A spectroscopic investigation of matrix isolated radical cations of para-dichlorobenzene, para-dimethoxybenzene, and nap...

MISSING IMAGE

Material Information

Title:
A spectroscopic investigation of matrix isolated radical cations of para-dichlorobenzene, para-dimethoxybenzene, and naphthalene
Physical Description:
xiii, 147 leaves : ill. ; 29 cm.
Language:
English
Creator:
Personette, William K., 1958-
Publication Date:

Subjects

Subjects / Keywords:
Matrix isolation spectroscopy   ( lcsh )
Cations   ( lcsh )
Chemistry thesis Ph. D   ( lcsh )
Dissertations, Academic -- Chemistry -- UF   ( lcsh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1993.
Bibliography:
Includes bibliographical references (leaves 141-146).
Statement of Responsibility:
by William K. Personette.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001950905
oclc - 31200891
notis - AKC7447
System ID:
AA00003671:00001

Full Text







A SPECTROSCOPIC INVESTIGATION OF
MATRIX ISOLATED RADICAL CATIONS
OF PARA-DICHLOROBENZENE,
PARA-DIMETHOXYBENZENE, AND NAPHTHALENE









By

WILLIAM K. PERSONETTE


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL
FULFILLMENT OF THE REQUIREMENT FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY




UNIVERSITY OF FLORIDA


1993























To my mother and brothers

and in memory of my father











ACKNOWLEDGEMENTS


The author wishes to express gratitude to Professor Martin T. Vala

for his willingness, patience and support to allow me to develop as an
independent researcher and for providing a thoughtful, reflective
atmosphere in which to work.
I am indebted to Dr. Jan Szczepanski for his experimental expertise
and kind words of encouragement through many difficult experiments.
Dr. Bryce Williamson is to be thanked, also, for his perspective and timely
advice.
FTICR/MS experiments were performed by Dr. Chris Barshick in
Professor John Eyler's laboratory and I thank them both. Theoretical

calculations of electronic states of para-dichlorobenzene cation were
performed by Professor Michael Zerner and vibrational calculations for

para-dichlorobenzene and para-dimethoxybenzene were performed by
Marshall Cory of Professor Zerner's group. I am indebted to both of them.
Previous members of the Vala group, Drs. Bob Pellow, T. M.
Chandrasekhar and Dennis Roser, and Marc Eyring made research
rewarding through fruitful discussion and shared experience.
Past and present members of Professor Colgate's group, Kyle Reed
and Drs. Evan House, Charles Simon and Ken McGill also contributed to
my outlook concerning research.









The support of all members of the Chemistry Department including
machinists, electronics technicians, glass blowers and secretaries as well
as all other professors, postdoctoral researchers and graduate students
whom I have failed to mention by name is appreciated. Also, all those
outside of the Chemistry Department with whom I have worked should be
thanked.
Professor Larry Eng-Wilmot is thanked for my present employment
at Rollins College as is Brie McDougall for typing this dissertation.
Finally, the positive influence my mother, father and brothers have
had on my decision to pursue a career in chemistry should not be forgotten
or underemphasized.












TABLE OF CONTENTS




page

ACKNOWLEDGEMENTS...............................................................iii

LIST OF TABLES .......................................................................vi

LIST OF FIGURES...................................................................... viii

ABSTRACT .................................................................................. xii

INTRODUCTION.........................................................................1

EXPERIMENTAL.........................................................................22

THEORY ................................................................................... 39

RESULTS AND DISCUSSION

Para-Dichlorobenzene ........................................................ 50

Para-Dimethoxybenzene.........................................................90

Naphthalene.......................................................................110

CONCLUSIONS AND FUTURE WORK...........................................137

REFERENCES .......... .................................................................141

BIOGRAPHICAL SKETCH ........................................................147











v








LIST OF TABLES


ITable P

1 Ionization Mechanisms in the Negative Glow
Region of an Abnormal Glow Discharge.................... 7
2 Colors of Glow Regions in a DC Gas Discharge..................... 9

3 Low-lying Metastable Energy Levels of Rare
Gas Atom s.......................................................... 10

4 Visible/UV Absorption Bands of PDCB Cation
in Ar Matrix at 12 K................................................54

5 Comparison of the PDCB Cation Optical Absorption
Transitions with the Photoelectron Bands .................55

6 Resonance Raman Bands of the PDCB Cation in
Ar Matrix at 12 K................................................ 65

7 Fluorescence Bands of PDCB Cation in Ar
M atrix at 12 K ...................... ......... ......................66

8 Calculated and Observed Spectra of PDCB
Cation (Energies in eV) ............................................

9 Calculated and Observed Vibrational
Frequencies (cm-1) for Neutral PDCB.........................73

10 Calculated and Observed Vibrational
Frequencies (cm-1) for PDCB Cation...........................74

11 Calculated and Observed Vibrational
Frequency Shifts Upon Ionization of PDCB ...............76

12 Calculated Intrinsic Stretching Frequencies
and Force Constants for Neutral PDCB......................78

13 Calculated Intrinsic Stretching Frequencies
and Force Constants for PDCB Cation........................78

14 Calculated and Observed IR Vibrational
Frequencies (cm-') for PDMOB Cation...................... 102








15 Calculated Intrinsic Stretching Frequencies
and Force Constants for Neutral PDMOB......................104

16 Calculated Intrinsic Stretching Frequencies
and Force Constants for PDMOB Cation.......................104

17 Theoretical and Experimental IR Bands
for Naphthalene Cation.............................................124








LIST OF FIGURES


Figure P

1 Representative Molecules of Interest.........................................

2 Astrophysical Processes of Interstellar Molecules
and Grains.................... ....................... ................. 6

3 Voltage/Current Characteristics of Gas Discharges ................ 12

4 General Experimental Set-up for Matrix-Isolation .................... 16

5 Matrix Preparation and Subsequent Treatments....................... 19

6 Cryostat with Vacuum Shroud and Furnace
Assembly Attached............ ................ ......................... 24

7 Schematic of Pulse-Valve Power Supply and
Driver Circuit ............................................................ 28

8 Pulsed-Glow Discharge Assembly........................................ 30

(A) Schematic of Pulsed-Glow Discharge/
Matrix Isolation Apparatus

(B) SIMION Calculated Trajectories of
p-Dimethoxybenzene Cations

9 General View of Experimental Set-Up for Matrix
Isolation Using Electron Bombardment Ionization............ 33

10 General View of Experimental Set-Up for FTICR Using
Glow Discharge Ionization.......................................... 37

11 General Types of Electronic Configurations Associated
with '2M Molecular Cations ........................................... 42

12 Fixed Nucleus Potential Well Calculation (solid line)
and its Harmonic Approximation (dashed line) ................ 46

13 Effect of Basis Set and Level of Theory on Potential
W ell Calculations...................................................... 48


viii








14 UV-Visible Spectrum of Neutral PDCB in Ar Matrix
at 12 K, Showing Band System at 280 nm and Visible
Spectrum of PDCB Radical Cation in Ar Matrix at
12 K, Showing Three Band Systems Marked
6, C and ...................................................... ...................... 52

15 FTIR Spectra of Neutral PDCB (top) and Neutral Plus
Cation PDCB (bottom) in Ar Matrix at 12 K..................... 57

16 Correlation of 520 nm Visible Absorption Band
Absorbance of PDCB Cation with Four
IR Band Absorbances... .............................................. 60

17 Fluorescence and Raman Bands of PDCB Cation
in Ar Matrix at 12 K Excited by 514.4 nm
Ar Ion Laser Radiation (blackened dots indicate
sharp Resonance-Raman bands while broad
bands are due to fluorescence)........................................ 62

18 Fluorescence of PDCB Cation in Ar Matrix at 12 K Excited
by 488.0 nm Ar Ion Laser Radiation, Showing
Emission from Two Sites (A and B)................................. 64

19 Schematic of HOMO (Highest Occupied Molecular Orbital)
of PDCB Cation.................................... ................. 72

20 Raman-Active (ag) Normal Modes Observed for
PDCB Cation (clockwise from upper left:
modes 2, 3, 4 and 6, same orientation as in
Fig. 19, bonds are removed for clarity)............................ 81

21 Infrared-Active Normal Modes Observed for PDCB
Cation (clockwise from upper left: modes 11, 13, 21
and 28, same orientation as Fig. 19 except mode 28
which is viewed along the C-C1 axis, bonds are
excluded for clarity).............................................. 83

22 FTICR Mass Spectrum of Pulsed-Glow
Discharged PDCB Seeded in Ar.................................... 85

23 Visible/UV Spectra of PDMOB Neutral (k < 300 nm) and
Radical Cation (300 nm < X < 470 nm) in an
Ar M atrix at 12 K......................... ............................. 92








24 Infrared Spectra of Discharged PDMOB/CC14/Ar
Mixture (top) Showing Neutral and
Radical Cation (black circles) Bands
and Only Neutral PDMOB Molecules (bottom)
in an Ar Matrix at 12 K. Bands with Black
Inverted Triangles are Unassigned................................ 95

