An experimental investigation of the nf Rydberg states of carbon disulfide
J.-P. Berger, S. Couris, and D. Gauyacq
Citation: J. Chem. Phys. 107, 8866 (1997); doi: 10.1063/1.475178
View online: http://dx.doi.org/10.1063/1.475178
View Table of Contents: http://jcp.aip.org/resource/1/JCPSA6/v107/i21
Published by the American Institute of Physics.
Additional information on J. Chem. Phys.
Journal Homepage: http://jcp.aip.org/
Journal Information: http://jcp.aip.org/about/about_the_journal
Top downloads: http://jcp.aip.org/features/most_downloaded
Information for Authors: http://jcp.aip.org/authors
Downloaded 03 Oct 2012 to 150.140.158.220. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/about/rights_and_permissions
An experimental investigation of the nf Rydberg states of carbon disulfide
J.-P. Berger and S. Courisa)
Foundation for Research and Technology-Hellas (FO.R.T.H.), Institute of Electronic Structure and Laser
(IESL), P.O. Box 1527, 71110 Heraklion, Crete, Greece
D. Gauyacq
Laboratoire de Photophysique Moléculaire du C.N.R.S and Institut de Physico-Chimie Moléculaire,
Bãt. 210, Université de Paris-Sud, 91405 Orsay Cedex, France
~Received 3 July 1997; accepted 22 August 1997!
The ( 2 P g3/2,1/2)n f Rydberg states of CS2 are investigated by means of (311) resonance enhanced
multiphoton ionization ~REMPI! time-of-flight ~TOF! spectroscopy. The excitation spectrum of
jet-cooled carbon disulfide has been obtained in the 74 000– 81 000 cm21 energy region. From the
CS1
2 mass selected REMPI spectra, the n f Rydberg series have been clearly identified and they have
2
been found converging to the two spin–orbit components of the CS1
2 (X P g ) ground state
1
1
1
corresponding to the n54 – 11 members. Comparison of the CS2 , CS , S mass selected REMPI
spectra gave a better insight of the competition between dissociation and ionization processes. The
use of linearly and circularly polarized laser light, selection rules and quantum defect considerations
have led to a preliminary analysis of the measured n f complexes. © 1997 American Institute of
Physics. @S0021-9606~97!03744-6#
I. INTRODUCTION
The spectroscopy and the dynamics of the excited states
of carbon disulfide have since long time attracted the interest
of experimentalists and have been the object of several studies over the last 60 years. Optical absorption spectroscopy,1–6 electron impact energy loss spectroscopy,7–9 magnetic circular dichroism,10 and multiphoton ionization
spectroscopy11–15 have been used in order to understand the
very interesting and rich structure in the UV-VUV energy
region and unravel the molecular structure and dynamics.
CS2 being a typical example of a 16-valence electrons triatomic has a linear ground state geometry with an electronic
configuration ...(5 s u ) 2 (2 p u ) 4 (2 p g ) 4 ; X 1 S 1
g , where the
highest occupied molecular orbital 2 p g is essentially of nonbonding character. One electron removal from this orbital
generates the (...2p g ) 3 X 2 P g cationic state, which is inverted with a spin–orbit splitting of 440 cm21. Recently
resonance enhanced multiphoton ionization ~REMPI!
schemes have permitted the identification of different ns,
n p, nd, and n f Rydberg states. In the two photon resonant
spectra,11,12 transitions to ns and nd gerade Rydberg states
are allowed and indeed have been observed, while in the
three photon resonant spectra13–15 the np and n f ungerade
Rydberg states are expected to appear and have been also
observed.
The 54 000– 73 000 cm21 energy region of CS2 has been
recently investigated in a series of papers11–15 using various
multiphoton ionization schemes @e.g., (211), (111 8 )11
and (311) REMPI-TOF#. Most of the transitions observed
have been assigned to the following Rydberg states:
1,3
1,3
1,3
4s s g ( 1,3P g ), 4pl u ( 1 S 1
u , P u , D u ), 3d, 5s s g ( P g ).
15
More recently, Morgan et al. have used REMPI-PES to
further investigate the 56 000– 81 000 cm21 energy region.
a!
Electronic mail: [email protected]
8866
J. Chem. Phys. 107 (21), 1 December 1997
Aided by the photoelectron spectra that they have got, they
have essentially confirmed our previous results and they
have successfully located the 5p Rydberg states. Above
74 000 cm21 their (311) REMPI spectrum of CS2 showed a
dense progression of sharp bands which have been attributed
to the @ 2 P g,1/2,3/2 # n f Rydberg states series forming two well
defined series converging to the two spin-orbit components
of the ionic core, respectively, in analogy with what was
found previously for the n f Rydberg states of CO2. 16,17 No
more detailed analysis was proposed for the very complex
structure of the excitation spectra. In fact Morgan et al. have
recorded the (311) REMPI spectrum of a near room temperature sample of CS2 monitoring the total number of photoelectrons emitted in the interaction region of the magnetic
bottle spectrometer. However in a three photon excitation
process up to 14 components for each n f member can be
observed according to the 3-photon selection rules, namely,
1,3
( 2 P g ) f s u 1,3P u , ( 2 P g ) f p 1,3S 1
D u , ( 2 P g ) f d 1,3P u ,
u ,
1,3
1,3
1,3
2
F u and ( P g ) f w D u , G u . Furthermore since all the
produced photoelectrons are measured, the spectra might be
strongly influenced by complementary and/or competing
processes which occurs simultaneously and can not be easily
evaluated a priori; as for example ionization of the dissociation or predissociation products either below or above the
ionization limit, etc. In similar cases the recording of the
parent molecule ion can decisively clarify the analysis of the
observed spectra and can even give an insight of the other
possible mechanisms which can mask the REMPI spectra.
