of 17-nm NiCu NPs with a 5:1 ratio of Ni:Cu are shown in Figure 2d. These ASAXS curves were measured below the Ni-K edge with very similar shapes and slight intensity shifts. The decomposition method was used to calculate the Ni resonant scattering curve (red dots). The Ni ASAXS intensity included the spatial distribution of Ni atoms and a structural model of a NiCu alloyed core with a Ni-enriched shell and a NiO outer shell. This model consisted of a LogNorm distributed core and two shells (the inner shell 1 and the outer shell 2). Core radii and shell thicknesses of NiCu NPs derived from ASAXS measurements are shown in

Table 2. Core radii and shell thicknesses of NiCu nanoparticles derived from ASAXS measurements at 8330 eV [54].

8330 eV	R Core [nm]	Shell 1 [nm]	Shell 2 [nm]	R Total [nm]
5NiCu-17	5.40	1.53	2.46	9.39
3NiCu-17	5.63	1.94	2.56	10.24
1NiCu-19	7.00	0.96	2.49	10.45
1NiCu-15	6.66	0.41	1.66	8.73
6NiCu-13	4.24	1.23	1.27	6.74
15NiCu-8	3.17	0.40	1.20	4.77
2NiCu-8	2.27	0.70	1.47	4.44

.

Two different structural model fits approximated on the ASAXS resonant curve of 5NiCu-17 obtained from the scattering at the photon energies 8004 eV, 8304 eV, and 8330 eV (Figure 2e). The two structural model fits are distinguished only by the occurrence of Ni in the core. The first structural model fit (red curve) is based on the core–shell–shell model. The second structural model fit (blue curve) is based on a model with a Ni core resulting in a hollow-shell–shell model. Moreover, the SiO₂-supported NiCu core–shell NPs showed outstanding CO selectivity (>99%) in the catalytic CO₂RR (Figure 2f).

3.2. ASAXS Applications in Other Catalyst Fields

The ASAXS technique is very suitable to observe the in situ evolution of particle size and geometric surface area distributions with changing electrochemical conditions (e.g., time, potential cycling). The ASAXS applications below are categorized with different target elements in the catalysts.

3.2.1. Pt Element

Pt-based catalysts have become a kind of significant activity material for fuel cells or supported-metal catalysts [34,50,70,71]. They can be classified into three main groups: Pt, Pt alloys, and core–shell Pt structure. It is found that activity and stability are related to particle structure and geometry. The ASAXS technique has the unique advantage of complex systems such as multi-metal alloys and core–shell nanostructures. The ASAXS intensities and the atomic form factor are strongly changed when the incident X-ray energy is near the Pt-L₃ absorption edge (11,564 eV).

Despite a great deal of research using ASAXS, to our knowledge, the first paper to apply this technique to a catalytic system is about Pt electrocatalysts supported on porous carbon [42]. Pt/C catalysts are widely welcomed for their high activity and relative stability, but the degradation of the electrochemically active surface area of Pt nanoparticles hinders their cost-effective realization. Gilbert et al. demonstrated that the particle size distributions of Pt electrocatalysts underwent periodic intervals during the different potential cycling [35]. Changes in the particle size distribution, mean diameter, and geometric surface area determined the mechanism behind the coarsening of Pt nanoparticles in the aqueous environment. The ASAXS data were analyzed by using the Modeling II macro of Irean software. The Pt scattering intensities in the range of 0.02–0.35 Å⁻¹ at E₁ = 11.38 keV and E₂ = 11.53 keV were fitted using a log-normal distribution, assuming the Pt nanoparticles were polydisperse spheres and scattered independently (Figure 3a). The log-normal function is given using the equation:

$$p\left(r\right)=\frac{1}{\sqrt{2\pi }\sigma r}\exp \left(-\frac{{\left(\ln r-\mu \right)}^2}{2{\sigma }^2}\right)$$

