Skip to main content
NCBI home page
ACS AuthorChoice logoLink to ACS AuthorChoice
. 2026 May 19;148(21):22036–22044. doi: 10.1021/jacs.6c03877

Torsional Ordering as a Prerequisite for Zeolite Crystallization Revealed by X‑ray Emission Spectroscopy

Kakeru Ninomiya †,‡,§, Ralph Ugalino ∥,⊥, Koki Itamoto ‡, Zhong Yin †, Hisao Kiuchi #, Yoshihisa Harada #, Hideyuki Magara §,, Kenji Tsuda §,, Masami Terauchi §,, Toshiyuki Yokoi ∇, Maiko Nishibori †,‡,§,*
PMCID: PMC13244451  PMID: 42156329

Abstract

Understanding how ordered crystals emerge from disordered precursors is a grand challenge in materials science. Resolving this question is critical for tailoring the properties of network materials, particularly zeolitic silicates, which are indispensable in energy and environmental catalysis. This challenge has been particularly acute as the key three-dimensional (3D) ordering of the silicate network has remained experimentally inaccessible. However, detailed knowledge of crystallization would enable tailoring of key properties for improved applications. Here, we provide direct observation of this process using O 1s X-ray emission spectroscopy (XES) as a spectroscopic ruler for the 3D network geometry. Supported by ab initio simulations, our results reveal the systematic evolution of the O–Si–O–Si torsion angles during the crystallization of an MWW-type zeolite. We discovered that the progressive ordering of these angles into a staggered conformation is a decisive early-stage event that precedes the formation of the crystalline framework. This work establishes O 1s XES as a powerful tool for probing local network topology, enabling quantitative access to key geometric parameters of 3D silicate networks, including bond distances, bond angles, and torsion angles. This reveals a fundamental principle: topological ordering is a distinct, precursorial event that occurs prior to the emergence of long-range crystalline order. This “topology-first” insight fundamentally advances the rational design of zeolites and other complex network materials with atomic-level precision.

graphic file with name ja6c03877_0010.jpg

graphic file with name ja6c03877_0008.jpg

Introduction

Understanding how disordered matter transforms into ordered crystalline structures is a fundamental challenge in science, underpinning fields from materials engineering to biomineralization. In the synthesis of zeolitescrystalline porous materials vital for catalysis and separations − this challenge is particularly acute. For decades, zeolite synthesis has largely been advanced through empirical hydrothermal exploration, leaving a critical bottleneck: the atomistic structural evolution from the amorphous precursor gel to the final crystalline framework remains incompletely understood. − Transitioning from chance discovery to predictive design demands atomistic insight into nucleation and growth; as such, resolving these events remains a preeminent challenge in materials chemistry.

Elucidating this amorphous-to-crystalline transformation is fundamentally a challenge of three-dimensional (3D) characterization. While established techniques have provided key insight, they remain insensitive to the geometric parameters that define 3D network topology. − Methods such as X-ray diffraction (XRD), electron microscopy, and scattering (SAXS/total scattering) effectively determine long-range order, particle size, and average atomic distances. Pair distribution function (PDF) analysis, for instance, has successfully tracked the evolution of Si–O and Si–Si bond distances. − However, these powerful techniques share a fundamental limitation: they collapse complex 3D architectures into one-dimensional distance information. By radially averaging the structure, they become inherently insensitive to the parameters that define 3D network topologymost critically, the O–Si–O–Si torsion angles. This information loss has precluded the direct observation of how a disordered precursor network organizes into a specific zeolite framework, leaving the atomic origins of crystallization elusive.

Here, we overcome this long-standing limitation by employing O 1s XES as a direct, element-selective probe of local 3D network geometry. XES is known to be highly sensitive to the electronic structure of complex materials, especially when combined with theory. − Building on this foundation, we establish O 1s X-ray emission spectroscopy (XES) as a quantitative spectroscopic ruler for the O–Si–O–Si torsion angle. This capability allows us to track the evolution of network topology throughout the entire crystallization process. By integrating XES with X-ray absorption spectroscopy (XAS) and ab initio simulations, we provide an atomic-scale view of the structural transition from an amorphous gel to a crystalline Fe-containing MWW-type zeolite. Our findings reveal that the ordering of torsion angles is a critical, previously inaccessible stage in framework formation, offering a new guiding principle for the rational design of functional porous materials with atomic-level precision.

Results and Discussion

A Direct View of 3D Network Formation by X-ray Emission Spectroscopy

XES is a recently emerging technique in probing the electronic structure of materials, particularly through their manifold of occupied electronic states. It provides similar information as X-ray photoelectron spectroscopy, albeit with the advantages of a wider workable sample base able to handle bulk-phase and insulating materials and access to additional information such as orbital symmetry and hybridization. To directly visualize the 3D network evolution during zeolite crystallization, we employed O 1s XES, a probe uniquely sensitive to the local environment of the bridging oxygen atoms that constitute the network’s backbone. Figure presents the O 1s XES spectra of Fe-MWW zeolite samples at varying hydrothermal treatment stages (HT1d, HT3d, and HT7d), alongside iron-containing amorphous SiO2 (Fe:Amo-SiO2) and hematite (α-Fe2O3) as references. The XES spectra of the Fe MWW-type zeolites closely resembled that of the amorphous SiO2 precursor, albeit with slight intensity modulations, and we opted to limit the interpretation of the O 1s edge XES spectra to a minimal Si2O7 silica model. This overall high similarity is expected because all samples are dominated by the silicate network and share the same fundamental Si–O bonding framework. This observed spectral evolution reflects subtle but systematic changes in local 3D network geometry superimposed on this common spectral background. The repeatability of these subtle spectral variations was confirmed by independent measurements of the same HT7d sample performed at a different time, which showed markedly smaller variability than the synthesis-stage-dependent spectral evolution (Figure S1 and Table S1). With the spectral profiles showing negligible resemblance with that of the α-Fe2O3 additive, as indicated by the large deviation from the experimental difference spectra (Figure a), we also opted to neglect the spectral contribution of the few oxygens bonded with Fe sites. The O 1s XES spectra exhibit a systematic transformation as the synthesis progresses; specifically, distinct features labeled C, E, and G emerge and intensify significantly over time. This spectral evolution, further isolated in the difference spectra (Figure a), serves as a definitive fingerprint of the structural ordering involved in zeolite network formation. A feature-specific quantitative analysis of these spectral changes is provided in Figure S2 and Table S2. These observations lead to a central question: what specific atomic-scale rearrangement does this spectroscopic fingerprint represent?