25 Correlation Diagram between 460 nm Visible Band and
Four IR Peaks (Varying Absorbances
Produced by Different Photolysis Times)........................... 97

26 Visible/UV Spectra of

(A) Unphotolyzed PDMOB Radical Cation
Formed from Discharged PDMOB/CC1VAr
Mixture;

(B) Spectrum A after 25 Minutes Photolysis;

(C) Discharged CCl4/Ar Mixture before
Photolysis;

(D) Spectrum C after 25 Minutes
Photolysis. All in Ar at 12 K.................................100

27 FTICR Mass Spectrum of Pulsed-Glow Discharged
PDMOB Seeded in Ar ................................................107

28 Visible/UV Spectra of Neutral and Cationic
Naphthalene in Ar at 12 K in the
290-700 nm Spectral Region.......................................... 113

29 Infrared Spectra of Neutral (bottom) and
Neutral Plus Cationic Naphthalene
(top) in Ar at 12 K. New Peaks Due to
Cation are Marked with Solid Circles............................ 115

30 Correlation of 675 nm Visible Band Absorbance
of Naphthalene Cation with 1218 cm-1
IR Band Absorption. The Empty Square
Indicates Datum from an Experiment
with CCl4 Absent..................................................... 118








31 Visible/UV Spectrum of Neutral (N) and Cationic (N+)
Naphthalene in Ar at 12 K Prior to Photolysis (top);
Spectrum after Photolysis for Five Minutes with
Medium-Pressure Hg Lamp (bottom).
Note that the Intensity of the Band System in
the 300-500 nm Range Has Been Compressed
Two-fold for Visual Clarity ...........................................120

32 Infrared Spectra of Neutral and Cationic Naphthalene
in Ar, 12 K, Prior to Photolysis (top) and After
Photolysis (bottom); Same Sample as in Figure 31
([CC1]/[Ar] = 3/1000) ................................................... 122
33 Resonance Raman Spectrum of Cationic
Naphthalene in Ar (0.1 % CCl4) at 12 K
(50 mW, 614.4 nm, Rhodamine 590 excitation).................127

34 FTICR Mass Spectrum of Pulsed-Glow Discharged
Naphthalene Seeded in Ar..........................................130








Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the Requirements for
the Degree of Doctor of Philosophy


A SPECTROSCOPIC INVESTIGATION OF MATRIX ISOLATED
RADICAL CATIONS OF PARA-DICHLOROBENZENE,
PARA-DIMETHOXYBENZENE, AND NAPHTHALENE

By

William K. Personette
December 1993

Chairman: Martin T. Vala
Major Department: Chemistry
Matrix isolation spectroscopy provides a means by which relatively
unstable chemical species may be investigated. Radical cations derived
from aromatic, organic molecules represent an important class of such
species. Specifically, those derived from the title molecules, para-
dichlorobenzene (PDCB), para-dimethoxybenzene (PDMOB), and
naphthalene (N), were studied spectroscopically using the matrix isolation
technique.
Radical cations of the title molecules were generated by either pulsed-
glow discharge or electron bombardment and then trapped in argon at 12 K.
Correlation of known cationic visible absorption bands with infrared (IR)
bands observed for the same sample/matrix allowed assignment of IR
bands to each of the three radical cations studied. Ab initio Hartree-Fock
Self Consistent Field (HF/SCF) level calculations of harmonic vibrational
frequencies for neutral and cationic PDCB, PDMOB, and N provided
further evidence for assignments and were used to characterize vibrational
frequency shifts, and intensity and bonding changes evoked upon









ionization. IR-active vibrations for PDCB+ are assigned at 843, 986, 1110
and 1429 cm-1, while those attributed to PDMOB+ occur at 1309 (1304), 1342,
1388 and 1427cm-1. IR bands at 1016, 1023, 1218/1215, 1401, 1519 and
1525 cm-1 are assigned to N+.
Resonance-Raman (RR) spectra of PDCB+ and N+ and the
fluorescence/RR spectrum of PDCB+ were recorded. The fluorescence
spectrum matched well with previously reported gas phase emission of
PDCB+ while the RR spectra obtained compared favorably with previous RR
studies of these radical cations in alkylhalide glasses at 77 K. Raman-
active vibrations for PDCB+ are reported at 330, 1113, 1189 and 1598 cm-1
while those ofN+ occur at 510, 765 and 1400 cm-1.
The role of CCl4 (and its degradation products) as gas phase
ionization enhancer and matrix charge compensator was further clarified
through photolysis experiments. Correlation analysis allowed assignment
of a broad, featureless visible absorption centered at -420 nm to CC13 Cl
based on previous assignment of 927 and 1019 cm-1 IR absorbances to this
species.
Finally, preliminary Fourier Transform Ion Cyclotron Resonance

Mass Spectrometry (FTICR/MS) experiments of gas phase pulsed-glow
discharge generated plasmas indicated fragmentation of parent molecules
and these spectra were related to matrix composition determined
spectroscopically. The necessity for independent, complementary
experimental and/or theoretical methods is thus emphasized.


xiii










INTRODUCTION


The spectroscopy of molecular ions has been the subject of extensive
investigation'14 since the early realization of the important role such ions
play in physics, chemistry and biology. Many of these ions are obviously
quite stable when present as constituents of solids and in the solution phase
but are quite unstable and reactive when in the gas phase. Many such
species are also short-lived and transient in nature and thus difficult to
study by conventional means. One such class of molecular ions is the
radical cations derived from aromatic, organic molecules such as
substituted benzenes and fused benzene rings known as polycyclic aromatic
hydrocarbons (PAHs). Representative molecules are depicted in Figure 1.
The significance of this particular group lies in the fact that many form
ions from their neutral parent molecules upon exposure to specific clay
minerals present in soils.15-22 Hence, the ability to identify and characterize
spectroscopically this group of ions is of consequence to monitoring the fate
of this industrially important series of molecules in the environment.
Moreover, since these ions may play a role in the chemistry of their disposal
by combustion or other destructive means spectroscopic determinations will
assist in understanding the reaction mechanisms in which these
intermediates may be involved. Interestingly and in contrast to their
environmental role, it has been suggested that the neutral PAHs are an



























Figure 1


Representative Molecules of Interest






ci---g-Cl


@-OH


-FCH


D-CH3


H3CO-- OCH3


CH CH3


CH3 CH3


C1-- N-NH2


9


CH3

s


CHJ S 'CH3




NH2 21YaNH2


@-OCH3


& C2H5


--NH2

@- N(CH3)2


ClI--OCH3

(CHCH2)n


H3COO-@ -OCH3


N(CH3)2 -N(CH3)2


9-~cl


O O O


LOjj


mg@


&NHNH-g









important and significant component of the interstellar medium based on
unidentified infrared (UIR) spectral emission features observed by
astronomers23 at 885, 1149, 1282, 1613 and 3030 cm-1. Considering the harsh
interstellar environment, i.e., a very strong ultraviolet radiative flux,
others24 have suggested the PAH hypothesis may better match observations
if these molecules exist as ionized species rather than neutrals (Figure 2).
For this reason, the spectroscopy of radical cations derived from PAHs is of
astronomical as well as terrestrial importance. A brief review of previous
studies of molecular ion spectroscopy follows, emphasizing the techniques
by which ions are produced and investigated.
There are a wide variety of means by which atomic and molecular
ions are produced. Historically, flames25' 26 and electrical discharges27 were
the predominant environments of gaseous ion production. Further
refinements for analytical purposes consisted of essentially extracting and
optimizing one of the many ionization processes operating in these complex
systems (Table 1). Specifically, electron impact,28 and Penning,29 field,
thermal and chemical ionization are the most commonly employed
processes today.30 Photoionization is also a prominent method of ion
production.31 The method of ionization used is a function of ion production
efficiency, tolerance of molecular ion fragmentation, and amenability to the
techniques used to detect and characterize the ions produced. For example,
electron impact is well suited to mass spectrometry (MS) since it is
relatively efficient, vacuum requirements of the mass spectrometer are not
exceeded, and fragmentation can be accounted for. Recently, glow
discharges have been coupled to mass spectrometers32-35 employing


























Figure 2


Astrophysical Processes of Interstellar Molecules and Grains
















NIR


FIR


UV


.dam
off








Table 1


Ionization Mechanisms in the Negative Glow Region
of an Abnormal Glow Discharge
(adapted from Ref. 32)


I. Primary Ionization Mechanisms

A. Electron Impact:

M + e- -- M+ + 2e-

B. Penning Ionization:

M + Am* -* M+ + AO + e-


(MO = neutral molecule)



(Am* = discharged gas A in a
metastable state)


II. Secondary Ionization Mechanisms

A. 1. Nonsymmetric Charge Transfer:

A+ + M M+ + AO

2. Symmetric (Resonance) Charge Transfer:

B+fast + BOslow -- B+slow + Bfast

3. Dissociative Charge Transfer:

A+ + MX M+ + X + Ao

B. Associative Ionization:

Am* + M -* (AM)+ + e-

C. Photoionization:


M + hv --)M+ + e-









differential pumping to overcome pressure overloads. Another method of
avoiding pressure difficulties and generating metastable and/or ionic
species by glow discharge is to pulse the discharge gas into the mass
spectrometer.36 This method consists of applying a voltage, typically a few
kilovolts, to an electrode suspended in close proximity to another electrode
held at ground potential both of which are under high vacuum. Upon
introduction of a gas pulse, typically a rare gas, in the region between the
electrodes an electrical discharge is initiated which extinguishes itself once
the pressure has dropped below the breakdown potential of the gas used.
This glow discharge is a low-temperature plasma existing in a moderate
pressure (0.1 to 10 torr) of inert gas exhibiting currents intermediate
between those of Townsend and arc discharge, typically in the millampere
range. The characteristics of gas discharges37 are shown schematically in
Figure 3 in which Vb, Vn and Vd indicate breakdown, normal and

discharge potentials, respectively. Low energy electrons of a few eV and
densities of ~1012/cm3 cause excitation of rare gas atoms to metastable states
while a much smaller number of more energetic electrons (5-50 eV) are
also produced. The glow discharge is probably the most commonly
recognized form of gaseous discharge because of its ease of production and
commercial utility as use in "neon" lights. The characteristic colors38 are
given in Table 2. Atoms and molecules either doped into the gas or
introduced separately into the discharge region are excited and/or ionized
by collisions with ions, metastable gas atoms, electrons, or photons (Table
1). The ionization potential and energies of low-lying metastable states of
the rare gas discharged39 (Table 3) dictate the ability of this ionization









Table 2


Colors of Glow Regions in a DC Gas Discharge


Negative Glow

pale blue
blue

yellowish white

green

dark blue

orange


Positive Column

pink

red

pale yellow with pink center

red to violet

dark red

brick red


Gas

H2

N2

02

He

Ar

Ne









Table 3


Low-lying Metastable Energy Levels of Rare Gas Atoms
and Ionization Potentials of Molecules of Interest


Metastable Energy, eV

19.8, 20.7

16.6, 16.7

11.5, 11.7

9.9, 10.5

8.3, 9.4


Molecule

para-dichlorobenzene

para-dimethoxybenzene

naphthalene


Ionization Potential, eV

24.58

21.56

15.76

14.00

12.13


Ionization Potential, eV

8.98

7.9

8.2


Gas

He

Ne

Ar

Kr

Xe





































0



cI~
~a)









Za)
bO

Cu"

0







12




I -4
0






1-1

^-4
t

S 0
Ib
'-4





-~ S










I g

I j
I 0
I: i-I














--f-
















0
'I I
I


(A OA
(A) aS~p









method to produce the ions of interest, i.e., the ionization potential of the
atom or molecule must be overcome via the ionization process energies
available in the discharge.
Gas phase ion spectroscopy and mass spectrometry have provided a
wealth of information concerning molecular ion structure and reactivity.
The coupling of these methods through excitation wavelength specific
photoionization and fragmentation techniques followed by classical ion
detection has made it possible to infer optical absorption spectra at
sensitivity levels typically associated with laser induced fluorescence (LIF)
and standard mass spectrometry. In addition, cooling via supersonic
expansion has assisted in the interpretation of gas phase ion spectra by
decreasing line widths and under favorable conditions eliminating
vibrationally and rotationally "hot" bands. These advances are impressive
and will continue to contribute to gas phase ion spectroscopy. However,
because of their level of sophistication they are as technically demanding as
they are informative. An alternative which is technically less arduous yet
provides significant and substantial spectroscopic and chemical
information concerning molecular ions is the matrix isolation technique.40
The data obtained via this technique are often the starting point for high
resolution gas phase experiments.
The encapsulation of chemical species within some medium is not a
particularly new idea.41-44 Numerous examples exist naturally, one of
which is the trapping of the 83 anion within the aluminosilicate structure

of the mineral lapis lazula. A blue color is imparted to the mineral as a
result of optical absorption by the anion. The dissolution of molecules in









solvents followed by freezing and y-ray irradiation to form solid glasses

containing ions for low temperature studies is well established.45 Fast-flow
techniques46 have allowed monitoring of transient species, also, as have
other time resolution procedures. The purpose of the matrix isolation
technique as formally introduced by Whittle, Dows and Pimentel47 in 1954 is
to embed species of chemical interest in a medium exhibiting minimal
interaction with its "guest" as well as transparency in the spectral regions
used to probe the species under study. Once-transient species now isolated
in the matrix could then be investigated at a more leisurely pace since the
processes accounting for their instability have been greatly arrested if not
stopped completely. This ideal is never completely realized but is most
closely approached by isolation in solid rare gases. Since these materials
are produced at low temperatures, high vacuum and cryogenic techniques
are employed to create substrates onto which the mixture of excess matrix
gas and species of interest are deposited. The general experimental
arrangement is shown in Figure 4. The low temperatures required have
the additional benefit of simplifying spectra through line narrowing and
elimination of "hot" bands. The method of deposition varies depending on
the species studied and the spectroscopic technique employed.
Stable, volatile materials may simply be premixed with the matrix
gas and deposited either in a continuous or pulsed mode.48 Pulsing appears
to minimize multiple site formation in the matrix as well as aggregation of
the species to be isolated. It allows for more rapid depositions and, as a
result, less contamination by impurities. Matrices are usually clearer (less
scattering), also.49 Detection of these species by standard spectroscopic


























Figure 4


General Experimental Set-up for Matrix Isolation

















Cryostat under high vacuum



Sample introduction









techniques is not particularly difficult. The production and trapping of
highly involatile, transient chemical species offers an example of the other
extreme of sample preparation. High temperature furnaces and laser
ablation are used to produce gaseous atoms, molecules and ions of metals
and refractory materials while high energy processes, such as those
discussed for gaseous ion production, provide an avenue to unstable
molecular fragments and ions. This is shown in the upper portion of
Figure 5. Continuous flow deposition is best suited to those methods that
provide a continuous flow of sample while pulsing is appropriate when the
sample itself is generated in pulses as is the case with ablation via pulsed
high-powered lasers. Pulsing offers the added advantage of avoiding
pressure build-up in the vacuum chamber when differential pumping is
not available or when locally and temporally high gas pressures are
required. Unambiguous detection and characterization of these ephemeral
species is more difficult than that of stable species for which the
experimental and theoretical bases are much more extensive. The methods
discussed so far are primarily associated with formation of species just
prior to deposition. The pre-matrix mixture has been studied
spectroscopically and by mass spectrometry in an attempt to provide
evidence for and correlation with the observed and proposed entities found
in the matrix. This entails a detailed understanding of the processes
occurring on and near the substrate surface during deposition. Some
insight into these processes is gained by treatment of the matrix after it has
been formed. The focus now is manipulation of chemical species in the
matrix after deposition.



























Figure 5


Matrix Preparation and Subsequent Treatments














High temperature species


Matrix gas \


Matrix gas with volatile species


High energy process,
e.g., electrical discharge


Photolysis


Additional energy input


Annealing


Diffusion leading to aggregates


Matrix Isolated
Sample


New matrix: species









Once a matrix is formed it can be subjected to a variety of
perturbations to induce changes that can be monitored spectroscopically.
The most frequently employed methods are annealing of the matrix and
exposure to external radiation sources, i.e., photolysis. This serves the
purpose of generating new species in the matrix and providing insight into
the initial matrix composition and the thermodynamics, photochemistry
and physics associated with the transformation(s) within the matrix.
These processes are depicted in the lower portion of Figure 5. Annealing is
the controlled input of thermal energy into a matrix for the purpose of
eliminating multiple sites, and for inducing the formation of aggregates
and disintegration of highly reactive intermediates. These are low
activation energy processes which are put into action by the softening of the
matrix upon warming and subsequent diffusion and reaction of entrapped
atoms, molecules, and/or ions.
Photolysis is the controlled input of electromagnetic radiation for the
purpose of generating, identifying, and distinguishing species present in
the matrix. High energy photolysis, e.g., using y-ray or ultraviolet

radiation, is capable of ionizing molecules with high ionization potentials
in the matrix. Lower energy light, e.g., visible or near infrared, is better
suited to photodetachment of electrons captured only weakly by electron
traps intentionally added to the matrix to enhance the formation of positive
ions of interest. In both instances, monitoring and correlating the growth
and diminishment of spectroscopic bands tentatively assigned to matrix
species as a function of photolysis energy and duration make significant









contributions to conclusive assignment of bands to a specific chemical
species.
The purpose of the research presented in this dissertation was to
employ the matrix isolation technique to experimentally obtain the
vibrational absorption spectra of radical cations derived from para-
dichlorobenzene (PDCB), para-dimethoxybenzene (PDMOB), and
naphthalene (N). To this end, a pulsed-glow discharge/matrix isolation
technique was developed and employed as the primary method of ion
generation (electron bombardment was also used) and trapping. During
the course of this investigation, Thoma, Wurfel, Schlachta, Lask and
Bondybey50 reported on a very similar methodology, which is noted and
acknowledged. A description of experimental details, theoretical
considerations and results and discussion for each molecule studied
follows.