(10)

where $\sigma$ and $\mu$ are the standard deviation and the mean particle size, respectively. It was found that the loss mechanism of the Pt surface area was mainly due to the preferential dissolution or loss of the smallest particles in the first 80 potential cycles. Different degrees of dissolved species redeposited on existing particles, resulting in particle growth, depending on the potential distribution (Figure 3b). This research group also investigated the evolution of Pt nanoparticles of Pt/C catalysts in an aqueous electrolyte environment. Stagnant electrolytes, flow electrolytes, and high-temperature flow electrolytes were used to reveal the different degradation trends in polymer electrolyte membrane fuel cells and aqueous environments. The aim was to determine the correlation of aqueous measurements to the stability of Pt catalysts in fuel cells [48]. It showed that the Pt nanoparticle surface area loss was controlled by Pt dissolution, which was particle size dependent. The loss of dissolved Pt was redeposited onto larger particles. In addition, the Pt nanoparticle of the Pt/C catalyst also exhibited an increase during the oxidation process at about 1.1 V vs. Ag/AgCl in an electrochemical cell. A shell model of oxidized particles was deduced as having a Pt core and a 1-nm oxide surface layer [43].

To improve the start/stop stability of Pt-based catalysts, different support materials, such as mesoporous MCM-41 silica [47], zeolite LZ-M-5 [49], ATO [50,51] and TiO₂ [52], have been studied as alternatives to the usual carbon supports. ATO provides enough conductivity for the Pt catalyst. The Pt/ATO catalysts showed good stability in basic electrochemical tests. ASAXS was used to detect the Pt particle size distributions directly from the net Pt scattering curves [50]. The 5 wt% Pt/ATO catalyst had the narrowest size distribution. With Pt loading increasing on the ATO support, the average particle size and size distribution increased. However, for the same 10 wt% Pt loading, the average Pt nanoparticle size on the pre-reduced ATO was significantly smaller than the Pt particles deposited on the untreated ATO. Schmidt et al. obtained the Pt nanoparticle size distributions from scattering curves as a function of degradation potential cycling [51]. The net Pt ASAXS intensities were fitted to the standard model of spherical particles, which was extended by a scattering interference model between spherical Pt nanoparticles and spherical support nanoparticles, commonly used log-normal Pt particle size distributions:

$$p\left(D\right)=\frac{1}{\sqrt{2\pi }\sigma \left(D/2\right)}\exp \left(-\frac{{\left(\log \left(D/2\mu \right)\right)}^2}{2{\sigma }^2}\right)$$

(11)

where D denotes the spherical Pt particle diameter. Therefore, for further analysis, the surface area-weighted size distribution was:

$$P_A\left(D\right)=\frac{\pi D^2}{〈A〉}p\left(D\right)$$

(12)

where $〈A〉={\displaystyle \int \pi D^2}p\left(D\right)dD$ is the average surface area per Pt particle. Because the ASAXS analysis and data fitting process are complex and affected by many different parameters, it is difficult to estimate the error of the obtained quantity. Subsequently, the four types of degradation mechanisms that occur during high-potential cycling were discussed for the catalysts: Pt mass loss due to dissolution in the electrolyte, electrochemical Ostwald ripening, support corrosion, and Pt agglomeration, respectively. The change in particle size distribution in the direction of a larger particle size during the degradation process can be directly attributed to the electrochemical Ostwald ripening (Figure 3c).

