Rotationally resolved pulsed-field ionization photoelectron bands for O 2 + (A 2 Π u ,v +=0–12) in the energy range of 17.0–18.2 eVY. Song, M. Evans, C. Y. Ng, C.-W. Hsu, and G. K. Jarvis Citation: The Journal of Chemical Physics 112, 1271 (2000); doi: 10.1063/1.480596 View online: http://dx.doi.org/10.1063/1.480596 View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/112/3?ver=pdfcov Published by the AIP Publishing Articles you may be interested in Pulsed field ionization–photoelectron bands for CO 2 + (A 2 Π u and B 2 Σ u + ) in the energy range of17.2–19.0 eV: An experimental and theoretical study J. Chem. Phys. 113, 7988 (2000); 10.1063/1.1314354 Rotationally resolved pulsed field ionization photoelectron bands for O 2 + (a 4 Π u ,v + =0–18) in the energyrange of 16.0–18.0 eV J. Chem. Phys. 112, 1306 (2000); 10.1063/1.480682 Rotationally resolved pulsed field ionization photoelectron study of CO + (X 2 Σ + ,v + =0–42) in the energy rangeof 13.98–21.92 eV J. Chem. Phys. 111, 8879 (1999); 10.1063/1.480259 Rotational-resolved pulsed field ionization photoelectron study of NO + (a 3 Σ + ,v + =0–16) in the energy rangeof 15.6–18.2 eV J. Chem. Phys. 111, 1937 (1999); 10.1063/1.479462 Rotationally resolved pulsed field ionization photoelectron bands of O 2 + (X 2 Π 1/2,3/2g ,v + =0–38) in theenergy range of 12.05–18.15 eV J. Chem. Phys. 111, 1905 (1999); 10.1063/1.479459

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

131.181.251.130 On: Sat, 22 Nov 2014 07:56:34

Rotationally resolved pulsed-field ionization photoelectron bandsfor O 2

1„A 2Pu ,v 150 – 12… in the energy range of 17.0–18.2 eV

Y. Song, M. Evans, and C. Y. Nga)

Ames Laboratory, U.S. Department of Energy and Department of Chemistry, Iowa State University,Ames, Iowa 50011

C.-W. Hsu and G. K. JarvisChemical Science Division, Lawrence Berkeley National Laboratory, Berkeley, California 94720

~Received 10 August 1999; accepted 21 October 1999!

We have obtained rotationally resolved pulsed-field ionization photoelectron~PFI-PE! spectra forO2 in the energy range of 17.05–18.13 eV, covering the ionization transitions O2

1(A 2Pu ,v1

50 – 12,N1)←O2(X3Sg

2 ,v950,N9). Although these O21(A 2Pu ,v1) PFI-PE bands have

significant overlaps with vibrational bands for O21(a 4Pu) and O2

1(X 2Pg), we have identified allthe O2

1(A 2Pu ,v150 – 12) bands by simulation of spectra obtained using supersonically cooled O2

samples with rotational temperatures'20 and 220 K. While thesev150 – 12 PFI-PE bandsrepresent the first rotationally resolved photoelectron data for O2

1(A 2Pu), the PFI-PE bands forO2

1(A 2Pu ,v159 and 10) are the first rotationally resolved spectroscopic data for these levels. Thesimulation also allows the determination of accurate ionization energies, vibrational constants, androtational constants for O2

1(A 2Pu ,v150 – 12). The analysis of the PFI-PE spectra supports theconclusion of the previous emission study that the O2

1(A 2Pu ,v159 and 10) states are stronglyperturbed by a nearby electronic state. ©2000 American Institute of Physics.@S0021-9606~00!01703-7#

I. INTRODUCTION

In a recent series of synchrotron-based pulsed-field ion-ization ~PFI! photoelectron~PFI-PE! experiments, we haveobtained and analyzed rotational-resolved PFI-PE vibrationalbands for the O2

1(X 2Pg ,b 4Sg2,22Pu ,B 2Sg

2 ,2 Su2 , and

c 4Su2),1–6 NO1(X 1S1 anda 3S1),7,8 and CO1(X 2S1,

A 2P,D 2P,and 32S1)9–11 states. A major advantage ofsynchrotron studies is the capability to cover a wide photonenergy range, allowing a long vibrational progression of anelectronic state to be examined. This capability has beenclearly demonstrated in the recent comprehensive PFI-PEstudies of O2

1(X 2Pg ,b 4Sg2 , andB 2Sg

2),2,5,6 NO1(X 1S1

anda 3S1),7,8 and CO1(X 2S1 andA 2P),9,10 Synchrotron-based PFI-PE studies are complementary to vacuum ultravio-let ~VUV ! laser PFI-PE measurements,12 which often cover anarrower energy range, but can be performed at a higherresolution.13 In addition to providing accurate ionization en-ergies~IEs! and new spectroscopic constants, the rotational-resolved synchrotron-based PFI-PE studies have revealedfundamental insight into the threshold photoionizationdynamics12 of O2, NO, and CO, which cannot be obtained intraditional ion spectroscopic measurements based on absorp-tion and ion emission.14

The spectroscopy for O21(A 2Pu ,v1) is relatively well

known.14–17On the basis of a comprehensive analysis of theO2

1(A 2Pu)→O21(X 2Pg) emission system, Coxon and

Haley14 have obtained highly accurate spectroscopic con-

stants for O21(A 2Pu ,v150 – 8 and 11– 15). However, the

vibrational and rotational constants for the O21(A 2Pu ,v1

59 and 10) levels have only been interpolated in the latterstudy. It is well known from previous photoelectron studiesthat the vibrational bands for O2

1(A 2Pu) are in serious over-lap with those of the O2

1(X 2Pg anda 4 Pu) states.6,15,17This,together with the high multiplicities of the electronic statesinvolved, has made the clear identification of these vibronicbands difficult in previous photoelectron spectroscopicstudies.17–19 In a recent synchrotron-based threshold photo-electron~TPE! study employing the penetration field tech-nique, along with the rotational cooling of the O2 sampleusing the supersonic beam technique, Morioka and co-workers have successfully identified most of the vibrationalbands for the O2

1(X 2Pg ,a 4Pu , and A 2Pu ,) states in theenergy region of interest to the present study.15 Although thevibrational bands for O2

1(A 2Pu ,v1) at energies close to itsdissociation limit have been assigned in Ref. 15, the photo-electron bands are only recorded and analyzed to the vibra-tionally resolved level.

This report is one among a series of articles1–6 on thedetailed photoelectron study of O2. Here, we present theanalysis of the rotational-resolved PFI-PE data forO2

1(A 2Pu ,v150 – 12) obtained using an O2 sample at'220 K. Due to the weakness of the O2

1(A 2Pu ,v150)band, its clear identification is only possible by using a cold~'20 K! O2 sample achieved in a supersonic expansion. ThePFI-PE bands for O2

1(A 2Pu ,v151 – 3,5) have also beenexamined using the 20 K O2 sample for confirmation of thespectral simulation. The simulation of these PFI-PE data al-lows the determination of accurate IEs, vibrational constants,

a!Author to whom correspondence should be addressed; electronic mail:cyng@ameslab.gov

JOURNAL OF CHEMICAL PHYSICS VOLUME 112, NUMBER 3 15 JANUARY 2000

12710021-9606/2000/112(3)/1271/8/$17.00 © 2000 American Institute of Physics

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

131.181.251.130 On: Sat, 22 Nov 2014 07:56:34

and rotational constants for O21(A 2Pu ,v150 – 12).

II. EXPERIMENT

The design and performance of the high-resolution VUVphotoionization facility of the Chemical Dynamics Beamlineat the Advanced Light Source~ALS! has been describedpreviously.2,20–23 Briefly, the major components include a10-cm period undulator, a gas harmonic filter, a 6.65 m off-plane Eagle monochromator, and a photoion–photoelectronapparatus.2,20–22

In the present experiment, helium is used in the har-monic gas filter, where higher undulator harmonics with pho-ton energies greater than 24.59 eV are suppressed. The fun-damental light from the undulator is then directed into the6.65 m monochromator and dispersed by a 4800 lines/mmgrating (dispersion50.32 Å/mm) before entering the experi-mental apparatus. The PFI-PE bands were measured usingthe 4800 lines/mm grating and monochromator entrance/exitslits of 50–200mm, achieving wavelength resolutions of0.016–0.064 Å~full-width-at-half-maximum, FWHM!. Thephoton energy step size was varied in the range of 0.1–0.25meV.

The ALS storage ring is capable of filling 328 electronbuckets in a period of 656 ns. Each electron bucket emits alight pulse of 50 ps with a time separation of 2 ns betweensuccessive bunches.23 In each storage ring period, a dark gap~16 ns! consisting of eight consecutive unfilled buckets existsfor the ejection of cations from the orbit. A multibunch op-eration with a dark gap of 16 ns corresponds to 320 bunchesin the synchrotron orbit or a repetition rate of 488 MHz.

The procedures for PFI-PE measurements using thephotoion–photoelectron apparatus have been described pre-viously in detail.2,20–22 The electron spectrometer, whichconsists of a steradiancy analyzer and a hemispherical energyanalyzer arranged in tandem, was used to filter prompt elec-trons. For measurements of PFI-PE bands, a pulsed electricfield (height50.67 V/cm, width540 ns) was applied everythree synchrotron periods, i.e., 1.97ms. Thus, the repetitionrate of the electric field pulses was 0.51 MHz.

A continuous molecular beam of pure O2 was producedby supersonic expansion through a stainless-steel nozzle(diameter50.127 mm) at a stagnation pressure of 760 Torrand a nozzle temperature of 298 K. The molecular beam wasskimmed by a circular skimmer (diameter51 mm) beforeintersecting the monochromatized VUV light beam 7 cmdownstream in the photoionization region. As shown below,the simulation of PFI-PE spectra suggests that the rotationaltemperature for O2 achieved is'220 K in this experiment.

Recently, the O21(A 2Pu ,v150 – 3 and 5) PFI-PE bands

were re-examined in another experiment performed using acolder O2 molecular beam sample with a rotational tempera-ture of '20 K. A new synchrotron-based PFI-PE detectionscheme, which relies on the time-of-flight~TOF! selection ofPFI-PEs, is used in this recent study.24 By employing a syn-chrotron dark gap of>80 ns, we show that the TOF peak forPFI-PEs can be adjusted to arrive in the dark gap with littlecontamination by prompt electrons. This new scheme is notonly more effective in suppression of prompt electrons, but

also has a higher electron transmission. Using a 2400 lines/mm/grating and monochromator entrance/exit slits of 50–100 mm, the wavelength resolution achieved in this PFI-PEexperiment was 0.032–0.064 Å~FWHM!.

The absolute photon energy scale was calibrated usingthe Ar1(2P3/2), and Ne1(2P3/2) PFI-PE bands recorded un-der the same experimental conditions before and after eachscan.2,21–23This calibration procedure assumes that the Starkshift for ionization thresholds of O2 and the rare gases areidentical. On the basis of previous experiments,1–9,25 the ac-curacy of the energy calibration is believed to be within60.5 meV.

III. RESULTS AND DISCUSSION

The main electronic configuration for the O2(X3Sg

2)ground state isKK(2sg)2(2su)2(3sg)2(1pu)4(1pg)2. Theejection of an electron from a 1pu bonding orbital results inthe O2

1(a 4Pu) and O21(A 2Pu) excited states. Both these

excited states, together with the ground O21(X 2Pg) state,

converge to the first dissociation limit O(3P)1O1(4S0), butwith significantly shallower potential wells than that of theO2

1(X 2Pg) state.26

In the energy region~17.00–18.15 eV! of interest to thepresent experiment, we have identified rotational-resolvedvibrational bands for O2

1(X 2Pg ,v1528– 38),O2

1(a 4Pu ,v158 – 18), and O21(A 2Pu ,v150 – 12) @see

Fig. 2~b! of Ref. 6#. The analysis of the vibrational PFI-PEbands for O2

1(X 2Pg) has been reported recently.6 Generally,each of thev150 – 12 PFI-PE bands for the O2

1(A 2Pu)state overlaps with the respectivev158 – 20 PFI-PE vibra-tional bands of the O2

1(a 4Pu) state. While the overlaps be-tween the respective O2

1(a 4Pu ,v158 – 10) andO2

1(A 2Pu ,v150 – 2) bands are very serious, the structuresfor the O2

1(a 4Pu ,v1.17) state give only minor interfer-ence to the PFI-PE bands for O2

1(A 2Pu ,v159 – 12). TheO2

1(X 2Pg ,v1528– 32) PFI-PE bands also have significantoverlap with the respective O2

1(A 2Pu ,v150 – 4) bands.6

Taking into account the spectra reported in other publi-cations on the analysis of the O2

1(X 2Pg ,v1)6 andO2

1(a 4Pu ,v1)27 states, we have selected to show in Figs.1~a!–1~f! the respective mixed experimental PFI-PE bands~open circles, top spectra of lower panels! forO2

1(A 2Pu ,v153 – 8) obtained using an O2 sample with arotational temperature of'220 K. Figures 2~a!–2~d! depictthe experimental PFI-PE bands~open circles, top spectra! forO2

1(A 2Pu ,v150 – 3) measured using a cold O2 molecularbeam sample~rotational temperature'20 K!. The experi-mental spectrum of Fig. 2~a! was obtained by subtracting thedeconvoluted band for O2

1(a 4Pu ,v158) @bottom spectrum~solid line! shown in Fig. 5~b! of Ref. 6# from the mixedexperimental spectrum@top spectrum~open circles! shown inFig. 5~b! of Ref. 6#. We note that this subtracted spectrumconsists of contributions from both the O2

1(X 2Pg ,v1528)and O2

1(A 2Pu ,v150) bands. The 20 K spectra of Figs.2~a!–2~d! reveal essentially a single peak for theO2

1(A 2Pu ,v150 – 3) PFI-PE bands. The mixed experimen-tal PFI-PE bands ~open circles, upper spectra! forO2

1(A 2Pu ,v159,10,11,and 12) are plotted in Fig. 3. These

1272 J. Chem. Phys., Vol. 112, No. 3, 15 January 2000 Song et al.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

131.181.251.130 On: Sat, 22 Nov 2014 07:56:34

FIG. 1. Deconvolution of the PFI-PE bands for O21(A 2P1/2,3/2u). ~a! v153, ~b! v154, ~c! v155, ~d! v156, ~e! v157, and ~f! v158 from the

corresponding mixed bands by BOS simulation.~upper panels!: The deconvoluted PFI-PE bands for O21(A 2P1/2,3/2u ,v153 – 8). The rotational transitions to

the O21(A 2P1/2u ,v1,N1) and O2

1(A 2P3/2u ,v1,N1) from O2(X3Sg

2 ,v950,N9) are marked using downward-pointing and upward-pointing stick marks,respectively. TheDJ rotational branches are also marked. Lower panels: The top spectra~open circles! are the mixed experimental PFI-PE spectra obtainedusing an O2 sample with a rotational temperature of'220 K. The deconvoluted PFI-PE bands~solid lines! for O2

1(a 4Pu ,v1511– 16) andO2

1(X 2P1/2,3/2g ,v1531 and 32) are shown as the bottom spectra. The middle spectra~solid lines! represent the sums of the deconvoluted bands.

1273J. Chem. Phys., Vol. 112, No. 3, 15 January 2000 Photoelectron bands for O2

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

131.181.251.130 On: Sat, 22 Nov 2014 07:56:34

vibrational PFI-PE bands are aligned by their IEs determinedin the simulation as described below.

The PFI-PE intensities@ I (e2)# are normalized by theVUV photon intensities@ I (hn)# measured using a tungstenphotoelectric detector. The detection efficiency for the tung-sten detector as a function of VUV photon energy has beencorrected based on the known photoelectric yield oftungsten.28 The I (e2)/I (hn) scales for experimental spectraplotted within Figs. 1~a!–1~f! or 2~a!–2~d! or 3 show theactual relative PFI-PE intensities observed in the experiment.The intensity for the O2

1(A 2Pu ,v151) band is higher thanthat for O2

1(A 2Pu ,v150), but lower than that forO2

1(A 2Pu ,v152). The structures for the O21(A 2Pu ,v1

>3) bands are the dominant components of the mixedPFI-PE bands observed at higher energies. The intensity forvibrational bands of O2

1(A 2Pu ,v1) reaches a global maxi-mum at v153 and then decreases gradually fromv153toward higher vibrational levels up tov156 as expectedbased on the Franck–Condon consideration. However, thePFI-PE band intensity for O2

1(A 2Pu) exhibits two other lo-cal maxima atv157 and 10. We note that the maximumintensity for thev159 is also higher than that forv158.These band intensity deviations from the Franck–Condonbehavior are indicative of perturbative ionization mecha-nisms involving other electronic states.

A. Spectral simulation and assignment of rotationalbranches

The relative intensities for rotational structures resolvedin individual vibrational PFI-PE bands for O2

1(A 2Pu ,v1

50 – 12) were simulated using the Buckingham–Orr–Sichel~BOS! model,29 which was derived to predict rotational linestrength observed in single-photon ionization of diatomicmolecules. This procedure has been described in detailpreviously.2,6,8 The BOS model is a one-electron model anddoes not take into account any channel interactions. TheBOS coefficientsCl’s are predicted to be identical for all thevibrational levels and spin-orbit components. These predic-tions were not confirmed in the present and previous experi-ments, indicative of the existence of channel interactions dueto perturbation by other states. Thus, the BOS simulationemployed here should be viewed as empirical in nature.

The ionization transition O21(A 2P1/2,3/2u)←O2(X

3Sg2)

can be classified as a Hund’s case~b! to ~a! transition and issimilar to the O2

1(X 2P1/2,3/2g)←O2(X3Sg

2) system. Sincethe detailed discussion on simulation of the latter system hasbeen reported previously,6 we have omitted the redundantdescription here. As in previous studies, this simulation usesthe known spectroscopic constants for O2(X

3Sg2 ,v950).30

Furthermore, the simulation has taken into account transi-tions involving the three fine-structure sublevelsF1 , F2 , andF3 of O2(X

3Sg2 ,v950), despite the fact that the present

experiment cannot resolve the fine-structure transitions.2 Weassume that the rotational populations for O2 are governed bythe Boltzmann distribution characterized by a rotational tem-perature. Thus, the rotational temperature for O2, along withthe BOS coefficients and rotational constants forO2

1(A 2P1/2,3/2u ,v1), is a fitting parameter. On the basis ofspectral simulations, we estimate that the PFI-PE spectra of

FIG. 1. ~Continued!.

1274 J. Chem. Phys., Vol. 112, No. 3, 15 January 2000 Song et al.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

131.181.251.130 On: Sat, 22 Nov 2014 07:56:34

Figs. 1~a!–1~f! and 3 are measured using an O2 sample witha rotational temperature of 220620 K. For the spectra ofFigs. 2~a!–2~d!, we estimate that the rotational temperatureof the O2 sample is'20 K. A Gaussian linewidth of 5 cm21

~FWHM! is used as the instrumental resolution profile in thesimulation.

For a S to P transition (DL51), the first 3-j symbolrequires that the BOS coefficientC0 is zero for the photo-ionization process. The best-fitted BOS coefficients for theO2

1(A 2P1/2,3/2u ,v150 – 12) PFI-PE bands are listed inTable I. We have been able to simulate thev150 – 2, 6–8,11, and 12 PFI-PE bands using onlyC1 for both V compo-nents. All other experimental PFI-PE vibrational bands weresatisfactorily simulated using the BOS coefficients (C1 ,C2)for both V components, except forv1510, V51/2, whereC3 is also required. The observation of dominant (C1 ,C2)coefficients indicates that thep- and d-partial waves of theground electron wave function are the major contributors inthe ionizing transitions. Although the uncertainties for theCl values are relatively high~60.2!, the deviations of theBOS coefficients from the general trend are still discerniblefor thev159 and 10 bands. For thev159 band,C2 ~50.8!is significantly greater than that forC1 ~50.2!, whereasC3 is

nonzero for thev1510 band. This observation is indicativeof perturbation by other states in the neighborhood ofv1

59 and 10. As shown in Fig. 3, the experimental spectrumfor the v1510 band has a substantially higher intensity anda narrower width compared to those for the adjacent bands.

FIG. 2. Simulations~bottom spectra, solid lines! of experimental PFI-PEbands~top spectra, open circles! for O2

1(A 2Pu). ~a! v150, ~b! v151, ~c!v152, and~d! v153 obtained using an O2 sample with a rotational tem-perature of'20 K. The experimental PFI-PF band of~a! is obtained bysubtracting the deconvoluted band for O2

1(a 4Pu ,v158) from the mixedexperimental spectrum~see Ref. 6!. The deconvoluted spectrum~dashedcurve! for O2

1(a 2P3/2g ,v1528) is also shown in~a!. The rotational tran-sitions to the O2

1(A 2P1/2u ,v1,N1) and O21(A 2P3/2u ,v1,N1) from

O2(X3Sg

2 ,v950,N9) are marked using downward-pointing and upward-pointing stick marks, respectively. TheDJ rotational branches are alsomarked.

FIG. 3. Simulated and experimental PFI-PF spectra for O21(A 2P3/2u ,v1

59 – 12). The horizontal scale ishn2IE(v1), where hn is the photonenergy. For eachv1 state, the upper spectrum~open circles! is the experi-mental spectrum, while the simulated spectrum~solid line! is shown under-neath the experimental curve.

TABLE I. The best-fitted BOS coefficients (Ci , i 51,2!a obtained in thesimulation of the PFI-PE bands for O2

1(A 2P1/2,3/2u ,v1).

v1 C1 C2

0 1.00/1.00 0.00/0.001 1.00/1.00 0.00/0.002 1.00/1.00 0.00/0.003 0.50/1.00 0.50/0.004 0.50/1.00 0.50/0.005 0.50/0.50 0.50/0.506 1.00/1.00 0.00/0.007 1.00/1.00 0.00/0.008 1.00/1.00 0.00/0.009 0.20/0.20 0.80/0.80

10b 0.50/1.00 0.20/0.0011 1.00/1.00 0.00/0.0012 1.00/1.00 0.00/0.00

aAll Cl values have an uncertainty of60.2. The first and second values foreachCi( i 51,2) entry are for the2P1/2u and 2P3/2u components, respec-tively.

bFor O21(A 2P1/2u ,v1510), C350.30.

1275J. Chem. Phys., Vol. 112, No. 3, 15 January 2000 Photoelectron bands for O2

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

131.181.251.130 On: Sat, 22 Nov 2014 07:56:34

The deconvoluted PFI-PE bands~solid line! forO2

1(A 2P1/2,v153 – 8) obtained by simulation are shown inthe upper panels of Figs. 1~a!–1~f!, respectively. We alsoshow the deconvoluted PFI-PE bands~solid lines, bottomspectra! for O2

1(a 4Pu ,v1511– 16) and O21(X 2P3/2,1/2g ,

v1531 and 32) in the lower panels of Figs. 1~a! and 1~f!.The middle spectra~solid lines! shown in the lower panels ofFigs. 1~a! and 1~f! are composite simulated spectra, repre-senting the sums of the corresponding deconvoluted spectra.Good agreement is found between the composite simulatedspectra and the experimental PFI-PE bands forO2

1(A 2P1/2,v153 – 8). The simulated spectra~solid linesbelow the experimental spectra! for the O2

1(A 2P1/2u ,v1

59, 11, and 12) bands are also in reasonable agreement withthe experimental bands, except for O2

1(A 2P1/2u ,v1510).The poor BOS fit to thev1510 PFI-PE band can be taken asevidence of strong perturbation by other states. We note thatthe simulated spectra of O2

1(A 2P1/2u ,v159 and 10) havetaken into account the contribution from O2

1(a 4Pu ,v1

517and 18) bands, respectively. The simulated spectra forO2

1(a 2P3/2,1/2u ,v1511 and 12) have not included the re-spective contributions from that for O2

1(a 4Pu ,v1

519 and 20). Some weak features at the low energy tails ofthe O2

1(A 2P1/2u ,v1511 and 12) bands might have the ori-gin from the O2

1(a 4Pu ,v1519 and 20) bands. The BOS fitsfor the cold PFI-PE bands for O2

1(A 2P1/2,v150 – 3) areplotted as the lower spectra~solid lines! in Figs. 2~a!–2~d!,respectively. The deconvoluted spectrum forO2

1(X 2P132,v1528) are shown as the dashed curve in Fig.2~a!. We note that good fits to both the 20 and 220 K spectrafor O2

1(A 2P1/2,v151 – 3) are observed using the same setof BOS coefficients listed in Table I.

The simulation shows that prominent peaks observed inthe O2

1(A 2P1/2,3/2,v150 – 12) bands are mostly due to the

rotational branch heads~or the turn-around rotational transi-tions of rotational branches!. In Figs. 1~a!–1~f! and 2~a!–2~d!, the assignments of rotational transitions,O2

1(A 2P1/2u ,v1,N1)←O2(X3Sg

2 ,v950,N9) andO2

1(2P3/2u ,v1,N1)←O2(X3Sg

2 ,v950,N9), are indicatedby respective downward-pointing and upward-pointing stickmarks. The rotational branches,DJ527/2, 25/2, 23/2,21/2, 1/2, 3/2, 5/2, 7/2, and 9/2~corresponding to theM, N,O, P, Q, R, S, T, and U branches, respectively! are alsomarked in these figures. For individualDJ branches, the po-sitions of the first and second transitions are indicated byprogressively shorter stick marks, thus revealing the originand direction of the rotational branches. We note that for thev150 – 2, 6–8, 11, and 12 bands, whoseC2 is zero, theUbranch is not observed.

We consistently observed a mismatch between the posi-tion of the band maximum observed in the experimentalspectrum and that in the simulated spectra. As shown in Figs.1~b!, 1~c!, 1~e!, and 3, the band maximum observed in theexperimental spectra for O2

1(A 2Pu ,v154, 5, and 7 – 12)are found to be at slightly higher energies than those of thecorresponding simulated spectra. The intense, sharp peak at17.462 eV resolved in thev154 band is consistent with alocal enhancement by the near-resonant autoionizationmechanism. The structure observed for thev1510 band is

likely to result from a different mechanism. The rotationalfeatures at the low energy side of the main structure at17.979 eV of thev1510 band appears to be significantlysuppressed. This observation is manifested by the special setof BOS coefficients needed for the simulation of thev1

510 band.In the previous emission study,14 Coxon and Haley also

observed strong perturbation of the O21(A 2Pu ,v159 – 10)

states, which have prohibited the determination of spectro-scopic constants for these vibrational levels. They suggestedthe presence of a perturbative state nearv159. They alsopointed out that a possible candidate for such a perturbativestate is the2Su

1 state, which was predicted inab initio con-figuration interaction calculations of Beebeet al.31 and Mar-ian et al.32 This state has a shallow well and correlates to thefirst dissociation limit, O(3P)1O1(4S), as that for theO2

1(A 2Pu andX 2Pg) states. In this experiment, the specialsets of BOS coefficients needed for the simulation of theO2

1(A 2Pu ,v159 and 10! bands can be taken as support thatthese states are strongly perturbed. In view of the observationin the present experiment that thev1510 band is morestrongly perturbed than thev159 band, we would suggestthat such a perturbative state, if it exists, most likely liescloser to the O2

1(A 2Pu ,v1510) level at 17.98 eV.

B. Spectroscopic constants for O 21„A 2Pu ,v 150 – 12…

There has been a controversy about the spin-orbit cou-pling constants (Av

1) for O21(A 2Pu) in terms of both sign

and values. Stevens reported that theAv1 value for

O21(A 2Pu) at low v1 levels was regular with a value of

18.2 cm21.33 However, the later experiment of Bozokyet al.34 showed thatAv

1 is negative forv150, 5, and 6 andpositive for v158. A detailed discussion on this issue wasgiven by Albrittonet al.,35 who concluded that theAv

1 valuefor O2

1(A 2Pu) is most likely inverted at lowv1 but changesfrom inverted to regular betweenv155 and 6. Based on theleast-square analysis of the emission data, Coxon andHaley14 obtainedAv

1 values for v150 – 8 and 11–15 ofO2

1(A 2Pu), which are in accord with Albritton’s conclu-sion. We have listed theseAv

1 values for O21(A 2Pu ,v1

50 – 8, 11, and 12! in Table II. However, since theseAv1

values are small (uAv1u<6 cm21) and comparable to our in-

strumental resolution of 4–5 cm21 ~FWHM!, it is difficult tomake accurate determination ofAv1 for O2

1(A 2Pu) in thepresent study. For this reason, we have used theAv

1 data ofCoxon and Haley in the BOS simulation. It appears thatnegativeAv

1 values give a better fit for lowv1. We concludethat the PFI-PE data presented here are consistent withuAv

1u<6 cm21 for O21(A 2P3/2,1/2u ,v150 – 12).

The energy expressions used for O2(3Sg

2 ,v9,N9) andO2

1(A 2Pu ,v1,N1) are

E~v9,N9!5Ee91ve9~v911/2!2ve9xe9~v911/2!2

1Bv9N9~N911!, ~1!

and

1276 J. Chem. Phys., Vol. 112, No. 3, 15 January 2000 Song et al.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

131.181.251.130 On: Sat, 22 Nov 2014 07:56:34

E1~v1,N1!5Ee11ve

1~v111/2!2ve1xe

1~v111/2!2

1Bv1N1~N111!2Bv

1V261/2Av1 , ~2!

respectively, whereBv15Be

12a1(v111/2) and Bv95Be92a9(v911/2). In Eq.~2!, 21/2Av

1(11/2Av1) is for theV

51/2 component, and11/2Av1(21/2Av

1) is for the V53/2 component of the O2

1(A 2Pu) state for the regular~in-verted! case. The valuesBe951.445 63 cm21 and a950.015 93 cm21 are taken from Ref. 30. We calculate theionizing transition energies as

DE~v1,N1!5E~v1,N1!2E~v950,N951!

5Tv11Bv

1N1~N111!2Bv1V261/2Av

1 ,

~3!

whereTv15T(v1,v950,N951)5@Ee

12Ee91ve1(v111/2)

2ve1xe

1(v111/2)22ve9(1/2)11/4ve9xe922Bv9#, notingthat the IEs for differentv1 states correspond to energies forthe ionization transitions O2

1@A 2Pu ,F1(v1,N150)#←O2@X 3Sg

2 ,F2(v950,N951)#.

The Tv1 and Bv

1 values for O21(A 2Pu ,v150 – 12) ob-

tained by the BOS simulation are listed in Table II. We havealso calculated the vibrational spacings (Dv) as given in thetable. We have also included in Table II theBv

1 and Dvvalues obtained in the previous high-resolution emissionstudy for comparison with the present results. We estimatethat the uncertainties ofTv

1 andBv1 values determined here

are64 cm21 and60.005 cm21, respectively. From Table II,we can see excellent agreement ofDv between this work andthose determined in Refs. 14 and 15. TheBv

1 values derivedfrom the present work also show excellent agreement withthose from Ref. 14, with the largest discrepancy of 0.009cm21. We note that forv159 and 10, the IE andBv

1 valuesare determined for the first time.

Vibrational and rotational constants obtained were fittedto Eqs. ~4! and ~5!, respectively, to obtain the vibrational(ve

1 , ve1xe

1 , ve1ye

1 , andve1ze

1! and rotational (Be1 , ae

1 ,andge

1! Dunham coefficients.

IE~v1!5Te11ve

1~v111/2!2ve1xe

1~v111/2!2

1ve1ye

1~v111/2!31ve1ze

1~v111/2!4, ~4!

Bv15Be

12ae1~v111/2!1ge

1~v111/2!2. ~5!

These Dunham-type coefficients obtained by the least-squares fit are listed in Table III for comparison with thosereported in Refs. 14 and 15. In general, the vibrational androtational Dunham coefficients determined here are in excel-lent agreement with the literature values. Since theTe

1 ,ve1 ,ve

1xe1 ,ve

1ye1 determined in Ref. 15 are based only

on vibrationally resolved data, our present results should bemore accurate. However, we note that the rotational con-stants determined from the rotational analysis of theO2

1(X 2Pg)←O21(A 2Pu) emission bands are far more accu-

rate compared to those of the current PFI-PE measurement.We also performed the least-square polynomial fit of the

vibrational spacingDG(v1) as a function ofv111/2. Basedon the polynomial used in the fit ofDG(v1), we havesolved for the root ofDG(vmax

1 )50 and obtainedvmax1

531.5, wherevmax1 is the maximumv1 level sustained by

TABLE II. Spectroscopic constants for O21(A 2Pu ,v150 – 12).

v1

Tv(cm21) Dv(cm21) Bv1(cm21) Av1(cm21)

This work This work Ref. 14 This work Ref. 14 Ref. 14

0 137 518 869 871.5 1.050 1.0521 23.501 138 387 845 844.2 1.030 1.0233 23.162 139 230 817 817.0 1.020 1.0123 22.783 140 047 794 789.7 0.994 0.9920 22.284 140 841 759 762.5 0.975 0.9716 21.695 141 600 735 735.2 0.955 0.9509 20.996 142 335 712 708.1 0.939 0.9301 20.227 143 047 678 681.0 0.910 0.9088 0.68 143 725 658 654.6a 0.887 0.8872 2.69 144 383 625 627.6a 0.878 ¯ ¯

10 145 008 599 601.1a 0.870 ¯ ¯

11 145 607 571 573.8 0.820 0.822 5.512 146 187 ¯ 547.5 0.810 0.8011 6.1

aSincev159 and 10 were not resolved, these values are interpolated.

TABLE III. Comparison of vibrational and rotational constants determined here and in refs. 14 and 15 forO2

1(A 2Pu ,v150 – 12).a

Dunham constants This work~cm21! Ref. 14~cm21! Ref. 15~cm21!

Te 137 073.2~3.2! 137 068.5~4.2!b 137 069~8!40 572.5~0.2!c

ve 895.06~3.46! 898.10~0.33! 894.4~2.8!vexe 12.687~1.080! 13.22~0.15! 12.96~0.26!veye 20.081~0.013! 20.098~0.029! 20.021~0.007!veze 2.3231023(4.7631023) 9.431023(2.531023) ¯

Be 1.0598~0.0060! 1.061 900~31!ae 0.0180~0.000 23! 0.019 625 8~135!ge 21.60431024(1.71631024) 28.51(39)31025

aThe error bars are given in the parentheses.bThis Te value of for the O2

1(A 2Pu) state is the sum of theTe value@96 49664 cm21, Ref. 6# for O21(X 2Pg)

@measured with respect to the O2 ground potential# and theTe value @40 572.560.2 cm21, Ref. 14# forO2

1(A 2Pu) @measured with respect to the O21 ground potential#.

cThe Te value obtained from Ref. 14 for O21(A 2Pu) measured with respect to the O2

1 ground potential.

1277J. Chem. Phys., Vol. 112, No. 3, 15 January 2000 Photoelectron bands for O2

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

131.181.251.130 On: Sat, 22 Nov 2014 07:56:34

the potential well. We note that this result is in agreementwith vmax

1 531 derived by Akahoriet al.6 However, since thevibrational levels we accessed are only up tov1512, theresults of this experiment cannot verify the existence of asmall potential hump for the potential curve of O2

1(A 2Pu)above the dissociation limit O(3P)1O1(4S) as proposed byAkahori et al.

IV. CONCLUSIONS

We have obtained rotationally resolved PFI-PE bandsfor O2

1(A 2Py ,v150 – 12) in the energy range of 17.0–18.2eV. The IE values and spectroscopic constants forO2

1(A 2Py ,v150 – 8,11,and 12) obtained from the BOSsimulation of the PFI-PE data are in excellent agreementwith the previous emission and TPE studies. Spectral simu-lations have provided new spectroscopic constants for theO2

1(A 2P1/2,3/2u ,v159 and 10) states.A significant suppression of the rotational features is ob-

served for thev1510 band. This and the observed intensityprofile for the O2

1(A 2Pu ,v156 – 10) bands are consistentwith the previous suggestion that thev156 – 10 levels areperturbed by the2Su

1 state predicted byab initio calcula-tions. The PFI-PE data obtained in the present experimentindicate that uAv

1u<6 cm21 for O21(A 2P3/2,1/2u ,v1

50 – 12).

ACKNOWLEDGMENTS

This work was supported by the Director, Office of En-ergy Research, Office of Basic Energy Sciences, ChemicalSciences Division of the U.S. Department of Energy underContract No. DE-AC03-76SF00098 for the Lawrence Berke-ley National Laboratory and Contract No. W-7405-Eng-82for the Ames Laboratory. C. Y. N. acknowledges partial sup-port by AFOSR. Y. S. acknowledges the Wall Fellowship for1999–2000. M. E. acknowledges the GAANN Fellowshipfor 1996–1997 and the Dow Fellowship for 1997–1998.

1C.-W. Hsu, P. Heimann, M. Evans, S. Stimson, T. Fenn, and C. Y. Ng, J.Chem. Phys.106, 8931~1997!.

2C.-W. Hsu, P. Heimann, M. Evans, S. Stimson, and C. Y. Ng, Chem.Phys.231, 121–143~1998!.

3C.-W. Hsu, M. Evans, S. Stimson, and C. Y. Ng, J. Chem. Phys.108,4701 ~1998!.

4C.-W. Hsu, M. Evans, S. Stimson, and C. Y. Ng, J. Chem. Phys.109,1285 ~1998!.

5M. Evans, S. Stimson, C. Y. Ng, C.-W. Hsu, and G. K. Jarvis, J. Chem.Phys.110, 315 ~1999!.

6Y. Song, M. Evans, C. Y. Ng, C.-W. Hsu, and G. K. Jarvis, J. Chem.Phys.111, 1905~1999!.

7G. K. Jarvis, Y. Song, and C. Y. Ng, J. Chem. Phys.111, 1937~1999!.8G. K. Jarvis, M. Evans, C. Y. Ng, and K. Mitsuke, J. Chem. Phys.111,3058 ~1999!.

9M. Evans and C. Y. Ng, J. Chem. Phys.111, 8879~1999!.10D. Fedorov, M. Evans, M. Gordon, and C. Y. Ng, J. Chem. Phys.111,

6413 ~1999!.11R. C. Shiell, M. Evans, S. Stimson, C. Y. Ng, and J. W. Hepburn, Chem.

Phys. Lett.~to be published!.12‘‘High Resolution Laser Photoionization and Photoelectron Studies,’’ in

Wiley Series in Ion Chemistry and Physics, edited by I. Powis, T. Baer,and C. Y. Ng~Wiley, Chichester, 1995!, and references therein.

13H. Palm and F. Merkt, Phys. Rev. Lett.81, 1385~1998!.14J. A. Coxon and M. P. Haley, J. Mol. Spectrosc.108, 119 ~1984!.15T. Akahori, Y. Morioka, T. Tanaka, H. Yoshii, T. Hayaishi, and K. Ito, J.

Chem. Phys.107, 4875~1997!.16K. P. Huber and G. Herzberg,Molecular Spectra and Molecular Struc-

ture, Constants of Diatomic Molecules, Vol. IV~Van Nostrand, NewYork, 1974!.

17K. Ellis, R. I. Hall, L. Avaldi, G. Dawber, A. McConkey, L. Andric, andG. C. King, J. Phys. B27, 3415~1994!.

18O. Edqvust, E. Lindholm, L. E. Selin, and L. A˚ sbrink, Phys. Scr.1, 25~1970!.

19J. A. Samson and J. L. Gardner, J. Chem. Phys.67, 755 ~1977!.20C. Y. Ng, in ‘‘Photoionization and Photodetachment,’’ edited by C. Y. Ng

~World Scientific, Singapore, 2000!; Adv. Ser Phys. Chem.~in press!.21C.-W. Hsu, M. Evans, P. Heimann, K. T. Lu, and C. Y. Ng, J. Chem.

Phys.105, 3950~1996!.22C.-W. Hsu, M. Evans, C. Y. Ng, and P. Heimann, Rev. Sci. Instrum.68,

1694 ~1997!.23P. Heimann, M. Koike, C.-W. Hsu, M. Evans, K. T. Lu, C. Y. Ng, A.

Suits, and Y. T. Lee, Rev. Sci. Instrum.68, 1945~1997!.24G. K. Jarvis, Y. Song, and C. Y. Ng, Rev. Sci. Instrum.70, 2615~1999!.25S. Stimson, Y.-J. Chen, M. Evans, C.-L. Liao, C. Y. Ng, C.-W. Hsu, and

P. Heimann, Chem. Phys. Lett.289, 507 ~1998!.26P. H. Krupenie, J. Phys. Chem. Ref. Data1, 423 ~1972!.27Y. Song, M. Evans, C. Y. Ng, C.-W. Hsu, and G. K. Jarvis, J. Chem.

Phys.112, 1305~2000!, this issue.28R. B. Cairns and J. A. R. Samson, J. Opt. Soc. Am.56, 1568~1966!.29A. D. Buckingham, B. J. Orr, and J. M. Sichel, Philos. Trans. R. Soc.

London, Ser. A268, 147 ~1970!.30K. P. Huber and G. Herzberg,Molecular Spectra and Molecular Struc-

ture, Vol. IV, Constants of Diatomic Molecules~Van Nostrand, NewYork, 1979!.

31N. H. F. Beebe, E. W. Thulstrup, and A. Anderson, J. Chem. Phys.64,2080 ~1976!.

32C. M. Marian, R. Marian, S. D. Peyerimhoff, B. A. Hess, R. J. Buenker,and G. Seger, Mol. Phys.46, 779 ~1982!.

33D. S. Stevens, Phys. Rev.38, 1292~1931!.34L. Bozoky, Z. Phys.104, 275 ~1937!.35D. L. Albritton, W. J. Harrop, A. L. Schmeltekopf, and R. N. Zare, J. Mol.

Spectrosc.41, 89 ~1973!.

1278 J. Chem. Phys., Vol. 112, No. 3, 15 January 2000 Song et al.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

131.181.251.130 On: Sat, 22 Nov 2014 07:56:34