University of Southern California Dissertations and Theses

Ultrafast electronic deactivation of DNA bases in aqueous solution

(USC Thesis Other)

Ultrafast electronic deactivation of DNA bases in aqueous solution

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

ULTRAFAST ELECTRONIC DEACTIVATION OF DNA BASES
IN AQUEOUS SOLUTION

by

Askat Jailaubekov

A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
in Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(CHEMISTRY)

May 2007

Copyright 2007 Askat Jailaubekov

ii
Dedication

Dedicated to my parents, Nurgulsum Jailaubekova and Yerkin Jailaubekov.

Thank you for your constant love and support.

iii
Acknowledgements

I would like to thank my advisor, Professor Stephen Bradforth, for giving me
the opportunity to work in his research group. Every research project I was involved
in was challenging and exciting. Steve’s scientific vision and optimism about any
problem was always encouraging. I am grateful to him for his solid guidance and
valuable suggestions.
This experimental work is a group effort, and I would like to thank the people
directly involved in this work. I thank Delmar Larsen and Dorte Madsen for building
up the dispersed apparatus and making it running. I also would like to thank Dr.
Peter Qin for providing us with the DNA samples and using his lab equipment.
Christina To, who joined us at the beginning of the DNA experiments, helped me a
lot with the sample preparation. I am thankful to her for that.
I would like to address my special thanks to all the graduate students and
postdocs from the Bradforth’s research group. I thank Amy Moskun, Jerainne
Johnson and Cyril Chen for helping me to learn the experimental techniques. I am
grateful to Amy for giving me the opportunity to participate in her research project.
This was truly a rewarding experience. I thank Cyril, my office mate, for discussion
on various topics including science, economics and politics. I enjoyed our political
discussions on situations in China and Kazakhstan or around the world. I thank the
Piotr Pieniazek, Diana Suffern, Chris Rivera, Christi Chester, and Tom Zhang. All of

iv
them have given me a valuable assistance, especially during the last year of my
research.
Finally, I would like to thank all my friends: Misha and Lyuda Slipchenko,
Kirill Kuyanov, George Kumi, Dmitry Skvortsov, Sergey Malyk, Eugene
Polikarpov, and Sergey Zakharov. Thank you all for your friendship.

v
Table of Contents
Dedication ii

Acknowledgements iii

List of Tables vii

List of Figures viii

Abstract xix

Chapter 1. Introduction 1
1.1 Ultrafast Excited-State Dynamics in Aqueous Solution:
Electronic Deactivation of DNA Bases
1
1.2 References for Chapter 1 16

Chapter 2. Tunable 30-femtosecond pulses across the deep
ultraviolet
19
2.1 Introduction 19
2.2 Experimental and Results 21
2.3 Summary 29
2.4 Acknowledgements 30
2.5 Supplement: Basics of wave propagation and phase-
matching in hollow waveguides
30
2.6 References for Chapter 2 37

Chapter 3. Experimental: The UV Pump-Dispersed Probe
Apparatus
39
3.1 Introduction 39
3.2 Construction of the Pump-Dispersed Probe Apparatus 42
3.3 Characterization the UV Pump-Dispersed Probe Apparatus:
Temporal Resolution and Dispersion
47
3.4 Anisotropy Experiments 52
3.5 Applications in Ultrafast Spectroscopy and Future Work 54
3.6 References for Chapter 3 57

Chapter 4. Electronic Relaxation Dynamics of Uracil Derivatives 59
4.1 Excited State Dynamics of Thymidine 59
4.2 References for Chapter 4 70

vi
4.3 Postscript to Chapter 4: Experimental Methods and
Anisotropy Dynamics
72
4.3.A Introduction 72
4.3.B Experimental 75
4.3.C Dispersed Data (isotropic) 80
4.3.D Dispersed Anisotropy 85
4.3.E Ground State Bleach 87
4.3.F Global Analysis: Sequential Kinetic Model 89
4.3.G Ionization potentials 93
4.3.H Branching Model and Anisotropy Simulations 93
4.3.I References for Postscript to Chapter 4 102

Chapter 5. Ultrafast Transient Absorption Studies on Adenine
Derivatives
103
5.1 Introduction 103
5.2 Experimental 110
5.3 Results 111
5.4 Analysis and Discussion 124
5.4.A Global analysis of transient absorption data 124
5.4.B Comparison of TRPES with dispersed transient
absorption data
130
5.4.C Electronic relaxation pathway 134
5.5 Summary 136
5.6 References for Chapter 5 139

Chapter 6. Future Work 141
6.1 Excitation Energy Dependence 142
6.2 References for Chapter 6 148

Bibliography 149

vii
List of Tables

Table 4.1 Characteristic parameters of the ground state absorption,
fluorescence and the time-resolved transient spectra of uracil and
thymine derivatives. Peak wavelengths are reported if bands
shows clear peak. Steady-state fluorescence data is taken from
Onidas et al and Gustavsson et al.
5,6
(a) Values in parentheses
indicate the shift of the peak in comparison to the corresponding
nucleoside.
84

Table 5.1 Characteristic times (fs) of the transient absorption decay and rise
times at different probe wavelengths resulting from the
monoexponential fits with an instrument response function of 45
fs.
a)
Limited by the time-resolution of the system.
b)
Extracted
from decomposition of transient absorption and anisotropy signals
of ATP recorded with pump and probe pulses at 267 nm.
21

115

viii
List of Figures

Figure 1.1 The chemical structures and standard ring numbering for the five
nucleic acid bases.
2

Figure 1.2 A schematic diagram of excited-state photophysics in an isolated
nucleobase. A UV pump pulse prepares the molecule in its
optically bright, S
2
( ππ*) singlet excited state. The nucleic acid
bases have allowed a π → π* transition, which is responsible for
their first strong absorption band at ~270 nm. After the
excitation, the molecule returns to the ground state by internal
conversion (denoted by wavy solid line in the diagram), which
competes with fluorescence. A very rapid internal conversion (<
1ps) leads to direct reduction of the electronic excited state
lifetime. All nucleic acid bases are known to have such
radiationless deactivation, which results in their very low
fluorescence and phosphorescence quantum yields. Very close in
energy lie at least one other singlet state, the optically dark S
1

(nπ*) state. A widely discussed mechanism for nucleobase
electronic relaxation involves two sequential steps: ππ*→ nπ*
followed by nπ* → S
0
internal conversion to the ground state.
5

Figure 1.3 A schematic representation of excited-state dynamics and various
deactivation pathways for the nucleobase from the Franck-
Condon (FC) point after photoexcitation. The conical intersection
is the point where two potential energy surfaces cross. At the
point of intersection the two electronic states are degenerate, and
it is at this point the population of the excited state can decay
from the upper to the lower intersecting state. The more complete
description of the conical intersection can be distinguished in a
reduced space called branching space (two-dimensional space of
two special normal modes) where the equations for two surfaces
appear as the geometric equations for cone.
49
The left and right
endpoints are the S
1
/S
0
and the S
2
/S
0
conical intersections,
respectively. The crossing, which is close to FC region, is the
S
2
/S
1
conical intersection.
10

ix
Figure 1.4 Schematic description of pump-probe experiments in gas and
solution phases. In the TRPES experiment, Koopmans’
correlations allow for a mapping of different excited states onto
different ionic states. The known character of the excited states
permits assignment of the various photoelectron bands observed.
We directly compare TRPES in the gas-phase with our dispersed
transient absorption data in aqueous solution. Detailed
examination of the ESA bands of our analysis and their
photoelectron spectra shows that they have quite similar Franck-
Condon envelopes and similar spacing between bands. The
excellent agreement between the ESA and TRPES bands leads us
to speculate that optical absorption in the liquid from S
2
( ππ*)
and S
1
(nπ*) occurs to upper states with the same origin as in gas-
phase photoionization, namely Rydberg upper states with cation
cores D
0
( π
-1
) and D
1
(n
-1
). (State reorder) In aqueous solution, the
transition energy to the n π* state is expected to shift to higher
energy, while the ππ* state should shift only slightly to lower
energies since in protic solvents ππ* states are more stabilized by
dipole-dipole and hydrogen bonding interactions than n π* states.
(Right) Excited state absorption bands extracted from global
analysis of the transient absorption data.
14

Figure 2.1 Experimental setup. L1, L2 are plano-convex (f =10 cm) and
plano-concave (f = -5cm) lenses, respectively. BBO is a
nonlinear crystal (0.5 mm, type I) and L3, L4 are focusing lenses
(f = 25 cm) mounted on XYZ translation stages. An iris (ID) is
used to adjust coupling efficiency of pump beam. λ/2 is a half
wave plate. W1, W2 are 2-mm thick CaF
2
windows. L5 is a
collimating lens (f = 50 cm) and DM a dichroic mirror. PS is a
90-degree periscope system.
22

Figure 2.2 (Top) DUV output tuning curve (thick line; bottom and left hand
axes) where pulses are obtained by mixing either OPA idler ( ●)
or signal ( ▲) pulses as seed for the fiber. ( □) Corresponding
OPA pulse energies (thin line; right and top axes) measured at
the entrance to fiber. (Bottom) Generated pulse spectra with
fiber seeded by tunable IR pulses. Spectral broadening is ~5 nm
(FWHM) corresponding to transform-limited pulse durations <
20 fs.
25

x

Figure 2.3 (a) Output spectra of 266-nm light generated by FWM in a
hollow core fiber (thick line) and by SFM in an 80- μm-thick
BBO crystal (thin line, FWHM = 1.4 nm). Fiber parameters: 30
cm length, 75 μm diameter. Pulsewidths in parentheses are
transform-limited pulse durations corresponding to indicated
spectral bandwidths (assumed Gaussian). (b) Spectrum of UV
pulse centered at 233 nm generated by FWM. Experimental
autocorrelation (points) of (c) 266-nm and (d) 233-nm pulses
after prism compression; (solid curve) fit assuming Gaussian
pulse shape.
27

Figure 2.4 Calculated transmissions of the EH
11
fundamental mode and that
of the subsequent, linearly polarized, high-order modes EH
12

(u
12
=5.5201) and EH
13
(u
13
=8.6537) for a capillary waveguide
with a diameter 150 μm, at 800 nm, as a function of the capillary
length (in cm).
32

Figure 2.5 Calculated transmissions of the EH
11
fundamental mode for a
capillary waveguides with the diameters 75, 100 and 150 μm, at
800 nm, as a function of the capillary length (in cm).
33

Figure 2.6 Calculated EH
11
mode propagation in the fiber with the diameter
of 75 μm at three different wavelengths 267, 400 and 800 nm, as
a function of the capillary length (in cm).
34

Figure 3.1 Pump-probe experimental configuration. Arrow shows a beam
emerging from Ti:Sapphire regenerative amplifier (Hurricane,
Spectra-Physics, 1kHz). The 800 nm beam with horizontal
polarization is split into two beams using a beamsplitter (BS).
ND: neutral density wheels for pump beam (ND
1
) and probe
beam (ND
2
); L
1
, L
2
and L
3
are the 15 cm, 20 cm, and 12.5 cm
focusing lenses, respectively. (WP) a zero-order half-wave plate
designed for 266 nm. White light continuum is generated in a 2-
mm thick CaF
2
disk. The disk is mounted on motorized stage.
White light continuum is collimated with parabolic mirror P
1
and
focused into sample with parabolic mirror P
2
. Both mirrors are
aluminum coated 90 ° off-axis parabolic mirrors with effective
focusing length of 5 cm (P
1
) and 18 cm (P
2
). To avoid a strong
scattering of the intense 800 nm part of the continuum beam in
the spectrograph (S) a dichroic mirror (D) is used to reject 98%
of 800 nm light (bandwidth = 75 nm at 45° P-polarization).
43

xi

Figure 3.2 White-light continuum spectra in various materials. Four kinds of
material were tested in our lab: sapphire, BaF
2
, CaF
2
and LiF.
The spectra were taken with a visible grating. CaF
2
material is
chosen as it demonstrated several suitable characteristics for
probe light such as wide spectral range (290-700 nm), small
degradation among all fluoride materials, good shot-to-shot pulse
stability (0.3 –1% r.m.s.) across whole range, and excellent phase
characteristics. Inset graph shows continuum intensities on a log
scale at the blue edge.
45

Figure 3.3 Time-resolved transient absorption spectrum of neat water after
excitation at 267 nm with 35 fs pump pulses. Two-dimensional
data was fitted with a forth order polynomial fitting routine. (a)
Color contour plot of the raw spectrum. (b) Dispersion corrected
spectrum.
48

Figure 3.4 Experimental transient signals ΔOD( ω
probe,
t) (dots) obtained
from neat water, with pump pulse centered at 267 nm and two
probe wavelengths, 350 nm and 500 nm. Solid lines: best fit of
the signal with equation (2). The dash line is the cross-correlation
function of 40 fs FWHM width.
51

Figure 3.5 (a) Experimental magic angle pump-probe spectrum (dots) and a
magic angle signal calculated directly from the experimental
parallel and perpendicular signals (solid line) using the
relationship S
MA
( λ, 0) =(1/3)( S
PA
( λ, 0)+2 S
PE
( λ, 0)) at time
delay Δt=0. No adjustable parameters were used in this
comparison, which shows the quality of the polarization used in
the experiment. (b) Initial anisotropy spectrum R( λ, 0) for
instantaneous two-photon absorption in buffer solution.
53

xii

Figure 3.6 (Top) UV pump-UVprobe experimental configuration for
anisotropy measurements. L
1
, L
2
and L
3
are the 15 cm, 20 cm,
and 12.5 cm focusing lenses, respectively. Neutral density
wheels for pump beam (ND
1
). (WP) is a zero-order half-wave
plate designed for 266 nm and (CP) is a compensating plate to
equalize both optical paths for the same dispersion. (Bottom) The
setup used for the simultaneous detection of parallel and
perpendicular signals. The probe light is split into parallel and
perpendicular components with a Wollaston polarizer after the
sample, which are then simultaneously imaged onto a single
photo-diode array operated on a shot-by-shot basis. The
polarization of probe beam was set to 45° with respect to the
pump polarization and was separated into the parallel and
perpendicular components with a Wollaston polarizer.
56

Figure 4.1 Transient absorption data of thymidine in buffer solution after
excitation at 267 nm with 35 fs pump pulses at room
temperature. The two-dimensional data was analyzed with a
global fitting routine (details in the text). (a) Color contour plot
of the dispersed pump-probe data. (b) Transient spectra at
different probe times as a function of wavelength. Experimental
data (dots) and global fit (solid lines).
62

Figure 4.2 Decay associated spectra extracted from global fitting for
electronically excited S
2
( ππ*) and S
1
(nπ*) states and the long-
lived channel with time decay constants of 80, 750, and > 5 ps,
respectively. The S
2
( ππ*) spectrum consists of two parts: a
stimulated emission part (negative absorbance) from 300 to 370
nm and a broadband excited-state absorption from 370 and
extending to visible wavelengths. Open dots represents the
inverted and scaled steady-state fluorescence spectrum for
thymidine reproduced from Ref.
18
. (Insert) The simple
sequential kinetic model, used as a target to fit 2D dataset.
63

xiii
Figure 4.3 Excited state absorption bands assigned based on comparison to
TRPES as S
2
( ππ*) and S
1
(nπ*) extracted from sequential kinetic
model of the transient absorption data of thymidine in solution
after excitation at 267 nm. To extract the shape of the excited
state absorption of S
2
( ππ*) we assume that the stimulated
emission spectrum is static and matches the steady-state
fluorescence spectrum. Therefore, after the subtraction of
fluorescence band (dotted line) from the decay-associated spectra
of S
2
, we obtain the ESA band (dash line) from S
2
.
65

Figure 4.4 (Left) Comparison of the spectra of the excited states, assigned
as S
2
( ππ*) and S
1
(nπ*) obtained from (a) TRPES data of
thymine in gas phase after excitation at 250 nm (from Ref
9
) and
(b) transient absorption data of thymidine in solution after
excitation at 267 nm. The spectra are plotted with identical
energy scales. Dash lines connect the maximums of two bands
between gas and solution phase pictures. (Right) Schematic
representation of pump-probe experiments in gas and solution
phases (see text and Figure 4.5 for details).
66

Figure 4.5 Schematic energy level diagram for both experiments. The
corresponding vertical ionization potentials (from ground state
thymine) IP
0
= 9.2 eV and IP
1
= 10.1 eV are known from
photoelectron spectroscopy (binding energies) to D
0
( π
-1
) and
D
1
(n
-1
), respectively.
19,20
The pump laser prepares the optically
bright state S
2
. Due to ultrafast internal conversion, this state
converts to the state S
1
. The dashed arrows indicate electron
emission.
68

Figure 4.6 The chemical structures and standard ring numbering for the
uracil derivatives studied.
73

Figure 4.7 Normalized absorption spectra of all uracil and thymine
derivatives studied. (a) thymidine (Thd), thymine (Thy), 1-
methyl-thymine (MeThy), and thymidine monophosphate
(TMP), the spectra of TMP and Thd are identical in this graph;
(b) Uracil (Ura), Uridine (Urd), 1-Methyl-Uracil (MeUra); (c)
absorption spectrum shift between Urd and Thd. (d) first
absorption band of Thd and the spectrum of the 35 fs pump
pulse.
78

xiv
Figure 4.8 Time-resolved spectra of Thd, Thy, 1-Me-Thy, and TMP in
buffer solution after excitation at 267 nm with 35 fs pump pulses
at room temperature at different probe times as a function of
wavelength. Data is recorded with pump at magic angle with
respect to probe.
81

Figure 4.9 Time-resolved spectra of Urd, Ura, and 1-Me-Ura in buffer
solution after excitation at 267 nm with 35 fs pump pulses at
room temperature at different probe times as a function of
wavelength. Data is recorded with pump at magic angle with
respect to probe.
82

Figure 4.10 Time-resolved spectra, S
PA
( λ, t), S
PE
( λ, t), which correspond to
transient absorption signals with the polarization of the pump
pulse set parallel and perpendicular, respectively, with respect to
the polarization of the probe pulse. (a,b) for Thd, (c,d) Urd.
Arrows show the red-shift and blue-shift by the same energy
(0.26 eV) of the stimulated emission peak (from 335 nm to 360
nm) and excited state absorption peak (from 390 nm to 360 nm),
respectively. The thin lines in (a) and (c) are the steady state
fluorescence spectra for Thd and Urd.
86

Figure 4.11 UV pump-UV probe transient absorption of Thd and Urd at
267nm (bleaching of the ground state population). (a) Magic
angle data, (b) bleach anisotropy.
88

Figure 4.12 Decay associated spectra extracted from global fitting for
Uridine, uracil, and 1-Methyl-Uracil. Simple three-level
sequential kinetic scheme were used.
90

Figure 4.13 Comparison of the spectra of the excited states, assigned as
S
2
( ππ*) and S
1
(nπ*) obtained from (top) TRPES data of thymine
and uracil in the gas phase after at 250 nm (from Ref
12
) and
(bottom) transient absorption data of thymine and uracil in
solution after excitation at 267 nm. To extract the shape of the
excited state absorption of S
2
( ππ*) we assume that steady state
emission matches with dominant stimulated emission component
in S
2
spectrum. Therefore after the subtraction of emission band
from the decay-associated spectra of S
2
thick line (ESA) shows
clean S
2
absorption band.
91

xv
Figure 4.14 Comparison of the ESA of S
2
( ππ*) extracted from global fit
(solid lines) with the time-resolved spectrum, S
PE
( λ, t), which
correspond to transient absorption signals with the polarization
of the pump pulse set perpendicular with respect to the
polarization of the probe pulse (dots).
92

Figure 4.15 The ESA spectra for uracil, thymine and their 1-methylated
derivatives.
94

Figure 4.16 Comparison of the global analysis results with the experimental
ground state bleaching data for uridine. (a) Decay-associated
spectra and (b) population dynamics extracted from global fitting
using a simple three-level sequential kinetic scheme (see insert).
Following the sequential model, the S
2
and S
1
channel lifetimes
are short (70 fs and 500 fs, respectively) in comparison to the
long-lived channel (> 5ps). This yields 100% of the population to
be found in the long-lived channel after ~2ps. (c) UV pump-UV
probe transient absorption of uridine at 267 nm (bleaching of the
ground state population). The latter graph shows that only ~30%
of the population is recovered to the ground state by 30 ps, and
the remaining ~70% of the population is trapped as a long-lived
species. The bleaching experimental data is in contradiction with
the population dynamics extracted from global fit and, therefore,
in disagreement with our simple sequential kinetic scheme.
96

Figure 4.17 (a) Decay-associated spectra and (b) population dynamics
extracted from global fitting using a three-state branching model
(see insert) for uridine. The dynamics reflects that after
excitation the population decays into the long-lived channel with
~70% yield. This yield ratio was accomplished only by applying
the branching model and varying the decay-time constants, in
order to be in agreement with ground state recovery ratio
obtained from the bleaching data.
97

Figure 4.18 Comparison of the long-lived signals obtained from our transient
absorption data (solid line) and from literature (dashed line).
14,15

The dashed line was assigned as triplet state spectrum. The
overlaying of our data and literature triplet data clearly shows
that our data is not representative of the triplet state.
98

xvi
Figure 4.19 (a) Decay associated spectra extracted from global fitting for the
S
2
( ππ*) and S
1
(nπ*) states and the S
0
(hot). The S
2
( ππ*)
spectrum consists of two parts: a stimulated emission part
(negative absorbance) from 300 to 370 nm and a broadband
excited-state absorption from 370 and extending to visible
wavelengths. The black line represents the inverted and scaled
steady-state fluorescence spectrum for uridine reproduced from
Ref.
6
To extract the shape of the excited state absorption of
S
2
( ππ*) we assume that the stimulated emission spectrum is
static and matches the steady-state fluorescence spectrum.
Therefore, after the subtraction of fluorescence band (black line)
from the decay-associated spectra of S
2
, we obtain the ESA band
(blue line) from S
2
. (b) We further split the ESA band into two
bands, the contribution to the ESA from the UV and the visible
regions. This is required to fit the anisotropy data. Anisotropy
values assigned to the UV and visible ESA bands are -0.2 and
0.4, respectively.
100

Figure 4.20 Overlays of the experimental and simulated anisotropies at
different delay times in the 300-700 nm range.
101

Figure 5.1 The molecular structures for 9H-adenine and 7H-adenine
(adenine tautomers), 2-amino-purine (2AP), 2’-deoxy-adenosine
(Ado), and 2’-deoxy-adenosine 5’-monophosphate (AMP).
104

Figure 5.2 Normalized absorption spectra of nucleic acid building blocks.
(a) Absorption spectrum of adenosine-monophosphate (AMP)
and the spectrum of the 35 fs pump pulse at 267 nm.; (b)
Adenine (Ade) and 2-amino-purine (2AP); (c) Triphosphates:
Guanosine-triphosphate (GTP), Cytidine-triphosphate (CTP),
and Uridine-triphosphate (UTP). The dotted vertical shows the
central wavelength of excitation.
112

xvii
Figure 5.3 Pump dispersed probe data for ATP in aqueous solution excited
at 267 nm. The total transient optical density (mOD) is indicated
by the color according to the colorbar. The blue shifting band
between 290-400 nm has been previously assigned to
vibrationally hot S
0
.
6
It is now clear that there is a previously
unresolved stimulated emission contribution at 307 nm (negative
amplitude). The ESA band in the visible (500-700 nm) decays
with ~ 220 fs. Note, there is no solvated electron contribution at
long time. (Top) Absorption (solid line) and emission
9
(dash
line) spectrum of ATP in aqueous solution, and excitation
wavelength (arrow).
113

Figure 5.4 Transient spectra of ATP, adenosine, adenine, and AMP in
buffer solution after excitation at 267 nm at different probe
times.
117

Figure 5.5 Transient absorption data of ATP (similar traces with the same
time constants have been observed for adenosine, and AMP), in
buffer solution after excitation at 267 nm at different probe
wavelengths.
118

Figure 5.6 (a, b)Transient spectra of 2AP in buffer solution after excitation
at 267 nm at different probe times for early-time (from 100 fs to
4.5 ps) and long-time (from 1 –70 ps) dynamics. (c,d)
Corresponding contour plots of the dispersed pump-probe data.
120

Figure 5.7 UV pump-UV probe transient absorption of ATP, GTP, UTP,
and CTP at 267nm (bleaching of the ground state population).
122

Figure 5.8 UV pump-UV probe transient absorption of Adenosine at 267nm
in H
2
O (line) and D
2
O (dots).
123

Figure 5.9 Dispersion-adjusted transient absorption data for ATP in buffer
solution after excitation at 267 nm. The two-dimensional data
was analyzed with a global fitting routine. (a) Color contour plot
of the dispersed pump-probe data. (b) Transient spectra at
different probe times as a function of wavelength. Experimental
data (dots) and global fit (solid lines).
127

xviii
Figure 5.10 (a) Decay associated spectra for ATP, extracted from global
fitting for the electronically excited S
2
( ππ*) and S
1
states and
vibrationally excited ground state, S
0
*
, with time decay constants
of 55, 190, and 610 fs, respectively. The S
2
( ππ*) spectrum
consists of two parts: a stimulated emission part (negative
absorbance) from 290 to 330 nm and a broad excited-state
absorption from 330 and extending to visible wavelengths.
(Insert) The simple sequential kinetic model, used as a target to
fit the two-dimensional experimental dataset. (b) To extract the
shape of the ESA of the S
2
( ππ*) state we assume that the
stimulated emission spectrum is static and matches the steady-
state fluorescence spectrum.
9
Therefore, after the subtraction of
fluorescence band ( □) from the decay-associated spectra of S
2
,
we obtain the ESA band ( ○) from S
2
.
128

Figure 5.11 Comparison of the spectra of the excited states, S
2
( ππ*) and S
1

obtained from (a) TRPES data of 9-methyl-adenine in gas phase
after excitation at 267 nm (from Ref.
25
) and (b) transient
absorption data of ATP in solution after excitation at 267 nm.
The spectra are plotted with identical energy scales (eV).
132

Figure 6.1 Schematic diagrams of the electronic relaxation dynamics in
adenine in the gas phase reproduced from Ref.
2
See details in the
text.
143

Figure 6.2 Transient absorption data of ATP in solution after excitation at
267 nm. The spectra are plotted with identical energy scales
(eV).
146

xix
Abstract
One of the primary mechanisms of DNA damage occurs following irradiation
with high energy ultraviolet (UV) light. Consequently, the study of excited state
dynamics of nucleic acid bases upon UV excitation is essential towards
understanding and mediating DNA photodamage. Due to the extremely short sub-
picosecond (10
-12
s) lifetimes, most time-resolved studies on DNA bases are
hampered by the ability to generate and manipulate short UV laser pulses. The
development of an ultrashort UV-pulse (~30 fs) source reported here now makes it
possible to characterize very rapid dynamics that is simply not observable in
previous lower time resolution experiments. Results of experiments combining
broadband UV/Visible dispersed probing with simultaneous polarization resolution
are presented for isolated free adenine and uracil derivatives in aqueous solution at
room temperature. Both spectral and dynamical data is acquired providing the most
detailed view of the excited-state dynamics to date.
For thymidine, the spectra of all transient species have been identified and
compare surprisingly well with corresponding time-resolved photoelectron spectra
from the same intermediate states in the gas-phase. The proposed model for
electronic deactivation is, thus, analogous to gas-phase dynamics where there is an
intermediate state between the optically bright ππ* and the final ground state during
electronic relaxation. Additional experimental studies for uridine show ultrafast
branching in the initial ππ* state: a fraction of the excited-state population decays via

xx
internal conversion to the ground state, the rest of the population decays to the n π*
state.
For adenosine, ultrafast electronic relaxation from the excited ππ* states to
the ground state is likely to take place via a ππ*/ S
0
conical intersection. However,
gas-phase time-resolved photoelectron spectroscopy predict the same n π* state
intermediate relaxation model. Therefore, the deactivation mechanism is different for
the gas and solution phase dynamics. This provides direct evidence of the solvent
effect on the excited-state potential energy surfaces, especially in the region of
conical intersections.

