Charge-exchange dominates long-range interatomic Coulombic decay of excited metal-doped He nanodroplets
L. Ben Ltaief, M. Shcherbinin, S. Mandal, S. R. Krishnan, A. C. LaForge, R. Richter, S. Turchini, N. Zema, T. Pfeifer, E. Fasshauer, N. Sisourat, M. Mudrich
CCharge-exchange dominates long-rangeinteratomic Coulombic decay of excitedmetal-doped He nanodroplets
L. Ben Ltaief, † M. Shcherbinin, † S. Mandal, ‡ S. R. Krishnan, ¶ A. C. LaForge, § R.Richter, (cid:107)
S. Turchini, ⊥ N. Zema, ⊥ T. Pfeifer,
E. Fasshauer, † N. Sisourat, @ andM. Mudrich ∗ , † † Department of Physics and Astronomy, Aarhus University, 8000 Aarhus C, Denmark ‡ Indian Institute of Science Education and Research, Pune 411008, India ¶ Department of Physics, Indian Institute of Technology, Madras, Chennai 600 036, India § Department of Physics, University of Connecticut, Storrs, Connecticut, 06269, USA (cid:107)
Elettra-Sincrotrone Trieste, 34149 Basovizza, Trieste, Italy ⊥ Istituto Struttura della Materia-CNR (ISM-CNR), 00133 Roma, Italy
Max-Planck-Institut für Kernphysik, 69117 Heidelberg, Germany @ Sorbonne Université, CNRS, Laboratoire de Chimie Physique Matière et Rayonnement,UMR 7614, F-75005 Paris, France
E-mail: [email protected]
Abstract
Atoms and molecules attached to rare gas clusters are ionized by an interatomicautoionization process traditionally termed ‘Penning ionization’ when the host clus-ter is resonantly excited. Here we analyze this process in the light of the interatomicCoulombic decay (ICD) mechanism, which usually contains a contribution from charge a r X i v : . [ phy s i c s . a t m - c l u s ] O c t xchange at short interatomic distance, and one from virtual photon transfer at largeinteratomic distance. For helium (He) nanodroplets doped with alkali metal atoms(Li, Rb), we show that long-range and short-range contributions to the interatomic au-toionization can be clearly distinguished by detecting electrons and ions in coincidence.Surprisingly, ab initio calculations show that even for alkali metal atoms floating indimples at large distance from the nanodroplet surface, autoionization is largely dom-inated by charge exchange ICD. Furthermore, the measured electron spectra manifestultrafast internal relaxation of the droplet into mainly the 1s2s S state and partiallyinto the metastable 1s2s S state.
Interatomic decay processes have recently been found to play a crucial role in the in-teraction of biological matter with energetic radiation. Both free radicals and low-energyelectrons produced by ICD processes can induce irreparable damage of the genome (doublestrand breaks in DNA) causing cancer or cell death.
Upon electronic excitation, weaklybound systems such as van der Waals or hydrogen bonded complexes and clusters can relaxby interatomic autoionization if the excited state energy exceeds their adiabatic ionizationenergy. In the case of rare gas clusters doped with atomic or molecular impurities, thisprocess has traditionally been termed Penning ionization, in analogy to the collisionalautoionization occurring in crossed atomic beams involving excited atoms, mostly rare gasesprepared in metastable excited states. This process is mainly driven by charge exchangebetween two interacting atoms or molecules which come so close to one another that theirvalence orbitals overlap. However, already in the early days of systematic Penning ionizationstudies, it was realized that the autoionization rate contains a second contribution describingenergy transfer in the form of a virtual photon exchange. Since the seminal work by L. Cederbaum in 1997, such non-local autoionization processesinvolving two or more atomic or molecular centers have been formulated in the theoreti-cal framework termed interatomic/intermolecular Coulombic decay (ICD). This approachmainly refers to the autoionization of weakly bound systems that are inner-shell excited by2nergetic photons or electrons, rather than to thermal collisions of metastable atoms. In suchinner-shell excited complexes, the virtual photon transfer mechanism or direct ICD processmediated by energy transfer often dominates over the decay by charge exchange.
