spectroscopy and mass spectrometry of state-selected...
TRANSCRIPT
이학박사 학위논문
분광학과 질량분석법을 이용한 상태 선택된
벤젠족 분자 이온들의 연구
Spectroscopy and Mass Spectrometry of State-selected Benzenoid Molecular Ions
2004년 2월
서울대학교 대학원 화학부 물리화학전공
권 찬 호
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
분광학과 질량분석법을 이용한 상태 선택된
벤젠족 분자 이온들의 연구
Spectroscopy and Mass Spectrometry of State-selected Benzenoid Molecular Ions
지도교수 김 명 수
이 논문을 이학박사 학위논문으로 제출함
2003년 11월
서울대학교 대학원
화학부 물리화학전공
성 명 권 찬 호
권찬호의 이학박사 학위논문을 인준함
2003년 12월 5일
위 원 장
부위원장
위 원
위 원
위 원
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
A Ph. D. Dissertation
Spectroscopy and Mass Spectrometry of State-Selected Benzenoid Molecular Ions
By Chan Ho Kwon Supervisor : Prof. Myung Soo Kim
School of Chemistry Seoul National University
February 2004
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
Abstract
Presence of benzene cation in a long-lived (10 µsec or longer) excited
electronic state, presumably 2E2g, was found through photodissociation kinetics
and charge exchange ionization mass spectrometry. In a subsequent work, a
method based on charge exchange in collision cells of a modified double focusing
mass spectrometer was developed to search routinely the long-lived excited states
with conventional mass spectrometry. This is based on the criterion that charge
exchange between polyatomic species is efficient only when the energy of
reaction is close to zero or negative ( ), or the exoergicity rule. Therefrom,
the 2B2 states of chlorobenzene, bromobenzene, benzonitrile, and phenyl
acetylene cations were found to have long lifetimes (ten microseconds or longer)
while excited electronic states with long lifetime were not detected for
fluorobenzene, iodobenzene, toluene, nitrobenzene, and styrene cations. The long-
lived states found were those generated by removal of an electron from the in-
plane nonbonding p orbitals of halogens or in-plane π orbitals of the triple bonds,
displaying well-resolved vibrational structures in the photoelectron spectra.
B~
0≤∆E
B~
For spectroscopic investigation of the benzenoid cations in excited electronic
states, firstly, continuously tunable vacuum ultraviolet (VUV) light source in the
104 ∼ 125 nm range has been developed by utilizing four-wave sum frequency
mixing in Hg vapor. Then, vibrational spectra of benzenoid cations in the ground
electronic states ( X ) and in the excited electronic states ( ), which consisted
mostly of fundamentals with proper symmetries, have been measured by one-
photon mass-analyzed threshold ionization (MATI) spectroscopy using VUV
radiation generated by four-wave mixing in Kr gas or Hg vapor. Vibrational
assignments were made by referring to the previous results, comparing with
~ B~
i
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
calculated frequencies, and invoking the selection rule for one-photon process.
Geometric change upon ionization, was calculated quantum chemically and used
to explain the prominent overtones of some vibrational modes and combinations
involving these. Furthermore, Franck-Condon factors calculations to utilize
spectral intensity information, were found to be extremely useful for reliable
vibrational assignment.
While, to aid the spectral analysis for Jahn-Teller benzenoid cations, the Jahn-
Teller coupling parameters for four e2g modes of C6X6+ (X=H, D, F) in the ground
electronic state were calculated from the topographical data of the potential
energy surface at the density functional theory (DFT) level. These were used to
calculate the energies of the Jahn-Teller states and upgraded through the
multimode fit to the experimental data. Excellent agreement between the
experimental and calculated frequencies was achieved. The vibrations which are
not linear Jahn-Teller active were observed and could be assigned by referring to
the frequencies obtained at the DFT level.
Keywords : Benzenoid Cation, Long-lived Excited Electronic State, Jahn-
Teller Effect, Vibrational Spectra, Vacuum Ultraviolet Radiation, Mass-
analyzed Threshold Ionization Spectroscopy, Charge Exchange Ionization
Mass Spectrometry, Quantum Chemical Calculation, Photodissociation
Kinetics, Franck-Condon Factor, Potential Energy Surface.
Student Number : 99305-802
ii
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
Contents
Abstract (in English) ......................................................................... i
Contents ............................................................................................. iii
List of Tables ...................................................................................... viii
List of Figures .................................................................................... xii
Chapter
1 Introduction ........................................................................................ 1
1.1 Why the Excited State? ................................................................... 1
1.2 Mass Spectrometry .......................................................................... 3
1.2.1 Ionization Methods ................................................................... 5
1.2.2 Mass Analyzers ......................................................................... 10
1.2.3 Theory of Mass Spectra ............................................................ 10
1.2.4 Long-lived Excited Electronic State ......................................... 12
1.3 MATI and ZEKE Spectroscopies ................................................... 13
1.3.1 Historical Background .............................................................. 13
1.3.2 General Principles ..................................................................... 16
1.3.3 One-photon MATI Scheme and Apparatus ............................... 19
1.3.4 Application to Excited Electronic State of Ions ........................ 26
References ............................................................................................. 28
2 Discovery of Isolated Electronic State by Mass Spectrometry 34
iii
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
2.1 Initial Discovery : Benzene Cation ................................................. 35
2.1.1 Experimental Setup ................................................................... 36
2.1.2 Energetics of Benzene Ion ........................................................ 40
2.1.3 Photodissociation Kinetics ........................................................ 42
2.1.4 Quenching of Photodissociation ............................................... 47
2.1.5 Charge Exchange Ionization by Benzene Cation ...................... 53
2.1.6 Conclusions ............................................................................... 58
2.2 Method to Detect Isolated Electronic States ................................. 60
2.2.1 Experimental Setup ................................................................... 61
2.2.2 Principle of Method .................................................................. 63
2.2.3 Utilization of Method ................................................................ 64
2.2.4 Conclusions ............................................................................... 74
2.3 Monosubstituted Benzene Cations ................................................. 74
2.3.1 Charge Exchange Ionization and Exoergicity Rule .................. 75
2.3.2 Experimental Setup ................................................................... 76
2.3.3 Results ....................................................................................... 77
2.3.4 Conclusions ............................................................................... 93
References .............................................................................................. 95
3 Coherent Vacuum Ultraviolet Radiation ................................. 101 3.1 VUV Generation in Gaseous Nonlinear Medium ......................... 101
3.1.1 General Principles ..................................................................... 101
3.1.2 Wavelength Calibration ............................................................. 103
3.1.3 Measurements of VUV Intensity .............................................. 105
3.2 Four-wave Difference Frequency Mixing in Kr Gas .................... 107
3.3 Four-wave Sum Frequency Mixing in Hg Vapor .......................... 107
3.4 Development of Coherent VUV source at 104 – 108 nm .............. 109
iv
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
3.4.1 Experimental Setup ................................................................... 112
3.4.2 VUV Generation at 104 – 108 nm ............................................ 115
References .............................................................................................. 119
4 VUV-MATI Spectroscopy of Benzenoid Molecules ............. 121 4.1 Determination of Ionization Energies ............................................ 122
4.2 Selection Rules in One-photon MATI Spectra .............................. 125
4.3 Franck-Condon Factor Calculations ............................................. 128
4.4 VUV-MATI Spectroscopy of Monohalobenzenes ......................... 129
4.4.1 Vibrational Spectra in the Ground Electronic States, ........... 131 X~
4.4.1.1 Chlorobenzene Cation .................................................. 131
4.4.1.2 Bromobenzene Cation .................................................. 137
4.4.1.3 Iodobenzene Cation ...................................................... 138
4.4.1.4 Fluorobenzene Cation..................................................... 141
4.4.2 Vibrational Spectra in the Excited Electronic States, ........... 144 B~
4.4.2.1 Chlorobenzene Cation .................................................. 144
4.4.2.2 Bromobenzene Cation .................................................. 151
4.4.2.3 Iodobenzene Cation ...................................................... 151
4.4.3 Conclusions ............................................................................... 155
4.5 VUV-MATI Spectroscopy of Difluorobenzenes ............................ 158
4.5.1 Computational ........................................................................... 158
4.5.2 Molecular Geometry Calculation .............................................. 159
4.5.3 Ionization Energies .................................................................... 160
4.5.4 p- difluorobenzene Cation ......................................................... 164
4.5.5 m- difluorobenzene Cation ........................................................ 165
4.5.6 o- difluorobenzene Cation ......................................................... 172
4.5.7 Geometrical Change upon Ionization and Vibrational Progressions
v
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
..................................................................................................... 179
4.5.8 Conclusions ............................................................................... 181
4.6 VUV-MATI Spectroscopy of Phenylacetylene and Benzonitrile . 182
4.6.1 Quantum Chemical Calculations .............................................. 183
4.6.2 Ionization Energies .................................................................... 184
4.6.3 Phenylacetylene Cation ............................................................. 188
4.6.4 Benzonitrile Cation ................................................................... 195
4.6.5 Conclusions ............................................................................... 201
References .............................................................................................. 202
5 The Jahn-Teller Effect in Benzenoid Cations ......................... 206 5.1 General Descriptions ....................................................................... 206
5.2 Computation .................................................................................... 208
5.2.1 The Jahn-Teller Potential Energy Surfaces and Coupling Constants..................................................................................................... 208
5.2.2 The Jahn-Teller Parameters and Vibronic Energies .................. 210
5.3 C6H6+ and C6D6
+ in the State ................................................... 211 X~
5.3.1 Ionization Energies...................................................................... 213
5.3.2 Jahn-Teller Effect and Vibronic Splitting ................................. 214
5.3.3 Vibrational Analysis .................................................................. 215
5.3.4 Conclusions ............................................................................... 228
5.4 C6H6+ and C6D6
+ in the State ................................................... 229 B~
5.4.1 Jahn-Teller Effect and Vibronic Splitting ................................... 231
5.4.2 Selection Rule ........................................................................... 232
5.4.3 Vibrational Analysis .................................................................. 233
5.4.4 Radiationless B~ 2E2g → X~ 2E1g transition .............................. 245
5.4.5 Conclusions ............................................................................... 247
vi
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
5.5 C6F6+ in the State ...................................................................... 247 X~
5.5.1 MATI Spectrum and Ionization Energy .................................... 248
5.5.2 Calculated Results ..................................................................... 250
5.5.3 Vibrational Analysis .................................................................. 255
5.5.4 Conclusions ............................................................................... 262
References .............................................................................................. 264
Abstract (in Korea) ........................................................................... 267
Acknowledgement ............................................................................. 269
vii
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
List of Tables
1.1
1.2
2.1
2.2
2.3
2.4
2.5
3.1
3.2
4.1
4.2
Classification of components consisting of mass spectrometer. ............ 4
Recombination energies of gaseous ions. ............................................... 8
Ion source pressure (P), collision frequency (Zc), source residence time (tR),
and number of collisions (Ncoll) suffered by ions exiting the ion source at
some benzene pressures. ......................................................................... 50
Ionization Energies and the ratios of molecular ion intensities generated by
charge exchange ionization (CI) with benzene ion and by electron ionization
(EI). ........................................................................................................ 57
Collision gases, their ionization energies (IE) in eV, and success/failure to
generate their ions by charge exchange with some precursor ions. ....... 79
Recombination energies of the X~ 2B1, A~ 2A2, B~ 2B2, and C~ 2B1 electronic
states of some monosubstituted benzene cations and the calculated oscillator
strengths of the radiative transitions from the B~ 2B2 states. .................. 80
Recombination energies (RE) of some excited hole states of fluorobenzene
cation and the calculated oscillator strengths of some radiative transitions.
.................................................................................................................. 81
Tunable generation in rare gases. ........................................................... 104
Tunable generation in metal vapors. ...................................................... 104
Ionization energies (IE) to the ground ( X~ 2B1) and B~ 2B2 excited states of
chloro-, bromo-, iodo-, and fluorobenzene cations, in eV. ..................... 132
Vibrational frequencies (in cm-1) and their assignments for the ground state
viii
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
( X~ 2B1) chlorobenzene cation. ............................................................... 134
4.3
4.4
4.5
4.6
4.7
4.8
4.9
4.10
4.11
4.12
Vibrational frequencies (in cm-1) and their assignments for the ground state
( X~ 2B1) bromobenzene cation. ............................................................... 140
Vibrational frequencies (in cm-1) and their assignments for the ground state
( X~ 2B1) iodobenzene cation. .................................................................. 143
Vibrational frequencies (in cm-1) and their assignments for the ground state
( X~ 2B1) fluorobenzene cation. ................................................................ 146
Vibrational frequencies (in cm-1) and their assignments for the
chlorobenzene cation in the B~ 2B2 excited state. .................................. 150
Vibrational frequencies (in cm-1) and their assignments for the
bromobenzene cation in the B~ 2B2 excited state. .................................. 153
Geometrical parameters of p-difluorobenzene in the ground state ( X~ 1Ag)
and those of the cation in the ground state ( X~ 2B2g) calculated at the
B3LYP/6-311++G (2df, 2pd) level. ........................................................ 161
Geometrical parameters of m-difluorobenzene in the ground state ( X~ 1A1)
and those of the cation in the ground state ( X~ 2B1) calculated at the
B3LYP/6-311++G (2df, 2pd) level. ........................................................ 162
Geometrical parameters of o-difluorobenzene in the ground state ( X~ 1A1)
and those of the cation in the ground state ( X~ 2B1) calculated at the
B3LYP/6-311++G (2df, 2pd) level. ........................................................ 163
Ionization energies (IE) to the ground states of p-, m-, and o-difluorobenzene
cations, in eV. ......................................................................................... 166
Frequencies (in cm-1) of the totally symmetric modes of the p-
ix
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
difluorobenzene cation in the ground electronic state ( X~ 2B2g) calculated at
the B3LYP level with various basis sets. ................................................ 167
4.13
4.14
4.15
4.16
4.17
4.18
4.19
4.20
5.1
5.2
Vibrational frequencies (in cm-1) and their assignments for the p-
difluorobenzene cation in the electronic ground state ( X~ 2B2g).. ........... 169
Vibrational frequencies (in cm-1) and their assignments for the m-
difluorobenzene cation in the ground electronic state ( X~ 2B1). ............. 174
Vibrational frequencies (in cm-1) and their assignments for the o-
difluorobenzene cation in the ground electronic state ( X~ 2B1). ............. 177
Vibrational frequencies (in cm-1) of phenylacetylene neutral and cation in the
ground electronic states calculated at the B3LYP, B3PW91, and BP86 levels
with the 6-311++G (2df, 2pd) basis set and experimental data for the neutral.
.................................................................................................................. 185
Vibrational frequencies (in cm-1) of benzonitrile neutral and cation in the
ground electronic states calculated at the B3LYP, B3PW91, and BP86 levels
with the 6-311++G (2df, 2pd) basis set and experimental data for the neutral.
.................................................................................................................. 186
Ionization energies (IE) of phenylacetylene and benzonitrile, in eV. ..... 188
Vibrational frequencies (in cm-1) and their assignments for phenylacetylene
cation in the ground electronic state ( X~ 2B1). ........................................ 191
Vibrational frequencies (in cm-1) and their assignments for benzonitrile
cation in the ground electronic state ( X~ 2B1). ........................................ 198
Ionization energies (IE) to the ground states of C6H6+ and C6D6
+, in eV. 215
Vibrational frequencies (in cm-1) and assignments for C6H6+ in the ground
electronic state ( X~ 2E1g). ........................................................................ 215
x
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
Vibrational frequencies (in cm-1) and assignments for C6D6+ in the ground
electronic state ( X~ 2E1g). ........................................................................ 222
5.3
5.4
5.5
5.6
5.7
5.8
5.9
Ionization energies (IE) to the excited states of C6H6+ and C6D6
+, in eV. 234
Vibrational frequencies (in cm-1) and assignments for C6D6+ in the excited
electronic state ( B~ 2E2g). ......................................................................... 238
Vibrational frequencies (in cm-1) and assignments for C6H6+ in the excited
electronic state ( B~ 2E2g). ......................................................................... 243
Ionization energies (IE) of hexafluorobenzene, in eV. ........................... 251
Vibrational frequencies (in cm-1) of hexafluorobenzene cation in the ground
electronic state ( X~ 2E1g) measured by the one-photon MATI and their
assignments. ........................................................................................... 253
Calculated and experimental Jahn-Teller coupling parameters for the four e2g
vibrational modes of C6F6+. .................................................................... 260
xi
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
List of Figures
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1.9
1.10
Illustrated diagram of photophysical processes. Radiation by straight arrows
(F=fluorescence, P=phosphorescence) and radiationless processes by wavy
arrows (IC=internal conversion, ISC=intersystem crossing, VR=vibrational
relaxation)................................................................................................. 2
Diagram of EI source.. ............................................................................. 6
The (a) EI and (b) self-CI mass spectra of benzene. ................................ 9
Comparison between the ZEKE and the photoelectron spectroscopies by
one-photon and resonance-enhanced multiphoton ionization schemes. .. 15
Schematics of the ZEKE/MATI process. ................................................. 17
Ionization threshold and higher states having their own individual Rydberg
series. High Rydberg states near n=100 are converted into special ZEKE
states which have an abnormal lifetime by external fields....................... 18
Experimental scheme for perpendicular TOF mass spectrometer. 4 mm × 50
mm size slit-electrode assemblies were used to enhance ion collection
efficiency. ................................................................................................. 20
Timing sequence for various pulses adopting according to (a) broadband
MATI scheme or (b) narrowband MATI scheme. .................................. 22
Schematic drawing of the four-wave mixing Kr cell and VUV MATI
instrument. (a) top and (b) side views. ................................................... 24
Schematic drawing of the four-wave mixing Hg cell and VUV MATI
instrument. (a) top and (b) side views. ................................................... 25
xii
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
2.1
2.2
2.3
2.4
2.5
2.6
Schematic diagram of the double focusing mass spectrometer with reversed
geometry (VG ZAB-E) modified for photodissociation study. The inset
shows the details of the electrode assembly. .......................................... 37
Circuit diagram devised to pulse the electron beam. MA3104-003D indicates
the main board circuits controlling the ion source of ZAB-E mass
spectrometer . ........................................................................................... 39
Energy diagram of the benzene molecular ion. The lowest reaction threshold
(E0) is 3.66 eV for C6H6+•→C6H5
++H•. ktot denotes the total rate constant
for dissociation in the ground electronic state predicted from previous studies.
.................................................................................................................. 41
PD-MIKE profile for the production of C4H4+• from the benzene ion at
357nm obtained with 2.1kV applied on the electrode assembly. Experimental
result is shown as filled circles. Reproduction of the profile using the rate
constant distribution centered at 6.3×107 s-1 obtained by experimental data is
shown as the solid curve. The positions marked A and B are the kinetic
energies of products generated at the position of photoexcitation and after
exiting the ground electrode, respectively. ............................................. 44
The total RRKM dissociation rate constant of benzene ion as a function of
the internal energy calculated with molecular parameters in ref. 8. The
internal energies corresponding to the dissociation rate constants of
( 5 . 5±1 .1 )×107 and (5±3 )×106 s-1 for PDs at 357 and 488.0 nm,
respectively, are marked. ........................................................................ 46
Pressure dependences of the precursor (C6H6+•) intensity (–––) and
photoproduct (C4H4+•) intensities at 357 (·····) and 488.0 (–––) nm. Pressure
in the CI source was varied continuously to obtain these data. The abscissa
shows the pressure read by an ionization gauge located outside of the source.
xiii
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
The inside source pressures estimated using eqn. (1) at three ionization
gauge readings are marked. The scale for the precursor intensity is different
from that for photoproduct intensities. ..................................................... 54
2.7
2.8
2.9
2.10
2.11
The ratios of molecular ion intensities generated by charge exchange
ionization (CI) with benzene ion and by electron ionization (EI) are plotted
as a function of the sample ionization energy. • and o are for CI intensities
measured at the ion source pressure of 0.013 and 0.09 Torr, respectively.
.................................................................................................................. 56
Schematic diagram of the double focusing mass spectrometer with reversed
geometry (VG ZAB-E). The inset shows details of the first collision cell
assembly modified for charge exchange study. ...................................... 62
MIKE spectrum of the C6H6+• primary ion generated by EI. CS2 was
introduced into the second collision cell. The acceleration energy for C6H6+•
was 4002 eV. The collision cell potential was 3902 V. The peak types are
denoted. The peak marked 77+(MID) is due to the metastable ion
decomposition of C6H6+• to C6H5
+ occurring in the field-free region between
the magnetic and electric sectors. ........................................................... 66
Mass spectra obtained under the single-focusing condition. The acceleration
energy in the source was 4004 eV and the collision cell potential was 3929 V.
C6H6 and CS2 were introduced into the ion source and the first collision cell,
respectively. C6H6 was ionized (a) by EI and (b) by CI at the 0.02 Torr
source pressure. (c) C6D6 and CH3Cl were introduced into the ion source and
the first collision cell, respectively and C6D6 was ionized by EI. The
instrument was tuned to maximize the type III ion signals. The types of the
major signals are denoted. ...................................................................... 68
MIKE spectrum recorded by setting the magnetic field to transmit the
xiv
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
C32S2+•(III) ion in Figure 3(b) and scanning the electric sector. ............ 69
2.12
2.13
2.14
2.15
2.16
Relative yields of the reagent gas ions, I(A+•)/I(C6H6+•), vs. the primary ion
translational energy. Benzene ions were generated by CI at (a) 0.02 Torr
(CI1) and at (b) 0.1 Torr (CI2) source pressures. Charge exchange ionization
was done in the first collision cell. For consistency, the instrument was tuned
to maximize the primary ion signal. In (a), CH3F+• and CH4+• signals were
hardly detectable at low energy while CS2+•, CH3Cl+•, CH3F+•, and CH4
+•
were not detectable at low energy in (b). ............................................... 71
Partial mass spectrum of C6H5Cl generated by 20 eV EI recorded under the
single focusing condition with 4006 eV acceleration energy is shown in (a).
(b) and (c) are mass spectra in the same range recorded with CH3Cl in the
collision cell floated at 3910 and 3960 V, respectively. Type II signals at m/z
49.3 and 50.3 in (b) and at m/z 49.6 and 50.6 in (c) are due to collision-
induced dissociation of C6H5Cl+• to C4H2+• and C4H3
+, respectively. The
peaks at m/z 50.6 in (b) and at m/z 50.8 in (c) are due to collision-induced
dissociation of C6H5+
to C4H3+. .............................................................. 82
Ion kinetic energy spectrum recorded by setting the magnetic field to
transmit the type III CH335Cl+• ion in Fig. 2 (b) and scanning the electric
sector. ...................................................................................................... 83
Partial mass spectrum obtained under the single focusing condition with
C6H5Br and CH3Br introduced into the ion source and collision cell,
respectively. C6H5Br was ionized by 20 eV EI and acceleration energy was
4008 eV. Collision cell was floated at 3907 V. ....................................... 85
Partial mass spectrum obtained under the single focusing condition with
C6H5CN and CH3Cl introduced into the ion source and collision cell,
respectively. C6H5CN was ionized by 20 eV EI and acceleration energy was
xv
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
4007 eV. Collision cell was floated at 3910 V. Type II signals at m/z 49.3,
50.3, and 51.3 are due to collision-induced dissociation of C6H5CN+• to
C4H2+•, C4H3
+, and C4H4+•, respectively. Those at m/z 49.6 and 50.6 are due
to collision-induced dissociation of C6H4+• to C4H2
+• and C4H3+, respectively.
.................................................................................................................. 88
2.17
3.1
3.2
3.3
3.4
3.5
Ion kinetic energy spectra recorded by introducing C6H5Br+• ((a) and (b))
and C6H5CH3+• ((c)) in the second cell filled with CH3Br. The molecular ions
were accelerated to 4 keV in the ion source. The second collision cell was
floated at (a) 3910, (b) 3943, and (c) 3910 V. Arrows indicate the expected
positions for ions from collision gases generated by charge exchange with
the precursor ions. The major peaks appearing at 3957 and 3974 eV in (a)
and (b), respectively, are due to collision-induced dissociation of C6H5Br+• to
C6H5+. The major peak appearing at 3960 eV in (c) is due to unimolecular
dissociation of C7H8+• to C7H7
+ occurring outside the collision cell, but
between the magnetic and electric sectors. ............................................ 89
The number of photoions by ion chamber currents measured as function of
the pressure of No/He. Voltage between two electrodes in ion chamber is 50
V. ............................................................................................................ 106
Schematic diagram for four-wave difference frequency mixing in Kr gas. 108
Schematic diagram for four-wave sum frequency mixing in Hg vapor. .. 110
Apparatus for VUV generation by four-wave sum frequency mixing in Hg
vapor. The laser beams were aligned off-center of LiF lens to separate the
VUV light from the residual UV and visible lights at the interaction region
with the molecular beam. ....................................................................... 111
Schematic diagram of the experimental apparatus including the Hg cell with
xvi
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
a cone type glass capillary, monochromator, and photoionization chamber.
.................................................................................................................. 114
(a) Spectral profile of VUV generated by frequency tripling in Hg at ω1
~312.8 nm. (b) and (c) show spectral profiles of VUV generated by FWSM
via 71S0 - 61S0 transition in Hg recorded using the Au plate monitor and
photoionization of benzene, respectively. PHg ∼ 1 torr and PHe ∼ 2 torr. .. 116
3.6
3.7
4.1
4.2
4.3
4.4
4.5
4.6
4.7
4.8
Spectral profiles of VUV generated by FWSM via 61D2 - 61S0 measured
using (a) the Au plate monitor and (b) photoionization of benzene. PHg ∼ 1
torr and PAr ∼ 0.5 torr. ............................................................................. 118
Lowering of the ionization potential due to an electric field. ................ 124
Illustration of the Franck-Condon principle and intensity distributions for
small and large displacement, respectively. ........................................... 130
Ground state one-photon MATI spectra recorded by monitoring (a)
C6H535Cl+• and (b) C6H5
37Cl+•. ............................................................... 133
Ground state one-photon MATI spectra recorded by monitoring (a)
C6H579Br+• and (b) C6H5
81Br+•. .............................................................. 139
Ground state one-photon MATI spectrum recorded by monitoring C6H5I+•.
.................................................................................................................. 142
Ground state one-photon MATI spectrum recorded by monitoring C6H5F+•.
.................................................................................................................. 145
B~ 2B2 state one-photon MATI spectrum of C6H535Cl+•. The x-axis of the
inset, ion energy, denotes energy scale referred to the position of the 0-0
band. ....................................................................................................... 149
B~ 2B2 state one-photon MATI spectrum of C6H579Br+•. The x-axis of the
xvii
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
inset, ion energy, denotes energy scale referred to the position of the 0-0
band. ....................................................................................................... 152
4.9
4.10
4.11
4.12
4.13
4.14
4.15
B~ 2B2 state one-photon MATI spectrum of C6H5I+•. .............................. 154
Equilibrium geometries of (a) C6H5F and C6H5F+• and (b) C6H5Cl and
C6H5Cl+• calculated at the B3LYP/6-311++G** level. Atomic displacements
upon ionization are drawn as broken arrows in the drawings of the neutrals.
The 6a eigenvectors of the cations are drawn as arrows in the drawings of the
cations. Bond lengths in Å and angles in degree. ................................... 156
One-photon MATI spectrum of p-C6H4F2 recorded by monitoring p-
C6H4F2+• in the ground electronic state. The x-scale at the top of the figure
corresponds to the vibrational frequency scale for the cation. ............... 168
One-photon MATI spectrum of m-C6H4F2 recorded by monitoring m-
C6H4F2+• in the ground electronic state. The x-scale at the top of the figure
corresponds to the vibrational frequency scale for the cation. ............... 173
One-photon MATI spectrum of o-C6H4F2 recorded by monitoring o-C6H4F2+•
in the ground electronic state. The x-scale at the top of the figure
corresponds to the vibrational frequency scale for the cation. Its origin is at
the 0-0 band position. Spectrum in the 100~800 cm-1 region magnified by
30 is shown as an inset to demonstrate the quality of the MATI spectra
obtained in this work. ............................................................................. 176
Equilibrium geometries of (a) neutral and (b) cation of p-C6H4F2. Arrows
in (a) indicate atomic displacements upon ionization magnified by 10.
Arrows in (b) indicate the eigenvector of the mode 6 of the cation. ...... 180
One-photon MATI spectrum of C6H5C≡CH recorded by monitoring
C6H5C≡CH+ in the ground electronic state. The x-scale at the top of the
xviii
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
figure corresponds to the vibrational frequency scale for the cation whose
origin is at the 0-0 band position. Spectrum in the 50 ∼ 2500 cm-1 region
magnified by 15 is shown as an inset to demonstrate the quality of the MATI
spectrum obtained in this work. ............................................................... 190
4.16
5.1
5.2
5.3
5.4
One-photon MATI spectrum of C6H5C≡N recorded by monitoring
C6H5C≡N+ in the ground electronic state. The x-scale at the top of the figure
corresponds to the vibrational frequency scale for the cation whose origin is
at the 0-0 band position. Spectrum in the 50 ∼ 2200 cm-1 region magnified
by 15 is shown as an inset to demonstrate the quality of the MATI spectrum
obtained in this work. ............................................................................. 197
One-photon MATI spectrum of C6H6 recorded by monitoring C6H6+ in the
ground electronic state. The x-scale at the top of the figure corresponds to
the vibrational frequency scale for the cation. Its origin is at the 0-0 band
position. Spectrum in the 100~2100 cm-1 region magnified by 30 is shown
as an inset to demonstrate the quality of the MATI spectrum obtained in this
work. Relative intensity of the peak marked by asterisk (*) changed with
the beam expansion condition. ............................................................... 216
One-photon MATI spectrum of C6D6 recorded by monitoring C6D6+ in the
ground electronic state. The x-scale at the top of the figure corresponds to
the vibrational frequency scale for the cation. Its origin is at the 0-0 band
position. Spectrum in the 100~2100 cm-1 region magnified by 40 is shown
as an inset to demonstrate the quality of the MATI spectrum obtained in this
work. ....................................................................................................... 221
Photoionization spectrum of C6H6 measured as a function of the VUV
photon energy (in cm-1). ......................................................................... 236
One-photon MATI spectrum of C6D6 recorded by monitoring C6D6+ in the
xix
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
excited electronic state B~ 2E2g. The x-scale at the top of the figure
corresponds to the vibrational frequency scale for the cation, or the ion
internal energy. Its origin is at the 0-0 band position. Spectrum in the
0~1900 cm-1 region magnified by 10 is shown as an inset to demonstrate the
quality of the MATI spectrum obtained in this work. ............................ 237
One-photon MATI spectrum of C6H6 recorded by monitoring C6H6+ in the
excited electronic state B~ 2E2g. The x-scale at the top of the figure
corresponds to the vibrational frequency scale for the cation, or the ion
internal energy. Its origin is at the 0-0 band position. Spectrum in the
0~1900 cm-1 region magnified by 10 is shown as an inset to demonstrate the
quality of the MATI spectrum obtained in this work. ............................ 242
5.5
5.6
5.7
5.8
One-photon MATI spectrum of C6F6 recorded by monitoring C6F6+ in the
ground electronic state. The x-scale at the top of the figure corresponds to
the vibrational frequency scale for the cation whose origin is at the 0-0 band
position. Spectrum in the 50 ~ 1900 cm-1 region magnified by 20 is shown as
an inset to demonstrate the quality of the MATI spectrum obtained in this
work. ....................................................................................................... 249
The optimized geometries for the (a) 2E1g (D6h), (b) 2B2g (D2h), and (c) 2B3g
(D2h) states of C6F6+ calculated at the B3LYP/6-311++G (2df) level. Values
in parentheses in (a) are the bond lengths in the neutral. ....................... 252
The Jahn-Teller potential energy surfaces along each normal coordinate of
the four e2g modes of C6F6+ in the ground electronic state. Only the portions
in the 2B3g side are drawn. ...................................................................... 259
xx
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
Chapter 1
Introduction
1.1 Why the Excited State?
Controlling the outcome of chemical reaction has been a long-standing goal of
chemists. The reaction dynamics has attempted to understand a chemical reaction
at the single molecule level and considerably been progressed together with rapid
development in laser technology. Particularly, photochemistry has long been a
prominent area of reaction dynamics.1-3 Many experimental and theoretical
approaches have been proposed to comprehend and even control a photochemical
process.4-13
In general, most photochemical processes involve first the electronic transition
by absorption or scattering of one or more photons, followed by nuclear motion
on the excited electronic state(s), crossed back to the ground electronic state, and
finally relaxed into minima representing photoproducts. Practically, scattering and
multiphoton processes require the intense light fields available only from lasers.
Multiphoton and light scattering processes are called nonlinear optical phenomena,
because unlike stimulated absorption, they do not depend linearly on the
excitation power.
Most excited electronic states have only a short lifetime, namely the excess
energy of an excited state can be dissipated through unimolecular photophysical
processes which are illustrated schematically in Fig. 1.1. Followings are some
important processes on the excited electronic states:
1
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
S1
S0
T1
hv
P
F
ISC
VRVR
IC
VR
E
PhotoproductsPhotoproducts
Fig. 1.1 Illustrated diagram of photophysical processes. Radiation by straight arrows (F=fluorescence,
P=phosphorescence) and radiationless processes by wavy arrows (IC=internal conversion, ISC=intersystem
crossing, VR=vibrational relaxation).
2
Copyright(c)2002 by Seoul N
ational University L
ibrary. All rights reserved.(http://library.snu.ac.kr) 2004/01/28 17:41:01
• Fluorescence : stimulated and spontaneous emission.
• Phosphorescence : lower energy spontaneous emission occurring from
longer-lived excited states, usually triplet multiplicity.
• Intersystem Crossing : transitions between excited electronic state manifolds
of different spin multiplicity.
• Internal Conversion : transitions from higher to lower lying excited states or
ground state, having the same spin multiplicity.
• Vibrational Relaxation : dissipation such as heat to the vibrational ground
state.
• Photochemical Reactions : photodissociation, photosynthesis, photoinduced
electron-transfer.
After comprehending about the creation and evolution of excited electronic
states, selectively preparing and storing these excited state ions may be especially
useful for the investigation of the state-specific unimolecular and bimolecular
reactions, even though it was difficult to prepare state-selected ions due to various
limitations. Furthermore, the study of the structure and dynamics of excited
electronic states14-18 will be useful to predict a photochemical process, to induce
new reaction control scheme, even to completely control a chemical reaction,
because of their important roles in chemical reactions.
1.2 Mass Spectrometry
Mass spectrometry19-21 has progressed considerably and rapidly during the last
decade and raised to an outstanding position for analytical methods. The
development of modern mass spectrometers has been related to advances in laser
and signal-processing technologies, ion optics and detection, fast and high-
3
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
performing electronics, etc. This has been particularly true for the time-of-flight
(TOF) mass spectrometry,20 which recently plays an increasingly important role in
all of the biological research areas as well as ion reaction dynamics field.
Table 1.1 Classification of components consisting of mass spectrometer.20
Ionization Methods Mass Analyzers Detectors
Electron ionization (EI) Magnetic (B) Faraday cup
Chemical ionization (CI) Double-focusing (EB) Electron multiplier (EM)
Fast atom bombardment (FAB) Ion cyclotron resonance (ICR) Photon multiplier (PM)
Field Ionization (FI) Quadrupole (Q) Microchannel plate (MCP)
Photoionization (PI) Quadrupole ion trap (ITMS) Array detectors
Multiphoton ionization (MPI) Time-of-flight (TOF) Image currents
Electrospray (ESI) Fourier transform (FTMS)
Matrix-assisted laser
desorption/ionization (MALDI)
Mass spectrometers can be characterized by the ionization sources, mass
analyzers, and detectors that are used. In general, they have been made up of
appropriate components according to the features of interesting system and
continuous/pulsed types, even possible to hybrid. For example, double focusing
mass spectrometer with reversed geometry (EI/CI-BE-PM) and TOF mass
spectrometer (PI/MATI-TOF-MCP) are used to study state-selected benzenoid
molecular ions.
4
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
1.2.1 Ionization Methods
A molecule (M) can be generally ionized through interaction with the
energetic electrons, particles, or photons.
−+ +→+ eMEM (1.1)
Here, ionization energy (IE) is defined as the minimal energy required on removal
of an electron in molecule. Ionizations can occur through various methods such as
electron ionization (EI), charge exchange ionization (CI), photoionization (PI),
electrospray ionization (ESI), matrix-assisted laser desorption ionization
(MALDI) that can be chosen, depending on the volatility, thermal stability, and
physical state of molecule.21
• Electron Ionization (EI): This is the oldest and most widely used method in
organic mass spectrometry. As shown in Figure 1.1, sample molecules (M) be
volatilized into the ionization chamber are bombarded with electrons which are
obtained from a heated filament in vacuum and accelerated by voltage (V), thus
have energy eV, where e is the electronic charge. Standard mass spectra are
usually obtained at 70 V because of maximum ionization efficiency and good
reproducibility for most organic molecules near this voltage.
(1.2) −•+− +→+ eMeM 2
However, this technique induces extensive fragmentation due to excess internal
energy in molecular ion formed by high electron energy, even though it can be
useful to identify its structure. Therefore, EI method capable to form the
molecular ion with large internal energy is not adequate in state-selective or
energy-selective reaction study.
• Charge Exchange Ionization22 (CI): The interaction of a reactant ion (A+) with
a neutral molecule (M) may lead to charge exchange, if the recombination energy
5
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
Ion accelerating potential
Feedback by trap current
Filament heater potential
Electron accelerating
potential
Repeller
e-
Filament
Trap
M+•
Fig. 1.2 Diagram of EI source.
6
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
(RE) of the reactant ion is greater than the ionization energy (IE) of the neutral
molecule.
AMAM +→+ ++ =∆ (1.3) )RE(AIE(M)E +−
The RE of A+ is defined as the electron affinity and normally has the same
numerical value of the IE to ground state of A+. In some cases, ions of reactant or
product may be formed in an excited electronic state. A selection of RE is
presented in Table 1.2. Molecular ion formed by the CI will have an internal
energy determined by the exothermicity (∆E) and hence, the product distribution
from the CI will depend on the ∆E of the reaction. This is ‘soft’ method of
ionization, resulting in little fragmentation as shown in Figure 1.3.
Charge exchange ionizations of polyatomic molecules could have been
utilized to measure the rate constants of unimolecular reactions, namely, to
energy-selective study. However, although the CI can forms the molecular ions
with less excess of internal energy than the case in the EI, it is not effective
method to select a state of molecule.
• Photoionization (PI): Photoionization method (PI) employs photons to ionize a
molecule. The photons are generated by high-power lamps or by lasers. Ionization
processes depend on the photon energy, with either direct transition to the ionic
state A+ or multiphoton absorptions, namely,
−+ +→ν+ eMhM 1 (1.4)
or
−+ +→ν+ eMnhM 2 . (1.5)
Practically, single-photon ionization occurs in the vacuum ultraviolet (VUV)
region and is the most general and cleanest photoionization method. In addition,
7
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
Table 1.2 Recombination energies of gaseous ions.
Ion RE, eVa
Kr+•
2P3/2 14.0
2P1/2 14.7
CO+• 14.0
CO2+•
13.8
Xe+• 2P3/2 12.1
2P1/2 13.4
COS+• 11.2
CS2+•
X~ 10.08
A~ 12.84
C6H6+• 9.243
a From ref. 22
8
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
30 40 50 60 70 80m/z
Inte
nsity
(arb
itrar
yun
its)
30 40 50 60 70 80m/z
Inte
nsity
(arb
itrar
yun
its)
(b)
(a)
Fig. 1.3 The (a) EI and (b) self-CI mass spectra of benzene.
9
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
the resolution of photon energy reached to rotational levels as well as the
vibrational levels of molecular systems. However, regardless of extremely high
energy resolution achieved by development of laser technique, it still remains the
problem in the resolution of photoionization because of removal electron with
significant kinetic energy. For the study of state-selected molecular ion, new
ionization technique which produces the threshold photoelectron with near zero
kinetic energy by one-photon PI, would be required.23
1.2.2 Mass Analyzers
The ions which are produced from an ion source, can be separated to their
masses by many different analyzers. Scanning analyzers successively transmit
ions of different masses along a time scale, which are either magnetic or
quadrupole analyzers. While, some simultaneously allow the transmission of all
ions such as the time-of-flight, the ion trap, the ion cyclotron resonance analyzers.
Also, the analyzers can be distinguished by three main characteristics, the upper
mass limit, the transmission, and the resolution. Considering the purpose and the
interested system, instruments combined several analyzers in sequence, namely,
the tandem mass spectrometers (MS/MS), are increasingly common because a
mass spectrum resulting from the dissociation of an ion selected in the first
analyzer can be obtained.
1.2.3 Theory of Mass Spectra
Theory of mass spectra, or RRKM (Rice-Ramsperger-Kassel-Marcus)-QET
(quasi- equilibrium theory), has been successful in explaining fragmentation of
polyatomic (four atoms or more) molecular ions produced by various means and
mass spectra formed thereby. One of the main assumptions of RRKM-QET is that
the internal energy, either electronic or vibrational, acquired at the time of
10
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
molecular ion formation redistributes rapidly prior to fragmentation.24 In
particular, rapid redistribution of the former means that internal conversion from
the excited electronic states to the ground state occurs efficiently for polyatomic
ions such that all the dissociation reactions occur in the latter state, even though
cases to the contrary are frequently observed for simpler diatomic and some
triatomic ions. From the very inception of the theory, efforts have been made to
find evidences against the assumption of rapid internal conversion such as the
occurrence of dissociation in an excited electronic state, or ‘isolated’ state, and
rapid radiative decay of the electronic energy. Various mass spectrometric and
spectroscopic techniques have been used for this purpose such as photoelectron-
photoion coincidence spectrometry,25-34 tandem mass spectrometry utilizing
collision-induced dissociation35-38 or photodissociation,39-44 and emission
spectroscopy.45-48 Tens of cases of isolated state dissociation have been reported,
most abundant having been dissociations in repulsive electronic states.26-33,39-43
Rapid dissociations in such states can compete effectively against the internal
conversion and the nonstatistical nature of the reactions is manifest in the
experimental breakdown graph, kinetic energy release distribution, or dissociation
anisotropy. Cases25,36-38 of slow internal conversion have been reported also,
CD2O+• ( A ) → CDO~ + + D• being the best known example. To measure the
lifetime, both radiative and nonradiative, of an excited electronic state lying below
the dissociation threshold, optical emission and laser-induced fluorescence
spectroscopies have been used.45-48 When the transition from the first excited
electronic state to the ground state is electric dipole allowed, absence of emission
is an indication of very rapid internal conversion. Similar conclusions have often
been drawn45-48 when emission is absent from an excited state whose radiative
transitions to the lower states are electric dipole forbidden, such as the B~ 2E2g
state of benzene cation, even though extremely slow decay of such a state is also
11
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
compatible with the experimental observation.14-18
1.2.4 Long-lived Excited Electronic State
There are some cases where the hypothesis of rapid conversion from an
excited to ground electronic states prior to dissociation is not valid. The most
frequently observed are the cases of direct dissociation in a repulsive excited
electronic state. 26-33,39-43 Predissociation via bound-to-repulsive transition has
been observed for dissociation of simple molecular cations also. Dissociation in a
bound excited state, or an ‘isolated’ electronic state, has been hardly observed in
ionic cases.14-18 Hence searching for isolated electronic states of polyatomic
cations remains a mission yet to be fulfilled.
In our recent investigation on photodissociation of benzene cation,14 We found
evidences for a very long lifetime (20 µsec or longer) of its B~ 2E2g state. In a
subsequent work,15 a method based on charge exchange in collision cells of a
modified double focusing mass spectrometer was developed for routine search for
long-lived excited states with conventional mass spectrometry. Then, the overall
scheme was applied to find long-lived excited electronic states of the molecular
ions of some benzene derivatives generated by electron ionization.17 It was found
that the B~ 2B2 excited electronic states of chlorobenzene, bromobenzene,
benzonitrile, and phenylacetylene cations had very long lifetimes (tens of
microsecond or longer). These are the states generated by removal of an electron
from the in-plane nonbonding p orbitals of halogens or in-plane π orbitals of the
triple bonds. Halogen p orbitals or π orbitals of triple bonds in ethene derivatives
also split into two just as in the benzene derivatives, one in-plane and the other
out-of-plane. All of the long-lived excited electronic states of benzenoid
molecular cations investigated by mass spectrometry are described with details in
next chapter.
12
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
1.3 MATI and ZEKE spectroscopies
1.3.1 Historical Background
Spectroscopy of molecular cations is of important for the study of combustion,
atmospheric chemistry, cosmochemistry, etc.49-51 Analysis of a vibrational
spectrum, with rotational resolution in particular, provides useful structural
information. Availability of reliable ionic spectral database is also essential for the
development and test of quantum chemical methods especially for the open shell
systems.
Unlike neutrals, it is difficult to perform infrared or Raman spectroscopy to
obtain vibrational spectra of molecular cations.52,53 Optical emission spectroscopy
is not generally applicable either because emission is not observed for most of the
polyatomic ionic species.54 Probably the simplest way to obtain vibrational
information of a molecular cation is to record the photoelectron spectrum (PES)55
of the corresponding neutral. However, the resolution of this technique is rather
poor (typically ∼10 meV, or 80 cm-1, for high resolution photoelectron
spectrometer)56-58 and hence is not adequate to obtain detailed vibrational
information. A different form of PES on the basis of the photoionization
experiments which already utilized a photon monochromator, called threshold
photoelectron spectroscopy (TPES),23 was developed with slightly better
resolution than that of the PES. The PES and TPES spectra are linked by the
energy conservation law:
)(eE)(ME)(eE)(ME)(MEhv kikki−+−++ +≈++= (1.6)
where Ei (M+) is the internal energy of the photoion, Ek (M+) is the kinetic energy
of the ion, while Ek (e-) is the kinetic energy of the departing electron. By
conservation of momentum, Ek (M+) ≈ 0 and becomes identically zero in TPES for
13
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
which Ek (e-) = 0. Namely, this method is to detect the electrons with zero kinetic
energy produced from an ionic state when the photon energy matches an
ionization threshold. Since these electrons are prompted with no kinetic energy,
the angular dispersion is negligible, which enables to increase the collection
efficiency of electron energy analyzer. Most of the work in TPES has been done
with synchrotron light sources, which are excellent sources of continuously
tunable short-wavelength light. However, TPES of which resolution is limited
largely by the synchrotron light source, is also not sufficient method to well-
resolve the molecular ionic states. A related technique in parallel with the
development of TPES using synchrotron radiation is the resonance-enhanced
multiphoton ionization (REMPI)-photoelectron spectroscopy (PES)59,60 which
ionizes a neutral by REMPI using coherent laser pulses of high sensitive and
resolution and determines the ion internal energy by measuring the electron
kinetic energy. Even though REMPI-PES has some advantages over PES such as
the intermediate state selection, spectral resolutions of the two are comparable.
A new technique for extreme high-resolution photoelectron spectroscopy which
combines REMPI and TPES, was developed by Schlag and coworkers. The
technique, called the pulsed-field ionization zero kinetic energy (PFI-ZEKE)61-63
photoelectron spectroscopy, takes advantages of both the high photon energy
resolution and the pulsed excitation provided by tunable pulsed dye lasers, in
addition, high photoelectron energy resolution due to the delayed pulsed field
ionization used for threshold detection. Here, one can detect the field-ionized
electrons from high Rydberg states or the corresponding positive ions. The latter
type of detection scheme first developed by Johnson has come to be known as
mass-analyzed threshold ionization (MATI) spectroscopy.64-66 Even though the
resolution of MATI is inferior to that of ZEKE at the moment, capability to identify
the ionic species responsible for a spectral peak is the advantage of the former.
14
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
M
M*
M+
hv1
hv2hvHe(I)
e-
e-
PES
REMPI-PES
M
M*
M+
hvVUV
hv1
hv2
e- (Ek=0)
ZEKE (TPES)
Fig. 1.4 Comparison between the ZEKE and the photoelectron spectroscopies
by one-photon and resonance-enhanced multiphoton ionization schemes.
15
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
1.3.2 General Principles
ZEKE spectroscopy67,68 was originally developed as extremely high resolution
version of TPES, adopting the pulsed nature of REMPI to introduce a delay time
between excitation to high Rydberg states of neutrals and extraction of the
Rydberg electrons by small pulsed electric field. The basic idea of ZEKE was that
after any electrons formed with small kinetic energy would drift away from the
excitation position during the delay time, electrons with zero kinetic energy only
leaved to await the extraction and detection. This results the stimulated study on
the real mechanism of ZEKE process and Schlag and coworkers found that the
responsible ZEKE signals come not from electrons formed by photoionization at
just above threshold but Rydberg electrons which are bound weakly to very high n
Rydberg states just below the ionization threshold by the pulsed field ionization.
Namely, high Rydberg state electrons generated by photon absorption in field-free
region, are so loosely bound that it is easily extracted by pulsed field ionization.
This ZEKE process is shown schematically in Fig. 1.5. The resolution in ZEKE
spectroscopy is generally determined by not the bandwidth of exciting laser but
the depth of field ionization. In the presence of the electric field, ionization
threshold is lowered by FE α=∆ and the value of is between 4 and 6,
depending on whether the field ionization is diabatic or adiabatic. While, it must
be remembered not only that Rydberg series of n>100 at ionization threshold lead
up to a narrow set of ZEKE states in a typical 8 cm
α
-1 bandwidth by external field,
but also that this is true for all the individual excited ionic states above the IP, as
shown in Fig. 1.6. Hence, the electrons from the ZEKE states of individual ionic
states are extracted by the pulsed field ionization and these ZEKE signals
recorded as function of photon energy are called the ZEKE spectrum.
