... n

. a.
.ﬁ;
J"; x.
5

a“?

k...

- ...! -
22?; a
1.1.. N. S <
. ... i. ‘

. .. .. £31,9-.
ﬁg»

A

mm“
5%.... . , ..

.. u...

n:

.2.
a .

L?"

....w
L3

v... a

.Lmﬁ. .
xww...
2.x
, a. a.
wﬁ. ...
..vwn

. «5..., s
a: .

khan“?

rawﬁ
§%

.1
.

 

 

v... .
a .
...n‘t...
z r...r
4. PCs.
. .
.. Y A...
. Fara...
\K‘ ..412 . ~
:=:71~ - ._I_:— . ‘ .
__ :5... x.“ s .. .x._,..u..§...
. : . . .2 . < ‘3 ..

 

 

u ... u
11/70.

:5.

THE? . s.

C?

F‘Q

This is to certify that the

dissertation entitled

UNDERSTANDING INTRAMOLECULAR AND

INTERMOLECULAR CONTRIBUTIONS

TO ENERGY RELAXATION
presentedby

Scott Nelson Goldie

has been accepted towards fulﬁllment
of the requirements for

Ph. D. degreein Analytical Chemistry

ﬂW

' . (
Major professor

Date 24 January 2001

MS U is an Afﬁrmative Action/Equal Opportunity Institution 0-12771

 

f 4 ”#4 _

 

 

LIBRARY
Michigan State
University

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJCIRCIDdoDue.p65-n15

 

UNDERSTANDING INTRAMOLECULAR AND INTERMOLECULAR
CONTRIBUTIONS TO ENERGY RELAXATION

By

Scott Nelson Goldie

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Chemistry

2001

ABSTRACT

UNDERSTANDING INTRAMOLECULAR AND INTERMOLECULAR
CONTRIBUTIONS TO ENERGY RELAXATION

By

Scott Nelson Goldie

A fundamental understanding of intermolecular interactions is required to achieve
predictability in liquid-phase systems. Using a pump-probe laser spectrometer to detect
stimulated emission, we have studied solution phase local organization by measuring
induced orientational anisotropy and vibrational population relaxation dynamics. We
have studied perylene and l—methylperylene in a series of normal alkanols (methanol to
decanol), and selected aliphatic aldehydes and ketones, to probe local organization in low
viscosity liquids. The reorientation behavior of the chromophores in the alkanol solvents
is different than that seen in n-alkanes. The T1 data for the ring breathing vibrational
modes of these same chromophores points to efﬁcient and solvent-dependent relaxation.
We can infer from these data the type of local organization characteristic of the alkanols.

In both the aliphatic aldehydes and the ketones, there is a change in the nature of
the solvent-solute interactions with both probe molecules as the hydrodynamic volume of
the solvent molecules approaches that of the probes. The solvent-dependent change in
behavior occurs for different solvent aliphatic chain lengths in the two systems.
Vibrational population relaxation measurements for two probe molecule vibrational
modes reveal efﬁcient intermolecular coupling in all cases. The data point to the role of

the solvent aldehyde proton in mediating solvent-solvent and solvent-solute interactions,

leading to chromophore dynamics reminiscent of those seen in the n-alcohols.
Reorientation of the two probes in the ketones yields results that are more consistent with
those seen in the alkanes, suggesting the relative weakness of dipole-induced dipole
interactions relative to H-bonding interactions in the solvation of perylene and 1-
methylperylene.

We have also applied transient spectroscopies to the characterization of polymer
curing for a deep-UV, chemically ampliﬁed polymer photoresist used in the fabrication
of semiconductors. We characterized changes in polymer free volume as a function of
cure using ground state recovery measurements of crystal violet, a probe molecule doped
in the polymer matrix. These data, in conjunction with linear spectroscopy, will
ultimately be used to provide insight into the diffusional behavior of protons in the

polymer matrix as a function of cure.

To everyone who believed in me before I knew enough to believe in myself...

iv

ACKNOWLEDGMENTS

My sincere thanks and genuine appreciation goes to...

My entire family, especially my parents, Don and Virginia, my brother Shawn
and his wife Andrea, for their love and support.

Michelle for her love, support and understanding; I wish her own success in her
goals, dreams and aspiration.

To Micheal, Kathy, Allison Marie, and the entire May family for their love and
support.

Professor Julie L. P. Jessop and Chair Alec B. Scranton, for the opportunities
provided by the photoresist project.

The Blanchard research group members, both past, Punit, Wendy, Jennifer, Patty,
Jeff and Selezion and present, Stephen, John, Lee, Shawn, Jay and Joe for their direct and
indirect contributions to this dissertation.

My committee, consisting of Professors Dantus, Berglund and Cukier, for their
constructive comments and guidance.

Joseph S. Ward, 111 for providing guidance and invaluable information to the ﬁrst
chapter of this dissertation, and for showing me a thing or two computing and chemistry.

Geoff Hoekstra, Dr. Tom Cullen, Dr. Tom Carter, Dr. Emily Brown, and Dr. Eric
Saari for assistance in the design of the LabView® virtual instrument presented in the
Appendix.

The past and present members of the Machine Shop, Glass Shop, and Electronics

Shop, including Sam, Glenn, Tom, Dick, Richard, Scott, Manfred, Ron, Scott and David;

all were helpful with technical information, a quick wit, or a quick turnaround for
projects when needed.

My students, all of them, have made an indelible impression on me, as I hope I
have on them. Thanks should especially go to Ilene Cohen, Troy M. Ferguson, Allison
Fouch, Lori Hallman, and Edmund Tillet, who gave me an encouraging word, accolade
or accomplishment that motivated me to continue working toward my goal. Dr. Paul
Hunter, Mr. Steve Poulios, and Ms. Wendy N. Whitford have supported the teaching
endeavors that I have undertaken during the end of this work. Their encouragement and
assistance has allowed me to improve as an instructor as I completed this dissertation.

Professor David S. Karpovich, for his genuine friendship and opinions that helped
to bring things back into perspective. I am blessed and honored to be his ﬁiend and him
to me. May we continue to “ﬁsh” for more than just salmon.

Professor Gary J. Blanchard for the science, for the encouragement, and for the
guidance, but more importantly for the boat rides in the Cadillac, for the coffee, for the
jokes, for the politics and for knowing when to kick me in the teeth and knowing when it
was enough. The sunrise over the marsh in St. Charles or in Ovid helped to bring me
back to center, clear my head, and allow me to “hunt” for a higher level of understanding
and knowledge. It is those memories that I will have for the rest of my life, and for this
reason I am deeply indebted to him and his family. Thank you Gary, for being so much

more than a research advisor.

vi

TABLE OF CONTENTS

List of Tables ..................................................................................................................... ix
List of Figures .................................................................................................................... xi
Introduction .......................................................................................................................... 1

Literature Cited .............................................................................................................. 22

Chapter 1 — Synthetic Pathways to l-Methylperylene

Introduction .................................................................................................................... 27
Synthetic Scheme 1 ........................................................................................................ 30
Synthetic Scheme 2 ........................................................................................................ 33
Literature Cited .............................................................................................................. 39

Chapter 2 - Orientational and Vibrational Relaxation Dynamics of Perylene and 1-

Methylperylene in n-Alkanols: Subtle Differences in Solvent-Solute Interactions

Introduction .................................................................................................................... 40
Experimental .................................................................................................................. 44
Results and Discussion .................................................................................................. 48
Conclusions .................................................................................................................... 74
Literature Cited .............................................................................................................. 75

Chapter 3 — Orientational and Vibrational Relaxation Dynamics of Perylene and 1-

Methylperylene in Aldehydes and Ketones

Introduction .................................................................................................................... 79
Experimental .................................................................................................................. 82
Results and Discussion .................................................................................................. 84
Conclusions .................................................................................................................. 1 14
Literature Cited ............................................................................................................ 116

Chapter 4 — In-situ Quantiﬁcation of Acid Generation in a Deep UV Photoresist Through

the Use of Probe Molecules

Introduction .................................................................................................................. 120
Experimental ................................................................................................................ l 23
Results and Discussion ................................................................................................ 124
Conclusions .................................................................................................................. 1 30
Literature Cited ............................................................................................................ 134

vii

Chapter 5 — In-situ Characterization of Free Volume Generation During Post Exposure
Bake (PEB) in a Commercial Photoresist

Introduction .................................................................................................................. 1 35
Experimental ................................................................................................................ 1 3 7
Results and Discussion ................................................................................................ 138
Conclusions .................................................................................................................. 147
Literature Cited ............................................................................................................ 148
Chapter 6 — Conclusions and Future Work ...................................................................... 149
Appendix ............................................................ 152
Appendix A — National Instruments Labview® Based Data Acquisition Software for
Pump-Probe Laser Spectrometer ............................................................................. 153

viii

LIST OF TABLES

Table l -— Reorientation times and zero-time anisotropies for perylene in the n-alkanols.
The data are the best ﬁt results of the data to the function R(t) = Rl(0)exp( 4/1.) +
R2(0)exp(-t/rz). Times are given in ps and the uncertainties listed are standard
deviations (ilo) for at least six determinations of each quantity. ............................ 55

Table 2 — Reorientation times and zero-time anisotropies for l-methylperylene in the n-
alkanols. The data are the best ﬁt results of the data to the function R(t) ==
R1(0)exp(-t/t1) + R2(0)exp(-t/t2). Times are given in ps and the uncertainties listed
are standard deviations (ilo) for at least six determinations of each quantity. ....... 56

Table 3 — Cartesian components of the rotational diffusion constant, D, for perylene in
the n-alkanols. These quantities and their uncertainties are derived from the ﬁtted
time constants shown in Table 1. .............................................................................. 60

Table 4 — Cartesian components of the rotational diffusion constant, D, for 1-
methylperylene in the n-alkanols. These quantities and their uncertainties are
derived from the ﬁtted time constants shown in Table 2. ......................................... 61

Table 5 — Vibrational population relaxation times, T], for perylene and l-methylperylene
in the n-alkanols. Relaxation times are determined from at least six individual data
sets and the uncertainties are standard deviations. Times are values from ﬁts of the
experimental data to Equation (10) and are given in ps. Times marked with an
asterisk are the results of ﬁts to Equation (11). ........................................................ 69

Table 6 — Reorientation Times and Zero-Time Anisotropies for Perylene in the n—
Aldehydes.a ............................................................................................................... 98

Table 7 — Reorientation Times and Zero—Time Anisotropies for l—Methylperylene in the
n-Aldehydes.a ......................................................................................................... 99

Table 8 — Reorientation Times and Zero-Time Anisotropies for Perylene in the Ketonesa
................................................................................................................................. 100

Table 9 — Reorientation Times and Zero-Time Anisotropies for l-Methylperylene in the
Ketones.a ................................................................................................................. 101

Table 10 — Cartesian components of the rotational diffusion constant, D, for perylene in
the aldehydes.a ........................................................................................................ 103

Table 11 - Cartesian components of the rotational diffusion constant, D, for 1-
methylperylene in the aldehydes.a .......................................................................... 104

Table 12 - Cartesian components of the rotational diffusion constant, D, for perylene in
the ketones.a ............................................................................................................ 105

ix

Table 13 — Cartesian components of the rotational diffusion constant, D, for 1-
methylperylene in the ketones.a .............................................................................. 106

Table 14 — Vibrational population relaxation times, T1, for perylene and 1-
methylperylene in the ketones. ............................................................................... 1 10

Table 15 — Table of sub-virtual instruments used in the pump-probe data acquisition
virtual instrument. ................................................................................................... 170

LIST OF FIGURES

Figure 1 — Schematic of picosecond pump-probe laser spectrometer-used in this work. 5

Figure 2 —~ Experimental I"(t) and I i(t) scans for l-methylperylene in n-pentanol, along
with the instrumental response function. These data are typical of those recorded for
reorientation measurements. ..................................................................................... 10

Figure 3 — Anisotropy function, R(t), for the data shown in Figure 2. The decay is ﬁt to
the function R(t) = R.(0)exp(-t/tl) + R2(0)exp(-t/tz) ................................................ 11

Figure 4 — Prolate ellipsoid with Cartesian coordinate system superimposed. See text for
discussion. ................................................................................................................. 12

Figure 5 — Oblate ellipsoid with Cartesian coordinate system superimposed. See text for
a discussion. .............................................................................................................. 12

Figure 6 — Three level system model for vibrational population relaxation measurements.
................................................................................................................................... 18

Figure 7 — Magic angle data for perylene in methanol with instrument response ﬁinction.
................................................................................................................................... 19

Figure 8 — Laboratory coordinate system applied to perylene. A similar laboratory
coordinate system was applied to l-methylperylene. ............................................... 29

Figure 9 - 300 MHz lH NMR of perylene in d-CH2C12. The structure of the perylene
molecule is shown in the inset. ................................................................................. 36

Figure 10 — 300 MHz 1H NMR of 1-methylpery1ene in d-CHzClz. The structure of the 1-
methylperylene molecule is shown in the inset. ....................................................... 37

Figure 11 -— Absorbance and emission spectra for perylene in methanol. Data has been
normalized for presentation. ..................................................................................... 38

Figure 12 — Absorption and emission spectra for 1-methylperylene in methanol. Data has
been normalized for presentation. ............................................................................. 38

Figure 13 — Absorption and emission spectra for perylene in n-pentanol. Data has been
normalized for presentation. ..................................................................................... 46

Figure 14 — Absorption and emission spectra for l-methylperylene in n-pentanol. Data
has been normalized for presentation. ...................................................................... 47

Figure 15 —— Dependence of perylene and l-methylperylene absorption maxima on solvent
aliphatic chain length. When normalized for wavelength shift relative to their

xi

absorption spectra in methanol, the solvent dependencies are identical for both
chromophores ............................................................................................................ 50

Figure 16 — Prolate ellipsoid with Cartesian coordinate system superimposed. See text
for discussion. ........................................................................................................... 53

Figure 17 —- Oblate ellipsoid with Cartesian coordinate system superimposed. See text for
a discussion. .............................................................................................................. 53

Figure 18 — Experimental I”(t) and I i(t) scans for l-methylperylene in n-pentanol, along
with the instrumental response function. These data are typical of those recorded for
reorientation measurements. ..................................................................................... 57

Figure 19 — Anisotropy function, R(t), for the data shown in Figure 18. The decay is ﬁt
to the function R(t) = R.(0)exp(-t/tl) + R2(0)exp(-t/tz) ............................................ 58

Figure 20 —- Reorientation time constants plotted as a function of solvent viscosity for
perylene (0) and l-methylperylene (I). The reorientation time constants are
identical to within the experimental uncertainty for the two chromophores in the
solvents used. The fast reorientation time is essentially solvent viscosity
independent and the dependence of the slow reorientation time does not depend
linearly on viscosity. ................................................................................................. 59

Figure 21 — Cartesian components of the rotational diffusion constant as a function of
solvent aliphatic chain length for perylene (DZ = O, Dx = O) and l-methylperylene
(D2 = I, DM = D) in the n-alkanols. .......................................................................... 62

Figure 22 — The ratio DZ/Dx as a function of solvent chain length for perylene (O) and l-
methylperylene (I). These data indicate that the environment experienced by both
chromophores is essentially identical and the structural anisotropy in the solvent
cage increases with solvent aliphatic chain length. .................................................. 63

Figure 23 — Vibrational population relaxation times for perylene (O) and 1-
methylperylene (Cl) as a function of solvent aliphatic chain length. ........................ 70

Figure 24 — Magic angle data for perylene in n-pentanol with instrument response
function. See text for a discussion. .......................................................................... 71

Figure 25 — Magic angle data for perylene in n-decanol. Note the difference in the
functionality of the population decay. See text for a discussion .............................. 72

Figure 26 — Normalized absorption and emission spectra of perylene in 3-octanone. ..... 85

Figure 27 — Normalized absorption and emission spectra of l-methylperylene in 3-
octanone. ................................................................................................................... 86

Figure 28 — Experimental I”(t) and I i(t) scans for l-methylperylene in 2-nonanone, along
with the instrumental response function. .................................................................. 89

xii

Figure 29 — Anisotropy function, R(t), generated from the raw data shown in Figure 28.
The decay is ﬁt to the function R(t) = R1(0)exp(-t/t1) + R2(0)exp(-t/tz), with the
results of the ﬁt given in Figure 28. .......................................................................... 90

Figure 30 — Reorientation time constants plotted as a function of carbon chain length for
perylene (0,0) and l-methylperylene (I,C|) in the aldehydes. .............................. 94

Figure 31 — Reorientation time constants plotted as a function of solvent viscosity for
perylene (0,0) and l-methylperylene ( I,E1) in the ketones. The boxed regions
indicate speciﬁc solvent lengths. .............................................................................. 96

Figure 32 — Normalized vibrational population relaxation data (probe polarized at 54.7°
with respect to the pump) and instrument response function for perylene in 3-
octanone. ................................................................................................................. 108

Figure 33 —— Vibrational population relaxation times of the 1375 cm'1 mode for perylene
(O) and l-methylperylene (I) as a function of solvent aliphatic chain length in the
ketones. ................................................................................................................... 111

Figure 34 — Vibrational population relaxation times of the 1733 cm'] mode for perylene
(O) and l-methylperylene (I) as a function of solvent aliphatic chain length in the
ketones. ................................................................................................................... 112

Figure 35 — The linear response of the three-protonation states of crystal violet. The acid
concentration is presented near the maximum of each absorption spectrum .......... 125

Figure 36 — a value calibration curve for crystal violet as a function of pH. The a values
provide concentration data by relating absorbance. ................................................ 131

Figure 37 — Changes in the steady state absorption response of crystal violet doped into a
polymer ﬁlm as a function of processing. See text for a discussion. ..................... 132

Figure 38 — Changes in pH in-Situ as a function of processing time. ............................. 133

Figure 39 — Instrument response function and ground state recovery measurement of
crystal violet in unprocessed spin cast ﬁlm. The solid line through the data is the
best-ﬁt double exponential decay function. ............................................................ 140

Figure 40 — Ground state recovery dynamics of crystal violet in ethylene glycol solvent.
The instrument response function is also provided for reference. .......................... 142

Figure 41 — Ground state recovery dynamics of crystal violet in glycol solvent. The
instrument response function is also provided for reference. ................................. 143

Figure 42 — Viscous matrix calibration curve relating viscosity of media to ground state
recovery of crystal violet. The line presented is the linear best ﬁt to the data ....... 144

xiii

Figure 43 — Instrument response ﬁmction and ground state recovery behavior of crystal
violet in processed (post application bake, UV exposure, post exposure bake) and
unprocessed (post application bake only) ﬁlms. ..................................................... 145

Figure 44 — Pump-probe laser spectrometer schematic. ................................................. 158
Figure 45 — Front panel of LabView® pump-probe data acquisition virtual instrument. 159

Figure 46 —— Complete circuit diagram for LabView® pump-probe data acquisition virtual
instrument. .............................................................................................................. 160

Figure 47 — Complete circuit diagram of pump-probe data acquisition virtual instrument,
showing second frame of secondary sequence loop. .............................................. 161

Figure 48 — Hi gh-resolution stage forward function ....................................................... 162

Figure 49 — Final frame of secondary sequence loop nested in ﬁrst frame of primary
sequence loop. ......................................................................................................... 163

Figure 50 — Alternatives for electronic time constant settings in secondary sequence loop

panel presented in Figure 49. .................................................................................. 164
Figure 51 — Second frame of primary sequence loop, with low resolution move stage \

forward function ...................................................................................................... 165
Figure 52 - Hi gh-resolution stage rewind function. ....................................................... 166
Figure 53 - Third frame of primary sequence loop. ........................................................ 167
Figure 54 — Fourth (ﬁnal) frame of primary sequence loop. .......................................... 168

Figure 55 — Alternative to true/false state function in Figure 54. This frame is
intentionally left empty, so that no processing occurs. ........................................... 169

xiv

INTRODUCTION

Understanding solvation has been a subject of considerable interest because of the
effect that solvent identity has on chemical processes and reactions. To have any hope of
controlling these processes and reactions, a fundamental understanding of the molecular
interactions, including dynamics and the energetics, that proceed between a solute
molecule and adjacent solvent molecules is necessary. Energy relaxation can occur
within a molecule initially, or between two molecules by electronic, vibrational,
rotational, or translational (collisional) channels.

The liquid phase is a widely used medium for chemical manipulation and
reaction. A further understanding of intermolecular interactions in this state would be of
value to the chemical community. Working in liquid media holds inherent challenges to
examining energy transfer processes. Interactions between solutes and their surrounding
solvent molecules are difﬁcult to resolve because, unlike solids, the spatial relationships
between the molecules are not ﬁxed on time scales that can be accessed using structural
measurements such as X-ray diffraction or multidimensional NMR spectrometry. The
existence of the liquid phase is a consequence of strong intermolecular interactions that
are difﬁcult to examine directly because of their combined long-range distance and short-
time duration within the medium. Collisional processes necessarily occur over short
distances, where there is actual “contact” between molecules. Long-range interactions
include electronic-electronic and vibrational-vibrational non-collisional energy transfer,
and these processes can proceed over tens of Angstrom distances. It is possible to probe
these interactions by measuring molecular motion and energy relaxation through the use

of time domain spectroscopies. With the advent of picosecond lasers, the study of liquids

expanded greatly because many of the dynamics relevant to solvation processes proceed
on the picosecond or longer timescale, especially in polar, strongly associative systems.
For motional processes, information about how a molecule physically interacts within its’
surroundings can be extracted by interpretation of the data within the framework of a
well-developed model. With an understanding of how an excited state population
relaxes, a complementary picture of local organization can be inferred, albeit at a lower
level at this point owing to the less mature nature of the theory available to treat this
problem. Local organization information is obtained from the efﬁciency of energy
transfer in a donor-acceptor pair, which depends on average donor-acceptor separation,
relative orientation and the nature of the coupling between the two species.

There are essentially three time-resolved spectroscopic approaches to the
measurement of solvent-solute interactions. These are measurement of the time-delayed

11-31

ﬂuorescence Stokes shift (TDFSS) of selected solutes,"lo rotational diffusion and

vibrational population relaxation.3244

Of these, TDF SS measurements have provided the
most insight into the early-time separation of inertial and diffusive contributions to
solvent reorganization about excited solute molecules, and at the same time these reports

have proven to be the most controversial in terms of their interpretation.‘”'49

Perhaps the
most signiﬁcant issue that needs to be addressed regarding TDFSS measurements is the
lack of generality of the effect. Despite the broad conclusions drawn from these
experimental results, only a handful of chromophores with complicated spectroscopic
responses are known to exhibit resolvable transient Stokes shiftsz’50 The two other

approaches to the measurement of solvent-solute interactions do not offer the same time-

resolution as the TDFSS measurements, but the phenomena of rotational diffusion and

vibrational relaxation are not limited to a few chromophores - these are general effects,
and the characteristic time scales over which they proceed are consonant with chemical
reaction processes.

Molecular probes that exhibit a change in an optical property in response to a
change in the local environment can provide valuable in situ information about chemical
and physical processes.5| For example, there are a number of molecules whose optical
response is sensitive to the pH of the local environment.52 Typically, these compounds
possess two or more different protolytic forms, each with a unique absorption and/or
ﬂuorescence spectrum. Examples of such probe molecules include xanthene dyes,
aromatic azines and coumarin dyes. These probes have been used in various biological
and medical applications to detect or quantify acid concentration in a system.53 Recently,
acid-sensitive dyes have been used to investigate photoacid generation in
microlithographic systems. For example, Pohlers et al. have reported the use of such
dyes to characterize the efﬁciency of photoacid generators.54 These authors observed a
shift in the absorbance spectrum of these dyes upon protonation from the neutral state to
the singly-charged state. In prior work, Pohler et al. observed shifts in the absorbance
spectra and large changes in ﬂuorescence intensity upon protonation of aromatic

monoazines.55

Dentinger et a1. evaluated several xanthene dyes for monitoring and
imaging photoacid production.SO Finally, Okoroanyanwu et al. reported the use of an
aromatic monoazine to monitor photoacid generation in ﬁlms and solutions.57

In addition to measuring the concentration of acid, ﬂuorescence can be used to

probe the local viscosity or free volume of the system. There are a number of probe

molecules that exhibit an increase in the ﬂuorescence signal and excited-state lifetime as

the local viscosity is increased. As the intramolecular isomerization or rotation becomes
hindered, non-radiative pathways for energy transfer are less efﬁcient; therefore, a greater
fraction of the excited-state energy must be dissipated through radiative (ﬂuorescent)
pathways. Crystal violet and malachite green are two such “propeller” probe molecules
that have been used to investigate free volume in solid matrices. Other free volume
sensitive-probes are based upon photoisomerization and intramolecular excimer
formation. With a small amount of a viscosity-sensitive molecule, the microviscosity or
free volume of the system can be measured. For example, Francis Wang et al. have used
1,3-bis-(1-pyrene) (BPP) to monitor the polymerization of methyl methacrylate in situ.58
Moorjani and coworkers further demonstrated the utility of ﬂuorescence monitoring
when dealing with diffusion-controlled reactions in media of different viscosity.59 Victor
and Torkelson have also used molecular probes to measure the free volume available in
polymer structures“)

We have undertaken a series of studies in an attempt to better understand solute-
solvent interactions for a variety of media, ranging from neat liquids to reactive polymer
matrices. We have used molecular reorientation, ﬂuorescence lifetime, vibrational
population relaxation and steady-state ﬂuorescence measurements, and report our
ﬁndings in this dissertation.

The picosecond pump-probe laser spectrometer used for most of the laser-based
experiments has been described in detail previously,"1 and we present only a brief outline
of its performance characteristics here. A graphical representation of the instrument is

presented in Figure 1. A mode-locked CW Nd:YAG laser (Coherent Antares 76-S)

pulse

  

 

 
  

    
  

 

 

 

3m 2w ,- diagnostics
ﬂ g j I
-H~w— 1.: It? ’
modulation ‘,
'-‘ s“: ,
cavity dumped dye lasers , ’ time delay
410 nm - 900 nm ? control

polarizatiq
detector sample EC: - o /

it 'I’ -l<‘. -_ :3: .
\ I, '

.~3lilllflilllliilllila‘P ....................

 

detection
electron iCS

 

Figure l — Schematic of picosecond pump-probe laser spectrometer used in this
work.

produces 30 W of average power (1064 nm, 100 ps pulses, 76 MHz repetition rate). The
output of this laser is either frequency-tripled to produce ~1.3 W of average power at
354.7 nm for experiments involving perylene and l-methylperylene as probe molecules,
or frequency-doubled to produce ~2.1 W of average power at 532 nm, for the
ﬂuorescence lifetime experiments where crystal violet is used as a probe molecule. In
both situations, this output is used to excite two cavity-dumped dye lasers synchronously
(Coherent 702-3). The laser dye used in both lasers is dependent upon the wavelengths
of interest. For experiments involving perylene and l-methylperylene as probe
molecules, both lasers operate with Stilbene 420 laser dye (Exciton). Rhodamine 6G
laser dye (Kodak) was used in both laser systems for the free volume characterization
experiments. The output of each laser, regardless of the dye used is ~10-100 mW
average power at 8 MHz repetition rate with a pulse that produces a 7 ps FWHM
autocorrelation trace using a three plate bireﬁ'ingent ﬁlter.

