# Fileset

[Atomic and Electronic Structures on a Mordenite Zeolite.pdf](https://mdr.nims.go.jp/filesets/b808c6c2-836b-4cf6-b216-63c5e73fcc5e/download)

## Creator

Shinya Hosokawa, Hitoshi Sato, Yasuhisa Tezuka, Jun-ichi Adachi, Koji Kimura, Koichi Hayashi, [Shinji Kohara](https://orcid.org/0000-0001-9596-2680), Hiroo Tajiri, Kentaro Kobayashi, Akihide Koura, Fuyuki Shimojo

## Rights

[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)

## Other metadata

[Atomic and Electronic Structures on a Mordenite Zeolite](https://mdr.nims.go.jp/datasets/0013be7f-6e09-4f6d-af92-156c4490005c)

## Fulltext

e-Journal of Surface Science and Nanotechnology 22, 25–31 (2024)Atomic and Electronic Structures on aMordenite ZeoliteShinya Hosokawa,a, † Hitoshi Sato,b Yasuhisa Tezuka,c Jun-ichi Adachi,d Koji Kimura,e Koichi Hayashi,eShinji Kohara,f Hiroo Tajiri,g Kentaro Kobayashi,a, ‡ Akihide Koura,h Fuyuki Shimojo ha Institute of Industrial Nanomaterials, Kumamoto University, Kumamoto 860-8555, Japanb Hiroshima Synchrotron Radiation Center, Hiroshima University, Higashi-Hiroshima 739-0046, Japanc Graduate School of Science and Technology, Hirosaki University, Hirosaki 036-8561, Japand Institute of Materials Structure Science (IMSS), High Energy Accelerator Research Organization (KEK), Tsukuba, Japane Department of Physical Science and Engineering, Nagoya Institute of Technology, Nagoya 466-8555, Japanf National Institute for Materials Science (NIMS), Tsukuba 305-0047, Japang Japan Synchrotron Radiation Research Institute (JASRI), Sayo 679-5198, Japanh Department of Physics, Kumamoto University, Kumamoto 860-8555, Japan† Corresponding author: shhosokawa@kumamoto-u.ac.jp‡ Present address: Faculty of Materials for Energy, Shimane University, Matsue 690-8504, JapanReceived: 1 June, 2023; Accepted: 11 August, 2023; J-STAGE Advance Publication: 21 September, 2023; Published: 21 September, 2023Atomic structures of an insulating and hydrophobic zeolite sampleof mordenite were measured by high-energy X-ray diffraction, and apair distribution function analysis was carried out. Valence- and con-duction-band O 2p partial densities of states (DOSs) in a mordenitewere measured by soft X-ray emission and absorption spectroscopies(SXES and SXAS), respectively. The SXAS spectrum for the con-duction band O 2p orbital has characteristic structures like that ofcrystalline SiO2, while pre-shoulders are observed in mordenite. Bychoosing characteristic energies in the SXAS spectrum for the incidentphoton energies, SXES spectra were obtained, in which a large peakand three small peaks or shoulders can be assigned by a lone pairorbital and bonding (σ) ones, respectively. A density functional theorywas applied to determine the exact atomic structures and electronicstates, and they are in good agreement with the corresponding experiments. It is concluded that the O 2p partial DOS is mainly O-Si covalent bonds, and the Al and Na atoms have minor contributions for them. From this study, it was found that the fundamentalproperties of complex zeolites can only be obtained in combination of experimental and theoretical investigations as mentionedabove, which can open feasibilities to uncover the origin of active sites in functional zeolites.Keywords Hydrophobic zeolite; Atomic structures; Partial electronic density of states; Density functional theoryI. INTRODUCTIONZeolites are functional materials that the nature manufac-tures by self-organizations. They are mostly composed ofaluminosilicate networks with a considerable number ofpores with a nanometer size in diameter [1], and only onegram of zeolite possesses a spacious surface of pores with anarea wider than that of a tennis court. Zeolites have manyapplications, such as classical ones of molecular sieves andadsorbents for moistures, radioactive Cs ions, etc., or recentlyhigh durable exhaust gas purification devices by P-includingones, petrochemical catalysts by hydrophobic ones, antibac-terial agents by Ag ions in their pores, and so on. More thantwo million tons of zeolites are manufactured per year allover the world, and thus, nowadays human lives do not workwithout zeolites.To consider functions in zeolites, three structural partsshould be carefully discussed, i.e., 1) frameworks made ofmainly Si, Al, and O atoms, 2) cations located at the poresurface like Na+ compensating the fourfold Al ions, and 3)molecules or nano-scaled atomic groups adsorbing into thepores. A variety of atomic groups can be adsorbed in thepores such as water, CO2 [2], NOx [3], Na [4], Se [5], Se-Te[6], etc., for which physical properties were investigated bylimited experimental methods so far. Although the applica-tions of zeolites are widely spread in human lives, basicRegular Papere-J. Surf. Sci. Nanotechnol. 