Time-resolved formation of excited atomic and molecular states in XUV-induced nanoplasmas in ammonia clusters
Rupert Michiels, Aaron Cristopher LaForge, Matthias Bohlen, Carlo Callegari, Andrew Clark, Aaron von Conta, Marcello Coreno, Michele Di Fraia, Marcel Drabbels, Paola Finetti, Martin Huppert, Veronica Oliver Álvarez de Lara, Oksana Plekan, Kevin Charles Prince, Stefano Stranges, Vit Svoboda, Hans Jakob Wörner, Frank Stienkemeier
TTime-resolved formation of excited atomic and molecular states in XUV-inducednanoplasmas in ammonia clusters
R. Michiels, ∗ A. C. LaForge,
1, 2
M. Bohlen, C. Callegari, A. Clark, A. von Conta, M. Coreno, M. Di Fraia, M. Drabbels, P. Finetti, M. Huppert, V. Oliver, O.Plekan, K. C. Prince, S. Stranges, V. Svoboda, H. J. W¨orner, and F. Stienkemeier Institute of Physics, University of Freiburg, 79104 Freiburg, Germany Department of Physics, University of Connecticut, Storrs, Connecticut, 06269, USA Elettra-Sincrotrone Trieste, 34149 Basovizza, Trieste, Italy Laboratory of Molecular Nanodynamics, Ecole Polytechnique F´ed´erale de Lausanne, 1015 Lausanne, Switzerland Laboratorium f¨ur Physikalische Chemie, ETH Z¨urich, 8093 Z¨urich, Switzerland ISM-CNR, Istituto di Struttura della Materia, LD2 Unit, 34149 Trieste, Italy Department of Chemistry and Drug Technologies, University Sapienza,00185 Rome, Italy, and Tasc IOM-CNR, Basovizza, Trieste, Italy (Dated: March 13, 2020)High intensity XUV radiation from a free-electron (FEL) was used to create a nanoplasma insideammonia clusters with the intent of studying the resulting electron-ion interactions and their in-terplay with plasma evolution. In a plasma-like state, electrons with kinetic energy lower than thelocal collective Coulomb potential of the positive ionic core are trapped in the cluster and take partin secondary processes (e.g. electron-impact excitation/ionization and electron-ion recombination)which lead to subsequent excited and neutral molecular fragmentation. Using a time-delayed UVlaser, the dynamics of the excited atomic and molecular states are probed from -0.1 ps to 18 ps. Weidentify three different phases of molecular fragmentation that are clearly distinguished by the effectof the probe laser on the ionic and electronic yield. We propose a simple model to rationalize ourdata and further identify two separate channels leading to the formation of excited hydrogen.
I. INTRODUCTION
Through the absorption of intense radiation, plasma-like states can be formed in condensed systems withina few femtoseconds. In particular, the study of laser-induced nanoplasmas in rare gas clusters has attractedconsiderable interest [1, 2] over a wide range of wave-lengths, from strong-field ionization in the infraredto multiphoton ionization in the X-ray. Specificallyfor short wavelength radiation, the mechanism ofnanoplasma formation is as follows: first, electronemission is due to direct single photon absorption untila positive ionic core is built up within the cluster.Afterwards, electrons are ionized from individual atomsin the cluster, but are trapped by the surroundingelectrostatic cluster potential leading to frustrationin the emission process and a plasma-like state beingformed. [3–7] At later times, the ions in the nanoplasmamove apart under the combined action of hydrodynamicforces and of Coulomb repulsion. Previous research onnanoplasmas, focusing primarily on rare gas clusters,has shown that frustration of electron emission canindeed be achieved with intense XUV pulses, and thatinelastic collisions and recombination processes play animportant role in the nanoplasma evolution. [8–15]For molecular clusters, as compared to their atomiccounterparts, the presence of both intra- and intermolec-ular bonds complicate the nanoplasma dynamics andthe underlying fragmentation processes. Specifically, the ejection of lightweight hydrogen in molecular clusters of-fers an efficient pathway for both cooling and charge dis-sipation. [16–18] The study of highly ionized molecularclusters has previously been investigated using intenseradiation from a tabletop laser [19–22] and a FEL, [23]but, so far, those results fail to resolve the temporal evo-lution of the underlying mechanisms and, therefore, littleis known about the ultrafast nanoplasma dynamics. Wereport on time-resolved measurements on ammonia clus-ters using intense, femtosecond XUV FEL pulses to cre-ate a highly-ionized, plasma-like state. Using a 266 nmUV laser, we probed the evolution of the nanoplasmawith a specific focus on the atomic and molecular ex-cited states. In the ion and electron yields, we observethe formation of excited hydrogen as a major contribu-tion (Fig. 2). From the time-resolved data we can identifytwo dominant pathways for the creation of excited hydro-gen. These can be created either via direct recombinationof ions and electrons, or by dissociation of highly excitedmolecules.
II. EXPERIMENTAL SETUP AND METHODS
The experiment was performed at the Low DensityMatter (LDM) endstation [24] at the seeded FELFERMI in Trieste, Italy. [25, 26] The photon energywas set to hν = 24.0 eV and the pulse length wasapproximately 70 fs FWHM. The FEL pulse energy at a r X i v : . [ phy s i c s . a t m - c l u s ] M a r - 1 . 0 - 0 . 5 0 . 0 0 . 5 1 . 0 1 . 5 2 . 0 2 . 5 3 . 0 3 . 5 4 . 0 4 . 5 5 . 00 . 00 . 20 . 40 . 60 . 81 . 0 - 1 . 0 - 0 . 5 0 . 0 0 . 5 1 . 0 1 . 5 2 . 0 2 . 5 3 . 0 3 . 5 4 . 0 4 . 5 5 . 00 . 00 . 20 . 40 . 60 . 81 . 0 U V32 P r o b e H + i o n s Ion-yield (arb. units) s t p h a s e 3 r d p h a s e t (cid:1) t S u m o f p r o b e e - Intensity (arb. units)
P u m p - p r o b e d e l a y i n p s2 n d p h a s e
Figure 1. Top: Schematic of the cluster explosion illustratingthe three phases. Middle: Probe yield of H + ions with vary-ing pump-probe delay. The shaded areas indicate the threephases of the cluster explosion as discussed in the text: 1)Direct pulse overlap phase. 2) Recombination and proton-radiation phase. 3) Molecular dissociation phase. Bottom:Absolute sum of the probe photoelectron yield. the target was approximately 70 µ J, calculated fromthe value measured upstream on a shot-by-shot basisby gas ionization and the nominal reflectivity of theoptical elements in the beam transport system. TheFEL photon energy of 24.0 eV was chosen in order tomaximize the product of photon flux and ionization crosssection within the technical limitations posed by theFEL. The FEL spot size in the interaction region was20 µ m FWHM. The pump-probe scheme was realizedusing a UV pulse produced from a frequency-tripledTi:Sapphire laser ( hν = 4.75 eV) with a pulse energy of about 45 µ J and a spot size of 80 µ m FWHM. Thecross-correlation between the two pulses, measured byresonant two photon ionization of helium, was about200 fs FWHM. A supersonic jet of ammonia clusterswas produced by expansion of pressurized ammoniathrough a custom, pulsed nozzle. By varying thebacking pressure of the ammonia, we controlled themean cluster size in the range of 10 – 3 × ammoniamolecules. For this experiment, a mean cluster size of2000 molecules was chosen giving the best signal. Themean cluster sizes where determined using a modifiedscaling law of the Hagena type. [27] The cluster beamwas perpendicularly crossed by the FEL/UV beam atthe focus of a velocity-map-imaging (VMI) spectrometer,and of an ion time-of-flight mass spectrometer. [24] Theelectron kinetic energy and angular distributions werereconstructed from velocity-map images using the Maxi-mum Entropy Legendre Reconstruction method. [28] Toemphasize the prevalence of cluster processes as opposedto processes involving isolated molecules we compare ourresults with the interaction of XUV light with moleculargas phase ammonia. These interactions are well studiedexperimentally and theoretically and the partial ioniza-tion cross sections leading to different ionic fragmentsare known from experiment [29] and calculations. [30–32] III. RESULTS AND DISCUSSIONA. Time-resolved probe ion yields
Irradiating a gas jet of molecular ammonia employingthe above-stated FEL pulse parameters results in asample ionization rate that is near saturation ( > + ions since it is known they playa strong role in the evolutionary behavior of molecularclusters. [16] The probe yield of H + ions as a function ofpump-probe delay is shown in Fig. 1. The proton yieldfor t < t is close to zero, since the 266 nm pulse, due tolow intensity, cannot directly create a significant fractionof H + ions. In the pulse overlap, around t = 0, the probeyield drastically increases, and the observed half-widthresembles the cross correlation of the two laser pulses,shown in the supplementary information. This is thefirst phase where the cluster is being multiply ionizedleading to the formation of a nanoplasma. As sequentialionization proceeds, quasi-free electrons are trapped bythe positive cluster potential leading to a frustrationin ionization, a defining property of XUV-inducednanoplasmas. [4–7] To characterize this behavior, weintroduce the frustration parameter, which can beestimated as α = N tot /q full , the ratio of total ionizationevents to the number of primary ionizations prior tofull frustration q full = ( (cid:126) ω − I p ) r s N / / (1 .
44 eV ˚A) ( r s the Wigner-Seitz radius) for a spherical homogeneouslycharged cluster. [5] The frustration parameter for thisexperiment is α >
15. Since the frustration parameteris much larger than one, electrons are trapped insidethe cluster potential and hence recombination with theions is strongly enhanced. After ionization frustrationsets in, the second phase begins where the H + probeyield decreases linearly reaching a minimum value atroughly 700 fs pump-probe delay. We attribute thedecrease to the quick decay of excited molecular stateswhich are populated by electron-ion recombinationand molecular formation within the cluster. A similarbehavior was observed for the fragmentation of smallexcited ammonia clusters and attributed to the lifetimeof the excited ˜A state of ammonia. [33] Furthermore,charge equalization takes place during the second phase,as protons leave the cluster and carry away the excesscharge. Electrons that have not recombined with ionswill successively leave the cluster as low kinetic energyelectrons once the local potential barrier is lower thanthe kinetic energy of the electrons. Eventually, there islittle excess charge in the cluster and the third phasefollows. In phase three the H + probe yield shown inFig. 1 increases, converging asymptotically to a higherfinal value than the intermediate maximum. The thirdphase is dominated by relaxation and dissociationprocesses of excited free molecules, atoms and clusterfragments. Complimentary to the H + ion yields, thesum of the probe electron yield, in the bottom panel ofFig. 1, shows identical behavior further prompting ourthree phase model interpretation. Additionally, similarbehavior is observed in many other ion yields, which areshown in Fig. 4 and in the supplementary information † . B. Time-resolved probe photoelectrons
Next, we focus on the UV-ionized probe electronspectrum (UV-PES) at a time delay where the ultrafastelectronic and nuclear dynamics have ended. TheUV-PES, shown in the top panel of Fig. 2 containsonly the probe photoelectrons which were extracted bysubtracting an XUV only PES from a XUV+UV PESat a pump-probe delay of 18 ps. The prominent featureat 1 . ± .
05 eV has a narrow width of 0 . ± .