In the present study, we have investigated the
(311)-photon mass selected REMPI spectra of a jet-cooled
CS2 sample in the one photon 370–406 nm laser wavelength
range corresponding to the three photon 74 000–
1
1
81 000 cm21 energy region. The CS1
2 , CS , and S ions
have been monitored simultaneously and from the spectra
recorded for each ionic mass channel possible fragmentation
mechanisms are discussed. Moreover, since the spectra re-
0021-9606/97/107(21)/8866/8/$10.00
© 1997 American Institute of Physics
Downloaded 03 Oct 2012 to 150.140.158.220. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/about/rights_and_permissions
8867
Berger, Couris, and Gauyacq: The nf Rydberg states of carbon disulfide
corded in the CS1
2 mass channel were much more resolved
than any previously reported spectra, we have been able with
the use of linear and circular polarized light and quantum
defect considerations to propose tentative assignments of the
n f (n54 – 11) Rydberg states.
II. EXPERIMENT
A standard experimental setup has been used consisting
of a XeCl excimer pumped dye laser ~Lambda Physik EMG
201 MSC/FL2002!, operating with QUI and PBBO dyes and
delivering 3–10 mJ pulses within the spectral region of interest ~370–406 nm!. The laser beam was focused by means
of a 15 cm focal length quartz lens into the vacuum chamber
of a homemade TOF mass spectrometer. The CS2 vapor ~50
mbar! was mixed with He as buffer gas ~up to 1 bar! and was
expanded into the vacuum chamber through a pulsed valve
~0.3 mm nozzle! perpendicularly to the laser beam direction.
In this jet source, CS2 molecules were rotationally cooled to
about 70 K. The ions produced by the MPI process were
repelled into the TOF-mass spectrometer and detected by a
pair of microchannel plates. The resulting signal was amplified and fed into boxcar integrators ~SRS 250! which were
interfaced to a PC for data storage. The polarization of the
laser ~linear or circular! was set by means of a Soleil–
Babinet compensator. Wavelength calibration was achieved
by comparison to the numerous sulfur atomic lines18 as they
appeared in the S1 ion channel. The accuracy of our measurements is within 61 cm21.
III. RESULTS AND DISCUSSION
A. Fragmentation of CS2
The (311) REMPI spectra of CS2 which have been
recorded by gating the three main ion mass channels of CS1
2,
CS1, and S1 are presented in Fig. 1 for the 370.4–405.4 nm
laser scan wavelength region ~corresponding to the
74 000– 81 000 cm21 three photon energy region!. The spectra have not been corrected for laser intensity variations over
the spectral range. The laser beam energy variation with
wavelength is shown ~dashed lines! on the same figure. As
shown in Fig. 1, spectra ~a! and ~b! recorded in the S1 and
CS1 ion mass channels are very similar both in intensity and
spectral width except for the numerous atomic sulfur lines
which are present in the S1 channel. They are also very
similar to the spectrum obtained by Morgan et al.15 using an
effusive beam of pure CS2 ~Fig. 2 of Ref. 15!. These spectra
consist of a dense and complex band structure, apparently
superimposed on an intense rather structured continuum. In
Fig. 1~c!, the REMPI spectrum corresponding to the molecule parent ion channel displays a number of well resolved
bands, generally much narrower than the corresponding
bands recorded in the other two channels and without any
underlying continuum.
In Table I several atomic sulfur lines are reported. They
are observed in the REMPI spectrum recorded in the S1 ion
mass channel and are marked by asterisks in Fig. 1~a!. These
transitions have been used for the wavelength calibration of
FIG. 1. The (311) REMPI spectra of CS2 in the 74 000– 81 000 cm21
energy region obtained using linearly polarized light in the ~a! S1, ~b! CS1,
and ~c! CS1
2 mass channels. The dashed lines indicate the laser energy
variation and the arrow indicates the change of dye solution. The asterisks
indicate atomic sulfur lines, given in Table I.
our spectra. Most of these atomic lines correspond18 to transitions from the 3 P J50,1,2 levels of sulfur ~with only exception the line at 76 375 cm21 arising from a 1 D 2 level! and
they correspond to (311) multiphoton ionization processes.
In Fig. 2 a characteristic mass spectrum is shown for the
excitation wavelength 74 163 cm21. The prominent signals
observed in the CS1 and S1 mass channels which are at least
comparable with the CS1
2 are indicative of the strong fragmentation mechanisms present during the excitation of the
neutral molecule and/or the molecular ion. It is well known
that dissociation ~or predissociation! mechanisms which can
take place either below or above the ionization limit of a
molecule can compete efficiently against the REMPI process
and can reduce dramatically or even eliminate completely
the parent molecule ion signal. In the wavelength region of
interest of this study four photons are required to ionize the
molecule and consequently fragmentation of CS2 might occur at the level of first, second, third, fourth or even higher
TABLE I. Observed (311) REMPI atomic sulfur transitions.
