ELECTRONIC STRUCTURE AND EXCITED STATE DYNAMICS OF
CHROMIUM(III) COMPLEXES

VOLUME 1
By

Joel Nicholas Schrauben

A DISSERTATION
Submitted to
Michigan State University
in partial fulﬁllment Of the requirements
for the degree Of
DOCTOR OF PHILOSOPHY
CHEMISTRY

2010

ABSTRACT

ELECTRONIC STRUCTURE AND EXCITED STATE DYNAMICS OF
CHROMIUM(III) COMPLEXES

By
Joel Nicholas Schrauben
Interest in fundamental aspects of transition metal photophysics and
photochemistry stems from potential application of such systems to technologies
such as solar cells, photocatalysts and molecular machines. Chromium(III) offers
a convenient platform for the fundamental study of transition metal photophysics
due to its relatively simple ligand-ﬁeld electronic structure. The work presented in
this dissertation deals with understanding the ground and excited state electronic
structure and dynamics of chromium(III) complexes, ranging from high-symmetry
derivatives of tris(acetylacetonato)chromium(III) (Cr(acac)3) to low symmetry
chromium-semiquinone complexes of the form [(tren)Cr(III)-SQ]+2 (where tren is
tris(2-aminoethyl)amine, a tetradentate amine capping ligand enabling only one
moiety of the orthosemiquinone (SQ) to chelate to the chromium(III) ion). This
effort can be thought of in terms of building up the additional interactions
(lowered symmetry and spin exchange) in a piecewise fashion by ﬁrst considering
the electronic structure and dynamics of high-symmetry systems, then lowering
the symmetry while maintaining the quartet spin nature of the high-symmetry
system by studying the chromium(III)-catechol systems. Finally, spin exchange

can be introduced via the chromium(III)-semiquinone system. In general, these

complexes represent dramatic changes from the high-symmetry complexes in
several ways: 1) the local symmetry of the chromium(III) ion is reduced from
high-symmetry, pseudo-octahedral ligation to a CZV-like N402 coordination,
effectively breaking the degeneracy of the ligand ﬁeld T and E states; 2) unpaired
spin of the semiquinone ligand interacts via Heisenberg Spin-exchange with the
unpaired spins of the chromium(III) ion, resulting in substantial changes in the
absorption spectrum indicative of radically different electronic structure of both
the ground and excited states. Studies of the excited-state dynamics were ﬁrst
carried out on derivatives of the archetypal complex Cr(acac)3 to gain an
understanding of correlations between electronic structure, geometry, and excited
state dynamics. These studies revealed an empirical correlation between low-
frequency modes of the molecule and the rate of ultrafast intersystem crossing in
the ligand ﬁeld manifold. Efforts on the lower symmetry catechol and
semiquinone complexes are focused mainly on synthesis and characterization of
the electronic structure. The ground states of these systems are characterized
primarily using electron paramagnetic resonance techniques, revealing the rich
nature of these spin systems. For these studies, gallium(III)-semiquinones are
employed as a structural analog to study spin density distribution in the absence of
the chromium(III) ion. The concepts learned from these studies provide a useful
backdrop to the eventual study of the excited state dynamics of the aforementioned

chromium(III)-catechol and -Semiquinone complexes.

COPyright by
Joel Nicholas Schrauben
20 10

For my mother, who made sure I did my homework.

ACKNOWLEDGEIVIENTS

I would ﬁrst like to thank my boss, Jim, for setting me up with a project that
encompasses so many of the aspects of inorganic chemistry that drew me to the
area: electronic structure, magnetism, and synthesis, and mixing ultrafast
spectroscopy into the equation. It’s been a pleasure working with you.

Many members of the McCuSker group have contributed to various aspects
of the work presented in this dissertation through their expertise, hard work, and
helpful discussions: Dr. Dong Guo (x-ray structures, synthetic mastermind), Dr.
Richard Fehir, Jr. (synthetic and magnetism wizard, masterﬁil tailgater), Allison
Brown (femto enabler, travel buddy), Dr. Troy Knight (low-T experiments, 40’s),
Drew Kouzelos (low-T, calculations, protector of group traditions), Lisa (e-chem,
Riv) as well as just about everyone else for listening at one time Or another. I’ve
also had the good fortune to collaborate with the McCracken group during my
time at MSU, speciﬁcally Matthew Krzyaniak (expert home-brewer). Also thanks

to Kevin Dillman (coherence!) from the Beck group for work on Cr(acac)3.

My family and ﬁ'iends have done a good job of keeping me grounded and
being patient during my grad school career. Thanks guys. Finally, thanks to Sarah

for waiting.

vi

TABLE OF CONTENTS

List of Tables .................................................................................. x
List of Figures ................................................................................ xi
Chapter 1. Introduction, Historical Perspective and Theory ......................... l
1.1
Introduction .................................................................... 1
1.2 Historical Perspective ................................................... l
1.3 Electronic Structure, Kinetics, and Application of Nonradiative
Decay Theory to Complexes of Chromium(III) ..................... 4
1.3.1 Electronic Structure .......................................... 4
1.3.2 Nonradiative Decay Theory ................................ 12
1.3.3 Applications of Nonradiative Decay Theory to
Complexes of Chromium(III) ............................... 22
1.4 Dissertation Outline ................................................... 29
1.5 References ............................................................... 31
Chapter 2. Ground State Electronic and Magnetic Structure of Gallium(III)-

Semiquinone Complexes and Quartet Complexes of

Chromium(III) .............................................................. 38
2.1 Introduction ............................................................ 38
2.2 Experimental Section .................................................. 45
2.2.1 Synthetic methods ........................................... 45
2.2.2 Physical Measurements on Gallium-Semiquinone
Complexes ................................................... 47
2.2.3 Physical Measurements of Chromium complexes ...... 48
2.2.4 Calculations: Gallium-Semiquinone Complexes ....... 48
2.2.5 Calculations: Chromium Complexes ..................... 50
2.3 Results and Discussion. Gallium(III)-Semiquinone
Complexes .............................................................. 52
2.3.1 [Ga2(tren)2(CAsq’°at)](BPh4)2(BF4) (2) ................... 52
2.3.2 Observation of a Triplet State .............................. 61
2.3.3 ENDOR of the Model Complex ........................... 67
2.3.4 Further Studies ............................................... 72
2.3.5 DFT Calculations of Ga-SQ complexes .................. 72
2.3.6 Conclusions for Ga-SQ complexes ........................ 79
2.4 Results and Discussion. Quartet Complexes of
Chromium(III) .......................................................... 79
2.4.1 Theory. Electronic Structure .............................. 79
2.4.2 Theory. Selection Rules and State Mixing .............. 86

vii

Chapter 3.

Chapter 4.

2.4.3 High-Symmetry complexes of Chromium(III) ......... 87

2.4.4 Low-Symmetry complexes of Chromium(III): X-band
EPR spectra and simulations of
[GaCr(tren)2(CA°a"cm)](BPh4)2 (3) and

[Cr(tren)(DTBCat)](PF6) ................................... 92
2.4.5 Effect of Zero-Field Splitting Parameters on the Energy
Proﬁle of a Quartet Spin System ........................ 97
2.4.6 Conclusions ................................................ 132
2.5 Final Remarks ......................................................... 132
2.6 References ............................................................ 134
Electronic Structure and Ligand Field Dynamics of High-Symmetry
Complexes of Chromium(III) ............................................ 139
3.1 Introduction ........................................................... 139
3.2 Background: Ultrafast Transient Absorption Spectroscopy....142
3.3 Experimental Section ................................................ 153
3.3.1 Physical methods .......................................... 153
3.3.2 Computational Methods ................................. 158
3.3.3. Synthesis 1. 3-Substituted Complexes of 2,4-
Pentanedione ............................................... 158
3.3.4 Synthesis II. Substituted Complexes of 1,3-
Propanedione .............................................. 164
3.4 Results and Discussion .............................................. 170
3.4.1 Cr(acac)3 ................................................... 170
3.4.2 3-Substituted Complexes of 2,4-pentanedione ........ 189
3.4.3 1,3-Substituted Complexes of 1,3-propanedione.....220
3.5 Comparison of Complexes ......................................... 244
3.6 Final Remarks ........................................................ 250
3.7 References ............................................................ 251
Electronic Structure and Nonradiative Dynamics of Heisenberg Spin
Exchange Complexes of Chromium(III) .............................. 257
4.1 Introduction ........................................................... 257
4.2 Experimental Section ................................................ 262
4.2.1 Synthetic Methods ......................................... 262
4.2.2 Physical Measurements ................................... 269
4.3 Theory of Heisenberg Spin Exchange .............................. 269
4.4 Results and Discussion .............................................. 275
4.4.1 Previous Results ........................................... 275
4.4.2 [Cr(tren)(3,6-R-Q)]""+2 series ........................... 283
4.4.3 Spin Exchange Members of the [M1M2(tren)2(CA"')]m+
Series ......................................................... 300

viii

Chapter 5.

Appendices.

4.5 Final Remarks ...................................................... 312

4.6 References ........................................................... 314
Future Work ............................................................. 319
5.1 Introduction ......................................................... 319
5.2 High-Symmetry Complexes ....................................... 319
5.3 Low-Symmetry and Spin-Exchange Complexes ............... 326
5.4 References ............................................................ 329
Appendix A: Additional Figures for Chapter 1 ....................... 331
Appendix B: Development of Characteristic Equations for Isolated
Spin Systems ............................................. 333
Appendix C: Crystallographic Data for Chapters 3 and 4 ........... 337

ix

Table 2-1.

Table 2-2.

Table 2-3.

Table 3-1.

Table 3-2.

Table C-l.

Table C-2.

Table C-3.

Table C—4.

LIST OF TABLES

Experimental and calculated spin densities for
[Ga2(tren)2(CAsq’°at)]3+ (2), [Ga2(tren)2(DHBQ)]3+ (2a), and the free
ligands [CAsq’ca‘f' and [DHBQ]3’ ........................................ 78

Spin Hamiltonian parameters for several substituted complexes of
Cr(acac)3 obtained in a 4 K glass ofbutyronitrilezpropionitrile......91

Spin Hamiltonian parameters, as obtained through simulation, for 3
and its model complex, [Cr(tren)(DTBCat)](PF6) ..................... 94

Relevant average bond lengths and bond angles for Cr(DBM)3 and
Ga(DBM)3. Calculated values were obtained at the UB3LYP/6-

311 g** level employing a CPCM solvent model of
dichloromethane .......................................................... 210

Relevant average bond lengths and bond angles for Cr(DBM)3 and
Ga(DBM)3. Calculated values were obtained at the UB3LYP/6-
311 g" level employing a CPCM solvent model of

dichloromethane .......................................................... 236
Select bond lengths and angles for Cr(3-N02ac)3 .................... 337
Crystallographic data for Cr(3-NOzac)3 and Cr(3-Phac)3 ........... 338
Crystallographic data for Cr(DBM)3 and Ga(DBM)3 ............... 339

Crystallographic data for [Cr(tren)(pycat)](BPh4) and [Cr(tren)(3,6-
1,2-orthocatecholate)](BPh4) ........................................... 340

Figure 1-1

Figure 1-2.

Figure 1-3.

Figure 1-4.

Figure 1-5.

Figure 1-6.

Figure 1-7.

Figure 1-8.

Figure 2-1.

Figure 2-2.

Figure 2—3.

LIST OF FIGURES

. Tanabe-Sugano diagram for a d3 ion in 0 symmetry and one-electron

representations of the relevant electronic levels in the ligand-ﬁeld
manifold of the Chromium(III) ion in an Oh environment ............. 5

Charge redistribution in d3 system as a result of promotion to an eg“
type orbital .................................................................... 8

Excited state (4T2) distortions of hexaamminechromiumﬂll), as
determined by Wilson and Solomon ...................................... 8

Jablonski diagram of photophysical processes that occur in a (13 ion
under 0 symmetry ......................................................... 11

Normal region—as AB is decreased the rate of electron transfer
decreases as a result of decreased vibrational overlap between
component vibrational wavefunctions ................................... 17

Inverted region—as AE is decreased the rate of electron transfer
increases as a result of increased vibrational overlap between

component vibrational wavefunctions ................................... 17

Proposed mechanisms for thermally activated 2E state deactivation,

reproduced from reference. (a. = direct reaction, b. = back
intersystem crossing into the quartet manifold and c. = surface
crossing to a ground state intermediate.) ................................ 26

Sterically constraining ligands that helped to elucidate modes of 2E
deactivation .................................................................. 28

Bridging forms of chloranilic acid. CAcat‘cat refers to the
diamagnetic dicatecholate form, while CASq’ca‘ is the semiquinone-
catecholate form, and is paramagnetic (S = 1/2). The resonance forms
show that SOMO (singly-occupied molecular orbital) density should

be distributed among the ketone-like C-O moieties ................... 40

Members of the bimetallic [M1M2(tren)2(CA"')]m+ series ............ 41

The complex [GaCr(tren)2(CA°at‘°at)](BPh4)2 (3) and its model

complex [Cr(tren)(DTBCat)](PF6) ....................................... 43
xi

Figure 2-4.

Figure 2-5.

Figure 2-6.

Figure 2-7.

Figure 2-8.

Figure 2-9.

Figure 2-10.

(A) Experimental X-band spectrum of complex 2 acquired in a
butyronitrile/propionitrile (9:2) glass at 4 K (0.63 uW, 9.458 GHz).
This spectrum was simulated (B) using the following parameters:
gxx = 2.0100, gyy = 2.0097, gZZ = 2.0060, axx(Ga) = 4.902 G, ayy(Ga)
= 4.124 G, aZZ(Ga) = 3.167 G. (C) is the experimental X-band
spectrum of [Ga2(tren)2(DHSQ)](BPh4)2(BF4), acquired at 4.4 K,
63 nW, 9.624 GHz ......................................................... 53

Calculated SOMO (a.) and spin density (b.) of complex 2 at the
UB3LYP/6-31 1G** level ................................................. 55

Room temperature X-band Spectrum of 2 (black), acquired at 9.696
GHz. The spectrum was simulated (red) with g = 2.00858, Z =
3.474 G (A( 9Ga) = 3.136 G and A(7‘Ga) = 3.980 G)..................60

Room temperature X-band spectrum of the model complex (black),
acquired at 9.695 GHz. The spectrum was simulated (red) with g =
2.00906, Z = 3.697 G (A(69Ga) = 3.337 G and A(7‘Ga) = 4.240
G) ............................................................................. 61

EPR resonances produced by the triplet state of a biradical complex.
The spectrum of 2 (black) was acquired at 4 K (9.624 GHz, 0.63
mW), and simulation of the sidebands (red) was carried out
assuming a triplet state, and with the Spin Hamiltonian parameters g
=2.009andD=150G ..................................................... 63

Energy level diagram of the Zeeman splitting of the triplet state
formed by ferromagnetic exchange between two molecules of 2.
The diagrams were produced from the Simulation parameters g =
2.009 and D = 150 G, and resonances are Shown at X-band
frequency (~ 0.3 cm") ..................................................... 64

(A) W-band data of 2 acquired at 94.158 GHz, 5 nW, 10 K. The
ﬁrst derivative was taken by applying a pseudo-modulation of 10
kHz (B). The main peak belongs to the monoradical 2, and was
simulated with an anisotropic g tensor (2.0077, 2.0062, 2.0053)
with widths (11, 5, 4) G (C). The sidebands were simulated
assuming a triplet electronic state formed by interaction of two
monoradical species (D). Simulated with g = 2.006 and D = 150
G .............................................................................. 66

xii

Figure 2-1 1.

Figure 2-12.

Figure 2-13.

Figure 2-14.

Figure 2-15.

Figure 2-16.

Figure 2-17.

Figure 2-18.

Comparison of the experimental X-band spectrum of complex 2
(red) acquired in a butyronitrile/propionitrile (9:2) glass at 4 K (0.63
uW, 9.458 GHz) with another sample saturated with
tetrabutylammonium tetraphenylborate salt (black), collected under
the same conditions. The diminished intensity of the sidebands for
the spectrum of solution containing the TBA salt reveals that these
peaks are likely due to an electronic state formed by aggregation of
complex 2 in solution ...................................................... 67

ENDOR spectrum of the model complex

[Ga2(tren)2(DHSQ)](BPh4)2(BF 4) acquired at X-band frequency and
10 K in a 9:2 butryonitrile/propionitrile glass7l

The ground quartet state of a Cr(III) ion under Oh symmetry (4A2g)
consists of two Kramer’s doublets split by 2|D|. As the symmetry is

lowered the rhombic zero ﬁeld splitting parameter E affects the
splitting, and as the system is introduced to an external magnetic
ﬁeld (B) the Kramer’s doublets split further according to the Zeeman
interaction ................................................................... 82

Experimental (black) and simulated (red) continuous-wave EPR
spectra of tris(3-phenyl-2,4-pentanedionato)chromium(III) (Cr(3-
Phacac)3) obtained in a 4 K glass of 9:2 butyronitrile:propionitrile.
Experimental conditions: v = 9.4775 GHz, modulation amplitude =
20 G, conversion time = 80 ms, power = 63 uW. Simulation
parameters are indicated in the ﬁgure .................................... 90

Experimental (black) and simulated spectra (red) for (a)

[Cr(tren)(DTBCat)](PF6) and (b) Complex 3. The spectra were
acquired at 4 K, with a power of 0.32 mW at 9.4595 GHz. The
transition roadmap of 3 is shown above ................................. 93

Impurity signals in the experimental X-band spectrum of complex 3
due to the homometallic species 6 and 2 (inset). All spectra were
acquired at 4 K and 9.45 GHz ............................................ 95

Schematic representation of an angular dependence diagram for a
simple S = 1/2 system under rhombic symmetry ....................... 97

Energy level diagrams showing how the experimental axial (D) and

rhombic (E) zero ﬁeld splitting parameters affect the energy proﬁle
of the ground 4A state for the magnetic ﬁeld parallel to the principal

xiii

Figure 2-19.

Figure 2-20.

Figure 2-21.

Figure 2-22.

axes of the system. In an isotropic system the two Kramer’s
doublets of the ground state are degenerate, and the energy proﬁles
are the same for every orientation with respect to the magnetic ﬁeld.
For an axial system the Kramer’s doublets are split by 2D, and the
x and y axes are degenerate. For rhombic systems an additional zero
ﬁeld splitting parameter is needed to describe the splitting between
the Kramer’s doublets, and the energy proﬁles for the magnetic ﬁeld
along the x, y, and z axes are different ................................. 100

Transition diagrams for quartet systems. Leﬁ: Resonant ﬁeld of
transition at X- Band frequency for an axial quartet spin system with
variable D. The labels indicate the origin of the particular transition
as either the magnetic ﬁeld parallel to the molecular z or xy axes,
which are degenerate under axial symmetry. The line drawn across
corresponds to the experimental value of D for the model complex,
0.396 cm". Right: Resonant ﬁeld of transition for a quartet system
under rhombic symmetry for D = 0.396 cm'1 and variable E/D. The
line drawn across the plot corresponds to the experimental value of
E/D for the model complex, 0.308 ...................................... 101

“Transition roadmap” of an isotropic quartet spin system. The
diagram shows the position of transitions as a function of orientation
with respect to the magnetic ﬁeld. There are only three possible
transitions in an isotropic system, and only the transition at g’ = 2
(around 3300 G) is formally allowed ................................... 102

Simulated spectra of quartet spin systems showing the effect of the
axial and rhombic zero ﬁeld splitting parameters. Left: Simulated
spectra for variable D, from 0 to 0.5 cm". The spectrum in red
corresponds most closely to the experimental value of D for the
model complex. Right: Variable E/D for D = 0.396 cm", the
experimental value of the model complex. The spectrum in red

corresponds closely to the experimental E/D value for the model
complex (0.308) ........................................................... 104

“Transition roadmap” of an axially symmetric quartet spin system,
where D = 0.396 cm], the experimentally observed value of the
model system. The transitions are labeled according to the energy-
level diagram on the right ............................................... 106

Figure 2-23. Calculated X-band EPR spectra for an axial quartet spin system with

variable D (in cm'l) ....................................................... 109

xiv

Figure 2-24.

Figure 2-25.

Figure 2-26.

Figure 2-27.

Figure 2-28.

Figure 2-29.

Energy level diagrams for B parallel to the x (or y), and z magnetic
axes for an axial quartet spin system with variable D (in cm'l)....110

Resonant ﬁeld of transition at X- Band frequency for an axial quartet
spin system with variable D. These transitions are color-coded to
those of Figure 2-24 ...................................................... 111

Resonant ﬁeld of transition diagram along the X, Y, and Z magnetic
axes for an axial quartet spin system with variable D (in cm“). The
resonances are at W-Band frequency, and are color-coded to the
energy-level diagram (Figure 2-27) .................................... 112

Energy level diagrams along the X, Y, and Z magnetic axes for an
axial quartet spin system with variable D (in cm") .................. 113

“Transition roadmap” of a rhombic quartet spin system, where D =
0.396 cm'land E = 0.124 cm“, the experimentally observed values of
the model system. The transitions are labeled according to the
energy-level diagram on the right ....................................... 115

Calculated X-band EPR spectra for quartet spin system with D =
0.05 and 0.10 cm" and variable E/D .................................. 120

Figure 2—30. Calculated X-band EPR spectra for quartet spin system with D =

Figure 2-31.

Figure 2-32.

Figure 2-33.

Figure 2-34.

0.15 and 0.20 cm" and variable E/D ................................... 121

Energy level diagrams along the X, Y, and Z magnetic axes for a
quartet spin system with D = 0.10 cm'1 and variable E/D. The

resonances are at X-band frequency, and are color-coded to Figure
2-32 ......................................................................... 122

Resonant ﬁeld of transition at X- Band frequency for a quartet spin
system with D = 0.10 cm'1 and variable E/D ........................... 123

Calculated X-band EPR spectra for quartet Spin system with D =
0.25 and 0.30 cm" and variable E/D ................................... 124

Energy level diagrams along the X, Y, and Z magnetic axes for a
quartet spin system with D = 0.30 cm'1 and variable E/D. The

resonances are at X-band frequency, and are color-coded to Figure
2-35 .......................................................................... 125

XV

Figure 2-35.

Figure 2-36.

Figure 2-37.

Figure 2-38.

Figure 2-39.

Figure 2-40.

Figure 3-1.

Figure 3-2.

Figure 3-3.

Figure 3-4.

Resonant ﬁeld of transition at X- Band frequency for a quartet spin
system with D = 0.30 cm'1 and variable E/D ........................... 126

Calculated X-band EPR spectra for quartet spin system with D =
0.35 and 0.40 cm" and variable E/D ................................... 127

Calculated X-band EPR spectra for quartet spin system with D =
0.45 and 0.50 cm" and variable E/D ................................... 128

Calculated X-band EPR spectra for quartet spin system with D =
1.00 cm'1 and variable E/D .............................................. 129

Energy level diagrams for B parallel to the X, Y, and Z magnetic
axes for a quartet spin system with D = 0.50 cm'1 and variable E/D.
The resonances are at X-band frequency, and are color-coded to

Figure 2-40 ................................................................ 130
Resonant ﬁeld of transition at X- Band frequency for a quartet spin

system with D = 0.50 cm'l and variable E/D ........................... 131
Molecules of interest in this study ........................................ 141

The transient absorption experiment employs a pump pulse, which
prepares an excited state population, and a probe pulse, which
interrogates dynamics within the excited state. Time resolution of
the experiment is achieved by varying the distance of the paths
traversed between the pump and probe pulses ........................ 143

Cartoon showing ground state absorption (GSA) and excited state
absorption (ESA), plotted in units of extinction coefﬁcient (a, M'1
cm'l) (left graph). The transient absorption experiment effectively
measures the difference in the extinction coefﬁcient of these two
states (right graph), leading to the possibility of both positive
(“BSA”) and negative (“bleach”) features in the transient
spectrum .................................................................... 144

Case of two identical electronic potential wells with small relative
displacement along the nuclear coordinate. Higher-lying vibrational
levels can absorb readily at many wavelengths (hv(l)), but as SI
vibrationally cools the number of available transitions is limited

xvi

Figure 3-5.

Figure 3-6.

Figure 3-7.

Figure 3-8.

Figure 3-9.

Figure 3-10.

Figure 3—1 1.

Figure 3-12.

Figure 3-13.

Figure 3-14.

Figure 3—15.

Figure 3-16.

(hv(2)), and the spectrum of $2 accordingly narrows. Amamx does not
shiﬁ due to the identical Slopes of S. and 82 .......................... 147

Potential energy surface diagrams detailing expected shifts of Ammex
for different slopes of S1 and S2 ......................................... 149

Cartoon representing transient full spectra that are red-shifting with
time. Black represents the earliest time excited state absorption
spectrum while red represents that of the therrnalized, metastable
state ......................................................................... 150

Generalized picture revealing how an intersystem crossing event can
affect the observed transient spectra ................................... 152

Transient ﬁJll spectra of a surface crossing event, with characteristic
spectra from the initial and ﬁnal electronic states. . 153

Ultrafast transient absorption setup which can achieve ~ 100 fs
pulses in the visible ....................................................... 155

Absorption and emission spectra for Cr(acaC)3, acquired in
dichloromethane and a low-temperature (80 K) optical glass of 2-
methyltetrahydrofuran, respectively .................................... 171

Orbital parentage of the 4LMCT transition ........................... 172
Orbital parentage of the 4(3IL) transition .............................. 174
Semi-qualitative energy level diagram (left) and diagram showing

the orbital parentage of the relevant ligand ﬁeld electronic states
(right) for Cr(acac)3 ...................................................... 176

Data revealing vibrational cooling dynamics on the 2B surface of
Cr(acac)3 aﬁer excitation of the sample at 625 nm, adapted from
reference 6. See text for details ......................................... 177

Jablonski diagram summarizing the nonradiative rates in the low-
lying ligand ﬁeld manifold of Cr(acac)3 ................................ 179

Kinetic trace for Cr(acac)3 pumped at 600 nm and probed at 592 i 2
nm. A monexponential rise (1 = 320 fs) in excited state absorption
was subtracted from the raw data (top panel) to obtain a ﬂat baseline.

xvii

ll

Figure 3-17.

Figure 3—18.

Figure 3-19.

Figure 3-20.

Figure 3-21.

Figure 3-22.

Figure 3—23.

Figure 3-24.

Figure 3-25.

The resulting data (bottom panel) are ﬁt with a rapidly damped
oscillatory component (164 cm“, ‘L’ = 70 fs) and one weaker
component (75 cm“) ...................................................... 181

Kinetic trace for Cr(acac)3 pumped at 600 nm and probed at 608 :1: 2
nm. A 1.6 ps rise in excited state absorption was subtracted ﬁ'om the
raw data (top panel) to obtain a ﬂat baseline. The resulting data
(bottom panel) are ﬁt with a rapidly damped oscillatory component
(165 cm" , T = 70 fs) and one weaker component (28 cm'l).........182

Kinetic trace for Cr(acac)3 pumped at 600 nm and utilizing the
integrated probe pulse centered at 600 nm (top). The data was ﬁt
with a biexponential rise in ESA (1'1 = 50 fs, I 2 = 1.3 ps) ........... 183

Ground state frequency calculation for Cr(acac)3 at 250 cm'l at the
UBLYP/6-3llg** level, employing a CPCM solvent model for
acetonitrile ................................................................. 1 87

UV-vis spectrum recorded in dichloromethane and emission
spectrum recorded in a low temperature glass of 2-methyl-
tetrahydrofuran for Cr(3-NOzac)3 ....................................... 191

Full spectral ground state recovery dynamics for Cr(3-N02ac)3 and
long-time single wavelength kinetic trace for Cr(3-N02ac)3 aﬁer
excitation at 633 nm ...................................................... 193

Full spectral and single wavelength kinetic data of Cr(3-N02ac)3 in
dichloromethane excited at 630 nm and 633 nm, respectively. 194

Observed lifetime as a ﬁlnction of pump energy at 480 nm probe.
All measurements were carried out in dichloromethane ............ 194

Absorption and emission spectra for Cr(3-Brac)3 recorded in
dichloromethane and an 80 K optical glass of 2-
methyltetrahydrofuran, respectively. See experimental section for
details ....................................................................... 196

Excitation scan of Cr(3-Brac)3 in an 80 K optical glass of 2-
methyltetrahydrofuran. The excitation scan reports the emission
intensity (probed near the emission maximum at 804 nm) as a
function of the excitation wavelength ................................. 198

xviii

Figure 3-26.

Figure 3-27.

Figure 3-28.

Figure 3-29.

Figure 3-30.

Figure 3-31.

Figure 3-32.

Figure 3-33.

Figure 3-34.

Figure 3-35.

Figure 3-36.

Full spectra for Cr(3-Brac)3 at 630 nm. Anomalies in the full spectra
around 630 nm are due to scatter of the pump pulse ................. 199

Representative single wavelength kinetic data of Cr(3-Brac)3 in
dichloromethane excited at 610 nm, with probe wavelengths
reported in the ﬁgure. The data were ﬁt with a monoexponential
function with lifetime 1.4 d: 0.1 ps ...................................... 200

Absorption and emission Spectra for Cr(3-methylac)3 recorded in
dichloromethane and an 80 K optical glass of 2-
methyltetrahydrofuran, respectively. See experimental section for
details ....................................................................... 202

UV-vis absorption (black and red traces) and emission spectra of
Cr(SCNac)3, collected in dichloromethane and an optical glass of 2-
methyltetrahydrofuran, respectively. See text for details ........... 205

Full spectra transient absorption spectra for Cr(3-SCNac)3 aﬁer 635
nm excitation .............................................................. 207

Representative single wavelength kinetics for Cr(3-SCNac)3, excited

at 515 nm. All ﬁts were carried out with a monoexponential decay
function to yield a lifetime of 1.65 i 0.1 ps ........................... 207

Vibrational cooling lifetime as a function of excitation energy for
Cr(3-SCNac)3 .............................................................. 208

Crystal structure of Cr(3-Phac)3. Representative bond lengths and
angles are shown in the ﬁgure .......................................... 209

UV-vis absorption (black and red traces) and emission spectra of
Cr(3-Phac)3, collected in dichloromethane and an optical glass of 2-
methyltetrahydrofuran, respectively. See text for details ........... 211

Representative data of ground state recovery for Cr(3-Phac)3. This
data was acquired with an excitation wavelength of 610 nm and a
probe wavelength of 480 nm. The value of 1.7 - 1.9 ns ground state
recovery time was found for all pump/probe combinations. . . . .....212

Full spectra transient absorption spectra of Cr(3-Phac)3 at 633 nm
pump in dichloromethane ................................................ 214

xix

Figure 3-37.

Figure 3-38.

Figure 3-39.

Figure 3-40.

Figure 3-41.

Figure 3-42.

Figure 3-43.

Figure 3-44.

Figure 3-45.

Figure 3-46.

Single wavelength kinetics for Cr(3-Ph-acac)3 excited at 560 nm
(top) and 630 nm (bottom). All ﬁts at 480 nm probe were carried
out with a monoexponential decay function yielding 1.30 at 0.05
ps ........................... . ................................................. 215

Lifetime of transient absorption dynamics on the blue edge of the
excited state absorption as a function of excitation wavelength. . ..216

Potential energy surfaces diagram highlighting the charge-transfer
transition that is employed to monitor dynamics on the 2E surface. If
the phenyl group of the 3-phenyl-acetylacetonate ligand rotates more
in-plane with the core of the acetylacetonate ligand the energy Of the
charge-transfer transition employed as a probe lowers with respect to
the ligand ﬁeld manifold, resulting in a red-shift of the excited state
spectrum... ...................................................................................... 218

Absorption spectrum of Cr(3-mesac)3 in dichloromethane. . . . . ....220

UV-vis and emission spectra of Cr(tbutylac)3 in dichloromethane
and an 80 K optical glass of 2-methyltetrahydrofuran .............. 221

Full spectra of Cr(t-butylac)3 after excitation 633 nm and select
single wavelength kinetic traces after 610 nm excitation. Similar
evolution of the full spectra were observed for other excitation
wavelengths across the 4A2 —> 4T2 transition .......................... 223

Gaussian deconvolution of the transient full spectra. An example ﬁt
for the transient spectrum at 2 ps is shown. See text for details of the
ﬁts ........................................................................... 224

Absorption and emission spectra of Cr(prop)3 recording in
dichloromethane and an 80 K optical glass of 2-methyl-
tetrahydrofuran, respectively ............................................ 227

Absorption spectrum of Cr(prop)3 in dichloromethane, revealing the
vibronic structure with ~ 1000 cm'1 splitting on the red edge of the
charge transfer band ....................................................... 228

Short timescale dynamics of a ﬂowing sample of Cr(prop)3 in
acetonitrile after excitation at 633 nm ................................. 229

XX

Figure 3-47.

Figure 3-48.

Figure 3-49.

Figure 3-50.

Figure 3-51.

Figure 3-52.

Figure 3-53.

Figure 3-54.

Figure 4-1.

Figure 4-2.

Figure 4-3.

Figure 4-4.

UV-vis and emission in dichloromethane and a low-temperature
glass of 2-methyltetrahydrofuran ....................................... 232

Single wavelength kinetic data for Cr(DCM)3 excited at 515 nm
(right) and 610 nm (leﬁ). All data were ﬁt with monoexponential
decay functions. The data at 515 nm excitation reveal lifetimes of a.)
1.78 :1: 0.50 (530 nm probe), b.) 1.93 d: 0.25 ps (480 nm probe) and
c.) 1.02 i 0.14 ps (640 nm probe). At 610 nm excitation the
lifetimes are a.) 1.43 i: 0.26 pS (480 nm probe), b.) 1.40 i 0.25 pS
(560 nm probe) and c.) 0.49 :1: 0.12 ps (520 nm probe). See text for
details ....................................................................... 233
Cr(DBM)3 crystal structure ............................................. 235

UV-vis absorption and emission for Cr(DBM)3 acquired in
dichloromethane solution and a low-temperature optical glass of 2-
methyltetrahydrofuran ................................................... 23 8

Full spectra of Cr(dbm)3 in dichloromethane solution after excitation
at 515 nm. The red spectrum near the baseline is at negative time,
and the other spectra are spaced by 30 fs, with red being the earliest
and blue the latest, starting at 0.1 ps and ending at 1.1 ps. Similar
full spectra were observed for other excitation wavelengths across
the 4A2 -—> 4T2 absorption ................................................ 239

Representative single wavelength kinetic traces for Cr(dbm)3 in
dichloromethane for excitation at 610 nm ............................. 241

Lifetime values for various excitation energies across the 4T2
absorption for 700 nm probe ............................................ 241

UV-vis absorption spectrum for Cr(hfac)3 in dichloromethane. . ..242
The redox states of the ortho-quinone ligands ........................ 259

Members of the [Cr(tren)(3,6-R-Q)]+”+2 series employed in this
study ........................................................................ 260

Members of the bimetallic [M1M2(tren)2(CA"')]m+ series ........... 261
Antiferromagnetic exchange arises from non-orthogonality of

constituent spin orbitals, resulting in the stabilization of the lowest
total spin state ............................................................. 273

xxi

Ill

Figure 4-5.

Figure 4-6.

Figure 4-7.

Figure 4-8.

Figure 4-9.

Figure4-l0.

Figure 4-1 1.

Figure 4-12.

Figure 4-13.

Figure 4-14.

Figure 4-15.

Ferromagnetic exchange arises from complete orthogonality of
constituent spin orbitals. The highest total spin state is stabilized
according to Hund’s rule274

Possible redox formulations of the Chromium(III)-semiquinone
exchange-coupled species ................................................ 275

UV-vis absorption spectra of [Cr(tren)(3,6DTBCat)](PF6) (a) and its
semiquinone analog, [Cr(tren)(3,6DTBSQ)](PF6)2 (b) ............. 276

Energy-level diagram of a Chromium(III) system under sz
symmetry, showing the three lowest-lying ligand ﬁeld excited
states ........................................................................ 277

Splittings of the sz electronic components as a result of anti-
ferromagnetic coupling between the chromium ligand ﬁeld states
(left side) and the semiquinone 7t magnetic orbital. The diagram
reveals that many spin-allowed transitions are possible as a result of
the Spin-exchange interaction, whereas energetically similar

transitions in a non-exchanged system are forbidden ................ 280
Ultrafast transient absorption dynamics of
[Cr(tren)(3,6DTBCat)](PF6) and [Cr(tren)(3,6DTBSQ)](PF6)2 in
acetonitrile solution aﬁer excitation at 333 nm ....................... 283

Crystal structure and selected bond lengths and angles for
[Cr(tren)(pycat)]BPh4 .................................................... 285

UV-vis absorption spectrum of [Cr(tren)(pyrocatecholate)]BPh4 in
acetonitrile ................................................................. 287

Cyclic voltammogram of [Cr(tren)(pyrocatecholate)]BPh4 in

acetonitrile. The values are references of

Ag/AgCl ............................................................. . ....... 288

Crystal structure and selected bond lengths and angles for

[Cr(tren)(3,6-CN-1,2-catecholate)]BPh4 ............................... 290

UV-vis absorption spectrum of [Cr(tren)(3,6-CN-l,2-

catecholate)]BPh4 in acetonitrile ....................................... 291
xxii

Figure 4-16.

Figure 4-17.

Figure 4-18.

Figure 4-19.

Figure 4-20.

Figure 4-21.

Figure 4-22.

Figure 4-23.

Figure 4-24.

Figure 4-25.

Cyclic voltammogram of [Cr(tren)(3,6-CN-l,2-
orthocatecholate)]BPh4 in acetonitrile. Potentials are referenced
versus Ag/AgCl ........................................................... 292
UV-vis absorption spectrum of [Cr(tren)(3,6-dinitro-1,2-
orthocatecholate)]BPh4 in acetonitrile ................................ 294

Cyclic voltammogram of [Cr(tren)(3,6-NOz-l,2-
orthocatecholate)]BPh4 in acetonitrile. Potentials are referenced
versus Ag/AgCl ........................................................... 296
UV-vis absorption spectrum of [Cr(tren)(3,6-diamino-1,2-
orthocatecholate)]BPh4 in acetonitrile ................................. 298
Cyclic voltammogram of [Cr(tren)(3,6-NH2-1,2-
orthocatecholate)]BPh4 in acetonitrile. Potentials are referenced
versus Ag/AgCl ........................................................... 299

Comparison of bimetallic semiquinone complexes, showing that the
signals of both the parallel mode and perpendicular mode spectra of
complex 4 can be attributed to impurities in the sample.
Experimental conditions: Complex 6: 9.4595 GHz, 0.30 mW, 4K;
Complex 4 perpendicular mode: 9.6250 GHz, 62.8 mW, 4 K;
Complex 4 parallel mode: 9.3941 GHz, 63 mW, 4 K; Complex 6:
9.4595 GHz, 0.63 mW .................................................... 303

X-band perpendicular mode EPR spectrum of 5 at 4 K. This
spectrum is plotted against the corresponding X-band spectrum for
6, revealing that many of the observed features in the spectrum of 5
can be attributed to complex 6 as an impurity ......................... 305

Boltzmann distribution of spin states as a function of temperature for
complex 5 assuming an antiferromagnetic coupling of J = 4 cm'1 and
no further zero ﬁeld splitting of the m5 sublevels .................... 307

Perpendicular mode X-band spectrum of compound 6 (9.458 GHz, 4
K, 0.3 mW). The spectrum was simulated assuming an isolated
sextet ground state using the following parameters: gxx = 1.98, gyy =
1.975, gzz = 1.97; D = 0.458 cm“, E = 0.1008 cm" .................. 310

Parallel mode X-band spectrum of compound 6 (9.394 GHz, 4 K,
3.16 mW). The spectrum was simulated assuming an isolated sextet

xxiii

Figure 4-26.

Figure 5-1.

Figure 5-2.
Figure 5-3.

Figure 5-4.

Figure 5-5.
Figure 5-6.
Figure A-l.
Figure A-2.

Figure C-l.

ground state using the following parameters: gxx = 1.98, gyy = 1.975,
g, = 1.97; D = 0.458 cm", E = 0.1008 cm" .......................... 311

Energy level diagram of the ground state of complex 6 as a function
of magnetic ﬁeld, calculated from the simulation parameters ...... 312

The lowest-energy spin-allowed transition for Cr(acac)3, as
calculated at the UB3LYP/6-311g** level (employing a CPCM
solvent model for acetonitrile) reveals that 9201 and 9301 molecular
orbitals (major contributors to the HOMO) have n-antibonding
symmetry while the 9701 and 98 01 orbitals (major contributors to this
transition) have sigma-antibonding symmetry. This is qualitatively
in line with a transition resulting in population of an eg* orbital.
Geometry minimization of these MOS effectively constitutes the 4T2
geometry, upon which a frequency calculation can be carried

out ........................................................................... 32 1
Proposed synthesis of dg-acac ........................................... 322
“Heavy” analogs of Cr(acac)3 ............................................ 322

Ground state DF T frequency calculations for Cr(acac)3 and
isotropically-enriched analogs of Cr(acac)3, referenced to the Raman
spectrum ..................................................................... 323
[I\J6-chromium(III)]3+ series of variable ligand rigidity .............. 324

A series of increasing constrained acetylacetonate-type ligands ...325

Tanabe-Sugano diagram of a d3 system ................................ 331
Tanabe-Sugano diagram of a (16 system ................................ 332
Crystal Structure for Cr(3-NOzac)3 ..................................... 337

Images in this dissertation are presented in color.

xxiv

Chapter 1: Introduction, Historical Perspective and Theory

1.1 Introduction

This dissertation concerns itself with fundamental questions underlying the
photophysical and physicochemical properties of transition metal compounds, and
employs chromium(III) as a platform for these studies. This chapter will present
the necessary background, both historical and theoretical, for understanding
excited state photophysical processes in transition metal systems, including an
overview of electronic structure and kinetics of Chromium(III) complexes and
nonradiative decay theory. Other theories pertaining to electronic structure,

dynamics, and magnetism are covered throughout this dissertation as necessitated.

1.2 Historical Perspective

Rational design of applications such as solar cell technology, molecular machines,
and artiﬁcial photosynthesis demands a strong understanding of the electronic
structure and dynamics that constitute the photophysical properties of the
molecules employed in these applications."2 The study of these fundamental
properties has a far reaching impact in the area of quantum chemistry and
Spectroscopy. The rich photochemistry and photophysics of transition metal
complexes has attracted researchers for many years, with the oldest studies being
performed primarily on ionic solids, such as chromium(III) impurities. The

emission spectra of these salts were described by Becquerel in 1867,3 whose

..Q '—

photochemical work on chromium salts was essential in the development of
photoengraving and lithography techniques.4 Since then, chromium(III) has been
extensively studied and characterized both photophysically and
photochemically,“6 and therefore makes an ideal probe for answering fundamental
questions about the electronic structure of transition metal complexes.

After the work of Becquerel the ﬁeld essentially lay dormant until 1940
when Van Vleck analyzed the absorption spectra potassium chrome alum
(KCr(SO4)2 -12 H20) in terms of crystal ﬁeld theory.7 Many more studies of this
type were carried out, most notably Sugano and Tanabe’s extensive study of
Cr(III) in A1203.8 Later on, after the advent of ligand ﬁeld theory in the early

ﬁfties, many studies of transition metal complexes in solution were carried out, but
it was not until the early sixties that the luminescence of a chromium(III) complex,

Cr(acac)3 (where acac is the monodeprotonated form of acetylacetone), was ﬁrst
reported by Forster and DeArmond.9 The assignment of the low-energy narrow
lineshape emission as originating from the lowest-energy 2E state was based in
part on single-crystal polarized absorption measurements on Cr(ox)3 (ox =
oxalate) carried out the previous year by Piper and Carlin,IO who also later carried
out the ﬁrst polarized single-crystal spectrum of Cr(acac)3.ll Many other studies
were carried out, most notably by Forster and DeArmond,12’13 as well as

theoretical advances such as the development of nonradiative decay theory that

began to lead to an understanding of the dynamical processes occurring in these
complexes.

Of course, technological advances also play a role in this story, most
notably with the development of the ruby laser by Ted Maiman in 196014—a
technological feat that not only changed the course of spectroscopy but also
ignited a large amount of interest in Chromium(III) photophysics (the gain
medium, ruby, is chromium doped corundum—a form of aluminum oxide).
Several decades afterward, in the 1990’s, spectroscopy was again fundamentally
altered with the advent of ultrafast spectroscopy, which enabled the scientist to
observe chemical and photophysical events on the lifespan of molecular

vibrations. Many early studies focused on organic's'm

or fully inorganic systems,
such as Zewail’s gas-phase experiments on iodine and salts of iodine.”26 With the
development of a dye-sensitized solar cell in 1991 employing nanoparticle TiOz
by Gratzel, which employed complexes of ruthenium(II) as the dye species,

7’28 In 1996, it was shown that

interest in transition metal photophysics expanded.2
electron injection into the conduction band of a dye-sensitized solar cell occurred
with T <- 500 fs.29 At this time ultrafast transition metal photophysics became
interesting not only from a purely scientiﬁc viewpoint, involving challenges in
spectroscOpy and theory, but also in the realm of applications. [Ru(bpy)3]2+,
tris(2,2’-bipyridine)ruthenium(II), eventually became the paradigm for ultrafast

spectroscopy of transition metal complexes, and has been extensively studied.”35

U

While the excited state dynamics of second and third row transition metal
complexes are concerned almost entirely with charge-transfer states, complexes of
ﬁrst-row transition metal elements have ligand-ﬁeld based states as their lowest-
energy electronic state. Therefore, unlike most other studies of excited state
processes, the dynamics of the complexes presented herein are occurring entirely
in the ligand ﬁeld manifold, i.e. only d-orbital based multielectronic
wavefunctions play a role in the observed dynamics. Up to this point, the extent of
published ultrafast spectroscopic data of chromium(III) complexes has been

conﬁned to a handful of studies on tris(acetylacetonato)chromium(III),36’39 as

described in chapter 3, as well as some photochemical'w’4| and donor-acceptor
studies.42 Some unpublished results are also relevant, which are reviewed in
chapters 3 and 4.38 Indeed, the ﬁeld of ultrafast dynamics of ﬁrst row transition

metal complexes remains largely uncultivated.43’44

1.3 Electronic Structure, Kinetics, and Application of Nonradiative Decay
Theory to Complexes of Chromium(III).

1.3.1 Electronic Structure. Chromium(III) complexes of high symmetry are
ideal for the study of photophysics and photochemistry of transition metal
containing systems due to the simplicity of the ligand ﬁeld manifold in an Oh

environment (compare, for example, the Tanabe-Sugano diagrams of d3 and d6

transition ions, Appendix A).45 Furthermore, the wealth of extant literature on

the photophysical propertiess’f"46

of this ion provides the researcher with an
invaluable resource for evaluation and context in which to place one’s results.
The Tanabe-Sugano diagram for a (13 species in an octahedral environment

is shown below in Figure 1-1. Using the common “one electron” molecular orbital

representation, the relevant electronic states of chromium(III) are highlighted.

 

 

 

 

 

 

  

 

 

80 4T‘9 2A29
* 1
e _. _
70. 2 9 4
60. 4T19 29 — - ‘—
501 4T
m 29 e * _ _
m 4° 9 1 L 1 ’E
30 2T29 129 — _ '—
T1
2 20' / L 2E9 *
10‘ t 1 1 1 A2
0F 4A 29 _ _ _
. , - . 2
o 10 20 30 40 50 g
Ao/B

 

 

 

Figure 1-1: Tanabe-Sugano diagram for a (13 ion in 0 symmetry.47 One-electron
representations of the relevant electronic levels in the ligand-ﬁeld manifold of the

Chromium(III) ion in an Oh environment.

This representation is strictly not correct because the electronic states of any (1 > 1
species are in fact multielectronic wavefunctions, but this formalism remains
useﬁil for gaining a qualitative understanding of the relevant electronic states.
Chromium(III) is a (13 ion, with a 4A2 ((tzg)3 in an inﬁnitely strong ﬁeld) ground
state. The free-ion (no imposed crystal or ligand ﬁeld potentials) ground state
term is 4F, which splits into the aforementioned 4A2 state as well as the ﬁrst spin-
allowed excited states, 4T2 or 4T1, under pseudo-octahedral symmetry. Transition
from the ground state to the low-lying quartet excited states corresponds to the
orbital (“one electron”) transition (tzg)3 —> (tzg)2(eg*). The ﬁnal quartet ligand ﬁeld
state, the upper lying 4T1, derives from the 4P term. Repulsion, which can occur
between states of the same irreducible representation, occurs between this state
and the lower-lying 4T1 (derived from 4F), leading to the non-linear energy of
these electronic states as a function of the ligand ﬁeld strength.

As can be inferred from the Tanabe-Sugano diagram, in the majority of
chromium(III) complexes the 2T1 States lies about 500 cm'] above the 2E state, so
that the states can be treated kinetically as a single state, which we will call 2E.46
In general, if the symmetry of the molecule is O (the pure rotational subgroup of
Oh, so the subscripts g and u can be dropped) or can be approximated as such, two

cases can be considered: 1) if the energy of the intraconﬁgurational spin ﬂip is

less than 10 Dq (the ligand ﬁeld strength), then 2E lies below 4T2, or 2) the ligand

 

 

ﬁeld strength is small enough so that the energy of the spin ﬂip exceeds the ligand
ﬁeld strength, and 2E lies above 4T2.

Considering a chromium(III) ion under the inﬂuence of an octahedral
ligand ﬁeld, one can see that from simple molecular orbital considerations that
formation of any quartet ligand ﬁeld state must result in antibonding metal-ligand
character to be introduced. In an excellent review, Kirk describes the effect of
promotion of an electron to the eg“ set (Figure 1-2), speciﬁcally dxy —+ dx2-y2, dxz
-—> dzz-x2, and dyz —> d22-y2 in the following manner: [the transition effectively]
“constitutes a rotation of charge distribution by 45° in one or another of the three
orthogonal planes containing the ligands. Because of the antibonding electron
density on two of the the Cr-L bonding axes in the quartet excited state, relaxation
will occur to a new geometry; a tetragonal distortion is suggested...some theories
have allowed for trigonal distortions.”6 That said, it is assumed from this model
that geometry distortions with respect to the ground state in the
intraconﬁgurational 2E state are negligible. The small geometrical change in this
state with respect to the ground state is in fact manifested by the narrow emission
spectrum from the 2E state (discussed in Chapter 3).

In 1978, Wilson and Solomon’s high-resolution polarized single-crystal
spectroscopic study of hexaamminechromiumﬂll), and their tour-de-force
application of ligand ﬁeld theory allowed for an estimation of the extent of the

J ahn-Teller distortion in the 4T2g state of this complex.48 They found, in

 

 

 

 

 

 

Figure 1-2: Charge redistribution in d3 system as a result of promotion to an eg"

 

 

 

 

 

type orbital.6
F NH3 ‘ 3+ _ 3+
NH3 1
H3N///,,h(:| “\\\\\NH3 H3NH’Hllu, C; .\‘\\\l\\\\NH3
. r' ,_ .. r'
H3N/ l \NH3 H3N/ l \NH3
_ NH3 4 L NHa J
4 A (Oh) 4T2g (OI1 notation, D4}l symmetry)
2g 12 pm lengthening of equatorial bonds
2 pm Shortening of axial bonds

 

 

 

Figure 1-3: Excited state (4T2) distortions of hexaamminechromium(III), as

' determined by Wilson and Solomon.48

accordance with the Simpliﬁed picture presented above, that the equatorial
chromium-nitrogen bonds lengthened by 12 pm, while the axial bond lengths are
shortened by 2 pm from the ground state (and presumably 2E) geometry Of 206
pm, representing a ~6 % change in the equatorial bond length. Furthermore, the

authors point out that their study required a low temperature single crystal, and

 

that the magnitude of the excited state distortions may increase in a solution
environment, as vibrational studies in the ground state have Shown a 5-10%

49,50

decrease in the force constant in the solution phase. In an octahedral system,

distortions of both the A1 g and Eg normal modes (Figure 1-3) contribute to the
excited state geometry distortions.

Forster provides an excellent overview of various kinetic processes that can
occur in chromium(III) systems.“46 Upon excitation into the Frank-Condon state,
only a handful of kinetics processes can ensue to provide relaxation back to the
ground 4A2 state. A J ablonski diagram of the kinetic processes that can occur
within a photochemically stable chromium(III) species wherein the energy of the
2E state is below that of the 4T2 is shown in Figure 1-4. Upon excitation into the
4T2 state a variety of processes ensue which dissipate the absorbed energy. From
the Frank-Condom state the lone radiative mechanism is ﬂuorescence (FL),
emission between states of the same spin multiplicity. The nonradiative
mechanism of energy dissipation from this state include internal conversion (1C),
which is a nonradiative decay mechanism between states of the same spin
multiplicity, and intersystem crossing (ISC) an isoenergetic process between
electronic states of different spin multiplicity. Fluorescence and internal
conversion both lead to ground state formation, while intersystem crossing results
in the formation of the 2E state, which is generally long-lived. If enough thermal

energy is present and the 4T2 and 2E states are close in energy, back intersystem

crossing (BISC) can occur. In a system where the 4T2 lies lowest in energy the
system will undergo internal conversion from this state to repopulate the quartet
ground state.

The 2E state can decay via phosphorescence (PH), a radiative emission
between states of different spin multiplicities, or by ISC into the ground 4A2
manifold. These are the various processes that occur between different electronic
states. However, one must keep in mind that other nonradiative events
(vibrational cooling (VC), redistribution of vibrational energy) are occurring
within the electronic state before the formation of the thermalized, metastable

state. These processes will be discussed at great length later.

10

ll

 

 

.1;
NH
(7)
O
'<
O

 

l

5

Energy

 

 

(up————-———-———————

 

 

 

 

 

 

----- nonradiative decay: internal
conversion, intersystem
crossing, vibrational cooling

radiative decay: ﬂuorescence,
phosphorescence

 

 

 

Figure 1-4: Jablonski diagram of photophysical processes that occur in a (13 ion

under 0 symmetry.

Finally, while various photochemical events can quench radiative (FL, PH) and
nonradiative (IC, ISC, VC) processes from the excited electronic states, we will
concern ourselves only with photophysical processes, i.e. those that cause no
chemical change to the system. This is, in fact, a good approximation given the
low photochemical quantum yields for the complexes employed in this study,51

although the researcher must be vigilant not to discount these mechanisms even in

11

seemingly photostable complexes. The rates of excited state decay are intimately
tied into the nuclear displacement (Q) and relative energies (E0) of the potential
energy surfaces of the two states in question. All of these factors are dealt with

within the formalism of nonradiative decay theory.

1.3.2 Nonradiative Decay Theory. As described in the previous section, when a
system absorbs a photon, the ensuing processes that relax the molecule back to the
ground state can be classiﬁed as either radiative (ﬂuorescence, phosphorescence)
or nonradiative (vibrational cooling, internal conversion, intersystem crossing,
quenching mechanisms). Nonradiative and radiative dynamics can be summarized

by F ermi’s “Golden rule” (equation 1.1).
Zﬂ' (1) 2
k = —h—|H I p(E) (1.1)

P12

In this equation quantiﬁes the coupling of the participating vibrational

and electronic wavefunctions and p(E) is a density-of-states term. Starting from
F ermi’s Golden rule, nonradiative decay theory was developed in the 1960’s and
70’s through the work of Jortner, Freed, El-Sayed, and others.52‘53 An eventual
goal of this research program is ascertaining whether this theory is generally
applicable to ultrafast processes of transition metal systems.

A general treatment of nonradiative decay theory is presented here, and

several speciﬁc examples are presented later in the text. The Bom-Oppenheimer

12

approximation allows one to decouple the electronic and nuclear components due
to the relative timescales on which they operate (this approximation has been
shown to breakdown for some ultrafast processes where strong vibronic coupling
is implicated, as discussed later in this dissertation). The resulting total
wavefunction for any electronic state is then the product of the electronic part, 01,

and all 3N-6 vibrational modes, x. We consider two states, the initial (promoting)

and ﬁnal (accepting):
3N —6 l
w1=w?(r)nz§)(Q) 0.2)
i=1
3N -6

W2 =w§(r)1—1[z§3>(g) ‘ 03)

Applying the Golden Rule of Fermi, the nonradiative rate takes the form

2
k

n

27:
r - 7- p(E) (1.4)

 

 

* /\
Jl/lz H l/lldT

Note that the coupling term now takes the form of a transition moment integral.
AS the Bom-Oppenheimer approximation demands that the electronic and
vibrational wavefunctions are distinct, the operator is expanded to include separate

electronic and nuclear operators

H = HElec+ HVib (1 -5)

l3

where H E186 operates only on the electronic terms and 1A1 Vib operates on the
vibrational components. By putting the wavefunctions and operator in the full
form and rearranging, one obtains a form that includes both electronic and
vibrational transition moment integrals:

N—6
kn, =— 2” hngomrzecwldr J) H Z(Z)(Q)HVb H 1,.(1)(Q)dr

2

 

This equation can be simpliﬁed by eliminating non-participating terms, or by
considering a single accepting mode, continuum of modes, etc. Manipulation and
determination of this result is the work of all the theory on nonradiative decay
(again, except for those cases where the Born-Oppenheimer approximation fails).
The driving force for nonradiative decay can be understood by examining
equation 1.6, the basic equation governing this phenomenon. For nonradiative
decay the vibrational overlap between the participating states, called Frank-
Condon factors, which relate to the geometry of the system, is an important factor
determining the rate. The rate is proportional to the matrix element of the

vibrational F rank-Condon factors:
knr 0C <xrlzf> (1.7)
Where I i corresponds to the vibrational mode (or modes) from which the

transition is originating, the so-called promoting mode, and I f is the mode (or

14

,0(E) (1 -6)

modes) to which the energy is transferred, called the accepting mode. The
vibrational overlap, in turn, depends on two different factors, the relative nuclear
displacements of the electronic states (AQ) and the difference in the zero-point
energies between the two states (AE). To understand this, we will employ some of
the terminology from Marcus’s theory of electron transfer and describe two
limiting cases of AQ positions, the “normal” region and the “inverted” region,
where changing AE while maintaining the same nuclear displacement results in

two opposing trends in the rate of km.

In the following discussion we will consider a single promoting mode and a
single accepting mode, however theories have been developed which consider
multiple promoting and accepting modes. The vibrational wavefunctions under
consideration are the lowest vibrational component of the upper state (the
promoting mode—assuming that this upper state is thermalized), and the
isoenergetic vibrational wavefunction of the lower potential (accepting mode).
Figure 1-5 shows two harmonic potentials corresponding to electronic potential
wells with the vibrational component wavefunctions superimposed on the
potentials. The upper potential is displaced with respect to the lower state along
the nuclear coordinate, Q. As the energetic separation of the two potential wells
decreases, the extent of vibrational overlap between the lowest component of the
upper well and the isoenergetic level of the lower well decreases; a concomitant
decrease in the rate of nonradiative decay between these two states results

according to equation 1.7. This describes the situation of the “normal” region,

15

 

which arises due to the greater amplitude on the edges of the potential for higher-
energy component vibrational wavefunctions.

Figure 1-6 again shows two harmonic potentials, but now with minimal
relative displacement along the nuclear coordinate axis. In this case the dearth of
amplitude of upper lying vibrational wavefunctions in the center of the potential
leads to a poor overlap between the lowest-lying vibrational wavefunction of the
upper state and isoenergetic levels of the lower state. AS the energetic separation
is reduced, vibrational overlap increases, and an associated increase in km occurs
according to equation 1.7. This situation is descriptive of the “inverted” region.

Finally, another general concept of nonradiative decay theory is the role of
entropy. Entropy plays an important role in driving nonradiative decay: the term

p(E) of equation 1.1 quantiﬁes the density of states in the system. A ﬁnal state

that has a greater density of vibrational levels will entropically drive the transition,
while under similar conditions a ﬁnal state with a more dilute manifold of
vibrational states will have a correspondingly lower nonradiative rate. This

concept was proposed early on in the development of the theory?"55

The role of conformational changes (modifying AE and AQ) between the
ground and excited states on the rate of nonradiative decay has been studied
extensively. Some of the earliest studies on nonradiative dynamics involved
monitoring the formation of the ground state via the lowest-energy excited state, a

triplet state, in hydrocarbon systems. The majority of these systems operate in the

16

 

Decrease

 

 

Q Q

 

 

 

Figure 1-5: Normal region—as AB is decreased the rate of electron transfer
decreases as a result of decreased vibrational overlap between component

vibrational wavefunctions.

 

E .- ; Decrease E _
, . AB 3

AE ,. ‘\ _

IAE

 

 

 

 

v
‘1

Q Q

 

 

 

Figure 1-6: Inverted region—as AB is decreased the rate of electron transfer
increases as a result of increased vibrational overlap between component

vibrational wavefunctions.

l7

 

 

inverted region; a paper published in 1966 by Siebrand56 explains “a crude

correlation between t and the triplet energy ET has been noted and traced back to

9

the Franck-Condon factor F(oc l/ r) of the transition...’ A year later, through a

series of deuterated hydrocarbons, Siebrand et.al. showed that the nonradiative

decay from the triplet state occurs efﬁciently through the C-H (C-D) modes.57

58,59

Later studies expanded the theory for hydrocarbons and there are also

noteworthy examples of studies of aromatic systems.60 Since these early studies,
the general concepts outlined above been applied to a broad range of chemical
systems for nonradiative transitions between the emissive, lowest-lying excited
state and the ground state. Several examples relevant to transition metal
photophysics are presented below.

Interesting examples of the role of conformational changes can be found in
organic and biological systems, where the phenomenon of ultrafast internal
conversion, radiationless decay between electronic states of the same spin
multiplicity, is particularly relevant. This has been observed in such biological
molecules as DNA and RNA nucleosides and the green-ﬂuorescent protein (GFP)
chromophore.61 Kohler and coworkers62 found excited state lifetimes under 1 ps
for all nucleosides, and those nucleosides with faster internal conversion rates had
a lower propensity toward photochemical damage. They suggest that this property
likely played an essential role in the early evolution of life on earth. The GFP

chromophore is composed of two halves, and semi—empirical quantum mechanical

18

44“ l

calculations suggest that the two halves are planar in the ground S0 state, while
they are perpendicular in the 81 state.63 This large conformation change means the
SO and 81 potential wells are displaced along the coordinate corresponding to the
axis of rotation, and correspond to the “normal region.” This displacement of the
wells results in a much larger Frank-Condon overlap (vibrational overlap) and thus
a much more efﬁcient internal conversion than if the wells were in the Marcus
inverted region. This ultrafast decay is responsible for the stability and low
quantum yields for photochemistry. This suggests a means to controlling
photochemical and photophysical events by controlling the environment, and thus
the relative displacements of the electronic states involved, and ultimately the
F rank-Condon overlap. In fact, these large conformational changes in organic and
biological systems, and the associated rapid nonradiative rates, may be the bridge
between organic photophysics and transition metal photophysics, where large
conformational changes are quite common.64

The theory is expanded by considering the so-called “weak coupling limit.”
In this limit the nonradiative decay rate is the product of the electronic ([30) and

vibrational overlap (F, Franck-Condon) factors (equation 1.8).

 

knr =IBOF (1'8)

[30 = Cfrok % (1.9)

F=ZHI<ZfIZI>2 (1.10)
19

-m“ g-

The electronic factors includes the promoting mode, 00K, from which the transition
originates, and a constant Ck, which includes contributions from vibronic
coupling, which acts to make formally symmetry-forbidden (LaPorte forbidden)
transitions allowed, and spin-orbit coupling, which increases the allowedness of
formally spin-forbidden transitions. For an intersystem crossing (spin-forbidden)
event a non-zero value of BO is obtained only if spin-orbit coupling contributions
are considered. The vibrational factor, F, is accepting-mode dependent and will
take on different forms depending on the approximation: is there a Single

53,65

accepting mode, a continuum of modes, or a ladder of modes? Note that

equations 1.8 through 1.10 effectively constitute the quantitative result of equation
1.7.

A popular model that has found success in describing nonradiative rates
between lowest-energy excited states and ground states is the so-called polaron
(also known as the spin-boson) model. In this model, many vibrational states are
playing a role in the nonradiative transition and the following equation for the

vibrational overlap factor, F, results:46

 

1 1/2 E 1 2(AV )2
F=£ ] exp —SM——y——+[L+—] —1/2— (1.11)
hmME th hmM 16m2
y=ln[—E—]—l (1.12)
SthM
20

1:1

In equations 1.11 and 1.12 E is the energy separation between the two states, SM is
the Huang-Rhys factor, which describes the displacement of the potential minima
of the promoting and accepting electronic states along the accepting mode
coordinate, 00M is the dominant accepting mode, and A111 )2 is the full width at half-
maximum of the emission spectrum. In the context of electron transfer one simply

needs to replace E with the corresponding quantities familiar to the theory of

electron transfer :66’67
_ 0
E—jAG l—AO (1.13)
2
__(AVV2) = 301.37 (1.14)
16ln2

Equations 1.13 and 1.14 form of the “Energy Gap Law” a limiting case of
nonradiative decay theory in the inverted region, which predicts a linear

relationship between the energy gap (E) and In km. Meyer et al. have conﬁrmed

the “Energy Gap Law” between the emissive state and the ground state in several

68.69

series of substituted Os(II), Ru(II), and Re(I) complexes. Employing the spin-

boson model, and considering only one vibrational mode ((0) and an equilibrium

displacement (AQ), the reorganizational energy for this mode is:
f 2
2.5% (Aqe) (1.15)

Where f = ,uco2 is the force constant. This is related to the Huang-Rhys factor,

also called the electron vibrational coupling constant, a dimensionless quantity

21

that takes into account the equilibrium nuclear displacement (AQ) and the

reorganization energy:

2 2
S: ’11“ =f(Aqe) =”w(Aqe) (1.16)
km 21160 272

 

Meyer et. al. have used the concepts outlined above to ﬁt emission spectra of
Ru(II) and Os(II) polypyridyl complexes. From these ﬁts, which utilize concepts
of nonradiative decay theory, they were able to obtain kinetic information on these

7074
molecules.

1.3.3 Applications of Nonradiative Decay Theory to Complexes of
Chromium(III). The ﬁrst theoretical application of nonradiative decay theory
speciﬁcally to transition metal complexes was carried out by Robins and Thomson
in 1973,75 wherein they applied a qualitative, symmetry-based approach to
nonradiative decay theory to describe the nonradiative 2E —> 4A2 conversion in a

series of Chromium(III) complexes previously studied by Forster and coworkers.
The majority of theoretical work in the ﬁeld up to that point was concerned with

organic systems, however this theory had been applied in several papers—’6’78 to

transition metal systems, with varying success. The approach adopted by Robins
and Thomson was motivated by the inherent high symmetry of many metal
complexes, such that symmetry-based selection rules likely play a large role in the

coupling terms affecting the rates of nonradiative relaxation. This concept arose a

22

few years prior via the work of Gardner and Kasha,79 who suggested that
molecules that display slow radiationless decay are “vibrationally deﬁcient,”
meaning that the molecules lack promoting and accepting modes of the same
symmetry to facilitate rapid nonradiative decay. Using this symmetry based
approached, they determined for octahedral and pseudo-octahedral complexes 1.)
that metal-ligand modes are likely not active in 2E ——r 4A2 nonradiative conversion
and 2.) that the rate of nonradiative decay was linearly dependent on the number
of hydrogen atoms attached to the diketonate skeletal framework: the more
hydrogen atoms bound directly to the 71: system of the ligand, the faster the rate of
intersystem crossing. The authors also note that comparison with systems that
have aliphatic ligands suggests that coupling to the 7t system leads to more
efﬁcient nonradiative decay. This result is likely not general for state changes in
transition metal systems, and probably reﬂects the intraconﬁgurational nature of
the state change that they were describing, where both states can be described in
terms of orbitals of 1: symmetry. These symmetry-based selection rules were later

. . . . . . 4 4 . .
applied to describe lntemal converSIOn (radiationless T2 —a A2 converSIOn) In

various Chromium(III) doped glasses.80

From the 19708 onwards many studies appeared which attempted to address
the mechanism of decay of the lowest energy excited state in simple
chromium(III) systems. In general, at low temperatures (< 100 K), the relaxation

was insensitive to the matrix and temperature, however various studies showed

23

- _.-.;._.!

that the decay depended on high frequency vibrations of ligated atoms, spin-orbit

coupling, as well as low-frequency modes.“83

A separate regime of dynamics
was found at higher temperatures, where dynamics vary with temperature and

solvent, such that the decay of the 2E state is given by:
k=kLT+kHT(T) (1.17)

Where kLT = kr + km and kHT(T) is the temperature dependent additional dynamics
observed at higher temperatures.84 At the time, researchers were attempting to
determine a uniﬁed model for this so—called “thermally activated relaxation,” and
three mechanisms were put forth to account for the decay of the 2E state: 1.)
quenching of the excited doublet state by direct chemical reaction, 2.) back
intersystem crossing to a low-lying quartet state, which can undergo internal
conversion to form the ground state or 3.) crossing to the potential energy surface
of a “ground state intermediate,” facilitated by low-frequency solvent and/or
normal modes of the molecule (Figure 1-7).85 Many studies were interested in
determining the dominant mechanism in various systems; most of the major
studies of this time employed am(m)ine complexes of chromium(III). Early on,
quenching of the 2E state by direct reaction was the favored candidate for the
major relaxation pathway for most complexes of this type. This arose from studies
of trans-Cr(NH3)4XY and trans-(Cr[14]aneN4)XY (where X and Y are simple

ligands such as SCN, CN and NH, and [l4]aneN4 is 1,4,8,11-

tetraazacyclotetradecane).86'88 These studies showed high yields of

24

photosubstitution for the trans-Cr(NH3)4XY type complexes but very low yields
of photosubstitution for the trans-(Cr[l4]aneN4)XY complexes. This was
attributed to the closed ring structure of the [14]aneN4 ligand (i.e. cyclam, Figure
1-8), which ostensibly prevented photosubstitution at the equatorial coordination
Sites, bolstering support for the direct reaction quenching mechanism. Support for
the mechanism wherein state crossing was facilitated by low frequency modes
came mainly from variable temperature/solvent studies, which showed that
freezing of the skeletal vibrations of the molecule, that apparently acted as
promoting modes, hindered decay of the 2E state.89 This question was ultimately
addressed by Ramasami et al. in a study where [Cr(en)3]3+ and [Cr(sep)]3+ (sep =
(S)-1,3,6,8,10,13,16,l9-octaazabicyclo[6.6.6]eicosane, see Figure 1-8) were
compared.85 The sep ligand fully encapsulates the chromium(III) ion, so direct
reaction is completely discounted. If the direct reaction mechanism was the
dominant mechanism for 2E decay then this complex would have a very long 2E
lifetime relative to the electronically similar [Cr(en)3]3+. Furthermore, the authors
note that the back intersystem crossing mechanism is anticipated to be highly
inefﬁcient in this system because of the large energy gap. The authors found that
the lifetime of the 2E state of [Cr(sep)]3+ was only slightly longer than that of
[Cr(en)3]3+, making direct reaction an unlikely candidate and supporting
intersystem crossing as a deactivation pathway. Endicott et al. later noted that the

same mechanism, namely coupling of low-frequency modes, would account for

25

 

4T2

2E

4A2

Energy

nr

intermediate

 

 

 

V

Nuclear coordinate

 

 

 

Figure 1-7: Proposed mechanisms for thermally activated 2E state deactivation,

reproduced from reference 84. (a. = direct reaction, b. = back intersystem crossing

into the quartet manifold and c. = surface crossing to a ground state intermediate.)
both the direct reaction deactivation as well as intersystem crossing to the ground

state intermediate, and relative contributions of each pathway are determined by

nuclear conﬁguration.90 In this sense, they added that the direct reaction pathway

26

 

should be considered a limiting case of a mechanism involving deactivation
promoted by low frequency normal modes or solvent modes.
The role of stereochemistry in these thermally activated relaxation events

91,92

was ﬁrst proposed by Kane-Maguire et al. Theoretical aspects of this were

studied by Vanquickenbome and coworkers.93

They showed that trigonal
distortions, which lower the symmetry of the system, mix d—orbitals creating
microstates of doubly ﬁlled d-orbitals. These doubly ﬁlled d-orbitals decrease
electronic repulsion in the excited state, thus providing a facile mechanism of
achieving the ground state electronic conﬁguration. Later, experimental evidence
began to arise which implicated trigonal distortions as playing an important role in
facilitating intersystem crossing: these studies compared amine complexes to
analogous constrained amine ligands, mostly derivatives of 1,4,7-
triazacyclononane (TACN) (see Figure 1.8). 82’84’94'96

In the 1990s, as a forerunner to this dissertation, ultrafast spectroscopy was
beginning to be applied to the study of chromium(III) photophysics.”38 These
studies, which were carried out with ~ 100 fs optical pulses, are reviewed
extensively in chapters 4 and 5. For the archetypal complex Cr(acac)3 it was
found that intersystem crossing between the ﬁrst spin-allowed 4T2 state and
lowest-energy 2E occurred with kisc > 10”, and an ~ 1 ps lifetime was observed

which was assigned as vibrational cooling within the 2E state. These dynamics

represented a new observation in ﬁeld of chromium(III) photophysics: one of the

27

 

rapidly evolving, non-thermalized state, explored in great detail in Chapter 3 of

this dissertation.

(1:, H. [NH a] (Ni. .1...)

 

2”§z

 

NH HN NHNHHN
TACN N
Cyclam Sep

//\N
:17
Eyf HN ”N

:8 for $21...

H2N

NH
Sen

TAE[9]aneN3
([9]aneN3CH2-)2

 

 

 

Figure 1-8: Sterically constraining ligands that helped to elucidate modes of 2 E

deactivation.

28

1.4 Dissertation Outline

The work presented herein will include studies on a variety of Chromium(III)
complexes, including n—delocalized ligand systems, low-symmetry ligand ﬁelds,
and spin exchange systems. The aim of this work is to fully characterize the
electronic, magnetic, and geometrical structures of these complexes with the goal
of correlating these structural changes to the observed excited state dynamics,
which are nonradiative in nature. Studies on complexes of gallium(III), an
effective analog of chromium(III) which provides useful information in the
absence of unpaired Spin, are also presented. Finally, Heisenberg spin exchange
complexes of chromium(III) are explored. The outline of this dissertation is as
follows:

- In Chapter 2, the electronic and magnetic structures of the ground states of
various systems are explored via electron spin resonance techniques. Gallium
semiquinones will be explored and issues relevant to understanding and
controlling spin distribution in such systems will be discussed. The ground state
magnetic structures of quartet complexes of chromium(III) will also be explored,
and an extensive investigation of the effect of zero ﬁeld splitting on the
appearance of spin resonance spectra is presented.

- Chapter 3 focuses on the electronic structure and excited state dynamics of high-
symmetry complexes of chromium(III). Spectroscopic techniques, both static and
time-resolved, are employed to characterize the excited electronic structure and

dynamics therein. This work includes ultrafast transient absorption results on

29

high-symmetry Chromium(III) complexes, which aims to address fundamental
questions vis-a-vis mechanisms of nonradiative decay in these systems. These
studies were carried out with various time resolutions, employing ultrafast optical
pulses typically of 100 fs duration.

' Chapter 4 explores the electronic structure and dynamics of spin exchange
complexes of chromium(III), employing the same techniques as those of Chapters
3 and 4. The results of the previous chapters are employed to aid in the
characterization of these systems.

' Chapter 5 highlights future work.

30

 

1.5 References

(1)

(2)
(3)

(4)

(5)
(6)
(7)

(8)

(9)

(10)
(11)

(12)
(13)
(14)
(15)

(16)

Hagfeldt, A.; Gratzel, M. Accounts of Chtmical Research 2000, 33, 269-
277.

Meyer, T. J. Accounts of Chemical Research 1989, 22, 163-170.

Becquerel, E. La lumiere, ses causes et ses effects; Gauthier-Villars: Paris,
1867.

Chromium in the natural and human environments; Wiley: New York,
1988; Vol. 20.

Forster, L. S. Chemical Reviews 1990, 90, 331-353.
Kirk, A. D. Chemical Reviews 1999, 99, 1607-1640.

Finkelstein, R.; Vleck, J. H. V. Journal of Chemical Physics 1940, 8, 790-
797.

Sugano, S.; Tanabe, Y. Journal of the Physical Society of Japan 1958, 13,
880-899.

Forster, L. S.; DeArmond, K. Journal of Chemical Physics 1961, 34, 2193-
94.

Piper, T. S.; Carlin, R. L. Journal of Chemical Physics 1960, 33, 608-609.

Piper, T. S.; Carlin, R. L. The Journal of Chemical Physics 1962, 36, 3330-
3332.

DeArrnond, K.; Forster, L. S. Spectrochimica Acta 1963, 19, 1403-1406.
DeArmond, K.; Forster, L. S. Spectrochimica Acta 1963, 19, 1687-1693.
Maiman, T. H. Nature 1960, 187, 493-494.

Bakker, H. J .; Planken, P. C. M.; Lagendijk, A. Journal Of Chemical
Physics 1991, 94, 6007-6013.

Akesson, E.; Walker, G. C.; Barbara, P. F. Journal Of Chemical Physics
1991, 95, 4188-4194.

31

 

 

 

(17) Elsaesser, T.; Kaiser, W. Annual Review Of Physical Chemistry 1991, 42,
83-107.

(18) Tominaga, K.; Walker, G. C.; Jarzeba, W.; Barbara, P. F. Journal Of
Physical Chemistry 1991, 95, 10475-10485.

(19) Willberg, D. M.; Breen, J. J.; Gutmann, M.; Zewail, A. H. Journal Of
Physical Chemistry 1991, 95, 7136-7138.

(20) Bowman, R. M.; Dantus, M.; Zewail, A. H. Chemical Physics Letters 1989,
161, 297-302.

(21) Gruebele, M.; Roberts, G.; Dantus, M.; Bowman, R. M.; Zewail, A. H.
Chemical Physics Letters 1990, 166, 459-469.

(22) Breen, J. J.; Willberg, D. M.; Gutmann, M.; Zewail, A. H. Journal Of
Chemical Physics 1990, 93, 9180-9184.

(23) Potter, E. D.; Herek, J. L.; Pedersen, S.; Liu, Q.; Zewail, A. H. Nature
1992, 355, 66-68.

(24) Dantus, M.; Bowman, R. M.; Zewail, A. H. Nature 1990, 343, 737-739.

(25) Dantus, M.; Bowman, R. M.; Gruebele, M.; Zewail, A. H. Journal Of
Chemical Physics 1989, 91, 7437—7450.

(26) Dantus, M.; Rosker, M. J.; Zewail, A. H. Journal Of Chemical Physics
1988, 89, 6128-6140.

(27) O'Regan, B.; Gratzel, M. Nature 1991, 353, 737-740.

(28) Hagfeldt, A.; Gratzel, M. Accounts Of Chemical Research 2000, 33, 269-
277.

(29) Tachibana, Y.; Moser, J. E.; Gratzel, M.; Klug, D. R.; Durrant, J. R.
Journal Of Physical Chemistry 1996, 100, 20056-20062.

(30) Yeh, A. T.; Shank, C. V.; McCusker, J. K. Science 2000, 289, 935-938.
(31) Vlcek, A. Coordination Chemistry Reviews 2000, 200, 933-97 7.

(32) Saes, M.; Bressler, C.; Abela, R.; Grolimund, D.; Johnson, S. L.; Heimann,
P. A.; Chergui, M. Physical Review Letters 2003, 90.

32

(33)

(34)

(35)

(36)

(37)

(38)

(39)

(40)

(41)

(42)

(43)

(44)

(45)

(46)

Henry, W.; Coates, C. G.; Brady, C.; Ronayne, K. L.; Matousek, P.;
Towrie, M.; Botchway, S. W.; Parker, A. W.; Vos, J. G.; Browne, W. R.;
McGarvey, J. J. Journal Of Physical Chemistry A 2008, 112, 4537-4544.

Yoon, S.; Kukura, P.; Stuart, C. M.; Mathies, R. A. Molecular Physics
2006, 104, 1275-1282.

Damrauer, N. H.; Cerullo, G.; Yeh, A.; Boussie, T. R.; Shank, C. V.;
McCusker, J. K. Science 1997, 275, 54-57.

Juban, E.; McCusker, J. Journal of the American Chemical Society 2005,
127, 6857-6865.

Juban, E. A.; Smeigh, A. L.; Monat, J. E.; McCusker, J. K. Coordination
Chemistry Reviews 2006, 250, 1783-1791.

Juban, E., Ph.D. Thesis, University of California, 2006.

Macoas, E. M. S.; Kananavicius, R.; Myllyperkio, P.; Pettersson, M.;
Kunttu, H. Journal of the American Chemical Society 2007, 129, 8934-
8935.

Inamo, M.; Okabe, C.; Nakabayashi, T.; Nishi, N.; Hoshino, M. Chemical
Physics Letters 2007, 445, 167-172.

To, T. T.; Heilweil, E. J.; Duke, C. B.; Burkey, T. J. Journal Of Physical
ChemistryA 2007, 111, 6933-6937.

Neuman, D.; Ostrowski, A. D.; Mikhailovsky, A. A.; Absalonson, R. 0.;
Strouse, G. F .; Ford, P. C. Journal of the American Chemical Society 2007,
130, 168.

Kane-Maguire, N. A. P. In Photochemistry And Photophysics 0f
Coordination Compounds 1; Springer-Verlag Berlin: Berlin, 2007; Vol.
280, p 37-67.

Forster, L. S. Coordination Chemistry Reviews 2006, 250, 2023-2033.
Juban, E. A., PhD. Thesis, University of California, Berkeley, 2006.

Forster, L. S. Coordination Chemistry Reviews 2006, 250, 2023-2033.

33

(47)
(48)
(49)

(50)

(51)

(52)

(53)

(54)

(55)

(56)

(57)

(58)

(59)

(60)

(61)
(62)

(63)

Bethune, A. http://www.albion.edu/chemistry/abbethune, 2009.
Wilson, R. B.; Solomon, E. 1. Inorganic Chemistry 1978, 17, 1729-1736.
Schmidt, K. H.; Mueller, A. Inorganic Chemistry 1975, I 4, 2183-2187.

Nakamoto, K. "Infrared Spectra of Inorganic and Coordination
Compounds "; 2 ed.; Wiley: New York, 1971.

Fleming, C. A.; Thornton, D. A. Journal of Molecular Structure 1975, 25,
271-279.

Freed, K.; Jortner, J. Journal of Chemical Physics 1970, 52, 6272-6291.

Avouris, P.; Gelbart, W.; El-Sayed, M. A. Chemical Reviews 1977, 77,
794-833.

Douglas, A. E. Journal Of Chemical Physics 1966, 45, 1007.

Chock, D. P.; Jortner, J .; Rice, S. A. Journal Of Chemical Physics 1968,
49, 610.

Siebrand, W. Journal of Chemical Physics 1966, 44, 4055-4056.

Siebrand, W.; Williams, D. F. Journal of Chemical Physics 1967, 46, 403-
404.

Siebrand, W. Journal of Chemical Physics 1967, 47, 2411-2424.

Siebrand, W.; Williams, D. F. Journal of Chemical Physics 1968, 49, 1860-
1871.

Kistiakowsy, G.; Parmenter, C. Journal Of Chemical Physics 1965, 42,
2942.

Zimmer, M. Chemical Reviews 2002, 102, 759-781.

Pecourt, J .-M. L.; Peon, J .; Kohler, B. Journal of the American Chemical
Society 2001, 123, 10370-10378.

Voityuk, A. A.; Michel-Beyerle, M.-E.; Rosch, N. Chemical Physics
Letters 1998, 296, 269-276.

34

(64)

(65)
(66)

(67)

(68)

(69)

(70)

(71)

(72)

(73)

(74)
(75)
(76)
(77)

(78)

(79)

Fidder, H.; Rini, M.; Nibbering, E. T. J. Journal Of The American
Chemical Society 2004, 126, 3789-3794.

Bixon, M.; Jortner, J. Journal of Chemical Physics 1968, 48, 715-726.
Chen, P.; Meyer, T. J. Chemical Reviews 1998, 98, 1439-1477.

Barbara, P. F.; Meyer, T. J .; Ratner, M. A. Journal of Physical Chemistry
1996, 100,13148-13168.

Caspar, J. V.; Kober, E. M.; Sullivan, B. P.; Meyer, T. J. Journal of the
American Chemical Society 1982, 104, 630-632.

Caspar, J. V.; Meyer, T. J. Journal of the American Chemical Society 1983,
87, 952-957.

Murtaza, Z.; Zipp, A. P.; Worl, L. A.; Graff, D.; Jr, W. E. J.; Bates, W. D.;
Meyer, T. J. Journal of the American Chemical Society 1991, 113, 5113-
5114.

Murtaza, Z.; Graff, D. K.; Zipp, A. P.; Worl, L. A.; Jr, W. E. J .; Bates, W.
D.; Meyer, T. J. Journal of Physical Chemistry 1994, 98, 10504-10513.

Liang, Y. Y.; Baba, A. 1.; Kim, W. Y.; Atherton, S. J.; Schmehl, R. H.
Journal of Physical Chemistry 1996, 100, 18408-18414.

Kober, E. M.; Caspar, J. V.; Lumpkin, R. S.; Meyer, T. J. Journal of
Physical Chemistry 1986, 90, 3722-3734.

Caspar, J. V.; Meyer, T. J. Inorganic Chemistry 1983, 22, 2444-2453.
Robbins, D. J .; Thomson, A. J. Molecular Physics 1973, 25, 1103-1119.
DeArmond, M. K.; Hllis, J. B. Journal of Chemical Physics 1971, 54, 2247.
Mitchell, W. J .; DeArmond, M. K. Journal of Luminescence 1971, 4, 137.

Watts, R. J.; Crosby, G. A. Journal of the American Chemical Society
1972, 94, 2606.

Gardner, P. J .; Kasha, M. Journal of Chemical Physics 1969, 50, 1543.

35

 

(30)

(81)

(82)

(83)

(84)

(85)

(86)

(87)

(88)

(89)

(90)

(91)

(92)

(93)

(94)

Andrews, L. J .; Lempicki, A.; McCollum, B. C. Chemical Physics Letters
1980, 74, 404-408.

Kuhn, K.; Wasgestian, F .; Kupka, H. Journal Of Physical Chemistry 1981,
85, 665-670.

Endicott, J. F .; Lessard, R. B.; Lynch, D.; Perkovic, M. W.; Ryu, C. K.
Coordination Chemistry Reviews 1990, 97, 65-79.

Forster, L. S. Chemical Reviews 1990, 90, 331-353.

Perkovic, M. W.; Heeg, M. J .; Endicott, J. F. Inorganic Chemistry 1991,
30, 3140-3147.

Ramasami, T.; Endicott, J. F.; Brubaker, G. R. Journal Of Physical
Chemistry 1983, 87, 5057-5059.

Gutierrez, A. R.; Adamson, A. W. The Journal of Physical Chemistry 1978,
82, 902.

Kirk, A. D. The Journal of Physical Chemistry 1981, 85, 3205.

Kutal, C.; Adamson, A. W. Journal of the American Chemical Society
1971, 93, 5581.

Pyke, S. C.; Ogasawara, M.; Kevan, L.; Endicott, J. F. Journal Of Physical
Chemistry 1978, 82, 302-307.

Endicott, J. F.; Tamilarasan, R.; Lessard, R. B. Chemical Physics Letters
1984, 112, 381-386.

Kane-Maguire, N. A. P.; Crippen, W. S.; Miller, P. K. Inorganic Chemistry
1983, 22, 696.

Kane-Maguire, N. A. P.; Wallace, K. C.; Miller, D. B. Inorganic Chemistry
1985, 24, 597.

Ceulemans, A.; Bongaerts, N.; Vanquickenbome, L. G. Inorganic
Chemistry 1987, 26, 1566.

Ramasami, T.; Endicott, J. F.; Brubaker, G. R. The Journal of Physical
Chemistry 1983, 87, 5057.

36

(95)

(96)

Perkovic, M. W.; Endicott, J. F. The Journal of Physical Chemistry 1990,
94,1217.

Lessard, R. B.; Heeg, M. J.; Buranda, T.; Perkovic, M. W.; Schwarz, C. L.;
Yang, R. D.; Endicott, J. F. Inorganic Chemistry 1992, 31, 3091-3103.

37

Chapter 2: Ground State Electronic and Magnetic Structure of Gallium(III)-

Semiquinone Complexes and Quartet Complexes of Chromium(III)

2.1 Introduction

Rational design of molecular magnetic materials requires a thorough
understanding of all factors contributing to the electronic and magnetic structures
of the system. To this end, Guo et. al. have reported the synthesis and
spectroscopic properties of the [M1M2(tren)2(CA"')]m+ series, where M is
gallium(III) or chromium(III), tren = tris(2-aminoethyl)amine and CA"' is the
chloranilate anion, the bridging chelate between the two metal ions, which takes

681,081

on a tetraanionic dicatecholate (CA ) or trianionic semiquinone-catecholate

radical (CAsq’cat) form in this series (Figure 2-1).I Systematic incorporation of

chromium(III), (d3, S = % ground state) and/or gallium(III) (dm, diamagnetic)
coupled with either bridging form yielded six complexes (Figure 2-2). The initial
study of these complexes characterized this series via magnetic susceptibility,
cyclic voltammetry, steady-state emission, and static absorption. The magnetic

susceptibility data revealed a variety of magnetic ground states, ranging from the

singlet (S = 0) ground state of 1 to the sextet (5 = %) ground state of 6. This

chapter will present ground state characterization of the magnetically dilute, non-
spin-exchange complexes of this series utilizing electron spin resonance

techniques. These experiments use magnetic ﬁelds to break the degeneracy of

38

electronic states with unpaired spin. Once these states are split by the applied
magnetic ﬁeld, a probe is employed—either continuous-wave or pulsed
microwave radiation—to characterize the interaction of the unpaired spin with the
applied magnetic ﬁeld. Using this general picture many different electron spin
resonance techniques have been developed which allow one to characterize not
only the spin of interest but also interactions with magnetic nuclei. Further
additional spin resonance spectra are presented which are relevant to the high-
symmetry systems explored in chapter 3. Chapter 4 will explore the properties of
the spin-exchange members of this series.

This research is focused on gaining an understanding of the effects of
Heisenberg spin exchange interactions on the photophysical properties of a
molecule. Metal-quinone complexes are ideal for the study of physicochemical
properties of exchange-coupled molecules because the redox activity of the
quinone ligand effectively provides a mechanism for turning the exchange
interaction on or off: in this series one is able to selectively turn on spin exchange

with incorporation of the paramagnetic species chromium(III), and CAsq‘cat

, or turn
off exchange interactions by substituting in the diamagnetic analogs gallium(III)

and CAcat‘cat. In this manner the series of molecules provides the ability to

differentiate properties inherent to the constituents of the molecule from those

which arise due to exchange interactions between the constituents.

39

 

Cat, Cat Sq, Cat
Cl Cl Cl
'0 o‘ _ 'O O 'O o
-1 e
+1 e'
‘O O O O 'O O
Cl Cl Cl

 

 

 

 

 

caLcat

Figure 2-1: Bridging forms of chloranilic acid. CA refers to the diamagnetic

sq,cat

dicatecholate form, while CA is the semiquinone-catecholate form, and is

paramagnetic (S = 1/2). The resonance forms show that SOMO (singly-occupied

molecular orbital) density should be distributed among the ketone-like C-O

. . 2-4
mOIetles.

A logical starting point for the characterization of the magnetic ground
states of the molecules in this series is complex 2. This molecule is the most
magnetically simple paramagnetic member of the series, consisting of two
diamagnetic metal centers bridged by the semiquinone ligand radical CASq‘Ca’,
resulting in a total S = 1/2 for the system. The goal of the work presented in the

ﬁrst section of this chapter is the characterization of the magnetic ground state of

this complex via electron paramagnetic resonance (EPR) Spectroscopy.

40

 

 

 

 

 

 

 

 

7 Cl ‘ 2+
,0 0\ Mi =M2=Ga(III) <1)
(tren)Mt ,M2(tren) M1 = Ga(III), M2 = Cr(III) (3)
O 0 M1 = M2 = Cr(III) (5)

' C1 1 3+
)0 \ Ml =M2=GaaII> (2)
(tren)M1 ,M2(tren) M1 = Ga(III), M2 = Cr(III) (4)
O 0 M1 = M2 = Cr(III) (6)

(\NHZ
F
tren = C NH2
NHz
Figure 2-2: Members of the bimetallic [M1M2(tren)2(CA"')]m+ series.
Additionally, the ground state of a model complex,

[Ga2(tren)2(DHBQ)](BPh4)2(BF4) (where DHBQ = 2,5-dihydroxy-l,4-
benzosemiquinone and BPh4 = tetraphenylborate anion), which was employed to
aid in the assignment of EPR spectral features of 2, was also explored.
Furthermore, several additional complexes of the form [(tren)Ga(III)PSQ]2+
(where PSQ is a substituted phenanthrenesemiquinone ligand) will be discussed.

These simple systems, which were also explored by our group4 reveal various

41

mechanisms by which control of spin distribution can be achieved. EPR has been
implemented to characterize the ground states of other metal-semiquinone
complexes with diamagnetic metals, like 2. Most signiﬁcantly, gallium(III)
complexes of the 3, 5-di-tert-butyl-1,2-benzosemiquinonate anion species have
been prepared and studied via X-band EPR by the Tuck group,5 and Wieghardt
and coworkers have synthesized and studied the magnetic properties of a series of
metal complexes, including gallium(III), with a phenoxy radical ligand.6 More
recently, Baker et al. have carried out synthetic, EPR and electron nuclear double
resonance (ENDOR) studies on a series of gallium(III), indium(III), and

aluminum(III) complexes of the radical anion of diazabutadiene.7’9 In all cases

27 69, 71 115
Al,

hyperﬁne coupling from the metal nuclear spin (either Ga or In) was
observed, and room temperature X-band spectra of the gallium(III) complexes
were simulated with different isotropic hyperﬁne couplings from the different
Gallium isotopes. Other semiquinone species have been prepared with

diamagnetic metal cations such as closed-shell alkali and alkaline earth cationsm’

H A common theme in all studies mentioned above is that Spin is delocalized to

some extent (generally much less than 1%) onto the metal center from the
primarily organic based (rt-based) radical. This is manifested by the weak
hyperﬁne couplings (HFC) stemming from the interaction of the electronic spin

with the nuclear spin.

Different challenges arise in systems with S > 1/2. In non-doublet spin

42

systems zero-ﬁeld splitting, i.e. loss of degeneracy of ms spin sublevels, arises,
and small changes in zero ﬁeld splitting can drastically affect the appearance of an
EPR spectrum. After exploring the (magnetically) simple gallium(III)-
semiquinone complexes, this chapter continues on by characterizing the
paramagnetic ground state of the quartet complexes of chromium(III), starting
with high-symmetry Cr(acac)3-type (where Cr(acac)3 is
tris(acetylacetonato)chromium(III)) complexes, then exploring low-symrnetry
complexes such as the bimetallic complex [GaCr(tren)2(CA°at’°at)](BPh4)2 (3)
(Figure 2-3). Electron paramagnetic resonance is employed to characterize the
ground state. A mononuclear CrIII complex, [Cr(tren)(DTBCat)](PF6) (where
DTBCat = 3,5-di-tert-butylorthocatechol), was employed to aid in the

characterization of 3.

 

 

 

F CI _2+
('1‘ ”\ 0 0 T
N
(N); 0 >8; >
\/N/ \0 0 N N-
._ 3 _+
0 T
\ /N
Cr
0 O/jjgll

 

 

 

 

 

Figure 2-3: The complex [GaCr(tren)2(CA°a"cat)](BPh4)2 (3) and its model

complex [Cr(tren)(DTBCat)](PF6).

43

EPR has been implemented to characterize mononuclear chromium(III)
complexes in a variety of ligand environments, including bis-bipyridine and bis-

13

phenanthroline complexes,'2’ corroles,l4 and Crm-alkyl Species.15 Cr(acac)3

doped in Al(acac)3 has been probed by EPR at 110 K,16 which revealed slightly
different zero ﬁeld splitting values, all around 1.2 cm], for the three different
crystal sites. These values were recently conﬁrmed using luminescence and
excitation line narrowing spectroscopy.” Wang and Pillbrow18 have carried out an
extensive study of the symmetry relationships between the different eigenstates
within the quartet manifold, and Mabbs has presented a similar study.‘9

To lead to a greater understanding of the ground state electronic structure of
quartet systems and the origin of perpendicular-mode transitions within quartet
systems, the effect of the axial (D) and rhombic (E) zero ﬁeld splitting parameters
on the energy proﬁle of the ground state, and thus the appearance of the low
temperature EPR spectrum, is explored in the ﬁnal sections of this chapter.
Furthermore, a more extensive and systematic study of the effect of zero ﬁeld
splitting on the appearance of the EPR spectra of powder axial and rhombic
quartet systems at low temperatures (4 K) was carried out.

Beyond the realm of simple metal-centered paramagnetic systems, we also
expand on previous reports and our results by providing simulated spectra and
energy level diagrams with different D and E/D values relevant to other quartet

spin systems, such as quartet ground state organic compounds such as 2,3,5,6-

44

tetraﬂuorophenylnitren-4-y120‘ 21 and high spin clusters of C60.22 This report would

also serve to aid in the characterization of exchange coupled metal compounds
with isolated quartet ground states (IJI >> kT), such as trinuclear copper(II)

complexes,23 and metal-semiquinone complexes.24 It is the hope of the author that

this subsection can serve as a general tool to aid in simple and quick

characterization of quartet magnetic systems via X-band EPR spectroscopy.

2.2 Experimental Section

2.2.1 Synthetic Methods. All reagents and materials were used as received unless
otherwise noted. Solvents were purchased from Aldrich Chemical Co. and distilled
and degassed by the freeze-pump-thaw method. The ligand tris(2-
aminoethyl)amine (tren) was purchased from Aldrich and vacuum-distilled prior
to use. All synthetic procedures involving hydrochloranilic acid and 1, 2, 4, 5-

tetrahydroxybenzene were performed under an inert atmosphere. [F eCp*2](BF4)
was prepared according to literature methods.25 Hydrochloranilic acid and l, 2, 4,

5-tetrahydroxybenzene was synthesized by a modiﬁcation of a previously reported

26, 2 28

method. 7 Ga(tren)(NO3)3 was prepared according to our previous paper.

[Ga2(tren)2(CA’q’°’t)](BPh4)2(BF4) (2). Complex 2 was achieved by Dr. Dong

Guo, and synthesized according to previously published procedures.” 29

45

[Ga2(tren)2(DHBQ)](BPh4)2. Preparation of this complex was carried out by Dr.
Dong Guo. Under N2, 1, 2, 4, 5-tetrahydroxybenzene (1.00 mmol, 0.142 g) and
triethylamine (4.00 mmol, 0.404 g) were dissolved in 100 mL of degassed
methanol. This solution was added dropwise into a solution of Ga(tren)(NO3)3
(1.80 mmol, 0.724 g) in 80 mL of methanol. Following ﬁltration of a green
precipitate, excess sodium tetraphenylborate (4.00 mmol, 1.37 g) in 20 mL
methanol solution was added to the ﬁltrate. After standing overnight, white

microcrystals were obtained. Yield: 0.53 g (45%).

[Ga2(tren)2(DHBQ)](BPh4)2(BF4). Preparation of this complex was carried out
by Dr. Dong Guo. Under N2, [Ga2(tren)2(DHBQ)](BPh4)2 (0.141 mmol, 0.171 g)
and [FeCp*2](BF4) (0.15 mmol, 0.062 g) were dissolved in 2 mL of acetonitrile
and stirred for 30 min, resulting in a green solution. After ﬁltration, 5 mL of
dichloromethane and 5 mL of ether was added to the ﬁltrate yielding a yellow
solid. The product was ﬁltered and washed with ether (3 X 20 mL) and
dichloromethane (3 X 20 mL) and dried. Yield: 0.100 g (59%). Anal. Calcd

Ga2B3F4N3C66H7gO4: C, 61.2; H, 6.1; N: 8.7. Found: C, 61.4; H, 6.0; N, 9.0.

Synthesis of Chromium complexes. Complex 3 and the model complex
[Cr(tren)(DTBCat)](PF6) were synthesized according to previously published

28, 29
procedures.

46

2.2.2 Physical Measurements on Gallium-Semiquinone Complexes. Unless
otherwise noted, X-band EPR samples were prepared in a dry, inert atmosphere
(N2) by dilution in a mixture of butyronitrile and propionitrile (9:2), both of which
had been distilled from CaHz, degassed, and stored under an inert atmosphere.

The resulting sample concentratiOns were approximately of millimolar
concentration. The sample holder was a typical quartz tube of an appropriate inner
diameter (~0.6 mm) for X-band experiments. Low temperature continuous-wave
(cw) X-band EPR spectra were obtained with a Bruker ESP300E spectrometer
using two different resonant cavities: a perpendicular-mode cavity (Bruker
ST4102/8943) with resonances around 9.46 GHz and a duel-mode cavity (Bruker
DM1466) with perpendicular-mode resonances at 9.6 GHz and parallel-mode
resonances around 9.4 GHz. For each compound an appropriate power was
chosen to avoid saturation of the signal. An Oxford ESR A900 helium cryostat
was employed to maintain the temperature at 4 K. X-Band spectra were also
acquired at room temperature for several complexes. Samples were prepared as
described above, but acquisition was carried out using a ﬂat cell and a room
temperature cavity (Bruker 4103TM).

Electron nuclear double resonance (ENDOR) of the model complex of 2
was carried out at 10 K using a typical Davies-type ENDOR pulse sequence.30
Radio frequencies were scanned from 2 — 28 MHz and samples were prepared as

described above. W-band measurements of 2 were carried out at 4 K on a Bruker

47

e680 W-band instrument with a typical two-pulse sequence of a 7r/2 pulse (68 ns)
followed by a 7: pulse (136 ns). The frequency was 94.158 GHz and a pseudo-
modulation of 10 KHz was applied to the signal prior to Simulation to obtain the

characteristic ﬁrst derivative shape.

2.2.3 Physical Measurements of Chromium complexes. X-band samples were
prepared in a dry, inert atmosphere (N2) by dilution in a mixture of butyronitrile
and propionitrile (9:2), both of which had been distilled from CaHz, degassed, and
stored under an inert atmosphere. The resulting sample concentration was
approximately millimolar. Continuous-wave X-band EPR spectra were obtained
with a Bruker ESP300E spectrometer using a perpendicular-mode cavity (Bruker
4102ST/9516) with resonances around 9.458 GHz. For each compound an
appropriate power was chosen to avoid saturation of the Signal. All Oxford ESR

A900 helium cryostat was employed to maintain the temperature at 4 K.

2.2.4 Calculations: Gallium-Semiquinone Complexes. All simulations of X-
band perpendicular mode EPR spectra of were performed using Bruker XSophe

31'” in order to obtain spin Hamiltonian parameters. Resonant magnetic ﬁelds

~

were calculated assuming that g and A have coincident principal axes,“ 35

which is true for spin centers of D2.1 symmetry. Once a reasonable ﬁt for the EPR
spectrum was obtained, the parameters from the EPR simulations were applied to

calculate energy-level diagrams using programs of local origin.

48

[I

ll.

In order to simulate the room temperature data, which prominently displays
69’71Ga hyperﬁne coupling, both isotopes must be taken into account. The natural
abundances for the two isotopes are 60.2% for 69Ga and 39.8% for 7lGa. An
accurate simulation of the spectrum must take into account the three possible
combinations of isotopes, which are 69Ga-69Ga (with a probability of 36.24%),
71Ga-“Ga (15.84 %), and 69Ga-"Ga (47.90%), and weigh the spectrum of each
combination by its probability prior to summing them. XSophe is capable of
doing this, and presents a weighted average of the HFC values to the user. In
order to extract the different HF C values for each isotope, one simply has to work

backward from the weighted average. The isotropic HFC is proportional to the

nuclear g-value (gn) and the spin density at the nucleus (p):

Aiso oc gnp (2.1)
For complex 2, the symmetry of the molecule dictates that the same amount of
spin will be delocalized onto both Ga nuclei, whether it happens to be 69Ga or

7l . .
Ga. Because of thls we can wrlte:

69 69
A( Ga)_ gn( Ga)_1.344393 _0787

 

_ _ _ . 2.2
A(7‘Ga) gn(7lGa) 1.708180 ( )
The weighted average is:
Z = A(69Ga) w(69Ga) + A(7‘Ga) w(7‘Ga) (2.3)

49

whereA is the average, and W(69Ga) and w(7'Ga) are the weights (natural

abundances) of 69Ga and 7'Ga. Using equations 2.2 and 2.3 one can work
backward and determine the HFC for each isotope, assuming that the same amount
of spin is delocalized onto both metal centers.

Time-independent calculations of the electronic structures of complex 2,
the model complex [Ga2(tren)2(DHBQ)](BPh4)2(BF4), and the two ﬁ'ee bridging
ligands in the trianionic semiquinone form were carried out using density
functional theory as implemented in Gaussian 9836 using the UB3LYP/6-311G**
level of theory. The initial geometry of the molecule was generated using

GaussView37 and optimized using the UB3LYP functional and a 6-311G** basis
set with imposed symmetries of C2V and Dzh- Single-point calculations were

carried out using the unrestricted UB3LYP functional and a 6-311G** basis,

assuming a doublet ground state and a molecular charge of 3+ for the Ga"1 dimers

and 3- for the free ligands.

2.2.5 Calculations: Chromium Complexes. The X-band perpendicular mode

31-33

EPR spectra were simulated using Bruker XSophe in order to obtain spin

Hamiltonian parameters. A partition number greater than 100 was used for all
simulations to reduce the “noise” associated with the numerical integration. For
XSophe, it should be noted that although a segment number of 2 is suitable for

32

doublet systems, more complex systems (S > l/2) require a segment number of 4

50

or 8. Once a reasonable ﬁt for the EPR spectrum was obtained, the parameters
from the EPR simulations were applied to calculate energy-level diagrams and the
angular dependence of EPR transitions for the compounds using either XSophe or
homemade programs.

The effect of zero-ﬁeld splitting on the appearance of the X-Band EPR
spectra of a mononuclear Chromium(III) complex was explored. Simulations
were carried out in Bruker XSophe, with line widths equal to those determined
from the simulation of the experimental data. For the general study of the effects
of zero ﬁeld splitting, the linewidths were arbitrarily set at 10 G. Transition
roadmaps, which illustrate how the ﬁeld position of transitions changes as the spin
system is rotated with respect to the applied magnetic ﬁeld, were also calculated
using XSophe. Calculations of energy level diagrams and transition diagrams
were carried out in Mathematica (vide infra).38 Transition diagrams, which map
out the ﬁeld of transition as a ﬁmction of the zero ﬁeld splitting parameters, were
calculated by taking the difference between two eigenvalues of a given orientation
as the energy of transition (about 0.3 cm" at X-Band frequency and 3.3 cm'l at W-
Band frequency) and solving for D or E. At X-band frequency solutions were

discarded which did not have transition probably within the window |D|< 1 cm"

and 0 < B < 10000 G (1 T) for the axial system and 0 < E/D < 1/3 and 0 < B <

10000 G for the rhombic case.

51

I1.

2.3 Results and Discussion. Gallium(III)—Semiquinone Complexes.
2.3.1 [Ga2(tren)2(CAsq’°")](BPh4)2(BF4) (2). The X-band EPR spectrum

of [Ga2(tren)2(CASq’°at)](BPh4)2(BF4) (2) at 4 K is shown in Figure 2-4(A). The

spectrum shows a main peak around g’ = 2 (where g'= (1%63) with multiple lines

due to hyperﬁne splitting. The rhombic nature of the g-tensor and hyperﬁne
tensor are clearly apparent from the dissymmetry of this main peak. All of the
complexes reported herein have either D211 or sz local symmetry at the magnetic
centers, and therefore possess “rhombic” EPR symmetry,19 meaning that, in
general the principal axes (the diagonal components of the g—tensor) are
nonequivalent, i.e. gxx 76 gW 7E gZZ as a result of different magnetic ﬁeld projections
along the principal axes. Two sidebands of much lower intensity are also
observed, and no half-ﬁeld transition (Am, = 2 ) is observed, as seen for other
Gaul-semiquinone complexes, as will be discussed later.“ 10

This molecule is the most magnetically simple of the series, consisting of
two diamagnetic metal centers bridged by a semiquinone ligand radical, resulting
in a total S = ‘/2 for the system. Magnetic susceptibility data for this complex from
2 - 350 K, which reveal a temperature-independent magnetic moment of pm: 1.64
i 0.04 113, (Ilspin-0n1y for S = '/2 is 1.73 mg) conﬁrm this formulation. From the
sq.cat

magnetic susceptibility data and previous investigations of the isolated CA

bridge,39 one would expect the EPR spectrum to exhibit a simple doublet

52

 

spectrum, with possible HFC contributions from the bridging atoms (13C, 35‘37Cl)

and the gallium (69’71Ga) nuclei.

 

 

 

 

 

 

fl) A
- U 3‘ B
.1 I/ \
LL 1' X C
105 2.00 2 05
gvalue

 

 

 

Figure 2-4: (A) Experimental X-band spectrum of complex 2 acquired in a
butyronitrile/propionitrile (9:2) glass at 4 K (0.63 11W, 9.458 GHz). This spectrum
was simulated (B) using the following parameters: gxx = 2.0100, gyy = 2.0097, gZZ
= 2.0060, axx(Ga) = 4.902 G, ayy(Ga) = 4.124 G, am(Ga) = 3.167 G. (C) is the
experimental X-band spectrum of [Ga2(tren)2(DHSQ)](BPh4)2(BF4), acquired at

4.4 K, 63 nW, 9.624 GHz.

A DFT calculation of the SOMO (Singly Occupied Molecular Orbital) of

complex 2 (a-SOMO, Figure 2-5a) and the corresponding spin density distribution

53

(2.5b) is useful in the interpretation of this spectrum. The B-SOMO is at a much
greater energy than the a—SOMO (molecular orbital 1068 for the B—SOMO versus
163 for the a-SOMO), so that the spin density can be described entirely in terms of
the a-SOMO. Figure 2-5 reveals that although the SOMO density is greatest at the
four equivalent carbon atoms and four equivalent oxygen atoms, with a node along
the central C2 axis that runs through the two chlorine atoms, the spin density is
present at every position on the CA bridge as a result of spin polarization: the 01-
spin populated SOMO orbitals induce B-spin at adjacent positions as a result of the

. . . . 40,41
sprn-polarrzatlon mechanlsm.

The bridging ligand has B1 g symmetry in D2h,
which can not support degeneracy, so that a rhombic g—tensor and hyperﬁne tensor
are expected. In terms Of hyperﬁne coupling (HFC), which gauges the interaction
between the unpaired electron spin and nuclear spins, weak and perhaps
unobservable peaks are expected from the four equivalent carbon atoms because

13C (I = 1/2) only has a natural abundance of 1.1%. For the four equivalent oxygen
atoms which display SOMO density in Figure 2-5, there is no expected observable
contribution to the hyperﬁne splitting because 170 (I = %) has a natural
abundance of only 0.037%. For S = ‘/2 systems, isotropic hyperﬁne splitting results
in (2nj1j+l) peaks for nj nuclei of spin Ij, so that, if indeed the peaks are observable,

ﬁve peaks are expected for the four equivalent central carbon atoms (with a
positive HF C because a-spin resides about these atoms) and three peaks for the

other two carbon atoms (with a negative HFC as a result of B-spin). Hyperﬁne

54

 

..
s!

interactions with Gallium (”Ga and “Ga, both of which have 1 = % and account

for 100% natural abundance) or Chlorine nuclei (”Cl and 37Cl, with I = %

account for 100% natural abundance) are potentially important. Previous
investigations of Ga-semiquinones, as noted in the introduction to this report,

corroborate the assignment of HF C as due to Gallium nuclei.

 

 

 

 

 

Figure 2-5: Calculated SOMO (a.) and spin density (b.) of complex 2 at the

UB3LYP/6-3l lG"“'I level.

The sidebands are of the correct number—assuming splitting from four

equivalent l3C nuclei, as the SOMO calculation suggests—and intensity to result
from '3 C splitting. While this explanation is justiﬁed by the position of the

calculated SOMO density, a previous EPR investigation of the CAsq’Cat bridge,39

55

 

in addition to our own investigation (vide infra) revealed HF C values that are
much to small to account for the sidebands, separated from the main peak by
approximately 70 G. Furthermore, this description does not take into account the
carbon atoms adjacent to the chlorine atoms, which have an excess of |3-spin. The
isolated chloranilic trianion species in alkaline solution (6 N NaOH) has been
previously studied using EPR spectroscopy.39 The spectrum revealed HF C of 2.71
and -8.17 gauss, resulting in 4 sidebands, and chlorine hyperﬁne interaction of
0.18 gauss. The origin of these sidebands will be discussed later.

Given the DFT picture, consideration of the dominant spin Hamiltonian
parameters, natural abundances of magnetic nuclei, the splitting rules for a S = ‘/2
system, and previously published hyperﬁne constants for similar compounds,42 we
can now begin to assign the features of the spectrum of this compound. The EPR
spectrum of the main peak, corresponding to 2, was calculated according to the

spin Hamiltonian

~ —0

ri=neé-g.§+i-A.s (2.4)
The ﬁrst term on the right side is the Zeeman splitting of the spin states upon

application of a magnetic ﬁeld B, whereé is the g-tensor. In general, if the g-
value corresponds to that of an organic-based radical, a small deviation from the

free-electron g-value (ge = 2.0023), on the order of 103, is expected.43 The

~

second term on the right side is the hyperﬁne coupling term, with A being the

hyperﬁne tensor, which gauges the magnitude of the interaction between the

56

—.

nuclear and electronic spins. I is a vector describing the nuclear spin, and S is

the electronic spin vector. The nuclear Zeeman term (ﬂeB ' 8n '1 ) is excluded as

an approximation, because the interaction is only about 1/2000 of the electronic
Zeeman interaction, as a result of the smaller magnetic moment of nuclei as
compared to that of the electron.

The simulation, Figure 2-4(B), conﬁrms the expected spin Hamiltonian
parameters. The g-tensors and hyperﬁne tensors are rhombic. The spectrum was

simulated assuming that hyperﬁne splitting from 69’"

Ga nuclei are responsible for
the Splitting pattern observed on the central-ﬁeld feature, although we were unable
to ﬁt this peak with different HFC values for both Ga isotopes because of the lack
of structure on the peak, due to the expression of the anisotropy in the HFC tensor
in frozen solutions. Although this simulation may not provide accurate numbers
for the diagonal components of the HFC tensor for both 69Ga and 7'Ga it does
allow us to put limits on the anisotropy present in the g-tensor. The value of Ag (
= gxx - gzz) of 0.004 is a reasonable value for a organic-based radical, and any
larger value of Ag would likely result in a much broader Spectrum than that of
Figure 2-4(A). The room temperature EPR spectrum of 2, presented in Figure 2-6,
Shows a dynamically averaged (isotropic) HFC as a result of tumbling in solution,

and the isotropic HFC for both gallium isotopes can be extracted. This spectrum

was simulated with the isotropic HFCs A(69Ga) = 3.136 G, A(7'Ga) = 3.980 G.

57

As mentioned previously, the splitting of the main peak to produce seven
peaks is likely the result of HF C to the gallium nuclei or the chlorine nuclei, both
of which have I = 3/2. To aid in the assignment of these peaks as due to gallium or
chlorine HFC, the compound [Ga2(tren)2(DHBQ)](BPh4)2(BF 4) was synthesized.
This compound is a useful control because it has no chlorine atoms, so a pattern of
hyperﬁne splitting for this complex similar to that of compound 2 would indicate
splitting due to interaction of the unpaired electron with the gallium nuclei. The X-
band spectrum of [Ga2(tren)2(DHBQ)](BPh4)2(BF4) is presented in Figure 2-4(C).
The spectrum, when acquired at low microwave powers, reveals hyperﬁne
splitting at the main peak, which manifests itself only as slight shoulders on this
peak. To conﬁrm that broadening of the spectrum relative to that of 2 and apparent
lack of sidebands was not the result of concentration effects due intermolecular
spin exchange, further spectra were collected at a lower sample concentration,
yielding the same results. Although the spectrum at 4 K leaves some doubt as to
the nature of the HFC in this complex, the room temperature isotropic spectrum,
Figure 2-6, clearly conﬁrms the origin of the splitting on the main peak as due to
HFC with Gallium nuclei, and was simulated with A(69Ga) = 3.337 G, A(7'Ga) =
4.240 G.

For both complexes at 4 K we were unable to ﬁt the spectra with different
69Ga and 71Ga HFC values because of the anisotropy inherent to powder spectra,

which can often result in one broadened line, as is observed for the model complex

58

at high microwave powers. As aforementioned, when the power is low some ﬁne
structure, which reveals itself as shoulders on the main peak, can be resolved.
69’7'Ga HF C values of about 4 G ﬁt this peak relatively well, revealing that about
0.08% of the spin is delocalized onto the Ga centers (a0 = 4356 G for 69Ga and
5535 G for 71Ga), and that > 99% of the spin density is ligand centered. This is
similar to other Gaul-semiquinone complexes, where in general less than 0.1 % of
the total spin density is delocalized onto the metal.9 Proton HFC is also likely
buried under this peak. The X-band EPR spectrum of the trianionic semiquinone
species DHBQ, acquired at 298 K, has been studied previously, both times in basic
aqueous solution, revealing isotropic hyperﬁne couplings of An = 0.79 G, AC“) =
2.82 G and qu) = 6.62 G (where AC“) corresponds to the hyperﬁne coupling of
the '3C nuclei adjacent to the oxygen atoms, and AC(2) corresponds to the
hyperﬁne coupling of the 13C nuclei adjacent to the hydrogen atoms).44 A later
report by Das and Fraenkel conﬁrms and reﬁnes these values, and the authors
apply the appropriate Sign to the hyperﬁne constant depending on the assumed
spin density.45 We also carried out this experiment, and in this case only the
isotropic lH HFC, which will be useful for assigning the features of the ENDOR

spectrum of the model complex, were ﬁt with g = 2.00923, an = 0.801 G, again

indicating the expected organic radical nature of the species. This topic will be

further explored when the ENDOR spectra are examined.

59

 

 

 

 

'3400 '3420 '34h0 '3460 '3480 '3500 '
BIG]

 

 

 

Figure 2-6: Room temperature X-band spectrum of 2 (black), acquired at 9.696
GHz. The spectrum was simulated (red) with g = 2.00858, A = 3.474 G (A(69Ga)

= 3.136 G and A(7‘Ga) = 3.980 G).

60

 

 

34'00 ' 34'20 ' 34'40 ' 3460 ' 34'80 ' 3500
BIG]

 

 

 

Figure 2-7: Room temperature X-band spectrum of the model complex (black),
acquired at 9.695 GHz. The spectrum was simulated (red) with g = 2.00906, A =

3.697 G (A(69Ga) = 3.337 G and A(7'Ga) = 4.240 G).

2.3.2 Observation of a Triplet State. It was mentioned previously that the
sidebands are of the correct number to correspond to HF C of four equivalent l3C

, I . . .
Sq ca unit of 2. This Idea was

nuclei, which are present in the bridging CA
discarded, however, because the observed HFC to 13C nuclei in a previous report
of the CASq’cat ligand revealed HFC values much too small to account for the

observed Splitting in the low-temperature X-band spectrum of 2. A more likely

possibility, given the 13C HFC revealed by the isolated semiquinones are much

61

_l:--'-'-:"

smaller than the splitting observed for 2 (2.4 and 8.2 G vs. ~140 G), is that another
electronic state is contributing to this spectrum. This has been observed before in
other metal-semiquinone complexes, including Gam, In"I and Tlm complexes of
the 3, 5-di-tert-butyl-1,2-benzosemiquinonate anion,5 Mg", Zn", CdII and Ball
complexes of the same anion,10 and others,46 where the side bands are different
resonances within a triplet state, due to the formation of a biradical (S = 1)
complex. An unequivocal test for the presence of a triplet electronic state would
be the presence of a half-ﬁeld (AmS = 2) resonance, although we were unable to
observe this resonance in either our cw-EPR experiments at 4 K or room
temperature.

Although the furthest sidebands can not be observed in the spectrum
presented in Figure 2.4, increasing the power reveals that these side bands are
present (Figure 2-8). The sidebands were simulated assuming a triplet (S = 1)

state, using the following spin Hamiltonian:

A

H=6eé.g.§+§-D-§ (2.5)
with deﬁnitions for the ﬁrst term as described previously. The right-hand term
accounts for zero-ﬁeld splitting effects, where D is a matrix describing the zero-
ﬁeld splitting in three dimensions and S is a spin vector. The triplet state was
simulated with g = 2.009 and D = 150 G (422 MHz or 0.012 cm”), and the
simulation parameters were applied to calculate energy level diagrams for this

state for the orientations of the magnetic ﬁeld perpendicular and parallel to the z

62

axis (Figure 2-9, utilizing the secular determinant shown in Appendix B).

Resonances are Shown at X-band frequency (~ 0.3 cm”).

 

7’ l,
/ '1 her—~-

l l
1' _ ‘ l l
- l‘ I \\ l _. _
V r , -

l
l
l
l
l
l
p l
.‘ V ‘ /
i
l
l
l

' 32'00 ' 33'00 ' 34'00 ' 3500 ' 3600 7 3700
BIG]

 

 

 

Figure 2-8: EPR resonances produced by the triplet state of a biradical complex.
The spectrum of 2 (black) was acquired at 4 K (9.624 GHz, 0.63 mW), and
simulation of the sidebands (red) was carried out assuming a triplet state, and with

the spin Hamiltonian parameters g = 2.009 and D = 150 G.

The intensity of the bands can be understood by the population distribution
of different orientations of the molecules with respect to the direction of the
magnetic ﬁeld. In general, there are many more orientations with the molecular z-
axis perpendicular to the direction of the magnetic ﬁeld, while there is only one
orientation that is perfectly parallel to the ﬁeld. The greater population of a given

orientation results in a greater signal. The energy level diagrams reveal that

63

 

molecules with an orientation where z is perpendicular to the magnetic ﬁeld
direction have resonances closer to g’ = 2 than those with parallel orientation,

resulting in the observed pattern of sidebands.

 

 

Energy (cm-1)

 

0 2000 4000 6000
BIG]

 

 

 

Figure 2-9: Energy level diagram of the Zeeman Splitting of the triplet state
formed by ferromagnetic exchange between two molecules of 2. The diagrams
were produced from the simulation parameters g = 2.009 and D = 150 G, and

resonances are shown at X-band frequency (~ 0.3 cm").

As further conﬁrmation of the spin Hamiltonian parameters, a W-band (~
95 GHz) spectrum of 2 was acquired, and is presented in Figure 2-10(A). In these

spectra g-values may not be accurate, as the magnetic ﬁeld could be incorrect by

64

as much as 50 G, due mainly to the large magnetic ﬁelds required to achieve
resonance at W-Band frequency. The main peak, which is lacking HF C from
gallium nuclei, again reveals the rhombic nature of the ground state of 2, with g =
(2.0077, 2.0062, 2.0053). The simulation is Shown in Figure 2-10(C). The W-
band spectra conﬁrm that the sidebands are indeed the result of a biradical (S = 1)
complex, as the intensities are much too great to ﬁt the sidebands with 13C HFC,
given the natural abundance of this isotope. These sidebands were simulated with
g = 2.006 and D = 150 G, indicating the detection of the same triplet state.

To investigate whether the observed triplet electronic state is the result of
intramolecular or intermolecular Heisenberg spin exchange coupling between two
unpaired electrons a sample was prepared and saturated with tetrabutylammonium
tetraphenylborate. If intermolecular exchange is the mechanism, the salt should
act to inhibit any aggregation of the molecules, and thus eliminate the observed
triplet signal, or at least decrease the intensity of the signal. This was in fact
observed (Figure 2-11). This result is as expected, given that cyclic voltammetry
presented in the previous report revealed a quasi-reversibility in the sq,cat/sq,sq
couple, suggesting degradation upon oxidation of CASq’c‘lIt to the dianionic CAsq‘SCI
form. The inability of this form to chelate is also evidenced by the many

sq.sq

unsuccessful syntheses of bimetallic complexes employing the CA species as

the bridging ligand. Furthermore, stabilization of a biradical within the fully

cat, cat

aromatic CA form seems unlikely as well. This result allows us to

65

conclusively ascribe the origin of this state as due to intermolecular exchange

coupling.

 

 

 

 

 

 

33300 33400 33500 33600 33700
BIG]

 

 

 

Figure 2-10: (A) W-band data of 2 acquired at 94.158 GHz, 5 nW, 10 K. The
ﬁrst derivative was taken by applying a pseudo-modulation of 10 kHz (B). The
main peak belongs to the monoradical 2, and was simulated with an anisotropic g
tensor (2.0077, 2.0062, 2.0053) with widths (11, 5, 4) G (C). The sidebands were
simulated assuming a triplet electronic state formed by interaction of two

monoradical species (D). Simulated with g = 2.006 and D = 150 G.

66

 

 

 

3200 5 3300 I 3400 I 3500
B[G]

 

 

 

Figure 2-11: Comparison of the experimental X-band spectrum of complex 2
(red) acquired in a butyronitrile/propionitrile (9:2) glass at 4 K (0.63 uW, 9.458
GHz) with another sample saturated with tetrabutylammonium tetraphenylborate
salt (black), collected under the same conditions. The diminished intensity of the
sidebands for the spectrum of solution containing the TBA salt reveals that these
peaks are likely due to an electronic state formed by aggregation of complex 2 in

solution.

2.3.3 ENDOR of the Model Complex. To conﬁrm the values of the hyperﬁne
constants for both 2 and its model complex, ENDOR was carried out. ENDOR of
2 Yielded no peaks but ENDOR of the model complex at 10 K (Figure 2-12)
revealed a series of peaks centered around 14 MHz and a broad featureless peak
near 8 M12. ENDOR is implemented to provide information on the weaker

interactions between the unpaired electron and the remote magnetic nuclei which

67

EPR cannot at times provide. This is especially true of unpaired electrons in 1t-
based organic systems, where hyperﬁne interactions can be unresolved due to
small anisotropy in the g- and A-tensors and delocalization over many nuclei.9
ENDOR also offers a signiﬁcant enhancement in resolution as compared to
standard cw-EPR techniques. ENDOR has been employed successfully to
elucidate the HFC of 1H on the peripheral tert-Butyl and aryl groups of several
GaIII radical complexes, including [(But-DAB)Ga{E(SiMe3)2}I] (E = N, P, As and
DAB is the radical anion of diazabutadiene),7 [(R-DAB)MX2] (R = Aryl or t-
Butyl, M = All", Ga"I or In"I and X = l or Br).9 In all cases the HFC from the
peripheral tert-butyl and aryl groups in the RT isotropic ﬂuid spectra were not
observable, but cw-ENDOR revealed the HFC from the protons on these moieties.
For the case of [(Ar-DAB)GaIz] the EPR signal was too broad to resolve any HFC,
but cw-ENDOR at X-band frequency and 10 K revealed HFC from not only the

III

aryl groups, but also from the nuclei of the Ga center of the complex in the 20-

55 MHz region. While the spectroscopic features assigned as Ga HFC were
difﬁcult to assign due to many complicating factors, including overlapping
features due to anisotropy in A and Q tensors, different isotopic abundances,
magnetic moments, quadrupole moments and isotropic HF C constants for the two
Gallium isotopes, the preliminary results of Simulation of the spectrum suggests A
= (67, 74, 86) MHz and a value for Q, the isotropic quadrupolar moment, of ~ 0.1

MHz. These values are over an order of magnitude greater than the HFC values

68

 

we obtained via cw X-band EPR spectroscopy, indicating a greater degree of spin
delocalization onto the gallium atom (per gallium) for [(Ar-DAB)GaIz] as
compared to 2.

In general, for a doublet ground state system in an isotropic ﬂuid
interacting with a nucleus with I = 1/2, two ENDOR lines are expected at the

resonance frequencies

VENDOR = Iv” i 4am, (2.6)

where “on is the Larmor frequency of the nucleus (for 1H the Larmor frequency is
14.689 MHz at 3500 G) and aiso is the isotropic hyperﬁne coupling constant. If the
Larmor frequency of the nucleus is less than 1/2 aiso, resonances will be separated

by 21)n and centered at ago/2

(2.7)

VENDOR = '5“ 1'30 i an

If the symmetry of the system is lowered, by having a system that inherently

possesses lower than cubic symmetry and sampling a powder or ﬁ'ozen matrix

rather than an isotropic ﬂuid, aiso can be replaced with A,, one of the principal

components of the hyperﬁne tensor. All different orientations must then be taken
into account, vastly complicating the interpretation of the spectrum.

The most prominent feature of the ENDOR spectrum of the model complex

is the broad quartet of peaks centered on 14.5 MHz, which can unequivocally be

assigned as 1H HF C given the proton Larmor frequency and isotropic HFC value

69

determined by exploration of the free ligand (2.25 MHz, where 1 G = 2.81 MHz),
so that the peaks fall into the regime governed by equation 2.6. The quartet
reveals an axial symmetry, with A..L z 0.7 MHz and A” z 3.4 MHz. The
dynamically average isotropic HFC of the free ligand falls into this range,
indicating a negligible perturbation to the organic-based spin system upon bonding
to the GaIII metal centers.

AS discussed above, the ENDOR spectra corresponding to gallium nuclei

can be extremely difﬁcult to interpret. For nuclei that possess a quadrupole, i.e.

those with I > 1, ENDOR peaks are found at the resonant frequencies
VENDOR = 1A,. i 1),, : %Qi(2m1 +1) (2.8)

where Q, is one of the principal components of the quadrupole tensor and mi can
take the values I, I-1, ...-I. The Larmor frequency for 69Ga and 7'Ga are 3.5867
MHz and 4.6913 MHz (at 3500 G), respectively. Given this, and the 69’7'Ga HFC
values determined from the X-band EPR spectrum of the model complex (~ 4 G or

11.2 MHz), and considering the quadrupole contribution to be negligible, the

peaks corresponding to Ga HFC should be found in the region governed by

equation 2.7: if the Larmor frequencies of the Gallium nuclei are less than 1/2 aiso,

which appears to be the case, the peaks will be centered around 5.6 MHz and
separated by on, so that, considering anisotropy as well, a broad peak will be

present from about 7 — 10 MHz, as an estimate, with its partner peak in the range 1.

70

— 4 MHz. However, if the nuclear quadrupole interactions are not negligible,

which may be the case in this system, the ENDOR spectrum will be vastly

affected, and much more difﬁcult to interpret.

 

 

 

 

 

 

14.4
15.1
7.6 13'1 16.5
0 10 20 30
RF [MHZ]
Figure 2-12: ENDOR spectrum of the model complex

[Ga2(tren)2(DHSQ)](BPh4)2(BF4) acquired at X-band frequency and 10 K in a 9:2

butryonitrile/propionitrile glass.

Despite the problematic assignment of the peak at 7.6 MHz via the
formalism presented above, we can rule out every other atom in this molecule as

being the source of the HFC given the Larmor frequencies and experimentally

71

determined HFC values. At this time a tentative assignment of the peak at 7.6

69’71Ga HF C is made, and as to the best of our knowledge is

MHz as resulting from
only the second incidence of a ENDOR signal corresponding to Gallium nuclei in

a discrete molecular metal complex system.9

2.3.4 Further Studies. EPR and ESEEM investigations of the nitro-substituted
gallium(III)-semiquinones [Ga(tren)(2,7-di-N02-PSQ)](C1)(BF4) and

[Ga(tren)(3,6-di-NOz-PSQ)](BF4)2 (where PSQ = phenanthrenesemiquinone) were

47, 48

carried out. This work has been presented elsewhere, and the reader is

directed to the references. The results add a great deal to the work presented
above by providing a more detailed look at the spin distribution in these
complexes, as well as elucidating the mechanisms of spin distribution. Perhaps
most importantly, it was shown that spin density at the gallium(III) metal center
could be modiﬁed by varying the substituent position on the ligand. This
observation opens up exciting avenues toward control of spin distribution and
magnetic properties of single molecules, which is discussed more ﬁJlly in the ﬁnal

chapter of this dissertation.

2.3.5 DFT Calculations of Ga-SQ complexes. The application of DFT in
determining HFC parameters, in particular the B3LYP functional,49 has enjoyed

reasonable success recently, as evidenced by the Larsen group’s work on VO2+

72

50,51 47,48,52

complexes as well as our own group’s work. It has also been reported

that in the study of organic radicals, DF T is far superior to unrestricted Hartree-

49,53,54

Fock calculations in terms of accurately predicting spin densities. Of note

and signiﬁcance to this study are recent studies by Tuononen et. al. wherein
different non-relativistic, quasi-relativistic fully relativistic functionals were
implemented to calculate HFC parameters for a series of Gallium-dipnictogen

55,56

diazabutadiene racials. These values were then used as initial guesses for

simulations, which employed iterative least-squares ﬁtting routines, of the EPR
spectra of the molecules in this series. This approach proved to be useful in these

instances, and in general is especially useful for compounds containing atoms

beyond the second period, which can often have I > 1/2 and/or possess different

isotopes with different nuclear magnetic moments.

Herein we wish to test the ability of DFT, and speciﬁcally the UB3LYP/6-
311G** level of theory, to accurately predict the spin density about a given atom
within this system. To do this, we relate the calculated spin densities from the
natural population analysis (NPA), as given by DFT, to the experimental spin
densities, which are calculated from the isotropic HFC constants. The isotropic
HF C values only give information on the spin at the nucleus, i.e. Fermi contact,
due to the fact that in room temperature isotropic media, which produces a
dynamically averaged value for the HFC, the dipolar (i.e. anisotropic) component

collapses to zero.57 As described earlier, Fermi contact coupling is the result of

73

unpaired electron spin density residing in an orbital (atomic or molecular) that has
a non-zero amplitude at a nucleus where I 315 0. The isotropic HFC for nucleus N is

calculated from equation 2.9:
47: ..1
aiso(N) = Tgegnﬂeﬂn (S2) p(N) (2.9)

where g6 and gn are the electronic and nuclear g-factors, Be and [3,, are the electronic
and nuclear Bohr magneton. p(N) is the Fermi contact integral, which
corresponds to the spin density at N. By replacing all the constants in equation 2.9

Nguyen et. al.58 produced the following conversion factors, where aiso(N) is given

in Gauss:
aiso(1H)=1594.9p(lH) (2.10a)
aiso(13C)= 401.lp('3C) (2.10b)
aim 35C1) = 52.lp(35C1) (2.10c)

Using these conversion factors, we were easily able to calculate the spin density at
these nuclei (i.e. the Fermi contact) by using the experimentally determined
isotropic HFC values. In the case of the 'H HFC in the model complex, which
was only observed in the low-temperature ENDOR spectrum, the isotropic HF C
value must be extracted from the hyperﬁne tensor. From this experiment it was

determined that A1. =1 0.7 MHz and A” z 3.4 MHz, leading to the following A-

ICIISOI'Z

74

 

"0.7 0 0
0 0.7 0
0 0 3.4

b— ‘

 

 

The isotropic HF C value is simply one-third the trace of this tensor, or i 1.6 MHz
(0.57 G). Naturally, there is some error in this measurement due to the noise level
of the ENDOR spectrum; this is to be compared to the value of i 0.79 G found for
the trianionic semiquinone DHBQ species. To determine the spin density at the
Gallium nuclei a weighted average of the HFC for the two different isotopes must

be carried out, assuming the following relation, as reported by Weltner,35 holds

true:
aiso =a0p(N) (3.11)
Here a0 is the HFC if the spin density at N is unity, and equals 4356 G for 6()Ga

and 5535 G for 7'Ca.

The lower half of Table 2.1 gives the comparison between the experimental
and calculated spin densities about the gallium nuclei in both 2 and the model
complex. For both systems the calculation overestimates the magnitude of the
Fermi contact by less than a factor of two. The negative value of the Fermi
coupling is explained as follows. Unpaired spin density at the nucleus may arise
via s-orbital admixture into the molecular orbital containing the unpaired spin or
by Spin polarization of ﬁlled inner orbitals by outer orbitals which contain an

unpaired spin. The calculated magnitude of the spin density for Gallium in 2 was

75

found to be 0.00072, corresponding to the s-orbital spin density. By summing up
the contributions of the s-orbitals, the calculation gives a value of -0.0012 for 2
and -0.0013 for the model complex. The calculated Fermi contact coupling of -
6.74 G for 6“)Ga in complex 2 and -6.87 G for the model complex indicates B-spin
density of the core s-orbitals, as indicated by the sum of the spin density in the s-
orbital contribution to the molecular orbital. The B—spin density of the s-orbital
contribution is a result of spin polarization by the u-spin populated p and d
orbitals.

The upper half of Table 2.1 explores the relationship between spin density,
Fermi contact coupling, and HFC for the six carbon atoms in the trianionic
semiquinone forms of the bridging ligands CA and DHBQ. The values for the

39, 45

HF C were obtained from previously published results. The calculated Fermi

contact couplings compare well with the experimental HFC results, and provides a
good example of the anisotropic components (p- and d-orbital contributions)
collapsing to zero in a dynamically averaged system: although the main
contributions to the spin density arise from the p-orbitals, these cancel out in
solution. The experimental spin density and the sum of the total s-orbital spin
densities differ by no more than a factor of 3.1 and no less than a factor of 1.5 for
all carbon atoms in both molecules.

The upper half of Table 2.1 also explores the spin density about the

chlorine atoms in 2 and CA3' and the protons in the model complex and DHBQ3'.

76

In both cases the atoms are in the same position with respect to the oxygen atoms
and lie along a C2 axis. In general, for the three systems for which experimental
HFC values were available, the magnitudes of the calculated and experimental
spin densities are remarkably Similar. Perhaps the most interesting aspect of the
data presented in this section of Table 2.1 is that the calculation indicates B-spin
density on the chlorine atom in CA3" while a-spin density is indicated on the
proton for both the model complex and DHBQ3'. Figure 2.5 indicates that for
complex 2 polarization of the spin density by the a-SOMO leaves residual B—spin
density on the carbon atom adjacent to the chlorines (or protons in the case of the
model complex). This indicates that chlorine atoms of 2 induce spin
delocalization, while in the model complex a spin polarization mechanism is

occurring.

77

Table 2.1: Experimental and calculated spin densities for [Ga2(tren)2(CAsq’cat)]3+
(2), [Ga2(tren)2(DHBQ)]3+ (2a), and the free ligands [CAsq’cm]3' and [DHBQ]3'.
(Key: a. calculated using equations 3.9 and 3.10 based on exp. values; b. summed
contributions for s, p, and d orbitals. Only 8 contributions are relevant for Fermi
contact; c. ref 39; (1. applies to 35Cl only (see text for details); e. ref 45; f. ref 44; g.

applies to 69Ga only (see text for details)).

 

 

 

 

 

 

 

Experimental Calculated
b 0
Also (G) p(ma S pb db F 91111] (G)
[CAsq,cat]3'
C1 2.71c 0.00656 0.00441 0.13085 0.00044 0.59
CZ -8170 0.0204 -0.00660 0.06705 0.00034 0.39
0.18 0.00345 0.00023 0.00310 0.00020 0.51241
[DHBQ13‘
C1 263e 0.00656 0.00469 0.13831 0.00034 0.82
c2 _6 66c -0.0166 0.0053 0.07359 0.00024 .7.01
1
H 0.79f 0.00050 0.00070 1°13674
2
6903 3.136 0.00072 0.00120 0.00176 0.00031 -6.74169g
713,, 3.980
35.370 0.00017 -0.006l4 0.00008 ’°°3'7°7
2a
69Ga 3.337 0.00077 0.00130 0.00205 0.00029 -6.86987g
71Ga 4.240
1H 0.570 0.00036 0.000130 0.03196

 

78

 

2.3.6 Conclusions for Ga-SQ complexes. The doublet ground state of
[Ga2(tren)2(CAsq’cat)](BPh4)2(BF4) (2) has been characterized by several
techniques, including EPR, ENDOR, and application of density functional theory.
A triplet electronic state was observed and characterized as arising due to the
interaction of a pair of molecules in solution. Comparison of experimental Fermi
contact couplings with those calculated by DFT has proven this method to be an
effective tool for predicting spin densities in both the gallium dimer systems and
CA and DHBQ semiquinones. This molecule is the simplest paramagnetic species
within the bimetallic [M1M2(tren)2(CA"')]"1+ series, and therefore the work
presented herein, in addition to the results of the previous report, form the
foundation for understanding the paramagnetic ground states of the molecules in

this series. The remainder of this chapter will focus on the more complex
paramagnetic species within this series: ﬁrst the mononuclear CrIII species will be
explored, and from there the next topic of interest will be the introduction of
exchange coupling by turning “on” or “off’ the semiquinone nature of the bridging

ligand.

2.4 Results and Discussion. Quartet Complexes of Chromium(III).
2.4.1 Theory. Electronic Structure. A quartet (S = 3/2) state has 2S + 1 = 4

degeneracy, which is manifested in any system of cubic or lower symmetry as two

Kramer’s doublets, 1115 = i 1/2 and mS = :1: 3/2 (Figure 2-13). This is a result of

79

Kramer’s rule, which states that if an ion has an odd number of unpaired electrons
the degeneracy of every level must remain at least twofold in the absence of a
magnetic ﬁeld. The splitting of these doublets is called zero ﬁeld splitting (zfs),
which arises from the point group being unable to support the four-fold
degeneracy of the state, and originates from magnetic dipole-dipole interactions
between the unpaired electrons. In our target molecules the spin density resides on
the Cr"I center, which has local cis-N402 (C2,) symmetry, and as a result should
have rhombic spin Hamiltonian parameters (vide infra), meaning in general that
the tensors describing the interactions in these systems will be anisotropic. In the
presence of applied magnetic ﬁeld, the most general spin Hamiltonian for a S = 3/2

system in the absence of hyperﬁne coupling is

A

H=ﬂeB-g-§+S-D-S (2.12)
The leftmost term on the right describes the familiar Zeeman interaction (Figure

2.13d), which breaks the degeneracy of i 1/2 and i 3/; Kramer’s doublets, resulting

in a linear stabilization or destabilization (depending on the Sign of the InS level) as
a function of the magnetic ﬁeld, B. The right-hand term accounts for zero ﬁeld
splitting effects, and is represented here in matrix form, where D is the zero ﬁeld
splitting tensor, a 3x3 matrix which quantiﬁes the interaction between the Spin

vectorsS. Because the zero-ﬁeld splitting matrix has a zero trace (X + Y + Z = 0,
where X, Y, and Z are the diagonal components of the zero ﬁeld splitting tensor),

the zero ﬁeld splitting can be given by two values,

80

«In

F.-

D = 7(X + Y) — Z and E = "£‘(X — Y) , resulting in the spin Hamiltonian of

equation 2.13.

A

H=ﬂeP-g-S+D[Szz—%S(S+1)]+E(Sf—SJ2,) (2.13)

where the parameter D describes axial splitting (gxx = gyy i gzz) of the zero-ﬁeld
states, while the parameter E is required to account for rhombic zero-ﬁeld effects
(ng 75 gyy 76 gm) (Figure 2-13c).18 While the presence of the parameter E indicates
a lower than axial symmetry, the magnitude of B does not in itself reﬂect a given
symmetry. The complexes reported herein have either D3, or C2,, local symmetry
at the magnetic centers, and therefore possess a “rhombic” EPR symmetry,19 that
is, in general the principal axes (the diagonal components of the g-tensor) are
nonequivalent, i.e gxx 9E gyy i gZZ as a result of different magnetic ﬁeld projections
along the principal axes. With different g-values for the principal axes EPR
transitions will occur at different positions in the spectrum, and can have a
dramatic impact on the appearance of the EPR spectrum, especially when zero
ﬁeld splitting is present.

Both 3 and its model complex have a d3 paramagnetic center, which has a
4A2 ground state in C2v symmetry. This ground state in isolation should exhibit no
zero-ﬁeld effects due to quenching of the orbital angular momentum, though
second-order spin-orbit effects, which result from admixture of excited electronic

states into the ground state, account for a small zero-ﬁeld splitting, even in Oh

81

D-‘-.hi

symmetry. Macfarlane59 has carried out calculations in trigonal symmetry and
concluded that the major contributions to the zero-ﬁeld splitting come from the

a4T1, a’Tz, and szz, with 4T2 contributing very little.

 

 

 

 

8- Free b.Oh c. Lower id. Field On
Ion g g Symmetryi
4A29
5 (egg) ........... l 3933
4F l
—— ZD ‘

 

- - — — - — -
a"
,.
,."
‘0
..
o
c"
..
-
-

 

 

 

 

 

é li%> é ............... 1 ................ g 968

ii 2) 5

i D>0 D>0 : 9,
E>0 B

 

 

 

Figure 2-13: The ground quartet state of a Cr(III) ion under Oh symmetry (4A2g)

consists of two Kramer’s doublets split by 2|D|. As the symmetry is lowered the

rhombic zero ﬁeld splitting parameter E affects the splitting, and as the system is
introduced to an external magnetic ﬁeld (B) the Kramer’s doublets split further

according to the Zeeman interaction.

82

 

__!.

FT — 1'"?th 7

 

Basis functions for the quartet spin system can be represented with respect

to the one electron spin functions 01 and [3 as:

1+ %) = laaa) (214a)

[4%) = %laaﬂ+ aﬂa + ﬂaa) (2.141))
|—%) = VIE-lam + 6616 + and) (2.14c)
|_%> = Imp) (2.14d)

For the spin Hamiltonian of equation 2.13 and the symmetrical quartet spin basis

 

 

set the secular determinant is: ‘9’ 6O
+%gzzﬂeBn + D — e gﬁeBg J35 0
AggﬂeBg+ é‘gzzﬂeB" — D _ 5 ﬂeBg— J31”: = O
l
\51‘: ﬂeBg+ —7gzz,6eBn-D—8 ‘JziﬂeBg-
0 JEE ‘JziﬂeBg+ —%gzzﬂeBn + D — g
(2.15)

where I: sin6cos¢, m = sint9sin¢, n = 0036 and g: = (gxxl iigyym). l, m, and

n are the direction cosines. This secular determinant is derived in Appendix B. In

all calculations presented below, the principal g-values are replaced by the

III

isotropic g value (1.98 for Cr ). Be is the electronic Bohr magneton (4.65334 x

83

Shh—'0

ll

10'5 cm'1 G”), and B is the magnetic induction (in Gauss). D and E are scalar
quantities and are the axial and rhombic zero ﬁeld splitting parameters,
respectively. Applying the appropriate angles and solving for 3 yields solutions
for the magnetic ﬁeld along the z-axis (0 = 0; (p = 0), x-axis (0 = n/2; (p = 0), or y-
axis (0 = TE/2; (p = u/2). Of course, one can solve for the eigenvalues of any other
orientation with respect to the magnetic ﬁeld by plugging in the correct angles.

For the axial case we need to only consider solutions along the z axis and
either the x or y axis, due to the degeneracy of these axes in axial symmetry.
Transitions that are not expected given the usual selection rule (i.e. Am, = i l)
arise due to the lowered symmetry of the system. To understand the origin of
formally forbidden EPR transitions occurring in this system we lOok to the secular
determinant, equation 2.15. A corresponding axial secular determinant is obtained

by setting gxx and gyy to the same value (called g i, called g-perpendicular and gZZ
takes the name g“, g-parallel) and by simply setting the rhombic splitting

parameter, E, to zero. The diagonal terms contain g” and the off-diagonal terms

contain g i. For B parallel to the z-axis the eigenvalues are

Wi% = Di%(gﬂeB) (2.16a)
Wiy2 = —Di%(g,6eB) (2.16b)

84

 

and the resulting zero-ﬁeld splitting is 2)D|. Off-diagonal terms all contain 819 so

that the different eigenvalues comprising the quartet state in a magnetic ﬁeld along
the x and y axes involve mixing of different pure Spin States.35

In the case of a quartet spin system in rhombic symmetry, the same
procedure used for axial symmetry is carried out. Namely, that the secular
determinant is solved for orientations corresponding to the x, y, and z axes. The
difference for rhombic systems is all three axes are non-degenerate and must be
considered individually. Again looking to the secular determinant, equation 2.15,
it is seen that further off-diagonal terms arise when E 7t 0. This additional state
mixing in rhombic system is important for describing the origin of perpendicular-

mode transitions given the typical EPR selection rule.

As described above, the magnitude of the zero ﬁeld splitting in axially

symmetric quartet systems between the m, = :l: 3/2 and mS = i 1/2 Kramer’s
doublets is 21D| (vide infra). D can be positive, resulting in the m, = i 1/2

Kramer’s doublet being stabilized, or negative so that the m, = ﬂ: 3/2 states are
stabilized. For rhombic quartet systems, the ratio E/D inﬂuences the magnitude of

the zero ﬁeld splitting, giving rise to Kramer’s doublets separated by“

AE=2|D|,/1+3(%)2 (2.17)

as shown above in Figure 2-13. Finally, it should be noted that the validity of m,

as a good quantum number depends on the magnitude of the zero ﬁeld splitting.

85

In general, for small zero ﬁeld splitting (|zfsl << h v) extensive mixing between the

various m, levels occurs, and m, is no longer valid. For large magnitudes of zero
ﬁeld splitting m, becomes a good quantum number. Because of this state mixing,

3), and

 

2),

 

the magnetic sublevels of quartet electronic states will be called [1),

[4) rather than by their pure Spin kets |i %) and li%) .

2.4.2 Theory. Selection Rules and State Mixing. In a perpendicular-mode EPR
experiment, microwave radiation is applied perpendicular to the magnetic force
lines. The magnetic component of the microwaves (Bhv) couples with the
magnetic moment of the Spin system, resulting in the transition. The probability

of a transition from a state m, to another m,’ is given by the modulus squared of
A)

moment, which along the x-axis is# = —g,BSx , assuming an isotropic g value.

2

 

 

the element of the secular determinant, P = |<ms ll‘Bl . Here u is the magnetic

A

When 5, operates on an m, function the value is

$WQJ%+%%-l (no

 

and such that the modulus squared of the transition element is zero unless m, = m,
:l: 1, that is, the ﬁnal state must differ by one in m, for the transition to be allowed.

This is the commonly known selection rule for EPR.

86

For parallel mode the microwave radiation is applied along the same axis as

the magnetic ﬁeld, and the magnetic moment is given by,” =‘gﬂSz , and

operates on the m, states as:
Szlms)=mslms) (2.19)

The result is that the probability is non-vanishing only is m, = m,’, that is, the
states must “look” the same, whether by having the same m, values or taking on
Similar m, values as a result of state mixing. In the case of small zero ﬁeld
splitting, state mixing occurs. In this case m, would cease to be a good quantum
number, as described above, with the result that formally forbidden EPR
transitions arise. This is a concept important to the accurate simulation of EPR

spectra.

2.4.3 High-Symmetry complexes of Chromium(III). The electron paramagnetic
resonance spectra of chromium(III) ions have been recorded in many different
high symmetry environments. In Single crystals, the ion must be doped in at ~ 1 %
concentration in order to avoid spin exchange interactions between the
paramagnetic ions. While in most EPR studies Chromium(III) has been doped into
many different ionic solids, most notably ruby,62 this section will focus on
coordination complexes of chromium(III). Tris(acetylacetonato)chromium(III),

Cr(acac)3, forms an isomorphous series with both Al(acac)3 and Ga(acac)3 and can

87

be doped into these materials in order to carry out single-crystal EPR studies.” 63’

64 In general, in all of these studies the EPR spectra can be accounted for with the

spin Hamiltonian of equation 2.13, with g values around 1.98, IDI = 0.6 cm”, and

IEI = 0.015 cm'] at low temperature. The axial splitting parameter, D, has been

reported as 0.7 cm'] for Cr(hfac)3 (hfac = monoanion of l,1,1,5,5,5-hexaﬂuoro-
2,4-pentandione).65 At room temperature the magnitude of the rhombic zero ﬁeld
splitting parameter E for Cr(acac)3 decreases to ~ 0.008 cm“, about half the low-
temperature value. One interesting application of this temperature-dependence of
the E parameter was carried out by Andriessen, who used Cr(acac)3 doped into
Al(acac)3 or Co(acac)3 to monitor phase transitions in these materials.'6 Finally,
electron-nuclear hyperﬁne interactions of Cr(acac)3 have been explored by
McGarvey on isotopically enriched samples of Cr53(acac)3, revealing hyperﬁne

- - -4 -l 66
Interactions on 1116 order Of 10 cm .

Given these spin Hamiltonian values, the
high symmetry systems can be treated with the same formalization outlined below
for lower symmetry (N402 coordination) systems.

Several reports on single crystal doped samples of Cr(acac)3 have employed
high resolution spectroscopic techniques (polarized emission, optically detected

magnetic resonance, excitation line narrowing spectroscopy), to obtain state

splittings and g-values of the lowest-energy, luminescent excited state and the

ground state. In all cases a value of ~ 1.2 cm'1 was obtained for the ground state

88

 

 

splitting (consistent with IDI = 0.6 cm']; in one such study MObius and coworkers

have assigned D as negative for Cr(acac)3).”’ 67’ 68 Over the years these studies
have led to some disagreement in the literature as to the nature and the degree of
splitting of the emissive state. As this topic pertains to the excited electronic
structure, it will be explored fully in Chapter 3.

Chapter 3 of this dissertation employs substituted complexes of Cr(acac)3 in
order to address fundamental issues of nonradiative decay in transition metal
complexes. X-band EPR spectra of several of these high symmetry complexes
were obtained to address the ground state Splittings. The spectra were collected in
a 9:2 butyronitrile:propopnitrile glass at 4 K, and a representative spectrum is
shown in Figure 2-14. This spectrum was simulated using the spin Hamiltonian of
equation 2.12. The values obtained for Cr(acac)3 and Cr(3-Brac)3 correspond to
the values observed in doped single crystals (described above), revealing very
little impact by the molecular environment on the spin Hamiltonian values. For all
complexes studied the axial zero ﬁeld splitting parameter was between 0.5 and 1.0
cm”, leading to a less than 2 cm'1 splitting in the ground state. The similarity of
the spin Hamiltonian parameters reﬂects the Similarity of the coordination

environment and electronic structure across this series.

89

 

 

A g = 1.9796
.~ D = 0.668 cm‘1

E/D = 0.0249

 

 

0 ' 10'00 ' 20'00 ' 30'00 ' 4000 I
B[G]

 

 

 

Figure 2-14: Experimental (black) and simulated (red) continuous-wave EPR
spectra of tris(3-phenyl-2,4-pentanedionato)chromium(lII) (Cr(3-Phac)3) obtained
in a 4 K glass of 9:2 butyronitrile:propionitrile. Experimental conditions: v =
9.4775 GHz, modulation amplitude = 20 G, conversion time = 80 ms, power = 63

11W. Simulation parameters are indicated in the ﬁgure.

90

 

Table 2.2: Spin Hamiltonian parameters for several substituted complexes of

Cr(acac)3 obtained in a 4 K glass of butyronitrile:propionitrile.

 

gm D 1cm") E 1cm“)

 

Cr(acac) 1.9792 0.576 0.0080

C’(3'Bra“)3 1.9796 0.664 0.0043

Cr(3'Phac)3 1.9796 0.668 0.0166

 

 

 

Finally, of additional interest is a time-resolved EPR (TR-EPR) study of
ruby and Cr(acac)3/Al(acac)3.62 The samples were excited into the ﬁrst spin-
allowed ligand ﬁeld state using a frequency doubled NszAG source (532 nm).
The experiment had a time resolution of about 0.5 us. In ruby, which has a 3 ms
lifetime for the lowest-energy excited state, 2E, the authors noted no signal
attributable to this state. The dynamics they observed, which were assigned
entirely to dynamics between spin sub-levels of the 4A2 ground state, where
initially emissive and then switched to absorptive after about 700 us, indicating
changing populations of spin sublevels within the 4A2 ground state. This
essentially relates to changing spin distribution in the ground state in a time-
resolve sense. The Cr(acac)3/Al(acac)3 system showed only an emissive signal,

but this result may be due to the much lower S/N recorded for this measurement.

91

 

2.4.4 Low-Symmetry complexes of Chromium(III): X-band EPR spectra and
simulations of [GaCr(tren)2(CA°“‘°at)](BPh4)2 (3) and
[Cr(tren)(DTBCat)](PF6). Magnetic susceptibility data published in our previous
report for complex 3 from 2 — 350 K reveal a temperature-independent magnetic
moment of “eff: 3.44 i 0.09 pg (u,p;n-,nly for S = 3/2 is 3.87 pm), indicating the

expected S = 3/2 ground state of Crm.

In general, CrIII centers with a cis-NaOz
local symmetry have g values around 1.98 and zero-ﬁeld splitting parameters of
[D] _<_ 0.5 cm'1 and ED of about 0.3, as revealed by an investigation of a series of
Cr"' complexes including [Cr(NH3)4XY]Z, cis-[Cr(2,2'-bpy)2XY]Z and cis-
[Cr(l,10-phen)2XY]Z (XY = HZO-HZO, HzO-OH', oxalate or acetylacetonate, and

I3, 69

Z is the counter-anion). These values provided useful initial guesses for our

simulations. The X-band spectra of [Cr(tren)(DTBCat)](PF6) and 3 were
simulated according to the spin Hamiltonian of equation 2.13.

The X-band EPR Spectra and simulations of [GaCr(tren)2(CAcat’ca’)](BPh4)2
(3) and its model complex, [Cr(tren)(DTBCat)](PF6) are shown in Figure 2-15.
The spin Hamiltonian parameters for both species are presented in Table 2-3. The
Spectra of the complex and its model are similar in almost all regards, and the
differences are almost certainly due to impurities in 3, of which one could be the S

28, 70 .
The spln

= 1/2 complex 2 and another the S = 5/2 complex 6 (Figure 2-16).
Hamiltonian parameters for the two systems fall within the range of values

r ePorted for CrIII complexes with a local cis-N4Oz ligand environment.

92

 

 

 

 

 

 

 

 

 

 

 

 

 

354W

‘ I‘M)

 

 

 

1000 I 2600 ' 30'00 ' 4000 I 5000 ' 60'00
B[G]

 

 

 

Figure 2-15: Experimental (black) and simulated spectra (red) for (a)
[Cr(tren)(DTBCat)](PF6) and (b) Complex 3. The spectra were acquired at 4 K,
with a power of 0.32 mW at 9.4595 GHz. The transition roadmap of 3 is shown

above,

93

 

The spectra Show no temperature dependence up to 60 K. Application of

equation 2-17 results in a splitting of 0.894 cm'1 for |D| = 0.392 cm"1 and |E| =

0.124 cm”, the experimentally determined values for complex 3. Here, in our
simulations and calculations, it is assumed that D > 0, although we have not been
able to prove this experimentally because the small magnitude of D allows for a
nearly equal Boltzmann distribution of both quartet and doublet Zeeman states,
even at 4 K.

The parameters were also applied to calculate energy-level diagrams for
these two molecules by solving the secular determinant along the magnetic x, y,

and z axes (vide infra).18 The energy level diagrams reveal signiﬁcant mixing
between the I: %) and [i %) manifolds, so that formally forbidden transitions are

allowed. The ﬁeld positions of the transitions, indicated by a red arrow, are

consistent with the observed spectrum.

Table 2-3: Spin Hamiltonian parameters, as obtained through simulation, for 3

and its model complex, [Cr(tren)(DTBCat)](PF6).

 

g“. gyy, gzz chm") E (cm-1)

 

model 1.985,1.982,1.970 0.396 0.122

 

3 1.980, 1.975, 1.970 0.392 0.124

 

 

 

 

 

 

94

 

 

 

 

  

 

 

 

 

        

ﬂn'ﬂffltﬂlml‘n‘vmmma-a-a-p

— ' I .0 ' 0 . . o O o~~*.-‘q

.l

 

 

 

i Ga(lll)-SQ-Ga(lll) (2)

- ii Cr(lll)-Cat-Ga(||l) (3) -
. J - l, . I . Cr(III)-SQl-erll) (6)
0.0 0.1 0.2 0.3 0.4 0.5 0.6

BIT]

 

 

 

Figure 2-16: Impurity signals in the experimental X-band spectrum of complex 3
due to the homometallic species 6 and 2 (inset). All spectra were acquired at 4 K

and 9.45 GHz.

The transition roadmap, the angular variation of the resonance position plot,
which is presented above the EPR spectrum (Figure 2-15), shows the ﬁeld of
transition (abscissa, in Gauss) with respect to orientation (ordinate). Lines on the
diagram plot the ﬁeld of resonance for a particular transition as a ﬁlnction of the

orientation of the spin system (the polar and azimuthal angles) with respect to the

95

 

direction of the magnetic ﬁeld. The transition which each line corresponds to is
shown with reference to the energy level diagram. The path follows the edges of
an octet of a sphere, and is as follows: starting at the z-axis the angle 0 is varied so
that orientations in the xz plane are sampled until the x-axis is reached, then (1) is
varied so that orientations in the xy plane are sampled until the y-axis is reached,
and ﬁnally rotating in the yz plane to return to the z axis. To gain an

understanding of the angular variation of the position of resonance, this concept is

shown schematically in Figure 2-17 for a simple S = 1/2 system under rhombic

symmetry, where the position of the I-%) —>I+}é) transition with respect to the

applied magnetic ﬁeld varies with angular orientation. In the diagram of Figure 2-
17 the variation with orientation is due to the anisotropy in the g-tensor. However,
our Cr(III) systems are essentially isotropic with respect to the g-tensor, and are
treated as such in the calculations, with an isotropic g value of 1.98. The
orientation dependence in these systems arises from mixing of the pure m, states
due to the relatively small magnitude of the zero ﬁeld splitting. The diagram of
Figure 2-15 reveals extremely wide resonances, which are common for transition

metal systems.

96

 

 

 

 

 

 

 

 

 

 

 

 

m»

Figure 2-17: Schematic representation of an angular dependence diagram for a

 

1

simple S = 1/2 system under rhombic symmetry.

2.4.5 Effect of Zero-Field Splitting Parameters on the Energy Profile of a

Quartet Spin System. To best understand the origin of the transitions observed

 

for both 3 and the model complex, it is paramount to build up the spectra piece by
piece using the Hamiltonian of equation 2.13. The simplest quartet spin system is
isotropically symmetric, meaning that the scalar quantities D and E are zero, and
only the electronic Zeeman term is operative (we are also considering an isotropic
g tensor, with g = 1.98). The zero ﬁeld splitting is zero because the symmetry of

the system does not allow for any zero ﬁeld splitting, as the zero ﬁeld splitting

tensor ( D) is traceless. The result is the energy level diagram on the far leﬂ of

Figure 2-18, where all allowed (Ams = i1 ) transitions are at the same ﬁeld and

the energy proﬁle is the same for all possible orientation of the molecule with
respect to the applied magnetic ﬁeld. Figure 2-19, left, shows the transition
diagram for an axially symmetric quartet spin system. When D = 0, the system is

isotropic, and one can see that all allowed transitions for the plotted orientations

are occurring at the same ﬁeld, with the exception being the g’ = 4 (Am, = —2 )

transition). All possible transitions in an isotropically symmetric quartet spin
system are shown in the transition roadmap, Figure 2-20. As expected from the
symmetry of the system, the ﬁeld at which a transition occurs does not change
with the orientation of the spin system with respect to the direction of the applied
magnetic ﬁeld. The transitions listed on the transition roadmap correspond to the
states of the energy level diagram to the right of the roadmap. The simulated
Spectrum (Figure 2-21, top, D = 0) does not Show the forbidden transitions

revealed by the roadmap as their transition probabilities are extremely small.

98

 

Figure 2-18: (Fasting page) Energy level diagrams showing how the experimental

- f 9 v T -

1:3: 1)) as; $657.21: (E) zero ﬁeld splitting parameters affect the energy profile

\

(f it: g?’:-nr.i ‘A state for the magnetic ﬁeld parallel to the principal axes of the
7:: r. an isotrOpiC system the mo Kramer's doublets of the ground state are
0-53:2: ' id the energy proﬁles are the same for every orientation with respect
or i.e magnetic ﬁeld. For an axial system the Kramer's doublets are split by 2D.
and the x and y axes are degenerate. For rhombic systems an additional zero ﬁeld
splitting parameter is needed to describe the splitting between the Kramer's

doublets. and the energy proﬁles for the magnetic ﬁeld along the x. y. and z axes

99

 

A

y (cm'1

0')

Ener

 

1.0
0.5

0.0

-0.5

-1.0

-1.5

Isotropic

D = 0.396 cm-1

:5:

4 cm'l

 

Axia

 

 

Tevery 6, (p

.-

r

I

I

 

—

I Y

I

.1

IMJ_J_L 1

 

 

0.0 0.2 0.4
BIT]

Energy (cm‘1)

2.0
1.5
1.0

0.5
_‘ 0.0 -
‘ 0.5

‘ -10
0.60.810

1.0 -X and y .
0.5 - .
0.0 — J
0.5 - -
-1.0 - .
,. J

-1.5 - 1
1

L
r

 

*I I I Y

I j I ' I '

N

l- -

 

 

 

 

l a l A I A l

 

0.0 0.2 0.4 0.6L0.8‘1.0
B1T]

Energy (cm‘1)

V r I V I V
. .
ﬂ
4 >
d b
I D
A l n l J l A

100

0.12

2.0
1.5
1.0
0.5
0.0

-0.5
-1.0

1.5
1.0
0.5
0.0

. -0.5

-1.0

-1.5

1.0
0.5
0.0
-0.5
-1.0
-1.5
-2.0

r

thmbie. .

 

I“ I '

r J
I- I4
I .
r 1
l- 1 4

, 2 DZ+3E2)1/2 .
1
‘ \ l

.Lml Pl.

r

I

I

I

 

 

v l v T j’ I I I v

u

I

r TTTTTTI
1 .441

I

 

 

h

1

d

 

 

 

0.00.20.40.608 1.0
BIT]

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1.0 -- ------- . .
. 2. l-%>—>|+‘2> 030
22 I‘X2lfl‘y2l .
O8 21 23 |:y2)—>|:%) 025 x1 x2
‘2 0.6 0.20 -
E Xyl m Y1
22 0.15
04 ——————————————————————— -_1 Z1
0.10
. Z3
0.2
XY2 23 0.05;
0’ 0.2 0.4 0.6 0.8 1.0 0 f 0.2 04 0.60.8 '10
BIT] 131T]

 

 

 

Figure 2-19: Transition diagrams for quartet systems. Left: Resonant ﬁeld Of
transition at X- Band frequency for an axial quartet spin system with variable D.
The labels indicate the origin of the particular transition as either the magnetic
ﬁeld parallel to the molecular z or xy axes, which are degenerate under axial
symmetry. The line drawn across corresponds to the experimental value of D for
the model complex, 0.396 cm“. Right: Resonant ﬁeld of transition for a quartet
system under rhombic symmetry for D = 0.396 cm"1 and variable E/D. The line
drawn across the plot corresponds to the experimental value of E/D for the model

complex, 0.308.

101

 

 

0

 

(P

' o o .1 —-
g '01 'o '01
(rule) Aﬁleua

L.
o

 

ILJmlmimM

0 0.2 0.4 0.6 0.8 1.0
BIT]

 

 

 

0

y > I
I Ams=3 l4>-12> |3
|

 

 

 

 

 

 

0.1 0.15 0.2 0.25 0.3 0.35
8m

 

 

 

Figure 2-20: “Transition roadmap” of an isotropic quartet spin system. The
diagram shows the position of transitions as a function of orientation with respect
to the magnetic ﬁeld. There are only three possible transitions in an isotropic

system, and only the transition at g’ = 2 (around 3300 G) is formally allowed.

102

 

A. e‘; ‘31.
.. .. _. .«r' ‘3‘... ‘t.w‘n"li' be

 

Figure 2-21: (Facing page) Simulated Spectra of quartet spin systems showing the
effect of the axial and rhombic zero ﬁeld splitting parameters. Top: Simulated
Spectra for variable D, from 0 to 0.5 cm". The spectrum in red corresponds most
closely to the experimental value of D for the model complex. Bottom: Variable
E/D for D = 0.396 cm'], the experimental value of the model complex. The
Spectrum in red corresponds closely to the experimental E/D value for the model

complex (0.308).

103

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

BIT]

104

 

 

 

A. Introduction of Axial Zero Field Splitting:

i. D = 0.396 cm”, the experimental value: Introducing the experimentally
determined axial zero ﬁeld splitting parameter (D = 0.396 cm”) drastically
changes the spectrum. This is shown in the energy level diagrams of Figure 2-18
(middle column). Now the energy proﬁle is dependent on the azimuthal angle
(6), but all orientations in the xy plane are degenerate. The effect of this
particular symmetry on the ﬁeld at which EPR transitions occur is seem most
clearly in the transition roadmap of this system, Figure 2-22, where the ﬁeld of
transition for the two observed transitions in the xy plane does not change as

orientations in the xy plane (changing ¢) are sampled. To gain an insight into the

effect of increasing D, one can look to the transition diagram for an axial quartet
system (Figure 2-19, left) or the simulated spectra (Figure 2-21, top). As
described above, for D = 0 (isotropic system), all allowed transitions occur at the
same applied magnetic ﬁeld. As D is increased these transitions diverge from g’ =
2. The principal axes from which these transitions originate (either 2 or xy) are
labeled. In Figure 2-19 (left) a line is drawn across the diagram which
corresponds to the experimentally determined value of D for these Cr(III) systems.
The spectrum that most closely matches this value is Shown in red in Figure 2-21
(tOp). The transition roadmap (Figure 2-22) reveals wide resonances, and the
energy proﬁle changes dramatically with orientation as pure spin states mix. It

Should be noted that while the transition diagrams and transition roadmaps provide

105

information on the positions of transitions, it must be kept in mind that all
transitions Shown will not have the same transition probabilities. The simulations
take this probability into account in order to construct the spectrum. Due to this, it
is important to consider ﬁrst the various simulations, which take into account the

probability of a transition over all angles.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Z I j T ,2” r a- 2.0
0 __ .1
l3)—*|2) f mi 05
Y r - f :0
~ ‘l --0.5 m
. , a
- |4) -1.0 L”
(p l2)—>|1) :?i111;:,:: ‘0
I "“9 £11 3
x - 5 W. 0-5 V‘
.” j 0
|3) j-O.5
0' - 1 -1.0
l4)—>|3) i 4) 4.5
0 0.2 0.‘4 0.6 08 10 1.2 00.120401608110
Bm BlT]

 

Figure 2-22: “Transition roadmap” of an axially symmetric quartet spin system,
where D = 0.396 cm”, the experimentally observed value of the model system.

The transitions are labeled according to the energy-level diagram on the right.

106

ii. General study, 0 _<_ D S, 0.5 cm". Simulations of axial quartet systems
(E = 0) were carried out over the range 0 S D S 0.5 cm", and are plotted in Figure
2-23. In the limit of no zero ﬁeld splitting (D = 0 cm") only one peak can be

observed, the g’ = 2 resonance, which in this case is actually three different
resonances occurring at the same ﬁeld: |-%)—>|—V2), |-%)-+l+%), and

l+%)—>l+%). In the limit of an isotropic g-tensor these transitions are

completely orientation independent. Any introduction of zero ﬁeld splitting then
signiﬁcantly complicates the spectrum. For example, multiple resonances which
are nearly equally spaced are observed at D = 0.05 cm”, and as D increases these
lines begin to move apart. The spectrum then simpliﬁes again for larger values of
D. This can be understood most simply by monitoring the transitions across a
series of energy-level diagrams of the quartet ground state (Figure 2-24). At the
limit of no zero ﬁeld splitting the x, y, and z axes are equivalent, that is, the
system is isotropic, and there are three different transitions occurring at the same
magnetic ﬁeld. Along the magnetic z axis, as D increases these transitions

diverge, producing transitions both upﬁeld and downﬁeld of the transition around

3300 G (black arrow), which remains constant, and is best described as l-%) —+

|+%> , although, as discussed above (vide supra) when hum z D this description
may not be the most accurate due to mixing of the pure spin states. Eventually the

transitions present at low zero ﬁeld splitting, which are best described as I“%) ——r

107

|—%) and |+%)_, |+%) (green arrows) fall out of resonance, and the I715) —>

|-%> transition (again, this may not be the best description) is reintroduced at

higher values of D (turquoise arrow). The situation is similar along the x and y

axes, although this time the states can not be expressed as having pure 1/2 or 3/2

spin (vide supra). The transitions for the axial case are summarized in a transition
diagram (Figure 2-25). In this diagram the transitions are labeled according to
their origin, for B either parallel to the z axis (labels 2], 22, and 23) or the x or y
axes, which are degenerate in the case of an axially symmetric system (labels xy)

and xyz). The diagram clearly shows only one transition in the limit of no zero

ﬁeld splitting (with the exception of the spin-disallowed |—%> —* 14%) (Am. = 2)
transition, which is plotted but not observed experimentally) and a further
simpliﬁcation of the spectral appearance occurs for D > 0.50 cm", where only
resonances at g’z 2 and g’ z 4 (where g ' = hV/ﬁ'pB) remain within the range 0 S B
_<_ 7000 G.

Figures 2-26 and 2-27 Shows the energy level diagram and transition diagram
for an axially symmetric system at W-band frequency (resonances at about 3.3 cm'
1). Although the pattern shown in the transition diagram is the same as that for the
transitions at X-band frequency, the important point is that the range of the plot is
Up to 8 cm", meaning that one’s ability to assign the proper Sign and magnitude of

the zero ﬁeld Splitting is greatly enhanced at this frequency.

108

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

l ——1.00
I r"
J ——0.50
r , -
J ——0.45
r " '
j - ——0.40
p , ,
—0.35
mom-1) , . ——0.30
l g ——0.25
r F '
| g g —0.20
IT I ' ,
I l L i - ——0.15
1* IT I r
g 1 In . ——0.10
IT r l r
.s 111 n, —-0.05
l —0.00
r I I r I l r l ' r ' I '
o 2000 4000 6000 8000 10000
816]

 

 

 

Figure 2-23: Calculated X-band EPR spectra for an axial quartet spin system

with variable D (in cm”).

109

 

1.59
1.03
0.51
0.01

0.51

-101

1.5«
1.01
0.51
0.0-

0.51

Energy (cm'l)

20'
1.53
1.03
0.51
0.03
0.51
-1.03

 

 

-1.03

 

1.5-
1.01
05‘
0.0-

05'

-103

-1.5-'

 

 

III ' I ' I ' '

BIT]

1.03
0.5
0.0-
0.51
-10.
-1.53

1.0-1
0.51
0.01
0.51
-1.03
-1.53
1.5-
1.01
0.53
0.01
0.53
-1.03
-1.53
1.5-
1.03
0.53
0.03
0.53
-1.01

 

0.1

 

5 ....... . a. .
0.0 0.2 0.4 0.6 0.8 1.0

BIT]

 

Figure 2-24: Energy level diagrams for B parallel to the x (01' y), and z magnetic

axes for an axial quartet spin system with variable D (in cm”).

110

 

 

 

1'0 '2. mwm
22 l_%>_’ |_%>

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.8 - 21 23 li%>->|i%>
‘TE 0.6
8 xy1
D 22
0.4
0.2
XY2 Z3

 

 

 

 

 

 

** o 0.2 0.4 ‘06 0.8 1.0
BIT]

 

 

 

Figure 2-25: Resonant ﬁeld of transition at X-Band frequency for an axial quartet
spin system with variable D. These transitions are color-coded to those of Figure

2.24.

111

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_~ _._.- _.__ ._ _L _A__ _.i_. _ (w W V f7.-

0 ' 20000 #40000 60000“ 80000
BIG]

 

 

 

Figure 2-26: Resonant ﬁeld of transition diagram along the X, Y, and Z magnetic
axes for an axial quartet spin system with variable D (in cm'l). The resonances are

at W-Band frequency, and are color-coded to the energy-level diagram (Figure 2-

27).

112

 

 

 

-101 -15
15? 101
6" 3‘. 0 30
g 5‘. << -5:
v — - 3'04
. \ J
>. .
9 1?, "5*
m .
C
m

 

a. _.
0010010
I
A' _x
.<?.S".9.€”.<?
R A
0

-5;

40+? \ -10-

'''''''''''

 

 

 

 

 

 

 

Figure 2-27: Energy level diagrams along the X, Y, and Z magnetic axes for an
axial quartet spin system with variable D (in cm'l). Resonances are shown at W-

band frequency.

113

 

B. Rhombic model:

i. D = 0.396 cm'1 and E = 0.124 cm], the experimental values:
Introduction of the rhombic zero ﬁeld splitting parameter, E, results in subtle
changes to the energy proﬁle, as modifying the polar angle (¢) now results in
different energy proﬁles, and the principal x and y axes are now nondegenerate.
The outcome of these two factors can be seen by comparing Figure 2-28, the
transition roadmap for the experimentally determined values of D and E, to Figure
2-22, the corresponding diagram for the axial system. In the rhombic case, while

the two separate [4) —>|3) transitions blend into one, the |3) —> |2) transition,

which had resonances along the principal z-axis in the axial case, is now a closed
loop, with transitions only for O < 6 < 7r/2. Additionally, while changes in the

|2) —> [1) and [4) —> ‘3) transitions may appear subtle, the result is a drastically

different EPR spectrum (Figure 2-21, lower, in red). This is due primarily to
variation with the polar angle, creating extremely wide resonances.

The right side of Figure 2-19 coupled with the series of spectra of Figure 2-
20 demonstrate the effect of changing the ratio E/D while keeping the value of D
ﬁxed (in this instance, for the experimentally determined value of D = 0.396 cm").
For the transition diagrams (Figure 2-19) it must be kept in mind that only the
orientations along the principal axes for which the magnetic ﬁeld is applied are

PIOtted. Therefore, the looping [3) —> l2) transition observed between the principal

Z and y axes in Figure 2-28 will not appear on this diagram. However, the

114

 

transition diagram still has some utility: the red dashed line drawn across the top
of Figure 2-19 corresponds to the experimental E/D value for our systems, and

predicts very well the positions and origins of the observed EPR transitions.

 

 

 

 

 

 

 

 

 

 

 

 

 

r__*..._‘____ _,_,__ _.. ’ :7: a r ﬂ T‘ ' '1): :2.0
/'/,_M ‘ ‘ L ~10
/ \\|3>—)|2> T '2) _
K T -0.0
3;“— '2>"" ’" ﬂ 3-10
K “Eu“:w :xti.-t:; :;;:1'0
f ’fﬂ _ ‘ . m
Jig—43> I If H ' ﬂ _‘ (3
\ / T 3 '2) 10.0 3
\ / .- . a
/ : 4) 1-10 3
'f “ . ......... — c:
z! . Py. - . ....... _10
/ \ . -.
\ ~_ 3 _
\\""““r L I > L-1.0
“RENE—R“ "x. g L 4 > _‘
r \\ L J #,_A+ ‘

 

0.21 0.0 08 1:0 1.2 ‘0 0.2 0.4 0.6 0.81.7)
3m Bm

 

 

 

Figure 2-28: “Transition roadmap” of a rhombic quartet spin system, where D =
0.396 cm'land E = 0.124 cm", the experimentally observed values of the model

system. The transitions are labeled according to the energy-level diagram on the

right.

ii. General study of rhombic quartet systems. For the case of a rhombic
quartet spin system a series of spectra were obtained by selecting a value of D,

ranging from 0.05 cm'1 to 0.50 cm", and the ratio E/D was changed until the

115

 

theoretical maximum of 1/3 was reached, producing the series of spectra and
transition diagrams presented in Figures 2-29 through 2-40. This maximum value
is the result of a convention which places the x-axis at the highest energy.18 For
the transition diagram the eigenvalues are solved for a certain value of D (O < D <
1 cm") and viable solutions are simply divided by D so that the ratio E/D can be
plotted. In the case of rhombic systems the spectra are much more difﬁcult to
describe than in the axial case. Three separate cases will be examined to gain an
understanding of the origin of transitions across this series: 1.) spectra up to D =
0.15 cm], with a detailed look at energy-level and transitions diagrams at D = 0.10
cm], 2.) evolution of the spectra from D = 0.20 cm'l to D = 0.30 cm", with a
detailed examination of the spectra for D = 0.30 cm], and 3.) evolution from D =
0.35 to 0.50 cm].

The most striking feature of the spectra for D S 0.15 cm'l (Figures 2-29 and

2-30) is the high density of transitions between 1000 and 6000 G, with relatively

few or no transitions at higher magnetic ﬁelds. This is due to the fact that for D S

0.15 cm", the zero ﬁeld splitting of 2|D| is less than or equal to the energy of

resonance, ~0.30 cm], with the result that certain transitions are accessible which
are not accessible at higher zero ﬁeld splitting. At D = 0.05 em], all of the
transitions are contained within the window 1000 < B < 5000 G, and there is very
little change in the spectra as E/D is varied from 0.00 to 0.33, reﬂecting very small

Changes in the energy proﬁle. In general, characteristic spectra at this value of D

116

 

will display a number of transitions in the g = 2 region, with. several transitions at
lower magnetic ﬁelds.

As D is increased to 0.10 cm'1 a pronounced spreading out of the peaks
away from the g = 2 region is observed, although the transitions are still contained
within the window 0 < B < 6000 G. A characteristic peak is observed around 600
G. The energy level diagram, Figure 2-31, which is summarized by the transition
diagram, Figure 2-32, reveals that a large number of transitions, representing every
possible combination of initial and ﬁnal spin states, are capable of being probed at
this value of D. As D is increased to 0.15 cm'1 certain transitions with large
oscillator strengths are pushed to lower ﬁelds, as the splitting of the Kramer’s
doublets now equals the microwave quantum at which spin resonance occurs. The
observed spreading of the spectrum continues.

For D _>_ 0.15 cm'1 the spectra begin to simplify, to some extent, as certain
transitions fall out of resonance due to the magnitude of the zero ﬁeld splitting.
Another pattern emerges which is most easily observable in a transition diagram—
the solutions for B parallel to the z-axis have a parabolic shape, and solutions for
B parallel to the x-axis converge at E/D = 0.33. This means that for E/D = 0 two
resonances exist which begin to converge around g’ = 2 as E/D is increased. This
is observed in the series of spectra for D = 0.20 cm], where resonances at 2000
and 4700 G for E/D = 0 converge around g’ = 2 at the maximum value of E/D.

The same pattern is observed for larger values of D, but as D increases the initial

117

 

position of the higher ﬁeld transition moves to ever higher ﬁelds. In the energy-
level diagram and transition diagram for D = 0.30 cm'1 (Figure 2-34 and 2-35,
respectively) one can observe the parabolic transition along the z-axis (red
arrows), which falls out of resonance around E/D = 0.29, and the convergence of
the solutions along the x-axis as E/D is increased (green and orange arrows).

As D increases beyond 0.30 cm'l the transitions attributed to the z-axis
move to higher magnetic ﬁelds, and a second parabolic transition emerges. Again,
the most conspicuous feature of this series of spectra (Figures 2-36 through 2-38)
is that as E/D approaches the limit of 1/3, several transitions converge around g’=
2, while two others converge around g’ = 4. The energy level diagram for D =
0.50 cm], Figure 2-39, reveals that this is indeed the case: the transitions
converging at g’ = 2 can be attributed to transitions along the x-axis and those
converging at g’ = 4 correspond to transitions along the z and y axes. The
transitions that converge around g’ = 4 vary little with changing E/D, as they are
contained within a single split Kramer’s doublet. The parabolic transitions of the
z-axis fall out of resonance at E/D = 0.17 and 0.27. The transition diagram of
Figure 2-40 summarizes the allowed perpendicular-mode transitions within this
system.

The series of spectra for D = 1.00 cm], Figure 2-3 8, reveals that the trends
continue even with very large values of D. Again transitions converge around g’ =

4, attesting to the fact that these transitions occur within a single split Kramer’s

118

 

doublet, and their ﬁeld of resonance should therefore, to a ﬁrst approximation, be

independent of the zero ﬁeld splitting.

119

 

 

 

 

 

 

 

#-
f
,
P
w
0)

 

 

 

 

 

 

 

 

* t .. E _ —0.20
E/D "
. LU ‘ﬁﬁ ——0.15

 

 

 

 

 

 

 

 

0112-- ———0.10
'1
h A —0.05
, J. .I ——0.00
I r ﬂ 1 r l r l 1
o = 0.10 cm'1

 

 

 

 

ll. ..1 . ——0.33

1. I I“ —0.30

1.. .. .1“, .. —0.25

- ||,,. '11‘ ——0.20
E/D *'

 

 

 

 

 

 

 

 

 

 

 

 

|.. . [.1 L . . —0.15
T

. .. I 1‘ ‘4'“? - ——0.10

1 , M 1 1% _ —0.05

. 1040 —0.00

.. r.

.I ,
0 02 014 0.6 0.8 1'.0
BIT]

 

 

 

Figure 2-29: Calculated X-band EPR spectra for quartet spin system with D =

0.05 and 0.10 cm" and variable E/D.

120

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

D=0.15cm'1
J, . —0.33
{L . g g —0.30
I 1% .- - ——0.25
, -. —0.20
E/D r
s s ——0.15
L4 ﬁ__‘ﬁ a __0.10
Lu. #1# —-—-0.05
r r'r r1
II ll #4 A ——0.00
T ['_r I I
I I I I f 1 u r
D=0.20 cm“l
I I h -—-—0.33
010.03. ——0.30
l W T '
_Ln Mg, . —, 0.25
EID __‘L~L_,+ﬁ +4.; I —-—-0.20
490+ ,nrﬁ —0.15
W1)
WA ———0.05
W2
. I J , —0.00
. I ' I: ' I I. 'l I I ﬂ f '
0 0.2 0.4 0.6 0.8 1.0
BIT]

 

Figure 2-30: Calculated X-band EPR spectra for quartet spin system with D =

0.15 and 0.20 cm" and variable E/D.

121

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Z x y EID
2- 1
1« 1
11 0,
o- 1 0.3
0‘ - / _14
-1.
—1T -2"
21 1
1 1 11 1
1 14 01 0.2
1 A < 0:
‘TE 01 V? -11
3 -1. '11
a 1 _2.
(T) 21 ‘ 1.
C 11
m 1‘ g 0 1
1 / 0 04 .
‘ 01 < 1 Q? _1.
1 -11
1 '11
1‘ 21 11 / 11 0
. 1 /
1 1 01 ‘ O
1 0+ 4
1 is '1" -1‘
1 __14 1
» 0.0 (14 0T8 020 ' 04 018 T 010 0.4 0.8 T
, BIT]

 

Figure 2-31: Energy level diagrams along the X, Y, and Z magnetic axes for a
quartet spin system with D = 0.10 cm'I and variable E/D. The resonances are at

X-band frequency, and are color-coded to Figure 2-32.

122

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.303 *1"
0.25% 22 x3
[ Y2
020:.-
Q Y3 .
LIJ .
Y1
0.15“.
z \
1 1
1 22
0.10 L X2 ‘
0.05;-
0 0.1 “0.27 10.13: 0.4 0.8 0.6
BIT]

 

 

 

Figure 2-32: Resonant ﬁeld of transition at X- Band frequency for a quartet spin

SYStem with D = 0.10 cm" and variable E/D.

123

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

D=0.25 cm'1
. ——-0.33
11. M I -
.12.... . , 79.30
_lL—qu—Jﬁ-a l —.9'25
E/D __1L_.1,__4... . - 9'20
4” ———0.15
_LL—Jrw .- j - -
| 1 l- A -——0.10
r" ”I '
‘ Jg —0.05
_211 0 .TC-
1 4 , —--0.00
r 1 r
I ' I ' l ' l I '
o=0.30 cm'1
1 1 1 g —0.30
1 | 1, , ————0.30
r0 j '
I1 I ‘ . ‘ --—0.25
r7 j V
- _ —0.20
Em _aLLJr - . - .
1| 1 - A —-—0.15
F ' *0 *
. , ,gﬁ 7—0.10
1) p —0.05
r '
J , —0.00
r' l
I ' ﬁ f l 1 T ‘ I 1 l '
O 0.2 0.4 0.6 0.8 1.0
BIT]

 

 

Figure 2-33: Calculated X-band EPR spectra for quartet spin system with D =

0.25 and 0.30 cm'1 and variable E/D.

124

 

Energy (cm'1)

 

ﬁﬁMN

 

 

020 ' 014 020T

 

 

\ 129.7 . AZ.._,,.. 0 -1‘
_::“— .‘Vv-".

 

 

v.-
--—“’
. 1‘
‘x...
3”"; WV
2"?"
mi
.,\\
._f—_.4-'">
1 ‘ ~ T F
-_\‘

 

 

0.0T 07: T08

BIT]

 

0T0 034 V 0.8

EID

0.3

0.2

0.1

 

Figure 2-34: Energy level diagrams along the X, Y, and Z magnetic axes for a

quartet spin system with D = 0.30 cm'1 and variable E/D. The resonances are at

X—band frequency, and are color-coded to Figure 2-35.

125

 

 

0.30 '*

 

 

 

Y1

 

0.25 V

 

 

 

 

 

 

 

 

0.20 ? X1 y2

E/ D

0.15:

 

0.10 2

 

 

 

 

 

 

0.05 i Z

 

0 0.2 0.4 0.6 0.8 1.0
BIT]

 

 

 

Figure 2-35: Resonant ﬁeld of transition at X- Band frequency for a quartet spin

System with D = 0.30 cm"1 and variable E/D.

126

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0:035 em'1
1 I 1 g 033*
1 |1‘ - 1 0.30
.1 ll 4 4 0.25
1 W ‘
J1 -—-———0.20
EID - . re 1 e
-1 Jr? _g ——0.15
I r ' I
F - g, ' ———,—0.10
ll Z W. —0.05
F
1 —0.00
1 [I r I T II * I I *
o=0.40 cm1
L 4? - 0.33
L erﬁ . 0.30
_11.__.1rl_,lﬁ - —,—0-25
EID ,L +L ' JH j 0.20
.1 J l - ——0.15
r
14 V, ._°-10
11 Z —0.05
F
1 , ,—0.00
I ' rI ' I ' I I" I F F 1
O 0.2 0.4 0.6 0.8 1.0
BIT]

 

Figure 2-36: Calculated X-band EPR spectra for quartet spin system with D =

0.35 and 0.40 cm" and variable E/D.

127

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I L_ mg _—"__0.15

[J L —0.10
r— - -

T ———0.05

7—000

1 T j I T I I
D = 0.50 cm 1

L + .- ——0.3e

 

L If — Q39
—L——i—+‘ . _ 0.25

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.1 g ——-— 0.20
E/D r ‘ r "F 0
10F 14 g —:0.15
b — 0.10
r C .
11 —:— 0.05
F‘
1 g 0.00
r. - -

 

0 0.2 0.4 0.6 0.8 1.0
BIT]

 

 

 

Figure 2-37: Calculated X-band EPR spectra for quartet spin system with D =

0.45 and 0.50 cm'1 and variable E/D.

128

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o=1.0 00cm“1

1 +T ——0.33

1 + -——-—0.30

.1 AF ——0.25

EID .11 1 .g ———0.20
1V —0.15

10 . ———0.10

11r . ——0.05

1 —0.00
. . . r. . , . , . T . . .
O 0.2 0.4 0.6 0.8 1.0

BIT1

 

 

Figure 2-38: Calculated X-band EPR spectra for quartet spin system with D =

1.00 cm'I and variable E/D.

129

 

 

 

 

 

 

 

 

Z X y EID
1
2‘ 1 1~ -. 1,,
I 7 , -
1< , ‘ o.
0‘ * I -1
. _ - -1‘ ,
-1. l g -21
2. f 1‘ . 11 '7’,
q " f I _
1* ‘ ‘ .
A 1 ...1/ 0 O - I 02
'E 01 - — 1711... -1: '1
3 _1. ” ‘ ‘ ,
3 2 '2
E 1 ' 1, I 11 _ if I
”-1 1: 1 I 0‘ ~ 01 0.1
0 I
. I: I -1. -11
-1. -
2 " 1‘ _J I 11 1 . 0
11 / O O
0 . _ _ _
.. F, 4’»- 1 1
-‘| - . . 1+4 f . . . . - . -
0.0 0.4 10.8 0.0 0.4 0.8 0.0 0.4 0.8
BIT]

 

 

Figure 2-39: Energy level diagrams for B parallel to the X, Y, and Z magnetic
axes for a quartet spin system with D = 0.50 cm'1 and variable E/D. The

resonances are at X-band frequency, and are color-coded to Figure 2-40.

130

 

 

0.30 ,

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

, X2
0.25 x1
, 22
Q 0.20. Y1
UJ ’ .
0.15 z1 ' 23
0.10’
Y2
0.057

0‘ ' 0.2' ’ 0.4 ' 0.6 0.8 1.0
BIT]

 

 

 

Figure 2-40: Resonant ﬁeld of transition at X- Band frequency for a quartet spin

system with D = 0.50 cm'1 and variable E/D.

131

2.4.6 Conclusions. Electron paramagnetic resonance (EPR) is employed to
explore the ground state spin Hamiltonian parameters of 3 and a model complex,
[Cr(tren)(DTBCat)](PF6), as well as several higher symmetry complexes of
chromium(III). Using the values obtained and previously reported spin
Hamiltonian values for other Chromium(III) systems, a systematic study of both
axial and rhombic examples of quartet systems was carried out to elucidate the
origin of the perpendicular mode EPR transitions. These were explored via energy
level diagrams and transition diagrams. Speciﬁcally, an approach was
implemented to build up the spectra in terms of increasing complexity of the zero
ﬁeld splitting spin Hamiltonian: ﬁrst a fully isotropic system was explored, then
the axial and rhombic zero ﬁeld splitting parameters were introduced. It is our
hope that this study will provide a simple reference tool for understanding and

assigning EPR parameters of quartet spin systems.

2.5 Final Remarks

The work presented in this chapter represents the foundation for the rest of the
work presented in this dissertation: not only in the sense of characterizing the
electronic structures of these complexes from the ground up, but also in the
piecewise approach that has been implemented to ultimately understand the
electronic structure and dynamics of spin-exchange complexes of chromium(III).
This strategy aims to understand physical properties in the absence of spin

exchange and thus quantify the effects of the introduction of spin exchange on

132

electronic structure and kinetics. Chapter 3 characterizes the nonradiative
dynamics of quartet Chromium(III) complexes, and Chapter 4 utilizes the wealth of
knowledge developed in Chapters 2 and 3 (as well as in the literature) to

understand the observed physical properties of spin-exchange complexes of

chromium(III).

133

2.6 References

(1-)
(2-)

(3-)

(4-)
(5-)

(6.)

(7-)

(8-)

(9-)

(10.)

(11.)

(12.)

(13.)

(14.)

Guo, D., McCusker, J. K. Inorganic. Chemistry 2007, 46, 3257-3274.

Wheeler, D. E.; Rodriguez, J. H.; McCusker, J. K. Journal Of Physical
ChemistryA 1999, 103, (20), 4101-4112.

F ehir, R. J .; McCusker, J. K. Journal Of Physical Chemistry A 2009, 113,
(32), 9249-9260.

Fehir, R. J ., Ph.D. Thesis, Michigan State University, East Lansing, 2009.

Brown, M. A.; El-Hadad, A. A.; McGarvey, B. R.; Sung, R. C. W.; Trikha,
A. K.; Tuck, D. G. Inorganica Chimica Acta 2000, 300-302, 613-621.

Mueller, J .; Kikuchi, A.; Bill, E.; Weyerrnueller, T.; Hildebrandt, P.; Ould-
Moussa, L.; Wieghardt, K. Inorganica Chimica Acta 2000, 297, 265-277.

Antcliff, K. L.; Baker, R. J.; Jones, C.; Murphey, D. M.; Rose, R. P.
Inorganic Chemistry 2005, 44, 2098-2105.

Baker, R. J.; Farley, R. D.; Jones, C.; Kloth, M.; Murphey, D. M. Jaurnal
of the Chemical Society, Dalton Transactions 2002, 3844-3 850.

Baker, R. J .; Farley, R. D.; Jones, C.; Mills, D. P.; Kloth, M.; Murphey, D.
M. Chemistry: A European Journal 2005, 11, 2972-2982.

Ozarowski, A.; McGarvey, B. R.; Peppe, C.; Tuck, D. G. Journal of the
American Chemical Society 1991, 113, 3288—3293.

Soedjak, H. S.; Cano, R. E.; Tran, L.; Bales, B. L.; Hajdu, J. Biochimica et
Biophysica Acta 1997, 1335, 305-314.

Weeks, C. L.; Levina, A.; Dillon, C. T.; Turner, R; Fenton, R. R.; Lay, P.
A. Inorganic Chemistry 2004, 43, 7844-7856.

Andriessen, W. T. M.; Meuldijk, J. Inorganic Chemistry 1976, 15, (9),
2044- 2047.

Meier-Callahan, A. E.; Bilio, A. J. D.; Simkhovich, L.; Mahammed, A.;

Goldberg, 1.; Gray, H. B.; Gross, Z. Inorganic Chemistry 2001, 40, (6788-
6793)

134

(15.)
(16.)

(17.)
(18.)

(19.)

(20.)

(21.)

(22.)

(23.)

(24.)

(25.)

(26.)

(27.)

(28.)

(29.)

Barrera, J. A.; Wilcox, D. E. Inorganic Chemistry 1992, 31, 1745-1752.

Andriessen, W. T. M. Journal of Physics and Chemistry of Solids 1976, 37,
189-196.

Riesen, H.; Dubicki, L. Journal of Physical Chemistry A 2008, ASAP.

Wang, D. M.; Pilbrow, J. R. Journal of Magnetic Resonance 1988, 77,
411-423.

Mabbs, F.; Collison, D., Electron Paramagnetic Resonance of d Transition
Metal Compounds. Elsevier: 1992.

Sander, W.; Grote, D.; Kossmann, S.; Neese, F. Journal of the American
Chemical Society 2008, 130, (13), 4396-4403.

Hans Henning Wenk, W. S. Angewandte Chemie International Edition
2002, 41, (15), 2742-2745.

Nakazawa, S.; Sato, K.; Shiomi, D.; Franco, M. L. T. M. B.; Lazana, M. C.
R. L. R.; Shohoji, M. C. B. L.; Itoh, K.; Takui, T. Inorganica Chimica
Acta 2008, 361, (14-15), 4031-4037.

Fleischhauer, P.; Gehring, S.; Saal, C.; Haase, W.; Tomkowicz, Z.;
Zanchini, C.; Gatteschi, D.; Davidov, D.; Barra, A. L. Journal of
Magnetism and Magnetic Materials 1996, 159, 166-174.

Benelli, C.; Dei, A.; Gatteschi, D.; Pardi, L. Inorganic Chemistry 1988,
27, 2831-2836.

Miller, J. S.; Calabrese, J. C.; Rommelmann, H.; Chittipeddi, S. R.; Zhang,
J. H.; Reiff, W. M. Journal of the American Chemical Society 1987, 109,
769.

Graebe, Justus Liebigs Annalen der Chemie 1868, 146, 31.

Weider, P. R.; Hegedus, L. S.; Asada, H.; D'Andreq, S. V. Journal of
Organometallic Chemistry 1985, 50, 4276.

Guo, D.; McCusker, J. K. Inorganic Chemistry 2007, 46, 3257-3274.

Wheeler, D. E.; McCusker, J. K. Inorganic Chemistry 1998, 37, 2296—2307.

135

(30.)

(31.)

(32.)

(33.)

(34.)

(35.)

(36.)

(37.)

(38.)

(39.)

(40.)

Schweiger, A.; Jeschke, G., Principles of Pulse Electron Paramagnetic
Resonance. Oxford University Press: New York, 2001; p 157.

XSophe Computer Simulation Software Suite, 1.1.4; Centre for Magnetic
Resonance and Department of Mathematics, The University of Queensland
(Brisbane, Queensland, Australia 4072) and Bruker Biospin GMBH
(formerly Bruker Analytik; Rheinstetten, Germany): 1993-2004.

Wang, D.; Hanson, G. Journal of Magnetic Resonance, Series A 1995, 117,
1-8.

Hanson, G. R.; Gates, K. E.; Noble, C. J.; Grifﬁn, M.; Mitchell, A.;
Benson, S. Journal of Inorganic Biochemistry 2004, 98, 903-916.

McClung, R. E. D. Canadian Journal of Physics 1968, 46, 2271.

Weltner, W., Magnetic Atoms and Molecules. Van Nostrand Reinhold:
New York, 1983; p 105.

Frisch, M. J .; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.;
Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A.; Stratmann, R. E.;
Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin, K. N.;
Strain, M. C.; Farkas, 0.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi, R.;
Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.;
Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.;
Rabuck, A. D.; Raghavachari, K.; F oresman, J. B.; Cioslowski, J .; Ortiz, J.
V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, 1.;
Gomperts, R.; Martin, R. L.; Fox, D. J .; Keith, T.; Al-Laham, M. A.; Peng,
C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.;
Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.; Head-Gordon, M.;
Replogle, E. S.; Pople, J. A. Gaussian 98, revision A.4; Gaussian, Inc.:
Pittsburgh, PA, 1998.

Gauss View 2.], Gaussian, Inc.: Pittsburgh, PA, 2000.

Wolfram Research, I., Mathematica. Version 5.2 ed.; Wolfram Research,
Inc. Champaign, Illinois, 2005.

Broze, M.; Luz, Z. The Journal of Physical Chemistry 1967, 71, (11), 3690-
3693.

Cano, J.; Ruiz, E.; Alvarez, S.; Verdaguer, M. Comments on Inorganic
Chemistry 1998, 20, (1), 27-56.

136

(41.)

(42.)

(43.)

(44.)

(45.)

(46.)

(47.)

(48.)

(49.)
(50.)

(51.)

(52.)

(53.)

(54.)

(55.)

Ruiz, E.; Cirera, J.; Alvarez, S. Coordination Chemistry Reviews 2005, 249,
2649-2660.

CRC Handbook of EPR Spectra of Quinones and Quinols. CRC Press:
Boca Raton, F1, 1985; p 382.

Drago, R. S., Physical Methods for Chemists. Second ed.; Surfside:
Gainesville, 1992; p 381.

Reitz, D. C.; Dravnieks, F.; Wertz, J. E. Journal of Chemical Physics 1960,
33, 1880-1881.

Das, M. R.; Fraenkel, G. R. Journal of Chemical Physics 1965, 42, (4),
1350-1360.

Brown, M. A.; McGarvey, B. R; Tuck, D. G. Journal of the Chemical
Society, Dalton Transactions 1998, 1371-1375.

Fehir Jr., R. J. PhD Thesis, Michigan State University, East Lansing, 2009.

Fehir, R. J .; Krzyaniak, M.; Schrauben, J. N.; McCracken, J. L.; McCusker,
J. K. Journal of Physical Chemistry B 2010, submitted.

O'Malley, P. J. Antioxidants & Redox Signaling 2001, 3, (5), 825-838.
Larson, S. C. Journal of Physical ChemistryA 2001, 105, 8333-8338.

Saladino, A. G; Larsen, S. C. Journal of Physical Chemistry A 2003, 107,
1872-1878.

Fehir, R. J .; McCusker, J. K. Journal of Physical Chemistry A 2009, 113,
9249-9260.

Ressouche, E.; Schwizer, J ., Molecular Magnets: Recent Highlights.
Springer: New York, 2003; p 119-137.

Barone, V., Recent Advances in Density Functional Methods. World
Scientiﬁc Publishing: Singapore, 1995.

Tuononen, H. M.; Armstrong, A. F. Inorganic Chemistry 2005, 44, 8277-
8284.

137

(56.)

(57.)

(58.)

(59.)

(60.)

(61.)

(62.)

(63.)

(64.)

(65.)
(66.)
(67.)

(653-)

(69.)

(70.)

Tuononen, H. M.; Chivers, T.; Armstrong, A.; Fedorchuk, C.; Boere, R. T.
Journal of Organometallic Chemistry 2007, 692, 2705-2715.

Drago, R. 8., Physical Methods for Chemists. second ed.; Surfside
Scientiﬁc Publishers: University of Florida, Gainesville, 1992; p 385.

Nguyen, M. T.; Creve, S.; Vanquickenbome, L. G. Journal of Physical
ChemistryA 1997, 101, 3174-3181.

Macfarlane, R. M. Journal of Chemical Physics 1967, 47, (6), 2066-2073.

Brickman, J.; Kothe, G. The Journal of Chemical Physics 1973, 59, (6),
2807-2814.

Lanciano, P.; Savoyant, A.; Grimaldi, S.; Magalon, A.; Guigliarelli, B.;
Bertrand, P. Journal of Physical Chemistry B 2007, ASAP.

Ikagawa, T.; Otsuka, T.; Kaizu, Y. Journal of Physical Chemistry A 1998,
102, 649-653.

Singer, L. S. Journal of Chemical Physics 1955, 23, 379-388.

Elbers, G.; Remme, S.; Lehmann, G. Inorganic Chemistry 1986, 25, (7),
896.

Jarrett, H. S. Journal of Chemical Physics 1957, 27, (6), 1298-1304.
McGarvey, B. R. The Journal of Chemical Physics 1964, 40, (3), 809-812.
Atanasov, M.; Schoenherr, T. Inorganic Chemistry 1990, 29, (22), 4545.

Fields, R. A.; Haindl, E.; Winscom, C. J .; Khan, Z. H.; Plato, M.; Mobius,
K. The Journal of Chemical Physics 1984, 80, (7), 3082-3086.

Andriessen, W. T. M.; Groenewege, M. P. Inorganic Chemistry 1976, 15,
(3), 621-626.

Schrauben, J. N.; Guo, D.; McCracken, J. L.; McCusker, J. K. Inorganica
Chimica Acta 2008, 361, 3539-3547.

138

Chapter 3: Electronic Structure and Ligand Field Dynamics of High-

Symmetry Complexes of Chromium(III)

3.1 Introduction

The work presented in this chapter employs transient absorption spectroscopy to
study a series of substituted complexes of the archetypal complex
tris(acetylacetonato)chromium(lII) in an effort to elucidate mechanisms of
nonradiative dynamics. As described in chapter 1, when a system absorbs a
photon, the ensuing processes that relax the molecule back to the ground state can
be classiﬁed as either radiative (ﬂuorescence, phosphorescence) or nonradiative
(vibrational cooling, internal conversion, nonradiative intersystem crossing,
quenching mechanisms). To achieve rational design of technologies such as solar
cell technology, molecular machines, and artiﬁcial photosynthesis,"2 one must be
able to predict and control the initial photophysical events after absorption of a
photon. The goal of nonradiative decay theory is the prediction of the rates of
state crossing events. in general, for inorganic systems there is a great deal of
understanding of the dynamics between the lowest energy excited state, which
according to Kasha’s rule is the emissive state, and the ground state. However, the
theory has never been conﬁrmed for intersystem between two excited states, and
furthermore no one has tested the applicability of nonradiative decay theory to
nonthermalized excited states. Moreover, there is no general conceptual

framework in place describing the ultrafast dynamics in transition metal systems.

139

A goal of the study of this series is to prove whether the current formalism of
nonradiative decay theory is applicable to transitions between nonthermalized,
rapidly evolving excited states. In the course of this study it is our hope to
establish the foundation of a framework which successfully describes the ultrafast
ligand ﬁeld dynamics in transition metal systems. To determine whether
nonradiative decay theory can describe decay dynamics between excited states, a
series of molecules with relatively simple ligand ﬁeld manifolds is required to
minimize the number of excited states that could potentially be participating in the
photophysics. A high photostability is also desirable to avoid photochemistry.
The molecules of study are a series of Chromium(III) tris-chelated derivatives of
mono-deprotonated 2,4-pentanedione (acac), present in Figure 3-1. Formally, the
symmetry of these molecules is D3, but the local environment around the
Chromium(III) center, 06, is effectively Oh. This series was designed with the
hope that with modifying the ligand the relative energies (AE) and displacements
(AQ) of the participating states will change enough—thus changing the coupling

integral—in order to observe signiﬁcant changes in km. Ligands with n-effects
(e.g. phenyl substituted acac) couple to tzg orbitals and will thus modify the energy
of the 2E state, according to the nephelauxetic effect,3 while o-modifying groups
(CF3 or CH3 substituted acac) will inductively modify the energy of the 4T2 state.

These systems also provide a conceptual stepping stone to understanding the

excited state dynamics of low-symmetry and Heisenberg spin exchange complexes

140

of chromium(III), discussed in Chapter 4 of this dissertation. In order to observe
the anticipated sub-100 femtosecond dynamics of the intersystem crossing in these

systems we implemented femtosecond transient absorption.4 The experiment is

described below.6

 

 

 

O—
//
Cr\ R R = H, Br, N02, SCN, phenyl, mesityl, methyl
\ /
O
_ _ 3
R. T
O___
A .
Cr R' = H, CH3, CF3, phenyl, mesutyl, cyclohexyl, t-butyl
\ /
\
O

 

 

 

 

 

Figure 3-1: Molecules of interest in this study.

For the remainder of this chapter the following abbreviations are employed
to represent the mono-anionic, chelating forms of the ligand: acac = of 2,4-
pentandione, 3-N02ac = 3-nitro-2,4-pentandione, 3-Brac = 3-bromo-2,4-
pentandione, hfac = 1,1,1,5,5,5-hexaﬂuoro-2,4-pentandione, dbm = 1,3-benzyl-

1,3-propandione, 3-methylac = 3-methyl-2,4-pentandione; 3-Phac = 3-phenyl-2,4-

141

pentandione; 3-SCNac = 3-thiocyanato-2,4-pentandione; tbutylac = 2,2,6,6-
tetramethyl-3,5-heptandione; dcm = 1,3-cyclohexyl-l,3-propanedione; prop = 1,3-

propanedione; 3-mesac = 3-mesityl-2,4-pentanedione.

3.2 Background: Ultrafast Transient Absorption Spectroscopy

The concept of the transient absorption experiment is fairly straightforward: an
excitation or “pump” pulse prepares an excited state in the sample, and a second
“probe” pulse follows which effectively takes a snapshot of the excited state
absorption. In reality the experiment is quite difﬁcult and requires a tremendous
amount of technology to successfully carry out on the picosecond and sub-
picosecond timescales, the timescale of molecular vibrational motion. While
many technologies are required (described brieﬂy in the experimental section in
the description of the instrument), the arrival of titanium sapphire lasers, whose
large bandwidth enabled creation of sub-picosecond pulses, was the ﬁrst crucial
step.5 After ampliﬁcation and manipulation of these pulses, the beam is split into
pump and probe beams. Eventually the pulses from both beam paths end up
overlapped at the sample, and the distance in beam paths gives the timing of the
experiment (Figure 3-2). In our case we vary the pump beam path while the probe
path stays constant (for stability reasons), so that as the pump beam path is
decreased the time between which the two pulses reach the sample increases. The
timing nomenclature of a transient absorption experiment is as follows: before the

pump pulse hits the sample, but after the probe pulse has, is considered “negative

142

 

 

 

 

Energy

 

 

 

 

 

 

 

 

J

 

Nuclear Coordinate

 

 

 

Figure 3-2: The transient absorption experiment employs a pump pulse, which
prepares an excited state population, and a probe pulse, which interrogates
dynamics within the excited state. Time resolution of the experiment is achieved
by varying the distance of the paths traversed between the pump and probe pulses.

The potential energy surfaces diagram is reproduced from reference 6.

143

time,” “time zero” occurs when the pump and probe pulses are completely
temporally overlapped within the sample, and “positive time” is when the pump
pulse proceeds the probe pulse. Data are referenced to the ground state absorption
so that changes in the extinction coefﬁcient between the ground and excited state

are what are effectively being measured, and the data are thus reported as AA

(Figure 3-3).

 

ESA .
As

e ESA — GSA
GSA >

 

 

Wavelength L1:

4- £ex< €gs

 

 

 

Wavelength

 

 

 

Figure 3—3: Cartoon showing ground state absorption (GSA) and excited state
absorption (ESA), plotted in units of extinction coefﬁcient (a, M'1 cm") (left
graph). The transient absorption experiment effectively measures the difference in
the extinction coefﬁcient of these two states (right graph), leading to the
possibility of both positive (“ESA”) and negative (“bleach”) features in the

transient spectrum.
As one becomes familiar with the transient absorption experiment, it

becomes clear that transient full spectra, i.e. full excited state absorption or bleach

features, provide information that is not easily gleaned from a series of single-

144

wavelength kinetic traces. Discussed below are various scenarios for spectral
evolution, which are put into context in terms of the kinetic processes that they
represent. A useful gedanken experiment is to consider one-dimensional potential
energy surfaces—and the characteristics of these surfaces, i.e. relative energies,
displacements and depths of the potential wells—and predict the evolution of the
full spectra based on these characteristics. (Whether this exercise is physically
meaningful depends on the number of normal modes which are actually involved
in facilitating nonradiative decay of the electronic state in question. Nevertheless,
this remains a useful tool for developing an understanding of the physical
processes, and the observation of these events, that are monitored in the transient
absorption experiment.) The possible observable full spectral dynamics depend
primarily on the relative displacement and slopes, that is whether one potential is
softer than the other, of the evolving electronic surface and the higher-lying
electronic state to which the probe pulse absorbs (cf. Figure 3-2). The examples
presented below are intended for the reader to gain an understanding of some of
the important electronic factors that can inﬂuence the observables of the transient
absorption experiment. Of course, beside these examples, various other
combinations of potential well shape, displacement, and energy separation can be
imagined. Considered ﬁrst are dynamics within a single electronic state which
does not evolve to form another electronic state, but rather thermalizes

(vibrationally cools) to form a metastable excited state.

145

In Figure 3-4 are shown two potential energy surfaces with no displacement
along the nuclear coordinate. These two wells are identical in depth and separate
energetically such that visible light induces transitions from the lower state to
upper state. Consider now that the lower state, S], is excited into a higher-lying
vibrational level, such that a transition from this state to the upper-lying state, 82,
will have maximum probability at energy hv(1). The probability of transition is
greatest at this energy because of optimum overlap of the constituent vibrational
wavefunctions of S1 and S2.

In the case of the transient absorption experiment hv(l) would correspond
to a monochromatic probe of S1, and because the transition probability is highest
would also correspond to the maximum intensity of the excited state absorption,
Xmamx. As the state thermalizes to form the lowest energy vibrational state, kmamx
will not change due to the similar slopes of the potential wells. However, the
excited state absorption is expected to narrow as the state thermalizes due to the

relatively small overlap of the v = 0 vibrational wavefunction of S1 with those of
the S2, limiting energetically the breadth of the transition. Therefore, a narrowing

single band without a shifting peak is indicative of vibrational relaxation within a
single electronic state. The narrowing of the excited state absorption will manifest

itself as a decay in the ESA at most probe wavelengths.

146

 

Energy

 

 

 

Nuclear Coordinate

 

 

 

Figure 3-4: Case of two identical electronic potential wells with small relative
displacement along the nuclear coordinate. Higher-lying Vibrational levels can
absorb readily at many wavelengths (hv(l)), but as S1 vibrationally cools the
number of available transitions is limited (hv(2)), and the spectrum of S2

accordingly narrows. Xmamx does not shift due to the identical slopes of S] and S2.

Again considering vibrational relaxation within a single electronic state,
one can envision scenarios in which Imam, will vary as S1 thermalizes. Figure 3-5
presents two scenarios in which S1 is vibrationally cooling but 31 and 32 have
different well depths, i.e. one potential is softer than the other. One can see that if

81 represents a softer potential than S2 a red shift of lemma is expected as S. cools

147

(Figure 3-5, upper panel), while a blue shift of kmamx occurs if the 82 potential is
softer (Figure 3-5, lower panel). In both cases the spectrum is expected to narrow
as S1 cools. In the context of the transient absorption experiment this represents
one case in which the dynamics are probe-wavelength dependent. If one were to
monitor 10 nm wide window of the full spectra, which is essentially what is
carried out when single-wavelength kinetics are collected on the instrument
presented in the experimental section of this chapter, one would expect decay in
the excited state absorption at one end (window 1 in Figure 3-6), a rise in the other
(window 3), and a combination of kinetics in the middle. Therefore, a shift in the
transient spectrum, concomitant with a narrowing of this spectrum, is also
indicative of decay within a single electronic state.

Electronic surface crossings carry different signatures in full-spectral
evolution. Figure 3-7 presents the example of a nonradiative intersystem crossing
process, a nonradiative transition between states of different spin, in this case a
singlet state and a triplet state. In this diagram the states are displaced along the
nuclear coordinate—but this is not necessary to the description of the dynamics in
this case but is rather shown this way to more easily distinguish between the
states. The singlet state IS. is prepared in an excited vibrational level. Laser
pulses probing this conﬁguration at energy hv(l) can induce transitions only into
the upper singlet state, lS2, due to the low transition probability between states of

different spin. As the state cools, similar dynamics to those described above are

148

 

Energy

 

 

 

Nuclear Coordinate

 

 

 

 

 

 

 

Nuclear Coordinate

 

 

 

Figure 3-5: Potential energy surface diagrams detailing expected shifts of Xmamx

for different slopes of S1 and S2.

149

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

V

Wavelength

 

 

 

Figure 3-6: Cartoon representing transient full spectra that are red-shifting
with time. Black represents the earliest time excited state absorption spectrum

while red represents that of the thermalized, metastable state.

observed, namely a narrowing of spectrum without any expected shift in 1mm“.
In the scenario of Figure 3-7, the thermalized lS1 has strong vibrational overlap
with a higher-lying vibrational level of the 3S3 state; therefore, as the 1S1 cools
population will be transferred nonradiatively to the triplet state 3S3. As this
population decays, the spectral features characteristic of 1S] will also decay and
features associated with the newly formed 3 S3 will arise. 3S3 only has signiﬁcant
transition probability to another triplet state, 3S4, at hv(2). In this scenario a shift

in 1mm“ is expected. Figure 3-8 represents this case: one band decays while

150

another grows, clearly indicative of a surface crossing to a new electronic state.
Again, if one were to monitor probe windows, at one probe window a decay in the
excited state absorption is observed, and at another a rise is observed, exactly the
same as the case described above for relaxation within a single electronic state.
Therefore, single-wavelength kinetics similar to those of Figure 3-6 would be
observed, where decay in one electronic state results in a shift of kmamx. Here,
however, the black transition is decaying concomitant with the growth of the red
transition, so that in window 1 (far left) a monoexponential decay is observed
while on the red edge (window 3) and rise is observed. The fact that the single
wavelength kinetics are the same for the two disparate processes of Figures 3-6
and 3-8 emphasizes the need for well-resolved full spectra, the only means of

distinguishing between these two cases.

151

 

 

 

 

hv(2) —

 

 

J I 384

 

Energy

 

1S1

 

 

. 3s3

 

 

 

Nuclear Coordinate

 

 

 

Figure 3-7: Generalized picture revealing how an intersystem crossing event can

affect the observed transient spectra.

152

 

 

 

AA

 

 

 

 

 

 

 

 

 

 

 

 

H

Wavelength

 

 

 

Figure 3-8: Transient full spectra of a surface crossing event, with characteristic

spectra from the initial and ﬁnal electronic states.

3.3 Experimental Section

3.3.1 Physical methods. UV-vis absorption was carried out on a Cary-50.
Spectrographic grade dichloromethane or acetonitrile (Jade) was used for all
measurements. Emission and excitation spectra were collected on a Spex
Fluoromax Spectroﬂuorometer. All samples were prepared in an inert
environment, and such that the absorbance at the excitation wavelength was
between 0.1 - 0.2. All emission spectra reported in this dissertation were collected
at low temperature (80 K) using a Janis SVT-IOO optical dewar. The emission

spectra of tris(acetylacetonato)chromium(III) (Cr(acac)3) and related complexes

153

was collected in an 80 K optical glass of 2-methyltetrahydrofuran. These
complexes emit near 800 nm, and a 650 nm long-pass ﬁlter was employed. For all
measurements the emission slits were set at 0.5 mm, and the excitation slits were
set between 0.5 - 1 mm. It should be noted that the response of the
photomultiplier tube which is employed for detection in these measurements is
quite poor for wavelengths longer than ~ 810 nm. The spectra were corrected for
lamp intensity, and spectra plotted in units of energy (cm'l) were further corrected
as per Parker and Rees.6

All time-resolved spectroscopic measurements were carried out in
dichloromethane, unless the samples were shown to photodegrade when exposed
to ultrashort pulses. In this case, the sample was ﬂowed and dichloromethane
could not be used because it would dissolve the tubing that was employed in the
experiment; acetonitrile was used instead. Ultrafast transient absorption data with
~ 100 fs pulse duration were collected as follows. A Tizsapphire oscillator
(Coherent Mira) is pumped at 5.4 W (Coherent Verdi), producing modelocked
pulses centered at 803 nm with a FWHM of 13.5 nm at an repetition rate of about
76 MHz. These pulses are fed into a cavity containing a stretcher/compressor and
regenerative ampliﬁer (Positive Light Spitfire), pumped by a Coherent Evolution.
The output of the regenerative ampliﬁer is again 800 nm, ~100 fs pulses at
kilohertz repetition rate. Approximately 70% is fed into an OPA (TOPAS) for

generating the appropriate pump wavelength, while the remaining 30% is used to

154

generate the white light continuum (CaF2 or sapphire) used for the probe pulse.
The pump traverses a delay line (Aerotech ATSlOO-ZOO with a Unidex 100 motion
controller), providing the timing for the experiment. The full white light
continuum can be used as a probe for full spectral transient studies, or a 10 nm
bandpass ﬁlter can be employed to select a probe wavelength for a “single-

wavelength” kinetic trace.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8
:5 <
> D:
E
TOPAS {:18 C 8
REGEN 10mm “”9“" K n.
' c L <1 1 1
.g ’ Delay Line
2 AC. \
‘ 0
Lil Lon Time Dela 4 y?
r g? y . 7D \ ’/ﬁ

 

 

 

 

 

 

 

 

Figure 3-9: Ultrafast transient absorption setup which can achieve ~ 100 fs pulses

in the visible.

In the visible, pump powers of 6-7 uJ/pulse are used, while exciting in the
ultraviolet requires smaller energies, approximately 1-3 uJ/pulse. Samples were
prepared in 1 mm pathlength cells so that the absorbance is 0.4 - 0.7 at the

excitation wavelength. For certain complexes a ﬂowing sample was used to avoid

155

photodecomposition of the sample. A lock-in ampliﬁer and chopper are employed
to achieve reasonable SN, and detection is carried out with two avalanche
photodiodes. Labview programs of local origin were used to process all data.7 In
general, ﬁtting of single wavelength kinetic traces with t > 500 fs was carried out
with an exponential function with an additional y-offset. Fitting was started at 300
fs, past the solvent cross-correlation signal. Although not used in this dissertation,
faster dynamics can be modeled via convolution of the kinetics and the instrument
response function, given by the solvent cross-correlation signal.

Transient full spectra were collected using a ﬁber-optic cable connected to
a Hamarnatsu HC233-0900 spectrograph. The spectrograph employs a diode array
(C5964 NMOS), upon which the collected photons are imaged after dispersion via
a grating. The data was collected in the absence of lock-in detection. Scans of
background pump scatter, called dark scans, are collected before the experiment is
carried out. Background scans, corresponding to negative times, are collected at
the beginning of each scan, and the dark and background scans are both subtracted
to calculate the ﬁnal transient spectra. Sub-ps full spectra must be corrected for the

chirp associated with the probe pulse. This is described in detail elsewhere.7
Coherence data on Cr(acac)3 were collected on an instrument housed by the

Beck group at Michigan State University. The sample was a ﬂowing solution of
acetonitrile with an absorbance of ~ 0.7. Femtosecond pump-probe transients with

impulsive excitation were recorded using the dynamic-absorption technique, in

156

which the probe beam is dispersed in a grating monochromator after passing
through the sample.8'IO The pump and probe pulses were obtained from the
signal-beam output of an optical parametric ampliﬁer (Coherent OPA 9450),
which was pumped by an ampliﬁed Tizsapphire laser (Coherent Mira-seed
oscillator and a modiﬁed Coherent RegA 9050 regenerative ampliﬁer, with
Coherent Verdi V5 and V10 pump lasers, respectively). The laser was operated at
a repetition rate of 250 kHz. The pump and probe pulses were corrected for
group-delay dispersion on the way to the sample by a SFlO Brewster prism-pair
pulse compressor. The pump-probe time delay was scanned using a rapid-scanning
delay stage (Clark-MXR, ODL-lSO) in a modiﬁed Mach-Zehnder interferometer
with confocal sample and autocorrelation-crystal positions. Calcite polarizers and
wave plates in the pump and probe beams set their planes of polarization at 90°;
after passing through the sample, the probe beam was analyzed by another calcite
polarizer oriented 90° relative to the pump-beam's plane of polarization, and then
it was passed through a monochromator (Spex 270M, 4-nm bandpass) and
detected by an ampliﬁed photodiode (Thorlabs PDA55). The slits of the
monochromater were adjusted to obtain a fairly narrow bandpass (4 nm) compared
to the width of the laser's spectrum. This approach is similar to that used by
Champion and co-workers in their studies of low-frequency vibrational coherence

. . 11,12
1n heme proteins.

The pump-probe signal was obtained from the photodiode
signal using a lock-in ampliﬁer (Femto LIA-MV-ZOO-H); the pump beam was

modulated at 50 kHz by a photoelastic modulator (Hinds Instrumentation). Fitting

157

of the oscillatory component (the residual after subtraction of the exponential
components) was accomplished with a program of local origin, and has been
described in detail elsewhere.13 Brieﬂy, the oscillatory residuals were ﬁt with a
multicomponent model that contains both slowly damped and rapidly damped
components. The slowly damped components were modeled as a simple damped
cosinusoid, while the rapidly damped components were modeled in the time
domain with inhomogeneously broadened components with asymmetric Gaussian

lineshapes.

3.3.2 Computational Methods. All calculations were carried out in Gaussian
03.14 Crystal structure geometries were used as a starting point for geometry
optimization calculations using the B3LYP functional and LANLZDZ basis set to
generate a minimized geometry. This geometry was then subjected to a further
geometry optimization at the unrestricted B3LYP/6-311g** level, employing a
dielectric continuum solvent model for dichloromethane (a = 8.93) or acetonitrile
(e = 35.688). The optimized geometry was used for time-dependent calculations
also employing a dichloromethane dielectric continuum using the unrestricted
B3LYP functional and 6-311g** basis set. Images of molecular orbitals and

. . . . 15
Vibrational modes were prepared 1n GaussVrew.

3.3.3. Synthesis 1. 3-Substituted Complexes of 2,4—peutanedione

158

i. tris(2,4-pentanediono)chromium(III), Cr(acac)3. Cr(acac)3 was prepared by
adapting a previous literature method.16 In 15 ml of water was added 2 g (7.51
mmol) of Chromium(III) chloride hexahydrate, 4.5 g (45 mmol) of acetylacetone
(2,4-pentanedione) and 4.5 g (45 mmol) of triethylamine. The solution was stirred
with gentle heating and allowed to react overnight. Purple crystals fall out of
solution which can be ﬁltered, dried, and recrystallized by dissolving in benzene
and slowly adding three volumes of petroleum ether. Yield: 0.57 g (21 %). Anal.
Calcd for C15H2106Cr expected: C, 51.57; H, 6.06; N, 0.00. Found: C, 51.65; H,

5.94; N, 0.02.

ii. tris(3-phenyl-2,4-pentanedionato)chromium(IIl), Cr(3-phenylacac)3. This
preparation was modiﬁed from a previously reported procedure.'7 0.6 g (10
mmol) of urea was stirred into 1 ml of water and 3 ml of dioxane. 0.176 g (1
mmol) of 3-phenyl-2,4-pentanedione (TCI) was added, then 0.089 g (0.33 mmol)
of Chromium(III) chloride hexahydrate was added. The reaction was reﬂuxed and
stirred for three days. Upon cooling, the product was washed on a ﬁne fritted
ﬁlter with copious amounts of water to remove any ammonium salts and dried in
vacuo. Anal. Calc. for C33H33O6Cr expected: C, 68.62; H, 5.76; N, 0.00. Found:
C, 72.60; H, 6.20; N, 0.10. This complex has also previously been prepared using

a method developed by Girolami and coworkers.“19

159

iii. tris(3-methyl-2,4-pentanedionato)chromium(III), Cr(3-methylacac)3. 0.6 g
(10 mmol) of urea was stirred into 2 ml of water and 2 m1 of dioxane. 0.114 g (1
mmol, 0.116 ml) of 3-methyl-2,4-pentanedione (Aldrich) was added, then 0.089 g
(0.33 mmol) of Chromium(III) chloride hexahydrate was added. The reaction was
reﬂuxed and stirred for three days. Upon cooling, the product was washed on a
ﬁne fritted ﬁlter with copious amounts of water to remove any ammonium salts
and dried in vacuo. The product was recrystallized by dissolving the solid in a 5:1
methanol/water solution, and allowing the methanol to slowly evaporate. This
yielded brown-green needle-like crystals. Anal. Calc. for C13H27O6Cr expected:

C, 55.24; H, 6.95; N, 0.00. Found: C, 55.26; H, 7.10; N, -0.02.

iv. tris(3-nitro-2,4-pentandiono)chromium(III), Cr(3-N02acac)3. This
complex was prepared according to literature methods.” A slurry of ﬁnely
divided copper(II) nitrate 2.5-hydrate (2.15 g) in 40 ml of acetic anhydride was
stirred for 15 minutes at 0 °C. 1 g (2.87 mmol) of Cr(acac)3 was added and
stirring maintained at 0 0C for two hours. After two hours the solution was
removed from the ice bath and allowed to stir at room temperature for one hour.
The slurry was decomposed by adding 120 ml of water, 120 ml of ice, and 3 g of
sodium acetate. The red precipitate was collected, washed with ethanol and water,

and air dried. The product is recrystallized from ethanol/chloroform. Yield: 0.57

160

g (21 %). Anal. Calc. for C15H13N3012Cr expected: C, 37.29; H, 3.75; N, 8.68.

Found: C, 37.26; H, 3.24; N, 8.84.

v. tris(3-bromo-2,4—pentandiono)chromium(III), Cr(3-Bracac)3. Cr(Br-acac)3
was prepared according to literature methods.21 A stirring solution of 0.6 g (1.7
mmol) of Cr(acac)3, 0.51 g sodium acetate, and 15 ml of glacial acetic acid was
prepared and cooled to 30 °C. To it was added a solution of 1.21 g (7.6 mmol) of
Br2 in 2 ml of glacial acetic acid. After 5 minutes the precipitates were collected

and washed with water, aqueous sodium bicarbonate, aqueous sodium sulfate, and
ﬁnally water. The brown substance was recrystallized twice from

benzene/heptane. Yield: 0.751 g (0.13 mmol, 74.6 %). Anal. Calc. for
C15ngO6Br3Cr expected: C, 30.74; H, 3.10; N, 0.00. Found: C, 31.03; H, 3.15;

N, 0.04.

vi. tris(3-thiocyano-2,4-pentandiono)chromium(III), Cr(3-SCNacac)3. The
procedure22’23 involves preparation of three different solutions, all of which must
be cooled in a dry ice/ acetone bath prior to mixing: 1) a vigorously stirring
suspension of potassium thiocyanate (10 g, 100 mmol) in absolute methanol; 2) 8
g bromine (200 mmol, 2.66 ml) in absolute methanol; 3) a solution of Cr(acac)3

(5.8 g, 16.5 mmol) in absolute methanol. The bromine solution was added slowly

to KSCN/methanol suspension so that the temperature does not go above -60 C.

161

Then the Cr(acac)3 is added all at once. The mixture is stirred for one hour in the

dry ice/acetone bath, then stirring is continued and the solution is allowed to reach
room temperature. The room temperature solution is poured onto crushed ice and
the precipitate is collected after one hour. The purple solid is recrystallized from
benzene/petroleum ether. Yield 3.79 g (7.28 mmol, 44.1%). Anal. Calc. for
C18H1306N3S3Cr expected: C, 41.53; H, 3.49; N, 8.07. Found: C, 42.27; H, 3.38;
N, 6.65. An infrared spectrum of this compound shows the characteristic 4.64 pm
SCN stretch, a single absorption at 6.40 um in the 6.25-6.75 region and a lack of

peak at 8.25 pm, associated with unsubstituted Cr(acac)3.

vii. 3-mesityl-2,4-pentanedione. This ligand was prepared according to
previously published procedures, with slight modiﬁcations. 24 Activated copper
powder (64 g) was prepared in a 1 L roundbottom ﬂask by suspending copper
powder in glacial acetic acid for 15 minutes, decanting the acid and replacing it.
After 15 minutes more, the acid was decanted and reagent grade acetone was used
to displace the acid. These washings continued until the acetone was colorless,

then the copper powder was dried in vacuo. After the powder is dry, 5 g of Cr(3-
Bracac)3, prepared as described above, and 10.5 g of iodomesitylene (Alfa Aesar)
were stirred into the copper powder. The mixture was heated under positive
pressure of Ar for 4 hours at 240 0C. After cooling, the product was broken up

and 430 ml of 50% HZSO4 solution was added and the mixture heated for 3 hours.

162

After cooling, the mixture was ﬁltered, and the ﬁltrate washed with water. The
dry ﬁltrate was extracted with benzene in a Soxhelet extractor for three hours.
Meanwhile, the acidic mother liquor is extracted with benzene, and these organic
extracts are washed three times with water and dried over NazSO4. The extracts
were combined, the solvent removed, and the residue separated with benzene on a
silica column. The second product to come off the column (rf ~ 0.5) is the desired
product. The solvent is removed, and a crude solid is obtained by dissolving the
oily product in methanol, adding a few drops of water, and letting the methanol
slowly evaporate. This brown solid was sublimed under high vacuum at ~ 80 0C
to yield a yellow crystalline solid, the desired product. Yield: 120 mg, (2.1 %).
‘H NMR (300 MHz, CDCl3); 1.739 (s, 6 H, -CH3), 2.075 p (s, 6 H, phenyl-CH3),

2.282 (s, 3H, phenyl-CH3), 6.910 (s, 2 H, aromatic), 16.597 (1 H, -OH).

viii. tris(3-mesityl-2,4-pentanedionato)chromium(III), Cr(3-mesac)3: This
complex, which is purple in color, was achieved according to the method
established by Girolami and coworkers,18’19 (ESI +, m/z = 704.4 MH+), but also
contained a signiﬁcant amount of an unknown green solid (ESI+, m/z = 1003.5).
A cleaner preparation was carried out: 65 mg (0.298 mmol) of 3-mesityl-2,4-
pentanedione was added to a stirring solution of 26.5 mg of Chromium(III)

chloride hexahydrate (9.93 x 10'5 mol) and 0.6 g urea (IO-fold excess). Enough

methanol was added to the solution to dissolve the ligand (approximately 12 ml)

163

and the solution was reﬂuxed for 24 hours. A purple precipitate forms during the
course of the reaction. After removing the reaction from heat, a stream of nitrogen
was blown over the solution and more purple precipitate forms along with the
unreacted ligand. The cooled solution is ﬁltered and washed with large amounts
of water and methanol. MS (E81 +), m/z = 704.5 (MH+, 100%). Anal. Calcd for
C42H51CrO6 expected: C, 71.67; H, 7.30; N, 0.00. Found: C, 71.33; H, 7.23; N,

0.00.

3.3.4 Synthesis II. Substituted Complexes of 1,3-Propanedione.

i. tris(propandiono)chromium(III). This preparation is modiﬁed from a
literature procedure.25 A mixture of 5.084 g (0.031 mol) of tetramethoxypropane,
5 ml water and 2 m1 of l M HCl were heated in a steam bath until the phases
mixed and the solution became lemon yellow. Overheating to produce a darker
color will decrease the yield. The solution was cooled to 0 °C in an ice bath then
2.52 g triethylamine was added, which was dried over 4 A molecular sieves and
distilled from P205. The solution rapidly turns a red-violet color. This solution
was added to a 0 °C (again cooled in an ice bath) suspension of 1.30 g anhydrous
CrCl3 ( mol), 1.635 g Zn dust and 100 ml of ether. After stirring for 30 minutes,
the ether layer becomes intensely violet; it is decanted, replaced, and this

procedure is repeated until the reaction is exhausted. The solvent is removed and

the red residue is dissolved into ~ 25 ml of hot benzene. The solution is allowed

164

to cool and then poured onto a column of neutral alumina packed with benzene.
The red/violet band is eluted with dichloromethane, leaving an orange band
remaining on the column. Recrystallization can be carried out by dissolving in
isopropanol, adding three times the volume of hot heptane and placing in a freezer.
Crystals can be easily grown by slow evaporation of an ether solution. Anal.

Calcd for C9H9CrO6: C, 40.77; H, 3.42; N, 0.00. Found C, 41.15; H, 3.36; N, -

0.02. MS (E81 +), m/z = 266.0 (MH+).

ii. 1,3-dicyclohexyl-l,3-propanedione. This synthesis is modiﬁed ﬁom literature
procedures.26 A 500 ml two neck round bottom ﬂask was equipped with a reﬂux
condenser and a rubber septum. 2.74 ml (20 mmol) of cyclohexyl methyl ketone
(Alfa Aesar) was put into solution with 200 ml of dry diethyl ether and bubble
degassed with nitrogen for 1/2 hour while stirring. 20 mmol of 2 M LDA (10 ml)
was then added dropwise over 10 minutes, and the orange-brown solution was
allowed to stir for 10 minutes. Meanwhile, 2.86 ml (20 mmol) of methyl
cyclohexanecarboxylate (Aldrich) was bubble degassed in 100 ml of dry ether and
added to the solution via cannula transfer. After reﬂuxing for two hours, the
solvent was removed and the product was put in water. The product was
neutralized with dilute HCl, extracted with ether, and dried over sodium sulfate.
This yields the crude product. To purify, put the product in a methanol solution: a

white precipitate will form which was shown by NMR to be an aliphatic side

165

product. The precipitate was ﬁltered off, and to the yellow methanol solution was
added a very hot concentrated solution of copper(II) acetate in water. While
cooling, a dark precipitate will form which is bis(1,3-d.icyclohexyl-1,3-
propanedionato)copper(II). Filter and wash with water and a large amount of
methanol. (Anal. Calcd for C30H46Cu04: C 67.45, H 8.68; found C 67.30, H 8.86).
To liberate the ligand, dissolve the copper complex in ether and add about 50 ml
of 10% sulfuric acid (aqueous) and shake until the green color of the copper
complex is replaced by the light yellow of the ligand. Extract the ligand with
ether, dry over sodium sulfate. The ligand can be recrystallized by dissolving in
methanol and adding a small amount of water. Yield: 0.99g (21.2%). Anal. Calcd

for C15H24Ozz C 76.23, H 10.24; Found C 76.41 H 10.46.

iii. tris(l,3-dibenzoylmethanato)chromium(III), Cr(dbm)3. The synthesis was
carried out via a previously published method.27 Brieﬂy, Cr(acac)3 and a slight
excess of 1,3-dibenzoylmethane were ground together, then heated at 175 °C for 4
hrs under a slight stream of N2. The cooled product was broken up, heated with

methanol to extract the starting materials, cooled, then ﬁltered. After washing
with acetone the product was recrystallized twice from benzene/cyclohexane.

Anal. Calcd for C45H33CrO6: C 74.90, H 4.58; Found C 74.16, H 4.18.

166

iv. tris(1,3-dibenzoylmethanato)gallium(III), Ga(dbm)3. 1 mmol (0.256 g) of
Ga(NO3)3 hydrate was dissolved in 20 ml methanol, and to it was added dropwise
3 mmol (0.672 g) of dibenzoylmethane in methanol over 15 minutes. A slurry of
sodium bicarbonate (0.5 g in 20 ml) in methanol was added dropwise and the
reaction was allowed to stir overnight, over which time the product precipitates.
The product was ﬁltered and washed with water. To recystallize, dissolve in
dichloromethane and slowly add methanol. Yield: 0.35 g, 47%. Elemental
analysis: Calculated for C45H33Ga06: C 73.09, H 4.50; Found C 71.69, H 4.50.
Note: the unreacted ligand can be recovered by adding water to the mother liquor

to crash it out.

v. tris(1,3-dicyclohexyl-1,3-propandionato)chromium(III), Cr(dcm)3. The
synthesis can be easily carried out using previously published methods for

8,19

tris( 1 ,1 ,1 ,5 ,5 ,5-hexaﬂuoroacetylacetonato)chromium(III).l Brieﬂy, under inert

conditions, CrClz (0.1 g, 8.46 x 10'5 mol) was added to a solution of two times
excess lithium bis(trimethylsilyl)amine (0.28 g) in 5 ml THF. After 15 minutes of
stirring the solution should turn from colorless to purple, and ﬁnally to blue after
about one hour. The reaction mixture was ﬁltered to remove LiCl that had
precipitated, and the mother liquor was cooled with an acetone/dry ice bath. Once

cooled, a solution of 1,3-dicyclohexyl-1,3-propandione (60 mg, 2.54 x 10'4 mol)

dissolved in a minimum amount of THF was added dropwise to the cold solution.

167

The color changes from blue to yellow. The solution is removed from the drybox
and allowed to stir while exposed to the atmosphere, and the color changes from
yellow to green. The solvent was removed, and the crude product washed with
methanol and ether, and recrystallized via slow evaporation of 3:1
CHzClz/methanol solution. Yield. 19.7 mg (31%). Anal. Calcd for C45H69CrO6 : C

71.30, H 9.19, Found: C 70.66, H 9.52

vi. tris(1,3-dicyclohexylmethanato)gallium(III), Ga(dcm)3. 0.114 g (0.564
mmol) of Ga(NO3)3 hydrate was dissolved in 10 ml methanol, and to it was added
dropwise to 0.4 g (1.69 mmol) of 1,3-dicyclohexyl-l,3-propanedione in methanol
over 15 minutes. A slurry of sodium bicarbonate (0.5 g (excess) in 20 ml) was
added dropwise and the reaction was allowed to stir overnight, over which time
the product precipitates. The product was ﬁltered, and washed with copious
amounts of methanol and water. To recystallize, dissolve in dichloromethane and
slowly add methanol. Yield: 0.24 g (55%). Anal. Calcd for C45H69GaO6: C
69.67, H 8.97; Found C 69.58, H 9.22. Note: the unreacted ligand can be

recovered by adding water to the mother liquor to crash it out.

vii. tris(2,2,6,6-tetramethyl-3,5-heptanedionato)chromium(III), Cr(tbuylac)3.

To a solution of 2 ml water and 2 ml dioxane was added 0.6 g urea (excess), 0.184

g (1 mmol, 0.208 ml) of 2,2,6,6-tetramethyl-3,5-heptanedione, and 0.089 g (0.33

168

mmol) Chromium(III) chloride hexahydrate. The reaction was reﬂuxed for 3 days
and the product washed with water and recrystallized from benzene/petroleum

ether. Anal. Calcd for C33H57CrO6: C 65.86, H 9.55; Found C 65.85, H 9.29.

viii. tris(l,1,1,5,5,5-hexafluoro-2,4-pentandiono)chromium(III). This method
was adapted from the synthesis of tris(1,1,1-triﬂuoro-2,4-
pentandionato)chromium(lII).28 0.535 g (2 mmol) of chromium(III) chloride
hexahydrate is dissolved in about 10 ml of water, to which 3.5 ml of 2 M NH3 in
ethanol is added slowly with shaking. The solution is allowed to stand for 10
minutes, after which the turquoise product, Cr(OH)3, is gravity ﬁltered and washed
with water. The moist Cr(OH)3 is added to a solution of 1 m1 l,1,1,5,5,5-
hexaﬂuoro-2,4-pentandione (excess) in 20 ml of dimethoxypropane. 0.34 ml of
glacial acetic acid is added, and the solution should feel cool while it turns purple
as the Cr(OH)3 dissolves. The solution is stoppered and allowed to react overnight.
The product is ﬁltered, air dried, and recrystallized by heating the product in
ethanol, ﬁltering the solution, and adding hexanes to the ﬁltrate. Yield 0.164 g.
The product in stored in an inert atmosphere in the dark, as prolonged exposure
results in decomposition of the product. This complex can also be made from a

literature preparation utilizing a Cr(II) precursor.18

169

3.4 Results and Discussion

3.4.1 Cr(acac)3. Although the absorption and emission spectra of this complex
have been examined previously, this section will explore the excited state
electronic structure of Cr(acac), in great detail, as the results and concepts

developed here will be extremely useful for analysis and assignment of transitions
for the other members of this series. The absorption spectrum, shown in Figure 3-

10, consists of several transitions in the ultraviolet, centered at 250 nm (8 ~ 11100
M'1 cm'l), 270 nm (8 ~ 10000 M'1 cm") and a more intense band at 330 nm (8 ~
15000 M" cm'l). Another transition of much lower intensity, centered at 374 nm
with 3 ~ 430 M 1 cm'l, is observed on the shoulder of the intense band at 330 nm.
In the visible one observes a very broad band of low intensity centered at 560 nm
with 8 ~ 65 M‘1 cm']. The emission spectrum is quite narrow with maximum
intensity at 775 nm (12907 em").

Starting in the UV, the transitions at 250 nm and 270 nm are assigned as an
intraligand 11: —> rt‘l transition based on the absorption spectrum of Ga(acac)3. In
this complex ligand ﬁeld and charge transfer transitions are precluded due to the
closed shell nature of the Ga(III) ion, leaving only the possibility of intraligand
transitions. The transition, centered at 280 nm, is quite intense with 8 ~ 31000 M'1
cm'l. Given this assignment, the other intense transition in the UV, at 330 nm, is
assigned as a charge-transfer transition based on its intensity. Given the relatively

high oxidation potential of the Chromium(III) ion this transition is likely a ligand-

170

to-metal charge transfer (LMCT) transition, whose orbital parentage is depicted in

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure3-11.
3-5 '1 l l ' l ' l ' F f T 12
. —Cr(acac)3 r
3-0 ‘ n e x 50, Cr(acac):3 _1,0
4" 2.5_ a, Ga(acac)3 2‘
o , —Emission -O.8 g
x 2.04 h .03
‘— - -0 6 E
6 _ ' a
‘7 .5
E P0.4 79
.. E
' o
—0.2 2
300 400 500 600 700 800
Wavelength (nm)

 

 

 

Figure 3-10: Absorption and emission spectra for Cr(acac)3, acquired in

dichloromethane and a low-temperature (80 K) optical glass of 2-

methyltetrahydrofuran, respectively.

In this drawing, the it orbitals shown on the right are ligand (acac') based. These
assignments have been conﬁrmed in the literature. Barnum29 assigns the band

near 330 nm to a CT type transition despite previous assignment of a It —> 1t*

171

transition due to the high intensity. All transitions higher in energy he assigns to it
—+ if“ transitions. This corresponds well to the transitions of Ga(acac)3, which can
have no CT type transitions due to the redox-inert nature of the gallium(III) ion,
and show no strong absorptions near 330 nm. Fleming and Thornton3O agree with
this assignment of the ultraviolet transitions in Cr(acac)3. The transitions of
Cr(acac)3 near 345 nm have been assigned in the literature as both metal to ligand
charge transfer“ and ligand to metal charge transfer. However, as stated above,

the stability of Chromium(III) to oxidation makes a LMCT far more likely (Figure

 

 

311).
_na: _nt
eg* __ eg" __
9
’l ’l ’l ’l ’l ’l
t29—__ 11 t29_—L— 1
4A2 —n 4LMCT —n

 

 

 

Figure 3-11: Orbital parentage of the 4LMCT transition.

Cr(acac)3 shows a shoulder on this charge-transfer band around 380 nm
with intensities 7-8 times that of the 4A2 —- 4T2. Fleming and Thornton assign this
as 4A2 —> 4T], but Barnum hesitated to assign the shoulder to a ligand ﬁeld
transition due to the relatively large intensity. This band has even been assigned
as a charge-transfer type transition.3| Hanazaki and coworkers have ascribed the

higher extinction coefﬁcient of this transition as due to coupling between the

172

ligand ﬁeld and ligand based electrons.32 This coupling will now be described in
detail. Many complexes of aromatic ligands bound to Chromium(III) have
moderate-strength absorptions (300-2000 M'lcm'l) which can not easily be
assigned as ligand ﬁeld (generally 8 < 100 M'] cm'l), charge transfer, or ligand
based (> 103 — 104 M'1 cm'l). These transitions have been observed in

33-35 . 35
phenanthrolme, as well as

Chromium(III) complexes of bipyridine,
substituted acetylacetonate complexes of chromium(III). In the late 1960’s
Nagakura and coworkers suggested that these transitions were to an excited state
constructed from both triplet intraligand (31L) character and ligand-ﬁeld character,
arising from direct coupling between the unpaired electrons on the ligand and

36-38
metal center.

Twenty years later Hanazaki and coworkers re-examined this
problem,32 with the results presented below.

To estimate the relative probability of a so-called 4(31L) transition (i.e. a

total spin-allowed transition as a result of coupling between triplet ligand and
metal-based spins) the authors start by constructing the relevant zero-order
wavefunctions, which include both metal and ligand electrons. As an example, the

ground state and one possible charge-transfer state can be expressed as comprised

of both d-orbital and ligand-based orbitals. Using this formalism the states 4('IL)
and 4(31L) can be constructed, corresponding to states with total quartet spin which
include the ligand ﬁontier orbitals. While the 4('IL) state simply corresponds to a

ligand localized 1: —> if“ transition, the 4(31L) state arises from coupling with the

173

ligand triplet states, and transition from the ground state to this state is forbidden
due to the triplet character of the ligand. However, these states can become
optically allowed if 4CT character is mixed in. Using this intensity-borrowing
mechanism, the authors were able to develop an equation which gives the ratio of
the oscillator strengths of the easily-assignable 4('IL) to the more ambiguous
4(3IL), using only the energies of the 4(11L), 4(31L), and charge transfer states. The
equation they developed results in a ratio of 0.04 for Cr(bpy)3, in reasonable
agreement with many of the moderate-strength absorptions observed in both
Cr(acac)3 and Cr(bpy)3-type complexes. Therefore, this moderate—strength
absorption has been assigned as a transition to a quartet state that possesses quartet
charge-transfer and triplet intraligand character, 4(3IL), the orbital parentage of

which is depicted in Figure 3-12.

 

 

_7:* L117"
eg“ __ ’ 69* __
t ’1’] ’l t ’1’] L
29_—_ ”ll, 29_—— ’1
4A2 —1t 4(3IL) --1t

 

 

 

Figure 3-12: Orbital parentage of the 4(31L) transition.

In terms of purely ligand ﬁeld transitions, the weak absorption whose maximum

is at 560 nm is assi ed as the 4A2 —+ 4T2 transition, the lowest-ener s in-
gn gy P

allowed transition for a d3 ion. This transition is well separated from the others

174

discussed above, allowing for selective excitation into the 4T2 electronic state
using a broadband ultrafast laser pulse. The next highest spin allowed ligand ﬁeld
state in energy is the 4T1, which is likely buried under the 4(31L) state near 375 nm.
Both states share a (tzg)2(eg"‘)I orbital parentage, leading to a displacement of the

electronic potential wells along the metal-ligand coordinate with respect to the
ground state (Figure 3-13). This displacement leads to the broad absorptions

observed in the visible region.

The lowest-lying excited state, 2E, is emissive in low temperature optical
glasses. For Cr(acac)3 this emission occurs at 775 nm and is extremely narrow.
The narrowness of this transition, as with the breadth of the 4T2 absorption
described in the preceding paragraph, can be attributed to the relative geometries
of the metal coordination sphere between the two electronic states. For the case of
2E the orbital parentage is the same as the ground state 4A2 (both are (tzgﬁ Figure

3-13), leading to very small geometric distortions between these electronic states,

and thus the narrow emission spectrum.

Our group has already explored the nonradiative dynamics of Cr(acac)3
using optical ultrafast transient absorption techniques with ~ 100 fs pulses.4’7 It is
well established that the 2E state recovers to the ground state with I ~ 700 ps, and
Figure 3-14, upper panel, shows that the full transient spectrum at 5 ps
corresponds to the nanosecond full spectrum, proving that the 2E state is fully

established by 5 ps. Figure 3-14, lower panel, shows full transient spectra over the

175

ﬁrst 5 ps after excitation. These chirp-corrected spectra reveal essentially the
same lineshape over this time frame, even as early as 400 fs, showing that the 2E
state is established faster than the resolution of the measurement. The slight
narrowing and loss of intensity of the spectrum over the ﬁrst 5 ps was assigned as
due to vibrational cooling in the 2E; the signatures of this process were discussed

above and will not be elaborated on here. The slight decay of the spectrum was ﬁt

to a lifetime of 1.1 i 0.1 ps.

 

 

 

 

.l

 

Cr(acac)3 - a

 

 

.h
0

 

00
O

 

 

 

Energy (cm'1x 103)
8 B

 

 

 

 

 

 

 

 

Figure 3-13: Semi-qualitative energy level diagram (left) and diagram showing
the orbital parentage of the relevant ligand ﬁeld electronic states (right) for

Cr(acac)3.

176

 

 

 

 

 

 

 

 

 

20 - -
15 - -
10 l -
5 - a
8"
db 0* -
‘— i ; i i i i
25,
g .5- -
1O ' —-350 fs ‘
, —400 fs .
— 1.1 ps \
5 - — 1.5 ‘w -
—2.5 ~
t _ 5 "n, "m...
o- :- s .- —. < e
450 500 550 600 650 700
Wavelength

 

 

 

Figure 3-14: Data revealing vibrational cooling dynamics on the 2B surface of
Cr(acac)3 after excitation of the sample at 625 nm, adapted from reference 6. See

text for details.

177

This picture of nonradiative decay in these systems was slightly modiﬁed by a
report by Kunttu and coworkers,39 wherein they reported the relaxation dynamics
of Cr(acac)3 as probed by ultrafast transient infrared spectroscopy. The question
they wished to address was whether relaxation via the 2E constitutes a major
relaxation pathway. By monitoring the uCO and uC=C stretches after pumping
into the 4T] (400 nm) and 4LMCT (345 nm) excited states they were able to
address this question, by assuming that the vibrational characteristics of the 2E
state are distinct from those of the ground 4A; state. The transient data shows
complete recovery of the ground state. The time evolution of the transient uC=C
stretch (pump at 400 nm and probe at 1521 cm") was ﬁt with biphasic kinetics
with time constants 1:] = 15 ps and 12 = 760 ps. While the 760 ps component
corresponds to ground state recovery via the 2E state, the authors explain the faster
dynamics as follows: after pumping into a high excited state and prompt
intersystem crossing into the doublet manifold, redistribution of the excess energy
into vibrations leads to a heating of the molecule. This in turn enhances BISC into
the quartet manifold until vibrational cooling leads to a trapping of the remaining
population in the 2E state. Most of the ground state (70-85%) population recovers
with a time constant of 15 ps, attributed to internal conversion from the 4T2. These
results are not at odds with the data acquired in the optical ultrafast transient
absorption experiment, and serve to indicate the limitations of the transient

absorption experiment in the absence of a deﬁnite 4T2 feature. The dynamics of

178

the low-lying ligand ﬁeld manifold are summarized in the Jablonski diagram of

 

 

 

 

 

 

 

 

 

Figure 3-15.
A
4 7315C < 10013
T2 I ——————— 9 I'm =1 lps
> | TISC <50fS' :
U) I
o : —i— 2E
If, : rm =15ps ’,
h I l
I ,’
I ’1
V I rm ~ 700ps
I
4 LI
A2 —

 

 

 

 

Figure 3-15: Jablonski diagram summarizing the nonradiative rates in the low-

lying ligand ﬁeld manifold of Cr(acac)3.

In an effort to resolve the ultrafast intersystem crossing event for this system,
transient absorption experiments were carried out using the ultrafast system of the
Beck group at Michigan State University, which produces ~ 50 fs pulses. The
sample was prepared in acetonitrile with an absorbance of ~0.7, and ﬂowed during
the experiment to prevent any possible signals from photodecomposition. Initial
experiments employed excitation at 600 nm (15 nm FWHM), resulting in the

preparation of a relatively cool vibrational Frank-Condom state of the 4T2 surface.

179

The probe pulse was also 600 nm, and a monochromator was employed to select
probe wavelengths of 592 i 2 nm (Figure 3-16) or 608 :l: 2 nm (Figure 3-17). The
entire probe pulse was also employed as a probe (Figure 3-18). The kinetic trace
of Figure 3-16 was obtained by subtracting an exponential decrease in excited
state absorption with a time constant of 320 fs to obtain a ﬂat baseline. This time
constant is on the appropriate time scale for intramolecular vibrational
redistribution processes. Fitting of the oscillatory component (the residual after
subtraction of the exponential components) was accomplished with a program of
local origin, and has been described in detail elsewhere.40’4l Brieﬂy, the
oscillatory residuals were ﬁt with a multicomponent model that contains both
slowly damped and rapidly damped components. The slowly damped components
can be modeled as a simple damped cosinusoid, while the rapidly damped
components are modeled in the time domain with inhomogeneously broadened
components with asymmetric Gaussian lineshapes. The data of Figure 3-16
reveals a rapidly damped oscillation which was ﬁt with two oscillatory

components: (1) a frequency of 164 cm'1 that damps with 1: = 70 fs, and (2) a
weaker component that oscillates with 75 cm'l. This weaker, 75 cm'1 component

has less than 5% of the amplitude of the 164 cm'1 oscillation, and functionally

serves to model all low-lying modes of the system. The redder probe (Figure 3-

17) reveals similar oscillatory kinetics as Figure 3-16, and a 1.6 ps ESA decay was

180

subtracted from this data. The data acquired with the integrated probe pulse

(Figure 3-18) was ﬁt with a biexponential function (1'1: 50 fs, 172: 1.3 ps).

 

 

6
4. :
2-

 

 

 

 

 

 

 

 

 

0 ' 2'00' 400' 600' 8'00'1000
Time (is)

 

 

 

Figure 3-16: Kinetic trace for Cr(acac)3 pumped at 600 nm and probed at 592 :l: 2
nm. A monexponential rise (t = 320 fs) in excited state absorption was subtracted

from the raw data (top panel) to obtain a ﬂat baseline. The resulting data (bottom

panel) are ﬁt with a rapidly damped oscillatory component (164 cm", I = 70 fs)

and one weaker component (75' cm'l).

181

 

 

 

 

 

-2- '
t: . , . , , , . é, -
|- -500 0 500 1000 1500 2000
<1 0.5

 

 

 

 

 

 

0 ' 200 ' 400T600T 800 T1000
Time (fs)

 

 

 

Figure 3-17: Kinetic trace for Cr(acac)3 pumped at 600 nm and probed at 608 i 2
nm. A 1.6 ps rise in excited state absorption was subtracted from the raw data (top
panel) to obtain a ﬂat baseline. The resulting data (bottom panel) are ﬁt with a

rapidly damped oscillatory component (165 cm], t = 70 fs) and one weaker

component (28 cm'l).

182

 

 

3.0 4 :'

2.0 -

1.0-

 

 

AT/T
I

 

 

 

 

 

0 ' 10'00 ' 20'00 ' 3000
Time (fs)

 

 

 

Figure 3-18: Kinetic trace for Cr(acac)3 pumped at 600 nm and utilizing the
integrated probe pulse centered at 600 nm (top). The data was ﬁt with a

biexponential rise in ESA (II = 50 fs, 1 2 = 1.3 ps).
These observations likely represents vibrational coherence is this system. In

recent years, vibrational coherence has been observed in many transition metal

. . . . . . 41,42
coordination systems: many cases eXist of coherence in Zinc porphyrins, heme

183

1 1,1243 45,46

complexes, other metalloproteins44 and mixed valence species. To date,

observation of vibrational coherence is rare for simple coordination complexes,

although recently coherence has been reported for [F e(bpy)3]2+.47 In most cases

coherence is observed on a single electronic state. This state can be the initially
populated excited state (so called impulsive absorption), although coherence on

the ground state, formed via impulsive stimulated Raman scattering, is also quite

common.‘“'"”’46 Control of relative contributions of excited state and ground state

coherence can be achieved by varying excitation pulse duration48 and introducing

42,49

chirp. Finally, coherence can be produced on other electronic states via

1 1,50-53

indirect coherence mechanisms, involving either photochemistry, rapid

geometric changes,47 or formation of some other type of product state (e.g. an

acceptor state in an electron transfer54 or proton transfer55 reaction). A different
phenomenon involves so—called “retention of coherence,” wherein a vibrational
wavepacket is formed on one electronic surface and propagates through a barrier
to populate another electronic state, with the wavepacket still intact. Some
interesting examples include work on photosynthetic bacterial reaction centers and
the FMO photosynthetic model complex, which has shown that coherence is
maintained between different excitonic states in these systems, disproving the

56-5

previously assumed incoherent hopping mechanism. 8 This phenomenon has

. . . . . 59
also been observed during internal conversmn 1n cyanine dye molecules,

184

photochemical rearrangements,60 photoisomerization (although many such

. . . . . 61,62 . . 45
examples involve barrierless isomerization), and mixed valence spec1es.

The time scales of ground state recovery (via the 4T2 and 2E states, as
discussed above) are much longer than the observed kinetics, suggesting an
excited state coherence. Given the previous studies that have been carried out on
this molecule (vide supra), it is reasonable to think that a conservative estimate of
the lifetime of the 4T2 ——> 2E ISC event is < 100 fs. Again, looking to the kinetic
trace of Figure 3-16, it is seen that signiﬁcant amplitude of the oscillatory
component is present at later times (> 400 fs). This observation, coupled with the
estimated lifetime for ISC, suggest that coherence is retained during this ISC
event, i.e. coherence is taking place on the Franck-Condon state and is maintained
during the ultrafast intersystem crossing event that immediately follows excitation,
a ﬁrst for a simple coordination compound. This may be the ﬁrst reported account
of retention of coherence for a spin-forbidden (i.e. intersystem crossing) event.

If retention of coherence is occurring during intersystem crossing, then the
frequency of the coherent vibration suggests the nature of the vibrational mode
that is coupled to the 4T2 —> 2E conversion (vide infra), revealing signiﬁcant
implications for the mechanism of ultrafast ISC in this complex. Vibrational
spectra of tris(acetylacetonato)metal(III) complexes have been extensively studied

both experimentally and theoretically.6""71 The Cr-O stretch is reported as a wide

absorption near 250 cm'l, and there are several other low-frequency vibrations of

185

Cr(acac)3, although no reports (infrared or Raman) of a vibrational mode near 164
cm], the observed frequency of both the 592 nm and 608 nm probes, again
suggesting an excited state coherence. This value may correspond to the
vibrational frequency of the Cr-O stretch in the 4T2 state: this excited ligand ﬁeld
state has occupation of an eg' orbital and thus weakened Cr-O bonds are expected.
A ground state frequency calculation at the UBLYP/6-3 1 lg" level, employing a
CPCM solvent model for acetonitrile (Figure 3-19), reveals that the 250 cm'l
mode is indeed metal-ligand active, but also includes large amplitude oscillations
involving the peripheral methyl groups of the acac ligand (the calculation also
reveals several vibrations near 164 cm", but these are not metal-ligand active).
This observation, along with recent results on substituted complexes of Cr(acac)3,
highlighted below, imply a signiﬁcant role for the structure of ligand backbone in
the ultrafast 4T; —+ 2E conversion.

Although the timescales for the various dynamical events reported above point
to retention of coherence, one can not discount coherence on the 2B surface as a
result of an indirect coherence mechanism. Such a mechanism has been reported
recently for [Fe(bpy)3]2+, where vibrational coherence was reported on the lowest-
energy excited state, 5T2.47 The authors note that the observed coherent frequency
of 130 cm'1 was likely due to N-Fe-N bending modes which were coherently
excited by the impulsive Fe-N stretch during the change from a low spin to a high

spin state. In the case of Cr(acac)3 one would also expect signiﬁcant geometric

186

distortions along the Cr-O bonds in the 4T2 excited state. Thus one would
anticipate the coherent mode to be metal-ligand active. The question remains
whether this coherence is formed on the Franck-Condon state or formed on the 2E
state as a result of an impulsive stretching of the Cr-O bonds on the same
timescale as intersystem crossing. Again, given the timescales observed for this

system the former is preferred.

 

 

 

 

 

 

Figure 3-19: Ground state frequency calculation for Cr(acac)3 at 250 cm'1 at the

UBLYP/6-3l l g** level, employing a CPCM solvent model for acetonitrile.

This observation has signiﬁcant implications pertaining to the mechanism
of the ultrafast ISC event. Retention of coherence during this event suggests that

the 4T2 and 2E states are very strongly coupled via the observed low frequency

187

mode (164 cm'l). This further implies that this low frequency mode (or perhaps
low-frequency modes) plays a signiﬁcant role in facilitating ultrafast ISC in this
complex. Work by Perkovic and Endicott in the early 19903 on vibrationally
constrained complexes of Chromium(III) (primarily derivatives of 1,4,7-
triazacyclononane, TACN) has implicated low-frequency modes as playing an
important role in 2 E ——+ 4A2 conversion.72 In a non-constrained system, low-
frequency modes would produce a trigonal distortion; this distortion would then
allow mixing of d-orbital character, making the way for rapid intersystem
crossing. The constraining complexes inhibit low-frequency modes, preventing a
trigonal distortion, and effectively slowing nonradiative decay via this pathway. It
is reasonable to assume that some similar mechanism, coupled with the relative
displacement and energies of the 4T2 and 2E states, may be facilitating ultrafast 4T2
—r 2E.

This description of intense vibronic coupling between two electronic states
to facilitate an ultrafast intersystem crossing event is essentially that of a spin-
forbidden “conical intersection.” Conical intersection simply describes the
molecular geometry points at which two potential energy surfaces—of the same
spatial and spin symmetry—intersect. These constitute degenerate and
intersecting points among the two surfaces, which are characterized by strong
vibronic coupling, resulting in an effective breakdown of the separation of

electronic and vibrational components of the wavefunction (Bom-Oppenheimer

188

approximation). This strong vibronic coupling provides a facile mechanism for
ultrafast interstate crossing (i.e. internal conversion, as observed mainly for

organic systems and biological systems73’74

or intersystem crossing). As a result
of this property, conical intersections have also been called “photochemical
funnels,” because the ultrafast processes that they induce effectively quench other
kinetic pathways, and funnel the energy toward a certain photochemical or
photophysical path. The breakdown of the Bom-Oppenheimer approximation is
especially relevant for large polyatomic molecules, which are characterized by a
large number of close-lying electronic states and a high number of degrees of
freedom, leading to a high density of vibrational states. 75

This study of coherence in these systems, and the implications of the

mechanisms of nonradiative decay that they suggest, are an exciting new avenue

for this research and is covered extensively in the ﬁiture work chapter, Chapter 5.

3.4.2 3-substituted Complexes of 2,4-pentanedione.

Cr(3-NOzac)3. This complex forms purple crystals, and is purple/pink as a
powder and in solution. Assignments for the UV-vis absorption spectrum, shown
in Figure 3-20, are fairly straightforward due to the similarity of this spectrum to

that of Cr(acac)3 (vide supra). Starting in the UV there is a transition centered at

272 nm (3 ~ 17000 M'1 cm'l), another of slightly lower intensity at 330 nm (8 ~

189

13400 M'1 cm!) and an anomalous transition centered near 366 nm (8 ~ 4000 M'1
cm'l). In the visible is a low-intensity broad transition centered at 560 nm.

The transition at 272 nm matches well what has been observed in Cr(acac)3
as an intraligand 7: ——> 71‘ type transition: Cr(acac)3 shows two peaks in this region
at 255 nm and 275 nm. A lower energy transition as predicted by the
nephelauxetic effect is not observed, indicating that the nitro group is not coupling
well into the inner 7: system of the acetylacetonate ligand. This is evident from the
x-ray crystal structure of this complex, presented in Appendix C, which reveals
that the it orbitals of the nitro groups are close to 90° with respect to the it system
of the acetylacetonate core. The apparent orthogonality of the nitro 11: system to
that of the acetylacetonate ligand is further manifested by the rest of the absorption
spectrum as well. The absorption at 330 nm is assigned as a ligand-to—metal
charge transfer transition based on a similar absorption observed for Cr(acac)3
centered at 333 nm. Any coupling between the nitro and acac 1!: systems should
result in a red shift of this peak relative to Cr(acac)3, which is not observed. In
fact, a new transition arises centered at 366 run that has an oscillator strength in
the range of charge transfer transitions. It is hypothesized that this transition is
due to a charge transfer transition directly from the nitro substituent to the metal
center (LMCT). The broad transition at 560 nm is easily assigned as absorption to

the spin-allowed 4T2 ligand ﬁeld state.

190

 

 

 

 

 

 

 

 

 

 

 

 

 

20 I r I ' ﬁ I l ' I I I
' -1.0
15' 8 -08
A 1 8 x10 N >
“o — Emission ‘ E
x 10 _ *- 0.6 g
‘7 'c
e *
‘7 P04 3
co
- 0.2
I I r I I ﬁ ﬁ ﬁ ' I ' I 0-0
300 400 500 600 700 800
Wavelength (nm)

 

 

 

Figure 3-20: UV-vis spectrum recorded in dichloromethane and emission
spectrum recorded in a low temperature glass of 2-methyl-tetrahydrofuran for

Cr(3-NOzac)3.

Considering the electronic similarities between this complex and Cr(acac)3
one would intuitively expect similar dynamics to those of Cr(acac)3. This is in

fact the case. Initial experiments were carried out with ~ 100 fs pulse duration,

resulting in an effective resolution (cross correlation of solvent signal) of ~ 250 fs.

191

Starting with longer timescale data (Figure 3-21), it is observed that ground state
recovery occurs with a lifetime of ~730 ps; the value for Cr(acac)3 is ~ 760 ps.
The full spectral excited state absorption dynamics (Figure 3-21, left) reveal an
excited state absorption that decreases in intensity with time without changing
bandshape, indicating absorption from the same excited state at all time points
with loss of this state to the ground state. From previous work on Cr(acac)-3 this
absorption is assigned as originating from the lowest energy excited state, 2E, and
the observed dynamics are due to loss of this state via intersystem crossing to the
ground state, 4A2. With regard to the bandshape of the 2E spectrum it is interesting
that it is clearly composed of two different Gaussian lineshapes, perhaps
corresponding to transition to two distinct charge-transfer states as discussed
above.

Moving on to shorter timescale dynamics, it is seen that the early time full
spectra (Figure 3-22, left) are again composed of a very broad absorption with
maximum excited state absorption around 510 nm. This bandshape is the same as
that observed in the longer time spectra (above), representative of absorption from ,
the 2E state. This bandshape is also observed in the full spectrum at 300 fs (not
shown), the effective resolution of our instrument, indicating that the 2E state is
formed faster than the resolution of the experiment. The early time (1 ps) full
spectra are much broader than ﬁill spectra for later times: the narrowing of the full

spectra with time is indicative of vibrational cooling, which is occurring on the 2E

192

surface. Single wavelength kinetics (Figure 3-22, right) show that this cooling
occurs with t = 1.65 i 0.03 ps, slightly slower than the same process for Cr(acac)3.
These observations were invariant with excitation wavelength across the 4T2
absorption: at all pump wavelengths the ﬁill spectra narrowed with time over the
ﬁrst ~ 10 ps, and the lifetime of this process as a function of pump wavelength was
essentially invariant (Figure 3-23). All of these observations suggest very similar

mechanisms of nonradiative decay to those observed for Cr(acac)3.

 

 

 

    

15...... 25 ..-E,
_ -1 to 300 p8 q r 633nm excitation ‘
12. _ 20F 480nmprobe ,
15- -

l
10F a

 

 

5- t=726¢13ps ~

 

 

 

 

 

Ox _ 0.3.. ........................................... -

450 T500 ' 550 ' 600 0 .400 ' 800 '1200
Wavelength (nm) Time (ps)

 

 

 

 

 

 

Figure 3-21: Full spectral ground state recovery dynamics for Cr(3-NOzacac)3
and long-time single wavelength kinetic trace for Cr(3-NOzac)3 after excitation at

633 nm.

193

 

 

 

20 - --' - — - 7— -F T ' I I ﬁ ﬁ ﬁ
Cr(3—N02ac)3 ' CI(3'N0230)3
' 630 nm excitation . 30 - 633 nm excitation _
i
- 480 nmt

_ 16510.03 ps -

 

I

 

 

 

 

 

 

 

 

560 nm
.§O§|§.§§€
l
' 680 nm ‘
AAAAAAAAAA
450 500 550 600 O 10 20

Wavelength Time (ps)

 

 

 

Figure 3-22: Full spectral and single wavelength kinetic data of Cr(3-NOzac)3 in

dichloromethane excited at 630 nm and 633 nm, respectively.

 

 

 

 

 

21: T
1.8 - -
1.5: I i 1,
g 1.31 I 1
t: 36' Cr(3—N02ac)3 1
' f 480 nm probe 1
0.3L 1

00 i r l . i I 1 g

16 17 18 19 20
Pump Energy (cm'1 x 103)

 

 

 

Figure 3-23: Observed lifetime as a ﬁinction of pump energy at 480 nm probe.

All measurements were carried out in dichloromethane.

194

Cr(3-Brac)3. This complex forms brown crystals, and is brown/yellow as a

powder and in solution. The ground state absorption spectrum is similar in most

respects to what has been observed for this class of coordination compounds. In
the ultraviolet three transitions are observed which are at 270 nm (8 ~ 6600 M'1
cm'l), 355 nm (8 ~ 8600 M'1 cm'l), and a low-intensity transition centered at 400
nm (8 ~ 480 M'1 cm'l) that presents as a shoulder on the much more intense
transition at 355 nm. In the visible one observes a low-intensity broad transition
centered at 565 nm (3 ~ 65 NT1 cm'l).

As shown previously for this class of compounds, the absorption centered
at 270 nm is ligand (1t ——> 7;) based. The similarity of this value to that of
Cr(acac)3 (272 nm) suggests relatively little perturbation in the ligand It system as
a result of substitution of Br (a It-donor and o-acceptor) in the 3-position of the
ligand backbone. The absorption at 355 nm is assigned as a LMCT type transition
based on typical assignments for this class of compounds, as well as literature
assignments. This value is red shifted by ~ 25 nm with respect to the
corresponding transition in Cr(acac)3. Again, the broad absorption centered at 565
nm is assigned as a transition into the 4T2 state. Here Amax is red shifted by 5 nm
with respect to the same transition in Cr(acac)3. This makes sense given the o-
accepting and It-donating nature of the Br substituent, which acts to decrease lODq

by lowering the energy of the cg. set while raising tzg.

195

 

 

Cr(3-Bracac)3

8

e X 10
emission

    
   

 

 

 

 

 

Normalized Intensity

 

 

 

 

 

'300'400'500'600'700'800
Wavelength (nm)

 

 

 

Figure 3-24: Absorption and emission spectra for Cr(3-Brac)3 recorded in

dichloromethane and an 80 K optical glass of 2-methyltetrahydrofuran,

respectively. See experimental section for details.

The maximum wavelength of emission for this complex has been reported
previously by Forster and DeArmond, who recorded the spectrum in a low-
temperature glass (5:5:2 ethyl ether: isopentane: ethanol) and reported a value of
12390 cm].76 This compares well to our result (obtained in a low temperature
optical glass of 2-methyltetrahydrofuran) of 12450 cm']. A recent spectroscopic

study employing luminescence and excitation line narrowing spectroscopy has

196

recently been carried out, yielding further insight into the electronic structure of
both the 4A2 and 2E electronic states, 77 revealing a 2E splitting of 138 cm'1 due to
a J ahn-Teller interaction and a 4A2 splitting of 1.4 cm'1 (corresponding perfectly to
the value of D ~ 0.7 cm'1 reported in Chapter 2, where zero-ﬁeld splitting is given
by 2D). Of general interest, and discussed by Forster and DeArmond,78 is that the
quantum yield of luminescence in this family of complexes is excitation
wavelength dependent, i.e. excitation into the ligand ﬁeld results in a different
quantum yield of emission than excitation into the charge-transfer or intraligand
transitions. For Cr(3-Brac)3, the authors report a factor of two increase in the
radiative quantum yield from the 2E state for excitation into the 4(31L) / 4T.
transition versus excitation into the 4T2 state, while pumping into the LMCT
transition provided a quantum yield in between these two values. A comparison of
the excitation scan, (monitoring at 804 nm) and the absorption spectrum (Figure 3-
25) reveals that the quantum yield of emission is deﬁnitely excitation wavelength
dependent: excitation into states with relatively low oscillator strength (such as
4(31L) / 4T1 and 4T2) emit competitively with states of much greater oscillator
strength (LMCT and IL transitions). This suggests that excitation into these states
of lower absorption intensity results in a higher quantum yield of emission.
Furthermore, assuming that the 4T1 and 4T2 states have the same oscillator

strength, the approximately two-fold increase in the excitation scan intensity of the

197

4T1 versus the 4T2 state is qualitatively in line with the report of quantum yield of
luminescence reported by Forster.

Nonradiative decay from the 2E state to recover the ground state was
probed using ultrafast transient absorption spectroscopy with ~ 100 fs pulses.
Values for ground state recovery between 1.9 — 2.5 ns were obtained for all
pump/probe combinations. The relatively large error on this measurement results
from the inability of our instrument to scan to times longer than ~ 1.2 ns. Despite
this error, it is clear that this value represents about a factor of three increase in the

ground state recovery lifetime via the 2E as compared to Cr(acac)3.

 

 

  
  

 

 

 

 

 

 

 

10 [ I I ‘ l ' I
A . —1.0 3.
m _ _ ._
$2 8 8 . . -0.8 g
x excutatlon scan _ .9
6- c:

E .

o 4" -O.4 ﬂ
‘7 ‘ - '5
g 2- -0.2 g
co ' 0
0+ ....................................... M-OD Z

300 400 500 600

Wavelength (nm)

 

 

 

Figure 3-25: Excitation scan of Cr(3-Brac)3 in an 80 K optical glass of 2-

methyltetrahydrofuran. The excitation scan reports the emission intensity (probed

near the emission maximum at 804 nm) as a function of the excitation wavelength.

198

Early time full spectra after excitation at 630 nm, the low-energy should of the 4T2
absorption, are shown in Figure 3-26. The full-spectral bandshape at 1 ps is
essentially retained over the entire time frame of the experiment, although within

the ﬁrst 10 ps the spectrum narrows.

 

 

 

 

 

 

 

400 ' 500 ' 600 ' 700
Wavelength (nm)

 

 

 

Figure 3-26: Full spectra for Cr(3-Brac)3 at 630 nm. Anomalies in the full

spectra around 630 nm are due to scatter of the pump pulse.

199

 

 

 

 

 

 

 

 

 

 

 

 

32'. $8,, Hi (a) 480 nm '
10'. if ’ . - - L
5:, l
0 I": ............ ............ . ............ ............ ‘ ........... , .......................... _ .................... .
32: 1 ' T ' ' T
{5‘ 24.— riﬁﬁﬁﬁﬁﬁﬁéﬁimnﬁoﬁﬂﬂﬂumnmﬂw-
S 16? a
5 8- (b) 550 nm 4
< OT": .......................................................................................................................... 1
< ' I T r -
1* (b) 650 nm "
miﬁiﬁﬁ‘ﬁﬁﬁ‘iiﬁeihuﬁuqeﬁqéﬁﬁﬁﬁgﬁ 1
't
..-...“ ......................................................................................................................... -
’ 2 4 T 6 8 10
Time (ps)

 

 

 

Figure 3-27: Representative single wavelength kinetic data of Cr(3-Brac)3 in

dichloromethane excited at 610 nm, with probe wavelengths reported in the ﬁgure.

The data were ﬁt with a monoexponential function with lifetime 1.4 i 0.1 ps.

This indicates an excited state absorbance from the long-lived, lowest
energy excited state 2E, implying that intersystem crossing from the initially
formed 4T2 occurs faster than the resolution of our experiment. The dynamics that

are observed in the ﬁrst 10 ps are occurring on the 2E surface, and are assigned as

vibrational cooling within this state. Single wavelength kinetics (Figure 3-27),

200

which corroborate the full spectral dynamics, reveal that this process occurs on a
slightly longer timescale than the corresponding event in Cr(acac)3:
approximately 1.4 ps for Cr(3-Brac)3 versus 1.1 ps for Cr(acac)3. This value was

obtained for all excitation wavelengths across the entire 4T2 absorption proﬁle.

Cr(3-methylac)3. Cr(3-methylac)3' forms dark green crystals and powders. The
UV-vis absorption and emission spectra of this complex are shown in Figure 3-28.
In the ultraviolet one observes two distinct transitions at 270 nm (8 ~ 9080 M 1 cm'
I) and 350 nm (8 ~ 10860 M'1 cm'l). On the low-energy shoulder of the 350 nm
absorption is an absorption of much lower intensity centered around 395 nm with 8
~ 320 M'1 cm]. In the optical region is a broad absorption of low intensity
centered around 574 nm with 8 ~ 70 M1 cm". The energy of the transition at 272

nm, which is similar to the value obtained for Cr(acac)3 and assigned as an
intraligand transition, is evidence that the ligand 1: structure is not varying much as
a result of the methyl substituent. While this is expected given the o—donating
nature of the methyl substituent, there is ample evidence that the relative energies
of the Chromium(III) based states and ligand It-based states are shifting as a result
of this substitution: the absorption at 350 nm, assigned as a LMCT transition, is
red shifted by 20 nm with respect to the same transition in Cr(acac)3. This implies
a destabilization of the ligand n-based orbitals with respect to chromium d-based

orbitals (or alternatively a stabilization of the chromium orbitals with respect to

201

the ligand it set). The moderate intensity transition at 395 nm is assigned as 4(3IL),
while the broad weak absorption at 574 nm is assigned as absorption into the 4T2
ligand ﬁeld state. This state is red-shifted by ~ 15 nm with respect to the same

transition in Cr(acac)3.

 

 

 

 

 

 

 

 

 

12 T I ' l ' F ' I ' I T l
-1.0
10- ll .
a
”A 8 8X10 -O.8 3s
2 —Emission Z’
x 1 (D
‘T 6" ”0.6 E
8 . g
‘7; 4n "-O.4-é
(O l- 5
t z
21 -0.2

 

 

. K t
o-

300 ' 400 ' 500 ' 600 ' 700 '800
Wavelength (nm)

 

 

0.0

 

 

 

 

Figure 3—28: Absorption and emission spectra for Cr(3-methylac)3 recorded in
dichloromethane and an 80 K optical glass of 2-methyltetrahydrofuran,

respectively. See experimental section for details.

202

The emission spectrum (blue trace, Figure 3-28) was recorded in a low-
temperature (80 K) optical glass of 2-methyltetrahydrofuran. It has a maximum

intensity at 12280 cm'1 (814 nm). This signiﬁes a considerable red shift of the

emission maximum from Cr(acac)3, implying that the 3-methylacetoacetonate
ligand acts as a better delocalizer of 7: density that acetylacetonate, according to
the nephelauxetic effect.

The stabilization of the 2E may have major implications for ground state
recovery dynamics via this state. These two states are nested, and, according to
nonradiative decay theory, stabilizing the 2E with respect to the ground state
(lowering AE) would increase the vibrational overlap facilitating faster
nonradiative decay. Initial ground state recovery measurements seemed to be
conﬁrming this, with monoexponential ﬁts revealing lifetimes on the order of ~
450 ps. These ﬁts, however, all showed a slight offset ﬁ'om baseline, indicating
that the original ground state was not recovered during this process, as a result of
either 1) formation of a photoproduct or 2) a completely different set of dynamics
from what has been previously observed for these systems, with ground state
recovery occurring on a much longer timescale than the experiment. There were
no other outward signs of photochemistry, i.e. precipitation of a photoproduct or
observation of different pre- and post-experimental absorption spectra. To test

these possibilities, future experiments must be carried under ﬂowing conditions.

203

 

Crl

spe

56F

“‘11

5110

abs

abs

410

[THE

abst

Cr(3-SCNac)3. Cr(3-SCNac)3 forms purple crystals and powders. The absorption
spectrum of Cr(3-SCNac)3 conforms completely with the other members of this
series, Figure 3-29. In the ultraviolet the spectrum tails up until the solvent
window is closed (~220-230 nm for dichloromethane), and on this rise is a
shoulder centered around 260 nm with 8 ~ 10200 M‘1 cm". A well-defmed
absorption is centered at 326 nm (8 ~ 10300 M"1 cm'l). On the red edge of this
absorption are shoulders of much lower intensity at 370 nm (8 ~ 480 M'1 cm!) and
410 nm (8 ~ 130 M'I cm"). In the visible one observes a broad low-intensity
transition centered around 560 nm with 3 ~ 75 M'1 cm". Given the previous
absorption spectra for this class of complexes (discussed above), the assignments
are straightforward. The shoulder at 260 nm is assigned as a ligand localized
transition originating from the delocalized 1t structure of the acetylacetonate
ligand. The absorption at 326 nm is assigned as a LMCT transition, while the
lower intensity absorptions on the shoulder of the LMCT transition are assigned as
4(31L) at 370 nm and absorption into the 4T] at 410 nm. The broad absorption
centered at 560 nm is again assigned as a transition to the lowest-energy spin

allowed ligand ﬁeld state, 4T2. All of these values match very closely those
observed for Cr(acac)3, representing only small perturbations to the electronic

structure as a result of substituting thiocyanate into the ligand backbone.

204

 

 

 

 

 

 

 

 

 

 

 

 

8 -1t)
8 x10
emission
A L0.8 g
('0 (D
e _ 8
x ~0t32§
‘T '0
g “N’
‘7 . -O.4 E
E? 4- E
to 1 Z
2‘ L -0.2
'0‘ i r00

 

T300 '400 '500 '600'700 '800
Wavelength (nm)

 

 

 

Figure 3-29: UV-vis absorption (black and red traces) and emission spectra of
Cr(SCNaC)3, collected in dichloromethane and an optical glass of 2-

methyltetrahydrofuran, respectively. See text for details.

Like the other complexes in this series, Cr(3-SCNac)3 emits in a low-

temperature optical glass. The emission spectrum (blue trace, Figure 3-29),
recorded in an 80 K optical glass of 2-methyltetrahydrofuran, has a maximum

intensity at 12920 cm'1 (774 nm) and is quite narrow. This value is equivalent to

205

that observed for Cr(acac)3, suggesting that the thiocyanate group does little to
change the it delocalizing ability of the ligand.

Femtosecond transient absorption data are presented in Figures 3-30 and 3-
31. A kinetic trace out to 1 ns, after 630 nm excitation, on the low-energy
shoulder of the 4A2 ——> 4T2 absorption was collected (not shown). This data was ﬁt
with a monoexponential function that has no offset from the baseline at long times,
indicating ground state recovery, and decays with a lifetime of 930 i 100 ps.
Early time full transient spectra are shown in Figure 3-30. Despite the noise
associated with the measurement, general features can still be gleaned from this
data. The data show a broad excited state absorption with a maximum at about
500 nm. Over the ﬁrst 10 ps the spectra narrow and decrease in intensity, without
the bandshape changing signiﬁcantly. These observations are consistent with
absorption from a single electronic state. Single-wavelength kinetic traces (Figure
3-31) conﬁrm the dynamics of the full spectra, revealing a decay in the excited
state absorption across the entire probe region. These data were ﬁt with a
monoexponential decay function yielding a lifetime of 1.65 :t 0.1 ps. These data
are entirely consistent with the picture developed for the nonradiative decay

observed in Cr(acac)3, namely ultrafast (k > 1013 s") intersystem crossing to the 2E
state and vibrational cooling on the 2E state with a 1-2 ps lifetime. This
vibrational cooling lifetime is constant for all excitation energies across the 4T2

proﬁle, Figure 3-32.

206

 

 

AA (x10'3)

 

 

 

 

0
400 450 500 550 600
Wavelength (nm)

 

 

 

 

Figure 3-30: Full spectra transient absorption spectra for Cr(3-SCNac)3 after 635

nm excitation.

 

 

 

 

 

 

15'- -
a? 121 (a). 480nm 1
b 9- i ‘
‘— ' (b).5 Onm ‘
:3 6: ﬂ"it-Htire-rteit"éiiiiﬁ'ttatm-asr'ltr- _‘
E 3; N ,,,,,,,,,,,,,,,,,,, .-- A(c).650 nm .'

L A 43...“, - “weary: ram ,

o“ ....................................................................................................... -

0 2 T 4 6 8 10
Time (ps)

 

 

 

Figure 3-31: Representative single wavelength kinetics for Cr(3-SCNac)3,
excited at 515 nm. All ﬁts were carried out with a monoexponential decay

function to yield a lifetime of 1.65 i 0.1 ps.

207

 

 

2.0 . . r . , E , -

 

 

 

l i g § -
1.6: § % T
13 1.2 - .
8- .
P 0.8 _ Cr(3-SCNac)3
- 480nm probe
0.4 - -
00 I r I . I r I 4

16 17 18 19

Pump Energy (cm'1 x 103)

 

 

 

Figure 3-32: Vibrational cooling lifetime as a function of excitation energy for

Cr(3-SCNac)3.

Cr(3-Phac)3. Cr(Phac)3 forms blue/green powders and blue crystals. The crystal
structure, shown in Figure 3-33, shows a highly symmetric environment about the
Chromium(III) metal center, with typical Cr-O bond lengths of 1.95 A. Further
crystallographic data is presented in Table 3-1 as well as Table C-2 of Appendix
C. Of interest to the electronic structure of this complex are the dihedral angles
that the phenyl groups make with respect to the inner It system of the
acetylacetonate ligand. These three values (75.880, 6829", and 87.400) indicate
that the It structures of the acetylacetonate and phenyl rings are able to couple to
some extent. Of course these values are affected by crystal packing forces and all

of the subsequent measurements are carried out in solution. The fact that the color

208

 

of

C01

 

 

 

(2,.)

of the solution is dissimilar to that of Cr(acac)3 indicates that the phenyl rings do

couple to the acetylacetonate 7t system system.

 

\ C20

‘ C19

 

9
C17 Q ‘ 18

C14
C16 ‘ C15 " C12

" ‘C13

    
  
      
     

 
   

 

      
  
 

 
     
 
 

    

   
  
   

 

O4 03
01 05
C24 C23
H m;
C. C6 C3 "a 02 06 ‘ C25 C29
y C4 C28 '
C1o‘ ‘ § C27 ‘ C30
r C5 C33 '0
‘ I
“ . C7 " ‘C31
C9
C8 C32

 

Figure 3-33: Crystal structure of Cr(3-Phac)3. Representative bond lengths and

angles are given in Table 3-1.

The absorption and emission spectra for Cr(3-Phac)3 are shown in Figure 3-

34. In the ultraviolet is an intense sub-300 nm transition that increases in intensity
until the solvent window is closed, while a well-deﬁned transition is centered at

346 nm (8 ~ 11400 M"1 0m!) with a shoulder of much lower intensity on the red

edge (394 nm, (3 ~ 450 M" cm“). In the visible is a low-intensity transition

209

centered at 571 nm with 8 ~ 105 M'1 cm']. The absorptions correspond broadly in
shape and position to those observed for the rest of this series, and the assignments
are made accordingly. Namely, in the ultraviolet, the intense sub-300 nm
absorptions are assigned as intraligand transitions, the band at 346 nm as a charge-
transfer transition, and the moderate intensity band at 394 nm as a 4(31L) type
transition. The energy of the charge transfer transition (346 nm) represents an ~
20 nm red Shift of this peak with respect to the same peak for Cr(acac)3, which is
likely due to the lower-lying 11: structure of the ligand as a result of the phenyl
substituent of Cr(3-Phacac)3 coupling into the inner 7t system of the

acetylacetonate ligand.

Table 3-1: Select bond lengths and angles for Cr(3-Phac)3.

 

 

 

 

 

 

 

Cr1-O1 1 .9466

Cr-O Cr1-02 1 .9555
Bonds Cr1-03 1 .9567
(A) Cr1-O4 1 .9502
Cr1-05 1 .9391

Cr1-06 1 .9502

O1-Cr1-05 178.63

02-Cr1-03 176.51

O-Cr—O O4-Cr1-06 177.56
Angles O1-Cr1-O4 92.69
(0) O1-Cr1-02 88.76
OZ-Cr1-O6 91 .57

02-Cr1-05 90.52

Phenyl C2-CB-C6-C1 1 75.88
Dihedral C15-C14-C17-C18 68.29
(0) 024-025-C28-C29 87.40

 

210

This increased delocalization of the ligand 1t system is also manifested in the low-

temperature emission spectrum, whose maximum intensity is at 12523 nm (798.5
nm), also red-shifted with respect to the emission spectrum of Cr(acac)3. This is

not surprising given the moderate coupling of the phenyl and acetylacetonate 7t

systems predicted from the x-ray crystal structure.

 

 

 

 

 

 

 

 

 

 

 

 

 

16 I w l ' F ' l ' l ' l
14_ 91.0
12- 8 ~08
A ‘ n 8X10 ' g
"o 10‘ —Emission 02)
‘7 b '0
E ‘ 0ND
‘70 6‘ ~04?
2 ‘ E
i T: 4‘ P 2
21 -O.2
°‘*"“.‘“": """ . """ : ' ."'": """ I , . .' 0-0
300 400 500 600 700 800
Wavelength (nm)

 

 

 

Figure 3-34: UV-vis absorption (black and red traces) and emission spectra of
Cr(3-Phac)3, collected in dichloromethane and an optical glass of 2-

methyltetrahydrofuran, respectively. See text for details.

211

Representative long-timescale (ps-ns) time resolved data is shown in Figure
3-35. For all pump/probe combinations a lifetime of 1.7 - 1.9 ns was obtained
after ﬁtting the data with a monoexponential function. Again, this value exceeds
the limit to which our instrument can scan, resulting in the large error on the
measurement. Shorter timescale data are shown in Figure 3-36 (full spectra over

the ﬁrst 5 ps) and Figure 3-37 (single wavelength kinetic traces).

 

 

 

 

 

 

 

16-' ' ' ' ' ' T " -
141 i _‘
121 l
”A 10’- J
E at _
ii 61 _
E 4L _
21. _
o i ............. ,,,,,,,,,,,, ,,
o I 300 I 600 I 900 .1200
Time (ps)

 

Figure 3-35: Representative data of ground state recovery for Cr(3-Phac)3. This

data was acquired with an excitation wavelength of 610 nm and a probe
wavelength of 480 nm. The value of 1.7 - 1.9 ns ground state recovery time was

found for all pump/probe combinations.

212

The full spectra data of Figure 3-36 reveals an excited state absorption that is quite
broad and dramatically redshifts and narrows over the ﬁrst three picoseconds,
from a band maximum of 535 nm to 560 nm. This prominent shifting of the full
spectrum manifests itself accordingly in single-wavelength kinetics (Figure 3-37):
on the blue edge the kinetics reveal a decay in the excited state absorption, while a
rise in the excited state absorption is observed on the red edge as the absorption
shifts. The lifetime of the decay on the blue edge is constant for excitation across

the entire excited state absorption proﬁle, Figure 3-3 8.

213

 

 

 

 

30— / ‘ —
l .
25+ ~ 4
00620?
o .
‘_
515 —':‘|pS-
l
<<E1o— _—§ _
' —4
5- —5 -
l
of" i

 

 

 

450 500 550 600
Wavelength (nm)

 

 

 

Figure 3-36: Full spectra transient absorption spectra of Cr(3-Phac)3 at 633 nm

excitation in dichloromethane.

214

 

 

 

 

 

 

 

0

x15* 560nm ‘

v1 1 WQOQQ§CQIQ§§§§§Q.IQQT

2 °’ 1
5- “I 680nm -

- OOOOOOOOOOOOOOOOO...‘

()_". ................................................................................. -

l l l n l

O 1 0 20 3O

 

 

 

 

 

 

Figure 3-37: Single wavelength kinetics for Cr(3-Ph-acac)3 excited at 560 nm

(top) and 630 nm (bottom). All ﬁts at 480 nm probe were carried out with a

monoexponential decay function yielding 1.30 :t 0.05 ps.

215

 

 

1.51 i 1
1.21 ' i

9,- 09" Cr(3—Phac)3

P 0.6 ~ 480 nm probe
0.3}
0'0 116 T 1? 1‘8 1‘9 ' 20
Pump Energy (cm"1 x 103)

 

 

 

 

 

 

Figure 3-38: Lifetime of transient absorption dynamics on the blue edge of the

excited state absorption as a function of excitation wavelength.

The dynamics revealed on long timescales likely correspond to ground state
recovery via the 2E state, as observed in other members of this series. The value
of ~ 1.8 ns corresponds to a 2-3 fold increase in ground state recovery lifetime as
compared to Cr(acac)3. The shorter timescale dynamics are different than those
observed for the rest of this series. The apparent shift in the full spectra could be
due to one of the following events: 1) the shifting and narrowing of the spectrum
is due to cooling on a single surface, where the surface being probed represents a
softer potential than that of the state to which the probe induces a transition,
similar to the situation presented in Figure 3-5; 2) the dynamics are occurring on a
single electronic surface, and the observed shifting of the hill spectra is due to

geometrical changes during cooling in this state, while the narrowing is due to

216

changing Franck-Condon factors as the system cools; 3.) the observed dynamics
are the result of a surface crossing in which the initially excited 4T2 state absorbs
bluer than the long-lived 2E. Scenario 3 involves changing populations of the 4T2
and 2E states occurring over the ﬁrst few picoseconds causing the apparent shift of
the spectrum.

Scenarios 1 and 2 are preferred given the observed dynamics across this
series, as well as considering the mode by which the ultrafast intersystem crossing
is likely occurring, as revealed by coherence measurements on Cr(acac)3 (vide
supra): namely, via low-frequency modes involving the methyl groups of the
ligand backbone. Scenario 2 can be accounted for if there was some energetic
beneﬁt to the phenyl ring on the ligand backbone to be more in-plane with the
acetylacetonate 1: structure in the 2E state versus the 4A2. In this model the 4T2 —>
2E ISC occurs faster than the resolution of the instrument, so that only dynamics
ﬁom the 2E state are observed. As the state vibrationally cools on the ~ 1 ps
timescale the phenyl group begins to rotate in-plane, increasing the coupling
between the phenyl group and 11: system of the acetylacetonate ligand. As this
occurs, the excited-state electronic structure changes as well, lowering the energy
of the charge-transfer manifold, the state into which excited state dynamics are
probed, resulting in the red shift of the spectrum over the ﬁrst few picoseconds.
(The 2E energy would be lowered as well according to the nephelauxetic effect,

but to a lesser extent than the charge-transfer energy.) This phenyl-rotation

217

 

 

 

 

Energy (cm'1 x 103)
B 8

_x
Q

 

 

 

 

 

‘ M-L Bond Distance
|

Figure 3-39: Potential energy surfaces diagram highlighting the charge-transfer

 

transition that is employed to monitor dynamics on the 2B surface. If the phenyl
group of the 3-phenyl-acetylacetonate ligand rotates more in-plane with the core
of the acetylacetonate ligand the energy of the charge-transfer transition employed
as a probe lowers with respect to the ligand ﬁeld manifold, resulting in a red-shiﬁ

of the excited state spectrum.

218

hypothesis will be tested by using Cr(3-mesac)3, a complex in which rotation of

the phenyl rings would be sterically hindered by the ortho-methyl groups of the
mesityl ring. Initial characterization of the electronic structure of this complex is

presented in the following section.

Cr(3-mesac)3. This complex forms purple powders and crystals similar in color
to Cr(acac)3. Single crystals can easily be grown by slow evaporation of an
approximately 5:1 dichloromethane/methanol solution. While a crystal structure is
not presented (collection is currently underway during writing of this section), the
fact that the color of this complex is more similar to that of Cr(acac); (purple) than
that of Cr(3-Phac)3 (blue/green) suggests that the mesityl group is more orthogonal
to the acetylacetonate 1r system than the unsubstituted phenyl ring. This is
expected from simple steric arguments, but nevertheless a crystal structure is
required to deﬁnitively make this assignment.

The UV-vis absorption spectrum of this complex is shown in Figure 3-40.
The main features of the absorption spectrum are an intense absorption at 354 nm
(8 ~ 12311 M'1 cm'l), a moderate intensity absorption on the low-energy onset of
the intense transition (390 nm, 8 ~ 550 M'1 cm'l), and a broad, weak absorption in
the visible at 564 nm with 8 ~ 100 M'1 cm'l. From characterization of members of
this series, these features are assigned as charge transfer, 4(31L), and 4A2 —> 4T2

transitions, respectively.

219

 

 

 

 

 

 

 

 

 

L 300 L400 ‘ 500 I600‘ 700 A 800
Wavelength (nm)

 

 

 

Figure 3-40: Absorption spectrum of Cr(3-mesac)3 in dichloromethane.

3.4.3 1,3-substituted complexes of 1,3-propanedione

Cr(t-butylac)3. The absorption and emission spectra of Cr(t-butylac)3 are shown
in Figure 3-41. The appearance of these spectra should be fairly familiar to the
reader of the previous section. In the ultraviolet one ﬁnds two absorption peaks of

large intensity at 279 nm (10300 M" cm") and 335 nm (13100 M" cm") and a

moderate strength absorption at 379 nm (450 M'] cm'l). From previous

220

_—

assignments, and given that the 7t structure of the ligand is essentially unchanged
from acetylacetonate, these absorptions are assigned as ligand II —+ 1t: 4CT, and

4(31L), respectively.

 

 

 

 

 

 

 

 

 

 

 

16...,1..,2,.r
PLO
n —8
12
A 8x10 ”0'8 >
”3 —Emission E
x -0.693
‘T 8- E
:E’ 0475’
I I— . a
3 E
w 4- o
+0.22
0- ---o.o

 

 

 

' 300 ' 460 I 560 600 ' 760 800
Wavelength (nm)

 

 

 

Figure 3-41: UV-vis and emission spectra of Cr(tbutylac)3 in dichloromethane

and 2-methyltetrahydrofuran.

In fact, the energies of the absorption maximum reported above represent a very

small perturbation to the ligand 71: structure. In the visible one observes two low-

intensity transitions: one on the low-energy shoulder of the 4(31L) absorption and ’

221

*9?-

another, more well-deﬁned peak centered at 570 nm with 8 ~ 95 M"1 cm'l. These
features are assigned as the spin allowed 4T2 and 4T1 transitions, respectively. An
emission spectrum was collected in an 80 K optical glass of 2—methyl-
tetrahydrofuran, and is shown in Figure 3-41, blue trace. The spectrum is quite

narrow, consistent with emission from the intraconﬁgurational 2E state, with the

emission maximum at 12389 cm'1 (807 nm). This value represents a ~500 cm"1

redshift to that of the emission maximum of Cr(acac)3 at 775 nm. This
observation is curious because a shift of the emission spectrum from one complex
to another is usually explained in terms of the ir-delocalizing ability of the ligand
vis-a-vis the nephelauxetic effect: a red-shift of the spectrum represents a more it-
delocalizing ligand. In this case the aforementioned argument can not be applied
as the t-butyl groups have no effective n—delocalizing ability.

Representative data of the early time dynamics, collected with ~ 100 fs
optical pulses, is shown in Figure 3-42. These data were excited at 633 nm (full
spectra) or 610 nm (single-wavelength kinetics), on the red edge of the 4T2
absorption. The transient full spectra show a broad spectrum at early times (300
fs), which narrows, red shifts, and increases in intensity with a maximum
absorption at 530 nm. Single-wavelength kinetic traces conﬁrm that these
dynamics are real: probing at 450 nm reveals a decay in the excited state
absorption, and probing at 550 nm reveals a rise in the excited state absorption.

Both the rise and the decay components are occurring with 1 ~ 2 ps. The proﬁle of

222

"0 are -Woﬁ

.>

the spectrum at 12 ps is also observed at much later times, conﬁrming that this
transient spectrum originates out of the long-lived 2E state. There are two distinct

possibilities to account for the dynamics observed in the transient full spectra: 1)
the dynamics are taking place on a single electronic surface, and the observed
narrowing, red-shifting, and grth of the signal in the red represents vibrational
cooling dynamics on this surface; 2) the dynamics represent absorption from two
different excited states, one whose excited state absorption is decaying with time

and another that has increasing excited state absorption with time.

 

 

 

I I I 1 I I ﬂ U l
I— -l

i 610 nm excitation
T ......... -

    
 
 

    

.......

  
 
 
  
 

550 nm probe, -
1:: 1.58 :l: 0.13 ps

450 nm probe, -
I: 2.00 :l: 0.17 ps

 

 

 

 

 

 

 

 

 

 

~e====a=a=§

................................................. 1

091+

450 500 550 600 ' o 4 8 12 16 20
Wavelength (nm) Time (ps)

 

Figure 3-42: Full spectra of Cr(tbutylac)3 after excitation 633 nm and select
single wavelength kinetic traces after 610 nm excitation. Similar evolution of the
full spectra were observed for other excitation wavelengths across the 4A2 —> 4T2

transition.

223

 

 

 

 

 

 

 

 

 

 

 

 

—1ps
—2
24~ —6 ‘ -
—12 *
(VPA
020— -
>< ' 0‘
v \
g 16 _ \ q
12- g, -
OT n_ l l k l L+ L A,
23f —300fs I -
—1ps ,
—2

24- —6 ‘ -
—12 4

(TBA

020- 1

x .~

v \

£16-//\ -
12‘ , [I _(L

OE j l I l . l 1 l s l A g
24 23 22 21 20 19 18
Energy (cm'1 x 103)

 

 

 

 

 

 

Figure 3-43: Gaussian deconvolution of the transient full spectra. An example ﬁt

for the transient spectrum at 2 ps is shown. See text for details of the ﬁts.

224

Gaussian deconvolution of the transient spectra was carried out by
assuming absorption from two different electronic states, with the additional
assumption that the peak maximum and the width of these absorption proﬁles are
essentially unchanged during the dynamics. These assumptions are crude and
unfortunate given that the evolution of an excited state necessarily leads to
changing absorption cross-section, affecting both the bandwidth and peak position
of the absorption proﬁle. However, despite all of the crude assumptions employed
in the ﬁtting procedure, all of the time slices where ﬁt reasonably well with this
model. Fitting of a transient spectrum at much later times (~ 20 ps), which
corresponds to absorption from the 2E only, revealed the peak position and
bandwidth of the transient absorption ﬁ'om this state, as well as a long-time offset
of AA (= 0.011), which was employed as the offset for both Gaussian features.

The ﬁts, ﬁgure 3-43, reveal one Gaussian component centered at 21500 cm'1 (465
nm), with a full width at half-maximum of 1750 cm'], which decays over time, and
a second component at 19050 cm"1 (525 nm), whose FWHM is between 1100-
1300 cm'1 and increases in intensity over time. At 300 fs these two Gaussians
have nearly equal amplitude, and by 12 ps the Gaussian at 465 run only makes up
a very small contribution of the total signal. The success of this ﬁtting procedure
strengthens the argument that absorption from two different excited states is
observed. If indeed this is the case, the bluer feature is absorption from the 4T2

state, and decay of this feature represents depopulation of this state. The

225

IT“- ---.,--.--—.

-—-

concomitant grth of the redder feature therefore represents population of 2E

state and likely encompasses vibrational cooling dynamics on this state as well.

Cr(prop)3. This complex forms reddish-purple crystals and powders, similar to
Cr(acac)3. The UV-Vis and low temperature emission spectra are reported in
Figure 3-44. In the ultraviolet one observes a series of peaks of high intensity,
with a maximum at 265 nm (8900 cm'1 M'l), and another well-deﬁned intense
peak at 345 nm (11300 cm'1 M'l). The series of peaks are assigned as intraligand
7t —> n‘ transitions based on their energies, while the peak at 345 nm is assigned as
a 4CT transition. The ﬁne structure observed in Cr(prop)3 is apparent both in the
250 — 350 nm range as well as on the low energy shoulder of the charge transfer
transition. The absorption spectrum in these ranges, plotted in energy units, is
reproduced in greater detail in Figure 3-44. In the ultraviolet, the splitting
between the peaks is ~ 2700 cm], likely due to a C-H stretch of the ligand

backbone.

In the visible there are several transitions. As aforementioned, there is a
moderate strength absorption composed of a vibronic progression of 1000 cm'1
splitting on the low-energy shoulder of the charge transfer transition, centered at
about 400 nm with 8 ~ 380 M’1 cm". It should be noted that the lowest-energy

component may be primarily 4T1 in nature, and only coincidentally at the

appropriate energy to ﬁt into the vibronic progression. This splitting of ~ 1000

226

cm'l likely represents C=C stretches of the ligand ring system: DF T calculations
reveal modes near 1000 cm'1 that are due to bending and stretching of the ligand
backbone. The well-resolved low intensity feature at 555 nm (60 M'] cm']) is

assigned as absorption into the 4T2.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

14‘ 1.0
8
a x25 . —O.8
—Emrssron z,
or)" '17)
C
3 -0.6 a
x 5
‘TE 8
.5
‘70 -o.4 E
E 4- r 5
"° ~ -o.2 Z
2;
,
ll 0‘ L\ J -0.0
300 400 500 600 700 800
Wavelength (nm)

 

 

 

Figure 3-44: Absorption and emission spectra of Cr(prop); recording in

dichloromethane and an 80 K optical glass of 2-methyltetrahydrofuran,

respectively.

227

An emission spectrum was collected in a low-temperature (80 K) optical glass of
2-methyltetrahydrofuran. The spectrum has one resolvable peak centered at 12330
cm'1 (811 nm) and another, more intense peak that is beyond the trustworthy range

of our detector (>815 nm).

 

 

 

  
   
  

 

 
   
 

 
      

 

 

 

 

 

 

 

 

 

12 :— 8 ' 8 ‘
or," 10 i 8 x 25 ’ 2600 cm'1 «
S i _1 2900 crn‘1
x 8 - 1000 cm ~ -
‘7 : 0 i
E 6 - — -
o l-
: 4 - - 4L -
w , . .
2 f] - 2 j —
or n l A l L O I A l n l L
1.5 2.0 2.5 3.0 3.5 3.0 3.5 4.0 4.5
Energy (cm'1 x 104) Energy (cm'1 x 104)

 

 

 

 

Figure 3-45: Absorption spectrum of Cr(prop)3 in dichloromethane, revealing the

vibronic structure with ~ 1000 cm'1 splitting on the red edge of the charge transfer

band.

Ultrafast transient absorption spectroscopy of this complex, utilizing ~ 100
fs Optical pulses, was initially carried out in dichloromethane in a static quartz

cuvette; however the sample showed signs of photodecomposition after a few

228

minutes. The solvent was changed to acetonitrile so that the sample could be
ﬂowed. Long time single wavelength kinetics occur on a much longer timescale
than our ultrafast instrument (I is likely greater than 5 ns). Preliminary study of
shorter timescale dynamics (Figure 3-46) reveal biphasic kinetics, with a fast
decay component (I ~ 700 fs) and an order of magnitude slower rise in the excited
state absorption (1 ~ 7 ps). At bluer wavelengths only the fast component is
observed, and the slower component is observed only in a narrow window due to

the small change in amplitude.

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

1O ’- 4 10 l l l
i
8 _ _
“3A 6 L _
o a:
2’. 4 i- . 480 nm 1
2 _ l Fits: ~700 fs decay q
I ~ 7 ps rise
I
0 -.i ......................................... .. ' 20 6 52.5 536 5?5
_ i ‘ 0 ¥ ..........................
-2 __! l i 1 4 1 . J . 1 . 1
O 1 O 20 30 450 500 550 600
Time (ps) Wavelength (nm)

 

 

 

Figure 3-46: Short timescale dynamics of a ﬂowing sample of Cr(prop)3 in

acetonitrile after excitation at 633 nm.

229

Given the paucity of electronic states involved, as well as the low-energy
excitation of the 4T2 state, only a handful of scenarios exist to account for the
observed dynamics. The biexponential kinetics suggest dynamics originating from
two different electronic states. One such scenario would involve the ~0.7 ps
component corresponding to depopulation of the 4T2 state—either via internal
conversion to reform the ground state or via intersystem crossing to form the 2E
state. Given the trends observed for this series, namely that the intersystem
crossing time is varying with substitution, it is not unreasonable that the faster 700
fs component is due to 4T2 —> 2E intersystem crossing. The long time offset of the
kinetics suggests population of the long-lived 2E state, therefore the slower ~7 ps
component is likely vibrational cooling in 2E, which is much slower than that of
Cr(acac)3. The disparate timescale for this complex as compared to Cr(acac)3 may
be due to a of lack of low-frequency modes by which to distribute excess
vibrational energy. This relates to the notion of “vibrationally deﬁcient” molecules
which was introduced in chapter 1. It has been suggested that this slower
component may be due to a photochemical reaction involving chelation of a
solvent acetonitrile molecule to the metal center. While displacement of one of
the chelating ligand oxygen sites is unlikely, Kirk79 points to the possibility of 7-
coordinate photochemical intermediates. Such a photochemical intermediate is
much more likely for this complex than others in the series due to the relatively

open coordination sphere as a result of the sterically smaller propanedione ligands.

230

Future experiments must employ a non-bonding solvent in order to test this

possibility.

Cr(dcm)3. The absorption and low-temperature emission spectra of Cr(dcm)3 are
shown in Figure 3-47. In the ultraviolet one observes a high-intensity peak at 280
nm (12770 M‘1 cm") which matches closely to that of the gallium(III) analog,
which has an absorption maximum at 282 (43800 M'1 cm!) The peaks in this
region are therefore assigned as intraligand transitions. Also observed in the
ultraviolet is an intense peak at 338 nm (17200 M'1 cm'l). Based on previous
assignments of similar peaks across this series, as ‘well comparison with the
gallium(III) analog, this absorption is assigned as transition to a 4C1“ state. On the
edge of the visible spectrum one observes a moderate intensity transition at 380
nm with 8 ~ 540 M'1 cm", assigned as absorption into the 4(31L). A low-intensity
shoulder, assigned as absorption into the 4T1 state, is also observed. Finally,
absorption into the 4T2 is seen at 570 nm with 6 ~ 90 M 1 cm". All of these values
represent very small deviations from the values observed for Cr(acac)3. The
emission spectrum (blue trace, Figure 3-47), collected in an 80 K optical glass of
2-methyltetrahydrofuran, is quite narrow, consistent with emission from the
intraconﬁgurational 2E state. The emission maximum is at 12500 cm'1 (800 nm),
redshiﬁed by several hundred wavenumbers from the emission maximum of

Cr(acac)3.

231

 

 

 

 

 

 

 

 

  
 

 

 

 

 

 

. — 1.0
a, Cr(dcm)
4- n 3 :

a x20, Cr(dcm)3 P0 8
6;“ e, Ga(dcm)3 g
$3 3- . . r g
x — Emussuon, Cr(dcm)3 10.6 BE)
5
:5, ~04 E

-0.2

L 0.6

 

 

 

300 400 I 560 I 600 ' 760 I 800 '

Wavelength (nm)

 

 

 

Figure 3-47: UV-vis and emission of Cr(dcm)3 in dichloromethane and a low-
temperature glass of 2-methyltetrahydroﬁiran. The electronic absorption spectrum

of Ga(dcm)3 reveals the relative energy of the intraligand transition.

Long-time (ps-ns) transient absorption kinetics (not shown) were ﬁt with a
monoexponential decay function which showed a negligible offset from the
baseline, indicating ground state recovery with a lifetime of 1.22 i 0.03 ns.

Shorter timescale (-l to 10 ps after excitation) single wavelength data for

232

 

610 nm excitation 515 nm excitation

 

 

 

 

 

       
 

 

 

 

 

 

 

10 t , . 20 i i
' )480 . ’ T ‘HTIEPIE: " 1
8 nm 1 11 ll. 10
8- M -15'i Ta)530nim‘
A r . 0 m . ‘ i
C? 6 - ‘1n. 7 12 h ' 4 ﬂ
2 + i b) 480 nm
25, 4 _ c) 520 nmq 8 t ‘
E 4- - _
2_£ _ .13 c)640nmW
1 . o I ................................................. ,
I l
0 ifg ................................................... - Li
1 1 1 1 1 _4 1 1 1 1
0

 

 

 

Figure 3-48: Single wavelength kinetic data for Cr(DCM)3 excited at 515 nm
(right) and 610 nm (left). All data were ﬁt with monoexponential decay functions.
The data at 515 nm excitation reveal lifetimes of a.) 1.78 i 0.50 (530 nm probe),
b.) 1.93 :t 0.25 ps (480 nm probe) and c.) 1.02 i 0.14 ps (640 nm probe). At 610
nm excitation the lifetimes are a.) 1.43 i 0.26 ps (480 nm probe), b.) 1.40 i 0.25

ps (560 nm probe) and c.) 0.49 i 0.12 ps (520 nm probe). See text for details.

excitation at 515 nm and 610 nm, at the blue and red edges of the 4T2 absorption,

are shown in Figure 3-48. These data reveal, in general, decay in the ESA with a
lifetime between 1.3 - 2 ps for bluer probes, and a rise in the ESA for probe

wavelengths between 500-600 nm, which has an ~1 ps lifetime. A decay in the

233

excited state absorption is observed for probe wavelengths to the red of 600 nm.
Certain pump/probe combinations reveal a 500 fs rise component. These dynamics
are consistent with those observed for Cr(t-butylac)3, where a decay is observed on
the blue edge of the excited state absorption, and a rise is observed for redder
portions, and is consistent with a two-state dynamical process. Unfortunately,
collection of full spectra proved to be very difﬁcult, so the assignment of

intersystem crossing dynamics remains speculative for this complex.

Cr(dbm)3. This complex forms brown crystals, brown/yellow powders and is
brown/yellow in solution. Crystals for this complex can be grown by slow
evaporation from a methanol/dichloromethane solution for both Cr(dbm)3 and its
gallium(III) analog. The crystal structure of Cr(dbm)3 is shown in Figure 3-49,
crystallographic data for Cr(dbm)3 and Ga(dbm)3 are given in Table C-3 of
appendix C, and relevant bond lengths and angles are presented in Table 3-2. The
crystal structure reveals a highly symmetric environment around the metal center,
with all M-O bond lengths 1.95 i 0.02 A, and O-M—O bond angles averaging ~ 900
and ~ 177°. Another important observation with pertinence to the electronic
structure is the dihedral angle of the phenyl groups of the ligand to the inner 7t
structure of the ligand. This shallow angle (~ 20°) ensures that the phenyl rings

are coupling effectively into the inner 1! system of the ligand, drastically changing

the ligand electronic structure (with respect to acetylacetonate), as discussed

234

below. A DFT calculation was carried out at the UB3LYP/6-311g** level
employing a CPCM solvent model of dichloromethane in order to determine the
solution-phase structure of this complex. The bond lengths and angles of the
minimized geometry correlate well to the average values of the crystal structure

geometry (Table 3-2).

 

Q "
:. Ligand2
\
\ ’
I»:
Ligand 1
. Ligand3

Figure 3-49: Cr(DBM)3 crystal structure.

8

  

 

 

 

Absorption and emission spectra for this complex, as well as the absorption
spectrum of the gallium(III) analog, are shown in Figure 3-50. The gallium(III)
complex of this ligand absorbs intensely out to about 400 nm. In the
Chromium(III) complex this intensity is quenched and further high-intensity

transitions arise which extend out to about 450 nm. The high-intensity band

235

centered at 400 nm is likely due to a low-energy charge transfer transitiOn,

facilitated by the low-lying 11: structure of the ligands. This band resides where the

Table 3-2: Relevant average bond lengths and bond angles for Cr(DBM)3 and
Ga(DBM)3. Calculated values were obtained at the UB3LYP/6-311g** level

employing a CPCM solvent model of dichloromethane.

 

 

Ga(dbm)3 Cr(dbm)3 Cr(dbm)3
calc
M-O 1.954567 1.9568 1.9568
(avg. A)

O—M-O 90.014 177.689 89.86584
(avg 0) 177.03 90.00373 179.5245
Ligand 1 -25.68 -26.2762

Phenyl 1 .05 4.7167
Dihedral Ligand 2 19.83 5.6976 333
(°) 28.45 -8.3421 '
Ligand 3 25.99 29.7017
15.59 28.4038

 

 

 

moderate-strength transitions of Cr(acac); and Cr(dcm)3 are found (as well as
many other member of this series), and in fact low-temperature absorption spectra
of Cr(dbm)3 reveal a structured shoulder on the low-energy edge of the charge-
transfer band whose spacing is ~ 1300 cm", corresponding to the C=C stretch of

the monodeprotonated dibenzoylmethane ligand. The implications of this ﬁne

236

structure have been discussed (vide supra), leading to an assignment of absorption
in this region as being partially due to a transition into the mixed 4(31L) state.
Similar to Cr(acac)3, the 4A2 —> 4T2 absorption occurs near 560 nm, although the
extinction coefﬁcient is around 100 M'1 cm'l, larger than the ~60 M 1 cm'1 of both
Cr(acac)3 and Cr(dcm)3. The 2E —> 4A2 emission is nearly 1000 cm'1 red of
Cr(acac)3 emission, which is expected from the extended 1: structure of the ligand

according to the nephelauxetic effect. This emission appears broader than that of
other members of this series, but this is likely an artifact of the low detector

response to the red of ~ 810 nm.

237

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8 _ T I ' I ' I ' I ' I ' I
. ———Cr(dbm):3 — 1.0
71 A — Absorption x50 ﬂ .
6 ‘ Ga(dbm)3 _ 0.8
A . 1 Cr(dbm)3 emission Z
v 5 .. 2
9 . ~ 0 6 3
x 4 _ ' El,
‘7 . iii
LE) 3 - e 0 4 2
IE 2 I- g
7; - _ 0.2 ‘4’.-
1 _ <
. i
O ‘1 -------------------- - 0.0
300 400 500 600 700 800
Wavelength (nm)

 

 

Figure 3-50: UV-vis absorption and emission for Cr(DBM)3 acquired in

dichloromethane solution and a low-temperature optical glass of 2-methyl-

tetrahydrofuran.

Transient absorption spectroscopy reveals that a ground state recovery
lifetime for Cr(DBM)3 3-6 times as long as that of Cr(acac)3 (approximately 3.5 ns
versus approximately 0.7 ns). Clearly better data must be acquired in order to
better compare the ground state recovery times for these three molecules, but these
time constants fall within the “dark” range of our current laser lab setup, as

discussed above. Shorter timescale dynamics are shown in Figures 3-51 and 3-52.

238

Full excited state transient absorption spectra, Figure 3-51, clearly show two

different absorption bands.

 

 

 

 

 

 

 

 

 

0.10—

09

o

8 0.08—

e

8

0 0.04—

0)

8’

cu 0.02~

5 Af/
0'00 I I— I I I I

500 550 600 650 700 750
Wavelength (nm)

 

 

 

Figure 3-51: Full spectra of Cr(dbm)3 in dichloromethane solution after
excitation at 515 nm. The red spectrum near the baseline is at negative time, and
the other spectra are spaced by 30 fs, with red being the earliest and blue the latest,
starting at 0.1 ps and ending at 1.1 ps. Similar full spectra were observed for other

excitation wavelengths across the 4A2 —> 4T2 absorption.

The bluer band, centered around 500 nm, decays very quickly while retaining its
shape. A second band, centered at 700 nm grows in over a much longer timescale.
Single-wavelength kinetics conﬁrm this picture, Figure 3-52: the data reveals
instrument-limited formation dynamics for probes bluer than ~550 nm and a ~1.3

ps rise in the excited state absorption for redder probes. The lifetime of the rise of

239

this redder component was found to be consistent for all excitation wavelengths
across the 4T2 absorption, Figure 3-53.

The full spectra dynamics share many of the characteristics of the other
members of this series, namely a feature that decays and gives rise to a redder
feature that grows in on a longer timescale than the ﬁrst component. Given the
results presented above for other members of this series the ﬁll] spectra dynamics
are remarkable only in the fact that the timescales are so different for the evolution
of the two features. One possible explanation of the slower component is that it is
due to rotation of the phenyl rings. To explore this possibility, single wavelength
kinetic traces at 530 nm pump and 700 nm probe were collected in two different
solvents: dichloromethane (p. = 0.4) and the more viscous N-methylpyrollidinone
(1.1 = 1.4). If the phenyl rings were rotating, then presumably the more viscous
solvent would inhibit rotation and slow down this process. While the data for N-
methylpyrollidinone did reveal a slightly slower process, the ﬁts were not outside
the margin of error for the measurement. Regardless of interpretation, this
complex presents the clearest example from this series of excited state absorption

from two different excited states.

240

 

 

 

 

 

 

 

 

 

 

6 _j f r T T j 1 I I _4 4 I I T
i & '
5 - - i 500 nm probe . ~
l ' 3 ~
(6.1 4 1 i i - 700 nm probe ