INSIGHTS INTO THE PHOTOPHYSICAL REACTION COORDINATE OF 1
ST
 
ROW TRANSITION 
METAL CHROMOPHORES FROM VIBRATIONAL COHERENCES
 
 
By
 
 
Bryan Christopher Paulus
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
A DISSERTATION
 
 
Submitted to
 
Michigan State 
University
 
in partial fulfillmen
t of the requirements
 
for the degree of
 
 
Chemistry
 

 
Doctor of Philosophy
 
2020
 
ABSTRACT
 
INSIGHTS INTO THE PHOTOPHYSICAL REACTION COORDINATE OF 1
ST
 
ROW TRANSITION 
METAL CHROMOPHORES FROM VIBRATIONAL COHERENCES
 
By
 
Bryan Christopher Paulus
 
 
Controlling the 
processes which occur following absorption of photon is beneficial for any 
conceivable application which seeks to convert light into chemical potent
ial. Transition metal 
chromophores often undergo ultrafast photoinduced transformations which involve signif
icant nuclear 
motion. This intricate relationship between electronic and nuclear degrees of freedom suggests significant 
mixing of their wavefunctio
ns and an interdependence of these molecular properties that can map onto the 
photophysical dynamics. In sys
tems such as these, the rich dynamical information encoded in vibrational 
coherences can in principle provide unique insight into the nature of this
 
coupling. Furthermore, this has 
the potential to provide subtle clues for how one could exert some degree o
f kinetic control through 
informed structural, compositional, and environmental modifications. This dissertation describes efforts to 
use excited st
ate vibrational coherences to glean mechanistic information about ultrafast excited state decay 
in several f
irst
-
row transition metal chromophore systems and, further, to exploit that information to 
reengineer the chromophore to alter the photophysical pro
perties.
 
This dissertation begins by using ultrafast transient absorption measurements to provide an updated
 
kinetic model for a series of chromium(III) tris
-
betadiketonate compounds. Following prompt 
4
T
2
 

2
E 
ISC to the lowest energy excited state, therma
lly activated back
-
intersystem crossing repopulates the 
4
T
2
 
state which internally converts to the ground state on the ~ 1
 
ps timescale. Steric bulk in the periphery of 
the molecule reduces the rate of in
ternal conversion resulting in significantly differe
nt spectral evolution. 
Identical low frequency symmetric breathing modes with dephasing times ranging from ~200
-
2500
 
fs were 
identified in the 
2
E excited state of each molecule. The more rapid dephasing t
imes are likely due to IVR. 
Similar methodology was 
then used to characterize the excited state dynamics in a structurally related series 
of cobalt(III) tris
-
betadiketonate compounds. Following ligand field excitation into the 
1
T
1g
 
state, each 
compound had
 
essentially identical biphasic kinetics with ground
 
state recovery occurring with a ~2
 
ps time 
constant from an excited state of (
t
2g
)
5
(
e
g
*
) electron configuration. Low frequency metal
-
ligand breathing 
modes similar to those observed in the chromium syste
ms were observed with dephasing times consistently 
o
n the order of ~200
 
fs.
 
Finally, two methods to elongate MLCT lifetimes of iron(II) polypyridyl compounds were 
demonstrated. In the first, the vibrational modes which drive the ultrafast, sub
-
200
 
fs MLCT 
deactivation 
to the lower energy metal
-
centered exci
ted states in a [Fe(cage)]
2+
 
control molecule were identified by their 
coherent oscillations in a transient absorption experiment. These modes were subsequently hindered by 
incorporation of electronically
 
benign copper(I) atoms to the N
4
 
coordination envir
onments in the periphery 
of the ligand scaffold resulting in a > 20
-
fold increase in the MLCT lifetime. In the final study, it was shown 
that the excited state properties of iron(II) polypyridyl type syst
ems can be systematically and dramatically 
tuned by 
swapping out a single bidentate phenanthroline ligand of a [Fe(phen)
3
]
2+
 
control molecule with 
stronger
-
field cyanide or acyclic dicarbene ligands. This modular approach to tuning the kinetics resulted 
in
 
a > 25
-
fold increase in the MLCT lifetime and a fun
damentally different decay pathway. 
 
 
iv
 
ACKNOWLEDGMENTS
 
 
 
I first want to thank Jim for allowing me into his group, giving me a cool project, funding my way 
for ACS and ICP trips, and allowing me to work on
 
projects that I was interested in. Next, I 
would like to 
thank the previous McCusker group ultrafast laser members Eileen, Jennie, and Monica for not only 
teaching me how to use the laser and set up experiments but also for coming to the rescue when I cou

fix things. I would also like to thank
 
Hayden for coming along and taking the laser reigns. To all the 
synthesis lab members, thank you for all your advice and helping hands. Thank you for letting me borrow 
your various hand
-
prepared solutions to use
 
in syntheses, and in several cases, making 
molecules that I 
could study with TA. I would further like to thank all McCusker group members for being willing to talk 
science, to read through my writings to make me sound less foolish, and for listening to my
 
presentations 
to make me a better presenter
. Finally, I would like to thank my family for always supporting me.
 
 
 
 
v
 
TABLE OF CONTENTS
 
 

..
.
.
 
vi
i
i
 

.

 
x
i
 
KEY TO SYMB
OLS AND ABBREVIATIONS 


 
xx
ii
 
Chapter 1: Coherent Vibrational Dynamics of Inorganic Molecular Chromophores
 

 
1
 
1.1
.
 

 
1
 
1.2
.
 



 
1
 
1.3
.
 
T

 
4
 
1.4
.
 

 
7
 
1.5
.
 


 
1
3
 
REFERENCES 


 
14
 

3
 

 
16
 
2.1
.
 

 
16
 
2.2
.
 


...
 
19
 
2.3
.
 

 
31
 
2.3.1
.
 
Synthesis of Cr(acac)
3
 

 
31
 
2.3.2
.
 

..
 
33
 
2.3.3
.
 


..
 
3
3
 
2.4
.
 


 
43
 
2.
4.1
.
 
Cr(acac)
3
 

..
 
43
 
2.4.2
.
 
Cr(3
-
NO
2
ac)
3
 


 
62
 
2.4.3
.
 
Cr(3
-
Clac)
3
 

...
 
65
 
2.4.4
.
 
Cr(3
-
Brac)
3
 


..

 
70
 
2.4.5
.
 
Cr(3
-
Iac)
3
 

..

 
74
 
2.4.6
.
 
Conclusions on 3
-

3
 

..

 
80
 
2.4.7
.
 
Effects of Increasing 

..

 
8
3
 
2.4.8
.
 
Cr(HD)
3
 


..

 
84
 
2.4.9
.
 
Cr(DMHD)
3
 

...
..
 
87
 
2.4.10
.
 

 
89
 
2.4.11
.
 
Cr(TMHD)
3
 


 
90
 
2.4.3
.
 
Cr(4
-
Cl
-
TMHD)
3
 


 
96
 
2.5
.
 

3
 

..
 
100
 

 
103
 
Appendix 2.A. Crystal Structures 

 
10
4
 


 
11
8
 
Appendix 2.C. Fit of Cr(acac)
3
 

 
1
19
 

 
120
 
Chapter 3: Excited State Dynamics o

3
 

 
12
4
 
3.1
.
 


 
12
4
 
3.2
.
 

 
12
8
 
3.3
.
 


 
13
2
 
3.3.1
.
 

3
 

 
13
2
 
3.3.2
.
 


...
..
 
13
2
 
 
vi
 

 
13
2
 
3.3.4
.
 

..
..
 
13
6
 
3.4
.
 

...

 
13
6
 
3.4.1
.
 
Co(acac)
3
 


...

 
13
6
 
3.4.2
.
 
Co(3
-
Iac)
3
 

..

 
15
3
 
3.4.3
.
 
Co(TMHD)
3
 



 
15
7
 
3.5
.
 

3
 

..

 
16
0
 


 
16
4
 


 
16
5
 

16
7
 

 
17
4
 
Chapter 4: Geometric Control of Ultrafast Charge
-
Trans
fer Relaxation Dynamics in Fe(II)
-
Based 
Chromopho

 
17
8
 
4.1
.
 


 
17
8
 
4.2
.
 

...
.
 
1
79
 
4.2.1
.
 


...
.
 
1
79
 
4.2.2
.
 

...
 
18
0
 
4.2.3
.
 

 
18
0
 
4.3
.
 

...
 
18
2
 
APPENDICE

 
19
8
 

 
199
 
Appendix 4.B. Reorganization Energy Calculation 
f
or [Fe(bpy)
3
]
2+

 
20
0
 
Appendix 4.C. Additional Normal Mode Assignment for
 

 
20
1
 

 
20
2
 

 
20
7
 
Chapter 5: A Mod
ular Approach to Controlling Excited State Properties of Iron(II) Pol

 
21
1
 
5.1
.
 
Introduction
 

 
21
1
 

 
21
3
 
5.2.1
.
 


 
21
3
 
5.2.2
.
 


 
21
4
 
5.3
.
 

..
 
21
5
 
5.3.1
.
 
Ground
-

 
21
5
 
5.3.2
.
 

 
21
6
 
5.3.3
.
 
Excited State Dynamics of 

 
2
19
 
5.3.4
.
 

 
22
3
 
5.4
.
 

 
22
8
 

.
 
2
29
 
Appendix 5.A. Singular Value Decomposition MAT
LAB Script for Full Spectrum TA Analysis
 


 
23
0
 
Appendix 5.B. Si

 
23
3
 

 
23
4
 
Appendix 5.D. Additional TA Data wit

 
23
7
 

 
24
2
 
Chapter 

 
24
5
 
6.1
.
 

..

 
24
5
 
6.2
.
 

3
 

...

 
24
5
 
6.3
.
 
Excited State Dynamics of Six
-
Coordinate Cobalt(III) 

24
7
 
6.4
.
 
Coherence as an Indicator 

...

 
24
8
 
 
vii
 
6.5
.
 
Modular Manipulation of Excited St

 
2
49
 

 
25
0
 
 
 
 
 
viii
 
LIST OF TABLES
 
 
Table
 
1.1.
 
Summary of the extinction coefficients, 

, and oscillator strengths, 
f
, typically observed for 
various electronic transitio

 
3
 
Table
 
2.1.
 
Summary of oscillatory components observed for variable excitation wavelength TA 
experiments 
conducted by Foszcz
12
 
on Cr(acac)
3
.
 
Errors are reported as standard deviations of all the data 
lis

 
30
 
Table
 
2.2.
 
Summary of coherent vibrational modes of Cr(acac)
3
 
observed during TA experiments in various 
solvents
. Reported IRFs are rounded to the nearest integer multiple of 5
 





 
50
 
Table
 
2.3.
 
Summary of the coherent vibrational modes of 
Cr(3
-
NO
2
ac)
3
 
following 
4
A
2
 

 
4
T
2
 
excitation. 
All measurements performed in Dioxane used 4 and 0.4 

J pump and probe pulses respectively while the 
THF measurement employed 4.7 and 0.5 

J pump a
nd probe pulses. The dioxane and THF experiments 
had 40 and 35 f
s IRF

 
6
2
 
Table 2.4.
 
Summary of observed coherent vibrational modes of Cr(3
-
Clac)
3
 
in toluene following excitation 
into the 
4
T
2
 
state. The sub
-
100 cm
-
1
 
oscillatory components picked up by the LPSVD fitting program were 
included for transparency, however, they are likely due to poor exponential fitting of the data and not to 
coherent vibrational motion. All experiments employed 4 and 0.4 

J pump and prob
e
 
pulses respectively 


 
6
8
 
Table
 
2.5.
 
Summary of the coherent oscillations of Cr(3
-
Brac)
3
 
following excitation into the 
4
T
2
 
state. 
Experimental parameters such as pump and probe wavelengths and pulse ener
g
ies as well as IRF FWHMs 
are also included for each experiment. 
...
....................................................................................................
 
7
4
 
Table
 
2.6.
 
Summary of the coherent oscillations of Cr(3
-
Iac)
3
 
following 600
 
nm excit
a
tion into the 
4
T
2
 
state. Each measurement employed 4 and 0.4 

J pump and probe pulse energies respectively with a 40 fs 


 
7
9
 
Table
 
2.7.
 
Summary of the oscillatory components observed for Cr(TMHD)
3
 
following exci
t
ation into the 
4
T
2
 
state across several solvents. Experimental details such as the IRF and pump and probe wavelengths 
and pulse energies are included for each experiment. Probe pulse energies were not measured when using 
a continuum probe and are labeled 
b

 

 
9
1
 
Table
 
2.8
.
 
Summary of coherent oscillations observed in TA data on Cr(4
-
Cl
-
TMHD)
3
 
following 600
 
nm 
excitation into the 
4
T
2
 
state while monitoring at 600
 
nm in dichloromethane. The sub
-
100
 
cm
-
1
 
oscillatory 
components
 
were included for transparency; however, it is likely that they arise from poor exponential 
fitting and do not truly represent coherent vibrational motion of the 
solute. Experimental details such as the 
IRF and the pump and probe pulse energies are includ

 
9
8
 
Table
 
2.B.1.
 
Frequencies and dephasing times of common solvent vibrational modes as determined from 
transient absorption spe

 
11
8
 
Table
 
3.1.
 
Summary of the observed kinetics in refe
rence 8

 
12
8
 
Table
 
3.2.
 
Summary of kinetic components from single wavelength transient absorption experiments on 
Co(acac)
3
 
in acetonitrile following 600
 
nm excitation into the 
1
T
1g
 

 
1
39
 
 
ix
 
Table
 
3.3.
 
Summary of the coherent v
ibrational modes observed for Co(acac)3
 
in acetonitrile. Pump
 
and 
probe pulse energies, E
pu,
 
and E
pr
, are provided for each experiment. The IRF was 40
 
fs for each experiment
 

 
1
4
2
 
Table
 
3.4.
 
Comparison of the oscilla
tion amplitude and the preexponential factor for the 510
 
fs kinetic 
component of Co(acac)
3
 
observed during TA experiments. Oscillation amplitudes and preexponential 
factors were either both directly acquired from LPSVD fits or sep
arately acquired from an e
xponential fit 

 
1
5
1
 
Table
 
3.5.
 
Summary o
f kinetic components from single wavelength transient absorption experiments on 
Co(3
-
Iac)
3
 
in toluen
e following 600
 
nm excitat
ion into the 
1
T
1g
 

 
1
5
4
 
Table.
 
3.6.
 
Summary of the coherent vibrational modes for Co(3
-
Iac)
3
 
in toluene following 600
 
nm 
excitation into the 
1
T
1g
 
excited state. All measurements had 40
 
fs IRF and employed 2.9 a
nd 0.4 

J pump 
and probe p
ulse energies respectively.
 


 
1
5
6
 
Table
 
3.7.
 
Summary of kinetic components of Co(TMHD)
3
 

 
1
59
 
Table
 
3.8.
 
Summary of coherent oscillations of Co(TMHD)
3
 
in DCM following 
600
 

 
1
59
 
T
able
 
3.9.
 

3
 

 
1
6
0
 
Table
 
4.1.
 
Summary of the coherent vibrational modes of 
[Fe(L)](PF
6
)
2
 
(
1
) observed during
 
TA 
experiments following 600
 

 
1
8
7
 
Table
 
4.2.
 
Summary of the coherent vibrational modes of [FeCu
2
(L)(PF
6
)
4
 
(
2
) in different solvents. All 
errors are expressed as the s
tandard deviation of multiple tri

 
1
9
1
 
Table
 
4.3.
 
Summary kinetic components of [FeCu
2
(L)]
2+
 
in various solvents and with PF
6
-
 
and Bar
F
-
 

 
1
9
5
 
Ta
ble
 
4.D.1.
 
Coherent oscillatory c
omponents observed for individual TA datasets on [Fe(L)[(PF
6
)
2
 
in 
acetonitrile solvent following 600
 

 
20
2
 
Table
 
4.D.2.
 
Kinetic components and fit parameters for individual TA datasets on 
[FeCu
2
(L)](PF
6
)
4
 
in 
various solve
nts following 600
 
nm excitation. A 7
 
ns kinetic component was locked in to describe the slow 

 
20
3
 
Table
 
4.D.3.
 
Coherent oscillatory component
s observed in individual TA datas
ets on [FeCu
2
(L)](PF
6
)
4
 
in 
various solvents following 600
 
nm excitation. The observed sub
-
100
 
cm
-
1
 
components are not expected to 
be physically meaningful.
 

 
20
4
 
Table
 
4.D.4.
 
Kinetic components and fit parameters for individual
 
TA datasets on [FeCu
2
(L)](Bar
F
)
4
 
in 
various solvents following 600
 
nm excitation and while monitoring at 700
 
nm. A 7 ns kinetic component 
was locked in and used to describe the slow ground state recover
y process and account for any baseline 


 
20
5
 
Table
 
4.D.5.
 
Results of individual kinetic fits of [Fe(L)](PF
6
)
2
 
TA data following 600
 
nm excitation in 

 
20
6
 
T
a
ble
 
5.1.
 
Electrochemical and spectroscopic parameter
s of [Fe(phen)
3
]
2+
 
(
1
), [Fe(phen)
2
(CN)
2
] (
2
), and 
[Fe(phen)
2
(C
4
H
10
N
4
)] (
3

 
2
1
6
 
 
x
 
Table
 
5.2.
 
Summary of the kinetic components of [Fe(phen)
3
](BF
4
)
2
 
(
1
), [Fe(phen)
2
(CN)
2
] (
2
), and 
[Fe(phen)
2
(C
4
H
10
N
4
)
](PF
6
)
2
 
(
3

 
2
2
8
 
Table
 
5.D
.
1
.
 
Summary of coherent oscillations of [Fe(phen)
3
](BF
4
)
2
 
(
1
) observed in TA experiments in 
methanol following 520
 
nm 
excitation. Experiments were conducted similarly in both wavelength re
gions, 
except that in the 610
-
630
 
nm region, several separate experiments were conducted individually at 610, 
620, and 630
 
nm and the results of the LPSVD fits from each of these probe w
avelengths were averaged 


 
2
3
8
 
Table
 
5.D
.
2.
 
Summary of the coherent oscillations of [Fe(phen)
2
(C
4
H
10
N
4
)](PF
6
)
2
 
(
3
) in methanol and 
acetonitrile following 600
 
nm excitation. The presented results are from a sing
le dataset for each probe 
wavelength and solvent. As such, no error ba
rs in the form of a standard deviation are presented for these 

 
2
4
0
 
Table
 
5.D
.
3. 
Summary of the kinetic components of [Fe(phen)
2
(C
4
H
10
N
4
)](
PF
6
)
2
 
(
3
) in methanol and 
acetonitrile following 600
 
nm excitation. Er
rors are reported as the standard deviation of multiple trials.
 


 
2
4
1
 
 
 
 
xi
 
LIST OF FIGURES
 
 
Figure
 
1.1.
 
Jablonski diagram depicting various excited state processes including vibrational relaxation 
(VR), internal conver
sion (IC), intersystem crossing (ISC), fluorescence (FL), and phosphorescence (PH). 
Radiative transitions are shown as solid lines and nonrad

 
3
 
Figure
 
1.2.
 
Energy level diagram showing possible signals in a p
ump
-
probe transient absorption 
experiment including ground state bleach (GSB), excited state absorption (ESA) and stimulated emission 

 
5
 
Figure
 
1.3.
 
Illustra
tion of the different ways pump
-
probe TA data can be 
viewed. In the 2
-
dimensional data 
set shown in (a) the entire transient spectrum is collected at each time delay. A vertical slice of this data 
(black) contains the differential absorption at a single wa
velength across a series of time delays. This is 
illu
strated in the black boxed kinetics trace of (b). A horizontal slice shown in red represents the transient 
spectrum at a specific pump
-

 
6
 
Figure
 
1.4.
 
Resonance as an analogy to coherence. 
The ground state wavefunction of benzene is a linear 

 
7
 
Figure
 
1.5.
 
Generation of a coherent superposition state, or wavepacket, as
 
a linear combination of 


 
8
 
Figure
 
1.6.
 
Potential energy surface diagram depicting the process of forming a wavepacket on the ground 

 
9
 
Figure
 
1
.7.
 
Simulated wavepacket propagation on an excited st
ate potential energy surface which (a) has 
the same force constant as the ground state, (b) has a different force constant than the ground state, (c) is 
anharmonic, and (d) is dissociative. The initial p
rojected wavepacket is shown in red. Wavepackets that
 

..
 
10
 
Figure
 
1.8.
 
(a) Effect of coherent vibrational motion on the entire transient absorption spectrum. Band 
oscillat
es according to the frequencies of the activated vibr
ations. (b) Manifestation of coherent vibrational 
motion on single wavelength TA measurements. Oscillatory features are superimposed upon the population 
dynamics. (c) Fourier transformation (blue) of osc
illatory components of TA data results in a vibration
al 
spectrum of the molecule. LPSVD fit of TA data provides not only frequencies of coherent vibrational 

 
1
1
 
Figure 
2
.1.
 
Qualitative potenti
al energy diagram representing the pertinent electron
ic states of Cr(acac)
3
. 
The 
2
T
1g
 
and 
2

2

electron configurations in the 1
-
electron picture are represented for each 

 
1
8
 
Figure
 
2.2.
 
Electron dens
ity distributions for the pertinent electronic ground and excited states of d
3
 
compounds in an O
h
 
ligand field. Distributions were calculated using the LIGFIELD program with the help 
of Jon Yarranton.
 




 
1
9
 
Figure
 
2.3.
 
(lef
t) Transient absorption spectrum of Cr(acac)
3
 
in acetonitrile solution at 

t
 
= 5 ps (solid line) 
along with normalized absorption amplitudes obtained from single
-
wavelength nanosecond transient 
absorptio
n data (red markers) collected in a 9:2 propionitrile/butyronitrile glass at 90
 
K. Close agreement 
between these two datasets s
uggests that Cr(acac)
3
 
is in the fully thermalized lowest energy excited state, 
the 
2
E state, by 

t
 
= 5 ps. (right) Transient abs
orption spectrum of Cr(acac)
3
 
over the course of the first 5 
 
xii
 
ps following 625 nm excitation. Reprinted with permission from 
r
ef
erence 7.
 
© (2005) American Chemical 
Society.
 


 
20
 
Figure
 
2.4.
 
(a) Single
-
wavelength kinetic trace of Cr(acac)
3
 
monitored at 592 
±
 
2 nm following 600 nm 
excitation in acetonitrile. An exponential increase in excited s
tate absorbance (

 
= 320 fs) was subtracted 
from the data to better illustrate the oscillatory compon
ent of the data. The solid line corresponds to a fit to 
two exponentially damped oscillatory components with frequencies and damping times of 164 
±
 
20 cm
-
1
 
(

damp
 
= 70 fs) 
and 75 
±
 
25 cm
-
1
 
(

damp
 
= 1.6 ps). (b) Density functional theory calculated vibratio
nal mode 
on the ground electronic state. The excited state wavepacket was proposed to exhibit a similar motion to 
this 184
 
cm
-
1
 
calculated symmetric breathing mode. 
Figure 
reproduced from reference 10 with permission 
of the Royal Society of Chemistry. 



 
2
1
 
Figure
 
2.5.
 
(a) Transient absorption spectra of Cr(TMHD)
3
 
in dicloromethane following 633 nm excitation. 
(b) Single
-
wavelength kinetic trace col
lected at 450 nm and (c) 550 nm following 610
 
nm excitation. Solid 
lines represent s
ingle exponential fits (with 

obs
 
= 2.0 
±
 
0.2 ps and 1.6 
±
 
0.2 ps for (b) and (c) respectively).
 
Figure reproduced from reference 10 with permission of the Royal Society of Chemistry



 
2
2
 
Figure
 
2.6.
 
(a) Plot of the elements of the Duschinsky displacement vector for the distorti
on of the ground 
state geometry to 
4
T
2
 
state equilibrium geometry as a function of vibrational modes on the ground electronic 
state inset showing an overlay of these two geomet
ries. (b) Time dependent populations calculated using 
wavepacket analysis. (c) P
lot of the position expectation value as a function of time during wavepacket 


 
2
3
 
Figure
 
2.7.
 
(a) Steady state (lower trace) and transient difference spectra
 
(upper traces) following 400
 
nm 
excitation of Cr(acac)
3
 
in tetrachloroethylene. (b) Evolution of the 1521 cm
-
1
 
bleach recovery (

C=C) 
following 400 nm excitation. Solid line 
corresponds to a biexponential fit with 


 
= 15 ps and 


 
= 760 ps. 
Inset superp
oses the bleach and the inverse of the excited state absorption feature at 1509 cm
-
1
. The sharp 
feature at early time is part of the coherent artifact.
 
Figure reprinted with pe
rmission from reference 13. © 


 
2
4
 
Figure
 
2.8.
 
(a) Time dependent shifting for 

CO and 

C=C stretching ESA bands of Cr(acac)
3
 
in 
tetrachloroethylene following 
4
A
2
 

 
4
T
1
 
(400
 
nm) and 
4
A
2
 

 
4
LMCT (
345
 
nm) excitation. (b) Time 
dependent narrowing of the for 

CO and 

CC ESA features. (c) Bleach recovery kinetics of the 

C=C 
stretching ba
nd following excitation into the 0 

 
1 transition at 1521 cm
-
1
 
with fit shown in red. The inset 
shows a comparison of the decay dynamics following excitation at various wavelengths. Data in inset are 
normalized to the value at 10 ps. 
Reprinted with
 
with pe
rmission from reference 14.
 
© (2015) American 
Chemical Society
.
 


 
2
5
 
Figure
 
2.9.
 
Simplified Energy level diagram for Cr(acac)
3
 
depicting the relaxation processes
 
proposed in 
reference 14



 
2
6
 
Figure
 
2.10.
 
Bleach recovery dynamics of Cr(acac)
3
 
in t
etrachloroethylene with experimental data 
(symbols) and simulated (solid lines) ground state populations. 
Reprinted with with permission from 
reference 14. © (2015) American Chemical Society.




 
2
8
 
Figure
 
2.11.
 
Early time full spectrum transient absorption data on (a) Cr(3
-
Br
-
acac)
3
, (b) Cr(3
-
SCN
-
acac)
3
, 
(c) Cr(3
-
NO
2
-
acac)
3
, and (d) Cr(3
-
Ph
-
acac)
3
 
following 
4
A
2
 

 
4
T
2
 
excitation.
 
Reproduced from data in 
Reference 15



 
2
9
 
Figure
 
2.12.
 
Representative OKE trace collected in methanol with 600
 
nm degenerate pump and probe. 

 
35
 
 
xiii
 
Figure
 
2.13.
 
(a) Repre
sentative TA data exhibiting coherent oscillations (blue). The corresponding LPSVD 
fit (red) and exponential components acquired from LPSVD fit (black) are superimposed on the original 
dataset. The Residuals of each of these with respect to the r
aw data ar
e shown above in red and black 

 
38
 
Figure
 
2.14.
 
Illustration of methodology used to systematically identify origin of coherent oscillations in 
TA data.
 
TA data a
nd corresponding LPSVD fits of (a) Cr(TMHD)
3
 
dissolved in benzene in a quartz 
cuvette, (b) neat benzene in a quartz cuvette, and (c) an empty quartz cuvette. (d) Superimposed LPSVD 
magnitude spectra normalized to the common peak at 500
 
cm
-
1



 
39
 
Figure
 
2.15. 

 
4
2
 
Figure
 
2.16.
 
(a) Full spectrum transient absorption data on Cr(acac)
3
 
following 
4
A
2
 

 
4
T
1
 
(400
 
nm) 
excitation in acetonitrile. (b) Decay associated difference
 
spectra (
DADS) associated with the global fit of 
the data in (a). A 1
 
ns lifetime associated with the longtime offset was locked in but had no bearing on the 

 
4
4
 
Figure
 
2.17
.
 
(a) Qualitative 
energy lev
el diagram depicting the ground 
4
A
2
 
state (



), 
2
E state (



), 
4
T
2
 
state (



), and charge transfer manifold (



) that the 
2
E and 
4
T
2
 
states can be promoted to through 
resonance with the probe beam. (b) 
Double sided Feynman diagrams depicting stimulated emission (SE), 
ground state bleachi
ng (GSB) and excited state absorption (ESA) processes. Following excitation, to the 
4
T
2
 
state, the s
ystem may evolve. Depending on the time delay between pump and probe, a population may 
exist in the 
4
T
2
 
or 
2
E states. This relaxation pathway is denoted by 
a gray box. (c) The relationship between 
signal amplitude (third order response function) and transi
tion dipole moments for relevant field 
interactions for SE, GSB, and ESA.
 

 
4
8
 
Figure
 
2.18.
 
(a) Representative TA data on Cr(acac
)
3
 
collected in acetonitrile following 600
 
nm excitation 
and monitoring at 592
 
nm. (b) Comparison of
 
results of |FFT| (blue) and LPSVD (red) analyses of data 
from part a. (c) Analogous data collected while monitoring at 480
 
nm. (d) Comparison of |FFT| and 
LPSVD 
analyses for 480
 

 
51
 
Figure
 
2.19.
 
(a) Representati
ve TA data on Cr(acac)
3
 
collected in tetrahydrofuran (THF) following 600
 
nm 
excitation while monitoring at the red edge of the excited state absorption feat
ure. (b) Corresponding FFT 
and LPSVD analyses. (c) Analogous data collected in 1,4
-
dioxane and (d) c
orresponding |FFT| and LPSVD 
vibrational spectra. It is possible that the 271
 
cm
-
1
 
band in part
 
b is a vibrational mode of THF. The  
~500
 
cm
-
1
 
component in 
part
 

 
5
2
 
Figure
 
2.20.
 
(a) TA data collected on an empty 1
 
mm glass cuvette. The observed coherent oscillations are 
assigned to ground state vibrational coherence in the cuvette itself. (b) Fast Fo
urier transform and LPSVD 
spectrum of the coherent oscillations of (a), showing a
 
~ 500
 
cm
-
1
 

 
5
3
 
Figure
 
2.21.
 
Resonance Raman spectra collected on neat THF (black) and Cr(acac)
3
 
(red). The solvent 
background was subtracted from the Cr(
acac)
3
 
dataset. The 488
 
nm excitation source is resonant with the 
blue edge of th
e 
4
A
2
 

 
4
T
2
 

 
5
4
 
Figure
 
2
.
22.
 
(a) Raman spectrum of neat tetrahydrofuran (blue) with 488
 
nm excitation source. A 
Lorentzian fit to a band centered at 913
 
cm
-
1
 
(red) shows a FWHM of 12.9
 
cm
-
1
 
corresponding to a 
dephasing
 
time 
of 820 
±
 
30
 
fs. (b) TA data on neat THF with 600
 
nm excitation source and 584
 
nm probe 
(blue) along with LPSVD fit of the dataset (red). The average dephasing time measured with four datasets 
is 950 
±
 
140
 

 
5
7
 
Figure
 
2.
23
.
 
Nuclear deformations associated with the 193
 
cm
-
1
 
(a) and 461
 
cm
-
1
 
(b) vibrational modes of 
 
xiv
 
Cr(acac)
3
.
 

 
5
9
 
Figure
 
2.24.
 

3
 

 
61
 
Figure 2.
25. 
Representative TA data on Cr(3
-
NO
2
ac)
3
 
following 600 nm excitation into the 
4
T
2
 
state in 1,4
-
dioxane while monitoring a 4 nm slice of the probe spectrum at (a) 608 nm, (c) 600 nm, and (
e) 592 nm. 
FFT and LPSVD analyses corresponding to the (b) 608 nm, 

 
6
3
 
Figure 2.26.
 
Nuclear deformation associated with the 150 cm
-
1
 
vibrational mode of Cr(3
-
NO
2
ac)
3

 
6
4
 
Figure
 
2.27. 
Absorption spectrum of Cr(3
-
C
lac)
3
 

 
6
5
 
Figure
 
2.28.
 
(a) Early t
ime full spectrum TA data on Cr(3
-
Clac)
3
 
in toluene following 600
 
nm excitation 
into the 
4
T
2
 
state. (b) Decay associated difference spectrum (DADS) associated with the global fit of the 
TA data to a sequential biexponential decay. The long
-
time offset was 
modeled by fixing a 1 ns exponential 
decay to describe the slower ground state recovery proc
ess. (c) Evolution associated difference spectra 

 
6
6
 
Figure 2.29.
 
Representative single wavelength
 
TA data on Cr(3
-
Clac)
3
 
in toluene following 600
 
nm 
excitation into the 
4
T
2
 
state while moni
toring at (a) 490 nm, (b) 550 nm, and (c) 650 nm. The two bluer 
wavelengths show ~1.2
 

 
6
7
 
Fig
ure
 
2.30.
 
Representative TA traces and FFT and LPSVD analyses of coherent oscillations of Cr
(3
-
Clac)
3
 
in toluene following excitation into the 
4
T
2
 
state. Coherent oscillations were monitored at (a) 608 
nm, (c) 600 nm, and (e) 592 nm. Corresponding FFT and 
LPSVD vibrational spectra are shown in (b), (d), 


 
6
9
 
Figure
 
2.31.
 
Nuclear deformations associated with the 149 cm
-
1
 
vibrational mode of Cr(3
-
Clac)
3

 
70
 
Figure
 
2.32.
 
(a) Early time full spectrum TA data on C
r(3
-
Brac)
3
 
in toluene following 600
 
nm excitation 
into the 
4
T
2
 
state. The 570
-
630
 
nm spectra
l region was omitted due to large contamination from pump 
scatter. (b) Decay associated difference spectrum (DADS) associated with the global fit of the TA data to 
a sequential biexponential decay. The long
-
time offset was modeled by fixing a 1 ns exponent
ial decay to 
describe the slower ground state recovery process. (c) Evolution associated difference spectra (EADS) 


 
7
1
 
Figure
 
2
.
33
.
 
Representative single wavelength TA traces collected on Cr(3
-
B
rac)
3
 
in toluene following 
excitation into the 
4
T
2
 

 
7
2
 
Figure
 
2.34.
 
Representative TA traces 
and FFT and LPSVD analyses of coherent oscillations of Cr(3
-
Brac)
3
 
in toluene following excitation into the 
4
T
2
 
state. Coherent oscillations were monitored at (a) 608 
nm and (c) 592 nm. Corresponding FFT and LPSVD vibrational spec
tra are shown in (b) and (

 
7
3
 
Figure
 
2.35.
 
Nuclear deformations associated with the 123 cm
-
1
 
vibrational mode of Cr(3
-
Brac)
3

 
7
4
 
Figure
 
2.36.
 
Absorption spectrum of Cr(3
-
Iac)
3

 
7
5
 
Figure
 
2.37.
 
Crystal Structure of Cr(3
-
Iac)
3
 
collected by Sara Adelma
n and solved by Professor Richard 

 
7
5
 
Figure
 
2.38.
 
(a) Early time full spectrum TA data on Cr(3
-
Iac)
3
 
in toluene following 600
 
nm excit
ation into 
the 
4
T
2
 
state. (b) Decay associated difference
 
spectra (DADS) associated with the global fit of the TA data 
 
xv
 
to a sequential biexponential decay. The long
-
time offset was modeled by fixing a 1
 
ns exponential decay 
to describe the slower ground st
ate recovery process. (c) Evolution associated difference
 
spectra (EADS) 
associated with the global fit. Distortions around 600
 
nm in the data, DADS, and EADS are due to residual 

 
7
6
 
Figure
 
2.39.
 
Represent
ative single wavelength TA data on Cr(3
-
Iac)
3
 
in toluene 
following 600
 
nm 
excitation into the 
4
T
2
 
state while monitoring at (a) 490
 
nm, (b) 550
 
nm, and (c) 650
 

 
7
7
 
Figure
 
2.40.
 
Representative TA traces and FFT and LPSVD analyses of coherent osci
llations of Cr(3
-
Iac)
3
 
in toluene following excitation in
to the 
4
T
2
 
state. Coherent oscillations were monitored at (a) 608 nm, (c) 
600 nm, and (e) 592
 
nm. Corresponding FFT and LPSVD vibrational spectra are shown in (b), (d), and (f).
 



 
7
8
 
Figure
 
2.41.
 
Nuclear deformation
s associated with the 101 cm
-
1
 
vibrational mode of Cr(3
-
Iac)
3

 
80
 
Figure
 
2.
42
.
 
Summary of the results on the 3
-

3
 
series. Internal conversion lifetimes 
for Cr(acac)
3
 
and Cr(3
-
NO
2
ac)
3
 
were acquired from references 7 and 15 respectiv
ely.
 

 
8
1
 
Figure
 
2
.
43
.
 
Depiction of the two possible mechanisms for excited state decay at early times in 
Cr(TMHD)
3
. Mechanism
 
1 involves a slow 1.8
 
ps 
4
T
2
 

 
2
E ISC. Mechanism
 
2 involves a rapid 
equilibration of the 
4
T
2
 
and 
2
E states. As the molecule
 
begins to vibrationally coo
l the 
2

4
T
2
 
equilibrium 
shifts to the left and results in a 1.8
 
ps ISC. IC within the quartet manifold can occur, although it is not the 

 
83
 
Figure
 
2.44.
 
(a) 
Early time full spectrum TA 
data on Cr(HD)
3
 
in dichloromethane following 600
 
nm 
excitation into the 
4
T
2
 
state. (b) Decay associated difference spectra (DADS) associated with the global fit 
of the TA data to a sequential biexponential decay. The long
-
time o
ffset was modeled by fixing 
a 1
 
ns 
exponential decay to describe the slower ground state recovery process. (c) Evolution associated difference 
spectra (EADS) associated with the global fit. Distortions in the transient spectrum near 600
 
nm are due to 
resid


 
8
5
 
Figure
 
2.45.
 
Representative single wavelength TA data on Cr(HD)
3
 
in acetonitrile (a) and dichloromethane 
(b and c) following 600
 
nm excitation into the 
4
T
2
 
state while monitoring at (a) 490
 
nm, (b) 550
 
nm, and 
(c) 650
 


 
8
6
 
Figure
 
2.46.
 
(a) Early time full spectrum TA data on Cr(DMHD)
3
 
in dichloromethane following 600
 
nm 
excitation into the 
4
T
2
 
state. (b) Decay associated difference spectra (DADS) associated with t
he global fit 
of the TA data
 
to a sequential biexponential decay. The long
-
time offset was modeled by fixing a 1
 
ns 
exponential decay to describe the slower ground state recovery process. (c) Evolution associated difference 
spectra (EADS) associated with t
he global fit. Distortions i
n the transient spectrum near 600
 
nm are due to 

 
8
7
 
Figure
 
2.47.
 
Representative single wavelength TA data on Cr(DMHD)
3
 
in dichloromethane (a and b) and 
acetonitrile (c) fol
lowing 600
 
nm excitation int
o the 
4
T
2
 
state while monitoring at (a) 490
 
nm, (b) 550
 
nm, 
and (c) 586
 

 
8
8
 
Figure
 
2.48.
 
Early time full spectrum TA data on (a) Cr(acac)
3
, (b) Cr(HD)
3
, (c) Cr(DMHD)
3
, and (d) 
Cr(TMHD)
3
 
fo
llowing excitation into the 
4
T
2
 
excited state.
 
Parts (a) and (d) were reproduced with data from 
references 7 and 15 respectively.


 
90
 
Figure
 
2.49.
 
Representative transient absorption data on Cr(TMHD)
3
 
in 1,4
-
dioxane following 
exc
itation 
into the 
4
T
2
 
excited state while monitoring at (a) 592 nm and (c) 550
 
nm with corresponding vibrational 
 
xvi
 

 
9
2
 
Figure
 
2.50.
 
Representative transient absorption data on Cr(TM
HD)
3
 
in tetrahydrofuran following 
excitation into the 
4
T
2
 
excited state while monitoring at (a) 568 nm and (c) 545
 
nm with corresponding 

 
9
3
 
Figure
 
2.51.
 
Stacked plot of Ra
man
 
data on Cr(acac)
3
 
and Cr(TMHD)
3
 
in THF with 488
 
nm excitation 
source in resonance with the blue edge of the 
4
A
2
 

4
T
2
 

 
9
3
 
Figure
 
2.52.
 
Nuclear deformations associated with the (a) 173
 
cm
-
1
 
and (b) 249
 
cm
-
1
 
experimentally 
o
bse
rved coherent vibrational modes of Cr(TMHD)
3

 
9
4
 
Figure
 
2.53.
 
(a) Early time full spectrum TA data on Cr(4
-
Cl
-
TMHD)
3
 
in dichloromethane following 
600
 
nm excitation into the 
4
T
2
 
state. (b) Global fit of the TA data to a sequential b
iex
ponential decay. The 
long
-
time offset was modeled by fixing a 1
 
ns exponential decay to describe the ground state recovery 
process. (c) Decay associated difference spectra (DADS) and (d) Evolution associated difference spectra 
(EADS) associated with the
 
gl

 
9
6
 
Figure
 
2.54.
 
(a) Representative single wavelength TA data on Cr(4
-
Cl
-
TMHD)
3
 
following 600
 
nm 
excitation into the 
4
T
2
 
state in dichloromethane. Time constants for the fits of specific data sets are included 
on th
e g
raph itself, however the average time constant measured across all datasets is 1.7 
±
 
0.2
 
ps. (b) 
Zoomed in portion of the 600
 
nm probe data shown in part a with exponential and LPSVD fits and (c) 
corresponding vibrational spectrum resulting from FFT 
and LS

 
9
7
 
Figure
 
2.55.
 
Nuclear deformations associated with the 118
 
cm
-
1
 
experimentally observed coherent 
oscillations of Cr(4
-
Cl
-
TMHD)
3

 
9
8
 
Figure
 
2.56.
 
Summary of the frequencies and dephasing times 
of the
 
low frequency symmetric breathing 

3
 

 
10
1
 
Figure
 
2.A
.
1.
 
Crystal structure for Cr(DMHD)
3
 
that was collected by Sara Adelman and solved by 


 
10
6
 
Figure
 
2.A
.
2.
 
Crystal structure for Cr(4
-
Cl
-
TMHD)
3
 
that was collected by Sara Adelman and solved by 
Professor Richard Staples. The structure is 

 
11
7
 
Figure
 
2.C
.
1.
 
Resonance Raman spectrum 
of Cr(acac)
3
 
in THF solution (blue) in the vicinity of the    
455
 
cm
-
1
 
solute mode. The 488
 
nm excitation source is resonant with the blue edge of the 
4
A
2
 

4
T
2
 
transition. A fit of the 455
 
cm
-
1
 
band to a Lorentzian function with a polynomial background (red) shows a 
FWHM of 8.5 
±
 
0.1
 
cm
-
1
 
corresponding to a dephasing time of 1.25 
±
 
0.2
 

11
9
 
Figure
 
3.1.
 
Series of cobalt complexes to be studied 

 
1
2
5
 
Figure
 
3.2.
 
(a) Tanabe
-
Sugano diagram for Co(acac)
3
 
calculated by McCusker group Member Jonathan 
Yarranton using the dtansug program written by Professor Joshua Telser of Roosevelt University and using 
the Racah B and C
 
interelectronic repulsion parameters 
from reference
 
5
. The rightmost dashed line 
corres
ponds to the value of Dq presented in that paper and the left most dashed line corresponds better to 
the vertical transition energies presented therein. 
(b) Vertical tr
ansition energies acquired from 
reference 
5. 
The 
5
T
2g
 
band was not observed experimental
ly but 
calculated using the LIGFIELD program of reference 
6. (c) Qualitative potential energy surface diagram based on the Tanabe
-
Sugano diagram in part a. (d) One
-
elec

 
12
6
 
 
xvii
 
Figure
 
3.3.
 
(a
) Transient 
absorption spectrum for an aqueous solution of [Co(tppn)](ClO
4
)
3
 
following 
~500
 
fs excitation at 314 nm. (b) Transient absorption data measured at 390
 
nm wi
th a biexponential fit. 
Reprinted with permission from reference 8. © (1993) American Ch
emical Society.
 

 
1
29
 
Figure
 
3.4.
 
Potential energy surface diagram for the 
1
T
1g
, 
3
T
1g
, 
5
T
2g
 
and 
1
A
1g
 
states of [Co(NH
3
)
6
]
3+
 
along 
the 

(a
1g
) vibrational coordinat
e. The 
5
T
2g
 
surface is depicted as several dashed lines to represent a range of 
possible positions and a solid line to represent the position with largest overlap with the relaxed 
3
T
1g
 
state. 
Digitized from data in
 
reference 13


 
1
29
 
Figure
 
3.5.
 
Full spectrum transient absorption data on Co(acac)
3
 
in acetonitrile solution following 
excitation at (a) 365
 
nm, (b) 390 nm, (c) 650 nm, and (d) 580
 
nm. Transient absorption data averaged ± 5 
nm from center wavelengths of (e) 710 nm and (
f) 520
 
nm f
ollowing 390
 
nm excitation. (h) and (g) 
r
epresent analogous datasets following 365
 
nm excitation. 
Reprinted with permission from reference 4. © 

 
1
3
1
 
Fig
ure
 
3.6.
 
(a
) A pictorial representation of a cross product between two vectors, 
a
 
and 
b
 
separated by 
angle 

. The cross product, 
a
×
b
, is perpendicular to the plane that contains 
a
 
and 
b
. The vector, 
n
, is a unit 
vector in the same direction as 
a
×
b
. (b) A p
ictorial re
presentation of the process by which each atom is 
projected onto the plane perpendicular to the unit vector determined by the sum of the mass weighted cross 
products of the position vectors and that contains the center of mass of the molecule. T
his is to e
liminate 
extra contributions to the moment of inertia to ensure an accurate angle, 

, is calculated which is needed to 

 
1
3
3
 
Figure
 
3.7.
 
Ground state absorptio
n spectrum of Co(acac)
3
 
in acetonitrile.
 

 
13
7
 
Figure
 
3.8.
 
(a) Transient absorption data from 460
-
610
 
nm for Co(acac)
3
 
in acetonitrile following 600
 
nm 
excitation into the 
1
T
1g
 
state. (b) Decay associated difference spectru
m associated with a glo
bal fit to two 
sequentially decaying exponential kinetic components. (c) Evolution associated difference spectra for the 
two kinetic components extracted from a global fit of the dataset.
 

 
13
8
 
Figure
 
3.9.
 
Additional f
ull spectrum transient 
absorption data on Co(acac)
3
 
in acetonitrile following 
600
 
nm excitation

 
1
39
 
Figure
 
3.10
.
 
Representative TA traces and FFT and LPSVD analyses of coherent oscillations of 
Co(acac)
3
 
in acetonitrile f
ollowing 600
 
nm excitat
ion into the 
1
T
1g
 
state. Kinetics were monitored at (a) 
620
 
nm, (c) 600
 
nm, and (e) 580
 
nm using a probe beam with a spectrum centered at 600
 
nm. Insets show 
coherent oscillations at early times. Corresponding FFT and LPSVD vibration
al spectra are shown in
 
(b), 

 
1
4
0
 
Figure
 
3.11.
 
(a) Transient absorption data on Co(acac)
3
 
measured at the peak of the high energy band of 
the transient spectrum at 520
 
nm following 600
 
nm excitation into th
e 
1
T
1g
 
state. The inset
 
shows the 
coherent oscillations at early time and an LPSVD fit. (b) Corresponding LPSVD and FFT vibrational 

 
1
4
1
 
Figure
 
3.12.
 
Plots of the calculated inner
-
sphere reorganization energy as a
 
function of norm
al mode 
frequency for pertinent electronic transitions of Co(acac)
3
. (a) Plot of the reorganization energy for the 
3
T
1g
 

1
A
1g
 
transition as a function of the frequencies of the vibrational modes of the 
3
T
1g
 
electronic excited 
state. (b) P
lot of the reorganization energy for the 
5
T
2g
 

1
A
1g
 
transition as a function of the frequencies of 
the vibrational modes of the 
5
T
2g
 
el
ectronic excited state. Reorganization energies were calculated from the 
nuclear displacements of the Duschinsky displac
ement vectors for the corresponding electronic transitions.
 

 
1
4
3
 
 
xviii
 
Figure
 
3.13.
 
Nuclear deformati
ons associated with the (a) 162 cm
-
1
, (b) 440 cm
-
1
, and (c) 466 cm
-
1
 
experimentally observed 
vibrational modes of Co(aca
c)
3

.
 
1
4
5
 
Figure
 
3.14. 
Possible assignments for the 263
 
cm
-
1
 
experimentally observed coherent vibrational mode of 
Co(ac
ac)
3
 
along with the corresponding inner
-
sphere reorganization energies for each mode along the 
3
T
1g
 

1
A
1g
 
reaction coor

 
1
4
6
 
Figure
 
3.15.
 
Ground state absorption spectrum of Co(3
-
Iac)
3
 
in t

 
1
5
3
 
Figure
 
3.16.
 
Difference spectrum of Co(3
-
Iac)
3
 
in toluene before and after significant light exposure
...
 
1
5
3
 
Figure
 
3.17.
 
Full spectrum TA data on Co(3
-
Iac)
3
 
in toluene following 600
 
nm excitation into the 
1
T
1g
 
state. 
The region 
from ~570
-
640
 
nm was omitted due to contamination from pump scatter


 
1
5
4
 
Figure
 
3.18.
 
Representative single wavelength TA traces and FFT and LPSVD analyses of coherent 
oscillations of Co(3
-
Iac)
3
 
in toluene following 600
 
nm excitation into the 
1
T
1
g
 
state. Kinetics were 
monitored at (a) 620
 
nm, (c) 600
 
nm, and (e) 580
 
n
m using a probe beam with a spectrum centered at 
600
 
nm. Insets show coherent oscillations at early times. Corresponding FFT and LPSVD vibrational 
spectra are shown in (b), (d), and (
f). LPSVD contributions are omitted from the 580
 
nm probe data due to 
lac

 
1
5
5
 
Figure
 
3.19.
 
DFT calculated symmetric breathing mode of Co(3
-
Iac)
3
 
on (a) the 
1
A
1g
 
ground state, (b) the 
3
T
1g
 
lowest energ
y triplet excited state, and (c) the 
5
T
2g
 

 
1
5
6
 
Figure
 
3.20.
 
Ground state absorption spectrum of Co(TMHD)
3
 

 
15
7
 
Figure
 
3.21.
 
(a) Full spectrum TA data on Co(TMHD)
3
 
in DCM followi
ng 600
 
nm 
1
A
1g
 

1
T
1g
 
excitation. 
Anomalies around 600
 
nm are due to excess pump scatter. (b) DADS associated with a global fit of the data 


 
15
8
 
Figure
 
3.22.
 
(a) and (c) Single wavelength TA data on Co(
TMHD)
3
 
in DCM following 600
 
nm 
1
A
1g
 

1
T
1g
 

 
1
59
 
Figure
 
3.23.
 
Nuclear distortion associated with t
he coherent vibrational mode of Co(TMHD)
3

 
1
6
0
 
Figure
 
3.24.
 
(a
) Two
-
level potential energy surface diagram with displaced minima for use in wavepacket 
calculation. Diabatic and adiabatic representations are shown as dashed and solid lines respectivel
y. The 
initial gaussian wavepacket is placed at 0 Å on the upper sur
face and allowed to propagate. (b) Contour 
plot showing the sum of the amplitude of the corresponding wavepackets on both excited states in position 
space as a function of time. (c) Two
-
le
vel potential energy surface diagram with nested potentials for use 
in wavepacket simulations. The initial gaussian wavepacket is placed at 0 Å on the upper surface. (d) 
Contour plot showing the sum of the amplitude of the corresponding wavepackets on both
 
excited states. 
Despite continuously cycling population from one su
rface to the other, the wavepacket retains a well
-

 
1
6
2
 
Figure
 
3.A.1.
 
NMR spectrum of Co(acac)
3
 
in CDCl
3

 
1
6
5
 
Figure
 
3.A.2.
 
NMR spectrum of Co(3
-
Iac)
3
 


 
1
6
6
 
Figure
 
3.A.3.
 
NMR spectrum of Co(TMHD)
3
 
in CDCl
3

 
1
6
6
 
Figure
 
4.1.
 
Drawing of cage ligand (
L
) used in this study.
 

 
1
8
2
 
 
xix
 
Figure
 
4.2
. 
Single
-
crystal x
-
ray structures of the cations of 
(a) [Fe(L)](PF
6
)
2
 
(
1
)
, 
(b) [FeCu
2
(L)](PF
6
)
4
 
(
2
), 
and (c) [ZnCu
2
(L)](PF
6
)
4
 
(
3
). The hydrogen atoms have been removed for clarity. For metal ions: red
 
=
 
Fe, 
green
 
=
 
Cu, light blue
 
=
 
Zn.
 


 
1
8
3
 
Figure
 
4.3.
 
Ground state electronic absorption spectra of [
Fe(L)](PF
6
)
2
 
(
1
), [FeCu
2
(L)](PF
6
)
4
 
(
2
), and 
[ZnCu
2
(L)](PF
6
)
4
 
(
3
). The spectra were acquired at room temperature in CH
3

 
1
8
4
 
Figure
 
4.4
.
 
Femtosecond time
-
resolved absorption data of [Fe(L)](PF
6
)
2
 
(
1
) acquired in CH
3
CN solution.  
(a) Full s
pectrum data collected over the first few ps following MLCT excitation at 600 nm. The anomalous 
features near 600
 
nm are due to scattered light from the 
excitation source. (b) Single wavelength kinetics 
trace monitored at 608
 
nm with LPSVD fit. (c) Resulti
ng vibrational spectrum corresponding to the FFT of 
the residuals of an exponential fit (blue) and the oscillations extracted through LPSVD fitting (red)
. (d) 
Single wavelength kinetics trace monitored at 658
 
nm with LPSVD fit. (e) Vibrational spectrum of 
the 
kinetic data collected at 658
 
nm.
 

 
1
8
5
 
Figure 4.5.
 
Spectroelectrochemical data on [Fe(L)]
2+
 
(
1
) collected by Sara Adelman

 
1
8
6
 
Figure
 
4.6. 
DFT calculated vector displacement diagrams of several of the coherently exc
ited vibrational 
modes of [Fe(L)](PF
6
)
2
 
(
1

 
1
8
8
 
Figure
 
4.7.
 
Overlay of the Fe
-
N
6
 
region of [Fe(L)](PF
6
)
2
 
(
1
) and [FeCu
2
(
L)](PF
6
)
4
 
(
2
) crystal structures. 
The complex drawn with red bonds is complex 
1
 
and the complex with pu
rple bonds is complex 
2

 
1
89
 
Figure
 
4.8.
 
Representative TA data on [FeCu
2
(L)](PF
6
)
2
 
(
2
) in acetonitrile following 600
 
nm excitation 
while 
monitoring at (a) 700 nm and (c) 650
 
nm (blue) with fit to equation 4.1 (red) and inset showing 
LPSVD fit to equati
on 4.2. Corresponding vibrational spectra (b) and (d) resulting from LPSVD fit (red) 
and FFT of the residuals of the exponential fit (blue). 
The appearance of an additional peak in the low 
frequency region in the LPSVD spectrum in part d can be attributed 
to error in exponential fitting in the 

 
1
9
0
 
Fig
ure
 
4.9.
 
Representative TA data on [FeCu
2
(L)](PF
6
)
4
 
(
2
) in (a) butyronitrile and (b) acetone following 
600
 
nm excit
ation and monitoring at 700
 
nm with insets showing raw data (blue) and LPSVD fits (red) over 
the first few picoseconds. (c) and (d) show the 
vibrational spectra resulting from the FFT and LPSVD 
analyses of the data in butyronitrile and acetone respectively

 
1
9
1
 
Figure
 
4.10.
 
DFT calculated vector displacement diagrams of the coherent modes of 
2
 
observed in this 


 
1
9
2
 
Figure
 
4.11.
 
Femtosecond time
-
resolved absorption spectra of [FeCu
2
(L)](PF
6
)
4
 
(
2
)
 
following 600
 
nm 

 
1
9
3
 
Figure
 
4.12.
 
Representative datasets of [FeCu
2
(L)](Bar
F
)
4
 
in (a) acetonitrile, (b) acetone, and (c) 
butyronitrile following 600
 
nm excitation while monitored at 700
 
nm. 

 
1
9
4
 
Figure
 
4.13.
 
Plots of the calculated inner
-
sphere reorganization energy as a function of normal mode 
frequency for the 
5
T
2
g
 

1
A
1g
 
transitions of (a) [Fe(L)]
2+
 
and (b) [FeCu
2
(L)]
4+
. The normal modes used are 
inherent to the 
5
T
2g
 
excited state. Reorganization energies were calculated from the nuclear displacements 
of the Duschinsky displacement vectors for the corresponding ele

 
1
9
6
 
Figure
 
4.A
.
1.
 
(a) TA data of [FeCu
2
(L)](PF
6
)
4
 
(
2
) in acetonitrile while 
exciting at 600
 
nm and monitoring 
at 396
 
nm (blue) with LPSVD fit to equation 4.2 (red). (b) LPSVD vibrational spectrum (red) with 
corresponding FFT of
 

 
199
 
 
xx
 
Figure
 
4.B
.
1.
 
Calculated reorganizatio
n energy for the 
5
T
2g
 

1
A
1g
 
transition of [Fe(bpy)
3
]
2+

...

 
20
0
 
Figure
 
4.C
.
1.
 
DFT calculated vector displacement diagram of the 123
 
cm
-
1
 
vibrational mode of [Fe(L)]
2+
.
 

 
20
1
 
Figure
 
5.1.
 
Qualitative potential energy surfaces for six
-
coordinate (a) iron(II) and (b) ruthenium(II) 

 
21
1
 
Figure
 
5.2.
 
Radial distribution function for (a) the n = 3 orbitals of iron(II) and (b) the n = 4 orbitals of
 


 
21
2
 
Figure
 
5.3
.
 
Drawing of the three compounds investigated in this study.
 

 
2
1
3
 
Figure
 
5.4.
 
Ground state absorption spectra of [Fe(phen)
3
](BF
4
)
2
 
(
1
), [Fe(phen)
2
(CN)
2
] (
2
), and 
[Fe(phen)
2
(C
4
H
10
N
4
)](PF
6
)
2
 
(
3
), collected by Karl Nielsen of t

 
2
1
6
 
Figure
 
5.5.
 
(a) Full spectrum TA data on [Fe(phen)
3
](BF
4
)
2
 
(
1
) in methanol following 570
 
nm excitation. 
Irregularities at 570
 
nm can be attributed to uncorrected
 
pump scatter. (b) First (black) and second (blue) 
right eigen
vectors from SVD of full spectrum dataset with 970
 
ps single exponential fits (red). (c) EADS 

 
2
1
7
 
Figure
 
5.6.
 
Sing
le wavelength kinetics traces collected on [Fe(phen)
3
](BF
4
)
2
 
(
1
) in methanol at 520
 
nm 
following 520
 
nm excitation. This particular dataset was fit to a 1250 
±
 
150
 
ps time constant with the error 


 
2
1
8
 
Figure
 
5.7.
 
Single wavelength kinetics trace c
ollected on [Fe(phen)
3
](BF
4
)
2
 
(
1
) in methanol at 610
 
nm 
following 520
 
nm excitation. Inset includes a fast Fourier transform of the residuals of an exponential fit to 
the population dynamics (bl
ue) overlaid with a vibrational spectrum produced via LPSVD of
 
the TA data 

 
2
1
8
 
Figure
 
5.8.
 
(a) Full spectrum TA data on [Fe(phen)
2
(CN)
2
] (
2
) in methanol following 570
 
nm excitation. 
(b) First (black), second (green), and third (blue) right eigen vectors from SVD of 
2
 
and
 
a biexponential fit 
(red) of the first right eigenvector to 6.6 and 315
 
ps kinetic components. (c) Simulated excited state 
populations using the time constants acquired from the right eigenvector fits and assuming sequential decay 
dynamics as shown in equ
ation
 
5.3. (d) EADS calculated for the two kinetic components using equation 
5.2. The red (F1) and blue (F2) spectra corresp
ond to the fast and slowly decaying species respectively.
 

 
2
19
 
Figure
 
5.9.
 
(a) Representat
ive TA data of [Fe(phen)
2
(CN
)
2
] (
2
) in methanol following 590
 
nm excitation 
and monitored at 550
 
nm in the center of the gr
ound state bleach. (b) Representative TA data on 
2
 
following 
600
 
nm excitation and monitored at 590
 
nm, the red edge of the ground s
tate bleach. The two data sets show 

 
2
2
0
 
Figure
 
5.10.
 
Single 
wavelength kinetics trace of [Fe(phen)
2
(CN)
2
] (
2
) monitored at 620
 
nm following 
550
 
nm excitation. Oscillatory features near the coh
erent artifact can be attributed to vibrational coherences.
 

 
2
2
1
 
Figure
 
5.11.
 
(a) 
Chirp corrected full spectrum TA data on [Fe(phen)
2
(C
4
H
10
N
4
)](PF
6
)
2
 
(
3
) following 
600
 
nm excitation in methanol. The region between 
590 and 610
 
nm has been removed and interpolated to 
remove the contribution from excess pump scatter. (b) First four right e
igenvectors from SVD of TA data 
of compound 
3
 
with triexponential fits (red) to 0.15, 1.2, and 5.3
 
ps time constants. An inset is in
cluded to 
highlight the dynamics over the first 10
 
ps. (c) Simulated excited state populations using the time constants 
 
xxi
 
acqu
ired from the right eigenvector fits and assuming sequential decay dynamics as shown in equation
 
5.4. 
(d) EADS calculated from the t
ime constants measured from SVD analysis assuming sequential decay 

 
2
2
4
 
Figur
e
 
5.12.
 
Single wavelength kinetics traces on [Fe(phen)
2
(C
4
H
10
N
4
)](PF
6
)
2
 
(
3
) in methanol monitored 
at (a) 650
 
nm abd (b) 550
 
nm follo
wing 600
 
nm excitation. Insets show first 5
 
ps and highlight observed 


 
2
2
5
 
Figure
 
5.B
.
1.
 
Simulated MLCT spectrum of (a) [Fe(phen)
3
]
2+
 
(
1
), (b) [Fe(phen)
2
(CN)
2
] (
2
), and (c) 
[Fe(phen)
2
(C
4
H
10
N
4
)]
2+
 
(
3
) i

 
2
3
3
 
Figu
re
 
5.C
.
1.
 
(a) Chirp corrected global fit using OPTIMUS fitting software on full spectrum data of 
[Fe(phen)
3
]
2+
 
(
1
) to a single 940
 
ps exponential kinetic component. (b) DADS fr
om global fit of 
1
 
full 
spectrum data. Since there was only a single kinetic component, t

 

 
2
3
4
 
Figure
 
5.C
.
2.
 
(a) Chirp corrected global fit using OPTIMUS fitting software on full
 
spectrum data of 
[Fe(phen)
2
(CN)
2
] (
2
) to two 5.5
 
ps and 290
 
ps exponential kinetic compo
nents. (b) DADS from the global 

 
2
3
5
 
Figure
 
5.C
.
3.
 
(a) Chirp corrected global fit us
ing OPTIMUS fitting software on full spectrum data of 
[Fe(phen)
2
(C
4
H
10
N
4
)](PF
6
)
2
 
(
3
) to 
three 0.12, 0.66, and 4.7
 
ps exponential kinetic components. (b) DADS 

 
2
3
6
 
F
igure
 
5.D
.
1.
 
(a) Early time TA data (blue) on [Fe(phen)
3
](BF
4
)
2
 
(
1
) in methanol monitore
d at 610
 
nm 
following 520
 
nm excitation. An exponential fit (red) and a LPSVD fit (black) are superimposed on the 
data. (b) FFT of the residual of the exponential fit o
f the 610
 
nm TA data (blue) and LPSVD vibrational 
spectrum (red). (c) Early time TA data
 
(blue) of compound (
1
) monitored in the bleach region at 520
 
nm 
with exponential (red) and LPSVD (black) fits. (d) Corresponding FFT (blue) and LPSVD vibrational 
spect

 
2
3
7
 
Figure
 
5.D
.
2
.
 
(a) Early time TA data (bl
ue) on [Fe(phen)
2
(C
4
H
10
N
4
)
](PF
6
)
2
 
(
3
) in methanol monitored at 
550
 
nm following 600
 
nm excitation. Exponential fit (red) and LPSVD fit (black) are superimposed on the 
d
ataset. (b) Corresponding FFT of the residual of the exponential fit (blue) and LPSVD vi
brational spectrum 
(red). (c) Analogous data monitoring at 650 nm and (d) corresponding FFT (blue) and LPSVD (red) 
vibrational spectra.
 

..
 
2
39
 
Figure
 
5.
D
.
3.
 
(a) Early time TA data (blue) on [Fe(phen)
2
(C
4
H
10
N
4
)
](PF
6
)
2
 
(
3
) 
in acetonitrile monitored at 
550
 
nm following 600
 
nm excitation. Exponential fit(red) and LPSVD fit (black) are superimposed on the 
dataset. (b) Corresponding FFT of the r
esidual of the exponential fit (blue) and LPSVD vibrational spectrum 
(red). (c) Analo
gous TA data while monitoring at 650
 
nm and (d) corresponding FFT (blue) and LPSVD 
(red) vibrational spectra.
 

 
2
4
0
 
Figure
 
6.1.
 
Structures o

 
2
4
8
 
 
 
 
xxii
 
KEY TO SYMBOLS AND
 
ABBREVIATIONS
 
 
acac: 
A
cetylacetonate
 

 
A
exp
: 
A
mplitude of exponential decay 
 
A

: 
A
mplitude of oscillatory component
 

A: Change in abso
rption
 
bISC: 
B
ack
-
intersystem crossing
 

-
B
ipyridine
 
Calc.: 
C
alculated
 
CDCl
3
: Deuterated chloroform
 
c
gs
: 
G
round state concentration
 
DADS: Decay associated difference spectrum
 
d
axial
: 
A
xial metal ligand bond length
 
d
eq
: 
E
quatorial metal ligand bond l
ength
 
DFT: 
D
ensity functional theory
 
DMHD: 2,6
-
D
imethyl
-
3,5
-
heptanedione
 
DSSC: Dye sensitized solar cell
 
E
a
: 
A
ctivation energy
 
EADS: 
E
volution associated difference spectrum
 
E
n
: Energy of n
th
 
vibrational level
 
en: 
E
thylenediamine
 
ESA: 
E
xcited state absorpt
ion.
 
Exp.: 
E
xperimental
 
f
: Oscillator strength
 
 
xxiii
 
F
FT: 
F
ast Fourier transform
 
GSB: 
G
round state bleach
 
HD: 3,5
-
H
eptanedione
 
I: 
M
oment of inertia
 
IC: 
I
nternal conversion
 
IR: Infrared
 
IRF: 
I
nstrument response function
 
ISC: 
I
ntersystem crossing
 
ISRS: 
I
mpulsively
 
stimulated Raman scattering
 
IVR: 
Intramolecular vibrational redistribution
 
K
: Duschinsky displacement vector
 
k
: Rate constant
 
k
b
: Boltzmann constant
 
k
bISC
: 
B
ack
-
intersystem crossing rate constant
 
k
IC
: 
I
nternal conversion rate constant
 
k
p
: 
P
hosphorescence 
rate constant
 
k
VC
: 
V
ibrational cooling rate constant
 
k
VR
: 
V
ibrational relaxation rate constant
 
L: Cage ligand
 
LMCT: Ligand to metal charge transfer
 
LPSVD: 
L
inear predictive singular value decomposition
 
L
r
: 
M
atrix that diagonalizes the mass weighted
 
Cartesi
an Hessian of the re
actant geometry
 



: Angular momentum
 
M
: 
D
iagonal mass matrix
 
m
i
: 
M
ass of ith atom,
 
MLCT: Metal to ligand charge transfer
 
 
xxiv
 
[MLCT
a
]: Initial MLCT excited state concentration
 
[MLCT
b
cool
]: 
V
ibrationally cool MLCT state concentration
 
[MLCT
b
hot
]: 
V
ibrationally excited MLCT state concentration
 
R
nl
(r)
: Probability density function
 








: Cartesian coordinate of ith reactant atom
 









: Cartesian coordinate of ith product atom
 

R
: Cartesian displacement vector
 
SE: 
S
timulated emissi
on
 
SOC: 
Spin
-
orbit coupling
 
SVD: 
S
ingular value decomposition
 
T
(
t
): Time dependent effective temperature
 
T
: 
R
otation matrix
 
t
: Time
 
T
1,elec
: Electronic population relaxation
 
T
1,vib
: 
V
ibrational population relaxation
 
T
1
: 
P
opulation relaxation time constant
 
T
2
*
: 
P
ure dephasing
 
T
2
: 
H
omogeneous dephasing time
 
TA: 
T
ransient absorption
 
TMHD: 2,2,6,6
-
T
etramethyl
-
3,5
-
heptanedione
 
T
x
: 
I
nitial effective temperature
 
VC: Vibrational cooling
 
VR: Vibrational relaxation
 

: 
F
luctuation magnitude
 


: 
F
ull width at half maximum of
 
Raman peak.
 
 
xxv
 


: 
D
ephasing time
 

: 
E
ffective fluctuation frequency
 

: Extinction coefficient
 

: Oscillation phase
 

i
: Fraction of excited state population 
i
 

: Reorganization energy
 

: Wavelength
 

I
RF
: Standard deviation of gaussian IRF
 



A
ngle
 

: 
Time constant
 

: Angular frequency
 

: Wavefunction
 
 
1
 
Chapter 1: Coherent Vibrational Dynamics of Inorganic Molecular Chromophores
 
1.1
.
 
Introduction
 
 
This dissertation discusses the ultrafast 
photoinduced ex
cited state dynamics in several different 
first
-
row transition metal molecular systems ranging from six coordinate chromium(III) and cobalt(III) 
betadiketonate complexes to iron(II) polypyridyl and carbene compounds. Much of this dissertatio
n places 
a part
icular emphasis on characterizing the coherent vibrational dynamics occurring in each chromophore 
for the goal of using information from coherences to not only glean important dynamical information for 
excited state characterization but also
 
to suggest way
s in which synthetic modifications can alter important 
vibrational
-
electronic couplings to tune photophysical properties. In this work, the ultrashort timescales 
characteristic of impulsive vibrational dynamics as well as many excited state 
transformations
 
necessitates 
experimental techniques with similarly short time resolution. As such, much of the work herein will utilize 
ultrafast transient absorption (TA) spectroscopy, which uses pulses of light on the order of a millionth of a 
billionth
 
of a second to
 
both induce and monitor excited state processes. In what follows, we will briefly 
discuss the basics of transition metal photophysical properties, transient absorption spectroscopy, and 
finally coherences and their manifestations in TA sign
als.
 
1.2
.
 
Trans
ition Metal Photophysical Properties
 
 
Transition metal chromophores have numerous types of electronic transitions available to them 
including 
d
-
d
 
and various types of charge transfer transitions. An excellent discussion of these topics can 
b
e seen in Figgi

1
 
and will briefly be summarized here. To start, the intensity or 
allowedness of electronic transitions depends on the transition moment integral, 
Q
 
(equation 1.1), in w
hich 
 
 
 
1.1
 

1
 
and 

2
 
are the wavefunctions for the initial and final states and 

 
is the dipole moment operator. A 
wavefunction itself is built of a product of multiple component wavefunctions as shown in equation 1.2. 
 
 
 
1.2
 
 
2
 
This leads to s
everal factors which can affect the magnitude of 
Q
. First, if there is a change in the spin 
moment between the final and initial wavefunctions (i.e. 

S
 

 
0), then 
Q
 
= 0. This spin selection rule says 
that an electronic dipole induced transition cannot indu
ce a change in spin state. These forbidden transit
ions 
can become somewhat allowed by spin
-
orbit coupling interactions that intermix an amount of different spin 
character to the final or initial states involved in the transition. Nevertheless, this spin
-
se
lection rule results 
in substantial 100
-
1000
-
fold 
reductions in intensities relative to spin
-
allowed transitions.
1
 
Despite their low 
intensity these transitions are an important aspect of spectroscopy in t
ransition metal complexes, and have 
even been util
ized in the ruby laser.
2
 
 
Another important selection ru
le in the spectroscopy of transition metal compounds arises from 
orbital symmetry. Because the transition moment integral, 
Q
, is performed over symmetric boundaries, this 
implies that 
Q
 
= 0 if the product 

1
 
× 

 

2
 
is an odd function. Thus, since the dipole moment operator, 

, is antisymmetric (ungerade) with respect to the inversion center of a system and is therefore an odd 
function, it follows that 

1
 
and 

2
 
must have opposite parity for the overall
 

1
 
× 

 

2
 
p
roduct to be an 
even function (gerade) and, by extension, to have a nonzero transition moment integral. Thus, for an 
electronic transition between two 
d
 
orbitals (which are even functions) the integrand of the transition 
moment integral is an odd
 
function 
and 
Q
 
= 0. These parity
-
 
or Laporte
-
forbidden transitions are quite 
weak, yet nevertheless give many transition metal compounds their color. The parity
-
selection rule can be 
overcome when a molecule lacks an inversion center. For example, in comp
ounds of 
T
d
 
symmetry, several 
of the 
d
 
and 
p
 
orbitals have the same irreducible representation, which allows for mixing and a relaxation 
of the Laporte
-
selection rule. Alternatively, since the wavefunction is a product of many component 
wavefunctions, a 
d
-
d
 
transiti
on may couple with an antisymmetric vibration to become vibronically 
allowed.
 
 
The last important electronic transition common to transition metal compounds are charge transfer 
transitions. The first type of charge transfer involves the transfer 
of an elec
tron from a molecular orbital of 
high ligand orbital character to an orbital of high metal character. This ligand to metal charge transfer 
 
3
 
(LMCT) is common in systems with high energy, filled ligand bonding or nonbonding orbitals and 
relatively 
low energy unfilled orbitals localized on the metal center. Alternatively, the reverse can occur 
where an electron in a largely metal based molecular 
orbital is transferred to a largely ligand centered 
orbital. These metal to ligand charge transfer (MLCT) 
occur in low valent metal complexes with high
-
lying, 
populated metal centered orbitals and low
-
lying empty ligand
-
based orbitals. Because both charge 
transfer 
type transitions involve the gerade 
d
 
orbitals of the metal and the ungerade 
p
 
orbitals of the li
gands, they 
are fully allowed and hence much more intense than 
d
-
d
 
transitions. Likewise, ligand centered transitions 
are often between states of opposite parity, resulting in highly intense electronic transitions. A summary of 
the typical extinction coeff
icients and oscillator strengths for transitions commonly seen in 
transition metal 
complexes can be seen in Table
 
1.1.
1
 
Transition Type
 


(M
-
1
cm
-
1
)

f
 
d
-
d
, spin forbidden
 
0.1
-
1
 
10
-
7
-
10
-
6
 
d
-
d
, spin allowed
 
10
-
100
 
10
-
5
-
10
-
4
 
Charge Transfer
 
1000
-
10,000
 
10
-
3
-
10
-
2
 
 
Table
 
1.1.
 
Summary of the extinction 
coefficients, 

, and oscillator strengths, 
f
, typically 
observed for various electronic transitions of transition metal complexes.
 
Figure
 
1.1.
 
Jablonski diagram depicting various excited state processes including 
vibrational relaxation (VR), internal conv
ersion (IC), intersystem crossing (ISC), 
fluorescence (FL), and phosphorescence (PH). Radiative transitions are 
shown as solid lines 
and nonradiative transitions are shown as dashed lines.
 
 
4
 
 
Following absorption of a photon, a molecule can undergo several types of photophys
ical pr
ocesses 
to dissipate energy. A summary of these processes is shown in the Jablonski diagram in Figure
 
1.1. After 
the initial photoexcitation (shown in purple in Figure
 
1.1) a molecule is promoted from the ground 
electronic state to a higher energy e
xcited 
state which is often of the same spin multiplicity as the ground 
state. In some cases, this also produces a vibrationally excited molecule which can dissipate this vibrational 
energy in two ways. The molecule can undergo vibrational cooling (VC) in 
which e
xcess vibrational energy 
is dissipated towards the solvent bath, or intramolecular vibrational redistribution (IVR) could occur in 
which vibrational energy is spread from one vibrational mode to another. These are often referred to 
collectively as v
ibratio
nal relaxation (VR) which is shown in red in Figure
 
1.1. Additionally, the excited 
molecule can convert to another state of the same spin multiplicity either radiatively in the case of 
fluorescence (FL) or nonradiatively via internal conversion (IC)
. Durin
g FL, excess energy is dissipated as 
electromagnetic radiation, whereas during IC excess energy is either converted to vibrational energy on the 
newly formed electronic state or dissipated as heat towards the solvent bath. Instead of converting to a
 
new 
el
ectronic state of similar spin, the molecule can also undergo a nonradiative transition to an excited state 
of different spin. This is referred to as intersystem crossing (ISC). Just like in the case of IC, ISC can 
dissipate thermal energy outwards 
towards
 
the solvent bath or inwards to make a vibrationally excited state 
which can again undergo VR. Finally, the molecule can nonradiatively ISC to the ground state of different 
multiplicity or radiatively phosphoresce (PH). All these processes can be la
belled 
collectively under the 
term photophysics.  In addition to photophysical transformations, electromagnetic radiation can result in 
the formation of a new chemical species. This is referred to as photochemistry. The work performed in this 
dissertation 
will se
ek to avoid photochemical reactions to make analysis of the photophysical properties of 
each system easier to interpret. 
 
1.3
.
 
Transient Absorption Spectroscopy
 
 
Due to the short timescales of many excited state processes, experiments need exception
ally sh
ort 
time resolution to track photoinduced dynamics in many systems. The work in this dissertation will employ 
pump
-
probe transient absorption (TA) spectroscopy. The basic premise of this experiment is to irradiate 
 
5
 
the sample with an ultrashort laser
 
pulse 
(pump) to create an excited state population in a small fraction of 
the molecules and subsequently, after a set time delay, interrogate the absorptive properties of the sample 
with a second weaker laser pulse (probe) by measuring its transmittance t
hrough 
the sample. By 
systematically varying the time delay between pump and probe pulses one can track the absorptive changes 
in the sample and in effect the excited state dynamics. Because the fraction of molecules that have been 
excited is generally qui
te smal
l, the difference in absorbance between the sample that has been irradiated 
with the pump pulse and the sample that has not been pumped is measured. This can be described using 
equation 1.3 where 
I
(

,
t
)
pumped
 
is the time dependent transmittance of t
he pump
ed sample and 
I
(

)
unpumped
 
is
 
 
 
1.3
 
the transmittance before the pump pulse arrives. The magnitude of the signal is related to the difference 
between the extinction coefficients of the excited state, 

i
, and ground state, 

GS
, and also relates to the 
pathlength, 
b
, the initial concentration 
of the chromophore, 
c
GS
, and the time dependent fraction of the 
i
th 
excited state population, 

i
(
t
).
3,4
 
It is clear from equation
 
1.3 that large amplitude transient signals arise 
when the extinction coefficients or cross
-
sections are very different bet
ween the ground and excited state.
 
 
Figure
 
1.2 shows an energy level diagram with possible signal contributions in a p
ump
-
probe 
transient absorption experiment. First, since the pump pulse reduces the number of absorbing molecules in 
Figure
 
1.2.
 
Energy level diagra
m showing possible signals in a pum
p
-
probe transient absorption 
experiment including ground state bleach (GSB), excited state absorption (ESA) and stimulated 
emission (SE).
 
 
6
 
the ground electronic
 
state, then a probe pulse with frequencies overlapping the ground state absorption 
spectrum will experience a decrease
 
in absorption. This is referred to as ground state bleach (GSB). Second, 
since a pump photon can create an excited state population, a p
robe photon can stimulate the molecule to 
emit a photon and return to the ground state. This stimulated emission (SE) results in a higher perceived 
transmittance and thus a more negative 

A. Finally, the probe pulse can induce an absorption from the 
popul
ated excited state to another excited sta
te. This excited state absorption (ESA) results in a decreased 
Wavelength (nm)
 
Time (ps)
 
a)
 
b)
 
c)
 
Figure
 
1.3.
 
Illustration of the different ways pump
-
probe TA data can be view
ed. In the 2
-
dimensional data set shown in (a) the entire transient spectrum is collected at each time delay. A 
vertical slice of this data (black) contains the differential absorption at a 
single wavelength across a 
series of time delays. This is illustra
ted in the black boxed kinetics trace of (b). A horizontal slice 
shown in red represents the transient spectrum at a specific pump
-
probe delay time and is shown in 
the red boxed spectrum (c)
.
 
 
7
 
transmittance of the probe pulse and hence an increase in 

A. Each of these types of signal may or may not 
overlap spectrally. Thus, the total transient
 
signal at a given wavelength may have co
ntributions from 
multiple processes.
 
 
Transient absorption data can be viewed and collected in several ways as illustrated in Figure
 
1.3. 
In some cases a spectrally broad probe beam (often a white light continuum ge
nerated via self
-
phase 
modulation) is dir
ected into a spectrometer following the sample to be dispersed across a multichannel 
detector so that the frequency resolved differential absorption spectrum can be measured across the entire 
probe spectrum at once.
 
Collection of this data across multiple 
pump
-
probe delay times results in a dataset 
like the one shown in Figure
 
1.3a. A horizontal slice of this dataset shows the transient difference spectrum 
at a single pump
-
probe delay time as shown in the red boxed s
pectrum of Figure
 
1.3c. A vertical slice 
of 
this 2
-
dimensional dataset shows the change in absorption at a single wavelength as a function of time as 
shown in the black boxed dataset in Figure
 
1.3b. This data style may also be collected using a 
monochromat
or to select out a single slice of the en
tire probe spectrum to measure or by using a spectrally 
narrower probe beam. 
 
1.4
.
 
Coherences and Impulsive Vibrational Spectroscopy
 
 
We will next discuss the concept of coherence and impulsive vibrational spectros
copy. Before 
diving right in, howe
ver, we can first utilize an analogous concept: the idea of resonance from organic 
chemistry.
5
 
Benzene for instance has two main resonance structures denoted by 

1
 
a
nd 

2
 
in Figure
 
1.4. 






a


 
+
 
b




Figure
 
1.4.
 
Resonance as an analogy to cohe
rence. The ground state wavefunction of 
benzene is a linear combination of multiple component wavefunctions.
 
 
8
 
Neither of th
ese two Kekul
é
 
structures correspond to the true wavefunction of benzene. Furthermore, the 
molecule is not flickering back and forth between them. The true wavefunction is a superposition or a linear 
combination of each Ke
kul
é
 
wavefunction: 
a

1
 
+ 
b

2
. Wave
functions that are comprised of a linear 
combination of multiple component wavefunctions are commonly observed in transient absorption 
experiments. Ultrashort laser pulses with bandwidths that are broad enough to span mult
iple vibrational 
levels on an elec
tronic excited state can excite several vibrational levels simultaneously. The resulting 
wavefunction is a linear combination of each of the constituent wavefunctions as shown in equation 1.4 
where 

(
R
,
t
) is the net
 
 
 
1.4
 
wavefunction, 
c
n
 
is the population coefficient, 

n
(
R
) is the wavefunction in of the 
n
th vibrational level, and 
E
n
 
is the energy of the 
n
th v
ibrational level.
6
 
A p
ictorial representation of this can be seen in Figure
 
1.5, in 
which the net wavefunction,
 
or wavepacket (blue), is the sum of each of the first five vibrational 
wavefunctions resulting in a much more localized amplitude in position space due to constructiv
e and 
destructive interferences. The oscillatory nature of each wavefunction, originating
 
from the exponential 
term in equation 1.4, shifts the interference pattern causing the wavepacket to migrate back and forth across 
the potential surface at the freque
ncy of the activated vibration.
 




















tot



Figure
 
1.5.
 
Generation of a coherent superposition state, or wavepacket, as a linear combination 
of oscillating w
avefunctions.
 
 
9
 
 
Vibrational wavepackets, or coherences, can also be gen
erated on the ground electronic state. To
 
explain how, it is helpful to view the pump
-
probe TA experiment in terms of a three
-
field interaction with 
the sample. Excited state coherences are generated when the first two field interactions of the ultrashort 
pump pulse create an excited state popula
tion wavepacket which can oscillate back and forth across the 
excited state potential energy surface. Ground state vibrational coherences are generated when the first field 
interaction of the pump pulse generates an
 
excited state polarization wavepacket an
d then, after some time 
delay, the second field interaction projects the wavepacket onto the ground state, displaced from its initial 
position.
7
 
This process is known as impulsively stimulated Raman scattering (ISRS).
8
 
The oscillation 
amplitude of these ground state wavepack
ets strongly depends on the ability of th
e excited state polarization 
wavepacket to propagate long enough to become significantly displaced from the hole on the ground state. 
When the pump pulse is too short this condition is not met and ground state vibra
tional features are not very 
significant.
7
 
 
The curvature of a potential surface 
has an important role in the propagation 
of a wavepacket. To 
illustrate this a figure from the book 
Coherent Vibrational Dynamics
7
 
was reproduced by simulating a 
Figure
 
1.6.
 
Potential energy surface diagram depicting the process of forming a 
wavepacket on the ground electronic state. 
 
 
10
 
wavepacket on several types of  potential energy surfaces (Figure
 
1.7) using the split operator method
9
 
in a 
homebuilt MATLAB prog
ram.
 
When the excited and ground state potential surfaces have the same 
curvature the wavepacket maintains its shape as it slides back and forth across the excited state surface 
(Figure
 
1.7a). In contrast, when the 
curvature is different between the final 
and initial states, the wavepacket 
expands or contracts according the forces experienced on the new potential surface (Figure
 
1.7b). For 
example, when the excited state has a smaller force constant, the restoring fo
rces near the now flatter 
equilibrium geo
metry are insufficient to hold the wavepacket together so it spreads out. Then, at the turning 
point, the combination of positive momentum and a negative restoring force recompress the wavepacket 
a)
 
b)
 
d)
 
c)
 
Figure
 
1.7.
 
Simulated wavepacket propagation on an excited state potential energy surface which 
(a) 
has the same force constant as the ground state, (b) has a different force constant than the 
ground state, (c) is anharmonic, and (d) is dissociative. The initial proj
ected wavepacket is shown 
in red. Wavepackets that have propagated on the excited state s
urface are shown in dashed lines.
 
 
11
 
back to its original shape. In the anharmonic potential (Fi
gure
 
1.7c), the different vibrational spacing at 
higher and lower energy portions of the potential surface cause nonuniform oscilla
tion frequencies which 

.
7
 
Finally, a dissociative potential 
surface (Figure
 
1.7d) results in the migration of the wavepacket away from t
he Franck Condon position 
while spreading out.
 
 
The propagation of wavepackets has important photophysical ramifications, and can e
ven, 
according to the time
-
dependent theory of spectroscopy, determine the absorption and emission spectrum 
Figure
 
1.8.
 
(a) Effect of coherent vibrational motion on the entire transient absorption spectrum. 
Band oscillates according to the frequencies of the activated vibrations. (b) Manifestation of coherent 
vibrational motion
 
on single wavelength TA measuremen
ts. Oscillatory features are superimposed 
upon the population dynamics. (c) Fourier transformation (blue) of oscillatory components of TA 
data results in a vibrational spectrum of the molecule. LPSVD fit of TA data provid
es not only 
frequencies of coherent
 
vibrational modes but also their dephasing times.
 
a)
 
b
)
 
c
)
 
 
12
 
of a chromophore.
10
 
In a TA experiment, the migration of a wavepacket along a potential energy surface 
causes periodic modifications to the Franck
-
Condon factors for absorption or emission. This manifests as 
oscillatory components in th
e TA data that occur at the same freque
ncy as the activated vibrations. This can 
be seen as a shifting of the transient spectrum to higher and lower wavelengths
11
 
as s
hown in Figure
 
1.8a. 
The amplitudes of 
these oscillations are dependent on the curvature of the transient absorption profile. 
Probing at the apex of the absorption feature where the derivative, 
d

A/d



is zero tends to wash out 
oscillations, whereas probin
g at inflection points where the slope 
is largest tends to show the strongest 
oscillations.
12,13
 
Ground state vibrational coherences of the solvent bath can als
o
 
lead to oscillatory features 
in TA data even when the probe is not resonant with an electronic transition. The synchronized vibrational 
motion of the solvent molecules induced by the pump pulse periodically modifies the refractive index and, 
in turn, the
 
frequency content of the probe spectrum.
11
 
The resultant shifting of the probe spectrum to higher 
and lower frequencies induces an oscillating transmittance of the pr
o
be pulse at each wavelength across its 
spectrum. When performing single wavelength measurements as shown in Figure
 
1.8b, coherences of both 
the solvent and solute manifest as oscillations that are superimposed upon the population dynamics of the 
molecule.
 
Fourier transformation of the isolated oscillatory components results in a vibrational spectrum of 
the system consisting only of the coherently activated vibrational modes. Additionally, using linear 
predictive singular value decomposition (LPSVD)
1
4
 
to fit the TA data to a summation of exponentially 
damped cosines as seen in equation 1.5, allows one to not only extract frequency information similar to a 
Fourier transform
 
 
 
1.5
 
analysis, but also the phase of the oscillat
i
on, 

i
, its amplitude, 
A
i
, and its dephasing time, 

i
. The quantitative 
characterization of coherent nuclear motion afforded by LPSVD analysis then provides the experimenter 
with an intricate view of the coherent excited state dynamics just after 
absorptio
n of a photon.
 
 
 
13
 
 
1.5
.
 
Contents of This Dissertation
 
 
The work presented in the remainder of this dissertation discusses the ultrafast excited state 
dynamics in several different first
-
row transition metal chromophores, with an emphasis on characte
rizing 
an
d extracting dynamical information from excited state vibrational coherences. We begin in chapter
 
2 with 
a series of chromium(III) tris
-
betadiketonate compounds. A review of the relevant literature surrounding 
this class of chromophores is present
ed and an
 
updated excited state deactivation mechanism is proposed 
to describe the dynamics in each molecule of the series. The coherent nuclear motion is characterized and 
used to help build the kinetic model.
 
 
In chapter
 
3, we will use transient absorpti
on spectr
oscopy to characterize the excited state 
dynamics in a small, related series of cobalt(III) tris
-
betadiketonate compounds. The coherent nuclear 
motion will again be characterized to aid in describing the excited state dynamics of the system. 
Furth
ermore, t
he types of motion and their dephasing times will be compared with the structurally similar 
chromium systems.
 
 
The subsequent two chapters address two separate methods for extending the MLCT lifetime of 
iron(II) polypyridyl type complexes. First, 
in chapte
r 4 we seek to identify and disrupt the specific normal 
modes that drive the ultrafast MLCT deactivation in this class of compounds. By gleaning information from 
excited state vibrational coherences, a synthetic modification to hinder a specific s
ubset of 
the 3N
-
6 normal 
modes of a control molecule is proposed which results in a dramatically altered MLCT lifetime. In 
chapter
 
5, we show that the photophysical properties of iron(II) chromophores can be systematically and 
modularly tuned by swapping a
 
single b
identate polypyridyl ligand with stronger field cyanide or acyclic 
dicarbene ligands. This approach results in over an order of magnitude change in both the ground state 
recovery and MLCT lifetimes. 
 
 
Finally, in chapter
 
6 of this dissertation we 
discuss p
ossible future efforts in each of the areas 
investigated.
 
 
 
14
 
 
 
 
 
 
 
 
 
 
 
 
 
REFERENCES
 
 
 
 
15
 
REFERENCES
 
 
(1)
 
Figgis, B. N.; Hitchman, M. A
. Ligand Field Theory and Its Applications
; Wiley
-
VCH: New York, 
2000.
 
(2)
 
Maiman, T. H. Stimulated Optical Radiation in Ruby. 
Nature
 
1960
, 
187
, 493

494.
 
(3)
 
Berera, R.; van Grondelle, R.; Kennis, J. T. M. Ultrafast Transient Absorption Spectroscopy: 
Principles and Application to Photosynthetic Systems. 
Photosynth. Res.
 
2009
, 
101
, 105

118.
 
(4)
 
Balzani, V.; Juris, A.; Ceroni, P. 
Ph
otochemistry and Photophysics: Concepts, Research, 
Applications
; Wiley
-
VCH: Weinheim, 2014.
 
(5)
 
Scholes, G. D.; Fleming, G. R.; Chen, L. X.; Aspuru
-
Guzik, A.; Buchleitner, A.; Coker, D. F.; 
Engel, G. S.; 
van Grondelle, R.; Ishizaki, A.; Jonas, D. M.; 
et al
. Using Coherence to Enhance 
Function in Chemical and Biophysical Systems. 
Nature
 
2017
, 
543
, 647

656.
 
(6)
 
Zewail, A. H. Femtochemistry: Atomic
-
Scale Dynamics of the Chemical Bond. 
J. Phys. Chem. A
 
2000
, 
104
, 5660

5694.
 
(7)
 
De Silvestri, S.; Cerullo, G.; La
nzani, G. 
Coherent Vibrational Dynamics
; CRC Press: Boca Raton, 
FL, 2008.
 
(8)
 
Ruhman, S.; Joly, A. G.; Nelson, K. A. Time Resolved Observations of Coherent Molecular 
Vibrational Motion and the General Occurrence of Impulsive Stimulated Scattering. 
J. Chem.
 
Phys
. 
1987
, 
86
, 6563

6565.
 
(9)
 
Al., F. E. Solution of the Schrodinger Equatio
n by a Spectral Method. 
J. Comput. Phys
. 
1982
, 
47
, 
412

433.
 
(10)
 
 
Heller, E. J. The Semiclassical Way to Molecular Spectroscopy. 
Acc. Chem. Res
. 
1981
, 
14
, 368

375.
 
(11)
 
Liebel, M
.; Schnedermann, C.; Wende, T.; Kukura, P. Principles and Applications of Broa
dband 
Impulsive Vibrational Spectroscopy. 
J. Chem. Phys. A
 
2015
, 
119
, 9506

9517.
 
(12)
 
Yang, T.
-
S.; Chang, M.
-
S.; Chang, R.; Hayashi, M.; Lin, S. H.; V

hringer, P.; Dietz, W.; Sc
herer, 
N. F. Femtosecond Pump
-
Probe Study of Molecular Vibronic Structures and
 
Dynamics of a 
Cyanine Dye in Solution. 
J. Chem. Phys
. 
1999
, 
11
0
, 12070.
 
(13)
 
Pollard,
 
W. T.; Decheimer, S. L.; Wang, Q.; Peteanu, L. A.; Shank, C. V.; Mathies, R. A. Theory 
of Dynamic Absorption Spectroscopy of Nonstationary States. 4. Application to 12
-
f
s Impulsive 
Raman Spectroscopy of Bacteriorhodopsin. 
J. Phys. Chem.
 
1992
, 
96
, 6147

615
8.
 
(14)
 
Rosca, F.; Kumar, A. T. N.; Ye, X.; Sjodin, T.; Demidov, A. A.; Champion, P. M. Investigations 
of Coherent Vibrational Oscillations in Myoglobin. 
J. Phys. Chem. A
 
2000
, 
104
, 4280

4290.
 
 
 
 
16
 
Chapter 2: Electronic and Vibrational Excited State Dynamics 

3
 
Derivatives
 
2.1
.
 
Introduction to Chromium(III)
 
Transition metal chromophores can often show stability across a wide range of oxidation states an
d 
the presence of many different types of excited states only adds to their unique uti
lity towards 
photochemical and photophysical applications. Charge transfer excited states for example, which involve 
oxidation of the metal center and reduction of one of
 
the ligands for metal to ligand charge transfer states 
or vice versa for ligand to me
tal charge transfer excited states, are of broad interest for use in electron 
transfer based applications such as dye sensitized solar cells and photoredox catalysis. Lig
and field excited 
states, which essentially derive from redistribution of electrons am
ongst the valance 
d
-
orbitals of the metal 
center, are capable of redox chemistry
1
 
and are often also conducive to substitution and isomerization 
reaction
s due to the significant antibonding nature of some 
d
 
orbitals.
2
 
Furthermore, in the case of 
chromium, these ligand field excited states are involved in the original ruby laser.
3
 
To date, chromium chromophores are the most well studied out of any transition
 
metal. Even as far 
back as 1857, the phosphorescent lifetimes of chromium compounds w
ere measured by Becquerel and his 
phosphoroscope.
4
 

mophoric reputation goes back to its discovery in 1798 by 
French chemist Nicholas Louis Vauquelin, who noticed the wide range of colors that the element could 
produce and
 
hence named the newly discovered element chromium deriving from the Greek root 
chroma
 
meaning color.
5
 
More recent investigations, however, have likely been more inspired fr

in the first laser, the ruby laser, by Theodore Maiman in 1960.
3
 
Additionally, from a practical perspective, 
d
3
 
chromium com
pounds have a number of convenient properties. The 
d
3
 
electron configuration is relatively 
kinetically inert with respect to ligand substitution reactions,
6
 
thus allowing a wealth of homoleptic 
octahedral chromium(III) compounds in the literature. These compounds are also often lumi
nescent, 
providing a nice handle with which to investigate photophysical properties. R
elative to other 
d
-
electron 
configurations, the electronic structure of chromium(III) compounds are quite simple as well, with a limited 
number of excited states to 
worry about while performing photophysical investigations.
 
 
17
 
In this work, the photophysical 
properties of Cr(acac)
3
 
and its electronically and sterically perturbed 
analogues will be investigated with regard to how nuclear motion maps onto the excited state 
decay 
mechanism as a function of electronic and steric perturbation. The Cr(acac)
3
 
molecula
r platform in 
particular has several properties that make it convenient to study, including the low quantum yield for 
photosubstitution reactions and the availabilit
y of high oscillator LMCT transitions from the excited 
electronic state that improve the si
gnal amplitude for transient absorption experiments.
7
 
The electr
onic 
structure is simple as suggested earlier, with only a few electronic states involved i
n the photodynamics 
upon visible light excitation. While Cr(acac)
3
 
is formally of 
D
3
 
symmetry, it is often helpful to express its 
electronic states using 
O
 
symmetry 
labels. The pertinent electronic states of Cr(acac)
3
, expressed using 
O
 
symmetry labels and
 
depicted using the 1 electron picture are presented in Figure
 
2.1. The ground electronic 
state (
4
A
2
) has a (
t
2g
)
3
 
electron configuration and consists of an electron
 
of parallel spin occupying each of 
the three valence 
t
2g
 
orbitals of chromium. The lowest 
energy excited states of Cr(acac)
3
 
are the 
2
E and 
2
T
1
 
states which both have the (
t
2g
)
3
 
electron configuration. The 
2
E excited state involves an intraconfigurational
 
spin
-
flip transition of one of 
t
2g
 
electrons relative to the ground state, while the 
2
T
1
 
s
tate consists of one 
singly occupied 
t
2g
 
orbital and one doubly occupied 
t
2g
 
orbital on the metal center. Since these electronic 
states are close in energy and behav
e similarly to modifications in the ligand field strength, they are often 
referred to colle
ctively as the 
2
E state. This dissertation will similarly henceforth refer to both states 
collectively as the 
2
E state. Because the 
2
E state is isoconfigurational wi
th the ground state, its energy is 
relatively insensitive to changes in the overall ligand 
field strength of the molecule. Due to the spin
-
flip 
nature of this transition, however, its energy is sensitive to the electron density on the metal center and the 
interelectronic repulsion, and thus can be tuned through nephelauxetic effects. Additionall
y, due to the 
intraconfigurational nature of the 
4
A
2
 

2
E transition, the geometries of these states are typically quite 
similar resulting in nested potential energy surfaces and narrow absorption and emission profiles. The 
4
T
2
 
state has a (
t
2g
)
2
(
e
g
)
1
 
electron configuration and because it involves the promo
tion
 
of a 
t
2g
 
electron to an 
e
g
*
 
antibonding orbital, its energy relative to the ground electronic state is highly dependent on the energy gap 
between these orbital sets, i.e. the ligand
-
field strength of the molecule. Furthermore, since the 
e
g
*
 
orbital 
 
18
 
ha
s su
bstantial metal
-
ligand antibonding character, this results in significantly distorted geometries relative 
to the ground electronic state. Rather than a symmetric expansion across all Cr
-
O bonds, however, the 
spatial degeneracy of the 
4
T
2
 
state dictates
 
tha
t a symmetry breaking distortion will occur to stabilize the 
complex according to the Jahn
-
Teller theorem. Quantum chemical calculations performed by Ando 
et al
. 
have shown that for Cr(acac)
3
 
this results in an axial compression and equatorial expansio
n of
 
the Cr
-
O 
bonds.
8
 
 
In reality, the electronic states depicted in Figure 2.1 cannot always be expressed perfectly in the one 
electron picture. To provide a more complete descri
ption of the wavefunctions for the pertinent elect
ronic 
states of the 
d
3
 
Cr(acac)
3
,the LIGFIELD program
9
 
was used with the help of Jon Yarranton to calculate the 
electron density distributions for each of the electronic states li
sted in Figure
 
2.1. The resulting electron 
density
 
distributions are illustrated in Figure
 
2.2. Since the 
2
E and 
2
T
1
 
states result from intraconfigurational 
Figure 
2
.1.
 
Qualitative potential energy diagram representing the pertinent electronic states of 
Cr(acac)
3
. The 
2
T
1g
 
and 
2
E potential energy surfaces have be

2

simplicity. The electron configurations in the 1
-
electron 
picture are represented for each state at right.
 
 
19
 
transitions from the 
4
A
2
 
ground state, each of these wavefunctions appear essentially identical. T
he 
4
T
2
 
and 
4
T
1
 
states, however, result from interconfigurational transitions from the 
4
A
2
 
ground state and hence look 
quite different. The 
4
T
2 
state results from the promotion of a 
t
2g
 
electron to an 
e
g
*
 
orbital. More specifically, 
this arises from one of 
three one
-
electron transitions (
d
xy
 

d
x
2
-
y
2
, 
d
xz
 

z
2
-
x
2
, or 
d
yz
 

d
z
2
-
y
2
) which 
effectively cause a 45
°
 
rotation of the electron density about one the cartesian axes.
2
 
The 
4
T
1
 
state is more 
complex and arises from partial population of both of the 
e
g
*
 
set of orbitals. With partial population in each 
of the 
d
-
orbitals, the 
4
T
1
 
electron density distribution is muc
h more spherically symmetric relative to each 
of the electronic sta
tes discussed previously.
 
2.2
.
 
Summary of Previous Results
 
The Cr(acac)
3
 
molecule has had a long history both within and outside of the McCusker group. We 
begin our summary of the photophys
ics of Cr(acac)
3
 
with the work of Juban 
et al.
 
which represents the
 
first 
photophysical investigation to study this molecule on ultrafast timescales.
7
 
Juba
n 
et al
. used ultrafast and 
low temperature nanosecond visible pump
 
visible probe spectroscopy as well as low temperature emission 
measurements to characterize the dynamics in this system. The narrow emission feature presented within 
this work as well as t
he congruence between the low temperature ns TA spectrum with the T
A spectrum 
collected at 5
 
ps, unequivocally showed the 
2
E state has formed by 5
 
ps and is the lowest energy electronic 
state from which the ~700
 
ps room temperature ground state recovery oc
curs. The apparent narrowing of 
4
A
2
, 
2
E, 
2
T
1
 
4
T
2
 
4
T
1
 
Figure
 
2.2.
 
Electron density distributions for the pertinent electronic ground and ex
cited states of d
3
 
compounds in an
 
O
h
 
ligand field. Distributions were calculated using the LIGFIELD program with the 
help of Jon Yarranton.
 
 
20
 
the TA spectrum (Figure 2.3) with a time constant of 1.1
 
ps was assigned to vibrational cooling on the 
2
E 
state and a lower limit for the rate of 
4
T
2
 

 
2
E ISC was assigned as 
k
 
> 1 x 10
13
 
s
-
1
 
(

 
< 100 fs) based o
n 
their ins
trument response function. Interestingly, the time constant for this fast, 1.1
 
ps, process was found 
to be invariant with respect to excitation energy, however the magnitude for the decay was found to increase 
as excitation energy increased. At 
the time, t
his was rationalized based on the wider range of absorptions 
available from a higher lying vibrational level than a lower lying vibrational level.
 
 
A follow up study performed by Joel Schrauben and Kevin Dillman on the Beck group laser system at
 
Michigan S
tate University attempted to leverage the shorter laser pulses of that system to observe the 
4
T
2
 

 
2
E ISC event in Cr(acac)
3
.
10
 
H
e instead observed a rapidly damped 164
 
cm
-
1
 
wavepacket on the excited 
electronic state (Figure 2.4a). Since coherent oscillations were observed well after the ISC event occurred, 
these wave packets were said to have been generated on the initially populat
ed 
4
T
2
 
surface and survived the 
ISC to the 
2
E state. Furthermore, it was hypothesized that the rapid dephasing time was an indicator that 
the observed vibrational mode was intricately involved in the ISC dynamics. Given that the electronic state 
on which t
he wavepacket was initially formed exhibited
 
a significant expansion of the Cr
-
O bond distance, 
it was proposed that the observed mode was similar to the 184
 
cm
-
1
 
symmetric breathing mode of the ground 
Figure
 
2.3.
 
(left) Transient abs
orption spectrum of Cr(acac)
3
 
in acetonitrile solution at 

t
 
= 5 ps (solid 
line) along with normalized absorption amplitudes obtained from single
-
wavelength nanosecond 
transient absorption data (red markers) collected in a 9:2 propionitrile/butyronitrile glass at 90
 
K. Close 
agreement between these two datasets s
ug
gests that Cr(acac)
3
 
is in the fully thermalized lowest energy 
excited state, the 
2
E state, by 

t
 
= 5 ps. (right) Transient absorption spectrum of Cr(acac)
3
 
over the 
course of the first 5 ps following 625 nm excitation. Reprinted with permission from ref
er
ence 7. © 
(2005) American Chemical Society.
 
 
21
 
state as calculated from density functional theory ca
lculations (Figure
 
2.4b). Upon inspection of motions 
inherent to this 184 cm
-
1
 
calculated mode and several other low frequency modes, it was observed that 
many of these normal modes involved signif
icant distortion of the Cr
-
O bonds coupled with large ampli
tude 
motions of the periphery methyl groups. It was subsequently hypothesized that disruption of this mode via 
synthetic modifications to make the peripheral groups more sterically encumbered would
 
result in changes 
to the photophysics of the compound. To 
test this hypothesis, the methyl groups were replaced with 
t
-
butyl 
groups. Instead of a slight narrowing of the spectrum as observed for Cr(acac)
3
, the TA spectrum of the 
Cr(TMHD)
3
 
compound (Figure
 
2.5) exhibited a decay in the bluer portion of the excited
 
state absorption 
region and a concomitant rise in the redder portion of the ESA region with an average time constant of 1.8 
± 0.4 ps and representing a significant change to the overall spectral p
rofile. Based on the principle that 
vibrational relaxation 
should not result in a significant overall change to the visible transient difference 
spectrum,
11
 
these results were interpreted to depict a change in electronic state implying a 1.8
 
ps 
4
T
2
 

 
2
E 
ISC. This represents a significant reduction to the ISC rat
e relative to the parent Cr(acac)
3
 
molecule and 
Figure
 
2.4.
 
(a) Single
-
wavelength kinetic trace of Cr(acac)
3
 
monitored at 592 
±
 
2 nm following 
600 nm excitation in acetonitrile. An exponential increase in excited state absorbance (

 
= 320 fs) 
was subtracted
 
from the data to better illustrate the oscillatory compon
ent of the data. The solid line 
corresponds to a fit to two exponentially damped oscillatory components with frequencies and 
damping times of 164 
±
 
20 cm
-
1
 
(

damp
 
= 70 fs) and 75 
±
 
25 cm
-
1
 
(

damp
 
= 
1.6 ps). (b) Density 
functional theory calculated vibratio
nal mode on the ground electronic state. The excited state 
wavepacket was proposed to exhibit a similar motion to this 184
 
cm
-
1
 
calculated symmetric 
breathing mode. Figure reproduced from reference 
10 with permission of the Royal Society of 
Chemistry.
 
a)
 
b)
 
 
22
 
seems to support the hypothesis that the rapid dephasing time of the 164 cm
-
1
 
mode is indicative that the 
nuclear motion occurring du
ring this vibration is somehow involved in and perturbed by the intersyste
m 
crossing process and that synthetic modifications inspired by the identification of these  rapidly dephased 
modes can lead to altered photophysics. These results are seminal to bot
h the work presented in the 
dissertation of Dr. Eileen Foszcz
12
 
whi
ch will be summarized shortly and the bulk of the work presented in 
this dissertation in which the veracity of these claims will be investigated and this approach will be applied 
in 
an effort to tailor the photophysics of a different chemical system.
 
 
The 
next advance in the understanding of the photophysics of Cr(acac)
3
 
came from the 
computational work of Ando 
et al
.
8
 
This group calculated the geometries of the ground and lowest en
ergy 
quartet excited state. After approximating the 
2
E geometry as the ground state geometry they were able to 
calculate the displacement of each normal mode of the 
2
E state required to distort the complex from the 
lowest energy geometry of the 
2
E state to
 
the lowest energy geometry of the 
4
T
2
 
state. In other words, they 
calculated the Duschinsky displacement vector (Figure
 
2.6). This in turn allowed the group to const
ruct a 
one
-
dimensional linear reaction coordinate describing the least motion path between
 
the geometries of the 
2
E and 
4
T
2
 
states. By calculating the potential energy surfaces for the pertinent doublet and quartet excited 
Figure
 
2.5.
 
(a) Transient absorption spectra of 
Cr(TMHD)
3
 
in dicloromethane following 633 nm 
excitation. (b) Single
-
wavelength kinetic trace collected at 450 nm and (c) 550 nm following 610
 
nm 
excitation. Solid lines represent single exponential fits (with 

obs
 
= 2.0 
±
 
0.2 ps and 1.6 
±
 
0.2 ps for 
(b) an
d (c) respectively). Figure reproduced from reference 10 with permission of the Royal Society 
of Chemistry.
 
a)
 
b)
 
c)
 
 
23
 
states as well as SOC constants along this reaction coordinate the group was able to then perform 
wavepacket simulations t
o model the excited state dynamics. These simulations had shown ~ 40 % of the 
4
T
2
 
population transfer to the doublet manifol
d within 250
 
fs, in large part due to the curve crossing and 
large spin orbit coupling between the 
4
T
2
 
and 
2
T
1
 
states. Interestingly
 
these simulations show a relatively 
compact standard deviation of the expectation value of the position along this one
-
dime
nsional coordinate 
during the first ~ 200
 
fs and then a subsequent ballooning of this standard deviation suggesting a broadening 
or 
dephasing of the wavepacket, consistent with the short dephasing times measured for the symmetric 
breathing mode observed by
 
Schrauben
10
, Foszcz
12
, and work described later in this dissertation. It is also 
noteworthy that this simulation shows a 
significant 
4
T
2
 
population remaining by 300
 
fs which does not 
match with the interpretations of the ultrafast work conducted by Juba
n and McCusker.
7
 
While 
this may be 
effected by the lack of an internal conversion pathway within the doublet manifold, and only using a single 
nuclear coor
dinate, it may hint at a nuance in the photophysical properties of Cr(acac)
3
.
 
a)
 
b)
 
c)
 
Figure
 
2.6.
 
(a) Plot of the elements of the Duschinsky displacement vector for the distortion of the ground 
state geometry to 
4
T
2
 
state equilibrium geometry as a function of vibrational modes on the ground electronic 
state inset showing an overlay of these
 
two geometr
ies. (b) Time dependent populations calculated using 
wavepacket analysis. (c) Plot of the position expectation value as a function of time during wavepacket 
simulation (solid) and standard deviation (shaded). Reprinted with permission from Refe
rence 8.
 
 
 
24
 
 
The next series of works conducted by Kunttu and coworkers,
13,14
 
used ultrafast time resolved 
infrared sp
ectroscopy to investigate the electronic relaxation pathways of Cr(acac)
3
 
through the lens of 
vibrational dynamics. In the first study,
13
 
Kunttu and coworkers exci
ted at 400
 
nm into the higher energy 
4
T
1
 
band of Cr(acac)
3
 
and performed transient absorption studies while m
onitoring across the 

CO and 

C=C stretching bands (Figure 2.7). Biphasic bleach recovery was observed with time constants of 15 and 
760 ps. The sl
ow, 760
 
ps, kinetic component was assigned as ground state recovery from the relaxed 
2
E 
state which matches w
ell with the ~800
 
ps ground state recovery measured by Juban and McCusker.
7
 
The 
15
 
ps process which produces 70
-
85 % of the ground state recovery is ascrib
ed to ISC back to the quartet 
manifold, likely initially to the 
4
T
2
 
state. This 15
 
ps process was unobserved in the visible probe TA data 
of Juban 
and McCusker.
 
 
The next study by Kunttu and coworkers
14
 
sought to further clarify the details of the excited state
 
dynamics of Cr(acac)
3
. From the work of Juban and McCusker,
7
 
it was already known that excitation into 
the 
4
T
2
, 
4
T
1
, and 
4
LMCT excited states results in s
ub
-
100
 
fs decay to the 
2
E state and that by 100
 
ps decay 
was exclusively from t
he 
2
E state. Kunttu and coworkers sought to describe the details of the dynamics 
a
)
 
b)
 
Figure
 
2.7.
 
(a) Steady state (lower trace) and transient difference spectra (upper traces) following 
400
 
nm excitation of Cr(acac)
3
 
in tetrachloroethylene. (b) Evolution of the 1521 cm
-
1
 
bleach recovery 
(

C=C) following 400 nm excitation. Solid lin
e corresponds to a
 
biexponential fit with 


 
= 15 ps and 


 
= 760 ps. Inset superposes the bleach and the inverse of the excited state absorption feature at 1509 
cm
-
1
. The sharp feature at early time is part of the coherent artifact. Figure reprinted with 
permission 
from re
ference 13. © (2007) American Chemical Society.
 
 
25
 
between these times by invoking a more detailed analysis of time resolved infrared data both af
ter infrared 
and electronic excitation and to fit the data to an updated kinetic model. By analysis of the time dependent 
shifting and narrowing of the excited s
tate absorption for both the 

CO and 

C=C stretching bands for 
4
LMCT (345
 
nm) and 
4
T
1
 
(400 nm) excitation, a consistent 7
-
10
 
ps time constant was observed that was 
assigned to vibrational relaxation on the excited electronic state (Figure
 
2.8). Following t
his shifting and 
narrowing pr
ocess, the ESA band kinetics were similar to the bleach. The difference between this 7
-
10
 
ps 
lifetime and the 15
 
ps bleach recovery was ascribed to a bottleneck state in the electronic decay cascade. 
Additionally, by monitoring
 
bleach recovery dynamics fol
lowing infrared excitation of the fundamental 

CO and 

C=C stretching transitions, a 12 ps kinetic component was observed that is representative of the 
vibrational relaxation rate on the ground electronic state as well as a ~ 3
00
-
700 fs component that was 
assigned as intramolecular vibrational redistr
ibution (IVR). Given this 12
 
ps lifetime for vibrational cooling 
Figure
 
2.8.
 
(a) Time dependent shifting for 

CO and 

C=C stretching ESA bands of 
Cr(acac)
3
 
in tetrachloroethylene following 
4
A
2
 

 
4
T
1
 
(400
 
nm) and 
4
A
2
 

 
4
LMCT 
(345
 
nm) excitation. (b) Time 
dependent narrowing of the for 

CO and 

CC ESA 
features. (c) Bleach recovery kinetics of the 

C=C stretching band following excitation 
into the 0 

 
1 transition at 1521 cm
-
1
 
with fit shown in red. The inset shows a 
compar
ison of the decay dynamics followin
g excitation at various wavelengths. Data 
in inset are normalized to the value at 10 ps. Reprinted with permission from reference 
14. © (2015) American Chemical Society.
 
a)
 
b)
 
c)
 
 
26
 
on the ground electronic state and the superposable bleach recovery dynamics at longer timescales for IR 
and electronic excitation (Figure
 
2.8c inset), it was proposed that ground state reco
very following electronic 
excitation was limited by vibratio
nal relaxation on the ground electronic state and that the difference 
between the rates following IR and electronic excitation could be attributed to additional electronic 
contribution to the deca
y process.
14
 
With the above insights, Kunttu and coworkers,
14
 
proposed
 
the kinetic model depicted in Figure
 
2.9 
and described mathematically by the set of differential equations in equation 2.1.
 
 
2.1a
 
 
2.1b
 
 
2.1c
 
 
2.1d
 
Figure
 
2.9.
 
Simplified Energy level diagram for Cr(acac)
3
 
depicting the r
elaxation processes 
proposed in referenc
e
 
14.
 
 
27
 
In t
his model, the initial 
4
T
2
 

2
E ISC is considered prompt so that all the excess energy is converted 
directly to vibrational energy on the 
2
E state, consistent with the conclusion of the initial ultrafast work of 
Juban and McCusker that ISC is kinetically competitive with vibrational
 
re
laxation on the 
4
T
2
 
state.
7
 
Since 
the lifetime of the 
2
E state exhibits an A
rr
henius behavior, this can be described by a thermally activated 
back
-
inter
sys
tem crossing (bISC) to the 
4
T
2
 
state governed by the time dependent effective temperature of 
the 
2
E state population, 
T(t)
, described in equation 2.2. and the activation energy, 
E
a
.
 
 
2.2
 
The time
 
dependent
 
effective temperature
 
is a function of the initial effective temperature of the 
2
E state 
population, 
T
x
, and the vibrational cooling rate constant, 
k
VC
. At high intramolecular temperatures, the 
2
E 
state population can undergo efficient bISC to the 
4
T
2
 
sta
te and subsequently un
dergo IC to the ground 
electronic state according to 
k
IC
. At this point, the hot 
4
A
2
 
population undergoes vibrational relaxation 
according to 
k
VR
 
to reform a fully relaxed ground state. Phosphorescence from the 
2
E state was included i
n 
the model with 
k
p
, h
owever the slow relative rate has little impact on the model at room temperature. It 
should be noted that with lower effective temperatures, the bISC process is significantly less efficient, 
supporting the differences in the observati
ons between the lower 
energy excitations of Juban and 
McCusker
7
 
and the higher energy excitations of Kunttu and coworkers
14
.
 
 
 
The kinetic model described above
 
was able to simulate 
the bleach recovery dynamics well for both 
4
A
2
 

4
T
1
 
and 
4
A
2
 

4
LMCT excitation (Figure
 
2.10) and yielded 
E
a
 
= 23.0 kJ mol
-
1
,
 
A
=2.9 × 10
13
 
s
-
1
 
(1/
A
 
~ 35 fs), and 
k
IC
 
= 0.64 × 10
12
 
s
-
1
 
(

IC
 
= 1.56
 
ps). Interestingly the 1.56
 
ps time constant 
for internal co
nversion from the 
4
T
2
 
state, matches well with the 1.1
 
ps time constant observed in the earlier 
visible pump probe work of Juban and McCusker
7
 
which was as
signed to vibrational relaxation on the 
2
E 
state suggesting a misassignment for this kinetic component. While the comparison between the 
interpretations of Juban
7
 
and Kunttu
13,14
 
are made more difficu
lt by the use of different solvents and 
excitation energies, it is compelling to consider the merits of both and how each interpretation would affect 
the conclusions drawn in other bodies 
of work on Cr(acac)
3
, namely those of Joel Schrauben whose work 
 
28
 
was
 
published prior to these discoveries.
10,15
 
For the moment, we will leave the question of which 
interpretation has more merit an
d the subsequent implications to be addressed with the re
sults shown later 
in this chapter.
 
 
It seems also necessary to pause and reflect on the rapid, ~300
-
700
 
fs component that was observed 
by Kunttu and coworkers
14
 
following infrared excitation and assigned to IVR. This fast component was 
observed on the ground el
ectronic state in each of Cr(acac)
3
, Al(acac)
3
 
and Fe(acac)
3
 
as well as in different 
solvents, strongly suggesting it is inherent to the M(acac)
3
 
structure
. Interestingly, although the magnitude 
of this component was significant in both the 

CO and 

C=C mo
des studied, it also varied between them, 
showing that the observation of IVR is vibrational mode dependent. Finally, this rapid ~300
 
fs IVR process 
seems 
also likely to result in dephasing of any vibrational coherences of these normal modes thus raising 
q
uestions about the hypothesis presented in the work of Schrauben 
et al.
 
that short dephasing times are an 
indicator for vibronic mixing of the nuclear and 
electronic degrees of freedom. These questions will be 
discussed alongside the results presented in t
his dissertation shortly.
 
 
In addition to observing coherent oscillations in the excited state of Cr(acac)
3
, Joel Schrauben also 
investigated the ultrafast
 
dynamics of a series of Cr(acac)
3
 
analogues involving substitutions either in the 3 
position of the 
acac backbone or at each of the methyl groups in the periphery of the molecule.
15
 
Early time 
Figure
 
2.10.
 
Bleach recovery dynamics of Cr(acac)
3
 
in tetrachloroethylene with experimental 
data (symbols) and simulated (solid lines) ground state populations. Reprinted with permission 
fr
om reference 14. © (
2015) American Chemical Society.
 
 
 
29
 
full spectrum TA data on several 
3
-
substituted Cr(acac)
3
 

work had shown that each of the 3
-
substituted compounds behaved very similarly to the par
ent Cr(acac)
3
 
compound, exhibiting a 1
-
2 ps decay in the blue edge of the spectrum that was des
cribed as a narrowing 
and redshift of the excited state absorption feature. Given the similarity between these spectral changes and 
those observed for the unsubs
tituted compound, these 1
-
2 ps kinetic components were also assigned to 
vibrational relaxation 
on the 
2
E excited state. While it was acknowledged that the exaggerated shifting 
observed for the phenyl
-
substituted derivative (Figure 2.11 d) could be interpre
ted in several ways such as 
being reflective of 
4
T
2
 

2
E ISC, it was ultimately attributed to a
 
rotation of the phenyl group. Attempts 
were made at investigating the rotationally hindered mesityl substituted compound although was 
incomplete by the time Joel graduated. Given that these data were collected p
rior to the work of Kunttu and 
a)
 
b)
 
c)
 
d)
 
Figure
 
2.11.
 
Early time full spectrum transient absorption data on (a) Cr(3
-
Br
-
acac)
3
, (b) Cr(3
-
SCN
-
acac)
3
, (c) Cr(3
-
NO
2
-
acac)
3
, and (d) Cr(3
-
Ph
-
acac)
3
 
following 
4
A
2
 

 
4
T
2
 
excitation.
 
Reproduced from data 
in Reference 15.
 
 
30
 
coworkers,
14
 
which claimed a back
-
intersystem crossing decay mechanism, these assignments may benefit 
from further consideration.
 
 
 
The final body of work to be summarized comes from previous McCusker group member, Eileen 
Foszcz.
12
 

vestigate the influence of excitation wavelength on the observed 
vibrational coherences of Cr(acac)
3
 
and Cr(TMHD)
3
. Despite the ~3000 cm
-
1
 
range in excitation 
wavelength applied across the 
4
A
2
 

4
T
2
 
absorption band, for Cr(acac)
3
 
and Cr(TMHD)
3
 
no significa
nt 
differences in the oscillatory components were observed, both with respect to frequency and damping times. 
A summary of the observed oscillatory components of the Cr(acac)
3
 
datasets are pre
sented in Table
 
2.1. It 
should be noted that a ~500 cm
-
1
 
compone
nt has been omitted from Table
 
2.1 since it will be shown shortly 
that this oscillatory component does not originate from Cr(acac)
3
. An important observation from the results 
of this body of w
ork is that among the numerous oscillatory components listed, th
e 187 and 457 cm
-
1
 
modes 
appear most robustly throughout the collection of data sets and are easily assigned to metal ligand stretching 
modes, with the 187 cm
-
1
 
mode resembling the DFT calcula
ted mode from the work of Schrauben 
et al
.
10
 
Interestingly the frequency and dephasing time for this mode are significantly diffe
rent between both bodies 
of work. Given the better frequency mat
ching of the 187 cm
-
1
 
average collected by Fosczcz,
12
 
it is likely 
that this is more correct. The substantially slower dephasing time observed by Foszcz for this mode raise 
questions ab

ge standard error in several 
of the presented damping times, however, make this difficult to speculate on. If two of the ~ps damping 
times observed for the 187 cm
-
1
 
mode are deemed as outliers
 
as was proposed by Foszcz, the average 
damping time is reduced 
to ~420 fs. These standard deviations and differences between the two studies 
highlight the difficulty in acquiring reproducible data with this experiment and sample. Furthermore, in 
Oscillatio
n Frequency, cm
-
1
 
70 (13)
 
111 (11)
 
187 (11)
 
221 (8)
 
256 (7)
 
282 (4)
 
333 (8)
 
457 (9)
 
553 (15)
 
Oscillation Damping Time, fs
 
240 
(250)
 
2300 
(2900)
 
580 
(400)
 
340 
(140)
 
2300 
(3000)
 
160 
(160)
 
580 
(700)
 
1200 
(1200)
 
650 
(560)
 
 
Tabl
e
 
2.1.
 
Summary of oscillatory components observed for variable excitation wavelength TA 
experiments 
conducted by Foszcz
12
 
on Cr(acac)
3
.
 
Errors are reported as standard deviations of all 

 
 
31
 
addition 
to the difficulties already inherent to this experimental techni
que, the low solubility of Cr(TMHD)
3
 
in acetonitrile, a solvent which exhibits relatively minor background oscillatory signal, prompted Foszcz to 
attempt to monitor coherent oscillations in di
chloromethane, which exhibits large amplitude oscillatory 
compon
ents that mask those attributable to the solute.
 
 
Despite the numerous investigations into Cr(acac)
3
 
and its synthetic derivatives, there seems to be 
several open questions regarding the photo
physics of this set of compounds. First, the nature of the ~1
 
ps
 
process must be identified. Several studies by previous McCusker group members,
7,10,15
 
have assigned this 
a
s vibrational relaxation in the 
2
E state, whereas Kunttu and coworkers
14
 
suggest this is 
4
T
2
 

4
A
2
 
ground 
state recovery preceded by thermally activated back
-
intersystem crossing. Next, we must address the 
question of which vibrational modes are coherently acti
ve in Cr(acac)
3
 
and its derivatives and how these 
synthetic modifications affect which 
vibrational modes are observed and their influence on the 
photophysical dynamics. Furthermore, we must attribute each of these vibrational modes to specific 
electronic s
tates. Finally, we will attempt to address the question of whether dephasing times can 
be an 
indicator for involvement of a given vibrational mode in the reaction coordinate for a photophysical process. 
Each of these questions will be addressed in the subs
equent work of this dissertation. 
 
2.3
.
 
Experimental Section
 
2.3.1
.
 
Synthesis of Cr(aca
c)
3
 
Derivatives
 
i. tris(2,4
-
pentanediono)chromium(III), Cr(acac).
 
Cr(acac)
3
 
was prepared according to the procedure of 
Schrauben.
15
 
ii. tris(3
-
chloro
-
2,4
-
pentanediono)chromium(III), Cr(3
-
Clac)
3
. 
Cr(3
-
Cl
-
ac)
3
 
was prepared from 
Cr(acac)
3
 
according to the method of Collman 
et al.
16
 
Anal. calcd. for C
15
H
18
O
6
Cl
3
Cr (452.64 g mol
-
1
): C 
39.8, H 4.0, N 0; found: C 40.88, H 4.01, N 0.20.
 
iii. tris(3
-
bromo
-
2,4
-
pentanediono)chromium(III), Cr
(3
-
Brac)
3
. 
Cr(3
-
Brac)
3
 
was synthesized 
previously by Joel Schrauben.
15
 
 
32
 
iv. tris(3
-
iodo
-
2,4
-
pentanediono)chr
omium(III), Cr(3
-
Iac)
3
.
 
Cr(3
-
Iac)
3
 
was prepared from Cr(acac)
3
 
according to the method of Collman 
et al.
16
 
Anal. calcd. for C
15
H
18
O
6
I
3
Cr (726.99 g mol
-
1
): C 24.8, H 2.5, 
N 0; found: C 25.08, H 2.58, N 0.085. Crystal structure included in appendix.
 
v. tris(3
-
nitro
-
2,4
-
pe
ntanediono)chromium(III), Cr(3
-
NO
2
ac)
3
.
 
Cr(3
-
NO
2
ac)
3
 
was s
ynthesized 
previously by Joel Schrauben.
15
 
vi. tris(3,5
-
heptanediono)chromium(III), Cr(HD)
3
. 
Cr(HD)
3
 
was synthesized in a manner similar to 
Cr(acac)
3
 
but with slight modification.
15
 
Heptanedione (4.116
 
g) was added to a round bottom flask with 
CrCl
3
·6H
2
O (2.854
 
g), urea (19.897
 
g), and 100
 
mL of DI water followed by an
other 100
 
mL DI water. 
Solution was refluxed for 16
 
h and cooled to room temperature before filtering on glass frit while washing 
with a copious amount of water. Remaining solid was dissolved in ace
tone and filtered into a separate 
container and then filte
red again through a silica plug with acetone eluent. The solvent was evaporated and 
then sample was dissolved in a small amount of chloroform, washed with brine, extracted into chloroform 
and dried 
over MgSO
4
. Solvent was evaporated off to afford a purple 
product.
 
vii. tris(2,6
-
dimethyl
-
3,5
-
heptanediono)chromium(III), Cr(DMHD)
3
.
 
Cr(DMHD)
3
 
was prepared in a 
manner similar to Cr(acac)
3
 
but with slight modification.
15
 
CrCl
3
·6H
2
O (0.424
 
g), dimethylheptanedione 
(2.79
0
 
g), and urea (0.752
 
g) were added to a 50
 
mL round bottom flask followed by
 
10 mL 1,4
-
dioxane 
and 9
 
mL DI water and brought to reflux for 22
 
h. Solid was collected via vacuum filtration subsequently 
filtered through a neutral alumina plug with acetone el
uent. Solvent was evaporated to afford a purple 
product. Crystal structure is
 
included in appendix.
 
viii. tris(2,2,6,6
-
tetramethyl
-
3,5
-
heptanediono)chromium(III), Cr(TMHD)
3
.
 
Cr(TMHD)
3
 
was prepared 
according to the procedure of Schrauben.
15
 
ix. tris(4
-
chloro
-
2,2,6,6
-
tetramethyl
-
3,5
-
heptan
ediono)chromium(III), Cr(4
-
Cl
-
TMHD)
3
.
 
Cr(4
-
Cl
-
TMHD)
3
 
was prepared from Cr(TMH
D)
3
 
from a slightly modified method of Collman 
et all
.
16
 
To a 
suspensi
on of Cr(TMHD)
3
 
(0.596 g) in glacial acetic acid, n
-
chlorosuccinimide (6.045 
g) was added with 
vigorous stirring in four parts separated by a few minutes each. Solution was then allowed to stir another 
30
 
min before being filtered over celite and washed wi
th copious amounts of water, sat. NaHCO
3(aq)
, 
 
33
 
sat.
 
Na
2
S
2
O
3(aq)
, and then more
 
water. Pulled solid into new filter flask with DCM then evaporated solvent 
and recrystallized from hexanes to afford green product. Crystal structure included in appendix.
 
2.3.2
.
 
Computational Methods
 
All calculations were performed with Gaussian 09 softw
are
17
 
o
n the Michigan State University 
High Performance Computing Center servers. Previously optimized geometries for Cr(acac)
3
 
and 
Cr(TMHD)
3
 
calculated by former McCusker group member, 
Eileen Foszcz
12
, were used as starting 
geometries for Cr(acac)
3
 
and Cr
(TMHD)
3
 
ground state optimizations, and all subsequent optimizations for 
the various Cr(acac)
3
 
derivatives were started from one of these geometries. Optimizations were performed 
using the unrestricted B3LYP functional, 6
-
311G
**
 
basis set, tight convergenc
e criteria, and a polarizable 
continuum model for the solvent. Calculations for Cr(3
-
Iac)
3
 
were unique in that the LANL2DZ basis set 
was used. Choice of solvent used for the polar
izable continuum model had little effect on the results in the 
frequency regi
on of interest. Cr(acac)
3
, Cr(HD)
3
, and Cr(DMHD)
3
 
calculations were performed in 
acetonitrile, Cr(TMHD)
3
 
was performed in tetrahydrofuran, Cr(3
-
Clac)
3
, Cr(3
-
Iac)
3
, and Cr(4
-
ClTMHD)
3
 
were performed in dichloromethane, Cr(3
-
Brac)
3
 
was performed in toluene, a
nd Cr(3
-
NO
2
ac)
3
 
was 
performed in acetone with solvent choice in part inspired by solubilities of the var
ious compounds and the 
solvents used for TA experiments. Mass weighted vector displacement diagrams of each pertinent 
vibrational mode were created in t
he Visual Molecular Dynamics program.
18
 
2.3.3
.
 
Time R
esolved Spectroscopic Methods
 
Our ultrafast transient absorption spectroscopy setup has been described i
n detail previously.
12,19
 
A 
Ti:sapphire oscillator (Coherent: Mantis) provides a ~400
 
mW, 80
 
MHz s
eed beam centered at ~800
 
nm 
for use in a regenerative amplifier (Coherent: Legend Elite) with a 527
 
nm Nd:YLF pump (Coherent: 
Evolution) producing a 1
 
KHz pulse train of approximately 35
 
fs 1
 
mJ pulses at 800
 
nm. This output is split 
using a 70/30 beam sp
litter with the higher energy portion used to pump a two
-
stage optical parametric 
amplifier (OPA) (Coherent: OperA Solo) to generate a visible pump beam
 
which is used as the excitation 
source during the transient absorption experiment. This pump beam is ag
ain split with a 90:10 beam splitter 
with the lower energy portion being used as a probe beam for degenerate pump
-
probe experiments. The 
 
34
 
remaining 30% o
f the 800 nm output from the regenerative amplifier is either used to pump a second OPA 
(Coherent: Opera
 
Solo) to generate a probe beam for single wavelength nondegenerate pump
-
probe 
experiments or used to make a white light continuum by attenuating the po
wer and focusing ~1
 

into a 3 mm sapphire window. This white light continuum can again be used 
for single wavelength 
nondegenerate pump
-
probe experiments or a large swath of the continuum can be measured using full 
spectrum measurements. Group vel
ocity dispersion for each OPA output is corrected using a folded prism 
compressor. Pump
-
probe delay is a
djusted by routing the pump OPA output into a computer
-
controlled 
translation stage (Aerotech). Pump and probe polarizations are set to ~54.7° relative 
to each other to 
eliminate anisotropic contributions to the signal and focused into a 1
 
mm quartz cuvett
e containing the 
sample. The pump pulse energies were set in the range of 2
-

by halving the pump power usin
g a variable neutral density filter wheel. Probe pulse energies were set to 
roughly 1/10
th
 
of the pump.
 
During single wavelength measurements, the pump beam was passed through a 446
 
Hz mechanical 
chopper. A monochromator (Jarrell Ash: MonoSpec 18; 1200 groove/mm grating, blaze 500 nm) was 
placed after the sample and used to select a 3.8
 
nm FWHM slice of the 
probe beam spectrum to be measured 
on an amplified silicon photodiode (Thor laboratories: PDA55). The center wavelength of this slice of the 
probe spectrum is specified for each single wavelength TA trace. A reference beam is produced from a 
reflection of 
the probe beam off a microscope slide cover slip inserted into the beam path prior to the 
sample. This reference beam is measured on another amplified silicon photodiode and attenuated with 
neutral density filters and an iris until its intensity matches th
at of the signal beam. Both the signal and 
reference beams are coupled to a lock
-
in amplifier (Stanford Research: SR810) which is synchronized to 
the chopper modulating the pump beam. Connection of the lock
-
in amplifier to a data acquisition card 
allows th
e signal to be converted from analogue to digital. A data workup program written in LabView by 
previous McCusker group members converts the data to a change in absorbance. 
 
During full spectrum measurements the white light continuum probe was collimated an
d 
subsequently focused into the sample using spherical mirrors. A lens placed after the sample was then used 
 
35
 
to focus the transmitted continuum probe into a liquid light guide 
which was coupled to a spectrometer 
(Spex
 
270 M) with 1
 
mm entrance slits. A gr
ating (300 grooves/mm; blaze 600
 
nm) within the spectrometer 
dispersed the probe beam onto a diode array detector (Hamamatsu: HC233
-
0900; C5964 NMOS, 1x512 
pixel array) affordin
g a roughly 300
 
nm spectral window. Data is acquired using a LabView program 
wr

pump beam scatter reaching the detector in the absence of the probe beam. Nex
t, a background is collected 
by measuring the signal with both pump and probe i
ncident on the sample at negative time delay. This 
effectively measures the transmitted spectrum in the absence of excited state dynamics. A LabView 
program subsequently subtrac
ts the dark counts spectrum to account for pump scatter. Data collection then 
p
roceeds by moving the pump delay line and measuring the transmitted intensity and dividing by the 
background signal. 
 
In this chapter all single wavelength TA experiments used t
o measure coherent oscillations 
employed degenerate pump and probe beams except
 
when the specified probe wavelength was 480
 
nm, in 
which case a white light continuum was used as the probe source. Single wavelength measurements 
employing a white light conti
nuum probe were also used to measure the kinetics of several systems at 490, 
55
0, and 650
 
nm. The instrument response function (IRF) for single wavelength kinetics traces was 
Figure
 
2.12.
 
Represen
tative OKE trace collected in methanol with 600
 
nm degenerate pump 
and probe. The first half of the dataset (black circles) was fit to a Gaussian to determine the 
FWHM of the IRF.
 
 
36
 
determined from the optical Kerr effect (OKE) in methanol by fitting the early ha
lf of the signal to a 
Gaussian and extracting the full width at half maximum.
20
 
An example dataset is shown in Figure
 
2.12. To 
set up an OKE experime
nt, a polarizer is placed between the sample and the monochromator and set such
 
that the transmitted polarization is 90
°
 
off the probe polarization. Since the polarizer blocks all the light 
reaching the signal photodiode, the reference photodiode is blocke
d during the experiment. The experiment 
is conducted by monitoring the intensit
y at the signal photodiode while scanning pump probe delay times 
from shortly before to shortly after the time at which the pump and probe pulses overlap in the sample. 
When bot
h pump and probe are overlapped in the sample, the induced birefringence rotate
s the polarization 
of the probe beam and allows a small fraction to transmit through the polarizer to the detector. This response 
is a result of two types of contributions: elec
tronic and nuclear. The essentially instantaneous electronic 
response results f
rom the distortion of electron clouds of the molecules experiencing the intense electric 
field. The nuclear response arises from the reorientation of molecules to align in the d
irection of the electric 
field. The transient birefringence arising from the nu
clear response decays as the molecules spontaneously 
rotate back to equilibrium positions.
21
 
When this reorientational motion is slow, a tail can be seen on the 

-
lived 
nuclear response makes it ideal for quantifying the 
duration of pump
-
probe over
lap in the sample. For this reason, the IRF of single wavelength measurements 
will be quantified using OKE measurements performed in methanol as opposed to the solvent used to 
d
issolve the sample of interest. In addition to the OKE measurement, a cross
-
cor
relation experiment can be 
performed in the same solvent used in the sample of interest to characterize the TA signals not due to the 
solute. The cross
-
correlation setup is simi
lar to the OKE experiment except that the polarizer between the 
sample and mono
chromator is removed and the signal and reference photodiode signals in the absence of 
the pump beam are nonzero and matched. During full spectrum measurements, the IRF and disp
ersion are 
measured by collecting TA data on the neat solvent and then used to 
chirp correct the sample with solute.
 
 
Full spectrum datasets were often globally fit using the free to download OPTIMUS
22
 
fitting 
software developed by Dr. Chavdar Slavov of Goethe Univ
ersity in Frankfurt. This program allowed for 
chirp correction, fitting of the 
coherent artifact, and fitting of the population dynamics to a series of 
 
37
 
exponential decay functions convolved with the IRF. One additional useful feature was the ability to 
cal
culate decay associated difference spectra (DADS). These plotted the amplitude 
of each exponential 
component as a function of wavelength to produce a spectrum that describes spectral changes arising from 
each kinetic component. The evolution associated dif
ference spectra (EADS) could be calculated for cases 
of sequential decay dynami
cs as well. These plot the effective spectrum of transient species and are also 
useful for characterizing the excited state dynamics. 
 
 
Single wavelength data was fit in the Igo
r pro program. Population dynamics were fit to a 
summation of exponential funct
ions, with only enough exponential functions to adequately describe the 
data. These fits were started after the point where the coherent artifact of the neat solvent was found t
o 
return to baseline. A fast Fourier transform (FFT) was then performed on the 
residuals of this fit to produce 
a magnitude spectrum. No zero
-
padding, artificial resampling of the datapoints, or spectral window 
functions were used in the FFT analysis. The 
FFT was thus effectively performed using a rectangular 
window spanning from the
 
initial residual point to the last point of the residuals which still had a constant 
time spacing. Additionally, oscillatory components were fit using a linear predictive singu
lar value 
decomposition (LPSVD) program written in MATLAB originally by Dr. And
rey Demidov and Paul 
Champion
23
 
and edited by Dan Roscioli of the Beck group at Michigan State University. This program f
its 
the data to a series of exponentially damped cosinusoids as shown in equati
on 2.3.
 
 
 
2.3
 
This program allows the user to fit population dynamics using zero frequency components as well as extract 
exponentially damped 
oscillations. Additionally, by making use of the Fourier relationship between 
damping ti
me and width in the frequency domain, the LPSVD program makes a vibrational spectrum that 
corelates to the LPSVD fit in the time domain. This LPSVD vibrational spectrum
 
is then normalized and 
compared with the normalized FFT of the residuals of a fit of th
e population dynamics to help ground the 
result of a mathematical fit to reality.
 
 
 
38
 
 
To illustrate the SVD portion of the LPSVD process, we can perform LPSVD on the TA data on 
Cr(TMHD)
3
 
in benzene in a 1
 
mm quartz cuvette with a 40
 
fs IRF featured in Figur
e
 
2.13a. In add
ition to 
an exponential rise, this data exhibited numerous oscillatory components. The LPSVD program is able to 
decompose the dataset into a linear combination of exponentially damped cososines as described in 
equation
 
2.3. The exponential k
inetics of the 
system are modeled as zero
-
frequency components and are 
shown as the black line superimposed on the data in Figure
 
2.13a. The difference between the complete 
data and this exponential component contains the remaining oscillatory components a
s seen in the b
lack 
residuals of Figure
 
2.13a. These oscillations are further decomposed into several frequency components 
Figure
 
2.13.
 
(a) Representative TA data exhibiting coherent oscillations (b
lue). T
he 
corresponding LPSVD fit (red) and exponential components acquired from LPSVD fit (black) 
are superimposed on the original dataset. The Residuals of each of these with respect to the 
raw data are shown above in red and black respectively. (b) Indi
vidual 
oscillatory components 
parsed out of raw data using LPSVD.
 
a)
 
b)
 
 
39
 
each with their own exponential damping times (Figure
 
2.13b). The sum of each of these oscillatory and 
exponential components gives 
the total LPSVD fit shown in red, superimposed upon the data in 
Figure
 
2.13a. The residuals of this fit contains un
accounted for oscillatory components (if any) and noise, 
although it can be warped if the exponential fitting is significantly off. 
 
 
Once an
 
LPSVD fit is obtained, the next step is to identify the origin of each of the oscillatory 
components of the datase
t. The numerous possible sources of coherent oscillations in a TA experiment can 
make the complete assignment of coherences a difficult task.
 
To alleviate some of this difficulty, control 
studies are performed on the neat solvent and, in some cases, the em
pty cuvette, in addition to the 
a)
 
b)
 
c)
 
d)
 
Figure
 
2.14.
 
Illustration of methodology used to systematically identify origin of coherent 
oscillations in TA data. TA data and corresponding LPSVD fits of (a) Cr(T
MHD)
3
 
dissolved in 
benzene in a quartz cuvette, (b) neat benzene in a quartz cuvette, and (
c) an empty quartz cuvette. 
(d) Superimposed LPSVD magnitude spectra normalized to the common peak at 500
 
cm
-
1
.
 
 
40
 
experiment conducted on the sample of interest. By analyzing the freqeuncy content of each of these data 
sets
 
separately (either via FFT or LPSVD) the data can be compared for common oscillatory components. 
An example of thi
s is shown in Figure
 
2.14 for the same dataset shown in Figure
 
2.13. Here, TA data along 
with LPSVD fits are shown for Cr(TMHD)
3
 
in benzene, 
neat benzene, and an empty quartz cuvette. 
Comparison of the frequency domain spectra in Figure
 
2.14d clearly shows
 
an approximately 500
 
cm
-
1
 
mode that is common to each dataset indicating it likely originates from the quartz cuvette. Similarly, the 
overla
p of the 995
 
cm
-
1
 
peak in the samples with and without solute (but not in the empty cuvette) clearly 
indicate that 
this is a mode of the the solvent. This type of assignment can often also be supported by 
looking at tables of vibrational modes for common s
olvents or by performing DFT frequency calculations. 
In the present case, we can readily assign the 995
 
cm
-
1
 
mode t
o the strongly Raman active symmetric ring 
stretch mode of benzene. Finally the two additional lower frequency modes seen only for the non co
ntroll 
data must originate from the solute itself. In this example it does not seem as though a coherently active 
m
ode of the solute overlaps with a background frequency, however this is not always the case. In such cases 
it can be helpful to perform a sim
ilar experiment in a solvent with different frequency vibrations. If the 
mode still appears when the solute is in t
he newly chosen solvent but not in the new solvent control, it is 
likely that the solute does have a mode at that frequency. One could also, 
in theory, subtract the solvent 
response from the TA data on the sample of interest and observe if there is still f
requency content in the 
region in question. This second method was not performed in the work of this dissertation. Nevertheless, 
we see that 
by performing careful control experiments every time experimental conditions are altered, one 
can relatively confid
ently identify vibrational modes of the solute.
 
When deciding how to set up the data collection parameters, several important factors were 
co
nsidered. Since the coherent oscillations observed in this work are often of relatively low amplitude, it is 
necess
ary to have a high signal to noise ratio to acquire reliable LPSVD fits. For control experiments on 
neat solvents with little additional back
ground signals, this can be achieved in relatively few scans. For the 
work presented in this dissertation solvent c
ontrol experiments were the result of at least 4 averaged scans. 
Data on the actual sample of interest was collected over a substantially lon
ger period of time with a 
 
41
 
minimum of 8 scans and often greater than 20 or 30. The individual scans were compared an
d any scans 
which showed a significant reduction to the signal magnitude or shifting of the coherent artifact were 
removed before averaging. 
Since both the LPSVD program and a general FFT require constant spacing 
between datapoints in the time domain, expe
riments were conducted such that the early time portions had 
constant time steps. When population dynamics were significantly slower than the
 
oscillatory decay 
kinetics, both a high
-
density region of points at earlier times and a lower density of points we
re collected 
thereafter to adequately capture both the oscillatory and exponential kinetics. Typically, oscillations were 
fit over the first 
2
-
3 ps of data (over which time the data would be collected with constant time spacing 
between data points), howeve
r in some instances, data was collected such that larger amounts of time could 
be fit. Typically, only compounds with longer dephasing times 
required more than 2
-
3
 
ps worth of data to 
perform a vibrational analysis on. For compounds with rapid dephasing ti
mes, collecting data over the first 
2 ps should be more than enough since the oscillations will be completely damped out over most of this 
ti
me. When collecting data for vibrational analysis, step distances of between 4 and 10 fs were used 
depending on the
 
sample, its kinetics, and the length of time that needed to be collected to get a good 
LPSVD fit. The Nyquist frequency for each of these sa
mpling rates is well above 1000 cm
-
1
 
insuring that 
all the oscillatory components described in this dissertation are as adequately sampled as physically 
allowable. The resolution of the non
-
zero
-
padded FFT spectra presented in this chapter is typically bet
ween 
5 and 20 cm
-
1
 
depending on the 
step distance in the time domain and the number of datapoints included in 
the FFT. The resolution in the frequency domain could have been improved by up to a factor of 2 by zero
-
padding the data enough to double the numb
er of datapoints,
24
 
however, this was not performed because 
the FFT was meant only to provide a qualitative sense of the frequency content of the data with the LPSV
D 
analysis providing more quantitati
ve results.
 
 
 
The duration of the IRF has an important influence on the ability to observe coherent oscillations 
in the time domain. It is rather intuitive that as the IRF becomes longer, the ability to observe periodic 
modulations to TA signal from vibrat
ional coherences become diminished. Ground state vibrational 
coherences also have the additional caveat that they require sufficient time delay between the first and 
 
42
 
second field interactions of the pump pulse to show si
gnificant modulation amplitude. Igno
ring this caveat 
for ground state vibrational coherences, we can get an idea of the so
-
called impulsive window describing 
the range of observable frequencies by convolving a gaussian IRF with a cosine function. The ampli
tude of 
the resulting oscillatory fu
nction is proportional to exp(
-

IRF
2

2
/2), where 

IRF
 
is the standard deviation of 
the IRF and 

 
is the angular frequency of the vibration.
25
 
It should be noted that an identical result can be 
obtained by simply taking 
the Fourier transform of the gaussia
n IRF. A simulation
 
of the impulsive window 
for various IRFs can be seen in Figure
 
2.15. The specific IRFs in this chapter varied depending on the pump 
probe 
combination and the wavelength re
gion of each. As described a
bove the vast majority of the 
coherence data presented in this chapter are from degenerate pump probe TA experiments where the probe 
beam is a replica of the pump pulse generated in a beam splitter. When these pump and probe pul
ses were 
generated via secon
d harmonic of the signal (i.e. when the center of the pump spectrum was between 580 
and 700
 
nm) the IRF was roughly 45
 
fs with slight fluctuations on the order of 10
 
fs from one experimental 
to another. The McCusker group laser 
system produces its shortest
 
laser pulses in this wavelength region 
and, as such, much of the experiments in this dissertation seeking to observe coherent oscillations employed 
laser pulses in this wavelength range. In contrast, degenerate pump probe exper
iments employing sum 
frequen
cy of the idler (~570
-
530
 
nm) or sum frequency of the signal (~520
-
460 nm) had significantly longer 
IRFs which were roughly 70
 
fs. Experiments using a white light continuum probe had an IRF of roughly 
Figure
 
2.1
5
.
 
Approximated impulsive windows for vari
ous IRFs.
 
 
43
 
100
-
160 fs. Whether using a
n IRF of ~45
 
fs or ~75
 
fs, t
he impulsive windows simulated in Figure
 
2.15 
show that vibrational frequencies of 200
 
cm
-
1
 
or less should be reliably visible in TA experiments. It is 
interesting to note that many of the datasets presented in this chapter show
 
relatively high frequency s
olvent 
vibrational modes despite, according to the approximated impulsive windows in Figure
 
2.15, having 
significant, or in some circumstances, enormous attenuation factors. It is unknown to this author why 
exactly these modes a
re so reproducibly visible i
n the following TA data; however it may be a result of the 
intense electric fields of the laser pulses used in these experiments and the high Raman activity of the 
observed vibrational modes.
 
Ground state absorption spectra were
 
collected on a Varion Cary 
50 UV
-
vis spectrometer before 
and after TA experiments to check for sample degradation. Samples were made to have an absorbance 
between 0.2 and 0.8 at the center wavelength of the pump spectrum in a 1
-
mm pathlength cuvette. Inst
ead 
of creating a blank for 
each ground state absorption measurement, the spectrometer was zeroed with nothing 
in the sample holder then an absorption spectrum was measured. The baseline offset at ~800
 
nm, a region 
in which ground state absorption was abse
nt, was subtracted from the 
rest of the data set to approximate 
the actual absorption of the sample.
 
2.4
.
 
Results and Discussion
 
2.4.1
.
 
Cr(acac)
3
 
Addressing the Question of the Nature of the 1
 
ps Process
 
One of the open questions raised during the 
summary of the previous results on C
r(acac)
3
 
was the 
nature of the excited state decay mechanism. Initial ultrafast visible transient absorption experiments 
conducted by Juban had assigned a ~1 ps narrowing and redshifting of the transient spectrum to vibr
ational 
relaxation on the 
2
E excited
 
state following unobserved, ultrafast 
4
T
2
 

2
E ISC.
7
 
These conclusions were 
made in part based on the qualitative rete
ntion of the bandshape as well as th
e increased decay amplitude 
in the wings of the ESA spectrum following lower wavelength excitation. Kunttu and coworkers used a 
combination of IR pump IR probe and visible pump IR probe experiments to arrive at a mechani
sm 
involving impulsive 
4
T
2
 

2
E ISC followed by thermally activated bISC and ~1
 
ps 
4
T
2
 

4
A
2
 
IC, with 
 
44
 
~7
 
ps vibrational relaxation on the 
2
E state.
13,14
 
The interpretations within these two studies are in conflict 
and should be addressed before progressing to the focus of this work regarding the information that can be 
obtai
ned from vibrational coherences and their dephasing times. While the comparison of these two studies 
is complicated by the dif
ferent solvents used in each study as well as the vastly different energies associated 
with the initially generated excited states
 
(conceivably opening the door to additional decay pathways), the 
increased sensitivity of the IR probe to vibrational dynamic
s suggests that the 7
-
10
 
ps vibrational relaxation 
time observed by Kunttu and coworkers is more plausible than the ~1
 
ps time con
stant assigned using a 
visible probe which is considerably less diagnostic or selective towards vibrational dynamics.
 
 
In addi
tion to a retrospective comparison of the two above interpretations,
7,14
 
we can gain further 
evidence for which kinetic model is most appropriate by repeating the experiment of Kunttu 
and 
coworkers
14
 
with a visible probe beam, thus allowing for a more direct comparison of the physical 
observables. Figure
 
2.16a shows f
ull spectrum TA data on Cr(acac)
3
 
following 400
 
nm excitation into the 
4
T
1
 
state, similar to what was performed by Kunttu. Dec
ay associated difference spectra (DADS) associated 
with a global fit of the dataset using OPTIMUS fitting software
22
 
can be seen in Figure
 
2.16b. The 1
 
ps time 
Figure
 
2.16.
 
(a) Full spectrum transient absorption data on Cr(acac)
3
 
following 
4
A
2
 

 
4
T
1
 
(400
 
nm) 
excitation in acetonitrile. (b) Decay associated difference spectra (DADS) associated with the global 
fit of the data in (a). A 1
 
ns lifetime asso
ciated with the longtime offset was locked in but had no 
bearing on the value of the shorter time co
nstant.
 
a)
 
b)
 
 
45
 

7
 
work involving the lowest energy spin
-
allowed excitation into the 
4
T
2
 

14
 
work involving 
excitation into the higher energy 
4
T
1
 
state and use of an
 
infrared probe. Given the agreement between all 
three of these experiments, it is highly likely that the same process is being monitored. Initial examination 
of the TA data in Figure
 
2.16 shows a general decrease in intensity of the ESA spectrum while ret
a
ining 
the overall band shape. Closer inspection of the 1
 
ps DADS, however, shows a flat decay amplitude from 
~460
 
nm to the edge of the probe window at 600
 
nm. Given that the 
2
E state is formed in less than 100
 
fs,
7
 
and that the spectral evolution following 400
 
nm excitation does not resemble spectral narrowing, we can 
unequivocally say that the 1
 
ps process does not represent vibrational relaxation on the 
2
E state. Thus, since 
ISC has already occurred, and the observed spectral evolution is not associated with vibrational relaxation, 
then the most likely origin of the spectral evolution is a net loss of excited state population. Furthermore, 
because the 1
 
p
s
 
DADS does not resemble the spectrum of the 
2
E state represented by the spectrum at 10
 
ps, 
we can say that the TA data in Figure
 
2.16a do not show the loss of the 
2
E state population exclusively. The 
presence of an additional excited state population can 
t
hus be inferred. This means that the spectrum of a 
quartet excited state, most likely the lowest energy 
4
T
2
 
excited state, is superimposed upon the 
2
E state 

k
inetic 
model.
14
 
The
 
large amplitude decay across the entire spectrum suggests that at high excitation energy (low 
excitation wavelength), the 
4
T
2
 
state serves to effectively siphon 
2
E state population and direct recovery of 
the ground electronic state. The dim
inished amplitu
de of the decay following lower energy 
4
A
2
 

4
T
2
 
excitation,
7
 
is consistent with a less efficient thermally activated bISC.
 
From the above analysis, it was shown that the most likely excited state 
decay mechanism in 
Cr(acac)
3
 
at early
 
times following excitation into the quartet manifold involves impulsive ISC to the 
2
E 
state followed by thermally activated bISC to the 
4
T
2
 
state and IC to the ground state. The portion of the 
2
E 
state population that 
remains trapped following vibrational
 
relaxation, then decays at a much slower 
timescale dependent on the temperature of the surrounding bath. It stands to reason that since many of the 
3
-

3
 
derivatives shared very similar kinetics and 
spectral evolution as the unsubstitut
ed 
 
46
 
parent molecule, that each of these derivatives and the unsubstituted Cr(acac)
3
 
undergo a similar decay 
pathway. Extending these interpretations to the more sterically encumbered derivatives such as 
Cr(TMHD)
3
, howeve
r, requires more caution. The spectra
l evolution for Cr(TMHD)
3
 
(Figure
 
2.5) involves 
a decay of the ESA at ~465
 
nm with a concomitant rise centered around ~525
 
nm. Given the LMCT nature 
of the excited state absorption, the extinction coefficient for the ES
A feature at 525
 
nm is likely several
 
orders of magnitude larger than that for the Laporte forbidden ground state absorption, hence we can still 
say that the observed dynamics do not exclusively represent repopulation of the ground electronic state as 
this
 
would result in a net decrease in 

A
. Furthermore, the significant changes to the absorption profile 
suggest that vibrational relaxation is not a primary factor causing the observed spectral changes. Thus, a 
significant portion of the spectral evolution m
ust be due to a change in electronic 
state, specifically, 
formation of the lowest energy, 
2
E, excited state. This result could be achieved in two ways. In the first 
option, the initial 
4
T
2
 

2
E ISC event could be slowed relative to Cr(acac)
3
 
and the observ
ed 1.8
 
ps process 
could be the first 
and only change in excited state population that occurs. This route would require a 
moderately slower IC rate relative to Cr(acac)
3
 
as well as a substantially slower ISC rate. An alternative 
route may involve a rapid 
4
T
2
 

2
E ISC and subsequent bISC (simil
ar to Cr(acac)
3
) with a less efficient IC 
to recover the electronic ground state. If the 
4
T
2
 
population cannot efficiently undergo IC, then it must 
continue to decay back to the lowest energy 
2
E excited state as the exce
ss thermal energy is dissipated. At 
present, given the lack of information it is difficult to determine which mechanism better describes the 
dynamics of Cr(TMHD)
3
. One way to experimentally gain more insight into which mechanism has more 
merit may be a var
iable excitation wavelength experime
nt. Since the second mechanism involves a 
thermally activated bISC, then the variable excitation energy should modulate the amount of 
4
T
2
 
population 
formed from bISC and by extension the amount of conversion to the 
2
E st
ate as the system vibrationally 
cool
s. The first mechanism, however, does not predict any difference in the net amplitude change as a 
function of excitation energy. A further discussion of these possible mechanisms will be continued again 
later in this cha
pter.
 
 
 
47
 
Observation and Interpretatio
n of Coherent Oscillations
 
 
As discussed in the summary of previous investigations into Cr(acac)
3

10
 
had 
proposed the notion that dephasing times that are intermediate between the sub
-
100
 
fs dephasing times of 
electronic coherences
26,27
 
and the often reported as ~p
s dephasing times of purely vibrational 
coherences
24,27

31
 
may serve as an indicator for vibroni
c mixing and involvement of a particular vibrational 
mode in an ultrafast photophysical process. Furthermore, it was proposed that the identification of these 
vibrational
 
modes with short dephasing times, could help inspire synthetic modifications that cou
ld alter 
the photophysical properties of the system. Much of the remainder of the discussion in this chapter will 
seek to address these claims and to further understand t
he role of nuclear motion in the excited state 
evolution of Cr(acac)
3
 
and its derivati
ves. Specifically, an attempt will be made to (1) identify the coherently 
active vibrational modes of Cr(acac)
3
, (2) assign each of these vibrational modes to a specific 
electronic 
state, (3) quantify the dephasing times of each vibrational mode, (4) ident
ify the origin of the dephasing 
mechanisms, and (5) address the merit of the of the hypothesis that short dephasing times can be indicative 
of involvement of specific vib
rational modes in an ultrafast transformation. To achieve this, coherent 
vibrational m
odes will be observed as outlined in section
 
2.3.3.
 
 
In much of the following report, coherent oscillations in TA data will be analyzed to provide 
information on excited 
state dynamics. Because there are many mechanisms by which a coherence can be 
created 
in a sample during a pump
-
probe transient absorption experiment, it is important to assess the 
specific origin of each oscillatory component carefully. One can often expl
oit the disparity of time scales 
for dephasing between purely electronic and purely vi
brational coherences to identify the type of coherence 
being monitored, i.e. electronic or vibrational. Electronic coherences typically dephase on sub
-
100
 
fs 
timescales,
26,27
 
whereas vibrat
ional coherences can require several picoseconds to completely dephase.
24,28

30
 
In the system of interest, oscillations persist from several hun
dreds of femtoseconds to several 
picoseconds,
10,12
 
excluding purely electronic coherences as a likely component of the signal. This leaves 
three p
ossible origins for the oscillatory components in the data: (1) ground state vibration
al coherence in 
 
48
 
the solvent, (2) ground state vibrational coherence in the solute, or (3) vibrational coherence in the excited 
electronic state of the solute. In general, a vibrational coherence can be created in the ground state of the 
solvent or solute 
via impulsive
 
stimulated Raman scattering (ISRS). The solvent contribution can often 
simply be identified by performing the pump
-
probe experiment on neat solvent or performing a frequency 
domain Raman experiment.
 
 
Discerning the nature of the solute cohere
nce is in general more difficult. For this, it is helpful to 
consult the Feynman diagrams seen in Figure
 
2.17 depicting stimulated emission (SE), ground state 
bleaching (GSB) and excited state absorption (ESA) proc
esses. It should be noted that the mechani
sm by 
which ground state vibrational coherence in the solute is generated (ISRS) can be grouped with GSB. The 
magnitude of the signal from each of these processes is a function of the probability of each necessary 
field 
interaction occurring, which is prop
ortional to the transition dipole moment of the transition it induces. In 
the system of interest, both SE and GSB involve transitions solely between ligand field states, either 
between the 
4
A
2
 
and 
4
T
2
 
states or the
 
4
A
2
 
and 
2
E states. Due to the Laporte for
bidden, and in some cases, 
Figure
 
2.17
.
 
(a) Qualitative energy level diagram depicting the ground 
4
A
2
 
state (



), 
2
E state (



), 
4
T
2
 
state (



), and charge t
ransfer manifold (



) that the 
2
E and 
4
T
2
 
states can be promoted to through 
resonance with the probe beam. (b) 
Double sided Feynman diagrams depicting stimulated emission 
(SE), ground state bleaching (GSB) and excited state absorption (ESA) processes. F
ollowing excitation, 
to the 
4
T
2
 
state, the system may evolve. Depending on the time delay between pump and probe, 
a 
population may exist in the 
4
T
2
 
or 
2
E states. This relaxation pathway is denoted by a gray box. (c) The 
relationship between signal amplitud
e (third order response function) and transition dipole moments for 
relevant field interactions for SE, GSB, and E
SA.
 
 
49
 
spin forbidden nature of these transitions, the transition dipole moments are particularly small. Transitions 
between the 
2
E or 
4
T
2
 
states and the charge transfer manifold, however, shoul
d exhibit a substantially larger 
transitio
n dipole owing to the spin and parity allowed nature of these transitions. Thus, it is clear that the 
amplitude of ESA signal will be much greater than for SE or GSB. Furthermore, we can acquire a rough 
estimate of
 
the magnitude of this disparity by approx
imating the squared transition dipole moments for the 
4
A
2
 

4
T
2
 

coefficients of Cr(acac)
3
 
(~100
 
L mol
-
1 
cm
-
1
)
7
 
and [Fe(
bpy)
3
]
2+
 
(~10,000 L mol
-
1 
cm
-
1
)
32
 
respectively. From 
the third order response functions listed in Figure
 
2.17, these values imply that the ESA signal is 
approximately 100
-
fol
d larger than those for GSB or SE! This is further
 
supported by the observation that 
the TA spectrum for the systems of interest are positive throughout the visible spectrum.
7,10,15
 
Consequently, 
the only means by which we can monitor vibrational coherence of the solute in this system is by excited 
state absorption, i.e. any observable coherence in Cr(acac)
3
 
or 
its derivatives must be in an electronic excited 
s
tate.
 
 
Now that we have established the nature of the signals that will be observed during a transient 
absorption experiment, it is important to address the mechanisms by which coherences dephase. In the 
op
tical Bloch picture, dephasing occurs by populatio
n relaxation (
T
1
) and pure dephasing (
T
2
*
) caused by 
random environmental fluctuations that cause a loss of phase information. The contributions of these two 
processes to the observed homogenous dephasing t
ime (
T
2
) are additive as seen in equation
 
2.4:
 
 
 
2.4
 
Where 
T
1,a
 
and 
T
1,b
 
are the population relaxation times of the states involved in the coherence. Vibrational 
coherences typically show dephasing times on the order of picosecond
s
24,28

30
 
and in solution phase this is 
often dominated by pure dephasing processes.
33
 
It is often observed that excited state vibration
al coherences 
dephase faster than their ground state analogues due to either higher anharmonicity of the potential surface 
or an increased density of states leading to faster population relaxation. An alter
native view of coherences 
involves their descripti
on as localized wave packets. In this picture, dephasing can in part be thought of as 
 
 
50
 
 
 

pu
, nm
 
(
E
pu
, 

J)
 

pr
, nm
 
(
E
pr
, 

J)
 
IRF, 
fs
 
Solvent
 
Frequency, cm
-
1
 
(Damping Time, fs)
 
600 (
3
)
 
592
 
(0.3)
 
45
 
MeCN
 
 
 
 
286 (119)
 
462 
(1142)
 
600 (
5
)
 
592
 
(0.3)
 
50
 
MeCN
 
98 
(512)
 
192 
(412)
 
 
 
463 
(487)
 
600 (
3.5
)
a
 
592
 
(0.3)
a
 
50
 
MeCN
 
83 
(376)
 
 
 
 
458 
(1887)
 
600 (
3.5
)
b
 
592
 
(0.3)
b
 
50
 
MeCN
 
 
195 
(411)
 
242 
(1074)
 
 
 
600 (
3
)
 
592
 
(0.3)
 
50
 
MeCN
 
 
203 
(418)
 
 
 
462 
(665)
 
600 (
3
)
 
592 (0.3)
 
60
 
MeCN
 
 
200 
(404)
 
 
314 (300)
 
457 
(2146)
 
600 (
3
)
 
592 (0.3)
 
60
 
MeCN
 
 
193 
(283)
 
 
 
458 
(1180)
 
600 (
3
)
 
480 (na)
 
160
 
MeCN
 
102 
(265)
 
189 
(435)
 
 
303 (348)
 
468 
(736)
 
600 (
3
)
 
480 (na)
 
100
 
MeCN
 
 
185 
(342)
 
239 
(250)
 
 
457 
(359)
 
600 (
3
)
 
480 (na)
 
75
 
MeCN
 
98 
(374)
 
 
 
 
457 
(1460)
 
600 
(3
)
 
592 
(0.3)
 
45
 
Dioxane
 
 
205 
(332)
 
 
 
461 
(774)
 
600 (
3
)
 
480 (na)
 
100
 
Dioxane
 
 
175 
(266)
 
 
295 (358)
 
 
600 (
3
)
 
595 (0.3)
 
50
 
THF
 
 
208 
(261)
 
 
271 
(12,900)
c
 
458 
(1320)
 
600 (
3
)
 
592 (0.3)
 
50
 
THF
 
 
187 
(228)
 
 
 
477 
(1260)
 
600 (
4.8
)
 
586 (0.5)
 
40
 
THF
 
 
186 
(90)
 
 
 
460 
(1359)
 
580 (
4.4
)
 
592
 
(0.5)
 
45
 
THF
 
 
 
 
 
458 
(1051)
 
560 (
5
)
 
546
 
(0.5)
 
65
 
THF
 
 
187 
(645)
 
 
 
 
Average Frequency, cm
-
1
 
95
 
193
 
241
 
300
 
461
 
Standard Deviation, cm
-
1
 
8
 
9
 
2
 
12
 
6
 
Average Damping Time, fs
 
380
 
350
 
660
 
280
 
1130
 
Standard Deviation, fs
 
100
 
130
 
580
 
11
0
 
500
 
(a) Experiment performed with parallel pump and probe polarizations. (b) Experiment performed with 
perpendicular pump and probe polarizations. (c) This dephasing time was significantly larger than the 
remainder of the values in the series and 
thus w
as deemed as an outlier and not used to calculate an 
average.
 
Table
 
2.2.
 
Summary of coherent vibrational modes of Cr(acac)
3
 
observed during TA experiments in 
various solvents. Reported IRFs are rounded to the nearest integer multiple of 5 fs.
 
 
51
 
a spatial dispersion of the wave packet. This can occur in a highly anharmonic portion of the potential 
surface or 
as a result of multiple 
population transfer and back
-
transfer events between potential surfaces.
 
 
 
The vibrational coherences of Cr(acac)
3
 
and their corresponding dephasing times following 
4
A
2
 

 
4
T
2
 
excitation, were monitored using transient absorption spe
ctroscopy. Studies were 
conducted in 
acetonitrile, 1,4
-
dioxane, and tetrahydrofuran while probing on the red and blue edges of the excited state 
absorption feature. The results are summarized in Table
 
2.2 and show good agreement with the previous 
results c
ollected by Foszcz.
1
2
 
Inspection of the data in Table
 
2.2 show that both the frequencies observed 
Figure
 
2.18
.
 
(a) Representative TA data on Cr(acac)
3
 
collected in acetonitrile following 600
 
nm 
excitation and monitoring at 592
 
nm. (b) Comparison of results of |FFT| (blue) and LPSVD (red) 
analyses of data from part a. (c) Analogous data collected while
 
monitoring 
at 480
 
nm. (d) 
Comparison of |FFT| and LPSVD analyses for 480
 
nm probe data.
 
a)
 
b)
 
d)
 
c)
 
 
52
 
and their corresponding dephasing times are relatively insensitive to both the probe wavelength and the 
choice of solvent and hence were combined when
 
reporting 
averages and error bars. Representative datasets 
in acetonitrile can be seen in Figure
 
2.18 and in 1,4
-
dioxane and tetrahydrofuran in Figure
 
2.19. It is 
important to note that a ~500
 
cm
-
1
 
oscillatory component with a ~200
 
fs damping time was obs
erved in 
se

12
 
a similarly rapidly damped ~500
 
cm
-
1
 
oscillatory component was observed and assigned as a vibrational mode of Cr(acac)
3
. Because DFT 
a)
 
Figure
 
2.19.
 
(a) Representative TA data on Cr(acac)
3
 
collected in tetrahydrofuran (THF) 
following 600
 
nm excitation while monitoring at the red edge of the ex
cited sta
te absorption 
feature. (b) Corresponding FFT and LPSVD analyses. (c) Analogous data collected in 1,4
-
dioxane and (d) corresponding |FFT| and LPSVD vibrational spectra. It is possible that the 
271
 
cm
-
1
 
band in part
 
b is a vibrational mode of THF. T
he ~500
 
c
m
-
1
 
component in part
 
d can be 
attributed to the glass of the cuvette.
 
b)
 
c)
 
d)
 
 
53
 
calculations on the ground electron
ic state did not show any frequencies around 500
 
cm
-
1
 
this was assigned 
to a vibrational mode on the 
4
T
2
 
excited state. TA data collected on an empty glass cuvette (Figure
 
2.20), 
however, show a ~500
 
cm
-
1
 
oscillatory com
ponent, prompting a reassignment of 
this mode to vibrational 
coherence in the cuvette itself generated by off
-
resonant impulsively stimulated Raman scattering. Work 
by Furukawa and coworkers also show an intense Raman band in the range of 400
-
600
 
cm
-
1
 
of s
ilica based 
glasses that is assigned
 
to a stretching/bending mode across bridging Si
-
O
-
Si oxygens.
34
 
It should also be 
noted that the 271
 
cm
-
1
 
mode in the THF dataset in Table 2.2. and Figure
 
2.19 may be due to a vibrationa
l 
mode of the solvent. Both DFT calc
ulations and frequency domain Raman experiments on neat THF show 
a vibrational mode in this frequency region.
 
 
In addition to the 270cm
-
1
 
and 500
 
cm
-
1
 
modes discussed above, a cautionary note should also be 
placed upon t
he 95
 
cm
-
1
 
frequency listed in Table
 
2.2. Given the presence of multiple oscillatory 
components superimposed on population dynamics in this system, it can often be difficult to acquire a 
perfect exponential fit of the data. Typically, incorrectly fit exponential dec
ays show a fit line that o
scillates 
both above and below the data itself appearing as a low frequency oscillatory component. Additionally, 
given that this mode is damped out at a similar timescale to the period of the oscillation, it may be the case 
Figure
 
2.20.
 
(a) TA data collected on an emp
ty 1
 
mm glass cuvette. The observed coherent 
oscillations are assigned to ground state vibrational coherence in the cuv
ette itself. (b) Fast 
Fourier transform and LPSVD spectrum of the coherent oscillations of (a), showing a ~ 
500
 
cm
-
1
 
component.
 
a)
 
b)
 
 
54
 
that t
his component was picked u
p by the LPSVD program to compensate for a poor exponential fit at early 
times. Because of the ambiguities associated with sub
-
100
 
cm
-
1
 
oscillatory components, it is sometimes 
recommended to ignore them or take extra care to ensur
e exponential fitting is a
ccurate.
35
 
 
In addition to the difficulties associated with the often low
-
amplitude of the oscillatory features 
associated with vibrational c
oherences in a TA spectrum
, the numerous sources of oscillatory components 
to the data sets as outlined above highlight the need for caution when interpreting results. As such, the 
remainder of the discussion will focus on vibrational modes that appear mos
t robustly in the data. In
spection 
of Table
 
2.2. clearly indicates that the 193 and 461
 
cm
-
1
 
vibrational modes appear most routinely, and hence 
will be the focus of analysis. 
 
 
The lower frequency
 
193
 
cm
-
1
 
mode observed here is likely a more accurate value of the fr
equency 
observe
d originally by Schrauben
10
 
and shows good agreement with the value obtained by Foszcz
12
. The 
dephasing tim
e observed in t
his study (350
 
±
 
130
 
fs) is also significantly longer than reported originally 
(70
 
fs)
10
, however it is still subs
tantially short
er than the ~ps dephasing times typically thought of for 
vibrational coherences. The 461
 
cm
-
1
 
mode, on the other hand, shows a dephasing time (1130 ±
 
500
 
fs) that 
fits well with traditional values for vibrational co
 
herences. The observation
 
of these modes
 
in the TA 
experiments is complemented by their inclusion in the resonance Raman spectrum featured in Figure
 
2.21 
Figure
 
2.21.
 
Resonance Raman spectra collected on neat THF (black) and Cr(aca
c)
3
 
(red). The solvent 
background was subt
racted from the Cr(acac)
3
 
dataset. The 488
 
nm excitation source is resonant with 
the blue edge of the 
4
A
2
 

 
4
T
2
 
transition.
 
 
55
 
which was collected on the Proshlyakov group instrument at Michigan State University
 
by professor 
Proshlyakov.
. By tuning the e
xcitation sourc
e in the Raman experiment to be in resonance with the 
4
A
2
 

 
4
T
2
 
absorption band, the ground state vibrational modes that are coupled to that transition will be 
selectively enhanced. Because these modes are strongly coupled to the interconfig
urational 
4
A
2
 

 
4
T
2
 
transition, it makes sense that they may also be cou
pled to the interconfigurational 
4
T
2
 

 
2
E transitions.
 
 
We can gain further insight into the dephasing mechanisms for the vibrational modes of Cr(acac)
3
 
by looking at the linewidth of the peaks in the resonant Raman spectrum. The linewidth of the peaks in 
the 
Raman spectra are linked to vibrational dephasing times via a Fourier transform relationship resulting in 
the relationship shown in equation 2.5.
 
 
 
2.5
 
where 


 
is the full width at half maximum of the fit to the Raman peak an
d 


 
is the dephasing time for 

the distribution 
of solvent environments from which the oscillator can be found induce a gaussian spread 
in vibrational f
requency. The fluctuation of the vibrational energy gap can then be described by two 
parameters, a fluctuation magnitude, 

, and an effective fluctuatio
n frequency, 

. The corresponding 
lineshape function becomes (equation
 
2.6)
 
 
 
2.6
 
with coherence decay described as exp[
-
g
(
t
)]. When fluctuations are slow relative to the fluctuation 
magnitude, exp[
-
g
(
t
)] is 
approximately gaussian and inhomogeneously broadened, thus containing little 
dynamical information. In the fast modulation lim
it (



) where thermal fluctuations rapidly allow each 
oscillator to sample the distribution of frequencies, the coherences decay 
exponentially as an ensemble 
producing a Lorentzian lineshape.
36,37
 
In the present experiment the ~460
 
cm
-
1
 
peak in the Cr(acac)
3
 
Raman 
spectrum fits well to a Lorentzian lineshape despite being convolved with a estimated ~4.2
 
cm
-
1
 
square 
wave pulse due to the experimental setu
p. Because of this, the fast modulation limit, i.e. exponential 
 
56
 
coherence decay, is l
ikely a reasonable approximation. A Lorentzian can also be fit to the ~190
 
cm
-
1
 
peak, 
however the smaller amplitude of the signal in concert with the jagged background mak
e it more difficult 
to get a reliable width. The homogeneous dephasing times (
T
2
) for
 
coherences produced in the Raman and 
TA experiments can now be described more specifically using the form seen in equation
 
2.4. Since the 
Raman experiment produces a cohe
rence between the first excited vibrational level and a nondecaying 
ground state, the
 
homogeneous dephasing time is described as seen in equation 2.7.
 
 
 
2.7
 
The form of the homogeneous dephasing relation for excited state coherenc
e in our system requires some 
slight modification due to the additional sources of po
pulation decay present within this system. Here, 
coherences may decay from both electronic and vibrational population relaxation. Furthermore, since the 
excitation transit
ion is Laporte
-
forbidden and because the initial excited state is strongly displaced 
relative 
to the ground state, it requires that each portion of the superposition state is initially vibrationally excited. 
If we assume a uniform vibrational and electroni
c relaxation time for each vibronic level, the homogeneous 
dephasing time for excited
 
state vibrational coherences in our system can be described by equation 2.8,
 
 
 
2.8
 
where 
T
1,elec
 
and 
T
1,vib
 
describe electronic and vibrational 
population relaxation times respectively.
 
 
With equations 2.7 and 2.8 in hand, we can
 
now begin to parse out the origins of dephasing for the 
vibrational modes of Cr(acac)
3
 
by comparing dephasing times of the ground state vibrational modes 
obtained with th
e frequency domain, Raman, experiment and the dephasing times measured on an excited 
electronic state obtained with the time domain, TA, experiment. Before proceeding, however, it is 
worthwhile to test how well the frequency and time domain experiments agr
ee.
 
To do this, we can compare 
the ground state dephasing time of a solvent vibration
al mode calculated from the linewidth in the Raman 
spectrum and measured directly using TA. Figure
 
2.22 illustrates this approach with the 913
 
cm
-
1
 
symmetric 
ring breathin
g mode of tetrahydrofuran. A Lorentzian fit to the peak in the Raman spectrum shows a
 
FWHM 
 
57
 
of 12.9
 
± 0.4 cm
-
1
 
corresponding to a dephasi
ng time of 820 ± 30 fs according to equation
 
2.5. In the time 
domain, dephasing times in the neat solvent for a 913 ± 6
 
cm
-
1
 
oscillatory component were 950 ±
 
140
 
fs. 
While dephasing times are not perfectly in sync between the time and frequency domain meas
urements, 
they are still within error of each other and suggest that this approach may be valid for a sufficiently well
 
resolved peak in the Raman spectrum.
 
 
 
With the method established, we can begin investigating the dephasing mechanism of the      
461
 
c
m
-
1
 
mode of Cr(acac)
3
 
with ~1
 
ps damping time (Table
 
2.2). Due to the differences in kinetic 
assignments between the in
itial ultrafast visible pump visible probe experiments conducted by the 
McCusker group
7
 
and the visible pump IR probe experiments performed by Kunttu and c
oworkers
13,14
 
we 
must proceed with some degre
e of caution. Although it has been established that the kinetic model described 
by Kunttu and coworkers is most likely the more appropria
te model, it can still be beneficial to see how the 
current results fit into both kinetic models. Thus, the following a
nalyses will be based on several different 
permutations of kinetic assignments between these studies. First, we consider the case in whic
h vibrational 
relaxation in the doublet excited state is described by a 1.1
 
ps time constant as assigned originally by 
the 
McCusker group (Case
 
1).
7
 
Here, 
the agreement between the dephasing time observed in the TA experiment 
Figure
 
2
.
22.
 
(a) Raman spectrum of neat tetrahydrofuran (blue) with 488
 
nm excitation source. 
A Lorentzian fit to a band centered at 913
 
cm
-
1
 
(red) shows a FWHM of 12.9
 
cm
-
1
 
corresponding 
to a dephasing time of 820 
±
 
30
 
fs. (b) TA data o
n neat THF with 600
 
nm excitation source and 
584
 
nm probe (blue) along with LPSVD fit of the dataset (red). The average dephasing time 
measured with four datasets is 950 
±
 
140
 
fs.
 
a)
 
b)
 
 
58
 
and the vibrational relaxation time suggest that
 
dephasing in the excited state is largely determined by 
vibrational relaxation and further that the electronic population relaxation cor
responding to 
2
E
 

 
4
A
2
 
conversion on the nanosecond timescale and pure dephasing do not significantly contribute. The 
c
omparable 1.25 ± 0.02
 
ps homogeneous dephasing time suggested by the 8.5
 
± 0.1
 
cm
-
1
 
Raman linewidth 
for this mode (Figure
 
2.C
.
1) and the 
differences in equations 2.7 and 2.8, however, suggest that either the 
pure dephasing time significantly decreases or t
hat the vibrational relaxation time approximately halves 
(i.e. that vibrational relaxation in the ground electronic state becomes faster)
. Both scenarios are considered 
unlikely. In the second case to consider (Case
 
2), we will use the vibrational relaxati
on time constants 
measured by Kunttu and coworkers
14
 
using the more structurally sensitive IR probe. Here, we use 
vibrational relaxation times
 
of 7 and 12
 
ps for the excited and ground electronic states respectively and an 
electronic relaxation time on the orde
r of a nanosecond describing 
2
E
 

 
4
A
2
 
conversion. Again, in this 
scenario, 
T
1,elec
 
is too large to contribute to the observed dephasing t
ime in the excited electronic state. Here, 
the slower vibrational relaxation implies that the bulk of the dephasing occ
urs via 
T
2
*
 
with a time constant 
of 1.3
 
ps in agreement with the standard observation that most dephasing in liquids is due to pure 
dephasing.
33
 
Furthermore, this agrees well with the 1.2
 
ps dephasi
ng time cal
culated for the Raman 
experiment using the same 
T
2
*
 
as determined above and a 12
 
ps vibrational relaxation time and it also 
matches the 1.25
 
± 0.02
 
ps dephasing time determined experimentally from the Raman linewidth. Because 
of this, Case 2 see
ms to by a 
likely hypothesis. Finally, in Case
 
3, we must consider the alternative 
assignment for the
 
1.1
 
ps kinetic component suggested by the Kunttu group
14
 
and corroborated with the full 
spectrum dataset shown in the previous section to be 
4
T
2
 

4
A
2
 
IC. Here
, we will a
gain assume 7 and 12
 
ps 
vibrational relaxation times in the excited and ground states respectively. In this scenario, we must consider 
the possibility of coherence loss due to fast electronic population relaxation from the 1.1
 
ps IC. The 
inclusi
on of this 
as a significant relaxation pathway implies that the bulk of the dephasing seen for this 
mode in the TA experiment is due to 
T
1,elec
 
from IC and that the pure dephasing is much slower as is the 
case for Case
 
1. This again yields similar problems
 
when compa
ring to the Raman estimated dephasing 
time as in Case
 
1 and thus suggests that IC is not a very prevalent decay pathway following excitation into 
 
59
 
the 
4
T
2
 
state. Case 2 is thus considered the most likely evaluation of the dephasing mechanism for 
this 
vibrat
ional mode, namely the bulk of the dephasing is due to 1.3
 
ps pure dephasing processes within the 
2
E 
excited state.
 
 
In principle, a similar analysis can be conducted on the ~190
 
cm
-
1
 
vibrational mode. Due to the 
significantly lower amplitude of
 
this peak 
in the Raman spectrum and the significant distortion of the 
spectral shape from the background signal, the conclusions from this analysis would be tentative at best. 
Attempts to fit this feature to a Lorentzian profile resulted in a 10.1 ± 0.9
 
c
m
-
1
 
FWHM co
rresponding to a 
1.05 ± 0.1
 
ps dephasing time. While the results of this fitting are included for the sake of transparency, an 
analysis is omitted due to the potential errors associated with the fitting process.
 
 
Next, using density functional 
theory (DFT) calculations, we can start to 
acquire a better idea of 
the nature of each vibrational mode observed in the TA experiments and further define the type of motions 
active during the ISC process. It is important to reiterate that the vibrational m
odes in Cr(acac)
3
 
are 
predominantly sampled
 
from the 
2
E state which has the same electron configuration as the ground state and 
as such, is expected to have similar vibrational frequencies. The vibrational modes calculated for the ground 
electronic state s
hould thus serve as an adequate substitute 
for vibrational modes in the excited state of 
Cr(acac)
3
 
and simplify the calculations. On the basis of frequency matching the 193
 
cm
-
1
 
and 461
 
cm
-
1
 
Figure
 
2.
23
.
 
Nuclear deformations associated with the 193
 
cm
-
1
 
(a) 
and 
461
 
cm
-
1
 
(b) vibrational modes of Cr(acac)
3
.
 
a)
 
b)
 
Calc. 185 cm
-
1
 
Exp.  193 cm
-
1
 
Calc. 453 cm
-
1
 
Exp.  461 cm
-
1
 
 
60
 
modes can be assigned to the pair of metal
-
ligand bond stretching
 
vibrations calculated at 185 and      
453
 
cm
-
1
 
(Figure
 
2.23). Given the sigma antibonding character of the initially generated 
4
T
2
 
state and the 
significant displacement along these types of coordinates on the 
4
T
2
 

2
E trajectory it also makes intuitive 
sense that these modes involve 
significant distortion of the metal
-
ligand bond distance. These also match 
the assignments from Schrauben
10
 
and Foszcz
12
 
f
or these vibrational modes. In the lower frequency   
193
 
cm
-
1
 
mode the entire ligand pushes toward the metal center while the two oxygen atoms splay outwards. 
In the 461
 
cm
-
1
 
mode, eac
h ligand compresses to make a smaller footprint and decrease the bite an
gle of 
the ligand. The distance between the metal center and the central carbon atom on each acac ligand is largely 
unchanged. Furthermore, because neither of these vibrational modes i
nvolve the oxygen atoms moving 
directly towards the metal center, but ra
ther involve the change in bight angle to some degree, there is also 
a small torsional element to each of these vibrational modes. The simplest way to view this torsional element 
is to
 
view each vibrational mode along the 
C
3
 
rotation axis as shown in Figu
re
 
2.23. If a triangle is formed 
with each of the three oxygen atoms nearest to the observer at a vertex and a separate triangle is formed 
with the three oxygen atoms furthest from the
 
observer, it can be seen from the displacement vectors that 
the oxygen 
atoms that form each triangle are displaced such that each triangle is rotated in a different 
direction, i.e. the triangles are rotated so that they approach eclipsing. This torsional 
element is likely a 
result of tethering the oxygen atoms together. 
 
It s
eems likely that a twisting motion would be effective in relaxing the Laporte selection rule and 
thus enhance the efficiency of an interconfigurational 
d
-
d
 
transition. Indeed, evidence
 
of the importance of 
torsional motion in transition metal complexes has
 
been observed in several systems. Endicott and 
coworkers have shown trigonal twisting distortions to facilitate 
2

4
A
2
 
ISC in octahedral Cr(III) 
complexes.
38,3
9
 
Additionally, Purcell has proposed that Bailar and Ray
-
Dutt twisting deformations may be 
involved in the race
mization dynamics of [Fe(phen)
3
]
2+
 
(where phen = 1,10
-
ortho
-
phenanthroline) by 
expediting 
1
A
1
 

5
T
2
 
ISC.
40
 
These results 
were supported theoretically by Vanquickenborne and Pierloot 
who used the angular overlap model to show Bailar and Ray
-
Dutt twisting motions are coupled to spin state 
conversion in strong field six coordinat
e Fe(II) species
41
 
and experimentally by McCusker 
et al.
42
 
who 
 
61
 
showed that 
1
A
1
 

5
T
2
 
conversion is tied to mobility about the 
C
3
 
rotation axis in a series of Fe(II) spin 
crossover compounds. Finally, Hörner and coworkers have shown that both trigonal twisting and ligand 
breathing motions are coupled to the 
1
A
1
 

5
T
2
 
conversion in co
nstrained spin crossover Fe(II) systems.
43
 
 
Conclusions on Cr(acac)
3
 
 
The previous two sections focusing on Cr(aca
c)
3
 
have answered several of the open questions 
brought up at the end of the summary of previous results. By analysis of the visible
 
pump visible probe TA 
data following 400 nm excitation into the high energy 
4
T
1
 
state, it was shown that vibrational relaxa
tion is 
not occurring with a 1 ps time constant as was originally reported by the McCusker group,
7
 
but rather a 
mechanism proposed by Kunttu and coworkers
14
 
inv
olving back
-
intersystem crossing followed by internal 
conversion is more appropriate. Two coherently active symmetric metal
-
ligand b
ond stretching vibrations 
were observed in single wavelength TA data following excitation into the 
4
T
2
 
excited state. The 46
1 cm
-
1
 
mode with 1130 fs dephasing time was shown to dephase by a pure dephasing mechanism on the 
2
E excited 
state. The following sections will seek to clarify what effects various substitutions on the acac ligand have 
on the photophysical dynamics followi
ng exc
itation into the 
4
T
2
 
excited state with an emphasis on 
characterizing the coherent nuclear motions of each compound. This work will first investigate the 

3
 
derivatives which have been substituted in the 3
-
position of 
the ac
ac 
backbone. Presumably, since substitutions in this position can couple into the pi
-
system of the acac ligand, 
Figure
 
2.24.
 

3
 
derivatives to be studied in this report.
 
 
 
62
 
this should afford a glimpse into the effect of electronic perturbations to the coherently active vibrational 
modes in this system. Additi
onally
, since the report of Schrauben 
et al.
10
 
proposed that the sterically 
encumbered 
t
-
butyl substitution had major photophysic
al ram
ifications, a series of increasingly sterically 
hindered molecules, building up to the 
t
-
butyl substituted acac will be studied. Finally, a molecule with 
both an electronic and a steric perturbing substitution will be investigated. A full list of the
 
molec
ules 
studied in this chapter can be seen in Figure
 
2.24.
 
3
-
Substituted Derivatives of Cr(acac)
3
 
2.4.2
.
 
Cr(3
-
NO
2
ac)
3
 
 
The photophysical characterization of Cr(3
-
NO
2
ac)
3
 
was conducted in the dissertation work of 
previous McCusker group member Joel Sch
rauben
15
 
and the full spectrum t
ransient absorption data from 
his work can be seen in Figure 2.11. The 1.65 ps narrowing and shifting of the transient spectrum was 
originally assigned to vibrational relaxation on the 
2
E state, however in light of the more recent w
ork of 
Kunttu and cowork
ers
14
 

reassign this as 
4
T
2
 

4
A
2
 
IC following bISC from the 
2
E state similar to the mechanism described for the 
unsubstituted parent compound. The remai
nder of the discussion f
or this compound will be centered around 
the coherent vibrational dynamics. The coherent vibrational modes of Cr(3
-
NO
2
ac)
3
 
are summarized in 
Table 2.3, with example datasets shown in Figure
 
2.25. Interestingly, now only a single coh
erent vibrational 
 
Pump (probe), nm
 
So
lvent
 
Frequency, cm
-
1
 
(Damping Time, fs)
 
600 (608)
 
Dioxane
 
151 (178)
 
600 (600)
 
Dioxane
 
153 (182)
 
600 (600)
 
Dioxane
 
146 (192)
 
600 (600)
 
Dioxane
 
151 (203)
 
600 (592)
 
Dioxane
 
149 (184)
 
600 (592)
 
Dioxane
 
145 (193)
 
600 (592)
 
THF
 
154 (171)
 
Average Frequen
cy, cm
-
1
 
150
 
Standard Deviation, cm
-
1
 
3
 
Average Damping Time, fs
 
186
 
Standard Deviation, fs
 
11
 
 
Table
 
2.3.
 
Summary of the coherent vibrational modes of Cr(3
-
NO
2
ac)
3
 
following 
4
A
2
 

 
4
T
2
 
excitation. All measurements performed in Dioxane used 4 and 0.4 

J pump and probe pulses 
respectively while the THF measurement employed 4.7 and 0.5 

J pump and probe
 
pulses. The 
dioxane and THF experiments had 40 and 35 fs IRFs respectively.
 
 
 
63
 
 
 
Figu
re 2.25. 
Representative TA data on Cr(3
-
NO
2
ac)
3
 
following 600 nm excitation into the 
4
T
2
 
state 
in 1,4
-
dioxane while monitoring a 4 nm slice of the probe spectrum at (a) 608 nm, (c) 600 nm, and 
(e) 592 nm. FFT and LPSVD analyses co
rresponding to the (b) 608
 
nm, (d) 600 nm, and (f) 592 nm 
data sets.
 
a)
 
c)
 
e)
 
f)
 
d)
 
b)
 
 
64
 
mo
de is observed with a 150 cm
-
1
 
frequency and 186 fs dephasing time, the latter of which is much shorter 
than the 350 fs dephasing time of the low frequency, 193 cm
-
1
 
mode, of Cr(acac)
3
. At this point it is difficult 
to speculate why this might be so. The r
apid dephasing time observed b
oth here and in the parent Cr(acac)
3
 
molecule certainly fit with the hypothesis that rapid dephasing times may indicate involvement of a 
vibrational mode in an ultrafast change in electronic state, however there is simply not 
enough evidence at 
present to 
confirm this. It is just as plausible that this mode undergoes rapid intramolecular vibrational 
redistribution similar to what was observed for Cr(acac)
3
 
by Kunttu and coworkers.
14
 
In this light, the faster 
dephasing time here coul
d be rationalized in terms of the loose bolt effect
44
. Here, the nitro group may be 
expected to rattle and rapidly dissipate ene
rgy akin to a loose bolt on a moving part of a mach
ine. At this 
point, both explanations seem plausible, so any further judgement must wait until more evidence is gained 

3
 
systems.
 
 
 
To visualize the 150 cm
-
1
 
mode of Cr(
3
-
NO
2
ac)
3
 
we can again turn to density functional t
heory 
calculations. In the parent Cr(acac)
3
 
we were able to confirm that the 461 cm
-
1
 
mode was observed in the 
2
E excited state and could only speculate that the 193 cm
-
1
 
mode was as well. Given that the 
2
E excited state 
and the 
4
A
2
 
ground state have the s
ame (
t
2g
)
3
 
orbital population, the geometry and vibrational frequencies 
are likely very similar between the two states. Thus, the frequency calculation for this compound was 
Calc. 148 cm
-
1
 
Exp. 150 cm
-
1
 
Figure 2.26.
 
Nuclear deformation associated with the 150 cm
-
1
 
vibrational mode of Cr(3
-
NO
2
ac)
3
.
 
 
65
 
performed on the ground electronic state for ease of computation. These calculati
ons reveal a   148
 
cm
-
1
 
metal ligand stretching vibration in good agreeme
nt with the 150 ± 4 cm
-
1
 
frequency observed in the TA 
experiments. A picture of this vibrational mode can be seen in Figure
 
2.26. This vibrational mode is 
essentially identical to the
 
193
 
cm
-
1
 
mode of the Cr(acac)
3
 
control molecule suggesting that this par
ticular 
vibration is essential for producing the geometric changes associated with 
4
T
2
 

 
2
E conversion.
 
2.4.3. Cr(3
-
Clac)
3
 
 
Cr(3
-
Clac)
3
 
forms light brown needle like crystals. The groun
d state absorption spectrum for this 
compound can be seen in Figure
 
2.27. The 
4
A
2
 

4
T
2
 
absorption band (564 nm, 79 M
-
1
cm
-
1
) is only slightly 
redshifted fro
m the 560 nm maximum in Cr(acac)
3
,
15
 
consistent with the increased 

-
basicity of the chloro
-
substituted ligand raising the energy of the 
t
2g
 
orbital set and thus marginally decreasing the ligand field 
strength of the compound by ~125 cm
-
1
. Based on analogous assignments on Cr(acac)
3
, the bands centered 
at 272 and 358
 
nm c
a
n be assigned to an intraligand 




*
 
transition and a LMCT type transition 
respectively.
15
 
 
Early time full spectrum TA data on Cr(3
-
Clac)
3
 
following 600 nm excitation into red edge of the 
4
A
2
 

4
T
2
 
ground state absorption band can be seen in Figure
 
2.28 along with the decay associate
d
 
difference spectra (DADS) and evolution associated difference spectra (EADS) resulting from a global fit 
of the full spectrum data to a sequential biexponential decay using OPTIMUS fitting software.
22
 
This data 
Figure
 
2.27. 
Absorption spectrum of Cr(3
-
Clac)
3
 
in dichlorome
thane.
 
 
66
 
set shows a clear decay in the blue and red edges o
f the transie
nt absorption band similar to what was 
observed for the unsubstituted Cr(acac)
3
 
parent molecule. Global fitting of the data to two sequentially 
decaying kinetic components reveals a 1.07 ps kinetic component responsible for the spectral change
s at 
early ti
mes (Figure
 
2.28 b), again similar to what was observed for Cr(acac)
3
. Given the similar kinetics 
and observations between the two compounds, the ~1.1 ps kinetic component of Cr(3
-
Clac)
3
 
can also be 
assigned primarily to 
4
T
2
 

4
A
2
 
IC following
 
ISC and bISC. The EADS, describing the transient spectra 
before and after the 1.1 ps process is complete, can be seen in Figure
 
2.28 c. EADS2 resembles the transient 
a)
 
b)
 
c)
 
Figure
 
2.28.
 
(a) Early time full spectrum TA data on Cr(3
-
Clac)
3
 
in toluene following 600
 
nm 
excitation into the 
4
T
2
 
state. (b) Decay associated difference spectrum (DADS) associated with the 
global fit of the TA data to a sequential bi
exponential decay. The long
-
time
 
offset was modeled by 
fixing a 1 ns exponential decay to describe the slower ground state recovery process. (c) Evolution 
associated difference spectra (EADS) associated with the global fit.
 
 
67
 
spectrum at longer delays and hence shows the spectrum of the 
2
E excited sta
te. At early 
times, there is an 
equilibrium between the 
2
E and 
4
T
2
 
states and hence EADS1 is representative of a superposition of the 
transient spectra of the 
2
E and 
4
T
2
 
excited states. EADS1 is also quite flat, reminiscent of the DADS 
observed following 4
00
 
nm excitat
ion of Cr(acac)
3
, supporting the assessment that both excited state 
populations are present at early times.
 
 
 
Single wavelength datasets can be seen in Figure
 
2.29. Here, the datasets probing 
at 490 and 
550
 
nm, representing the blue edge of the transient 
absorption band, show an average time constant of 1.24 
± 0.07 ps, in good agreement with the full spectrum fit and quite similar to the time constant observed in 
Figure 2.29.
 
Representative sin
gle wavelength TA data on Cr(3
-
Clac)
3
 
in toluene following 600
 
nm 
excitation into the 
4
T
2
 
state whil
e monitoring at (a) 490 nm, (b) 550 nm, and (c) 650 nm. The two 
bluer wavelengths show ~1.2
 
ps fits while the 650 nm probe data may show additional dynamics
.
 
a)
 
b)
 
c)
 
 
68
 
the unsubstituted Cr(acac)
3
 
com
pound.  A single exponential fit of the 650
 
nm probe data (Figu
re
 
2.29 c) 
showed a longer time constant compared to the other two probe wavelength regions. A biexponential fit of 
the data looked better and gave 850 fs and 4.8 ps time constants which are re
miniscent of the internal 
conversion and vibrational cooling ti
me constants of Cr(acac)
3
, however considering the quality of the data, 
it is difficult to say whether a biexponential fit is appropriate. Given the larger amplitude decay in the bluer 
portion
 
of the spectrum, however, the time constants observed at 490 n
m and 550 nm are likely more 
representative of the timescale for IC.
 
 
Coherent oscillations of Cr(3
-
Clac)
3
 
were observed in a similar manner as the previously studied 
compounds. A summary of the results collected in toluene can be seen in Table 2.4 and re
presentative 
datasets can be s
een in Figure
 
2.30. The low frequency modes that were picked up during the LPSVD 
analysis of two of the datasets in Table
 
2.4 were included for the sake of transparency, however, it is 
expected that they are present to correct
 
for poor exponential fitting 
of the LPSVD program and do not 
represent anything physically meaningful. Additionally, a 449
 
cm
-
1
 
oscillatory component was observed in 
a single dataset. It is possible that this corresponds to a vibrational mode reminiscent 
of the 461 cm
-
1
 
mode 
of Cr(aca
c)
3
, however, since it was only observed once, it is not worth spending too much time speculatin
g 
over. Clearly the most robustly observed vibrational mode of these data is the 149 cm
-
1
 
mode with 1100 fs 
 
Pump (probe)
 
Frequency, cm
-
1
 
(Damping Time, fs)
 
600 (608)
 
 
 
144 (873)
 
449 (581)
 
600 (608)
 
45 (
-
13640)
 
93 (5450)
 
150 
(953)
 
 
600 (600)
 
65 (32560)
 
 
149 (1100)
 
 
600 (600)
 
 
 
150 (1224)
 
 
600 (592)
 
 
 
151 (1426)
 
 
600 (592)
 
 
 
149 (1015)
 
 
600 (
584)
 
 
 
151 (1100)
 
 
Average Frequency, cm
-
1
 
149
 
 
Standard Deviation, cm
-
1
 
2
 
 
Average Damping, fs
 
1100
 
 
Standard 
Deviation, fs
 
180
 
 
 
Table 2.4.
 
Summary of observed coherent vibrational modes of Cr(3
-
Clac)
3
 
in toluene following 
excitation into the 
4
T
2
 
s
tate. The sub
-
100 cm
-
1
 
oscillatory components picked up by the LPSVD fitting 
program were included for transparency, however, they are likely due to poor exponential fitting of 
the data and not to coherent vibrational motion. All experiments employed 4 and
 
0.4 

J pump and 
probe
 
pulses respectively with a 45 fs IRF.
 
 
69
 
 
 
a)
 
b)
 
c)
 
d)
 
f)
 
e)
 
Figure
 
2.30.
 
Representative TA traces and FFT and LPSVD analyses of coherent oscillations of Cr(3
-
Clac)
3
 
in toluene following excitation into the 
4
T
2
 
state. Coherent osc
illations were monitored at (a) 608 
nm, (c) 600 nm, and (e) 592 nm. Corresponding FFT a
nd LPSVD vibrational spectra are shown in (b), 
(d), and (f).
 
 
70
 
dephasing time. DF
T frequency calculations on the ground electronic state show a 153
 
cm
-
1
 
symmetric 
metal ligand breathing mode (Figure
 
2.31) similar to the 193 cm
-
1
 
mode observed for Cr(acac)
3
. It is 
interesting that each of the three compounds discussed thus far appear to
 
exhibit identical excited state 
decay 
mechanisms with the same coherent nuclear motion yet there is a wide range of dephasing times. 
This suggests that the dephasing times may not be as strong of an indicator of the involvement of a 
vibrational mode in a
n ultrafast photophysical transformation
 
as was suggested by the hypothesis derived 

10
 
A more complete
 
discussion of the implications of these
 
dephasing times 
will be provided after the presentation of each of the 3
-
substituted compound results
 
2.4.4. Cr(3
-
Brac)
3
 
 
General photophysical characterizations of Cr(3
-
Brac)
3
 
were conducted in the dissertation wor
k of 
Joel Schrauben,
15
 
and the e
arly time full spectrum TA data from his
 
work can be seen in Figure
 
2.11a. As 
such, only work pertaining to the early time dynamics of this compound will be reported here.
 
To start, 
another full spectrum TA dataset was obtained, which can be seen in Figure
 
2.
32
a with the DADS and 
EADS resulting 
from a sequential biexponential global fit using OPTIMUS
22
 
fitting software. The TA data 
are qualitatively similar to that collected by Joel on the same compound (Figure 2.11a)
15
 
and the TA data 
on Cr(acac)
3
 
(Figure 2.3).
7
 
Additionally, the global fit returned a ~1 ps kinetic component responsible for 
Calc: 153 cm
-
1
 
Exp: 149 cm
-
1
 
Figure
 
2.31.
 
Nuclear deformations associated with the 149 cm
-
1
 
vibrational mo
de of Cr(3
-
Clac)
3
.
 
 
71
 
the decay in the blue and red edges of the transient absorption spectrum. It is again evident from the 
qualit
atively identical spectral evolution and
 
kinetics that Cr(3
-
Brac)
3
 
undergoes a similar decay 
mechanism as the Cr(acac)
3
 
parent molecule involving rapid ISC/bISC and that th
e ~1 ps kinetic component 
can be attributed laregely to 
4
T
2
 

4
A
2
 
IC.
 
 
 
In addition
 
to the full spectrum dataset just discussed, the kinetics were also investigated using 
single wavelength measurements to get a more precise view of the time constants assoc
iated with excited 
state decay. Representative datasets can be seen in Figure
 
2.33.
 
While the fits on the blue edge of the 
a)
 
b)
 
c)
 
Figure
 
2.32.
 
(a) Early time full spectrum TA data on Cr(3
-
Brac)
3
 
in toluene following 600
 
nm 
excitation into the 
4
T
2
 
state. The 570
-
630
 
nm spectral region was omitted due to large contamination 
from pump scatter. (b) Decay a
ssociated difference spectrum (DADS) associated with the global 
fit of the TA data to a sequential biexponential decay. The long
-
time offset was mod
eled by fixing 
a 1 ns exponential decay to describe the slower ground state recovery process. (c) Evolution 
associated difference spectra (EADS) associated with the global fit.
 
 
72
 
transient spectrum show similar time constants, the fit of the 650
 
nm probe data shows a time constant that 
is quite different and a 
does not match the data as well. A biexponential fit of the 650
 
nm data (not shown)
 
produces a better fit to the data and produces ~1 and 6 ps time constants similar to the internal conversion 
and vibrational relaxation time 
constants measured for Cr(acac)
3
.
14
 
Just like in the Cr(3
-
Clac)
3
 
650
 
nm 
probe dataset (Figure
 
2.29c) it may be the cas
e that more than one process is contributing to the spectral 
evolution here. Without higher quality data in this region,
 
however, it is difficu
lt to say with certainty, thus 
the 650
 
nm probe datasets will not be included in the averaging. The average time 
constant for
 
4
T
2
 

4
A
2
 
IC measured at 490 nm and 550 nm probe wavelengths is 1.2 ± 0.1
 
ps, which is again, very similar to each 
of the previous
ly discussed members of the series.
 
a)
 
b)
 
c)
 
Figure
 
2
.
33
.
 
Representative single wavelength TA traces collected on Cr(3
-
Brac)
3
 
in toluene 
following excitation into the 
4
T
2
 
state and monitoring at (a) 490 nm, (b) 550 nm, 
and (c) 650 nm.
 
 
73
 
 
Next, we can turn to the coherent dynamics of Cr(3
-
Brac)
3
. Table
 
2.5 summarizes the coherent
 
vibrational modes of Cr(3
-
Brac)
3
 
observed in toluene and dich
loromethane following 600
 
nm excitation 
into the 
4
T
2
 
state. In addition, representative datasets can be seen in Figure 2.34. Both show a single solute
-
attributable oscillation with a 123
 
± 1 cm
-
1
 
frequency and a 2.5 ± 0.3 ps dephasing time. DFT calculations 
on the ground electronic state show a 122 cm
-
1
 
symmetric breathing mode (Figure
 
2.35) akin to the low 
frequency mode observed in the Cr(acac)
3
 
control molecule.
 
Figure
 
2.34.
 
Representative TA traces and FFT and LPSVD analyses of coherent oscillations of Cr(3
-
Brac)
3
 
in toluene following excitation into the 
4
T
2
 
state. Coherent oscillations were monitored at (a) 
608 nm and (c) 5
92 nm. Corresponding F
FT and LPSVD vibrational spectra are shown in (b) and (d).
 
a)
 
b)
 
d)
 
c)
 
 
74
 
 
On an experimental note, this
 
compound tended to crash out where the laser hit the sample. None 
of the presented results are expected to be greatly affected,
 
however, care had to be taken to prevent a 
buildup of precipitate reducing the amount of light transmitted through the sample.
 
This was achieved by 
caref
ul consideration of how many data points to include in each scan as well as physically translating the 
cuvette after each scan.
 
2.4.5. Cr(3
-
Iac)
3
 
 
Cr(3
-
Iac)
3
 
forms light brown needle like crystals. The ground state absorption spe
ctrum for this 
compound can be seen in Figure
 
2.36. The 
4
A
2
 

4
T
2
 
absorption band (564
 
nm, 95 M
-
1
cm
-
1
) is redshifted 
Calc: 122 cm
-
1
 
Exp: 123 cm
-
1
 
Figure
 
2.35.
 
Nuclear deformations associated with the 123 cm
-
1
 
vibrational mode of Cr(3
-
Brac)
3
.
 

pu
, nm
 
(
E
pu
, 

J)
 

pr
, nm
 
(
E
pr
, 

J)
 
IRF, fs
 
Solvent
 
F
requency, cm
-
1
 
(Damping Time, fs)
 
600 (
4.7
)
 
608
 
(0.5)
 
35
 
Toluene
 
121 (2780)
 
600 (
5
)
 
608
 
(0.5)
 
45
 
Toluene
 
122 (1390)
 
600 (
3.3
)
 
600
 
(0.4)
 
55
 
Toluene
 
125 
(2580)
 
600 (
4.7
)
 
592 (0.5)
 
35
 
Toluene
 
123 (2450)
 
600 (
3.3
)
 
592 (0.4)
 
55
 
Toluene
 
122 (2040)
 
600 (
3.3
)
 
600 (0.4)
 
55
 
DCM
 
123 (2920)
 
600 (
5
)
 
600 (0.4)
 
40
 
DCM
 
122 (3610)
 
Average Frequency, cm
-
1
 
123
 
Standard Deviation, cm
-
1
 
1
 
Average Damping Time, fs
 
2500
 
Standard Deviation, fs
 
300
 
 
Table
 
2.5.
 
Summary of the coherent oscillations of Cr(3
-
Brac)
3
 
followi
ng excitation into the 
4
T
2
 
state. 
Experimental parameters such as pump and probe wavelengths and pulse energies as well as IRF 
FWHMs are also included for each experiment.
 
 
75
 
from the 560
 
nm maximum of Cr(acac)
3
 
by the same amount as the Cr(3
-
Clac)
3
 
analogue. Given the 
similarity in the remaining spectr
al featur
es to those in Cr(3
-
Clac)
3
, we can again assign the bands centered 
at 272 and 363
 
nm to an intraligand 

 


*
 
transition and an LMCT type transition respectively. The 
redshift of the LMCT transition relative to the chloro
-
substituted derivative su
ggest tha
t the iodo
-
group is 
a better 

-
donor. A crystal structure of Cr(3
-
Iac)
3
 
was also obtained (Figure
 
2.37.) and solved by group 
member Sara Adelman and Professor Richard Staples of Michigan State University. The coordinates can 
be found in the append
ix.
 
 
Earl
y time full spectrum TA data on Cr(3
-
Iac)
3
 
following 600
 
nm excitation into the red edge of 
the 
4
A
2
 

 
4
T
2
 
ground state absorption band can be seen in Figure
 
2.38 along with the decay associated 
difference spectra (DADS) and evolution associated di
fference 
spectra (EADS) resulting from a global fit 
of the full spectrum data to a sequential biexponential decay using OPTIMUS fitting software.
22
 
This dataset 
again exhibits a decay in the blue edge of the spectrum similar to Cr(acac)
3
 
and its derivatives that ha
ve 
alread
y been discussed. The less apparent decay on the red edge of the spectrum is likely due to the smaller 
amplitude decay in that wavelength region in conjunction with the lower stability of the white light 
continuum in that wavelength region. Nevert
heless, a
 
decay on the red edge of the transient spectrum can 
be observed when using the more precise single wavelength measurements (Figure
 
2.39). Before 
Figure
 
2.36.
 
Absorption spectrum of 
Cr(3
-
Iac)
3
.
 
Figure
 
2.37.
 
C
rystal Structure of 
Cr(3
-
Iac)
3
 
collected by Sara Adelman and 
solved by Professor Richard Staples of 
Michigan State University.
 
 
76
 
progressing to the single wavelength data, however, it is worth noting the 1.2
 
ps decay acquired from the 
glo
bal fit of the full spectrum TA data (Figure
 
2.38) which qualitatively matches both the unsubstituted 
Cr(acac)
3
 
control molec
ule and each of its 3
-
substituted derivatives discussed previously. Like these other 
compounds, this 1.2 ps decay process can readi
ly be assigned to 
4
T
2
 

4
A
2
 
IC following equilibration of 
the 
4
T
2
 
and 
2
E states.
 
 
 
Next, in search of a more precise measurem
ent of the kinetics following 
4
A
2
 

4
T
2
 
excitation, 
single wavelength experiments were performed. Representative datasets are shown
 
in Figure
 
2.39 in which 
a)
 
b)
 
c)
 
Figure
 
2.38.
 
(a) Early time full spectrum TA data on Cr(3
-
Iac)
3
 
in toluene following 600
 
nm excitation 
into the 
4
T
2
 
state. (b) Decay ass
ociated difference spectra (DADS) associated with the global fit of the 
TA data to a sequential biexponential decay. The long
-
time o
ffset was modeled by fixing a 1
 
ns 
exponential decay to describe the slower ground state recovery process. (c) Evolution ass
ociated 
difference spectra (EADS) associated with the global fit. Distortions around 600
 
nm in the data, DADS, 
and EADS are due to r
esidual pump scatter.
 
 
77
 
the decay was monitored on the blue and red wings of the transient spectrum. We again see that the kinetics 
measured on the blue edge of the transient s
pectrum appear somewhat different than on the kinetics 
measured on the red edg
e at 650
 
nm which appear slower. This situation was present for each of the halogen 
derivatives presented earlier. We will thus continue to ignore the kinetics monitored at 650
 
n
m and report 
a time constant derived from the average of the 490
 
nm and 550
 
nm
 
experiments. This results in a 1.3 ± 
0.1
 
ps time constant for 
4
T
2
 

4
A
2
 
IC in good agreement with the results of the global fit of the full spectrum 

3
 
systems discussed thus far.
 
 
 
With an
 
idea of the kinetics for population relaxation in hand we can next turn
 
to the coherent 
vibrational dynamics in Cr(3
-
Iac)
3
. Table
 
2.6 summarizes the coherent oscillations of Cr(3
-
Iac)
3
 
following 
 
a)
 
b)
 
c)
 
Figure
 
2.39.
 
Representative single wavelength TA data on Cr(3
-
Iac)
3
 
in toluene fol
lowing 
600
 
nm 
excitation into the 
4
T
2
 
state while monitoring at (a) 490
 
nm, (b) 550
 
nm, and (c) 650
 
nm.
 
 
78
 
 
 
Figure
 
2.40.
 
Representative TA traces and FFT and LPSVD analyses of coherent oscillations of 
Cr(3
-
Iac)
3
 
in toluene following exc
itation into the 
4
T
2
 
state. Co
herent oscillations were monitored 
at (a) 608 nm, (c) 600 nm, and (e) 592
 
nm. Corresponding FFT and LPSVD vibrational spectra are 
shown in (b), (d), and (f).
 
a)
 
c)
 
e)
 
f)
 
d)
 
b)
 
 
79
 
600
 
nm excitation into the 
4
T
2
 
state and Figure
 
2.40 sh
ows representative datasets. Clearly, the primary 
c
oherently active vibrational mode is a 101 ±
 
2 cm
-
1
 
vibration with a 2.4 ± 0.6
 
ps dephasing time. This is 
readily apparent in the TA data provided in Figure
 
2.40 as a large amplitude oscillation that lasts
 
for several 
picoseconds. On two occasions the LPSV
D program picked up additional frequencies that were not 
observed in the neat solvent control experiment. Given the irreproducibility of these values it is highly 
unlikely that they correspond to actual so
lute vibrational modes. The 55 cm
-
1
 
oscillatory com
ponent 
obtained during one experiment is likely due to poor exponential fitting of the LPSVD program. It is also 
possible that the ~200
 
cm
-
1
 
component is a weakly Raman active out of plane bending mode of 
the methyl 
group of the toluene solvent.
 
 
 
We can a
gain turn to density functional theory calculations to gain insight into the nature of this 
coherently active vibrational mode. And, just like the previous members of the 3
-
substituted series of 

3
 
compounds, we can take advantage of the similarity
 
of the electron distribution between the 
4
A
2 
and 
2
E ground and excited states and perform the frequency calculations on the more computationally 
accessible ground electronic state. The resulting 102
 
cm
-
1
 
calculated symmetric breathing mode 
(Figure
 
2.41) s
hows excellent agreement with the experimental 101 ±
 
1
 
cm
-
1
 
frequency and is consistent 
with the vibrational modes observed for every member of the series thus far.
 
 
Probe, nm
 
Solvent
 
Frequency, cm
-
1
 
(Damping Time, fs)
 
615
 
Toluene
 
55 (206
0)
 
103 (3450)
 
204 (1420)
 
608
 
Toluene
 
 
101 (1800)
 
 
608
 
Toluene
 
 
100 (1780)
 
 
600
 
Toluene
 
 
102 (2520)
 
 
600
 
Toluene
 
 
103 (3090)
 
 
592
 
Toluene
 
 
102 (2130)
 
181 (200)
 
592
 
Toluene
 
 
103 (2020)
 
 
600
 
DCM
 
 
96 (2190)
 
 
Average Frequency, cm
-
1
 
 
101
 
 
Standard Devi
ation, cm
-
1
 
 
2
 
 
Average Damping Time, fs
 
 
2400
 
 
Standard Deviation, fs
 
 
600
 
 
 
Table
 
2.6.
 
Summary of the coherent oscillations of Cr(3
-
Iac)
3
 
following 600
 
nm excitation into the 
4
T
2
 
state. Each measurement employed 4 and 0.4 

J pump and probe pulse ener
gies respectively 
with a 40 fs IRF.
 
 
 
80
 
 
On an experimental note, Cr(3
-
Iac)
3
 
h
ad the tendency to crash out of solution where the laser hit. 
As such, care was taken when deciding the number of data points in each s
can to prevent a significant 
buildup of precipitate reducing the transmission of the probe beam through the sample. The c
uvette was 
also translated after each scan when performing these experiments. Additionally, the sample was observed 
to decompose after 
a while as the solution began to turn brown. This was marked by an absorption feature 
growing in at ~500
 
nm that is likel
y a result of a dissociated iodine species. Given the excellent agreement 
between the calculated and experimental frequencies and the o
bservation that the halogen atoms seem to 
dramatically and systematically change the vibrational frequency from ~190 cm
-
1
 
in unsubstituted Cr(acac)
3
 
to ~101 cm
-
1
 
in Cr(3
-
Iac)
3
, it is expected that this decomposition is not effecting the results shown here 
to 
any significant degree and that the observed oscillatory component can be attributed entirely to Cr(3
-
Iac)
3
.
 
2.4.6. Co
nclusions on 3
-

3
 
compounds
 
 
Now that the results on each member of the 3
-

3
 
series has been sh
own, we can 
gain further insight by comparing the results on each compound. First, it is clear by the qualitatively simil
ar 
spectral evolution across the series that the excited state decay mechanism is identical for each member, 
i.e. that following excita
tion into the 
4
T
2
 
state, the molecule rapidly undergoes ISC to the doublet manifold 
followed by bISC to create a quasi
-
eq
uilibrium between the 
4
T
2
 
and 
2
E states and the 
4
T
2
 
population can 
Calc: 102 cm
-
1
 
Exp: 101 cm
-
1
 
Figure
 
2.41.
 
Nuclear deformations associated with the 101 cm
-
1
 
vibrational mode of Cr(3
-
Iac)
3
.
 
 
81
 
undergo IC to reform the ground electronic state with a time constant of ~ 
1
 
ps. This further illustrates the 
insensitivity of the decay mechanism with respect to synthetic modifications.
 
 
Perhaps unsurprisingly, given the identical decay mechanisms across this series of 3
-
substituted 

3
 
compounds, the coherent vibrationa
l mode that is active in each compound is the same type of 
metal ligand breathing mode (Figure
 
2.42), which, due 
to the tethering of the oxygen atoms, results in a 
small torsional element as well. When comparing the dephasing times across this series to th
e IC time 
constants, we can gain important information regarding the electronic state that these vibrational mode
s are 
active on. From the Feynman diagrams discussed in section 2.4.1 we know that the coherences sampled 
during the TA experiments must be fro
m an electronic excited state (
4
T
2
 
or 
2
E). Given that the internal 
conversion time constant is smaller than the d
ephasing time for the heavier halogen
-
substituted derivatives 
then we can say unequivocally that this vibrational mode is not sampled from the 
4
T
2
 
state and that all of 
the coherently active vibrational modes (including the 461
 
cm
-
1
 
mode of the control mol
ecule discussed 
earlier) are sampled in the 
2
E excited state. This is further supported by the good agreement between 
experiment and theory whi
le using the 
4
A
2
 
ground state (a proxy for the 
2
E state) for frequency calculations. 
Additionally, these results 
again indisputably disprove the vibrational relaxation assignment for the ~1
 
ps 
process originally assigned by Juban and McCusker
7
 
as vibrational relaxatio
n would result in dephasing of 
this mode and, according to equation 2.8, could not be faster than the observed vibrational dephasing.
 
Cr(acac)
3
 
193 (10) cm
-
1
 
340 (140) fs
 

IC
 
= 1.1 ps
 
Cr(3
-
NO
2
ac)
3
 
150 (4) cm
-
1
 
186 (11) fs
 

IC
 
= 1.6 ps
 
Cr(3
-
Clac)
3
 
149 (2) cm
-
1
 
1100 (180) fs
 

IC
 
= 1.2 ps
 
Cr(3
-
Brac)
3
 
121 (1) cm
-
1
 
2500 (700) fs
 

IC
 
= 1.2 ps
 
Cr(3
-
Iac)
3
 
101 (2) cm
-
1
 
2400 (600) fs
 

IC
 
= 1.3 ps
 
Figure
 
2.
42
.
 
Summary of the resu
lts on the 3
-

3
 
series. Internal conversion 
lifetimes for Cr(acac)
3
 
and Cr(3
-
NO
2
ac)
3
 
were acquired from references 7 and 15 respectively.
 
 
82
 
 
We can n
ext bring our attention to the main question driving this body of work: can dephasing 
times be an indicator for i
nvolvement of a vibrational mode in an ultrafast photophysical transformation. It 
is clear that 
2
E state population is present at early times i
n each compound implying an ultrafast ISC. Yet, 
there also exists a wide range of dephasing times for the same vi
brational mode across the series (200
-
2500
 
fs) suggesting that the ISC rate and the dephasing times observed here are uncorrelated and further 
that dephasing times are a poor indicator for involvement in an ultrafast photophysical transformation in 
this sy
stem. It is important to stress, however, that this represents a single counter example and that this 
does not mean that in all cases electroni
c state changes cannot result in vibrational dephasing. The burden 
of proof of this statement has not been met by
 
this body of results. In fact, it may be the case that ISC does 
act to dephase these modes to some degree depending on the time constant for I
SC and the vibrational 
period. It could very well be the case that the ISC is so fast relative to the slow vibrat
ional periods of the 
heavier halogen derivatives that it is effectively instantaneous. It is easy to envision, however, that the 
shorter period
 
of the 193
 
cm
-
1
 
mode of the unsubstituted Cr(acac)
3
, may result in ISC events at staggered 
times across the ense
mble, effectively causing dephasing. Given that the dephasing time for each of these 
modes is slower than the ISC time, this would suggest that
 
if ISC is resulting in any dephasing, it does not 
result in the complete dephasing of the vibrational coherence 
and that additional factors must be at play. 
Unfortunately, without further evidence, it is impossible to know for certain at present. Indeed, 
intramolecular vibrational redistribution is also a plausible explanation for the short dephasing time in this 
sy
stem. Kunttu and coworkers
14
 
have already observed a ~300
-
700 fs IVR component in the ground 
electronic state of Cr(acac)
3
, Al(acac)
3
, and Fe(acac)
3
 
suggesting that rapid IVR may be inherent to the 
M(acac)
3
 
molecular structure. The sensitivity of the magnitude of IVR observed in that study to the 
vibrational frequency may also explain the wide range of dephasing times across the series shown here. 
How
e
ver, without further evidence, we cannot say with certainty what exactly causes dephasing in this 
vibrational mode.
 
 
 
 
83
 
2.4.7. Effects of Increasing Steric Bulk
 
 

3
 
syst
em is 
rel
atively insensitive with respect to synthetic modifications and that each of the 3
-

3
 
compounds undergo the same decay mechanism. We can now turn our attention to the effect of increased 
steric bulk in the periphery of the mole
cule such
 
as with the Cr(TMHD)
3
 
molecule first studied on ultrafast 
timescales by Schrauben.
10,15
 
The transient difference spectrum for this compound (Figure
 
2.5) exhibited a 
~1.8
 
ps ri
se center
ed around 525
 
nm with a concomitant decay around 465
 
nm, which, based on the earlier 
discussion in section 2.4.1, we know is a manifestation of 
4
T
2
 

2
E ISC. It was rationalized that this could 
occur in one of two ways. In Case
 
1, we have a simila
r mechani
sm as the original one proposed by 
Schrauben.
10
 
Following 
4
A
2
 

 
4
T
2
 
excitation, the molecule slowly trickles down to the
 
2
E state
 
with a 1.8
 
ps 
time constant necessitating a slower 
4
T
2
 

 
4
A
2
 
internal conversion rate relative to the Cr(acac)
3
 
control 
molecule and more importantly a more than 18
-
fold increase in the ISC lifetime. This latter point on the 
remarkable change in 
the ISC l
ifetime seems somewhat unrealistic given earlier insensitivity of the decay 
Figure
 
2
.
43
.
 
Depiction of the two possible mechanisms for excited state decay at early times 
in 
C
r(TMHD)
3
. Mechanism
 
1 involves a slow 1.8
 
ps 
4
T
2
 

 
2
E ISC. Mechanism
 
2 involves a rapid 
equilibration of the 
4
T
2
 
and 
2
E states. As the molecule begins to vibrationally cool the 
2

4
T
2
 
equilibrium shifts to the left and results in a 1.8
 
ps ISC. IC within the quartet manifold can occur, 
although it is n
ot the dominant source of population relaxation. 
 
 
84
 

3
 
backbone. In the alternative proposed mechanism, 
Case
 
2, Cr(TMHD)
3
 
behaves very similar to the Cr(acac)
3
 
parent mole
cule. Fol
lowing 
4
A
2
 

4
T
2
 
excitation, 
the molecule undergoes rapid 
4
T
2
 

2
E ISC followed by rapid thermally activated 
2

4
T
2
 
bISC to create 
a quasi
-
equilibrium between these two states. If the IC from the 
4
T
2
 
state is slowed relative to Cr(acac)
3
 
then as the molecule vibra
tionally cools the thermally activated bISC from the doublet manifold will 
eventually become slower than th
e ISC from the 
4
T
2
 
state resulting a net 
4
T
2
 

2
E ISC being observed. This 
mechanism implies that the primary difference between Cr(acac)
3
 
and Cr(TMH
D)
3
 
is the slower IC from
 
the 
4
T
2
 
state. A summary of these two possible mechanisms for excited state decay
 
in Cr(TMHD)
3
 
at early 
times can be seen in Figure
 
2.43
.
 
 
 
It seems that the major difference between these two mechanisms lies in the ISC rate. If 
M
echanism
 
1 is truly active then this implies a substantial reduction in the ISC rate relative to 
Cr(acac)
3
, 
whereas if Mechanism 2 is correct then the differences between Cr(acac)
3
 
and Cr(TMHD)
3
 
are much more 
subtle, primarily focused on the IC rate from t
he 
4
T
2
 
state. To determine which of these mechanisms is most 
appropriate we can slowly build up t
o the Cr(TMHD)
3
 
compound and watch for changes in the dynamics. 
If a substantial change in the ISC rate is truly occurring this should be observed clearly in t
he progression 
of compounds depicted in Figure
 
2.24. In the following sections, we will thus inve
stigate the dynamics in 
Cr(HD)
3
 
and Cr(DMHD)
3
 
and compare them to what is known about Cr(acac)
3
 
and Cr(TMHD)
3
 
to create 
a physical model that represents the dy
namics in all systems.
 
 
2.4.8. Cr(HD)
3
 
 
Transient absorption data on Cr(HD)
3
 
at early times in dichloromethane following 600
 
nm 
excitation can be seen in Figure
 
2.44. Although the data is a bit noisy there is a clear decay on the blue edge 
of the transient
 
spectrum with no obvious rising feature indicating that there is no major s
hift in the ISC 
lifetime in this compound. Global fitting of the full spectrum dataset using the OPTIMUS fitting software,
22
 
reveals a 3
 
ps decay in both the red and blue edges. Overall, th
e spectral evolution appears very similar to 
Cr(acac)
3
, thus we can assign t
he 3
 
ps process to internal conversion from the 
4
T
2
 
state.
 
 
 
 
 
85
 
 
 
a)
 
b)
 
c)
 
Figure
 
2.44.
 
(a) Early time full spectrum TA data on Cr(HD)
3
 
in dichloromethane following 600
 
nm 
excitation into the 
4
T
2
 
state. (b) Decay associated difference spectra (DADS) a
ssociated with th
e 
global fit of the TA data to a sequential biexponential decay. The long
-
time offset was modeled by 
fixing a 1
 
ns exponential decay to describe the slower ground state recovery process. (c) Evolution 
associated difference spectra (EADS) a
ssociated with th
e global fit. Distortions in the transient 
spectrum near 600
 
nm are due to residual pump scatter.
 
 
86
 
 
We can acquire a more precise estimate of the IC lifetime by performing single wavelength TA 
expe
riments on Cr(HD
)
3
 
(Figure
 
2.45). These datasets qualitatively match the full spectrum datasets 
showing a 1.7 ± 0.4
 
ps decay on both the red and blue edges of the transient spectrum, assigned as internal 
conversion from the 
4
T
2
 
state. Unfortunately, no coh
erence of the so
lute was observed in this system, 
although it is likely that the same low frequency metal ligand stretching mode is active in this system as 
well (predicted to be ~182 cm
-
1
 
by DFT calculations). It is interesting to note that the IC lifetim
e in this 
compou
nd is longer than in Cr(acac)
3
, which was one of the required kinetic changes to achieve both 
Mechanisms 1 and 2 described earlier.
 
 
 
a)
 
b)
 
c)
 
Figure
 
2.45.
 
Representative single wavelength TA data on Cr(HD)
3
 
in acetonitrile (a) and 
dichloromethane (b and c) following 600
 
nm excitation in
to the 
4
T
2
 
state while monitoring at (a) 
490
 
nm, (b) 550
 
nm, and (c) 650
 
nm.
 
 
87
 
2.4.9. Cr(DMHD)
3
 
 
Full spectrum transient absorptio
n data on Cr(DMHD)
3
 
at early times in dichloromethane following 
600
 
nm
 
excitation into the 
4
T
2
 
state can be seen in Figure
 
2.46. Although the data is noisy, there is a clear 
decay in the blue edge of the spectrum and what appears to be a slight rise betwee
n ~550
-
590
 
nm. A global 
fit of the dataset to a biexponential decay re
veals 1.3
 
ps kinetic component which, according to the negative 
feature in the DADS, results in a slight rise between ~530
-
590
 
nm. This is also shown in the EADS. If this 
concomitant ris
e is truly present then this means Cr(DMHD)
3
 
is behaving similarly to 
the slightly more 
sterically encumbered Cr(TMHD)
3
 
compound.
 
 
a)
 
b)
 
c)
 
Figure
 
2.46.
 
(a) Early time full spectrum TA data on Cr(DMHD)
3
 
in dichloromethane follow
ing 
600
 
nm excitation into the 
4
T
2
 
state. (b) Decay associated difference spec
tra (DADS) associated 
with the global fit of the TA data to a sequential biexponential decay. The long
-
time offset was 
modeled by fixing a 1
 
ns exponential decay to describe the 
slower ground state recovery process. 
(c) Evolution associated difference spec
tra (EADS) associated with the global fit. Distortions in the 
transient spectrum near 600
 
nm are due to residual pump scatter.
 
 
88
 
 
To confirm these spectral features and to g
ain a better idea of the time constants associated with 
them, single wavelength measurements were performed on Cr(DMHD)
3
 
to monitor the kinetics at th
e red 
and blue edges and at the center of the transient spectrum (Figure
 
2.47.). The single wavelength dat
asets 
reveal a decay on the blue edge of the transient spectrum at 490
 
nm with concomitant rising features at the 
center (550
 
nm) and red edges of the
 
spectrum (586
 
nm), each with similar time constants (1.7 ±
 
2
 
ps). The 
single wavelength datasets thus con
firm the qualitative spectral changes observed in Figure
 
2.46 and by 
extension show that Cr(DMHD)
3
 
behaves similar to Cr(TMHD)
3
 
in terms of both spect
ral evolution and 
kinetics (

Cr(TMHD)3
 
= 1.8 ps)
10
. Coherent vibrational modes of Cr(DMHD)
3
 
were unable to be observed.
 
 
a)
 
b)
 
c)
 
Figure
 
2.47.
 
Representative single wavelength TA data on Cr(DMHD)
3
 
in dichloromethane (a and 
b) and acetonitrile (c)
 
following 600
 
nm excitation into the 
4
T
2
 
state while monitoring at (a) 
490
 
nm, (b) 550
 
nm, and (c) 586
 
nm.
 
 
89
 
2.4.10. Concluding Comments on Effec
ts of Steric Bulk
 
We can now assess which of the two excited state decay mecha
nisms outlined in section 2.4.7 is 
most appropriate when steric bulk is added to the periphery of the acac ligand. In the first possible 
mechanism, steric bulk causes a drastic i
ncrease in the lifetime for 
4
T
2
 

2
E ISC and a slight increase in 
the lifetime
 
for IC from the 
4
T
2
 
state. The second possible mechanism is very similar to that for the 
unsubstituted Cr(acac)
3
 
molecule. Upon excitation to the 
4
T
2
 
state, the molecule rapidly forms a quasi
-
equilibrium betw
een the 
2
E and 
4
T
2
 
states. As the system dissipates excess vibrational energy the efficiency 
of thermally activated bISC from the 
2
E state is diminished resulting in a shift of the 
2

4
T
2
 
equilib
rium 
to the left and the observation of ISC from the 
4
T
2
 
st
ate. The only substantive change relative to Cr(acac)
3
 
is the slightly slower IC from the 
4
T
2
 
state. We can parse out which of these excited state decay mechanisms 
is most realistic by comparing sp
ectral evolution seen in the early time full spectrum data 
for each 
compound in Figure
 
2.48 including data on Cr(acac)
3 
and Cr(TMHD)
3
 
collected by Eric Juban
7
 
and Joel 
Schrauben
10,15
 

-
substituted series that the excited state decay 

3
 
is relatively insensitive to synthetic mo
difications. Furthermore, comparison of 
each of the dataset
s in Figure
 
2.48 shows only very subtle changes throughout the steric series, giving no 
indication of a 
quantum leap
 
in the ISC lifetime. It is thus highly unlikely that the excited state decay 
mec
hanism (Mechanism 1) first proposed by Schrauben correctly 
describes the dynamics in this series of 
compounds. Mechanism 2, however, can describe the dynamics in every member in both the 3
-
substituted 
and steric
-
bulk acac series discussed thus far with sma
ll changes to the internal conversion lifetime. This 
modula
tion of the IC lifetime as a function of steric bulk is supported by the slower IC time constant 
observed for Cr(HD)
3
 
relative to the Cr(acac)
3
 
control molecule. Addition of another set of methyl g
roups 
to the periphery if the molecule is enough to make th
e shift in the 
2

4
T
2
 
equilibrium outpace the IC from 
the 
4
T
2
 
state. This mechanism also makes intuitive sense when considering the vibrational overlap between 
relevant electronic states. For exa
mple, small changes in the geometry of the 
4
T
2
 
state would not be expected 
to significantly alter the presumably abundant vibrational overlap with the 
2
E electronic state and hence 

ilibrium position of the 
4
T
2
 
state as 
 
90
 
a result of increased steric encumbr
ance, however, would be expected to modulate the sparser vibrational 
overlap with the 
4
A
2
 
ground state and in effect modi
fy the IC rate.
 
2.4.11. Cr(TMHD)
3
 
 
The previous section has just illustrated the decay mechanism of Cr(TMHD)
3
, and the kinetics of 
this
 
compound have already been investigated.
10,15
 
This work will thus focus on characterizing the coherent 
vibrational motion of Cr(TMHD)
3
. A summary of the oscillatory components observed
 
during TA 
experiments on Cr(TMHD)
3
 
can be seen in Table
 
2.7 with representative datasets in 1,4
-
dioxane 
and
tetrahydrofuran shown in Figures 2.49 and 2.50 respectively. Frequencies bellow 100 cm
-
1
 
were not 
reported as they were likely due to poor exponenti
al fitting or poor signal stability. Furthermore, many of 
the values in Table
 
2.7 have large error bars aris
ing from the low magnitude of the oscillatory signal, the 
a)
 
b)
 
d)
 
c)
 
Figure
 
2.48.
 
Early time full spectrum TA data on (a) Cr(acac)
3
, (b) Cr(HD)
3
, (c) 
Cr(DMHD)
3
, and (d) Cr(TMHD)
3
 
following excitation 
into the 
4
T
2
 
excited state. Parts (a) 
and (d) were reproduced with data from references 7 and 15 respectively.
 
 
91
 

pu
, nm
 
 
(E
pu
, 

J)
 

pr
, nm
 
 
(E
pr
, 

J)
 
IRF, fs
 
Solvent
 
Frequency, cm
-
1
 
(Damping Time, fs)
 
600 (3)
 
592 (0.3)
 
45
 
Dioxane
 
 
166 (683)
 
 
 
 
 
 
600 (3)
 
592 (0.3)
 
45
 
Dioxane
 
115 (543)
 
168 (6102)
a
 
 
246 (485)
 
285 (1609)
 
 
 
600 (2)
 
592 (0.2)
 
45
 
Dioxane
 
124 (542)
 
 
 
 
 
 
 
600 (2)
 
592 (0.2)
 
45
 
Dioxane
 
137 (1485)
 
 
221 (1028)
 
 
 
 
 
600 (2.8)
 
592 (0.3)
 
40
 
Dioxane
 
 
162 (840)
 
 
 
 
 
 
600 (3)
 
480 (na)
 
100
 
Dioxane
 
 
157 (12
21)
 
 
 
 
 
 
600 (3)
 
480 (na)
 
125
 
Dioxane
 
 
201 (138)
 
 
 
 
 
 
600 (3)
 
480 (na)
 
165
 
Dioxane
 
122 (606)
 
177 (175)
 
 
256 (428)
 
 
 
 
560 (3)
 
550 (0.3)
 
50
 
Dioxane
 
 
168 (431)
 
 
244 (929)
 
 
 
 
520 (5)
 
514 (0.4)
 
75
 
Dioxane
 
 
177 (327)
 
 
 
298 (882)
 
 
 
600 (3)
 
592 (0.3)
 
45
 
Benze
ne
 
128 (255)
 
173 
(1052)
 
 
 
 
 
 
600 (3)
 
595 (0.3)
 
50
 
THF
 
111 (373)
 
 
221 (229)
 
 
 
346 (1026)
 
 
600 (3)
 
592 (0.3)
 
50
 
THF
 
 
 
 
242 (396)
 
 
 
 
600 (3)
 
592 (0.3)
 
45
 
THF
 
 
186 (1639)
 
 
245 (2387)
 
 
323 (2199)
 
372 (1151)
 
600 (3)
 
592 (0.3)
 
50
 
THF
 
139 (499)
 
 
 
 
 
 
 
600 (5)
 
584 (0.5)
 
35
 
THF
 
 
162 (1371)
 
 
 
 
 
 
600 (3)
 
480 (na)
 
165
 
THF
 
126 (1143)
 
195 (383)
 
 
256 (820)
 
 
 
 
580 (5)
 
568 (0.5)
 
45
 
THF
 
 
162 (559)
 
 
 
 
 
 
560 (3)
 
550 (0.3)
 
50
 
THF
 
 
172 (638)
 
 
 
 
 
 
560 (5)
 
545 (0.5)
 
55
 
THF
 
 
177 (855)
 
 
252 (1242)
 
 
 
 
560 (5)
 
546 (0.5)
 
65
 
THF
 
 
166 
(734)
 
209 (3928)
 
 
 
 
375 (995)
 
520 (5)
 
508 (0.5)
 
70
 
THF
 
 
174 (1300)
 
197 (769)
 
 
 
 
 
Average Frequency, cm
-
1
 
125
 
173
 
212
 
249
 
292
 
335
 
374
 
Standard Deviation, cm
-
1
 
10
 
12
 
11
 
6
 
9
 
16
 
2
 
Average Damping, fs
 
680
 
770
 
1490
 
960
 
1250
 
1600
 
1070
 
Standard 
Deviatio
n, fs
 
420
 
450
 
1660
 
700
 
510
 
830
 
110
 
a. The slow nature of this decay coupled with the additional oscillatory components present made i
t 
difficult for the LPSVD program 
to get a reasonable dephasing time for this vibrational mode. This value was 
thus exclud
ed from the averaging process.
 
Table
 
2.7.
 
Summary of the oscillatory components observed for Cr(TMHD)
3
 
following excitation into the 
4
T
2
 
state across several 
solvents. Experimental details such as the IRF and pump and probe wavelengths and pulse energies 
are included for each experiment. 

 
 
92
 
presence of oscillatory background signals from the solvent, and a generally congested
 
vibrational 

3
 
derivatives (i.e. mos
tly only the low frequency symmetric breathing mode) it is likely that several of the 
frequencies extracted from LP
SVD analyses are simply experimental artifacts and not attributable to 
anything physically meaningful. Inspection of Table 2.7 shows that the
 
125, 173, and 249
 
cm
-
1
 
frequencies 
appear most frequently across the provided datasets and thus are most likely to
 
actually represent excited 
state vibrational coherences. To gain more insight into the legitimacy of these vibrational modes we can 
use the 
resonance Raman spectrum of Cr(TMHD)
3
 
shown in Figure 2.51. Since the 125
 
cm
-
1
 
feature from 
 
a)
 
b)
 
d)
 
c)
 
Figure
 
2.49.
 
Representative transient absorption data on Cr(TMHD)
3
 
in 1,4
-
dioxane 
following excitation int
o the 
4
T
2
 
excited state while monitoring at (a) 592 nm and (c) 550
 
nm 
with corresponding vibrational spectra (b and d) resulting 
from FFT and LPSVD analyses.
 
 
93
 
 
 
a)
 
b)
 
d)
 
c)
 
Figure
 
2.50.
 
Representative transient absorption data on Cr(TMHD)
3
 
in tetrahydrofuran 
following excitation into the 
4
T
2
 
excited state while monitoring at (a) 568 nm and (c) 545
 
nm 
with corresponding vibrational spec
tra (b and d) resulting from FFT and LPSVD analyses.
 
162
 
56
 
Figure
 
2.51.
 
Stacked plot of Raman data on Cr(acac)
3
 
and Cr(TMHD)
3
 
in THF with 488
 
nm 
excitation 
source in resonance with the blue edge of the 
4
A
2
 

4
T
2
 
absorption band.
 
 
94
 
Table
 
2.7 is too c
lose to the Rayleigh line, its presence in the Raman spectrum of Figure
 
2.51
 
could not be 
confirmed or denied. Considering the large error bars on the 173
 
(12) cm
-
1
 
mode it seems that this could be 
attributed to the 156 cm
-
1
 
peak in the Raman spectrum. The
 
249 cm
-
1
 
mode, on the other hand is in good 
agreement with the 250 cm
-
1
 
pea
k observed in the Raman data. The 212, 292, 335, and 374
 
cm
-
1
 
frequencies 
observed in the TA data, however, do not have corresponding peaks in the Raman spectrum. While this 
could 
be a symptom of selection rules, it seems unlikely given the irregularity of
 
their occurrence in the TA
 
data that they arise from coherences of Cr(TMHD)
3
. The 212, 292, 335, and 374
 
cm
-
1
 
oscillatory 
components are thus assigned as experimental artifacts.
 
 
Now that the 125, 173, and 249
 
cm
-
1
 
frequencies observed during TA experime
nts on Cr(TMHD)
3
 
have been shown to possibly be legitimate vibrational modes, the nuclear deformations associated with 
these modes can be characterized with DFT calculations. We ca
n first assign the 173
 
cm
-
1
 
experimental 
frequency to a calculated 162
 
cm
-
1
 
symmetric breathing mode (Figure
 
2.52) that is essentially identical to 

3
 
derivatives studied thus far. While the frequency 
m
atching between the 173 (12)
 
cm
-
1
 
experimental and 162
 
cm
-
1
 
calculated frequ
encies is not very good, the 
calculated mode does fall within the error bars. This is also in good agreement with the 156
 
cm
-
1
 
peak in 
the resonance Raman spectrum (Figure
 
2.51). A
ssigning the 249 cm
-
1
 
vibrational mode is more complicated 
a)
 
b)
 
Calc: 162 cm
-
1
 
Exp: 173 cm
-
1
 
Raman: 156 cm
-
1
 
Calc: 245 cm
-
1
 
Exp: 249 cm
-
1
 
Raman: 250 cm
-
1
 
Figure
 
2.52.
 
Nuclear deformations associated with the (a) 173
 
cm
-
1
 
and (b) 249
 
cm
-
1
 
experimen
tally observed coherent vibrational modes of Cr(TMHD)
3
.
 
 
95
 
given the density
 
of states in this frequency region and the lack of precedent to compare to. Nevertheless, 
this mode is tentatively assigned to a    245
 
cm
-
1
 
calculated vibrational mode (Figure
 
2.
52) which exhibits 
a symmetric compression of the TMHD ligands towards the m
etal center concomitant with each oxygen 
atom splaying outwards in the plane of the ligand backbone. This particular motion was chosen in 
consideration of the symmetry of the mode 
(and the lack thereof in several of the other modes in this 
frequency region
) as well as the calculated Raman activity of this mode. While the selection rules for 
Raman, resonance Raman, and excited state coherent vibrational modes observable in this TA ex
periment 
are all likely different, we have seen that every observable cohere
nt vibrational mode in this system thus 
far has been highly symmetric and Raman active. We can thus make the assumption that this holds for the 
experimental 249
 
cm
-
1
 
mode as well g
iven that it is also observed in the resonance Raman spectrum 
(Figure
 
2.51).
 
While this vibrational mode is somewhat similar to the symmetric breathing mode that has 

3
 
derivatives, this mode exhibits a bit more em
phasis on 
distortion of the periphery of the molecule as opposed to distorti
on of the primary coordination sphere. It 
may be the case that the additional atoms in this molecule result in several different flavors of this type of 
motion appearing at differe
nt frequency regions. Finally, the 125 (10) cm
-
1
 
mode again falls in a regio
n of 
high density of vibrational states. With no precedent, large error bars, and no obvious way to assess the 
viability of each of these candidate modes, no assignment will be mad
e here.  
 
 
When comparing the dephasing times of the low frequency symmetric
 
breathing mode across the 

3
 
series, it can be seen that the more sterically encumbered Cr(TMHD)
3
 
exhibits dephasing times 
(770
 
fs) that are intermediate between the unsub
stituted Cr(acac)
3
 
(340
 
fs) and halogenated Cr(3
-
Clac)
3
 
(1100
 
fs) compounds.
 
While it was established that the dephasing times in this system are not very 
informative with regards to contribution to an electronic state change, it may yet be enlightening to
 
further 
probe why the dephasing times are changing across this series of co
mpounds. Apart from Cr(3
-
NO
2
ac)
3
 
the 
dephasing time increased as the vibrational frequency decreased. This may suggest a correlation between 
the vibrational frequency and the susce
ptibility of a normal mode to dephasing via excited state dynamics 
or hint a
t a subtle relationship between the collection of vibrational frequencies of a molecule and the rate 
 
96
 
of IVR. To investigate this we can study one more molecule, Cr(4
-
Cl
-
TMHD)
3
 
whic
h should have a reduced 
vibrational frequency of the symmetric breathing mod
e relative to Cr(TMHD)
3
.
 
2.4.12. Cr(4
-
Cl
-
TMHD)
3
 
 
The goal of this section is to characterize the early time ultrafast dynamics of Cr(4
-
Cl
-
TMHD)
3
 
and 
to glean more information on th
e origin of dephasing in these types of systems. As such the discussion will 
largely focus on these endeavors. The crystal structure information is shown in the appendix and was 
collected by McCusker group member Sara Ad
elman and solved by Professor Richar
d Staples of Michigan 
State University.
 
 
The early time full spectrum TA data on Cr(4
-
Cl
-
TMHD)
3
 
in DCM following excitation into the 
4
T
2
 
state is shown in Figure
 
2.
53
a.
 
While this dataset is unfortunately quite noisy, the global fit 
a)
 
b)
 
d)
 
c)
 
Figure
 
2.53.
 
(a) Early time full spectrum TA data on Cr(4
-
Cl
-
TMHD)
3
 
in dichloromethane 
following 600
 
nm excitation into the 
4
T
2
 
state. (b) Global fit of the TA data to a sequential 
bi
exponential decay. The long
-
time offset was modeled by fixing a 1
 
ns exponential decay to 
describe the ground state recovery process. (c) De
cay associated difference spectra (DADS) and (d) 
Evolution associated difference spectra (EADS) associated with the 
global fit.
 
 
97
 
(Figure
 
2.
53
b) cleans
 
up nicely and seams to closely resemble the spectral evolution of Cr(TMHD)
3
 
in 
Figure 2.48d. The fit reveals a 1.7
 
ps kinetic component which exhibits a decay in the blue edgeof the 
transient spectrum with a concomitant rise from ~550
-
650
 
nm, which is aga
in similar to Cr(TMHD)
3
 
in 
terms of kinetics and spectral evolution. These s
pectral features are confirmed in single wavelength datasets 
(Figure
 
2.54a) showing a rising feature in the red edge of the transient spectrum at 600
 
nm and a 
concomitant decaying 
feature in the blue edge of the spectrum at 510
 
nm and an average time const
ant of 
1.7
 
±
 
0.2
 
ps. The early time kinetics of Cr(4
-
Cl
-
TMHD)
3
 
are thus essentially identical to those of 
Cr(DMHD)
3
 
(
1.7
 
±
 
0.2
 
ps) and Cr(TMHD)
3
 
(
1.8
 
ps)
10
.
 
a)
 
b)
 
c)
 
Figure
 
2.54.
 
(a) Representative single wavelength TA data on Cr(4
-
Cl
-
TMHD)
3
 
following 600
 
nm 
excitation into the 
4
T
2
 
state in dichloromethane. Time constants for the fits of
 
specific data sets are 
included on the graph itself, howe
ver the average time constant measured across all datasets is 1.7 
±
 
0.2
 
ps. (b) Zoomed in portion of the 600
 
nm probe data shown in part a with exponential and 
LPSVD fits and (c) corresponding vibra
tional spectrum resulting from FFT and LSVD analyses.
 
 
98
 
 
In addition to confirming the qualitative spectral evolution of Cr(4
-
Cl
-
TMHD)
3
, the single 
wavelngth datasets also showed coherent oscillations suppe
rimposed on the population dynamics 
(Figure
 
2.54b) which were analysed with FFT 
and LPSVD techniques (Figure
 
2.54c). A summary of these 
preliminary results can be seen in Table
 
2.8. It should be noted that the sub
-
100
 
cm
-
1
 
oscillatory components 
presented 
in Table 2.8 likely do not actually represent coherent vibrational motion of Cr(
4
-
Cl
-
TMHD)
3
 
but 
rather an artifact of poor exponential fitting of the LPSVDS program. It is readilly apparent from each of 

3
derivatives thus far that one of these 
modes is likely a symmetric breathing mode. This is 
confirmed by DFT calculation
s showing a calculated 118 cm
-
1
 
symmetric breathing mode (Figure
 
2.55) in 
excellent agreement with the experimental results. Unfortunately, the high degree of uncertainty in th
e 
E
pu
 
(E
pr
), 

J
 
IRF, fs
 
Frequency, cm
-
1
 
(Damping Time, fs)
 
2.2 (0.3)
 
55
 
48 (279)
 
120 (478)
 
153 (235)
 
3.6 (0.4)
 
45
 
85 (529)
 
115 (342)
 
169 (381)
 
Average Frequency, cm
-
1
 
67
 
118
 
161
 
Standard Deviation, cm
-
1
 
26
 
4
 
11
 
Average Damping Time, fs
 
400
 
400
 
300
 
Standar
d Deviation, fs
 
200
 
100
 
100
 
 
Table
 
2.8.
 
Summary of coherent oscillations observed in TA data on Cr(4
-
Cl
-
TMHD)
3
 
following 
600
 
nm excitation into the 
4
T
2
 
state while monitoring at 600
 
nm in dichloromethane. The sub
-
100
 
cm
-
1
 
oscillatory components were inc
luded for transparency; however, it is likely that they arise 
from poor exponential fitting and do not truly represent coherent vibrational motion of the 
solute. 
Experimental details such as the IRF and the pump and probe pulse energies are included for ea
ch 
measurement.
 
Figure
 
2.55.
 
Nuclear deformations associated with the 118
 
cm
-
1
 
experimentally observed 
coherent oscillations of Cr(4
-
Cl
-
TMHD)
3
.
 
Calc:
 
118 cm
-
1
 
Exp: 118 cm
-
1
 
 
99
 
161
 
cm
-
1
 
component coupled with the high density of states in that frequency r
egion precludes any 
definitive assignment of that mode. Additional TA experiments are needed to narrow the selection of 
possible vibrational modes before any meaningful assignm
ent can be made.
 
 
 
We can now turn our attention to the dephasing time of the 
low frequency symmetric breathing 
mode. In the 3
-

3
 
compounds it was observed that halogen substitution served 
to both reduce the vibrational freque
ncy of the symmetric breathing mode and extend the dephasing time. 
Replacement
 
of the methyl groups in the periphery of the structure with bulkier 
t
-
Butyl groups was also 
shown to decrease the vibrational frequency while increasing the dephasing time relat
ive to the Cr(acac)
3
 
control molecule. This trend prompts the question of whet
her there is an inverse relationship between the 
vibrational frequency (and by extension the vibrational period) and the susceptibility of a vibrational mode 
to be dephased by ex
cited state evolution or if there is a more subtle relationship between the co
llection of 
vibrational frequencies of a molecule and its IVR rate. Until now, it seemed that Cr(3
-
NO
2
ac)
3
 
which 
exhibited a similar frequency to Cr(3
-
Clac)
3
 
yet a shorter dephas
ing time than Cr(acac)
3
 
was the only outlier 
to the trend of increased dephasi
ng times concomitant with lower vibrational frequencies. Now, however, 
the Cr(4
-
Cl
-
TMHD)
3
 
results affords several additional comparisons. First, when comparing the ~190 cm
-
1
 
mode
 
of Cr(acac)
3
 
with 380
 
fs damping time to this 118
 
cm
-
1
 
mode of Cr(4
-
Cl
-
TMHD)
3
 
with 400
 
fs 
damping time, we see there is a drastic reduction in frequency with essentially no change in dephasing time, 
which suggests that there is no inverse relationship bet
ween dephasing and vibrational frequency that is 
active in this system. Furthe
rmore, when comparing the Cr(4
-
Cl
-
TMHD)
3
 
results with the ~160
 
cm
-
1
 
mode 
(770
 
fs dephasing time) of the unchlorinated Cr(TMHD)
3
 
compound, we see a decrease in the dephasing 
time 
upon a decrease in the vibrational frequency further disproving the above hypo
thesis. This thus points 
to a more complicated relationship between the dephasing of this mode and its vibrational frequency, 
whereby the rate of dephasing is dictated by the rel
ative energies of the symmetric breathing mode and the 
accepting vibrational m
odes that are excited via IVR.
 
 
 
 
100
 

3
 
Systems
 
 
The work presented in this chapter has built both upon the works of previous McCusker group 
membe
rs
7,10,12,15
 
and other research groups
8,13,14
 
to build a comprehensive photophysical model of the 
excited state dynami
cs of Cr(acac)
3
 

-
substituted and more sterically encumbered derivati
ves. These 
results show that following excitation into a quartet excited state, a rapid quasi
-
equilibrium is established 
between the 
2
E and 
4
T
2
 
excited states. This equilibrium i
s perturbed by 
4
T
2
 

4
A
2
 
IC with a rate that is 
modulated by steric bulk in the periphery of the molecule, and by vibrational cooling which shifts the 
2
E 

4
T
2
 
equilibrium to the left. A previous assignment of 1.1
 
ps
 
vibrational relaxation
7
 
on the 
2
E state of 
Cr(acac)
3
 
was disproven separately using an analysis of full spectrum TA data following higher energy 
excitation into the 
4
T
1
 
excited state; analysis of the resonance Ra
man li
newidth and the dephasing time of 
the 461
 
cm
-
1
 
mode of Cr(acac)
3

3
 
derivatives exceeding the previously proposed vibrational cooling time constant. This ~1
 
ps process was 
then rea
ssigne
d to internal conversion from the 
4
T
2
 
state. It was also consistently shown that the same low 
frequency symmetric metal
-

3
 
compound studied. The 
unsubstituted Cr(acac)
3
 
compound also exhibited an 
additi
onal coherent metal
-
ligand stretching mode at 
461 cm
-
1
. Both of these normal modes were attributed to the 
2
E excited state on the basis of low oscillator 
strength preventing the observation of wavepackets on the ground electronic state during TA expe
riment
s; 
similar dephasing times in the excited state measured by TA and on the ground state extracted by an analysis 
of the Raman linewidth; and damping times that were slower than the internal conversion from the 
4
T
2
 
state. 
Finally, this work addressed t
he pro
position that short dephasing times can be an indicator for involvement 
of specific vibrational modes in an ultrafast electronic state change. Observation of dephasing times of the 
same vibrational mode that varied from ~200
 
fs to 2.5
 
ps with no obse
rvable
 
change in the ultrafast ISC 
rate suggested that in this system, dephasing times are a rather poor indicator for involvement in the ultrafast 
ISC. Furthermore, the subtle patterns in dephasing times in relation to vibrational frequencies suggests tha
t 
IVR 
may be contributing to the observed dephasing in this system and that its efficiency is dependent on 
the relative energies of the 3N
-
6 vibrational modes of the molecule. A summary of the frequencies and 
 
101
 
dephasing times of the low frequency symmetric
 
breathing mode observed in each complex can be seen in 
Figure
 
2
.56.
 
 
Further studies on the Cr(acac)
3
 
system may benefit from variable temperature and variable 
excitation wavelength measurements. It was proposed in the work of Kunttu and coworkers
14
 
that relax
ation 
from the 
2
E excited state at long times followed an Arrhen
ius relationship resulting from thermally activated 
bISC to the 
4
T
2
 
state from which ground state recovery could occur via 
4
T
2
 

4
A
2
 
IC. If so then variable 
temperature experiments on the set of compounds studied here could provide thermodynamic properties to 
help describe the 
2

4
T
2
 
equilibrium which may be useful for a more complete quantitative modeling of 
the excited state 
dynami
cs. These results could be combined with variable excitation wavelength 

3
 
system 
which in turn determine the quantum yield of ground state recovery at ear
ly tim
es due to IC within the 
Cr(TMHD)
3
 
173 (12) cm
-
1
 
770 (450)
 
fs
 
Cr(4
-
Cl
-
TMH
D)
3
 
118 (4) cm
-
1
 
400 (100)
 
fs
 
Cr(acac)
3
 
193 (10) cm
-
1
 
340 (140)
 
fs
 
Cr(3
-
NO
2
ac)
3
 
150 (4) cm
-
1
 
186 (11)
 
fs
 
Cr(3
-
Clac)
3
 
149 (2) cm
-
1
 
1100 (180)
 
fs
 
Cr(3
-
Brac)
3
 
123 (1) cm
-
1
 
2500 (700)
 
fs
 
Cr(3
-
Iac)
3
 
101 (2) cm
-
1
 
2400 (600)
 
fs
 
Figure
 
2.56.
 
Summary of the frequencies and dephasing times 
of the low frequency 

3
 
derivative.
 
 
102
 
quartet manifold. The combination of these two sets of experiments may provide enough information to 
accurately and quantitatively describe the excited state dynamics in this class of compounds.
 
 
 
 
 
103
 
 
 
 
 
 
 
 
 
 
 
 
 
APPENDICES
 
 
 
 
104
 
App
endix 2.A. Crystal Structures
 
The following is the PDB file for the crystal structure of Cr(3
-
Iac)
3
.
 
HEADER    CSD ENTRY sla_bcp2005
 
CRYST1  12.8403  10.1064  16.2581  90.00 105.04  90.00     P21/n
 
SCALE1      0.077880  0.000000  0.020928       0.000000
 
SC
ALE2      0.000000  0.098947  0.000000       0.000000
 
SCALE3      0.000000  0.000000  0.063690       0.000000
 
HETATM    1 I001 UNK     1       4.300   0.876  
-
0.286  1.00  2.22           I  
 
ANISOU    1 I001 UNK     1      324    290    231    
-
15    
-
4
5  
  
-
92       I  
 
HETATM    2 I002 UNK     1       5.954   8.926   3.495  1.00  2.11           I  
 
ANISOU    2 I002 UNK     1      304    147    349    
-
46    
-
13     56       I  
 
HETATM    3 I003 UNK     1       5.373   1.115   9.099  1.00  2.60         
  
I
  
 
ANISOU    3 I003 UNK     1      526    256    207    
-
52    
-
60     65       I  
 
HETATM    4 Cr04 UNK     1       5.363   3.587   4.269  1.00  1.08          Cr  
 
ANISOU    4 Cr04 UNK     1      156    105    151      7    
-
10     
-
2      Cr  
 
HETATM    
5 O005 UNK     1       4.062   2.238   3.734  1.00  1.45           O  
 
ANISOU    5 O005 UNK     1      176    159    215    
-
32     27     
-
2       O  
 
HETATM    6 O
006 UNK     1       6.695   4.929   4.763  1.00  1.50           O  
 
ANISOU    6 O006 UNK   
  
1      206    141    225      4    
-
44     
-
7       O  
 
HETATM    7 O007 UNK     1       4.387   3.794   5.945  1.00  1.58           O  
 
ANISOU    7 O007 UNK     1
      
244    180    177     41      6     
-
9       O  
 
HETATM    8 O008 UNK     1       4.2
97   5.001   3.439  1.00  1.51           O  
 
ANISOU    8 O008 UNK     1      187    159    228      0    
-
45     13       O  
 
HETATM    9 O009 UNK     1       6.447 
  
2.192   5.114  1.00  1.57           O  
 
ANISOU    9 O009 UNK     1      216    175    207
     
28     
-
1     
-
7       O  
 
HETATM   10 O00A UNK     1       6.269   3.361   2.558  1.00  1.50           O  
 
ANISOU   10 O00A UNK     1      180    187    202   
 
-
34      5    
-
25       O  
 
HETATM   11 C00B UNK     1       5.745   6.848   3.764  1.00  
1.30           C  
 
ANISOU   11 C00B UNK     1      229     90    177    
-
17     11     34       C  
 
HETATM   12 C00C UNK     1       4.573   6.233   3.277  1.00  1.3
3           C  
 
ANISOU   12 C00C UNK     1      204    140    161     12     
-
3      9     
  
C  
 
HETATM   13 C00D UNK     1       3.919   1.640   2.627  1.00  1.36           C  
 
ANISOU   13 C00D UNK     1      139    140    236     28    
-
45     
-
2       C
  
 
HETATM   14 C00E UNK     1       6.761   6.159   4.459  1.00  1.26           C  
 
ANISOU 
  
14 C00E UNK     1      175    160    142    
-
30     34     
-
7       C  
 
HETATM   15 C00F UNK     1       5.884   2.718   1.521  1.00  1.42           C  
 
ANISOU   1
5 C00F UNK     1      179    150    211     46    
-
17     18       C  
 
HETATM   16 C00G UNK
     
1       4.447   3.082   7.007  1.00  1.41           C  
 
ANISOU   16 C00G UNK     1      184    180    173    
-
49      3    
-
40       C  
 
HETATM   17 C00H UNK   
  
1       4.768   1.872   1.512  1.00  1.44           C  
 
ANISOU   17 C00H UNK     1      2
02    190    153     43    
-
38    
-
44       C  
 
HETATM   18 C00I UNK     1       6.365   1.699   6.290  1.00  1.52           C  
 
ANISOU   18 C00I UNK     1      214     90    275     14    
-
59    
-
13       C  
 
HETATM   19 C00J UNK     1       5.388   2.082
   
7.218  1.00  1.58           C  
 
ANISOU   19 C
00J UNK     1      275    170    154    
-
30    
-
53     74       C  
 
HETATM   20 C00K UNK     1       3.518   7.004   2.511  1.00  2.16           C  
 
ANISOU   20 C00K UNK     1      249    240    330    
-
14   
-
109     70       C  
 
HETATM   21 H00A UNK     1
       
2.921   6.375   2.054  1.00  3.24           H  
 
HETATM   22 H00B UNK     1       3.952   7.582   1.850  1.00  3.24           H  
 
 
105
 
HETATM   23 H00C UNK     1       2.997   7.553   3.135  1.00  3.24      
     
H  
 
HETATM   24 C00L UNK     1       3.397 
  
3.428   8.025  1.00  2.35           C  
 
ANISOU   24 C00L UNK     1      331    350    214     47     74     18       C  
 
HETATM   25 H00D UNK     1       2.832   4.151   7.680  1.00  3.55           H  
 
HETA
TM   26 H00E UNK     1       3.829   3.719   8.8
55  1.00  3.55           H  
 
HETATM   27 H00F UNK     1       2.844   2.639   8.205  1.00  3.55           H  
 
HETATM   28 C00M UNK     1       2.780   0.664   2.573  1.00  1.98           C  
 
ANISOU   28 C00M 
UNK     1      240    150    360    
-
31    
-
17  
  
-
44       C  
 
HETATM   29 H00G UNK     1       2.470   0.474   3.484  1.00  3.00           H  
 
HETATM   30 H00H UNK     1       3.082  
-
0.167   2.151  1.00  3.00           H  
 
HETATM   31 H00I UNK     1    
   
2.045   1.049   2.051  1.00  3.00           H
  
 
HETATM   32 C00N UNK     1       8.013   6.857   4.885  1.00  2.03           C  
 
ANISOU   32 C00N UNK     1      242    220    310    
-
64    
-
33     
-
1       C  
 
HETATM   33 H00J UNK     1       7.806   7.
489   5.605  1.00  3.00           H  
 
HETATM   3
4 H00K UNK     1       8.392   7.343   4.123  1.00  3.00           H  
 
HETATM   35 H00L UNK     1       8.662   6.197   5.207  1.00  3.00           H  
 
HETATM   36 C00O UNK     1       6.743   2.949   0.303  
1.00  2.13           C  
 
ANISOU   36 C00O UNK   
  
1      249    330    231    
-
16     25    
-
18       C  
 
HETATM   37 H00M UNK     1       6.196   3.329  
-
0.416  1.00  3.24           H  
 
HETATM   38 H00N UNK     1       7.125   2.097   0.006  1.00  3.24   
        
H  
 
HETATM   39 H00O UNK     1       7.467   3.571   0.526  1.00  3.24           H  
 
HETATM   40 C00P UNK     1       7.409   0.659   6.593  1.00  2.40           C  
 
ANISOU   40 C00P UNK     1      363    230    320     54   
-
100     30       C  
 
H
ETATM   41 H
00P UNK     1       6.977  
-
0.210   6.725  1.00  3.63           H  
 
HETATM   42 H00Q UNK     1       7.894   0.910   7.408  1.00  3.63           H  
 
HETATM   43 H00R UNK     1       8.039   0.600   5.844  1.00  3.63           H  
 
CONECT    1   
17
 
CONECT   
 
2   11
 
CONECT    3   19
 
CONECT    4    5    6    7    8
 
CONECT    4    9   10
 
CONECT    5    4   13
 
CONECT    6    4   14
 
CONECT    7    4   16
 
CONECT    8    4   12
 
CONECT    9    4   18
 
CONECT   10    4   15
 
CONECT   11    2   12   14
 
CONECT
   
12    8  
 
11   20
 
CONECT   13    5   17   28
 
CONECT   14    6   11   32
 
CONECT   15   10   17   36
 
CONECT   16    7   19   24
 
CONECT   17    1   13   15
 
CONECT   18    9   19   40
 
CONECT   19    3   16   18
 
CONECT   20   12   21   22   23
 
CONECT   21   
20
 
CONECT   22   20
 
CONECT   23   20
 
CONECT   24   16   25   26   27
 
 
106
 
CONECT   25   24
 
CONECT   26   24
 
CONECT   27   24
 
CONECT   28   13   29   30   31
 
CONECT   29   28
 
CONECT   30   28
 
CONECT   31   28
 
CONECT   32   14   33   34   35
 
CONECT   33   32
 
CONE
CT 
  
34   32
 
CONECT   35   32
 
CONECT   36   15   37   38   39
 
CONECT   37   36
 
CONECT   38   36
 
CONECT   39   36
 
CONECT   40   18   41   42   43
 
CONECT   41   40
 
CONECT   42   40
 
CONECT   43   40
 
MASTER        0    0    0    0    0    0    0    3   43    0
   
44    0
 
END
 
 
The Following is the PDB file for the crystal structure of Cr(DMHD)
3
.
 
HEADER    CSD ENTRY sla_bcp_1142
 
CRYST1  
18.7029  15.7164  20.7137  90.00  90.00  90.00      Pbca
 
SCALE1      0.053468  0.000000  0.000000       0.000000
 
SCALE2      0.0
00000  0.063628  0.000000       0.000000
 
SCALE3      0.000000  0.000000  0.048277       0.000000
 
HETATM    1 Cr01 UNK     1     
  
4.895   4.757   6.752  1.00  2.20          Cr  
 
Figure
 
2.A.1.
 
Crystal structure for Cr(DMHD)
3
 
that was collected by Sara Adelman and solved 
by Professor Richard Staples.
 
 
107
 
ANISOU    1 Cr01 UNK     1      352    255    228    
-
32    
-
58     11      Cr
  
 
HETATM    2 O002 UNK     1       3.923   6.224   5.911  1.00  2.54           O  
 
ANISOU    2 O002 UNK     1      404    287  
  
274     10    
-
38      4       O  
 
HETATM    3 O003 UNK     1       6.158   5.949   7.665  1.00  2.52           O  
 
ANISOU    
3 O003 UNK     1      373    342    241    
-
71    
-
36     
-
9       O  
 
HETATM    4 O004 UNK     1       6.108   4.680   5.219  1
.00  2.81           O  
 
ANISOU    4 O004 UNK     1      401    380    288     31    
-
33    
-
22       O  
 
HETATM    5 O005 UNK   
  
1       3.678   4.850   8.287  1.00  2.83           O  
 
ANISOU    5 O005 UNK     1      382    428    264    
-
93    
-
22     32
       
O  
 
HETATM    6 O006 UNK     1       3.665   3.536   5.858  1.00  2.77           O  
 
ANISOU    6 O006 UNK     1      423 
   
289    341    
-
39   
-
122      9       O  
 
HETATM    7 O007 UNK     1       5.830   3.257   7.592  1.00  3.00           O  
 
AN
ISOU    7 O007 UNK     1      461    311    368    
-
22   
-
139     55       O  
 
HETATM    8 C008 UNK     1       3.591   2.274   
6.012  1.00  2.51           C  
 
ANISOU    8 C008 UNK     1      389    306    259    
-
45     44    
-
41       C  
 
HETATM    9 C00
9 UNK     1       3.887   6.475   4.653  1.00  2.70           C  
 
ANISOU    9 C009 UNK     1      443    270    311    
-
61    
-
8
6     15       C  
 
HETATM   10 C00A UNK     1       6.219   6.124   8.928  1.00  2.63           C  
 
ANISOU   10 C00A UNK     1      375    330    295     17    
-
52    
-
34       C  
 
HETATM   11 C00B UNK     1       3.982   5.251   9.457  1.00  2.86         
  
C  
 
ANISOU   11 C
00B UNK     1      422    364    300    
-
11      9     34       C  
 
HETATM   12 C00C UNK     1       5.515   2.023   7.523  1.00  2.61           C  
 
ANISOU   12 C00C UNK     1      406    313    273     41     48     20       C  
 
HETATM 
  
13 C00D UNK     1
       
4.450   1.508   6.802  1.00  2.71           C  
 
ANISOU   13 C00D UNK     1      410    269    351     
-
1      1    
-
15       C  
 
HETATM   14 H00D UNK     1       4.292   0.572   6.848  1.00  3.24           H  
 
HETATM   15 C00E UNK
     
1       5.894 
  
5.192   4.069  1.00  2.94           C  
 
ANISOU   15 C00E UNK     1      464    395    259    
-
56      5    
-
36       C  
 
HETATM   16 C00F UNK     1       7.525   6.686   9.440  1.00  3.35           C  
 
ANISOU   16 C00F UNK     1      3
90    560    322   
 
-
52    
-
65    
-
88       C  
 
HETATM   17 H00F UNK     1       7.418   6.928  10.404  1.00  4.03           H  
 
HETATM   18 C00G UNK     1       4.811   5.993   3.742  1.00  3.21           C  
 
ANISOU   18 C00G UNK     1      535    442    
244      3    
-
49  
   
46       C  
 
HETATM   19 H00G UNK     1       4.694   6.228   2.829  1.00  3.87           H  
 
HETATM   20 C00H UNK     1       5.195   5.818   9.811  1.00  3.11           C  
 
ANISOU   20 C00H UNK     1      432    499    251    
-
37   
 
-
26    
-
40       C
  
 
HETATM   21 H00H UNK     1       5.338   6.012  10.730  1.00  3.71           H  
 
HETATM   22 C00I UNK     1       6.412   1.089   8.322  1.00  3.33           C  
 
ANISOU   22 C00I UNK     1      540    332    393     84    
-
80     22  
     
C  
 
HETATM   2
3 H00I UNK     1       6.146   0.146   8.124  1.00  4.03           H  
 
HETATM   24 C00J UNK     1       2.906   5.039  10.513  1.00  3.66           C  
 
ANISOU   24 C00J UNK     1      499    556    337    
-
90     87      6       C  
 
HETA
TM   25 H00J UNK   
  
1       3.086   5.667  11.270  1.00  4.42           H  
 
HETATM   26 C00K UNK     1       2.729   7.362   4.242  1.00  3.46           C  
 
ANISOU   26 C00K UNK     1      542    369    402     56   
-
138      8       C  
 
HETATM   27 H00K 
UNK     1       2.805   8.215   4.758  1.00  4.18           H  
 
HETATM   28 C00L UNK     1       2.470   1.616   5.236  1.00  3.37           C  
 
ANISOU   28 C
00L UNK     1      508    375    396   
-
123    
-
91     
-
1       C  
 
HETATM   29 H00L UNK     1    
   
2.435   0.648   5.481  1.00  4.03           H  
 
HETATM   30 C00M UNK     1       6.951   4.891   3.027  1.00  4.50           C  
 
ANISOU   30 C00M UNK     1
      
626    720    363    126    103    
-
16       C  
 
 
108
 
HETATM   31 H00M UNK     1       6.632   5.
237   2.146  1.00  5.37           H  
 
HETATM   32 C00N UNK     1       7.938   7.929   8.662  1.00  4.63           C  
 
ANISOU   32 C00N UNK     1      494    
573    691   
-
175   
-
127    
-
76       C  
 
HETATM   33 H00A UNK     1       7.986   7.715   7.708  
1.00  6.95           H  
 
HETATM   34 H00B UNK     1       8.817   8.232   8.973  1.00  6.95           H  
 
HETATM   35 H00C UNK     1       7.277   8.639   8.8
07  1.00  6.95           H  
 
HETATM   36 C00O UNK     1       8.593   5.602   9.318  1.00  5.25   
        
C  
 
ANISOU   36 C00O UNK     1      450    786    758     92   
-
196    
-
62       C  
 
HETATM   37 H00E UNK     1       8.325   4.819   9.841  1.00  7.9
0           H  
 
HETATM   38 H00N UNK     1       9.447   5.944   9.657  1.00  7.90           H  
 
H
ETATM   39 H00O UNK     1       8.693   5.347   8.377  1.00  7.90           H  
 
HETATM   40 C00P UNK     1       1.141   2.260   5.589  1.00  5.03           C
  
 
ANISOU   40 C00P UNK     1      448    593    870    
-
88   
-
172     79       C  
 
HETATM   41 H0
0P UNK     1       1.001   2.207   6.558  1.00  7.58           H  
 
HETATM   42 H00Q UNK     1       0.417   1.789   5.127  1.00  7.58           H  
 
HETATM   4
3 H00R UNK     1       1.148   3.200   5.311  1.00  7.58           H  
 
HETATM   44 C00Q UNK     1 
      
7.868   1.276   7.917  1.00  5.21           C  
 
ANISOU   44 C00Q UNK     1      519    591    870    115   
-
150     75       C  
 
HETATM   45 H00S UNK   
  
1       7.966   1.095   6.958  1.00  7.82           H  
 
HETATM   46 H00T UNK     1       8.431  
 
0.655   8.424  1.00  7.82           H  
 
HETATM   47 H00U UNK     1       8.143   2.197   8.105  1.00  7.82           H  
 
HETATM   48 C00R UNK     1       1.522   5.328   9.984  1.00  5.61           C  
 
ANISOU   48 C00R UNK     1      500    790    840    
 
68    214    233       C  
 
HETATM   49 H00V UNK     1       1.498   6.235   9.611  1.00  8.45           H  
 
HETATM   50 H
00W UNK     1       0.873   5.257  10.715  1.00  8.45           H  
 
HETATM   51 H00X UNK     1       1.298   4.680   9.284  1.00  8.45
           
H  
 
HETATM   52 C00S UNK     1       2.685   7.727   2.758  1.00  4.96           C  
 
ANISOU   52 C00S UNK     1
      
820    563    502    133   
-
244    126       C  
 
HETATM   53 H00Y UNK     1       3.505   8.206   2.515  1.00  7.42           H 
 
 
HETATM   54  H   UNK     1       1.908   8.300   2.585  1.00  7.42           H  
 
HETATM   55  HA  UNK     1       2.613 
  
6.909   2.223  1.00  7.42           H  
 
HETATM   56 C00T UNK     1       1.432   6.684   4.664  1.00  5.66           C  
 
ANISOU   56
 
C00T UNK     1      461    850    840    119   
-
123    168       C  
 
HETATM   57 H00Z UNK     1       1.348   5.823   4.2
05  1.00  8.45           H  
 
HETATM   58  HB  UNK     1       0.673   7.256   4.424  1.00  8.45           H  
 
HETATM   59  HC  UNK    
 
1       1.440   6.540   5.632  1.00  8.45           H  
 
HETATM   60 C00U UNK     1       2.758   1.735   3.744  1.00  5.5
0           C  
 
ANISOU   60 C00U UNK     1     1060    633    395   
-
214   
-
188    
-
83       C  
 
HETATM   61  H00 UNK     1       2.83
7   2.681   3.501  1.00  8.21           H  
 
HETATM   62  HD  UNK     1       2.026   1.328   3.238  1.00  8.21           H
  
 
HETATM   63  HE  UNK     1       3.597   1.273   3.536  1.00  8.21           H  
 
HETATM   64 C00V UNK     1       3.007   3.627  11
.039  1.00  5.42           C  
 
ANISOU   64 C00V UNK     1      710    696    655      2    167    301       C  
 
HETATM   6
5 H0AA UNK     1       2.850   2.994  10.307  1.00  8.13           H  
 
HETATM   66  HF  UNK     1       2.334   3.491  11.738  1.00  8
.13           H  
 
HETATM   67  HG  UNK     1       3.901   3.481  11.413  1.00  8.13           H  
 
HETATM   68 C00W UNK   
  
1       6.204   1.349   9.803  1.00  5.97           C  
 
ANISOU   68 C00W UNK     1     1250    635    382    323   
-
136     87      
 
C  
 
HETATM   69 H1AA UNK     1       6.424   2.282  10.005  1.00  8.92           H  
 
HETATM   70  HH  UNK     1       6.785   0.756  10.324  1.00  8.92           H  
 
HETATM   71  HI  UNK     1       5.269   1.177  10.036  1.00  8.92           H  
 
 
109
 
HETATM  
 
72 C00X UNK     1       7.178   3.420   2.900  1.00  6.97           C  
 
ANISOU   72 C
00X UNK     1     1070    840    740    230    180   
-
323       C  
 
HETATM   73 H2AA UNK     1       7.485   3.060   3.760  1.00 10.42           H  
 
HETATM   74  HJ  UNK 
    
1       7.861   3.252   2.216  1.00 10.42           H  
 
HETATM   75  HK  UNK     1
       
6.342   2.980   2.643  1.00 10.42           H  
 
HETATM   76 C00Y UNK     1       8.217   5.614   3.387  1.00  9.19           C  
 
ANISOU   76 C00Y UNK     1      75
0   1110   1630   
-
250    690   
-
350       C  
 
HETATM   77 H3AA UNK     1       8.055 
  
6.580   3.374  1.00 13.74           H  
 
HETATM   78  HL  UNK     1       8.916   5.391   2.738  1.00 13.74           H  
 
HETATM   79  HM  UNK     1       8.504   5.342 
  
4.284  1.00 13.74           H  
 
CONECT    1    2    3    4    5
 
CONECT    1    6    
7
 
CONECT    2    1    9
 
CONECT    3    1   10
 
CONECT    4    1   15
 
CONECT    5    1   11
 
CONECT    6    1    8
 
CONECT    7    1   12
 
CONECT    8    6   13   28
 
CONECT   
 
9    2   18   26
 
CONECT   10    3   16   20
 
CONECT   11    5   20   24
 
CONECT   12   
 
7   13   22
 
CONECT   13    8   12   14
 
CONECT   14   13
 
CONECT   15    4   18   30
 
CONECT   16   10   17   32   36
 
CONECT   17   16
 
CONECT   18    9   15   19
 
CONECT   1
9   18
 
CONECT   20   10   11   21
 
CONECT   21   20
 
CONECT   22   12   23   44   68
 
CONECT   23   22
 
CONECT   24   11   25   48   64
 
CONECT   25   24
 
CONECT   26    9   27   52   56
 
CONECT   27   26
 
CONECT   28    8   29   40   60
 
CONECT   29   28
 
CONECT   
30   15   31   72   76
 
CONECT   31   30
 
CONECT   32   16   33   34   35
 
CONECT   33
   
32
 
CONECT   34   32
 
CONECT   35   32
 
CONECT   36   16   37   38   39
 
CONECT   37   36
 
CONECT   38   36
 
CONECT   39   36
 
CONECT   40   28   41   42   43
 
 
110
 
CONECT   41   40
 
C
ONECT   42   40
 
CONECT   43   40
 
CONECT   44   22   45   46   47
 
CONECT   45   44
 
CONECT   46   44
 
CONECT   47   44
 
CONECT   48   24   49   50   51
 
CONECT   49   48
 
CONECT   50   48
 
CONECT   51   48
 
CONECT   52   26   53   54   55
 
CONECT   53   52
 
CONECT  
 
54   52
 
CONECT   55   52
 
CONECT   56   26   57   58   59
 
CONECT   57   56
 
CONECT 
  
58   56
 
CONECT   59   56
 
CONECT   60   28   61   62   63
 
CONECT   61   60
 
CONECT   62   60
 
CONECT   63   60
 
CONECT   64   24   65   66   67
 
CONECT   65   64
 
CONECT   66   6
4
 
CONECT   67   64
 
CONECT   68   22   69   70   71
 
CONECT   69   68
 
CONECT   70   
68
 
CONECT   71   68
 
CONECT   72   30   73   74   75
 
CONECT   73   72
 
CONECT   74   72
 
CONECT   75   72
 
CONECT   76   30   77   78   79
 
CONECT   77   76
 
CONECT   78   76
 
CONEC
T   79   76
 
MASTER        0    0    0    0    0    0    0    3   79    0   80    
0
 
END
 
The Following is a PDB file for the crystal structure of Cr(4
-
Cl
-
TMHD)
3
. The structure is quite disordered 
but still shows that the compound was made.
 
HEADER    CSD ENTR
Y jkm519a
 
CRYST1  18.4350  12.4657  18.8857  90.00 118.96  90.00     P21/n
 
SCALE1
      
0.054245  0.000000  0.030024       0.000000
 
SCALE2      0.000000  0.080220  0.000000       0.000000
 
SCALE3      0.000000  0.000000  0.060520       0.000000
 
 
111
 
HETATM    1 C
r1  UNK     1       2.600   4.347  12.216  1.00  4.13          Cr  
 
ANISOU    1 C
r1  UNK     1      518    401    650     47      8    
-
15      Cr  
 
HETATM    2 Cl1  UNK     1       2.177   9.356  12.105  1.00 10.24          Cl  
 
ANISOU    2 Cl1  UNK     1
     
2495    437    960    278      6     26      Cl  
 
HETATM    3  O1  UNK     1
       
2.772   5.774  13.511  1.00  4.37           O  
 
ANISOU    3  O1  UNK     1      662    462    536     67    
-
38     27       O  
 
HETATM    4  O2  UNK     1       2.800 
  
5.719  10.859  1.00  4.44           O  
 
ANISOU    4  O2  UNK     1      626    
524    537     73      6    
-
14       O  
 
HETATM    5  C1  UNK     1       2.508   7.009  13.397  1.00  4.42           C  
 
ANISOU    5  C1  UNK     1      552    511    617   
  
64     22    
-
29       C  
 
HETATM    6  C2  UNK     1       2.404   7.617  12.143  1.00  4.70           C  
 
ANISOU    6  C2  UNK     1      641    508    635     64    
-
26     16       C  
 
HETATM    7  C3  UNK     1       2.567   6.960  10.914  1.00  4.6
0           C  
 
ANISOU    7  
C3  UNK     1      568    564    617     51     21     58       C  
 
HETATM    8  C4  UNK     1       2.265   7.696  14.749  1.00  5.01           C  
 
ANISOU    8  C4  UNK     1      666    589    650     84    
-
26    
-
68       C
  
 
HETATM    9  C5  UNK     1
       
2.284   6.631  15.844  1.00  9.32           C  
 
ANISOU    9  C5  UNK     1     1851   1110    580    376     68    
-
40       C  
 
HETATM   10  H5A UNK     1       1.670   5.906  15.605  1.00 13.98           H  
 
HETATM   1
1  H5B UNK     1       2.004 
  
7.030  16.694  1.00 13.98           H  
 
HETATM   12  H5C UNK     1       3.192   6.273  15.935  1.00 13.98           H  
 
HETATM   13  C6  UNK     1       3.329   8.730  15.053  1.00  8.29           C  
 
ANISOU   13  C6  UNK   
  
1      811   1240   1100   
   
1   
-
152   
-
580       C  
 
HETATM   14  H6A UNK     1       4.183   8.281  15.218  1.00 12.48           H  
 
HETATM   15  H6B UNK     1       3.069   9.241  15.849  1.00 12.48           H  
 
HETATM   16  H6C UNK     1       3.4
20   9.339  14.291  1.00 12.4
8           H  
 
HETATM   17  C7  UNK     1       0.864   8.337  14.795  1.00  7.37           C  
 
ANISOU   17  C7  UNK     1      711   1050   1040     90    156   
-
410       C  
 
HETATM   18  H7A UNK     1       0.804   9.040  1
4.115  1.00 11.05           H
  
 
HETATM   19  H7B UNK     1       0.712   8.727  15.682  1.00 11.05           H  
 
HETATM   20  H7C UNK     1       0.186   7.652  14.619  1.00 11.05           H  
 
HETATM   21  C8  UNK     1       2.451   7.618   9.518  1.00  
5.30           C  
 
ANISOU   2
1  C8  UNK     1      702    650    660     46    
-
50    100       C  
 
HETATM   22  C9  UNK     1       3.591   8.621   9.293  1.00  8.06           C  
 
ANISOU   22  C9  UNK     1     1063   1180    820   
-
192    
-
42    450     
  
C  
 
HETATM   23  H9A UNK   
  
1       3.473   9.388   9.892  1.00 12.08           H  
 
HETATM   24  H9B UNK     1       3.576   8.929   8.363  1.00 12.08           H  
 
HETATM   25  H9C UNK     1       4.449   8.188   9.481  1.00 12.08           H  
 
HETATM 
  
26  C10 UNK     1       1.079   8.262   9.328  1.00  7.83           C  
 
ANISOU   26  C10 UNK     1      983   1040    950    210   
-
160    240       C  
 
HETATM   27 H10A UNK     1       0.382   7.640   9.625  1.00 11.76           H  
 
HETATM   28 H
10B UNK
     
1       0.946   8.472   8.379  1.00 11.76           H  
 
HETATM   29 H10C UNK     1       1.029   9.086   9.854  1.00 11.76           H  
 
HETATM   30  C11 UNK     1       2.576   6.491   8.445  1.00  7.83           C  
 
ANISOU   30  C11 UNK     1
     
12
44   1200    530    132      4     50       C  
 
HETATM   31 H11A UNK     1       3.415   6.004   8.580  1.00 11.69           H  
 
HETATM   32 H11B UNK     1       2.568   6.890   7.550  1.00 11.69           H  
 
HETATM   33 H11C UNK     1       1.821 
  
5.872
   
8.533  1.00 11.69           H  
 
HETATM   34 Cl2  UNK     1       4.675   0.015  13.140  0.75  7.50          Cl  
 
ANISOU   34 Cl2  UNK     1      944    485   1419    199    483    132      Cl  
 
HETATM   35 Cl3  UNK     1      
-
1.407   2.057  10.0
93  0.7
5  7.95          Cl  
 
 
112
 
ANISOU   35 Cl3  UNK     1      659   1014   1349   
-
108     18   
-
408      Cl  
 
HETATM   36  O3  UNK     1       2.311   3.040  13.620  0.75  4.71           O  
 
ANISOU   36  O3  UNK     1      490    440    860     13      4  
  
100  
     
O  
 
HETATM   37  O4  UNK     1       4.482   3.852  12.356  0.75  4.16           O  
 
ANISOU   37  O4  UNK     1      508    434    640     17    
-
46     10       O  
 
HETATM   38  O5  UNK     1       0.673   4.556  12.254  0.75  4.36           O
  
 
ANIS
OU   38  O5  UNK     1      537    480    640     86     34     10       O  
 
HETATM   39  O6  UNK     1       2.274   2.959  10.892  0.75  4.58           O  
 
ANISOU   39  O6  UNK     1      479    480    780    
-
42     71   
-
130       O  
 
HETATM   4
0  C12 
UNK     1       2.940   1.968  13.863  0.75  4.46           C  
 
ANISOU   40  C12 UNK     1      535    482    679     28    112     40       C  
 
HETATM   41  C13 UNK     1       4.132   1.687  13.184  0.75  4.52           C  
 
ANISOU   41  C13 UNK   
  
1    
  
541    489    687     53    104     26       C  
 
HETATM   42  C14 UNK     1       4.927   2.679  12.581  0.75  4.57           C  
 
ANISOU   42  C14 UNK     1      522    533    682     35    115      2       C  
 
HETATM   43  C15 UNK     1       2.233   1.
017  14.865  0.75  4.56           C  
 
ANISOU   43  C15 UNK     1      548    495    690     13    105     63       C  
 
HETATM   44  C16 UNK     1       1.169   1.827  15.618  0.75  5.52           C  
 
ANISOU   44  
C16 UNK     1      758    640    700     44
    
184    210       C  
 
HETATM   45 H16A UNK     1       1.602   2.545  16.124  0.75  8.29           H  
 
HETATM   46 H16B UNK     1       0.686   1.238  16.236  0.75  8.29           H  
 
HETATM   47 H16C UNK     1
       
0.537   2.215  14.977  0.75  8.29   
        
H  
 
HETATM   48  C17 UNK     1       1.494  
-
0.026  14.014  0.75  6.26           C  
 
ANISOU   48  C17 UNK     1      948    630    800   
-
211    220     30       C  
 
HETATM   49 H17A UNK     1       0.770 
  
0.411  13.517  0.75  9.32           H  
 
H
ETATM   50 H17B UNK     1       1.118  
-
0.718  14.598  0.75  9.32           H  
 
HETATM   51 H17C UNK     1       2.121  
-
0.437  13.383  0.75  9.32           H  
 
HETATM   52  C18 UNK     1       3.186   0.425  15.9
04  0.75  6.17           C  
 
ANISOU   52  C
18 UNK     1      805    760    780     73     48    260       C  
 
HETATM   53 H18A UNK     1       3.751  
-
0.253  15.479  0.75  9.24           H  
 
HETATM   54 H18B UNK     1       2.666   0.012  16.625  0.75  9.2
4           H  
 
HETATM   55 H18C UNK     1 
      
3.749   1.137  16.272  0.75  9.24           H  
 
HETATM   56  C19 UNK     1       6.389   2.498  12.191  0.75  4.66           C  
 
ANISOU   56  C19 UNK     1      516    565    690     30    120      8       C
  
 
HETATM   57  C20 UNK     1       7.094  
 
1.720  13.310  0.75  7.75           C  
 
ANISOU   57  C20 UNK     1     1264    810    870    178     
-
1     70       C  
 
HETATM   58 H20A UNK     1       7.099   2.259  14.129  0.75 11.61           H  
 
HETATM   5
9 H20B UNK     1       8.017   1.526  13.04
2  0.75 11.61           H  
 
HETATM   60 H20C UNK     1       6.617   0.880  13.474  0.75 11.61           H  
 
HETATM   61  C21 UNK     1       6.458   1.631  10.904  0.75  6.15           C  
 
ANISOU   61  C21 UNK   
  
1      567    900    870     45    239   
 
-
60       C  
 
HETATM   62 H21A UNK     1       6.360   0.685  11.141  0.75  9.24           H  
 
HETATM   63 H21B UNK     1       7.323   1.767  10.463  0.75  9.24           H  
 
HETATM   64 H21C UNK     1       5.737   1.891  10.295  0.75  9.24           H 
 
 
HETATM   65  C22 UNK     1       7.083   3.795  12.031  0.75  6.90           C  
 
ANISOU   65  C22 UNK     1      560    820   1240    
-
56    
-
23     80       C  
 
HETATM   66 H
22A UNK     1       6.727   4.260  11.244  0.75 10.34           H  
 
HETATM   67
 
H22B UNK     1       8.043   3.640  11.912  0.75 10.34           H  
 
HETATM   68 H22C UNK     1       6.936   4.344  12.828  0.75 10.34           H  
 
HETATM   69  C23 UNK     1
      
-
0.264   3.851  11.771  0.75  4.71           C  
 
ANISOU   69  C23 UNK    
 
1      530    572    687     20     82      7       C  
 
 
113
 
HETATM   70  C24 UNK     1      
-
0.008   2.852  10.833  0.75  4.76           C  
 
ANISOU   70  C24 UNK     1      556    
554    700    
-
20     66     
-
4       C  
 
HETATM   71  C25 UNK     1       1.26
6   2.487  10.390  0.75  4.76           C  
 
ANISOU   71  C25 UNK     1      566    545    699    
-
10     73    
-
34       C  
 
HETATM   72  C26 UNK     1      
-
1.657   4.149  12.3
93  0.75  4.88           C  
 
ANISOU   72  C26 UNK     1      555    600    700 
    
38     69     
-
5       C  
 
HETATM   73  C27 UNK     1      
-
1.446   5.257  13.460  0.75  7.77           C  
 
ANISOU   73  C27 UNK     1      652   1430    870    211     79  
 
-
320       C  
 
HETATM   74 H27A UNK     1      
-
0.888   4.909  14.187  0.75 11
.69           H  
 
HETATM   75 H27B UNK     1      
-
2.314   5.535  13.820  0.75 11.69           H  
 
HETATM   76 H27C UNK     1      
-
1.001   6.027  13.048  0.75 11.69           H
  
 
HETATM   77  C28 UNK     1      
-
2.585   4.773  11.342  0.75  6.78          
 
C  
 
ANISOU   77  C28 UNK     1      677    980    920    216    
-
63    
-
60       C  
 
HETATM   78 H28A UNK     1      
-
2.133   5.530  10.914  0.75 10.19           H  
 
HETATM   7
9 H28B UNK     1      
-
3.407   5.086  11.775  0.75 10.19           H  
 
HETATM  
 
80 H28C UNK     1      
-
2.809   4.101  10.664  0.75 10.19           H  
 
HETATM   81  C29 UNK     1      
-
2.285   2.947  13.022  0.75  7.77           C  
 
ANISOU   81  C29 UNK   
  
1      732   1240    980     48    201    140       C  
 
HETATM   82 H29A UNK 
    
1      
-
2.935   2.555  12.401  0.75 11.69           H  
 
HETATM   83 H29B UNK     1      
-
2.741   3.210  13.848  0.75 11.69           H  
 
HETATM   84 H29C UNK     1      
-
1.591   2.286  13.228  0.75 11.69           H  
 
HETATM   85  C30 UNK     1       1
.558   1.466   9.266  0.75  5.01           C  
 
ANISOU   85  C30 UNK     1      614    569    720     10     48    
-
61       C  
 
HETATM   86  
C31 UNK     1       0.904   1.948   7.900  0.75  7.02           C  
 
ANISOU   86  C31 UNK     1     1236    850    5
80   
-
161     45    
-
40       C  
 
HETATM   87 H31A UNK     1      
-
0.062   2.066   8.024  0.75 10.58           H  
 
HETATM   88 H31B UNK     1
       
1.063   1.275   7.205  0.75 10.58           H  
 
HETATM   89 H31C UNK     1       1.304   2.800   7.628  0.75
 
10.58           H  
 
HETATM   90  C32 UNK     1       1.059   0.067   9.648  0.75  6.34           C  
 
ANISOU   90  C32 UNK     1      918    
620    870    142   
-
173    
-
70       C  
 
HETATM   91 H32A UNK     1       1.425  
-
0.183  10.522  0.75  9.47       
    
H  
 
HETATM   92 H32B UNK     1       1.354  
-
0.580   8.973  0.75  9.47           H  
 
HETATM   93 H32C UNK     1       0.080   0.070   9.6
92  0.75  9.47           H  
 
HETATM   94  C33 UNK     1       3.056   1.421   9.014  0.75  7.78           C  
 
ANISO
U   94  C33 UNK     1      805   1100   1050    
-
21    139   
-
520       C  
 
HETATM   95 H33A UNK     1       3.398   2.334   8.920  0.75 11.6
9           H  
 
HETATM   96 H33B UNK     1       3.233   0.919   8.191  0.75 11.69           H  
 
HETATM   97 H33C U
NK     1       3.501   0.981   9.768  0.75 11.69           H  
 
HETATM   98 Cl2B UNK     1       4.564   0.014  11.072  0.25  7.50          Cl
  
 
ANISOU   98 Cl2B UNK     1      944    485   1419    199    483    132      Cl  
 
HETATM   99 Cl3B UNK     1     
 
-
1.213   2.132  14.366  0.25  7.95          Cl  
 
ANISOU   99 Cl3B UNK     1      659   1014   1349   
-
108     18   
-
408      Cl  
 
HETATM  10
0  O3B UNK     1       2.539   3.091  13.698  0.25  4.71           O  
 
ANISOU  100  O3B UNK     1      490    440  
  
860     30      4     90       O  
 
HETATM  101  O4B UNK     1       0.707   4.774  12.112  0.25  4.36           O  
 
ANISOU  101  O4B UNK   
  
1      537    480    640     86     34     10       O  
 
HETATM  102  O5B UNK     1       4.462   3.827  11.918  0
.25  4.28           O  
 
ANISOU  102  O5B UNK     1      536    440    650     17    
-
63     10       O  
 
HETATM  103  O6B UNK     1       2.454   3.191  10.654  0.25  4.54           O  
 
ANISOU  103  O6B UNK     1      466    480    780    
-
37     83   
-
140
       
O  
 
HETATM  104 C12B UNK     1       1.446   2.474  13.855  0.25  4.63           C  
 
 
114
 
ANISOU  104 C
12B UNK     1      560    507    690     20     94     47       C  
 
HETATM  105 C13B UNK     1       0.152   2.916  13.577  0.25  4.83           C  
 
AN
ISOU  105 C13B UNK     1      584    550    700     18     70     30       C  
 
HETATM  106 C14B UNK     1
      
-
0.152   3.985  12.614  0.25  4.84           C  
 
ANISOU  106 C14B UNK     1      557    580    700     27     70      9       C  
 
HETATM  107 C15
B UNK     1       1.761   1.295  14.830  0.25  4.61           C  
 
ANISOU  107 C15B UNK     1      563    
498    690     18    103     55       C  
 
HETATM  108 C16B UNK     1       1.288  
-
0.087  14.359  0.25  5.75           C  
 
ANISOU  108 C16B UNK     1  
    
956    560    670   
-
107    155     70       C  
 
HETATM  109 H16D UNK     1       0.471  
-
0.328  14.8
52  0.25  8.69           H  
 
HETATM  110 H16E UNK     1       1.985  
-
0.746  14.547  0.25  8.69           H  
 
HETATM  111 H16F UNK     1       1.100  
-
0.058  13.405  0.25  8.69           H  
 
HETATM  112 C17B UNK     1       3.286   1.195  14.980  0.25  6.9
0           C  
 
ANISOU  112 C17B UNK     1     1062    690    870     
-
8   
-
162    180       C  
 
HETATM  113 H17D UNK     1       3.715   1.475  14.145
  
0.25 10.34           H  
 
HETATM  114 H17E UNK     1       3.534   0.268  15.181  0.25 10.34           H
  
 
HETATM  115 H17F UNK     1       3.581   1.777  15.711  0.25 10.34           H  
 
HETATM  116 C18B UNK     1       1.190   1.558  16.215  0.25  6.16 
          
C  
 
ANISOU  116 C18B UNK     1      939    680    720    179    
-
33    110       C  
 
HETATM  11
7 H18D UNK     1       1.913   1.530  16.876  0.25  9.24           H  
 
HETATM  118 H18E UNK     1       0.519   0.884  16.428  0.25  9.24           H  
 
HETATM  119 H18F UNK     1       0.775   2.451  16.229  0.25  9.24           H  
 
HETATM  120 C19B UNK   
  
1      
-
1.558   4.510  12.247  0.25  4.85           C  
 
ANISOU  120 C19B UNK     1      551    590    700     27     75      5       C  
 
HETATM  121 
C20B UNK     1      
-
1.428   5.683  11.324  0.25  6.99           C  
 
ANISOU  121 C20B UNK     1      865    880    910    144   
-
184     70       C  
 
HETATM  122 H20D UNK     1      
-
1.007   6.430  11.800  0.25 10.50           H  
 
HETATM  123 H20E UNK     
1      
-
2.316   5.953  11.011  0.25 10.50           H  
 
HETATM  124 H
20F UNK     1      
-
0.873   5.433  10.554  0.25 10.50           H  
 
HETATM  125 C21B UNK     1      
-
2.443   4.862  13.430  0.25  7.05           C  
 
ANISOU  125 C21B UNK     1      720   
1080    880     22     62    
-
20       C  
 
HETATM  126 H21D UNK     1
      
-
2.213   4.283  14.191  0.25 10.58           H  
 
HETATM  127 H21E UNK     1      
-
3.383   4.715  13.188  0.25 10.58           H  
 
HETATM  128 H21F UNK     1      
-
2.308   5.795  13.
677  0.25 10.58           H  
 
HETATM  129 C22B UNK     1      
-
2.285 
  
3.386  11.466  0.25  7.64           C  
 
ANISOU  129 C22B UNK     1     1024    920    960     22   
-
211    
-
40       C  
 
HETATM  130 H22D UNK     1      
-
1.815   3.215  10.623  0.25 11.
45           H  
 
HETATM  131 H22E UNK     1      
-
3.206   3.663  11.2
76  0.25 11.45           H  
 
HETATM  132 H22F UNK     1      
-
2.295   2.567  12.005  0.25 11.45           H  
 
HETATM  133 C23B UNK     1       4.975   2.699  11.611  0.25  4.66           
C  
 
ANISOU  133 C23B UNK     1      530    552    690     29    111  
    
3       C  
 
HETATM  134 C24B UNK     1       4.113   1.729  11.044  0.25  4.77           C  
 
ANISOU  134 C24B UNK     1      565    547    700     22     87    
-
13       C  
 
HETATM  1
35 C25B UNK     1       2.917   2.090  10.398  0.25  4.89           C
  
 
ANISOU  135 C25B UNK     1      590    557    710      4     61    
-
39       C  
 
HETATM  136 C26B UNK     1       6.466   2.418  11.963  0.25  4.65           C  
 
ANISOU  136 C26B UNK  
   
1      518    558    690     33    120      5       C  
 
HETATM  13
7 C27B UNK     1       6.818   3.913  11.590  0.25  4.84           C  
 
ANISOU  137 C27B UNK     1      291    710    840     94     95    120       C  
 
HETATM  138 H27D UNK     1       6.
202   4.231  10.897  0.25  7.26           H  
 
HETATM  139 H27E UNK   
  
1       7.738   3.959  11.253  0.25  7.26           H  
 
 
115
 
HETATM  140 H27F UNK     1       6.734   4.478  12.386  0.25  7.26           H  
 
HETATM  141 C28B UNK     1       7.337   2.169  
13.244  0.25  4.87           C  
 
ANISOU  141 C28B UNK     1      350    690    810     36    203    100       C  
 
HETATM  142 H28D UNK     1       7.083   2.816  13.939  0.25  7.34           H  
 
HETATM  143 H28E UNK     1       8.284   2.289  13.023  0.25 
 
7.34           H  
 
HETATM  144 H
28F UNK     1       7.186   1.262  13.571  0.25  7.34           H  
 
HETATM  145 C29B UNK     1       7.232   1.645  10.854  0.25  6.62           C  
 
ANISOU  145 C29B UNK     1      743    980    790     74     66     10    
   
C  
 
HETATM  146 H29D UNK     1
       
6.618   1.043  10.393  0.25  9.95           H  
 
HETATM  147 H29E UNK     1       7.961   1.134  11.258  0.25  9.95           H  
 
HETATM  148 H29F UNK     1       7.605   2.290  10.210  0.25  9.95           H  
 
HETATM
  
149 C30B UNK     1       2.210 
  
1.190   9.387  0.25  4.92           C  
 
ANISOU  149 C30B UNK     1      599    560    710      8     53    
-
50       C  
 
HETATM  150 C31B UNK     1       3.121   0.598   8.293  0.25  7.71           C  
 
ANISOU  150 C31B UN
K     1      941   1060    930   
 
136    
-
69   
-
390       C  
 
HETATM  151 H31D UNK     1       3.648   1.323   7.887  0.25 11.61           H  
 
HETATM  152 H31E UNK     1       2.570   0.174   7.606  0.25 11.61           H  
 
HETATM  153 H31F UNK     1      
 
3.723  
-
0.059   8.691  0.25 11.6
1           H  
 
HETATM  154 C32B UNK     1       1.484   0.025  10.048  0.25  6.31           C  
 
ANISOU  154 C32B UNK     1      968    570    860    
-
34   
-
133     40       C  
 
HETATM  155 H32D UNK     1       2.092  
-
0.43
3  10.666  0.25  9.47           H
  
 
HETATM  156 H32E UNK     1       1.185  
-
0.606   9.359  0.25  9.47           H  
 
HETATM  157 H32F UNK     1       0.707   0.360  10.542  0.25  9.47           H  
 
HETATM  158 C33B UNK     1       1.153   2.006   8.619  0.
25  5.78           C  
 
ANISOU  15
8 C33B UNK     1      916    630    650    
-
86    
-
28     10       C  
 
HETATM  159 H33D UNK     1       0.572   2.471   9.255  0.25  8.69           H  
 
HETATM  160 H33E UNK     1       0.614   1.402   8.065  0.25  8.69     
      
H  
 
HETATM  161 H33F UNK   
  
1       1.600   2.662   8.044  0.25  8.69           H  
 
CONECT    1    3    4   36   37
 
CONECT    1   38   39
 
CONECT    2    6
 
CONECT    3    1    5
 
CONECT    4    1    7
 
CONECT    5    3    6    8
 
CONECT    6    2    5  
  
7
 
CONECT    7    4    6   21
 
CONECT    8    5    9   13   17
 
CONECT    9    8   10   11   12
 
CONECT   10    9
 
CONECT   11    9
 
CONECT   12    9
 
CONECT   13    8   14   15   16
 
CONECT   14   13
 
CONECT   15   13
 
CONECT   16   13
 
CONECT   17    8   18   19 
  
20
 
CONECT   18   17
 
CONECT   
19   17
 
CONECT   20   17
 
CONECT   21    7   22   26   30
 
CONECT   22   21   23   24   25
 
 
116
 
CONECT   23   22
 
CONECT   24   22
 
CONECT   25   22
 
CONECT   26   21   27   28   29
 
CONECT   27   26
 
CONECT   28   26
 
CONECT   29   26
 
CO
NECT   30   21   31   32   33
 
CONECT   31   30
 
CONECT   32   30
 
CONECT   33   30
 
CONECT   34   41
 
CONECT   35   70
 
CONECT   36    1   40
 
CONECT   37    1   42
 
CONECT   38    1   69
 
CONECT   39    1   71
 
CONECT   40   36   41   43
 
CONECT   41   34   40   42
 
CONECT   42   37   41   56
 
CO
NECT   43   40   44   48   52
 
CONECT   44   43   45   46   47
 
CONECT   45   44
 
CONECT   46   44
 
CONECT   47   44
 
CONECT   48   43   49   50   51
 
CONECT   49   48
 
CONECT   50   48
 
CONECT   51   48
 
CONECT   52   43   53   54   5
5
 
CONECT   53   52
 
CONECT   54
   
52
 
CONECT   55   52
 
CONECT   56   42   57   61   65
 
CONECT   57   56   58   59   60
 
CONECT   58   57
 
CONECT   59   57
 
CONECT   60   57
 
CONECT   61   56   62   63   64
 
CONECT   62   61
 
CONECT   63   61
 
CONECT   64   61
 
CONEC
T   65   56   66   67   68
 
CONECT   66   65
 
CONECT   67   65
 
CONECT   68   65
 
CONECT   69   38   70   72
 
CONECT   70   35   69   71
 
CONECT   71   39   70   85
 
CONECT   72   69   73   77   81
 
CONECT   73   72   74   75   76
 
 
117
 
CONECT   74   73
 
CONECT   75   73
 
CONECT   76   73
 
CONECT   
77   72   78   79   80
 
CONECT   78   77
 
CONECT   79   77
 
CONECT   80   77
 
CONECT   81   72   82   83   84
 
CONECT   82   81
 
CONECT   83   81
 
CONECT   84   81
 
CONECT   85   71   86   90   94
 
CONECT   86   85   87   88   89
 
CONECT  
 
87   86
 
CONECT   88   86
 
CONECT   89   86
 
CONECT   90   85   91   92   93
 
CONECT   91   90
 
CONECT   92   90
 
CONECT   93   90
 
CONECT   94   85   95   96   97
 
CONECT   95   94
 
CONECT   96   94
 
CONECT   97   94
 
MASTER        0    0    0    0    0    0    0  
  
3  161    0   98    0
 
EN
D
 
 
 
Figure
 
2.A.2.
 
Crystal structure for Cr(4
-
Cl
-
TM
HD)
3
 
that was collected by Sara Adelman and 
solved by Professor Richard Staples. The structure is highly disordered but still shows the desired 
compound.
 
 
 
118
 
Appendix 2.B. Dephasing Times of Common Solvent 
Modes
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Solvent
 
Coherent Mode and Dephasing Time
 
THF
 
 
913 (6) cm
-
1
 
 
 
950 (140) fs
 
 
MeCN
 
377 (11) cm
-
1
 
92
2 (11) cm
-
1
 
 
600 (250) fs
 
1300 (900) fs
 
 
Dioxane
 
425 (5) cm
-
1
 
836 (4)
 
 
440 (200) fs
 
2100 (1100) fs
 
 
Benzene
*
 
 
995 cm
-
1
 
 
 
~3000 fs
 
 
*
Only from a single measurement
 
Table
 
2.B.1.
 
Frequencies and dephasing times of common solvent vibrational 
modes as
 
determined from transient absorption spectroscopy.
 
 
119
 
Appendix 2.C. Fit of Cr(acac)
3
 
Raman Spectrum
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 
2.C.1.
 
Resonance Raman spectrum of Cr(acac)
3
 
in THF solution (blue) in 
the vicinity of the 455 cm
-
1
 
solute mode. The 488
 
nm excitation source is resonant 
with the blue edge of the 
4
A
2
 

4
T
2
 
transition. A fit of the 455 cm
-
1
 
band to a 
Lorentzian function with a polynomial background (red) shows a FWHM of 8.5 
±
 
0.1 cm
-
1
 
corresponding to a dephasing time o
f 1.25 
±
 
0.02 ps. 
 
 
120
 
 
 
 
 
 
 
 
 
 
 
 
 
REFERENCES
 
 
 
 
121
 
REFERENCES
 
 
(1)
 
Bolletta, F.; Maestri, M.; Moggi, L.; Balz
ani, V. Bimolecular Electron Transfer Reactions of 
Electronically Excited States of Co
-
Ordination Compounds. 
J. Chem. Soc., Chem. Commun
. 
1975
, 
901

902.
 
(2)
 
Kirk, A. D. Photochemistry and Photophysics of Chromium(I
II) Complexes. 
Chem. Rev
. 
1999
, 
99
, 
1607

1
640.
 
(3)
 
Maiman, T. H. Stimulated Optical Radiation in Ruby. 
Nature
 
1960
, 
187
, 493

494.
 
(4)
 
Becquerel, E. La Lumiere, Ses Causes et Ses Effects. Gauthier
-
Villars:Paris. 
1867
.
 
(5)
 
Chromium https://www.rsc.org/period
ic
-
table/element/24/chromium.
 
(6)
 
Shriver,
 
D. F.; Atkins, P. W.; Langford, C. H.; Overton, T. L.; Rourke, J. P.; Weller, M. T.; 
Armstrong, F. A. Inorganic Chemistry, 5th ed.; W. H. Freeman and Company: New York, 2010.
 
(7)
 
Juban, E. A.; McCusker, J. K. Ultr
afast Dynamics of 
2
E State Formation in Cr
(acac)
3
. 
J. Am. Chem. 
Soc.
 
2005
, 
127
, 6857

6865.
 
(8)
 
Ando, H.; Iuchi, S.; Sato, H. Theoretical Study on Ultrafast Intersystem Crossing of Chromium(III) 
Acetylacetonate. 
Chem. Phys. Lett
. 
2012
, 
535
, 177

181.
 
(9)
 
Bendix, J.; Borson, M.; Schäffer, C. E. Accur
ate Empirical Spin
-
Orbit Coupling Parameters. 

nd
 
for
 
Gaseous nd
q
 
Transition Metal Ions. The Parametrical Multiplet Term Model. 
Inorg. Chem
. 
1993
, 
32
, 2838

2849.
 
(10)
 
Schrauben, J. N.; Dillman, K. L.; Beck, W. F.; McCusker, J. K. Vibrational Coherence in t
he 
Excited State Dynamics of Cr(acac)
3
: Probing the R
eaction Coordinate for Ultrafast Intersystem 
Crossing. 
Chem. Sci
. 
2010
, 
1
, 405

410.
 
(11)
 
McCusker, J. K. Femtosecond Absorption Spectroscopy of Transition Metal Charge
-
Transfer 
Complexes. 
Acc. Chem. Res
.
 
2003
, 
36
, 879

887.
 
(12)
 
Foszcz, E. D. Understanding 
the Interplay Between Geometry and Ultrafast Dynamics in Ligand 
Field Excited States of Inorganic Chromophores, Michigan State University, 2015.
 
(13)
 
Ma


as, E. M. S.; Kananavicius, R.; Myllyperkiö, P.
; Pettersson, M.; Kunttu, H. Relaxation 
Dynamics of C
r(acac)
3
 
Probed by Ultrafast Infrared Spectroscopy. 
J. Am. Chem. Soc.
 
2007
, 
129
, 
8934

8935.
 
(14)
 
Ma


as, E. M. S.; Mustalahti, S.; Myllyperkiö, P.; Kunttu, H.; Pettersson, M. Role of Vibrational 
Dynami
cs in Electronic Relaxation of Cr(acac)
3
. 
J. Phys. Ch
em. A
 
2015
, 
119
, 2727

2734.
 
(15)
 
Schrauben, J. N. Electronic Structure and Excited State Dynamics of Chromium(III) Complexes, 
Michigan State University, 2010.
 
(16)
 
Collman, J. P.; Moss, R. A.; Maltz, H.;
 
Heindel, C. C. The Reaction of Metal Chelates. I. 
Ha
logenation of Metal Chelates of 1,3
-
Diketones. 
J. Am. Chem. Soc.
 
1961
, 
83
, 531

534.
 
 
122
 
(17)
 
Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; 
Scalmani, G.; Baron
e, V.; Petersson, G. A.; Nakatsuji, H.; 
et al
. Gaussi
an, Inc.: Wallingford CT 
2009.
 
(18)
 
Humphrey, W.; Dalke, A.; Schulten, K. VMD: Visual Molecular Dynamics. 
J. Mol. Graph. Model.
 
1996
, 
14
, 33

38.
 
(19)
 
Brown, A. M.; McCusker, C. E.; Carey, M. C.; Blanco
-
R
odríguez, A. M.; Towrie, M.; Clark, I. P.; 

.; McCusker, J. K. Vibrational Relaxation and Redistribution Dynamics in Ruthenium(II) 
Polypyridyl
-
Based Charge
-
Transfer Excited State: A Combined Ultrafast Electronic and Infrared 
Absorption Study. 
J. C
hem. Phys. A
 
2018
, 
122
, 7941

7953.
 
(20)
 
Albrecht, H.;
 
Heist, P.; Kleinschmidt, J.; Lap, D. V. Single
-
Shot Measurement of Ultraviolet and 
Visible Femtosecond Pulses Using the Optical Kerr Effect. 
Appl. Opt.
 
1993
, 
32
, 6659

6663.
 
(21)
 
Righini, R. Ultrafast Op
tical Kerr Effect in Liquids and Solids. 
Science
 
1993
, 
262
, 1386
-
1390.
 
(22)
 
Slavov, C.; Hartmann, C.; Wachtveitl, J. Implementation and Evaluation of Data Analysis 
Stratefies for Time
-
Resolved Optical Spectroscopy. 
Ann. Chem.
 
2018
, 
87
, 2328

2336.
 
(23)
 
Rosc
a, F.; Kumar, A. T. N.; Ye, X.; Sjodin, T.; Demid
ov, A. A.; Champion, P. M. Investigations 
of Coherent Vibrational Oscillations in Myoglobin. 
J. Phys. Chem. A
 
2000
, 
104
, 4280

4290.
 
(24)
 
Hamm, P.; Zanni, M. Concepts and Methods of 2D Infrared Spectroscopy; 
Cambridge University 
Press: Cambridge, 2011.
 
(25)
 
Kim, J.; Kim, D. O.; Joo, T. Excited
-
State Dynamics of Thioflavin T: Planar Stable Intermediate 
Revealed by Nuclear Wave Packet Spectroscopies. 
J. Phys. Chem. A
 
2018
, 
122
, 1283

1290.
 
(26)
 
De Silvestri, S.; 
Cerullo, G.; Lanzani, G. Coherent Vibrational
 
Dynamics; CRC Press: Boca Raton, 
FL, 2008.
 
(27)
 
Reid, P. J.; Silva, C.; Barbara, P. F. Electronic Coherence, Vibrational Coherence, and Solvent 
Degrees of Freedom in the Femtosecond Spectroscopy of Mixed
-
Valenc
e Metal Dimers in H2O 
and D2O.
 
J. Phys. Chem.
 
1995
,
 
99
, 2609

2616.
 
(28)
 
Monni, R.; Auböck, G.; Kinschel, D.; Aziz
-

Conservation of Vibrational Coherence in Ultrafast Electronic Relaxation: The Case of Dipla
tinum 
Complexes in Solution. 
Chem. Phys. Lett
.
 
2017
, 
683
, 112

120.
 
(29)
 
Wang, Q.; Schoenlein, R. W.; Peteanu, L. A.; Mathies, R. A.; Shank, C. V. Vibrationally Coherent 
Photochemistry in the Femtosecond Primary Event of Vision. 
Science
 
1994
, 
266
, 422

424.
 
(30)
 
Rafiq, S.; Scholes, G. D. Slow Intramole
cular Vibrational Rela
xation Leads to Long
-
Lived Excited
-
State Wavepackets. 
J. Chem. Phys. A
 
2016
, 
120
, 6792

6799.
 
(31)
 
Rafiq, S.; Scholes, G. D. From Fundamental Theories to Quantum Coherences in Electron Transfer. 
J. Am. Chem. Soc
. 
2019
, 
141
, 708

722.
 
(3
2)
 
Cho, H.; Strader, M
. L.; Hong, K.; Jamula, L.; Gullikson, E. M.; Kim, T. K.; de Groot, F. M. F.; 
McCusker, J. K. Ligand
-
Field Symmetry Effects in Fe(II) Polypyridyl Compounds Probed by 
Transient X
-
Ray Absorption Spectroscopy. 
Faraday Discuss
. 
2012
, 
157
,
 
463

474.
 
 
123
 
(33)
 
Gershgo
ren, E.; Wang, Z.; Ruhman, S.; Vala, J.; Kosloff, R. Investigating Pure Vibrational 
Dephasing of I
3
-
 
in Solution: Temperature Dependence of T
2
* for the Fundamental and First 
Harmonic of 

1
. 
J. Chem. Phys
. 
2003
, 
118
, 3660

3667.
 
(34)
 
Fu
rukawa, T.; Fox, K. E.
; White, W. B. Raman Spectroscopic Investigation of the Structure of 
Silicate Glasses III. Raman Intensities and Structural Units in Sodium Silicate Glasses. 
Chem. Phys.
 
1981
, 
75
, 3226

3237.
 
(35)
 
Liebel, M.; Schnedermann, C.; Wende, T
.; Kukura, P. Principl
es and Applications of Broadband 
Impulsive Vibrational Spectroscopy. 
J. Chem. Phys. A
 
2015
, 
119
, 9506

9517.
 
(36)
 
Myers, A. B. MOLECULAR ELECTRONIC SPECTRAL BROADENING IN LIQUIDS AND 
GLASSES. 
Annu. Rev. Phys. Chem
. 
1998
, 
49
, 267

295.
 
(37)
 
Kubo, R. Fluctuati
on, Relaxation and Resonance in Magnetic Systems; ter Haar, D.; Plenum Press: 
New York, 1962; pp 23

69.
 
(37)
 
Perk
ovic, M. W.; Endicott, J. F. Stereochemical Tuning of Chromium(III) Photophysics With 

-
tris(alkylamine)
-
1,4,7
-
triazacyclononane complexe
s. 
J. Phys. Chem
. 
1990
, 
94
, 1217

1219.
 
(39)
 
Perkovic, M. W.; Heeg, M. J.; Endicott, J. F. Stereochemical Perturba
tions of the Relaxation 
Behavior of (2E)Chromium(III). Ground
-
State x
-
Ray Crystal Structure, Photophysics, and 
Molecular Mechanics Simulations 
of the Quasi
-

-
Ethylidynetris(3
-
Azabutan
-
1
-
Amine)]Chromium Tribromid. 
Inorg. Chem.
 
1991
, 
30
,
 
3140

3147.
 
(40)
 
Purcell, K. F. Pseudorotational Intersystem Crossing in d
6
 
Complexes. 
J. Am. Chem. Soc. 
1979
, 
101
, 5147

5152.
 
(41)
 
Vanquickenb
orne, L. G.; Pierloot, K. Role of Spin Change in the Stereomobile Reactions of Strong
-
Field d
6
 
Transition
-
Metal C
omplexes. 
Inorg. Chem
. 
1981
, 
20
, 3673

3677.
 
(42)
 
McCusker, J. K.; Rheingold, A. L.; Hendrickson, D. N. Variable
-
Temperature Studies of Laser
-
In
itiated 
5
T
2
 

1
A
1
 
Intersystem Crossing in Spin
-
Crossover Complexes: Empirical Correlations 
between Activation Par
ameters and Ligand Structure in a Series of Polypyridyl Ferrous Complexes. 
Inorg. Chem
. 
1996
, 
35
, 2100

2112.
 
(43)
 
Stock, P.; Deck, E.; Hohnstei
n, S.; Korzekwa, J.; Meyer, K.; Heinemann, F. W.; Breher, F.; Hörner, 
G. Molecular Spin Crossover in Slow Motion:
 
Light
-
Induced Spin
-
State Transitions in Trigonal 
Prismatic Iron(II) Complexes. 
Inorg. Chem
. 
2016
,
 
55
, 5254

5265.
 
(44)
 
Lewis, G. N.; Calvin, M.
 
The Color of Organic Substances. 
Chem. Rev
. 
1939
, 
25
, 273

328.
 
 
 
 
124
 
Chapter 3: Excited State Dynamics of a 

3
 
Derivatives
 
3.1
.
 
Introduction
 
 
The previous chapter established a kinetic model to describe the excited state dynamics in a series
 

3
 
derivatives. Furthermore, the relationship between dephasing times of the coherent vibrational 

3
 
complexes and their dynamical role in the 
photophysics was discussed. It also illustra
ted the simila

3
 
compounds studied following excitation into the lowest energy spin allowed ligand field band. We next 
assess how these results transfer to other transition metal systems.
 
 
Aside from 
d
3
 
chr
omium compound
s, the next most studied electronic configuration for transition 
metal chromophores is the 
d
6
 
configuration. Much of this work has been focused on ruthenium(II) 
complexes due to their applicability towards photoinduced electron transfer based
 
applications 
arising from 
their long lived (ns
-

s)
1
 
metal to ligand ch
arge transfer (MLCT) excited state lifetimes. Some important 
applications, such as dye sensitized solar cells (DSSCs), however, rely on light harvesting, an inherently 
material intensive process due to the relatively low s
olar flux on Earth. This makes the
 
use of ruthenium, 
one of the rarest substances on the planet,
2
 
a
 
poor choice when scaling up these applications. This has 
prompted the search for alternative transition metals with charge separated excited states that are long 
lasting enough to be coupled to electron transfer processes
. While several such systems exist
, such as the 
class of 
d
10
 
copper(I) polypyridyl compounds, perhaps the easiest conceptual alternative to 
d
6
 
ruthenium(II) 
compounds are the complexes of their vastly more earth abundant 
d
6
 
iron(II) congener. A crucial 
con
sequence of switching from a 4
d
 
to
 
a 3
d
 
metal, however, is the reduced ligand field splitting (an effect 
that will be discussed in more detail in chapter 5). Whereas the lowest energy excited states of many 
ruthenium(II) polypyridyl compounds are MLCT in n
ature, the lowest energy excited s
tates of the iron(II) 
analogues are typically ligand field states. These low
-
lying metal
-
centered excited states provide for 
extremely rapid deactivation pathways from the initially generated charge transfer states.
1,3
 
Significant 
effort from many research groups has been put forth to understa
nd why these deactivation pathways
 
are so 
efficient and how to control the rate of MLCT state decay. Two such methods will be investigated in detail 
 
125
 
in chapter 4 and 5 of this dissertation. Given that much of the excited state dynamics in these compounds 
o
ccurs within ligand field excited 
states, it would be desirable to investigate the dynamics upon direct 
excitation into specific ligand field states. Unfortunately, the vastly higher oscillator strengths of the MLCT 
transitions of the ground state often sw
amp out the Laporte forbidden liga
nd field transitions. The higher 
oxidation potential of the isoelectronic cobalt(III) analogues, however, result in a significant blue shift of 
these charge transfer features and allow for the clean observation of several 
ligand field bands in the ground 
s
tate absorption spectrum. Furthermore, due to the considerably lower extinction coefficients of the ground 
state electronic transitions, any excited state absorption features in the visible region may be more 
distinguishab

d on an intense ground state bleach, as is the case for many 
iron(II) polypyridyls. Thus, cobalt(III) compounds may offer a more direct way to probe the excited state 
dynamics within the ligand field manifold for 
d
6
 
electr
onic configurations.
 
The work in 
this chapter will focus on understanding the excited state 
dynamics of a series 
cobalt(III) compounds in hopes of establishing the underpinnings of a roadmap to six
-
coordinate cobalt(III) 
photophysics. To achieve this, transient absorption experiments will
 
be performed on a series of 
cobalt(III)trisbetadiketonate 
compounds (Figure
 
3.1) which have a number of desirable properties. First, 
the synthesis of these compounds is quite simple, with a preparation for the control Co(acac)
3
 
molecule 
found in the Journ
al of Chemical Education,
4
 
and various substitution reactions 
available for adaptation 
from the work on chromium in the p
revious chapter. Additionally, the presence of high oscillator strength 

3
 
systems (similar to their chromium(III) 
Co(acac)
3
 
Co(3
-
Iac)
3
 
Co(TMHD)
3
0000000000
 
 
Figure
 
3.1.
 
Series of cobalt complexes to be studied in this chapter.
 
 
126
 
analogues) provides f
or ample signal when performing transient absorption experi
ments. Furthermore, the 
redshifted 

3
 
compounds relative to other six
-
coordinate Co(III) 
alternatives such as [Co(bpy)
3
]
3+
 
make this system both instrumentally easie
r to study due to the ease with 
which laser pulses in this spectral region can be generated and compressed on the McCusker group
 
laser 
system and allow for compensation of the increase in ligand field strength relative to [Fe(bpy)
3
]
2+
 
upon 
Figure
 
3.2.
 
(a) Tanabe
-
Sugano diagram for Co(acac)
3
 
calculated by McCusker group 
Member Jonathan Yarranton using the dtansug program written by Professor Joshua Telser 
of
 
Roosevelt University and using the Racah B and C interelectronic repulsion parameters 
from reference
 
5
. The rightmost dashed line corres
ponds to the value of Dq presented in that 
paper and the left most dashed line corresponds better to the vertical trans
ition energies 
presented therein. 
(b) Vertical transition energies acquired from 
reference 
5. 
The 
5
T
2g
 
band 
was not observed experimental
ly but 
calculated using the LIGFIELD program of reference 
6. (c) Qualitative potential energy surface diagram based on 
the Tanabe
-
Sugano diagram in 
part a. (d) One
-
electron pictures of relevant electronic states.
 



 


 



 


 



 


 



 


 



 



 







 



 
a)
 
c)
 
d)
 
1
A
1g

1
T
1g
: 16850 cm
-
1
 
1
A
1g

3
T
2g
: 12200 cm
-
1
 
1
A
1g

3
T
1g
: 9250 cm
-
1
 
1
A
1g

5
T
2g
: 8800 cm
-
1
*
 
*Calculated using LIGFIELD
 
b)
 
 
127
 
increasing the o

3
 
compounds discussed in chapter 2 may al
low for some interesting comparisons regarding the nuclear 
motion and the vibrational dephasing times. The two derivatives shown
 
in Figure
 
3.1 were chosen to 
observe the effects of both a change in vibrational frequency of the symmetric breathing mode and 
change 
in steric encumbrance, which in the chromium systems resulted in significant changes to the vibrational 
dephasing times a
nd the observed spectral evolution respectively.
 
The electronic structure of Co(acac)
3
 
can be summarized in Figure
 
3.2. The Tana
be
-
Sugano 
diagram shown in Figure
 
3.2a was calculated by McCusker group member Jonathan Yarranton using the 
dtansug program writ
ten by professor Joshua Telser of Roosevelt university and using the Racah B    
(425
 
cm
-
1
) and C (3650
 
cm
-
1
) interelectronic rep
ulsion parameters from the work of Tsiamis 
et al.
5
 
Additionally, a full 
d
6
 
analysis was performed on Co(acac)
3
 
using the LIGFIELD pro
gram
6
 
and the energies 
of the 
1
A
1g
 

1
T
1g
, 
3
T
2g
, 
3
T
1g
 
transitions from the work of Tsiamis 
et al.
5
 
to redetermine the ligand field 
strength (19243
 
cm
-
1
) as well as the Racah B (455
 
cm
-
1
) and C (3678
 
cm
-
1
) parameters. These values were 
subsequently u
s
ed to rediagonalize the energy matrices to provide the predicted 
1
A
1g
 

5
T
2g
 
vertical 
excitation energy listed in Figure
 
3.2b. 
In addition to illustrating the relative energies of each excited state 
at the Frank Condon position, the slopes of each line on
 
the Tanabe
-
Sugano diagram are also very 
informative about the nature of each excited state. Lines with approximately zero slope (no such lines are 
observed for the strong
-
field portion of the Tanabe
-
Sugano diagram shown in Figure
 
3.2a) are rather 
insensit
i
ve to the ligand field strength of the molecule and hence typically correspond to 
intraconfigurational or spin
-
flip transitions. Lines with nonzero slopes, however, do exhibit a dependence 
on the ligand field strength of the molecule and thus are typicall
y
 
interconfigurational transitions in which 
electrons are moving between the 
t
2g
 
and 
e
g
*
 
sets of orbitals on the metal center. Because of the population 
of 
e
g
*
 
antibonding orbitals, electronic states resulting from interconfigurational transitions exhibit 
g
eometric 
distortions and, by extension, displaced potential energy surfaces relative to the ground state. Lines with 
steeper slopes correspond to transitions that place more electrons in the 
e
g
*
 
set of orbitals and are hence 
more distorted. Thus, the 
5
T
2g
 
state, which has the steepest slope on the Tanabe
-
Sugano diagram has the 
 
128
 
most displaced potential energy surface. Each of these deductions is consistent with the potential energy 
surface diagram and electronic configurations (in the 1
-
electron picture) sh
o
wn in Figure 3.2c and d. We 
see that the lowest energy excited state in this system is the 
5
T
2g
 
state, similar to what is observed in many 
iron(II) polypyridyl complexes.
1,7
 
3.2
.
 
Summary of Previous 
Results
 
 
In contrast to the photophysical literature on 
six
-
coordinate ruthenium(II) or iron(II) compounds, 
the literature concerning the photophysics of six
-
coordinate cobalt(III) compounds is quite sparse, with 
only a small handful of publications showin
g transient absorption data.
4,8

10
 
There have been numerous 
investigations into the photochemistry of cobalt(III) compounds that do help paint a photophysical picture 
of these chromophores, however these studies will not be described in significant detail in t
his dissertation 
and the relevant results will only be b
rought up when needed. This short review of previous results will 
focus on a transient absorption investigation on several CoN
6
 
compounds performed by McCusker 
et al
.
8
 
and another study on Co
(acac)
3
 
publish
ed by Ferrari 
et al.
4
 
at the time of writing th
is dissertation
. 
 
We may first discuss the study of McCusker 
et al.
8
 
on a series of CoN
6
 
coordination compounds: 
[Co(en)
3
](ClO
4
)
3
, [Co(tpen)](ClO
4
)
3
, and [Co(tppn)](ClO
4
)
3
. Transient absorption experiments were 
performed by using a 314
 
nm ~500 fs excitation lase
r pulse, corres
ponding to 
1
A
1g
 

1
LMCT excitation 
for both [Co(tpen)]
3+
 
and [Co(tppn)]
3+
 
or 
1
A
1g
 

1
T
2g
 
excitation for [Co(en)
3
]
3+
. Each compound showed 
biphasic kinetics (Table 3.1). A representative dataset on [Co(tppn)]
3+
 
is shown in Figure
 
3.3. On the 
basis 
of lack of emission in t
hese complexes ([Co(CN)
6
]
3
-
, a compound of sufficiently higher ligand field strength 
than the compounds discussed here to produce a triplet metal
-
centered lowest energy excited state, exhibits 
emission from the 
3
T
1g
 
state)
11,12
 
an
d a single crystal absorption spectrum and vibrational fine structure 
 
[Co(en)
3
]
3+
 
[Co(tpen)]
3+
 
[Co(tppn)]
3+
 


, ps
 
2 ± 1
 
4 ± 2
 
3 ± 1
 


, ps
 
450 ± 100
 
44 ± 5
 
51 ± 3
 
 
Table
 
3.1.
 
Summary of the observed ki
netics in reference 8.
 
 
129
 
analysis showing a 
5
T
2g
 
lowest energy excited state for a similar ligand field strength CoN
6
 
compound 
(Figure
 
3.4),
13
 
they assigned the slower 


 
component
 
to ground state recovery from the 
5
T
2g
 
s
tate. 
Additionally, studies on quenching of 
photosubstitution reactions of a  series of [Co(NH
3
)
5
X]
n+
 
compounds 
Figure
 
3.4.
 
Potential energy surface diagram for the 
1
T
1g
, 
3
T
1g
, 
5
T
2g
 
and 
1
A
1g
 
states of 
[Co(NH
3
)
6
]
3+
 
along the 

(a
1g
) vibrational coordinate. The 
5
T
2g
 
surface is depicted as several 
dashed li
nes to represent a range of possible positions and a solid line to represent the position 
with largest overlap with the relaxed 
3
T
1g
 
state. Digitized from data in 
reference 13
.
 
Figure
 
3.3.
 
(a) Transient absorption spectrum for an aqueous solution of 
[Co(tppn)](ClO
4
)
3
 
following ~500
 
fs excitation at 314 nm. (b) Transient absorption 
data measured at 390
 
nm with a biexponential fit. Reprinted with permission 
from 
reference 8. © (1993) American
 
Chemical Society.
 
 
a)
 
b)
 
 
130
 
was also ascribed to conversion to the 
5
T
2g
 
state which was of sufficiently low energy such that ligan
d 
substitution could not be achieved.
14,15
 
This is consi
stent with the assessment by McCusker 
et al.
8
 
that the 
5
T
2g
 
state is formed in each of the three compounds shown in Table
 
3.1. Next, in trying to assign the origin 
of the fast kinetic component it was noted that the peak in the transient spectrum for [Co(en)
3
]
3+
 
(not shown 
here) corresponded almost exactly to the 
difference in vertical excitation energies between the 
1
A
1g
 

1
LMCT and 
1
A
1g
 

1
T
1g
 
transitions suggesting that the origin of the observed excited state absorption band 
may be the 
1
T
1g
 

1
LMCT transition
 
from the Frack Condon position of the 
1
T
1g
 
state. Th
e fast component 
for [Co(en)
3
]
3+
 
was then tentatively assigned to depopulation of the 
1
T
1g
 
state. It should be noted that a 
singular valued decomposition (SVD)  of the full spectrum dataset had shown tha
t the entirety of the 
transient spectrum could be des
cribed by a single basis spectrum at all time delays, thus somewhat 
weakening the above argument that the excited state absorption band corresponds to the 
1
T
1g
 

1
LMCT 
transition. Finally, the biphasic nature of the excited state deactivation in each compo
und, regardless of the 
fact that different excited states were produced initially between [Co(en)
3
]
3+
 
(
1
T
2g
) and both [Co(tpen)]
3+
 
and [Co(tppn)]
3+
 
(
1
LMCT), su
ggested a similar kinetic origin for the time constants in each compound and, 
by extension, that 
no dynamics within the LMCT manifold were observed for the latter two compounds. In 
summary, the study by McCusker 
et al.
8
 
had established that LMCT excited state li
fetimes in a few CoN
6
 
cobalt complexes are less than a picosecond and that, in each of the compou
nds studied, the 
5
T
2g
 
state was 
formed and subsequently depopulated with lifetimes of several tens of picoseconds to several hundred 
picoseconds. An additional
 
faster kinetic component was tentatively assigned to the loss of the lowest 
energy singlet excit
ed state.
 
The next work to be summarized comes from Ferrari 
et al.
4
 
concerning the first ultrafast transient 
absorption study of Co(acac)
3
 
that was published at the time of writing this dissertation. Figure
 
3.5 shows 
the 
relevant transient absorption data acquired from the study. The dynamics following both charge tr
ansfer 
and ligand field excitation showed relatively similar biphasic kinetics with the charge transfer excitation 
showing a 70 ± 40
 
fs rise time followed by a
 
1.5 ± 0.3
 
ps decay to baseline and the ligand field excitation 
showing a 70 ± 40
 
fs rise time fo
llowed by a 1.8 ± 0.3 ps decay to baseline. The absorption spectra of the 
 
131
 
lowest energy triplet and quintet excited states were calculated using time
-
dependen
t
 
density functional 
theory calculations (not shown) to help identify which excited states were contributing to the observed 
spectral features. It was found that the absorption spectrum of the lowest energy triplet excited state (
3
T
1g
) 
qualitatively matche
d
 
the dual peaks seen in the transient data, although the energetic matching could be 
better. In contrast, the spectrum of the 
5
T
2g
 
state showed only a single band centered approximately at the 
energy of the high energy band shown in Figure
 
3.5, although w
i
th an order of magnitude reduction in the 
oscillator strength relative to the same band in the triplet absorption spectrum. Based on this, the ~1.8
 
ps 
feature was assigned as depopulation of the 
3
T
1g
 
state to reform the ground 
1
A
1g
 
state. The fast rise ti
m
e was 
assigned as a convolution of several processes including vibrational cooling and intersystem crossing 
resulting in the formation of the triplet metal
-
centered excited state.
 
 
a)
 
b)
 
d)
 
c)
 
e)
 
f)
 
g)
 
h)
 
Figure
 
3.5.
 
Full spectrum transient absorption
 
data on Co(acac)
3
 
in acetonitrile solution 
following excitation at (a) 365
 
nm, (b) 390 nm, (c) 650 nm, and (d) 580
 
nm. Transient 
absorption data averaged ± 5 nm from center wavelengths of (e) 710 nm and (f) 520
 
nm 
following 390
 
nm excitation. (h) and (g) 
r
epresent analogou
s datasets following 365
 
nm 
excitation
. Reprinted with permission from reference 4. © 2019 Ferrari, Satta, Palma, Di Mario, 

 
 
132
 
3.3
.
 
Experimental Methods
 
3.3.1
.
 

3
 
Compounds
 
i. tris
(2,4
-
pentanediono)cobalt(III), Co(acac)
3
. 
This compound was made according to literature 
procedures without modification.
16
 
1
H NMR (CDCl
3
) 

 
5.54 (s, 3 H), 2.20 (s, 18
 
H).
 
ii. tris(3
-
iodo
-
2,4
-
pentanediono)cobalt(III), Co(3
-
Iac)
3
.
 
This compound was synthes
ized from Co(acac)
3
 
according to literature procedures.
17
 
1
H NMR (toluene
-
d
8
) 

 
2.41 (s, 18
 
H).
 
iii. tris(2,2,6,6
-
tetramethyl
-
3
-
5
-
heptanediono)coba
lt(III), Co(TMHD)
3
. 
This compound was 
synthesized according to literature procedures.
16
 
1
H NMR (CDCl
3
) 

 
5.64 (s, 3
 
H), 1.10 (s, 48
 
H).
 
3.3
.2
.
 
Computational Methods
 
All calculations were performed on Gaussian 16 software
18
 
on the Michigan State University High 
Performance Computing Center servers. Optimizations were perf
ormed using the unrestricted B3LYP 
functional, 6
-
311G (d,p) basis set, tight convergence criteria, and a polarizable
 
continuum model for the 
acetonitrile solvent for the Co(acac)
3
 
and Co(TMHD)
3
 
compounds. The basis set and solvent were switched 
to LANL2DZ 
and toluene respectively for Co(3
-
Iac)
3
. Optimizations and frequency calculations were 
performed on the lowest energ
y singlet (
1
A
1g
), triplet (
3
T
1g
), and quintet (
5
T
2g
) electronic states. Mass 
weighted vector displacement diagrams of each 
pertinent vibrational mode were created in the Visual 
Molecular Dynamics program.
19
 
 
3.3.3
.
 
Duschinsky Displacement Vector Calculations
 
 
A Duschinsky displacement vector is a vector that
 
describes the distortions in each of the molecular 
normal modes that are necessary to distort a molecule from one geometry
 
to another. Or, in simpler terms, 
it is a normal mode representation for a molecular distortion. This vector was calculated accordin
g to the 
equation described in the work of Ando 
et al.
20
 
and shown in equation 3.1.
 
 
 
3.1
 
Here, the Duschinsky displacement vector, 
K
, is determined fr
om 
M
, a diagonal matrix containing the 
atomic masses of each of the atoms in the molecule, 
L
r
, the matrix that diagonalizes
 
the mass weighted 
 
133
 
Cartesian Hessian of the reactant geometry, and 

R
, the Cartesian displacement vector between the reactant 
and pro
duct geometry.
 
 
An important requirement of the above calculation is that the reactant and product coordinates must 
have th
e proper orientation such that there is no linear or angular momentum between the two geometries. 
A procedure for performing this reo
rientation is described in the work of Miller 
et al.
21
 
It is relativ
ely simple 
to ensure that there is no linear momentum between the reactant and product geometries by setting the 
origin of 
each axis system at the center of mass of each. If this is not satisfied initially, then the position 
vector describing the center of
 
mass for a given geometry must be subtracted from the position vector of 
each atom in the molecule of interest. Ensuring t
hat there is no angular momentum between the reactant 
and product geometries is far less trivial. The work of Miller 
et al.
21
 
describes several equations that must 
hold for this condition to be satisfi
ed and they also describe how to actually achieve this with a 2
-
dimensional molecule. Unfortunately, adding a third dimensi
on to a molecule makes 
this methodology 
cease to work. Because no literature procedure could be found that works on a three dimensio
nal molecule 



 

 


 



 














 



 
a)
 
b)
 
Figure
 
3.6.
 
(a) A pict
orial representation of a cross product between two vectors, 
a
 
and 
b
 
separated by angle 

. The cross product, 
a
×
b
, is perpendicular to the plane that contains 
a
 
and 
b
. The vector, 
n
, is a unit vector in the same direction as 
a
×
b
. (b) A pictorial representa
tion of 
the process by which each atom is projected onto the plane perpendicular to the unit vector 
determined by
 
the sum of the mass weighted cross products of the position vectors and that 
contains the center of mass of the molecule. This is to eliminate
 
extra contributions to the 
moment of inertia to ensure an accurate angle, 

, is calculated which is needed to 
rotate the 
structure to nullify the angular momentum between two geometries.
 
 
134
 
and because all of the relevant molecules in this dissert
ation are three dimensional, a new more general 
procedure for ensuring that there is no angular momentum between two geometries was developed here and 
will be described below.
 
 
 
To ensur
e that there is no angular momentum between the reactant and product g
eometries, we 
must first perform the procedure to eliminate linear momentum outlined above. Next, we must try to achieve 
the conditions in equation
 
3.2
21
 
 
 
3.2
 
Where 



, is the angular momentum, a function of
 
m
i
, the mass of t
he 
i
th atom, and 








 
and 









 
the cartesian 
coordinates for the 
i
th
 
atom of the reactant and product geometries respectively. Typically, the right
-
hand 
side of equation 3.2 is initially nonzero, i.e. it is a nonzero vector. If we multiply 
this vector by the reciprocal 
of its norm, it becomes a unit vector pointing in the 
same direction as the angular momentum vector. This 
is shown in equation
 
3.3.
 
 
 
3.3
 
Next, it is helpful to remember that the cross product between
 
two vectors is another vector perpendicular 
to the plane that contains the two original vectors. A pictorial representation of this is shown in Figure
 
3.6a. 
If the two vectors being crossed are
 
parallel, i.e. the angle between them is zero, then the cross
 
product will 
be the zero vector and equation 3.2 will be satisfied. Thus, to eliminate the angular momentum, we need to 
rotate each 





 
by 

 
about the unit vector 



 
so that the sum of the m
ass weighted cross products is the zero 
vector. We kno
w from any introductory physics course that the angular momentum is proportional to the 
moment of inertia, 
I
, and the angular speed, 

, via 
L
=
I

. The angular speed can also be related to the ratio 
of th
e cartesian displacement and the distance from the axi
s of rotation. For small displacements, the angular 
speed can be replaced with tan(

). It follows that 

 
can be found by equation 3.4.
 
 
135
 
 
 
3.4a
 
 
 
3.4b
 
 
 
3.4c
 
 
In order to find 

, w
e must find the moment of inertia. Since the molecule is rotated about the axis 
determined by the unit vector 



, then the position of each atom along the 



 
coordinate does not contribute 
to the moment of inertia. Thus, in order to not overcount the m
o
ment of inertia in the denominator of 
equation 3.4a and b, we must first project the position vectors of each atom in the reactant and product 
geometries onto the plane perpendicular to 



 
(and by extension the residual angular momentum that we 
seek to 
eli
minate) and that contains the center of mass. This is performed in equation 3.4c and depicted 
pictorially in Figure
 
3.6b.
 
Now that we have determined the angle of rotation required to nullify the angular momentum 
between the reactant and product geometr
ies
, we need to perform the actual rotation. In principle, either the 
reactant or the product position vectors could be rotated, however, since we will later use the Hessian of 
the reactant geometry to determine the Duschinsky displacement vector, it is si
mpl
er to rotate the position 
vectors of the product geometry about the unit vector 



. This is performed in equation 3.5a with the general 
rotation matrix, T, defined in equations 3.5b
-
f.
 
 
 
3.5a
 
 
 
3.5b
 
 
136
 
 
 
3.5c
 
 
 
3.5d
 
 
 
3.5e
 
 
 
3.5f
 
Following this matrix manipulation, the reactant and product geometries will have zero relative 
linear or angular momentum a
nd th
ey will be ready to be used in the Duschinsky displacement vector 
calculation described in equation 3.1. These calculations were performed in a MATLAB script that will be 
added to the appendix of this chapter. Instructions for its use are provided in 
annot
ations throughout the 
document.
 
3.3.4
.
 
Time Resolved Spectroscopic Methods
 
 
All spectroscopic measurements were performed on the same laser system and with the same 
procedures as outlined in chapter 2 of this dissertation. Each sample will be excited 
at 600
 
nm, in resonance 
with the 
1
A
1g
 

1
T
1g
 
lowest energy spin allowed ligand fi
eld transition. All full spectrum experiments used 
a white light continuum probe generated in a sapphire window. A single TA experiment performed on 
Co(acac)
3
 
utilized nondegenerate pump and probe beams. In this 
experiment the probe beam with center 
wavele
ngth of 520
 
nm was generated in a second OPA and a small slice of this spectrum centered at 520
 
nm 
was measured for the TA experiment. The IRF for this experiment was ~110
 
fs. All other single wavelength 
measurem
ents performed in this chapter were degenera
te pump probe experiments with ~45
 
fs IRFs where 
the probe pulse was a replica of the pump pulse. Small slices of this transmitted probe spectrum were 
measured at specified wavelengths.
 
3.4
.
 
Results and Discussio
n
 
3.4.1. Co(acac)
3
 
 
The ground state absorpt
ion spectrum of Co(acac)
3
 
can be seen in Figure
 
3.7. The band centered at 
594
 
nm (127 M
-
1
cm
-
1
) can be assigned to the lowest energy spin allowed ligand field transition. More 
 
137
 
spe
cifically, this corresponds to th
e 
1
A
1g
 

1
T
1g
 
transition. The shoulder featured at 324
 
nm (~7400
 
M
-
1
cm
-
1
) has been assigned to a MLCT transition by Hadjikostas and coworkers
5
 
and Barnum
22
. Additionally, the 
lower energy hump at ~375
 
nm has been assigned to an LMCT type transition by Ferrari 
et al
.
4
 
using density 
functional theory calculations. The work in this dissertation wil
l focus solely on the excited state dynamics 
following
 
1
A
1g
 

1
T
1g
 
ligand field excitation.
 
 
Although only a single ligand field transition was highlighted in Figure
 
3.7, we already have a 
rather detailed picture of the energetic ordering of many of the 
electronic states pertinent to this system. We 
know from
 
the sharp lines in the NMR spectrum (Figure
 
3.A.1) that Co(acac)
3
 
has a diamagnetic, 
1
A
1g
 
ground state. Next, as discussed earlier, the series of absorption bands presented in the work of Hadjikostas
 
and coworkers
5
 
and summarized in Figure
 
3.2 directly provides the energies of two additional spin 
forbidden transitions to the 
3
T
2g
 
and 
3
T
1g
 
states. The similar (
t
2g
)
5
(
e
*
g
) strong
-
field el
ectronic configurations 
of each of the 
1
T
1g
, 
3
T
2g
, and 
3
T
1g
 
states suggests their equilibrium geometries are quite similar and by 
extension that the ordering of the vertical transition energies to the
se states translates to the ordering of their 
zero
-
point
 
energies as well. Finally, the vertical transition energy for the 
1
A
1g
 

5
T
2g
 
transition as calculated 
by McCusker group member Jon Yarranton using the LIGFIELD program
6
 
y
et falls slightly lower in energy 
despite the larger distortions and hence more displaced potential energy surface inherent to 
the (
t
2g
)
4
(
e
*
g
)
2
 
Figure
 
3.7.
 
Ground state absorption spectrum of Co(acac)
3
 
in aceton
itrile.
 
 
138
 
electron configuration. This establishes the 
5
T
2g
 
state as the lowest energy excited state of the system and 
an energetic
 
ordering for the pertinent electronic excited states of E(
5
T
2g
) < E(
3
T
1g
) < E(
3
T
2g
) < E(
1
T
1g
).
 
 
With an idea of the 
excited state ordering in hand, we can now turn to investigating the 
photophysical properties of Co(acac)
3
. Figure
 
3.8 shows the transient 
absorption spectrum from 460
-
610
 
nm for Co(acac)
3
 
following 600
 
nm excitation into the 
1
T
1g
 
state. The best global fi
t of this dataset 
requires two exponentially decaying kinetic components with time constants of 0.3 and 1.9
 
ps. The resulting 
DADS and EADS
 
can be seen in Figure
 
3.8b and c respectively. The corresponding experiment recently 
performed by Ferrari 
et al
.
4
 
exhibited a 70 ± 40 fs rise time and a 1.8 ± 0.3 ps decay. While
 
the 70
 
fs rise 
Figure
 
3.8.
 
(a) Transient absorption data from 460
-
610
 
nm for Co(acac)
3
 
in acetonitrile following 
600
 
nm excitation into the 
1
T
1g
 
state. (b) Decay associated difference spectrum associated with a 
global fit to two sequentially decaying exponential
 
kinetic components.
 
(c) Evolution associated 
difference spectra for the two kinetic components extracted from a global fit of the dataset.
 
a)
 
b)
 
c)
 
 
139
 
time was too close to the instrument response function to be observed here, the 1.8 ± 0.3 ps time constant 
is in excellent agreement with the 1.9
 
ps decay observed in Figure
 
3.8. Interestingly, we also observe 
an 
additional 0.3
 
ps kinetic 
component. By inspection of the DADS and EADS we see that this faster process 
results in a decrease of the magnitude of the transient signal while not significantly perturbing the shape of 
the spectrum and the slower process re
turns the spectrum back to ba
seline. Additional transient absorption 
data over a broader wavelength range can be seen in Figure
 
3.9 which shows good qualitative agreement to 
the literature data shown in Figure 3.5.
4
 
 
To get a more precise idea of the kinetics of this system, single wavelength measurements 
were 
performed. Several representative 
datasets using a degenerate pump and probe are shown in Figure
 
3.10 
 
Probe, nm
 

1
, fs
 

2
, fs
 
620
 
780
 
2200
 
620
a
 
385
a
 
1930
a
 
600
 
620
 
2200
 
600
 
540
 
2080
 
580
 
370
 
1700
 
580
 
440
 
1700
 
520
 
430
 
2100
 
Average (

)
 
510 (150)
 
2000 (200)
 
a
 
Time constants acquired from LPSVD fit.
 
Table
 
3.2.
 
Summary of kinetic comp
onents from single wavelength transient absorption 
experiments on Co(acac)
3
 
in acetonitrile following 600
 
nm excitation into the 
1
T
1g
 
state.
 
Figure
 
3.9.
 
Additional full spectrum transient absorption data on Co(acac)
3
 
in 
acetonitrile following 6
00
 
nm excitation
.
 
 
140
 
 
a)
 
b)
 
c)
 
d)
 
e)
 
f)
 
Figure
 
3.10
.
 
Representative TA traces and FFT and LPSVD analyses of coherent oscillations 
of 
Co(acac)
3
 
in acetonitrile following 600
 
nm excitation into the 
1
T
1g
 
state. Kinetics were monitored at 
(a) 620
 
nm, (c) 600
 
nm, and (e) 580
 
nm using a probe beam with a spectrum centered at 600
 
nm. 
Insets show coherent oscillations at early times. Corresp
onding FFT a
nd LPSVD vibrational spectra 
are shown in (b), (d), and (f).
 
 
141
 
and a summary of the observed kinetics is shown in Table
 
3.2. An additional non
-
degenerate pump probe 
experiment was performed in which the probe beam spectrum was centered at 520
 
nm 
(Figure
 
3.11), the 
maximum of the high energy peak in the transient sp
ectrum, and the transient signal was collected with 
sufficient pump
-
probe delay such that the signal returned to baseline. These single wavelength datasets 
again reveal biphasic kinetics
 
with 510 ± 150 fs and 2000 ± 200 fs time constants in good agreement 
with 
the global fits of the full spectrum data. Additionally, coherent oscillations can be observed at early times 
in each dataset in Figures 3.10 and 3.11. A summary of these coherences
 
as well as their dephasing times 
can be seen in Table
 
3.3. The above 
results contain observations that were absent in the recently published 
investigation into the photophysics of Co(acac)
3
 
by Ferrari 
et al.
4
 
and will be discussed in detail below. We 
will begin with a discussion of the coherent oscillations and end 
with an analysis of the population dynamics 
for this system.
 
 
When ana
lyzing the coherent oscillations observed for Co(acac)
3
, we can benefit from the 
groundwork already laid out for the related Cr(acac)
3
 
system in Chapter 2 of this dissertation. First, ba
sed 
on the magnitude of the dephasing times, we can readily conclude t
hat these coherences survive too long 
to be purely electronic coherences. The two remaining possible origins for the coherent oscillations are then 
a)
 
b)
 
Figure
 
3.11.
 
(a) Transient absorption data on Co(acac)
3
 
measured at the peak of the high energy 
band of the transient spectrum at 520
 
nm following 600
 
nm excitation into the 
1
T
1g
 
state. The 
inset shows the coherent oscillations at early time and an LPSVD fit. (b) Corresponding LPSVD 
and FFT vibrational spectru
m.
 
 
142
 
vi
brational coherences on an excited 
electronic state or vibrational coherences on the ground electronic 
state generated via impulsively stimulated Raman scattering. Because the oscillator strength for the 
Laporte
-
forbidden ground state absorption is so smal
l relative to the excited state abs
orption features 
resulting from a much more allowed charge transfer transition, we can unequivocally conclude that all 
vibrational coherences of the solute must be on an excited electronic state. A similar argument was ma
de 
for Cr(acac)
3
 
in the previous ch
apter.
 
 
Before we make assignments of vibrational coherences to specific excited states, it may be fruitful 
to predict which types of vibrational modes might be expected on each electronic state. One way to achieve 
this i
s to use the Duschinsky displacemen
t vector described in section 3.3.3. This displacement vector is a 
normal mode representation for a change in molecular geometry. In, other words, the Duschinsky 
displacement vector describes the difference between two ge
ometries in terms of displacement a
long the 
3N
-
6 vibrational modes of a molecule. Because wave packets, at least initially, move in the direction of 
steepest descent along a potential energy surface,
23
 
then by identifyi
ng which vibrational modes show the
 
highest reorganization energy for a given electronic transition, we may be able to identify which vibrational 
modes are likely to be coherently active in a particular electronic state. 
 
We can first group the electronic 
excited states into two groups: tho
se with the (
t
2g
)
5
(
e
g
*
)
1
 
electronic 
configuration (
1
T
1g
, 
3
T
2g
, and 
3
T
1g
 
states) and those with the (
t
2g
)
4
(
e
g
*
)
2
 
electronic configuration (
5
T
2g
 
state). 
We assume that electronic states within the same group will make similar nuclear distortions so that only
 
a 
Probe, nm
 
E
pu
 
(
E
pr
), 

J
 
Frequency, cm
-
1
 
(Damping Time, fs)
 
620
 
3 (0.3)
 
163 (148)
 
269 (591)
 
436 
(250)
 
463 (1489)
 
6
20
 
2.9 (0.4)
 
154 (115)
 
 
 
462 (5738)
 
600
 
3 (0.3)
 
159 (129)
 
272 (275)
 
 
473 (730)
 
600
 
2.9 (0.4)
 
177 (76)
 
 
 
 
580
 
3 (0.3)
 
153 (102)
 
259 (456)
 
444 (405)
 
465 (1945)
 
580
 
2.9 (0.4)
 
163 (248)
 
251 (475)
 
441 (346)
 
465 (1642)
 
520
 
3 (0.3)
 
162 
(205)
 
 
 
 
Average Freq
uency
 
162 ± 8
 
263 ± 10
 
440 ± 4
 
466 ± 4
 
Average Damping Time
 
150 ± 60
 
450 ± 130
 
330 ± 80
 
1450 ± 520
 
 
Table
 
3.3.
 
Summary of the coherent vibrational modes observed for Co(acac)
3
 
in acetonitrile. Pump 
and probe pulse energies, E
pu
 
and E
pr
, are provided for
 
each experiment. The IRF was 40 fs for each 
experiment except the 520nm probe dataset which had a 110 fs IRF.
 
 
143
 
single displacement vector for each
 
group need be calculated. Plots of the Duschinsky displacement vectors 
in terms of inner
-
sphere reorganization energy for the 
3
T
1g
 

1
A
1g
 
and 
5
T
2g
 

1
A
1g
 
transitions are shown 
in Figure
 
3.12. These vectors were calculated as a function of the normal modes
 
on the 
3
T
1g
 
and 
5
T
2g
 
excited 
states for Figure
 
3.12 a and b respectively so that we may better portray the vibrational modes to be
 
expected 
on excited states of (
t
2g
)
5
(
e
g
*
)
1
 
and (
t
2g
)
4
(
e
g
*
)
2
 
electronic configurations. An initial inspection of both 
reorganization energy plots in Figure
 
3.12 shows that most of the distortions characterizing the 
1
A
1g
 

3
T
1g
 
and 
1
A
1g
 

5
T
2g
 
transitions a
re low frequency modes which can be attributed to deformatio
ns of the primary 
coordination sphere. Additionally, the overall reorganization energy (i.e. the sum of the reorganization 
energy contributions along each normal mode) is larger for the 
5
T
2g
 

1
A
1
g
 
transition than it is for the 
3
T
1g
 

1
A
1g
 
transition. This
 
is consistent with the greater number of electrons occupying 

-
antibonding orbitals 
in the 
5
T
2g
 
state. Finally, we observe that there are simply fewer distorted modes for the 
5
T
2g
 

1
A
1g
 
transiti
on. This reflects the relative symmetry between the final an
d initial states for each transition. The 
Figure
 
3.12.
 
Plots of the calculated inner
-
sphere reorganization energy as a function of normal 
mode frequency for pertinent 
electronic transiti
ons of Co(acac)
3
. (a) Plot of the reorganization 
energy for the 
3
T
1g
 

1
A
1g
 
transition as a function of the frequencies of the vibrational modes of 
the 
3
T
1g
 
electronic excited state. (b) Plot of the reorganization energy for the 
5
T
2g
 

1
A
1g
 
transition 
as a function of the frequencies of the vibrational modes of the 
5
T
2g
 
el
ectronic excited state. 
Reorganization energies were calculated from the nuclear displacements of the Duschinsky 
displacement vectors for the corresponding electronic transitions.
 
3
T
1g
 

1
A
1g
 
5
T
2
g
 

1
A
1g
 
a)
 
b)
 
 
144
 
5
T
2g
 
state results from a symmetric expansion of the ground state geometry, whereas the Jahn
-
Teller active 
3
T
1g
 
state exhibits an axial compression and eq
uatorial expansion and hence additional vibrational modes 
to
 
achieve the necessary geometry.
 
We can now begin to compare the collection of experimentally observed coherent nuclear 
vibrations listed in Table
 
3.3 to the nuclear distortions depicted by the Du
schinsky displacement vectors in 
Figure
 
3.12. Both the exper
imentally observed modes at 162
 
and 440
 
cm
-
1
 
match well with the peaks around 
180 and 440 cm
-
1
 
in each plot of the Duschinsky displacement vector for the 
3
T
1g
 

1
A
1g
 
and 
5
T
2g
 

1
A
1g
 
transitions. H
owever, only the 
3
T
1g
 

1
A
1g
 
transition exhibits distortions in vibrational modes around 450 
and 250
 
cm
-
1
, which are necessary to explain the experimentally observed 263
 
and 466
 
cm
-
1
 
oscillatory 
components. We thus see that the observed coherent nuclear mo
tion most resembles the dist
ortions expected 
on an electronic state of (
t
2g
)
5
(
e
g
*
)
1
 
character, i.e. the observed modes are likely active on the 
1
T
1g
, 
3
T
2g
, or 
3
T
1g
 
excited states and not on the 
5
T
2g
 
state. This observation marks a significant departure fro
m the 
dynamics seen in the r
elated 
d
6
 
compound, [Fe(bpy)
3
]
2+
, in which coherent nuclear motion occurs on the 
5
T
2g
 
lowest energy excited state after rapidly depopulating (< 200
 
fs) the initially formed MLCT excited 
state.
24,25
 
Now that we have narrowed down the possible electronic states that the experimentally observed 
vibrational coherences could be active on, we can begin to assign them to specific vibrational modes 
calculated on the 
3
T
1g
 
excited st
ate. We may first start wit
h the 162 ± 8
 
cm
-
1
 
mode which appears most 
robustly in the TA data. This mode is in relatively good agreement with the calculated 178 cm
-
1
 
mode with 
the largest inner
-
sphere reorganization energy shown in Figure
 
3.12a. We can thu
s assign this 162
 
cm
-
1
 
expe
rimental frequency to the calculated 178
 
cm
-
1
 
symmetric breathing mode shown in Figure
 
3.13a. 
Although there is a large uncertainty in the experimental frequency of this mode due to the rapid dephasing 
time and the relatively lon
g vibrational period, the d
ecent frequency matching between experimental and 
calculated modes, the large calculated inner
-
sphere reorganization energy for the assigned mode and the 
similarity of the observed motion to that observed for the analogous Cr(aca
c)
3
 
and each of its derivat
ives 
discussed in Chapter 2 of this dissertation all support this assignment. Next, we can readily assign the 
 
145
 
experimentally observed 466
 
cm
-
1
 
mode to the 451
 
cm
-
1
 
calculated mode (Figure
 
3.13c) corresponding to a 
Co
-
O stretchin
g mode with noticeable reorganization energy depicted in Figure
 
3.12a. Not only is the 
general motion of this vibrational mode similar to what was observed for Cr(acac)
3
 
but the magnitudes of
 
the observed and calculated frequencies are almost identical as 
well (453
 
cm
-
1
 
calculated and 461
 
cm
-
1
 
experimental). The experimentally observed 440
 
cm
-
1
 
mode can next be assigned to the 444
 
cm
-
1
 
calculated 
mode (Figure
 
3.13b) which also shows noteworthy
 
reorganization energy in Figure
 
3.12a. Interestingly, 
this Co
-
O 
stretching mode is antisymmetric, showing a compression of the Cr
-
O bonds of the bottom two 
acac ligands seen in Figure
 
3.13b while the top acac ligand shows increasing M
-
L bond lengths.
 
The 
final, 263
 
±
 
10
 
cm
-
1
 
mode is more challenging to assign. There is
 
not only a relatively high 
density of states in this frequency region but also a large uncertainty in the frequency of the vibrational 
mode. Several plausible assignments can be seen in Figu
re
 
3.14 along with the calculated inner
-
sphere 
reorganization ene
rgy along each coordinate according to the Duschinsky displacement vector shown in 
Figure
 
3.12a. We see that the frequency matching between the observed 263
 
cm
-
1
 
oscillation and the 
265
 
cm
-
1
 
calculated mode in Figure
 
3.14 is quite good, however the predict
ed reorganization energy along 
this coordinate is zero. While this is not definitive proof that this assignment is incorrect, it does suggest 
that substantial migration along this coordinate 
immediately after photoexcitation is rather unlikely. In this 
lig
ht, both the 259
 
cm
-
1
 
and 254
 
cm
-
1
 
modes in Figure
 
3.14 also show minute reorganization energies. In 
Calc. 178 cm
-
1
 
Exp. 162 cm
-
1
 
Calc. 444 c
m
-
1
 
Exp. 440 cm
-
1
 
Calc. 451 cm
-
1
 
Exp. 466 cm
-
1
 
a)
 
b)
 
c)
 
Figure
 
3.13.
 
Nuclear deformations associated with the (a) 162 cm
-
1
, (b) 440 cm
-
1
, 
and (c) 466 cm
-
1
 
experimentally observed vibrational modes of Co(acac)
3
. 
 
 
146
 
contrast, the remaining 248
 
cm
-
1
 
mode shown in Figure
 
3.14 has a substantially larger reorganization energy 
and is still relatively close to the experimen
tally observed 263
 
cm
-
1
 
frequency. We thus, t
entatively assign 
the 248
 
cm
-
1
 
vibrational mode as the observed nuclear motion. 
 
With an assignment of the nuclear motion associated with each of the coherently active vibrational 
modes in Table 3.3, we can begi
n a preliminary discussion of their dephasing
 
times. With the exception of 
the 466
 
cm
-
1
 
mode with 1450
 
±
 
520
 
fs dephasing time, each normal mode has a dephasing time well under 
a picosecond. It seems that the dephasing time of the 466
 
cm
-
1
 
mode for Co(acac
)
3
 
is also on par with the 
1.1
 
ps dephasing t
ime of the analogous mode for Cr(acac)
3
 
shown in Chapter
 
2. The 162
 
cm
-
1
 
breathing 

 
= 248 cm
-
1
 

 
= 210 cm
-
1
 

 
= 254 cm
-
1
 

 
= 30 cm
-
1
 

 
= 259 cm
-
1
 

 
= 15 cm
-
1
 

 
= 265 cm
-
1
 

 
= 0 cm
-
1
 
Figure
 
3.14. 
Possible assignments for the 263
 
cm
-
1
 
experimentally observed coherent 
vibrational mode of Co(acac)
3
 
along with the corresponding inner
-
sphere reorganization 
energies for each m
ode along the 
3
T
1g
 

1
A
1g
 
reaction coordi
nate determined from the 
Duschinsky displacement vector.
 
 
147
 
mode of Co(acac)
3
 
however has a very short 150
 
±
 
60
 
fs dephasing time that is even faster than the 
analogous mode for Cr(acac)
3
.
 
While the origin of the dephasing of this mo
de was assigned to rapid IVR 
in Cr(acac)
3
 
it is difficult to discern if this is true for the cobalt system without further information. This 
point will be revisited following a comparison of the vibrational cohe

3
 
derivatives 
later 
in this chapter.
 
The above discussion has both identified the initial nuclear motion of Co(acac)
3
 
upon 
photoexcitation and narrowed down the populated excited states at very early times to those with (
t
2g
)
5
(
e
g
*
)
 
electronic configurations, thus 
providing a nice starting point to analyze the population dynamics. We know 
from Figures 3.8 and 3.9 that there are no significant changes to the initial qualitative shape of the transient 
spectrum over the course of the ex
cited state dynamics. While we ma
y expect that an intraconfigurational 
electronic transition may only lead to subtle changes in a transient absorption spectrum, we would expect 
much more dramatic changes for an interconfigurational transition. Thus, since 
no such dramatic change in 
the tr
ansient spectrum is observed and because we know that, at least during the initial few hundred 
femtoseconds, an excited state of (
t
2g
)
5
(
e
g
*
) electronic configuration is populated (
1
T
1g
, 
3
T
2g
, or 
3
T
1g
), then 
we know that the
 
observed transient spectrum must
 
not be representative of the 
5
T
2g
 
state. This is further 
supported by the qualitative agreement between the 
3
T
1g
 
absorption spectrum calculated by Ferrari 
et al.
4
 
and the observed TA data as well as the poor matching of the calculated 
5
T
2g
 
spectrum in the same paper. 
Additionally, we
 
note that the energy difference 
between the charge transfer band in the ground state 
absorption spectrum
5
 
of Co(acac)
3
 
at 30,100
 
cm
-
1
 
and the excited state absorptio
n
 
band at ~480
 
nm 
(20,830
 
cm
-
1
) in Figure
 
3.5 is 9,270
 
cm
-
1
, remarkably close to the vertical transition energy of the 
3
T
1g
 
state 
listed in Figure
 
3.2. This may suggest that the observed transient species is the 
3
T
1g
 
state, however, the 
broad maximum in Fig
ure
 
3.5 and the shift to slightly longer wavelengths in the dataset provided in 
Figure
 
3.9 make this assignment tenuous at best. 
 
Inspection of the TA data in Figures 3.8 and 3.9 show that the 2.0
 
ps kinetic compone
nt results in 
a complete return to baseli
ne and thus must represent either ground state recovery or formation of a lower 
energy dark state. We can further disprove that the 2.0
 
ps process is representative of loss of the lowest 
 
148
 
energy 
5
T
2g
 
excited state by
 
considering the timescale for this suppo
sed 
5
T
2g
 

1
A
1g
 
process. In the work 
of McCusker 
et al.
8
 
that was summarized earlier in this chapter, several six
-
coordinate cobalt(III) 
compounds were observed to have 
5
T
2g
 

1
A
1g
 
lifetimes of ~40
-
450
 
ps, much slower tha
n the observed 
2.0
 
ps time constant in th
is study. Likewise, recent variable temperature experiments on several octahedral 
d
6
 
Fe(II) compounds had shown the extracted barrierless 
5
T
2g
 

1
A
1g
 
lifetimes on the order of multi
-
hundred 
picoseconds.
26
 
The 1
-
2 orders of magnitude shorter time constant observed in this study suggest it is likely 
from a fundamentally different process. Coincidentally, an additional faste
r single digit picosecond kinetic 
compone
nt observed in the cobalt work of McCusker,
8
 
was assigned to depopulation of an excited state 
with (
t
2g
)
5
(
e
g
*
) electronic configuration (specifically the 
1
T
1g
 
state). This benchmark value for depopulation 
of the manifold 
of states with (
t
2g
)
5
(
e
g
*
) electron confi
guration is much closer to the 2.0
 
ps ground state 
recovery observed here suggesting that it too may be a result of depopulation of an excited state with 
(
t
2g
)
5
(
e
g
*
) electron configuration. Additionally, this occurs on a comparable timescale as the 1.1
 
ps 
4
T
2g
 

4
A
2g
 
internal conversion observed for Cr(acac)
3
 
in which a similar process of removing a single electron 
from the 
e
g
*
 
orbital set and placing it in the 
t
2g
 
o
rbitals occurs.
 
We now know from consideration of the time constant for return of the transi
ent signal to baseline, 
a comparison of the calculated and experimental transient spectra, and lack of significant change of the 
spectral shape after initially pers
isting on an excited state of (
t
2g
)
5
(
e
g
*
) electronic configuration that the 
2.0
 
ps kinetic c
omponent of Co(acac)
3
 
either represents ground state recovery from one of the 
1
T
1g
, 
3
T
2g
, or 
3
T
1g
 
excited states, or formation of the 
5
T
2g
 
excited state which must 
be effectively dark to avoid showing 
any transient signal. In what follows, we will assess t

5
T
2g
 
state. Certainly, the 
intensity of charge transfer transitions from the 
5
T
2g
 
excited state are likely to be reduced relative to simila
r 
transitions in excited states of shorter M
-
L bond lengths due to reduced orbital overlap. 
This phenomenon 
has been shown quite clearly in the absorption spectra of the 
d
6
 
low
-
spin [Fe(tren(py)
3
)]
2+
 
and high
-
spin 
[Fe(tren(6
-
Me
-
py)
3
)]
2+
 
compounds which show approximately an order of magnitude reduction in the 
extinction 
coefficient of the MLCT a
bsorption band.
27
 
Despite this reduction in signal magnitude, however, 
spectral changes in TA experiments in the visible reg
ion resulting from vibrational relaxation within the 
 
149
 
5
T
2g
 
state have been observed in [Fe(tren(py)
3
)]
2+
 
implying that this state is not always
 
completely dark in 
TA experiments.
27
 
Furthermore, a noticeable baseline offset in the visible region was observed in the 
sp
ectrum of the 
5
T
2g
 
state for [Co(tppn)]
3+
 
in Figure
 
3.3, again showing that the 
5
T
2g
 
state of a cobalt(III) 
compou
nd is not necessarily devoid of signal in a TA experiment.
8
 
Density functional theory calculations 
of the excited state absorption spectra performed
 
by Ferrari 
et al.
4
 
show that the spectrum of the 
3
T
1g
 
state 
q
ualitatively matches the double peaked transient spectrum that decays with a 2
 
ps time constant whereas 
the calcul
ated 
5
T
2g
 
spectrum only has a single peak in the visible spectrum centered at approximately the 
energy of the high energy band in the TA spect
rum. Although the calculated oscillator strength for the band 
in the 
5
T
2g
 
state spectrum was approximately an orde
r of magnitude lower than that for the 
3
T
1g
 
state, the 
lack of any baseline offset while probing at 520
 
nm (Figure
 
3.11), the maximum of the c
alculated 
5
T
2g
 
band, 
strongly suggests that the 
5
T
2g
 
state is not being populated by the 2
 
ps process, and by exte
nsion that the 
2.0
 
ps process is representative of ground state recovery from one of the 
1
T
1g
, 
3
T
2g
, or 
3
T
1g
 
states.
 
 
We next turn to the 
510
 
fs kinetic component listed in Table
 
3.2 to characterize the excited state 
dynamics of Co(acac)
3
. This kinetic com
ponent was not observed by Ferrari 
et al.
4
 
however, they do no
te 
a change in the relative peak heights during the first picosecond which could be due to a kinetic process 
which they did not fit for. I
n the data presented here, it also seems as though no significant changes to the 
spectral shape are occurring during t
his 510
 
fs process, but rather a decrease in the magnitude of the signal. 
This spectral change is likely either attributed to nuclear moti
on (vibrational cooling or a Jahn
-
Teller 
distortion) or an electronic state change. We will address the likelihood of 
each of these possibilities.
 
 
We may first deliberate the merits of attributing the 510
 
fs kinetic component to nuclear motion, 
starting w
ith the prospect of vibrational cooling and ending with the notion of a Jahn
-
Teller distortion. 
Considering that vibra
tional cooling on the structurally similar Cr(acac)
3
 
occurs on the 7
-
10
 
ps timescale,
28
 
a 510
 
fs time constant for vibrational cooling in Co(ac
a
c)
3
 
seems rather fast. Although the electronic state 
on which this process occurs is quite different for each compound (in Cr(acac)
3
 
all 
d
 
electrons reside within 
t
2g
 
orbitals, whereas in the excited state of Co(acac)
3
 
an electron is populating the antibo
n
ding 
e
g
*
 
orbital), 
it seems unlikely that vibrational cooling would occur with such disparate rates between the two structurally 
 
150
 
similar compounds. A Jahn
-
Teller distortion, however, may occur as fast as the frequency of the normal 
modes driving it allow.
 
According to density functional theory calculations, the 
3
T
1g
 
excited state (and 
presumably the 
1
T
1g
 
and 
3
T
2g
 
excited states with similar electronic configurations) shows an equatorial 
elongation relative to the ground state with average bond lengths of d
a
xial
 
= 1.89 Å and d
eq
 
= 1.99
 
Å. This is 
in contrast to the uniform bond lengths calculated for the 
1
A
1g
 
ground state (1.90
 
Å) or 
5
T
2g
 
excited state 
(1.99
 
Å). One might expect that the reduced net overlap upon settling into the equilibrium geometry of the 
3
T
1g
 
(or 
1
T
1g
 
or 
3
T
2g
) state should result in diminished oscillator strength for a CT transition, consistent with 
the observations in the TA spectra. Additionally, inspection of the dephasing times shown in Table 3.3, 
show that several of the dephasing tim
e
s are qualitatively similar to this 510
 
fs time constant. It seems 
plausible that these modes are dephased as the molecule settles into its equilibrium geometry similar to how 
the flattening distortion of Cu(I) polypyridyls is dephased as the flattening i
s
 
completed.
29
 
A closer look at 
the Duschinsky displacement vector diagram in Figure
 
3.12a shows that the primary vibrational mode 
responsible for this distortion is the 162
 
cm
-
1
 
symmetric breathing mode depicted in Figure
 
3.13a. In theory, 
we may be able to gain i
n
sight into the relevance of nuclear motion to the 510
 
fs spectral changes by 
comparing the magnitude of the coherent oscillations to the magnitude of the preexponential factor for the 
510
 
fs kinetic component. As, the excited state wavepacket travels from
 
one turning point to the other, the 

the oscillations in the TA data. If this oscillation amplitude is significantly less than the preexponential 
facto
r
 
for the 510
 
fs kinetic component, then it is unlikely that it results from nuclear motion since the 
extrema of the nuclear motion are not enough to produce the necessary 


A. A summary of this comparison 
is shown in Table
 
3.4. Unfortunately, it is rather
 
inconclusive. When comparing oscillation and exponential 
decay amplitudes acquired directly from the LPSVD fit of each dataset, we see that the 162
 
cm
-
1
 
oscillation 
can describe 100
 
% of the exponential decay, albeit with a large standard deviation. In co
n
trast, the same 
values acquired through direct exponential fitting of each dataset show that only ~50
 
% of the exponential 
decay could be described from nuclear motion. Thus, at this point it is rather inconclusive if the 510 fs 
kinetic component of Co(ac
a
c)
3
 
is due solely to distortions of the molecule.
 
 
151
 
 
An alternative assignment for the 510
 
fs kinetic component of Co(acac)
3
 
besides distortio
n of the 
molecule would be an electronic state change. It was noted that the spectrum after the first picosecond 
qu
alitatively resembled a calculated spectrum of the 
3
T
1g
 
state in the work of Ferrari 
et al
.
4
 
and that the 
energy differences between the high energy band maximum in the TA spectrum and the charge transfe
r 
band in the ground state absorption spectrum
5
 
is fortuitously close to the vertical transition energy of the 
3
T
1g
 
state, both sugge
sting that the 
3
T
1g
 
state might be the sampled in the transient absorption experiment 
and necessitating an electron
ic state change from the initially populated 
1
T
1g
 
state. Additionally, since each 
of the plausibly accessible excited states (
1
T
1g
, 
3
T
2g
, and
 
3
T
1g
) have a similar electronic configuration, we 
may expect only subtle changes to the absorption spectrum if tra
nsitions between these states were to occur, 
fitting with the subtle changes arising from the 510
 
fs kinetic component. If a single change in
 
electronic 
state is observed then it must either be 
1
T
1g
 

3
T
2g
 
intersystem crossing from the initially populated 
1
T
1g
 
state (followed by a fast IC to form the 
3
T
1g
 
state) or 
3
T
2g
 

3
T
1g
 
internal conversion (implying a prompt, 
unobservable ISC). It stands to reason that the slower process is the most observable
. We can thus evaluate 
both options on the basis of expected rates.
 
 
We can deduce from the similarity in the 1
-
electron picture e
lectronic configurations for each of 
the 
1
T
1g
, 
3
T
2g
, and 
3
T
1g
 
states, as well as the similar slopes on the Tanabe Sugano diagr
am in Figure
 
3.2 that 
each of these electronic states should show fairly similar distortions from the ground state equilibrium 
Probe Wavelength, nm
 
LP
S
VD: A

/A
exp
 
Residuals: A

/A
exp
 
620
 
97 %
 
59 %
 
620
 
165 %
 
*
 
600
 
55 %
 
21%
 
600
 
280 %
 
44%
 
580
 
11%
 
117%
 
580
 
68%
 
54%
 
520
 
27%
 
26%
 
Aver
age (

)
 
100 (90) %
 
50 (30) %
 
* Could not 
acquire value for this dataset
 
Table
 
3.4.
 
Comparison of the oscillation amplitude and the preexponential factor for 
the 510
 
fs kinetic component of Co(acac)
3
 
observed during TA experiments. Oscillation 
amplitudes 
and preexponential factors were either bo
th directly acquired from LPSVD 
fits or separately acquired from an exponential fit to the data and the oscillation 
magnitude of the residual of the exponential fit.
 
 
152
 
geo
metry and thus their potential energy surfaces should be fairly nested relative to each other. With regards 
to the energy gap 
law as applied to weakly coupled systems
30,
31
 
(i.e. nested potentials) a smaller energy gap 
between a pair of nested states should result in a larger rate constant for nonr
adiative decay. Since the 
potential energy surfaces of interest here are all nested and likely exhibit similar force constants
, we may 
approximate the energy gaps based on the vertical transition energies listed in Figure
 
3.2. We see that the 
energy gap be
tween the 
1
T
1g
 
and 
3
T
2g
 
states is larger than that for the 
3
T
2g
 
and 
3
T
1g
 
states alluding to slower 
ISC. Furthermore, we can in
fer from spin selection rules that the coupling between the 
1
T
1g
 
and 
3
T
2g
 
excited 
states is substantially weaker than that between
 
the two triplet excited states. All of this suggests that 
1
T
1g
 

3
T
2g
 
ISC is slower than IC within the triplet manifold. Thus
, we can say that the likely slowest observable 
transition is the 
1
T
1g
 

3
T
2g
 
transition. If the 510
 
fs kinetic component of Co(ac
ac)
3
 
is indeed due to an 
electronic state change then it is most likely ISC from the initially populated singlet excited state
.
 
 
We have at this point described several plausible scenarios to describe the observed kinetics of the 
Co(acac)
3
 
system following 
1
A
1g
 

1
T
1g
 
excitation. It seems evident that the 
5
T
2g
 
lowest energy excited 
state remains unpopulated throughout the course 
of the excited state dynamics, in stark contrast to many 
iron(II) polypyridyl systems.
1
 
We further know that the initial nuclear motion 
following 
1
A
1g
 

1
T
1g
 
excitation resembles the distortions expected on an excited state of (
t
2g
)
5
(
e
g
*
) character and that ground state 
recovery occurs from one of these states (
1
T
1g
, 
3
T
2g
, 
3
T
1g
) with a 2.0
 
ps time constant. An additional 510
 
fs 
process is 
assigned eithe
r to settling into the equilibrium geometry of this manifold of states or to 
intersystem crossing from the initially populated 
1
T
1g
 
state. Unfortunately, there is not enough evidence at 
present to definitively prove either of these scenarios 
so additional 
experiments are needed to help further 
refine the kinetic model of Co(acac)
3
. Ideally, ultrafast X
-
ray fluorescence spectroscopy could be used to 
measure the transient K

 
emission spectrum, which is highly sensitive to the 3
d
 
spin moment and 
hence 
could be
 
used to definitively identify the presence of a singlet or triplet excited state. Similar experiments 
were recently performed on [Fe(bpy)
3
]
2+
 
and [Fe(dccp)
2
]
2+
 
to quantify the influence of the 
3
T
1g
 
state on the 
excited state dynamics.
32,33
 
Such an experiment, however, is 
difficult to get access to. We will thus instead 
 
153
 
attempt t
o glean more information about this chemical system by studying the small series of derivatives 
displayed in Figure
 
3.1.
 
3.4.2. Co(3
-
Iac)
3
 
 
The ground state absorption spectrum of Co(3
-
Iac)
3
 
can be seen in Figure
 
3.15. The band centered 
at 595
 
nm can be a
ssigned to the 
1
A
1g
 

1
T
1g
 
lowest energy spin
-
allowed ligand field transition, similar to 
what was observed for the unsubstituted Co(acac)
3
 
system. The identical peak position between these t
wo 
compounds suggests that the effect of the iodine substitution on the ligand field
 
strength is quite small, in 

3
 
systems. The remainder of this section will discuss 
the dynamics following 600
 
nm 
1
A
1g
 

 
1
T
1g
 
excitation.
 
 
On an experimental note, the Co(3
-
Iac)
3
 
system was not as photost
able as the unsubstituted 
Co(acac)
3
 
system and a significant change in the absorption spectrum could be seen if a solution of Co(3
-
Iac)
3
 
was exposed to light for too long. A representative difference spectrum highlighting the sample 
degradation can be seen
 
in Figure
 
3.16, in which a significant increase in absorbance around 500
 
nm occurs 
resulting in an increasingly brownish color to the s
olution. This is likely attributable to the release of iodine 
from the ligand backbone. This is not expected to have sig
nificantly affected the results shown in this 
dissertation since the observed frequencies of the coherent oscillations discussed shortly
 
are exceptionally 
Figure
 
3.16.
 
Difference spectrum of 
Co(3
-
Iac)
3
 
i
n toluene before and after
 
significant light exposure.
 
Figure
 
3.15.
 
Ground state absorption 
spectrum of Co(3
-
Iac)
3
 
in toluene.
 
 
154
 
close to the calculated frequencies for the u
ndegraded compound and are highly dependent on the pres
ence 
of iodine on the ligand backbone.
 
 
Full spectrum transient absorption data on Co(3
-
Iac)
3
 
in toluene following 600
 
nm excitation into 
the 
1
T
1g
 
state is shown in Figure
 
3.17. The data is qualitatively similar to the data for the unsubstituted 
Co(acac)
3
 
and show that most of the transient signal is lost in the first six picoseconds. Single wavel
ength 
TA measurements were used to acquire a more precise view of the kinetics of this system. A summary of 
the kinetics can be seen in Table
 
3.5 with representat
ive data sets shown in Figure
 
3.18. We again see 
biphasic kinetics in this system with 180 ± 1
50
 
fs and 2070 ± 160
 
fs time constants. Additionally, a 94 ± 
1
 
cm
-
1
 
coherent oscillation with 320 ± 100 fs dephasing time attributable to the excited state of the
 
solute 
was observed in several datasets as summarized in Table 3.6. 
 
Probe, nm
 

1, 
fs
 

2
, fs
 
620
 
110
 
2000
 
620
 
90
 
1900
 
600
 
400
 
2280
 
580
 
110
 
2100
 
Average (

)
 
180 (150)
 
2070 (160)
 
 
Table
 
3.5.
 
Summary o
f kinetic 
components from single wavelength transient abs
orption 
experiments on Co(3
-
Iac)
3
 
in toluene following 600
 
nm excitation into the 
1
T
1g
 
state. 
 
Figure
 
3.17.
 
Full spectrum TA data on Co(3
-
Iac)
3
 
in toluene following 600
 
nm excitation into the 
1
T
1g
 
state. The region from ~5
70
-
640
 
nm was omitted due to contamination from pump scatter.
 
 
155
 
 
 
Figure
 
3.18.
 
Represen
tative single wavelength TA traces and FFT and LPSVD analyses of coherent 
oscillations of Co(3
-
Iac)
3
 
in toluene following 600
 
nm excitation
 
into the 
1
T
1g
 
state. Kinetics were 
monitored at (a) 620
 
nm, (c) 600
 
nm, and (e) 580
 
nm using a probe beam with a spe
ctrum centered at 
600
 
nm. Insets show coherent oscillations at early times. Corresponding FFT and LPSVD vibrational 
spectra are shown in (b
), (d), and (f). LPSVD contributions are omitted from the 580
 
nm probe data 
due to lack of solute oscillations.
 
a)
 
b
)
 
d
)
 
c
)
 
f
)
 
e
)
 
 
156
 
 
The above results in Fi
gure
 
3.17 and Table
 
3.5 show that both the spectral evolution and kinetics 
for ground state recovery of Co(3
-
Iac)
3
 
are essentially identical to the unsubstitu
ted Co(acac)
3
 
control 
molecule. This further suggests a similar decay cascade between the two comp
ounds. While the faster 
kinetic component seen here is significantly shorter than what was observed in Co(acac)
3
, it is possible that 
the large amplitude, low
-
frequency oscillatory component occurring concomitant with the exponential 
decay distorted the me
asured time constant. The insensitivity of the excited state dynamics to the iodo
-
substituent in the 3
-
position of the ligand backbone is also consistent with
 
the lack of effect observed in the 
Cr(acac)
3
 
systems discussed in chapter
 
2 of this dissertation.
 
 
As summarized in Table
 
3.6, only a single coherent vibrational mode of Co(3
-
Iac)
3
 
could be 
observed. Given the frequency region and the types of motion alr
eady observed in Co(acac)
3
 
as well as each 

3
 
compounds discussed in chapter
 
2, it is clear that the observed 94 ± 1
 
cm
-
1
 
vibration of 
Co(3
-
Iac)
3
 
corresponds to a symm
etric breathing mode. Figure
 
3.19 shows this symmetric breathing mode 
Probe, nm
 
Frequency, cm
-
1
 
(Damping Time, fs)
 
620
 
95 (216)
 
620
 
93 (320)
 
600
 
93 (411)
 
Avg. Frequency
 
94 ± 1
 
Avg. Dam
ping
 
320 ± 100
 
 
Table.
 
3.6.
 
Summary of the coherent vibrational modes observed for Co(3
-
Iac)
3
 
in 
toluene following 600
 
nm excitation into the 
1
T
1g
 
state.
 
All measurements had a 40 fs 
IRF and employed 2.9 and 0.4 

J pump and probe pulse energies respectiv
ely.
 
Figure
 
3.19.
 
DFT calculated symmetric breathing mode of Co(3
-
Iac)
3
 
on (a) the 
1
A
1g
 
ground state, (b) the 
3
T
1g
 
lowest energy triplet excited state, and (c) the 
5
T
2g
 
lowest energy 
excited state of the system.
 
1
A
1g
 
102 cm
-
1
 
3
T
1g
 
95 cm
-
1
 
5
T
2g
 
99 cm
-
1
 
a)
 
b
)
 
c
)
 
 
157
 
as c
alculated on the 
1
A
1g
 
ground state, the 
3
T
1g
 
lowest energy triplet excited state, and the 
5
T
2g
 
lowest energy 
excited state of the system. We already know from consideration of the lo
w extinction coefficient for the 
initial excitation that a vibrational coh
erence on the ground electronic state is not observed in this system. 
Furthermore, from consideration of the discussion of the unsubstituted Co(acac)
3
 
compound, it is unlikely 
that t
he 
5
T
2g
 
lowest energy excited state of this system is accessed. These find
ings are confirmed by the 
excellent agreement between the 94
 
±
 
1
 
cm
-
1
 
observed and 95
 
cm
-
1
 
calculated frequency of the 
3
T
1g
 
excited 
state in Figure
 

t this mode is observed on the 
3
T
1g
 
state specifically, it 
does indicate t
hat it is observed on an excited state of (
t
2g
)
5
(
e
g
*
) electronic configuration (
1
T
1g
, 
3
T
2g
, 
1
T
1g
), 
consistent with the dynamics in Co(acac)
3
.
 
 
 
What is perhaps more 
interesting here is the dephasing time of this mode. In the chromium systems, 
iodo
-
substitu
tion led to a dramatic order of magnitude increase in the vibrational dephasing time relative to 
Cr(acac)
3
, which was attributed to a decrease in efficiency of IVR. 
In contrast, iodo
-
substitution seems to 
have very little effect on the dephasing time in th
ese cobalt systems. This could imply a fundamentally 
different dephasing mechanism between these two sets of compounds. A more complete discussion of this 
will be ma
de after the results of the final Co(TMHD)
3
 
compound are revealed. 
 
3.4.3. Co(TMHD)
3
 
Figure
 
3.20.
 
Ground state absorption spectrum of Co(TMHD)
3
 
in dichloromethane solution.
 
 
158
 
 
The
 
visible region of the ground state absorption spectrum of Co(TMHD)
3
 
can be seen
 
in 
Figure
 
3.20. A peak centered at 594
 
nm can be assigned to the 
1
A
1g
 

1
T
1g
 
lowest energy spin allowed 

3
 
compounds. The spectral 
shape and position are unperturbed relativ
e to the Co(acac)
3
 
control molecule suggesting that this 
substitution 
has little effect on the ligand field strength or the initial geometry of the molecule. The 
following transient absorption experiments on Co(TMHD)
3
 
will employ an excitation pulse that i
s in 
resonance with this 
1
A
1g
 

1
T
1g
 
transition.
 
 
Full spectrum transi
ent absorption data on Co(TMHD)
3
 
is shown in Figure
 
3.21 along with the 
DADS associated with a global fit of the data to a single exponentially decaying kinetic component. The 
timescale 
and shape of the spectral evolution is qualitatively similar to that f
or Co(acac)
3
 
and Co(3
-
Iac)
3
 
suggesting a similar decay cascade. Single wavelength TA measurements were conducted to get a more 
precise idea of the kinetics of this system. These data are
 
shown in Figure
 
3.22 along with the corresponding 
vibrational spectra
 
from FFT and LPSVD analyses. A summary of the kinetics and coherent oscillations is 
shown in Table 3.7 and Table 3.8 respectively. We again see that the dynamics in this system are biph
asic 
with time constants that are essentially identical to the Co(acac
)
3
 
control molecule. An analysis of the 
coherent vibrational motion reveals a 163
 
cm
-
1
 
frequency with large error bars due to the rapid dephasing 
 
Figure
 
3.21.
 
(a) Full spectrum TA data on Co(TMHD)
3
 
in DCM following 600
 
nm 
1
A
1g
 

1
T
1g
 
excitation. Anomalies around 600
 
nm are due to excess pump scatter. (b) DADS associated with a 
global fit of the data to a single exponential kinetic component.
 
a)
 
b
)
 
 
159
 
 
 
a)
 
b)
 
c)
 
d)
 
Figure
 
3.22.
 
(a) and (c) Single wavelength TA data on Co(TMHD)
3
 
in DCM 
following 600
 
nm 
1
A
1g
 

1
T
1g
 
excitation. (b) and (d) corresponding vibrational spectra 
from FFT and LPSVD analysis.
 
Probe, nm
 

1
, fs
 

2
, fs
 
600
 
530
 
2000
 
600
 
500
 
1900
 
Avg (

)
 
520 (20)
 
1950 (70
)
 
 
Table
 
3.7.
 
Summary of kinetic 
components of Co(TMHD)
3
 
in DCM 
following 600 nm excitation.
 
IRF, fs
 
E
pu
 
(
E
pr
), 

J
 
Frequency, cm
-
1
 
 
(Damping Time, fs)
 
45
 
3.6 (0.4)
 
 
146 (120)
 
55
 
2 (0.3)
 
75 (118)
 
180 
(148)
 
Avg. Frequency
 
-
 
163 ± 24
 
Avg. Damping
 
-
 
130
 
± 20
 
 
Table
 
3.8.
 
Summary of coherent oscillations of 
Co(TMHD)
3
 
monitored at 600 nm in DCM following 
600
 
nm excitation.
 
 
160
 
times and difficulty in fitting. This is likely the symmetric breathing mode calculated at 174
 
cm
-
1
 
on t
he 
3
T
1g
 
excited state and depicted in Figure
 
3.23. A second 75
 
cm
-
1
 
oscillatory component was acquired in
 
a 
single data set but is likely due to poor exponential fitting rather than coherent vibrational motion. 
 

3
 
Dyna
mics
 
 
The excited state dynamics of a series of cobalt(III) trisbetadiketonate complexes have been 
investigated in this chapter. A summary of the kinetics and coherent oscillations is shown in Table
 
3.9. We 
see that each of the compounds studied exhibit id
entica
l spectral evolution in the visible region following 
1
A
1g
 

1
T
1g
 
excitation. Furthermore, the measured kinetics for Co(acac)
3
, Co(3
-
Iac)
3
, and Co(TMHD)
3
 
are 
all within error, except for the fast component of Co(3
-
Iac)
3
 
which may have been influenced 
by the
 
large 
amplitude oscillatory component. Because of these findings, we can conclude that the iodo and 
t
-
butyl 
substitutions have little effect on the excited state dynamics (a finding that is consistent with the data on the 
related chromium compounds 
of the
 
previous chapter) and that each compound has an identical decay 
Figure
 
3.23.
 
Nuclear distortion associated with the coherent vibrational mod
e of Co(TMHD)
3
.
 
calc:
 
174
 
cm
-
1
 
exp: 163
 
cm
-
1
 
Compound
 

1
, fs
 

2
, fs
 

breath,obs
, c
m
-
1
 

breath,calc
, cm
-
1
 

dephase
, fs
 
Co(acac)
3
 
510 (150)
 
2000 (200)
 
162 (8)
 
178
 
150 (60)
 
Co(3
-
Iac)
3
 
180 
(150)
 
2070 (160)
 
94 (1)
 
95
 
320 (100)
 
Co(TMHD)
3
 
520 (20)
 
1950 (70)
 
163 (23)
 
174
 
130 (20)
 
 
Table
 
3.9.
 
Summary of the kinetic components and coherent o
scillations of each of the 

3
 
compounds investigated in this chapter.
 
 
161
 
cascade. The discussions in sections 3.4.1 and 3.4.2 have shown that the 
5
T
2g
 
lowest energy excited state 
remains unpopulated during the course of the excited state dynamics, in stark c
ontras
t to many 
d
6
 
six
-
coordinate iron(II) polypyridyl systems.
1
 
Furthermore, following 
1
A
1g
 

1
T
1g
 
excitation, the system 
remains in excited states of (
t
2g
)
5
(
e
g
*
) electronic configuration (
1
T
1g
, 
3
T
2g
, 
3
T
1g
) until the ground state is 
recovered with a ~2
 
ps time constant. The dynamics 
prior to ground state recovery are more ambiguous. 
The fast tim
e constant was proposed to represent either a change in electronic state (
1
T
1g
 

3
T
1,2g
 
ISC) or a 
change in geometry.
 
 
An additional important observation is that the dephasing time of the 
symmetric breathing mode of 
each cobalt compound was quite rapid. I
n contrast, both synthetic modifications to the acac ligand shown 
here resulted in significant, up to an order of magnitude, changes in the dephasing times of the symmetric 
breathing mode i

3
 
compounds discussed in chapte
r
 
2. Whereas the 
changes to the dephasing times in the chromium systems were ascribed to changes in IVR rates resulting 
from changes to the vibrational density of states, the invariance of 
the dephasing times in the cobalt systems 
suggests a fundamentally 
different dephasing mechanism. While it is difficult to conclusively identify the 
exact origin for dephasing in these compounds, we may be able to rule out a notion that was suggested 
origi

3
 
systems
34
 
(and later disproven for that particular system by the results of 
chapter 2 of this dissertati
on), i.e. that the rapid dephasing is indicative of such modes driv
ing an electronic 
state change. To achieve this, we can perform wave packet simulations to model how state changes affect 
the dispersion or dephasing of the wavepacket. We will employ the s
plit operator method, first developed 
by Feit, Fleck, and Steiger.
35
 
For the purposes of this dissertation a homebuilt MATLAB script for this 
procedure was written based on the work of Tanner,
36
 
and Reber 
et al.
37
 
A   detailed description of the 
mechanics of this meth
odology is beyond the scope of this dissertation, and the intereste
d reader is directed 
to the references just listed. However, in brief, the split operator method uses propagators to advance the 
wave function (or wave packet) in time. The principle idea o
f the split operator method is to split the 
Hamiltonian operator in
to propagators which are diagonal in the basis of position or momentum. By 
performing half
-
propagation steps in momentum and then position space and then another set of half
-
steps 
 
162
 
in the re
verse order the wavepacket is advanced in time. The Fourier transfo
rm relationship between 
position and momentum space makes it easy to convert the wavepacket into the proper representation. In 
what follows, we will show how the relative displacement of ex
cited state potential energy surfaces affects 
dephasing times. 
 
Fi
gure
 
3.24a shows a two
-
level system in which the potential energy surfaces are significantly 
displaced relative to each other. An initial gaussian wavepacket was placed at 0
 
Å on the upper surface and 
allowed to propagate as shown in Figure
 
3.24b. The wave
packet immediately splits to the upper and lower 
a)
 
b)
 
d)
 
c)
 
Figure
 
3.24.
 
(a) Two
-
level potential energy surface diagram with displaced minima for use in 
wavepacket calculation. Diabatic and adiabatic representations are s
hown as dashed and solid 
lines respectively. The initial gaussian wavepacket is placed at 0 Å on the upper 
surface and 
allowed to propagate. (b) Contour plot showing the sum of the amplitude of the corresponding 
wavepackets on both excited states in positi
on space as a function of time. (c) Two
-
level 
potential energy surface diagram with nested potentials for u
se in wavepacket simulations. The 
initial gaussian wavepacket is placed at 0 Å on the upper surface. (d) Contour plot showing the 
sum of the amplitud
e of the corresponding wavepackets on both excited states. Despite 
continuously cycling population from one
 
surface to the other, the wavepacket retains a well
-
defined phase relationship.
 
 
163
 
surfaces with the bulk of the amplitude on the lower potential. As each wavelet reaches the initial crossing 
point more ampl
itude is transferred to the other surface thus distributing the phase of the overal
l wavepacket 
resulting in significant dephasing in only a few hundred femtoseconds. In contrast, nesting of the potential 
surfaces as seen in Figure
 
3.24c results in a drama
tically different dephasing rate. Here, despite 
continuously cycling from the upper
 
and lower states, the identical forces on each potential surface keep 
each wavelet propagating identically thus retaining their phase relationship. We thus see that when po
tential 
energy surfaces of identical curvature are nested along a particular nuclea
r coordinate, electronic state 
changes do not dephase a vibrational coherence. Thus, since each of the 
1
T
1g
, 
3
T
2g
, and 
3
T
1g
 
excited states 

3
 
are nested, then the
 
rapid vibrational dephasing times must not be due to electronic state 
changes.
 
 
Th
e results of this chapter have shed new insight into the ultrafast excited state dynamics of six
-
coordinate cobalt(III) compounds, however additional experiments are needed 
to conclusively characterize 
these ultrafast processes. Ideally measurement techniq
ues that are sensitive to the spin
-
state of the metal 
center such as ultrafast X
-
ray emission spectroscopy could be used to clear up some of the ambiguities of 
spin
-
state as
signment to build a more complete picture of the decay cascade. Furthermore, with a
 
better 
understanding of why the lowest energy 
5
T
2g
 
excited state of this system is not populated on ultrafast 
timescales, we may be able to engineer six
-
coordinate Fe(II) c
ompounds that do not rapidly depopulate 
their MLCT states in favor of this lower en
ergy ligand field state. Additional future efforts on cobalt(III) 
compounds may focus on extending the LMCT state lifetimes which have been shown to be exceedingly 
fast in b
oth Co(acac)
3
 
by Ferrari 
et al.
4
 
and Several CoN
6
 
compounds by McCusker 
et al.
8
 
Long lasting 
LMCT excited states of earth abundant cobalt compounds cou
ld then potentially have wide
-
ranging 
applications for use in photo
-
induced electron transfer chemistry.
 
 
 
 
164
 
 
 
 
 
 
 
 
 
 
 
 
 
APPENDICES
 
 
 
 
 
 
 
 
 
 
 
 
 
 
165
 
Appendix 3.A. NMR Spectra
 
 
 
 
 
 
 
 
 
 
Figure
 
3.A.1.
 
NMR spectrum of Co(acac)
3
 
in CDCl
3
.
 
 
 
166
 
 
 
 
 
 
 
 
Figure
 
3.A.2.
 
NMR spectrum of Co(3
-
Iac)
3
 
in deuterated toluene.
 
 
Figure
 
3.A.3.
 
NMR spectrum of Co(TMHD)
3
 
in CDCl
3
.
 
 
167
 
Appendix 3.B. Duschinsky Displacement Vector Calculation MATLAB Script
 
The following contains the code for the MATLAB script that performed the Duschinsky 
displacement vector calculations used for this 
dissertation work. It should be noted that, due to formatting 
restrictions, the code below was pasted directly from a workin
g MATLAB script. Because of this, there 
may be several instances in which the written code runs beyond the margins used here and hen
ce wraps 
around to the following line. As such, this code may not work properly until these disjointed portions of the 
code 
are brought back together on the same line.
 
 
clear 
all
 
clc
 
 
 
%This is a script that uploads the coordinates from excel files 
that were
 
%created from pbd files and reorients them both so that they have no
 
%relative linear or angular momentum. The procedure 
of [Miller, w. H.; Ruf,
 
%B. A.; Chang, Y.
-
T. J chem phys. 89 (10), 1988] was followed for this.
 
%Then once there is no Linear
 
or Angular momentum between these two
 
%geometries, it uses the hessian calculated for one geometry to calculate
 
%the Duschinsky di
splacement vector that describes the displacement of each
 
%vibrational mode necessary to distort the reactant state to the pr
oduct
 
%state geometry. The procedure of [Ando, S.; Iuchi, S.; Sato, H. Chem.
 
%Phys. Let. 2012, 535, 177
-
181.] was used for this. Fi
nally, The reduced
 
%masses for each vibrational mode as well as the force constant was
 
%calculated so that the innersphere re
orgonization energy for each normal
 
%mode could be obtained using lamda_i_tot=sum(0.5*f_j*(K_j)^2). 
 
 
 
 
 
%*************************
********************************************
 
%___________________________DIRECTIONS________________________________
 
%This scr
ipt will ask for three files before doing any calculations. It
 
%will ask for the coordinates of the reactant state followed by the
 
%coordinates of the product state followed by the hessian for the reactant
 
%state. These files must be in the correct format 
for this script to work.
 
%To make the coordinate files, first create a pdb file that contains the
 
%geometry of the molecule and the
n import the data into excell so that not
 
%all of the data is in the first column. This will initially create 8
 
%columns of values. Delete the first, third and last (eigth) columns by
 
%right clicking the top of the column in excel and selecting delete. Now
,
 
%delete the top row by selecting row 1 on the side of the excel doc and
 
%selecting delete. Next, scroll down t
o the end of the coordinates and
 
%delete all of the values that describe the connectivity of atoms. We now
 
%have a file that has columns with at
om index, atom type, x coordinate, y
 
%coordinate, and z coordinate. We also need to add the masses of each of
 
%t
hese atoms in amu. Add this in by hand after the z coordinate column.
 
%Save it as an excell file and its ready for use. To get the Hessian, we
 
%
first need to create a formatted checkpoint file for the frequency
 
 
168
 
%calculation of the reactant state. to do thi
s, log onto the terminal in
 
%HPCC and go to the directory that contains the checkpoint file (.chk) for
 
%the relevant frequency calculation. next
, type "formchk 
-
3 filename.chk
 
%newfilename.fchk" without the quotes. This creates a formated checkpoint
 
%file.
 
Now, open this document in excel and delete everything but the
 
%cartesian force constants and save the file. Now open up a new excel doc
 
%and i
mport these cartesian force constants using fixed width cells to get
 
%five columns with all of the lower 
triangle of the symmetric hessian
 
%matrix. Delete everything but these hessian values (ie no text should be
 
%present, just the five columns of data). S
ave this as an excel file and it
 
%should be ready for use.
 
%*********************************************
*****************************
 
 
 
%First we will import an excel file with the reactant state geometry and
 
%save it as Reactant.
 
[filename, pathname]=uig
etfile;
 
s=strcat(pathname,filename);
 
Reactant=xlsread(s);
 
%Next, we will do the same with an excel 
file with the product state
 
%geometry and save it as Product.
 
[filename, pathname]=uigetfile;
 
ss=strcat(pathname,filename);
 
Product=xlsread(ss);
 
%Note, These
 
excel files are arranged in columns from left to right as atom
 
%index number, atom type (ie C, H, 
etc), x coord, y coord, z coord, and
 
%mass (in amu).
 
 
 
%The following makes column vectors of the x, y, z components of each atom
 
%first for the reactant sta
te and then for the product state.
 
Rrinix=Reactant(:,3);
 
Rriniy=Reactant(:,4);
 
Rriniz=Reactant(:,5)
;
 
mir=Reactant(:,6); 
%masses of each reactant atom
 
 
 
Rpinix=Product(:,3);
 
Rpiniy=Product(:,4);
 
Rpiniz=Product(:,5);
 
mip=Product(:,6);
%masses of each product 
atom
 
 
 
%we get rid of any linear momentum between reactant and product states by
 
%setting the origi
n of each of their axis systems at the center of mass of
 
%each. The result is colum vectors of the x, y, and z components
 
%(separately) of all the atoms
 
mRri
nix=Rrinix.*mir;
 
mRriniy=Rriniy.*mir;
 
mRriniz=Rriniz.*mir;
 
 
 
mRpinix=Rpinix.*mip;
 
mRpiniy=Rpiniy.*mip;
 
mRpiniz=Rpiniz.*mip;
 
 
 
 
169
 
Rcomrx=sum(mRrinix)/sum(mir);
 
Rcomry=sum(mRriniy)/sum(mir);
 
Rcomrz=sum(mRriniz)/sum(mir);
 
 
 
Rcompx=sum(mRpinix)/sum(mip);
 
Rcompy=s
um(mRpiniy)/sum(mip);
 
Rcompz=sum(mRpiniz)/sum(mip);
 
 
 
Rrfinal
x=Rrinix
-
Rcomrx;
 
Rrfinaly=Rriniy
-
Rcomry;
 
Rrfinalz=Rriniz
-
Rcomrz;
 
 
 
Rpfinalx=Rpinix
-
Rcompx;
 
Rpfinaly=Rpiniy
-
Rcompy;
 
Rpfinalz=Rpiniz
-
Rcompz;
 
 
 
%Now we must insure that there is no angular momentum 
between the two
 
%structures. To do this, we must ensure that 
the sum of the cross products
 
%of each of the position vectors for each element multiplied by the scalar
 
%of their respective masses is zero.
 
%We use the procedure of [Miller, w. H.; Ruf,B. A.; C
hang, Y.
-
T. J.
 
%chem. phys. 89 (10), 1988] for this. 
 
 
Rri_z0=[0,0,0];
 
 
Rpi_z0=[0,0,0];
 
 
 
 
%First, we find the current angular momentum of the system by summing the
 
 
%mass weighted cross products of the position vectors.
 
 
 
for
 
i = 1: length(Rrfinalx);
 
    
Rri=[Rrfinalx(i
),Rrfinaly(i),Rrfinalz(i)];
 
    
Rpi=[Rpfinalx(i),Rpfinaly(i),Rpfinalz(i)];
 
    
mcross_sub=mir(i)*cross(Rri,Rpi);
 
    
cross_sub=cross(Rri,Rpi);
 
    
if
 
i==1
 
        
mcrossA=mcross_sub;
 
        
crossA=cross_sub;
 
        
mcrossB=mcrossA;
 
       
 
crossB=crossA;
 
    
else
 
        
mcrossA=mcrossB;
 
        
crossA=crossB;
 
        
mcrossB=mcrossA+mcross_sub;
 
        
crossB=crossA+cross_sub;
 
    
end
 
  
 
end
 
%Next, we create a unit vector, N, that points in the direction fo 
the sum
 
%of the mass weighted cr
oss product of the position vectors
 
nx=mcrossB(1)/sqrt(mcrossB(1)^2+mcrossB(2)^2+mcrossB(3)^2);
 
ny=mcrossB(2)/sqrt(mcrossB(1)^2+mcrossB(2)^2+mcrossB(3)^2);
 
nz=mcrossB(3)/sqrt(mcrossB(1)^2+mcrossB(2)^2+mcrossB(3)^2);
 
 
170
 
N=[nx,ny,nz];
 
%Now, before we determine 
the angle that the position vectors of one system
 
%needs to be rotated by (along the N unit vector) we must first project all
 
%atoms onto the plane perpendi
cular to the N unit vector. This will ensure
 
%that there is no overcounting towards the sum of the m
ass weighted dot
 
%product of the position vectors which is used to determine the angle of
 
%rotation.
 
for
 
i=1: length(Rrfinalx);
 
    
Rri=[Rrfinalx(i),Rrfinal
y(i),Rrfinalz(i)];
 
    
Rpi=[Rpfinalx(i),Rpfinaly(i),Rpfinalz(i)];
 
    
Rri_projection=Rri
-
dot(Rri,N)*
N;
 
    
Rpi_projection=Rpi
-
dot(Rpi,N)*N;
 
    
mdot(i)=mir(i)*(dot(Rri_projection,Rpi_projection));
 
end
 
   
 
%phi= angle that Rr is needed to be rotated about t
he vector sum of the mass
 
%weighted cross products of Rr and Rp
 
phi=
-
atan(sqrt(dot(mcrossB,mcrossB))
/sum(mdot));
 
 
 
C=cos(phi);
 
S=sin(phi);
 
t=1
-
cos(phi);
 
T=[ t*(nx^2)+C t*nx*ny
-
S*nz t*nx*nz+S*ny; t*nx*ny+S*nz t*(ny^2)+C t*ny*nz
-
S*nx; t*nx*nz
-
S*ny 
t*ny*nz+S*
nx t*(nz^2)+C];
 
%T is the rotation matrix that rotates Rp about mcrossB by angle phi
 
Rpi_rotx=zeros(
size(Rrfinalx));
 
Rpi_roty=zeros(size(Rrfinaly));
 
Rpi_rotz=zeros(size(Rrfinalz));
 
for
 
i =1: length(Rrfinalx);
 
    
Rpi=[Rpfinalx(i),Rpfinaly(i),Rpfinalz(i)];
 
    
Rpi_rot=T*Rpi.';
 
    
Rpi_rotx(i)=Rpi_rot(1);
 
    
Rpi_roty(i)=Rpi_rot(2);
 
    
Rpi_rotz(i)=Rpi_rot
(3);
 
end
 
 
 
%Now we will check that the rotation is done correctly.
 
for
 
i=1: length(Rrfinalx);
 
    
Rri=[Rrfina
lx(i),Rrfinaly(i),Rrfinalz(i)];
 
    
Rpi=[Rpi_rotx(i),Rpi_roty(i),Rpi_rotz(i)];
 
    
mcross_sub=mir(i)*cross(Rri,Rpi);
 
    
if
 
i==1
 
        
mcrossA=mc
ross_sub;
 
        
mcrossB=mcrossA;
 
    
else
 
        
mcrossA=mcrossB;
 
        
mcrossB=mcrossA+mcross_sub;
 
    
end
 
    
mdot(i)=mir(i)*dot(Rri,Rpi);
 
end
 
 
 
 
171
 
%Next we import the data necessary to create the hessian matrix by
 
%importing an excel file with the low
er triangular elements. This file
 
%should have all of the data from a formated checkpoint file where each
 
%matrix element is listed in five columns.
 
[filename, pathname]=uigetfile;
 
Hess=strcat(pathname,filename);
 
H_xl=xlsread(Hess);
 
hessLTcount=(((3*length
(Rrfinalx)
-
1)*(3*length(Rrfinalx)))/2)+(3*length(Rrfinalx));
 
%First, we create one long column array with 
all of the hessian elements
 
for
 
i=1:ceil(hessLTcount/5);
 
    
if
 
i==1
 
        
Hvect=[H_xl(i,1);H_xl(i,2);H_xl(i,3);H_xl(i,4);H_xl(i,5)];
 
        
 
    
e
lse
 
        
Htempsub=[H_xl(i,1);H_xl(i,2);H_xl(i,3);H_xl(i,4);H_xl(i,5)];
 
        
Hvect=vertcat(Hvect,Htempsub);
 
    
end
 
end
 
%Because of how Hvect was just made, if hessLTcount is not a multiple of 5
 
%then there will be several NAN values in Hvect. We remo
ve t
hose using the
 
%following line.
 
Hvect(find(isnan(Hvect)))=[];
 
numb=0;
 
HessLD=zeros(3*length(Rrfinalx));
 
%Next, we create the lower diagonal Hessian matrix
 
for
 
i=1:3*length(Rrfinalx)
 
    
for
 
j=1:i
 
       
numb=numb+1;
 
        
HessLD(i,j)=Hvect(numb);
 
   
 
end
 
end
 
%now, we make the full hessian.
 
Hessfull=HessLD+tril(HessLD,
-
1).';
 
 
 
%Now we create the mass weighted cartesian hessian which we then will
 
%diagonallize to determine the appropriate eigen vectors (vibrational
 
%modes).
 
 
 
%Lets first create 
another 
array that carries all the masses of our system
 
for
 
i=1:length(mir)
 
    
if
 
i==1
 
        
Mbig_sub=[mir(i);mir(i);mir(i)];
 
        
Mbig=Mbig_sub;
 
    
else
 
        
Mbig_sub=[mir(i);mir(i);mir(i)];
 
        
Mbig=vertcat(Mbig,Mbig_sub);
 
    
end
 
end
 
%Lets now gen
erate the mass weighted cartesian coordinates
 
Hm=zeros(size(Hessfull));
 
for
 
i=1:(3*length(Rrfinalx));
 
 
172
 
    
for
 
j=1:(3*length(Rrfinalx));
 
        
Hm(i,j)=Hessfull(i,j)/sqrt(Mbig(i)*Mbig(j));
 
    
end
 
end
 
 
 
%Now we diagonalize the mass weighted ca
rtesian coord
inates, Hm, and
 
%extract out the eigen vectors ,L.
 
 
 
[V,D,L]=eig(Hm);
 
 
 
%now, we generate the displacement vector that discribes the displacement
 
%between the two equilibrium geometries.
 
for
 
i=1:length(Rrfinalx)
 
    
if
 
i==1
 
    
Rdiff_sub=[Rpi_rotx(i)
-
Rrfin
alx(i);Rpi_roty(i)
-
Rrfinaly(i);Rpi_rotz(i)
-
Rrfinalz(i)];
 
    
Rdiff=Rdiff_sub;
 
    
else
 
    
Rdiff_sub=[Rpi_rotx(i)
-
Rrfinalx(i);Rpi_roty(i)
-
Rrfinaly(i);Rpi_rotz(i)
-
Rrfinalz(i)];
 
    
Rdiff=vertcat(Rdiff,Rdiff_sub);
 
    
end
 
end
 
 
 
freq=real(1.890152771.*(diag(D
).*(4.35974e
-
18)*(1e20)*(6.022e23)*1000/(4*(pi()^2)*(299792548^2)*(1e4))).^(0.5));
%array of frequencies in units 
of cm
-
1
 
 
 
%Now, we calculate the reduced masses, mu, of each vibrational mode
 
lcart=zeros(length(Mbig),length(M
big));
 
for
 
i=1 :length(Mbig);
 
  
  
for
 
j=1:length(Mbig);
 
   
lcart_small(j)=(L(j,i)^2)/Mbig(j);
 
    
end
 
    
lcartsum(i)=sum(lcart_small);
 
    
mu(i)=(lcartsum(i)^(
-
1));
% reduced masses
 
end
 
 
 
 
 
K=((diag(Mbig).^(0.5))*L)'*Rdiff; 
%Duschinsky displacement vector
 
kf=diag(D).*mu.';
%Force constant
s
 
lamda=(219470)*0.5.*kf.*K.^2;
%Reorganization energies in cm^
-
1 I think.
 
%The ~220000 number is to convert from hartrees to wavenumbers.
 
 
 
%Now, we 
remove the rotational and transaltional modes.
 
Kcut=K(7:length(K));
 
freqcut=freq(7:length(freq));
 
lamdacut=
lamda(7:length(lamda));
 
 
 
figure(1)
 
bar(abs(Kcut))
 
xlabel(
'Normal Mode Number'
);
 
ylabel(
'Elements of the Duschinsky K vector in absolute value (Aamu^
1^/^2)'
)
 
 
173
 
 
 
figure(2)
 
stem(freqcut,abs(Kcut))
 
xlabel(
'Frequency (cm^
-
^1)'
);
 
ylabel(
'Elements of the Duschins
ky K vector in absolute value (Aamu^1^/^2)'
);
 
 
 
figure(3)
 
stem(freqcut,Kcut)
 
xlabel(
'Frequency (cm^
-
^1)'
);
 
ylabel(
'Elements of the Duschinsky K vecto
r (Aamu^1^/^2)'
);
 
 
 
figure(4)
 
stem(freqcut,lamdacut)
 
xlabel(
'Frequency (cm^
-
^1)'
);
 
ylabel(
'Reorgonization E
nergy (cm^
-
^1)'
)
 
 
 
lamdatot=sum(lamdacut)
%Total Reorganization energy in cm^
-
1
 
Kabs=abs(Kcut);
 
%Uses the absolute displacements to compute a 
weighted average frequency
 
ABSweightavg=sum(abs(Kcut).*freqcut)/sum(abs(Kcut))
 
rmsfreq_a=(sum((Kabs.^2).*freqcut)/s
um(Kabs.^2))
 
rmsfreq_b=(sum((Kabs.*freqcut).^2)/sum(Kabs.^2))^(0.5)
 
 
 
ReorgWieghtedFreq=sum(lamdacut.*freqcut)/sum(lamdacut)
 
 
 
 
174
 
 
 
 
 
 
 
 
 
 
 
 
 
RE
FERENCES
 
 
 
 
175
 
REFERENCES
 
 
(1)
 
Wenger, O. S. Is Iron the New Ruthenium? 
Chem. Eur. J
. 
2019
, 
25
, 6043

6052.
 
(2)
 
Hexel, G
. B.; Hedrick, J. B.; Orris, G. J. Rare Earth Elements
-
Critical Resources for High 
Technology 
https://pubs.usgs.gov/fs/2002/fs087
-
02.
 
(3)
 
Auböck, Gerald; Chergui, M. Sub
-
50
-
Fs Photoinduced Spin Crossover in [Fe(bpy)
3
]
2+
. 
Nat Chem.
 
2015
, 
7
, 629

633.
 
(4)
 
Fer

Prosperi, T.; 
Turchini, S. A Fast Transient Absorption Study of Co(acac)
3
. 
Front. Chem
. 
2019
, 
7
, 1

12.
 
(5)
 

-
dionato) Cobalt(III) Chelates. 
Structure
-
Redox Potential Relations. 
Inorg. Chem.
 
1987
, 
26
, 26

32.
 
(6)
 
Bendix, J.; Borson, M.; Schäffer, C. E. Accurate Empirical Spin
-
Orbit Coupling Par
ameters 

nd
 
for 
Gaseous nd
q
 
Transition Metal Ions. The Parametrical Multiplet Term Model. 
Inorg. Chem
. 
1993
, 
32
, 2838

2849.
 
(7)
 
McCusker, J. K.; Walda, K
. N.; Dunn, R. C.; Simon, J. D.; Magde, D.; Hendrickson, D. N. 
Subpicosecond 
1

5
T
2
 
Intersystem Cr
ossing of Low
-
Spin Polypyridyl Ferrous Complexes. 
J. Am. Chem. Soc.
 
1993
, 
115
, 298

307.
 
(8)
 
McCusker, J. K.; Walda, K. N.; Magde, D.; Hendrickson, D. N. 
Picosecond Excited
-
State 
Dynamics in Octahedral Cobalt(III) Complexes: Intersystem Crossing versus Inte
rnal Conversion. 
Inorg. Chem
. 
1993
, 
32
, 394

399.
 
(9)
 
Langford, C. H.; Malkhasian, A. Y. S.; Sharma, D. K. Subnanosecond Transients in the Spectra of 
Coba
lt(III) Amine Complexes. 
J. Am. Chem. Soc
. 
1984
, 
106
, 2727

2728.
 
(10)
 
Kirk, A. D.; Langford, C. H.; Sha
rma, D. K. Comparison of Subnanosecond Transients from 
Cobalt(III) Linkage Isomers: Ligand Field Triplets versus the Radical Pair. 
Inorg. Chem
. 
1989
, 
28
,
 
974

975.
 
(11)
 
Hipps, K. W.; Crosby, G. A. Geometrical Distortion and Spin
-
Orbit Splitting in Excited T
ransition 
Metal Complexes. 
Inorg. Chem
. 
1974
, 
13
, 1543

1544.
 
(12)
 
Miskowski, V. M.; Gray, H. B.; Wilson, R. B.; Solomon, E. I. Position of the 
3
T
1g
 

1
A
1
g
 
Transition 
in Hexacyanocobaltate(III). Analysis of Absorption and Emission Results. 
Inorg. Chem
. 
1979
, 
18
, 
1410

1412.
 
(13)
 
Wilson, R. B.; Solomon, E. I. Spectroscopic Studies of Photochemically Important Transition 
Metal Excited States. 2. The 
1
T
1g
, 
3
T
1g
, and 
5
T
2g
 
Excited States of Hexaamminecobalt(III). 
J. Am. 
Chem. Soc
. 
1980
, 
102
, 4085

4095.
 
(14)
 
Langford, C. H. Quantum Yiel
d Variation over Narrow Spectral Regions in Coordination 
Photochemistry. 
Acc. Chem. Res
. 
1984
, 
17
, 96

103.
 
 
176
 
(15)
 
Langford, C. 
H.; Malkhasiah, A. Y. S. Wavelength Dependence over the Envelope of a Single 
Electronic Band in the Photochemistry of the Penta
-
Amm
inechlorocobalt(III) Ion. 
J. Chem. Soc., 
Chem. Commun
. 
1982
, 1210

1211.
 
(16)
 
Shalhoub, G. M. Co(acac)
3
: Synthesis, Reactions,
 
and Spectra An Experiment for General 
Chemistry. 
J. Chem. Ed.
 
1980
, 
57
, 525

526.
 
(17)
 
Collman, J. P.; Moss, R. A.; Maltz, H.; Hein
del, C. C. The Reaction of Metal Chelates. I. 
Halogenation of Metal Chelates of 1,3
-
Diketones. 
J. Am. Chem. Soc
. 
1961
, 
83
, 53
1

534.
 
(18)
 
Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; 
Scalmani, G.; Barone, V.
; Petersson, G. A.; Nakatsuji, H.; Li, X.; Caricato, M.; Marenich, A. V.; 
Bloino, J.; Janesko, B. G.; Gomperts, R.; Mennucci,
 
B.; Hratch, D. J. Gaussian 16, Revision C.01. 
2016.
 
(19)
 
Humphrey, W.; Dalke, A.; Schulten, K. VMD: Visual Molecular Dynamics. 
J. 
Mol. Graph. Model
. 
1996
, 
14
, 33

38.
 
(20)
 
Ando, H.; Iuchi, S.; Sato, H. Theoretical Study on Ultrafast Intersystem Crossing of
 
Chromium(III) 
Acetylacetonate. 
Chem. Phys. Lett.
 
2012
, 
535
, 177

181.
 
(21)
 
Miller, W. H.; Ruf, B. A.; Chang, Y.
-
T. A Diabatic Reaction Path Hamiltonian. 
J. Chem. Phys.
 
1988
, 
89
, 6298

6304.
 
(22)
 
Barnum, D. W. Electronic Absorption Spectra of Acetyl
-
Acetonat
o Complexes. 1. Complexes With 
Trivalent Transition Metal Ions. 
J. Inorg. N
ucl. Chem.
 
1961
, 
21
, 221

237.
 
(23)
 
Heller, E. J. The Semiclassical Way to Molecular Spectroscopy. 
Acc. Chem. Res.
 
1981
, 
14
, 368

375.
 
(24)
 
Lemke, H. T.; Kjaer, K. S.; Hartsock, R.; v
an Driel, T. B.; Chollet, M.; Glownia, J. M.; Song, S.; 
Zhu, D.; Pace, E.; 
Matar, S. F.; Nielsen, M. M.; Benfatto, M.; Gaffney, K. J.; Collet, E.; Cammarata, 
M. Coherent Structural Trapping through Wave Packet Dispersion during Photoinduced Spin State 
Swit
ching. 
Nat. Commun.
 
2017
, 
8
, 15342.
 
(25)
 
Auböck, G.; Chergui, M. Sub
-
50
-
Fs 
Photoinduced Spin Crossover in [Fe(bpy)
3
]
2+
. 
Nat. Chem.
 
2015
, 
7
, 629

633.
 
(26)
 
Carey, M. C.; Adelman, S. L.; McCusker, J. K. Insights into the Excited State Dynamics of Fe(II) 
Polyp
yridyl Complexes from Variable
-
Temperature Ultrafast Spectroscopy. 
Chem. Sci.
 
2019
, 
10
, 
134

144.
 
(27)
 
Monat, J. E.; McCusker, J. K. Femtosecond Excited
-
St
ate Dynamics of an Iron(II) Polypyridyl 
Solar Cell Sensitizer Model. 
J. Am. Chem. Soc.
 
2000
, 
122
, 4092

4097.
 
(28)
 
Ma


as, E. M. S.; Mustalahti, S.; Myllyperkiö, P.; Kunttu, H.; Pettersson, M. Role of Vibrational 
Dynamics in Electronic Relaxation of Cr(ac
ac)
3
. 
J. Phys. Chem. A
 
2015
, 
119
, 2727

2734.
 
(29)
 
Iwamura, M.; Takeuchi, S.; Tahara, T. Ultrafast Exci
ted
-
State Dynamics of Copper(I) Complexes. 
Acc. Chem. Res.
 
2015
, 
48
, 782

791.
 
(30)
 
Bixon, M.; Jortner, J. Intramolecular Radiationless Transitions. 
J. Che
m. Phys.
 
1968
, 
78
, 715

726.
 
 
177
 
(31)
 
Englman, R.; Jortner, J. The Energy Gap Law for Radiationless Transit
ions in Large Molecules. 
Mol. Phys.
 
1970
, 
18
, 145

164.
 
(32)
 
Zhang, W.; Alonso
-
Mori, R.; Bergmann, U.; Bressler, C.; Chollet, M.; Galler, A.; Gawelda, W.; 
Hadt, R. G.; Hartsock, R. W.; Kroll, T.; 
et al
. Tracking Excited
-
State Charge and Spin
-
Dynamics 
in Iro
n Coordination Complexes. 
Nature
 
2014
, 
509
, 345

348.
 
(33)
 
Britz, A.; Gawelda, W.; Assefa, T. A.; Jamula, L. L.; Yarranton, J. T.; Galler, A.; Khakhulin, D
.; 
Diez, M.; Harder, M.; Doumy, G.; 
et al
. Using Ultrafast X
-
Ray Spectroscopy to Address Questions 
in 
Ligand
-
Field Theory: The Excited State Spin and Structure of [Fe(dcpp)
2
]
2+
. 
Inorg. Chem.
 
2019
, 
58
, 9341

9350.1
 
(34)
 
Schrauben, J. N.; Dillman, K. L.; Beck
, W. F.; McCusker, J. K. Vibrational Coherence in the 
Excited State Dynamics of Cr(acac)
3
: Probing the
 
Reaction Coordinate for Ultrafast Intersystem 
Crossing. 
Chem. Sci.
 
2010
, 
1
, 405

410.
 
(35)
 
Feit, M. D.; Fleck, J. A.; Steiger, A. Solution of the Schrodin
ger Equation by a Spectral Method. 
J. 
Comput. Phys
. 
1982
, 
47
, 412

433.
 
(36)
 
Tanner, J. J. A Computer
-
B
ased Approach to Teaching Quantum Dynamics. 
J. Chem. Ed.
 
1990
, 
67
, 917
-
921.
 
(37)
 
Reber, C.; Zink, J. I. Interference Dips in Molecular Absorption Spectra Calculated for Coupled 
Electronic State Potential Surfaces. 
J. Chem. Phys.
 
1992
, 
96
, 2681
-
2690.
 
 
 
 
178
 
Chap
ter 4: Geometric Control of Ultrafast Charge
-
T
ransfer Relaxation Dynamics in Fe(I
I)
-
B
ased 
Chromophores: Using Coherence to Inform Synthetic Design
 
The work presented in this chapter was a collaborative effort between this author and Sara Adelman and 
was su
bmitted for publication in 
Nature
 
under the following citation: [Paulus, B. C.; A
delman; Jamula, L. 
L.; McCusker, J. K. 
Nature
 
2020
, xx, xx
-
xx]. BCP collected and analyzed all ultrafast TA data, performed 
DFT optimizations and frequency calculations, and m
ade vibrational mode figures. SLA synthesized all 
complexes and carried out groun
d
-
state and nanosecond time
-
resolved characterizations. LLJ developed 
the initial synthetic route to produce the main ligand and compound 
2
.
 
4.1. Introduction
 
 
Quantum mechani
cal coherences have received significant attention for their potential role in 
co
upling to and/or directing photophysical processes.
1
 
Coherences have been observed and analyzed in a 
variety of chemic
al systems ranging from small molecules to proteins in the context of characteriz
ing 
ultrafast dynamics.
2

8
 
These studies represent pioneering efforts that are providing new, fundamental 
insights into the quantum mechanical nature of photo
-
induced processes. As suggested r
ecently,
1
 
observation of
 
coherences during an electronic transition also has the potential to inspire synthetic design 
to access historically difficult transformations due to the fact that in many ch
emical systems there exists an 
intricate relationship between electronic and nucl
ear degrees of freedom. This leads to significant mixing 
of their wavefunctions and an interdependence of these molecular properties that can map onto the 
photophysical and ph
otochemical dynamics. In systems such as these, vibronic (as opposed to purely 
el
ectronic and purely vibrational) coherences can in principle provide unique insight into the nature of this 
coupling, potentially affording subtle clues for how one could exer
t some degree of control through 
informed structural, compositional, and environm
ental modifications.
9

12
 
 
Molecules for which a design approach targeting the interplay between electronic and vibrational 
degree
s
 
of freedom might be particularly advantageous are those involving first
-
row transition metal ions. 
Compounds of this class are actively being pursued as earth
-
abundant replacements for applications that 
leverage photo
-
induced electron transfer reactivity
,
13,14
 
but such systems have been challenging to develop. 
 
179
 
Ruthenium
-
bas
ed compounds, for example, have proven to be extremely effective for light
-
actuated redox 
chemistry, but their isoelectronic iron(II) analogs typically undergo ultrafast (< 200
 
fs) conversion 
from the 
initially formed metal
-
to
-
ligand charge transfer (MLCT)
 
excited state to low
-
lying ligand field (LF) excited 
states.
15

17
 
These latter states are localized on the metal center and are therefore not characterized by charge 
separation, a fa
ct that significantly limits their ability to engage in electron
 
transfer reactions. The 
underlying reasons for this dramatic difference in photo
-
induced excited
-
state reactivity between first
-
row 
and second
-
 
and third
-
row
-
based chromophores have recently 
been discussed in detail;
18
 
the fact that the 
origin of th
ese effects stems from fundamental properties of atomic orbital wavefunctions that lie at the 
very core of quantum mechanics underscores the difficulty associated with circumventing this probl
em.
 
 
Despite the challenging nature of this problem, the desirab
ility of translating the photo
-
induced 
reactivity endemic to complexes employing some of the rarest elements on Earth to their more abundant 
congeners in the first transition series has cataly
zed research geared toward altering this paradigm. Due in 
part t
o its isoelectronic relationship with ruthenium, complexes of iron have attracted considerable attention. 
The most successful efforts in this area have focused on energetically reordering the 
LF and charge
-
transfer 
excited states 

 
primarily through the us
e of strongly 

-
donating ligands 

 
leading to the preparation of 
molecules exhibiting charge
-
transfer excited
-
state lifetimes on the order of a nanosecond.
19

22
 
With this 
report,
 
we introduce an alternate approach, one that exploits the ultra
fast nature of MLCT
-
state deactivation 
as well as the large amplitude geometric distortions that accompanies it. Specifically, we demonstrate that 
the observation and identification of excited
-
state coherence can provide critical insight into the geometric
 
nature of the reaction coordinate associated with MLCT
-
to
-
LF conversion in such a way as to inform 
synthetic design, in effect creating new strategies for imparting synthetic control over ult
rafast dynamics.
 
4.2
.
 
Experimental Methods
 
4.2.1
.
 
Synthesis
 
All 
synthesis and ground state characterization was performed by Sara Adelman of the McCusker 
group.
 
 
 
180
 
4.2.2
.
 
Density Functional Theory Calculations
 
Vibrational frequencies were 
calculated on the high spin (HS) and low spin (LS) states of [Fe(L)]
2+
 
(
1
) as well 
as the LS ground state of [FeCu
2
(L)]
4+
 
(
2
) using the B3LYP functional and 6
-
31G basis set. 
Frequency calculations of the LS 
1
 
were also conducted with polarizability functio
ns and showed little 
change in the calculated frequencies. All calculations were co
nducted using Gaussian16 code
23
 
on the High 
Performanc
e Computing Center (HPCC) at Michigan State University. Mass weighted vector displa
cement 
diagrams of each pertinent vibrational mode were created in the Visual Molecular Dynamics (VMD) 
program.
24
 
4.2.3
.
 
Ultrafast Transient Absorption Spectroscopy
 
The ultrafast transient absorption spectroscopy setup was des
cribed in Chapter 2 of this dissertation.
 
 
Samples were dissolved in spectrophotome
tric grade acetonitrile in an argon
-
filled glovebox such 
that they had an absorbance between 0.3 and 0.8 at the center of the pump wavelength in a 1
 
mm air
-
free 
cuvette. Sam
ples were checked for degradation by comparing UV
-
Vis spectra taken before and afte
r the 
experiment. The instrument response function (IRF) for  single wavelength kinetics traces was determined 
from the optical Kerr effect in methanol by fitting the early 
half of the signal to a Gaussian and extracting 
the full width at half maximum (FWH
M).
25
 
During full spectrum measurements, the IRF and dispersion ar
e 
measured by collecting TA data on the neat solvent and then used to chirp correct
 
the sample with solute. 
During single wavelength measurements, the probe beam was directed into a monochromator placed after 
the sample. The monochromator selected a 3.8
 
nm
 
FWHM slice of the probe beam spectrum to be 
measured. The center wavelength of thi
s slice of the probe spectrum is specified for each single wavelength 
TA trace. Each experiment was repeated at least three times and each value listed is the average of all
 
values 
obtained by fit, while the error is listed as one standard deviation.
 
 
Popu
lation dynamics were modeled by use of equation 4.1, in which the total signal, S, is described 
by the summation of several exponentially decaying components convolved with 
a Gaussian IRF (G).
 
 
 
4.1
 
 
181
 
As mentioned above, for single 
wavelength measurements, the IRF was determined by use of the optical 
Kerr effect in methanol. Next, each dataset was fit to as few exponential components as necessary to 
adequately reproduce the data. For 
1
, this involved tw
o components: one to describe t
he ultrafast MLCT 

5
T
2g
 

1
A
1g
 
ground state recovery which effectively 
describes the y
-
offset present following MLCT decay. This long time component was locked in at 25
 
ns as 
determ
ined by Sara Adelman of the McC
usker group using nanosecond transient absorption spectroscopy; 
however its exact value does not have a significant impact on the fit as the measured window in time is 
much slower than the rate of this process. For 
2
, we use 
three exponential components wi
th the third and 
slowest component describing 
5
T
2g
 

1
A
1g
 
ground state recovery locked in at 7
 
ns (determined separately 
on the ns TA system by Sara Adelman of the McCusker group).
 
 
For fitting oscillatory components, we empl
oy a linear predictive singular value decomposition 
(LPSVD) program written in MA
TLAB originally by Dr. Andrey Demidov and Paul Champion and edited 
by Dan Roscioli of the Beck group at MSU. This program fits the data to a series of exponentially damped 
cos
inusoids as shown in equation 4.2.
 
 
 
4.2
 
This program a
llows the user to fit population dynamics using zero frequency components as well as extract 
exponentially damped oscillations. Due to the drastically different time scales be
tween some of the 
population dynamics and the damping times of the vibrational co
herences, we use LPSVD to fit the 
oscillations on the first ~2
 
ps of data and use the exponential fitting described above for extracting 
population dynamics. The program makes
 
use of the Fourier relationship between damping time and width 
in the frequency 
domain to produce a vibrational spectrum which is then compared to as fast Fouier 
transform (FFT) of the residuals of a fit to the population dynamics. To identify coherent os
cillations due 
to the solvent and cuvette, TA experiments were performed on neat 
solvent and an empty cuvette.
 
 
 
 
182
 
4.3
.
 
Results and Discussion
 
 
The excited state ultimately formed subsequent to ML
CT
-
LF state conversion in Fe(II) polypyridyl 
complexes is characterized by a significant (~0.2 
Å
) lengthening of Fe
-
N bonds,
26
 
which in turn has fueled 
a long
-
standing assumption that the Fe
-
N breathing mode dominates the excited state reaction 
coordinate.
15,16
 
While structural changes along the Fe
-
N coordinate unquestionably occur, the available data 
do not allow one to differentiate between causation, in which the sp
ecific mode(s) is kinetically competent 
for determ
ining the dynamics, versus correlation in which the structural change is a consequence of a 
process that is driven by other degrees of freedom. There is a significant body of work going back to 

e of the angular overlap model to rationalize the 
rapid racemization of [Fe(phen)
3
]
2+
, suggesting 
that torsional modes may play a role in both the ground
-
state recovery (GSR) process as well as MLCT
-
state deactivation.
27

29
 
When viewed through this lens, the possibility t
hat other degrees of freedom could 
be involved in 
defining the reaction coordinate for the overall photophysical transformation of these 
chromophores must be considered.
 
 
With these ideas in mind, we designed a new iron(II) complex based on a polypyridyl l
igand (L) 
that could impose stricter steric 
rigidity about both Fe
-
N breathing and torsional degrees of freedom that 
tris
-
bpy while still maintaining some level of structural tunability. This structural motif employs a 
macrocyclic cage ligand first describ
ed by Lehn
30
 
and is composed of three bipyridine (bpy) units tethered 
together by tertiary amine
-
capped triimines (Figure
 
4.1). The bpy
-
based cored of the assembly is well suited 
Figure
 
4.1.
 
Drawing of cage ligand (
L
) used in this study.
 
 
183
 
to bind Fe(II), whereas the fo
ur nitrogen atoms that comprise the imine ca
ps provide for the introduction of 
additional ions that can impart added rigidity to the overall framework.
 
 
The synthesis and characterization of the three compounds prepared for this study was performed 
by Sara
 
Adelman of the McCusker group and will not 
be shown here. Single
-
crystal x
-
ray structures for 
[Fe(L)]
2+
  
(
1
), [FeCu
2
(L)]
4+
 
(
2
), and [ZnCu
2
(L)]
4+
 
(
3
) are shown in Figures
 
4.2 a
-
c respectively. Focusing 
initially on compound 
1
, we find that the primary coor
dination sphere about the Fe(II) center is 
u
nremarkable and indeed typical for low
-
spin Fe(II) polypyridyl complexes with the exception that the 
ligand framework imparts a local symmetry about the iron center closer to D
3h
 
than the D
3
 
symmetry that 
charact
erizes most tris
-
bidentate coordination comp
ounds. The optical properties of compound 
1
 
are 
dominated in the v
isible by an intense, broad feature centered at 580
 
nm (Figure
 
4.3) assigned to an MLCT 
absorption envelope. Although we do note an attenuation i
n the intensity of the absorpt
ion envelope relative 
to [Fe(bpy)
3
]
2+
 
-
 
a consequence, we believe, of the aforementioned change in symmetry 

 
the overall band 
shape is essentially the same for the two compounds. The absorption spectrum for compound 
3
 
reveals
 
features toward the blue regi
on associated with the Cu
+
 
chromophores, but these absorptions do not 
significantly overlap with the red portion of the Fe(II)
-
based MLCT transition that will be the focus of the 
experiments described below.
 
 
a
)
 
b
)
 
c
)
 
Figure
 
4.2
. 
Single
-
crystal x
-
ray structures of the cations of 
(a) [Fe(L)](PF
6
)
2
 
(
1
)
, 
(b) 
[FeCu
2
(L)](PF
6
)
4
 
(
2
), and (c) [ZnCu
2
(L)](PF
6
)
4
 
(
3
). The hydrogen atoms have 
been 
removed for clarity. For metal ions: red
 
=
 
Fe, green
 
=
 
Cu, light blue
 
=
 
Zn.
 
 
184
 
 
Time
-
resolved e
lectronic absorption data acquired on [Fe(L)](PF
6
)
2
 
(
1
) in CH
3
CN solution 
following 
1
A
1g
 

1
MLCT excitation at 600
 
nm are illustrated in Figure
 
4.4. As suggested by previous work 
on low
-
spin iron(II) polypyridyls
31
 
as well as s
pectroelectrochemical measurements on compound 
1
 
specifically 
(Figure
 
4.5), the ligand
-
based radical anion 

 
which serves as an optical marker for the presence 
of the MLCT state in this class of compounds 

 
will manifest as a broad absorption in the far re
d region of 
the visible spectrum. The absence of such a featur
e at 

 
> 600
 
nm in Figure
 
4.4a indicates that the initially 
formed charge
-
transfer excited state for this compound is depopulated in less than 200
 
fs following 
excitation. This is confirmed by s
ingle
-
wavelength kinetic data (Figures
 
4.4 b and d) that allow
 
us to ascribe 
a time constant for MLCT deactivation of 110
 
±
 
30
 
fs under these conditions. This time scale is fairly 
typical for MLCT relaxation in Fe(II) polypyridyl complexes;
15,16,32
 
in a simila
r vein, we believe that the 
MLCT state is rapidly converting to the ligand
-
field manifold of the transition metal ion, ultimately forming 
the high
-
spin 
5
T
2g
 
state as the lowest energy excited st
ate of the system. Nanosecond time
-
resolved 
electronic absorpt
ion measurements performed by Sara Adelman of the McCusker group reveal single 
exponential kinetics leading to ground
-
state recovery with a time constant of 24 
±
 
2 ns.
 
Figure
 
4.3.
 
Ground state ele
ctronic absorpti
on spectra of [Fe(L)](PF
6
)
2
 
(
1
), [FeCu
2
(L)](PF
6
)
4
 
(
2
), and 
[ZnCu
2
(L)](PF
6
)
4
 
(
3
). The spectra were acquired at room temperature in CH
3
CN solutions.
 
 
(
3
)
 
 
185
 
a)
 
b)
 
c)
 
d)
 
e)
 
Figure
 
4.4
.
 
Femtosecond time
-
resolved absorption data of [Fe(L)](
PF
6
)
2
 
(
1
) acquired in CH
3
CN 
solution.  (a) Full spectrum data collected over the first few ps following MLCT excitation at 600 nm. 
The anomalous features near 600
 
nm are due to scattered light from the 
excitation source. (b) Single 
wavelength kinetics trac
e monitored at 608
 
nm with LPSVD fit. (c) Resulting vibrational spectrum 
corresponding to the FFT of the residuals of an exponential fit (blue) and the oscillations extracted 
through LPSVD fitting (red)
. (d) Single wavelength kinetics trace monitored at 65
8
 
nm with LPSVD 
fit. (e) Vibrational spectrum of the kinetic data collected at 658
 
nm.
 
 
186
 
 
Time
-
resolved absorptio
n da
ta acquired on compound 
1
 
revealed coherent oscillations 
superimposed on the single
-
wavelength kinetics. The frequencies giving rise to these oscillations were 
extracted via both fast Fourier transform (FFT) and linear predictive singular value decompo
siti
on 
(LPSVD) analyses (Figure 4.4c and e); the latter also afforded the dephasing times associated with these 
oscillations (vida infra). The strongly allowed nature of the MLCT absorption in compound 
1
 
requires that 
we consider both ground
-
 
and excited
-
s
tate
 
origins for these features. In this regard, we argue that the 
observed dephasing times are too rapid to be attributable to purely vibrational coherences on the ground 
electronic state from impulsively stimulated Raman scattering where picosecond depha
sing
 
times are more 
typical.
9,33

35
 
The notion that instead these coherences are generated directly on the MLCT excited
-
state 
surface and survive the transfer to the ligand
-
field manifold is supported by the fac
t th
at our observations 
are similar to those reported by Chergui (127 and 157 cm
-
1
, 

OSC
 
= 400 
±
 
100 fs)
16
 
and Gaffney (126
 
cm
-
1
, 

OSC
 
= 310 
±
 
10
 
fs),
15
 
both of whom associated these featur
es to the excited state(s) of [Fe(bpy)
3
]
2+
. 
Accordingly, we assign the oscillations to vibronic coherences in the excited state(s) of
 
compound 
1
. We 
note the similarities in the MLCT decay rates, frequencies, and dephasing times of the coherently activated
 
vibrational modes between compound 
1
 
and [Fe(bpy)
3
]
2+
, suggesting that compound 
1

[Fe(bpy)
3
]
2+
 
extends not only to t
he ground
-
state absorption spectrum and MLCT decay kinetics, but likely 
to the nature of the reaction coordinate for MLCT
-
t
o
-
LF conversion as well.
 
Figure 4.5.
 
Spectroelectrochemical data on [Fe(L)]
2+
 
(
1
) collected by Sara Adelman.
 
 
187
 
 
Coherent oscillations can be observed for degrees of freedom that are linked or orthogonal to a 
reaction c
oordinate of interest and are therefore not necessarily diagnost
ic in and of themselves. We 
hypothesized that one marker for identifying coherences that are vibronic in nature 

 
and therefore 
potentially coupled to an excited
-
state evolution trajectory def
ined by significant geometric changes 

 
could 
be their dephasing
 
times. Coherence that is purely electronic in origin is expected to dephase on a sub
-
100
 
fs time scale in most chemical systems.
36
 
In contrast, pure vibrational coherences can last for several 
picoseconds under certain conditions.
9,33

35,37,38
 
Since a vibronic feature implies a mixing of electronic and 
vibrational degrees of freedom, we posited that coherent oscillations that were vibronic in origin would 
exhibit dephasing times intermed
i
ate between the two extremes just mentioned, e.g., on a time scale of 
hundreds of fs. We therefore proposed to use this hypothesis as a basis for selecting which subset of the 
3N
-
6 vibrational degrees of freedom of the chromophore to target for chemical m
o
dification. The ultimate 
test of whether these modes are indeed part of the reaction coordinate would come from experiment, i.e., if 
the modes we identify in this manner are helping to define the reaction coordinate for ultrafast MLCT
-
state 
relaxation, we
 
should observe a concomitant effect on the MLCT
-
state lifetime upon synthetic tailoring of 
these specific degrees of freedom.
 
 
 
The data presented in Figure 4.4 reveal features at 127 (5) cm
-
1
, 156 (2)
 
cm
-
1
, 173 (3)
 
cm
-
1
, and 215 
(12)
 
cm
-
1
, all of which e
x
hibit dephasing times that we would consider intermediate in nature (Table
 
4.1). 
DFT frequency calculations carried out on the high
-
spin (i.e. 
5
T
2g
) excited state of compound 
1
 
shows 
reasonable agreement with these experimental values, lending further sup
p
ort to our assertion that the 
observed oscillations are associated with the excited
-
state manifold of compound 
1
. We can immediately 
assign the experimentally observed 173 (3) cm
-
1
 
mode to the pair of energetically isolated doubly 
 
Coherent Vibrational Modes of 
1
 
Frequency, cm
-
1
 
12
7
 
(
5
)
 
15
6
 
(
2
)
 
17
3
 
(
3
)
 
21
5
 
(1
2
)
 
Dephasing Time, fs
 
230 (90)
 
300 (30)
 
360 (230)
 
730 (600)
 
 
 
Table
 
4.1.
 
Summary of the coherent vibrational modes of [Fe(L)](PF
6
)
2
 
(
1
) 
observed 
during TA experiments following 600
 
nm excitation in acetonitrile solution.
 
 
188
 
degenerate modes 
(calcul
ated at 174
 
cm
-
1
) illustrated in Figure
 
4.6 b and c, corresponding to a wagging 
motion of the imine nitrogen atoms of the capping fragment concomitant with in
-
plane rocking of the bpy 
moieties. Possible candidates for the 156
 
(2)
 
cm
-
1
 
mode are three
 
featur
es calculated at 151
 
cm
-
1
, 152
 
cm
-
1
, 
and 157
 
cm
-
1
. All three of these calculated vibrations are characterized by symmetric torsional motion about 
the tren cap coupled with a distortion of the primary coordination sphere of the metal center in which 
the 
lig
ated nitrogen atoms tilt (Figure
 
4.6a). The overall distortion is reminiscent of a modified Bailar twist 
recently described by the Jakubikova group.
29
 
The assignment of the 127 (5) cm
-
1
 
feature is more 
challenging due to the high density of calculated modes in this region and the fact that the vibrational period 
for this frequency range is longer than the obser
ved dep
hasing time (thereby resulting in larger uncertainty 
in the measured frequency). A similar challenge arises with the 215 (12)
 
cm
-
1
 
mode due to its relatively 
small amplitude.
 
 
The ambiguity associated with assigning these latter two modes notwithsta
nding, 
the coherence 
data on compound 
1
 
is clearly revealing large amplitude motion of both the primary coordination sphere 
around the iron center that is propagating across the supramolecular structure to the polyimine capping 
Calc. 157
 
cm
-
1
 
Exp. 156 cm
-
1
 
Calc
. 174
 
cm
-
1
 
Exp. 173 cm
-
1
 
Calc. 174
 
cm
-
1
 
Exp. 173 cm
-
1
 
a)
 
b)
 
c)
 
Figure
 
4.6. 
DFT calculated vector displacement diagrams of several of the coherently excited 
vibrational modes of [Fe(L)](PF
6
)
2
 
(
1
).
 
 
189
 
groups. This information imm
ediatel
y suggested an approach whereby we introduce an ion that can be 
sequestered by the N
4
 
coordination environments on the periphery of the ligand framework. The new metal
-
ligand interactions in this region should attenuate the degrees of freedom identi
fied fr
om the coherence data 
and illustrated in Figure
 
4.6: if these modes are important with regard to the reaction coordinate for MLCT
-
to
-
LF state conversion, we should see an increase in the observed lifetime of the MLCT excited state of the 
chromophore.
 
 
Fig
ure
 
4.2b shows the single crystal x
-
ray
 
structure of the cation of [FeCu
2
(L)](PF
6
)
4
 
(
2
) in which 
Cu(I) ions have been introduced into the two capping N
4
 
cavities of the ligand. A comparison of the Fe
-
(bpy)
3
 
cores of compounds 
1
 
and 
2
 
(Figure
 
4.7) confirms 
that the presence of the Cu(I) ions is 
not 
significantly perturbing the primary coordination sphere of the Fe(II) chromophore in the ground state, 
implying that the MLCT excited state of the Fe(II) species in compound 
2
 
should be essentially identical 
to t
hat of compound 
1
. To further establish
 
the electronically benign nature of the Cu(I) ions in compound 
2
, we also prepared [ZnCu
2
(L)]
4+
 
(
3
, Figure 4.2c) where the replacement of Fe(II) by Zn(II) allows us to 
look selectively at the spectral features associ
ated with the Cu(I) ions due to the spe
ctroscopically silent 
nature of Zn(II). It can be seen that absorbances from the Cu(I) ions are at much higher energy than those 
of the Fe(II) site, consistent with the yellow color of this compound (Figure
 
4.3). Thes
e near
-
UV features 
are most likely MLCT
 
in nature involving the imine groups of the capping ligand. Excitation of compound 
2
 
at 600
 
nm 

 
the same wavelength used to acquire the data on compound 
1
 

 
is therefore cleanly selective 
for the Fe(II) chromophore.
 
Both the structural and ground
-
state s
pectroscopic data on compounds 
1
 
and 
2
 
 
Figure
 
4.7.
 
Overlay of the Fe
-
N
6
 
region of [Fe(L)](P
F
6
)
2
 
(
1
) and [FeCu
2
(L)](PF
6
)
4
 
(
2
) crystal structures. 
The complex drawn with red bonds is complex 
1
 
and the complex with purple bonds is complex 
2
.
 
 
190
 
therefore establishes that excitation at 600
 
nm creates the same type of excited state in both chromophores, 
i.e. a 
1
MLCT state associated with the Fe(II)
-
bpy fragment.
 
 
Ultrafa
st time
-
resolved absorption data were collected on compound 
2
 
and are shown in 
Figure
 
4.8. 
As with compound 
1
, compound 
2
 
exhibits coherent oscillations superimposed on the kinetics that define 
its population dynamics (Figure
 
4.8a, 4.8c, 4.A1a). An analysi
s of these features yields frequencies that are 
quite similar to those observed
 
for compound 
1
 
(Figures 4.8b, 4.8d, 4.A1b). We observe a slight lengthening 
of the dephasing times of these coherent oscillations relative to compound 
1
, which may indicate a p
artial 
decoupling of these degrees of freedom from the ultrafast electronic
-
sta
te evolution of compound 
2
.  Similar 
dephasing times are observed following MLCT excitation of compound 
2
 
in several different solvents 
Figure
 
4.8.
 
Representative TA data on [FeCu
2
(L)](PF
6
)
2
 
(
2
) in acetonitrile following 
600
 
nm excitation whil
e monitoring at (a) 700 nm and (c) 650
 
nm (blue) with fit to equation 
4.1 (red) and inset showing LPSVD fit
 
to equation 4.2. Corresponding vibrational spectra 
(b) and (d) resulting from LPSVD fit (red) and FFT of the residuals of the exponential fit 
(blue)
. The appearance of an additional peak in the low frequency region in the LPSVD 
spectrum in part d can be a
ttributed to error in exponential fitting in the LPSVD program, 
likely arising from fitting a small portion of the overall data set.
 
a)
 
b
)
 
d
)
 
c
)
 
 
191
 
(Figure
 
4.9), which strongly supports an
 
intramolecular origin for the dephasing mechanism. A summary 
of the coherent oscillations of compound 
2
 
observed across each solvent can be seen in Table
 
4.2. Analyzing 
these data in a manner analogous to compound 
1
, DFT calculations allow us to ascribe t
he 219
 
cm
-
1
 
mode 
of compound 
2
 
as one of three modes calculated between 225 and 227
 
cm
-
1
, each of which exhibit a rocking 
Solvent
 
 
Coherent Vibrational Modes of 
2
 
MeCN
 
Frequency, cm
-
1
 
116 (3)
 
183 (
3
)
 
2
19
 
(
5
)
 
Dephasing Time, fs
 
600
 
(
210
)
 
810
 
(
270
)
 
5
7
0 (
19
0)
 
BuCN
 
Frequency, cm
-
1
 
114 (1)
 
182 (2)
 
219 (4)
 
Dephasing Time, fs
 
800 (140)
 
580 (140)
 
450 (170)
 
Acetone
 
Frequency, cm
-
1
 
117 (1)
 
183 (2)
 
222 (2)
 
Dephasing Time, fs
 
660 (200)
 
870 (400)
 
540 (140)
 
 
Table
 
4.2.
 
Summary of the coherent v
ibrational modes of [FeCu
2
(L)(PF
6
)
4
 
(
2
) in 
different solvents. All errors are expressed as the standard deviation of multiple trials.
 
Figure
 
4.9.
 
Representative TA data on [FeCu
2
(L)](PF
6
)
4
 
(
2
) in (a) butyronitrile and (b) acetone 
following 600
 
nm exci
tation and monitoring at 700
 
nm with insets showing raw data (blue) and 
LPSVD fits (red) over the first few picoseconds. (c) and (d) show 
the vibrational spectra resulting 
from the FFT and LPSVD analyses of the data in butyronitrile and acetone respectivel
y.
 
a)
 
b)
 
c)
 
d)
 
 
192
 
motion of the bpy moieties along with an antisymmetric Cu
-
N stretch in the imine cap (Figure
 
4.10). A 
mode calculated at 186
 
cm
-
1
, w
hich is characterized by mot
ions similar to those seen for the 174
 
cm
-
1
 
mode 
of compound 
1
, is likely the mode detected at 183
 
cm
-
1
. Finally, the energetic match between the mode 
detected at 116 cm
-
1
 
with a metal ligand breathing mode identified by Gaffney
 
and co
-
workers
15
 
and a set 
of three near
-
de
generate modes calculated be
tween 118
 
and 119
 
cm
-
1
 
suggests a similar assignment to that 
of Gaffney.
 
 
Full spectrum time
-
resolved data on compound 
2
 
are illustrated in Figure
 
4.11.
 
Whereas spectral 
features associated with the radical anion of the ligand r
ed of 600
 
nm are absent for 

t > 200
 
fs
 
in compound 
1
, the spectra in Figure
 
4.11 clearly show the presence of a strong absorption in this region that persists for 
several picoseco
nds following MLCT excitation of compound 
2
. Fitting of the single
-
wavelengt
h data in 
this region reveals biphasic kinetics with time constants of 

1
 
= 430 
±
 
50
 
fs and 

2
 
= 2.6 
±
 
0.1
 
ps. 
Importantly, the longer component defines the eventual loss of absorp
tion in the region of the bpy
-
 
chromophore: we therefore assign the decay of
 
the MLCT state of compound 
2
 
to its ligand
-
field manifold 
to 

2
. This represents a > 20
-
fold increase in the lifetime of the MLCT state of this system and is to our 
Calc
.
 
118
 
cm
-
1
 
Exp.
 
116
 
cm
-
1
 
Calc
.
 
186
 
cm
-
1
 
Exp.
 
183
 
cm
-
1
 
Calc
.
 
186
 
cm
-
1
 
Exp. 
183
 
cm
-
1
 
Cal
c. 
225
 
cm
-
1
 
Exp. 219
 
cm
-
1
 
a
)
 
b
)
 
c
)
 
d
)
 
Figure
 
4.10.
 
DFT calculated vector displacement diagrams of the coherent modes of 
2
 
observed in this study.
 
 
193
 
knowledge the longest
-
lived charge
-
transfer excited state for a low
-
spin Fe(II) tris
-
poly
pyridyl complex in 
fluid soluti
on. 
We believe this is the first example in which coherence data has been used to inform the 
synthetic redesign of a molecule to affect a significant, measurable change in the ultrafast excited
-
state 
dynamics of a charge
-
tran
sfer chromophore
. Specifically, we suggest that the increase in lifetime of the 
MLCT excited state of compound 
2
 
relative to compound 
1
 
is a direct consequence of the chemical 
modifications made to the ligand framework of the system, modifications that wer
e informed by th
e analysis 
of the coherence featues observed in the excited
-
state dynamics of compound 
1
.
 
 
The origin of the faster component detected in the decay kinetics is unclear at the present time. We 
suspect that it is associated with solvent libra
tional motion in
duced by the large redistribution of charge 
upon formation of the MLCT excited state. These sorts of perturbations have been observed in the ultrafast 
time
-
resolved measurements of compounds like [Ru(bpy)
3
]
2+
,
39,40
 
but are typically not seen in Fe(I
I)
-
based 
analogs
 
most likely because the time scale for decay of the MLCT state in these systems is usually faster 
than the inertial response times of most polar solvents. Finally, nanosecond TA data collected by Sara 
Adelman of the McCusker group shows th
at compound 
2
 
ev
entually returns to the ground
-
state with a time 
constant of 7 
±
 
3 ns (not shown). 
 
To better characterize the origin of the large change in the photophysical properties of compound 
2
 
relative to compound 
1
, additional TA measurements were 
conducted. First
, we performed TA 
Figure
 
4.11.
 
Femtosecond time
-
resolved absorption spectra of [FeCu
2
(L)](PF
6
)
4
 
(
2
) 
following 600
 
nm excitation in acetonitrile soluti
on.
 
 
194
 
measurements on 
2
 
(Figure
 
4.A
.
1) while probing in t
he region of copper based MLCT absorption gleaned 
from the absorption spectra featured in Figure
 
4.3. We observed a bleach signal with small amplitude 2.6
 
ps 
decay to mor
e negative 

A with large amplitude 116
 
cm
-
1
 
oscillations, similar to those observed at longer 
wavelengths. The lack of a large amplitude decay to further negative 

A suggests that the copper atoms are 
not involved in reductive quenching of the iron based M
LCT state as this process would lead to a l
oss of 
copper based MLCT absorption in this region. Additionally, TA experiments were performed on 
2
 
in 
several different solvents (Figures 4.8, 4.9) as well as with a different counteranion (BAr
F
-
) (Figure
 
4.12),
 
in an effort to characterize the influence
 
of interaction with the solute from the solvent bath and 
counteranion. The different solvents were chosen to observe the effect of a change in solvent dielectric (~21 
for acetone and butyronitrile and ~38 for ace
tonitrile) as well as a change in inertial 
response of the solvent. 
These data clearly show the loss of the transient MLCT state absorption over the course of several 
picoseconds. A summary of the kinetics across each solvent and counter anion can be seen 
in Table
 
4.3. 
Figure
 
4.12.
 
Representative datasets of [FeCu
2
(L)](Bar
F
)
4
 
in (a) acetonitrile, (b) acetone, and 
(c) butyronitrile following 600
 
nm excitation while monitored at 700
 
nm.
 
a)
 
b)
 
c)
 
 
195
 
The fast, multi
-
hundred femtosecond component that we tentatively assigned to solvent librational motion 
above are all within error in Table 4.3, although the error bars are 
quite large relative to the magnitude of 
the time constant making i
t difficult to derive definitive conclusions about this fast process. Further work 
to conclusively identify this process is currently being pursued. The MLCT lifetimes across each solvent 
a
nd counter anion are all within error. The insensitivity of the MLC
T decay kinetics to counter anion suggest 
that ion pairing is not inducing the longer MLCT lifetime in compound 
2
, despite the 4+ charge on the 
cation. This is further supported by the iden
tical MLCT lifetimes in different solvents despite the large 
variat
ion in solvent dielectric.
 
The above results rule out factors such as reductive quenching of the MLCT state of compound 
2
 
as well as solute
-
solvent or ion
-
pairing interactions as a major re
ason for the significant extension of the 
MLCT lifetime of 
2
 
relati
ve to 
1
. These findings thus bolster the assertion that the reason for the change in 
photophysical properties of 
2
 
are from the hinderance of the vibrational modes that drive MLCT decay in 
1
. We can further prove this claim by performing Duschinsky displac
ement vector calculations to quantify 
the displacement of the lowest energy 
5
T
2g
 
excited state. Figure
 
4.13 shows the calculated reorganization 
energies for the 
5
T
2g
 

1
A
1g
 
transitions of [
Fe(L)]
2+
 
(Figure
 
4.13a) and [FeCu
2
(L)]
4+
 
(Figure
 
4.13b). We see 
tha
t the reorganization energy of compound 
2
 
(9,320
 
cm
-
1
) is significantly smaller than for compound
 
1
 
(12,100
 
cm
-
1
) which is on par with the reorganization energy calculated in the same manne
r for the 
prototypical [Fe(bpy)
3
]
2+
 
(13,050
 
cm
-
1
) (Figure
 
4.B
.
1). I
t should be noted that while the calculated 
reorganization energies for the reverse 
1
A
1g
 

5
T
2g
 
reaction are significantly different (19,870
 
cm
-
1
 
for 
Solvent
 
Counter Ion
 
Solven
t Librational 
Motion, fs
 
MLCT

LF, ps
 
Aceto
nitrile
 
PF
6
-
 
4
3
0 (50)
 
2.6 (0.1)
 
Butyronitrile
 
PF
6
-
 
350 (150)
 
2.8 (0.4)
 
Acetone
 
PF
6
-
 
320 (80)
 
2.7 (0.4)
 
Acetonitrile
 
BAr
F
-
 
260 (200)
 
2.6 (0.4)
 
Butyronitrile
 
BAr
F
-
 
270 (50)
 
2.6 (0.1)
 
Acetone
 
BAr
F
-
 
210 (90)
 
2.9 (0.2)
 
 
Table
 
4.3.
 
Summary kinetic compone
nts of [FeCu
2
(L)]
2+
 
in various solvents and with PF
6
-
 
and Bar
F
-
 
counteranions. Errors are expressed as standard deviations from multiple trials.
 
 
196
 
compound 
1
, 6,820
 
cm
-
1
 
for compound 
2
, and 18,820
 
cm
-
1
 
for [Fe(bpy)
3
]
2+
), a testament to the different 
motions and force cons
tants inherent to each electronic state, the overarching trend remai
ns the same. We 
thus find that the 
5
T
2g
 
excited state of compound 
2
 
is less displaced than that of the compound 
1
 
control 
molecule or the prototypical [Fe(bpy)
3
]
2+
, confirming that the syn
thetic modification did in fact hinder the 
vibrational modes of nece
ssary for ultrafast MLCT decay in this class of compounds.
 
In summary, with this report we have demonstrated that it is possible to leverage the information 
afforded from excited
-
state coh
erence observations toward the synthetic redesign of a molecular 
chr
omophore. Large
-
amplitude distortions associated with MLCT
-
state excitation of an Fe(II)
-
based tris
-
bipyridine assembly that exhibited dephasing times intermediate between what is expected
 
for pure 
electronic and pure vibrational coherences provided the im
petus to target these degrees of freedom for 
altering the ultrafast excited
-
state dynamics of the compound. Ultimately, the fact that synthetic 
modification of these modes resulted in a 20
-
fold change in the observed kinetics confirmed that these 
specific 
degrees of freedom and the MLCT relaxation dynamics of the chromophore are indeed linked. We 
believe that these results suggest the viability of a fundamentally different approach to the m
anipulation of 
a)
 
b)
 
Figure
 
4.13.
 
Plots of the calculated inner
-
sphere reorganization energy as a function of normal mode 
f
requency for the 
5
T
2g
 

1
A
1g
 
transitions of (a) [Fe(L)]
2+
 
and (b) [FeCu
2
(L)]
4+
. The normal modes 
used are inherent to the 
5
T
2g
 
excited state. Reorganization energies 
were calculated from the nuclear 
displacements of the Duschinsky displacement vectors for 
the corresponding electronic transition.
 
 
197
 
excited
-
state dynamics wherein phenomena observed thr
ough ultrafast spectroscopic techniques can be 
exploited for the creation of chromophores with design
-
specific properties.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
198
 
 
 
 
 
 
 
 
 
 
 
 
 
APPENDICES
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
199
 
Appendix 4.A. Additional Transient Absorption Data on Compound 2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
a)
 
b)
 
Figure
 
4.A.1.
 
(a) TA data of [FeCu
2
(L)](PF
6
)
4
 
(
2
) in acetonitrile while exciting at 600
 
nm 
and monitoring at 396
 
nm (blue) with LPSVD fit to equation 4.2 (red). (b) LPSVD vibrational 
spectrum (red) with cor
responding FFT of the residuals to an 
exponential fit (blue).
 
 
200
 
Appendix 4.B. Reorganization Energy Calculation 
f
or [Fe(bpy)
3
]
2+
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure
 
4.B.1.
 
Calculated reorganization 
energy for the 
5
T
2g
 

1
A
1g
 
transition of 
[Fe(bpy)
3
]
2+
.
 
 
201
 
Appendix 4.C. Additional Normal Mode Assignment for Compound 1
 
 
Since there was a large uncertainty in the experimentally observed frequency of the 127 (5)
 
cm
-
1
 
mode of 
compound
 
1
 
and a high density of vibrational states in that frequency region, an assignment during 
the initial characterization was forgone. Now that the vibrational modes of 
2
 
have been assigned, however, 
we can turn back and assign the 123
 
cm
-
1
 
mode of 
1
. In com
pound 
2
 
the low frequency mode corresponded 
to an Fe
-
N breathing mode in which the tren
-
caps compressed inwards towards the iron
-
center and 
elongated the Fe
-
N bond distance. In compound 
1
 
we find an analogous motion at 113
 
cm
-
1
 
on the high 
spin 
5
T
2
g
 
excited state, in reasonable agreement with the observed 127
 
cm
-
1
 
frequency.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Calc. 113
 
cm
-
1
 
Exp. 12
7
 
cm
-
1
 
Figure
 
4.C.1.
 
DFT calculated vector 
displacement diagram of the 127
 
cm
-
1
 
vibrational mode of [Fe(L)]
2+
.
 
 
202
 
Appendix 4.D. 
Results of Kinetic and Oscillatory Analyses on Each Individual Dataset
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Probe, nm
 
IRF, fs
 
Frequency, cm
-
1
 
(Damping Time, fs)
 
650
 
85
 
 
 
169 (241)
 
227 (1130)
 
700
 
65
 
 
 
 
204 (588)
 
650
 
60
 
 
 
173 (197)
 
 
603
 
45
 
122 (155)
 
 
 
204 (263)
 
594
 
45
 
131 
(388)
 
 
173 (341)
 
248 (1133)
 
600
 
50
 
132 (198)
 
 
 
229 (279)
 
592
 
50
 
127 (250)
 
 
 
207 (349)
 
608
 
50
 
123 (180)
 
 
176 (705)
 
 
650
 
85
 
 
157 (276)
 
 
 
642
 
85
 
 
 
173 (293)
 
 
658
 
85
 
 
154
 
(320)
 
 
219 (1754)
 
Average Frequency, cm
-
1
 
127
 
156
 
173
 
215
 
Standard 
Deviation, cm
-
1
 
5
 
2
 
3
 
12
 
Average Damping Time, fs
 
230
 
300
 
360
 
730
 
Standard Deviation, fs
 
90
 
30
 
230
 
600
 
 
Table
 
4.D.1.
 
Coherent oscillatory components observed for individual TA datase
ts on [Fe(L)](PF
6
)
2
 
in acetonitrile 
solvent following 600
 
nm excitation.
 
 
203
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Probe, nm
 
Solvent
 
IRF, fs
 

1
, fs
 

2
, fs
 
A
1
 
x 10
3
 
A
2
 
x 10
3
 
A
3
 
x 10
3
 
650
 
MeCN
 
85
 
450
 
2530
 
14.0
 
24.6
 
-
12.0
 
650
 
MeCN
 
60
 
500
 
2680
 
12.5
 
21.3
 
-
17.7
 
650
 
MeCN
 
90
 
340
 
2610
 
6.9
 
12.8
 
-
9.1
 
65
0
 
MeCN
 
115
 
425
 
2610
 
6.9
 
11.8
 
-
9.5
 
650
 
MeCN
 
70
 
330
 
2370
 
8.4
 
18.7
 
-
14.2
 
700
 
MeCN
 
65
 
454
 
2660
 
4.5
 
14.9
 
-
1.6
 
700
 
MeCN
 
100
 
450
 
2710
 
4.8
 
10.9
 
0.9
 
700
 
MeCN
 
100
 
470
 
2640
 
4.4
 
10.5
 
0.4
 
700
 
MeCN
 
110
 
424
 
2755
 
3.3
 
7.5
 
0.5
 
700
 
MeCN
 
90
 
410
 
2460
 
6.2
 
11.6
 
4
 
700
 
MeCN
 
90
 
475
 
2850
 
7.4
 
12.7
 
-
0.5
 
400
 
MeCN
 
280
 
-
 
2515
 
-
 
-
 
-
 
396
 
MeCN
 
280
 
-
 
2645
 
-
 
-
 
-
 
650
 
Acetone
 
115
 
250
 
2460
 
11.2
 
14.1
 
-
11.6
 
650
 
Acetone
 
70
 
250
 
2340
 
11.9
 
15.1
 
-
13.4
 
700
 
Acetone
 
90
 
380
 
2780
 
7.0
 
12.0
 
5
 
700
 
Acetone
 
90
 
400
 
3140
 
5.7
 
8.7
 
0.0
 
650
 
BuCN
 
115
 
280
 
2
460
 
11.9
 
12.4
 
-
12.4
 
650
 
BuCN
 
70
 
360
 
3100
 
14.9
 
13.5
 
-
13.1
 
700
 
BuCN
 
90
 
210
 
2390
 
10.3
 
10.7
 
2.2
 
700
 
BuCN
 
90
 
560
 
3070
 
11.2
 
10.7
 
-
0.6
 
MeCN Average (

), fs
 
430
 
2600
 
 
 
 
MeCN Standard Deviation, fs
 
50
 
100
 
 
 
 
Acetone Average, fs
 
320
 
2700
 
 
 
 
Acetone Standard D
eviation, fs
 
80
 
400
 
 
 
 
BuCN Average, fs
 
350
 
2800
 
 
 
 
BuCN Standard Deviation, fs
 
150
 
400
 
 
 
 
 
Table
 
4.D.2.
 
Kinetic components and fit parameters for individual TA datasets on [FeCu
2
(L)](PF
6
)
4
 
in various solvents following 
600
 
nm excitation. A 7
 
ns kineti
c component was locked in to describe the slow ground state recovery process and account for any 
baseline of
fset.
 
 
204
 
Probe, nm
 
E
pu
 
(E
pr
), 

J
 
IRF, fs
 
Solvent
 
Frequency, cm
-
1
 
(Damping Time, fs)
 
700
 
4 (0.2)
 
65
 
MeCN
 
 
117 (744)
 
177 (731)
 
 
700
 
4 (0.4)
 
100
 
MeCN
 
 
117 (613)
 
180 (818)
 
220 (639)
 
700
 
4 (0.4)
 
100
 
MeCN
 
54 (840)
 
118 
(902)
 
182 (645)
 
225 (465)
 
700
 
4 (0.4)
 
110
 
MeCN
 
 
120 (934)
 
184 (671)
 
224 (257)
 
700
 
4 (0.4)
 
90
 
MeCN
 
 
123 (357)
 
 
 
700
 
3.6 (0.4)
 
90
 
MeCN
 
 
112 (419)
 
186 (651)
 
238
a
 
(403)
 
650
 
4 (0.4)
 
85
 
MeCN
 
38
 
(1040)
 
115 (581)
 
181 (421)
 
217 (528)
 
650
 
4 (0.4)
 
60
 
MeCN
 
49 
(562)
 
115 (720)
 
185 (767)
 
223 (743)
 
650
 
4 (0.4)
 
90
 
MeCN
 
62 (456)
 
114 (610)
 
184 (963)
 
220 (683)
 
650
 
4 (0.4)
 
115
 
MeCN
 
 
117 (676)
 
185 (529)
 
221 (369)
 
650
 
4 (0.4)
 
70
 
MeCN
 
 
119 (748)
 
184 (1225)
 
 
610
 
4 (0.4)
 
45
 
MeCN
 
 
117 (1122)
 
183 (1358)
 
220 (850)
 
603
 
4 
(0.4)
 
45
 
MeCN
 
 
115 (625)
 
183 (1058)
 
210 (429)
 
594
 
4 (0.4)
 
45
 
MeCN
 
 
112 (419)
 
181 (640)
 
212 (732)
 
400
 
3.4 (0.2)
 
280
 
MeCN
 
 
114 (696)
 
 
 
396
 
3.3 (0.2)
 
280
 
MeCN
 
 
113 (757)
 
 
 
700
 
4 (0.4)
 
110
 
Acetone
 
 
117 (433)
 
181 (454)
 
223 (508)
 
700
 
4 (0.4)
 
90
 
Acetone
 
 
116 (466)
 
185 (1183)
 
224 (626)
 
700
 
3.6 (0.4)
 
90
 
Acetone
 
 
115 (664)
 
184 (592)
 
223 (699)
 
650
 
4 (0.4)
 
90
 
Acetone
 
 
117 (667)
 
183 (751)
 
220 (518)
 
650
 
4 (0.4)
 
115
 
Acetone
 
 
118 (757)
 
185 (688)
 
224 (331)
 
650
 
4 (0.4)
 
70
 
Acetone
 
 
118 (992)
 
180 
(1523)
 
219 (1794)
b
 
700
 
4 (0.4)
 
110
 
BuCN
 
 
113 (584)
 
184 (385)
 
218 (360)
 
700
 
4 (0.4)
 
90
 
BuCN
 
 
112 (815)
 
180 (686)
 
226 (315)
 
700
 
3.6 (0.4)
 
90
 
BuCN
 
 
113 (701)
 
182 (690)
 
221 (752)
 
650
 
4 (0.4)
 
90
 
BuCN
 
 
114 (821)
 
181 (58
4)
 
217 (509)
 
650
 
4 (0.4)
 
115
 
BuCN
 
 
114 (926)
 
181 (445)
 
218 (376)
 
650
 
4 (0.4)
 
70
 
BuCN
 
 
115 (986)
 
182 (707)
 
216 (362)
 
a. 238 cm
-
1
 
frequency was an outlier and not used in averaging
 
b. 1794 fs dephasing time was an outlier and not used in averaging.
 
Table
 
4.D.3.
 
Coherent oscillatory components observed in individual TA datasets on [FeCu
2
(L)](PF
6
)
4
 
in various solvents following 
600
 
nm excitation. The observed sub
-
100
 
cm
-
1
 
components are not expected to be physically meaningful.
 
 
205
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Solvent
 
IRF, fs
 

1
, fs
 

2
,
 
f
s
 
A
1
 
x 10
3
 
A
2
 
x 10
3
 
A
3
 
x 10
3
 
MeCN
 
103
 
120
 
2300
 
9.4
 
5.4
 
-
0.5
 
MeCN
 
84
 
400
 
2870
 
1.9
 
3.1
 
-
0.2
 
MeCN
 
110
 
230
 
2580
 
3.5
 
6.3
 
-
1.0
 
MeCN
 
110
 
300
 
2650
 
3.4
 
6.7
 
-
0.3
 
Acetone
 
84
 
105
 
2580
 
5.8
 
3.4
 
-
0.01
 
Acetone
 
110
 
280
 
3040
 
4.0
 
5.6
 
-
0.2
 
Acetone
 
110
 
250
 
2960
 
4.7
 
6.6
 
-
0.1
 
BuCN
 
110
 
270
 
2590
 
4.3
 
5.1
 
-
0.3
 
BuCN
 
110
 
270
 
2580
 
1.2
 
1.5
 
-
0.1
 
DCM
 
103
 
170
 
1360
 
12.8
 
4.9
 
-
0.1
 
DCM
 
84
 
210
 
1390
 
4.6
 
2.3
 
-
0.2
 
 
Table
 
4.D.4.
 
Kinetic components and fit parameters for individual TA datasets on [FeCu
2
(L)](BAr
F
)
4
 
in various solvents fol
lowing 
600
 
nm excitation and while monitoring at 700
 
nm. A 7
 
ns kinetic component was locked in and used to describe the slow ground 
state recover
y process and account for any baseline offset.
 
 
206
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Probe, nm
 
IRF, fs
 

1
, fs
 

2
, fs
 
650
 
85
 
120
 
1400
 
650
 
60
 
133
 
1
798
 
650
 
90
 
98
 
1285
 
603
 
45
 
123
 
 
594
 
48
 
99
 
 
592
 
48
 
110
 
 
608
 
48
 
50
 
 
650
 
85
 
134
 
 
642
 
85
 
140
 
 
658
 
85
 
115
 
 
 
Table
 
4.D.5.
 
Results of individual kinetic fits of [Fe(L)](PF
6
)
2
 
TA data following 600
 
nm excitation in acetonitrile 
solvent. 
 
 
207
 
 
 
 
 
 
 
 
 
 
 
 
 
REFERENCES
 
 
 
 
208
 
REFERENCES
 
 
(1)
 
Scholes, G. D.; Fleming, G. R.; Chen, L. X.; Aspuru
-
Guzik, A.; Buchleitner, A.; Coker, D. F.; 
Engel, G. S.; van Grondelle, R.; Ishizaki, A.; Jonas, D. M.; 
et al
. Using Coherence to Enhance 
Function in Chemica
l and Biophysical Systems. 
Nature
 
2017
,
 
543
, 647

656.
 
(2)
 
Zewail, A. H. Femtochemistry: Atomic
-
Scale Dynamics of the Chemical Bond. 
J. Phys. Chem. A
 
2000
, 
104
, 5660

5694.
 
(3)
 
Bishop, M. M.; Roscioli, J. D.; Ghosh, S.; Mueller, J. J.; Shepherd, N. C.; Beck,
 
W. F. Vibrationally 
Coherent Preparation of the
 
Transition State for Photoisomerization of the Cyanine Dye Cy5 in 
Water. 
J. Phys. Chem. B
 
2015
, 
119
 
(23), 6905

6915.
 
(4)
 
Schoenlein, R. W.; Peteanu, L. A.; Mathies, R. A.; Schank, C. V. The First Step in Vis
ion: 
Femtosecond Isomerization of Rhodopsin. 
Sci
ence
 
1991
, 
254
, 412

415.
 
(5)
 
Zhu, L.; Sage, J. T.; Champion, P. M. Observation of Coherent Reaction Dynamics in Heme 
Proteins. 
Science
 
1994
, 
266
, 629

632.
 
(6)
 
Chenu, A.; Scholes, G. D. Coherence in Energy Tra
nsfer and Photosynthesis. 
Annu. Rev. Phys. 
Chem.
 
2015
, 
66
, 69

96.
 
(7)
 
Delor, M.; Archer, S. A.; Keane, T.; Meijer, A. J. H. M.; Sazanovich, I. V.; Greetham, G. M.; 
Towrie, M.; Weinstein, J. A. Directing the Path of Light
-
Induced Electron Transfer at a Mole
cular 
Fork Using Vibrational Excitation. 
Nat. Chem.
 
2017
,
 
9
, 1099

1104
.
 
(8)
 
Zerdane, S.; Wilbraham, L.; Cammarata, M.; Iasco, O.; Riviere, E.; Boillot, M.
-
L.; Ciofini, I.; 
Collet, E. Comparison of Structural Dynamics and Coherence of d
-
d and MLCT Light
-
Ind
uced 
Spin State Trapping. 
Chem. Sci.
 
2017
, 
8
, 4978

4986.
 
(9)
 
Monni, R.
; Auböck, G.; Kinschel, D.; Aziz
-

Conservation of Vibrational Coherence in Ultrafast Electronic Relaxation: The Case of Diplatinum 
Compl
exes in Solution.
 
Chem. Phys. Lett.
 
2017
, 
683
, 112

120.
 
(10)
 
Kim, P.; 
Kelley, M. S.; Chakraborty, A.; Wong, N. L.; Van Duyne, R. P.; Schatz, G. C.; Castellano, 
F. N.; Chen, L. X. Coherent Vibrational Wavepacket Dynamics in Platinum(II) Dimers and Their 
Imp
lications. 
J. Phys. Chem. C
 
2018
, 
122
, 14195

14204.
 
(11)
 
Gueye, M.; Ma
nathunga, M.; Agathangelou, D.; Orozco, Y.; Paolino, M.; Fusi, S.; Haacke, S.; 
Olivucci, M.; Léonard, J. Engineering the Vibrational Coherence of Vision into a Synthetic 
Molecular Device
. 
Nat. Commun.
 
2018
, 
9
, 313.
 
(12)
 
Iwamura, M.; Takeuchi, S.; Tahara, T
. Ultrafast Excited
-
State Dynamics of Copper(I) Complexes. 
Acc. Chem. Res
. 
2015
, 
48
, 782

791.
 
(13)
 
Wenger, O. S. Is Iron the New Ruthenium? 
Chem. Eur. J
. 
2019
, 
25
, 6043

6052.
 
(14)
 
Wenger
, O. S. Photoactive Complexes With Earth
-
Abundant Metals. 
J. Am. Chem.
 
Soc
. 
2018
, 
140
, 
13522

13533.
 
 
209
 
(15)
 
Lemke, H. T.; Kjaer, K. S.; Hartsock, R.; van Driel, T. B.; Chollet, M.; Glownia, J. M.; Song, S.; 
Zhu, D.; Pace, E.; Matar, S. F.; Nielsen, M. M.; Ben
fatto, M.; Gaffney, K. J.; Collet, E.; Cammarata, 
M. Coherent Structur
al Trapping through Wave Packet Dispersion during Photoinduced Spin State 
Switching. 
Nat. Commun.
 
2017
, 
8
, 15342.
 
(16)
 
Auböck, Gerald; Chergui, M. Sub
-
50
-
Fs Photoinduced Spin Crossover i
n [Fe(Bpy)3]2+. 
Nat. 
Chem.
 
2015
, 
7
, 629

633.
 
(17)
 
Monat, J. E.; McCusker, J. K. Femtosecond Excited
-
State Dynamics of an Iron(II) Polypyridyl 
Solar Cell Sensitizer Model. 
J. Am. Chem. Soc.
 
2000
, 
122
, 4092

4097.
 
(18)
 
McCusker, J. K. Electronic Structure in 
the Transition Metal Block and Its Implications for Light 
Harvesting. 
Science
 
2019
,
 
363
, 484

488.
 
(19)
 
L
iu
, L.; Duchanois, T.; Etienne, T.; Monari, A.; Beley, M.; Assfeld, X.; Haacke, S.; G
ros, P. C. A 
New Record Excited State 3MLCT Lifetime for Metalorgan
ic Iron(II) Complexes. 
Phys. Chem. 
Chem. Phys
. 
2016
, 
18
, 12550

12556.
 
(20)
 
Chabera, P.; Kjaer, K. S.; Prakash, O.; Honarfar, A.; Liu, Y.; Fredin, L. A.; Harlang, T. C. B.; Lidin, 
S.; Uhlig,
 
J.; Sundstro, V.; 
et al

x with a 528 ps Metal
-
to
-
 
Ligand Charge
-
Transfer Excited
-
State Lifetime. 
Phys. Chem. Lett.
 
2018
, 
9
, 459
-
463.
 
(21)
 
Steube, J.; Burkhardt, L.; Päpcke, A.; Moll, J.; Zimmer, P.; Schoch, R.; Wö
lper, C.; Heinze, K.; 
Lockbrunner, S.; Bauer, M. Excited
-
State Kine
tics of an Air
-
Stable Cyclometalated Iron(II) 
Complex. 
Chem. Eur. J
. 
2019
, 
25
, 11826

11830.
 
(22)
 
Kjær, K. S.; Kaul, N.; Prakash, O.; Chábera, P.; Rosemann, N. W.; Honarfar, A.; Gordivska, O
.; 
Fredin, L. A.; Bergquist, K.
-
E.; Häggström, L.; Ericsson, T.; 
et
 
al. 
Luminescence and Reactivity of 
a Charge
-
Transfer Excited Iron Complex with Nanosecond Lifetime. 
Science
 
2019
, 
363
, 249

253.
 
(23)
 
Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria
, G. E.; Robb, M. A.; Cheeseman, J. R.; 
Scalmani, G.; Barone, V.; P
etersson, G. A.; Nakatsuji, H.; Li, X.; Caricato, M.; Marenich, A. V.; 
Bloino, J.; Janesko, B. G.; Gomperts, R.; Mennucci, B.; Hratch, D. J. Gaussian 16, Revision C.01. 
2016
.
 
(24)
 
Humphrey,
 
W.; Dalke, A.; Schulten, K. VMD: Visual Molecular Dynamics. 
J. Mol
. Graph. Model
. 
1996
, 
14
, 33

38.
 
(25)
 
Albrecht, H.; Heist, P.; Kleinschmidt, J.; Lap, D. V. Single
-
Shot Measurement of Ultraviolet and 
Visible Femtosecond Pulses Using the Optical Kerr Effe
ct. 
Appl. Opt
. 
1993
, 
32
, 6659

6663.
 
(26)
 
Gütlich, P.; Goodwin, H. A
. Spin Crossover in Transition Metals III; Springer
-
Verlag: Berlin, 
Heidelberg, 
2004
.
 
(27)
 
Purcell, K. F. Pseudorotational Intersystem Crossing in 
d
6
 
Complexes.
 
J. Am. Chem. Soc.
 
1979
, 
101, 5147

5152.
 
(28)
 
Mc
C
usker, J. K.; Rheingold, A. L.; Hendrickson, D.
 
N. Variable
-
Temperature Studies of Laser
-
Initiated 5 T 2 f 1 A 1 Intersystem Crossing in S
pin
-
Crossover Complexes: Empirical Correlations 
between Activation Parameters and Ligand Structure in a Series of Polypyridyl Ferrous Complexes. 
Inorg. Chem
. 
1996
, 
3
5
, 2100

2112.
 
 
210
 
(29)
 
Ashley, D. C.; Jakubikova, E. Ray
-
Dutt and Bailar Twists in 
Fe(II)
-

-
b
ipyridine): Spin 
States, Sterics, and Fe
-
N Bond Strengths. 
Inorg. Chem
. 
2018
, 
57
, 5585

5596.
 
(30)
 
D
e Mendoza, J.; Mesa, E.; Rodriguez
-
Ubis, J.
-
C.; Vázquez, P.;
 
Vögtle, F.; Windscheif, P.
-
M.; 
Rissanen, K.; Lehn, J.
-
M.; Lilienbaum, D.; Zies
sel, R. A New Macrobicyclic Tris
-
Bipyridine 
Ligand and Its Cu(
I
)
2
 
and Ag(I)
3
 
Complexes. 
Angew. Chem. Int. Ed. Engl
. 
1991
, 
30
, 1331

1333.
 
(31)
 
Brown, A. M.; McCusker, C. E.; McCu
sker, J. K. Spectroelectrochemical Identification of Charge
-
Transfer Excited St
ates in Transition Metal
-
Based Polypyridyl Complexes. 
Dalt. Trans
. 
2014
, 
17635

17646.
 
(32)
 
Smeigh, A. L.; Creelman, M.; Mathies, R. A.; McCusker, J. K. Femtosecond Time
-
Resolved
 
Optical and Raman Spectroscopy of Photoinduced Spin Crossover: Temporal Resolu
tion of Low
-
to
-
High Spin Optical Switching. 
J. Am. Chem. Soc
. 
2008
, 
130
, 14105

14107.
 
(33)
 
 
Wang, Q.; Schoenlein, R. W.; Peteanu, L. A.; Mathies, R. A.; Shank, C. V. Vibrational
ly Coherent 
Photochemistry in the Femtosecond Primary Event of Vision 
Science
 
1
994
, 
266
, 422

424.
 
(34)
 
Hamm, P.; Zanni, M. Concepts and Methods of 2D Infrared Spectroscopy; Cambridge University 
Press: Cambridge, 
2011
.
 
(35)
 
Rafiq, S.; Scholes, G. D. Slow In
tramolecular Vibrational Relaxation Leads to Long
-
Lived. 
J. 
Chem. Phys. A
 
2016
,
 
120
, 6792

6799.
 
(36)
 
De Silvestri, S.; Cerullo, G.; Lanzani, G. Coherent Vibrational Dynamics; CRC Press: Boca Raton, 
FL, 
2008
.
 
(37)
 
Reid, P. J
.; Silva, C.; Barbara, P. F. Elec
tronic Coherence, Vibrational Coherence, and Solvent 
Degrees of Freedom in the Femtosecond Spectroscopy of Mixed
-
Valence Metal Dimers in H
2
O 
and D
2
O. 
J. Phys. Chem. 
1995
,
 
99
,
 
2609

2616.
 
(38)
 
Rafiq, S.; Scholes, G. D. From F
undamental Theories to Quantum Co
herences in Electron Transfer. 
J. Am. Chem. Soc
. 
2019
, 
141
, 708

722.
 
(39)
 
Damrauer, N. H.; Cerullo, G.; Yeh, A.; Boussie, T. R.; Shank, C. V.; McCusker, J. K. Femtosecond 
Dynamics of Excited
-
State Evolution in [Ru(
b
py)
3
]
2+
.
 
Science
 
1997
, 
275
, 54

57.
 
(40)
 
Y
eh, A. T.; Shank, C. V.; McCusker, J. K. Ultrafast Electron Localization Dynamics Following 
Photo
-
Induced Charge Transfer. 
Science
 
2000
, 
289
, 935

938.
 
 
 
 
211
 
Chapter 5: A Modular Approach to Controlling Excited State Properties 
of Iron(II) Polypyridyls
 
The work
 
presented in this chapter was a collaborative effort between this author, and McCusker group 
members Karl C. Nielsen, Dr. Christopher R. Tichnell, and Dr. Monica C. Carey. The initial synthesis and 
impetus for the work 
was driven by CRT, with KCN performi
ng all remaining synthesis and ground state 
characterizations. MCC performed an initial preliminary transient absorption experiment to show a 
compound had promising photophysical properties. All transient absorption data
 
shown herein was 
collected and anal
yzed by the author of this dissertation.
 
5.1. Introduction
 
 
Photoinduced electron transfer reactions are of interest for a wide range of applications. A 
significant portion of the work in this regard has focused
 
on ruthenium(II) polypyridyl complexes whic
h 
often have long
-
lived (ns
-

s)
1
 
charge separated excited states, which survive long enough to be coupled to 
other process
es. Some important applications of charge se
parated states, such as solar energy conversion, 
rely on light harvesting, an inherently material intensive process due to the relatively low solar flux on 
Earth. This makes the use of ruthenium, one of the rares
t substances on the planet,
2
 
a poor choice when 
scaling up these 
applications. This has prompted the search f
or alternative first row transition metal 
chromophores with charge separated excited states that are long lasting enough to be coupled to electron 
Iron(II)
 
Ruthen
ium(II)
 
a)
 
b)
 
Figure
 
5.1.
 
Qualitative potential energy surfaces for six
-
coordinate (a) iron(II) and (b) 
ruthenium(II) compounds.
 
 
212
 
transfer processes. While several such systems exist, such as the
 
class of 
d
10
 
copper(I) polypyridyl 
compound
s, perhaps the easiest conceptual alternative to 
d
6
 
ruthenium(II) compounds are the complexes of 
their vastly more earth abundant 
d
6
 
iron(II) congener. Unfortunately, the exceedingly rapid MLCT 
deactivation of ir
on(II) polypyridyl compounds, (often sub
-
200
 
fs)
3
 
precludes their u
se in photo
-
actuated 
electron transfer reactions.
 
 
The origin for these sh
ort MLCT excited state lifetimes in iron
-
based compounds can be traced 
back to the presence of lower energy metal centered excited states that are less energetically accessible in 
t
heir ruthenium counterparts (Figure
 
5.1). Framed differently, the smaller l
igand field splitting inherent to 
3
d
 
metals (a consequence of the shorter radial extension of the 3
d
 
orbitals relative to the 4
d
 
orbitals) is 
responsible for their lackluster perfor
mance. The lack of a radial node in the 3
d
 
orbitals of iron results in a 
contraction of this orbital relative to its 4
d
 
counterpart in ruthenium (Figure
 
5.2). The larger radial extension 
of the 4
d
 
orbita
ls of ruthenium then allow for increased overlap with
 
ligand
-
based orbitals before 
electrostatic repulsion from the 
p
 
and 
s
 
electrons dominates. This in turn results in smaller ligand field 
splitting in 3
d
 
transition metal systems than their 4 and 5
d
 
conge
ners. Several research groups have thus 
sought to rem
edy this problem by utilizing strongly 

-
donating ligands to raise the lower lying ligand field 
excited states of iron above the MLCT manifold. This approach has been rather successful in producing 
iron(
II) compounds with significantly longer lasting MLCT 
excited states.
4

7
 
Figure
 
5.2.
 
Radial distribution function for (a) the n = 3 orbitals of iron(II) and (b) the n = 4 orbitals of 
ruthenium(II)
.
 
a)
 
b)
 
 
213
 
 
This work seeks to illustrate the benefits of a more modular approach to 
ligand design whereby the 
ligand field strength and, by extension, the photophysical properties can be systematically tuned by 
swapping a single phenanthr
oline ligand with stronger
-
field cyanide or dicarbene ligands, allowing for 
semi
-
orthogonal modulation
 
of the ligand field. Figure
 
5.3 shows the series investigated in this work, which 
terminates with an acyclic carbene complex that was first synthesized 4
0 years ago.
8
 
The metal
-
templated 
synthesis avoids the instability inherent to other carbene lig
ands which require 
in situ
 
activation and 
coordination. This opens the door to further tuning of the s
pectral properties of these complexes by synthetic 
modification of the two remaining polypyridyl ligands which are intimately involved in the MLCT
-
based 
f
eatures in the ground state absorption spectrum for each compound.
 
5.2
.
 
Experimental Methods
 
5.2.1
.
 
Sy
nthesis and Characterization
 
 
All synthesis and ground state characterization on compounds 
1
, 
2
, and 
3
 
were performed by 
McCusker group members Karl Niels
en and Dr. Christopher Tichnell according to literature procedures.
8,9
 
 
All electrochemical and spectroelectrochemical e
xperiments were conducted under an argon 
atmosphere in a drybox by Karl Nielsen. Cyclic voltammetry (CV) and differential pulse voltammetry 
(DPV) experiments were performed using an Electrochemical Workst
ation (CH
-
Instruments) potentiostat. 
The three
-
elect
rode setup included a platinum disk working electrode, a silver wire pseudo
-
reference 
electrode, and a platinum wire counter electrode. A saturated solution of tetrabutylammonium perchlorate 
(TBAClO
4
 
from
 
Sigma Aldrich) was used in methanol. Ferrocene was 
used as an internal standard and all 
(1)
 
(2)
 
(3)
 
Figure
 
5.3
.
 
Drawing of the three compounds investigated in this study.
 
 
214
 
potentials are reported against this reference herein. All spectroelectrochemical measurements were also 
conducted under an argon atmosphere in a drybox and performed 
in a manner similar to those described for 
electroch
emical experiments. A Honeycomb spectroelectrochemical cell (Pine Research) was used. Gold 
honeycomb electrodes (which includes working and reference electrodes) were utilized. A silver wire 
pseudo
-
refere
nce electrode in a glass fritted jacket was filled w
ith the methanolic TBAClO
4
 
electrolyte. 
Bulk electrolysis was performed by holding potentials ~100
 
mV positive or negative of the anodic and 
cathodic current maxima for oxidation and reduction respectivel
y. A 5.8
 
Spectrograph (Spectral 
Instruments Incorpor
ated) was used to monitor changes in absorbance. Reductive and oxidative spectra 
from early electrolysis times were added together to simulate MLCT spectral features.
10
 
5.2.2
.
 
Ultrafast Transient Absorption Spectroscopy
 
 
Detail
s of the transient absorption set
-
up and methodology are provided in Chapter 2 of this 
dissertation. Samples were dissolved in HPLC grade methanol such that the absorbance at the center 
wavelength of the 
pump spectrum was between 0.3 and 0.8. A ground stat
e UV
-
Vis spectrum was acquired 
before and after use to check for any degradation of the sample. Additionally, the signal for each compound 
was checked for linearity before data collection. All errors pres
ented for single wavelength measurements 
are from th
e standard deviation of multiple datasets except where explicitly stated.
 
 
Full spectrum data was chirp corrected and fit using global analysis in OPTIMUS, a free, 
downloadable program for analysis of tim
e resolved spectroscopic data.
11
 
Decay associated difference 
spectra (DADS) and evolution associated difference spectra (EADS) were also calculated from OPTIMUS 
fits assuming sequential decay dynamics. Additionally, the chirp corrected data was subjected to singula
r 
value decomposition in MATLAB and the right eigenv
ectors, describing the time
-
dependent nature of the 
data, were subsequently fit with exponential functions to discern the kinetics of the underlying 
photophysical processes. The time constants acquired fr
om fitting the right eigenvectors were implemented 
i
nto the solution to a set of differential equations describing the kinetic model for a particular molecule to 
produce a concentration matrix, 
C
, in which each column contains the concentration of a partic
ular excited 
state as a function of time. Assuming t
he experimental data can be described in terms of a linear 
 
215
 
combination of basis spectra for each chemical component, the data matrix, 
A
, can be written in terms of a 
basis spectrum matrix, 
F
, and the conc
entration matrix as in equation 5.1.
 
 
 
5.1
 
If the product C
T
C is invertible, then F can be calculated according to equation
 
5.2,
12
 
 
 
5.2
 
thus 
affording the EADS for each transient species as columns of the basis spectrum matrix, 
F
. For this 
investigation we employ both the SVD analysis and the global fitting from the OPTIMUS software to 
provide
 
a complimentary view of the kinetics and spectrosco
pic changes as a function of time. The 
MATLAB script used to perform SVD analysis is included in appendix A.
 
 
The full spectrum measurements in this study serve as a convenient way to qualitatively unders
tand 
the spectral changes as a function of time, and
 
the fitting techniques outlined above serve to estimate the 
timescale of the dynamics. These measurements, however, can sometimes suffer from a low signal to noise 
ratio, in part due to the limited inten
sity of the white light continuum probe, and in part
 
due to the lack of 
lock
-
in amplification for this measurement on our laser system. For more precise kinetic measurements, we 
employ a second optical parametric amplifier (OPA) to generate a probe pulse, 
often of higher intensity 
than can be generated in a
 
continuum and employ lock
-
in amplification to increase the signal to noise ratio. 
Coherent oscillations were observed in some datasets at short pump
-
probe delays. The oscillatory features 
in these data s
ets were analyzed the manner outlined in chapter 2 o
f this dissertation. 
 
5.3
.
 
Results and Discussion
 
5.3.1
.
 
Ground
-
State Properties of Compounds 1, 2, and 3
 
 
The ground state absorption spectrum of compounds 
1
, 
2
, and 
3
 
can be seen in Figure
 
5.4, in which 
a broad, intense metal
-
to
-
ligand charge transfer (M
LCT) based absorption band for each compound is 
observed. Replacement of a single phen ligand with the incr
easingly strong field cyanide and carbene 
ligands in compounds 
2
 
and 
3
 
results in consecutive redshifts of the MLCT band, in good agreement with 
the 
MLCT energies calculated from electrochemistry (Table
 
5.1). The redshift of the MLCT absorption 
 
216
 
feature in conjunction with the increasing ligand field strengths of compounds 
2
 
and 
3
 
suggest a decreasing 
energy gap between the MLCT and lowest energy ligan
d field excited state across the series. In what 
follows, we will investigate the resulting effect of these changes to the ligand field strength and MLCT 
energies on the excited state properties.
 
5.3.2
.
 
Exited State Dynamics of 1
 
 
Transient absorption dat
a on compound 
1
 
are illustrated in Figure
 
5.5. By ~500
 
fs following the 
initial photoexcitation, we observe that the TA data on 
1
 
do not resemble the predicted MLCT spectrum as 
calculated using spectroelectrochemical measurements on similar compounds
10
 
as well as this compound 
specifically (Figure
 
B1a). Furthermore, the congruence between the TA data and the inverted ground state 
absorption spectrum suggest that compound 
1
 
is in the 
5
T
2g
 
excited state by 500
 
fs as is typical for
 
many 
iron(II) 
polypyridyl compounds.
13

16
 
SVD analysis was performed on compound 
1
 
(Figure
 
5.5 b and c) 
Compound
 

max,MLCT,
 
nm
 
Fe
2+/3+
,
 
V vs. FC
0/+
 
Phen
-
/0
,
 
V vs. FC
0/+
 
Calculated 
MLCT
, nm
a
 
1
 
508
 
0.648
 
-
1.78
b
 
510
 
2
 
550
 
0.082
 
-
2.16
b
 
553
 
3
 
567
 
0.200
 
-
1.92
b
 
584
 
a
 
Calculated from the difference in oxidation and reduction potential obtaine
d electrochemically.
 
b
 
Transition was irreversible.
 
Ta
ble
 
5.1.
 
Electrochemical and spectroscopic parameters of [Fe(phen)
3
]
2+
 
(
1
), [Fe(phen)
2
(CN)
2
] 
(
2
), and [Fe(phen)
2
(C
4
H
10
N
4
)] 
(
3
) collected by Karl Nielsen of the McCusker group.
 
Figure
 
5.4.
 
Ground state absorption spectra of [Fe(phen)
3
](BF
4
)
2
 
(
1
), [Fe(phen)
2
(CN)
2
] (
2
), and 
[Fe(phen)
2
(C
4
H
10
N
4
)](PF
6
)
2
 
(
3
), collected by Karl Nielsen of t
he McCusker group.
 
 
217
 
alongside a global fit of the data (Figure 5.C1) in an effort to assess the spectral changes that occur and 
their timescales.
 
A 970
 
ps single exponential fit of the first and second right eigenvectors in the SVD 
analysis (Figure
 
5.5b) agree well with the 9
40
 
ps single exponential fit from global analysis (Figure
 
5.C1a). 
Both procedures also produce similar EADS that resemble the
 
inverse of the ground state absorption 
spectrum, as would be expected for a 
5
T
2g
 
excited state population, which often have MLCT b
ands in the 
visible region of significantly lower oscillator strength than the MLCT transitions from the low
-
spin state. 
This
 
kinetic process which results in a return to baseline across the entire spectrum can thus readily be 
Figure
 
5.5.
 
(a) Full spe
ctrum TA data on [Fe(phen)
3
](BF
4
)
2
 
(
1
) in methanol following 
570
 
nm excitation. Irregularities at 570
 
nm can be at
tributed to uncorrected pump scatter. 
(b) First (black) and second (blue) right eigenvectors from SVD of full spectrum dataset 
with 970
 
ps sin
gle exponential fits (red). (c) EADS for the single kinetic component 
obtained from SVD analysis.
 
a)
 
b)
 
c)
 
 
218
 
assigned to 
5
T
2g
 

1
A
1g
 
ground state recovery. A more precise single wavelength measurement of this 
process (Figure
 
5.6) resulted in a 1,250 
± 150
 
ps tim
e constant, in good agreement with the full spectrum 
dataset.
 
 
To better quantify the rate of MLCT deactivation in this system, the kinetics were monitored in a 
region that was predicted by spectroelectrochemical measurements (Figure
 
B1a) and confirmed in 
the 
earliest time trace of the full spectrum dataset to show 
a significant amount of excited state absorption 
(ESA) of the phenanthroline radical anion of the MLCT state (Figure
 
5.7). We observe a 180 
±
 
30
 
fs decay 
that can be assigned to the loss of the M
LCT excited state population. Interestingly, there appear to 
be 115 
±
 
5
 
cm
-
1
 
coherent oscillations superimposed upon the population dynamics which exhibit probe wavelength 
dependent dephasing times ranging from 250 
±
 
50
 
fs when monitored in the red portion
 
of the spectrum 
(Figure
 
5.7) to 670 
±
 
20 fs when monitored i
n the center of the bleach feature (Figure
 
D1). Given that 
similar observations were made on the closely related [Fe(bpy)
3
]
2+
 
compound by Chergui
14
 
and Gaffney
13
 
we also assign these coherent oscillations as vibrational co
herences on the excited electronic stat
e. A more 
complete discussion of the coherent oscillations can be seen in appendix D.
 
Figure
 
5.6.
 
Single wavelen
gth kinetics 
traces collected on [Fe(phen)
3
](BF
4
)
2
 
(
1
) in 
methanol at 520
 
nm following 520
 
nm 
excitation. This particu
lar dataset was fit to a 
1250 
±
 
150
 
ps time constant with the error 
bar representative of this single dataset.
 
Figure
 
5.7.
 
Single wavel
ength kinetics trace 
collected on [Fe(phen)
3
](BF
4
)
2
 
(
1
) in methanol at 
610
 
nm following 520
 
nm excitation. Inset 
include
s a fast Fourier transform of the residuals 
of an exponential fit to the population dynamics 
(blue) overlaid with a vibrational spectrum
 
produced via LPSVD of the TA data (red).
 
 
219
 
 
The results shown here are in good agreement with the work of Tribollet 
et al.
16
 
on [Fe(phen)
3
]
2+
 
in a different solvent
 
(acetonitrile). The findings therein show an ultrafast deactivation of the MLCT state 
to the 
5
T
2g
 
lowest energy excited state with a 220 
±
 
40
 
fs time constant and subsequent ground state recovery 
with a 1.1 
±
 
0.1
 
ns 
time constant. These findings indicate 
that compound 
1
 
behaves as a typical iron(II) 
polypyridyl complex with an ultrafast MLCT lifetime and a ground state recovery lifetime on the order of 
a nanosecond.
 
5.3.3
.
 
Excited State Dynamics of Compound 2
 
Figure
 
5.8.
 
(a) Full spectrum TA data on [Fe(phen
)
2
(CN)
2
] (
2
) in methanol fol
lowing 570
 
nm 
excitation. (b) First (black), second (green), and third (blue) right eigen vectors from SVD of 
2
 
and a biexponential fit (red) of the first right eigenvector to 6.6 and 315
 
ps kinetic components. 
(c) Simulated exc
ited state populations using
 
the time constants acquired from the right 
eigenvector fits and assuming sequential decay dynamics as shown in equation
 
5.3. (d) EADS 
calculated for the two kinetic components using equation 5.2. The red (F1) and blue (F2) spec
tra 
correspond to the fast a
nd slowly decaying species respectively.
 
a)
 
b)
 
d)
 
c)
 
 
220
 
 
Full spectrum TA data on compound 
2
 
can be se
en in Figure
 
5.8, in which we observe a slight 
decrease in the magnitude of the ground state bleach over the first few picoseconds followed by multi
-
hundred picosecond ground state recovery. SVD analysis shows that the first right eigenvector is 
by far the
 
most significant in describing the spectral evolution of the system (Figure
 
5.8b), suggesting that the spectral 
evolution can largely be described with a single basis spectrum. A fit of this eigenvector to a biexponential 
decay produces 6.6
 
ps a
nd 310
 
ps 
time constants in good agreement with a global fit of the data to two 
5.5
 
ps and 290
 
ps kinetic components using OPTIMUS fitting software (Figure
 
5.C2). Given that the TA 
data does not match the calculated MLCT spectrum (Figure
 
5.B
.
1b) by 500
 
fs,
 
we can ag
ain say that the 
MLCT state is short lived and decays by this time.
 
We can next benefit from the increased signal to noise ratio and ti
me resolution of single 
wavelength measurements to refine the above time constants and determine the lifetime of the sho
rt lived 
MLCT excited state. Single wavelength measurements while monitoring in the bleach region again reveal 
biphasic kinetics with gr
ound state recovery occurring with a 355 
±
 
7
 
ps time constant. The faster kinetic 
component exhibits a probe wavelength 
dependent time constant ranging from 2.6 
±
 
0.8
 
ps when monitored 
at the maximum of the bleach (Figure
 
5.9a) to 11 
±
 
2
 
ps when monitored 
towards the red edge of the bleach 
at 590
 
nm (Figure
 
5.9b). While not an exact match, these time constants are in good a
greement with the full 
a)
 
b)
 
Figure
 
5.9.
 
(a) Representative TA data of [Fe(phen)
2
(CN
)
2
] (
2
) in methanol following 590
 
nm 
excitation and monitored at 550
 
nm in the center of the ground state ble
ach. (b) Representative 
T
A data on 
2
 
following 600
 
nm excitation and monitored at 590
 
nm, the red edge of the ground 
state bleach. The two data sets show different kinetics at early times.
 
 
221
 
spectrum measurements discussed above. An additional 250 
±
 
90
 
fs decay component can be observed at 
very early times when monitoring at 620
 
nm (
Figure
 
5.10). By inspection of the simulated MLCT spectrum 
produced via spectroelectroch
emical measurements (Figure
 
5.B
.
1b) we see that the absorption feature 
above 600
 
nm from the reduced phenanthroline ligand can serve as a convenient optical marker for 
the 
MLCT state population, similar to compound 
1
. We thus assign this rapid 250
 
±
 
90
 
fs 
component to decay 
of the MLCT state.
 
We next turn to the slowest 355
 
±
 
6
 
ps kinetic component which must result in the formation of the 
ground electronic state. We kno
w that the vast majority of low
-
spin Fe(II) polypyridyl compounds undergo 
rapid initial 
decay from the initially populated MLCT state to the 
5
T
2g
 
metal centered excited state and 
subsequently undergo 
5
T
2g
 

1
A
1g
 
ISC to recover the ground electronic state with time constants on the 
order of nanoseconds.
17
 
According to the energy gap law for strongly coupled systems (i.e. displaced 
potential s
urfaces
),
18,19
 
we would predict that the r
ate of 
the nonradiative transition between the strongly 
displaced 
5
T
2g
 
excited state and the 
1
A
1g
 
ground state will increase as the energy gap between these two 
states and, by extension, the ligand field strength of the molecule increases. Thus, considerin
g the i
ncreased 
ligand field strength of the cyanide ligands of compound 
2
 
relative to the phenanthroline ligand of 
compound 
1
, a faster rate for ground state recovery from the 
5
T
2g
 
state would be entirely consistent with 
Figure
 
5.10.
 
Single wavelength kinetics trace of [Fe(phen)
2
(CN)
2
] (
2
) monitored at 620
 
nm f
ollowing 
550
 
nm excitation. Oscillatory features near the coherent artifact can be attributed to vibrational 
coherences.
 
 
222
 
our observations. Furthermore, rec
ent var
iable temperature ultrafast TA measurements have shown that the 
barrierless rate for 
5
T
2g
 

1
A
1g
 
ISC in several tris bidentate iron(II) polypyridyl compounds is ~230
 
ps.
20
 
Given that the rate for ground state recovery of compound 
2
 
is between the 1
250
 
ps time constant of 
1
 
and 
the ~230
 
ps barrierless time constants determined for several other tris
-
bidentate iron(II) polypyridyl 
analogues as predicted by Jortner
 
and Marcus theory, we assign the 355
 
ps process of compound 
2
 
as the 

S = 2, 
5
T
2g
 

1
A
1g
. We can further eliminate the alternative 

S = 1, 
3
T
1,2g
 

1
A
1g
 
assignment as the change 
in 

S would be expected to lead to an even more dramatic rate increase for gr
ound state recovery relative 
to compound 
1
. Thus far, iron(II) compounds that have been i
dentified to decay via a triplet metal centered 
(MC) state, do so with time constants on the order of ~10
 
ps,
6,21
 
much faster than is observed for compoun
d 
2
 
in this study.
 
 
Finally, we consider the remaining 2.6
 
ps kinetic component. Given our previously 
discussed assignments of 250 
±
 
90
 

 
ps 
5
T
2g
 

1
A
1g
 
ISC, we can narrow the 
possible kinetic assignments to a change in electronic state (i.e. 
3
T
1,2g
 

5
T
2g
) or vibrational relaxation 
within the 
5
T
2g
 
excited state. We first address the possibility of a change in
 
electronic state by consulting 
the full spectrum TA data and DADS in Figure
 
5.C2b. Considering the (
t
2g
)
5
(
e
g
*
) and (
t
2g
)
4
(
e
g
*
)
2
 
electron 
configurations of the 
3
T
1,2g
 
and 
5
T
2g
 
states respectively, we would expect a significant increase in the metal 
ligand 
bond length during a 
3
T
1,2g
 

5
T
2g
 
electronic transition. This increase in the metal ligand bond length 
should lead to a significant decrease in the mixing between the 
t
2g
 
and ligand 

 
symmetry orbitals and hence 
a reduction of the oscillator strength for 
a high intensity charge tra
nsfer type transition. The rising feature 
observed for compound 
2
 
does not match this prediction and suggests that the 2.6
 
ps is not due to a change 
in electronic state. Alternatively, vibrational relaxation commonly occurs at th
is timescale and a 
vibratio
nally relaxed 
5
T
2g
 
state population may be expected to exhibit increased oscillator strength for a 
MLCT type transition for similar reasons as stated above. We thus assign the 2.6
 
ps process to vibrational 
relaxation on the 
5
T
2g
 
state. Each of these assign
ments also agree well with what has been observed for the 
closely related [Fe(bpy)
2
(CN)
2
] complex by Gaffney and coworkers.
22
 
 
223
 
With a kinetic model in hand, we can no
w further analyze the full spe
ctrum TA data presented in 
Figure
 
5.8. By application of the kinetic model described above, we can create a population matrix 
describing the excited state populations as a function of time, which, when applied to the full spec
trum TA 
data, will allow us to
 
extract out the basis spectra or EADS for each chemical species. We note, however, 
that since the ultrafast MLCT deactivation was too fast to be observed in the full spectrum dataset we must 
omit that kinetic component from t
he analysis. The kinetics can 
thus be described according to the set of 
differential equations in equation
 
5.3 which describe the sequential decay of the hot and cool 
5
T
2g
 
state 
populations.
 
 
 
5.3a
 
 
 
5.3b
 
It should be noted that due t
o the disparate timescales of vibrational relaxation and ground state recovery 
in this system that a parallel decay pathway would yield similar results.  The solutions to these differential 
equations are plotted in Figure 5.8c
 
where the rapidly decaying re
d trace represents the population of the 
vibrationally excited 
5
T
2g
 
state and the blue trace represents the vibrationally cool 
5
T
2g
 
population that 
eventually converts to the ground state. Using these concentrations and the TA
 
data, the basis spectra were 
calculated as outlined in equation
 
5.2. The resulting basis spectra for the hot and cold 
5
T
2g
 
populations are 
shown as the F1 (red) and F2 (blue) spectra in Figure
 
5.8d. 
 
5.3.4
.
 
Excited State Dynamics of Compound 3
 
 
Full spect
rum TA data on
 
compound 
3
 
are 
illustrated in Figure
 
5.11. Initial inspection of 
Figure
 
5.11a shows markedly different features from compounds 
1
 
and 
2
, most notably, significant excited 
state absorption above 620
 
nm as well as decay of most of the transient
 
signal by 20
 
ps. Fits of the 
first two 
right eigenvectors from SVD analysis (Figure
 
5.11b) show three kinetic components at 0.15
 
ps, 1.2
 
ps, and 
5.3
 
ps, in qualitative agreement to the global fit shown in Figure
 
5.C3. The TA data clearly show a decay 
to l
ess positive 

A in the red
-
mos
t portion of the spectrum with a concomitant rise in the bleach region and 
 
224
 
a blueshifting of the intercept with the x
-
axis around 650
 
nm during the initial rapid process. The 
intermediate proc
ess then results in similar features but with lower amplitude c
hanges. Finally, the slowest 
kinetic component results in a return of the signal to baseline and ground state recovery of 
3
. These 
observations are also reflected in the EADS resulting from the
 
global fit shown in Figure
 
5.C3c (which 
assume sequential deca
y dynamics). It is interesting to note that the transient spectrum resembles the 
calculated MLCT spectrum from spectroelectrochemical measurements (Figure
 
5.B
.
1c) at all delay times. 
a)
 
b)
 
d)
 
c)
 
Figure
 
5.11.
 
(a) Chirp corrected full spectrum TA data on [Fe(phen)
2
(C
4
H
10
N
4
)](PF
6
)
2
 
(
3
) following 
600
 
nm excitation i
n methanol. The region between 590 and 610
 
nm has been removed and interpolated 
to remove the contribution from excess pump scatter. (b) First four right eigenvectors from SVD of TA 
data of compound 
3
 
with triexponential fits
 
(red) to 0.15, 1.2, and 5.3
 
ps
 
time constants. An inset is 
included to highlight the dynamics over the first 10
 
ps. (c) Simulated excited state populations using the 
time constants acquired from the right eigenvector fits and assuming sequential decay dyn
amics as 
shown in equation
 
5.4.
 
(d) EADS calculated from the time constants measured from SVD analysis 
assuming sequential decay dynamics.
 
 
225
 
To obtain more accurate
 
time constants, single wavelength measurements w
ere also conducted 
(Figure
 
5.12) while monitoring the ESA feature at 650
 
nm and ground state bleach at 550
 
nm. These data 
reveal triexponential decay with 260 
±
 
40
 
fs, 3.4 
±
 
1.2
 
ps, and 7.4 
±
 
0.9
 
ps kinetic components in relatively 
good agreement with the 
SVD results.
 
 
As mentioned above, the changes to the TA spectrum during the initial ultrafast process can be 
described as a decay in the magnitude of the signal with a slight blue shifting of
 
the ESA feature above 
600
 
nm. Interestingly, several photophysic
al studies on low
-
spin six
-
coordinate iron(II) carbene 
compounds exhibit kinetic components with time constants on the same order of magnitude. These have 
been assigned to several different p
rocesses from 
1

3
MLCT ISC,
21
 
to localization of the MLCT 
excited state.
23
 
It is tempting to add solvent reorganization to the list of plausible possibilities as well. Given
 
the presence of substantial ESA above 600
 
nm following this init
ial fast process and the agreement between 
the intermediate and late time transient spectra (or EADS1 and 2 from Figure
 
5.C3c) with the 
spectroelectrochemically approximated MLCT spectrum (Fi
gure
 
5.B
.
1c), we can say that this initial rapid 
process does not
 
represent MLCT deactivation, thus marking a substantial change in the excited state 
dynamics relative to compounds 
1
 
and 
2
 
which behave largely as typical low
-
spin iron(II) polypyridyl 
compl
exes. It is more difficult to discern between the other possible 
assignments for this process. 
a)
 
b)
 
Figure
 
5.12.
 
Single wavelength kinetics traces on [Fe(phen)
2
(C
4
H
10
N
4
)](PF
6
)
2
 
(
3
) in methanol monitored 
at (a) 650
 
nm abd (b) 550
 
nm followi
ng 600
 
nm excitation. Insets show first 5
 
ps and highlight observed 
coherent oscillations.
 
 
226
 
Considering that 
1

3
MLCT ISC often occurs on the order of ~30
 
fs in divalent transition metal 
polypyridyl compounds
24,25
 
and that the IS
C processes that have been assigned for carbene complexes have 
al
so been sub
-
100
 
fs, we cautiously rule out ISC as a possible origin for the 260 
±
 
40
 
fs kinetic component. 
The remaining excited state localization and solvent librational motion options both
 
seem viable. The large 
redistribution of charge upon formation o
f the MLCT excited state would be expected to induce significant 
reorganization of the solvent which occurs at similar rates.
26,27
 
However, without further evidence, we are 
presently unable to definit
ively assign either of these two possible assignments and suggest
 
both as plausible 
possibilities.
 
 
The intermediate 3.4
 
ps kinetic component is a bit more straightforward to assign. Inspection of the 
TA data in Figures
 
5.11a and 5.12 as well as the DADS i
n Figure
 
5.C3b show that this kinetic component 
is responsible fo
r a slight rise to less negative 

A in the region of the ground state bleach, with smaller 
amplitude decay in the ESA region above 600
 
nm. Considering that the transient spectrum resembles th
e 
predicted MLCT spectrum even at large pump
-
probe delay times, w
e find it unlikely that this process 
involves loss of an MLCT excited state. Alternatively, vibrational relaxation, which often occurs at similar 
timescales in analogous iron
-
carbene complexe
s,
5,23
 
seems more reasonable. In fact, many of the iron 
carbene compounds for which similar timescale vibrational relaxation occur in the MLCT excited state 
show similar spectra
l distortions. We thus also assign this 3.4
 
ps kinetic component 
to vibrational relaxation 
on an MLCT excited state.
 
 
Finally, the third, 7.4 
±
 
0.9
 
ps kinetic component is responsible for the return of the transient signal 
to baseline and, by extension, th
e deactivation of the MLCT excited state and recovery of the grou
nd 
electronic state. This represents a >25
-
fold increase in the MLCT lifetime relative to compounds
 
1
 
and 
2
, 
simply by swapping out a single phenanthroline ligand with an acyclic dicarbene. I
t is interesting to note 
the lack of a long lived ground state bl
each as is typical of many iron(II) polypyridyl compounds including 
1
 
and 
2
 
which decay to the ground state via the 
5
T
2g
 
excited state on timescales on the order of several 
hundreds of picose
conds to nanoseconds.
17
 
We consider it unlikely that compound 
3
 
undergoes 7.4
 
ps 

5
T
2g
 
ISC followed by subsequent 
5
T
2g
 

1
A
1g
 
ISC on a significantly shorter timescale. 
 
227
 
Furthermore, it 

1
A
1g
 
ISC occurs on such a s
hort timescale as well, 
considering many isoelectronic ruthenium(II) polypyridyl compounds require many nanoseconds to 
microseconds to depopulate their MLCT excited states and often proceed via an intermediate triplet metal 
centered 
(MC) excited state.
3
 
It seems more likely that the ground state in t
his system is also reco
vered via 
an intermediate MC excited state. This leaves a 
3
MC excited state as the only remaining alternative. Thus, 
since the excited state absorption and bleach decay with the same time constant, we propose a 7.4
 
ps MLCT 

3
MC conv
ersion followed by a fa
ster, undetected 
3

1
A
1g
 
ISC to recover the ground electronic state. 
 
 
We next apply our kinetic model to the full spectrum TA data in a manner similar to what was done 
for compound 
2
. According to the above discussion, we have th
at compound 
3
 
has an in
itial rapid 
localization of the excited state or solvent inertial response followed by loss of a vibrationally excited 
MLCT population largely due to vibrational relaxation, and finally loss of the vibrationally cool MLCT 
state. The 
excited state populatio
ns can thus be described by the set of differential equations shown in 
equation 5.4.
 
 
 
5.4a
 
 
 
5.4b
 
 
 
5.4c
 
We note two important points. First, since the 
3
MC intermediate decayed 
faster than it was populated and 
because it was not observed in the full spectrum data, then we do not need to include it in the model 
presented in equation 5.4. Second, we model the d
ynamics as sequential here, i.e. that every molecule must 
undergo each d
ecay process and cannot bypass them. For example, a molecule must vibrationally relax 
prior to conversion to a 
3
MC state in the above model. An alternative parallel decay kinetic model
 
can be 
made with minor changes, however due to the significantly differ
ent timescales for each process, the results 
are largely identical. The solution set to equation 5.4 is plotted in Figure
 
5.11c with the basis spectrum for 
 
228
 
each chemical species presen
ted in Figure
 
5.11d. The resulting basis spectra clearly show the excite
d state 
evolution and unequivocally show that the last observed transient species is an MLCT excited state.
 
5.4
.
 
Concluding Thoughts
 
 
 
The photophysical properties of a series of low
-
spin six
-
coordinate iron(II) polypyridyl complexes 
have been investigate
d. By swapping out a single phenanthroline ligand with the stronger field cyanide or 
dicarbene ligands the photophysical properties have
 
been dramatically altered as summarized in Table
 
5.2. 
For the intermediate ligand field strength of compound 
2
, it was 
shown that the overall decay pathway was 
similar to the control molecule, compound 
1
, however the rate of ground state recovery has been
 
increased 
by more than a factor of two. The even greater ligand field strength of compound 
3
 
was illustrated to not 
onl
y exhibit a > 25
-
fold increase in the MLCT lifetime relative to compounds 
1
 
and 
2
, but also to undergo 
a fundamentally different excited
 
state decay pathway. This work illustrates a modular approach to tunning 
the photophysical properties of iron(II) polyp
yridyl complexes that not only is capable of demonstrating 
long enough MLCT lifetimes to be viable for electron transfer reactions, but 
also to allow for additional 
synthetic modifications of the polypyridyl ligand to tune, MLCT energetics, electron locali
zation, and 
ability to bind to a semiconductor for interfacial electron transfer.
 
 
 
Compound
 

Undefined
, fs
 

vib
, ps
 

MLCT
, fs
 

5T2g
, ps
 
1
 
 
 
180 
±
 
30
 
1250
a
 
2
 
 
2.6 
±
 
0.8
 
250 
±
 
90
 
355 
±
 
7
 
3
 
260 
±
 
40
 
3.4 
±
 
1.2
 
7400 
±
 
900
 
 
a
 
Value was the result of a single measurement and does not have an error bar
 
Table
 
5.2.
 
Summary of the kinetic components of [Fe(phen)
3
](BF
4
)
2
 
(
1
), [Fe(phen)
2
(CN)
2
] (
2
), 
and [Fe(phen)
2
(C
4
H
10
N
4
)](PF
6
)
2
 
(
3
) in methanol solution.
 
 
229
 
 
 
 
 
 
 
 
 
 
 
 
 
APPENDICES
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
230
 
Appendix 5.A. Singular Value Decomposition MATLAB Script 
f
or Full Spectrum TA Analysis
 
%This matlab script was written by Bryan Paulus o
f the McCusker group at
 
%Michigan State University.
 
clear 
all
 
clc
 
[filename, pathname]=uigetfile;
 
s=strcat(pathname,file
name);
 
FullSpecData=xlsread(s);
 
cut=50;
%number of time points to cut***********
 
r=3;
%number of singular vectors to keep for reconstructi
on of Data Matrix
 
 
 
 
 
 
 
%Grabs Full spectrum data as an excell file where the first column is the
 
%wavelength wave (with
 
a zero place holder in the 1,1 position), the top row is
 
%the time trace (again, with a zero place holder in the 1,1 position)
 
%and TA 
data everywhere else (ie starting at cell 2,2).
 
t=FullSpecData(1,:);
 
t=t(2:length(t));
%time array
 
%Grabs the first row f
rom the uploaded excel file and makes into an
 
%array called "t" then removed the first element of the array so that only
 
%time poins are
 
present
 
w=FullSpecData(:,1);
 
w=w(2:length(w));
 
wt=w';
%transpose of wavelength vector
 
%Grabs the first column of the 
uploaded excel file and makes it into a
 
%wavelength array similar to how the time array was made above
 
A=FullSpecData;
 
A(1,:)=[]; 
 
A(:,1)=[
]; 
%Creates a matrix of the delta A data points
 
A(A==inf) = 0; 
%turns any infinite values to 0
 
A(isnan(A)) = 0; 
%turns any NAN values to 0
 
 
 
 
 
%Next we remove any data corresponding to time points we don't wish to
 
if
 
cut==0
 
    
tcut=t;
 
    
Acut=A;
 
else
 
   
 
tcut=t(cut:length(t));
 
Acut=A(:,cut:end);
 
end
 
 
 
[U S V] = svd(Acut); 
%perfoms SVD
 
A_svd = U*S*
V'; 
%recombines the decomposed vectors
 
figure(1)
 
surf(tcut,w,Acut,
'EdgeColor'
,
'none'
);
 
title(
'Data'
)
 
xlabel(
'Time (ps)'
)
 
ylabel(
'Wavelength (nm)'
)
 
zlabel(
'Change
 
in Absorbance'
)
 
 
231
 
figure(2)
 
plot(tcut,
-
V(:,1),tcut,V(:,2),tcut,
-
V(:,3),tcut,V(:,4),tcut,
-
V(:,5))
 
title(
'Vectors of Data Weight Matrix'
)
 
xlabel(
'Time (ps)'
)
 
ylabel(
'Data Weight'
)
 
legend(
'v1'
,
'v2'
,
'v3'
,
'v4'
,
'v5'
)
 
refline(0,0)
%Adds zero line
 
 
 
figure(3)
 
plot(w
,
-
U(:,1),w,U(:,2),w,U(:,3),w,U(:,4),w,
-
U(:,5))
 
title(
'Basis Spectra'
)
 
xlabel(
'Wavelength (nm)'
)
 
ylabel(
'Change in Absorbance (a.u.)'
)
 
legend(
'u1'
,
'u2'
,
'u3'
,
'u4'
,
'u5'
)
 
refline(0,0)
% Adds a zero line
 
 
 
N=[1:1:length(tcut)];
 
figure(4)
 
scatter(N,diag(S))
 
title
(
'Scree Plot'
)
 
xlabel(
'Component Number'
)
 
ylabel(
'Significance'
)
 
xlim([0 10])
 
hold
 
plot(N,diag(S))
 
hold 
off
 
 
 
figure(5)
 
scatter(N,log(diag(S)))
 
title(
'Scree Plot'
)
 
xlabel(
'Component Number'
)
 
ylabel(
'log(Significance)'
)
 
xlim([0 10])
 
hold
 
plot(N,log(diag(S))
)
 
hold 
off
 
%******************************************************************
 
%If you want to reconstruct the data from a certain number of the most
 
%significant components then use the following where "r" represents the
 
%number of sign
ificant components.
 
Replace "r" with a number. "r" can be
 
%found at the begining of this script.
 
 
 
Arec = U(:,1:r)*S(1:r,1:r)*V(:,1:r)'; 
%reconstruction of data set from significant vectors
 
Usmall=U(:,1:r);
 
Ssmall=S(1:r,1:r);
 
Vsmall=V(:,1:r);
 
figure(6)
 
sur
f(tcut,w,Arec,
'Edge
Color'
,
'none'
);
 
title(
'Reconstructed Data'
)
 
xlabel(
'Time (ps)'
)
 
ylabel(
'Wavelength (nm)'
)
 
 
232
 
zlabel(
'Change in Absorbance'
)
 
 
 
Ares=zeros(size(Acut));
 
Ares=Acut
-
Arec;
%Residual of Data matrix and reconstructed data matrix
 
figure(7)
 
surf(tcut,w,Ares,
'Edgecolor'
,
'none'
);
 
title(
'Residual'
)
 
xlabel(
'Time (ps)'
)
 
ylabel(
'Wavelength (nm)'
)
 
zlabel(
'Change in Absorbance'
)
 
figure(8)
 
plot(tcut,Ares(150,:))
 
title(
'Residual At Single Wavelength'
)
 
xlabel(
'Time (ps)'
)
 
ylabel(
'Residual'
)
 
%Change the number in the above parenthes
es to plot the data at a different
 
%wavelength pixel.
 
figure(9)
 
plot(w,Ares(:,10))
 
xlabel(
'Wavelength (nm)'
)
 
ylabel(
'Residual'
)
 
title(
'Residual At a Single Time'
)
 
%Change the number in the above parentheses to plot 
the data at a different
 
%time step.
 
SVT=S
*transpose(V); 
%Right singular vectors weighted by singular values.
 
%The rows of SVT correspond to
 
 
 
 
 
figure(10)
 
plot(tcut,SVT(1,:),tcut,SVT(2,:))
 
xlabel(
'Time (ps)'
)
 
ylabel(
'SV^T'
)
 
 
 
 
233
 
Appendix 5.B. Simulated MLCT Spectra
 
 
 
 
 
 
 
 
 
 
 
Figure
 
5.B.1.
 
Simulated M
LCT spectrum of (a) [Fe(phen)
3
]
2+
 
(
1
), (b) [Fe(phen)
2
(CN)
2
] (
2
), and 
(c) [Fe(phen)
2
(C
4
H
10
N
4
)]
2+
 
(
3
) in methanol, predicted by summing the oxidative and reductive 
spectra.
 
a)
 
b)
 
c)
 
 
234
 
Appendix 5.
C. Global Fi
ts of Full Spectrum TA Data
 
 
 
 
 
 
 
 
 
 
 
 
a)
 
b)
 
Figu
re
 
5.C.1.
 
(a) Chirp corrected global fit using OPTIMUS fitti
ng software on full spectrum 
data of [Fe(phen)
3
]
2+
 
(
1
) to a single 940
 
ps exponential kinetic component. (b) DADS from 
global fit of 
1
 
full spectrum data. Since there was only a single kinetic component, t
he DADS is 
equivalent to the EADS.
 
 
235
 
 
 
 
 
 
 
 
 
 
a)
 
b)
 
c)
 
Fig
ure
 
5.C.2.
 
(a) Chirp corrected global fit using OPTIMUS fitting software on full spectrum 
data of [Fe(phen)
2
(CN)
2
] (
2
) to two 5.5
 
ps and 290
 
ps exponential kinetic components. (b) 
DADS from the global fit. (c) EADS from the global fit assuming sequential d
ecay dynam
ics.
 
 
236
 
 
 
 
 
 
 
 
 
a)
 
b)
 
c)
 
Figure
 
5.C.3.
 
(a) Chirp corrected global fit using OPTIMUS fitting software on full spectrum 
data of [Fe(phen)
2
(C
4
H
10
N
4
)](PF
6
)
2
 
(
3
) to three 0.12, 0.66, and 4.7
 
ps exponential kinetic 
components. (b) DADS from the 
global fit. (c
) EADS from the global fit assuming sequential 
decay dynamics.
 
 
237
 
Appendix 5.D. Additional TA Data with Coherent Oscillations
 
 
Several of the compounds investigated in this study exhibited coherent osci
llations in the transient 
absorption data. Example datasets 
illustrating the coherent oscillations of compound
 
1
 
can be seen in 
Figure
 
5.D1 in which both FFT and LPSVD analyses are displayed side by side. A summary of the 
frequencies and dephasing times ob
served for 
1
 
for two different probe wavelength regions can 
be seen in 
Table
 
B1. Interestingly, a probe wavelength dependence was observed in the dephasing times of these 
Figure
 
5.D.1.
 
(a) Early time TA data (blue) on [Fe(phen)
3
](BF
4
)
2
 
(
1
) in methanol monitored at 
610
 
nm following 520
 
nm excitation. An exponential fit (red) and a 
LPSVD fit (black) are 
superimpo
sed on the data. (b) FFT of the residual of the exponential fit of the 610
 
nm TA data 
(blue) and LPSVD vibrational spectrum (red). (c) Early time TA data (blue) of compound (
1
) 
monitored in the bleach region at 520
 
nm with ex
ponential (red) and LPSVD (blac
k) fits. (d) 
Corresponding FFT (blue) and LPSVD vibrational spectra.
 
a)
 
c)
 
b)
 
d)
 
 
238
 
coherently active vibrational modes. Analogous results on a related compound, [Fe(bpy
)
3
]
2+
, collected by 
Chergui and coworkers
14
 
had shown similar vibrational frequenc
ies (127
 
cm
-
1
 
and 157
 
cm
-
1
) assigned as 
vibrational coherences in the 
5
T
2g
 
excited state and also exhibited probe wavelength dependent dephasing 
times with the dephasing times observed in the red
-
most portion of the measured spectrum exhibiting the 
shortes
t dephasing times. These observations were rationalized in terms of a hot 
5
T
2g
 
state population with 
an initially small 
5
T
2g
 

5
MLCT energy gap rapidly falling out of resonance as the wave packet moved 
away from its initial position. Given the photophysical and structural similarities between compound 
1
 
and 
[Fe(bpy)
3
]
2+
 
it is li
kely that the origins of the coherences and their 
dephasing times are similar in both 
complexes.
 
 
 
Coherent oscillations were also observed in the TA data of compound 
3
. Coherences were observed 
in both a spectral region dominated by excited state absorption at 650
 
nm and at 550
 
nm, a region exhibiting 
p
rominent ground state bleach. Additionally, analog
ous datasets were collected in methanol (Figure
 
D2) 
and acetonitrile (Figure
 
5.D3) to ascertain the influence of the solvation shell on the observed vibrational 
modes and dephasing times. The results are su
mmarized in Table 5.D2. We see that aside from the
 
additional 225
 
cm
-
1
 
oscillatory component observed in the methanolic sample, the frequencies and their 
dephasing times are similar in both solvents, suggesting that the origin for the dephasing process is 
largely 
intramolecular. Given that these signals a
re observed at 650
 
nm, a region that is dominated by excited state 
absorption, it is likely that these coherent vibrational modes are generated in the MLCT excited state, as 
opposed to the ground electronic
 
state which can be produced via impulsively stimu
lated Raman scattering. 
 
 
Probe Region
 
Frequency, cm
-
1
 
(Damping 
T
ime, fs)
 
520
 
nm
 
117 ± 1 (670 ± 20)
 
156 
±
 
7 (1300 
±
 
100)
 
610 

 
630 nm
 
115 
±
 
5 (250 
±
 
50)
 

 
 
Table
 
5.D.1
.
 
Summary of coherent osc
illations of [Fe(phen)
3
](BF
4
)
2
 
(
1
) observed in TA experiments 
in methanol following 520
 
nm excitation. Experiments were conducted similarly in both wavelength 
regions, except that in the 610
-
630
 
nm region, several separate 
experiments were conducted 
indivi
dually at 610, 620, and 630
 
nm and the results of the LPSVD fits from each of these probe 
wavelengths were averaged together.
 
 
239
 
 
 
 
 
 
 
 
a)
 
Figure
 
5.D.2
.
 
(a) Early time TA data (blue) on [Fe(phen)
2
(C
4
H
10
N
4
)
](PF
6
)
2
 
(
3
) in methanol 
monitored at 550
 
nm following 600
 
n
m excitation. Exponential fit (red) and LPSVD fit (black) 
are superimposed on the dataset. (b) C
orresponding FFT of the residual of the exponential fit 
(blue) and LPSVD vibrational spectrum (red). (c) Analogous data monitoring at 650 nm and (d) 
correspondi
ng FFT (blue) and LPSVD (red) vibrational spectra.
 
b)
 
c)
 
d)
 
 
240
 
 
 
 
Figure
 
5.
D
.
3.
 
(a) Early time TA data (blue) on [Fe(phen)
2
(C
4
H
10
N
4
)
](PF
6
)
2
 
(
3
) in acetonitrile 
monitored at 550
 
nm following 600
 
nm excitation. Exponential fit(red) and LPSVD fit (black) 
are su
perimposed on the dataset. (b) Corresponding FFT of the residual of the 
exponential fit 
(blue) and LPSVD vibrational spectrum (red). (c) Analogous TA data while monitoring at 
650
 
nm and (d) corresponding FFT (blue) and LPSVD (red) vibrational spectra.
 
a)
 
c)
 
b)
 
d)
 
Probe 

 
Solvent
 
Frequency, cm
-
1
 
(Damping Time, fs)
 
550
 
nm
 
MeOH
 
66 (730)
 
115 (280)
 
159 (290)
 
 
 
MeCN
 
50 (324)
 
120 (381)
 
159 (370)
 
 
 
650
 
nm
 
MeOH
 
 
103 (400)
 
163 (570)
 
225 (940)
 
408 (760)
 
MeCN
 
 
108 (370)
 
156 (560)
 
 
412 (850)
 
 
Table
 
5.D.2.
 
Summary of the coher
ent oscillations of [Fe(phen)
2
(C
4
H
10
N
4
)](PF
6
)
2
 
(
3
) in methanol and 
acetonitrile following 600
 
nm excitation. The presented results are from a single dataset for each probe 
wavelength and solvent. As such, no error bars in the form of a
 
standard deviation a
re presented for these 
values.
 
 
241
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Solvent
 

1
,
 
fs
 

2
,
 
fs
 

3
,
 
fs
 
MeCN
 
210 
± 30
 
2300 
± 300
 
8900 
± 600
 
MeOH
 
260 
± 40
 
3400 
± 1200
 
7400 
± 900
 
 
Table
 
5.D.3. 
Summary of the kinetic components of [Fe(phen)
2
(C
4
H
10
N
4
)](PF
6
)
2
 
(
3
) in 
methanol 
and acetonitrile following 600
 
nm excitation. Errors are reported as the standard deviation of 
multiple trials.
 
 
242
 
 
 
 
 
 
 
 
 
 
 
 
 
REFERENCES
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
243
 
REFERENCES
 
 
(1)
 
Juris, A.; Balzani, V.; Barigelletti, F.; Campagna, S.; Belser, P.; Von Zelewsky, A. Ru(II) 
Polypyridine Complexes: Photophysics, Photochemistry, Electrochemistry, and 
Chemiluminescence. 
Coord. Chem. Rev.
 
1988
, 
84
, 85

277.
 
(2)
 
Hexel, G. B.; Hedrick, J. B.;
 
Orris, G. J. Rare Earth Elements
-
Critical Resources for High 
Technology https://pubs.usgs.g
ov/fs/2002/fs087
-
02.
 
(3)
 
Wenger, O. S. Is Iron the New Ruthenium? 
Chem. Eur. J
.
 
2019
, 
25
, 6043

6052.
 
(4)
 
L., L.; Duchanois, T.; Etienne, T.; Monari, A.; Beley, M.; 
Assfeld, X.; Haacke, S.; Gros, P. C. A 
New Record Excited State 3MLCT Lifetime for Metalorga
nic Iron(II) Complexes. 
Phys. Chem. 
Chem. Phys
.
 
2016
, 
18
, 12550

12556.
 
(5)
 
Chabera, P.; Kjaer, K. S.; Prakash, O.; Honarfar, A.; Liu, Y.; Fredin, L. A.; Harlang, T.
 
C. B.; Lidin, 
S.; Uhlig, J.; Sundstro, V.; 
et al
. Fe
II
 

x with a 528 Ps Metal
-
to
-
 
Ligand Charge
-
Transfer Excited
-
State Lifetime. 
J. Phys. Chem. Lett.
 
2018
, 
9
, 459

463.
 
(6)
 
Steube, J.; Burkhardt, L.; Päpcke, A.; Moll, J.; Zimmer, P.; Schoch, R.; Wölper, C.; Heinze, K.; 
Lockbr
unner, S.; Bauer, M. Excited
-
State Kinetics of an Air
-
Stable Cyclometalated Iron(II) 
Complex. 
Chem. Eur. J.
 
2019
, 
25
, 11826

11830.
 
(7)
 
Kjær, K. S.; Kaul,
 
N.; Prakash, O.; Chábera, P.; Rosemann, N. W.; Honarfar, A.; Gordivska, O.; 
Fredin, L. A.; Bergquist, 
K.
-
E.; Häggström, L.; 
et al
. Luminescence and Reactivity of a Charge
-
Transfer Excited Iron Complex with Nanosecond Lifetime. 
Science
 
2019
, 
363
, 249

253.
 
(8)
 
Doonan, D. J.; Balch, A. L. Carbene Complexes Resulting from the Addition of Various Amines 
to Isoc
yanide Complexes of Iron(II) and Ruthenium(II). 
Inorg. Chem
. 
1974
, 
13
, 921

927.
 
(9)
 
Schilt, A, A. Mixed Ligand Complexes of Iron(II) and (III) with Cyani
de and Aromatic Di
-
Immines. 
J. Am. Chem. Soc.
 
1960
, 
82
, 3000

3005.
 
(10)
 
Brown, A. M.; McCusker, C. E.; 
McCusker, J. K. Spectroelectrochemical Identification of Charge
-
Transfer Excited States in Transition Metal
-
Based Polypyridyl Complexes. 
Dalt. Trans.
 
201
4
, 
17635

17646.
 
(11)
 
Slavov, C.; Hartmann, C.; Wachtveitl, J. Implementation and Evaluation of Data Ana
lysis 
Strategies for Time
-
Resolved Optical Spectroscopy. 
Ann. Chem.
 
2018
, 
87
, 2328

2336.
 
(12)
 
Henry, E. R.; Hofrichter, J. Singular Value Decomposition: 
Application to Analysis of 
Experimental Data. 
Meothods Enzym
. 
1992
, 
210
, 129

192.
 
(13)
 
Lemke, H. T.; Kj
aer, K. S.; Hartsock, R.; van Driel, T. B.; Chollet, M.; Glownia, J. M.; Song, S.; 
Zhu, D.; Pace, E.; Matar, S. F.; Nielsen, M. M.; Benfatto, M.; Gaffney
, K. J.; Collet, E.; Cammarata, 
M. Coherent Structural Trapping through Wave Packet Dispersion during P
hotoinduced Spin State 
Switching. 
Nat. Commun
. 
2017
, 
8
, 15342.
 
(14)
 
Auböck, Gerald; Chergui, M. Sub
-
50
-
Fs Photoinduced Spin Crossover in [Fe(bpy)
3
]
2+
. 
Na
t. Chem.
 
2015
, 
7
, 629

633.
 
 
244
 
(15)
 
Monat, J. E.; McCusker, J. K. Femtosecond Excited
-
State Dynamics of an 
Iron(II) Polypyridyl 
Solar Cell Sensitizer Model. 
J. Am. Chem. Soc
. 
2000
, 
122
, 4092

4097.
 
(16)
 
Tribollet, J.; Galle, G.; Jonusauskas, G.; Deldicque, D.; 
Tondusson, M.; Letard, J. F.; Freysz, E. 
Transient Absorption Spectroscopy of the Iron(II) [Fe(phen)
3
]
2+
 

-
Radiative Relaxation of an Isolated Iron(II) Complex. 
Chem. Phys
. 
2011
, 
513
, 42

47.
 
(17)
 
Gütlich, P.; Goodwin, H. A. Spin Cr
ossover in Transition Metals III; Springer
-
Verlag: Berlin, 
Heidelberg, 
2004
.
 
(18)
 
Bixon, M.; Jortner, 
J. Intramolecular Radiationless Transitions. 
J. Chem. Phys
. 
1968
, 
78
, 715

726.
 
(19)
 
Englman, R.; Jortner, J. The Energy Gap Law for Radiationless Transitions in Large Molecules. 
Mol. Phys
. 
1970
, 
18
, 145

164.
 
(20)
 
Carey, M. C.; Adelman, S. L.; McCusker, J. 
K. Insights into the Excited State Dynamics of Fe(II) 
Polypyridyl Complexes from Var
iable
-
Temperature Ultrafast Spectroscopy. 
Chem. Sci
. 
2019
, 
10
, 
134

144.
 
(21)
 
Liu, Y.; Harlang, T.; Canton, S. E.; Chabera, P.; Saurez
-
Alcantara, K.; Fleckhaus, A.; Vithanag
e, 
D. A.; Goransson, E.; Corani, A.; Lomoth, R.; 
et al.
 
Towards Longer
-
Lived Metal
-
t
o
-
Ligand 
Charge Transfer States of Iron(II) Complexes: An N
-
Heterocyclic Carbene Approach. 
Chem. 
Commun
. 
2013
, 
49
, 6412

6414.
 
(22)
 
Kjær, K. S.; Zhang, W.; Alonso
-
Mori, R.; 
Bergmann, U.; Chollet, M.; Hadt, R. G.; Hartsock, R. 
W.; Harlang, T.; Kroll, T.; Kub

et al
. Ligand Manipulation of Charge Transfer Excited 

-
bipyridine)
2
(CN)
2
]. 
Struct. Dyn
. 
2017
, 
4
, 44030.
 
(23)
 
Liu, Y.; Kjær, K. S.; Fredin, L. A.; Chabera, P.; Harlang, T.; Canton, S. E.; Lidin, S.; Z
hang, J.; 
Lomoth, R.; Bergquist, K.; 
et al
. A Heteroleptic Ferrous Complex with Mesoionic Bis(1,2,3
-
triazol
-
5
-
ylidene ) Ligands: Taming the MLCT Excited State of Iron(II). 
2015
, 
21
, 3628

3639.
 
(24)
 
Cannizzo, A.; van Mourik, F.; Gawelda, W.; 
Zgrablic, G.; B
ressler, C.; Chergui, M. Broadband 
Femtosecond Fluorescence Spectroscopy of [Ru(bpy)
3
]
2+
. 
Angew.
 
2006
, 
45
, 3174

3176.
 
(25)
 
Gawelda, W.; Cannizzo, A.; Pham, V.
-
T.; van Mourik, F.; Bressler, C.; Chergui, M. Ultrafast 
Nonadiabatic Dynamics of [
Fe
II
(bpy)
3
]
2+
 
i
n Solution. 
J. Am. Chem. Soc
. 
2007
, 
129
, 8199

8206.
 
(26)
 
Damrauer, N. H.; Cerullo, G.; Yeh, A.; Boussie, T. R.; Shank, C. V.; McCusker, J. K. Femtosecond 
Dynamics of Excited
-
State Evolution in [Ru(bpy)
3
]
2+
. 
Science
 
1997
, 
275
, 54

57.
 
(27)
 
Yeh
, A. T.; Shank,
 
C. V.; McCusker, J. K. Ultrafast Electron Localization Dynamics Following 
Photo
-
Induced Charge Transfer. 
Science
 
2000
, 
289
, 935

938.
 
 
 
 
245
 
Chapter 6:
 
Future Work
 
6.1
.
 
Introduction
 
 
The work in this dissertation has spanned numerous 
transition metals, electron
ic configurations and 
ligand scaffolds in an effort to understand and manipulated ultrafast excited state dynamics in a variety of 
chemical systems. Coherent vibrational dynamics have been used in several ways ranging from an aid
 
in 
characterizing the exci
ted state dynamics to a tool with which to synthetically redesign the chromophore 
around. Although the results described here have expanded what is known about the excited state dynamics 
of several transition metal systems and th
e role of coherent vibratio
nal dynamics, there are still numerous 
questions to be answered or additional research paths to be pursued. In what follows, we will provide a 
brief outlook for future work stemming from the work in each chapter of this dissertat
ion.
 
6.2
.
 
Detailed Kinetic 

3
 
Excited State Dynamics
 
 
The results of chapter
 

3
 
systems involve an initial ultrafast 
4
T
2
 

2
E ISC followed by a thermally activa
ted 
2

4
T
2
 
bISC and 
subsequently by 
4
T
2
 

4
A
2
 
IC to recover the ground electronic state. Kunttu and coworkers
1
 
developed a 
kinetic model to describe the observations of their visible pump IR probe TA experiments shown in 
e
quation 2.1. However, because they fit a
pplied their model to the recovery of the ground state bleach as 
opposed to decay of the excited state absorption signal (which is composed of a superposition of absorption 
from the 
2
E and 
4
T
2
 
states), they were able
 
to simplify their kinetic model to desc
ribe only the net 
2

4
T
2
 
bISC as opposed to the rates of both forwards and backwards reactions. Thus, the kinetic model developed 
by Kunttu and coworkers cannot describe the delayed 
4
T
2
 

2
E ISC after dissipation of some of the thermal 
energy in bulkier 
members of the Cr(a

3
 
series such as Cr(TMHD)
3
. To describe these dynamics more 
completely, a modified kinetic model is outlined in equation 6.1 in which an additional thermally activated 
ISC term has been added to the differential equation describing the concentration o
f the 
4
T
2
 
state. The 
remainder 
of the parameters are identical to what was describe in equation 2.1 with the exception the time 
dependent effective temperature which is rewritten in equation 6.2 to include a generic bath temperature, 
T
bath
, to allow for va
riable temperature measurements
.
 
 
246
 
 
 
6.1a
 
 
 
6.1b
 
 
 
6.1c
 
 
 
6.2
 
 

3
 
s
eries 
would allow for the most detailed description of the excited state dynamics of this class of compounds to 
date and would allow for an intricate view of how each synthetic modification employed in chapter 2 
physically affects the excited state dynam
ic
s. The implementation of such a kinetic model will likely be 
somewhat difficult due to the large number of parameters to fit for. To make the situation easier, three types 
of experiments could be used to gain useful thermodynamic information on each comp
ou
nd. First, since it 
was insinuated by Kunttu and coworkers
1
 
that gr
ound state recovery of Cr(acac)
3
, even at longer timescales, 
involves thermally activated bISC to the quartet manifold, then variable temperature ground state recovery 
measurements would afford t
he activation energy and preexponential factor for bISC. Addi
tionally, variable 
temperature measurements could be employed to monitor the early time dynamics as the temperature of the 
bath plays a role in the time dependent effective temperature in equatio
n 6.2. Finally, variable excitation 
wavelength measurements w
hile monitoring at the same wavelength across the experiment set would vary 
the initial effective temperature, 
T
x
. The combination of each of these experiment types may allow for a fit 
to equatio
n 6.1, and in turn, an extremely detailed view of the excited
 
state dynamics in this class of 
compounds.
 
 
247
 
6.3
.
 
Excited State Dynamics of Six
-
Coordinate Cobalt(III) Compounds
 
 
Chapter 3 of this dissertation discussed the excited state dynamics of a series of
 

3
 
type 
complexes. A lot of the ambiguities from the
se results arose from assignments of spin states. Since visible 
TA experiments are not particularly sensitive to spin state, it is recommended that a complementary 
experiment be used to extract s
uch information such as ultrafast k

 
fluorescence measurement
s which are 
highly sensitive to the spin state on the metal center. However, given that the such a measurement technique 
requires instrumentation that is difficult to get timely access to, altern
ative methods may be more practical 
for future endeavors. One
 
method which could be used to confirm that ground state recovery occurs on a 
2
 
ps timescale in these systems (as opposed to ~2
 
ps formation of the 
5
T
2g
 
state) would be to perform full 
spectrum m
easurements while monitoring more in the UV. This may afford 
relatively larger signals for the 
5
T
2g
 
state as the 

 


*
 
bands shift. This was not performed in the present work for two reasons. First, the 
low extinction coefficients of the ligand field band
s necessitated a very high sample concentration which 
effectively makes the UV portion of the spectrum completely 
opaque. A remedy for this could be to use 
higher energy excitation using a band that has higher extinction coefficients so a lower concentrati
on could 
be used. Second, the white light continuum generated in the sapphire window in the current experiment 
doe
s not generate sufficient UV light to perform the experiment in question. Replacement of the sapphire 
window with a CaF
2
 
window on a translati
on state (to prevent damage) may allow for high enough UV 
output to perform the desired experiment.
 
 
Additionally,
 
we know that cobalt(III) compounds of significantly higher ligand field strength than 

3
 
compounds studied show much longer groun
d state recovery lifetimes (presumably because 
they populate the 
5
T
2g
 
state). It may be interesting to systematica
lly alter the ligand field strength of six
-
coordinate cobalt(III) systems to monitor the onset of this extended lifetime. This could be achiev
ed in 
numerous ways such as using the homoleptic series of Co(acac)
3
 

3
 

3
]
3+
, or using 
a more modular approach  of Co(acac)
3
 

2
(bpy)]
+
 

2
(CN)
2
]
-
 
(or some other strong 
field ligand set) similar to what was performed in chapter 5 of this dissertation, or simply adding 
 
248
 
incrementally hi
gher numbers of nitrogen donors such as Co(acac)
3
 

2
(en)]
+
 

[Co(acac)(tren)]
2+
. Each of these series has t
he benefit of a 

-
system to tap into for high intensity charge 
transfer transitions from the excited state so that a signal can be seen in 
TA measurements.
 
6.4
.
 
Coherence as an Indicator and Constrained Iron(II) Compounds
 
 
The main point of chapter 4 of this dissertation was to illustrate that coherence information could 
point the investigator towards synthetic modifications tha
t could alter
 
photophysical properties. Future 
work in this regard would involve using this same process towards tuning the photophysics of a different 
chemical system. This would require that the system undergoes an ultrafast transformation that is driven
 
by 
large amp
litude nuclear motion. Although short dephasing times were used as an indicator for specific 
modes to target, the reliability of this indicator is still somewhat in question given the rapid dephasing times 
observed for some solvents or the qua
rtz cuvette o
r even the broad range of dephasing times observed in 

3
 
series which each had seemingly identical photophysical properties. However, regardless of 
the meaningfulness of individual dephasing times, excited state coherences relay im
portant infor
mation
 
about the nuclear motion upon photoexcitation which, when coupled with other information as well as 
chemical intuition, could provide important insights with regards to tuning the excited state dynamics of a 
chemical system.
 
Figure
 
6.1.
 
Structures of proposed condensed macrocyclic cage ligands.
 
 
 
 
249
 
 
An additio
nal important
 
insight that this work has brought forth is that highly strained iron(II) 
polypyridyls have much longer MLCT lifetimes. This concept could be extended to other macrocylic cage 
ligands as well such as the ligands shown in Figure
 
6.1. Both of t
hese ligand s
ets have the potential to make 
a highly strained iron(II) chromophore with extended MLCT lifetimes.
 
6.5
.
 
Modular Manipulation of Excited State Dynamics of Iron(II) Chromophores
 
 
The work in chapter 5 of this dissertation has shown that the pho
tophysical pr
operties of iron(II) 
polypyridyl type complexes can be systematically tuned via switching a single bidentate polypyridyl ligand 
with other ligands of varying ligand field strength thus providing enough tunability to alter the ground state 
reco
very and MLCT
 
lifetimes by over an order of magnitude. Future efforts in this area could probe how 
the excited state properties are further perturbed by substituents on the polypyridyl ligand or by swapping 
additional polypyridyl ligands with stronger fiel
d ligands suc
h as the dicarbene ligand used in this work. 
Considering the substantial affect that a single dicarbene ligand has had on the MLCT lifetime, the addition 
of another such ligand has the potential to create a substantially longer lived MLCT exci
ted state wit
h 
spectral properties readily tunable by substituents on the remaining polypyridyl ligand. Furthermore, the 
carbene ligand used in this study has further room for synthetic modifications that can potentially result in 
dramatic changes to its l
igand field s
trength. By swapping out the peripheral methyl group with more 
electron donating or withdrawing groups one can, in theory, continue to significantly alter the electronic 
properties of the carbene carbon and, by extension, the metal center.
 
 
 
 
250
 
 
 
 
 
 
 
 
 
 
 
 
 
RE
FERENCES
 
 
 
 
251
 
REFERENCES
 
 
(1)
 
Ma


as, E. M. S.; Mustalahti, S.; Myllyperkiö, P.; Kunttu, H.; Pettersson, M. Role of Vibrational 
Dynamics in Electronic Relaxation of Cr(
a
cac)
3
. 
J. Phys. Chem. A
 
2015
, 
119
, 2727

273
.