The pump wavelength for experiments with perylene and l-methylperylene was
chosen to access the 0-0 transition of the chromophore and the probe wavelength to
stimulate emission from the Slv=0 state to the S0":1 state, where the vibrational resonance
of interest is the 1375 cm'1 v=1 ring breathing mode, or the 1730 cm'1 v==l combination
mode. These wavelengths were used for the vibrational relaxation and the reorientation
measurements. It has been demonstrated previously, that the ground state and excited
state reorientation behavior of perylene are identical30 and we assume that this is also the
case for l-methylperylene. The probe beam was polarized at 0°, 90°, or 54.7° with
respect to the pump beam depending upon the experiment being performed. For

rotational diffusion measurements, the 0° and 90° polarizations are of interest, while the

vibrational population relaxation measurements use a probe beam polarization of 54.7° to
eliminate molecular reorientation contributions to the signal.

For the free volume characterization experiments, the pump laser wavelength was
set to 589.5 nm, coincident with the absorption maximum of the ﬁrst singlet transition of
crystal violet. The probe wavelength was chosen to lie within the absorption band of
crystal violet. The absorption spectrum of crystal violet is presented in Chapter 4. The
probe beam was polarized at 54.7° with respect to the pump beam to eliminate molecular
reorientation contributions to the signal.

Detection of the transient signals, regardless of the experiment, was accomplished
using a radio and audio frequency triple-modulation scheme, with synchronous
demodulation detection.62'64 Each reported time constant is the average of at least ﬁve
individual determinations that are themselves the average of seven to ten time-scans.

The an'ival time of the probe pulse relative to the pump is controlled through the
use of a computer interfaced translation stage and stepper motor controller. The distance
that the probe pulse travels is changed through the positioning of the translation stage,
resulting in a change of the arrival time of the probe laser pulse train. The time resolution
of this system, ~10 ps, is determined by the cross-correlation between the pump and
probe laser pulse trains.

In rotational diffusion measurements, the probe molecule sweeps out a speciﬁc
volume. The solvent cage formed around the solute can inﬂuence the ability of the solute
to rotate about speciﬁc axes and can determine the shape of the volume swept out. This
information provides for some insight into how inter-molecular interactions inﬂuence the

motion of the probe, and by inference the average organization of the solvent molecules

around the solute. The solute/solvent system can also be studied to determine its energy
relaxation behavior. The key to this approach is to identify systems where inter-
molecular energy transfer dominates as the initial relaxation step, and the other properties
of the system (inter-molecular orientation and distance) can be examined.

To examine the inter-molecular interactions in solution we use two
complementary methods that share a single experimental apparatus. Induced orientational
anisotropy measurements provide information about the forces that serve to inhibit
molecular motion by the examining the time required for an ensemble of molecules to
randomize. Vibrational population relaxation measurements provide information about
the local environment by revealing how energy is transferred to the surrounding solvent
bath over distances longer than those characteristic for inelastic collisions. Information
about the density of donor-acceptor sites, orientation, and coupling can help to construct
a preliminary model of how the solute and solvent molecules interact in solution.

The examination of rotational diffusion dynamics is a widely used method to

,26,6 -69 -
25 5 The theoretical framework

probe the nature of solvent-solute interactions.
needed to understand these interactions is well documented.70'75 From the application of
the theory to the experimental data, the solute rotational diffusion constant, the angle
between the excited and probed transition moments and, in some cases, a sense of the
dielectric properties of the local environment can be obtained. This information can then
be used to extend the understanding of the average local organization of the solvent
around the solute.

Using linearly polarized light, an ensemble of molecules with transition dipole

moments oriented parallel to the polarization of the incident electric ﬁeld can be excited,

inducing a transient anisotropy in the system. This induced anisotropy will decay in time
due to the rotational dynamics of the system. The induced orientational anisotropy

function is extracted from the raw data through Equation (1).

= Inn-Ila) 1
1|,(z)+21l(z) ( )

 

R(t)

I"(t) and I i(t) are the intensities of the response when the polarizations of the

two beams are oriented parallel and orthogonal relative to each other. We present a
representative sample data set in Figure 2. When these data are related according to
Equation (1), an exponential decay curve results (Figure 3).

Chuang and Eisenthal have formulated the relationship between R(t) and solute
properties such as the angle between the excited and probed one photon transition
moments and the actual anisotropy in the diffusion constant."4 They found that R(t) can
decay with as many as ﬁve exponential components, although the most common case is
that of a single exponential. In cases where more than one decay component is resolved,
it is possible to interpret the details of the reorientation dynamics in substantial detail.

For the purposes of the following discussion, we consider that the chromophores
have their n-system in the approximate xy plane with x being the long, in-plane axis, and
z is perpendicular to the rt-system plane (Figure 8.) For each chromophore our intention
was to assign the coordinate system such that the x-axis coincides with the absorption
transition moment. For perylene this assignment is coincident with speciﬁc bond axes
because of its high symmetry while for l-methylperylene it is not since the n-system for
this chromophore is non-planar. Using these approximations and Chuang and Eisenthal’s

equations,64 we can relate the experimental reorientation times to the shape of the volume

signal intensity (a.u.)

 

 

 

l n l n l n l J l

-100 0 100 200 300 400 500
delay time (ps)

Figure 2 — Experimental I“(t) and I i(t) scans for l-methylperylene in n-pentanol,
along with the instrumental response function. These data are typical of those
recorded for reorientation measurements.

10

 

 

 

0.3 P
0.2 -
£34
0.1 -
0.0
1 l 1 I J l 4 l A 1 # I
-100 0 100 200 300 400 500

delay time (ps)

Figure 3 — Anisotropy function, R(t), for the data shown in Figure 2. The decay is fit
to the function R(t) = R1(0)exp(-t/r.) + R2(0)exp(-t/'c2)

ll

swept out by the rotating molecule. We describe this volume in terms of ellipsoids,
where a prolate ellipsoid, Equation (2) and Figure 4, is characterized by rotation
primarily about its long in-plane axis, D. > Dy ~ D2 and an oblate ellipsoid, Equation (3)
and Figure 5, by is characterized by rotation about the axis perpendicular to the
chromophore n-system, DZ > Dx ~ Dy.

Prolate: R(t) = 0.4exp(—6th) (2)

Oblate: R(t) = 0.3exp(—(2DX +41), )z) +0.1exp(—6Dxt) (3)

Dx

    

DY

Figure 4 — Prolate ellipsoid with Cartesian coordinate system superimposed. See
text for discussion.

Dz

DY

Figure 5 - Oblate ellipsoid with Cartesian coordinate system superimposed. See
text for a discussion.

A single exponential decay ﬁt of the reorientation anisotropy function R(t) is the
most common observation, but this observation can be attributed to either a spherical

rotor shape or a prolate rotor shape being swept out of solution by the probe molecule.

We model this behavior as a prolate rotor, since both perylene and l-methylperylene are

not spherical molecules in shape. A double exponential decay ﬁt is described to be an
oblate rotor shape being swept out of solution by the probe chromophore. The rotor
shape of the chromophore and changes in the rotational behavior result from
intermolecular and intramolecular interactions between the solute and the solvent in
solution, and determination of the rotor shape can elucidate information regarding these
solvation processes.

Vibration population relaxation dynamics can also be used as a probe of solvation
processes. Vibrational population relaxation provides information about the speciﬁc
relaxation pathways that are used to dissipate energy in a molecule. Rotational diffusion
measurements sense the entire molecule and its’ average response to a perturbation.
When a molecule is excited, it is generally true that there are multiple ways for it to
dissipate that energy, with many solvent cage conﬁgurations that lie within a narrow
energetic window, capable of interconversion on a fast time scale. A determination of
how molecules at room temperature dissipate energy to surroundings can give
information on how molecules interact. In gases, the spacing between molecules can be
determined statistically and differs from that in solution. Analogous arguments can be
made for solids. In these systems, the dissipation of energy is not random — but occurs
through speciﬁc pathways and over relatively deﬁned distances. In solutions, both
collisional and non-collisional energy transfer processes have to be taken into account.76
Non-collisional energy transfer can occur over relatively long distances. These types of
energy transfer are signiﬁcant in solution.77

Energy transfer can be described in terms of coupling of transition moments that

scale as r‘n where the speciﬁc type of interaction determines the value of n. Dipole-

l3

dipole, dipole-induced dipole and induced dipole-induced dipole interactions all scale
over r'b, and it is this sort of coupling that Forster discussed most thoroughly. Coupling
between higher order multipole moments (quadrupole, octupole) scale with r'" where n 2
8. The interaction energy falls off more rapidly the higher the order of the multipole,78
and this consideration is important for our work.

77.79

F'o'rster described this for dipole-dipole coupling. Experimentally, an excited

donor molecule (DI) transfers energy to a ground state acceptor molecule (A) with a

characteristic rate constant km:

D'+A—*”—">D+ A'

km =§ {HEW

where “(Q is the radiative lifetime of D’, R0 is the critical transfer distance which

(4)

incorporates the spectral overlap term and the magnitudes of the transition dipole
moments, R is the D-A intermolecular distance and K2 describes the orientation of the

donor with respect to the acceptor:

2

K = (sinQD sinOA cos¢—2cosBD c036,) (5)

The strength of the coupling is dependent upon the spectral overlap considerations
(detuning) between the donor and acceptor and the average distance between the two
species and their relative orientations. The information about the efﬁciency of the
coupling then provides additional complementary information to the rotational diffusion
measurements, since both address the separation distance and orientation of the donor-
acceptor pair. For electronic-electronic transitions, R0 is approximately 25-50A and

~10A for vibrational-vibrational transfer. The angles DD and 0A indicate the orientation

14

of the donor and acceptor transition moments with respect to the plane containing the two
species, and (I) is the relative angle between the transition moments along the connecting
vector.

Under our two-pulse measurement scheme for vibrational population relaxation
molecules possessing a center of inversion, such as perylene, vibrational modes that are
infrared active will be Raman inactive and vice versa. In l-methylperylene, the center of
inversion is removed by the addition of the methyl group to the ring, and this result is that
all modes are both infrared and Raman active.80 Infrared active vibrations modulate the
dipole moment of the molecule and Raman active vibrations modulate the polarizability.
These two types of transitions may involve different types of coupling with
characteristically different interaction lengths. If a donor mode is infrared active (Au),
and the acceptor mode is also IR active, the coupling will scale as r’6. If Raman active
donor resonances are not also 1R active, coupling to the acceptor through a higher
moment is operative, and such coupling operates over a shorter distance than dipolar
coupling. This change in the range probed makes it difﬁcult to relate such data to other
types of information increasing the number of possible explanations. Rotational
dynamics and vibrational population relaxation measurements that probe approximately
the same distance surrounding the donor molecule simplify the explanations involved.

This method‘m’41 examines the depopulation of a vibrational mode of the
chromophore on the picosecond time scale. The time constraints for relaxation are highly
dependent upon the mode in question and the solute/solvent system. If this system-
dependency can be understood, signiﬁcant information can be gained on local

organization in solution.

15

In this method the chromophore under examination is treated as a strongly
coupled three level system. We present a schematic in Figure 6. Spontaneous and
stimulated emission from the S1 state to the 80 state, stimulated absorption from an
excited vibrational level in the So electronic manifold to the SI state, and depopulation of
the excited vibrational level to the ground vibrational level in the So state all inﬂuence the
experimental signal. The difference between the pump laser frequency and the probe
laser frequency is set to a Raman resonance in the solute. The solvent system is chosen
such that it possesses vibrational resonances in close energetic proximity of the solute
(AVDA~0) providing a speciﬁc, hopefully efﬁcient pathway for energy transfer via long
range polar vibrational-vibrational coupling. These choices bias the measurement to
make intra-molecular relaxation a non-dominant pathway. Here, perylene and
l-methylperylene have ring breathing modes at ~1375 cm”. The terminal methyl groups
of the alkanols, aldehydes and ketones used in this work exhibit rocking motions at
~1378 cm", essentially degenerate with the solvent mode. In addition to this
donor/acceptor system, the donor solute molecules also possess a vibrational mode near
~1730 cm'l. This mode is degenerate with the carbonyl-stretching mode in the aldehydes
and ketones (~1727 and 1716 cm", respectively). These donor/acceptor modes will be
examined throughout this work to understand the role of molecular structure and
functionality in mediating solution phase vibrational excitation transport.

As a function of time, the population of state A will exhibit an exponential
ﬂuorescence decay, while the population of state B will build up and reach a steady—state,
if the decay rate constant from the intermediate vibrational level is much larger than the

decay rate constant from state A. We have chosen our solute/solvent system such that

16

this is the case. If the population changes of states A and B are measured as a function of
time, a buildup, related to the populating of state B, followed by long time decay related
to the ﬂuorescence depopulation of state A will result. We present a typical dataset that
demonstrates this in Figure 7.

The time constant for population relaxation gives information about the
conditions necessary to transfer energy from the donor to the acceptor. This information
is then used to infer the presence or absence of local organization in these systems. By
itself, this information does not provide a complete picture. When taken with
complementary rotational diffusion data, however, some insight into local organization
can be obtained.

These principles of interactions between solutes and solvents can be used to
elucidate information about applied chemical systems. Probe molecules can be used to
determine changes in the local environment of solutions. Crystal violet is a useful probe
of microviscosity in solution. Crystal violet (Scheme 4, Chapter 4 —— this work), a
member of the triphenylmethane dye family, is predominantly a D3 “propeller” shaped
molecule in solution with three identical N,N-dimethylaniline rings bound to a central
carbon atom tilted slightly out of the molecular plane. Ring rotation around the bonds
between the phenyl rings to the central carbon atom in what is believed to be a barrierless
process along the excited state potential energy surface. Studies have indicated a strong
dependence of the excited state lifetime on solvent viscosity, which is indicative of
essentially no intra-molecular rotational barrier to ring rotation.“82 The characteristically
low ﬂuorescence quantum yield of this family of molecules prevents us from using a

ﬂuorescence lifetime measurement to explore these changes. Instead, we examine the

17

 

Pump kA
= 0 Spontaneous
V0-0 VP .
ump ° Stlmulated
kV

 

 

 

A SW

Probe

V0-0 _ VProbe = vProbe

kE

B 85*"

 

c 55*

 

Figure 6 — Three level system model for vibrational population relaxation

measurements.

18

1.0

0.8

0.6

0.4

signal intensity (a.u.)

0.2

0.0

 

 

 

l n l L l r l 1 J 1 l L l A l n l

-100 0 100 200 300 400 500 600 700 800

delay time (ps)

Figure 7 — Magic angle data for perylene in methanol with instrument response

function.

19

ground state population by detecting the transient change in absorption of the probe pulse
resulting from excitation by the pump laser pulse, to sense changes in the local
environment in the photoresist thin ﬁlms.

Chapter 1 of this dissertation outlines two procedures attempted in our laboratory
to synthesize l-methylperylene, one of the solute chromophores studied in chapters 2 and
3. We compare the methods of Ziegler83 and Peake84 et a1. and provide our results using
their procedures, with characterization and puriﬁcation data.

Chapters 2 and 3 of this dissertation examine the rotational diffusion and
vibrational population dynamics of perylene and l-methylperylene in a series of normal
aliphatic alcohols and a series of aldehydes and selected ketones. We have found that the
presence of a labile proton that can facilitate hydrogen bonding in alcohols and aldehydes
has a greater effect on the local organization in the condensed phase than dipole-induced
dipole interactions, such as those found in the ketones.

Chapters 4 and 5 of this dissertation deal with the characterization of photoresist
materials used in the semiconductor industry. We describe a methodology to quantify the
production of acid molecules in a photoresist material used in the microlithographic
production of semiconductors using the probe molecule crystal violet in Chapter 4. In
Chapter 5, we use the same probe molecule to examine the interactions between solute
(probe molecule) and the solvent (photoresist) as a function of processing step.

Chapter 6 provides some ﬁnal conclusions as well as some suggestions for future
work in both the solution phase characterization of perylene and l-methylperylene and
the continued characterization of photoresist materials used in microlithographic

applications.

20

In the appendix, a description of the hardware used and the function of each of the
parts of the virtual instrument created to operate the pump-probe laser spectrometer
instrument used in most of the experiments of this dissertation is recorded for future

reference.

21

10.
ll.

12.

13.
14.
15.
16.
17.
18.
19.
20.

21.

Literature Cited

Shapiro, S. L.; Winn, K. R.; Chem. Phys. Lett., 1980, 71, 440.

Maroncelli, M.; Fleming, G. R.; J. Chem. Phys, 1987, 86, 6221.

Huppert, D.; Ittah, V.; Kosower, E.; Chem. Phys. Lett., 1989, 159, 267.
Chapman, C. F.; Fee, R. S.; Maroncelli, M.; J. Phys. Chem, 1990, 94, 4929.
Huppert, D.; Ittah, V.; Chem. Phys. Lett., 1990, 1 73, 496.

Jarzeba, W.; Walker, G. C.; Johnson, A. E.; Barbara, P. F.; Chem. Phys, 1991, 152,
57.

Wagener, A.; Richert, R.; Chem. Phys. Lett., 1991, 176, 329.

Fee, R. S.; Milsom, J. A.; Maroncelli, M.; J. Phys. Chem, 1991, 95, 5170.
Yip, R. W.; Wen, Y. X.; Szabo, A. G.; J. Phys. Chem1993, 97, 10458.
Fee, R. S.; Maroncelli, M.; Chem. Phys, 1994, I83, 235.

Sanders, M. J.; Wirth, M. J.;Chem. Phys. Lett., 1983, 101, 361.

Gudgin-Templeton, E. F .; Quitevis, E. L.; Kenney-Wallace, G. A.; J. Phys. Chem,
1985, 89, 3238.

Von Jena, A.; Lessing, H. E.; Chem. Phys, 1979, 40, 245.

Von Jena, A.; Lessing, H. E.; Ber. Bunsen-Ges Phys. Chem, 1979, 83, 181.

Von Jena, A.; Lessing, H. E.; Chem. Phys. Lett., 1981, 78, 187.

Eisenthal, K. B.; Acc. Chem. Res, 1975, 8, 118.

Fleming, G. R.; Morris, J. M.; Robinson, G. W.; Chem. Phys, 1976, 17, 91.
Shank, C. V.; Ippen, E. P.; Appl. Phys. Lett., 1975, 26, 62.

Millar, D. P.; Shah, R.; Zewail, A. H.; Chem. Phys. Lett., 1979, 66, 435.
Gudgin-Templeton, E. F .; Kenney-Wallace, G. A.; J. Phys. Chem, 1986, 90, 2896.

Blanchard, G. J .; Wirth, M. J .; J. Phys. Chem, 1986, 90, 2521.

22

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35

36.

37.

38.

39.

40.

41.

42.

Blanchard, G. J .; J. Chem. Phys, 1987, 87, 6802.

Blanchard, G. J .; Cihal, C. A.; J. Phys. Chem, 1988, 92, 5950.

Blanchard, G. J .; J. Phys. Chem, 1988, 92, 6303.

Blanchard, G. J .; J. Phys. Chem, 1989, 93, 4315.

Blanchard, G. J.; Anal. Chem, 1989, 61, 2394.

Alavi, D. S.; Hartman, R. S.; Waldeck, D. H.; J. Phys. Chem, 1991, 95, 6770.
Hartman, R. S.; Alavi, D. S.; Waldeck, D. H.; J. Phys. Chem, 1991, 95, 7872.
Jiang, Y.; Blanchard, G. J .; J. Phys. Chem, 1994, 98, 6436.

Brocklehurst, R; Young, R. N.; J. Phys. Chem, 1995, 99, 40.

Pauls, S. W.; Hedstrom, J. F .; Johnson, C. K.; Chem. Phys, 1998, 23 7, 205.
Elsaesser, T.; Kaiser, W.; Annu. Rev. Phys. Chem, 1991, 42, 83.

Lingle, R., Jr.; Xu, X.; Yu, S. C.; Zhu, H.; Hopkins, J. B.; J. Chem. Phys, 1990, 93,
5667.

Anﬁnrud, P. A.; Han, C.; Lian, T.; Hochstrasser, R. M.; J. Phys. Chem, 1990, 94,
1180.

. Heilweil, E. J .; Casassa, M. P.; Cavanagh, R. R.; Stephenson, J. C.; Ann. Rev. Phys.

Chem, 1989, 40, 143.

Heilweil, E. J .; Cavanagh, R. R.; Stephenson, J. C.; Chem. Phys. Lett., 1987, 134,
181.

Heilweil, E. J .; Cavanagh, R. R.; Stephenson, J. C.; J. Chem. Phys, 1989, 89, 230.

Heilweil, E. J .; Casassa, M. P.; Cavanagh, R. R.; Stephenson, J. C.; J. Chem. Phys,
1986, 85, 5004.

Hambir, S. A.; Jiang, Y.; Blanchard, G. J.; J. Chem. Phys, 1993, 98, 6075.
Jiang, Y.; Blanchard, G. J .; J. Phys. Chem, 1994, 98, 9411.
Jiang, Y.; Blanchard, G. J .; J. Phys. Chem, 1994, 98, 9417.

Jiang, Y.; Blanchard, G. J .; J. Phys. Chem, 1995, 99, 7904.

23

43.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

McCarthy, P. K.; Blanchard, G. J .; J. Phys Chem, 1995, 99, 17748.

. McCarthy, P. K.; Blanchard, G. J .; J. Phys. Chem, 1996, 100, 5182.

Agmon, N.; J. Phys. Chem, 1990, 94, 2959.

Maroncelli, M.; Fee, R. S.; Chapman, C. F.; Fleming, G. R.; J. Phys. Chem, 1991,
95, 1012.

Blanchard, G. J .; .1. Chem. Phys, 1991, 95, 6317.

Jiang, Y.; McCarthy, P. K.; Blanchard, G. J .; Chem. Phys, 1994, 183, 249.
Flory, W. C.; Blanchard, G. J .; Appl. Spec, 1998, 52, 82.

Castner, E. W.; Maroncelli, M.; Fleming, G. R.; J. Chem. Phys, 1987, 86, 1090.

Jessop, J .L.; Goldie, S.N.; Scranton, A.B.; Blanchard, G.J.; Rangarajan, B.;
Capodieci, L.; Subramanian, R.; Templeton, M.K. SPIE Proc. 1999, 3678, 914.

Haugland, R. P. In Handbook of Fluorescent Probes and Research Chemicals, 6th
ed.; Spence, M. T. Z; Johnson, 1. 1D. Eds; Molecular Probes: USA, 1996.

Slavik, J .; Fluorescent Probes in Cellular and Molecular Biology, CRC Press, Inc.:
Ann Arbor, MI, 1994.

Pohlers, G.; Scaiano, J .C.; Sinta, R.; Chem. Mater. 1997, 9, 3222.
Pohlers, G.; Virdee, S.; Scaiano, J .C.; Sinta, R.; Chem. Mater. 1996, 8, 2654.

Dentinger, P.M.; Lu, 3.; Taylor, J.W.; Bukofsky, S. J.; Feke, G.D.; Hessman, D.;
Grober, R.D.; J. Vac. Sci. T echnol. B, 1998, I6, 3767.

Okoroanyanwu, U.; Byers, J.D.; Cao, T.; Webber, S.E. Willson, CG; Micro and
Nano-patterning Polymers; Ito, H.; Reichmanis, E.; Nalamasu, O.; Ueno, Y. Eds;
American Chemical Society: Washington, DC, 1998, 174.

Wang, F.W.; Lowry, R.E.; Grant, W.H.; Polymer, 1984, 25, 690.

Moorjani, S.K.; Rangarajan, B.; Scranton, A.B.; “Effect of viscosity on the rate of
photosensitization of diaryliodonium salts by anthracene,” In Photopolymerization:
Fundamentals and Applications; Scranton, A.B.; Bowman, C .N.; Peiffer, R.W. Eds.
American Chemical Society: Washington, DC, 1997, 95.

Victor, J.G.; Torkelson, J .M.; Macromolecules, 1987, 20, 2241.

24

61.

62.

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

81.

82.

Jiang, Y.; Hambir, S. A.; Blanchard, G. J .; Opt. Commun, 1993, 99, 216.
Bado, P.; Wilson, S. B.; Wilson, K. R., Rev. Sci. Instrum 1982, 53, 606.

Andor, L.; Lorincz, A.; Siemion, J .; Smith, D. D.; Rice, S. A., Rev. Sci. Instrum
1984, 55, 64.

Blanchard, G. J.; Wirth, M. J.; Anal. Chem. 1986, 56, 532.

Labhart, H.; Pantke, E. R. Chem. Phys. Lett. 1973, 23, 482.

Zinsli, P. E. Chem. Phys. 1977, 20, 299.

Klein, U. K. A.; Haar, H. P. Chem. Phys. Lett. 1979, 63, 40.
Christensen, R. L.; Drake, R. C.; Phillips, D. J. Phys. Chem. 1986, 90, 5960.
Ben-Amotz, D. Scott, T. W. J. Chem. Phys. 1987, 87, 3739.

Debye, P. Polar Molecules; Chemical Catalog Co.: New York, 1929; 84.
Perrin, F. J. Phys. Radium 1936, 7, 1.

Chuang, T. J.; Eisenthal, K. B. J. Chem. Phys. 1972, 57, 5094-5097.

Hu, C. M.; Zwanzig, R. J. Chem. Phys. 1974, 60, 4354.

Youngren, G. K.; Acrivos, A. .1. Chem. Phys. 1975, 63, 3846.

Zwanzig, R.; Harrison, A. K. J. Phys. Chem. 1985, 83, 5861.

Yardley, J. T. Introduction to Molecular Energy Transfer; Academic: New York,
1980.

Forster , T. Ann. Phys. (Liepzig) 1948, 2, 55.
Akins, P. W. Physical Chemistry, 4th ed.; Freeman: New York, 1990.

Forster, T. In Comparative Effects of Radiation; Burton, M., Kirby-Smith, J. S.,
Magee J. C., Ed; Wiley: New York, 1960; p 300.

Cotton, F. A. Chemical Applications of Group Theory, 3rd ed. Wiley: New York,
1990.

Ben-Amotz, D.; Harris, C. B.; J. Chem. Phys. 1987, 86, 4856.

Ben—Amotz, D.; Harris, C. B.; J. Chem. Phys. 1987, 86, 5433.

25

83. Zieger, H. E.; Laski, E. M. Tetrahedron Lett. 1966, 32, 1966

84. Peake, D. A.; Oyler, A. R.; Heikkila, K. E; Liukkonen, R. J; Engroff, E. C.; Carlson,
R. M. Synth. Commun. 1983, I3, 21.

26

Chapter 1

SYNTHETIC PATHWAYS TO l-METHYLPERYLENE

Introduction

Much of the work in this dissertation concerns the behavior of perylene and
l-methylperylene in solution. These two probe molecules were chosen for examination
because perylene is well characterized"8 and the ease of synthesis of l-methylperylene, a
molecule with similar electronic spectroscopic properties, but lower symmetry. The
lower symmetry of l-methylperylene affects the linear response, vibrational population
relaxation times and rotational difﬁision, (vide infra). These changes can help to
elucidate information about the solution phase behavior of these chromophores, which
has yet to be resolved satisfactorily in the literature.

Perylene is a member of the D2}, point group. The addition of a single methyl
group in the 1 position reduces the symmetry of l-methylperylene to the C1 point group,
removing the center of inversion. The methyl group on l-methylperylene causes the two
naphthalene moieties to be twisted out of planarity by ~ 20°.“”10 The methyl group added
to form l-methylperylene results in a reduction in the conjugation of the tit-system, Iwhich
causes a blue shift, relative to the linear response of perylene. The reduction in symmetry
also causes additional vibrational transitions to become allowed. These additional
transitions are important to our understanding of these data for several reasons. Under
our two-pulse measurement scheme for vibrational population relaxation, used in many
experiments in this dissertation, molecules possessing a center of inversion, such as

perylene, the lowest order multipole moment that can be modulated using our excitation

27

scheme is the quadrupole moment. The solute donor mode will couple to the solvent bath
modes through quadrupole-dipole interactions, assuming an IR-active acceptor mode.
This coupling mechanism depends on donor-acceptor spacing as {8.5 For a donor that
does not possess a center of inversion, such as l-methylperylene, the excited vibrational
resonance is a dipolar modulation and the donor-acceptor coupling will depend on
intermolecular distance as r'b. The Stokes shift of l-methylperylene relative to perylene
is indicative of a small but ﬁnite permanent dipole moment in l-methylperylene. PM3
semi—empirical calculations indicate u ~ 0.24 D in the SO and 0.53 D in the S1 for
l -methylperylene. ' 1

Measurement of the vibrational population relaxation, T1, of perylene and
l-methylperylene as a function of solvent identity can provide information on the spatial
proximity and relative orientation of the donor and acceptor speciess’l2 Quadrupole-
dipole and dipole-dipole coupling processes operate over different length scales (r'8 vs.
r") and thus it is not possible to compare results directly. Despite the differences of
length scale, the solvent-dependent trends observed in the T; data for the two molecules
provide insight into the average spatial relationship between donor and acceptor
functionalities in solution.

Perylene and l-methylperylene have ring breathing modes at ~1375 cm”. The
terminal methyl groups of the alkanols, aldehydes and ketones used in this work exhibit
rocking motions at ~1378 cm'l, essentially degenerate with the solvent mode. In addition
to this donor/acceptor system, the donor solute molecules also possess a vibrational mode
near ~1730 cm]. This mode is degenerate with the carbonyl-stretching mode in the

aldehydes and ketones (~1727 and 1716 cm", respectively). These donor/acceptor

28

modes will be examined throughout this work to understand the role of molecular
structure and functionality in mediating solution phase vibrational excitation transport.

To simplify the treatment of our rotational diffusion data, we assign a coordinate
system to each chromophore such that the x-axis resides in approximately in the xy plane

of the n-system and it is coincident with the long, in-plane axis of the chromophore. The

z-axis in our coordinate system is perpendicular to the n-system plane, shown in Figure 8.

 

 

Figure 8 — Laboratory coordinate system applied to perylene. A similar laboratory
coordinate system was applied to 1-methylperylene.

The x-axis also coincides with the absorption transition moment of perylene because of
its high symmetry, but this does not apply for l-methylperylene it is not since the it-
system for this chromophore is non-planar. Comparing the reorientation behavior of
perylene with that of l-methylperylene allows us to evaluate whether or not measurable
contributions from weak dipolar solvation effects are present in addition to the dominant
van der Waals interactions.

l-Methylperylene (CZIHM, F W = 266.34) can be synthesized from perylene in one
of two ways. The ﬁrst method used in our laboratory, developed by Zieger and Laski,l3
involved the addition of methyl lithium to perylene in benzene with heating to obtain a

reported yield of 33%. The second method, by Peake, et al.”, contrasts the ﬁrst method

29

by cooling to -78°C, in dry tetrahydrofuran (THF) in an attempt to slow the kinetics of
the alkyl addition. The reported yield of this second synthesis scheme was 85 i 5%.
Both methods were attempted in our laboratory, and the method reported by Peake et al.
was preferred, because of higher yield and “one-pot” (single reaction vessel) procedure.
The experimental schemes, procedures, and our results are outlined here. Spectroscopic
data and separation techniques are also supplied in an attempt to provide a more complete
synthetic procedure.

Synthesis Scheme 1

Scheme 1 — Synthesis of l-methylperylene through alkylation of perylene with
methyllithium.

CH3Li Li

H3C\ r—\ /CH3 H
t ‘i
CH3 CH3 : CH; 5% Pd on Carbon
Benzene H T
80°C
24hr

Scheme 1 illustrates the method of Zieger and Laski.l3 In this method, a 500mL,

 

 

three necked round bottom ﬂask is equipped with a condensor sealed with a rubber
septum and a second rubber septum equipped with a nitrogen inlet, and an addition
funnel. A magnetic stir bar, heating mantel and magnetic stirring plate were also used.
The reaction vessel is charged with perylene (5 g, CZOHM, F W = 252.31, 20 mmol) in 400
mL of benzene and purged with nitrogen for 30 min. 5 mL of methyl lithium in benzene
(1.4 M, 7 mmol) was added through the septum in the condensing column using a
nitrogen-purged syringe. 10 mL of N, N, N’, N’-tetramethylethylenediamine (FW =
116.21, 66 mmol) was added to the reaction using a nitrogen-purged syringe, to serve as

an electron donor. The solution color turned to a dark green upon the addition of the N,

30

N, N’, N’—tetramethylethylenediamine. The reaction was reﬂuxed under nitrogen for 24
h, with stirring. The reaction subsequently quenched using 100 mL of 2 M HCl. The
organic layer was then separated, dried using sodium sulfate, and concentrated. The
crude product was crystallized from hexane and ethanol. Subsequent recrystallization
from ethanol, afforded an orange solid. This powder was redissolved in ethanol and
heated. To this orange colored solution, ~0.3 g of 5% wt. palladium on activated carbon
was added to the solution and stirred for 30 min. The carbon was removed through
vacuum ﬁltration. The orange solution was observed to have turned yellow in color.
Concentration of this solution afforded a dark orange solid. 1H NMR of the recovered
solid showed that a peak at 2.9 ppm (s, 3H) indicative of the methyl group added to the
perylene system.

Separation of the product and the starting material was not a trivial task. The
change in the polarity from perylene to l-methylperylene is not signiﬁcant enough to
achieve baseline resolution between the two peaks in any of the chromatographic
separation schemes attempted in our laboratory. Thin layer chromatography was
attempted using a series of solvent systems suggested through consultations with the
members of the MSU Organic area. Hexane/ethyl acetate, cyclohexane/ethyl acetate, and
petroleum ether/ethyl acetate were examined, along with reverse phase thin layer
chromatography, with 95/5 (v/v) acetonitrile/water as the mobile phase. 95/5 (v/v)
cyclohexane/ethyl acetate together with extended elution times had the best results for
thin layer chromatography. This process is labor intensive, and the recovered product for
each separation is very small. Several hundred milligrams are required for the rotational

diffusion and vibrational population relaxation measurements in this dissertation, and the

31

throughput, coupled with the reported and observed low yields was insufﬁcient to
produce enough material in a timely manner. A suggestion was made from the MSU
Organic area to attempt to separate the l-methylperylene and perylene using column
chromatography. A 97/3 (v/v) petroleum ether/ethyl acetate mobile phase were found to
work best for with ZOO-mesh silica gel as the stationary phase. Because the crude product
was only slightly soluble in petroleum ether, the sample was added to the column by
creating a slurry with a minimal amount of tetrahydrofuran (THF), and silica gel and
dried under vacuum. This was deposited at the top of the column, and then eluted with
the solvent system speciﬁed above. Because the reaction has an inherent low yield,
separation of the product from the starting material was still difﬁcult, and did not produce
sufﬁcient quantities of material for the spectroscopic experiments. Consultations with the
MSU Organic area resulted in another literature search for additional information about a
more efﬁcient methodology.

The challenges associated with the method of Zieger and Laski13 are numerous.
The low yield and separation make it a labor intensive, inefﬁcient process. With this
methodology, the introduction of water anywhere in the system will react with the methyl
lithium and quench it. The system must be devoid of all water, including solvents,
glassware, and syringes used. Proper handling of air sensitive reagents must be
maintained. Failure to do so will adversely affect the yield, as shown here. Few of these
techniques were observed during the ﬁrst attempt using the procedure of Zieger and
Laski, and this resulted in a lower yield being obtained in our laboratory than what was

reported in the literature, and contributed to the inability to recover any useful product.

32

Synthesis Scheme 2
With these challenges, a second search for alkylation methodology was
performed. A second method, developed by Peake et al.'4 was found, and is presented in
Scheme 2. This method, was reported to have a higher yield than the Zieger and Laskil3
method (85 i 5% vs. 33%), and was performed in a single reaction vessel, an

improvement over the ﬁrst method.

Scheme 2 - The alkylation of perylene using methyl lithium/lithium bromide to form
l-methylperylene.

2.: CH3LI/LIBI' l (CXCCSS)
THF 2 3N32$203
-78°C 85%
Overnight

In this method, a 1000 mL oven-dried three necked round bottom ﬂask is charged
with perylene (2 g, CZOHM, FW = 252.31, 8 mmol) in 500 mL of dry tetrahydrofuran
(THF). The THF was distilled from sodium metal and benzophenone ketyl. The reaction
vessel is equipped with three rubber septa, and a magnetic stir bar with a magnetic stirrer.
A positive pressure of nitrogen was maintained throughout the experimental procedure.
The reaction was cooled to ~ -78°C using a dry ice/acetone bath. 11 mL of methyl
lithium/lithium bromide complex (1.5 M, 20 mmol) was added using a dry, purged
syringe and a syringe pump over the course of 1 hr with stirring. The solution color
turned to a dark green upon addition of the lithium complex. 1 hr after the addition of the
lithium complex, the cooling bath was removed. The reaction vessel was allowed to
slowly warm to room temperature overnight, with stirring. The reaction was then cooled

to 0°C and 7 g of iodine was added in one portion. The reaction continued to stir for 6.75

33

h after the iodine addition. This is much longer than the speciﬁed procedure14 of 15 min.
The time was extended as 1H NMR of initial experiments (not shown) indicated the
existence of possible intermediates that could have been reacted with longer
decomposition times.

The solution was then diluted with 500 mL CHzClz to minimize the size of the
emulsion that will form with THF/water mixtures. To decompose any excess iodine, the
solution was washed with saturated aqueous sodium thiosulfate (2x 400 mL). The
recovered organic phase was then washed with water (2x 200 mL), and brine solution (2x
200 mL). The organic phase dried over MgSO4 and concentrated. The crude product
was then puriﬁed using silica gel column chromatography. The solvent system of 70/30
(v/v) hexane/methylene chloride suggested by Peake et al.'4 did not separate the products
satisfactorily. Testing determined that a better solvent system for the column
chromatography was 95/ 5 (v/v) petroleum ether/ethyl acetate.

The fractions containing the desired product were then combined, and
concentrated. 1H NMR of the concentrated product indicated the desired singlet at 2.9
ppm, attributed to the methyl group and the multiplet between 7.37 and 8.18 ppm, in the
aromatic region had the appropriate integration of 10.73. The crude product was then
recrystallized twice from hexane and ethanol affording the desired product as
yellow/orange plates, 0.3185 g, 16% yield. The difference in yield to that reported14 is
attributed to the difference in ability level of the members of our laboratory and the
research team of Peake et al.

The lH NMR for pure perylene and the puriﬁed l-methylperylene product are

presented in Figure 9 and Figure 10, respectively. The melting point of the puriﬁed

34

product was determined to be 116-120°C, in accordance with published values for
l-methylperylene.l3 Mass spectrometry of the puriﬁed l-methylperylene product shows a
parent peak at m/z = 266.2 and a peak at 252.2 in a 7.25:1 ratio. These peaks are in
accordance with published data.'3 A typical absorption/emission set for perylene and
l-methylperylene in methanol respectively, are provided in Figure 11 and Figure 12 for
reference. The absorption and emission maxima were observed to be 433.5 and 435.5 nm
for perylene in methanol, and 425.5 and 436.0 nm for l-methylperylene in methanol.

5 .
13" The lrnear

These values are within experimental agreement with similar solvents.
response is also observed to be broader for l-methylperylene than for perylene in the
same solvent.

The second synthetic scheme of Peake et al.14 was applied in our laboratory to
obtain a sufﬁcient amount of desired product in a short amount of time. The single
reaction vessel, and straightforward puriﬁcation procedure provided for a reasonable

throughput so that these probe molecules can be used with sufﬁcient speciﬁcity and

purity in the spectroscopic experiments outlined elsewhere in this work.

35

00

05

10

20

IIIIIITITIIIIIII
15

25

:lIIITI

3O

35

PIIIIII

'TI

I
l

50 45 40

55

60

IIVIVTIIIIIIIIIIIIIIIII

65

IIIII.

70

 

- I} 98
[x
:r 7'8

 

80

I
I
l
2 Lit.
IIIIIIFIIIIIITI
5

l
85

-' I

l

l I
l

90

 

 

lllll

95

Figure 9 - 300 MHz 1H NMR of perylene in d-CHzClz. The structure of the
perylene molecule is shown in the inset.

36

0.5

1.0

r. 1::
2.0 1.5

I E

2.5

 

I:

08

3.0

IIITIT
45 40 3.5

l

5.0

WENT
55

Ariirri
6.0

6.5

'Vll

II‘EII
7.0

 

‘ "
II
iii A)?“ '

lb]
7
IO

1:"

s 1'}

““7.
80

Li 1,4 L____

l
I
8.5

90

 

9.5

Figure 10 — 300 MHz 1H NMR of 1-methylperylene in d-CHzClz. The structure of
the l-methylperylene molecule is shown in the inset.

37

08 -

06 ~

0A~~

 

 

02 -

 

(L0

 

normalized signal intensity, a.u.

I A I 4 I A I A I L I

300 350 400 450 500 550 600
wavelength, nm

Figure 11 - Absorbance and emission spectra for perylene in methanol. Data has
been normalized for presentation.

L0 -
08 ~
06 P

04 -

 

02 -

 

00

 

normalized signal intensity, a.u.

I A I L I A I A I I I

300 350 400 450 500 550 600

wavelength, nm

Figure 12 — Absorption and emission spectra for l-methylperylene in methanol.
Data has been normalized for presentation.

38

10.

ll.

12.

l3.

14.

15.

Literature Cited

. Jiang, Y.; Blanchard, G. J.; J. Phys. Chem, 1994, 98, 6436.

Brocklehurst, 8.; Young, R. N.; J. Phys. Chem, 1995, 99, 40.

Pauls, S. W.; Hedstrom, J. F.; Johnson, C. K.; Chem. Phys, 1998, 237, 205
Hambir, S. A.; Jiang, Y.; Blanchard, G. J .; J. Chem. Phys, 1993, 98, 6075.
Jiang, Y.; Blanchard, G. J .; J. Phys. Chem, 1994, 98, 9411.

Jiang, Y.; Blanchard, G. J.; J. Phys. Chem, 1994, 98, 9417.

McCarthy, P. K.; Blanchard, G. J .; J. Phys. Chem, 1995, 99, 17748.
McCarthy, P. K.; Blanchard, G. J .; J. Phys. Chem, 1996, 100, 5182.

Lewitzka, L.; Lohmannsroben, H.-G.; Strauch, M .; Luttke, W.; J. Photochem.
Photobiol. A: Chem, 1991, 61, 191.

Grimme, S.; Lohmannsroben, H.—G.; J. Phys. Chem, 1992, 96, 7005.
Calculated using Hyperchem V 4.0.

Jiang, Y.; Blanchard, G. J. J. Phys. Chem. 1995, 99, 7904.

Zieger, H. E.; Laski, E. M. Tetrahedron Lett. 1966, 32, 1966

Peake, D. A.; Oyler, A. R.; Heikkila, K. E; Liukkonen, R. J; Engroff, E. C.; Carlson,
R. M. Synth. Commun. 1983, 13, 21.

Campbell, A. D.; Elder, R. S.; Emerson, C. W. J. Chem. Soc. 1959, 3526.

39

Chapter 2

ORIENTATIONAL AND VIBRATIONAL RELAXATION DYNAMICS OF
PERYLENE AND l-METHYLPERYLENE IN N-ALKANOLS: SUBTLE
DIFFERENCES IN SOLVENT-SOLUTE INTERACTIONS‘

Introduction

Understanding the details of interactions between solvents and solutes has
received a great deal of attention by the chemistry community because of the importance
of these interactions in determining properties such as chemical reaction yield and
kinetics, or the ability to isolate one compound from another. Interactions between
solutes and their surrounding solvent molecules are difﬁcult to resolve because, unlike
solids, the spatial relationships between the molecules are not ﬁxed on time scales that
can be accessed using structural measurements such as X-ray diffraction or
multidimensional NMR spectrometry. lnterrnolecular interactions in the liquid phase are
more complex than those in the gas phase because of their characteristic strength, the
property that gives rise to the liquid phase and at the same time prevents a simple
statistical description of collisional interactions from providing adequate insight. For
these reasons, time domain spectroscopies have found particular favor in the investigation
of solvation processes.

There are essentially three time-resolved spectroscopic approaches to the
measurement of solvent-solute interactions. These are measurement of the time-delayed

j-
232 and

ﬂuorescence Stokes shift (TDFSS) of selected solutes,2'” rotational diffusion
vibrational population relaxation.3345 Of these, TDFSS measurements have provided the

most insight into the early-time separation of inertial and diffusive contributions to

40

solvent reorganization about excited solute molecules, and at the same time these reports
have proven to be the most controversial in terms of their interpretation.‘“"50 Perhaps the
most signiﬁcant issue that needs to be addressed regarding TDF SS measurements is the
lack of generality of the effect. Despite the broad conclusions drawn from these
experimental results, only a handful of chromophores with complicated spectroscopic
responses are known to exhibit resolvable transient Stokes shiftsz’51 The two other
approaches to the measurement of solvent-solute interactions do not offer the same time-
resolution as the TDFSS measurements, but the phenomena of rotational diffusion and
vibrational relaxation are not limited to a few chromophores - these are general effects,
and the characteristic time scales over which they proceed are consonant with chemical
reaction processes. We concentrate here on the rotational diffusion and vibrational
population relaxation dynamics of perylene and l-methylperylene in the n-alkanols
methanol through n-decanol. We have chosen these probe molecules because they are

relatively well characterized30'32“‘0'45

and because literature reports on their solution phase
dynamical behavior leave open some signiﬁcant questions.

Perylene has been studied more extensively than l-methylperylene. Several
reports on the reorientation dynamics of perylene point to solvents in which a two-

3 1'32 and others where there is only a single decay.30 It

component anisotropy decay is seen
is important to understand the fundamental basis for these results and the data we present
here provide some insight into this matter. Comparing the reorientation behavior of
perylene with that of l-methylperylene allows us to evaluate whether or not there is a

measurable contribution from (weak) dipolar solvation effects in additional to the

dominant van der Waals interactions. We reported previously that the reorientation of

41

perylene in n-alkanes pentane through hexadecane yielded single exponential anisotropy
decays30 and the nature of the solvent-solute frictional interactions changed as the length
of the solvent molecule became similar to that of the solute. For l-methylperylene in
these same solvents, we found that there was a clear break in the anisotropy behavior at
the same point, but the measured change was not simply a change in the frictional
boundary condition, but a shift in the functionality of the anisotropy decay from single to
double exponential.43 Such a change is consistent with a substantial alteration of the way
in which the l-methylperylene chromophore reorients in the longer alkane solvents.
These data point to a qualitative change in the nature of solvent-solute interactions that
depends on the relative lengths of the molecules, but a broader picture of the solvation of
these two molecules has yet to emerge.

We have measured vibrational population relaxation from the ~1375 cm'l ring
breathing modes of perylene and l-methylperylene to understand the role of “structural”
factors in mediating solution phase vibrational excitation transport. We have found that
both the spatial separation between vibrational donor and acceptor moieties and their
relative orientation are important in mediating vibrational excitation transport.45
Interestingly, it is not the acceptor density that determines the efﬁciency of transport.
The ability of the donor and acceptor vibrational coordinates to align with one another
mediates vibrational excitation transport for l-methylperylene in the alkanes. In contrast
to the reorientation measurements, the vibrational population relaxation dynamics of
these two chromophores cannot be compared directly because the efﬁciency of donor-
acceptor coupling is determined by the symmetry of the chromophores.43 Using our two-

pulse measurement scheme, systems that possess a center of inversion, such as perylene,

42

will necessarily couple to the solvent bath modes according to the modulation of the
solute’s quadrupole moment. For donors without a center of inversion, such as 1-
methylperylene, the modulation of the solute dipole moment by vibrational motion
mediates excitation transport. These two coupling processes operate over different length
scales (r'8 vs. r'6) and thus it is not possible to compare results directly. Despite the
differences of scale, the solvent-dependent trends observed in the T1 data for the two
molecules provide insight into the average spatial relationship between donor and
acceptor functionalities in solution.

Comparison of the reorientation and vibrational population relaxation dynamics
we report here for perylene and l-methylperylene in the n-alkanols yields signiﬁcant
information on the nature of the solvent-solute interactions the chromophores experience.
A key difference between the n-alkanes and n-alkanols is the extent to which solvent self-

_ 2 . .
22 5 We ﬁnd from the reorientation measurements

association determines their properties.
that self-association of the alcohols determines the immediate environment of the
chromophores to a signiﬁcant extent, especially for the longer n-alkanes. The vibrational
population relaxation data suggest that, for the longer n-alkanols, perylene appears to
associate with the non-polar aliphatic chains while the more polar l-methylperylene
interacts with the solvent alcohol functionality. These results are not surprising but do

reveal the point at which solvent size, relative to the chromophore, becomes important for

amphiphilic systems.

43

Experimental

Laser system: The picosecond pump-probe laser spectrometer used for both the
reorientation and vibrational population relaxation measurements has been described in
detail previously,” and we present only a brief outline of its performance characteristics
here. A mode-locked CW Nd:YAG laser (Coherent Antares 76-S) produces 30 W of
average power (1064 nm, 100 ps pulses, 76 MHz repetition rate). The output of this laser
is frequency-tripled to produce ~1.3 W of average power at 354.7 nm. The third
harmonic light is used to excite two cavity-dumped dye lasers (Coherent 702)
synchronously. Both lasers operate with Stilbene 420 laser dye (Exciton). The output of
each laser is ~100 mW average power at 8 MHz repetition rate with a pulse that produces
a 7 ps FWHM autocorrelation trace using a three plate birefringent ﬁlter. The pump laser
wavelength was set between 429.7 nm and 438.9 nm, depending on the chromophore and
solvent, while the probe laser was set in the range of 456.7 nm to 463.1 nm. The pump
wavelength was chosen to access the 0-0 transition of the chromophore and the probe
wavelength to stimulate emission from the S1v=0 state to the So":1 state, where the
vibrational resonance of interest is the 1375 cm'l v=l ring breathing mode. These
wavelengths were used for the vibrational relaxation and the reorientation measurements.
Our previous work has demonstrated that the ground state and excited state reorientation
behavior of perylene are identical30 and we assume that this is also the case for l-
methylperylene. The probe laser polarization was set alternately to 0° and 90° relative to
the pump laser polarization for the reorientation measurements, and to 54.7° for the
vibrational relaxation measurements. The time resolution of this system, ~10 ps, is

determined by the cross-correlation between the pump and probe laser pulse trains.

44

Detection of the transient signals was accomplished using a radio and audio frequency

triple-modulation scheme, with synchronous demodulation detection.“56

Each reported
time constant is the average of at least six individual determinations that are themselves
the average of seven to ten time-scans.

Steady State Spectroscopy: The steady state absorption spectra of both
chromophores in the n—alkanols were recorded with 1 nm resolution using a Hitachi U-
4001 spectrometer. The spontaneous emission spectra for the same solutions was
obtained with 1 nm resolution using a SPEX Fluorolog2 Model F1 1 lAT spectrometer.
These data were used to determine the appropriate laser wavelengths for each
chromophore/solvent pair. The absorption and emission spectra for perylene and l-
methylperylene in n-pentanol are shown in Figure 13 and Figure 14. We understand the
details of these spectra] proﬁles and describe them in the next section.

Chemicals and Sample Handling: Perylene (99.5%, sublimed) was obtained from
Aldrich Chemical Company and used as received. l-Methylperylene was synthesized by
using the procedure of Peake et al.,57 which is selective for methylation at the l-position,
and outlined in Chapter 1 of this dissertation. All other solvents were obtained from
Aldrich Chemical Company at the highest purity possible, and used as received. To
minimize thermal lensing considerations, sample solutions were ﬂowed through a 1 mm

path length quartz ﬂow cell. The sample temperature was controlled at 300 i 0.1 K using

a therrnostated bath.

45

 

 

 

 

linear response (a.u.)

 

 

A k A I . - 4

300 350 400 450 500 550 600

wavelength (nm)

 

Figure 13 — Absorption and emission spectra for perylene in n-pentanol. Data has
been normalized for presentation.

46

 

 

linear response (a.u.)

 

I n I n A I A l __

300 350 400 450 500 550 600

wavelength (nm)

 

Figure 14 — Absorption and emission spectra for l-methylperylene in n-pentanol.
Data has been normalized for presentation.

47

Results and Discussion

The focus of this chapter is on understanding the solvent-solute interactions of
perylene and l-methylperylene in the n-alkanols. Placing our results in the context of
previous work on these same chromophores will allow us to gain insight into the
dominant intermolecular interactions of these probe molecules. While the majority of the
information on the solvation behavior of these probes is obtained from the time domain
measurements, it is important to understand their steady state optical responses because
they are reﬂective of the average environment they experience.

Steady State Linear Spectroscopy. Among the reasons for using perylene and 1-
methylperylene for these studies is the well-behaved optical response of each molecule.
The quasi-mirror image absorption and emission spectra demonstrate the relatively small
role that vibronic coupling plays in these data58 and the Stokes shift of l-methylperylene
compared to that of perylene points to the presence of a small but ﬁnite permanent dipole
moment in l-methylperylene. PM3 semi-empirical calculations indicate u ~ 0.24 D in
the S0 and 0.53 D in the S1 for l-methylperylene.59

The linear responses of solutes can provide signiﬁcant insight into the dielectric
properties of the solvent. For example, the absorption maximum of the polar
chromophore oxazine 725 exhibits a solvent polarity-dependent red shift in n-alkanol
solvents up to n-heptanol.25 Comparison of those data to the oxazine 725 absorption
maximum in DMSO reveals that at least one contribution to the polarity-dependent
spectral shift is solvent H-bonding. For the chromophores we use here, H-bonding
almost certainly does not play the same role, but it is clear that there is a solvent-

dependence to their absorption maxima. Before considering this point, we need to

48

account for the differences in the linear response of perylene and l-methylperylene. The
absorption maximum of perylene is red shifted from l-methylperylene for a given solvent
and the individual features in the l-methylperylene spectra are less well-deﬁned. We
understand these differences as arising from the fact that the methyl group on 1-
methylperylene causes the two naphthalene moieties to be cocked at ~ 20° with respect to
one another.60’°' The resulting slight break in conjugation causes the blue shift and the
reduction in symmetry causes additional vibrational resonances to become allowed,
adding to the greater width of the absorption and emission spectral features. The solvent
dependence of the absorption maxima of perylene and l-methylperylene are shown in
Figure 15. First, we note that the solvent-dependence for perylene and l-methylperylene
appear to be identical and, if plotted in terms of relative shift with respect to the band
position in methanol (not shown) the results for the two chromophores are identical. We
understand these shifts in the context of the chromophore ground and excited states being
solvated more efﬁciently with increasing aliphatic chain length of the alkanol solvent.
The excited states of the chromophores are solvated more efﬁciently than the ground
states, presumably because van der Waals interactions between the solvent aliphatic
chains and the chromophore n' states are stronger than those for the ground state. These
data are reminiscent of those for oxazine 725 in the same solvents, except that we do not
see a saturation in the spectral red shift with increasing aliphatic chain length. We
interpret this difference in terms of the afﬁnity for the ionic dye oxazine 725 for the
alcohol functionality and the preference of perylene and l-methylperylene for the

aliphatic portion of the solvent. These data suggest that the balance of interactions

49

 

 

438 - . I
I I
A I I
E 436 h ' Perylene
3 I
I
S 434 - l
E
x 432 -
cc
5 .
D 430 - 0
U C I
c: o
“3 o
"e 428 r 0 l-Methylperylene
8 .
.o o
‘9 426 h
o
L I 4 I l I L I A I
0 2 4 6 8 10

solvent aliphatic chain length (no. of C's)

Figure 15 — Dependence of perylene and l-methylperylene absorption maxima on
solvent aliphatic chain length. When normalized for wavelength shift relative to
their absorption spectra in methanol, the solvent dependencies are identical for both
chromophores.

50

between the different functionalities of these solvents will be important in understanding
the transient data we consider next.

Molecular reorientation. Of the time domain spectroscopies used in the study of
solvent-solute interactions, molecular reorientation measurements are supported best by a
theoretical framework for the interpretation of the experimental data'62’67 The starting
point for relating reorientation data to solvent and solute properties is the modiﬁed

Debye-Stokes-Einstein (DSE) equation,°2'°3'°5*’7

= M
k BTS (6)

 

T0R

In Equation (6), n is the solvent viscosity, V the solute hydrodynamic volume
(225 A3 for perylene and 243 A3 for l-methylperylene),68 f is a frictional interaction term
to account for ﬁictional contributions to solvent-solute interactions, k3 is the Boltzmann
constant, T is the temperature and S is a shape factor to account for the non-spherical
shape of the solute. The DSE equation is clearly not intended to account for speciﬁc
molecular-level interactions between molecules, and it is clear that any description of
solvation phenomena based on solvent bulk properties must be incomplete. These
limitations notwithstanding, the modiﬁed DSE model provides remarkably close
agreement with many experimental results and it thus serves as a useful starting point in
any discussion of reorientation measurements.

Experimentally, the time-dependencies of the stimulated gain signal on the probe
beam, polarized parallel (1”(t)) and perpendicular (Ii(t)) to the pump beam, are used to
extract information on the re-randomization of the anisotropic distribution of

chromophores selected by the pump pulse, Equation (7):

51

___ IIIUJ‘IIU)

R t
( ) I"(t)+21i(t) (7)

 

The most common result for reorientation measurements is to recover an
experimental R(t) function that decays as a single exponential, with the time constant of
this decay being taken as TOR. Such data are often interpreted in the context of Equation
(6). There are a number of experiments, however, where R(t) is found to decay with a

multiple exponential functionality,23 '3 1 ‘42:“)

or depends on the manner in which the
chromophore is excited.”'32 In these cases, Equation (6) is clearly not sufﬁcient. Chuang
and Eisenthal have formulated the relationship between R(t) and solute properties such as
the angle between the excited and probed one photon transition moments and the actual
anisotropy in the diffusion constant.64 They found that R(t) can decay with as many as
ﬁve exponential components, although the most common case is that of a single
exponential. In cases where more than one decay component is resolved, it is possible to
interpret the details of the reorientation dynamics in substantial detail.

For the purposes of the following discussion, we consider that the chromophores
have their n-system in the approximate xy plane with x being the long, in-plane axis, and
z is perpendicular to the tt-system plane. For each chromophore our intention was to
assign the coordinate system such that the x-axis coincides with the absorption transition
moment. For perylene this assignment is coincident with speciﬁc bond axes because of
its high symmetry while for l-methylperylene it is not since the tt-system for this
chromophore is non-planar. Using these approximations and Chuang and Eisenthal’s
equations,"4 we can relate the experimental reorientation times to the shape of the volume

swept out by the rotating molecule. We describe this volume in terms of ellipsoids,

52

where a prolate ellipsoid, Equation (8) and Figure 16, is characterized by rotation
primarily about its long in-plane axis, D. > Dy ~ D2 and an oblate ellipsoid, Equation (9)
and Figure 17, by is characterized by rotation about the axis perpendicular to the
chromophore tt—system, DZ > Dx ~ Dy.

Prolate: R(t)=0.4exp(—6th) (8)

Oblate: R(t) = 0.3exp(—(2DX + 40, )t) + 0.1exp(—6Dxt) (9)

    

DY

Figure 16 — Prolate ellipsoid with Cartesian coordinate system superimposed. See
text for discussion.

 

 

Figure 17 — Oblate ellipsoid with Cartesian coordinate system superimposed. See
text for a discussion.

53

For both perylene and l-methylperylene in the n-alkanols, we measure
experimental R(t) functions that are ﬁt best by the sum of two exponential decays (Table
1, Table 2, Figure 18, and Figure 19.) The solvent dependence of the ﬁtted times is
presented in Figure 20. Using Equation (8) to interpret our data, we can obtain the
quantities D2 and D,. We show the diffusion constants as a function of solvent aliphatic
chain length in Figure 21, and their ratio in Figure 22. These data on l-methylperylene
are consistent with those reported in the n-alkanes because, in both cases, they trend
toward an increasing ellipsoidal aspect ratio with increasing solvent aliphatic chain
length. In our previous work,43 for n-alkanes octane and shorter, we recovered ratio
Dz/Dx < l and in this work, Dz/Dx ranges from ~4 to ~85. Also, in the n-alkanes, we see
that the ellipsoid aspect ratio saturates at ~ 8.5 for sufﬁciently long alkanes.43 For 1-
methylperylene in n-decanol, Dz/Dx ~ 8.5. It is not clear that this is a saturation value
because we have not used longer n-alkanols because n-dodecanol has been shown to form
a mesophase between 24°C and 30°C .22 Clearly the data on l-methylperylene in the two
families of solvents are not identical but this is not surprising because of the important
role that the alcohol functionality plays in mediating persistent organizationsz‘70 The
close correspondence between the two bodies of data, however, suggests that the same
fundamental physics are operative in both systems.

The data on perylene in the alkanes and alkanols provide less in the way of direct
correspondence. For n-alkanes up to n-hexadecane,32 methanol and ethanol we recover a
single exponential decay and for n-alkanols propanol through decanol we ﬁnd a two

component anisotropy decay. Perylene reorients as a prolate rotor in alkanes and as an

54

Table l — Reorientation times and zero-time anisotropies for perylene in the n-
alkanols. The data are the best fit results of the data to the function R(t) =
R‘(0)exp(-t/r,) + R2(0)exp(-t/1:2). Times are given in ps and the uncertainties listed
are standard deviations (il a) for at least six determinations of each quantity.

 

 

 

Solvent Rl(0) I] (ps) R2(0) 12 (ps)

CH3OH 0.23i0.06 l7i2 -- --
C2H50H 0.11i0.02 30i6 -- --

C 3H7OH 0.20:0.05 l6i4 0.10i0.03 59i17
C4H90H 0.17:0.05 27i8 0.06i0.03 101:39
C5H1 10H 0.19i0.09 21i7 0.11:0.03 82:16
C6H13OH 0.19i0.07 28:6 0.1 11003 1 10i22
C7H150H 0.2232005 35i7 0.10i0.03 l68i29
CanOH 0.21i0.04 32i5 0.11i0.01 157i15
C9H19OH 0.20i0.05 33.—1:15 0.14i0.04 146:23
CIOHZIOH 0.2031005 37i7 0.151002 217i45

 

 

55

 

Table 2 — Reorientation times and zero-time anisotropies for l-methylperylene in
the n-alkanols. The data are the best fit results of the data to the function R(t) =
R1(0)exp(-t/t.) + Rz(0)exp(-t/'cz). Times are given in ps and the uncertainties listed
are standard deviations (110') for at least six determinations of each quantity.

 

 

 

Solvent R1(0) 1'] (ps) R2(0) t2 (ps)

CH3OH 0.16i0.04 lli2 0.06i0.01 34 i 5
CszOH 0.25i0.05 l7i2 0.07i0.03 53 i 13
C3H7OH 0.28i0.03 18i2 0.09i0.03 62 i 6
C4H90H 0.19i0.05 17i3 0.12i0.04 69 j: 14
C5HHOH 0.19i0.06 21i7 0.12i0.05 81 i 27
C6H13OH 0.21i0.04 23:7 0.13i0.04 107 i 26
C7H15OH 0.23i0.03 29i4 0.14:0.02 139 i 15
CanOH 0.18i0.02 291-5 0.14i0.02 133 i 15
C9H190H 0.16i0.04 28i5 0.17i0.03 132 i 20
CIOHZIOH 0.21i0.02 40i6 0.1132002 206 i 21

 

 

56

 

signal intensity (a.u.)

 

 

 

 

 

I n I n J 1 I L I . I

-100 0 100 200 300 400 500
delay time (ps)

Figure 18 — Experimental I"(t) and li(t) scans for l-methylperylene in n-pentanol,
along with the instrumental response function. These data are typical of those
recorded for reorientation measurements.

57

03 -

02,-

R(t)

OJ r

 

OI)

 

I n I n I n I L I n J

-100 0 100 200 300 400 500
delay time (ps)

 

Figure 19 — Anisotropy function, R(t), for the data shown in Figure 18. The decay is
fit to the function R(t) = R1(0)exp(-t/t.) + R2(0)exp(-t/1:2)

58

300 -

250 r

I

200

150 '-

10R(p8)

100 -

IO-IH-I-I—O—A

.l—DIH

iii 1‘

O A I I I A I n I l I A I

0 2 4 6 8 10 12
solvent viscosity (cP)

Ur
O
' l
C III
IIOIl—I-l

 

Figure 20 — Reorientation time constants plotted as a function of solvent viscosity
for perylene (0) and l-methylperylene (I). The reorientation time constants are
identical to within the experimental uncertainty for the two chromophores in the
solvents used. The fast reorientation time is essentially solvent viscosity independent
and the dependence of the slow reorientation time does not depend linearly on
viscosity.

59

Table 3 — Cartesian components of the rotational diffusion constant, D, for perylene
in the n-alkanols. These quantities and their uncertainties are derived from the
fitted time constants shown in Table l.

 

 

 

 

Solvent Dx (GHz) Dz (GHz) DZ/Dx
CH3OH —— 9.8 i- 1.2 < l
CszOH -- 5.6 i 1.1 < 1
C3H7OH 2.8 i 0.8 14.2 i 3.6 5.0 i 0.4
C4H90H 1.7 i 0.6 8.4 i- 2.5 5.1 i 0.5
CSHHOH 2.0 i 0.4 10.9 i 3.6 5.4 i 0.4
C6H13OH 1.5 i 0.3 8.2 i 1.8 5.4 i 0.3
C7H150H 1.0 i 0.2 6.7 i 1.3 6.7 i 0.3
CanOH 1.1i0.1 731- 1.1 69:02
C9H|90H 1.1 i- 0.2 7.0 i: 3.2 6.1 i 0.5
CIoniOH 0.8 i 0.2 6.4 i 1.2 8.3 i 0.3

 

60

 

Table 4 — Cartesian components of the rotational diffusion constant, D, for l-
methylperylene in the n-alkanols. These quantities and their uncertainties are
derived from the fitted time constants shown in Table 2.

 

 

 

Solvent Dx (GHz) DZ (GHz) DZ/Dx
CH3OH 4.9 i 0.7 20.3 i 0.7 4.1 i 0.2
C2H50H 3.2 i 0.8 13.1 i 1.5 4.2 _+_ 0.3
C3H7OH 2.7 1r 0.3 12.5 i 1.4 4.7 i 0.1
C4H90H 2.4 i 0.5 13.5 i 2.4 5.6 i 0.3
C5HnOH 2.1 i 0.7 10.9 i 3.6 5.3 i 0.5
C6H13OH 1.6 i 0.4 10.1 i 3.1 6.5 i 0.4
C7H15OH 1.2 i 0.1 8.0 i 1.1 6.7 i 0.2
CanOH 1.3 i 0.1 8.0 i 1.4 6.4 i 0.2
C9HI9OH 1.3 i 0.1 8.3 i 1.5 6.6 i 0.2
CIOHZIOH 0.8 i 0.1 5.8 i 0.9 7.2 i 0.2

 

 

61

 

X

 

 

 

Q

825'

QN

..rzo-

C2

0).

CI

8.15-

e. i

8

210-} i
53’ i i
(D

85w;

0

g' igiu,,,.
..0

E LAILIIJA‘
“:0 2 4 6 8 10
Q

Solvent aliphatic chain length (no. of C's)

Figure 21 — Cartesian components of the rotational diffusion constant as a function
of solvent aliphatic chain length for perylene (D2 = O, D, = O) and l-methylperylene
(D1 = I, D, = CI) in the n-alkanols.

62

x7t' I
2N6: {Ii

5'. l I
Ii

I I I I I I I I I

0 2 4 6 8 10
Solvent aliphatic chain length (no. of C's)

f

A
I

 

Figure 22 — The ratio Dz/Dx as a function of solvent chain length for perylene (I)
and l-methylperylene (I). These data indicate that the environment experienced by
both chromophores is essentially identical and the structural anisotropy in the
solvent cage increases with solvent aliphatic chain length.

63

oblate rotor in the alkanols n-propanol and longer. In methanol and ethanol, we recover
what appears to be a single exponential decay, and note that our ability to distinguish
between a fast second component and a single decay is limited by temporal resolution
and signal-to—noise ratio. The apparently discontinuous change in D2 for perylene in
ethanol and n-propanol (Table 3) suggests that we have not resolved the fast second
component in R(t) for the smaller solvents. We measure Dz/Dx ~ 8.5 for perylene in n-
decanol and DZ/Dx < l for perylene in n-decane. This remarkable difference in aspect
ratios for one chromophore in the different solvents underscores the important role that
the presence of the alcohol group plays in the organization of the solvent.

It is interesting to note that, for both chromophores, the increasingly anisotropic
environment formed with increasing solvent aliphatic chain length is characterized by a
strong dependence of D2 on solvent identity and a much weaker solvent-dependence of
D,( (Figure 21). The dominant constraint imposed on the chromophores is for
reorientation out. of plane, suggesting a layered or quasi-lamellar environment. This
ﬁnding is consistent with our results on l-methylperylene in the n-alkanes and the fact
that we observe this phenomenon in all solvents argues for persistent local organization
over a length-scale that is at least on the order of the chromophore (vide infra). We note
also that this result is in excellent agreement with a recent report by the Johnson group on
two photon excitation induced anisotropy decays, where the form of the R(t) function was
found to depend measurably on the excitation polarization conditions.32

It is clear that the reorientation behavior of these two chromophores depends
sensitively on the solvent in which they are dissolved. With only the reorientation data, it

is possible to draw correlations between bulk solvent properties and microscopic

64

reorientation behavior, but the extent to which these correlations can provide real insight
into the relevant intermolecular interactions is limited. For this reason we have also
studied the vibrational population relaxation behavior of these chromOphores in the same
alkanol solvents. Those data, in conjunction with the reorientation measurements,
provide additional insight into the average organization of solvents about the solutes.

Before we discuss our vibrational population relaxation data, we consider what is
known about the solvent alcohols. NMR experiments have shown that, at room
temperature, the dominant form of the n-alkanols is the trimer,52 which is believed to
exist in a ring-bound conﬁguration.22 This ﬁnding indicates that the solvent bath mode is
heterogeneous on the length scale sensed by the solvent. Such local heterogeneity can, in
principle, be sensed by vibrational population relaxation as we discuss below. Another
signiﬁcant implication of the presence of trimers in solution is that it makes the dominant
form of the solvent have a hydrodynamic volume three times larger than that predicted
from calculations of monomer solvent.68 This is true in the limit of the characteristic
lifetime of the solvent trimers being on the same order as the dynamics we measure. For
the alkanols, the longitudinal relaxation time, TL, ranges from tens of ps to hundreds of ps
as the aliphatic chain length increases.70 With the larger effective solvent species, it is
reasonable to expect more restriction of the chromophores, consistent with our ﬁndings
of an effective oblate rotor shape in the n-alkanols and in the longer n-alkanes for l-
methylperylene.

Vibrational population relaxation measurements. Vibrational population
relaxation measurements provide information that is complementary to reorientation

measurements. While these two measurements sense fundamentally different

65

phenomena, both depend on the properties of the chromophore local environment. The
reorientation data we present here indicate that, in the n-alkanols, both perylene and 1-
methylperylene exist in environments that are restricted in such a way as to allow
chromophore rotation primarily about an axis perpendicular to the n-system plane. The
type of solvent local organization most consistent with these ﬁndings is that of a quasi-
lamellar medium. As we had mentioned above, the alkanol solvents likely exist primarily
as H-bonded trimers, and such a conﬁguration will likely give rise to a micro-
heterogeneous environment. With this picture in mind, we present our T1 data.

We have focused on the vibrational relaxation behavior of the perylene 1375 cm'1
and the l-methylperylene 1370 cm'1 ring-breathing modes.4045 We have chosen these
modes because the terminal CH3 group on the alkanol exhibits a rocking motion at ~1378
cm", essentially degenerate with the solute donor mode. In this manner, measurement of
T; as a function of alkanol identity can provide information on the spatial proximity and
relative orientation of the donor and acceptor vibrational coordinates. The details of how

the experimental technique operates have been presented elsewhere‘m'n’72

and we forego a
discussion of this point here. The vibrational population relaxation time constants we
report here are signiﬁcantly longer for perylene than for l-methylperylene in methanol
through n-heptanol. We understand this effect in the context of the nature of the coupling
between the donor and acceptor.“43 For molecules possessing a center of inversion, the
lowest order multipole moment that can be modulated using our excitation scheme is the
quadrupole moment. Assuming an IR-active acceptor mode, the intermolecular coupling

will decay with donor-acceptor spacing as f8.“ For a donor that does not possess a center

of inversion, the excited vibrational resonance will modulate the dipole moment and the

66

donor-acceptor coupling will depend on intermolecular distance as Eb.“ We have
synthesized and use l—methylperylene precisely because it is of lower symmetry than
perylene, and we have shown that the coupling between the ring breathing mode and the
solvent bath will be stronger than it is for perylene. Because the theoretical framework
for the interpretation of vibrational population relaxation in liquids is not well developed,
we do not attempt to extract solvent-solute interaction information from single T1 data
points. Rather, we are interested in the solvent—dependent behavior of these time
constants (Table 5 and Figure 23.)

We observe a break in the T1 trend for both chromophores at n-octanol. For 1-
methylperylenc, we see an abrupt increase in T1 with increasing alkanol chain length.
We interpret this effect as a change in the proximity of the solvent acceptor mode relative
to the solute donor mode. This ﬁnding suggests the association of the l-methylperylene
chromophore with the solvent alcohol functionalities. We can rationalize this assertion
based on the fact that l-methylperylene has a non-zero dipole moment and dipolar
interactions must therefore contribute to the solvent-solute interactions at some level.
When the alkanol chain becomes long enough that the terminal CH3 group can, on
average, be relatively isolated from the l-methylperylene chromophore, the donor-
acceptor coupling weakens, accounting for our experimental ﬁndings.

For perylene, the situation is not directly comparable. This is because the
functionality of the population relaxation dynamics data changes for n-octanol through n-
decanol. For solvents methanol through n-heptanol, we recover the expected form of the

signal shown in Figure 24 and Equation (10),

so) =aexp(—t/r...)—bexp(—t/TI) (10)

67

but for the longer alkanols the signal functionality changes to be a sum of exponentials,
shown in Figure 25 and Equation (11).
S(t)=aexp(—t/z's,e)+bexp(—t/r2) (11)

where 12 is the fast decay time. We have encountered this solvent-dependent change-
over previously for tetracene in n-hexadecane,44 and understand it in the context of the
solvent-solute interactions providing or mediating an additional non-radiative decay
pathway. In the previous work, we conjectured that intersystem crossing was responsible
for the observed solvent-dependent change in the functionality of S(t) because tetracene

has a relatively low (Iln and a modest (Disc-“’71"

For perylene, (in is ~ 0.94 and depends
only weakly on solvent.75 Intersystem crossing is thus less likely to account for our result
here. Because the origin of this new relaxation pathway is not known, there is ambiguity
in the assignment of the physical meaning of the recovered time constant. We therefore
will not attempt to cloud the discussion presented here with speculative interpretations of
these data. The fact that we see a change in the interactions between perylene and its
immediate environment at the same point that we observe a change for l-methylperylene
indicates that perylene is sensing either the same local environment as l-methylperylene
or its complement. For l-methylperylene, the T1 data are consistent with a measurable
contribution from dipole-dipole solvent-solute interactions, but for perylene, there is no
permanent dipole moment, precluding the contribution of such interactions. The

complement to the environment sensed by l-methylperylene would be for perylene to

interact most strongly with the aliphatic portion of the solvent, and this

68

Table 5 — Vibrational population relaxation times, T1, for perylene and 1-
methylperylene in the n-alkanols. Relaxation times are determined from at least six
individual data sets and the uncertainties are standard deviations. Times are values
from fits of the experimental data to Equation (10) and are given in ps. Times
marked with an asterisk are the results of ﬁts to Equation (11).

 

 

 

 

Perylene 1 -methylperylene

Solvent Tl (ps) TI (ps)
CH30H 209 i 68 40 i 18
CszOH 374 i 109 31. i 7
C3H70H 443 i 111 48 i ll
C4HqOH 389 i 119 126 i 20
CsHIIOH 397:109 36i11
C6HI3OH 180i52 60i11
C7H150H 439 i109 41 i 12
CanOH 54 i24* 160 i- 41
C9H190H 721-]6“ 2141225
CIOHZIOH 185i74* 99: 38

 

 

 

69

 

600 -

500 '-

 

 

I I I L L I I I _I j

0 2 4 6 8 10
Solvent aliphatic chain length (no. of C's)

Figure 23 — Vibrational population relaxation times for perylene (I) and 1-
methylperylene (El) as a function of solvent aliphatic chain length.

70

0.8 '-

elec

0.6 -
0.4 L

0.2 '-

4h...

IIIIILIIIIIII4+II

-100 0 100 200 300 400 500 600 700 800
delay time (ps)

signal intensity (a.u.)

 

 

 

 

 

Figure 24 — Magic angle data for perylene in n-pentanol with instrument response
function. See text for a discussion.

71

1.0 *-
0.8 -
0.6 ‘-
0.4 -

0.2 -

00—“1 e

II I4 I4I I I I III I I I I_I

~100 0 100 200 300 400 500 600 700 800
delay time (ps)

signal intensity (a.u.)

 

 

 

 

 

Figure 25 — Magic angle data for perylene in n-decanol. Note the difference in the
functionality of the population decay. See text for a discussion.

72

suggestion is in intuitive agreement with the dominance of van der Waals interactions
between (symmetric) PAHs and aliphatic solvents. Indeed, we observe this same
behavior for tetracene only in long-chain aliphatic solvents.

Taken collectively, the reorientation and vibrational population relaxation data for
perylene and l-methylperylene in the alkanols point to the importance of solvent self-
organization in determining solvent-solute interactions. Alcohols are known to form
persistent H-bonded networks with the lifetime of the network being proportional to the
aliphatic chain length.52‘7| NMR data point to the dominance of H-bonded trimers in
solution,52 and with the assumption that the characteristic time constant of the H-bonded
network is the same as the average lifetime for the trimers, it is likely that these species
dominate the solvent-solute interaction dynamics we sense. The reorientation data
suggest that both chromophores are constrained to reorient in an environment that is best
described as quasi-lamellar.43 The presence of solvent trimers implies structural
heterogeneity on the length scale of the solvent molecules, with relatively hydrophobic
and hydrophilic microenvironments in close spatial proximity. The vibrational
population relaxation measurements suggest that l-methylperylene (associates with the
more closely with polar portions of the solvent and this segregation becomes pronounced
for sufﬁciently large alkanol solvents. For perylene, it is likely that the chromophore
associates with the less polar environment, although it would be necessary to assign some
speciﬁc physical meaning to the second exponential decay seen for n-octanol through It-

decanol to augment this assertion.

73

Conclusions

We have studied the rotational diffusion and vibrational population relaxation
dynamics for perylene and l-methylperylene in a series of n-alcohols from methanol
through n-decanol. We ﬁnd evidence for solvent local organization, consistent with the
relatively long-lived hydrogen bonded complexes that characterize the alkanol solvents.
This local organization constrains the chromophores to reorient as oblate ellipsoids. The
reorientation times measured in each n-alkanol are the same for both chromophores. We
conclude that in these systems, the speciﬁc characteristics of the chromophore have less
of an effect on the rotor shape than the hydrogen bond-mediated solvent transient
structure.

The vibrational population relaxation times for perylene indicate that vibrational
energy transfer is less efﬁcient than it is for l-methylperylene in n-alkanols. This is an
expected result and we understand it based on the differences in the nature of donor-
acceptor coupling for the two chromophores. For the alkanols, it is the strong persistent
solute-solute interactions that mediate the solvent-solute interactions we measure. For
alkanes, the interactions between solvent and solute are on the same order as solute-solute
interactions, allowing the structure of the solute to play a more deterministic role in the
form of the recovered transient optical responses. We expect that, in systems where
dipole-dipole interactions are signiﬁcant but where H-bonding interactions are largely
mitigated, the difference in dynamics between these two chromophores will become more

pronounced.

74

10.
11.
12.

l3.

14.
15.
l6.
l7.
l8.
19.
20.

21.

Literature Cited

. This work was previously published - Goldie, S. N.; Blanchard, G. J. J. Phys. Chem.

A. 1999, 103, 999.

Shapiro, s. L.; Winn, K. R.; Chem. Phys. Lett., 1980, 71,440.

. Maroncelli, M.; Fleming, G. R.; J. Chem. Phys, 1987, 86, 6221.

Huppert, D.; Ittah, V.; Kosower, E.; Chem. Phys. Lett. 1989, 159, 267.
Chapman, C. F.; Fee, R. S.; Maroncelli, M.; J. Phys. Chem, 1990, 94, 4929.
Huppert, D.; Ittah, V.; Chem. Phys. Lett., 1990, 1 73, 496.

Jarzeba, W.; Walker, G. C .; Johnson, A. E.; Barbara, P. F.; Chem. Phys, 1991, 152,
57.

Wagener, A.; Richert, R.; Chem. Phys. Lett., 1991, 176, 329.

Fee, R. S.; Milsom, J. A.; Maroncelli, M.; J. Phys. Chem, 1991, 95, 5170.
Yip, R. W.; Wen, Y. X.; Szabo, A. G.; J. Phys. Chem1993, 97, 10458.
Fee, R. S.; Maroncelli, M.; Chem. Phys, 1994, 183, 235.

Sanders, M. J.; Wirth, M. J.;Chem. Phys. Lett., 1983, 101, 361.

Gudgin-Templeton, E. F .; Quitevis, E. L.; Kenney-Wallace, G. A.; J. Phys. Chem,
1985, 89, 3238.

Von Jena, A.; Lessing, H. E.; Chem. Phys, 1979, 40, 245.

Von Jena, A.; Lessing, H. E.; Ber. Bunsen-Ges. Phys. Chem, 1979, 83, 181.
Von Jena, A.; Lessing, H. E.; Chem. Phys. Lett., 1981, 78, 187.

Eisenthal, K. B.; Acc. Chem. Res, 1975, 8, 118.

Fleming, G. R.; Morris, J. M.; Robinson, G. W.; Chem. Phys, 1976, I 7, 91.
Shank, C. V.; Ippen, E. P.; Appl. Phys. Lett., 1975, 26, 62.

Millar, D. P.; Shah, R.; Zewail, A. H.; Chem. Phys. Lett., 1979, 66, 435.

Gudgin—Templeton, E. F.; Kenney-Wallace, G. A.; J. Phys. Chem, 1986, 90, 2896.

75

22. Blanchard, G. J .; Wirth, M. J.; J. Phys. Chem, 1986, 90, 2521.

23. Blanchard, G. J .; J. Chem. Phys, 1987, 87, 6802.

24. Blanchard, G. J .; Cihal, C. A.; J. Phys. Chem, 1988, 92, 5950.

25. Blanchard, G. J .; J. Phys. Chem, 1988, 92, 6303.

26. Blanchard, G. J .; J. Phys. Chem, 1989, 93, 4315.

27. Blanchard, G. J .; Anal. Chem, 1989, 61, 2394.

28. Alavi, D. S.; Hartman, R. S.; Waldeck, D. H.; J. Phys. Chem, 1991, 95, 6770.
29. Hartman, R. S.; Alavi, D. S.; Waldeck, D. H.; J. Phys. Chem, 1991, 95, 7872.
30. Jiang, Y.; Blanchard, G. J.; J. Phys. Chem, 1994, 98, 6436.

31. Brocklehurst, B.; Young, R. N.; J. Phys. Chem, 1995, 99, 40.

32. Pauls, S. W.; Hedstrom, J. F.; Johnson, C. K.; Chem. Phys, 1998, 23 7, 205.
33. Elsaesser, T.; Kaiser, W.; Annu. Rev. Phys. Chem, 1991, 42, 83.

34. Lingle, R., Jr.; Xu, X.; Yu, S. C.; Zhu, H.; Hopkins, J. B.; J. Chem. Phys, 1990, 93,
5667.

35. Anﬁnrud, P. A.; Han, C.; Lian, T.; Hochstrasser, R. M.; J. Phys. Chem, 1990, 94,
1180.

36. Heilweil, E. J .; Casassa, M. P.; Cavanagh, R. R.; Stephenson, J. C.; Ann. Rev. Phys.
Chem, 1989, 40, 143.

37. Heilweil, E. J .; Cavanagh, R. R.; Stephenson, J. C.; Chem. Phys. Lett., 1987, 134,
181.

38. Heilweil, E. J .; Cavanagh, R. R.; Stephenson, J. C.; J. Chem. Phys, 1989, 89, 230.

39. Heilweil, E. J .; Casassa, M. P.; Cavanagh, R. R.; Stephenson, J. C.; J. Chem. Phys,
1986, 85, 5004.

40. Hambir, S. A.; J iang, Y.; Blanchard, G. J.; J. Chem. Phys, 1993, 98, 6075.
41. Jiang, Y.; Blanchard, G. J.; J. Phys Chem, 1994, 98, 9411.

42. Jiang, Y.; Blanchard, G. J.; J. Phys. Chem, 1994, 98, 9417.

76

43

44.

45.

46.

47.

48.
49.
50.
51.
52.
53.
54.

55.

56.

57.

58.
59.

60.

61.

62.

63.

64

Jiang, Y.; Blanchard, G. J.; J. Phys. Chem, 1995, 99, 7904.
McCarthy, P. K.; Blanchard, G. J.; J. Phys. Chem, 1995, 99, 17748.
McCarthy, P. K.; Blanchard, G. J .; J. Phys. Chem, 1996, 100, 5182.
Agrnon, N.; J. Phys. Chem, 1990, 94, 2959.

Maroncelli, M.; Fee, R. S.; Chapman, C. F.; Fleming, G. R.; J. Phys. Chem, 1991,
95, 1012.

Blanchard, G. J.; J. Chem. Phys, 1991, 95, 6317.

Jiang, Y.; McCarthy, P. K.; Blanchard, G. J .; Chem. Phys, 1994, I83, 249.
Flory, W. C.; Blanchard, G. J .; Appl. Spec, 1998, 52, 82.

Castner, E. W.; Maroncelli, M.; Fleming, G. R.; J. Chem. Phys, 1987, 86, 1090.
Makarov, M. G.; Gus’kov, A. K.; Shvets, V- F.; Zh. F iz. Khim., 1982, 56, 71.
Jiang, Y.; Hambir, S. A.; Blanchard, G. J.; Opt. C ommun., 1993, 99, 216.

Bado, P.; Wilson, S. B.; Wilson, K. R. Rev. Sci. Instrum. 1982, 53, 706.

Andor, L.; Lorincz, A.; Siemion, J .; Smith, D. D.; Rice, S. A.; Rev. Sci. Instrum.
1984, 55, 64

Blanchard, G. J.; Wirth, M. J.; Anal. Chem, 1986, 56, 532.

Peake, D. A.; Oyler, A. R.; Heikkila, K. E.; Liukkonen, R. J .; Engroff, E. C.; Carlson,
R. M. Synth. Commun. 1983, 13, 21.

Karpovich, D. S.; Blanchard, G. J .; J. Phys. Chem, 1995, 99, 3951.
Calculated using Hyperchem V 4.0.

Lewitzka, L.; Lohmannsroben, H.-G.; Strauch, M.; Luttke, W.; J. Photochem.
Photobiol. A: Chem, 1991, 61, 191.

Grimme, S.; Lohmannsroben, H.-G.; J. Phys. Chem, 1992, 96, 7005.
Debye, P. Polar Molecules; Chemical Catalog Co.: New York, 1929; p. 84.
Perrin, F. J. Phys. Radium 1936, 7, l.

Chuang, T. J.; Eisenthal, K. B. J. Chem. Phys. 1972, 5 7, 5094.

77

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

75.

Hu, C. M.; Zwanzig, R. J. Chem. Phys. 1974, 60, 4354.

Youngren, G. K.; Acrivos, A. J. Chem. Phys. 1975, 63, 3846.
Zwanzig, R.; Harrison, A. K. J. Chem. Phys. 1985, 83, 5861.

Edward, J. T.; J. Chem. Ed, 1970, 47, 261.

Piston, D. W.; Bilash, T.; Gratton, E.; J. Phys. Chem, 1989, 93, 3963.
Garg, S. K.; Smyth, C. P.; J. Phys. Chem, 1965, 69, 1294.
Blanchard, G. J .; Rev. Sci. Instrum., 1996, 67, 4085.

Blanchard, G. J .; Anal. Chem, 1997, 69, 351A.

Birks, J. B.; Photophysics of Aromatic Molecules, Wiley-Interscience, London
(1970)

Douris, R. G.; Ann. Chim, 1959, 31, 479.

Berlman, I. B.; Handbook of Fluorescence Spectra of Aromatic Molecules, 2nd
Edition, Academic Press, 1971, p. 399.

78

Chapter 3

ORIENTATIONAL AND VIBRATIONAL RELAXATION DYNAMICS OF
PERYLENE AND l-METHYLPERYLENE IN ALDEHYDES AND KETONESl

Introduction

Over the past three decades, the scientiﬁc community has made enormous strides
in understanding intermolecular interactions in the liquid phase. Despite these gains, our
understanding of liquids lags behind that of solids and gases because of the associative
and dynamic nature of the liquid phase. Indeed, for a given solvent-solute system, there
are many solvent cage conﬁgurations that lie within a very narrow energetic window,
capable of interconversion on a fast timescale. Because of this intrinsic complexity,
gaining an understanding of the average environment that a solute molecule experiences
in a given solvent system is not a simple matter.

With the advent of picosecond lasers, the study of liquids expanded greatly
because many of the dynamics relevant to solvation processes proceed on the picosecond
or longer timescale, especially in polar, strongly associative systems. There have been
many methods devised for studying intermolecular interactions in solution and among the
most successful has been the measurement of probe molecule rotational motion. The
study of molecular reorientation has been aided by the existence of a well-established
theoretical framework for the interpretation of experimental data, and much insight into
solvent-solute interactions has been gained through such studiesz'22 One limitation of
rotational diffusion measurements lies in the amount of information that is available from
typical experimental data and also that the probe molecules used in such studies are

invariably larger than the length scale over which any solvent organization persists. As a

79

result of these limitations we have shown that by comparing the results of reorientation
and vibrational population relaxation measurements, we can gain additional insight into
the details of intermolecular interactions.”35 This is because, for many systems, dipolar
and/or hydrogen-bonding interactions dominate both the reorientation and vibrational
population relaxation events, but vibrational population relaxation operates over
characteristically shorter distances than are sensed by the reorientation measurements.

Of central importance to gaining an understanding of intermolecular interactions
is the appropriate choice of probe molecule. There have been many chromophores
examined in such studies, with many of them being polar or ionic dye molecules. The
information gained from such work is useful but in many cases less than ideal because of
the strength of the solvent-solute interactions. In an attempt to use a limited number of
relatively well-characterized chromophores in selected solvent systems, we have reported
recently on the reorientation and vibrational population relaxation dynamics of perylene
and l-methylperylene in a variety of non-polar and amphiphilic solvents. We have used
these two probe molecules because of their symmetries. Perylene is of D2}, symmetry,
possessing a center of inversion. It thus has no permanent dipole moment and, for the
vibrational population relaxation measurements we report here, the coupling between
solvent and solute is expected to scale as r'8 in the absence of substantial perturbation of
the chromophore by the solvent.36 l-Methylperylene is a probe molecule that is
structurally and spectroscopically similar to perylene, but because of the presence of the
methyl group at the 1— position, the center of inversion has been lifted and all vibrational
modes in this molecule are both infrared- and Raman-active. Consequently,

intermolecular coupling processes are expected to scale with r'6 for this molecule.36

80

Using both molecules allows us, in principle, to examine intermolecular interactions over
a range of length scales while the reorientation dynamics of the two remains very similar
in most solvent systems.

As part of our ongoing effort to understand intermolecular interactions in the
liquid phase, we report here on the reorientation and vibrational population relaxation
dynamics of perylene and l-methylperylene in a series of aliphatic aldehydes and
ketones. These two probe molecules exhibit relaxation dynamics that are the same for a
given solvent system, but their behavior in the different families of solvents provides
substantial insight into the dominant forces responsible for solvation of these molecules.
The reorientation of these probes in the aldehydes is reminiscent of their behavior in the
n-alcohols while the ketones data offer qualitative similarities to that seen in the n-
alkanes. From these ﬁndings, in conjunction with the vibrational population relaxation
measurements, we consider that longer-lived associative interactions between solvent
molecules (e. g. H-bonding) play a greater role than solvent dipole moment in determining
the dynamics of these probes.

In examining the data we report here, it is important to place the interpretation in
the context of other data that exist for these molecules. Our prior work on perylene and
l-methylperylene in n-alcohols showed that both molecules yielded the same dynamics to
within the experimental uncertainty for n-propanol through n-decanol, characterized by a
double exponential decay of the induced orientational anisotropy. Perylene and 1-
methylperylene behaved differently in the n-alkanes. For l-methylperylene we observed
a change in dynamical behavior between C3 and C9, as we see here for the ketones. We

also report here that the aldehydes and ketones exhibit a break in their dynamical

81

response at C5, whereas for the n-alcohols, the analogous break is seen either between C2
and C3 (perylene) or is not seen (l-methylperylene). We view these data as indicating
that the ketones and alkanes behave similarly, as do the aldehydes and n-alcohols. The
aldehydes and n-alcohols both possess a relatively labile proton capable of substantial
intermolecular interaction, although the aldehyde proton is less labile than the n-alcohol
proton. The presence of the labile protons in these solvents provides the opportunity for
signiﬁcant intermolecular interactions and thus more persistent solvent local
organization. Because the ketones and alkanes do not possess analogous protons, the
interactions between solvent molecules are weaker, thereby providing less conﬁnement
of the solute. A key ﬁnding of this work is that the solvent dipole moment plays a
relatively less important role than the presence of labile protons for the solvation of
essentially nonpolar species.
Experimental

Laser system: The picosecond pump-probe laser spectrometer used for both the
reorientation and vibrational population relaxation measurements has been described in
detail previously,37 and we present only a brief outline of its properties here. A mode-
locked CW Nd:YAG laser (Coherent Antares 76-S) produces 30 W of average power
(1064 nm, 100 ps pulses, 76 MHz repetition rate). The output of this laser is frequency—
tripled to produce ~1.2 W of average power at 355 nm. The third harmonic light is used
to excite two cavity-dumped dye lasers (Coherent 702-3) synchronously. Both lasers
operate with Stilbene 420 laser dye (Exciton). The output of each laser is ~60 mW
average power at 8 MHz repetition rate with a pulse that produces a 7 ps FWHM

autocorrelation trace using a three plate birefringent ﬁlter. The pump laser wavelength

82

was set between 433.0 nm and 440.1 nm, depending on the chromophore and solvent,
while the probe laser was set in the range of 460.8 nm to 468.4 nm for vibrational
population relaxation measurements of the solute ~1375 cm'l ring breathing modes, and
468.5 nm to 476.5 nm for measurements that examined the solute ~1733 cm'l
combination modes. The pump wavelength was chosen to access the 0—0 transition of the
chromophore and the probe wavelength to stimulate emission from the 81":0 state to the
S0":1 state of interest. The same wavelengths were used for both vibrational population
relaxation and reorientation measurements. Our previous work has demonstrated that the
ground state and excited state reorientation behavior of perylene are identical30 and we
take this to be the case for l-methylperylene as well. The probe laser polarization was set
alternately to 0° and 90° relative to the pump laser polarization for the reorientation
measurements, and to 54.7° for the vibrational population relaxation measurements. The
time resolution of this system, ~10 ps, is determined by the cross-correlation between the
pump and probe laser pulse trains. Detection of the transient signals was accomplished
using a radio and audio frequency triple—modulation scheme, with synchronous
demodulation detection.”40 Each reported time constant is the average of at least ﬁve
individual determinations that are themselves the average of ten to twenty-ﬁve time-
scans.

Steady State Spectroscopy: The steady state absorption spectra of both
chromophores in the aldehydes and ketones were recorded with 1 nm resolution using a
Hitachi U-4001 spectrometer. The spontaneous emission spectra for the same solutions
were obtained with 1 nm resolution using a Hitachi F-4500 spectrometer. These data

were used to determine the appropriate laser wavelengths for each chromophore/solvent

83

pair. The absorption and emission spectra for perylene and l-methylperylene in 3-
octanone are shown in Figure 26 and Figure 27.

Chemicals and Sample Handling: Perylene (99.5%, sublimed) was obtained from
Aldrich Chemical Company and used as received. l-Methylperylene was synthesized
using the procedure of Peake et al.,4| which is more than 95% selective for methylation at
the l-position, described in Chapter 1 of this disseration. All chemicals used in the
synthesis were purchased from Aldrich in the highest purity grade available.

Solvents used in this work were obtained from Aldrich in the highest purity grade
available. Solutions (~10“1 M) were made fresh before each set of experiments to
minimize oxidative effects. A 1 cm quartz cuvette with a magnetic stir bar was used to
contain and mix the sample during analysis, to eliminate the effects of thermal lensing. A

jacketed brass cuvette holder connected to a re-circulating water bath (300.0 :1: 0.1K),
fabricated in house, was used to maintain the sample temperature. Results from previous
experiments in our laboratory show no discemable difference between this sample
handling method and the use of a ﬂowing sample cell.
Results and Discussion

The focus of this work is on understanding the solvent-solute interactions of
Perylene and l-methylperylene in selected aldehydes and ketones, and to place these
results in context with respect to the behavior of these same probe molecules in other
solvent systems. In this work, it is the functionality of the induced orientational
anisotropy decay function, R(t), and in particular, how R(t) varies with the identity of the

S01Vent, that provides insight into the intermolecular interactions experienced by these

84

 

 

 

 

 

linear response (normalized)

4 I . r I

350 400 450 500 550

wavelength (nm)

Figure 26 — Normalized absorption and emission spectra of perylene in 3-octanone.

85

 

 

 

linear response (normalized)

I
' . 4 I . 1

350 400 450 500 550

wavelength (nm)

Figure 27 — Normalized absorption and emission spectra of l-methylperylene in 3-
octanone.

86

probe molecules. Vibrational population relaxation data indicate strong solvent-solute
interactions for both probe molecules. Before considering the dynamical data, we review
the linear responses of the probe molecules.

Steady State Linear Spectroscopy. Perylene is of D2,, symmetry with the ground
electronic state being of 1Ag symmetry and the ﬁrst excited singlet state being of 'B3u
symmetry.42 l-Methylperylene, by virtue of the presence of the methyl group, is of CI
symmetry, with all states having A character. These molecules exhibit qualitatively
similar linear responses; the quasi-mirror image absorption and emission spectra (Figure
26 and Figure 27) indicate that vibronic coupling plays a minor role in determining their
linear responses.43 l-Methylperylene exhibits a slightly larger Stokes shift than perylene,
reﬂecting the presence of a small permanent dipole moment. Semi-empirical calculations
indicate that it ~ 0.2 D in the So and ~ 0.5 D in the S1 for l-methylperylene.57 The
absorption maximum of perylene is red-shifted from l-methylperylene in a given solvent,
and the individual features in the l-methylperylene spectra are less well deﬁned. These
differences result from the presence of the methyl group on l-methylperylene, which
causes the two naphthalene moieties to be cocked at ~20o with respect to one another.“’45
The resulting slight break in conjugation causes the blue shift and the reduction in
symmetry causes additional vibrational resonances to become allowed, adding to the
greater width of the absorption and emission features. To within the experimental
resolution (1 nm), the absorption responses of both perylene and l-methylperylene are
solvent-independent for the solvents studied here.

Rotational diffusion measurements. Rotational diffusion has been used

extensrvely as a tool for understanding rnterrnolecular Interactions in solutron.‘2 ‘0 One

87

reason for our use of this technique is that it is supported by a sound theoretical
framework for the interpretation of the experimental data.46'5' The modiﬁed Debye-
Stokes-Einstein (DSE) (Equation (12)) is the usual starting point for relating reorientation

data to solvent and solute properties,°2‘(’3'°5'°7

an

 

70R =

where n is the solvent viscosity, V the solute hydrodynamic volume (225 A3 for perylene
and 243 A3 for l-methylperylene),52 f is a frictional interaction term to account for
frictional contributions to solvent-solute interactionsfia’ k3 is the Boltzmann constant, T
is the temperature and S is a shape factor to account for the non-spherical shape of the
solute.63 In this model, a single exponential decay of the induced orientational anisotropy,
R(t), is expected. For cases where a more complicated ﬁmctionality is seen in R(t), it is
clear that this simplistic but remarkably useful model cannot describe the data
adequately.

The experimental stimulated gain signal transients, polarized parallel (1”(t)) and

perpendicular (II(t)) to the pump beam, are related to the function R(t) through Equation

(13), and representative experimental l”(t), I i(t) and R(t) data are shown in Figure 28 and

 

Figure 29.
1 (I)-Ii(t)
Rt = “
() I“(t)+21i(t) (13)

The typical result for reorientation measurements is to recover an experimental

R(t) function that decays as a single exponential. There are a number of experiments,

however, where R(t) is found to decay with a multiple exponential functionality,”"’3’53‘54

or the measured decay is found to depend on the pump and/or probe wavelengthslz‘ss’56

88

 

 

 

 

1.0 r
0.8 L
’1‘ 0.6 r
3
EL
I“ ..
I: 0.4
<
0.2 '
...U

wvi‘f

4

4L

 

0

200 300 400 500

scan time (ps)

100

Figure 28 — Experimental I"(t) and I i(t) scans for l-methylperylene in 2-nonanone,
along with the instrumental response function.

89

 

 

 

0.3 r

0.2 '-
5? 0.1 -

H
‘1
.‘Ul " I
00 . ‘III....-.._. I. I'LII". I, 'I I1 I'
0 100 200 300 400 500

scan time (ps)

Figure 29 — Anisotropy function, R(t), generated from the raw data shown in Figure
28. The decay is ﬁt to the function R(t) = R1(0)exp(-t/tl) + Rz(0)exp(-t/tz), with the
results of the fit given in Figure 28.

90

In these cases, a more thorough study of the anisotropy decay dynamics is required.
Chuang and Eisenthal have derived expressions for I“(t) and I i(t) that depend on the angle
between the excited and probed transition dipole moments and the Cartesian components
of the rotational diffusion constant.“ R(t) can decay with as many as ﬁve exponential
components, and in cases where more than one decay component is resolved, it is
possible to interpret the reorientation dynamics in substantial detail.

To facilitate the treatment of our experimental data using Chuang and Eisenthal’s
formulation, we assign the Cartesian axes as follows: The chromophore it-systems are in
the (approximate) xy plane and z is perpendicular to the xy plane. The x-axis coincides
with the absorption transition dipole moment. For perylene this assignment is coincident
with speciﬁc bond axes, consistent with its D2}, symmetry. For l-methylperylene, the
assigned x-axis is not necessarily coincident with any bonds because the rt-system for this
chromophore is non-planar and the presence of the methyl group at the 1- position will
affect the orientation of the transition dipole moment relative to the molecular
framework. Using these approximations with Chuang and Eisenthal’s equations,“ we
can relate the experimental reorientation times to the shape of the volume swept out by
the rotating molecule. Because the experimental data do not possess sufﬁcient
information to separate the three Cartesian components of the rotational diffusion
constant unambiguously, we describe this volume in terms of prolate or oblate ellipsoids.
For a prolate ellipsoid, characterized by rotation primarily about its long in-plane axis, Dx
> Dy = Dz. An oblate ellipsoid is characterized by rotation about the axis perpendicular to

the chromophore n-system, DZ > Dx = Dy. Using these geometric approximations, we

91

obtain expressions for R(t) that can be used directly in the interpretation of our
experimental data.
Oblate: R(t) = 0.3exp(—(ZD,r + 40, )t) + 0.1exp(—6Dxt) (14)
Prolate: R(t) = 0.4exp(—6th) (15)

Where the excited and probed transition dipole moments are assumed to be parallel and
lie along the x-axis (vide infra). In our previous work on the reorientation dynamics of
perylene and l-methylperylene in alkanes and n-alcohols, we have interpreted much of
the data in the context of Equations (14) and (15). We found that, for reorientation of
perylene in alkanes, we always recovered a single exponential anisotropy decay, but the
nature of the frictional contributions to the reorientation of the chromophore (f in
Equation ( 12)) decreased with increasing solvent aliphatic chain length.30 The
reorientation of l-methylperylene in the alkanes appears to be signiﬁcantly different.43
We observed a single exponential anisotropy decay in pentane, hexane, heptane and
octane. For nonane, we observed a two-component R(t) decay, and interpreted these
results in terms of a solvent-dependent change in the effective rotor shape of the
chromophores. As the solvent aliphatic chain length increased, we quantitated the aspect
ratio of the ellipsoid of rotation, Dz/Dx, saturated at a value of ~8.5, suggesting that, as
the solvent increases in length, the conﬁnement of the l-methylperylene chromophore
becomes more pronounced.

For the reorientation of perylene and l-methylperylene in the n-alcohols, we
observed a two-component decay of R(t) for all cases save for perylene in methanol and
ethanol. We argued that the observation of a single component decay for these two

solutions is the result of limited time resolution achievable with our system rather than an

92

explicit change in effective rotor shape. With this assertion, the results for probe
reorientation in the n-alcohols suggest an oblate rotor shape in all cases and thus
substantial solvent conﬁnement of the probe. This interpretation is consistent with the
known strong solvent-solvent interactions present in these solvents. As with the alkanes,
for long chain n-alcohols we observe a saturation in the ratio Dz/Dx at ~8.5, arguing for
substantial solute conﬁnement. Our results for the reorientation of perylene and l-
methylperylene in aldehydes and ketones are consistent with our ﬁndings in the alkanes
and n-alcohols and, when these bodies of information are viewed as a whole, it is clear
that the role of labile solvent protons plays a more signiﬁcant role in solvation processes
than does solvent permanent dipole moment.

We present the reorientation time constants for perylene and l-methylperylene in
the aldehydes in Figure 30. The data show that the reorientation dynamics of the two
probe molecules are identical in each aldehyde studied, to within our experimental
uncertainty. We do not view this ﬁnding as surprising given the structural similarity of
the two probes. Of perhaps greater signiﬁcance is the clear change in behavior from
single exponential to double exponential decay in going from pentanal to hexanal. As
discussed above, these data indicate a change in the shape of the volume swept out by the
probe molecule(s) from prolate to oblate. In contrast to our previous measurements, we
observe that the DZ/Dx ratio does not change substantially in going from hexanal to
decanal, and its limiting value in the aldehydes is ~ 5.5. These two ﬁndings, a change in
effective rotor shape between C5 and C6, and a DZ/Dx ratio of ~5.5, suggests (but does not
prove) the existence of substantial solvent-solvent interactions in the aldehydes. We will

return to a discussion of this point later.

93

180 ‘-
160;
140; I
120': i

100

 

 

80

rOR(pS)
' I
II
l—-l-OI—l
,5

60

20:1; i I it i I

0 L n i A L l I J I k I L I A I

3 4 5 6 7 8 9 10
solvent carbon chain length (no. of C's)

I I rﬁ

I

 

Figure 30 — Reorientation time constants plotted as a function of carbon chain
length for perylene (0,0) and 1-methylperylene(I,El) in the aldehydes.

94

The reorientation dynamics of perylene and l-methylperylene in selected ketones are
shown in Figure 31. These data are for ketones of three different lengths; C7, C3 and C9,
and the chemical variable in this series of data is the location of the carbonyl
functionality. As with the aldehydes, the reorientation behavior of both probe molecules
is identical to within the experimental uncertainty for a given solvent. We note that, for
the ketones, the reorientation behavior exhibits a change between C3 and C9 total solvent
length, and this ﬁnding is reminiscent of the experimental reorientation data for 1-
methylperylene in the n-alkanes. For the C9 ketones, we ﬁnd that the ratio of Dz/Dx does
depend on the location of the carbonyl functionality, with 2-nonanone exhibiting a Dz/Dx
ratio of ~4.5, substantially less than the DZ/Dx ~ 5.5 ratio seen for 3-, 4- and 5-nonanone
(Table 4). While the ketones possess signiﬁcant permanent dipole moments at their
carbonyl functionality (calculated p. = 2.5 D),57 and this interaction appears to play a role
in the carbonyl position-dependent DZ/Dx ratio seen for the ketones, we do not see
evidence for strong solvent-solute interactions between the probe molecules and these
solvents with the reorientation measurements.

Previous studies of the rotational dynamics of l-methylperylene in n-alkanes
exhibited similar behavior to what we have observed here.43 The alkanes do, indeed, exert
a substantial structural inﬂuence in the “free volume” that is accessible to the
chromophore molecules. The break in behavior in going from C3 to C9 in both the
alkanes and the ketones reﬂects the balance between the strength of solvent-solvent and
solvent-solute interactions. The change in behavior from prolate rotor to oblate rotor is
consistent with conﬁnement of the probe molecules in a layered or quasi-lamellar

environment.

95

 

120p
110 -
100 -
90 L
80:
70L

1331 I II
33: ! MM

1
r

 

 

 

IOR(ps)

 

 

 

 

 

0 I I I I I I I I I
2 3 4 2 3 2 3 4 5
Heptanone Octanone Nonanone

carbonyl position (carbon number)

Figure 31 — Reorientation time constants plotted as a function of solvent viscosity
for perylene (0,0) and l-methylperylene (I,El) in the ketones. The boxed regions
indicate specific solvent lengths.

96

The experimental anisotropy data provide information on both the reorientation
and spectroscopic properties of the probe molecules. Equations (14) and (15) were
derived with the assumption that the transition moments of the excited and probed
transitions are parallel and lie along the x-axis. For these conditions, we expect a zero
time anisotropy of 0.4 for an oblate rotor and zero-time components of 0.1 and 0.3 for an
oblate rotor. Experimentally, we observe values close to the theoretical predictions
(perylene in aldehydes: Table 6; l-methylperylene in aldehydes: Table 7; perylene in
ketones: Table 8; l-methylperylene in ketones: Table 9), but there are noticeable
deviations in the zero-time anisotropies in all cases. The Johnson group has noted similar
results for the reorientation of perylene in 2-propanol and cyclohexane42 and these data

30‘43'59 J as et al. argue that at least a portion of the

are consistent with our earlier reports.
experimental deviation of the R(O) values stems from contributions to the data from
excited state absorption. This explanation is consistent with the known excited state
absorption response of perylene.58 Unfortunately, we do not have experimental data on
the excited state absorption spectra of perylene or l-methylperylene in the solvents we
report on here and thus cannot infer the relative contribution of this effect.

We have touched upon the issue of diffusion constant ratio above, but a more
detailed discussion of these data is in order. We present the calculated values for
perylene (Table 6) and 1-methylperylene (Table 7) for the aldehydes and Table 8 and
Table 9 for the ketones. It is useful to compare these results to our earlier data. For 1-

methylperylene in n-alkanes octane and shorter, we recovered ratio DZ/Dx < 1, and a

DZ/Dx approaching ~8.5 in the longer chain solvents, and found that this ellipsoid aspect

97

Table 6 — Reorientation Times and Zero-Time Anisotropies for Perylene in the n-

 

 

 

Aldehydes.‘
Perylene

Solvent R1(0) I] (ps) R2(0) 12 (ps)
C 2H5CHO 0.23:0.04 18:5 -- --
C3H7CHO 0.28:0.01 24:4 -- --
C4H9CHO 0.30:0.05 21:4 -- --
C5H1 .CHO 0.24:0.10 21:8 0.20:0.10 86:4
CbH l 3CHO 0.21:0.10 21:6 0.12:0.05 79:13
C7H15CHO 0.19:0.05 24:7 0.11:0.04 116:36
CanCHO 0.15:0.05 17:7 0.19:0.02 92:15
C9H19CHO 0.23:0.06 22:4 0.14:0.02 104:16

 

 

 

“The data are the best ﬁt results of the data to the function R(t) = RI(0)exp(-t/tl) or R(t) =

R1(0)exp(-t/tl) + R2(0)exp(—t/t2). Times are given in ps and the uncertainties listed are

standard deviations (:16) for at least six determinations of each quantity.

98

 

Table 7 — Reorientation Times and Zero-Time Anisotropies for l-Methylperylene in
the n-Aldehydes.a

 

l-Methylperylene

Solvent R1(0) 1:, (ps) R2(0) 1:; (ps)
CZHSCHO 0.26:0.14 25:10 -- --
C3H7CHO 0.25:0.04 20:3 -- --
C4H9CHO 0.23:0.04 34:6 -- --
Cer ICHO 0.25:0.11 17:5 0.12:0.05 102 : 28
C6H13CHO 0.26:0.08 21:8 0.10:0.06 98 : 19
C7H15CHO 0.15:0.04 21:4 0.12:0.04 84 : 2
CanCHO 0.16:0.07 17:7 0.20:0.02 97 : 10
CqHIgCHO 0.23:0.04 24:12 0.17:0.06 138 : 27

 

 

 

 

 

 

8The data are the best ﬁt results of the data to the function R(t) = R1(0)exp(-t/tl) or R(t) =
R1(0)exp(-t/tl) + R2(0)exp(-L/12). Times are given in ps and the uncertainties listed are

standard deviations (:10) for at least six determinations of each quantity.

99

Table 8 — Reorientation Times and Zero-Time Anisotropies for Perylene in the
Ketones.’

 

Perylene
Solvent R1(0) 1] (ps) R2(0) 12 (ps)
2-heptanone 0.24:0.05 31:5 -- --
3-heptanone 0.34:0.03 25:3 -- --
4-heptanone 0.20:0.01 26:3 -- --
2-octanone 0.19:0.03 27:2 -- --

 

 

3-octanone 0.20:0.02 33:3 -- --

2-nonanone 0.23:0.03 17:2 0.12:0.01 57:10
3-nonanone 0.29:0.09 19:5 0.09:0.04 88:10
4-nonanone 0.30:0.17 18:6 0.12:0.06 74:21
5-nonanone 0.31:0.12 25:5 0.09:0.05 93:23

 

 

 

 

aThe data are the best ﬁt results of the data to the function R(t) = R.(0)exp(-t/rl) or R(t) =
Rl(0)exp(-t/tl) + R2(0)exp(-t/t2). Times are given in ps and the uncertainties listed are

standard deviations (:10) for at least six determinations of each quantity.

100

Table 9 - Reorientation Times and Zero-Time Anisotropies for l-Methylperylene in
the Ketones.‘l

 

l-methylperylene

Solvent R1(0) 1'] (ps) R2(0) 132 (ps)
2-heptanone 0.27:0.05 11:2 -- ~-
3-heptanone 0.26:0.05 1 7:2 -- --
4-heptanone 0.24:0.03 1 8:2 -- --
2-octanone 0.27:0.05 17:3 -- --
3-octanone 0.34:0.06 21:7 -- --
2-nonanone 0.25:0.04 14:7 0.12:0.07 54 : 9
3-nonanone 0.31:0.03 17:8 0.14:0.05 71 : 17
4-nonanone 0.32:0.10 20:8 0.11:0.05 98 : 15
5-nonanone 0.37:0.20 17:6 0.13:0.05 94 : 22

 

 

 

 

 

 

3The data are the best ﬁt results of the data to the function R(t) = R1(0)exp(-t/rl) or R(t) =
R1(0)exp(-t/tl) + R2(0)exp(-t/‘cz). Times are given in ps and the uncertainties listed are

standard deviations (:10) for at least six determinations of each quantity.

101

ratio saturates at ~ 8.5 for sufﬁciently long alkanes.43

For both perylene and l-
methylperylene in the n-alcohols, we recovered Dz/Dx ranges from ~4 in methanol to
~85 in n-decanol.59 It is not clear that Dz/Dx ~ 8.5 is a saturation value because we are
limited in the length of the n-alcohols we can use. n-Dodecanol has been shown to form
a mesophase between 24°C and 30°C,60 and we must therefore expect similar phenomena
for even longer chain n-alcohols near their freezing point. In the work we present here,
we recover DZ/Dx < 1 for l-methylperylene in aldehydes pentanal and smaller, and in the
C7 and C3 ketones, regardless of carbonyl functionality position. In the longer chain
aldehydes and ketones, the ratio of DZ/Dx does not reach the same level seen for l-
methylperylene in n-alkanes and n-alcohols, but does appear to saturate (Table 10, Table
11, Table 12, Table 13).

It is important to note that long—chain solvents are not the only ones that exhibit
double exponential anisotropy decay behavior. In a recent study by J as et al.,42 perylene
in cyclohexane and 2-propanol were shown to exhibit double exponential decay
anisotropy decays, and the authors concluded that the fast reorientation about the z-axis
resulted from conﬁnement of the perylene molecule. Their conclusions are in full
agreement with ours. The phenomenon of (planar or quasi-planar) solute conﬁnement
within solvent lamellar-like structures appears to be reasonably general and it will be
important to elucidate the nature of the solvent-solvent intermolecular forces that give
rise to this sort of solute conﬁnement.

Despite the wealth of information available from reorientation measurements, we
would like to achieve an understanding of solvent—solute interactions that these

measurements alone cannot provide. In an attempt to garner further insight into solvation

102

Table 10 — Cartesian components of the rotational diffusion constant, D, for
perylene in the aldehydes.“

 

 

 

Perylene
Solvent Dx (GHz) D. (GHz) Dz/Dxb
CszCHO -- 9.0 : 2.6 < 1
C3H7CHO -- 7.0 : 1.1 < 1
C4H9CHO -- 7.9 : 1.4 < 1

C5HHCHO l.9:0.1 10.7:4.2 5.5 :0.4
C6H13CHO 2.1 : 0.3 10.9 : 3.2 5.2 : 0.3
C7H15CHO 1.4 : 0.5 9.8 : 2.8 6.8 : 0.4
CanCHO 1.8 : 0.3 13.9 : 5.5 7.6 : 0.4
C9HI9CHO 1.6 : 0.2 10.8 : 2.0 6.7: 0.2

 

 

 

 

aThese quantities and their uncertainties are derived from the ﬁtted time constants shown
in Table 6. bThe uncertainties in the quantities Dz/Dx were calculated from the standard

deviations in the DZ and the DK data according to the formula

 

0070. =(Dz/D.)\[0'p, /Dx)2 +(0'D/DZ)2 .

103

Table 11 — Cartesian components of the rotational diffusion constant, D, for 1-
methylperylene in the aldehydes.a

 

1 -methylperylene

 

 

Solvent Dx (GHz) Dz (GHz) DZ/Dxb
C2H5CHO -- 6.6 : 2.6 < 1
C3H7CHO -- 8.3 : 1.2 < 1
C4H9CHO -- 4.9 : 1.0 < 1

C5Hl ICHO 1.6 : 0.4 13.6 : 3.9 8.4 : 0.4
C6H|3CHO 1.7 : 0.3 11.2 : 4.5 6.6 : 0.4
C7H15CHO 2.0 : 0.1 11.0 : 2.0 5.6 : 0.2
CanCHO 1.7 : 0.2 13.5 : 5.1 7.9 : 0.4
C9H19CHO 1.2 : 0.2 9.8 : 4.9 8.1 : 0.5

 

 

 

 

aThese quantities and their uncertainties are derived from the ﬁtted time constants shown
in Table 7. bThe uncertainties in the quantities DZ/Dx were calculated from the standard

deviations in the DZ and the Dx data according to the formula

 

6WD} = (D, /D, )\/(a,,‘ /D, )2 +(al, /D, )2 .

104

Table 12 — Cartesian components of the rotational diffusion constant, D, for
perylene in the ketones.II

 

 

 

Perylene
Solvent Dx (GHz) Dz (GHz) Dz/DJ’
2-heptanone -- 5.3 : 0.9 < 1
3-heptanone -- 6.7 : 0.7 < 1
4-heptanone -- 6.3 : 0.6 < 1
2-octanone -- 6.3 : 0.5 < 1
3-octanone -- 5.1 : 0.5 < 1

2-nonanone 3.0 : 0.5 13.4 : 1.4 4.5 : 0.2
3-nonanone 1.9 : 0.2 12.0 : 3.4 6.4 : 0.3
4-nonanone 2.3 : 0.7 12.4 : 4.4 5.5 : 0.5
5-nonanone 1.8 : 0.5 9.2 : 2.0 5.1 : 0.3

 

 

 

 

aThese quantities and their uncertainties are derived from the ﬁtted time constants shown
in Table 8. bThe uncertainties in the quantities Dz/Dx were calculated from the standard

deviations in the DZ and the Dx data according to the formula

 

GDJD. =(Dz/D_,)\/(0'D‘ /Dx)2 +(ch_/D_,)2 .

105

Table 13 — Cartesian components of the rotational diffusion constant, D, for 1-
methylperylene in the ketones.a

 

l -methylperylene

 

 

Solvent D, (GHz) D. (GHz) DZ/D,B
2-heptanone -- 5.4 : 1.1 < 1
3-heptanone -- 5.5 : 1.4 < 1
4-heptanone -- 6.7 : 1.0 < 1
2-octanone -- 5.0 : 1.1 < 1
3-octanone -- 5.7 : 0.6 < 1

2-nonanone 3.1 : 0.5 16.0 : 7.7 5.2 : 0.5
3-nonanone 2.3 : 0.6 13.6 : 6.5 5.8 : 0.5
4-nonanone 1.8 : 0.3 12.7 : 6.5 7.2 : 0.5
5-nonanone 1.8 : 0.4 14.0 : 5.2 7.8: 0.4

 

 

 

 

aThese quantities and their uncertainties are derived from the ﬁtted time constants shown
in Table 9. bThe uncertainties in the quantities Dz/Dx were calculated from the standard

deviations in the D2 and the DK data according to the formula

 

any, = (D, /D,)\/(o,, /D, )2 +(ol, /D, ) .

106

processes in these systems, we consider the vibrational population relaxation dynamics of
perylene and l-methylperylene in the aldehydes and ketones next.

Vibrational population relaxation measurements. Vibrational population
relaxation measurements provide information that is complementary to reorientation
measurements. While reorientation and vibrational population relaxation sense
fundamentally different phenomena, both depend on the properties of the chromophore
local environment. We have focused on the vibrational population relaxation of the
perylene 1375 cm'l and the l-methylperylene 1370 cm‘I ring breathing modes.43 We have
chosen these modes because the solvent terminal CH3 groups exhibit a rocking motion at
~1390 cm"1 in the aldehydes, and ~1375 cm'1 in the ketones. In addition to this
donor/acceptor pair, perylene and l—methylperylene both possess a second ring-breathing
mode, a combination mode (1375 cm'l + 358 cm“), at ~1733 cm’1 and ~1730 cm",
respectively, that is essentially degenerate with the aldehyde and ketone carbonyl
stretching modes at ~l 727 cm'1 and ~l7l7 cm", respectively. The measurement of T. as
a function of solvent identity for these modes can provide information on the spatial
proximity and relative orientation of the donor and acceptor vibrational coordinates. We
have chosen the series of ketones to address the proximity of the C=O and -CH3 solvent
functionalities to the selected chromophore vibrational coordinates. The experimental

“‘62 and we forego a discussion of

details of this technique have been presented elsewhere
this matter here. We Show a representative Tl data set and instrument response function

in Figure 32. The vibrational pOpulation relaxation time constants we report here are

Similar for the two chromophores in each of the ketones studied. We present

107

 

.0
O\
m

AT/T (a.u.)
O
4:.

 

 

 

 

_l_ I I I

0 100 200 300 400 500
delay time (ps)

k

Figure 32 — Normalized vibrational population relaxation data (probe polarized at
54.7° with respect to the pump) and instrument response function for perylene in 3-
octanone.

108

these T. values in Table 14 and Figure 33 and Figure 34. For all of the aldehydes
examined, the T. times were unresolvably fast, implying strong solvent-solute coupling.
For the ketones, the similarity of the T1 data for the two probe molecules is unexpected
because perylene possesses a center of inversion and l-methylperylene does not.
Molecules possessing a center of inversion couple with their immediate environment with
a distance—dependence of r'g, and molecules lacking a center of inversion couple with a
distance-dependence of {6.36 The fact that the T] values are so similar for the two probe
molecules combined with the similarity in their reorientation times argues for a
substantial perturbation of the chromophore symmetry by the solvent environment.
Because the theoretical framework for the interpretation of vibrational population
relaxation in liquids is not well developed, we do not attempt to extract information on
solvent-solute interactions from individual T; values. Rather, we are interested in the
solvent-dependence of these time constants (Figure 33 and Figure 34). The observed
trends for the ketone carbonyl acceptor mode closely follow those of the terminal methyl
group acceptor mode, and this is an expected result since the donor mode (~1733 cm“) is
a combination of the 1375 cm'1 ring breathing mode and the 358 cm'1 in-plane stretching
mode. The ketone terminal methyl groups are in close proximity to the solute in all
cases, and any additional coupling between the solute (donor) combination mode and the
solvent carbonyl stretch is inefﬁcient because of the facile pathway for relaxation through
the 1375 cm'I ring breathing mode.
The picture that emerges from the reorientation and T. data is that, for all of the ketones,

the solvent terminal methyl groups remain in relatively close proximity to the solute; that

109

Table 14 — Vibrational population relaxation times, T1, for perylene and 1-
methylperylene in the ketones.

 

 

 

Perylene 1 -methylperylene
Solvent T. — 1370 cm" Tl — 1733 cm" I, — 1375 cm" T. — 1733 cm'T
3-pentanone 15:10 13:6 11:10 20:14
3-heptanone 16 : 6 40 : 21 20 : 11 22 : 2
3-octanone 34 : 17 47 : 30 8 : 5 50 : 11
3-nonanone 21 : 7 26 :14 35 :14 35 :13

 

 

 

 

 

Relaxation times are determined from at least six individual data sets and the
uncertainties are standard deviations. Times are values from ﬁts of the experimental data

to the function:
f(t) : Al exp(_t/TSIE)— A2 CXp(—l/Tl)
where rs”; is the stimulated emission decay time constant and T1 is the vibrational

population relaxation time constant. The TI times in are given in ps.

110

 

 

 

 

60 - .-
L .-
a 50 u ..
3*
ﬁg 40 I l
E I
"(E 30 P
o l-
? 20 - ¢
M
[—7 10 -
0 I l L
0° co co“ 00°
69““ e9“ 9““ 00°?“
"5’8 25"“ ”V ’5’

Figure 33 — Vibrational population relaxation times of the 1375 cm'1 mode for
perylene (O) and 1-methylperylene (I) as a function of solvent aliphatic chain
length in the ketones.

111

80-

 

50

I
OI

 

T

 

 

ii
30 - I
I
I

7

Il (1733 cm" mode) (ps)

10-

 

 

Figure 34 - Vibrational population relaxation times of the 1733 cm'1 mode for
perylene (O) and l-methylperylene (I) as a function of solvent aliphatic chain
length in the ketones.

112

the interaction between the solute and solvent is of sufﬁcient strength to effect a
structural distortion that lifts the perylene center of inversion and that the solute is
conﬁned in a quasi-lamellar environment in solvents of sufﬁcient length. It is important
to note here that the extent of the distortion of the perylene molecule is not related in a
linear fashion to the apparently large change in measured T1 times relative to the

41
unperturbed case.

Any break in the symmetry of the chromophore will give rise to
stronger solvent-solute coupling. The proportionality between any such structural
distortion and a change in T1 has not been established. If the details of the putative
structural distortion were known and the potential energy gradient for the solute along
this coordinate were determined, then an estimate of the change in induced dipole
moment as a ﬁrnction of distortion along the appropriate coordinate could be made.
Given the limited information available, such an estimate is not possible. These ﬁndings
do suggest, however, that the perylene chromophore is not as able to sequester itself away
from the solvent polar group in the ketones to the extent that it can in the alcohols.

We now consider the T1 results for perylene and l-methylperylene in the
aldehydes. For all cases we observe that the T] relaxation is fast, indicating high
efﬁciency vibrational energy transfer from the solute to the solvent. We observe efﬁcient
solvent-solute coupling for both chromophores and, consistent with the data for the
ketones, the strong coupling is necessarily due to the close proximity of the donor and
acceptor moieties. These ﬁndings are consistent with the reorientation data. Because of
the similarity of the perylene and l—methylperylene reorientation data in the aldehydes

and n-alcohols, substantial solvent-solvent interactions involving the terminal

functionalities dominate the solvent organization that the chromophores sense. Our data

113

indicate that the aldehyde proton can H-bond efﬁciently with other solvent molecules,
presumably at the C=O functionality.

The dynamical data for perylene and l-methylperylene in the aldehydes are
reminiscent of the behavior of these chromophores in the n-alcohols, and there is a
similarity between the l-methylperylene behavior in the alkanes and the ketones. For all
cases, we see quasi-lamellar conﬁnement of the probe molecules in longer solvents and
the T. times for these systems are uniformly short. These data argue for strong solvent-
solute interactions, and a central issue is the nature of these interactions. If dipole-
induced dipole interactions were dominant, then we would expect to see Similar behavior
for these probe molecules in the aldehydes and ketones because of their Similar dipole
moments. We do not observe this experimentally. If H-bonding interactions were most
signiﬁcant, then the expected correspondence would be between the aldehydes and
alcohols, and this is what we observe experimentally. The correspondence between the
ketone and alkane data suggests that dipolar interactions are relatively unimportant for
these probe molecules.

Conclusions

We have studied the rotational diffusion and vibrational population relaxation
dynamics for perylene and l-methylperylene in a series of aldehydes from propanal to
decanal, and a series of ketones that have an aliphatic substituent similar in length to that
of the major axis of the chromophore. We ﬁnd evidence for solvent local organization in
each solvent. In the aldehyde solvents, we observe changes in the rotational diffusion
behavior that are consistent with hydrogen-bonding solvent-solvent interactions by virtue

of the similarity of these data to the behavior of the same chromophores in n-alcohols.

114

For the ketones, we observe chromophore behavior that is reminiscent of that seen for
these same molecules in n-alkanes. The central conclusions of this work are that the
interactions between solvent and solute are sufﬁciently strong to perturb the symmetry of
the chromophores, based on our T1 data; the solvent-solvent interactions are sufﬁciently
strong to provide a quasi-lamellar, conﬁning environment for these chromophores, as
manifested in the reorientation data. The balance of solvent-solvent and solvent-solute
forces depends on the speciﬁc system studied, and for the chromophores perylene and l-
methylperylene, H-bonding solvent-solvent interactions are more important in

determining solvation properties than dipole-induced dipole interactions.

115

8.

9.

10.

11.

12.

13.

I4.

15.

16.

17.

18.

19.

20.

Literature Cited

. This work was submitted for review and publication to J. Phys. Chem. A.

Sanders, M. J.; Wirth, M. J.; Chem. Phys. Lett., 1983, 101, 361.
Gudgin-Templeton, E. F.; Quitevis, E. L.; Kenney-Wallace, G. A.; J. Phys. Chem,
1985, 89, 3238.
Von Jena, A.; Lessing, H. E.; Chem. Phys, 1979, 40, 245.
Von Jena, A.; Lessing, H. E.; Ber. Bunsen-Ges Phys. Chem, 1979, 83, 181.
Von Jena, A.; Lessing, H. E.; Chem. Phys. Lett., 1981, 78, 187.
Eisenthal, K. 8.; Acc. Chem. Res, 1975, 8, 118.
Fleming, G. R.; Morris, J. M.; Robinson, G. W.; Chem. Phys, 1976, 1 7, 91.
Shank, C. V.; Ippen, E. P.; Appl. Phys. Lett., 1975, 26, 62.
Millar, D. P.; Shah, R.; Zewail, A. H.; Chem. Phys. Lett., 1979, 66, 435.
Gudgin-Templeton, E. F .; Kenney-Wallace, G. A.; J. Phys. Chem, 1986, 90, 2896.
Blanchard, G. J.; Wirth, M. J.; J. Phys. Chem, 1986, 90. 2521.
Blanchard, G. J.; J. Chem. Phys, 1987, 87, 6802.
Blanchard, G. J.; Cihal, C. A.; J. Phys. Chem, 1988, 92, 5950.
Blanchard, G. J.; J. Phys. Chem, 1988, 92, 6303.
Blanchard, G. J.; J. Phys. Chem, 1989, 93, 4315.
Blanchard, G. J.; Anal. Chem, 1989, 61, 2394.
Alavi, D. S.; Hartman, R. S.; Waldeck, D. H.; J. Phys. Chem, 1991, 95, 6770.
Hartman, R. S.; Alavi, D. S.; Waldeck, D. H.; J. Phys. Chem, 1991, 95, 7872.

Jiang, Y.; Blanchard, G. J.; J. Phys. Chem, 1994, 98, 6436.

116

21

22

23

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

Brocklehurst, B.; Young, R. N.; J. Phys. Chem,l995, 99, 40.

Pauls, S. W.; Hedstrom, J. F .; Johnson, C. K.; Chem. Phys, 1998, 23 7, 205.
Elsaesser, T.; Kaiser, W.; Annu. Rev. Phys. Chem, 1991, 42, 83.

Lingle, R., Jr.; Xu, X.; Yu, S. C.; Zhu, H.; Hopkins, J. B.; J. Chem. Phys, 1990, 93,
5667.

Anﬁnrud, P. A.; Han, C.; Lian, T.; Hochstrasser, R. M.; J. Phys. Chem, 1990, 94,
1180.

Heilweil, E. J .; Casassa, M. P.; Cavanagh, R. R.; Stephenson, J. C.; Ann. Rev. Phys.
Chem, 1989, 40, 143.

Heilweil, E. J .; Cavanagh, R. R.; Stephenson, J. C.; Chem. Phys. Lett., 1987, 134,
181.

Heilweil, E. J .; Cavanagh, R. R.; Stephenson, J. C.; J. Chem. Phys, 1989, 89, 230.
Heilweil, E. J .; Casassa, M. P.; Cavanagh, R. R.; Stephenson, J. C.; J. Chem. Phys,
1986, 85, 5004.

Hambir, S. A.; Jiang, Y.; Blanchard, G. J .; J. Chem. Phys, 1993, 98, 6075.

Jiang, Y.; Blanchard, G. J .; J. Phys. Chem, 1994, 98, 941 1.

Jiang, Y.; Blanchard, G. J .; J. Phys. Chem, 1994, 98, 9417.

Jiang, Y.; Blanchard, G. J .; J. Phys. Chem, 1995, 99, 7904.

McCarthy, P. K.; Blanchard, G. J .; J. Phys. Chem, 1995, 99, 17748.

McCarthy, P. K.; Blanchard, G. J .; J. Phys. Chem, 1996, 100, 5182.

McCarthy, P. K.; Blanchard, G. J .; J. Phys. Chem, 1996, 100, 14592.

Jiang, Y.; Hambir, S. A.; Blanchard, G. J .; Opt. Commun., 1993, 99, 216.

117

38

39

4O

41

42

43

44

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

Bado, P.; Wilson, S. B.; Wilson, K. R. Rev. Sci. Instrum. 1982, 53, 706.

Andor, L.; Lorincz, A.; Siemion, J .; Smith, D. D.; Rice, S. A.; Rev. Sci. Instrum.
1984, 55, 64

Blanchard, G. J.; Wirth, M. J.; Anal. Chem, 1986, 56, 532.

Peake, D. A.; Oyler, A. R.; Heikkila, K. E.; Liukkonen, R. J .; Engroff, E. C.;
Carlson, R. M. Synth. Commun. 1983, 13, 21.

Jas, G. S.; Larson, E. J .; Johnson, C. K.; Kuczera, K.; J. Phys. Chem. A, 2000, 104,
9841.

Karpovich, D. S.; Blanchard, G. J. J. Phys. Chem. 1995, 99, 3951.

Lewitzka, L; Lohmannsroben, H.-G Strauch, M.; Luttke, W. Photochem Photobiol.
A: Chem. 1991, 61,191.

Grimme, S.; Lohmannsroben, H.-G J. Phys. Chem. 1992, 96, 7005.

Debye, P. Polar Molecules; Chemical Catalog Co.: New York, 1929; p. 84.
Perrin, F. J. Phys. Radium 1936, 7, l.

Chuang, T. J.; Eisenthal, K. B. J. Chem. Phys. 1972, 57, 5094.

Hu, C. M.; Zwanzig, R. J. Chem. Phys. 1974, 60, 4354.

Youngren, G. K.; Acrivos, A. J. Chem. Phys. 1975, 63, 3846.

Zwanzig, R.; Harrison, A. K. J. Chem. Phys. 1985, 83, 5861.

Edward, J. T.; J. Chem. Ed., 1970, 47, 261.

Brocklehurst, 8.; Young, R. N. J. Chem. Phys. 1995, 99, 40.

Piston, D. W.; Bilash, T.; Gratton, E.; J. Phys. Chem, 1989, 93, 3963.

Blanchard, G. J.; Wirth, M. J.; J. Chem. Phys, 1985, 82, 39.

118

56. Pauls, S. W.; Hedstrom, J. F; Johnson, C. K. Chem. Phys. 1998, 23 7, 205.
57. The permanent dipole moments for S0 and SI l-methylperylene and for ground state
solvent molecules was calculated using Hyperchem 6.0®, performing a semi-

empirical calculation with PM-3 parameterization.

58. Matsunuma, S.; Akamatsu, N,; Kamisuki, T.; Adachi, Y.; J. Chem. Phys, 1988, 88,

2956.

59. Goldie, S. N.; Blanchard, G. J. J. Phys. Chem. A. 1999, 103, 999.
60. Blanchard, G. J.; Wirth, M. J. J. Phys. Chem. 1986, 90, 2521.
61. Blanchard, G. J .; Rev. Sci. Instrum., 1996, 67, 4085.

62. Blanchard, G. J.; Anal. Chem, 1997, 69, 351A.

119

Chapter 4

IN-SITU QUANTIFICATION OF ACID GENERATION IN A DEEP UV
PHOTORESIST THROUGH THE USE OF PROBE MOLECULESI

Introduction

One of the goals in the semiconductor industry is to increase the feature density of
computer semiconductor chips. As feature density increases, so does the processing
capability of these devices. To accomplish this goal, the microlithographic methods used
today have been moving toward the use of shorter wavelengths, which translates to
smaller features on a semiconductor chip. The circuit patterns of a microprocessor are
transferred to silicon dioxide layers deposited on a silicon wafer using light and a
quartz/chromium photomask. The intensity of commonly used short wavelength sources
is characteristically low. In the manufacturing process, a single photon will at best result
in a single photoevent. The limited number of short wavelength photons results in a low
manufacturing efficiency.

There are two main problems that gave rise the development of chemically
ampliﬁed resists. The first is the lack of photon ﬂux in the deep UV (DUV) region of the
spectrum of commonly used sources. The second problem is the relative insensitivity, or
low quantum efﬁciency, of traditional resists and their photoactive compounds in the
deep UV part of the spectral region.

To circumvent these problems, photoresists have been formulated with additives
that will chemically amplify the number of photon absorbtion events in a polymeric
matrix. One of these means is through the use of photoinitiators. These additives serve

as catalysts in the ampliﬁcation of the photons. A photoinitiator is activated by a photon.

120

which then dissociates to generate a photoacid molecule. In the polymer, the degradation
process involves cleavage of pendant groups from the polymeric backbone through the
use of this photoacid. (Scheme 3)

Scheme 3 - Deprotection reaction representative of photoresist exposure. The gases

isobutylene and carbon dioxide are liberated, and the proton regenerated as a result
of exposure to UV radiation.

 

 

— R — R R R
—C———C— ~-C--------"-C-'-
+
<9 L <3 <3 wt <g>+c02<g>+u+
OH O OH OH
e _n [=0
0
Me Me
Me

As a result of this process, carbon dioxide and isobutylene gases are generated
and diffuse from the polymer. The photoacid is then regenerated and through a reaction-
diffusion process causes many cleavage events to occur as the result of a single photon
being absorbed. The cleavage of the pendant groups causes the polymer to be soluble in
an alkaline solution, allowing it to be dissolved and removed, leaving a void where
further chemical processing (i.e. etching, deposition) can occur. In a chemically
ampliﬁed resist, the sensitivity to deep UV photons is increased. Instead of a single
photoevent that results from a single photon, a single photon can result in several
deprotection reactions.

Though this process enables the use of DUV photons for smaller feature sizes, it
can cause another problem through the uncontrolled diffusion of the photoacid outside of
the desired regions of the circuit paths. For this reason, it is important to understand the

diffusional characteristics of the photoacid. The two processing steps that have the

121

greatest impact on the concentration and diffusional mobility of the photoacid are
exposure and post exposure bake (PEB). The acid is produced by the photolysis of an
onium salt photoacid generator (PAG). For a given PAG concentration, the exposure
dose will dictate the amount of acid that is generated. The temperature and duration of
the PEB both inﬂuence the diffusional mobility of the acid and the rate of the
deprotection reaction. The PEB is the time in which the actual acid diffusion and
deprotection reaction take place, and therefore the PEB conditions have the greatest
impact on the resulting linewidth. Higher PEB temperatures increase the sensitivity of
the resist but also increase the diffusivity of the acid, increasing susceptibility to process
variations. Both reaction and diffusion are very sensitive to temperature and time, and
even slight non-uniformity on hot plates or a delay in reaching a cool plate (and therefore
a delay in quenching of the reaction) can cause signiﬁcant changes in critical dimensions
(CDs). These nonlinear interactions between exposure light intensity, chemical reaction
and non-constant diffusion make optimizing and running a stable process a challenge.
This creates non-uniform line shapes, which can result in incorrect or imperfect patterns
on a chip. To understand and account for these processes, it is necessary to characterize
the amount of acid generated and to determine the location of the photoacid with high
resolution. We present a method here that will accomplish these tasks with relative ease

and little equipment overhead.

122

Experimental

Steady State Spectroscopy -— in-situ samples

APEX E-2408 DUV photoresist, (Shipley Co.), stored in the dark at 4°C, was
doped with 0.1 wt% crystal violet (Sigma-Aldrich, used as received). This solution was
spin coated at 2000 rpm onto 1” x 2” quartz slides using a PWM32 spinner (Headway
Research Inc.) A ﬁlm layer of ~1 um thickness resulted. The slide was then heated at
90°C for 60 seconds (post application bake, PAB) to evaporate the solvent (propylene
glycol mono-methyl ether acetate, PGMEA). The samples were then divided into
exposed (subject to UV exposure and further processing) and control samples. Exposed
samples were subjected to the output of a 200 W Hg(Xe) lamp (Oriel) for 30 seconds and
then a post exposure bake (PEB) for 90 seconds at 90°C. Absorption measurements for
all samples were obtained using a HP 8452-A diode array spectrometer, with a special
sample holder manufactured in-house for ﬁlm studies.

Steady State Spectroscopy — liquid phase samples

The steady state absorption spectra of crystal violet as a function of pH in aqueous
solution were recorded with 1 nm resolution using a Hitachi U-4001 spectrometer. A
constant crystal violet concentration of 6 x 10'6 M was maintained for all absorption
measurements by dilution from a single stock solution. The pH of the solutions was set
using a series of 0.2 M potassium chloride (Spectrum, used as received) and 0.2 M
hydrochloric acid (Columbus Chemical Industries, used as received) buffer systems.
Outside the buffer range of this acid/base pair, aqueous solutions of 0.2 M potassium
chloride or 0.2 M hydrochloric acid were used to maintain the pH. A Fisher Scientiﬁc

Accumet pH meter and combination electrode with automatic temperature correction was

123

used for all pH determinations. The ﬁnal pH of the solutions ranged from 0.26 to 5.12;
all pH measurements were performed after the absorption measurements. The ionic
strength of our solutions was not explicitly controlled.
Results and Discussion
Crystal violet is a widely used and studiedz‘8 acid/base indicator. The structure of
crystal violet (Scheme 4) is characterized by three p-dimethylaniline rings, bonded to a

central carbon atom. As the level of protonation increases on the amine

Scheme 4 - The structure of crystal violet, with protonation levels and equilibria.
The native form of crystal violet is shown on the left. See text for further discussion.

H+ H+

\ / \ \ /'

N N’

E? .. E? * ?
MOO/v— (Oily 0 ﬁt

|
groups,9 the steady state response of crystal violet changes. The native form of crystal

 

 

 

violet (CV+) is violet in color, with its So -) Sl absorption maximum centered about 590
nm. Protonation of one of the amino groups to form HCV2+ shifts the absorption
maximum to 630 nm, which produces a solution that is green in color. Additional
protonation to form HZCV3+ produces a yellow solution, with the absorption maximum
centered at 430 nm.

We were able to characterize the various forms of crystal violet in solution
through the use of buffered pH systems10 from pH~5 to pH~0, corresponding to proton
concentration of 10'5 M to 1 M. The changes in absorption spectra are presented in

Figure 35. We have found these absorption bands correspond to similar behavior

124

(L60

0.55
0.50
0.45
A. 0.40
3
3 0.35
q, 0.30
8 0.25
m
.3 0.20
8 0.15
..D
m 0.10

(105
(100
4105

[H] = 7.59x10'° M
."\
.’ '\
.’ 'i
.’ i [H‘]=1.55x10"M
‘./\

I I V I V I V I

/

I'IWI'I'I'T'I'T

   

 

 

350 400 450 500 550 600 650 700

wavelength (nm)

Figure 35 - The linear response of the three-protonation states of crystal violet. The
acid concentration is presented near the maximum of each absorption spectrum.

125

of crystal violet in-situ. The native form of CV+, present at low proton concentrations,
absorbs strongly at 590 nm. As the pH decreases, and the proton concentration increases,
the onset of a second absorption band centered at 630 nm is observed with a simultaneous
decrease in absorbance at 590 nm. Continued increase in proton concentration causes a
continued decrease in the 590 and the 630 nm features, while the absorption feature at
430 nm increases.

This protonation process of crystal violet is an equilibrium process. The

equilibria and the equilibrium constants expressions are presented in equations (16) to

 

(18).
HZCV3+ <—'i-->HCV3'+H+ <——K“—>CV‘+H+ (16)
Where the equilibrium constants are expressed in the following manner:
HCV2+ [H+
.=[ 13. 1 (17)
[H.CV 1
V H‘
24C 112‘] (18)
[HCV ]

From these expressions and these data, we determined the constants for
protonation/deprotonation equilibria, by applying Beer’s law to each of the solutions and

solving a system of simultaneous equations, presented in equations (19) to (21),

Ab b b 09)
A 590 :8 ::)I(())I,E'1‘ bC 2281131. + (5195)th begglyiﬁN +8;'9E)LLOW bc 5Y()[:)LL()\V (20)
Ah“) :82/318LET bC:~:8LET + 2:)EEN bCSI‘JBEEN + Z;[;.)LLOVV bczzi‘glIJW' (2 1 )

where A; is the total absorbance at wavelength 1 (nm), 5::wa is the molar absorptivityl ‘

of the indicated colored species at it for each level of protonation, b is the path length (1.0

126

("OLOR
it

cm) and c is the concentration of each species of each protonation condition. As a

check, the sum of these concentrations was compared to the total concentration of the
crystal violet stock solution used in these experiments. The calculated total concentration
was found to be within experimental uncertainty of the total concentration of the stock
solution. We calculated the equilibrium constant values K1 and K; as presented in
equations (17) and (18). K1 was determined to be (2.5 3c 1.5) x 10'2 M and K2 was
calculated to be (6.7 i 2.5) x 10'3 M. These values correspond to pK. = 1.67 i 0.29 and
pKz = 2.20 i 0.19, which is in reasonable agreement with the ﬁrst pKa value of a
structurally related triphenylmethane dye, malachite green, reported to be ~1.3 in the
literature.2

With these data, it is possible to determine the relative concentrations of each
species in any solution of crystal violet as a function of solution pH based on absorbance
measurements, through the use of (1 values. a-Values are unitless ratios of the
concentration of individual species to the total analytical concentration of all the related
species.12 The sum of all the fractions in a system must equal unity. The a values are
readily expressed in terms of [H+] and the equilibrium constants K and K2, previously

determined here, by rearranging the equilibrium constant expressions:

 

 

a():[H2CV3+]: 2 [If] 2 (22)
CT [H] +K,[H*] +KIK,
HCV“ H* K1
1 L. I J .23)

 

' CT [Ir]2 +Kl [tr]2 +K,K,

127

:[CV‘L

KlK,
2 CT [tr]2 +Kl [Pr]2 +K,K,

 

(24)

where [H2CV3+] is the concentration of the diprotonated yellow colored species, [HCV2+]
is the concentration of the monoprotonated green colored species, and [CV] is the
concentration of the native form of crystal violet, which is purple in color. The total
concentration, CT, is the sum of these three concentrations of crystal violet in various
forms of protonation. It is then possible to relate the measurable quantities of pH and the
absorption characteristics to the concentration of each species and its level of protonation.
We present our calibration curve for crystal violet in Figure 36. Using this set of curves,
we can then apply measured absorbance quantities to proton concentration as the result of
exposure, and determine the fraction of each species as a ﬁJnction of pH.

We have examined the changes in the steady state absorption response of crystal
violet doped into the photoresist as a function of processing and present these data in
Figure 37. Initially, the absorption spectrum is characterized with a single dominant
feature at 590 nm, attributed to the native form of crystal violet [CV+]. Immediately after
exposure of the entire slide to DUV radiation and post exposure bake, the feature at 590
nm has diminished and a feature at 430 nm dominates the spectrum. Assuming that the
data from our solution studies can be used to interpret the polymer data, this corresponds
to the diprotonated form [HZCV3+] and is indicative of a high proton concentration.

Subsequent absorption spectra collected showed a decrease in the 430 nm feature
and an increase in the 630 nm feature followed by a recovery at 590 nm. We ascribe this
trend to the evaporation or neutralization of the photoacid in the ﬁlm. One possible

. . . - 13.14
mechamsm for proton loss is long-term exposure of the ﬁlm to airborne amines.

128

This trend corresponds to a multistep process in the photoresist ﬁlm. The ﬁrst
step is protonation of the crystal violet due to photacid production from the photoinitiator
in the photoresist formulation. This process occurs within a second of exposure to the
source. As the protons react with airborne bases, the crystal violet probe molecule level
of protonation will decrease, with the corresponding changes in the absorption spectra.
These changes will be similar to the changes observed in the liquid state. Kinetically,
this process is slower, as it occurs over hours to days (vide infra). We present an
example of these changes as a function of time in Figure 38. A major source of amines
that contribute to airborne molecular contamination is the human factor, as many amines
are generated and are released as a result of metabolism in the human body. Bases that
have been found in semiconductor manufacturing facilities include airborne ammonia, n-
methyl pyrrolidinone (NMP), 1,4-diaminobutane, and 1,5-diaminopentane. Each of these
airborne bases can be absorbed into the polymer matrix, causing defects in the device.

In this set of experiments, the entire slide was exposed to DUV radiation. This is
different from the methodology of semiconductor manufacturing, in which the
illumination of the regions of the photoresist is carefully controlled through the use of
quartz/chromium photomasks. These features that are transferred to the silicon dioxide
layers have critical dimensions on the order of microns in size. It is possible through the
use of proper near ﬁeld optics (NFO’s) to build an absorption spectrometer that would
sample a microscopic area, on the same length scale as these critical dimensions, and
interrogate the absorption behavior of a probe molecule doped into the region. From
these changes, as we have demonstrated here, it is possible to determine the changes in

the pH of the local environment, and calculate the amount of photoacid generated in a

129

speciﬁc region. With sufﬁcient spatial resolution, the extent of a photoacid gradient
could be determined, and the acid mobility as a function of time and processing
conditions could also be examined. This information is important to the semiconductor
fabrication community.
Conclusions

With appropriate near ﬁeld optics (NFO’s), it may be possible to determine the
absorbance of crystal violet at a speciﬁc point and then relate that observation to acid
concentration using the data presented here. This Will help to accomplish our goal of
quantifying the amount of the photoacid generated by the photoinitiator, and then provide
information about the location and diffusion behavior of the photoacid. This information
can be used to model more accurately the photoacid behavior in photoresists used in

microlithographic applications.

130

 

 

 

 

or fraction

 

”I

 

O'OM.l.l‘l‘i -
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0

pH

Figure 36 — a value calibration curve for crystal violet as a function of pH. The a
values provide concentration data by relating absorbance.

131

 

 

0.14 -
. After spin casting
0.12 r
0.10 - 2 h post-PEB
A . 5 d post-PEB
3 -
«3 0'08 After post
v ' ex osure bake PEB '
8 0.06 - p ( ) x1."-
‘: , d.v"‘-\. '. ' . ‘.
N r’ .' "'N
«D 1- ooooo . .... ..
,_ 0.04 ,"~\ ...." '. L' “E,
8 ' IT‘! """ . h1i‘- I . -, ’fr\’\
'3 002 FI'O/Nd\ -/ \-‘.?.\\ p’Vs Ira, "I
I"' ’ \ ' " " 1
/
. l
_0 A 1 n l n l A l n l 1 I A l

 

.02
350 400 450 500 550 600 650 700

wavelength (nm)

Figure 37 — Changes in the steady state absorption response of crystal violet doped
into a polymer film as a function of processing. See text for a discussion.

132

b)
O
I'I'l

 

 

 

 

 

 

 

Figure 38 — Changes in pH in-situ as a function of processing time.

133

9.

Literature Cited

This work, and the work of Chapter 5, will be submitted for publication to J. Phys.
Chem. B.

“Indicators for Non-Aqueous Acid-Base Titrations”; Bishop, E. In Indicators;
Bishop, E., Ed.; Pergamon: New York, 1972.

Duxbury, D. F. Chem. Rev. 1993, 93, 381.
Ben-Amotz, D.; Harris, C. B. Chem. Phys. Lett. 1985, 119, 305.
Ben—Amotz, D.; Harris, C. B. J. Chem. Phys. 1987, 86, 4856.

Ishikawa, M.; Ye, J. Y.; Maruyama, Y; Nakatsuka, H.; J. Phys. Chem. A. 1999, 103,
4319.

Maruyama, Y.; Ishikawa, M.; Satozono, H.; J. Am. Chem. Soc. 1996, 118, 6257.

Maruyama, Y.; Magnin, O. Satozono, H.; Ishiakawa, M. J. Phys. Chem. A. 1999, 103,
5629.

Tuthill, S. M.; Kolling, O. W.; Roberts, K. H. Anal. Chem. 1960, 32, 1678.

10. Lange 's Handbook of Chemistry, 14th ed.; Dean, J. A., Ed.; McGraw-Hill: New York,

11.

12.

13.

14.

1992; 8.106.
Conant, J. B.; Werner, T. H.; J. Am. Chem. Soc., 1930, 52, 4436.

Skoog, D. A.; West, D. M.; Holler, F. J. Fundamentals of Analytical Chemistry, 5th
ed. Saunders: Chicago, 1988, 227.

Thompson, L. F.; “Resist Processing” In Introduction to Microlithography, 2nd ed.;
Thompson, L. F.; Willson, C. G.; Bowden, M. J. Eds; ACS Professional Reference
Book: Washington, DC, 1994, 269.

Mack, C. A.; “Lithographic Effects of Acid Diffusion in Chemically Ampliﬁed
Resists” In Microelectronics Technology: Polymers for Advanced Imaging and
Packaging; Reischmanis, E.; Ober, C. K.; MacDonald, S. A.; Iwayanagi, T.;
Nishikubo, T. Eds; American Chemical Society: Washington, DC, 1995, 56.

134

Chapter 5

IN-SIT U CHARACTERIZATION OF FREE VOLUME GENERATION DURING
POST EXPOSURE BAKE (PEB) IN A COMMERICAL PHOTORESISTl

Introduction

As described in the previous chapter, photoresists are used to facilitate the transfer
of circuit patterns of a microprocessor to silicon dioxide layers deposited on a silicon
wafer using light and a quartz/chromium photomask. These photoresists have been
formulated with additives that will amplify the number of photon absorption events in a
polymeric matrix using chemical means thereby improving sensitivity and efﬁciency, so
that the shorter DUV wavelengths can be used to produce smaller feature sizes. One of
these means is through the use of photoinitiators. A photoinitiator is activated by a DUV
photon, which then dissociates to generate a photoacid molecule. In the polymer, the
degradation process involves cleavage of pendant groups from the polymeric backbone
through the use of this photoacid. These photoacids serve as catalysts in the
ampliﬁcation of the photons, since a single photon can generate several photoacid
molecules that are regenerated in the course of the deprotection reaction.

As a result of this reaction, carbon dioxide and isobutylene gases are generated
and diffuse from the polymer. This process creates free volume in the polymer matrix,
through which the regenerated photoacid can diffuse and cause many additional cleavage
events to occur. This chain of events results from the absorption of a single photon.
Though this process enables the use of shorter wavelengths for device manufacture, it can
cause problems through the uncontrolled diffusion of the photoacid outside of the

intended circuit pattern. For this reason, it is important to understand the diffusional

135

characteristics of the photoacid. The two processing steps that have the greatest impact
on the concentration and diffusional mobility of the photoacid are photo-exposure and
post exposure bake (PEB). We have described a novel method for determining the
concentration of the photoacid generated in the previous chapter. In this chapter, we are
interested in characterizing the diffusional characteristics of the photoacid in the polymer
matrix.

The temperature and duration of the PEB both influence the diffusional mobility
of the acid and the rate of the deprotection reaction. The PEB is the time in which the
actual acid diffusion and deprotection reactions take place, and therefore the PEB
conditions have the greatest impact on the resulting device feature dimensions. Higher
PEB temperatures increase the sensitivity of the resist and also increase the diffusivity of
the acid, giving rise to results that depend sensitively on process variations. Both
reaction and photoacid diffusion are very sensitive to temperature and time, and even
slight non-uniformity on hot plates or a delay in reaching a cool plate (and therefore a
delay in quenching of the reaction) can cause signiﬁcant changes in device structure
critical dimensions (CDs). These nonlinear interactions between exposure light intensity,
chemical reaction and a non-constant diffusion make optimizing and running a stable
process a challenge. This situation gives rise to non-uniform line shapes, which can
result in incorrect or imperfect circuit patterns on a chip. To understand and account for
these processes, it is necessary to characterize the free volume generated as a function of
processing. We present a new method here that will accomplish these tasks in situ and

allow for better mathematical modeling of these processes.

136

Experimental

The picosecond pump-probe laser spectrometer used for the ground state recovery
measurements has been described in detail previously,2 and we present only a brief
outline of its performance characteristics here. A mode-locked CW Nd:YAG laser
(Coherent Antares 76-S) produces 30 W of average power (1064 nm, 100 ps pulses, 76
MHz repetition rate). The output of this laser is frequency-doubled to produce ~2.1 W of
average power at 532 nm. The second harmonic light is used to excite two cavity-
dumped dye lasers synchronously (Coherent 702-3). Both lasers operate with Rhodamine
6G laser dye (Kodak). The output of each laser is ~100 mW average power at 8 MHz
repetition rate with a pulse that produces a 7 ps FWHM autocorrelation trace using a
three plate bireﬁ'ingent ﬁlter. The time resolution of this system, ~10 ps, is determined
by the cross-correlation between the pump and probe laser pulse trains. The pump laser
wavelength was set to 589.5 nm, coincident with the absorption maximum of the ﬁrst
singlet transition of crystal violet. The probe wavelength was chosen to lie within the
absorption band of crystal violet. The probe beam was polarized at 54.7° with respect to
the pump beam to eliminate molecular reorientation contributions to the signal.
Detection of the transient signals was accomplished using a radio and audio frequency
triple-modulation scheme, with synchronous demodulation detection.3’5 Each reported
time constant is the average of at least ﬁve individual determinations that are themselves
the average of seven to ten time-scans.

Samples for these experiments were prepared by spin coating APEX E-2408
DUV photoresist, (Shipley Co.) doped with 0.1 wt% crystal violet (Sigma-Aldrich) at

2000 rpm onto 1” x 2” quartz slides using a PWM 32 spinner (Headway Research Inc.).

137

A ﬁlm layer of ~1 um resulted. The slide was then heated at 90°C for 60 seconds (post
application bake, PAB) to evaporate the solvent (propylene glycol, mono-methyl ether
acetate, PGMEA). The samples were then divided into two sets. One set of samples
experienced no further processing subsequent to spin-coating and post-application bake,
while a second set of samples was exposed to 200 W Hg(Xe) lamp (Oriel) for 30 seconds
and then baked for 90 seconds at 90°C. The samples were mounted in a manually driven
translation stage and the two laser beams were co-focused on them. To minimize the
effects of photobleaching and thermal degradation, the slide was translated as the sample
was scanned, and neutral density ﬁlters were used to attenuate the incident light.

Liquid phase samples of ~1 x 10“t M crystal violet in ethylene glycol, 1,3-
propanediol, propylene glycol, 1,4-butanediol, 1,3-butanediol, and glycerol were made
using chemicals of the highest purity available (Aldrich, Spectrum). All chemicals were
used as received. The ground state recovery of each liquid phase sample was measured
using a flow cell system to minimize thermal lensing considerations. The sample
solutions were ﬂowed through a 1 mm path length quartz ﬂow cell, and the sample
temperature was controlled at 300 :1: 0.1 K using a thermostated bath.

Solution viscosities were determined using a Brookﬁeld Model DV-I+ viscometer
with spindle #18 for the diol solutions and #34 for glycerol. The viscosities of ethylene

glycol and propylene glycol were acquired from the literature.6
Results and Discussion
In addition to being a well-characterized acid-base indicator, crystal violet is a
Widely used probe of microviscosity in solution. Crystal violet (Scheme 4, Chapter 4 —

this work), a member of the triphenylmethane dye family, is predominantly a D3

138

 

 

“propeller” shaped molecule in solution with three identical N,N-dimethylaniline rings
bound to a central carbon atom tilted slightly out of the molecular plane. Ring rotation
around the bonds between the phenyl rings to the central carbon atom in what is believed
to be a barrierless process along the excited state potential energy surface. Studies have
indicated a strong dependence of the excited state lifetime on solvent viscosity, which is
indicative of essentially no intra-molecular rotational barrier to ring rotation.7’8 The
characteristically low ﬂuorescence quantum yield of this family of molecules prevents us
from using a ﬂuorescence lifetime measurement to explore these changes. Instead, we
examine the recovery of the ground state population by detecting the transient change in
absorption of the probe pulse resulting from excitation by the pump laser pulse, to sense
changes in the local environment in the photoresist thin ﬁlms. We are able to
characterize two ﬁlm properties using this widely available chromophore because of its
combined sensitivity to local acid concentration (Chapter 4, this work) and viscosity.

We observe an exponential decay of the ground-state recovery signal for the
un-exposed, spin-coated, doped resist ﬁlm, and we show these data in Figure 39. A ﬁt of
the decay to a sum of two exponentials results in a 25 i 4 ps short time component (11)
and a longer 120 i 18 ps time (Q) component. The double exponential decay behavior
observed is similar to ﬂuorescence decay experiments using malachite green, another
triphenylmethane dye at similar temperatures in highly viscous systems, such as a solid
SiOZ-ZrOZ matrix9 and in polymethyl methacrylate (PMMA) and polyvinyl alcohol

(PVA) ﬁlmslo‘l "‘2’” However, the physical meaning of the observation of two
exponential decay components for a triphenylmethane dye is the subject of an ongoing

debate in the literature. None-the-less, there is a well-established relationship between

139

 

1.0

0.8

0.6

0.4

0.2

signal intensity (a.u.)

0.0

 

 

 

 

 

 

0 100 200 300 400
delay time (ps)

Figure 39 — Instrument response function and ground state recovery measurement
of crystal violet in unprocessed spin cast film. The solid line through the data is the
best-ﬁt double exponential decay function.

140

the values of both decay time constants and the bulk viscosity of the medium, with the
constants increasing with increasing viscosity. The values of the two time constants we
observed are similar to the time constants obtained by Ben-Amotz and Harris,7 Canva et
al.9 and Ippen, Shank and Bergman.l4 Ippen et al. '4 have developed a theory that relates
the recovery time to the viscosity of the matrix.

