J . Phys. Chem. 1985, 89, 1647-1653
1647
Multlphoton Ionization of NO-Rare Gas van der Waals Species John C. Miller* and Wu-Chieh Chengt Chemical Physics Section, Health and Safety Research Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831 (Received: November 5, 1984)
Multiphoton ionization spectra are presented for several NO-rare gas van der Waals molecules. Structured excitation spectra of the ArNO cluster associated with the B21T ( v = 9), C21T (v = 0, 1, 2, 3), and D22+ (v = 0) states of NO are observed as well as an unstructured, presumably dissociative, spectrum near the A22+ (v = 0) state. The C state was also observed for NeNO and KrNO. The series of clusters Ar,NO+ (1 d n d 7) was observed but yielded only structureless excitation spectra.
Introduction
The spectroscopic study of electronic states of van der Waals clusters of molecules and rare gas (RG) atoms’-3 has accelerated since the advent of the supersonic n ~ z z l e . ~In, ~particular, clusters involving combinations of diatomic molecules and rare gases offer the simplest systems for study and hold out the promise of a complete description of the spectroscopy, photophysics, and photochemistry of these species. The pioneering studies of Levy et al.s used laser-induced fluorescence (LIF) to study the RG-I, system, both establishing the techniques required and setting the standards for all such subsequent investigations. Surprisingly, however, it was several years following the first reports of He-I, in 1976 before the observation of the electronic spectra of any other such diatomic-RG cluster. In 1981 Langridge-Smith et a1.6 observed a dissociative state of ArNO, and Brinza et al.’ studied NeC12 in 1983. Both of these studies used L I F techniques. Very recently the multiphoton ionization (MPI) technique has been applied to the ArNO cluster by Sato, Achiba, and Kimuraa8 This work shows several of the advantages of the MPI technique relative to the LIF method. Specifically, mass analysis can provide identification of the new species formed as well as allowing discrimination against other species (the parent monomer or higher clusters) while recording spectra. Furthermore, the use of MPI photoelectron spectroscopy provides data on the states of the cluster ion as well. The disadvantages of MPI, however, include relatively low resolution as well as the line broadening and shifting mechanisms due to the use of high-powered lasers. The MPI technique has another advantage which is particularly important in the study of van der Waals molecules. This is the ability to study Rydberg states much more easily than LIF or other techniques. Because Rydberg states are highly polarizable, van der Waals forces are enhanced. Furthermore, the influence of the positively charged core allows the stronger dipole forces to become more important contributors to the binding energy. Thus, the binding energy of molecule-rare gas complexes should increase in the order ground state < valence state < Rydberg state < ionic ground state. Nitric oxide is the “alkali atom” of molecules, having one unpaired electron outside of filled molecular orbitals. Consequently, its spectroscopy is dominated by Rydberg transitions which have been very well c h a r a ~ t e r i z e d . ~ J ~ We report here, studies of the (2+1) and (2+2) MPI spectroscopy of RG-NO clusters which complement the work of Sato et aL8 and Langridge-Smith et aL6 and also probe several new states of the RG-NO complex. The series of cluster ions Ar,NO+ (1 d n d 7) was also observed. Experimental Section
The experimental arrangement is a modified version of that used previously for MPI mass analysis” and MPI photoelectron spectroscopy12 and is shown in Figure 1. Briefly, the pulsed supersonic nozzle (Quanta Ray Model PSV-2), mounted on an X,Y,Z translatable flange, is operated Department of Physics, Paine College, Augusta, GA.