1
Chapter 1. Introduction

1.1 Ultrafast Excited-State Dynamics in Aqueous Solution: Electronic
Deactivation of DNA Bases
Deoxyribonucleic acid (DNA) is the genetic material of living organisms. In
human cells, environmental factors such as ultraviolet (UV) exposure from sun light
can cause DNA photodamage, resulting in individual molecular lesions. Many of
these lesions produce structural damage to the DNA molecule and can change or
eliminate the cell's ability to copy the genetic code. Other lesions induce potentially
harmful mutations in the cell's genome. Consequently, a complete knowledge of
DNA photochemistry induced by UV light is crucial to the understanding of
mutation-generating photochemical reactions.
The DNA molecule is a long biopolymer, and its monomer constituents are
the nucleic acid bases (Figure 1.1). The bases absorb UV light of wavelengths
shorter than 300 nm strongly and are the dominant UV absorbers in many cells, i.e.
the most important UV biochromophores.
1
The nucleic acid bases are amazingly
photostable due to their ultrafast deactivation of excited singlet states to the ground
state (< 1ps) via radiationless internal conversion.
2
In spite of this remarkable
protection mechanism from photochemical damage, the UV induced lesions still
occur in the large DNA macromolecule. One of the key questions in DNA
photophysics and photochemistry is to understand what happens to the excitation

2

Figure 1.1 The chemical structures and standard ring numbering for the five
nucleic acid bases.

N O
N
O
CH
3
Thymine
N O
N
O
H
H H
H
H
H H
Adenine Guanine
N
N
NH
2
N
N
H
H
H
N
N
N H
2
N
N
O
H
H
H
N O
N
NH
2
H
H
H
1
2
3
4
5
6
1
2
3
4
5
6
7
8
9
Uracil Cytosine

3
energy deposited in a nucleobase by UV light and where the photodamage occurs in
the DNA. The first step in answering these questions is determining how the
nucleobase responds to light, i.e. what is the initial excited state dynamics of an
isolated nucleobase.
Recently, there has been renewed interest in the spectroscopy of isolated
nucleobases in the gas and condensed phases.
2
This is mainly due to advances in
subpicosecond ultrafast spectroscopy in the deep UV,
3-14
molecular beam studies of
nucleobases and their complexes,
15-18
and new methods and the increasing power of
quantum calculations.
19-36
One of the main goals of these investigations is to
understand the mechanism of the radiationless decay process, which leads to rapid
deactivation of individual bases from the UV excited state. So far, the excited-state
dynamics and mechanisms of the radiationless decay have been found to be
complicated and sensitive to factors such as phase and excitation energy. Clearly,
several electronic states, which are in a close proximity to the optically bright state,
may strongly interact with each other through vibronic coupling. However, even the
energy levels of these electronic states are still unresolved.
The most significant recent results have been for room temperature bases and
nucleosides (nucleobases with deoxyribose rings) in a biologically important
environment, aqueous solution. After many attempts over the past two decades,
several groups have successfully employed ultrafast time-resolved spectroscopy and
have directly measured the excited state lifetimes. In such measurements, a
femtosecond UV pump pulse prepares the molecule in its optically bright, excited S
2

4
( ππ*) singlet state (Figure 1.2). Data from Kohler’s group, for example, for the
nucleosides show the following decay times: 0.3 ps (adenosine), 0.5 ps (guanosine),
0.5 ps (thymidine), and 0.7 ps (cytidine).
13,14
These lifetimes have been measured by
detecting excited-state absorption (ESA) to higher lying singlet states using
femtosecond transient absorption spectroscopy. Similar lifetimes have been detected
by measuring emission decays in fluorescence upconversion measuments
6,8,37,38
. In
general, the time resolution in all studies to date is still comparable to the time scales
that are being reported.
In the femtosecond transient absorption experiments
13,14
studied by Kohler
and co-workers, the S
2
( ππ*) singlet state is initially excited by a 150 fs UV pump
pulse at 263 nm. Since the singlet excited states lie more closely as the energy is
increased, the ESA transitions usually lie at longer wavelengths than absorption
bands of the ground state. Therefore, the excited-state absorption is probed at a
visible wavelength ( λ
ESA
~570 nm). For example, the transient absorption signal for
adenosine has been described by monoexponential decay with a 290 fs time constant.
The transients at several separate probe wavelengths in the visible ( λ~ 450-700 nm)
have shown identical time constants, whereas for the transients at near-UV probe
wavelengths ( λ~ 340-270 nm), the decay time increases dramatically as the probe
wavelength decreases. The time constant increases from 0.4 ps at 340 nm to 2.0 ps at
270 nm. The authors originally assigned the visible band to the ESA from the S
2

state with the lifetime of 290 fs. The near-UV band has been assigned to absorption

5

Figure 1.2 A schematic diagram of excited-state photophysics in an isolated
nucleobase. A UV pump pulse prepares the molecule in its optically bright, S
2
(ππ*)
singlet excited state. The nucleic acid bases have allowed a π → π* transition, which
is responsible for their first strong absorption band at ~270 nm. After the excitation,
the molecule returns to the ground state by internal conversion (denoted by wavy
solid line in the diagram), which competes with fluorescence. A very rapid internal
conversion (< 1ps) leads to direct reduction of the electronic excited state lifetime.
All nucleic acid bases are known to have such radiationless deactivation, which
results in their very low fluorescence and phosphorescence quantum yields. Very
close in energy lie at least one other singlet state, the optically dark S
1
(nπ*) state. A
widely discussed mechanism for nucleobase electronic relaxation involves two
sequential steps: ππ*→ nπ* followed by nπ* → S
0
internal conversion to the ground
state.
Allowed
S
0

S
2
(ππ∗)
S
1
(n π ∗)
Absorption
Fluorescence
Phosphorescence
Intersystem
Crossing
T
1

Weakly allowed
Singlet Triplet

6
by hot ground state molecules formed after internal conversion. In Kohler’s picture,
during rapid deactivation, molecules, which internally convert from S
2
→ S
0
*
are
vibrationally very highly excited in the ground state. The high vibrational energy
gives rise to a red-shift of the ground-state absorption (S
0
*
→S
2
). Therefore,
compared to normal (room temperature) ground-state absorption which ends at ~ 290
nm, the hot ground-state absorption is proposed to extend up to 400 nm. Once the
vibrational energy dissipates into the solvent the absorption spectrum returns to the
normal shape.
Although Kohler and co-workers have succeeded in directly observing ESA
in nucleosides using improved femtosecond transient absorption spectroscopy, the
results have been difficult to interpret due to the single wavelength analysis. With a
limited number of transients recorded at several separate probe wavelengths, no
spectral dynamics can be distinguished. For instance, the near-UV band has been
entirely assigned to absorption by the hot ground state. However, higher lying ESA
bands, which have absorption in the near-UV are expected. These bands are likely to
contribute to the strong transient signals observed in the experiments. Further, a
fraction of molecules initially excited to the S
2
state emits before it decays
nonradiatively. Since the emission spectrum has maximum values at 305-335 nm,
the stimulated emission should be observable. However, in the solution phase
experiments with poor time resolution, the initial dynamics of all the UV transients
are buried under a huge instrument-limited spike at time zero arising from two-
photon pump-probe absorption process in the solvent.

7
Comprehensive fluorescence studies of nucleosides in aqueous solution have
been performed by Gustavsson and co-workers using the upconversion technique,
37

providing both fluorescence decays and fluorescence anisotropy data. The work
revealed that the fluorescence decays are complex and cannot be described by single
exponentials. The biexponential fits showed one very fast component and a longer
one. Here, it is useful to note that absorption spectrum of each nucleoside consists of
several overlapping bands corresponding to different electronic transitions. For
example, a 267-nm pump pulse populates the two ππ* excited states in adenosine.
Based on this argument it is possible to consider that the biexponential behavior is
due to involvement of both ππ* excited states in the fluorescence decay process.
However, for another nucleoside, thymidine, this argument is not valid because 267-
nm excitation populates only one ππ* state. Again, as in transient absorption studies,
the fluorescence decays have been recorded at several individual emission
wavelengths. The authors have concluded that to understand the excited state
dynamics in detail, the time-resolved emission spectra in a broad probe window have
to be obtained.
In contrast to the mechanism described in the recent literature for the solution
phase, the electronic deactivation has been thought by most investigators in the gas
phase to take place to the ground state through the optically dark S
1
(nπ*) state.
Broo
39
suggested that the S
2
( ππ*) and the S
1
(nπ*) states are almost isoenergetic. A
very small energy barrier separates the singlet states. The minimum of the n π* state
is slightly below the minimum of the ππ* state. Thus, after excitation of the

8
nucleobase molecule to the Franck-Condon (FC) region of the ππ* state, an efficient
nonadiabatic transition takes the molecule down to a new minimum, which has
already the n π* electronic character. The relaxation to the ground state has been
proposed by Lim
40
who suggested that the vibronic coupling between the ππ* and
nπ* states can decrease the energy gap between those states and the ground state. In
addition, the increased Franck-Condon factors lead to efficient internal conversion to
the ground state.
While the femtosecond transient absorption
13,14
and fluorescence
upconversion studies
37
have brought some insight in the description of the excited-
state dynamics, it is clear that temporal and spectral information of the excited states
are important to understand the precise mechanism of ultrafast deactivation. Perhaps
the most prominent recent results where both dynamical and spectral characterization
of the excited states have been acquired to reveal the mechanism of ultrafast
deactivation are the gas phase studies of isolated nucleobases using time-resolved
photoelectron spectroscopy (TRPES) presented by Stolow and co-workers.
3,4,41

TRPES has the ability to directly observe the dynamics of both the bright (ππ*) and
dark (n π*) electronic excited states involved in ultrafast process because the time-
and energy-resolved photoelectron spectra of the ππ* and n π* states can be well
separated. Since ionization is always allowed, the optically dark states can be probed.
From these measurements Stolow has concluded that the electronic relaxation in the
gas-phase involves the n π* state for all nucleobases.
3
The relaxation pathway can be
viewed as two sequential steps: S
2
( ππ*) −
τ1
→ S
1
(nπ*) −
τ2
→ S
0
, where S
2
is the

9
initially populated bright excited state. For adenine, the TRPES studies
41
show
ultrashort excited-state lifetimes: τ
1
= 70 fs and τ
2
= 1.1 ps. Moreover, an excitation
energy dependent branching ratio of the relaxation pathways has been observed for
adenine. At 267 nm pump wavelength, an additional relaxation pathway (S
2
( ππ*) →
πσ* → S
0
) has been found, which involves another optically dark, dissociative πσ*
state. This is consistent with previously proposed theoretical work by Sobolewski
and Domcke.
22,42

The exact mechanism responsible for the very rapid electronic relaxation is
not completely understood, however, many theoretical studies
20,23-36
suggest that
conical intersections play an important role in electronic relaxation dynamics and
provide a quick path for a nonradiative decay to the ground state (Figure 1.3). After
the excitation into a Franck-Condon region, the excited-state population begins to
evolve on the potential energy surface, and it has a high chance of finding a region
with conical intersections. Such crossings are not rare, and, in fact, many organic
photochemical reactions have now been shown to involve conical intersections.
43-46

In particular, for nucleobases, the nuclear motions that can lead to the conical
intersections are molecular ring deformations in the out-of-plane coordinate.
Radiationless decay from the upper to the lower intersecting state can occur on the
timescale of a single molecular vibration when the system moves near conical
intersection.
Theoretical studies
20,22,25,42
to locate conical intersections suggested a
ππ*/n π* ( ππ*/ πσ*) conical intersection followed by a n π*/S
0
( πσ*/S
0
) conical

10

Figure 1.3 A schematic representation of excited-state dynamics and various
deactivation pathways for the nucleobase from the Franck-Condon (FC) point after
photoexcitation. The conical intersection is the point where two potential energy
surfaces cross. At the point of intersection the two electronic states are degenerate,
and it is at this point the population of the excited state can decay from the upper to
the lower intersecting state. The more complete description of the conical
intersection can be distinguished in a reduced space called branching space (two-
dimensional space of two special normal modes) where the equations for two
surfaces appear as the geometric equations for cone.
47
The left and right endpoints
are the S
1
/S
0
and the S
2
/S
0
conical intersections, respectively. The crossing, which is
close to FC region, is the S
2
/S
1
conical intersection.

Conical
Intersection
S
0

S
2
(ππ ∗)
hν
Conical
Intersection
hν
S
1
(n π ∗)

11

intersection. Other quantum calculations
23,26
found the ππ* lower in energy and
proposed only one conical intersection, ππ*/S
0
. The involvement of three-state
conical intersections in the photophysics and radiationless decay processes of the
nucleobases has been also investigated.
24
These theoretical studies have contributed
to the effort of elucidating the details of the dynamics. However, the studies are
relevant to gas phase nucleobases, and the effect of solvent has been neglected in
these quantum calculations, though it is expected to play a major role. Indeed, the
time-resolved studies in aqueous solution
6,8,13,14,37,38,48
demonstrate shorter lifetimes
than those in the gas-phase
4,41,49
. Also, there is no excitation energy dependence in
solution phase
38
, in contrast to the gas phase observations.
49
In future theoretical
work, solvation models will be used to add the effect of the solvent. Meanwhile,
more experimental information is required in order to help developing the
appropriate theoretical models to describe the solution phase excited-state
deactivation of the nucleobases.
In this work we present high-time resolution studies of adenine and uracil
derivatives in aqueous solution at room temperature using dispersed transient
absorption spectroscopy. Adenine and uracil derivatives are the most extensively
studied DNA monomers. The wealth of gas phase and theoretical literature allows a
detailed comparison of the solution phase with the isolated molecule dynamics.
To initiate photophysics of a nucleic acid molecule with high-time resolution
requires <50 fs pulses in the deep ultraviolet region (260-270 nm), a difficult spectral

12
range for femtosecond sources. The femtosecond UV-pump pulses in all previous
studies to date were still 150-200 fs. We developed a unique 30-femtosecond tunable
UV laser source
50
based on phase-matched four-wave mixing in a hollow core fiber.
The implementation is presented in Chapter 2. The fiber source generates 30 fs UV
pump pulses at 267 nm and is continuously tunable over 225-240 nm range. These
are the shortest deep UV pump pulses to date originating from a commercial 100-fs
Ti:Sapphire lasers. With the help of these ultrashort UV pump pulses, the solvent
spike at time zero arising from two-photon pump-probe absorption process in the
solvent will be significantly shortened in time domain. Consequently, the near-UV
transient spectrum is better resolved and then deconvoluted to distinguish between
stimulated emission, excited state absorption, and vibrationally hot ground state.
Further, the tunability of the UV source in the 225-240 nm range can be exploited for
excitation energy dependence studies, which is poorly understood in solution phase.
This interesting future research is described in detail in Chapter 6.
To acquire both spectral and dynamical characterization providing the most
detailed view of the excited-state dynamics, a new UV pump-dispersed probe
apparatus is implemented to study ultrafast excited-state dynamics of individual
adenine and uracil derivatives in solution over a wide spectral range with
unprecedented time resolution (~ 40fs). Therefore, its development is essential in this
research. In Chapter 3 we describe the construction and characterization of the UV
pump-dispersed probe apparatus. We show that the temporal resolution of the
apparatus is almost the same across the whole range of the probe wavelengths

13
(typically, 290-700 nm) and estimated to be 40 fs. This parameter is important
because, the transient absorption data are then analyzed using target spectral
analysis.
Chapter 4 presents a detailed picture of ultrafast excited-state dynamics of
thymidine in solution. The spectra of all transient species have been identified. We
made direct comparison between TRPES in the gas-phase and our dispersed transient
absorption data in aqueous solution (Figure 1.4). Based on these results we believe
we have developed a methodology to map the electronic character associated with
the specific excited state as well as the dynamical timescales in the relaxation of
nucleobases. New experimental data presented in the Postscript to Chapter 4 reveal
the complexity in the electronic relaxation mechanism. Our results for uridine and
thymidine show ultrafast branching in the initial ππ* state; a fraction of the excited-
state population decays via internal conversion to the ground state (S
2
( ππ*) → S
0
),
the rest of the population decays to the n π* state, which decays to S
0
on slower time
scale (S
2
( ππ*) −
fs
→ S
1
(n π*) −
ps
→ S
0
).
For adenosine, the longest band in the ground state absorption spectrum is, in
fact, due to two energetically close ππ* states. We have predicted the following
relaxation pathway: S
3
( ππ*) → S
2
( ππ*) → S
0
. Thus, ultrafast electronic relaxation of
adenosine molecules from the excited ππ* states to the S
0
ground state is likely to
take place via a ππ*/ S
0
conical intersection. Therefore, the proposed model for

14

Figure 1.4 Schematic description of pump-probe experiments in gas and solution
phases. In the TRPES experiment, Koopmans’ correlations allow for a mapping of
different excited states onto different ionic states. The known character of the excited
states permits assignment of the various photoelectron bands observed. We directly
compare TRPES in the gas-phase with our dispersed transient absorption data in
aqueous solution. Detailed examination of the ESA bands of our analysis and their
photoelectron spectra shows that they have quite similar Franck-Condon envelopes
and similar spacing between bands. The excellent agreement between the ESA and
TRPES bands leads us to speculate that optical absorption in the liquid from S
2
( ππ*)
and S
1
(nπ*) occurs to upper states with the same origin as in gas-phase
photoionization, namely Rydberg upper states with cation cores D
0
( π
-1
) and D
1
(n
-1
).
(State reorder) In aqueous solution, the transition energy to the n π* state is expected
to shift to higher energy, while the ππ* state should shift only slightly to lower
energies since in protic solvents ππ* states are more stabilized by dipole-dipole and
hydrogen bonding interactions than n π* states. (Right) Excited state absorption
bands extracted from global analysis of the transient absorption data.

Ryd (D
1
)
Ryd (D
0
)
Probe
Energy
ESA
State reorder
Δt
D
1
(n
-1
)
D
0
( π
-1
)
S
1
(n π∗)
S
2
( ππ ∗)
S
0

267 nm
D
2
( π
-1
)
D
3
(n
-1
)
200 nm
S
2
( π π ∗)
Δt
267 nm
S
1
(n π∗)
Gas Phase Aqueous Solution
TRPES
Dispersed Transient
Absorption
Ryd (D
2
( π
-1
))
Ryd (D
3
(n
-1
))

15
electronic deactivation is not analogous to gas-phase dynamics where n π* state acts
as an intermediate state between optically bright ππ* and ground state during
electronic relaxation.

16
1.2 References for Chapter 1

(1) C. R. Cantor and P. R. Schimmel, Biophysical Chemistry. (W.H. Freeman
and Co., San Francisco, 1980).

(2) C. E. Crespo-Hernandez, B. Cohen, P. M. Hare, and B. Kohler, Chem. Rev.
104, 1977 (2004).

(3) S. Ullrich, T. Schultz, M. Z. Zgierski, and A. Stolow, Phys. Chem. Chem.
Phys. 6, 2796 (2004).

(4) S. Ullrich, T. Schultz, M. Z. Zgierski, and A. Stolow, J. Am. Chem. Soc. 126,
2262 (2004).

(5) S. K. Pal, J. Peon, and A. H. Zewail, Chem. Phys. Lett. 363, 57 (2002).

(6) J. Peon and A. H. Zewail, Chem. Phys. Lett. 348, 255 (2001).

(7) T. Gustavsson, A. Sharonov, and D. Markovitsi, Chem. Phys. Lett. 351, 195
(2002).

(8) T. Gustavsson, A. Sharonov, D. Onidas, and D. Markovitsi, Chem. Phys.
Lett. 356, 49 (2002).

(9) D. Markovitsi, A. Sharonov, D. Onidas, and T. Gustavsson, ChemPhysChem
4, 303 (2003).

(10) A. Sharonov, T. Gustavsson, V. Carre, E. Renault, and D. Markovitsi, Chem.
Phys. Lett. 380, 173 (2003).

(11) B. Cohen, P. M. Hare, and B. Kohler, J. Am. Chem. Soc. 125, 13594 (2003).

(12) C. E. Crespo-Hernandez and B. Kohler, J. Phys. Chem. B 108, 11182 (2004).

(13) J.-M. Pecourt, J. Peon, and B. Kohler, J. Am. Chem. Soc. 122, 9348 (2000).

(14) J.-M. L. Pecourt, J. Peon, and B. Kohler, J. Am. Chem. Soc. 123, 10370
(2001).

(15) E. Nir, K. Kleinermanns, L. Grace, and M. S. de Vries, J. Phys. Chem. A 105,
5106 (2001).

17
(16) E. Nir, M. Muller, L. I. Grace, and M. S. de Vries, Chem. Phys. Lett. 355, 59
(2002).

(17) E. Nir, C. Plutzer, K. Kleinermanns, and M. de Vries, Eur. Phys. J. D 20, 317
(2002).

(18) H. Kang, B. Jung, and S. K. Kim, J. Chem. Phys. 118, 6717 (2003).

(19) B. Mennucci, A. Toniolo, and J. Tomasi, J. Phys. Chem. A 105, 7126 (2001).

(20) N. Ismail, L. Blancafort, M. Olivucci, B. Kohler, and M. A. Robb, J. Am.
Chem. Soc. 124, 6818 (2002).

(21) A. Holmen, A. Broo, B. Albinsson, and B. Norden, J. Am. Chem. Soc. 119,
12240 (1997).

(22) A. L. Sobolewski and W. Domcke, Eur. Phys. J. D 20, 369 (2002).

(23) M. Merchan and L. Serrano-Andres, J. Am. Chem. Soc. 125, 8108 (2003).

(24) S. Matsika, J. Phys. Chem. A 109, 7538 (2005).

(25) S. Matsika, J. Phys. Chem. A 108, 7584 (2004).

(26) T. Gustavsson, A. Banyasz, E. Lazzarotto, D. Markovitsi, G. Scalmani, M. J.
Frisch, V. Barone, and R. Improta, J. Am. Chem. Soc. 128, 607 (2006).

(27) H. R. Hudock, B. G. Levine, H. Satzger, S. Ullrich, A. Stolow, and T. J.
Martínez, J. Am. Chem. Soc., submitted (2006).

(28) R. Improta and V. Barone, J. Am. Chem. Soc. 126, 14320 (2004).

(29) M. Z. Zgierski, S. Patchkovskii, and E. C. Lim, J. Chem. Phys. 123, 081101
(2005).

(30) M. Z. Zgierski, S. Patchkovskii, T. Fujiwara, and E. C. Lim, J. Phys. Chem. A
109, 9384 (2005).

(31) S. B. Nielsen and T. I. Sølling, ChemPhysChem 6, 1276 (2005).

(32) S. Perun, A. L. Sobolewski, and W. Domcke, J. Am. Chem. Soc. 127, 6257
(2005).

(33) S. Perun, A. L. Sobolewski, and W. Domcke, Chem. Phys. 313, 107 (2005).

18

(34) L. Blancafort and M. A. Robb, J. Phys. Chem. A 108, 10609 (2004).

(35) C. Marian, D. Nolting, and R. Weinkauf, Phys. Chem. Chem. Phys. 7, 3306
(2005).

(36) C. M. Marian, J. Chem. Phys. 122, 104314 (2005).

(37) D. Onidas, D. Markovitsi, S. Marguet, A. Sharonov, and T. Gustavsson, J.
Phys. Chem. B 106, 11367 (2002).

(38) T. Pancur, N. K. Schwalb, F. Renth, and F. Temps, Chem. Phys. 313, 199
(2005).

(39) A. Broo, J. Phys. Chem. A 102, 526 (1998).

(40) E. C. Lim, J. Phys. Chem. 90, 6770 (1986).

(41) H. Satzger, D. Townsend, M. Z. Zgierski, S. Patchkovskii, S. Ullrich, and A.
Stolow, Proc. Natl. Acad. Sci. USA 103, 10196 (2006).

(42) A. L. Sobolewski, W. Domcke, C. Dedonder-Lardeux, and C. Jouvet, Phys.
Chem. Chem. Phys. 4, 1093 (2002).

(43) M. Klessinger and J. Michl, Excited States and Photochemistry of Organic
Molecules. (VCH Publishers Inc., New York, 1995).

(44) M. A. Robb, F. Bernardi, and M. Olivucci, Pure Appl. Chem. 67, 783 (1995).

(45) D. R. Yarkony, J. Phys. Chem. A 105, 6277 (2001).

(46) M. Ben-Nun and T. J. Martinez, Adv. Chem. Phys. 121, 439 (2002).

(47) D. R. Yarkony, Rev. Mod. Phys. 68, 985 (1996).

(48) W.-M. Kwok, C. Ma, and D. L. Phillips, J. Am. Chem. Soc. 128, 11894
(2006).

(49) H. Kang, K. T. Lee, B. Jung, Y. J. Ko, and S. K. Kim, J. Am. Chem. Soc. 124,
12958 (2002).

(50) A. E. Jailaubekov and S. E. Bradforth, Appl. Phys. Lett. 87, 021107 (2005).

19
Chapter 2. Tunable 30-femtosecond pulses across the deep ultraviolet*

2.1 Introduction
In recent years, new sources of tunable sub-50 fs pulses in the ultraviolet
(UV) spectral range have been developed based on non-collinear optical parametric
amplification (NOPA) and sum frequency mixing techniques.
1-4
Recently, Baum et
al. have demonstrated the generation of pulses as short as 7-fs by achromatic
frequency doubling of NOPA output.
2
Sub-10-fs pulses in the 275-335 nm range can
be generated in relatively thick β-barium borate (BBO) crystals to ~0.25 μJ in this
scheme. However, for many experimental studies in photochemistry
5-7
and
photobiology
8,9
in particular would substantially benefit from more intense sources
tunable into the deep ultraviolet (DUV) while maintaining ultrashort pulse durations.
Generation of DUV ( ≤ 266 nm) pulses in the same way would assume an achromatic
sum-frequency process, which has not so far been demonstrated.
Alternative frequency conversion approaches using noble gases as nonlinear
optical media can extend the available UV spectral range to shorter wavelengths and
avoid absorption in the generation medium.
10
To obtain high efficiency conversion,
the Kapteyn and Murnane group has demonstrated the phase-matched generation of
UV light in long argon-filled hollow core fibers.
11
Phase matching for a given
wavelength is achieved by variation of the pressure of argon inside the fiber. Using a

* Appears as A. E. Jailaubekov and S. E. Bradforth, Appl. Phys. Lett. 87, 021107 (2005).

20

parametric four-wave mixing (FWM) process, pumped by 20 fs pulses from a
Ti:Sapphire multi-pass amplifier, they were able to generate intense 8 fs pulses at
270 nm with up to 2.5 μJ.
12
Further, a preliminary report from the same group
suggested the possibility of tunable output in the DUV (214-244 nm) by introduction
of an optical parametric amplifier (OPA) idler seed.
13
However, no characterization
of the pulse energy and duration, or demonstration of the tunability of DUV pulses
so generated has been reported to date.
In this chapter, we demonstrate that phase-matched optical parametric
generation in argon-filled waveguides can be efficient even with ≥100 fs
fundamental pump pulses, making this method suitable for use with standard
commercial regenerative amplifier systems. By increasing the peak intensity in the
waveguide we are able to achieve the similar conversion efficiency as obtained
previously by Durfee et al.
12
with 20 fs pump pulses. The generation and
characterization of DUV pulses in the 220-240 nm range is also presented. Both self-
phase and cross-phase modulation inside the waveguide are responsible for
generation of ultra-broadband DUV pulses with bandwidths that are sufficient to
support sub-20 fs pulses. We show that it is possible to compress these pulses with a
simple and high transmission prism-pair arrangement to achieve 25 fs pulses at 266
nm and 31 fs pulses at 233 nm.