1.

1

Open in a new tab

O 1s XES spectra measured as a function of hydrothermal synthesis time at an incident X-ray energy of 550 eV. The spectra were normalized to a baseline of 0 and a maximum intensity of 1. As the synthesis proceeds, distinct spectral features labeled C, E, and G progressively emerge and intensify, indicating a systematic evolution of the local oxygen environment during zeolite crystallization. Reference spectra of both iron-containing and iron-free amorphous SiO2 (Fe:Amo.-SiO2 and Amo.-SiO2), together with α-FeOOH and α-Fe2O3, are shown for comparison. The close resemblance between the spectra of Fe-MWW zeolites and Fe:Amo.-SiO2, together with the negligible contribution from Fe–O species, indicates that the observed spectral evolution predominantly reflects structural changes within the silicate network.

4.

4

Open in a new tab

Direct spectroscopic evidence for the ordering of 3D network topology during zeolite crystallization. (a) Experimental O 1s XES difference spectra as a function of hydrothermal synthesis time, referenced to Fe-containing amorphous SiO2 (Fe:Amo.-SiO2), highlighting the systematic emergence of features associated with framework ordering. (b) Simulated O 1s XES difference spectra for representative torsion angles (ω = 23.1° and 37.6°), referenced to the Boltzmann-weighted simulated spectrum. The close correspondence between the experimental and simulated difference spectra indicates a progressive shift toward an ordered, staggered-rich torsional configuration. (c) Crystal structure of the MWW-type zeolite viewed along the {100} direction, illustrating the framework geometry characterized by staggered O–Si–O–Si torsion angles that give rise to the observed XES spectral features. (d) Calculated potential energy surface of the Si2O7 dimer model as a function of the Si–O–Si bond angle (θ) and the O–Si–O–Si torsion angle (ω). The relative total energies are plotted with the global minimum set to 0. The Si–O bond length (l) was fixed at 1.75 Å to maintain consistency with the XES spectral simulations. The energy landscape shows that the staggered conformation (ω = 60°) is energetically more stable than the eclipsed conformation (ω = 0°), the latter being destabilized by steric hindrance and electronic repulsion between terminal oxygen atoms. These results support a thermodynamic basis for torsional ordering during the structural evolution toward the MWW-type zeolite framework.

To decipher the structural origin of this spectral fingerprint, we performed ab initio electronic structure calculations on a minimal Si2O7 dimer model illustrated in Figure , systematically calculating the XES spectra as a function of Si–O bond length (l), Si–O–Si bond angle (θ), and O–Si–O–Si torsion angle (ω). Our simulations demonstrated that variations in bond length and Si–O–Si bond angle failed to reproduce the specific emergence of features C, E, and G without simultaneously introducing broader spectral changes, particularly around peaks A and B (see also Supporting Information, Figures S3 and S4). In contrast, increasing ω selectively enhances the intensities of features C, E, and G, while peak A remains essentially unchanged and peak B shows only a comparatively minor change, mainly a slight shift toward higher energy (Figure a,b). These results identify torsional variation as the structural parameter most closely associated with the experimentally observed spectral evolution. This behavior indicates that the dominant spectral changes arise primarily from progressive local symmetry breaking at the bridging oxygen site, rather than from simple energy shifts in the ground-state electronic levels (Figure c).

2.

2

Open in a new tab

Minimal computational model used to elucidate the structural origin of the experimental O 1s XES spectral evolution. The model consists of a Si2O7 dimer, which captures the essential local connectivity of the silicate network while enabling systematic control of key geometric parameters: the Si–O bond length (l), the Si–O–Si bond angle (θ), and, most critically, the O–Si–O–Si torsion angle (ω). This model provides a well-defined platform for isolating the influence of individual structural degrees of freedom on the XES spectral features, enabling direct correlation between local geometric distortions and changes in the O 1s XES spectral profile.

3.

3

Open in a new tab

Theoretical demonstration that the O–Si–O–Si torsion angle governs the key XES spectral features. (a) Simulated O 1s XES spectra as a function of torsion angle (ω). The Si–O bond length (l) and the Si–O–Si bond angle (θ) were fixed at 1.75 Å and 150.8°, respectively. The spectra were normalized to a baseline of 0 and a maximum intensity of 1. The intensities of peaks c, e, and g are uniquely sensitive to ω. (b) Difference spectra at each torsion angle, referenced to the Boltzmann-weighted simulated spectrum. (c) Calculated orbital energy diagram as a function of torsion angle. The torsion-angle dependence of the energy levels is small, indicating that the observed spectral evolution is not driven by energy shifts. (d) Spatial distribution of the valence molecular orbital (MO 31) of the Si2O7 dimer model for representative torsion-angle conformations. In the eclipsed geometry (ω ≈ 0°), the high local symmetry promotes electron delocalization across the dimer, resulting in reduced electron density on the bridging oxygen atom. In contrast, in the staggered geometry (ω ≈ 60°), symmetry lowering suppresses delocalization and enhances electron localization on the bridging oxygen, increasing the asymmetry of the π-type orbital. This increased localization directly enhances the O 1s XES transition probability associated with feature C.

Ab initio calculations reveal that feature C derives from valence molecular orbital 31 (Figure d), while features E and G originate from a dense manifold of higher-lying valence molecular orbitals. In the eclipsed conformation (ω ≈ 0°), the high structural symmetry promotes electron delocalization across the entire Si2O7 dimer. For example, the central oxygen in MO 31 adopts a “bridging-like” motif that facilitates delocalization between SiO4 tetrahedral units; consequently, the electron density is only 3.4% localized at the central oxygen, leading to lower XES intensity. Conversely, in the staggered conformation (ω ≈ 60°), the lowering of symmetry suppresses this delocalization. The oxygen in MO 31 adopts a “vertex-like” motif, increasing its localization to 7.2%. Similarly, the higher-energy states contributing to peaks E and G experience increased orbital asymmetry and electron localization. Consequently, the O 1s XES transition probabilities for this entire manifold of π-type molecular orbitals are enhanced, leading to the observed intensification of features C, E, and G.