0022-3654/85/2089-1647$01.50/0
as a free jet at 5-10 Hz. The pulse is approximately 60 ps long for H e expansions. The nozzle and gas reservoir can be heated to -100 OC or cooled to --lo OC. The latter is accomplished by passing liquid N2 through copper coils in thermal contact with the nozzle assembly. Typically, 5% RG-NO mixtures were expanded at backing pressures of 2-10 atm through either a 0.5or 0.75-mm hole. Nitric oxide-argon mixtures of 1% or 5% were obtained from Matheson. Other mixtures were prepared in a simple vacuum system from pure NO, He, Ne, Ar, Kr, and Xe components. Tunable light from an excimer-pumped dye laser (Lambda Physik EMGlO1, FL2000E) was focused with a 75- or 100-mm quartz lens to a spot -3 cm downstream from the nozzle. A master pulse generator provided start pulses to the nozzle at 1-10 Hz, and a second pulse generator with a variable delay triggered the laser 120 ps later. Ions, created by MPI, are extracted with 100 to 200 V/cm field and allowed to drift through a short (- 20 cm) time-of-flight (TOF) mass spectrometer with resolution ( M / A m ) of about 50. Ions are detected with a high current channeltron whose output is amplified, averaged with a boxcar integrator (Stanford Research Systems), and displayed on an oscilloscope. Energy analysis of electrons or positive ions is accomplished with a spherical sector electrostatic energy analyzer described previously.12Frequency doubling or triplingI3 is possible as are two-color experiments using a synchronously triggered N 2
- -
(1) B. L. BIaney and G. E. Ewing, Annu. Reu. Phys. Chem., 27, 553 (1976). (2) W. Klemperer, J . Mol. Srruct. 59, 161 (1980). (3) D. H. Levy in ‘Photoselective Chemistry”, J. Jortner, R. D. Levine, and S. A. Rice, Eds., Wiley, New York, 1981, p 323. (4) D. H.Levy, Annu. Rev. Phys. Chem., 31, 197 (1980). (5) R. E. Smalley, D. H.Levy, and L. Wharton, J. Chem. Phys., 64 3266 (1976); M. S. Kim, R. E. Smalley, L. Wharton, and D. H. Levy, J . Chem. Phys. 65, 1216 (1976); R. E. Smalley, L. Wharton, and D. H.Levy, Chem. Phys. Letr., 51,392 (1977); K. E. Johnson, L. Wharton, and D. H. Levy, J . Chem. Phys., 69,2719 (1978); G. Kubiak, P. S. H. Fitch, L.Wharton, and D. H.Levy, J . Chem. Phys., 68,4477 (1978); R. E. Smalley, L. Wharton, and D. H. Levy, J . Chem. Phys., 68,671 (1978); W. Sharfin, K. E. Johnson, L. Wharton, and D. H. Levy, J. Chem. Phys., 71, 1292 (1979); J. A. Brazy, B. M. DeKoven, T. D. Russell, and D. H. Levy, J . Chem. Phys., 72, 2439 (1980); J. E. Kenny, K. E. Johnson, W. Sharfin, and D. H.Levy, J . Chem. Phys., 72, 1109 (1980); J. E. Kenny, T. D. Russell, and D. H.Levy, J. Chem. Phys., 73,3607 (1980); K. E. Johnson, W. Sharfin, and D. H. Levy, J . Chem. Phys., 74, 163 (1981); K. E. Johnson and D. H. Levy, J. Chem. Phys., 74, 1506 (1981). (6fP. R: R. Langridge-Smith, E. Carrasquillo M., and D. H.Levy, J. Chem. Phys., 74, 6513 (1981). (7) D. E. Brinza, C. M. Western, D. D. Evard, F. Thommen, B. A. Swartz, and K. C. Janda, J . Phys. t h e m . , 88, 2004 (1984). (8) K.Sato, Y. Achiba, and K. Kimura, J. Chem. Phys., 81, 57 (1984). (9) E. Miescher and K. P. Huber In “International Review of Science: Physical Chemistry and Spectroscopy”, D. A. Ramsay, Ed., Ser. 2, Vol. 3, Buttersworth, London, 1976, pp 37-73. (IO) K. P. Huber and G. Herzberg, “Constants of Diatomic Molecules”, Van Nortrand-Reinhold. New York. 1979. (11) C. D. Cooper, A’. D. Williamson, J. C. Miller, and R. N. Compton, J . Chem. Phys., 73, 1527 (1980). (12) J. C. Miller and R. N. Compton, J . Chem. Phys., 75, 22 (1981); J . Chem. Phys., 75, 2020 (1981); Chem. Phys. Lett., 93, 453 (1982). (13) J. C. Miller, R. N. Compton, and C. D. Cooper, J . Chem. Phys., 76, 3967 (1982).
0 1985 American Chemical Society
-
1648 The Journal of Physical Chemistry, Vola89, No. 9, 1985
Miller and Cheng
GAS
0
1
r-l
2
1
3
4
5 6 7 Ar,NOt
+ TOF MASS SPECTROMETER
SPHERICAL SECTOR ENERGY ANALYZER
LASER BEAM
=-
y ' NeNO'
NO'
r NO' ArNO'
A = 383.8 nm
NO' l
l
l
r
l
l
r
5
l
I 20
TIME OF FLIGHT ( p s )
k = 382.4 nm
I
1 40
I
Figure 1. Schematic diagram of the apparatus used for the generation and multiphoton ionization detection of rare gas-NO van der Waals molecules.