05 eVFWHM, primarily due to the experimental resolution.From the vertical binding energy, and comparison withour previous experiments, we know that these electronscome from excited hydrogen in the n = 2 state. [34] Wealso observe higher lying excited states of hydrogen, with
Pump-probe delay in ps
E l e c t r o n k i n e t i c e n e r g y ( e V )
I n t e n s i t y
H * ( n = 2 ) D S P N * c o n t i n u u m
E l e c t r o n k i n e t i c e n e r g y ( e V )
Intensity
H * ( n = 3 ) H * ( n = 4 )
H * ( n = 5 )
N * 2 s ² 2 p ² ( ³ P ) 3 p P N * 2 s ² 2 p ² ( D ) 3 s
Figure 2. Top: Probe photoelectron intensities for ammo-nia clusters are shown versus electron kinetic energy at anXUV-UV pump-probe delay of 18 ps and 24.0 eV FEL pho-ton energy. The contributions from excited hydrogen statesand their expected binding energies are marked in blue andthe excited nitrogen states are marked in red. Bottom: Colorplot of the probe photoelectron intensities with pump-probedelay on the y-axis and electron kinetic energy on the x-axis.Dashed black lines are used to highlight the first and secondformation channel of H ∗ . kinetic energy > E coll ). Taking the ionization cross section forthe 266 nm probe into account, we expect the relativeobserved photoelectron intensities for excited hydrogenwith principal quantum number n = 2/n = 3/n = 4 tobe 1/0.24/0.08 when E Coll = 0 and 1/0.11/0.02 when E Coll = 15 eV. The relative intensities in our experimentare 1/0.18 ± ± † , which is dominated by protons(86%) and nitrogen ions (7%). This suggests that amajor part of the remaining photoelectrons comes fromthe excited states of atomic nitrogen. A contributionfrom solvated electrons as observed in [34] would beobserved at 3 eV electron kinetic energy, which stronglyoverlaps with photoelectrons from excited hydrogen andnitrogen. For this reason, a contribution from solvatedelectrons can neither be excluded nor included.Complementary to the ion yields, shown in Fig. 1,the time-resolved UV-PES gives information about theevolution of the respective electronic binding potentials.Horizontal projections of the lower panel of Fig. 2 fordifferent pump-probe delay are shown in separate panelsin Fig. 3. Initially, within the first few hundreds offemtoseconds, the UV-PES is dominated by a largepeak of low kinetic energy electrons with a broaddistribution extending to higher energies; no additionalpeaks are observed since the cluster potential screensthe binding energies of the ionized states. The observedfeatures reach maximum intensity at 150 fs pump-probedelay, coinciding with the end of phase one. Around apump-probe delay of 500 fs, shown as red curve in themiddle panel of Fig. 3, we observe the first well-definedpeak from the n = 2 state of excited hydrogen. The peakis slightly offset from its nominal value by 300 meV. Inthe following 500 fs the peak converges to the kineticenergy precisely matching the vertical binding energyof H ∗ (n = 2) (blue curve). In addition to initially beingoffset, the peak is also broadened; both of these featuresindicate the influence of the local cluster potentialon the atomic hydrogen binding energy. As the peakemerges immediately with the decaying nanoplasma,it seems reasonable to assign this first channel to adirect recombination of protons and electrons in thenanoplasma. In the further evolution of the UV-PESin the bottom panel of Fig. 2, we can see a new pho-toelectron line appear at roughly 1.2 ps pump-probedelay. The kinetic energy of these electrons is at first < ∗ (n = 2) photoelectron peak as shown in theblack and red curve in the right panel of Fig. 3. In thefollowing picoseconds, the new peak gains in intensityand converges to the expected kinetic energy for excitedhydrogen photoelectrons. From the clear differences inthe timescale, we conclude that two separate processesforming excited hydrogen are observed. C. Formation of excited hydrogen by moleculardissociation
In this section, we discuss the second contribution ofthe observed excited hydrogen using rigorous analysis of the photoelectron and ion time-of-flight spectra. A firststep will be to identify possible precursor molecules orcluster fragments from which the H ∗ dissociates. Thetime-resolved probe ion yields of NH +4 , the dimer, andthe larger cluster fragments are shown in Fig. S4 in thesupplementary information † . Despite the considerabletemporal evolution in the first and second phase of thenanoplasma, the probe ion yield of all larger fragmentsstays constant in the third phase. From this observationwe conclude that the precursor molecule is not a clusterfragment (NH ) n H m with n > m any integer.The next important observation is that the convergenceof the photoelectron kinetic energy, as illustrated bydashed black lines in Fig. 2, evolves from lower kineticenergy towards higher kinetic energy. We estimatethat the first clear observation of photoelectrons fromthe second channel is at 1.4 ps pump-probe delay, witha mean electron kinetic energy of 400 meV. This isshifted by 900 meV compared to the final convergencevalue of 1.3 eV. Most neutrally dissociating potentialenergy surfaces of molecules evolve in the directionof higher electron binding energy. [37–39] The risingkinetic energy observed in the experimental spectrumindicates that the ionization energy is dropping duringthe dissociation. Since the observed change is too largeto be explained by dissociation of any neutral ammoniafragment, we deduce that the precursor is a positivelycharged molecule. In conclusion, the precursor is amolecule which dissociates into unknown fragments plusH ∗ , with at least one of the unknown fragments beingionic.With these presumptions, we next identify fragmentsthat show an increase in yield during the third phase.This is the case for H + (Fig. 3), N ++ , NH +2 (Fig. 4),and N + (Fig. S5 † ). Of these three, only NH +2 has anoverall negative probe yield, converging to zero for longpump-probe delays. As such, the NH +2 yield reproducesexactly the dynamic expected for a product producedby the undisturbed dissociation, i.e. without the probelaser. This observation restricts us to two possible parentmolecules, NH + ∗ and NH + ∗ , since we have H ∗ + NH +2 asproducts. Complementary, we look for ions whose probeyield reduces in phase three. These ions are created bythe UV laser lifting the excited molecule onto a differentdissociative potential energy surface and/or ionizing itprior to dissociation. Only two ions show a decrease inprobe ion yield in phase three. These two ions are H +2 and NH + , shown in Fig. 5. We notice that they matchthe products NH +2 and H ∗ in numbers of H and N atoms.We thus propose the dissociation channel: N H + ∗ → H ∗ + N H +2 (1)dissociating with a half-life time of τ = (1.6 ± < t <
10 ps, the UV laser
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 ) 0 . 2 p s p u m p - p r o b e d e l a y 0 . 0 p s- P h a s e 1 - 0 1 2 3 40 . 00 . 20 . 40 . 60 . 81 . 0 - P h a s e 2 / 3 - 1 . 0 p s 0 . 5 p s 0 . 3 5 p s
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 )H * ( n = 2 ) 0 1 2 3 40 . 00 . 20 . 40 . 60 . 81 . 0 - P h a s e 3 - H * ( n = 4 / 5 )H * ( n = 3 )H * ( n = 2 ) 1 3 . 0 p s 2 . 5 p s 1 . 5 p s
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 )
Figure 3. Selected cuts from the probe photoelectron yield. Left Panel: Probe photoelectrons for zero and 200 fs pump-probedelay. Middle Panel: Pump-probe delay from 350 fs to 1 ps. Right Panel: Pump-probe delay > interacting with the precursor molecule leads to differentions, partially governed by the reaction: N H + ∗ UV −−→ N H + + H +2 + e − (2)This interpretation is consistent with the observed ionyields. Nonetheless, it is tentative, since in the observedsystem it is hardly possible to exclude all alternatives.Please note, that all of the above discussion is also validif additional hydrogens are ejected as neutral fragments.In our experiment only charged fragments are detected,thus a differentiation of the possible parents NH + ∗ from NH + ∗ is not possible. Furthermore, the absolutechanges in ion yield clearly show that Eq. 2 is not theonly possible fragmentation channel when ionizing NH + ∗ prior to dissociation. This is not surprising, since amolecule with so much excess energy is very likely tohave multiple dissociation pathways. A similar channelwith NH + ∗ as parent molecule, i.e. NH + ∗ → H ∗ + N H + is possible and cannot be distinguished through the ionyields.The dissociation of ammonia cations has beenextensively studied using photoionization, ab-initio cal-culations, [29, 40–42] as well as single-electron captureemploying PEPIPICO. [43] From these studies, theenergetic dissociation threshold for NH + ∗ → NH +2 +H is known to be 15.76 eV with respect to the groundstate of ammonia, and is associated with the ˜A Eexcited state of NH + ∗ . In a very simple approximation,adding 10.2 eV excitation energy to reach H ∗ (n = 2),one determines a minimum energy of 26 eV requiredfor the dissociation NH + ∗ → NH +2 + H ∗ (n = 2). Thedouble ionization threshold of ammonia is 33.7 eV [44]and the ˜B excited state of the ammonia cation can bereached when ionizing the 2a valence shell of ammoniawith 27.7 eV. [45] As such, the dissociation yieldingH ∗ is energetically possible after electron capture byNH ++3 and from the ˜B excited state. Nonetheless, the - 1 0 1 2 3 4 5- 3- 2- 101 - 1 0 1 2 3 4 5 - 6- 5- 4- 3- 2- 101 P r o b e H +2 y i e l d Intensity (arb.) t = 700 f s a ) - 1 0 1 2 3 4 501234 P r o b e N H + y i e l d Intensity (arb.)
P u m p - p r o b e d e l a y i n p sb )
Intensity (arb.)
P r o b e N H +2 y i e l d t = 700 f s c ) - 1 0 1 2 3 4 5 0 . 00 . 40 . 81 . 21 . 6 P r o b e N + + y i e l d
Intensity (arb.)
P u m p - p r o b e d e l a y i n p sd )
Figure 4. Probe ion yields versus pump-probe delay for H +2 (a), NH + (b), NH +2 (c), and N ++ (d). The blue vertical lineat 700 fs pump-probe delay marks the transition between thesecond and third phase of the nanoplasma expansion. possibility of a fragmentation channel producing excitedhydrogen is not discussed in the literature, even thoughit is generally known that excited states of the aminoradical fragment are produced. [46] In Fig. 5, the mostimportant states are shown in a highly simplified statescheme with the y-axis reflecting on the thermodynamicenergy threshold for certain products. The proposeddissociation channel of NH + ∗ is marked as dotted redlines, with dotted blue lines representing the knownchannels. We can see that after electron-capture from + H + H + B EX A ''2 B A E N H + +3 HH * N H N H + H +2 N H + N H +2 N H + + *3
N H + + *3 X A N H +3 N H + *3
N H + *3
N H +2 N H N H +2 energy (eV) H +2 N HU Ve - c a p t u r e H x N + + Figure 5. Schematic of the known excited states of NH +3 andNH ++3 , and their fragmentation channels (blue dotted lines)from [40, 43, 44, 47]. For all shown states, the y-axis positionreflects their thermodynamic energy threshold with respectto the NH ground state. The red dotted line symbolizes theproposed new decay channel yielding excited hydrogen. Blueand red arrows show the electron capture from the doublyionized state and the re-pumping with the UV probe laser.Please notice that the shown initial and final states are notcomplete and do not intend to show all possible state evolu-tions. NH ++3 , dissociation into NH +2 and H ∗ is energeticallypossible. Absorbing one or two photons from the UVprobe laser, the dissociative energy surfaces coming fromthe doubly ionized ammonia can be reached. To explainour data, it is crucial that some of these channels donot create H + in their dissociation, as is the case whendissociating into NH + + H +2 . The state scheme in Fig. 5shows only the most important contributors and isneglecting, for example, doubly excited states, and manyof the known dissociative recombination channels. [46]From the state scheme, our interpretation of excitedhydrogen forming via dissociative recombination seemsjustified. Nonetheless, the overall high intensity in thechannel is surprising. We assume that many of the dis-sociation channels governing the ion dynamics could notbe identified in this work. It is known that dissociativerecombination of nitrogen containing molecules ofteninvolves excited and ground-state neutrals [46] and itcan be safely assumed that this is also possible for thevery high lying states of NH ++ ∗ . IV. CONCLUSIONS
In this experiment, we have used an intense, XUVpulse to create a nanoplasma inside ammonia clusters.Using a time-delayed UV pulse, we probed the resultingexcited states of molecules and atoms, thereby gaininginformation on their temporal evolution. We usedsimultaneous ion time-of-flight and electron velocity-map-imaging detection to map the formation of excitedstates and detect all charged fragments. Surprisingly,despite many possible secondary processes, the dominantfragment channels follow universal principles and wereidentified. The dynamics are governed by sequentialone-photon ionization, electron-ion recombination, andelectron-impact excitation creating various charged andneutral fragments. We characterize the cluster evolutionafter XUV irradiation as a three phase process. Thefirst phase is dominated by rising cluster potentialwith fast electrons being trapped inside the ionic core.This is followed by the second phase, during whichfast protons leave the cluster resulting in a drop of theelectronic potential and detection of the UV-ionizedelectrons. In the third and final phase, the UV laserprobes the molecular dissociation dynamics of excitedand neutral cluster fragments. We observe two differentformation channels for excited hydrogen. The firstprocess is assigned to recombination of electrons andprotons in the nanoplasma, whilst the second processis interpreted as dissociative recombination of doublycharged ammonia molecules with electrons.
CONFLICTS OF INTEREST “There are no conflicts to declare”.
ACKNOWLEDGEMENTS
Funding from the Deutsche Forschungsgemeinschaft(STI 125/19-1, GRK 2079) Carl-Zeiss-Stiftung, grants200021 146598 and 200020 162434 from the SwissNational Science Foundation, as well as the Depart-ment of Excellence, Department of Chemistry andTechnologies of drugs, and Progetto Ateneo-2016 (prot.n. RG116154C8E02882) of Sapienza University aregratefully acknowledged. ∗ [email protected] [1] T. Fennel, K.-H. Meiwes-Broer, J. Tiggesb¨aumker, P.-G.Reinhard, P. M. Dinh and E. Suraud, Reviews of modernphysics , 2010, , 1793.[2] U. Saalmann, C. Siedschlag and J. Rost, Journal ofPhysics B: Atomic, Molecular and Optical Physics , 2006, , R39.[3] H. Wabnitz, L. Bittner, A. De Castro, R. D¨ohrmann,P. G¨urtler, T. Laarmann, W. Laasch, J. Schulz,A. Swiderski, K. von Haeften et al. , Nature , 2002, ,482–485.[4] C. Bostedt, H. Thomas, M. Hoener, E. Eremina, T. Fen-nel, K.-H. Meiwes-Broer, H. Wabnitz, M. Kuhlmann,E. Pl¨onjes, K. Tiedtke et al. , Physical review letters , 2008, , 133401.[5] M. Arbeiter and T. Fennel,
New Journal of Physics , 2011, , 053022.[6] M. Arbeiter and T. Fennel, Physical Review A , 2010, ,013201.[7] H. Iwayama, K. Nagaya, M. Yao, H. Fukuzawa,X. Liu, G. Pr¨umper, K. Motomura, K. Ueda, N. Saito,A. Rudenko et al. , Journal of Physics B: Atomic, Molec-ular and Optical Physics , 2013, , 164019.[8] C. Jungreuthmayer, L. Ramunno, J. Zanghellini andT. Brabec, Journal of Physics B: Atomic, Molecular andOptical Physics , 2005, , 3029.[9] M. M¨uller, L. Schroedter, T. Oelze, L. N¨osel, A. Przys-tawik, A. Kickermann, M. Adolph, T. Gorkhover,L. Fl¨uckiger, M. Krikunova et al. , Journal of PhysicsB: Atomic, Molecular and Optical Physics , 2015, ,174002.[10] B. Sch¨utte, F. Campi, M. Arbeiter, T. Fennel,M. Vrakking and A. Rouz´ee, Physical review letters , 2014, , 253401.[11] T. Oelze, B. Sch¨utte, M. M¨uller, J. P. M¨uller,M. Wieland, U. Fr¨uhling, M. Drescher, A. Al-Shemmary,T. Golz, N. Stojanovic et al. , Scientific reports , 2017, ,40736.[12] B. Sch¨utte, M. J. Vrakking and A. Rouz´ee, Physical Re-view A , 2017, , 063417.[13] T. Pohl, D. Vrinceanu and H. Sadeghpour, Physical re-view letters , 2008, , 223201.[14] H. Iwayama, J. Harries and E. Shigemasa,
Physical Re-view A , 2015, , 021402.[15] B. Sch¨utte, T. Oelze, M. Krikunova, M. Arbeiter, T. Fen-nel, M. J. Vrakking and A. Rouz´ee, Journal of PhysicsB: Atomic, Molecular and Optical Physics , 2015, ,185101.[16] P. Di Cintio, U. Saalmann and J.-M. Rost, Physical re-view letters , 2013, , 123401.[17] A. E. Kaplan, B. Y. Dubetsky and P. Shkolnikov,
Phys-ical review letters , 2003, , 143401.[18] I. Last and J. Jortner, Physical Review A , 2001, ,063201.[19] E. Snyder, S. Wei, J. Purnell, S. Buzza and A. Castle-man Jr, Chemical physics letters , 1996, , 1–7.[20] D. Card, D. Folmer, S. Sato, S. Buzza and A. Castleman,
The Journal of Physical Chemistry A , 1997, , 3417–3423.[21] D. Niu, H. Li, F. Liang, L. Wen and X. Luo,