This theoretical insight establishes a quantitative framework for interpreting the experimental XES data. To further evaluate the thermodynamic driving force for these structural transitions, we calculated the potential energy surface of the Si2O7 dimer model as a function of the Si–O–Si bond angle and the O–Si–O–Si torsion angle (Figure d). The energy landscape reveals that the eclipsed conformation (ω ≈ 0°) is energetically unfavorable due to steric hindrance and electronic repulsion between terminal oxygen atoms. In contrast, the total energy significantly decreases as the torsion angle evolves toward a minimum near the staggered conformation (ω ≈ 60°). This provides a thermodynamic basis for using the torsion angle as a key structural descriptor: the transition from a disordered, eclipsed-rich amorphous state to a more ordered, staggered-rich framework is consistent with energetic stabilization. While the specific torsion angle in the final MWW-type framework (ω ≈ 37.6°) is constrained by its 3D topology, the tendency toward torsional ordering appears to reflect an intrinsic feature of silicate network formation.

To define a reference representative of the amorphous SiO2 precursor, we constructed a Boltzmann-weighted ensemble average XES spectrum by sampling a broad distribution of Si–O–Si bond angles and O–Si–O–Si torsion angles at the synthesis temperature (150 °C). By comparing this amorphous-like reference with calculated spectra of specific torsion-angle configurations, we achieved a parameter-resolved correlation between the overall spectral evolution and local network geometry.

This remarkable agreement provides strong evidence that, at the atomic level, zeolite crystallization is closely associated with the ordering of the O–Si–O–Si torsion angle. Notably, the interlayer ten-membered rings of the final crystalline MWW-type framework (Figure c) exhibit an average torsion angle of ω ≈ 37.6°, aligning precisely with the conformational region that induces the characteristic features observed in the crystalline-phase XES spectrum. Consequently, the systematic development of these spectral features, particularly the emergence of peak C, serves as a direct spectroscopic signature of ordered 3D framework formation. More broadly, our results demonstrate that O 1s XES, by probing the electronic states of bridging oxygen atoms, provides intrinsic sensitivity to 3D network geometry. While this study identifies torsional ordering as the primary structural driver, our simulations suggest that XES spectral profiles also respond to variations in bond angles and lengths, underscoring its potential as a versatile probe for the 3D architecture of complex network materials. Ultimately, this approach establishes a new paradigm for decoding the hidden structural evolution of amorphous-to-crystalline transitions, where XES acts as a direct bridge between local electronic states and long-range topological order.

A Unified Mechanism for Zeolite Crystallization

Our direct observation of 3D network topology provides the missing link required to construct a unified mechanism for zeolite crystallization. By integrating insights from a comprehensive suite of spectroscopic and diffraction probes, we can now provide a comprehensive atomistic account of the transformation from an amorphous gel to a crystalline solid. The hierarchical process unfolds in three distinct stages, as illustrated in Figure . Crucially, these stages are defined not by arbitrary synthesis time, but by specific structural “triggers” identified through our complementary multiprobe analysis: Stage I is marked by the initiation of local torsional ordering (O 1s XES); Stage II by the emergence of long-range periodic order (XRD); and Stage III by the rapid acceleration of framework heteroatom incorporation (Fe K-edge XAFS). This phenomenon-based classification ensures the robustness of our model across varying experimental conditions (see Figure S8 and Table S5).

  • (1)

    Stage I: seeding of topological order (<1 day).

5.

5

Open in a new tab

Schematic illustration of the hierarchical, multistep crystallization mechanism of Fe-containing MWW-type zeolites. This model highlights that local torsional ordering of the silicate network precedes and facilitates long-range crystallization, while heteroatom incorporation becomes decisive only at later stages.

The transformation initiates not with long-range crystallization, but with the emergence of local 3D order within the amorphous gel. Crucially, our O 1s XES data reveal that this is the precise time frame in which O–Si–O–Si torsion angles begin to evolve from a random distribution toward ordered, staggered conformations (Figure ). In stark contrast, XRD (Figure ) remains featureless at 1 day, while Fe K-edge XAS (Figures and S9) and UV–vis spectroscopy (Figure S10) show no significant heteroatom incorporation into the silicate network. − These findings demonstrate that 3D topological ordering is a precursorial event that precedes long-range order and occurs independently of heteroatom incorporation. Long before crystallization becomes detectable by X-ray diffraction, the fundamental topology of the zeolite framework is established through early torsional orderinga network-intrinsic pathway for reducing configurational disorder and establishing 3D connectivity.

6.

6

Open in a new tab

XRD patterns of Fe-MWW zeolites collected at different synthesis times, shown over (a) a wide 2θ range and (b) an expanded view of the low-angle region. While pronounced local structural evolution is already detected by O 1s XES at Day 1, no Bragg reflections are observed at this stage. The first diffraction peaks appear only after 3 days of hydrothermal treatment, indicating the onset of long-range periodic order. This discrepancy between early local ordering and delayed diffraction signatures highlights the complementary sensitivity of X-ray emission spectroscopy to prenucleation structural organization that remains invisible to conventional diffraction techniques.

7.

7

Open in a new tab

Delayed incorporation of Fe heteroatoms tracked by Fe K-edge XAFS. (a) Evolution of Fe K-edge XANES spectra as a function of hydrothermal synthesis time. (b) Quantitative evolution of the local Fe coordination environment, expressed as the fraction of tetrahedrally coordinated Fe (FeO4). A pronounced increase in the FeO4 fraction is observed after Day 3, indicating that substantial heteroatom incorporation occurs predominantly during the post-nucleation crystal growth stage. These results demonstrate that Fe incorporation is decoupled from the initial torsional ordering of the silicate network and becomes significant only after the establishment of long-range framework order.

Our potential energy surface analysis (Figure d) provides mechanistic support for the Stage I ordering process. The calculations show that staggered-like torsional configurations are energetically favored over eclipsed ones, indicating a thermodynamic basis for the evolution of the silicate network toward more ordered local conformations. Although the present model does not directly quantify crystallization kinetics, these results support the view that torsional ordering is not merely a secondary consequence of crystallization, but rather a fundamental structural condition that facilitates the emergence of the MWW-type framework. In this sense, Stage I can be understood as a network-intrinsic ordering process that prepares the local topology for subsequent long-range organization.

  • (2)

    Stage II: nucleation and framework crystallization (1–3 days).

Once a critical density of topologically ordered domains is established, they coalesce to serve as nuclei for long-range crystallization. This transition is evidenced by the initial emergence of Bragg peaks in the XRD patterns on day 3 (Figure ). Notably, these early diffraction peaks exhibit shorter interplanar spacings compared to those observed after 5 days, suggesting that the day-3 framework exists as a structurally contracted intermediate, distinct from the fully matured zeolite. Concurrently, Si K-edge XAS (Figure S11) confirms the development of a more ordered silicate environment. In contrast, Fe K-edge XAS reveals that Fe incorporation remains minimal, indicating that this stage is predominantly governed by the condensation of the SiO2 network (Figure ). Thus, Stage II represents a transitional crystallization regime where a pure silicate framework establishes a long-range intermediate structure, effectively bridging early topological ordering with subsequent heteroatom-driven maturation.

  • (3)

    Stage III: accelerated crystal growth and decisive heteroatom incorporation (>3 days).

In the final stage, following the establishment of a continuous silicate framework, crystal growth accelerates dramatically, yielding bulk 3D zeolite crystals (Figure ). Consistently, SEM observations also capture the progression of crystal growth; platelet-like particles characteristic of MWW-type zeolites become visible by day 3 (Figure S12). This rapid expansion coincides with a pronounced surge in Fe incorporation into the framework, as evidenced by Fe K-edge XAS (Figure ). This contrast with Stages I and IIwhere framework formation proceeds largely independently of the heteroatomsignals a qualitative mechanistic shift. In Stage III, heteroatom incorporation emerges as the decisive driver for framework maturation and the completion of the crystalline structure. Collectively, these observations reveal a hierarchical crystallization process: initial topological ordering and framework construction are governed by the SiO2 network, whereas heteroatom incorporation becomes critical for driving accelerated crystal growth and structural maturation at later stages.

While this mechanism was elucidated using the Fe-MWW system, we consider this hierarchical pathwaywhere topological ordering precedes heteroatom incorporationto be a plausible representative scenario in zeolite synthesis. The identity of the heteroatom (e.g., Al, Ti, or Sn) may modulate the crystallization kinetics, but the tendency toward torsional ordering is likely a general feature of the evolving silicate network. The XES-based approach demonstrated here thus provides a general methodology for exploring the prenucleation stages of diverse framework materials.

In summary, by employing O 1s XES as a spectroscopic ruler for the O–Si–O–Si torsion angle, we have provided an unprecedented atomistic view into the inception of zeolite crystallization. Our findings reveal that crystallization is not merely the expansion of ordered domains, but a hierarchical pathway initiated by network-intrinsic topological ordering. Crucially, this torsional alignment precedes both the emergence of long-range order and substantial heteroatom incorporation.

This pivotal first stepthe ordering of the 3D network geometryremains invisible to conventional probes such as XRD and PDF analysis, which has historically rendered the early stages of synthesis a “black box”. By directly visualizing this hidden transition, our work redefines the early events of crystallization and elucidates the distinct, stage-dependent roles of the silicate framework and heteroatoms. Beyond zeolites, this approach establishes a broadly applicable strategy for unraveling the formation mechanisms of diverse complex network materials, including glasses, metal–organic frameworks, and other covalently bonded network solids.

Conclusion

This work fundamentally reframes our understanding of the birth of a crystal. By establishing XES as a probe for the O–Si–O–Si torsion angle, we have unlocked a new dimension in crystallization studies, moving the field beyond a one-dimensional, distance-based view of structural evolution. We have demonstrated that the ordering of this torsion angle is not simply a consequence of crystallization, but an important early stage structural condition associated with the emergence of the crystalline framework.

The implications of this paradigm shift are far-reaching. This methodology provides a roadmap to unravel, and ultimately control, the complex synthesis pathways of a vast array of materials, from industrial catalysts and glasses to advanced electronic materials, with atomic-level precision. Future efforts will focus on systematically tracking how different structure-directing agents, heteroatoms, and synthesis conditions modulate the 3D network topology. Ultimately, this approach provides the analytical foundation to transform trial-and-error synthesis into a truly rational predictive science, enabling the bottom-up design of next-generation framework materials.

Methods

Experiments

MWW-type zeolites, composed of SiO4 units assembled into secondary building units, were investigated in this study. Fe-containing MWW-type zeolites were synthesized according to a previously reported procedure, as outlined in Scheme S1. Precursor gels were prepared by introducing fumed silica (α-SiO2, Aerosil 200 V, Evonik Industries) and α-FeOOH powder (Kojundo Chemical Lab Co., Ltd.) into an aqueous solution containing N,N,N-trimethyl-1-adamantylammonium hydroxide (TMAdaOH, 25 wt % aqueous solution, Sachem), hexamethylenimine (HMI, Kanto Chemical Co. Inc.), and distilled water. The molar composition of the synthesis gel was adjusted to 1.0 SiO2: 0.033 α-FeOOH: 0.25 TMAdaOH: 0.31 HMI: 44 H2O, with the total volume fixed at 15 mL. The resulting gel was stirred at room temperature for 2 h and then transferred into a Teflon-lined vessel (Parr Instrument Company). Hydrothermal synthesis was carried out by sealing the vessel in an autoclave and heating at 150 °C for 1, 3, 5, 7, or 14 days under continuous rotation at 60 rpm. After synthesis, the solid products were recovered by centrifugation at 9000 rpm for 30 min. The collected solids were washed repeatedly with distilled water by centrifugation at 9000 rpm for 15 min, repeated four times, and subsequently dried at 80 °C for 12 h. The dried samples were finally calcined in air at 580 °C for 3 h to obtain the Fe-containing MWW-type zeolite products.