21
2.2 Experimental and Results
In the first experiment, we used difference-frequency FWM of second-
harmonic light with the fundamental (3 ω=2 ω+2 ω- ω) to reproduce experiments
described in ref.
12
and verify whether broad bandwidth pulses could be generated
with high conversion efficiency with relatively long driving pulses. The experimental
setup
14
is shown in Figure 2.1. 800 μJ, 110 fs pulses at a repetition rate of 1 kHz are
generated from a Ti:Sapphire regenerative amplifier system (Spectra Physics
Hurricane). A fraction of the 800 nm laser output (240 μJ) is down-collimated using
a Galilean telescope (2:1 ratio) to generate the pump light (70 μJ) at frequency 2 ω by
second harmonic generation in a 0.5 mm thick BBO crystal. After separation of the
two colors with a dichroic beam splitter, the 400-nm and 800-nm beams are then
recombined and focused into a 30 cm long flexible fused silica capillary (ID = 75
μm, OD = 363 μm, Polymicro, TSP075375). To keep the capillary straight, it is
threaded inside hard-wall glass tubing (ID = 0.4 mm, OD = 6 mm), which was in
turn evacuated and filled with argon gas. CaF
2
windows and ultra-torr fittings are
mounted on both ends of the tubing. The argon pressure is adjusted to maximize the
phase-matched UV signal.
11,12,14

The fiber input (output) pulse energies under vacuum conditions for the 400-
nm and 800-nm pulses were measured to be 65 μJ (28 μJ) and 64 μJ (4 μJ),
respectively. For the dimensions of the waveguide, the maximum theoretical
throughputs are 68% and 20% for the second-harmonic and fundamental beams,

22

Figure 2.1 Experimental setup. L1, L2 are plano-convex (f =10 cm) and plano-
concave (f = -5cm) lenses, respectively. BBO is a nonlinear crystal (0.5 mm, type I)
and L3, L4 are focusing lenses (f = 25 cm) mounted on XYZ translation stages. An
iris (ID) is used to adjust coupling efficiency of pump beam. λ/2 is a half wave plate.
W1, W2 are 2-mm thick CaF
2
windows. L5 is a collimating lens (f = 50 cm) and DM
a dichroic mirror. PS is a 90-degree periscope system.
(1200-2200 nm)
Ti:S

110 fs
800 nm
1 kHz
OPA
L1
L2
L3
L4
L5
BBO
λ/2
30:70
W1
W2
(400 nm)
ID
DM
PS
Argon
Autocorrelator

23
which propagate in the lowest order mode EH
11
.
14-16
The lower throughputs
experimentally achieved reflect that incomplete mode coupling, rather than
propagation inside the fiber, is currently limiting. Pinhole measurements show that
the focused spot size of the pump beam is approximately 100 μm, therefore, a small
fraction of the 400-nm power is clipped by the waveguide walls. However, at tighter
focusing conditions more power is coupled to the higher order modes, thereby
decreasing overall throughput.
The optimized argon pressure (253 torr) for phase-matched 266-nm light
(3 ω) is in excellent agreement with the predicted value assuming that all beams are
propagating in their lowest order mode EH
11
.
11
The measured UV pulse energies
were up to 8.5 μJ (before the prism compressor), which corresponds to 30%
conversion efficiency expressed as a fraction of 400-nm pump light coming out of
the fiber. The result is comparable to conversion efficiencies reported in ref.
12
with a
140 μm diameter fiber pumped by 35 fs pulses. In general, the third harmonic power
can be estimated by the simple relationship, P
3 ω
~ L
2
/a
4
, where L and a are the length
and the radius of the fiber, respectively.
17
Thus, if we use a 140 μm diameter fiber
we generate 12 times less power. Alternatively, further increasing the peak intensity
of the pump beam by taking a fiber diameter smaller than 75 μm results in lower
coupling efficiency and a larger propagation loss.
Having shown excellent power can indeed be achieved at 266 nm with a
commercial regeneratively amplified 800 nm source, we now turn our attention to
generating tunable DUV via FWM of 400 nm pulses mixed with continuously

24
tunable light from an OPA. Here, infrared pulses (1200-2200 nm) are used as the
seed pulses. The output is given by ω
DUV
= ω
pump
+ ω
pump
- ω
seed
, giving rise to DUV
light in the range 220-240 nm. The experimental configuration (Figure 2.1) for this
frequency-mixing scheme is as follows. 560 μJ of the 800-nm laser output is used to
pump a collinear double pass OPA (Spectra-Physics, OPA-800C), to give tunable
infrared (IR) light in the range of 1200-1600 nm (signal) and 1600-2200 nm (idler).
The pump and seed beams are focused and recombined into the hollow core fiber,
which is now 10 cm long (shortened due to diminished propagation length of mid-IR
in the fiber).
Figure 2.2 shows the DUV pulse energy generated by phase-matched FWM.
The phase-matching pressure was adjusted through 300-600 Torr as the OPA is
tuned. The maximum output energy of 2.5 μJ at 237 nm was measured after four
dichroic mirrors (used to remove residual 400-nm light), but before the prism
compressor. Pulse energies > 1 μJ were obtained throughout the 224-240 nm range.
Figure 2.2 also shows the OPA seeding energy in front of the fiber. The coupling
efficiency of the IR beam has not been as well optimized as the 800-nm beam earlier.
As the seed wavelength lengthens, the focus waist at the fiber entrance increases.
Approximately 30% of the IR pulse energy is thought to be currently coupled into
the fiber. This likely accounts for the steeper roll off in generated DUV power on the
short wavelength side of Figure 2.2. We note that a study of the DUV output power
dependence showed that seed energies as low as 10 μJ are sufficient to generate 1-2
μJ DUV pulses.
14
Figure 2.2 also shows the spectra of the DUV pulses across the

25

Figure 2.2 (Top) DUV output tuning curve (thick line; bottom and left hand
axes) where pulses are obtained by mixing either OPA idler ( ●) or signal ( ▲) pulses
as seed for the fiber. ( □) Corresponding OPA pulse energies (thin line; right and top
axes) measured at the entrance to fiber. (Bottom) Generated pulse spectra with fiber
seeded by tunable IR pulses. Spectral broadening is ~5 nm (FWHM) corresponding
to transform-limited pulse durations < 20 fs.

2200 2000 1800 1600 1400 1200
10
20
30
40
50
60
220 225 230 235 240
0.0
0.5
1.0
1.5
2.0
2.5
Wavelength, nm
220 230 240 250
OPA pulse energy, μJ
OPA wavelength, nm
DUV pulse energy, μJ
DUV pulse wavelength, nm
DUV
OPA
idler
OPA
signal

26
tuning range. According to the broad spectral bandwidth obtained, sub-20 fs pulses
tunable in the DUV can be expected if full compression is achieved.
The broadening in the output spectra is shown in greater detail in Figure 2.3;
spectral broadening in the fiber has been analyzed for the case of FWM mixing
generation at 266 nm. The measured bandwidths for the incoming 400-nm and 800-
nm pulses are 2.3 nm (140 cm
-1
) and 11.7 nm (180 cm
-1
), respectively. Due to self-
phase modulation, the bandwidth of 400-nm pulses emerging from the fiber is
broadened to 7.6 nm (470 cm
-1
). The spectrum of the generated 266 nm light is
shown in Figure 2.3(a). The full width at half maximum (FWHM) is 7.4 nm (1050
cm
-1
), which is considerably broader than that expected from conversion of the
broadened 400 nm in the perturbative limit.
12
The additional broadening occurs
because of the cross-phase modulation
12,17
driven by the 400 nm pulses, which are
the most intense in the waveguide. Figure 2.3(a) also shows a direct comparison of
pulses generated when standard sum frequency mixing (SFM; 3 ω=2 ω+ ω) in a thin
BBO crystal is substituted for the fiber FWM approach. The spectra of the UV pulses
generated in the hollow core fiber are > 4 times broader than those produced by SFM
in the 80 μm thick crystal. The reduced upconverted bandwidth in the χ
(2)
crystal
conversion process is well understood and is due to group-velocity mismatch
between pulses in the finite length crystal. The walkoff problem is effectively
eliminated in the hollow core fiber approach by use of a rare gas as the propagation
medium.

27
Figure 2.3 (a) Output spectra of 266-nm light generated by FWM in a hollow
core fiber (thick line) and by SFM in an 80- μm-thick BBO crystal (thin line, FWHM
= 1.4 nm). Fiber parameters: 30 cm length, 75 μm diameter. Pulsewidths in
parentheses are transform-limited pulse durations corresponding to indicated spectral
bandwidths (assumed Gaussian). (b) Spectrum of UV pulse centered at 233 nm
generated by FWM. Experimental autocorrelation (points) of (c) 266-nm and (d)
233-nm pulses after prism compression; (solid curve) fit assuming Gaussian pulse
shape.

257 262 267 272 277
0.0
0.5
1.0

7.4 nm
(14 fs)
(a)

Amplitude, a.u.
Wavelength, nm
223 228 233 238 243

6.1 nm
(13 fs)
(b)

Wavelength, nm
-150 -100 -50 0 50 100
0.0
0.2
0.4
0.6
0.8
1.0
τ
pulse
= 25 fs
(c)

Absorption, a.u.
Time delay, fs
-150 -100 -50 0 50 100
(d)
τ
pulse
= 31 fs

Time delay, fs

28
Pulse compression has so far been performed using an isosceles Brewster-
angle CaF
2
prism pair (69
o
apex angle). It is a convenient and low loss method for
compression (75% throughput for double-pass transmission), however, a substantial
residual third order dispersion (TOD) is expected. Grating pair compression is
typically preferred to avoid TOD, but in the DUV high reflective losses from
gratings are inevitable. An autocorrelation technique exploiting two-photon
absorption in a ~50 μm flowing liquid water film
18
was chosen to characterize the
ultrashort 266-nm and DUV pulses. The autocorrelation FWHM after optimizing of
the prism compressor is 35 fs (Figure 2.3(c)) for 266-nm pulses and 44 fs (Figure
2.3(d)) for 233-nm pulses corresponding to pulsewidths of 25 fs and 31 fs,
respectively, assuming a Gaussian pulse shape. The shortest UV pulses at 266 nm
and 233 nm were obtained at a prism separation of 28.5 cm and 18 cm, respectively.
The experimental trace clearly deviates from Gaussian; the pedestal which is most
prominent at 266 nm in the observed autocorrelation originates from the substantial
residual TOD expected using the prism-based compressor.
The only other techniques for generating tunable DUV below 250 nm of
which we are aware have been based on sum frequency generation in BBO and
suffer greatly from temporal walkoff of the three frequencies, as well as crystal
damage due to absorptive losses. So far, achromatic upconversion techniques, which
show strong promise for wavelengths longer than 270 nm,
2
have not been extended
to a sum frequency process. The capabilities of the hollow core fiber approach to
tunable ultrashort ultraviolet generation are thus highly complementary in the

29
spectral range accessible to achromatic NOPA doubling. Additional advantages,
which include an order of magnitude greater pulse energies, ease and cost of
implementation (as no adaptive optics are required), excellent stability (0.6% r.m.s.
at 266nm) and spectrally and spatially homogeneous beam quality suggest that this
alternate technology will make ultrashort DUV pulses accessible for a broad base of
scientific applications.

2.3 Summary
In conclusion, we have confirmed that the FWM approach in capillary
waveguides is compatible with commercial 100 fs regenerative amplifier technology
and demonstrated its extension to produce tunable ultrashort pulses in the 220 – 240
nm region with pulse energy exceeding 1 μJ. The fiber waveguide can withstand
high intensity focused beams without significant optical damage. Long term stability
is superior to the use of non-linear crystals. We expect that, based on the excellent
spectral characteristics of the generated pulses, simple instrumental improvements in
compression can be made to achieve more nearly transform-limited pulses. We note
that extending the range of tunability of this apparatus to longer wavelength (241 –
299 nm) can be achieved by use of the frequency doubled OPA signal or idler (600 –
1100 nm) as a seed for the FWM. The achromatic doubling technique of Baum et al.
however is likely preferable for broad bandwidths longer than 270 nm.
2

The exceptional characteristics of this source in terms of pulse-to-pulse
energy stability, mode quality and tunability will enable pump-probe and non-linear

30
spectroscopic measurements with unprecedented time resolution. Studies on small
molecule photochemistry and photoionization in both the gas
19,20
and condensed
phase
5,21
will now be possible in a wavelength region where hitherto sub-30 fs pulse
have not been available.

2.4 Acknowledgements
We thank Lino Misoguti and Margaret Murnane for their guidance and
encouragement in the initial construction of the hollow core fiber apparatus. This
work is supported by the National Science Foundation (CHE-0311814) and by the
Packard Foundation.

2.5 Supplement: Basics of wave propagation and phase-matching in hollow
waveguides
Hollow core waveguides differ significantly from conventional optical fibers:
hollow dielectric waveguides (capillaries) guide laser beams through Fresnel
(grazing incidence) reflections at the inner wall of the capillary instead of by total
internal reflection. Whereas the mode structure is similar to that of a step-index
optical fiber, all the modes are inherently lossy because of partial transmission at the
walls of the capillary.
15
The throughput of the hollow core fiber, the ratio of the
output to the input intensities, is given by:
) exp(
0
l
I
I
nm
α − =
(1)

31
The field loss rate α
nm
(or attenuation coefficient) for the mode with azimuthal and
radial modal indices m and n in a capillary of radius a is
1
1
2
2
2
3
2
2
−
+
⎟
⎠
⎞
⎜
⎝
⎛
=
ν
ν λ
π
α
a
u
nm
nm
air
glass
n
n
= ν
(2)
where u
nm
is the modal constant, λ is the wavelength, ν is the ratio between refractive
indices of the external (fused silica) and internal (gas) media. The lowest-order mode
in a capillary waveguide is the linearly polarized EH
11
mode and its modal constant
u
11
=2.405. The mode EH
11
is closest to the free-space TEM
00
Gaussian mode. As it
is seen in Figure 2.4, for sufficiently long waveguides the mode discrimination is
very high so that only the fundamental mode EH
11
can propagate. The strong
dependence of the transmission from the capillary radius and the laser wavelength
(Figure 2.5, 2.6) should also be considered when capillary length is selected.
By countering diffraction of a wave, a waveguide adds a geometrical
component to the wavevector.
22
The propagation constant for a capillary filled with a
medium of index n is given by
()
()
2
2
2
4
2 2
2 2
1
1
2
a
u
P
a
u n
k
nm
nm
π
λ
λ δ
λ
π
λ
π
π
λ
λ
λ π
− + ≈
≈
⎥
⎥
⎦
⎤
⎢
⎢
⎣
⎡
⎟
⎠
⎞
⎜
⎝
⎛
− =
(3)
In the approximate form for propagation constant above, the index of refraction for a
gas is written in the form n( λ) = 1 + P δ( λ), where P is the pressure of the gas, δ( λ)
contains the gas-dispersion function. Thus, in this approximation the k-vector now

32

Figure 2.4 Calculated transmissions of the EH
11
fundamental mode and that of
the subsequent, linearly polarized, high-order modes EH
12
(u
12
=5.5201) and EH
13

(u
13
=8.6537) for a capillary waveguide with a diameter 150 μm, at 800 nm, as a
function of the capillary length (in cm).
0 2040 6080 100
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1
0
% 800 1.45332 , 150 , l , 2.4048 , ()
% 800 1.45332 , 150 , l , 5.5201 , ()
% 800 1.45332 , 150 , l , 8.6537 , ()
100 0l
11
EH
12
EH
13
EH
nm m 800 , 150 μ

33

Figure 2.5 Calculated transmissions of the EH
11
fundamental mode for a
capillary waveguides with the diameters 75, 100 and 150 μm, at 800 nm, as a
function of the capillary length (in cm).
0 2040 6080 100
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1
0
% 800 1.45332 , 150 , l , 2.4048 , ()
% 800 1.45332 , 100 , l , 2.4048 , ()
% 800 1.45332 , 75 , l , 2.4048 , ()
100 0l
m μ 150
m μ 100
m μ 75
11
EH

34
0 2040 6080 100
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1
0
% 800 1.45332 , 75 , l , 2.4048 , ()
% 400 1.45332 , 75 , l , 2.4048 , ()
% 267 1.45332 , 75 , l , 2.4048 , ()
100 0l

Figure 2.6 Calculated EH
11
mode propagation in the fiber with the diameter of
75 μm at three different wavelengths 267, 400 and 800 nm, as a function of the
capillary length (in cm).
11
, 75 EH m μ
nm 267
nm 400
nm 800

35

consists of three sources of dispersion: vacuum, gas and waveguide. For a particular
nonlinear mixing process, the total phase mismatch ( Δk) is given by the vector sum
of all waves involved, where the k-vector for each wave is given by equation (2). By
adjusting the gas pressure, waveguide radius and spatial mode in which the beams
propagates, the phase mismatch can be tuned to achieve the phase-matching
condition of Δk = 0. For example, for our particular case of difference-frequency
four-wave mixing process, ω
signal
= 2 ω
pump
− ω
seed
, the phase mismatch is Δk =
2k
pump
− k
seed
− k
signal
, or, where all beams in the same spatial mode ( u
nm
= u )
()
⎥
⎦
⎤
⎢
⎣
⎡
− − −
⎥
⎥
⎦
⎤
⎢
⎢
⎣
⎡
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
− − = Δ
signal seed pump
signal
signal
seed
seed
pump
pump
a
u
P k λ λ λ
π λ
δ
λ
δ
λ
δ
π 2
4
2
2
2
2
= Δk
gas
− Δk
mode
(4)
The phase mismatch results from a gas dispersion term ( Δk
gas
∝ P) minus a
modal dispersion term ( Δk
mode
∝ 1/a
2
). Since Δk
mode
>0, and in gases with normal
dispersion Δk
gas
>0, there will exist an optimum pressure P
opt
, for which Δk = 0.
Thus, the advantage of the phase-matching technique using hollow waveguides is
that the mode of the generated light is determined by phase matching, making it
possible to balance the modal and the material phase mismatches.
The signal field in non-depleted-pump approximation
17,22

[ ]
3
3
* 2 ) 3 (
3
1 ) exp(
) (
k
z k i E E ik
z E
seed pump
signal
Δ
− Δ
=
χ
(5)

36
where z is interaction length. Or
⎟
⎠
⎞
⎜
⎝
⎛ Δ
=
2
2 2 2
kz
inc s z E
signal
(6)
Hence, the intensity of the signal depends quadratically on interaction length. At the
same time, the phase mismatch Δk is dependent on the pressure P, the wavelength λ,
the modal constant u and the core radius a, therefore, the intensity is also dependent
indirectly on the pressure:
()
()
⎟
⎠
⎞
⎜
⎝
⎛ Δ
=
2
;
2 2 2
z P k
inc s z P z E
signal
(7)

37
2.6 References for Chapter 2

(1) I. Z. Kozma, P. Baum, S. Lochbrunner, and E. Riedle, Opt. Express 11, 3110
(2003).

(2) P. Baum, S. Lochbrunner, and E. Riedle, Opt. Lett. 29, 1686 (2004).

(3) A. Kummrow, M. Wittmann, F. Tschirschwitz, G. Korn, and E. T. J.
Nibbering, Appl. Phys. B 71, 885 (2000).

(4) P. Tzankov, T. Fiebig, and I. Buchvarov, Appl. Phys. Lett. 82, 517 (2003).

(5) A. C. Moskun and S. E. Bradforth, J. Chem. Phys. 119, 4500 (2003).

(6) V. H. Vilchiz, J. A. Kloepfer, A. C. Germaine, V. A. Lenchenkov, and S. E.
Bradforth, J. Phys. Chem. A 105, 1711 (2001).

(7) N. A. Anderson, C. G. Durfee, M. M. Murnane, H. C. Kapteyn, and R. J.
Sension, Chem. Phys. Lett. 323, 365 (2000).

(8) S. Ullrich, T. Schultz, M. Z. Zgierski, and A. Stolow, J. Am. Chem. Soc. 126,
2262 (2004).

(9) J.-M. L. Pecourt, J. Peon, and B. Kohler, J. Am. Chem. Soc. 123, 10370
(2001).

(10) S. Backus, J. Peatross, Z. Zeek, A. Rundquist, G. Taft, M. M. Murnane, and
H. C. Kapteyn, Opt. Lett. 21, 665 (1996).

(11) C. G. Durfee, S. Backus, M. M. Murnane, and H. C. Kapteyn, Opt. Lett. 22,
1565 (1997).

(12) C. G. Durfee, S. Backus, H. C. Kapteyn, and M. M. Murnane, Opt. Lett. 24,
697 (1999).

(13) L. Misoguti, S. Backus, C. G. Durfee, R. Bartels, M. M. Murnane, and H. C.
Kapteyn, in Ultrafast Phenomena XII, edited by T. Elsaesser, S. Mukamel,
M. M. Murnane, and N. F. Scherer (Springer-Verlag, Berlin, 2001), pp. 112.

(14) A. E. Jailaubekov, Master's thesis, University of Southern California, 2003.

(15) E. A. J. Marcatili and R. A. Schmeltzer, Bell Syst. Tech. J. 43, 1783 (1964).

38

(16) M. Nisoli, S. Stagira, S. D. Silvestri, O.Svelto, S. Sartania, Z. Cheng, M.
Lenxner, C. Spielmann, and F. Krausz, Appl. Phys. B 65, 189 (1997).

(17) A. M. Zheltikov, Phys. Usp. 45, 687 (2002).

(18) M. J. Tauber, R. A. Mathies, X. Y. Chen, and S. E. Bradforth, Rev. Sci.
Instrum. 74, 4958 (2003).

(19) A. Stolow, A. Bragg, and D. Neumark, Chem. Rev. 104, 1719 (2004).

(20) J. Davies, R. Continetti, D. Chandler, and C. Hayden, Phys. Rev. Lett. 84,
5983 (2000).

(21) M. Bauer, C. Lei, R. Tobey, M. M. Murnane, and H. C. Kapteyn, Surf. Sci.
532, 1159 (2003).

(22) C. G. Durfee, L. Misoguti, S. Backus, H. C. Kapteyn, and M. M. Murnane, J.
Opt. Soc. Am. B 19, 822 (2002).

39
Chapter 3. Experimental: The UV Pump-Dispersed Probe Apparatus

3.1 Introduction
Femtosecond transient absorption spectroscopy has proven to be a powerful
technique to study ultrafast chemical reaction dynamics in solution. In a typical
transient absorption experiment a reaction is initiated by the absorption of a strong
pump pulse, and its dynamics is followed by recording the absorbance of a weak
probe pulse at a single wavelength as a function of the time delay between the pump
and probe pulses, producing

one-dimensional data. If several transient species with
different lifetimes are formed after excitation and they all absorb light at the
wavelength of the probe pulse the data will show a multiexponential behavior. The
analysis of this data can be complicated, as a result, the reaction dynamics can be
interpreted differently.
It is possible, however, to study the reaction over a wide range of probing
wavelengths. In practice, one often employs a so-called white-light continuum (or
broadband) pulse as a probe. Focusing of an intense ultrashort laser pulse into a
dense and transparent medium such as glass or water generates the white-light
continuum. The white-light continuum can be made to extend from the near
ultraviolet to the near infrared.
1
After the sample, the white-light continuum (WLC)
is dispersed by a spectrograph and detected by a multichannel detector which allows
for the simultaneous measurement of the intensities of the probe pulse for all

40
wavelengths present in the continuum. Again, by recording the absorbance of the
WLC probe pulse as a function of the time delay between the pump and probe pulses
one can now obtain a two-dimensional

time-resolved spectrum.
The dispersed (or sometimes, broadband, in the literature) pump-probe using
the white-light continuum is an efficient technique and in the last decade it has been
used for many experimental studies in photochemistry
2-7
and photobiology
8-11
and
more recently for DNA photophysics
12,13
. In addition to the rich data supplied by the
dispersed-probing technique, it is technically advantageous due to the following
reasons. Firstly, since the probe pulse contains a broad range of wavelengths, the
whole transient absorption spectrum is recorded within a single laser pulse shot. This
feature not only decreases the data acquisition time, but also eliminates the variation
of the data related to inaccurate beam alignment and laser intensity fluctuation,
which is unavoidable for the single-wavelength probing where measurements at
different wavelengths are performed one by one.
Secondly, in order to maintain high time resolution of the experiment in a
single-wavelength probing setup, the probe pulse should be compressed. A separate
compression optical line, which includes a pair of prisms, is designed for that
purpose. The prism compressor adjusts group velocity dispersion or temporal chirp
of the femtosecond probe pulse so that it becomes equal to zero, i.e. the spectral
components of the pulse have no shift in time domain. The prism compressor should
be perfectly aligned and optimized via pulsewidth measurements. The same routine
should be repeated every time the wavelength of the probe light is changed. In the

41
dispersed-probe setup we can get rid of the prism compressor making the optical
setup more simple and robust. To achieve high time resolution, the group velocity
dispersion is adjusted mathematically by correction of the time-dependent shift of all
spectral components. This is possible to perform only in two-dimensional time-
resolved data.
In general, the overall time resolution of the femtosecond pump-probe
experiment in the condensed phase is determined by the three parameters: the
pulsewidths of both pump and probe pulses, and temporal walkoff between pump
and probe pulses due to their group velocity mismatch. Obviously, these three
parameters should be minimized in order to improve the time resolution. Our UV
pump-dispersed probe apparatus is equipped with a hollow core fiber setup, which
generates 25-30 fs UV pump pulses at 266 nm and is continuously tunable over 225-
240 nm range. These are the shortest deep UV pump pulses to date originating from
a commercial 100-fs Ti:Sapphire lasers. The pulsewidth of the WLC probe pulses is
“compressed” by performing the dispersion correction as described earlier. In this
case, again, we obtain the shortest probe pulses possible, which are limited only by
high-order dispersion.
The last parameter, the temporal walkoff, becomes critical when the pump
and probe pulses are deep UV and visible wavelengths, respectively. The walkoff
can be minimized by reducing the thickness of the sample. A wire-guided gravity-
driven jet was developed in our lab to produce optically stable thin films of liquids of
6-100 μm.
14
The temporal walkoff is a wavelength-dependent parameter, therefore, it

42
is also important that variation of this parameter across the range of the probe
wavelengths (typically, 290-700 nm) is small. For example, the temporal walkoff of
the 266-nm pump and the 700-nm (290-nm) probe pulses in the 50 μm jet is 22 fs (7
fs), which is less than the pump pulsewidth.
The UV pump-dispersed probe apparatus enables us to study ultrafast
excited-state dynamics of individual DNA bases in solution over a wide spectral
range with unprecedented time resolution (~ 40fs), therefore its development is
essential in this research. In this Chapter we describe the construction and
characterization of a new UV pump-dispersed probe apparatus. Further
developments of the apparatus will be also presented which are necessary to study
ultrafast dynamics in condensed phase.

3.2 Construction of the Pump-Dispersed Probe Apparatus
Figure 3.1 shows the experimental configuration of the UV pump-dispersed
probe apparatus. A commercial 1mJ Ti:Sapphire regenerative amplifier laser system
(Hurricane, Spectra-Physics) operating at 1 kHz produces short duration (~100 fs full
width at half-maximum) laser pulses at 800 nm. These laser pulses (~250 μJ/pulse)
are subsequently split into two parts. The first portion is sent to drive hollow core
fiber setup, producing excitation light between 220-240 nm or 266 nm and is used as
an excitation source. The hollow core fiber setup has been described in detail in
Chapter 2. Briefly, the 266nm pump pulses were obtained by converting 800 nm
laser pulses using a phase-matched four-wave mixing scheme in the fiber. An

43

Figure 3.1 Pump-probe experimental configuration. Arrow shows a beam
emerging from Ti:Sapphire regenerative amplifier (Hurricane, Spectra-Physics,
1kHz). The 800 nm beam with horizontal polarization is split into two beams using
a beamsplitter (BS). ND: neutral density wheels for pump beam (ND
1
) and probe
beam (ND
2
); L
1
, L
2
and L
3
are the 15 cm, 20 cm, and 12.5 cm focusing lenses,
respectively. (WP) a zero-order half-wave plate designed for 266 nm. White light
continuum is generated in a 2-mm thick CaF
2
disk. The disk is mounted on
motorized stage. White light continuum is collimated with parabolic mirror P
1
and
focused into sample with parabolic mirror P
2
. Both mirrors are aluminum coated 90°
off-axis parabolic mirrors with effective focusing length of 5 cm (P
1
) and 18 cm (P
2
).
To avoid a strong scattering of the intense 800 nm part of the continuum beam in the
spectrograph (S) a dichroic mirror (D) is used to reject 98% of 800 nm light
(bandwidth = 75 nm at 45° P-polarization).
Hollow core
fiber setup
Probe delay
BS
ND
2
266 nm
L
1
WP L
2

CaF
2

Chopper
P
1
Sample
D
S
L
3

P
2
ND
1
OPA

44
additional tunable IR seed from a collinear two-stage infrared OPA is required to
obtain UV pulses in the 220-240 nm range. The fiber was filled with argon gas, the
pressure of which was optimized for best conversion efficiency into the third
harmonic. The duration of the pump pulses at the sample is measured at ~30 fs. The
intensity of the pump light at the sample was maintained at 0.2-1.5 μJ/pulse with
neutral density filters. The polarization of the pump was set to a magic angle relative
to that of the probe by using a half-wave plate.
After going through an optical delay translation stage, a second portion (2-3
μJ) of the 800-nm laser pulses is used to generate a white-light continuum (Figure
3.2) by focusing into a 2 mm thick CaF
2
plate with a 15 cm singlet lens. CaF
2

material is chosen as it gives rise to considerable extension at the blue edge (inset of
Figure 3.2) To prevent burning of the CaF
2
plate, the plate was slowly (~0.5 cm/s)
translated in the plane, which is perpendicular to the focused beam. The CaF
2
plate is
mounted on a motorized stage making a circular motion. The white-light continuum
extends to 290 nm and is used as a probe beam. The probe light is achromatically
collimated and focused (spot size ~ 80 μm) in the sample with the use of two off-axis
(90°) parabolic mirrors (5 cm and 18 cm, effective focusing length) to reduce the
group velocity dispersion.
The pump beam was focused in the sample (spot size ~200-300 μm) with a
20 cm focal length lens at a crossing angle of 5-10°. The sample was delivered using
a wire-guided gravity-driven flow jet of ~50-100 μm thickness. After the spatial and
certain temporal overlap of the two beams in the sample, the white-light probe beam

45
300 400 500 600 700
0
100
200
300
400
500
600
Wavelength, nm
250 300 350 400 450 500
10
100

Log (Amplitude), mV
BaF
2
LiF
Sapphire
CaF
2
Amplitude, mV
Wavelength, nm

Figure 3.2 White-light continuum spectra in various materials. Four kinds of
material were tested in our lab: sapphire, BaF
2
, CaF
2
and LiF. The spectra were taken
with a visible grating. CaF
2
material is chosen as it demonstrated several suitable
characteristics for probe light such as wide spectral range (290-700 nm), small
degradation among all fluoride materials, good shot-to-shot pulse stability (0.3 –1%
r.m.s.) across whole range, and excellent phase characteristics. Inset graph shows
continuum intensities on a log scale at the blue edge.

46
passes through a dichroic mirror (800 nm high reflector, R>99%, bandwidth ~76 nm,
CVI Laser, TLM2-800-45-P-1012) to reject the intense infrared around 800 nm. The
probe beam is then focused on the entrance slit (100 μm) of a spectrograph (Oriel,
Spectra-Physics, 0.125m). The instrument spectral resolution is ~ 2.5 nm, an
estimate based on focusing the probe light to a spot size of ~ 100 μm, twice as wide
as the pixel size. The spectrograph is equipped with two gratings with different
blaze angles designed to have high reflectance efficiencies in the UV (300-400 nm)
and visible range (400-700 nm). Both gratings disperse approximately 300 nm of the
spectrum onto a single photo-diode array. The linear diode array contains 256 pixels
(Hamamatsu, part number: S3901-256Q) with superior characteristics such as low
dark current, high saturation charge, good linearity, and wide spectral response (200-
1000 nm). The spectrum is calibrated using a set of interference filters at 355, 400,
500, 550 and 600 nm.
The spectrum from the diode array is digitized at the full repetition rate of the
laser with a 512 kSample/second 16 bit A/D (United Electronic Industries, PDL-
mFPD2M FS4-500/16) and processed on shot-by-shot basis in a computer.
Significant filtering of the data can be achieved in this way to reject spectra for
outlier pump pulses or sample film glitches. A spectra selection is performed in real
time by the LabView program by removing those that do not fall within the standard
deviation from an average spectrum. Typically, the selection ratio was around 85%,
and ~2000 transient absorption spectra (2 seconds) were averaged per time delay.
Thus a complete time-resolved spectral development can be acquired in ~30 minutes.

47
A phase-locked chopper was used to block every other pump pulse (500 Hz). The
maximum changes in absorption in our experiments were 5-7 mOD, with a noise
level of 0.05 mOD. A transient absorption spectrum with the minimum signals of
less than 100 μOD can be recorded with decent signal-to-noise ratio (noise level,
5%) at a specific time delay within 10-15 min. This procedure allows to study the
long-time dynamics (after 5-10 ps, no dispersion correction required) of transient
species which have a very low extinction coefficient.

3.3 Characterization the UV Pump-Dispersed Probe Apparatus:
Temporal Resolution and Dispersion
The white light continuum is chirped due to self-phase modulation and group
velocity dispersion in the material in which it is generated. Therefore different
spectral components of the probe enter the sample at different times for any specific
setting of the pump-probe delay. Figure 3.3 (a) shows a typical time-resolved
transient absorption spectrum of pure water after excitation at 267 nm before the
dispersion correction procedure. The trace of the cross-phase modulation and two-
photon absorption is clearly seen on this spectrum. Dispersion-corrected data (Figure
3.3(b)) is then constructed by fitting the cross-phase modulation signals, calculating
the dispersion of the white light probe pulse, and then applying the determined
wavelength-dependent time shift to the raw data. The shift error in the dispersion
corrected data is less than 10 fs. The dispersion of the continuum is estimated at

48

-2
0
2
4
6
8
10
12
300 350 400 450 500 550
0
0.1
0.2
0.3
0.4
0.5
Wavelength (nm)
Probe Time (ps)

-2
0
2
4
6
8
10
12
300 350 400 450 500 550
-0.1
0
0.1
0.2
0.3
0.4
0.5
Wavelength (nm)
Probe Time (ps)

Figure 3.3 Time-resolved transient absorption spectrum of neat water after
excitation at 267 nm with 35 fs pump pulses. Two-dimensional data was fitted with a
forth order polynomial fitting routine. (a) Color contour plot of the raw spectrum. (b)
Dispersion corrected spectrum.

49
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
− ×
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
− ≈ Δ
4
2
2 2
2
) (
2
1
sin
) (
exp ) , (
pump
probe
pump pump
probe
probe
t t
t OD
βτ
ω
βτ τ
ω
ω
~300 fs over the range of 300-600 nm. This is quite small compared to comparable
apparatuses reported in the literature
5
.
The theoretical and experimental background of the pump-dispersed-probe
technique where the continuum is treated as a single chirped pulse is presented
elsewhere.
2,15-17
In the nonresonant case with linearly chirped probe pulse one can
obtain the following expression
2
for the transient signal:

Here ΔOD(ω
probe,
t), ω
probe
, t(ω
probe
), τ
pump
, τ
probe
, β, and t
0
(ω
probe
) are the
experimentally measured pump-induced change in optical density, the spectral
component of the probe, the frequency-dependent pump-probe time delay, the pump
and probe pulse durations, the chirp rate, and the time-zero function, which is fitted
at the peak of the cross-phase modulation signals, respectively. Our experimental
parameters are τ
pump
= 30 fs, τ
probe
≈ 300 fs. Under this condition τ
pump
<< τ
probe
the
above equation can be simplified to

. (1)

The expression (1) should be compared with the cross-correlation function between
the pump and the spectral component of the WLC probe:
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
− − ×
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
− ≈ Δ
2 2
0
4
2
2 2
2
) ( ) ( ) (
2
1
sin
) (
exp ) , (
probe pump
probe probe
pump
probe
pump pump
probe
probe
t t t t
t OD
τ βτ
ω ω
βτ
ω
βτ τ
ω
ω
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
− ≈
2
2
) (
exp ) , (
cc
probe
probe cc
t
t F
τ
ω
ω

50
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
− ×
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
− ≈ Δ
4
2
2 2
2
) (
2
1
sin
) (
exp ) , (
cc
probe
cc cc
probe
probe
t t
t OD
βτ
ω
βτ τ
ω
ω
Here τ
cc
is the cross-correlation width, i.e. the temporal resolution of the experiment
at a specific wavelength. Therefore, the equation (1) can be modified as:
(2)

In this form equation (2) is used to fit the cross-phase modulation signals as shown in
Figure 3.4. The particular shape of the cross-phase modulation, with positive
maximum and negative wings before and after time-zero, is due to the sine function
in equation (2). Based on this fit, the cross-correlation at full width half maximum
(FWHM) is estimated to be ~ 40 fs for all probe wavelengths between 290 and 700
nm. Hence, the cross-correlation function is not wavelength dependent, which
implies that the temporal walkoff contribution is negligible. If the walkoff
contribution was not negligible, we would expect the higher values for the cross-
correlation widths at longer than at shorter wavelengths. Indeed, the group velocity
mismatch of the 266-nm pump and the 700-nm and 290-nm probe pulses in the 50
μm jet is 22 fs and 7 fs, respectively. In addition, the obtained value for the cross-
correlation width (40 fs) is close to the pump pulsewidth (30 fs). This is in
accordance with equation (1), which states that the cross-phase modulation width is
equal to the pump pulsewidth. The difference between τ
pump
and τ
cc
implies that
WLC probe has high-order dispersion.
Although equation (2) allows us to obtain a reasonable estimation for the
cross-correlation width, it is clear that the amplitudes of the transient signal cannot

51
-80 -60 -40 -20 0 20 40 60 80
-1
0
1
2
3
(mOD)
Time delay (fs)
350 nm

-80 -60 -40 -20 0 20 40 60 80
-0.2
-0.1
0.0
0.1
0.2
0.3
0.4
(mOD)
Time delay (fs)
500 nm

Figure 3.4 Experimental transient signals ΔOD(ω
probe,
t) (dots) obtained from
neat water, with pump pulse centered at 267 nm and two probe wavelengths, 350 nm
and 500 nm. Solid lines: best fit of the signal with equation (2). The dash line is the
cross-correlation function of 40 fs FWHM width.

52
be well fitted. As mentioned before, we have considered only the nonresonant case.
For cross phase modulation, the time-integrated signal at any probe wavelength is
zero. A two-photon absorption contribution has to be accounted for in the observed
deviation from the fit. This contribution monotonically increases as the probe
wavelength decreases. The detailed treatment of other contributions such as Raman
and two-photon absorption is given in Ref.
2
. Here, we limit ourselves only by the
electronic contribution to the transient signal, which is usually several times higher
in magnitude.

3.4 Anisotropy Experiments
Anisotropy experiments were performed by taking the separately measured
time-resolved spectra, S
PA
(λ, t), S
PE
(λ, t), and S
MA
(λ, t) which correspond to transient
absorption signals with the polarization of the pump pulse set parallel, perpendicular,
and at magic angle respectively, with respect to the polarization of the probe pulse
(horizontal). The pump half wave plate was rotated to obtain a certain polarization.
In this case, the anisotropy measurements were back-to-back scans, performed with
constant laser parameters (e.g. pump pulse intensity) and liquid jet conditions. To
test the polarized time-resolved spectra we reconstruct the magic angle data from the
experimental parallel and perpendicular signals using the relationship S
MA
(λ, t)
=(1/3)( S
PA
(λ, t)+2 S
PE
(λ, t)). The experimental magic angle and reconstructed data
were then compared to show the quality of the measured data (Figure 3.5 (a)).

53
Figure 3.5 (a) Experimental magic angle pump-probe spectrum (dots) and a
magic angle signal calculated directly from the experimental parallel and
perpendicular signals (solid line) using the relationship S
MA
(λ, 0) =(1/3)( S
PA
(λ, 0)+2
S
PE
(λ, 0)) at time delay Δt=0. No adjustable parameters were used in this
comparison, which shows the quality of the polarization used in the experiment. (b)
Initial anisotropy spectrum R(λ, 0) for instantaneous two-photon absorption in buffer
solution.
300 350 400 450 500 550 600
0
2
4
6
8
10
Experimental
Reconstructed

(mOD)
Wavelength (nm)
a)
300 400 500 600
0.0
0.2
0.4
0.6
0.8

Anisotropy, R(λ, 0)
Wavelength (nm)
b)

54
Immediately prior to depolarization measurements, care was taken to get the
cleanest possible polarization at the sample. The extinction and transmission of the
two polarization beams were tested with a polarizer placed at the sample. Extinction
was at least 3000:1 for pump polarization. The pump wave plate angle could
reproducibly be set to within a precision of 1 degree. The polarization purity of the
white-light continuum was not so clean as pump polarization. It was measured to be
~ 200:1 across the whole continuum. Despite the poor polarization of the probe we
still obtain good anisotropy data. Figure 3.5(b) shows instantaneous two-photon
absorption anisotropy values across whole probe light spectrum for buffer solution.
We checked that the initial anisotropy of buffer solution matches the value of 0.38
from UV pump-UV probe anisotropy value at 266 nm.
18

3.5 Applications in Ultrafast Spectroscopy and Future Work
In summary, a UV pump-dispersed probe apparatus has been developed. All
optical components are robust and the apparatus produces two-dimensional time-
resolved spectra with excellent signal-to-noise data (1-5%, r.m.s.). The time
resolution of the pump-dispersed probe apparatus is measured to be ~40 fs.
Broadband anisotropy measurements have also been implemented and show this to
be a promising technique for new ultrafast spectroscopy in the condensed phase.
Further developments of the setup are currently under construction in our lab
to improve data collection for dispersed anisotropy experiments. For simultaneous
detection of parallel and perpendicular polarization components of the probe beam, a

55
Wollaston polarizer will be used in combination with the spectrometer and two-diode
arrays. Currently, we have developed single color prototype of this technique. Figure
3.6 shows the UV pump-UV probe experimental configuration for anisotropy
measurements. Since the probe image on the diode array contains both parallel and
perpendicular signals in this case, the anisotropy is recorded within a single laser
pulse shot. This decreases the data acquisition time and removes errors in the data
related with laser intensity fluctuation. Other things to improve are related to each
optical component, i.e. the hollow core fiber, liquid jet, white-light continuum
generation. These will include: 1) the implementation of UV chirped mirrors to
improve the compression of pump pulses down to 15-20 fs; 2) the development of
the microfluidics systems to deliver microvolume samples, such as DNA oligomers;
3) the implementation of deep UV or far IR continuum light sources for dispersed
pump-probe setup.

56

Figure 3.6 (Top) UV pump-UVprobe experimental configuration for anisotropy
measurements. L
1
, L
2
and L
3
are the 15 cm, 20 cm, and 12.5 cm focusing lenses,
respectively. Neutral density wheels for pump beam (ND
1
). (WP) is a zero-order
half-wave plate designed for 266 nm and (CP) is a compensating plate to equalize
both optical paths for the same dispersion. (Bottom) The setup used for the
simultaneous detection of parallel and perpendicular signals. The probe light is split
into parallel and perpendicular components with a Wollaston polarizer after the
sample, which are then simultaneously imaged onto a single photo-diode array
operated on a shot-by-shot basis. The polarization of probe beam was set to 45° with
respect to the pump polarization and was separated into the parallel and
perpendicular components with a Wollaston polarizer.
Probe delay
CP
266 nm
L
1
WP
L
2

Chopper
Sample
L
3

ND
1
Inset
50 μm
Liquid jet
266 nm
pump
Diode Array
266 nm
Probe
S
⊥
S
ll
Wollaston
Polarizer

57
3.6 References for Chapter 3

(1) R. L. Fork, C. V. Shank, C. Hirlimann, R. Yen, and W. J. Tomlinson, Opt.
Lett. 8, 1 (1983).

(2) S. A. Kovalenko, A. L. Dobryakov, J. Ruthmann, and N. P. Ernsting, Phys.
Rev. A 59, 2369 (1999).

(3) S. A. Kovalenko, R. Schanz, H. Hennig, and N. P. Ernsting, J. Chem. Phys.
115, 3256 (2001).

(4) S. A. Kovalenko, N. Eilers-Konig, T. A. Senyushkina, and N. P. Ernsting, J.
Phys. Chem. A 105, 4834 (2001).

(5) M. Raytchev, E. Pandurski, I. Buchvarov, C. Modrakowski, and T. Fiebig, J.
Phys. Chem. A 107, 4592 (2003).

(6) H. Satzger, S. Sporlein, C. Root, J. Wachtveitl, W. Zinth, and P. Gilch,
Chem. Phys. Lett. 372, 216 (2003).

(7) H. Satzger, C. Root, and M. Braun, J. Phys. Chem. A 108, 6265 (2004).

(8) C. C. Gradinaru, I. H. M. van Stokkum, A. A. Pascal, R. van Grondelle, and
H. van Amerongen, J. Phys. Chem. B 104, 9330 (2000).

(9) D. S. Larsen, M. Vengris, I. H. M. van Stokkum, M. A. van der Horst, R. A.
Cordfunke, K. J. Hellingwerf, and R. van Grondelle, Chem. Phys. Lett. 369,
563 (2003).

(10) D. S. Larsen, M. Vengris, I. H. M. van Stokkum, M. A. van der Horst, F. L.
de Weerd, K. J. Hellingwerf, and R. van Grondelle, Biophys. J. 86, 2538
(2004).

(11) E. Papagiannakis, I. H. M. van Stokkum, M. Vengris, R. J. Cogdell, R. van
Grondelle, and D. S. Larsen, J. Phys. Chem. B 110, 5727 (2006).

(12) W.-M. Kwok, C. Ma, and D. L. Phillips, J. Am. Chem. Soc. 128, 11894
(2006).

(13) L. Valis, Q. Wang, M. Raytchev, I. Buchvarov, H.-A. Wagenknecht, and T.
Fiebig, Proc. Natl. Acad. Sci. 103, 10192 (2006).

58
(14) M. J. Tauber, R. A. Mathies, X. Y. Chen, and S. E. Bradforth, Rev. Sci.
Instrum. 74, 4958 (2003).

(15) K. Ekvall, P. van der Meulen, C. Dhollande, L.-E. Berg, S. Pommeret, R.
Naskrecki, and J.-C. Mialocq, J. Appl. Phys. 87, 2340 (2000).

(16) M. Ziolek, R. Naskrecki, M. Lorenc, J. Karolczak, J. Kubicki, and A.
Maciejewski, Optics Comm. 197, 467 (2001).

(17) M. Lorenc, M. Ziolek, R. Naskrecki, J. Karolczak, J. Kubicki, and A.
Maciejewski, Appl. Phys. B 74, 19 (2002).

(18) P. R. Monson and W. M. McClain, J. Chem. Phys. 53, 29 (1970).

59
Chapter 4. Electronic Relaxation Dynamics of Uracil Derivatives ∗

4.1 Excited State Dynamics of Thymidine
Elucidating the rapid excited state dynamics of DNA bases is crucial for
understanding the subsequent photochemistry of the nucleic acids.
1,2
After excitation
in the UV the nucleobases are known to undergo ultrafast internal conversion. In
particular, for pyrimidines several distinctly different mechanisms for electronic
relaxation pathway have been suggested involving close-lying first ππ* and n π*
states after excitation in the first absorption band at ~260 nm.
3-5
Several experimental
and theoretical studies pointed out the importance of state ordering of these states on
relaxation dynamics.
6
Recently, gas phase photophysics of DNA bases have been
presented using TRPES and resonant ionization.
7,8
Being able to identify the
character of the state Ullrich et al. have reported an ultrafast relaxation pathway from
bright S
2
( ππ*) which decays rapidly (<50 fs) to the S
1
(nπ*) for all pyrimidine bases.
9

This is consistent with the order of the states from ab initio calculations which show
that S
1
(nπ*) is the lowest singlet state for uracils. (Ref.
2
and references therein)
In aqueous solution, the transition energy to the n π* state is expected to shift
to higher energy, while the ππ* state should shift only slightly to lower energies
since, in a protic solvents, ππ* states are more stabilized by dipole-dipole and

∗ A. E. Jailaubekov, D. S. Larsen, C. Rivera, C. Chester, and S. E. Bradforth, J. Am. Chem. Soc.,
(Communication paper) in preparation (2007)

60
hydrogen bonding interactions than n π* states. In contrast to the gas phase, quantum
mechanical calculations for uracils that model solvation predict the ππ* state and the
nπ* state to be nearly degenerate in aqueous solution or even reverse the singlet state
ordering, with the former state slightly lower in energy.
10-13
Therefore, Gustavsson et
al. have suggested that a conical intersection between the lowest ππ* and ground
state is instead responsible for ultrafast electronic relaxation of thymine in solution.
11

This two-state picture completely ignores the involvement of the n π* state in
nonradiative decay. Due to low time resolution and single-wavelength analyses, all
previous time-resolved studies in solution using transient absorption or fluorescence
upconversion techniques were unable to observe the change in electronic character
resulting in incomplete picture of excited state dynamics.
14-18
Here, we present a
dispersed UV-Vis spectral analysis of the excited-state absorption of the pyrimidine
nucleoside, thymidine, in aqueous solution observed by femtosecond pump-probe
spectroscopy with 45 fs time resolution providing the most detailed view of the
excited state dynamics. We directly compare time-resolved photoelectron
spectroscopy (TRPES) in the gas-phase with our dispersed transient absorption data
in aqueous solution.
The femtosecond transient absorption setup (see Postscript for details)
consists of a home-built hollow core fiber UV pulse generator delivering 35 fs pump
pulses at 267 nm and a CaF
2
based UV-vis white-light continuum generator which
provides with broadband (290-700 nm) probe pulses. Transient absorption signals of
the entire probe spectrum were measured simultaneously on a diode array operated

61
on a shot-by-shot basis at 1 kHz. Thymidine solution was prepared in neutral (pH=7)
phosphate buffer with concentration of 1 OD in 100 μm (12 mM). The thickness of
the liquid jet was ~100 μm. The measured cross-correlation width across the whole
bandwidth of the probe pulse was 45 fs fwhm. The experimental data is collected at
magic angle.
Figure 4.1 shows time-resolved transient absorption spectrum of thymidine at
an excitation wavelength 267 nm. A global target analysis of the data with a multi-
exponential trial function required several components to describe the observed
spectrum. Initially we use the simplest three-state sequential kinetic model as our
“first order approximation”. From gas-phase studies using TRPES it is known that
the time evolution of the photoelectron spectrum of excited state thymine can be fit
using three exponential decay times of 50, 490, and 6400 fs.
9
If we accept the gas-
phase model as a starting point, we use these values as initial guess values for a
global fitting procedure. The so determined time decay constants that best fit the
solution data are 80 fs, 750 fs, and >5 ps. The decay associated spectrum for each of
kinetic intermediate is shown in Figure 4.2. The first two states with lifetimes of 80
fs and 750 fs would then be assigned to electronically excited S
2
( ππ*) and S
1
(nπ*),
respectively, consistent with the TRPES assignment of the excited-state dynamics.
We also observe the long-lived component with lifetime of >5ps which was reported
previously in the literature
4,9
and was assigned to a triplet state in gas phase
4
.

62

Figure 4.1 Transient absorption data of thymidine in buffer solution after
excitation at 267 nm with 35 fs pump pulses at room temperature. The two-
dimensional data was analyzed with a global fitting routine (details in the text). (a)
Color contour plot of the dispersed pump-probe data. (b) Transient spectra at
different probe times as a function of wavelength. Experimental data (dots) and
global fit (solid lines).
300 400 500 600 700
-4
-3
-2
-1
0
1
2
3
4

mOD
Wavelength (nm)
50 fs
100 fs
200 fs
500 fs
1 ps
b)
-4
-3
-2
-1
0
1
2
3
4
300 400 500 600 700
0
0.2
0.4
0.6
0.8
1
Wavelength (nm)
Probe Time (ps)

63

Figure 4.2 Decay associated spectra extracted from global fitting for
electronically excited S
2
( ππ*) and S
1
(nπ*) states and the long-lived channel with
time decay constants of 80, 750, and > 5 ps, respectively. The S
2
( ππ*) spectrum
consists of two parts: a stimulated emission part (negative absorbance) from 300 to
370 nm and a broadband excited-state absorption from 370 and extending to visible
wavelengths. Open dots represents the inverted and scaled steady-state fluorescence
spectrum for thymidine reproduced from Ref.
18
. (Insert) The simple sequential
kinetic model, used as a target to fit 2D dataset.
300 400 500 600 700
-10
-5
0
5
S
2
S
1
long-lived
fluorescence
S
1
S
2
80 fs
750 fs
> 5 ps
S
1
S
2
80 fs
750 fs
> 5 ps

S
1
Amplitude (mOD)
Wavelength (nm)
S
2

Fluorescence

64
The spectrum of the S
2
( ππ*) state shows a clear minimum at 331 nm which is
due to stimulated emission contribution to the spectrum. An unambiguous
assignment of the S
2
( ππ*) state is further supported by the comparison of its
spectrum with steady-state fluorescence spectrum in aqueous solution. The peak of
the stimulated emission part (331 nm) is in excellent agreement with that of the
fluorescence spectrum (330 nm).
18
The spectrum tentatively assigned to the S
1
(nπ*)
state consists of two excited-state absorption (ESA) bands, a strong band, which rises
steeply to 300 nm, and weak and broad band in the 500-700 nm range. In contrast to
Ref.
14
, our work reveals that the ESA decay in the visible wavelengths is biphasic
and cannot be described by a single exponential. In fact, both intermediates appear to
have ESA in the visible. To estimate the position of the maximum of ESA for the S
2

state we subtract the steady state emission from S
2
spectrum. As shown on Figure
4.3, we obtain ESA band of S
2
located between 350-700 nm with a peak at 380 nm.
The shape of this band is in excellent agreement with polarized transient absorption
data (Postscript), which proves the validity of the subtraction.
Figure 4.4 shows a direct comparison of the observed spectra originating
from the thymine excited states from gas-phase photoelectron studies and against
those from our work in aqueous solution. Detailed examination of the two ESA
bands of our analysis and their photoelectron spectra shows that they have quite
similar Franck-Condon envelopes and similar spacing between bands (~1 eV). The
gas phase assignment of the excited-state dynamics using a Koopman’s correlation is

65

Figure 4.3 Excited state absorption bands assigned based on comparison to
TRPES as S
2
( ππ*) and S
1
(nπ*) extracted from sequential kinetic model of the
transient absorption data of thymidine in solution after excitation at 267 nm. To
extract the shape of the excited state absorption of S
2
( ππ*) we assume that the
stimulated emission spectrum is static and matches the steady-state fluorescence
spectrum. Therefore, after the subtraction of fluorescence band (dotted line) from the
decay-associated spectra of S
2
, we obtain the ESA band (dash line) from S
2
.
300 500 700
-5
0
5
4.5 4.0 3.5 3.0 2.5 2.0 1.5
ESA
Fluorescence
S
1
Amplitude (mOD)
Wavelength (nm)
S
2

400
Probe photon energy (eV)

66

Figure 4.4 (Left) Comparison of the spectra of the excited states, assigned as
S
2
( ππ*) and S
1
(nπ*) obtained from (a) TRPES data of thymine in gas phase after
excitation at 250 nm (from Ref
9
) and (b) transient absorption data of thymidine in
solution after excitation at 267 nm. The spectra are plotted with identical energy
scales. Dash lines connect the maximums of two bands between gas and solution
phase pictures. (Right) Schematic representation of pump-probe experiments in gas
and solution phases (see text and Figure 4.5 for details).
300 500 700
0
5
4.5 4.0 3.5 3.0 2.5 2.0 1.5
S
1
Amplitude (mOD)
Wavelength (nm)
S
2

400
S
2
( ππ*)
S
1
(n→π*)
D
0
( π
-1
)
D
1
(n
-1
)
0.3 eV
1.3 eV
3.2 eV
4.2 eV
S
2
( ππ*)
S
1
(n→π*)
D
0
( π
-1
)
D
1
(n
-1
)
0.3 eV
1.3 eV
200 nm
(6 eV) probe

67
rather powerful as passage through each electronic surface produces a characteristic
signature in the probe spectrum. The excellent agreement between the ESA and
TRPES bands leads us to speculate that optical absorption in the liquid from S
2
( ππ*)
and S
1
(nπ*) occurs to upper states with the same origin as in gas-phase
photoionization, namely Rydberg upper states with cation cores D
0
( π
-1
) and D
1
(n
-1
).
(see diagram, Figure 4.5). Using the values of transition energies of ESA peaks (3.2
and 4.2 eV) in liquid phase experiments, electron kinetic energies (1.3 and 0.3 eV) in
TRPES peaks, and photon energy of ionization pulse (6.2 eV) in the gas-phase
experiments, it is possible to calculate a solvent shift of ~1.7 eV for both cationic
states D
0
( π
-1
) and D
1
(n
-1
).
The similarity between photoelectron and transient absorption spectra
provides further evidence for our assignment of electronic deactivation mechanism
for the thymine base in both gas-phase and aqueous solution. Based on the gas-phase
assignment then we would conclude that after excitation at 267 nm, S
2
( ππ*)
undergoes an ultrafast (80 fs) internal conversion to S
1
(nπ*), which, in turn,
deactivates within 750 fs. Recently, however, Martinez and co-workers have
suggested a reassignment of the TRPES results of Stolow for thymine.
21
They have
speculated that the changing photoelectron spectrum is not due to a change in
electronic state but rather nuclear wavepacket dynamics entirely on the S
2
( ππ*)
surface. In this picture, the S
1
(nπ*) state need not be visited to explain the TRPES

68

Figure 4.5 Schematic energy level diagram for both experiments. The
corresponding vertical ionization potentials (from ground state thymine) IP
0
= 9.2 eV
and IP
1
= 10.1 eV are known from photoelectron spectroscopy (binding energies) to
D
0
( π
-1
) and D
1
(n
-1
), respectively.
19,20
The pump laser prepares the optically bright
state S
2
. Due to ultrafast internal conversion, this state converts to the state S
1
. The
dashed arrows indicate electron emission.
Probe photon energy
Pump pulse
S
1
(nπ*) S
2
( π π*)
D
0
( π
-1
)
S
0
200 nm
(6.2 eV)
0.3 eV
1.3 eV
3.2 eV
4.2 eV
D
1
(n
-1
)
D
0
solv
D
1
solv
Electron kinetic energy

69
data. Therefore, the assignment of the solution pathway will be changed, although it
remains a clear result of this work the parallel in timescale and spectroscopy of the
transient decays in both phases.
In conclusion, we have reported new and a detailed picture of ultrafast
excited state dynamics of thymidine in solution. In combination with unprecedented
time resolution and broadband probe detection, the demonstrated technique opens up
a new level of information and analysis of excited sate dynamics in nucleobases. The
spectra of all transient species have been identified and compare surprisingly well
with corresponding photoelectron spectra from the same intermediate states in the
gas-phase. The proposed model for electronic deactivation is, thus, analogous to gas-
phase dynamics where there is an intermediate state between the optically bright ππ*
and ground state during electronic relaxation. To gain further insight into the
electronic character of the intermediates in this decay, and thus to identify the
important conical intersections in the dynamics of deactivation, we have initiated
additional experimental studies. It is clear that further theoretical studies are required
to locate the conical intersections which are responsible for fast nonadiabatic
processes in solution.

70
4.2 References for Chapter 4

(1) B. P. Ruzsicska and D. G. E. Lemaire, in CRC Handbook of Organic
Photochemistry and Photobiology, edited by W. M. Horspool and P.-S. Song
(CRC Press:, Boca Raton, FL, 1995), pp. 1289.

(2) C. E. Crespo-Hernandez, B. Cohen, P. M. Hare, and B. Kohler, Chem. Rev.
104, 1977 (2004).

(3) E. Nir, K. Kleinermanns, L. Grace, and M. S. de Vries, J. Phys. Chem. A 105,
5106 (2001).

(4) H. Kang, K. T. Lee, B. Jung, Y. J. Ko, and S. K. Kim, J. Am. Chem. Soc. 124,
12958 (2002).

(5) Y. He, C. Wu, and W. Kong, J. Phys. Chem. A 107, 5145 (2003).

(6) N. Ismail, L. Blancafort, M. Olivucci, B. Kohler, and M. A. Robb, J. Am.
Chem. Soc. 124, 6818 (2002).

(7) S. Ullrich, T. Schultz, M. Z. Zgierski, and A. Stolow, J. Am. Chem. Soc. 126,
2262 (2004).

(8) C. Canuel, M. Mons, F. Piuzzi, B. Tardivel, I. Dimicoli, and M. Elhanine, J.
Chem. Phys. 122, 074316 (2005).

(9) S. Ullrich, T. Schultz, M. Z. Zgierski, and A. Stolow, Phys. Chem. Chem.
Phys. 6, 2796 (2004).

(10) M. K. Shukla and J. Leszcynski, J. Phys. Chem. A 106, 8642 (2002).

(11) T. Gustavsson, A. Banyasz, E. Lazzarotto, D. Markovitsi, G. Scalmani, M. J.
Frisch, V. Barone, and R. Improta, J. Am. Chem. Soc. 128, 607 (2006).

(12) C. M. Marian, F. Schneider, M. Kleinschmidt, and J. Tatchen, Eur. Phys. J. D
20, 357 (2002).

(13) R. Improta and V. Barone, J. Am. Chem. Soc. 126, 14320 (2004).

(14) J.-M. L. Pecourt, J. Peon, and B. Kohler, J. Am. Chem. Soc. 123, 10370
(2001).

71
(15) B. Cohen, C. E. Crespo-Hernandez, and B. Kohler, Faraday Discuss. 127,
137 (2004).

(16) J. Peon and A. H. Zewail, Chem. Phys. Lett. 348, 255 (2001).

(17) T. Gustavsson, A. Sharonov, and D. Markovitsi, Chem. Phys. Lett. 351, 195
(2002).

(18) D. Onidas, D. Markovitsi, S. Marguet, A. Sharonov, and T. Gustavsson, J.
Phys. Chem. B 106, 11367 (2002).

(19) A. Padva, T. J. O'Donnell, and P. R. LeBreton, Chem. Phys. Lett. 41, 278
(1976).

(20) D. Dougherty, K. Wittel, J. Meeks, and S. P. McGlynn, J. Am. Chem. Soc.
98, 3815 (1975).

(21) H. R. Hudock, B. G. Levine, H. Satzger, S. Ullrich, A. Stolow, and T. J.
Martínez, J. Am. Chem. Soc., submitted (2006).

72
4.3 Postscript to Chapter 4: Experimental Methods and Anisotropy
Dynamics ∗
The remainder of this chapter will describe the experiments in more detail.
Additional results will be provided, such as experiments for other uracil derivatives
and new aspects of the relaxation dynamics uncovered with a much higher level of
global analysis. Polarized pump-dispersed-probe experiments give further insight
into the electronic character of the excited state and timescales for the deactivation
pathways.

4.3.A Introduction
In recent years, studies of the photophysics of the nucleobases have paid
more attention to excited state dynamics of uracil derivatives (uracils, see Figure
4.6). This is due to their importance in understanding DNA photodamage. It is
known that UV radiation causes damage to DNA primarily at sites of adjacent
pyrimidines (e.g. T-T and C-C dimers).
1-3
Another reason to study the photophysics
of uracils is the ability to make more detailed corrections to theoretical calculations,
of which there has been an explosion in recent years. In comparison with purines
(adenine, guanine) the uracil derivatives are smaller molecular systems, which
requires less expensive quantum calculations. Also the first absorption band of uracil
consists of only one bright singlet state.

∗ The first half of this chapter was prepared as a Communication paper to the Journal of the American
Chemical Society, intended for general audience and limited in length. The Postscript material serves
as supporting information to this paper.

73
N O
N
O
CH
3
Thymine
N O
N
O
H
H
H
H
H
H
H
1
2
3
4
5
6
Uracil
N O
N
O
CH
3
CH
3
1-Methyl-Thymine
H
H
N O
N
O
CH
3
1-Methyl-Uracil
H H
H
O
H OH
H H
H H
OH
N
N
O
O
CH
3
H
H
Thymidine
(deoxyribonucleoside)
N
O
O N
O
OH OH
H H
H H
OH
Uridine
(ribonucleoside)
H H
H

Figure 4.6 The chemical structures and standard ring numbering for the uracil
derivatives studied.

74
Recently, Martinez and co-workers have suggested a reassignment of the
TRPES results of Stolow for thymine.
4
They have speculated that the changing
photoelectron spectrum is not due to a change in electronic state but rather nuclear
wavepacket dynamics entirely on the S
2
( ππ*) surface. The authors have been
proposed that the fast time component corresponds to relaxation from the Franck-
Condon (FC) region into a well-defined minimum on S
2
( ππ*). In addition, they
propose that picosecond components of 6.4 and 2.4 ps observed in the gas phase are
the times required for electronic relaxation out of S
2
into S
1
, for thymine and uracil
respectively. These results clearly contradict the assignment made in recent TRPES
studies. However, the relaxation mechanism still implicates S
1
(nπ*) as an
intermediate state in deactivation. Thymine and uracil decay at different rates in the
gas phase studies, and this effect has been explained in Martinez’ resent work. It is
interesting to compare the dynamics of uracil and thymine in solution phase.
From solution fluorescence upconversion studies the emission decay times
are different for uracil and thymine.
5
Uracil shows a very short excited state lifetime
~100 fs, while thymine has biexponential decay with time constants of ~200 fs and
630 fs. Based on these excited-state lifetimes and theoretical calculations,
Gustavsson and co-workers have proposed that the relaxation mechanism for uracils
in aqueous solution does not involve S
1
(nπ*) as an intermediate state in deactivation
process. The population of the S
2
( ππ*) state decays directly through the S
2
( ππ*)/S
0

conical intersection. This intersection is characterized by pyramidalization and out of
plane motion of the substituents on the C5 atom. A thorough analysis of the excited-

75
state potential energy surfaces have shown that the energy barrier separating the S
2

( ππ*) minimum from the conical intersection increases going from uracil through
thymine.
In the previous section we made no attempt to reconcile the results of time-
resolved fluorescence studies of thymidine in solution
6
with our transient absorption
(TA) results. The fluorescence decay shows biexponential behavior with one very
fast component (150 fs) and a longer one (720 fs) which is not consistent within our
model, because S
1
(750 fs) should be a dark state. New polarization experiments on
thymidine and uridine have helped us to reconcile these observations.
Finally, we report time-resolved spectra for uracil, thymine and their 1-
methylated derivatives. We performed these studies to support the hypothesis made
in section 4.1 that there is a connection between the ESA band position and the
ionization potentials. It is known that these is a shift between the bare nucleobase
and the nucleoside aqueous phase ionization potentials.
7-9

4.3.B Experimental
Experiments were performed using our femtosecond pump-dispersed-probe
setup, as described in detail in Chapter 3 (Figure 3.1). Briefly, UV pump pulses with
a center wavelength of 267 nm were obtained from the hollow core fiber via four-
wave mixing scheme. The fiber setup was pumped by 800 nm laser pulses from a
Ti:Sapphire regenerative amplifier laser system (Hurricane, Spectra-Physics)
operating at 1 kHz. The UV pump pulses were then compressed down to 35 fs using

76
a prism pair before they hit the sample. The white-light continuum, used for
broadband probing, was generated by focusing a weak 800-nm beam into a slowly
translating CaF
2
crystal. Reflective optics were used to focus the probe beam into the
sample, which minimizes the group velocity dispersion down to 300 fs over the 300-
700 nm probe range. The sample solution was interrogated by the pump and probe
beams in a wire gravity drop jet. This flow system produces films with highly
laminar flow and jet thicknesses from ~6-100 μm. For the experiments presented, the
jet thickness was ~50-100 μm. After the spatial overlap of the beams in the sample,
the white-light probe pulse was dispersed with a spectrograph onto a home-built
diode array detector.
All samples were prepared in neutral (pH 7) 50 mM phosphate buffer
solution. The concentration of the solutions were chosen so that an absorbance of ~ 1
OD is achieved in the 100 μm path length at the pump wavelength. This corresponds
to concentrations of ~ 10-12 mM for different solutes. These high concentrations are
required to obtain good quality signal-to-noise data since the sample is delivered by
gravity drop jet, where thickness is only 50-100 μm. Lower concentrations of 0.5-1
mM are typically used in transient absorption experiments by other groups.
However, long path length cells (~1mm) with CaF
2
windows were used in others
experiments, which significantly reduce the time resolution due to group velocity
mismatch between pump and probe pulses. UV-Vis absorption spectra were taken
before and after laser runs to ensure that no degradation is observed and no
photoproducts are formed during photoexcitation.

77
All samples, thymidine (Thd), thymine (Thy), 1-methyl-thymine (MeThy),
thymidine monophosphate (TMP), uracil (Ura), uridine (Urd), and 1-methyl-uracil
(MeUra) were purchased from Sigma-Aldrich and were used as received from
manufacturer.
The steady state absorption spectra of all uracil derivatives are shown in
Figure 4.7. They have their lowest singlet absorption band peaking between 260-273
nm. The absorption spectra for both the thymine and uracil series show the same
trend: the nucleosides, thymidine and uridine, are peaking at 267 nm and 262 nm,
respectively, and their corresponding bases are blue-shifted by 2-3 nm, while the
methylated bases are red-shifted by ~5 nm.
In the femtosecond experiments, pump pulse intensities were typically 5-10
GW/cm
2
at the sample. The 50-100 nJ pump pulses (~30 fs) were focused into a spot
of 200 μm in diameter. Pump pulse intensities were set to minimize solvated electron
contribution, but at the same time, obtain reasonable signal-to-noise transient data. In
general, the maximum values for transients were 4-5 mOD. In the experiments with
neat buffer solution, at pump intensities of 10 GW/cm
2
the maximum solvated
electron signal formed from two-photon ionization of the solvent was 0.2 mOD at
700 nm. High optical density of the concentrated solutions should, in principle,
decrease the solvated electron signal generated from solvent. The value of this signal
is estimated to be 0.03 mOD in the assumption that the sample jet thickness does not
change.

78

Figure 4.7 Normalized absorption spectra of all uracil and thymine derivatives
studied. (a) thymidine (Thd), thymine (Thy), 1-methyl-thymine (MeThy), and
thymidine monophosphate (TMP), the spectra of TMP and Thd are identical in this
graph; (b) Uracil (Ura), Uridine (Urd), 1-Methyl-Uracil (MeUra); (c) absorption
spectrum shift between Urd and Thd. (d) first absorption band of Thd and the
spectrum of the 35 fs pump pulse.
220 240 260 280 300 320
0.0
0.2
0.4
0.6
0.8
1.0
1.2
Normalized Absorption
Wavelength (nm)
Thd
Thy
MeThy
TMP
220 240 260 280 300 320
0.0
0.2
0.4
0.6
0.8
1.0
1.2
Wavelength (nm)
Urd
Ura
MeUra
(b)
220 240 260 280 300 320
0.0
0.2
0.4
0.6
0.8
1.0
1.2
Urd
Thd
Normalized Absorption
Wavelength (nm)
(c)
220 240 260 280 300 320
0.0
0.2
0.4
0.6
0.8
1.0
1.2
Wavelength (nm)
Thd
pump pulse
(a)
(d)

79
Anisotropy experiments were performed by taking the separately measured
time-resolved spectra, S
PA
( λ, t), S
PE
( λ, t), and S
MA
( λ, t) which correspond to transient
absorption signals with the polarization of the pump pulse set parallel, perpendicular,
and at magic angle respectively, with respect to the polarization of the probe pulse.
At pump intensities of 10 GW/cm
2
the experimental and reconstructed data did not
match, showing lower values in reconstructed data. This is due to saturation effect in
the signal, the saturation is larger in the parallel signal.
10
To avoid this effect, the
pump intensity was lowered by 4-5 times. However, this procedure obviously led to
an increase in the data collection time.
The probe beam diameter at the sample was ~80 μm. This value was obtained
by measuring the total throughput of the white-light continuum through two different
pinholes, 100 and 30 μm diameters. Using the same pinholes, we investigate the
chromatic aberration in the focused white-light continuum. The pinhole was
translated in the longitudinal axis (the axis along the white-light beam) and the
spectrum of the continuum was recorded. Despite the fact that we used two off-axis
parabolic mirrors that should minimize chromatic aberration, we still observed a
pronounced effect. The focusing spots of the blue and red components are separated
by 7-10 mm. We note that such a problem can arise from intrinsic difference in the
divergences of the various colors generated in the CaF
2
disk. We made no attempt to
improve the mode of the 800 nm beam before white-light generation or to optimize
the focal length. These parameters are crucial to achieve a good mode in the
continuum. During the time-resolved experiments the position of the sample was

80
located at the midpoint of the red and blue focus. Despite this problem in chromatic
aberration, the use of a ~3 times larger pump spot size seems to eliminate major
distortions in the transient spectra. For example, in a control experiment with neat
water at high pump intensities we confirmed that the observed spectrum at 10 ps
matches the well-documented spectrum of the solvated electron, which spans from
300 nm to near IR.
All two-dimensional transient absorption data S( λ, t) were dispersion-
corrected and analyzed with a global fitting routine (PDP Fit software from Mikas
Vengris, Vilnius, Lithuania). The full width at half maximum (FWHM) of the
instrument response function was 45 fs, based on fitting of the cross-phase
modulated signals at time-zero. Typically, the simplest sequential kinetic scheme
(three-level system) was sufficient to fit the time-resolved spectrum unless otherwise
indicated. The cross-phase modulation and two-photon absorption spike at zero time
delay were omitted from the fitting procedure by imposing zero weights around time
zero with the width of the IRF window.

4.3.C Dispersed Data (isotropic)
The relative polarization of the pump and probe pulses was set to the magic
angle (54.7 °) for isotropic measurements. Figure 4.8 and Figure 4.9 show the time-
resolved transient absorption spectra of thymine and uracil derivatives, respectively,
in buffer solution recorded at magic angle. There are three main features present

81

Figure 4.8 Time-resolved spectra of Thd, Thy, 1-Me-Thy, and TMP in buffer
solution after excitation at 267 nm with 35 fs pump pulses at room temperature at
different probe times as a function of wavelength. Data is recorded with pump at
magic angle with respect to probe.

300 400 500 600 700
-4
-3
-2
-1
0
1
2
3
4
50 fs
100 fs
200 fs
500 fs
1 ps
5 ps
(mOD)
Wavelength (nm)
Thymidine
300 350 400 450 500 550 600
-2
-1
0
1
2
50 fs
100 fs
200 fs
500 fs
1 ps
5 ps
mOD
Wavelength (nm)
Thymine
300 400 500 600 700
-3
-2
-1
0
1
2
3
4
50 fs
100 fs
200 fs
500 fs
1 ps
5 ps
(mOD)
Wavelength (nm)
1-Me-Thymine
300 400 500 600 700 800
-4
-3
-2
-1
0
1
2
3
4
50 fs
100 fs
200 fs
500 fs
1 ps
5 ps
TMP
(mOD)
Wavelength (nm)

82

Figure 4.9 Time-resolved spectra of Urd, Ura, and 1-Me-Ura in buffer solution
after excitation at 267 nm with 35 fs pump pulses at room temperature at different
probe times as a function of wavelength. Data is recorded with pump at magic angle
with respect to probe.

300 400 500 600 700
-4
-3
-2
-1
0
1
2
50 fs
100 fs
200 fs
500 fs
1 ps
5 ps
mOD
Wavelength (nm)
Uridine
250 300 350 400 450 500 550 600 650 700
-3
-2
-1
0
1
50 fs
100 fs
200 fs
500 fs
1 ps
5 ps
(mOD)
Wavelength (nm)
Uracil
300 400 500 600 700
-4
-3
-2
-1
0
1
2
50 fs
100 fs
200 fs
500 fs
1 ps
5 ps
mOD
Wavelength (nm)
1MeUracil

83
in all transient spectra: a strong absorption band in the UV probe wavelengths (300-
325 nm), stimulated emission signals in the 300-400 nm range with negative
amplitudes, and broad excited state absorption (ESA) band which spans 400 to 700
nm. The shapes of the ESA bands for the various derivatives are different. Only
thymidine and TMP show the same spectral dynamics. This implies that the
electronic structure and photophysics of nucleoside and nucleotide are almost
identical; therefore, the addition of the phosphate group doesn’t change the excited
state dynamics significantly.
The main characteristics of the experimental data are represented in Table
4.1. Our assignment of the stimulated emission is supported by comparison to
steady-state fluorescence maxima. All thymine derivatives show excellent agreement
with fluorescence data. Moreover, our data for stimulated emission maxima is well
correlated with the shift of the steady-state first absorption band. For instance, the
stimulated emission peaks are blue-shifted and red-shifted by 4 and 5 nm for Thy
and 1-Me-Thy, respectively, relative to peak position of the Thd. The same shift is
observed in absorption spectra.
Another connection to the ground-state absorption band can be made with our
transient data. As just stated, the 1-Me-Thy absorption is red-shifted relative to that
of Thd. Therefore, one might expect a larger contribution from the vibrationally hot
ground state of 1-Me-Thy. The spectrum of the vibrationally hot S
0
state has no
absorption in the visible, and is expected to extend out around 300 nm. The red edge

84

Absorption
λ
max
a)
(nm)
Stimulated
Emission
λ
max

(nm)
Steady-state
Fluorescence
λ
max

(nm)
ESA (nm)
λ
max

(nm)
Isosbestic
point (nm)
Thd 268 330 330 -
Thy 265 (-3) 326 (-4) 329 400
1-Me-Thy 273 (+5) 335 (+5) 332 435
TMP 268 331 330 -
Urd 262 316 - 435
Ura 259 (-3) 316 (+0) 312 500
1-Me-Ura 267 (+5) 324 (+8) 315 425

Table 4.1 Characteristic parameters of the ground state absorption, fluorescence
and the time-resolved transient spectra of uracil and thymine derivatives. Peak
wavelengths are reported if bands shows clear peak. Steady-state fluorescence data is
taken from Onidas et al and Gustavsson et al.
5,6
(a) Values in parentheses indicate
the shift of the peak in comparison to the corresponding nucleoside.

85
of the hot ground state near 300 nm was observed previously for adenosine and
cytidine in solution.
11
As it is shown in our data the magnitude of the UV band is the
largest for 1-Me-Thy among the thymine derivatives.
A distinct difference between the two groups of derivatives, uracil and
thymine, is the presence of the isosbestic point in the transient absorption spectra for
uracil derivatives. The isosbestic point is usually the indication of two principal
absorbing components.

4.3.D Dispersed Anisotropy
In order to understand in detail the electronic relaxation dynamics of
thymidine and uridine transient absorption anisotropy measurements were carried out
with dispersed pump-probe technique. Figure 4.10 shows transient absorption spectra
for Thd and Urd with pump-pulse polarization set parallel and perpendicular to that
of the horizontal probe pulse. Both signals S
PA
( λ, t) and S
PE
( λ, t) are highly
polarized. Indeed, the parallel data shows strong signals for stimulated emission,
while the perpendicular data is dominated by positive amplitude signals, i.e. excited
state absorption. Since the 267-nm pump pulse is exciting the lowest ππ* state, the
pump pulse photoselects molecules with the π→π* transition dipole moment aligned
with the polarization of the pump pulse. For stimulated emission, a probe pulse with
same polarization as the pump pulse has the highest probability to stimulate emission
back to the ground state, as the transition dipole is (in absence of vibronic effects)

86

Figure 4.10 Time-resolved spectra, S
PA
( λ, t), S
PE
( λ, t), which correspond to
transient absorption signals with the polarization of the pump pulse set parallel and
perpendicular, respectively, with respect to the polarization of the probe pulse. (a,b)
for Thd, (c,d) Urd. Arrows show the red-shift and blue-shift by the same energy
(0.26 eV) of the stimulated emission peak (from 335 nm to 360 nm) and excited state
absorption peak (from 390 nm to 360 nm), respectively. The thin lines in (a) and (c)
are the steady state fluorescence spectra for Thd and Urd.
300 350 400 450 500 550 600
-6
-4
-2
0
2
4
50 fs
100 fs
200 fs
500 fs
1 ps
5 ps
mOD
Wavelength (nm)
Parallel Thd (a)
300 350 400 450 500 550 600 650
-1
0
1
2
3
50 fs
100 fs
200 fs
500 fs
1 ps
5 ps
mOD
Wavelength (nm)
Perpendicular
Thd
(b)
300 400 500 600
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
Parallel
50 fs
100 fs
200 fs
500 fs
1 ps
mOD
Wavelength (nm)
Urd
(c)
300 400 500 600
-0.4
-0.3
-0.2
-0.1
0.0
0.1
0.2
0.3
0.4
Perpendicular
mOD
Wavelength (nm)
Urd
(d)

87

the same. Therefore, the dynamics of stimulated emission or population dynamics
from S
2
( ππ*) surface can be observed as a function of time. Any lower n π* state is
assumed to yield negligible stimulated emission. In the case of the Thd the initial
(~50 fs) envelope of the emission profile resembles a mirror of the ground state
absorption spectrum. After a few hundred femtoseconds the stimulated emission
signal shifts to the red from 335 nm to 360 nm. Interestingly, that the ESA peak on
the perpendicular-polarized data shifts to the blue. By 0.5 ps the transient spectrum
has two minima. A parallel-polarized spectrum allows us to see the spectral
evolution of the stimulated emission. Moreover, the timescale for the emission is ~1
ps. In comparison, there is no spectral evolution observed in the Urd data, since
emission dynamics is fast and already complete by 200 fs.

4.3.E Ground State Bleach
A complementary set of information can be retrieved from UV pump-UV
probe experiments at 267 nm. In this case if we assume that excited state absorption
is negligible compared with large extinction coefficient of ground state absorption,
the transient absorption largely portrays the bleaching of the ground state population
and the bleach recovery. Figure 4.11 shows the magic angle bleach data and bleach
anisotropy of Thd and Urd. Both nucleosides show only partial recovery to the
ground state. In contrast, purines (adenosine and guanosine) show total bleach

88

Figure 4.11 UV pump-UV probe transient absorption of Thd and Urd at 267nm
(bleaching of the ground state population). (a) Magic angle data, (b) bleach
anisotropy.

0 1 2 3 4 5 10 20 30 40 50
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
Thd
Urd
Magic Angle Bleach
Normalized
(ps)
0.0 0.5 1.0 1.5 10 20 30 40 50
0.0
0.2
0.4
0.6
0.8
1.0
1.2
Thd
Urd
Bleach Anisotropy
r(t)
(ps)

89
recovery (data in Chapter 5). About 80% of Thd is recovered by 10 ps with time
constant of 1.7 ps. Relative to the Thd data, the Urd has only ~30% bleach
recovered, the time constant is estimated to be 1.1 ps. Recall, that the decay time of
the hot ground state should be then consistent with the time constant of ground state
bleach recovery observed in our data probing at 267 nm.

4.3.F Global Analysis: Sequential Kinetic Model
As it has been demonstrated in section 4.1, we use the simplest three-state
sequential kinetic model to fit our transient absorption data for all uracils studied.
The determined three time decay constants that best fit the solution data are 60-90 fs,
0.5-1 ps, and >5 ps. The decay associated spectrum for each of kinetic intermediate
is shown in Figure 4.12. The first two states with lifetimes would then be assigned to
electronically excited S
2
( ππ*) and S
1
(nπ*), respectively, consistent with the TRPES
assignment of the excited-state dynamics. In section 4.1 we have compared the
TRPES data for thymine with the TA data for thymidine. Now we are able to
compare the data only for the bases. Again, detailed examination (Figure 4.13) of the
two ESA bands of our analysis and their photoelectron spectra shows that they have
quite similar Franck-Condon envelopes and similar spacing between bands (~1 eV).
Further, the shape of the ESA band is in excellent agreement with polarized transient
absorption data (Figure 4.14), which proves the validity of the subtraction of
emission band from the decay-associated spectra of S
2
to obtain clean S
2
absorption
band.

90

Figure 4.12 Decay associated spectra extracted from global fitting for Uridine,
uracil, and 1-Methyl-Uracil. Simple three-level sequential kinetic scheme were used.
300 400 500 600 700
-10
-8
-6
-4
-2
0
2
4
Amplitude
Wavelength (nm)
80 fs
500 fs
> 5 ps
Steady-state Fluorescence
Uridine
300 400 500 600 700
-8
-6
-4
-2
0
2
4
Amplitude
Wavelength (nm)
90 fs
800 fs
>5 ps
Steady-state Fluorescence
1-Me-Uracil
300 400 500 600 700
-8
-6
-4
-2
0
2
Amplitude
Wavelength (nm)
60 fs (70 fs)
700 fs (430 fs)
>5 ps
Steady-state Fluorescence
Uracil
250 300 350 400 450 500 550 600
-5
-4
-3
-2
-1
0
1
2
70 fs
1 ps
5 ps
Steady-state Fluorescence
Amplitude
Wavelength (nm)
Thymine
250 300 350 400 450 500 550 600 650 700 750
-8
-7
-6
-5
-4
-3
-2
-1
0
1
2
3
4
5
70 fs
1 ps
5 ps
Steady-state Fluorescence
Amplitude
Wavelength (nm)
1-Me-Thymine
250 300 350 400 450 500 550 600 650 700 750
-8
-7
-6
-5
-4
-3
-2
-1
0
1
2
3
4
5
80 fs
750 fs
>5 ps
Steady-state Fluorescence
Amplitude
Wavelength (nm)
Thymidine

91

Figure 4.13 Comparison of the spectra of the excited states, assigned as S
2
( ππ*)
and S
1
(nπ*) obtained from (top) TRPES data of thymine and uracil in the gas phase
after at 250 nm (from Ref
12
) and (bottom) transient absorption data of thymine and
uracil in solution after excitation at 267 nm. To extract the shape of the excited state
absorption of S
2
( ππ*) we assume that steady state emission matches with dominant
stimulated emission component in S
2
spectrum. Therefore after the subtraction of
emission band from the decay-associated spectra of S
2
thick line (ESA) shows clean
S
2
absorption band.
300 500 700
-8
-6
-4
-2
0
2
4
4.5 4.0 3.5 3.0 2.5 2.0 1.5
ESA(S
2
)
Fluor
S
1
Amplitude (mOD)
Wavelength (nm)
S
2

400
Uracil
Probe photon energy (eV)
300 500 700
-4
-2
0
2
4
4.5 4.0 3.5 3.0 2.5 2.0 1.5
ESA(S
2
)
Fluor
S
1
Amplitude (mOD)
Wavelength (nm)
S
2

400
Thymine
Probe photon energy (eV)
Electron Energy (eV)
0.
0
0.
5
1.
0
1.5 2.0 2.5 3.0
0.0
0.2
0.4
0.6
0.8
1.0
Electron Energy (eV)
0 1 2 3 4
0.0
0.2
0.4
0.6
0.8
1.0

92

Figure 4.14 Comparison of the ESA of S
2
( ππ*) extracted from global fit (solid
lines) with the time-resolved spectrum, S
PE
( λ, t), which correspond to transient
absorption signals with the polarization of the pump pulse set perpendicular with
respect to the polarization of the probe pulse (dots).
300 350 400 450 500 550 600 650
-1
0
1
2
3
50 fs
100 fs
scaled ESA from gobal fit
ESA
mOD
Wavelength (nm)
Perpendicular and ESA
Thd
300 400 500 600
-0.2
-0.1
0.0
0.1
0.2
0.3
100 fs
200 fs
scaled ESA from fit
Perpendicular and ESA
mOD
Wavelength (nm)
Urd

93
4.3.G Ionization potentials
Figure 4.15 shows the ESA spectra for uracil, thymine and their 1-methylated
derivatives. These studies were performed to support the hypothesis that there is a
connection between the ESA band position and the ionization potentials. It is known
that these is a shift between the bare nucleobase and the nucleoside aqueous phase
ionization potentials.
7-9
The gas-solution shifts of Rydberg (cation) states are the net
result of at least two different contributions: electronic polarization and IP difference
between thymine and thymidine. The latter contribution can be estimated from the
reduction of the IP due to glycosidic bond formation. For thymidine IP of the first π
orbital is 0.5 eV less than that of the thymine.
8
Taking into account this reduction
value we can predict the value of gas-solution binding energy shift of 1.2 eV caused
by polarization screening of the liquid environment around the molecular core. From
experimental studies of ionization potentials for thymine clusters, relative to the
neutral thymine monomer, hydration by three water molecules decreases the IP by
0.65 eV.
9
The estimated shift in the range of 1-2 eV is general in the photoemission
of condensed systems.
13

4.3.H Branching Model and Anisotropy Simulations
In previous sections we used the simple sequential kinetic scheme in order to
fit our transient absorption data and compare it with the gas-phase data acquired
from TRPES. Although the gas-phase data agreed with our data initially, upon

94

Figure 4.15 The ESA spectra for uracil, thymine and their 1-methylated
derivatives.