Atomic sulfur
transitions (cm21)
75 952
76 324
76 375
76 721
77 181
78 692
79 185
79 612
79 788
80 125
80 182
80 521
80 688
80 931
81 084
Configurations
2
4
2
3 4 0
3s 3 p – 3s 3p ( S )4d
3s 2 3 p 4 – 3s 2 3p 3 ( 4 S 0 )6s
3s 2 3 p 4 – 3s 2 3p 3 ( 2 D 0 )3d?
3s 2 3 p 4 – 3s 2 3p 3 ( 4 S 0 )6s
3s 2 3 p 4 – 3s 2 3p 3 ( 3 P 0 )4s
3s 2 3 p 4 – 3s 2 3p 3 ( 4 S 0 )5d
3s 2 3 p 4 – 3s 2 3p 3 ( 4 S 0 )7s
3s 2 3 p 4 – 3s 2 3p 3 ( 4 S 0 )6d
3s 2 3 p 4 – 3s 2 3p 3 ( 4 S 0 )6d
3s 2 3 p 4 – 3s 2 3p 3 ( 4 S 0 )8s
3s 2 3 p 4 – 3s 2 3p 3 ( 4 S 0 )6d
3s 2 3 p 4 – 3s 2 3p 3 ( 4 S 0 )8s
3s 2 3 p 4 – 3s 2 3p 3 ( 4 S 0 )7d
3s 2 3 p 4 – 3s 2 3p 3 ( 4 S 0 )9s
3s 2 3 p 4 – 3s 2 3p 3 ( 4 S 0 )7d
Terms
J values
3
2–2
1–1
2–2
2–1
2–2
2–3
2–1
0–1
1–2
1–1
2–3
2–1
1–1
1–1
2–1
3
0
P– D
P – 3S 0
1
D – 1D 0?
3
P – 3S 0
3
P – 3P0
3
P – 3D 0
3
P – 3S 0
3
P – 3D 0
3
P – 3D 0
3
P – 3 So
3
P – 3D 0
3
P – 3 So
3
P – 3D 0
3
P – 3 So
3
P – 3D 0
3
J. Chem. Phys., Vol. 107, No. 21, 1 December 1997
Downloaded 03 Oct 2012 to 150.140.158.220. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/about/rights_and_permissions
8868
Berger, Couris, and Gauyacq: The nf Rydberg states of carbon disulfide
FIG. 2. Mass spectrum of CS2 at an excitation energy of 74 163 cm21.
photon absorption. In our experiment the one photon energy
lies within the 3.05–3.35 eV energy range. Reactions ~1! and
~2!,
this fragmentation through reactions ~3! and ~4! requires absorption of five and six photon respectively. Hence, dissociative ionization channels ~3! and ~4! should be less probable
as compared with processes ~1! and ~2!. In addition, the S1
@from reaction ~3!# and CS1 @from reaction ~4!# ion signals
arising from dissociative ionization should exhibit a structureless continuum over the entire spectral region of interest
~unless accidental multiphoton resonances occurred in the S1
and CS1 molecular ions!. No such behavior is observed in
our spectra indicating that mechanisms ~3! and ~4! are not
important under our experimental conditions.
Based on the above considerations, we conclude that the
CS and S fragments are mostly formed by dissociation of
neutral CS2 below the ionization limit following reactions ~1!
and ~2!. The neutral fragments are then ionized by four photons ~the ionization potentials of CS and S being19 11.33 eV
and 10.36 eV, respectively!. This conclusion is further supported by the fact that the REMPI spectra recorded in the
CS1 and S1 ion mass channels are basically identical.
4.46 eV
CS2 ——→ CS~ X 1 S 1 ! 1S~ 3 P !
~1!
B. The nf Rydberg states of CS2
~2!
Among the 16-valence electrons systems, the CO2 molecule has been the most carefully investigated, especially on
what concern the n f Rydberg states. The first high resolution
study of the n f Rydberg states ~from n54 to 32! of the CO2
molecule was performed by Cossart-Magos et al.16 using
VUV absorption spectroscopy. In summary, they have observed two sets of bands ~for each n value!, separated by
'160 cm21 ~corresponding to the spin–orbit coupling constant of the 2 P g ground ionic state! and characterized by
small quantum defects ( d '0.01– 0.10). Their ab initio calculations carried out in the frozen core approximation, for
n54 and 5, have led to similar values of the quantum defects ('0.01) for all the components of the 4 f complex,
while the experimentally determined quantum defects
showed larger variations. Later on Wu and Johnson25,26 investigated the same n f complexes of CO2, by (311)-photon
REMPI spectroscopy and essentially confirmed the one photon absorption results. Similar results have been obtained by
Dobber et al.27 However their (311) REMPI spectra, recorded by collecting the total ion signal, were strongly affected by predissociation and/or fragmentation. Such spectra
exhibit much more complicated structure than the mass selected CO1
2 MPI spectrum, as nicely illustrated in Figs. 3 and
4 of Ref. 25 and in Figs. 1–4 of Ref. 27.
The (311) REMPI spectra of CS2 of Fig. 1 for the
1
1
CS1
2 , CS , and S mass channels show exactly the same
behavior as discussed above. However, at the wavelengths of
the most intense peaks of the spectrum of Fig. 1~c!, the signal was found to follow a cubic dependence vs laser energy
indicating that the resonances occurred at the three photon
energy level as shown in Fig. 3.
As shown in Figs. 4 and 5, in the 75 000– 81 000 cm21
energy region, most of the intense bands can be grouped into
three sets, each set consisting of two series separated by an
energy of about 440 cm21, corresponding to the spin–orbit
7.85 eV
CS2 ——→ CS~ a 3 P ! 1S~ 3 P !
are known19 to occur during CS2 photolysis with threshold
energies of 4.46 eV (35 970 cm –1) and 7.85 eV
(63 310 cm21), respectively. Thus dissociation of CS2 according to paths ~1! and ~2! can be reached via at least two
and three photon absorption respectively.
Two photon absorption in the above spectral region
gives rise to excitation of the 1 B 2 state manifold
(49 196– 54 035 cm21) which is well known to exhibit severe dissociation following reaction ~1!. The spectroscopy of
the 1 B 2 – X 1 S 1
g band system has been extensively studied
by both one photon absorption20,21 and more recently by ~1
11! and (212) multiphoton ionization spectroscopy.22,23
Indeed, numerous broad bands appearing in our CS1 and S1
mass channels REMPI spectra have been found to correspond to 1 B 2 – X 1 S 1
g transition system reported in these previous studies. These bands appear very clearly in the
75 200– 76 500 cm21 three photon energy region of Fig. 1
where the CS1 and S1 ion signals exhibit a strong and dense
structure while there is almost no signal in the CS1
2 mass
channel.
From gas-phase photolysis of CS2 Black et al.24 have
measured a quantum yield close to unity for reaction ~2! in
the 125–140 nm range which corresponds to the
71 400– 80 000 cm21 three photon energy region of the
present work. This implies that, when dissociation of CS2
occurs in this energy range, reaction ~2! should be highly
favored.
Fragmentation of the CS1
2 cation can also occur. From
energy threshold considerations,19
14.852 eV
1 4 0
1 1
CS1
2 ——→ CS~ X S ! 1S ~ S ! ,
~3!
15.83 eV
3
1
2 1
CS1
2 ——→ CS ~ X P ! 1S~ P ! ,
~4!
J. Chem. Phys., Vol. 107, No. 21, 1 December 1997
Downloaded 03 Oct 2012 to 150.140.158.220. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/about/rights_and_permissions
8869
Berger, Couris, and Gauyacq: The nf Rydberg states of carbon disulfide
FIG. 3. Intensity dependence curves of some of the members of the n f
Rydberg states.
2
constant of the CS1
2 (X P g ) ground state. This indicates
that each series belonging to one of these three sets clearly
converge to one or the other spin–orbit component ~V c
53/2 or 1/2! of the ground state of the ionic core. In other
words, these series show upper state electronic structure approaching Hund’s case ~e! in which V c is a good quantum
number, while the electron spin S becomes a mixed quantum
number. This situation is not surprising for high Rydberg
states of CS2 because of the large spin–orbit constant of the
ion ground state. It was actually mentioned in an earlier work
on CS2 ~Ref. 9! that the singlet–triplet designations are valid
only for the lower values of the principal quantum number n,
typically for n,5. More recently, a complete description of
Hund’s case ~e! has been proposed in order to understand
both the electronic and the rotational structure of the n f Rydberg states of CO2. 17 At this point, one can point out a
difference between the two molecules due to their different
rotational and spin–orbit constants in the ion core ~the ratio
23
B 1 /A 1 is equal to 2.531024 in CS1
in
2 and to 2.3310
1
CO2 !. This difference might lead to a slightly different struc-
FIG. 4. The (311) REMPI spectra in the CS1
2 mass channel using linearly
polarized light ~bottom trace! and circularly polarized light ~top trace!.
FIG. 5. The (311) REMPI spectrum in the CS1
2 mass channel with circularly polarized light and assignments of the n f Rydberg states (n
54 – 10). Series A, B, and C classification is explained in the text.
ture within each member of the observed n f series for both
species. In what concerns the electronic structure of the CS2
n f complexes, we will adopt the case ~e! notation (V c )nl.
The three sets of Rydberg series are labeled A, B, and C
in Table II. The series A and C are clearly observed in the
TABLE II. The energies ~in cm21! of the n f Rydberg transitions (n
54 – 11) and their quantum defects. In parentheses are the energies of these
states given in Ref. 15.
˜ 2P
→X
3/2
˜ 2P
→X
1/2
74 103
76 636
78 060
78 926
79 489
79 870
80 142
74 545
77 087
78 504
79 366
79 929
80 310
74 163
76 772
~76 730!
78 119
~78 144!
79 001
~78 969!
79 543
~79 536!
79 911
~79 902!
80 170
~80 172!
74 641
77 208
~77 160!
78 604
~78 570!
79 448
~79 440!
79 980
~79 980!
80 353
~80 352!
80 611
~80 610!
80 807
74 607
76 973
78 275
79 067
79 586
79 939
75 034
77 410
78 720
?
