Enhanced ionization of embedded clusters by Electron Transfer Mediated Decay in helium nanodroplets
A. C. LaForge, V. Stumpf, K. Gokhberg, J. von Vangerow, N. V. Kryzhevoi, P. O'Keeffe, A. Ciavardini, S. R. Krishnan, M. Coreno, K. C. Prince, R. Richter, R. Moshammer, T. Pfeifer, L. S. Cederbaum, F. Stienkemeier, M. Mudrich
EEnhanced ionization of embedded clusters by Electron Transfer Mediated Decay inhelium nanodroplets
A. C. LaForge, ∗ V. Stumpf, K. Gokhberg, J. von Vangerow, F. Stienkemeier, N. V. Kryzhevoi, P. O’Keeffe, A. Ciavardini, S. R. Krishnan, M. Coreno, K. C.Prince, R. Richter, R. Moshammer, T. Pfeifer, L. S. Cederbaum, and M. Mudrich Physikalisches Institut, Universit¨at Freiburg, 79104 Freiburg, Germany Physikalisch-Chemisches Institut, Universit¨at Heidelberg, 69120 Heidelberg, Germany CNR - Istituto di Struttura della Materia, CP10, 00016 Monterotondo Scalo, Italy Department of Physics, Indian Institute of Technology - Madras, Chennai 600 036, India CNR-Istituto di Struttura della Materia, CP10, 00016 Monterotondo Scalo, Italy Elettra-Sincrotrone Trieste, 34149 Basovizza, Trieste, Italy Max-Planck-Institut f¨ur Kernphysik, 69117 Heidelberg, Germany (Dated: November 8, 2018)We report the observation of electron transfer mediated decay (ETMD) involving magnesium(Mg) clusters embedded in helium (He) nanodroplets. ETMD is initiated by the ionization of Hefollowed by removal of two electrons from the Mg clusters of which one is transferred to the Heenvironment while the other electron is emitted into the continuum. The process is shown to bethe dominant ionization mechanism for embedded clusters for photon energies above the ionizationpotential of He. For Mg clusters larger than 5 atoms we observe stable doubly-ionized clusters.Thus, ETMD provides an efficient pathway to the formation of doubly-ionized cold species in dopednanodroplets.
The interplay between electrons after single photon ab-sorption has been a stimulating topic in atomic physicssince its foundation. Specifically, processes such as shake-off in single photon double ionization [1], post collisioninteraction in Auger processes [2], and autoionizationof doubly-excited states [3] have been a fertile groundfor studying electron correlation. Systems consisting ofmany weakly-interacting atoms or molecules additionallyoffer a unique environment for studying said correlationwhere new decay mechanisms, unavailable in atomic sys-tems, become accessible between the constituents. Inparticular, Cederbaum and coworkers [4] theoreticallypredicted a new decay mechanism, known as interatomicCoulombic decay (ICD), available in electronically ex-cited weakly bound systems. In the case where localelectronic decay is energetically forbidden, ICD offers anew ultrafast decay path typically on the femtosecondtimescale where energy is exchanged with a neighbor-ing atom leading to its ionization. Since its proposi-tion [4] and experimental confirmation [5, 6], ICD hasbeen observed in a wide variety of weakly-bound systemsincluding, for example, He dimers [7, 8] and water clus-ters [9, 10]. For reviews, see [11, 12].Electron transfer mediated decay (ETMD), theoret-ically predicted [13] and recently experimentally ob-served [14, 15], is another interatomic decay mechanismaccessible in weakly-bound systems. Charge transfer(CT) from the neighbor to the ion releases energy whichis utilized to either directly emit a second electron leadingto double ionization, ETMD(2), or the released energy istransferred to a second neighbor leading to a single ion-ization of the two neighbors, ETMD(3). Importantly,ETMD is a much stronger decay channel than its ra- diative counterpart. Like in ICD, the original versionof ETMD applies to ions which possess sufficient excessenergy to ionize their neighbors. In sharp contrast toboth cases, recently, a new variant of ETMD has beendisclosed which does not require excess energy and caneven be triggered from the ground state [16].Recently, Stumpf et al. [17] predicted ETMD to dra-matically enhance ( ∼ and neutral He. Sur-prisingly, due to ETMD, the theoretical cross section for double ionization of Mg is even higher than that of direct single ionization and is comparable to that