# Fileset

[Amrit-2025-Gap opening of image potential stat.pdf](https://mdr.nims.go.jp/filesets/1e2e688f-4f37-4cea-be3a-1f71ba9cfbc1/download)

## Creator

Pratyay Amrit, Naoya Kawakami, Noriaki Takagi, Hiroshi Ishida, Chun-Liang Lin, [Ryuichi Arafune](https://orcid.org/0000-0003-4371-6116)

## Rights

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

## Other metadata

[Gap opening of image potential state on reconstructed Ir(001) surface](https://mdr.nims.go.jp/datasets/60ab20b2-4e7d-4c16-ae52-d48dc7707c7d)

## Fulltext

Gap opening of image potential state on reconstructed Ir(001) surfacePHYSICAL REVIEW RESEARCH 7, 023294 (2025)Gap opening of image potential state on reconstructed Ir(001) surfacePratyay Amrit ,1 Naoya Kawakami ,1 Noriaki Takagi ,2 Hiroshi Ishida ,3Chun-Liang Lin ,1,* and Ryuichi Arafune 4,†1Department of Electrophysics, National Yang Ming Chiao Tung University, 1001 University Rd., Hsinchu 300, Taiwan2Graduate School of Human and Environmental Studies, Kyoto University, Kyoto 606-8501, Japan3College of Humanities and Sciences, Nihon University, Tokyo, Japan4Research Center for Materials Nanoarchitectonics (MANA), National Institute for Materials Science (NIMS),1-1 Namiki, Ibaraki 304-0044, Japan(Received 23 February 2025; accepted 3 June 2025; published 23 June 2025)We have investigated the image potential states (IPSs) for the unreconstructed Ir(001)-(1×1) and thereconstructed Ir(001)-(5×1) surfaces by using two-photon photoemission spectroscopy. We reveal that theIr(001)-(5×1) reconstructed surface covered by Xe adatoms exhibits a band gap of 100 meV, which arisesfrom the new periodic potential created by the reconstruction, whereas no band gap was observed on theIr(001)-(1×1) surface within the measured momentum range. The experimental results are compared with thetheoretical ones obtained within density functional theory for both the clean and Xe-covered Ir(001)-(5×1)reconstructed surfaces. The planar averaged charge density distributions of the IPS for both the surfaces showthat the Xe adsorption does not significantly alter the positions of charge density maxima, which rationalizes whythe band gap arising from the substrate superlattice potential does not change significantly upon Xe adsorption.The agreement between the experimentally observed band gap for the IPS on the Xe-covered surface and thetheoretical calculations highlights the robustness of IPS behavior under Xe adsorption.DOI: 10.1103/46k5-fft2I. INTRODUCTIONSurface reconstruction resulting from the rearrangementof surface atoms usually plays an important role in alteringthe electronic properties of materials. Experimental studieshave shown that surface reconstruction affects the occu-pied electronic states [1–5]. However, the effect of surfacereconstruction on the unoccupied states, particularly image-potential states (IPSs), remains largely unexplored. IPSs areunique surface-bound electronic states that exist near thevacuum level [6,7]. They exhibit remarkable characteristics,such as long lifetimes and free electronlike effective mass. Adeeper understanding of how atomic rearrangements influenceIPSs could shed light on fundamental electron dynamics atsurfaces and pave the way for potential applications in deviceslike quantum dot transistors [8–10] and ultrafast photodetec-tors [11].Previously, many studies have demonstrated the effectsof adsorbates on metal surfaces and their impact on IPSs[12–21]. For example, Muller et al. have deposited FePc onAg(111) and observed a strong modification of IPSs [22],indicating that IPSs can be altered by a periodic potential ofthe molecular array. Rather than adding additional adsorbates*Contact author: clin@nycu.edu.tw†Contact author: arafune.ryuichi@nims.go.jpPublished by the American Physical Society under the terms of theCreative Commons Attribution 4.0 International license. Furtherdistribution of this work must maintain attribution to the author(s)and the published article’s title, journal citation, and DOI.on the surface, a self-reconstructed surface is much morefavorable to study the influence of how the rearrangementof atoms can affect the IPSs while excluding the additionalcharge transfer from adsorbates.In this study, we will investigate the effect of surface re-construction on IPSs by applying two-photon photoemission(2PPE) spectroscopy on Ir(001) surface, which forms (1×1)metastable unreconstructed and (5×1) reconstructed struc-tures. Here, Xe atoms are used to adjust the work functionof Ir(001) surface to fit the photon energy of 2PPE experi-ments. Our findings revealed that the IPS on the Xe-coveredIr(001)-(5×1) surface exhibits a band gap, whereas the corre-sponding one on the Xe-covered Ir(001)-(1×1) surface doesnot. The difference is attributed to the distinct reconstructionsof the two surfaces, i.e., IPSs are sensitive to the variation ofthe potential parallel to the surface caused by the symmetrybreaking. We further examined the influence of Xe adsorptionon IPSs by using theoretical calculations. The size of bandgap does not show significant change upon Xe adsorption,indicating that Xe primarily lowers the