DirectICD relies on the excited state being coupled to the ground state by an electric dipole-allowedtransition and its decay rate scales as ∝ R − with the interparticle distance R , in contrastto the exponential scaling ∝ exp ( − R/a ) of the charge exchange ICD term, where parameter a depends on the spatial extension of the involved orbitals. Therefore, the ICD by en-ergy transfer occurs at rather large interatomic distances. However, charge exchange cansignificantly contribute to the ICD rate of excited dimers, especially at short interatomicdistances where the valence electron orbitals overlap thereby enhancing the decay rate byorders of magnitude.
Thus, the autoionization of electronically excited clusters due tointeratomic electronic couplings, traditionally named Penning ionization, can be regardedas one member of the family of ICD processes, and we argue that both terms are equallyappropriate in the case presented here.In experiments, the contributions to the non-local autoionization rate by charge andenergy exchange are usually indistinguishable, since the final products (electrons and ions)are the same. However, T. Jahnke and co-workers have recently shown experimentally thatthe two contributions can actually be disentangled for the cases of Ne and HeNe dimersby inferring the interatomic distance where ICD takes place. In the current work, we presenta method to discern the long and short-range contributions to ICD for a very different system– helium nanodroplets doped with alkali metal atoms. This method relies on the property ofclusters and nanodroplets to solvate ions, provided they are formed with low kinetic energyto prevent their detaching from the cluster. We find that ICD occurring at large dopant-helium distances ( R (cid:38) Å) strongly dominates over ICD at short range ( R (cid:46) Å), where theatomic energy levels are measurably shifted. Nevertheless, ab initio calculations show thatICD proceeds nearly exclusively by charge exchange for all experimentally relevant distances.Furthermore, detailed insights into the internal relaxation of the excited He nanodroplets3rior to ICD are gained.
01 02 03 04 05 0 1 2 1 3 1 4 1 5 1 6 1 7 1 80 . 00 . 51 . 0 05 01 0 01 5 02 0 02 5 0 1 3 1 4 1 5 1 6 1 7 1 8 1 90 . 00 . 51 . 0 S S P a ) L i d o p i n gT o t a l e l e c t r o n s h n S S P h n b ) E l e c t r o n s i n c o i n c i d e n c ew i t h L i + Electron counts [arb. u.]
E l e c t r o n k i n e t i c e n e r g y [ e V ]
P I E S c ) R b d o p i n gT o t a l e l e c t r o n s d ) E l e c t r o n s i n c o i n c i d e n c ew i t h
R b + E l e c t r o n k i n e t i c e n e r g y [ e V ]
C o n v o l . P I E S
C o n v o l . P I E S C o n v o l . s h i f t e d P I E S
Figure 1: Electron spectra recorded for He nanodroplets doped with Li atoms [a) and b)]and with Rb atoms [c) and d)] at a photon energy hν = 21 . eV. a) and c): Spectra of allelectrons emitted from the doped He nanodroplet. b) and d): Spectra recorded in coincidencewith Li + and Rb + dopant ions. The spectra in a) and b) as well as those in c) and d) arenormalized to the same scaling factors to preserve their relative intensities, respectively. Thered lines in b) and d) show Penning ionization electron spectra (PIES) measured in crossedatomic beams. The light blue lines show these PIES for adjusted relative contributions of S and S components convoluted with the instrument function of the spectrometer (andslightly shifted in the case of Rb). The dashed vertical lines indicate expected electronenergies based on the atomic term values and ionization potentials.To perform these experiments, a beam of He nanodroplets containing on average ¯ N He =2 . × He atoms per droplet is doped with alkali metal atoms M (Li, Rb) and irradiatedby extreme-ultraviolet (EUV) light generated by a synchrotron light source. The EUV lightis tuned to the strongest absorption band of He nanodroplets correlating to the 1s2p P stateof He at a photon energy hν = 21 . eV and to higher-lying bands. Subsequently, the dopeddroplets autoionize in the reaction (He) ∗ N + M → He N + M + + e − . The electrons e − andions M + are detected in coincidence using a velocity-map imaging (VMI) spectrometer tomeasure the electron kinetic energy K e . hν = 21 . eV. These spectrarecorded for different average droplet sizes ( ¯ N = 2500 and × ) are essentially identicalto those in Fig. 1. The top panels, a) and c), show the spectra recorded when measuringall electrons emitted from the droplets, while