0
0
l
l
40
TIME OF F L I G H T ( p s )
Figure 2. Time-of-flight mass spectra showing the multiphoton ionization detection of NeNO, ArNO, and KrNO van der Waals molecules. The figure is traced from photographs of the oscilloscope image.
laser pumped dye laser (Molectron UV-400, DL200). Results and Discussion Mass Analysis. Figure 2 shows the TOF mass spectra of the triatomic RG-NO van der Waals molecules observed in this study. The spectra are tracings of photographs taken of the oscilloscope display. The laser wavelength used to ionize is indicated on each. The mass scale is calibrated by using NO+, Ar', Kr+, and Xe+ ions produced by MPI at well-known resonances. With careful calibration the cluster ion is observed within f l mass unit of its actual mass. Alternately, all of the rare gas ions (including He+ and Xe2+)can be observed when the laser is tuned to the very strong (2+1) MPI resonance of N O via the C211(u=O) X21111z(u=O)transition. Presumably the ions result from laserinduced plasma formation in the high-pressure RG-NO mixture. At these wavelengths the NO' signal itself is severely saturated, appears distorted, and shows a shift and broadening in its TOF which is characteristic of space-charge effects. Further studies of this phenomenon are in progress.
-
Figure 3. Time-of-flightmass spectra showing the multiphoton ionization detection of larger RG-NO van der Waals molecules. The figure is traced from photographs of the oscilloscope image.
In all of the wavelength spectra to be discussed later the ion under study is monitored by setting the boxcar gate on the mass of interest. However, when the cluster ion signal is very small, compared to a very large NO+ ion signal at the same wavelength, the saturated NO+ signal can produce a distortion at the mass of the cluster which interferes with taking mass-resolved spectra. This distortion leads to a cluster ion signal whose wavelength dependence depends on both its intrinsic spectrum plus the corresponding spectrum of the uncomplexed NO. This effect can be mitigated by pulsing a voltage on an extra electrode between the end of the flight tube and the channeltron. When at ground potential, no ions are transmitted; when pulsed negative, the ions are transmitted and detected. Thus, very large NO+ signals can be discriminated against. Of course, the space charge effects of the large number of NO+ ions in the interaction region are not changed. The source of the NO+ peaks in Figure 2 depends on the wavelength. In Figure 2 the NeNO+ resonance overlaps the C211 - X2113,ztransition from warm background NO. In b and c of Figure 2 the bulk of the NO+ signal arises from the broad-band amplified stimulated emission (ASE) of the laser which is sufficiently intense to produce NO+ via three-photon ionization at the C state as well as from nonresonant MPI. In other wavelength regions, to be discussed in more detail later, the NO+ originates from the dissociation of neutral or ionic ArNO molecules. When the gas composition was reduced to 1% N O in Ar, the interaction region changed from about 50 to about 100 nozzle diameters from the orifice, and the laser tuned in the vicinity of the NOC211(u=O) state, the mass spectrum of Figure 3 is obtained. Although very weak, the series of cluster ions Ar,NO+ with 1 < n < 7 is clearly evident. In addition, the Ar+ and Arz+ ions are observed which require N O for their formation (Le., they are not observed in pure Ar). Unfortunately, these ions (with the exception of ArNO+), while clearly produced via the laser, exhibit no excitation spectrum other than the tuning curve of the dye. The possibility that they are created by electron impact ionization by electrons accelerated in the extraction field appears to be ruled out by the following arguments. First, the intensity of these ions does not vary with wavelength although the number of electrons produced by MPI of N O changes dramatically on tuning the laser. Second, retro-reflection of the laser beam onto the surface of the extraction plates in order to simulate scattered light effects produces no increase in the ion yield. The low intensity and structureless excitation spectrum are consistent with nonresonant MPI or MPI resonant with repulsive parts of the various cluster potentials. Similar nonresonant MPI of the ArNO molecules near the A state of N O is described below. Unfortunately, a spectrum such as that of Figure 3 contains no structural or spectroscopic information and serves only to demonstrate the production of these species in our expansion (true even if these species are produced by accelerated electrons).
Multiphoton Ionization of NO-Rate Gas Species
The Journal of Physical Chemistry, Vol. 89, No. 9, 1985 1649 V'
3
2
!
0
I
I
I
I
I-
+ U
I 383.5
I
383.0
-
I
I
384.0
384.5
WAVELENGTH (nml
Figure 5. Multiphoton ionization spectrum (2+1) of ArNO associated with the C211(u=O) X211112(u=O)transition of NO.
-
TABLE I: Transition Energies and Vibrational Assignments for ArNO Associated with the Czll X2111,zTransition of NO
laser transition energy, vibrational wavelength, nm cm-' in vacuum spacing, cm-' NO(u'=O) ArNO(u'=O)" ArNO(u'=l)" ArNO(u'=2)" ArNO(u'=3)"
381.77 384.22 383.79 383.41 383.07
52 373 52038 52 096 52 149 52 194
0 58 53 4s
NO(u'= 1 ) ArNO(u'=O)" ArNO(u'=l)" ArNO(u'=2)" ArNO(u'-3)"
365.59 367.73 367,.36 367.01 366.71
54 690 54 372 54 427 54 479 54 523
0 55 52 44
NO(u'-2) ArNO(u'=O)" ArNO(u'=l)' ArNO(u'=2)"
350.35 353.41 353.12 352.86
57 069
56 576 56 622 56 663
46 41
NO(u'=3) ArNO(u'=O)" ArNO(u'=I)' ArNO(u'=2)@ ArNO(v'=3)"
337.69 338.64 338.38 338.16 337.98
59 208 59 043 59 087 59 126 59 158
0 44 39 32
u'
refers to the (NO)-Ar stretching vibration.
