Interatomic-Coulombic-decay-induced recapture of photoelectrons in helium dimers
P. Burzynski, F. Trinter, J. B. Williams, M. Weller, M. Waitz, M. Pitzer, J. Voigtsberger, C. Schober, G. Kastirke, C. Müller, C. Goihl, F. Wiegandt, R. Wallauer, A. Kalinin, L. Ph. H. Schmidt, M. Schöffler, G. Schiwietz, N. Sisourat, T. Jahnke, R. Dörner
aa r X i v : . [ phy s i c s . a t m - c l u s ] S e p Phys. Rev. A 90, 022515 (2014)
Interatomic-Coulombic-Decay induced recapture of photoelectrons in Helium Dimers
P. Burzynski , F. Trinter , J. B. Williams , M. Weller , M. Waitz , M. Pitzer , J. Voigtsberger ,C. Schober , G. Kastirke , C. M¨uller , C. Goihl , F. Wiegandt , R. Wallauer , A. Kalinin , L.Ph. H. Schmidt , M. Sch¨offler , G. Schiwietz , N. Sisourat , , T. Jahnke , and R. D¨orner ∗ Institut f¨ur Kernphysik, J. W. Goethe-Universit¨at,Max-von-Laue-Str. 1, 60438 Frankfurt am Main, Germany Helmholtz-Zentrum Berlin f. Materialien u. Energie,Institute G-ISRR, Hahn-Meitner-Platz 1,14109 Berlin, Germany Sorbonne Universit´es, UMR 7614,Laboratoire de Chimie Physique Mati`ere et Rayonnement,F-75005 Paris, Franceand CNRS, UMR 7614,Laboratoire de Chimie Physique Mati`ere et Rayonnement,F-75005 Paris, France
We investigate the onset of photoionization shakeup induced interatomic Coulombic decay (ICD)in He at the He + ∗ ( n = 2) threshold by detecting two He + ions in coincidence. We find this thresholdto be shifted towards higher energies compared to the same threshold in the monomer. The shiftedonset of ion pairs created by ICD is attributed to a recapture of the threshold photoelectron afterthe emission of the faster ICD electron. PACS numbers:
Excited ions can get rid of their excess energy via theemission of a photon or an electron. If, however, the ex-cited atom is spatially close to other atoms and the ex-citation energy is above the ionization threshold of thisneighbor, the excess energy can also be transfered to theneighbor where it leads to emission of an electron. Thisenergy transfer process is termed interatomic Coulom-bic decay (ICD). It was introduced by Cederbaum andcoworkers in 1997 [1] and was demonstrated experimen-tally first for Neon clusters [2] and Neon dimers [3].The related interatomic Auger transitions in solid matterhave ofen been discussed, but broad valence bands, sur-face/bulk differences and significant electron energy-lossprocesses do typically preclude a clear assignment of thisprocess for solids. Many studies have shown since thenthat ICD is a very general phenomenon occuring in vander Waals bound (see e.g. [2–5]) and hydrogen boundsystems (see e.g. [6–8]). It can be induced by photoion-ization (see e.g. [2, 3]), photoexcitation [9–12], Augerdecay [13, 14], ion impact [15–17], and electron impact[18] or as in the present case after shakeup [19]. Themost extreme system for which ICD has been reportedis the Helium dimer [20, 21]. The neutral He dimer isvery weakly bound (about 95 neV) and the internucleardistance extends to very large distances, with the mean ∗ Electronic address: [email protected] distance of about 52 ˚A. ICD in He can occur when oneof the Helium atoms is ionized and its remaining elec-tron is shaken up to any excited state (He + ∗ (n=2,3...)).In the next step, the He + ∗ (n=2)He contracts and duringthat nuclear motion it undergoes ICD. The electron ofthe exited He + relaxes to the ground state and the en-ergy is transfered to the neutral neighbor where the ICDelectron is emitted. Finally, the two He + ions Coulombexplode back-to-back:He-He hν −→ He + ∗ -He + e ph ICD −−−→ He + + He + + e ph +e ICD
Usually, ICD and the subsequent Coulomb explosionis discussed in a two step picture, where the decay isindependent from the initial ionization/excitation pro-cess. In the present work, we show that close to the ion-ization/excitation threshold this two step approximationbreaks down. Photoelectron and ICD electron interac-tion can lead to recapture of the photoelectron into abound state of one of the two ions, which quenches theCoulomb explosion. A direct link between the ionizationprocess and ICD has been discussed in two contexts inthe literature so far. The first is the recoil effect, whereit has been shown experimentally [22] and theoretically[23] that the recoil momentum of the photoelectron or anAuger electron can induce nuclear motion, which in turnmodifies the ICD energy spectrum. The second contextis closely related to the present work, where Trinter etal. [24] have seen a shift of the photoelectron energy dueto post collision interaction (PCI) with the ICD electron[25, 26]. In a time dependent picture the photoelectronis originally created in the potential of a singly chargedspecies. After some delay the ICD electron is emitted,but then it surpasses the slow photoelectron, which fromthen on feels the attractive potential of a doubly chargedion. This slows down the photoelectron. This streakingtowards lower photoelectron energies depends on the timedelay between the photoabsorption and the ICD electronand can easily be modeled. It has recently been usedto make the first ’movie’ of nuclear motion during ICD[24]. The same process of post collision interaction leads,for very small photoelectron energies, to a recapture ofa part of the photoelectron wave packet, which is theeffect we study here. For atomic Auger decay followinginnershell ionization, PCI is well understood [25, 26]. Re-cently Sch¨utte et al. [27] have experimentally verified thetime dependent picture we have just discussed, by show-ing that the energy exchange between photoelectron andAuger electron indeed depends on the Auger emissiontime delay. For atomic multiple ionization, such recap-ture of the photoelectron by PCI, leads to a shifted onsetof the production of higher charge states, as e.g. reportedfor Ar [28, 29].The present experiment has been performed at beamline UE112-PGM-1 in the synchrotron radiation facilityBESSY (Berlin) during single bunch operation using aCOLTRIMS reaction microscope [30–32]. The photonbeam was intersected with a supersonic He gas jet inthe center of the COLTRIMS spectrometer. A 7.5 V/cmhomogeneous electric field guided the ions towards a po-sition sensitive micro channel plate detector with hexago-nal delay-line readout (RoentDek HEX90) [33]. A nozzletemperature of 21 K at 2.5 bar driving pressure resultedin a fraction of about 1-2% He in the atomic gas jet. Thephoton energy was scanned across the He + (n=2) thresh-old. In the offline analysis the momentum vectors of theions were obtained from the position of impact at thedetector and the time-of-flight. We analyzed ion pairsemitted back-to-back and the simultaneously measuredHe + (n=2) monomer ions from the atomic helium frac-tion of the gas jet. The back-to-back events dominatethe recoil pair-correlation function slightly above thresh-old. This fact is related to a small fraction of sequentialpair production due to the limited photon density of theenergy-filtered synchrotron beam, to an extremly smallprobability of the correlated electron knock-off transi-tions close to the threshold and to a small fraction ofsecondary ion/atom collisions in the gas jet. He + (n=2)ions have been discriminated from the He + (n=1) ions bythe ion momentum vector (see [34] Fig. 1). Close tothe He + ( n = 2) threshold the He + (n=1) ions carry the1.7 a.u. recoil momentum of the photoelectron while theHe + ∗ (n=2) ions are accompanied by a zero kinetic energyphotoelectron and hence have almost no recoil momen-tum.Fig. 1 shows the measured kinetic energy release(KER, summed over both recoil-ion energies) as functionof the photon energy. The KER above the He + ∗ (n=2) FIG. 1: (A) Kinetic energy release (KER) of He + ion pairsas function of photon energy, red arrow shows the thresholdfor creating He + ∗ (n=2) for a helium atom. (B) Red full line:Projection of (A) onto the y-axis for photon energy range65.41-65.42 eV. Black dotted line: Theoretical KER distribu-tion. The two curves are normalized to the maximum. threshold (E γ range=65.41-65.42 eV, Fig. 1b) is in ex-cellent agreement with published work [20]. It shows avibrational structure from the contracting dimer with amaximum between 8-9.5 eV, which results from nuclearwave packet hitting the inner turning point on the poten-tial energy surface of the excited dimer ion (see [21, 35]for a detailed analysis).Fig. 2 shows the photon energy dependence of the ionpair count rate, i.e. a projection of the data from Fig.1a onto the horizontal axis (filled circles) and count ratefrom He + ∗ (n=2) monomers (open circles). Both datasetsare normalized to the highest energy point. A constantbackground for the ion pairs probably resulting from aknock off process [36, 37] and higher harmonics from thebeam line has been subtracted.Owing to the resolution of the beam line, the countrates are not a step function at the He + ∗ (n=2) thresh-old. The resolution of the beam line is σ = 4.5 meVwhich is extracted from the data of the monomers. Thesemonomers come from atomic helium. The black dashedline of Fig. 2 is a convolution of the step function (blackdotted) with a Gaussian function with a FWHM of 4.5meV. This line follows very well the monomer ion countrate. The ion pair count rate shows a significant energy c o un t s ( a r b . U n i t s ) ❍ He +* (n=2) monomer ● He + /He + from ICD FIG. 2: Photon energy dependence of the count rate for:two He + ions (filled circles) (projection of data in Fig. 1aonto the x-axis) and He + ∗ (n=2) ions from ionization of themonomer (open circles). Black dotted step function: thresh-old for He + ∗ (n=2), black dashed line: step function convo-luted with a Gaussian function ( σ = 4.5 meV) for the energyresolution of the beam line. Green full line: expected onsetof ICD without taking PCI into account (see text), red dashdotted line: model including recapture of photoelectrons byPCI using calculated widths from Table I (see text) as well asthe experimental resolution of σ = 4.5 meV. The data pointsare normalized at the highest points. offset compared to the monomers. Note that monomersand ion pairs are measured simultaneous in our setup,which excludes any possible systematic error on the pho-ton energy as the origin of this energy offset. We willshow now that the shifted onset of the ion pair produc-tion is caused by recapture of the slow near thresholdphotoelectron after ICD, which neutralizes one of thetwo ions. A classical modeling of this recapture process,summing over all calculated ICD channels using calcu-lated ICD lifetimes, results in the dash dotted red line inFig. 2, which nicely reproduces the shifted onset. Thereare no free parameters in this calculation. The model ispresented in the following.In the present data, the photoelectron energy is below50 meV, while the ICD electron has an energy of 6-15 eV.Even though there is a time delay ( t ICD ) between theemission of the photoelectron and ICD electron, the ICDelectron will always overtake the photoelectron in thevicinity of the residual ion(s), because t ICD is small andthe ICD electron is much faster than the photoelectron.Note that t ICD is not the lifetime of the respective ICDchannel, but the time at which ICD occurs for this indi-vidual event. For the present case the travel time of theICD electron is negligible compared to t ICD . The pho-toelectron therefore starts its way to the continuum ini- tially leaving behind a singly charged species. However,when the ICD electron is emitted the ion charge state in-creases from singly to doubly charged, which changes thepotential the photoeletron feels, at least after it has beenovertaken by the ICD electron, from − /r e to − /r e (inatomic units), where r e is the distance the photoelectronhas reached at time t ICD .We calculate the trajectory of the photoelectron clas-sically by starting the electron at a distance r es in aCoulomb potential with an initial kinetic energy E γ − E IP X + 1 /r es where E γ is the photon energy and E IP X is the ionization potential of He plus the energy it takesto excite He + from its ground state to He + ∗ (n=2). Wehave chosen r es = 10 a.u. and have verified that theresults are insensitive of this choice over a wide range.Without the ICD electron the photoelectron would es-cape to the continuum. However, if ICD occurs, the elec-tron loses the energy − /r e and might get trapped inthe ionic potential. For every particular electronic andvibrational He + ∗ (n=2)He state ICD occurs with an ex-ponential time dependence. We calculate the fraction ofrecapture photoelectrons using this exponential distribu-tion of t ICD for each photon energy. We then sum overall electronic and vibrational states. The electronic statesare weighted with their statistical weight and the vibra-tional states with their Franck-Condon overlap with the He ground state. Franck-Condon factors, energies andICD lifetimes for each state are given in Table I.There are two counteracting effects included in this cal-culation. First the threshold for each state He + ∗ (n=2)Hewill be slightly below the He + ∗ (n=2) threshold for themonomer, because of the larger size of the combined elec-tronic potential due to both heavy constituents of thedimer (see Table I for binding energies). Thus neglectingthe recapture the threshold for ion pair creation wouldbe slightly smeared to lower energies, as shown by thefull green curve in Fig. 2. If one includes the recaptureas described above the full green curve is modified to thedash dotted red curve in Fig. 2, which is shifted towardshigher photon energies. This dash dotted red curve alsoincludes the photon energy resolution which we obtainedin situ from the monomer. This calculation captures themain effect seen in the experiment. Experimentally, wealso see a small contribution of ion pairs below threshold,which is not reproduced by our calculation. The origin ofthese contributions below threshold is unclear, but mightbe related to details of the angular distribution of ejectedICD and photoelectrons or to the accuracy of the com-puted negative energy shift.In conclusion, we have shown that close to thresholdfragment creation by ICD cannot be treated indepen-dently of the ionization process. Post collision interactionof the photoelectron with the ICD electron can even leadto recapture of the photoelectron. This observed effectis similar to PCI between photoelectron and fast Augerelectrons in atomic species. While we have studied onlythe monopole term of this post collision here, we expectthat the angular distributions of the electrons will also be ν Σ + g : 2 p z , s Σ + u : 2 p z , s Π g : 2 p x,y Π u : 2 p x,y Franck-Condon factors . · − . · − . · − . · − . · − . · − . · − . · − . · − . · − . · − . · − . · − . · − . · − . · − . · − . · − . · − . · − . · − . · − . · − . · − . · − . · − . · − . · − . · − . · − . · − . · − . · − . · − . · − - lifetimes [fs] negative shifts of the vibrational levelcompared to the monomer [eV] · − · − · − · − · − · − · − · − · − · − · − · − · − · − · − · − · − · − · − · − · − · − · − · − · − · − · − · − · − · − · − · − · − · − · − -dissociation limit 65.393 eVTABLE I: Calculated characteristics of the He + ∗ (n=2)Hestates: Franck-Condon factors for the overlap with the He ground state, vibrational state energies and ICD lifetimes [35]. altered, an effect studied recently for the atomic Augercase [26]. The discussed recapture will also be active atthe threshold for creation of He + ∗ in n=3 and higher. Inthis case the angular distributions of photoelectron andICD electron will be different(see [38]). We believe thatPCI in the continuum will in the future become a majortool for ultrafast time resolved studies, as shown recentlyin pioneering work by Trinter et al. [24]. Acknowledgments
We thank the staff of BESSY II for experimen-tal support. This work was funded by the DeutscheForschungsgemeinschaft and supported by RoentDekHandels GmbH. [1] L.S. Cederbaum, J. Zobeley and F. Tarantelli,
Phys. Rev.Lett.
Phys. Rev. Lett.
Phys. Rev. Lett.
Phys. Rev. Lett.
Phys. Rev. Lett.
J. Chem. Phys.
NaturePhysics
139 (2010).[8] M. Mucke, M. Braune, S. Barth, M. F¨orstel, T. Lischke,
V. Ulrich, T. Arion, U. Becker, A. Bradshaw and U. Her-genhahn,
Nature Physics
143 (2010).[9] F. Trinter, J. B. Williams, M. Weller, M. Waitz, M.Pitzer, J. Voigtsberger, C. Schober, G. Kastirke, C.M¨uller, C. Goihl, P. Burzynski, F. Wiegandt, R. Wal-lauer, A. Kalinin, L. Ph. H. Schmidt, M. Sch¨offler, Y.-C.Chiang, K. Gokhberg, T. Jahnke, R. D¨orner,
Phys. Rev.Lett.
J. Phys. Chem. Lett. Europhys. Lett.
228 (2005).[12] B. Najjari, A.B. Voitkiv. and C. M¨uller,
Phys. Rev. Lett.
Phys. Rev. Lett.
Phys. Rev. A.
PNAS
Phys. Rev. A.
Phys. Rev. Lett.
Phys. Rev. A.
Phys. Rev. Lett.
Phys. Rev. Lett.
Nature Physics
508 (2010).[22] K. Kreidi, Ph. Demekhin, T. Jahnke, Th. Weber,T. Havermeier, X. Liu, Y. Morisita, S. Sch¨ossler, L.Schmidt, M. Sch¨offler, M. Odenweller, N. Neumann, L.Foucar, J. Titze, B. Ulrich, F. Sturm, C. Stuck, R.Wallauer, S. Voss, I. Lauter, H. Kim, M. Rudloff, H.Fukuzawa, G. Pr¨umper, N. Saito, K. Ueda, A. Czasch, O. Jagutzki, H. Schmidt-B¨ocking, S. Scheit, L. Cederbaumand R. D¨orner,
Phys. Rev. Lett.
J. Chem.Phys.
Phys. Rev. Lett.
J. Phys. B: At., Mol. Opt.Phys.
Phys. Rev. Lett.
Phys. Rev. Lett.
Phys. Rev. A.
Phys. Rev. Lett.
Rep. Prog. Phys.
PhysicsReports
J. Electron Spectrosc.Relelat. Phenom.
229 (2004).[33] O. Jagutzki, A. Cerezo, A. Czasch, R. D¨orner, M. Hat-tass, M. Huang, V. Mergel, U. Spillmann, K. Ullmann-Pfleger, Th. Weber, H. Schmidt-B¨ocking and G.D.W.Smith,
IEEE Transact. on Nucl. Science
Phys. Rev. Lett.
Phys. Rev. A.
Phys. Rev. Lett.
Phys. Rev. A.
Phys. Rev. A.82,