Highly efficient double ionization of mixed alkali dimers by intermolecular Coulombic decay
A. C. LaForge, M. Shcherbinin, F. Stienkemeier, R. Richter, R. Moshammer, T. Pfeifer, M. Mudrich
HHighly efficient double ionization of mixed alkali dimers by intermolecular decay
A. C. LaForge,
1, 2, ∗ M. Shcherbinin, F. Stienkemeier, R. Richter, R. Moshammer, T. Pfeifer, and M. Mudrich Physikalisches Institut, Universit¨at Freiburg, 79104 Freiburg, Germany Department of Physics, University of Connecticut, Storrs, Connecticut, 06269, USA Department of Physics and Astronomy, Aarhus University, 8000 Aarhus C, Denmark Elettra-Sincrotrone Trieste, 34149 Basovizza, Trieste, Italy Max-Planck-Institut f¨ur Kernphysik, 69117 Heidelberg, Germany
As opposed to purely molecular systems where electron dynamics proceed only through in-tramolecular processes, weakly-bound complexes like helium droplets offer an environment wherelocal excitations can interact with neighboring embedded molecules leading to new intermolecular relaxation mechanisms. Here, we report on a new decay mechanism leading to the double ionizationof alkali dimers attached to helium droplets by intermolecular energy transfer. From the electronspectra, the process is similar to the well-known shakeoff mechanism observed in double Auger decayand single photon double ionization [1, 2], however, in this case, the process is dominant, occurringwith efficiencies equal to, or greater than, single ionization by energy transfer. Although an alkalidimer attached to a helium droplet is a model case, the decay mechanism is relevant for any systemwhere the excitation energy of one constituent exceeds the double ionization potential of anotherneighboring molecule. The process is, in particular, relevant for biological systems, where radicalsand slow electrons are known to cause radiation damage [3].
INTRODUCTION
The correlated action of multiple electrons after pho-ton absorption in atomic and molecular systems has ledto the discovery of a variety radiation-induced decay pro-cesses (e.g. multiple excitation/ionization or various au-toionization channels.) [1, 4, 5]. When the system com-plexity is increased to larger, more complex systems, newand diverse intermolecular decay mechanisms open up.In particular, processes such as intermolecular Coulombicdecay (ICD) [6] where energy is exchanged between elec-tronically excited atoms or molecules and their neighborshave been of broad interest; for reviews, see [7, 8]. ICDand related intermolecular processes are a potentially im-portant channel for radiation damage of biologically rel-evant systems [9, 10]. Recently, ICD was measured forthe first time in a hydrated biomolecular system [11].Due to its simple electronic structure and high ion-ization potential, weakly-bound helium (He) complexeshave served as a model system for studying intermolec-ular processes such as ICD [12–14] and electron trans-fer mediated decay (ETMD) [15, 16]. Additionally, Hecomplexes have been used to observe a related type ofICD. In this case, when He droplets [17, 18] or He-Nedimers [19] are resonantly excited, the energy can betransferred to neighboring constituents leading to theirionization. An example of such a process is shown in theelectron kinetic energy distribution in Fig. 1d) for potas-sium atoms (K) attached to the surface of He droplets.A photon ( hν = 21.6 eV) is initially absorbed by a Henanodroplet at the resonance correlating to the 1 s p P -state of atomic He [20]. Through ultrafast intraband re-laxation within the droplet [21, 22], an excited 1 s s S He atom ( E e = 20.6 eV) is formed. The excess energy isthen transferred by ICD to the K atom leading to its ionization while the He atom relaxes to its ground state.The characteristic electron kinetic energy is the differencebetween the He excited state and the acceptor’s ioniza-tion potential. For the case of K atoms, this results in akinetic energy of about 16.3 eV, which matches the posi-tion of the pronounced peak in the spectrum (black linein Fig. 1d)). A question which arises is how the situationchanges for systems where double ionization is energet-ically allowed. For endofullerenes, it has been theoret-ically proposed that double ICD (dICD) can become aviable decay mechanism [23]. Here, we show that, in-deed, dICD is not only a possible decay path, but caneven be the dominant decay mechanism occurring withefficiencies equal to, if not exceeding, single ionization. RESULTS AND DISCUSSION
