Evolution and ion kinetics of a XUV-induced nanoplasma in ammonia clusters
R. Michiels, A. C. LaForge, 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, H. J. Wörner, F. Stienkemeier
EEvolution and ion kinetics of a XUV-induced nanoplasma 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, 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 S.C.p.A., 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: October 21, 2020)High-intensity extreme ultraviolet (XUV) pulses from a free-electron laser can be used to createa nanoplasma in clusters. In Ref. [Michiels et al. PCCP, 2020; 22: 7828-7834] we investigated theformation of excited states in an XUV-induced nanoplasma in ammonia clusters. In the presentarticle we expand our previous study with a detailed analysis of the nanoplasma evolution and ionkinetics. We use a time-delayed UV laser as probe to ionize excited states of H and H +2 in the XUV-induced plasma. Employing covariance mapping techniques, we show that the correlated emissionof protons plays an important role in the plasma dynamics. The time-dependent kinetic energy ofthe ions created by the probe laser is measured, revealing the charge neutralization of the clusterhappens on a sub-picosecond timescale. Furthermore, we observe ro-vibrationally excited molecularhydrogen ions H + ∗ being ejected from the clusters. We rationalize our data through a qualitativemodel of a finite-size non-thermal plasma. I. INTRODUCTION
Laser-induced nanoplasmas have been an active fieldof research in recent years and combine high energyphysics [1] with atomic and molecular quantum dynamicson a nanoscale [2–4]. Research has been fueled by the ne-cessity to understand radiation damage and plasma for-mation in the single-shot-imaging of nanoparticles [5].Many intriguing physical processes have been discov-ered: for example, nanoplasmas have been investigatedas sources of high-energy particles [6–9] and coherent ra-diation [10, 11]. Nanoplasma research in rare-gas clustersled to the discovery of enhanced absorption by collectivequasi-particle resonance of the electrons [12]. Many in-teresting properties of nanoplasmas come from the en-hancement of recombination processes due to the largenumber of confined electrons and positive ions [13, 14].The decay of a nanoplasma is governed by a complex in-terplay of Coulomb explosion and hydrodynamic forcesleading to shock shells in the outer Debye layer [15–18].Single-shot X-ray diffraction imaging of the nanoplasmaevolution in rare-gas clusters revealed sub-picosecond dy-namics [19], and a core-shell structure [20].Nanoplasmas in clusters can be induced using longwavelength radiation (infrared (IR), or near infrared(NIR)) on one hand, or, on the other hand, short-wavelength extreme ultraviolet (XUV) radiation. Thetwo regimes can be distinguished using the Keldysh pa-rameter γ ∝ /λ , where λ is the wavelength of the radia-tion [21–23]. A Keldysh parameter of γ (cid:28) γ (cid:29) E ele = hν − IP .Here, hν is the energy of a single XUV photon and IP isthe first ionization potential of the species.Concerning plasma dynamics, molecular clusters withmultiple components differ significantly from homoge-neous atomic clusters. When hydrogen is among theconstituents, the nuclear dynamics speeds up and thelightweight protons offer an efficient pathway for cool-ing and charge neutralization [26–28]. Nanoplasmas inmolecular clusters have been studied previously withtabletop [29–32] as well as free-electron lasers [7, 33],showing significant fragmentation of the molecules andthe generation of high-energy ions. Calculations predictthe inner charge state and temperature in the plasma coreto be much lower in (CH ) n molecular clusters, whencompared to atomic C n clusters [28]. Previous pump-probe experiments with nanosecond lasers were unableto resolve the fast plasma dynamics happening on a subpicosecond timescale. In a recent femtosecond XUV-pump UV-probe time-resolved experiment on ammoniaclusters, we investigated the dynamics of molecular andatomic states upon nanoplasma formation [33].In the present work, we extend these studies with adetailed analysis of the nanoplasma evolution and ion ki-netics. First, we will address the kinetic energy of H + a r X i v : . [ phy s i c s . a t m - c l u s ] O c t emitted from the nanoplasma and use covariance map-ping to analyze how energy is dissipated from the clus-ters via high-kinetic-energy protons. Secondly, the pho-toionization of H ∗ and the photodissociation of H +2 bythe probe laser are discussed. Finally, we analyze time-dependent kinetic energy distributions of H + and photo-electrons upon UV-ionization of H ∗ ( n = 2). We discusshow the observations allow conclusions concerning thelifetime of the Coulomb potential at the cluster surface. II. EXPERIMENTAL SETUP AND METHODS
The experiment was performed at the Low DensityMatter (LDM) endstation [34] at the seeded FEL FERMIin Trieste, Italy [35]. Details on the experimental setupcan be found in Ref. [33] and Ref. [34]. A jet of neu-tral ammonia clusters was created via supersonic expan-sion using a home-built pulsed nozzle. The mean clus-ter size was (cid:104) N (cid:105) = 2000 molecules and the cluster sizedistribution is assumed to be a broad log-normal distri-bution [36]. The cluster jet was crossed perpendicularlywith the XUV laser and a 261 nm UV laser. The inter-action region was in the focus of a combined velocity-map-imaging (VMI) and Wiley-McLaren [37] type iontime-of-flight (ToF) spectrometer. High intensity XUVpulses were used to multiply ionize the ammonia clus-ters. The FEL pulse intensity in the interaction regionwas I XUV ≈ · W/cm ² at 14.3 eV photon energy, I XUV ≈ · W/cm ² at 19.2 eV, 23.8 eV, 28.6 eV and33.4 eV photon energy and I XUV ≈ · W/cm ² at42.9 eV photon energy. The UV probe pulse had an inten-sity of I UV ≈ · W/cm ² . Correlation maps were cal-culated as Pearson’s correlation coefficient [38, 39]. Weincluded partial covariances [40] to compensate for thetarget density fluctuations and FEL pulse energy fluc-tuations using the sum intensity of the ion-ToF signalper FEL shot as a control variable. The correlation mapwas calculated using 30 000 FEL shots. The shot-to-shotstandard deviation of the FEL pulse energy, as well as thestandard deviation of the total ion-ToF sum was ≈ p -value. The p -valueshows the probability of finding an equal or stronger cor-relation if the correlation were in fact zero.In the measured ion-ToF spectrum, the high-kinetic-energy protons create distinct forward and backwardpeaks corresponding to protons arriving earlier and laterthan those initially at rest. This process is illustratedin Fig. 1 a). The flight trajectories of three differentprotons originating from a highly charged cluster aresketched. The protons created with initial momentum inthe forward and backward direction of the spectrometercreate separate peaks in the ion-ToF spectrum. Ionswith initial velocity components perpendicular to theextraction direction may not be detected. Using ion V E x t r a c t o r e l e c t r o d e :I o n T o F t u b e :
Intensity (arb.)
T i m e o f f l i g h t ( n s )
E x p e r i m e n t S i m u l a t i o n ( f i t ) S i m u l a t i o n ( 1 9 e V ) b )a )
I o n M C PR e p e l l e r e l e c t r o d e : U U U Figure 1. a) Sketch illustrating the ion-ToF geometry andthe resulting proton flight paths for three different initial mo-menta (cid:126)k , (cid:126)k , (cid:126)k . b) Experimental proton-ToF spectrum (red)and best fit from the simulation (blue). The black curve is asimulation for one exact proton kinetic energy (19 eV), servingas an illustration of the double peak structure. trajectory simulations and the known geometry of theToF spectrometer, the initial kinetic energy of the ionscan be deduced from their arrival time. An example ofthe forward and backward peak can be seen in the ToFdistribution simulated for protons with 19 eV kineticenergy, which is shown as the black curve in the Fig. 1 b).The method used to calculate kinetic energies from theion-ToF spectra was adapted from Ref. [41]. We carriedout ion-ToF trajectory simulations for discrete integersteps in proton kinetic energy ( E H + = 0 , , , , ... eV)using Simion ® . The experimental spectra were fittedto a linear combination of the simulated spectra usingleast-squares fitting. An example ToF spectrum (red)and the resulting fit (blue) are shown in Fig. 1 b). To ac-count for the detector resolution, the simulated ion flighttimes were convoluted with a Gaussian function. Thedetector resolution was determined experimentally to be18 ns full width at half maximum (FWHM). From thefit parameters, we calculated the mean kinetic energy ofthe protons. Because the detection efficiency for pro-tons depends on the kinetic energy, we it into accountwhen calculating the mean kinetic energy. The detec-tion efficiency was obtained from the Simion simulationsby calculating the portion of protons arriving at the de-tector. Only about 10% of isotropically emitted protonswith 10 eV kinetic energy are detected. III. RESULTS AND DISCUSSIONA. Correlated emission of protons from thenanoplasma
A comparison of experimental ToF spectra for protonscreated with high-intensity XUV pulses (red) and onlythe UV pulse (blue) is shown in Fig. 2 a). The red curveis broader than the blue curve, and the shoulders indicatea large contribution from high-kinetic-energy protons. Inour experiment, ions are regarded as having a high kineticenergy if they are created with kinetic energies ≥ x -axis, and the flight time of the other proton on the y -axis. A positive value means that in a given spectrumcontaining a set of protons arriving at time t x , there is ahigher probability to also detect protons with arrival time t y . A negative value shows a reduced probability, i.e. thatthe arrival times t x and t y are anti-correlated. On the di-agonal line of the map in Fig. 2, the auto-correlation isseen, broadened by the detector resolution. The correla-tion coefficient is symmetric with respect to the diagonal,we show both sides in order to facilitate projections onthe spectrum in the top panel.In the upper left half of the map, we see a region ofstrong positive correlations centered at [ t , t ] = [1460 ns, Time of flight (ns)
T i m e o f f l i g h t ( n s ) - 2 . 0- 1 . 00 . 01 . 02 . 03 . 04 . 05 . 06 . 0
Correlation (%)
I o n k i n e t i c e n e r g y ( e V ) b ) E x p e r i m e n t : H + h i g h k i n e t i c e n e r g y H + l o w k i n e t i c e n e r g y Intensity a ) Figure 2. a) Proton peak from the ion-ToF spectrum obtainedby irradiation of ammonia clusters with high-intensity XUVpulses ( hν = 28 . E H + ≥ E H + ≈
20 eV.The single-pixel p -value of this correlation feature is p ≤ · − . Thus, the correlation is significant even forkinetic energies where the collection efficiency is < t , t ] = [1520 ns, 1560 ns] and [ t , t ] = [1440 ns,1470 ns]. These peaks show an anti-correlation of high-kinetic-energy protons with low-kinetic-energy protons.In the center of the image, the correlation is substan-tially lower than for the main peaks, converging to zerountil overwhelmed by the auto-correlation.Correlated emission of protons is only possible if thereis a common source. To interpret the data, we willlook at it in an event-based picture. The correlationsobserved in the experiment show that there is anunderlying event A, which creates high-kinetic-energyprotons with equal kinetic energy. Additionally, this Power exponent h v - I P2 0 1 0 4 2 0 2 4 1 0 2 002 04 06 08 01 0 01 2 01 4 01 6 0 H + Intenisty (arb.)
K i n e t i c e n e r g y ( e V )a )b ) E H + = 0 e V E H + = 7 e V E H + = 2 0 e V Intensity (arb.)
F E L i n t e n s i t y ( 1 0
W / c m ² )
Figure 3. a) Proton signal as a function of FEL intensityfor three different proton kinetic energies ( E H + ). b) Pro-ton intensity as a function of kinetic energy (blue curve) andthe power exponent k ( E H + ) (red curve) indicating the pump-power dependence of the proton intensity for different kineticenergies. Data was taken using a pump laser with 28.6 eVphoton energy. event A is negatively correlated to an event B, whichcreates low kinetic energy protons. Event A needs to bea multi-body fragmentation, either of a doubly chargedmolecule in the gas phase, or a highly charged cluster.Despite the fact that there are significant amounts ofresidual non-condensed molecules present in the clusterjet, the correlation observed in Fig. 2 b) cannot be causedby gas-phase molecules. This is primarily because bothdetected particles are protons and the parent moleculeis ammonia. The only possible fragmentation channelleading to two correlated protons is NH → NH + H + + H + , a channel which has so far not been observed inphotoion-photoion coincidence measurements [42, 43], orion-impact dissociation [44] on doubly-charged ammonia.Therefore, we assume that event A is associated a singlecluster being multiply ionized by absorption of n ≥ hν − IP [25], where IP isthe first ionization potential of the ammonia cluster(9.4 eV [46]). In general, charge ejection out of multiplycharged clusters happens either as a concerted Coulombexplosion, or as sequential emission of positively chargedions. In a sequential emission process, each clustergradually cools and creates positive correlation betweenall kinetic energies of the cooling cascade, including pos-itive correlations between high- and low-kinetic-energyprotons. In contrast, we observe a very pronouncedcorrelation between protons of equal kinetic energy,clearly pointing to a concerted ejection of charges.Consequently, the experimental correlations we observeprovide strong evidence for a pronounced core-shellnature of the Coulomb explosion of an XUV-inducednanoplasma in ammonia clusters. In this core-shellexplosion, a significant amount of charge and energyis taken away by protons with one specific kinetic energy.We now take a closer look at how the XUV pulse powerinfluences the kinetic energy of the detected protons.Fig. 3 b) shows a typical XUV-induced proton ToF peakwith strong broadening (blue curve). Additionally, thepower exponent k from a power fit: A ( I XUV ) = A I k XUV is shown for the different areas of the peak (red curve).Here, A is the ion intensity and I XUV is the XUV pulseintensity. The scaling constant A is a free parameterof the fit. Examples of the power fitting for different ki-netic energies are given in Fig. 3 a). Vertical dashed linesin Fig. 3 b) mark the area where the time of flight corre-sponds to a proton kinetic energy of E H + = hν − IP . Thepower coefficient is close to one and approximately con-stant for all protons with E H + ≤ hν − IP . A power coef-ficient of one shows a linear relation between the numberof photons in the XUV pulse and the number of protonsdetected. We conclude that the number of protons emit-ted from the nanoplasma rises linearly as a function ofXUV pulse intensity. This linear relation is surprisingconsidering the multi-photon nature of the nanoplasmaignition. However, it can be explained with the core-shellstructure of the nanoplasma. An increase in XUV pulseintensity cannot increase the plasma potential beyondfull frustration. Nonetheless, each additional photon ab-
104 08 01 2 01 6 0 2- 2- 10 1 0 1 5 2 0 2 5 3 0 3 5 4 0 4 50 . 02 . 04 . 06 . 08 . 01 0 . 0 1 224681 01 2
Intensity (arb.) m / z ( a t o m i c u n i t s )
P u m p p h o t o n e n e r g y : 1 4 . 3 e V 1 9 . 1 e V 2 8 . 6 e V a )P r o b e H + P u m p p h o t o n e n e r g y : 1 4 . 3 e V 1 9 . 1 e V 2 8 . 6 e V c )P r o b e H +2 Intensity (arb.) m / z ( a t o m i c u n i t s ) P r o b e H + P r o b e H +2 P u m p H +2 (cid:1) EH+ (eV)
P u m p p h o t o n e n e r g y ( e V ) e ) m / z ( a t o m i c u n i t s )
P u m p p h o t o n e n e r g y : 1 4 . 3 e V ( s c a l e d ) 1 9 . 1 e V ( s c a l e d ) 2 8 . 6 e V ( s c a l e d ) b )P u m p H + P u m p p h o t o n e n e r g y : 1 4 . 3 e V 1 9 . 1 e V 2 8 . 6 e V d )P u m p H +2 m / z ( a t o m i c u n i t s ) Figure 4. a) Proton-ToF peak created by the probe laser afternanoplasma ignition with different XUV pump wavelengths.b) Corresponding pump proton-ToF peak. c) H +2 -ToF peakcreated by the probe laser after nanoplasma ignition with dif-ferent XUV pump wavelengths. d) Corresponding pump H +2 -ToF peak. Note that the small peaks at the center are an ar-tifact from the subtraction of background-gas contributions.e) Mean kinetic energy difference of the H + and H +2 ions asa function of the XUV photon energy used for nanoplasmaignition. The values are displayed as difference compared tothe mean kinetic energy obtained by UV-ionizing the clus-ters. This is done in order to distinguish between broadeningeffects from the cluster environment and kinetic energy re-lease from Coulomb explosion. Red and blue: Probe ions.Green: Pump H +2 ions. All data shown were obtained at apump-probe delay of 18 ps. sorbed in the cluster will supply an energy of hν to thenanoplasma. A part of this additional energy is dissi-pated by emitting more protons, and the experimentaldata show that the relation between the photon flux andthe number of protons emitted is linear. Furthermore,we observe that emission of protons with E H + ≥ hν − IP is highly non-linear in XUV pulse intensity. From this we conclude that a large increase in XUV pulse intensity isrequired to create a plasma potential that is larger than hν − IP . B. Probing the nanoplasma with UV laserradiation
The previous section discussed the kinetics of protonsemitted from highly ionized ammonia clusters. In thefollowing, we will look at the difference in the ion spec-trum for pump-probe (XUV+UV pulse) and pump only(XUV pulse). The difference between pump-probe ionsand pump-only ions will be called probe ion yield. Theions that are created by the pump alone will be calledpump ion yield. The UV laser has a large single-photonionization cross section for the excited states of moleculesand atoms. Absorption of multiple photons is required toionize the electronic ground states of all involved atomsand molecules. Thus, the effect of the UV probe laseron the neutral molecules is negligibly small when com-pared to the ionization out of excited states. In thenanoplasma, excited states are created via recombinationof free electrons and ions, or by electron-impact excita-tion [25, 33].We will focus on H + and H +2 probe ion yields, startingwith the asymptotically converged spectrum at a pump-probe delay of 18 ps. The corresponding probe ion-ToFpeaks are shown in Fig. 4 a) and c), respectively. We ob-serve that the UV pulse can have two different effectson the ion yield, specifically, producing additional H + ions, while decreasing the H +2 ion yield. We observe thatthe H +2 ToF spectrum has a double peak structure. Incontrast, the H + ion peak is not significantly broadened.The double peak structure for the H +2 probe ion yieldis particularly pronounced, forming a local minimum inthe center, a feature that is not seen in the pump H +2 spectrum (see Fig. 4 d)). The decrease of H +2 ion yield inthe asymptotic difference can only be due to dissociationof H +2 by the probe laser. We conclude that, despite thefact that H +2 with low kinetic energy is created by theXUV pump laser, the probe laser primarily dissociatesthe high-kinetic-energy H +2 .We will now look at the photodissociation cross sec-tion of H +2 in order to explain how the probe laser se-lectively dissociates H +2 with kinetic energy. The dis-sociation probability in our experiment varies between10% at the center of the H +2 ToF peak, and 50% on theouter flanks (c.f. Fig. 4 c) and d)). To explain a dissoci-ation probability of 10% with the used probe laser in-tensity, a photodissociation cross section of σ ≈ . σ ≈ . +2 has a strongdependence on the ro-vibrational quantum state of themolecule [47]. The cross sections required for the exper-imentally observed dissociation probabilities of 10% and50% correspond to a ro-vibrational energy of 2500 K and8400 K, respectively [47]. From this selectivity, we candraw two conclusions: First, a large part of the H +2 cre-ated in the ammonia nanoplasma, particularly the H +2 with low kinetic energy, has ro-vibrational energies lowerthan 2500 K. Secondly, the H +2 ions that emerge from thenanoplasma with significant kinetic energy do also havea larger ro-vibrational energy.Quantitative values for the mean kinetic energy of theprobe H +2 and H + at a pump-probe delay of 18 ps andthe pump H +2 ions are shown in Fig. 4 e). The valuesare displayed as ∆ E H + , the difference compared to themean kinetic energy of protons in a spectrum obtainedby UV-ionizing the clusters. This is done in order todistinguish between broadening effects from the clusterenvironment and kinetic energy release from Coulombexplosion. On the x -axis, the photon energy of the XUVpump is varied in the range of 14.3 eV to 42.9 eV. Thekinetic energy of the probe H +2 ions (blue circles) rangesfrom 1 eV (12 000 K) to 9 eV (100 000 K), and is largerthan the ro-vibrational energy. Furthermore, the meankinetic energy of probe H +2 (blue circles) is significantlylarger when compared to that of the pump H +2 (green tri-angles). The mean kinetic energy of probe and pump H +2 rises as a function of hν . The XUV photon energy de-pendence of the mean kinetic energy of H +2 shows that ahigher XUV photon energy generates a deeper plasma po-tential. However, the change in mean kinetic energy doesnot directly correspond to the change in photon energy.This is not surprising, since the total photon flux andthe ionization cross section are peaked at 19 eV. Fromthe difference in the mean kinetic energy of probe andpump H +2 we conclude that the ro-vibrationally excitedH +2 is primarily created in the Debye layer of highly ion-ized clusters where the electric field is the strongest. Onthe other hand, the H +2 with lower ro-vibrational energyare emitted during the later stages of the nanoplasmaevolution.No significant kinetic energy release can be seen in theprobe H + (red rectangles in Fig. 4 e)). The primary por-tion of the probe H + signal is due to UV ionization outof excited states of atomic hydrogen [33, 48]. These ex-cited hydrogen atoms are formed according to the reac-tions [33]: H + + e − → H ∗ and (1)NH ∗ +3 → NH +2 + H ∗ . (2)From the negligible kinetic energy, we deduce that H ∗ isnot subject to a significant plasma potential at the mo-ment of ionization, i.e., at a pump-probe delay of 18 ps.High-kinetic-energy protons that are created through the dissociation of H +2 contribute only a minor part to theprobe H + yield, as can be seen by comparing the abso-lute scales of Fig. 4 a) and c). C. Time-resolved studies on Coulomb explosion
In the spectra shown in Fig. 4 a) and c) we observedthat H + ions created by the probe laser at a pump-probedelay of 18 ps do not have significant kinetic energy.For the case that H ∗ in the cluster is ionized whilethe nanoplasma is still active, there are two differentoptions. Either the excited hydrogen atom was formedin the bulk of the plasma, in which case the proton willremain inside the cluster; or, the excited hydrogen wasformed in the Debye layer of the plasma, in which casethe proton will acquire kinetic energy proportional tothe plasma potential at the time of ionization. Thisallows us to use the time-dependent kinetic energy ofthe probe protons as a probe for the plasma potential.Fig. 5 a) shows the mean kinetic energy of the probeH + as a function of pump-probe delay. The kineticenergy has a maximum in the vicinity of t . Afterwards,it decays rapidly and converges at 400 fs pump-probedelay. The decay time-constant is universal for allpump photon energies and only marginally larger thanthe temporal pulse overlap of the two laser pulses(depicted as the grey shaded area in Fig. 5 a)). Usingour ToF spectrometer, the kinetic energy of the protonscould only be determined for values larger than 500 meV.Complementary to the ions, we can detect photo-electrons from the probe ionization of H ∗ . The mostabundantly populated H ∗ state is the n = 2 state [33],yielding electrons with energy E ele = 1.35 eV when ion-ized with the 4.75 eV UV photons. If there is a Coulombpotential present, we observe a shift in the kineticenergy of the photoelectrons. The pump-probe delaydependent shift in the vertical binding energy (VBE)of the electrons (∆ E ele ) is displayed in Fig. 5 b). Theinset shows an example of the different peak positionsat pump-probe delays of 0.5 ps and 1 ps, respectively.A determination of the peak position was only possiblefor ∆ E ele < −
400 meV. For larger shifts, the peak isstrongly broadened and could not be distinguished fromthe overlapping low energy electrons emitted by thenanoplasma. The VBE shift at a pump-probe delay of300 fs is roughly -300 meV and converges to zero with ahalf-lifetime of (200 ±
30) fs.We conclude that the fast charge-equalization seenin the proton kinetic energy is governed by the high-energy Coulomb explosion in the Debye layer. After theCoulomb explosion, the clusters are only mildly chargedand hydrodynamic forces are dominant, explaining theslower decay of the remaining -300 meV plasma potential. - 2 0 0 0 2 0 0 4 0 0 6 0 0 8 0 0 1 0 0 0 1 2 0 0 1 4 0 0 1 6 0 0 1 8 0 0- 2- 10123456789 - 2 0 0 0 2 0 0 4 0 0 6 0 0 8 0 0 1 0 0 0 1 2 0 0 1 4 0 0 1 6 0 0 1 8 0 0- 0 . 5- 0 . 4- 0 . 3- 0 . 2- 0 . 10 . 0 0 . 9 1 . 1 1 . 3 1 . 5 1 . 70 . 40 . 60 . 81 . 01 . 2
P u m p p h o t o n e n e r g y : 1 4 . 3 e V 1 9 . 1 e V 2 3 . 8 e V
Proton (cid:1) kE (eV)
P u m p - p r o b e d e l a y i n f s P u l s e o v e r l a p c )b ) P u m p p h o t o n e n e r g y : 1 4 . 3 e V 1 9 . 1 e V 2 3 . 8 e V
Electron (cid:1) kE in eV
P u m p - p r o b e d e l a y i n f sa )
P u m p - p r o b e d e l a y : 1 p s 0 . 5 p s
Intensity (arb.) e k E i n e V (cid:1) k E
Figure 5. a) Mean kinetic energy of probe-laser-induced H + versus pump-probe delay for nanoplasma ignition with differ-ent XUV photon energies. The shaded area shows the crosscorrelation of the two laser pulses deduced from a 1+1 (cid:48) ioniza-tion of helium atoms. b) Shift in the vertical binding energyof probe laser ionized H ∗ (n = 2) as a function of pump-probedelay. The inset shows an example of the photoelectron peakfor two different pump-probe delays. c) Sketch depicting theevolution of the highly ionized ammonia cluster in two phases. At a pump-probe delay of 1 ps, the plasma potential iscompletely neutralized. The two-phase expansion of thecluster is illustrated in Fig. 5 d).
IV. CONCLUSION
We induced a nanoplasma in ammonia clusters us-ing high power XUV radiation from the FERMI free-electron laser. Using simultaneous photoelecton and iondetection we have shown that emission of high-kinetic-energy ions plays a significant role in the evolution ofXUV-induced nanoplasmas. Using shot-to-shot covari-ance mapping, we show that protons with kinetic energy4 eV ≤ E H + ≤
20 eV are emitted from the nanoplasmain a correlated way. Furthermore, we use a delayed UVlaser as probe to ionize excited states of H and H +2 in theplasma. From the pump-probe dependent mean kineticenergy of the probe-laser induced H + ions, we get infor-mation on the lifetime of the plasma confining potential.We found that the nanoplasma decays in a two-stage pro-cess. In the first stage, the potential reduces drasticallythrough the concerted emission of protons. These ob-servations show that the highly-ionized ammonia clusteracts as a core-shell system where the Debye layer of thenanoplasma dissipates a large fraction of the energy con-tained in the nanoplasma through a concerted Coulombexplosion of protons and other light, positively-chargedions. After the Coulomb explosion, the cluster expansionslows down and hydrodynamic forces become dominant.Using UV-laser-induced dissociation of H +2 , we show thatthe ion temperatures in the core of the XUV-inducednanoplasma are less than 2500 K. ACKNOWLEDGEMENTS
Funding from the Deutsche Forschungsgemeinschaft(STI 125/19-2, GRK 2079) Carl-Zeiss-Stiftung, grants200021 146598 and 200020 162434 from the Swiss Na-tional Science Foundation, as well as the Departmentof Excellence, Department of Chemistry and Technolo-gies of drugs, and Progetto Ateneo-2016 (prot. n.RG116154C8E02882) of Sapienza University are grate-fully acknowledged. ∗ [email protected][1] T. Ditmire, J. Zweiback, V. Yanovsky, T. Cowan,G. Hays and K. Wharton, Nature , 1999, , 489–492.[2] T. Fennel, K.-H. Meiwes-Broer, J. Tiggesb¨aumker, P.-G.Reinhard, P. M. Dinh and E. Suraud,
Reviews of modernphysics , 2010, , 1793. [3] U. Saalmann, C. Siedschlag and J. Rost, Journal ofPhysics B: Atomic, Molecular and Optical Physics , 2006, , R39.[4] U. Saalmann, Journal of Physics B: Atomic, Molecularand Optical Physics , 2010, , 194012.[5] R. Neutze, R. Wouts, D. Van der Spoel, E. Weckert andJ. Hajdu, Nature , 2000, , 752–757.[6] T. M. Trivikram, R. Rajeev, K. Rishad, J. Jha andM. Krishnamurthy,
Physical review letters , 2013, ,143401.[7] 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.[8] I. Last and J. Jortner, Physical Review A , 2001, ,063201.[9] M. Hohenberger, D. Symes, K. Madison, A. Sumeruk,G. Dyer, A. Edens, W. Grigsby, G. Hays, M. Teichmannand T. Ditmire, Physical review letters , 2005, , 195003.[10] J. Tisch, T. Ditmire, D. Fraser, N. Hay, M. Mason,E. Springate, J. Marangos and M. Hutchinson, Journal ofPhysics B: Atomic, Molecular and Optical Physics , 1997, , L709.[11] B. B´odi, M. Aladi, P. R´acz, I. B. F¨oldes and P. Dombi, Optics express , 2019, , 26721–26727.[12] J. Zweiback, T. Ditmire and M. Perry, Physical ReviewA , 1999, , R3166.[13] B. Sch¨utte, F. Campi, M. Arbeiter, T. Fennel,M. Vrakking and A. Rouz´ee, Physical review letters , 2014, , 253401.[14] B. Sch¨utte, T. Oelze, M. Krikunova, M. Arbeiter, T. Fen-nel, M. J. Vrakking and A. Rouz´ee,
New Journal ofPhysics , 2015, , 033043.[15] A. E. Kaplan, B. Y. Dubetsky and P. Shkolnikov, Phys-ical review letters , 2003, , 143401.[16] M. Arbeiter and T. Fennel, New Journal of Physics , 2011, , 053022.[17] 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.[18] B. Sch¨utte, M. J. Vrakking and A. Rouz´ee, Physical Re-view A , 2017, , 063417.[19] T. Gorkhover, S. Schorb, R. Coffee, M. Adolph, L. Fou-car, D. Rupp, A. Aquila, J. D. Bozek, S. W. Epp, B. Erk et al. , Nature photonics , 2016, , 93.[20] D. Rupp, L. Fl¨uckiger, M. Adolph, A. Colombo,T. Gorkhover, M. Harmand, M. Krikunova, J. P. M¨uller,T. Oelze, Y. Ovcharenko et al. , Structural Dynamics ,2020, , 034303.[21] L. Keldysh, Soviet Physics—JETP , 1964, , 1945.[22] F. Faisal and L. Dimou, in Topics in Atomic and NuclearCollisions , Springer, 1994, pp. 361–370.[23] H. R. Reiss and V. P. Krainov, ICONO’95: Fundamentalsof Laser-Matter Interaction, 1996, pp. 39–44.[24] A. Zheltikov,
Physical Review A , 2016, , 043412.[25] M. Arbeiter and T. Fennel, Physical Review A , 2010, ,013201. [26] I. Last and J. Jortner, The Journal of chemical physics ,2004, , 1336–1347.[27] A. Andreev, P. Nickles and K. Y. Platonov,
Physics ofPlasmas , 2010, , 023110.[28] P. Di Cintio, U. Saalmann and J.-M. Rost, Physical re-view letters , 2013, , 123401.[29] E. Snyder, S. Wei, J. Purnell, S. Buzza and A. Castle-man Jr,
Chemical physics letters , 1996, , 1–7.[30] D. Card, D. Folmer, S. Sato, S. Buzza and A. Castleman,
The Journal of Physical Chemistry A , 1997, , 3417–3423.[31] D. Niu, H. Li, F. Liang, L. Wen and X. Luo,
Chinesescience bulletin , 2005, , 2115–2117.[32] W. Wang, H. Li, D. Niu, L. Wen and N. Zhang, ChemicalPhysics , 2008, , 111–116.[33] R. Michiels, A. C. LaForge, M. Bohlen, C. Callegari,A. Clark, A. von Conta, M. Coreno, M. Di Fraia,M. Drabbels, P. Finetti et al. , Physical Chemistry Chem-ical Physics , 2020, , 7828–7834.[34] 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.[35] 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.[36] C. Bobbert, S. Sch¨utte, C. Steinbach and U. Buck, TheEuropean Physical Journal D-Atomic, Molecular, Opticaland Plasma Physics , 2002, , 183–192.[37] W. Wiley and I. H. McLaren, Review of scientific instru-ments , 1955, , 1150–1157.[38] K. Pearson, Proceedings of the Royal Society of London ,1895, , 240–242.[39] P. Schober, C. Boer and L. A. Schwarte, Anesthesia &Analgesia , 2018, , 1763–1768.[40] L. J. Frasinski,
Journal of Physics B: Atomic, Molecularand Optical Physics , 2016, , 152004.[41] B. Bapat and V. Sharma, International Journal of MassSpectrometry , 2006, , 10–15.[42] D. Winkoun and G. Dujardin,
Zeitschrift f¨ur Physik DAtoms, Molecules and Clusters , 1986, , 57–64.[43] M. Stankiewicz, P. Hatherly, L. Frasinski, K. Codling andD. Holland, Journal of Physics B: Atomic, Molecular andOptical Physics , 1989, , 21.[44] P. Bhatt, T. Sairam, A. Kumar, H. Kumar and C. Safvan, Physical Review A , 2017, , 022710.[45] J. Eland, Laser Chemistry , 1991, , 259–263.[46] A. Lindblad, H. Bergersen, W. Pokapanich, M. Tchap-lyguine, G. ¨Ohrwall and O. Bj¨orneholm, Physical Chem-istry Chemical Physics , 2009, , 1758–1764.[47] V. S. Lebedev, L. P. Presnyakov and I. I. Sobel’man, Physics-Uspekhi , 2003, , 473.[48] 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 ,2019,122