Chinesescience bulletin , 2005, , 2115–2117.[22] W. Wang, H. Li, D. Niu, L. Wen and N. Zhang, ChemicalPhysics , 2008, , 111–116.[23] B. Iwan, J. Andreasson, M. Bergh, S. Schorb, H. Thomas,D. Rupp, T. Gorkhover, M. Adolph, T. M¨oller, C. Bost- edt et al. , Physical Review A , 2012, , 033201.[24] V. Lyamayev, Y. Ovcharenko, R. Katzy, M. De-vetta, L. Bruder, A. LaForge, M. Mudrich, U. Person,F. Stienkemeier, M. Krikunova et al. , Journal of PhysicsB: Atomic, Molecular and Optical Physics , 2013, ,164007.[25] E. Allaria, R. Appio, L. Badano, W. Barletta, S. Bas-sanese, S. Biedron, A. Borga, E. Busetto, D. Castronovo,P. Cinquegrana et al. , Nature Photonics , 2012, , 699.[26] E. Allaria, A. Battistoni, F. Bencivenga, R. Borghes,C. Callegari, F. Capotondi, D. Castronovo, P. Cinque-grana, D. Cocco, M. Coreno et al. , New Journal ofPhysics , 2012, , 113009.[27] C. Bobbert, S. Sch¨utte, C. Steinbach and U. Buck, TheEuropean Physical Journal D-Atomic, Molecular, Opticaland Plasma Physics , 2002, , 183–192.[28] B. Dick, Physical Chemistry Chemical Physics , 2014, ,570–580.[29] J. A. Samson, G. Haddad and L. Kilcoyne, The Journalof chemical physics , 1987, , 6416–6422.[30] Y. Li and A. Varandas, The Journal of Physical Chem-istry A , 2010, , 6669–6680.[31] J. Biesner, L. Schnieder, G. Ahlers, X. Xie, K. Welge,M. Ashfold and R. Dixon,
The Journal of chemicalphysics , 1989, , 2901–2911.[32] A. Bach, J. M. Hutchison, R. J. Holiday and F. F. Crim, The Journal of Physical Chemistry A , 2003, , 10490–10496.[33] T. Freudenberg, W. Radloff, H.-H. Ritze, V. Stert,K. Weyers, F. Noack and I. Hertel,
Zeitschrift f¨ur PhysikD Atoms, Molecules and Clusters , 1996, , 349–364.[34] A. C. LaForge, R. Michiels, M. Bohlen, C. Callegari,A. Clark, A. von Conta, M. Coreno, M. Di Fraia,M. Drabbels, M. Huppert et al. , Physical review letters ,2019, , 133001.[35] D. R. Bates,
Atomic and molecular processes , Elsevier,2012, vol. 13.[36] A. Burgess,
Memoirs of the Royal Astronomical Society ,1965, , 1.[37] J. Davies, J. LeClaire, R. Continetti and C. Hayden, TheJournal of chemical physics , 1999, , 1–4.[38] C. Meier and V. Engel,
Physical Chemistry ChemicalPhysics , 2002, , 5014–5019.[39] M. Tsubouchi, C. A. de Lange and T. Suzuki, Journal ofelectron spectroscopy and related phenomena , 2005, ,193–205.[40] D. Winkoun and G. Dujardin,
Zeitschrift f¨ur Physik DAtoms, Molecules and Clusters , 1986, , 57–64.[41] A. Viel, W. Eisfeld, S. Neumann, W. Domcke andU. Manthe, The Journal of chemical physics , 2006, ,214306.[42] A. Viel, W. Eisfeld, C. R. Evenhuis and U. Manthe,
Chemical Physics , 2008, , 331–339.[43] B. Leyh and A. Hoxha,
Chemical physics , 1995, , 65–77.[44] M. Stankiewicz, P. Hatherly, L. Frasinski, K. Codling andD. Holland,
Journal of Physics B: Atomic, Molecular andOptical Physics , 1989, , 21.[45] M. S. Banna, H. Kossmann and V. Schmidt, Chemicalphysics , 1987, , 157–163.[46] A. Florescu-Mitchell and J. Mitchell,
Physics reports ,2006, , 277–374.[47] M. Langford, F. Harris, P. Fournier and J. Fournier,
In-ternational journal of mass spectrometry and ion pro- cesses , 1992,116