Structural characterization was performed by XRD (Rigaku Corporation, SmartLab, Cu–Kα), SEM (JEOL Ltd., JSM-7800F, operated at an accelerating voltage of 0.5 kV at room temperature) and the chemical state of Fe was analyzed using UV–visible (UV–vis) spectroscopy (JASCO Corporation, V-770TWK). To assess the validity of the computational conditions, Fourier transform infrared (FT-IR) spectroscopy was performed using a JASCO FT/IR-4200 spectrometer. Spectra were collected over the range of 400–8000 cm–1 with a resolution of 1 cm–1. For comparison, an Fe:Amo.-SiO2 reference powder containing tetracoordinated Fe was prepared mechanochemically by ball-milling a mixture of amorphous SiO2 and α-FeOOH for 48 h.

For X-ray spectroscopy in the soft X-ray region (O K-edge and Si K-edge), samples were prepared as pellets pressed onto aluminum pans. O K-edge XAS and O 1s XES measurements were performed at SPring-8 BL27SU and NanoTerasu BL07U. Horizontally polarized X-rays (10 × 200 μm2 beam size) were used at a 45° incident angle, with a silicon drift detector (SDD) placed at a 90° takeoff angle. O K-edge XAS spectra were normalized using the average intensity in the 545–550 eV range (Figure S13). Non-resonant O 1s XES spectra were acquired at an excitation energy of 550 eV with an integration time of 30 min and an energy resolution of approximately 0.5 eV (E/ΔE ∼ 1000). XES spectra were normalized by the baseline and maximum peak intensity. To minimize X-ray-induced damage, the sample stage was continuously scanned during the XES measurements.

Si K-edge XAS measurements were conducted at NanoTerasu BL08U using partial fluorescence yield (PFY) mode by monitoring the Si Kα emission. The incident X-ray beam was set at 45° relative to the sample surface, with an exposure time of 1 s per energy point, and an SDD was positioned at a 45° takeoff angle (Figure S11). Spectra were normalized by setting the pre-edge intensity to zero and the average post-edge intensity (1885–1890 eV) to unity. Fe K-edge XAS measurements were performed in the transmission mode at SPring-8 BL01B1 and NanoTerasu BL08W (Figures and S9). Pellets were prepared using only the zeolite samples, without dilution or mixing with BN. For the Fourier transform, data in the k-range of 3–10 Å–1 were used, and the fit window was 1–5 Å–1.

Theory

Theoretical calculations were performed using the ORCA quantum chemistry package. To analyze variations in bond lengths, bond angles, and torsion angles at the smallest structural unit, we constructed a Si2O7 dimer model for O 1s XES spectral simulations. To compensate for the cluster charge, the terminal oxygen atoms of the SiO4 tetrahedra were capped with H+. Amorphous SiO2 XES was modeled as a Boltzmann-weighted ensemble average over representative local-structure clusters. The ensemble-averaged spectrum was computed as S(ω) = ∑ i w i S i (ω), with w i ∝ exp­[-(E i -E min)/k B T], where E i is the total energy of cluster i and T was set to the synthesis temperature (150 °C).

Geometry optimization, vibrational frequency analysis, and XES simulations were carried out using the B3LYP functional − in combination with the def2-TZVP basis set and the def2-TZVP/J auxiliary basis sets. , The accuracy of the structural optimization calculations was validated by comparing the optimized amorphous SiO2 structural parameters with those reported in prior experimental and computational studies (Table S3) and by confirming consistency with the experimental IR spectrum (Figure S7). − Solvent effects were incorporated using the conductor-like screening model (COSMO). −

XES spectra were simulated using the density functional theory/restricted open-shell configuration interaction with singles (DFT-ROCIS) approach. The excitation space was defined by two donor spaces and one acceptor space. The first donor space included the O 1s core orbital of the central oxygen atom (Ob) in the structural model, while the second donor space comprised the O 2s, O 2p, and Si 2s and 2p orbitals that primarily define the bonding orbitals of the Si2O7 dimer. The acceptor space included the virtual orbitals required to describe the non-resonant O 1s XES transitions. The extent of electron localization at the central oxygen atom for selected valence orbital states was estimated from reduced Löwdin orbital populations.

To reproduce the overall line shape of the experimental spectra, the XES simulations with systematically varied bonds and torsion angles employed a model in which the Si–O bond lengths were set to 1.75 Å, which is 0.12 Å longer than the value obtained from geometry optimization (Figure S3 and Table S4). This empirical adjustment was introduced to improve agreement with the experimental line shape and to facilitate comparison of the spectral features. To assess the reliability of this approach, additional simulations were also performed using a geometry-optimized bond length of 1.63 Å (Figures S5 and S6). These results indicate that, although the absolute peak positions are sensitive to the bond length, the relative intensity modulations and the overall shape of the difference spectra remain consistent. This suggests that the identified spectroscopic signatures are robust and primarily reflect the torsional ordering of the network, rather than arising from the empirical bond-length adjustment.

For feature-specific analysis of the O 1s XES difference spectra, the energy positions of features A, B, D, and F were determined from the XES spectra together with their derivative spectra, whereas those of features C, E, and G were defined from the maxima in the difference spectra (Figure S2 and Table S2). The integrated areas of features C, E, and G were evaluated within fixed energy windows using the trapezoidal rule. For quantitative comparison between experiment and simulation, a single global energy correction was applied to the calculated spectra such that feature F of the Boltzmann-weighted average spectrum matched the corresponding feature in the experimental XES spectrum. The same fixed energy windows were then used for both the experimental and simulated difference spectra.

We note that the Si2O7 dimer model used in our simulations is a simplified representation of the complex 3D silicate network. Nevertheless, it is appropriate for the present analysis because O 1s XES is an element-selective probe that is primarily sensitive to the local electronic structure of bridging oxygen atoms. This minimal cluster captures the essential local connectivity and provides a well-defined platform for isolating the influence of the individual geometric parameters. We acknowledge that the model does not explicitly include long-range periodic constraints or the full electrostatic environment of the extended framework, such as the Madelung potential, which may influence the absolute energy levels and bond distances. Future studies using larger clusters or periodic density functional theory (DFT) will be valuable for further refining the correlation between electronic states and complex 3D architectures.