Many efforts have been made to understand the mechanism of Stark ionization
16
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
High Rydberg stateHigh Rydberg state
M+e-
hvhvCoulomb Coulomb potentialpotential
EE
M+e-
Low Rydberg stateLow Rydberg state
e-M+e-
M+M+ e-Continuum (Cation)
Direct ion(electron)Direct ion(electron)
++ ––
ee--
ee--
ee--
RRMM++
MM++
RR
RR
++ ––
ee--
ee--
ee--
MM++
MM++
MM++
MM++ee--
++ ––
MM++ee--ZEKEZEKE
MATIMATI
0V
Laser pulse
Delay time, ~µs
Spoil Field
PFI
TOF spectrumdirect electrons
PFIPFI--ZEKE signal ZEKE signal
Pulsed field ionization (PFI)
Fig. 1.5 Schematics of the ZEKE/MATI process.
17
Copyright(c)2002 by Seoul N
ational University L
ibrary. All rights reserved.(http://library.snu.ac.kr) 2004/01/28 17:41:01
g
n=100
n=50
n=30
E=hv
ZEKE states
ZEKE states
ZEKE states
ZEKE states
ZE
KE
spec
trum
n=200Ground
state
v1
v2
v3
0-0
v 1v 2
v 3
Fig. 1.6 Ionization threshold and higher states having their own individual Rydberg series. High Rydberg states
near n=100 are converted into special ZEKE states which have an abnormal lifetime by external fields.
18
Copyright(c)2002 by Seoul N
ational University L
ibrary. All rights reserved.(http://library.snu.ac.kr) 2004/01/28 17:41:01
of Rydberg state. ZEKE state surviving for several µs, has been another issue
and investigated both theoretically and experimentally, even though not
conclusive thought.68-70
1.3.3 One-photon MATI Scheme and Apparatus
Two color 1+1’ excitation is the most widely used scheme in ZEKE or MATI
spectroscopy, namely excitation of a molecule to a high-lying Rydberg state via
an intermediate state, which is followed by pulsed-field ionization.61-66 In this
scheme, the measured spectral transitions would be governed by the nature of
the intermediate state. This sometimes helps spectral assignment. The fact that
detailed spectroscopic information on the excited electronic states is not
available for most of the molecules is one of the difficulties in routine
application of the 1+1’ scheme. Even though difficulty lies in the fact that the
first excited electronic states are located 6.2 eV (200 nm) or more above the
ground states and not accessible by one photon absorption of a commercial dye
laser output in most of the cases. Also to be mentioned in that excitation to these
states often result in diffuse spectra either due to rapid dissociation or relaxation.
In such cases, one can not expect good ZEKE or MATI signals when the 1+1’
scheme is used. Recently, we developed one photon MATI scheme using
vacuum ultraviolet irradiation and studied the spectra of the polyatomic ions.16
Presence of an appropriate intermediate state is no longer a requirement in this
scheme because the molecules are prepared in the Rydberg states by direct one-
photon absorption from the ground state. In addition, every possible dipole
allowed transition makes up complex spectra, which show rich vibrational
structures.
The TOF mass spectrometer consists of two differentially pumped chambers
and a 60 cm flight tube with a dual microchannel plate detector equipped at the
19
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
skimmer
VUV laserE3
E2
E1
temperature-controlled pulsed nozzle
G
MCP
TOF
Fig. 1.7 Experimental scheme for perpendicular TOF mass spectrometer. 4
mm × 50 mm size slit-electrode assemblies were used to enhance ion
collection efficiency.
20
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
end. A sample seeded in He at the stagnation pressure of ∼2 atm was
supersonically expanded through a temperature–controlled pulsed nozzle (dia.
500 µm, General Valve) and introduced to the ionization chamber through a
skimmer (dia. 2 mm, Beam Dynamics) placed about 3 cm downstream from the
nozzle orifice. The background pressure in the ionization chamber was typically
∼ 10-8 Torr. The VUV laser pulse was collinearly overlapped with the molecular
beam in a counter-propagation manner to maximize the laser-molecular beam
interaction volume. Instead of the usual circular aperture, 4 mm × 50 mm size
slit-electrode assemblies were used to enhance ion collection efficiency, Fig. 1.7.
Spoil field of 0 ~ 0.2 V/cm was applied in the ionization region to remove
directly produced ions. To achieve pulsed-field ionization (PFI) of neutrals in the
ZEKE state, an electric field of 10 ~ 80 V/cm was applied with the field
direction perpendicular to that of the molecular and laser beams. Then, ions were
accelerated, flied through a field-free region, and were detected by a dual
microchannel plate (MCP) detector. It is well known that the spoil field must be
kept low to obtain a MATI spectrum with good resolution, which requires use of
a long time delay between VUV absorption and PFI. Use of a time delay longer
than 10 ns led to rapid decay of MATI signal in our apparatus. We could
lengthen the lifetime of the VUV-excited neutrals tremendously, however, by
applying a short pulse of scrambling field at the laser irradiation time. This
allowed the use of a very long delay time, ~ 25 µs, and low spoil field. Timing
sequence for various pulses is shown in Fig. 1.8. MATI signal detected by MCP
was preamplified and A/D converted by a digital storage oscilloscope (LeCroy
9370L). Either the full time-of-flight mass spectrum or selected regions of the
spectrum as needed were transferred to a personal computer in real time.
The tunable VUV light was generated by resonant four-wave mixing (ωVUV±
= 2ωUV ± ωS) in Kr and Hg in the region of 130 ∼ 143 nm and 106.6 ∼ 117 nm,
21
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
(a) Broad-band MATI scheme
E1
E2
E31500V
1250V
-0.3V
photon
25 µs
(b) Narrow-band MATI scheme
E1
E2
E3 20V
2000V
-0.3V
photon
25 µs
1750V
Fig. 1.8 Timing sequence for various pulses adopting according to (a)
broadband MATI scheme or (b) narrowband MATI scheme
22
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
respectively. VUV generation in Kr will be described with details in Chapter 4.
A schematic diagram of the experimental apparatus is in Fig. 1.9. Briefly, to
excite the Kr 5p[1/2]0 – 4p6 transition, the light at 212.5 nm (∼0.5 mJ/pulse) was
generated by frequency tripling of 637.6 nm output of a dye laser (Continuum
ND6000) pumped by the second harmonic of an Nd:YAG laser (Continuum
PL8000). Another dye laser output (410 ∼ 570 nm) pumped by the third
harmonic of the second Nd:YAG laser (Continuum Surelite II) was combined
with the 212.5 nm light and loosely focused with a fused silica lens (f = 50 cm)
in the Kr cell to generate the VUV light tunable in the 130 ∼ 143 nm range. A
MgF2 lens (f = 25 cm) was placed at the exit of the Kr cell and the laser beams
were aligned off-centered at the lens to separate the residual light beams (UV
and visible) from the VUV light that was focused onto the molecular beam. The
optimized Kr pressure in the cell was 5 ∼ 15 Torr.
VUV in the region of 106.6 ∼ 117 nm used to measure the spectra of the
cations in the excited states was generated by four-wave sum frequency mixing
in Hg. A schematic diagram of the experimental apparatus is in Fig. 1.10. The
UV light (ωUV = 312.8 nm, ∼ 2 mJ/pulse), which excites the Hg 61S0 – 71S0
transition via the two-photon resonance, was generated by frequency-doubling of
an output of a dye laser (Continuum ND6000) pumped by the second harmonic
of an Nd:YAG laser (Continuum PL8000) with ∼ 7 nsec pulse duration and 10
Hz repetition rates. ωS (2 ∼ 6 mJ/pulse) at 335 ∼ 470 nm was generated by the
second dye laser (Lambda Physik SCANMATE 2E) pumped by the second or
third harmonic of another Nd:YAG laser (Continuum PL8010). The two laser
beams were combined with a dichroic mirror and tightly focused using an
achromatic lens (f = 20 cm) onto the Hg vapor. The four-wave mixing Hg cell
was designed similar to that of Hilbig et. al.74 A LiF lens (f = 20 cm) was placed
23
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
molecular beamVUVE3
E2E1G
TOF
Kr cell
MgF2 lens(R=103mm)
Dichroic mirror
PI chamber
50cm lens
(a) top view
(b) side viewT
OF
MCP
Au plate
ωUV
ωS
Temp.-controlledPulsed valve
Fig. 1.9 Schematic drawing of the four-wave mixing Kr cell and VUV
MATI instrument. (a) top and (b) side views.
24
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
molecular beam
PI chamber
(b) side viewT
OF
MCP
LiF lens(R=103mm)
(a) top view
Achromaticlens(f=200mm)
Fused silicawindow
Hg cell
Dichroicmirror
ωUV
ωS
Temp.-controlledPulsed valve
VUV E3
E2E1G
TOF
Hg
Heatingblock
Ar
Water inWater in
Fig. 1.10 Schematic drawing of the four-wave mixing Hg cell and VUV
MATI instrument. (a) top and (b) side views.
25
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
at the exit of the Hg cell and the laser beams were aligned off-center to separate
the VUV light from the residual UV and visible lights at the interaction region
with the molecular beam. The VUV output in the 106.6 ∼ 117 nm region was
optimized at the Hg vapor pressure close to 0.9 Torr with Ar buffer (1 ∼ 2 Torr).
The spectral resolution was ∼ 1 cm-1 and 1010 ∼ 1011 photons /pulse were
generated. A small portion of a dye laser output was used to calibrate its
frequency based on the optogalvanic effect in a Fe/Ne hollow cathode lamp. Its
precision was ± 0.5 cm-1 in the visible region. A gold wire was placed in the
VUV beam path as a beam monitor. Its output was used to normalize the
intensity of each vibrational peak in the MATI spectra.
1.3.4 Application to Excited Electronic States of Ions
When vibrational peaks in the PES band of an excited electronic state of a
molecular cation are resolved,56-58 one can expect to obtain a well resolved
vibrational spectrum for this state. The states of benzene and chlorobenzene
cations studied previously by various methods are such states. The above are the
hole states generated by removal of an electron from an orbital lying below the
highest occupied orbital of the neutral. Various difficulties appear when one
attempts ‘1+1’ ZEKE or MATI for such a case. A useful technique here is to
prepare molecular cations in the ground electronic state and record the
multiphoton dissociation spectrum occurring via resonance excitation to the
excited electronic state of interest, or REMPD.
B~
B~
75,76 An alternative is to prepare
neutrals in highly excited Rydberg states and observe their ionization induced by
another laser, or photoinduced Rydberg ionization (PIRI) spectroscopy.77,78
REMPD and PIRI can suffer some complications when the transition involved is
electric dipole forbidden, such as the ← transitions for the benzene and
chlorobenzene cations. Here again, the simplest approach is to obtain ZEKE or
X~
26
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
MATI spectrum with the one-photon scheme using appropriate VUV radiation.
The one-photon MATI scheme has been found to be especially useful to
obtain the cation ground state spectrum because knowledge on the neutral
intermediate states is not required. In the study of excited electronic states, the
fact that ionization occurs only to the hole states in MATI is an additional
advantage. One-photon ZEKE scheme has been utilized already to obtain
vibrational spectra of simple cations in the ground and excited states.16,71-73 The
present study has demonstrated that one-photon MATI scheme can be routinely
used to obtain vibrational spectra of polyatomic cations also once coherent VUV
radiation becomes available over a wide spectral range.
27
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
References
1. B. D. Bartolo, Spectroscopy of the Excited State, (Plenum Press, New York,
1976).
2. M. Klessinger and J. Michl, Excited States and Photochemistry of Organic
Molecules, (VCH Publishers, New York, 1995).
3. J. Laane, H. Takahashi, and A. Bandrauk, Structure and Dynamics of
Electronic Excited States, (Springer, Berlin, 1999).
4. Raff, L. M., Thompson, D. L., Theory of Chemical Reaction Dynamics Vol.
III, edited by Baer, M. (CRC, Florida, 1985).
5. F. F. Crim, M. C. Hsiao, J. L. Scott, A. Sinha, and R. L. van der Wal, Phil.
Trans. Roy. Soc. A 332, 259 (1990)
6. P. Brumer and M. Shapiro, Ann. Rev. Phys. Chem. 43, 257 (1992)
7. K. K. Lehmann, G. Scoles, and B. H. Pate, Ann. Rev. Phys. Chem. 45, 241
(1994)
8. Y. S. Choi and C. B. Moore, J. Chem. Phys. 103, 9981 (1995)
9. D. J. Nesbitt and R. W. Field, J. Phys. Chem. 100, 12735 (1996)
10. M. Bergt, T. Brixner, and B. Kiefer, B., et al, J. Phys. Chem. A 103, 10381
(1999)
11. R. P. A. Bettens and M. A. Collins, J. Chem. Phys. 111, 816 (1999)
12. T. Ishida, and G. C. Schatz, Chem. Phys. Lett. 314, 369 (1999)
13. T. Hollebeek, T. S. Ho, and H. Rabitz, Annu. Rev. Phys. Chem. 50, 537
(1999)
14. M. S. Kim, C. H. Kwon, and J. C. Choe, J. Chem. Phys. 113, 9532 (2000).
15. C. H. Kwon, M. S. Kim, and J. C. Choe, J. Am. Soc. Mass Spectrom. 12,
28
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
1120 (2001).
16. C. H. Kwon, H. L. Kim, and M. S. Kim, J. Chem. Phys. 116, 10361 (2002).
17. Y. Y. Youn, C. H. Kwon, J. C. Choe, and M. S. Kim, J. Chem. Phys. 117,
2538 (2002).
18. C. H. Kwon, H. L. Kim, and M. S. Kim, J. Chem. Phys. 119, 4305 (2003).
19. E. Constantin and A. Schnell, Mass Spectrometry, (Ellis Horwood, New
York, 1991).
20. R. J. Cotter, Time-of-Flight Mass Spectrometry: Instrumentation and
Applications in Biological Research, (ACS, Washington, DC, 1997).
21. E. de Hoffmann and V. Stroobant, Mass Spectrometry: Principles and
Applications, (John Wiley & Sons, Chichester, 2001).
22. A. G. Harrison, Chemical Ionization Mass Spectrometry (CRC, Boca Raton,
1992).
23. I. Powis, T. Baer, and C. Ng, high Resolution Laser Photoionization and
Photoelectron Studies: in Wiley Series in Ion Chemistry and Physics, (John
Wiley & Sons, Chichester, 1995).
24. R. A. Marcus, J. Chem. Phys. 20, 364 (1952).
25. R. Bombach, J. Dannacher, J. P. Stadelmann, and J. Vogt, Chem. Phys. Lett.
77, 399 (1981).
26. B. E. Miller and T. Baer, Chem. Phys. 85, 39 (1984).
27. I. G. Simm, C. J. Danby, and J. H. D. Eland, Int. J. Mass Spectrom. Ion
Phys. 14, 285 (1974).
28. J. H. D. Eland, R. Frey, A. Kuestler, H. Schulte, and B. Brehm, Int. J. Mass
Spectrom. Ion Phys. 22, 155 (1976).
29. J. Dannacher, A. Schmelzer, J. P. Stadelmann, and J. Vogt, Int. J. Mass
29
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
Spectrom. Ion Phys. 31, 175 (1979).
30. G. K. Jarvis, K. J. Boyle, C. A. Mayhew, and R. P. Tuckett, J. Phys. Chem.
A 102, 3219 (1998).
31. T. Nishimura, P. R. Das, and G. G. Meisels, J. Chem. Phys. 84, 6190 (1986).
32. T. Nishimura, Q. Zha, and G. G. Meisels, J. Chem. Phys. 87, 4589 (1987).
33. K. M. Weitzel, F. Güthe, J. Mähnert, R. Locht, and H. Baumgärtel, Chem.
Phys. 201, 287 (1995).
34. J. Dannacher, Org. Mass Spectrom. 19, 253 (1984).
35. S. A. McLuckey and R. G. Cooks, Int. J. Mass Spectrom. Ion Proc. 56, 223
(1984).
36. J. Biggerstaff, K. Qian, S. Howard, A. Shukla, and J. Futrell, Chem. Phys.
Lett. 151, 507 (1988).
37. R. Chawla, M. Krishnamurthy, A. K. Shukla, and J. Futrell, Chem. Phys.
Lett. 301, 531 (1999).
38. X. Zhou, J. Wang, A. Shukla, and J. Futrell, Int. J. Mass Spectrom. 194, 171
(2000).
39. R. E. Krailler and D. H. Russell, Int. J. Mass Spectrom. Ion Proc. 66, 339
(1985).
40. T. L. Bunn and T. Baer, J. Chem. Phys. 85, 6361 (1986).
41. D. Y. Kim, J. C. Choe, and M. S. Kim, J. Chem. Phys. 113, 1714 (2000).
42. D. S Won, M. S. Kim, J. C. Choe, and T. K. Ha, J. Chem. Phys. 115, 5454
(2001).
43. K. Suto, Y. Sato, Y. Matsumi, and M. Kawasaki, J. Phys. Chem. A 101,
1227 (1997).
44. M. Hamdan, F. M. Harris, and J. H. Beynon, Int. J. Mass Spectrom. Ion
30
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
Proc. 74, 303 (1986).
45. J. P. Maier, P. Ausloos, ed., Kinetics of Ion-Molecule Reactions (Plenum
Press, New York, 1979).
46. M. Allan and J. P. Maier, Chem. Phys. Lett. 34, 442 (1975).
47. M. Allan, J. P. Maier, and O. Marthaler, Chem. Phys. 26, 131 (1977).
48. J. P. Maier and O. Marthaler, Chem. Phys. 32, 419 (1978).
49. Glassman, Combustion (Academic Press, San Diego, 1996).
50. G. P. Brasseur, J. J. Orlando, and G. S. Tyndall, Atmospheric Chemistry and
Global Change (Oxford University Press, New York, 1999).
51. C. R. Cowley, An Introduction to Cosmochemistry (Cambridge University
Press, Cambridge, 1995).
52. A. Carrington, in Molecular Ions: Geometric and Electronic Structures,
edited by J. Berkowitz and K. Groeneveld (Plenum Press, New York, 1983).
53. A. Carrington and T. P. Softley, in Molecular Ions: Spectroscopy, Structure,
and Chemistry, edited by T. A. Miller and V. E. Bondybey (North-Holland
Publishing Com., New York, 1983).
54. J. P. Maier, in Kinetics of Ion-Molecule Reactions, edited by P. Ausloos
(Plenum Press, New York, 1979).
55. K. Kimura, S. Katsumata, Y. Achiba, T. Yamazaki, and S. Iwata, Handbook
of HeI Photoelectron Spectra of Fundamental Organic Molecules (Japan
Scientific Societies Press, Tokyo, 1981).
56. A. W. Potts, D. Edvardsson, L. Karlsson, D. M. P. Holland, M. A.
MacDonald, M. A. Hayes, R. Maripuu, K. Siegbahn, and W. von Niessen,
Chem. Phys. 254, 385 (2000).
57. D. M. P. Holland, D. Edvardsson, L. Karlsson, R. Maripuu, K. Siegbahn, A.
W. Potts, and W. von Niessen, Chem. Phys. 252, 257 (2000).
31
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
58. D. M. P. Holland, D. Edvardsson, L. Karlsson, R. Maripuu, K. Siegbahn, A.
W. Potts, and W. von Niessen, Chem. Phys. 253, 133 (2000).
59. K. Walter, K. Scherm, and U. Boesl, Chem. Phys. Lett. 161, 473 (1989).
60. K. Walter, K. Scherm, and U. Boesl, J. Phys. Chem. 95, 1188 (1991).
61. K. Müller-Dethlefs, M. Sander, and E.W. Schlag, Chem. Phys. Lett. 112,
291 (1984).
62. K. Müller-Dethlefs and E. W. Schlag, Annu. Rev. Phys. Chem. 42, 109
(1991).
63. I. Fischer, R. Lindner, and K. Müller-Dethlefs, J. Chem. Soc. Faraday Trans.
90, 2425 (1994).
64. L. Zhu and P. Johnson, J. Chem. Phys. 94, 5769 (1991).
65. H. Krause and H. J. Neusser, J. Chem. Phys. 97, 5923 (1992).
66. H. Krause and H. J. Neusser, J. Chem. Phys. 99, 6278 (1993).
67. J. W. Hepburn, Chem. Soc. Rev. 25, 281 (1996).
68. E. W. Schlag, ZEKE Spectroscopy (Cambridge University Press, Cambridge,
1998).
69. A. Held, L. Y. Baranov, H. L. Selzle, and E. W. Schlag, Chem. Phys. Lett.
291, 318 (1998).
70. A. Held, U. Aigner, L. Y. Baranov, H. L. Selzle, and E. W. Schlag, Chem.
Phys. Lett. 299, 110 (1999).
71. C. H. Kwon, H. L. Kim, and M. S. Kim, J. Chem. Phys. 118, 6327 (2003).
72. C. H. Kwon, H. L. Kim, and M. S. Kim, J. Chem. Phys. 119, 215 (2003).
73. C. H. Kwon, H. L. Kim, and M. S. Kim, J. Chem. Phys. 119, 4311 (2003).
74. R. Hilbig and R. Wallenstein, IEEE J. Quantum Electron. QE-19, 1759
(1983).
32
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
75. X. Ripoche, I. Dimicoli, J. LeCalve, F. Piuzzi, and R. Botter, Chem. Phys.
124, 305 (1988).
76. K. Walter, U. Boesl, and E. W. Schlag, Chem. Phys. Lett. 162, 261 (1989).
77. J. G. Goode, J. D. Hofstein, and P. M. Johnson, J. Chem. Phys. 107, 1703
(1997).
78. R. Anand, J. D. Hofstein, J. E. LeClaire, and P. M. Johnson, J. Phys. Chem.
A 103, 8927 (1999).
33
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
Chapter 2
Discovery of Isolated Electronic State by
Mass Spectrometry
One of the main assumptions in the original formulation of the theory of
mass spectra1 was that the rate of dissociation of an ion be slow relative to the
rate of redistribution of energy of the internal degrees of freedom, both
electronic and vibrational. Namely, a molecular ion generated in an excited
electronic state would undergo a rapid radiationless transition to the ground
electronic state. Then, statistical distribution of the internal energy, which is now
in the form of the vibrational energy in the ground electronic state, would
determine the reaction rate. With this assumption, the theory becomes equivalent
to the Rice-Ramsperger-Kassel-Marcus (RRKM) theory2 for unimolecular
reaction in the ground electronic state. The theory has been successful to explain
most of the mass spectral features. Regardless, validity of the assumption of the
rapid conversion to the ground electronic state has been the focus of
investigations over the years.3-5 Cases to the contrary, namely the presence of
isolated electronic states, have been reported. The most obvious of these is
repulsive dissociation from an excited electronic state.6-8 Also, elaborate
experiments using techniques such as the photoelectron-photoion coincidence
spectrometry provided evidence for the presence of the isolated electronic states
for some simple systems.9-10
34
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
2.1 Initial Discovery: Benzene Cation
The benzene molecular ion is one of the most extensively studied ionic
systems. Various experimental and theoretical methods were used to investigate
the structure and the dissociation dynamics of this ion. It is well established now
that the C6H6+• ion generated by electron ionization of the neutral benzene
precursor retains the benzene structure.11 On the other hand, the nature of its
dissociation dynamics has been a controversial subject since the report by
Andlauer and Ottinger12 that the dissociations to C6H5+ and C4H4
+• were not in
competition according to their charge exchange kinetics data. Rosenstock and
coworkers13 analyzed the photoionization data and suggested that C6H5+and
C6H4+• be formed from the benzene ion ground state while C4H4
+• and C3H3+
from the first excited state. Further investigations14-18 carried out thereafter,
however, showed the statistical character of the benzene ion dissociation. Most
conspicuously, Neusser and coworkers15 found that the production rate constants
for the above four fragment ions measured by photodissociation of the benzene
ion generated by resonance-enhanced two-photon ionization were identical. It
was concluded that the benzene ion excited to the C or electronic
state undergoes a rapid internal conversion to the ground state where the
dissociations occur statistically. Now, the statistical dissociation of the benzene
ion in the ground state is generally accepted. A successful Rice-Ramsperger-
Kassel-Marcus (RRKM) theory modeling of the dissociation rate has been
reported,
~2u
2 A D~ 1u2 E
18 which is in excellent agreement with the recent experimental
data.15,17,18
Various experimental and theoretical efforts have been made also to study
the excited electronic states of benzene ion, a part of the interest arising from the
possible presence of an isolated electronic state mentioned above.19-35 In addition
35
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
to the ground electronic state X , several excited electronic states (
, and E~ ) have been identified by the photoelectron
spectroscopy.
~
2uB
g2E
1g2 E B~ g2
2 E ,
C~ 2u2 A , D~ 1u
2 E 2
2
30-32 Compared to the neutral benzene36 and the halogen-substituted
benzene ions,37,38 the first excited electronic state ( ) of benzene ion has
been found to be very difficult to investigate by conventional absorption and
emission spectroscopies.
B~ g22 E
19-22 In particular, lack of fluorescence from this state
has been the subject of great interest,9,12 which forbids the use of the highly
sensitive laser-induced fluorescence technique. This has been attributed to the
rapid internal conversion to the ground electronic state.22 Recently, high
resolution spectra of the B~ state have been obtained with the use of the
modern spectroscopic techniques such as the resonance-enhanced multiphoton
dissociation spectroscopy23 and the photo-induced Rydberg ionization
spectroscopy24 and the properties of this state have begun to be unveiled.
In the course of our photodissociation investigation of the benzene molecular
ion generated by electron impact, we observed an unexpected strong absorption
in the near ultraviolet. A combined kinetic and charge exchange ionization study
showed the presence of a long-lived excited electronic state of benzene ion. The
results from our investigation are discussed in this section.
2.1.1 Experimental Setup
A schematic diagram of the double focusing mass spectrometer with reversed
geometry (VG ZAB-E) modified for photodissociation (PD) study is shown in
Fig. 2.1. Samples were introduced into the ion source via a septum inlet and
ionized using 20 eV electron energy. Ions generated in the ion source were
accelerated to 8 keV and analyzed by conventional and photodissociation mass
36
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
Magnetic sector
Ion source
Electrodeassembly
Electric sector
PMT
Chopper
Lens
Argon ion laser
Phase-sensitive detection
Laser beam Prism
Laser beam
Collision cell
R1 R2 R3 R6R4 R7R5
Ion beam
Ion beam
Fig. 2.1 Schematic diagram of the double focusing mass spectrometer with reversed geometry (VG ZAB-E)
modified for photodissociation study. The inset shows the details of the electrode assembly.
37
Copyright(c)2002 by Seoul N
ational University L
ibrary. All rights reserved.(http://library.snu.ac.kr) 2004/01/28 17:41:01
spectrometries. Two different ion sources were used. To generate ions only by
electron impact, a source with a large conductance was used, which will be
called the electron ionization (EI) source. A source with smaller conductance, a
chemical ionization (CI)39 source, was used to generate ions via ion-molecule
reactions in the presence of the ionizing electron beam. The ion source
temperature was maintained at 150 °C. The CI experiment was done not only for
benzene itself but also for other samples, using benzene as the reagent gas.
Pressure of a sample other than benzene was maintained at ∼1 % of the reagent
gas.
Pressure was monitored with an ionization gauge located outside the ion
source. The pressure inside the CI source, P, was estimated from the ionization
gauge reading, Pig, using the relationship.40
ig)(= PgZS
P (2.1)
Here, Z is the conductance of the source evaluated in the usual way41, S is the
pumping speed estimated according to ref. 40, and g is the gas correction factor
of the ionization gauge available in the literature.42 Also measured was the
average source residence time of an ion exiting the CI source, which was used in
the work to estimate the number of collisions suffered by an ion. For this
purpose, the electron beam was pulsed and the time delay between the electron
and ion pulses was measured. Circuit diagram devised to pulse the electron beam
is shown in Fig. 2.2. The source residence time under the EI condition was
estimated by ion-optical trajectory calculations.43 Adding this to the measured
time delay difference between the CI and EI experiments, the source residence
time was estimated at various ion source pressures.
For the photodissociation study, benzene ion beam was selected by the
38
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
Fig. 2.2 Circuit diagram devised to pulse the electron beam. MA3104-
003D indicates the main board circuits controlling the ion source of ZAB-
E mass spectrometer.
39
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
magnetic sector and was crossed perpendicularly with a laser beam inside an
electrode assembly located near the intermediate focal point of the instrument as
shown in Fig. 2.1. The UV multiple line with the mean wavelength of 357 nm of
an argon ion laser (Spectra Physics Beamlok 2065-7S) and the 488.0 and 514.5
nm lines of another argon ion laser (Spectra Physics 164-09) were used. The
translational kinetic energy of product ions was analyzed by the electric sector.
Recording the kinetic energy of product ions generated by the dissociation of
mass-analyzed precursor ions is called the mass-analyzed ion kinetic energy
spectrometry (MIKES).44 A MIKE spectrum for PD, or a PD-MIKE spectrum, is
often contaminated by contributions from the same reactions occurring
unimolecularly or by collision of the precursor ions with the residual gas. Hence,
phase-sensitive detection was adopted to record a MIKE spectrum originating
only from PD. To improve the quality of PD-MIKE spectra, signal averaging
was carried out for repetitive scans.
2.1.2 Energetics of Benzene Ion
Adiabatic ionization energy to X was determined recently by
photoelectron spectroscopy
~
22 E
2u
1g2 E
g
31 as 9.243 eV which is essentially the same as 9.244
eV measured by zero kinetic energy spectroscopy.35 Information on the excited
states is limited to those created by loss of an electron from the occupied orbitals
as observed by photoelectron spectroscopy (PES). The second PES band, namely
the band corresponding to the ionic B state, is well resolved. A recent PES
study
~
2 B
31 reported the adiabatic ionization energy of 11.488 eV for this state.
Those to C , and E~ are known less accurately due to the
broad and overlapping nature of the photoelectron bands as ~12, ~13.8 and
~14.7 eV, respectively. No experimental energetics data are available for the
~2u
2 A , D~ 1u2 E
40
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
X 2E1g (ground state)
B 2E2g
C 2A2u
D 2E1u
E(eV)
0
5
3
2
4
Dissociation( Products )
~
~
~
~
E 2B2u
~
Electronic states( C6H6
+• )
ktot ~ 107s-1
ktot ~ 104s-1
C6H6+•→ C6H5
+ + H•
6
1
Fig. 2.3 Energy diagram of the benzene molecular ion. The lowest reaction
threshold (E0) is 3.66 eV for C6H6+•→C6H5
++H•. ktot denotes the total rate
constant for dissociation in the ground electronic state predicted from
previous studies.18
41
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
excited electronic states associated with the elevation of an electron to
unoccupied orbitals. However, these states are thought to lie well above the
state according to the spectroscopic data for neutral benzene. The lowest
reaction threshold for benzene ion, namely the critical energy
B~ g22 E
18 for the
production of C6H5+ and H• from the ground state benzene ion, is 3.66 eV. These
energetics data are shown schematically in Fig. 2.3.
2.1.3 Photodissociation Kinetics
When the benzene ions generated by EI were irradiated with visible (514.5
and 488.0 nm) or near UV (357 nm) lasers with the corresponding photon
energies of 2.41,
2.54, and 3.47 eV, respectively, noticeable amounts of C4H4+• were produced
with C6H5+, C6H4
+•, C5H3
+, and C3H3+ appearing as minor products. The electron
energy used for ionization hardly affected the photodissociation features, even
though data obtained at 20 eV EI will be presented here for comparison with
other data which will be discussed later. To obtain a rough estimate of the PD
cross section at 357 nm, the PD yields were measured for benzene ion, 1,3,5-
hexatriene, cyclopropylbenzene, and n-butylbenzene ions. The latter three are
the ions whose PD cross sections at 357 nm are available in the literature.46-48
Care was taken such that all the measurements were done under the same
experimental condition, especially the same ion beam-laser beam overlap. Both
the precursor and PD fragment ion intensities were measured to obtain the
relative PD yield in each case. Comparing the PD yield of benzene ion with
those of the other three and using the PD cross sections of the latter ions in the
literature, the PD cross section of benzene ion at 357 nm was estimated. This
was ~9×10-19 cm2, corresponding to a fairly strong dipole-allowed transition. The
42
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
PD yields at 488.0 and 514.5 nm were roughly 30 and 20 %, respectively, of that
at 357 nm.
According to previous photodissociation studies of benzene ion trapped in an
ion cyclotron resonance (ICR) spectrometer reported by Freiser and
Beauchamp33 and also by others,34,35 a benzene ion can absorb two visible
photons sequentially via the dipole-allowed C ← transition
( ← is dipole-forbidden), namely via two successive C~ ←
transitions with a rapid internal conversion in-between, and dissociate. Such a
sequential two-photon excitation mechanism is not applicable here because the
ion transit time (~1ns) across the light field is far shorter than in the ICR
experiment (~1s). Even for other dipole-allowed one-photon photodissociation
processes, the product yields detected by the present apparatus are less than
0.001 % of the parent ion intensity. This forms another argument against the
sequential two-photon absorption. Then, the fact that dissociation of benzene ion
could be induced by absorption of one photon with energy less than the reaction
threshold was puzzling. Hence, an attempt was made in this work to estimate the
internal energy of benzene ion undergoing dissociation using the
photodissociation kinetics technique developed previously.
~2u
2 A X~ 1g2 E
B~ g22 E X~ 1g
2 E X~
49-51
The PD-MIKE spectrum at 357 nm was recorded for benzene ion with 2.1
kV applied to the electrode R2 and the remaining electrodes of the electrode
assembly (Fig. 2.1) grounded. With the electric field present in the
photoexcitation region, the kinetic energy of a product ion will change
depending on the position of its formation. The MIKE spectral region
corresponding to the production of C4H4+• is shown in Fig. 2.4. The positions
marked A and B in this figure are the kinetic energies of products generated at
the position of photoexcitation and after exiting the ground electrode (R3 in this
43
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
5300 5500 5700 5900
B
A
Inte
nsity
Translational energy, eV
Fig.2.4 PD-MIKE profile for the production of C4H4+• from the benzene
ion at 357nm obtained with 2.1kV applied on the electrode assembly.
Experimental result is shown as filled circles. Reproduction of the profile
using the rate constant distribution centered at 6.3×107 s-1 obtained by
experimental data is shown as the solid curve. The positions marked A and
B are the kinetic energies of products generated at the position of
photoexcitation and after exiting the ground electrode, respectively.
44
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
case), respectively. The asymmetric tail in the lower energy side of A is due to
dissociation occurring before R3 on a nanosecond time scale. The method to
evaluate the rate constant (k) by analyzing this time-resolved PD-MIKE profile
is well established now.49-51 The most probable rate constant determined from
several duplicate experiments are (5.5±1.1)×107 s-1. The MIKE profiles for
C6H5+ and C6H4
+• were not well resolved due to the interference from the
precursor signal. The MIKE profiles for C5H3+ and C3H3
+ were very similar to
that in Fig. 2.4, indicating the competitive dissociations. If all the five
dissociation channels are competitive as is currently accepted, the measured rate
constant corresponds to the sum of the individual rate constants. Then, the
average internal energy content of the photodissociating benzene ion can be
determined by comparing with the previous experimental rate-energy data or
with their RRKM fit. Since the present rate constant is a little out of range from
the previous rate-energy data,15,17 the RRKM calculation of Grebner and
Neusser18 was extended to higher internal energy. Production of C5H3+, which
was less than 10 % of the total product intensity, was ignored since the
parameters needed for calculation were not reported by the above investigators.
The RRKM expression of the rate constant for each channel, ki (E), at the
reactant internal energy E is given by
)()-()( 0
EhEEWEk ii
ii ρσ
‡
= (2.2)
Here, h is the Planck constant, ρ is the density of states of the reactant, Wi‡ is the
state sum at the transition state for the ith channel with the critical energy E0i,
and σi is the reaction path degeneracy. Then, the total dissociation rate constant
is obtained by adding rate constants for all the competing channels.
45
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
4 5 6 7 8
2
4
6
8
10357 nm PD
488.0 nm PD
log
k, k
in s-1
Internal energy, eV
Fig. 2.5 The total RRKM dissociation rate constant of benzene ion as a
function of the internal energy calculated with molecular parameters in ref.
18. The internal energies corresponding to the dissociation rate constants
of ( 5 . 5±1 .1 )×107 and (5±3 )×106 s-1 for PDs at 357 and 488.0 nm,
respectively, are marked.
46
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
∑4
1=total )(=)(
ii EkEk (2.3)
The total rate constant calculated over the internal energy range of interest using
the molecular parameters in ref. 29 is shown in Fig. 2.5. The calculated rate
constants at two internal energy values are marked in Fig. 2.3 also to emphasize
the fact that the photodissociation can be observed with the present nanosecond
apparatus when the internal energy of the photo-excited benzene ion is higher
than the reaction threshold by a few eV. The average internal energy of the
photodissociating benzene ion corresponding to the observed rate constant of
(5.5±1.1)×107 s-1 was read from the calculated rate-energy data, which was
6.1±0.1 eV. Subtracting the 357 nm photon energy (3.47 eV) from this value
results in the average pre-excitation internal energy of 2.6±0.1 eV for benzene
ion undergoing photodissociation. Dissociations in the visible were slower and
the rate constants were beyond the limit that can be determined reliably by the
present technique. My best estimate for PD at 488.0 nm was k= (5±3)×106 s-1
which corresponds to the internal energy of 5.5±0.1 eV. Subtracting the photon
energy of 2.54 eV, the pre-excitation internal energy of benzene ion undergoing
PD at this wavelength becomes 3.0±0.1 eV. A routine explanation for the above
pre-excitation internal energies is that the benzene ions generated in various
excited electronic states by EI undergo rapid internal conversion and the
electronic energies of these excited states are converted to the vibrational energy
in the ground electronic state. Validity of such an explanation can be checked by
observing the decrease in the PD signal as the vibrational energy of the ground
electronic state benzene ion is relaxed in the high pressure ion source.
2.1.4 Quenching of Photodissociation
As an attempt to induce the quenching of the PD signal by collisional
47
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
relaxation of the vibrationally hot benzene ions, we introduced a benzene/argon
mixture (benzene:argon=1:50) into the source and recorded the PD signal as a
function of pressure. Even though the Ar+• ion generated by electron ionization
can ionize the neutral benzene by charge exchange, the benzene ion thus
generated would have the internal energy as large as 6.5 eV and dissociate
immediately, not affecting the photodissociation carried out at 22 µs after ion
formation. We could not achieve efficient relaxation of the vibrationally hot
benzene ion in this way, however. Namely, the PD signals increased with the
mixed gas pressure in the ion source almost to the point where the intensity of
the precursor benzene ion itself began to decrease due to various collisional
processes such as deflection and subsequent wall collision. Some of the data
obtained in this experiment will be discussed later.
In the case of molecular ions, it is well known that the resonant charge
exchange17,52 with their neutral counterparts is more efficient for vibrational
relaxation than the usual inelastic processes. In this process, the molecular ion is
converted to a neutral retaining the vibrational energy while the neutral collision
partner becomes an ion with much less vibrational energy. Namely,
C6H6+•* + C6H6 C→ 6H6* + C6H6
+• (2.4)
As an attempt to induce the quenching of the PD signal via charge exchange
collision of vibrationally hot benzene ions in the ion source, we introduced
benzene to the CI source to much higher pressure than normally used in the
electron ionization (EI) mass spectrometry and measured the PD signal as a
function of the benzene pressure. Fig. 2.6 shows the precursor (C6H6+•) intensity
and photoproduct (C4H4+•) intensities at 357 and 488.0 nm as functions of the
pressure measured by an ionization gauge located outside of the source. The
pressure dependence of the photoproduct signal at 514.5 nm was essentially the
48
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
same as that at 488.0 nm. At high pressure, the ionization gauge reading can be
converted to the source pressure following the method described in a previous
section. Pressure data at some important points are listed in Table 2.1. Tuning of
the instrument changed a little with the source pressure, the effect of which was
not corrected for in the data. The precursor intensity increases initially as the
pressure increases, as expected. This reaches the maximum at the ionization
gauge reading of 3.5×10-5 Torr which corresponds to the ion source pressure of
∼0.04 Torr and then begins to decrease with the pressure. Various physical and
chemical processes must be responsible for the overall pressure dependence in
the high pressure region. The photoproduct ion signals also rise and fall with the
pressure increase, with the maxima at pressures lower than for the precursor. The
most remarkable observation here is the fact that the maximum in 357 nm PD
occurs at much higher pressure than that in 488.0 nm one. Even in the high
pressure region where the 488.0 nm PD signal disappears completely, the 357
nm signal keeps increasing with the pressure.
Absorption of a photon at 357 and 488.0 nm increases the benzene ion
internal energy by 3.47 and 2.54 eV, respectively. Considering that the critical
energies for the production of the major fragment ions are ∼4 eV,29 rather close
to the 357 nm photon energy, one may attempt to explain the above observation
with the argument that PD at 357 nm would not be easily quenched because the
pre-excitation internal energy needed for dissociation is not large. Such an
argument is not valid in the present experiment, however, because the photo-
excited benzene ion must possess internal energy a few eV above the critical
energy for its dissociation to be detected on a nanosecond time scale (Figs. 2.3
and 2.5). Specifically, the pre-excitation internal energies needed for PDs at 357
and 488.0 nm are comparable, 2.6±0.1 and 3.0±0.1 eV, according to the PD
kinetics data presented above.
49
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
Table 2.1. Ion source pressure (P), collision frequency (Zc), source residence
time (tR), and number of collisions (Ncoll) suffered by ions exiting the ion source
at some benzene pressures.
Pig/Torr P/Torr Zc/µs-1 tR/µs Ncolla
4×10-6 0.0051 0.13 4.2 0.6 (1.4)
1×10-5 0.013 0.33 5.8 1.9 (4.5)
2×10-5 0.025 0.63 7.6 4.8 (12)
3×10-5 0.038 0.96 9.0 8.6 (23)
5×10-5 0.063 1.59 11.2 18 (49)
7×10-5 0.088 2.23 13.0 29 (82)
a Numbers in the parentheses are for the Ar reagent gas.
To see if the vibrational relaxation by charge exchange collision, reaction
(2.4), was operative in the actual experiment, we estimated the average number
of collisions suffered by benzene ion exiting the ion source by multiplying the
collision frequency (Zc) by the average residence time in the ion source.
uZ ccc 2= σρ (2.5)
Here, ρc and <u> are the number density and the average speed, respectively. σc
is the cross section for ion-dipole collision which was estimated roughly using
the Langevin formula.53
21
r0
2
c E2'
=
πεαπσ e (2.6)
50
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
Here, α′ is the polarizability volume (1.032×10-29 and 1.66×10-30 m3 for C6H6
and Ar, respectively)54 and Er is the relative translational energy which can be
approximated as 23 kT. Collision frequencies, average source residence time, and
average number of collisions suffered by ions exiting the source estimated at
several pressures are listed in Table 2.1. We expect that the number of collisions
thus estimated is accurate within a factor of 2∼3. At the ionization gauge reading
of ∼4×10-6 Torr corresponding to the maximum in the PD intensity at 488.0 nm,
the average number of collisions was 0.60. On the other hand, the same PD
intensity hardly decreased even after 10∼20 collisions in the quenching
experiment with Ar mentioned previously (data not shown). This indicates that
the efficient vibrational relaxation by charge exchange collision be operative in
the ion source at high benzene pressure. In contrast, the average number of
collisions at the ionization gauge reading of 2.5×10-5 Torr corresponding to the
maximum in the PD intensity at 357 nm was 6.7. Namely, the quenching of the
PD signal at 357 nm by charge exchange collision in the ion source is much less
efficient than in the 488.0 nm PD case. Considering that every charge exchange
collision is effective for vibrational relaxation (488.0 nm PD case), very
inefficient quenching of PD at 357 nm can not be explained simply based on the
smaller requirement of the pre-excitation internal energy in this case.
The fact that the collisional quenching behaviors of PDs at 357 and 488.0 nm
are dramatically different can be explained only by assuming that the C6H6+•
precursors for 357 and 488.0 nm PDs be different, either in structure or state. As
was mentioned in a previous section, benzene ion is the most stable of the
C6H6+• isomers and retains its structure. Also, the fact that the recombination
energy of 9.2 eV was found valid for C6H6+• generated from benzene at even
higher source pressure (0.4 Torr)55 than used in this work supports that the
51
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
C6H6+• ion investigated here has the benzene structure. This is also supported by
the charge exchange ionization data to be presented later. Then, the remaining
possibility is that the electronic state of benzene ion responsible for PD at 357
nm is different from that at 488.0 nm. If the C ← transition for
benzene ion in the ground electronic state with a large vibrational energy is
responsible for PD at 488.0 nm, the PD at 357 nm must involve a transition from
an electronic state other than the ground state. The fact that no peak was
observed in this spectral region in the absorption spectrum of benzene ion
trapped in the argon matrix
~2u
2 A X~ 1g2 E
20,21 is in agreement with such an assignment.
The assumption that the initial state for PD at 357 nm be an excited
electronic state has some difficulties of its own. The first is that this state should
survive for ∼10µs in the source and also for a few tens of microseconds needed
for the flight from the source to the laser-ion interaction region. This can be
reconciled only by assuming that the electronic state involved is long-lived, with
the lifetime of 10 µs or longer. The second difficulty of the above assumption is
that the collisional relaxation of this excited electronic state is not as efficient as
the vibrational relaxation in the ground electronic state (PD at 488.0 nm). This
can be explained readily, however, considering that the most efficient process in
the collision of the electronically excited benzene ion (C6H6+•†) with the neutral
benzene would be the resonant charge exchange.
C6H6+•† + C6H6 C→ 6H6 + C6H6
+•† (2.7)
The PD yield at 357 nm is maintained at high source pressure because the
resonant charge exchange of C6H6+•† regenerates C6H6
+•† and hence does not
affect the C6H6+• population in the excited electronic state. This is in contrast
with the resonant charge exchange of C6H6+•* which reduces its population.
52
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
��
Relaxation of the electronically excited benzene ion by inelastic collisions
must be occurring also, however, considering that decay of the 357 nm PD
signal occurs at lower pressure than that of the total benzene ion current in Fig.
2.6. Then, benzene ion would be mostly in the ground electronic state at a very
high source pressure, for example at ∼ 7×10-5 Torr ionization gauge reading in
Fig. 2.6 which corresponds to ∼ 0.09 Torr source pressure. The relative 357 nm
PD cross section at this pressure was less than 0.1 % of that at much lower
pressure region suggesting that most of the benzene ions are in the ground
electronic state. This may be the reason why the benzene ion recombination
energy of 9.2 eV, which is equivalent to the ionization energy of benzene to the
ionic ground state, was observed by Rao and Fenselau55 in the charge exchange
experiment performed at the benzene pressure of ∼ 0.4 Torr. Two experimental
results, Rao and Fenselau’s and ours, can not be compared quantitatively,
however, because the methods used to measure the pressure might not be the
same.
2.1.5 Charge Exchange Ionization by Benzene Ion in the Excited
Electronic State
Benzene is a well-known reagent in the chemical ionization mass
spectrometry.39 As mentioned above, benzene ion is the reagent ion in this
ionization scheme and ionizes compounds with the ionization energy less than
9.243 eV by charge exchange at a very high source pressure of ∼ 0.4 Torr.
C6H6+• + A → C6H6 + A+• (2.8)
If the benzene ion population in the long-lived excited electronic state is
maintained at a moderately high source pressure such as 1.0×10-5 Torr ionization
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
10-6 10-5 10-4
0
0.09 Torr
0.04 Torr
0.013 Torr
Rel
ativ
e in
tens
ity
Ionization gauge reading, Torr
Fig. 2.6 Pressure dependences of the precursor (C6H6+•) intensity (–––)
and photoproduct (C4H4+•) intensities at 357 (·····) and 488.0 (–––) nm.
Pressure in the CI source was varied continuously to obtain these data. The
abscissa shows the pressure read by an ionization gauge located outside of
the source. The inside source pressures estimated using eqn. (1) at three
ionization gauge readings are marked. The scale for the precursor intensity
is different from that for photoproduct intensities.
54
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
gauge reading (∼0.013 Torr in source pressure), it should be possible to
determine its recombination energy through charge exchange experiments.
Namely, samples with ionization energy less than the recombination energy of
the electronically excited benzene ion would be ionized efficiently by charge
exchange while those with higher ionization energy would not be ionized.