300 500 700
0
2
4
4.5 4.0 3.5 3.0 2.5 2.0 1.5
Ura
Urd
1-Me-Ura
ESA(S
2
)
Amplitude (mOD)
Wavelength (nm)

400
Probe photon energy (eV)
300 500 700
0
2
4
4.5 4.0 3.5 3.0 2.5 2.0 1.5
Thy
Thd
1-MeT
ESA(S
2
)
Amplitude (mOD)
Wavelength (nm)

400
Probe photon energy (eV)

95
acquiring further data it was evident that results from the global analysis using
simple sequential model was not consistent with ground state bleach measurements
(Figure 4.16). Therefore, we propose a new model to reconcile both the dispersed
transient absorption and ground state bleach data. Figure 4.17 shows the decay-
associated spectra and population dynamics extracted from global fitting using the
following three-state branching model for uridine. After excitation into the S
2
(FC)
state the population decays into two separate channels, which are referred to as
channel S
0
(hot) and S
1
, with time constants of 230 fs and 110 fs. The channel S
0
(hot)
decays into the ground state within 1 ps, and the channel S
1
(long-lived channel) has
effectively infinite lifetime (in the fitting program the time constant was set to 100
ps). Note that this long-lived channel cannot be assigned as a triplet state. Figure
4.18 clearly shows the two species have different transient spectra. The dynamics
reflect that after excitation the population decays into the long-lived channel with
~70% yield. This is now in agreement with ground state recovery dynamics.
Since the experimental anisotropic data is available for uridine and thymidine
it provides another test to check the validity of the suggested model. In the case of
thymidine the stimulated emission signal shifts to the red from 335 nm to 360 nm
within a few hundred femtoseconds. Therefore, analysis of the anisotropic data is
rather complicated. We will compare the experimental anisotropic data with the
simulated anisotropy for uridine and following steps explains how it is carried out.
As it was done in the first section of the chapter, the decay-associated spectrum of S
2

96

Figure 4.16 Comparison of the global analysis results with the experimental
ground state bleaching data for uridine. (a) Decay-associated spectra and (b)
population dynamics extracted from global fitting using a simple three-level
sequential kinetic scheme (see insert). Following the sequential model, the S
2
and S
1

channel lifetimes are short (70 fs and 500 fs, respectively) in comparison to the long-
lived channel (> 5ps). This yields 100% of the population to be found in the long-
lived channel after ~2ps. (c) UV pump-UV probe transient absorption of uridine at
267 nm (bleaching of the ground state population). The latter graph shows that only
~30% of the population is recovered to the ground state by 30 ps, and the remaining
~70% of the population is trapped as a long-lived species. The bleaching
experimental data is in contradiction with the population dynamics extracted from
global fit and, therefore, in disagreement with our simple sequential kinetic scheme.
300 400 500 600 700
-10
-8
-6
-4
-2
0
2
4
70 fs
500 fs
70 fs
500 fs
mOD
Wavelength (nm)
(a)
-1 0 1 2 3 4 5
0.0
0.2
0.4
0.6
0.8
1.0
Probe Time (ps)
(b)
0 1 2 3 4 5 10 20 30 40 50
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
Urd
Normalized
(ps)
(c)

97

Figure 4.17 (a) Decay-associated spectra and (b) population dynamics extracted
from global fitting using a three-state branching model (see insert) for uridine. The
dynamics reflects that after excitation the population decays into the long-lived
channel with ~70% yield. This yield ratio was accomplished only by applying the
branching model and varying the decay-time constants, in order to be in agreement
with ground state recovery ratio obtained from the bleaching data.

300 400 500 600 700
-8
-4
0
4
230 fs
1 ps
inf
S
2
S
0
(hot)
230 fs
1 ps
inf
S
2
S
0
(hot)
mOD
Wavelength (nm)
S
2
S
1
S
0
(hot)
(a)
-1 0 123 45
0.0
0.3
0.6
Population
Probe Time (ps)
(b)

98

Figure 4.18 Comparison of the long-lived signals obtained from our transient
absorption data (solid line) and from literature (dashed line).
14,15
The dashed line was
assigned as triplet state spectrum. The overlaying of our data and literature triplet
data clearly shows that our data is not representative of the triplet state.

300 400 500 600 700
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Normalized
Wavelength (nm)
long lived signal
triplet

99
is decomposed into its constituents of ESA and stimulated emission (SE). Figure
4.19 shows the spectra for each constituent. ESA band is further decomposed into the
two bands, from the UV and visible regions (Figure 4.19(b)). This is necessary
because in this case we can assign different anisotropy values for the two extracted
bands (UV and visible). These anisotropy values are found to be the best-fit for
simulated anisotropy: R(SE)=0.4, R(S
0
)=0.4, R(S
1
)=0.2, R(UV)=-0.2, R(VIS)=0.4.
Figure 4.20 shows the excellent fit of the experimental and simulated anisotropies at
different delay times in the 300-700 nm range.
We have now both isotropic and anisotropic data is fitted with our proposed
model. Our results for uridine show ultrafast branching in the initial ππ* state: a
fraction of the excited-state population decays via internal conversion to the ground
state (S
2
( ππ*) → S
0
), the rest of the population decays to the n π* state, which decays
to S
0
on slower time scale (S
2
( ππ*) −
fs
→ S
1
(nπ*) −
ps
→ S
0
).

100

Figure 4.19 (a) Decay associated spectra extracted from global fitting for the
S
2
( ππ*) and S
1
(nπ*) states and the S
0
(hot). The S
2
( ππ*) spectrum consists of two
parts: a stimulated emission part (negative absorbance) from 300 to 370 nm and a
broadband excited-state absorption from 370 and extending to visible wavelengths.
The black line represents the inverted and scaled steady-state fluorescence spectrum
for uridine reproduced from Ref.
6
To extract the shape of the excited state absorption
of S
2
( ππ*) we assume that the stimulated emission spectrum is static and matches the
steady-state fluorescence spectrum. Therefore, after the subtraction of fluorescence
band (black line) from the decay-associated spectra of S
2
, we obtain the ESA band
(blue line) from S
2
. (b) We further split the ESA band into two bands, the
contribution to the ESA from the UV and the visible regions. This is required to fit
the anisotropy data. Anisotropy values assigned to the UV and visible ESA bands are
-0.2 and 0.4, respectively.

300 400 500 600 700
-10
-8
-6
-4
-2
0
2
4
6
mOD
Wavelength (nm)
SE
S0
S1
ESA
(a)
300 400 500 600 700
-10
-8
-6
-4
-2
0
2
4
6
mOD
Wavelength (nm)
SE
S0
S1
ESA
VIS
UV
(b)

101

Figure 4.20 Overlays of the experimental and simulated anisotropies at different
delay times in the 300-700 nm range.
300 350 400 450 500 550 600
-2
0
2
4
Wavelength (nm)
Anisotropy, R(t, λ)
50 fs
300 400 500 600
-2
-1
0
1
Anisotropy, R(t, λ)
Wavelength (nm)
100 fs
300 350 400 450 500
-0.9
0.0
0.9
Anisotropy, R(t, λ)
Wavelength (nm)
150 fs
300 350 400 450 500
-0.5
0.0
0.5
1.0
Anisotropy, R(t, λ)
Wavelength (nm)
200 fs
300 400 500
-0.2
0.0
0.2
0.4
0.6
Anisotropy, R(t, λ)
X Axis Title
500 fs
300 350 400 450 500 550
-0.2
-0.1
0.0
0.1
0.2
0.3
Anisotropy, R(t, λ)
Wavelength (nm)
1 ps
300 350 400 450 500 550
-0.2
0.0
0.2
0.4
Anisotropy, R(t, λ)
Wavelength (nm)
1.4 ps

102
4.3.I References for Postscript to Chapter 4

(1) Kenneth H. Kraemer, Proc. Natl. Acad. Sci. USA 94, 11 (1997).

(2) Y. H. You, P. E. Szabo, and G. P. Pfeifer, Carcinogenesis 21, 2113 (2000).

(3) G. P. Pfeifer, Y. H. You, and A. Besaratinia, Mutat. Res. 571, 19 (2005).

(4) H. R. Hudock, B. G. Levine, H. Satzger, S. Ullrich, A. Stolow, and T. J.
Martínez, J. Am. Chem. Soc., submitted (2006).

(5) T. Gustavsson, A. Banyasz, E. Lazzarotto, D. Markovitsi, G. Scalmani, M. J.
Frisch, V. Barone, and R. Improta, J. Am. Chem. Soc. 128, 607 (2006).

(6) D. Onidas, D. Markovitsi, S. Marguet, A. Sharonov, and T. Gustavsson, J.
Phys. Chem. B 106, 11367 (2002).

(7) H. Fernando, G. A. Papadantonakis, N. S. Kim, and L. P.R., Proc. Natl.
Acad. Sci. USA 95, 5550 (1998).

(8) C. Yu, T. J. O'Donnell, and P. R. LeBreton, J. Phys. Chem. 85, 3851 (1981).

(9) S. K. Kim, W. Lee, and D. R. Herschbach, 100, 7933 (1996).

(10) D. Magde, J. Chem. Phys. 68, 3717 (1978).

(11) J.-M. L. Pecourt, J. Peon, and B. Kohler, J. Am. Chem. Soc. 123, 10370
(2001).

(12) S. Ullrich, T. Schultz, M. Z. Zgierski, and A. Stolow, Phys. Chem. Chem.
Phys. 6, 2796 (2004).

(13) B. Winter, R. Weber, W. Widdra, M. Dittmar, M. Faubel, and I. V. Hertel, J.
Phys. Chem. A 108, 2625 (2004).

(14) P. D. Wood and R. W. Redmond, J. Am. Chem. Soc. 118, 4256 (1996).

(15) I. G. Gut, P. D. Wood, and R. W. Redmond, J. Am. Chem. Soc. 118, 2366
(1996).

103
Chapter 5. Ultrafast Transient Absorption Studies on Adenine Derivatives

5.1 Introduction
Adenine derivatives (Figure 5.1) are the most extensively studied DNA
monomers.
1
The wealth of gas phase and theoretical literature allows a detailed
comparison of the solution phase with the isolated molecule dynamics. There has
been much recent progress in understanding primary electronic relaxation processes
in the isolated nucleobase adenine in the gas phase using time-resolved photoelectron
spectroscopy (TRPES).
2-4
Photoelectron spectroscopy can allow observation both
bright (ππ*) and dark (n π*) electronic excited states as they transiently evolve in the
ultrafast deactivation process. The identification of the electronic character is
important because with this information it is possible to predict the electronic
relaxation pathway. However, the electronic relaxation of adenine derivatives in the
biologically important environment, aqueous solution, is poorly understood. Most of
the time-resolved studies in aqueous solution have only temporal information about
excited-state dynamics, whereas the spectral information is required to associate the
excited-state absorption spectrum to a specific electronic excited state. Moreover, the
decay times of the excited state recorded to date are ~100 fs in solution phase.
1
In
order to obtain the spectral dynamics and properly capture the fastest dynamics, one
should have high time resolution (<50 fs), which has been lacked in previous time-
resolved experiments.

104
N
N
NH
2
N
N
H
H
H
1
2
3
4
5
6
7
8
9
N
N
NH
2
N
N
7H-adenine
H
H
H
N
N
N
N
H
2
N
2-amino-purine (2AP)
9H-adenine
H
H
H
N
N
N
N
NH
2
O
H OH
H H
H
H
2
C
H
OH
Adenosine
AMP
N
N
N
N
NH
2
O
H OH
H H
H
H
2
C
H
O P O
-
O
-
O
1'
2' 3'
4'
5'

Figure 5.1 The molecular structures for 9H-adenine and 7H-adenine (adenine
tautomers), 2-amino-purine (2AP), 2’-deoxy-adenosine (Ado), and 2’-deoxy-
adenosine 5’-monophosphate (AMP).

105
The excited-state dynamics of adenosine
5,6
and monomethylated adenines
7
in
solution at room temperature were studied by Kohler and co-workers using
femtosecond pump-probe spectroscopy. (Adenosine, AMP, and ATP are chosen in
the solution phase studies to avoid the complexity related with the presence of the
two (9H and 7H) tautomers in adenine: by attaching the ribose group at the 9
position, the issue of the 7H tautomer is removed). Transient absorption data at a
limited number of probe wavelengths have shown two spectrally distinct regions of
transient absorption. Excited-state absorption (ESA) by the initially populated ππ*
state of adenosine is characterized by weak and broad absorption band centered near
600 nm, which decays in 290 fs by internal conversion. Following internal
conversion, vibrationally highly excited ground state molecules have been detected
by hot ground state absorption at ultraviolet probe wavelengths. Vibrational cooling
is reported to be ~2 ps at 270 nm. This decay time was assigned to intermolecular
vibrational energy transfer to the solvent. The transients between 270 nm and 340
nm showed complex spectral dynamics: the decay times and signal amplitude
increase as the probe wavelength decreases. A contribution from stimulated
emission, which is expected to also be present in that spectral region, has not been
observed before. A prominent instrument-limited spike at time zero has complicated
the explanation of the excited state dynamics at shorter wavelengths.
The fluorescence studies of DNA nucleosides and nucleotides in aqueous
solution have also been studied by femtosecond fluorescence upconversion.
8-11

Gustavsson and co-workers showed that all fluorescence decays for adenine

106
derivatives are complex and cannot be described by monoexponential functions. The
biexponential fits showed one very fast component ( τ
1
) and a longer one ( τ
2
). In the
case of adenosine and AMP, the τ
1
value was estimated to be 100 fs (instrument-
limited) for both derivatives and the τ
2
value was 0.42 ps and 0.52 ps for adenosine
and AMP, respectively. The time-zero fluorescence anisotropy of adenosine and
AMP was found to be the same with a value around 0.24. This is different from the
0.4 value expected for parallel absorption and emission transition moments. Recall
that the lowest absorption band in the spectrum consists of two overlapping bands
corresponding to different electronic transitions.
12
An excitation at 267 nm populates
the two ππ* (L
a
and L
b
) excited states. The upper L
a
transition is strongly allowed
and has higher oscillator strength than the lower L
b
state. Therefore, biexponential
decay may indicate involvement of both ππ* states.
13
The 0.24 anisotropy value can
be understood assuming that fluorescence takes place only from the lower ππ* state.
Based on the estimated radiative lifetimes for adenosine and AMP, which are quite
similar, and the same time-zero anisotropy values, the authors have concluded that
the phosphate group does not affect the nature of the emitting state. This finding
suggested that electronic relaxation in the two derivatives is the same, however,
more experimental proof is required to clarify this issue.
The comparison of the lifetime values resulting from femtosecond transient
absorption and fluorescence upconversion decays shows different timescales for the
excited-state lifetimes. In addition, only monoexponential lifetimes are given in
transient absorption studies (290 fs), whereas, the fluorescence upconversion data is

107
described by biexponential functions ( τ
1
= 100 fs; τ
2
~ 0.4-0.5 ps). This problem
complicates the understanding of the excited state dynamics. It is also argued that
femtosecond transient absorption and fluorescence upconversion studies show
different features of the excited state dynamics because they detect different
electronic transitions. However, recent work by Phillips and co-workers
14
showed
that it is possible to connect the results from the two studies. By performing
broadband Kerr-gated time-resolved fluorescence (KTRF) and transient absorption
(TA) experiments, they reported the first femtosecond combined time- and
wavelength-resolved study on adenosine in aqueous solution. With the advantage of
a broad spectral probe window, the results from both techniques enable temporal and
spectral characterization of the excited-state relaxation processes. The temporal
dynamics of KTRF spectra shows there are two emitting components with lifetimes
of ~0.13 ps and ~0.45 ps associated with L
a
and L
b
ππ* excited states, respectively.
Global analysis of TA data in the 400-650 nm probe window shows that the temporal
decays can be described by the same two decay time constants. The general
agreement in the spectral dynamics obtained by the TA and KTRF techniques shows
that they are probing the same excited-state species.
Despite the valuable experimental information obtained in the combined
KTRF and TA study, the clear separation of the spectra associated with different
decay times of excited states has not been demonstrated. In addition, the processes
controlling the temporal evolution of the TA spectrum in the 300-400 nm probe
window remain unclear. Again, the decay times increase as the probe wavelength

108
decreases. The wavelength-dependent dynamics in this blue region was explained by
a previously proposed idea that the spectral evolution is due to the combined
contributions of the excited-state and the vibrationally hot ground state absorptions.
The authors argue that compared to the normal ground state absorption which ends at
~290 nm, the hot ground state TA spectra extend further down to 400 nm. However,
no TA spectrum associated with the hot ground state has been presented. In order to
obtain complementary information on the electronic deactivation mechanism, the TA
spectra usually have to be deconvoluted to distinguish between stimulated emission,
excited state absorption, and vibrationally hot ground state.
As mentioned above, recent TRPES have shown convincing results for the
observation of both bright ( ππ*) and dark (n π*) electronic excited states involved in
the ultrafast relaxation. As a result, the gas-phase mechanism for adenine electronic
relaxation involves the lowest n π* state. The relaxation pathway can be viewed as
two sequential steps: S
2
(ππ*) −
τ1
→ S
1
(n π*) −
τ2
→ S
0
, where S
2
is the initially
populated bright excited state. Could this model be applied for electronic relaxation
in aqueous solution? Although the gas-phase studies
2,4,15
show ultrashort excited-
state lifetimes (e.g., in Ref.
4
, τ
1
= 0.07 ps; τ
2
= 1.1 ps), the time-resolved studies in
aqueous solution
5,6,8-11,14
considered above exhibit even shorter lifetimes. Also, there
is no excitation energy dependence in solution phase
11
, in contrast to the gas phase
observations of excitation energy dependent dynamics.
15
This indicates that the
electronic relaxation pathways might be different in aqueous solution, and the
solvent effects are important during excited-state relaxation. For example, it is

109
known that the solvent can reorder the ππ* and n π* states.
1,16-18
The n π* state is less
stabilized in polar solvents than the ππ* state. As a result, in the solution phase the
n π* state may lie higher than the ππ* state. The definitive mechanism responsible for
the electronic relaxation in the solution phase remains an unresolved problem.
Theoretical studies by Sobolewski and co-workers propose that the πσ* state
plays a crucial role in the ultrafast electronic relaxation of adenine both in the gas
and solution phases.
19,20
According to this model, the optically dark repulsive πσ*
state has conical intersections with the initially excited ππ* state and the ground
state, and serves as an intermediate state. The gas phase TRPES studies confirmed
the πσ* state’s participation in the electronic deactivation process in adenine.
2-4

Therefore, adenine in the gas phase has two fast relaxation channels from the ππ*
state, via the n π* and πσ* states. Both channels are excitation energy dependent and
competing with each other in the region of excitation wavelength of 250-277 nm. For
example at 267 nm excitation, the πσ* states plays a major role. We note that
methylation at the N9-H position blocks the πσ* state, making this channel
inaccessible. Indeed, the experimental data showed that 9-methyl adenine has only
one relaxation channel, through the n π* state.
Here we present high-time resolution studies of adenine derivatives in
aqueous solution using dispersed transient absorption spectroscopy. Broad detection
over a wide spectral range and unprecedented time resolution (~ 40fs) allows us to
improve spectral characterization. The TA spectra have been deconvoluted to

110
distinguish between stimulated emission, excited state absorption, and vibrationally
hot ground state. Our results permit the direct comparison between TRPES in the gas
phase and dispersed TA data in aqueous solution. Based on this comparison, we are
able to obtain additional information on the electronic deactivation mechanism. We
propose a mechanism responsible for ultrafast electronic deactivation of adenosine
and AMP in aqueous solution.

5.2 Experimental
The laser system and optical apparatus composing the ultrafast dispersed
pump-probe apparatus have been described in detail previously in Chapter 3. Briefly,
three techniques are used to obtain the experimental data. For isotropic
measurements of the transient absorption spectra, the relative polarization of the
pump and probe pulses was set to the magic angle (54.7 °). Anisotropy measurements
were performed by combining two separately measured transient spectra. In this
case, the polarization of the pump light is set parallel or perpendicular with respect to
the polarization of the probe light. Finally, the single color UV pump-UV probe
isotropic and anisotropic measurements at 267 nm have been used with simultaneous
detection of parallel and perpendicular polarization components of the probe beam.
Unless otherwise stated, the experimental conditions and parameters were
analogous to that described in Chapter 4. All samples, adenine (Ade), 2-amino-
purine (2AP), adenosine (Ado), adenosine-monophosphate (AMP), adenosine-
triphosphate (ATP), guanosine-triphosphate (GTP), cytidine-triphosphate (CTP), and

111
uridine-triphosphate (UTP) were purchased from Sigma-Aldrich and were used as
received from manufacturer. The solutions were prepared in neutral (pH 7) 50 mM
phosphate buffer solution. The concentration of the solutions with adenine
derivatives was 5-7 mM. This is chosen so that an absorbance of ~ 1 OD is achieved
in 100 μm path length, a typical jet thickness in our experiments.
The steady state absorption spectrum measurements were carried out using a
Cary 50 UV/Visible spectrometer. Absorption spectra of all adenine derivatives and
other DNA building blocks are shown in Figure 5.2. Nucleic acid monomers have
their first absorption band maxima in the 250-270 nm range. The adenine base,
nucleoside, nucleotide and triphosphate have nearly identical absorption envelope in
the 240-290 nm range. The lowest singlet absorption band for 2-amino-purine is red-
shifted with its maximum absorption at 307 nm.

5.3 Results
In Figure 5.3, ATP is excited with 35 fs, 267 nm pulses and probed with
near-UV/visible dispersed probe (290-700 nm). The dataset shows two clear regions
of net transient absorption – around 650 nm and between 290 and 400 nm. The
weaker visible band assigned to excited state absorption (ESA) is in good agreement
with ref.
6
. However, the near-UV band is much better spectrally resolved than in
earlier work which had only ~300 fs resolution and was constructed from several
individual probe wavelengths.
6
The near-UV band has been previously assigned to

112
220 240 260 280 300
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
AMP
pump pulse
Normalized Absorption
Wavelength (nm)
a)
b)
220 240 260 280 300 320 340 360
0.0
0.2
0.4
0.6
0.8
1.0
1.2
Ade
2AP
Wavelength (nm)
220 240 260 280 300
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
GTP
CTP
UTP
Normalized Absorption
Wavelength (nm)
c)

Figure 5.2 Normalized absorption spectra of nucleic acid building blocks. (a)
Absorption spectrum of adenosine-monophosphate (AMP) and the spectrum of the
35 fs pump pulse at 267 nm.; (b) Adenine (Ade) and 2-amino-purine (2AP); (c)
Triphosphates: Guanosine-triphosphate (GTP), Cytidine-triphosphate (CTP), and
Uridine-triphosphate (UTP). The dotted vertical shows the central wavelength of
excitation.

113

Figure 5.3 Pump dispersed probe data for ATP in aqueous solution excited at
267 nm. The total transient optical density (mOD) is indicated by the color according
to the colorbar. The blue shifting band between 290-400 nm has been previously
assigned to vibrationally hot S
0
.
6
It is now clear that there is a previously unresolved
stimulated emission contribution at 307 nm (negative amplitude). The ESA band in
the visible (500-700 nm) decays with ~ 220 fs. Note, there is no solvated electron
contribution at long time. (Top) Absorption (solid line) and emission
9
(dash line)
spectrum of ATP in aqueous solution, and excitation wavelength (arrow).
300 400 500 600 700
0.0
0.5
1.0
Abs
Fluorescence
Normalized
Wavelength (nm)
-0.5
0
0.5
1
1.5
2
2.5
3
3.5
300 350 400 450 500 550 600 650 700
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
Wavelength (nm)
Probe Time (ps)

114
absorption of the adenosine molecules returning to the ground state after internal
conversion but still very highly vibrationally excited. These hot molecules (T ~
1000 K) have highly red-shifted absorption spectra and rapid cooling of the
vibrational energy into the water H-bonding network leads to fast blue shifting of the
absorption band back toward the 300 K absorption spectrum. Significantly, Figure
5.3 also shows evidence for a stimulated emission band (negative change in optical
density) around its expected maximum location 307 nm,
9,10
not previously observed
and right in the middle of the blue shifting hot ground state absorption.
Table 5.1 shows characteristic times of the transient absorption decays at
different probe wavelengths. There is a general agreement between our data and
those presented in earlier reports on transient absorption in aqueous solution for the
decay timescales in the visible ESA band.
6,14
The decay times are 290 fs and 230 fs
obtained by Kohler’s and Phillips’ groups, respectively. (Phillips’ group has actually
reported the bi-exponential decay for the visible band with 130 fs and 450 fs time
constants. To compare their results with other groups they have provided the one-
exponential time constant of 230 fs). We have measured decay time of 220 fs for the
visible band (500-700nm). The overestimated values obtained by other groups is
likely due to the solvated electron subtraction procedure, which were performed by
other groups to obtain clear ESA signal. In their study, the contribution of the
solvated electron signal was present, which may lead to overestimated value.
Further, the decay time of the near-UV band from our TA spectra at 290 nm (700 fs)

115

Probe wavelength Decay time Rise time
500-700 nm 220 fs Instantaneous
a)
380 nm 220 fs Instantaneous
a)

320 nm 450 fs 120 fs
290 nm 700 fs 260 fs
267 nm 220 fs
b)
-

Table 5.1 Characteristic times (fs) of the transient absorption decay and rise
times at different probe wavelengths resulting from the monoexponential fits with an
instrument response function of 45 fs.
a)
Limited by the time-resolution of the
system.
b)
Extracted from decomposition of transient absorption and anisotropy
signals of ATP recorded with pump and probe pulses at 267 nm.
21

116
coincides with the time constant of 820 fs at 300 nm in TA study by Phillips and co-
workers. This time constants are approximately 2-3 times shorter than 2 ps time
constant at 270 nm reported by Kohler and co-workers.
Figure 5.4 shows the transient absorption spectra of ATP, adenosine, adenine,
and AMP in buffer solution.* All samples demonstrate the same basic spectral
characteristics: the transient spectra consist of two excited-state absorption (ESA)
bands, a strong band (300-450 nm), and weak band in the 500-700 nm range. Both
ESA bands for the ATP, AMP, and Ado transient spectra decay with approximately
220 fs time constant. However, our high time resolution data reveals that the decay
has the biphasic behavior (Figure 5.5). This means the involvement of an even faster
component (<100 fs) in the excited-state dynamics.
The temporal evolution of the transient absorption data shows a slower decay
in the wavelength range from 400 to 290 nm. The decay time constant is estimated to
be ~700-800 fs at 290 nm. It has been proposed previously
6
that the UV region
spectral evolution is due to contributions from ESA and the red edge of the
vibrationally hot ground state. The ultrafast electronic relaxation process dumps most
of the UV photon energy into a vibrationally hot ground state. This should extend the
ground state absorption spectrum further into the 300-350 nm range. The room

* NOTE: the data for adenine, 2AP and AMP are not at the magic angle data. Due to an
error at the time the data was collected, the magic angle between pump and probe polarizations was
set incorrectly. This can be seen by noticing the lower transient signals in the visible probe
wavelengths (500-700nm) in comparison that of in the near-UV range (300-400 nm). In the typical
magic angle data set, the ratio of UV and visible excited-state absorption band intensities is 4:1 at 50
fs time delay. The transient absorption spectra of ATP and adenosine are properly recorded at magic
angle.

