Multiphoton ionization spectroscopy of hydrogen iodide Scott A. Wright and J. Douglas McDonald School of Chemical Sciences, University of Illinois, Urbana, Illinois 61801

(Received 10 February 1994; accepted 16 March 1994)

The spectroscopy of hydrogen iodide Rydberg .states is investigated by resonantly enhanced multiphoton ionization (REMPI), utilizing time-of-flight mass spectrometry for detection. A (2-t 1) - REMPI excitation mechanism provides ‘access to electronic states in the region 64 641-71 352 cm- t . Two-photon selection rules allow the observation of transitions forbidden in the HI absorption experiments. Many of the assigned Rydberg states contain underlying, unassigned structure with anomalous intensity fluctuations and apparent breaking off of branches. Additionally, we-observe autoionized leveIs, most likely accessed by either a (3+ 1) or a direct four-photon excitation. Molecular constants are calculated for all HI bands observed. Also included is the first REMPI study of iodine -atoms coming from photolyzed HI precursor. ._

I. INTRODUCTION

The hydrogen halides are very important chemical spe- cies which have broad application in chemistry. They pro- vide a basis for the study of smoothly changing chemical and spectroscopic properties within a periodic group of stable molecules. Many disciplines in chemistry benefit from the unique characteristics of these molecuIes. Their relatively small size makes them a good prototype for theoretical mod- eling. The heavier members of this group are of particular interest to dynamicists because of their reactivity, due in part to their polarity and low bond strength.

The large mass of hydrogen iodide leads to significant spin-orbit interaction. Electronic structure reflects this fact by showing very strong tendencies toward (a,~) angular momentum coupling, also known as Hund’s case (c). The analogous hydrogen halides become progressively closer to pure Russell-Saunders coupling, or Hund’s case (a), with decreasing mass. The spectroscopist has the archetypical op- portunity to study the transition between case (a) and case (c) coupling and the resulting effects on electronic structure. In- teresting dynamical effects which result directly from spin- orbit interactions include predissociation and autoionization, each reflected by interesting spectroscopy.

Structured absorption spectra of hydrogen iodide, as well as the other hydrogen halides, were first reported by Price.’ Electronic structure calculations by Mulliken2*3 facilitated further understanding of the species. The discussion in these early studies focused on bonding character and angular mo- mentum coupling trends. Re-examination with absorption experiments by Ginter et a1.4-6 allowed a more detailed look at electronic structure and term series, as well as rotational assignment of the bands.

All of the preceding investigations of hydrogen iodide have been constrained by single-photon selection rules, as shown below for two of Hund’s coupling cases. [Rotational selection rules for case (a) are obtained by replacing each ti in rule 3 by A.]

(1) (Case (a): AR=O,+l, AS=O, C++++C-; (2) (Case (c): A$?,=O,+l, O’*O-; (3) Rotational, case (c): AJ=O,+l for fi’=fi”=O;

AJ=O,?l for R>O.

Multiphoton experiments allow a closer look at single- photon disallowed transitions, thus providing additional ex- perimental information about electronic structure. Multipho- ton selection rules may be arrived at by successive applications of single-photon selection rules. For example, for two-photon transitions between states represented by Hund’s case (c), selection rules are shown by rules 4 and 5:

(4) Case (c), two photons: Afi=O,+_1,?2, O+*O-; (5) Rotational, case (c), two photons: AJ=O,+2

for 92’=fl”=O, AJ=O,+1,+2, for fi>O. Selection rules for two-photon transitions are, therefore, re- laxed from those required for single-photon excitation. Ad- ditional rotational structure enables more thorough and accu- rate calculation of molecular constants.

Resonantly enhanced multiphoton ionization (REMPI) has been used to study the electronic spectroscopy of hydro- gen chloride’-l4 and hydrogen bromide.t5,i6 Such studies have lended to a better understanding of the Rydberg, va- lence, and superexcited states of the halogen acids. Further understanding of the hydrogen halides may be gained by applying REMPI spectroscopic probes to HI. Other than add- ing to the pool of general knowledge regarding hydrogen halides, HI should be an interesting molecule to study for reasons described above.

II. EXPERIMENT

Experiments are conducted in a double-differentially pumped vacuum system. The first stage, the source chamber, is evacuated by a 10 in. diffusion pump, backed by a 100 cfm liquid Ns trapped forepump. The second stage is also evacu- ated by a 10 in. diffusion pump, while the third stage, con- taining the interaction region, is pumped by a turbomolecular pump.

The initial chemical expansion occurs in the source chamber through a specially designed corrosion-resistant pulsed solenoid valve.i7 The design features isolation of the high-pressure gas from any metallic parts, provided by ce- ramic and fluoroplastic construction for those components in contact with the gas flow. This device enables reliable chemi- ca1 delivery of severe corrosives for months at a time with negligible maintenance. The expansion is skimmed at each

238 J. Chem. Phys. 101 (l), 1 July 1994 0021-9606/94/i 01(1)1238/8/$6.00 @ 1994 American Institute of Physics

S. A. Wright and J. D. McDonald: Multiphoton ionization of hydrogen iodide

chamber interface using conical nickel skimmers (orifice diameter=0.070 in.) before entering the interaction region in the third chamber.

The second stage has opposing laser ports and a 360” rotatable top flange. Mounted 180” apart on the rotatable fiange are the third vacuum stage and a photomultiplier. The third stage houses a Wiley-McLaurenr8 time-of-flight mass spectrometer (TOFMS) with a nominal flight distance of 1 m, capable of mass resolution of 1: 110 amu. The source of the TOFMS contains opposing laser ports (CaF, or A1203), as well as electron gun hardware. Positive ions are detected in all of the experiments of this study, although this mass spectrometer is capable of accelerating either positive or negative ions. In general, parent ions make up -95% of the total peak area detected. Where this rule is violated, it is discussed.

Ions are detected at the top of the TOFMS with a cas- caded pair of leaded glass microchannel plate electron mul- tipliers. The resulting signal is amplified and transmitted to a 2048 channel; transient digitizer (Biomation 8100). A per- sonal computer controls all experimental apparatus and handles digitized data input through a digital I/O board (DIO; Metrabyte). Outgoing TTL pulses trigger an electronic variable time delay at 12 Hz, sequentially firing the pulsed valve, the laser, and the mass spectrometer. The laser grating is stepped independently by the DIO.

Scans are initiated by collecting a “preliminary” mass spectrum, either by electron impact or photoionization. The transient signal from the TOFMS is displayed and updated, in real time, on an atomic mass scale by using the appropri- ate time-to-mass calibration. After collecting this initial mass spectrum, the user selects ranges of mass channels to monitor during spectral scanning. The ability to select multiple mass peaks is important in this study, as the lone appearance’-of either molecular or fragment ions, as well as the coincidence of both, are distinct spectral signatures (discussed in Sec. IV). At -130 amu, ion signal from one mass channel over- laps the neighborihg mass only when very large signals oc-

._ cur. Tunable light is supplied by a Nd:YAG (Quanta Ray

DCR-1) pumped dye laser (Lambda Physik FL 2002). The energy range studied was covered by the dye curves of rhodamine 590, rhodamine 610, rhodamine 640, and DCM (Exciton). The output of the dye laser is frequency doubled in KDP and the resulting ultraviolet radiation is focused into the interaction region inside the vacuum system. All spec- troscopy described was done with linearly polarized light. Laser bandwidth varies between 0.5 and 1.0 cm-*.

Laser energy ranges from 50 to 300 ,uJ, but remains constant across most spectra to less than 10%. Ion signal was found to be a nonlinear function of laser power, inferring that the process being studied is not facilitated by merely single- photon excitation. Exact functional dependence of signal in- tensity on laser power was not established due to laser power fluctuations, however this nonlinearity is not attributable to a threshold-type effect, as the total integrated peak area does not undergo any abrupt changes with power, and mass peak ratio (molecular ions: atomic ionsj remains constant. It

65950 66000 66050 66100 66150 66200 Two-photon energy (cd)

PIG. 1. Mass 128 REMPI spectrum of g ‘Z&-(0+), v’=O.

should be noted that the spectra presented represent unnur- malized spectra spanning several dye curves.

Neon atomic lines were used for wavelength calibration over the region being addressed, to allow reliable conversion of dye laser calibrated wavelength to actual wave numbers. There may be a systematic error in our absolute wave num- ber calibration over time. This would be seen as a shift in the band origins, but does not affect relative peak positions. All energies are given in vacuumcorrected wave numbers.

Commercially available hydrogen iodide (Matheson, 99.9%) was used without further purification. Dilution to 30% in helium provided ample molecular density for these experiments without adding significant rotational cooling. Some difficulty was experienced in reproducing beam tem- peratures due. to the lack of an adequate corrosion-resistant gas regulator.

Ill. RESULTS

REMPI spectra of hydrogen iodide, appearing in Figs. l-8, have been collected using the methods and- equipment described in Sec. II. Across the energy range of 64 641- 71 352 cm-‘, ten vibronic bands are observed. These repre-

I.,.,...*,,,,,.I,.,,,,.,,,...*,,,,, 67680 67700 67720 67740

Two-photon e"erm'(cm-')

FIG. 2. Mass 128 REMPI spectrum of g 32-(Of), v’=l.

J. Chem. Phys., Vol. 101, No. 1, 1 July 1994

240 S. A. Wright and J. D. McDonald: Multiphoton ionization of hydrogen iodide

8 40 -a : : : : 7 :

Jl I.....,,,.,,........,,.~,,..,.....-

69000 69050 69100 69150 69200 69250 69300 69: I....,.,..,.,....,.,l...,..,..l.,,r

66600 66650 66700 66750 Tim-photon energy (cm-‘) Two-photon energy (cm-‘)

FIG. 3. Mass 128 REMPI spectrum of d 3110(Ot). Asterisks represent un- assigned lines (see the text).

PIG. 6. Mass 128 RFMPI spectrum of previously unreported CL=2 state. Asterisks represent unassigned lines (see the text).

sent all the molecular spectra detected in the region. Nine of the spectra appear at mass 128 and will be the focus of this section. The remaining spectrum, at mass 127, is discussed in Sec. IV C.

Spectral line positions, reported in Tables I and II, are designated at the average wave number of the two half- intensity points of each line to avoid inconsistencies caused by asymmetric or diffuse line shapes. In order to determine v~,~B~, and DJ for each band, line data are fitted to FQ. (6) using a nonlinear least squares fitting routine. (6) E,,-E,~I=~~+B,IJ’(J’+~)-D,~J’~(~’+~)~. Subtracting the ground state energy from the fitting equation brings correctly assigned lines of like J’ (in different branches) to identical energy. Iterations of band simulation and fitting help fine-tune peak assignments. This method works quite well, especially on states with insufficient data to do a combination difference analysis. Molecular constants used for the ground state, X ‘x+(0+), are Bs=6.426 365 cm-’ and D~r=O.O002069 cm-‘.19 Lines which are ob- scured or of very weak intensity are omitted from the fit. Our

70625 70650 70875 70900 70925 Two-photon energy (cm-‘)

FIG. 4. Mass 128 REMPI spectrum of E ‘s’(O+). Asterisks represent un- assigned lines (see the textj.

SP 1 6 2 RI

,: 3 4 5

I

,......,*.,,,*,,,,,,~ 70150 70200 702! 50

Two-photon energy (cm-‘)

64650 64675 64700 64725 64750 Two-photon energy (cm-‘)

64775 6, 460(

FIG. 7. Mass 128 REMPI spectrum of F ‘A(CL=2). Single asterisks repre- sent unassigned HI lines. V marks autoionized HI at three-or-greater photon energy (see the text). FIG. 5- Mass 128 REMPI spectrum off 3A,(C&=2).

J. Chem. Phys., Vol. 101, No. 1, 1 July 1994

S. A. Wright and J. D. McDonald: Multiphoton ionization of hydrogen iodide 241

1...I.....I....,I..II,..,,,,,,I.,..,., 35500 35550 35600 35650

Single-photon energy (cm”)

PIG. 8. Mass 127 RBMPI spectra of an autoionized HI state at three-or- &reater photon energies (see the text). Lines narrower than 5 cm-’ (single- photon energy) are believed to be due to systematic noise rather than fine structure. The diffuse progression appears to be rotational structure.

TABLE I. Line positions, 0=0 states. w=weak, s=shoulder, o=obscured, b=blended; not rotationally resolved.

J” 0 Q S

g 3x-(o+), v’=O 0 1 2 3 4 5 6 7 8 9

10 11 12

g 9x0+), v’=l 0 1 2 3 4 5 6

d ‘rb+@+) 0 1 2 3 4 5 6 7 8

E %+(O+) 0 1 2 3 4 5

65985.2 65958.3 65931.3

67667.4(o)

69122.6 69096.2 69069.2 69041.8 69013.8

b b b b

66017.6 66014.3 66010.4 66005.9 66001.0 65995.1 65988.2 65981.2 65973.1

67706.2 67703.9 67701.8 67695.9 67688.2 67679.3 67667.4

69160.2 69159.2 69157.5 69154.8 69151.4 69148.0 69143.9 69138.8(b)

04 (b)

70853.7 70850.5 70844.1 70832.9

66059.8 66084.3 66107.0 66129.8 66151.1 66172.0 66191.8(w) 66212.8(w) 66231.1(w)

67740.4

69198.5 69222.2 69245.6 69268.0 69289.8 69310.8 69331.6

TABLE II. Line positions, C?DO states. w=weak, s=shoulder, o=obscured, b=blended; not rotationally resolved.

JN 0 P Q R s

f 3Aa(fi=2) 0 1 2 3 64655.2 4 5 6

Cl=2 0 1 2 3 66569.8 4 66555.6 5 6 7 8

F ‘A(fl=2) 0 1 2 3 70194.5 4 70143.6(w) 70181.3 5 70116.4(w) 70167.1 6 70153.5 7 70140.0

64731.4 64718.9 64759.5

64693.6 64733.6(s) 64787.9 64695.5 64749.2 64698.2 64765.3 64701.2 64781.3(w) 64704.3

66647.0 66634.4 66671.4

66608.6 66645.3(s) 66694.9 66607.2 66656.0 (66717.0 66604.7 66665.5 66738.6(b) 66600.9 66758.7(b) 66596.4(w) 66590.5(w) 66584.0(w)

(b) (b) @I (b) 04 iv

70270.9 70258.6 70296.0 70270.9(o) 70321.0 70283.0 70345.3 70294.6 70306.2 70317.6 70327.7(w)

8 70126.0(w) 9 70110.9(w)

04 70341.0(w)

fitted constants are compared to those obtained in absorption experiments,s,6 however it should be noted that slightly dif- ferent (and somewhat outdated) ground state constants were usedinthose fits (B~1=6.4275 cm-‘, Dg=O.O00205 cm-‘). Molecular constants and electronic assignments appear in Table III. Since HI more closely follows case (c) than case (a) coupling, fi notation is more valid than A, S notation. The latter is noted, as assigned by Ginter et al.,5*6 throughout the text and tables as a reference to term series ancestry. This nomenclature allows easy spectral comparison of HI with the other hydrogen halides.