Samples used in the experiment are listed in Table 2.2 together with their
ionization energies. Since some of the samples are ionized by electron impact
even under the chemical ionization (high source pressure) condition, the
molecular ion intensities were measured under the CI and EI conditions and their
ratios (CI/EI ratio) were evaluated. At the high ion source pressure of 0.09 Torr
(7×10-5 Torr ionization gauge reading), samples with ionization energy larger
than the recombination energy (9.243 eV) of the ground state benzene ion were
hardly ionized, as listed in Table 2.2. This is in agreement with the report by Rao
and Fenselau55 and also with the near complete decay of the electronically
excited benzene ion at this pressure indicated in Fig. 2.6. At lower pressure,
however, the molecular ion intensities of some samples with ionization energy
larger than 9.243 eV increased with the source pressure. The CI/EI ratios for the
molecular ion intensities measured at the low ion source pressure of 0.013 Torr
(1.0×10-5 Torr ionization gauge reading) are listed in Table 2.2. Data at 0.013
Torr are drawn in Fig. 2.7 also. It is remarkable to note that the samples with
ionization energy less than ~11.5 eV can be ionized efficiently by benzene ion at
this pressure while those with ionization energy a little larger than this are
difficult to ionize. This is a strong evidence supporting our proposition that there
exists a long-lived excited electronic state of benzene ion. Also to be mentioned
is that the above experimental recombination energy of ~11.5 eV is much larger
than those for other C6H6+• ions which have been investigated in the
isomerization study.56 Namely, other isomers of C6H6+• can not ionize samples
55
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
9 10 11 12 13
0
2
4
6 9.243 eV 11.5 eV
CI/E
I rat
io
Ionization energy, eV
Fig. 2.7 The ratios of molecular ion intensities generated by charge
exchange ionization (CI) with benzene ion and by electron ionization (EI)
are plotted as a function of the sample ionization energy. • and o are for CI
intensities measured at the ion source pressure of 0.013 and 0.09 Torr,
respectively.
56
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
Table 2.2. Ionization Energies and the ratios of molecular ion intensities
generated by charge exchange ionization (CI) with benzene ion and by electron
ionization (EI).
Compounds IEa (eV) CI/EI(0.013 Torr)b CI/EI(0.09 Torr)b
Chlorobenzene 9.06 3.6 3.5
Fluorobenzene 9.20 3.9 1.4
Benzonitrile 9.62 5.3 0.06
Chloropentafluorobenzene 9.72 4.7 0.01
Nitrobenzene 9.86 3.8 0.06
Hexafluorobenzene 9.91 2.5 0.02
Ethylene 10.51 3.0 0.02
Methylene chloride 11.32 4.4 0.04
Chloroform 11.37 4.7 0.03
Carbon tetrachloride 11.47 3.4 0.06
Ethane 11.52 0.09 ~0
Dichlorofluoromethane 11.75 0.05 0.04
1-chloro-1,1-difluoroethane 11.98 0.16 0.01
Chlorodifluoromethane 12.2 0.09 0.05
Methane 12.51 0.24 ~0
a All the ionization energy (IE) data are from ref. 57 except that for
dichlorofluoro-methane, which is from ref. 58.
b Ratio of the molecular ion intensities measured under the CI and EI conditions.
The pressure in the parenthesis is the ion source pressure in the CI experiment.
EI experiment was done at much lower pressure to avoid ion-molecule reaction.
57
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
with ionization energy close to 11.5 eV by charge exchange. This is another
evidence that the precursor in the PD at 357 nm has the benzene structure. The
difference between the experimentally estimated recombination energy of ~11.5
eV for the excited benzene ion and the first ionization energy of benzene is ~2.3
eV. This is also in remarkable agreement with the average pre-excitation internal
energy of 2.6 eV estimated by the photodissociation kinetics at 357 nm
considering that the benzene ion would possess a few tenth of an eV of thermal
vibrational energy in addition to the electronic energy.
As a further test, we also measured the PD yield at 357 nm with the addition
of a small amount (~1%) of scavenger which can decrease the population of the
excited electronic state of benzene ion. For example, addition of benzonitrile
which has the ionization energy of 9.62 eV decreased the PD yield to ~60% even
though the total precursor current was hardly affected. This is also in agreement
with our proposition.
2.1.6 Conclusions
Photodissociation kinetics and charge exchange ionization mass spectrometry
have shown that there exists a long-lived excited electronic state of benzene ion
at ~2.3 eV above the ground state. The lifetime of the state seems to be longer
than 10 µs, maybe much longer. Even though its presence seems to be certain,
the exact identity of the state is a matter of conjecture because not much is
known about the electronic states of benzene ion. According to the photoelectron
spectroscopy, and C are the electronic states in this energy region
which are accessible by one electron removal from an occupied molecular
orbital of the neutral. is not a likely candidate because the corresponding
photoelectron peak appears as a broad structureless band indicating a rapid
B~ g22 E ~
2uA
2u2 A
C~ 2
58
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
radiationless process in this state. The assumption of the rapid internal
conversion of this state needed to interpret the photodissociation in the visible33
as the sequential two-photon process is also against the assignment. This leaves
as the best candidate for the long-lived excited electronic state observed
in this work. In particular, ~2.3 eV energy observed for the long-lived state well
matches with the energy difference of 2.245 eV between B and
Such an assignment is not absolutely certain, however, because of the
controversy concerning the lifetime of the B~ state. Leach and coworkers
B~
X~
g22 E
~g2
2 E
E~
X~
g2
1g2 E .
g22 E
g
22
established that fluorescence from this state is not observable and proposed that
a rapid (>8×1010 s-1) internal conversion to the ground state be responsible.
Köppel27 found theoretically that a new type of conical intersection between the
and states triggered by the Jahn-Teller effect may be responsible for the
rapid internal conversion. However, failure to fluoresce may have an entirely
different origin, namely a very long lifetime. Further spectroscopic study on this
state would be useful, such as the acquisition of a high resolution spectrum and
the determination of the lifetime from the spectral width. If the B~ is the
initial state of the present photo-excitation at 357 nm, the transition involved can
be either ← or ← both of which are electric dipole-
allowed. Appearance of the prominent photodissociation signal at 357 nm is in
agreement with such assignments. Whatever the transition is, the benzene ion
further excited by the absorption of a 357 nm photon must undergo a rapid
internal conversion to the ground state for the RRKM theory to be applicable as
assumed in this work. The broad spectral profiles of the D and bands in the
photoelectron spectrum suggests the feasibility of such an assumption. Even
though not applicable in the present case, a situation may arise in which the
B~
2 E
D~ 1u2 E B~ g2
2 E E~ 2u2 B B~ 2E2
~
59
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
dissociation occurs in the B~ state. Then, its kinetics would not be
compatible with the currently accepted model of statistical dissociation in the
ground state.
g22 E
The fact that the observed excited electronic state is very long-lived suggests
the possibility of studying the physics and chemistry of this state by various
means. A useful further development may be the preferential generation and
storage of ions in this state. Then, study of the excited state chemistry on a
practical time scale may become a reality.
2.2 Method to Detect Isolated Electronic States
Using the same techniques introduced in previous section, we attempted to
find the evidence for the presence of the isolated electronic states in other ionic
systems. We have found, however, that the above techniques are not generally
applicable. For the photodissociation kinetic technique to be applicable, the
molecular ion of interest must absorb at the wavelength range covered by strong
continuous wave lasers and also must dissociate on the nanosecond ~ low
microsecond time scale. Furthermore, reliable rate-energy data, which are
needed to compare the experimental results with, are available only for limited
number of ionic systems. In the case of the charge exchange ionization in the ion
source, various complications can arise because both the reagent and sample are
mixed in the ion source and ionized together.39 In fact, the previous study on the
benzene system should be regarded as fortunate because benzene is a well-
known reagent for chemical ionization.55 A better method than the charge
exchange ionization in the ion source would be to separate spatially the
ionization and charge exchange regions. Such apparatuses were built in early
days of the charge exchange ionization mass spectrometry and used to measure
60
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
accurately the cross sections, appearance energies, etc.59,60 Instead of building a
sophisticated instrument, we simply modified a commercial double focusing
mass spectrometer, which turned out to be adequate for the present purpose. The
results from the investigation are reported in this paper.
2.2.1 Experimental Setup
A schematic diagram of the double focusing mass spectrometer with reversed
geometry (VG ZAB-E; Manchester, UK) is shown in Fig. 2.8. Benzene was
introduced into the ion source via a glass capillary connected to a reservoir
(septum inlet) and ionized under the electron ionization (EI) or chemical
ionization (CI) condition using 20 eV electron energy. The ion source
temperature was maintained at 140 °C. The CI experiments were done at the
benzene pressure of 0.02 or 0.1 Torr in the CI source. Ions generated in the
source were accelerated with high voltage (VS) of 4 kV. 1,3-Butadiene was
purchased from Matheson (Parsippany, NJ) and other chemicals from Aldrich
(Milwaukee, WI). All the chemicals were of the highest purity commercially
available and were used without further purification.
Collision Cells : To investigate the charge exchange ionization of reagent gases
by benzene ion, two collision cells were utilized, the first cell located between
the ion source and the magnetic sector and the second cell between the magnetic
and electric sectors. The second cell is as designed by the manufacturer and can
be floated at a high voltage. In the original design, the first collision cell was
located after the Y-focusing electrodes. For this work, we redesigned the
assembly such that the collision cell was located in front of the Y-focusing
electrodes (see the inset in Fig. 2.8). Also, the collision cell was modified to be
floatable at a high voltage and a repeller plate was installed. Namely, the first
collision cell in its present form functions as the ion source for the reagent gas
61
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
Collision cell assembly
Ion beam
Ion source
Magnetic sector Conversion dynode
EM
Electric sector
Collision Cell Y-lens
Repeller
IonSource
Collision cell assembly
Conversion dynode
PM
SecondCollision
cell
Fig. 2.8 Schematic diagram of the double focusing mass spectrometer with reversed geometry (VG ZAB-E). The
inset shows details of the first collision cell assembly modified for charge exchange study.
62
Copyright(c)2002 by Seoul N
ational University L
ibrary. All rights reserved.(http://library.snu.ac.kr) 2004/01/28 17:41:01
with the ionization achieved by collision with the primary ions entering the cell.
2.2.2 Principle of the Method
This subsection is to show evidence for the presence of primary ions in long-
lived excited electronic states by observing charge exchange ionization (eqn.
(2.8)) of various reagent gases with different ionization energies occurring in the
collision cell. When the primary ion has large translational kinetic energy, not
only the charge exchange ionization but also other processes such as the
collisional impact ionization can generate ions from the reagent gas.61
Decelerating the primary ion would be necessary to eliminate the latters. However,
when the translational energy of the primary ion is reduced to near thermal level,
it diverges in the collision cell or even before entering the cell, which results in
poor reagent ion signal. It will be seen in the next section that deceleration of the
primary ion to near thermal energy is not needed and fairly good results can be
obtained with deceleration to 50∼100 eV.
When a primary ion, m1+, with translational energy eVS is injected into a
collision cell containing a reagent gas, three types (Ⅰ, Ⅱ, and Ⅲ) of ions may
exit the cell, the primary ion itself (typeⅠ), its collision-induced dissociation
product (m2+, typeⅡ), and the ions generated from the reagent gas (typeⅢ). When
a high voltage, VC ( < VS), is applied on the collision cell, the translational
energies of these ions exiting the cell are as follows.62
TypeⅠ, KⅠ = eVS (2.9)
TypeⅡ, KⅡ = e[VC + (m2/m1)(VS - VC)] (2.10)
TypeⅢ, KⅢ = eVC (2.11)
63
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
When the second collision cell is used, these three types of ions can be readily
distinguished by measuring their translational energies with the electric sector.
However, it is not possible to determine whether the typeⅢ ions generated are the
molecular ion of the reagent gas, its fragments or reaction products, or a mixture
of these.
Identity of the typeⅢ ions can be determined by performing experiment using
the first collision cell and analyzing ions with the magnetic sector. The mass (m)
of the ion with the translational energy eV and the magnetic field B of the sector
with radius r needed to transmit it are related63 by
m/z = B2r2e/2V (2.12)
Then, using the translational energies in eqns. (2.9) ∼ (2.11), the peaks in the mass
spectrum recorded by scanning the magnetic field can be identified. As a further
check, one can set the magnetic field to transmit a particular peak in the mass
spectrum and measure its translational energy by scanning the electric sector
potential as in the usual mass-analyzed ion kinetic energy spectrometry
(MIKES)64. All the typeⅢ ions must appear at the trasnslational energy of eVC
(eqn. (2.10)) in the MIKE spectrum. One may also record the typeⅢ ions only by
setting the electric sector potential to transmit ions with the translational energy
eVC and scanning the magnetic field. This is not possible yet with the present
apparatus.
2.2.3 Utilization of Method
Since the ionization energies to the ground and long-lived excited electronic
states of benzene ion are 9.243 and ∼11.5 eV, respectively, we chose five
molecules with ionization energies slightly below or above these values as the
64
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
reagents. These are 1,3-C4H6 (butadiene)65, CS2, CH3Cl, CH3F, and CH4 with the
ionization energy66 of 9.08, 10.07, 11.22, 12.47, and 12.51 eV, respectively.
Energetics consideration dictates that only 1,3-C4H6 can be ionized by charge
exchange with the ground state benzene ion while 1,3-C4H6, CS2, and CH3Cl can
be ionized by the benzene ion in the long-lived excited state. Neither of the
benzene states have sufficient electronic energy to ionize CH3F and CH4 by
charge exchange.
We first attempted to study charge exchange ionization of CS2 introduced into
the second collision cell. Benzene ion was generated by 20 eV EI in the ion source,
accelerated to 4002 eV, mass-selected by the magnetic sector, decelerated by
collision cell potential of 3902 V, and introduced to the cell. Then, benzene ion
would have the pre-collision translational energy of 100 eV. The pressure of CS2
in the second collision cell was adjusted to attenuate the primary ion beam
(C6H6+•) intensity by 20 %. A partial MIKE spectrum thus obtained is shown in
Fig. 2.9. Most of the peaks in this spectrum arose from unimolecular (metastable
ion decomposition) and collision-induced dissociations of benzene ion and could
be identified using eqn. (2.9). A peak, marked Ⅲ in the figure, appeared at the
translational energy of 3902 eV which corresponded to the acceleration energy
due to the voltage applied on the collision cell. The position of this peak changed
with the cell potential exactly as dictated by eqn. (2.10), showing that the peak
originated from the reagent gas. As mentioned earlier, however, ions responsible
for this peak can not be identified by MIKE spectrometry. Peaks at the same
position were also observed with 1,3-C4H6 and CH3Cl reagent gases but not with
CH3F and CH4, indicating that charge exchange ionization occurred in the former
three cases.
Similar experiment was done by introducing CS2 in the first collision cell. Fig.
2.10 (a) shows the mass spectrum recorded under the single-focusing condition,
65
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
3900 3930 3960 3990
77+(MID)
II
II
II
III
Inte
nsity
Translational Energy, eV
Fig. 2.9 MIKE spectrum of the C6H6+• primary ion generated by EI. CS2 was
introduced into the second collision cell. The acceleration energy for C6H6+•
was 4002 eV. The collision cell potential was 3902 V. The peak types are
denoted. The peak marked 77+(MID) is due to the metastable ion
decomposition of C6H6+• to C6H5
+ occurring in the field-free region between
the magnetic and electric sectors.
66
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
namely without using the electric sector. The acceleration voltage in the ion
source was 4004 V and 3929 V was applied on the first collision cell in this case.
The mass scale shown in the spectrum is the one calculated with eqn. (2.12) using
the acceleration voltage in the source, 4004 V. Hence, the peaks appearing near
the integer masses correspond to ions generated in the source by electron
ionization. Two peaks appear at the nominal masses of 74.6 and 76.5. Inserting
the collision cell potential of 3929 V into eqn. (2.12), these can be identified as 12C32S2
+• and 12C32S 34S
+• generated in the collision cell. It is not certain at this
point that these CS2+• ions are generated by charge exchange with C6H6
+•. First of
all, these ions may have been generated by charge exchange with fragment ions
produced together with C6H6+• by 20 eV EI of benzene. Most of the fragment ions
appearing in the EI spectra are even-electron species and their recombination
energies, or ionization energy of the corresponding odd-electron radicals, are
usually lower than the ionization energies of the reagent gases used in this work.
Namely, charge exchange ionization of the reagent gases by the even-electron
fragments can be safely neglected. The same argument does not apply to the odd-
electron species. Among the major odd-electron fragment ions (C6H4+•, C4H4
+•,
and C4H2+•) appearing in the EI spectrum of benzene, only C4H2
+• has literature
recombination energy66 larger than the ionization energy of CS2. In this work, its
intensity was reduced to a negligible level by carrying out electron ionization at
20 eV. The second complication arises from the fact that both the ion source and
the first collision cell are evacuated by the same pumping system in the present
apparatus and hence some CS2 enters the ion source and is ionized there. Even
though the primary CS2+• ion beam intensity thus generated would not be strong,
it may contribute significantly to charge exchange ionization of CS2 in the
collision cell because the reaction is symmetric. Presence of the primary CS2+•
(typeⅠ) can not be judged from the mass spectrum in Fig. 2.10(a) because CS2+•
67
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
72 74 76 78 800
50
100
Ι
(b)
(c)
(a)x 5
Ι
Ι
ΙΙΙ
C32S2+· (ΙΙΙ)
Rela
tive
Inte
nsity
, %
m/z
48 50 52 54 56 580
50
100
ΙΙΙΙ
Ι
Ι
CH335Cl+· (ΙΙΙ)
Rela
tive
Inte
nsity
, %
m/z
72 74 76 78 800
50
100
ΙΙΙΙ
x 5
x 5
Ι
C32S2+· (ΙΙΙ)
Rela
tive
Inte
nsity
, %
m/z
Fig. 2.10 Mass spectra obtained under the single-focusing condition. The acceleration energy in the source was 4004 eV and the collision cell potential was 3929 V. C6H6
and CS2 were introduced into the ion source and the first collision cell, respectively. C6H6 was ionized (a) by EI and (b) by CI at the 0.02 Torr source pressure. (c) C6D6
and CH3Cl were introduced into the ion source and the first collision cell, respectively and C6D6 was ionized by EI. The instrument was tuned to maximize the type III ion signals. The types of the major signals are denoted.
68
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
3900 3915 3930 3945 3960
Inte
nsity
Translational Energy, eV
Fig. 2.11 MIKE spectrum recorded by setting the magnetic field to transmit
the C32S2+•(III) ion in Figure 2.10(b) and scanning the electric sector.
69
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
and its isotopomer overlap with benzene ion and its fragment at m/z 78 and 76. To
see the effect of the reagent gas leakage to the ion source more clearly, we
recorded the 20 eV EI mass spectrum of C6D6 with CH3Cl in the first collision
cell, Fig. 2.10(c). In this spectrum, m/z 52, 54, and 56 correspond to the typeⅠ
ions of C4D2+•, C4D3
+, and C4D4+• and m/z 48.7 and 50.6 to the typeⅢ CH3
35Cl+•
and CH337Cl+•, respectively. m/z 50 can be the typeⅠ ions of C4D+, CH3
35Cl+•, or
their mixture. By comparing with mass spectrum recorded in the absence of the
reagent gas, we estimated that 75 % of the m/z 50 intensity was C4D+ and the
remainder CH335Cl+•. Namely, the typeⅠ CH3Cl+• is weaker than the typeⅢ one
in the mass spectrum and is not likely to contribute significantly to the typeⅢ
signal.
An easy way to eliminate the above complications, for the benzene case at
least, is to use a CI source and increase the benzene pressure in the source. Fig.
2.10(b) shows the mass spectrum recorded at the benzene pressure of 0.02 Torr,
so-called CI1 condition in previous section, with CS2 in the first collision cell. Not
only the fragment ions from benzene but also the typeⅠ CS2+• (m/z 76) ions are
virtually absent in this spectrum. Regardless, the typeⅢ CS2+• ions appear
prominently at m/z 74.6. Then, these ions must have been generated by reaction of
CS2 with benzene ion. Further evidences showing that this reaction is the charge
exchange with benzene ion in the long-lived excited electronic state will be
presented later. We will just point out here that this long-lived state persists in the
CI1 condition, namely at the source benzene pressure of 0.02 Torr, according to
previous section. We also recorded the MIKE spectrum with the magnetic sector
set to transmit the typeⅢ CS2+• ion, Fig. 2.11. A single peak appears at 3929 eV
in the MIKE spectrum, as expected for a typeⅢ ion generated in the collision cell
floated at 3929 V.
70
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
0 200 400 600 800 100010-6
10-5
10-4
10-3
10-2
10-1
(b) 1,3-C4H6+
CS2+
CH3Cl+
CH3F+
CH4+
Rela
tive
Yiel
d, Ι
(A+. ) /
Ι(C 6H
6+. )
Primary Ion Translational Energy, eV
0 200 400 600 80010-6
10-5
10-4
10-3
10-2
10-1
(a)
Rela
tive
Yiel
d, Ι
(A+. ) /
Ι(C 6H
6+. )
1,3-C4H6+
CS2+
CH3Cl+
CH3F+
CH4+
Primary Ion Translational Energy, eV
Fig. 2.12 Relative yields of the reagent gas ions, I(A+•)/I(C6H6
+•), vs. the primary ion translational energy. Benzene ions were generated by CI at (a) 0.02 Torr (CI1) and at (b) 0.1 Torr (CI2) source pressures. Charge exchange ionization was done in the first collision cell. For consistency, the instrument was tuned to maximize the primary ion signal. In (a), CH3F+• and CH4
+• signals were hardly detectable at low energy while CS2
+•, CH3Cl+•, CH3F+•, and CH4+• were not detectable at low
energy in (b).
71
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
We have mentioned earlier that not only the charge exchange but also other
processes such as collisional impact ionization may contribute to the above typeⅢ
ion signal. Checking such a possibility is all the more important because the mass
spectral data presented so far were obtained at the inside-the-cell primary ion
translational energy (K = eVS - eVC) of 50 ~ 100 eV, which is much larger than
thermal. In this regard, we measured the relative yield of the reagent gas ions,
I(A+)/I(C6H6+•), as a function of the primary ion translational energy for the five
reagent gases adopted in this work. The results obtained with benzene ions
generated under the CI1 condition are shown in Fig. 2.12(a). It is to be noted that
the relative yields of 1,3-C4H6+•, CS2
+•, and CH3Cl+•, which can be generated via
electronically exoergic charge exchange with C6H6+• in the long-lived excited
state, remain high up to ~ 100 eV of the primary ion translational energy and then
decrease at higher energy. On the other hand, charge exchanges of CH3F and CH4
with C6H6+• in the long-lived excited state are electronically endoergic and the
corresponding products, CH3F+• and CH4+•, are hardly detectable at low energy.
Their production at higher collision energy (≥400 eV) with much smaller cross
sections than the above exoergic cases is likely due to collision-induced endoergic
charge exchange or collisional impact ionization.67 We are not much interested in
the exact nature68,69 of the mechanism involved in the generation of the reagent
gas ions at high energy. The main highlight of the data in this figure is that
reagent ions which can be generated by exoergic charge exchange with C6H6+• in
the excited state are produced with good yields at the primary ion energy of 50 ~
100 eV and not for ions that require endoergic charge exchange.
As a further test of the method, we repeated the same experiment, but using
benzene ions generated at the higher CI source pressure of 0.1 Torr. This was
called the CI2 condition in the previous work which found that benzene ions were
completely relaxed to the ground electronic state under this condition. The data in
72
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
Fig. 2.12(b) show that 1,3-C4H6+• is generated with high yield at ~50 eV of the
primary ion translational energy while the molecular ions of the remaining four
reagents were not observable. This is compatible with the fact that charge
exchange of 1,3-C4H6+• with benzene ion in the ground state is electronically
exoergic while those for the other reagents are endoergic. The results in Figures
5(a) and 5(b) strongly suggest that prominent reagent ion signals are due to the
electronically exoergic charge exchange, even at 50 eV of primary ion
translational energy.
Atomic charge exchange processes have been heavily investigated over the
years. 61,70-72 It is well known that the cross section for a symmetric charge
exchange is the maximum at the thermal energy and decreases with increasing
impact velocity. On the other hand, the cross section for non-symmetric charge
exchange is small at low impact velocity, rises to a maximum, and falls off as the
velocity increases. It is also known that the impact velocity at the maximum cross
section is well predicted by the Massey’s adiabatic criterion. Polyatomic charge
exchange has not been as much investigated.59,60 However, it is well known in the
field of chemical ionization mass spectrometry39 that the electronically exoergic
charge exchange is especially efficient. It is thought that the negative energy
defect, or exoergicity, is compensated by vibrational excitation of products,
rendering the process energetically symmetric, or near resonant. We have shown
here that the same reasoning may also hold for the charge exchange by ions in the
excited electronic state. To test the exoergicity rule, we performed additional
charge exchange experiments for the reagent gases used in this work with the
primary ions generated from these gases, in various combinations. The only
exception to the exoergicity rule was the electronically endoergic reaction CH3F+•
+ CH4 → CH3F + CH4+•, which showed a good yield at low translational energy.
This is not surprising, however, because the cross section for this reaction with
73
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
very small endoergicity (0.04 eV) is expected to be large even at small primary
ion translational energy according to the Massey’s adiabatic criterion.73
2.2.4 Conclusions
Most of the experimental methods employed so far to find the presence of
isolated electronic states rely on the measurement of dissociation or fluorescence
from these states. These methods are often useless when the excited state of
interest is non-fluorescing and lies well below the dissociation threshold such as
the benzene ion in the B state. Charge exchange ionization in collision cells
investigated in this work can be useful in such cases. The present observation that
the cross sections for electronically exoergic charge exchange reactions are much
larger than and hence can be distinguished from endoergic ones will be
particularly useful to make a reasonable estimate for the energy level of the
excited state involved. Also the fact that large difference in cross sections can be
observed at 50 ~ 100 eV of the primary ion translational energy means that a
simple instrumentation can be used for such an investigation. Even though the
method can detect the presence of long-lived excited electronic states, it is not
adequate to determine the identity of the state. Development of spectroscopy-
based techniques will be needed for this purpose.
2g2E~
2.3 Monosubstituted Benzene Cations
In previous section on photodissociation of benzene cation, we found
evidences for a very long lifetime (20 µsec or longer) of its A~ 2E2g state. In a
subsequent section, a method based on charge exchange in collision cells of a
modified double focusing mass spectrometer was developed for routine search for
long-lived excited states with conventional mass spectrometry. Observations
74
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
made were summarized as a hypothesis termed the exoergicity rule (will be
discussed in following subsection).
We have searched for other polyatomic molecular ions in long-lived excited
electronic states using the charge exchange method introduced in the previous
section. Molecules displaying well-resolved vibrational structure in an excited
state photoelectron peak have been chosen as candidates. For molecules
displaying broad excited electronic state photoelectron peaks, rapid internal
conversion of the molecular ions in the excited electronic states generated initially
by electron ionization would result in molecular ions in the ground electronic state,
which can be probed by the above charge exchange method. Some of such
molecules were also investigated as counter examples. Results from investigation
on monosubstituted benzene cations are shown in this section.
2.3.1 Charge Exchange Ionization and Exoergicity Rule
Charge exchange is an important class of ion-molecule reactions. Charge
exchange between atomic species has been heavily investigated over the years.
The cross sections of atomic charge exchange processes are known to be very
sensitive to the absolute value of the change of internal energy (∆E), for the
reaction (2.13):
A+ + B → A + B+, ∆E. (2.13)
∆E = IE(B) – RE(A+) (2.14)
Here, IE and RE denote the ionization and recombination energies, respectively,
both of which are defined as positive quantities. For a resonant (∆E=0) charge
exchange, the cross section is maximum at low collision energy and decreases
rapidly as the collision energy increases. For endoergic (∆E>0) or exoergic
(∆E<0) charge exchange, the cross section increases with the collision energy,
75
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
reaches a maximum, then decreases at higher collision energy. The adiabatic
maximum rule, which is a propensity rule based on the analysis of a large volume
of experimental data, can predict the impact velocity (ν) at which the cross section
is maximum with surprising accuracy5 as in (2.15):
hEa ∆
=ν (2.15)
Use of 7 Å for the adiabatic parameter, a, has been reported to be adequate.
2.3.2 Experimental Setup
A schematic diagram of the double focusing mass spectrometer with reversed
geometry (VG Analytical ZAB-E) modified for the present charge exchange study
is shown in Fig. 2.8. Details of the experimental method were described in
previous section.
There are two technical difficulties in the present method of finding the
presence of a long-lived excited electronic state based on ∆ of the charge
exchange reaction. One is the charge exchange by fragment ions generated in the
ion source which have recombination energies comparable to or larger than that of
the excited state being probed. These are usually odd-electron species which are
not as abundant as even-electron species. By lowering the electron energy used
for ionization, say to 20 eV, interference from high recombination energy
fragment ions could be avoided in this work. The second complication arises from
collision gas leakage to the ion source and its ionization in the source. When this
ion enters the collision cell, symmetric charge exchange occurs in the cell and
hence generates collision gas ion. Possibility of such a process must be suspected
when collision gas ion signal generated in the ion source is much larger (by 10 or
more) than the same ion signal generated in the collision cell. As shown in the
E
76
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
inset of Fig. 2.8, we installed an additional slit between the ion source chamber
and the collision cell assembly evacuated by analyzer pumping system to improve
the differential pumping. Care was taken in all the measurements such that the
signal due to collision gas ion generated in the ion source is less than or negligible
compared to the signal of the same ion generated in the collision cell.
Both of the above complications can be eliminated by selecting the precursor
ion beam with the magnetic sector and observing charge exchange occurring in
the collision cell (‘second cell’) located between the magnetic and electric sectors
and floated at high voltage (VC′). Then, ions originating from the collision cell
would have the corresponding kinetic energy (eVC′) and can be identified by the
electric sector even though their exact identity cannot be determined. The second
cell experiment, which is complementary to the above first cell one, was also
performed in this work.
Iodobenzene and benzonitrile were purchased from TCI (Tokyo) and WAKO
Pure Chemical (Osaka), respectively. 1,3-butadiene and ethane were purchased
from Matheson (Parsippany, NJ). Other chemicals were purchased from Aldrich
(Milwaukee, WI). All the chemicals were of the highest purity commercially
available and were used without further purification.
2.3.3 Results
As was mentioned in a previous section, three types (I, II, and III) of ions
appear in the mass spectrum obtained by scanning the magnetic sector with charge
exchange reagent gas in the collision cell floated at high voltage (VC). These are
ions generated in the ion source (type I), their collision-induced dissociation
products (type II) generated in the collision cell, and ions originating from the
collision gas (type III) generated by charge exchange. Taking VS as the
77
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
acceleration voltage in the ion source, the translational energies after exiting the
collision cell can be expressed as eqns. (2.9) ~ (2.11).
Since the translational energies of type II and III ions differ from that of the
ordinary ions (type I), they will not appear at their ordinary m/z positions in the
spectrum. Measuring their effective m/z positions and analyzing with eqns. (2.9) ~
(2.11), their correct m/z values can be obtained. Also, the fact that their positions
in the spectrum move with the collision cell potential, eqns. (2.9) and (2.10), helps
positive identification of these ions. Ionization energies of collision gases used are
listed in Table 2.3. For each candidate ion, some of the collision gases in the table
were chosen as needed. Recombination energies of the candidate ionic states are
listed in Tables 2.4 and 2.5.
Chlorobenzene was the first molecule chosen for investigation in this work. Its
high resolution photoelectron spectrum has been reported recently,74 which shows
well- resolved vibrational structures for the ground state peak (3b1)−1 X~
B~
2B1 and
the third and fourth peaks corresponding to the hole states (6b2)−1 2B2 and
(2b1)−1 C~ 2B1. The vibrational bandwidths for the B~ 2B2 state were comparable to
those for X~ 2B1 indicating that the band broadening was mostly due to apparatus
rather than rapid relaxation. On the other hand, the vibrational bands for the C~ 2B1
peak were broader than the above states. Hence, the B~ 2B2 state was the prime
candidate in the present search for long-lived excited electronic states. The
recombination energies of the X~
~
2B1, B~ 2B2, and C~ 2B1 states are 9.066, 11.330,
and 11.699 eV,74 respectively, as listed in Table 2.4. Fragment ions with m/z 77,
76, 52, 51, and 50 appear in the 20 eV electron ionization (EI) mass spectrum.
Among these ions, m/z 50 has the highest recombination energy of 10.17 eV
which is less than that of the B 2B2 state of C6H5Cl+• and hence would not
interfere with our investigation. Fig. 2.13 shows the partial mass spectra of
78
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
Table 2.3. Collision gases, their ionization energies (IE) in eV, and success/failurea to generate their ions by charge
exchange with some precursor ions.
Precursor ions Collision gas IEb, eV C6H5Cl+• C C6H5Br+• C6H5CN+•
6H5CCH+• C6H5I+• C6H5F+•
(CH3)2CHNH2 8.72 O O O O
1,3-C4H6(butadiene) 9.07
O X O
CS2 10.07 O
CH3Br 10.54 O O O X X
C2H5Cl 10.98 X
CH3Cl 11.28 O X O X
C2H6 11.52 X O
O2 12.07 X
Xe 12.12 X X X
CHF3 13.86 X
a Success and failure indicated by O and X, respectively. Symbols are not drawn when experiments were not done.
b IEs of (CH3)2CHNH2 and CH3Cl taken from ref. 75 and others from ref. 79.
79
Copyright(c)2002 by Seoul N
ational University L
ibrary. All rights reserved.(http://library.snu.ac.kr) 2004/01/28 17:41:01
Table 2.4 Recombination energiesa of the X~ 2B1, A~ 2A2, B~ 2B2, and C~ 2B1 electronic states of some monosubstituted
benzene cations and the calculated oscillator strengthsb of the radiative transitions from the ~B 2B2 states.
State c C6H5Cl+• C C C6H5Br+•6H5I+•
6H5CN+• C6H5CCH+•
X~ 2B1 9.066 (0.0000000)
8.991 (0.0000000)
8.754 (0.0000000)
9.71 (0.0000000)
8.75 (0.0000000)
A~ 2A2 9.707 (0.0000008)
9.663 (0.0000001)
9.505 (0.0000000)
10.17 (0.0000010)
9.34 (0.0000004)
B~ 2B2 11.330
10.633 9.771 11.84 10.36
C~ 2B1 11.699 11.188 10.541 12.09 11.03
a Taken from refs. 74, 77, 78, 79, and 80 for C6H5Cl+•, C6H5Br+•, C6H5I+•, C6H5CN+•, and C6H5CCH+•,
respectively. In eV. b Results from TDDFT/UB3LYP calculations with the 6-31G** basis set. LanL2DZ used for the iodine atom.
Shown inside the parentheses.
c X~ 2B1 and A~ 2A2 are the states formed by removal of an electron from the b1 and a2 benzene π orbitals of the
neutral, respectively. ~B 2B2 and ~C 2B1 are the states formed by removal of an electron from the in-plane and out-
of-plane halogen nonbonding or triple bond π orbitals of the neutral, respectively.
80
Copyright(c)2002 by Seoul N
ational University L
ibrary. All rights reserved.(http://library.snu.ac.kr) 2004/01/28 17:41:01
Table 2.5. Recombination energies (RE)a of some excited hole states of
fluorobenzene cation and the calculated oscillator strengthsb of some radiative
transitions.
Oscillator strengthb Statec Characterd RE, eV
From E~ 2B2 state From J~ 2B2 state
X~ 2B1 π2 9.20 0.0000000 0.0000000
A~ 2A2 π3 9.81 0.0000261 0.0000943
B~ 2B2 B(19) 12.24 0.0298871 0.0049450
C~ 2B1 π1 12.24 0.0000000 0.0000000
D~ 2A1 B(18) 13.04 0.0029268 0.0120808
E~ 2B2 B(14), n(F2p‖) 13.89 0.1020018
F~ 2B2 B(16) 14.62 0.0016411
G~ 2A1 B(13) 15.17 0.0126313
H~ 2B1 n(F2p⊥) 16.31 0.0000000
I~ 2A1 B(15) 16.31 0.0000091
J~ 2B2 B(14) , n(F2p‖) 16.70
a Taken from ref. 79. b Calculated at the TDDFT/UB3LYP level with the 6-311G** basis set. c Hole states will be called , , … in order of increasing energy.X~ A~ B~
d π means an electron removal from a benzene π orbital, B(i) from ith orbital of
benzene, n(F2p‖) from in-plane nonbonding 2p orbital of fluorine, and n(F2p⊥)
from out-of-plane nonbonding 2p orbital of fluorine.
81
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
0
50
100
+
+
+. .
(a)
C 4H4
C 4H3
C 4H2
I
I
I
0
50
100
+
+
+++
++
.. .
.
(b)
C 4H4
CH337
Cl
C 4H2
CH335
Cl
C 4H3
CH237
Cl
CH235
Cl
III
III
III
III
II II II I
I
I
Rela
tive
Inte
nsity
48 49 50 51 520
50
100
+
+
+
+
+
+
+
.
..
.
m/z
(c)
CH 237
Cl
II
II
C 4H4
C 4H3
C 4H2
CH337
Cl
CH335
Cl
CH235
Cl
II/II
I
III
III
III I
I
I
Fig. 2.13 Partial mass spectrum of C6H5Cl generated by 20 eV EI recorded under the single focusing condition with 4006 eV acceleration energy is shown in (a). (b) and (c) are mass spectra in the same range recorded with CH3Cl in the collision cell floated at 3910 and 3960 V, respectively. Type II signals at m/z 49.3 and 50.3 in (b) and at m/z 49.6 and 50.6 in (c) are due to collision-induced dissociation of C6H5Cl+• to C4H2
+• and C4H3+, respectively. The peaks at m/z 50.6 in (b) and at m/z
50.8 in (c) are due to collision-induced dissociation of C6H5+
to C4H3+.
82
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
3880 3900 3920 39400
50
100
Rela
tive
Inte
nsity
Translational Energy, eV
Fig. 2.14 Ion kinetic energy spectrum recorded by setting the magnetic field
to transmit the type III CH335Cl+• ion in Fig. 2.13 (b) and scanning the
electric sector.
83
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
chlorobenzene recorded without (Fig. 2.13 (a)) and with (Figs. 2.13 (b) and 2.13
(c)) CH3Cl in the collision cell. The collision cell was floated at 3910 and 3960 V,
respectively, to obtain Figs. 2.13 (b) and 2.13 (c). All the peaks in the figures have
been identified based on the method described previously. Movement of type II
and III peaks, as evident when Figs. 2.13 (b) and 2.13 (c) are compared, has been
checked also. We also varied the acceleration voltage in the ion source (VS) and
confirmed that only the type II peaks moved with VS (compare eqns. (2.9) and
(2.10)). For further confirmation of type III CH3Cl+•, magnetic field of the sector
was set to transmit this ion in Fig. 2.13 b) and the electric sector potential was
scanned to measure the ion kinetic energy, Fig. 2.14. Excellent agreement
between this measurement and the potential applied to the cell, 3910 V, showed
that this ion was generated from the collision gas indeed. Since the ionization
energy of CH3Cl is 11.28 eV,75 appearance of type III CH3Cl+• and its ion-
molecule reaction product,76 CH2Cl+, means that there is an ionic species in the
precursor ion beam which has the recombination energy equal to or larger than
11.28 eV. As was mentioned earlier, some of the type III ions in this spectrum
might have been generated from CH3Cl+• in the precursor beam which was
produced from CH3Cl leaked into the ion source. CH335Cl+• and CH3
37Cl+• in the
precursor beam which did not suffer collision would appear at m/z 50 and 52,
overlapped with the C4H2+• and C4H4
+• precursor signals. The C4H2+• /C4H3
+ and
C4H4+• /C4H3
+ ratios in Figs. 2.13 (b) and 2.13 (c) are similar to the same ratio in
Fig. 2.13 (a), which indicates that the fraction of CH3Cl+• species in the precursor
beam is negligible. The fact that the collision gas leakage into the ion source is not
significant will be shown with another spectrum to be presented later. We also
attempted to obtain charge exchange signal using C2H6 (IE=11.52 eV) as collision
gas. No type III ions were generated whatsoever. To summarize, type III signals
appear when the ionization energies of collision gases are 11.28 eV or lower while
84
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
88 90 92 94 96
0
50
100
+
++
+
..
CH381
Br
CH379
Br
CH281
Br
CH279
Br
III
III
III
III
Rela
tive
Inte
nsity
m/z
Fig. 2.15 Partial mass spectrum obtained under the single focusing
condition with C6H5Br and CH3Br introduced into the ion source and
collision cell, respectively. C6H5Br was ionized by 20 eV EI and
acceleration energy was 4008 eV. Collision cell was floated at 3907 V.
85
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
they are absent when the ionization energies are 11.52 eV or larger. Based on the
exoergicity rule ( ) established in the previous section, it is concluded that
the precursor beam contains C
0≤∆E
6H5Cl+• in a long-lived excited electronic state with
the recombination energy in the range 11.28 ~ 11.52 eV, B~ 2B2 state with the
recombination energy 11.330 eV being the best candidate.
We used the same guideline, narrow vibrational bands for an excited
electronic state peak in high resolution photoelectron spectrum, to search for other
long-lived excited electronic states. The B~ 2B2 and C~ 2B1 states of bromobenzene
ion with the recombination energies of 10.633 and 11.188 eV,77 respectively, were
our next targets. A partial mass spectrum of bromobenzene recorded with CH3Br
(IE=10.54 eV) in the collision cell is shown in Fig. 2.15. Type III CH3Br+•
appears prominently in this spectrum together with its ion-molecule reaction
product, CH2Br+, indicating the presence of an ionic species in the precursor beam
with recombination energy 10.54 eV or larger. Also to be noted is that type I
CH3Br+• signals expected at m/z 94 and 96, are hardly observable indicating that
collision gas leak into the ion source is not significant. On the other hand, charge
exchange signal was not obtained with C2H5Cl (IE=10.98 eV), indicating the
presence of a long-lived state with the recombination energy of 10.54 ~ 10.98 eV.
Obviously, B~ 2B2 is such a state. Even though the C~ 2B1 state shows narrow
vibrational bandwidth in the high resolution photoelectron spectrum, this turned
out not to be long-lived in the present work. Namely, we cannot predict long
lifetime of an excited electronic state solely based on the well-resolved vibrational
structure in the photoelectron spectrum. This is due to rather poor resolution of the
high resolution photoelectron spectrometers, ~ 8 meV (64 cm−1).
It is interesting to note that both of the long-lived states found, B~ 2B2 states of
C6H5Cl+• and C6H5Br+•, have the same character, namely elimination of an
86
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
electron from a molecular orbital which is essentially halogen nonbonding p
orbital parallel to the benzene ring, n(Cl3p‖) or n(Br4p‖).74,77 The same state of
iodobenzene ion has the recombination energy of 9.771 eV.78 Even though the
vibrational structure is resolved in the corresponding photoelectron peak, each
vibrational band is noticeably broader than that of the ground state band.78 As
expected, none of the collision gas underwent charge exchange with C6H5I+•
except (CH3)2CHNH2 which has ionization energy (8.72 eV)75 smaller than the
recombination energy of the ground state C6H5I+• (8.754 eV).78 In the case of
fluorobenzene, n(F2p‖) is mixed with a σ orbital of the benzene ring resulting in
two photoelectron peaks with partial n(F2p‖) character at 13.89 and ~16.7 eV,
Table 2.5. Even though vibrational splitting is indicated for the former, each
vibrational band is rather broad.79 Failure to observe charge exchange signals
from CH3Cl (IE=11.28 eV),75 Xe (IE=12.12 eV), and CHF3 (IE=13.86 eV) is in
agreement with the above photoelectron spectral feature. A strong charge
exchange signal was observed, of course, with 1,3-butadiene which has ionization
energy (9.07 eV) maller than the recombination energy of C6H5F+• in the ground
state, 9.20 eV.79
Benzonitrile and phenyl acetylene are similar to halobenzenes in the sense that
the cylindrical symmetry of the π electron system of the triple bonds are broken
due to the presence of the benzene ring, resulting in two distinct photoelectron
peaks. Here again, the states corresponding to elimination of an electron from the
in-plane π orbitals of the triple bonds, π(C N≡ ‖) and π(C C≡
≡
‖), seem to show
narrower vibrational bands than those from the out-of-plane π orbitals, π(C N≡ ⊥)
and π(C C≡ ⊥).79,80 Recombination energies of the π(C N‖)−1 B~ 2B2 and
π(C N≡ ⊥)−1 C~ 2B1 states of benzonitrile cation are 11.84 and 12.09 eV,
respectively.79 Then, the charge exchange results in Table 2.3 that collision gases
87
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
47 48 49 50 51 52
0
50
100+
+
+
+
+
+
.
.
.
.
C 4H4
CH337
Cl
C 4H2
CH237
Cl
CH335
Cl
CH235
Cl
IIIII
III
III
III
IIIIII
II
I
I
Re
lativ
e In
tens
ity
m/z
Fig. 2.16 Partial mass spectrum obtained under the single focusing
condition with C6H5CN and CH3Cl introduced into the ion source and
collision cell, respectively. C6H5CN was ionized by 20 eV EI and
acceleration energy was 4007 eV. Collision cell was floated at 3910 V.
Type II signals at m/z 49.3, 50.3, and 51.3 are due to collision-induced
dissociation of C6H5CN+• to C4H2+•, C4H3
+, and C4H4+•, respectively. Those
at m/z 49.6 and 50.6 are due to collision-induced dissociation of C6H4+• to
C4H2+• and C4H3
+, respectively.
88
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
0
50
100
(a)
0
50
100
(b)
Rela
tive
Inte
nsity
3800 3850 3900 39500
50
100
(c)
Translational Energy, eV Fig. 2.17 Ion kinetic energy spectra recorded by introducing C6H5Br+• ((a) and (b)) and C6H5CH3
+• ((c)) in the second cell filled with CH3Br. The molecular ions were accelerated to 4 keV in the ion source. The second collision cell was floated at (a) 3910, (b) 3943, and (c) 3910 V. Arrows indicate the expected positions for ions from collision gases generated by charge exchange with the precursor ions. The major peaks appearing at 3957 and 3974 eV in (a) and (b), respectively, are due to collision-induced dissociation of C6H5Br+• to C6H5
+. The major peak appearing at 3960 eV in (c) is due to unimolecular dissociation of C7H8
+• to C7H7+
occurring outside the collision cell, but between the magnetic and electric sectors.
89
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
As counter examples, we investigated charge exchange by molecular ions of
toluene, nitrobenzene, and styrene. Photoelectron spectra of these molecules do
not show excited state peaks with well-resolved vibrational structure. We will not
show mass spectra of these molecules because the experimental results were
simple and predictable. Namely, charge exchange signals were observed only
when the ionization energy of a collision gas was smaller than the recombination
energy of the ground state ion, or the first ionization energy of the corresponding
neutral.
We also performed second collision cell experiments with various gases
introduced into the cell. Molecular ions generated in the source were accelerated
to 4 keV, mass-selected by the magnetic sector, and decelerated to 50~100 eV by
floating the cell at 3900~3950 V. Fig. 2.17 shows the ionic kinetic energy spectra
obtained by introducing bromobenzene and toluene molecular ions to the cell
filled with CH3Br and by scanning the electric sector potential. Arrows in the
spectra indicate the kinetic energy corresponding to the cell potential, or ions
originating from the collision cell. A prominent peak appeared at this position and
moved with the cell potential when bromobenzene ion was introduced into the cell
(Figs. 2.17(a) and 2.17(b)) while such a peak was absent when toluene ion was
introduced (Fig. 2.17(c)). These are in agreement with the first cell results that
CH3Br is ionized by bromobenzene ion in a long-lived state but not by toluene ion.
The time between bromobenzene ion formation in the ion source and its arrival at
the second cell is ~ 40 µsec. Then, the above observation means that some of the
bromobenzene ions in the B~ 2B2 state have survived as long as 40 µsec. We also
performed similar experiments for other ions. Since the results were the same as
those of the first cell experiments, no further spectra will be presented.
For an excited electronic state to have a very long lifetime (ten microseconds
or longer), neither radiative decay nor nonradiative decay should be efficient.
90
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
Even though efficiency of the latter is not easy to investigate, that of the former
can be estimated through the symmetry selection rule and calculation of the
oscillator strength. For chlorobenzene, bromobenzene, benzonitrile, and phenyl
acetylene cations, two electronic states lie below the long-lived excited states
B~
B~
2B2. These are (b1)−1 X~ 2B1 and (a2)−1 A~
~
2A2 correlating with the X~ 2E1g state of
benzene cation. It is to be noted that the electric dipole transition B~ 2B2 → X~
X~
2B1 is
symmetry forbidden while B~ 2B2 → A~ 2A2 is allowed. Very long lifetimes of the 2B2 states require that the oscillator strengths of the latter transitions be very
small even though they are symmetry allowed. In this regard, the oscillator
strengths were calculated with the GAUSSIAN 98 package. Geometries of the
molecular cations in the ground electronic states were optimized at the
UB3LYP/6-31G** level and oscillator strengths were obtained through the time-
dependent density functional theory (TDDFT) calculation. The oscillator strengths
for the B~ 2B2 → A~ 2A2 transitions thus obtained are listed in Table 2.4. These are
10−6 or smaller in all four cases, as required. Similar calculation was done for
iodobenzene cation using the LanL2DZ basis set for the iodine atom, which also
showed negligible oscillator strength for the B~ 2B2 → A~
~
2A2 transition. Then, the
broad vibrational bandwidth for the B peak in the photoelectron spectrum of
iodobenzene and our finding that the B~ state of iodobenzene cation is not very
long-lived must be due to rapid internal conversion. We do not have explanation
at the moment why the internal conversion from the B 2B2 states is slow for the
former four cases while it is fast for iodobenzene cation. We also calculated the
oscillator strengths for the transitions from the C~ 2B1 states. Transitions with
significant strengths were found for all five cases, especially C~ 2B1→2B1.
Namely, consideration of the radiative decay does not allow long lifetime for the
91
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
C~
B~
states. This does not mean that the C~ 2B1→ X~ 2B1 radiative transitions can be
observed for these cations because internal conversion from the C~ states may be
even faster. Similar calculations were done for the radiative transitions from the
excited states of fluorobenzene cation with the fluorine nonbonding 2p character.
Both of these states showed radiative decay channels with significant strength,
Table 2.5. Then, absence of any emission from the fluorobenzene cation generated
by electron impact as reported previously19,37 means that nonradiative decay of
these states is even faster.
We also calculated the lowest energy doublet and quartet states with an
electron in the lowest unoccupied molecular orbital (LUMO) of chlorobenzene,
bromobenzene, iodobenzene, benzonitrile, and phenyl acetylene cations at the
TDDFT/UB3LYP level. This was to check the possibility that proximity of these
states to the B~ 2B2 states would affect the lifetime of the latter states. The lowest
doublet states with an electron in LUMO were found at 4.48, 4.39, 4.28, 4.13, and
4.28 eV above the ground states while the lowest quartet states were found at 4.17,
4.39, 3.91, 3.59, and 3.95 eV above the ground states for chlorobenzene,
bromobenzene, iodobenzene, benzonitrile, and phenyl acetylene cations,
respectively. Namely, all these states were found to be located substantially (> 1.5
eV) above the B~ 2B2 states and are not expected to affect the B~ 2B2 state lifetimes.
The most intriguing of the present observations is that the B~ 2B2 state of
iodobenzene cation is not long-lived while the corresponding states of
chlorobenzene, bromobenzene, benzonitile, and phenyl acetylene cations are.
Looking at the photoelectron specta,74,77-80 one finds that the 0−0 band of the 2B2 state is well separated from the lower-lying A
~
~
2A2 state continuum for the
latter four cases while the vibrational bands of the B 2B2 state are overlapped with
the A~ 2A2 state continuum for the iodobenzene cation. Then, efficient internal
92
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
conversion of the B~ 2B2 state of iodobenzene cation to the A~ 2A2 state, and
eventually to the X~ 2B1 state, may be responsible for the vibrational band
broadening in the photoelectron spectrum and for the depletion of the B~ 2B2 state
population observed in this work.
2.3.4 Conclusions
Searching for isolated electronic states of polyatomic ions has been one of the
active research areas in the fields of ion chemistry and mass spectrometry. Most
of such states found so far were repulsive states in which dissociation could occur
rapidly prior to internal conversion. For radiative bound states, spectroscopic
techniques have been used to measure the efficiency of their internal conversion
to the ground electronic states. The most difficult to study are the nonradiative
bound states lying below the dissociation thresholds. Absence of emission from
such a state has been usually interpreted as due to rapid internal conversion even
though very long lifetime (ten microseconds or longer) can be an alternative
explanation as suggested for the B~ 2E2g state of benzene cation in previous section.
Measurement of vibrational bandwidth in photoelectron spectra is not helpful to
judge whether the lifetime of the concerned state is very long (ten microseconds
or longer) or very short (picoseconds or shorter) due to rather poor resolution of
the method.
In this section, we have used the charge exchange method to judge whether
some excited states of monosubstituted benzene cations chosen based on narrow
vibrational bandwidths in the photoelectron spectra have long lifetimes. The
method has been found simple and decisive, even though time consuming, in the
sense that the long-lived states were located nearly exactly at the recombination
energies expected from the photoelectron spectra. Present results are further
93
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
confirmation on the validity of the exoergicity rule established previously that
charge exchange between polyatomic species is efficient when ∆ , and not
otherwise.
0≤E
Among the monosubstituted benzene cations investigated, C6H5Cl+•, C6H5Br+•,
C6H5CN+•, and C6H5CCH+• have been found to possess long-lived excited
electronic states, all of which are the B~ 2B2 states and show well-resolved
vibrational structures in the photoelectron spectra. It is interesting to note that
these states arise from removal of one electron from in-plane halogen nonbonding
p orbitals or in-plane π orbitals of triple bonds. It is known that these in-plane
orbitals have almost pure halogen p or triple bond π character while the
corresponding out-of-plane orbitals possess some benzene π character.74,77,79,80 We
do not know at the moment whether or how these orbital characters are related to
the efficiency of internal conversion. Regardless, the above correlation suggests
possible presence of long-lived excited states for unsaturated aliphatic molecular
ions with substituents such as halogen and nitrile, which is under investigation in
this laboratory.
94
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
References
1. H. M. Rosenstock, M. B. Wallenstein, A. L. Wahrhaftig, and H. Eyring, Proc.
Nat. Acad. Sci. 38, 667 (1952).
2. R. A. Marcus, J. Chem. Phys. 20, 359 (1952).
3. R. L. LeRoy, J. Chem. Phys. 53, 846 (1970).
4. J. D. Rynbrandt and B. S. Rabinovitch, J. Phys. Chem. 75, 2164 (1971).
5. A. Lee, R. L. LeRoy, Z. Herman, R. Wolfgang, and J. C. Tully, Chem. Phys.
Lett. 12, 569 (1972).
6. M. Kawasaki, K. Kasatani, H. Sato, H. Shinohara, and N. Nishi, Chem. Phys.
88, 135 (1984).
7. Y. Lee and S. J. Lin, J. Chem. Phys. 108, 134 (1998).
8. D. Y. Kim, J. C. Choe, and M. S. Kim, J. Chem. Phys. 113, 1714 (2000).
9. C. Lifshitz and F. A. Long, J. Phys. Chem. 69, 3746 (1965).
10. J. H. D. Eland, R. Frey, A. Kuestler, H. Schulte, and B. Brehm, Int. J. Mass
Spectrom. Ion Phys. 22, 155 (1976).
11. W. J. van der Hart, L. J. de Koning, N. M. M. Nibbering, and M. L. Gross,
Int. J. Mass. Spectrom. Ion Processes 72, 99 (1986).
12. B. Andlauer and Ch. Ottinger, J. Chem. Phys. 55, 1471 (1971).
13. H. M. Rosenstock, J. T. Larkins, and J. A. Walker, Int. J. Mass Spectrom. Ion
Phys. 11, 309 (1973).
14. T. Baer, G. D. Willett, D. Smith, and J. S. Phillips, J. Chem. Phys. 70, 4076
(1979).
15. H. Kühlewind, A. Kiermeier, and H. J. Neusser, J. Chem. Phys. 85, 4427
(1986).
95
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
16. Kiermeier, H. Kühlewind, H. J. Neusser, E. W. Schlag, and S. H. Lin, J.
Chem. Phys. 88, 6182 (1988).
17. S. J. Klippenstein, J. D. Faulk, and R. C. Dunbar, J. Chem. Phys. 98, 243
(1993).
18. Th. L. Grebner and H. J. Neusser, Int. J. Mass. Spectrom. 185/186/187, 517
(1999).
19. M. Allan, J. P. Maier, and O. Marthaler, Chem. Phys. 26, 131 (1977).
20. J. H. Miller and L. Andrews, Chem. Phys. Lett. 72, 90 (1980).
21. J. H. Miller, L. Andrews, P. A. Lund, and P. N. Schatz, J. Chem. Phys. 73,
4932 (1980).
22. Braltbart, E. Castellucci, G. Dujardin, and S. Leach, J. Phys. Chem. 87, 4799
(1983).
23. K. Walter, R. Weinkauf, U. Boesl, and E. W. Schlag, Chem. Phys. Lett. 155,
8 (1989).
24. J. G. Goode, J. D. Hofstein, and P. M. Johnson, J. Chem. Phys. 107, 1703
(1997).
25. W. von Niessen, L. S. Cederbaum, and W. P. Kraemer, J. Chem. Phys. 65,
1378 (1976).
26. H. Köppel, L. S. Cederbaum, and W. Domcke, J. Chem. Phys. 89, 2023
(1988).
27. H. Köppel, Chem. Phys. Lett. 205, 361 (1993).
28. M. Döscher and H. Köppel, Chem. Phys. 225, 93 (1997).
29. (a) U. Müller and G. Stock, J. Chem. Phys. 108, 7516 (1998); (b) K. Muller-
Dethlefs and J. B. Peel, J. Chem. Phys. 111, 10550 (1999).
30. S. R. Long, J. T. Meek, and J. P. Reilly, J. Chem. Phys. 79, 3206 (1983).
96
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
31. P. Baltzer, L. Karlsson, B. Wannberg, G. Öhrwall, D. M. P. Holland, M. A.
MacDonald, M. A. Hayes, and W. von Niessen, Chem. Phys. 224, 95 (1997).
32. E. E. Rennie, C. A. F. Johnson, J. E. Parker, D. M. P. Holland, D. A. Shaw,
and M. A. Hayes, Chem. Phys. 229, 107 (1998).
33. B. S. Freiser and J. L. Beauchamp, Chem. Phys. Lett. 35, 35 (1975).
34. P. N. T. van Velzen and W. J. van der Hart, Chem. Phys. 61, 325 (1981).
35. R. C. Dunbar, Chem. Phys. Lett. 125, 543 (1986).
36. (a) R. L. Whetten, K. J. Fu, and E. R. Grant, J. Chem. Phys. 79, 2626 (1983);
(b) K. Siglow, R. Neuhauser, and H. J. Neusser, J. Chem. Phys. 110, 5589
(1999); (c) K. Siglow and H. J. Neusser, J. Chem. Phys. 112, 647 (2000).
37. M. Allan and J. P. Maier, Chem. Phys. Lett. 34, 442 (1975).
38. J. P. Maier, O. Marthaler, M. Mohraz, and R. H. Shiley, Chem. Phys. 47, 295
(1980).
39. A. G. Harrison, Chemical Ionization Mass Spectrometry (CRC, Boca Raton,
1992).
40. M. S. Kim, M. Rabrenović, and J. H. Beynon, Int. J. Mass. Spectrom. Ion
Processes 56, 71 (1984).
41. G. F. Weston, Ultrahigh Vacuum Practice (Butterworth, London, 1985).
42. R. L. Summers, NASA Tech. Note TND-5285 (National Aeronautics and
Space Administration, Washington, DC, 1969).
43. D. A. Dahl and J. E. Delmore, (EGG-CS-7233, 1988, Rev. 2).
44. R. G. Cooks, J. H. Beynon, R. M. Caprioli, and G. R. Lester, Metastable Ions
(Elsevier, Amsterdam, 1973).
45. R. G. Neuhauser, K. Siglow, and H. J. Neusser, J. Chem. Phys. 106, 896
(1997).
97
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
46. R. C. Dunbar and H. H. Teng, J. Am. Chem. Soc. 100, 2279 (1978).
47. E. W. Fu and R. C. Dunbar, J. Am. Chem. Soc. 100, 2283 (1978).
48. J. H. Chen, J. D. Hays, and R. C. Dunbar, J. Phys. Chem. 88, 4759 (1984).
49. J. C. Choe and M. S. Kim, J. Phys. Chem. 95, 50 (1991).
50. S. H. Lim, J. C. Choe, and M. S. Kim, J. Phys. Chem. A 102, 7375 (1998).
51. W. G. Hwang, J. H. Moon, J. C. Choe, and M. S. Kim, J. Phys. Chem. A 102,
7512 (1998).
52. Y. H. Yim and M. S. Kim, Int. J. Mass. Spectrom. Ion Processes 123, 133
(1993).
53. T. A. Lehman and M. M. Bursey, Ion Cyclotron Resonance Spectrometry
(John Wiley & Sons, Inc., New York, 1976).
54. R. C. Weast (ed.), Handbook of Chemistry and Physics (CRC Press, Boca
Raton, 1989).
55. S. C. S. Rao and C. Fenselau, Anal. Chem. 50, 511 (1978).
56. W. J. van der Hart, Int. J. Mass. Spectrom. Ion Processes 130, 173 (1994).
57. S. G. Lias, J. E. Bartmess, J. F. Liebman, J. L. Holmes, R. D. Levin, and W.
G. Mallard, J. Phys. Chem. Ref. Data 17, Suppl. No. 1 (1988).
58. NIST Standard Reference Database Number 69 (NIST chemistry webbook,
November 1998 Release).
59. E. Lindholm, Charge Exchange and Ion-Molecule Reactions Observed in
Double Mass Spectrometers, In Ion-Molecule Reactions in the Gas Phase; P.
Ausloos, Ed.; (American Chemical Society: Washington, 1966).
60. (a) J. H. Futrell and C. D. Miller, Rev. Sci. Instr. 1966, 37, 1521-1526. (b)
Hughes, B. M.; Tiernan, T. O. J. Chem. Phys. 1971, 55, 3419-3426.
61. Massey, H. S. W.; Gilbody, H. B. Electronic and Ionic Impact Phenomena,
98
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
Vol. IV; Oxford University Press.: London, 1974.
62. Kim, Y. H.; Kim, M.S. Rapid Commun. Mass Spectrom. 1991, 5, 25-29.
63. McDowell, C. A. Mass Spectrometry; McGraw-Hill: New York, 1963; p 7.
64. Cooks, R. G.; Beynon, J. H.; Caprioli, R. M.; Lester, G. R. Metastable Ions;
Elsevier: Amsterdam, 1973; p 44.
65. NIST Standard Reference Database Number 69 (NIST chemistry webbook,
February 2000 Release).
66. Lias, S. G.; Bartmess, J. E.; Liebman, J. F.; Holmes, J. L.; Levin, R. D.;
Mallard, W. G. J. Phys. Chem. Ref. Data 1988, 17, (Suppl. No. 1).
67. Maier II, W. B. J. Chem. Phys. 1965, 42, 1790-1804.
68. Marx, R. Charge Transfers at Thermal Energies: Energy Disposal and
Reaction Mechanisms, In Kinetics of Ion-Molecule Reactions; Ausloos, P.,
Ed.; Plenum Press.: New York, 1979; p 103.
69. (a) Mauclaire, G.; Dera, R.; Fenistein, S.; Marx, R. J. Chem. Phys. 1979, 70,
4017-4022. (b) Mauclaire, G.; Dera, R.; Fenistein, S.; Marx, R. J. Chem.
Phys. 1979, 70, 4023-4026.
70. Giese, C. F.; Maier II, W. B. J. Chem. Phys. 1963, 39, 197-200.
71. Hasted, J. B. Physics of Atomic Collisions; Butterworths: Washington, 1964.
72. Lehrle, R. S.; Parker, J. E.; Robb, J. C.; Scarborough, J. Int. J. Mass
Spectrom. Ion Phys. 1968, 1, 455-469.
73. See M. S. Kim, Org. Mass Spectrom. 1991, 26, 565-574 for sample
calculations.
74. A. W. Potts, D. Edvardsson, L. Karlsson, D. M. P. Holland, M. A.
MacDonald, M. A. Hayes, R. Maripuu, K. Siegbahn, and W. von Niessen,
Chem. Phys. 254, 385 (2000).
99
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
75. K. Watanabe, T. Nakayama, and J. Mottl, J. Quant. Spectrosc. Radiat.
Transfer 2, 369 (1962).
76. A. A. Herod, A. G. Harrison, and N. A. McAskill, Can. J. Chem. 49, 2217
(1971).
77. D. M. P. Holland, D. Edvardsson, L. Karlsson, R. Maripuu, K. Siegbahn, A.
W. Potts, and W. von Niessen, Chem. Phys. 252, 257 (2000).
78. D. M. P. Holland, D. Edvardsson, L. Karlsson, R. Maripuu, K. Siegbahn, A.
W. Potts, and W. von Niessen, Chem. Phys. 253, 133 (2000).
79. K. Kimura, S. Katsumata, Y. Achiba, T. Yamazaki, and S. Iwata, Handbook
of HeI Photoelectron Spectra of Fundamental Organic Molecules (Japan
Scientific Societies Press, Tokyo, 1981).
80. J. W. Rabalais and R. J. Colton, J. Electron Spectrosc. Relat. Phenom. 1, 83
(1972/73).
100
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
Chapter 3
Coherent Vacuum Ultraviolet Radiation
One-photon ionization which is the most general and cleanest photoionization
method, usually occurs in the vacuum ultraviolet (VUV) region and the progress in the
field of VUV spectroscopy has been limited by the available VUV light source.1-3
Because there is currently no media able to provide suitable light at shorter
wavelengths than 190 nm, the coherent and tunable VUV radiations have been
developed, which also hold promise for many new applications in atomic and
molecular studies.4-6 The easiest way to generate coherent VUV light is by third
harmonic generation (THG) or four-wave mixing (FWM) in a nonlinear gaseous
medium.1 Of course, all instruments used to study these sources must be evacuated, as
vacuum ultraviolet (VUV) can not travel freely in air. In addition, it is necessary to
separate and monitor the VUV light of interest from the unwanted and residuals.
In this chapter, the basic concepts of nonlinear optics and the generations of
coherent and powerful VUV pulse at 104 ~ 143 nm by four-wave mixing in nonlinear
medium are summarized.
3.1 VUV Generation in Gaseous Nonlinear Medium
3.1.1 General Principles
When a light consisting of electric and magnetic fields propagates through matter, a
macroscopic polarization P which is induced by the electric field E of incident light
101
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
may be described by a power series expansion in )( iE ω :
, (3.1) K+++= 3)3(3
2)2(2
)1( )()()( iii EEEPiii
ωχωχωχ ωωω
Here, is the linear susceptibility of the medium and is related to the refractive
index n by n . The quantities are the nonlinear susceptibilities
and they describe the nonlinear-optical properties of the medium. The importance of
the induced polarization can be understood from the fact that any oscillating dipole
also emits radiation, at the frequency of oscillation. Hence, as a result of the nonlinear
effects, the radiating dipoles can be used to generate light at new frequencies.
(1)iωχ
(1)2i
1 ωχ+= K,, )3(3
)2(2 ii ωω χχ
In general, for a material with inversion symmetry, for example, atomic gases such
as Xe, Kr, or Ne and metal vapors such as Hg, Cd, or Ze, there are no even powers of
the field in the expansion of the polarization for symmetry reason. The lowest-order
nonlinearity is then the cubic term in eqn. (3.1) and this term is responsible for all four-
wave mixing processes (FWM) or third-harmonic generation (THG). Then, in the gas
phase, the number density, , is so low that local field effects are small and hence,
macroscopic susceptibility describing nonlinear polarization induced by the laser is
simply and this polarization may be introduced into the wave equation as a
nonlinear source term, (THG) or (FWM). Furthermore,
the VUV intensity (I
N
VUV =
)3(χN
i3ωω 21VUV 2 ω+ω=ω
VUV) of the radiation generated at frequency is VUVω
)()()()( 3i
2VUV
32VUV kbFINI )( ∆ωωχ∝ω ; THG (3.2)
or
)()()()()( 22
12
VUV32
VUV kbFIINI )( ∆ωωωχ∝ω ; FWM (3.3)
where I is the laser intensity, is the phase-matching factor, b is the beam
confocal parameter (Rayleigh length of the focused Gaussian beam), and
)( kbF ∆
102
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
iVUV 3 ωω −=∆ kkk or ∆ is the wave vector mismatch between the
generated and the incident waves. In a focused beam condition (b<<L), where L is the
length of the nonlinear medium, can be approximated into
21VUV 2 ωωω −−= kkkk
)( kbF ∆
)()( 2∆π=∆ expkbkbF
)(22 1VUV1 nnk −νπ=∆
)0()( <∆∆ kkb
=0 0( ≥ . (3.4) )∆ k
Also, the wavevector mismatch ∆k between the waves is given by
[ ])(2 2VUV2 nn −ν+ . (3.5)
However, one can compensate for the phase mismatch, choosing the harmonic
radiation to lie in the region of anomalous dispersion on the high energy side of a
resonance line and adding a second normally dispersive gas.
Although the nonlinear susceptibilities for gases are generally much smaller than
the corresponding values for metals, the nonlinear processes in rare gases can be easily
accomplished, considering the experimental difficulty in preparation the metal vapors
of high density. Extensive wavelength tunability with rare gases has been achieved by
Hilbig and Wallenstein7 and the typical conversion efficiency for Kr or Xe, was
reported by 10-6. The VUV radiations generated in rare gases are listed in Table 3.1.
While, four-wave mixing (FWM) in metal vapor10 offers advantage of high efficiency,
which can be achieved by using relatively low photon energy and intensity of incident
radiation. Especially, Hg or Mg vapor, was found to be a very efficient nonlinear
medium which demonstrated conversion efficiency of ~ 10-3. The VUV radiations
generated in metal vapors are also listed in Table 3.2.
3.1.2 Wavelength Calibration5
Whenever one performs high-resolution spectroscopy experiments, one should
calibrate the wavelength in the VUV light. The easiest way for wavelength calibration
103
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
Table 3.1 Tunable generation in rare gases.
Wavelength Nonlinear medium Processes
195 ~ 163 Xe 2×266±λs, λia
147 ~ 118 Xe 2×266±λs
147 ~ 140 Xe:Kr 3λDye9
130 ~ 110b Kr 2λUV+λL10
123.5 ~ 120 Kr:Ar 3λDye9
106c Xe 2×31811
a Parametric oscillator with signal and idler wavelengths λs, λi. Ref.7.
b λUV from a range of laser dyes.
c Tunable over a small region.
Table 3.2 Tunable generation in metal vapors.
Wavelength Nonlinear medium Processes Primary Laser
174 ~ 145 Mg 2×459.7+λDye N2-Dye12
160 ~ 140 Mg 2×431.0+λ Dye N2-Dye13
140 ~ 106 Zn 2×358.5+λ Dye KrF-Dye14
125 ~ 117 Hg 2×312.8+λ Dye Nd:YAG-Dye15
115 ~ 104.5 Hg 2×268.8+λ Dye Nd:YAG-Dye16
104
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
is to utilize the optogalvanic effect. The method is to peel off a part of the mixed beam
and pass it through a hollow cathode source. This is useful technique because of
considerably broaden range of available lines to sue as standards and providing the
sharp lines in the UV and VUV. One may use either sealed lamps or open, depending
on the wavelength range. Hybrid fillings of rare gas-metal combinations (for example,
Fe/Ne) are available with the sealed units.
3.1.3 Measurements of VUV Intensity17
To measure the absolute intensity of radiation of any wavelength, a detector of
which gain is known, must be used. This can be calibrated with a standard source
whose intensity of radiation is accurately informed. Calorimetirc measurement of heat
produced by photon would be absolute if photon is completely absorbed and
transformed into heat. However, the fact that the intensity of VUV generated is usually
low partially restricts the use of this measurement due to insensitive. Solar blind
detectors which are inert to UV or visible, can be used with requirement of any
calibration.
The most accurate and reproducible method of measuring absolute intensities is
that utilizing the principle of photoionization of a suitable gas. Actually, any gas would
suffice provided its photoionization yield known. The photoionization yield of a gas is
defined as the number of ions produced per photon absorbed by the gas.
We used single ion chamber built in our laboratory to measure the intensity of
VUV light generated by four-wave mixing in Kr gas or Hg vapor. Single ion chamber
current will be monotonically increased by absorption of VUV photons as the pressure
of photoionization gas (NO in He) increases. However, due to distance between
collection electrodes and collimation lens, collected ions reached maximum signal and
then began to decrease as pressure increased. It can be calculated as follows. As VUV
photon passes absorbing gas, the number of photons N (x) will decrease according to
105
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
���
Fig. 3.1 The number of photoions by ion chamber currents measured as function
of the pressure of No/He. Voltage between two electrodes in ion chamber is 50 V.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
Beer’s law,
)exp()( 0 xdNxN ε−= (3.6)
where x denotes the position along beam path, d denotes the number density of gas,
and ε denotes absorption coefficient. Fig. 3.1 shows experimental data and its fit. The
maximum ion signal in data is 4.03×109 ions/pulse and the absolute number of VUV
photons ( ) obtained from the curve fit is 9.2×100N 9 photon/pulse.
3.2 Four-Wave Difference Frequency Mixing in Kr Gas
To excite the Kr 5p[1/2]0 – 4p6 transition for VUV generation by four wave
difference mixing, (as shown in Fig. 3.2) the light at 212.5 nm (∼0.5 mJ/pulse) was
generated by frequency tripling of 637.6 nm output of a dye laser (Continuum
ND6000) pumped by the second harmonic of an Nd:YAG laser (Continuum PL8000).
Another dye laser output (420 ∼ 800 nm) pumped by the second or third harmonic of
the second Nd:YAG laser (Continuum Surelite II) was combined with the 212.5 nm
light and loosely focused with a fused silica lens (f = 50 cm) in the Kr cell to generate
the VUV light tunable in the 123 ∼ 142 nm range. A MgF2 lens (f = 25 cm) was placed
at the exit of the Kr cell and the laser beams were aligned off-centered at the lens to
separate the residual light beams (UV and visible) from the VUV light, which was
focused onto the molecular beam. (as shown in Fig. 1.9) The optimized Kr pressure in
the cell was 0.1 ∼ 13 Torr. Its precision was ±0.5 cm-1 in the visible region.
3.3 Four-Wave Sum Frequency Mixing in Hg vapor
VUV in the region of 107 ∼ 127 nm used to measure the spectra of the cations in
the excited states was generated by four-wave sum frequency mixing in Hg. A
schematic diagram of the experimental apparatus is in Fig. 3.4. The UV light (ωUV =
107
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
4p6
5p[5/2]2
λ1=212.6nm λ1=202.3nm
5p[1/2]0
λVUV=123 ~ 142nm
λ2=400 ~ 800nm
λVUV=120 ~ 123nm
λ2=573 ~ 630nm
Fig. 3.2 Schematic diagram for four-wave difference frequency mixing in Kr
gas.
108
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
312.8 nm, ∼ 2 mJ/pulse), which excites the Hg 61S0 – 71S0 transition via the two-
photon resonance, was generated by frequency-doubling of an output of a dye laser
(Continuum ND6000) pumped by the second harmonic of an Nd:YAG laser
(Continuum PL8000) with ∼ 7 nsec pulse duration and 10 Hz repetition rates. ωS (2 ∼ 6
mJ/pulse) at 339 ∼ 675 nm was generated by the second dye laser (Lambda Physik
SCANMATE 2E) pumped by the second or third harmonic of another Nd:YAG laser
(Continuum PL8010). The two laser beams were combined with a dichroic mirror and
tightly focused using an achromatic lens (f = 20 cm) onto the Hg vapor. The four-wave
mixing Hg cell was designed similar to that of Hilbig et. al.8 A LiF lens (f = 20 cm)
was placed at the exit of the Hg cell and the laser beams were aligned off-center to
separate the VUV light from the residual UV and visible lights at the interaction region
with the molecular beam. The VUV output in the 107 ∼ 127 nm region was optimized
at the Hg vapor pressure close to 0.9 Torr with Ar buffer (1 ∼ 2 Torr). The spectral
resolution was ∼ 1 cm-1 and 1010 ∼ 1012 photons /pulse were generated. A small portion
of a dye laser output was used to calibrate its frequency based on the optogalvanic
effect in a Fe/Ne hollow cathode lamp. Its precision was ±0.5 cm-1 in the visible region.
3.4 Development of Coherent VUV Source at 104 – 108 nm
Powerful coherent VUV radiation can be obtained by third harmonic generation
and four wave mixing in gaseous nonlinear medium.7,8,18,19 Two tunable outputs from
pulsed dye laser, one in the ultraviolet (ω1) and the other in the visible (ω2), are used in
the latter technique to generate VUV by sum or difference mixing (ωVUV± =2ω1 ±
ω2).7,8 Capability to generate high power VUV radiation which is tunable over a wide
spectral range is the main advantage of this technique. For example, continuously
tunable VUV radiation in the region of 123∼143 nm generated by difference frequency
mixing (FWDM) in Kr and that of 106.6 ~ 125 nm generated by sum frequency mixing
109
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
61S0
71S061D0
λ1=312.78nm λ1=280.28nm
λ2=320~700nmλ2=405~440nm
λVUV=105~128nm λVUV=104~106nm
Fig. 3.3 Schematic diagram for four-wave sum frequency mixing in Hg vapor.
110
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
LiF lens
Achromaticlens
ωUV, ωS
Hg
Heatingblock
Ar/He
Water inWater in
Out Out
ωVUV
Fig. 3.4 Apparatus for VUV generation by four-wave sum frequency mixing in
Hg vapor. The laser beams were aligned off-center of LiF lens to separate the
VUV light from the residual UV and visible lights at the interaction region with
the molecular beam.
111
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
(FWSM) in Hg have been used in this laboratory for MATI study of molecular ions.
VUV photon flux measured inside the MATI apparatus was 109 photons/pulse or larger
over the above spectral ranges. The low wavelength limit of 106.6 nm is set by the
transmission of the LiF window mounted at the rear end of the Hg cell. To obtain VUV
below this limit, several windowless schemes have been devised, utilizing rotating
pinhole, supersonic jet, or glass capillary array.20-22 For a windowless Hg cell to be
useful as a routine VUV source for ZEKE and MATI spectrometries and especially for
dynamics study using these techniques, its output must be around 5×108 photons/pulse
(1nJ) or higher measured at the interaction region of the main apparatus. In addition,
the device must be free from various troubles such as clogging of flow restrictors,
excessive loss of Hg, etc. such that a prolonged operation (several hours or longer) is
possible. Hardly any contamination of the main apparatus and pumps by Hg is another
important requirement. An improved windowless Hg cell which uses a single capillary
as the flow restrictor has been designed and constructed with the above requirements in
mind. Details of the design and its performance in the 104 ~ 108 nm range will be
presented in this note.
3.4.1 Experimental Setup
A schematic diagram of the apparatus is shown in Fig. 3.5. The Hg vapor is
generated inside the heating block (2 cm length along the beam path). 14 cm long
water-cooled arms with baffles inside are attached on the front and rear sides of the
heating block to reduce diffusion of the Hg vapor. A cone-type glass capillary (0.8 and
2 mm I.D. with 70 mm length) is installed along the beam path near the rear end of the
cell. The optical transmission through the capillary was larger than 80 %. Conductance
of the buffer gas (He) through the capillary was estimated to be ∼0.4 l/min. A copper
cold trap cooled by liquid nitrogen is installed between the heating block and the
capillary to minimize contamination of the capillary by Hg and also to minimize
112
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
reabsorption of VUV by Hg.
A differential pumping system has been devised to insure high vacuum in the
monochromator chamber and eventually very high vacuum in the PI chamber. An
intermediate chamber is installed between the Hg cell and the monochromator chamber.
A 40 mm long glass tube with 2.5 mm I.D. connects the intermediate and the
monochromator chambers. The intermediate chamber is evacuated by a mechanical
pump (180 l/min). A copper cold trap is installed in this chamber also to remove
residual Hg and pump oil. The monochromator chamber is evacuated by a
turbomolecular pump (50 l/s). Finally, an aperture (2.5 mm diameter, 3mm length)
separates the monochromator and PI chambers. In a typical operating condition, the Hg
pressure in the cell is ∼1 torr and Ar or He buffer gas with the pressure of several torrs
is introduced to the cell. Then, the pressures in the intermediate, monochromator, and
PI chambers are maintained at ∼3×10-2, ∼5×10-5, and <10-7 torr, respectively. It is to be
mentioned that hardly any trace of Hg is visible in the intermediate chamber after a
prolonged operation.
Two-photon resonant 71S0 - 61S0 or 61D2 - 61S0 transition in Hg has been utilized
for FWSM. The ultraviolet laser (ω1=312.8 or 280.3 nm, respectively, with ~3
mJ/pulse) was generated by frequency-doubling of a dye laser (Continuum ND6000)
output pumped by the second harmonic of an Nd:YAG laser (Continuum PL8000). ω2
at 320 ~ 355 nm (2 ~ 3 mJ/pulse) or 405 ~ 436 nm (3 ~ 6 mJ/pulse) was generated by
another dye laser (Lambda Physik SCANMATE 2E) pumped by the second or third
harmonic of an Nd:YAG laser (Continuum PL8010). Two laser beams were spatially
and temporally overlapped and focused at the center of the Hg cell with an achromatic
lens (f=20 cm).
The home-built monochromator consists of an Al-MgF2 coated concave diffraction
grating (R=0.5 m, 2400 gr/mm, Jobin Yvon) with dispersion of 0.69 nm/mm and an
113
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
Achromaticlens
ωUV, ωS
m=1
PI chamber
m=0
m=-1
Au plate
Monochromator(pumping with TMP)
Concave gratingR=0.5M
Mechanicalpump
Temperature-controlledpulsed valve
VUV+UV
N2(l)
Hg
Heatingblock
N2(l)Buffer
gas
Water in
Capillary
Aperture(dia.=2.5mm)
Cold finger(Cu)
Hg cell with glass cone type capillary
Fig. 3.5 Schematic diagram of the experimental apparatus including the Hg cell with a cone type glass capillary,
monochromator, and photoionization chamber. 114
Copyright(c)2002 by Seoul N
ational University L
ibrary. All rights reserved.(http://library.snu.ac.kr) 2004/01/28 17:41:01
Au plate working as a VUV monitor. VUV beam diffracted at m=-1 was introduced to
the PI chamber while the beam at m=1 was monitored by the Au plate. The grating
must be rotated when VUV wavelength is changed significantly (by changing ω2). This
is not needed when the scan range is narrow (< 500 cm-1).
3.4.2 VUV Generation at 104 – 108 nm
The systematic procedure to generate VUV by FWSM in Hg and introduce it to the
main apparatus is as follows. The two-photon resonance frequency, ω1, for transitions
71S0 - 61S0 and 61D2 - 61S0 of Hg was determined by observing the third harmonic
generation at 104.3 and 93.4 nm, respectively, using the Au plate monitor. A third
harmonic spectrum near 104.3 nm is shown in Fig. 3.6(a). Then, ω2 was overlapped
with ω1 observing VUV signal increase in the Au plate monitor. For VUV generation
via 71S0 - 61S0 transition, the spectral profile measured with the monitor is shown in
Fig. 3.6(b). It is to be mentioned that intensity of the ω1 laser was adjusted to suppress
the third harmonic signal. Finally, VUV was introduced to the main apparatus and
ionized benzene in the supersonic beam. The VUV spectrum recorded by measuring
benzene ion signal is shown in Fig. 3.6(c). This is essentially the same as the VUV
spectrum recorded using the Au plate monitor, Fig. 3.6(b). Similar spectra obtained for
VUV generated via 61D2 - 61S0 transition are shown in Fig. 3.7.
Previously, we estimated the number of VUV photons per pulse exiting a Hg cell
with LiF window using a standard photoionization cell with NO.17 Here, we measured
the benzene photoionization signals using VUV at 108 nm from the Hg cell with LiF
window and from the present windowless cell. Combining the above data, the number
of VUV photons per pulse at 108 nm from the windowless cell was estimated.
Correction for the wavelength dependence of the photoionization efficiency of benzene
was not made because variation was less than 15 % in the 104 ~ 108 nm spectral
region.23 Typical VUV power measured were 4×109 photons/pulse (8 nJ/pulse) at
115
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
104.20 104.22 104.24 104.26 104.28 104.30
(a)In
tens
ity
105 106 107 108
106.24nm
106.68nm(b)
Inte
nsity
105 106 107 108
(c)
Inte
nsity
Wavelength, nm
Fig. 3.6 (a) Spectral profile of VUV generated by frequency tripling in Hg at ω1
~312.8 nm. (b) and (c) show spectral profiles of VUV generated by FWSM via 71S0 -
61S0 transition in Hg recorded using the Au plate monitor and photoionization of
benzene, respectively. PHg ∼ 1 torr and PHe ∼ 2 torr.
116
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
106.6 nm generated via 71S0 - 61S0 with the He buffer (2 torr) and 6×108 photons/pulse
(1nJ/pulse) at 104.3nm generated via 61D2 - 61S0 with the Ar buffer (0.5 torr). These
are the values at the photoionization region of the main apparatus. The actual power
generated by the Hg cell is thought to be larger by an order of magnitude.
In the 105 ~ 108 nm spectrum of VUV generated via 71S0 - 61S0, Fig. 3.6(c), a dip
appears at around 106.24 nm. This may be the same one as observed by Koudoumas
and Efthimiopoulos at 106.28 nm.24 The latter was attributed to the absorption of the
VUV radiation by Hg+ and Hg2 by the above investigators. The same authors reported
enhanced VUV generation near 106.68 nm, as observed in this work also, which was
attributed to transition to the autoionizing 6p' 3D1 state. A second dip observed near
this wavelength which was attributed to reabsorption by Hg combined with low
conversion efficiency at high Hg pressure does not appear in the present result. Low
conversion efficiency was observed near 107 nm due to the positive ∆k, a phase
mismatch near the resonance as pointed out by Koudoumas and Efthimiopoulos.
In the 104 ~ 106 nm spectrum of VUV generated via 61D2 - 61S0, Fig. 3.7(b), a dip
appears at 104.8 nm which is due to absorption by Ar 4s1[1/2]1 - 3p6 1S0 transition.25
This dip disappeared when He was used as the buffer gas. However, the conversion
efficiency got lower by a factor of 3 ~ 5 than that using the Ar buffer.
In summary, continuously tunable coherent VUV radiation in the 104 ~ 108 nm
region has been generated by four wave sum frequency mixing with a windowless Hg
cell using a cone-type glass capillary as the flow restrictor. Several nJ/pulse of VUV
power was measured at the interaction region of the main apparatus, which seems to be
higher by an order of magnitude at least than previously reported. The device could be
run for a prolonged period without any trouble (such as blocking of the capillary by
Hg) or need for maintenance. In addition, near absence of Hg contamination outside
the cell makes this a useful VUV light source for routine spectroscopic work.
117
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
104.0 104.5 105.0 105.5 106.0
104.78nm
(a)
Inte
nsity
104.0 104.5 105.0 105.5 106.0
(b)
Inte
nsity
Wavelength, nm
Fig. 3.7 Spectral profiles of VUV generated by FWSM via 61D2 - 61S0 measured
using (a) the Au plate monitor and (b) photoionization of benzene. PHg ∼ 1 torr
and PAr ∼ 0.5 torr.
118
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
References
1. W. Jamroz and B. P. Stoicheff, in Progress in Optics XX, edited by E. Wolf
(North-Holland Publishing Co., Amsterdam, 1983).
2. S. P. McGlynn, G. L. Findley, and R. H. Huebner, Photophysics and
photochemistry in the vacuum ultraviolet (Kluwer Academic Publishers, 1985).
3. F. J. Wuilleumier, Y. Petroff, and I. Nenner, Vacuum Ultraviolet Radiation
Physics (World Scientific, Singapore, 1992).
4. S. P. McGlynn, G. L. Findley, and R. H. Huebner, Photophysics and
photochemistry in the vacuum ultraviolet (Kluwer Academic Publishers, 1985).
5. U. Becker and D. A. Shirley, VUV and soft X-ray photoionization (Plenum press,
New York, 1996).
6. J. A. R. Samson and D. L. Ederer, Vacuum Ultraviolet Spectroscopy I and II
(Academic Press, San Diego, 1998).
7. R. Hilbig and R. Wallenstein, Appl. Phys. 21, 913 (1982).
8. R. Hilbig and R. Wallenstein, IEEE J. Quantum Electron. QE-19, 1759 (1983).
9. R. Hilbig and R. Wallenstein, IEEE J. Quantum Electron. QE-17, 1566 (1981).
10. D. Cotter, Opt. Commun. 31, 397 (1979).
11. W. Zapka, D.Cotter, and U. Brackman, Opt. Commun. 36, 79 (1981).
12. B. P. Stoicheff, J. R. Banic, P. Herman, W. Jamroz, P. E. Larocque, and R. H.
Lipson, in Proc. Laser Techniques for Extreme Ultraviolet Spectroscopy,
(American Institute of Physics, New York, 1982).
13. S. C. Wallace and G. Zdasiuk, Appl. Phys. Lett. 28, 449 (1976).
14. W. Jamroz, P. E. Larocque, and B. P. Stoicheff, Opt. Lett. 7, 617 (1982).
15. R. Mahon and F. S. Tomkins, IEEE J. Quantum Electron. QE-18, 913 (1982).
119
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
16. R. R. Freeman, R. M. Jopson, and J. Bokor, in Proc. Laser Techniques for
Extreme Ultraviolet Spectroscopy, (American Institute of Physics, New York,
1982).
17. J. A. R. Samson, Techniques of Vacuum Ultraviolet Spectroscopy (Wiley, New
York, 1967).
18. J. C. Miller and R. N. Compton, Phys. Rev. A 25, 2056 (1982).
19. J. P. Marangos, N. Shen, H. Ma, M. H. R. Hutchinson, and J. P. Connerade, J. Opt.
Soc. Am. B 7, 1254 (1990).
20. K. D. Bonin and T. J. Mcllrath, J. Opt. Soc. Am. B 2, 527 (1985).
21. J. Bokor, P. H. Buchsbaum, and P. R. Freeman, Opt. Lett. 8, 217 (1983).
22. P. R. Herman and B. P. Stoicheff, Opt. Lett. 10, 502 (1985).
23. V. H. Dibeler and R. M. Reese, J. Res. Natl. Bur. Std., 68A, 409 (1964).
24. T. Efthimiopoulos and E. Koudoumas, Appl. Phys. B 55, 355 (1992).
25. A. A. Radzig and B. M. Smirnov, Reference Data on Atoms, Molecules, and Ion
(Springer-Verlag, Berlin, Heidelberg, 1985).
120
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
121
Chapter 4
VUV-MATI Spectroscopy of Benzenoid
Molecules
Structure, thermochemical properties, and dynamics of ions are of
fundamental interest in relation to studies of combustion, atmospheric chemistry,
cosmochemistry, etc.1-3 Information on ionic vibrational structures is especially
useful to probe ions in complex mixtures or follow complicated reaction processes.
Nowadays, conventional spectroscopic techniques such as high-resolution
photoelectron or laser-induced fluorescence spectroscopies have been popular in
characterizing polyatomic ions.4,5 However, obtaining vibrational spectra of
polyatomic ions with these techniques is a formidable job because of their limited
capabilities. The resolution of photoelectron spectroscopy (PES), which is
typically 10 meV (80 cm-1), is not good enough to obtain vibrational information
on polyatomic cations, even though PES is useful to investigate electronic states.6
The laser-induced fluorescence spectroscopy, which usually has higher resolution
than PES and can resolve vibrational peaks, is not generally applicable because
most of the excited electronic states of polyatomic cations do not fluoresce.
Zero kinetic energy (ZEKE) photoelectron spectroscopy has a much better
resolution than ordinary PES and hence can obtain even rotational information for
simple molecular ions.7-9 Mass-analyzed threshold ionization (MATI) basically
employs the same principle as ZEKE except for detecting ions instead of electrons
and hence providing mass-selectivity in the spectra.10-11 Generally adopted in the
ZEKE or MATI spectroscopies is a two-color 1+1′ scheme. Namely, excitation to
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
122
a Rydberg state is achieved in two steps via an intermediate state. However, since
the first excited states of most of the molecules are located in the region beyond
commercial dye laser outputs ( > 210 nm) and these states in many cases are either
unbound or relax rapidly, use of this scheme suffers the major limitation of low
transition probability to the Rydberg state. One-photon ZEKE/MATI using tunable
vacuum ultraviolet (VUV) radiation can overcome such difficulties because the
transition occurs directly from the ground state to a Rydberg state, not mediated
by an excited electronic state of the neutral.12-20
One-photon ZEKE scheme has been utilized already to obtain vibrational
spectra of simple cations in the ground and excited states.12-15 The present work
will demonstrate that one-photon MATI scheme can be routinely used to obtain
vibrational spectra of polyatomic cations also once coherent VUV radiation
becomes available over a wide spectral range.
4.1 Determination of Ionization Energies
MATI or ZEKE spectroscopies utilize high Rydberg states of neutrals
conversing to the states of corresponding ions. The energies nlmE of a series of
Rydberg states are described by the Rydberg formula.
2)( l
Mnlm n
REδ−
−= (4.1)
Here n, l, and m are the usual quantum numbers for H-like atoms, RM is the mass-
dependent Rydberg constant (R∞ = 109737 cm-1), and δl is the l-dependent
quantum defect. A Rydberg state with the n quantum number of 200 lies ~3 cm-1
below the corresponding series (ionization) limit. Due to using the pulsed field
ionization, the ionization energy measured by these spectroscopies always gives
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
123
the value just below the ionization threshold. Hence, the position of MATI peak
depends on the field strength employed as follows:
FE α=∆ (4.2)
where F is electric field (in V/cm) and theory predicts the adiabatic value of α
to be 6 cm-1 (V/cm)-½. Lowering of the ionization potential due to an electric field
is shown in Fig. 4.1. When molecule is in field-free region, Rydberg electron feels
attraction by ion core and it can be regarded as Coulomb attraction between
positive and negative charge of e’s. When external electric field is applied, it acts
as a perturbation and will distort Coulomb potential, lowering the ionization
energy by E∆ . The perturbed potential energy of electron is
eFzr
eE −πε
−=4
2
(4.3)
where r is distance from the ion core and z is its component along applied
electric field direction. Differentiation of this with respect to z gives the
location of ionization barrier.
eFz
e−
πε= 2
2
40 (4.4)
or,
Fezπε
=4
(4.5)
Then, the lowering of ionization energy from unperturbed value, 0, is given by,
FeF
eeFe
FeE ⋅πε
−=πε
−πε
πε−=∆
42
44
4
32
(4.6)
The coefficient calculated with physical constants are 6.12 cm-1 (V/cm)-½.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
���
Fig. 4.1 Lowering of the ionization potential due to an electric field.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
125
4.2 Selection Rules in One-photon MATI Spectra
Peaks with widely different intensities appeared in one-photon MATI spectra,
which were assigned to the fundamentals, overtones, or combinations. Vibrational
selection rules for one-photon transitions from the neutral ground state to Rydberg
states were helpful for spectral assignment.
• D6h Symmetry Group19,20
The transition dipole moment between the ground and Rydberg states of
neutral is as follows.
Rev
GevGR ΨµΨ=µrr (4.7)
Here GevΨ and R
evΨ are the vibronic wavefunctions of the ground and Rydberg
states, respectively, and µr is the dipole moment operator. RevΨ can be further
separated into the vibronic wavefunction of the ion core, CevΨ , and the
wavefunction of the Rydberg electron, ReΨ . Namely,
Re
Cev
GevGR ΨΨµΨ=µ
rr (4.8)
Electronic and vibrational parts in the ground state and ion core may be further
separated with the Born-Oppenheimer approximation.
Cv
Gv
Re
Ce
GeGR ΨΨΨΨµΨ=µ
rr (4.9)
In the production of the benzene cation in the ground state from the ground state
neutral, GeΨ and C
eΨ belong to a1g and e1g symmetries, respectively. µr in D6h
symmetry has a2u ⊕ e1u species. Rydberg orbitals can be classified in terms of
spherical symmetry, which correlate with the symmetry species in D6h as follows.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
126
s orbital: a1g
p orbitals: a2u (pz); e1u (px,y)
d orbitals: a1g ( 2zd ); e1g (dxz,yz); e2g (dxy, 22 yxd − )
f orbitals: a2u ( )35( 22 rzzf − ); e1u ( )5( 22 rzxf − , )5( 22 rzyf − ); e2u ( )( 22 yxzf − , xyzf );
b1u ⊕ b2u ( )3( 22 yxxf − , )3( 22 yxyf − )
In the ground state benzene, e1g is the highest occupied molecular orbital, which
correlates with a d orbital in spherical symmetry. Then, an electron in e1g can be
promoted to one of the p or f Rydberg orbitals according to the selection rule for
H-like atom (∆l = ± 1). In D6h symmetry, transitions to some of these are electric
dipole-allowed while the others are forbidden. For example, transition to px,y is
allowed because a1g ( GeΨ ) ⊗ a2u (z) ⊗ e1g ( C
eΨ ) ⊗ e1u (px,y) = a1g ⊕ a2g ⊕ e2g.
Since benzene is prepared under the beam condition, it is initially in the zero-
point level, or GvΨ belongs to a1g. Hence, in an electric dipole-allowed transition,
CvΨ must also belong to a1g. Namely, fundamentals and all the overtones of the a1g
modes are allowed while ∆υ = 2, 4, 6, ··· selection rule holds for nontotally
symmetric modes, as have been well established in electronic spectroscopy. Then,
intensity of each vibrational band would be determined by the Franck-Condon
factor.
Fundamentals of nontotally symmetric vibrations which are dipole-forbidden
also appear in MATI spectra, even though weakly. In these cases, vibronic
mechanism must be invoked. For example, let us consider excitation of a b1g
mode by one quantum. Then, GevΨ and C
evΨ belong to a1g and e2g, respectively.
Selecting px,y (e1u) Rydberg orbital and µx,y (e1u), symmetry of the transition
moment becomes a1g ⊗ e1u ⊗ e2g ⊗ e1u = a1g ⊕ a2g ⊕ 3e2g, making the fundamental
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
127
transition of a b1g mode vibronically allowed. It can be shown that all the
fundamentals of nontotally symmetric gerade (g) vibrations of benzene are
vibronically allowed, even though dipole-forbidden, in one-photon excitation to a
Rydberg state and can appear weakly in a MATI spectrum.
If the benzene cation has the center of symmetry, as in D6h or D2h, one would
not expect to observe fundamentals of ungerade (u) modes. However, observation
of the v20 (e2u) fundamental was reported in the previous two-photon ZEKE and
MATI studies. One may invoke magnetic dipole or electric quadrupole
mechanism to account for the appearance of ungerade fundamentals. Such
mechanisms are highly unlike because a g,u-forbidden vibronic transition would
be extremely weak and hence essentially undetectable by ZEKE or MATI. Hence,
appearance of ungerade fundamentals in the two-photon ZEKE or MATI spectra
was attributed to the fact that the intermediate state used in this scheme, A~ 1B2u,
is slightly nonplanar. Such an intermediate state is not involved in the present one-
photon scheme. However, ungerade fundamentals appear, even though very
weakly, in the one-photon spectra as will be shown later. One plausible
mechanism for the very weak appearance of ungerade fundamentals is the l
mixing in the Rydberg states caused by the stray field inside the instrument or by
scrambling field applied.
