Geometric Dependence of Strong Field Enhanced Ionization in D_{2}O
Gregory A. McCracken, Andreas Kaldun, Chelsea Liekhus-Schmaltz, Philip H. Bucksbaum
aa r X i v : . [ phy s i c s . c h e m - ph ] A ug Geometric Dependence of Strong Field Enhanced Ionization in D O Gregory A. McCracken,
1, 2, a) Andreas Kaldun,
1, 3
Chelsea Liekhus-Schmaltz,
1, 3 and Philip H. Bucksbaum
1, 3, 2 Stanford PULSE Institute, SLAC National Accelerator Laboratory2575 Sand Hill Road, Menlo Park, CA 94025 Department of Applied Physics, Stanford University, Stanford, CA 94305 Department of Physics, Stanford University, Stanford, CA 94305 (Dated: 11 September 2018)
We have studied strong-field enhanced dissociative ionization of D O in 40 fs, 800 nm laserpulses with focused intensities of < − ×
15 W / cm by resolving the charged fragment mo-menta with respect to the laser polarization. We observe dication dissociation into OD + /D + dominates when the polarization is out of the plane of the molecule, whereas trication dis-sociation into O + /D + /D + is strongly dominant when the polarization is aligned along theD-D axis. Dication dissociation into O/D + /D + , and O + /D is not seen, nor is there anysignificant fragmentation into multiple ions when the laser is polarized along the C v symme-try axis of the molecule. Even below the saturation intensity for OD + /D + , the O + /D + /D + channel has higher yield. By analyzing how the laser field is oriented within the molecularframe for both channels, we show that enhanced ionization is driving the triply charged threebody breakup, but is not active for the doubly charged two body breakup. We conclude thatlaser-induced distortion of the molecular potential suppresses multiple ionization along theC v axis, but enhances ionization along the D-D direction. I. INTRODUCTION
The ionization of molecules in strong ultrafast laserfields is a fundamental subject in laser-matter interac-tions, and also has relevance for laser chemistry andlaser-induced plasmas. , Strong-field molecular ion-ization is affected by vibrations and rotations: Laser-induced bond stretching tends to enhance ionization, while laser-driven bond alignment changes the polariza-tion angle-dependent laser-electron coupling. Becauseof the strong-field coupling to these additional degrees offreedom, molecular ionization cannot be understood fromsimple low frequency tunneling theories developed foratomic systems. Modifications to tunneling ionizationappropriate to molecular orbitals have been proposed, but these become complicated in lighter molecules, wherestrong-field interactions with bonding orbitals can initi-ate fast nuclear motion leading to new phenomena suchas induced multiple ionization. To date, almost all experiments investigating the in-terplay between nuclear motion and ionization have in-volved linear molecules or symmetric tops . An ex-ception is water, where strong-field induced electron re-moval from the inner-valence HOMO-1 bonding orbitalinitiates bending motion on the timescale of 800 nm laserfield oscillations. Studies comparing D O and H O haveshown that this affects the spectrum in high harmonicgeneration (HHG). In this paper we have studied the dissociative mul-tiple ionization of water in strong 800 nm focused laserfields. We examine dissociation channels where two orthree electrons have been removed, and reconstruct thefull momentum of all dissociating ions in coincidence, aswell as the kinetic energy released in the coulomb explo-sion. Our results show how nuclear bending motion in a) Electronic mail: [email protected] the water molecule leads to dramatically enhanced mul-tiple ionization in particular geometries, validating thegeneral predictions of strong-field enhanced ionization. The enhancement is strongest when the laser is polarizedalong the D-D axis, where it couples most strongly toelectrons in the HOMO-2 valence orbital (see Fig. 1).At the same time, we find that when the laser is polar-ized along the C v axis where it couples most strongly tothe HOMO-1 orbital, multiple ionization is nearly com-pletely suppressed. While this second finding seems tocontradict the general principles of enhanced ionizationin strong fields, it may be due to a laser-induced shift in FIG. 1. The three valence orbitals of D O are shown withthe laser polarization aligned along the direction that has thehighest probability of strong field ionization. The dominantchannel is found to be O+/D+/D+ at each intensity stud-ied in this paper. Hits from this channel are shown on thedetector. The molecular frame analysis suggests that theO+/D+/D+ channel is driven by alignment with the D-Dbond, indicating the importance of the HOMO-2 orbital. orbital energy and geometry as the molecule unbends.
II. EXPERIMENTAL METHODS
The experiments are performed in a vacuum chamberat a base pressure of 3 × − mbar fitted with elec-trostatic focusing elements and a Roentdek time- andposition-sensitive hex-anode delay-line detector. Thetarget is D O gas at 300K in the tight focus of a 40 fs,800 nm laser pulse. The focal volume and gas densityyield fewer than one ionized molecule per laser pulse onaverage. The electrostatic lenses direct all charged frag-ments to the detector, and the arrival time and positionof each ion determines its mass and momentum.The intensity in the laser focus is scanned between7 ×
14 W / cm and 2 ×
15 W / cm , calibrated by compar-ing charge state ratios of argon in the same focus. Thelaser propagation axis and polarization are in the detec-tor plane, as shown in Fig.1. Data used in this analysiswere accumulated over 250 million laser pulses at a 1 kHzrepetition rate (69 hours). All ion hits were separatelyrecorded for later analysis.
III. RESULTS AND ANALYSIS
The data were analyzed to identify all ionization chan-nels that produce 2 + or 3 + charge states of D O. Falsecoincidences, in which ions from more than one moleculewere detected in the same shot, were efficiently elim-inated by requiring that the momenta sum to nearlyzero. This also misses partially ionized channels, includ-ing three-particle breakup where one of the particles wasneutral such as O/D + /D + ; but those can be recoveredin a different way, described below. Of the four possiblechannels with non-neutral fragments, only two are seen:OD + /D + and O + /D + /D + . Both channels can be seeneasily in Fig. 2a, which plots the kinetic energy release(KER) against the angle θ between the momentum vec-tor of one of the D + particles and the laser polarization.These two channels occupy distinct regions in this plot.The two-body dication decay O + /D +2 is not seen,although it been noted in previous excitation studies:O + /H +2 is a significant decay channel in 1 s core excta-tion of water, and also occurs in VUV excitation of thewater cation. This channel has also been reported ina previous strong-field ionization study, but the branch-ing fraction was small, and could only be seen for theshortest pulses. We also looked for the dication breakup channelO/D + /D + . Although the two-body decay OD + /D + isthe most energetically favorable dissociation for doubleionization, the three-body O/D + /D + channel is knownto be a significant dissociation pathway from some ex-cited states of the dication. For example, studies ofsingle-photon double ionization of HOD at 40 eV observeO/H + /D + breakup with a branching ratio of approxi-mately 20%. . If both of the ionized electrons are re-moved from inner valence orbitals via a strong-field pro-cess, then we might expect to see the O/D + /D + channel. Such partially ionized channels may be identified bytheir KER and orientation dependence, since the neutralparticles are not detected. To look for this, we plot allD + /D + coincidences in Fig.2b. To filter out false coinci-dences with H +2 , only events where each particle has morethan 300 meV of kinetic energy are shown. Additionally,coincidences where the momentum sum of the D + par-ticles is zero are shown in grey, and all other events areshown in red. One prominent feature are coincidencesfrom O + /D + /D + , which can be seen in red at high KER.There is one other channel at lower KER in grey, indi-cating the momentum sum is zero for these events. Thedistribution is practically identical to that of the H + /H + channel measured in the same experiment, which canonly be from the coulomb explosion of H +2 . Therefore,the channel in grey is from the coulomb explosion of D +2 ,which is degassed from the heavy water sample. Thereare no other features in Fig.2b indicating the presence ofa O/D + /D + channel at this particular intensity, or forthe range of intensities studied in this experiment.The dominance of the OD + /D + channel suggestsionization to either the ground state or the first ex-cited state in the dication, since these are the statesthat dissociate predominantly to OD + /D + . These twostates have electronic configurations with HOMO vacan-cies: ... (1 b ) (3 a ) (1 b ) and ... (1 b ) (3 a ) (1 b ) re-spectively. Therefore we expect these fragmentationproducts following strong field ionization of electronsfrom the HOMO.The angle-dependent coupling of the molecular orbitalsto the laser in the low-frequency limit is proportional to F ˆ ǫ · ~µ , where F ˆ ǫ is the laser field vector and ~µ is thedipole moment of the orbital. For the p -type valenceorbitals of water, the dipole moments are perpendicularto the nodal plane of the orbital. As shown in Fig.1, thismeans that the (1 b ) HOMO orbital couples to the laserpolarization along ˆ x M , perpendicular to the molecularplane; the (3 a ) HOMO-1 orbital couples to the laser po-larization along ˆ z M , the C v axis; and the (1 b ) HOMO-2orbital couples to the laser polarization along ˆ y M , the D-D axis. These three perpendicular axes ˆ x M , ˆ y M , and ˆ z M in the molecular frame define the molecular alignment.For the OD + /D + channel, Fig.2a shows that the disso-ciation axis is predominantly perpendicular to the laserpolarization, i.e. the strong field of the laser is orthog-onal to the molecular plane. Since the polarizability ofthe lone-pair HOMO is along this direction, ionizationfrom this orbital likely plays a role (see Fig.1). Ioniza-tion from the HOMO is involved in the excitation of thelowest dication states. Furthermore, the KER of theOD + /D + channel is similar to that found in PIPICOstudies using single photon excitation near the verticalexcitation energies of the lowest dication states. TheKER also has a relatively narrow distribution. This in-dicates that there is minimal nuclear motion during theionization. Indeed, removal of a HOMO electron preparesthe cation in the ground state which has a very similarequilibrium geometry to the neutral ground state.
Therefore, alignment of the laser field out of the molec-ular plane drives ionization of the HOMO, and leads toOD + /D + dissociation.While OD + /D + dissociation is driven by alignment FIG. 2. a) True coincidences from the OD + /D + (blue) andO + /D + /D + (red) channels at an intensity of 2.1I , whereI = 7 ×
14 W / cm . The kinetic energy release (KER) isplotted against θ , the angle between the laser field and thetrajectory of a randomly selected D + particle from the event.True coincidences are sorted for via momentum sums. b) θ versus the kinetic energy sum of D + fragments is plotted forall D + /D + coincidences at 2.1I . Grey indicates events wherethe momentum sum goes to zero, and the KER is less than12 eV. These events are attributed to coulomb explosion ofD , which is present in the D O sample. All other events arein red. The main contribution is from O + /D + /D + , as seenby comparison with Fig.2a. There are no clear signatures ofO/D + /D + channels at lower KER. of the strong field normal to the molecular plane,O + /D + /D + dissociation is driven by alignment of thefield in the molecular plane. To show dependence onalignment, we reconstruct the molecular frame for eachevent by taking advantage of the symmetry of the dis-sociation. The orientation with respect to the laser fieldcan then be determined. Fig.3a shows the normalizedO + momentum versus the difference of normalized D + momentum for each event in the channel. The verticalline at a D + momentum difference of zero indicates thatthe breakup is symmetric. Because of the symmetry, wecan use the momenta of the fragments to reconstruct themolecular axes in a straightforward manner.The molecular axes in the lab frame are then deter-mined by the O + momentum (ˆ z M ), and the cross prod-uct of D + momenta (ˆ x M ). ˆ y M is found by completingthe right handed coordinate system. For each event, thelaser polarization is oriented within the molecular frame. FIG. 3. a) The O + momentum norm is plotted against thedifference between D + momentum norms for the O + /D + /D + channel at an intensity of 2.1 I . Both axes are normalized tosum of all momentum vector norms, P tot . The vertical linecentered at zero indicates a symmetric breakup. As the rela-tive O + momentum approaches zero, the molecule dissociatesin an increasingly linear geometry. b) Bivariate histogramof strong field orientation in the molecular frame for eachO + /D + /D + event at 2.1I . This channel has a strong pref-erence for alignment along the D-D bond. Orientation of thelaser field with respect to the molecule is described by a polarangle, α M , and an azimuthal angle β M . Fig.3b shows the results, along with a description of thespherical coordinate system used to describe the orien-tation of the laser field. The analysis shows that thepolarization is primarily in the molecular plane for theO + /D + /D + channel, and preferentially along the D-Dbond. Therefore, in plane alignment of the strong fieldleads to triple ionization and a symmetric dissociation,and out of plane alignment leads to double ionization anda asymmetric fragmentation into OD + /D + .We have seen that the field orientation plays a largerole in the yield for both channels. More strikingly, dou-ble ionization is significantly less probable than triple ion-ization over the whole range of intensities used here. Theprobabilities of the OD + /D + and O + /D + /D + channelsacross different intensities are shown in Fig. 4. Sincethese are the only two channels with appreciable yieldfor each intensity, and there are no stable states of thedication, the yield of each channel gives the probabil-ity for double and triple ionization. Triple ionization isalways more probable than double ionization, even be- FIG. 4. Probability for OD + /D + and O + /D + /D + in D Oversus intensity. Since OD + /D + and O + /D + /D + are the onlydissociation channels that were detected, these are the prob-abilities for double and triple ionization. Surprisingly, evenbelow saturation for double ionization, triple ionization is al-ways more likely. Intensity is in units of I = 7 ×
14 W / cm . fore the double ionization saturation point between 1.5I and 2.0I . This is a dramatic departure from what isexpected in low frequency, strong field ionization. Be-low saturation for removing N-1 electrons, ionization ofN-1 electrons should be more likely than ionization of Nelectrons. This is not the case for D O.The unexpectedly high yield for the O + /D + /D + chan-nel could be explained by enhanced ionization. Enhancedionization is driven by the laser field’s interaction withthe polarizability of the molecule. The interaction, whichis strongest when the field is aligned with the polarizabil-ity, must occur along nuclear degrees of freedom. For theOD + /D + channel, the laser is predominantly orientedalong ˆ x M , so cannot effectively induce nuclear motion.Conversely, O + /D + /D + is driven by in plane alignmentof the field. Therefore, the OD + /D + channel does notinvolve enhanced ionization, while the effect is likely ac-tive for O + /D + /D + . Evidence of enhanced ionizationhas been seen before in water, where stretching ofthe angular bond was shown to be important. Fig. 5ashows the distribution of dissociation angles, γ , for theO + /D + /D + channel. This dissociation angle is related tothe bond angle upon coulomb explosion. Significant an-gular stretching can be seen to occur, especially at higherintensity. Since the laser field is predominantly alignedto the D-D bond in this channel, it can interact with thepolarizability of the HOMO-2 and cause the molecule tobecome more linear.Sub-structure can be seen in the γ distribution, indi-cating multiple pathways to enhanced ionization. Fig. 5bshows that the KER varies with γ in a non-monotonicmanner. This further suggests that there are differentpathways to enhanced ionization with unique critical val-ues for both angular and O-D bonds.Although enhanced ionization may explain the highyield of O + /D + /D + relative to OD + /D + , it does notexplain why O/D + /D + does not have appreciable prob- ability. Across all intensities, there is scare evidence ofO/D + /D + , much like in Fig.2b. This channel is knownto be a significant pathway for dication states that are ex-cited by the removal of electrons from lower orbitals. Since the laser field is in the molecular plane for enhancedionization, ionization involves removal of the electronsfrom inner orbitals of D O, so the dications created arein excited states. Below the saturation point for tripleionization, which is around 2.1I according to Fig. 4,there should be appreciable probability for O/D + /D + .Since this is not seen, removal of a third electron may bedriven by autoionization, or multiphoton resonant ion-ization pathways.Enhanced ionization is also suppressed for field align-ment along ˆ z M , as shown in Fig.3b. This is peculiar sincefield alignment along ˆ z M would allow for stretching of theOD bonds due to interaction with the HOMO-1. Studiesof isoelectronic H S in high intensity 800 nm pulses haveshown that ionization rates are also significantly lowerwith the laser polarization along the symmetry axis. Effects such as the re-ordering of orbital energies dueto the strong field driven nuclear distortion should beconsidered in understanding the suppression of enhancedionization in bent triatomics along a particular in-planeaxis.
FIG. 5. a) The dissociation angle, γ , measured in theO + /D + /D + channel for 2.1I , related to the bond angle upondissociation. There is a substructure indicating multiple dis-sociation pathways, with most pathways involving increase inthe bond angle. b) The mean KER for different dissociationangles obtained by Gaussian fits to the KER spectrum for 10degree bins. The non-monotonic scaling with angle furtherindicates there are different subchannels.
IV. CONCLUSION
In this work, we saw two dominant ionization pathwaysfor strong-field double and triple ionization from D O: anasymmetric OD + /D + channel driven by laser field align-ment orthogonal to the molecular plane, and a symmetricO + /D + /D + channel driven by field alignment along theD-D bond. Triple ionization was found to be more likelythan double ionization, even below the saturation inten-sity for double ionization. This is partly due to enhancedionization only being active for the O + /D + /D + channel,where the laser field can effectively induce bond stretch-ing. However, it is curious why the O/D + /D + channelcannot be a product of enhanced ionization as it is a pri-mary dissociation pathway in the dication. Mechanismssuch as autoionization or multiphoton resonant ioniza-tion might be involved in the ionization process whenthe strong field is in the molecular plane. Furthermore,the apparent suppression of enhanced ionization alongthe symmetry axis is intriguing. It is worth investigatingif suppression of enhanced ionization along a particularaxis is a general feature in bent triatomics.This work was supported by the National ScienceFoundation under Grant No. PHY-0649578. T. Seideman, M. Y. Ivanov, and P. B. Corkum,Physical Review Letters , 2819 (1995). E. Constant, H. Stapelfeldt, and P. B. Corkum,Physical Review Letters , 4140 (1996). X. Gong, Q. Song, Q. Ji, H. Pan, J. Ding, J. Wu, and H. Zeng,Physical Review Letters (2014), 10.1103/PhysRevLett.112.243001. X. Gong, Q. Song, Q. Ji, H. Pan, J. Ding, J. Wu, and H. Zeng,Physical Review Letters (2014), 10.1103/PhysRevLett.112.243001. C. Huang, Physical Review A (2014), 10.1103/PhysRevA.90.043420. W. Lai, Physical Review A (2016), 10.1103/PhysRevA.93.043401. B. Schtte, P. Ye, S. Patchkovskii, D. R. Austin, C. Brahms,C. Strber, T. Witting, M. Y. Ivanov, J. W. G. Tisch, and J. P.Marangos, Scientific Reports , srep39664 (2016). T. Zuo and A. D. Bandrauk,Physical Review A , R2511 (1995). H. Stapelfeldt and T. Seideman,Reviews of Modern Physics , 543 (2003). X. M. Tong, Z. X. Zhao, A. S. Alnaser,S. Voss, C. L. Cocke, and C. D. Lin,Journal of Physics B: Atomic, Molecular and Optical Physics , 333 (2005). M. Ammosov, N. Delone, and V. Krainov, Sov. Phys. JETP ,1191 (1986). X. M. Tong, Z. X. Zhao, and C. D. Lin,Physical Review A (2002), 10.1103/PhysRevA.66.033402. L. Christiansen, Physical Review A (2016), 10.1103/PhysRevA.93.023411. C. Cornaggia, Journal of Physics B: Atomic, Molecular and Optical Physics , 19LT01 (2016). J. P. Farrell, S. Petretti, J. Frster, B. K. Mc-Farland, L. S. Spector, Y. V. Vanne, P. De-cleva, P. H. Bucksbaum, A. Saenz, and M. Ghr,Physical Review Letters (2011), 10.1103/PhysRevLett.107.083001. O. Jagutzki, A. Cerezo, A. Czasch, R. Dorner,M. Hattas, M. Huang, V. Mergel, U. Spill-mann, K. Ullmann-Pfleger, T. Weber, and et al.,IEEE Transactions on Nuclear Science , 24772483 (2002). W. A. Bryan, S. L. Stebbings, J. McKenna, E. M. L. English,M. Suresh, J. Wood, B. Srigengan, I. C. E. Turcu, J. M. Smith,E. J. Divall, C. J. Hooker, A. J. Langley, J. L. Collier, I. D.Williams, and W. R. Newell, Nature Physics , 379 (2006). J. Laksman, E. P. Mnsson, A. Sankari,D. Colin, M. Gisselbrecht, and S. L. Sorensen,Physical Chemistry Chemical Physics , 19322 (2013). H. B. Pedersen, C. Domesle, L. Lammich, S. Dziarzhyt-ski, N. Guerassimova, R. Treusch, L. S. Harbo,O. Heber, B. Jordon-Thaden, T. Arion, M. Frs-tel, M. Stier, U. Hergenhahn, and A. Wolf,Physical Review A (2013), 10.1103/PhysRevA.87.013402. F. A. Rajgara, A. K. Dharmadhikari, D. Mathur, and C. P.Safvan, The Journal of Chemical Physics , 231104 (2009). Z. Streeter, F. Yip, D. Reedy, A. Landers, and W. McCurdy,Bulletin of the American Physical Society , D1.00135 (2017). P. J. Richardson, J. H. D. Eland, P. G. Fournier, and D. L.Cooper, The Journal of Chemical Physics , 3189 (1986). B. Gervais, E. Giglio, L. Adoui, A. Cassimi, D. Duflot, and M. E.Galassi, The Journal of Chemical Physics , 024302 (2009). T. J. Van Huis, S. S. Wesolowski, Y. Yamaguchi, and H. F.Schaefer, The Journal of Chemical Physics , 11856 (1999). K. Nobusada and K. Tanaka,The Journal of Chemical Physics , 7437 (2000). D. Winkoun, G. Dujardin, L. Hellner, and M. J. Besnard,Journal of Physics B: Atomic, Molecular and Optical Physics , 1385 (1988). F. Schneider, F. Di Giacomo, and F. A. Gianturco,The Journal of Chemical Physics , 7560 (1996). C. R. Brundle and D. W. Turner,Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences , 27 (1968). J. H. Sanderson, A. El-Zein, W. A. Bryan, W. R. Newell, A. J.Langley, and P. F. Taday, Physical Review A , R2567 (1999). F. Lgar, K. F. Lee, I. V. Litvinyuk, P. W. Dooley,S. S. Wesolowski, P. R. Bunker, P. Dombi, F. Krausz,A. D. Bandrauk, D. M. Villeneuve, and P. B. Corkum,Physical Review A (2005), 10.1103/PhysRevA.71.013415. D. L. Cooper, J. Gerratt, M. Raimondi, and M. Sironi,The Journal of Chemical Physics , 1666 (1987). A. Hishikawa, E. J. Takahashi, and A. Matsuda,Physical Review Letters97