EXPERIMENTAL


The matrix isolation technique requires a vacuum system, cryostat
and gas/sample inlet port. The general setup is shown in Figure 4. The
cryostat with vacuum shroud and furnace assembly attached is shown in
Figure 6. For the purposes of the research presented here the gate valves

and furnace assembly were removed and replaced with a pulsed-discharge
or electron bombardment assembly for ion generation (to be described later),
a window transparent to the spectral region under investigation or a port
for evacuation depending on the method of ion production used or other
experimental constraints.
The system was evacuated by a two inch air-cooled oil diffusion pump

(Alcatel PDR 250, rated at 200 L/sec) back-pumped by a mechanical
roughing pump (Alcatel 2020, rated at 450 L/min) and separated from the
sample matrix by a liquid nitrogen cooled cold trap. All surface-to-surface
interfaces were sealed with O-rings. The pressure was monitored with an
ion gauge (Model 4336P, Kurt J. Lesker Co.) interfaced with a Granville-
Phillips 270 Gauge Controller. The system was pumped to the 4 7 x 10'
torr range (without a full cold trap) then to 1 4 x 106 torr (with a full cold
trap) before any depositions were performed. The pressure measured
decreased to 4 7 x 10-7 torr once the cryostat had reached matrix deposition
temperature (-12 K). Typically the system was pumped on overnight,
followed by a matrix experiment the next day. When a pulse valve was
























Figure 6


Cryostat with Vacuum Shroud and Furnace Assembly Attached












ELECTRICAL --

He GAS
He GAS


THERMOCOUPLE
and
HEATING WIRES


EXPANDER
Ist STAGE
2nd STAGE-



COPPER COLD TIP

TARGET WINDOW-

RADIATION SHIELD


JOINT


ASSEMBLY








employed it was either left on during the pumping cycle or the gas manifold
line was evacuated via a bypass valve (cf. Figure 4). The bypass valve also
allowed introduction of dry nitrogen gas into the vacuum system to expedite
cryostat warm-up and subsequent disassembly and reassembly.
The cryostat employed consisted of a two stage closed-cycle Air
Products Displex (Model DE 202) helium refrigerator. The first stage was
attached to a copper radiation shield held at 40 60K while a copper cold
finger was attached to the second stage. A copper window container was
then screwed into the cold finger (a polished aluminum block was used for
Raman studies). All metal-metal and metal-window interfaces were joined
using indium gaskets. The window temperature was measured using a
gold-chromel (0.07% Fe) thermocouple compressed between indium at the
cold finger-copper window container interface. An Air Products APD-B
temperature controller was used for temperature read-out and was
calibrated against liquid nitrogen. Temperatures were maintained at 12 K.
Argon (Research purity, 99.9995% pure, Matheson Gas Products)
matrices were deposited either in a pulsed or continuous mode depending
on the method of ion production. In either case CCl4 (if added) was
premixed in a gas manifold or bulb by conventional manometric techniques
to a ratio of 100/1 to 500/1 (Ar/CC14). When the pulsed mode was employed
the sample was premixed from a side arm down stream from the matrix
gas just prior to the pulse valve. For PDCB and N samples, room
temperature provided a sufficient sample vapor pressure for deposition. In
the case of PDMOB the side arm was heated to ~600C. When a continuous
flow mode was employed as in the case of electron bombardment of N, the









sample was inlet and mixed with the matrix gas just inside the vacuum
chamber just prior to deposition.
Pulsed-glow discharge deposition was the primary method employed
for producing and trapping ions under investigation. The pulse valve
(General Valve, Series 9) was powered and driven by the home-made circuit
shown schematically in Figure 7. It consists of a power supply (D-4, 5, 6)
which supplies +28 volts (V) DC to the 2N3055 transistor (C-2) and +5 V DC
to a pulse generator (A-C, 4-6), a TIL 112 optoisolater (B3), and a delay
circuit (C, 2-3) used to couple to another system, e.g., a pulsed Nd-YAG
laser. The 100 KA variable resistor (C5) determines the +5 V output
duration while the 2 MQ variable resistor (B5) determines the 0 V (ground)
output duration. Together they determine the duty-cycle of the pulse
generator. A 5 msec, +5 V pulse at 1 Hz frequency was used. This pulse
triggered the optoisolator to drive the pulse valve solenoid (labeled SOL, B1)
and open the pulse valve. At that time the gas/sample mixture entered the
vacuum system through a 0.8 mm stainless steel (SS) tube held at ground
potential and initiated formation of a plasma between the tube and a thin
copper wire attached to a hemispherically-shaped copper grid held at +3
KV potential (see Figure 8A). The primary plasma was generated in the -2
mm gap between the SS tube and the Cu wire. A weaker, secondary
plasma was generated within the cylindrical lense held at +150-200 V. As
mentioned, the pulse rate was set to 1 Hz but formation of the discharge
increased the frequency to ~2 Hz. As deposition proceeded, typically two
hours, the needle valve had to be opened further and/or the backing
pressure in the manifold increased to allow reliable plasma formation.



































a)








a)

0




a)
U'

0
e- .)

CU


C.)
CI2

























III


3>

























Figure 8


Pulsed-Glow Discharge Assembly
(A) Schematic of Pulsed-Glow Discharge/Matrix Isolation Apparatus
(B) SIMION Calculated Trajectories of p-Dimethoxybenzene Cations





(A)


to pulsed
valve


IRV ^

+ 3 kV


grid


"":.. ;. ." "
cylindrical
lens


1

10K
window


lens


window


(B)


CD


tr
grid


1 1 1 [ 1 1. 1 1 11 1 1 1









Deposition rates were =50 torr/hr with a manifold pressure in the 500 760
torr range. The system pressure pulsed over a 3 6 x 10-6 torr range during
deposition.
The trajectories of PDMOB cations were simulated by the SIMION
program and are shown in Figure 8B. The initial position for each
trajectory was chosen to lie on a line spanning the hemispherical grid
space and the direction of flight was varied from 0 (directly at the window)
to 3600 by 300 intervals. The shape of the grid and the high positive potential
on it resulted in the ions taking only one pathway. The trajectories shown
are for a sample window with zero or a small positive potential. The overall
behavior of ions in this discharge arrangement is very much like that of
hollow-cathode electrodes in discharge lamps. However, since the cations
are deposited in a matrix the potential of the matrix becomes increasingly
positive to a point at which necessary deposition of anions occurs to insure
charge neutrality.
The method of pulsed-glow discharge was used to generate PDCB,
PDMOB, and N cations for matrix isolation and was also coupled to a
Fourier Transform Ion Cyclotron Resonance (FTICR) Mass Spectrometer
for preliminary investigation of possible fragmentation accompanying
ionization (to be discussed later). Another ionization method employed for
matrix isolation investigation of N cation and other cations not discussed
here was electron bombardment.

The ionization of N was accomplished by use of the specially-
constructed electron bombardment source shown in Figure 9. The
tungsten filament (0.1 mm diameter) was heated by a current of














0
0

N

~0


II
o0
0

0


0




cI

0

0

W

0
v-i













'-I



























4m
+


w -
o I


0 <


g
m

m
iii
gv,
b

~









1.20 1.45 A (at 6 8 V) and electrons were emitted. The anode potential
(UA) was held at +20 to +50V, while the cathode potential (UB) was
maintained at -50 to 200 V. A copper ring situated above and adjacent to the
matrix window was held at +50 V (Uo) and collected electrons. A
background current of 2 pA was registered which increased to 20 30 pA
during deposition.. Ionization occurred as the result of the vapor-phase
mixture of argon, CCl4, and naphthalene passing through the electron
beam just prior to deposition.
Investigation of UV-vis and IR absorption of matrix isolated cations
was accomplished by ionization of the sample (as discussed) followed by
deposition onto a BaF2 window. Barium fluoride was used since it is
transparent over the 210 nm to 14 gm range and is thus usable in both the
UV-vis and IR regions. For either ionization method the system was
configured in such a way as to establish two unobstructed light paths
perpendicular to each other intersecting at the matrix window (cf. Fig. 9).
The infrared spectra were run on a Nicolet 7199 Fourier transform
infrared spectrometer (typically with 300 scans, 1 cm-1 resolution) while the
UV-vis scans were done on a Cary-17 spectrophotometer (0.2 0.6 nm
resolution over the 240 700 nm range). To allow for sequential scans in the
UV-vis and IR by a simple 900 rotation the sample cryostat was positioned
in a specially-constructed housing as part of the sample compartment of
the Cary-17. The IR beam from the FTIR spectrometer passed
perpendicular to the UV-vis beam, through a set of KBr windows in the
cryostat's outer shroud, reflected off a set of steering mirrors and impinged









onto the detector (TGS or MCT). The whole setup was purged with dry N2 by
use of connecting tubes between the FTIR and UV-vis spectrometers.
Fluorescence and resonance Raman spectra were collected on a
SPEX 1700 3/4 m double monochrometer using a cooled RCA C3034 photon
multiplier tube with standard photon counting electronics. Data collection,
reduction and storage were done on an HP 3000 computer with a 7475
digital plotter. Samples were prepared by pulsed-discharge deposition onto
an angled, polished aluminum block. The fluorescence resonance Raman
spectra of PDCB cation were measured using Argon ion laser excitation
(Spectra Physics, Model 2020), while the N cation resonance Raman
spectrum was measured using a dye laser (Spectra Physics, Model 375B)
pumped by the aforementioned Argon ion laser. The dye employed was
Rhodamine 590.
Photolysis experiments involving N cations were performed using an
unfiltered, unfocused medium-pressure Hg lamp while the same lamp
with a cutoff filter (BG23 VEB JENA, with transmission between 350 and
550 nm) was used for experiments involving PDMOB cations, and a filter
with cutoff X <420 nm was used for PDCB cations.
FTICR/MS experiments were performed by Dr. Chris Barshick in the
laboratory of Professor John Eyler at the University of Florida with the
assistance of the author. Preliminary FTICR/MS experiments were
performed by attaching the pulsed-discharge assembly described previously
to the system shown in Figure 10. The continuous glow discharge source
just to the right of the turbo pump was replaced with the pulsed-discharge
source. The pulse valve and ion analyzer cell were controlled by a Nicolet

















0


N












rI3

0M
-a

I;






xr
C4
o0


I-








37






(DI
0







o a


o0-
4- 3 0













o
r.


C)


0
a:1











0 OD







L4


0L
n E _0

Oo2









FTMS-1000 electronics console. To avoid electronic interference a +1.5 KV
potential was applied to the hemispherical grid (greater voltages resulted in
shutdown of the instrument) and +200 V to the cylindrical lens of the
pulsed-discharge source. A series of four electrostatic lenses was employed
to guide the cations from the discharge region toward the ion analyzer cell.
The voltages were adjusted to prevent secondary discharges between each
other and to optimize signal. Proceeding from the discharge region toward
the analyzer cell the lenses were set at -762 V, -511 V, -773 V and -792 V.
Sensitivity was enhanced by selective ejection of ions with m/z less than 40.
Typically, 50 scans were collected and processed.
Further experimental details specific to each molecule studied are
presented within the Results and Discussion chapter.









THEORY


The interpretation of experimental results is aided through
computation of physical observables such as electronic and vibrational
absorption spectra. Increasingly sophisticated levels of theory have become
available through a variety of computer program packages, e.g.,
GAMESS.1' 52 The increase in sophistication (and accuracy of computed
results) is accompanied by an increase in computing time and power
required. An experimentalist must decide what level of theory is required
for their particular applicationss. To make this decision the limitations of
both experimental and computational results must be understood so that
judicious use of computing facilities is achieved. The following outlines
limitations associated with the calculation of electronic and vibrational
absorption spectra of the radical cations (2M+) studied.
Photoelectron (PE) spectroscopy of neutral molecules has been
employed to predict the electronic absorption (EA) spectra of 2M's derived
from the parent molecule. The theoretical link between PE spectra of
neutral molecules and the electronic structure of 2M* species lies in
Koopman's theorem53 54 which states that the energy of a molecular orbital
(MO) of a neutral molecule obtained from a Hartree-Fock (HF) procedure is
equal to the negative of the energy required to remove an electron from that
MO. By measuring the energy differences between PE bands observed the









energies of electronic transitions of 2M*s can be estimated. There are
limitations, however, in that only states that involve ejection of one electron
from one MO, i.e., no simultaneous ejection of an electron and promotion of
another electron to another MO, are observed in PE spectroscopy. The four
general types of electron configurations possible for 2Ms are given in
Figure 11. Of these, only G and A-types are the observable Koopmans
configurations while B and C-types are non-Koopmans configurations (not
observable in PE spectra). The G-type is the ground state configuration, the
A-type involves electron promotion from a doubly to singly (d-+s) occupied

MO while types B and C involve promotion to virtual (v) MOs. A major
limitation is that all of these states are potentially observable in EA
spectroscopy and, therefore, not all EA bands observed can be accounted for
by energy differences in PE bands observed.
The presence of EA bands in the visible region for radical cations has
usually been explained by the availability of an electron "hole" in the
highest occupied molecular orbital (HOMO) generated upon ionization of
the neutral parent. This allows for low energy transitions from doubly
occupied MOs to this singly occupied HOMO not available to the neutral
molecule. The lowest energy transition for the neutral molecule must be
from a doubly occupied MO to the lowest unoccupied molecular orbital
(LUMO). However, the HOMO -4 LUMO gap in the ion may red-shift and

also account for low energy EA bands. Finally, configuration interaction
(CI) may lead to strong interactions between various states of the same
symmetry and result in long wavelength absorption.


























Figure 11


General Types of Electronic Configurations Associated
with 2M Molecular Cations
(adapted from Ref. 14)



















d
E 0 9 0













G A B Ca Cb
(d-: s) (s-v) (d--v)









Hiickel Molecular Orbital (HMO) theory serves as an adequate
starting point in the computation of the electronic structure of planar,
conjugated hydrocarbons. Typically, the order of the lowest energy
transitions are predicted correctly and if they are of differing symmetries
CI will not affect this order but may influence their relative energies by
interaction with higher energy configurations. If a poor correspondence
between experiment and theory is realized higher levels of theory which
incorporate CI (simple HMO theory does not) may be employed.
The accuracy of computed vibrational frequencies and intensities is
limited by computer time and power available. The accuracy required for
comparison to experiment should be considered and a level of theory chosen
that adequately serves the purpose of the investigation. The vibrational
spectra of stable molecules, for which geometries and reasonable force
constants are known, may be computed by Wilson's method of F and G
matrices55 with subsequent refinement (of force constants) by minimizing
the differences between experimental and computed frequencies. However,
since there are usually more force constants than vibrational frequencies a
unique fit is not guaranteed and a variety of force constant combinations
may yield equally plausible results. Ab initio calculations may indicate
which set is the most physically reasonable and, importantly, may be the
only method by which to predict geometries and force constants of transient
species for which no molecular analogies are available. Once the
geometries and force constants of these species are computed Wilson's
method may be applied with greater confidence."6









The harmonic approximation states that the potential energy of a
molecule is a quadratic function of the nuclear displacements of the
constituent atoms during the course of a vibration. This is shown in Figure
12 (dashed line) for a diatomic molecule in which the nuclear displacement
is simply compression and elongation of the bond about the equilibrium
bond distance. The actual or anharmonic potential energy is also shown in
Figure 12 (solid line). At small displacements the harmonic approximation
holds but deviates at larger displacements. Since the quadratic force
constant is the second derivative or curvature of the potential well surface
with respect to mass-weighted cartesian displacements, evaluated at the
equilibrium nuclear configuration,57 use of the harmonic approximation
typically over-estimates force constants and subsequent frequencies.
Although anharmonicity can be accounted for,"accumulated experience
suggests that the accuracy of calculated vibrational frequencies is affected
more by the quality of the calculated potential than by the inclusion of
anharmonicity."56 In Figure 13, it is shown that as the size of the basis set
increases for each Self Consistent Field (SCF) calculation with the eventual
inclusion of electron correlation (dotted to dashed to solid line) departure
from the "experimental" curve decreases. Hence, even if the harmonic
approximation is employed over-estimation of computed frequencies will
decrease as sophistication of the calculation increases. Frequencies are
typically over-estimated by -10% and various scaling factors are used to
align experimental and computed frequencies.5" A constant scaling factor
was applied in this research but others have used mode-by-mode methods of
scaling.


























Figure 12

Fixed Nucleus Potential Well Calculation (solid line) and
its Harmonic Approximation (dashed line)













I I
r I
I I
I I
I: I


Bond Length


Hl

Q



























Figure 13

Effect of Basis Set and Level of Theory on Potential Well Calculations















*"




*s


...... = SCF, small basis set
----- = SCF, large basis set
-- = Large basis with
electron correlation


Bond Length >


II



SI
s)
H:






49


The emphasis of this chapter was to delineate the limitations of
computational methods employed during the course of this research.
Other, pertinent details concerning theoretical aspects of this research will
be presented in the context of the Results and Discussion chapter which
follows.










RESULTS AND DISCUSSION
Para-Dichlorobenzene


The material presented here consists of frontier studies done on the
vibrational spectra of para-dichlorobenzene,59 para-dimethoxybenzene,60
and naphthalene61 cations as appear in articles published during the
course of this research. This investigation has been extended to perylene62
and anthracene63 cations as part of an ongoing program to systematically
study the vibrational spectra of PAH cations.
The radical cation of PDCB was chosen for the initial investigation
since its electronic absorption64 (10 K, Ar matrices), gas phase electronic
emission,65 and resonance Raman66 (77 K, freon mixture matrix) spectra
have been reported previously, but its infrared spectrum has not been. The
electronic absorption, fluorescence, resonance Raman and infrared spectra
of the PDCB radical cation in an argon matrix at 12 K are reported here.
Also, the results of an intermediate neglect of differential overlap
parametized for spectroscopy (INDO/S) calculation of its electronic excited
states and an ab initio Hartree-Fock Self Consistent Field (SCF) level
calculation of the harmonic frequencies of neutral and cationic PDCB are
presented.
The ultraviolet (UV) absorption spectrum of neutral PDCB in argon
at 12 K appears in Figure 14 with the lowest energy transition occurring at
280 nm. Two band systems between 550 and 280 nm become apparent














00



cIq
.6


ww





o b
ca









E-4
ho~a


cq
Z




i i





o M





~~ 0l


a)












































0
co






I I I 1 I I I I I
- 0 o


0
O





0


0
In
nLO

o






o




E
0 C5
T I
H-
z
w
-J
_.1
0 m
(YI


33NVaaOSV


e

ago









after deposition of the pulsed-glow discharge plasma. Friedman, Kelsall
and Andrews6 (FKA) had previously observed bands which correspond
very closely with the prominent bands at 520 and 322 nm and had assigned
them to the PDCB cation. The 520 nm band system was assigned to the
&2Bj 4- 2B allowed transition.64 65 FKA assign all the vibronic bands in

the 520 440 nm region to one electronic transition and those in the 322 283
nm region to another transition. Based on the known photoelectron
spectrum of PDCB65 the 520 440 nm region is reassigned to contain two
electronic transitions. Table 4 lists the vibronic bands of the PDCB cation
along with their assignments while a comparison of optical absorption
bands and photoelectron bands is presented in Table 5. The energies of the
observed optical bands match very well with those predicted by subtraction
of the ionization energies of the ground and excited states of PDCB.
The infrared (IR) spectrum of neutral PDCB (in an Ar matrix at 12
K) is shown in Figure 15 (bottom), exhibiting prominent bands at 820, 1012,
1090 and 1480 cm-1. The IR spectrum of the same sample matrix showing
the 520 and 322 nm optical bands discussed earlier is presented in Figure 15
(top). Bands not observed for neutral PDCB appear at 843, 986, 1110 and 1429
cm-1. If the newly observed IR bands are due to the PDCB cation and the 520
nm optical band has been correctly attributed to the same species, then a
linear correlation should exist between the relative intensities of the optical
and IR bands. To establish this relationship a series of experiments were
performed with varying deposition times, pulsed-glow discharge pulse
rates and discharge configurations. The ratio of the 520 nm band









Table 4


Visible/UV Absorption Bands of PDCB Cation in Ar Matrix at 12 K


X/nm P/cm-1 AD/cm-1 Assignment


528.0
520.5
511.7
503.8
498.6


490.6
482.6
475.9
467.8
456.0
449.6


322.4
315.0
311.5
308.1
305.0
301.3
298.5
295.2
292.3
283.5


18 939
19 212
19543
19 849
20056


20383
20721
21013
21376
21930
22242


31017
31746
32102
32457
32787
33189
33501
33875
34602
35273


-273
0
331
637
844


0
338
630
993
1547
1859


0
729
1085
1440
1770
2172
2484
2858
3585
4256


Site
Origin:
V6
2v6



Origin:
V6
2V6
3v6
v12;
V2 + V6


B2B3u


D2B3g


Origin:
V5
V4
2v
V5 + V4
31V5
V4 + 2V5
4v5
5v5 or 2v4 + 2v,
6vs








Table 5

Comparison of the PDCB Cation Optical Absorption
Transitions with the Photoelectron Bands


Optical
transition in
State 1Ei / eVa 1Ei IE(2B2g) Ar matrix

X2B2g 8.98 0.0


A2Big 9.87 0.89 eV Not obs.;
(1393 nm, 7178 cm-') Forbidden


t2B3, 11.36 2.38 eV 2.38 eV
(520.9 nm, 19 195 cm-1) (520.5 nm, 19 212 cm-1)


C2BU 11.49 2.51 eV 2.53 eV
(494.0 nm, 20 243 cm-') (490.6 nm, 20 383 cm-1)


S12.8 3.82 eV 3.84 eV
(324.6 nm, 30 808 cm-1) (322.4 nm, 31 017 cm-1)


a From Ref. 65.
















cq


*-I














'O
0











CZ
0
ca




























E-4
II

a)


Ca




u



'I






a






57

0
0
0


U
,,) E
'IT uI

OD






0-- Z0
c I




L0 a.
(7
Eu O z O

S0

c0
0
O 7




e) C'j 1
I I









S0 0 0

33NVd OSV









absorbance to the 1429 cm-1 band absorbance was found to be consistently
-16. A correlation diagram for the four IR bands suspected of being due to
the PDCB cation is presented in Figure 16. For the eight different
experiments performed, an excellent correlation was found. To further test
the validity of this method, correlation of the 520 nm band absorbance to the
neutral PDCB IR band absorbances observed over this series of experiments
was checked and, as would be expected, was found to be poor. Another IR
band at 951 cm-1 (marked as an inverted triangle in Figure 15) was observed
upon pulsed-glow deposition but it did not correlate with the 520 nm PDCB
cation band and thus is unassigned at present.
Further confirmation of assignment of the four IR bands to the PDCB
cation was established through a photolysis experiment. A pulsed-glow
discharge matrix was irradiated using a medium pressure Hg lamp (with
a X<420 nm cutoff filter) for 20 minutes. This resulted in an approximately
twofold decrease of both the 520 nm band and the four IR band intensities,
while the 280 nm neutral PDCB band intensity remained essentially
constant. Therefore, the four IR bands initially suspected of being due to
the PDCB cation, can be conclusively assigned to this matrix species.
Shown in Figures 17 and 18 are the fluorescence and resonance
Raman spectra of PDCB cation. Both broad (fluorescence from the 92Bg

state) and sharp (resonance Raman lines) features are apparent in the
spectrum shown in Figure 17 (excited by 10 mW, 514.5 nm Ar ion laser
radiation). Previous Raman work by Kato, Muraki and Shida66 is compared
to the present results in Table 6. The broad bands from Figure 17 are
analyzed in Table 7 and it can be seen that the vibrational intervals










U




co
0




e0






0





o

Q
0




C4-
o
- U














CO
o
S










0
sed

F4-

a








0
U)
0
CL

'.3








$4
0

o
*^
&,

Iw
o
3l
*a


k

0
CM










ojE
o o
.L

(QU
oz
oco
E EEE E -
Su U 0
CO0 O qo co

S0 0 o


(D I Vd 0 V

,,oix(H I)33N vado0S 9V










0



C U
CU 1





'-I
4) 4
02 C.
CU 0

-G 9
42k

4)






SCU



r CU 02

'-4







CU d
00




CU U



CU 0


4) t,
U c.

El CU
4)
U ..
02E

4)





E.4












e
606


96L-


o
(1
\
U c)


861L


69 6LL 1t7t7L

68LL crl


AIISN31NI


I
E
u

Iz7


















as
0






00
E4










cl,



co
bi)
L40











C.)




cIE
a)i
1-
0O


















0

0






o











0 <



0
LO
0












co
0
00 0








0

m Cu o
II3N
r --, -- ",- O








Table 6


Resonance Raman Bands of the PDCB Cation in Ar Matrix at 12 K
(AP/cm-'1)

In freon mixture, 77 K, In argon matrix, 12K, pulsed-
y irradiation (P/cm') glow discharge (/cm1) Assignment

330 V6
666 660 2v6
990 990 3v6
1115 1113 V4
1195 1189 13
1447 1443 v4+ v6
1601 1598 V2
1932 1928 v2 + V6
2258 v2 + 2v6
2714 2711 V, + v4
3196 2V2


a 538.27 nm laser excitation; Ref. 66.
b 514.5 nm laser excitation; present work.










Table 7


Fluorescence Bands of PDCB Cation in Ar Matrix at 12 K


514 nm exc. 488.0 nm exc.

A Series B Series

A / cm' v/cm' A'v/bcml A i/'cm-' w/cm-' A ~dcm-1 A 7Ccm"' vicm-1 A /ecm-' Assign'

384 19062 0 1402 19089 0 1619 18873 0
716 18720 332 1704 18789 300 1947 18545 328 vs
1040 18396 332+324 2027 18465 300+324 2294 18198 328+347 2vs
1485 17951 1101 2508 17984 1105 2766 17726 1147 V4
1988 17448 1604 2989 17503 1586 3231 17260 1612 V2
2318 17124 1604+324 3304 17188 1586+318 3564 16928 1612+332 V2 + V6
2643 16793 1604+324 3632 16860 1586+318 V2 + 2v6
+331 +328
3094 16342 1604+1106 V2 + v6

a Displacement from laser excitation line at 514.5 nm (19 436 cm-').
b Displacement from 19 052 cm-' (0-0) band.
c Displacement from laser exc. line 488.0 nm (20 492 cm-1).
d Displacement from 19 089 cm-1 (0-0) line (A series origin).
e Displacement from 18 873 cm-1 (0-0) line (B series origin).
f Assignment for all three columns: 514.5 nm excitation, 488.0 nm excitation; series A
and B.









(compared to the Raman lines observed) match quite closely. Only broad
band emission is observed upon excitation by the 488 nm Ar ion laser line
(cf. Fig. 18 and Table 7). The spectrum in Figure 18 indicates the presence
of "extra" fluorescence bands (labeled as series B) as compared to those
apparent in Figure 17. The existence of two distinct sites in the matrix,
giving rise to the 520 (A series) and 528 (B series) nm optical absorbances,
can account for this observation. Simultaneous excitation of PDCB cations
in both sites by the 488 nm radiation could occur resulting in dual emission.
Also, FKA have assigned the 528 nm absorption as a "site" band. Excitation
by the 514.5 nm line of the B site is apparently less probable.
Theoretical calculations were performed on the excited electronic
states of PDCB cation as an aid to the confirmation of previously assigned
transitions and/or assignment of unknown transitions observed (cf. Fig.
14). The geometrical structures for neutral and cationic PDCB were
optimized using the INDO/1 method.67'8 Bond distances obtained for
neutral cationicc) PDCB were rc.H = 1.093 A (1.101A), rc.ci = 1.777 A (1.738
A), rc.c = 1.381 A (1.408 A) for the C-C bonds adjacent to the C-C1 bonds and
r'c.c = 1.395 A (1.357 A) for the central C-C bonds.
The INDO/S model69 70 was used to determine the electronic spectra
of both neutral and cationic PDCB. A IB2u (3 x 104) state at 37,500 cm-1, a
Ba, state at 39,100 cm-1, and a 1B3, (0.0177) state at 40,200 cm-1 (calculated
oscillator strength in parentheses) were computed for neutral PDCB. The
lowest energy band with ma, = 280 nm (=35, 760 cm-1) shown in Figure 14 is
in generally good agreement with the weakly allowed 'B2u state computed to

occur at 37,500 cm-1. For the PDCB cation, since the calculation (which









includes configuration interaction) refers to the 2B2g ground state, the

calculated values should be compared to the band maxima in Figure 14.
The calculated spectrum for the cation is shown in Table 8 and the order of
low-lying states is in accord with that suggested earlier.65
The first excited state of the cation is a 2Big state calculated at 0.68 eV;

it is observed in the ionization but not the absorption spectrum. The
2B, +- 2B2g allowed transition is calculated to be 0.5 eV higher in energy

than that observed (2.89 instead of 2.40 eV). The first of two possible
assignments for the higher-lying transitions is suggested in Table 8. The
assignment of A through 05 is in agreement with previous assignments65
and is also consistent with observed intensities (cf. Fig. 14). The observed
features at 2.72 and 3.97 eV are assigned to D (2B3g) and E (2B,),

corresponding to computed energies of 3.23 and 3.41 eV, respectively. The
second suggested assignment results from the observation both 2B3u at 2.89
eV and 2B2u at 3.14 eV are calculated to be 0.5 0.6 eV higher in energy than

observed. If this is a consistent trend then the E state might be assigned to
the third calculated 2B3u state at 4.26 eV while the broad, structureless
feature observed at 2.72 eV could be assigned to the second calculated 2B3u

state at 3.41 eV. Optical polarization data would aid in distinguishing
between these two possibilities.
To determine the effect of ionization on the vibrational force constants
of specific bonds in PDCB, an ab initio Hartree-Fock SCF level calculation
(6-31 G basis; for open shells, ROHF) was performed. Initially, the
geometries of the neutral parent and the radical cation were optimized
using the GAMESS program.51 52 The optimized geometries found are










Table 8


Calculated and Observed Spectra of PDCB Cation (Energies in eV)


Observed (cf. Table 5)
A/Ea Opt. (Ar
matrixb)


0.89


2.38 2.40
2.51 2.58
2.72
3.82 3.97


Calculated (INDO/S)
E (Koopmans) CI Assignment
neutral catione, d


0.20
2.65
2.45
2.55
3.35


0.68
2.89(0.156)
3.14
3.23
3.41(0.081)
3.52
4.26(0.054)
4.87(0.185)
4.88(0.000)
5.31(0.094)


2Big

2B3u
2B2U

2B3
2B
2A9
2B3U
2Au

2Blu
2A,


a The first IE of XiB is 8.98 eV. All others are relative to this; see Ref. 65.
b See Fig. 14. These are estimated vx.
c Relative to the first molecular orbital eigenvalue of 9.74 eV.
d The numbers in parentheses are calculated oscillator strengths from the
2Big ground state of the cation.
e The first IE is calculated at 9.00 eV. All others are relative to this.


YB
f2


E









(radical cation figures in parentheses) rc.H = 1.070 A (1.070 A), rc-c = 1.805
A (1.742 A), rc.c = 1.380 A (1.419 A) for the C-C bonds adjacent to the C-Cl
bonds, and r'c.c = 1.388 A (1.359 A) for the central C-C bonds. The values
are in good agreement with those found using the INDO/1 method. The
relative decrease of the rc.c, and r'c.c lengths and the increase of the rc.c
adjacent bonds is consistent with the removal of an electron from the
highest occupied molecular orbital (HOMO) of PDCB. This x molecular
orbital (cf. Fig. 19) is antibonding with respect to the C-Cl and central C-C
bonds and bonding with respect to the adjacent C-C bonds.
Harmonic vibrational frequencies for PDCB neutral and radical
cation species were calculated using the GAMESS program. Vibrational
frequencies calculated at the SCF level are typically overestimated by -10%
because of the harmonic approximation used and the neglect of electron
correlation effects.56-58'71 To account for this, a scaling factor of 0.9 has been
adopted to adjust the calculated frequencies. Previous experience has
shown that the match with experimental vibrational frequencies using
such a scaling factor is usually 50 cm-1.56"58, 71 Calculated, scaled, and
experimental frequencies for neutral PDCB72-75 and its radical cation are
given in Tables 9 and 10, respectively. Good agreement is found between the
scaled and experimental frequencies for both species. The observation that
the lower-frequency modes need not be scaled as severely as the higher-
frequency modes is in agreement with a SCF calculation (6-31 G* basis)
done by Rohlfing et al.75 The presence of polarization functions (d orbitals)
on carbon and chlorine in 6-31 G* generally decreases the calculated out-of-
plane mode frequencies (a., bg, b2g, and b.) relative to the values obtained



























Figure 19


Schematic of HOMO (Highest Occupied Molecular Orbital) of PDCB Cation
(adapted from Ref. 66)





72
z








Table 9


Calculated and Observed Vibrational Frequencies (cm-1)
for Neutral PDCB


Mode Sym. Cale. Scaled (0.9) Expt.a


3403
1777
1329
1186
804
344
1169
478
987
3385
1668
1208
1135
554
1168
826
333
3400
1533
1331
1228
223
3386
1783
1479
712
375
998
577
112


3063
1599
1196
1067
724
310
1052
430
888
3046
1501
1087
1021
499
1051
743
300
3060
1398
1198
1105
201
3047
1605
1331
641
337
898
519
101


3072
1574
1169
1096
747
328
951
405
815
3078
1477
1090
1015
550
934
687
298
3087
1394
1220
1107
226
3065
1577
1290
626
350
819
485
122


a Reference 72 75.









Table 10


Calculated and Observed Vibrational Frequencies (cm-1) for PDCB Cation


Mode Sym. Calc. Scaled (0.9) Expt.a


3417
1814
1354
1204
801
346
1190
419
947
3402
1596
1247
1178
575
1181
796
273
3415
1624
1423
1070
238
3402
1553
1336
622
385
997
558
89


3075
1633
1219
1084
721
311
1071
377
852
3062
1436
1122
1060
517
1063
716
246
3073
1462
1281
963
214
3062
1398
1239
560
346
897
502
80


1598
1189
1113

330




1429

1110







986






843


a Present work.









with the 6-31 G basis. This may be due to the ability of the polarization
functions to better accommodate charge build up in the n system during
out-of-plane motion. Interestingly, the calculated IR intensities indicate
that the two bl, modes (1429 and 1110 cm-1) of the ionic form should be the

most intense and indeed these are the most prominent experimental ionic
bands.
In attempting a mode-by-mode comparison between the neutral and
ionic species, problems arise. One expects a simple inverse relationship
between bond lengths and stretching force constants. However, normal
coordinate frequency shifts cannot be so simply described because coupling
of various internal coordinates varies as a function of the geometry changes
and the force constant changes which result from ionization. Nevertheless,
comparisons can be made by first coupling calculated neutral and ionic
vibrational frequencies by symmetry and then by energy, within each
symmetry block. Finally, after scaling, the energy shifts can be compared
to the experimental values.
Calculated and scaled neutral-to-ionic frequency shifts are compared
in Table 11. The direction of the frequency shift is correctly predicted for all
observed bands except for the b3, mode. This mode differs from the others

since it is the only out-of-plane mode observed. The shifts for the Raman-
active totally symmetric modes match better than the IR-active modes.
This may be due to the greater interaction with the host matrix of
transitions involving a dipole moment change with subsequent polarization
of the matrix.








Table 11


Calculated and Observed Vibrational Frequency Shifts
Upon Ionization of PDCB


Shift (ion -V neutral; cm-1)
Mode Sym.
Calc. Scaled (0.9) Expt.

2 ag +37 +33 +24
3 +25 +22 +20
4 +18 +16 +17
6 +2 +2 +2
11 bl, -72 -65 -48
13 +43 +39 +95
21 ba -158 -142 -121
28 b3 -1 -1 +24









In order to discuss the effect of ionization on specific bonds, it is
necessary to transform the normal coordinates to internal coordinates.
Boatz and Gordon76 have introduced such a vibrational decomposition
scheme in which symmetry-adapted normal coordinate frequencies are
transformed into intrinsic frequencies. These latter represent the sum of
the contributions of all normal modes of a vibration to a particular internal
coordinate. Calculated intrinsic frequencies may vary depending on the
specific internal coordinates used, but usually do not differ by more than
~10 cm-1, particularly for bond stretches. Only intrinsic stretching
frequencies and force constants are presented here. The quantities
calculated for PDCB, using the method of Boatz and Gordon, are given in
Tables 12 and 13, respectively. It may be seen that, upon ionization, the C-
Cl force constant rises from 2.509 to 3.237 mdyn/A, the central C-C force
constants also rise from 6.002 to 7.289 mdyn/A, while the adjacent C-C force
constants drop from 6.237 to 4.577 mdyn/A. From the optimized geometries,
described previously, it has already been noted that, upon ionization, an
increase in the C-C1 and central C-C and decrease in the adjacent C-C bond
strengths is expected. This expectation is fully borne out by the calculated
intrinsic frequencies and force constants described here.
A similar trend was also found by Ernstbrunner and coworkers46 in a
normal coordinate analysis of neutral and cationic para-
dimethoxybenzene. They found the ring C-O force constant increased 25%
(compared to 29% for C-Cl in PDCB), while the central C-C bond force
constant increased 6% (20% for PDCB), and the adjacent C-C bond force
constant decreased 10% (compared to 25% for PDCB).









Table 12


Calculated Intrinsic Stretching Frequencies
and Force Constants for Neutral PDCB


Frequency (cm-1) Force constant (mdyn/A)
Bond
Calc. Scaled (0.9) Calc. Scaled (0.9)

C-H 3388 3049 6.287 5.658
C-Cl 728 655 2.788 2.509
C-Ca 1373 1236 6.669 6.002
C-C -1400 1260 -6.930 6.237

a The central C-C bonds.









Table 13


Calculated Intrinsic Stretching Frequencies
and Force Constants for PDCB Cation


Frequency (cm-1) Force Constant (mdyn/A)
Bond
Calc. Scaled (0.9) Calc. Scaled (0.9)

C-H 3401 3061 6.335 5.701
C-C1 827 744 3.597 3.237
C-Ca 1514 1363 8.099 7.289
C-C -1198 1078 -5.086 4.577


a The central C-C bonds.









The aforementioned intrinsic frequency analysis makes clear why
certain bands blue shift and others red shift upon ionization. Consider
modes 2 and 11: examination of the eigenvectors for neutral PDCB indicate
they are associated primarily with central C-C and adjacent C-C bond
stretching motions, respectively. Since the force constant for the former
increases and for the latter decreases, the blue shift of 24 cm-1 for mode 2
and the red shift of-48 cm-1 for mode 11 is expected. The ratio of the
absolute magnitudes of the frequency shifts for modes 11 and 2 is two and
this tracks exactly the number of C-C bond stretches involved in each mode,
i.e., there are twice as many adjacent as there are central C-C bonds. This
is also consistent with the removal of an electron from the HOMO which is
bonding with respect to the adjacent C-C bonds but antibonding with respect
to the central C-C bonds (cf. Fig. 19). For further clarification the Raman-
active (ag) and IR-active normal modes (cf. Table 11) observed for PDCB
cation are sketched in Figures 20 and 21, respectively.
The FTICR mass spectrum of pulsed-glow discharged para-
dichlorobenzene seeded in Ar is shown in Figure 22. Prominent peaks are
observed at (relative intensities in parentheses) 186 (18), 146 (42), 111 (95), 75
(50) and 40 (100) amu. Isotopic 37C1 and 35C1 multiplets are associated with
the first three peaks and are assigned as the chlorine-containing species
ArC6H4Cl2+, C6H4C12+ and C6H4C1, respectively. The 75 amu peak
corresponds to C6H34 while that at 40 amu results from incomplete ejection
of Ar*. The preponderance of non-parent ion species (C6H4C1* and C6Hs*)
present in this mass spectrum provokes the question: Is the parent ion the
sole species present in the matrix under study? Spectroscopic evidence for



























Figure 20

Raman-Active (ag) Normal Modes Observed for PDCB Cation
(clockwise from upper left: modes 2, 3, 4 and 6,
same orientation as in Fig. 19, bonds are removed for clarity)







Mode 2


Mode 2
0


o9o?o0
O


Mode 4
0
Cr
0 Q0

0


Mode 3
0

S0

0


























Figure 21

Infrared-Active Normal Modes Observed for PDCB Cation
(clockwise from upper left: modes 11, 13, 21 and 28,
same orientation as Fig. 19 except mode 28 which is viewed
along the C-C1 axis, bonds are excluded for clarity)







Mode 11
0




0


Mode 13
0




0


Mode 21
0


-0 O-
6-0Q 0
0






















'*


Qe

-o




0


C4-
0

.1






cfl

0
H





00



SC
(0





-o
*^ _.
-----------------------------------------------------------= -
*^




-- = -

o.C


OS
1ISN31NI 3AIlV93a


001


0





0



0x



K
z3









PDCB neutral and/or ionic fragments in the matrices investigated is absent
but the possibility of fragmentation appears in the study of para-
dimethoxybenzene and naphthalene videe infra). Recombination of
fragments, e.g., C6H4C1+ + Cl (Cl) -4 C6H4C12+ (C6H4C12) may occur as the

matrix forms, regenerating parent, cationic (neutral) PDCB, and changing
the matrix composition (relative to the gas phase). Because of this
possibility, correlation of mass spectra with subsequent spectroscopic
investigation of matrix isolated species is problematic. Emphasis has been
placed, rather, on correlating different spectroscopic observations, e.g.,
UV-vis and IR spectra, made on the same matrix as well as incorporating
theoretical computations of chemical species under study.
Pulsed-glow discharge/matrix isolation has been employed to obtain
and stabilize quantities of PDCB parent ion sufficient to observe directly
several infrared and Raman-active fundamental vibrational modes.
Correlation analysis of known and well characterized visible bands with
suspected IR absorbances due to PDCB+ coupled with theoretical
computations have allowed assignments to be made with confidence. In
addition, changes in bonding and subsequent shifts in vibrational
frequencies as well as intensity patterns have been addressed. Preliminary
FTICR/MS experiments performed, however, suggest the complexity of the
gas phase plasma formed and invite inquires concerning the composition of
the matrix formed. All this leads to a few very important questions. (1) Is
this ion generation method gentle enough to produce parent ion solely? (2)
If not, how can degradation products be accounted for or can the sought-









after species be selectively investigated in the presence of other matrix
constituents? (3) How does the efficiency of this method compare to
methods previously reported and how might ion production be enhanced?
These questions are addressed now for PDCB and are considered further
during the discussions on para-dimethoxybenzene and naphthalene.
The first two questions are very much related and are answered first.
FKA64 recorded the visible absorption spectrum of PDCB* by two methods:
(1) Filtered high-pressure Hg arc photolysis of the pre-formed matrix
containing PDCB neutral and (2) microwave-powered discharge of flowing
Ar directed at the matrix window during deposition. In either case, the
spectra obtained were identical (except for absolute intensities) and
indistinguishable from those obtained by pulsed-glow discharge.
Comparison of the resonance Raman spectrum presented previously
(in Ar at 12 K) with that reported by Kato, Muraki and Shida6 (in freon
mixture at 77 K with y-irradiation) lends further support for production (or

at least selective probing) of PDCB parent ion only. As shown in Table 6,
the bands observed are very similar despite differences in ion production
method, matrix media and temperatures.
The gas phase emission spectrum reported by Maier and Marthaler65
(produced by controlled electron impact) compares favorably with the
matrix fluorescence spectrum obtained (cf. their Fig. 1 and our Fig. 17),
exhibiting vibronic structure involving 330 and 1600 cm-1 modes. The vapor
phase and matrix origins occur at 19,620 and 19,130 cm-1, respectively. A
red gas-to-matrix shift of -500 cm-1 is not unreasonable for solvation of the
ion. Based on these comparisons, it is concluded that PDCB parent ion is