To gain detailed insight into the process, we turn ourinvestigation to alkali dimers, the simplest metal cluster.Furthermore, to circumvent issues with detector deadtimes, the dimers are composed of alkali atoms of dif-ferent mass, K and Rb (rubidium). The process of dICDis schematically shown in Fig. 2 along with the potentialenergy curves of free K-Rb dimers in the ground X Σ + state (black line) [24] and dicationic state (red line). Thedicationic curve was calculated using a Coulomb poten-tial shifted to match the asymptotic ionization energyof the free atoms, which are given as dashed lines inFig. 2. Similar to the ICD ionization of K atoms de-scribed above, dICD occurs by a transfer of energy fromthe excited 1 s s S He atom ( E e = 20.6 eV) to the K-Rbdimer. However, in this case, the double ionization po-tential of the K-Rb dimer is energetically less than theexcited He atom resulting in the emission of two electrons a r X i v : . [ phy s i c s . a t m - c l u s ] D ec K + m o m e n t u m ( a . u . ) R b + e -K + R b K + + R b + i o n 1 t i m e - o f - f l i g h t ( m s ) K + + K +2 b ) d ) c ) a ) ion 2 time-of-flight ( m s) K + + R b + I C D electron intensity (arb. units) e l e c t r o n k i n e t i c e n e r g y ( e V ) K + + R b + K + d I C D FCF
F C F f o r K - R b ( S + ) i o n k i n e t i c e n e r g y ( e V ) ion intensity (arb. units) K + R b + F C F f o r K - R b ( S + ) K + R b + FCF
Figure 1. Coincidence spectra for discriminating possible involved decay mechanisms applied while measuring the energetics ofthe constituent ions and electrons. a) The ion-ion coincidence time-of-flight spectrum for K-Rb dimers attached to the surfaceof He droplets. The ionization process is triggered by energy transfer from the excited 1 s s S He atom ( E e = 20.6 eV). b) TheK + , Rb + , and electron VMIs taken in triple (e − , K, Rb) coincidence. The momenta scales are given in atomic units. c)The kinetic energy distribution for the K + (blue line) and Rb + (gray line) taken in triple (e − , K, Rb) coincidence. d) Theelectron kinetic energy distributions taken in triple (e − , K, Rb) coincidence (red line) and double (e − , K) coincidence(black line). Note that the black line in d) was a separate measurement where single K atoms were attached to the surfaceof He droplets. Its expected kinetic energy is given by a dashed vertical line. The filled lines in c) and d) correspond to therespective ion and electron kinetic energy distributions for K-Rb dimers calculated from Franck-Condon factor simulations (seetext for details). along with the dicationic dissociation of the ions.Direct evidence for dICD is thus determined by mea-suring multiple coincidences of electrons and ions pro-duced by this process. Fig. 1 a) shows the electron-ion-ion coincidence time-of-flight spectrum for K-Rb dimersattached to the surface of He droplets for a photon en-ergy of 21.6 eV. In general, distributions observed in ion-ion coincidence maps identify ions created by multipleionization while the shape of the distribution gives infor-mation about the dissociation process [25]. In this case,the coincidence map is centered around the respectivemasses of K and Rb where several sharp, negative slopingfeatures are observed. These distributions indicate thatfragmentation occurs through dicationic dissociation ofthe dimers leading to back-to-back emission of the ions.The primary ion pair originates from dimers of K and Rb while the neighboring distributions come from theisotopes, K and Rb. There are additional, weakerdistributions due to complexes of an alkali ion with a fewHe atoms attached. For cases where the He droplet isnot resonantly excited, no such distributions are observedindicating that ionization proceeds through excited Heatoms. Additionally, using electron-ion-ion coincidenceimaging techniques, one can extract electron/ion kinetic energy spectra from the individual ion pairs in the coin-cidence map (see Fig. 1 c) and d)).Fig. 1 b) shows the raw velocity map images (VMIs)of ions and electrons measured in triple (e − , K, Rb)coincidence. VMIs are two-dimensional projections ofthe charged particle’s momentum sphere which are theninverted to obtain kinetic energy distributions for the re-spective electrons and ions [26]. The left and middleimages show VMIs of K + and Rb + ions, and the rightimage shows the electron VMI. The clearly visible ringstructure in all VMIs indicates a non-zero kinetic energycomponent.From the VMIs, we determine the kinetic energy dis-tributions by Abel inversion [26]. Fig. 1 c) shows the ionkinetic energy distributions for the K ion (blue line)and Rb ion (gray line) measured in triple (e − , K, Rb) coincidence. The ions have broad kinetic energiescentered around 3.75 eV and 1.5 eV, respectively. Thesum of these energies corresponds to the kinetic energyrelease of the ion pair in the dicationic state as illustratedin Fig. 2. To assess this conjecture, we have performedFranck-Condon factor (FCF) simulations of the ion andelectron kinetic energy distributions assuming verticaltransitions between the potential energy curves given in K + +Rb + Rb + P o t en t i a l ene r g y ( e V ) Interatomic distance R (Å) [K-Rb] K-Rb( S + ) K + K-Rb + He * ® [K-Rb] + He + 2e - Figure 2. The potential energy curve of K-Rb dimers inthe ground (black line) and dicationic (red line) state. Theasymptotic limit of the dicationic state is given by the dashedred line along with the individual ionization potentials of K(dashed blue line) and Rb (dashed gray line). A schematic ofthe process is given where the K-Rb dimer is represented byblue and gray spheres, the He atom by a red sphere, and theHe droplet by a yellow sphere. The photon ( hν = 21.6 eV)is initially absorbed by the He droplet at the 1 s p P res-onance. Through ultrafast intraband relaxation within thedroplet [21, 22], an excited 1 s s S He atom ( E e = 20.6 eV)is formed. The excess energy is then transferred by dICD tothe K-Rb dimer leading to its double ionization while the Heatom relaxes to its ground state. Fig. 2. The initial state is assumed to be an excited Heatom in the 1 s s S -state ( E e = 20.6 eV) interacting withthe alkali dimer in its vibronic ground state. [17, 18]. Theresults are shown as filled peaks. Note that the 1 s s S -state of a He atom in a droplet is still dipole-coupled tothe ground state [20], thereby allowing ICD-like energytransfer to occur. The kinetic energy release from theFCF simulations, shown in Fig. 1 c), gives quantitativelysimilar results to the measured values, but drastically un-derestimates the width. Broadening of the experimentaldistributions is likely due to perturbations of the initialand final, ionic state by dopant-He droplet interactions.In particular, the transient attachment of the localizedexcited He atom to the K-Rb dimer may lead to its sta-bilization. Depending on the configuration of the state,dICD proceeds at different internuclear distances of theK-Rb dimer resulting in a broader distribution of thefragmented ions.Fig. 1 d) shows the electron kinetic energy distribution(red line) measured in triple (e − , K, Rb) coincidence.The spectrum shows two peaks centered at 0 eV and 8 eV,which arise from double ionization of alkali dimers. Thesimulated excess electron energy for double ionization ofK-Rb dimers is 8 eV (filled peak) fitting well with thesum electron energy. The measured kinetic energy spec-trum shows a U-shaped distribution indicating one elec- a ) d I C D
L i + I C D
L i + + L i + b ) electron intensity (arb. units) N a + N a + + N a + c ) K + K + + K + d ) R b + E l e c t r o n e n e r g y ( e V ) R b + + R b + Figure 3. Electron kinetic energy distributions from the ion-ization of small, homogeneous clusters of alkali metals at-tached to the surface of a He droplet. Independent spectrafor a) Li, b) Na, c) K, and d) Rb. The black filled lines weretaken in double (e − , Ak + ) coincidence and the red filled lineswere taken in triple (e − , Ak + , Ak + ) coincidence, where Akare the alkali metals shown in a)-d). The excited 1 s s S Heatom ( E e = 20.6 eV) triggers the energy transfer process. tron takes nearly all of the excess energy while the sec-ond electron is emitted with nearly zero kinetic energy.Similar distributions have previously been observed insingle-photon double ionization of atoms (SPDI) [2, 5]and double Auger decay (DAD) [1, 27]. In those cases,the mechanism, known as shakeoff, is due to the suddenremoval of the primary electron leaving the system in aperturbed ionic state; the secondary electron then has aprobability of relaxing to an unbound state resulting inan unequal sharing of the excess energy. The electronenergy distribution in Fig. 1 d) shows a similar distribu-tion to shakeoff, but, in contrast, occurs relatively closeto the double ionization threshold. This could, in part,be due to the low ionization potential of the valence elec-tron for alkali atoms. The overall similarity to shakeoffindicates that dICD proceeds through a one-step processas opposed to other two-step electron impact ionizationmechanisms in SPDI such as knockout [2].We have verified dICD for several mixed alkali dimersystems (K-Rb, Na-K, Na-Rb), small homogeneous al-kali clusters (Li, Na, K, Rb, and Cs), and even alkalineearth atoms (Ba). Fig. 3 shows the electron kinetic en-ergy distributions of small, homogeneous clusters of a)Li, b) Na, c) K, and d) Rb attached to the surface of aHe droplet. The excited 1 s s S He atom ( E e = 20.6 eV)triggers the energy transfer process. The black filled lineswere taken in double (e − , Ak + ) coincidence and the redfilled lines were taken in triple (e − , Ak + , Ak + ) coinci-dence with the alkali metal ions, Ak + , where Ak denotesLi, Na, K, Rb. The red filled lines show electrons emit-ted from dICD while the black filled lines show electronsemitted from ICD, occurring at higher kinetic energies,as well as dICD. Due to the comparable electronic struc-ture and ionization potentials of alkali metals, their dis-tributions exhibit similar features. In all cases shown,dICD is a prominent decay channel leading one to con-clude the process is rather ubiquitous and not limitedspecifically to K-Rb dimers where the excited ionic stateof Rb could also lead to double ionization through a cas-cade mechanism. In particular, the asymmetric distribu-tion observed in dICD is evident of a similar one-step,shakeoff-type ionization mechanism.Surprisingly, as can be seen in Fig. 3, dICD is a highlyefficient process showing comparable, if not larger, ion-ization rates to ICD. In contrast, for SPDI near thresh-old, the branching ratio to single ionization is much lessthan 1% for atoms [28] and small molecules [29]. ForDAD, the branching ratio to Auger decay is typically afew percent for atoms [30]. As such, one can concludethat dICD can even be the dominant process in weakly-bound systems for cases where it is energetically allowed.In general, dICD should not be limited to outer valenceshell excited atoms; Auger forbidden inner valence shellexcited/ionized atoms, which have even higher excitationenergies, have the potential for dICD as well. Addition-ally, the multiple ions and electrons formed in the processof dICD should play an important role in biological sys-tems. For instance, core shell-ionization of a solvatedmagnesium dication leads to a variety of cascade chan-nels where Auger and intermolecular decay processes oc-cur [31]. For each step where ICD is allowed, dICD wouldalso be an energetically open decay channel leading to anenhancement in the production of neighboring water ionsand low energy electrons. The subsequent ionization ofwater typically leads to proton transfer and the forma- tion of the hydroxyl radical, a highly reactive damagecenter [32], while the production of low energy electronsis a known source of radiation damage for proteins andDNA [3]. METHODSExperimental setup
The experiment was performed using a mobile Hedroplet machine attached to a velocity map imaging pho-toelectron photoion coincidence spectrometer [33] at theGasPhase beamline of Elettra-Sincrotrone Trieste, Italy.The setup has been described in some detail earlier [18],and only the significant points will be addressed here.In short, a beam of He nanodroplets is produced bycontinuously expanding pressurized He (50 bar) of highpurity (He 6.0) out of a cold nozzle (T = 7 −
40 K)with a diameter of 5 µ m into vacuum. Under these ex-pansion conditions, the mean droplet sizes range from10 to 10 He atoms per droplet [34]. After passing askimmer (0.4 mm) and a mechanical beam chopper usedfor discriminating the droplet beam signal from the Hebackground, the droplets were doped using the “pick-up” technique [35] with subsequent heated doping cellsfilled with alkali metals. While most atomic and molec-ular species become submerged into the interior of Henanodroplets, alkali atoms remain weakly-bound on thesurface [36]. The He droplet beam next crosses the syn-chrotron beam inside of a PEPICO detector consisting ofan ion time-of-flight detector and velocity map imagingdetector (5% ∆
E/E resolution). With this setup, onecan record either electron or ion kinetic energy distribu-tions, depending on the polarity, in coincidence with onespecific ion mass or with several ion masses in multicoin-cidence mode [33]. When electrons are recorded on theVMI, only one electron for each coincidence event can bedetected. The kinetic energy distributions were recon-structed using the Maximum Entropy Legendre Recon-struction method [26]. The polarization axis was perpen-dicular to the VMI axis to ensure cylindrical symmetrywhich is required for the inversion process. The pho-ton energy was set to 21.6 eV and a tin filter was usedto eliminate any higher-order light contamination. Thepulse repetition rate was 500 MHz.
DATA AVAILABILITY
The data that support the plots within this paper andother findings of this study are available from the corre-sponding authors on request.
AUTHOR CONTRIBUTIONS STATEMENT
A.C.L. and M.M. conceived the experiment. A.C.L,M.S., and R.R. conducted the experiment. A.C.L, M.S.,and M.M. analyzed the data. M.M. performed the FCFcalculations. A.C.L. interpreted the data with help fromR.R., F.S., R.M., T.P., and M.M.. A.C.L. wrote thepaper. All authors reviewed the manuscript. ∗ [email protected][1] T. A. Carlson, Phys. Rev. , 142 (1967).[2] T. Schneider, P. L. Chocian, and J.-M. Rost, Phys. Rev.Lett. , 073002 (2002).[3] B. Bouda¨ıffa, P. Cloutier, D. Hunting, M. A. Huels, andL. Sanche, Science , 1658 (2000).[4] R. Madden and K. Codling, Phys. Rev. Lett. , 516(1963).[5] R. Wehlitz, F. Heiser, O. Hemmers, B. Langer, A. Men-zel, and U. Becker, Phys. Rev. Lett. , 3764 (1991).[6] L. S. Cederbaum, J. Zobeley, and F. Tarantelli, Phys.Rev. Lett. , 4778 (1997).[7] U. Hergenhahn, J. Electron. Spectrosc. Relat. Phenom. , 78 (2011).[8] T. Jahnke, J. Phys. B: At., Mol. Opt. Phys. , 082001(2015).[9] K. Gokhberg, P. Kolorenˇc, A. I. Kuleff, and L. S. Ceder-baum, Nature , 661 (2014).[10] F. Trinter, M. Sch¨offler, H.-K. Kim, F. Sturm, K. Cole,N. Neumann, A. Vredenborg, J. Williams, I. Bocharova,R. Guillemin, et al. , Nature , 664 (2014).[11] X. Ren, E. Wang, A. D. Skitnevskaya, A. B. Trofimov,K. Gokhberg, and A. Dorn, Nat. Phys. , 1062 (2018).[12] N. Sisourat et al. , Nat. Phys. , 508 (2010).[13] T. Havermeier et al. , Phys. Rev. Lett. , 133401(2010).[14] M. Shcherbinin, A. LaForge, V. Sharma, M. Devetta,R. Richter, R. Moshammer, T. Pfeifer, and M. Mudrich,Phys. Rev. A , 013407 (2017).[15] V. Stumpf, N. Kryzhevoi, K. Gokhberg, and L. Ceder-baum, Phys. Rev. Lett. , 193001 (2014). [16] A. LaForge, V. Stumpf, K. Gokhberg, J. von Vangerow,F. Stienkemeier, N. Kryzhevoi, P. O’Keeffe, A. Ciavar-dini, S. Krishnan, M. Coreno, et al. , Phys. Rev. Lett. , 203001 (2016).[17] C. C. Wang et al. , J. Phys. Chem. A , 9356 (2008).[18] D. Buchta et al. , J. Phys. Chem. A , 4394 (2013).[19] F. Trinter et al. , Phys. Rev. Lett. , 233004 (2013).[20] M. Joppien, R. Karnbach, and T. M¨oller, Phys. Rev.Lett. , 2654 (1993).[21] O. Kornilov, O. B¨unermann, D. J. Haxton, S. R. Leone,D. M. Neumark, and O. Gessner, J. Phys. Chem. A ,7891 (2011).[22] M. P. Ziemkiewicz, C. Bacellar, K. R. Siefermann, S. R.Leone, D. M. Neumark, and O. Gessner, J. Chem. Phys. , 174306 (2014).[23] V. Averbukh and L. S. Cederbaum, Phys. Rev. Lett. ,053401 (2006).[24] S. Rousseau, A. Allouche, and M. Aubert-Fr´econ, J. Mol.Spectrosc. , 235 (2000).[25] J. Eland, Laser. Chem. , 259 (1991).[26] B. Dick, Phys. Chem. Chem. Phys. , 570 (2014).[27] J. Viefhaus, A. N. Grum-Grzhimailo, N. M. Kabachnik,and U. Becker, J. Electron Spectrosc. Relat. Phenom. , 121 (2004).[28] J. A. Samson, W. C. Stolte, Z.-X. He, J. N. Cutler, Y. Lu,and R. Bartlett, Phys. Rev. A , 1906 (1998).[29] T. Masuoka, Physical Review A , 3886 (1994).[30] P. Kolorenˇc, V. Averbukh, R. Feifel, and J. Eland, J.Phys. B: At., Mol. Opt. Phys. , 082001 (2016).[31] V. Stumpf, K. Gokhberg, and L. S. Cederbaum, Nat.Chem. , 237 (2016).[32] C. D. Jonah and B. M. Rao, Radiation chemistry: presentstatus and future trends , Vol. 87 (Elsevier, 2001).[33] P. O’Keeffe et al. , Rev. of Sci. Instrum. , 033109(2011).[34] J. P. Toennies and A. F. Vilesov, Angew. Chem. Int. Ed. , 2622 (2004).[35] T. Gough, M. Mengel, P. Rowntree, and G. Scoles, J.Chem. Phys. , 4958 (1985).[36] M. Barranco, R. Guardiola, S. Hern´andez, R. Mayol,J. Navarro, and M. Pi, J. Low Temp. Phys. , 1 (2006).[37] R. J. LeRoy and G. T. Kraemer, Chemical Physics Re-search Report No. CP-650R , University of Waterloo, University of Waterloo