Ignition of a helium nanoplasma by x-ray multiple ionization of a heavy rare-gas core
D. Schomas, C. Medina, L. Ben Ltaief, R. B. Fink, S. Mandal, S. R. Krishnan, R. Michiels, M. Debatin, F. Stienkemeier, S. Toleikis, C. Passow, N. Ekanayake, C. Ott, R. Moshammer, T. Pfeifer, A. Heidenreich, M. Mudrich
IIgnition of a helium nanoplasma by x-ray multiple ionization of a heavy rare-gas core
D. Schomas, C. Medina, L. Ben Ltaief, R. B. Fink, S. Mandal, S. R. Krishnan, R. Michiels, M. Debatin, F. Stienkemeier, S. Toleikis, C. Passow, N. Ekanayake, C. Ott, R. Moshammer, T. Pfeifer, A. Heidenreich,
7, 8 and M. Mudrich Physikalisches Institut, Universit¨at Freiburg, 79104 Freiburg, Germany Department of Physics and Astronomy, Aarhus University, 8000 Aarhus C, Denmark Indian Institute of Science Education and Research, Pune 411008, India Indian Institute of Technology Madras, Chennai 600036, India Deutsches Elektronen-Synchrotron, DESY, 22607 Hamburg, Germany Max-Plank-Institut f¨ur Kernphysik, 69117 Heidelberg, Germany Kimika Fakultatea, Euskal Herriko Unibertsitatea (UPV/EHU) and DonostiaInternational Physics Center (DIPC), P.K. 1072, 20080 Donostia, Spain IKERBASQUE, Basque Foundation for Science, 48011 Bilbao, Spain (Dated: May 7, 2020)The dynamics of an x-ray-ionized two-component core-shell nanosystem is probed using dopedhelium (He) nanodroplets. First, a soft x-ray pump pulse selectively inner-shell ionizes the corecluster formed of heavier rare-gas atoms, causing electron migration from the He shell to the highly-charged core. This ignites a He nanoplasma which is then driven by an intense near-infrared probepulse. The ultrafast charge redistribution, evidenced by the rise of He + and He ion yields fromthe nanoplasma within (cid:46)
70 fs, leads to strong damping of the core cluster expansion. Thus, Hedroplets act as efficient tampers that reduce the radiation damage of embedded nanostructures, aproperty that could be exploited for improving coherent diffraction images.
Free nanometer-sized atomic clusters and nanodropletsare attractive test objects for studying the intricate inter-action of photons, electrons, and ions in strongly driven,highly correlated regimes of light-matter interaction. De-tailed knowledge of the interaction processes occurringinside highly ionized nanosystems, so-called nanoplas-mas, is essential for predicting and controlling the evo-lution of targets subjected to x-ray coherent diffractionimaging (CDI) [1–3]. Recording highly resolved coher-ent diffraction images of isolated nanoparticles and evenindividual molecules by single ultrashort x-ray pulses isone of the primary goals of existing and upcoming x-rayfree electron lasers (FELs) [4–6].Helium nanodroplets are particularly well suited tar-gets for studying the nanoplasma dynamics of heteroge-neous systems due to their unique properties: He dropletsefficiently pick up impurity atoms or molecules of anyspecies; owing to the superfluid nature of the He droplets,these dopants are highly mobile inside a He nanodropletwhere they aggregate into a cluster which is spatially andenergetically well separated from the He environment.Since He droplets are transparent from the infrared upto extreme ultraviolet (XUV) spectral regions, dopantscan be selectively excited or ionized without directly ex-citing the He host droplet [7–9].When He nanodroplets are exposed to intense near-infrared (NIR) laser pulses, they ionize in an avalanche-like process, where the presence of dopants in the dropletsdrastically lowers the intensity threshold for ionization.The evolution and systematics of dopant-induced, strong-field NIR-driven He nanoplasmas have been studied indetail [10–15]. In doped He nanodroplets, seed electrons were created by tunnel ionization of the dopants, andthe He shell is ignited by laser-driven electron impactionization (EII) [16].In the present work, we exploit for the first time an-other unique property of He nanodroplets – their neartransparency to x-ray radiation. Using short soft x-raypulses, we inner-shell photoionize heavier rare-gas atomsaggregated inside the nanodroplets