Dataset and article "Helium 1s photoemission and photon stimulated desorption of He+ ions by double excitations from adsorbed helium layers: Zero-point motion and matrix effects"

 View Online  Export Citation RESEARCH ARTICLE | SEPTEMBER 01 2024 Helium 1s photoemission and photon stimulated desorption of He+ ions by double excitations from adsorbed helium layers: Zero-point motion and matrix effects Special Collection: Physics and chemistry of low temperature atoms and molecules S. J. Riepl; S. Kossler; J. Braun; J. Minár; J. V. Barth; P. Feulner  Low Temp. Phys. 50, 713–721 (2024) https://doi.org/10.1063/10.0028136 Articles You May Be Interested In Effects of core space and excitation levels on ground-state correlation and photoionization dynamics of Be and Ne J. Chem. Phys. (February 2019) Graphene enhanced resonant Raman spectroscopy of gallium nitride nanocrystals Appl. Phys. Lett. (June 2025) 30 October 2025 09:20:02 Helium 1 s photoemission and photon stimulated desorption of He + ions by double excitations from adsorbed helium layers: Zero-point motion and matrix effects Cite as: Fiz. Nizk. Temp. 50,796–805 (September 2024); doi: 10.1063/10.0028136 View Online Export Citation CrossMar k Submitted: 20 July 2024 S. J. Riepl, 1 S. Kossler, 1 J. Braun, 2 J. Minár, 3 J. V. Barth, 1 and P. Feulner 1,a) AFFILIATIONS 1 School of Natural Sciences, Physik E20, Technische Universität München, D-85748 Garching, Germany 2 Department Chemie, Ludwig-Maximilians-Universität München, D-81377 München, Germany 3 University of West Bohemia, New Technologies-Research Center, Pilsen, Czech Republic a) Author to whom correspondence should be addressed: [email protected] ABSTRACT Excited with p-polarized light, the near-edge He 1sphotoemission signal from monolayers of He adsorbed on the close-packed surfaces of silver, copper, ruthenium, and platinum shows periodic splitting with photoelectron momentum. By applying a simple single scattering model, we explain this effect by zero-point motion induced variation of the photo hole’s image charge screening and interference of the photoelectron’s final state wavefunction. Relativistic one-step photoemission calculations support this interpretation. In the second part of our study, we investigate neutral double excitations of He bilayers adsorbed on platinum by monitoring the emission of He + ions. We identify strong matrix and zero-point motion effects, namely resonances absent for isolated He, density and light polarization dependence, line broadening and a correlation of excitation and kinetic ion-energy. © 2024 Author(s). All article content, except where otherwise noted, is licensed under a Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). https://doi.org/10.1063/10.0028136 1. INTRODUCTION Quantum effects are important for condensed helium. The liquid gets superfluid at the lambda point, and solidification occurs only at elevated pressure due to the large zero-point motion energy. Neutron scattering yields a zero-point energy of ∼26 K for solid 4 He, 1 more than the depth of the potential well of ∼11 K for He atoms. 2 For 3 He quantum effects are even more important (see the review of Beamish and Balibar 2 for more zero-point motion and other quantum effects of solid He). Even for more strongly bound 3 He and 4 He on metal surfaces the zero-point energy is appreciable: for a monolayer of 4 He on the Ag(111) surface, one of our examples described below, it equals about one-quarter of the depth of the physisorption potential. 3 Zero-point motion can have a large effect on electronic excitation energies measured for adsorbed or condensed He. In photoemission studies of He layers on metals, energy shifts by image potential screening of the photo hole will depend on the exact vertical position of the He atom at the instant of ionization, widening the range of recorded kinetic electron energies. For neutral electronic excitations line broadening exists as well. The electron affinity of condensed He of 1.3 eV is large, 4 and for excited species the enlarged wavefunctions and electronic spread account for strong Pauli repulsion with the closed-shell atoms of the matrix. For singly excited He bilayers on Ru(0001) we have observed blue-shifted lines as a function of the displacement from the equilibrium position as