117

Figure 5.4 Transient spectra of ATP, adenosine, adenine, and AMP in buffer
solution after excitation at 267 nm at different probe times.

300 400 500 600 700
0
1
2
3
4
5

50 fs
100 fs
200 fs
400 fs
600 fs
(mOD)
Wavelength (nm)
ATP
300 400 500 600 700
0
1
2
3
4

50 fs
100 fs
200 fs
400 fs
600 fs
(mOD)
Wavelength (nm)
Adenosine
300 400 500 600 700
0
1
2
3
4

50 fs
100 fs
200 fs
400 fs
600 fs
(mOD)
Wavelength (nm)
Adenine
300 400 500 600 700
0
1
2
3
4
5

100 fs
200 fs
300 fs
400 fs
500 fs
(mOD)
Wavelength (nm)
AMP

118

Figure 5.5 Transient absorption data of ATP (similar traces with the same time
constants have been observed for adenosine, and AMP), in buffer solution after
excitation at 267 nm at different probe wavelengths.

-0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
(mOD)
Time delay (ps)
300 nm
350 nm
400 nm
500 nm
680 nm

119

temperature absorption spectrum ends at 290 nm. The time constant at 300 nm is in
good agreement with a cooling time of 0.82 ps from recent transient absorption
studies by Kwok et. al.
14
In general, all three transient spectra show the same shape
and spectral dynamics, which implies that the electronic structure and the
photophysics of these derivatives is determined by the nucleoside group, and the
phosphate groups have no influence on the excited state dynamics.
Figure 5.6 shows the transient absorption spectra of the fluorescent modified
base, 2AP. The dynamics and the transient spectra are dramatically different from
that of the normal adenine derivatives. The ESA is attributed to the absorption from
the lowest S
1
( ππ*) state which has a lifetime of several nanoseconds.
22
A dip in the
ESA in the 350-400 nm range is due to stimulated emission. The negative signals in
the UV region are due to strong bleaching of the ground state, since the maximum of
the ground state absorption is at 310 nm. The ESA dynamics can be separated into
two time scales. In the first timescale the ESA rises, except in the region from 400-
450 nm. The remarkable features of this spectral evolution are two isosbestic points
at 415 nm and 460 nm. The signal of the ESA reaches its maximum at 5 ps, and
slowly decays on the second timescale (5-70 ps). The decay time constants are not
the same across the probe wavelengths. The ESA signal in the visible decays slower
than that in the near UV.

120

Figure 5.6 (a, b)Transient spectra of 2AP in buffer solution after excitation at 267
nm at different probe times for early-time (from 100 fs to 4.5 ps) and long-time
(from 1 –70 ps) dynamics. (c,d) Corresponding contour plots of the dispersed pump-
probe data.

300 400 500 600
0.0
0.5
1.0
1.5
2.0

100 fs
500 fs
1 ps
4.5 ps
(mOD)
Wavelength (nm)
2AP
a)
300 400 500 600
0.0
0.5
1.0
1.5
2.0

1 ps
23 fs
70 ps
(mOD)
Wavelength (nm)
2AP
b)
350 400 450 500 550
-0.1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
350 400 450 500 550
0
20
40
60
80

121

Figure 5.7 shows magic angle UV pump-probe data at 267 nm for the
following triphosphates: ATP, GTP, UTP, and CTP. The data represents primarily
the ground state bleach recovery; it is assumed that ESA is weaker than ground state
absorption at 267 nm, although changes in ESA clearly contribute to the dynamics.
The recovery times for the purine derivatives, ATP and GTP, are much faster than
that for the pyrimidine derivatives, CTP and UTP. In fact, the ATP and GTP fully
recover within 5 ps with time constants of 750 fs and 850 fs, whereas the CTP and
UTP have the residual bleach for at least 30 ps. The bleach recovery data for Ado,
AMP, and ATP are identical (Figure 5.7). The ground state recovery dynamics has
been reported previously by Kohler and co-workers at 250 nm where a recovery time
constant of ~2 ps was determined for AMP.
23
This is not in agreement with our
measurements.

122

Figure 5.7 UV pump-UV probe transient absorption of ATP, GTP, UTP, and
CTP at 267nm (bleaching of the ground state population).
01234510 20 30
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
GTP
CTP
UTP
ATP
Magic Angle Bleach
UTP
CTP
GTP
ATP
(Normalized)
(ps)

123

Figure 5.8 UV pump-UV probe transient absorption of Adenosine at 267nm in
H
2
O (line) and D
2
O (dots).
012345
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
Ado in D
2
O
Ado in H
2
O
Normalized Absorption
ps

124
5.4 Analysis and Discussion

5.4.A Global analysis of transient absorption data
The isotropic transient absorption data are analyzed by global analysis.
24
The
isotropic signal S
iso
( λ, t) is described by a sum of mono-exponential decays. Each
decay channel has a spectrum A
i
( λ) associated with it (or sometimes in the literature,
decay-associated difference spectrum, DADS) and concentration c
i
(t). For a system
with n channels the isotropic two-dimensional signal dependent from delay time t
and wavelength λ is given by:
) ( ) ( ) ( ) , (
1
t IRF t c A t S
n
i
i i iso
⊗
⎥
⎦
⎤
⎢
⎣
⎡
⋅ =
∑
=
λ λ ,
where IRF(t) is the instrument response function and ⊗ stands for convolution. The
concentration of each channel is given by single exponent decay:
i
t
i
e t c
τ /
) (
−
=
where τ
i
is the time constant (lifetime) of the channel i.
In target global analysis one picks a suitable physical model to fit the
dispersed pump-probe data. The transient absorption data presented in this chapter is
fitted using a simple sequential kinetic scheme. As we used in Chapter 4, in a two-
level system, the channel 1 evolves with time-constant τ
1
into channel 2 that
subsequently decays with time constant τ
2
. The corresponding concentrations c
1
and
c
2
for the two channels are given by:

125
1
/
1
) (
τ t
e t c
−
= , and
1
2
1
1
/ / 1
1
2
) (
) (
1 1
− −
− − −
−
−
=
τ τ
τ
τ τ t t
e e
t c
When performing the global fit using PDP Fit software (from Mikas Vengris,
Vilnius, Lithuania), there are several important factors should be taken into
consideration. The values obtained for the time constants using the sequential model
do not guarantee the fit is unique. In fact, there are several combinations of time
constants which give rise to a reasonable fit. Therefore, we impose certain additional
criteria to determine the correct time constants. Since the spectrum A
i
( λ) and the
concentration c
i
(t) are coupled, the search for best-fit time constants eventually leads
to an accurate shape for each spectrum A
i
( λ). The first criterion is to check that the
time constants satisfy the following condition: τ
1
< τ
2
< τ
3
. This will ensure that the
lifetime of the upper state is shorter than the lower. This not always the case, when,
for example, some of the population is trapped on the upper state. However, the
trapping can be better described with a parallel model when dynamics has a clear
physical branching of population.
The next step is to check the spectra (or amplitudes) of each contributing
channel A
i
( λ). The amplitudes should be comparable in magnitude (or no more than
one order of magnitude different) and should be positive everywhere except when
stimulated emission or ground state bleach from a species is known to be
contributing to the channel. Finally, the model is tested for robustness by perturbing
each time constant around its best-fit value. For a stable model the best-fit time
constants should be reproducible.

126
In special cases, the time constants can be fixed if we know additional
information about the dynamics. For example, the time constant for long-lived
species or a triplet state can be fixed to a large value (e.g., nanoseconds). The
instrument response function (IRF) of 45 fs is also constrained. The coherent
artifacts at zero time delay, such as the cross-phase modulation and solvent two-
photon absorption, were omitted from the fitting procedure by imposing zero weights
around time zero within the width of the IRF window.
We begin our global analysis from fitting the transient absorption data for
ATP shown in Figure 5.9. From gas-phase studies using TRPES, the time evolution
of the photoelectron spectrum of excited state adenine is fitted using three
exponential decay times of <50, 750 fs and several nanoseconds for adenine
2,3
, and
more recently for 9-methyl-adenine
25
, decay times of 70 fs and 1.1 ps have been
obtained for the first two decay times. Accepting the gas-phase model, we adopt the
three-state sequential model for the excited-state dynamics of ATP in aqueous
solution also. The time decay constants determined that best fit the solution data are
55 fs, 190 fs, and 610 fs. The decay-associated spectra and sequential scheme are
shown in Figure 5.10 (a). The short-lived ( τ
1
= 55 fs) spectrum (S
2
) consists of two
excited-state absorption (ESA) bands centered at ~ 400 nm and ~650 nm, and a
negative absorption peaking at 305 nm. The latter cannot be due to the ground state
bleaching because the longest wavelength in the absorption band of ATP is 290 nm.
Therefore, we assign this negative absorbance to stimulated emission (SE). The

127

Figure 5.9 Dispersion-adjusted transient absorption data for ATP in buffer
solution after excitation at 267 nm. The two-dimensional data was analyzed with a
global fitting routine. (a) Color contour plot of the dispersed pump-probe data. (b)
Transient spectra at different probe times as a function of wavelength. Experimental
data (dots) and global fit (solid lines).
300 400 500 600 700
0
1
2
3
4

(mOD)
Wavelength (nm)
40 fs
100 fs
200 fs
400 fs
600 fs
1000 fs
b)
-0.5
0
0.5
1
1.5
2
2.5
3
3.5
300 350 400 450 500 550 600 650 700
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
Wavelength (nm)
Probe Time (ps)

128

Figure 5.10 (a) Decay associated spectra for ATP, extracted from global fitting for
the electronically excited S
2
(ππ*) and S
1
states and vibrationally excited ground
state, S
0
*
, with time decay constants of 55, 190, and 610 fs, respectively. The
S
2
(ππ*) spectrum consists of two parts: a stimulated emission part (negative
absorbance) from 290 to 330 nm and a broad excited-state absorption from 330 and
extending to visible wavelengths. (Insert) The simple sequential kinetic model, used
as a target to fit the two-dimensional experimental dataset. (b) To extract the shape
of the ESA of the S
2
(ππ*) state we assume that the stimulated emission spectrum is
static and matches the steady-state fluorescence spectrum.
9
Therefore, after the
subtraction of fluorescence band ( □) from the decay-associated spectra of S
2
, we
obtain the ESA band ( ○) from S
2
.
300 400 500 600 700
-2
-1
0
1
2
3
4
5
S
*
0
S
2
S
1
S
0
(hot)

S
1
S
2
(ππ*)
S
0
*
(hot)
55 fs
190 fs
610 fs
S
1
S
2
(ππ*)
S
0
*
(hot)
55 fs
190 fs
610 fs
S
1
S
2
Amplitude (mOD)
Wavelength (nm)
a)
300 500 700
-2
-1
0
1
2
3
4
5
4.5 4.0 3.5 3.0 2.5 2.0 1.5
S
*
0
Fluorescence
ESA (S
2
)
400
S
1
Amplitude (mOD)
Wavelength (nm)
Probe energy (eV)
b)

129

steady-state fluorescence maximum is at 306 nm for AMP.
9
The second decay-
associated spectrum S
1
( τ
2
= 190 fs) has approximately the same ESA features as the
S
2
spectrum, the strong band in the near UV and weak band in the visible. However,
the strong band is blue-shifted by ~ 40 nm, while the visible ESA band remains
unchanged. The third decay-associated spectrum (S
0
*
, τ
3
= 610 fs) steeply rises to 290
nm and has almost zero amplitude in the visible. The time constant of S
0
*
is close to
the ground state bleach recovery time of 750 fs as seen in Figure 5.7.
The assignment of the demonstrated target analysis is the following. The first
two states are assigned to the electronically excited singlet states S
2
(ππ*) and S
1

(Figure 5.10 (insert)). We also assign the electronic character of the S
2
state to the
ππ* character, because this state is initially populated. We observe stimulated
emission from the S
2
state which is consistent with strongly allowed π- π* transitions.
The possible electronic character of the intermediate S
1
singlet state will be
discussed later. The third state can be assigned to the vibrationally excited ground
state S
0
*
. The form of its spectrum is the red edge of the absorption from a
vibrationally hot ground state. Thus, our three-state sequential scheme results in the
following electronic relaxation pathway: S
0
−
h ν
→ S
2
(ππ*) −
τ1
→ S
1
−
τ2
→ S
0
(hot)
−
τ3
→ S
0
. After excitation at 267 nm, the optically bright S
2
(ππ*) undergoes an
ultrafast ( τ
1
= 55 fs) internal conversion to intermediate S
1
which has a lifetime of τ
2

= 190 fs. The S
1
state relaxes to the vibrationally hot S
0
*
state which, in turn, relaxes

130
within τ
3
= 610 fs. The bottom of the ground state is recovered in ~750 fs as seen in
bleach recovery data.
Figure 5.10 (b) shows our estimation for the position of the maximum of
ESA for the S
2
state. As mentioned before the spectrum of the S
2
state includes
contributions from both ESA and SE. To extract the shape of the clean S
2
ESA we
assume that the stimulated emission spectrum is static and matches the steady-state
fluorescence spectrum.
9
The fluorescence spectrum is then taken with negative sign
and scaled to the minimum in the stimulated emission at 305 nm. The steady-state
fluorescence is subtracted from the S
2
transient spectrum. We obtain the resultant S
2

ESA band located between 320-700 nm with a peak at 385 nm (3.2 eV). The results
are also plotted as a function of probe energy in eV. The separation of the ESA
maxima of the S
2
and S
1
states is ~0.25 eV. The bandwidths of both ESA bands are
0.8 and 1 eV for the S
2
and S
1
states, respectively.

5.4.B Comparison of TRPES with dispersed transient absorption data
In Chapter 4 we have compared time-resolved photoelectron spectroscopy
(TRPES) in the gas-phase with our dispersed transient absorption data in aqueous
solution for the pyrimidine nucleoside, thymidine. The direct comparison of the ESA
of S
2
and S
1
and their photoelectron spectra shows that they have quite similar
bandwidths and similar spacing between bands. Therefore, we proposed that optical
absorption in the liquid from S
2
(ππ*) and S
1
(n π*) occurs to upper Rydberg states
with cation cores D
0
( π
-1
) and D
1
(n
-1
), the same states involved in gas-phase

131
photoionization. A similar comparison is applied here for the adenine derivatives.
Figure 5.11 (a) shows the spectra of the excited states, S
2
(ππ*) and S
1
(n π*) obtained
from TRPES data for 9-methyl-adenine in gas phase after excitation at 267 nm.
25

The reason we have chosen 9-methyl-adenine instead of adenine, is that the 9N-H
stretching coordinate for 9-methyl-adenine is locked by a heavy methyl group. In the
same way the sugar group locks the 9N-H stretching in the ATP molecule. The
dissociative character of the 9N-H stretch is considered to be very important in
adenine because it opens up an additional electronic relaxation pathway via πσ*
state
2,3,19,20,25
.
The electronic relaxation dynamics of 9-methyl-adenine in the gas phase is
proposed to be the following
25
: S
2
(ππ*) −
τ1
→ S
1
(n π*) −
τ2
→ S
0
. The time constants τ
1

and τ
2
are 70 fs and 1.1 ps, respectively. A global fitting of the TRPES data using
these time constants and the suggested model resulted in decay-associated
photoelectron spectra, represented in Figure 5.11 (a). The photoelectron spectra
assigned to the S
2
(ππ*) and S
1
(n π*) states are well separated by approximately 1 eV.
According to Koopmans correlations, the S
2
(ππ*) state ionizes into cationic π-hole
states, D
0
( π
-1
) and D
2
( π
-1
), while the S
1
(n π*) state ionizes into cationic n-hole states,
D
1
(n
-1
) and D
3
(n
-1
). Figure 5.11 (b) shows the Gaussian fits (from Ref.
25
) to the
decay-associated photoelectron spectra of the S
2
( ππ*) state. The two Gaussians are
centered at 8.8 eV and 10.5 eV for D
0
( π
-1
) and D
2
( π
-1
), respectively.

132

Figure 5.11 Comparison of the spectra of the excited states, S
2
(ππ*) and S
1

obtained from (a) TRPES data of 9-methyl-adenine in gas phase after excitation at
267 nm (from Ref.
25
) and (b) transient absorption data of ATP in solution after
excitation at 267 nm. The spectra are plotted with identical energy scales (eV).
(b)
78 9 10 11
0
1
2
3
4
23 4 5
Sum of n π* → D
1
(n
-1
), D
3
(n
-1
)
π π* → D
2
( π
-1
)
ππ* → D
0
( π
-1
)
S
1
Electron Binding Energy (eV)
Probe Photon Energy (eV)
ESA (S
2
)

Amplitude (a.u.)
Probe photon energy
Pump
S 2 ( ππ*)
D
0
( π
-1
)
S 0
7.8 eV
4.6 eV
D 0
solv
Electron Binding
Energy
ESA (S
2
)
3.2 eV
8.8 eV
(a)

133
Since we determined the shape and the amplitude of the ESA of the S
2
state
in the dispersed TA data in aqueous solution, and we know that it has ππ* electronic
character, we compare it with the spectrum of the lowest lying cationic state
(D
0
( π
-1
)) of 9-methyl-adenine in the gas phase. In order to match the two bands and
to plot them on the same scale as a function of the electron binding energy, the
ESA(S
2
) band is shifted by 5.6 eV. The schematic energy level diagram (Figure 5.11
(c)) can explain that the value of the shift consists of two parts: the pump pulse
energy (4.6 eV) and a solvent shift of ~ 1 eV. For comparison, solvent shifts of ~1.7
eV for cationic states D
0
( π
-1
) and D
1
(n
-1
) were estimated in the comparison of the gas
phase TRPES and aqueous phase TA data for thymidine.
Detailed examination of S
2
(ππ*) ESA and its photoelectron spectrum shows
that they have quite similar Franck-Condon envelopes. Recall that the two peaks of
theses spectra have been deliberately matched, however, both spectra show a
minimum in transition probability at 9.8 eV.
The decay-associated spectrum of the S
1
state obtained from dispersed TA is
also shifted by the same 5.6 eV and plotted as a function of the electron binding
energy. Recall a common solvent stabilization of D
0
( π
-1
) and D
1
(n
-1
) could be
employed for thymidine. Here, however, we observe a discrepancy between the
photoelectron and S
1
ESA spectrum. The peaks from the different measurements are
separated by 0.7 eV. The lowest lying cationic state (S
1
(n π*) → D
0
(n
-1
)) has
maximum value at 9.8 eV, whereas shifted ESA (S
1
) peaks at 9.1 eV. Clearly, the
gas-phase spectrum for S
1
, which is assigned to n π* character, and S
1
ESA for liquid

134
phase do not match. Under the assumption that energy gap between the excited state
and cationic state do not change from the gas phase to liquid phase, the S
1
state in
aqueous phase data cannot then be attributed to have n π* electronic character.

5.4.C Electronic relaxation pathway
In Chapter 4 we have considered two basic models for excited state
deactivation. Even though we do not (in fact, should not) expect the pyrimidine and
purine derivatives (e.g. thymidine and adenosine) to have similar excited state
dynamics, the general principles are still valid, therefore, we apply the same
treatment. The first model, S
2
(ππ*) −
τ1
→ S
1
(n π*) −
τ2
→ S
0
, is the gas phase picture
for the electronic relaxation of both nucleobases. It involves the S
1
(n π*) state as a
intermediate state. Both time constants τ
1
and τ
2
represent the switch from one
electronic state to another. In the second model, [ππ*; S
2
(FC)] −
τ1
→ [ ππ*; S
2
(min)]
−
τ2
→ S
0
, the electronic deactivation to the ground state S
0
takes place directly from
the optically bright S
2
state. The decay-associated spectra in this case represent the
ESA of S
2
(FC) in the Franck-Condon region and the ESA from the relaxed S
2
(min).
The fist time constant, τ
1
, corresponds to the time required for the wavepacket to
move from the FC region, and τ
2
is the time for internal conversion to the ground
state. As mentioned before, the lowest UV ground state absorption band of the
adenine derivatives is due to absorption of two energetically close ππ* state.
According to the band shape decomposition of the absorption spectrum of

135
adenosine
12
, the lower state has smaller oscillator strength. At the wavelength of
excitation (267 nm), the extinction coefficients for the lower and upper ππ* states are
~2,500 and ~9,000 M
-1
cm
-1
. Thus, the upper state is predominantly populated during
the excitation. For simplicity, the S
3
state will stand for the upper ππ* state, and S
2

for the lower ππ* state. The last model can be modified as: [ ππ*; S
3
/S
2
(FC)] −
τ1
→
[ ππ*; S
2
(min)] −
τ2
→ S
0
.
The presented relaxation models differ significantly only in the nature of the
intermediate state. If one identifies the electronic character, it is possible then to
predict the proper excited state dynamics. In the previous section, we have compared
the TRPES and dispersed TA spectra, and it has been concluded that the intermediate
state S
1
cannot be assigned to n π* state. In fact, the shape of the decay-associated
spectra of S
1
resembles that of the S
2
. They have approximately the same ESA
features in the visible and a strong band in the near UV. The strong band is blue-
shifted by only by 40 nm (0.25 eV), while the visible ESA band remains unchanged.
This similarity is a possible indication of wavepacket motion on the same potential
energy surface, i.e. the electronic character ππ* remains unchanged and the selection
rules are unchanged as well. This means that the second relaxation model, [ππ*;
S
3
/S
2
(FC)] −
τ1
→ [ ππ*; S
2
(min)] −
τ2
→ S
0
, is more appropriate, and the intermediate
state is, actually, relaxed S
2
(min) state. Since S
2
(min) is an optically bright state, one
should expect emission from this state for a lifetime τ
2
= 190 fs. We should note that
we only see SE from the S
3
/S
2
(FC) or initial state and that the ESA overwhelms any

136
emission from the relaxed S
2
. However, recall that the oscillator strength of the S
2
-S
0

transition (the lower ππ*) is approximately 5 times smaller than the transition to the
upper S
3
state.
12
In addition, Martinez and co-workers have noted that the emission
transition probability can be substantially reduced when a wavepacket moves away
from the FC region of the ππ* surface in thymine and uracil.
26

The 190 fs timescale recorded is consistent with fluorescence upconversion
data for adenosine in aqueous solution
9
, which shows two decay components, a fast
(100 fs) and slow (420 fs) one in the fluorescence decay. Because of the limited IRF
of the fluorescence upconversion experiments, the very rapid decay of the initial
bright state will not be observable. The origin of the 420 fs component is not
immediately clear when comparing to our dataset.

5.5 Summary
Dispersed transient absorption is employed to investigate the detailed
ultrafast excited-state dynamics of adenine derivatives in aqueous solution. The TA
dataset is rich and shows that excited-state dynamics is complex. The spectral
evolution reveals additional intermediate. The data also shows evidence for a
stimulated emission band, which is not previously observed. Therefore, both ESA
and SE are contributing to the TA signals in the near UV. The time constant of hot
ground state S
0
*
is close to the ground state bleach recovery time of 750 fs. This
consistency helps to assign the decay-associated spectrum for the S
0
*
state.
The UV vibrational cooling rate is faster than previously reported in aqueous

137
solution for adenosine. Moreover, vibrational cooling rates for ATP and GTP
(purines) are much faster than those for CTP and UTP (pyrimidines). Both the bleach
recovery and dispersed TA data for Ado, AMP, and ATP are identical. This implies
that the phosphate group doesn’t change the electronic structure much, thus excited-
state dynamics for these systems is the same.
The excited state dynamics is faster in adenosine than in thymidine. The first
two time decay constants determined that best fit the solution data are 55 fs, 190 fs.
In contrast to thymidine data, which shows 80 fs and 750 fs time-constants. Direct
comparison of our data with the TRPES spectra help us to suggest the electronic
relaxation pathway for adenosine in aqueous solution: S
0
−
h ν
→ S
3
(ππ*, FC) −
55 fs
→
S
2
(ππ*, min) −
190 fs
→ S
0
(hot) −
610 fs
→ S
0
. After excitation at 267 nm, the optically
bright S
3
(ππ*) undergoes an ultrafast ( τ
1
= 55 fs) internal conversion to intermediate
S
2
which has a lifetime of τ
2
= 190 fs. The S
2
state relaxes to the vibrationally hot S
0
*

state which, in turn, relaxes within τ
3
= 610 fs. The bottom of the ground state is
recovered in ~750 fs as seen in bleach recovery data. Thus, ultrafast electronic
relaxation of adenosine molecules from the excited ππ* states to the S
0
ground state
is likely to take place via a ππ*/ S
0
conical intersection. Therefore, the proposed
model for electronic deactivation is not analogous to gas-phase dynamics where n π*
state acts as an intermediate state between optically bright ππ* and ground state
during electronic relaxation.
Deactivation mechanism is similar to previously proposed by other

138
groups
11,14
and supported by recent theoretical studies
27-30
and different from gas
phase We suspect that the solvent effect, such as the energetic reordering of vertical
states, is responsible for changing the electronic deactivation in solution. Further, our
data have shown no evidence for branching in the deactivation pathway. Recall, the
latter was observed for uridine. However, involvement of optically dark states πσ*
and n π* is not completely ruled out. Further experimental studies using anisotropy
which is sensitive to transition dipole moment and theoretical work with solvation
models are required to elucidate the electronic deactivation of nucleic acid bases.

139
5.6 References for Chapter 5

(1) C. E. Crespo-Hernandez, B. Cohen, P. M. Hare, and B. Kohler, Chem. Rev.
104, 1977 (2004).

(2) S. Ullrich, T. Schultz, M. Z. Zgierski, and A. Stolow, J. Am. Chem. Soc. 126,
2262 (2004).

(3) S. Ullrich, T. Schultz, M. Z. Zgierski, and A. Stolow, Phys. Chem. Chem.
Phys. 6, 2796 (2004).

(4) H. Satzger, D. Townsend, M. Z. Zgierski, S. Patchkovskii, S. Ullrich, and A.
Stolow, Proc. Natl. Acad. Sci. USA 103, 10196 (2006).

(5) J.-M. Pecourt, J. Peon, and B. Kohler, J. Am. Chem. Soc. 122, 9348 (2000).

(6) J.-M. L. Pecourt, J. Peon, and B. Kohler, J. Am. Chem. Soc. 123, 10370
(2001).

(7) B. Cohen, P. M. Hare, and B. Kohler, J. Am. Chem. Soc. 125, 13594 (2003).

(8) T. Gustavsson, A. Sharonov, D. Onidas, and D. Markovitsi, Chem. Phys.
Lett. 356, 49 (2002).

(9) D. Onidas, D. Markovitsi, S. Marguet, A. Sharonov, and T. Gustavsson, J.
Phys. Chem. B 106, 11367 (2002).

(10) J. Peon and A. H. Zewail, Chem. Phys. Lett. 348, 255 (2001).

(11) T. Pancur, N. K. Schwalb, F. Renth, and F. Temps, Chem. Phys. 313, 199
(2005).

(12) B. Bouvier, T. Gustavsson, D. Markovitsi, and P. Millie, Chem. Phys. 275, 75
(2002).

(13) B. Cohen, C. E. Crespo-Hernandez, and B. Kohler, Faraday Discuss. 127,
137 (2004).

(14) W.-M. Kwok, C. Ma, and D. L. Phillips, J. Am. Chem. Soc. 128, 11894
(2006).

(15) H. Kang, K. T. Lee, B. Jung, Y. J. Ko, and S. K. Kim, J. Am. Chem. Soc. 124,

140
12958 (2002).