Pt m nanoparticle alloys, such as PtCo [36], PtRh [42], PtNi [53], and PtCu [34,71] have high activity and stability as an attractive and active bimetallic nanoparticle system. Understanding the composition, morphology, and atomic-scale structure evolution of the nanoparticles during the catalytic process is crucial for more active and more stable catalysis designing. However, it is particularly difficult because of the multi-component, the large dispersion in particle size, and the unclear and likely inhomogeneous spatial distribution of the constituents in the nanoparticles. The ASAXS-determined Pt particle size distribution evolutions for the Pt₃Co/C catalyst subjected to potential cycling were analyzed in the four cases of different environments and wave potential cycling [36]. It was found that the main mechanism behind the surface area loss and average diameter increase in Pt₃Co is the dissolution of the smallest Pt particles. The presence of Co accelerated the dissolution, followed by redeposition to the larger particles. The EXAFS and ASAXS data were consistent with the initial intraparticle structure, which consisted of a slightly Pt-rich shell that evolved into a structure with a heteroatomic bond slightly more dominant than the homoatomic bond (used as the potential cycles), such as a structure with a Pt-rich shell, a Co-rich underlayer, and a Pt-Co alloy core. In situ ASAXS studied the nanostructural evolution of PtNi₆ and PtNi₃ alloy catalysts [53]. The exposure of PtNi₆ nanoparticles to an acidic electrolyte resulted in a decrease in the average Ni particle size from 4.6 to 3.9 nm, and then an increase to 4.0 nm (Figure 3d). It indicated that owing to the structural transformation from Ni-rich solid solution to a Pt-rich Pt-Ni alloy phase, the initial Ni surface dissolved rapidly and was followed by lattice expansion. In contrast, the Pt size distribution was still unchanged. However, over the first 1000 potential cycles, both Pt and Ni species’ diameters increased (Figure 3e). The Pt and Ni sizes increased from a value of 3.4 to 3.8 nm and from 3.0 to 3.5 nm, respectively. The Ni size was reduced to a 3.1 nm final value during an additional 500 cycles. The difference in Pt and Ni sizes indicated that Pt and Ni rearranged from initially chemically homogeneous alloy nanoparticles to nanoparticles with a core–shell structure. A better understanding of the structural transformations of PtNi₆ alloy nanoparticles was characterized by Pt atoms dissolved in the Ni matrix; Subsequent electrochemical cycling led to the surface dissolution of Ni and a rapid rearrangement of Pt and residual Ni atoms until an ordered Pt₃Ni alloy structure type was formed (Figure 3f(i)). In addition, electrochemical treatment of the original disordered PtNi₃ nanoparticles stimulated the rapid dissolution of the Ni surface, resulting in a slightly smaller size of the PtNi core/Pt shell nanoparticles (Figure 3f(ii)).

The class of Cu-rich Pt₂₅Cu₇₅ alloy nanoparticle electrocatalysts was investigated via ASAXS [34]. The ASAXS curves were fitted with a monodisperse ensemble of randomly oriented spherical Pt and Cu particles and log-normal distribution function:

$$p\left(r\right)=\exp \left[-\frac{1}{2}\frac{{\left(\ln \frac{r}{R_0}\right)}^2}{{\sigma }^2}\right]\frac{\exp \left(-\frac{1}{2}{\sigma }^2\right)}{r\sigma \sqrt{2\pi }}$$

(13)

where $R_0$ and $\sigma$ are the mean particle radius of the distribution and the log-normal dispersion or width in size, respectively. Although there is no prior argument as to why the experimental SAXS datum of lognormal distribution fittings is superior to Gaussian functions, they often provide a good fit in many cases. For example, the Pt ASAXS intensities fitted well with log-normal distributions. Electrochemical Cu dissolution and dealloying processes of an electrocatalyst precursor in acidic electrolytes showed the larger q shift in the Cu ASAXS scattering curve, indicating that the Cu atom removal from the carbon-supported Pt₂₅Cu₇₅ alloy particle surface suggested the formation of a Pt-Cu core surrounded by a Pt enriched Pt shell.

3.2.2. Ni Element

Ni-based catalysts have attracted much attention due to their controllable morphology and high dispersion, which are widely used in energy production and environmental remediation. The properties of Ni-based catalysts are mainly governed by compositions, structural morphology, and surface area. ASAXS was recently used to determine the particle size distributions of Pt(Ni)/TiO₂ [52], PtNi [53], Ni/SiO₂ [55], NiAl [56], and Ni/MgAl₂O₄ [57]. Berg Rasmussen et al. [55] studied a Ni/SiO₂ tablet sample with ASAXS. The Ni nanoparticles were highly dispersed on the porous silica and assumed to be solid spheres with a free-form determination of the size distribution or with a log-normal distribution at the low radius part. The best Ni ASAXS fits were obtained by two distributions (Figure 4a). The main difference between the two size distributions was the tail of larger particles observed using the free-form determination (Figure 4b). It illustrated from the obtained fitting that this tail was very meaningful and showed the importance of not assuming a specific form of size distribution.

The “pure resonant term” of ASAXS and the spatial distribution of Ni nanoparticles in a NiAl alloy catalyst was presented [56]. Log-normal particle size distribution of the Raney-type catalyst was obtained using the equation:

$$p\left(r\right)=\frac{1}{\sqrt{2\pi }\sigma r}\exp \left(-\frac{{\ln }^2\frac{r}{R_0}}{2{\sigma }^2}\right)$$

(14)

The above model involves the correlation of adjacent globular compact units (radius r), which are almost linearly displaced. The best-fit result is a size distribution with a characteristic radius of 20.8 Å ($\sigma$ = 0.24). The number of connected spheres is two, and their distance is 41.6 Å, which is approximately twice the characteristic radius value. It can be assumed that Ni exists in the NiAl alloy catalyst in the form of particle pairs. In addition, the main difference in the formulas 10, 11, 13, and 14 is the different forms of numerators in the exponential functions.

The Decomposition Method of ASAXS was used to resolve the morphology of the fresh and reduced nickel/spinel catalyst particles [57]. Using a spherical core–shell particle model, the separated Ni ASAXS intensity was analyzed by using the SASfit software. The particle size distribution was a log-normal distribution of Equation (10). The particle shapes were approximated with spherical symmetry for ASAXS analysis. The separated ASAXS intensities of the fresh Ni/NiO nanoparticles and the same nanoparticles after reduction at 773 K were displayed together with the best fits (Figure 4c). The parameters obtained using the least-squares fit showed that the average radius of the fresh Ni/NiO core–shell particles was 3.7 ± 2.0 nm. The ASAXS analysis of Ni nanoparticles at 773 K showed that the nanoparticles had a uniform electron density. It was concluded that the NiO shell layer had been completely reduced and the nanoparticle was composed of pure Ni (Figure 4c inset).

3.2.3. Au Element

Au-based catalysts have great application value due to their unique activity and selectivity for many kinds of catalytic reactions. The special catalytic performance of supported Au nanoparticles mainly depends on the nanoparticle morphology, namely, size, shape, thickness, and carrier effect [72]. The ASAXS technique was able to contribute primary quantitative information on the Au-based catalysts [41,58,73,74], and is now considered an important technique in this field. Stefano Polizzi et al. investigated the Au (0.2 wt%)/C sample, finding that the Au size distribution can be easily obtained by using an analytical function to fit the experimental data through an optimization procedure [41]. The ASAXS contrast variation was performed at different energies near the Au L_III = 11,918 eV absorption edges. In the Au/C samples studied, a bimodal distribution was used to describe the scattering intensity,

$$p\left(r\right)=\left[p_1\left(r\right)+K{p}_2\left(r\right)\right]/\left(1+K\right)$$

(15)

where K is the ratio of the total number of smaller particles to that of the larger ones and $p_i\left(r\right)$ are the normalized Schulz distributions.

$$p\left(r\right)=\frac{1}{\mathsf{\Gamma }\left(z+1\right)}{\left(\frac{z+1}{〈r〉}\right)}^{\left(z+1\right)}{r}^z\exp \left(-\frac{z+1}{〈r〉}r\right)$$

(16)

where z (with z > −1) is the form parameter of the distribution. The separated ASAXS data of the Au/C catalyst could be obtained and fitted with a bimodal particle size Schulz distribution [73,74]. The ASAXS analysis of the Au/C catalyst was able to discriminate two different populations of particles: the 14-nm larger ones and approximately 2-nm clusters [74].

ASAXS was used to investigate the spatial distribution of the Au precursor inside the Au and Ag-loaded P4VP core of the PS₁₁₁-b-P4VP₉₆ micelles [58]. The scattering contribution only related to the Au-loaded domain and was given by subtracting the scattering curve (E₃ = 11,915 eV) from another curve (E₁ = 11,557 eV) (Figure 5a). The Au contribution can be fitted with a spherical scattering model (Figure 5b). ASAXS proved the uniform distribution of the Au precursor throughout the silver-loaded P4VP micellar core.

3.2.4. Cu Element

Copper (Cu), a low-cost transition metal, has attracted a lot of attention from researchers in recent years. Due to the high catalytic activity of Cu nanoparticles, these nanoparticles are ideal candidates for electrochemical catalysts, especially in CO₂RR. Cu-based catalysts were commonly used to change electrochemical CO₂ reduction to CO, methane (CH₄), ethylene (C₂H₄), and ethanol (C₂H₅OH). The ASAXS technique has seldom been used to investigate Cu nanoparticles in the CO₂RR catalytic field. However, this technique was found to be capable of successfully handling a PtCu alloy [34], CuO/microcrystalline cellulose [59], and CuO/SiO₂ catalyst [60]. The CuO/microcrystalline cellulose nanocomposite was measured at three or five energies below the Cu K absorption edge (8979 eV) via ASAXS [59]. The size distribution of the CuO particles was determined using a Monte Carlo fitting [75]. The Monte Carlo method was not assumed with any predetermined distribution shape and was not very sensitive to data noise. The Cu nanoparticle structures played an important role in catalysis. ASAXS could provide deeper insights into the Cu particle morphology and the Cu spatial distribution in the CuO(Cl)/SiO₂ catalyst [60]. ASAXS was performed at five energies, 353, 127, 47.5, 17.5 and 6.5 eV, below the Cu K-edge (Figure 6a). The angularly independent incoherent scattering at a high q range was used as the background. The ASAXS intensities subtracted from the background are shown in Figure 6b. The ASAXS data are represented by stars and crosses, and the fitting result displayed as a line is overlaid in the magnification inset. It can be seen that the scattering intensity gradually decreases as the energy approaches the Cu absorption edge. The ASAXS curve at 127 eV below the Cu edge could be explained with three different contributions (Figure 6c). The first contribution (gray) was the Porod curve at large q-values, and the background showed no resonant effect. The second contribution (blue) was the support nanostructure, which was simulated with a broad size distribution of spherical objects. The third one (red) was related to Cu nanoparticles of about 1.4 nm, with a logarithmic size distribution shown in the inset of Figure 6c.

3.2.5. Co, Se, Mo, Ru, Rh, Pd, Ce Elements

A catalyst with a Pt:Co atomic ratio of 3:1 has been selected, which has been proven to have the highest oxygen reduction reaction activity and activity stability [36]. The mean diameter (4.9 nm) of the Co distribution was obtained from ASAXS analysis of the catalytic Pt₃Co/C powder at the Co K absorption edge (7709 eV). The increase in the average diameters of Pt and Co measured using ASAXS showed the same conditional dependence and was linearly related to an increase in the Co average diameter that was approximately twice as large as Pt. S. Humbert et al. investigated the microstructure and size distribution of two cobalt-based nanoparticles and aggregates [61]. The ASAXS intensity could be fitted with only one size distribution function because of two inflections of the scattered signal (Figure 7a). It indicated two kinds of Co nanoparticles; one was crystalline particles at a small scale and the other was aggregates at a large scale. The ASAXS fit showed that the high q-value part curve is essentially due to the contribution of the small-size particles, whereas the low q-value part curve is due to the aggregate contribution. This conclusion was further confirmed through the ASAXS fitting (based on Equation (10)). S. Humbert’s experimental group also synthesized a series of CoMoP/Al₂O₃ catalysts with a multi-scale organization of the active phase [38]. ASAXS curves were obtained from the various activation and the different sulfidation techniques to target a specific chemical element, Mo, in these cases. The scattered intensity was measured at different energies slightly below the Mo K absorption edge (20,000 eV). According to the shape of the scattering curves, two kinds of Mo objects had to be considered. Size distributions of slabs and aggregates were represented by log-normal distributions (Equation (10)). The size distribution of slab stacks simulated by discs and the size distribution of slab aggregates simulated by spheroid concerned only the axes. Based on fitting the ASAXS curves, it concluded that the slab morphology (length and stacking) had been affected significantly by the additives, and the liquid phase sulfidation had a dispersing effect compared to the gas phase one.

The research team of Sebastian Fiechter and Sylvio Haas revealed the nanostructure of carbon-supported, Ru-Se-based catalysts in fuel cells via ASAXS [46,62]. The scattering intensities were collected at five energies below the Se (12,658 eV) and the Ru (22,117 eV) K absorption edges. The scattering curve could be decomposed magically into the partial scattering intensities of Ru, Se, and the carbon support (“Black Pearls”). Their sum fitted the measured curve very well (Figure 7b), or the five scattering intensities could be resolved with three scattering contributions, I₀(q), I_0R(q), and I_R(q), by solving linear Equation (9) (Figure 7c). Figure 7d shows the volume-weighted size distributions of the three structural components. The structural model of catalytic nanoparticles had been derived (Figure 7e). This model indicated that the surface of nearly 2.5-nm spherical Ru particles was decorated with an approximately 0.6-nm Se-containing cluster.