22, 25–31 (2024) | DOI: 10.1380/ejssnt.2023-063 25mailto:shhosokawa@kumamoto-u.ac.jphttps://doi.org/10.1380/ejssnt.2023-063knowledges on these materials are not so rich, and thus,higher advancements have not been achieved on the basisof the fundamental understandings of basic scientific proper-ties so far. This was because in addition to complex structuresof zeolites, most of active sites for the functions, such asdoped atoms in the frameworks, cations at the pore surfaces,and adsorbed atomic groups, have no long-range periodicity,which makes difficult to exactly determine atomic positionsby usual diffraction methods. Moreover, it is difficult tomeasure electronic structures of zeolites by photoemissionspectroscopy because they are highly insulating.From the experimental viewpoint of atomic structures,zeolites are one of the amazing as well as complex materialsbecause of the aforementioned combination. However, a pairdistribution function (pdf ) analysis becomes very effectiveby using high energy X-rays or neutrons from presentlydeveloped synchrotron or pulsed neutron facilities, whichutilized for powder diffraction experiments on polycrystallinematerials [7, 8], particularly on zeolites [9, 10]. From theexperimental viewpoint of electronic structures, soft X-rayemission and absorption spectroscopy (SXES and SXAS)techniques are available for measuring the electronic densityof states (DOS) of insulating materials [11]. Both involvesthe element and azimuthal quantum number-selective meth-ods. In our previous reports, these methods were employedsuccessfully to measure the O 2p partial DOS of typicallyinsulating crystalline materials such as SiO2 and B2O3glasses [12, 13]. Moreover, theoretical calculations wereperformed by density functional theory (DFT) to verify theexperimental data.Mordenite (Na2O·Al2O3·20SiO2) is a typical hydrophobiczeolite. A considerable amount of mordenite is naturallyproduced, and thus, it is considered as one of thermally stablezeolites, and often chosen as a standard to judge the qualityof other zeolites. Mordenite has one-dimensional pores withan averaged diameter of about 0.66 nm. Accordingly, it isconsidered as a simple zeolite among the complex materialgroup for basic scientists, and often used for confined mate-rials sciences such as water [14] and one-dimensional mate-rials [5, 6]. Thus, we chose mordenite as a typical zeolite forinvestigating atomic and electronic structures by above-men-tioned recent sophisticated experimental methods using syn-chrotron radiation.In this article, we report results of high-energy X-raydiffraction (HEXRD) for investigating atomic structures,and SXAS and SXES for studying electronic structures ofO 2p partial DOS. The experimental results are discussed tobe reasonable by comparing with the DFT results, and thefeasibility of the combination of these methods are clarifiedto investigate the fundamental natures of atomic and elec-tronic structures of zeolites in general.II. EXPERIMENTAL AND THEORETI-CAL PROCEDURESSynthetic mordenite powder with a crystal size of about1 µm was supplied by Toyo Soda Manufacturing Co., Ltd.(courtesy of Tosoh Inc.) (No. TSZ 640NAA). The mordenitepowder was washed by distilled water and the powder sam-ple was pressed to make a pellet with a diameter of 10mmand a thickness of about 1mm.The HEXRD experiment was performed at BL47XU ofthe SPring-8, Sayo, Japan. The diffractometer installed at thebeamline was originally designed for anomalous X-ray scat-tering experiments. To shorten the angle scanning time, thedetecting system has three scintillation counters separatedone another by about 30°. Each detector has an analyzercrystal of LiF with the (002) reflection to discriminate onlyelastic scattering signals from fluorescent and Compton scat-tering contributions [15] with a resolution energy of about30 eV at the incident X-ray energy of about 25.5 keV for thisexperiment. The incident X-ray energy was chosen to be at25.481 keV near the Ag Kα absorption edge, covering thewave number (Q) range up to about 190 nm−1.SXES and SXAS measurements were carried out atBL13A of the Photon Factory in the High-Energy Acceler-ator Research Organization (PF-KEK), Tsukuba, Japan.There, synchrotron radiation from a linear undulator devicewas monochromatized by a valiable-included-angle Monk-Gilliesen-type monochromator with a varied-line-spacinggrating with a line density of 1000 linesmm−1 [16]. TheSXAS spectra of O 2p conduction-band partial DOS weremeasured around the O K absorption edge in total electronyield mode. The energy resolution of SXAS was higher than0.1 eV full-width at half-maximum (FWHM).The SXES spectra were measured using a Rowland-typemonochromator with a spherical grating with a radius of 5mand a line density of 1200 linesmm−1 and a photon detectorwith a CsI-coated multichannel plate. The energy resolutionof SXES was about 0.3 eV FWHM. Details of the experi-mental setups are given elsewhere [17]. The SXES spectra ofO 2p valence-band partial DOS were obtained with incidentphoton energies of 535–550 eV beyond the O K absorptionedge. The incident photon energies were selected at sevenones where the SXAS spectrum has characteristic structures.All the SXES and SXAS experiments were carried out atroom temperature under ultrahigh vacuum condition of lessthan 10−7 Pa.The DFT calculation was performed with a generalizedgradient approximation [18] for exchange-correlation energy.Projector augmented wave potentials [19] were employed forthe electron-ion interaction with the valence electron config-urations of 3s13p03d0, 3s23p13d0, 3s23p23d0, and 2s22p4 forthe Na, Al, Si, and O elements, respectively. The electronicwavefunctions and electron density were expanded usingplain wave basis sets with cutoff energies of 30 and 250Ryd, respectively.We used 148 atoms (4Na, 4Al, 44Si, and 96O) under a peri-odic boundary condition with a supercell size of 1:8131 nm ×2:0507 nm × 0:75221 nm equal to the unit cell size [20, 21].Due to small compositions of Na and Al atoms, size effectmay cause errors in the results concerning atomic and elec-tronic structures, which should be carefully discussed. Sim-ulations were carried out with a constant-temperature con-Regular Papere-J. Surf. Sci. Nanotechnol. 22, 25–31 (2024) | DOI: 10.1380/ejssnt.2023-063 26https://doi.org/10.1380/ejssnt.2023-063stant-volume canonical ensemble at 300K for 8000 steps intime steps of 1.2 fs (totally 29 ps with 24000 steps). Van derWaals interactions were included using a Grimme correction[22, 23] (DFT-D method). We used our own moleculardynamics program, details of which are given in Ref. 24.The initial condition was set to be the XRD result [20, 21],and only the Na positions were set at the center of pores asexplained later.III. RESULTSFigure 1 shows a logarithmic plot of the structure factor,SðQÞ, of a mordenite crystal obtained by the present HEXRDexperiment. The scattering data collected with three detectorsare indicated by different colors in the figure. As seen in thefigure, a number of Bragg peaks are observed in the diffrac-tion pattern. The largest peak height is about 40 in the low Qrange of about 5–10 nm−1. The positions and heights of theBragg peaks in the low Q region are very similar to theprevious laboratory XRD works [25]. In the higher Q rangebeyond 70 nm−1, the diffuse scattering signals dominate theSðQÞ spectrum. A recent theory reveals that SðQÞ values atthe Q ! 1 limit is converged to be unity owing to a multi-phonon effect even in a crystal phase [26].Figure 2 shows the pair distribution function, gðrÞ, ob-tained from the present pdf analysis. As seen in the figure,the first prominent peak is observed at 0.162 nm with a heightof about 7, corresponding the first neighboring Si-O bondlength of the crystal [20]. The second distinct peak is seen at0.261 nm with a height of about 1.6, reflecting the O-Osecond neighboring distances in the crystal [20]. Details willbe discussed later by associating with the DFT results.The solid curve in Figure 3 shows the SXAS spectrum ofmordenite near the O K absorption edge, corresponding tothe O 2p conduction-band partial DOS. At a glance, thespectral feature is similar to that of single crystal (c-) SiO2(quartz) [12] shown as the dotted curve in the figure. In thespectrum of c-SiO2, a pre-shoulder is observed at about 536eV at the lower energy of the main peak at about 538 eV.These spectral features for the unoccupied O p DOS in c-SiO2 was investigated by an electron energy-loss spectrosco-py and a band structure calculation well reproduces theexperimental data, where the pre-shoulder and main peakare made of O p and Si sp3 antibonding states [27]. This issupported by the DFT calculations shown in Figure 7(a) ofRef. 