expected, but also peak splitting due to symmetry breaking; 5 for superfluid 4 He droplets line broadening was found as well. 6 In the present study, we show how these line broadening effects by zero-point motion can at least partly be resolved with improved spectral resolution. Our first example is near-threshold photoemission from He monolayers on close-packed metal surfaces, resulting in low kinetic energies of the photoelectrons. We excited the He 1slevel with photons polarized along the surface normal (A z -light), yielding a Low Temperature Physics ARTICLE pubs.aip.org/aip/ltp Low Temp. Phys. 50, 713 (2024); doi: 10.1063/10.0028136 50, 713 ©Author(s)2024 30 October 2025 09:20:02 p z -wave with two lobes, one directed towards the vacuum and one towards the surface. For low electron energies, the elastic backscattering cross section is large, 7 and the lobe directed toward the vacuum interferes with that backscattered from the surface. Due to image potential screening the photoelectron energy depends on the exact vertical position at the instant of ionization, 8 which maps the zero-point distribution onto the photoelectron’s kinetic energy distribution and the interference pattern of the two parts of the p z -wave. Our second example addresses double excitations of He bilayers on Pt(111). Since the seminal experiment of Madden and Codling 9 that showed only one strong Rydberg series instead of three as expected for non-correlated electrons, and the first explanation of these findings by the theoretical paper of Cooper, Fano, and Pratts, 10 doubly excited He atoms (He**) became a paradigm system for strong electron correlation. Experimental and theoretical work over many years has led to a detailed and fully quantitative understanding of this system, see, e.g., the review by Tanner, Richter, and Rost. 11 Later the energy range with overlapping Rydberg states close to the limit of complete fragmentation at 79 eV caught attention as an example for quantum chaos, 12 with analogies to celestial mechanics. 13 Isolated He** atoms decay via autoionization (AI) that interferes with direct ionization, leading to Fano line shapes. More recently, experiments with attosecond time resolution on He** enabled the analysis of the temporal build-up of Fano profiles for the first time. 14 Concepts for future experiments on attosecond electron dynamics of He** are described in Ref. 15. Experiments on double excitations of condensed He are rare. LaForge et al. 16 investigated electron and ion emission from He droplets of different sizes by excitation with synchrotron radiation. They found broadened and blue-shifted absorption lines compared to He gas, with contributions from dipole-forbidden transitions, but still with Fano profiles. A very recent study by Bastian et al. 17 investigated the decay of the first doubly excited resonances of He droplets, analyzing in detail contributions of AI, interatomic Coulombic decay (ICD) and secondary processes by inelastic electron scattering. 17,18 Here, we investigate the stimulated emission of He + ions from He bilayers on Pt(111) by electronic double excitations. We find many resonances hitherto not observed in droplets. The arbitrary orientation of the light’sE-vector with respect to the surface normal enables the investigation of polarization effects. For several resonances we find excitation and kinetic ion-energy correlated, providing insights in the desorption mechanism and a better spectral resolution of excitation energies compared to kinetic energy-integrated data. Unfortunately, a rigorous theoretical treatment as for the singly excited states 5 could not be achieved as yet, several trials failed, hence our explanations will be occasionally tentative. We trust that our data will stimulate further efforts in this direction. 2. EXPERIMENT Data have been acquired at the UE112-PGM beamline of the synchrotron radiation (SR) source BESSY-II, Berlin, during single bunch operation of the storage ring. A time-of-flight (TOF) spectrometer enabling the detection of electrons, ions, metastable particles and fluorescence photons was used for data acquisition. 19 Investigations of He layers physisorbed on metal surfaces require low substrate temperatures and shielding of the 300 K blackbody radiation from the environment to prevent rapid infrared-induced desorption of the adsorbate. 