Grunwaldt et al. [63] demonstrated that ASAXS measurements of soft X-rays provided structural information on the CeO₂ deposit at the Ce M-edge (884 eV). The ASAXS scattering intensity appeared on two shoulders at 0.021 Å⁻¹ and 0.04 Å⁻¹ (Figure 7f), and the intensity was less pronounced after annealing (Figure 7g). The distance distribution functions showed bimodal distributions, and the two gyration radii of the bimodal distribution are 70 Å⁻¹ and 170 Å⁻¹, respectively. The bimodal distribution was maintained after annealing (Figure 7g), which indicated no change in the CeO₂ dimensions. In addition, the Rh [41] and Pd [41,74] elements were also investigated through the ASAXS technique. ASAXS scattering cross sections can be fitted by the size distribution for the Rh catalyst particle near the Rh-K edge (23,220 eV). The main results of this ASAXS analysis in the Pd (2.9 wt%)/SiO₂ sample were: (i) two particle populations approximately 3 and 13 nm; (ii) smaller particles accounted for approximately 80% of the whole metal; (iii) the Pd content was 2.7 wt%.

4. Summary and Outlook

The catalyst reconstruction and its influencing factors in electrocatalytic CO₂RRs have not been understood in detail; it is hoped that a new perspective for fully understanding the reconstruction behavior of electrocatalysts under operando conditions and controllably tuning the catalysts to exhibit higher activity and stability can be implemented. Although we focus on the electrocatalytic CO₂RRs, ASAXS is actually universal for the dynamic reconstruction process investigation of all kinds of catalysts, such as photocatalytic CO₂RRs, catalysts in a hydrogen evolution reaction (HER), and an oxygen evolution reaction (OER) in electrolysis water, and so on. Common catalysts for HER and OER include metals such as platinum, palladium, nickel, molybdenum, as well as their oxides, sulfides, and nitrides. The electrochemical processes of HER and OER catalyst reactions involve complex structural changes, including changes in surface active site, oxidation state, crystal structure, and so on. So, the nanostructural design and synthesis of efficient HER and OER catalysts are key to improving the energy efficiency of water electrolysis for hydrogen and oxygen production. Therefore, there is an urgent need for multi-dimensional in situ characterization techniques to detect the size, spatial distribution, active site spatial structure, electronic structure, and surface reaction intermediate group changes in CO₂ reduction, HER, and OER catalysts under operando conditions at multiple time scales and multiple spatial scales, to further reveal the complete structure–activity relationship.

The ASAXS technique is nowadays suitable for the characterization of catalysts and is well suitable for a multi-phase system (voids, multi-metal, matrix, and nanoparticles), and its advantage is that it can characterize objects between one nanometer and several hundred nanometers in size. However, the ASAXS technique is nowadays rarely used to determine the structural information of the active phases of CO₂RR, HER, and OER catalysts. We reviewed the recent development of in situ SAXS(ASAXS) applied in CO₂-related catalysis and discussed some other element catalytic applications of ASAXS. In situ ASAXS can probe the nanoparticle size and size distribution, nanostructure, and structural evolution under real reaction conditions to finally rationalize the design and performance of electrocatalysts. A pressing issue for ASAXS is the difficulty of data analysis. The advanced analysis programs can perform more quantitative determinations of the ASAXS and collect valuable structural information such as nanoparticle size, size distribution, and hierarchical structure. We hope that these applications of ASAXS techniques will help to further understand the modes of CO₂RR, HER, and OER catalysts, evaluate the effectiveness of these catalysts, optimize the nanostructural design of catalytic materials, fine-tune the structure, and master the mechanism of structure tuning to improve the catalytic performance.

With the development of new generation (e.g., 4th) synchrotron X-ray sources with higher flux, stronger coherence, and better time resolution, ASAXS can be used to probe new frontiers in electrocatalysts and structural reconstruction to boost the knowledge of the structure–property relations for the design of high-activity CO₂RR electrocatalysts. The abnormal USAXS–SAXS–WAXS combined technology will be gradually established by using the new beam source to be built in Beijing. This combined technique would be called A(USAXS–SAXS–WAXS) technology, and it would be very important to further improve the spatial and temporal resolution and detection ability of multi-dimensional in situ synchrotron radiation technology for establishing the complete structure–activity relationship of catalysts.