12. On the other hand, the SXAS spectrum of mordeniteshows several additional pre-shoulders at about 533, 535,and 537 eV. They are located at slightly different energypositions for that of c-SiO2, and may be related to thecovalent-ionic bondings of O atoms with Al and/or Na ions.Besides, there are several characteristic structures as indi-cated by arrows, which were used for the incident photonenergies, E0, for the subsequent SXES measurement.The solid curves in Figure 4 shows the SXES spectra ofmordenite, corresponding to the O 2p valence-band partialDOS, at several E0 values from 533.7 to 550 eV. The SXESspectra exhibit four characteristic peaks or shoulders as in-dicated by dashed lines. Most of the structures do not changewith varying E0. The present SXES spectra of the mordenitesample resemble well that of a single crystal SiO2 (quartz)sample [12] shown as the dotted curve in the figure, wherethe prominent peak is made of O lone-pair (LP) electrons andthree structures at the lower energies correspond to the O-Sibonding (�) states. Thus, the O 2p partial electronic struc-tures in mordenite are dominated by O LP and O-Si �Figure 1: Logarithmic plot of SðQÞ spectrum on mordenite ob-tained from the present HEXRD experiment.Figure 2: gðrÞ of mordenite obtained from the present pdf analysis.SXAS intensity  (Arb. units)550545540535530Energy  (eV)Mordenitec-SiO2Figure 3: SXAS spectra of mordenite (solid curve) and c-SiO2(quartz) (dotted curve) near the O K absorption edge. For clarity,the spectra are displaced. Down arrows show the incident photonenergies for the following SXES experiments. The up arrow indi-cates the pre-shoulder in c-SiO2. The data of c-SiO2 is redrawn fromRef. 12 with a permission by Physical Society of Japan.Regular Papere-J. Surf. Sci. Nanotechnol. 22, 25–31 (2024) | DOI: 10.1380/ejssnt.2023-063 27https://doi.org/10.1380/ejssnt.2023-063orbitals, although a small amount of Al and Na ions arecontained in the mordenite sample. In combination with theSXAS spectrum in Figure 3, the band gap is estimated to beabout 6 eV, smaller than that of SiO2 crystal of about 8 eVdue to the existence of the pre-shoulders at the bottom of theconduction band.At the lower two E0 values of 533.7 and 535.2 eV near theO K edge, however, the prominent peak at about 526 eVshifts towards the lower energies, and the shoulder at about524 eV changes the spectral shapes. Note that these incidentphoton energies correspond to the energies of the pre-should-ers in the SXAS spectrum shown in Figure 3. Such spectralshifts in SXES measured at E0 near the band gap energy werediscussed on a cubic BN data by Agui et al. [28], whoargued the relationship of this phenomenon to the indirectgap of this sample. A Raman shift [29] for SXES spectra nearthe bang gap energy is another possibility to explain thespectral shifts to the lower peak energies. To examine theorigin of these shifts, further studies, such as electronic bandcalculations, are necessary, and the details cannot be dis-cussed in this paper.VI. DISCUSSION—COMPARISONWITH DFT CALCULATIONFigure 5 shows three-dimensional atomic configurationsobtained (a) before and (b) after the structural relaxation bythe DFT calculation. Bonds are indicated when the inter-atomic distance is less than 0.20 nm. Here, we explain theatomic configurations of mordenite in detail by using initialconfigurations [20, 21] shown in Figure 5(a). The Si and Alatoms are fourfold coordinated with the twofold coordinatedO atoms to form the zeolite frame. The O atoms have 10different crystallographic sites as shown by the numbers inFigure 5(a). The Si/Al atoms have 4 different sites as givenby the alphabets in the figure. Two Al atoms out of four arelocated at the ‘a’ site and the other two at the ‘c’ sites as seenin Figure 5(a) so that the Al atoms have long distances ofabout 1 nm between them. Four Na atoms stand at the centerof 12-membered rings of the pores, in which two are at thecentral pore and other two are in another pore in Figure 5(a).The aluminosilicate framework changes very slightly bythe DFT relaxation, whereas the positions of two Na atoms inthe pore center largely move to the surface of the poreshaving short interatomic distances with O atoms as shownin Figure 5(b). We performed the calculations for severaltimes by changing the initial atomic configurations alongthe pore directions in Figure 5(a), and the resultant Napositions are at mostly the same pockets on the pore surfaceas shown in Figure 5(b). This would be because the calcu-lation temperature of 300K is high and the Na atoms haveenough free volume for the movements around them in thepores. Note that the Na atoms seem to locate apart from theAl atoms, which contradicts the preliminary idea of thecharge distributions in zeolites, i.e., Na+ and Al− [30]. Sincediffraction methods can determine only the average positionsin the lattice, and their results do not guarantee the realatomic positions [31]. Therefore, the obtained result fromthe DFT relaxation is reasonable.Figure 6(a) shows total X-ray pair distribution functions,gðrÞ, obtained by the present DFT calculation, multiplied bythe averaged weighting factors of X-ray scattering to com-pare with the experimental XRD data shown by the thin solidcurve. The plenary peak is located at 0.164 nm, which is ingood agreement with the experimental result at 0.162 nm,although the peak height is much higher. The second peakappears at 0.268 nm, which is again in good agreement with4.54.03.53.02.52.01.51.00.50.0Reduced SXES intensity  (Arb. units)535530525520515510Energy  (eV)MordeniteE0 (eV) =550.0543.5540.0537.5536.7535.2533.7c-SiO2560.0Figure 4: SXES spectra of mordenite (red solid curves), corre-sponding to the O 2p valence-band partial DOS, at several E0values from 533.7 to 550 eV, together with that of c-SiO2 (quartz)(blue dotted curve). The energy positions of prominent structuresare shown by dashed lines. For clarity, the spectra are displaced eachother by 0.5. The data of c-SiO2 is redrawn from Ref. 12 with apermission by Physical Society of Japan.Figure 5: Three-dimensional atomic configurations (a) before and(b) after the DFT relaxation. The numbers and alphabets in (a)indicate ten and four different sites for the O and Si/Al atoms,respectively.Regular Papere-J. Surf. Sci. Nanotechnol. 22, 25–31 (2024) | DOI: 10.1380/ejssnt.2023-063 28https://doi.org/10.1380/ejssnt.2023-063the experiment at 0.261 nm. Relatively broad and asymmetricthird and fourth peaks are observed at about 0.32 and0.42 nm, respectively. These peaks are also observed exper-imentally in Figure 2.To clarify the element-selective information about thesestructures in gðrÞ, selected partial pair distribution functions,gijðrÞ, are given in Figure 6(b, c). In (b), the correlations withO are summarized as solid, dashed, and chain curves for Si-O, Al-O, and Na-O, respectively. As clearly seen in thefigure, the prominent peak for gðrÞ is mostly composed ofthe Si-O partial with the position of 0.164 nm. In addition,the main part of the fourth broad and asymmetric peak atabout 0.42 nm seems to be made of the Si-O correlation. TheAl-O bond length is determined to be 0.176 nm, which wasalso obtain in the Na 4A type zeolite [10]. Owing to a smallcomposition of Al, i.e., Al/Si = 1/11 in the DFT calculation,however, the Al-O peak cannot be visualized in Figure 6(a).On the other hand, the Na-O interatomic distance is found tobe 0.250 nm, indicating that the Na atoms are not located inthe frame of the zeolite. Due to the low concentration of Naatoms, this peak is not clearly observed in Figure 6(a).Figure 6(c) shows the Si-Si (solid curve) and O-O (dottedcurve) partials. The O-O interatomic distance is given to be0.268 nm, corresponding to the second peak in gðrÞ at thesame position. This peak corresponds to the O-O distance inthe SiO4 tetrahedra, and mostly coincides with the experi-mental peak position of 0.261 nm shown in Figure 2. The Si-Si correlation appears as an asymmetric peak at the topposition of 0.314 nm. This Si-Si interatomic distance corre-sponds to that between the SiO4 tetrahedra, and the distribu-tion is reflected by a variety of the Si-O-Si bond angles in thecrystal [20, 21] as shown in Figure 5. The other five gijðrÞfunctions are negligible in the total gðrÞ owing to the smallcompositional fractions of Al and Na.Figure 7 shows the theoretical result for (a) total DOS,DðEÞ, the elementally partial ones, D�ðEÞ, and the orbitalangular momentum contributions, Dl�ðEÞ, of (b) Si, (c) O, (d)Na, and (e) Al obtained by the DFT calculation as a functionof energies with respect to the Fermi energy, EF. Here, EF isdefined approximately at the center of the band gap in DðEÞ.The DðEÞ function shown in (a) by the black solid curve isseparated into D�ðEÞs given by colored curves. The D�ðEÞsare divided into Dl�ðEÞs with orbital angular momentumnumbers l, such as s, p, and d electrons corresponding tol = 0, 1, and 2, respectively. The present experimental SXASand SXES data should correspond to l ¼ 1 in (c) O partials(the dotted curve). The band gap energy obtained from thepresent DFT calculation is about 4 eV, smaller than the ex-perimental value. It is well-known that DFT calculationsusually underestimate the magnitude of band gap [32], inparticular oxide materials.The present theoretical result of the O 2p valence-bandDOS in the shallow energy range between −2 and −13 eV isin good agreement with the present SXES data, which showsthe distinct LP band at about −4 eV and three � statesbetween −5 and −12 eV at the proper energy positions. Thehybridizations of the O 2p states mostly take place with theSi 3s and 3p electrons depending on the (binding) energy, E,which is a typical feature of the sp3 hybridization of Si inSiO2 [12, 33]. The contributions of the Na and Al atoms arevery small, because the spectra in Figure 7(d, e) are smalland the compositions of these elements are also small. Notethat the vertical axis in this figure is shown DOSs per oneatom.DFTXRDFigure 6: (a) gðrÞ, (b) giOðrÞ, and (c) giiðrÞ functions obtained fromthe present DFT calculation. The thin solid curve in (a) shows theresult of HEXRD experiment (the same as in Figure 2) for thecomparison.0123TotalSiONaAl0l = 0l = 1l = 201D(E), D(E)Dl (E)012-25 -20 -15 -10 -5 0 501(e) Al(d) Na(c) O(b) Si(a) Total E - EF (eV)Figure 7: DFT results for (a) total DOS, DðEÞ, and the elementallypartial ones, D�ðEÞ, and the orbital angular momentum contribu-tions, Dl�ðEÞ, of (b) Si, (c) O, (d) Na, and (e) Al.Regular Papere-J. Surf. Sci. Nanotechnol. 22, 25–31 (2024) | DOI: 10.1380/ejssnt.2023-063 29https://doi.org/10.1380/ejssnt.2023-063Concerning the conduction-band DOS, the calculatedO 2p partial DOS is very small. On the other hand, theSi 3s and 3p states largely contribute the main part of theconduction band at energies beyond about 3 eV. This featureis reasonable when assuming the typical feature that theconduction band is composed of Si-O anti-bonding stateand the fraction of Si sp3 states is much larger than theO 2p state in the conduction band region [33, 34].The SXAS spectra basically originate from only the O 1s-2p excitations, and thus, it is necessary to observe the theo-retical O 2p contributions. Figure 8 shows the enlarged spec-trum of the O contribution, DOðEÞ, in the conduction bandrange. The red dashed curve shows the corresponding O pelectron DOS. The main conduction band edge is approx-imately located at 3 eVas indicated by the arrow, correspond-ing the experimental SXAS data at about 535.5 eV in Fig-ure 3. The structural features basically coincide with theexperimental SXAS data although the 4 eV dip is buriedprobably because of the inaccuracies of the DFT calculationand the insufficient energy resolution of experiment. Con-cerning the pre-shoulders, there are two small and broadpeaks between 1.5–3.0 eV above EF as seen in the figure,which proves the existence of the pre-shoulders in the ex-perimental SXAS data in Figure 3. As seen in Figure 7(d),the Na 3s electrons contribute the pre-shoulders, while theSi 4s electrons may also intervene there because the Sicomposition is large. The O 1s-2p SXAS spectrum doesnot originate only from the simple 1s-2p direct optical tran-sition, but it would be necessary to consider the effect of thecore hole in the final state, which may also the reason whythe coincidence of the experimental and theoretical results onthe conduction band are not enough satisfactory. Thus, wejudged at present that the discussion about the comparisonbetween the results of the SXAS experiment and DFT cal-culation is difficult to proceed further.V. CONCLUSION AND PERSPECTIVEIn this paper, we present atomic structures by HEXRDwith the pdf analysis, O 2p partial electronic DOSs on aninsulating mordenite sample obtained by SXAS and SXESexperiments, and a DFT calculation. The atomic structuresare in excellent agreement with the present DFT results. TheSAXS spectrum for the O 2p conduction band is very similarto that of SiO2 crystal, indicating that the O 2p partialelectronic states in the mordenite is composed of mostly O-Si interactions. The SXES spectra for the O 2p valence bandhave characteristic structures of LP and � orbitals, whichresemble well those of SiO2 crystal. According to the presentresults, we concluded that the experimental and theoreticalmethods used in the present studies are very feasible andpromising to investigate the fundamental properties of atomicand electronic structures on structurally complex zeolitematerials.Here, we mention our future plan how to solve the presentdiscrepancies between the experiment and the DFT calcula-tion, As pointed out above, the calculation size was only oneunit cell, and only four Na and Al atoms were included in thepresent calculation, which may induce the disagreementswith the experimental data of the electronic structures. There-fore, a larger calculation box of at least four unit cell con-taining about 600 atoms or more preferably 32 unit cells withabout 5000 atoms would be required, whereas a typicalcomputational power in the laboratory forbids such calcula-tions with such large scales. To overcome the computationallimitations, an algorithm of artificial neural network (ANN)potential was proposed by Behler and Parrinello [35], whichcreates an ANN empirical potential by training an empiricalpotential with a limited DFT result using a super computer.By using this method, the calculation could be conductedwith a first-principles accuracy but over 104 times shortercomputing time [36], which is helpful to confirm the presentknowledge on the structurally complex zeolites.The aim of this paper is to study the fundamental proper-ties of atomic and electronic structures of mordenite as areference for studying further complex zeolites with func-tional properties. For the applicational uses of zeolites, sub-stitutional exchanges of functional elements are made ineither zeolite frames, compensated cations, or pore holes.The examples of the applications are described in the firstparagraph of the Introduction section, and the substitution ofSn atoms into mordenite is highly related to the CO2 absorp-tion. For the first stage of these subsequent studies, it isimportant to obtain information about a reference zeolitewithout dopant elements for the further investigations to finda role of dopant element as an active site. Our first trial wasvery recently published on a Ag-containing 4A type zeolitehaving an antibacterial function [10].AcknowledgmentThe HEXRD experiment was carried out at BL47XU of theSPring-8 (No. 2023A1346). The SXAS and SXES experimentswere performed at BL13A in the PF-KEK (Nos. 2018G597 and2020G541). This work was supported by JSPS Grant-in-Aid forTransformative Research Areas (A) ‘Hyper-Ordered Structures Sci-ence’ (No. 21H05569 for SH and KKo, 23H04117 for SH, and20H05881 for KKi, KH, SK, and HT) and for Scientific Research1 2 3 4 5 E - EF (eV)0.000.050.10DO(E), DOl (E)DO(E)l = 0l = 1Figure 8: The enlarged spectra of the O contributions in the con-duction band range. The solid, dotted, and dashed curves representthe total, s, and p partial O spectra, respectively.Regular Papere-J. Surf. Sci. Nanotechnol. 22, 25–31 (2024) | DOI: 10.1380/ejssnt.2023-063 30https://doi.org/10.1380/ejssnt.2023-063(C) (No. 22K12662 for SH), and the Japan Science and TechnologyAgency (JST) CREST (Nos. JPMJCR1861 for SH and KKo and JP-MJCR18I2 for AK and FS).References[1] D. W. Breck, Zeolite Molecular Sieves: Structure, Chemistry,and Use (Wiley, New York, 1974).[2] D. Bonenfant, M. Kharoune, P. Niquette, M. Mimeault, and R.Hausler, Sci. Technol. Adv. Mater. 9, 013007 (2008).[3] M. Colombo, I. Nova, and E. Tronconi, Catal. Today 151, 223(2010).[4] T. Ikeda, T. Nakano, and Y. Nozue, J. Phys. Chem. C 118, 23202(2014).[5] K. Tamura, S. Hosokawa, H. Endo, S. Yamasaki, and H.Oyanagi, J. Phys. Soc. Jpn. 55, 528 (1986).[6] M. Inui, M. Yao, and H. Endo, J. Phys. Soc. Jpn. 57, 553 (1988).[7] V. Petkov, I.-K. Jeong, J. S. Chung, M. F. Thorpe, S. Kycia, andS. J. L. Billinge, Phys. Rev. Lett. 83, 4089 (1999).[8] T. Egami and S. J. L. Billinge, Underneath the Bragg Peaks:Structural Analysis of Complex Materials (Pergamon, Amsterdam,2003).[9] M. M. Martínez-Iñesta, I. Peral, T. Proffen, and R. F. Lobo, Stud.Surf. Sci. Catal. 154, 1393 (2004).[10] S. Hosokawa, K. Kobayashi, A. Koura, F. Shimojo, Y. Tezuka,J. Adachi, Y. Onodera, S. Kohara, H. Tajiri, A. Chokkalingam, andT. Wakihara, Microporous Mesoporous Mater. 359, 112662 (2023).[11] A. Kotani and S. Shin, Rev. Mod. Phys. 73, 203 (2001).[12] S. Hosokawa, H. Sato, K. Mimura, Y. Tezuka, D. Fukunaga, Y.Matsuzaki, and F. Shimojo, J. Phys. Soc. Jpn. 84, 024605 (2015).[13] S. Hosokawa, H. Sato, K. Mimura, Y. Tezuka, D. Fukunaga, K.Shimamura, and F. Shimojo, J. Phys. Soc. Jpn. 83, 114601 (2014).[14] V. Dondur and R. Dimitrijevic, Stud. Surf. Sci. Catal. 24, 345(1985).[15] S. Kohara, H. Tajiri, C. H. Song, K. Ohara, L. Temleitner, K.Sugimito, A. Fujiwara, L. Pusztai, T. Usuki, S. Hosokawa, Y.Benino, N. Kitamura, and K. Fukumi, J. Phys. Conf. Ser. 502,012014 (2014).[16] A. Toyoshima, T. Kikuchi, H. Tanaka, K. Mase, K. Amemiya,and K. Ozawa, J. Phys. Conf. Ser. 425, 152019 (2013).[17] S. Shin, A. Agui, M. Fujisawa, Y. Tezuka, T. Ishii, and N.Hirai, Rev. Sci. Instrum. 66, 1584 (1995).[18] J. P. Perdew, K. Burke, and M. Ernzerhof, Phys. Rev. Lett. 77,3865 (1996).[19] P. E. Blöchl, Phys. Rev. B 50, 17953 (1994).[20] P. Simoncic and T. Armbruster, Am. Mineral. 89, 421 (2004).[21] K. Shiokawa, M. Ito, and K. Itabashi, Zeolites 9, 170 (1989).[22] S. Grimme, J. Antony, S. Ehrlich, and H. Krieg, J. Chem. Phys.132, 154104 (2010).[23] S. Grimme, S. Ehrlich, and L. Goerigk, J. Comput. Chem. 32,1456 (2011).[24] F. Shimojo, S. Fukushima, H. Kumazoe, M. Misawa, S.Ohmura, P. Rajak, K. Shimamura, L. B. Oftelie, S. Tiwari, R. K.Kalia, A. Nakano, and P. Vashishta, SoftwareX 10, 100307 (2019).[25] V. Gramlich, Untersuchung und Verfeinerung Pseudosymme-trischer Strukturen, Ph.D. thesis, ETH, Zurich, 1971: M. M. J.Treacy and J. B. Higgins (Eds.), Collection of Simulated XRDPowder Patterns for Zeolites (Elsevier, Amsterdam, 2001), pp. 244-245.[26] H. Kitamura, Acta Crystallogr. A 78, 415 (2022).[27] L. A. J. Garvie, P. Rez, J. R. Alvarez, P. R. Buseck, A. J.Craven, and R. Brydson, Am. Mineral. 85, 732 (2000).[28] A. Agui, S. Shin, M. Fujisawa, Y. Tezuka, T. Ishii, Y.Muramatsu, O. Mishima, and K. Era, Phys. Rev. B 55, 2073 (1997).[29] H. Ishii, Y. Ishiwata, R. Eguchi, Y. Harada, M. Watanabe, A.Chainani, and S. Shin, J. Phys. Soc. Jpn. 70, 1813 (2001).[30] J. Sauer, Chem. Rev. 89, 199 (1989).[31] For example, S. Hosokawa, N. Happo, S. Senba, T. Ozaki, T.Matsushita, A. Koura, F. Shimojo, and K. Hayashi, J. Phys. Soc.Jpn. 83, 124602 (2014).[32] K. A. Johnson and N. W. Ashcroft, Phys. Rev. B 58, 15548(1998).[33] J. R. Chelikowsky and M. Schlüter, Phys. Rev. B 15, 4020(1977).[34] Z. Y. Wu, F. Jollet, and F. Seifert, J. Phys. Condens. Matter 10,8083 (1998).[35] J. Behler and M. Parrinello, Phys. Rev. Lett. 98, 146401(2007).[36] K. Shimamura, S. Fukushima, A. Koura, F. Shimojo, M.Misawa, R. K. Kalia, A. Nakano, P. Vashishta, T. Matsubara, and S.Tanaka, J. Chem. Phys. 151, 124303 (2019).All articles published on e-J. Surf. Sci. Nanotechnol. are licensedunder the Creative Commons Attribution 4.0 International (CC BY4.0). You are free to copy and redistribute articles in any medium orformat and also free to remix, transform, and build upon articles forany purpose (including a commercial use) as long as you giveappropriate credit to the original source and provide a link to theCreative Commons (CC) license. If you modify the material, youmust indicate changes in a proper way.Copyright: ©2024 The author(s)Published by The Japan Society of Vacuum and Surface ScienceRegular Papere-J. Surf. Sci. Nanotechnol. 22, 25–31 (2024) | DOI: 10.1380/ejssnt.2023-063 31https://doi.org/10.1088/1468-6996/9/1/013007https://doi.org/10.1016/j.cattod.2010.01.010https://doi.org/10.1016/j.cattod.2010.01.010https://doi.org/10.1021/jp507894uhttps://doi.org/10.1021/jp507894uhttps://doi.org/10.1143/JPSJ.55.528https://doi.org/10.1143/JPSJ.57.553https://doi.org/10.1103/PhysRevLett.83.4089https://doi.org/10.1016/S0167-2991(04)80655-4https://doi.org/10.1016/S0167-2991(04)80655-4https://doi.org/10.1016/j.micromeso.2023.112662https://doi.org/10.1103/RevModPhys.73.203https://doi.org/10.7566/JPSJ.84.024605https://doi.org/10.7566/JPSJ.83.114601https://doi.org/10.1016/S0167-2991(08)65300-8https://doi.org/10.1016/S0167-2991(08)65300-8https://doi.org/10.1088/1742-6596/502/1/012014https://doi.org/10.1088/1742-6596/502/1/012014https://doi.org/10.1088/1742-6596/425/15/152019https://doi.org/10.1063/1.1146469https://doi.org/10.1103/PhysRevLett.77.3865https://doi.org/10.1103/PhysRevLett.77.3865https://doi.org/10.1103/PhysRevB.50.17953https://doi.org/10.2138/am-2004-2-323https://doi.org/10.1016/0144-2449(89)90021-3https://doi.org/10.1063/1.3382344https://doi.org/10.1063/1.3382344https://doi.org/10.1002/jcc.21759https://doi.org/10.1002/jcc.21759https://doi.org/10.1016/j.softx.2019.100307https://doi.org/10.1107/S2053273322006441https://doi.org/10.2138/am-2000-5-611https://doi.org/10.1103/PhysRevB.55.2073https://doi.org/10.1143/JPSJ.70.1813https://doi.org/10.1021/cr00091a006https://doi.org/10.7566/JPSJ.83.124602https://doi.org/10.7566/JPSJ.83.124602https://doi.org/10.1103/PhysRevB.58.15548https://doi.org/10.1103/PhysRevB.58.15548https://doi.org/10.1103/PhysRevB.15.4020https://doi.org/10.1103/PhysRevB.15.4020https://doi.org/10.1088/0953-8984/10/36/016https://doi.org/10.1088/0953-8984/10/36/016https://doi.org/10.1103/PhysRevLett.98.146401https://doi.org/10.1103/PhysRevLett.98.146401https://doi.org/10.1063/1.5116420https://doi.org/10.1380/ejssnt.2023-063