The two-photon selection rules in F& (5) allow discrimi- nation of Cl=0 states by the absence of P and R branches. States with such characteristics are shown in Figs. l-4 and assigned as g 38-(Of); g 38-(0+), u’=l; d 3110(O+), and E 'x'(O+). In addition, k 3111,(Of) and H ‘x+(0+) are ob- served with no rotational structure (see Sec. IV A), hence their spectra are excluded. It should be noted that across the spectrum of Fig. 3, laser power varied by 25%, being weak- est on the lower wave number end.

All states lack lines of J<fi, as required by angular momentum conservation, thereby clearly marking states of CC-0 by their omission of low J’ lines. C4>0 states in the region are shown in Figs. 5-7 and are assigned as F ‘A(CI=2), f 3A2[C&=2), and a previously unreported state of fi=2. No Cl-type doubling was observed.

J. Chem. Phys., Vol. 101, No. 1, 1 July 1994

242 S. A. Wright and J. D. McDonald: Multiljhoton ionization of hydrogen iodide

TABLE III. Molecular states and constants. n notation is shown’ in parenthesis following the 11; S designation.

‘. ~-_ - Literature State v. (cm-‘) B; (cm-‘) D, (X104 cm-‘) t-e (A) constants’ Notes

x ‘Is+ lo+) . . . . . . . . . . . . . . . 1.60gb f- 3Az

0.0, 6.426 365, 2.069 (2) 64 691 6.80 1-0.03 17 26 1.580 64 693.9, 6.737, 10.6

g 3x- (0+), u’=O 66 023 6.118 to.010 2.9 Cl.2 1.666 66 022.6, 6.110, 2.5

$) , 66 610 6.17 kO.01 15 +3 1.659 . . . Extra lines

g 3X- v’=l 67 706 5.59 20.05 20 210 1.743 k3&

67 704.4, 5.62, 28. ifJO+) 68 113 1.650b 68 110.7, 6.24, 3.0

H ‘2’ co+) Single Q-head

68 279 1.714b 68 277.3, 5.78, 5.0 d 3&+ co+)

Single Q-head 69 161 6.116 LO.016 1 rt2 1.667 69 157.8, 6.117, 2.1 Extra lines

F ‘A 0.1 70 234 6.320 +0.015 1 ‘22 1.640 Extra lines E Ix+

70 228.3, 6.30. 1.26 io+j 70 856 6.212 kO.002 183.8 LO.6 1.654 70 850.5, 6.002, 128 Weak; only Q

branch; extra lines

‘Literature constants are in the same units as the fitted constants. Excited state constants taken from Ref. 6. Ground state constants are from Ref. 19. %. calculated from literature constants.

IV. DISCUSSION

A partial energy level diagram for hydrogen iodide is shown in Fig. 9. Excitations representing one, two, three, and four photons are shown to aid in discerning the mechanism responsible for the observed ionization spectra. The ioniza- tion potential for HI is 83 745.84 ~k”‘.‘~

Single-photon energies used in .this study are between 32 320 and 35 676 cm-‘. This spectral region features con- tinuous absorption starting at -28 000 cm-’ and peaking at -46 000 cm-1,1g*20 resulting from ekitation to the dissocia- tive $aJes_A Lt’I, a 311~~ ? and eW3111,. Single-p&ton e&ar

-tiG ~-thZ%re provides ph%todikociation &Oughout otir

IL : Distance (A) ‘_ : i. a* I! I

FIG. 9. Selected potential energy curves of. hydrogen iodide. Arrows indi- cate the range of multiphoton energies accessed in this study. (The left arrow fepksents the low energy end and the right-atrow displays the high energy end.) Constants for g ‘m-(0+) and V %‘(O’) are taken from,Ref. 19, the former being shown as a representative Rydberg state. B ‘E+, X 2113,2, and A *Z+ come from Ref. 28, *&, is assumed to have same shape as ‘II,,, shifted in energy by he value in Ref. 30. a(311~+ ,%J and A(‘rI) are from Ref. 2. 411 i$ &awn assuining the same shape as repulsive molecular states, but using re of the ground ionic state. . . . .I

. .? study. This is a very relevant point, discussed further in Sec. IV D, but does not account for the structured molecular spec- tra seen. Further, photon fluxes are sufficiently low (due to long laser pulses) to make successive, single-photon excita- tion ‘through this continuum to a higher excited (r&o&t) state a very unlikely process.