80 025
80 344
D( n 1 2 n 2 )
n
d 3/2
d 1/2
Series A
442
451
444
440
440
440
4
5
6
7
8
10
0.09
0.14
0.17
0.18
0.18
0.20
0.205
0.09
0.14
0.16
0.18
0.18
0.20
Series B
478
436
4
5
0.075
0.07
0.065
0.07
485
6
0.054
0.07
447
7
0.07
0.07
437
8
0.065
0.065
442
9
0.06
0.06
441
10
0.08
0.08
11
Series C
427
436
445
422
439
405
4
5
6
7
8
9
0.07
20.05
20.04
20.04
20.03
20.03
0
J. Chem. Phys., Vol. 107, No. 21, 1 December 1997
Downloaded 03 Oct 2012 to 150.140.158.220. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/about/rights_and_permissions
20.05
20.04
20.04
0
20.03
20.09
8870
Berger, Couris, and Gauyacq: The nf Rydberg states of carbon disulfide
REMPI spectrum recorded with circularly polarized light
~e.g., Fig. 5! while series B exhibits intense bands in the
spectrum recorded with linearly polarized light ~e.g., Fig. 4!.
Polarization effects will be discussed below. The series B has
been previously observed by Morgan et al.15 However, their
reported positions ~shown in parentheses in the first two columns of Table II! present large energy shifts ('40 cm21)
for the first members of the series whereas the agreement
becomes very good for the higher members. Part of these
differences could be due to the fact that they were collecting
the total number of photoelectrons produced, resulting in
broader band profiles, hence leading to higher uncertainties
in locating the band centers compared to our spectra.
The most important classification criterion for these
three sets of bands A, B, and C is that each of them is
associated with a characteristic quantum defect d, given in
the fifth and sixth columns of Table II. Theses quantum defects have been obtained by fitting the corresponding band
frequencies with the Rydberg formula. The Rydberg constant
was set at 109 736.5 cm21 ~corrected by taking into account
the mass of the CS2 molecule! and the lowest ionization
potential and spin–orbit splitting, recently determined by
high resolution ZEKE spectroscopy28 ~in good agreement
with earlier studies29,30!, i.e., 81 286 cm21 and 440 cm21, respectively, have been used for this calculation. Edlen plots of
the quantum defect d vs the difference ~IP-n! are shown in
Fig. 6. A linear behavior is observed within each series, as
expected ~except for the value corresponding at 78 119 cm21
band of series B!. The IP value was checked from our series
limits by varying it in the Edlen plots as shown in Fig. 6
~open circles and triangles in this figure!. This shows that the
IP value determined by Fischer et al.28 is in very good agreement with our series limits. However the present work gives
an estimate of the ionization potential value with a better
accuracy (62 cm21! than in Ref. 28.
There are typical ranges of the quantum defect values d,
corresponding to each n l series depending on the atoms
which compose the molecule. It is generally accepted that
the quantum defects for molecules composed of third-row
atoms ~e.g., CS2! are relatively large and are approximately
equal to 2.0, 1.5, 0.4, and ;0.0 for the s-, p-, d- and f -type
Rydberg molecular orbitals respectively. Consequently, since
only np and n f Rydberg states are three-photon allowed, and
from the magnitude of the (n2 d ) values obtained from our
experimental results ~see Table II!, it can be unambiguously
deduced that these series arise from f orbitals. Such an assignment is reinforced by the fact that the observed peaks
exhibit fairly sharp structures. This later observation indicates that these states have relatively long lifetimes and are
weakly perturbed by dissociation/fragmentation mechanisms.
As it has been previously shown, in particular for CO2, n f
Rydberg states are weakly predissociated while the n p ones
are heavily predissociated.
Figure 7 shows the detailed structure of the n f – X transitions between n54 and n59 for both polarization conditions. As shown in this figure, the number of n f complex
components decreases as n increases. For n510 and 11, only
FIG. 6. Edlen’s plots. Calculated quantum defects d vs (IP2 n ) for the n f
Rydberg series observed. The best value for the IP ~j! is indicated by the
linear fits ~straight lines!.
one component ~from the series B! appears in our spectra.
From n56 to 9, all observed bands are associated with series
A, B or C ~with the exception of the transition at
78 119 cm21!. The 5 f complex exhibits a more complex
structure but the separation between the ( 2 P 3/2)5 f and the
( 2 P 1/2)5 f components still appears clearly. The picture becomes very complicated for n54 partly because ( 2 P 3/2)4 f
and ( 2 P 1/2)4 f manifolds overlap. Comparable observations
have been done for the n f complexes of CO2. 25 The apparent
complexity of the 4 f band structure might originate from
two facts; ~i! in this Rydberg complex, Hund’s case ~e! is not
completely achieved ~ii! some of the bands appearing in the
4 f Rydberg energy region might not belong to these complexes but to other transitions which they become allowed
e.g., because of vibronic couplings. Calculations of the type
of those carried out in Ref. 17 can substantially contribute to
the understanding of the observed spectral complexity.