A guideline for vibrational assignment gained from consideration of the
selection rules can be summarized as follows. Fundamentals and overtones of the
a1g modes may appear prominently in the one-photon MATI spectra of benzene
while nontotally symmetric gerade fundamentals may appear weakly. Ungerade
fundamentals may appear also, even though very weakly.
• C2v Symmetry Group17,18
Rydberg orbitals can be classified in terms of spherical symmetry, which
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
128
correlate with the symmetry species in C2v as follows.
s orbital: a1
p orbitals: a1(pz); b1(px); b2(py)
d orbitals: a1( 2zd ); a2(dxy); b1(dxz); b2(dyz); a1( 22 yxd − )
f orbitals: a1( )35( 22 rzzf − ); b1( )5( 22 rzxf − ); b2( )5( 22 rzyf − ); a1( )( 22 yxzf − );
a2( xyzf ); b1( )3( 22 yxxf − ); b2( )3( 22 yxyf − )
Then, the symmetry selection rules for the electric dipole allowed transition
become
1a)()()()( =ΨΓ⊗ΨΓ⊗µΓ⊗ΨΓ Re
Ce
Ge
r . (4.10)
GeΨ belongs to a1 symmetry species for the both phenylacetylene and
benzonitrile. CeΨ is essentially the same as the electronic wavefunction of the
ground state cation and has b1 symmetry in both cases. µr has symmetry species
a1, b1, and b2 in C2v. It can be shown that Rydberg ← ground transition in the
neutral is allowed for various selections of the Rydberg orbital. Most of the
molecules prepared under the supersonic jet condition are in the zero-point
vibrational level which is totally symmetric, a1. Hence, in an electric dipole-
allowed transition, CvΨ should also belong to a1. Then, fundamentals and all the
overtones of a1 modes are allowed while ∆υ = 2, 4, 6, … selection rule holds for
nontotally symmetric modes. Under the Born-Oppenheimer approximation, the
relative intensity of each vibrational peak is determined by the Franck-Condon
factor.
4.3 Franck-Condon Factor Calculations
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
129
Assuming that the properties of the ion core in a high Rydberg state be well
approximated by those of the cation, Franck-Condon factors for the Rydberg ←
ground transition were calculated with the above quantum chemical data using the
method of Sharp and Rosenstock.21 A brief account of the method is as follows.
The vibrational wavefunction under harmonic approximation can be written as
ψv(Q) = Nv exp( –21 Q† Γ Q) Hv( Γ1/2 Q ) (4.11)
Here v is the vibrational state vector designating the quantum number for each
mode, Q is the normal coordinate vector, Γ is the diagonal matrix of reduced
frequencies 4π2νi/h, H is the Hermite function, and Nv is the normalization
constant. In the calculation of the Franck-Condon factor for the vC ← vG
component in the Rydberg ( C ) ← ground ( G ) transition.
2CGC ]d)()([q GCGC QQQ vvvv ΨΨ= ∫ (4.12)
Here QC and QG are the normal coordinates in the Rydberg and ground states,
respectively, which can be related by
QG= JQC + K (4.13)
With the common coordinates established, integration in eqn. (4.6) can be carried
out. Analytical expression for the Franck-Condon integral in the harmonic limit
available in the literature,22 which is given as a function of J, K, ΓC, and ΓG. In
addition, the intensities dictate about QC vs. QG, the displacement.
4.4 VUV-MATI Spectroscopy of Monohalobenzenes
MATI spectra of C6H5Cl+• and C6H5Br+•, C6H5I+•, and C6H5F+• in the ground
electronic states and those of C6H5Cl+• and C6H5Br+• in the B~ 2B2 excited states
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
���
Fig. 4.2 Illustration of the Franck-Condon principle and intensity distributions
for small and large displacement, respectively.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
131
have been obtained via one-photon excitation. The one-photon MATI scheme has
been found to be especially useful to obtain the cation ground and excited
electronic state spectra because knowledge on the neutral intermediate states is
not required.
4.4.1 Vibrational Spectra in the Ground Electronic States, X~
4.4.1.1 Chlorobenzene Cation
Excitation of chlorobenzene to Rydberg states below the first ionization
threshold followed by PFI produces two isotopomers of the molecular ion,
C6H535Cl+• and C6H5
37Cl+•, in the mass spectrum. By monitoring one of these
isotopomers and scanning the VUV wavelength, MATI spectrum for each
isotopomer was recorded, Figs. 4.3(a) and 4.3(b). Checking the direct ion yield vs.
photon energy, or the photoionization efficiency curve (not shown), the first
intense peaks appearing at ∼73180 cm-1 in these figures were assigned to the 0-0
band. The energy of a 0-0 band appearing in a MATI spectrum is usually smaller
than the true ionization energy. This is because molecules in ZEKE states a few
cm-1 below the threshold can also be ionized due to the high PFI field used. Here,
the position of the 0-0 band was measured using various values of PFI field
without the spoil field and the accurate ionization energy was estimated by
extrapolation to the zero field limit. The ionization energy of C6H535Cl thus
obtained was 73177±5 cm-1, in excellent agreement with 73173±5 cm-1 reported
by Lembach and Brutschy.23 The ionization energy of C6H537Cl obtained similarly
was hardly different from the above. The ionization energies to the ground and
first excited electronic states of halobenzene cations measured in this work are
listed in Table 4.1.
The vibrational frequency of each band was estimated simply as the difference
of its position from that of the 0-0 band. Vibrational frequencies calculated from
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
132
Table 4.1 Ionization energies (IE) to the ground ( X~ 2B1) and B~ 2B2 excited states
of chloro-, bromo-, iodo-, and fluorobenzene cations, in eV.
IE ( X~ 2B1) IE ( B~ 2B2) Ref.
Chlorobenzene 9.0728±0.0006 11.3327±0.0006 This work
9.0723±0.0006 23
9.0720±0.0006 25
9.066±0.008 11.330±0.008 27
Bromobenzene 8.9976±0.0006 10.6406±0.0006 This work
8.991±0.008 10.633±0.008 28
8.98±0.02 26
Iodobenzene 8.7580±0.0006 This work
8.754±0.008 29
8.77±0.02 30
Fluorobenzene 9.2033±0.0006 This work
9.2033±0.0006 31
9.2044±0.0005 32
9.18±0.02 26
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
133
73000 74000 75000 76000
6a26b1
8a1121
16b1
41
7a2
6a17a1
6a1121
7a1121
8a1
8b1
6a16b1
6b1
10b1
111 18b1
1118a1
7a1
9a1
19a1
121
6an(a)Io
n Si
gnal
Photon Energy, cm-1
73000 74000 75000 76000
8a1121
6a26b16a16b1
41 7a212
12
6an
0-0
9a111
7a1121
6a17a1
6a11217a1
19a1 8b1
8a118a1
10b1
1216b116b1
111
0-0
(b)
Ion
Sign
al
Photon Energy, cm-1
Fig. 4.3 The ground state one-photon MATI spectra recorded by monitoring
(a) C6H535Cl+• and (b) C6H5
37Cl+•.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
134
Table 4.2 Vibrational frequencies (in cm-1) and their assignments for the ground
state ( X~ 2B1) chlorobenzene cation.
This work Modea (Wilson)
SymmetryNeutrala PESb MPI-PESc MATId ZEKEe C6H5
35Cl+• C6H537Cl+•
1 a1 1003 950 971 975 974 972 4 b1 685 600(?) 600(?) 6a a1 417 427 422 420 422 419 415 6b b2 615 510 526 531 527 530 7a a1 1093 1121 1100 1115 1116 1118 1114 8a a1 1586 1554 1554 8b b2 1598 1593 1592 9a a1 1153 1180 1194 1200 1193 1193
10b b1 741 771 771 11 b1 197 141 139 12 a1 706 720 714 716 713 710 16b b1 467 393 394 482 482 18a a1 1026 960 992 995 991 991 18b b2 287 311 286 19a a1 1482 1429 1411f 1408f
6a2 838 829 6a3 1260 1246 6a4 1677 1661 6a5 2097 2078 7a2 2235 2225
6a16b1 950 950 6a1121 1135 1131 6a26b1 1368 1360 6a111 1394 1392
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
135
6a17a1 1533 1527 7a1121 1828 1821 8a1121 2280 2277
a Vibrational assignments in Wilson notation and frequencies for the ground state
neutral taken from ref. 26.
b ref. 27.
c ref. 26.
d ref. 23.
e ref. 25.
f These peaks may be assigned alternatively to 6a118a1.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
136
the MATI peaks in Figs. 4.3(a) and 4.3(b) are summarized and compared with
previous results in Table 4.2. Also listed in the table are the vibrational
frequencies and their assignments (Wilson notation and symmetry)24 of the
ground state C6H5Cl neutral reported by various investigators as summarized by
Walter and coworkers. A more recent report by Wright and coworkers25
interchanged the assignments for the a1-type normal modes 2 and 20a and also for
the b2-type normal modes 3 and 14 of the C6H5Cl neutral. We are not in a position
to judge which of the two assignments are correct. Here we adopt the assignments
by Walter and coworkers26 because the assignments for the C6H5Br and C6H5F
neutrals are also included in that work.
The MATI spectra in Fig. 4.3 recorded by ionization to the ground electronic
state of C6H5Cl+•, or the ground state MATI spectra, are in excellent agreement
with the previous spectra reported by Lembach and Brutschy, even though the
present ones are of higher quality and extend over wider spectral range. The most
striking feature in these MATI spectra is appearance of the prominent 6an
progression. Appearance of these peaks at 419, 838, 1260, 1677, and 2097 cm-1
for C6H535Cl+• (Fig. 4.3(a)) and at 415, 829, 1246, 1661, and 2078 cm-1 for
C6H537Cl+• (Fig. 4.3(b)) clearly shows the presence of the isotope shift for this
substituent–sensitive mode. Wright and coworkers carried out ab initio
calculations for the ground state neutral and cation at the Hartree-Fock (HF) level
with the 6-31G** basis set. It was found that the cation geometry was distorted
from that of the neutral mostly along the direction of the 6a normal mode vector,
in agreement with the appearance of 6an progressions. We also observed the 7a1
and 7a2 progressions in the MATI spectra, 1118 and 2235 cm-1 for C6H535Cl+• and
1114 and 2225cm-1 for C6H537Cl+•. Appearance of these progressions can be
explained as above because the 6a and 7a modes differ only in the phase of atomic
motion.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
137
When transitions start from the ground vibrational state of the neutral, it is
known that the totally symmetric vibrational states of the cation, or the states with
a1 symmetry in the present case, produce prominent ZEKE or MATI peaks.33 In
agreement with this propensity rule, prominent peaks are observed for the a1-type
modes 1, 8a, 9a, 12, 18a, and 19a at 974, 1554, 1193, 713, 991, and 1411 cm-1,
respectively, in addition to the 6an and 7an progressions in the MATI spectrum of
C6H535Cl+•. The b2-type modes such as 6b, 8b, and 18b appear distinctly at 527,
1593, and 286 cm-1, respectively. The b1-type modes also appear, even though less
distinctly, such as 10b and 11 at 771 and 141 cm-1, respectively. In contrast, the
a2-type modes are hardly observable. Various combination bands appear,
especially in the 1600∼2800 cm-1 region. Some of the combination bands, namely
those at 950, 1135, 1533, and 1828 cm-1 have been assigned to the 6a16b1, 6a1121,
6a17a1, and 7a1121 states. For major peaks assigned previously, there is no
difference between the present and previous assignments.
4.4.1.2 Bromobenzene Cation
Vibrational spectra of bromobenzene cation in the X~ 2B1 state were obtained
previously by photoelectron spectroscopy (PES)28 and by MPI-PES.29 Unlike
C6H5Cl+•, no high resolution spectrum utilizing ZEKE or MATI technique has
been reported so far. Ionization of C6H5Br near the threshold to the ground
electronic state of the cation generates two isotopomers, C6H579Br+• and
C6H581Br+•, in the mass spectrum. The ground MATI spectra recorded by
scanning the VUV wavelength and monitoring these ions are shown in Figs.
4.4(a) and 4.4(b). As before, the first peaks appearing at ∼72564 cm-1 in the
spectra were identified as the 0-0 bands based on the photoionization efficiency
curve. The first ionization energy determined by extrapolation was 72570±5 cm-1
(8.9976±0.0006 eV), which compares well with the previous PES results of
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
138
8.98±0.0229 and 8.991±0.008 eV.28
Vibrational frequencies of the C6H5Br+• in the X~ 2B1 state estimated as the
difference of each peak position from that of the 0-0 peak are listed in Table 4.3
together with other relevant information. The ground state vibrational spectrum of
C6H5Br+• is similar to that of C6H5Cl+• and can be assigned by referring to the
assignments for the latter. In particular, prominent 6an progression is observed at
331, 659, 987, 1322, and 1653 cm-1. Also, both the fundamental and first overtone
of the 7a mode appear prominently at 1073 and 2142 cm-1. In addition, the a1-type
modes 8a, 12, and 18a appear distinctly at 1577, 678, and 1008 cm-1, respectively.
The peak at 3083 cm-1 can assigned either to the mode 2 or to 20a because both
are the a1-type. Also, distinct are the b2-type modes 6b, 8b, and 18b at 593, 1523,
and 257 cm-1, respectively, while the b1-type modes, 10b at 791 cm-1 and 11 at
126 cm-1, are less distinct. The 6a mode participates heavily to generate various
combination peaks. Readily recognizable is the 6an7a1 progression at 1402, 1734,
and 2061 cm-1. Also distinct are the combination peaks with the 7a participation.
Except for the weak peak at 3083 cm-1, assignments of the vibrational peaks in the
ground state spectrum are rather straightforward. Finally, it is to be mentioned that
the isotope shift in the vibrational spectrum of C6H5Br+• is less than that of
C6H5Cl+•, as expected.
4.4.1.3 Iodobenzene Cation
The vibrational spectrum of iodobenzene cation has not been reported except
for the low resolution PES spectrum where the fundamentals, overtones, and
combination bands of the 6a and 7a modes were identified.29 The ground state
MATI spectrum of C6H5I+• obtained in this work is shown in Fig. 4.5. The first
intense peak at 70633 cm-1 in this spectrum has been identified as the 0-0 band as
before. The accurate first ionization energy determined by extrapolation of its
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
139
73000 74000 75000
19a1
10b1
111 21(?)
7a1121
8a1
6a17a28b1 6a18a1
6a28a1
6an7a1
6an
141
9a1
18a17a1
121
6b1
16a118b1
0-0
(b)
Ion
Sign
al
Photon Energy, cm-1
73000 74000 75000
19a1
10b121(?)6a17a2
7a1121
6an
6a28a1
6a18a1
6an7a1
141
18a1
8a1
8b1
0-0
11118b1
16a16b1
121
9a1
7a1
(a)Io
n Si
gnal
Photon Energy, cm-1
Fig. 4.4 The ground state one-photon MATI spectra recorded by monitoring
(a) C6H579Br+• and (b) C6H5
81Br+•.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
140
Table 4.3 Vibrational frequencies (in cm-1) and their assignments for the
ground state ( X~ 2B1) bromobenzene cation.
This work Modea
(Wilson) Symmetry Neutrala PESb MPI-PESc
C6H579Br+• C6H5
81Br+• 1 a1 1001 950 2 a1 3065 3083(?) 3083(?) 6a a1 314 331 320 331 329 6b b2 614 540 593 593 7a a1 1070 1100 1073 1073 8a a1 1578 1530 1577 1577 8b b2 1523 1523 9a a1 1176 1193 1193 9b b2 1158 1180
10b b1 736 791 791 11 b1 181 126 126 12 a1 671 720 678 678 14 b2 1321 1307 1307 16a a2 409 396 394 18a a1 1020 1016 980 1008 1008 18b b2 257 257 19a a1 1472 1466 1466 20a a1 3067 3083(?) 3083(?) 6a2 659 659 6a3 987 986 6a4 1322 1320 6a5 1653 1649
6a16b1 928 6a7a 1402 1399 6a27a 1734 1729 7a12 1754 1750 6a8a 1911 1907 6a37a 2061 2058 6a28a 2241 2239 6a7a2 2474 2471
a Vibrational assignments in Wilson notation and frequencies for the ground state neutral taken from ref. 26. b ref. 28. c ref. 26.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
141
position was 70638±5 cm-1 (8.7580±0.0006 eV). This compares well with the
previous PES results of 8.754±0.00829 and 8.77±0.02 eV.30
The vibrational frequencies obtained from the ground state MATI spectrum
are summarized in Table 4.4 together with relevant information. Vibrational
frequencies of the iodobenzene neutral in the ground electronic state have been
taken from ref. 34. The vibrational assignments in this reference has been changed
by referring to the differences in assignments for C6H5Cl and C6H5Br between
Varsanyi34 and Walter and coworker.26 The general feature of the ground state
MATI spectrum of C6H5I+• is quite similar to those of C6H5Cl+• and C6H5Br+• and
vibrational assignments can be made by referring to those for the latters. Here
again, the 6an progression appears prominently at 284, 567, 848, and 1129 cm-1.
7a1 appears also at 1036 cm-1, but without further progression. Other a1-type
modes appearing distinctly are 8a, 12, and 18a modes at 1575, 661, and 1015 cm-1,
respectively. The b2-type modes 8b and 18b also appear distinctly at 1517 and 242
cm-1, respectively, while the b1-type modes 11 and 16b appear weakly at 127 and
406 cm-1, respectively. More prominent than the non-a1-type modes are the
combination bands involving the 6a mode such as 6a1121, 6a111, 6a118a1, 6a17a1,
and 6a211 at 943, 1269, 1296, 1310, and 1548 cm-1, respectively.
4.4.1.4 Fluorobenzene Cation
The first ionization energy of C6H5F was measured with two-color (1+1′)
ZEKE by Shinohara et al.32 and with two-color (1+1′) MATI by Lembach and
Brutschy.31 The latter work also reported the vibrational frequencies in the X~ 2B1
ground state of C6H5F+• and their assignments. The ground state MATI spectrum
via one-photon VUV absorption obtained in this work is shown in Fig. 4.6. The
dominant peak at 74221 cm-1 in this spectrum was identified as the 0-0 band as
before. The first ionization energy determined from its position was 74229±5 cm-1
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
142
71000 72000 73000
8a1
116b1
16b1
16a1
111
18b1
16a1 16
b1
121 18a1 6a2116a1121
7a1
6a118a1
6a17a111121
6an
0-0
Ion
Sign
al
Photon Energy, cm-1
Fig. 4.5 The ground state one-photon MATI spectrum recorded by
monitoring C6H5I+•.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
143
Table 4.4 Vibrational frequencies (in cm-1) and their assignments for the ground
state ( X~ 2B1) iodobenzene cation.
Modea
(Wilson) Symmetry Neutrala PESb This work
1 a1 998 990 6a a1 268 282 284 6b b2 612 538 7a a1 1063 1036 8a a1 1575 1575 8b b2 1517 10b b1 729 808 11 b1 167 127 12 a1 654 661 16a a2 398 357 16b b1 421 406 17b b1 903 903 18a a1 1015 1016 1015 18b b2 220 242 6a2 567 6a3 848 6a4 1129
6a1121 943 18b111 1226 6a111 1269
6a118a1 1296 6a17a1 1310 6a211 1548 6a27a1 1594 11121 1648
12118a1 1676 7a1121 1695 6a1121 2256
a Vibrational assignments in Wilson notation and frequencies for the ground
state neutral taken from ref. 34. b ref. 29.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
144
(9.2033±0.0006 eV). This is in excellent agreement with the two-photon MATI
result of Lembach and Brutschy, 74229±5 cm-1, but deviates somewhat from the
ZEKE result, 74238±4 cm-1.
The vibrational frequencies measured from the spectrum are listed in Table 4.5
together with relevant information. The present results are in excellent agreement
with those reported by Lembach and Brutschy and correlate well with the
vibrational frequencies of the ground state neutral. A noticeable difference from
the two-photon MATI spectrum, and also from the ground state one-photon MATI
spectra of C6H5Cl+• and C6H5Br+• presented above, is the complete absence of 6a
overtones. Other than this, the general spectral feature in the ground state MATI
spectrum of C6H5F+• is similar to those of C6H5Cl+• and C6H5Br+•. The a1-type
modes 6a, 7a, 8a, 9a, and 19a appear prominently at 500, 1274, 1610, 1168, and
1502 cm-1, respectively. The b2-type modes 6b, 9b, 14, and 19b appear distinctly
at 606, 1106, 1339, and 1464 cm-1, respectively, while the b1-type modes 10b and
11 appear weakly at 763 and 182 cm-1, respectively.
4.4.2 Vibrational Spectra in the Excited Electronic State, B~
4.4.2.1 Chlorobenzene Cation
According to the previous high resolution PES study,27 the ionization energy
to the B~ 2B2 state of C6H5Cl+• is 11.330±0.008 eV, or ∼109.4 nm in VUV
wavelength. The VUV-MATI spectrum of C6H535Cl+• obtained in the 107∼109.8
nm spectral region is shown in Fig. 4.7. The dominant first peak in the spectrum
can be identified as the origin. The ionization energy to the B~ state determined
from its peak position is 91404±5 cm-1 (11.3327±0.0006 eV), in excellent
agreement with the above PES result.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
145
74000 75000 76000 77000
9a2
9b1
6a18a1
9a1121
6a1141
6a19a1
8a1
19a1
19b1
7a11419a1
151
10b1
6b1
16b1
18b1
6a1
121111
0-0
Ion
Sign
al
Photon Energy, cm-1
Fig. 4.6 The ground state one-photon MATI spectrum recorded by
monitoring C6H5F+•.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
146
Table 4.5 Vibrational frequencies (in cm-1) and their assignments for the
ground state ( X~ 2B1) fluorobenzene cation.
Modea
(Wilson) Symmetry Neutrala MPI-PESb MATIc This work
3 b2 1301 1299 6a a1 517 500 500 500 6b b2 615 510 505 606 7a a1 1232 1274 8a a1 1604 1620 1610 8b b2 1597 1574 9a a1 1156 1170 1164 1168 9b b2 1128 1106 10b b1 754 763 11 b1 249 181 182 12 a1 809 810 795 804 14 b2 1326 1339 15 b2 1066 1071 16b b1 498 479 18b b2 400 410 400 402 19a a1 1500 1502 19b b2 1460 1464
6a19a1 1668 6a131 1797 6a1141 1842 6a18a1 2109 9a1121 1968 9a19b1 2282
9a2 2343
a Vibrational assignments in Wilson notation and frequencies for the ground
state neutral taken from ref. 26. b ref. 26. c ref. 31.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
147
Unlike in the ground state MATI spectrum, the 0-0 band almost totally
dominates the excited state MATI spectrum, indicating that the equilibrium
geometries of the ground state neutral and the B~ state cation are very similar.
This is in agreement with the previous identification of this state as formed by
removal of an electron from the chlorine nonbonding p orbital parallel to the
benzene ring n(Cl3p‖), or a state with n(Cl3p‖) character.27 In contrast, Anand
and coworkers35 suggested that the B~ state probed in their PIRI study had the σ
character. In the PIRI experiment, the C6H5Cl neutral prepared in a Rydberg state
correlating with the cation ground state by two-photon absorption was further
excited and autoionization or fragmentation in the B~ or higher electronic states
was observed. Similarly, multi-photon absorption mediated by the B~ 2B2←X~ 2B1
resonance was assumed in the REMPD study of the B~ state of C6H5Cl+• by
Ripoche and coworkers.36 One of the problems related to the multi-photon
excitation of the ground state cation (REMPD) or the ion core correlating with the
ionic ground state (PIRI) is that the B~ 2B2←X~ 2B1 transition is electric dipole
forbidden and hence there is no guarantee that the peaks observed reflect the
B~ 2B2←X~ 2B1 resonance. Complication may also arise from the possible presence
of an excited state(s) formed by promotion of an electron in an occupied orbital to
an unoccupied orbital, which is not a hole state observed in PES. Anand and
coworkers argued based on their experimental results that the excited state
observed in their PIRI experiment was the electric dipole allowed 2B1 rather than 2B2. Namely, one can not rule out the possibility that the states probed by PIRI or
REMPD may have been different from the B~ 2B2 state appearing in the PES
spectrum. In contrast, the present one-photon MATI technique probes hole states
only and does not suffer from the complication related to multi-photon absorption.
Hence, it is highly likely that the vibrational structure observed in the 107~109.8
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
148
nm region in this work is that of the B~ 2B2 state with the n(Cl3p‖) character
observed in PES.
Vibrational frequencies measured from the excited state MATI spectrum in
Fig. 4.7 are listed in Table 4.6 together with relevant information. Vibrational
frequencies reported in the REMPD and PIRI works are listed also for comparison.
Considering the n(Cl3p‖) character of the ionic state accessed, one expects good
correlation of ionic vibrational frequencies with those of the ground state neutral.
It is to be noted that the vibrational frequencies measured in this work display
better correlation with the neutral data than the REMPD or PIRI results do.
Two broad bands appear in the photoelectron spectrum at 42 and 120 meV (340
and 970 cm-1, respectively) above the 0-0 band of the B~ 2B2 state, which were
assigned to the normal modes 6a and 1, respectively. In the present MATI
spectrum, major peaks other than the 0-0 peak appear at 329, 382, and 961 cm-1. It
is straightforward to assign the peak at 961 cm-1 to the normal mode 1. We will
assign the peak at 382 cm-1 to the normal mode 6a. In addition to better
correlation with the corresponding frequency in the neutral, such an assignment
can explain the appearance of some combination bands and an overtone (Table
4.6). It is not easy to assign the 329 cm-1 peak. We will assign it to the 16a mode
simply based on its proximity to the corresponding frequency in the neutral. Other
peaks display decent correlation with those in the neutral and can be easily
assigned. Here again, majority of the peaks are due to the a1-type modes such as
6a, 12, 1, 18a, 7a, and 9a at 382, 725, 961, 1009, 1080, and 1173 cm-1. Some b2-
type modes (6b, 3, and 14) appear also distinctly at 546, 1279, and 1338 cm-1
while b1-type modes (4, 10b, and 17b) are slightly less distinct. The prominent
16a mode at 329 cm-1 is the only a2-type mode, which casts some doubt on the
validity of its assignment. An alternative assignment may be 18b which is a b2-
type mode.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
149
500 1000 1500
18b1
18a1 314112110b1 17b19a17a1
11
6b1
6a1
16a1
Ion Energy, cm-1
91000 91500 92000 92500 93000 93500
0-0
Ion
Sign
al
Photon Energy, cm-1
Fig. 4.7 The B~ 2B2 state one-photon MATI spectrum of C6H535Cl+•. The x-
axis of the inset, ion energy, denotes energy scale referred to the position of
the 0-0 band.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
150
Table 4.6 Vibrational frequencies (in cm-1) and their assignments for the
chlorobenzene cation in the B~ 2B2 excited state.
Modea
(Wilson) Symmetry Neutrala PESb REMPDSc PIRId This work
1 a1 1003 970 869 1010 961 3 b2 1271 1279 4 b1 682 667 5 b1 985 730 6a a1 420 340 387 384 382 6b b2 616 562 546 7a a1 1085 943 1131 1080 9a a1 1174 1263 1173 10a a2 830 761 10b b1 740 759e
11 b1 196 153 12 a1 701 636 725 16a a2 400 313 223 329 16b b1 467 218 439 17b b1 902 899 18a a1 1026 866 1009 18b b2 297 260 329 246
6a116a1 709 6b116a1 870
a Vibrational assignments in Wilson notation and frequencies for the ground
state neutral taken from ref. 26. b ref. 27. c ref. 36. d ref. 35. e This peak may be assigned alternatively to 6a2.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
151
4.4.2.2 Bromobenzene Cation
Vibrational spectrum of bromobenzene cation in the B~ 2B2 state has been
obtained for the first time in the present MATI work (Fig. 4.8). The strongest peak
in the spectrum is the 0-0 band as in the B~ 2B2 MATI spectrum of C6H5Cl+•. Its
unambiguous assignment also allows accurate determination of the ionization
energy to this state, 85822±5 cm-1 (10.6406±0.0006 eV) as determined through
extrapolation. This is in good agreement with 10.633±0.008 eV determined from
high resolution photoelectron spectrum.28
Vibrational frequencies obtained from the MATI spectrum are listed in Table 4.7
together with relevant information. Vibrational frequencies of C6H5Br+• in the
B~ 2B2 state display good correlation with those of the neutral as expected. By
comparing the frequencies of the substituent-sensitive modes of C6H5Cl+• and
C6H5Br+•, one finds that mass dependences of these frequencies are similar to
those observed for the neutrals. Here again, prominent peaks are mostly a1-type
such as 1, 9a, and 12 modes at 959, 1180, and 622 cm-1. The b2-type modes
appear less prominently such as 3 and 9b at 1251 and 1130 cm-1, respectively. It is
interesting to note that the 6a mode, which has appeared prominently in all the
spectra analyzed so far, is absent in the B~ 2B2 MATI spectrum of C6H5Br+•.
4.4.2.3 Iodobenzene Cation
In a recent high resolution photoelectron spectrum of C6H5I, the B~ 2B2 state
appears overlapped with the A~ 2A2 state.29 Some vibrational bands are observed
but are much broader than those in the X~ 2B1 ground state. We could measure the
MATI signal in the 126.3~127.3 nm region of VUV which is near the ionization
threshold to the B~ 2B2 state. However, the MATI spectrum obtained was a
structureless broad band as seen in Fig. 4.9. This is due to a very short lifetime
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
152
85500 86000 86500 87000 87500
0-0
Ion
Sign
al
Photon Energy, cm-1
500 1000 1500
17b1141 8a1
19a1319a19b1
7a1
18a111
121
6b1
Ion Energy, cm-1
Fig. 4.8 The B~ 2B2 state one-photon MATI spectrum of C6H579Br+•. The x-
axis of the inset, ion energy, denotes energy scale referred to the position of
the 0-0 band.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
153
Table 4.7 Vibrational frequencies (in cm-1) and their assignments for the
bromobenzene cation in the B~ 2B2 excited state.
Modea
(Wilson) Symmetry Neutrala PESb This work
1 a1 1001 970 959
3 b2 1264 1251
6b b2 614 542
7a a1 1070 1015
8a a1 1578 1571
9a a1 1176 1180
9b b2 1158 1130
12 a1 671 620 622
14 b2 1321 1333
17b b1 904 889
18a a1 1020 982
19a a1 1472 1419
a Vibrational assignments in Wilson notation and frequencies for the ground
state neutral taken from ref. 26.
b ref. 28.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
154
78600 78800 79000 79200
Ion
Sign
al
Photon Energy, cm-1
Fig. 4.9 The B~ 2B2 state one-photon MATI spectrum of C6H5I+•.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
155
(<10-13 sec) of the B~ 2B2 state possibly caused by rapid internal conversion to
lower electronic states.
4.4.3 Conclusions
An interesting feature common to the ground state MATI spectra of chloro-,
bromo-, and iodobenzene cations is the presence of strong 6an progression. In
contrast, such a progression is absent for C6H5F+• even though 6a1 appears
prominently. Also, the 0-0 band is especially dominant in the C6H5F+• spectrum
compared to the spectra of the other cations. Through ab initio calculations at the
HF/6-31G** level, Wright and coworkers found that the equilibrium geometry of
C6H5Cl+• in the ground state was distorted from that of the neutral mostly along
the 6a mode eigenvector. To see if the change in molecular geometry upon
ionization can also explain the presence/absence of 6a progression for the
halobenzene cations, similar calculations have been done in this work. Since the
equilibrium geometries of the ground state neutrals and cations are only slightly
different from C2v, results calculated with the C2v symmetry will be presented.
Calculations were done at the Hartree-Fock (HF), the second order Møller-Plesset
perturbation theory (MP2), and the B3LYP density functional theory levels with
various basis sets using GAUSSIAN 98 suite37 of programs. MP2 and B3LYP
calculations gave similar results. Also, the results for bromo- and iodobenzene
were qualitatively similar to those for chlorobenzene. Hence, only the results for
fluoro- and chlorobenzene obtained at the B3LYP/6-311++G** level will be
compared here. Fig. 4.10 shows the equilibrium geometries of fluoro- and
chlorobenzene neutrals and cations in their ground states obtained at this level.
Geometrical changes upon ionization magnified by 20 are drawn as arrows in the
drawings of the neutral geometries. The 6a eigenvectors for the cations are shown,
also as arrows, in the drawings of the cation geometries. As was pointed out by
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
���
Fig. 4.10 Equilibrium geometries of (a) C6H5F and C6H5F+• and (b) C6H5Cl
and C6H5Cl+• calculated at the B3LYP/6-311++G** level. Atomic
displacements upon ionization are drawn as broken arrows in the drawings
of the neutrals. The 6a eigenvectors of the cations are drawn as arrows in the
drawings of the cations. Bond lengths in Å and angles in degree.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
157
Wright and coworkers, the main geometrical change upon ionization of C6H5X
(X=F, Cl) is the contraction of the C-X bond, contraction of the center two C-C
bonds, and elongation of the other C-C bonds. Also, the atomic displacements
upon ionization resemble the 6a eigenvectors as was pointed out by Wright and
coworkers. The 6a eigenvector of the fluorobenzene cation is different from the
others mostly in the significant movement of the C(1) atom where the halogen
atom is attached. Then, the fact that C(1) is displaced only slightly upon
ionization would lead to less overlap between the displacement and the 6a
eigenvector for the fluorobenzene cation. This may be the explanation for the
experimental results that the 6an progression is extensive for the chloro-, bromo-,
and iodobenzene cations and not for C6H5F+•.
In Chapter 2, for charge exchange study of halobenzene cations, the B~ 2B2
states of C6H5Cl+• and C6H5Br+• were found to be long-lived (microseconds or
longer) while those of C6H5I+• and C6H5F+• underwent rapid internal conversion
to lower electronic states.38 Structureless broad band observed for the B~ 2B2
states of C6H5I+• in this work is compatible with the above result. Appearance of
well resolved vibrational peaks in the excited state MATI spectra of C6H5Cl+• and
C6H5Br+• is also compatible with the above charge exchange results. The
vibrational bandwidth of ~10 cm-1 measured under the best experimental
condition in this work corresponds to a lifetime of ~500 fsec. However the
bandwidth seems to be mostly due to factors other than lifetime broadening such
as ionization over a range of Rydberg states in MATI and participation of various
rotational states. For example, the 0-0 bands in the excited state spectra have
essentially the same widths as those in the ground state spectra. Namely, present
MATI spectra alone cannot prove or disprove metastability of C6H5Cl+• and
C6H5Br+• in the B~ 2B2 state. Since the VUV wavelength shorter than the LiF
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
158
cutoff limit was not available, we could not record the MATI spectrum of C6H5F+•
in the B~ 2B2 state. We do not expect to observe a MATI spectrum with
vibrational structure in this case, however, because the corresponding peak looks
broad in the photoelectron spectrum.
4.5 VUV-MATI Spectroscopy of Difluorobenzenes
The structure of p-difluorobenzene cation has been extensively investigated by
analyzing ZEKE and MATI spectra obtained with the two color 1+1' scheme.39-41
Reiser et al. concluded that the ionic ground state has the D2h symmetry just as the
neutral in the ground state.40 On the other hand, Fujii et al. suggested that the
molecular symmetry in the Rydberg state is reduced to C2h based on the fact that p
and f Rydberg series were not observed.42 For o- and m-difluorobenzene cations,
accurate ionization energies, vibrational frequencies, and their structures have not
been reported.
In this section, we present the vibrational spectra of o-, m-, and p-
difluorobenzene cations in the ground electronic states obtained by using the one-
photon VUV-MATI technique. Accurate ionization energies and complete
vibrational assignments are presented. Symmetries of the cations in the ground
electronic states will be discussed based on the spectral features and quantum
chemical results.
4.5.1 Computational
Calculations of equilibrium geometries and vibrational frequencies of o-, m-,
and p-difluorobenzene neutrals and cations in the ground states were performed at
the Hartree-Fock (HF), the second order Møller-Plesset perturbation theory (MP2),
and the B3PW91 and B3LYP density functional theory levels with various basis
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
159
sets using GAUSSIAN 98 suite of programs.37 At all the levels adopted, the
geometrical changes upon ionization look qualitatively the same. Hence, only the
geometrical parameters obtained at the MP2 and B3LYP levels with the highest
basis set used the 6-311+G (df, p), will be tabulated and discussed. For the
calculated vibrational frequencies of difluorobenzene cations, the results obtained
at the B3LYP/6-311+G (df, p) levels will be quoted, which displayed the best
agreement with the experimental frequencies. Frequencies as obtained by
calculation will be presented, namely without scaling.
4.5.2 Molecular Geometry Calculation
As has been mentioned earlier, equilibrium geometries of p-, m-, and o-
difluorobenzene neutrals and cations in the ground electronic states have been
calculated at the HF, MP2, and DFT levels using various basis sets. Regardless of
the levels and basis sets adopted, the point group symmetries of the equilibrium
geometries were the same, namely D2h for the neutral and cation of p-
difluorobenzene and C2v for the neutrals and cations of m- and o-difluorobenzenes.
Symmetries of the neutrals in the ground electronic states, X~ 1A1g for para and
X~ 1A1 for meta and ortho, are in agreement with previous experimental results.34
Among the cations, symmetry is known only for the para isomer in the ground
state ( X~ 2B2g), which is D2h in agreement with the present calculated result.39,40
Structural data obtained at the B3LYP level displayed better agreement with the
experimental results than those at other levels. The B3LYP results obtained with
the largest basis set used, 6-311++G (2df, 2pd), are listed in Tables 4.8 - 4.10. Full
structural data determined by electron diffraction are available for p-
difluorobenzene neutral and partially for the meta isomer.43,44 These are listed in
Tables 4.8 and 4.9 also. It is to be noted that the calculated results are in good
agreement with the experimental data for the para and meta isomers. Similar trend
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
160
was observed for the calculated vibrational frequencies in the case of p-
difluorobenzene cation. Hence, only B3LYP results will be quoted in the
vibrational assignment to be presented later.
4.5.3 Ionization Energies
MATI spectra of p-, m-, and o-difluorobenzenes are shown in Figs. 4.11~4.13.
The intense peaks appearing at the lowest photon energy, namely at around 73853,
75326, and 74994 cm-1 in Figs. 4.11~4.13, respectively, correspond to the 0-0
bands. The position of the 0-0 band in a one photon MATI spectrum is equivalent
to the ionization energy of the molecule. However, the ionization energy thus
measured is usually a little smaller than the correct value because the molecules in
ZEKE states a few cm-1 below the threshold can also be ionized when a high PFI
field is used. To correct for this effect, the 0-0 band position was measured using
various PFI fields and the accurate ionization energy was estimated by
extrapolation to the zero field limit. Spoil field was not used in such
measurements.
The ionization energies to the ground electronic states of p-, m-, and o-
difluorobenzene cations measured from the MATI spectra in this work are listed
in Table 4.11 together with the previous measurements.40,41,46 The ionization
energy of p-difluorobenzene obtained in this work is smaller than the previous
MATI and ZEKE results26,23 by a few cm-1. Our previous measurements17,47 of the
ionization energies of chlorobenzene and iodoethane were in excellent agreement
with the ZEKE results.25,48 Hence, we do not have an explanation for the above
discrepancy at the moment. Accurate ionization energies of m- and o-
difluorobenzenes obtained in this work are in excellent agreement with the
previous data obtained by photoelectron spectroscopy49 and Rydberg
spectroscopy50 which are not as accurate.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
161
Table 4.8 Geometrical parameters of p-difluorobenzene in the ground state
( X~ 1Ag) and those of the cation in the ground state ( X~ 2B2g) calculated at the
B3LYP/6-311++G (2df, 2pd) level.
B3LYP Exp.a
D2h
X~ 1Ag X~ 2B2g X~ 1Ag
Bond Length (Å)
C1-F1 1.350 1.298 1.354
C1-C2 1.384 1.420 1.388
C2-C4 1.390 1.362 1.400
C2-H2 1.080 1.081 1.088
Bond Angle (º )
C3-C1-C2 122.2 123.6 123.5
C1-C2-H2 119.8 119.2 123.3
a The experimental values by electron diffraction in ref. 43.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
162
Table 4.9 Geometrical parameters of m-difluorobenzene in the ground state
( X~ 1A1) and those of the cation in the ground state ( X~ 2B1) calculated at the
B3LYP/6-311++G (2df, 2pd) level.
B3LYP Exp.a
C2v X~ 1A1 X~ 2B1 X~ 1A1
Bond Length (Å)
C2-F2 1.347 1.300 1.351b
C1-C2 1.384 1.380
C2-C4 1.384 1.442
C4-C6 1.390 1.384
C1-H1 1.079 1.080
C4-H4 1.080 1.082
C6-H6 1.081 1.080
Bond Angle (º )
C3-C1-C2 117.0 115.4 116.0
C1-C2-C4 122.7 123.3 123.9
C1-C2-F2 118.3 119.9
C2-C4-C6 118.2 119.4 117.9
C2-C4-H4 119.8 118.1
C4-C6-C5 121.0 119.1 120.4
a The experimental values by electron diffraction in ref. 44. b The experimental values through the combined use of NMR, electron
diffraction, and microwave spectroscopy data in ref. 45.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
163
Table 4.10 Geometrical parameters of o-difluorobenzene in the ground state
( X~ 1A1) and those of the cation in the ground state ( X~ 2B1) calculated at the
B3LYP/6-311++G (2df, 2pd) level.
B3LYP C2v
X~ 1A1 X~ 2B1
Bond Length(Å)
C1-F1 1.342 1.295
C1-C2 1.388 1.454
C1-C3 1.382 1.386
C3-C5 1.391 1.378
C5-C6 1.390 1.440
C3-H3 1.081 1.081
C5-H5 1.081 1.081
Bond Angle( º )
C2-C1-C3 120.5 121.1
C2-C1-F1 119.2 117.0
C1-C3-C5 119.3 117.4
C1-C3-H3 118.9 119.8
C3-C5-C6 120.2 121.6
C3-C5-H5 119.5 119.8
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
164
4.5.4 p-Difluorobenzene Cation
Six normal modes of the p-difluorobenzene cation with D2h symmetry are
totally symmetric (ag), the normal modes 1~6 in Mulliken notation. In the MATI
spectrum of p-difluorobenzene, Fig.4.11, many peaks appear prominently in
addition to the 0-0 band, namely at 368, 441, 839, 881, 1152, 1278, 1319, 1379,
1590, 1640, 1816, 2031, 2077, 2470, 2518, 2908, 2968, and 3015 cm-1 in terms of
the vibrational frequencies of the cation. In the beginning, it is assumed that all
these peaks are due to fundamentals, overtones, or combinations of the totally
symmetric modes. Frequencies of the ag modes calculated at the B3LYP level
with various basis sets are shown in Table 4.12. Frequencies calculated with
different basis sets tend to be quite similar, especially with the basis set 6-311+G
(df, p) or larger. Then, by comparing with the calculated frequencies, it is rather
straightforward to find the fundamentals from the prominent peaks mentioned
above, namely peaks at 3015, 1640, 1379, 1152, 839, and 441 cm-1 being
assignable to the modes 1~6, respectively. All the remaining peaks in the above
set except the peak at 368 cm-1 can be easily assigned to the overtones and
combinations of the normal modes 2~6. For example, the peaks at 881(441×2)
and 1319(441×3) are 62 and 63 overtones, respectively, and the peaks at 1590,
2031, 2470, and 2908 cm-1 are the 6n41 combinations. Successful assignments of
the overtones and combinations of the ag modes support the above assignments of
the fundamentals. It is obvious that the distinct peak at 368 cm-1 can not be
assigned to an ag mode and must be assigned to some other symmetry.
The excellent correlation between the frequencies of the ag modes calculated at
the B3LYP level and the experimental results leads one to expect similarly decent
correlation for non-totally symmetric modes also. Complete vibrational
assignments of the peaks in the MATI spectrum of p-difluorobenzene thus made
are listed in Table 4.13. Symmetry of each mode and its Mulliken notation are
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
165
provided in the table together with the Wilson notation. Also listed are vibrational
frequencies of the neutral, frequencies of the cation calculated at the B3LYP/6-
311++G (2df, 2pd) level, and frequencies measured in the previous and present
experiments. It is to be noted that the present assignment is in complete agreement
with the previous ones. This means that the B3LYP/6-311++G (2df, 2pd) results
have been excellent guidelines in the present case.
The most remarkable feature in the MATI spectrum of p-difluorobenzene is
the appearance of the 6n progression and its combinations with other ag modes,
namely 6n21, 6n31, 6n41, and 6n51. This indicates that the molecular geometry
changes along the direction of the mode 6 eigenvector upon ionization. This will
be discussed later. Other than predominance of the fundamentals, overtones, and
combinations of the ag modes, no further symmetry selectivity is observed in the
MATI spectrum. In fact, all the vibrational modes of p-difluorobenzene cation
except 11, 12, 18, 27, and 28 are observable in the MATI spectrum.
The most remarkable feature in the MATI spectrum of p-difluorobenzene is
the appearance of the 6n progression and its combinations with other ag modes,
namely 6n21, 6n31, 6n41, and 6n51. This indicates that the molecular geometry
changes along the direction of the mode 6 eigenvector upon ionization. This will
be discussed later. Other than predominance of the fundamentals, overtones, and
combinations of the ag modes, no further symmetry selectivity is observed in the
MATI spectrum. In fact, all the vibrational modes of p-difluorobenzene cation
except 11, 12, 18, 27, and 28 are observable in the MATI spectrum.
4.5.5 m-Difluorobenzene Cation
A low resolution photoelectron spectrum is the only available spectral
information for the m-difluorobenzene cation.49 Namely, the symmetry and
vibrational frequencies of this ion in the ground electronic state are not known. As
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
166
Table 4.11 Ionization energies (IE) to the ground states of p-, m-, and o-
difluorobenzene cations, in eV.
IE ( X~ ) Ref.
p-difluorobenzene 9.1576±0.0006 This work
9.1590±0.0004 40
9.1589±0.0006 41
9.161±0.002 46
m-difluorobenzene 9.3400±0.0006 This work
9.32±0.02 49
9.34 50
o-difluorobenzene 9.2992±0.0006 This work
9.30±0.02 49
9.30 50
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
167
Table 4.12 Frequencies (in cm-1) of the totally symmetric modes of the p-difluorobenzene cation in the ground
electronic state ( X~ 2B2g) calculated at the B3LYP level with various basis sets.
Modea Symm. Neutrala 4-31G(d,p) 6-31G(d,p) 6-311+G(df,p) 6-311++G(df,pd) 6-311++G(2df,2pd) Exp.b
1[2] ag 3084 3238 3240 3218 3218 3219 3015
2[8a] ag 1617 1697 1694 1679 1681 1680 1640
3[7a] ag 1245 1425 1417 1399 1399 1396 1379
4[9a] ag 1142 1176 1173 1175 1177 1174 1152
5[1] ag 858 855 853 848 848 849 839
6[6a] ag 451 449 447 449 449 449 441
a Vibrational assignments in Mulliken notation (Wilson notation in square bracket) and frequencies for the
ground state neutral taken from ref. 34.
b This work.
Copyright(c)2002 by Seoul N
ational University L
ibrary. All rights reserved.(http://library.snu.ac.kr) 2004/01/28 17:41:01
168
0 1000 2000 3000
74000 75000 76000 77000
61131
3151
2141
21211
91
261 101
6n21
6n31
6n41
51 41
21
31
61291161291
171
6n
314161
30181
6n51
11
0-0
Ion
Sign
al
Photon Energy, cm-1
Fig. 4.11 One-photon MATI spectrum of p-C6H4F2 recorded by monitoring
p-C6H4F2+• in the ground electronic state. The x-scale at the top of the
figure corresponds to the vibrational frequency scale for the cation. Its
origin is at the 0-0 band position.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
169
Table 4.13 Vibrational frequencies (in cm-1) and their assignments for the
p-difluorobenzene cation in the electronic ground state ( X~ 2B2g).