Two-photon energies lie between 64 641 and 71 352 cm-‘. This region has been explored by absorption experimen@ and is known to be rich in structured elec- tronic states. An additional photon is necessary to ionize the excited molecule. Three photons contain sufficient energy to directly ionize HI(96 961-107 028 cm-‘). The single-photon spectrum in this region is continuous’l from 95 000 to 104 WO’cm-‘. Additionally, any mass 128 discrete structure above the ionization potential can be discounted, as lines of bound ionic states will be broadened by autoionization; this is not the case in our observed molecular spectra. The ion- ization we see at mass 128, therefore, must be achieved by a (2+ 1) REMPI process.

A. &I=0 states

This category includes the valence state,= V ‘Z$+(O+), in addition to Rydberg states. V is bound, with a very large equilibrium internuclear distance, small vibrational spacing, and rotational constant abok half that of the Rydberg states. It is important to consider the valence state when deciphering anomalies in the Of Rydberg states, as all Of states in HI interact to some degree. Such mixing is prevalent among Rydberg states, as well as between states of Rydberg and valence character. Rydberg-valence interactions in the halo- gen acids are evidenced spectroscopi&lly by perturbations in intensity and line positions.6*‘3*‘4

Ginter et a1.5*6 studied all the vibrational levels in HI V ‘X+(0+), using absorption methods, from u ’ = m through u ’ = m + 14, excluding u ’ = m + 7, which may be obscured. From this data, they derive an empirical equation which fits vibrational band origins throughout V. By extrapolation of data from the other hydrogen halides, they also note that u ’ =m is probably not the lowest vibrational level of the V state. Shown in Table IV are calculated and reported vibra- tional band origins of V in the region of this study, repro-

J. Chem. Phys., Vol. 101, No. 1, 1 July’1994

TABLE IV. Observed and calculated band origins of HI V ‘x+(0+) vibra- tioeal levels, jp.the region of this study (from Ref. 6).

V’

m-3a 8, m -2 V*-la”

‘hi t?Z+l... mf2 ._ mf3 m-f4 m-f-5 m+6 : m+7n %. Ill + 8 71920

ralculated from the etipirical equation: v,(cm-‘)=68 000 + 505n - 1 On’ with u’=rn+n.

I.

duced for the convenience of this discussion. REMPI line intensities of the HI valence state may be extremely weak, if the trend set by the other hydrogen halides is followed. In HCI, the vibrational levels of V are observed as among the strongest multiphoton transitions in the spectrum.7g8 The cor- responding lines of HBr are, in general, relatively weak when detected by REMPI methods.‘5’16

Mass-selective ionization methods enable a whole new approach regarding studies of Rydberg:valence mixing. In a REMPI study of HCI, Spiglanin et aL9 report that E fZ’(O’) and LT 3S- O+) Rydberg states and the valence state ( i V ‘x+(0+) are marked by simultaneous detection of molecu- lar and atomic ions.’ Wallace et al. l4 report that the parent and fragment ion pe;-&!s are of comparable intensities in these excitations,.The mechanism for the process suggests a (2+ 1) REMPI excitation through V or states of correct symmetry to mix with V,. to a superexcited state (at three-photon ener- gies), in which-the processes of autoionization and predisso- ciation are in competition.13*‘4,U Three-photon energies are sufficient to produce Cl* and H(2S) in”predissociatibn, with an addition’al photon providing ionization of the excited chlorine atoms. Callaghan et al. mention similar effects in their REMPI study of HBr,‘” however they note that the molecular ions of the -V state constitute only -2 10% of the total signal. If a similar mechanism exists for HI, one would expect an even greater contribution from the dissociative mechanism23 in .the predissociationlautoionization competi- tion of the superexcited state due to -the greater spin-orbit coupling of HI. Thus, seeing ex&sively atomic ions would not be umeasonablei 2

None of the assigned O+ states in this ‘study display atomic ions. E and ‘H show strange effects, as described below, while the g state shows a fairly large fluctuation in rotational constant with vibrational quantum number. The two II states in this category, d and k, also exhibit unusual ,behaviors. Many of the mass 128 spectra presented contain lines which, at this time, remain unassigned. These show anomalous perturbations including what seems to be a break- ing off of bands and so-called f‘intensity borrowing” in the region of the assigned state. The state or states represented by these unassigned lines produce only molecular ions. If

these unassigned lines are constituents of the V ‘z’(O+) state, then in these cases HI is not showing a mechanism

vn (cm-‘) similar to that described above for HCI and HBr. .I

66,385 ii 66 950 67 485 68 004 68 489

I. d3tl,

68 927 69 419

.I -i In HI absorption experiments, the parity-forbidden 69 910 d 311a(O-) is observed as a diffuse Q branch.5 This occur- 70512 -’ rence is likened to that of b 3110(0-), which is seen as a 70 949 single “Q head. ‘76 Each transition is apparently allowed 71045 through rotational distortion, a J-dependent state mixing be-

tween levels of Ai2= ?I 1 .Z,25 In this case, the mixing occurs between 3110, s,H’;‘and 3112. _,_