Supplementary Material

ja6c03877_si_001.pdf (2.1MB, pdf)

Acknowledgments

The authors thank Prof. A. Muramatsu, Prof. R. Osuga, M. Hodota, and A. Ogihara of Tohoku University for their helpful discussions and technical support. This work was financially supported by a Grant-in-Aid for Scientific Research (S) (KAKENHI, JP21H05011 and JP25H00900), a Grant-in-Aid for Scientific Research (B) (KAKENHI, JP23K23032). Synchrotron X-ray measurements were performed at the BL01B1 beamline (proposal Nos. 2022B1784 and 2023B1678) and the BL27SU beamline (proposal Nos. 2022A1030, 2023B1560, 2023B1033, 2023B1678, 2024B1515, 2025A1210, and 2025B1312) of SPring-8. Part of the synchrotron X-ray experiments in this study were conducted at the NanoTerasu BL07U, BL08U, and BL08W beamlines of the Photon Science Innovation Center (PhoSIC) under the Coalition Program. Part of the simulation results in this research was obtained using the Super Computer AOBA Subsystem AOBA-S, under the NanoTerasu DX Project supported by the Cyber Science Center, Tohoku University.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacs.6c03877.

  • Repeatability analysis of O 1s XES measurements; feature-specific quantitative analysis of O 1s XES and difference spectra; validation and robustness tests of the Si2O7 dimer model and XES simulations, including bond-length, bond-angle, and torsion-angle dependence; stage classification and mechanistic support from multiprobe analysis; and supporting experimental characterization by Fe K-edge EXAFS, UV–vis, Si K-edge XAFS, SEM, O K-edge XAFS, and synthetic procedure (PDF)

The authors declare no competing financial interest.