Modea Symm. Neutrala PESb MATIc ZEKEd Calculated This work 1[2] ag 3084 3219 3015 2[8a] ag 1617 1597 1640 1680 1640 3[7a] ag 1245 1339 1375 1396 1379 4[9a] ag 1142 1113 1150 1149 1174 1152 5[1] ag 858 831 836 836 849 839 6[6a] ag 451 427 439 440 449 441 7[17a] au 943 1001 1015e
8[16a] au 347 359 358 374 368f 9[10a] b1g 800 780 768 10[20a] b1u 3088 3209 3094 11[19a] b1u 1511 1507 12[13] b1u 1212 1317 13[18a] b1u 1012 985 983 14[12] b1u 737 749 743/731g 15[5] b2g 928 1005 1015e
16[4] b2g 692 743 743/731g
17[10b] b2g 375 303 302 306 303 18[20b] b2u 3028 3218 19[19b] b2u 1437 1518 1526 20[14] b2u 1285 1352 1351 21[18b] b2u 1085 1131 1113 22[15] b2u 350 368 368f
23[7b] b3g 3084 3210 3054 24[8b] b3g 1432 1459 25[3] b3g 1285 1277 1278h
26[6b] b3g 635 591 583 27[9b] b3g 434 430 436 28[17b] b3u 833 859 891 29[16b] b3u 504 508 524 515
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
170
30[11] b3u 166 129 127 126 129 124 61301 565 6181 799 803 62 880 881
61291 954 61261 1025
61(141/161) 1169 61(141/161) 1182
6191 1207 6281 1237 1244 5161 1276 1278
63 1319 1319 61131 1419 4161 1589 1590 52 1674 1671
5162 1714 1716 64 1759 1761
3161 1816 1816 62131 1863 6271 1891 4151 1988 1992 4162 2028 2031 2161 2077 2077 5261 2114 2112 5163 2154 2152 3151 2219 3162 2257 2257 42 2302 2302
6371 2329 31131 2362 415161 2426 2431 4163 2471 2470 2162 2517 2518
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
171
5262 2552 2550 5164 2590 3163 2697 21211 2753 2141 2789
415162 2867 4164 2908
314161 2968
a Vibrational assignments in Mulliken notation (Wilson notation in square
bracket) and frequencies for the ground state neutral taken from ref. 34. b ref. 46. c ref. 41. d ref. 40. e This peak may be assigned alternatively to 71 or 151.
f This peak may be assigned alternatively to 81 or 221.
g This peak may be assigned alternatively to 141 or 161. h This peak may be assigned alternatively to 251 or 6151.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
172
has been mentioned earlier, quantum chemical calculations performed at various
levels in this work suggest that the ground state cation belongs to the C2v point
group. Then, eleven normal modes of the cation, modes 1~11, are totally
symmetric (a1). By comparing the vibrational frequencies of the prominent peaks
in Fig.4.12 with the calculated frequencies, fundamentals, overtones, and
combinations of the a1 modes could be easily identified. Among the a1 modes,
the modes 1~3 which are due to C-H stretching do not appear distinctly in the
MATI spectrum and can not be identified. Among the prominent peaks, those at
379, 885, 909, and 1496 cm-1 could not be assigned to a1 symmetry. By
comparing with the calculated frequencies, these were assigned to 19(b1), 16(b1),
12(a2), and 23(b2), respectively. Complete assignment for the vibrational peaks
in the MATI spectrum of m-difluorobenzene is shown in Table 4.14. A
noticeable feature in the spectrum is the appearance of the 6n and 10n
progressions. If the symmetry of the cation is lowered from C2v, some of the
modes become totally symmetric and may occur prominently in the MATI
spectrum. For example, if the symmetry is lowered from C2v to Cs, b2 symmetry
in C2v becomes a′ which is totally symmetric. There are ten normal modes with
b2 symmetry, modes 21-30, in m-difluorobenzene cation. Among these, only the
mode 23 appears distinctly, as has been mentioned earlier. In fact, the b2 modes
are not particularly more prominent than the a2 and b1 modes. This suggests that
the cation retains the C2v symmetry of the neutral as found in quantum chemical
calculations.
4.5.6 o-Difluorobenzene Cation
Just as for the m-difluorobenzene cation, a low resolution photoelectron
spectrum is the only spectral information available for the o-difluorobenzene
cation.49 C2v symmetry is also found in this case from quantum chemical
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
173
0 1000 2000 3000
76000 77000 78000
4181
51121611116191
161
301201
121
181
61231
5161
51101
231
91101171
101111
131
191
6110n
71 4151
81
111 91
6n
141
10n
0-0
Ion
Sign
al
Photon Energy, cm-1
Fig. 4.12 One-photon MATI spectrum of m-C6H4F2 recorded by
monitoring m-C6H4F2+• in the ground electronic state. The x-scale at the
top of the figure corresponds to the vibrational frequency scale for the
cation. Its origin is at the 0-0 band position.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
174
Table 4.14 Vibrational frequencies (in cm-1) and their assignments for the m-
difluorobenzene cation in the ground electronic state ( X~ 2B1).
Modea Symm. Neutrala Calculated This work 1[20a] a1 3222 2[2] a1 3087 3223 3[7a] a1 3049 3206 4[8a] a1 1605 1572 1605 5[19a] a1 1449 1538 1559 6[13] a1 1277 1374 1355 7[18a] a1 1066 1111 1092 8[12] a1 1008 997 987 9[1] a1 735 740 728
10[6a] a1 524 519 505 11[9a] a1 331 346 343
12[17a] a2 879 933 909 13[16a] a2 599 622 614 14[10a] a2 251 204 199 15[5] b1 978 1006
16[17b] b1 850 904 885 17[11] b1 769 802 784 18[4] b1 672 599 587
19[16b] b1 458 399 379 20[10b] b1 230 190 190 21[20b] b2 3096 3209 22[8b] b2 1613 1432 1429 23[19b] b2 1490 1523 1496 24[3] b2 1337 1279 25[14] b2 1260 1399 1377 26[9b] b2 1158 1119 27[18b] b2 1120 1228 1258 28[7b] b2 954 928 923 29[6b] b2 514 489 467 30[15] b2 478 421 418
142 393
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
175
111141 543 112 686
111301 759 111291 806 101111 846
102 1008 101111141 1047
811201 1175 91101 1231 101171 1289 91181 1316 81111 1330 103 1513
61111 1694 8191 1713 91102 61191
1734
102171 1794 61101 1859 101251 1876 101231 1999 51101 2060 6191 2085 61102 2363 51121 2465 4181 2585 62 2706
61231 2848 61103 2864 5161 2910
a Vibrational assignments in Mulliken notation (Wilson notation in square
bracket) and frequencies for the ground state neutral taken from ref. 34.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
176
0 1000 2000 3000
75000 76000 77000 78000
41111
251
271
10n111
171 28
1
301
151
201291
71 101
41101
3110121
6171 41 3181
91
11n
161
61101
10n
0-0
Ion
Sign
al
Photon Energy, cm-1
200 400 600 800
141
111301
112291
191
151
301161
201
Ion Energy, cm-1
Fig. 4.13 One-photon MATI spectrum of o-C6H4F2 recorded by
monitoring o-C6H4F2+• in the ground electronic state. The x-scale at the top
of the figure corresponds to the vibrational frequency scale for the cation.
Its origin is at the 0-0 band position. Spectrum in the 100~800 cm-1 region
magnified by 30 is shown as an inset to demonstrate the quality of the
MATI spectra obtained in this work.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
177
Table 4.15 Vibrational frequencies (in cm-1) and their assignments for the o-
difluorobenzene cation in the ground electronic state ( X~ 2B1).
Modea Symm. Neutrala Calculated This work 1[2] a1 3081 3221 3089
2[20b] a1 3045 3210 3032 3[8a] a1 1605 1575 1548
4[19b] a1 1514 1516 1487 5[14] a1 1292 1399 1397b 6[7a] a1 1272 1364 1342c 7[9a] a1 1152 1196 1171
8[18b] a1 1025 979 966 9[1] a1 762 762 749
10[6a] a1 568 572 553 11[15] a1 287 297 294 12[5] a2 970 1012
13[17a] a2 840 883 887d 14[4] a2 703 746 704
15[16a] a2 588 423 413 16[10b] a2 196 150 148 17[17b] b1 929 992 994 18[11] b1 749 782 799
19[16b] b1 450 456 442e 20[10a] b1 275 262 254 21[7b] b2 3218 22[20a] b2 3060 3203 23[8b] b2 1618 1534 1523 24[19a] b2 1472 1346 1342c 25[3] b2 1253 1206 1202 26[13] b2 1206 1473 1435 27[18a] b2 1103 1124 1065 28[12] b2 857 856 29[6b] b2 546 533 505 30[9b] b2 440 394 382 111161 442
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
178
112 588 111301 677 101111 848 291301 887
101301 930 101291 1051
102 1104 81111 1261
102111 1397
103 1658 71101 1723
101251 1755 41111 1778 171281 1842 61101 1896 51101 1949 41101 2039 41101 2099 7181 2138 4191 2237 71102 2278 6181 2308 4181 2452 6171 2509 41102 2591
a Vibrational assignments in Mulliken notation (Wilson notation in square
bracket) and frequencies for the ground state neutral taken from ref. 34. b This peak may be assigned alternatively to 102111.
c This peak may be assigned alternatively to 61 or 241. d This peak may be assigned alternatively to 291301. e This peak may be assigned alternatively to 111161.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
���
calculations. All the a1 vibrational modes except for the modes 1 and 2 which are
due to C-H stretching appear prominently in the MATI spectrum. Experimental
and calculated vibrational frequencies and assignments are listed in Table 4.15.
In this case overtones and combinations do not appear as abundantly and
intensely as in the case of para and meta isomers, even though 10n and 11n
progressions are observed. Frequency of the peak at 848 cm-1 is close to the
calculated frequency of the mode 28(b2). Then, it becomes the only b2 mode
with prominent intensity. Considering appearance the 10n and 11n progressions,
however, we would rather assign it to the combination 101111. Then, none of the
b2 modes which become totally symmetric upon symmetry reduction to Cs
appear prominently in the MATI spectrum of o-difluorobenzene. This suggests
that the C2v symmetry of the neutral is retained in the cation.
4.5.7 Geometrical Change upon Ionization and Vibrational
Progressions
A remarkable feature common to the MATI spectra of three isomers of
difluorobenzene is the appearance of strong overtone progressions of some
totally symmetric modes. These are 6n for the para, 6n and 10n for the meta, and
10n and 11n for the ortho isomers. Progressions of combinations involving these
overtones are observed abundantly also. Appearance of overtones of a particular
totally symmetric mode, or large Franck-Condon factors for such overtones,
indicates that the geometrical change upon ionization is well overlapped with the
eigenvector of the mode involved. As an example, atomic displacements upon
ionization of p-difluorobenzene evaluated from the quantum chemical results in
Table 4.8 are compared with the eigenvector of the mode 6 of the cation in
Fig.4.14. Similarity of the two vectors is apparent in the figure. As a more
quantitative measure of the overlap, the atomic displacement vector was mass-
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
���
Fig. 4.14 Equilibrium geometries of (a) neutral and (b) cation of p-C6H4F2.
Arrows in (a) indicate atomic displacements upon ionization magnified by
10. Arrows in (b) indicate the eigenvector of the mode 6 of the cation.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
181
weighted and its projection on the eigenvector of each totally symmetric mode
was calculated. In the case of the para isomer, the absolute values of the
projection were 0.001(1), 0.13(2), 0.13(3), 0.12(4), 0.11(5), and 0.42(6) with the
numbers in the parentheses indicating the normal mode number. Large
projection along the mode 6 is in agreement with the prominent 6n progression
observed in the MATI spectrum. In the case of the meta isomer, these were
0.001(1), 0.001(2), 0.002(3), 0.12(4), 0.04(5), 0.15(6), 0.03(7), 0.10(8), 0.09(9),
0.32(10), and 0.12(11). Appearance of strong fundamentals of the modes 10 and
6 and their overtones and combinations is in good agreement with the
projections calculated. For the ortho isomer, the calculated absolute projections
of 0.000(1), 0.001(2), 0.13(3), 0.10(4), 0.08(5), 0.06(6), 0.08(7), 0.06(8), 0.08(9),
0.29(10), and 0.20(11) are in agreement with the strong fundamentals, overtones,
and combinations of the modes 10 and 11. Successful explanation of the above
spectral features indicates reliability of the results from the B3LYP/6-311++G
(2df, 2pd) calculation in the present cases.
4.5.8 Conclusions
VUV-MATI spectra of o-, m-, and p-difluorobenzene cations have been
obtained and vibrational frequencies have been assigned by comparing to those
calculated at B3LYP/6-311+G (df, p) level. Even though the structure and
vibrational frequencies of p-difluorobenzene cation have been completely
analyzed by ZEKE-PES spectroscopy using two color 1+1' scheme, one-photon
MATI spectra in our present work would have provided much more information
on vibrations of the fundamentals, their combinations and overtones. By
comparing the measured vibrational frequencies to the calculated, it is concluded
that the symmetries of the o-, m-, and p-difluorobenzene cations belong to the
point group of C2v, C2v, and D2h, respectively, same as those of neutrals. The 6an
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
182
progressions prominently appear as p-C6H4F2+• >m-C6H4F2
+• >o-C6H4F2+• in the
MATI spectra. Considering the geometrical change upon ionization from
calculations, it has been found that the overlaps between the atomic
displacements upon ionization and eigenvectors of the 6a mode are expected to
be substantial and decrease in turn p-C6H4F2+•, m-C6H4F2
+•, and o-C6H4F2+•. In
addition, the most common feature in the structure of all difluorobenzene cations
studied in the ground state is the contraction of the C-F bonds upon ionization
resulting from removal of a π electron in benzene ring.
4.6 UV-MATI Spectroscopy of Phenylacetylene and
Benzonitrile
One-photon MATI spectra of halobenzenes reported recently by this
laboratory showed well-resolved vibrational peaks of the corresponding cations,
which consisted mostly of fundamentals with proper symmetries. Vibrational
assignments were made by referring to the previous results, comparing with
calculated frequencies, and invoking the selection rule for one-photon process.
Difference in the geometry between the neutral and cation, or geometric change
upon ionization, was calculated quantum chemically and used to explain the
prominent overtones of some vibrational modes and combinations involving
these. As a more rigorous attempt to utilize spectral intensity information for
vibrational assignment, Franck-Condon factors were calculated from the
quantum chemical results in our recent study of an aliphatic halide.51 Theoretical
prediction of the intensities for transitions to all the totally symmetric vibrational
states was found to be extremely useful for reliable assignment.
In this section, we report the vibrational spectra of phenylacetylene and
benzonitrile cations obtained by one-photon MATI spectroscopy. Successful
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
183
vibrational assignments made based on the above strategy, namely by utilizing
the calculated frequencies and Franck-Condon factors and symmetry selection
rule, will be presented also.
4.6.1 Quantum Chemical Calculations
Calculations of equilibrium geometries and vibrational frequencies of
phenylacetylene and benzonitrile and their cations in the ground states were
performed at the density functional theory (DFT) levels, B3LYP, B3PW91, and
BP86, with various basis sets using GAUSSIAN 98 suite of programs. Size of
the basis set was systematically increased until the basis set dependence
disappeared. Hence, only those obtained with the largest basis set used, the 6-
311++G (2df, 2pd), will be listed and discussed. For the vibrational frequencies,
the results obtained at the BP86/6-311++G (2df, 2pd) level showed the best
agreement with the experimental data. Frequencies obtained by these
calculations are presented without scaling in Tables 4.16 and 4.17. At all the
levels used in the calculation, equilibrium geometries of the phenylacetylene and
benzonitrile neutrals and cations belong to the C2v symmetry. Invoking the
symmetry selection rule, prominent peaks in the MATI spectra must be mostly
due to transitions to the a1 vibrational states. The Franck-Condon factors for
such transitions, either fundamentals, overtones, or combinations, calculated at
three density functional theory (DFT) levels, B3LYP, B3PW91, and BP86 using
the 6-311++G (2df, 2pd) basis set were rather similar. The Franck-Condon
factors calculated at the BP86/6-311++G (2df, 2pd) level and normalized to that
of the 0-0 transition are listed in Tables 4.19 and 4.20 also. Comparing the
experimental and calculated frequencies and Franck-Condon factors was helpful
to identify a1 peaks. The remaining weak peaks in one-photon MATI spectra
must be due to electric dipole-forbidden but vibronically allowed transitions.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
184
Only the calculated frequencies, not intensities, can be used to assign these
nontotally symmetric transitions.
4.6.2 Ionization Energies
MATI spectra of phenylacetylene and benzonitrile recorded by monitoring
C6H5C≡CH+ and C6H5C≡N+ in the ground electronic states are shown in Figs.
4.15 and 4.16, respectively. The spectra magnified by 15 are also shown as
insets in the figures. The intense peaks appearing at the lowest photon energy,
namely at 71127 and 78461 cm-1 in Figs. 4.15 and 4.16, respectively, correspond
to the 0-0 bands. The position of the 0-0 band in a one-photon MATI spectrum is
equivalent to the ionization energy of the molecule. However, the ionization
energy thus measured is usually a little smaller than the correct value because
the molecules in ZEKE states9 some cm-1 below the threshold can also be
ionized when a high PFI field is applied. To correct for this effect, the 0-0 band
position was measured using various PFI fields and the accurate ionization
energy was estimated by extrapolation to the zero field limit. Spoil field was not
used in such measurements. The ionization energies to the ground electronic
states of phenylacetylene and benzonitrile cations measured in this work are
listed in Table 4.18 together with previous results.52-55 There has been no
previous report on the accurate ionization energy of phenylacetylene measured
by ZEKE or MATI. The ionization energy of phenylacetylene determined in this
work, 8.8195 ± 0.0006 eV, is a little different form 8.825 ± 0.001 eV measured
by threshold photoelectron spectroscopy (TPES).52 More annoying is that the
ionization energy of benzonitrile obtained in this work, 9.7288 ± 0.0006 eV, is
smaller than a previous ZEKE result54 by ∼ 20 cm-1. My experience is that the
present MATI technique which uses high voltage electronics tends to under-
estimate ionization energies by 0 ∼ 5 cm-1.52,53 We do not have an explanation
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
185
Table 4.16. Vibrational frequencies (in cm-1) of phenylacetylene neutral and
cation in the ground electronic states calculated at the B3LYP, B3PW91, and
BP86 levels with the 6-311++G (2df, 2pd) basis set and experimental data for
the neutral.
Neutral Cation Mode Symm.Exp.a B3LYP B3PW91 BP86 B3LYP B3PW91 BP86
1 a1 763 777 781 756 766 767 746 2 a1 3064 3202 3210 3125 3221 3228 3144 3 b2 1283 1310 1325 1290 1298 1304 1257 4 b1 691 708 707 683 741 738 699 5 b1 986 1014 1011 968 1045 1040 999 6a a1 467 474 471 459 470 467 455 6b b2 619 638 648 616 581 578 560 7a a1 3035 3172 3179 3094 3196 3201 3119 7b b2 3198 3206 3121 3219 3226 3142 8a a1 1598 1640 1651 1587 1643 1652 1592 8b b2 1573 1609 1621 1558 1536 1538 1483 9a a1 1178 1202 1198 1163 1210 1206 1172 9b b2 1158 1185 1180 1149 1175 1174 1139 10a a2 842 860 859 825 827 824 789 10b b1 165 142 142 136 120 118 115 11 b1 756 782 777 752 817 815 787 12 a1 1000 1018 1017 987 998 999 971 13 a1 1192 1223 1228 1193 1266 1271 1228 14 b2 1331 1358 1357 1325 1387 1393 1348 15 b2 161 159 152 153 151 145 16a a2 418 413 409 395 363 358 345 16b b1 531 558 558 536 481 479 457 17a a2 971 998 996 953 1027 1024 982 17b b1 918 946 944 904 991 988 949
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
186
18a a1 1028 1050 1053 1020 1009 1011 981 18b b2 1071 1102 1103 1069 1113 1115 1081 19a a1 1489 1526 1525 1474 1486 1483 1435 19b b2 1444 1478 1477 1429 1436 1442 1399 20a a1 3083 3191 3198 3113 3210 3216 3133 20b b2 3058 3180 3188 3103 3208 3214 3130 βCC b2 516 539 538 517 531 528 507 βCH b2 653 689 690 635 694 695 645 vCC a1 2118 2202 2209 2131 2107 2108 2053 vCH a1 3291 3468 3472 3393 3409 3412 3334 γCC b1 352 371 369 354 324 322 311 γCH b1 610 645 633 588 645 642 621
a Ref. 34.
Table 4.17. Vibrational frequencies (in cm-1) of benzonitrile neutral and cation
in the ground electronic states calculated at the B3LYP, B3PW91, and BP86
levels with the 6-311++G (2df, 2pd) basis set and experimental data for the
neutral.
Neutral Cation Mode Symm. Exp.a B3LYP B3PW91 BP86 B3LYP B3PW91 BP86
1 a1 769 774 774 752 755 756 736 2 a1 3071 3196 3215 3130 3210 3228 3143 3 b2 1289 1319 1334 1298 1416 1426 1382 4 b1 686 706 704 682 631 628 607 5 b1 987 1021 1016 974 1042 1035 994 6a a1 461 467 465 452 459 456 445 6b b2 629 641 637 619 504 500 483 7a a1 3042 3178 3185 3100 3196 3201 3118
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
187
7b b2 3027 3204 3212 3127 3219 3225 3141 8a a1 1599 1641 1654 1588 1660 1669 1603 8b b2 1584 1615 1627 1563 1275 1279 1235 9a a1 1178 1203 1200 1165 1210 1207 1172 9b b2 1163 1188 1184 1153 1148 1150 1110 10a a2 848 863 862 828 814 811 775 10b b1 172 147 146 140 118 116 111 11 b1 758 781 780 751 815 813 784 12 a1 1001 1019 1019 989 1001 1000 971 13 a1 1191 1220 1226 1188 1252 1256 1215 14 b2 1337 1361 1361 1329 1392 1389 1344 15 b2 162 169 166 160 157 155 148 16a a2 401 410 406 392 353 347 335 16b b1 548 573 574 552 449 442 413 17a a2 978 1002 1000 957 1029 1025 983 17b b1 925 954 952 912 990 985 947 18a a1 1027 1050 1053 1021 984 988 962 18b b2 1071 1105 1106 1073 1088 1093 1054 19a a1 1492 1528 1527 1475 1474 1469 1425 19b b2 1448 1481 1480 1431 1528 1529 1475 20a a1 3080 3207 3203 3118 3221 3216 3132 20b b2 3039 3188 3195 3110 3209 3214 3131 βCN b2 551 570 570 549 568 570 548 vCN a1 2232 2332 2341 2236 2192 2196 2124 γCN b1 381 392 390 375 317 312 298
a Ref. 34.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
188
Table 4.18. Ionization energies (IE) of phenylacetylene and benzonitrile, in eV.
IE ( X~ ) Ref.
Phenylacetylene 8.8195 ± 0.0006 This work
8.825 ± 0.001 TPES52
8.82 ± 0.02 PES53
Benzonitrile 9.7288 ± 0.0006 This work
9.7315 ± 0.0002 ZEKE54
9.71 ± 0.01 PI55
for the above discrepancy at the moment even though we would like to point out
that quality of the present MATI spectrum is superior to the previous ZEKE
spectrum54. Assuming that the shift of a vibrational peak in a MATI spectrum
due to the applied electric fields is similar to that of the 0-0 band, the vibrational
frequency corresponding to each peak can be determined simply by taking the
difference of its position from that of the 0-0 band. Vibrational frequency scales
with origins at the 0-0 band positions are also drawn in Figs. 4.15 and 4.16.
Vibrational frequencies of C6H5C≡CH+ and C6H5C≡N+ in the ground electronic
state calculated at the BP86/6-311++G (2df, 2pd) level are compared with the
experimental data in Tables 16 and 17, respectively. Also listed in the tables are
intensities of each peak in the MATI spectra normalized to that of the 0-0 band.
Frequencies of some vibrations measured previously by TPES52 and ZEKE54 are
also included in the tables.
4.6.3 Phenylacetylene Cation
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
189
The phenylacetylene cation with C2v symmetry has 36 nondegenerate normal
modes, thirteen of which belong to the a1 symmetry species, three to a2, eight to
b1, and twelve to b2. Among the a1 type modes, 2, 7a, 20a, and νCH in Wilson
notation33 are due to CH stretching and have frequencies of ∼ 3000 cm-1.
According to our previous study on the CH stretching modes, they are not
expected to appear distinctly in the one-photon MATI spectrum. Then,
prominent peaks in the spectral region 0 ∼ 1700 cm-1 and near ∼ 2000 cm-1 can
mostly be assigned to the fundamentals of the remaining a1 modes, 1, 6a, 8a, 9a,
12, 13, 18a, 19a, and vCC, as well as their overtones and combinations. Among
these, the calculated Franck-Condon factors are especially significant for the
fundamentals of 6a, 1, 12, 9a, 8a, and vCC whose calculated frequencies at the
BP86 level are 455, 746, 971, 1172, 1592, and 2053 cm-1, respectively. Hence,
the prominent peaks at 458, 747, 979, 1185, 1604, and 2040 cm-1 in the MATI
spectrum of phenylacetylene can readily be assigned to 6a1, 11, 121, 9a1, 8a1, and
vCC1, respectively. Even though the fundamental of v13, which is another a1-type
mode, is expected to be weak according to its calculated Franck-Condon factor,
it appears distinctly near its calculated frequency, at 1249 cm-1. 19a1 appears
rather distinctly at 1435 cm-1 even though its calculated Franck-Condon is only
0.004. We would rather assign this to a composite of 19a1 and 6a1121 based on
the calculated frequencies and Franck-Condon factors. A weak shoulder peak at
989 cm-1 is close to 981 cm-1 calculated for 18a1. We are reluctant to make such
an assignment because the calculated Franck-Condon factor for this transition is
extremely small, 0.0002. It is to be mentioned that the harmonic frequencies
calculated at the BP86 level are usually a little smaller than the experimental
ones while those at the B3LYP and B3PW91 levels are larger. The same trend
holds for most of the fundamentals of the phenylacetylene and benzonitrile
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
190
Fig. 4.15 One-photon MATI spectrum of C6H5C≡CH recorded by
monitoring C6H5C≡CH+ in the ground electronic state. The x-scale at the
top of the figure corresponds to the vibrational frequency scale for the
cation whose origin is at the 0-0 band position. Spectrum in the 50 ∼ 2500
cm-1 region magnified by 15 is shown as an inset to demonstrate the
quality of the MATI spectrum obtained in this work.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
191
Table 4.19 Vibrational frequencies (in cm-1) and their assignments for
phenylacetylene cation in the ground electronic state ( X~ 2B1).
Calculatedc This work Modea Symm. TPESb
Frequency Intensityd Frequency Intensityd
Fundamentals 1 a1 759 746 0.131 747 0.076 2 a1 3144 5×10-6 3 b2 1257 0 1287 0.007 4 b1 699 0 706 0.017 5 b1 999 0 996 0.008 6a a1 460 455 0.353 458 0.438 6b b2 560 0 561 0.008 7a a1 3119 3×10-6 7b b2 3142 0 8a a1 1592 0.096 1604 0.073 8b b2 1483 0 1505 0.006 9a a1 1172 0.081 1185 0.102 9b b2 1139 0 1158 0.018 10a a2 789 0 795e 0.015 10b b1 115 0 110 0.008 11 b1 787 0 795e 0.015 12 a1 971 0.069 979 0.066 13 a1 1228 0.010 1249 0.087 14 b2 1348 0 15 b2 145 0 143 0.005 16a a2 345 0 346 0.011 16b b1 457 0 17a a2 982 0 989 0.006 17b b1 949 0
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
192
18a a1 981 0.0002 18b b2 1081 0 1076 0.007 19a a1 1435 0.004 1435f 0.044 19b b2 1399 0 20a a1 3133 0.0003 20b b2 3130 0 βCC b2 504 507 0 499 0.012 βCH b2 645 0 658 0.008 vCC a1 2053 0.101 2040 0.042 vCH a1 3334 0.0001 γCC b1 311 0 303 0.010 γCH b1 621 0 622 0.010
Overtones and Combinations 10b2 a1 230 0.002 221 0.010 152 a1 290 0.0002 286 0.007
10b1γCC1 a1 426 0.002 409 0.034 6a110b1γCC1 a1 881 0.0007 865 0.006
6a2 a1 910 0.065 914 0.063 41γCC1 a1 1010 0.0003 1009 0.012 6b1βCC1 a1 1067 0.0004 1064 0.005 6a111 a1 1201 0.045 1205 0.082 6a3 a1 1365 0.008 1370 0.022 42 a1 1398 0.008 1407 0.028 6a1121 a1 1426 0.026 1435f 0.044 6a117a1 a2 1437 0 1448 0.009 121βCC1 b2 1478 0 1465 0.007 6a19a1 a1 1627 0.029 1643 0.017 6a1131 a1 1683 0.004 1706 0.025 1117a1 a2 1728 0 1739 0.010 6a119a1 a1 1890 0.002 1895 0.017
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
193
6a18a1 a1 2047 0.035 2058 0.017 9a1121 a1 2143 0.005 2164 0.017 121131 a1 2199 0.001 2228 0.013 118a1 a1 2338 0.011 2356 0.016 6a1vCC1 a1 2508 0.038 2496 0.027 8a1121 a1 2563 0.005 2582 0.012 11vCC1 a1 2764 0.008 8a19a1 a1 2799 0.012
2789 0.016
19a119b1 b2 2834 0 2837 0.007 6a2vCC1 a1 2963 0.007 2953 0.011 121vCC1 a1 3024 0.008 3022 0.010
6a18a1121 a1 3018 0.002 3032 0.008 8a2 a1 3184 0.006 3206 0.014 9a1vCC1 a1 3224 0.010 3227 0.011 131vCC1 a1 3281 0.002 3292 0.010
a Wilson notation. b Ref. 52. c BP86/6-311++G (2df, 2pd) level. d Normalized to the intensity of the 0-0 band. e A composite of 10a1 and 111. f A composite of 19a1 and 6a1121.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
194
neutrals as can be seen in Tables 4.16 and 4.17. Even though the results must
rise from different error cancellations at these levels,56 the correlation can be
used advantageously in the peak assignments.
The fact that 6a1 appeared most prominently in the one-photon MATI
spectrum indicates that the geometrical change upon ionization occurs mostly
along the 6a eigenvector. This has been confirmed by calculation, even though
not shown here, by projecting the geometrical change vector on the 6a
eigenvector. This also suggests that 6an overtones and combinations of 6a and
other a1 modes would appear prominently in the one-photon MATI spectrum.
Accordingly, the prominent peaks at 914 and 1370 cm-1 can be assigned to 6a2
and 6a3, respectively. Also, the distinct peaks at 1205, 1435, 1643, 1706, 2058,
and 2496 can be assigned to 6a combinations, 6a111, 6a1121, 6a19a1, 6a1131,
6a18a1, and 6a1vCC1, respectively, from the calculated frequencies and Franck-
Condon factors. The peak at 1435 cm-1 can be alternatively assigned to 19a1 as
has been mentioned earlier. Other combinations of the a1 modes, 1118a1, 9a1121,
and 118a1 appeared distinctly also at 1739, 2164, and 2356 cm-1, respectively.
The Franck-Condon factors for 11vCC1 and 8a19a1 calculated at the BP86 level
are 0.012 and 0.008, respectively. Hence, the peak at 2789 cm-1, which can be
assigned either to 11vCC1 or to 8a19a1 based on the frequencies may better be
assigned to a composite of the two transitions, 11vCC1/8a19a1. The calculated
Franck-Condon factors for the fundamentals of the CH stretching modes 2, 7a,
20a, and νCH are negligible. This is understandable because the lengths of all
the CH bonds hardly change upon ionization. Extremely weak peaks appeared in
the 3000 ∼ 3200 cm-1 region of the present MATI spectrum (Fig. 4.16). We are
reluctant to assign them to 21, 7a1, 20a1, or νCH1 because their calculated
Franck-Condon factors are very small. Instead, it is more likely that they are
overtones or combinations. Thus, the very weak peaks at 3022, 3206, and 3292
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
195
were assigned to 121vCC1, 8a2, and 131vCC1, respectively. Some distinct peaks
may be assigned to combinations of nontotally symmetric modes with a1 overall
symmetry. Hence, the distinct peaks at 221, 286, 409, 1009, and 1407 were
assigned to 10b2, 152, 10b1γCC1, 41γCC1, and 42, respectively.
The fundamentals of nontotally symmetric modes do not appear in the
simulated spectrum because the Franck-Condon factors calculated under the
Born-Oppenheimer approximation are zero for these transitions. They still
appeared in the actual spectrum, even though very weakly, through vibronic
mechanism. Hence, very weak peaks at 303, 346, 499, 561, 622, 706, and 795
cm-1 can be assigned to γCC1, 16a1, βCC1, 6b1, γCH1, 41 and 10a1/111 by
comparing with the calculated frequencies of 311, 345, 507, 560, 621, 699, and
789/787 cm-1, respectively. A very weak 16b1 is expected at ~457 cm-1 but must
have been buried in the strong 6a1 transition. Similarly, 17a1 and 17b1 expected
at 950~980 cm-1 may have been buried as shoulders of 121. The weak shoulder
peak at 989 cm-1 mentioned previously may be 17a1 rather than 18a1.
4.6.4 Benzonitrile Cation
Kimura and coworkers reported 1+1' ZEKE spectrum of benzonitrile in the 0
~ 1200 cm-1 vibrational energy region.54 Tentative assignments were made by
comparing with results from PM3 semi-empirical calculations. They adopted
Mulliken notation for vibration modes, with some minor errors. We reinterpreted
their assignments using the DFT frequencies calculated in this work and listed
them in Table 4.20.
A spectrum with much higher quality than the above was obtained in the 0 ~
2500 cm-1 region (78200 ~ 81000 cm-1 in photon energy) by one-photon MATI
in this work. The 81000 ~ 81500 cm-1 region could not be recorded due to a dip
in VUV output. According to our experience in MATI of phenylacetylene in this
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
196
work and halobenzenes in previous studies, this region is not expected to be
important because the CH stretching fundamentals would not be observed
anyway. Symmetry selection rule and frequencies and Franck-Condon factors
calculated at the DFT levels will be utilized to assign the observed vibrational
peaks as has been done for phenylacetylene.
Quantum chemical calculations performed at various levels for this molecule
suggest that the cation belongs to the C2v point group in the ground electronic
state. Twelve normal modes are a1-type as listed in Table 4.17. Among these, 2,
7a, and 20a are high frequency vibrations with CH stretching character and can
be neglected in this work. Among the a1-type vibrations, the fundamental of 6a
displays the largest Franck-Condon factor and is readily identified as the peak at
438 cm-1. Franck-Condon factors are significant for 11, 18a1, 121, 9a1, 8a1, and
vCN1 with the calculated frequencies of 736, 962, 971, 1172, 1603, and 2124
cm-1, respectively. The prominent peaks at 737, 959, 968, 1168, 1612, and 2136
cm-1 in the MATI spectrum correspond to these transitions. The remaining a1-
type fundamentals are 131 and 19a1 with the calculated frequencies of 1215 and
1425 cm-1, respectively, and the calculated Franck-Condon factors of only 0.001
for each. Even though one might assign the peak at 1413 cm-1 to 19a1 based on
the frequency, 6a1121 seems to be a better assignment when the intensity is
considered also. For 131, the peak at 1223 cm-1 is the only candidate even though
the calculated and observed intensities show a substantial discrepancy. For the
peak at 1223 cm-1, no alternative assignment is possible, either to a fundamental,
overtone, or combination. Namely, its assignment to 131 is unavoidable and its
small Franck-Condon factor must be attributed to inaccuracy in the calculation.
It is to be noted from Table 4.20 that the Franck-Condon factor for the same
transition in phenylacetylene is also much smaller than the experimental
intensity, even though not as dramatically as in benzonitrile. It is to be
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
197
Fig. 4.16 One-photon MATI spectrum of C6H5C≡N recorded by
monitoring C6H5C≡N+ in the ground electronic state. The x-scale at the top
of the figure corresponds to the vibrational frequency scale for the cation
whose origin is at the 0-0 band position. Spectrum in the 50 ∼ 2200 cm-1
region magnified by 15 is shown as an inset to demonstrate the quality of
the MATI spectrum obtained in this work.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
198
Table 4.20 Vibrational frequencies (in cm-1) and their assignments for
benzonitrile cation in the ground electronic state ( X~ 2B1).
Calculatedc This work Modea Symm. ZEKEb
Frequency Intensityd Frequency Intensityd
Fundamentals 1 a1 736 0.213 737 0.098 2 a1 3143 0.0001 3 b2 1382 0 1386 0.011 4 b1 607 0 606 0.022 5 b1 1036 994 0 1034 0.047 6a a1 447 445 0.295 438 0.376 6b b2 559 483 0 7a a1 3118 2×10-7 7b b2 3141 0 8a a1 1603 0.143 1612 0.032 8b b2 1235 0 9a a1 1172 0.098 1168 0.072 9b b2 1110 0 1111 0.025 10a a2 854 775 0 771e 0.017 10b b1 110 111 0 111 0.016 11 b1 784 0 771e 0.019 12 a1 971 0.036 968 0.039 13 a1 1215 0.001 1223 0.024 14 b2 1344 0 1353 0.020 15 b2 144 148 0 143 0.011 16a a2 332 335 0 336 0.011 16b b1 405 413 0 412 0.022 17a a2 1005 983 0 984 0.011 17b b1 920 947 0 922f 0.025 18a a1 962 0.081 959 0.047 18b b2 1054 0 19a a1 1425 0.001 1413g 0.017
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
199
19b b2 1475 0 1470h 0.011 20a a1 3132 1×10-5 20b b2 3131 0 βCN b2 537 548 0 536 0.015 vCN a1 2124 0.055 2136 0.037 γCN b1 298 0 291 0.018
Overtones and Combinations 10b2 a1 219 222 0.005 223 0.015 10b2151 b2 370 0 361 0.037 6a110b2 a1 667 0.002 662 0.011 4110b1 a1 718 0.001 712 0.022 16b2 a1 826 0.017 6a2 a1 890 0.045
854 0.091
6b2 a1 966 0.009 922f 0.023 42 a1 1214 0.0004 1212 0.015 6a1121 a1 1416 0.012 1413g 0.017 12 a1 1472 0.020 1470h 0.011 6a1131 a1 1660 0.0003 1647 0.008 1118a1 a1 1698 0.017 11121 a1 1707 0.006
1724 0.034
119b1 b2 1846 0 1841 0.011 119a1 a1 1908 0.018 1903 0.013 13 a1 2208 0.001 2205 0.036
a Wilson notation. b Ref. 55. c BP86/6-311++G (2df, 2pd) level. d Normalized to the intensity of the 0-0 band. Intensity of the peaks above 1700 cm-1 is not accurate due to very weak VUV power. e A composite of 10a1 and 111. f A composite of 17b1 and 6b2. g A composite of 19a1 and 6a1121. h A composite of 19b1 and 12.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
200
mentioned that Kimura and coworkers did not identify the a1-type fundamentals,
except for 6a1, in their 1+1' ZEKE work.54
Strong fundamental of the 6a mode suggests that its overtones and
combinations, especially with other prominent a1-type modes 1, 8a, 9a, 12, 13,
18a, and vCN, would appear distinctly in the one-photon MATI spectrum. 6a2
and 6a3 overtones are expected to appear at somewhat lower than 876 and 1314
cm-1. Their calculated Franck-Condon factors are 0.045 and 0.005, respectively.
They are not easy to identify, however, because of the presence of strong
spectral features nearly. The strong peak at 854 cm-1 can not be matched with the
calculated frequency of any fundamental. It may be a composite consisting of
6a2 and 16b2. I do not attempt to identify 6a3 because of its small Franck-Condon
factor. a1-type combinations involving 6a1, namely 6a1121 and 6a1131, are
identified at 1413 and 1647 cm-1, respectively. The mode 1 which has substantial
fundamental intensity also shows an overtone 12 at 1470 cm-1. The a1-type
overtones and combinations involving nontotally symmetric modes are also
observed even though with weaker intensities than the a1 fundamentals. These
are 10b2, 6a110b2, 4110b1, and 42 at 223, 662, 712, and 1212 cm-1, respectively.
Fundamentals of some nontotally symmetric modes also appear probably
through vibronic mechanism. The most noticeable among these are the
fundamental, overtone, and combinations of the 10b mode, namely 10b1, 10b2,
10b2151, 6a110b2, and 4110b1. 10b1 appeared very prominently in the 1+1' ZEKE
spectrum reported by Kimura and coworkers.54 It is not as prominent in the
present one-photon MATI spectrum. The fact that the initial electronic states
involved in the transition moment integral are different in the two processes
must be responsible for the above difference. The remaining distinct feature in
the MATI spectrum is the peak at 1034 cm-1. Comparing with the frequencies
calculated at the three DFT levels, it seems to be logical to assign this to 51, in
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
201
agreement with the assignment by Kimura and coworkers. Even though
6b1βCN1 may be an alternative based on the frequency alone, very small Franck-
Condon factor calculated for the latter, 0.001, is incompatible with the
observation.
4.6.5 Conclusions
For a molecule with a large number of vibrational degrees of freedom such as
C6H5C≡CH+ and C6H5C≡N+, assigning its vibrational spectrum can be a
formidable job, especially when no additional information is available in the
literature. In the case of two-photon ZEKE/MATI, further information can often
be obtained through intermediate state selection. This is not the case in the one-
photon scheme, even though the fact that excitation via an intermediate is not
needed is its clear advantage in experimental terms. It was demonstrated that use
of the selection rule and calculated frequencies and Franck-Condon factors,
especially those at the DFT levels, led to nearly complete vibrational
assignments for the cations of phenylacetylene and benzonitrile. It is to be
emphasized that the frequencies obtained at the DFT levels, especially BP86,
provided nearly quantitative fit to the experimental data even though harmonic
approximations were adopted for all the vibrations. Cancellation of various
errors must have acted favorably to result in such good fits.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
202
References
1. I. Glassman, Combustion (Academic Press, San Diego, 1996).
2. G. P. Brasseur, J. J. Orlando, and G. S. Tyndall, Atmospheric Chemistry and
Global Change (Oxford University Press, New York, 1999).
3. C. R. Cowley, An Introduction to Cosmochemistry (Cambridge University
Press, Cambridge, 1995).
4. T. A. Miller and V. E. Bondybey, Molecular Ions: Spectroscopy, Structure,
and Chemistry, (North-Holland Publishing Com., New York, 1983).
5. J. P. Maier, in Kinetics of Ion-Molecule Reactions, edited by P. Ausloos
(Plenum Press, New York, 1979).
6. K. Kimura, S. Katsumata, Y. Achiba, T. Yamazaki, and S. Iwata, Handbook
of HeI Photoelectron Spectra of Fundamental Organic Molecules (Japan
Scientific Societies Press, Tokyo, 1981).
7. K. Müller-Dethlefs, M. Sander, and E.W. Schlag, Chem. Phys. Lett. 112,
291 (1984).
8. J. W. Hepburn, Chem. Soc. Rev. 25, 281 (1996).
9. E. W. Schlag, ZEKE Spectroscopy (Cambridge University Press, Cambridge,
1998).
10. L. Zhu and P. Johnson, J. Chem. Phys. 94, 5769 (1991).
11. H. Krause and H. J. Neusser, J. Chem. Phys. 97, 5923 (1992).
12. C. Y. Ng, Annu. Rev. Phys. Chem. 53, 101 (2002).
13. R. Seiler, U. Hollenstein, T. P. Softley, and F. Merkt, J. Chem. Phys. 118,
10024 (2003).
14. W. Kong, D. Rodgers, and J. W. Hepburn, J. Chem. Phys. 99, 8571 (1993).
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
203
15. A. Mank, T. Nguyen, J. D. D. Martin, and J. W. Hepburn, Phys. Rev. A 51,
R1 (1995).
16. J. W. Hepburn, J. Chem. Phys. 107, 7106 (1997).
17. C. H. Kwon, H. L. Kim, and M. S. Kim, J. Chem. Phys. 116, 10361 (2002).
18. C. H. Kwon, H. L. Kim, and M. S. Kim, J. Chem. Phys. 118, 6327 (2003).
19. C. H. Kwon, H. L. Kim, and M. S. Kim, J. Chem. Phys. 119, 215 (2003).
20. C. H. Kwon, H. L. Kim, and M. S. Kim, J. Chem. Phys. 119, 4305 (2003).
21. T. E. Sharp and H. M. Rosenstock, J. Chem. Phys. 41, 3453 (1963).
22. F. Duschinsky, Acta Physicochim. URSS 7, 551 (1937).
23. G. Lembach and B. Brutschy, Chem. Phys. Lett. 273, 421 (1997).
24. E. B. Wilson, Jr., Phys. Rev. 45, 706 (1934).
25. T. G. Wright, S. I. Panov, and T. A. Miller, J. Chem. Phys. 102, 4793
(1995).
26. K. Walter, K. Scherm, and U. Boesl, J. Phys. Chem. 95, 1188 (1991).
27. A. W. Potts, D. Edvardsson, L. Karlsson, D. M. P. Holland, M. A.
MacDonald, M. A. Hayes, R. Maripuu, K. Siegbahn, and W. von Niessen,
Chem. Phys. 254, 385 (2000).
28. D. M. P. Holland, D. Edvardsson, L. Karlsson, R. Maripuu, K. Siegbahn, A.
W. Potts, and W. von Niessen, Chem. Phys. 252, 257 (2000).
29. D. M. P. Holland, D. Edvardsson, L. Karlsson, R. Maripuu, K. Siegbahn, A.
W. Potts, and W. von Niessen, Chem. Phys. 253, 133 (2000).
30. T. Baer, B. P. Tsai, D. Smith, and P. T. Murray, J. Chem. Phys. 64, 2460
(1976).
31. G. Lembach and B. Brutschy, J. Phys. Chem. 100, 19758 (1996).
32. H. Shinohara, S. Sato, and K. Kimura, J. Phys. Chem. A 101, 6736 (1997).
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
204
33. W. G. Scherzer, H. L. Selzle, and E. W. Schlag, Phys. Rev. Lett. 72, 1435
(1994).
34. G. Varsanyi, Assignments for vibrational spectra of seven hundred benzene
derivatives (Adams Higer, London, 1974).
35. R. Anand, J. D. Hofstein, J. E. LeClaire, and P. M. Johnson, J. Phys. Chem.
A 103, 8927 (1999).
36. X. Ripoche, I. Dimicoli, J. LeCalve, F. Piuzzi, and R. Botter, Chem. Phys.
124, 305 (1988).
37. M. J. Frisch, G. W. Trucks, H. B. Schlegel et al., GAUSSIAN 98, Revision
A.6 Gaussian, Inc., Pittsburgh, Pennsylvania, 1998.
38. Y. Y. Yoon, C. H. Kwon, J. C. Choe, and M. S. Kim, J. Chem. Phys. 117,
2538 (2002).
39. D. Rieger, G. Reiser, K. Müller-Dethlefs, and E. W. Schlag, J. Phys. Chem.
96, 12 (1992).
40. G. Reiser, D. Rieger, T. G. Wright, K. Müller-Dethlefs, and E. W. Schlag, J.
Phys. Chem. 97, 4335 (1993).
41. G. Lembach and B. Brutschy, J. Phys. Chem. A 102, 6068 (1998).
42. M. Fujii, T. Kakinuma, N. Mikami, and M. Ito, Chem. Phys. Lett. 127, 297
(1986).
43. A. Domenicano, G. Schultz, and I. Hargittai, J. Mol. Struct. 78, 97 (1982).
44. E. J. H. Van Schaick, H. J. Geise, F. C. Mijlhoff, and G. Renes, J. Mol.
Struct. 16, 389 (1973).
45. G. J. Den Otter, J. Gerritsen, and C. MacLean, J. Mol. Struct. 16, 379
(1973).
46. E. Sekreta, K. S. Viswanathan, and J. P. Reilly, J. Chem. Phys. 90, 5349
(1989).
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
205
47. S. T. Park, S. K. Kim, and M. S. Kim, J. Chem. Phys. 114, 5568 (2001).
48. N. Knoblauch, A. Strobel, I. Fischer, and V. E. Bondybey, J. Chem. Phys.
103, 5417 (1995).
49. D. G. Streets and G. P. Ceasar, Mol. Phys. 26, 1037 (1973).
50. R. Gilbert and C. Sandorfy, Chem. Phys. Lett. 27, 457 (1974).
51. M. Lee and M. S. Kim, J. Chem. Phys. 119, 5085 (2003).
52. J. M. Dyke, H. Ozeki, M. Takahashi, M. C. R. Cockett, and K. Kimura, J.
Chem. Phys. 97, 8926 (1992).
53. D. L. Lichtenberger, S. K. Renshaw, and R. M. Bullock, J. Am. Chem. Soc.
115, 3276 (1993).
54. M. Araki, S. Sato, and K. Kimura, J. Phys. Chem. 100, 10542 (1996).
55. K. Watanabe, T. Nakayama, and J. Mottl, J. Quant. Spectry. Radiative
Transfer 2, 369 (1962).
56. J. Neugebauer and B. A. Hess, J. Chem. Phys. 118, 7215 (2003).
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
206
Chapter 5
The Jahn-Teller Effect in Benzenoid Cations
The Jahn-Teller effect,1 which distorts the symmetry of a nonlinear molecule
in a degenerate electronic state through vibronic interaction, has been a subject
of great interest in spectroscopy and chemistry. Polyatomic cations have been
actively studied in this regard because removal of an electron from a degenerate
orbital in a neutral results in the cation in the orbitally degenerate electronic state.
In fact, the cations of benzene and hexafluorobenzene have been prototypes in
the study of the Jahn-Teller effect.
5.1 General Descriptions
Acquisition of a vibrationally resolved spectrum is a prerequisite for the
Jahn-Teller investigation. Even though the photoelectron spectroscopy may be
used for the Jahn-Teller study in the cations, detailed vibrational data are not
often available because of its poor resolution. More often than not, optical
spectroscopic techniques are not useful either because most of the polyatomic
cations in excited electronic states relax nonradiatively very fast. In the case of
the benzene cation, even though the first excited state ( B~ 2E2g) is very long-lived,
the vibrational spectra in the first excited and the ground ( X~ 2E1g) states can not
still be recorded by optical spectroscopy because the B~ ↔ X~ transition is
optically forbidden. Recently developed zero kinetic energy (ZEKE)
photoelectron spectroscopy, mass-analyzed threshold ionization (MATI)
spectroscopy, and photo-induced Rydberg ionization spectroscopy have been
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
207
useful to obtain vibrational spectra in the X~ and B~ states of C6H6+. In the
case of the hexafluorobenzene cation (C6F6+), B~ 2A2u ↔ X~ 2E1g is allowed and
emission is observed. Early investigation on the Jahn-Teller effect in the ground
state of C6F6+ reported by Leach and coworkers2 utilized this emission spectrum
measured in the discharge medium. Subsequently, spectra with much better
quality were reported by Miller and coworkers,3 which were obtained by
measuring dispersed fluorescence induced by laser for C6F6+ trapped in inert gas
matrices. Similar spectrum for the gas phase cation was also reported. However,
its quality was rather poor compared to the matrix spectra.
Early theoretical developments to elucidate the Jahn-Teller effect can be
accessed through excellent reviews. One of the recent efforts was to evaluate the
spectroscopic Jahn-Teller parameters from topographical features of the
potential energy surface obtained by quantum chemical calculations.4,5 Then, the
energies of various vibronic states calculated with these parameters were utilized
for spectral assignment. Since the Jahn-Teller effect lifts the electronic
degeneracy in the ground states of C6H6+ and C6F6
+, its proper treatment requires
the use of a multiconfiguration treatment. Barckholtz and Miller4 suggested to
calculate the average of the two Jahn-Teller surfaces by using a complete active
space self-consistent field (CASSCF) wavefunction. Various difficulties are
involved in this approach such as the inaccuracy of energy obtained at moderate
CASSCF levels and difficulty in evaluating reliable normal mode eigenvectors.
Johnson5 suggested to obtain the topographical parameters through the single
configuration density functional theory (DFT) calculation and to use the
eigenvectors at the global energy minimum. A decent agreement with the
experimental results for C6H6+ in the X~ 2E1g state was reported. Influence of the
spin-orbit coupling was ignored in the two approaches. The quantum mechanical
method to calculate the vibronic energy levels in the presence of the Jahn-Teller
effect and the spin-orbit coupling has been developed also and a software
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
208
package (SOCJT) is available from Miller’s laboratory.6
5.2 Computation
The Jahn-Teller effect distorts the equilibrium geometry of C6F6+ (or C6H6
+)
in the ground electronic state ( X~ 2E1g) from D6h to D2h symmetry and splits the
electronic state into B2g and B3g. The optimized geometries and energies in these
states were calculated at the B3LYP density functional theory level using the
GAUSSIAN 98 suite of programs. Size of the basis set was systematically
increased until the basis set dependence became insignificant. All the results
reported in this paper were obtained with the 6-311++G (2df) basis set. The
vibrational frequencies and eigenvectors were obtained at the global minimum
also. To follow the Johnson’s method,5 the geometry and energy of the
undistorted structure are needed, which correspond to those of the conical
intersection, or D6h cusp, in the Jahn-Teller potential energy surface. Its
geometry was optimized at the B3LYP/6-311++G (2df) level enforcing the D6h
symmetry.
5.2.1 The Jahn-Teller Potential Energy Surfaces and Coupling Constants
The linear Jahn-Teller effect lifts the electronic degeneracy and results in the
potential energy surface (PES) looking like a ‘Mexican hat’ along the normal
coordinate involved. A circularly symmetric moat may describe the shape of the
PES near the equilibrium geometry. The depth of the moat along the ith mode,
εi(1), is the stabilization energy due to the linear Jahn-Teller effect by this
vibration and is related to the linear coupling constant, Di, as follows.
εi(1) = Diωe,i (5.1)
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
209
Here ωe,i is the harmonic frequency of this mode. The quadratic Jahn-Teller
effect breaks the circular symmetry of the moat, resulting in a maximum and a
minimum in the moat. A half of the energy difference between the two
corresponds to the additional stabilization energy due to the quadratic effect, εi(2),
and is related to the quadratic coupling constant, Ki, as follows.
εi(2) = Diωe,i Ki (5.2)
The total Jahn-Teller stabilization energy, εT, can be evaluated as the energy
difference between the D6h cusp (2E1g) and the global minimum (2B3g).
εT = E (2E1g) - E (2B3g) (5.3)
It is not possible to apportion this into each Jahn-Teller active mode because the
geometry at the global minimum does not necessarily correspond to
simultaneous minima along all the Jahn-Teller active modes. The usual approach
to evaluate εi(1) and εi
(2) is to start from the undistorted geometry, the energy
minimum in D6h or the Jahn-Teller cusp, and calculate the potential energy along
each Jahn-Teller active normal coordinate. Following Johnson,5 the eigenvector
of the ag component of the Jahn-Teller active e2g mode obtained at the global
minimum (2B3g) was used in the calculation. Energies at twenty points each in
the positive and negative directions along the eigenvector were calculated by
B3LYP/6-311++G(2df) and fit quadratically such that the two potential energy
curves meet at the cusp. εi(1) and εi
(2) were evaluated from the energies at the
cusp and two minima. Sum of the stabilization energy along each Jahn-Teller
active mode would correspond to the total stabilization energy under the ideal
situation.
εT ≈ ∑∑==
+=+p
iiiei
p
iii KDεε
1,
1
)2()1( )1(ω)( (5.4)
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
210
5.2.2 Jahn-Teller Parameters and Vibronic Energies
The Jahn-Teller splitting is frequently described in terms of the vibronic
angular momentum quantum number j defined below.7
j = l + 1/2 Λ (5.5)
Here l is the vibrational angular momentum and Λ is the quantum number
designated to differentiate the two components of the degenerate electronic state.
Four e2g normal modes in X~ 2E1g, namely v15 ~ v18 (to be explained), are
linear Jahn-Teller active, which results in the splitting of each singly excited
level into j = ± 1/2 and ± 3/2 states. The quadratic effect further mixes j = 3/2
and -3/2 states and lifts the degeneracy. It is usual to use the ii l,,υΛ bases.
Then, the matrix elements for the Jahn-Teller Hamiltonian are given in terms of
ωe,i, Di, and Ki. Even when mode-mode coupling is not incorporated explicitly,
the Jahn-Teller active modes are coupled through the zero-point level, which
necessitates a simultaneous multimode calculation. In the multimode case, l
represents the total vibrational angular momentum ∑i
il . The multimode
package available from Miller’s laboratory, SOCJT,6 was used in this work.
Spin-orbit coupling was ignored.
Initially, single mode calculation was performed for each linear Jahn-Teller
active mode to get rough estimates of its vibronic energy levels. Then, two-mode
calculation was done for two linear Jahn-Teller active modes, v17 and v18, which
determine the spectral pattern in the low energy region. Finally, all the four
linear Jahn-Teller modes were included to get the complete vibronic levels. The
maximum j was set at 9/2 and the maximum vibrational quantum number used in
the four-mode calculation were 3, 3, 5, and 8 for v15, v16, v17, and v18,
respectively. The number of the bases for each j was 37841 in the four-mode
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
211
calculation, leading to 37841×37841 Hamiltonian matrix. The multimode results
were compared with the experimental data for the final assignments. Then, the
assigned vibronic peak positions were fed into SOCJT to evaluate the Jahn-
Teller parameters for the linearly active modes through nonlinear regression
under the same multimode condition as above.
5.3 C6H6+ and C6D6
+ in the X~ State
Benzene cation has been the focus of intensive research effort, both
experimental8-17 and theoretical,18-25 over the years. Interest in this system arises
mainly from the fact that it is a prototype molecular system displaying Jahn-
Teller distortion because its ground electronic state, X~ 2E1g, is orbitally
degenerate.
Unlike the symmetrically substituted halobenzene cations such as 1,3,5-
C6H3F3+ and C6F6
+, the benzene cation in the gas phase does not have sufficient
quantum yield of fluorescence.26-29 Hence, emission spectroscopy and laser-
induced fluorescence which were used to study various benzene derivative
cations could not be performed for the benzene cation. Even though
photoelectron spectroscopy was used to study the benzene cation, the technique
is not generally adequate to resolve vibrational structures embedded in the
spectrum.8,10 By improving spectral resolution with the use of one color ‘1+1’
laser photoelectron spectroscopy, however, Reilly and coworkers1 could observe
the vibrational splitting of a Jahn-Teller active mode ν18 (Mulliken notation,30 to
be used in this paper) in C6H6+ and C6D6
+ and suggested that the ground
electronic state of the benzene cation is 2E1g belonging to the D6h point group. At
nearly the same time, Iwasaki and coworkers31 performed the electron spin
resonance study of the benzene cation trapped in 4.2 K freon matrix. Evidence
was observed that the orbital degeneracy of the X~ 2E1g state was lifted due to
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
212
static distortion for the trapped benzene cation and the unpaired electron
occupied the b2g orbital with D2h symmetry. Even though the experimental data
were not sufficient, a number of theoretical calculations were performed.18-20
Stabilization energy and linear and quadratic Jahn-Teller coupling constants
were obtained by quantum chemical calculations at various levels to investigate
the extent of Jahn-Teller distortion to the equilibrium geometry. Even though
higher resolution spectral data have become available recently,11-15 theoretical
endeavor is still continuing because analysis of the complicated spectral data is a
formidable task.25
Zero kinetic energy (ZEKE) photoelectron spectroscopy and mass-analyzed
threshold ionization (MATI) spectroscopy which detects ions rather than
electrons in ZEKE are useful techniques to study structure and dynamics of
molecular cations. Considering the theoretical importance of the benzene cation,
it is not surprising that this system has been the focus of many ZEKE and MATI
investigations. Müller-Dethlefs and coworkers12 obtained rotationally resolved
photoelectron spectrum of benzene by ZEKE and concluded that the benzene
cation in the ground electronic state has the D6h symmetry. The vibrational peaks
arising from the Jahn-Teller active mode 18 (e2g) were identified and analyzed.
D6h symmetry was confirmed at a time scale of pseudorotation, even though the
structure in the ground state would be fluxional and dynamically coupled to the
two D2h structures.13 Krause and Neusser14 recorded MATI spectra and found
evidence for the quadratic Jahn-Teller effect, namely splitting of j = ± 3/2
vibronic states of the mode 18. Johnson and coworkers15 recently measured
MATI spectra of C6H6 and C6D6 and analyzed them considering the linear Jahn-
Teller effect in the mode 18 and the quadratic effect in the mode 20 (e2u). In
addition to ZEKE and MATI, infrared (IR) absorption spectroscopy was also
attempted by measuring IR-induced photodissociation of van der Waals
complexes between the benzene cation and rare gas atoms.32,33 The observed
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
213
spectra appeared very similar to the IR spectra expected for the bare benzene
cations, with little deviation in peak positions (~1%), according to Meijer and
coworkers. Most of these spectral data have been utilized in a recent theoretical
investigation of the Jahn-Teller effect in the benzene cation in the ground
electronic state by Applegate and Miller.25
In all the ZEKE and MATI studies of the ground state benzene cation made
so far, two-color ‘1+1’ scheme was used. The first photon in this scheme
prepares neutral benzene in the 181 vibrational state of the excited electronic
state A~ 1B2u, which is further excited to a high Rydberg state by the second
photon. It is widely acknowledged that capability to select a particular
vibrational state of the intermediate electronic state in the two-photon scheme is
helpful to assign the ZEKE or MATI spectra. A general drawback of this scheme
is the difficulty to find an appropriate intermediate state which is accessible with
a commercial dye laser output (> 200 nm) and displays well-resolved peaks upon
transition from the ground state. In the case of benzene, ν18 is the vibrational
mode which shows serious Jahn-Teller distortion and splitting in the ground
state of the cation. ZEKE and MATI intermediated by the 181 vibrational state of
A~ 1B2u result in various overtone, combination, and difference excitations
involving the ν18 mode which is Jahn-Teller active. The resulting spectra are
very complicated and hence difficult to assign. This is one of the reasons why
the vibrational assignment for and the elucidation of the Jahn-Teller effect in the
benzene cation are still an outstanding problem even after tremendous efforts
made over the years.
5.3.1 Ionization Energies
One-photon MATI spectra of C6H6 and C6D6 are shown in Figs. 5.1 and 5.2.
The intense peaks appearing at the lowest photon energy, namely at around
74545 and 74568 cm-1 in Figs. 5.1 and 5.2, respectively, correspond to the 0-0
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
214
bands. The position of the 0-0 band in a one- photon MATI spectrum is
equivalent to the ionization energy of the molecule. However, the ionization
energy thus measured is usually a little smaller than the correct value because
the molecules in ZEKE states at several cm-1 below the ionization threshold can
also be ionized when a high PFI field is applied. To correct this effect, the 0-0
band positions were measured using various PFI fields and the accurate
ionization energy was estimated by extrapolation to the zero-field limit. The
spoil field was not used in such measurements.
The ionization energies to the ground electronic states of C6H6+ and C6D6
+
measured from the one-photon MATI spectra in this work are listed in Table 5.1
together with the previous measurements. Ionization energies of C6H6 and C6D6
obtained in this work tend to be slightly lower than the previous data obtained by
ZEKE9,35 and MATI11,15 techniques using the two-photon scheme even though
the differences, ~0.0005 eV, are not serious considering the random errors in the
measurement. Presence of tiny stray field in the apparatus may be responsible
for such differences.
5.3.2 Jahn-Teller Effect and Vibronic Splitting
Benzene cation in the ground electronic state X~ 2E1g with D6h symmetry
possesses a1g (1, 2), a2g (3), a2u (4), b1u (5, 6), b2g (7, 8), b2u (9, 10), e1g (11), e1u
(12, 13, 14), e2g (15, 16, 17, 18), and e2u (19, 20) modes. Numbers in the
parentheses indicate the mode number designated by Mulliken notation. Among
these, e2g modes, four in total, are active under linear and quadratic Jahn-Teller
coupling while e1g, e1u, and e2u modes are active under quadratic coupling only.
With a single vibrational excitation in the X~ 2E1g state, an e2 mode splits into b1,
b2, and e1 vibronic species while an e1 mode into a1, a2, and e2. The linear Jahn-
Teller coupling in a singly excited e2g mode results in splitting into j = ± 1/2 and
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
215
Table 5.1 Ionization energies (IE) to the ground states of C6H6+ and C6D6
+, in eV.
IE ( X~ ) Ref.
C6H6+ 9.2432±0.0006 This work
9.24365±0.00005 9
9.24371±0.00006 11
9.24381±0.0001 15
9.243841±0.000006 35
C6D6
+ 9.2466±0.0006 This work
9.24732±0.0001 15
9.247181±0.000006 35
± 3/2 states. The quadratic effect further splits j = 3/2 and –3/2 states. It is
important to note that neither the vibrational quantum number v nor l is a good
quantum number while j is a good quantum number when only the linear Jahn-
Teller effect is considered. Also to be noted is that the Jahn-Teller active modes
are coupled through the zero-point level, which make complete analysis
extremely difficult.
5.3.3 Vibrational Analysis
Even though benzene was used as a prototype molecule in the development
of ZEKE, and subsequently MATI, comprehensive analysis of the cation
vibration has not been attempted until recently. In one of the early MATI studies
of benzene, Krause and Neusser11 reported tentative assignment of MATI peaks,
even though only up to 1325 cm-1 in vibrational energy. The most recent IR
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
���
Fig. 5.1 One-photon MATI spectrum of C6H6 recorded by monitoring
C6H6+ in the ground electronic state. The x-scale at the top of the figure
corresponds to the vibrational frequency scale for the cation. Its origin is at
the 0-0 band position. Spectrum in the 100~2100 cm-1 region magnified by
30 is shown as an inset to demonstrate the quality of the MATI spectrum
obtained in this work. Relative intensity of the peak marked by asterisk (*)
changed with the beam expansion condition.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
217
Table 5.2 Vibrational frequencies (in cm-1) and assignments for C6H6+ in the
ground electronic state ( X~ 2E1g).
Modea Symm. Neutralb PESc MATId Millere This work 1[2] a1g 3073 3082 2[1] a1g 993 976 967 969 967 3[3] a2g 1350 1393 4[11] a2u 673 659 660 5[13] b1u 3057 3022 6[12] b1u 1010 883 878 7[5] b2g 990 934 8[4] b2g 707 415 416 418 420 9[14] b2u 1309 1351 1357 10[15] b2u 1146 1180 1183f
11[10] a2g(1g) 983 986g
11[10] e2g 846 843 843
11[10] a1g(2g) 724 724
12[20] e1u 3064 3109 13[19] e1u 1482 1420 1420 14[18] e1u 1037 948
15(1/2)[7] e1g 3056 3062 15(±3/2) b2g(1g) 2848
16(1/2)[8] e1g 1599 1573 1648 16(±3/2) b2g(1g) 1563 16(±3/2) b1g(2g) 1522 1519
17(1/2)[9] e1g 1178 1234 1228 1257 1255 17(±3/2) b2g(1g) 1183f
17(±3/2) b1g(2g) 1157 18(1/2)[6] e1g 606 673 674 677 18(± 3/2) b2g(1g) 363 367 363 18(± 3/2) b1g(2g) 355 343 347 350
19[17] e2u 967 994 986g
20[16] b2u(1u) 328 327
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
218
20[16] e1u 398 319 303 306 305 20[16] b1u(2u) 285 289 296
202 586/608 596 (1/2,3) 763 761 (1/2,4) 1073 1072 191201 1283 21181 1328 1322 101201 1475
21181201 1613 21111 1786 112 1825
111171 1881 22 1934
161181 1999 161181 2007
21111181 2046 21111201 2085
22201 2238 111161 2365 217181 2445 21112 2508 21161 2524 21161 2621 31171 2639
23 2903 a Vibrational modes in Mulliken notation (Wilson notation in square bracket). b From ref. 34. c From ref. 8. d Peak assignments for the two-photon MATI spectrum reported in ref. 4. Similar
assignments were also reported in ref. 15. e Assignment of the two-photon ZEKE13 and IR-induced photodissociation33
spectra reported by Applegate and Miller in ref. 25. f This peak may be assigned alternatively to 101 or 171(±3/2). g This peak may be assigned alternatively to 111 or 191.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
219
photodissociation study of benzene-Ne, Ar reported by Meijer and coworkers33
showed vibrational peaks up to 1470 cm-1 and a peak at 3097 cm-1 but only the
peaks below 1000 cm-1 were assigned. Applegate and Miller25 used all these data
and some unpublished results and reported assignment up to 1815 cm-1 by
comparing with their own calculated results. It is to be mentioned that there are
far more peaks in the original ZEKE spectrum, for example in the 1000 ~ 1500
cm-1 region, than assigned by these investigators. Namely, not all the ZEKE
peaks were assigned. The number of peaks in the present one-photon MATI
spectra is much less than that in ZEKE as has been mentioned earlier. Complete
vibrational assignment, even though tentative, up to the C-H stretching region
(~3000 cm-1) is attempted in this work. Assignments by Applegate and Miller
have been adhered to when available. Also taken into account were the positions
of some Jahn-Teller components predicted by these investigators through
calculation. For the peaks without any previous information, assignments have
been attempted by referring to the mode frequencies in neutral benzene. In
particular, the ratio of the frequency of a particular mode in C6D6 to that in C6H6,
namely the isotope ratio, was calculated for the neutral and cation. Similar
isotope ratios in the neutral and cation were taken as an evidence for successful
assignment. This is based on the assumption that the force fields in the benzene
neutral and cation are rather similar. Finally, it is to be mentioned that effort was
made to identify the hot bands, especially in the region of low vibrational
frequency, by changing the beam condition drastically. Relative intensity of the
very weak peak at 463 cm-1 marked by an asterisk in the inset of Fig. 5.1
changed with the beam condition. Assignments for the peaks in the one-photon
MATI spectra, Figs. 5.1 and 5.2, are listed in Tables 5.2 and 5.3, respectively,
together with the previous spectral data.
In the one-photon MATI spectrum of C6H6, Fig. 5.1, the most intense peak
other than the 0-0 peak appears at 967 cm-1, the v2 (a1g) fundamental according
to previous assignments.8,11,13 Two peaks at 1934 and 2903 cm-1 can be assigned
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
220
readily to its overtones, 22 and 23, respectively. Corresponding peaks are
observed at 926, 1849, and 2772 cm-1 in the MATI spectrum of C6D6, Fig. 5.2.
The isotope ratio of this mode, 926/967=0.96, is in excellent agreement with
0.95 for the neutral.34 Appearance of the strong 2n progression suggests that the
geometrical change upon ionization of neutral benzene occurs mostly along this
normal mode. This is a part of the proof for the preservation of D6h symmetry in
the cation. Fundamentals of the other a1g mode, v1, are observed at 3082 and
2350 cm-1 for C6H6 and C6D6, respectively, not much different from 3073 and
2303 cm-1 observed in the neutral.
The other nondegenerate gerade modes are v3 (a2g), v7 (b2g), and v8 (b2g).
Among these, v8 can be readily identified by referring to the previous
assignments, namely peaks at 420 and 343 cm-1 for C6H6 and C6D6, respectively.
Applegate and Miller identified the peaks at 934 and 754 cm-1 in the ZEKE
spectra of C6H6 and C6D6 as the v7 fundamentals. A weak peak was found at 750
cm-1 for C6D6 accordingly, even though the corresponding peak for C6H6 could
not be seen due to the presence of an intense peak nearby. The v3 fundamentals
have not been assigned before. We assign them to the peaks at 1393 and 1061
cm-1 in the one-photon spectra of C6H6 and C6D6, respectively, based on their
proximity to those in the neutrals, 1350 and 1059 cm-1, and the adequate isotope
ratio.
The remaining gerade modes are v11 (e1g) which shows quadratic Jahn-Teller
effect and v15 ~ v18 (e2g) for which linear coupling is present and mode mixing
can be important also. Applegate and Miller25 assigned the peaks at 724, 843,
and 983 cm-1 in the ZEKE spectrum of C6H6 to v11. The first two peaks are found
at the same positions in the present one-photon MATI spectrum. The third may
correspond to a shoulder peak at ~986 cm-1. Even though Applegate and Miller
assigned 562 and 673 cm-1 to v11 of C6D6+, it is difficult to find them in Fig. 5.2
due to the presence of broad bands. They may be buried in the asymmetric tails
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
���
Fig. 5.2 One-photon MATI spectrum of C6D6 recorded by monitoring
C6D6+ in the ground electronic state. The x-scale at the top of the figure
corresponds to the vibrational frequency scale for the cation. Its origin is at
the 0-0 band position. Spectrum in the 100~2100 cm-1 region magnified by
40 is shown as an inset to demonstrate the quality of the MATI spectrum
obtained in this work.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
222
Table 5.3 Vibrational frequencies (in cm-1) and assignments for C6D6+ in the
ground electronic state ( X~ 2E1g).
Modea Symm. Neutralb PESc MATId Millere This work 1[2] a1g 2303 2350 2[1] a1g 945 928 926 926 3[3] a2g 1059 1061 4[11] a2u 496 488 488 491 5[13] b1u 2285 6[12] b1u 970 877 840f
7[5] b2g 829 754 750 8[4] b2g 599 351 344 343 9[14] b2u 1282 1327 10[15] b2u 824 833 840f
11[10] a2g(1g) 11[10] e2g 660 673 658(?) 11[10] a1g(2g) 562 550(?) 12[20] e1u 2288 2331 13[19] e1u 1333 1229 1230 14[18] e1u 814 773 766
15(1/2)[7] e1g 2274 2292 15(±3/2) b2g(1g) 2168
16(1/2)[8] e1g 1558 1552 16(±3/2) b2g(1g) 1453 16(±3/2) b1g(2g) 1409
17(1/2)[9] e1g 869 877 17(±3/2) b2g(1g) 821
17(±3/2) b1g(2g) 802g
18(1/2)[6] e1g 579 637 634 636 18(± 3/2) b2g(1g) 356 357 18(± 3/2) b1g(2g) 343 335 338 334
19[17] e2u 787 810 802g
20[16] b2u(1u) 285 289 289 20[16] e1u 345 278 262 265 263 20[16] b1u(2u) 245 249 252
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
223
202 509 202 550(?) 82 681
4181 820 821 111201 975
191201, (1/2,5) 1058 1061 (3/2,5) 1119 1117 (1/2,6) 1154
81171, (1/2,7) 1164 1161 21201 1189
2181, (1/2,8) 1266 1267 21202 1431 21111 1581 2171 1663
21141 1691 21171 1708 21171 1744 21171 1795
22 1849 21171201 2006 21171201 2033
22201 2109 2281 2196
21161 2468 22171 2653 22171 2717
23 2772 a Vibrational modes in Mulliken notation (Wilson notation in square bracket). b From ref. 34. c From ref. 8. d Peak assignments for the two-photon MATI spectrum reported in ref. 15. e Assignment of the two-photon ZEKE13 and IR-induced photodissociation33
spectra reported by Applegate and Miller in ref. 15. f This peak may be assigned alternatively to 61 or 101. g This peak may be assigned alternatively to 171(±3/2) or 191.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
224
of these bands. In all the previous two-photon ZEKE and MATI investigations
of C6H6, vibronic peaks correlated to v18 (e2g, v6 in Wilson notation35) appeared
prominently at 347, 367, and 677 cm-1 and were the focus of the study of the
linear and quadratic Jahn-Teller effect in C6H6+. In the present one-photon
MATI spectrum, no peak can be found near 677 cm-1 which can be assigned to
181 (1/2). Peaks are observed at 350 and 363 cm-1, even though weakly, which
can be attributed to 181 (± 3/2). Dramatic difference in the intensities of the 181
vibronic bands between the one- and two-photon spectra is not unexpected
because the two-photon scheme involves excitation of this mode in the neutral
intermediate state. More surprising is the fact that observations made in the
photoelectron spectrum16 are opposite to the present results. Namely, the j = 1/2
peak appears prominently in the photoelectron spectrum while the j = ± 3/2
peaks are missing. We do not have an explanation for these results at the
moment. We will just mention that this spectral region in the one-photon MATI
spectrum has been reproduced and checked several times to assure that the
results are not due to some artifacts. In addition to the above peaks, Applegate
and Miller assigned ten e2g vibronic peaks up to 1522 cm-1. Among these, six
were correlated to prominent peaks in the two-photon ZEKE spectrum, namely
at 763, 1073, 1245, 1257, 1408, and 1435 cm-1. Assignment of each peak to a
particular normal mode was not made probably because of severe mode coupling.
Only the j quantum number for each peak was given. Since the v18 mode has a
resonance frequency which is much lower than those of the other e2g modes,
some of the aboves must have main contribution from the v18 overtones and
could appear prominently in the two-photon spectrum intermediated by 181
excitation. Then, the fact that many of these are not prominent in the one-photon
spectrum is not surprising. A prominent peak at 1255 cm-1 in the one-photon
spectrum can be correlated to the peak at 1257 cm-1 in the two-photon spectrum
which was assigned to a j = 1/2 peak by Applegate and Miller. Considering the
harmonic frequencies calculated by these investigators, this peak can be assigned
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
225
to 171 (1/2). Two peaks were observed in the two-photon ZEKE of C6D6, namely
at 868 and 877 cm-1, which were assigned to 171 (1/2) and 61 by Applegate and
Miller. Considering that the peak at 877 cm-1 appears very prominently and that
v6 is an ungerade mode, assignment to 171 (1/2) would be more appropriate for
this peak. Then, the isotope ratio for this mode becomes 0.70 (=877/1255) which
is not too different from 0.74 (=869/1178) in the neutral. There are two choices
for the 171 (±3/2) pair in C6H6+, 1052 and 1072 or 1157 and 1183 cm-1 pairs.
Considering appearance of a pair at 802 and 821 cm-1 in the one-photon
spectrum of C6D6, We would prefer the second pair, 1157 and 1183 cm-1, which
result in the isotope ratio of 0.69.
Since the previous ZEKE and MATI studies of benzene were not performed
above 1500 cm-1, there was no report on the assignment of v16. Reilly and
coworkers8 suggested 161 (1/2) as the transition responsible for the peak at
around 195 meV (1570 cm-1) in the laser photoelectron spectrum. This mode has
the frequencies 1599 and 1558 cm-1 in neutral C6H6 and C6D6 with the isotope
ratio 0.97. Strong peaks at 1648 and 1552 cm-1 in the one-photon spectra of C6H6
and C6D6, respectively, with the isotope ratio 0.94 can be assigned to this
transition. Doublets appearing at 1519 and 1563 cm-1 for C6H6 and at 1409 and
1453 cm-1 for C6D6 can be similarly assigned to 161 (±3/2). It is to be mentioned
that such assignments are in agreement with the prediction made by Applegate
and Miller through theoretical calculations. The peaks at 3062 and 2292 cm-1 in
the one-photon spectra of C6H6 and C6D6, respectively, may be assigned to
151(1/2) by referring to the corresponding frequencies in the neutrals, 3056 and
2274 cm-1. Then, assignment of the peaks at 2848 and 2168 cm-1 in the one-
photon MATI spectra of C6H6 and C6D6, respectively, to 151(±3/2) completes
the assignment for the gerade modes.
Among the ten ungerade modes four vibrations are IR active under the D6h
symmetry, namely v4 belonging to a2u and v12 ~ v14 belonging to e1u.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
226
Fundamentals of these modes appear prominently in the IR-induced
photodissociation spectra33 of C6H6+-rare gas complexes. Some combination
peaks also appeared prominently in the above spectra aided by Fermi resonance.
For example, two strongest peaks at 629 and 697 cm-1 in the IR-induced
photodissociation spectrum of C6H6+-Ne were attributed to 181201 components
with the e1u symmetry which are in Fermi resonance with 41. In addition, some
prominent peaks were assigned to the fundamentals and combinations belonging
to other vibrational symmetries by invoking vibronic symmetry. It looks a little
strange to note, however, that 81141 belonging to e2u is more intense in the IR
spectrum than 131 which is vibrationally allowed (e1u). Peaks belonging to b1u
and b2u which are vibronically forbidden were also observed. In addition, some
gerade modes, 181 and 202 were also reported to have been observed. All these
seem to suggest the importance the symmetry breaking effect arising from the
attachment of a rare gas atom. Some prominent peaks in the two-photon ZEKE
spectrum13 were assigned to the fundamentals of the ungerade modes. For
example, peaks at 306, 660, and 994 cm-1 assigned to 201 (e2u), 41 (a2u), and 191
(e2u), respectively, have intensities nearly comparable to the strongest two-
photon ZEKE peaks. Whether C6H6+ has the D6h symmetry or is slightly
perturbed to D2h, the center of symmetry is maintained in the present one-photon
MATI scheme. Accordingly, the ungerade fundamentals appear very weakly in
the one-photon spectrum, the strongest among which is the weak 201
fundamental at 305 and 263 cm-1 in C6H6 and C6D6, respectively. Plausible
mechanisms for the appearance of the ungerade fundamentals have been
presented already.
Among the nondegenerate ungerade fundamentals, 61 and 91 can be assigned
to peaks at 878 and 1357 cm-1 in Fig.5.1 by referring to the previous
assignments.18 41 which appeared prominently at 660 cm-1 in the two-photon
spectrum and the IR-induced photodissociation spectrum does not appear at all,
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
227
even though a very weak feature at around 491 cm-1 in the one-photon spectrum
of C6D6 can be correlated to the peak at 488 cm-1 in the two-photon spectrum.
The peak at 1183 cm-1 which has been assigned to 171 (±3/2) above can be
alternatively assigned to 101. Among the fundamentals of three e1u modes, 131
can be assigned to the peaks at 1420 and 1230 cm-1 in the C6H6 and C6D6 spectra.
141 can be found for C6D6 only at 766 cm-1 while 121 can be assigned to the
peaks at 3109 and 2331 cm-1 in the C6H6 and C6D6 spectra, respectively. Distinct
peaks at 305 and 327 cm-1 and a shoulder peak at 296 cm-1 in the one-photon
spectrum of C6H6 can be assigned to 201 by referring to the previous
assignments. Similarly, very weak shoulder peaks at 986 and 802 cm-1 in the
one-photon spectra of C6H6 and C6D6 can be assigned to 191. It is to be
mentioned that assignments for the ungerade fundamentals are in excellent
agreement with those by Applegate and Miller.
With almost all the fundamentals assigned, an attempt has been made to
assign the remaining peaks to overtones and combinations. The prominent peaks
at 1283 and 1061 cm-1 in the one-photon spectra of C6H6 and C6D6, respectively,
can be attributed to 191201 which contains a1g species. Even though the weak
peak at 1475 cm-1 is close to the strong peak at 1470 cm-1 in the IR-induced
photodissociation spectrum, we would rather assign the former to a 101201
component belonging to a1g. Distinct peaks at 1786, 1881, 1999, 2238, 2365, and
2524 cm-1 in the C6H6 spectrum and those at 1708, 1795, and 2006 cm-1 in the
C6D6 spectrum have been assigned as combinations also. Even though these
assignments are tentative only, the data are presented here as an aid for future
theoretical study. Even after the above effort, there still remain some peaks, even
though extremely weak, which are left unassigned. These are the peaks at 1052,
2569, 2738, and 2754 cm-1 in the C6H6 spectrum and at 411, 1306, 1360, 1626,
2399, and 2623 cm-1 in the C6D6 spectrum. It is to be noted that all the peaks
below 1700 cm-1 in the one-photon MATI spectrum of C6H6 except the peak at
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
228
1052 cm-1 have been reasonably assigned in this work. In contrast, only around
60 % of the peaks appearing below 1500 cm-1 and around 30 % in the 1000 ~
1500 cm-1 region in the two-photon ZEKE spectrum were assigned.18
5.3.4 Conclusions
One-photon MATI spectra of C6H6 and C6D6 have been obtained by using
tunable and coherent VUV radiation generated by four-wave difference
frequency mixing in Kr. The spectra were simple compared to the two-photon
ZEKE and MATI spectra reported previously. Vibrational data for the neutral
benzene and the previous assignment for the cation made by utilizing two-
photon ZEKE peaks, IR-induced photodissociation peaks, and calculated results
were useful to assign the peaks in the one-photon spectra. Also useful was the
selection rule for Rydberg transition which classifies the transitions into three
groups in terms of intensity, electric dipole-allowed, vibronically allowed, and
g,u-forbidden vibronic transitions. Exceptions to this selection rule have been
found to be rare in the one-photon spectra while surprises were encountered
frequently in the two-photon ZEKE or MATI and IR-induced photodissociation
spectra. Compared to the two-photon spectra obtained via transition to the 181
vibrational state of A~ 1B2u of the neutral, the one-photon spectra were simpler
and hence easier to analyze. In particular, almost all the peaks below 1700 cm-1
in the one-photon spectrum of C6H6 could be assigned while around 70 % of the
peaks in the 1000 ~ 1500 cm-1 region of the two-photon spectrum are left
unassigned. This suggests that analysis of the one-photon spectrum, rather than
two-photon spectrum, can be the useful first step in the vibrational study of a
polyatomic cation especially when complications are involved such as those
arising from the Jahn-Teller effect.
Even though the peaks in the one-photon spectra could be assigned extremely
well by referring to the previous assignments, some changes had to be made.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
229
More fundamentals have been assigned than before, some of which being to
vibronic levels arising from Jahn-Teller splitting. Incorporation of the results in
the theoretical study is expected to lead to a better understanding of the Jahn-
Teller effect in the benzene cation.
5.4 C6H6+ and C6D6
+ in the B~ State
Considering only the three highest energy occupied orbitals, electron
configuration of the neutral benzene in the ground electronic state ( X~ 1A1g) is
(1a2u)2(3e2g)4(1e1g)4. Removing an electron from these orbitals results in the hole
states of the cation X~ 2E1g, B~ 2E2g, and C~ 2A2u. Ionization energies to the first
two states are 9.243 and 11.488 eV, respectively, according to a recent PES
study16 of C6H6. The C~ band appears broad and overlapped with the B~ band
in the photoelectron spectrum. The C~ state onset is thought to lie a few tenths
of an eV above that of the B~ state.
Benzene cation in the ground electronic state has been heavily investigated
over the years because the orbital degeneracy makes it a prototypical system for
the Jahn-Teller (JT) effect. Situation for the first excited state, B~ 2E2g, is even
more interesting because of its proximity to the C~ 2A2u state. Namely, in
addition to the usual intrastate JT interaction, the pseudo Jahn-Teller (pseudo-
JT) interaction between B~ 2E2g and C~ 2A2u is also possible.15,36,37
Another interesting feature in the benzene cation system is that no emission
is observed at all even when many excited hole states are populated by electron
ionization or photoionization. This is in contrast with the observation of strong
emission from the B~ state of various benzene derivative cations. For benzene
cation with D6h symmetry, B~ 2E2g→ X~ 2E1g transition is electric dipole-
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
230
forbidden. A general consensus has been that the same transition occurring
nonradiatively, namely via internal conversion, is very efficient such that the
radiative transition is almost completely quenched. Köppel and coworkers
carried out extensive quantum chemical and dynamical calculations on the
benzene cation system.36,37 Many conical intersections were found
computationally, including those between B~ and X~ . Since the latters are
located somewhat above the zero-point level of the B~ state, however, there is a
possibility that some of the benzene cations prepared in this state do not undergo
rapid internal conversion. In fact, we found through charge exchange and other
mass spectrometric experiments that some of the benzene cations in the B~ state
have a very long lifetime, 10 microsecond or possibly much longer. Then, lack
of emission from the B~ state indicates extremely poor probability for the B~ -
X~ radiative transition, which has been a serious obstacle to the use of
conventional spectroscopic techniques for the study of the benzene cation system.
A dipole-forbidden transition such as B~ ←X~ can occur vibronically, even
though very weakly. However, absence of emission from the B~ state means
that one needs a method, other than the measurement of emission, to confirm the
absorption of a photon. Schlag and coworkers obtained the first optical spectrum
for the B~ ← X~ transition by measuring C6H5+ generated by resonance-
enhanced multiphoton dissociation (REMPD) of the benzene cation in the
ground state.38 Resonance-enhanced multiphoton ionization was used to
selectively generate the ground state ion. Initial state selection was further
improved by Johnson and coworkers by introducing photoinduced Rydberg
ionization (PIRI) spectroscopy.15 Here neutral benzene was selectively prepared
in some vibrational states of a high Rydberg state converging to the ionic ground
state via two-photon excitation. Then ion core of the Rydberg state was further
excited by the third photon and C6H6+ formed by autoionization or its fragment
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
231
were detected. Since gerade vibrational states of the Rydberg state could be
populated efficiently, PIRI spectra recording ungerade vibrations in the ionic B~
state were obtained with high quality. This was useful to investigate pseudo-JT
interaction induced by e2u vibrations. Gerade vibrations were also recorded,
which are needed to study the intrastate JT interaction. The spectral quality was
not as good as those for the ungerade modes, however.
Complications due to multiphoton effect can be present in the above REMPD
and PIRI schemes because a high power laser is used to induce dipole-forbidden
B~ ←X~ transition. A better scheme to investigate the B~ state is to excite the
neutral to a Rydberg state converging to this state and record ZEKE or MATI
spectra, namely without going through the ionic (or ion core) ground state. With
the usual two-photon ZEKE or MATI scheme, this involves exciting an electron
in the e2g orbital to an unoccupied orbital. No spectroscopic information is
available for such neutral states. In practical terms, such states would lie high
above the ground state and not be accessible with a commercial dye laser.
5.4.1 Jahn-Teller Effect and Vibronic Splitting
If the benzene cation in the B~ 2E2g electronic state has D6h symmetry, its
vibrational modes can be classified into a1g (1, 2), a2g (3), a2u (4), b1u (5, 6), b2g
(7, 8), b2u (9, 10), e1g (11), e1u (12, 13, 14), e2g (15, 16, 17, 18), and e2u (19, 20)
modes. Numbers in the parentheses indicate the mode number designated by
Mulliken notation. For e-type degenerate electronic states in D6h, e2g modes are
active under linear and quadratic JT coupling while e1g, e1u, and e2u modes are
active under quadratic coupling only. A singly excited e2 mode in B~ 2E2g splits
into a1, a2, and e2 vibronic species while an e1mode into b1, b2, and e1 species.
An alternative description of vibronic species is to designate the vibronic angular
momentum quantum number j which is the sum of the vibrational (l) and
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
232
electronic (Λ) parts. There has been some controversy concerning the definition
of Λ. Goode et al. suggested Λ = ± 2 for the B~ 2E2g state and predicted weak
intrastate JT effect.15 Recent theoretical work reported by the same laboratory
found evidence for very strong JT effect in this state, however. We will follow
Miller6 and put Λ = ± 1. Then, j becomes
Λ+= 21lj (5.6)
Since l = ± 1 for a singly excited (v = 1) degenerate mode, j becomes ± 1/2 and ±
3/2. Linear JT interaction splits j = ± 3/2 from ±1/2 states. Quadratic JT further
splits j =3/2 and –3/2. The above is only a simplified picture because v is not a
good quantum number in the presence of JT coupling and interaction among
zero-order states with different υ cannot be neglected. Mode coupling further
complicates the situation and makes vibrational analysis extremely difficult. e2u
modes were reported to be active in pseudo-JT interaction between B~ 2E2g and
C~ 2A2u. Pseudo-JT effect will not be considered here, however, because the
peaks observed in the one-photon MATI spectra are mostly gerade fundamentals.
According to a recent quantum chemical study by Johnson,4 the calculated
linear JT coupling parameters were very large for the e2g modes v16 and v18 while
those for v15 and v17 were rather small. With a substantial JT effect along a
normal coordinate, electronic degeneracy is essentially lifted and energy
minimum moves along this coordinate, resulting in symmetry reduction to D2h.
Then, an e2g mode in B~ 2E2g splits into ag ⊕ b3g, which result in (ag ⊕ b3g) ⊗ (ag
⊕ b3g) = 2ag ⊕ 2b3g vibronic species. Namely, a1g vibronic species in D6h is
converted to ag in D2h, a2g to b3g, and e2g to ag ⊕ b3g.
5.4.2 Selection Rule
The selection rule for one-photon MATI from supersonically cooled benzene
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
233
neutral (zero-point level in X~ 1A1g) to X~ 2E1g of the cation was described in
previous section. Following the same procedure, identical selection rule is
obtained for one-photon MATI to B~ 2E2g of the cation. Results can be
summarized as follows. Vibronic transitions appearing in one-photon MATI can
be classified into three categories, electric dipole-allowed, vibronically allowed,
and g,u - forbidden vibronic transitions. A1g fundamentals can appear
prominently because the transition involved is dipole-allowed. Fundamentals of
all the other gerade modes may appear less prominently via dipole-forbidden
vibronic transitions. Even though ungerade fundamentals are not expected when
the center of symmetry is present in the cation, instrumental imperfection or high
order effect may allow the processes, even though very weakly, as was found in
one-photon MATI to X~ 2E1g. According to the previous quantum chemical
calculations,4 the geometrical change upon B~ 2E2g ← X~ 1A1g ionization is not
significant. Hence overtones and combinations are not expected to appear
prominently in one-photon MATI. This is in contrast with the PIRI spectra of
gerade states reported previously, in which the transitions started from the
201(e2u) vibrational state of X~ 2E1g. Accordingly, most of the prominent peaks in
the PIRI spectra were assigned to the overtones and combinations involving v20.
Finally, it may be helpful to study the changes in symmetry selection rule
when D6h→D2h symmetry reduction occurs. In this case, the normal modes of
interest are a1g and e2g vibrations in D6h, which generate the totally symmetric ag
vibrations in D2h. Namely, e2g vibrations may occur prominently upon symmetry
reduction in addition to a1g vibrations.
5.4.3 Vibrational Analysis
One-photon MATI spectra of C6D6 and C6H6 measured by exciting the
neutrals to the high Rydberg states converging to the ionic B~ 2E2g state are
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
234
shown in Figs. 5.4 and 5.5, respectively. Magnified spectra are shown as insets
to demonstrate the quality of the spectra. A dip at ~ 94110 cm-1 appears in both
spectra, which is due to the dip in VUV power (Fig. 5.3). Quality of these
spectra looks much better than the corresponding PIRI spectra for the gerade
states reported by Johnson and coworkers.15 The 0-0 bands appearing at around
92938 and 92664 cm-1 in Figs. 5.4 and 5.5, respectively, are the most intense
peaks in each spectrum while the corresponding peaks were relatively weak in
the PIRI spectra. X~ 2E1g and B~ 2E2g cation states are generated by removal of a
bonding e1g or antibonding e2g electron, respectively, from the neutral ground
state X~ 1A1g. Hence, the geometrical change in B~ 2E2g ← X~ 2E1g would be
larger than that in B~ 2E2g ← X~ 1A1g. In addition, the zero vibrational states in
the PIRI spectra were accessed via an excited vibrational state (201) in X~ 2E1g
such that the fundamental and overtones of v20 would be favored. The aboves
may explain the remarkable difference in the relative intensities of the zero
vibrational bands between the one-photon MATI and three-photon PIRI spectra.
5.4.3.1 Ionization Energies
Table 5.4 Ionization energies (IE) to the excited states of C6H6+ and C6D6
+, in eV.
IE ( B~ ) Ref.
C6H6+ 11.4897±0.0006 This work
11.4900±0.0001 15
11.488±0.003 16
C6D6
+ 11.5235±0.0006 This work
11.5240±0.0001 15
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
235
The position of the 0-0 band in one-photon MATI spectrum is equivalent to
the ionization energy to the ionic state which the Rydberg states accessed by
VUV are converging into. The 0-0 band position appearing in a spectrum is
usually a little smaller than the correct ionization energy because molecules in
ZEKE states several cm-1 below the threshold can also be ionized when a high
PFI field is used. To correct for this effect, the 0-0 band position was measured
using various PFI fields and the results were extrapolated to the zero field limit.
Spoil field was not used in such measurements. The ionization energies to
B~ 2E2g of C6H6+ and C6D6
+ thus obtained are listed in Table 5.4 together with the
previous PIRI15 and PES16 measurements.
5.4.3.2 Vibrational Assignment
Vibrational patterns of the one-photon MATI spectra obtained in this work
are quite different from the gerade state PIRI spectra. The main reason may be
that overtones and combinations involving v20 (e2u) are prevalent in the PIRI
spectra because transitions start from 201 in X~ 2E1g while peaks in the one-
photon MATI spectra are expected to be fundamentals mostly. The guidelines
adopted to assign the one-photon MATI peaks are as follows.
Vibrational frequencies in B~ 2E2g were calculated by time-dependent density
functional theory (TDDFT) at the B3LYP/6-311G(2d, p) level by Johnson
recently.15 For modes other than e2g, calculated frequencies appeared reasonable
and resembled the corresponding values for the neutral. Calculated frequencies
for these modes will be utilized for assignments as much as possible. Among the
e2g modes, v15 lies outside the spectral range covered in this work. Since the
calculated linear JT coupling constant for v17 was very small, the fundamental
may be searched in the vicinity of the calculated frequency. These leave the e2g
modes v16 and v18. Very large calculated linear JT coupling constants for these
modes indicate that their assignments will be a formidable job. The second
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
236
93000 93500 94000 94500
Sign
al
Photon Energy, cm-1
Fig. 5.3 Photoionization spectrum of C6H6 measured as a function of the
VUV photon energy (in cm-1).
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
���
Fig. 5.4 One-photon MATI spectrum of C6D6 recorded by monitoring
C6D6+ in the excited electronic state B
~ 2E2g. The x-scale at the top of the
figure corresponds to the vibrational frequency scale for the cation, or the
ion internal energy. Its origin is at the 0-0 band position. Spectrum in the
0~1900 cm-1 region magnified by 10 is shown as an inset to demonstrate
the quality of the MATI spectrum obtained in this work.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
238
Table 5.5 Vibrational frequencies (in cm-1) and assignments for C6D6+ in the
excited electronic state ( B~ 2E2g).
Assignmentc Observed Neutrala PIRIb
Empiricald JT fite
154 201 |3/2, 1⟩ 373 366 181(±3/2) 481 486 181(±3/2) |1/2, 2⟩
574 599 596 81 597 579 181(±1/2) |1/2, 3⟩ 729 728 171(±3/2) |3/2, 3⟩ 745 171(±3/2) |3/2, 4⟩
780 787 776 191
849 869 742 171(±1/2) |1/2, 6⟩ 934 945 21
1206 |3/2, 10⟩ 1222 171181 |3/2, 11⟩ 1293 |3/2, 12⟩
1373 181191 1443 171181 |3/2, 15⟩
1493 |1/2, 16⟩
1580 |1/2, 19⟩
a From ref. 34. b From ref. 39. c Vibrational modes in Mulliken notation. d Assignment made by referring to the calculated frequencies in ref. 4. Small
D18 assumed for v18. e Three-mode (v16, v17, v18) JT fit assuming large D16 and D18. See text for
details.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
239
guideline adopted was the selection rule. Peaks were assigned to gerade
vibrations as much as possible, except for some very weak peaks. For a very
prominent peak, assignment to an a1g or e2g mode was regarded as a higher
priority. Finally, the fact that the calculated frequency of each non-e2g mode in
B~ 2E2g is similar to the corresponding frequency in the neutral suggests that the
force fields in the cation B~ 2E2g and in the neutral X~ 1A1g are rather similar.
Hence, a pair of peaks in the one-photon MATI spectra of C6H6 and C6D6 was
chosen such that the isotope ratio (C6D6/C6H6 frequency ratio) in the cations was
similar to that in the neutral.34 The results of the vibrational assignments for
C6D6+ and C6H6
+ are listed in Tables 5.5 and 5.6, respectively.
Of the two a1g modes, v1 fundamental is outside the spectral range covered in
this work. v2 fundamental for C6D6+ can be readily identified to the strong peak
at 934 cm-1 in excellent agreement with the calculated frequency and the
frequency in the neutral. Since its isotope ratio in the neutral is 0.95 (945/993),
the corresponding fundamental in C6H6+ is expected at ~ 980 cm-1. Two weak
peaks appear in this region of Fig. 5.5, namely at 971 and 1034 cm-1. The former
is apparently a better candidate even though its intensity seems to be very weak
as an a1g fundamental.