Figure 3 contains a very intense mass 128 structure un- derlying d 3110(0-+) which, at the time of submission of this article, remains unassigned. It was first believed to be -me .I.. perturbed O- partner of d(O+). Assuming this assignment, the branch structure of d(O-) includes 0(2)-O(5) and S(O)- S(l), with a perfectly. normal breaking off at J’=3. This data yield molecular constants of v,,=69.150 ‘cm-“ and B,=6.3 cm-‘, 11 cm-‘.

making v, of-q-.redshifted from that of O+ by These values compare fairly well with Ginter’s

derived values of v,=69 149.5 cm-‘, B,=6.091 cm-‘, and a degeneracy splitting of 8 cm-‘. Discrepancy in the rotational constants might be attributable to a misnumbering of J val- ues in the. single-photon experiments, where the O- Q branch appeared very weak and diffuse, and ultimately transitions were numbered ~such that the rotational constants of the O+ and O_states were most nearly equal.

r There remain problems with this assignment, however. The 0 branch lines are much larger than those in the S branch. Perturbations which appear in one branch should be ..~ present-m the other branches, as well, at identical J’, assum- ing. a nonperturbed lower state. If the spectrum were normal- ized for laser power, the difference between 0 and S branch . . intensities, would be even more dramatic, as much lower power was used on the low wave number end of this scan. Accidental double resonance at two and three photon ener- gies could account for the selective enhancement of the 0 branch. The most serious .barrier to this assignment is that the progression has -eight members, equally spaced at 26 cm:‘, not merely the obvious six lines used in the fit for d(O-).

Another possibility of assignment is V(O+),. u ’ = m + 3 (see Table IV). As mentioned~earlier, we expect weak signals from two-photon V+-X transitions. Intensification could be attributed to Rydberg-valence mixing with the strong d(0’) state. Using, m.olecular constants reported by Ginter. et aL6 for=band simulation, the unassigned lines appear to be almost # . . ..~ assignable to alternating 0 and Q branches; however, the equal spacing between the lines suggests that the state in question has a similar rotational constant to that of the ground state. Since V(O+) has a rotational constant around 3 cm -’ compared to 6.4 cm-’ for X(0+), it follows that i/+-X rotational line spacings diverge relatively steeply with in- creasing J. Assignment of the intense unassigned progression in Fig. 3 to either V(O+ j, u ’ = m + 3, or d(O-) is ambiguous, at best.

S. A. Wright and J. D: McDonald: Multiphoton ionization of hydrogen iodide 243

J. Chem. Phys., Vol. 101, No. 1, 1 July I.994

244 S. A. Wright and J. D. McDonald: Multiphoton ionization of hydrogen iodide

2. E ‘X+(0+) Only a few weak lines are observed for this state, shown

in Fig. 4. J numbering was chosen such that the fitted rota- tional constant, B,, most closely matches that of ‘IIn ion core states. Our fitted band origin is somewhat shifted from that of Ginter et aL6 They calculate the band origin from only a few lines in the P and R branches; each of our J values could be off by one or two. An underlying state is evidenced by au additional very weak HI progression, but remains unassigned at this time.

3. H’S+(O+) and k311,(O+) These two states are grouped because of their similar

spectroscopic appearance and close proximity. We observe a single, sharp, symmetric Q head for each of these states, despite the difference in rotational constants relative to the ground state. No branch structure is present. Bandwidths of H(O+) and k(O+) are 3.5 and 2.5 cm-‘,FWHM, respectively. In Ginter’s absorption experiments,6 these states were ob- served as perturbed bands with “many irregularities in struc- ture and intensity” and showing “pronounced mixing of V(O+), k(O+), H(O+), etc., a feature also evident in the com- parable transitions of HCl, DCl, HBr, DBr, and DI.” k could very possibly be displaying rotational distortion, as described earlier for d 3110(Of) and b 3110(0’) states.

B. lb-0 states

Line positions of all the R=2 states fit very well to the rotational constants reported in Table III, and peak intensities follow the expected trend, suggesting unperturbed states. However, Fig. 6 contains au underlying electronic state, evi- denced by lines at small rotational spacing (7-8 cm-‘) in the S-branch region of the assigned state. The unassigned lines show no noticeable change in intensity with rotational tem- perature relative to the a=2 lines; however new, presumably higher J lines, appear under hot beam conditions. Once again, there seems to be a breaking off of bands in the unas- signed state. The region of F ‘A(fi=2), shown in Fig. 7, also displays several very weak mass 128 lines, too weak to es- tablish a periodic progression.

4 few additional comments are necessary for F ‘A&=2). Inspection of Table III reveals a shift in our fitted band origin from that reported in the preceding absorp- tion experiments. This state was recorded in those previous studies as being “very weak and partially obscured at low J values.” It is very possible that their incomplete data led to misassignment of lines by one or two J values. The medium- sized asymmetric peak between the Q and P branches is actually a very large signal at mass 127, large enough to overlap mass 128 mass window. Inspection of the 127 amu signal reveals a Fano profile,26 representative of autoioniza- tion of our molecule. This is discussed further in Sec. IV C.