References

  1. Corma A.. Inorganic Solid Acids and Their Use in Acid-Catalyzed Hydrocarbon Reactions. Chem. Rev. 1995;95:559–614. doi: 10.1021/cr00035a006. [DOI] [Google Scholar]
  2. Davis M. E.. Ordered porous materials for emerging applications. Nature. 2002;417:813–821. doi: 10.1038/nature00785. [DOI] [PubMed] [Google Scholar]
  3. Olsbye U., Svelle S., Bjørgen M., Beato P., Janssens T. V. W., Joensen F., Bordiga S., Lillerud K. P.. Conversion of Methanol to Hydrocarbons: How Zeolite Cavity and Pore Size Controls Product Selectivity. Angew. Chem., Int. Ed. 2012;51:5810–5831. doi: 10.1002/anie.201103657. [DOI] [PubMed] [Google Scholar]
  4. Davis M. E.. Zeolites from a Materials Chemistry Perspective. Chem. Mater. 2014;26(1):239–245. doi: 10.1021/cm401914u. [DOI] [Google Scholar]
  5. Zaarour M., Dong B., Naydenova I., Retoux R., Mintova S.. Progress in zeolite synthesis promotes advanced applications. Micropor. Mesopor. Mater. 2014;189:11–21. doi: 10.1016/j.micromeso.2013.08.014. [DOI] [Google Scholar]
  6. Liang J., Liang Z., Zou R., Zhao Y.. Heterogeneous Catalysis in Zeolites, Mesoporous Silica, and Metal–Organic Frameworks. Adv. Mater. 2017;29:1701139. doi: 10.1002/adma.201701139. [DOI] [PubMed] [Google Scholar]
  7. Li Y., Li L., Yu J.. Applications of Zeolites in Sustainable Chemistry. Chem. 2017;3:928–949. doi: 10.1016/j.chempr.2017.10.009. [DOI] [Google Scholar]
  8. Pan X., Jiao F., Miao D., Bao X.. Oxide-Zeolite-Based Composite Catalyst Concept That Enables Syngas Chemistry beyond Fischer–Tropsch Synthesis. Chem. Rev. 2021;121:6588–6609. doi: 10.1021/acs.chemrev.0c01012. [DOI] [PubMed] [Google Scholar]
  9. Rangnekar N., Mittal N., Elyassi B., Caro J., Tsapatsis M.. Zeolite membranes – a review and comparison with MOFs. Chem. Soc. Rev. 2015;44:7128–7154. doi: 10.1039/C5CS00292C. [DOI] [PubMed] [Google Scholar]
  10. Zhang J., Tang X., Yi H., Yu Q., Zhang Y., Wei J., Yuan Y.. Synthesis, characterization and application of Fe-zeolite: A review. Appl. Catal., A. 2022;630:118467. doi: 10.1016/j.apcata.2021.118467. [DOI] [Google Scholar]
  11. Chen J., Huang W., Bao S., Zhang W., Liang T., Zheng S., Yi L., Guo L., Wu X.. A review on the characterization of metal active sites over Cu-based and Fe-based zeolites for NH3-SCR. RSC Adv. 2022;12:27746–27765. doi: 10.1039/D2RA05107A. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Cundy C. S., Cox P. A.. The Hydrothermal Synthesis of Zeolites: History and Development from the Earliest Days to the Present Time. Chem. Rev. 2003;103:663–701. doi: 10.1021/cr020060i. [DOI] [PubMed] [Google Scholar]
  13. Kirschhock C. E. A., Ravishankar R., Looveren L. V., Jacobs P. A., Martens J. A.. Mechanism of Transformation of Precursors into Nanoslabs in the Early Stages of MFI and MEL Zeolite Formation from TPAOH–TEOS–H2O and TBAOH–TEOS–H2O Mixtures. J. Phys. Chem. B. 1999;103(24):4972–4978. doi: 10.1021/jp990298j. [DOI] [Google Scholar]
  14. Davis T. M., Drews T. O., Ramanan H., He C., Dong J., Schnablegger H., Katsoulakis M. A., Kokkoli E., Mccormick A. V., Penn R. L., Tsapatsis M.. Mechanistic principles of nanoparticle evolution to zeolite crystals. Nat. Mater. 2006;5:400–408. doi: 10.1038/nmat1636. [DOI] [PubMed] [Google Scholar]
  15. Chen F., Li C., Wang B., Meng X., Wu Q., Xiao F. S.. Nonclassical Approaches and Behaviors in Synthesis, Structure Characterization, and Catalysis of Zeolites. J. Phys. Chem. C. 2023;127:3377–3388. doi: 10.1021/acs.jpcc.2c07934. [DOI] [Google Scholar]
  16. Yamada H., Tominaka S., Ohara K., Liu Z., Okubo T., Wakihara T.. J. Phys. Chem. C. 2019;123:28419–28426. doi: 10.1021/acs.jpcc.9b07232. [DOI] [Google Scholar]
  17. Lobo R. F., Zones S. I., Medrud R. C.. Synthesis and Rietveld Refinement of the Small-Pore Zeolite SSZ-16. Chem. Mater. 1996;8(10):2409–2411. doi: 10.1021/cm960289c. [DOI] [Google Scholar]
  18. Agostini G., Lamberti C., Palin L., Milanesio M., Danilina N., Xu B., Janousch M., van Bokhoven J. A.. In Situ XAS and XRPD Parametric Rietveld Refinement To Understand Dealumination of Y Zeolite Catalyst. J. Am. Chem. Soc. 2010;132(2):667–678. doi: 10.1021/ja907696h. [DOI] [PubMed] [Google Scholar]
  19. Andersen C. W., Bremholm M., Vennestrom P. N. R., Blichfeld A. B., Lundegaard L. F., Iversen B. B.. Location of Cu2+ in CHA zeolite investigated by X-ray diffraction using the Rietveld/maximum entropy method. IUCrJ. 2014;1:382–386. doi: 10.1107/S2052252514020181. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Tang X., Ye J., Guo L., Pu T., Cheng L., Cao X.-M., Guo Y., Wang L., Guo Y., Zhan W., Dai S.. Atomic Insights into the Cu Species Supported on Zeolite for Direct Oxidation of Methane to Methanol via Low-Damage HAADF-STEM. Adv. Mater. 2023;35(25):2208504. doi: 10.1002/adma.202208504. [DOI] [PubMed] [Google Scholar]
  21. Mayoral A., Carey T., Anderson P. A., Lubk A., Diaz I.. Atomic Resolution Analysis of Silver Ion-Exchanged Zeolite A. Angew. Chem., Int. Ed. 2011;50:11230–11233. doi: 10.1002/anie.201105450. [DOI] [PubMed] [Google Scholar]
  22. Fan W., Ogura M., Sankar G., Okubo T.. In situ Small-Angle and Wide-Angle X-ray Scattering Investigation on Nucleation and Crystal Growth of Nanosized Zeolite A. Chem. Mater. 2007;19(8):1906–1917. doi: 10.1021/cm062827j. [DOI] [Google Scholar]
  23. de Moor P.-P. E. A., Beelen T. P. M., van Santen R. A.. In situ Observation of Nucleation and Crystal Growth in Zeolite Synthesis. A Small-Angle X-ray Scattering Investigation on Si-TPA-MFI. J. Phys. Chem. B. 1999;103(10):1639–1650. doi: 10.1021/jp982553q. [DOI] [Google Scholar]
  24. Ren L., Li C., Fan F., Guo Q., Liang D., Feng Z., Li C., Li S., Xiao F.-S.. UV–Raman and NMR Spectroscopic Studies on the Crystallization of Zeolite A and a New Synthetic Route. Chem.Eur. J. 2011;17:6162–6169. doi: 10.1002/chem.201100098. [DOI] [PubMed] [Google Scholar]
  25. Lesthaeghe D., Vansteenkiste P., Verstraelen T., Ghysels A., Kirschhock C. E. A., Martens J. A., Speybroeck V. V., Waroquier M.. MFI Fingerprint: How Pentasil-Induced IR Bands Shift during Zeolite Nanogrowth. J. Phys. Chem. C. 2008;112:9186–9191. doi: 10.1021/jp711550s. [DOI] [Google Scholar]
  26. Wakihara T., Kohara S., Sankar G., Saito S., Sanchez-Sanchez M., Overweg A. R., Fan W., Ogura M., Okubo T.. A new approach to the determination of atomic-architecture of amorphous zeolite precursors by high-energy X-ray diffraction technique. Phys. Chem. Chem. Phys. 2006;8:224–227. doi: 10.1039/B512840D. [DOI] [PubMed] [Google Scholar]
  27. Wakihara T., Okubo T.. Hydrothermal Synthesis and Characterization of Zeolites. Chem. Lett. 2005;34(3):276–281. doi: 10.1246/cl.2005.276. [DOI] [Google Scholar]
  28. Ikuno T., Chaikittisilp W., Liu Z., Iida T., Yanaba Y., Yoshikawa T., Kohara S., Wakihara T., Okubo T.. Structure-Directing Behaviors of Tetraethylammonium Cations toward Zeolite Beta Revealed by the Evolution of Aluminosilicate Species Formed during the Crystallization Process. J. Am. Chem. Soc. 2015;137(45):14533–14544. doi: 10.1021/jacs.5b11046. [DOI] [PubMed] [Google Scholar]
  29. Minami A., Hu P., Sada Y., Yamada H., Ohara K., Yonezawa Y., Sasaki Y., Yanaba Y., Takemoto M., Yoshida Y., Okubo T., Wakihara T.. Tracking Sub-Nano-Scale Structural Evolution in Zeolite Synthesis by in situ High-Energy X-Ray Total Scattering Measurement with Pair Distribution Function Analysis. J. Am. Chem. Soc. 2022;144:23313–23320. doi: 10.1021/jacs.2c05722. [DOI] [PubMed] [Google Scholar]
  30. Nilsson A., Pettersson L. G. M.. Chemical bonding on surfaces probed by X-ray emission spectroscopy and density functional theory. Surf. Sci. Rep. 2004;55:49–167. doi: 10.1016/j.surfrep.2004.06.002. [DOI] [Google Scholar]
  31. Bergmann U., Glatzel P.. X-ray emission spectroscopy. Photosynth. Res. 2009;102:255–266. doi: 10.1007/s11120-009-9483-6. [DOI] [PubMed] [Google Scholar]
  32. Itamoto K., Ninomiya K., Fukushi H., Taguchi H., Nakajima K., Uchiyama E., Miyazawa Y., Ando M., Fukushima S., Nishibori M.. Compositional and Structural Analysis of Iron Phosphate Conversion Coatings by Soft X-ray Emission Spectroscopy. Mater. Trans. 2024;65(12):1560–1565. doi: 10.2320/matertrans.MT-M2024092. [DOI] [Google Scholar]
  33. Matsuno R., Nutthon Y., Miyano A., Ninomiya K., Nishibori M., Kiuchi H., Fujikawa S., Harada Y., Takahara A.. Electronic Structure of Carbon Dioxide in Sylgard-184 Evaluated by Using X-ray Emission Spectroscopy. Chem. Lett. 2022;51(6):650–653. doi: 10.1246/cl.220113. [DOI] [Google Scholar]
  34. Maganas D., DeBeer S., Neese F.. A Restricted Open Configuration Interaction with Singles Method To Calculate Valence-to-Core Resonant X-ray Emission Spectra: A Case Study. Inorg. Chem. 2017;56(19):11819–11836. doi: 10.1021/acs.inorgchem.7b01810. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Osuga R., Tanaka G., Yabushita M., Ninomiya K., Maki S., Nishibori M., Kanie K., Muramatsu A.. Development of Synthetic Route for Fe-substituted MWW-type Zeolites Using Mechanochemical Method. J. Jpn. Petrol. Inst. 2022;65:67–77. doi: 10.1627/jpi.65.67. [DOI] [Google Scholar]
  36. Schwidder M., Kumar M. S., Klementiev K., Pohl M. M., Brückner A., Grünert W.. J. Catal. 2005;231:314–331. doi: 10.1016/j.jcat.2005.01.031. [DOI] [Google Scholar]
  37. Pérez-Ramírez J., Kapteijn F., Brückner A.. J. Catal. 2003;218:234–239. doi: 10.1016/S0021-9517(03)00087-3. [DOI] [Google Scholar]
  38. Mansour H., Omri K., Bargougui R., Ammar S.. Appl. Phys. A: Mater. Sci. Process. 2020;126:152. [Google Scholar]
  39. Liu G., Zhu D., Zhou W., Liao S., Cui J., Wu K., Hamilton D.. Appl. Surf. Sci. 2010;256:2546–2552. doi: 10.1016/j.apsusc.2009.10.102. [DOI] [Google Scholar]
  40. Li D., Bancroft G. M., Fleet M. E., Feng X. H.. Silicon K-edge XANES Spectra of Silicate Minerals. Phys. Chem. Minerals. 1995;22:115–122. doi: 10.1007/bf00202471. [DOI] [Google Scholar]
  41. Neese F.. The ORCA program system. WIREs Comput. Mol. Sci. 2012;2:73–78. doi: 10.1002/wcms.81. [DOI] [Google Scholar]
  42. Becke A. D.. Density-functional exchange-energy approximation with correct asymptotic behavior. Phys. Rev. A:At., Mol., Opt. Phys. 1988;38:3098–3100. doi: 10.1103/PhysRevA.38.3098. [DOI] [PubMed] [Google Scholar]
  43. Lee C., Yang W., Parr R. G.. Development of the Colle–Salvetti correlation-energy formula into a functional of the electron density. Phys. Rev. B:Condens. Matter. 1988;37:785–789. doi: 10.1103/PhysRevB.37.785. [DOI] [PubMed] [Google Scholar]
  44. Becke A. D.. Density-functional thermochemistry. III. The role of exact exchange. J. Chem. Phys. 1993;98:5648–5652. doi: 10.1063/1.464913. [DOI] [Google Scholar]
  45. Feyereisen M., Fitzgerald G., Komornicki A.. Use of Approximate Integrals in Ab Initio Theory. An Application in MP2 Energy Calculations. Chem. Phys. Lett. 1993;208:359–363. doi: 10.1016/0009-2614(93)87156-w. [DOI] [Google Scholar]
  46. Kendall R. A., Früchtl H. A.. The impact of the resolution of the identity approximate integral method on modern ab initio algorithm development. Theor. Chem. Acc. 1997;97:158–163. doi: 10.1007/s002140050249. [DOI] [Google Scholar]
  47. Mozzi R. L., Warren B. E.. J. Appl. Crystallogr. 1969;2:164–172. doi: 10.1107/S0021889869006868. [DOI] [Google Scholar]
  48. Munetoh S., Motooka T., Moriguchi K., Shintani A.. Comput. Mater. Sci. 2007;39:334–339. doi: 10.1016/j.commatsci.2006.06.010. [DOI] [Google Scholar]
  49. Zhu W., Zheng G., Cao S., He H.. Thermal conductivity of amorphous SiO2 thin film: A molecular dynamics study. Sci. Rep. 2018;8:10537. doi: 10.1038/s41598-018-28925-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Klamt A.. Conductor-Like Screening Model for Real Solvents: A New Approach to the Quantitative Calculation of Solvation Phenomena. J. Phys. Chem. 1995;99(7):2224–2235. doi: 10.1021/j100007a062. [DOI] [Google Scholar]
  51. Yang C. S., Mora-Fonz J. M., Catlow C. R. A.. Stability and Structures of Aluminosilicate Clusters. J. Phys. Chem. C. 2011;115(49):24102–24114. doi: 10.1021/jp202394w. [DOI] [Google Scholar]
  52. Yang C. S., Mora-Fonz J. M., Catlow C. R. A.. Modeling the Polymerization of Aluminosilicate Clusters. J. Phys. Chem. C. 2012;116(42):22121–22128. doi: 10.1021/jp3049197. [DOI] [Google Scholar]
  53. Roemelt M., Maganas D., DeBeer S., Neese F.. A combined DFT and restricted open-shell configuration interaction method including spin-orbit coupling: Application to transition metal L-edge X-ray absorption spectroscopy. J. Chem. Phys. 2013;138:204101. doi: 10.1063/1.4804607. [DOI] [PubMed] [Google Scholar]
  54. Yin Z., Inhester L., Thekku Veedu S., Quevedo W., Pietzsch A., Wernet P., Groenhof G., Fö hlisch A., Grubmüller H., Techert S.. Cationic and Anionic Impact on the Electronic Structure of Liquid Water. J. Phys. Chem. Lett. 2017;8(16):3759–3764. doi: 10.1021/acs.jpclett.7b01392. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

ja6c03877_si_001.pdf (2.1MB, pdf)

Articles from Journal of the American Chemical Society are provided here courtesy of American Chemical Society

RESOURCES