panels b) and d) show the same measurementstaken in coincidence with Li + and Rb + , respectively. The vertical dashed lines illustratethe electron energies K e = E He − E i evaluated for the level energies E He of He atoms in theexcited states 1s2s S, 1s2s S, and 1s2p P assuming relaxation of the droplet excited states toatomic levels. The ionization energies of Li and Rb dopants are taken as the atomic values E i = 5 . eV and . eV, respectively. The vertical line marked by hν depicts the electronenergy K e = hν − E i one would measure if ICD occurred directly from the resonantly excitedstate of the He droplet. However, we find that the excited He droplets initially relax intolower-lying levels near those of free He atoms prior to the ICD. The relaxation dynamicswas recently directly mapped using EUV-UV pump-probe and a relaxation time of about1 ps was found. Electrons emitted by direct photoionization of the alkali atoms in theresidual gas or on He nanodroplets, which would add to the signal at K e = hν − E i , do notcontribute due to the low absorption cross section of the atoms.One main finding here is that ICD predominantly occurs out of the 1s2s S state as seenfrom the sharp peaks in the total electrons spectra shown in the top panels a) and c) of Fig. 1.The finite widths of the peaks are due to the limited resolution of our VMI spectrometer. Thisfinding is in agreement with the time-resolved relaxation measurements which showed thatmost of the He droplets initially excited into the 1s2p P band end up as nearly unperturbedHe atoms in the 1s2s S state residing in void bubbles inside the droplets or as free 1s2s Satoms ejected out of the droplets. The fact that the peaks are nearly unshifted with respectto the atomic 1s2s S level indicates that the dopant-droplet autoionization occurs at largeinteratomic distance R (cid:38) Å where the energy levels are only weakly perturbed. In contrast,in those events that generate both an electron and a free dopant ion [bottom panels b) and5)], the electrons are emitted at slightly lower kinetic energies. Note that we measure 30and 200 times more electrons in total as compared to the number of electron-ion coincidenceevents involving Li + and Rb + ions, respectively. Assuming a detection efficiency for ions of30 %, this results in ratios of total emitted electrons to coincident electrons of 9 and 60 forLi + and Rb + ions, respectively. This ratio is larger for Rb than for Li because the heavyRb ion is less likely to detach from the droplet surface compared to the light Li ion. Thiswas seen in similar experiments for alkali atoms and dimers attached to He nanodroplets,where the proportion of dopant ions ejected by ICD was lower for the heavier species whencomparing to the ion signal due to double ICD of the dimers. Detecting those ions that arefully solvated inside the He nanodroplets would require the ion spectrometer to be placed inline with the He droplet beam and the detector to be sensitive for large ion masses, as e. g. in Ref. The double-peak structure present in all coincidence electron spectra [Figs. 1 b) and d)]resembles the Penning ionization electron spectrum (PIES) measured by M. W. Ruf andco-workers using crossed atomic beams, shown as red lines. This PIES is characteristic fortraditional Penning ionization occurring in collisions of He atoms prepared in both 1s2s S and S metastable states with groundstate atoms of a different species at thermal beam velocities.In such collisions, the interaction time of the colliding atoms is extremely short ( (cid:28) ps)such that the autoionization is entirely dominated by the charge exchange mechanism whichis active at short interatomic distance. The light blue lines show the results of convolving the PIES with the instrument functionof the VMI spectrometer. In the Rb case, the 1s2s S and 1s2s S peaks are shifted by − . and − . eV, respectively, in order to match the experimental and modeled spectra.A Gaussian function centered around eV is added to account for an additional signalcomponent due to double ICD. The electron energy shifts are likely due to elastic scatteringof the emitted electron with the surrounding He atoms. We mention that the Penningelectron spectra for molecules embedded in the droplet interior are even more strongly shifted6nd broadened. In addition, the amplitudes of the S component was scaled by factors 0.38(Li) and 0.24 (Rb) relative to the ones of the S component to fit the experimental spectra.
Potential energy [eV]
I n t e r a t o m i c d i s t a n c e R [ Å ] H e * ( S ) + L i H e * ( S ) + L i H e + L i + + e - S h o r t - r a n g e I C D L o n g - r a n g e I C D
Figure 2: Potential energy curves involved in the ICD process of Li and excited He ∗ , takenfrom. ICD taking place at large interatomic distance R (cid:38) Å is represented by the pinkvertical arrow. ICD occurring after the contraction of the He ∗ Li dimer to short distance R = R , the electron is emitted with an energy given by the atomic 1s2s S level. While for thefree He atom the transition from this state to the 1s
S ground state is dipole forbidden,this transition becomes partly allowed for the He ∗ Li dimer due to the breaking of the atomicsymmetry (in the molecular symmetry the Σ → Σ -transition is allowed).When the He and Li atoms move towards each other along the attractive potential energycurves prior to decaying, the potential energy difference taken up by the electron as kineticenergy is reduced. At short distance R (cid:46) Å, the non-local autoionization process is dom-inated by charge exchange ICD. Therefore, it can occur both for the partly allowed 1s2s Sstate and for the 1s2s S state which remains metastable with respect to optical de-excitation7ue to the spin selection rule. Surprisingly, the double peak structure of the coincidenceelectron spectra (Fig. 1) shows that a considerable fraction of excited He droplets relax intothe 1s2s S state which requires a change of the electron spin multiplicity. Previously, fluores-cence measurements had indicated that triplet states would only be populated by electron-ionrecombination following He droplet autoionization occurring at hν > eV. Thus, we findthat droplet relaxation pertains not only to the atomic motion (bubble formation) and tothe orbital angular momentum state of the excited-state electron (interband relaxation ),but even to the spin degree of freedom.From the large ratio of unshifted electrons created by ICD at long range to those shiftedto lower energies, we conclude that in most cases, ICD proceeds before the atoms havenotably moved towards each other. But why is the coincidence detection of electrons anddopant ions selective to the short-range process, whereas the majority of electrons are createdat long range? This is related to the peculiar property of He nanodroplets that excitedelectrons are repelled from their local He environment thereby leading to the formation ofa void bubble, whereas ions tend to be attracted towards the He and form strongly boundsnowball complexes with a dense shell of He atoms around the ion. As a result, an alkaliatom ionized at the droplet surface is likely to sink into the droplet and thus eludes itsdetection.
Counter-intuitively, this happens when the atom is ionized at large distance(pink dashed arrow in Fig. 2), because in that case the ion is created at the droplet surfacenearly at rest. In contrast, when ionization occurs near the minimum of the potential energywell (blue dashed arrows in Fig. 2) where the dopant atom has acquired a maximum ofkinetic energy, the ion bounces back off the He and escapes from the droplet as the atom’skinetic energy is conserved in the ICD. This implies that the detected ratio of S versus S atoms is probably larger than the actual ratio of populations in these states. Since theHe ∗ Li potential energy curve has a deeper well in the S state than in the S state, thekinetic energy and therefore the escape probability of the ions is enhanced for the S state.This is likely the reason why the S contribution is seen in the coincidence spectra but not8n the total electron spectra. This picture is supported by the dopant ion kinetic energydistributions measured in ion VMI mode, see the Supplemental Material. (cid:1) [ R ] × dN/dR [meV/Å] dN/dR [Å-1] R [ Å ] R [ Å ] S S Figure 3: Calculated average ICD rates for the He ∗ +Li reaction involving the two lowestexcited states of He ∗ , 1s2s , Σ , weighted by the number of He atoms present at distance R from the Li atom, N ( R ) , in a He droplet of radius 63 Å, see inset.The clear distinction between ICD occurring at long and short ranges from the photo-electron spectra might lead one to conclude that in this way, the contributions from virtualphoton and charge exchange ICD can be separated experimentally, as it was demonstratedfor the neon dimer. To assess this conjecture, we have carried out ab initio calculationsof the ICD rates Γ for the 1s2s S and 1s2s S states using the Fano-CI-Stieltjes method. Computational details will be given in a future publication. The results show that this isnot the case for the He ∗ Li dimer. In this system, charge exchange ICD dominates over directICD for interatomic distances ranging at least up to 12 Å. In this range, Γ follows the ex-pected exponential scaling with R . Only for R > Å can Γ be extrapolated by the powerlaw ∝ R − for the S state, as direct ICD by virtual photon transfer is possible via electricquadrupole transition. Γ S , S D = Γ S , S × dN/dR weighted by the numberof He atoms present in a He droplet at distance R from the Li atom, dN ( R ) /dR , see theinset. For a spherical shape of the He nanodroplet, dN ( R ) /dR is given by dN ( R ) dR = 2 πRV D (cid:18) R D − R R Li + R − R Li (cid:19) such that (cid:82) dN ( R ) /dR dR = 1 , i. e. one He ∗ is excited at an arbitrary position in thedroplet. We assume