C. Ionic states

In the same two-photon region as d(O+) at mass 128, there is an onset of an ionization channel at mass 127. The lowest laser power used in this study occurred where the I+ signal is very large, therefore this is not an effect caused by

excessive laser power. The signal appears as continuum that begins at a two-photon energy of -69 000 cm-‘, peaks around 69 250 cm-t, and transforms into extremely diffuse structure in the region of F(fi=2) around 70 250 cm-‘. At 70 930 cm-’ the diffuseness subsides somewhat to reveal a much more defined state structure, shown in Fig. 8. This complex mesh of spectral lines seems to contain diffuse, blended rotational structure, however rotational assignment has not been successful to this point. No mass 128 ions are detected in the region of Fig. 8.

The appearance of If at HI valence state resonances would correlate to REMPI studies of HCl and HBr, as de- scribed in Sec. Iv A. In fact, this two-photon region is pre- cisely where V(O+), u ’ = m + 7 is calculated (see Table IV; recall that u ’ = m + 7 is the only vibrational level of V which was not observed in HI absorption experiments). No other vibrational levels of V are observed in our study. If the state shown in Fig. 8 is V(O+), u’ =m+7, then its appearance in the absence of other vibrational levels of V may be revealing something important about the nature of the state mixing in V. However, question remains regarding the assignment. The appearance of atomic ions may represent a superexcited state contribution to the two-photon resonance, but broaden- ing of these lines must represent predissociation of the reso- nant state. This is a very important clue as to the type of state to which these lines belong. If this progression does repre- sent the V state, such a predissociation would happen through t, a or A, as these are the only dissociative curves in range to mix: none of these states contain I+ in their sepa- rated atom limit. Absorption of three more photons is neces- sary to produce I+.

States with such separated-atom limits occur only above the first HI ionization limit (see Fig. 9). Four photons allow access to 4 2Z$+, which is predissociated by “II to 1 H(2S)+I+(3P). Autoionizing Rydberg states below “II are also known to predissociate (to H+I*), as evidenced by the continuous absorption of HI, centered around 99 500 cm -‘.21.27,28 Three photons are required for the latter excita- tion, with au additional photon necessary to produce I+. [This (3 + n) ionization mechanism is similar to that pro- posed to explain the appearance of Cl+ ions in valence state HCl excitation at two-photon energies, as discussed earlier.]

The diffuse structure and the lone appearance of If sug- gest that this spectrum represents an excitation to a predis- sociative superexcited state at either three-photon (v,=106 845 cm-‘) or four-photon energy (~~~142 460 cm-‘). One cannot discount the fact that two, three, and four photon excitations may all play a role in this “band.” The separate contributions are very difficult to distinguish, given the nature of single-color photoionization spectroscopy. The broadness definitely arises from three and/or four photon ex- citation, as there is no continuum at the two-photon level. Sharp two photon resonances may be buried by the broad features of higher energy absorptions. The If continuum ob- served at lower wave numbers than the diffuse structure must represent some four-photon excitation. State assignments, along with the mechanisms responsible for these observa- tions, cannot be made more definite at this time. More stud-

J. Chem. Phys., Vol. 101, No. 1, 1 July 1994

S. A. Wright and J. D. McDonald: Multiphoton ionization of hydrogen iodide 245

ies are necessary, along with additional information about the superexcited state structure of HI.

D. Atomic absorptions

By simultaneously monitoring the 127 amu window one is able to perform a concurrent photodissociation experiment on HI, using REMPI to monitor iodine atom yield. The en- ergy range of this study is sufficient to provide single-photon absorption to dissociative states of HI, creating H(‘S) and I(2P3,2 or ‘P1,2) atoms [see Fig. 9). Two-photon energies are then sufficient to electronically excite the nascent iodine at- oms, with an additional photon providihg the photoioniza- tion, a (2fl) REMPI. Our laser pulses are sufficiently long (13 ns) to perform these one-color pump/probe experiments.

Such REMPI studies have previously been performed with HBr” and HC17, but never for HI. The photodissocia- tion of HI has been studied previously using two-photon laser-induced fluorescence on the I atom productF9 The HI fluorescence experiments used 248 mn light to photolyze prior to the probe. At that wavelength the population ratio, I(2Pllz)/I(2P3,2), is 0.4720.03. [Transitions from spin-orbit excited iodine (*P,,J are redshifted relative to the atomic lines by 7603 cm-‘.]

In our single-photon wavelength range (309.4-280.3 mn), the dissociative HI states corresponding to excited state iodine atoms are accessible, though with very low extinction coefficients. Two atomic iodine transitions are observed in this region: at two-pepton energies of 65 644 and 67 062 cm-‘. These lines are assigned to 2D51zczP,,2 and 4p3,2+-2p3/2, respectively. Although not all accessible ground-state iodine lines were observed, none of the transi- tions observed come from excited state iodine.

V. CONCLUSIONS

All the electronic spectra at mass 128 represent excita- tion of hydrogen iodide by a (2f 1) REMPI mechanism. We have identified only n=O and a=2 states. One of the a=2 states has never before been reported, another observed pre- viously only by a very weak forbidden transition. Through- out the assigned electronic spectra we observe underlying, unassigned states, all of which seem to display J-dependent perturbations.

The mass 127 spectra have provided a view of electronic states above the ionization limit, as well as a one-color pump/probe photodissociation experiment on HI. The former has revealed possible rotational structure in a superexcited state. Results from the latter suggest that only ground state atomic iodine (2P3/2) is produced by photodissociation at our single-photon energies.

ACKNOWLEDGMENTS

This work was funded by the National Science Founda- tion under Grant No. CHE91-07804. The authors wish to thank Matthew Verce for early work on this experiment, in- cluding the wavelength calibration, and the referee for help- ful comments.

‘W. Price, Proc. Roy. Sot., London Ser. A 167, 216 (1938). *R. S. Mulliken, Phys. Rev. 51, 310 (1937). ‘R. S. Mulliken, Phys. Rev. 61, 277 (1942). 4S. G. Tilford, M. L. Ginter, and A. M. Bass, J. Mol. Spectrosc. 34, 327

(1970). ‘M. L. Ginter, S. G. Tilford, and A. M. Bass, J. Mol. Spectrosc. 57, 271

(1975). 6D. S. Ginter. M. L. Ginter, and S. G. Tllford, I. Mol. Spectrosc. 92, 40

(1982). ,.,

7S. Aiepalli, N. Presser, D. C. Robie, and R. J. Gordon, Chem. Phys. Lett. 118, 88 (1985).

‘R. Callaghan, S. Arepalli, and R. J. Gordon, J. Chem. Phys. 86, 5273 (1987).

‘T. A. Spigl@n, D. W. Chandler, and D. H. Parker, Chem. Phys. Lett. 137, 414 (1987).

“D. S. Green, G. A. Bickel, and S. C. Wallace, J. Mol. Spectrosc. 150, 303 0991).

“D. S. Green, G. A. Bickel, and S. C. Wallace, J. MoLZSpectrosc. 150, 388 (1991).

12Y. Xie, P. T. A. Reilly, S. Chilukuri, and R. Gordon, J. Chem. Phys. 95, 854 (1991). -

13D. S. Green, G. A. Bickel, and S. C. Wallace, J. Mol. Spectros& 150, 354 (1991).

14D. S. Green and S. C. WalIace, J. Chem. Phys. 96,5857 (1992). “S Arepalli, N. Presser, D. C. Robie, and R J. Gordon, Chem. Phys. Lett.

li7, 64 (1985). 16R. Callaghan and R. J. Gordon, J. Chem. Phys. 93, 4624 (1990). 17S. A. Wright and J. D. McDonald, Rev. Sci. Instrum. 65, 265 (1994). ‘*W. C. WiLy and I. H. McLauren, Rev. Sci. Instmm. 26, 1150 (1955). 19K P Huber and G. Herzberg, Molecular Spectra and Molecular Structure, . .

Vol. IV, Constants of Diatomic Molecules (Van Nostrand Reinhold, New York, 1979).

‘OH. Okabe, Photochemistry of Small Molecules (Wiley, New York, 1978). ‘ID. T. Terwilliger and A. L. Smith, J. Chem. Phys. 63, 1008 (1975). “Two valencelstates exist in the hydrogen halides: t 3Z’(l,0-) and V

’ Ii + O+) t is dissociative and will be excluded from this discussion. 23H. Lefebire-Brian and F. Keller, 5. Chem. Phys. 90, 7176 (1989). %H. Lefebvre-Brion and R. W. Field, Perturbations in the Spectra of Di-

atomic Molecules (Academic, New York, 1986). “M. L. Gmter, J. Mol. Spectrosc. 20, 240 (1966). 2”U. Fano, Phys. Rev. 124, 1866 (1961). “7N. BGwering, H.-W. Klausing, M, Miiller, M. Salzmann, and U.

Heinzmann, Chem. Phys. Lett. 189, 467 (1992). 28H J . . Lempka T R. Passmore, and W. C. Price, Proc. R. Sot., London Ser. , ,

A 304, 53 (1968). “P. Brewer, P. Das, G. Ondrey, and R& Bersohn, J. Chem. Phys. 79, 720

i1983). 3oJ. H. D. Eland and J. Berkowitz, J. Chem. Phys. 67, 5034 (1977).

J. Chem. Phys., Vol. 101, No. 1, 1 July 1994

The Journal of Chemical Physics is copyrighted by the American Institute of Physics (AIP). Redistribution of

journal material is subject to the AIP online journal license and/or AIP copyright. For more information, see

http://ojps.aip.org/jcpo/jcpcr/jsp