1

ANewElectron‐IonCoincidence3DMomentum‐ImagingMethodandItsApplicationinProbingStrongFieldDynamicsof2‐Phenylethyl‐N,N‐Dimethylamine

LinFan1,SukKyoungLee1,Yi‐JungTu1,BenoîtMignolet2,DavidCouch3,KevinDorney3,QuynhNguyen3,LauraWooldridge3, MargretMurnane3, Françoise

Remacle2,H.BernhardSchlegel1andWenLi1*

1DepartmentofChemistry,WayneStateUniversity,Detroit,48202 2DepartmentofChemistry,B6c,UniversityofLiege,B4000Liege,Belgium

3JILAandUniversityofColoradoatBoulder,Boulder,CO,80303wli@chem.wayne.edu

We report thedevelopmentof a new three‐dimensional (3D)momentum‐imagingsetup based on conventional velocity map imaging (VMI) to achieve coincidencemeasurementofphotoelectronsandphoto‐ions.This setupusesonlyone imagingdetector(microchannelplates/phosphorscreen)butthevoltagesonelectrodesarepulsedtopushbothelectronsandionstowardthesamedetector.Theion‐electroncoincidenceisachievedusingtwocamerastocaptureimagesofionsandelectronsseparately. The time‐of‐flight (TOF) of ions and electrons are read out fromMCPusing a digitizer. We demonstrate this new system by studying the dissociativesingle and double ionization of PENNA (2-phenylethyl-N,N-dimethylamine). Wefurthershowthecamera‐based3Dimagingsystemcanoperateat10kHzrepetitionrate.

Introduction

http://dx.doi.org/10.1063/1.4981526

2

The coincidence detection of photoelectrons and photoions arising from a singleatom/molecule in gas phase is a powerful tool for untangling multi‐channelionization/dissociation dynamics. The early implementation of coincidencetechnique only provided energy information of each particle using time‐of‐flightmethods.1‐4 Various position‐ and time‐sensitive detectors were introduced tomeasureboththepositionandtheTOFandthusenabled3Dmomentumdetectionof all particles.5‐12 Successful coincidence detection requires the count rate of theexperiment to be less than one event per driving pulse (electron/ion/photon) tosuppress false coincidence events. Velocity mapped imaging13‐15 was not initiallydevelopedtoachievecoincidencedetectionbecauseitdidn’tprovidehighresolutionTOF informationof individualparticles. Instead, itusedan imagingdetectorandacamera to measure the positions of particles with a relative large TOF range(nanosecond to one microsecond), which identifies the mass of the particles.Nonetheless,VMIhasbecomeaverypopularmethodinreactiondynamics,ultrafastspectroscopy and other fields, because of its relatively easy implementation andsimultaneousmeasurementofenergyandangulardistributionofchargeparticles.Recently, a new type of VMI imaging system was developed to achieve 3Dmomentum detectionwith a conventional imaging detector.16,17 In this system, astandard video camera was replaced with a fast frame CMOS camera (>1kFrames/s)andthecameraexposureissynchronizedwiththelaserpulses(>1kHz)togetherwithawaveformdigitizer,whichreadsouttheMCPpulsesignal.Inthesecoincidenceexperiments, because the count rate is lowso that a true coincidencebetweenhitpositionsfromthecameraandTOFsextractedfromthedigitizercanbe

http://dx.doi.org/10.1063/1.4981526

3

establishedtoyieldthreecoordinates(X,Y, t)requiredfor3Dmomentum(Px,Py,Pz)imaging.Thedemonstratedtimeresolutionforthissystemisexcellent(~30ps).With this system, sliced velocitymapping of photoelectronswas achieved for thefirst timewith an imagingdetector. Itwas also demonstrated that photoelectron‐photoion coincidence could be achieved by accelerating electrons and ionssimultaneously inadouble‐sidedVMIsetup towards two imagingdetectorsat theopposite ends of the spectrometer and applying the same 3D measurementscheme.18 However, a conventional VMI apparatus features only one imagingdetectorandauni‐directionalspectrometer.Isitpossibletoconvertsuchapparatustoaphotoelectron‐photoioncoincidence imagingapparatus? In thiswork,wewillshowthatthisispossibleandthesolutionisquitesimplewiththecamera‐based3Dimagingsystem.

PENNAisabifunctionalmoleculewitha‐CH2‐CH2‐bridgelinkingthechromophoregroup(phenyl)andaminegroup.Inthepastdecade,therehavebeengreatinterestsinstudyingthephotoionizationdynamicsofPENNA.ThisismainlybecausePENNAis oneof the firstmolecules thatwere identified to support anew typeof chargemigrationprocessthattakesplaceatanextremelyshorttimescale(afewhundredsof attoseconds to a few femtoseconds) 19‐24. This charge transfer arises from acoherentsuperpositionofafewelectronicstatesofthecationanditoccursbeforeany significant nuclear motion. Resonant two-photon ionization (R2PI) and RydbergFingerprintSpectroscopyhavebeenusedtostudytheslowerintramolecularchargetransfer (~80 fs) at different laser wavelengths. 19,22,25,26 However, such studieswere limited to tens of femtosecond time resolution. A strong field IR‐pump‐IR‐

http://dx.doi.org/10.1063/1.4981526

4

probemethodhasbeenproposedtoobservetheultrafastchargemigrationwitharesolution of few femtoseconds.24 So far experimental investigations on thephotoionization dynamics of PENNA under intense laser field have been scarce.Furthermore,thedynamicsofdissociationfollowingionization(singleanddouble)are crucial for measuring molecular/recoil frame ionization rates because theyshowwhether the axial recoil approximation is valid or not.With the axial recoilapproximation,therecoil‐frameionizationratemeasurementismuchsimplerthanthosemethodsthatrequirepre‐alignmentofmolecules.Here,weappliedournew3D electron‐ion coincidence technique to study the strong fieldionization/dissociationofPENNAwithonelaser.Wehaveidentifiedthedissociationpathways following the strong field double ionization of PENNA. This result willprovidebackgroundinformationforfuturetime‐resolvedstudies.

Experimentalsetupandcomputationalmethods

Experimentalsetup

AnamplifiedTi:Sapphire femtosecond lasersystemwasusedto ionizethePENNAmoleculesinmid‐intensitystrongfield(~8×1013W/cm2).Thewavelengthwas800nm and the repetition ratewas1500Hz. The ultrashort laser pulse (~30 fs)waslinearly polarized along the time‐of‐flight axis. For the one‐camera setting, theexperiment was carried out with circularly polarized laser. PENNA (purity 98%)waspurchasedfromSigma‐Aldrichanditwasseededinheliumtobesentintothesourcechamberbyagas jetwith20μmorifice.ThePENNAsamplewasheated to~30°Candthegas jetwasheatedto~70°Cfromoutsideofthesourcechamber.A

http://dx.doi.org/10.1063/1.4981526

5

skimmer with 0.5 mm orifice was used to skim the center part of the expandedmolecularbeambeforeitenteredthemainchamber.Thelaserbeamwassentintothemainchamberperpendicularlywith themolecularbeamandthe time‐of‐flightaxis, and itwas focusedonto themolecular beamby a concavemirror (f =5 cm)insidethemainchamber.Afour‐plateelectron‐ionopticswasbuilttovelocitymaptheelectronsandionsarisingfromstrong‐fieldionization.AshortTOFlength(~10cm) was adopted to allow detection of high‐energy electrons. The timings ofelectronsandionshittingthedetectorwerepickedofffromthefrontMCPplatebyahigh‐speeddigitizer(NationalInstruments,PXIe5162)throughasignal‐decouplingcircuit. The maximum sampling rate of the digitizer is 5 GHz. The positions ofelectrons hitting the detector were recorded from the phosphor by acomplementary metal‐oxide semiconductors (CMOS) camera (Basler, acA640‐750μm).Thiscamera(e‐)wassetat~65cmawayfromthedetector,pointingatthecenter of phosphor screen with a normal angle. Another CMOS camera (XIMEA,MQ013MG)was employed to record the ionpositions. Itwas located at~120 cmawayfromthedetectorandslightlyoff‐centerofthescreen.Thiscamera’spointingdirectionhada~6°angledeviation to thescreennormal.The justification for theuseoftwocamerasisdiscussedintheResultsandDiscussionsection.AschematicoftheexperimentalsetupisshowninFigure1(A).

Computationalmethods

ElectronicstructurecalculationswerecarriedoutwiththedevelopmentversionofGaussian27 using the wB97XD functional.28 Relaxed potential energy scans were

http://dx.doi.org/10.1063/1.4981526

6

executed with the 6‐31G(d) basis set.29, 30 The geometries of neutral PENNA,PENNA+, and PENNA2+, transition states, products were further optimized by 6‐311++G(d,p)basissetandtheirSCFenergieswerealsocalculatedwiththisbasisset.All optimized structures were checked by normal mode vibrational analysis, andwavefunctionsweretestedforSCFstability.ForPENNAdications,theidentitiesoftheopen‐shellsingletandtripletelectronicconfigurationswereconfirmedbyspin‐squaredexpectationvalues()andspindensitypopulations.GaussView31wasusedtovisualize isodensityplotsof thespinpopulations(isovalue=0.004au).ToexplorethepotentialenergysurfacesfordissociationofPENNA2+,relaxedpotentialenergysurfacescanswereperformedbystretchingtheC–Cbondandoptimizingtheremainingcoordinates.Thetransitionstatestructuresontheopen‐shellsingletandtriplet surfaces of the dication were confirmed to have only one imaginaryfrequencybyvibrationalmodeanalysis.Toexploretheclosed‐shellsingletpotentialenergy surface of the dication,we startedwith ground state geometry of PENNA,vertically ionized to the closed‐shell singlet state of PENNA2+ and tracked thedissociationusingDampedVelocityVerlet(DVV)reactionpathfollowing.Uptothetransition state, the closed shell singlet dication calculations had a closed‐shell toopen‐shell instability and the wavefunction optimized to the lower energy open‐shellsingletdication.

The photoelectron spectrumhas been computed bymodeling the photoexcitationand photoionization dynamics induced by a 22 fs 800nm IR pulse with a fieldstrength of 1012W/cm2 in the PENNA molecule at a frozen geometry. The time‐dependentSchrödingerequationwasnumericallyintegratedusingabasisofneutral

http://dx.doi.org/10.1063/1.4981526

7

andionizedelectronicstates.24,32Theelectronicstructuresofthelowest30neutralstates were computed at the TDDFT wB97xD/6–311++G(d,p) level. A densemanifold of excited states is required to accurately describe the multiphotonexcitation and ionization of the PENNA ground state. The ionized states aredescribedastheanti‐symmetrizedproductofthefield‐freecationicelectronicstatesand the wavefunction of the ionized electron described by a plane wave. Thisapproximation can affect the photoelectron spectrum, especially at low kineticenergybecausetheinteractionbetweentheionizedelectronandthecationiccoreisneglectedandoverthebarrierandtunnelionizationarenotaccountedfor.ThetenlowestcationicstatescomputedatthewB97xD/6–311++G(d,p)level.Theionizationcontinuaarediscretizedbothinenergy(from0to25eV)andangulardistribution.Intotal, thereare18000ionizedstatesand30neutralstatesinthebasis.Toaccountfortherandomorientationofthemolecules,wecomputedthedynamicsforasetof50 randomly oriented molecules. From the amplitudes of the ionized states, wecomputedthephotoelectronspectrumaveragedovertheorientation.

ResultsandDiscussion

Achieving3Dcoincidencemeasurementwithasingleimagingdetector

ToupgradeaconventionalVMItoacoincidence3Dmomentumimagingapparatus,we need to address two issues associated with it. The first issue is the uni‐directionalspectrometerofaconventionalVMIapparatus.Itcannormallyberunineither photoelectron or photoion mode by applying different voltages on theelectrodes.To imageboth ionsandelectronsat thesamedetector, thevoltagesof

http://dx.doi.org/10.1063/1.4981526

8

electrodesshouldbeswitchedveryquicklytoacceleratebothparticlestowardthesamedirection.Thishasbeendemonstratedby Janssenandcoworkersbypulsingthe electrodes and employing a delay line detector.33 Owing to the large massdifferencebetweenelectronsandions,theelectroncanmaintainthesameimagingcondition as with non‐pulsing electrodes while ions suffer minimum momentumblurring (SIMIONsimulation~1%).Once ions andelectronsarrive at the imagingdetector, they both produce flashes on the phosphor screen, which are thencapturedbythecamerainthesameframe.However,howtoassociatethepositionsandmeasured TOFs is not trivial. Previously, in either ion or electronmode, thebrightnessofthecameraflashshowsastrongcorrelationwiththeintensityoftheTOFpeakinthedigitizedwaveform.16Thiswasexploitedtoassociatethepositionsand TOFs in a multi‐hit event. In Figure 1(B), it is shown that because thecorrelationslopesaredifferentforelectronsandions,thisschemecanleadtosomemis‐assignment of ions in the electron image (or vice versa) when processingframescontainingbothelectronsandions.

Figure 1. (A) The schematic of the two‐camera VMI setup that is capable ofcoincidence 3D momentum‐imaging. The electrodes are pulsed to push both

http://dx.doi.org/10.1063/1.4981526

9

electronsand ions toward the imagingdetector. (B)Withaone‐camerasetup, theseparation between electrons and ions are not complete. The main features arephotoelectrons from strong field ionization of PENNA using circularly polarizedlight.

The solution to this is to addone additional camera. Because the largedifferencebetweentheTOFsoftheelectrons(1us),onecameracanbetriggeredtoonlyexposeforthefirst200nsafterthelaserpulsetocaptureelectronswhilethesecondcamerastartstoexposeafter500nstoonlyimageions.Withthisconfiguration(Figure1(A)),theassociationbetweenthetimeandpositionof the charged particles becomes quite easy and self‐evident: the positions ofparticles intheelectroncameraareassociatedwiththeelectronTOFswhilethosepositions on the ion camera with the ion TOFs. Both TOFs are measured bydigitizing theMCPoutputwith a single high‐speeddigitizer. If there aremulti‐hiteventssuchasindissociativedoubleionization,thebrightness‐intensitycorrelationcannowbeappliedtoeachcameraframeseparately,asshownpreviouslyineitherelectron or ion detectionmode.With this new scheme, we can achieve completeseparation of electrons and ions and this enables ion‐electron coincidence 3Dimaging measurement for the first time using a single imaging detector. Weestimated thespatialand temporal resolutions tobe6%and1% for ionsand3%and3%forelectrons,respectively.InFigures2and4weshowedtheresultsofion‐electrondoublecoincidencemeasurementofPENNAsingleionizationandelectron‐ion‐iontriplecoincidenceofPENNAdissociativedoubleionization.

http://dx.doi.org/10.1063/1.4981526

10

SingleionizationofPENNA

Previous ion mass spectrum of PENNA in an intense laser field showed thatdimethylaminemonocation(N(CH3)2CH2+,mass58)isthemajorproductwhiletheyieldofparentionismuchlowerthanN(CH3)2CH2+(~0.05),whichsuggestedmostparent cations are unstable.34 Figure 2(A) shows the momentum distribution ofN(CH3)2CH2+intheplaneperpendiculartothelaserpolarizationwhileFigure2(B)and (C) show the momentum distributions of electron in coincidence withN(CH3)2CH2+ in the plane of perpendicular and parallel to the laser polarization,respectively.Theelectronenergydistribution isshowninFigure3(A). Ithasbeenshowed previously that the single ionization of PENNA could populate threedifferentelectronic statesD0,D1andD2,with thedifference in ionizationenergiesbeinglessthanonephotonenergy(1.6eV).34ThedissociationsoftheexcitedstatesD1andD2arethroughaseriesofconicalintersectionstotheD0stateandthemainproductisN(CH3)2CH2+.Thekineticenergyrelease(KER)ofionmatchedwellwiththatofcalculations.34InFigure3(A),weshowtheelectronenergydistributionandits comparison with the photoelectron spectrum resulting from frozen geometrysimulationsoftheelectrondynamicsofrandomlyorientedmolecules,whichtakesintoaccountthemultiphotonexcitationandionizationbythelaserpulse,formoremethoddetails see ref. 24Theagreement isgood forhighkineticenergyelectrons(above 4 eV). The discrepancy at low kinetic photoelectron energy can beunderstood from the limitations of themodel used in the simulations. Themodelonly includes themultiphoton photoionization process and neglects the Coulombinteractionbetweenionizedelectronandioniccore,whichisknowntoproducelow

http://dx.doi.org/10.1063/1.4981526

11

energy electrons by over‐the‐barrier or tunneling ionization. This result furthersuggests that in strong field ionization, the photoelectron spectrum alone isinsufficienttoidentifytheproducedcationelectronicstates.Acompletetheoreticalmodeling of the strong field photoionization of big molecules such as PENNA iscurrentlyoutofreach.

Figure 2. (A) XY momentum distribution of dimethyl amine monocation(N(CH3)2CH2+). (B) XY momentum distribution of photoelectrons in coincidencewithdimethylaminemonocation(N(CH3)2CH2+).(C)Ytmomentumdistribution(tisthe TOF axis) of coincidence photoelectron in coincidence with dimethyl aminemonocation(N(CH3)2CH2+).

With the electron‐ion coincidence measurement capability, we can produce thecorrelation map between the KER of the dissociated fragments and the electronkinetic energy (eKE) and this is shown in Figure 3(B). Interestingly, an apparentcorrelationcanbeseen(thediagonalstructure).However,becausetheenergyscaleis different between the KER and eKE, it is unlikely that an energy conservationmechanism (as in single photon ionization) is at play. Upon further investigation,such a structure persists even for non‐dissociative channels such aswater single

http://dx.doi.org/10.1063/1.4981526

12

ionization,whichsuggeststheobservedfeatureisnottruecorrelation.However,wedid observe significant difference in momentum distributions of electrons incoincidence with monocations and dissociative dications (see Figure 2(C) andFigure4(C)).Thisvalidatesourmethodfordetectingcoincidenceevents.

Figure 3. (A) The photoelectron spectroscopy of single ionization (black) andcomparison with theoretical simulations (red). (B) The energy correlation mapbetweenionKERandelectroneKE.

DoubleionizationanddissociationdynamicsofPENNA

PENNAdicationswerenotobservedinthemassspectrumandthissuggeststhatalldications dissociate after production by the strong laser field. Triple coincidence(ion‐ion‐electron)wasusedtoidentifythedissociationproducts.Wecouldemployquadrupole coincidence (ion‐ion‐electron‐electron)because the currentmethod iscapable of highly efficient detection of two electrons35. However, two‐electronmeasurements do not provide further insight to the scope of this study while

http://dx.doi.org/10.1063/1.4981526

13

require much longer acquisition time. Therefore, we will focus on the triplecoincidencedata(Figure4).

Figure 4. XY momentum distributions of coincidence ion pairs of (A). dimethylaminemonocation(N(CH3)2CH2+,mass58)and(B).benzylmonocation(C6H5‐CH2+,mass91).(C)Ytmomentumdistributionofphotoelectronsincoincidencewiththeionpairs.

It canbe readily identified from thephotoion‐photoion coincidence (PIPICO)map(Figure 5(A)) that the major dissociation channel leads to dimethyl aminemonocation(N(CH3)2CH2+,mass58)andbenzylmonocation(C6H5‐CH2+,mass91).Afterapplyingmomentumconservationcriteriatoremovefalsecoincidence,wecancleanly select the ion pairs that arise from the dissociation of PENNA dication.Figure 4(A) and (B) showmomentumdistributions of ionsmass 58 andmass 91after applyingmomentum conservation criteria in the plane perpendicular to thelaser polarizationwhile Figure 4(C) is themomentumdistribution of coincidenceelectronsintheplaneparalleltothelaserpolarization.Theangulardistributionsofboth ions are isotropic (spherical), which suggests that the dissociation time is

http://dx.doi.org/10.1063/1.4981526

14

longerthanrotationperiod.ThetotalKERoftheionpairsispeakedaround2.9eVandhasacutoffextendingbeyond4eV(Figure5(B)).

Figure 5. (A) Photoion‐photoion coincidence (PIPICO)map of dissociative doubleionizationofPENNA.(B)Thekineticenergyreleasedistributionofmass58and91ionpairsfromdoubleionization.

Nowweturntotheorytohelp identifythedicationstates thatare involvedinthedissociation processes. PENNAdication has three possible electronic structures:closed‐shellsinglet,open‐shellsinglet,andopen‐shelltriplet.Inordertofigureoutthe electronic structure of PENNA2+, transition states, and reaction pathways ofdouble ionization, density functional theory calculations were carried out. Westarted with the ground state of neutral PENNA that was confirmed in previousstudies. 19, 24 From the neutral PENNA molecule, we calculated the adiabaticallyoptimizedgeometriesofthePENNAmonocationanddications.Then,relaxedscansby stretching the C‐C bond of PENNA dications were employed to explore thedissociationpotentialenergysurfaces.

http://dx.doi.org/10.1063/1.4981526

15

Closed‐shell singlet PENNA dications are produced with two electrons removedfromthesameorbital,leadingtoanexcitedstatewhileopen‐shellsingletdicationshave two electrons removed from different orbitals, resulting in diradicals. TherelaxedscansbystretchingtheC‐CbondofPENNA2+showthatclosed‐shellsingletPENNAdicationsoverlapwiththesingletopen‐shellafteritgoesthroughtheenergybarrier, leading to the same products (Figure 6 inset). The energy of the tripletpotentialsurfaceishigherthanthesingletpotentialsurface(Figure6).Thesinglettransition state has a shorter C‐C bond (1.92Å) than the triplet transition state(2.17Å),andtheenergyis0.65eVlower.

Fromtheenergydiagram,wecanseethereversebarrierfortheopen‐shellsingletstate is 3.85 eV,which is close to themeasuredmaximum kinetic energy release(KER). The open‐shell triplet state has a reverse barrier of only 2.67 eV,which isevensmallerthanthepeakvalueofthemeasuredKER.Theopen‐shellsingletstatealsohas loweractivationbarrier(0.34eV) thanthatof theopen‐shell tripletstate(0.96 eV). Both facts prompt us to conclude that dissociation through open‐shellsingletisthedominantdissociationchannelofmetastablePENNAdications.Thisinturnalsosuggests thatstrong fielddouble ionizationproduceddominantlysingletopen‐shell dications.Closed‐shell dicationsareunlikely toplaya roledue to theirhigh excitation energies and barrierless reaction pathways, which should lead toanisotropicangulardistributionofthefragments.

http://dx.doi.org/10.1063/1.4981526

16

Figure 6. Energy levels of PENNA dication open‐shell states, transition states andfinalproducts.TheinsetshowstherelaxedpotentialenergysurfacesalongtheC‐Cbondfordifferentdicationsstates.

Achieving3Dmomentum‐imagingofelectronsat10kHzandbeyond

Finally, we demonstrate another major improvement of the camera‐basedmomentum‐imagingsystem.Foratypicalcoincidencemeasurement,becauseofthelowcount rate required forachieving true coincidence, it ispreferred tohave thewholesystemrunningatarepetitionrateashighaspossible.Thelimitingfactorisusually the laser system.WhileonekHz laser is verypopularandsuitable for thecamera‐based3Dmomentum‐imagingsetup,higherrepetitionlasersrunningat10kHzor100kHzdoexistandarebeingused inmanystrong fieldexperiments.Byexploiting a simple fact of all coincidence measurement, we can improve the

http://dx.doi.org/10.1063/1.4981526

17

repetitionrateofthecurrentimagingsystembyfivetotenfoldswithoutupgradingthecamera.The lowcountrateofacoincidenceexperimentmeans thatnoteverycameraframehasevent‐hits.Forexample,roughly80%of thecamera frameswillbewithoutevent‐hitifthecameraframerateisthesameasthelaserrepetitionrateandthecountrateiskeptbelow0.2events/lasershot.Thissuggeststhatthecameraframerateisnotfullyutilizedinthisway.Adifferentwayofrunningexperimentsistoexposeacameraframeformorethanonelasershots.Aslongastheaveragehitinthecameraisclosetoonepercameraframe,thecameraeventanddigitizereventcanbecorrelatedtoprovidethethreecoordinatesfor3Dmomentumimaging.Evenif therearea fewevents inonecamera frame, thebrightness‐intensitycorrelationwillbeabletocorrelatethetimeswiththepositionsoftheevents.Withthismethod,thecameraframeratecannowbefullyutilizedwhilethesystemrepetitionrateisincreasedfivetotentimesbeyondthehighestcameraframerate.Wedemonstratedthis using a 10 kHz laser located in the Kapteyn/Murnane group at University ofColoradoofBoulder.Thecamerawasrunningat2kFrames/secondandthelaserat10kHz.AstandardVMIsystemwithathree‐lensspectrometerwasusedtodetectthe photoelectrons arising from strong field ionization of krypton. Because onlyelectronswereof interest,noelectrode‐pulsingwasemployed.Figure7showsthe3DmomentumdistributionsoftheelectronNewtoncloud.Theachievedspatialandtemporal resolutionwasgoodwhile it tookonly fiveminutes toaccumulate theseevents.

http://dx.doi.org/10.1063/1.4981526

18

Figure 7.3Dmomentum distributions of photoelectrons arising from strong fieldionization of krypton by linearly polarized laser beams running at a 10 kHzrepetitionrate.Thecamera‐basedimagingsystemrunsat2kFrames/second.

Conclusion

We demonstrate a new method to convert a standard VMI apparatus to acoincidence 3D momentum‐imaging setup without modifying parts inside thevacuum chamber or the imaging detector. It should be noted this setup isautomatically capable of slicing the electron Newton sphere due to its excellenttemporal resolution.The additional cost for adding a second camera isminimum.The current setup requires a high repetition laser in order to expedite the dataacquisition.Furtherimprovementofthemulti‐hitcapabilitymightenablethistobeusedwithlowerrepetitionlasers.Furthermore,wehaveshownthatahighsystemrepetitionratebeyondthecameraframeratecanbeachieved.

With this new imaging setup, by measuring the KER of dissociative doubleionizationandcomparingitwithdensityfunctionalcalculations,weshowthemainproducts of strong field double ionization of PENNA are singlet diradicals, whichdissociateintoN(CH3)2CH2+andC6H5‐CH2+.Theisotropicdistributionsof fragment

http://dx.doi.org/10.1063/1.4981526

19

ions suggest a long lifetime of the parent dications, which poses a challenge forfutureexperimentsthataimformolecular/recoilframeionizationrate.

Acknowledgement

ResearchsupportedbytheChemicalSciences,Geosciences,andBiosciencesDivision,OfficeofBasicEnergySciences,OfficeofScience,U.S.DepartmentofEnergy,undergrant numberDE-SC0012628. Computational resources were provided byWayneState University Grid and the Consortium des Équipements de Calcul Intensif (CÉCI), fundedby theFondsde laRechercheScientifiquedeBelgique(F.R.S.‐FNRS)undergrantnumber2.5020.11.W.L.ispartiallysupportedasaSloanResearchFellow.F.R.andB.M.gratefullyacknowledgesupportoftheFondsdeNationaldelaRechercheScientifique(FNRS,Belgium).

[1] C.J.Danby,andJ.H.D.Eland,Intl.J.MassSpectrom.Ion.Phys.8,153(1972).[2] I.Powis,P.I.Mansell,andC.J.Danby,Intl.J.MassSpectrom.Ion.Phys.32,15(1979).[3] T.Baer,Intl.J.MassSpectrom.200,443(2000).[4] B.Sztáray,andT.Baer,Rev.Sci.Instrum.74,3763(2003).[5] A.V.Golovin,F.Heiser,C. J.K.Quayle,P.Morin,M.Simon,O.Gessner,P.M.Guyon,andU.Becker,Phys.Rev.Lett.79,4554(1997).[6] J.A.Davies,J.E.LeClaire,R.E.Continetti,andC.C.Hayden,J.Chem.Phys.111,1(1999).

http://dx.doi.org/10.1063/1.4981526

20

[7] D.Rolles,Z.D.Pešić,M.Perri,R.C.Bilodeau,G.D.Ackerman,B.S.Rude,A.L.D. Kilcoyne, J. D. Bozek, N. Berrah, Nucl. Instrum.Methods Phys. Res. B261, 170(2007).[8] W.Thetal.,J.Phys.B‐At.Mol.Opt.Phys.34,3669(2001).[9] M.Lebech,J.C.Houver,andD.Dowek,Rev.Sci.Instrum.73,1866(2002).[10] T.Osipov,C.L.Cocke,M.H.Prior,A.Landers,T.Weber,O.Jagutzki,L.Schmidt,H.Schmidt‐Böcking,andR.Dörner,Phys.Rev.Lett.90,233002(2003).[11] A.M.Rijs,M.H.M.Janssen,E.t.H.Chrysostom,andC.C.Hayden,Phys.Rev.Lett.92,123002(2004).[12] A.Vredenborg,W.G.Roeterdink,andM.H.M.Janssen,Rev.Sci.Instrum.79,063108(2008).[13] A.T.J.B.Eppink,andD.H.Parker,Rev.Sci.Instrum.68,3477(1997).[14] D.Townsend,M.P.Minitti,andA.G.Suits,Rev.Sci.Instrum.74,2530(2003).[15] M.Ryazanov,andH.Reisler,J.Chem.Phys.138,144201(2013).[16] S.K.Lee,F.Cudry,Y.F.Lin,S.Lingenfelter,A.H.Winney,L.Fan,andW.Li,Rev.Sci.Instrum.85,123303(2014).[17] S.K. Lee,Y. F. Lin, S. Lingenfelter, L. Fan,A.H.Winney, andW.Li, J. Chem.Phys.141,221101(2014).[18] A. H. Winney, Y. F. Lin, S. K. Lee, P. Adhikari, andW. Li, Phys. Rev. A 93,031402(2016).[19] R.Weinkauf,L.Lehr,andA.Metsala,J.Phys.Chem.A107,2787(2003).[20] S.Lünnemann,A. I.Kuleff,andL.S.Cederbaum,Chem.Phys.Lett.450,232(2008).

http://dx.doi.org/10.1063/1.4981526

21

[21] S.Lünnemann,A.I.Kuleff,andL.S.Cederbaum,J.Chem.Phys.129,104305(2008).[22] J.C.Bush,M.P.Minitti,andP.M.Weber,J.Phys.Chem.A114,11078(2010).[23] A. J. Jenkins,M.Vacher,M. J.Bearpark, andM.A.Robb, J.Chem.Phys.144,104110(2016).[24] B.Mignolet,R.D.Levine,andF.Remacle,J.Phys.B47,124011(2014).[25] W.Cheng,N.Kuthirummal,J.L.Gosselin,T.I.Sølling,R.Weinkauf,andP.M.Weber,J.Phys.Chem.A109,1920(2005).[26] L.Lehr,T.Horneff,R.Weinkauf,andE.W.Schlag,J.Phys.Chem.A109,8074(2005).[27] M.J.Frischetal.,Gaussian,Inc.,WallingfordCT,2016).[28] J.D.Chai,andM.Head‐Gordon,Phys.Chem.Chem.Phys.10,6615(2008).[29] M.M.Francl,W.J.Pietro,W.J.Hehre,J.S.Binkley,M.S.Gordon,D.J.Defrees,andJ.A.Pople,J.Chem.Phys.77,3654(1982).[30] H.P.C.,andJ.A.Pople,Theor.Chim.Acta.28,213(1973).[31] R.Dennington,T.Keith,andJ.Millam,(SemichemInc.,ShawneeMission,KS,2009).[32] B.Mignolet,R.D.Levine,andF.Remacle,Phys.Rev.A89,021403(2014).[33] C.S.Lehmann,N.B.Ram,andM.H.M.Janssen,Rev.Sci.Instrum.83,093103(2012).[34] S.Sun,B.Mignolet,L.Fan,W.Li,R.D.Levine,andF.Remacle,J.Phys.Chem.A121,1442(2017).

http://dx.doi.org/10.1063/1.4981526

22

[35] Y.F.Lin,S.K.Lee,P.Adhikari,T.Herath,S.Lingenfelter,A.H.Winney,andW.Li,Rev.Sci.Instrum.86,096110(2015).

http://dx.doi.org/10.1063/1.4981526

CMOS Camera (e-)

Continuous beam

SkimmerPhosphor/MCP

Computer

High-speedDigitizer

Concave mirror

Image

TOF

Signal decoupler

Linearly polarized light

CMOS Camera (ion)

BA

Ion mis-assignment

YX

t

http://dx.doi.org/10.1063/1.4981526

0

20

40

60

-60

-40

-20

P y (a

.u.)

0 20 40 60-60 -40 -20

Px (a.u.)0.0 0.2 0.4 0.6 0.8-0.8 -0.6 -0.4 -0.2

0.0

0.2

0.4

0.6

0.8

-0.8

-0.6

-0.4

-0.2

Px (a.u.)

P y (a

.u.)

A B CIon Electron Electron

0.0 0.2 0.4 0.6 0.8-0.8 -0.6 -0.4 -0.2

0.0

0.2

0.4

0.6

0.8

-0.8

-0.6

-0.4

-0.2

Pt (a.u.)

P y (a

.u.)

0510

25

45

80

http://dx.doi.org/10.1063/1.4981526

400

250

130

130

0.0 4.0 8.0 12.0 16.0 20.0 24.0 28.0

eKE (eV)

0.00

0.25

0.50

0.75

1.00

1.25

1.50

1.75

KER

(eV

)

BAExp.Theory

Coun

ts

1600

0.00

200

400

600

800

1000

1200

1400

eKE (eV)15.00 2.5 5.0 7.5 10.0 12.5

http://dx.doi.org/10.1063/1.4981526

0 50 100 150-150 -100 -50

0

50

100

150

-150

-100

-50

Px (a.u.)

P y (a

.u.)

0.0 0.2 0.4 0.6 0.8-0.8 -0.6 -0.4 -0.2

0.0

0.2

0.4

0.6

0.8

-0.8

-0.6

-0.4

-0.2

Pt (a.u.)

P y (a

.u.)

B C

0 50 100 150-150 -100 -50

0

50

100

150

-150

-100

-50

Px (a.u.)

P y (a

.u.)

×1.5

A Mass=58 Mass=91 Electron

01.53

6

16

24

×1.5

http://dx.doi.org/10.1063/1.4981526

600 800 1000 1200 1400 1600 1800 2000600

800

1000

1200

1400

1600

BA

Ion TOF (bin/2ns)

Ion

TOF

(bin

/2ns

)

0.0 1.0 2.0 3.0 4.0 5.0

KER (eV)

0

20

40

60

80

100

Coun

ts

01.7

17

34

53

http://dx.doi.org/10.1063/1.4981526

2.67

3.85

0.34 eV

0.96 eV

-3.51 eV

-1.71 eVProducts

TSPENNA2+

19.33 eV, triplet19.30 eV, singlet

PENNA+

7.43 eV

Neutral GS

20

18

16

14

12

10

8

6

4

2

0

Ener

gy (e

V) 21.0

17.5

18.0

18.5

19.0

19.5

20.0

20.5

C-C Bond Length (Å)1.5 2.0 2.5 3.0 3.5 4.0

Ener

gy (e

V)

PENNA2+

Singlet surface

Triplet surface

Open-shellsinglet

Open-shell triplet

Closed-shell singlet

http://dx.doi.org/10.1063/1.4981526

0.0 0.2 0.4 0.6-0.6 -0.4 -0.2

Px (a.u.)

A

0.0

0.2

0.4

-0.4

-0.2

P y (a

.u.)

0.0 0.2 0.4 0.6-0.6 -0.4 -0.2

Px (a.u.)

B

0.0

0.2

0.4

-0.4

-0.2

P z (a

.u.)

0.0 0.2 0.4-0.4 -0.2

Pz (a.u.)

C

0.0

0.2

0.4

-0.4

-0.2

P y (a

.u.)

051025

45

80

http://dx.doi.org/10.1063/1.4981526

Article File1234567