work function withoutsignificantly altering the unoccupied electronic states.II. EXPERIMENTAL AND THEORETICAL METHODSThe sample preparation was performed in an ultrahigh vac-uum (UHV) chamber with a base pressure of 1×10−10 mbar.The (5×1) reconstructed surface was prepared by the follow-ing treatments: First, the sample surface was sputtered with1.5 keV Ar ions for 30 mins, followed by exposure to 6×10−8mbar of O2 with heating at 1220 K for 10 mins to remove thecarbon impurities. O2 reacts with the carbon impurities seg-2643-1564/2025/7(2)/023294(8) 023294-1 Published by the American Physical Societyhttps://orcid.org/0000-0003-3664-8768https://orcid.org/0000-0001-9500-5557https://orcid.org/0000-0002-0799-9772https://orcid.org/0000-0003-2080-1561https://orcid.org/0000-0001-8781-3650https://orcid.org/0000-0003-4371-6116https://ror.org/00se2k293https://ror.org/02kpeqv85https://ror.org/05jk51a88https://ror.org/026v1ze26https://crossmark.crossref.org/dialog/?doi=10.1103/46k5-fft2&domain=pdf&date_stamp=2025-06-23https://doi.org/10.1103/46k5-fft2https://creativecommons.org/licenses/by/4.0/PRATYAY AMRIT et al. PHYSICAL REVIEW RESEARCH 7, 023294 (2025)FIG. 1. (a) Schematic illustration of the 2PPE process and the role of Xe adsorption in lowering the work function, thereby enablingpopulation of the IPSs. LEED patterns obtained with Eb = 99.5 eV for (b) the (1×1) surface and (c) the (5×1) surface. The series of 2PPEspectra of (d) the Xe/(1×1) and (e) the Xe/(5×1). The green and red solid curves in (d) and (e) represent the fitted curve (see text). 2PPEspectrum at k|| = 0.23 Å−1 for (f) the Xe/(1×1) and (g) the Xe/(5×1). Black dots represent the experimental data, and the red solid curve isthe fitted curve (see text). The pink and light blue curve represent the deconvoluted peak, and the blue line is a linear baseline.regated from the bulk and removes it from the surface [1,23–25]. Then, flash annealing in the UHV at 1420 K for 1 minresults in the (5×1) reconstruction. When preparing the (1×1)metastable surface, the reconstructed (5×1) structure waseliminated by annealing at 530 K for 10 mins, followed by O2adsorption at 1 ×10−7 mbar. The O2-exposed surface was thenrapidly annealed in the UHV environment for 30 secs at 750K. To completely remove the O2 from the surface, the samplewas exposed to H2 at 1 ×10−7 mbar for 5 mins at 530 K,resulting in a clean (1×1) surface [26,27].After the sample processing, the sample was transferred tothe analysis chamber constructed of μ metal, with a base pres-sure better than 8 ×10−11 mbar. The sample was cooled downto ∼12 K using liquid helium. The 2PPE experimental setupconsisted of a Ti:sapphire laser oscillator (Coherent, Mira900)with a pulse duration of 170 fs and a repetition rate of 80 MHz,and a hemispherical electron energy analyzer (Specs GmbH,Phoibos 100) equipped with a two-dimensional detector. Ul-traviolet light with an energy of 5.032 eV was generated byfrequency-tripling using β-BaB2O4 nonlinear crystals. All themeasurements were conducted using the p-polarized light,with the laser beam incident at 45◦ off normal to the sam-ple surface. The photoelectrons were analyzed with the passenergy set at 5 eV. Angle-resolved measurements were madealong the �̄ − X̄ direction. Due to insufficient photon energyin our setup to populate the IPS, Xe was adsorbed on the sam-ple prior to the 2PPE measurement to lower the work functionby exposing the sample to 1.0×10−8 mbar Xe gas for 5 minsat 40 K [12,28–32]. A schematic diagram of the 2PPE processand the role of Xe adsorption in lowering the work functionis illustrated in Fig. 1(a). By monitoring the work functionchange as a function of exposure time, we have judged that themonolayer Xe was formed. For the (5×1) surface, the workfunction decreased monotonically from 5.86 eV to 5.40 ±0.03 eV over 5 mins. For the (1×1) surface, the work functiondecreased to 5.24 ± 0.03 eV, and after that, it did not changesignificantly. In this study, the coverage of Xe (θ ) is defined asthe ratio between the Xe atoms and underlying Ir(001)-(1×1)atoms. The monolayer of Xe corresponds to θ = 0.45 fromthe model strucutre [24].We performed theoretical calculations using the computercode referenced in [33] within the framework of density func-tional theory (DFT), combined with the embedded Green’sfunction (EGF) technique [34] and the full-potential lin-earized augmented plane-wave method [35]. This approachwas used to calculate the electronic structure of semi-infiniteIr(001)-(5×1) and −(1×1) surfaces [36–38]. We used theDFT-LDA exchange-correlation energy functional since DFT-GGA underestimates work functions of 5d metals typicallyby a few tenths of eV. Furthermore, in order to be able toreproduce IPSs, the planar average of the short-range DFTone-electron potential, V̄eff (z) with z being the surface normalcoordinate, is admixed gradually with a model image potentialby using an interpolation function proposed by Nekovee andInglesfield [39],V (z) = [1 − ρ(z)]V̄eff (z) + ρ(z)[EV − 14(z − zim)], (1)where EV denotes the vacuum level, and p(z) varies smoothlyfrom 0 at z = zim (the image plane), to 1 at z = zb (theembedding surface on the vacuum side). The effects of theasymptotic Coulomb potential beyond zb are incorporated bythe embedding potential acting on z = zb. The image plane(zim) position is determined by the center of mass of thecharge density induced by a weak static electric field, withthe topmost Ir atom position defined as z = 0 [40–43].023294-2GAP OPENING OF IMAGE POTENTIAL STATE ON … PHYSICAL REVIEW RESEARCH 7, 023294 (2025)The input data for the aforementioned EGF program isthe atomic coordinates of each surface. For the clean (1×1)and (5×1) surfaces, we adopt the same atomic geometryas described in Ref. [38]. For the Xe-covered surfaces, weoptimized the atomic heights of Xe adatoms relative to theIr(001) substrate by using the VASP program [44,45]. Todo so, we employed the rev-VDW-DF2 functional [46,47] toapproximately account for van der Waals interactions.III. RESULTS AND DISCUSSIONSThe low energy electron diffraction (LEED) patterns inFigs. 1(b) and 1(c) confirm the high-quality Ir(001) surfaces,with clear diffraction spots for both the (1×1) and the (5×1)surfaces. Figures 1(d) and 1(e) show a series of 2PPE spec-tra of the IPS (n = 1) for the Xe-covered Ir(001)-(1×1)[Xe/(1×1)] and the Xe-covered Ir(001)-(5×1) [Xe/(5×1)],respectively. The spectra for the Xe/(1×1) consist of a singlepeak across all momentum. In contrast, the spectra of theXe/(5×1) exhibits a shoulder at k|| = 0.23 Å−1, which is thefirst Brillouin zone (BZ) boundary in the �̄ − X̄ direction forthe (5×1) surface, highlighted by the red curve.Figure 1(f) shows the spectrum for the Xe/(1×1) at k|| =0.23 Å−1. The black dots represent the experimental data,and the red curve indicates the fitted curve. The fitting wasperformed using the sum of the Lorentzian function (pink)and a linear baseline (blue), multiplied by the Fermi-Diracfunction to account for the Fermi level cutoff. The resultingfunction was then convoluted with the Gaussian function thatrepresents the instrument function. The full width at half max-imum of the Gaussian function from the fitting was 30 meV.The curve was well fitted using the single Lorentzian function,confirming that there is no gap in the IPS of the Xe/(1×1).The curve at the BZ boundary for the Xe/(5×1) is plottedin Fig. 1(g). The curves are well fitted by introducing twoLorentzian functions, as shown by the pink and light bluecurves. The peak positions for the pink and light blue curvesare 4.75 eV and 4.85 eV, respectively, corresponding to thetop and bottom of the lower and upper bands. These peakpositions observed in 2PPE experiments indicate a band gapof 100 ± 10 meV. Therefore, the spectra of the Xe/(5×1)exhibit a band gap in the IPS, whereas those of Xe/(1×1)do not.Figures 2(a) and 2(b) show the 2D band structure data ofthe IPS (n = 1) for the Xe/(1×1) and Xe/(5×1), using themaximum curvature method to highlight the curves [48]. Thered dashed curve in Fig. S1 represents the parabolic fitting[49], as shown in the Supplemental Material [50]. From thefitting, the effective mass was evaluated to be the same as themass of the free electron, indicating that the IPS retain theirfree-electron-like dispersion. The blue curve in Fig. S2(d) rep-resents the band structure calculated from the Kronig-Penney(K-P) model [51]. Briefly, we assume the one-dimensionalfree-electron Hamiltonian including the periodic rectangularpotential with a height of 0.14 eV and a width of 6.77 Å,shown in the Supplemental Material [50], which well char-acterizes the surface potential distribution calculated from theDFT calculations for the (5×1) surface. We obtain a 100 meVenergy gap. We do not observe the spectral intensity of thefolded band (k|| = 0.00 − 0.20 Å−1 for the upper branch andFIG. 2. 2D Band structure of the IPS for (a) the Xe/(1×1) and(b) the Xe/(5×1), measured via 2PPE at 12 K and shown using themaximum curvature method.k|| = 0.25 − 0.40 Å−1 for the lower branch) for the Xe/(5×1)in Fig. 2(b). The reason is unclear, but the faint or missingspectral intensities in the other half of the folded bands areconsistent with previous studies [17]. Note that, when ref-erenced to the parabolic fit plotted in Fig. S1, the deviationof the band above k|| = 0.34 Å−1 in Figs. 2(a) and 2(b)are the artifacts by the maximum curvature method, whichenhances the intensity change by the Fermi-Dirac distribu-tion. Our experimental results, supported by the K-P model,provide explicit evidence that the surface potential changedue to the surface reconstruction significantly influences theelectronic properties of the unoccupied states, particularly thefree-electron-like IPSs.It is well known that the Xe atoms form an overlayerincommensurate with the Ir(001) surface, modifying the elec-tronic structure by lowering the work function [28,52,53].Besides, the Xe layer may push the IPS away from the surface,weakening the effect of potential corrugation of the recon-structed surface to the IPS. Despite these, we experimentallyobserved a sizable band gap at the BZ boundary of the (5×1)lattice. To investigate the impact of Xe adsorption on theIPSs on both (1×1) and (5×1) surfaces, we performed DFTcalculations, including the image potential correction [54].Figure 3(a) shows the relaxed structural model for the clean(5×1) surface. The top view highlights the quasihexagonaltop layer formed by (5×1) reconstruction [27,55–57], with theunit cell marked by a black rectangle. The side view shows thecorrugation of the topmost layer [27,38]. Figure 3(b) showsthe DFT calculated band structure of the IPS for the (5×1)surface along the �̄- X̄ direction, where we used a bandunfolding method [58] to calculate the band structure beyondthe X̄ point. The calculated work function is 5.63 eV, and theenergy of the IPS (n = 1) at the �̄ point is 4.95 eV. A 60 meVband gap, highlighted in the zoomed-in image, is observed,resulting from surface reconstruction. In contrast, no bandgap is observed in the IPS for the (1×1) surface, as shownin Fig. 3(c). These results are in good agreement with theexperimental results and the K-P model fitting, as discussedabove.To elucidate whether the band gap of the clean recon-structed surface is affected by the Xe layer or not, weperformed the DFT calculations for the Xe/(5×1) surface.023294-3PRATYAY AMRIT et al. PHYSICAL REVIEW RESEARCH 7, 023294 (2025)FIG. 3. (a) Ball model showing the top and side view for the clean (5×1) reconstructed surface, with the unit cell highlighted by ablack rectangle box. Calculated band structure for (b) the clean (5×1) reconstructed surface and (c) the clean (1×1) unreconstructed surface,determined along the �̄-X̄ direction. The band structures are shown in logarithmic intensity. The zoomed-in area in (b) highlights the band gappresent.Since the monolayer Xe is incommensurate with the under-lying (5×1) surface, unfortunately, it is not feasible for DFTcalculation. Instead, we used a commensurate (5×2) structure[Xe/(5×2)] to investigate the effect of Xe adsorption, corre-sponding to θ = 0.3. Figure 4(a) shows the relaxed structuralmodel for the top and side view of the Xe/(5×2) surface, withXe atoms (green) and the unit cell (black rectangle). The Xeatoms adsorb on top of the Ir atoms with the Ir-Xe interplanedistance of 3.39 Å.Figure 4(b) shows the DFT calculated band structure of theIPS for the Xe/(5×2) surface. The work function calculatedfor the Xe/(5×2) surface is 4.74 eV. On comparing with theXe-free (5×1) surface, the work function is 0.89 eV loweron the Xe/(5×2) surface. The energy of the IPS (n = 1) atthe �̄ point is 3.95 eV, which is 0.10 eV deeper than thaton the Xe-free (5×1) surface. A band gap, highlighted inthe zoomed-in image, of 45 meV is observed, which is 15meV smaller than that of the Xe-free (5×1) surface. Theseresults show that the band gap persists even after Xe ad-sorption. As for a Xe-covered unreconstructed surface, weconsidered a (√2×√2) lattice of Xe adatoms on unrecon-structed Ir(001) [Xe/(√2×√2)] with the coverage of θ = 0.5[59]. The work function, calculated from DFT, is 4.99 eV,which is 0.92 eV lower than that of the Xe-free (1×1) surface.The energy of the IPS (n = 1) at the �̄ point is 4.36 eV. TheXe/Ir(001)-(√2×√2) surface, shown in Fig. 4(c), does notexhibit a band gap. This suggests that the surface reconstruc-tion is the dominant factor in band gap formation in the IPS,even in the presence of the Xe layer.To further understand the effect of the Xe adsorption onthe IPS, we examined the change in the charge density distri-bution of the IPS (n = 1) near the surface. The blue curvein Fig. 5(a) shows the planar averaged charge density dis-tribution of the IPS at the �̄ point for the Xe-free (1×1)surface, with z = 0 corresponding to the outermost Ir atoms(Ir1, indicated by purple triangle). The blue dashed line at1.64 Å indicates the zim position obtained from the DFTcalculations for the Xe-free (1×1) surface. The red curvein Fig. 5(a) shows the charge density distribution of the IPSfor the Xe/(√2×√2) with the outermost Xe atom (greentriangle). The zim (red dashed line) position obtained fromthe calculation is −0.57 Å inside the Xe atomic layer. Thezim position in the interior of the Xe layer is particularlyinteresting and we will discuss it later. Despite the Xe layerbeing 3.39 Å above the topmost Ir1 atom, the maximum IPScharge density occurs nearly at the same positions in bothsurfaces: z = 5.62 Å for Xe free (1×1) and z = 6.15 Åfor Xe/(√2×√2). This similarity in the maxima positionsof charge density suggests that Xe adsorption does notFIG. 4. (a) Ball model showing the top and side view for the Xe/(5×2) surface, with the unit cell marked by a black rectangular box.The Xe atoms are represented by green spheres. Calculated band structure for (b) the Xe/(5×2) surface and (c) the Xe/(√2×√2) surface,determined along the �̄-X̄ direction. The band structure is shown in logarithmic intensity. The zoomed-in area in (b) highlights the band gap.023294-4GAP OPENING OF IMAGE POTENTIAL STATE ON … PHYSICAL REVIEW RESEARCH 7, 023294 (2025)FIG. 5. Planar average charge density (ρ ) distribution at the �point of the IPS state for (a) the clean Ir(001)-(1×1) surface (blue)and the Xe/(√2×√2) surface (red). (b) The charge density for theclean Ir(001)-(5×1) surface (blue) and the Xe/(5×2) surface (red).In both (a) and (b), the image plane (zim ) position for each structure isshown by red and blue dashed lines, respectively. In addition, the po-sition of the topmost Ir (Ir1) and Xe atoms are marked by the purpleand green triangles, respectively. (c) Calculated integrated density ofstates (DOS) projected onto the Xe adatoms (blue) and the Ir atoms(red) in the reconstructed topmost Ir layer of the Xe/Ir(001)-(5×1)surface.significantly affect the charge density distribution of the IPSon the unreconstructed surfaces.Figure 5(b) shows the planar-averaged charge density dis-tributions at the �̄ point of the IPS (n = 1) for the Xe-free(5×1) surface (blue curve) and the Xe/(5×2) surfaces (redcurve). Regarding the Xe-covered surface, the IPS at �̄ islocated within a projected bulk band, and the charge-densityoscillations in the interior of the Ir substrate seen in Fig. 5(b)arise from the bulk states in the same energy range and arenot due to the IPS. It is technically not possible to extractonly the charge density of the IPS from the calculated Green’sfunction. By comparing the charge densities in Figs. 5(a) and5(b), we observe that the charge density maxima move closerto the metal surface in the reconstructed surfaces than the un-reconstructed ones. This may occur because the reconstructedsurface has a 20% higher atomic density than the unrecon-structed one, which results in higher electron densities and themore attractive exchange-correlation one-electron potential atthe surface. Again, zim position for Xe-covered surface (reddashed line) is located at the interior of the topmost Xe layer.The overall shape of the blue and red curves are very similar,except for a shoulder in the red curve at z ≈ 5.15 Å, which isattributed to the charge density from the Xe atoms. The nearlyidentical position of the charge density maxima across bothcurves confirm that Xe adsorption has a minor effect on thecharge distribution of the IPS, especially on the reconstructedsurfaces.While one might expect Xe adsorption to significantly shiftthe charge density maxima away from the Ir surface, thisis not observed. Two major factors may contribute to thisbehavior: the positive electron affinity of Xe and the imageplane position with respect to the Xe overlayer.Previous studies have demonstrated that the interactionbetween a metal surface and a dielectric layer induces chargeredistribution across the interface, creating a dipole that shiftsthe Fermi level (EF ) in the metal [29,52,60,61] and polarizesthe Xe atoms [62]. The positive electron affinity of Xe playsa crucial role in its interaction with the IPS. Unlike materialswith negative electron affinity, such as Ar, which provide arepulsive barrier and push the IPS wave function toward thevacuum region [63], the positive electron affinity of Xe weak-ens the screening effect. The reduced screening allows partialpenetration of the IPS wave function into the Xe layer [15,64].As a result, the charge density remains largely unaffected,with the IPS still strongly bound near the surface.The zim position observed for the Xe-covered surfaces fur-ther supports these results. The image planes of low-indexsurfaces of elemental metals are typically located at 1–2 Åon the vacuum side of the topmost lattice plane [40,42,43].For clean Ir(001), our DFT calculation yielded zim = 1.59 Åfor the (1×1) surface and zim = 1.63 Å as measured fromthe averaged z coordinate of the corrugated reconstructedlayer for the (5×1) surface. Similarly, in the jellium model,where the nuclear charge is treated as the uniform positivebackground charge, the zim is not located exactly at the edgeof this uniform background but is displaced by approximately0.6–0.8 Å toward the vacuum, depending on the bulk electrondensity parameter [40,42,43]. However, for the Xe-coveredsurfaces, the image plane is located at the interior of the Xelayer, indicating that the shift of zim by Xe adsorption is lessthan expected. This can be attributed to the dielectric nature ofthe Xe layer, which weakens the strength of the image forcecompared to the clean metal surface, limiting its ability topush zim further away from the surface [62]. It indicates that023294-5PRATYAY AMRIT et al. PHYSICAL REVIEW RESEARCH 7, 023294 (2025)Xe is less polarized than the Ir substrate. To further supportthis argument, we calculated the integrated density of states(DOS) for the Xe-covered (5×2) surface. The blue and redlines in Fig. 5(c) show the DOS projected on the Xe adatomsand Ir atoms in the reconstructed Ir topmost layer, respec-tively. Within linear response theory, the screening chargeinduced by an external field is created by electronic statesnear the Fermi level. The nearly vanishing DOS near theFermi level of the Xe atoms shown in Fig. 5(c) indicates thatthe Xe overlayer cannot screen the external field effectively.Therefore, the image plane position is located at the interiorof the topmost Xe layer.The combination of weak positive electron affinity and di-electric screening explains why the charge density distributionremains unchanged after the Xe adsorption. Since the chargedensity distribution of the IPS remains nearly unchanged intheir position relative to the reconstructed surface after Xeadsorption, the in-plane potential also remains largely unaf-fected. Therefore, we conclude that the band gap observedin the IPS is an intrinsic feature of the Ir(001) caused bysurface reconstruction and not significantly influenced by theXe adsorption.Finally, we comment on the possible role of Xe-inducedeffects beyond simple work function modulation. As dis-cussed by Güdde and Höfer [15], rare-gas layers can introduceburied image-potential-like interface states or modify the IPSvia electrostatic potential corrugation. These effects are moreprominent in thicker layers with negative electron affinitygases. In contrast, we use only monolayer Xe of positiveelectron affinity and our DFT results do not show the in-terface states. Furthermore, the energy gap is observed evenwithout Xe, reinforcing that the gap originates from the un-derlying (5×1) reconstruction rather than any hybridizationor interface-localized states. We thus conclude that the mono-layer Xe serves primarily to reduce the work function andplays no essential role in the origin of the band gap. It wouldbe interesting to detect buried interface states by using athicker Xe sample.IV. CONCLUSIONIn this study, we investigated the impact of surface re-construction on the IPSs using the unreconstructed and thereconstructed Ir(001) surfaces by 2PPE spectroscopy andDFT calculations. Our 2PPE results experimentally revealedthe band gap in the IPS of the Xe/Ir(001)-(5×1) surface,which was absent in the Xe/Ir(001)-(1×1) surface. Thisband gap in the unoccupied state is caused by the changein the potential distribution due to surface reconstruction.DFT calculations confirmed that the band gap persists inthe reconstructed surface with or without Xe adsorption.The results of this work demonstrate the robustness of sur-face reconstruction as a tool for tuning the unoccupiedstates.ACKNOWLEDGMENTSThis work was financially supported by JSPS KAK-ENHI (Grant No. 22H01960) and the World PremierInternational Research Center Initiative (WPI) on Mate-rials Nanoarchitectonics, MEXT, Japan. It is also par-tially supported by the National Science and Technol-ogy Council, Taiwan under NSTC 114–2923-M-A49-001-MY2, NSTC 113–2124-M-A49-007, NSTC 113–2628-M-A49-006-MY3, and NSTC T-Star Center Project: Fu-ture Semiconductor Technology Research Center underNSTC 114–2634-F-A49-001.DATA AVAILABILITYThe data that support the findings of this article are notpublicly available upon publication because it is not techni-cally feasible and/or the cost of preparing, depositing, andhosting the data would be prohibitive within the terms of thisresearch project. The data are available from the authors uponreasonable request.[1] T. N. Rhodin and G. Brodén, Preparation and chemisorp-tive properties of the clean normal and reconstructed sur-faces of Ir(100)—role of multiplets, Surf. Sci. 60, 466(1976).[2] D. Spišák and J. Hafner, Reconstruction and de-reconstructionof the Ir(1 0 0) surface and ultrathin Fe/Ir(1 0 0) Films,Surf. Sci. 546, 27 (2003).[3] A. Nuber, M. Higashiguchi, F. Forster, P. Blaha, K. Shimada,and F. Reinert, Influence of reconstruction on the surface stateof Au(110), Phys. Rev. B 78, 195412 (2008).[4] N. Memmel, Monitoring and modifying properties of metalsurfaces by electronic surface states, Surf. Sci. Rep. 32, 91(1998).[5] A. Bendounan, H. Cercellier, Y. Fagot-Revurat, B. Kierren,V. Y. Yurov, and D. Malterre, Modification of shockley statesinduced by surface reconstruction in epitaxial Ag films onCu(111), Phys. Rev. B 67, 165412 (2003).[6] P. M. Echenique and J. B. Pendry, The existence and detectionof rydberg states at surfaces, J. Phys. C: Solid State Phys. 11,2065 (1978).[7] P. M. Echenique and J. B. Pendry, Theory of image states atmetal surfaces, Prog. Surf. Sci. 32, 111 (1989).[8] Y. Zhang, Z. Yang, X. Zhang, B. Lin, G. Lin, and J. Chen,Coulomb-coupled quantum-dot thermal transistors, Europhys.Lett. 122, 17002 (2018).[9] F. Craes, S. Runte, J. Klinkhammer, M. Kralj, T. Michely, andC. Busse, Mapping image potential states on graphene quantumdots, Phys. Rev. Lett. 111, 056804 (2013).[10] Y. Zhang, Q. Chen, A. P. Alivisatos, and M. Salmeron, Dynamiccharge carrier trapping in quantum dot field effect transistors,Nano Lett. 15, 4657 (2015).[11] Q. Wang et al., Ultrafast broadband photodetectors based onthree-dimensional dirac semimetal Cd3As2, Nano Lett. 17, 834(2017).[12] K. Wandelt and J. E. Hulse, Xenon adsorption on palladium.I. the homogeneous (110), (100), and (111) Surfaces, J. Chem.Phys. 80, 1340 (1983).[13] J. L. F. Da Silva, C. Stampfl, and M. Scheffler, Xe adsorptionon metal surfaces: First-principles investigations, Phys. Rev. B72, 075424 (2005).023294-6https://doi.org/10.1016/0039-6028(76)90329-0https://doi.org/10.1016/j.susc.2003.08.052https://doi.org/10.1103/PhysRevB.78.195412https://doi.org/10.1016/S0167-5729(98)00006-5https://doi.org/10.1103/PhysRevB.67.165412https://doi.org/10.1088/0022-3719/11/10/017https://doi.org/10.1016/0079-6816(89)90015-4https://doi.org/10.1209/0295-5075/122/17002https://doi.org/10.1103/PhysRevLett.111.056804https://doi.org/10.1021/acs.nanolett.5b01429https://doi.org/10.1021/acs.nanolett.6b04084https://doi.org/10.1063/1.446815https://doi.org/10.1103/PhysRevB.72.075424GAP OPENING OF IMAGE POTENTIAL STATE ON … PHYSICAL REVIEW RESEARCH 7, 023294 (2025)[14] J. L. F. de Silva, C. Stampfl, and M. Scheffler, Adsorption ofXe atoms on metal surfaces: New insights from first-principlescalculations, Phys. Rev. Lett. 90, 066104 (2003).[15] J. Güdde and U. Höfer, Femtosecond time-resolved studiesof image-potential states at surfaces and interfaces of rare-gasadlayers, Prog. Surf. Sci. 80, 49 (2005).[16] T. Yamada, N. Kawakita, C. Okui, and T. Munakata, Hy-bridization of an unoccupied molecular orbital with an imagepotential state at a lead phthalocyanine/graphite interface,J. Phys.: Condens. Matter 31, 044004 (2018).[17] B. W. Caplins, A. J. Shearer, D. E. Suich, E. A. Muller,and C. B. Harris, Measuring the electronic corrugation at themetal/organic interface, Phys. Rev. B 89, 155422 (2014).[18] N. Armbrust, J. Güdde, P. Jakob, and U. Höfer, Time-resolvedtwo-photon photoemission of unoccupied electronic states ofperiodically rippled graphene on Ru(0001), Phys. Rev. Lett.108, 056801 (2012).[19] X. Wang, X. Shen, R. Osgood, R. Haight, and F. Himpsel,Observation of lateral superlattice effects on stepped Cu(001),Phys. Rev. B 53, 15738 (1996).[20] I. G. Hill and A. B. Mc Lean, Strongly anisotropic band dis-persion of an image state located above metallic nanowires,Phys. Rev. Lett. 82, 2155 (1999).[21] S. Bode, K. Starke, P. Rech, and G. Kaindl, Highly spin-polarized, nearly free-electron states in front of Co(101¯0),Phys. Rev. Lett. 72, 1072 (1994).[22] E. A. Muller, J. E. Johns, B. W. Caplins, and C. B. Harris,Quantum confinement and anisotropy in thin-film molecularsemiconductors, Phys. Rev. B 83, 165422 (2011).[23] X. Cao et al., Revisiting oxygen adsorption on Ir(100), J. Phys.Chem. C 126, 10035 (2022).[24] A. Ignatiev, A. V. Jones, and T. N. Rhodin, Leed investigationsof xenon single crystal films and their use in studying theIr(100) surface, Surf. Sci. 30, 573 (1972).[25] K. Anic, A. V. Bukhtiyarov, H. Li, C. Rameshan, and G.Rupprechter, CO adsorption on reconstructed Ir(100) surfacesfrom UHV to mbar pressure: A LEED, TPD, and PM-IRASStudy, J. Phys. Chem. C 120, 10838 (2016).[26] K. Heinz, G. Schmidt, L. Hammer, and K. Müller, dynamics ofthe reconstruction process Ir(100) 1×1 → 1×5, Phys. Rev. B32, 6214 (1985).[27] A. Schmidt, W. Meier, L. Hammer, and K. Heinz, deep-goingreconstruction of Ir(100)-5×1, J. Phys.: Condens. Matter 14,12353 (2002).[28] J. Ross Macdonald and C. A. Barlow, Work function change onmonolayer adsorption, J. Chem. Phys. 39, 412 (1963).[29] K. Wandelt, The local work function: Concept and implications,Appl. Surf. Sci. 111, 1 (1997).[30] T. Fauster and W. Steinmann, in Electromagnetic Waves: RecentDevelopments in Research, 1st ed. (Elsevier, Amsterdam, 1995),Vol. 2, Chap. 8, pp. 347–411.[31] N. D. Lang and W. Kohn, Theory of metal surfaces: Workfunction, Phys. Rev. B 3, 1215 (1971).[32] J. Küppers, H. Michel, F. Nitschké, K. Wandelt, and G. Ertl,Xenon adsorption as a tool for local surface structure determi-nation at Ir(100) surfaces, Surf. Sci. 89, 361 (1979).[33] H. Ishida, Surface-embedded green-function method: A for-mulation using a linearized-augmented-plane-wave basis set,Phys. Rev. B 63, 165409 (2001).[34] J. E. Inglesfield, A method of embedding, J. Phys. C Solid StatePhys. 14, 3795 (1981).[35] D. J. Singh and L. Nordstrom, Planewaves, Pseudopotentials,and the LAPW Method (Springer Science & Business Media,New York, 2006).[36] H. Ishida, Rashba spin splitting of shockley surface states onsemi-infinite crystals, Phys. Rev. B 90, 235422 (2014).[37] T. Nakazawa, N. Takagi, M. Kawai, H. Ishida, and R. Arafune,Rashba splitting in an image potential state investigated bycircular dichroism two-photon photoemission spectroscopy,Phys. Rev. B 94, 115412 (2016).[38] H. Ishida, Surface band structure of the reconstructed Ir(001)-(5×1) surface, Surf. Sci. 744, 122472 (2024).[39] M. Nekovee, Theory of image states at magnetic surfaces,Prog. Surf. Sci. 50, 149 (1995).[40] N. D. Lang and W. Kohn, Theory of metal surfaces: Inducedsurface charge and image potential, Phys. Rev. B 7, 3541(1973).[41] N. D. Lang, Interaction between closed-shell systems and metalsurfaces, Phys. Rev. Lett. 46, 842 (1981).[42] R. O. Jones, P. J. Jennings, and O. Jepsen, Surface barrier inmetals: A new model with application to W(001), Phys. Rev. B29, 6474 (1984).[43] P. J. Jennings, R. O. Jones, and M. Weinert, Surface barrier forelectrons in metals, Phys. Rev. B 37, 6113 (1988).[44] G. Kresse and J. Hafner, Ab initio molecular dynamics for liquidmetals, Phys. Rev. B 47, 558 (1993).[45] G. Kresse and J. Furthmüller, Efficiency of ab-initio to-tal energy calculations for metals and semiconductors us-ing a plane-wave basis set, Comput. Mater. Sci. 6, 15(1996).[46] J. Klimeš, D. R. Bowler, and A. Michaelides, Van Der Waalsdensity functionals applied to solids, Phys. Rev. B 83, 195131(2011).[47] I. Hamada, Van Der Waals density functional made accurate,Phys. Rev. B 89, 121103 (2014).[48] P. Zhang, P. Richard, T. Qian, Y.-M. Xu, X. Dai, and H. Ding,A precise method for visualizing dispersive features in imageplots, Rev. Sci. Instrum. 82, 043712 (2011).[49] T. Nakazawa, R. Arafune, N. Takagi, and M. Kawai, Linewidthanalysis of image potential states on noble metal surfaces withhigh-energy resolved two-photon photoemission spectroscopy,Surf. Interface Anal. 48, 1194 (2016).[50] See Supplemental Material at http://link.aps.org/supplemental/10.1103/46k5-fft2 for details on the parabolic fitting of the IPSon the Xe/(1×1) surface and the band structure calculationusing the Kronig-Penney model for the Ir(001)-(5×1) surface.[51] R. D. L. Kronig and W. G. Penney, Quantum mechanics ofelectrons in crystal lattices, Proc. R. Soc. London Ser. A 130,499 (1931).[52] K. Wandelt, Surface and Interface Science (John Wiley & Sons,Hoboken, NJ, 2016), Vol. 5.[53] C. Hückstädt, S. Schmidt, S. Hüfner, F. Forster, F. Reinert,and M. Springborg, Work function studies of rare-gas/noblemetal adsorption systems using a kelvin probe, Phys. Rev. B73, 075409 (2006).[54] G. Butti, S. Caravati, G. P. Brivio, M. I. Trioni, and H. Ishida,Image potential states and electronic structure of Na/Cu(111),Phys. Rev. B 72, 125402 (2005).023294-7https://doi.org/10.1103/PhysRevLett.90.066104https://doi.org/10.1016/j.progsurf.2005.10.003https://doi.org/10.1088/1361-648X/aaf08ehttps://doi.org/10.1103/PhysRevB.89.155422https://doi.org/10.1103/PhysRevLett.108.056801https://doi.org/10.1103/PhysRevB.53.15738https://doi.org/10.1103/PhysRevLett.82.2155https://doi.org/10.1103/PhysRevLett.72.1072https://doi.org/10.1103/PhysRevB.83.165422https://doi.org/10.1021/acs.jpcc.2c01237https://doi.org/10.1016/0039-6028(72)90047-7https://doi.org/10.1021/acs.jpcc.5b12494https://doi.org/10.1103/PhysRevB.32.6214https://doi.org/10.1088/0953-8984/14/47/310https://doi.org/10.1063/1.1734263https://doi.org/10.1016/S0169-4332(96)00692-7https://doi.org/10.1103/PhysRevB.3.1215https://doi.org/10.1016/0039-6028(79)90622-8https://doi.org/10.1103/PhysRevB.63.165409https://doi.org/10.1088/0022-3719/14/26/015https://doi.org/10.1103/PhysRevB.90.235422https://doi.org/10.1103/PhysRevB.94.115412https://doi.org/10.1016/j.susc.2024.122472https://doi.org/10.1016/0079-6816(95)00050-Xhttps://doi.org/10.1103/PhysRevB.7.3541https://doi.org/10.1103/PhysRevLett.46.842https://doi.org/10.1103/PhysRevB.29.6474https://doi.org/10.1103/PhysRevB.37.6113https://doi.org/10.1103/PhysRevB.47.558https://doi.org/10.1016/0927-0256(96)00008-0https://doi.org/10.1103/PhysRevB.83.195131https://doi.org/10.1103/PhysRevB.89.121103https://doi.org/10.1063/1.3585113https://doi.org/10.1002/sia.6083http://link.aps.org/supplemental/10.1103/46k5-fft2https://doi.org/10.1098/rspa.1931.0019https://doi.org/10.1103/PhysRevB.73.075409https://doi.org/10.1103/PhysRevB.72.125402PRATYAY AMRIT et al. PHYSICAL REVIEW RESEARCH 7, 023294 (2025)[55] M. A. Van Hove, R. J. Koestner, P. C. Stair, J. P. Bibérian, L.L. Kesmodel, I. Bartoš, and G. A. Somorjai, The surface recon-structions of the (100) crystal faces of iridium, platinum andgold. II. structural determination by LEED intensity analysis,Surf. Sci. 103, 218 (1981).[56] H. C. Poon, D. K. Saldin, D. Lerch, W. Meier, A. Schmidt,A. Klein, S. Müller, L. Hammer, and K. Heinz, Spontaneoussymmetry breaking of the Ir(100)-(5×1)-Hex surface inducedby hydrogen adsorption, Phys. Rev. B 74, 125413 (2006).[57] G. Gilarowski, J. Méndez, and H. Niehus, Initial growth of Cuon Ir(100)-(5×1), Surf. Sci. 448, 290 (2000).[58] H. Ishida, R. Arafune, and N. Takagi, First-principles cal-culation of the graphene dirac band on semi-infinite Ir(111),Phys. Rev. B 102, 195425 (2020).[59] This structure is a hypothetically assumed model and cannot berealized in practice because the lattice constant of the Xe latticebecomes shorter than that of the close-packed Xe Lattice.[60] S. Prada, U. Martinez, and G. Pacchioni, Work function changesinduced by deposition of ultrathin dielectric films on metals: Atheoretical analysis, Phys. Rev. B 78, 235423 (2008).[61] T. Engel and R. Gomer, Adsorption of inert gases on tungsten:Measurements on single crystal planes, J. Chem. Phys. 52, 5572(1970).[62] W. R. Merry, R. E. Jordan, D. F. Padowitz, and C. B. Harris,Electrons at metal-insulator interfaces: I. The effect of Xemonolayers on the image potential states of Ag(111), Surf. Sci.295, 393 (1993).[63] P. M. Echenique, R. Berndt, E. V. Chulkov, T. Fauster, A.Goldmann, and U. Höfer, Decay of electronic excitations atmetal surfaces, Surf. Sci. Rep. 52, 219 (2004).[64] W. Berthold, F. Rebentrost, P. Feulner, and U. Höfer, Influ-ence of Ar, Kr, and Xe layers on the energies and lifetimesof image-potential states on Cu(100), Appl. Phys. A 78, 131(2004).023294-8https://doi.org/10.1016/0039-6028(81)90108-4https://doi.org/10.1103/PhysRevB.74.125413https://doi.org/10.1016/S0039-6028(99)01255-8https://doi.org/10.1103/PhysRevB.102.195425https://doi.org/10.1103/PhysRevB.78.235423https://doi.org/10.1063/1.1672827https://doi.org/10.1016/0039-6028(93)90286-Shttps://doi.org/10.1016/j.surfrep.2004.02.002https://doi.org/10.1007/s00339-003-2310-6