Charge separation in transition metal and quantum dot systems

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

CHARGE SEPARATION IN TRANSITION METAL AND QUANTUM DOT
SYSTEMS

by

Diana Masayo Suffern

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

May 2011

Copyright 2010 Diana Masayo Suffern

ii

Dedication

This dissertation is dedicated to my mother

Marsha Tama Suffern

iii

Acknowledgements

To my Advisor, Stephen Bradforth, I thank you many times, for years you led me in my
research and taught me to be a better scientist. You are a great mentor.

To Laura Edwards and Jessica Quinn, we will always be the PChem Girls!

To Everyone in the Bradforth Group, past and present, whom I have worked with, Amy
Moskun, Xiyi Chen, Jerainne Johnson, Askat Jailaubekov, Piotr Pieniazek, Delmar
Larsen, Chris Elles, Chris Rivera, Christi Schroeder, Tom Zhang, Andrew Horning,
Rupesh Parbhoo, Saptaparna Das, Anirban Roy, and Sean Roberts, thank you for all your
help. It’s been great working together. Victor Lenchenkov, thank you for your work on
the copper project.

To Chris Rivera, I’m happy we’ve managed to remain friends! Thank you for
everything. You’ve always been there for me. (I’m sorry about the blanket incident…)

To Professors and Staff of the Chemistry Department, thank you for your support and
guidance.

iv
To Graduate Students, Post-docs, and Researchers in the Chemistry and Other
Departments at USC whom I’ve had the chance to befriend or work with, I cannot list
everyone, but I will miss being among all of you.

To the Jay Nadeau Group at McGill University, Jay Nadeau, Sam Clarke, Lina Carlini,
and Dan Cooper, it’s been great working with you on the quantum dots project.

To Friends near and far, it would be a lonely life without you.

To Ares, you are just a little Dachshund, but you have the biggest heart and personality,
and are always excited to see me when I got home. You’re my baby. Mina, you are a
crazy cat that I inherited through marriage, but you somehow have grown on me.

To All my Family, I love you and thank you for everything.

To my Husband, Andrew Warren, I could never wish upon you all that you do for me. It
is so much. I love you with all my heart.

v

Table of Contents

Dedication ii

Acknowledgements iii

List of Tables viii

List of Figures ix

Abstract xxi

Chapter 1. Introduction 1

1.1 Charge Separation in Molecular and Nanoparticle 1
Systems
1.2 References for Chapter 1 14

Chapter 2. Photoinduced Excited State Dynamics of
Tribromocuprate(I) Anion and Transient Spectral
Evolution 17

2.1 Abstract 17
2.2 Introduction 18
2.3 Experimental 23
2.3.1 Ultrafast Pump-Probe Spectroscopy
2.3.2 Solution Preparation and Sample Delivery
2.4 Results 28
2.4.1 CuBr
3
2-
Ground State
2.4.2 Broadband Transient Spectra
2.4.3 Transient Band 1 Near 350 – 400 nm
2.4.4 UV Band 1 Absorption Quenching
2.4.5 Transient Visible Band 2
2.4.6 Low Intensity Experiments
2.5 Discussion 41
2.5.1 Reported Transients on the Nanosecond Timescale
2.5.2 UV Band 1 Spectral Evolution
2.5.3 Kinetic Model
2.5.4 Global Analysis
2.5.5 Nature of the CuBr
3
2-
CTTS State
2.6 Conclusions 66
2.7 References for Chapter 2 68
vi
2A Supplement 2A 73
2B Supplement 2B 75

Chapter 3. Oscillatory Transient Absorption of Photoexcited
Tribromocuprate(I) Anion: Observation of Vibrational
Wavepackets 78

3.1 Abstract 78
3.2 Introduction 79
3.2.1 Wavepackets
3.2.2 Copper Bromide Complexes
3.2.3 Structure and Geometry of Copper Complexes
3.3 Experimental 89
3.4 Results 89
3.4.1 Oscillatory Transient Signal
3.4.2 Discrete Fourier Transform
3.5 Discussion 94
3.5.1 Vibrational Wavepacket Motion
3.5.2 Assignments
3.6 Conclusions 105
3.7 References for Chapter 3 106

Chapter 4. Luminescence Lifetimes of CuBr
3
2-
Measured by Time-
Correlated-Single-Photon-Counting 109

4.1 Abstract 109
4.2 Introduction 110
4.3 Experimental 113
4.4 Results 116
4.5 Discussion 122
4.6 Conclusions 126
4.7 References for Chapter 4 127

Chapter 5. The Effects of β-Mercaptoethanol on the Photoluminescence
Lifetimes of CdSe/ZnS, CdSe, and CdTe Quantum Dots 129

5.1 Abstract 129
5.2 Introduction 130
5.2.1 Quantum Dot Background
5.2.2 Photophysics
5.2.3 Surface
5.2.4 Environment
5.3 Experimental 140
5.4 Results 143
5.5 Discussion 146
vii
5.5.1 CdSe/ZnS and CdSe
5.5.2 CdTe
5.5.3 Mechanism
5.6 Conclusions 151
5.7 References for Chapter 5 153

Chapter 6. Conclusions and Future Work 156

6.1 Tribromocuprate(I) Anion, CuBr
3
2-
156
6.2 CdSe/ZnS, CdSe, and CdTe Quantum Dots 160
6.3 References for Chapter 6 163

Bibliography 165

Appendix Mathcad Simulation for CuBr
3
2-
Global Analysis 175

viii

List of Tables

Table 2.5.1 Approximate timescales of dynamics observed in the 55
broadband data labeled in Figures 2.5.2, 2.5.4, and
2.5.5. R = rise, D = decay. Each of the time segments
are labeled as τ1, τ2, τ3, τ4.

Table 3.2.1 Literature values for experimental and calculated 84
bending and stretching normal modes of Cu(I) and
Cu(II) dibromocuprate and Cu(I) tribromocuprate
complexes. G: Gas phase, S: Solution Phase, C:
Crystal.

Table 3.4.1 Time of oscillation peaks following 2PA in 92
femtoseconds for probe wavelengths 301, 350, 400,
500, and 550 nm.

Table 3.4.2 Difference between oscillation peaks listed in Table 92
3.4.1 in femtoseconds. Δ
1,2
is the difference between
Peak 1 and Peak 2, and so forth. Conversions of the
peak to peak time differences to frequency domain (in
wavenumbers) are in parenthesis.

Table 4.2.1 Luminescence lifetimes of 266 nm excited CuBr
3
2ˉ
112
under various ligand concentrations, pH, and ionic
strength conditions. The first 5 rows are taken from
Reference 9 and last 4 rows from Reference 10.
Decays are reported as single exponential and detection
was at 480 nm. (Bold is comparison to Figure 4.4.1
results).

ix

List of Figures

Figure 1.1 Schematic of a solar cell mechanism. 4

Figure 1.2 Basic structure of an OLED. 6

Figure 1.3 A photon is used to produce an electron (blue dot) – 8
hole (red dot) pair. In competition with recombination
of the electron and hole, transfer can occur to a
molecule in a surrounding environment. O represents
an electron acceptor, or oxidizing agent, and R
represent a hole acceptor, or reducing agent.

Figure 1.4 Simple schematic of an atomic or molecular anion that 10
undergoes charge-transfer-to-solvent (CTTS)
photodetachment. The blue figures represent water. P
represents an atom or molecule. a) The electron
occupies a diffuse space defined by the solvent cavity.
This is known as the CTTS state. b) One detachment
mechanism that can occur is diffusion of the electron
away from its parent atom or molecule. This type of
CTTS detachment is a good model for detachment from
singly charged halide anions.

Figure 1.5 A doubly charged anion is photoexcited to produce 12
ground state singly charged species resembling a
reaction intermediate in an I
d
ligand exchange
mechanism.

Figure 1.6 Fluorescence imaging of live cells. CdSe/ZnS quantum 13
dots conjugated to dopamine are within these cells via
uptake by the dopamine receptor transport in the cell
membrane. The QDs fluoresce brighter in regions of
high oxidation potential. Taken from Reference 1 with
permission.

x
Figure 2.2.1 Scheme 1. Recreation of the proposed kinetic scheme 21
published in Reference 2. Scheme applies to aqueous
halocuprates(I), X = Cl
−
, Br
−
, I
−
. For the Br
−
system at
[Br
−
] = 5 M, the ground state equilibrium constant is
9.45 and the triplet equilibrium constant is 73, both in
favor of the tribromo- species.

Figure 2.3.1 1 kHz repetition rate laser system and experimental 25
layout.

Figure 2.3.2 250 kHz repetition rate laser system and experimental 27
layout.

Figure 2.4.1 UV/visible absorption spectra of 0.004 M ground state 30
CuBr
3
2−
in water with a small contribution from CuBr
2
−
(see text). NaBr concentration in solution is 0.4 M.

Figure 2.4.2 a) Two dimensional contour plot of transient broadband 31
absorption following 266 nm excitation of aqueous
CuBr
3
2-
. Colors represent magnitude of optical density
in mOD. b) Broadband probe wavelength cuts of
aqueous CuBr
3
2-
from 0.5 ps to 402 ps.

Figure 2.4.3 266 nm excited aqueous CuBr
3
2-
. Transient broadband 32
absorption spectral cuts of 295 nm to 600 nm measured
with the 1 kHz laser system. a) 84 – 174 fs b) Colored
traces are 204 fs – 2 ps. The 204 fs delay absorption
spectra breaks the isosbestic point. The black traces are
those plotted in (a).

Figure 2.4.4 266 nm excited transient absorption of CuBr
3
2−
for 33
single wavelengths as a function of delay time.
Wavelengths are 320 nm to 550 nm. a) 0 to 250 fs.
The initial drop across all wavelengths is part of the
2PA signal. b) 0 to 2 ps. c) 0 to 500 ps.

xi
Figure 2.4.5 a,b) Spectral cuts at ~ 200 ps (black traces) and ~ 500 35
ps (red traces) with different quenchers. Data
normalized using absorption strength at early delays
following 2PA. The discrepancy in the 290 spectral
region between the unquenched data sets in (a) and (b)
could result from the data not taken back to back,
thereby introducing slight experimental variations. a)
CuBr
3
2−
with 0.2 M H
+
. b) CuBr
3
2−
with 0.4 M NO
3
-
.
c,d) Spectral cuts with no quenchers (not normalized)
taken under same conditions as acidic (c) and nitrate (d)
solutions.

Figure 2.4.6 a) Spectral cuts of 266 nm excited aqueous CuBr
3
2−
37
with HBr at [H
+
] = 0.2. b) Spectral cuts of transient
absorption for 266 nm excited neat water. c)
Broadband transient absorption of 267 nm excitation of
0.4 M Br
−
aqueous solution from pre-time = 0 to 300
ps. There is a prominent band centered at 360 nm that
grows in on a ~ 100 ps timescale. This peak is similar
to the absorption spectrum of Br
2
●−
in aqueous solution
and the growing in time constant corresponds to the
bimolecular Br
●
+ Br
−
reaction rate to form Br
2
●−
. d)
Kinetic trances for the Br
−
solution and neat water. The
600 nm trace for neat water and Br
−
solution have the
same dynamics, as also seen at early times (inset). The
360 nm kinetic traces show a rise for the Br
−
solution
while water is level on this mOD scale.

Figure 2.4.7 Two color pump-probe 266 nm excitation, 500 nm and 39
900 nm probes taken on 250 kHz repetition rate laser
system. Black trace: 900 nm detection of neat water.
Blue trace: 900 nm of aqueous CuBr
3
2−
. Green trace:
266 nm/500 nm of CuBr
3
2−
for comparison. No signal
was detected from 0.4 M Br
−
solution (not shown). (a)
0-5 ps. (b) 0-500 ps.

Figure 2.5.1 Transient absorption results from nanosecond laser 42
flash photolysis of aqueous 0.1 mM CuBr
3
2−
in 5 M
[Br
−
] at 50 ns delay time. Excitation with 7 ns 266 nm
laser pulse.

xii
Figure 2.5.2 Early isosbestic point as shown in Figure 2.4.3a above. 47
There must be two species contributing; they are
labeled i and ii. i is assigned to the CuBr
3
2−
CTTS
state. The most likely assignment for species ii is the
dibromo- product of Br
−
dissociation.

Figure 2.5.3 Reproduced CuBr
3
2−
triplet spectrum from Stevenson, 51
et. al.
3
obtained from 266 nm laser flash photolysis of
CuBr
3
2−
in aqueous 5.0 M [Br
−
]. Spectrum was
deconvoluted from absorption of other transient species
at 50 ns, including Br
2
●−
.

Figure 2.5.4 Spectral cuts as shown in Figure 2.4.3b above. Mid 52
time species labeled as iii is the triplet CuBr
3
2−
state.

Figure 2.5.5 Spectra l cuts from 2 ps to 472 ps of same data set 54
shown in Figures 2.5.2 and 2.5.4. The spectral shifting
and rise on the blue side is indicative of one species
decaying while directly populating another species with
a higher extinction coefficient.

Figure 2.5.6 Scheme 2.2. Simplest kinetic scheme describing the 56
UV band dynamics as presented in Section 2.5.3. It
does not include any solvated electron. Not plotted on
an energy scale. i is tribromo- *[
1
CuBr
3
2−
], ii is
dibromo- *[
1
CuBr
2
−
], iii is tribromo- *[
3
CuBr
3
2−
], and
iv is dibromo- *[
3
CuBr
2
−
]. The early isosbestic point is
associated with τ1, the ISC to the tribromo triplet is τ2,
and τ3 is the dibromo CTTS ISC from singlet to triplet.

Figure 2.5.7 a) Spectra for each transient species, i: CuBr
3
2−
CTTS 60
(blue), iii: CuBr
3
2−
triplet (red), ii: CuBr
2−
CTTS
(green), iv: CuBr
2−
CTTS triplet (magenta). Smaller
blue trace is the additional compartment needed for best
global analysis results.

Figure 2.5.8 Concentration profiles for each of the transient species 61
obtained by global analysis. a) From t = 0 to 5 ps. b)
From t = 0 to 500 ps.

Figure 2.5.9 a) Absorption spectra of Br
●
(Ref
4
), Br
2
●−
(Ref
5
), and 61
the solvated electron. b) Fits overlaid with the Br
−

spectral data using the known solvated electron
spectrum and kinetics, and the spectra of Br
●
and Br
2
●−
.
The formation time of Br
2
●−
from the fit is 150 ps
-1
.
xiii

Figure 2.5.10 Global analysis results and spectral cuts of 62
experimental data. The global analysis curves are
offset (below) that of the actual data to show the
features; without translation the global analysis
overlays the data. a) 74 – 154 fs. b) 1002 fs – 1804 fs.
c) 102 ps – 442 ps.

Figure 2.5.11 Scheme 2.3. Kinetic scheme that includes the UV band 63
dynamics in Scheme 2.2 and also another compartment
necessary to populate a significant amount of triplet
state. This compartment is not yet assigned. Not
plotted on an energy scale.

Figure 2.5.12 a) 950 nm probe of CuBr
3
2−
with (red trace) and 66
without (black trace) 0.4 M nitrate. b) 400 nm probe
and 520 nm probe for CuBr32- with (red and blue
traces) and without (black and green traces) 0.4 M
nitrate. Data normalized according to UV/Vis
absorption at 266 nm. (See supplement 2.1 for UV/vis).

Figure 2A.1 a) CuBr
3
2−
transient 400 nm time traces without 73
quencher (green trace) and with [H
+
] = 0.2 M (black
trace). Data normalized at signal following 2PA. Red
trace is fit. b) 400 nm time traces for CuBr
3
2−
with
(black trace) and without (green trace) and NO
3
-
. Red
trace is fit.

Figure 2A.2 a) Back to back scans for 700 nm transient absorption 73
of CuBr
3
2−
solution without quencher and with acid
(taken from broadband data in Section 2.2, Figure (a)
and 2.2.4b). 680 nm time cut for CuBr
3
2−
with nitrate,
not back to back with other two scans. The nitrate
solution has been normalized to the unquenched
solution. Black traces are experimental data and red
traces are fits. b) Time traces of 950 nm probe un-
normalized. Black trace is CuBr
3
2−
solution, red trace is
CuBr
3
2−
with 0.4 M NO
3
−
and blue trace is neat water.

xiv
Figure 2A.3 Difference plots of 266 nm excited, broadband probe 74
aqueous CuBr
3
2−
in 0.4 M Br
−
at neutral pH. Plots
show both UV grating and visible grating detection.
Positive values indicate increased absorption and
negative values indicate decreased absorption. a) 100
fs spectral cut subtracted from 150 fs spectral cut
reveals a band in the visible decreases while a dual
band centered at approximately 350 nm and further in
the blue grows in. b) Two traces are 500 fs subtracted
from 1 ps and 1.5 ps subtracted from 2 ps. There is
continued decrease in the visible and increase of the
dual UV band and a shoulder at about 400 nm. c) A
narrow band at approximately 340 nm grows in with
decrease at approximately 440 nm. d) Difference plots
for neat water.

Figure 2B.1 Black traces: 266 nm 3.4 mW excitation of copper 77
bromide in aqueous solution of 0.4 M Br
-
at neutral pH.
Red Traces: 255 nm 1.5 mW excitation of copper
bromide in aqueous solution of 0.3 M Br
-
at neutral pH.
Detection wavelength is 900 nm and experiments were
done on 250 kHz repetition rate laser system.

Figure 3.2.1 Vertical electronic excitation schematic for a diatomic 81
molecule. R represents internuclear distance, the
parabolas are potential energy surfaces (PES) for the
ground and excited states, and the Gaussians are
electronic wavefunctions. a) Ground and excited PES
with same internuclear minima. b) Excited PES has
longer equilibrium internuclear distance than ground
PES.

Figure 3.2.2 a) Schematic of probe wavelength absorption of a 82
propagating wavepacket. The colored lines represent
various energies of resonant photons that promote the
vibrational wavepacket from the excited state PES to
the probing surface. b) Observed spectral signal with
probe center corresponding to blue energy arrow in (a).
Damping is induced by wavepacket dephasing and/or
vibrational energy relaxation by interaction with the
solvent.

xv
Figure 3.2.3 Experimental IR crystallographic spectra for 85
[PMePh
3
]
2
[CuBr
3
]. Values are listed in Table 3.2.1.
The copper species in this crystal has been assigned by
crystallography as Cu
I
Br
3
2-
.

Figure 3.2.4 In these figures, the green ball represents Cu, and 85
purple the Br
-
ligands. Arrows indicate the direction of
movement. a) Vibrational normal modes of three
coordinate complex molecule with D
3h
symmetry.

1
(a
1
') is Raman active, 

(a
2
") and both 
 
and 
 
e')
modes are IR active. b) Normal modes of dibromo
complex with D
∞h
symmetry. Here, the 
1
symmetric
stretch ( 
g
+
) is Raman active and both the 
3

asymmetric stretch ( 
u

) and 
2
bend ( 
u
) are IR active.

Figure 3.2.5 Crystal Field Diagram for Trigonal Planar Complexes 87
with occupation shown for d
10
copper(I).

Figure 3.4.1 266 nm excited CuBr
3
2-
transient time cuts. Vertically 90
offset for comparison. a) Time traces for probe
wavelengths from 301 nm to 700 nm. b) Time cuts for
wavelength range from 380 nm to 430 nm, labeled on
graph in nm.

Figure 3.4.2 a) From Chapter 2, figure is reproduced to aid in 90
assignment. Broadband spectral cuts from 0.5 ps to 400
ps of aqueous CuBr
3
2-

after photo-excitation at 266 nm.
Arrow points to region we observed least modulation of
the transient absorption signal. b) Raw 350 nm (blue)
and 501 nm (red) time traces to emphasize modulation
depth. Data not offset.

Figure 3.4.3 a) Kinetic traces of 266 nm excited CuBr
3
2-
from 301 91
nm to 550 nm. b) Labels of oscillatory peaks
corresponding to Table 3.4.1.

Figure 3.4.4 a) Discrete Fourier transform for 350 nm, 410 nm, and 93
500 nm kinetic traces for the time window of 0 to 2 ps.
b) Discreet Fourier transform phase for 350 nm (blue
triangles) and 500 nm (green squares), left axis.
Dashed lines are FT from (a), right axis.

xvi
Figure 3.5.1 a) UV/visible absorption spectra of 0.004 M ground 97
state CuBr
3
2-
in water. b) One color 266nm
excite/266nm probe data taken on the 1 kHz repetition
rate laser system. 266nm/266nm experiments on the
250 kHz laser system (data not shown) yield the same
results.

Figure 3.5.2 Physical model for impulsive dissociation of CuBr
3
2-
, 100
prompting a bending vibration of CuBr
2
-
. Arrows
represent the forces delivered in the bond breaking.

Figure 3.5.3 a) Scenario describing the impulsive model in which 102
the starting ground state has an equilibrium Br-Cu-Br
bond angle of 120° and the photoproduct has an
equilibrium bond angle of 180°. The probing surface is
probably linear (top solid parabola) with a different
force constant (based on CuCl
2
calculations and
assuming Br follows the same trends). The dashed
double well represents the PES if the probing surface
has a bent equilibrium geometry. b) The CuBr
2
-

product has a shorter Cu-Br bond length than CuBr
3
2-

and Br
-
dissociation would lead to a prompting of the
stretching mode. The probing surface depicted in this
cartoon is dissociative; however, the true PES is
unknown.

Figure 3.5.4 a) Two Morse oscillators with similar curvatures and 104
with the probing surface (blue trace) displaced to the
longer bond length side. b) Difference potential of the
two potentials in (a).

Figure 4.2.1 Proposed kinetic scheme as described in Chapter 2. 111

Figure 4.2.2 a) Luminescence spectra of aqueous CuBr
3
2ˉ
. The 112
absorption and emission band maxima are 281 nm and
475 nm, respectively. b) Two scenarios for the
observed TCSPC signal. The rise in the black trace is
an illustration of an instrument limited emission signal
in which the luminescing species has been populated
within the time of the instrument response. The red
trace is the emission of a species that has a population
growing in slower than the instrument resolution.

xvii
Figure 4.4.1 TCSPC measurement of 266 nm excited aqueous 117
CuBr
3
2ˉ
in 2 M Br
ˉ
concentration under nitrogen (black
trace). Fit to exponential decays (red trace). (inset)
Time window of <0 ps to 600 ps. Experimental
conditions were ~230 µW excitation power, 0.6 mm slit
width, magic angle detection with a 475 nm
interference filter before monochromator to further
block stray photon scatter.

Figure 4.4.2 a) TCSPC data for 0.4 M Br
ˉ
CuBr
3
2ˉ
aqueous solution 119
and scattering control solution excited at 266 nm.
Luminescence was detected at 475 nm (green trace),
360 nm (red trace), and 335 nm (blue trace). Different
interference filters were placed before the
monochromator entrance for each detection wavelength
and the slit width was 2.4 mm throughout. The 360 and
335 nm data is translated vertically for ease of viewing.
The scattering solution used for IRF (black trace) was
detected at 266 nm. Excitation power was 850 – 950
µW. The faster than expected long component shown
in the green trace is due to oxygen quenching of the
triplet state since these samples were under ambient
conditions following preparation. (Inset) Time window
of -50 ps to 2000 ps (same color scheme, the 360 nm
dataset is not translated in this panel). b) Raw TCSPC
data for CuBr
3
2ˉ
solution at 0.4 M Br
ˉ
ligand
concentration and ~0.4 M ionic strength compared with
a scattering solution. 335 nm detection at magic angle
with a 340 nm interference filter placed before the
monochromator entrance slit of 2.4 mm. Collection
time 600 s. Excitation for both samples is 950 µW of
266 nm.

Figure 4.4.3 a) Back to back TCSPC measurements of 1 M Br
ˉ
120
CuBr
3
2ˉ
solution (red trace) taken for a duration of 1200
seconds and scattering control solution measurement
taken for 600 seconds (black trace). ~230 µW 266 nm
excitation, 475 nm detection at magic angle, 475 nm
filter before monochromator entrance slit width of 0.6
mm, 5 ns time window. b) Neat water excited with
260 µW of 266 nm and detected at 360 nm with no
polarizer and no filter before monochromator entrance
slit width of 2.4 mm. Collection time was 1200 s with a
50 ns time window. Both a) and b) solutions were
under ambient conditions following preparation.
xviii

Figure 4.4.4 Luminescence lifetime traces for 280 nm excited, 475 122
nm detection of ~0.004 M CuBr
3
2ˉ
in 0.2 M Br
ˉ

concentration without quencher (black trace) and with
0.2 M KNO
3
(red trace) recorded back to back. Inset is
normalized and on a log scale and includes fit to
quenched solution (smooth black trace). The fit is
biexponential with a 20 ps and 1250 ps time. Solutions
were under ambient conditions.

Figure 4.5.1 a) 266 nm excited CuBr
3
2ˉ
0.4 M Brˉ solution at ~ 0.4 M 125
ionic strength detected at 335 nm (black dots) and 266
nm excited 266 nm detected scattering solution (red
trace) as IRF measurement. Time is offset from zero;
peak of photon counts taken as time zero. b) Overlay
of TCSPC CuBr
3
2ˉ
data with the convolution of the data
with exponential functions. c) Absolute value of
convolution of IRF with varying exponential decays.
Decreasing the exponential decay time decreases its
contribution to the overall signal to the limit where the
instrument is insensitive to the fluorescence and no
signal will be observed. Solutions kept under ambient
conditions following preparation.

Figure 5.2.1 Example of aqueous nanoparticle (CdSe/ZnS QD) 133
absorption (black line) and fluorescence spectra (red
line). The exciton peak is approximately at 585 nm and
the fluorescence peaks at 595 nm. There are features
toward the blue side of the absorption band, unlike the
bulk absorption spectrum that resembles a step function
(not shown).

Figure 5.2.2 Formation and relaxation pathways for a quantum dot 134
exciton. An electron is promoted by a photon into the
conduction band, leaving behind a hole in the valence
band. 1) The electron and hole relax to the band edges.
2) The electron radiatively recombines with the hole.
3) The electron falls into a shallow trap state near the
band edge and then either radiatively or non-radiatively
recombines with the hole, and/or the band edge hole
falls into a trap state near the valence band edge. 4)
The electron falls into a deep trap state and radiatively
recombines with the hole.

xix
Figure 5.2.3 CdSe/ZnS quantum dots are capped with 3- 139
mercaptoacetic acid (MAA). Dopamine is coupled to
the MAA via an amide bond. In this picture there is
just one dopamine conjugated to the QD, however, the
actual system has 100s of dopamine per QD.

Figure 5.4.1 a) 1 minute TCSPC data collection of 84 nM of 144
CdSe/ZnS with (red trace) and without (black trace)
BME. Excitation was 2.45 mW of 400 nm and
detection at magic angle with a 200 ns time window.
0.6 mm slits at the monochromator entrance were used.
b) Sum of 20 1 minute scans of the same samples and
under the same conditions as (a). The small spike at ~
75 and 180 ns delay time is an artifact of the TCSPC
instrument and not due to the QD photodynamics.

Figure 5.4.2 a,b) Normalized TCSPC scan of CdSe Core with (red 145
trace) and without (black trace) BME. In this
experiment, 30 second scans were obtained. Shown is
the sum of the first two scans (a) to compare 1 minute
collection times as with the other samples. Because the
QY is small, there are poor statistics in the initial scans
and it is impossible to determine a difference in
lifetimes between QD only and QD ΒME, however, the
20 minute scan (b) shows an enhancement. c) CdTe
normalized 1 minute scan. The solution without ΒME
(black) has a luminescence lifetime near single
exponential, however, adding ΒME (red) makes the
decay more multi-exponential. d) CdTe sum of 20
scans, normalized.

Figure 5.5.1 Relative energy levels of valence bands and conduction 149
band of CdSe, CdSe/ZnS, and CdTe semiconductors in
aqueous solution and β-mercaptoethanol (BME).
Redox potentials for semiconductors are taken from
published values obtained by cyclic voltammetry for
nanoparticles comparable in size to the ones we used in
this study. Redox potential of BME taken from
Reference 6 and is for pH = 7.

xx
Figure 5.5.2 Schematic of CdSe/ZnS QD solubilized with MPA. 150
There is a BME attached to an exposed Cd. A hole
trapped at a nearby exposed Se site transfers to the
BME. In this figure, there is also an electron trapped at
an exposed Cd site. If MPA or BME were bonded to
this Cd, the electron would not be able to trap here.
BME acts to passivate the Cd dangling bond but may
also accept a hole. These are the two competing
pathways leading to enhancement or quenching of
luminescence.

Figure 6.1 Fluorescence upconversion. a) The fluorescence signal 158
is collected and directed to a nonlinear crystal where it
is overlapped with a probe or “gate” pulse, usually 800
nm in wavelength. The ultraviolet SFG signal is
collected with a photon counting PMT. b) The gate
pulse is delayed and the intensity of the SFG as a
function of delay time is a measure of the sample
fluorescence lifetime.

Figure 6.2 400 nm excited, broadband probe of ~ 30 nM aqueous 161
is collected and directed to a nonlinear crystal where it
is overlapped with a probe or “gate” pulse, usually 800
nm in wavelength. The ultraviolet SFG signal is
collected with a photon counting PMT. b) The gate
pulse is delayed and the intensity of the SFG as a
function of delay time is a measure of the sample
fluorescence lifetime.

xxi

Abstract

Ultrafast pump-probe and pump-broadband probe spectroscopy and time-
correlated-single-photon-counting (TCSPC) techniques are used to study two different
systems that undergo charge separation upon UV excitation. The first system is a model
transition metal coordination compound, aqueous tribromocuprate(I) anion (CuBr
3
2-
).
CuBr
3
2-
has an absorption band centered at 280 nm, assigned as charge-transfer-to-
solvent (CTTS) since resonant excitation in this band produces solvated electrons as
reported in the literature. However, unlike most CTTS systems where ejection occurs on
a femtosecond timescale, electron ejection has previously been reported for CuBr
3
2-
to
occur over nanoseconds, and the anion also undergoes intersystem crossing from its
initially populated CTTS state to a triplet state. The spectral and kinetic features obtained
in the current ultrafast experiments reveal that UV excited CuBr
3
2-
has complex
femtosecond/picosecond excited state dynamics and a surprisingly small quantum yield
of prompt solvated electrons via its charge-transfer-to-solvent (CTTS) state. Analysis of
data from a combination of broadband, low intensity, and photon counting experiments
contributes to a proposed kinetic model, which includes some of the transient species
reported for nanosecond flash photolysis. Aside from transient kinetics, an oscillatory
signal was analyzed and it is determined that a vibrational wavepacket on the potential
energy surface of the CuBr
3
2-
CTTS state is launched via resonant excitation by the pump
pulse. This vibrational coherence survives intersystem crossing into a CuBr
3
2-
triplet
state.
xxii
A second set of systems studied include CdSe/ZnS, CdSe, and CdTe quantum
dots (QDs), where charge separation to an electron – hole pair precedes emission. These
QDs in the presence of β-mercaptoethanol (BME) are characterized by time-correlated-
single-photon-counting (TCSPC) and it is observed that BME has an effect on the
photoluminescence lifetimes in all three of these QD systems. Varying the nature of the
QD surface and also the band gap energy of the QD changes the luminescence lifetime of
the QD – BME system. It is proposed that the presence of BME leads to competition
between two mechanisms: surface passivation of Cd electron traps resulting in
photoenhancement, and hole transfer between Se and BME resulting in luminescence
quenching.

1

Chapter 1
General Introduction

1.1 Charge Separation in Molecular and Nanoparticle Systems

The ability of a system to undergo charge separation is valuable for many
applications ranging from solar energy conversion
7-9
to biological labeling.
1, 10-12

Probably one of the most visible areas of research today involves conversion of solar
energy into other forms of energy that are useable by society. This is a major goal in
today’s world since other natural resources are being depleted, while our need for power
sources is increasing.
13
Yet charge separation systems can also aid in charge transfer
control, optimization of light emitting diodes used for display screens, and the many
biological uses of quantum dots. This section will first introduce a few relevant systems
and applications, and then give a brief description of the two systems that have been
studied and will be presented in detail in this Thesis. It will be shown that simpler
models can be studied to shed some light on the mechanisms of charge separation.

2
1.2 Applications for Solar Energy Conversion

The sun delivers energy to the earth’s surface at a rate four times that of the rate
of energy consumption by humans.
8
It is the most abundant source of energy available
on earth.
9
While the technology for converting sunlight into useable energy has existed
for decades, it is still inefficient and costly. Currently, much attention is focused on
increasing the yield of energy harnessed per solar photon collected. The forms of energy
we are interested in are chemical, electrical, or mechanical energy. Examples of systems
that convert light into these forms of energy are natural and artificial photosynthesis
structures,
8, 9
molecular machines,
14, 15
and solar cells.
7

1. Solar Energy Conversion into Chemical Energy. Photosynthesis is an excellent
example of the conversion of light into chemical energy. Nature has taken advantage of
sunlight by implementing a range of efficient photosynthesis systems for organisms in
varying environments.
9, 16
All these systems entail light harvesting, charge separation,
catalysis, and photoprotection. Light harvesting involves a chromophore, or “antenna”,
which absorbs a photon generating an exciton, which then migrates to a reaction center
where charge separation occurs. These reaction centers can be thought of as nano-
photovoltaics since they transfer electrons from donors to acceptors, creating charge
separated states.
8
The reduced and oxidized molecules created by reaction center charge
separation result in a trans-membrane potential, eventually used for water oxidation and
fuel production in living organisms.
Since natural photosynthesis is the best method for solar energy harvesting known
on the planet, it is an excellent model system for artificial photosynthesis engineering.
8

3
An example of an artificial reaction center is the carotenoid-porphyrin-fullerene triad that
absorbs a photon and induces charge separation.
8
Often, artificial antennas are coupled to
reaction centers, such as a molecular hexad, that strongly absorbs photons and rapidly
and efficiently transfers energy to the reaction center.
8
Photoelectrochemical cells
(PECs) also convert light to chemical energy. One example reported by Li et. al.
17
is a
PEC device that utilizes a Ru
II
visible absorber as a photosensitizer, ([Ru(bpy)
2
(4,4’-
(PO
3
H
2
)
2
bpy)]
2+
). This complex is attached to TiO
2
nanostructures that comprise the
anode and a Pt surface in the device which acts as the cathode. Upon photoexcitation,
this PEC can oxidize water, leading to the production of H
2
and O
2
, thereby
accomplishing the splitting of water. Water splitting by visible light is a promising
pathway for conversion of solar energy into a fuel.
17

2. Solar Energy Conversion into Mechanical Energy. Molecular machines
perform a specific motion when initiated by an energy source and can either be synthetic
or naturally occurring as some proteins undergo folding upon chemical activation.
18

Saha, et. al.
14
give a good description of synthetic molecular machines and review a
number of them, which are of great interest to the nanotechnology community. Examples
include a Cu
I
/Cu
II
complex that undergoes linear motion of its ligand upon oxidation or
reduction, a Ru complex involved in rotational motion, and other systems that take part in
molecular shuttling, isomerizational mechanical switching, and macroscopic transport. In
all of these examples, absorption of a photon induces a chemical change that causes a
discrete movement of the molecule that may or may not be reversible. Both chemical and
electrochemical energy can also be used as stimuli,
14
however, these types of power
4
sources result in waste products whose accumulation will result in a hindrance of the
motion of the molecular machine,
15
making photons the best source of power.
14, 15

Figure 1.1. Schematic of a solar cell mechanism.

3. Solar Energy Conversion into Electrical Energy. Charge separation is key to
the function of solar cells. Conversion of light to electrical energy in a photovoltaic solar
device consists of several steps as depicted in Figure 1.1. First, a strongly absorbing
molecule (or substrate) is photoexcited to generate an electron-hole pair, or what is
known as an exciton. The exciton then diffuses randomly and if it reaches a donor-
acceptor interface before it relaxes, charge transfer can occur and a geminate pair
(donor
+
, acceptor
-
) is formed. The charges migrate to the electrodes and are collected.
Optical absorption and
exciton formation
Exciton migration to the
donor-acceptor interface
Exciton dissociation into
charge carriers
Charge carrier mobility
Charge Collection
1
5
2
3
4
5
There is currently much research being done on optimizing the efficiency of solar cells,
and implementing organic molecules into the design. One method is to introduce a
photosensitizer that can improve initial light collection, such as organic donor-bridge-
acceptor dyads,
19
organic dyes,
19
and even chlorophyll.
20
Other methods include
reducing the pathlength necessary for the exciton to reach the donor-acceptor interface,
and introducing triplet sensitizers to create longer lived excitons.
21

1.3 Other Charge Separation Systems

Several other charge transfer systems discussed in this section include charge
transfer control for device engineering, quantum dots for biological labeling, and
intramolecular charge transfer for light emitting diode technology.
1. Charge Separation Control. Benniston et. al.
22
report on the advancements
made in controlling the dynamics and direction of charge transfer in donor-spacer-
acceptor molecular assemblies. Key aspects to this research include the conformation of
the molecular units, promotion of fast energy transfer along the molecular axis, and
preferential direction of transfer. The ultimate goal of this research is to create artificial
neural networks that work under light illumination.
2. Organic Light Emitting Diodes. Among the many applications of transition
metal complexes, such as dye-sensitized solar cells, sensors, biological labels, and
catalysts, transition metal complexes that undergo fast intersystem crossing (ISC) have
been applied to organic light emitting diodes (OLEDs) technology. Current commercial
6
products the use OLED technology are cell phones, MP3 players, and TV monitors.
23
A
simple structure of an OLED includes a stack of multiple thin organic layers in between a
transparent anode and a metallic cathode.
23
A representation of an OLED is shown in
Figure 1.2. When a potential is applied, a positive charge is injected into the anode and a
negative charge into the cathode. The two charges migrate through the hole transporting
layer and electron transporting layer, respectively. Upon reaching the organic layer, the
electron will occupy the LUMO of the organic molecule and the hole will transfer to the
HOMO. The electron will then relax, and then the organic molecule either undergoes
internal conversion or intersystem crossing to a triplet state.

Cathode
Electron Transport Layer
Emitting Layer
Hole Transporting Layer
Transparent Conducting Anode
Glass or Plastic Substrate
e
-
h
+
Cathode
Electron Transport Layer
Emitting Layer
Hole Transporting Layer
Transparent Conducting Anode
Glass or Plastic Substrate
e
-
e
-
h
+
h
+

Figure 1.2. Basic structure of an OLED. Adapted from Reference 23.

For heavy transition metals, such as iridium and ruthenium, large spin-orbit
coupling facilitates intersystem crossing (ISC) so that ISC can compete with fast
7
fluorescence, instead resulting in strong phosphorescence.
24
One example of an organic
LED complex is fac-tris(2-phenylpyridine) iridium(III) [Ir(ppy)3], which has ~ 100%
phosphorescence quantum yield.
24
Other transition metals utilized include ruthenium and
osmium.
12, 25
The emission color tuneability can be changed by varying the ligand and
some devices utilize several emissive layers of different colors to obtain white light.
26

While OLEDs are an example of bringing two charges together, and the theme of
this document is charge separation, the important characteristic of the organic molecules
utilized in OLEDs is its optical charge transfer bands. Transition metals complexed to
organic ligands can undergo metal-to-ligand-charge-transfer (MLCT) or ligand-to-metal-
charge-transfer (LMCT) depending on the nature of the metal center and the ligands. The
electron injected into the LUMO and hole transferred to the HOMO of the transition
metal complex creates this MLCT state and vice versa for a LMCT state.
3. Semiconductor Nanoparticles. Colloidal semiconductor nanoparticles, named
quantum dots (QDs), have been employed as chemical sensors,
11
photocatalysts,
27
solar
cell absorbers,
10
lasers,
28
and cellular labels,
29
among others. As a semiconductor,
excitation leads to the generation of an electron in the conduction band and a hole in the
valence band. As a nanoparticle, confinement effects introduce valuable characteristics,
such as emission tuneability, narrow band emission yet wide absorption spectra, and a
large effective stokes shift. Another consequence of the small size of QDs is that they
have a large surface area ratio. A more thorough description and background on quantum
dots and its photophysics is presented in Section 5.2 of this Thesis. The main discussion
in this Section refer to the applications that result because of the ability of the QD to
transfer electrons or holes to surrounding molecules, thereby increasing the distance
8
between the photogenerated electrons and holes. A schematic is presented in Figure 1.3,
which describes a scenario in which an electron – hole pair in a QD can transfer an
electron or hole to a reducing or oxidizing agent in solution.

CB
h  h 
O
R
O, R
E
CB
h  h 
O
R
O, R
E

Figure 1.3. A photon is used to produce an electron (blue dot) – hole (red dot) pair. In competition
with recombination of the electron and hole, transfer can occur to a molecule in a surrounding
environment. O represents an electron acceptor, or oxidizing agent, and R represent a hole acceptor,
or reducing agent.

Dimitrijevic et. al.
10
have utilized TiO
2
quantum dots as an absorber and charge
donor, and dopamine (DA) and pyrroloquinoline quinine (PQQ) as charge acceptors for
solar conversion to electrical energy. In their system, photoexcitation generates an
electron – hole pair in the QD, then the hole transfers to dopamine while the electron
transfers to PQQ. The intent is to collect the charges in a similar manner as solar cell
devices. Harris et. al.
27
have studied a CdSe QD – TiO
2
– methyl viologen (MV
+
)
system. The CdSe acts as a photon acceptor, then transfers an electron to TiO
2
and
subsequently to MV
+
, leading to photocatalysis. Their study shows the potential to
develop technology for solar energy to hydrogen production and also photocatalytic
remediation. Other groups have increased the distance between a photogenerated
9
electron and hole in a QD by using electron or hole accepting conjugates spaced by
linkers.
10, 30

1.4 Systems Presented in this Thesis

1. Tribromocuprate(I) Anion. Tribromocuprate, CuBr
3
2-
, has several features
worth exploring. First, it was studied to continue a systematic study in our lab of a series
of model multiply charged anions that may undergo photodetachment directly (photo-
induced charge separation into the liquid) or via charge-transfer-to-solvent (CTTS).
31-36

Second, it is a good candidate for transition state spectroscopy (TSS) in the solution
phase and possibly producing observable quantum beats in the transient absorption signal
with ~70 fs time resolution available to us. Third, CuBr
3
2-
is a simple transition metal
complex that has been reported to not only avoid prompt electron detachment after
excitation into its CTTS state, but to undergo intersystem crossing. This is surprising and
we sought to confirm its curious dynamics following UV excitation.
CTTS photodetachment is much like the MLCT and LMCT processes in OLED
transition metal complexes; however, the transfer is from the molecule or atom to the
solvent.
37
This type of photodetachment generally is known to produce a very short lived
excited state species before electron ejection occurs (See Figure 1.4). Separation of the
electron from the atomic molecular moiety is important for studying both the dynamics of
the parent species and the solvated electron in solution. Most studies have focused on the
CTTS states of simple atomic anions.
34, 38-42
Here we further the study to a transition
10
metal coordination compound and find some surprising results regarding the lifetime and
therefore stability of the CTTS state.

a b
P
e
-
P
e
-

P
e
-
P
e
-

Figure 1.4. Simple schematic of an atomic or molecular anion that undergoes charge-transfer-to-
solvent (CTTS) photodetachment. The blue figures represent water. P represents an atom or
molecule. a) The electron occupies a diffuse space defined by the solvent cavity. This is known as
the CTTS state. b) One detachment mechanism that can occur is diffusion of the electron away from
its parent atom or molecule. This type of CTTS detachment is a good model for detachment from
singly charged halide anions.

CuBr
3
2-
photodetachment also potentially applies transition state spectroscopy
(TSS) to the solution phase. TSS involves electron detachment from an anion to
“instantaneously” create a transition state in a known chemical reaction.
43
The dynamics
of the detached molecule resembling the transition state can be probed spectroscopically
to help understand the species in this region of the reaction coordinate. TSS is commonly
performed on gas phase molecules, yet our molecule of interest, CuBr
3
2-
, has features
favorable to TSS in solution. First, based on ferrocyanide photodetachment studies
performed in our laboratory, a multiply charged molecule will eject an electron at a
further distance from the parent compared to detachment from singly charged
11
molecules.
35
The greater distance for detachment removes the electron from interfering
in the subsequent dynamics taking place in the detached parent molecule. Second,
photodetachment is postulated to occur from the copper center. This allows us to take
advantage of the fact that CuBr
3
2-
has a Cu in an oxidation state of +1 and detachment
produces a +2 Cu species. The Born-Oppenheimer approximation tells us that the
detached CuBr
3
-
geometry is “frozen” on the time scale of electron ejection. While the
geometry of CuBr
3
2-
is favorable for a Cu(I) species, it is unfavorable for a Cu(II)
species.
Figure 1.5 is a schematic of the process of TSS, although instead of a true
transition state generated by the photon, we speculate that CuBr
3
2-
photodetachment
would generate a reaction intermediate for an assumed dissociative interchange (I
d
)
reaction. In the CuBr
3
2-
system, vibrational wavepacket motion is prompted because of
the Cu
I
/Cu
II
structural preferences discussed above. Due to the relatively large mass of
the Br
-
ligands and the 50 fs time resolution of our pump-probe experiment, an induced
vibration is detectable in our transient spectra. Analysis of wavepacket motion is
beneficial for understanding the potential energy surface involved in photoexcited
tribromocuprate.
A very interesting finding in the CuBr
3
2-
excited state system is that it undergoes
intersystem crossing, much like other high spin-orbit coupling complexes, yet this
intersystem crossing is faster than Cu complexes described by other groups.
44
The rate of
intersystem crossing in fact determines the lifetime of the initially excited CTTS state,
which turns out to be much longer than most other CTTS state reported in the literature.
34,
37, 41, 45, 46
Only two other examples of transition metal complexes with CTTS states
12
living longer than a few hundred femtoseconds have been reported both by Rentzepis and
coworkers.
47, 48

ABC
2-
ABC
-
A+BC
AB+C
Anion (2-)
Reaction
Intermediate
Anion (1-)
ABC
2-
ABC
-
A+BC
AB+C
Anion (2-)
Reaction
Intermediate
Anion (1-)

Figure 1.5. A doubly charged anion is photoexcited to produce a ground state singly charged species
resembling a reaction intermediate in an I
d
ligand exchange mechanism.

2. CdSe/ZnS, CdSe, and CdTe Quantum Dots. QDs are often conjugated to
molecules that can provide additional functionality to the QD. One example is
conjugation to biologically active molecules, such as dopamine (DA). A system
presented in this Thesis is CdSe/ZnS QDs used as cellular labels. When conjugated to
DA, the QD can undergo uptake into a live cell. Charge transfer from the photoexcited
QD to the dopamine conjugate, and subsequent charge transfer from agents in the cell to
the dopamine changes the QD fluorescence quantum yields.
1
Thus, the brightness of the
QD is an indication of the redox potential in various regions of the cell (See Figure 1.6).
1

13
In this example, charge transfer is across the interface of the QD to its conjugate and
between the conjugate and its environment. Other QDs that we have previously studied
include CdSe and CdTe, also conjugated to dopamine.
49

Figure 1.6. Fluorescence imaging of live cells. CdSe/ZnS quantum dots conjugated to dopamine are
within these cells via uptake by the dopamine receptor transport in the cell membrane. The QDs
fluoresce brighter in regions of high oxidation potential. Taken from Reference 1 with permission.

While the interaction between the QD and the environment is valuable, problems
occur when there are defect states on the surface of the QD. The most common defect
states are exposed Cd and Se atoms, or in the case of CdTe, Cd and Te. Often a shell or
passivating molecules are used to “cover” the defects on the surface, thereby shutting off
its undesirable fast electron-hole recombination. (A detailed description is presented in
the Introduction of Chapter 5, Section 5.2). In our own experiments, in order to
determine if the QD itself (and not only DA) is interacting with the surroundings, such as
a reducing agent, experiments have been performed to complement our published bio-
relevant studies
1, 49, 50
and these are reported in this Thesis.
14
1.4 References for Chapter 1

1. S. J. Clarke, C. A. Hollmann, Z. J. Zhang, D. Suffern, S. E. Bradforth, N. M.
Dimitrijevic, W. G. Minarik and J. L. Nadeau, Nat. Mater., 2006, 5, 409-417.

2. K. L. Stevenson, R. S. Dhawale, A. Horvath and O. Horvath, J. Phys. Chem. A,
1997, 101, 3670-3676.

3. K. L. Stevenson, D. W. Knorr and A. Horvath, Inorg. Chem., 1996, 35, 835-839.

4. A. Treinin and E. Hayon, J. Am. Chem. Soc., 1975, 97, 1716-1721.

5. D. Zehavi and J. Rabani, J. Phys. Chem., 1972, 76, 312-&.

6. B. Mickey and J. Howard, 1995, 130, 909-917.

7. Y. J. Chang and T. J. Chow, Tetrahedron, 2009, 65, 4726-4734.

8. D. Gust, T. A. Moore and A. L. Moore, Acc. Chem. Res., 2009, 42, 1890-1898.

9. I. McConnell, G. H. Li and G. W. Brudvig, Chem. Biol., 2010, 17, 434-447.

10. N. M. Dimitrijevic, O. G. Poluektov, Z. V. Saponjic and T. Rajh, J. Phys. Chem.
B, 2006, 110, 25392-25398.

11. S. Impellizzeri, S. Monaco, I. Yildiz, M. Amelia, A. Credi and F. M. Raymo, J.
Phys. Chem. C, 2010, 114, 7007-7013.

12. A. T. Yeh, C. V. Shank and J. K. McCusker, Science, 2000, 289, 935-938.

13. J. S. Wilson, A. S. Dhoot, A. J. A. B. Seeley, M. S. Khan, A. Kohler and R. H.
Friend, Nature, 2001, 413, 828-831.

14. S. Saha and J. F. Stoddart, Chem. Soc. Rev., 2007, 36, 77-92.

15. V. Balzani, A. Credi, F. Marchioni and J. F. Stoddart, Chem. Commun., 2001,
1860-1861.

16. Y. Saga, Y. Shibata and H. Tamiaki, J. Photoch. Photobiol. C. Photoch. Rev.,
2010, 11, 15-24.

17. L. Li, L. L. Duan, Y. H. Xu, M. Gorlov, A. Hagfeldt and L. C. Sun, Chem.
Comm., 2010, 46, 7307-7309.
15
18. A. A. Nickson and J. Clarke, Methods, 2010, 52, 38-50.

19. Y. J. Cheng, S. H. Yang and C. S. Hsu, Chem. Rev., 2009, 109, 5868-5923.

20. X. F. Wang, Y. Koyama, O. Kitao, Y. Wada, S. Sasaki, H. Tamiaki and H. S.
Zhou, Biosens. Bioelectron., 2010, 25, 1970-1976.

21. C. S. K. Mak, H. L. Wong, Q. Y. Leung, W. Y. Tam, W. K. Chan and A. B.
Djurisic, J. Organomet. Chem., 2009, 694, 2770-2776.

22. A. C. Benniston and A. Harriman, Coord. Chem. Rev., 2008, 252, 2528-2539.

23. J. Shinar and R. Shinar, J. Phys. D. Appl. Phys., 2008, 41, 133001-133001-
133001-133026.

24. G. J. Hedley, A. Ruseckas and I. D. W. Samuel, Chem. Phys. Lett., 2008, 450,
292-296.

25. P. T. Chou and Y. Chi, Eur. J. Inorg. Chem., 2006, 3319-3332.

26. T. Tsuboi, J. Non.-Cryst. Solids, 2010, 356, 1919-1927.

27. C. Harris and P. V. Kamat, ACS Nano, 2009, 3, 682-690.

28. L. V. Asryan, J. Nanophotonics, 2009, 3, -.

29. A. P. Alivisatos, W. W. Gu and C. Larabell, Annu. Rev. Biom. Eng., 2005, 7, 55-
76.

30. N. A. Anderson and T. Q. Lian, Annu. Rev. Phys. Chem., 2005, 56, 491-519.

31. R. A. Crowell, R. Lian, I. A. Shkrob, D. M. Bartels, X. Y. Chen and S. E.
Bradforth, J. Chem. Phys., 2004, 120, 11712-11725.

32. J. A. Kloepfer, V. H. Vilchiz, V. A. Lenchenkov and S. E. Bradforth, Chem. Phys.
Lett., 1998, 298, 120-128.

33. J. A. Kloepfer, V. H. Vilchiz, V. A. Lenchenkov, X. Y. Chen and S. E. Bradforth,
J. Chem. Phys., 2002, 117, 766-778.

34. J. A. Kloepfer, V. H. Vilchiz, V. A. Lenchenkov, A. C. Germaine and S. E.
Bradforth, J. Chem. Phys., 2000, 113, 6288-6307.

35. V. Lenchenkov, J. Kloepfer, V. Vilchiz and S. E. Bradforth, Chem. Phys. Lett.,
2001, 342, 277-286.
16
36. M. C. Sauer, I. A. Shkrob, R. Lian, R. A. Crowell, D. M. Bartels, X. Y. Chen, D.
Suffern and S. E. Bradforth, J. Phys. Chem. A, 2004, 108, 10414-10425.

37. X. Y. Chen and S. E. Bradforth, Annu. Rev. Phys. Chem., 2008, 59, 203-231.

38. E. R. Barthel, I. B. Martini and B. J. Schwartz, J. Phys. Chem. B, 2001, 105,
12230-12241.

39. S. E. Bradforth and P. Jungwirth, J. Phys. Chem. A, 2002, 106, 1286-1298.

40. M. C. Cavanagh, R. M. Young and B. J. Schwartz, J. Chem. Phys., 2008, 129,
134503-134501-134503-134510.

41. M. K. Fischer, A. Laubereau and H. Iglev, Phys. Chem. Chem. Phys., 2009, 11,
10939-10944.

42. W. J. Glover, R. E. Larsen and B. J. Schwartz, J. Chem. Phys., 2010, 132.

43. D. M. Neumark, Phys. Chem. Chem. Phys., 2005, 7, 433-442.

44. G. B. Shaw, C. D. Grant, H. Shirota, E. W. Castner, G. J. Meyer and L. X. Chen,
J. Am. Chem. Soc., 2007, 129, 2147-2160.

45. A. Kammrath, J. R. R. Verlet, A. E. Bragg, G. B. Griffin and D. M. Neumark, J.
Phys. Chem. A, 2005, 109, 11475-11483.

46. F. H. Long, H. Lu, X. L. Shi and K. B. Eisenthal, Chem. Phys. Lett., 1990, 169,
165-171.

47. J. Chen, H. Zhang, I. V. Tomov, X. L. Ding and P. M. Rentzepis, P Natl. Acad.
Sci. USA, 2008, 105, 15235-15240.

48. J. Chen, H. Zhang, I. V. Tomov and P. M. Rentzepis, Inorg. Chem., 2008, 47,
2024-2032.

49. D. R. Cooper, D. Suffern, L. Carlini, S. J. Clarke, R. Parbhoo, S. E. Bradforth and
J. L. Nadeau, Phys. Chem. Chem. Phys., 2009, 11, 4298-4310.

50. D. Suffern, S. J. Clarke, C. A. Hollmann, D. Bahcheli, S. E. Bradforth and J. L.
Nadeau, 2006, 6096, O960-O960

17

Chapter 2
Photoinduced Excited State Dynamics of Tribromocuprate(I) Anion
and Transient Spectral Evolution

2.1 Abstract for Chapter 2

The excited state evolution of UV excited tribromocuprate (CuBr
3
2−
) is studied via
ultrafast pump-probe and pump-broadband probe spectroscopy. The complex spectral
and kinetic features of the transient absorption are presented and a kinetic scheme
proposed that reflects some of the previously reported transients of UV excited CuBr
3
2−

in nanosecond laser flash photolysis results. While a prominent contribution of electron
photodetachment was expected, a surprisingly small (~ 5%) quantum yield of prompt
electron formation was observed. Even though the QY of electron ejection is small, it
was found that the CTTS electron is diffuse, as in normal inorganic CTTS systems as
judged by its static quenching by nitrate.

18
2.2 Introduction

Copper coordination compounds have received much attention over the past 50
years because of their significance in many areas of research, such as biology,
1
energy
conversion,
2
and molecular devices.
3
Copper is a first row transition metal usually
occurring in an oxidation state of +1 or +2. Cu(I) has a filled d-orbital shell with a
tendency to give up an electron and Cu(II) has an empty d-orbital and makes an excellent
electron acceptor. Nature often takes advantage of the facile redox action of the
Cu(I)/Cu(II) pair
1, 4
and copper protein complexes are responsible for many of the
electron transport processes in living organisms.
5
Intramolecular electron transfer
between the metal center and ligands is an inner sphere process while intermolecular
processes between copper and a nearby molecule are called outer sphere.
Copper complexes are also studied for their potential application in solar energy
conversion and display devices.
2
The advantages of copper over the more traditional
ruthenium and iridium compounds are cost effectiveness,
3
as copper is an earth-abundant
element, and copper compounds also have simple synthesis procedures
6, 7
. Common
systems are Cu(I) complexed to organic ligands containing low energy empty π orbitals.
Examples include [Cu
I
(dmp)
2
]
+
(copper(I) phenanthroline) and [Cu
I
-(NN)
2
]
+
(copper(I)
diimine).
2, 8, 9
Visible photoexcitation of these complexes leads to charge transfer from
the metal to the ligand,
3
creating metal-to-ligand-charge-transfer (MLCT) states. The
MLCT excited state has copper in the +2 oxidation state (d
9
). Information on excited
state dynamics including electronic state crossings and molecular structural
19
rearrangements is sought after for optimization of these systems to function as part of the
above mentioned devices.
2, 10

A class of copper systems that is particularly interesting to theoreticians is the
copper(II) halides. In particular, Cu
II
Cl
2
, has received a considerable amount of attention
in the past few decades because it is a simple example of a Cu(II) coordination compound
that can be studied in the gas phase and also by theory
11
. Since the copper in this
complex has an empty d orbital and closed-shell halide ligands that cannot accept
electrons but are excellent donors, excitation leads to charge transfer from the halide to
the copper. This event is referred to as ligand-to-metal-charge-transfer or LMCT and
produces formally a d
10
Cu(I) excited state. Other copper halides such as Cu
II
Br
2
are
found to possess similar electronic structures as the copper chlorides,
11, 12
and, in addition
to their optical spectroscopy, can be probed by photoelectron spectroscopy of the
corresponding reduced anion, Cu
I
Cl
2
−
or Cu
I
Br
2
−
.
A third possible process for low oxidation state metals such as Ru(II), Fe(II), and
Cu(I) is charge-transfer-to-solvent (CTTS).
13, 14
Cu(I) systems with closed-shell ligands,
such as the halides, are good candidates for CTTS electron ejection because MLCT and
LMCT processes are not facile due to orbital occupancy and the only allowable
intramolecular transitions are extremely high in energy. Other examples of molecules
and atoms that undergo the CTTS mechanism are the group I alkalides, halides, and
pseudo halides.
14
In these systems, a pre-existing solvent polarization allows for diffuse
CTTS states that are partly defined by the solvent cavity. Because the excited electron is
weakly bound it is usually very unstable with respect to rearrangement in the solvent
shell. Therefore, for almost all of the systems reported to date, the CTTS state is short
20
lived, on the order of a few hundred femtoseconds, and supported both by experiment and
theory.
14-18
Decay of the optically prepared CTTS state always leads to electron ejection
and production of a solvated electron. The exceptions to the traditional behavior of
CTTS states are two other metal complexes, [Co
III
(C
2
O
4
)
3
]
3−
and [Fe
III
(C
2
O
4
)
3
]
3−
,
reported in 2008 by Rentzepis and coworkers,
19, 20
that are addressed in the Discussion of
this Chapter, Section 2.5.
Stevenson and Horvath and coworkers have extensively studied the
photochemistry of aqueous Cu(I) halocuprates and cyanocuprates with nanosecond (ns)
laser flash photolysis and fluorimetry
21-38
motivated by these compounds’ potential for
solar energy conversion application since UV excitation in acidic solution leads to
formation of H
2
.
23, 31
Because solvated electrons are observed, the absorption bands for
these Cu(I) halides have been assigned as CTTS, and one of the first steps in the series of
photoinitiated chemical reactions is ejection from the Cu(I) center.
27, 32, 34
Another
feature of this system is that there is a blue shift in the absorption band upon increase of
ionic strength, typical for CTTS bands.
39
Yet unlike other CTTS systems reported by
ourselves and others up until recently,
14-18
the laser flash photolysis results for Cu(I) halo-
and cyanohalocuprates indicate two timescales for electron ejection, the first within the
duration of the nanosecond excitation laser pulse, and the second on the order of tens of
nanoseconds. In addition, a luminescent triplet species evolving from the initially
prepared CTTS state has been identified. This triplet species is known to be the delayed
electron precursor based on the coincidence of times for solvated electron formation and
luminescence decay.
32
Luminescence in itself is unusual for CTTS type inorganic
systems.
21

CuX
3
2-
CTTS
CuX
3
2-
3
CuX
2
- 3
CuX
3
2-
h ν
Br
-
+
CuX
2
-
X
-
+
CTTS
CuX
2
-
Bimolecular
Reaction
products
CuX
3
-
e
(aq)
-
+ CuX
2
-
+ e
(aq)
-
CuX
3
2-
CuX
3
2-
CTTS
CuX
3
2- CTTS
CuX
3
2-
3
CuX
2
- 3
CuX
2
- 3
CuX
3
2- 3
CuX
3
2-
h ν
Br
-
+
CuX
2
-
CuX
2
-
X
-
+
CTTS
CuX
2
- CTTS
CuX
2
-
Bimolecular
Reaction
products
Bimolecular
Reaction
products
CuX
3
-
e
(aq)
-
+ CuX
2
-
+ e
(aq)
-
CuX
3
2-
CTTS
CuX
3
2-
3
CuX
2
- 3
CuX
3
2-
h ν
Br
-
+
CuX
2
-
X
-
+
CTTS
CuX
2
-
Bimolecular
Reaction
products
CuX
3
-
e
(aq)
-
+ CuX
2
-
+ e
(aq)
-
CuX
3
2-
CuX
3
2-
CTTS
CuX
3
2- CTTS
CuX
3
2-
3
CuX
2
- 3
CuX
2
- 3
CuX
3
2- 3
CuX
3
2-
h ν
Br
-
+
CuX
2
-
CuX
2
-
X
-
+
CTTS
CuX
2
- CTTS
CuX
2
-
Bimolecular
Reaction
products
Bimolecular
Reaction
products
CuX
3
-
e
(aq)
-
+ CuX
2
-
+ e
(aq)
-

Figure 2.2.1. Scheme 1. Recreation of the proposed kinetic scheme published in Reference 32.
Scheme applies to aqueous halocuprates(I), X = Cl
−
, Br
−
, I
−
. For the Br
−
system at [Br
−
] = 5 M, the
ground state equilibrium constant is 9.45 and the triplet equilibrium constant is 73, both in favor of the
tribromo- species.

Solutions of ground state CuX (X = Cl, Br, I, CN) in excess X
-
ligand exhibit
speciation equilibria of several species of the form CuX
n
(n-1)−
. In the case of CuBr, at
high ligand concentrations, i.e. 0.5 – 5 M, the dominant equilibrium is between CuBr
2
−

and CuBr
3
2−
and significantly favoring the tribromocuprate species. These two species
have shifted absorption bands with different oscillator strength, with respect to the other,
and the UV absorption bands of both systems are assigned as CTTS. The kinetic model
22
for the CuX
3
2−
(X= Cl, Br, I, CN) systems is shown in Figure 2.2.1. All pseudo halide
and halide systems are believed to follow the same dynamics. Photoexcitation promotes
the ground state to the CuX
3
2−
CTTS state, which has three decay pathways, including
internal conversion back to the ground state, electron ejection and intersystem crossing
into a CuBr
3
2−
triplet state. The triplet state is luminescent and can eject an electron or
form other products via bimolecular reactions. The phosphorescence of this triplet state
has been well studied
27, 30, 32
and is the subject of Chapter 4. The CuX
3
2−
and CuX
2
−

triplet states are in equilibrium and are referred to as an “exciplex”.
27

In an effort to explore the unique behavior of the Cu(I) halocuprate CTTS state,
we performed sub-100 femtosecond ultrafast pump-probe spectroscopy on one of the
model systems studied by Stevenson and Horvath, aqueous CuBr
3
2−
. The
tribromocuprate species is easy to favor because of the equilibrium constant, and
excitation of the dibromo- ground state is greatly reduced. Excitation of the dibromo-
ground state would introduce excited states with inhomogeneous kinetics from that of the
CuBr
3
2−
system and complicate the experimental results. The sophistication of our
experimental setup allows us to collect transient absorption spectra as early as 50 fs
following laser excitation. This allows us to resolve early dynamical information
involving the evolution of the tribromocuprate(I) anion CTTS state and formation of
other transient species. In this work, we combine ultrafast pump-probe spectroscopy and
time-correlated-single-photon-counting (TCSPC) and we propose a kinetic model that
can explain our results.

23
2.3 Experimental

Ultrafast pump-probe spectroscopy was used to measure transient absorption of
266 nm excited aqueous CuBr
3
2−
. Two different experimental setups provided either a
broadband probe or a single color probe. The pump/broadband probe experiments were
done with a 1 kHz repetition rate laser system and the two color experiments were done
with a 250 kHz repetition rate laser system. Transient absorption was measured as a
function of probe delay time.

2.3.1 Optical Instrumentation

The broadband experiments utilized a Ti-sapphire regenerative amplifier (Spectra
Physics Hurricane) operating at a 1 kHz repetition rate. The experimental layout is
depicted in figure 2.3.1. The 1 mJ ~ 100 fs 800 nm output of the amplifier was separated
into the pump and probe paths by a 90/10 beamsplitter and the pump beam was focused
into a 0.5 mm thick BBO ( -barium borate) crystal to generate second harmonic. 266 nm
pulses were produced by a hollow core fiber four-wave mixing scheme.
40
The second
harmonic and fundamental are combined and focused into a fused silica hollow capillary
filled with argon gas at an optimized pressure of ~ 253 Torr for best third harmonic
conversion efficiency, with an output frequency given by ω
uv
= ω
pump
+ ω
pump
− ω
seed
.
The 266 nm pulses, with bandwidth of 5nm, were compressed with a CaF
2
prism pair and
passed through a 500 Hz chopper that blocked every other laser pulse. 2 µJ of 266 nm
24
beam was focused down to a spot with ~ 200 µm FWHM. The probe portion of 800 nm
was reflected off of a retroreflector mounted on a delay stage to vary the delay time
between pump and probe. The stage was carefully aligned to ensure no more than 5%
spatial walkoff between the pump and probe beams over the 500 ps probe window,
determined by measurement of the laser intensity leak through of a ~ 50 µm pinhole
while scanning the translation stage. The probe was then focused into a CaF
2
plate,
which was translating in x and y directions to generate stable white light. The white light
was collimated and focused by a pair of aluminum coated off-axis parabolic mirrors and
sent to the sample. The transmitted probe beam was dispersed by a monochromator onto
a 256-pixel silicon diode array, where the spectrum 290 − 700 nm and 900 − 1100 nm
was recorded for each pulse. Two gratings were used, a UV grating (for ~ 290 − 600 nm)
and a visible grating (for ~ 500 − 1100 nm) in the monochromator. The wavelength
resolution is ~ 2.5 nm at best, estimated by focusing the probe light to a spot size of ~ 100
µm, which is twice as wide as the pixel size. The shot-to-shot transient absorption was
calculated based on the probe signal with and without the pump present. Two
spectrograph gratings were used to cover the 290 − 710 nm spectrum. The sample was
delivered with a wire-guided gravity-driven (WGGD) flow jet.
41
The jet thickness was ~
80 µm at the laser pulse overlap and produced an instrument response of ~ 50 fs
measured by the water 2 photon absorption at zero delay. The sample reservoir was kept
under nitrogen pressure and the sample flowed through once only to avoid
photodegrading the CuBr
3
2−
.

25
Sample
Wire guided
gravity jet
1kHz Ti:Sapphire Amplifier
Chopper
Monochromator
BS
CaF
2
PM
CaF
2
Prisms
BBO
Ar Filled Fiber
Translation
Delay Stage
Sample
Wire guided
gravity jet
1kHz Ti:Sapphire Amplifier
Chopper
Monochromator
BS
CaF
2
PM
CaF
2
Prisms
BBO
Ar Filled Fiber
Translation
Delay Stage

Figure 2.3.1. 1 kHz repetition rate laser system and experimental layout.

Two-color pump-probe experimental data were obtained by use of a 200 − 250
kHz repetition rate Ti:sapphire ultrafast system (Coherent Rega 9050). The amplifier
produces ~ 60 fs FWHM 800 nm pulses and this output is separated into a pump and
26
probe path by a beamsplitter. A schematic is presented in Figure 2.3.2. For the UV
pump beam, a 400 nm pumped tunable optical parametric amplifier (OPA 9450,
Coherent) is used. The OPA produces 510 nm or 532 nm, depending on the pump
wavelength needed for the experiment. The green pulses were compressed with a fused
silica prism pair and then focused into a 0.08 mm thick BBO crystal for second harmonic
generation (SHG).
42
The UV pulses so generated were compressed with a CaF
2
prism
pair and sent to a chopper operating at a frequency of ~ 1800 Hz. The pump beam was ~
1 mW and focused to a spot size of ~ 50 µm at the sample with nJ pulse energies. The
probe beam was sent to a retroreflector mounted on a delay stage, carefully aligned to
ensure less than 5% walkoff at the sample over a 500 ps time window. The probe beam
was focused into a sapphire crystal to produce white light continuum. In the experiments
described here, an interference filter was used to transmit only a single color probe from
the continuum (500 and 900 nm) or 400 nm derived from frequency doubling the 800 nm
were focused to a diameter less than (and overlapped with) the pump spot at the sample.
The transmitted probe beam was detected by a photodiode and the signal recovered by a
SR810 lock-in amplifier reference to the chopper. The liquid sample was delivered by a
fast flowing jet using a horizontally positioned stainless steel nozzle with an opening slit
width of ~ 200 µm and producing a comparable sample thickness. The sample was
circulated and the reservoir and jet kept under nitrogen pressure. The instrument
response was determined by the FWHM of the difference frequency generation between
the pump and probe pulses or by water ionization 2PA and was close to and under 200 fs.

27

Sample
Fast Jet
250 kHz Ti:Sapphire
Amplifier
Chopper
PD
BS
Sapphire
BBO
CaF
2
Prisms
Delay Stage
OPA
Lock-in
Filter
λ/2
FS Prisms
Sample
Fast Jet
250 kHz Ti:Sapphire
Amplifier
Chopper
PD
BS
Sapphire
BBO
CaF
2
Prisms
Delay Stage
OPA
Lock-in
Filter
λ/2
FS Prisms

Figure 2.3.2 250 kHz repetition rate laser system and experimental layout.

2.3.2 Solution Preparation and Sample Delivery

Aqueous CuBr
3
2−
solutions were prepared as follows. A desired volume of
deionized water was placed in an Erlenmeyer flask and deaerated with N
2
for ~ 2 hours.
Sample volumes prepared varied from 200 mL to 1 L. Sodium bromide (NaBr, Sigma-
28
Aldrich, 99+% reagent grade) was used to make the desired Br
-
ligand concentration,
usually 0.4 M. In the dark, copper(I) bromide (Cu
I
Br, Avocado Res. Chem., 98%) was
added to the flask containing sodium bromide in the dark and under N
2
to make a solution
of 4 mM CuBr. The flask was sealed under positive N
2
pressure, stirred and gently
heated for ~3 hours. The flask was removed from the stir/hot plate and left to sit
overnight for undissolved Cu
I
Br to settle to the bottom of the flask.
For the quenching experiments, potassium nitrate crystals (KNO
3
, Mallinckrodt,
100.0% assay grade) was added after the NaBr for a total of 0.4 M NO
3
−
. For constant
ionic strength experiments, sodium perchlorate (NaClO
4
, Sigma-Aldrich, 98+% reagent
grade) was used and added to the flask with NaBr.

2.4 Results

There is an abundance of experimental results presented in this section. Unless
noted with a II superscript following the Cu in the chemical formula, the copper complex
is in the +1 oxidation state. To summarize, the UV/Visible absorption of ground state
CuBr
3
2−
in equilibrium with a small amount of ground state CuBr
2
−
will be reviewed
(Figure 2.4.1). Second, the broadband transient absorption from 290 nm through 700 nm
will be presented (Figures 2.4.2). This transient data contains two bands. Band 1 and its
dynamics will be presented in detail including quenching experiments (Figures 2.4.3 and
2.4.5). Band 2 will be assigned and some IR probe and quenching results shown to prove
its assignment (Figure 2.4.6). Lastly, low intensity 900 nm transient data is presented to
29
examine the evidence for ultrafast solvated electron detachment from copper (Figure
2.4.8).

2.4.1 CuBr
3
2-
Ground State

Aqueous tribromocuprate(I) anion, CuBr
3
2−
, has an absorption band centered at
279 nm with a peak extinction coefficient of 9170 M
-1
cm
-1
assigned as a CTTS band.
43

Figure 2.4.1 shows the steady state UV/visible absorption spectrum of CuBr
3
2−
in an
aqueous solution of 0.4 M Br
−
at room temperature and neutral pH. CuBr
3
2−
is favored in
the equilibrium with its dibromocuprate counterpart

CuBr
2
−
+ Br
−
↔ CuBr
3
2−
(2.1)

with an equilibrium constant of 10.2 at an ionic strength of 2 M.
44
At the ligand
concentration used in our experiments, 0.4 M, we expect a CuBr
3
2−
vs. CuBr
2
−
ratio of
2.17 (the bromide is in excess).
At 266 nm (the excitation wavelength used for the majority of the CuBr
3
2-

experiments in this work) the ratio of optical density for CuBr
3
2-
to CuBr
2
-
is 1.33 in
favor of the tribromo form.
43
Combining the higher extinction coefficient for the
tribromocuprate anion and the higher equilibrium concentration, 90% of the light being
absorbed by a copper complex is due to absorption of CuBr
3
2-
and only 10% due to
CuBr
2
-
.

30
225 250 275 300 325 350
0.0
0.5
1.0
1.5
2.0

OD
Wavelength, nm

Figure 2.4.1. UV/visible absorption spectra of 0.004 M ground state CuBr
3
2−
in water with a small
contribution from CuBr
2
−
(see text). NaBr concentration in solution is 0.4 M.

2.4.2 Broadband Transient Absorption

The broadband probe data sets shown either used a UV grating for ~ 290 – 600
nm detection, or a visible grating for ~ 500 – 700 nm and ~ 900 – 1100 nm detection.
Both the UV and visible data sets were cut at 500 nm, with the bluer wavelengths taken
from the UV data set, and red wavelengths taken from the visible data set. Figure 2.4.2a
shows a 2D contour plot of transient absorption over the spectral range from 350 nm to
600 nm (UV blazed grating), and time delays from 0 to 450 ps. A series of spectral cuts
using two additional datasets employing both gratings have been plotted in Figure 2.4.2b.
The prominent features of the transient spectra are two broad bands, one peaking near
350 nm by 500 ps and one peaking further to the red than the experimental spectral
window. These bands are labeled as (1) and (2) in the plot. Band 2 is assigned primarily
31
to the solvated electron that is unavoidably generated from the solvent and free Br
-
and is
discussed toward the end of this section.

a b
Wavelength, nm
Delay time, ps
350 400 450 500 550
450
400
350
300
250
200
150
100
50
16
14
12
10
8
6
4
mOD
Wavelength, nm
Delay time, ps
350 400 450 500 550
450
400
350
300
250
200
150
100
50
16
14
12
10
8
6
4
mOD

300 400 500 600
2
4
6
8
10
12
0.5 ps
1 ps
2 ps
52 ps
102 ps
202 ps
302 ps
402 ps
2
1 1
1
mOD
Wavelength, nm

Figure 2.4.2. a) Two dimensional contour plot of transient broadband absorption following 266 nm
excitation of aqueous CuBr
3
2-
. Colors represent magnitude of optical density in mOD. b) Broadband
probe wavelength cuts of aqueous CuBr
3
2-
from 0.5 ps to 402 ps.

2.4.3 Transient Band 1 Near 350 – 400 nm

Spectral cuts at delay times earlier than those in Figure 2.4.2b reveal complex spectral
dynamics which is now shown in Figure 2.4.3a and b. The transient absorption signal
does not drop to baseline following coherent 2PA (see Figure 2.4.4a). 2PA is the
simultaneous absorption of 1 pump + 1 probe photon with contributions both from the
solvent and free bromide (non-resonant with the pump pulse) and from the solute
(resonant with the pump). Following 2PA, this early transient absorption displays an
isosbestic point around 420 nm (Figure 2.4.3a), indicating that two species interchange
population over the course of ~ 100 fs. Following this first phase, there is a separate
32
dynamics occurring from 200 fs to 2 ps. This is characterized by growth of a band
centered close to 390 nm (Band 1) and a second absorption band peaking further to the
blue than the experimental window. At longer times shown in Figure 2.4.3b, Band 1 still
continues to grow and blue shift. By ~ 500 ps, the band peaks at 350 nm and the spectral
shape differs again from that of earlier delay times (Figure 2.4.2b).

a b
300 400 500 600
2
4
6
8
10
mOD
Wavelength, nm
Delay time, fs
84
94
104
114
124
134
144
154
164
174

300 400 500 600
2
4
6
8
10
mOD
Wavelength, nm
Delay time
204 fs
504 fs
754 fs
1 ps
1.25 ps
1.5 ps
2 ps

Figure 2.4.3. 266 nm excited aqueous CuBr
3
2-
. Transient broadband absorption spectral cuts of 295
nm to 600 nm measured with the 1 kHz laser system. a) 84 – 174 fs b) Colored traces are 204 fs – 2
ps. The 204 fs delay absorption spectra breaks the isosbestic point. The black traces are those plotted
in (a).

33

a
0.0 0.1 0.2
3
4
5
6
7
8
9
Absorption, mOD
Time, ps
0.0 0.1 0.2
3
4
5
6
7
8
9
Absorption, mOD
Time, ps

Wavelength, nm
550
526
501
475
451
426
401
392
360
340
320

b c
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8
5
10
Absorption, mOD
Time, ps

0 100 200 300 400
5
10
Absorption, mOD
Time, ps

Figure 2.4.4. 266 nm excited transient absorption of CuBr
3
2−
for single wavelengths as a function of
delay time. Wavelengths are 320 nm to 550 nm. a) 0 to 250 fs. The initial drop across all
wavelengths is part of the 2PA signal. b) 0 to 2 ps. c) 0 to 500 ps.

34
2.4.4 Quenching UV transient Band 1

Either added H
+
or NO
3
−
partly quench Band 1 as observed in the broadband data
shown in Figures 2.4.5a and 2.4.5b. Band 1 has the same behavior for both quenched and
unquenched solutions until ~ 200 ps delay time, at which point the absorption begins to
decrease across all probe wavelengths. For H
+
quenching, the loss of signal appears to
affect the transient spectrum uniformly but this is not the case for NO
3
−
quenching.
Figures 2.4.5a and b show decreases in signal across the UV and visible from 200 ps to
500 ps for CuBr
3
2−
with 0.2 M HBr (a) and CuBr
3
2−
with 0.4 M KNO
3
(b). For
comparison, unquenched data sets are shown in Figures 2.4.5c and d. Fits to a single
exponential decay of the nitrate and acid quenched 400 nm transient data from 30 – 500
ps (NO
3
-
) and 100 – 500 ps (H
+
) are shown along with the kinetic traces in Supplement
2A. At 0.2 M H
+
, the fit of the 400 nm decay is 8 ns. This exponential decay is
consistent with the known H
+
quenching rate constant for a triplet CuBr
3
2−
transient
species, which is 6.2 x 10
8
M
-1
s
-1
, measured by the slope of [H
+
] concentration vs. decay
constant of luminescence in a 1.5 x 10
-4
M CuBr
3
2−
solution.
27, 30
In 0.4 M nitrate, the fit
is ~ 600 ps, which works out to a ~ 4 x 10
9
M
-1
s
-1
rate constant. To our knowledge, the
quenching rates of UV excited CuBr
3
2-
transients by NO
3
−
are not known.
35

a b
300 350 400 450 500 550
0
2
4
6
8
10
12
14
16
500 ps
200 ps

mOD
Wavelength, nm

300 350 400 450 500 550 600
0
5
10
15
493 ps
204 ps

mOD
Wavelength, nm

c d
300 350 400 450 500 550 600
2
4
6
8
10
12
492 ps

mOD
Wavelength, nm
202 ps

300 350 400 450 500 550 600
0
5
10
15
493 ps
204 ps

mOD
Wavelength, nm

Figure 2.4.5. a,b) Spectral cuts at ~ 200 ps (black traces) and ~ 500 ps (red traces) with different
quenchers. Data normalized using absorption strength at early delays following 2PA. The
discrepancy in the 290 spectral region between the unquenched data sets in (a) and (b) could result
from the data not taken back to back, thereby introducing experimental variations in the pump-probe
spatial overlap. a) CuBr
3
2−
with 0.2 M H
+
. b) CuBr
3
2−
with 0.4 M NO
3
-
. c,d) Spectral cuts with no
quenchers (unnormalized) taken under same conditions as acidic (c) and nitrate (d) solutions.

36
2.4.5 Quenching of Transient Visible Band 2

Band 2 completely grows to its peak absorption by 1 ps and then begins to slowly
decay over the next few hundred ps (see Figure 2.2.4b above). This spectral behavior is
consistent with formation of the solvated electron
17, 45
and known electron quenchers, H
+

and NO
3
−
, confirm the assignment this band since diffusive quenching rates agree with
the bimolecular rate constants for Reactions 2.2 and 2.3.
46, 47

H
+
+ e
-
 H· (2.2)

NO
3
−
+ e
−
 NO
3
2−
(2.3)

Figure 2.4.6a shows that Band 2 for 266 nm excited CuBr
3
2−
with a [HBr] of 0.2
M virtually decays to zero by 500 ps. The 700 nm kinetics for the acidic sample can be
fit with an additional exponential decay of 217 ps, in agreement with the H
+
electron
quenching rate constant of 2.3 x 10
10
M
-1
s
-1
.
48
Similarly, the kinetic fit of CuBr
3
2−
with
0.4 M KNO
3
−
contains an additional exponential decay of 258 ps, which is consistent
with the NO
3
−
scavenging rate constant of 9.7 x 10
9
M
-1
s
-1
.
48
Time cuts and fits are
shown in Supplement 2A.
37

a b
400 450 500 550 600 650 700
0
2
4
6
8
10
12
14

100 ps
50 ps
500 ps
mOD
Wavelength, nm
2 ps

300 400 500 600 700
0
2
4
6
8
100 fs
200 fs
300 fs
500 fs
1 ps
2 ps
100 ps
150 ps
450 ps

mOD
Wavelength, nm

c
350 400 450 500 550 600
0
1
2
3
4
5

mOD
Wavelength, nm
pre time 0
500 fs
1 ps
2 ps
5 ps
20 ps
50 ps
100 ps
300 ps

0 100 200 300 400 500
0
1
2
3
4
5
6
7
0 2 4
0
2
4

mOD
Delay Time, ps

360 nm H
2
O
600 nm H
2
O
600 nm Br
-
mOD
Delay Time, ps
360 nm Br
-

Figure 2.4.6. a) Spectral cuts of 266 nm excited aqueous CuBr
3
2−
with HBr at [H
+
] = 0.2. b) Spectral
cuts of transient absorption for 266 nm excited neat water. c) Broadband transient absorption of 267
nm excitation of 0.4 M Br
−
aqueous solution from pre-time = 0 to 300 ps. There is a prominent band
centered at 360 nm that grows in on a ~ 100 ps timescale. This peak is similar to the absorption
spectrum of Br
2
●−
in aqueous solution and the growing in time constant corresponds to the bimolecular
Br
●
+ Br
−
reaction rate to form Br
2
●−
. d) Kinetic trances for the Br
−
solution and neat water. The 600
nm trace for neat water and Br
−
solution have the same dynamics, as also seen at early times (inset).
The 360 nm kinetic traces show a rise for the Br
−
solution while water is level on this mOD scale.

The solvated electrons probed in the 500 nm and longer region could originate
from the copper complex or from the medium (water and excess bromide ions). The
following assesses the contribution from both. The transient signal from neat water
38
(Figure 2.4.6b) also shows a broad matching visible band, suggesting that much of Band
2 solvated electrons in the CuBr
3
2−
data is generated from two-photon ionization of the
water. High laser intensity of the 1 kHz repetition rate laser system produces solvated
electrons from water via two-photon absorption of 266 nm photons.
49
At wavelengths
longer than 900 nm, it was assumed that the largely dominant absorbing transient species
is the solvated electron. The 950 nm time cuts (Supplement 2A) show that neat water
and CuBr
3
2−
solutions give rise to almost the same transient absorption intensity. These
experiments were done back to back under the same experimental conditions. This
prompted us to later collect transient broadband signal from 266 nm excited aqueous Br
−

at 0.4 M concentration. This experiment was done at a different time from the data
presented in Figures 2.4.6a and b. Results of the Br
−
experiments are shown in Figure
2.4.6c and d. A water scan was taken back to back with the Br
−
solution under the same
experimental conditions and it can be seen that 0.4 M aqueous Br
−
yields a solvated
electron signal 1.7 times greater than water alone. As both water and Br
−
are ionized
with 2 photons, this suggests that the ultrafast e
−

(aq)
signal in the CuBr
3
2−
experiment is
almost entirely due to the environment and not the copper complex.

2.4.6 Low Intensity Experiments

Based on the results of the previous section, it was necessary to do an experiment
to eliminate the background electrons produced from ionization of the medium in order to
quantify whether there is any prompt electron ejection from CuBr
3
2−
itself. Single probe
39
a b
-1 0 1 2 3 4 5
0
10
20
30
40
50
60
70

Lock-in signal ( V)
Delay time (fs)

0 100 200 300 400 500
0
10
20
30
40
50
60
70

Lock-in signal ( V)
Delay time (fs)

Figure 2.4.7. Two color pump-probe 266 nm excitation, 500 nm and 900 nm probes taken on 250 kHz
repetition rate laser system. Black trace: 900 nm detection of neat water. Blue trace: 900 nm of
aqueous CuBr
3
2−
. Green trace: 266 nm/500 nm of CuBr
3
2−
for comparison. No signal was detected
from 0.4 M Br
−
solution (not shown). (a) 0-5 ps. (b) 0-500 ps

wavelength transient absorption experiments were done using a 250 kHz repetition rate
laser system to reduce or completely eliminate background solvated electrons originating
from two-photon ionization of water or the large free bromide ion concentration (~ 0.4
M). This is effective because the higher repetition rate laser system has considerably
lower pump pulse intensity, by a factor of ~ 1000. Since photo-ionization of both water
and Br
−
is a two-photon process and scales as pump intensity squared and CuBr
3
2−

resonant excitation is a one-photon process and linear with pump intensity, lowering the
pulse intensity will reduce electron ejection from both water or the free bromide to a
much larger extent than the excitation of CuBr
3
2−
. Single color probe results are plotted
in Figures 2.4.7a and b. The excitation intensity was reduced to 1.9 x 10
8
W/cm
2
until
within the noise negligible signal was detected from neat water or from an aqueous
40
solution of 0.4 M Br
−
. The 250 kHz pump-probe spectrometer is not capable of
broadband probe detection, and therefore we collected single wavelength probe
absorption at 500 and 900 nm for comparison.
As seen in Figure 2.4.7a, the prompt lock-in amplifier signal for a 500 nm probe
is approximately 5 times higher than the signal with a 900 nm probe. We can take into
account the different intensity of the two probe colors and the different photodetector
responsivities at 500 nm and 900 by use of a measured ratio of ~ 1.2, determined as the
lock-in signal when chopping the whole probe beam (61 µV : 57 µV, 500 nm : 900 nm).
This ratio and the known solvated electron absorption band
50
allows us to obtain an
approximate QY of prompt electron detachment from all copper species, which is
determined to be < 5% by fitting the spectral data with the global analysis method
discussed in Section 2.5.4. In this fitting, an amount of electrons were added into the fit
to produce an absorption ratio of 1.2. This left a difference in absorption at 500 nm that
must be due to the copper transient. Based on the results for neat water and Br
−
solutions
and also the low intensity experiments, there are three sources of solvated electrons in our
broadband transient data, one being a very small fraction from a bromocuprate complex,
and the other two from the solvent and free Br
−
. In supplement 2b we will consider
whether this small prompt electron yield can be assigned to the minor CuBr
2
-
population
in the solution.

41
2.5 Discussion

2.5.1 Reported Transients on the Nanosecond Timescale

The transient absorption from our ultrafast broadband experiments resembles that
reported by Stevenson and Horvath, and coworkers, from laser flash photolysis as shown
in Figure 2.5.1.
27
The flash photolysis experiments contain a transient band at ~ 350 nm
and visible band of similar absorption intensity assigned to the solvated electron. The
biexponential decaying ~ 350 nm band was assigned to three species, the CuBr
3
2−
triplet,
Br
2
●−
, and Cu(0). The excited state evolution and chemistry in our experiments should
lead to the same photoproducts present at nanoseconds following excitation, barring any
experimental differences that can prompt unique pathways. While we also assign the
visible band (Band 2) in our own experimental data to the solvated electron, we have
found that only a small fraction of solvated electron originates from prompt photoejection
from CuBr
3
2−
. The majority is from 2 photon ionization of water and Br
−
due to our high
laser pulse intensity. 2PA ionization would not occur in a flash photolysis experiment,
and actually, the electrons in the ns experiments were reported to originate from ejection
of the CuBr
3
2−
triplet species. As this triplet has a ns decay lifetime (further discussed
below) the triplet ejected electrons would not be observed at our maximum measured
delay time of 500 ps. The authors also report a significant fraction of prompt ejection
from CuBr
3
2−
UV excitation. Because we are not seeing clear evidence for electrons
ejected from the copper complex inside 500 ps, we suspect the occurrence in their
experiments could be a consequence of a bi-photonic process in which an evolved state
42
absorbs a second pump photon within the duration of the pump laser pulse, as they have
described for a similar system (copper(I) cyano-halide) in a subsequent report.
26
Since
our pump pulses are very short (~ 50 fs), we would not induce such a process.

300 400 500 600 700 800 900
0.00
0.05
0.10
0.15
0.20
0.25

Absorbance
Wavelength, nm

Figure 2.5.1. Transient absorption results from nanosecond laser flash photolysis of aqueous 0.1 mM
CuBr
3
2−
in 5 M [Br
−
] at 50 ns delay time. Excitation with 7 ns 266 nm laser pulse. Data reproduced
from Reference 27.

UV excitation of CuBr
3
2−
induces luminescence, which was assigned to the triplet
species because of the ~ 200 nm emission red shift and long lifetime.
27, 30, 32
We have
confirmed the presence of this triplet species in our transient absorption time window
using TCSPC, described in detail in Chapter 4. Since the lifetime of the triplet state is
longer than our experimental time window, decay of its spectrum in our transient data
43
will not be observed. Our TCSPC measurements (chapter 4) have also confirmed that the
triplet species is formed within ~ 8 ps.
The slower decaying species in the UV band of the flash photolysis data was
assigned to Br
2
●−
, formed in several ways as described by Stevenson, et. al.
27, 34
The first
is by the reaction of an excited copper(I) species with Br
−
ion by the following
mechanisms.

Br
●
+ Br
−
 Br
2
●−
(2.4)

Br
●
radical is formed upon fast electron ejection from the Br
−
ligand instead of ejection
from the copper center, resulting in a Br
●
dissociating from the CTTS state,

*[CuBr
3
2−
] → CuBr
2
−
+ Br
●
+ e
−
(aq)
CuBr
3
2−
(2.5)

Since we do not have a significant amount of prompt electron detachment from copper,
the source of Br
●
from reaction 2.3 is ruled out in our own experiments. However, 2PA
of free Br
−
in solution would not only form prompt electrons, it would also provide Br
●

for reaction 2.2, and we do indeed see a contribution of Br
2
●−
in our transient data (Figure
2.4.6a). The other two sources of Br
●
and Br
2
●−
proposed by Stevenson, et. al.
34
include
the following two reactions, respectively, which would not occur on the timescale of our
experiments.

44

3
CuBr
3
2−
+ Br
−
→ CuBr
3
3−
+ Br
●
(2.6)

3
CuBr
3
2−
+ Br
−
→ CuBr
2
2−
+ Br
2
●−
(2.7)

The third species in the ns experiments is Cu(0), which is produced via electron
scavenging by ground state bromocuprates. Cu(I) electron scavenging occurs on a long
timescale and will not be observed before 500 ps, therefore it is not one of our transient
species. Thus far, in this Discussion, we acknowledge the presence of three species in
our broadband data that are also in the ns flash photolysis data, the CuBr
3
2−
triplet state,
Br
2
●−
which is formed independent from CuBr
3
2−
UV excitation, and the solvated
electron, most of which is independent from excitation of CuBr
3
2−
.

2.5.2 UV Band 1 Spectral Evolution

We now discuss the spectral dynamics observed in our ultrafast experiments.
Important observables for the assignment of the transient bands include small
photodetachment originating from copper, an early isosbestic point, mid time 400 nm
band, long time isosbestic point, UV acid and nitrate quenching, oscillatory beats
(detailed in Chapter 3), and luminescence lifetimes and quenching (detailed in Chapter
4). Despite the complexity of the transient spectra, the most plausible kinetic scheme is
proposed and the unusual nature of the CTTS state of CuBr
3
2−
will also be discussed.
Since the visible Band 2 has already been assigned to the solvated electron and it has
45
been established that only a small fraction of electrons actually originate from prompt
photoejection from copper complexes, we will save its discussion for later in this Section.
Band 1 contains a wealth of information and will be explored first.
As mentioned in Section 2.1, since copper(I) and Br
−
are both closed-shell,
neither MLCT or LMCT transitions are accessible with UV excitation and calculations
show that electronic transitions for copper halides are above ~ 4.5 eV,
12, 51, 52
leaving only
one CuBr
3
2−
excited electronic state that can be populated by absorption of a UV photon.
This state is the CuBr
3
2−
CTTS state in Scheme 1 (Figure 2.2.1).
27, 32
If a spin flip occurs
the CuBr
3
2−
triplet forms. Aside from the excited electronic states present (CTTS states),
other transient species can exist if detachment, dissociation, or bimolecular chemistry
occurs.
By inspection, the main spectral features of Band 1 can easily be
compartmentalized into three time segments, labeled as Early, Mid, and Long time.
Early time spans from 0 fs to 500 fs and includes the 2PA and isosbestic point, Mid time
includes the growing in of a band centered near 400 nm and from 500 fs to 2 ps, and
Long time is from 2 ps to the end of the experimental time window, 500 ps. The Long
time dynamics include blue shifting and a change in the spectral structure of Band 1.
1. Early Time, 0 – 500 fs. A positive absorption signal following the end of the
2PA spike at ~ 60 fs indicates that a transient species is prepared within the time of the
pump laser pulse (see Figure 2.4.4a, above). Not considering electron photodetachment,
the only transient species that can be fully prepared within the pulse duration are the
tribromocuprate CTTS singlet state (denoted as
*
[
1
CuBr
3
2−
]), or products of fast
unimolecular photodissociation. Very little spectral contribution from water background
46
before 100 fs following excitation is seen in the neat water data shown in Figure 2.4.6b,
which means the solvated electron and H
2
O
+
/OH products of photolysis are ruled out.
The product of prompt photodetachment from Br
−
, Br
●
, absorbs at 275 nm.
53

Direct dissociation of a Br
−
from CuBr
3
2−
ligand would probably occur on the
time scale of half a Cu – Br vibrational period. Our calculated asymmetric stretch
frequencies for Cu
II
Br
3
−
(an approximation for the CuBr
3
2−
CTTS state) are 274 and 256
cm
-1
.
12, 54-56
The average of the two, ~ 265 cm
-1
, corresponds to a period of ~ 125 fs with
half a period > 60 fs. Therefore the species formed within the pulse duration cannot be a
dissociation product of CuBr
3
2−
as we would expect the transient signal to drop much
closer to zero than is observed. While there is no absorption spectra reported for the
*
[
1
CuBr
3
2−
] state, the electronic structure of the excited Cu(I) center can be approximated
as Cu(II) because the CTTS electron resides far from its underlying CuBr
3
−
. With this
approach, we might expect similar absorption bands as Cu
II
Br
2
, for which a band is
reported near ~ 530 nm.
12, 57
The CTTS state,
*
[
1
CuBr
3
2−
], especially since it has already
been reported as the initial excited state species,
27, 32
is the most likely candidate.
Another observable in the Early time regime is the ~ 420 nm isosbestic point as
seen in Figure 2.5.2. The species on the red side of the ~ 420 nm isosbestic point is the
one prepared during the excitation laser pulse,
*
[
1
CuBr
3
2−
]. The species that forms on the
blue side of the isosbestic point must be directly populated by the CTTS state. Three
scenarios are considered. The first is Br
−
ligand dissociation to form CuBr
2
−
, the second
is fast ISC to form the CuBr
3
2−
triplet state, and lastly, a concerted dissociation along
with formation of Br
2
●−
.

47
300 400 500 600
2
4
6
8
10
mOD
Wavelength, nm
Delay time, fs
84
94
104
114
124
134
144
154
164
174
i
ii
300 400 500 600
2
4
6
8
10
mOD
Wavelength, nm
Delay time, fs
84
94
104
114
124
134
144
154
164
174
i
ii

Figure 2.5.2. Early isosbestic point as shown in Figure 2.4.3a above. There must be two species
contributing; they are labeled i and ii. i is assigned to the CuBr
3
2−
CTTS state. The most likely
assignment for species ii is the dibromo- product of Br
−
dissociation.

It is likely that the dissociation product, CuBr
2
−
, would also be a CTTS state
(denoted
*
[
1
CuBr
2
−
]). Electronic structure calculations for copper halide coordination
complexes have shown that negative charge is centered on the ligands, rather than across
the metal complex.
52
It is reasonable to assume that the ligand leaves in this same form,
and the electron residing in the CTTS orbital remains. If a Br radical departed the
complex with the CTTS electron, the product would be ground state CuBr
2
-
. This species
has an absorption band at 280 nm and no absorption further to the red. Furthermore, the
resulting CuBr
2
-
CTTS state is a singlet because ISC would not have occurred by this
time. If Br
●
, and not Br
−
is dissociating from the complex, then the result would be as in
48
Equation 2.3 and a solvated electron will accompany the reaction, which has already been
ruled out.
The second process considered is the initially prepared CTTS state undergoing
ISC on the timescale of the isosbestic point. Specific transition metal coordination
compounds with bulky ligands have been shown to undergo ultrafast intersystem crossing
(ISC) following metal-to-ligand-charge-transfer (MLCT) or ligand-to-metal-charge-
transfer (LMCT). Chen and coworkers have studied the photodynamics of copper(I)
phenanthrolines and reported timescales for T  S intersystem crossing and structural
rearrangement.
2
The ISC of these copper complexes occurs on ~15 ps timescale.
Although their system has large bulky ligands as opposed to our halide ligands, ISC for
the CuBr
3
2−
system should be on a comparable timescale. There are two other reasons
rule out the CuBr
3
2−
triplet (
3
CuBr
3
2−
). The reported absorption spectrum for the triplet
state is shown in Figure 2.5.3 and has a band maximum of ~ 390 – 400 nm. It is unlikely
that the triplet would form at 350 nm and then red shift to 400 nm. The oscillatory signal
prominent in Figure 2.4.4b also helps to refute the assignment of the blue side of the
isosbestic point to the CuBr
3
2−
triplet state. Oscillatory transient signal is the main topic
of Chapter 3 and details can be found therein. These oscillations are assigned to the
CTTS singlet and triplet species, and it can be shown that if the species growing in at 350
nm is oscillating, the isosbestic point will not exist.
For Br
2
●−
to form fast enough to take part in the isosbestic point, there must be a
concerted dissociation and formation of Br
2
●−
leaving the CuBr
3
2−
CTTS state by the
reaction

49
*
[
1
CuBr
3
2−
]  CuBr + Br
2
●−
(2.8)

The absorption spectrum of aqueous Br
2
●−
is reported to have a band maximum at 350
nm in early publications (1950s – 1960s),
58, 59
and absorption maxima between 360-370
nm in more recent articles (1970s – 2000s).
60-63
Peak extinction coefficients reported in
these references are varied, 7800, 9600, 1200, and 35000 M
-1
s
-1
. We also observe Br
2
●−

growing in on a diffusive timescale in our Br
−
experiments and show that it has an
absorption band centered at 360 nm (Figure 2.4.7). Other than the slight discrepancy in
peak absorption, there are several reasons to rule out the fast formation of Br
2
●−
.
As will be shown in the Long time dynamics, the species that grows in on the blue
side of the early isosbestic point must decay on the ~ 100 ps timescale, and the species it
populates must have an absorption band close to 340 nm (see Figure 2.5.5). The known
decay reactions for Br
2
●−
are the following:

Br
2
●−
+ Br
−
 Br
3
−
(2.9)

Br
2
●−
+ e-  2Br
−
(2.10)

Br
2
●−
+ H
●
 H
+
+ 2Br
−
(2.9)

Br
2
●−
+ CuBr
3
2−
 Br
−
+ CuBr
4
2−
(2.10)

50
Br
3
−
has an absorption band centered at 270 nm
60
and Br
−
absorbs even further in
the UV,
64
outside of our transient probe window, ruling out equations 2.7, 2.8, and 2.9.
The CuBr
4
2−
absorption spectrum has a band at 655 nm, some bands in the IR, and one at
350 nm (28570 M
-1
cm
-1
)
65
, however, Stevenson and coworkers have claimed to see Br
2
●−

as one of their transients in which they see decay without anything else rising in the UV
or visible.
27

Another reason Br
2
●−
is unfavored is because other transition metal complexes
ligated to halides do not form the dihalide radical upon dissociation. Gas phase
photodissociation of IrBr
6
2−
probed via photodissociation action spectroscopy does not
yield Br
2
●−
radical and therefore the reaction pathway,

IrBr
6
2−
+ hv  IrBr
4
−
+ Br
2
●−
(2.11)

was ruled out as a possible scheme.
66
Gas phase time-resolved photoelectron
spectroscopy showed that the IrBr
6
2−
dissociates to release Br
−
and does not form Br
2
●−
.
67

Photolysis of PtBr
6
2−
in a methanol matrix showed no evidence for formation of Br
2
●−
in
ns flash photolysis experiments and Br
3
●−
was also not a possible product.
68
As a last
example, Firedrich, et. al. performed gas phase photodissociation on IrX
6
2−
, (X = Cl and
Br), and saw no formation of X
2
●−
radicals.
69

It must be noted that quantum beating of the initially prepared CuBr
3
2−
CTTS
state could appear as an isosbestic point in the transient spectral data. Depending on the
excited state surfaces, vibrational coherences may appear as spectral oscillations. We
indeed observe oscillatory signal, which is discussed in detail in Chapter 3, although
51
doubt that a vibrational wavepacket is the contributor to the isosbestic point since it is not
observed in a simple oscillating band model (not shown). However, a more complete
analysis needs to be done in order to confidently rule out quantum beating.
2. Mid Time, 500 fs – 2 ps. As the tribromocuprate CTTS triplet,
3
CuBr
3
2−
, has an
absorption band that peaks near ~ 390 nm (Figure 2.5.3), it is a likely candidate for the
spectral feature at Mid time, which has already been established by TCSPC to be in our
time window. The CTTS triplet is not a new idea. Theory often utilizes the triplet CTTS
electronic state for simplification of calculations and simulations in clusters and in the
bulk.
70
However, to our knowledge, the triplet CTTS has never been experimentally
accessed in other systems since the intersystem crossing between singlet and triplet CTTS
states cannot compete with the usually fast depopulation of the singlet CTTS via electron
ejection. Figure 2.5.4 shows spectral cuts for Mid time.
300 325 350 375 400 425 450
0.0
0.1
0.2
0.3
0.4
0.5
0.6

Absorption
Wavelength, nm

Figure 2.5.3. Reproduced CuBr
3
2−
triplet spectrum from Stevenson, et. al.
27
obtained from 266 nm
laser flash photolysis of CuBr
3
2−
in aqueous 5.0 M [Br
−
]. Spectrum was deconvoluted from
absorption of other transient species at 50 ns, including Br
2
●−
.

52
iii
300 400 500 600
2
4
6
8
10
mOD
Wavelength, nm
Delay time
204 fs
504 fs
754 fs
1 ps
1.25 ps
1.5 ps
2 ps
iii
300 400 500 600
2
4
6
8
10
mOD
Wavelength, nm
Delay time
204 fs
504 fs
754 fs
1 ps
1.25 ps
1.5 ps
2 ps

Figure 2.5.4. Spectral cuts as shown in Figure 2.4.3b above. Mid time species labeled as iii is the
triplet CuBr
3
2−
state.

3. Long Time, 2 – 500 ps. Spectral shifting on the timescale of ~100 ps is an
indication of one species decaying into another species with significant spectral overlap.
In addition to spectral blue shifting, Band 1 also increases in absorption intensity, and so
the species growing in must have a higher extinction coefficient than the one decaying.
Aside from solvated electrons, the singlet dibromocuprate CTTS (
*
[
1
CuBr
2
−
]) and triplet
tribromocuprate CTTS (
*
[
3
CuBr
3
2−
]) are the only species present at 2 ps (species i and ii).
*
[
3
CuBr
3
2−
] decays on the nanosecond timescale and therefore cannot be losing
population within 500 ps. This leaves
*
[
1
CuBr
2
−
] as the only candidate for the decaying
species. Since the excited state scheme reported for the dibromocuprate manifold mirrors
53
that of the tribromocuprate,
32
and one available assignment is ISC of
*
[
1
CuBr
2
−
] into the
triplet state,
*
[
3
CuBr
2
−
], yet we do acknowledge that ~ 100 ps ISC rate is much slower
than the ISC rate for tribromocuprate. We also considered the two following reactions,

*
[
1
CuBr
2
−
] + H
2
O  [CuBr
2
(H
2
O)]
−

(2.12)

*
[
1
CuBr
2
−
] + Br
−
 CuBr
3
2−
(2.13)

Reaction 2.12 is the association of water with the Cu(I) complex. The product may be
ground or excited state. As an argument for aquation, a description of Cu(I) as an atom is
as follows: It is known that the d
10
s
0
electron configuration of Cu(I) has a different
behavior than transition metals with vacancies in the d-shell orbitals. It results that Cu(I)
can exist in linear, trigonal, or tetrahedral coordination,
71
or even coexist in a mixture of
these in a solid since the energy barriers between the different coordination states are
relatively low.
72
At room temperature and pressure, aquated Cu
+
ions are unstable and
becomes Cu(s) and Cu
2+
, however, the Cu(I) state can be stabilized by aquation.
73
This
suggests that association of water with the excited state CuBr
2
−
is possible; however we
cannot support aquation without a more relevant example. Reaction 2.13 is not likely
because it would have to form an excited state CuBr
3
2−
, which is either the singlet CTTS
or the triplet CTTS. If it forms a ground state CuBr
3
2−
, there wouldn’t be a transient
absorption at ~ 330 nm, but rather a bleach recovery at 266 nm. In fact, all of the
transient copper species must be excited states, since the ground state absorption of
54
CuBr
3
2−
and CuBr
2
−
are at ~ 280 nm. Figure 2.5.5 shows the spectral cuts from 2 ps to
472 ps and is labeled with species ii and iv.

300 400 500 600
2
4
6
8
10
12
mOD
Wavelength, nm
Delay time, ps
2
52
102
202
302
402
472
ii
iv
300 400 500 600
2
4
6
8
10
12
mOD
Wavelength, nm
Delay time, ps
2
52
102
202
302
402
472
ii
iv

Figure 2.5.5. Spectral cuts from 2 ps to 472 ps of same data set shown in Figures 2.5.2 and 2.5.4. The
spectral shifting and rise on the blue side is indicative of one species decaying while directly
populating another species with a higher extinction coefficient.

Table 2.5.1 summarizes the rise and decay kinetics of transient absorption across
the experimental time window of < 100 fs to ~ 500 ps. The first two species contribute to
the early isosbestic point in the τ1 time region, which are the CuBr
3
2−
and CuBr
3
2−
CTTS
states. τ2 is the formation of the CuBr
3
2−
triplet state, which has a broad absorption that
stretched across the UV and visible, confirmed by the work of Stevenson and
55
coworkers.
27
τ3 is the intersystem crossing from the singlet dibromo CTTS state,
*[
1
CuBr
2
−
], to the dibromo triplet state *[
3
CuBr
2
−
].

Table 2.5.1. Approximate timescales of dynamics observed in the broadband data
labeled in Figures 2.5.2, 2.5.4, and 2.5.5. R = rise, D = decay. Each of the time segments
are labeled as τ1, τ2, τ3, τ4.

Time Component (ps) UV 400 nm Vis
Fast, τ1, 70 fs – 150 fs R R D
Mid, τ2, 150 fs – 1 ps R R R*
Long, τ3, 2 ps – 500 ps R D D

*There is a slight decay in the visible region following the rise across the spectral window. This will
be addressed following in this Section.

2.5.3 Kinetic Model

In summary of the spectral features presented in Section 2.5.3, some main points and
assumptions relevant to the analysis of the transient data are the following:
1. The broad absorption following the 2PA spike is the *[
1
CuBr
3
2−
] CTTS state,
which loses a Br
−
ligand in a unimolecular dissociation reaction to form CuBr
2
−
. CuBr
2
−

cannot be in the ground state and therefore we assign it as an excited state, the CTTS
singlet, *[
1
CuBr
2
−
].
2. The formation of the *[
3
CuBr
3
2−
] CTTS occurs within ~ 8 ps, and actual
growing in of its spectral signature is observed as early as 200 fs. Since care was taken to
56
keep the solution under nitrogen in the 1 kHz pump-probe experiments, decay of the
tribromocuprate triplet will not be observed in our broadband experimental time window.
The small ~ 520 nm decay simultaneous with continued growing in of the triplet state is
the loss of singlet CTTS population upon ISC.
3. The Long time continued decay of the red edge of the UV band into a 330 nm
absorbing species is the ISC from the singlet dibromo CTTS, *[
1
CuBr
2
−
] to the dibromo
triplet, *[
3
CuBr
2
−
].
4. It is assumed that all excited state dynamics observed in the transient data is
resultant from absorption by ground state CuBr
3
2−
and not CuBr
2
−
since the
tribromocuprate – dibromocuprate equilibrium in the solutions strongly favors the
tribromocuprate species.

Figure 2.5.6. Scheme 2.2. Simplest kinetic scheme describing the UV band dynamics as presented in
Section 2.5.3. It does not include any solvated electron. Not plotted on an energy scale. i is tribromo-
*[
1
CuBr
3
2−
], ii is dibromo- *[
1
CuBr
2
−
], iii is tribromo- *[
3
CuBr
3
2−
], and iv is dibromo- *[
3
CuBr
2
−
].
The early isosbestic point is associated with τ1, the ISC to the tribromo triplet is τ2, and τ3 is the
dibromo CTTS ISC from singlet to triplet.
Ground

τ2
τ1
i
τ3
ii
iii
iv
57
Although the data proves to be complex, a simple kinetic model that fits our experimental
data has been assembled and is shown in Scheme 2.2 in Figure 2.5.6 with species i to iv
and the three time constants, τ1, τ2, and τ3.
5. The solvated electrons are treated as independent pathways by just adding a fit
to the spectral evolution of the aqueous Br
−
solution (background). This accounts for
both the electrons originating from water and Br
−
. The background fit also includes the
products of Br
−
detachment, which are Br
●
and diffusion limited formation of Br
2
●−
.

2.5.4 Global Analysis

Global analysis is a constrained fitting procedure that uses assigned absorption
spectra for each transient species and is a powerful tool for deconvolution of complex
transient data.
74, 75
The transient absorption at any given probe delay time is the sum of
all transient species present that absorb in the spectral window. The absorption as a
function of time and wavelength is given in the following equation




n
j
j j
t c t TA
1
), ( ) ( ) , (   
(2.14)

where TA is the transient absorption, c
j
(t) is the concentration of the j
th
species at time t,
and ε
j
( λ) is the extinction coefficient of the j
th
species at wavelength λ. The procedures of
58
global analysis are as follows. The change in population of each transient species can be
written by the following equation.




n
j
j ij
i
t c k
dt
t dc
1
) (
) (
(2.15)

where c
i
(t) is the concentration of the i
th
species at time t, k
ij
is the rate constant for the i
th

species transferring population to the j
th
species. These are called concentration profiles
and are calculated based on the rate constants for the given kinetic model. Absorption
spectra are then assigned to each species. The transient absorption can be found at any
given wavelength and delay time and the global analysis results are compared to data.
Parameters can be adjusted to better fit the data; however, keeping known rate constants
and known absorption spectra fixed produces more meaningful results. The following
parameters were held constant in our global analysis.

 Absorption spectrum of triplet approximated from Stevenson, et. al. shown in
Figure 2.5.3
27

 Upper limit for the appearance time of *[
3
CuBr
3
2−
] of ~ 8 ps
 All copper transient spectra have similar extinction coefficients
 Gaussian absorption bands in energy domain (cm
-1
)

59
In the kinetic model, there are four transient species and three rate constants. As
another starting point, the rate constants are estimated based on observation of the
transient broadband data. Free fitting to exponential rises and decays were attempted,
however, because of the complexity of the system, it was best to rely on the global
analysis to produce the rate constants. The absorption spectra for each of the four species
were estimated based on difference spectra taken from the data (shown in Supplement
2.1). Difference spectra reveal the change in absorption between two selected delay
times.
The global analysis was initiated with 100% of the population as *[
1
CuBr
3
2−
] CTTS
state. The parallel Br
−
and water background kinetics were kept independent of the
CuBr
3
2−
kinetics. Kinetic rate constants and absorption spectra of species i to iv, aside
from the spectral shape of the triplet
3
CuBr
3
2−
, were adjusted according to inspection of
the spectral simulations in comparison to the transient data. The solvated electron, Br
●
,
and Br
2
●−
spectra are all known and were fixed (Figure 2.5.7a). After making
adjustments to the population kinetics and to the spectra for Species i – iv, a final set of
absorption spectra that best fit the data was obtained (Figure 2.5.7). Extinction
coefficients are relative since the absolute concentration of each species in solution is not
known. As seen in Figure 2.5.7, another species needed to be added to the scheme to
hold the kinetic partition for reasonable population of the triplet state (red trace) and so
four rates were actually obtained. The final rate constants are listed below and the
resulting concentration profiles are shown in Figures 2.5.8a and b. The spectrum and
evolution of the solvated electron obtained by fitting the Br
−
data, independently added
into the global analysis, is shown in Figure 2.5.9b.
60

τ1 = 85 fs
-1

τ2a = 200 fs
-1
τ2b = 850 fs
-1

τ3 = 180 ps
-1

iii
i
ii
iv
500 600 400 300
700
Wavelength, nm
Arb.
iii
i
ii
iv
500 600 400 300
700
Wavelength, nm
Arb.

Figure 2.5.7. a) Spectra for each transient species, i: CuBr
3
2−
CTTS (blue), iii: CuBr
3
2−
triplet (red),
ii: CuBr
2−
CTTS (green), iv: CuBr
2−
CTTS triplet (magenta). Smaller blue trace is the additional
compartment needed for best global analysis results.

61
0
0.2
0.4
0.6
0.8
500 0
0
1
0.4
0.8
0.6
0.2
1000 5000
Time, fs
Fraction Population
iii
i
ii
0
0.2
0.4
0.6
0.8
500 0
0
1
0.4
0.8
0.6
0.2
1000 5000
Time, fs
Fraction Population
iii
i
ii

0
0.2
0.4
0.6
0.8
iv
100 0
0
1
0.4
0.8
0.6
0.2
Time, ps
Fraction Population
iii
i
ii
200 300 500 400
0
0.2
0.4
0.6
0.8
iv
100 0
0
1
0.4
0.8
0.6
0.2
Time, ps
Fraction Population
iii
i
ii
200 300 500 400

a b
Figure 2.5.8. Concentration profiles for each of the transient species obtained by global analysis. a)
From t = 0 to 5 ps. b) From t = 0 to 500 ps.

a b
0
3
4
4 e −
(aq)
Br ●
500 600 400 300
700
Wavelength, nm
ε, M
-1
cm
-1
Br2 ● −
0
5000
10000
15000
0
3
4
4 e −
(aq)
Br ●
500 600 400 300
700
Wavelength, nm
ε, M
-1
cm
-1
Br2 ● −
0
5000
10000
15000

500 600 400 300
1
4
3
2
0
10 ps
100 ps
400 ps
300 ps
Wavelength, nm
mOD
500 600 400 300
1
4
3
2
0
10 ps
100 ps
400 ps
300 ps
Wavelength, nm
mOD

Figure 2.5.9. a) Absorption spectra of Br
●
(Ref
53
), Br
2
●−
(Ref
63
), and the solvated electron. b) Fits
overlaid with the transient absorption data for control aqueous Br
−
solution using the known solvated
electron spectrum and kinetics, and the spectra of Br
●
and Br
2
●−
. The formation time of Br
2
●−
from the
fit is 150 ps
-1
.

62
Concentrations
500 600 400 300
Wavelength, nm
mOD
0
74 fs
104 fs
124 fs
154 fs
4
10
6
8
12
Concentrations
500 600 400 300
Wavelength, nm
mOD
0
74 fs
104 fs
124 fs
154 fs
4
10
6
8
12

1002 fs
1204 fs
1404 fs
1804 fs
500 600 400 300
Wavelength, nm
mOD
4
10
0
6
8
12
1002 fs
1204 fs
1404 fs
1804 fs
500 600 400 300
Wavelength, nm
mOD
4
10
0
6
8
12

500 600 400 300
Wavelength, nm
mOD
5
10
0
102 ps
200 ps
302 ps
442 ps
500 600 400 300
Wavelength, nm
mOD
5
10
0
102 ps
200 ps
302 ps
442 ps

Figure 2.5.10. Global analysis results and spectral cuts of experimental data. The global analysis
curves are offset (below) that of the actual data to show the features; without translation the global
analysis overlays the data. a) 74 – 154 fs. b) 1002 fs – 1804 fs. c) 102 ps – 442 ps.

63
Ground
τ2a, relax
τ1, – Br
−
*[
1
CuBr
3
2 −
]
τ3, ISC
*[
3
CuBr
3
2 −
]
*[
1
CuBr
2
−
]
*[
3
CuBr
2
−
]
τ2b, ISC
Ground
τ2a, relax
τ1, – Br
−
*[
1
CuBr
3
2 −
]
τ3, ISC
*[
3
CuBr
3
2 −
]
*[
1
CuBr
2
−
]
*[
3
CuBr
2
−
]
τ2b, ISC

Figure 2.5.11. Scheme 2.3 Kinetic scheme that includes the UV band dynamics in Scheme 2.2 and
also another compartment necessary to populate a significant amount of triplet state. This
compartment is not yet assigned. Not plotted on an energy scale.

The global analysis results and transient data are shown in Figures 2.5.10a – c.
They are offset by subtraction of 2 mOD to show the spectral features, yet the global
analysis results overlay the data well when not offset. It can be seen that our model is a
good representation of the broadband spectral data.

2.5.5 Nature of the CuBr
3
2−
CTTS State

The effort to understand CTTS states has focused mainly on the halide anions.
14

Atomic anions have been used as a model system since only an atomic core and the

64
solvent need to be considered, making them the simplest CTTS systems that can lend
insight to more complex polyatomic anions.
14
In the halide model, the CTTS state has a
quasi-bound electron with an orbital filling the first solvent shell of the parent atom.
15

Fluctuations in the solvent allow the electron to separate from the parent atom and reside
within the same solvent cavity, forming a contact pair. However, the behavior of
polyatomic and/or multiply charged anions deviates from the halides.
76
For example,
excitation in the CTTS band of ferrocyanide ejects an electron at a distance of five
solvent shells, bypassing the entire contact pair step of the halide CTTS mechanism.
77

The UV absorption band of ground state CuBr
3
2−
peaking at 279 nm has been
assigned as CTTS since resonant excitation in this band prompts oxidation of the Cu(I)
center and formation of a solvated electron.
27
The electronic transition is either a Cu
CTTS 3d
10
or Cu 4s
1
3d
10
transition with the Cu 4s
1
is a metal centered Rydberg state
and taken to be the same as a CTTS state.
25
The ≤ 5% prompt electron ejected from the
CuBr
3
2−
CTTS state is surprising given the fast and efficient detachment from
ferrocyanide.
77
Lifetimes for CTTS states of Br
−
, OH
−
, and [Fe(CN)
6
]
4−
have upper limits
of 200 fs,
15
150-300 fs,
78
and <170 fs,
77
respectively, based on appearance of solvated
electrons but it is suspected that the CTTS lifetime occurs below 100 fs. In other words,
observation of CTTS state lifetimes has only been measured indirectly by product
formation.
14

In other systems, bands that were once assigned as LMCT have been reassigned
to CTTS, as in the cases of iron and cobalt oxalate complexes reported by Rentzepis and
coworkers.
79
They have also reported that the CTTS state results in a very small electron
formation quantum yield, 0.05 for ferrioxalate
20
and 0.10 for trisoxalato cobaltate(III).
19

65
This later system was also reported to have a ~ 800 fs CTTS lifetime. Similar to our
CuBr
3
2−
compound, these metal complexes undergo dissociation that competes with the
electron ejection.
To further characterize the CuBr
3
2−
CTTS state nitrate was used to determine if
the CuBr
3
2−
CTTS state is spatially diffuse. Aside from a reaction controlled by
diffusion, nitrate quenches solvated electrons statically by the electron transfer reaction
between nitrate and the solvated electron precursor to form nitrite.
46

e
−
(precursor)
+ NO
3
−
 NO
3
2−
(2.16)

The quantum yield of quenching is dependent on the spatial extent of the precursor
electronic wavefunction. 0.4 M nitrate static quenching of the water solvated electron
precursor prepared by 2PA drops electron formation by approximately 50%. CTTS
states, traditionally also solvated electron precursors, have been shown to react with
nitrate statically. The iodide CTTS state is quenched by ~ 30% with 0.4 M NO
3
-
,
measured by decreased quantum yield of solvated electron formation.
42
Figure 2.5.12
shows the 950 nm probe kinetic traces for CuBr
3
2−
with and without 0.4 M nitrate. There
is approximately a 50% drop in signal following the 2PA spike. Yet the decrease of the
solvated electron formation cannot be deconvoluted from the huge contribution of Br
−

and water. As the nitrate was already shown to diffusively quench the CuBr
3
2−
triplet
state, and fortunately the CTTS state does have a band in our spectral window, we
checked whether or not the excited singlet state was also being quenched. Figure 2.5.9b
shows kinetic traces for 400 nm and 520 nm probe wavelengths. To ensure that both of
66
the CuBr
3
2−
solutions, with and without KNO
3
, had the same concentration of CuBr
3
2−
,
the transient data was normalized based on a fitting of the UV/Vis absorption spectra for
each solution (see Supplement 2A). Indeed, it can be seen in Figure 2.5.12b that the
CuBr
3
2−
CTTS state is being statically quenched.

a b
0 1 2 3 4 5
-2
0
2
4
6
8
10
12
With KNO
3
Without KNO
3

mOD
Time, ps
0.0 0.5 1.0 1.5
0
5
10
15
20

520 nm
400 nm
mOD
Time, ps
Figure 2.5.12. a) 950 nm probe of CuBr
3
2-
with (red trace) and without (black trace) 0.4 M nitrate. b)
400 nm probe and 520 nm probe for CuBr
3
2-
with (red and blue traces) and without (black and green
traces) 0.4 M nitrate. Data normalized according to UV/Vis absorption at 266 nm. (See supplement
2.1 for UV/vis).

2.6 Conclusions

It has been shown that UV excitation into the CTTS assigned absorption band of
ground state CuBr
3
2−
produces a complex excited state evolution and the broadband
transient data presents several dynamical features. The appearance time of the previously
reported CuBr
3
2−
triplet state was determined to be ~ 850 fs based on global analysis of
67
the transient broadband data. Surprisingly, there is only a small, ~ 5%, quantum yield of
solvated electron ejection from the copper complex. This was difficult to determine since
the water and bromide background produced such a large solvated electron signal,
however, the low intensity experiments showed that indeed only a very small fraction of
electrons originate from copper complexes and it is still unclear how much comes from
CuBr
3
2−
itself (see Supplement 2b). The singlet CTTS state appears to decay by ISC and
dissociation and not by loss of an electron. The triplet is responsible for most of the
electrons ejected in this system. Yet because the CuBr
3
2−
singlet CTTS is quenched by
nitrate, the CTTS electron can still be considered diffuse, although it is stable enough to
survive several hundred fs.

68
2.7 References for Chapter 2

1. E. I. Solomon, U. M. Sundaram and T. E. Machonkin, Chem. Rev., 1996, 96,
2563-2605.

2. G. B. Shaw, C. D. Grant, H. Shirota, E. W. Castner, G. J. Meyer and L. X. Chen,
J. Am. Chem. Soc., 2007, 129, 2147-2160.

3. N. Armaroli, Chem. Soc. Rev., 2001, 30, 113-124.

4. E. I. Solomon and M. D. Lowery, Science, 1993, 259, 1575-1581.

5. R. H. Holm, P. Kennepohl and E. I. Solomon, Chem. Rev., 1996, 96, 2239-2314.

6. D. G. Cuttell, S. M. Kuang, P. E. Fanwick, D. R. McMillin and R. A. Walton, J.
Am. Chem. Soc., 2002, 124, 6-7.

7. T. Gunaratne, M. A. J. Rodgers, D. Felder, J. F. Nierengarten, G. Accorsi and N.
Armaroli, Chem. Comm., 2003, 3010-3011.

8. A. Tsuboyama, K. Kuge, M. Furugori, S. Okada, M. Hoshino and K. Ueno, Inorg.
Chem., 2007, 46, 1992-2001.

9. Z. A. Siddique, Y. Yamamoto, T. Ohno and K. Nozaki, Inorg. Chem., 2003, 42,
6366-6378.

10. W. F. Fu, X. Gan, J. Jiao, Y. Chen, M. Yuan, S. M. Chi, M. M. Yu and S. X.
Xiong, Inorg. Chim. Acta, 2007, 360, 2758-2766.

11. W. L. Zou and J. E. Boggs, J. Chem. Phys., 2009, 130.

12. X. B. Wang, L. S. Wang, R. Brown, P. Schwerdtfeger, D. Schroder and H.
Schwarz, J. Chem. Phys., 2001, 114, 7388-7395.

13. O. Horvath, Stevenson, K. L., Charge Transfer Photochemistry of Coordination
Compounds, VCH Publishers, Inc., New York, 1993.

14. X. Y. Chen and S. E. Bradforth, Annu. Rev. Phys. Chem., 2008, 59, 203-231.

15. M. K. Fischer, A. Laubereau and H. Iglev, Phys. Chem. Chem. Phys., 2009, 11,
10939-10944.

69
16. A. Kammrath, J. R. R. Verlet, A. E. Bragg, G. B. Griffin and D. M. Neumark, J.
Phys. Chem. A, 2005, 109, 11475-11483.

17. J. A. Kloepfer, V. H. Vilchiz, V. A. Lenchenkov, A. C. Germaine and S. E.
Bradforth, J. Chem. Phys., 2000, 113, 6288-6307.

18. F. H. Long, H. Lu, X. L. Shi and K. B. Eisenthal, Chem. Phys. Lett., 1990, 169,
165-171.

19. J. Chen, H. Zhang, I. V. Tomov, X. L. Ding and P. M. Rentzepis, Proc. Nat.
Acad. Sci. U.S.A, 2008, 105, 15235-15240.

20. J. Chen, H. Zhang, I. V. Tomov and P. M. Rentzepis, Inorg. Chem., 2008, 47,
2024-2032.

21. K. L. Stevenson and J. H. Jarboe, J. Photoch. Photobio. A. Chem., 2002, 150, 49-
57.

22. K. L. Stevenson, P. B. Bell and R. E. Watson, Coord. Chem. Rev., 2002, 229,
133-146.

23. A. Horvath and K. L. Stevenson, Inorg. Chem., 1993, 32, 2225-2227.

24. K. L. Stevenson, J. H. Jarboe, S. A. Langmeyer and T. W. Acra, Inorg. Chem.,
2003, 42, 3559-3564.

25. A. Horvath, C. E. Wood and K. L. Stevenson, J. Phys. Chem., 1994, 98, 6490-
6495.

26. A. Horvath and K. L. Stevenson, Coord. Chem. Rev., 2000, 208, 139-151.

27. K. L. Stevenson, D. W. Knorr and A. Horvath, Inorg. Chem., 1996, 35, 835-839.

28. O. Horvath and J. H. Fendler, J. Photoch. Photobio. A. Chem., 1993, 71, 33-37.

29. O. Horvath, J. H. Fendler and K. L. Stevenson, Inorg. Chem., 1993, 32, 227-230.

30. K. L. Stevenson, R. M. Berger, M. M. Grush, J. C. Stayanoff, A. Horvath and O.
Horvath, J. Photoch. Photobio. A. Chem., 1991, 60, 215-227.

31. D. D. Davis, Stevenson, K. L., Davis, C. R., J. Am. Chem. Soc., 1978, 100, 6.

32. K. L. Stevenson, R. S. Dhawale, A. Horvath and O. Horvath, J. Phys. Chem. A,
1997, 101, 3670-3676.
70
33. K. L. Stevenson, P. B. Bell, O. Horvath and A. Horvath, J. Am. Chem. Soc., 1998,
120, 4234-4235.

34. K. L. Stevenson, P. B. Bell, R. S. Dhawale, O. Horvath and A. Horvath, Rad.
Phys. Chem., 1999, 55, 489-496.

35. A. Horvath, O. Horvath and K. L. Stevenson, J. Photoch. Photobio. A. Chem.,
1992, 68, 155-163.

36. A. Horvath and K. L. Stevenson, Coord. Chem. Rev., 1996, 153, 57-82.

37. O. Horvath and K. L. Stevenson, Inorg. Chem., 1989, 28, 2548-2551.

38. C. R. Davis and K. L. Stevenson, Inorg. Chem., 1982, 21, 2514-2516.

39. Blandamer.M.J., Fox, Chem. Rev., 1970, 70, 59.

40. A. E. Jailaubekov and S. E. Bradforth, Appl. Phys. Lett., 2005, 87.

41. M. J. Tauber, R. A. Mathies, X. Y. Chen and S. E. Bradforth, Rev. Sci. Instru.,
2003, 74, 4958-4960.

42. J. A. Kloepfer, V. H. Vilchiz, V. A. Lenchenkov and S. E. Bradforth, Chem. Phys.
Lett., 1998, 298, 120-128.

43. K. L. Stevenson, M. M. Grush and K. S. Kurtz, Inorg. Chem., 1990, 29, 3150-
3153.

44. D. G. Peters and R. L. Caldwell, Inorg. Chem., 1967, 6, 1478.

45. V. H. Vilchiz, J. A. Kloepfer, A. C. Germaine, V. A. Lenchenkov and S. E.
Bradforth, J. Phys. Chem. A, 2001, 105, 1711-1723.

46. T. W. Kee, D. H. Son, P. Kambhampati and P. F. Barbara, J. Phys. Chem. A,
2001, 105, 8434-8439.

47. C. L. Thomsen, D. Madsen, S. R. Keiding, J. Thogersen and O. Christiansen, J.
Chem. Phys., 1999, 110, 3453-3462.

48. S. M. Pimblott and J. A. LaVerne, 1998, 102, 2967-2975.

49. C. G. Elles, A. E. Jailaubekov, R. A. Crowell and S. E. Bradforth, J. Chem. Phys.,
2006, 125.

50. F. Y. Jou and G. R. Freeman, J. Phys. Chem., 1977, 81, 909-915.
71
51. S. F. Ruzankin, V. F. Anufrienko, S. A. Yashnik and Z. R. Ismagilov, J. Struct.
Chem., 2006, 47, 404-412.

52. M. Lorenz, N. Caspary, W. Foeller, J. Agreiter, A. M. Smith and V. E. Bondybey,
Mol. Phys., 1997, 91, 483-493.

53. A. Treinin and E. Hayon, J. Am. Chem. Soc., 1975, 97, 1716-1721.

54. G. A. Bowmaker, Brocklis.Ld and R. Whiting, Aust. J. Chem., 1973, 26, 29-42.

55. G. A. Bowmaker, G. R. Clark, D. A. Rogers, A. Camus and N. Marsich, Dalton
T., 1984, 37-45.

56. I. Persson, M. Sandstrom, A. T. Steel, M. J. Zapatero and R. Akesson, Inorg.
Chem., 1991, 30, 4075-4081.

57. C. W. Dekock and D. M. Gruen, J. Chem. Phys., 1966, 44, 4387.

58. P. D. Fleischauer, Ph.D. Thesis, University of Southern California, 1968.

59. L. I. Grossweiner and M. S. Matheson, J. Phys. Chem., 1957, 61, 1089-1095.

60. M. Dangelantonio, M. Venturi and Q. G. Mulazzani, Rad. Phys. Chem., 1988, 32,
319-324.

61. J. Grodkowski and P. Neta, J. Phys. Chem. A, 2002, 106, 11130-11134.

62. J. Lind, X. H. Shen, T. E. Eriksen, G. Merenyi and L. Eberson, J. Am. Chem. Soc.,
1991, 113, 4629-4633.

63. D. Zehavi and J. Rabani, J. Phys. Chem., 1972, 76, 312-&.

64. M. F. Fox, Farad. Trans. I, 1977, 73, 872-882.

65. P. S. Braterman, Inorg. Chem., 1963, 2, 448.

66. J. C. Marcum and J. M. Weber, J. Chem. Phys., 2009, 131.

67. C. Rensing, O. T. Ehrler, J. P. Yang, A. N. Unterreiner and M. M. Kappes, J.
Chem. Phys., 2009, 130.

68. E. M. Glebov, V. F. Plyusnin, A. B. Venediktov and S. V. Korenev, Russ. Chem.
B+, 2003, 52, 1305-1311.

72
69. J. Friedrich, S. Gilb, O. T. Ehrler, A. Behrendt and M. M. Kappes, J. Chem.
Phys., 2002, 117, 2635-2644.

70. S. E. Bradforth and P. Jungwirth, J. Phys. Chem. A, 2002, 106, 1286-1298.

71. J. K. Burdett and O. Eisenstein, 1992, 31, 1758-1762.

72. J. Blumberger, L. Bernasconi, I. Tavernelli, R. Vuilleumier and M. Sprik, J. Am.
Chem. Soc., 2004, 126, 3928-3938.

73. J. Brugger, B. Etschmann, W. Liu, D. Testemale, J. L. Hazemann, H. Emerich,
W. van Beek and O. Proux, Geochim. Cosmochim. Ac., 2007, 71, 4920-4941.

74. A. R. Holzwarth, Data Analysis in Time-resolved measurements, Kluwer,
Dordrecht, 1996.

75. I. H. M. L. van Stokkum, D. S.; van Grondelle, R., R. Biochim. Biophys. Acta -
Bioenerg., 2004, 82.

76. R. Lian, D. A. Oulianov, R. A. Crowell, I. A. Shkrob, X. Y. Chen and S. E.
Bradforth, J. Phys. Chem. A, 2006, 110, 9071-9078.

77. V. Lenchenkov, J. Kloepfer, V. Vilchiz and S. E. Bradforth, Chem. Phys. Lett.,
2001, 342, 277-286.

78. R. A. Crowell, R. Lian, I. A. Shkrob, D. M. Bartels, X. Y. Chen and S. E.
Bradforth, J. Chem. Phys., 2004, 120, 11712-11725.

79. H. Zhang, J. Chen, I. V. Tomov, A. S. Dvornikov and P. M. Rentzepis, J. Phys.
Chem. A, 2007, 111, 11584-11588.

73

Supplement 2A

Nitrate and Acid Quenching Kinetic Fits.
a b
100 200 300 400 500
0
5
10
15

with H
+
mOD
Time, ps
no H
+

100 200 300 400 500
5
10
15
20
with KNO
3
no KNO
3

mOD
Time, ps
Figure 2A.1. a) CuBr
3
2 −
transient 400 nm time traces without quencher (green trace) and with [H
+
] =
0.2 M (black trace). Data normalized at signal following 2PA. Red trace is fit. b) 400 nm time traces
for CuBr
3
2 −
with (black trace) and without (green trace) and NO
3
-
. Red trace is fit.

a b
0 100 200 300 400 500
0
2
4
6
8
10
12
with 0.4 M NO
3
-

with [H+]=0.2 M
no quencher
mOD
Wavelength, nm

0 100 200 300 400 500
-2
0
2
4
6
8
0 2
0
2
4
6
8
10
12

m O D
Time, ps

mOD
Time, ps

Figure 2A.2. a) Back to back scans for 700 nm transient absorption of CuBr
3
2 −
solution without
quencher and with acid (taken from broadband data in Section 2.2, Figure (a) and 2.2.4b). 680 nm
time cut for CuBr
3
2 −
with nitrate, not back to back with other two scans. The nitrate solution has been
normalized to the unquenched solution. Black traces are experimental data and red traces are fits. b)
Time traces of 950 nm probe un-normalized. Black trace is CuBr
3
2 −
solution, red trace is CuBr
3
2 −

with 0.4 M NO
3
−
and blue trace is neat water.
74
Subtraction Spectra
a b
300 350 400 450 500 550 600 650
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
100 fs subtracted from 150 fs
450 fs subtracted
from 500 fs

mOD
Wavelength, nm

300 350 400 450 500 550 600 650
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
1.5 ps subtracted from 2 ps
500 fs subtracted from 1 ps

c d
300 350 400 450 500 550 600 650
-2
-1
0
1
2
2 ps subtraced from 102 ps
200 ps subtraced from 450 ps

mOD
Wavelength, nm

300 350 400 450 500 550 600 650
-2
-1
0
1
2
2 ps subtracted
from 100 ps

200 ps subtracted from 450 ps
mOD
Wavelength, nm
500 fs subtracted from 1 ps

Figure 2A.3. Difference plots of 266 nm excited, broadband probe aqueous CuBr
3
2 −
in 0.4 M Br
−
at
neutral pH. Plots show both UV grating and visible grating detection. Positive values indicate
increased absorption and negative values indicate decreased absorption. a) 100 fs spectral cut
subtracted from 150 fs spectral cut reveals a band in the visible decreases while a dual band centered
at approximately 350 nm and further in the blue grows in. b) Two traces are 500 fs subtracted from 1
ps and 1.5 ps subtracted from 2 ps. There is continued decrease in the visible and increase of the dual
UV band and a shoulder at about 400 nm. c) A narrow band at approximately 340 nm grows in with
decrease at approximately 440 nm. d) Difference plots for neat water.
75

Supplement 2B

CuBr
3
2
ˉ vs. CuBr
2
ˉ

The hypothesis we are testing is that there is no early time detachment of
electrons from the tri-bromo form. Because the CuBr
2
ˉ absorption band is superimposed
on the CuBr
3
2
ˉ absorption band it is difficult to isolate only the tribromocuprate species.
In addition, pushing the equilibrium of CuBr
3
2
ˉ and CuBr
2
ˉ toward the dibromo species
requires a smaller amount of Brˉ in sample preparation. Reducing the ligand
concentration also reduces the solubility of the starting solid CuBr reagent and the
concentration of both dibromo and tribromo are reduced at lower ligand concentrations.
Our broadband experiments were done at 266 nm excitation and at 0.4 M ligand
concentrations. In the experiments described here, we utilized the 250 kHz repetition rate
laser system (see Section 2.3) to provide lower intensity excitation pulses, preventing
significant photodetachment via water or Brˉ 2PA. The pump was changed to 255 nm in
order to excite at a wavelength more favorable to CuBr
2
ˉ than 266 nm, and the Brˉ
concentration was reduced to 0.3 M, the smallest concentration we could use that still
gave us a sufficient concentration of dibromo- and tribromo- species to see transient
signal. At 0.3 M Brˉ, concentrations of copper species are 2.8 x 10
-4
M CuBr
2
ˉ and 8.6 x
10
-4
M CuBr
3
2
ˉ; the ratio of di- vs. tribromocuprate is 0.32. At 255 nm, the extinction
coefficient is approximately 1:1 di- vs. tribromocuprate (labeled Exp A). At 0.4 M ligand
concentration (labeled Exp B) the concentrations are 3.0 x 10
-4
M CuBr
2
ˉ and 1.2 x 10
-3

76
M CuBr
3
2
ˉ. The ratio of di- vs. tribromocuprate is 0.25 and the extinction coefficient is
approximately 0.75:1 di- vs. tribromocuprate at 255 nm.
Due to the solubility dependence on ligand concentration, the ODs for the two
solutions were 0.12 and 0.6 for the 0.3 M Brˉ solution and 0.4 M Brˉ solution,
respectively, in 1 mm cuvettes. Based on the dibromo-tribromo equilibrium, the OD of
CuBr
2
ˉ was 0.038 and 0.11 for 0.3 M Brˉ and 0.4 M Brˉ, respectively. If the only signal
at 900 nm was due to solvated electron and the pathlengths and excitation powers were
constant, the transient absorption of Exp A would still be 3 times smaller than that of Exp
B, despite our efforts to favor CuBr
2
ˉ absorption.
There were many experimental differences between Exp A and Exp B. The
sample pathways for the two experiments differed since one incorporated a flow cell
(Exp A, 50 µm) and the other a liquid jet (Exp B, 100 µm pathlength). Exp A had an
excitation power of 1.5 mW (255 nm) and Exp B had an excitation power of 3.4 mW
(266 nm). If normalized for excitation power and pathlength, the data for Exp A would
show at least 4 times higher signal than Exp B. The fact that the two transient signals
have the same intensity suggests that CuBr
2
ˉ is not the only species ejecting a prompt
electron. However, other factors, such as spot size and absorption by the flow cell need
to be taken into account and more controlled experiments performed in order to come to a
confident conclusion.

77
a b
0 50 100 150 200
0
5
10
15
20

Lock-in Signal (microV)
Time, fs

-1 0 1 2 3 4 5
0
5
10
15
20

Lock-in Signal (microV)
Time, fs

Figure 2B.1. Black traces: 266 nm 3.4 mW excitation of copper bromide in aqueous solution of 0.4 M
Br
-
at neutral pH. Red Traces: 255 nm 1.5 mW excitation of copper bromide in aqueous solution of
0.3 M Br
-
at neutral pH. Detection wavelength is 900 nm and experiments were done on 250 kHz
repetition rate laser system.

78

Chapter 3
Oscillatory Transient Absorption of Photoexcited Tribromocuprate(I) Anion:
Observation of Vibrational Wavepackets

3.1 Abstract for Chapter 3

Broadband transient absorption of UV excited CuBr
3
2-
shows oscillations in the signal
that spans the UV and visible spectrum. These oscillations are assigned to quantum beats
corresponding to a vibrational wavepacket that is launched by the pump laser pulse.
Analysis of the oscillations reveals that the most likely candidate producing these beats is
the CuBr
3
2-
CTTS state, which transfers its vibrational coherence to the CuBr
3
2-
triplet
species through ISC and both a bending mode and stretching mode are being observed.
Jahn-Teller distortion of the CuBr
3
2-
is responsible for the vibrational coherence and
occurs because the Cu(I) center becomes Cu(II)-like upon photoexcitation into the
CuBr
3
2-
CTTS state.

79
3.2 Introduction

3.2.1 Wavepackets

UV excitation of aqueous CuBr
3
2-
yields an oscillatory transient absorption signal at
probe wavelengths spanning the UV and visible spectrum. As we shall see, features in
the oscillations point toward vibrational coherence (a wavepacket). A vibrational
wavepacket is a coherent superposition of vibrational states that propagates along an
electronic potential energy surface (PES). There are three mechanisms that can launch a
vibrational wavepacket. First, a vibrational wavepacket can be prepared directly via
vertical photoexcitation from the ground state to an excited
1, 2
or ionized
3-5
state. In this
case, the laser pulse must have a bandwidth that is large enough to excite several
vibrational states within the Franck-Condon envelope. The second mechanism involves a
non-adiabatic transition that impulsively changes the electronic surface that the molecule
is undergoing nuclear motions on in the same way a short laser pulse changes the surface.
Examples include autodetachment of an electron or internal conversion. The third
mechanism is by way of bond-breaking delivering an impulsive force that excites the
photoproduct into a vibrational coherence.
6

A simple illustration of wavepacket motion is in a vibrating diatomic molecule.
According to the Born-Oppenheimer approximation,
7
electrons move much faster than
atoms; therefore, when an ultrashort laser pulse “instantaneously” excites a diatomic
molecule from its ground state to an excited state, the molecular geometry remains
“frozen”. Depending on the equilibrium geometries of the two electronic states,
80
vibrational motion may or may not result from the electronic transition. In Figure 3.2.1a,
populating the upper PES with a short pulse will not induce vibrational motion because
the vertically prepared (Franck-Condon) geometry of the excited molecule matches the
equilibrium structure. In contrast in Figure 3.2.1b, the upper electronic state has a PES
minimum with an equilibrium distance that is displaced from that of the ground state.
Vertical photoexcitation with a laser pulse of sufficient bandwidth will prepare a
wavepacket of vibrational states with Franck-Condon activity. In this cartoon, three
quantum states of the stretching mode have been superposed. The wavepacket
propagates along the PES and, if it is bound, will make one or more recurrences. For a
wavepacket propagating under the influence of the dissipative medium of the solution,
dephasing and relaxation also need to be accounted for.
8

Vibrational wavepackets can be observed by means of pump-probe transient
spectroscopy. Banin and coworkers
6
observed transient oscillations in probe absorption
signal following 308 nm excitation of I
3
-
and partly attributed the oscillating signal to
vibrationally hot I
2
-
. Takeuchi and coworkers
2
observed vibrational wavepacket motion
on the S
2
state of diphenylcyclopropenone in solution following 295 nm excitation and
Consani and coworkers observed wavepacket motion in the transition metal complex
[Fe
II
(bpy)
3
]
2+
.
1
The above examples illustrate vibrational coherence induced by
dissociation (first) and by short pulse electronic excitation (later two). An example of
wavepacket motion resulting from photoionization is that of D
2
+
and H
2
+
.
3-5

81
a b
Ground State PES
Excited State PES
R
Ground State PES
Excited State PES
R

R
Excited State PES
Ground State PES
R
Excited State PES
Ground State PES

Figure 3.2.1. Vertical electronic excitation schematic for a diatomic molecule. R represents
internuclear distance, the parabolas are potential energy surfaces (PES) for the ground and excited
states, and the Gaussians are electronic wavefunctions. a) Ground and excited PES with same
internuclear minima. b) Excited PES has longer equilibrium internuclear distance than ground PES.
Adapted from Reference 9.

Figure 3.2.2a illustrates how a propagating vibrational wavepacket is observed in
a transient spectral signal. Three electronic potential energy surfaces need to be
considered in wavepacket observation. The first two are the ground state and excited
state PES as shown in Figures 3.2.1a and 3.2.1b, and the third is the electronic state that
is accessed by absorption of the probe pulse (probing surface). It has already been
established that the ground and excited state PES minima need to be displaced with
respect to the other for a vibrational wavepacket to be launched by the pump pulse. In
order for a particular vibrational wavepacket motion to display an oscillatory feature in
the time-domain, excited states should be bound and the difference potential between the
82
probing and excited state surface must vary with the vibrational coordinate.
2
The higher
probing surface does not need to be bound, as is depicted in Figure 3.2.2a but, if it is, it
should be displaced from the excited state PES or have a very different curvature. In this
scenario, at time = 0, the difference potential is the largest and probe wavelengths toward
the blue will absorb most. As the wavepacket propagates, there is increased absorption at
redder wavelengths and decreased absorption in the blue. Recurrences result as the
wavepacket oscillates between turning points on the PES. Oscillatory damping is caused
by dephasing of the wavepacket, as well as collisions with surrounding molecules. A
corresponding transient absorption signal is depicted in Figure 3.2.2b.

a b
R
Excited State PES
Upper PES
R
Excited State PES
Upper PES

Time
T ran sient Absorption
Time
T ran sient Absorption

Figure 3.2.2. a) Schematic of probe absorption by a propagating wavepacket. The colored lines
represent various resonant transitions that promote the vibrational wavepacket from the excited state
PES to the probing (upper) surface. b) Observed spectral signal with probe center corresponding to
blue energy arrow in (a). Damping is induced by wavepacket dephasing and/or vibrational energy
relaxation by interaction with the solvent.

83
3.2.2 Copper Bromide Complexes

The complex under study, CuBr
3
2-
, is a relatively small molecule with heavy
ligands (low vibrational frequencies), making it a good candidate for resolving
vibrational wavepackets in the pump-probe signal. The small number of atoms offers a
limited number of possible vibrational modes. The normal modes of both
tribromocuprate and dibromocuprate are shown in Figures 3.2.4a and b. In D
3h
, CuBr
3
2-

has 4 normal modes that include two degenerate vibrations. All vibrations except the a
1
'
mode is IR active. In contrast, D
∞h
CuBr
2
-
has 3 normal modes where all but the
symmetric stretch is IR active. In electronic pump-probe spectroscopy, modes that are
Franck Condon active in the pump transition may launch wavepackets if the excitation
pulse has sufficient bandwidth.
10
These are usually totally symmetric modes, or if there
is a change in equilibrium point group symmetry in the upper state (for example, by Jahn-
Teller distortion), then modes along the symmetry distortion coordinate are also
activated.
The vibrational frequencies of Cu(I) and Cu(II) dibromo-and tribromo- complexes
determined in the literature range from 40 cm
-1
to >300 cm
-1
(See Table 3.2.1). In both
oxidation states of the dibromo complex, the vibrational frequencies can be seen to be
~80, 190 and 320 cm
-1
. On going to the tri-bromo complexes, it is well known that the
vibrational frequencies of transition metal complexes decrease as the ligand number
increases.
11, 12
The values in Table 3.2.1 are consistent with this expectation, notably the
IR active asymmetric stretch frequency drops from 320 to 185 cm
-1
.
Figure 3.2.3 shows
the original experimental IR spectra for CuBr
3
2-
in crystals of [PMePh
3
]
2
[CuBr
3
] and is
84
taken from Reference 11. The strong IR band at 185 cm
-1
is assigned to the asymmetric
e'

Table 3.2.1. Literature values for experimental and calculated bending and stretching
normal modes of Cu(I) and Cu(II) dibromocuprate and Cu(I) tribromocuprate complexes.
G: Gas phase, S: Solution Phase, C: Crystal.

Species Mode Theory
(cm
-1
)
Experiment
(cm
-1
)
IR/
Raman

Ref
Cu
I
Br
2
-
Bend  ) 78 81
a
(Solution)
81,77
b

(Crystal)
IR
13
,
14

15
,
11

Sym Str ( +) 187 193 (Solution)
191
b
(Crystal)
Raman
13
,
14

15

Asym Str ( -) 317 322
a
(Solution)
IR
13
,
14

15
,
11

321, 313
b

(Crystal)

Cu
II
Br
2
Bend  ) 86 --
13

Sym Str ( +) 188 --
13

Asym Str ( -) 341 --
13

Cu
I
Br
3
2-
In plane bend
(e')
47
c
40, 59
(Crystal)
IR
11

Deform (a
2
") 75
c
90 IR
11

Sym Stretch
(a
1
')
146
c
- - c
Asym Stretch
(e')
167
c
185 IR
11

Cu
II
Br
3
-
Sym Stretch

180
c
-- c
Asym Stretch
(a
1
)
274 -- c
Asym Stretch
(b
2
)
256 -- c

a
We believe the authors of Ref
14
inadvertently reversed the assignment of the bend and asymmetric
stretch; we have used the assignment from Ref
13
throughout.
b
No vibrational assignments made in these papers.
c

B3LYP/PW91/6-31g* this work. Solvent model is CHEMSOL.
85
0 50 100 150 200 250 300 350 400
40
90
59
185

Wavenumber, cm
-1

Figure 3.2.3. Experimental IR crystallographic spectra for [PMePh
3
]
2
[CuBr
3
] taken from Reference
11, replotted. Values are listed in Table 3.2.1. The copper species in this crystal has been assigned by
crystallography as Cu
I
Br
3
2-
.

a
Symmetric Stretch A
1
’
Asymmetric Bend E
’
Asymmetric Stretch E
’
Out of Plane Bend A
2
’’
ν
1
ν
2
ν
4
ν
3
Symmetric Stretch A
1
’
Asymmetric Bend E
’
Asymmetric Stretch E
’
Out of Plane Bend A
2
’’
Symmetric Stretch A
1
’
Asymmetric Bend E
’
Asymmetric Stretch E
’
Out of Plane Bend A
2
’’
ν
1
ν
2
ν
4
ν
3

b
Symmetric Stretch
Bend
Asymmetric Stretch ν
1
ν
2
ν
3
Symmetric Stretch
Bend
Asymmetric Stretch
Symmetric Stretch
Bend
Asymmetric Stretch ν
1
ν
2
ν
3

Figure 3.2.4. In these figures, the green ball represents Cu, and purple the Br
-
ligands. Arrows
indicate the direction of movement. a) Vibrational normal modes of three coordinate complex
molecule with D
3h
symmetry. 
1
(a
1
') is Raman active, 

(a
2
") and both 
 
and 
 
e') modes are IR
active. b) Normal modes of dibromo complex with D
∞h
symmetry. Here, the 
1
symmetric stretch
( 
g
+
) is Raman active and both the 
3
asymmetric stretch ( 
u

) and 
2
bend ( 
u
) are IR active.

86

stretch and the other three bands at lower frequency are assigned as bends. The e' bend is
likely split due to a distortion in symmetry in the crystal. Following the analysis of CuI
3
2-

in the same reference, 3 bands near 60 cm
-1
in the IR spectra of CuI
3
2-
were assigned as
bending modes, including the a
2
" and split e' due to distortion from D
3h
symmetry. We
have also carried out DFT calculations for CuBr
3
2-
with a model solvent environment
necessary to stabilize the double negative charge on the anion. The computed
frequencies are consistent with the infrared bands in the crystal but also predict the IR-
forbidden symmetric stretch at ~145 cm
-1
.

3.2.3 Structure and Geometry of Copper Complexes

It is well known that d
10
transition metal ions (e.g., Cu
+
, Ag
+
, Au
+
and Hg
2+
) have
a tendency to form linear complexes.
15
Based on the IR and Raman spectra reported by
Persson et. al. and the rule of mutual exclusion for spectra of centrosymmetric molecules,
CuBr
2
-
has been confirmed to be linear D
∞h
with only slight structural and vibrational
perturbations due to solvent effects.
14
In aqueous solution, copper(I) di-halides are linear
and do not have water in their inner sphere.
16
The reported bond length in CuBr
2
-
is 2.21-
2.23 Å.
14, 17-24
In copper (I), the highest coordination state reported is three. CuBr
3
2-
has
been observed in crystals, and the complex ion has been reported to have a Cu – Br bond
distance 2.365 Å with D
3h
symmetry.
20
In aqueous solution, copper(I) tri-halides are
reported by Stevenson et. al.
16
to be trigonal planar for triscoordinated complexes with no
water in its inner sphere based on findings by Kappenstein et. al.
25
on the Raman
87
spectroscopy of Cu(CN)
n
n-1
(for n = 2,3,4) species. Our DFT calculations predict a D
3h

geometry for the aqueous ion (B3LYP/PW91/6-31g with a CHEMSOL model for the
solvent
26
) and with a bond length of 2.41 Å.

Figure 3.2.5. Crystal Field Diagram for Trigonal Planar Complexes with occupation shown for d
10

copper(I).

In contrast, copper(II) normally exhibits a higher coordination number, usually
four or six and, because the copper is d
9
,
copper(II) complexes are Jahn-Teller distorted
in these coordinations states. Examples of trigonal copper(II) complexes can be found in
the literature
27
but the copper is considered coordinatively unsaturated. As for four and
six coordinate geometries, if a D
3h
Cu(II) species such as CuBr
3
-
was formed, the orbital
splitting in a trigonal field (see Figure 3.2.5) are such that the d
x
2
-y
2 and d
xy
degeneracy
(the e' HOMO) will be lifted to reduce the degeneracy of the ground state electronic
2 2
y x
d


xy
d

xy
d

xy
d

2
z
d

z
y
x
88
wavefunction and thus CuBr
3
-
would be expected to undergo Jahn-Teller distortion to C
2v

(in plane deformation). We note that pyramidalization to C
3v
does not break the
degeneracy unless the three-fold symmetry is reduced as well. By distortion to C
2v
, the
ground electronic state symmetry is reduced from
2
E' in D
3h
to
2
A
1
+
2
B
2
. Our own DFT
calculations (B3LYP/PW91/6-31g*) confirm that the equilibrium structure of CuBr
3
-
in
the gas phase is C
2v
with all Cu-Br having very similar bond lengths (2.32 Å), but larger
angular deformations in plane - the unique bond angle is 110
o
. We note that the Cu-Br
bond length is shorter than in the copper(I) form.
This introductory information has served to show that a change in oxidation state
in the copper center of CuBr
3
2-
should give rise to distortions in at least two modes, the
symmetric stretch and the in-plane bend. If dissociation of the complex in the excited
state with loss of a bromide ion occurs, vibrational modes in the dibromocuprate(II) also
become relevant. By use of novel UV pulse generation techniques, sufficient bandwidth
of the UV excitation pulses in the current ultrafast pump-probe experiment allows us to
create wavepackets in these two modes and read out these vibrational motions.
Typically, 4 nm FWHM pulses centered at 266 nm are used in our experiments, yielding
565 cm
-1
of bandwidth for the excitation pulse, sufficient to superimpose more than one
vibrational quantum state in a wavepacket. To resolve a wavepacket crossing the PES a
time resolution of at least half the vibrational period is required.
10
For relevant
frequencies in both molecular systems, 80 cm
-1
is equivalent to a period of 417 fs in the
time domain, and 340 cm
-1
is a 98 fs vibrational period. Thus, oscillatory signals are
potentially resolvable for all vibrations present in these molecules in an experiment that
has 50 fs resolution.
89
3.3 Experimental

The broadband pump-probe experiments reported in this Chapter were performed
using a 1 kHz repetition rate laser system (Spectra Physics Hurricane), details of which
are described in Section 2.3 and depicted in Figure 2.3.1. The sample was 0.004 M
aqueous CuBr
3
2-
in 0.4 M Br
-
ligand concentration and at ~0.4 M ionic strength as also
described in Section 2.3.

3.4 Results

3.4.1 Oscillatory Transient Signal

Figure 3.4.1a shows transient kinetic plots of 266 nm excited CuBr
3
2-
from 301 nm to 700
nm, which are extracted from the 2D time-wavelength plots. Oscillations are apparent at
all probe wavelengths following the 2PA until approximately 1.5 ps and are least
pronounced in the region around 410 nm (Figure 3.4.1b), which is near the peak of the
UV transient Band 1 that grows in before 1 ps (black trace in Figure 3.4.2a). Figures
3.4.2 reproduces data already shown in Chapter 2. The modulation depth in the 350 nm
and 501 nm probe absorption is ~10% of the overall signal, obtained by taking the
difference from highest peak to lowest trough relative to the transient absorption intensity
at the highest peak.

90

a b
-1 0 1 2 3 4 5 6
500 nm
550 nm
600 nm
650 nm
700 nm
301 nm
350 nm
400 nm
450 nm
Delay Time, ps

-1 0 1 2 3 4 5 6
Time, ps
430
425
420
415
410
405
400
395
385
380

Figure 3.4.1. 266 nm excited CuBr
3
2-
transient time cuts. Vertically offset for comparison. a) Time
traces for probe wavelengths from 301 nm to 700 nm. b) Time cuts for wavelength range from 380
nm to 430 nm, labeled on graph in nm.

a b
300 400 500 600
2
4
6
8
10
12
0.5 ps
1 ps
2 ps
52 ps
102 ps
202 ps
302 ps
402 ps mOD
Wavelength, nm

0.0 0.5 1.0 1.5 2.0
2
3
4
5
6
7
8
9
10
11
12
mOD
Time, ps

Figure 3.4.2. a) From Chapter 2, figure is reproduced to aid in assignment. Broadband spectral cuts
from 0.5 ps to 400 ps of aqueous CuBr
3
2-

after photo-excitation at 266 nm. Arrow points to region we
observed least modulation of the transient absorption signal. b) Raw 350 nm (blue) and 501 nm (red)
time traces to emphasize modulation depth. Data not offset.

91
The phase relationship between the different probe wavelengths is better observed
in Figure 3.4.3. The 301 nm and 350 nm probe absorption kinetic trace has its first peak
near 200 fs (the vertical lines in the figure start at 250 fs delay and then are spaced by 500
fs) while 200 fs is a minimum for 500 nm and 550 nm. The oscillations continue with
peaks on the blue side of 410 nm that correspond to troughs on the red side of 410 nm,
thus the two sides of the 400 nm transient band are out of phase by π.

a b
0 500 1000 1500 2000

500 nm
550 nm
301 nm
350 nm
400 nm
Delay Time, fs

0 500 1000 1500 2000
4
3
2
1
2
4 3

550 nm
350 nm
Delay Time, fs

Figure 3.4.3. a) Kinetic traces of 266 nm excited CuBr
3
2-
from 301 nm to 550 nm. b) Labels of
oscillatory peaks corresponding to Table 3.4.1.

Table 3.4.1 lists the time delays of the oscillatory peaks for probes shown in
Figure 3.4.3. The oscillations in the 400 nm data are not pronounced enough for values
to be obtained and included. The value for “Peak 1” at 550 nm may be an artifact of
population dynamics since the early time absorption band in this region in the broadband
transient data in Chapter 2 could account for what appears to be Peak 1. If so, Peak 1
92
would not be a true beat. Both the higher and lower frequency beats appear to have a 
phase shift across the absorption band.

Table 3.4.1. Time of oscillation peaks following 2PA in femtoseconds for probe
wavelengths 301, 350, 400, 500, and 550 nm.

Probe Wavelength Peak 1, fs Peak 2, fs Peak 3, fs Peak 4, fs
301 nm -- 164 344 764
350 nm -- 174 344 754
400 nm -- -- -- --
500 nm -- 254 474 894
550 nm 94 254 464 894

Table 3.4.2. Difference between oscillation peaks listed in Table 3.4.1 in femtoseconds.
Δ
1,2
is the difference between Peak 1 and Peak 2, and so forth. Conversions of the peak
to peak time differences to frequency domain (in wavenumbers) are in parenthesis.

Probe Wavelength Δ
1,2
, fs (cm
-1
) Δ
2,3
, fs (cm
-1
) Δ
3,4
, fs (cm
-1
)
301 nm -- 180 (185) 420 (79)
350 nm -- 170 (196) 410 (81)
400 nm -- -- --
500 nm -- 220 (152) 420 (79)
550 nm 160 (208) 210 (159) 430 (78)

3.4.2 Discrete Fourier Transform

The oscillatory component was also Fourier analyzed. To obtain a useful Fourier
transform which will reveal beat frequencies present in the data, the raw dataset was first
truncated at 2 ps so that all data is equally spaced in time. (All oscillations have died
93
away by this cut off point.) The data in the region of zero time delay (out to the first
minimum in the data) is replaced by zeroes to exclude the large signal due to solvent
2PA. Then the DC component of the signal is subtracted off by first fitting the data with a
bi-exponential rise. The desired residual from this operation is a beat oscillating either
side of zero amplitude. This process approximately removes the dynamics associated
with population changes (the zero frequency signal in the FFT). Finally to properly pack
the signal for FFT, the data is zero padded to 1024 points and reflected to eliminate an FT
offset. Wavelengths selected were 350 nm, 410 nm, and 500 nm and results of the
transform are shown in Figure 3.4.4. As the time window selected from the original data
is 2 ps, the frequency resolution is 16.7 cm
-1
.

0 100 200 300 400 500
0
50
100
150
200
182
155
81

FT Intensity
Frequency, cm
-1
350 nm
410 nm
500 nm

0 50 100 150 200 250
FT Amplitude, Arb. Units

Frequency, cm
-1

Phase
0
π/2
π
- π/2
- π
0 50 100 150 200 250
FT Amplitude, Arb. Units

Frequency, cm
-1

Phase
0
π/2
π
- π/2
- π

Figure 3.4.4. a) Discrete Fourier transform for 350 nm, 410 nm, and 500 nm kinetic traces for the
time window of 0 to 2 ps. b) Discrete Fourier transform phase for 350 nm (blue triangles) and 500 nm
(green squares), left axis. Dashed lines are FT from (a), right axis.

There is a major peak in the Fourier transform (FT) at 81 cm
-1
and a smaller peak
at 155 cm
-1
for both 350 nm and 500 nm. In addition, the 350 nm Fourier transform has a
94
shoulder near 50 cm
-1
and another peak at 182 cm
-1
. Peaks are of much smaller
amplitude at 410 nm as expected from inspecting the raw data. We suspect the peak near
25 cm
-1
is probably an artifact. 81 cm
-1
agrees well with the beat observed with all probe
colors (Table 3.4.1) while 155 cm
-1
agrees with the beat seen on the red side of the band
and 182 cm
-1
agrees with probes on the blue side. The phases from the Fourier transform
are shown in Figure 3.4.4b. For 350 nm, at 81, 155, and 182 cm-1, the phases are ~ π/2,
~ π/2, and π, respectively. The phases for 500 nm at the same FT frequencies are ~ π/2,
~π/2, and 0, respectively. The phases of each frequency at the two probe wavelengths are
all out of phase with respect to the other by π.

3.5 Discussion

3.5.1 Vibrational Wavepacket Motion

Analysis of the oscillatory features in combination with known vibrational
frequencies and structure makes possible an assignment of the species undergoing
vibrational motion and therefore the reaction dynamics following photoexcitation. The
important experimental observables in our transient data include: (i) the phase shift of the
oscillations as a function of probe wavelength on both sides of a region of reduced
oscillatory amplitude, (ii) the amplitude of the oscillations compared to the total intensity
of transient signal, and (iii) and the period of oscillations, fitting, and Fourier transform
analysis.
95
i) The phase shift of π between the wavelengths on the blue side and the red side
of the UV band is typical of a wavepacket oscillating within a potential well when the
difference potential with the probing surface is monotonic
1, 28
as described by the
example given in Figure 3.2.2. As such, this phase information tells us about the relative
shapes of the potential energy surfaces.
ii) Amplitude of Oscillations. Based on analysis by Kovalenko et. al.
29
on
electronic coherence in excited state fluorinated benzene, large amplitudes in oscillations
compared to total absorption intensity are indicative of an electronic coherence. Because
the modulation amplitude is ~10%, electronic wavepackets can be ruled out, supporting
our assignment of vibrational quantum beats.
iii) Period of Oscillations, Fits and Fourier Transform. The beat frequencies
derived from the oscillation period in the transient absorption signal across the
wavelength range of 301 nm to 550 nm are listed in Table 3.4.2. These frequencies are
confirmed via the Fourier analysis (Figure 3.4.4). The frequencies derived are in good
correspondence to known vibrational frequencies in dibromo- and tribromo- copper
complexes (Table 3.2.1 and Figure 3.2.3). Specifically, in the Fourier transform data of
the oscillatory transient signal, ~80 cm
-1
is similar to a Br-Cu-Br bending mode, and both
152 and 180 cm
-1
are close to a Cu-Br stretching mode. Also 152 cm
-1
could be the
second harmonic of beat in the 80 cm
-1
bending mode allowing for anharmonicity.

96
3.5.2 Assignments

We can use the information given in the Introduction to consider possible
wavepacket motion in ground, excited or product states. We note that coherent
oscillations persist longer that the early phase of population transfer as assigned in
Chapter 2. Since copper(I) complexes do not have water in its inner sphere,
16
it is
assumed crystal and gas phase structural and vibrational information is pertinent to the
aqueous bromocuprate complexes. In general CTTS states have their promoted electron
at long range from the molecular core rather like a gas phase Rydberg state of a neutral
molecule.
30, 31
Further as electron promotion in CuBr
3
2-
is thought to involve excitation
into a diffuse orbital on copper,
32
it is not unreasonable to assume that the molecular core
in CTTS photoexcited Cu
I
Br
2
-
and Cu
I
Br
3
2-
is similar to Cu
II
Br
2
and Cu
II
Br
3
-
. Unlike the
tribromocuprate species, the Cu
I
Br
2
-
anion and Cu
II
Br
2
neutral have only minor structural
and vibrational frequency differences.
13
(In what follows, unless noted with a valence
superscript in the chemical formula, the copper is assumed to be in the +1 oxidation
state). There are five mechanisms we will now consider for the assignment of the
observed oscillatory signal.
1. Ground state wavepackets in CuBr
3
2-
driven by resonant impulsive stimulated
Raman. IR and Raman data for CuBr
3
2-
show vibrational frequencies in good agreement
with the oscillatory signal and its Fourier transform from this experiment.
11
Yet, if the
ground state were contributing to the quantum beats, two requirements should be met.
First, a ground state electronic absorption needs to overlap the near-UV and visible
spectral region where we see beating here and beats should also be observed across the
97
strong ground state absorption band at 280 nm. There is no ground state absorption in the
visible, and one color pump-probe experiments at 266 nm show no clear oscillatory
features (see Figure 3.5.2).

a b
225 250 275 300 325 350
0
1
2

OD
Wavelength, nm

-0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0
0
1
2
3
4
5

mOD
Time, ps

Figure 3.5.1. a) UV/visible absorption spectra of 0.004 M ground state CuBr
3
2-
in water. b) One
color 266nm excite/266nm probe data taken on the 1 kHz repetition rate laser system. 266nm/266nm
experiments on the 250 kHz laser system (data not shown) yield the same results.

2. Full detachment of an electron from the metal center of CuBr
3
2-
to produce a
Cu(II) transient species, Cu
II
Br
3
-
. In this scenario, the photodetached product finds itself
in an unfavorable geometry, inducing wavepackets in the modes that are displaced from
their equilibrium geometry. Since it has been reported that the UV band of ground state
CuBr
3
2-
is a CTTS band, we originally presumed that photoexcitation leads to some
instantaneous electron ejection from the copper center, which is discussed in more detail
in the Introduction of Chapter 2, Section 2.2. Because crystallography determines the
98
structure of CuBr
3
2-
to be trigonal planar, D
3h
,
20
detachment to a Cu(II) (d
9
) complex,
would be expected to lead to a Jahn-Teller distortion along the in-plane bending
deformation mode to a C
2v
structure.
33-35
Also because the Cu-Br bonds are calculated to
shorten in the Cu(II) form, the symmetric stretch mode would also be excited. However,
the fact that ultrafast ejection of electrons is not observed in our pump-probe data is
strong evidence that the oscillatory signal is not due to the photodetached Cu
II
Br
3
-

product and thus this is not a major channel in the photophysics. The modulation depth
(~10%) is incompatible with our estimated upper limit for prompt solvated electrons (≤
5%)
3. Fast dissociation to produce the dibromo radical anion Br
2
● −
. In this
concerted mechanism, a Br
-
ligand and Br radical come off together to form vibrationally
excited Br
2
-
• and Cu
0
Br
−
. The observed quantum beats would be in the Br
2
●−
product.
This mechanism is suggested because Br
2
-
• has a known absorption band at 360 – 370
nm, which is in the region of the transient spectral UV band following photoexcitation of
CuBr
3
2-
.
8, 36-42
There are three reasons we believe Br
2
●−
is not the source of the observed
vibrational wavepacket. First, as reported in Section 2.5 of Chapter 2, concerted
photodissociation of CuBr
3
2-
into a CuBr
-
and Br
2
●−
product is not likely as this band is
not considered to have LMCT character. Second, the vibrational frequencies reported for
Br
2
●−
do not include the dominant ~80 cm
-1
seen in our experimental data.
43
Resonance
Raman data for Br
2
●−
in solution gives a fundamental frequency of 163 cm
-1
with only
small anharmonicity (0.7 cm
-1
)
42
so the wavepacket would have to be formed very high
in the well to exhibit an 80 cm
-1
average frequency and then the beat frequency should
increase with delay as the Br
2
-
cools.
44
However, the Br – Br distance in the CuBr
3
2-

99
complex is approximately 4 Å, based on geometries in the literature and our own
calculations.
13, 20
and the Br
2
●−
will also have 4 Å at the time of formation. Alexander et.
al.
45
reported a PES for ground state Br
2
●−
and at 4 Å, the energy is at ~ 0.7 eV. With a
ω
e
and ω
e
χ
e
of 167 cm
-1
and 0.7 cm
-1
, respectively, the most likely vibrational quantum
number for the nascent fragment would be v = 39 with a local frequency (spacing
between of v=39 and v=40) of 111 cm
-1
. As we do not see this occurring, the < 1 ps
dephasing rate would have to be substantially faster than the cooling rate which is
unlikely for such a highly excited vibration and low frequency mode in water. The third
reason Br
2
●−
is unlikely to be responsible for the quantum beats is that the peak of its
electronic absorption is considerably shorter than 410 nm,
36-39, 41
the probe wavelengths
with smallest oscillation amplitude in our data. Typically, as a wavepacket passes
through the middle of a bound potential well, the oscillation appears at double the
fundamental frequency and its amplitude is severely reduced if the detection time
resolution, as here, is not sufficient to resolve a beat at the second harmonic.
10
The
situation would be more complex if the photoproduct were formed very high in the
anharmonic potential and a blue-shifting transient band would be expected to accompany
an increasing beat frequency. However, in this scenario, isosbestic behavior in the
evolution of the transient band would not be expected.
4. Abrupt dissociation of a Br
−
ligand to form an impulsively excited CuBr
2
−

species. In this scenario, the prompt loss of a Br
-
ligand delivers an impulse on the Cu
atom and induces bending excitation in the CuBr
2
-
product with a beat frequency around
80 cm
-1
. In this scheme, the copper product must be an electronically excited Cu
I
Br
2
-

state, because ground state CuBr
2
-
has an absorption peak at ~275 nm and no absorption
100
in the 400 nm region.
46
Like the tribromo parent, the only low-lying electronically
excited state of CuBr
2
-

is a CTTS state and this would require an excited state absorption
band centered at 410 nm to the probing surface.
Figure 3.5.3 shows this scenario of ballistic dissociation of a Br
-
ligand with
impulsive excitation of the dibromocuprate product. Furthermore, we note that as the
equilibrium geometry of both CuBr
2
-
and Cu
II
Br
2
(our model for CTTS excited CuBr
2
-
-
the probing state) is linear,
13
and the CuBr
3
2-
parent starts out trigonal planar, means that
even if no impulse were delivered and the bend were a spectator mode, the initial Br-Cu-
Br bond angle 120° is strongly displaced relative to the linear equilibrium structure of the
CuBr
2
-
product. These two effects taken together suggest cartoon potential energy
surfaces for this scenario as depicted in Figure 3.5.4a).

Cu
Br
Br
Br
Cu
Br
Br
Cu
Br
Br
Br
Br
Br
Br
Cu
Br
Br
Cu
Br
Br
Br
Cu
Br
Br
Cu
Br
Br
Br
Br
Br
Br
Cu
Br
Br

Figure 3.5.2. Physical model for impulsive dissociation of CuBr
3
2-
, prompting a bending vibration of
CuBr
2
-
. Arrows represent the forces delivered in the bond breaking.

The second vibrational frequency observed would be a stretching mode. CuBr
3
2-

is calculated at BPW91 level of theory to have ~ 0.2 Å longer Cu – Br bonds (at 2.4 Å)
101
than in CuBr
2
-
(2.25 Å) or Cu
II
Br
2
(2.2 Å)
,
13
an effect which can be thought of as a copper
ion being better able to attract two Br
-
ligands than three Br
-
.
32
So dissociation of a
bromide ligand will produce the CuBr
2
-
species with a Cu-Br bond length displaced from
equilibrium and the symmetric stretching mode (~188 cm
-1
)
13
will therefore be excited.
Figure 3.5.4b shows the PES indicating the stretching mode displacement and its probing
surface.
However, in Figure 3.5.4a, the probing surface can have only two forms along the
bending coordinate depending on the accessed higher state. It can be either harmonic
centered at 180
o
(solid line) or a double well potential (dashed line) with a maximum at
180
o
. All calculated states for Cu
II
Cl
2
and Cu
II
Br
2
in fact have linear equilibrium
geometries consistent with the gas phase photoelectron spectra of CuCl
2
-
and CuBr
2
-
.
13
In
either case, a wavepacket moving in the photoproduct potential with 80 cm
-1
frequency,
because of the non-monotonic difference potential the fundamental bending frequency
will not be observed. It is suspected that only 2ω and 4ω will be observed in the
oscillatory transient signal.
To further cast doubt on this scenario, there is no evidence to suggest that the state
of CuBr
3
2-
initially excited by the pump pulse is dissociative. Although Jahn-Teller
distortion should induce a bending in the Cu(II) core of the excited state complex and
despite the promotion of an electron from a partially bonding in plane d orbital into a
non-bonding diffuse electron, the potential is not net anti-bonding along the Cu-Br bonds.
B3LYP/PW91 calculations indicate CuBr
3
-
is in fact a bound C
2v
molecule. Reports exist
for a CuBr
3
-
complex in acetonitrile solution.
47
Thus, we conclude that dissociation of a
bromide ligand also does not explain the observed data.
102

a b
PES
Probe λ
E
120 180
Angle Br-Cu-Br
240
PES
Probe λ
E
120 180
Angle Br-Cu-Br
240

PES
Probe λ
E
2.2
Internuclear Distance (Å) Cu-Br
2.4
PES
Probe λ
E
2.2
Internuclear Distance (Å) Cu-Br
2.4

Figure 3.5.3. a) Scenario describing the impulsive model in which the starting ground state has an
equilibrium Br-Cu-Br bond angle of 120° and the photoproduct has an equilibrium bond angle of
180°. The probing surface is probably linear (top solid parabola) with a different force constant
(based on CuCl
2
calculations and assuming Br follows the same trends). The dashed double well
represents the PES if the probing surface has a bent equilibrium geometry. b) The CuBr
2
-
product has
a shorter Cu-Br bond length than CuBr
3
2-
and Br
-
dissociation would lead to a prompting of the
stretching mode. The probing surface depicted in this cartoon is dissociative; however, the true PES is
unknown.

Note that a related pathway, dissociation of neutral Br
●
radical was not considered
for similar reasons and because in addition this would form Cu
0
Br
2
2-
, an unlikely product.
Such a process would also be characterized as an initial LMCT rather than CTTS
transition. Yet another related mechanism suggested in the literature is initial electron
detachment from the Br
-
ligand, then departure of bromine from the complex as a Br
●

radical leaving behind CuBr
2
-
.
48
However, we rule this out also as we see only a small
quantum yield of prompt electrons originating from the copper complex.
103
5. Launching of a wavepacket on the CuBr
3
2-
CTTS surface and coherent transfer
of the wavepacket motion onto the CuBr
3
2-
triplet CTTS state. A related scenario to that
described under (2) is that the wavepacket is launched by the short laser pulse in the
initial CTTS state itself and not in a Cu
II
photoproduct. The similarity is we envisage the
CTTS excited state as having a Cu(II) core and a diffuse 4s (or 4p) electron but without a
detached solvated electron. Thus, the Jahn-Teller distortion and symmetric bond
compression described there will also apply for vibrational coherences expected in the
CuBr
3
2-
CTTS singlet state. As we have determined that inter-system crossing takes place
in less than 1 ps (shorter than the dephasing of the vibrational coherences), the nuclear
wavepackets are transfered into to the triplet state. It is less likely that the vibrational
coherence is induced by the ISC process itself in the CuBr
3
2-
triplet, because the
intersystem crossing would have to be extremely fast (~100 fs) to induce beats in a 180
cm
-1
mode of the CuBr
3
2-
triplet. Also, we have documented reasons why these modes
are displaced in the optical transition and we expect the equilibrium geometries of the
singlet and triplet states to be very similar because the highest lying electron is in a
different diffuse CTTS orbital.
Based on our calculations and the assumption made about mapping the CTTS
excited state onto a detached state, the most likely active modes are the symmetric stretch
(B3LYP/PW91/6-31g* calculated frequency 179 cm
-1
) and the Jahn-Teller distortion
coordinates bringing the system to C
2v
. These are mainly the in-plane bend (calculated
gas-phase frequency ~77 cm
-1
) and, with less activity, the in-plane asymmetric stretch
pair (calculated as 256 and 274 cm
-1
). The latter mode has little displacement and the
highest frequency suggesting that it would be difficult to resolve a wavepacket motion in
104
this mode. We therefore assign the ~80 cm
-1
beat in the transient data FT to the in-plane
Jahn-Teller active bending mode, noticing that this frequency is not only consistent with
DFT calculations but similar to the bending frequency in tribromocuprate(I). We
recognize that the wavepacket motion in the in-plane bend is inherently a two-
dimensional problem. The bending mode two-dimensional wavepacket will move in a
“Mexican Hat” potential along the two components of the orginal e' normal mode
components. Considering the probing surface, the absorption spectra of the triplet state
near 400 nm could be a LMCT transition deriving from the copper(II) core and therefore
the upper PES would have a core d
10
. As a result, the surface minimum in the bending
mode will once again be at 120 degrees and therefore a bending 1ω and 2ω will be
observed. A full analysis of the signal resulting from motion in this 2D potential is
beyond the scope of this Chapter.

a b
R
Potential Energy
R
Potential Energy
0
4
5
5
R
Difference Potential
0
4
5
5
R
Difference Potential
Figure 3.5.4. a) Two Morse oscillators with similar curvatures and with the probing
surface (blue trace) displaced to the longer bond length side. b) Difference potential of
the two potentials in (a).

105
The beat in the symmetric stretching mode is more straightforward to understand.
A  phase shift across the probing spectrum will occur because the difference potential is
likely to be monotonic.

3.6 Conclusions

The pump-probe spectroscopy performed on CuBr
3
2-
in this work proves that we are
capable of observing vibrational wavepacket motion in the condensed phase. The sub
100 fs laser pulses in our experimental setup provide the time resolution needed to
resolve oscillatory signals and extract the vibrational frequencies of the CuBr
3
2-
CTTS
state. Another reason the wavepacket can be observed in this system is due to the heavy
ligands, which produce low frequency vibrational modes. There are many scenarios
presented in this Chapter for the assignment of the oscillating species and we have
provided arguments as to why the most likely scenario is wavepackets formed in the
CuBr
3
2-
singlet CTTS state which undergoes ISC into the triplet state while maintaining
vibrational coherence.

106
3.5 References for Chapter 3

1. C. Consani, M. Premont-Schwarz, A. ElNahhas, C. Bressler, F. van Mourik, A.
Cannizzo and M. Chergui, Angew. Chem. Int. Edit., 2009, 48, 7184-7187.

2. S. Takeuchi and T. Tahara, J. Chem. Phys., 2004, 120, 4768-4776.

3. M. T. Zanni, B. J. Greenblatt, A. V. Davis and D. M. Neumark, J. Chem. Phys.,
1999, 111, 2991-3003.

4. A. S. Alnaser, B. Ulrich, X. M. Tong, I. V. Litvinyuk, C. M. Maharjan, P.
Ranitovic, T. Osipov, R. Ali, S. Ghimire, Z. Chang, C. D. Lin and C. L. Cocke,
Phys. Rev. A, 2005, 72.

5. H. Niikura, F. Legare, R. Hasbani, M. Y. Ivanov, D. M. Villeneuve and P. B.
Corkum, Nature, 2003, 421, 826-829.

6. U. Banin and S. Ruhman, J. Chem. Phys., 1993, 98, 4391-4403.

7. M. Born and R. Oppenheimer, Ann. Phys-Berlin, 1927, 84, 0457-0484.

8. Y. Liu, A. S. Pimentel, Y. Antoku, B. J. Giles and J. R. Barker, J. Phys. Chem. A,
2002, 106, 11075-11082.

9. S. E. Bradforth, Ph.D. Thesis, University of California, Berkeley, 1992.

10. D. M. Jonas, S. E. Bradforth, S. A. Passino and G. R. Fleming, J. Phys. Chem.,
1995, 99, 2594-2608.

11. G. A. Bowmaker, G. R. Clark, D. A. Rogers, A. Camus and N. Marsich, J. Chem.
Soc. Dalton, 1984, 37-45.

12. D. M. Adams, Metal-Ligand and Related Vibrations, London, 1967.

13. X. B. Wang, L. S. Wang, R. Brown, P. Schwerdtfeger, D. Schroder and H.
Schwarz, J. Chem. Phys., 2001, 114, 7388-7395.

14. I. Persson, M. Sandstrom, A. T. Steel, M. J. Zapatero and R. Akesson, Inorg.
Chem., 1991, 30, 4075-4081.

15. G. A. Bowmaker, Brocklis.Ld and R. Whiting, Aust. J. Chem., 1973, 26, 29-42.

107
16. K. L. Stevenson, J. H. Jarboe, S. A. Langmeyer and T. W. Acra, Inorg. Chem.,
2003, 42, 3559-3564.

17. S. Andersson, S. Jagner, Acta Chem. Scand. A, 1986, 40, 177-181.

18. S. Andersson, S. Jagner, Acta Chem. Scand. A, 1986, 40, 210-217.

19. S. Andersson, S. Jagner, Acta Chem. Scand. A, 1986, 40, 52-57.

20. S. Andersson, S. Jagner, Acta Chem. Scand. A, 1987, 41, 230-236.

21. S. Andersson, S. Jagner, J. Cryst. Spectrosc., 1988, 18, 591-600.

22. S. Andersson, S. Jagner, Acta Chem. Scand., 1989, 43, 39-43.

23. S. Andersson, M. Hakansson and S. Jagner, J. Cryst. Spectrosc., 1989, 19, 147-
157.

24. S. Andersson, 1988.

25. C. Kappenstein, R. P. Hugel, A. J. P. Alix and J. L. Beaudoin, J. Chim. Phys.
PCB, 1978, 75, 427-443.

26. J. Florian and A. Warshel, J. Phys. Chem. B, 1997, 101, 5583-5595.

27. R. A. Howald and D. P. Keeton, Spectrochim. Acta., 1966, 22, 1211.

28. M. Guhr, M. Bargheer, M. Fushitani, T. Kiljunen and N. Schwentner, Phys.
Chem. Chem. Phys., 2007, 9, 779-801.

29. S. A. Kovalenko, A. L. Dobryakov and V. Farztdinov, Phys. Rev. Lett., 2006, 96.

30. Blandame.Mj, Chem. Rev., 1970, 70, 59.

31. X. Y. Chen and S. E. Bradforth, Annu. Rev. Phys. Chem., 2008, 59, 203-231.

32. K. L. Stevenson, R. S. Dhawale, A. Horvath and O. Horvath, J. Phys. Chem. A,
1997, 101, 3670-3676.

33. H. A. Jahn and E. Teller, P. Roy. Soc. Lond. A Mat., 1937, 161, 220-235.

34. L. X. Chen, G. Jennings, T. Liu, D. J. Gosztola, J. P. Hessler, D. V. Scaltrito and
G. J. Meyer, J. Am. Chem. Soc., 2002, 124, 10861-10867.

108
35. L. X. Chen, G. B. Shaw, I. Novozhilova, T. Liu, G. Jennings, K. Attenkofer, G. J.
Meyer and P. Coppens, J. Am. Chem. Soc., 2003, 125, 7022-7034.

36. J. Grodkowski and P. Neta, J. Phys. Chem. A, 2002, 106, 11130-11134.

37. M. Dangelantonio, M. Venturi and Q. G. Mulazzani, Rad. Phys. Chem., 1988, 32,
319-324.

38. J. Lind, X. H. Shen, T. E. Eriksen, G. Merenyi and L. Eberson, J. Am. Chem. Soc.,
1991, 113, 4629-4633.

39. D. Zehavi and J. Rabani, J. Phys. Chem., 1972, 76, 312.

40. B. Cercek, M. Ebert, A. J. Swallow and J. P. Keene, Science, 1964, 145, 919.

41. L. I. Grossweiner and M. S. Matheson, J. Phys. Chem., 1957, 61, 1089-1095.

42. G. N. R. Tripathi, R. H. Schuler and R. W. Fessenden, Chem. Phys. Lett., 1985,
113, 563-568.

43. R. Wilbrandt, N. H. Jensen, A. H. Sillesen and K. B. Hansen, Chem. Phys. Lett.,
1984, 106, 503-507.

44. B. J. Greenblatt, M. T. Zanni and D. M. Neumark, Faraday Discuss., 1997, 101-
113.

45. M. L. Alexander, N. E. Levinger, M. A. Johnson, D. Ray and W. C. Lineberger, J.
Chem. Phys., 1988, 88, 6200-6210.

46. O. Horvath, Stevenson, K., Charge Transfer Photochemistry of Coordination
Compounds, VCH Publishers, New York, 1993.

47. J. C. Barnes and D. N. Hume, Inorg. Chem., 1963, 2, 444.

48. K. L. Stevenson, D. W. Knorr and A. Horvath, Inorg. Chem., 1996, 35, 835-839.

109

Chapter 4
Luminescence Lifetimes of CuBr
3
2
ˉ Measured by Time-Correlated-Single-Photon-
Counting

4.1 Abstract for Chapter 4

Time-correlated-single-photon-counting (TCSPC) experiments were performed on UV
excited CuBr
3
2
ˉ to obtain an independent estimate on the intersystem crossing (ISC) rate
between the initial CuBr
3
2
ˉ charge-transfer-to-solvent (CTTS) state and the triplet state,
in the scheme proposed in Chapter 2 of this work. In addition to phosphorescence from
the triplet state, a fast component attributed to fluorescence of the CTTS state is
observed. Based on instrument limited decay of the fluorescence and instrument limited
rise in the phosphorescence components, an upper limit of 8 ps is obtained for the inverse
ISC rate. Because the fluorescence component is observed at all, suggests that ISC is not
more than an order of magnitude and other emission experiments that could be used to
obtain a more accurate determination are discussed. We also obtain confirmation that the
nitrate ion diffusively quenches the CuBr
3
2
ˉ triplet state and also quenches its precursor.
All results of this Chapter are consistent the assignment of the ultrafast pump-probe
broadband data of Chapter 2 and help to elucidate the nature of the CuBr
3
2
ˉ CTTS state.

110
4.2 Introduction

In Chapter 2 of this work, a kinetic scheme was proposed to explain the transient spectral
dynamics of photoexcited CuBr
3
2ˉ
obtained from ultrafast one and two color pump-probe
and pump-broadband probe experiments. The time window accessed in the ultrafast
experiments allows the observation of earlier transient dynamics of excited states and
products than previously seen in the copper bromide system or, to the author’s
knowledge, any other copper halide and copper pseudohalide system.
1-17
In agreement
with the literature,
9, 10
the proposed kinetic scheme based on the results in Chapter 2
includes ISC from the CuBr
3
2ˉ
singlet CTTS state to the CuBr
3
2ˉ
triplet state (Figure
4.2.1). Ultrafast time resolution and constrained global analysis data fitting revealed this
ISC value to be close to 1 ps.
It is known that 266 nm excitation of CuBr
3
2ˉ
induces emission with a band
maximum red shifted ~200 nm from that of the ground state absorption band (See Figure
4.2.2a) with a lifetime on the order of 10s to 100s ns depending on ligand concentration,
ionic strength, and pH
9, 10, 18
as shown in Table 4.2.1. Due to the large spectral shift of
the emission band relative to the absorption band, and the long decay kinetics, the
luminescent species was assigned to a CuBr
3
2ˉ
triplet state which is prepared via ISC
from the initially prepared CTTS state.
18
The time-correlated-single-photon-counting
(TCSPC) technique utilized in our laboratory has a superior time resolution of ~20 ps
compared to the previously reported ns fluorescence lifetime experiments.
9, 10, 18

Therefore, luminescence lifetimes detected by this method could confirm or refute the
CuBr
3
2ˉ
CTTS singlet to triplet ISC rate. If the intersystem crossing rate for photoexcited
111
CuBr
3
2ˉ
is slower than the instrument response, the TCSPC counts detected at the triplet
luminescence wavelength would show a slower than instrument limited rise, indicating a
growth of the triplet population (Figure 4.2.2b, red trace). If the triplet state is already
fully populated within the instrument response, a rise will still be observed, however it
will be instrument limited (Figure 4.2.2b, black trace).

Ground
τ2a, relax
τ1, – Br
−
*[
1
CuBr
3
2 −
]
τ3, ISC
*[
3
CuBr
3
2 −
]
*[
1
CuBr
2
−
]
*[
3
CuBr
2
−
]
τ2b, ISC
Ground
τ2a, relax
τ1, – Br
−
*[
1
CuBr
3
2 −
]
τ3, ISC
*[
3
CuBr
3
2 −
]
*[
1
CuBr
2
−
]
*[
3
CuBr
2
−
]
τ2b, ISC

Figure 4.2.1. Proposed kinetic scheme as described in Chapter 2.

Although normally a routine measurement, TCSPC measurements of UV excited
CuBr
3
2ˉ
proved to be difficult due to the low luminescence quantum yield (QY) combined
with the slow radiative rate of the triplet. The QY for photoexcited trichlorocuprate is
0.0035 at 5 M Cl
ˉ
ligand concentration and decreases to ~0.0005 at 0.5 M Cl
ˉ

concentration.
19

112
Table 4.2.1. Luminescence lifetimes of 266 nm excited CuBr
3
2ˉ
under various ligand
concentrations, pH, and ionic strength conditions. The first 5 rows are taken from
Reference 9 and last 4 rows from Reference 10. Decays are reported as single
exponential and detection was at 480 nm. (Bold is comparison to Figure 4.4.1 results).

[Br
ˉ
]

pH
Ionic
Strength (M)
Luminescence
Lifetime (ns)
0.1 7 5 17
0.2 7 5 35
0.4 7 5 71
0.7 7 5 135
1.0 7 5 223
2 7 2 33
3 7 3 100
5 3 5 548
5 7 5 814

a b

3 3
Time
Photon Counts
3 3 3 3
Time
Photon Counts

Figure 4.2.2. a) Luminescence spectra of aqueous CuBr
3
2ˉ
taken from Reference 18, re-plotted. The
absorption and emission band maxima are 281 nm and 475 nm, respectively. b) Two scenarios for the
observed TCSPC signal. The rise in the black trace is an illustration of an instrument limited emission
signal in which the luminescing species has been populated within the time of the instrument
response. The red trace is the emission of a species that has a population growing in slower than the
instrument resolution.

113
Based on steady state luminescence spectra found in the literature that plot both
the chloro and bromo species, CuBr
3
2ˉ
has a QY four times higher that of trichlorocuprate
at 5 M ligand concentration.
18
An actual value is not reported for lower concentrations of
Br
ˉ
, yet as shown in Table 4.2.1, the emission lifetime and therefore intensity of CuBr
3
2ˉ

also drops with lowered ligand concentration. As reported in Reference 10, on a relative
scale, luminescence QY at 0.4 M Br
ˉ
ion concentration is less than 10% of the QY for 5
M Br
ˉ
. With very low photon counts, factors such as excitation laser scatter, water
Raman signal, water impurities, and luminescence of the cuvette could not be neglected.
Reducing the excitation power and filtering the emission from scattering were necessary
but further lowered photon count rates, dramatically decreasing the signal to noise ratio.

4.3 Experimental

Emission lifetimes were obtained through TCSPC measurements. A 250 kHz
repetition rate Ti:sapphire ultrafast system (Coherent Rega 9050), producing ~60 fs
FWHM 800 nm pulses was used to pump an optical parametric amplifier (OPA 9450,
Coherent) that was tuned to produce 532 nm pulses with ~15 nm bandwidth. The 532 nm
pulses were compressed with a fused silica prism pair to a pulse width of ~50 fs and then
focused into a 0.08 mm thick BBO crystal to generate the second harmonic
20
for use as
excitation pulses. The UV pulses were compressed with a CaF
2
prism pair to a temporal
pulse width of ~100 fs FWHM and with ~2-3 nm of bandwidth measured with a REES
spectrophotometer. Quenching experiments utilized 280 nm excitation pulses which
114
were obtained by OPA tuning to 560 nm and SHG as described for the 266 nm pulse
generation. The beam was focused into the sample using a 5 cm focal length lens to
ensure the beam waist was within the cuvette to minimize fluorescence from the cell with
a minimum beam spot size of 17 µm for the unquenched experiments. For the quenched
solutions, a 100 cm focal length lens was used to ensure a soft focal waist. The
pathlength was 1 mm for all except the quenching experiments, which used a 1 cm
pathlength cell. The excitation pulse power was adjusted by varying neutral density (ND)
filters before the sample to minimize background signal from scatter and weak
fluorescence of the cuvette fused silica or quartz and was typically below 1 mW, which at
a 250 kHz repetition rate delivers a maximum value of 1.4 x 10
4
W/cm
2
for peak
intensity. As discussed in Section 2.5.1 in Chapter 2, there is less than 10% absorption of
photons by the dibromocuprate ground state species at 266 nm and an even smaller
percentage at 280 nm.
The luminescence of each sample was collected with a 1” diameter 3.5 cm focal
length lens placed perpendicular to the incident 266 nm beam and focused into a
monochromator with a 1” diameter 10 cm focal length lens. The monochromator was a
CVI CMSP112 double spectrograph with a 1/8 m total path length in negative dispersive
mode with a 600 groove/mm grating, and had an f number of 3.9. The slit width was
typically 0.6 mm, which together with a monochromator dispersion of 16 nm/mm
translates to ~10 nm bandpass for our experiments. Detection wavelengths were varied
by adjusting the angle of the monochromator grating and ranged from 266 nm to 600 nm.
Filters were placed before the entrance slit of the monochromator to reduce scattered
pump light. At detection wavelengths of 475 nm, 360 nm, and 335 nm, interference
115
filters with ~20 nm bandwidths and band maxima at 475 nm, 355 nm, and 340 nm were
used, respectively. A broadband polarizing cube was set at magic angle to eliminate
possible anisotropy in the luminescence signal. The PMT was a Hamamatsu RU3809
micro-channel plate detector biased at -3.0 kV and was mounted at the monochromator
exit slit.
Pulse pileup was avoided in these experiments simply because count rates were so
low, with a peak count ~100. The instrument response is ~20 picoseconds determined by
measuring scatter from a scattering solution. Samples were held in Spectrosil quartz
cuvettes purchased from Starna Cells with an internal path length of 1 mm or 1 cm. A
Becker and Hickl SPC-630 photon counting board recorded the amplified emission
signals. The reference signal was provided by a portion of the excitation beam sent to a
fast photodiode.
Sample preparation is described in the experimental section of Chapter 2. The
samples used in the TCSPC experiments had varied ligand concentrations of 0.4 M Br
ˉ
to
2 M Br
ˉ
and all were aqueous. Ionic strength was held constant when needed by addition
of NaClO
4
. KNO
3
was used in the quenching experiments and added to the solution to
yield a NO
3
ˉ
concentration of 0.4 M. Aside from the 2 M Br
ˉ
ligand solution (data shown
in Figure 4.4.1), following preparation, oxygen was not excluded from solutions. Since
oxygen is a known triplet quencher, the long time lifetime component has a faster decay
here than reported values (See Table 4.2.1). However, reproducing the triplet lifetime
was not the focus of the current experiments.

116
4.4 Results

Figure 4.4.1 is the TCSPC measurement of 266 nm excited CuBr
3
2ˉ
detected at
475 nm and magic angle polarization. The excitation power was 240 µW and the
solution was aqueous 2 M Br
ˉ
with ~2 M total ionic strength. Poor signal to noise in the
measurements presented in this section was unavoidable due to the very low count rates
as a result of small QY of these copper halide systems, as discussed in Section 4.2. For
the results in Figure 4.4.1, the collection was taken over the course of 4800 seconds in
order to increase counts while maintaining experimental conditions that minimize
background to obtain a total peak count rate of ~250 at zero time delay. For comparison,
a solution of 0.6 OD fluorene in cyclohexane, excited with just 14 µW at 266 nm,
detected at magic angle with the same slit width (0.6 mm), and with a collection time of
600 s, yields a peak count rate of >4000. With the same excitation power and collection
time used for the CuBr
3
2ˉ
sample, the peak photon count would be over 320,000 (which is
over the limit of the experimental collection capability) for the fluorene sample.
The CuBr
3
2ˉ
luminescence shows only two decay components. The fast decay
appears to be instrument-limited and has never been reported in the literature. More
analysis on this feature is presented below. The longer time component is consistent with
an exponential decay of 33 ns, consistent with the triple lifetime measured for a CuBr
3
2ˉ

solution under similar conditions (See Table 4.2.1, row 6 in bold).
9
However, the total
time range for the TCSPC experiments here was limited to ~5 ns in order to achieve the
best time resolution since the purpose of these experiments is to bracket the ISC rate. Fig.
4.4.1 confirms that triplet formation is faster than the instrument response. Details on
117
control experiments that confirm the fast component is not due to other effects besides
CuBr
3
2ˉ
luminescence is described in detail below.

0 1 2 3 4
0
50
100
150
200
250
0.0 0.5
0
200

Counts
Time, ns

Photon Counts
Time, ns

Figure 4.4.1. TCSPC measurement of 266 nm excited aqueous CuBr
3
2ˉ
in 2 M Br
ˉ
concentration
under nitrogen (black trace). Fit to exponential decays (red trace). (inset) Time window of <0 ps to
600 ps. Experimental conditions were ~230 µW excitation power, 0.6 mm slit width, magic angle
detection with a 475 nm interference filter before monochromator to further block stray photon scatter.

Measurement of the 2 M [Br
ˉ
] solution (Figure 4.4.1) confirms the long lifetime is
in agreement with literature values and that we are indeed observing the CuBr
3
2ˉ
triplet
state. It was also desirable to perform experiments under as close to the same conditions
as the broadband experiments as possible (described in Section 2.3 of Chapter 2). In
addition, of particular concern was to confirm that the fast time component is due to an
instrumentally limited fluorescence and not scatter (Rayleigh or Raman), water
118
impurities, or fast luminescence from the cuvette. Figure 4.4.2a shows the TCSPC data
for 266 nm excited aqueous CuBr
3
2ˉ
solution at ~0.4 M total ionic strength (other details
are in the caption). The detection wavelength was varied to determine if the fast
luminescence decay component could be separated from the long component and the fast
emission band uncovered. The long decay emission component almost completely
disappeared at both 335 nm and 360 nm detection, indicating that the fast decaying
species indeed emits at these shorter wavelengths. (In Figure 4.4.2a, the 360 nm data set
(blue trace) and 335 nm data set (red trace) are offset vertically in order to see the data
more clearly). The finding that this fast component is observed at all of the detection
wavelengths for CuBr
3
2ˉ
, 335 nm, 360 nm, and 475 nm, was surprising because it
indicates the early emission band is very broad and largely overlays the triplet band.
To verify that the results are due to CuBr
3
2ˉ
and not background scatter, we first
validate the fast component seen across all detection wavelengths. A scattering solution
was used to collect the instrument response function (IRF) of the TCSPC experimental
setup, using 266 nm excitation and 266 nm detection. The IRF is shown in Figure 4.4.2a
(black trace) and has a FWHM of ~25 ps. To determine if stray light scattered in the
monochromator was also responsible for the instrument limited component at 335 and
360 nm detection in Figure 4.4.2a, a control experiment was conducted using the same
scattering solution and emission detection wavelengths as CuBr
3
2ˉ
. Figure 4.4.2b
compares a CuBr
3
2ˉ
solution (red trace) with that of the scattering solution (black trace)
under identical experimental conditions, recorded back to back. As can be seen in the un-
normalized results, the photon counts at zero delay due to stray light of the scattering
solution is approximately 5% that from CuBr
3
2ˉ
. Similarly, detection of the scattering
119
solution compared to the CuBr
3
2ˉ
solution at 475 nm shows virtually no scatter
contribution to the time profile nor background luminescence from impurities or the cell
(Figure 4.4.3a). we note that the Raman response of water is not a contributing source of
signal as it should appear at 295 nm, and even with the broad monochromator bandpass
used in these experiments, it is certainly not contributing at 335 nm.

a b
0 1000 2000
0.0
0.2
0.4
0.6
0.8
1.0
1.2
-50 0 50 100
0.0
0.5
1.0

Counts
Time, ps

Normalized Photon Counts
Time, ps

0.0 0.5 1.0
0
50
100
150
CuBr
3
2-
Scattering Solution

Photon Counts (Raw)
Time, ns

Figure 4.4.2. a) TCSPC data for 0.4 M Br
ˉ
CuBr
3
2ˉ
aqueous solution and scattering control solution
excited at 266 nm. Luminescence was detected at 475 nm (green trace), 360 nm (red trace), and 335
nm (blue trace). Different interference filters were placed before the monochromator entrance for
each detection wavelength and the slit width was 2.4 mm throughout. The 360 and 335 nm data is
translated vertically for ease of viewing. The scattering solution used for IRF (black trace) was
detected at 266 nm. Excitation power was 850 – 950 µW. The faster than expected long component
shown in the green trace is due to oxygen quenching of the triplet state since these samples were under
ambient conditions following preparation. (Inset) Time window of -50 ps to 2000 ps (same color
scheme, the 360 nm dataset is not translated in this panel). b) Raw TCSPC data for CuBr
3
2ˉ
solution at
0.4 M Br
ˉ
ligand concentration and ~0.4 M ionic strength compared with a scattering solution. 335
nm detection at magic angle with a 340 nm interference filter placed before the monochromator
entrance slit of 2.4 mm. Collection time 600 s. Excitation for both samples is 950 µW of 266 nm.

120
Other than the sharp spike, there is a very small amplitude long-time decay
component in the 335 nm and 360 nm datasets (difficult to see in Figure 4.4.2a).
However, this has been determined to originate from the very small impurity emission
from the water used to make the CuBr
3
2ˉ
solutions (Figure 4.4.3b). However, it is
important to point out that this is not the case for 475 nm detection. Figure 4.4.3a clearly
shows that the water impurity contribution at 475 nm is small compared to the copper
complex phosporescence.

a b
0 1 2 3 4
0
20
40
60
80
Photon Counts
Time, ns

5 10 15 20
0
20
40
60
80

Counts
Time, ns

Figure 4.4.3. a) Back to back TCSPC measurements of 1 M Br
ˉ
CuBr
3
2ˉ
solution (red trace) taken for
a duration of 1200 seconds and scattering control solution measurement taken for 600 seconds (black
trace). ~230 µW 266 nm excitation, 475 nm detection at magic angle, 475 nm filter before
monochromator entrance slit width of 0.6 mm, 5 ns time window. b) Neat water excited with 260 µW
of 266 nm and detected at 360 nm with no polarizer and no filter before monochromator entrance slit
width of 2.4 mm. Collection time was 1200 s with a 50 ns time window. Both a) and b) solutions
were under ambient conditions following preparation.

The above results agree with the proposed schematic shown in Figure 4.2.1. The fast
time component seen at all detection wavelengths is attributed to fluorescence of the
121
CuBr
3
2ˉ
CTTS state and the long component at 475 nm is assigned to phosphorescence of
the CuBr
3
2ˉ
triplet state. These assignments are discussed further in Section 4.5.

The next set of TCSPC experiments involved nitrate emission quenching. We have
observed time-dependent reduction of the broadband transient absorption induced by H
+
ions, a known CuBr
3
2ˉ
triplet quencher.
18
We have also observed transient absorption
quenching of the band we have assigned to the triplet also with NO
3
ˉ
. However, there
were some questions due to superimposed products of the quenching reaction, and here
we can more clearly determine if the triplet phosphorescence is quenched by nitrate.

Figure 4.4.4 shows the TCSPC results of CuBr
3
2ˉ
with and without potassium nitrate. In
these quenching experiments, the excitation was at 280 nm and 1.24 mW, with detection
at 475 nm, magic angle, and 0.6 mm slits used at the entrance of the monochromator.
The scans were taken under the same conditions and back to back. While no effort was
made to minimize background signal from photon scatter, water impurities, or cell
luminescence as in the other TCSPC experiments in this Chapter, there still are three
important observations. First the nitrate increases the decay rate of phosphorescence.
Second, the initial quantum yield is much larger for the CuBr
3
2ˉ
solution without NO
3
ˉ

than with NO
3
ˉ
. Third, the rate constant of ~4 x 10
9
M
-1
s
-1
calculated by the 400 nm
decay in the pump-probe transient spectral data for CuBr
3
2ˉ
with 0.4 M NO
3
ˉ
(see Chapter
2, Section 2.5), yields a decay rate of ~1250 ps at 0.2 M NO
3
ˉ
, which fits the quenched
TCSPC data down to 5%, shown by the smooth black curve in the inset of Figure 4.4.4.
The duration of signal below 5% is likely due to solvent impurities (See Figure 4.4.3) or
122
cell fluorescence at these higher excitation intensities. (Note that for the red curve, the
number of photon counts after ~10 ns is virtually zero).

0 5 10 15 20
1000
2000
3000
0 10 20 30
0.01
0.1
1

Photon Counts, log scale
Time, ns

Photon Counts
Time, ns

Figure 4.4.4. Luminescence lifetime traces for 280 nm excited, 475 nm detection of ~0.004 M
CuBr
3
2ˉ
in 0.2 M Br
ˉ
concentration without quencher (black trace) and with 0.2 M KNO
3
(red trace)
recorded back to back. Inset is normalized and on a log scale and includes fit to quenched solution
(smooth black trace). The fit is biexponential with a 20 ps and 1250 ps time. Solutions were under
ambient conditions.

4.5 Discussion

Our results confirm that ISC in the UV excited CuBr
3
2ˉ
system occurs within the
instrument response (~20 ps) as no rise is observed in the triplet phosphorescence, but
that the process is slow enough to observe traces of singlet fluorescence. Although the
CuBr
3
2ˉ
early time fluorescence component and the instrument response function almost
overlay, as seen in Figure 4.5.1a, we can now find an upper limit for the decay constant
123
of the fluorescence and therefore the triplet formation time. As is well known in TCSPC,
a 20 ps IRF does not mean that a 20 ps single exponential decay is the minimum
resolvable lifetime. Figure 4.5.1b shows that we can determine an upper limit to the
fluorescence decay time by comparison with different exponential decays convoluted
with the measured IRF. It is obvious from Figure 4.5.1b that 20 ps, 15 ps, and 10 ps
decaying exponential do not reproduce the data, but based on the signal to noise, it is
difficult to decide between exponentials 8 ps or faster as best fits to the data. Therefore,
we place an upper limit for the triplet formation rate at 8 ps. However, as shown in
Figure 4.5.1c, decreasing the lifetime still further would decrease the absolute height of
the instrument limited spike until eventually no emission is detectable. This analysis
illustrates that the singlet state responsible for fluorescence here cannot have a lifetime
substantially shorter than 1 ps. However, without systematic comparisons to other
chromophores with known sub-picosecond lifetimes and similar radiative rates under the
same conditions, we cannot extract a lower limit to the triplet formation time.
The observation of fluorescence from the CuBr
3
2ˉ
system is interesting in several
ways. Most impressive is the actual observation of the fluorescence from a CTTS state.
This is a novel finding since primary CTTS states are generally very short lived, < 100 fs,
and collection of CTTS emission has not been reported before. The longer than usual
singlet CTTS lifetime of the CuBr
3
2ˉ
means that somehow the CTTS state here is less
fragile to water rearrangement than other atomic anion CTTS states studied to date.
21-25

Techniques with higher time resolution than TCSPC, such as fluorescence upconversion
spectroscopy or Kerr gating fluorescence might be able to directly recover the
fluorescence lifetime of the CuBr
3
2ˉ
singlet CTTS state. Amongst CTTS systems, this is
124
a great candidate for ~100 fs time resolution fluorescence measurements because of its
relatively long initial state lifetime (these ideas are expanded in Chapter 6 Future Work).
Another important outcome of this measurement is the confirmation that triplet formation
in copper coordination complexes can indeed be faster than the 10s ps reported by Shaw
et. al.
26
Other metal complexes such as iridium and ruthenium have been shown to
undergo sub-ps ISC rate.
The nitrate experiments confirm that the transient broadband quenching is in fact
due to the reaction of nitrate with the CuBr
3
2
ˉ triplet state. As seen in Figure 4.4.4, not
only is the lifetime of the triplet shortened, but the initial yield is much larger for the
CuBr
3
2
ˉ without NO
3
ˉ than for the solution with NO
3
ˉ. The increase of the
phosphorescence decay rate means that nitrate is diffusively reacting with the triplet state.
This helps to conclusively assign the spectral component of the UV broadband spectrum
in the pump-probe experiments that is influenced by nitrate to an excited state and
specifically the triplet. The agreement of the quenching rate constant of ~4 x 10
9
M
-1
s
-1

obtained from the pump-probe data (Chapter 2) with the decay of phosphorescence is
further support of this assignment. The reduction of prompt emission QY is a clear
indication that the triplet precursor (CuBr
3
2
ˉ CTTS state) is also quenched (static
quenching). This type of quenching involves the non-diffusive electron transfer reaction
of NO
3
ˉ
with excited states that have spatially extended orbitals, including the water
solvated electron precursor and CTTS states.
27
This is an indication that, while the
CuBr
3
2
ˉ CTTS state behaves differently than other known CTTS states, its excited state
wavefunction is still relatively diffuse.

125

a b
-50 0 50 100 150 200
0.0
0.5
1.0

Photon Counts, Normalized
Time, ps
CuBr
3
2-
IRF

-50 0 50 100 150 200
0.0
0.5
1.0

Normalized Photon Counts
Time, ps
CuBr
3
2-
20 ps
15 ps
10 ps
8 ps
5 ps

c
1
0
Absolute Scale
Time
20 ps
15 ps
10 ps
5 ps
1 ps

Figure 4.5.1. a) 266 nm excited CuBr
3
2ˉ
0.4 M Brˉ solution at ~0.4 M ionic strength detected at 335
nm (black dots) and 266 nm excited 266 nm detected scattering solution (red trace) as IRF
measurement. Time is offset from zero; peak of photon counts taken as time zero. b) Overlay of
TCSPC CuBr
3
2ˉ
data with the convolution of the data with exponential functions. c) Absolute value of
convolution of IRF with varying exponential decays. Decreasing the exponential decay time
decreases its contribution to the overall signal to the limit where the instrument is insensitive to the
fluorescence and no signal will be observed. Solutions kept under ambient conditions following
preparation.

126
4.6 Conclusions

In this work, both fluorescence and phosphorescence were observed and assigned to the
CuBr
3
2
ˉ CTTS singlet and triplet states. The upper limit to the intersystem crossing rate
between these two species was found. Contrary to other reports on copper complexes,
this intersystem crossing rate is less than 8 ps. Nitrate was found to quench the
phosphorescence lifetime and provide insight to the size of the CTTS orbital since NO
3
ˉ
statically reacts with species that have diffuse electronic orbitals. The TCSPC technique
provides a nice link between the ultrafast and nanosecond experiments reported in the
literature because it covers an intermediate time window. This time window, along with
the experimental results of TCSPC and pump-probe spectroscopy, helps confirm the
proposed scheme in Chapter 2.

It must be noted that Stevenson and Horvath reported a slight increase in luminescence
quantum yield and a 20 nm red shift in the emission spectrum of queous chlorocuprate at
lower ligand concentrations.
9
This led them to believe that the dichloro- species is also
luminescent. As the bromocuprates behave very similarly to other halocuprate systems,
Stevenson and Horvath concluded that the CuBr
2
ˉ species must also emit. However, at
the Brˉ ligand concentrations used for the experiments of this work, we do not believe
CuBr2ˉ to contribute significantly to emission or transient absorption results.

127
4.7 References for Chapter 4

1. C. R. Davis and K. L. Stevenson, Inorg. Chem., 1982, 21, 2514-2516.

2. D. D. Davis, Stevenson, K. L., Davis, C. R., J. Am. Chem. Soc., 1978, 100, 6.

3. T. F. Braish, R. E. Duncan, J. J. Harber, R. L. Steffen and K. L. Stevenson, Inorg.
Chem., 1984, 23, 4072-4075.

4. O. Horvath, J. H. Fendler and K. L. Stevenson, Inorg. Chem., 1993, 32, 227-230.

5. O. Horvath and J. H. Fendler, J. Photoch. Photobio. A, 1993, 71, 33-37.

6. A. Horvath, O. Horvath and K. L. Stevenson, J. Photoch. Photobio. A, 1992, 68,
155-163.

7. A. Horvath and K. L. Stevenson, Inorg. Chem., 1993, 32, 2225-2227.

8. O. Horvath and K. L. Stevenson, Inorg. Chem., 1989, 28, 2548-2551.

9. K. L. Stevenson, R. S. Dhawale, A. Horvath and O. Horvath, J. Phys. Chem. A,
1997, 101, 3670-3676.

10. K. L. Stevenson, D. W. Knorr and A. Horvath, Inorg. Chem., 1996, 35, 835-839.

11. A. Horvath and K. L. Stevenson, Coord. Chem. Rev., 1996, 153, 57-82.

12. K. L. Stevenson and J. H. Jarboe, J. Photoch. Photobio. A, 2002, 150, 49-57.

13. K. L. Stevenson, P. B. Bell, O. Horvath and A. Horvath, J. Am. Chem. Soc., 1998,
120, 4234-4235.

14. K. L. Stevenson, P. B. Bell, R. S. Dhawale, O. Horvath and A. Horvath, Radiat.
Phys. Chem., 1999, 55, 489-496.

15. A. Horvath and K. L. Stevenson, Coord. Chem. Rev., 2000, 208, 139-151.

16. K. L. Stevenson, P. B. Bell and R. E. Watson, Coord. Chem. Rev., 2002, 229,
133-146.

17. K. L. Stevenson, J. H. Jarboe, S. A. Langmeyer and T. W. Acra, Inorg. Chem.,
2003, 42, 3559-3564.

18. K. L. Stevenson, R. M. Berger, M. M. Grush, J. C. Stayanoff, A. Horvath and O.
Horvath, J. Photoch. Photobiol. A, 1991, 60, 215-227.
128
19. K. L. Stevenson, J. L. Braun, D. D. Davis, K. S. Kurtz and R. I. Sparks, Inorg.
Chem., 1988, 27, 3472-3476.

20. J. A. Kloepfer, V. H. Vilchiz, V. A. Lenchenkov and S. E. Bradforth, Chem. Phys.
Lett., 1998, 298, 120-128.

21. X. Y. Chen and S. E. Bradforth, Annu. Rev. Phys. Chem., 2008, 59, 203-231.

22. M. K. Fischer, A. Laubereau and H. Iglev, Phys. Chem. Chem. Phys., 2009, 11,
10939-10944.

23. A. Kammrath, J. R. R. Verlet, A. E. Bragg, G. B. Griffin and D. M. Neumark, J.
Phys. Chem. A, 2005, 109, 11475-11483.

24. J. A. Kloepfer, V. H. Vilchiz, V. A. Lenchenkov, A. C. Germaine and S. E.
Bradforth, J. Phys. Chem., 2000, 113, 6288-6307.

25. F. H. Long, H. Lu, X. L. Shi and K. B. Eisenthal, Chem. Phys. Lett., 1990, 169,
165-171.

26. G. B. Shaw, C. D. Grant, H. Shirota, E. W. Castner, G. J. Meyer and L. X. Chen,
J. Am. Chem. Soc., 2007, 129, 2147-2160.

27. T. W. Kee, D. H. Son, P. Kambhampati and P. F. Barbara, J. Phys. Chem. A,
2001, 105, 8434-8439.
129

Chapter 5
The Effects of β-Mercaptoethanol on the Luminescence Lifetimes of CdSe,
CdSe/ZnS, and CdTe Quantum Dots

5.1 Abstract

The luminescence lifetimes of aqueous solutions of CdSe/ZnS QDs solubilized
with mercaptopropionic acid (MPA) in the presence of β-mercaptoethanol (BME) was
measured by time-correlated-single-photon-counting (TCSPC). The purpose of this study
was to determine if BME itself has an effect on these QD systems as BME was used as an
addition reagent in prior experiments designed to test the photoluminescence lifetimes of
CdSe/ZnS QDs conjugated to dopamine in varying redox potential environments. Our
experiments mimic cellular conditions better by using pH levels similar to live biological
systems and MPA concentrations that do not kill live cells, than other studies on the
effect of thiolates on CdSe/ZnS QD photoluminescence. It was found that BME does in
fact quench the lifetime of CdSe/ZnS solubilized by MPA. TCSPC experiments on bare
CdSe QDs and CdTe QDs were performed to verify a proposed mechanism that includes
competition between enhancement and quenching in all three of these systems.

130
5.2 Introduction

5.2.1 Quantum Dot Background

Colloidal quantum dots (QD) are spherical semiconductor nanocrystals ranging
from ~2 to 5 nm in diameter. These nanoparticles display bulk-like properties, such as
large spin-orbit coupling and large dielectric constants.
1
However, QDs also possess
properties unlike the bulk, the most important being their size dependent features. Band
gap energies and emission wavelengths are dependent on the diameter of the quantum
dot, in which the energy of the band gap increases as the diameter of the QD decreases.
2

Exciton binding energy and exchange interactions are also size dependent.
1
Presently,
the most commonly studied QDs are CdSe, CdS, CdTe, InP, PbSe, and PbS.
1

The band gap in the bulk semiconductor is defined as the energy required to
create an electron and a hole.
3-5
The result of photoexcitation of an electron from the
valence band into the conduction band of a semiconductor can be described as a Wannier
exciton in which the system is approximated by the hydrogenic Hamiltonian.
6
The
effective Bohr radius of the exciton is the spatial separation between the electron and hole
and is given by the equation
2
2
e
a

 

5.1
where a is the Bohr radius, ћ is Planck’s constant, ε is the semiconductor dielectric
131
constant, and µ is the reduced mass of the electron and hole. The size of this radius
controls how large a crystal must be for its energy bands to be treated as continuous.
In 1983, synthesis of 10 Å diameter semiconductor crystals was achieved, and it
was discovered that these crystal “clusters” had band gap energies several tenths of an eV
higher than that of the same material in the bulk.
7
It was also observed in the initial
experiments on ZnSe clusters that the absorption spectra of small particles (under 100 Å)
displayed features toward the blue end of the spectra, unlike the typical absorption
spectrum of a bulk semiconductor resembling a step function.
4
Above 100 Å, the
particles showed bulk-like properties and when the size of the particles decreased to 40-
50 Å, the absorption band edge of the nanoparticle began to shift toward the ultraviolet.
At 20-30 Å, the blue shift was greater than 1 eV, and the blue end of the spectrum
became structured due to the discrete energy bands in the particle. Quantization of
nanoparticles had already been predicted in the literature by this time.
1
Other early
materials studied were CdS, Ag-halides, Zn and Cd phosphides and arsenides, CuCl, PbI
2

and Fe
3
O
2
layers.
4

In his 1984 article in the Journal of Chemical Physics, Brus showed that in the
limit of Angstrom sized semiconductors, the band gap energy becomes
...
8 . 1 1 1
2
*
2
2
2 2
 






  
R
e
m m R
E E
h e
g

 
5.2
where E* is the minimum energy needed to create an exciton, E
g
is the band gap energy
132
of the bulk, ћ is Planck’s constant, R is the distance between the excited electron and the
hole, m
e
is the effective mass of the electron, m
h
is the effective mass of the positive hole,
and ε is the semiconductor dielectric constant.
5
The second term is the quantum
localization energy and is proportional to 1/R
2
. The third term is formally the Coulomb
attraction, and higher terms correspond to energy lost due to solvation and is related to
polarization effects. This polarization energy is positive and smaller than the magnitude
of the Coulomb energy. Coulomb attraction favors the electron and hole to come
together and the polarization term favors the electron and hole to reside in the center of
the nanocrystal where there is maximum dielectric stabilization. However, both the
Coulomb and polarization effects are counteracted by increasing kinetic energy as the
electron and hole separation becomes smaller. The net result is a blue shift in the energy
of the nanocrystal semiconductor versus its bulk when its radius becomes smaller than
the Bohr exciton radius, as was experimentally shown by Rossetti, et. al.
7

As described above, the size of the semiconductor nanocrystal is responsible for
properties unique to quantum dots.
8
While the crystal lattice structure is unchanged, the
electronic wavefunction of the excited electron is different in the QD continuum than in
the bulk continuum because the electron and hole are limited to the space of the QD.
4

This is known as carrier confinement. Aside from a blue shift in band gap energy, the
radiative lifetimes of QDs become faster as the size gets smaller, due to a greater
electron/hole overlap. Another characteristic of quantum dots is a consequence of the
large surface to volume ratio, which leads to a heavy dependence of QD photophysics on
its surface properties.
2

133

Figure 5.2.1. Example of aqueous nanoparticle (CdSe/ZnS QD) absorption (black line) and
fluorescence spectra (red line). The exciton peak is approximately at 585 nm and the fluorescence
peaks at 595 nm. There are features toward the blue side of the absorption band, unlike the bulk
absorption spectrum that resembles a step function (not shown).

5.2.2 Photophysics

The one photon dynamics of quantum dots (QDs) are such that when a photon of
equal or higher energy than the bandgap of the QD is absorbed, an electron from the
valence band is promoted into the conduction band, forming an exciton, or electron-hole
pair (see Figure 5.2.2). Quantum dot electron (e
-
) – hole (h
+
) pairs are of a distinct class
of excitons that either can be thought of as a confined bulk-type or molecular excitation.
1

An ideal exciton, once formed, will undergo electron-hole recombination through varying
relaxation pathways and kinetics, both radiative and non-radiative. First, in an isolated
134
Shallow Trap
States
Conduction Band
+
e-
Valence Band
Deep Trap
States
Band Edge
Fluorescence
IR
Fluorescence
Radiative and
Non-radiative
Decay
E
h ν
Hole Traps
1
3
1
2
3
4
Shallow Trap
States
Conduction Band
+
e-
Valence Band
Deep Trap
States
Band Edge
Fluorescence
IR
Fluorescence
Radiative and
Non-radiative
Decay
E
h ν
Hole Traps
1
3
1
2
3
4

Figure 5.2.2. Formation and relaxation pathways for a quantum dot exciton. An electron is promoted
by a photon into the conduction band, leaving behind a hole in the valence band. 1) The electron and
hole relax to the band edges. 2) The electron radiatively recombines with the hole. 3) The electron
falls into a shallow trap state near the band edge and then either radiatively or non-radiatively
recombines with the hole, and/or the band edge hole falls into a trap state near the valence band edge.
4) The electron falls into a deep trap state and radiatively recombines with the hole.

singly excited QD model, the e
-
and the h
+
relax toward the conduction and valence band
edges, respectively, on an ultrafast timescale. From this point forward, there are various
competing pathways the system can undergo. One is emission as the e
-
and h
+
recombine
from the band edge. For CdSe QDs, this occurs on the 10s of nanoseconds time scale. A
faster recombination mechanism that can be either radiative or non-radiative is the
electron falling into a trap state that lies near the edge of the conduction band then
recombining with the h
+
. This latter process occurs faster than radiative decay from the
band edge. Similarly, the h
+
could get caught in a trap state near the valence band
135
edge. Another route is the electron entering a deep trap state, or mid-band gap state.
Radiation from this state is in the infrared and happens on a much slower time scale than
the above two pathways. Single exponential luminescence decay lifetimes of QD
excitons indicate deactivation through one pathway only and brightly fluorescing QDs
follow the band edge radiative relaxation. Multiexponential decay is attributed to a
combination of the above mentioned relaxation mechanisms.
9

5.2.3 Surface

The trap states described above and depicted in Figure 5.2.2 originate as a result
of the nature of the QD surface.
2
While the interior crystal lattice structure of a QD is the
same as in the bulk, on the surface there are breaks in the lattice, which give rise to these
surface trap states. Trap states lower the quantum yield of fluorescence by preventing
radiative recombination from the band edge. In the case of CdSe QDs, these sites are the
Cd and Se surface unsaturated bonds.
10
It has been reported that 70% of the surface
atoms are Cd and 30% are Se.
11

Contrary to the photoexcited electron, the hole is thought to reside at the center of
the QD core.
12
Yet because most surface Se are not passivated, there is much opportunity
for the hole to become trapped at the surface.
12
To our knowledge, the characteristics of
a trapped hole on a Se dangling bond are not fully understood. There are, however, ab
initio quantum chemical calculations
13
that show that a CdSe nanoparticle has a relaxed
surface (deviating from interior wurtzite lattice structure) that maintains the
136
localization of the HOMO on the Se atom, while the LUMO transitions from a Cd state to
an Se state. Positron annihilation spectroscopy
14
that suggests that the CdSe surface is
also Se rich. It is therefore shown that there is a greater propensity of the hole, rather
than the electron, to trap on the surface based on the large number of Se hole trapping
sites.
12

Photoenhancement and improved stability of quantum dots are often achieved by
the addition of a coating made out of a semiconductor material of higher band gap
energy. This process is based on concepts of electronics band-gap engineering.
15
The
shell material binds epitaxially, thus preserving internal symmetry of the QD.
16
The first
reported coating, or shell, for CdSe QDs was ZnS.
17
Shells provide a tunneling barrier
that lowers the electronic coupling between the CdSe states and ligands and also provides
a physical barrier between the QD and its environment, decreasing photodegradation and
enhancing quantum yield.
18
Aside from partial leakage of the exciton into the shell
material,
17
the electron-hole pair is mainly confined to the core semiconductor material.
19

Ligands such as trioctylphosphine oxide (TOPO), which are used for
solubilization of the QDs into organic solvents, exclusively bind to Cd atoms on the
surface of the QD, with either one or two ligands per Cd and none for Se.
20
The number
of ligands binding to cadmium ions was obtained from ab inito calculations of structure
and binding for Cd
2
Se
2
clusters.
20
From these calculations, it was also determined that
the Se atoms do not have any ligand binding. Since Se atoms in nanocrystals are less
electrophillic than for clusters, the nanoparticle surface should also not have any ligand
coordination to Se. It has been deduced that the influence of Cd is largely passivated,
137
or in other words, the hole trap is removed with ligand binding.
11
Passivation leads to
enhanced photoluminescence quantum yield since this closes off nonradiative channels.
11

While Se does not show to bind to ligands, there is still a possibility of Se passivation by
some molecules, such as oxygen and water.
11

5.2.4 Environment

Quantum dots are utilized in various biological applications because of their
relatively high fluorescence quantum yield, narrow band emission, large effective Stokes
shift, size tunability and low photobleaching.
21
All of these optical properties make
quantum dots advantageous over conventional organic and protein dyes for cellular
labeling. Information about cellular structure and molecular markers, cell motility, cell
lineage, in-vitro tracking, and the monitoring of physiological events in live cells can be
gathered using luminescent QDs.
2
However, as discussed above, much of the QD
behavior is influenced by surface trap states and the shells and ligands used as passivants.
There is also a variability of fluorescence steady state quantum yields and lifetimes based
on the environment, such as pH, ionic strength, redox potential, and light illumination.
22

Furthermore, these environment effects on the luminescence behavior are not always
predictable and can be problematic for certain applications where the absence of
fluorescence is assumed to be due to the absence of the QD itself.

138
The positive side to the environment sensitivity of quantum dots is that this
feature can be exploited. For example, conjugation to bio-recognized molecules is a
means to give quantum dots specificity
21
to receptor mediated uptake as opposed to non-
specific endocytosis.
2
In Jay Nadeau’s laboratory at McGill University, CdSe/ZnS
quantum dots were conjugated to dopamine, enabling cellular uptake via the dopamine
receptor channels in the cell membrane and it was observed that CdSe/ZnS QD-dopamine
conjugates within live cells fluoresce brighter in oxidizing regions of the cell, implicating
the use for biological redox sensing probes.
22
In this example, electron transfer occurs
from the QD to dopamine following UV excitation. The dopamine then interacts with
oxidizing or reducing agents in the environment. The CdSe/ZnS conjugate system is
depicted in Figure 5.2.3.
To confirm the mechanism of electron transfer and its effects on the
photodynamics of CdSe/ZnS – dopamine conjugates, in collaboration with Nadeau’s
group, we used time-correlated-single-photon-counting (TCSPC) to measure
luminescence lifetimes of quantum dots and their conjugates. An effect of a reducing
agent on the lifetime of the QD-dopamine conjugate was established in past published
work and we proposed that the oxidized dopamine was reduced by β-mercaptoethanol
(BME) in the aqueous medium following QD excitation. From these experiments and in
agreement with the literature, we also found other factors that contributed to lifetime
changes, including illumination time and exposure to oxygen. Further studies were
performed to classify and understand the mechanisms for these changes.
23

139
QD
S
O
S
R
O
S
R
O
S
R
O
S
R
O
HO
HO
NH
QD
S
O
S
R
O
S
R
O
S
R
O
S
R
O
HO
HO
NH

Figure 5.2.3. CdSe/ZnS quantum dots are capped with 3-mercaptoacetic acid (MAA). Dopamine is
coupled to the MAA via an amide bond. In this picture there is just one dopamine conjugated to the
QD, however, the actual system has 100s of dopamine per QD.

One factor that needed attention was the interaction of the un-conjugated
CdSe/ZnS QD itself with BME, since this was the solution-phase reducing agent used
throughout in the electron transfer experiments that involved QD-conjugates. Jeong, et.
al.
24
studied the effects of ΒME alone on CdSe/ZnS QD and found ΒME to both quench
and enhance photoluminescence. However, Jeong’s experiments were done on QDs
solubilized via a polymer encapsulation method and in solutions of ΒME at
concentrations from 25 mM to 2500 mM. Our QDs were solubilized with either
mercaptopropionic (MPA) acid or mercaptoacetic acid (MAA). Further, we sought to
study the effects of BME using the same lower concentration of ΒME (1mM) as in
140
our original QD – dopamine experiments. This concentration regime is quite distinct -
ΒME is utilized in order to observe changes in the QD – DA fluorescence QY to study
cytotoxicity without itself causing cellular damage.
22
Other experimental conditions,
such as pH, buffer agents, and laser fluence needed to be the same as in our previous
experiments.

5.3 Experimental

Chemicals were purchased from Sigma-Aldrich Canada and US and used as
received. A detailed description of synthesis, solubilization and conjugation is reported
in References 23 and 25. A method utilizing the solvent 1-octadecene (ODE), as reported
by Asokan and coworkers,
25
was used in the synthesis of CdSe quantum dots. For TOPO
capping, elemental Se was incubated for no less than 24 hours in trioctylphosphine oxide
(TOPO), generating trioctylphosphine-Se. This was then put into a solution of TOPO,
along with the Cd precursor, and baked for 30 seconds to 30 minutes at 250 to 350°C.
The QDs were washed several times in methanol and then dissolved into
dichloromethane. The QDs were purified and stored in the ODE/hexanes phase until
further use. A similar synthesis technique was used for CdTe QDs, although nucleation
occurred in the presence of Cd
0
as reported by Kloepfer and coworkers.
26

To make CdSe/ZnS core/shell QDs, a zinc sulfide (ZnS) precursor was prepared
and injected over a time course of 5 min during the desired stage of QD growth. The
141
temperature was then allowed to drop to 100°C and it was maintained at this temperature
for several hours. The ZnS shell included 2-3 monolayers, estimated based on the
standard synthesis procedure used.
27
Purified QDs were finally stored in the
ODE/hexane phase until further use.
The QDs were solubilized in water by mass action replacement of the TOPO by
the thiol end of the polar molecules 3-mercaptopropionic acid (MPA),
(HSCH
2
CH
2
COOH), or 3-mercaptoacetic acid (MAA), under pH conditions of 8-10. To
achieve this, ~0.5 mL of MAA or MPA was added to ~2.5 mL of TOPO-capped
CdSe/ZnS QDs in dichloromethane and rocked gently for 2 hours in the dark. ~2.5 mL
of phosphate buffered saline (PBS) (pH 7.5) was added to the solution and then the
solution was agitated vigorously, and the layers were allowed to separate. The colored
layer was removed and washed ~2-3 times by use of centrifugation and removal of the
supernatant. The resulting QDs were suspended in 1-2 mL of PBS and dialyzed against 2
L of PBS for 1 hour to remove excess MPA. The QDs were stored at room temperature
in the dark.
Deaerated samples were prepared under a nitrogen atmosphere. 500 mL of 7.4 pH
PBS and 50 µL of concentrated QD solution were placed into at AtmosBag glove bag
(Sigma) that was continuously flushed with N
2
gas; the PBS solution was bubbled with
N
2
gas for 60-90 minutes within the glove bag. The deaerated samples were placed into a
quartz cuvette flushed with N
2
and tightly capped before removal from the bag.
Measurements were done immediately.
142
Emission lifetimes were obtained by use of the time-correlated-single-photon-
counting (TCSPC) technique. 800 nm laser pulses out of a Coherent Rega 9050
Ti:sapphire regenerative amplifier operating at 250 kHz repetition rate was frequency
doubled in a BBO (β-BaB
2
O
4
) crystal to produce 400 nm pulses for excitation of the
samples. The temporal pulse width of the 400 nm pulses were approximately 100
femtoseconds at full width half maximum (FWHM) and approximately 3 nm in
frequency bandwidth at FWHM, assuming a Gaussian beam profile. The beam was
focused into the sample with the largest focal spot diameter of 0.785 mm. The excitation
power was ~ 2.4 mW, with peak pulse intensities at the sample were 1 x 10
7
W/cm
2
with
1 x10
-6
J/cm
2
fluence after attenuation of the 400 nm excitation beam with neutral density
filters placed before the focusing lens. The luminescence was collected with a 1 inch
diameter 3.5 cm focal length lens placed perpendicular to the incident 400 nm beam and
focused into a monochromator with a one inch diameter 10 cm focal length lens. The
monochromator had an f number of 3.9. The monochromator was a CVI CMSP112
double spectrograph with a 1/8 m total path length in negative dispersive mode with 600
groove/mm grating. Typically the slit widths were 0.6-1.2 mm, and based on a
monochromator dispersion of 16 nm/mm, provided 10-20 nm resolution. The PMT was a
Hamamatsu RU3809 micro-channel plate detector powered by a variable high voltage
power supply at -3.0 kV and was mounted on the monochromator box exit slit.
A Becker and Hickl SPC-630 photon counting board recorded the amplified
emission signals. The reference signal was provided by a portion of the excitation beam
sent to a fast photodiode.
143
5.4 Results

Figures 5.4.1a and 5.4.1b show normalized TCSPC lifetime measurements for 84
nM CdSe/ZnS quantum dots (fluorescent at 595 nm) in aqueous solution with and
without 1 mM BME. The excitation was 2.5 mW of 400 nm. Detection was at 595 nm
and magic angle. 0.6 mm slits before the monoch77romator was used and counts were
collected within a 200 ns window. Because we have already established that there is an
effect of illumination exposure time on fluorescence lifetimes,
23
we compared 1 minute
collection time scans (Figure 5.4.1a) to get the minimal effect of illumination time
possible while collecting photons. Although this does not yield the most precise lifetime
determination because of the short data averaging, qualitative trends can still be observed.
The CdSe/ZnS solution clearly has a faster lifetime with BME present than without. Also
shown to compare the lifetimes after 20 minutes of laser illumination is a sum of 20 one
minute scans in Figure 5.4.1b. The BME also has a quenching effect after this amount of
illumination time.
Results for CdSe quantum dots without a passivating ZnS shell (fluorescent at 540
nm) are presented in Figures 5.4.2a and 5.4.2b. Figure 5.4.2a is a 1 minute scan and
Figure 5.4.2b is the sum of twenty 1 minute scans. Because the quantum yield was low, a
1.2 mm slit was used before the monochromator decreasing the spectral resolution.
Excitation was at 400 nm and 2.4 mW. Detection was at magic angle and 540 nm. The
concentration of the bare core CdSe QD solution was approximately the same as for the
CdSe/ZnS QDs and BME was at 1 mM. The 1 minute scan does not show a
144
markedly different luminescence lifetime because it had a low fluorescence yield and
therefore poor signal to noise, however, the 20 minute scan shows an enhancement in
the lifetime of bare CdSe with BME compared to without BME. The results of 50 nM
aqueous CdTe (fluorescent at 610 nm) are presented in Figures 5.4.2c and d. As with
CdSe/ZnS, the 1 mM BME added to the solution had a quenching effect. Excitation was
also with 400 nm at 2.4 mW and detection at 610 nm and magic angle. 0.6 mm slits were
used before the monochromator.

a b
0 50 100 150
1E-4
1E-3
0.01
0.1
1

Normalized Counts
Time, ns

0 50 100 150
1E-4
1E-3
0.01
0.1
1

Time, ns

Figure 5.4.1. a) 1 minute TCSPC data collection of 84 nM of CdSe/ZnS with (red trace) and without
(black trace) BME. Excitation was 2.45 mW of 400 nm and detection at magic angle with a 200 ns
time window. 0.6 mm slits at the monochromator entrance were used. b) Sum of 20 1 minute scans
of the same samples and under the same conditions as (a). The small spike at ~ 75 and 180 ns delay
time is an artifact of the TCSPC instrument and not due to the QD photodynamics.

145
a b
0 50 100 150
1E-3
0.01
0.1
1

Normalized Counts
Time, ns

0 50 100 150
1E-3
0.01
0.1
1

Time, ns

c d
0 50 100 150
0.01
0.1
1

Normalized Counts
Time, ns

0 50 100 150
1E-3
0.01
0.1
1

Time, ns
Figure 5.4.2. a,b) Normalized TCSPC scan of CdSe Core with (red trace) and without (black trace)
BME. In this experiment, 30 second scans were obtained. Shown is the sum of the first two scans (a)
to compare 1 minute collection times as with the other samples. Because the QY is small, there are
poor statistics in the initial scans and it is impossible to determine a difference in lifetimes between
QD only and QD ΒME, however, the 20 minute scan (b) shows an enhancement. c) CdTe normalized
1 minute scan. The solution without ΒME (black) has a luminescence lifetime near single
exponential, however, adding ΒME (red) makes the decay more multi-exponential. d) CdTe sum of
20 scans, normalized.

146
5.5 Discussion

5.5.1 CdSe/ZnS and CdSe

The CdSe/ZnS core-shell QD used in our experiments is assumed to have 2 – 3
monolayers of ZnS shell, since the synthesis for the core-shell quantum dots used in these
experiments followed general practices.
25
It is known that at 2 – 3 monolayers, the CdSe
core is not completely covered by the shell, leaving exposed Cd and Se.
Mercaptopropionic acid (MPA) is attached either to exposed Cd or to the ZnS surface as
a solubilizing agent. While the shell increases emission quantum yield, MPA (a thiol)
acts to increase the radiationless decay and accordingly reduces the emission QY of CdSe
QDs. This quenching has been attributed to the ability of thiols to act as hole traps for
these QDs.
24, 28

As β-mercaptoethanol (BME) is also a thiol and a stronger reducing agent than
MPA, it can be hypothesized that BME will quench our CdSe/ZnS QDS. Yet the
question remains, what is the mechanism of quenching? Pong et. al.
28
found that methyl
thiols (CH
3
-SH) do not bind to the ZnS surface of CdSe/ZnS QDs. Instead, the thiol must
first lose its thiolic proton so that the bound state is in a thiolate form. Since bond
strengths for Zn
shell
– S
thiol
are 46.5 kcal/mol, and the S
shell
– S
thiol
is 25.1 kcal/mol, they
concluded that the formation energy for the Zn – BME or S – BME with a bound thiol
state is too low for thiols. It was also shown by Jeong et. al.
24
that the thiolate and not the
thiol is the species responsible for the effect on luminescence, yet it does not only
147
have a quenching effect, but can also increase emission QY at lower pH. A turning point
exists in which a small concentration of thiolate will enhance CdSe/ZnS luminescence,
while a large amount of thiolate will act as a quencher.
28
It has also been proposed in the
literature that there are specific sites on the QD surface in which the ΒME in its thiolate
form will act to passivate and specific sites in which it will act to quench.
24
Passivation
blocks the hole or electron trap sites and therefore removes its fast deactivation pathway.
At the pH level in our system, neutral to slightly basic, we have a mixture of both the
thiol and thiolate form of BME.
To aid in establishing a mechanism, we studied CdSe QDs without a ZnS
passivating shell. CdSe bare core QDs have a greater number of exposed Cd and Se
dangling bonds. Contrary to CdSe/ZnS QDs, ΒME has an enhancing effect on CdSe core
QDs (Figures 5.4.2a and b) and thus we observe the dual nature of ΒME as both an
enhancer and quencher. Based on these experimental results, it is hypothesized that there
are correlations between the number and type of surface defects and the effect of ΒME:
decreased lifetime with increased surface Se, and increased lifetime with increased
surface Cd. If this were true, because the CdSe/ZnS QD is already extensively passivated
by the ZnS shell and the amount of exposed Cd is limited, additional thiolate will act as a
quencher by hole transfer from the Se site. Without a ZnS shell, the bare CdSe core has
more open Cd sites in which the ΒME would attach to and enhance. The known
enhancement effect by thiolic Cd passivation is discussed in more detail below. One can
speculate that there will be an optimal concentration of the thiolate BME for maximum
148
enhancement and as the concentration of thiolate BME is further increased, luminescence
enhancement will begin to decrease.

5.5.2 CdTe

To further support that the quenching behavior of ΒME in our system is due to hole
transfer from an Se trap to ΒME, we looked at photoluminescence lifetimes of MPA
solubilized CdTe QDs under the same conditions. As with CdSe/ZnS, the lifetime of
CdTe fluorescence is quenched (Figure 5.4.3a and b), even though CdTe is normally
passivated by thiols due to the removal of the surface Cd trap by the binding of thiol.
24, 29

Studies have shown that while a wide variety of thiolates passivate CdTe, there are also
thiolates that are capable of quenching.
29, 30
For this to occur, the thiolate must have a
higher redox potential than the CdTe valence band edge and thus hole transfer from the
photoexcited CdTe to the thiolate is made possible. We know that CdTe has a valence
band top that lies at a higher redox potential than CdSe.
23
BME also has a higher redox
potential than the valence band top of CdTe, -0.28 V at pH = 7 for BME
31
and ~ -0.9 V
for CdTe.
23
This supports the model that ΒME quenches the QD luminescence via hole
scavenging and is accepting a hole from surface Te in CdTe QDs, and also from surface
Se in the CdSe/ZnS system. According to the energy level diagram in Figure 5.5.1, both
the hole in CdTe and CdSe can transfer to the BME.

149
BME BME

Figure 5.5.1 Relative energy levels of valence bands and conduction band of CdSe, CdSe/ZnS,
and CdTe semiconductors in aqueous solution and β-mercaptoethanol (BME).
Figure adapted from Reference 23. Redox potentials for semiconductors are taken
from published values obtained by cyclic voltammetry for nanoparticles comparable
in size to the ones we used in this study.
32-34
Redox potential of BME taken from
Reference 31 and is for pH = 7.

5.5.3 Mechanism

To summarize the mechanism of CdSe/ZnS luminescence quenching by BME,
there are two competing pathways, one leading to quenching, and the other to
enhancement. BME passivates Cd dangling bonds and reduces electron traps to enhance
the fluorescence QY. Quenching occurs by charge transfer of the hole trapped at a
surface Se or Te. In the CdSe/ZnS QD, BME is a stronger hole acceptor than MPA and
quenches luminescence to a greater extent than MPA. Hole trapping exceeds
150
passivation. Without a passivating ZnS shell, there is more opportunity for Cd binding
and the net effect is enhancement for the CdSe core QDs. In CdTe QDs, BME quenches
luminescence because its redox potential lies above the top of the CdTe valence band as
opposed to MPA, and hole trapping exceeds passivation. Figure 5.5.2 illustrates the two
mechanisms.

S
O
R
S
O
R
Zn
S Cd
S S
OH
S
Cd
Se
h
+
e
-
S
O
R
S
O
R
S
O
R
S
O
R
Zn
S Cd
S S
OH
S
S S
OH
S
Cd
Se
h
+
e
-
e
-
S
O
R
S
O
R

Figure 5.5.2 Schematic of CdSe/ZnS QD solubilized with MPA. There is a BME attached to an
exposed Cd. A hole trapped at a nearby exposed Se site transfers to the BME. In
this figure, there is also an electron trapped at an exposed Cd site. If MPA or BME
were bonded to this Cd, the electron would not be able to trap here. BME acts to
passivate the Cd dangling bond but may also accept a hole. These are the two
competing pathways leading to enhancement or quenching of luminescence.

The quenching effect of the thiolate as opposed to thiol could be a result of the
ability of thiolate to bind to the surface. There have been several reports of distance
151
dependence of hole transfer between CdSe and CdSe/ZnS QDs and thiolates. Zhang et.
al. looked at CdSe/ZnS and CdSe/CdS/CdZnS/ZnS QDs hole transfer with hole accepting
materials.
9
Not only have they found that hole transfer is from the valence band of the
QD to the HOMO of the hole acceptor and varies in relation to the HOMO level, but also
that shell thickness influences hole transfer rates due to the distance dependence.
Dorokhin et. al. found that hole transfer rates between CdSe/ZnS QDs and ferrocenyl
thiols are dependent on the length of a thiol linker to the QD.
35
A longer linker decreased
the hole transfer rate. In our QD systems, binding would result in the thiolate being in
closer proximity to the Te or Se hole trap.

5.6 Conclusions

We have shown that BME quenches luminescence lifetimes of CdSe/ZnS – MPA
QDs under the experimental conditions favorable to the quantum dot application of our
collaboration, namely labeling of live cells. In addition, BME quenches luminescence of
CdTe, but enhances CdSe core only QDs. The results of bare CdSe QDs and CdTe QDs
experiments, along with observables reported in the literature support a mechanism that
includes two competing pathways leading to enhancement and quenching. We postulate
that the interaction of available hole and trapping sites with BME determine whether
there is a net enhancement or quenching effect on CdSe core-shell and core only QDs.
The energy of the valence band edge and redox potential of the thiolate determines if
152
hole transfer occurs from the Se or Te trapping site to the reducing agent (BME). We
know that MPA or MAA bind to the Cd surface and enhances fluorescence via
eliminating the surface Cd as an electron trapping site. Thus, we suppose that thiolate
also binds to the QD. This can have an enhancing effect, but will also decrease the
distance between the trapped hole and increase the rate of hole transfer. Because BME
does show an effect on the QD systems we study, it is important to take this into
consideration when interpreting lifetime results in experiments with oxidizing or reducing
conjugates such as bioactive dopamine.
2, 21, 36, 37

153
5.7 References for Chapter 5

1. G. D. Scholes and G. Rumbles, Nat. Mater., 2006, 5, 683-696.

2. A. P. Alivisatos, W. W. Gu and C. Larabell, Annu. Rev. Biom. Eng., 2005, 7, 55-
76.

3. D. F. Blossey, Phys. Rev. B-Solid St., 1970, 2, 3976-3990.

4. L. Brus, J. Phys. Chem., 1986, 90, 2555-2560.

5. L. E. Brus, J. Chem. Phys., 1984, 80, 4403-4409.

6. G. H. Wannier, Phys. Rev., 1937, 52, 0191-0197.

7. R. Rossetti, S. Nakahara and L. E. Brus, J. Chem. Phys., 1983, 79, 1086-1088.

8. A. P. Alivisatos, J. Phys. Chem., 1996, 100, 13226-13239.

9. Y. L. Zhang, X. G. Kong, Y. Q. Qu, P. T. Jing, Q. H. Zeng, Y. J. Sun, A. Y.
Wang, J. L. Zhao and H. Zhang, J. Luminesc., 2009, 129, 1410-1414.

10. H. M. Gong, Z. K. Zhou, H. Song, Z. H. Hao, J. B. Han, Y. Y. Zhai, S. Xiao and
Q. Q. Wang, J. Fluoresc., 2007, 17, 715-720.

11. D. F. Underwood, T. Kippeny and S. J. Rosenthal, J. Phys. Chem. B, 2001, 105,
436-443.

12. M. Jones, S. S. Lo and G. D. Scholes, J. Phys. Chem. C, 2009, 113, 18632-18642.

13. A. Puzder, A. J. Williamson, F. Gygi and G. Galli, Phys. Rev. Lett., 2004, 92, -.

14. S. W. H. Eijt, A. Van Veen, H. Schut, P. E. Mijnarends, A. B. Denison, B.
Barbiellini and A. Bansil, Nat. Mater., 2006, 5, 23-26.

15. H. Mattoussi, J. M. Mauro, E. R. Goldman, G. P. Anderson, V. C. Sundar, F. V.
Mikulec and M. G. Bawendi, J. Am. Chem. Soc., 2000, 122, 12142-12150.

16. X. G. Peng, M. C. Schlamp, A. V. Kadavanich and A. P. Alivisatos, J. Am. Chem.
Soc., 1997, 119, 7019-7029.

17. P. Reiss, M. Protiere and L. Li, Small, 2009, 5, 154-168.

154
18. R. R. Cooney, S. L. Sewall, E. A. Dias, D. M. Sagar, K. E. H. Anderson and P.
Kambhampati, Phys. Rev. B, 2007, 75, -.

19. S. A. Ivanov, A. Piryatinski, J. Nanda, S. Tretiak, K. R. Zavadil, W. O. Wallace,
D. Werder and V. I. Klimov, J. Am. Chem. Soc., 2007, 129, 11708-11719.

20. P. Yang, S. Tretiak, A. E. Masunov and S. Ivanov, J. Chem. Phys., 2008, 129, -.

21. K. E. Sapsford, T. Pons, I. L. Medintz and H. Mattoussi, Sensors, 2006, 6, 925-
953.

22. S. J. Clarke, C. A. Hollmann, Z. J. Zhang, D. Suffern, S. E. Bradforth, N. M.
Dimitrijevic, W. G. Minarik and J. L. Nadeau, 2006, 5, 409-417.

23. D. R. Cooper, D. Suffern, L. Carlini, S. J. Clarke, R. Parbhoo, S. E. Bradforth and
J. L. Nadeau, Phys. Chem. Chem. Phys., 2009, 11, 4298-4310.

24. S. Jeong, M. Achermann, J. Nanda, S. Lvanov, V. I. Klimov and J. A.
Hollingsworth, J. Am. Chem. Soc., 2005, 127, 10126-10127.

25. S. Asokan, A. R. Carreon, Z. Z. Mu, K. M. Krueger, A. Alkhawaldeh, V. L.
Colvin and M. S. Wong, P. Soc. Photo-Opt. Ins., 2005, 5705, 60-67

26. J. A. Kloepfer, R. E. Mielke, M. S. Wong, K. H. Nealson, G. Stucky and J. L.
Nadeau, Appl. Environ. Microbiol., 2003, 69, 4205-4213.

27. B. O. Dabbousi, Rodriguez-Viejo, J., Mikulec, F. V., Heine, J. R., Mattoussi, H.,
Ober, R., Jensen, K. F., Bawendi, M. G. , J. Phys. Chem. B, 1997, 101, 9463-
9475.

28. B. K. Pong, B. L. Trout and J. Y. Lee, Langmuir, 2008, 24, 5270-5276.

29. Y. S. Xia and C. Q. Zhu, Microchim. Acta, 2009, 164, 29-34.

30. S. F. Wuister, C. D. Donega and A. Meijerink, J. Phys. Chem. B, 2004, 108,
17393-17397.

31. B. Mickey and J. Howard, J. Cell Biol., 1995, 130, 909-917.

32. A. J. Nozik and R. Memming, J. Phys. Chem., 1996, 100, 13061-13078.

33. J. Jasieniak, J. Pacifico, R. Signorini, A. Chiasera, M. Ferrari, A. Martucci and P.
Mulvaney, Adv. Funct. Mater., 2007, 17, 1654-1662.

155
34. S. Moeno, M. Idowu and T. Nyokong, Inorg. Chim. Acta, 2008, 361, 2950-2956.

35. D. Dorokhin, N. Tomczak, A. H. Velders, D. N. Reinhoudt and G. J. Vancso, J.
Phys. Chem. C, 2009, 113, 18676-18680.

36. N. M. Dimitrijevic, Saponjic, Z. V., Bartels, D. M., Thurnauer, M. C., Tiede, D.
M., Rajh, T., J. Phys. Chem. B, 2003, 107, 7368-7375.

37. N. Kaji, Tokeshi, M., Baba, Y., Anal. Sci., 2007, 23.

156

Chapter 6
Conclusions and Future Work

6.1 Tribromocuprate(I) Anion

Tribromocuprate(I) anion, CuBr
3
2-
, is a molecular aqueous system that has been
shown to achieve photoinduced charge separation via complex photophysics. Aside from
deviating from the expected CTTS behavior, CuBr
3
2-
is an example of a copper
coordination complex undergoing fast intersystem crossing after being photoinduced to a
charge transfer state. In the Introduction, Section 1, it was discussed that the ability of
transition metal complexes to populate a triplet state is important in organic light-emitting
technology and potentially organic photovoltaics. CuBr
3
2-
is a much simpler system in
that the ligands are closed shell halides and the copper has a full d orbital shell. The
photoexcited electron can only transfer to a CTTS state, as opposed to undergoing MLCT
or LMCT.
More information can be gathered regarding the kinetics of the CuBr
3
2-
CTTS
state, which would be interesting because it surprisingly has been shown to produce
relatively stable excited states prior to electron transfer to the solvent. The charge-
transfer-to-solvent (CTTS) states of atomic anions have been known to live for less than
100 fs.
1-5
To the author’s knowledge, only two other systems have been reported to have
“metastable” CTTS states (thioxalato cobaltate(III) and ferrioxalate(III)) and one of them
157
has a lifetime reported as 800 fs, measured by transient absorption.
6, 7
Tribromocuprate is
another system that has been found to have a CTTS state that persists for almost 1 ps and
actually lives long enough to undergo intersystem crossing before ejecting an electron,
and possibly also has a minor prompt ejection channel. It would be valuable to get an
independent measurement of this CTTS lifetime to complement transient absorption.
Ultrafast fluorescence can provide a means for this measurement and also eliminate
ground state products and the solvated electron from the signal since only excited states
fluoresce.
In most cases, an upper limit for the lifetime of a CTTS state is obtained by
observation of product formation, typically the solvated electron which is on the order of
a few hundred fs.
1-5
The longer lifetime of the CuBr
3
2-
CTTS state makes this system
ideal for direct CTTS lifetime measurement. Also crucial is the fact that the CuBr
3
2-

CTTS state has been found to be fluorescent, making ultrafast fluorescence techniques
applicable. Fluorescence measurements of higher time resolution than TCSPC must be
implemented to resolve the CuBr
3
2-
CTTS lifetime and intersystem crossing rate from the
CTTS singlet state to the triplet. One ultrafast technique is fluorescence upconversion,
which has a time resolution that is limited only by the laser pulsewidth.
8
Another
technique with a 200 – 400 fs time resolution is Kerr-gate fluorescence.
8

Fluorescence upconversion is a nonlinear spectroscopic technique. A UV laser
pulse is used to excite the sample and the fluorescence is collected by a collimating lens
and then focused into a nonlinear crystal. A second probe pulse is focused into the
crystal, and spatially and temporally overlapped with the fluorescence. The angle of the
nonlinear optical crystal is tuned to achieve phase matching and the sum frequency (SFG)
158
of the fluorescence and probe pulse is collected and detected by a PMT. A schematic
representation of fluorescence upconversion is illustrated in Figure 6.1.

a
SFG
Fluorescence
Gate Pulse
Nonlinear optical crystal
SFG
Fluorescence
Gate Pulse
Nonlinear optical crystal

b
0
Delay
Excitation
Pulse
Gate Pulse
Time
Fluorescence
0
Delay
Excitation
Pulse
Gate Pulse
Time
Fluorescence

Figure 6.1. Fluorescence upconversion, reproduced from Reference 8. a) The fluorescence signal is
collected and directed to a nonlinear crystal where it is overlapped with a probe or “gate” pulse,
usually 800 nm in wavelength. The ultraviolet SFG signal is collected with a photon counting PMT.
b) The gate pulse is delayed and the intensity of the SFG as a function of delay time is a measure of
the sample fluorescence lifetime.

One drawback to fluorescence upconversion is that, as a nonlinear process with
one of the fields being incoherent, the conversion efficiency is low (~ 0.1% in a 1 mm
BBO crystal
8
). This is particularly a problem if the radiative rate (but not the quantum
yield) of fluorescence is small. Yet this should not an obstacle when measuring
fluorescence of CuBr
3
2-
on the ultrafast timescale, as the radiative rate as judged by the
band extinction coefficient is reasonable.
159
In cases where the signal is too small to acquire, or if the time-dependent
broadband fluorescence spectrum is desired, Kerr-gate fluorescence is another option.
Kerr-gate fluorescence resembles upconversion, except it utilizes the Kerr effect as a gate
instead of the SFG process in upconversion. The Kerr material acts as a polarizer as
birefringence is induced by the probe pulse. Following the Kerr material, a polarizer is
used to collect only the photons at the polarization generated in the gate.
9
The time
resolution depends on the recovery time of the gate material.
9
Current methods result in
200 fs – 400 fs time resolution, sufficient for the ~1 ps decay of the CTTS state with
sensitivity similar to TCSPC. One major advantage to this type of fluorescence
spectroscopy is the ability to collect the total emission spectrum by dispersing the light
onto a multichannel detector, however, this is not necessary to determine the CTTS
lifetime of CuBr
3
2-
. The major drawback to the Kerr method is the usual wavelength
range possible for detection to date, 400-675 nm,
8
which is dependent on the properties of
the gating material. This only allows for detection in the range of CuBr
3
2-
emission that
contains both fluorescence and phosphorescence. This may produce a strong background
signal if the phosphorescent background is longer lived than the inverse repetition rate of
the laser amplifier. This last issue is probably not an obstacle for this specific system
because the time between laser pulses at 250 kHz repetition rate is 4 µs (our fastest
repetition rate laser system), and the longest phosphorescence decay time reported for
CuBr
3
2-
is ~ 800 ns.
10

160
6.2 CdSe/ZnS, CdSe, and CdTe Quantum Dots

It is known that CdSe/ZnS, CdSe, and CdTe quantum dots (QDs) can transfer
charge across the QD interface to its surrounding molecules. This feature is utilized for
many applications. In Chapter 5, we have shown that the QD systems we study interact
with β-mercaptoethanol (BME) in aqueous solution. In agreement with our data and
other reports, we have proposed a model that includes hole transfer from the CdSe/ZnS,
CdSe, and CdTe QDs to BME. To further confirm our scheme, it would be beneficial to
rule out the electron as a main contributor to the observed changes in the QD – BME
systems. Also, since it is speculated that the photogenerated hole first relaxes to the
valence band edge, then traps at a surface Se dangling bond, it would be interesting to
determine if the photogenerated hole is indeed first trapping on the surface and then
secondly transferring to BME, or if the hole transfer occurs before it reaches the band
edge.
Ultrafast pump-broadband probe spectroscopy has been utilized to examine
relaxation dynamics of QDs.
11
Preliminary experiments were done in our laboratory on
~30 nM CdSe/ZnS QDs in the presence of dopamine (See Figure 6.2). In this system, it
has been shown that dopamine conjugated to the QD transfers an electron to the photo-
generated QD hole (h
+
).
12
In these experiments it was not clear whether or not dopamine
was conjugated to the QD due to the age of the QDs available for the experiments, yet it
is certain that we are capable of observing the QD bleach and complex spectral dynamics
in the first few hundred femtoseconds following excitation. If the electron is responsible
for the changes we observe upon addition of BME to our QD systems, we should be able
161
to resolve a difference in transient broadband signal between CdSe/ZnS QDs with and
without β-mercaptoethanol (BME).

a b

350 400 450 500 550 600 650
-3
-2
-1
0
1
2
3

mOD
Wavelength, nm
119 fs
519 fs
2744 fs
3744 fs
CdSe/ZnS UV/vis

395 400 405 410 415
-60
-50
-40
-30
-20
-10

mOD
Wavelength, nm
119 fs
519 fs
2744 fs
3744 fs

Figure 6.2. 400 nm excited, broadband probe of ~ 30 nM aqueous CdSe/ZnS QDs with dopamine.
The colored traces are time delays. a) The black trace is the UV/vis of the ground state CdSe/ZnS
solution. By ~ 600 fs (green trace), the transient absorption has a negative signal that mirrors that of
the QD absorption spectrum indicating ground state QD bleach. The sharp negative signal ~ 400 nm
is also a bleach due to hole burning by the 400 nm excitation. b) Expanded portion of transient
spectrum to show the bleach recovery at ~ 400 nm (pump wavelength).

However, it is not possible to resolve hole dynamics by use of our pump-probe
spectroscopy technique because transient absorption is dominated by electron
populations, and the hole, as a spectator, has little effect.
13
This is due to the high
spectral density of the valence band states that results in a considerable extent of hole
populations over multiple levels.
13
Therefore, a negative result based on pump-probe
absorption experiments does not necessarily shed light on the hole dynamics.
Alternatively, the femtosecond dynamics of the hole can be observed with
ultrafast photoluminescence spectroscopy, such as fluorescence upconversion described
162
above in Section 6.1.
14, 15
Measurement of the fluorescence intensity as a function of
delay time would yield a growing in as the hole moves closer to the valence band edge
(cooling). A decay component of the upconversion signal is due to hole trapping, which
usually is on a slower time scale than the ~ 300 fs cooling.
15
Differences in this early rise
and decay lifetimes will give information on the hole transfer dynamics between the QDs
and BME. If a larger contribution of a decay component during the dominated rise in
signal from h
+
cooling (first few hundred fs) is necessary to fit the fluorescence data, then
there must be hole transfer occurring from the photoexcited QD to BME before trapping
on the QD surface.

163
6.3 References for Chapter 6

1. X. Y. Chen and S. E. Bradforth, Annual Review of Physical Chemistry, 2008, 59,
203-231.

2. F. H. Long, H. Lu, X. L. Shi and K. B. Eisenthal, Chemical Physics Letters, 1990,
169, 165-171.

3. M. K. Fischer, A. Laubereau and H. Iglev, Physical Chemistry Chemical Physics,
2009, 11, 10939-10944.

4. A. Kammrath, J. R. R. Verlet, A. E. Bragg, G. B. Griffin and D. M. Neumark,
Journal of Physical Chemistry A, 2005, 109, 11475-11483.

5. J. A. Kloepfer, V. H. Vilchiz, V. A. Lenchenkov, A. C. Germaine and S. E.
Bradforth, Journal of Chemical Physics, 2000, 113, 6288-6307.

6. J. Chen, H. Zhang, I. V. Tomov, X. L. Ding and P. M. Rentzepis, Proceedings of
the National Academy of Sciences of the United States of America, 2008, 105,
15235-15240.

7. J. Chen, H. Zhang, I. V. Tomov and P. M. Rentzepis, Inorganic Chemistry, 2008,
47, 2024-2032.

8. J. H. Xu and J. R. Knutson, Fluorescence Spectroscopy, 2008, 450, 159-183.

9. S. Arzhantsev and M. Maroncelli, Applied Spectroscopy, 2005, 59, 206-220.

10. K. L. Stevenson, D. W. Knorr and A. Horvath, Inorganic Chemistry, 1996, 35,
835-839.

11. R. R. Cooney, S. L. Sewall, E. A. Dias, D. M. Sagar, K. E. H. Anderson and P.
Kambhampati, Physical Review B, 2007, 75, -.

12. S. J. Clarke, C. A. Hollmann, Z. J. Zhang, D. Suffern, S. E. Bradforth, N. M.
Dimitrijevic, W. G. Minarik and J. L. Nadeau, Nature Materials, 2006, 5, 409-
417.

13. S. Xu, A. A. Mikhailovsky, J. A. Hollingsworth and V. I. Klimov, Physical
Review B, 2002, 65, -.

14. D. F. Underwood, T. Kippeny and S. J. Rosenthal, Journal of Physical Chemistry
B, 2001, 105, 436-443.
164
15. E. Hendry, M. Koeberg, F. Wang, H. Zhang, C. D. Donega, D. Vanmaekelbergh
and M. Bonn, Physical Review Letters, 2006, 96, -.

165

Bibliography

1. D. M. Adams, Metal-Ligand and Related Vibrations, London, 1967.

2. M. L. Alexander, N. E. Levinger, M. A. Johnson, D. Ray and W. C. Lineberger, J.
Chem. Phys., 1988, 88, 6200-6210.

3. P. Alivisatos, J. Phys. Chem., 1996, 100, 13226-13239.

4. P. Alivisatos, W. W. Gu and C. Larabell, Annu. Rev. Biom. Eng., 2005, 7, 55-76.

5. S. Alnaser, B. Ulrich, X. M. Tong, I. V. Litvinyuk, C. M. Maharjan, P. Ranitovic,
T. Osipov, R. Ali, S. Ghimire, Z. Chang, C. D. Lin and C. L. Cocke, Phys. Rev. A,
2005, 72.

6. S. Andersson, M. Hakansson and S. Jagner, Acta Chem. Scand. A, 1989, 19, 147-
157.

7. S. Andersson, S. Jagner, Acta Chem. Scand. A, 1986, 40, 177-181.

8. S. Andersson, S. Jagner, Acta Chem. Scand. A, 1986, 40, 210-217.

9. S. Andersson, S. Jagner, Acta Chem. Scand. A, 1986, 40, 52-57.

10. S. Andersson, S. Jagner, Acta Chem. Scand. A, 1987, 41, 230-236.

11. S. Andersson, S. Jagner, J. Cryst. Spectrosc., 1988, 18, 591-600.

12. S. Andersson, S. Jagner, Acta Chem. Scand., 1989, 43, 39-43.

13. N. A. Anderson, T. Q. Lian, Annu. Rev. Phys. Chem., 2005, 56, 491-519.

14. N. Armaroli, Chem. Soc. Rev., 2001, 30, 113-124.

15. S. Asokan, A. R. Carreon, Z. Z. Mu, K. M. Krueger, A. Alkhawaldeh, V. L.
Colvin and M. S. Wong, P. Soc. Photo-Opt. Ins., 2005, 5705, 60-67.

16. L. V. Asryan, J. Nanophotonics, 2009, 3.

166
17. V. Balzani, A. Credi, F. Marchioni and J. F. Stoddart, Chem. Commun., 2001,
1860-1861.

18. U. Banin and S. Ruhman, J. Chem. Phys., 1993, 98, 4391-4403.

19. Barbiellini and A. Bansil, Nat. Mater., 2006, 5, 23-26.

20. E. R. Barthel, I. B. Martini and B. J. Schwartz, J. Phys. Chem. B, 2001, 105,
12230-12241.

21. J. C. Barnes and D. N. Hume, Inorg. Chem., 1963, 2, 444.

22. C. Benniston and A. Harriman, Coord. Chem. Rev., 2008, 252, 2528-2539.

23. Blandamer,M. J., Fox, Chem. Rev., 1970, 70, 59.

24. D. F. Blossey, Phys. Rev. B-Sol. State, 1970, 2, 3976-3990.

25. J. Blumberger, L. Bernasconi, I. Tavernelli, R. Vuilleumier and M. Sprik, J. Am.
Chem. Soc., 2004, 126, 3928-3938.

26. M. Born and R. Oppenheimer, Ann. Phys-Berlin, 1927, 84, 0457-0484.

27. G. A. Bowmaker, Brocklis.Ld and R. Whiting, Aust. J. Chem., 1973, 26, 29-42.

28. G. A. Bowmaker, G. R. Clark, D. A. Rogers, A. Camus and N. Marsich, J. Chem.
Soc. Dalton, 1984, 37-45.

29. S. E. Bradforth, Ph.D. Thesis, University of California, Berkeley, 1992.

30. S. E. Bradforth and P. Jungwirth, J. Phys. Chem. A, 2002, 106, 1286-1298.

31. T. F. Braish, R. E. Duncan, J. J. Harber, R. L. Steffen and K. L. Stevenson, Inorg.
Chem., 1984, 23, 4072-4075.

32. P. S. Braterman, Inorg. Chem., 1963, 2, 448.

33. J. Brugger, B. Etschmann, W. Liu, D. Testemale, J. L. Hazemann, H. Emerich,
W. van Beek and O. Proux, Geochim. Cosmochim. Ac., 2007, 71, 4920-4941.

34. L. Brus, J. Phys. Chem., 1986, 90, 2555-2560.

35. L. E. Brus, J. Chem. Phys., 1984, 80, 4403-4409.

36. J. K. Burdett and O. Eisenstein, Inorg. Chem., 1992, 31, 1758-1762.
167

37. M. C. Cavanagh, R. M. Young and B. J. Schwartz, J. Chem. Phys., 2008, 129.

38. B. Cercek, M. Ebert, A. J. Swallow and J. P. Keene, Science, 1964, 145, 919.

39. Y. J. Chang and T. J. Chow, Tetrahedron, 2009, 65, 4726-4734.

40. X. Y. Chen and S. E. Bradforth, Annu. Rev. Phys. Chem., 2008, 59, 203-231.

41. L. X. Chen, G. Jennings, T. Liu, D. J. Gosztola, J. P. Hessler, D. V. Scaltrito and
G. J. Meyer, J. Am. Chem. Soc., 2002, 124, 10861-10867.

42. L. X. Chen, G. B. Shaw, I. Novozhilova, T. Liu, G. Jennings, K. Attenkofer, G. J.
Meyer and P. Coppens, J. Am. Chem. Soc., 2003, 125, 7022-7034.

43. J. Chen, H. Zhang, I. V. Tomov, X. L. Ding and P. M. Rentzepis, P Natl. Acad.
Sci. USA, 2008, 105, 15235-15240.

44. J. Chen, H. Zhang, I. V. Tomov and P. M. Rentzepis, P. Natl. Acad. Sci. USA,
2008, 47, 2024-2032.

45. Y. J. Cheng, S. H. Yang and C. S. Hsu, Chem. Rev., 2009, 109, 5868-5923.

46. P. T. Chou and Y. Chi, Eur. J. Inorg. Chem., 2006, 3319-3332.

47. S. J. Clarke, C. A. Hollmann, Z. J. Zhang, D. Suffern, S. E. Bradforth, N. M.
Dimitrijevic, W. G. Minarik and J. L. Nadeau, Nat. Mater., 2006, 5, 409-417.

48. C. Consani, M. Premont-Schwarz, A. ElNahhas, C. Bressler, F. van Mourik, A.
Cannizzo and M. Chergui, Angew. Chem. Int. Edit., 2009, 48, 7184-7187.

49. R. R. Cooney, S. L. Sewall, E. A. Dias, D. M. Sagar, K. E. H. Anderson and P.
Kambhampati, Phys. Rev. B, 2007, 75.

50. D. R. Cooper, D. Suffern, L. Carlini, S. J. Clarke, R. Parbhoo, S. E. Bradforth and
J. L. Nadeau, Phys. Chem. Chem. Phys., 2009, 11, 4298-4310.

51. R. A. Crowell, R. Lian, I. A. Shkrob, D. M. Bartels, X. Y. Chen and S. E.
Bradforth, J. Chem. Phys., 2004, 120, 11712-11725.

52. D. G. Cuttell, S. M. Kuang, P. E. Fanwick, D. R. McMillin and R. A. Walton, J.
Am. Chem. Soc., 2002, 124, 6-7.

53. B. O. Dabbousi, Rodriguez-Viejo, J., Mikulec, F. V., Heine, J. R., Mattoussi, H.,
Ober, R., Jensen, K. F., Bawendi, M. G., J. Phys. Chem. B, 1997, 101, 9463-9475.
168

54. M. Dangelantonio, M. Venturi and Q. G. Mulazzani, Rad. Phys. Chem., 1988, 32,
319-324.

55. C. R. Davis and K. L. Stevenson, Inorg. Chem., 1982, 21, 2514-2516.

56. D. D. Davis, Stevenson, K. L., Davis, C. R., J. Am. Chem. Soc., 1978, 100, 6.

57. C. W. Dekock and D. M. Gruen, J. Chem. Phys., 1966, 44, 4387.

58. N. M. Dimitrijevic, O. G. Poluektov, Z. V. Saponjic and T. Rajh, J. Phys. Chem.
B, 2006, 110, 25392-25398.

59. N. M. Dimitrijevic, Saponjic, Z. V., Bartels, D. M., Thurnauer, M. C., Tiede, D.
M., Rajh, T., J. Phys. Chem. B, 2003, 107, 7368-7375.

60. D. Dorokhin, N. Tomczak, A. H. Velders, D. N. Reinhoudt and G. J. Vancso, J.
Phys. Chem. C, 2009, 113, 18676-18680.

61. L. Li, L. L. Duan, Y. H. Xu, M. Gorlov, A. Hagfeldt and L. C. Sun, Chem.
Comm., 2010, 46, 7307-7309.

62. S. W. H. Eijt, A. Van Veen, H. Schut, P. E. Mijnarends, A. B. Denison, B.
Barbiellini and A. Bansil, Nat. Mater., 2006, 5, 23-26.

63. C. G. Elles, A. E. Jailaubekov, R. A. Crowell and S. E. Bradforth, J. Chem. Phys.,
2006, 125.

64. M. K. Fischer, A. Laubereau and H. Iglev, Phys. Chem. Chem. Phys., 2009, 11,
10939-10944.

65. P. D. Fleischauer, Ph.D. Thesis, University of Southern California, 1968.

66. J. Florian and A. Warshel, J. Phys. Chem. B, 1997, 101, 5583-5595.

67. M. F. Fox, Farad. Trans. I, 1977, 73, 872-882.

68. J. Friedrich, S. Gilb, O. T. Ehrler, A. Behrendt and M. M. Kappes, J. Chem.
Phys., 2002, 117, 2635-2644.

69. W. F. Fu, X. Gan, J. Jiao, Y. Chen, M. Yuan, S. M. Chi, M. M. Yu and S. X.
Xiong, Inorg. Chim. Acta, 2007, 360, 2758-2766.

70. E. M. Glebov, V. F. Plyusnin, A. B. Venediktov and S. V. Korenev, Russ. Chem.
B+, 2003, 52, 1305-1311.
169

71. W. J. Glover, R. E. Larsen and B. J. Schwartz, J. Chem. Phys., 2010, 132.

72. H. M. Gong, Z. K. Zhou, H. Song, Z. H. Hao, J. B. Han, Y. Y. Zhai, S. Xiao and
Q. Q. Wang, J. Fluoresc., 2007, 17, 715-720.

73. B. J. Greenblatt, M. T. Zanni and D. M. Neumark, Faraday Discuss., 1997, 101-
113.

74. J. Grodkowski and P. Neta, J. Phys. Chem. A, 2002, 106, 11130-11134.

75. L. I. Grossweiner and M. S. Matheson, J. Phys. Chem., 1957, 61, 1089-1095.

76. T. Gunaratne, M. A. J. Rodgers, D. Felder, J. F. Nierengarten, G. Accorsi and N.
Armaroli, Chem. Comm., 2003, 3010-3011.

77. M. Guhr, M. Bargheer, M. Fushitani, T. Kiljunen and N. Schwentner, Phys.
Chem. Chem. Phys., 2007, 9, 779-801.

78. D. Gust, T. A. Moore and A. L. Moore, Acc. Chem. Res., 2009, 42, 1890-1898.

79. C. Harris and P. V. Kamat, ACS Nano, 2009, 3, 682-690.

80. G. J. Hedley, A. Ruseckas and I. D. W. Samuel, Chem. Phys. Lett., 2008, 450,
292-296.

81. R. H. Holm, P. Kennepohl and E. I. Solomon, Chem. Rev., 1996, 96, 2239-2314.

82. R. Holzwarth, Data Analysis in Time-resolved measurements, Kluwer, Dordrecht,
1996.

83. O. Horvath and J. H. Fendler, J. Photoch. Photobio. A, 1993, 71, 33-37.

84. O. Horvath, J. H. Fendler and K. L. Stevenson, Inorg. Chem., 1993, 32, 227-230.

85. Horvath, O. Horvath and K. L. Stevenson, J. Photoch. Photobio. A. Chem., 1992,
68, 155-163.

86. O. Horvath, Stevenson, K., Charge Transfer Photochemistry of Coordination
Compounds, VCH Publishers, New York, 1993.

87. Horvath and K. L. Stevenson, Coord. Chem. Rev., 1996, 153, 57-82.

88. Horvath and K. L. Stevenson, Coord. Chem. Rev., 2000, 208, 139-151.

170
89. O. Horvath and K. L. Stevenson, Inorg. Chem., 1989, 28, 2548-2551.

90. Horvath and K. L. Stevenson, Inorg. Chem., 1993, 32, 2225-2227.

91. Horvath, C. E. Wood and K. L. Stevenson, J. Phys. Chem., 1994, 98, 6490-6495.

92. R. A. Howald and D. P. Keeton, Spectrochim. Acta., 1966, 22, 1211-&.

93. S. Impellizzeri, S. Monaco, I. Yildiz, M. Amelia, A. Credi and F. M. Raymo, J.
Phys. Chem. C, 2010, 114, 7007-7013.

94. S. A. Ivanov, A. Piryatinski, J. Nanda, S. Tretiak, K. R. Zavadil, W. O. Wallace,
D. Werder and V. I. Klimov, J. Am. Chem. Soc., 2007, 129, 11708-11719.

95. H. A. Jahn and E. Teller, P. Roy. Soc. Lond. A Mat., 1937, 161, 220-235.

96. E. Jailaubekov and S. E. Bradforth, Appl. Phys. Lett., 2005, 87.

97. J. Jasieniak, J. Pacifico, R. Signorini, A. Chiasera, M. Ferrari, A. Martucci and P.
Mulvaney, Adv. Funct. Mater., 2007, 17, 1654-1662.

98. S. Jeong, M. Achermann, J. Nanda, S. Lvanov, V. I. Klimov and J. A.
Hollingsworth, J. Am. Chem. Soc., 2005, 127, 10126-10127.

99. D. M. Jonas, S. E. Bradforth, S. A. Passino and G. R. Fleming, J. Phys. Chem.,
1995, 99, 2594-2608.

100. M. Jones, S. S. Lo and G. D. Scholes, J. Phys. Chem. C, 2009, 113, 18632-18642.

101. F. Y. Jou and G. R. Freeman, J. Phys. Chem., 1977, 81, 909-915.

102. N. Kaji, Tokeshi, M., Baba, Y., Anal. Sci., 2007, 23.

103. Kammrath, J. R. R. Verlet, A. E. Bragg, G. B. Griffin and D. M. Neumark, J.
Phys. Chem. A, 2005, 109, 11475-11483.

104. C. Kappenstein, R. P. Hugel, A. J. P. Alix and J. L. Beaudoin, J. Chim. Phys.
PCB, 1978, 75, 427-443.

105. T. W. Kee, D. H. Son, P. Kambhampati and P. F. Barbara, J. Phys. Chem. A,
2001, 105, 8434-8439.

106. J. A. Kloepfer, R. E. Mielke, M. S. Wong, K. H. Nealson, G. Stucky and J. L.
Nadeau, Appl. Environ. Microbiol., 2003, 69, 4205-4213.

171
107. J. A. Kloepfer, V. H. Vilchiz, V. A. Lenchenkov, X. Y. Chen and S. E. Bradforth,
J. Chem. Phys., 2002, 117, 766-778.

108. J. A. Kloepfer, V. H. Vilchiz, V. A. Lenchenkov, A. C. Germaine and S. E.
Bradforth, J. Chem. Phys., 2000, 113, 6288-6307.

109. J. A. Kloepfer, V. H. Vilchiz, V. A. Lenchenkov and S. E. Bradforth, Chem. Phys.
Lett., 1998, 298, 120-128.
110. S. A. Kovalenko, A. L. Dobryakov and V. Farztdinov, Phys. Rev. Lett., 2006, 96.

111. V. Lenchenkov, J. Kloepfer, V. Vilchiz and S. E. Bradforth, Chem. Phys. Lett.,
2001, 342, 277-286.

112. R. Lian, D. A. Oulianov, R. A. Crowell, I. A. Shkrob, X. Y. Chen and S. E.
Bradforth, J. Phys. Chem. A, 2006, 110, 9071-9078.

113. J. Lind, X. H. Shen, T. E. Eriksen, G. Merenyi and L. Eberson, J. Am. Chem. Soc.,
1991, 113, 4629-4633.

114. Y. Liu, A. S. Pimentel, Y. Antoku, B. J. Giles and J. R. Barker, J. Phys. Chem. A,
2002, 106, 11075-11082.

115. F. H. Long, H. Lu, X. L. Shi and K. B. Eisenthal, Chem. Phys. Lett., 1990, 169,
165-171.

116. M. Lorenz, N. Caspary, W. Foeller, J. Agreiter, A. M. Smith and V. E. Bondybey,
Mol. Phys., 1997, 91, 483-493.

117. S. K. Mak, H. L. Wong, Q. Y. Leung, W. Y. Tam, W. K. Chan and A. B. Djurisic,
J. Organomet. Chem., 2009, 694, 2770-2776.

118. J. C. Marcum and J. M. Weber, J. Chem. Phys., 2009, 131.

119. H. Mattoussi, J. M. Mauro, E. R. Goldman, G. P. Anderson, V. C. Sundar, F. V.
Mikulec and M. G. Bawendi, J. Am. Chem. Soc., 2000, 122, 12142-12150.

120. McConnell, G. H. Li and G. W. Brudvig, Chem. Biol., 2010, 17, 434-447.

121. B. Mickey and J. Howard, J. Cell Biol., 1995, 130, 909-917.

122. S. Moeno, M. Idowu and T. Nyokong, Inorg. Chim. Acta, 2008, 361, 2950-2956.

123. M. Neumark, Phys. Chem. Chem. Phys., 2005, 7, 433-442.

124. Nickson and J. Clarke, Methods, 2010, 52, 38-50.
172

125. H. Niikura, F. Legare, R. Hasbani, M. Y. Ivanov, D. M. Villeneuve and P. B.
Corkum, Nature, 2003, 421, 826-829.

126. A. J. Nozik and R. Memming, J. Phys. Chem., 1996, 100, 13061-13078.

127. X. G. Peng, M. C. Schlamp, A. V. Kadavanich and A. P. Alivisatos, J. Am. Chem.
Soc., 1997, 119, 7019-7029.

128. Persson, M. Sandstrom, A. T. Steel, M. J. Zapatero and R. Akesson, Inorg.
Chem., 1991, 30, 4075-4081.

129. G. Peters and R. L. Caldwell, Inorg. Chem., 1967, 6, 1478.

130. S. M. Pimblott and J. A. LaVerne, J. Phys. Chem. A, 1998, 102, 2967-2975.

131. B. K. Pong, B. L. Trout and J. Y. Lee, Langmuir, 2008, 24, 5270-5276.

132. A. Puzder, A. J. Williamson, F. Gygi and G. Galli, Phys. Rev. Lett., 2004, 92.

133. P. Reiss, M. Protiere and L. Li, Small, 2009, 5, 154-168.

134. Rensing, O. T. Ehrler, J. P. Yang, A. N. Unterreiner and M. M. Kappes, J. Chem.
Phys., 2009, 130.

135. R. Rossetti, S. Nakahara and L. E. Brus, J. Chem. Phys., 1983, 79, 1086-1088.

136. S. F. Ruzankin, V. F. Anufrienko, S. A. Yashnik and Z. R. Ismagilov, J. Struct.
Chem., 2006, 47, 404-412.

137. Y. Saga, Y. Shibata and H. Tamiaki, J. Photoch. Photobiol. C. Photoch. Rev.,
2010, 11, 15-24.

138. S. Saha and J. F. Stoddart, Chem. Soc. Rev., 2007, 36, 77-92.

139. K. E. Sapsford, T. Pons, I. L. Medintz and H. Mattoussi, Sensors, 2006, 6, 925-
953.

140. M. C. Sauer, I. A. Shkrob, R. Lian, R. A. Crowell, D. M. Bartels, X. Y. Chen, D.
Suffern and S. E. Bradforth, J. Phys. Chem. A, 2004, 108, 10414-10425.

141. G. D. Scholes and G. Rumbles, Nat. Mater., 2006, 5, 683-696.

173
142. G. B. Shaw, C. D. Grant, H. Shirota, E. W. Castner, G. J. Meyer and L. X. Chen,
J. Am. Chem. Soc., 2007, 129, 2147-2160.

143. Shinar and R. Shinar, J. Phys. D. Appl. Phys., 2008, 41.

144. Z. A. Siddique, Y. Yamamoto, T. Ohno and K. Nozaki, Inorg. Chem., 2003, 42,
6366-6378.

145. I. Solomon and M. D. Lowery, Science, 1993, 259, 1575-1581.

146. I. Solomon, U. M. Sundaram and T. E. Machonkin, Chem. Rev., 1996, 96, 2563-
2605.

147. L. Stevenson, P. B. Bell, R. S. Dhawale, O. Horvath and A. Horvath, Rad. Phys.
Chem., 1999, 55, 489-496.

148. L. Stevenson, P. B. Bell, O. Horvath and A. Horvath, J. Am. Chem. Soc., 1998,
120, 4234-4235.

149. K. L. Stevenson, P. B. Bell and R. E. Watson, Coord. Chem. Rev., 2002, 229,
133-146.

150. K. L. Stevenson, R. M. Berger, M. M. Grush, J. C. Stayanoff, A. Horvath and O.
Horvath, J. Photoch. Photobiol. A, 1991, 60, 215-227.

151. K. L. Stevenson, J. L. Braun, D. D. Davis, K. S. Kurtz and R. I. Sparks, Inorg.
Chem., 1988, 27, 3472-3476.

152. K. L. Stevenson, R. S. Dhawale, A. Horvath and O. Horvath, J. Phys. Chem. A,
1997, 101, 3670-3676.

153. K. L. Stevenson, M. M. Grush and K. S. Kurtz, Inorg. Chem., 1990, 29, 3150-
3153.

154. K. L. Stevenson and J. H. Jarboe, J. Photoch. Photobio. A. Chem., 2002, 150, 49-
57.

155. K. L. Stevenson, J. H. Jarboe, S. A. Langmeyer and T. W. Acra, Inorg. Chem.,
2003, 42, 3559-3564.

156. K. L. Stevenson, D. W. Knorr and A. Horvath, Inorg. Chem., 1996, 35, 835-839.

157. D. Suffern, S. J. Clarke, C. A. Hollmann, D. Bahcheli, S. E. Bradforth and J. L.
Nadeau, 2006, 6096, O960-O960.

174
158. S. Takeuchi and T. Tahara, J. Chem. Phys., 2004, 120, 4768-4776.

159. M. J. Tauber, R. A. Mathies, X. Y. Chen and S. E. Bradforth, Rev. Sci. Instru.,
2003, 74, 4958-4960.

160. C. L. Thomsen, D. Madsen, S. R. Keiding, J. Thogersen and O. Christiansen, J.
Chem. Phys., 1999, 110, 3453-3462.

161. A. Treinin and E. Hayon, J. Am. Chem. Soc., 1975, 97, 1716-1721.

162. T. Tsuboi, J. Non.-Cryst. Solids, 2010, 356, 1919-1927.

163. A. Tsuboyama, K. Kuge, M. Furugori, S. Okada, M. Hoshino and K. Ueno, Inorg.
Chem., 2007, 46, 1992-2001.

164. N. R. Tripathi, R. H. Schuler and R. W. Fessenden, Chem. Phys. Lett., 1985, 113,
563-568.

165. D. F. Underwood, T. Kippeny and S. J. Rosenthal, J. Phys. Chem. B, 2001, 105,
436-443.

166. H. M. L. van Stokkum, D. S.; van Grondelle, R., R. Biochim. Biophys. Acta -
Bioenerg., 2004, 82.

167. V. H. Vilchiz, J. A. Kloepfer, A. C. Germaine, V. A. Lenchenkov and S. E.
Bradforth, J. Phys. Chem. A, 2001, 105, 1711-1723.

168. X. F. Wang, Y. Koyama, O. Kitao, Y. Wada, S. Sasaki, H. Tamiaki and H. S.
Zhou, Biosens. Bioelectron., 2010, 25, 1970-1976.

169. X. B. Wang, L. S. Wang, R. Brown, P. Schwerdtfeger, D. Schroder and H.
Schwarz, J. Chem. Phys., 2001, 114, 7388-7395.

170. G. H. Wannier, Phys. Rev., 1937, 52, 0191-0197.

171. R. Wilbrandt, N. H. Jensen, A. H. Sillesen and K. B. Hansen, Chem. Phys. Lett.,
1984, 106, 503-507.

172. S. Wilson, A. S. Dhoot, A. J. A. B. Seeley, M. S. Khan, A. Kohler and R. H.
Friend, Nature, 2001, 413, 828-831.

173. S. F. Wuister, C. D. Donega and A. Meijerink, J. Phys. Chem. B, 2004, 108,
17393-17397.

174. Y. S. Xia and C. Q. Zhu, Microchim. Acta, 2009, 164, 29-34.
175

175. P. Yang, S. Tretiak, A. E. Masunov and S. Ivanov, J. Chem. Phys., 2008, 129, -.

176. T. Yeh, C. V. Shank and J. K. McCusker, Science, 2000, 289, 935-938.

177. T. Zanni, B. J. Greenblatt, A. V. Davis and D. M. Neumark, J. Chem. Phys., 1999,
111, 2991-3003.

178. Zehavi and J. Rabani, J. Phys. Chem., 1972, 76, 312.

179. H. Zhang, J. Chen, I. V. Tomov, A. S. Dvornikov and P. M. Rentzepis, J. Phys.
Chem. A, 2007, 111, 11584-11588.

180. Y. L. Zhang, X. G. Kong, Y. Q. Qu, P. T. Jing, Q. H. Zeng, Y. J. Sun, A. Y.
Wang, J. L. Zhao and H. Zhang, J. Luminesc., 2009, 129, 1410-1414.

181. W. L. Zou and J. E. Boggs, J. Chem. Phys., 2009, 130.

176

Appendix
Mathcad Simulation for CuBr
3
2-
Global Analysis

177

Kinetics Scheme
The kinetics are all first order. The
pump pulse prompts ligand
dissociation into the
dibromocuprate CTTS singlet and
also prepares the tribromocuprate
CTTS. The singlet tribromo CTTS
relaxes into the triplet tribromo
CTTS. The singlet dibromo CTTS
relaxes into the triplet dibromo
CTTS state.
This is the kinetic model:
Number of states: N 8 :=
i 0 N 1 − .. :=
CuBr32- evolution
f x y , ( ) 0 :=

W matrixN N , f , ( ) := 1 ----------- > 3
^
|
l
matrix function:
matrix(m,n,f)
m number of rows
n number of columns
f function of two variables
0 ----------- > 2 ------- > 4
^
|
l
Species
0: tri-CTTS
1: relaxed tri-CTTS
2: di-CTTS
3: tri-CTTS triplet
4: di-CTTS triplet
5: Br-
6: Br radical
7: Br2- radical
ground state CuBr32-
W
4 2 ,
180000 ( )
1 −
:=
W
2 0 ,
85 ( )
1 −
:=
W
3 1 ,
850 ( )
1 −
:=
W
1 0 ,
200 ( )
1 −
:=
W
6 5 ,
50 ( )
1 −
:= W
7 6 ,
150000 ( )
1 −
:=

178

W
i i ,
0
N 1 −
j
W
j i ,
−
∑
=
:=
Solving the set of differential equations for the kinetic scheme above.
W
0.017 −
5 10
3 −
×
0.012
0
0
0
0
0
0
1.176 − 10
3 −
×
0
1.176 10
3 −
×
0
0
0
0
0
0
5.556 − 10
6 −
×
0
5.556 10
6 −
×
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.02 −
0.02
0
0
0
0
0
0
0
6.667 − 10
6 −
×
6.667 10
6 −
×
0
0
0
0
0
0
0
0




























=

W0
1
0
0
0
0
1
0
0






















:=
These are our initial conditions. 0.9 weight begins in the
initially prepared singlet tribromocuprate CTTS, state labeled 0
and 0.1 weight begins in the Br- species, labeled 5.

179

These are the eigenvalues and
eigenvectors.
ξ eigenvals W ( ) := v eigenvecs W ( ) :=
ξ
0
0
1.176 − 10
3 −
×
5.556 − 10
6 −
×
0.017 −
0
6.667 − 10
6 −
×
0.02 −


























=
v
0
0
0
0
1
0
0
0
0
0
0
1
0
0
0
0
0
0.707
0
0.707 −
0
0
0
0
0
0
0.707
0
0.707 −
0
0
0
0.792
0.254 −
0.556 −
0.018
1.841 10
4 −
×
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0.707
0.707 −
0
0
0
0
0
0.707
0.707 −
2.357 10
4 −
×
























=
C v
1 −
W0 ⋅ :=
C
0.702
0.298
0.454
0.993
1.263
1
1.415
1.414






















=
p i t , ( )
0
N 1 −
j
C
j
e
ξ
j
t ⋅
v
i j ,
⋅




⋅




∑
=
:=
This is the population as a
function of time for each
species, i = 0 to 4.

180

Concentration profiles
0 500 1 10
3
× 1.5 10
3
×
0
0.2
0.4
0.6
0.8
Transient Concentration Profiles
Time, fs
Fraction Population
p 0 t , ( )
p 1 t , ( )
p 2 t , ( )
p 3 t , ( )
p 4 t , ( )
p 5 t , ( )
p 6 t , ( )
p 7 t , ( )
t

Early time: These are the concentration of each species as a function of time,
determined by solving the series of differential kinetic equations
above.

181

Mid time:
0 2 10
3
× 4 10
3
× 6 10
3
× 8 10
3
× 1 10
4
×
0
0.2
0.4
0.6
0.8
Transient Concentration Profiles
Time, fs
Fraction Population
p 0 t , ( )
p 1 t , ( )
p 2 t , ( )
p 3 t , ( )
p 4 t , ( )
p 5 t , ( )
p 6 t , ( )
p 7 t , ( )
t

182

0 1 10
5
× 2 10
5
× 3 10
5
× 4 10
5
× 5 10
5
×
0
0.2
0.4
0.6
Transient Concentration Profiles
Time, fs
Fraction Population
p 0 t , ( )
p 1 t , ( )
p 2 t , ( )
p 3 t , ( )
p 4 t , ( )
p 5 t , ( )
p 6 t , ( )
p 7 t , ( )
t

Long time:

183

Electron Spectrum including dynamics

Assumption: The dynamics of the detached electron follows the same dynamics and
has the same extinction coefficients, etc., as for ferrocyanide.
vsinf 13990 := position of the s-band peak at time infinity from Vitya mathcad sheet.
Position of the s-band peak at time infinity determined for 40 mM aqueous
solution of K
4
Fe(CN)
6
by adding 0.0096 eV (~ 77 cm
-1
) shift due to ionic
strength (interpolated from data by A. N. Asaad, N. Chandrasekhar, A. W.
Nashed, P. Krebs, J. Phys. Chem. A 103 (1999) 6339-43) to 13913 cm
-1
(center
of solvated electron absorption spectrum in H
2
O at 298 K reported by F.-Y. Jou
and G.R. Freeman):
esinf 19000 := extinction coefficient at peak of the s-band at time infinity.
tviscorr 570 := time it takes the visible (s-state) band to shift (thermalize)
vs0 9700 := position of the s-band peak at time zero
HWHM of the Gaussian side of the electron absorption spectrum s-
band at time inf - Joo and Freeman 0.355 eV H2O, 298 K
gshwhminf 2863 :=
HWHM of the Lorentzian side of the electron absorption spectrum s-
band (constant) - Joo and Freeman 0.488 eV, H2O 298 K
laus 3936 :=

184

vsinf2 13913 := The gshwhminf and laus are same for detachment Iodide and ferrocyanide.
trise 310 :=
Rise time and ws(t) from Vitya's sheet.
Geminate Recombination for CuBr
CuBrSrecomb t ( ) .702 0.22 exp
t −
1480






+ 0.073 exp
t −
10180






+ 0.2 exp
t −
76400






+ :=
ωs t ( ) 1 exp
t −
trise






−






:=
removed CuBrSrecomb(t) from this equation
(was multiplied)
vs t ( ) vs0 vsinf − ( ) exp
t −
tviscorr












vsinf + :=
shift of the peak maximum frequency with time
gsinf
gshwhminf
2 ln 2 ( ) ⋅
:=
σs t ( ) gsinf := no width variation with time of the gaussian side of spectrum

185

Shift of the entire absorpton spectrum with time
ν λ ( )
10
7
λ
:=
λ 100 110 , 1600 .. :=
ε λ t , ( ) esinf exp
ν λ ( ) vs t ( ) − ( )
2
−
2 σs t ( ) ( )
2








Φ vs t ( ) ν λ ( ) − ( ) ⋅








1 Φ vs t ( ) ν λ ( ) − ( ) − ( )
1
1
ν λ ( ) vs t ( ) − ( )
2
laus ( )
2
+










⋅ +












⋅ :=

S λ t , ( ) ωs t ( ) ε λ t , ( ) ⋅ :=
0 500 1 10
3
× 1.5 10
3
×
0
5 10
3
×
1 10
4
×
1.5 10
4
×
2 10
4
×
S λ 50 , ( )
S λ 100 , ( )
S λ 250 , ( )
S λ 500 , ( )
S λ 1000 , ( )
S λ 1500 , ( )
S λ 3000 , ( )
S λ 5000 , ( )
λ

Se λ t , ( ) S λ t , ( ) := Assumption: There is no electron recombination.

186

These parameters fit the 020206 water vis data

position of the s-band peak at time zero
h2otviscorr 500 :=
h2ovs0 11210 :=
time it takes the visible (s-state) band to shift (thermalize)
position of the s-band peak at time infinity (equilibrium)
extinction coefficient at peak of the s-band at time infinity.
h2ovsinf 14000 :=
h2oesinf 19000 :=
position of the p-band peak at time infinity (equilibrium) h2oes0 11205 − :=
h2ot ε 390 := extinction coefficient at peak of the s-band at time zero.
h2oepmax 16600 := HWHM of the gaussian side of the electron absorption spectrum at time zero
HWHM of the gaussian side of the electron absorption
spectrum s-band at time infinity
h2ovp0 7692 :=
h2ovpinf 8196 := time it takes for the extiction coefficient to achieved maximum value.
extinction coefficient at peak of the p-band. assume no change in time. h2ogsinf 2800 :=
position of the p-band peak at time zero h2ogs0 10400 :=

187

time it takes the Gaussian side of the spectrum to reach final width h2otgs 500 :=
HWHM of the Laurentsian side of the electron absorption spectrum s-band
(constant)
h2olaus 3600 :=
HWHM of the gaussian side of the electron absorption spectrum p-band at time
inf
h2ogpinf 880 :=
HWHM of the Laurentsian side of the electron absorption spectrum p- band
(constant)
h2olaup 1200 :=
Total diffusion coefficient.
determines the distance the electron was ejected originally
reaction radius
trapping or appearance time of the electron
Time it takes for the p-state electrons to relax to the s-state (2-state
model)
h2otstep 280 :=
h2otrise 150 :=
h2orxn2 2.6 :=
h2oto 12500 :=
h2oD 0.0007 :=

188

h2ovs t ( ) h2ovs0 h2ovsinf − ( ) exp
t −
h2otviscorr












h2ovsinf + :=
Spectral shift funtion for the s-band
h2ovinf
210
5.1 10
4
× =
m 0 210 .. :=
h2olinf
m
1
h2ovinf
m
10
7
⋅ :=
0 4 10
3
× 8 10
3
×
1 10
4
×
1.1 10
4
×
1.2 10
4
×
1.3 10
4
×
1.4 10
4
×
1.5 10
4
×
h2ovs t ( )
t
h2ovinf
m
9000 200 m ⋅ + :=
h2olinf
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
3
1.11110
3
1.08710
3
1.06410
3
1.04210
3
1.0210
3
110
980.392
961.538
943.396
925.926
909.091
892.857
877.193
862.069
847.458
...
=
0 4 10
3
× 8 10
3
×
0
0.2
0.4
0.6
0.8
h2ovs t ( ) h2ovsinf −
h2ovs0 h2ovsinf −
t

189

h2oσs t ( ) h2ogs0 h2ogsinf − ( )exp
t −
h2otgs






h2ogsinf + :=
HWHM variation with time of the gaussian side
Extinction coeficient of the thermalized solvated electron as a funtion of frequency
(solvated electron spectrum)
h2ol1ps
m
10
7
h2ov1ps
m
:=
h2ov1ps
m
h2ovinf
m
408 − :=
h2oε υ ( ) h2oesinf exp
υ h2ovsinf − ( )
2
−
2h2ogsinf
2
















Φ h2ovsinf υ − ( ) ⋅








1 Φ h2ovsinf υ − ( ) − ( )
1
1
υ h2ovsinf − ( )
2
h2olaus
2
+










⋅ +












⋅ :=
h2oe
m
h2oε h2ovinf
m
( )
:= in wavelength:
h2oε λ ( ) h2oesinf exp
10
7
λ
h2ovsinf −






2
−
2h2ogsinf
2
























Φ h2ovsinf
10
7
λ
−






⋅












1 Φ h2ovsinf
10
7
λ
−






−






1
1
10
7
λ
h2ovsinf −






2
h2olaus
2
+














⋅ +


















⋅ :=

190

0 500 1 10
3
× 1.5 10
3
× 2 10
3
×
0
5 10
3
×
1 10
4
×
1.5 10
4
×
h2oε λ ( )
λ
0 200 400 600 800 1 10
3
× 1.2 10
3
×
0.1
0.2
0.3
0.4
0.5
h2oε h2ovinf
m
( )
h2oε h2ovinf
m
( )
h2olinf
m
h2ol1ps
m
,

191

Extinction coeficient of the electron as a funtion of frequency and time. (spectral
shift of the electron) S-band
h2oes t ( ) h2oes0 h2oesinf − ( ) exp
t −
h2otε












h2oesinf + :=
change in time of the emax.
h2oε υ t , ( ) h2oesinf exp
υ h2ovs t ( ) − ( )
2
−
2 h2oσs t ( ) ( )
2
















Φ h2ovs t ( ) υ − ( ) ⋅








1 Φ h2ovs t ( ) υ − ( ) − ( )
1
1
υ h2ovs t ( ) − ( )
2
h2olaus ( )
2
+










⋅ +












⋅ :=
in wavelength:
h2oε λ t , ( ) h2oesinf exp
10
7
λ
h2ovs t ( ) −






2
−
2 h2oσs t ( ) ( )
2
























Φ h2ovs t ( )
10
7
λ
−






⋅












1 Φ h2ovs t ( )
10
7
λ
−






−






1
1
10
7
λ
h2ovs t ( ) −






2
h2olaus ( )
2
+














⋅ +


















⋅ :=

192

0 200 400 600 800 1 10
3
× 1.2 10
3
× 1.4 10
3
× 1.6 10
3
× 1.8 10
3
× 2 10
3
×
0
2 10
3
×
4 10
3
×
6 10
3
×
8 10
3
×
1 10
4
×
1.2 10
4
×
1.4 10
4
×
1.6 10
4
×
1.8 10
4
×
2 10
4
×
h2oε λ 0.0 , ( )
h2oε λ 100 , ( )
h2oε λ 200 , ( )
h2oε λ 300 , ( )
h2oε λ 400 , ( )
h2oε λ 550 , ( )
h2oε λ 750 , ( )
h2oε λ 1000 , ( )
h2oε λ 1500 , ( )
h2oε λ 3000 , ( )
h2oε λ 5000 , ( )
λ

193

0 5 10
3
× 1 10
4
×
0
2 10
3
×
1.5 10
3
×
1 10
3
×
500
0
h2oε 200 t , ( )
h2oε 400 t , ( )
h2oε 600 t , ( )
h2oε 800 t , ( )
h2oε 1000 t , ( )
h2oε 1200 t , ( )
h2oε 1400 t , ( )
h2oε 1600 t , ( )
t

194

Infrared (P-Band) of the solvated electron.
h2ovp t ( ) h2ovp0 h2ovpinf − ( ) exp
t −
h2otviscorr












h2ovpinf + :=
Spectral shift funtion for the p-band
Extinction coeficient of the thermalized solvated electron as a funtion of
frequency (solvated electron spectrum) p-band
h2oεp λ t , ( ) h2oepmax exp
10
7
λ
h2ovp t ( ) −






2
−
2h2ogpinf
2
























Φ h2ovp t ( )
10
7
λ
−






⋅












Φ
10
7
λ
h2ovp t ( ) −






1
1
10
7
λ
h2ovp t ( ) −






2
h2olaup
2
+














⋅ +


















⋅ :=

195

400 540 680 820 960 1.1 10
3
× 1.24 10
3
× 1.38 10
3
× 1.52 10
3
× 1.66 10
3
× 1.8 10
3
×
0
2 10
3
×
4 10
3
×
6 10
3
×
8 10
3
×
1 10
4
×
1.2 10
4
×
1.4 10
4
×
1.6 10
4
×
1.8 10
4
×
2 10
4
×
h2oεp λ 0.1 , ( )
h2oεp λ 300 , ( )
h2oεp λ 750 , ( )
h2oεp λ 1500 , ( )
h2oεp λ 3000 , ( )
λ

196

0 50 100 150 200
0
5 10
3
×
1 10
4
×
1.5 10
4
×
h2oεp 400 t , ( )
h2oεp 600 t , ( )
h2oεp 800 t , ( )
h2oεp 1000 t , ( )
h2oεp 1200 t , ( )
h2oεp 1500 t , ( )
h2oεp 2000 t , ( )
t

h2oGaussErf t ( ) 1 erf
h2orxn2
4 h2oD ⋅ h2oto ⋅






−
h2orxn2 e
h2orxn2
2
( )
−
4 h2oD ⋅ t h2oto + ( ) ⋅
⋅
3.142h2oD ⋅ t h2oto + ( ) ⋅
1 erf
h2orxn2
2
t ⋅
4 h2oD ⋅ h2oto ⋅ t h2oto + ( ) ⋅






−






⋅ + :=
h2oSrecomb t ( ) .702 0.22exp
t −
1480






+ 0.073exp
t −
10180






+ 0.2exp
t −
76400






+ :=

197

Recombination that fits 020206 water vis data
0 4 10
3
× 8 10
3
×
0
0.5
1
h2oωs t ( )
t

h2oSrecomb t ( ) .6 .5exp
t −
1480






+ 0.08exp
t −
10180






+ .25exp
t −
76400






+ :=
h2oωs t ( ) 1 exp
t −
h2otstep






−






h2oSrecomb t ( ) ⋅ :=
0 1 10
3
× 2 10
3
× 3 10
3
×
0
0.1
0.2
0.3
0.4
h2oωp t ( )
t

h2oωp t ( ) 1 exp
t −
h2otrise






−






exp
t −
h2otstep






⋅ :=
h2oωt t ( ) h2oωs t ( ) h2oωp t ( ) + :=

h2oSsband λ t , ( ) h2oωs t ( ) h2oε λ t , ( ) ⋅ :=
0 1 10
3
× 2 10
3
× 3 10
3
×
0
0.5
1
1.5
h2oωt t ( )
t

198

Other Spectra
Individual Gaussians (static)
Gaussians and sums of gaussians are assumed for each of the transient species. All extinction
coefficients are
similar orders of magnitude as expected for all similar species.
Gi... are the gaussians for the different species in the spectra.
a is the extinction coefficient
b is the peak wavelength
c is the FWHM
The triplet extinction coefficient is an assumption that it is equal or less than solvated electron,
from Stevenson and Horvath (1996).
In wavelength:
tribromo-CTTS triplet
KNOWN
dibromo-CTTS relaxed
tribromo-CTTS
tribromo-CTTS
sum of two gaussians sum of two gaussians
a33 10000 := a3 9500 := a22 1800 := a2 6500 := a1 4500 := a0 7000 :=
b3 405 := b22 290 := b2 350 := b33 290 :=
b1 530 := b0 530 :=
c0 120 :=
c3 55 := c22 20 := c2 55 := c33 50 := c1 120 :=

199

dibromo-CTTS triplet dibromo radical
KNOWN
Br radical
KNOWN
Not in our window
Br radical
KNOWN
a6 2900 :=
b6 275 :=
a5 0 :=
b5 0 :=
a4 7000 := a7 9900 :=
b7 360 :=
c7 40 := c6 30 := c5 0 :=
b4 335 :=
c4 40 :=
It is more accurate to assume spectral gaussians in the
energy domain. Here, the above gaussians in wavelengths
are converted to energy.
Converted to wavenumber:
b0ω
10
7
b0
:=
b2ω
10
7
b2
:= b1ω
10
7
b1
:=
c0ω
10
7
b0
c0
2
+
10
7
b0
c0
2
−
−










− :=
c2ω
10
7
b2
c2
2
+
10
7
b2
c2
2
−
−










− := c1ω
10
7
b1
c1
2
+
10
7
b1
c1
2
−
−










− :=
G0 ω ω ( ) a0 e
ω b0ω − ( )
2
−
2 c0ω ( )
2
⋅
⋅ :=
G1 ω ω ( ) a1 e
ω b1ω − ( )
2
−
2 c1ω ( )
2
⋅
⋅ := G2 ω ω ( ) a2 e
ω b2ω − ( )
2
−
2 c2ω ( )
2
⋅
⋅ :=

200

b22ω
10
7
b22
:= b3ω
10
7
b3
:= b6ω
10
7
b6
:=
c3ω
10
7
b3
c3
2
+
10
7
b3
c3
2
−
−










− :=
c22ω
10
7
b22
c22
2
+
10
7
b22
c22
2
−
−










− := c6ω
10
7
b6
c6
2
+
10
7
b6
c6
2
−
−










− :=
G22 ω ω ( ) a22 e
ω b22ω − ( )
2
−
2 c22ω ( )
2
⋅
⋅ := G6 ω ω ( ) a6 e
ω b6ω − ( )
2
−
2 c6ω ( )
2
⋅
⋅ :=
G3 ω ω ( ) a3 e
ω b3ω − ( )
2
−
2 c3ω ( )
2
⋅
⋅ :=
b33ω
10
7
b33
:=
G2 ω ω ( ) G2 ω ω ( ) G22 ω ω ( ) + :=
b7ω
10
7
b7
:=
b4ω
10
7
b4
:=
c7ω
10
7
b7
c7
2
+
10
7
b7
c7
2
−
−










− :=
c33ω
10
7
b1
c33
2
+
10
7
b1
c33
2
−
−










− :=
c4ω
10
7
b4
c4
2
+
10
7
b4
c4
2
−
−










− :=
G33 ω ω ( ) a3 e
ω b33ω − ( )
2
−
2 c33ω ( )
2
⋅
⋅ := G4 ω ω ( ) a4 e
ω b4ω − ( )
2
−
2 c4ω ( )
2
⋅
⋅ := G7 ω ω ( ) a7 e
ω b7ω − ( )
2
−
2 c7ω ( )
2
⋅
⋅ :=
G3 ω ω ( ) G3 ω ω ( ) G33 ω ω ( ) + :=

201

1 10
4
× 2 10
4
× 3 10
4
× 4 10
4
× 5 10
4
×
0
2 10
3
×
4 10
3
×
6 10
3
×
8 10
3
×
Transient Absorption Spectra in Energy
Wavenumber
Extinction Coefficient
G0ω ω ( )
G1ω ω ( )
G2ω ω ( )
G3ω ω ( )
G4ω ω ( )
G6ω ω ( )
G7ω ω ( )
ω

202

The spectra are converted back to wavelength in order to overlay data, which is plotted in
wavelength.
ω λ ( )
10
7
λ
:=
G0 λ ( ) a0 e
ω λ ( ) b0ω − ( )
2
−
2 c0ω
2
⋅
⋅ :=
G1 λ ( ) a1 e
ω λ ( ) b1ω − ( )
2
−
2 c1ω
2
⋅
⋅ :=
G2 λ ( ) a2 e
ω λ ( ) b2ω − ( )
2
−
2 c2ω
2
⋅
⋅ :=
G22 λ ( ) a22 e
ω λ ( ) b22ω − ( )
2
−
2 c22ω
2
⋅
⋅ :=
G2 λ ( ) G2 λ ( ) G22 λ ( ) + :=
G3 λ ( ) a3 e
ω λ ( ) b3ω − ( )
2
−
2 c3ω
2
⋅
⋅ :=
G33 λ ( ) a33 e
ω λ ( ) b33ω − ( )
2
−
2 c33ω
2
⋅
⋅ :=
G4 λ ( ) a4 e
ω λ ( ) b4ω − ( )
2
−
2 c4ω
2
⋅
⋅ :=
G3 λ ( ) G3 λ ( ) G33 λ ( ) + :=
G6 λ ( ) a6 e
ω λ ( ) b6ω − ( )
2
−
2 c6ω
2
⋅
⋅ :=
G7 λ ( ) a7 e
ω λ ( ) b7ω − ( )
2
−
2 c7ω
2
⋅
⋅ :=

203

SHTλ
309.70157
319.96545
329.99918
340.03934
350.17163
360.3027
370.29633
380.28928
390.05075
400.04224
410.21689
420.29938
430.10492
440.09274
450.22021










































:= SHTabs
0.39087
0.34577
0.30146
0.33475
0.3682
0.38732
0.4144
0.43351
0.43732
0.43891
0.42806
0.41658
0.39649
0.35393
0.32954










































:=
300 400 500 600 700
0.2
0.3
0.4
SHTabs
SHTλ

Left: SHTabs is the tribromo triplet
species as reported by Stevenson 1996.
Replotted. They did not report an
extinction coefficient, but says that it
is equal or less than that of the solvated
electron, which is ~ 19000 M
-1
cm
-1
.

204

300 400 500 600 700
0
2 10
3
×
4 10
3
×
6 10
3
×
8 10
3
×
1 10
4
×
Transient Spectra in Wavelength
Wavelength
Extinction Coefficient
SHTabs 21000 ⋅
G0 λ ( )
G1 λ ( )
G2 λ ( )
G3 λ ( )
G4 λ ( )
G6 λ ( )
G7 λ ( )
SHTλ λ ,

205

S is the equations for the spectrum with time, as a function of the concentration profile.
S0 λ t , ( ) G0 λ ( ) p 0 t , ( ) ⋅ := S4 λ t , ( ) G4 λ ( ) p 4 t , ( ) ⋅ := h2oS λ t , ( ) h2oωs t ( ) h2oε λ t , ( ) ⋅ h2oωp t ( ) h2oεp λ t , ( ) ⋅ + :=
S1 λ t , ( ) G1 λ ( ) p 1 t , ( ) ⋅ := S6 λ t , ( ) G6 λ ( ) p 6 t , ( ) ⋅ := Se λ t , ( ) ωs t ( ) ε λ t , ( ) ⋅ := Species
0: tri-CTTS
1: relaxed tri-CTTS
2: di-CTTS
3: tri-CTTS triplet
4: di-CTTS triplet
5: Br-
6: Br radical
7: Br2- radical
S2 λ t , ( ) G2 λ ( ) p 2 t , ( ) ⋅ := S7 λ t , ( ) G7 λ ( ) p 7 t , ( ) ⋅ :=
S3 λ t , ( ) G3 λ ( ) p 3 t , ( ) ⋅ :=
Data has been dispersion adjusted. Spec is the sum of all the transient spectra weighted by its
population. Electrons from 2PA Br- detachment and from 2PA water are added in. Since
spectral magnitude is in extinction coefficient, and data magnitude is mOD, Spec is reduced to fit
data. The weight of the solvated electrons is determined by the [H+] quenching fit shown below.
Spec λ t , ( ) S0 λ t , ( ) S1 λ t , ( ) + S2 λ t , ( ) + S3 λ t , ( ) + S4 λ t , ( ) + S6 λ t , ( ) .1 ⋅ + S7 λ t , ( ) .1 ⋅ + Se λ t , ( ) .1 ⋅ + h2oS λ t , ( ) .2 ⋅ + ( ) 0.0017 ⋅ :=
h2oSbr λ t , ( ) Se λ t , ( ) 0.4 ⋅ h2oS λ t , ( ) 2.1 ⋅ + S7 λ t , ( ) S6 λ t , ( ) + ( ) 1.3 ⋅ + [ ] 0.0017 ⋅ :=
Spec λ t , ( ) S0 λ t , ( ) S1 λ t , ( ) + S2 λ t , ( ) + S3 λ t , ( ) + S4 λ t , ( ) + h2oSbr λ t , ( ) 90 ⋅ ( ) + [ ] 0.0015 ⋅ :=

206

300 400 500 600 700
0
5
10
Data
Wavelength, nm
mOD
uv2000fs
uv12ps
uv52ps
uv102ps
uv202ps
uv302ps
uv402ps
uv442ps
vis1804fs
vis12ps
vis52ps
vis102ps
vis202ps
vis302ps
vis402ps
vis442ps
uvλ uvλ , uvλ , uvλ , uvλ , uvλ , uvλ , uvλ , visλ , visλ , visλ , visλ , visλ , visλ , visλ , visλ ,

207

300 400 500 600 700
0
5
10
15
Simulation
Wavelength, nm
Absorption
Spec λ 2000 , ( )
Spec λ 12000 , ( )
Spec λ 52000 , ( )
Spec λ 102000 , ( )
Spec λ 202000 , ( )
Spec λ 302000 , ( )
Spec λ 402000 , ( )
Spec λ 442000 , ( )
λ

208
Simulations have been offset to show features.

300 400 500 600
0
2
4
6
8
10
uv74fs
uv104fs
uv124fs
uv154fs
vis74fs
vis104fs
vis124fs
vis154fs
Spec λ 74 , ( ) 2 −
Spec λ 104 , ( ) 2 −
Spec λ 124 , ( ) 2 −
Spec λ 154 , ( ) 2 −
uvλ uvλ , uvλ , uvλ , visλ , visλ , visλ , visλ , λ , λ , λ , λ ,

209

300 400 500 600
0
2
4
6
8
10
uv174fs
uv204fs
uv254fs
uv274fs
vis174fs
vis204fs
vis254fs
vis274fs
Spec λ 174 , ( ) 2 −
Spec λ 204 , ( ) 2 −
Spec λ 224 , ( ) 2 −
Spec λ 254 , ( ) 2 −
uvλ uvλ , uvλ , uvλ , visλ , visλ , visλ , visλ , λ , λ , λ , λ ,

210

300 400 500 600
2
4
6
8
10
12
uv1004fs
uv1204fs
uv1404fs
uv1804fs
vis1004fs
vis1204fs
vis1404fs
vis1804fs
Spec λ 1004 , ( ) 2 −
Spec λ 1204 , ( ) 2 −
Spec λ 1404 , ( ) 2 −
Spec λ 1804 , ( ) 2 −
uvλ uvλ , uvλ , uvλ , visλ , visλ , visλ , visλ , λ ,

211

300 400 500 600
0
5
10
uv12ps
uv22ps
uv32ps
uv42ps
uv52ps
vis12ps
vis22ps
vis32ps
vis42ps
vis52ps
Spec λ 12000 , ( ) 2 −
Spec λ 22000 , ( ) 2 −
Spec λ 32000 , ( ) 2 −
Spec λ 42000 , ( ) 2 −
Spec λ 52000 , ( ) 2 −
uvλ uvλ , uvλ , uvλ , uvλ , visλ , visλ , visλ , visλ , visλ , λ , λ , λ , λ , λ ,

212

300 400 500 600
0
5
10
uv102ps
uv202ps
uv302ps
uv442ps
vis102ps
vis202ps
vis302ps
vis442ps
Spec λ 102000 , ( ) 2 −
Spec λ 202000 , ( ) 2 −
Spec λ 302000 , ( ) 2 −
Spec λ 442000 , ( ) 2 −
uvλ uvλ , uvλ , uvλ , visλ , visλ , visλ , visλ , λ , λ , λ , λ ,

213
Neat Water data for simulation. Data has been dispersion adjusted. Spec is the sum of all the
transient spectra weighted by its population. Electrons from the small CuBr32- detachment and from
water are added in. Since spectral magnitude is in extinction coefficient, and data magnitude is mOD,
Spec is reduced to fit data.
h2oSw λ t , ( ) h2oS λ t , ( ) 0.0004 ⋅ :=

214

300 400 500 600 700
0
2
4
6
8
10
Data
Wavelength, nm
mOD
uvw100fs
uvw300fs
uvw500fs
uvw1ps
uvw2ps
uvw100ps
uvw200ps
uvw450ps
visw100fs
visw300fs
visw500fs
visw1ps
visw2ps
visw100ps
visw200ps
visw450ps
uvwwave uvwwave , uvwwave , uvwwave , uvwwave , uvwwave , uvwwave , uvwwave , viswwave ,

215

400 600 800
0
2
4
6
8
10
Data
Wavelength, nm
mOD
uvw100fs
uvw300fs
uvw500fs
uvw1ps
uvw2ps
visw100fs
visw300fs
visw500fs
visw1ps
visw2ps
h2oSw λ 120 , ( )
h2oSw λ 300 , ( )
h2oSw λ 500 , ( )
h2oSw λ 1000 , ( )
h2oSw λ 2000 , ( )
uvwwave uvwwave , uvwwave , uvwwave , uvwwave , viswwave , viswwave , viswwave , viswwave , viswwave , λ ,

216

400 600 800
0
2
4
6
8
10
Data
Wavelength, nm
mOD
uvw2ps
uvw100ps
uvw200ps
uvw300ps
uvw450ps
visw2ps
visw100ps
visw200ps
visw300ps
visw450ps
h2oSw λ 2000 , ( )
h2oSw λ 100000 , ( )
h2oSw λ 200000 , ( )
h2oSw λ 300000 , ( )
h2oSw λ 450000 , ( )
uvwwave uvwwave , uvwwave , uvwwave , uvwwave , viswwave , viswwave , viswwave , viswwave , viswwave , λ ,

217
Bromide in water data and simulations

300 400 500 600
0
1
2
3
4
Data
Wavelength, nm
mOD
br10ps
br50ps
br100ps
br200ps
br300ps
br400ps
br480ps
h2oSbr λ 10000 , ( )
h2oSbr λ 50000 , ( )
h2oSbr λ 100000 , ( )
h2oSbr λ 200000 , ( )
h2oSbr λ 300000 , ( )
h2oSbr λ 400000 , ( )
h2oSbr λ 480000 , ( )
brwave brwave , brwave , brwave , brwave , brwave , brwave , λ ,

h2oSbr λ t , ( ) Se λ t , ( ) 0.4 ⋅ h2oS λ t , ( ) 2.1 ⋅ + S7 λ t , ( ) S6 λ t , ( ) + ( ) 1.3 ⋅ + [ ] 0.00015 ⋅ :=

Abstract (if available)

Abstract Ultrafast pump-probe and pump-broadband probe spectroscopy and timecorrelated-single-photon-counting (TCSPC) techniques are used to study two different systems that undergo charge separation upon UV excitation. The first system is a model transition metal coordination compound, aqueous tribromocuprate(I) anion (CuBr₃²⁻). CuBr₃²⁻ has an absorption band centered at 280 nm, assigned as charge-transfer-tosolvent (CTTS) since resonant excitation in this band produces solvated electrons as reported in the literature. However, unlike most CTTS systems where ejection occurs on a femtosecond timescale, electron ejection has previously been reported for CuBr₃²⁻ to occur over nanoseconds, and the anion also undergoes intersystem crossing from its initially populated CTTS state to a triplet state. The spectral and kinetic features obtained in the current ultrafast experiments reveal that UV excited CuBr₃²⁻ has complex femtosecond/picosecond excited state dynamics and a surprisingly small quantum yield of prompt solvated electrons via its charge-transfer-to-solvent (CTTS) state. Analysis of data from a combination of broadband, low intensity, and photon counting experiments contributes to a proposed kinetic model, which includes some of the transient species reported for nanosecond flash photolysis. Aside from transient kinetics, an oscillatory signal was analyzed and it is determined that a vibrational wavepacket on the potential energy surface of the CuBr₃²⁻ CTTS state is launched via resonant excitation by the pump pulse. This vibrational coherence survives intersystem crossing into a CuBr₃²⁻ triplet state.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Capturing the sun: leveraging excited state dynamics

Asset Metadata

Creator Suffern, Diana Masayo (author)

Core Title Charge separation in transition metal and quantum dot systems

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Chemistry

Publication Date 05/07/2011

Defense Date 11/19/2010

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

Tag CdSe,cdse/zns,cdte,charge separation,charge transfer,copper bromide,ctts,OAI-PMH Harvest,tribromocuprate

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Bradforth, Stephen E. (committee chair), Dappen, Werner (committee member), Reisler, Hannah (committee member)

Creator Email dianamwarren@gmail.com,suffern@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m3925

Unique identifier UC1435969

Identifier etd-Suffern-4234 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-474735 (legacy record id),usctheses-m3925 (legacy record id)

Legacy Identifier etd-Suffern-4234.pdf

Dmrecord 474735

Document Type Dissertation

Rights Suffern, Diana Masayo

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