that the He droplet has radius R D = 63 Å, and the Li atom is atdistance R Li = R D + d Li − d d from the droplet center, where d Li = 5 . Å is the distance ofthe Li atom from the droplet surface, and d d = 2 . Å is the depth of the dimple as givenin Ref. The volume of the He droplet is V D = 4 πR D / . For simplicity, we assume thatthe He density distribution has a sharp edge at distance d Li from the Li atom. For thisgeometry, the fraction of the total rate of exchange ICD versus the total rate of direct ICD(integral over Γ S D for R > Å) is × . Thus, for all realistic configurations involving aLi atom attached to a He cluster or nanodroplet, ICD proceeds nearly exclusively by chargeexchange despite the rather large interatomic distances. Most likely, this conclusion holdsfor all other alkali atoms attached to He droplets due to the similar structure of their valencestate orbitals. Taking into account that in some cases ICD occurs at shorter He ∗ -Li distancethan d Li due to atomic motion preceding the ICD, the overall proportion of exchange ICDis even further enhanced.When neglecting atomic motion, the droplet-weighted average ICD rate results in charac-teristic times t S D = 1 / Γ S D and t S D = 1 / Γ S D which fall in the µs range, given the considerablecontribution of large values of R to the weighted average. Thus, in the real droplet system,only He atoms in the first layer of the dimple next to the Li atom at R ∼ Å effectivelyundergo ICD with decay times t S = 900 ps and t S = 140 ps. Already for those He atomsexcited in the second layer of the dimple at R ∼ Å, the ICD times given by the decay rate Γ S , S are much longer, t S , S > ns, implying that massive changes of the local geom-10try take place prior to ICD, e. g. due to the formation of a bubble around the He ∗ , themigration of He ∗ to the droplet surface, He ∗ dimer formation, and other effects associatedwith atomic motion. b )a ) P2 s S2 s S Total electrons [arb. u.]Electrons in coinc. with Li+ [arb. u.]
E l e c t r o n k i n e t i c e n e r g y [ e V ]
Figure 4: Spectra in logarithmic scale of all electrons a) and of electrons recorded in co-incidence with Li + ions b) for doped He nanodroplets resonantly excited at various photonenergies.Finally, we address the question how the ICD signals depend on the excited states ofthe He nanodroplet. Fig. 4 shows electron spectra recorded for Li-doped He nanodropletsat various photon energies hν when all emitted electrons are measured (a), or only electron-Li + ion coincidences are detected (b). At hν = 21 . eV (blue lines), the He droplets areexcited directly to the 1s2s S droplet state, whereas at hν = 21 . and . eV (black andred lines), the 1s2p P droplet state is excited. At hν = 23 . and . eV (brown andpink lines), higher states correlating to the 1s3p and 1s4p levels are reached. Most notably,11t all values of hν , the electron spectra are dominated by ICD involving the 1s2s states.Only for hν > eV, a contribution of the 1s3s, p and 1s4s, p states to the ICD signal ofabout % of the corresponding signal from the 1s2s states is observed [Fig. 4 b)]. Thisindicates that internal droplet relaxation is much faster than ICD, i. e. the 1s2s state ispopulated in much less than t S = 900 ps. This is in agreement with relaxation times (cid:46) psmeasured directly using pure He nanodroplets. To directly map the dynamics of the variousICD channels, pump-probe experiments using tunable ultrashort EUV laser pulses would behighly desirable.In summary, we have shown that for He nanodroplets doped with alkali metal atoms(Li and Rb), interatomic autoionization induced by resonant excitation of the He dropletis predominantly driven by charge exchange ICD (equivalently termed Penning ionization),even though it proceeds at large interatomic distances given by the initial configuration ofthe doped He nanodroplet. This is due to the diffuse structure of the electron orbitals ofboth the alkali metal atom and the excited He ∗ atom. This case drastically differs from mostsystems where ICD was studied previously, e. g. by inner-shell excitation or ionization ofrare gas dimers and clusters. It is likely that charge exchange ICD is also the dominantautoionization mechanism in other systems involving valence-excited rare gas atoms, such asmost Penning reactions as well as the recently studied autoionization of multiply excitedrare gas clusters. Furthermore, we find that nearly irrespective of the initial level ofexcitation of the He nanodroplet, autoionization occurs out of the 1s2s-correlated He levelsdue to ultrafast droplet relaxation. While direct ICD mostly involves the droplet perturbed1s2s S state, charge exchange ICD occurs out of the 1s2s S and the 1s2s S states in closeanalogy to traditional Penning ionization of colliding atoms at thermal velocities. Besidesgiving insight into fundamental interatomic decay processes, the method of measuring co-incidence electron and ion spectra for surface-bound atoms may be useful for probing therelaxation dynamics of other types of clusters and nanoparticles as well.12 xperimental methods
The experiments are performed using a He nanodroplet apparatus combined with a velocity-map imaging photoelectron-photoion coincidence (VMI-PEPICO) detector at the GasPhaseand CiPo beamlines of Elettra-Sincrotrone Trieste, Italy. The apparatus has been describedin detail elsewhere.
