The Spectrum of Mulewla Oxygen Paul H. Krupenie Optical Physics Division, National Bureau of Standands, Washington, D.C. 20234 Ths i a tcl review end complain ofthe observed and prediied spectroscopic data on snd tens Og, Oj, end Of The oriole ible, infrared, Ramen, microwave, ané electron pare magne resonance spetira are included. Each electronic band ystems dcuseed in deta. and ables {bond orgae and beads are given The muerowave and EPR data are also tabulated. Special subjects tech asthe dsucation eneray of Or, perturbations. and predarocitions are discussed. Patetah here cures are Given, ae well a fnaes, FranceCondon integra, and other intensity factor. A Smmary table sts the molecule snstnte fr all own eletrome states of Os and O;, Eectonic ‘rutuce and theoreti elections are ele discussed. Kev words: Critical review elevtoni pestis moleela oxygen; potenti energy curves; rotational {pecrums spectoecoie constants Page 1, Imnroduetion . 88 cluding a2; 625.2 2. Blectronie structure of Ox and is fone. 426 naka progression ID), B*E;—X93; (Ta 21. Electronic structre of O. and O: 126 taka progression D, "ly a ty tly = 2.2, Numeral esieulations a 1p. Ogawe-Yamewahi transition 95° = a. Semiempirical - 427 HORS rn 436 Single configuration ab initio ert ee cae ° i oF, CL Ogg eee oae a ae 24, Electronic structure of O3... 428 B3(0F) —X2;{0.9780-660 ARR... 437 25. Ionization potentials, Rydberg series, pho BPE; 105) —X*3 10s) (650-600 AYR... $87 toionization, and photoelectron spectros- €Ey(Oj) XZ (Oz) (595-510 A)... 497 copy : 429 Second Negative system of 3, Electronic spectrum of O: and O3 :. 430 (530-1940 AR. 438 Sal a'dy-X Ey Infrared Atmonpheric sy . WEG w Tle; Fiat Negative system of tem (15,800-9240 AYR . 430 0, (8530-4990 AV 439 32, Bj oX 35 Atmospheric system O9T0— 318, c455-> 54%; Hopfield sytem of Of 5200 Jor. 431 (2000-1940 Ayy.. - 440 33. b23;—a Ay Noxon system 19,080 4)... 4313.34, Unclansfed bande : to 34. c!ZG—X 13; Herzberg Il system 4790- 4(02):Collision complex or Ox, sable dimer of 4490; 2715-2540 AYR 431 Os. High pressnre hands of oxygen and cima 35, Cha—X*2; Herzberg Il system (2630- taneous transitions in compressed and con 2570 A)R and high pressure bands 2925- ea ry mah i onygen.. - 5. Perturbations 443 on Car mia sew 900 RANE “s 3,7, A*3,—-X 82; Herzberg i system (a380- 5.2 Os, BSE; state. bed eee gon 53.03, @Thy state 43 38, B°3z-NI8; Schumann lunge system 6. Prediseocition of the BS; st 43 (5350-1750 Ay. versa 438 T. Dissociation energy of Or 4S; 43.9, Miscellaneous absorption transitions 1585- 8 Microwave spectrum of Os. 45 1140 A; Alberti, Ashby, Douglas bands in- 8.1 Rotational spectrum. : “5 8.2 Zeeman effect and EPR spectrum. 45 9, Raman and Zeeman effects. 46 9.1 Raman effect, 6 9.2 Zeeman effect in electron spectra... 447 4. Phys: Chem Ral Bote, Vl. 1, No. 2,1972428 Page 10, Potential energy curves a 1, Transition probability parameters. a7 Ta @°8,—X 85; Infrared Atmosphere 7 yg 112. B12; X53; Atmospheric system... 449 113. b'Z}—a'Ay Noxon system 49 Md, ASE} —1 9%; Herzberg | transition... 49 MLS. © Sr —X 8S; Hersherg Ml system; C8, !dy bands. cicrne 461 Band origins of the °5;—X°S; Heraberg 1 system 461 Band heads of the A-X system observed in emission (R). : 462 Band origine of the B °S;~X°2; Schumann: Runge system (Ri), 460 (a) emission 463 (b) absorption 464 Bande af the R35; 63; system of 0; 45 Band heads of the ¢ ™. : Miscellaneous unclassified bands. Observed fine structure (microwave) frequen- 3 b#35 system of OF 45 45 ies for Os (MH), 478 ‘Magnetic hyperfine structure lines. 419 ‘Magnetic dipole rotational spectrum of O... 479 Line widths (Av/p) in MHz/mm Hg. 479 Miscellaneous constants derived fromthe microwave spectrom, 90 Rotational constants for the X*%5 stat $80 Rotational constants for the ay state 481 Rotational constants for the 68; state... 481, Rotational constants for the ¢'S; state 481 40. 41 W 48 a 45 4K "1 6a THE SPECTRUM OF MOLECULAR OXYGEN 425 Pege |. Rotational constants for the C Ay state 481 Rotational constants for Whe A°Ej states. $01 Rotational constants for the BAZy states... 46 Rotational eonstants for the 8°; sta 82 Rotational constants for the aE; etatevecre 02 Rotational constants for the 93; state 482 Rotational constants for the 1Ay states ..re.e 482 Rotational constants for the "Ty state 492 Rotational constants for the X*fl, state of 0; 482 Rotational eanstants for the a Tl state of 0: Rotational constants for the A-Tly state. of 0: . sees 482, Rotational constants for the 62) state of 08, 483 Rotational constants forthe "Sq state of Of. 88 Pose 58, RKR potentials for electronic states uf O and on. Fi : 488 54, Lifetimes, Einstein coefficients, and oscillator strengths : 55. Absolute fvaluce for the O, B°S;—X*%; bands... 487 36, UV absorption; some special fesares. cee 487 5. Franck-Condon integrals... 488 58 Absolute band strengths Spe for the A"; — PTZ bandsaeesaeee 318 1, Three types of crossing between a bound and a repulsive state 519 2, Potential energy curves for O;, Oz, and Of... 520 Vol 1, No. 2, 1972ae PAUL H. KRUPENIE 1. Introdu ‘This report, the second in a series on the spectra of distomie molecules, includes & review of the literature (Ghrough February 1971) on the spectra of O7,+ Os, (02. and OF # and 2 compilation of eriticelly evaluated numerical data on band positions, molecular constent energy levels, potential energy curves, transition proba bilities, and other molecular properties derived from the spectrum. Though emphasis is on the gas phase, condensed oxygen is considered in discussions of Raman spectra, simultaneous transitions, eg. 2('3,) —> 2033), and the co called high preeeure bande. The cited references are mainly those from which the tabu: lated numerical data are taken, but include also some ‘which deal with interpretation and history (Os is the dominant molecule in thermochemistry and Js the second mast abundant constituent in the earth’s atmosphere. The dissociation, predissociation, and band Systenis of wuleculas oxygen play an Innporisc tole It the aurora, airglow, and nightslow.* Molecular oxygen is an important constituent of stellar atmospheres lineludins that of the sun) and possibly of planetary atmospheres. Meeks and Lilley {268° have susuested that the microwave spectrum (60 GHe) can be used es 2 probe of the thermal structure of the earth's atmosphere and a a space probe of planetary auospheres. In Taboratory studies below 2000 A, spectrographs have to be evacuated because of the strong absorption of lisht by the Os, B83;-X9S; Schumann-Runve transition; more than 99 percent of this absorption is due to the dissociation continu (1750-1300 A). The ionization potentials of O» are known from con: ventional spectroscopy of Os. from Rydbers series limits, photon impact (UV absorption), photoionization, electron impact, and photoelectron spectroscopy. An extremely useful review of electron spretmacony. electron impect, photoelectron spectroscopy. and Penning ionization has been given by Berry 38) Topics not considered in detail in this review include absorption coethicients, collision and photon cross sections, photoionization and absorption cross sections, collision detachment cross. sections, dissociative re ‘rambination coaficiente, ion electron rooombination, and charge transfer collision cross sections. Photoionization cross section measurements have been recently seviewed by Schoen [850]. Autvionization of Os Rydberg states has been qualitatively discussed by Price (319), Cook and Ching [93] have diseussed in detail the divergence of measured absorption coefficients. ir J relul seriens cwmverning spectee shall “md om a ne Of tea ym he i 4. Phys Char. Rl ato Vo 1, No 21972 bye mentioned: Wallace (39) has tabulated band head: for O, and OF transitions as well a erm valves ond vo tational eonstents forthe various states: Gilmore (161] he summarized a great deal uf information about the potential energy curves of Os and is ions. In addition, 2 caprae summary of sumry Or transitions and the oon tition under which they are observed i contained in 4 review by Herzberg [19] on forbidden transitions in Ainiomie malecdes. Sinlet molecular oxygen (prin pallya 8 and "33)has been reviewed by Wayne( 305) ‘The possibly important role of reactivity of oxygen wit pollutants is mentioned in «discussion of the relation 1 laboratory measurements 10 atmospheric chemist. Reproductions of specra ae not given here but ean be found in references cited in table I as well ay in “The Tuemudeaton of Mkealer Spetta™ by Teutve and Gaydon (316). 2, Blectronte structure of ©; and its fons 2.1, Eleetronie Structure of O; ond 0; ‘Tie onder of the molecular wabitale (O'S) for OF been given by Mulliken ° [284 Ve < Now < 2s < Dare < Bo < ite < It < 30 KOK 6 a 6 6 @ a ‘The bonding character of each orbital is designated as K inner, 6 (onding), rw tanitbondings 20 nd toy are virtually atomic Is orbitals: 244 and 20. are modified 2s; all other orbitals no longer resemble their separated tatomie) form. The electron configuration which gives rise to the ground state, 35, and also tothe observed states a8, and b;, is KK(204)!(2a)*(30y)" (As,)* The possible states originating from excited om figurations are summarized in able 2 (sis ane of the few paramagnetic molecules with an even number of electrons: the ground atate has a per manent magnetic moment due to unpaired apins. Pre diction of the ground state as °© was an early success of the molecular orbital method and an early failure of the Heitler-London method. Only by “reexamination” was Heitle-London theory able to prediet the correct sound state. Seven known stable states of the neutral molecule nse trom Ihe (Wo lowest electron congurations (able 1), A repulsive Tl state is inferred from ab initio caleu- lations and from the predissociation of the B 3; state. All hound etater which hhave been observed, This inference is based on the semi empirieal potenti curves forthe unobserved states de termined by Vanderslice et al. [387] and Gilmore [161], tnd from configuration interaction (Cl) calculations by 1o ground etateTHE SPECTRUM OF MOLECULAR OXYGEN Schavfer and Harris (844), (See also {187)) Five other sales, recently observed, arise from unknown or un ‘evtain excited election configurations. In addition these have heen observed numerous states belonging to Rydberg series. Bight states of Of are known. Configurs- tions for seven af them sre established The electronic structure for & number of states of 04 and Of has been given by Mulliken [286]. A weakly hound state at about 65-7 eV, dissociating to “P+ "Dy, was hinted at, but such state has not been observed. ations. of O» is Mullken's summary of the lower sven below. Electron ionized Oj state Very (fromXS;) XM, ee Grom X55) a Thy Ty Yay (rom '8,) Ml, (A= 195), 1s lzom "23) My (A 3a rom X #5) (oy) throm X35) ‘The configuration (I77)*(Im5)* includes three *Hl, states, only one of which as known (bound). All three should combine with the ground state of the ion. (See the recent discussion of these “Il, states by Dixon and Hull 115); thie is mentioned in see 7.5) Nordheim-Pischel [300], using valence bond approach, has estimated the relative separation between 4 number of low-lying states of O:, and has shown, qualitatively, that °H states would not lie at low ener: jes. Indeed, no stable Il state of O; has been identified. It was further concluded that 3; and #3; states arising from °P1 Dare repulsive. One of the two possible 3; states derivable from these atomie products, the B state, has heen observed and is stable, No Xj state is known, Miller etal [276] have shown from observed magnetic that the unpaired elec trons in the O; bond are primarily par rather than po and have coneluded that “electron conhgurations which put unpaired electrons in the po orbit are unimportant in the Og molecule.” Lal [249] roported cooing now bands in tho viciblo and UV regions which could not be fited into other known systems (e.g., BX, O: or A~X, 02). He there: fore inferred the existence of new state of O3, likely TI, which would perturb 4'll, around v= 6. No numer cal data were published. Such a #II state is unknown, and Lal's correlation of such state with “S,-+"D, is hyperfine structure of 10 2.2. Numerical Calculations Semlempicl Mofitt [280] has shown the need to include confi zation interaction to obtain reasonable excitation energies relative to the ground state and correct dissociation ar products. By using an intuitive modification of the single configuration MO method he predicted the relative fonder of Oy states deived from ain} and arjr} com figurations. Both 'S; and "&,) were predicted as bound (03; at higher energy), and both were predicted to lie below 48%;. Fyperimentally the tentative order is e 35, CAs, both below the A state, Both ¢ and C states dissociate to ground state atoms. Mofit also showed that the ionic state *2; dissociates to 'D+"P because of the non-erossing rule. ‘Watari [402], treating O. as a six[] electron problem, hhas predicted the correct order of several states of the two honest configurations. Linnett [254] has accounted for the double bond of Os and the bond order of O; as 25, Onville-Thomas [309], has used MO theory to hove that the O-0 band lenath in various oxveen-con: taining molecules decreases with increasing bond order. Fumi and Parr [150] have computed vertical excite tion energies for the states arising from the three configurations ajx;. mim}, and zim}. They used LCAO-MO wave fiinetinns plus CI between states ‘of the first and third configurations, but with neglect of ererlap and ompisioal rsies were in fair agreement with experiment and with the more complicated results of Moffit [280]. The ‘Ay was predicted to lie below "23, but it was not eertain whether these should be bound. Gilmore (161) has drawn them as slightly stable. Bassani, Montaldi, and Fumi [32[, using the method of Fumi and Parr, have calculated vertical excitation energies of I] and ® states of O/ relative to the ground state of the ion Gin the z-clectron approximation). Excitation energies of unobserved states 2y, Tl, 4, were indicated. The "by state fell near the + the other two lay far above it. The *y state would be about 1 eV above A Ily, ‘Thy several volts above *Mly, and %y several vols abuve “My. "y seems bound with D® about 2.5 eV. This state as sketched by Gilmore [261] would have a binding energy of about AE eV itis reve 1 e¥ abuve die suiaususn of Aly. The other two unulserved states are also stable in this approximation [321 Vanderslice etal. [3871 using a valence-hond approxi ion and experimental data, have estimated the poten: tials for 12 repulsive states derived from ground state atoms. These states were slightly altered in Gilmore's hngures (101) to aceount tor eonhguration interaction ine of momen Their rel Single cnfigurtion ob nite Itoh and Ohno [212] have used different sereening constants for the O atom and its jon (AO's for covalent and ionic structure) to calculate excitation energies (at 11.207 A) for welection states. In gener, their results agree with those of Mofitt [280] and Fumi and Parr (150). A reasonable value was obtained for the separation *3;~2E2. Contour diagrams were given of individual and total orbital electron densities for the 4 Phy: Chem Ret Da fol 1, No.2, 1972a8 PAUL H. KRUPENIE ground state. Bader et a. [25] have reproduced Wahl's more recent calculated individual and total charge density contours for the O; ground state at re (based on nearly H-F wave functions). Configuration iterocton abit Meckler [268] has built up LCAO MO funetions from 1 minimum basis set of Gaussian atomic functions to study the full CI for the X8%; and 'S} states. The ight atomic electrons were distributed among 12 spin- orbitals arising from 2p atomic levels, with na sp hnydridization. ‘Though his calculations give ro, D*, and w to within a few percent of the experimental values, larze potential maxima (~} D®) which peak at ~ 4A were calculated for both states. There is no ex- perimental evidence to eupport thie. In all cares Mocker assumed filled 1s and 2s shells, His results are less accurate than those of Kotani et al. [236], but are given 1s afunction of ‘Kotani et al. [286], have made very extensive ealeu- lations of electronic properties of O- and OF at r= 1.217 A by the LCAO method using « minimum basis set of Slates wibitele fon the all clcetron problem. Piftcen dimensional CI makes significant improvement in some ‘af these properties. The relative merits of various approximate treatments (for the ground state of O:) are compared. Properties calculated include, among, ‘others, D®, excitation energies, quadrupole ‘moment (Q), occillator strengths (f), and polariaabilities. The same nine states are considered an were treated by Fumi and Parr (150], and Itoh and Ohno [212]. Kotani et al. [236] have shown that excitation of 2s electrons (as was done by Meckler) produces significant changes. ‘A comparison is made of the energy levels calculated by Cl with more than a dozen different approximations A fifteen configuration calculation lowers. the total ‘every uf die proud stave by 5 eV, but sll ives «pre value for De. For in} configuration the *&y state was predicted to lie between *E; and "S;, as is tentatively found experimentally, Schaefer and Harris (344] have calculated potential ‘eneruy curves for 62 states of O- derived from the lowest states of O: #P, 'D, 'S. Twelve bound states were pre dicted, imeluding seven found experimentally, and several predicted easlier by others. Several new bound states are predicted. In the calculations, only @ mink rum STO hesie sot wae mend topather with » complete “valence CI” to give the correct dissociation products Screening. parameters were used appropriate to the SP ground state. The ¢ "Zz state was calculated to lie below both C#4, and 4 %8;. The new recults suggest ‘that the "Es state ('D+'S) has a larger binding energy than 14y(:D+"D), the reverse ofthe sketch by Gilmore [161]. The wditional bound states are $35 CP4D) a 59beV. 2A 2N6D+'D) a 823 eV 4~ 155 A 4. hye. Cherm, Bo: Date, Vel 1, No.2, 1972 AsCP+1S) at B71 eV, 7 ~ 2A. ‘More recent (unpublished) ealeulations by H. H. Michele are qualitatively the same. 2.3. Electronic Structure of OF° ‘The ion Of has been observed only in electron impact experiments. Porhane 8" cperten might he ‘observed in @ hollow eathode discharge in low pressure z, for Carroll and Hurley [77] have observed Nj* under those circumstances. Appearance potentials for Of were long ago given as 54.1 eV [163] and 50.1 eV [166] Recently, Hurley and Maslen [209] and Hurley [208] have theoretically predicted the energies, molec llar constants, and approximate potential curves for Of, derived from curves of Ne, ‘The ground state appearance potential was estimated as 35.92 eV, smaller by 15 eV than prior experimental values. Dorman and Morrison (118) have remeasured the appearance poten: tial of Of* to be 36.20.5 eV which lies close to the calculated value. The diserepaney between this value fat eusller, higher values Ion explatied. Daly ad Powell's [101] measured value for the appearance potential of the Of" (X state) is about 1 volt higher than that of Dorman and Morrison: the 4 Sj state enerey was found to he 24 eV above X "Ss. Hurley (208) has estimated the lifetime of the A state (O3*) for dis tion by tunneling at ~1 see., and 10 sec, for the X state (OF) 2.4. Electronic Structure of O; ‘The molecular ion Op is formed in gas discharges containing oxygen, and has been detected by mass portramotor in H. Oat 0.1 sum pressure (A. Sehmel tekopf, private communication). It is formed also in the D layer of the upper atmosphere, and is stable in solution or in ionic lattices [97 ‘The ground state configuration of O is KK (209)? (200)* (a4)! (*15)* (1m)? XM. ance oreo iet ‘The molecular ground state dissociates into ground state products OGP)-+O-CP®) which can form a total of DAstates. ‘The ground state dissociation energy has been crudely ‘estimated by Bates and Massey [34], by Creighton and Lippincott (9¢{ tram a semiempmical formula using the frequeney of a Raman line of KOs, and more earefully by Mulliken (286), By using an improved value for FAGQ,) {312} one ohtxine DUOz)=4.07 eV. Thie places 0, X My, about 04 eV below Os, X35. The dissocia- tion energy is cyclically related to several other quanti ties by the formule D4O,)-+EA(O;)= DOs) +EAO) ‘where the following values have been used: D%O3)=THE SPECTRUM OF MOLECULAR OXYGEN 420 6.116=0.002 eV) [62], BA(O.)= (0.40.02 eV {312}, and EA(O) = (1.478 + 0.002 eV)* (39). re for the ground state is tentatively given as 1.33 A uthich is a compromise between the values 1.28 A [2] and 1.32-1.35 A [171], both caleulated from crystal structure of KOz, and the value 1.377 A, estimated by the use of Badger's rule [49]. This approximate value of ris consistent with that expected by analogy with Fs (13-14 A), From the low-temperature fluorescence ‘emission spectra of O; in alkali halides Rolle (336) hhas obtained ©! =1152-1163 em-', wo%s =8.5 em, 6; (7802100) om, AG*Y)=1125-145 em” ‘These results were in good agreement with AG"(d) = (11452) cm” obtained from Raman spectra in solid KO, [97} Recently, the Raman spectrum of OF in alkali halo oryotats [104] he yielded an ontrapolated valvo for the isolated ion: AG"(4) ~ 1090 em~. This value leads tow ~ 1107 em! et! ~ 8-9 em Spence and Schulz [361] and Boness and Schulz [49] have. measured. vibrational spacings. of the ground state of O; produced by scattering of electrons by gaseous Os. They obtained te ~ 1080 em", exe ~ 12 cm! Combining this vibrational data with SOM) of Holzer et al. [194] gives compromise values we ~ 1113 fem"l, et; ~11.7 em"! (uncertainties are assumed a8 25. cm", wexe4 em"), There has een much speculation about a stable 4%) state of Oz which dissociates to OP) +O-(P) 184, 68, 211, 351, 92, 286}, but most likely no lowly stable excited states exist for O- (286, 252, S. Smith, private communication The Auorescence observed by Rolfe 36] has generally been assigned as "Tyy—> XT], bot an upper state sssignment as *Z; cannot be ruled out. Rolfe [336] had indicated that the upper state term value was 3.65 eV above the ground state of 0; (this was the longsvavelength limit of excitation, Hurst and Bortner (210) obtained evidence of excitation to an electronic state of OF (presumably *S,) which lies approximately 1 eV above the ground state. Recent calculations (see below) Inleate hae dhs excled state lies somewhat higher. ‘The lowestying configurations of O; which dissociate to grosindatate eonfiguratione af O-FO- thowgh: ppethaps not to ground-state products) are (2) Borg} (1m4)*(1rg)4 "Ths (b) (Ber9)*(Ur24) ()*(374) = "2, Bs, Me, Es, (6) (904)"(Ame)*(14)*(G00) "E2)s, PEA2), *B ay ay Bp EG, ye eb 2) Has anand ahr ce ac vale ch tele kore we ted oh Re inn (@) (Berg) ay) (rg) "35 (not to ground state ++ 0-) Until recently the structure of the possible states of 7 has been largely speculative; early estimates of the relative bonding of diferont states were ersde [34, 286}. A rough guide to these energies was provided by calculated values for analogous states of Fy (144, 192], Recent tentative results of SCF Cl ealeulations IM. Krauss, private communication, 2713] show numer- fous stable states besides the ground state ‘The stable states in order of increasing binding cenergy are: *5, #2}, Mys #25, 225, E21, ‘2p. Hey 48y, #5) 435, and X*Tl,, The states enclosed in brackets all have binding energies of ~1 eV. (Several af those states are included in figs 2) n Potentials, Rydberg Series, ination, and Phataalartran Spectenerany ‘The known ionization potentials of Oz have been determined spectroscopically (ie., from Rydberg series Tim his date on clectronic transitions in Of), by photoionization, electron impact, ion-molecule collisions, and by photoelectron (photoemission) spectroscopy. No known Rydhers series converges to the ground state of O}, though the first terms of such series have been tentatively identified (see sec. 3.9, 3.10, The lowestlying Rydberg term, expected at about 8 eV, probably synse) "Ty, 1s 2 dipole forbidden transition from the round state, and remains unobserved [161] In addition, interaction with other states causes mixing of configurations and contributes to the wealeness and difluseness of such forbidden Rydberg terms of excited Os, Configurations . . . mypoy and . . . xeupma do sive rise to allowed Rydberg series. Broadening of autoionization has been discussed by Price (319). A recent photoionization measurement of the lowest IP of Os by Samson and Cairns (342, 42], (12.0592 0.001) V—U7265 om-1 cotablishes the Of cnerey level scale. Samson and Caims have given a usefi summary of previous determinations of this IP by several methods. Details of the observed Rydherg series are given in secs. 3.9 and 3.10; only a few Rydberg states have hhad their fine structure analyzed, ‘Temuaive clessieatons of Rydberg states are given by Huber [197] and Leclereq (251); additional classi fications and related diseussions are to be found in apers which sive the experimental details (see also refs. 253, 125)). Even though the assignments of quan: tum numbers n and defects 8 for the lowest terms in the various series are open to some interpretation especially those of Yoshino and Tanaka [317)}, the series limite and Rydberg formulas still properly deseribe the Rydberg series. Rydhere states are known from transitions orig nating in the X and 6 states. The @*3;, a's, and 2. Phys Chom Rls Dat, Ve. 1, Nos 2, 1972a0 835 Rydberg states arising from forbidden transitions Ihave al theie fine structure audieds the other ohserved states or series are assigned as: 1435, BZ — Bo pow) Be, (apm) © 435 = Rou) (nda) My, (ns09) 2B For the a 'S;—2!2} and 2S; X@S5 trancitione, AG" and B” from fine structure analysis the vals forthe 2, X Ml state. ‘The electron impact method yields direetly the ver tical IP, not the adiabatic IP obtained by all ther methods. Frost and MeDowel 187] have measured by electron impact five states of O3, including ‘ly, ‘Ty Aly 25 and & new state at 21.58 eV. An OF term at this eneray was confirmed experimentally by Brion [61 ‘This encray may represent “removal of a rads or os2p lect with ot withust simultencows excitation of fnother electron in one of the other orbits” [147]. At 2248 eV evidence exists for a process leading to forma tion of O:: this may be a transition toa repulsive state of 3, with a minimum at 20.7 eV which is close to the Rydberg limit of 203 eV [147] and the limit +S Lichten [252] has tentatively identified @ metastable state of O- at~ 12 eV in electron bombardment studies, ‘The observed lifetime is > 10°%. Possible configuration ofthis state is 2p) (2p)? (ay). MeGowan and Kerwin [202} have observed four new states of O:, (produced by molecule-on olisions) at 23.9 eV, 27.9eV, 31.3 eV, and 1 eV uncertainty=0.2 eV). “Uwey of al (19. 11) wove the Bret to sty Of extensively by photoelectron spectroscopy (PES) More recently Turner and May [384], Turner [383] and especially Edgvst et al. (125] have extended these measurements, the Tatler even resolving overlapping vibrational struture of a Ty and Th. Price (see discussion by Dixon and Hull (115) has cmerved «nem bial ease with 3.3 eY Male wld 23.7 eV in the photoelectron spectrum of O: with 804 A photons. Dixon and Hull (115), on the basis of semiemvivical calculations. have auiened thi feature asa new “Ily state of Of, the uppermost “ly state derivable from a 3x3 configuration. It is not cert that this state is stable Edgvst et al. [125] have studied the FES of Os in the range 12-28 eV with both 584 A and 304 A photons They observed vibrational levels ofthe 4*Tly among the Iighor levelo of the a Tle stote-"The difwee state with vertical IP of 24.0 eV corresponds tothe state observed by Price Extensive vibrational structure was observed 1125] inthe following states: X#My, a Thy d Tly, B35, and B35, Geiger and Sebrider [158] have observed many of these features (as well as some whieh are still ux- identified) in the energy loss spectrum up to 21 eV. ‘Additional features [125] include a My state at 24.0 eV re close to PAUL H, KRUPENIE above this, ¢ 35, v=0-2; and at 27.4 eV, an unident- fied peak in an unexplained continuum extending from 25.5 to 28 eV. The features at 23.7 and 27-4 eV had heen seen earlier by McGowan and Kerwin 2621 ‘Turner [383] ohserved marked broadening for B*3;, »>3. He interpreted this broadening which begins at 207 eV, the dissociation limit for 8°;, as possibly ‘due to strong interaction between the B state (which he ‘assumes is Ez) and & state, Edgviet ot al. [125], with slightly better resolution, see sharp structure only ‘Samson [340] used photons ranging in energy from 27-66.6 eV to study the photorlactean apectrum of Os and found states at 28.5, 24.6, and 27.3 eV; uncertainty was estimated as 0.3 eV (eee also McGowan and Kerwin [262]). Signalto-noise ratio prevented resolution of possible structure above 27.3 eV. Samson speculated that the state at 23.5 eV (and possibly also the state at 28.6 eV) might arise from removal of a 2s electron, fand the state at 27.3 eV might arise from removal of 8 a72 electron, 2. Electronic Spectrum of O, and OF Os is a weak light emitter because, for most of its excited states, transitions to the ground state are strongly forbidden. Since "0 has zero nuclear spin, antisyn metric rotational levels of "Os are missing. ‘The B*3;—-X +2; Schumann-Runge transition domi: nates the spectrum of molecular oxygen. Diserete bands ‘of this allowed transition span the region 5850-1750 A; fan even stronger dissociation continuum spans. the region 1750-1300 A. Most other O2 transitions which have been observed under high resolution are torbudden hy electric dipole selection rules and are considerably weaker than &—X; the exceptions are transitions below 1800 A, mainly ta Rydherg states Retween 1800 & and 300 A continua overlap the diserete structure, Weak 6 °%}—X 8%; atmospheric bands (9970-5380 A) and a 'Q,—X*S; TR atmospheric bands (15,800-9240 A) have also heen observed. In contrast with Os, the three observed transitions for 0f are all electric dipole allowed: 4 Ml,—XM1, (2-), vosu-ivaw As 0°27—a°Nly) (I>), 830-3940 A; and c*E;—b*E; Hopfield bands, 2360-1940 A. Additional states of O; known from Rydberg series and photo spotty enable the Of dvublet and quartet states t0 be placed on a common energy scale 2A. wtp HOB y Infearcd Atmespherte 8ystem(16,/000- 9240 A)R ‘The 'A,~235 magnetic dipole intercombination tran- sition in“Os, predicted by Mulliken [245], was frst ‘observed by Ellis and Kneser in 1933 [126] in absorp:‘THE SPECTBLIM OF MOLECULAR OXYGEN sion by liquid oxygen. The fine structure was resolved and the transition assigned as 'A,—°S by Herzberg. {in 1984), who photographed these bands in the solar spectrum, Details of the tolar 0-0 and 1-0 bands were siven somewhat later by Hereberg and Herzberg [191 The spectra were taken with a three-me.er grating in first order (dispersion, 4.9 Ajmm). The observed a~X hand positions and rotational constants for the a state ate listed in tables 2 and 37, respectively. Rotational constants for theX state are in table 36. ‘The bands, inherently very weak, are characterized by 0, P, R,S, and Q-form branches; they showed two strong heads and no clear 2er0 gap. Relative aceuraey of the observed line postions in the 0-0 band is +0.03 em~! for unblended lines: in 1-0, it iy £0.07 can *. (See fimtonute Gta) fa so [291] regarding absolute :) The small difference between Fi=F, and F:—F; for the ground state accounts for verlanping @P and &R branches. (See also Babcock find Herzberg|[24] and tec, 8 af the microwave spectrum.) Of nine possible branches, only one pair remained unresolved: (@P@R). The other branches are SR, *R, MOQ. E GY ‘The 0-0 band has been observed in the day and night sirelows (165), and is weaker than might be expected ecause of reabsorption by the Tower atmosphere ‘The 0-1 band was found in emission from the night sky (91, 385]. Its fine structure was resolved, but the rnumerieal data has not been published[91). 3.2, b!Zj—X!%j Atmospheric System (9970-5380A)R ‘The D')—A°Zy magnetiv dipule imexcombnasion transition consists of doublet P and R branches, ie, P,Q, R, and #Q branches," the latter two forming & werk head Ohservations of these weak, red-devraded bbands date back to the work of Wollaston in 1802 and Fraunhofer in 1817. The most extensive and precise measurements of the b-X transition ate by Babeock and Herzberg [24] who studied absorption in the labor tory (with air paths up to 30 m) and solar absorption by the earth's atmosphere (100 km air paths). These meas- a jon of cates work by Boboock and collaborators [see especially Dieke and Babcock [112] ‘and Babcock [22] from which the transition was first identified as "2" by Mulliken (2831 The early work led Giauque and Johnston [160} to some of the earliest identifications of isotopes 1"'0"0, 00) by spectroscopy. The only sdditional detailed absorptio measurements are the fragmentary data on the 4-0 band by Ossenbriiggen [311], Band origins and heads for the 6-X transition are listed in tables 4 and 5, respect ivelyy rotational constants for the X and b states are listed in tables 36 and 38, respectively. ant ‘The O-X bands were first detected in the solar spectrum’ and later produced in the labor and 193 Babcock snd collaborators, culminated in a definitive work by Babcock and Herzberg [24] wherein are given eferences to the easlies measurenients, Baoils of this system have also been produced in emission from variety of sources, * but studies of these, with few exceptions. only provided identifieation of several add tional band heads not seen in absorption (able 5) Babcock and Herzbery [24] have meusured the fine structure of the following absorption bands: 0-0, 10, 2-0, 3-0, nL, 2-1, 3-1; "O"O and 8800, 0-0, 1-0. The spectre were produced using @ plane grating spectrograph of 30oot focus with recipro- cal dispersion of O88 9 019 AJmm For unblendad medium intensity lines, wavelengths have been deter: mined with an accuracy of about 0,001 A. Combination differences for bands with v"=0 were in agreement to ‘within 0.005 cm-'. These measurements have provided molecular constants for the X and b states for the three isotopes studied." w? was derived from AG"(), and the vibrational comstans of Curry and Herzberg [99], with AG"(1) = 156.3856 believed (then) to he accurate within 0.001 cm”. A least squares refiting by Albvtton ‘er al.[9] gives 1556.379= 0,006. 3.3. bX} —e 'y Noxon System (19,080 A) Noxon {801} observed emission of the Q branch of the b-a, 0-0 band at 19,080:=30 A in a low pressure discharge through He containing a trace of Os. This sare to have heen the frst shaervation af an electronic jon in 4 diatomic molecule hich is primarily (electric) quadrupole [301]. (This system is forbidden bby magnetic dipole selection rules since AA=2.) 34, 13; —X 35 Hersberg Il System (4790-4490; 2715-2540 A) R Absorption bands of this very weak forbidden transi: tion have been observed among the stronger AS; — XeZz bende by Uesabeg [109]. The ubsesved band origins are listed in table 6; predicted origins are given in table 7; rotational eonstants for the ¢'3; state are in tala 20 The epwctra wore sheared in fourth order of 2 2141. grating spectrograph: oxygen pressure was 2.7 atm. The molecular constants for the ¢ state were ‘obtained by assuming the values for the ground st from the work of Babcock and Herzberg (24) 4. Pays: Chem Ref. Date Vol. 1. Ne 2.1972asz ‘The ¢—X hands [with branches "P, "Q, "R, "Ql have structure similar to the =X system, but weaker by & factor 10%, and with much greater difference |B’-B"| because of the feeble binding energy of the c state. Unblended line positions are assumed uncertain by 0.1 om; numerous blended lines have considerably larger uncertainties. Long exposures resulted in 0.1 emt displacements between iron arc calibration spectra and <—¥line positions Herzberg [189] analyzed the fine structure of six absorption bands. Degen (108 has analyzed the structure of two feeble emission bands observed in an oxygen: largon aflerglow. These have been tentatively identified by Degen as the 0-7 (4491 A) and 0-8 (4791 A) members of the e=X system. The new »” numbering, stil tentative, is five unite larger than that provisionally assumed by Herzberg [189]. The spectra were taken by Degen on a 3 m grating spectrograph, exposure 100 hours, disper: sion 3 A/mm. Measured rotational line positions are tncertain by 0.1 A. The 0-7 band had heen mentioned, ‘earlier by Broida and Gaydon [68] but it had not been assigned: the 4791 A band is new. Many features of the nightglow spectrum te within several Angstroms of the calculated band positions [189, 103]. Broida and Gaydon [63] had observed some bands in an oxygen afterglow 4600-3600 A) which right be fitted into this Delandres array, but only with considerable uncertainty. Some bands observed in the UY airglow suggested an increase in Hereberg's pro: visional o* numbering by unity, but this had only been hinted at by Chamberlain (79) The numbering tsed by Degen {103} places the e state below both C and A; all three states derive from the eame electron configuration. A short extrapolation of vibrational terms makes it certain that the ¢13s state dissociates to ground state atoms. ‘The c-X transition, assumed allowed as electric dipole radiation, is only spin forbidden, and is likely so ‘weak because itis forbidden as electric dipole radiation fu the separated stots 3.5. C*A\—X*E; Hersberg Ill System (2630-2570 A) Rand High-Proteure Bonde (2024-2440 A) Fragments of two very weak, triple-headed bands have been observed in absorption by Herzberg [189], smhy wovd «350 pressure. These underlie the far stronger 4° j—X 8S, bands, Observed C-X subband origins are listed in table 83: rotational constants for the C®A, state are in table 40. Vibrational numbering in the € state is uncertain. Arbitrarily, Herzberg (189) assumed that the triplet ‘observed hy Herman [182], whose middle sub-band head fs at 2913 A, was the 0-0, and assigned the two frag ments he (Herzberg) observed to bands 5-0 and 6-0. The only earefully: determi AG(SY¥4)= 611.16 em-? is for Ay components. Assign 11 vibrational 4. Phys Chem: Data, Vol 1, No 2,1972 PAUL H. KRUPENIE ment of the upper state 58 °Ay it based on the assumed, lectron configuration. though the number of missing lines near the origin could not he determined. Thus, it could not be shown, unequivocally, that the upper state is not Il, &S, and R- and Porm branches were ‘detected. Both odd and even J were observed in the upper state, an indication that A> 0. Herzberg, [189] has assumed that the C-X bands are the freengae analogue af the diftue tiplotc, nearly acim cident with the positions of the 4X hnds, as observed by Wulf {415}, Finkelnburg and Steiner [141], and Finkelnburg [240]. These high-pressure bands (table 8b), fist observed under pressures of 60-600 atm, had bees, aseribed by Finkelnburg and Steiner (141), to the collisioninduced forbidden transition, *8—53. In 9 later paper Finkelnburg (140) aseribed tlese w vibra tions of @ collision complex. (See sec. 4 on condensed oxygen.) Observation of the diffuse structure sll nesrly the *P-43P dissociation limit [140] strongly suggests that the *A, state dissociates to ground state atoms. Experi- mental data is insufficient to allow a reliable potential feurve Ww be drawit Giluwse [261] tas skevehed tn a fragmentary curve just below that of A®E;, following Herzberg {189}. The C—X transition is forbidden only by the rule AA=0, =1. but is extraordinarily weak, likely because it is aleo forbidden inthe separated atoms. For the 0’=6 level, the triplet splitting between 3A) and 9A, is 145.9 em, indicating ease a coupling. Spliuing between middle ‘and short wavelength com: ponents of the high pressure triplets is 188 em (141, 182); the splitting between the middle and the long. wavelength cnmpanents ie 118 em! B.