without significantlyphotoionizing the surrounding shell of He atoms. In thisway, a core of highly charged ions and a distribution ofelectrons is created, which ignite a He nanoplasma drivenby an intense NIR pulse. The measured ultrafast build-up of the He nanoplasma ion yields in our soft x-ray-NIR pump-probe experiment reflects the ultrafast elec-tron transfer from the He shell to the dopant core, as re-vealed by molecular dynamics (MD) simulations. Whilethis electron transfer causes the He shell to become partlycharged and to expand, the dopant core is rapidly neu-tralized and its expansion is damped. Thus, we demon-strate that He droplets serve as tampers that efficientlyspatially confine x-ray ionized nanoclusters. This prop-erty of He nanodroplets could be exploited for improvingthe resolution of single-shot x-ray coherent diffraction im-ages of embedded molecules and nanoparticles [1, 17–20].The experiments were performed at the XUV-FEL fa-cility FLASH at DESY in Hamburg. A continuous Henanodroplet apparatus including a gas doping unit and acombined electron velocity-map imaging and ion time-of-flight mass spectrometer [21] was mounted at the open-port beamline BL2. The focused soft x-ray (5 nm) andNIR (800 nm) beams were collinearly superimposed us-ing a mirror with a centered hole mounted at the exit a r X i v : . [ phy s i c s . a t m - c l u s ] M a y of the beamline. The focal spot sizes (1 /e radius) ofthe x-ray and NIR beams in the interaction region ofthe spectrometer were 20 and 50 µ m, respectively. TheFWHM pulse durations were about 110 and 55 fs andpulse energies were 75 µ J and up to 10 mJ. In the ex-periments discussed here, the peak intensity of the x-raypulses was 2 . × Wcm − and that of the NIR pulseswas 1 . × Wcm − . The delay time between the twopulses was controlled by a mechanical delay stage placedin the NIR laser beam.The mean sizes of the He nanodroplets were varied inthe range (cid:104) N (cid:105) = 4000-11000 He atoms per droplet byvarying the temperature of the He nozzle. The meannumber of heavier rare-gas dopants embedded in the Hedroplets was controlled by the dopant partial pressurein the doping cell [7]. When He droplets are multiplydoped by rare gas atoms, the dopants form clusters insidethe He droplets, of which the binding energy is releasedinto the droplets [22]. The droplets subsequently shrinkby evaporation of He atoms. All droplet sizes specifiedin this work refer to the estimated reduced sizes due todopant aggregation [23].He nanodroplets are nearly transparent at a wave-length of 5 nm where the photoionization cross section ofHe (0.032 Mbarn) is by two orders of magnitude smallerthan that of Ar (4.8 Mbarn), Kr (5.4 Mbarn) and Xe(1.6 Mbarn). Therefore, the yield of He ions from pureHe nanodroplets irradiated by only the x-ray pulses fellbelow the detection threshold. Likewise, the NIR laserpulses alone did not substantially ionize pure He nan-odroplets due to the high ionization energy of He. Whendoping the He droplets with small amounts of the heav-ier rare gases Ar, Kr, or Xe, the NIR pulse alone yieldedonly low rates of He + and He +2 ions. However, when ap-plying both the x-ray and the NIR pulses simultaneouslyto doped He nanodroplets, large yields of He ions andelectrons were detected.Fig. 1 (a) shows a series of delay-dependent measure-ments of the yields of He + ions (symbols) for variousnumbers of Ar atoms embedded in He droplets of averagesize before doping (cid:104) N (cid:105) = 8000. The constant signal offsetdue to ionization by only the NIR pulse was subtracted.The delay-dependent yields of the most prominent ions,He + , as well as the yields of He , He +2 and of electronsfollow the same characteristic pump-probe delay depen-dence. Likewise, the pump-probe transients recorded forkrypton (Kr) and xenon dopants (not shown) displaythe same structure as those for Ar-doped He droplet.Fig. 1 (b) shows the He + yield measured for Kr-dopedHe droplets at various He droplet sizes. Followed by asteep rise within the temporal overlap of the two pulses,the He + yields reach a maximum value between 100 and200 fs, and fall off within about 1 ps. Note that no delaydependence is measured for pure He droplets [Fig. 1 (a),Ar ].To extract the characteristic features from each delay- Figure 1. He ion yields measured as a function of the de-lay between the x-ray pump and the NIR probe pulses. Inpanel (a), the He droplets initially consisting of about 8000He atoms were doped with various numbers of Ar atoms. In(b), He droplets of variable size were doped with Kr atomsas indicated. The insets show the fit results for the rise times τ , decay times τ , and maxima of the He + yields. dependent measurement, the data are fitted by the model s (∆ t ) = s [Θ(∆ t ) exp ( − ln 2 ∆ t/τ )] ◦ g (∆ t, τ ) , where g (∆ t, τ ) = exp (cid:0) − ln 2 ∆ t /τ (cid:1) (cid:112) ln 2 /π/τ is anormalized Gaussian function. This model is rationalizedby assuming that two infinitesimally short pulses inducea step-like signal rise Θ(∆ t ) at the delay ∆ t = 0, followedby an exponential signal decay ∝ exp ( − ln 2 ∆ t/τ ). Thefinite pulse durations are taken into account by convolv-ing this kernel with g (∆ t, τ ). Thus, in this model therise time of the measured signal, given by τ , should bedetermined by the width of the cross-correlation functionof the two laser pulses, whereas τ determines the intrin-sic decay time of the system. s denotes the height of thesignal maximum.Fig. 1 displays these fit results as insets in panels(a) and (b). The general trends can be summarized asfollows: The ion yields significantly rise for increasingdroplet size and doping level, as well as for rising FELand NIR pulse intensities (not shown). The two charac-teristic times τ and τ roughly double in value over thevaried ranges of droplet size and Ar doping level. Sur-prisingly, the shortest measured rise time of τ = 70 fs,measured for small droplets and low doping level, is sig-nificantly shorter than the width of the cross correlationof the two laser pulses, τ corr = 130 ±
20 fs. This find-ing leads us to conclude that the detected signal dependsnonlinearly on the pulse intensities, where the signal isstrongly enhanced in the delay range when the maximaof the two pulses nearly coincide. This is in line withthe concept that nanoplasmas form in an avalanche-likeionization process [24–26]. The ultrashort rise time ofthe ion yields indicates that small doped He droplets aremaximally activated by secondary ionizations within theduration of the x-ray pulse, whereas larger He dropletsare activated about 100 fs after the maximum of the x-ray pulse. A similar observation, recently made for pureXe clusters irradiated by hard x-ray pulses [27], was in-terpreted by the interplay between electron-impact ex-citation and depletion of excited states by interatomicCoulombic decay (ICD).To obtain deeper insights, we perform systematicmolecular dynamics (MD) simulations using the method-ology described elsewhere [12, 13]. The x-ray ionization isimplemented in an ad hoc procedure, where all Ar atomsare singly 3p or 2p photoionized, the latter triggering theemission of one or two Auger electrons with branchingratios reproducing the experimental relative charge stateabundances. These abundances (0 .
04 Ar + , 0 .
80 Ar and0 .
16 Ar ) are taken from our measurements of pure Aratoms and agree well with previous results [28]. Photoand Auger ionizations are assumed to proceed instanta-neously, generating a distribution of low and high-energyelectrons.Snapshots of a He droplet doped with an Ar clus-ter are shown on the top of Fig. 2. The Ar nanoplasmacreated at t = 0 (left panel) induces 100 more Ar and300 He ionizations during the first 200-300 fs, before thearrival of the NIR pulse. Most of these He + ions exit thedroplet within 200 fs (middle panel). Neutral He atoms,which acquire kinetic energy by electron-atom collisions,as well as Ar ions expand more slowly (right panel).When the NIR probe pulse arrives at ∆ t (cid:38) + and He is displayedin Fig. 2 (a) and (b), respectively, for He droplets of size (cid:104) N (cid:105) = 4300 doped with 34 Ar atoms. The simulatedpump-probe curves are obtained for the He Ar sys-tem. For each delay, the signal is averaged over 300 MDtrajectories and over the NIR laser intensities in the focalvolume similarly to Ref. [29].To test the response of the He shell in the extremecase where all photo and Auger electrons instantaneouslyleave the droplet at t = 0, simulations were carried outfor the corresponding highly cationic Ar clusters (opensymbols in Fig. 2). Strikingly, the two simulations nearlycoincide, indicating that the creation of multiply chargedAr ions inside the He nanodroplets is an efficient meansof igniting a nanoplasma. Furthermore, it shows thata droplet of 2000 He atoms is large enough to substan-tially stabilize even a multiply charged pure cation clus- Figure 2. Top row: Snapshots of the evolution of aHe Ar cluster x-ray ionized at t = 0. Balls of differentcolor indicate free electrons, neutral and ionized He atoms,and Ar ions. Experimental and simulated He + (a) and He (b) ion yields as a function of the pump-probe delay. ter. Only the He yield decays slightly faster when emit-ted electrons are removed, indicating that bare Ar q + ionsare more efficient in ionizing the He shell which then ex-pands faster.The rising edges in the simulations are steeper thanin the experiment due to the assumption of an instanta-neous x-ray ionization of the Ar core in the simulation.The drop of the simulated He + yield at ∆ t >
100 fs isslightly faster than the measured one; likely, in the ex-periment, more larger He droplets out of the broad Hedroplet size distribution contribute which feature longerdecay times τ . Larger He droplets contain more dopantsand are more susceptible to NIR strong-field ionization.Their expansion induced by x-ray ionization of the coreis delayed compared to small droplets. The ratio of He vs. He + yields (not shown) is slightly overestimated inthe simulation (0 .
40) compared to the experiment (0 .
01 0 02 0 03 0 002 04 06 08 0 0 2 0 4 0 6 0 8 0 1 0 0 1 2 0 1 4 051 01 52 02 53 03 5 ( a )
Ionizations
H e t o t a l A r t o t a l A r P E + A u g e r ( b )
F r e e A r c l u s t e r E m b e d d e d A r
Net charges in Ar cluster volume ( e ) T i m e ( f s ) ( c )
Ar cluster radius (Å)
F r e e A r c l u s t e r E m b e d d e d A r c l u s t e r
Figure 3. (a) Time evolution of the number of Ar andHe ionizations induced in a He Ar core-shell cluster byan x-ray pulse. (b) Sum of all charges inside the expandingAr cluster volume for a free Ar cluster (cyan line) and anHe Ar cluster (blue line). (c) Expansion dynamics of afree Ar cluster (black line) compared to an embedded Ar cluster (red line).
150 fs after x-ray ionization of an Ar cluster embeddedin a He droplet in the absence of the NIR pulse. Panel(a) depicts the number of inner ionizations [30] of theAr and the He atoms in response to creating Ar q + andelectrons. Initiated by the ionization of the Ar clusterby the x-ray pulse at t = 0 (cyan curve and area), theAr is further ionized mostly by EII (blue line and area)within the first two fs, creating mostly Ar and Ar ions. Electrons diffusing from the Ar nanoplasma into theHe shell generate He + and additional quasifree electronsby EII assisted by the electric field of the Ar ions (blacksolid line). This proceeds over a period of more than 200fs owing to the large number of He atoms surrounding theAr core. The large number of quasifree electrons createdby the highly charged Ar q + can subsequently be drivenby the NIR pulse to transform the whole core-shell clusterinto a nanoplasma.But how do the charges redistribute inside this core-shell system during the activation phase? This can beseen from the time-evolution of the total net charge Q of the cluster core, i. e. the sum of all ions and electroncharges inside the expanding Ar cluster volume. Fig. 3(b) shows the evolution of Q for an embedded Ar clus-ter and, for comparison, for a free Ar cluster (cyanline). The cyan shaded area shows that for the embed-ded Ar cluster, the initial Ar q + charges are substantiallyneutralized by electron transfer from the He shell to theAr cluster. This massive charge redistribution strongly Ar55 R (100fs)/R(0fs) N H e ( b )
Quasifree electrons (norm.)
T i m e ( f s )0 E m b e d d e d A r a l l e l e c t r o n s e ' s r e m o v e d ( a ) H e d r o p l e t s i z e N a l l e l e c t r o n s e ' s r e m o v e dF r e e A r c l u s t e r Figure 4. (a) Dependence of the expansion of the x-ray-ionized Ar cluster on the size of the He shell. Open symbolsshow the results of a simulation where the x-ray-emitted elec-trons are artificially removed. (b) He droplet size dependenceof the evolution of the number of nanoplasma electrons in theinitial volume of the droplet. affects the expansion dynamics of the system, as previ-ously observed for heavier rare gas clusters irradiated byXUV pulses [17]. Fig. 3 (c) shows the radius of the Arcluster, taken as the largest distance of an Ar atom fromthe Ar center of mass. The embedded Ar cluster ex-pands much more slowly (red line) compared to the freeAr cluster (black line), showing that the He droplet actsas a tamper that strongly dampens the expansion of theionized core cluster. The tamper effect features a strongdependence on the size of the He droplet, as previouslyshown [1, 17]. The increase of the Ar cluster radius at t = 100 fs relative to its initial size is displayed in Fig. 4(a). This factor amounts to 3.0 for a free Ar clusterand to 6.2 for a free Ar cluster if the x-ray-emitted elec-trons are instantaneously removed, see the marks on thevertical scale. For a He droplet of size N = 10 theexpansion of the embedded Ar cluster within 100 fs islargely suppressed (ratio of radii = 1 . τ of the pump-probe traces [insets inFig. 1 (a) and (b)] can be rationalized by the chargingdynamics of the x-ray-ionized Ar-He core-shell system.The NIR pulse optimally drives an ionization avalanchewhen a maximum of nanoplasma electrons has accumu-lated inside the nanodroplet. As the He shell is ionizedby the diffusive motion of electrons in the field of the Arions in the core [Fig. 3 (a)], we expect that the maximumof the He ionization is reached at later times for largerHe shells. Indeed, the number N e of nanoplasma elec-trons in the initial volume of the nanodroplet, shown inFig. 4 (b), reaches a maximum that shifts in time from t = 70 to 160 fs when the He droplet size increases from249 to 10149 He atoms per droplet. For better compa-rability, each N e ( t ) curve is normalized to its maximum.The drop of N e ( t ) at later times is due to the stagna-tion of EII by an interplay of decreasing electron kineticenergies and particle loss from the droplet as it startsto expand. The shifting of the N e ( t ) maxima is in goodagreement with the shifting of the maxima of the mea-sured pump-probe traces to longer delays, as quantifiedby τ (Fig. 1). To directly probe the electron-transferdynamics inside the nanodroplet and the resulting tam-pering of the core-cluster expansion, CDI experimentsshould be carried out using ultrashort x-ray pulses asthey are becoming available from modern FELs such asthe European XFEL.The authors gratefully acknowledge financial supportfrom the Basque Government (project IT1254-19) andfrom the Spanish Ministerio de Economia y Competivi-dad (project CTQ2015-67660-P), Deutsche Forschungs-gemeinschaft (DFG) within the project MU 2347/12-1 and STI 125/22-2 in the frame of the Priority Pro-gramme 1840 ‘Quantum Dynamics in Tailored IntenseFields’, and the Carlsberg Foundation. Computationaland manpower support provided by IZO-SGI SG Ikerof UPV/EHU and European funding (EDRF and ESF)is gratefully acknowledged. We would like to thank B.Manschwetus, R. Treusch, B. Erk and the whole FLASHteam for assistance and support during the experiment,and the group of K.