Using these theories, we have determined the ground state recovery (GSR)
behavior of crystal violet in a series of viscous matrices. We present representative data
for the two extremes of our calibration curve in Figure 40 and Figure 41. We present the
short-time component of the exponential decay for the series of viscous solutions in
Figure 42. The short time component shows a linear relationship with solvent viscosity.
Several of the solutions we used as matrices could only be ﬁt to a single exponential
decay. We believe that this is due to ﬁtting parameters used by our computer program
(Microcal Origin) and not indicative of a change in behavior of our probe molecule. By
using these data as a calibration curve, we are able to extrapolate that the local viscosity
experienced by crystal violet in the unexposed ﬁlm is approximately 150 CR

We have been able to apply these methods and data to the exposed photoresist
ﬁlm. We present the difference in the ground state recovery behavior of crystal violet in
the exposed and unexposed ﬁlm in Figure 43. The average double exponential decay
best ﬁt to a 30 2t 8 ps short time component (11) and a long 190 :t 75 ps (r2) component.
These values are beyond the limits of our calibration curve, but indicate that the viscosity

of the exposed matrix increases to approximately 180-200 cP. This is also indicated by
the comparison of the ground state recovery behavior of crystal violet in the two polymer

films, presented in Figure 43. The change in the behavior of the crystal violet is related

141

 

t==5pm

signal intensity, (a.u.)

 

 

l n l A l 4 J

75 100 125 150
delay time, (ps)

 

Figure 40 — Ground state recovery dynamics of crystal violet in ethylene glycol
solvent. The instrument response function is also provided for reference.

142

 

1.0-

   
 

=é 0.8-
Cd

5 P
7;. 0.6-
C

8

$2

'_: 0.4-
N

c: .
00

'5; 0.2- 12:115ps

 

 

 

 

0.0

 

L l n L n I 1 l 1 l J

100 200 300 400 500

delay time, ps

'41

Figure 41 — Ground state recovery dynamics of crystal violet in glycol solvent. The
instrument response function is also provided for reference.

143

 

25-

   
  
    
 
    
 

iGlycerol
20 -

1,4 Butanediol

l 1,3 Butanediol
I

Propylene Glycol

1,3 Propanediol

! Ethylene Glycol

 

n I n l l l
O 20 4O

60 80

 

A l n l A l L l
100 120 140 160
viscosity (cP)

Figure 42 —- Viscous matrix calibration curve relating viscosity of media to ground
state recovery of crystal violet. The line presented is the linear best fit to the data.

144

0.8 "

0.6 -
Processed film

0.4 f

 

 

signal intensity, (a.u.)

0.2 ' ‘ Unprocessed ﬁlm

 

00 V Kh‘J h

 

I n I n l 4 I n I

0 100 200 300 400
delay time (ps)

Figure 43 — Instrument response function and ground state recovery behavior of
crystal violet in processed (post application bake, UV exposure, post exposure bake)
and unprocessed (post application bake only) films.

145

 

to the changing viscosity of the photoresist matrix. The change in the viscosity is
attributed to the relaxation of the polymer chains into free volume created by the
diffusion of carbon dioxide and the isobutylene gases from the photoresist as the cleavage
of the pendant chains along the polymer backbone continues during the post-exposure
bake. The amount of free volume generated can be better determined through these data.

There are a number of probe molecules that exhibit an increase in the
ﬂuorescence signal and excited-state lifetime as the local viscosity is increased. As the
molecular rotations become hindered, non-radiative pathways for energy transfer are less
efﬁcient; therefore, more of the excited-state energy must be dispelled through radiative
(ﬂuorescent) pathways. Crystal violet and malachite green are two such “propeller”
probe molecules that have been used to investigate free volume in solid matrices. Other
free volume sensitive probes are based upon photoisomerization and intramolecular
excimer formation. With a small amount of a viscosity-sensitive molecule, the micro-
viscosity or free volume of the system can be measured. For example, Francis Wang et
al. have used 1,3-bis—(1-pyrene) (BPP) to monitor the polymerization of methyl
methacrylate in situ.15 Moorjani and coworkers ﬁmher demonstrated the utility of
ﬂuorescence monitoring when dealing with diffusion-controlled reactions in media of
different viscosity.16 Victor and Torkelson have also used molecular probes to measure
the free volume available in polymer structures.l7

Crystal violet (Scheme 4), a member of the triphenylmethane dye family, is
predominantly a D3 propeller shape molecule in solution with three identical N,N-
dimethylaniline rings bound to a central carbon atom tilted slightly out of the molecular

plane. In addition to being a widely used acid-base indicator, crystal violet is a useful

146

probe of microviscosity or free volume. Ring rotation around the bonds between the
phenyl rings to the central carbon atom in what is believed to be a barrierless process
along the excited state potential energy surface. Studies have indicated a strong
dependence of the excited state lifetime on solvent viscosity, which is indicative of
essentially no intra-molecular rotational barrier to ring rotation.7’8 The characteristically
low ﬂuorescence quantum yield of this family of molecules prevents us from using a
ﬂuorescence lifetime measurement to explore these changes. We instead examine the
recovery of the ground state population by detecting the transient change in absorption of
the probe pulse resulting from excitation by the pump laser pulse, to sense changes in the
local environment in the photoresist thin ﬁlms. We are able to characterize two ﬁlm
properties using this widely available chromophore because of its combined sensitivity to
local acid concentration and viscosity.
Conclusions

We have developed a pump-probe spectroscopic method to determine changes in
free volume in a photoresist matrix through the examination of changes in ground state
recovery times of crystal violet. We have quantiﬁed the change in the local viscosity in-
situ as a function of processing step. We attribute the changes in local viscosity as the
result of diffusion of gases from the polymer matrix that result when pendant chains are
removed through exposure to UV light in microlithographic applications. This
experimentally determined information can then be applied to mathematical models
currently used to describe this process in the microlithographic ﬁeld, to better understand

the manufacturing process.

147

10.

11.

12.

13.

14.

15.

16.

17.

Literature Cited

. This work, and the work of Chapter 4, will be submitted for publication to J. Phys.

Chem. B.
Jiang, Y.; Hambir, S. A.; Blanchard, G. J ., Opt. Commun. 1993, 99, 216.
Bado, P.; Wilson, S. B.; Wilson, K. R., Rev. Sci. Instrum. 1982, 53, 606.

Andor, L.; Lorincz, A.; Siemion, J .; Smith, D. D.; Rice, S. A., Rev. Sci. Instrum.
1984, 55, 64.

. Blanchard, G. J.; Wirth, M. J.; Anal. Chem. 1986, 56, 532.

Lange ’s Handbook of Chemistry, 14th ed.; Dean, J. A., Ed; McGraw-Hill: New York,
1992; 8.106.

Ben-Amotz, D.; Harris, C. B.; J. Chem. Phys. 1987, 86, 4856.
Ben-Amotz, D.; Harris, C. B.; J. Chem. Phys. 1987, 86, 5433.

Canva, M.; Le Saux, G.; Georges, P.; Brun, A.; Chaput, F.; Boilot, J. P.; Chem Phys.
Lett. 1991, 176, 495.

Abedin, K. M.; Ye, J. Y.; Inouye, H.; Hattori, T.; Sumi, H.; Nakatsuka, H.; J. Chem.
Phys. 1995, 103, 6414.

Abedin, K. M.; Ye, J. Y.; Inouye, H.; Hattori, T.; Nakatsuka, H.; J. of Luminescence.
1995, 64, 135.

Ye, J .Y.; Hattori, T.; Inouye, H.; Ueta, H.; Nakatsuka, H.; Maruyama, Y.; Ishikawa,
M.; Phys. Rev. B. 1996, 52, 8349.

Ye, J .Y.; Hattori, T.; Nakatsuka, H.; Maruyama, Y.; Ishikawa, M.; Phys. Rev. B.
1997, 56, 5286.

Ippen, E. P.; Shank, C. V.; Bergman, A.; Chem. Phys. Lett. 1976, 38, 611.
Wang, F .W.; Lowry, R.E.; Grant, W.H.; Polymer, 1984, 25, 690.

Moorjani, S.K.; Rangaraj an, B.; Scranton, A.B.; “Effect of viscosity on the rate of
photosensitization of diaryliodonium salts by anthracene,” In Photopolymerization:
Fundamentals and Applications; Scranton, A.B.; Bowman, C.N.; Peiffer, R.W. Eds.
American Chemical Society: Washington, DC, 95.

Victor, J .G.; Torkelson, J .M.; Macromolecules, 1987, 20, 2241.

148

Chapter 6

CONCLUSIONS AND FUTURE WORK

In this dissertation, we have studied the rotational and vibrational population
relaxation dynamics of perylene and l-methylperylene in n-alkanols, aliphatic aldehydes
and ketones. We have concluded that the presence of a labile proton is more important
than solvent dipole moments to the solvation processes of an essentially nonpolar species.
To better understand the effect of the presence of this labile proton on the rotational and
vibrational population relaxation dynamics of the two probe molecules in solution, a
more complete set of ketone solvents with a larger range of hydrodynamic volumes in
solution could add to the understanding from the perspective of the changing solvent
dipole moment. To continue to examine the effects of the labile hydrogen atom, these
same rotational and vibrational population relaxation dynamic measurements of perylene
and l-methylperylene could be made in a series of carboxylic acid solvents, where the
hydrodynamic volume is varied and the positioning of the carbonyl functional group
relative to the probe molecule. The carboxylic acid solvent would be the next logical step
in increasing complexity of the functional group, and would give additional insight into
the mediation of solvent structure around the solute.

In Chapter 3 of this dissertation, the deviation from the R(O) value was partially
attributed to the possible contributions from excited state absorption by perylene and 1-
methylperylene in aldehydes and ketones, but due to the lack of data, further inferences

could not be made. Excited state absorption measurements for the probe molecules in the

149

aldehydes and ketones would add to this body of work, and begin to address the possible
reasons why the value of R(O) does not reach the expected value of 0.40.

The work performed in chapters 4 and 5 of this dissertation, regarding the
characterization of photoresist materials used in microlithographic applications has the
greatest potential for continued study. The characterization of acid mobility in situ of
actual process samples would require the construction of a near field optical system that
would be able to interrogate the regions immediately surrounding the lines transferred to
the silicon wafer. The amount of acid and its location and the diffusion behavior of the
acid molecule has been shown to be valuable information to the microlithographic
community, and if a simple enough instrument could be constructed, it is suspected that
the instrument and the information that could be obtained would be well received.

It may be possible to construct a system that will use the uptake of airborne bases
as a monitor for exposure. The kinetics of airborne base uptake is dependent upon the
speciﬁc matrix that the indicator is doped into. A systematic study of the matrix used and
the effect on the kinetics of the uptake mechanism would be beneﬁcial to better
understand how the matrix affects the sensitivity to the airborne bases. Airborne
contaminants, usually in the form of bases have been shown to have detrimental effects
on the line shape of the semiconductor circuit patterns.

The methodology developed in Chapter 5 could be used to determine the
intermediate changes in free-volume as a function of processing step. The effect of DUV
dosage amount could be explored, as could the effect of changing the post exposure bake
time and temperature. These results would better improve the understanding of the

creation of free volume in the polymer during the manufacturing process.

150

Both of the techniques in Chapters 4 and 5 could be applied to other photoresists
used in semiconductor fabrication. The ﬁeld of microlithography is rapidly changing.
New materials that allow for smaller and smaller features have replaced the materials
used in this dissertation. These techniques and methods developed here could be used to
characterize the new materials, with the information obtained of similar value to the

community.

151

 

APPENDIX

152

APPENDIX A

NATIONAL INSTRUMENTS LABVIEW® BASED DATA ACQUISITION
SOFTWARE FOR PUMP-PROBE LASER SPECTROMETER

This section presents the National Instruments LabView® virtual instrument (VI)
written in our laboratory to control the pump-probe spectrophotometer (Figure 44) used
in many of the experiments presented in this dissertation. This VI was written for
operation in Microsoft Windows NT v4.0 operating system using, National Instruments
LabView® v5.0.1, and National Instruments NI-DAQ software v6.0 software packages.
The hardware used was a Pentium based personal computer with 64MB of memory, a
National Instruments SC-2040 simultaneous sample and hold card, a National
Instruments AT-MIO-16E-10 data acquisition card, and a National Instruments AT-
GPIB-NT card for instrument control.

The V1 is designed to obtain two analog voltage signals from the output of the
lock-in ampliﬁers used in the pump-probe experiment, and to process the data to account
for any gain applied by the electronics. The data are displayed for each individual
channel on the front panel (user interface). The V1 is also used to control the motion of
the translation stage located on the laser table, either low-resolution or high-resolution.
The translation stage controls the differential delay time between the pump and probe
laser pulses in this experiment. The control signals are sent to the stepping motor
controller using the GPIB interface. The V1 has ﬁle management functionality built in,
with the data acquisition parameters and ﬁle information included in the output ﬁle

automatically.

153

 

The V1 is based on a hardware timed, non-buffered technique of data acquisition,
where the scan clock on the data acquisition board controls the acquisition timing. The
front panel, which serves as the user interface, is presented in Figure 45. It is designed to
accept data acquisition parameters, including ﬁle and sample information, the total delay
time, the number of scans, the resolution of the stage expressed in ps/step, and the
number of time constants to wait at each step before the data run begins, and write these
parameters to the data ﬁle automatically. The front panel will also display the data from
each of the four input channels individually in the smaller graphical displays, while
displaying the signal averaged data from channel 0 (signal) and channel 1 (reference).

The overall circuit diagram for the V1 is presented in Figure 46. The dominant
feature of this ﬁrst ﬁgure is a primary for-next loop that will cause its contents to occur
for the number of cycles indicated by the number of scans control on the front panel. The
V1 will ﬁrst obtain the number of scans, the total delay, the stepper motor resolution, and
the system time constants to wait from the front panel (Figure 45). The V1 will then
create an array ﬁlled with zeros, the length of which is based on the total number of delay
points from the front panel. The delay between each step of the translation stage is also
obtained and set here, based on user input from the front panel. This delay between each
step is designed to ensure that enough time elapses before data acquisition to allow for
the mechanics of the system to settle properly. Before any new data are acquired, and
before the VI progresses beyond the initial actions of the primary for-next loop, the
graphical displays are cleared by sending a zero value to the history data attribute of each
display. The creation of a data ﬁle, with the sample information and run date and time,

then occurs, with all parameters separated with a carriage return and line feed to facilitate

154

 

data importation to other programs using ASCII format. At the conclusion of the for-next
loop, the output data array is appended to this ﬁle and transposed from row major
(horizontal) to column major (vertical) format.

Inside this primary for-next loop is a nested secondary sequence loop, consisting
of four total frames. The ﬁrst frame of this sequence loop, ﬁame 0 (Figure 46), contains
a secondary for-next loop and a three-framed secondary sequence loop nested in the
secondary for-next loop. The secondary sequence loop is designed to acquire the
individual data points that make up the total data set (Figure 46), move the stage forward
one step between each data point, (Figure 47 and Figure 48), and facilitates a time delay
for the mechanical settling after the stage has moved (Figure 49). This secondary
sequence loop will occur as long as the secondary for-next loop is not satisﬁed, which is
based on the total delay time of each individual data run set on the front panel. If this
condition is satisﬁed, this secondary for-next loop will terminate, and the primary
sequence loop will advance to frame 1 (Figure 47). Frame 0 in the secondary sequence
loop (Figure 46) is designed to acquire the data point from each of four channels
simultaneously, through the use of the SC-2040 simultaneous sample and hold card and
the AT-MIO-16E-10 data acquisition card. The raw data are processed in this frame,
where the gain applied by each of the lock-in ampliﬁers is accounted for in the
conversion of the raw data to voltages. Each of these data points is collected and
displayed on individual graphical displays, and the quotient of the sample and reference
channel is averaged over the dataset. This average is also displayed on a separate

graphical display.

155

 

In frame one of the secondary sequence loop, the movement of the translation
stage such that the delay between the pump pulse and the probe pulse is changed by one
picosecond using the GPIB interface. There is a switch on the front panel that is set to
match the resolution of the stage in use, and determine the string sent via the GPIB
interface to the stepper motor controller. The low state, or false state (Figure 47), is for
the low-resolution stage where a picosecond is 15 steps in length. The high state, or true
state (Figure 48), is for the high-resolution stepper motor controller, where a picosecond
is 150 steps in length.

After the stage is moved forward one step in the ﬁrst frame of the secondary
sequence loop, the VI is designed to wait a number of time constants as a delay to allow
for mechanical and electronic settling. The number of time constants is based on the
input from the user on the front panel and the actual time constant settings for the
electronics of the system. These settings are set on the front panel such that they are
equal to the physical settings of the reference audio frequency lock-in ampliﬁer. The
clock used to time this delay is CPU process dependent, and may need to be changed if
several processes are running simultaneously. The alternative settings for the time
constant indicators are presented in Figure 50, and are controlled by the indicator on the
front panel.

This completes the description of the secondary sequence loop functions. The
secondary sequence loop will occur until the total number of steps is equal to the total
delay time set on the front panel. This is determined through the use of the secondary
sequence for-next loop in the VI. When the secondary for-next loop completes its cycle,

the primary sequence loop will advance to frame 1 and rewind the stage to its original

156

 

position. The same switch on the front panel that determines the string sent to the stepper
motor controller via GPIB for forward motion is also used to determine the string sent for
the rewind ﬁmction. The false or low state is to rewind the low-resolution stage (Figure
51), while the true or high state is to rewind the hi gh-resolution stage (Figure 52.)

In frame two of the primary sequence loop (Figure 53), a delay to allow for
mechanical settling after stage rewind is built in to the VI. It is necessary, since the
program continues to progress after the command signals are sent via GPIB to the stepper
motor controller, and the mechanical stage requires some time to rewind. It was found
that 17500 cycles was sufﬁcient for 1000 ps. Additional time can be added through a
front panel control, and may be necessary for delay times longer than 1000 ps.

In the ﬁnal frame of the primary sequence loop (Figure 54), the history data
attribute of the sample, reference, and average signal graphical displays are cleared if the
number of scans requested does not equal the current scan number (false state).
Otherwise (true state, Figure 55), the loop will hold the values on the display until the VI
is re-run.

The primary sequence loop will continue as long as the primary for-next loop is
active. This is based on the condition that the scan number does not equal the number of
scans requested on the front panel. When the two are equal, the primary for-next loop
will terminate and the data array will be appended to the ﬁle, after it is transposed from
row major (horizontal) to column major (vertical) format.

A complete list of sub-virtual instruments used is presented in Table 15 for
reference. The disk locations are the default install locations on a typical installation of

the LabView® 5.0.1 program.

157

 

  
  
     
 
  
     

Frequency tripler

P_u|se
diagnostics

Mode-locked Nd:YAG laser Chopper 6 I

   

Delay time
control

 

Lock-in ampliﬁers

Figure 44 ~ Pump-probe laser spectrometer schematic.

158

 

mm<>>._h_0m ZOEm500< <._.<Q mmomn. QEDQ

9mm 0mm Om; 8m 9mm

8m 23 gm

 
 

1Iee m N Flee

  
                         

omm Dom omk can 0mm new 0mm Dom omv Dow 09 com own son DE. 09. am a
_ _ _ V _ _ _ A _ _ _ _ _ L _ _IMWO_W_H1

. _

1mmoom.

 

  

1 _H_ r
.256 oocwfom

.loolg

$55 0.9% m

 

11egm

.33 Sum .0 .5

    

/

  

5 23m. OP
E<mAOOF Z_
>>O¢G< 121—OE

umD

_.,_.1$1.1m.N,.\...U.K xx...
0451.. .11 04111....“

seem
>ozmommzmx

 

 

/

._> :2 2 .362 c. 26:0 E9. co v.96
a.u.uEqu :2 new 0353 Soon cozcéoi .39—
.”.dj co nozcoo 28:00.0 9 288:8 2 «3:8 .366,

 

 

 

 

 

 

1o. moo o 20.695 \.

1mmoo m ‘ 2.3 1. mm 82w 3m ..n3 2 «Ecacou as:

1mmoo v ﬂli. 23w .0 .3632

-m. mg m surﬁng“? Q“ as 3.533

1 m . mac m 111E“ 7.228 as. F Eoumw

. m moo _ Dcaﬁuﬂom

1 m. mom _ D£oc£o>m3 39m

1n mow _ U50c£o>~3 95¢

. m, we _ _ii11_§eaz

-39.: U383
. o q_ Q ass 1.
g .35; 3292 >26 coacoaxm cam. <3 ..E Eco <3 .3 ._
I .o ,! 558:2 cmow 9 .111M/Em1 r . 1 J.,
B _o D 73 2.0a 2:; E026 coacmmxm “616$: gram 1021.795 Q _

 

Figure 45 — Front panel of LabView® pump-probe data acquisition virtual

tru ment.

ms

159

 

 

=
‘iéHltlcuw Dale
:1
-

:1

     
   

 

1::1
'Hdory [1.3113

1.
E-‘EEJ @ mm
’8 Mammum

 

Figure 46 — Complete circuit diagram for LabView W pump-probe data acquisition virtual instrument.

160

 

 

 

m I-t m_‘

iHmow [lava
—
-EH1'|1‘11},1 [1.31.3
—

 

Figure 47 — Complete circuit diagram of pump—probe data acquisition virtual instrument, showing second frame 01' secondary sequence loop.

“—EHIWJW flat-a

161

  

 

 

 

 

 

 

 

 

"".""'r"'lIt'll'lrt'
. . 4

 

 

h‘snhot
‘\.“V

 

. ..
1‘:

‘I

 

vrre.rrrrr.r"¢.rr.r.rrr. rrr.r.vrrrrrrr
AAAa4:111‘00a'lllll'dl‘n‘o'nn‘d'o‘Iod‘AAA1 .o

Figure 48 — igh-resolution stage forward function.

III-I.I.II1I1I II.IIIIII.II IIIII .IIIII.I.I.I.I.III1I..I I.I.II.I.1I1.IIII.I11II.I I1.I1_I1.I1II..I_..I_I1IIIII1

 

   

 

 

 

 

. . . 1 . . 1 1_ . . . 1 _ 1. . 1 . . .1 ., 1H 1 . 1
I......I I I I I I I I I I I.I: IIIIII .1. I.I I I I... .I1..I.I.1I1I .I..I._I.I. I. I.I:.IrI....I1.1,I.I1I I I I.I11I1.I.1o

 

 

..I.I.I._I I.I,I.I..I_I I.I.I._I I I I I I I I I I.I.I I ."E“. .I I I.I I.I.I_I.I.I..I....I..I_.1I...I .I II I_.._.I...I.I.I.I.“.I1._.I._

@143

 

 

 

 

'-

 

 

 

 

- I I I I I I I I.I I I I I I I I I I I I I.I I I I I I I I I Inga“ I I IHI.I.I..I_..IH..I.IH.I.I I I.I...I I..I...I._I._I I I. I.I.I.I”_.I.I...I.I...1.

Figure 49

— Final frame of secondary sequence loop nested in first frame of primary

sequence loop.

163

 

Figure 50 — Alternatives for electronic time constant settings in secondary sequence
loop panel presented in Figure 49.

164

...“...mﬂlu... lam-w“... Ian-w...
.ﬂ

51:13I........Iéa.a.u.?a1=1I1u1a1a1u1¥n1I.I..g1a1I1I1!n§In.EH!xiii...-1I1I1IwIm-IEI.-,EI1I1I1I_1I11I1I1III_.

411-, Jim I. "11..I.EI%IEI.HMIﬁlwmIﬁIgrEIEImea.

IaIQIEIHIEIﬁuﬂlmlimit.[sum-BIKIEIHIHIHIEI a? 5....HIEIﬂcﬁaﬂmﬂmawﬂdﬂmimﬂllmin.

s I‘I‘
I]:

51311-
.1.II!I\‘

I'I'II L’
I.I.ii

...

I
1.
1.
1.
I
I
.
.Q
h
.I
L
.1.
L
I
h
I
1.
i.
k
1‘
1.
.I
1.
.I
I
I
t
I
1.
.I
L
I
L
I
I
a
a.
I
I
I
1
I

 

LI’I'I‘II‘ II'II‘IIII
on...

I I I‘I
| I-‘.‘ ‘5‘ 1333.1 I Q.‘

III-Ir

 

Second frame of primary sequence loop, with low resolution move stage
165

forward function.

Figure 51 —

 

 

 

 

11 .
’rpy'p,yvrI.rvv.0.vr.r.1.v.v,v.r.rvrr.r'.v.r,o.rr.r-.4r.v.vrv'r-yrurr-' 7".717"r_!1r.l""'.vlrrrrf 'llirvrrvlvrrrr,lrrr.rvr1'_r.
4
r.
4
v
4
r
4
r
4
1
r.
4
' Q
4 . 1
v
4
I.
4
r.
4
l
4
,
4
r
___..---___1 ‘
4

 

D]

 

1;..s'snssgrr

 

1 009,--

 

lt¢‘t\\'|l

I.IIII‘QI‘VIV‘I‘IIIIIII‘VV‘VHI
s1|sn 1r

 

.‘I‘V‘I

3,4114";It’dhlao‘lock'o'l‘uaa131-42244344)}413441‘144‘3}242'43‘4-33‘4‘4‘4‘4444'4'41441434'4II23143I23}334444)}4
1

4
~4
4
4
4
4
I
'4

'4

 

44:444IIJ'4

Figure 52 - High-resolution stage rewind function.

  

‘

bhtbh‘iII‘Ili‘bt‘lsbll‘lb‘s‘l‘

tlbhthbIthl§‘850\ﬁb§l\‘llll‘IIOOIII

 

  

. .

 

. I. _
.1IHI~....I.. I.._I._.I. I I.I I.I...I .I I I.I” I....I...I .I I.. I....I.......I .I._.IM I. I UI.....I....I "EHII.1I1..I....I1I.....1.I.x.H.mI.....I_.w.mI..I.I.I.I.IEILIﬂI“I.--13.3.9.3...I...:I .I....I...I”.1IM_WI.....I«_..

Figure 53 - Third frame of primary sequence loop.

167

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII-IIIIIIIIIIIIIIIIIIII.

 

 

........................1.......n—gan

 

 

 

 

IIlIIIIIIIIIIIIIIIIIIIIIIIIIIIII.

Figure 54 — Fourth (final) frame of primary sequence loop.

168

 

Figure 55 — Alternative to true/false state function in Figure 54. This frame is
intentionally left empty, so that no processing occurs.

169

Table

15 - Table of sub-virtual instruments used in the pump-probe data

acquisition virtual instrument.

 

 

 

 

 

M
CLERR

 

 

 

 

M
CONHG

 

 

 

M
S-SCRN

 

 

 

 

at
81991

 

 

 

 

 

 

 

 

 

 

 

 

 

General Error Handler.vi
C :\Program Files\National Instruments\LabVIEW\vi.lib\Utility\error.llb\General
Error Handler.vi

AI Clear.vi
C :\Program Files\National Instruments\LabVIEW\vi.lib\DAQ\AI.LLB\AI
Clear.vi

AI Config.vi
C :\Program Files\Nationa1 Instruments\LabVIEW\vi.lib\DAQ\AI.LLB\AI
Conﬁgvi

A] Single Scan.vi
C :\Program Files\Nationa1 Instruments\LabVIEW\vi.lib\DAQ\AI.LLB\AI Single
Scan.vi

AI Start.vi
C :\Program F iles\National Instruments\LabVIEW\vi.lib\DAQ\AI.LLB\AI Start.vi

Simple Error Handler.vi
C:\Program F iles\National Instruments\LabVIEW\vi.lib\Utility\error.llb\Simple
Error Handler.vi

Write To Spreadsheet File.vi
C:\Program F iles\National Instruments\LabVIEW\vi.lib\Utility\ﬁle.llb\Write To

Spreadsheet File.vi

Write Characters To File.vi
C :\Program Files\National Instruments\LabVIEW\vi.lib\Utility\ﬁle.llb\Write
Characters To File.vi

DataProcessing

C :\Program Files\National
Instruments\LabVIEW\user.lib\PumpProbeVI's.llb\DataProcessing

170

IIIIIIIIIIIIIII

1111111111111