(CA,) is assumed a good approximation to the true By. values. By assuming further that Bs(°A3)~Bs:) 1f°8s)—B4CA.), the former being unobserved, (82) has been approximately ealeulated (189 in System (4980-2700 A) Twenty-seven weak hands (table 9) observed in the airglow have been tentatively assigned by Chamberlain {81] to the intereombination, electric dipole transition, C2A,—a'ds (eee table I of ref [81] for a lst of band heads). Identification is uncertain because of low dispersion (21 A/mm). The quantum numbering for the upper slate is bese uit that tenatlvely assumed by Herzberg[189) 37. ASEs XS; Herabers | System (4880-7420 Ay ‘Very weak absorption hands ofthis forbidden electric dipole transition were first observed by Herzberg (185); additional bands were observed at longer wavelength by Herman [182]. A detailed fine structure analysis of these strongly red-degraded bands which showed no prom= nent head wens Inter given hy Hereberg (187) 4th order, 21-2 grating; absorbing paths up to 800 m-atm; 3rd‘THE SPECTRUM OF MOLECULAR OXYGEN axa der iron Hines served as calibration). This was the lint 3-3" transition identified. with like muliplchy tor both states. The dominant Q-branch lines const of six components: “Qs, “Ps, %Q;, Ris, Res + Par, and 'Q, whith was observed forthe strongest bands Fin bee Rar, Qs Paz, and "Pag were alee shacrved ‘trom bands (187, Otnerved 4-3 bands and rotational instants forthe A 3} state are given in tables 10 and I reapetively. A Deslandres array of predicted band ions ean be found inthe paper by Herabere (17) Duley {119} and others tentatively dented some bands ofthe UV night sirlow (8003100 A) a mem hers of the A-X system, Finally, Chamberlain (80} dently condrined these Wemtfestion by fine struc: ture assignments. Meanwhile, Broida and Gaydon [63] honed ACT han items from, erator afterglows (table 11). (See also Gaydon [153].) Observa- tion of new progression showed that Hersbergs 1982 9” numbering (187) should be increased by unity (ieraberg’s original numbering [185] was tated by 4 Barth (28,29, 30} obeerved this eystem in oxygen and snygenitrogen (air?) alterglows 4500-2500 A) sind wae able to past ly resolve rotational otructwree Unless further revision in numbering is necessary, the last vibrational level in A*3z before dissociation is 11 ‘A summary of Herzberg’s fine structure analysis 1187] follows. Relative wavenumbers of unblended lines fare assumed accurate to 0,04 em; very weak struc ture at longer wavelengths is accurate to 0.1 em! Rotational constants for the A state could wut be deter mined from combination differences because 20 few members of the O- and S- form branches were observed. Instead, differences B!—B and D;~D! obtained {rom Q= branch lines were used together with precise values of Bj and D5 (24). For the 1-0 band only Iragmentary branches were observed.) 2G values drop drastically with v's « polynomial of a few terms is @ poor fit to the AG's. Bi decreases nonlinearly with v'. Df increases even more sharply swith o's @ power suies fe « youn ft iy dis data, These complications together with the revised quantum numbering foro lead to the following tentative formulas fited tov’ =1 104 (see footnote to table 41) B= 0,910530.01416(0'+4)-0.00097(0"-+H)* Df =18.79-0.3060-44) 40210(0's )=I 10-4 G(w') =799.08(0'+3)-12..16(0' +4P—0.550(0" +3) Tot! A007 15-+7R6 S1'—12.9850'*-0.5500'9, ‘The short wavelength limit of these rapidly converg- ing bands made necessary. slight revision in the dis- sociation energy of Os (but see see. 3.8 and sec. 6 on ani ‘Bee and Gayo elt tn i on dissociation energy). In sec. 4 on high pressure and condensed oxygen spectra, there is briefly discussed ‘confusion of A-X bands with certain high-pressure bands, Triplet splitting for A®%y shows Fy>F,> Fe: for XA3p, beyond low Ny F,> Fy> Fy. Splitting of both states is fitted roughly by the formulas of Schlapr [846]. (See also theory of Present [318], The positive ‘quantity — (2\-+a) inereases significantly for high [8 to 11}, indicating increasing deviation from Hund’s cease b likely to ease cy near dissociation. y is virtually zero for v'=1 10 6 but becomes negative thereafter, ‘The 11-0 bend was observed till W’=15. A diffuse feature where N’=17 lines are expected likely indi: cates predissociation by rotation [187] 3.8, B°X5 X22 Schumann-Runge System (5350-1750 A) R Ihe #°27—A°27 transition isthe most extensively studied system of molecular oxygen, with more than 100, single-headed bands identified. These bands are red: degraded and show no prominent eade (origin head separation is only a few cm). The dissociation energy of O, (D"=41260-+ 15 em), one of the most precisely known for any diatomic molecule, is derived from the convergence limit of the upper state [62]. Band origins of is very intense system, for both emission and absorp: tion, are listed in table 12. Rotational constants for the Nand states are given in tables 36 und 42, sespectively ‘Schumann, in 1903, first observed 14 diserete absorp. tion bands of this system lying below 2000 A, though he ‘was not able to measure wavelenaths. Absorption bands were later observed by Steubing, Bloch and Bloch Ducleaux and Jeantet, Hopfield, and Leifson; these measurements are now of historical interest only. In 1921, Runge [338] first produced B-X bands in emission by using a high-voltage are in oxygen. A summary of the ‘more recent measurements on this system follows. Fuchtbauer and Holm (140) cboorved @ number of ‘band heads absorbed by heated oxygen (2000-2300 A, dispersion ~$ A/mm); their assigned o' values should be raised by two units. Only the 81 band, from among, their observations, has not heen recorded more recently. Ossenbriiggen [311] reported fine structure of a dozen bands measured from plates taken by Fiichtbauer and Tiolm (249); dese have been remeasurea more recently (table 12). LochteHoltgreven and Dieke [256] classified mi merous lines, including many meatured hy Fecefeldt 137] from plates taken by Runge [338] with a 6.5 m grating instrument in second order. The hands (4370- 3100 A), looking like © many-lined spectrum, were emitted from a high-voltage are in oxygen, and were characterized by low v’ and high v*. (o'=0, 1, 2; v'=11 to 21), P and R branch lines fell close to one another, giving the appearance of doublets. For N = 20, the lines split into two components, with the long 4. Pays ham Bal te Vo Y Na 2, 197244 wavelength component stronger At very sugh NV [the highest reported was 89] the stronger component split into two, producing a tnplet structure. In bands with »'=2, the short wavelength component was too wesk to be identified with certainty. Some of Fesefeldt’s meas: remente on two plater differed by up to § em! Line intensities quoted by LochteHoligreven and Dieke ‘were taken from Fesefeldt 137) Pulskamp [323] photographed weak bands of the v'"=0 and 1 progressions including the frst measure ments of the 0-0 and 3-1 band heads (dispersion 2.76 Amm). Vacuum wavelengths were given for these Ineads and some fine structure. A discharge ube ‘continuum was used as background source. Curry and Herzberg [99] measured bands observed in third onder of a 8 m grating snectroaranh (dispersion ‘Mmm; hydrogen lamp background source). Bands measured included 1-0 to 5-0, 41 to 7-1. The fine structure of the 1-0 and 4-1 hands was given for the hist time: these bands had been observed under 1ow resolution by Palskamp [323], P and R branches were separated, but no doublet or triplet separation was cheorved. Improved constonte far the Y and R atatoe were obtained. Sharp line positions were assumed ac: curate to 20.1 em (poor lines to 0.8 em") Knauss and Ballard [233] photographed absorption bands in the region 1925-1760 A produced by # 3:m vacuum spectrograph viewing a condensed discharge through a capillery. Rotational analysis was given for Ivands of the (v' 0} progression with o’—4 to 15. Their band origins together with those of Curry and Herzher [99] were fitted hy a formula which gave a poor prediction for the 0-0 position. Their formula representing Be values for the B state gave too small values for Ba and By, Lines of the 14-0 band were broadened; in 15-0 they were doubled, with splitting much greater than that found by Lochte-Holtgreven and Dieke (256). Very complicated structure was observed for 16-0." Millon and Herman {278} observed several band Iheads in an uineandenced diceharve in On ‘These were identified as 0-4 to 0-1; 1-10, 1-11. Feast [133] questioned some of the vibrational assignments, (See the discussion of Feast below.) “Lal [246], using a hgh-trequency discharge, produced ‘more than two dozen bends in the region 4490-2450 A which were attributed to the B-X transition. Details soncorning thece hands were never published. many-line structure of these nearly headless bands makes identification difficult. Lal's vibrational assign- ments were questioned by Fesst (182, 133] because of the irregular Deslandres array. Feast [132] observed bands in the same region and indicated that “the heads are so weak . . that analysis has been made by identi (259) mane OT A a apt pte re he mere sree Me et Bw PAUL H. KRUPENIE fication of the rotational structure of exch band and ‘heads’ that seem present with small dispersion are found, with higher dspersiom to be merely due to ‘chance overlapping of structure from several bands.” Lal’s bands ace listed in table 30 as uncertain excluding ide with previous meaeure- Feast (134] has given rotational analysis for bands in the region 3100-2500 A emitted from a high-voltage are in oxygen. The observed bends include 1-12, 0-11, 11, O-10, 1-10, 1-9, 2-9, 1-8, 2-8, 2-7. “There are 100 many blended lines inthe region 2-10 and 0-9 to Allo detailed analysis." Spin triplets were not resolved. ‘each band showed only P and R branches. Measure- iments were assumed accurate to 0.2 em Garton and Feast [152] gave a preliminary report on absorption from heated oxygen and assigned vibrational quantum numbere for bands below 2500 A. Higher J values than the data of Ossenbriiggen [311] were ob- Sesved. Line positon ain bn vin weve given a unpublished manuscript by Fesst and Garton (136). Origins are given (152) for bands 2-10, 0-9. 3-8, 3-7, 26, $6, 456, 2-5, 5-6, 3°5.4-5, 5-8, 7-3. 6-5, 5-h 6-4, and also fine structre of yet other bands (emission and absorption) whose origins and rotational constants hhad been published previously. Hereiog and Wicland (101) have extended vo longer wavelensth the absorotion studies of the B-X bands. They messured fine struce ofthe 1-8,2-8, 1-7, and 2-7 bands (80 atm, 1070 °C; cell lensth = 20 em. Feast and Garton (136) have given a detailed line list of BOX ands which they observed (S100-2125 A) in both absorption at 1900 °C and emission from a high- voltage aie These inchule 0-12, 0-11, I-11, I-12, 1-10, 0-10, 1-9, 2-9, 1-8, 2-8.2-7; bands not reported previously include 2-10, 0-9, 3-8, 3-7, 2-6, 3-6, 4-6, 5-6, 2-5, 3-5, and 4-5: remessurement of bands previously observed by Ossenbriiggen [311] include 5-5. 6-5, 7-5, 5-4, and 6-4. These bands extended to high J (~59 max} (Unidentified lines, generally weak are ncluled in die lst) Dispersion was 42 cm~mm at 2475 A and 36 em~/mm at 2200 A, Measurements were assumed accurate to 02 em-! (0.01 to 0.02 A) for sharo lines. Some lines of the hand identified as 1-7 by Hercang and Wieland (181] were included here as belonging tothe 3-8band. Darie [128] observed emission of more than 30 band heads im # water/fuorme fame Wow=2400 A). The several new bands were subsequently reported by Rakotoarjimy, Weniger, and Grenat (324). Durie’s band heads were sharp because the low effective rota. tional temperature resulted in less overlap of rotational structue than is usually encountered, Brix and Heraberg [62] have made the most precise measurements of the B-X Schumann-Hunge bands with high v', and from their convergence limit, have determined a seemingly definitive value for the dissocia- Sion eneray of Oz (but see the Emitation mentioned in sect. 7). The strongly red-degraded bands were photo:THE SPECTRUM OF MOLECULAR OXYGEN sve in fourth order of @ Sm vacuum spectrograph ipersion 0,65 0 04 A/mm; resolution 160,000, xyset' pressure 3 to 350 mm). A Lyman discharge ‘he was background source forthe 50 xm long absorp- w ube, The region sted i detail was 1804 4 (11-0) 2'1750°A, the link uf dinvicte abovptione o! Oaude seve observed for '=12 to 21. The last bound level, 22, was extrapolated. All measurements, were ‘wale against second order iron lines as. standard srivelengts, Relative wavenumbers of unblended lines sve accurate to 0.05 em; absolute error was 0.2 em” Fine structure, band origins, and, for the B ‘ate, Mbradonal quanta and routlunal constants were Electric dipole selection rules allow twelve posible sramchoot Pasar Rasen PQuny Pann "Oars "Ques MPa The six main branches with AJ=21 were identied Aiur low Nj also, some lines ofthe sx satellite branches (ANm=1, AJ=0.2)) as well as to electric dipole Tushidden branches "Ror and "Ps (AV=23, 8/1) cere identified. This was the fist case known where lurbidden branches with AV>1 have been observed {raw allowed (—¥) wansition Near the convergence limit, the triplet spliing was the same order of mogutude as the rotational tracture: The triplet opliting wae diceuoood by Brin and Heraberg (62) in terms of Schapp's formulas (346] tnd deviations thereirom. [Since then, more extensive theory of splitting for the ground state has been given by nkham and Strandberg [375] and others; sce sec. @ on microwave spectrum.) Previously, triplet spliuing hhad_ been observed only by LochieHoltgreven and Dieke (255) and Kauss and Ballard (298) For 0” = 16, perturbations in structure occur accom tvanied by the appearance of extra lines and intensity Nvgulace, Term vahwo for the Py componcate of the B°Sz state showed slight deviation from Schlapp’s formulas [346] T.—oy=Fe{N)=B.N(N+1)—D,N* (N+1)*. Short branches and large perturbations limited he use af combination diferences. ‘There remained approximately 100 weak unclassi: fied lines just t0 long wavelength of the dissociation continuum, Ferturbations prevented assignment of these lines. The perturbations indicate the presence of another stable state dissocitioning to °Ps-+!Ds, ‘Bands having v 2 19 ne tugher then the previously accepted dissociation limit. 022-0) is predicted to lie at 57127 em" (F3)-Some lines of 22-0 were detected, the quanto aqsignmente are cnenrain. A. chert ‘extrapolation beyond the origin of the 21-0 band gave 8 convergence limit 5712825 em”? which was in close agreement with one derived from the 4°2;~X"35 bands, provided that pairs of dissociation products included the Pz state. Ths led to D4O;)=41260215, em"? (6.115620.002 eV). The magnitude of the uncer fly reflects she diference bevwmeen he Fail all sociation limits, 158686 cm”, which represents ‘D— 435 “Pa, This agrees wit the values given by Moore 281 tovwithin about gem Ratotoarimy, Weniger, and Grenat [824] have extended observations in the UV, bain absorption td emission, by using a hightrequency dicharge In ih atumeyliie uf unpgun Bight nest bende one dented in absorption (270-1980 ) and 28 new bands in emission 6800-2100 A). Dispersion 4 A/mm in emis: Sion: 2 Alm in absorstion) Detail were not given, nly @ Deslandres table of band origin wavelenghs Herman, Herman, and Rakotoajimy (83), ving + sinilar emission source, obpered. numerous bands, nd reported tine srucre, rotational constant for Ih states, and band origin wavelengths Fine struture was given for IDL, 22ly 2-22, 2-23, 2-24, 2-25, Bao, 2°28, and 9-24, Origine wore lee given for 0-20, 3-25, 5-27, but nofinestructure. evs and Garvin 31 have reported several absorption beads produced allowing ash photolysis of NO Fitzsimmons and Bair [142] observed heads of 30 abeorption bends. with bigh o". These High energy states were proiuced in secondary processes following Povolyels of ovone. OF these bande all bat five had Previouly been sen in emision; the ther ve hed not been observed before. sere [308], using © quate epcetrography photo graphed the absorption spectrum excited by an se silent discharge (dispersion 1.46 AJmm at 2100 Aj. Seventeen bands of the B-X system were identified, several new. Rotational constants for both upper and lower states were obtained, including several for low v for which there had heen no previous determination, ‘The primary alm was the observation of the BX system from excited vibrational levels of the ground state and the clarifieation of discrepancies in rotational eanetante for the R atate The ohenrved region sens 2250-1970 A, with numerous overlapping lines in this manyrline spectrum. Resolving power was too low for ‘observation of electron spin splitting, Using previous ‘data as well as his own, Ogawa observed that the By versus » plot for BS, showed positive curvature for 0=0 10 15, and was virtually straight for v= 16 to 21. ‘A formula fitted tothe first group of levele wae given = 0.8184—0.01238(04+4) -3.3% 10-*(044)* 8.7 1K += 1.3 10-4(0 HH), Ogerea and Chang [306], using a3 m-vacvum spectro: rep Glnces sesiprocal operon, £08 Alam) phote fraphed 34 absorption bands in the region 2000-1770 A. ‘The observed bands, half of them new, include members of progressions with «"=0, 1.2, 3. detailed line list is given, logether with a table of band origins. For the (©) progression, the band origins differed by an average 0.3 cm from adjusted values of Knauss and Dalla (288) fume 13) More than 1000 new lines belonging to absorption 4 Oc emer et Bata Va. Na 2.1972436 bands in "=, 2 progressions have been measured by Hudson and Carter (19a) inthe region 2020-1750 A ‘hie puliched line lit pan the region 1095-1877 A (clave accuracy 20.02 A}. See also see 11.6) ‘Acksrman and Biaume (3) have remensured under high resolution bands 6-0 though 19-0. They used the fame instrument as did Brix and Hersberg (62) The measurements of Curry and Herzberg [99] and Brix tnd. Hersbere (62, where they overlap. AcKerman’s, Sein lowe agieetcnn ut cal ifercmees i ah fete of data are found for By and Da for s'™=12, 13 (See table #2 fostnote) ‘The O:. BOX. O-D transition enerev and the zero point eneriy ofthe B35 sate have ben debated (eed Teel) for moe than 30 years beeaure ofa misreadingof the ear teratre. (Compare, for example vee 162} and{9]) Improved Yaluce of bot quantities have been shasined from measurements of Ackerman and Biaume Bh 3.9. Miscellaneous Absorption Transitions 1585-1140 A Alberti, Ashby, Douglas Bands Including «i a1 Fiat Negative System of 0; (8530-4990 AV The viole-desraded b-a band svatem of Of was frst observed by Schuster and by Willner (1877-8) near the negative pole ofa discharge tube filed with oxygen. Additional low resolution observations were made by Steubing (1910, ‘The first partial fine structare resolution was by Frerichs (1926), who divided the tbserved lines int two branches per band, using Hol Tand's published measurement of spectra from hollow cathode emission taken by Cari, A band observed at 53900 A by Steuhing tnd shown Inter to the 2-2 band) was not detected by Fresichs because of overlapping Totational structure produced in the hollow cathode 439 Additional band heads have been observed in high frequency discharges (low rotational temperature cources) by Mulliken and Stevene [287], Roxiky and Schmid {54}, and Singh and Lal [354}. The most promi- rent heads are listed in table 27; the origins are in table 26. Rotational constants for the a and b states are given in tables 49 and 51, respectively. Mulliken (284) assigned these bands as ‘3)—‘Tly from the probable electronic configuration of O;, and assumed the vibrational numbering of Mulbken and Stevens [287]. Fine structure analysis of this system by Nevin (291] confirmed Muliken’s assignment ‘Nevin [291] ced both a diccharge through helium containing & small amount of oxygen (total pressure 3~$ 1mm; oxygen pressure ~0.1 mm) and a hollow cathode to produce the six hands he studied. The speetra were photographed in secund order of a 21-f grating spectro- graph (dispersion ~ 1.25 Afmm; resolving power ~ 180,000). The prominent band head is formed by the ‘Qs: branch (= 18 cm" trom the ong}. Wavelength calibrations were derived from neon in the discharge and an iron arc. Absolute wavelengths were assumed rorrect 19 O01 A: polative enonrary wae seetmed mnch better [29a Wavenumbers"” were given to three digits beyond the decimal; use of current conversion tables would change these values in the last digit. For the 0-3 band shifts of =0.08 em-" between plates ‘were not eliminated. Absolute values of wavenumbers were averaged from measurements of three plates [le]. Lines possibly belonging to the 1-4 bend were observed but not analyzed. Faint lines of unknown origin were observed near the °Qss head, ‘The *E rotational energies have been accounted for by the theory of Bude [64]." The experimental fine struc- ture parameters are €=0.1887 em! and y= 0.00033, em [2913]. Separations MF:,(J) and AFs(J) for the Inverted *T] states," ealeulated from modified formulas originally obtained by Brandt [56], are not in complete agieement with those derived from experiment [291 Nevin [9012] soncidered this devietion de toa probable perturbation of the 4Il state by *35, ‘My, or “dy derived from O° 2D.) + O(P,). The interval F5(J)-F¥()) is bout three times that of FS! (1) ~FY'(J). The perturba tion displaces the F, level away from Fs, and F: toward Fy, No detailed study of this perturbation has been made. ‘A-doubling is very small in ‘Tlyy and *TIy2; it increases linearly with ) in ‘Ty. In 1_,it increases at first, and then decreases for highest J. A-doubling for ‘Ilys and ATI {65] increases systematically with v. Formulas have been derived by Bud’ and Kovacs 65] for the A-doubling in a Il state, intermediate in coupling between Hund’s nT bate av en ped se ates fe el Se 4 Py: Chem Ret a40 PAUL H. KRUPENIE case a amu v, comparison veitn Nevin's data seems sdequate, ‘Avcase (@) Il state gives rise to 48 branches for @ ly transition (ease (b) gives 27 branches). Since level F/(N} is blended with F{(W), and level F{(W) is blended with F;(N), the branches involving them are blended, and the number of possible branches is re- duced to 40; all have heen observed. The intensity dis tribution in these branches isin agreement with Budé’s theory (6, assuming the a Ml, state is intermediate in ‘coupling between Hund’s esses (a} and (b) Dufay et al (122 have produced several new en bands (8847-6750 A) by highenergy proton bombard ment of oxygen. Recently, Wonigor [408] nstended monciramonts of this system into the photographic infrared region, He observed emission of ade discharge by using dispersion of 3 A/mm (assumed precision of about + 0.07 cm~' is ‘ot confirmed in the reexamination by Albritton et al {9). Observed hands include 0-3, 4,5, and 1~4,5, 6: de- teiled line lists were published only for 0-3 and I-4, For the 0 9 baad head; Weniger Hsted the Q, head rather than the most prominent °Qy; head. A-doubling is small (~ 0.10 em) forall eomponents of aTT except “Ts, ‘where itis several times larger. Weniger pointed out that the coupling constant 4 is negative, and has been mis- printed by both Herzberg [186] and Rosen (337), Kovies ‘and Weniger(238] have used all previous measurements tf Nevin and Wenlger co obcain values of 4 for all v= 6 (see footnote under a MI, n table 1 and also see. 5). 3.13, €°2) 7 6; Moptiold system of OF, (2360-1940 A) V In 1920, Hope (195) produced five single-headed, red-degraded bands in a condensed discharge in a He +0; mixture (2220-2130 A), He assumed the progression Delonged to a new system of Os. Herzberg (186, p. 559) Suggested thatthe lower state ofthis new system might, be the H state, tentatively assigned as°fl by Price and Collins 320). Tanaka, Jursa, and LeBlane 370] extended. Hopteli's vimcrvatins (2960 1940 A}. Dy using @ total pressure of 20 mm and Os pressure of I mm in the discharge they suppressed the stronger 4-X, Oz system which epanned tha cama ef appearance of the fine structure suggested a 3-3. transition, Tentative vibrational constants were obtained by assuming the 19403 A band was 0-0, Similar AG values for progression IL of Tanaka and Takamine (371) suggested that the upper state of progression TT ‘was possibly the lower state of the Hopfield system [370) LeBlanc [250] resolved the P and R branch fine structure for the 0-4, 0-5, and 0-6 bands; other bands Ihnd other spectra superimposed (dispersion ~ 1.5 Afmm at 2100 A). From the close agreement between the vibrational constants for this progression and values steal region The 4 Phy: Chem, et Doe, Wok. 1, a 21972 ‘optaiea trom fine structure on another system by Nevin [201a}, the lower ta lontifed as AS; O8 Assuming an allowed transition, LeBlane assigned the ‘upper state of the Hopfield bands as ¢ *3;. It was then understood that the slight diffuseness of the observed lines arose from the superposition of four lines in the quartet transition. ‘The energy of the c state (Ta) is then [417, 370 19809 om-'=24.561 eV above Os, X°E;, Codling and Madden [89] have recently observed new Rydberg series converging to both »=0 and 1 levels of the ¢*Zy state, For tbe estate AGG) ~ 1540 emt, close to that of the ground state of Os, The v=0, 1 series Timits are 19812530 em-! and 199665-+30 cm, These agree hin ‘experimental error [89]. ‘The cb band origins and heads are given in tables 28 and 29, respectively. Rotational constants for the 4b and states are listed in tables 51 and 52, respectively ith the priow data to 2.14, Unclassified Bands (Other sections mention bands which ae unclassified or whoveclsrfcation i in doubt; such bands are tssembled in table 30, Bands in the same eategory, but aot tabulated In his report, are brely discussed below Tn an oxygen afterglow, Broida and Gaydon (63 produced "tinier af wnat. singlsheated,” te Aegraded bands (4855-3655 A), Several bands didnot corclate ‘with known transitions; others sould toute tively bo antgned to trailons 4-X, eX, orto en ttberwise unobeetved transition, Ab. Band. head positions are uncenain by several Angsttoms. Una Diguous assignments to the transition A~b cannot be made hecause the observed bands have such an irreg: ular Franck-Condon locus. 4. (O:}z Collision Complex oF Os, Stuble Dimer of O:, High Pressure Bands of Oxygen and Simultaneous Transitions in Compressed and Condensed Oxygen. ‘This brief summary of the experimental data on the spectra and physical properties of compressed and Pandeneed maleeniae avyoen ie an attempt tn anewor the question: are the high pressure bands of oxygen due to 04? Inevitably, this leads to consideration of the possible existence of a stable dimer, O,. For this dis- cussion a dimer will he defined at “two molecules which interact over a period of time long compared to the time between intermolecular collisions so that the vibretional and rotational degroes of freedom cannot be treated as those of two independent particles” (234); ive. if two O. molecules are associated for a time long compared with vibrational period characteristic of this union, A collision complex (03): will he considered 4 similar pair of molecules whose interaction time is‘THE SPECTRUM OF MOLECULAR OXYGEN ‘ot long compared with the time between collisions nore strictly, not long compared with vibrational frlod characteristic of the interacting pair)” In the ‘erature, there is no general agreement on the meaning F the terms dimer and complex; in addition, (0); and Date alien need intorchangesbly While some experimentel results are consistent with hhe stability of O,, a stronger case can be made for a short-lived collision complex (Os}e to aceount for bands observed in high pressure oxygen and the numerous simultaneous transitions which are induced (made allowed) by intermolecular forces. It is dificult to Aescribe as @ chemical entity « colliding pair with binding energy of = 600 em-! (= 0,008 eV) [231]. Cashion [78], using a binding energy of 82 em, has bredieted 7 bound vibrational levels for the interacting pairs (Ou), assuming’s Lennard-Jones potential. A dimer of 3 was frst mentioned by C, Harries and ly E, Baur in 1907. It was, however, # paper by G. N. Lewis in 1328, that started serious inquiry nto the stability and characteristics of an O, molecule. Therein, the formation of Os was postulated to account for the decreased paramagnetiam with inerensed concentration of oxygen in liquid oxygen-nitrogen mixtures, Since then ‘measurements of the spectra and physical properties of ompressed and condensed oxygen have heen inter ‘eted both as supporting or denying the existence of a ible Oy molecule. Knobler [234] has recently written “that there has been no irrefutable proof of its exis: nee." Moro rooontly, in another thesisy Keys [991] concluded “that the formation of O,, irespective ofits ‘exact electronie configuration, is a reality.” ‘Most of the literature on these topics ean be traced through the cited references, especially the theses by Knobler and by Keys. Much earlier, Heller [179] had esti- mated a vibrational frequency and binding energy for the complons and; in addition, ted le dit ferent kinds of experimental evidence indicating the existence of (0,) (a van der Waals complex, he called it) ‘The observed spectrs of compressed and condensed ‘oxygen are divided into four groups which will be di ‘cussed roughly in the following order. (1) 8.3 jum and 6.4 um ; induced absorption eorre- sponding to the U, ground state fundamental and first overtone, (@) 12600-6800 4 ; intensity proportional to pres: sure. Insludes bands of O, (a X, & X) srancl tions and bands attributed to a complex. (8) 6800-3000 4 ; intensities of the broad, difiuse bands are roughly proportional to the sauare of. the pressure. Simultaneous transitions in cou pled Os molecules; eg, (y+ 'y)— P35+ #35). (4) 2900-2400 A ; wsiplet (60 called high pressure) bands whose frequencies are tentatively corre- Isted with *4, 235" Os, sd ew ast Observed frequencies are slightly displaced from those of the free molecule Os has no permanent dipole moment, and eonsequently under normal conditions has no vibraton-rtation spec- ‘um, However, Crawford, Welsh, and Locke {96} have tthearend aheariton norveenanding tthe Ox wbratiane] fundamental frequency in high pressure (up to 60 atm) and liquid oxygen. Shapiro and Gush [852] have observed both the collision induced fundamental snd first overtone bands in oxygen at pressures of several atmospheres. “It is significant that the perturbation is svficient to alter the transition probabilities but has nealigible effect ‘on the vibrational frequency" [96} Smith and Johnston [899], however, found a shoulder at 1610 em” in the in frared spectrum of condensed oxygen; this frequency is consistent with a possible vibration in Os. The spectral feature at 1610 em~" has @ more plausible explanstion, says Calms (73, 74), who observed the infrared spectre ofa and 8 oxygen; itis due oan Oz vibration-translation ‘combination band lin the solid oxygen lattice). In liguid oxygen Ellis and Kneser(126] frst observed bands (6800-3000 A) which were atsibuted to simul neous transitions (in two enupled exygrn melee) Some of the bands, observed more recently in the gas phase under pressures of les than 10 atmospheres (path lengths > 350 m) (187, 18], were diffuse and showed no fine sircture even with dispersion of 2 A/mm. Several ofthese bands, also atuibuted to a complex, have been observed in the absorption spectrum of the atmosphere Alovng the soting own (188,209, 120) In a series of articles (1955 10 1968) Dianov-Klckov studied the intensity variation in the absorption spectra of compressed and condensed oxygen (12600-3000 4), The spectra included o-X and OX bands of Os, end bands attbuted 10 an (0%), complex. Cho, Alin, and Welsh (87) have studied effects af pressure om absorp sion Tuteneiden hn aX td OX ane Os, Bot Dianov-Klokov (108) and Cho et al. (87) concluded that the pressure induced effects dominate the mag dipole effets above 1.5 atmospheres. an with the earlier observations in the atmosphere (182, 389, 120], Robin (383) has observed some of the bends at pressures up to 5000 atm. Cho, Allin, and. Welsh [86] have observed simula neous transitions 2('89)-2(°35), lying close t0 Os, 34 -X Fp, in the spectrum of OeNe liquid mixtures. ss Allin, wi Wels [249] have observed violet degraded bands (of simultaneous transitions) in all three forms of solid oxygen. The spectra (12600~ 3800 A) were internreted a Cy tranctone ennerpoeed on a continnous distribution of lative frequencies. The authors deny formation of an (02): complex. “since simultaneous transitions are a general feature of in duced absorption, particularly in condensed phases.” ‘This general feature refers to induced vibration-rotation transitions in condensed hydrogen. Whilow and Findlay 4. Phys Chom, Rel: Dot, Vl. 1, Me. 2.1972aa PAUL H. KRUPENIE [412], studying emission of compressed oxygen, found the intensity variation for the simultaneous transitions tw be propoitional tv the aquare of the concentration of the excited Oe species. A similar quadratic pressure dependence in absorption has most often been in ternreted as evidence for a dimer. But not all results show this dependence [234, 107). Dianov-Klokov [109] has estimated from band widths an upper limit to the lifetime of the complex which is Smaller than the Uoosely detined) ume herween col lisions at atmospheric pressure, Arnold, Browne, and Osryalo {17} have deduced an unbound complex from the tempersines dependence af the intenity of the band at 6340 A. Dianoy-Klokoy [107] has emphasized that “all the absorption bends in the 12600-3000 A region in liquid and condensed oxygen are hasically related to inter ‘molecular interaction,” and that the spectra correspond to dipole transitions in (Os complexes. Whether these sre called induced dipole transitions oF recult from a collision complex, weak interactions are responsible. Rettschnick and Hoytink [829] explain simultaneous transitions of the complex as due to electron exchange between the two exygen molecules. during collision. rikhotko et al. [821] have drawn no conclusion about ‘8 dimer from their study of the spectra of solid oxygen. Joules al 229) infested that there mae ane Oe in the ‘y-phase of solid oxygen from X-ray diffraction of single erystals. They concluded that there was no Oy in other solid phases or in solution, because of the similar structure of ~O, and B-F; which was inconsistent with ‘4 dimer, Cairns and Pimentel (74) have interpreted their results on the infrared spectra of solid a-oxygen as showing np evidence for O4, stressing thee unver tainty in previous identifieations, Barrett, Meyer. and Wasserman [27] drew similar conclusions trom their study of the erystal structure of oxygen. Blickens- derfer and Ewing [47] have found some spectroscopic evidence for hound dimers in dilute oxygen at low temperatures. Further studies are in progress. Finkelnburg and Steiner [41] rst observed the mgh pressure bands of oxygen (2900-2300 A), a long. pro sression of diffuse triplet maxima in absorption by oxy gen at pressures of 60-600 atm. Some of these are tentatively ascribed to Os, 24,—X*3; (187, 188]- The hands, headless and red-degraded, did not appear at lower pressures. The convergence limit of these bands ‘wat roughly coincident with the dissociation limit of Os. This, together with the quadratic increase of absorp tion with pressure, suggested that the bands originated funn a vallision indused forbidden transition im. Ou 9AyX 1S, Finkelnburg, however, later [140] inter preted these bands as arising from vibration in Os Herman [1821 studied the variation of absorption coefficient with pressure in the visible and UV spectra of compressed oxygen up to 30 atm. The triplet bands were shown to be independent of the A-X Herzberg ‘bands of O; (3000-2400 A), The triplet bands increased 4. Phys Chom. to, Vb 2, No 2,972 ancity with increased pressure: the AX hands of Oz disappeared with increased pressure. These triplets were attributed to (Os) Knobler [234) recently remeasured magnetic sus cceptbilty of liquid oxygen at temperatures of 65-90 K, and concluded that neither the chemical approach of Lewis nor antiferromaznetic exchange completely explained the interaction between oxyuen molecules. His review of other physical properties revealed n comelusive evidence for stable O, molecule Physical adsorption, magnetic susceptibility meas: ‘urements, and the use of E-P.R. techniques by Mulay and Keys [290] and Keys [81] have provided evidence for the existence of O,. A discussion of possible bond: ing in the dimer has also been given [231] Blickensderfer and Ewing [348a,b] have observed the collision-nduced simultaneous transitions in gaseous oxygen at low pressure (I-3 atm) and temperatures of 300 K and 87 K. Mechanisms of the absorption in duced by binary collisions wore di evidence for a bound state dimer was found. Krishna [243a} and Krishna and Cassen [243b] have recently developed a general theory of relative intensities for the oxygen simultaneous transitions (bimolecular induced transitions). Robinson [385] had earlier dis assed enhancement of forbidden transitions by weak teens sn [903] have more recently calculated the intensity of one double transition and fevalues for several double transitions. Tabisz etal. [368] studied intensity profiles fof both single and double transitions in compressed ‘oxygen in the near IR and visible regions. The bands ‘were interpreted as collision (pressure}induced clec- tronic transitions, with no evidence for the existence of ‘quasi-stable complexes. In summary: Many spectral features of compressed and condensed oxyzen are not found in the low pres sure gas. The simultaneous transitions can be con sidered induced by weak intermolecular forces ot due to formation of shortlived (unstable) collision complexes. Ihe virtually unchanged ntrared spectrum (eavelengths >3y) is not consistent with a dimei but the shoulder at 1610 em could represent vi bration in O,. ‘The triplet bands (2000-2400 A) ave tentatively assigned to a transition in O;, a8 was done in the original work by Finkelnburg and Steiner. The quadratic pressure or concentration dependence of intensities for simultaneous transitions (12600-3000 A) observed by some only indicates dependence on initial reactants (two bodies) but does not indicate mlntles v1 mot the Gnel product fs a stable dimen The slight displacement of band frequencies from those of low pressure Os indicates neslisible change in electron configuration, which suggests a weak interaction between pairs of O: molecules. Various physical properties provide some evidence for a stable dimer. At low temperatures it is expected that @ dimer would become important when is feeble succed. No directn8, and dissociatior products for the ele:tronic states of Op 3nd 3° Es [A [ ze ot a % nergy cama .ss7] 2x0 = 1940 | » | 2 | 200,09 (672) > | 6 lau, i Lereesi6 02206747 s9a19 jbo v ess0 - 4990 | » | 7 1.408220 -0205830 4460 [Avex R 6530 - 1940 | » | 23 taeisoa2 10432¢_ | 515.41 je—»| 20 uuri23 68912, 948.7 @ | 20 = 1535 7 - uss | | 4 | 7,300 - wo | s 308, 369 6304724) - 1750 | » 62, 302,3 -.00097 479 | ~30 |is95.08 - 2430 | = 3799 2 ERS lari ex, = Bo] 6 1391 | -.000740 393.09 4790-4490, 278-2540] 19080 018169303 | 1.294192 (-5) 9970 - 5300 | » | 5 orn tse00 - 9240 ‘ 01893268 » | er 9 24, 62, 302, 9TABLE |, Molecular constants, electron configurations, end di a fA onan ke co Were | eere wee Da 98098 ,e ae Brarerenl tsa0] : 172000 4 isa70z 4 |.6s97 Lae og s%0y w77s| 630 | ser | tea) 18700 Lae rg, - |isess6 | aerse Lae '05**5u nooe| 675 |use.es | 73455 is743s_|isr76 238 m5 ty 2} 60 | esa | 13.568 ese | iso1e 238 Pgs Su ares | 505 |1035.838 | 0321s s7ses [ores [2 2 2 4 1 pty saesi| 17.10 [1908.13 | 62018 ) HE; 0.) 20+ 198128 98-510 8) | = 19BLES -R/An-0.955)", 403, . 1595-5108) Bg (OF) ——X'E, (04) 20 = 146568 30-6608) ing | seas ‘ay | see7es oat | 736s. 41 ool wing | 762624 | 76019 au iso | user pray | 75450 | 75283 ‘ eos | ass7 ety | a9380.5 | 4979.33 toa rrr | sia: | 2.37 |ros.05779) 1o.sigea0 | -ossei24as | o23974908 sexy? | 3500715 | 35398.70 33 253 | os9| 3.01 {799.08 [ies | -.s80 sasig | 34738 bos wou | rz | 26s | 7500 sze6ai | 33038.4 fo33 tse | 909] 297 | 794.29 | x 0.2440 | 00088 wizg? |8120,9088 | 13196.314 fat sey, |zais9 | zoesz | 967 |iase.ces! 043 1268431 7es2.39 | 798.1 24. eget, [3370 | 24150 | 10.73 | ti809.) iais69 ° o je 22 4 2 tegete, |azeo | s20s7] 1.77 |1580.1932 o4TaTa73g -1.2727481(-3)| 14107535¢ TEPRIN-ED WITH PERMISSION FROM JOURNAL OF PHYSICAL AND CHEWICALREFERENCE DATA, VOL.1,8O.2, 972‘THE SPECTRUM OF MOLECULAR OXYGEN, inding energy exceeds AT. But even in studies of lid oxygen, the presence of stable dimers has not ‘een unequivocally established. 5. Perturbations Perturbations mentioned in section 3 were incidental tw the discussion of the electronic-vibrational tran sitions. In this section, attention will be focused on several specific perturbations, even though the data may be fragmentary, 5.1, 01, AZ; state In the 11-0, 4°5{—X*3y band, Herzberg [187] letected a strong perturbation above N=11. No letails were given. 3:2, On, B°Z, st Ackerman and Biaume {3} have implied « perturba- on of v—6 from a discontinuity in difforoncos \Goun.—AGexie obtained from their study of the 3 53;—X 3p fine structure. Albritton et al. [9], ‘sing measurements of B-X bands bv several au: hors, have found a discor plots of A2G, Measured values of ay(5-0] and co (6-D) are about 0.7 em™ larger and smaller respec: tively, than values which give a smooth plot for AG. Unpublished results of R. A. Howard and S. G. Ti lord show very broed lines for the 5-0 and 6-0 B-X hands, Albritton et sl consider the possibility of vibrational perturbations of levels »=5 and 6, per hhaps originating in crossing by a repulsive poten: tial curve desived from ground state atoms. Present accuracy of By values does not permit an unambiguous verification or denial that there is a vibrational perturhation {0} For the B-X transition, Brix and Herzberg [62] mentioned two perturbations: (a) In the 160 band ‘4 conspicuous perturbation occurs in R(7) and P(9), ie, v=16, J=8. (b) Normally, Py2, @) broadening of BX absorption. lines with v' >2, @) very weak fuorescence, disappeating at low pressute, a5 observed I served by Rasetti [327], where seemed deppalaten a eas RreeR ne pe sociation, was Teasonable, and (4) uorescence in ox izadiated by 18194 ccury ta adie to cnee formation. Feast [133] observed BX emission from bands with v3 and discounted such predissociati SPRY Chum. Bef nts Wal 1 te 9 seanans Volman [391} presented strong photochemical evidence in favor of i Wilkinson and Mulliken ($141 obtained high resolu tion B-X spectra which showed rotational line broaden: ing in the 12-0 absorption band, but not in 13-0 and ‘shove. A predissociation was proposed, caused by *Tly having avery shallow, broad well. crossing the left limb of the B state potential curve at 6.9 eV. It eould not be established whether broad structure in bands with 81-212 arose from predissocintion or was dite te. resolved tiplets (because triplet splitting decreases at lower 0’). Under lower resolution Rakotoarijimy et al, [324] observed diffuseness in B-X absorption bands haying ‘v'=4, Herman et al. (183] also claimed strong perturba: tion of 0 =3, 4 but not 5. It was specifically to deter ine wlicules levels uf DE, 0 <0 were predissuciatea that Carrol (76) reexamined the absorption spectra uf the B-X system under high resolution, He found all Tines in the 4-1 hand diffuse indieating onset of « ore- dissociation below v'=4, N=0. A weaker predissocie tion was also observed for lines having N ~ 25 in’ He concluded that possibly »'=5 was also predissociated. Reexamination of plates taken by Dix and Herzberg 62] indicated another, but weaker, predissociation Peaking at v’=11, Carroll (76] sugzested that Tl, frossed the right limb of the B state potential curve near 94, but could not determine whether B. o=11 was crossed by the same state. Lines in B-X bands 2-0 and 1-0. broader than expected, were attributed by Carroll {70} to blends of fine structure components. Vanderslice et al. (387) considered this broadening more likely due to predis ‘of the R state hy 9 neatly seciation of the many level tangent *I], curve. Carroll further found a minimum in line broadening for B, suzzesting (wo pre: dissociations with perhaps *E> predissociating By =I High temperature absorption messurements by Hudson and Carter [199%, 200] indicaie that the B state ip preiissuciated ie levels wD tu 17, ymasibly also in v= 2. Myers and Bartle [289] also presume that should be predissociated. Line widths measured hhy Hudson and Carter in B-X’ bands 3-0 tw 12-0 varied from 2.5 emt to 0.9 em: for 13-0 to 17-0, Tine widths were ~0.5 cm-1, Line widths for 2-0 were (0.50.2) cm, so predissociation was not conclusive. hese results were interpreted to mean that crossing by M1, occurred between v=2 and 3, on the right limb of the B state. No comment could be made about 1-0 for experimental reacons. Hudeon and Carter [200] have pointed out the important role in atmospheric chemistry of predissociation of vibrational levels of the B state. (See also ref. [201] ‘The predissociation is thus characterized by: (1) B, v=4, lines are brosdest possibly at v=2: (2) broadening peaks again at 0 11, wit svininsuen at 9, (9) v= 22, sudden triplet apliting ‘onset of ppredissociation 4 Pry Chem el Bote, Vl 1, Ne 21972 PAUL H. KRUPENIE inerease. The question remains whether v= 12 shows triplet splitting and not spparent broadening Riess and Ben-Aryeh [831] and Murrell and Tylor [288] have applied the Franck-Condon principle to predissociation broadening of B-X lines. They have shown that a maximum in predissociation probability tecurs near curve eressing just below v4 (at r between 1.84 and 1.92 A) and that subsidiary maxima may occur above the crossing point which do not necessarily indicate a crossing of the left limb of the B state, Tur: neling is responsible for predissociation broadening Delow the crossing point. Franck-Condon factors for B9Y;—T1, were found to be sensitive to the form of the repulsive curve, No additiona) crossing of the left limb of the B state was indicated, Child [83] has derived an analytical expression for the probability of prediccscintion we fam vibrational quantum number. Analysis of # predis: sociation pattern can then lead to the form of the appropriate repulsive potential. The theory was applied to the BY; state of Os, using the results of Murrell and Taylor. Ackerman and Biaume (3] have since observed line Divadening fox vO 16 19, incidental wo tele primasy objective which was fine structure analysis. They were not sure that the apparent peak in broadening for the B-X, 11-0 band does not arise from separation of the triplets, which makes line widths dificult to measure. ‘They cautioned against the use of photographic line widths as cross section data if the spectra hive been taken mainly for fine structure analysis, but believed that their results were not in complete support of the conclusions reached by Murrell and Taylor (288) ‘The nonempirieal calculations of Schnefer and Miller [34a] indicate that the repulsive curve deduced bby Murrell and Taylor (288) and Riess and Ben-Aryeh [331] is not Tl, but could be Tl, of; and still ofr « panial explanation of the ebserved predissociation of the B state. Schaefer and Miller speculated that the Tl, curve crosses the left limb of 04 and ure *Thy curve crosses te B state on he B state near fer, perhaps between ve=O and 1. The nearly parallel eal. culated end RAR potential curves for the B state imply that improved wavefunctions would any lower the cealeulated eurve. ‘A general conclusion drawn by Schaefer and Miller jis that high4ying excited valence states are poorly deseribed by a single electron configuration. They found that for the B state near r- three configurations were important: Cop)*(iraMkry), Gala), ars), and Gos)(Iy)*(lmr))'Bou). An important implica tion of the assumed crossing of the Jeft limb of the B state by ®my is that the #Il,—X°%y transition would dominate the ©; continuum in the region 2000- 1750 A. A more reliale determination of predissociation Tine widths [J depenstence] ix necessary to. resolve ning questions: eancesning the predissociation ofaaa Volman [391] presented strong photochemical evidence sn favor of. ‘Wilkinson and Mulliken [414] obtained high resolu tion B-X spectra which showed rotational line broaden in the 19-0 shearption hand, hut nat in 12-0 and above. A predissociation was proposed, caused by Ty having a very shallow, broad well, crossing the left limb of the B state potential curve at 6.9 eV. It eould not be established whether broad structure in bands with 0 < 12 arose from predissociation or was due to un: resolved triplets (because triplet splitting decreases at lower 0"). Under lower resolution Rekotoarijimy et al. (324) observed diffuseness in B-X absorption bands having v'=4, Herman etal. {183] also claimed strong perturba- tion of v”=3, 4 but not 5. It was specifically to deter rine whether levels of B#%3, »< 8 were predissociated that Carroll [76] reexamined the absorption spectra of the H-A system under high resolution. He found all lines in the 4-1 band diffuse, indicating onset of a pre- jssociation below 4. A weaker predissocia: tion was alo observed fr lines having N~ 25 inv =3. He concluded that possibly’ =5 was aso predissociated, Reexamination of plates taken by Brix and Herzberg, [62] indicated another, but weaker, predissoci Peaking at v1, Carroll [10] suggested that crossed the right limb of the B state potential curve near v=4, but could not determine whether B. 0= 11 was orocced by the ame etate Lines in B-X bands 2-0 and 1-0, broader than expected, were attributed by Carroll [76] to blends of fine structure components. Vanderslice et al. [387] considered this broadening more likely due to pred sociation of the many levels of the B state by a nearly tangent *[ly curve. Carroll further found a minimum dissociations with pethape °Ey predissociating By v=. High temperature absorption measurements by Hudson and Carter (199a, 200} indicate that the B state is predissociated in levels v=3 to 17, possibly also in p= 2. Myers and Bartle [289] also presume that 2 should be predissoclated. Line widths mewsured by Hudson and Carter in B-X bands 3-0 to 12-0 varied from 2.5 em-! to 0.9 emt; for 13-0 to 17-0, Tine widthe wore O05 cm-t Lina widthe far 910 wo (0.50.2) em-t, so predissociation was not conclusive. ‘These results were imerpreted to mean that crossing by Tl, occurred between 9=2 and 3, on the right limb of the B state, No comment could be made about 1-0. for experimental reasons. Hudson and Carter [200] have pointed out the important role in atmospheric chemistry of predissociation of vibrational levels of the B state. (See also ref 201],) ‘The predissociation is thus characterized by: (1) B. v=4, lines are broadest: onset of predissociation possibly at 12: @) broadening peaks again at v=8, 11, with minimum at 9; @) o=12, sudden triplet splitting ine broadening, for 2 4 Phy, Cham Re Det, Vol, No.2, 1972 PAUL H. KRUFENIE increase. The question remains whether »=12 shows {apletepliting and not apparent broadening Riess and Ben-Aryeh (881) and Murrell and Taylor {288} have applied the FranckCondon principle to tredissociation broadening of B-X lines. They have Shown that maximum in predissociation probability occurs near curve crossing just below v= 4 at r between 1.84 and 1.92 A) and that subsidiary maxima may oceur shove the crossing point which “dy nol necessarily indicate a crossing of the left limb of the B state. Tun ning is responsible for predissociation broadening helow the crossing point. Franck-Condon factors for B°S;—l, were found to be sensitive to the form of the repulsive curve. No additional crossing of the let limb of the B state was indicated. Chill [8] has derived an analytical expression for the probability of predissocition as a function of vibrational quantum number. Analysis of a predis: sociation patton can then lead to the form af the appropriate repulsive potential The theory was applied to the B23; state of Os, using the results of Murrell and Taylor. ‘Ackerman and Biaume [3] have since observed line broadening for v=0 to 13, incidental to their primary objective which was fine structure analysis. They were hot sure that te apparent peak in Drowdentng for tie B-X, 11-0 band does not ari from separation of the Uuiplets, which makes ine widths dificult to measure. ‘They cautioned against the ute of nhotoeraphie Tine Widthe as cross section data ifthe spectta have been taken mainly for fine structure analysis, but believed that theie results were not in complete support of the conclusions reached by Murrell and Taylor (285) The non-empirical calculations of Schaefer and Mile [344] indicate thatthe repulsive curve deduced bby Murrell and Taylor [208} and Rices and Bem Aryoh {331] is not iy, but could be *ly oF 8; and stil offer a partial explanation of the observed predissociation of the B state. Schaefer and Miller speculated that the I curve crosses the left imb of the B state near and the *Tly curve erosses the B state on the right, perhaps between »=0 and 1. The nearly parallel ea. ulated and RAR potential eurves for the B state Imply that improved wavelunctions would only lower the caleulated curve {A guneral sonslasion drawn by Schaafer and Miller is that highlying excited valence states are poorly deseribed by a single electron configuration. They found that for the B state near re three configura important: @o,}UraFn,¥, -Go,\lry ln, Ber), and. (o,)(Xm4)*(1m,)*(30.). An important implica tion of the assumed crossing of the left limb of the B atste by 2y iy that the ly—X°Ey transition would dominate’ the ‘Oz continuum in the region 2000 1750 A. A more reliable determination of predissociation line widths [J dependencel is necessary to, resolve remaining questions concerning the predissociation of BPG.‘THE SPECTRUM OF MOLECULAR OXYGEN 7. Dissoe ion Energy of O, The dissociation energy D* forthe ground state of Os ‘= 4126015) em- This value is determined from 11) the convergence limit of the B *E; state [62] at 7127.5 em?, 2) the separation between dissociation limits for the B and A *%j states as representing the enerey OCDs)-OGP2), and @) the assumption from the non-erossing rule that the atomic product common to ol B and A states is OCPs), The uncertainty takes ini0 scenunt the inability to rule out possible potential maxima of about 2 em. Details are given by Brix and Hersherg [62]; a snocinet summary ix given hy Gaydon N54) 8. Microwave Spectrum of O, A summary is given below of the fine structure spec trum of "Os, "0s, 900, and the hyperfine structure of 00. The observed line frequencies are collected in tables 31-33; line widths are given in table 34. Mis- cellaneous data derived from the microwave spectrum ore given in table 85. The magnetic resonance (Zeeman) spectrum is discussed. ror details ot the theoretical temperature: and pressure-dependence line widths, and the non-resonant (Debye) spectrum the reader is re- lvrred to papers by Artman and Cordon [19], Tinleham [374, 375], and papers since 1959 by M. Mizushima, and A. A. Maryott. Additional references are given in a com pilation by Wacker et al. [392] 8.1. Rotational Spectrum Through O, is electrically non-polar, it has a per- ‘manent magnetic moment associated with the aligned spins of two unpaired electrons. Consequently, mag: nelie dipole (resonance) transitions (AN: 21) ate allowed between te pins snukiplet comp of the ground state of Os, ie., components with the same value of N, but different values of J. The com: ponents! F. and Fy lie close tosether and below Fe by about 2 em-! G-mm wavelength transition). Beringer, in 1946, first observed this very weak ab: sorption without resolving Fs~F, and Fy-Fs structure, 4nd in the following year Van Vieck accounted, theo: retically, for the intensity distribution of this line whose existence he had proviously predicted, This line was rrerially rovclved by Lamont (im 2042) and by Strand. hers, Meng, and Ingersoll. Burkhalter et al. [69] first resolved the fine structure in the 5mm wavelength ‘egion (60 GH2) and found deviations from the theo ‘tical formulas of Schlapp [346], which were exten sions of early theory by Kramers and Hebb. (See Brix and Herzberg [62}) The position of the 1~ line near 25mm wavelength was first measured by Anderson, Johnson, and Gordy (14), SR PRS th kw pe or 445 Miller and Townes [275] and Mizushima and Hill [279] revised Schlapp’s formulas. Zimmerer and Mizushima [422] and West and Mizushima [311] ex- tended these formulas, and remeasured many fre- quencies with high accuracy. (Table 31 also includes more recent. measurements by MeKnight and Gordy [265] and Wilheit and Barrett [413],) Tinkham, in his thesis [374], and in three papers with Strandberg [375-377] has developed the most extensive theoretical treatment of the spin multiplet structure and intensities of the Os microwave spec: trum, These papers account for the fine structure theory of the Oz ground state [875], the magnetic field interactions Zeeman effect) (376), and the line widths [877], Tinkham’s fine structure parameters have been recently revised by Tischer [3781. (Compare also West and Mizushima [411] and Appendix C.) Magnetic dipole rotational transitions (AN=1), predicted by Tinkham [874], have been observed lunder high resolution by MeKnight and Gordy (205) = (123, 14.2, 16 em~). The latter two have been iden- tified in the solar spectrum by Gebbie et al. (157) Additional magnetic dipole entational trancitions ah. served in the laboratory and the solar spectrum are included in table 33. Miller, Javan, and Townes [274] measured the transitions in '90!0 and ™0,. Since none of these lines showed splitting due to magnetic hyperfine structure, it was assumed that "0 had zero nuclear pin (273, 274, 370], a result that has since become well established, “The hyperfine structure of "00 arises from mag. netic dipole and clectrie quadrupole interactions of the "0 nuclear spin with the unpaired electron spins. Miller and Townes (275) have confirmed the assignment of nuclear spin I=$ for "0 (10, 159] by obtaining ‘sprevitent between observed frequencies fur "O"O. ‘and the theoretical hyperfine spectrum calculated with the assumption that J('"0)=3 The theory of the hyperfine structure ix given hy Froech and Foley (1461 For the two parameters representing the magnetic interaction energy, the data of Miller and Townes [275] 140 MHz. (Miller and Townes also measured the fine structure of ™Us, "Uz, and 080.) ‘Zeeman Effect and EPR Spectrum ‘The Zeeman effect of Os was first studied in the 6-X system by Schmid, Buds, and Zemplen. This low resolu tion work was done with high magnetic fields which broke down internal coupling. An extended Zeeman theory® of the rotational levels of the "O. ground state developed by Henry was further improved by ‘Suranberg(976}. In the microwave region, Beringer and Castle first resolved lines of the magnetic 4. Phy Chem Ra. Dato, Vol, Ne 2, 1972a6 PAUL H KeNIDENIE dipole transitions between Zeeman levels of the ground vibrational state, using fields up to 9000 gauss. Some quantum numbers were assigned on the basis of Henry's theory. These observations were considerably extended by Tinkham and Strandberg [376] who observed over 200 lines and identified 40. For more than half the observed lines AJ=22. J breaks down as a good quantum number in magnetic fields of ~ 10,000 gauss ‘and all AJ transitions then become allowed. Hendrie and Kusch [180] also studied these Zeeman jons and obtained total g-values for the rotational states. Bowers, Kamper, and Lustig (51] repeated Tinkham and Strandberg’s measurement on 19 lines, Lines were measured with an accuracy of +5 ppm. Positions of absorption lines were fitted to values based on the Tinkham and Strandberg thoory to 2 ppm. (They also measured ""0"0.) Hill and Gordy (193), using magnetic fields of ~ 100 ‘gauss, introduced no decoupling of internal angular momenta, Zeeman splittings were observed for lines 1,3, and 3. ‘The electronic paramagnetic resonance (EPR) spectrum of the metastable aay state nas recently been studied by Falick et al. [129] who observed the four AM)=1 transitions * for the J=2 level. Miller [277] using areater precision, observed these in adic tion to four AM)=1 transitions for the J=8 level. Measured magnetic interaction parameters are included in table 35. Miller's value of By (1.41808 00020 em-1) Js the same as that obtained from the a'dy—X'S5 electronie spectra [191] within experimental uncertainty Miller's value includes the effect of electronic mass.) Tiecher [378] haa observed 27% magnetic dipole trans: tions between Zeeman components of the fine structure terms for the X*37_ state. Paramagnetie resonance absorption between N=3, J=4, M=—4 and N=5, J=5, M==4 levels were observed by Evenson et al {128} using 890 GH laser excitation (table 33), Observed zero-field energy difference between N=3, J=3 and N-5, levels is fitted slightly better by the theoretical parameters of West and Mizushima [411] than by those of Tischer [378]. Work in progress (Even son. private communication) on 12 ndalitional lines should lead to a consistent set of theoretical param: eters which will also reproduce the rotational line frequencies. (See Appendix C.) Several lines, observed high an the earth's atmosphere, have been assigned with some uncertainty to Ox27.8-51 em) (156) 9. Raman and Zeeman Effects 9.1. Roman Effect ‘The vibrauonal Raman emrect has een observed in Doth liquid and gaseous oxygen. Rotational Raman spec- tra have been observed only in the gas, Among the 42. Pry: Chem Re Data, Val, No.2, 1972 first examples of pure vibrational transitions in « homo nuclear diatomic molecule, were the observations of McLennan and MeLeod (in 1928) [266] in liquid oxygen, which was irradiated with two different mereury lines. Both 1-0 and 20 vibrational transitions were detected (the latter smaller than the gas phase quantum by 35 em). Both transitions are electric dipole forbidden.) Shortly thereafter, the fundamental frequency wes detected by Rasetti and others inthe sas at atmospheric pressure and above. Rasetti (326] also first observed the rotational Raman spectrum. Recently, Weber and ‘McGinnis [406] observed the vibrational and rotational Raman spectra under 5 A/mm dispersion and obtained values for AG(4) and By in close agreement with those derived from electronic spectra. Rotational spectro: ‘exams have heen nblished hy eeversl authors (405, 26a, ¢, 381, 406, 366,328} Bhagavantam [41], irradiating oxygen with visible light, had found that rotational structure disappeared above 30 atm pressure (no wavelengths were given). ‘Trumpy [381], using 4086.8 A He light, studied the pressure eflect on rotational Raman spectra at 15, 20; end 60 stm. Surprisingly, even at GD ati the lines were only broadened and not completely washed out Saha (339), using about the same dispersion as Trumpy (~12Almm), was unable to see discrete rotational structure in liquid oxygen, indicating far more broad ‘ening in the liquid than in the gas at 60 atm. Renschler et al [328] have measured triplet struc: sve in the rotational Ranen spectrum of Or, Positions of lines observed at 21.29 em, 14.30 em~, and 16.25, fem-? are in good agreement with magnetic rotation spectra measurements (compare tale 33) Depolarization ratio for Raman scattering was determined by Cabannes and Rousset [72] o be 0.261 = 0.010; Bhagavantam [41] had obtained p < 0.3. Recently ‘Yoshino and Bernstein (418) obtained 0.3 In a study of mercury light scattered by Os at at mospheric pressure, Rasetti [327] observed, in addi: tion to the Raman lines, » series of doublet linoe ‘spanning the region 3980-2365 A. This fluorescence spectrum, excited by the 1849 A line of Hz, corres: ponds to elements of two branches of B°Es—X 925, 8 through 22). The doublet separa corresponds to FY(13)=FI(11). The fluorescence was not observed at mm pressure, comay w eapectatons, “Ruse considered. these doublets to represent a. possible transition between the Raman effect and fluorescence which he called “selective Raman effect” (now called Resonance Raman effect). That is, where entire branches might be expected, only Raman lines were observed whose transition probability was very high. (See Rasetti's table z.) Depolarization ratio for Rayleigh seattering was determined as 0.014 by Weber et al. [406}, considerably fan previous valven (0.02-0.06) [214, 316, 60), presumably because of the actual isolation of the lower‘THE SPECTRUM OF MOLECULAR OXYGEN at Nayleizh line from the Reman lines. The new vatue represents an upper limit. The mean rate of change vo polarizability with inter-nuclear distance was de- termined by Stansbury et al. [364] as 1.4, from the vtio of Raman to Rayleigh intensities in the gas. %. Spectra Joeman spectrum of the b 2) -X *Yy transition vs heen produced by application of magneuc fields 110 26,000 g [347]. The absorption spectrum around (600 A includes extra lines which arise from splitting WF the ground state rotetional levels. Zeoman epliting was about 1 em! for # magnetic field of 21,400 g. Lines were noticeably broadened at pressures above 10 atm. The mst intense Zeeman lines lav to the red of the P vind R branch lines. Zeeman lines to the violet of the "Q and "Q lines were somewhat weaker, and displaced somewhat further from the fieldree lines. Maximum line. sharpness and minimum displacement occurred lor NT Hosiky and Schmid (55] produced measureable Zoeman splitting of linge belonging 10 the b 135 —> «MM, Of 1-0 band at 5631 A by use of magnetic fields sf 15,000 and 23,000 g. No detailed report has been Imiblished. (Zeeman effects in the microwave spectrum ol the ground state are discussed in section 8.2.) Zeeman Effeet in Electro 10. Potential Energy Curves Potential energy curves for electronic states of 0;, Os, and Of are shown in figure 2. Some of these ‘curves have been derived from spectroscopic data by the use of the RydbergKlein-Rees (RKR) procedure; for these, numerical data taken mainly from Albritton tal. (9) are given in table 58. For several ther states Sree data ty sparse the dyed curves ls fgwe 2 are cnly suggestive. For the ©; and CA, states of Os only tentative vibrational quantum numbers can be tnsiened. For Oz the arorosimate curves are desived from tentative ab initio quantum mechanical caleula: tions. (Bee see. 2.4) The curves for 12 repulsive states of O+ which dis- ‘toms had been estimated tal [387] and slightly adjusted by Gilmore [161] o account for the effects of w- Recent wonllgaration inter action calculations by Schaefer and Harris [345] indicate ‘an entirely different order for these states compared with the semiemoirical estimates. Schaefer and Harris, found that five unobserved states of Os are likely stable: two of these had been assumed stable in prior work and were included in Gilmore's curves [161]. (See nee. 220.) ‘The curves in Hig. 2 ditler from those ot Gilmore 1161] a8 follows: (1) the © "E> state now lies below Cra becouse: of more scent epectroscopte date, @) the shape of the X*Mly state of O3 is slightly altered wovause » has been increased by unity, 8) numer approximate curves have been deleted, and (4) appx ‘mate potential curves derived from CI caleulations are indicated for some stable states of O. ‘The various RKR results by different authors provide turning polis tae dif, ha geneva, by wo awore tht £0,003 A. The single exception has been the BeZy state which, in all published results, shows erratic inner fuming points ahowwe vee 15. RKR potential curves for O; were first calculated by Vandersice et al. [387] The irregularity of the left limb of the B*2z potential curve has since been examined by others including Gilmore (261), Richards and Barrow [330], Ginter and Battino (162), and Harris et al. 172}. The seatter re- ‘mained in the third decimal place in ray. The possible ‘causes of the seater mvlude (1) oselladons tn wuss constants, (2) sensitivity of the RKR method to inac- ‘curacies, especially in By values, near the dissociation Timit, and inertia “Albritton etal. [9] have recalculated the potential ‘curve for the B 32; state using the recent accurate data ‘of Ackerman and Biaume [3] up to v= 13; the use of the data of Brix and Herzberg [62] above v= 13. yielded ieregular rai Numerical experiments by Albritton 1 al. showed that monotonically varying rau could be biained by the use of adjusted Be values Ing wishin the experimental uncertaimies of the values quoted by Brix and Herzberg. To eliminate irregularities in ‘aig. it was concluded, By values accurate to 0.0005 em-* ‘would be necessary for v> 13. (There appears, however, to be no a priori reason why the left limb of « potential must always be monotonic (M. Krauss, private communication) TKR potentials have been calculated by Singh and Rai [355] for the X, a, A, and 8 states of Os. A change in viheational vannim numbering [47] vids their results for the O; ground state, but Asundi and Rama- ‘chandrarao [21] have recalculated the curve for the Of ground state with the corrected quantum numbering. Singh and Rai have commented that the A state curve appears to be approaching its convergence limit very rapidly (which suggests a possible potential maximum for this state). Albrition et al. [9] do not mention this possibilty Potential enorgy curves for Of* are briefly men- toned in a0. 2.35 det [209] and (208), ‘an he found in references 11, Trancition Probability Paramotore Tt is not uncommon to find extensive and reliable measurements of the positions of spectral lines origi nating in a given state of a molecule, the analysis of the spectra being in terms of a notation which is largely agreed upon. By contrast, in the area of spectroscopy loosely called quantitative epeotroccopy, the experimental varameters which measure transition 4 Pha: Cham Ref. Dat. Va. 1. Ne. 2.1972440 probabilities ere not nearly so extensively studied nor so well established. id the notation used in their description is sometimes ambiguous and not standard- ied. The notation by & given author may vary from paper to paper. In this section not all topics will be considered ‘quantitatively. Measured radiative lifetimes and ‘fralues for many transitions in Oz and Oy are sum- rmorived (table GH). Fralues for the Os, B°3;~X*35 transition are sum- marized in table 55. Absorption coefficients for selected features in the UV are summarized in table 56: most ‘ther absorption coefficients and integrated band in tensities are considered only qualitatively. Calculated Franck-Condon parameters (table 57) are considered, ‘with the reservations that results based on the Morse potential may, in some instances, be poor approxima- tions, and that the use of the r centroid approximation for the eatensive measurements of ing alectronie twamition mamonte. hae ions (214, 244], Nearly all of the data in table 57 are from the extensive calculations of Albritton etal. (9) For each eleetrome transition the radiatve btetime of the upper state is discussed together with the tran sition probability parameters corresponding to the tran sition. Ta general, the form af transition moments should be regarded as tentative. In a recent review by Nicholls [298], the general concepts used in application of ‘the Franck-Condon principle are briefly discussed, and a summary is given of recent measurements and caleulations of aero: ‘nomically important transition probability parameters. Dificutiee with the definition and meaning of electronic Fralues (fo) and electronic Einstein 4 coefficients for discrete spectra have been discussed by many authors, most recently by Tatum [372] and Schade [843], in addition to these conceptual dificult, different conventions in the use of degeneracies lend ‘confusion when several sets of data are compared. Wentink et ab. [410] Inve compared the two convent tions as applied to transition moments for AS=0 transi tions only where electronic transition moment Re (Nicholls)= Va'GR- (Mulliken) where 2 is the elec: tonic degeneracy of state v', ie. (2—94,9) 28+1) and 6" is the number of orbitals of state u*. Values of Jor oF Te are independent of convention since these parameters involve ratios e'g"=G'/l'; but absolute values of transition moment, transition moment matrix clements, and Ay depend on the convention used Batec [38] accumed come youre ago thet Re(e) did not vary rapidly with internuclear distance for electronic transitions with AA=0 (labelled parallel transitions). 41 Phy. Cham Re: Dot, Vol. 1, Ma 2, 1972 PAUL H. KRUPENIE “The position regarding perpendicular type transition tie., AA == 1) i more uncertain" [33]. In the Iterator ithas often been assumed that transitions vith AA === 1 will have Re(r) rapidly varying. These generaliztions are not rigorously based. The behavior of Re(r) has ‘lten heen ‘deduced trom crude band intensities. and the use of Morse based Franck-Condon factors. A brief comparison of the use of Morse and RKR potenti: Follows ‘The Morse potential is a three parameter analytic function (1861 whieh may represent a useful approxi ration, mainly for low vibrational levels of « given electronic state. For use with fragmentary data i Provides a qualitative description of electronic state and transitions. When used. to. describe electron! transitions where [r-—r®0.1A the Franck-Condc factors are often good approximations. For weak banc (those with (relatively) small Franck-Condon factors Vwaitnne heer [elma O1 then Morse based Franck-Condon factors. may be Faw lec unreliable. For an extreme case where |r=r2| ~ 0:4 A (O, B°S;—X *35) comparison between Morse and ( [KH results is striking [292, 310, 172, 9. ‘The RKR method of calculating potential curves {rom spectroscopic data is based on numerical integrs Yion of two phase integrals. Prom these integr ‘obtained the turning points fora given vibrational term value? The RKR potentials do not predetermine rela tions among molecular constants as do semiempirical Potentials. However, ifthe vibrational term values and mal constants are not knoven up to the dissocia: then the method yields only a portion of the curve hounded by the known tere values, ti 111. @'4,—X°3; Infrared Atmosphere System Badger et al. [26] have measured absolute intensities of diserete absorption of the a-X, 0-0 (1.26 pm) and 1-0 (1.065 jum) bands at a pressure of four atmos heres. The underlying continuous absorption, yr portional to pressure, is attributed to collision complex (On). A radiative lifetime for the 0-0 transition of the sd molecule was estimated as 45 minutes (see table in pure oxygen at one atmosphere, it is only 9.2 minutes, Additional discussion of effects of pressure, complexes, and comparison with other measurements is, five by Badges ot aly (see eopetially table 1 uf sei (206). Relative intensity ratio 1(0-0)/1-0) is estimated fs 3900: ratio af Franck-Candan factors ie. shoot 76 [296, 9], Integrated absorption eoeficients of continua n the region of these bands exceeds those for the discrete bands above 4 atm pressure. Badger et al cite references to other work on a much weaker con: tinuum underlying the 2-0 band, and a far weaker one inderlying the 3-0 band,probabilities are not nearly so extensively studied nor so well established, and the notation used in their description is sometimes ambiguous and not standard ied. ‘The notation by & given author may vary from Tn this section not all topics will be considered quantitatively. Measured radiative lifetimes and _fovalues for many transitions in Os and Oj are sum: marized (lable 54}; the extensive measurements of fFrvalues for the Oz, B*2s—X35 transition are sum. ‘marized in table 55. Absorption coefficients for selected Features in the UV ave susarived in table 56; must ther absorption coefficients and integrated band in tensities are considered only qualitatively. Calculated Franck-Condon parameters (table 5712 ate considered. with the reservations that results based on the Morse potential may, in some instances, be poor approxima tions, and that the use of the r centroid approximation tor determing electronic transition moments as some limitations (214, 243), Nearly all of the data in table 57 are from the extensive ealeulations of Albritton ov al (0) For each electronic transition the radiative lifetime of the upper state is discussed together with the tren sition probability parameters corresponding to the tra sition. In general, the form of transition moments should be regarded as tentative. In a recent review by Nicholls [298], the general concepts used in application of the Franek-Condon principle are briefly diseussed, and summary is given of recent measurements and calculations of aero: nomically important transition probability parameters. Difficulties with the definition and meaning of electronic ‘Fralues (fu) and electronic Einstein 4 coef for discrete spectra have been discussed by many ‘euthore, most recently by Tatum [879] and Schedce [343] In addition to these conceptual difficulties, different conventions in the use of degeneracies lend confusion when several sels of data are compared. ‘Wentink et al. [410] have compared the two conven: tions as applied to transition moments for 4S=0 transi tions only where electronic transition moment Re (Nichols)~ Va'GR; (Mulliken) where gis the elec tronic degeneracy of state o', ie. (2—64,0) (28-41) and 6" is the number of orbitals of state, Values of fos ar nu ave indopendent of camvontion sinee parameters involve ratios g’/g"=G'IG'; but absolute values of transition moment, transition moment matrix elements, and Avie depend on the convention used. Bates [38] assumed some years ago that Ke(r) did not vary rapidly with internuclear distance for electronie transitions with AA=0 (labelled parallel transition). PAUL H. KRUPENIE “The postion regarding perpendiculartype transitions (he., AA==1) fe more Uncertain” 35}. In the Iterature ‘thas often been essumed that transitions with A= == 1 wall have Ry(7) rapidly varying. These generalizations ire nat righranely based. ‘The behavior of RaCe) has ‘often been deduced from crude band intensities and the use of Morse based Franck-Condon factors. A brief comparison of the use of Morse and RKR potentials follows ‘The Morse potential is a three parameter analytic funetion (186) which may represent a useful approx: ration, mainly for low vibrational levels of a giver ectronic state. For use with fragmentary data provides a qualitative description of elecuonic states ‘and trensitions. When used to describe electronic transitions where |e!—rf|=0.LA the Franck-Condor factors are often good approximations. For weak bande {those with (celatively) small Franek-Condon factors tr tor lectrome transitions where |—ve] 2018 then Morse based Franck-Condon factors, may be unreliable. For an extreme ease where |ri=re|~ 04 Aio.. R352—¥ 935) omnariean hetwoon Morse and TRKR results is striking 292,310, 172, 9} The RKR method of calculating, potential eurves from spectroscopic data is based on numerical integra tion of two phase integrals. From these integrals re obtained the turning points for @ given vibrational term value The RKR potentials do not predetermine rela: tions among molecular oonttante se do semiempirical potentials. However, if the vibrational term values and rotational constante are not knows up to the dissoci tion limit then the method yields only « portion of the potential curve bounded by the known term valies. V1.1. @'4,—X 2%; Infrared Atmospheric System Badger et al. [26] have measured absolute intensities of diserete absorption of the aX, 0-0 (1.26 um) and 1-0 (1.065 jam) bands ata pressure of four atmos heres. The underlying continuous absorption, pro: portional to pressure, is attributed to collision complex (Oss. A radiative lifetime for the 0-0 transition of the inolated malooule wae catimated a0 46 minutes ace table 5); in pure oxygen at one atmosphere, iis only 922 minutes. Additional discussion of effects of pressure, complexes, and comparison wit other measurements is tiven by Badger et aly (see especially table of ref 20). Relative intensity ratio 10-0)/10-0) is estimated as 200% rau of Franck-Comlon factors is about 70 (296, 9} Integrated absorption coefcients of continue in the region of these bands exceeds those for the discrete bands shove + atm pressure. Badger etl. cite references to other work on a much weaker con- ‘inoum underlying the 2-0 band, and a far weaker one tnderying the $-O band,THE SPECTRUM OF MOLECULAR OXYGEN Vallance Jones and Harrison (385} account for absence 1 the 0-0 band in twilight emission as arising from ‘eabsorption by the lower atmosphere. "The emission ‘ate of the 0-0 band should be about 10 times that of the 01 band.” Morse based Franck-Condon factors have been oal valated for a small Deslandres array [168b, 296). RKR ‘alues of Albritton etal. (9] are given in table 57 ‘The nine branches ofthis system all have comparable imensties [191], confirming the theoretical formulas ¥ Van Vleck [388]. The intensity of the a-X system in The gas phase is less than 2 that of the b-¥ system; n liquid oxygen the aX transition is five times stronger than the b-X transition (126). Both transitions have an sbsolute intensity in the liquid which is greater than in ‘he gas, The 1-0 band in the liquid is nearly as stronz w= the 0-O; in the gas itis much weaker and corres: wns to Franck-Condon principle expectations. 11.2. BIS; 20; Atmosphere System The integrated absorption coeficient for the 6—X. 1-0 hand has recently been studied very extensively. Vorch ot ah {67} vbtained for the jotegrated mass sbsoxption coefficient a value of (090.25) g7) on? emt, smaller by 20 percent than the value of Wark and. Mercer [398]. Burch et al. considered the Hleets of line broadening half-width of sele-broadened vaygen lines decreased with J at one atm pressure) ine shape, and the dependence on temperature, Inessure, wink paths lengals Line positions word nese fom Wark and Mercer [398] who also tabulated ines for the isotopic molecules “OO and "OO. The absorption coefficent of Wark and Mercer was Inised on solar spectra, and an assumed constant line ‘width of 0.088 cm”. The band strength value based on. vatrapolation to low pressure from measurements x Gho etal. (81), prmanily on induced ansorpuion saves shout the same value as that of Burch et al. (who labeled their absorption coefficient as band strength). ‘The A value for the O-0 hand i 085 em! (67H) liable 58. The observed line strengths [67] were fied about as well by rotational weighting functions hhased on experiment [85] as by theoretical weighting Junctions of Schalpp [345, 340] based on strietly ease 6 oupling. Recently, Watson [403] derived weighting factors considering deviations from Hund’s case), Int for O, thooe aro virtually idontioal with Sohlapp’e results, Adiks and Dianov-Klokov [3] have determined the fovalue for the 0-0 band (corresponding to 11082 see“) from a low pressure limit of the integrated sdsorption coefficient. No significant J-dependence was detected. Relative distribution of intensities within ‘he band followed Schlapp's formulas [346]. Miller sak, [272] lnave als easudved in devel de live whe nud intensities of the 4X, 0-O band and obtained a slightly smaller value for the Einstein A coefficient ‘an that of Burch and Grvvnak (671. They concluded shat the theoretical HiinkLondon factors of Schlepp [346] and Watson [408] fitted their date mote closely than did the empirical ones of Childs and Mecke (85) For bands other then 0-0 the measured parameters are crude. If an estimate is made of the contribution to Zedve by the 0-1 band (ve.. (0.7) (085)=.006), the radiative lifetime () for level US, 2-—O i estimate as 11 seconds (table 54), which is slightly larger than earlier values [85, 398} Wallace and Hunten [207] have mosoured the airglow photon intensity ratio 10-0)/ (0-1) as 17:2; Noxon. [301] measured 20:+4 in the laboratory. Morse-based Franck-Condon factors (168), 296] give 23; RKR results Daive 14. 11.3, bY; ‘Tho absolute transition probability of the ba, 0-0 transition was estimated by Noxon [301] as 1.7% 10-8 5-2" The 0-0 emission band was observed superimposed fon a strong continuum, Observed band width was 20 A, In photon units of intensity, the ratio b-a, 0-0 t0 b-X, (0-0 was about 0.004. Noxon assumed that the Q branch represented 20 perrent af tha hand intansity Sax Noxon's paper for mention of earlier estimates of the probability of this electric quadrupole 2A, Noxon System Morse Franck-Condon factors have heen calculated by Nicholls [296], RKR results (table $7) are from Albritton etal. (9 11.4, ASE z— X85 Herzberg | Transition The 423;—X935 transition has a widely spread Condon locus which is characteristic of transition dominated by @ continuum — Segro™0.0138) (104) Jarmain and Nicholls 219} have calculated RKR Franck Gondon factors which are slightly different then those cited by Degen et al (104), These differ from Morse based values (168, 296] by as much as a factor of 3 Franck-Condon densities 219), calculated for the photo- dissociation continua of the "0,1, and 2 progressions (2500-1500 A), when used with the measured absorp- tion cross sections of Ditchburn and Young (114), have led to a transition moment Me which decreases slowly with r. This behavior was consistent with that found by Degen and Nichols (105) who deduced thee Re (for the range 1.33< fre< 1.531 from emission Ihand intensities and Morse hased Franck-Condon factors. (4880-2590 A). (No details were published) Jarman and Nicholls [219] tentatively estimated the continuum f value as 2%10-t, 2 orders of magnitude (2 Tate ey erm ChB Neh nd Bann "26 and450 smaller than previously thought [114] (see discussion in ref. (219]). The possibility of over-lapping continua near the 4X ban convergence iit (2300 A) may Wood to an overeninate of valves in the newb con Ditchburn and Young's [114] measured absorption croneveeins in the 4X comin saociated ih the 9": 12500-1850 A) did not obev Beers aw. The adhont! absent increasing pressure (2500-2000 A) was aseribed to formation of 0. Cro sectons (xtapolted to 10, pressure) howd pam tocar un By wo 9m en (2300 A} The 4-X continuum peak mas caused to be at 1870 As but col not be checked experimentally Recauanat Borst Degen [10] has mensoredintenses of A-X bands emitel inthe afergiow of «microwave dvcherge in tn OrAr mistre Os) in order to provide relaive Band sense fo the spnem, Both botopraphie and Plnsactte taturencnis more made, which were ra sandard Tungsten sp Blames TnsnredInouaen ed move ta lamp From bands (estimated uncertainties were 5~20%6) and the use of Morse based Franck-Condon factors and interpolated Degen deduced a linear decrease in tra i with r(vanation of about 20%). An approx ate method was used to obtain integrated band intenst ties from measured rotational line intensities in these bands, which heavily overlap one another (because 1a!" ~2u,). Because of this complication and the dependence on somewhat unreliable que and centroids, the transition moment and band strengths are likely not very reliable (4600-2800 A). Degen and Nicholls (106} have made photoelectric relative intensity measurements of 36 4-X bands produced in an argon oxygen afterzlow. In addition they tabulated calevlated relative band strengths using transition moment Re(F)= const. F-M# for the range 131A 0.1). 11.10. UV Absorption Cooffcions of; No detailed examination of the absorption coefficients for Or is intended, because this has been incorporated into a comprahencive roviow of UV photoabsorption rose sections by Hudson (198). Watanabe [399] had reviewed measurements made belore 1958. Sullivan and Holland [367] have tabulated absorption coeficient measurements made up to 1966, for wavelengths below 3000 A. Huffman [208], in a more recent summary of absorption cross sections, discussed discrete absorption, Dut tabulated only absorption of the continuum at wave lengths of solar lines below 1215 A The most recent measurements (which include cor rections for scattered light and pressure gradients where necessary), not always in agreement, are those of Hutiman et al. (208), Cook and Metzger 95], Matsunaga sand Watanabe (260), and Samson and Cairns [341]. The smallest baud wid woe fu these mcanucienss was 03 A. Observations include mumerous unclassified absorption maxima. A detailed discussion of the sources of the continua tie, which dissociation or ionization process) is to be found in the cited references [204, 95, 260, 341), Some highlights are discussed below. Table 56 gives @ succinet suuinary of tome special absorptlon features; the cited numerical values should not be regarded as definitive, ‘Thompson et al. 13731 detected no Os absorption in the region 4000-2050 A, and concluded that absorption cocficients were less than 10 em, their sensitivity ‘The B-X Schumenn-Runge continuum dominates the region 1/90-1300 A, tne peak absorption of th asymmetric continuum occurring near 1420 A. Absorp- tion by the BAX transition has been discussed separately indetai (ser. 11.6) Wilkinson and Mulliken [414] have detected weak ‘continuum arising from a transition to a repulsive "Tl state by measurements in windows of the 1780-1850 A eon. The only two measurements give k= 0.44 em"! 2. Phy. Chom. Ro. Dat, Vo 1, Ne 2,1972486 at 1781 A and 0.28 emt at 17964, The f-value of this weak continuum is eshmated as 001, Diffuse absorption peaks at 1293 A, 1382 A, and 1352 A, detected hy Tanake [369], have also heen cheered in absorption snofficiont masenremente hy Watanabe et al. [400] end Goldstein and Mastrup (264) Tanaka has suggested that the feature at 1298 A repre- sents a continuum arising from products *P-+'S; the other two features may arise from 'D-+"D or P+. ‘The region 1300-1050 A includes complex absorption. Price and Collins [829] observed @ weak continuum beginning at 1105.A and extending te chorter wavelength, ‘Ogavea [304] has measured shsorption coefficients in the region of the Lyman a-doublet (1217.8-12168 A). Forthe doublet way [tem 20.02 anses | oan Measurements were made at pressures 0.3 torr, and the results extrapolated to zero pressure. A 3-m vacuum spectrograph was used (dispersion (1.45-0,95)A/mm). Ogawa (204) hae summarized results of earlier maser: ‘ments, Gaily [151] has measured absorption coefficients (1215.67-1212.57 A) at 760 tor, for Lyman-a. He obtain ‘ed values roughly twice that of Ogawa. Other windows {k<1) have also been studied (1300-1000 A) by Watan- abe et al, 400], Watanabe [399], and Cook and Ching [94]. (See also measurements by Ogawa and Yamawaki [07] atthe 1187-,1271-, and 1207—A windows) ‘The region of the Os spectrum below 1300 A can then be loosely categorized as fallow: =1900.A. at minduma(h 200m 1200-1100 A: weak continuum ( ~ 50-100 m=") 1100-1020 A: increasing strength of dissociation con tinuum; complex spectrum. Rydberg series converging toX,O3 should ie in this region; none have heen confirmed. 1020-840 A: intense diffuse Hopfield c-b bands Uf Of. Rydberg series, inlzation continua overlapped by dissociation continua (E ~ 300 em~") 300 em”), Absorption maximum at ‘900 As another at 300 A, ‘The f-value for this region (0 < 1300 A) is ~6 (245, 407, 113" At 1094 A, 1069 A, and 1044 A, Watanabe and "Seer a ewe aes 4. Phys Chm Rel Bete Vo. 1, No 2, 1972 PAUL H. KRUPENIE Marmo [401) observed absorption minima, with values somewhat pressure dependent. This was interpreted a ing feeble but unresolved structure, iacellancous (transition probebility pperemetors) 201,855) Doolitle et al. [117], studying photoionization of Oz with (584-858) photons, have observed possible predissociation of OF, B'2;. Predissociation lifetime, short relative to the radiative lifetime which is estimated as 10-* s, was assumed to explain the absence of the allowed transition B°S;— "I. bo ly crude estimates exist for the lifetime of Op Electron impact studies led Frost and MeDowell [147] to estimate r> 10° s. Conway [92] and Prasad and Craggs [S17] estimated the lifetime of excited Or oro Dianov-klakow has shtained ernde estimates of the lifetime of the collision complex (O,):. From band widths [108] he obtained @ lower Iimit 03%10- 5; the diffuseness of bands of the complex, hence the spacing of its energy levels, exceed the spacing of the ‘-X, 0: hands which are characterized by Be~ 1.44er-? 1 1 sch that rag =p ~ 2% 10 (109) Smith (358) has considered the effect of autoionization on the relative intensity af vibrational peaks in photo electron spectroscopy. Qualitative calculations ar ‘ven for O- transitions to the O7 ground state (compar ref. [0). The ariking effects of autivniaaton ave b Iustrated in experiments by Branton etal [59], who used photon energies below 21 eV (namely, the neon doublet 736 A and 744), onathan et al [222] have observed the photoelectron sectrum of O:, a!Ay (to 03, X"Ily. The relative tensity distribution was qualitatively ilferent from that ‘bvaied Inthe FES of Os, 42, co the same sate of theion Studies by Freund [145] on molecular translational snectroscapy suigues! thal the 8Tly siste rence by Liehten [252] at ~ 12 eV likely has a lifetime < 10-Ts (Cichten had assumed @ lifetime of > 10»). Freund, Jn addition, obtained quabtative information on ¢ repulsive "H, Rydberg stale of Os having. energy 22-23 eV Ogawa [305] has observed six absorption bands righting in the aA, tate (n new band was found x 1455.0 A). Tentative absorption coefficients were of‘THE SPECTRUM OF MOLECULAR OXYGEN 485 ‘ined (1486,0-1408.5 A). Preliminary cross sections for rxcitation of excited states of O; have also been ob ‘ined by Konishi etal. [285] 12. Miscellaneous (@) Bridge and Buckingham [60] have deter. mined the depolarization ratio at one atm and 20 °C; r= (3.09-40.02)10"% Polarizability is 1.598 A® for ae A (b) Molecular quadrupole moments (definitions and ‘numerical values) have been summarized most recently ly Bimbaum [44]. The various measurements for adisagree by up toa factor of 10, (c) Sehnepp and Dressler (348] and Boursey et al [50] have observed the Os, B-X transition in solid () Ben-Aryeh [36] has discussed emissivity of the BX transition for optically thick systems (where there sre departures feom Roar's Ine) (Gow alen Rlako ot el |16}) (6) Akimoto and Pitts (6] have observed Os, a-X, 0-0 emission, together with simultaneous transitions in sohd oxygen." Ihe hletime of Us (44) was estimated as © 10s at liquid helium temperature, 13. Summary and Conclusion For many electronie states of Os only a few vibra ‘ional levels have been studied under high resolution, ly fragmentary data exists fin sow states, sind in some instances, the vibrational quantum numbering is uncertain, Isotopic studies can be used in principle to provide the unique quantum numbering, but the for bidden transitions of interest are inherently weak. The states which are incompletely studied include Ay, e'Bj, C8, and A Zz. For study of all but the ay state very long absorption paths are necessary. A potentially strong charge transfer transition C24,—~6 3) has not been ohserved; a large Ar is a complicating factor. ‘The apparent irregularities above v= 15 in the left limb of the potential curve of the B*%z state ate stll under study. Potential curves calculated with CI wave functions indicate that the repulsive Tly state crosses, the left limb of the B°3; state. Predissociation of the B'S; state, primarily arising from spin-orbit coupling, likely involves more than one repulsive state, ‘The answers to remaining questions concerning predissaci tluw of the B state may depend on more caretul examin: tion of the J dependence of measured line widths. Several states of OF need further study: Cy eASs, REE; From among thous anly for the 18s has there been any rotational analysis, An electric dipole allowed transition B°%j—A Tl, has not been observed, but this sr. oF transition is not inherently strong. Fronounced changes in spin splitting are likely near the dissociation limits of states B'S; (Q,) and A'Tly (03); the former had heen discussed some years ago, ‘The most reliably measured of all oxyenn transitions remains b'3j—X 3; , with As F's disagreeing by at most 0.005 cm~. Recently computer least squares refitting ‘of these measurements which were published in 1948 hhas demonstrated that the early graphical fting gave too small estimates of uncertainties in B values and band origins. 4 Py: Cham Raf. Bat, Vo, Me. 2,1972a0 PAUL H. KRUPENIE Tone rete tin Fn ec is Teas | rom) ach ty lieeeee fern tae feces | mor] ers [usta a eo: ase Jace | oraz | sos |osoou | Ce Pe ne sea geme | eal | raat vse | sen | um | 7 aor SSS Samoa pope mem fe Jocosees|usasn} og st] lease | rome | 247 leaeste sees |-on0 |erues| eee ex fess) os hese hel Scare ee rgeam mccain Berea eh meal tem "or a arr hieg—t henna Seger ae eu aire () Misprint in valve of a ven by Degen [10 is corrected ‘V eicergctdowettntesetiraret gain Epa ia mare ty Ferte-h AE (a) Mlecular constants ited to #2, ‘aly fom electron configuration arguments, The leton Src ei eel th Sse a ea ‘ng J, as expected [374), since rotational energy increases de. implying # normal °8 state. The some erpument ic ra ta on 6 Rodeo iim ams sutoninntririnces 9 akeetincat at Ieeular rotations the would ser the quoted vals of B fdr te) Repructons of ooX spec ret 198,277 4. Phys ham. Rel Bete Vo. 1, No. 2, 1972 1) Fora ted y Dy okies «ten 4.14 00-9, {c} The state deviates lv Hand’ eave band areas ave ¢ nearest i,THE SPECTRUM OF somone for the econ tes of Os ond O:* (MOLECULAR OXYGEN 47 ve jae v] mene me Emo |e |e [amen ear lose lms vjrarsomin | mo = ofa fem an te | | | Soa | pee fee | bs e|> tee | xe wee oe | om va |e 0 = oj [tee i cyte jae Joon Suoe TS "meet fereeene enecen] 3 | | ns lm fern] [Gace ean te | tae 2 See ‘ (2 Tet sig ofthe Atte the reve of hein (0 Reprod 5X per 1908) ithe round sate hs Fa >Fs> Fe I8TE paameter 5, Vinal qantas enti gamer een eincauan “hea ee ded from even J eve nis a levee re (g) Reproduction fA-X pect rel 18,108) {3 Maecalr comtanc se fo a computer Fest! squares ft levels 18, The emi set eoeticients i ain tothowe ablated neue: Yoe=2.2067951 (3), Yom —1. 509057 (4, Yip daaahe(=6) for ibatonal tre vale ard YouG) = 1.5426 (8) Yo-= 6.1580 (6). Yip ==9,5818 (1), n= 2.90100(-8) for By als. (b) Concent ited by’ Ackerman and Bis [3] were hae erapieaty: {Postion of 122, the last level before dsocation een eaulted by Brix and Herbert (62) Fragmentary ‘rut of the 22-0 band was obcrvd, But the role 1) Yor" oe (e) Low order coins ted tow < ar flows fer 1.30), are, ~ OOH, wie =~ 0188732: eo (0 Reprodueton of B= specie, res (9, 288.62, 32, 306, 6.3) (9) Alberto al 1 dented this tate a (208 centdethe wtb By 2 Ozawa et (0) Ty is wbined from spproxiate ae, ase and exteap lain to 20. le) Elector confcurton and ibis! (a) 304) = 199-2 (6) Virion levels are perturbed wi only appeoinete 0. (8) w-bon(0-0) 631815 (a) De ie abnormally large selaive to that of mst eer states af Os. But D, Sor the fl and sates are O(10*) 000) seth te) (0-0) (lle) 81625. 1401-1105 Ryden sia f10;)—X0.) Ryder ces! Reproduction of spect. ret [93 ries: Reproduction of pec ef [83,27 tin of ent oa X°llg 19) Vibrathnl constants fied by the ator to 71 iil eonstant Sed by Leta anpebaked vy Seno "61 0° am ee ation et aL 1¥) A the eaty sermination by Stevens 35), 1s Vesoneidee of dealing se Stevens 36S. le) The wnssionnaton measurement of Sarton and ‘enactments an the ATX bande of O: 1 Male and Reo (2, 43} eve 97265 em! or 58 eV. Uncensimy af the photoionization ‘ange ncerany might be ataced vo thi ale sy Dele and Walker (11) becase of porally feslved_suuonged bands ip thi reson, a ‘io, the thermal nbn a sous emperors nde 0 peak popsiion for J=8: the resolved ratmional envelope ena induce an unceranty assumed ta correspond tothe lye ni The Chynetetea) #1 sate without spirit opting wold be HO cm" higher the separation af the ye Chem Rel Dat, Vo. 1, Ma 2, 1972458 4m os PAUL H. KRUPENIE Fount to table 1—Continged Nisholon [249 had lusined value for IPO) from iniationefcency curves which je virtully ‘estes that of Sameon and Cains (24), (0) PES)=20 0) nptspecuvacos) 16 1g) Reproduction of photeceeton spec ref, {1254 {a} Ths term vl is determined Crom the pion of (he b'Sy on deduced from Rydbecy sere S17, ‘ith an approximate coreton for hesdign ‘epration, and the O-0 bund oii for train, Onis. (0) Molecular contents ied to v=, 1e) 4 lon0 196) varies monotonically from —67.92 tn 48.06 [28]. (See slo [ee 1a Yano 10) Sqn PES)=20 [1251 tmeopestesenpy 1a) The term vali is derived frm the tsumied po tin of the p= 0 level of XL state wth pn salting pl the 0-0 origin (4070 cm") sl [Bed rom the dimted number of AA band ‘wigine table 29, A ft to the band heads be fates the OO doublet a 39879 and 4079 en ‘The extrapulsed band origin for 0-0 dies nt lee theads- and though Based om fee data pit, ie based on nore rlale date than the band bende ()A=62 [365]. Adoubling Ie nelle in the tecanectn free ED) Vibrational contante fied by the author to 98 Retaonal costae fied by Lafitte bse and Abit o [3]. (oy PRSI=16 (125) (0) AX, e(0-0) = 4000s ealeaated (a) Reproductions of O;, 4-X spre 213}. (a1 Term vahie deived fom int of Ryder serie [R17 ecu band orgs, based onan approx fists correcta tu obnerved band ed pone 0002 eV Band head mesrements weal ve Bt 14658 en by Moles constants fied by Albion et al [9] 1 Yoo. (i) enedPESI=6 (125), fe) oul0-0) (0). ba) = 1660.78 1 Rerodation a 0: Po een, (4D rele (2, 4. ths cher, Re, Dot, Va. 1, Ne 2, 1972 C24: The peal curve fr this atte hes been estimate by Gimore 1161) Jude [224 a recently ober fumercnce tubing amano ten, pes by 690 A radiation, With 625 A rdition (018 et aiference the Banescence paiern changed x0 th Sage ssued the psn a the By atte tb determined 21948 n within O15 e¥. Lind 25) ba dated x psile Ryderg sei fm as the enrgylin pes abvervedby Cag endSches TUS the tf at 19.96. {sh Term salve derived from Rydberg sees ii FAITE cAMeboary ot (0 ft garth Shs stem be Es by aogier ing states nN. (8) Uncertainty in = :5;uncensinty in onae~=2 {017 Simla vales wee sbtlned fm put ‘econ speciroscopy 125). There mains analy (ona T3i Ano (Assuming identcation of tis state is core thee seit bands fe B°S5 4M, wath (0-0) ~ 26160 co (90.5 A, (o) SPS) = 90 129. oncasied) 213 eV: The crm se heme of electron ipect Ineatrementa by Fit and MeDuwel [4] ond Sn 6 Sie an Lindi 57 eae ca Innenlecte ration. ie Ss: Gal pene eres it $121 for & state pl eb, (0-0 pation au extraced by LeBlane [250 from {he rnesignmeno the Hophels bende Rydberg teres of Coding and Madden (ve heels the tame mt Molectar conn forthe & stete a derived by Llane (230 ae a nicely tone dered hy Nevin abu 11) B64) fom Myer 6) emo PESI™=2 125) resevis i 154009) #12500, Possible new sats: Features tentatively denied os new states of (0; have been obnerd in electron and phon impact experince [ae 125, 3801 They occur atthe fai eres 225-94 assumed to he My by Edisto ab. (12), 286, 275-27.9, 21.3, 3, “Enlarged copies oft le 1 ay he obtained fom the author span"THE SPECTRUM OF MOLECULAR Uaruen 489 Tone 2,_Elecronic confgratos for molecular oe leo contain (ton) toe) ex) Gn) ty) lee) Om) Con) 12 4 2 eee AB 142 3p B85, C35 85 or 2 a2 Hh Ma A Mes "Ba 8s | roa xen, Te 21 Pgs Rre ae 26 a [nom za 3 pO, Me ATE BE, Ca 202 2 2 2 6 2 eP Oe Pye 5935.65, My My ay A985 235 2 {he 27, ate repaie nd rodanoites B ‘The H2.°3;, By wate are sgl bound, bt mot aerved spucrsopely For Othe C8, sates known rom pone Ryders nese and Saeeacence. Se fetntest tbe 1.) For O7 ee sc 24 Toms 4 Band origin of heb mosphere t Irae a « ” a con) Toate 8 Band oii of hea eX P37 tmp IR ate 8) " rm | nasiszan | wai [aes Reh oa | oaeone 1 ‘osm a a 365 feoom | sane ray rsa | Sos % 1 Sosis | irsnoma | 0 wma) | soe s Co sone | cass) ra) (1 band head observed in tight emsion by Valance Jones and Hareos 5) (+0, 1-0 band aii from igh resaiton measurements by Here ‘eng and Hersbere 191, ‘ean ne rom ngud oxygen absorption measurements oo, ‘Alin, and Welsh 80. 2-0,3-0,4-0 bands oberved i ai oxygen by Landa, Alin aod Wels [24]. These poitons are sifted to higher wavenanber ry wi10 wt eae senda gan pe pon ee tls Bis and Keser [126 ‘ass from Rabeock and Heraberg [2 refed using least squares ty Alston eta [9 " Obtained fom fragmentary dat by Osenbrgzen (810) wove | roan | rss | 00 caoist | 150626 10 Dats fom lshenek and Herbers 24) woro | roar | asizzses | 0-0 avese | 1400s re ats from Baboock and Herabers (28. 4 Phys. Chem. Re, Dot, Va. 1, Ne 2, 1972460 ‘TABLE 5, Band heals ef the ®S2=X 23> armoshere etn UR) dy on Ret a fen") own) sw: oF aw (c203) 1361 2s 196, 184 97.8 apo 12 196,184 (2223) 586 a 26,270 7982 517 “ 163 yemat | 2588.20 3 195,188 roa | esse 2 199,138 7398.78, 13165.14 0-0 196, 186 70 13008 x 10 om 4000 “ 18 (a 1378 2 18 en 158 Lo 8 eas 1928 2 98 Presence ofthe 0-2 band inthe nh sky hasbeen sogasted by Dufay [121]; Keessveky [240 usns higher deperson di not obeerve it Findlay 138, dd observe iin an alterlow fom an onyuen di i 0.0; explosions by Hornbeck and Hope 19 ond Herman et 1,845 Most nence were O-Dand 1-1 Herman et 18 produced tmany band in high frequeneydechage through Os a atmosphere Tins (196) by wp to 5 A. Carine [98] observed two bands ina sorption by the earth's stmonphere. Observed waveleath of 02, 2-3, and 0-1 bands ate amcertein 90 ae the ienietons A bond head 85074 har been idanted ae O- 106, 18%, 19), bat ae far from ste predeted portion. The wavelength ~ 865 A given by Kaylan (26) ant Mine 20) forthe band labeled by them 280-1 artespond tothe ors of that band: the head postion nen above ‘Taaus 7. Colenlated bond ersin watelenathe (A) of thee 35 PAUL H. KRUPENIE ‘Tanz 6. Bond origins ofthe eS; —N°E; Meebo Messen h) a vem bt ‘aemision 4015 sks o Data fom Depen 108, 1 absorion ams ‘75, oo ask 38581 oo 2002 ‘461.30 o 208.0 se023.70 0-9 Data of Heribers [1894 Tentative rales used are shore of Degen 1103), and nea ie ‘ereae fone hse proviinal ed by Herabee 13, ‘The broad Condon cus for this system esl ia weakening of the Bande st lnger wavelengths - 7 , 0 1 ais + 5 6 7 n 2 1 Sioea | Set | Sane | Sone | ea | Som? | Huts | tana | ace | smes | sass | sano 2 densa | anea | sna | uona | sro | awn | aanoa | suns | soos | sea | Sue | sss 3 on | sea | amo | sims | aos | amos | aga | some | sons | ont | sats | sz : Sans | Soret | hats | sna | aes | som | arse | somo | saa | sont | ma | soso é deota | moze | soca | sasns | ana | soma | soe | soma | asa | suse | aro | snos : ams | ares | sows | ota | sure | sere | siozo | use | sans | aes | se? | 9052 2 ues | ms | aoa | neo | soma | sets | sun | soos | ssto | cs | save | snes : Ho | zea | oa | dats | Sass | Sa | Sao | sro | Sess | sigs | sano | acs P fere9 | mrs | mito | ana | ates | sna | sea | asco | sera | amen | ats | sor M foots | fast | Seaa | ants | stees | san | susoo | aoont | airs | sma | sons | sou Bands enclosed inthe rectangle are tho abserved by Heraber stu by Deen (105), wlio pcan ths table wate 41 Phy: Cham Re. Date, Va. 1, No. 21972 sorption, Underinedenvies are hoe rm afters emissionTHE SPECTRUM OF MOLECULAR OXYGEN oc fo} Bond oriin ofthe C*8,—X°E- Herbere system (R) & (omy ow 251882 ‘6879 |p 2508.36 seen. 98 le 89385 ast sor Dat fom Herter (195). Vibrations) quantum sombeing is Wy Band hae A a a Oa TT 28 090 Bis aan oo sass ss016 me ssits 10 205 asm 7839 ssoi0 20 ne see mas sonal 0 Ema sors 209.8 see 203 mae +o aos aan 0827 as beset mo so 28 sao asa sow 2593 2594 oo 25828 same 25599 so0se 2535 sera 1 253? 293s 25084 39386 nr ane eo 10 29e2l nora 1020 2081 ane + ne oe ans oss 281 ass wo 2965 69 1 20 Sai Dna from Herman {182} and (a) Fikelnburg end Steiner (1) Fie quant esignments by Herzberg {188} are tentative, These ales absorption maxime mere fst observed by Finkel snd Pore [11] in aaygen a presses fram 60-60 stm path lene ‘sre2-26m, There bands have not been observed at low preset 461 {below 1 stm) More secent measurements by Herman [182 extended this system to longer wavelenth Herman observed tha here tilts Incretsed i srenah wi increased pressure, while the AX bande which they overlap, dssppeared. The ener absorption maxima (strongest) comet closely with the wavelengths of the 4X ban. Terman observed, in addon, feeble structure In the region 2000 3017 A, and around 3050 A. “The spacing and weakness ofthis ‘coninton of these pets ‘Tawue 9_Ponsible?4y-> 0, bande ae Tene a identification we ss a, tae = 2 sr mi iu a s we os 2 tn a s is a s az a ea e 3 so ee ” so a os 4 os = fa oa won ” os ‘on os ea 5 ot sen = ia so aa a seu + vs a ra = ae & oy am o 2 san a as Date from Chamber [o1) ean numbering is tentative, AME JU and arene ofthe A222 —N"RZ Merbera | Syatom (95007.15) o0 uae - 35780 05 1 5561.25 cay ese a 2 ‘0166 0 ay T2348 0 84 7 70 simozs 0 ‘as 00 e546 26 oo ansa94 1-0 ats from Herabere 117: values ate larger hy aity than thove wed by Herbers (se Broa and Gaydon (6). 0-0 and poiioe is ealelated. The 1-0 band poston derived irom relatively fae mentary dat, 4 Phys: Chem. Ref: oto, Vl. 1, No.2, 1972482 PAUL H. KRUPENIE ‘Tamu 11. Band ew ofthe ACK sate oserved in emiion (RI T T ee eee i ica eee eee i = ola lpescn fecal ws | | coma emma See ae Ses (ils m | ee | i]s Se | bes | i] |» Ba | Me | 1) |» S|) ge |i] 3 sa | om | 2] | 8 a. | Ba |G] S|» ar | im fe | SS te | oe’ |e) i se | ee 5 mS | me |S] a a es | om [il fo» mm | ae | 3) 8 se | mm {| oe we | ose | os | os Ser | Bea | 2] da |» a |e [i] 8 BS |B fe] e ms | oe |b] 8 m: | a. [2/8 | « m |e il 8 Be | wet | st mike |3) a mu | Boo [P| |» m | Me [i] a 3657" ust | 2 O8 30 2086, ‘37082 + er Bio | fms | P| |S mer | me | 2 | on wor | ze | | one ml mm | i | oe Be) Be |S] | ow tm | ae fa] a : ‘lg m |e [i] o ons —_— IS, 2563 39005 1 m1 oma | ome | «| EF] a far | Bae | S| | wr | Bat | | Bus | ocr | | ie |» end sd wv fo sd ayn ee supe ye a an Kan alerted Ennead syne emen ty "Mrs cece fm Bre tl aye a ine fr fod Kaplan [0) Tae were saluted by 2 formals of erters UUer} Laboratory afterglow lense re Tom Brida nd Gaydon 4. Phys Cham. Rafat, Vo. 1, Ne 2, 1972 to the stem, ith any quanta sasinmente voclengi reported. Soe rel (29, 0-H). vepeated bet‘THE SPECTRUM OF MOLECULAR OXYGEN ‘Tame 2, Band oriins othe B*2z~X 463 3) Sehumann-ange ste (R) = os =). =a are gree cas ny (em) (ay (om) Ay (cm) cdr [am] om bom fmme [owns | ow Pw daze | ema) os | Jee (ag fest] set] ce] s fas] ae) os Pee eee me ee | Gan] oe | [ee [as fee | gate) ce) ae Pee | es] oo | ae cp ee eee ce (oe (mse lf eee (ee a3 | 183 | soseo om23.0 | O13 | 256 Ls | 4095.9 | eT 136, a esle] aceasta fee cee |) |g am | i || ses | sien ais | 328 |] 25961 | arma | 26 | 136 £20) | lee | eeu ee | asta re eee ee] (oe a ee ioe tare (ge ea) ore ead (oes gee eet eee (pee res eee eee ey are faa | ee steal cee fee) as) | Soa eee sy ee et lly ee ors a ee aol (fice ctl ee [aes | re me mee | mart ae eee | eee | Sey) ee dee se [eee geet ee ee) mee) | ee | Pace ray ee ene ees | |e eee | eo (es | eid ast fees cep pee iesme eat els feo et ee eee oaeedl atl eeatesel oes peer all eas meer erates cee eae eee es ea a eesleeece lie mer | cet | oe | ef eet) see as | || aes] | | eas wd | aants’| aon [ime | wae | nase [oe | me sar treet] see [Sas | ae ts | me | aa) seme FS mes | eae Se fms | ames tre | om Tae | ee reat Proeeeg| eat ea eter eed nt] B so2]|res = 10-20, 3-25, 5-27 though no ie tue wae published Calelaed postions of elena long-wave: Teelogreven end Dieke 256) observed no low) lines for tol 215 ani 2-16 For 0-1 they oberved structure only for bigh "aes 8061-9128 Maketarny, Wenier Gena 52] sive only « Dslandres table “Nand arin weelengts, but no Be strctre rst 134 Mion and Herman (28) lst band beads only. Fest (55) Kate alle atemative quntem afgnmcnty fr nevcrel bend Feast and Garon 136, Woclboeck apd Bauer (85 resened rotational stractare fs the 2-1 band in an epecally intense source of the B-X binds in emiasion. measured line positions ifr by several em" from the nore ex tensive measurements of Feast and Garton (136), Many bands which have been observed under lv resluton (A/mm inn store oectroarat have heen atibuted to Ox 2X. Thee detentions by Vegard and Kee (30), with many bund aslaned > 20, are 4 Phys: Cham, Ba Oat, Vo. 1, Me. 2, 1972PAUL HL wf ‘Tamue 12 Band eres ofthe B 5X °F; Schumann Rance system (R) Continue 1b sbeoion a Ref » vow [nee ae oo few | Ret oy a ay | tem 16 v2 |] zsos | susos | a5 | ise |) waste] sis | so | 3 902.98 wa | zzme | sor | 55 | aoe f ivest) sams | 71 | 502 wae} ass | aster | os | awe f] imoa | sasso2 | m2 | ate ena uo ff sim? | same | a] am |] tcorse | sasciae | 0 | os 92.08 ua | rer | aoe | rs | aoe |] agora | soso | 21 | 06 214.00 wa ff aisos | aoateo | 5 | ame fp rover | samer | ze | 306 a9 we |} reso | somes | os | aoe sams |ae2 | so 65.67 wef) ass | sso | os | a sames.s | 9-1 | 306 wie wz |) aise | aoreo | os | ae saat | oo | 8 gras) sores | 210 | az |) ano2 | ais. | +s | ae sazins | 1e2 | 306 sm] assinas | ia | ia |} oss |. arsos7 aos | me suo. fot | ae sunee] awerst | oi | uaz ff one | arsrze | v3 | 300 sussezr | 7-0 | 3 swan} 2ensa0 | em | ue |} mots | asa | ss | we saa [ui | 06 sas] agero | sae | ue | ams | sams | 6 | se sis | eo | os aozor] mmaras | ote | use ff 20002 | amsaso | os | a ff sea | sume fr | oe ssazso| zoe | aie | 42 |} mso's | asst | oe | ae |] rans | susa07 fasa | 306 08.96] goonia7 | ai | a4 lf 2onn6 | ames | ros | soz |} resors| sieozrr | 0 | | 3 swoo.30) eset | 2-16 | a2 |) mses) aonstio | 2-1 | te |} anes | stoma | 15-1 | 200 soe | z9e08.25 | o-rs | 142 |] ass | ssa | rine | ae |} reres0| ssoso.90 f10-0 | 3 saas| aomass | os | 2 |] amos’) owe | 52 | ow |f imei | ssi fie | we sean} sows | 2a | a2 |] aorz2 | asso | os | am || raus| soaavo fue | “3 siso40) siorzao | 1-13 | 2 ff aoorss| aveon | a1 | “te |] ame] soatsr fimo | 3 sisiso| siose4e | sae | ise |} 99.0 | soos | o-s | am |) srazc0) semesar Jao | 3 Sioa | sza03.9s | o-t2 | 142 ff rovee | sooo2 | 2 | ame || imasr| seanar [eo | 6 suse) szrasa | 2s | use |) rover] saowsee | 1-0 | 3 |] reese) sessoss | iso | ‘uge-76| sogera7 | az | use ff resco} sonsse | 1 | 9 ff rasa] sorsso Jeo | @ goens0) szsors0 | on | ise ff r9ezo | sons fos | ame |] iiseou | seaseat [iro | wo | amass we | soser.1 | 2 | 906 |) atss79] senssss | 0 | 202 | soon.2 wast] smom | 2-0 | 3 |) sissae{ stows | w-0 |e Bai | sew por | sist [ates | om] Titec) Sromaas |2-0 | a pian | asm. rwwoss| sims26 | 51 | 9% || toss! srtacr [are | a pins | oss 9 vws82 | suo | 2 } so | crsosy| saa fone | @ rasa | doacne | a6 | ise |) srs) sissaa | so | 3 was | ames | 6 | ise | rows | sissas | 9-2 | a0 zsier | asiezs | 2s | 136 | 99925) sisee | or | oo» fh Hercang and Wieland (18); p=80 am, 7=1070 °C. Rotational ideatfed ak 6-28, 4-27, 5-28, 6-29. More recently weak ban consents of Guny and Herabere[ 9) were ned. ‘Oza 802) Many hands observed br Ozawa had previously been reported by Rekotariny, Wenier, sud Crent (24) and Feat sd Garon [136 ‘Donen and Chane Fiasimmune and Ha [142] observed numerous band heads which te included in tha table of ein) of seat excited exper produced by secndary processes in ash photorsis of enone. High 4 alues were populated. Some of thee had been jroduced eae iy the same pote bs Mah and Norrish BS}. New bende Ince 3°38, 1-16, 2-17, 3-1, 3-13. In addon, bands were ob served st wavelessthe longer than AST0A: thee ere tentatively 4 Phys Chom. el Beto, Wok, No.2, 1972 (-77 and I-17 were detested in sh phos of ome [264- b Caml [26 observed thee and head ‘Bens and Garin [Bl] observed seer! bad eas flowing phe Ackerman and Bisue (These estenive new messoement persed the carer work of Curry ond Hersherz 9] and Knauss nd Bard 283) Orin of the 22-0 band is rece by shut eateepolton. On feauneots ofthis ban. the last ele siwciton, were aierve by Bein andHersers‘THE SPECTRUM OF MOLECULAR OXYGEN Tasue 8, Band ofthe PE! NEF gue P) Orn” on Ret Tags, Bande ofthea %; ) ma fw] wm [osm] om | >] m0 | 7] mom ve 6 T0613 | 9] 107s | 43333] 1165 fo omar a] 4s of ns} s | nmm [1] mene | ar] mmf 5] ene |g | tao ‘ee | ae fe can] tie [2] em | of tone ime we] foe |S taf nee fie | smn | | tera ts n| gue | é cam | ne fir | eres | 2] tera tm af ome | wae | ne fiz | eas | tf tres ter | seme | | teen fia [nea | osm | 3 | uma Ie | sere | S| team fies [ame is | eae | S| eam io we | seme | | team fitas | aie fie | sua | 3 | laa rer i | tices | 2 | two [ita | te far | Sms | | tana ‘ot | aman | 3) ee ise | tis [ie | Sees | 0} tana tn | om | 2 | eae [ins | iim fr | omen | 0 | tae =| as (2) wean [as | nef | enee |] tae | dea | 1 | wan [ss | tne far | gai | 1 | tees B| de} ] Mean |s | US ]2 |) Shs | 8 tee B| gem | 1| wae jas | im fa | sea | 0 | teen »| are | 0] teow [zz | ime fa | sem fo | twas 2} gan | 0| tae fa | iin fa | sam |m| twas B) 8) Meas Ee | ne ye] Ge | | ies at | 68.076 o | 63s lesa aie f= 671,7610.015] 148 858-43 1090, =| gem |@| twm fra | te ae a | @ess || twee [ine © | sarecoam| | estes bel ar9] eaten (1) bad wert 365 i 4 a oo =| $ | ge fd >) mm jo] we [aml ue[t) 2 jt ; Say [eye | ae |: n| fee | 2 : Sf iteoee | eae | tae : f]sS@ |e] nape] gs | a] ieee fe | ase e sf meas Jase | Goer | 1] tease fica | ter 7 S| omar fest |e ie | rm | | toan fae ‘ @ | imam fine | nm | | sim] eo] war | asa % S| imie fice na Qt | eae) of war = te S| iBne |i | nat = = n 1 tan fisse | nw n T | tear fies | te (0) un i | smo | 1] tera ita | te | aa S| Gem || twas fas | tm fs | mom [0 [ me | ane a| oma [1] ames fins | no |e 2] oe af imee [me [im ft | mse | o[ wae | san =| omar |i | em [as | tim |e | eae | 0] tan | ese S| Gem 0 | trae |e | ime fio | om | 0] use fas | tue fin | am | o| wore | ose z | ono | 0 | wan [ase | tem | 0 | toon | wae =| can |e | twan [ine te] tar | a | tsoae | tea 2 | eons] | terse nim |S | Sedrreno] "| tomes zemant rnp cng mens ae fr ee om eet a Sad phd ha WS cm 4 Phys Carel Bot, Vol. 1, Na 2, 1972{THE SPECTRUM OF MOLECULAR OXYGEN « ee re or in b503)- X70) To pm)“Tames 21, Sire Ryders series BPS(O!)—N7E (0) T ef aa for Leow Po Paco fff nar Por een Jot Pace (0 bend ne (2-0) bad ses 3 +] ame) 12 39 nw ffs | on iat 5 | ‘Grae 6 | isn | sam | ns Is | “sae 5 issn | 5250 | aos + | esa 6 [ism | 30 | ins fe | otras + [iva | 330 | eos 8 | sataee 3 | twossi | ests | te |) 6 | sizer 3] isi | oan | is 9 | sass § | ira | ta | ano fo | osm 4 asses] Ta | ins 1 | ass 5 | tsease | ame | nes fr | errs 4 | tet | 9310 | ies 2 | Sane S| iwen fio | ine |e | asssr 3 | isease — froae™ | tos 13 | ison 1] meee |inas | ne ffs | esses 2 | iso |e | ta 15 | isa 5 | tess [isa | ino fs | oacors 2 [iam — fisss | toe 1 | sine s fasts fies | me fis | our 1) sis fies | ion | see 4 | wz [isan | mo fir | coxsoe 1 fsa fisse | 0 | case a fiese [183 | ne fps | oui 0 | issie fins 2 | ono 2] uss isa | tno If | eozerenons 15 ase 12 a | ans 1] isa five Ll 2 | ones 1] say | any 3 | Siw a | iste |S x | atten o | tsa | 226 spe | sam © | sto 000s 1s Net aioe [5 | fares ot + | sisase Sor -Obandoee wie fe | oot ah 9 | asso tan «| a 9 «3 ] | eos e333 5 | ‘vase a | 15 | sm | caso os 7 | ease 6 [imi | S308 | at eee hat 8 | orzo @ | teeost | ate i | saute ress 5 | snsawe & | iet@ — | S316 1s | sao Be i | as i figs | sae py | oss 7 te | saser 2 [ist |rosr | Setesooe ~ 3 | asin 2 | tsssor | nas 15 | amass a [ise iss we | mass 3 | team | sar 1 | este 1] tse fisas | mir o friar fina | ars o | tora [ies = | eszczoon |” | sansa 1s Dat from Youn ad Tana 47). Gilmore 160] asigned the O; sate hich the eine most kel) 25. 9=4 ers obneved by Tanks ond Takamine 373). 4 Phys Chem Rel Bo, 1, Ne 2, 1472