Briefly, a beam of He nanodroplets is produced by continuouslyexpanding pressurized He (50 bar) of high purity out of a cold nozzle (14 K) with a diameterof 5 µ m into vacuum, resulting in a mean droplet size of ¯ N He = 2 . × He atoms per droplet.Further downstream, the beam passes a mechanical beam chopper used for discriminatingdroplet-beam correlated signals from the background. The He droplets are doped with Liand Rb atoms by passing through two vapor cells containing elementary alkali metal heatedto 400 ◦ C and 90 ◦ C, respectively.In the detector chamber, the He droplet beam crosses the synchrotron beam in the centerof the VMI-PEPICO detector at right angles. By detecting either electrons or ions with theVMI detector in coincidence with the corresponding particles of opposite charge on the TOFdetector, we obtain either ion mass-correlated electron images or mass-selected ion images.Kinetic energy distributions of electrons and ions are obtained by Abel inversion of theimages. The energy resolution of the electron spectra obtained in this way is ∆ E/E = 6 %. Acknowledgement
M.M. and L.B.L. acknowledge financial support by Deutsche Forschungsgemeinschaft (DFG,German Research Foundation, projects MU 2347/10-1 and BE 6788/1-1:1) and by the Carls-berg Foundation. A.C.L gratefully acknowledges the support by Carl-Zeiss-Stiftung. S.R.K.thanks Max Planck Society and D.S.T., Govt. of India, for support. E.F. gratefully acknowl-edges support by the Villum Foundation. N.S. acknowledges financial support by AgenceNationale de la Recherche through the program ANR-16-CE29-0016-01. The research lead-ing to this result has been supported by the project CALIPSOplus under Grant Agreement1330872 from the EU Framework Programme for Research and Innovation HORIZON 2020.
References (1) Sanche, L.
Radiation protection dosimetry , , 57–62.(2) Trinter, F. et al. Nature , , 664.(3) Gokhberg, K.; Kolorenč, P.; Kuleff, A. I.; Cederbaum, L. S. Nature , , 661.(4) Kamke, W.; Kamke, B.; Kiefl, H. U.; Wang, Z.; Hertel, I. V. Chem. Phys. Lett. , , 399–403.(5) Fröchtenicht, R.; Henne, U.; Toennies, J. P.; Ding, A.; Fieber-Erdmann, M.;Drewello, T. J. Chem. Phys. , , 2548–2556.(6) Scheidemann, A. A.; Kresin, V. V.; Hess, H. J. Chem. Phys. , , 2839–2844.(7) Ren, Y.; Kresin, V. V. Phys. Rev. A , , 043204.(8) Wang, C. C.; Kornilov, O.; Gessner, O.; Kim, J. H.; Peterka, D. S.; Neumark, D. M. J.Phys. Chem. A , , 9356.(9) An der Lan, L.; Bartl, P.; Leidlmair, C.; Schöbel, H.; Jochum, R.; Denifl, S.;D. Märk, T.; M. Ellis, A.; Scheier, P. J. Chem. Phys. , , 044309.(10) Buchta, D.; Krishnan, S. R.; Brauer, N. B.; Drabbels, M.; O’Keeffe, P.; Devetta, M.;Di Fraia, M.; Callegari, C.; Richter, R.; Coreno, M.; Prince, K. C.; Stienkemeier, F.;Moshammer, R.; Mudrich, M. J. Phys. Chem. A , , 4394–4403.