The next strongest in the C6D6 MATI spectrum are the peak at 849 cm-1 and
the doublet at 729 and 745 cm-1. Among the gerade fundamentals, 171 (e2g) and
111 (e1g) are expected in this region. 111 (e1g) for C6D6+ is expected at 700 ~ 750
cm-1 according to the calculation while the corresponding frequencies for C6H6+
are expected at 860 ~ 960 cm-1. Having assigned the peak at 971 cm-1 in the
C6H6 spectrum to 21 already, only an extremely weak peak at 915 cm-1 remains
in this region. Considering the intensities of the above peaks, it seems to be more
appropriate to assign the peak at 849 cm-1 in the C6D6 MATI spectrum to 171
(±1/2) and the doublet at 729 and 745 cm-1 to 171 (±3/2). Then, corresponding
peaks are expected at ~1150, ~980, and ~1010 cm-1 in the C6H6 spectrum when
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
240
the isotope ratio (0.74) in the neutral is used. The peak at 1226 cm-1 and the
doublet at 1034 and 1065 cm-1 are assigned to 171 accordingly. Isotope ratios are
0.69 for the 849 and 1226 cm-1 pair, 0.70 for 745 and 1065 cm-1, and 0.70 for
725 and 1034 cm-1. Assignment of the intense peak at 1290 cm-1 in the C6H6
spectrum to 171 (±1/2) would have resulted in the isotope ratio of 0.66 which is
quite different from those for the doublets.
The remaining strong peak in the C6D6 spectrum appears at 597 cm-1. A very
weak peak appeared at 596 cm-1 in the gerade PIRI spectrum of C6D6+. Johnson
assigned this and a tiny feature at 656 cm-1 in the spectrum of C6H6+ to 81 (b2g)
because the frequencies observed were in excellent agreement with the
calculated values.39 A peak corresponding to the latter appears at 654 cm-1 in the
one-photon MATI spectrum of C6H6. In fact, this is the most intense band other
than the 0-0 band in this spectrum. The same peak appearing at 0.0808 eV (652
cm-1) above the 0-0 band of B~ 2E2g was the most intense fundamental in the
photoelectron spectrum of C6H6 and was assigned to 181 (e2g).16 Even though the
calculated frequencies for this strongly JT-active mode are not expected to be
dependable, its higher frequency components are expected at ~630 and ~590 cm-
1 for C6H6+ and C6D6
+, respectively. For neutrals, 181 appears at 606 and 579 cm-
1, respectively. Then, considering the intensities of the peaks at 597 and 654 cm-1
in the one-photon MATI spectra of C6D6+ and C6H6
+, respectively, their
assignment to 181 (±1/2) rather than to 81 may be more appropriate. Weak peaks
are embedded as lower frequency tails of these peaks, at ~574 and ~626 cm-1 in
Figs. 5.4 and 5.5, respectively, which may be assigned to 81.
Left unassigned in the low frequency region of the one-photon MATI
spectrum of C6D6 are two weak peaks at 373 and 481 cm-1. Referring to the
mode frequencies calculated by Johnson, these cannot be assigned to
nondegenerate fundamentals, either gerade or ungerade. Two peaks appeared
close to the above positions in the gerade PIRI spectrum, at 366 and 458 cm-1,
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
241
which were assigned to 202 by Johnson. The above two peaks in the one-photon
MATI spectrum might be assigned similarly. In the case of the gerade PIRI
spectrum, the B~ 2E2g ← X~ 2E1g transition started from the 201 vibrational state
in X~ 2E1g. Hence, the 20n overtone band is not unexpected. In the one-photon
MATI, B~ 2E2g ← X~ 1A1g, planar symmetry of the ion core is maintained
whether B~ 2E2g belongs to D6h or D2h. On the other hand, v20 (e2u) is an out-of-
plane ring deformation. Hence, it is highly unlikely that 20n overtones are
observed in the present one-photon MATI.
Having eliminated 202, there are some possible assignments for the peaks at
373 and 481 cm-1 in the one-photon MATI spectrum of C6D6. One is to assign
them to the j = ± 3/2 components of 181 already. It is to be noted that the peak at
597 cm-1 in the same spectrum has been assigned to the j = ± 1/2 components of
181. The linear JT coupling constant for v18, D18, expected from such
assignments would not be as large as was predicted quantum chemically (D18 =
0.663). Another possibility is that these are the JT components of v16 (e2g), for
which the calculated linear JT coupling constant was very large, D16 = 1.11.
There are several weak but distinct peaks at the 1200 ~ 1600 cm-1 region in
the one-photon MATI spectrum of C6D6, namely at 1222, 1293, 1443, 1493, and
1580 cm-1. Referring to the calculated frequencies again, these cannot be
assigned to nondegenerate gerade fundamentals or overtones. The peaks at 1222
and 1443 cm-1 may be assigned to combination bands, 849 + 373 or 745 + 481
and 849 + 597, respectively. It is more likely, however, that most of the peaks in
this region are due to the JT components of the linearly and quadratically active
modes.
Left unassigned below 600 cm-1 in the one-photon spectrum of C6H6 is the
doublet consisting of two peaks at 511 and 535 cm-1, which may be assigned to
the j = ± 3/2 components of 181 as in C6D6+ if the linear JT coupling constant is
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
���
Fig. 5.5 One-photon MATI spectrum of C6H6 recorded by monitoring
C6H6+ in the excited electronic state B
~ 2E2g. The x-scale at the top of the
figure corresponds to the vibrational frequency scale for the cation, or the
ion internal energy. Its origin is at the 0-0 band position. Spectrum in the
0~1900 cm-1 region magnified by 10 is shown as an inset to demonstrate
the quality of the MATI spectrum obtained in this work.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
243
Table 5.6 Vibrational frequencies (in cm-1) and assignments for C6H6+ in the
excited electronic state ( B~ 2E2g).
Assignmentd
Observed Neutrala PESb PIRIc
Empiricale JT fitf
229 234 225 201 |3/2, 1⟩
511 202 181(±3/2) 535 256 181(±3/2) |1/2, 2⟩
626 707 656 81
654 606 652 181(±1/2) |1/2, 3⟩ 971 993 976 21
1034 171(±3/2) |3/2, 5⟩ 1065 171(±3/2) |3/2, 6⟩
1115 |3/2, 7⟩ 1226 1178 171(±1/2) |1/2, 7⟩
1267 171201 |3/2, 9⟩ 1290 171201 |1/2, 9⟩ 1414 |3/2, 10⟩ 1483 21181 |1/2, 11⟩
1570 171181 |3/2, 12⟩
a From ref. 34. b From ref. 16. c From ref. 39. d Vibrational modes in Mulliken notation. e Assignment made by referring to the calculated frequencies in ref. 4. Small D18
assumed for v18. f Three-mode (v16, v17, v18) JT fit assuming large D16 and D18. See text for details.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
244
small. Above 1200 cm-1, the peaks at 1290, 1483, and 1570 cm-1 are distinct.
Even though, the v3 (a2g) and v7 (b2g) fundamental are expected in this region,
assignments to the JT components of the degenerate gerade modes are more
likely when the situation in C6D6+ is taking into account. Finally, we attempted
multi-mode JT calculations for e2g modes (v16, v17, and v18) to see if the peaks in
question could be assigned to the JT components. The SOCJT package
developed by Miller was used in the calculation.6 Initial guesses for the
unperturbed harmonic frequencies, ωi were made by referring to the calculated
frequencies for the cations and the neutral frequencies. Similarly, the linear JT
coupling constants, Di, calculated by Johnson were used initially. Small
quadratic JT coupling constants, Ki, were added for initial fit. These were all
optimized until the experimental peak positions could be reproduced by
calculation. Decent fits could be found for the MATI peaks of C6D6 and C6H6
which are suspected to be the JT components. For C6D6+, peaks at 481, 597, 729,
745, 849, 1206, 1222, 1293, 1443, and 1493 cm-1 could be reproduced at 471,
584, 728, 742, 868, 1203, 1211, 1318, 1452, and 1486 cm-1, respectively. The
optimized (ωi, Di, Ki) values were (1347, 1.07, -0.07) for v16, (741, 0.02, 0.02)
for v17, and (536, 1.22, -0.06) for v18. ωi is in cm-1 while Di and Ki are unitless.
Similarly, the C6H6 MATI peaks at 535, 654, 1034, 1065, 1226, 1267, 1290,
1414, 1483, and 1570 cm-1 could be reproduced by calculation at 531, 648, 1033,
1066, 1225, 1271, 1291, 1413, 1483, and 1572 cm-1, respectively. The optimized
(ωi, Di, Ki) values in this case were (1467, 1.09, 0.097) for v16, (1044, 0.002,
0.263) for v17, and (561, 0.991, -0.141) for v18. Then, the isotope ratios for the
unperturbed harmonic frequencies are 0.92, 0.71, and 0.96 for v16, v17, and v18,
respectively. These compare rather well with 0.97, 0.74, and 0.96 for the neutrals.
The assignments based on these calculations are listed as jn,j in Tables 5.5
and 5.6. nj means nj th value when the frequency eigenvalues are listed in the
order of increasing frequency.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
245
We also attempted the fit using the linear JT coupling constants much
smaller than the calculated values. Decent fits were possible with various
combinations of (ωi, Di, Ki), which then changed the assignments. Hence, it is to
be emphasized that the tentative assignments for the JT components presented
above can be meaningful only if the calculated linear JT coupling constants are
reliable. Even though the results from the TDDFT calculations at the B3LYP/6-
311G (2d, p) level by Johnson have been very helpful in the present work, it is
not clear whether the results are accurate enough for the fine details of the
excited state potential energy surfaces. Calculations at the configuration
interaction levels seem to be needed to confirm the results from the TDDFT
/B3LYP/6-311G (2d, p) calculations. Reliable JT calculations and hence
improved assignments would be possible thereafter. Calculations of the excited
state potential energy surface at such high levels would require tremendous
computational effort and are beyond the scope of the present work.
5.4.4 Radiationless B~ 2E2g → X~ 2E1g Transition
The widths of the 0-0 bands in the one-photon MATI spectra of C6H6 and
C6D6 are 21 cm-1, which is nearly the same as those observed in MATI to the
ground electronic state of the benzene cation. Insufficient rotational cooling and
use of high PFI voltage seem to be the main reasons for the rather poor spectral
resolution. More importantly, essentially the same width for the 0-0 bands in
MATI to X~ 2E1g and B~ 2E2g suggests that the benzene cation in the zero-point
level of B~ 2E2g does not undergo rapid relaxation to X~ 2E1g. This is in
agreement with the previous finding that some benzene cations prepared in the
B~ 2E2g state have very long lifetime, 10 microseconds or much longer.
Widths of some vibrational peaks recorded up to ~ 1600 cm-1 (0.2 eV) look a
little broader. However, it is more likely that these are due to the presence of
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
246
overlapping peaks rather than inherently broader width. For example, width of
the peak at 1570 cm-1 in the one-photon MATI spectrum of C6H6 is ~ 22 cm-1,
comparable to that of the 0-0 band. This indicates that the B~ 2E2g →X~ 2E1g
radiationless transition is not very rapid up to 0.2 eV vibrational energy in the
B~ 2E2g state. The spectral resolution in this work is not good enough to estimate
the natural lifetime at this vibrational energy. Even though the lower limit
estimated from the peak widths is ~ 1psec, actual lifetime may be much longer.
Köppel and coworkers carried out extensive ab initio calculations on the
potential energy surfaces of the benzene cation in various electronic states.23 The
outer valence Greens function (OVGF) and equation-of-motion ionization
potential coupled clusters singles and doubles (EOMIP-CCSD) methods were
used with the DZ+P or TZ2P basis set. It was found that the lowest conical
intersection between X~ and B~ lied a few tenths of an eV above the B~ state
minimum. This is in agreement with the present observation that the benzene
cation in the B~ state with vibrational energy less than 0.2 eV does not seem to
undergo very rapid relaxation to the ground state. Also performed by the same
research team was the quantum dynamics calculation for the C~ →B~ →X~ and
E~→D~ →B~ internal conversions.24 In the case of the former, the C~ →B~ and
B~ →X~ decays were found to occur on the time scales of ~ 20 and ~ 50 fsec,
respectively. The C~ state onset is thought to lie a few tenths of an eV above the
zero-point level of the B~ state, which is close to the upper limit of the internal
energy probed in this work. Hence we are not in a position to judge the
reliability of quantum dynamics results of Köppel and coworkers even though 50
fsec lifetime for the B~ →X~ decay seems to be much shorter than the 1 psec
lower limit set in this work. Quantum dynamics calculation starting from the B~
state with internal energy just below the C~ sate onset would be very interesting
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
247
in this regard.
5.4.5 Conclusions
One-photon MATI spectra of C6D6 and C6H6 converging to the ionic excited
state B~ 2E2g have been obtained by using coherent VUV radiation generated by
four-wave sum frequency mixing in Hg vapor. These are essentially the
vibrational spectra of the B~ 2E2g state. It is thought that the present one-photon
spectra are free from various complications such as those arising from multi-
photon effects. Quality of the spectra was much better than the gerade PIRI
spectra for the same state reported previously.15 Spectral patterns in the one-
photon MATI were different from those in the gerade PIRI which utilized three
sequential excitations by powerful lasers. The main reason for the difference
may be traced to the fact that an excited vibrational state of the ion core was
used in the gerade PIRI while the one-photon MATI started from the ground
state neutral.
By comparing with the calculated results, assignments have been possible for
some nondegenerate modes and degenerate modes which are not expected to be
affected much by the JT effect. To assign the JT components arising from the e2g
modes, multimode JT calculations have been carried out. Even though good fits
have been possible, we do not take the results as reliable because of the
arbitraries involved. High level calculations of the JT coupling constants in the
B~ 2E2g state of the benzene cation are called for in this regard.
5.5 C6F6+ in the X~ State
As has been mentioned already, a high quality dispersed fluorescence
spectrum for the B~ 2A2u → X~ 2E1g transition is not available for the gas phase
C6F6+ cation. A useful alternative is to record a ZEKE or MATI spectrum, which
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
248
often provides more decisive and trouble-free information than the emission
spectrum. No ZEKE or MATI spectrum has been reported for the ground state
C6F6+ cation yet, possibly due to some experimental difficulties involved in the
use of the usual two-photon scheme. In this section, a MATI spectrum for the
ground state C6F6+ cation measured with one-photon scheme utilizing coherent
vacuum ultraviolet radiation generated by four-wave difference frequency
mixing in Kr will be reported. Also reported will be the calculation of the
vibronic energy levels from the DFT results following the Johnson’s method5
and determination of the Jahn-Teller parameters through nonlinear regression of
the experimental results.
5.5.1 MATI Spectrum and Ionization Energy
The one-photon MATI spectrum of hexafluorobenzene recorded by
monitoring C6F6+ in the ground electronic state is shown in Fig. 5.6. The
spectrum in the 50 ~ 1900 cm-1 range magnified by 20 is also shown in the
figure as an inset to demonstrate the quality of the spectrum obtained in this
work. Quality of this spectrum is substantially better than any gas phase
spectrum reported so far2,40 and is comparable to or better than the dispersed
fluorescence spectrum recorded for C6F6+ trapped in the inert gas matrix. The
fact that the gas phase spectrum is free from environmental perturbation which
often plagues the matrix spectrum and that the vibrational peaks observed are
definitely due to C6F6+ as guaranteed by the mass selectivity of MATI are the
main advantage of the present approach. The fact that the electronic states
involved in the one-photon MATI are different from those in the emission in the
matrix provides a means to check the validity of the spectra obtained by the
latter. Also, one may expect to obtain structural information unavailable in the
matrix spectrum because the principles governing the two techniques are
different.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
249
79800 80300 80800 81300 81800
378
1798
1727
1650
119
160
958
1048
1108 12
1612
48
1408
1498
835
651/
671
632
362
1565
1260
183115
48
600/
615
734
759
1345
553
51847
5
909689
493
786
406
321/
333
145
214
233
266 389
283
Sign
al
Photon energy, cm-1
0 500 1000 1500
500 1000 1500
Ion energy, cm-1
Fig. 5.6 One-photon MATI spectrum of C6F6 recorded by monitoring
C6F6+ in the ground electronic state. The x-scale at the top of the figure
corresponds to the vibrational frequency scale for the cation whose origin
is at the 0-0 band position. Spectrum in the 50 ~ 1900 cm-1 region
magnified by 20 is shown as an inset to demonstrate the quality of the
MATI spectrum obtained in this work.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
250
The intense peak appearing at the lowest photon energy, namely at ~79931
cm-1 in Fig.5.6, corresponds to the 0-0 band. The position of the 0-0 band in a
one-photon MATI spectrum is usually a little smaller than the correct ionization
energy because the molecules in ZEKE states some cm-1 below the threshold can
also be ionized when high PFI field is applied. To correct this effect, the 0-0
band position was measured with various values of the PFI field and the accurate
ionization energy was estimated by extrapolation to the zero field limit. Spoil
field was not used in such measurements. The ionization energy to the ground
electronic state of C6F6+ measured in this work is listed in Table 5.7 together
with the previous results. Even though the random error in the measurement of
the ionization energy with the present MATI technique is ± 0.0001 ~ 0.0002 eV,
a larger error (± 0.0006 eV) is quoted here from our experience in the previous
measurements on other samples. There has been no report on the accurate
ionization energy measured by ZEKE or MATI. The present result, 9.9108 ±
0.0006 eV, agrees with the ionization energy measured by the threshold
photoelectron spectroscopy (TPES), 9.930 ± 0.045 eV, within the error limits.
Assuming that the shift of a vibrational peak in a MATI spectrum due to the
applied fields is similar to that of the 0-0 band, the vibrational frequency
corresponding to each peak can be determined as its distance from the 0-0 band
position. The vibrational frequency scale with the origin at the 0-0 band position
is also drawn in Fig. 5.6. The vibrational frequencies of each peak are listed in
Table 5.8 together with the previous dispersed fluorescence results.
5.5.2 Calculated Results
For C6F6+ in the ground electronic state, X~ 2E1g, with the D6h symmetry, the
vibrational modes can be classified into a1g (1, 2), a2g (3), a2u (4), b1u (5, 6), b2g
(7, 8), b2u (9, 10), e1g (11), e1u (12, 13, 14), e2g (15, 16, 17, 18), and e2u (19, 20).
Numbers in the parentheses indicate the mode number designated by Mulliken
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
251
notation. In X~ 2E1g, the e2g modes are active both in the linear and quadratic
Jahn-Teller effect while the e1g, e1u, and e2u modes are only quadratically active.
A singly excited e2g,u mode in X~ 2E1g splits into b1g,u, b2g,u, and e1g,u vibronic
species while an e1g,u mode into a1g,u, a2g,u, and e2g,u species. Only the Jahn-Teller
distortion due to the four e2g vibrations has been considered in this work.
Table 5.7 Ionization energies (IE) of hexafluorobenzene, in eV.
IE ( X~ ) Ref.
9.9108 ± 0.0006 This work
9.930 ± 0.045 TPES41
9.906 PES42
The optimized geometries for the 2E1g (D6h), 2B2g (D2h), and 2B3g (D2h) states
obtained by the method explained in the previous section are drawn in Fig. 5.7.
Energies at the 2B2g (D2h) and 2B3g (D2h) geometries referred to the D6h cusp
(2E1g) are -1011 and -1032 cm-1, respectively. The 2B3g minimum corresponds to
the global minimum under the overall Jahn-Teller effect. The 2B2g geometry
corresponds to a saddle point in the moat and has one negative eigenvalue.
The Jahn-Teller PESs along each normal coordinate of v15 ~ v18 calculated
following the method described in the previous section are shown in Fig. 5.8.
Only the portions in the 2B3g side are drawn. The 2B2g sides of PESs are virtually
the y-reflections of those in the figure except for the v18 mode, even though the
energies at the minima are slightly different from those in the 2B3g side. The
difference was more noticeable for v18 because both the linear and quadratic
parameters were significant. It is to be emphasized that the geometry which can
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
252
1.307Å
1.291Å1.448Å
1.389Å
117.6º121.2º
(b)
B2g(D2h)
1.407(1.388)Å
1.296(1.330)Å
120.0º
(a)
E1g(D6h)
1.302Å
1.286Å
1.370Å
1.427Å
122.4º118.8º
(c)
B3g(D2h)
Fig. 5.7 The optimized geometries for the (a) 2E1g (D6h), (b) 2B2g (D2h), and (c) 2B3g (D2h) states of C6F6+
calculated at the B3LYP/6-311++G (2df) level. Values in parentheses in (a) are the bond lengths in the neutral.
Copyright(c)2002 by Seoul N
ational University L
ibrary. All rights reserved.(http://library.snu.ac.kr) 2004/01/28 17:41:01
253
Table 5.8 Vibrational frequencies (in cm-1) of hexafluorobenzene cation in the
ground electronic state ( X~ 2E1g) measured by the one-photon MATI and their
assignments.
Jahn-Teller fitdObserved LIFa Calc.b Main characterc
Eigenvalue Eigenstate 0 0-0 0 |1/2, 1⟩
119 123 201 145 181(3/2) 145 |3/2, 1⟩
160 155
181(-3/2) 160 |3/2, 2⟩ 214 219 41 233 247 202 266 281 101 283 289 181(1/2) 277 |1/2, 2⟩ 321 326 171(3/2) 327 |3/2, 3⟩
333 335 182(-5/2)/171(-3/2) 333/340 |1/2, 3⟩/|3/2, 4⟩ 362 111 378 367 111 389 111 406 171(3/2)181(-1/2) 407 |1/2, 4⟩ 475 448 81 493 417 17υ(1/2)(υ=1~ 4) 494 |1/2, 5⟩ 518 498/508 183(7/2) 511/515 |1/2, 6⟩/|1/2, 7⟩ 553 554 560 21 600 609 61 / 171(3/2)181(1/2) 602 |3/2, 7⟩ 615 171(-3/2)181(1/2) 611 |3/2, 8⟩ 632 ? 651 ? 671 172(-5/2) 669 |1/2, 9⟩ 689 699 171(-3/2)182(3/2) 684 |1/2, 11⟩ 734 782 71 759 172(5/2)181(-1/2) 760 |3/2, 12⟩ 786 797 172(1/2)182(1/2) 788 |1/2, 13⟩
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
254
835 843 21181 909 172(1/2)182(5/2) 904 |1/2, 17⟩
958 21171181 / 172(-5/2)182(-1/2) 964 |1/2, 21⟩
1048 1052 21171 / 173(7/2)181(3/2) 1047 |3/2, 23⟩
1108 1107 1120 22 / 171(3/2)182(5/2) 1108 |1/2, 28⟩ 1216 161(±3/2) 1219 |3/2, 32⟩ 1248 1226 161(1/2) 1247 |1/2, 36⟩ 1260 161(±3/2) 1260 |3/2, 35⟩ 1345 1357 91 / 186(9/2) 1349 |3/2, 42⟩ 1408 1416 51 / 183(-5/2) 1409 |1/2, 46⟩ 1498 1466 121 / 188(1/2) 1494 |1/2, 48⟩ 1548 1554 11 / 151(±3/2) 1534 |3/2, 48⟩ 1565 173(-7/2)182(5/2) 1590 |3/2, 49⟩ 1650 172(-1/2)183(1/2) 1624 |1/2, 49⟩ 1727 174(9/2)184(-7/2) 1725 |1/2, 50⟩ 1798 21161(1/2) 1831 11181(1/2)
a Refs. 27 and 43.
b Vibrational frequencies at the global minimum (2B3g (D2h)) calculated at
the B3LYP/ 6-311++G (2df) level for modes other than v15 ~ v18. c Mulliken notation30 was used for the vibrational modes. For the four
linear Jahn-Teller active modes, the main characters were determined by
checking the decoupled states making the largest contribution in the four-
mode SOCJT output. Slashes ( / ) indicate two alternative assignments.
Further details for the Jahn-Teller components are listed in the next
columns. d Outputs from the four-mode SOCJT calculation after the best fit to the
experimental data. The eigenstate denotes | j, nj ⟩.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
255
be specified by the coordinates at the minima of PESs in Fig. 5.8 does not
coincide with the geometry at the global minimum because each PES represents
a cut along each normal coordinate. The Jahn-Teller stabilization energies due to
the linear (εi(1)) and quadratic (εi
(2)) coupling in each mode are listed in Table 5.8
together with the coupling constants calculated therefrom and compared with the
previous results by Miller and coworkers. The linear coupling constants
calculated here are not much different from the calculated and experimental data
by Miller and coworkers, which suggests that their assignment can be a useful
guideline for the present attempt. As was pointed out by Johnson, the magnitude
of the calculated quadratic coupling constants, and even their sign, can be in
serious error because of various approximations made such as in the choice of
the eigenvectors. These are used only as qualitative guidelines in this work.
5.5.3 Vibrational Analysis
Almost all the major vibrational peaks observed in the dispersed fluorescence
spectrum of C6F6+ in solid Ne were assigned to the Jahn-Teller states arising
from the fundamentals, overtones, and combinations of v15 ~ v18 by Bondybey
and Miller.27 Most of these, especially the ones in the low frequency region, find
their counterparts in the present one-photon MATI spectrum. Some additional
peaks appear in the latter, which are probably due to other vibrations.
The vibrational selection rule in the transition from the ground electronic
state of C6F6, X~ 1A1g, to Rydberg states converging to the ionic ground state,
X~ 2E1g, must be the same as that in the corresponding process in benzene
prepared under the beam condition. To summarize, fundamentals and all the
overtones of the a1g modes are allowed while ∆υ=2, 4, 6, … selection rule holds
for nontotally symmetric modes. Fundamentals of nontotally symmetric,
nondegenerate, and gerade vibrations which are dipole-forbidden can appear in a
MATI spectrum via vibronic mechanism, even though weakly. The vibrational
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
256
selection rule involving Jahn-Teller states in electronic transition between a 2A1
state and a 2E state was described in details by Barckholtz and Miller.6 Even
though the transition intensity to the overtones of degenerate modes is zero
under the harmonic oscillator potentials, Jahn-Teller coupling produces intensity.
The observable Jahn-Teller levels in the 2E state are those with | j | ≤ | ∑υ′′+i
i2/1 |.
When the transitions start from the vibrationless ground state, only 2/1±=j
levels are accessible. When quadratic Jahn-Teller coupling is introduced, the
2/1±=j levels are mixed with the 2/5m=j levels, rendering some intensity
to the transitions to the levels which are predominantly 2/5±=j . The fact that
the equilibrium geometries of the neutral ( X~ 1A1g) and the cation ( X~ 2E1g) are
substantially different especially along the eigenvectors of the Jahn-Teller active
modes also leads to the appearance of strong peaks associated with the
fundamentals, overtones, and combinations of the Jahn-Teller active vibrations.
Here we will first attempt to assign peaks arising from the e2g modes, v15 ~ v18,
which are linearly and quadratically Jahn-Teller active. The remaining modes
will be assigned thereafter.
The Jahn-Teller peaks were assigned by an iterative calculation-assignment-
fitting procedure as follows. The procedure started with the low frequency v17
and v18 modes which had large linear coupling constants in the calculation.
Using the linear and quadratic Jahn-Teller coupling constants for these modes
obtained by the quantum chemical calculation, we performed the two-mode
Jahn-Teller calculation with SOCJT. Similar calculation was done with Miller
and coworker’s parameters25 also. The results from the two calculations were
similar even though the former provided more fits to the experimental peak
positions than the latter. The major low frequency MATI peaks at 283, 406, 493,
and 518 cm-1 could be readily identified with the calculated Jahn-Teller
eigenvalues of 280, 424, 491, and 529 cm-1, respectively. Designating the Jahn-
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
257
Teller eigenstates as | j, nj ⟩, in which nj indicates the njth state with j in terms of
increasing energy, the aboves corresponded to |1/2, 2⟩, |1/2, 4⟩, |1/2, 5⟩, and |1/2,
6⟩, respectively, the 0-0 band being |1/2, 1⟩. From the coefficients in the linear
combination of decoupled states i∏ | υi, li ⟩, the main characters of these states
were identified. These were used to trace the evolution of the Jahn-Teller states
as the calculation got more complicated with the inclusion of v15 and v16. There
were some peaks in the low frequency region which could not be assigned to any
of the Jahn-Teller inactive fundamental such as those at 145 and 160 cm-1. These
two are close to the broad peak centered at ~ 155 cm-1 in the dispersed
fluorescence spectrum of C6F6+ in the gas phase by Sears et al.,43 which was
assigned to the 2/3=j state arising from 181, or 181(3/2). In the two-mode
calculation described above, a pair of states at 145 and 160 cm-1 were predicted
with the main character of 181(±3/2). Similarly, the pair at 321/333 cm-1 is close
the 171(3/2) pair at 326/335 cm-1 reported by Sears et al.43, while the present
two-mode calculation generated the 171(±3/2) pair at 313/317 cm-1. It is
surprising to note that the 2/3±=j states are observed in the one-photon
MATI spectrum, even though forbidden. Compared to the 145/160 cm-1, the
intensity of the 321/333 cm-1 pair is significant. It must contain a contribution
from the transition to |1/2, 3⟩ which has not been assigned yet. Using all the
above peaks, the linear and quadratic Jahn-Teller parameters for v17 and v18 were
calculated via nonlinear regression. The results were used to identify peaks up to
~ 1000 cm-1. Then, the Jahn-Teller parameters for v15 and v16 obtained from the
quantum chemical calculations were added and the four-mode calculation was
performed. Changes in the positions of the above peaks were checked. The main
character of each peak was useful for this. Other peaks with the v17 / v18 character
were identified. Since the linear Jahn-Teller parameter of v16 is very small, its
fundamental is expected to appear near the calculated frequency of 1248 cm-1. In
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
258
fact, a broad peak at ~ 1226 cm-1 in the dispersed fluorescence spectrum of C6F6+
in solid Ne was assigned to 161(1/2) by Bondybey and Miller.27 This was split
into three peaks at 1216, 1248, and 1260 cm-1 in the one-photon MATI spectrum.
According to the above four-mode calculation, the best assignments were
161(1/2) for the peak at 1248 cm-1 and 161(±3/2) for those at 1216 and 1260 cm-1.
It was difficult to identify peaks assignable to the fundamental of v15 because its
linear Jahn-Teller parameter was large and because mixing of the decoupled
states was rather extensive in the spectral region above 1500 cm-1. Finally,
eighteen peaks which were positively assigned to the Jahn-Teller states, namely
at 145, 160, 283, 321, 333, 406, 493, 518, 689, 786, 909, 1216, 1248, 1260, 1548,
1565, 1650, and 1727 cm-1, were fed into SOCJT to determine the Jahn-Teller
parameters by the four-mode fit. The parameters thus determined are compared
with those calculated by Applegate and Miller,25 those determined from the
experimental data by Sears and coworkers,43 and those obtained by the quantum
chemical calculation in this work in Table 5.9. The parameters determined from
the experimental data in this work are in reasonable agreement with those from
others. The eigenvalues of the Jahn-Teller states identified are listed in Table 5.8
together with the | j, nj ⟩ state designations and the main characters. The present
assignments are in good agreement with those by Sears and coworkers even
though there are some differences in specifying the main character of each peak.
For example, we designated the main characters of 171(3/2)181(-1/2) and
17υ(1/2)(υ=1 ~ 4) to the peaks at 406 and 493 cm-1, respectively. On the other
hand, 171(1/2) was designated to the peak at 417 cm-1 by Sears and coworkers,43
which can be better correlated to the peak at 406 cm-1 than at 493 cm-1 in the
MATI spectrum. Since a main character is simply determined by checking the
decoupled state making the most important contribution to a particular Jahn-
Teller state as found through the calculation, we do not have an explanation for
the above differences. It is to be emphasized that the designation of main
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
259
0.0 0.1 0.2
-600
-300
0
300
600
o
v17v15 v18v16
Displacement along each mode, A
Pote
ntia
l ene
rgy,
cm
-1
Fig. 5.8 The Jahn-Teller potential energy surfaces along each normal
coordinate of the four e2g modes of C6F6+ in the ground electronic state.
Only the portions in the 2B3g side are drawn.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
260
Table 5.9. Calculated and experimental Jahn-Teller coupling parameters for the
four e2g vibrational modes of C6F6+.
Miller This work Mode Calc.a Exp.b
Calc.c Exp.d
15 ωi 1629 1610 1704 1702 Di 0.34 0.23 0.261 0.246 Ki -0.0018 - -0.006 0.010 εi
(1)e 554 370 445 418 εi
(2)f -1 - -3
16 ωi 1210 1215 1248 1252 Di 0.032 0.05 0.017 0.011 Ki 0.075 - -0.321 -0.275 εi
(1) 39 61 21 13 εi
(2) 3 - -7
17 ωi 418 425 434 433 Di 0.72 0.68 0.805 0.751 Ki -0.019 ±0.006 0.017 0.030 εi
(1) 301 289 349 325 εi
(2) -6 ±1.7 6
18 ωi 256 265 265 256 Di 0.45 0.38 0.270 0.462 Ki 0.0096 - 0.057 0.057 εi
(1) 115 101 72 118 εi
(2) 1 - 4
εT 1019g 821 888g 875 a Ref. 25. b Ref. 43. c Determined from the B3LYP/6-311++G (2df) results. d Determined by the four-mode fit to the experimental data. e Stabilization energy due to the linear Jahn-Teller effect. f Stabilization energy due to the quadratic Jahn-Teller effect. g Total stabilization energy calculated by eqn. (4).
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
261
does not affect the Jahn-Teller fitting and analysis in practical terms. In fact,
such a designation is meaningless when the Jahn-Teller interaction is
significant,25 except in very low frequency region.
With all the significant Jahn-Teller peaks assigned, we set out to assign the
remaining vibrations. The most important among these are a1g vibrations, v1 and
v2. The v2 fundamental, 21, could be readily assigned to the strong peak at 553
cm-1, close to the calculated frequency of 560 cm-1. Its overtone, 22, appeared at
1108 cm-1. These are in agreement with the assignments for the strong peaks at
554 and 1107 cm-1 by Bondybey and Miller.27 The prominent peak at 1548 cm-1
was assigned to 11 based on its calculated frequency of 1554 cm-1, even though
the peak may contain contribution from Jahn-Teller states such as 151(±3/2). It is
interesting to note that the same type vibration of the benzene cation in X~ 2E1g,
namely the totally symmetric C-H stretching, appeared very weakly. This was
attributed to the fact that the C-H bond length changed less than 0.001 Å upon
ionization and presumably led to a very small Franck-Condon factor. Much
larger change in the C-F bond length, 0.034 Å in Fig. 5.7(a), in the present case
is compatible with the prominent 11. Among the remaining nondegenerate
fundamentals, 41 ~ 101 appeared near their calculated frequencies, even though
very weakly. Excellent agreement between the experimental and calculated
frequencies of the Jahn-Teller inactive vibrations shows that the potential energy
surface of C6F6+ calculated at the D2h global minimum is the excellent
representation of its structure in the ground electronic state. An interesting aspect
in the above assignments is that the fundamentals of the ungerade modes 4, 5, 6,
9, and 10 appear in the one-photon MATI spectrum even though they are doubly
forbidden. Similar transitions were observed in our previous one-photon MATI
study of benzene. The l mixing in the Rydberg states caused by the stray field
inside the instrument or by scrambling field applied was mentioned as a possible
mechanism.44
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
262
The vibrations which are Jahn-Teller active only quadratically are left
unassigned. The lowest frequency peak at 119 cm-1 can be assigned to 201 (e2u)
which has the calculated harmonic frequency of 123 cm-1. Obviously, no
alternative assignment can be found. Appearance of the more prominent
overtone, 202, at 233 cm-1 supports such an assignment. Similarly, the peaks at
362, 378, and 389 cm-1 are probably due to 111. Some peaks are left unassigned,
such as those at 632 and 651 cm-1, because there are various possibilities which
can not be confirmed.
As has been mentioned already, all the major peaks in the low frequency
range, namely those at 289, 417, 508, and 554 cm-1, observed in the dispersed
fluorescence spectrum of C6F6+ in the Ne matrix reported by Bondybey and
Miller27 find their counterparts in the one-photon MATI spectrum. This is not
always the case for the minor peaks appearing at the higher frequency range. For
example, even though the peak at 972 cm-1 is prominent in the matrix spectrum,
its counterpart can not be found in the MATI spectrum. Also, the MATI
spectrum shows more peaks than those observed in the matrix spectrum. It is to
be emphasized that most of these new peaks, even though weak, could be
identified properly by comparing with the calculated results. Most importantly,
the present data have been obtained mass selectively in the gas phase and hence
are free from the environmental effect caused by the matrix.
5.5.4 Conclusions
Quantum chemical calculation of the Jahn-Teller parameters from
topographical features of the potential energy surface requires evaluation of the
energy at the crossing point of two Jahn-Teller split states. This, in turn, requires
multiconfiguration calculations as suggested by Miller and coworkers,5 which
can be extremely time consuming for a polyatomic molecule as large as C6F6+,
especially when a large basis set is used to achieve good energy accuracy. It has
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
263
been found that a single configuration calculation at the B3LYP level, as
suggested by Johnson,4 with a rather extensive basis set is an adequate
alternative in the C6F6+ case.
The Jahn-Teller parameters obtained from the calculation were the starting
point in the assignments of vibrational peaks appearing in the one-photon MATI
spectrum. These were improved progressively until an excellent agreement
between the experimental data and the results from the four-mode Jahn-Teller
calculation was achieved. Compared to the previous vibrational analysis for the
same system by Bondybey and Miller,27 which utilized the spectral data obtained
in the matrix mostly, the number of modes assigned is larger and the frequencies
are expected to be more accurate. To say the least, the present results provide
experimental data useful for further theoretical study of the Jahn-Teller effect in
this system.
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
264
References
1. H. A. Jahn and E. Teller, Proc. R. Soc. London Ser. A 161, 220 (1937).
2. C. Cossart-Magos, D. Cossart, and S. Leach, J. Chem. Phys. 69, 4313
(1978).
3. V. E. Bondybey, T. A. Miller, and J. H. English, J. Am. Chem. Soc. 101,
1248 (1979).
4. T. A. Barckholtz and T. A. Miller, J. Phys. Chem. A 103, 2321 (1999).
5. P. M. Johnson, J. Chem. Phys. 117, 10001 (2002).
6. T. A. Barckholtz and T. A. Miller, Int. Rev. Phys. Chem. 17, 435 (1998).
7. H. C. Longuet-Higgins, U. Opik, M. H. L. Pryce, and R. A. Sack, Proc. R.
Soc. London Ser. A 244, 1 (1958).
8. S. R. Long, J. T. Meek, and J. P. Reilly, J. Chem. Phys. 79, 3206 (1983).
9. L. A. Chewter, M. Sander, K. Müller-Dethlefs, and E. W. Schlag, J. Chem.
Phys. 86, 4737 (1987).
10. G. Muller, J. Y. Fan, J. L. Lyman, W. E. Schmid, and K. L. Kompa, J.
Chem. Phys. 90, 3490 (1989).
11. H. Krause and H. J. Neusser, J. Chem. Phys. 97, 5923 (1992).
12. (a) R. Lindner, H. Sekiya, B. Beyl, and K. Müller-Dethlefs, Angew. Chem.
Int. Ed. Engl. 32, 603 (1993). (b) R. Lindner, H. Sekiya, and K. Müller-
Dethlefs, Angew. Chem. Int. Ed. Engl. 32, 1364 (1993).
13. R. Lindner, K. Müller-Dethlefs, E. Wedum, K. Haber, and E. R. Grant,
Science, 271, 1698 (1996).
14. R. Neuhauser and H. J. Neusser, Chem. Phys. Lett. 253, 151 (1996).
15. J. G. Goode, J. D. Hofstein, and P. M. Johson, J. Chem. Phys. 107, 1703
(1997).
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
265
16. P. Baltzer, L. Karlsson, B. Wannberg, G. Öhrwall, D. M. P. Holland, M. A.
MacDonald, M. A. Hayes, and W. von Niessen, Chem. Phys. 224, 95
(1997).
17. K. Siglow and H. J. Neusser, J. Electron Spectrosc. Relat. Phenom. 112,
199 (2000).
18. K. Raghavachari, R. C. Haddon, T. A. Miller, and V. E. Bondybey, J. Chem.
Phys. 79, 1387 (1983).
19. M. Huang and S. Lunell, J. Chem. Phys. 92, 6081 (1990).
20. J. Eiding, R. Schneider, W. Domcke, H. Köppel, and W. von Niessen,
Chem. Phys. Lett. 177, 345 (1991).
21. K. Takechita, J. Chem. Phys. 101, 2192 (1994).
22. K. Müller-Dethlefs and J. B. Peel, J. Chem. Phys. 111, 10550 (1999).
23. M. Döscher, H. Köppel, and P. G. Szalay, J. Chem. Phys. 117, 2645 (2002).
24. H. Köppel, M. Döscher, H. D. M. I. Bâldea, and P. G. Szalay, J. Chem.
Phys. 117, 2657 (2002).
25. B. E. Applegate and T. A. Miller, J. Chem. Phys. 117, 10654 (2002).
26. M. Allan, J. P. Maier, and O. Marthaler, Chem. Phys. 26, 131 (1977).
27. V. E. Bondybey and T. A. Miller, J. Chem. Phys. 73, 3053 (1980).
28. C. Cossart-Magos, D. Cossart, S. Leach, J. P. Maier, and L. Misev, J. Chem.
Phys. 78, 3673 (1983).
29. T. A. Miller and V. E. Bondybey, in Molecular Ions: Spectroscopy,
Structure, and Chemistry, edited by T. A. Miller and V. E. Bondybey
(North-Holland Publishing Com., New York, 1983).
30. R. S. Mulliken, J. Chem. Phys. 23, 1997 (1955).
31. M. Iwasaki, K. Toriyama, and K. Nunome, J. Chem. Soc. Chem. Commun.
320 (1983).
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
266
32. R. G. Satink, H. Piest, G. von Helden, and G. Meijer, J. Chem. Phys. 111,
10750 (1999).
33. J. M. Bakker, R. G. Satink, G. von Helden, and G. Meijer, Phys. Chem.
Chem. Phys. 4, 24 (2002).
34. G. Varsanyi, Assignments for vibrational spectra of seven hundred benzene
derivatives (Adams Higer, London, 1974).
35. E. B. Wilson, Jr., Phys. Rev. 41, 706 (1934).
36. J. Eiding, R. Schneider, W. Domcke, H. Köppel, and W. von Niessen,
Chem. Phys. Lett. 177, 345 (1991).
37. H. Köppel, L. S. Cederbaum, W. Domcke, J. Chem. Phys. 89, 2023 (1988).
38. K. Walter, R. Weinkauf, U. Boesl, and E. W. Schlag, Chem. Phys. Lett. 155,
8 (1989).
39. P. M. Johnson, J. Chem. Phys. 117, 9991 (2002).
40. T. J. Sears, T. A. Miller, and V. E. Bondybey, J. Am. Chem. Soc. 103, 326
(1981).
41. G. Dujardin, S. Leach, O. Dutuit, T. Govers, and P. M. Guyon, J. Chem.
Phys. 79, 644 (1983).
42. G. Bieri, L. Asbrink, and W. Von Niessen, J. Electron Spectrosc. Relat.
Phenom. 23, 281 (1981).
43. E. W. Schlag, ZEKE Spectroscopy (Cambridge University Press, Cambridge,
1998).
44. T. J. Sears, T. A. Miller, and V. E. Bondybey, J. Chem. Phys. 74, 3240
(1981).
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
국문 초록
전하교환 이온화 질량분석법과 광분해 반응속도론을 통해 벤젠분자
양이온의 수명이 긴 들뜬 전자상태를 발견하였다. 이런 전자상태는 그
수명이 수십 마이크로초정도이며 에너지론과 대칭 선택규칙에 의해 첫
번째 들뜬 전자상태로 예측된다. 다른 분자양이온 경우 수명이 긴 들뜬
전자상태를 효과적으로 탐색하기 위해 역배치 이중집중 질량분석계의
이온원 밖에 전하교환 이온화를 위한 충돌실을 설치하였다. 이는
이온원과 전하교환 이온화의 충돌실을 공간적으로 분리함으로써
이온/분자 회합반응을 피하고 대상 이온의 재결합 에너지를 전하교환
반응의 발열성 규칙에 근거하여 측정함으로써 수명이 긴 들뜬
전자상태의 분자 양이온을 발견하기 위함이다. 이렇게 개발된 방법을
이용하여 벤젠 분자양이온의 수명이 긴 들뜬 전자상태의 존재를
검증하고 염화벤젠, 브롬화벤젠, 시아노벤젠, 페닐아세틸렌 양이온들의
수명이 긴 들뜬 전자상태를 발견하였다. 이들 전자상태는 같은 평면의
각 할로겐원자의 비결합 p 오비탈이나 삼중결합의 π 오비탈로부터
전자를 제거해서 생긴 전자상태들로 바닥 전자상태의 π 오비탈과 서로
수직관계에 있음이 그 이유로 제안되었다.
들뜬 전자상태 벤젠족 양이온들의 분광학적 연구를 위해 수은
증기에서 사파장 혼합 방법으로 104-125nm 영역에서 수 백 nJ 세기의
가변 진공자외선 광원을 개발했다. 여기에 단광자 질량분석문턱이온화
방법을 접목하여 바닥 전자상태뿐만 아니라 들뜬 전자상태의 벤젠족
양이온들의 주로 특정 대칭모드들로 구성되는 진동스펙트럼들을 얻었다.
그 진동배정은 문헌값과 양자계산값을 참조하고 전이 선택성 규칙을
267
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
도입함으로써 이루어지며 양자화학계산으로 이온화시 기하구조 변화를
비교하여 진동구조 특징을 설명할 수 있다. 나아가 각 대칭진동모드의
세기 정보를 제공하는 Franck-Condon 인자 계산은 특히 진동배정에
효과적이다.
한편 바닥 전자상태의 Jahn-Teller 벤젠족 양이온(벤젠과 헥사불화
벤젠 양이온)들의 스펨트럼 분석을 위해 네 e2g 모드들의 위치에너지
표면을 양자화학계산을 통해 구축하고 그들 형태로부터 이론적인 Jahn-
Teller 요소들을 결정하였다. 이들을 Jahn-Teller 상태 에너지 계산에
이용해서 다중모드 회귀분석을 수행함으로써 실험과 이론 진동수들의
좋은 일치를 이루었다. 나머지 진동모드들은 양자화학계산을 통해 잘
배정하였다.
● 주요어 : 벤젠족 양이온, 수명이 긴 들뜬 전자상태, Jahn-Teller,
진동 스펙트라, 진공자외선, 질량분석문턱이온화 분광법, 전하 교환
이온화 질량분석법, 광분해 반응속도론, 양자화학계산, Franck-Condon
인자, 위치에너지 표면.
● 학번 : 99305-802
268
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01
감사의 글
오늘의 성과를 이루기까지 지난 학위과정동안 끊임없는 관심과 질타로
헌신적인 지도를 아끼지 않으셨던 김명수 교수님께 깊이 감사 드립니다.
선생님께서 보여주신 진정한 과학자로서의 자세와 가르침은 저를 일깨웠
을 뿐 아니라 제가 나아갈 인생의 중요한 길잡이가 될 것입니다. 아낌없는
조언과 용기를 주신 김홍래, 최중철 교수님께 감사 드립니다. 또한 바쁘신
중에도 저의 부족한 논문을 읽고 심사해주신 이순보, 장두전, 신석민 교수
님께 감사 드립니다.
힘들고 때로는 좌절도 했던 긴 연구기간동안 그래도 재미와 보람을 느
꼈던 것은 많은 선배님, 동료, 그리고 후배들 덕분입니다. 항상 따뜻한 말
씀으로 용기를 주신 영환 형, 성태 형, 영진 형, 연구에 매진할 수 있도록
도와 준 완구, 상태, 상현, 두영에게 깊이 감사 드립니다. 그리고 오랫동
안 실험실 생활을 함께 한 정희, ZAB-E 고치며 많은 밤 함께 고생한 동신
과 연호, 많은 도움을 주지도 못했는데 늘 열심히 따라 준 주연, 태훈, 오
규, 봉준, 승환, 지원, 효영, Long-lived ion 탐지법을 물려받아 마음 고생
많이 한 미진과 여영, MATI 팀을 이끌 미나, 그리고 함께 했던 다른 모든
후배들에게도 감사의 뜻을 전합니다.
이제 또 한 번의 과정을 마치고 새로운 시작을 하면서 너무나 아쉽고
부족했음을 알기에 앞으로 나아가는 길의 교훈으로 삼고자 합니다. 이런
저를 계속 지켜보고 관심을 가져주시면 큰 힘이 되리라 생각합니다.
마지막으로 지금까지 부족한 자식이었던 저를 항상 헌신적인 지지와 변
함없는 믿음으로 대해 주신 아버지, 어머니, 빙장, 빙모님, 그리고 다른
가족 모두에게 진심으로 깊이 감사 드립니다. 그리고 끊임 없는 애정으로
고락을 함께 해온 사랑하는 아내 명원과 딸 미주에게 이 영광을 돌립니다.
269
Copyright(c)2002 by Seoul National University Library. All rights reserved.(http://library.snu.ac.kr)
2004/01/28 17:41:01