of He. Over-all, the decay path and its predicted enhancement is notlimited to Mg in He clusters, but can be applied to anyembedded atoms or molecules which have a lower doubleionization potential than the single ionization potentialof the environment. Thus, the phenomenon is consideredto be of quite general relevance and can be used in Hedroplets as a new pathway to the formation of doubly-ionized cold species which are difficult to form otherwise.Here, we report on the first experimental observationof ETMD of particles embedded in superfluid He nan-odroplets. Following the initial ionization of a He atomwithin the droplet, ETMD leads to double ionization ofthe embedded Mg clusters. The electron kinetic energyspectra reveal a low energy ETMD peak at about 1 eVagreeing well with theory. The ETMD mechanism turnsout to be a dominant means to doubly ionize Mg clus-ters within the droplets allowing the investigation of thestability of doubly-ionized Mg clusters. a r X i v : . [ phy s i c s . c h e m - ph ] D ec He droplets have widely been used as a cold, weaklyperturbing matrices for studies in spectroscopy andchemical dynamics of embedded atoms and molecules [18,19]. While typically the He environment is inert to theembedded species, when the droplet is excited or ion-ized the situation is completely different and the dropletbecomes a highly reactive medium to the embeddedspecies [20]. Even doubly-ionized dopants have recentlybeen observed due to sequential collisions of metastableHe atoms produced in a single droplet by electron bom-bardment [21].The experiment was performed using a mobile Hedroplet machine attached to an imaging photoelectron-photoion coincidence (PEPICO) detector at the Gas-Phase beamline of Elettra-Sincrotrone Trieste, Italy. Thesetup has been described in some detail earlier [22, 23],and only the significant points will be addressed here. Inshort, a beam of He nanodroplets is produced by continu-ously expanding pressurized (50 bar), high purity He outof a cryogenic nozzle with 5 µ m diameter. Under theseexpansion conditions, the mean droplet sizes range from10 to 10 He atoms per droplet [18]. 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 [24] with an oven cell filled with Mg heated togenerate partial pressures where 1-25 Mg atoms were at-tached to the droplets. While most atomic and molecularspecies become submerged into the interior of He nan-odroplets, alkali earth atoms such as Mg remain weaklybound inside the surface layer [25]. The He droplet beamnext crosses the synchrotron beam at the focus of thePEPICO detector consisting of an ion time-of-flight de-tector and velocity map imaging detector operating incoincidence. With such a detection technique, one canrecord electron kinetic energy distributions in coincidencewith a specific ion masses in multicoincidence mode [22].The kinetic energy distributions were reconstructed us-ing a standard Abel inversion method [26]. The pho-ton energy was tuned by scanning the monochromatorand gap of the undulator simultaneously with a typicalstep size of 20 meV and energy resolution E/ ∆ E ≈ .The intensity of the radiation was monitored by a cali-brated photodiode and all photon energy dependent ionand electron spectra shown in this work are normalizedto this intensity signal.Fig. 1 shows the ion signal intensities of Mg + , Mg +2 ,Mg +3 , and He +2 as a function of photon energy. Thedroplet size is 5 000 He atoms with an average of 2-3Mg atoms attached. The photon energy was tuned from20 eV to 26 eV, which covers energies below the lowestdipole-allowed excitation energy (21.2 eV) to above theatomic ionization threshold (24.6 eV) of He droplets [27].The observed He ion signals are similar to previous syn-chrotron results [28, 29] where for energies below about23 eV (the adiabatic ionization potential of He droplets) ion intensity (arb. units) p h o t o n e n e r g y ( e V ) M g M g M g H e D r o p l e t s i z e : 5 , 0 0 0 a t o m sM g c l u s t e r s i z e : 3 a t o m s
Figure 1. Ion signal intensity of Mg (black), Mg (red), Mg (green), and He (blue) as a function of photon energy. Thedroplet size is 5 000 He atoms with an average of 2-3 Mg atomsattached. the signal is nearly zero. At energies higher than 23 eVbut below the ionization threshold of atomic He ( hν ≤ + orMg + n ion by a single photon. A single photon can alsodoubly ionize Mg n via ETMD [17]. The thus producedMg n can either fragment into two singly charged frag-ments or remain a stable dication.For the Mg ions below 21 eV photon energy, no signalwas observed; therefore, direct ionization of Mg is negligi-ble. Around 21.6 eV, there is a large peak in all three Mgion signals in Fig. 1 corresponding to ionization by Pen-ning or ICD processes [31, 32] as the He atoms are excitedto the droplet-equivalent of the 1s2p state [27]. At higherphoton energies, the Mg ion signals closely follow theHe +2 ion signal pointing at He-mediated ionization of Mg.Previous experiments with dopants (alkali metals) whichcannot undergo ETMD exhibit efficient dopant ionizationat the excitation energies of the droplet. However, ion-ization of the dopant was comparatively weak at higherphoton energies above the droplet’s ionization thresholdin contrast to the case of Mg presented here [23, 33]. This b ) mass-correlated electron intensity (arb. units) e K E ( e V ) M g M g +2 M g +3 h n = 4 0 e Va ) M g +3 M g +2 M g + ion intensity (rel. units) m a s s ( a m u ) h n ( e V ) 4 0 . 0 2 1 . 5M g D r o p l e t s i z e : 5 0 , 0 0 0 a t o m sM g c l u s t e r s i z e : 6 a t o m s
Figure 2. a) Mass spectra of He droplets doped with Mgclusters at photon energies of 40 eV (black) and 21.5 eV (red).The droplet size is 50 000 He atoms with an average of 5-6Mg atoms attached. b) Electron kinetic energy distributionsmeasured in coincidence with single ions for Mg n =7 , , (blackline), Mg +2 (red line), and Mg +3 (blue line). is surprising considering that the cross section for reso-nant excitation of He [28] is three times higher than theionization cross section near threshold [34] and that Mgis located close to the droplet’s surface similar to alklalimetals [35]. The question which arises is whether thestrong enhancement is due to ETMD.To address the differences in the ionization mechanismsdiscussed above, we show in Fig. 2 a) the mass spectrafor photon energies of 40 eV (black line) and 21.5 eV (redline) for droplets consisting of 50 000 He atoms with anaverage of 5-6 Mg atoms attached. For both energies, alarge contribution of Mg ions is observed in the massspectra, and, similar to Fig. 1, there are substantiallyhigher signals above the ionization threshold(h ν = 40eV). Here, Mg + -He complexes are observed at multiplesof 4 amu in the mass spectra following multiples of the Mg mass (24 amu).Importantly, at higher masses, broad peaks at half-integer values of the mass/charge ratio appear. Theseare due to the formation of doubly-ionized Mg clusterswith at least 5 atoms. The stability of doubly-ionizedclusters has previously been studied [36, 37] and it wasexperimentally shown that Mg clusters consisting of 5atoms or more are sufficiently long-lived to be detectedin a mass spectrometer. The signals in the mass spec-trum corresponding to integer numbers of Mg atoms inMg + n may, of course also be due to Mg n . However, it isimpossible to disentangle them from singly ionized clus-ters. The observation of doubly-ionized Mg clusters givesthe first direct evidence of ETMD for this system.In order to identify the various ionization mecha-nisms, mass-correlated electron spectra are shown inFig. 2 b) for Mg n (black line). As there was no dif-ference between the electron spectra correlated to thevarious doubly-ionized clusters, they were combined toincrease statistics. At 15.4 eV, one observes a large pho-toelectron peak resulting from the initial ionization of He( hν – E i (He) = 15.4 eV). The Mg n peak at low energy isdue to ETMD as the electrons emitted in this process liein the observed energy range (see supplementary materialand Ref. [17]) and there are no other mechanisms whichproduce an electron peak in this energy range. We shallargue below that ETMD is by far the dominant mecha-nism for producing doubly-ionized Mg clusters. Shownin Fig. 2 b) are also the mass-correlated electron spectrafor the Mg +2 (red line) and Mg +3 (blue line) ions. Sur-prisingly, they are similar to that of Mg n , exhibiting aphotoelectron peak at 15.4 eV and a low-energy peak atabout 1 eV. Therefore, not only is ETMD responsible forthe doubly-ionized clusters in the mass spectra but couldbe a primary mechanism for the production of singlycharged smaller clusters, which result upon fragmenta-tion of doubly-ionized unstable clusters. There is a slightdiscrepancy between the spectra for Mg n and those ofsmaller clusters, Mg +2 and Mg +3 ; namely, the ETMD peakfor Mg n extends to higher energies. The additional en-ergy for Mg n clusters is due to the clusters not under-going dissociation which requires additional energy. Thesupplementary material contains mass and electron spec-tra similar to those shown in Fig. 2 for 5 000 He atoms,but with an average of 2-3 Mg atoms attached. In thiscase, Mg can clearly be identified in the mass spectra.Fig. 3 a) shows the ion-ion coincidence time-of-flightspectrum for He droplets consisting of 50 000 atomsdoped with an average of 5-6 Mg atoms. The photon en-ergy was 40 eV. As the flight times of the respective ionsare symmetric, the coincidence map is folded along theaxis of symmetry. The peak shapes observed in the spec-tra give information about the dissociation process [38].Additionally, due to the dense He environment, the ki-netic energy of Coulomb-exploded molecules embeddedin He droplets is significantly damped. The coincidence M g +3 - M g +5 M g +2 - M g +4 M g +1 - M g +3 M g +2 - M g +3 ion 2 time-of-flight ( m s) i o n 1 t i m e - o f - f l i g h t ( m s )M g +1 - M g +2 a ) C o u l o m b - e x p l o d e d c l u s t e r s s i n g l e e m b e d d e d M g a t o m electron intensity (arb. untis) e K E ( e V )b )
Figure 3. a) Ion-ion coincidence time-of-flight spectrum andb) electron spectrum correlated to Mg + m -Mg + n ion-ion coinci-dences (black line). The photon energy is 40 eV. The dropletsize is 50 000 He atoms with an average of six Mg atoms at-tached. electron spectra correlated to a single Mg atom (redline) embedded in a droplet of size 5 000 atoms. map is centered around Coulomb-exploded Mg oligomers.Due to the dead time of the detector ( ≈ + m - − Mg + n where n (cid:54) = m . Overall, the coinci-dence map reveals a rich spectrum of doubly-ionized clus-ters up to Mg showing that many of the singly ionizedMg ions seen in Fig. 2 a) stem from larger doubly-ionizedclusters.Fig. 3 b) shows the electron spectra correlated to ion-ion coincidences (black line). Since there was no sub-stantial change (e.g. peak position and width or relativeratio of the ETMD to photoelectron peak) for the vari-ous Mg ion-ion pairs, the individual electron spectra weresummed for all large heterogeneous ion-ion pairs shown in the coincidence map in Fig. 3 a). Similar to the electronspectra shown in Fig. 2 b), the ETMD peak is also cen-tered at 0.9 eV. Here, the ratio of the integrated ETMDpeak to the photoelectron peak is 81% in this case, sig-nificantly higher than that shown in Fig. 2 b) (35-50%depending on the correlated ion) consistent that doubly-ionized clusters are a result of ETMD. The most likelyreason that the ETMD to photoelectron peak ratio isless than unity is due to the large background signal offalse coincidences from ionized He. Besides the photo-electron and ETMD peaks, there are no electron signalsfrom other mechanisms, which highlights that ETMD isthe dominant process for double ionization of Mg clus-ters. The ion-ion coincidence map and electron spectrafor He droplets consisting of 5 000 atoms doped with onaverage 2-3 Mg atoms are given in the supplementarymaterial and give consistent results with those in Fig.3.So far, we have solely focused on Mg clusters embed-ded in He nanodroplets where we have shown that theHe environment dramatically enhances the double ion-ization of the cluster due to ETMD. Finally, we addressthe enhancement of the double ionization efficiency ofa single Mg atom with a single He atom as a neighboras investigated by Stumpf et al. [17]. We observed nodoubly-ionized Mg atoms when only a single Mg atomis embedded in the nanodroplet; see supplementary ma-terial. The Mg + -correlated electron spectra for a singleMg atom embedded in a droplet consisting of 5 000 atomsis shown in Fig. 3 b). In this case, the ETMD electronpeak was absent suggesting that ionization proceeds ex-clusively through CT. Thus, ETMD appears to be inac-tive for single Mg atoms attached to He droplets.We attribute this to the ultrafast formation of He +2 in the droplet, which is predicted to occur in 60-80 fsafter the initial ionization [23]. Our calculations showclearly that the ETMD channel for a single Mg atomis closed once He +2 is formed. Additionally included inthe supplementary material are theoretical calculationsof the ETMD electron kinetic energy for the most likelyconfigurations between He +2 and Mg trimer where ETMDcan still occur. It is yet inconclusive whether ETMD canoccur for Mg dimers.In conclusion, electron transfer mediated decay was ob-served for Mg clusters embedded in He droplets. This de-cay channel was shown to be a dominant ionization mech-anism for energies above the ionization threshold of He.For clusters of 5 Mg atoms and greater, stable, doubly-ionized Mg clusters were observed after ETMD. For sin-gle Mg atoms embedded in nanodroplets, the ETMDchannel is closed due to the ultrafast formation of anequilibrated He dimer ion. In general, ETMD offers anovel method for producing doubly-ionized systems if theenvironment has a higher single ionization threshold. ∗ [email protected][1] T. Pattard, T. Schneider, and J. Rost, J. Phys. B: At.,Mol. Opt. Phys. , L189 (2003).[2] A. Russek and W. Mehlhorn, J. Phys. B: At., Mol. Opt.Phys. , 911 (1986).[3] R. Madden and K. Codling, Phys. Rev. Lett. , 516(1963).[4] L. S. Cederbaum, J. Zobeley, and F. Tarantelli, Phys.Rev. Lett. , 4778 (1997).[5] S. Marburger, O. Kugeler, U. Hergenhahn, andT. M¨oller, Phys. Rev. Lett. , 203401 (2003).[6] T. Jahnke et al. , Phys. Rev. Lett. , 163401 (2004).[7] N. Sisourat et al. , Nat. Phys. , 508 (2010).[8] T. Havermeier et al. , Phys. Rev. Lett. , 133401(2010).[9] T. Jahnke et al. , Nat. Phys. , 139 (2010).[10] M. Mucke et al. , Nat. Phys. , 143 (2010).[11] U. Hergenhahn, J. Electron. Spectrosc. Relat. Phenom. , 78 (2011).[12] T. Jahnke, J. Phys. B: At., Mol. Opt. Phys. , 082001(2015).[13] J. Zobeley, R. Santra, and L. S. Cederbaum, J. Chem.Phys. , 5076 (2001).[14] K. Sakai et al. , Phys. Rev. Lett. , 033401 (2011).[15] M. F¨orstel, M. Mucke, T. Arion, A. M. Bradshaw, andU. Hergenhahn, Phys. Rev. Lett. , 033402 (2011).[16] V. Stumpf, P. Kolorenˇc, K. Gokhberg, and L. S. Ceder-baum, Phys. Rev. Lett. , 258302 (2013).[17] V. Stumpf, N. Kryzhevoi, K. Gokhberg, and L. Ceder-baum, Phys. Rev. Lett. , 193001 (2014).[18] J. P. Toennies and A. F. Vilesov, Angew. Chem. Int. Ed. , 2622 (2004).[19] F. Stienkemeier and K. K. Lehmann, J. Phys. B: At.,Mol. Opt. Phys. , R127 (2006). [20] M. Mudrich and F. Stienkemeier, Int. Rev. Phys. Chem. , 301 (2014).[21] H. Sch¨obel et al. , Phys. Rev. Lett. , 243402 (2010).[22] P. O’Keeffe et al. , Rev. of Sci. Instrum. , 033109(2011).[23] D. Buchta et al. , J. Phys. Chem. A , 4394 (2013).[24] T. Gough, M. Mengel, P. Rowntree, and G. Scoles, J.Chem. Phys. , 4958 (1985).[25] M. Barranco, R. Guardiola, S. Hern´andez, R. Mayol,J. Navarro, and M. Pi, J. Low Temp. Phys. , 1 (2006).[26] G. A. Garcia, L. Nahon, and I. Powis, Rev. of Sci. In-strum. , 4989 (2004).[27] M. Joppien, R. Karnbach, and T. M¨oller, Phys. Rev.Lett. , 2654 (1993).[28] D. Buchta et al. , J. Chem. Phys. , 084301 (2013).[29] R. Fr¨ochtenicht, U. Henne, J. P. Toennies, A. Ding,M. Fieber-Erdmann, and T. Drewello, J. Chem. Phys. , 2548 (1996).[30] A. Kramida, Yu. Ralchenko, J. Reader, and and NISTASD Team, NIST Atomic Spectra Database (ver. 5.2),[Online]. Available: http://physics.nist.gov/asd [2015, July 14]. National Institute of Standards and Tech-nology, Gaithersburg, MD. (2014).[31] C. C. Wang et al. , J. Phys. Chem. A , 9356 (2008).[32] F. Trinter et al. , Phys. Rev. Lett. , 233004 (2013).[33] A. A. Scheidemann, V. V. Kresin, and H. Hess, J. Chem.Phys. , 2839 (1997).[34] G. Marr and J. West, At. Data Nucl. Data Tables ,497 (1976).[35] Y. Ren and V. V. Kresin, Phys. Rev. A , 043204(2007).[36] F. Reuse, S. N. Khanna, V. de Coulon, and J. Buttet,Phys. Rev. B , 11743 (1990).[37] T. Diederich, T. D¨oppner, T. Fennel, J. Tiggesb¨aumker,and K.-H. Meiwes-Broer, Phys. Rev. A , 023203(2005).[38] J. Eland, Laser. Chem.11