-H. Meiwes-Broer for loaning partsof the apparatus. [1] C. Gnodtke, U. Saalmann, and J. M. Rost, Phys. Rev.A , 041201 (2009).[2] 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 , 93 (2016).[3] L. Fl¨uckiger, D. Rupp, M. Adolph, T. Gorkhover,M. Krikunova, M. M¨uller, T. Oelze, Y. Ovcharenko,M. Sauppe, S. Schorb, et al. , New J. Phys. , 043017(2016).[4] R. Neutze, R. Wouts, D. Van der Spoel, E. Weckert, andJ. Hajdu, Nature , 752 (2000).[5] H. N. Chapman, P. Fromme, A. Barty, T. A. White,R. A. Kirian, A. Aquila, M. S. Hunter, J. Schulz, D. P.DePonte, U. Weierstall, et al. , Nature , 73 (2011).[6] A. Barty, J. K¨upper, and H. N. Chapman, Annu. rev.phys. chem. , 415 (2013).[7] J. P. Toennies and A. F. Vilesov, Angew. Chem. Int. Ed. , 2622 (2004).[8] F. Stienkemeier and K. Lehmann, J. Phys. B , R127(2006).[9] M. Mudrich and F. Stienkemeier, Int. Rev. Phys. Chem. , 301 (2014).[10] S. Krishnan, L. Fechner, M. Kremer, V. Sharma, B. Fis-cher, N. Camus, J. Jha, M. Krishnamurthy, T. Pfeifer,R. Moshammer, et al. , Phys. Rev. Lett. , 173402 (2011).[11] S. R. Krishnan, C. Peltz, L. Fechner, V. Sharma, M. Kre-mer, B. Fischer, N. Camus, T. Pfeifer, J. Jha, M. Krish-namurthy, C.-D. Schr¨oter, J. Ullrich, F. Stienkemeier,R. Moshammer, T. Fennel, and M. Mudrich, New J.Phys. , 075016 (2012).[12] A. Heidenreich, B. Gr¨uner, M. Rometsch, S. Krishnan,F. Stienkemeier, and M. Mudrich, New J. Phys. ,073046 (2016).[13] A. Heidenreich, B. Gr¨uner, D. Schomas, F. Stienkemeier,S. R. Krishnan, and M. Mudrich, .[14] M. Kelbg, A. Heidenreich, L. Kazak, M. Zabel, B. Krebs,K.-H. Meiwes-Broer, and J. Tiggesb¨aumker, J. Phys.Chem. A , 8107 (2018).[15] M. Kelbg, M. Zabel, B. Krebs, L. Kazak, K.-H. Meiwes-Broer, and J. Tiggesb¨aumker, J. Chem. Phys. ,204302 (2019).[16] A. Mikaberidze, U. Saalmann, and J. M. Rost, Phys.Rev. Lett. , 128102 (2009).[17] M. Hoener, C. Bostedt, H. Thomas, L. Landt, E. Erem-ina, H. Wabnitz, T. Laarmann, R. Treusch, A. De Castro,and T. M¨oller, J. Phys. B , 181001 (2008).[18] Z. Jurek and G. Faigel, Eur. Phys. J. D , 35 (2008).[19] S. P. Hau-Riege, S. Boutet, A. Barty, S. Bajt, M. J. Bo-gan, M. Frank, J. Andreasson, B. Iwan, M. M. Seibert,J. Hajdu, et al. , Phys. Rev. Lett. , 064801 (2010).[20] B. Ziaja, H. Chapman, R. Santra, T. Laarmann,E. Weckert, C. Bostedt, and T. M¨oller, Phys. Rev. A , 033201 (2011).[21] D. Schomas, N. Rendler, J. Krull, R. Richter, andM. Mudrich, J. Chem. Phys. , 013942 (2017).[22] M. Lewerenz, B. Schilling, and J. P. Toennies, J. Chem.Phys. , 8191 (1995).[23] S. Kuma, H. Goto, M. N. Slipchenko, A. F. Vilesov,A. Khramov, and T. Momose, J. Chem. Phys. ,214301 (2007).[24] T. Fennel, K.-H. Meiwes-Broer, J. Tiggesb¨aumker, P.-G.Reinhard, P. M. Dinh, and E. Suraud, Rev. Mod. Phys. , 1793 (2010).[25] T. D¨oppner, J. M¨uller, A. Przystawik, S. G¨ode,J. Tiggesb¨aumker, K. Meiwes-Broer, C. Varin, L. Ra-munno, T. Brabec, and T. Fennel, Phys. Rev. Lett. ,53401 (2010).[26] B. Sch¨utte, M. Arbeiter, A. Mermillod-Blondin, M. J.Vrakking, A. Rouz´ee, and T. Fennel, Phys. Rev. Lett. , 033001 (2016).[27] Y. Kumagai, H. Fukuzawa, K. Motomura, D. Iablonskyi,K. Nagaya, S.-i. Wada, Y. Ito, T. Takanashi, Y. Sakak-ibara, D. You, et al. , Phy. Rev. X , 031034 (2018).[28] P. Lablanquie, L. Andric, J. Palaudoux, U. Becker,M. Braune, J. Viefhaus, J. Eland, and F. Penent, J.Electron Spectrosc. Rel. Phenom. , 51 (2007).[29] A. Heidenreich and J. Jortner, J. Chem. Phys. ,074315 (2011).[30] I. Last and J. Jortner, Phys. Rev. A62