20 The home-made liquid 4 He (l-He) bath cryostat of our UHV chamber (base pressure of 5⋅10 −11 mbar) was operated at a He pressure of typically 5⋅10 −2 mbar corresponding to a l-He temperature below 1 K. 21 A radiation shield cooled to 80 K surrounded the cryostat, the crystal mount, and the spectrometer. The SR entered this shield through a small aperture. The orientation of the E-vector was either parallel to the surface (s-orA xy -polarization) or tilted by 10° with respect to the surface normal (mainly p-or A z -polarization; see Refs. 5,22, and 23 for more details of the experimental setup). The lifetime of one l-He filling (250 ml) was at least 3 h. Ru(0001) and Pt(111) crystals were attached to this cryostat by a monocrystalline tungsten rod with very good heat conductivity at low temperature. With this setup, sample temperatures below 1.2 K were routinely obtained. 5,22,23 The samples were heated either by thermal radiation or electron bombardment from a tungsten filament, and temperatures were measured with spot-welded K-type thermocouples. Ru(0001) and Pt(111) samples were cleaned by sputtering with Ne + , repeated heating to 1450 K [1100 K for Pt(111)] in 10 –7 mbar O 2 and final flashing to 1570 K (1300 K) for 60 s. 100 atomic layers thick Ag(111) films were prepared by thermal evaporation of high-purity silver (parts of an Ag monocrystal) onto the Ru(0001) substrate. 22,23 For all samples, crystallographic order was checked by low-energy electron diffraction (LEED), and the cleanliness by photoemission spectroscopy (PES) of core C 1s,O1sand valence levels. He films were prepared by dosing purified He gas (purity better than 6 N) through a capillary aiming at the sample. We compensated the unavoidable loss of adsorbed He due to photon stimulated desorption by a controlled He background pressure of 8⋅10 –10 mbar max., the adjustment of which enabled the variation of the adlayer density within the light spot, particularly the density of the second layer when investigating He bilayers (see below). 3. RESULTS AND INTERPRETATION 3.1. Quantum interference in near-threshold photoemission from He monolayers on metal surfaces Figure 1 shows He 1snear-threshold photoemission spectra from 4 He monolayers adsorbed on Ag(111) [Figs. 1(a) and 1(b)] and Ru(0001) [Fig. 1(c)]. We prepared monolayers by dosing saturated bilayers and desorbing the second layer by heating the sample until no He 1sfrom the second layer was seen in PES, cf. Fig. 1 of Ref. 5. The He 1strace splits periodically in electron momentum, see, e.g., the data for 4 He/Ag(111) in Figs. 1(a) and 1(b). The first splitting occurs at an electron energy of ∼3.8 eV, the second at ∼14.5 eV, i.e. at the 4-fold electron energy or the double electron momentum. We have observed this effect for Cu(111) and Pt(111) as well. The splitting is about 10% larger for 3 He compared to 4 He, pointing toward a zero-point motion effect, while the periodicity in momentum indicates interference. Low Temperature Physics ARTICLE pubs.aip.org/aip/ltp Low Temp. Phys. 50, 713 (2024); doi: 10.1063/10.0028136 50, 714 ©Author(s)2024 30 October 2025 09:20:02 We explain this splitting by the mechanism shown in Fig. 1(d). The A z -light with its E-vector polarized along the surface normal (we neglect the 10° misorientation according to our experimental geometry) excites the 1selectron into a p-wave with p z -symmetry, featuring two lobes, one directed toward the vacuum and the other towards the surface. The ionization is a sudden process, encompassing the spatial interval of the He atom along the surface normal due to its zero-point motion. The binding energy, however, and also the kinetic energy of the photoelectron depend on this position at the instant of ionization due to image charge screening of the photo hole, i.e. the z-distribution of the adsorbate position due to zero-point motion is projected onto the screening function; in a classical picture the energy gain of the photoelectron would be proportional to q 2 /4z (in atomic units) with the elementary charge q,or∼3.6 eV/Å. As shown in Fig. 1(d), this projection results in a total width Wof the photoemission line which is seen for He/Ag(111) at higher energy [Fig. 1(b)] and for Ru(0001) already close to