Dissociative electron attachment to pulsed supersonic O 2 jet : Violation of Σ + ⇌ Σ − selection rule and dependence on carrier gas proportion
aa r X i v : . [ phy s i c s . a t m - c l u s ] F e b Dissociative electron attachment to pulsed supersonic O jet : Violation of Σ + ⇋ Σ − selection rule and dependence on carrier gas proportion Irina Jana, Varun Ramaprasad, and Dhananjay Nandi ∗ Department of Physical Sciences,Indian Institute of Science Education and Research Kolkata,Mohanpur 741246, India
The formation of O − and O − ions via dissociative electron attachment to a pulsed supersonic jetof O molecules containing weakly bound small van der Waals clusters seeded in a beam of argon isreported. The energy dependence of the O − and O − yield exhibits three peaks near 7, 11 and 16eV incident electron energies. The 7 eV peak arises from the Π u state of O − whereas, the 11 and16 eV peaks are ascribed to two distinct resonance states: Σ + g and Σ + u states of O − , respectively,via a violation of the Σ + ⇋ Σ − selection rule. The dependence of the cross-section of these two newpeaks at ∼
11 and ∼
16 eV on the proportion of the carrier gas is also investigated and an optimumproportion has been observed experimentally which gives the lowest temperature of 14.86 K andhighest Mach number of 72.31 for the pulsed supersonic jet.
PACS numbers: 34.80.Ht, 31.15.A-
I. INTRODUCTION
Clusters of small molecules and atoms are importantintermediates between the gas phase and the condensedphase. These are of importance in many chemical andphysical processes starting from living tissues to atmo-spheric chemistry and hence is the topic of interest sincethe past few decades [1–3]. We study here the electronattachment process with isolated oxygen molecules andsmall van der Waals clusters of oxygen formed in pulsedsupersonic jet of oxygen molecules. Electron capture byan effusive oxygen beam is known to produce a resonanceat 6.5 eV incident electron energy following the reaction[4–7]: O ( Σ − g ) + e − → ( O − ) ∗ ( Π u ) → O − ( P ) + O ( P )(1)While electron attachment with oxygen clusters producesmainly two homologous series:( O ) − n and ( O ) n O − andcan be represented as :( O ) x + e − → ( O ) x − O − + O (2)which may be followed by subsequent isomerization [6].Marc et al. studied the electron attachment tooxygen clusters for the two homologous series ( O ) − n and( O ) n O − by expanding the gas through a 10 µm nozzle[6]. Other than the 6.5 eV peak for O − ion, they alsoreported the formation of a peak near 0 eV, forming dueto a non-dissociative attachment to O , while the dimer O − ion is seen to produce a peak ∼ et al. alsoreported vibrationally resolved electron attachment to ∗ [email protected] ( O ) n clusters from 0 to 2 eV incident electron energiesexpanding the O gas using a 20 µm nozzle [8]. Backin 1987, Azria et al. reported the formation of O − ionsfrom dissociative electron attachment (DEA) in electronstimulated desorption (ESD) to three-layer-thick thinfilm of O at two resonances near 7 and 13 eV incidentelectron energies. The 13 eV peak was ascribed due totwo distinct O − resonance states : Π u and Σ + g , via aviolation of the selection rule Σ − ⇋ Σ + [9]. Illenbergerin his review of electron attachment process to differentmolecular clusters claimed that DEA to ( O ) n clustersproduced two new distinct peaks near 8.3 and 14.5 eVincident electron energies for the dimer ( O ) − ion andalso O − ion [10].We report in this article a clear observation of two newdistinct resonance peaks at ∼
11 and ∼
16 eV incidentelectron energies along with the 7 eV peak, from O − and O − ions formed by dissociative electron attachment to apulsed supersonic jet of O molecules containing ( O ) n clusters seeded in a beam of argon.We also investigated for the first time the dependenceof the cross - section of these new symmetry forbiddenresonance peaks, at ∼
11 and ∼
16 eV incident electron en-ergies, on the proportion of the seeder gas. The change inthe pulsed supersonic beam temperature and Mach num-ber with change in argon proportion is also investigatedfor the first time and the results are explained.
II. EXPERIMENTAL SETUP
In order to have a supersonic jet of molecules, a highpressure gas is passed through a nozzle into a low pres-sure environment where it expands quasi-statically andadiabatically. In this set-up, the Amsterdam Cantilever
FIG. 1. Schematic with appropriate dimensions showing thepulsed valve holding the nozzle on the rail.
Piezo Valve (ACPV) is placed in an expansion chamberwith a 200 µm nozzle which can be controlled by an elec-tronic driver unity (EDU). The ACPV can be operatedboth in continuous and pulsed (0 - 5 kHz) mode givingeither a continuous or pulsed supersonic jet of molecules[11].A pulsed supersonic jet reduces the cost of pumpingby eliminating the need for a catcher chamber which isnecessary for a continuous supersonic jet. In the currentsetup, a pulsed supersonic jet of 37 µs (FWHM) is pro-duced by triggering the EDU with an input TTL pulseof 1.5 kHz using a master pulse generator. Another pulseoutput of 200 ns pulse width from the same master pulsegenerator goes to the input channel of an Ortec G8020delay generator. All other necessary pulses are takenfrom different outputs of this delay generator to have aconstant reference delay with respect to the molecularbeam pulse. A 138 µm skimmer, separating the twochambers, is placed inside the isentropic region, termedas the zone of silence , to extract the centreline beamconsisting of the coldest jet of molecules. The expansionchamber containing the pulsed valve is pumped with thehelp of a TC400 turbo molecular pump while, the mainchamber containing the electron gun, Faraday cup andthe time-of-flight (TOF) mass spectrometer is pumpedwith the help of a TC700 turbo molecular pump. Boththe chambers are maintained at a base pressure of ∼ − mbar with no gas and a pressure of ∼ − mbarin the expansion chamber and ∼ − mbar in the mainchamber with the gas inflow. To ensure the skimmerlies within the zone of silence and away from the Machdisc, the ACPV is placed on a rail allowing the distancebetween the nozzle and the skimmer to be changed.Details of the structure of a cold supersonic jet hasalready been explained in details earlier in many works[12, 13]. Once the pulsed centreline beam enters theinteraction region (in the main chamber), it is crossedwith a pulsed electron beam emitted from a home-madeelectron gun. FIG. 2. Pulse diagram showing different pulses (not to scale).
The electron gun has a tungsten filament heated witha 2.1 A constant current source which emits electronsthrough thermionic emission. The electron gun ispulsed with a 1.5 kHz TTL pulse of 200 ns pulse-widthcoming from the output of channel 1 of the G8020 delaygenerator. The pulsed supersonic jet takes a finite timeto travel from the point of origin i.e, the nozzle to theinteraction region. The delay between electron gunpulse and molecular beam pulse is varied from the delaygenerator to ensure proper synchronization betweenthe two pulses. The electron gun current is constantlymonitored with the help of a Faraday cup placed belowthe electron gun kept at a positive voltage of 40 V.The crossed pulsed electron beam and supersonic jetcollide under single collision condition producing nega-tive ions in all 4 π directions, thus forming the Newtonsphere of negative ions. The
Newton spheres formed arepushed from behind by the pusher plate having a pulsednegative voltage of -20 V. The pusher plate is pulsedwith a 4 µs pulse having a 300 ns delay with respect tothe electron gun. This delay ensures better extraction ofthe negative ions. The Newton spheres of negative ionsare then collimated with the help of the grounded pullerplater and the lens electrode having a positive polarityof 18 V. The Newton spheres then finally enter the 110mm long field-free flight tube having a positive polar-ity of 100 V, where they expand freely and finally reachthe MCP based detector. The detector consists of threemicro-channel plates (MCPs) placed in a Z-Stack con-figuration. The TOF of each ion reaching the detectorand the number of counts at a defined incident electronenergy are recorded with a TAC unit using LabVIEW.Details of the main chamber containing the electron gun,Faraday cup and time-of-flight mass spectrometer havealready been discussed before [14]. III. FORMATION OF AN OPTIMUM JET
Ideal gas condition of stationary flow is assumedthroughout where a high pressure gas is allowed topass through a small aperture and expand into a lowpressure region where it undergoes an isentropic wall-freeexpansion without condensation.The most important jet equation relating the stagna-tion enthalpy H , kinetic energy of the directed mass flow(E) and rest enthalpy (H) can be written as: H = E + H (3)where E = 12 m u (4)and H = U + P V (5)where u is velocity of the jet, U is internal energy, P ispressure and V is volume [13, 15].The temperature T and T before and after the aper-ture can be related using the specific heat at constantpressure ( c P ) using ideal gas behavior as: c P T = c P T + 12 m u (6)or TT = 1[1 + γ − M ] (7)Again, for an ideal, isentropic gas expansion with con-stant γ , we can write: PP = ( TT ) γγ − = [1 + γ − M ] − γγ − ρρ = nn = ( TT ) γ − = [1 + γ − M ] − γ − where P , P are the pressures before and after the aper-ture, ρ, ρ are densities before and after the apertureand n, n are number densities before and after theaperture.The above equations give one a clear quantitative ideaabout the Mach number ( M ) and local temperature ( T )along the streamline of the expanding gas jet. Conse-quently, all thermodynamic variables can be computedin the supersonic jet.The end of the isentropic region, known as the zone ofsilence is marked by the Mach disc [13]. The distanceof the Mach disc from the nozzle is related to the nozzlediameter as: x M = d (0 . r P P (8)where x M is the Mach disc location and d is theopening of nozzle. The location of the Mach disc wascalculated to be 857.47 mm with a 200 µ m nozzle, P was ∼ P was ∼ − mbar. It was ensuredthat the skimmer, to extract the centreline beam, wasplaced inside the zone of silence stretching to a distanceof 847.47 mm from the nozzle.In the present chamber, the pulsed valve fitted on arail can be moved to different positions changing thenozzle-to-skimmer distance 1. In order to investigatethe effect of nozzle-to-skimmer distance variation on theobserved beam interaction, the pulsed valve holding thenozzle was manually placed at 3 different positions andthe beam intensity of O − and O − ions were studiedkeeping the stagnation pressure fixed at 5 bar andexpansion chamber pressure at 7 × − mbar for all3 positions. As the nozzle-to-skimmer distance waschanged from screw points 1 to 3, it could be observedthat the O − cluster ions counts reached an optimum atscrew point 2. As the pulsed valve was moved to thefirst screw point on the rail, which is the closest to theskimmer, high jet densities formed at the nozzle reachthe skimmer. As a result, the O − ions counts increasedstrikingly while the formation of O − ion decreased to agreat extent. The strong barrel shock waves producedat the boundaries of the high density jets might not getattached to the skimmer and thus combine with theMach disc which lies perpendicular to the direction ofsupersonic flow scattering the jet particles. This resultsin a highly degraded supersonic beam which hinders theformation of the weakly bound van der Waals clustersi.e, the O − ions [16].On the other hand, when the pulsed valve was placedon the third screw position, the number of counts of O − ions were found to increase considerably in comparisonwith the O − ions but the total number of counts of both O − and O − ions was observed to be very low resultingin the remarkably low beam intensity at the interactionregion. In fact, when the distance between the nozzleand skimmer is large, the intensity of gas particles variesfollowing the classical scattering equation: I = I e m σ eff z (9)where, I and I are the ideal and observed beamintensities, respectively; m is the background numberdensity in the expansion chamber; z is the lengthover which scattering occurs and σ eff is the effectivescattering cross section [12, 17].Whereas, when the pulsed valve was placed at thesecond screw point, two noticeably different peaks in themass spectra were observed at masses 16 and 32 a.m.u.corresponding to O − and O − ions with sufficiently highbeam intensities. This supports the well known fact thatfor an optimum nozzle-to-skimmer distance, the beamintensity reaches the maximum. Shifting the nozzle,closer or further away from the optimum distance,results in a remarkable decrease in the beam intensity[12].In order to have proper synchronization between hepulsed molecular beam and pulsed electron beam, it isnecessary to know the time taken by the supersonic jetfrom the nozzle to reach to the interaction region. It hasto be assured that when the jet has reached the inter-action region then only the electron beam has been puton. The pusher plate pulse and the start pulse are wellsynchronized with the electron gun pulse. Schematic ofall the necessary pulses used in the experiment are shownin Figure 2. To find out the exact time taken by the jetto cover the distance between its point of origin and theinteraction region, the delay between the molecular beampulse and the electron gun pulse was varied slowly usingthe DGB 35 digital delay pulse generator from StanfordResearch Systems and the O − ion counts were noted at6.5 eV incident electron energy using LabVIEW, since O − is well known to have a resonant peak at 6.5 eV.The delay for which a maximum count was observed wasnoted. IV. CLUSTER FORMATION WITH A SEEDEDBEAM:
Velocity distribution (m/s)5100 5150 5200 5250 N u m be r o f c oun t s ( a r b . un i t s ) Mass (a.m.u.)16 20 25 30 32 35 40 N o . o f c oun t s ( a r b . un i t s ) FIG. 3. Velocity distribution fitted curves using Equation 11for different O :Ar ratios 1:4 (Blue line), 1:9 (Red line) and1:10 (Black line) compared together. For a mono-atomic gas final velocity of the supersonicjet can be expressed as [10]: v t = s k T M g (10)Here M g is the mass of the gas. This terminal velocity( v t ) is denoted by the peak position of the narrowvelocity distribution which is a characteristic feature ofthe supersonic jet.The situation becomes a little complicated for poly-atomic molecules as γ ∼
1, making the cooling processless effective. To address this problem, one can use aninert carrier gas like He or Ar. The molecules are dilutedin small proportion in a stream of inert carrier gas, argonin this case, such that the mass ( M g ) in Equation 10 isnow the mass of the carrier gas Ar. The carrier gas actsas a refrigerant taking away the excess heat from thejet and thus helping it to polymerize. Thus the use ofa carrier gas makes the cooling of the jet more effectivewhich in turn, results in better cluster formation.After the optimum distance between the nozzle andthe skimmer is identified, a beam of oxygen moleculesdiluted with Ar is then allowed to expand through thenozzle in the expansion chamber. The ratio of oxygenand argon has been varied over a wide range and themass spectra and ion-yield curves for both O − and O − ions have been noted. The difference in the abovementioned curves with change in argon proportion arethen investigated and reported here.The mass spectra with O : Ar =1:10 is shown in theinset of Figure 3. With the change of Ar proportion themass spectra do not show much visible change while,the number of counts and FWHM change. The firstpeak of the mass spectra corresponds to O − ions (16a.m.u) while the second peak corresponds to O − ions (32a.m.u).The TOF spectra can be converted to a velocity dis-tribution spectra considering the proper distance coveredby the jet. The velocity distribution of a collimated su-personic jet can be expressed using the equation: f ( v ) = a ( v − c ) exp ( − b (( v − u ) − c ) )) (11)where a is a normalization constant, b = m k T ,where m and T are mass and temperature of the jet,respectively; k is the Boltzmann constant; u is the flowvelocity and c is just a parameter shifting the center ofthe fit [12, 15].By fitting the velocity distribution spectra of O − ionsfor different O : Ar ratios, the parallel translationaltemperature of the supersonic jet ( T ) can be calculated.It is also to be noted here that the peak position of thenarrowed velocity distribution of the beam represents TABLE I. Table showing different fitting parameters forEquation 11 for O : Ar ratios 1:4, 1:9 and 1:10.a b c u R γ eff ), effective mass ( M eff ), velocity of sound ( C ),Mach number and jet temperature (T) with variation in argonproportion for O : Ar ratios 1:4, 1:9 and 1:10. γ eff M eff (a.m.u.) C ( ms ) M T(K)1:4 1.56 35.2 121.40 42.84 401:9 1.58 37.6 101.84 50.97 29.691:10 1.58 37.82 71.84 72.31 14.86 the terminal velocity of the beam. Thus by knowingthe peak position of the supersonic jet, the velocity andhence the Mach number of the jet can be calculated.The fitted curves for the velocity distribution of O − ions using Equation 11 for O : Ar ratios 1:4, 1:9 and1:10 are compared in Figure 3. The FWHM for O : Ar ratios 1:4, 1:9 and 1:10 are 108, 135 and 95, respectively.This shows that the 1:10 ratio has the lowest FWHMvalue implying the most narrow velocity distribution.The different fitting parameters for O : Ar ratios 1:4,1:9 and 1:10 are given in Table I. The calculated beamtemperatures and Mach numbers are shown in TableII where C denotes speed of sound in the respective O : Ar medium.It can be noted from Table II that as the proportionof argon is increased keeping the stagnant pressure ( P )same ( 5.5 bar), the Mach number increases notably for O : Ar ratios 1:4, 1:9 and 1:10.The temperature of the jet also keeps reducing withthe increase in Ar concentration. In the light of thisobservation, we predict that amongst the O : Ar ratios1:4, 1:9 and 1:10, the beam having th portion of argonforms the most effective clusters of oxygen molecules withthe highest Mach number and the lowest temperature. V. ION-YIELD CURVES
The ion yield curves for O − and O − ions after properbackground correction are shown in Figures 4 and 5,respectively for O : Ar ratio 1:4, 1:10 and 1:12. Thestagnant reservoir pressure was kept fixed at 5 bar, 5.5bar and 6 bar for O : Ar ratios 1:4, 1:10 and 1:12,respectively. Ion yield curve for pure O supersonic jetwithout any carrier shows a single resonance peak ∼ O − and O − ions. An additional peak near 0 eVis also known to be observed by previous groups which Incident electron energy (eV)0 5 10 15 20 I on C oun t s ( a r b . un i t s ) O :Ar=1:4O :Ar=1:10O :Ar=1:12 FIG. 4. Ion-yield curves of O − ions for O :Ar ratios 1:4, 1:10and 1:12. Symbols represent experimental data and lines theircorrespoding fits. Incident electron energy (eV)0 5 10 15 20 I on C oun t s ( a r b . un i t s ) O :Ar=1:4O :Ar=1:10O :Ar=1:12 FIG. 5. Ion-yield curves of O − ions for O :Ar ratios 1:4, 1:10and 1:12. Symbols represent experimental data and lines theircorrespoding fits. explains the large thermal-energy electron attachmentcross-section and low-temperature and high-pressureconditions [6, 8]. At stagnation pressure 5, 5.5 and 6bar for different O : Ar ratios: 1:4, 1:10 and 1:12 , weobserved two distinct new features near ∼
11 and ∼ O − and O − ionsas shown in the ion yield curves. The evolution of thesetwo new features has been reported only by Illenbergerpreviously who had observed two new peaks at ∼ ∼ O − ions at 3.5 bar stagnation pressure[10]. Illenberger also reported that the appearance ofthe two new features was more pronounced for O − ionsat even lower stagnation pressures (1.5 bar). Whilewe report here some completely new observations, we Argon fraction0.8 0.82 0.84 0.86 0.88 0.9 0.92 P ea k r a t i o FIG. 6. Ratio of different peaks with variation in Ar propor-tion for O − ions. The vertical lines represent error bars foreach peak ratio. Argon fraction0.8 0.82 0.84 0.86 0.88 0.9 0.92 P ea k r a t i o FIG. 7. Ratio of different peaks with variation in Ar propor-tion for O − ions. The vertical lines represent error bars foreach peak ratio. observe that the new features near ∼
11 and ∼
16 eVincident electron energies are comparatively much morepronounced in O − ion than O − ion for O : Ar ratios1:4 and 1:10. However, as the proportion of Argon isincreased to a even higher value of O :Ar = 1:12, itcould be observed that the peaks near ∼
11 and ∼
16 eVincident electron energies become more pronounced for O − ion as compared to O − ion.The appearance for these additional new features wasexplained by Illenberger in reference to the electronstimulated desorption (ESD) experiments performed bySanche et al. [10, 18]. The ground state of oxygen is a Σ − g state . Electron attachment to the neutral molecule at 6.5 eV incident electron energy occurs due to atransition of the molecule to a Π u state. According tothe σ − selection rule, the single electron wave functionshould be σ + in a single electron molecule frame ofreference. Thus, transitions from σ − to σ + states andvice-versa are not allowed as per the σ − selection rule.But the two new resonance peaks occurring at ∼
11 and ∼
16 eV incident electron energies support the fact thattransitions from Σ − g to Σ + g and Σ − g to Σ + u are givingrise to the peaks near ∼
11 and ∼
16 eV, respectively[10]. This points to the fact that in the condensedphase, the σ − selection rule is violated. Evidence forthe violation of the σ − selection rule at the condensedphase of materials have also been given previously forESD experiments [9, 18].In light of these findings of the symmetry-forbiddentransitions, another explanation can be provided. Theincoming electrons are initially inelastically scatteredin the cluster and slowed down. These slowed downelectrons in the vicinity of the cluster is resonantlycaptured by the molecule giving rise to the negativeions. Thus the secondary reactions may give rise tothe symmetry forbidden states in the ion-yield curves[10, 19].It can also be observed that these symmetry forbiddenstates are a function of the stagnant pressure and alsoa function of the proportion of the carrier gas Ar. Theratios of the second and third resonance peaks withrespect to the first peak are plotted in Figures 6 and 7for O − and O − ions, respectively.The ratios are clearlyseen to follow a particular trend with the change in theconcentration of the carrier gas. The O : Ar ration1:10 gives the optimum peak ratio. Shifting the argonproportion below or above this proportion results inconsiderable decrease in the number of counts of thetwo new resonance peaks at ∼
11 and ∼
16 eV incidentelectron energies. This points to the fact that theconcentration of the carrier gas has a very importantrole to play in cooling down of the supersonic jet andhence the effective formation of clusters.When a beam of polyatomic molecules, seeded in amonoatomic inert carrier gas, undergoes supersonic ex-pansion, the translational temperature of the carriergas falls to an extreme low value. The polyatomic gasmolecules undergo two-body collisions with this low- tem-perature bath created by the carrier gas resulting intranslational and rotational cooling of the polyatomic gasjet. The decrease in the proportion of the seeder gas be-low a certain amount definitely affects the cooling of thejet making the formation of clusters less effective. While,if the proportion of the carrier gas is too high, complexformation of the mono-atomic inert gas again sets a limiton the cooling of the supersonic jet. At some point, themolecules start colliding with the atoms and begin topolymerize giving off inter-atomic binding energy. Thisagain reheats the jet making the cooling less effective [20].
VI. CONCLUSION
The present results indicate the presence of symmetryforbidden electronic transitions in electron attachmentprocesses to isolated oxygen molecules in supersonic oxy-gen jet and also ( O ) n homologous clusters. Thus DEAto O supersonic beam containing neutral clusters canprovide a method by which resonances not observed inthe gas phase can be detected. The cross-section of such forbidden electronic transitions can also be controlled bychanging the concentration o the carrier gas. However,a time-dependent nuclear dynamics would be helpful tounderstand the mechanism responsible for such transi-tions. VII. ACKNOWLEDGMENTS
We gratefully acknowledge financial supports from“Science and Engineering Research Board (SERB)”for supporting this research under the Project“EMR/2014/000457” [1] P. W. Stephens and J. G. King, Physical review letters , 1538 (1983).[2] B. Loupias, M. Koenig, E. Falize, S. Bouquet, N. Ozaki,A. Benuzzi-Mounaix, T. Vinci, C. Michaut, M. R.Le Goahec, W. Nazarov, et al. , Physical review letters , 265001 (2007).[3] J. Lengyel, A. Pysanenko, V. Poterya, P. Slav´ıˇcek,M. F´arn´ık, J. Koˇciˇsek, and J. Fedor, Physical reviewletters , 113401 (2014).[4] T. Jaffke, M. Meinke, R. Hashemi, L. G. Christophorou,and E. Illenberger, Chemical physics letters , 62(1992).[5] W. Henderson, W. Fite, and R. Brackmann, PhysicalReview , 157 (1969).[6] T. M¨ark, K. Leiter, W. Ritter, and A. Stamatovic, Phys-ical review letters , 2559 (1985).[7] M. Tashiro, K. Morokuma, and J. Tennyson, PhysicalReview A , 052707 (2006).[8] S. Matejcik, A. Kiendler, P. Stampfli, A. Stamatovic,and T. M¨ark, Physical review letters , 3771 (1996).[9] R. Azria, L. Parenteau, and L. Sanche, Phys. Rev. Lett. , 638 (1987).[10] E. Illenberger, Chemical reviews , 1589 (1992). [11] D. Irimia, D. Dobrikov, R. Kortekaas, H. Voet, D. A.van den Ende, W. A. Groen, and M. H. Janssen, Reviewof Scientific Instruments , 113303 (2009).[12] B. R. Cameron, Investigations of the flow dynamics of su-personic molecular beams and the ionization of molecularclusters by electron impact , Ph.D. thesis, The Universityof Canterbury, Christchurch, New Zealand (November1993).[13] S. Zhang, X. Ma, H. Liu, B. Li, and X. Zhu, Science inChina Series G: Physics, Mechanics and Astronomy ,709 (2006).[14] P. Nag and D. Nandi, Measurement Science and Tech-nology , 095007 (2015).[15] H. Haberland, U. Buck, and M. Tolle, Review of scientificinstruments , 1712 (1985).[16] K.-S. Im, S.-K. Cheong, X. Liu, J. Wang, M.-C. Lai,M. W. Tate, A. Ercan, M. J. Renzi, D. R. Schuette, andS. M. Gruner, Physical review letters , 074501 (2009).[17] R. D. Levine, R. B. Bernstein, et al. , Molecular reactiondynamics (Clarendon Press; Oxford University Press,2005).[18] L. Sanche, Physical review letters , 1638 (1984).[19] G. Schulz and J. Dowell, Physical Review , 174(1962).[20] R. E. Smalley, L. Wharton, and D. H. Levy, Accountsof Chemical Research10