TABLE II: Transition Energies and Vibrational Assignments for NeNO and KrNO Associated with the C211 X2111,2Transition of NO + -
laser transition energy, vibrational wavelength, nm cm-l in vacuum spacing, cm-I NeNO(u'=O)" NeNO(u'=l)"
382.39 382.14
52 287 52 321
0 34
KrNO(u'=O)" KrNO(u'=l)' KrNO(v'=2)' KrNO(u'=3)' KrNO(u'=4)'
384.75 384.36 383.99 383.64 383.31
51 967 52019 52 070 52 118 52 162
52 51 48 44
a u'
(14) H.H.W. Thuis, S.Stolte, J. Reuss, J. J. H.van den Biesen, and C. J. N. van den Meijdenberg, Chem. Phys., 52, 21 1 (1980).
0
refers to the (NO)-RG stretching vibration.
spectrum for the u = 0 is displayed in Figure 5 and the peak positions are listed in Table I. The observation of peaks closer to the dissociation limit is difficult due to the large NO+ signal near the NO C state resonance and the decreasing Franck-Condon factors. The spectra for the higher vibrational levels are similar to those for the (Q-0) transition and show no evidence of broadening due to vibrational predissociation. However, the line widths observed may be broadened by saturation or Stark effects due to the higher power of the laser, which would obscure small increases due to other decay modes.
1650 The Journal of Physical Chemistry, Vol. 89, No. 9, 1985
-
TABLE III: Transition Energies and Vibrational Assignments for ArNO Associated with the w2 XzII,,2 Transition of NO
VI
0
i
Miller and Cheng
1 N ..
1 -Ne
h
C STATE
0
NO(u'=O) ArNO(u'=O)',b ArNO(u'=l)' ArNO(v'=2)' ArNO(v'=3)' ArNO(u'=4)" ArNO(u'=5)" ArNO(u'=6)" ArNO(u'=7)' ArNO(u'=8)' ArNO(u'=9)' ArNO(u'= 10)' ArNO(u'=ll)"
laser wavelength, nm 375.17 381.21 380.75 380.29 379.82 379.36 378.91 378.49 378.05 377.67 377.27 376.89 376.50
transition energy, vibrational cm-I in vacuum spacing, cm-' 53 294 52450 52513 52 577 52 641 52 705 52 768 52 826 52 888 52 942 52 997 53051 53 106
0
63 64 64 64 63 58 62 54 55 54 55
"u' refers to the (NO)-Ar stretching vibration. *Assumed origin (see text).
382.5
381.5 WAVELENGTH (nm)
-
-
Figure 6. Multiphoton ionization spectrum (2+1) of NeNO associated with the C211(u=O) X2111i2(u=0)transition of NO. " I
4
383
3
2
4
0
384
-
385
WAVELENGTH (nrn)
Figure 7. Multiphoton ionization spectrum (2+1) of KrNO associated X211,i2(u=O) transition of NO.
with the C211(u=O)
We were also able to observe the C-X transition for the NeNO and KrNO complexes, which are shown in Figures 6 and 7 and summarized in Table 11. The NeNO+ signal intensity is the same order of magnitude as the ArNO+ intensity relative to that of NO'. The KrNO' signal is, however, much weaker and barely observable. Attempts to synthesize HeNO and XeNO were unsuccessful. The weakness of the KrNO spectra and the nonobservation of XeNO are probably due to the exchange reaction RG-NO + RG N O + RG-RG. The Xe, dimer is more strongly
-
bound than XeNO and hence will displace any Xe initially bound to NO. The Kr2 dimer bonding is comparable to that estimated for KrNO. When either Kr or Xe is expanded with NO/Ar the presence of more than 5% or so of the heavier RG reduces the ArNO' signal observed. Several trends are apparent in the data of Tables I and 11. First, the shift of the v = 0 band of the RG-NO van der Waals molecules relative to that of N O is 90, 338, and 409 cm-', respectively, for NeNO, ArNO, and KrNO. This shift mirrors the difference between the ground-state and excited-state potential well depths. For ArNO where the ground-state well depth can be estimated as DO))N 120 cm-', the shift of 338 cm-I then establishes the excited-state potential well depth as D,,' E 458 cm-I. The different shifts for NeNO and KrNO indicate that the difference between D,' and DC is correspondingly less and greater relative to that of ArNO. The shifts, increasing with the increasing size, and hence polarizability of the rare gases, reflect increasing well depths for both the ground and excited states. Similar trends are observed for the rare gas dimers and H2-RG complexes whose constants are conveniently tabulated by Klemperer2 as well as for the RG-I, complexes studied by Levy et aL5 The same trend of increasing difference between the ground- and excited-state bonding should also lead to a similar trend in internuclear distances which would be reflected by Franck-Condon distributions being peaked at higher v'for the larger rare gas complexes. As relative intensities in our spectra are not reliable due to interference from the large wavelength dependent excess of NO+ ions this trend is difficult to observe definitely. Finally, increased binding for the heavier rare gas complexes lead to larger force constants for the stretching vibration observed here. This would be observed in the wo values listed in Tables I and I1 once the differing masses are taken into account. If the comparatively rigid N O is taken as a single entity and the RG-NO complex considered a diatomic molecule then ~ . force constants are then the force constant K = ( 2 7 r ~ , ) ~The 3.39 X lo-,, and 3.51 X los2m d y n l k for NeNO, 0.821 X ArNO, and KrNO, respectively. The force constant has increased in going from Ne to Ar to Kr in accord with the increased bonding reflected in the shifts as discussed above. Of course, access to all of the relevant constants Dl,D,",w,), and w,)' are necessary for detailed comparison. These trends in potential well depth and force constant should be typical of those for all RG-molecules. If N O is considered to be physically about the same size and polarizability as Kr (based on similar Do values for ArNO and A r b ) then the bond strengths and trends observed for the series HeKr, NeKr, ArKr, Kr,, and XeKr should be relevant. The ground-state De values for these RG-Kr complexes are 17.2, 51.8, 113.9, 138.0, and 159 cm-', respectively, and the vibrational force constants are 0.09 X 0.65 X 1.16 X lo-,, 1.31 X and 1.46 X mdyn/A, respectively.* Remembering that the force constants for the RG-NO molecules are for the more strongly bound excited state, we observe that both the magnitudes and trends for the RG-NO
Multiphoton Ionization of NO-Rate Gas Species and RG-Kr complexes are quite similar. Similar trends are also observed for the RG-H2 compounds.2 The spectra of ArNO assigned to higher vibrational levels of the uncomplexed N O C state show several irregularities, however. For the u' = 2 and u' = 3 states, Table I shows that the van der Waals molecules have lower frequency vibrations (46 and 44 cm-I, respectively) and erratic shifts relative to the uncomplexed N O (493 and 165 cm-', respectively). For u'= 0 and u' = 1 spectra, the frequencies are 58 and 55 cm-I and the shifts are 335 and 318 cm-I, respectively. These differences are attributed to well known homogeneous perturbations between the C211 and B211 state^.^ This Rydberg valence mixing is largest for the C211(vr=3)-BZII(v'=15) states and results in roughly a 1:l mixture of the two states. Such mixing would be expected to result in unusual parameters for the cluster transitions associated with the states. Several other aspects of the C-state spectra deserve comment. First, for the A r N O spectrum s_hownin Figure 5 and in Figure 2 of ref 8, additional structure is observed. Sat0 et aL8 observe each peak split into a doublet which is only barely resolved in our spectra. In addition, a series of smaller peaks are observed between the major peaks. The positions and relative intensities of these peaks are unaffected by heating the nozzle assembly and gas reservoir to 100 O C or cooling it to -10 OC. Similar structure is observed for NeNO and KrNO but those spectra are of overall poorer quality. It seems reasonable to assign the doublet splitting of each peak to partially resolved rotational branch heads. The other structure is somewhat puzzling. In spite of careful searching, no bands were observed to the red of the origin for ArNO. This would seem to argue against an analysis based on hot bands as does the inability to change their relative intensity on heating or cooling the gas prior to the expansion. A ground-state frequency of about 40 cm-l would, however, reasonably account for the extra bands if it is assumed that the 0 1 band is not observed. The sharp bands to the blue of each major peak would then represent the 1 1,2 1, and 3 1 transitions. The other weaker peaks 2 and u' 3 series but these spacings might then be the u' are less regular. For KrNO an additional peak is observed 66 cm-l to the red of the presumed origin band and side bands to the red of each major peak are appropriately spaced for the u' 1 series. However, a ground-state frequency of 66 cm-l would be greater than that of the excited state which is 52 cm-l. This seems to be contrary to the evidence of the overall red shift of the electronic transition from that of N O which implies that the ground state is less strongly bound than that of the excited state. Alternately, the van der Waals bending vibration, where the RG atom bends against the rigid N O bond axis, could appear in the spectrum built on each stretching mode and would be of somewhat lower frequency. However, the peaks do not seem regular enough for this interpretation. Another possibility would involve a metastable conformation different than the T s t r ~ c t u r e ' ~ of the majority of the molecules. A final possibility is that the ArNO' ions giving rise to the weaker peaks originate from higher clusters which are known to be present in the expansion (see Figure 3). If so, since the smaller peaks are not appreciably broader than the major peaks, they would represent bound states in the larger clusters and the cluster ion peak itself should appear in the mass spectrum, although possibly weakly. Furthermore, the extra structure would, in principal, have a different dependence on backing pressure or gas composition than that of the larger peaks. In fact, the relative intensities are independent of all experimental conditions, thus arguing that all of the structure is due to the same cluster species. Better resolved spectra are required to determine the origin of these satellite peaks although the hot band interpretation appears most reasonable. A final point concerns the observation of metastable ArNO+ ions following (2+1) MPI via the C state. A cluster of ions are observed at longer TOF than the normal NO+ position but with an ill-defined TOF. They are also observed in the spectra of Sat0 et aL8 These NO+ ions are probably produced from metastable
-
- - -- -
-
(15) B. J. Howard, C. M. Western, and P. D. Mills, Faraday Discuss., Chem. SOC.,73, 129 (1982).
The Journal of Physical Chemistry, Vol. 89, No. 9, 1985 1651
,N
i! 1-Ar 5 0
D STATE
E
e
c
-g >
t cn
P3
bz h
376
377
378
379
380
384
WAVELENGTH (nrn)
-
Figure 8. Multiphoton ionization spectrum (2+ 1) of ArNO associated with the D211(u=O) XzII,,2(u=O) transition of NO.
ArNO' ions. The excitation spectrum of these ions is identical with that of the cluster, and a smeared out TOF is characteristic of metastable decomposition of the ion in the region between the focal spot of the laser and the acceleration grids. The spread out, longer T O F reflects decomposition at various positions relative to the extraction grids and hence varying amounts of kinetic energy acquired from the field. As the ions are born as ArNO' their TOF is always longer than that of ions born as NO'. In contrast, neutral dissociation of the cluster followed by absorption of an additional photon by the N O fragment would produce NO' at its usual TOF as the NO' formation would have to take place within the laser focus. The possibility of absorption of additional photons by NOAr' to metastable states is ruled out by the lack of low-lying states of NO'. The lifetime of the metastable ion must be on the order of 1 ps. An alternate explanation could be collisional dissociation as the ions are extracted from the focal region, but this is considered unlikely due to the low number density at this distance from the nozzle. The origin of the metastability is somewhat puzzling. It is most likely due to vibrational predissociation of vibrationally excited ArNO' where the N O vibrational mode is greater than u = 0. The energy is such a mode (a:' = 2376.4 cm-' for free N 0 + ) l o is much greater than the van der Waals binding energy and, consequently, dissociation occurs following energy migration to the weak bond. Such processes have been well characterized in the studies of Levy et aL5 However, in this case, MPI photoelectron studies of both free NO12 and of A r N 0 8 indicate that both ions are formed with predominantly u = 0 for the NO+ vibrational mode due to Franck-Condon factors between the C state potential and that of the virtually identical positive ion. The DzZ+State. Between the D2Zf(u=O) and C211(u=O) transitions of N O a weak excitation spectrum due to ArNO is observed. This spectrum, shown in Figure 8, is associated with the D2Z+(v=O) X2111j2(u=O)transition of the van der Waals molecule. The peak positions are tabulated in Table 111. The spectrum is limited at both its long and short wavelength ends by interference from the very strong NO+ produced at the C and D state, respectively. N o vibrational numbering is given in the figure as further peaks on the low energy side of the spectrum may be obscured. If the first observed peak is the (0-0) band a shift of 844 cm-' from the D state of N O is calculated. With a ground-state Doof 120 cm-l the dissociation limit for the D state is at least 724 cm-l, nearly twice as large as for the C state. The observation of the 12 peaks spaced approximately 60 cm-' requires a well depth of at least 654 cm-I. The increased binding in the D state relative to the C state is also evidenced by the long Franck-Condon progression. No ionization was observed in this spectral region for any other of the RG-NO complexes. The lack of anharmonicity in the data of Figure 8 and Table I11 is somewhat puzzling and casts some doubt on the assignment, however. Weak van der Waals bonding usually leads to quite anharmonic potential
-
-
Miller and Cheng
1652 The Journal of Physical Chemistry, Vol. 89, No. 9, 1985
I
I
1
I 310
377
-
I
I
I
319
WAVELENGTH (nm)
Figure 9. Multiphoton ionization spectrum (2+1) of ArNO associated with the D2n(u=0) XzIIli2(u=O) transition of NO in the spectral region of Figure 8.
wells. No other electronic states of N O are reasonable candidates, however. The valence B211state has vibrational levels in this region but as discussed below appears to be much more weakly bound. The next highest Rydberg state, the E2Z+, is considerably higher in energy. Assignment to the D state produces a reasonable dissociation energy, however, and remains the most promising assignment. The reason for the weakness of this band system is apparent upon scanning the NO+ ion signal in this same region. As shown in Figure 9, a number of peaks are observed between the very intense D and C state spectra. These peaks correspond exactly to the peaks for ArNO given in Table 111. Furthermore, spectra of N O / H e mixtures show no such peaks. The ArNO molecule is clearly predissociating into neutral N O in the C state which is subsequently ionized with one additional photon. The steadily increasing bandwidths of Figure 8 show the dissociation rate is increasing with excess energy over that required to dissociate. If the ground-state potential well depth is taken to be -120 cm-l then the threshold for dissociation to the C state lies between the first and second peak of Figure 8. The first peak, which cannot predissociate, is quite narrow. Unfortunately, its true width is distorted by the large NO+ signal in the region. The width of the peaks increases by about a factor of 4 between the first and last peaks observed. The B211 State. All of the states discussed so far have been Rydberg states consistent with the earlier observation that MPI is particularly sensitive to these states. The B2111,2,3/2 states are the lowest valence states of N O which are optically connected to the ground state.9 Because the valence states of N O cannot be ionized in a one-electron step they appear only very weakly in MPI spectra. Their presence is felt, however, in the ubiquitous perturbations of many Rydberg states when there is accidental near d e g e n e r a ~ y . ~Very , ~ ~ recently, Achiba, Sato, and Kimura17 have published detailed MPI spectra of the u = 9 level of the B state and have probed the ionization mechanism through MPI photoelectron spectroscopy. In connection with this study, however, this level of the B state, which is unusually well separated from nearby Rydberg states provides an opportunity to compare the van der Waals bonding in a valence excited RG-NO complex to that of the Rydberg states previously discussed. The excitation spectrum of mass-resolved ArNO+ is shown in Figure 10. Also shown is the spectrum for NO+ which is similar to that of Figure 4 of Achiba et al." The NO+ spectrum is also reflected as dips (16) M.G. White, M. Seaver, W. A. Chupka, and S. D. Colson, Phys Reu. Lett., 49, 28 (1982). (17) Y . Achiba, K. Sato, and K. Kimura, J . Chern. Phys., in press.
+
0
z
rl
+z
360.5
369.0
369.5
WAVELENGTH (nm)
310.0
-
Figure 10. Multiphoton ionization spectrum (2+1) of (a) ArNO, (b) NO, and (c) N atoms in the region of B211~/2,,/2(U=9) X * ~ I / Z ( U = ~ ) transition of NO.
in the ArNO' spectrum due to the large disparity in the relative numbers of each ion. Two peaks of ArNO' appear to the red of the B2nl/2(V=9) state and one peak near the B2&/2(U=9) state. For the B2IIIl2state the shift of the (0-0) band from that of uncomplexed N O is 43 cm-'. When coupled with the ground-state well depth of 120 cm-l, one arrives at an excited state Dovalue of 163 cm-I. Thus, the valence state B2111/2is only sightly more strongly bound than the ground state and considerably less strongly bound than the C and D Rydberg excited states. The vibrational spacing of 30 cm-l also reflects weaker bonding than that for the Rydberg states. Interestingly, the mass-resolved T O F spectrum shows the production of Nf ions in fairly large quantities only at 369.03 nm, the origin of the B2IIIl2state of uncomplexed N O (Figure 10). The N+ cannot arise from the predissociation of NO+ as the lowest N+ + 0 asymptotic limit lies almost 11 eV above the NO+ ground state. Clearly, to produce N+, at least four more photons must be absorbed. The sharp, single peak in the N+ excitation spectrum suggests that a superexcited state of N O is predissociating to give N 0 which is then subsequently ionized by resonantly enhanced MPI of the N atom. Such hybrid resonances have been observed
+
The Journal of Physical Chemistry, Vol. 89, No. 9, 1985 1653
Multiphoton Ionization of NO-Rate Gas Species
-
TABLE I V Transition Energies and Vibrational Assignments for v=9) XzII1/zTransition of ArNO Associated with the Bzl13/z,1/2(
NO
laser transition energy, vibrational wavelength, nm cm-l in vacuum spacing, cm-l B2111/2(~=9)
i 1
ArNO(u'=O)' ArNO(u'-1)' NO
369.32 369.12
54 138 54 168
369.03
54 181
368.83
54 210
368.72
54 226
0 30
B2npp(U=9)
ArNO(u'-0)" ArNO(u'=l)" NO 'u'
refers to the (NO)-Ar stretching vibration.
TABLE V Summary of Spectroscopic Constants for the Known States of ArNO and ArNO'
To,cm-l NOAr X211112(u=0) NOAr A2Zt(u=O) NOAr B2IIIi2(u=9) NOAr C2111,2(u=0) NOAr D2Zt(u=O) NOAr XIZ+(u=O)
I
wg,
cm-l
0
(40)'
e
e
54 138b 52 038b 52450b 73 788"
30 58 63
Do,cm-l 120 e 163c 458c 964c 1040d
0.1
'Assuming a hot band assignment is correct (see text). bFirst observed band assumed to be the origin. cBased on a value of 120 cm-I for the ground state. dFrom ref 8. weakly bound or entirely repulsive.
I
I
I
3.65 (NOI-Ar INTERNUCLEAR DISTANCE (41
Figure 11. Schematic representation of the potential energy curves for
ArNO constructed as described in the text. in many diatomic and polyatomic molecules when the laser is simultaneously resonant with a dissociative transition of the molecule and with an atomic transition. A search of Charlotte Moore's tables for atomic nitrogen reveals that the 5s 2P, J = 3/2 2p3 2P0,J = 'I2 transition, which is three-photon allowed, would occur at exactly 369.03 nm. Furthermore, the 2P0,J = 'I2state is metastable relative to the 2p3 9,J = 3/2 ground state and the 2p3 2Do,J = 5/2, 3/2 excited state due to parity and spin selection rules. This unusually complex hybrid resonance thus requires a simultaneous triple resonance. The laser is resonant with the B state of NO, some superexcited dissociative or predissociative state of NO, and the (3+1) MPI transition of metastable excited atoms. The observation of such an overall seven-photon ionization is due to the very efficient (2+1) MPI of NO coupled with production of metastable excited nitrogen atoms which can act as an energy trap as they cannot decay to any lower states. The (3+1) MPI of 2Pnitrogen atoms, although weak, is then easily observed in mass-resolved spectra. The observation of this process also leads to an upper limit for the dissociation energy of ground-state NO. The threshold for production of 2P0,J = nitrogen is the Doof NO plus the energy of the 2P0, J = 'I2 4S0, J = 3 / 2 transition (28839.3 cm-'). Hence, the observation of Nt a t 369.03 nm sets an upper limit of 52428 cm-' for the Doof NO. This upper limit compares well with other limits set by recombination fluorescence (52400 f 10 cm-'),'* analysis of N O lasing transitions (52 410-52434 cm-'),I9 and the two-photon fluorescence excitation spectrum (52 40752411.2 cm-1).20
-
+
(18) T. W. Dingle, P. A. Freedman, B. Gelernt, W. J. Jones, and I. W. M. Smith, Chem. Phys., 8, 171 (1975). (19) E. Miescher, J . Mol. Spectrosc., 53, 302 (1974). (20) P.A. Freedman, Can. J . Phys., 55, 1387 (1977).
The nature of the superexcited state responsible for producing 2P nitrogen remains unknown. The B' and I states, postulated by Achiba et al." as responsible for ionization of the B state, correlate with 2Donitrogen.
Summary To date, five electronic states of ArNO are known as well as the ground state of the ion. Figure 11 represents a schematic summary of these data. The figure represents a slice through the four-dimensional potential surface along the (NO)-Ar bond stretch coordinate. The curves are meant to represent, to scale, the relative potential well depths of all of the states determined as described previously. The ground state of A r N O is "T" shaped with an internuclear distance of 3.65 and an angle of 95O to the N O axis as determined by molecular beam electric resonance R F and microwave ~pectroscopy.'~The other potential curves are situated in the figure at unspecified internuclear distances such that the Franck-Condon factors are qualitatively correct. The well depth of the B(v=9) state is assumed to be similar to that of the v=O level which is shown in the figure. The well depth of the ion is taken from Sat0 et ala8and its internuclear distance assumed to be slightly shorter than the D state. The number and spacing of the observed vibrational levels are shown approximately for each electronic state. Table V summarizes the spectroscopic constants deduced as described in previous sections for ArNO and ArNO+. The corresponding curves for N e N O and KrNO will be quantitatively the same, with well depths reduced or increased proportionately for each state. Acknowledgment. Research is sponsored by the Office of Health and Environmental Research, U S . Department of Energy under contract DE-AC05-840R21400 with Martin Marietta Energy Systems, Inc. Wu C. Cheng acknowledges the Oak Ridge Associated Universities for support during his summer stay at ORNL. The authors also acknowledge many fruitful discussions with Cornelius E. Klots and R. N . Compton. Registry No. NO, 10102-43-9; Ne, 7440-01-9; Kr, 7439-90-9; Ar, 7440-37- 1.