In addition to quantum defect considerations, polarization effects are very important in order to help the upper state
symmetry assignments. Polarization dependence of the intensity of the observed bands has also been carried out in the
present work. As was shown in earlier work25 multiphoton
transitions are polarization dependent ~contrary to one photon absorption! providing a route to determine the final state
symmetry. The REMPI spectra of CS2 recorded with linearly
J. Chem. Phys., Vol. 107, No. 21, 1 December 1997
Downloaded 03 Oct 2012 to 150.140.158.220. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/about/rights_and_permissions
Berger, Couris, and Gauyacq: The nf Rydberg states of carbon disulfide
8871
FIG. 7. Structure of the n f – X transitions for n54 – 9 for both linearly and circularly polarized laser light.
and circularly polarized light, as shown in Figs. 4 and 7,
exhibit large differences. By changing the polarization from
linear to circular, the most intense bands have almost vanished while other bands have significantly increased. It is
then convenient to define the polarization ratio V as being
the ratio of the band intensity recorded with circularly polarized light to that obtained with linearly polarized light. In
Table III the positions of the numerous sharp bands recorded
J. Chem. Phys., Vol. 107, No. 21, 1 December 1997
Downloaded 03 Oct 2012 to 150.140.158.220. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/about/rights_and_permissions
8872
Berger, Couris, and Gauyacq: The nf Rydberg states of carbon disulfide
TABLE III. Band positions of the (311) REMPI spectrum of CS2 in the
73 900– 81 000 cm21 energy region. The family assigned to each band is followed by
the effective quantum number n eff for the two IP limits. The measured polarization
ratio V and the corresponding assignments are also indicated.
Transition
energy ~cm21!a
73 933
73 969
74 023
74 103
74 163
74 303
74 402
74 486
74 545
74 607
74 641
74 748
74 828
74 958
75 034
75 252
76 345
76 405
76 580
76 637
76 663
76 700
76 715
76 772
76 786
76 861
76 969
77 084
77 150
77 208
77 318
77 390
77 410
77 835
78 060
78 119
78 184
78 275
78 504
78 540
78 603
78 720
78 926
79 001
79 069
79 365
79 448
79 489
79 489
79 542
79 585
79 929
79 980
80 025
79 870
79 911
79 939
80 310
80 353
80 170
80 611
Label
‘‘w’’
‘‘w’’
‘‘w’’
A4
B4
‘‘w’’
‘‘w’’
A 48
C4
B 84
neff 3/2
3.863
3.873
3.887
3.909
3.925
3.964
3.993
4.017
4.035
4.053
‘‘w’’
‘‘w’’
‘‘w’’
C 48
4.164
neff 1/2
3.844
3.871
3.893
3.909
3.926
3.935
3.966
3.988
4.027
4.049
1.
1.
1@
1@
1.
'1
1.
1.
1@
'1
1.
1.
1.
'1
'1
Assignment
( 2 P g 3/2 )4 f
( 2 P g 1/2 )4 f
‘‘w’’
‘‘w’’
‘‘w’’
‘‘w’’
‘‘w’’
~76 730!
‘‘w’’
~77 160!
‘‘w’’
A 5 2 11 (57)
A5
B 5 1 22 (109)
B 5 2 11 (57)
B5
C5
A 5 1447
B 5 2 11 (58)
B 5 1436
C 5 1437
~78 144!
‘‘w’’
~78 570!
~78 969!
~79 440!
~79 536!
~79 980!
~79 902!
~80 352!
~80 172!
~80 610!
A6
B6
C6
A 6 1444
B 6 2 11 (63)
B 6 1484
C 6 1445
A7
B7
C7
A 7 1439
B 7 1447
A8
Bg
C8
A 8 1440
B 8 1438
C 8 1440
A9
B9
C9
A 9 1440
B 9 1442
B 10
B 101442
4.858
4.872
4.892
4.643
4.930
4.938
4.980
5.042
5.11
4.706
5.187
5.321
5.639
5.832
5.886
5.948
6.037
6.283
6.395
6.539
6.819
6.930
7.035
7.558
7.727
7.814
7.814
7.932
8.032
8.993
9.329
8.803
8.933
9.026
4.803
4.862
4.897
4.928
4.989
5.031
5.042
5.311
5.471
5.566
5.639
5.838
5.869
5.928
6.042
6.260
6.346
6.426
6.817
6.941
7.004
7.814
7.928
8.032
7.690
8.803
8.940
9.916
9.921
10.84
80 806
80 471 ‘‘vw’’
80 513 ‘‘vw’’
80 527 ‘‘vw’’
80 750
80 781
80 973
81 007
a
Vcir/lin
8
B 11
10.92
.1
2.36
!1
!1
.1
,0.1
@1
!1
1.86
2.0
0.092
( 2 P g 3/2 )5 f
( 2 P g 1/2 )5 f
2.0
1.36
!1
!1
2.06
1.93
!1
0.09
2.10
1.19
0.17
2.42
2.
,0.1
@1
@1
0.27
1.93
1.31
0.42
2.05
1.96
,0.1
1.31
1.58
,0.1
0.71
1.03
( 2 P g 3/2 )6 f
( 2 P g 1/2 )6 f
( 2 P g 3/2 )7 f
( 2 P g 1/2 )7 f
( 2 P g 3/2 )8 f
( 2 P g 1/2 )8 f
( 2 P g 3/2 )9 f
( 2 P g 1/2 )9 f
( 2 P g 3/2 )10f
( 2 P g 1/2 )10f
( 2 P g 3/2 )11f
( 2 P g 3/2 )11f
Symbol ‘‘w’’ for bands appearing weak in spectra recorded with linearly
and circularly polarized light.
in the CS1
2 mass channel are reported together with their
polarization ratios. From previous polarization studies,25 it
can be summarized that for three photon resonance transitions, polarization ratio of D–S and F–S transitions is expected to be equal to 2.5 while those of S–S and P–S
transition should be much less than 2.5. These ratio values
can only be obtained for pure absorption processes. Since in
our experiments, absorption processes might be perturbed by
fragmentation mechanisms and/or saturation effects ~due to
high laser intensities! and/or contributions from residual linear polarization, it is not surprising to observe fluctuations in
V values. However, bands can be divided into two groups.
The first group is composed of the bands which have a polarization ratio much larger than one, implying that the upper
state symmetry is D or F. Bands composing the second
group have V values much less than one and thus, correspond to S–S or P–S transitions. Based on these arguments, we can assign the bands of series B as corresponding
to S–S or P–S transitions and bands of series A and C to
D–S and F–S transitions.
On the basis of very simple intensity considerations, one
can give a more detailed assignment of the electronic symmetry of the upper Rydberg components corresponding to
the intense series A, B, and C. Indeed, a simple model based
on the ‘‘semiunited atom’’ approximation allows to evaluate
the relative magnitude of the three-photon transitions moments for all symmetry components of the n f complexes,
and for both linear and circular polarization. This model has
been applied successfully for the 3d and 4s Rydberg complexes of acetylene which also correspond to a 2 P ionic core
symmetry.31,32 The ground state p g orbital has been considered as a d p g orbital, correlating with an atomic d orbital in
the ‘‘semiunited atom’’ limit, so that the electronic transition
moment is only governed by the vibronic tensor magnitudes
T 1 and T 3 , or in other words by only 1 parameter, i.e., the
ratio T 1 /T 3 . The remaining part of the transition moment,
involving the angular electronic and rotational factors, is obtained from standard angular momentum algebra in both polarization conditions. In a preliminary step, we have set arbitrarily the ratio T 1 /T 3 to 1, and calculated the relative
intensities for the transition towards the 6 electronic components of an n f complex, accessible by three-photon excitation, i.e., f s (P), f p (S,D), f d (P,F) and f f (D). In the
case of linearly polarized light, only one transition carries a
significant intensity, the transition to the f d P, more intense
than all the other bands by at least one or two orders of
magnitude. So we can assign series B to the transition to the
f d P component of the n f complex. This result is in good
agreement with the rather small quantum defect of this series
of about 0.07, as shown in Table II, since one expects a small
quantum defect for the nonpenetrating f d orbital as in
CO2. 17 In the case of circularly polarized light, there is not
only one prominent transition, which is actually in agreement
with the observed spectrum. On the basis of this simple
model calculation, and considering in addition the quantum
defects of Table II, one can tentatively assign series A to the
upper f p D component and series C to the upper f d F com-
J. Chem. Phys., Vol. 107, No. 21, 1 December 1997
Downloaded 03 Oct 2012 to 150.140.158.220. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/about/rights_and_permissions
Berger, Couris, and Gauyacq: The nf Rydberg states of carbon disulfide
ponent. These tentative assignments should be confirmed by
a more detailed calculation. Nonetheless, it has to be pointed
out that this very simple approach explains the small number
of observed n f components in the three-photon spectrum,
due to their very different relative intensities. This approach
also provides a reasonable assignment of the observed n f
components, without need of a large number of parameters
to be fitted ~intensity calculations are based on only one parameter, the ratio T 1 /T 3 !.
As said above, the 4 f complex exhibits a complex structure in the linear polarization spectrum, as compared with the
higher members (n55 – 9) spectra in the same figure. The
same band structure was observed by Morgan et al.15 Moreover, all bands belonging to this complex correspond to polarization ratios smaller than 1. Similar observations for the
4 f complex of CO2 have been done by Wu and Jonhson.24
They have found that the V values were comprising between
0.5 and 0.9. Therefore, in the case of the 4 f complex, we
could not use polarization effects in order to classify the
observed series into sets A, B, and C. Alternatively, we used
the quantum defect values for the different bands and compared them to the extrapolated values on the Edlen plots of
Fig. 6. Bands at 74 103 cm21, 74 163 cm21, and
74 607 cm21 ~see Table III and Fig. 7! could therefore be
assigned to series A, B, and C, respectively, converging to
the lowest spin–orbit limit ~with quantum defects of 0.09,
0.06, and 20.05, respectively!, while bands located at
74 545 cm21, 74 641 cm21, and 75 034 cm21 have been associated with series A, B, and C, respectively, converging to
the upper spin–orbit limit ~with quantum defects of 0.09,
0.075, and 20.05, respectively!.
We also recorded the (311) REMPI spectrum for the
34
SC32S isotope ~present with a natural abundance of about
4%! in the 75 000– 81 000 cm21 energy region. No frequency shift was observed between the most intense bands of
the spectra obtained by monitoring the 32SC32S and 34SC32S
cation signals indicating that these bands correspond to electronic origins.
IV. CONCLUSION
We have investigated the 74 000– 81 000 cm21 energy
region of CS2 using three photon resonance enhanced multi1
1
photon ionization spectroscopy. The CS1
mass
2 , CS , S
selected MPI spectra have been recorded and led to a better
understanding of the fragmentation mechanisms which compete against ionization process. From the comparison of
these three spectra, we concluded that dissociation of the
CS1
2 ions is negligible compared to the other processes and
that important predissociation of the CS2 molecule occurs at
two photon absorption, particularly in the 392.2–398.9 nm
wavelength range ~75 200– 76 500 cm21 three photon energy
region!. The nicely resolved spectrum recorded in the CS1
2
mass channel has been assigned to transitions to the n f
Rydberg complexes ~with n54 – 10! from selection rules and
quantum defect considerations. The use of linear and circular
polarized light enabled the identification of three Rydberg
8873
series, in terms of electronic f l L components of the n f
complex for n56 – 10. Those series are found to converge to
2
the two spin–orbit components of the CS1
2 (X P g ) ground
state with an adiabatic ionization potential of 81 286
in
good
agreement
with
previous
62 cm21,
measurements.28,29,30 Further theoretical work is definitely
required in order to study in more details this energy region
of CS2.
ACKNOWLEDGMENTS
The authors wish to acknowledge support from the Ultraviolet Laser Facility operating at FORTH-IESL within the
Large Installations Plan of the EU. J.-Ph.B. thanks the EU
for partial financial support under the HCM network Grant
No. CHRX-CT-94-0561. Support through the Greek–French
bilateral collaboration ~PICS No. 152! is also acknowledged.
1
W. C. Price and D. M. Simpson, Proc. R. Soc. London, Ser. A 165, 272
~1938!.
2
Y. Tanaka, A. S. Jursa, and F. J. LeBlanc, J. Chem. Phys. 32, 1205 ~1960!.
3
J. W. Rabalais, J. M. McDonald, V. Scherr, and S. P. McGlynn, Chem.
Rev. 71, 73 ~1971!.
4
F. R. Greening and G. W. King, J. Mol. Spectrosc. 59, 312 ~1976!.
5
R. McDiarmid and J. P. Doering, J. Chem. Phys. 91, 2010 ~1989!.
6
C. Cossart-Magos, M. Horani, M. Jungen, and F. Launay, J. Chem. Phys.
104, 7412 ~1996!.
7
D. G. Wilden and J. Comer, Chem. Phys. 53, 77 ~1980!.
8
M.-J. Hubin-Franskin, J. Delwiche, A. Poulin, B. Leclerc, P. Roy, and D.
Roy, J. Chem. Phys. 78, 1200 ~1983!.
9
J. P. Doering and R. McDiarmid, J. Phys. Chem. 87, 1822 ~1983!.
10
A. Gedanken, J. Phys. Chem. 92, 5862 ~1988!.
11
S. Couris, E. Patsilinakou, M. Lotz, E. R. Grant, C. Fotakis, C. CossartMagos, and M. Horani, J. Chem. Phys. 100, 3514 ~1994!.
12
J. Baker and S. Couris, J. Chem. Phys. 105, 62 ~1996!.
13
J. Baker and S. Couris, J. Chem. Phys. 103, 4847 ~1995!.
14
J. Baker, M. Konstantaki, and S. Couris, J. Chem. Phys. 103, 2436 ~1995!.
15
R. A. Morgan, M. A. Baldwin, A. J. Orr-Ewing, M. N. R. Ashfold, W. J.
Buma, J. B. Milan, and C. A. de Lange, J. Chem. Phys. 104, 6117 ~1996!.
16
C. Cossart-Magos, M. Jungen, and F. Launay, Mol. Phys. 61, 1077
~1987!.
17
C. Cossart-Magos, H. Lefebre-Brion, and M. Jungen, Mol. Phys. 85, 821
~1995!.
18
V. Kaufman and W. C. Martin, J. Phys. Chem. Ref. Data 22, 282 ~1993!.
19
E. K. Moltzen, K. J. Klabunde, and A. Senning, Chem. Rev. 88, 391
~1988!.
20
R. J. Hemley, D. J. Leopold, J. L. Roebber, and V. Vaida, J. Chem. Phys.
79, 5219 ~1983!.
21
J. L. Roebber and V. Vaida, J. Chem. Phys. 83, 2748 ~1985!.
22
D. J. Donalson, J. Chem. Phys. 91, 7455 ~1989!.
23
A. Mank, C. Starrs, M. N. Jego, and J. W. Hepburn, J. Chem. Phys. 104,
3609 ~1996!.
24
G. Black, R. L. Sharpless, and T. G. Slanger, J. Chem. Phys. 66, 2113
~1977!.
25
M. W. Wu and P. M. Johnson, J. Chem. Phys. 91, 7399 ~1989!.
26
M. W. Wu and P. M. Johnson, J. Chem. Phys. 95, 761 ~1991!.
27
M. R. Dobber, W. J. Buma, and C. A. de Lange, J. Chem. Phys. 101, 9303
~1994!.
28
I. Fischer, A. Lochschmidt, A. Strobel, G. Niedner-Schatteburg, K.
Muller-Dethlefs, and V. E. Bondybey, Chem. Phys. Lett. 202, 542 ~1993!.
29
A. W. Potts and G. H. Fatallah, J. Phys. B 13, 2545 ~1980!.
30
L. S. Wang, J. E. Reutt, Y. T. Lee, and D. A. Shirley, J. Electron Spectrosc. 47, 167 ~1988!.
31
J. H. Fillion, A. Campos, J. Pedersen, N. Shafizadeh, and D. Gauyacq, J.
Chem. Phys. 105, 22 ~1996!.
32
N. Shafizadeh, J. H. Fillion, D. Gauyacq, and S. Couris, Philos. Trans. R.
Soc. London, Ser. A 355, 1637 ~1997!.
J. Chem. Phys., Vol. 107, No. 21, 1 December 1997
Downloaded 03 Oct 2012 to 150.140.158.220. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/about/rights_and_permissions