(11) Siska, P. E. Rev. Mod. Phys. , , 337.(12) Katsuura, K. J. Chem. Phys. , , 3771–3775.(13) Cederbaum, L. S.; Zobeley, J.; Tarantelli, F. Phys. Rev. Lett. , , 4778–4781.1414) Hergenhahn, U. J. Electron Spectrosc. Rel. Phenom. , , 78–90.(15) Jahnke, T. J. Phys. B , , 082001.(16) Jahnke, T.; Czasch, A.; Schöffler, M.; Schössler, S.; Käsz, M.; Titze, J.; Kreidi, K.;Grisenti, R. E.; Staudte, A.; Jagutzki, O.; Schmidt, L. P. H.; Weber, T.; Schmidt-Böcking, H.; Ueda, K.; Dörner, R. Phys. Rev. Lett. , , 153401.(17) Sisourat, N.; Kryzhevoi, N. V.; Kolorenč, P.; Scheit, S.; Jahnke, T.; Cederbaum, L. S. Nature Physics , , 508.(18) Averbukh, V.; Müller, I. B.; Cederbaum, L. S. Phys. Rev. Lett. , , 263002.(19) Kuleff, A. I.; Gokhberg, K.; Kopelke, S.; Cederbaum, L. S. Phys. Rev. Lett. , ,043004.(20) Sann, H. et al. Chem. Phys. , , 221–225.(21) von Vangerow, J.; John, O.; Stienkemeier, F.; Mudrich, M. J. Chem. Phys. , ,034302.(22) Von Vangerow, J.; Coppens, F.; Leal, A.; Pi, M.; Barranco, M.; Halberstadt, N.;Stienkemeier, F.; Mudrich, M. J. Phys. Chem. Lett. , , 307–312.(23) Ruf, M.-W.; Yencha, A. J.; Hotop, H. Z. Phys. D , , 9–20.(24) Joppien, M.; Karnbach, R.; Möller, T. Phys. Rev. Lett. , , 2654–2657.(25) Eppink, A. T.; Parker, D. H. Rev. Sci. Instrum. , , 3477–3484.(26) Mudrich, M. et al. arXiv preprint arXiv:1905.04489 ,(27) LaForge, A.; Shcherbinin, M.; Stienkemeier, F.; Richter, R.; Moshammer, R.;Pfeifer, T.; Mudrich, M. Nature Physics , , 247.1528) Shcherbinin, M.; LaForge, A.; Hanif, M.; Richter, R.; Mudrich, M. , , 1855–1860.(29) Movre, M.; Thiel, L.; Meyer, W. J. Chem. Phys. , , 1484–1491.(30) Soldán, P.; Lee, E. P. F.; Lozeille, J.; Murrell, J. N.; Wright, T. G. Chem. Phys. Lett. , , 429–436.(31) von Haeften, K.; Castro, A. R. B. d.; Joppien, M.; Moussavizadeh, L.; Pietrowski, R. v.;Möller, T. Phys. Rev. Lett. , , 4371–4374.(32) Ziemkiewicz, M. P.; Neumark, D. M.; Gessner, O. Int. Rev. Phys. Chem. , ,239–267.(33) Müller, S.; Mudrich, M.; Stienkemeier, F. J. Chem. Phys. , , 044319.(34) Galli, D. E.; Buzzacchi, M.; Reatto, L. J. Chem. Phys. , , 10239–10247.(35) Miteva, T.; Kazandjian, S.; Sisourat, N. Chem. Phys. , , 208–215.(36) Hernando, A.; Barranco, M.; Pi, M.; Loginov, E.; Langlet, M.; Drabbels, M. Phys.Chem. Chem. Phys. , , 3996–4010.(37) Ovcharenko, Y. et al. Phys. Rev. Lett. , , 073401.(38) Iablonskyi, D. et al. Phys. Rev. Lett. , , 276806.(39) Serdobintsev, P. Y.; Melnikov, A. S.; Pastor, A. A.; Timofeev, N. A.; Khodor-kovskiy, M. A. J. Chem. Phys. , , 194301.(40) Buchta, D.; Krishnan, S. R.; Brauer, N. B.; Drabbels, M.; O’Keeffe, P.; Devetta, M.;Di Fraia, M.; Callegari, C.; Richter, R.; Coreno, M.; Prince, K. C.; Stienkemeier, F.;Ullrich, J.; Moshammer, R.; Mudrich, M. J. Chem. Phys. , , 084301.(41) Dick, B. Phys. Chem. Chem. Phys. ,16