the threshold [Fig. 1(c)]. In addition to this incoherent part, we also resolve a coherent contribution due to interference of direct and reflected wave. To verify this concept, we have calculated a simple single scattering model. We project a ± 0.5 eV region around the center binding energy (within a cos 4 window) onto the screening function to obtain the z-value at the instant of ionization. We then calculate the interference of the direct and reflected lobe, taking the medium bond length from Ref. 3and treating the inner potential and scattering phase as fit parameters to obtain the correct position of the first line splitting on the kinetic energy scale. Figure 2 shows that FIG. 1. He 1sphotoemission by A z -polarized light from a monolayer of 4 He on Ag(111) in (a) the near-edge region and (b) at higher photon energy and (c) from a monolayer of 4 He on Ru(0001) in the near-edge region. (d) Mechanism combining the zero-point distribution along the He-substrate distance z(red), image potential screening (blue), electron reflection from the surface (black), and interference of direct and reflected part of the p z -wave (green). Low Temperature Physics ARTICLE pubs.aip.org/aip/ltp Low Temp. Phys. 50, 713 (2024); doi: 10.1063/10.0028136 50, 715 ©Author(s)2024 30 October 2025 09:20:02 this simple procedure reproduces the splitting and its periodicity in momentum, as well as the different lengths of the splitting regions for the first and the second interference order. We also have performed relativistic photoemission calculations 24 for variable bond lengths assuming again image potential screening as discussed above. This one-step calculation reproduces the splitting of the coherent part like our simple model, apart from a shift of the kinetic energy scale since we did not implement fitting parameters (inset in Fig. 2). We expect that this coherent part will vanish under the incoherent background Wforhigherphotoelectronenergies when the backscattering cross section decreases, particularly for surfaces where loss processes are strong, compare, e.g., the ratios of coherent to incoherent signals obtained for He/Ag(111) at low [Fig. 1(a)] and higher photon energy [Fig. 1(b)]; and for He/Ag(111) and He/Ru(0001) in the near-threshold range [Figs. 1(a) and 1(c)]. We note that we did not include an EXAFS-like increase of the ionization cross section at the positions of the interference maxima, although such an effect is likely (the extended X-ray absorption fine structure approach uses the modulation of the photoionization cross section by the interference of back-scattered and forwardpropagating wave for structural analysis). 25 3.2. Doubly excited He** states in He bilayers on Pt(111) In Fig. 3 we compare spectra obtained at the lower part of the He double excitation range, i.e. below the second ionization threshold N= 2, from isolated He atoms, 26 large droplets, 16 He bilayers adsorbed on Pt(111) measured by us, and theoretical resonance energies of singlet 15 and triplet states 27,28 of He gas. For droplets the electron yield has been recorded, 16 for He gas the total ion yield, 26 and for the bilayer the emission of He + ions. The decay of He** states via AI to He + is evidenced by the Fano line shape of the gas phase spectrum, also visible for the droplets. 16 For He** in the condensed phase, AI produces He + and electrons with energies beyond the first ionization limit of He that can ionize further atoms in the matrix. ICD is an alternative process yielding He + and singly excited H*, most probably as He + –He* ionic excimers, 16,17 and slow decay electrons which cannot produce further ions. 29 We note already here that the emission of ions from the surface after AI or ICD is not selfevident. The ions need kinetic energy to overcome the image charge attraction by the metal substrate and/or attraction due to binding He 2 + states. 30 We will discuss possible desorption mechanisms below. We also measured the photon energy range between 55.6 and 59.5 eV where in non-zero momentum transfer electron-energy loss spectroscopy the first member of the 1Seseries was found (2(1, 0)þ 2in N(K,T)A nnotation) 11 Ref. 31 and the range from 64.5 to 70.0 eV. In neither region maxima beyond the noise floor could be detected. In analogy to the gas phase data (Fig. 3), we label the photon energy range below 62.0 eV photon energy as N, n =2, 2, the range from 62.0 to 63.8 eV as 2, 3, and that from 63.8 to 64.5 eV as 2, 4. FIG. 2. Result of a qualitative model calculation for He 1sline splitting by interference of direct and reflected photoelectron wave. The inset shows the result of the relativistic one-step photoemission calculation. FIG. 3. He double excitations below the second ionization limit recorded from different species and with different probes: Total ion yield from He gas (after Ref. 26), electron yield from He droplets containing 10 11 atoms (after Ref. 16), and He + ions desorbed from He bilayers on Pt(111) with A xy - and A z -polarized light, and for different coverages of the second He layer. Calculated gas phase values for singlet resonances from Ref. 15 are shown as points and for triplet states as stars. 27,28 Low Temperature Physics ARTICLE pubs.aip.org/aip/ltp Low Temp. Phys. 50, 713 (2024); doi: 10.1063/10.0028136 50, 716 ©Author(s)2024 30 October 2025 09:20:02 For He atoms the n+ states (using the nomenclature of Ref. 9) dominate, the n–maxima are about two orders of magnitude smaller and (2p,nd)-final state resonances are even less intense. 26,32 The first resonance in this energy range around 60 eV is the strongest, also for the He droplets, but here it is much broader and blueshifted with respect to the gas phase. The shoulder at the leading edge has been explained by the dipole-forbidden (2p2)1Deresonance that was also observed for the gas phase by an analysis of the asymmetry parameter. 33 The bilayer spectra resemble the droplet data for the N, n =2, 2 range. For both polarizations and the entire coverage range two maxima exist, the second one broader than the first one (we will present more details below when correlating excitation and kinetic ion energies). At photon energies beyond 62 eV, the bilayer spectra show a rather complex structure of overlapping maxima that changes with the coverage of the second layer (due to the continuous variation of our results with coverage, we assume a 2D-like gas phase for the second layer). Some maxima vanish with decreasing coverage, and all change their shapes. To obtain more detailed information we have recorded not only the yield but also the kinetic energy E kin of the desorbing He + ions, see Figs. 4 and 5. Due to the low kinetic energies of the He + ions we had to accelerate the ions by 70 eV to avoid ambiguities by mixing signals from different light pulses. In combination with the small geometry uncertainties of our TOF device, the discrimination of the exact origin of the kinetic-energy scale was not trivial. We have set the scale’s zero to the cut-off of the ions’kinetic-energy distribution that was constant for the entire photon energy range. We note, however, that we cannot exclude an error of the scale’s zero point by ± 50 meV at most; the slope error however is small, well below 1%. Due to the geometry of our TOF spectrometer, only the component of the kinetic energy along the surface normal of the sample is recorded. In the N, n = 2, 2 range two maxima are seen in Fig. 4 with both light polarizations, for the saturated bilayer the first at E ph ; E kin —values of 59.95;0.2 eV with A z - polarization and 59.95;0.28 eV with A xy -polarization, the second at 60.25;0.2 eV in A z and 60.4;0.28 eV in A xy . With A xy -light both maxima are of equal height, with A z the second is larger than the first. The first is broad, the second very broad on the E ph scale. No correlation between the excitation and the kinetic ion-energy exists, see the compilation in Table I. In the N, n = 2, 3 region, the spectrum at saturation is much more detailed. For A z five maxima at 62.60;0.18, 62.90;0.25, 63.10;0.39, 63.35;0.29, and 63.68;0.22 eV can be discerned, with the largest at 63.35 eV. E ph and E kin are correlated for the peaks at 62.90, 63.10, and 63.35 eV with dE ph /dE kin slopes of 0.7, 0.5, and 0.2, respectively. For A xy , 4 maxima exist at 62.90;0.27 eV (slope 0.4), 63.10;0.42 eV (slope 0.1), 63.40;0.34 eV (slope 0.4), and 63.60;0.34 eV. The last one appears as a shoulder at the trailing edge of the peak at 63.40 eV. In the N, n = 2, 4 range two maxima are found with A z ,as well as A xy -light, with A z at 63.95;0.22 and 64.10;0.4 eV (merely a shoulder of the maximum at 63.95 eV), and with A xy at 64.10;0.33 and 64.15;0.37 eV. At medium coverage (near 1/2 of saturation) only minor changes compared to full coverage are observed for the 2, 2 region. We find small variations in excitation and kinetic ion-energy (Table I) and, more pronounced, changes of the relative intensities of the maxima: The 59.95 to 60.30 eV peak ratio increases for A xy and the 2, 2 maxima decrease with respect to those of the 2, 3 and 2, 4 regions for A z . More significant changes are observed in the 2, 3 and 2, 4 regions. With A xy -light the peak at 63.10 eV loses intensity as those at 62.65 and 63.05 eV with A z . The spectral resolution gets better: At A z the maxima at 64.00 and 64.20 eV start to separate, and at A xy the maxima at 64.00 and 64.15 eV. At the lowest coverage of the second He layer (∼10% of saturation), the 2, 4 maxima separate even more. For A z three isolated peaks are seen at 63.99;0.24, 64.11;0.31, and 64.18;0.22 eV (see details in Fig. 5), and for A xy two maxima at 63.95;0.22 and 64.15;0.33 eV (Fig. 4, lowest panel). In the 2, 3 range, the maxima at 63.32;0.24 and 63.5;0.38 eV are now well resolved. Other resonances have lost intensity: The peak at 63.10 eV has vanished in A z and A xy , and the maximum at 62.65 eV in A z . The signal at 62.85;0.20 eV in A xy has become very weak. Obviously these resonances are dipole-forbidden for isolated atoms in the second layer and gain oscillator strength by lateral interaction with neighbors. The significant differences with respect to the gas phase indicate that other dipole-forbidden resonances survive due to horizontal coupling with the first He layer and the metal. If we compare the largest maxima to the strong n+ transitions of the gas phase, i.e. the peaks at 60.30 eV for 2, 2, 63.40 eV for 2, 3, and 64.10 eV for 2, 4, we find a blue-shift in the red, and a redshift in the blue end of the series, compatible with repulsive neighbor interaction and screening of the ionic series limit by the metal substrate. Because of the striking similarity of the droplet and the bilayer data we assign the two 2, 2 maxima as in Ref. 16 to the dipole-forbidden 1 D e -state with mainly (2p 2 )-configuration 15 (here 59.95 eV in A z and A xy ), and to the dipole-allowed 2 + 1 P o resonances (here, 60.25 eV in A z and 60.40 eV in A xy ), see marks in Fig. 3. We note, however, that such a direct correspondence of resonance energies calculated for the gas phase, i.e. a center-symmetric system with strong electron correlation, and the bilayer situation where polarization and coverage dependence indicate strongly perturbed electron correlation, cannot be expected, at least not for higher quantum numbers with extended orbitals stretching over more than one cell of nearest neighbors. For the 2, 2 region, this perturbation may be small, explaining the correspondence of gasphase (including the findings of Ref. 33), droplets 16 and bilayer results. Resorting to the (Z+ 1) equivalent-core analog Li for He* and the (Z+ 2) analog Ba for He** and their van der Waals radii of 1.82 Å 34 and 1.53 Å, 35 we conclude that He** will be not “larger” than He*, which is mostly contained within the lattice cell around the excited atoms, however with strong polarization towards the vacuum; 5 orbital plots in Ref. 15 show similar sizes. For the 2, 3 and 2, 4 regions the He** wavefunctions will be much larger, being strongly polarized towards the vacuum and influenced by the metallic substrate. A similar correspondence of resonances as for the 2, 2 range cannot be expected. One could try to assign the group of three peaks between 62.65 eV (A z ) and 63.05 eV to the 31 P o maximum with the two 1 S e resonances at both sides, and the 3 closely spaced 2, 4 maxima to the central 4 + 1 P o peak with either 1 S e or 1 D e neighboring states (Fig. 3), but such assignments would Low Temperature Physics ARTICLE pubs.aip.org/aip/ltp Low Temp. Phys. 50, 713 (2024); doi: 10.1063/10.0028136 50, 717 ©Author(s)2024 30 October 2025 09:20:02 FIG. 4. Kinetic energy distributions of He + ions desorbed from bilayers of He on Pt(111) by A z -polarized (right) and A xy -polarized light (left), for saturated, medium and low coverage of the second layer (from top to bottom). Low Temperature Physics ARTICLE pubs.aip.org/aip/ltp Low Temp. Phys. 50, 713 (2024); doi: 10.1063/10.0028136 50, 718 ©Author(s)2024 30 October 2025 09:20:02 be arbitrary without calculations. Admittedly, these will be demanding because electron correlation, interaction with the matrix and the metal substrate, and zero-point motion have to be included; standard DFT code will not suffice. We finally focus on the correlation of excitation and kinetic ion-energy that is seen for some of the resonances. It was a surprise for us that this correlation is seen only for some of these maxima and not for all. When starting these experiment, we thought that in the first step He atoms of the second layer would be excited, with increased excitation energy for reduced He–He distance due to zero-point motion (compare, e.g., the potential curves for He–He** in Ref. 17). Next, we expected desorption of He** due to the antibonding nature of these potentials, similar to the cavity ejection mechanism known from condensates of neon and argon, 36 and subsequent AI yielding He + . Here, we also had expected larger kinetic energies for smaller He–He distance because of stronger Pauli repulsion. We do see a correlation for some resonances, definitely for those which obviously are dipole-forbidden without perturbation by neighbors; in He gas, some of them have long lifetimes. 15 We speculated that for resonances which are strongly broadened by zero-point motion (mainly those from the 2, 2 region), desorption and excitation decay occur on similar time scales. Closely spaced atom distances would support fast desorption as well as fast ICD 37 and by perturbation possibly also faster AI, and these effects could compensate, leading to the result we see. For delayed ionization, however, the correlation would persist. We hope that future calculations will resolve this puzzle. Desorption scenarios different from the cavity ejection process must also be considered, e.g., ICD into antibonding He*–He + states 16,17 and the formation of He + –He + pairs by impact of fast decay electrons from AI inside the layer and subsequent Coulomb explosion, 18,38 a process that would yield ∼2 eV kinetic energy for each atom if nearest neighbors would be concerned. 18 In Ref. 17,the branching into ICD final states was determined for He droplets by theory and experiment to values in the low percentage range, so we expect this reaction to be a minority channel. In addition, ICD would ionize He atoms of the matrix, i.e. within the second or the first layer. Desorption of He + from the first layer on the metal is unlikely due to image potential attraction and fast neutralization. Even from the second layer, one would expect repulsion of He + rather in a lateral FIG. 5. Expanded high photon energy part of the kinetic energy distribution of He + ions desorbed by A z -light from a He bilayer on Pt(111) with low coverage of the second layer. Low Temperature Physics ARTICLE pubs.aip.org/aip/ltp Low Temp. Phys. 50, 713 (2024); doi: 10.1063/10.0028136 50, 719 ©Author(s)2024 30 October 2025 09:20:02 than normal direction, i.e. not compatible with our detection geometry. Desorption of He* by cavity ejection could happen but would not create He + , i.e., would not show up in our results. We also note that for all resonances no bimodal distributions of the kinetic ion-energy have been observed which would be the case if two processes would contribute equally. 39 Moreover, the distributions of kinetic energies show maxima in a rather limited range for all resonances, from 0.18 to 0.42 eV. We therefore think that a single desorption process dominates, most probably He** repulsion by the closed-shell matrix with subsequent AI. Coulomb explosion may cause the weak tails at high kinetic ion-energy (Figs. 4 and 5). In summary, we have demonstrated quantum effects for two rather different types of electronic processes in adsorbed He on metal surfaces. For adsorbed He atoms, the uncertainty of their distance from the substrate due to zero-point motion in combination with image potential screening, the sudden nature of the photoionization process, and electron scattering by the substrate causes interference of the photoelectron’s final state wavefunction, resulting in a periodic doubling of the apparent binding energy value. A simple scattering calculation can reproduce this effect, confirmed by state of the art one-step photoemission theory. He + desorption by double excitation of He in bilayers on a metal surface shows a variety of effects due to interaction with neighbors and zero-point motion, namely dependence the polarization of the light, line broadening, particular for excitations at the red end of the spectrum, and a multitude of resonances that were neither observed for isolated He nor for He droplets, pointing towards strong perturbation of the electron correlation prototypical for isolated He**. For some resonances, excitation energy and kinetic ion-energy are correlated, indicating coupling of excitation, expulsion and ionization processes. Compared to our first example, a detailed analysis by theory was not possible as yet. We hope that the results presented here will stimulate future efforts in this direction. ACKNOWLEDGMENTS We thank R. Schneider, N. Armbrust, and the staff of BESSY for help during the experiments. SJR, SK, JVB, and PF acknowledge support from the Deutsche Forschungsgemeinschaft (Project Fe 286/2-1, and the clusters of excellence Munich-Centre for Advanced Photonics, and e-conversion) and by the Helmholtz– Zentrum–Berlin. JM would like to thank the project Quantum materials for applications in sustainable technologies (QM4ST), funded as project No. CZ.02.01.01/00/22_008/0004572 by Programme Johannes Amos Commenius, call Excellent Research. Author Contributions S. J. Riepl and S. Kossler contributed equally to this work. REFERENCES 1 M. A. Adams, J. Mayers, O. Kirichek, and R. B. E. Down, Phys. Rev. Lett. 98, 085301 (2007). 2 J. Beamish and S. Balibar, Rev. Mod. Phys. 92, 045002 (2020). 3 E. Zaremba and W. Kohn, Phys. Rev. B 15, 1769 (1977). 4 W. T. Sommer, Phys. Rev. Lett. 12, 271 (1964). 5 S. Kossler, P. Feulner, and J.-P. Gauyacq, Phys. Rev. B 89, 165410 (2014). 6 M. Joppien, R. Karnbach, and T. Möller, Phys. Rev. Lett. 71, 2654 (1993). 7 W. Moritz, Physics of Solid Surfaces: Subvolume A (Springer, Berlin, 2015). 8 G. Kaindl, T.-C. Chiang, D. E. Eastman, and F. J. Himpsel, Phys. Rev. Lett. 45, 1808 (1980). 9 R. P. Madden and K. Codling, Phys. Rev. Lett. 10, 516 (1963). 10 J. W. Cooper, U. Fano, and F. Prats, Phys. Rev. Lett. 10, 518 (1963). 11 G. Tanner, K. Richter, and J. M. Rost, Rev. Mod. Phys. 72, 497 (2000). 12 R. Püttner, B. Grémaud, D. Delande, M. Domke, M. Martins, A. S. Schlachter, and G. Kaindl, Phys. Rev. Lett. 86, 3747 (2001); Y. H. Jiang, R. Püttner, D. Delande, M. Martins, and G. Kaindl, Phys. Rev. A 78, 021401 (2008); Y. H. Jiang, PhD-thesis, Freie Universität Berlin (2006). 13 A. S. Schlachter, Radiation Phys. Chem. 75, 2159 (2006). 14 V. Gruson, L. Barreau, A. Jiménez-Galan et al., Science 354, 734 (2016); A. Kaldun, A. Blättermann, V. Stooß et al., Science 354, 738 (2016). TABLE I. He + emission from He bilayers on Pt(111) by He** excitations: Photonand kinetic ion-energies at the maxima of the distributions, and correlations of excitation and ion energy, extracted from the data of Figs. 4 and 5. For resonances without a clear correlation of E ph and E kin no numbers are given. Values for not wellseparated shoulders of other maxima are put in parentheses. Range N, n E ph , eV E kin , eV dE ph / dE kin E ph , eV E kin , eV dE ph / dE kin A z , saturated 2nd layer A xy , saturated 2nd layer 2, 2 59.95 0.20 59.95 0.28 60.25 0.20 60.40 0.28 2, 3 62.60 0.18 –– – 62.90 0.25 0.7 62.90 0.27 0.4 63.10 0.39 0.5 63.10 0.42 0.1 63.35 0.29 0.2 63.40 0.34 0.4 63.68 0.22 (63.60) (0.34) 2, 4 63.95 0.22 64.10 0.33 (64.10) (0.40) 64.15 0.37 A z , medium coverage of 2nd layer A xy , medium coverage of 2nd layer 2, 2 59.95 0.29 59.95 0.32 60.18 0.29 60.30 0.32 2, 3 62.65 0.25 –– – 62.90 0.28 0.7 62.90 0.27 0.4 63.05 0.39 0.5 63.10 0.42 0.1 63.35 0.29 0.4 63.35 0.33 0.4 63.70 0.24 (63.55) (0.33) 2, 4 64.00 0.28 64.00 0.25 (64.20) (0.30) 64.15 0.38 A z , low coverage of 2nd layer A xy , low coverage of 2nd layer 2, 2 59.90 0.28 59.90 0.28 60.15 0.27 60.20 0.29 2, 3 –– 62.97 0.31 1 62.85 0.20 0.4 –– 63.33 0.27 0.2 63.32 0.24 0.3 63.61 0.23 63.50 0.38 2, 4 63.99 0.24 63.95 0.22 64.11 0.31 64.18 0.22 64.15 0.33 Low Temperature Physics ARTICLE pubs.aip.org/aip/ltp Low Temp. Phys. 50, 713 (2024); doi: 10.1063/10.0028136 50, 720 ©Author(s)2024 30 October 2025 09:20:02