(16) A. Holmen, A. Broo, B. Albinsson, and B. Norden, J. Am. Chem. Soc. 119,
12240 (1997).

(17) B. Albinsson and B. Norden, J. Am. Chem. Soc. 115, 223 (1993).

(18) B. Mennucci, A. Toniolo, and J. Tomasi, J. Phys. Chem. A 105, 4749 (2001).

(19) A. L. Sobolewski and W. Domcke, Eur. Phys. J. D 20, 369 (2002).

(20) A. L. Sobolewski, W. Domcke, C. Dedonder-Lardeux, and C. Jouvet, Phys.
Chem. Chem. Phys. 4, 1093 (2002).

(21) D. S. Larsen, A. E. Jailaubekov, and S. E. Bradforth, in preparation (2007).

(22) O. F. A. Larsen, I. H. M. van Stokkum, M.-L. Groot, J. T. M. Kennis, R. van
Grondelle, and H. van Amerongen, Chem. Phys. Lett. 371, 157 (2003).

(23) C. E. Crespo-Hernandez, B. Cohen, and B. Kohler, Nature 436, 1141 (2005).

(24) I. H. van Stokkum, T. Scherer, A. M. Brouwer, and J. W. Verhoeven, J. Phys.
Chem. 98, 852 (1994).

(25) H. Satzger, D. Townsend, and A. Stolow, Chem. Phys. Lett. 430, 144 (2006).

(26) H. R. Hudock, B. G. Levine, H. Satzger, S. Ullrich, A. Stolow, and T. J.
Martínez, J. Am. Chem. Soc., submitted (2006).

(27) S. Perun, A. L. Sobolewski, and W. Domcke, J. Am. Chem. Soc. 127, 6257
(2005).

(28) L. Blancafort, J. Am. Chem. Soc. 128, 210 (2006).

(29) C. M. Marian, J. Chem. Phys. 122, 104314 (2005).

(30) H. Chen and S. Li, J. Phys. Chem. A 109, 8443 (2005).

141
Chapter 6. Future Work

A new UV pump-dispersed probe apparatus was built and employed to study
ultrafast excited-state dynamics of individual DNA bases in solution over a wide
spectral range and with unprecedented time resolution (~ 40fs). In particular, we
have been able to provide a spectral characterization for the excited-state relaxation
dynamics of uracil and adenine derivatives. So far, molecules have been excited to
their bright ππ* electronic state(s) using excitation wavelength of 267 nm. This
results in excess vibrational energy above the origin of the ππ* state. Our results for
uridine and thymidine show ultrafast branching in the initial ππ* state: a fraction of
the excited-state population decays via internal conversion to the ground state
(S
2
( ππ*) → S
0
), the rest of the population decays to the n π* state, which decays to S
0

on slower time scale (S
2
( ππ*) −
fs
→ S
1
(nπ*) −
ps
→ S
0
). For adenosine, the longest
band in the ground state absorption spectrum is, in fact, due to two energetically
close ππ* states. We have predicted the following relaxation pathway S
3
( ππ*) →
S
2
( ππ*) → S
0
. Thus, ultrafast electronic relaxation of adenosine molecules from the
excited ππ* states to the S
0
ground state is likely to take place via a ππ*/ S
0
conical
intersection. Now we recall that for both DNA monomers in the gas phase Stolow
and co-workers predict the same relaxation model, namely, S
2
( ππ*) → S
1
(nπ*) →
S
0
, which involves the S
1
(nπ*) state as a intermediate state.
1
Therefore, we are lead
to the conclusion that there are different deactivation mechanisms for the gas and

142
solution phase dynamics. This provides direct evidence of the solvent effect on the
excited-state potential energy surfaces, especially in the region of conical
intersections.

6.1 Excitation Energy Dependence
To locate the regions of various conical intersections on the molecular
potential energy surfaces is key to understanding the ultrafast excited-state dynamics
of DNA monomers. The location of conical intersections can be gained from the
experimental studies where the excitation energy populating the initial state is tuned.
One of the prominent examples, which reveal the complexity of the excited state
dynamics in these systems, is the gas-phase study on adenine at different pump
wavelengths using time-resolved photoelectron spectroscopy (see Figure 6.1).
2

Excitation of adenine at 277 nm prepares the molecule at the minimum of the ππ*
state. The population decays with 9 ps time constant. The slow relaxation was
interpreted as resulting from the inaccessibility of the conical intersection with the
nπ* state. When the excitation wavelength is switched to 267 nm, the experimental
results show a branching in the initial ππ* state: a fraction (10-55%) of the excited-
state population decays sequentially, going to the n π* state and from there to the
ground state via conical intersections ( τ
1
=70 fs, τ
2
=1.1 ps). A major fraction (45-
90%) decays via the πσ* state directly to the ground state ( τ
1
=40 fs). At 250-nm
excitation, both channels are open for the deactivation of electronic energy, however,

143

Figure 6.1 Schematic diagrams of the electronic relaxation dynamics in adenine
in the gas phase reproduced from Ref.
2
See details in the text.
250 nm
267 nm
277 nm
S
0

n π∗
π π ∗
π σ ∗

144

due to high excess of vibrational energy the molecule decays primarily through n π*
channel.
2
In summary, at 267 nm excitation the πσ* state plays a major role. By
comparing the branching ratio as a function of excitation energy in the gas phase, it
has been possible to locate the ππ*/ πσ* conical intersection.
Femtosecond fluorescence upconversion spectroscopy has been used to study
excited state lifetimes of nucleobase adenine and its ribonucleoside adenosine in
aqueous solution.
3
The excitation energy was tuned in the 245-280 nm range. The
observed biexponential temporal fluorescence data of adenine revealed a fast decay
time between τ
1
=0.34 ps at the shorter excitation wavelength and τ
1
=0.67 ps at the
longer excitation wavelength. A slow decay time τ
2
=8.4 ps turned out to be
independent of the excitation wavelength. The two values were assigned to the 9H-
adenine tautomer ( τ
1
) and 7H-adenine tautomer ( τ
2
). The excited state lifetime of
9H-adenine in aqueous solution is thus sub-picosecond even at the minimum of the
ππ* state. In contrast, the population decays with a 9 ps time constant in the gas
phase. Another surprising feature that was found in this study is that excited state
lifetime for adenosine, which is assumed to have the same electronic structure as 9H-
adenine, is independent of excitation energy ( τ=0.31 ps) unlike 9H-adenine. The
obvious decrease of the excited state lifetime in the latter from 0.67 ps to 0.34 ps as a
function of excitation energy suggests an opening of additional electronic relaxation
channel. However, the lack of the spectral characterization for the excited-state

145
relaxation dynamics precludes making this conclusion as clearly as in the gas phase
TRPES studies. Therefore, I would argue that it is as yet unclear which deactivation
channel plays a major role in solution phase.
Excitation of adenine at longer pump wavelengths, 270-280 nm, is another
interesting avenue for future research. In Chapter 5 we have extracted the decay-
associated spectra of S
3
( ππ*) and S
2
( ππ*) for the 267 nm pump wavelength (see
Figure 6.2). At 267 nm the upper state S
3
( ππ*) is predominantly populated during
the excitation. In our time-resolved data we observe strong excited-state absorption
(ESA) of S
3
( ππ*). Now if one were to carry out an experiment in the 270-280 nm
excitation range, we expect the lower state S
2
( ππ*) to be predominantly populated,
therefore, the amplitude of the ESA of S
3
( ππ*) would be smaller in comparison with
that of S
2
( ππ*).
Further, being able to use the tunability range of our UV hollow-core fiber as
an excitation source in the 220-240 nm we could provide additional information on
excited state dynamics at higher excitation energies. For example it would be
interesting to resolve the disagreement of lifetimes between 9H-adenine and
adenosine in the 240-267 nm excitation range.
Based on the results presented in Chapter 4 and 5, we believe we have
developed a methodology to map the electronic character associated with the specific
excited state as well as the dynamical timescales in the relaxation of nucleobases.
Further, we have shown the value of the direct comparison between TRPES in the

146

Figure 6.2 Transient absorption data of ATP in solution after excitation at 267
nm. The spectra are plotted with identical energy scales (eV).

7 8 9 10 11
0
1
2
3
4
2 345
S
2
( ππ ∗)
Electron Binding Energy (eV)
Probe Photon Energy (eV)
S
3
( ππ ∗)

Amplitude (a.u.)

147
gas-phase and our dispersed transient absorption data in aqueous solution. The
studies suggested here, exploiting excitation energy dependence, would allow us to
obtain further support to our original hypothesis of mapping the electronic character.

148
6.2 References for Chapter 6

(1) S. Ullrich, T. Schultz, M. Z. Zgierski, and A. Stolow, Phys. Chem. Chem.
Phys. 6, 2796 (2004).

(2) H. Satzger, D. Townsend, M. Z. Zgierski, S. Patchkovskii, S. Ullrich, and A.
Stolow, Proc. Natl. Acad. Sci. USA 103, 10196 (2006).

(3) T. Pancur, N. K. Schwalb, F. Renth, and F. Temps, Chem. Phys. 313, 199
(2005).

149
Bibliography

(1) B. Albinsson and B. Norden, J. Am. Chem. Soc. 115, 223 (1993).

(2) N. A. Anderson, C. G. Durfee, M. M. Murnane, H. C. Kapteyn, and R. J.
Sension, Chem. Phys. Lett. 323, 365 (2000).

(3) S. Backus, J. Peatross, Z. Zeek, A. Rundquist, G. Taft, M. M. Murnane, and
H. C. Kapteyn, Opt. Lett. 21, 665 (1996).

(4) M. Bauer, C. Lei, R. Tobey, M. M. Murnane, and H. C. Kapteyn, Surf. Sci.
532, 1159 (2003).

(5) P. Baum, S. Lochbrunner, and E. Riedle, Opt. Lett. 29, 1686 (2004).

(6) M. Ben-Nun and T. J. Martinez, Adv. Chem. Phys. 121, 439 (2002).

(7) L. Blancafort, J. Am. Chem. Soc. 128, 210 (2006).

(8) L. Blancafort and M. A. Robb, J. Phys. Chem. A 108, 10609 (2004).

(9) B. Bouvier, T. Gustavsson, D. Markovitsi, and P. Millie, Chem. Phys. 275, 75
(2002).

(10) A. Broo, J. Phys. Chem. A 102, 526 (1998).

(11) C. R. Cantor and P. R. Schimmel, Biophysical Chemistry. (W.H. Freeman
and Co., San Francisco, 1980).

(12) C. Canuel, M. Mons, F. Piuzzi, B. Tardivel, I. Dimicoli, and M. Elhanine, J.
Chem. Phys. 122, 074316 (2005).

(13) H. Chen and S. Li, J. Phys. Chem. A 109, 8443 (2005).

(14) B. Cohen, C. E. Crespo-Hernandez, and B. Kohler, Faraday Discuss. 127,
137 (2004).

(15) B. Cohen, P. M. Hare, and B. Kohler, J. Am. Chem. Soc. 125, 13594 (2003).

(16) C. E. Crespo-Hernandez, B. Cohen, P. M. Hare, and B. Kohler, Chem. Rev.
104, 1977 (2004).

(17) C. E. Crespo-Hernandez, B. Cohen, and B. Kohler, Nature 436, 1141 (2005).

150

(18) C. E. Crespo-Hernandez and B. Kohler, J. Phys. Chem. B 108, 11182 (2004).

(19) J. Davies, R. Continetti, D. Chandler, and C. Hayden, Phys. Rev. Lett. 84,
5983 (2000).

(20) D. Dougherty, K. Wittel, J. Meeks, and S. P. McGlynn, J. Am. Chem. Soc.
98, 3815 (1975).

(21) C. G. Durfee, S. Backus, H. C. Kapteyn, and M. M. Murnane, Opt. Lett. 24,
697 (1999).

(22) C. G. Durfee, S. Backus, M. M. Murnane, and H. C. Kapteyn, Opt. Lett. 22,
1565 (1997).

(23) C. G. Durfee, L. Misoguti, S. Backus, H. C. Kapteyn, and M. M. Murnane, J.
Opt. Soc. Am. B 19, 822 (2002).

(24) K. Ekvall, P. van der Meulen, C. Dhollande, L.-E. Berg, S. Pommeret, R.
Naskrecki, and J.-C. Mialocq, J. Appl. Phys. 87, 2340 (2000).

(25) H. Fernando, G. A. Papadantonakis, N. S. Kim, and L. P.R., Proc. Natl.
Acad. Sci. USA 95, 5550 (1998).

(26) R. L. Fork, C. V. Shank, C. Hirlimann, R. Yen, and W. J. Tomlinson, Opt.
Lett. 8, 1 (1983).

(27) C. C. Gradinaru, I. H. M. van Stokkum, A. A. Pascal, R. van Grondelle, and
H. van Amerongen, J. Phys. Chem. B 104, 9330 (2000).

(28) T. Gustavsson, A. Banyasz, E. Lazzarotto, D. Markovitsi, G. Scalmani, M. J.
Frisch, V. Barone, and R. Improta, J. Am. Chem. Soc. 128, 607 (2006).

(29) T. Gustavsson, A. Sharonov, and D. Markovitsi, Chem. Phys. Lett. 351, 195
(2002).

(30) T. Gustavsson, A. Sharonov, D. Onidas, and D. Markovitsi, Chem. Phys.
Lett. 356, 49 (2002).

(31) I. G. Gut, P. D. Wood, and R. W. Redmond, J. Am. Chem. Soc. 118, 2366
(1996).

(32) Y. He, C. Wu, and W. Kong, J. Phys. Chem. A 107, 5145 (2003).

151
(33) A. Holmen, A. Broo, B. Albinsson, and B. Norden, J. Am. Chem. Soc. 119,
12240 (1997).

(34) H. R. Hudock, B. G. Levine, H. Satzger, S. Ullrich, A. Stolow, and T. J.
Martínez, J. Am. Chem. Soc., submitted (2006).

(35) R. Improta and V. Barone, J. Am. Chem. Soc. 126, 14320 (2004).

(36) N. Ismail, L. Blancafort, M. Olivucci, B. Kohler, and M. A. Robb, J. Am.
Chem. Soc. 124, 6818 (2002).

(37) A. E. Jailaubekov, Master's thesis, University of Southern California, 2003.

(38) A. E. Jailaubekov and S. E. Bradforth, Appl. Phys. Lett. 87, 021107 (2005).

(39) H. Kang, B. Jung, and S. K. Kim, J. Chem. Phys. 118, 6717 (2003).

(40) H. Kang, K. T. Lee, B. Jung, Y. J. Ko, and S. K. Kim, J. Am. Chem. Soc. 124,
12958 (2002).

(41) S. K. Kim, W. Lee, and D. R. Herschbach, 100, 7933 (1996).

(42) M. Klessinger and J. Michl, Excited States and Photochemistry of Organic
Molecules. (VCH Publishers Inc., New York, 1995).

(43) S. A. Kovalenko, A. L. Dobryakov, J. Ruthmann, and N. P. Ernsting, Phys.
Rev. A 59, 2369 (1999).

(44) S. A. Kovalenko, N. Eilers-Konig, T. A. Senyushkina, and N. P. Ernsting, J.
Phys. Chem. A 105, 4834 (2001).

(45) S. A. Kovalenko, R. Schanz, H. Hennig, and N. P. Ernsting, J. Chem. Phys.
115, 3256 (2001).

(46) I. Z. Kozma, P. Baum, S. Lochbrunner, and E. Riedle, Opt. Express 11, 3110
(2003).

(47) Kenneth H. Kraemer, Proc. Natl. Acad. Sci. USA 94, 11 (1997).

(48) A. Kummrow, M. Wittmann, F. Tschirschwitz, G. Korn, and E. T. J.
Nibbering, Appl. Phys. B 71, 885 (2000).

(49) W.-M. Kwok, C. Ma, and D. L. Phillips, J. Am. Chem. Soc. 128, 11894
(2006).

152

(50) D. S. Larsen, A. E. Jailaubekov, and S. E. Bradforth, in preparation (2007).

(51) D. S. Larsen, M. Vengris, I. H. M. van Stokkum, M. A. van der Horst, R. A.
Cordfunke, K. J. Hellingwerf, and R. van Grondelle, Chem. Phys. Lett. 369,
563 (2003).

(52) D. S. Larsen, M. Vengris, I. H. M. van Stokkum, M. A. van der Horst, F. L.
de Weerd, K. J. Hellingwerf, and R. van Grondelle, Biophys. J. 86, 2538
(2004).

(53) O. F. A. Larsen, I. H. M. van Stokkum, M.-L. Groot, J. T. M. Kennis, R. van
Grondelle, and H. van Amerongen, Chem. Phys. Lett. 371, 157 (2003).

(54) E. C. Lim, J. Phys. Chem. 90, 6770 (1986).

(55) M. Lorenc, M. Ziolek, R. Naskrecki, J. Karolczak, J. Kubicki, and A.
Maciejewski, Appl. Phys. B 74, 19 (2002).

(56) D. Magde, J. Chem. Phys. 68, 3717 (1978).

(57) E. A. J. Marcatili and R. A. Schmeltzer, Bell Syst. Tech. J. 43, 1783 (1964).

(58) C. Marian, D. Nolting, and R. Weinkauf, Phys. Chem. Chem. Phys. 7, 3306
(2005).

(59) C. M. Marian, J. Chem. Phys. 122, 104314 (2005).

(60) C. M. Marian, F. Schneider, M. Kleinschmidt, and J. Tatchen, Eur. Phys. J. D
20, 357 (2002).

(61) D. Markovitsi, A. Sharonov, D. Onidas, and T. Gustavsson, ChemPhysChem
4, 303 (2003).

(62) S. Matsika, J. Phys. Chem. A 108, 7584 (2004).

(63) S. Matsika, J. Phys. Chem. A 109, 7538 (2005).

(64) B. Mennucci, A. Toniolo, and J. Tomasi, J. Phys. Chem. A 105, 7126 (2001).

(65) B. Mennucci, A. Toniolo, and J. Tomasi, J. Phys. Chem. A 105, 4749 (2001).

(66) M. Merchan and L. Serrano-Andres, J. Am. Chem. Soc. 125, 8108 (2003).

153

(67) L. Misoguti, S. Backus, C. G. Durfee, R. Bartels, M. M. Murnane, and H. C.
Kapteyn, in Ultrafast Phenomena XII, edited by T. Elsaesser, S. Mukamel,
M. M. Murnane, and N. F. Scherer (Springer-Verlag, Berlin, 2001), pp. 112.

(68) P. R. Monson and W. M. McClain, J. Chem. Phys. 53, 29 (1970).

(69) A. C. Moskun and S. E. Bradforth, J. Chem. Phys. 119, 4500 (2003).

(70) S. B. Nielsen and T. I. Sølling, ChemPhysChem 6, 1276 (2005).

(71) E. Nir, K. Kleinermanns, L. Grace, and M. S. de Vries, J. Phys. Chem. A 105,
5106 (2001).

(72) E. Nir, M. Muller, L. I. Grace, and M. S. de Vries, Chem. Phys. Lett. 355, 59
(2002).

(73) E. Nir, C. Plutzer, K. Kleinermanns, and M. de Vries, Eur. Phys. J. D 20, 317
(2002).

(74) M. Nisoli, S. Stagira, S. D. Silvestri, O.Svelto, S. Sartania, Z. Cheng, M.
Lenxner, C. Spielmann, and F. Krausz, Appl. Phys. B 65, 189 (1997).

(75) D. Onidas, D. Markovitsi, S. Marguet, A. Sharonov, and T. Gustavsson, J.
Phys. Chem. B 106, 11367 (2002).

(76) A. Padva, T. J. O'Donnell, and P. R. LeBreton, Chem. Phys. Lett. 41, 278
(1976).

(77) S. K. Pal, J. Peon, and A. H. Zewail, Chem. Phys. Lett. 363, 57 (2002).

(78) T. Pancur, N. K. Schwalb, F. Renth, and F. Temps, Chem. Phys. 313, 199
(2005).

(79) E. Papagiannakis, I. H. M. van Stokkum, M. Vengris, R. J. Cogdell, R. van
Grondelle, and D. S. Larsen, J. Phys. Chem. B 110, 5727 (2006).

(80) J.-M. Pecourt, J. Peon, and B. Kohler, J. Am. Chem. Soc. 122, 9348 (2000).

(81) J.-M. L. Pecourt, J. Peon, and B. Kohler, J. Am. Chem. Soc. 123, 10370
(2001).

(82) J. Peon and A. H. Zewail, Chem. Phys. Lett. 348, 255 (2001).

154

(83) S. Perun, A. L. Sobolewski, and W. Domcke, J. Am. Chem. Soc. 127, 6257
(2005).

(84) S. Perun, A. L. Sobolewski, and W. Domcke, Chem. Phys. 313, 107 (2005).

(85) G. P. Pfeifer, Y. H. You, and A. Besaratinia, Mutat. Res. 571, 19 (2005).

(86) M. Raytchev, E. Pandurski, I. Buchvarov, C. Modrakowski, and T. Fiebig, J.
Phys. Chem. A 107, 4592 (2003).

(87) M. A. Robb, F. Bernardi, and M. Olivucci, Pure Appl. Chem. 67, 783 (1995).

(88) B. P. Ruzsicska and D. G. E. Lemaire, in CRC Handbook of Organic
Photochemistry and Photobiology, edited by W. M. Horspool and P.-S. Song
(CRC Press:, Boca Raton, FL, 1995), pp. 1289.

(89) H. Satzger, C. Root, and M. Braun, J. Phys. Chem. A 108, 6265 (2004).

(90) H. Satzger, S. Sporlein, C. Root, J. Wachtveitl, W. Zinth, and P. Gilch,
Chem. Phys. Lett. 372, 216 (2003).

(91) H. Satzger, D. Townsend, and A. Stolow, Chem. Phys. Lett. 430, 144 (2006).

(92) H. Satzger, D. Townsend, M. Z. Zgierski, S. Patchkovskii, S. Ullrich, and A.
Stolow, Proc. Natl. Acad. Sci. USA 103, 10196 (2006).

(93) A. Sharonov, T. Gustavsson, V. Carre, E. Renault, and D. Markovitsi, Chem.
Phys. Lett. 380, 173 (2003).

(94) M. K. Shukla and J. Leszcynski, J. Phys. Chem. A 106, 8642 (2002).

(95) A. L. Sobolewski and W. Domcke, Eur. Phys. J. D 20, 369 (2002).

(96) A. L. Sobolewski, W. Domcke, C. Dedonder-Lardeux, and C. Jouvet, Phys.
Chem. Chem. Phys. 4, 1093 (2002).

(97) A. Stolow, A. Bragg, and D. Neumark, Chem. Rev. 104, 1719 (2004).

(98) M. J. Tauber, R. A. Mathies, X. Y. Chen, and S. E. Bradforth, Rev. Sci.
Instrum. 74, 4958 (2003).

(99) P. Tzankov, T. Fiebig, and I. Buchvarov, Appl. Phys. Lett. 82, 517 (2003).

155
(100) S. Ullrich, T. Schultz, M. Z. Zgierski, and A. Stolow, J. Am. Chem. Soc. 126,
2262 (2004).

(101) S. Ullrich, T. Schultz, M. Z. Zgierski, and A. Stolow, Phys. Chem. Chem.
Phys. 6, 2796 (2004).

(102) L. Valis, Q. Wang, M. Raytchev, I. Buchvarov, H.-A. Wagenknecht, and T.
Fiebig, Proc. Natl. Acad. Sci. 103, 10192 (2006).

(103) I. H. van Stokkum, T. Scherer, A. M. Brouwer, and J. W. Verhoeven, J. Phys.
Chem. 98, 852 (1994).

(104) V. H. Vilchiz, J. A. Kloepfer, A. C. Germaine, V. A. Lenchenkov, and S. E.
Bradforth, J. Phys. Chem. A 105, 1711 (2001).

(105) B. Winter, R. Weber, W. Widdra, M. Dittmar, M. Faubel, and I. V. Hertel, J.
Phys. Chem. A 108, 2625 (2004).

(106) P. D. Wood and R. W. Redmond, J. Am. Chem. Soc. 118, 4256 (1996).

(107) D. R. Yarkony, J. Phys. Chem. A 105, 6277 (2001).

(108) D. R. Yarkony, Rev. Mod. Phys. 68, 985 (1996).

(109) Y. H. You, P. E. Szabo, and G. P. Pfeifer, Carcinogenesis 21, 2113 (2000).

(110) C. Yu, T. J. O'Donnell, and P. R. LeBreton, J. Phys. Chem. 85, 3851 (1981).

(111) M. Z. Zgierski, S. Patchkovskii, T. Fujiwara, and E. C. Lim, J. Phys. Chem. A
109, 9384 (2005).

(112) M. Z. Zgierski, S. Patchkovskii, and E. C. Lim, J. Chem. Phys. 123, 081101
(2005).

(113) A. M. Zheltikov, Phys. Usp. 45, 687 (2002).

(114) M. Ziolek, R. Naskrecki, M. Lorenc, J. Karolczak, J. Kubicki, and A.
Maciejewski, Optics Comm. 197, 467 (2001).

Abstract (if available)

Abstract One of the primary mechanisms of DNA damage occurs following irradiation with high energy ultraviolet (UV) light. Consequently, the study of excited state dynamics of nucleic acid bases upon UV excitation is essential towards understanding and mediating DNA photodamage. Due to the extremely short subpicosecond lifetimes, most time-resolved studies on DNA bases are hampered by the ability to generate and manipulate short UV laser pulses. The development of an ultrashort UV-pulse (~30 fs) source reported here now makes it possible to characterize very rapid dynamics that is simply not observable in previous lower time resolution experiments. Results of experiments combining broadband UV/Visible dispersed probing with simultaneous polarization resolution are presented for isolated free adenine and uracil derivatives in aqueous solution at room temperature. Both spectral and dynamical data is acquired providing the most detailed view of the excited-state dynamics to date.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Charge separation in transition metal and quantum dot systems

PDF

Electronic states and photodissociation dynamics of hydroxyalkyl radicals

PDF

Ultrafast spectroscopic interrogation and simulation of excited states and reactive surfaces

PDF

Quantifying aspects of DNA damage

PDF

Capturing the sun: leveraging excited state dynamics

PDF

Sensing sequence-specific DNA micro-environment with nucleotide-independent nitroxides

PDF

Following redox chemistry and excited state dynamics in solution using liquid jet photoelectron spectroscopy

PDF

Unraveling photodissociation pathways in pyruvic acid and the role of methylhydroxycarbene

PDF

Spectroscopy of hydrogen and water clusters in helium droplets

PDF

Two-photon absorption spectroscopy and excited state photochemistry of small molecules

PDF

Singlet fission in covalently-linked tetracenes

PDF

An investigation of morphology and transport in amorphous solid water via guest-host interactions

PDF

Photodissociation dynamics of atmospherically relevant small molecules in molecular beams

PDF

Dissociation energy and dynamics of HCl and aromatic-water clusters

PDF

A multifaceted investigation of tris(2-phenylpyridine)iridium

PDF

Liquid phase photoelectron spectroscopy as a tool to probe excess electrons in liquid ammonia along with Birch chemistry carbanions: from dilute electrolytic solutions to the metallic regime

PDF

Temperature-dependent photoionization and electron pairing in metal nanoclusters

PDF

Surface plasmon resonant enhancement of photocatalysis

PDF

The effect of microhydration on ionization energy and proton transfer in nucleobases: analysis and method development

$Study of rotation and phase separation in ³He, ⁴He, and mixed ³He/⁴He droplets by X-ray diffraction$

PDF

Study of rotation and phase separation in ³He, ⁴He, and mixed ³He/⁴He droplets by X-ray diffraction

Asset Metadata

Creator Jailaubekov, Askat (author)

Core Title Ultrafast electronic deactivation of DNA bases in aqueous solution

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Chemistry

Publication Date 03/02/2009

Defense Date 01/17/2007

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag DNA photochemistry,excited-state dynamics,femtosecond,laser spectroscopy,OAI-PMH Harvest,ultrafast

Language English

Advisor Bradforth, Stephen E. (committee chair), Kresin, Vitaly V. (committee member), Reisler, Hannah (committee member)

Creator Email askat@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m297

Unique identifier UC1459405

Identifier etd-Jailaubekov-20070302 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-206447 (legacy record id),usctheses-m297 (legacy record id)

Legacy Identifier etd-Jailaubekov-20070302.pdf

Dmrecord 206447

Document Type Dissertation

Rights Jailaubekov, Askat

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu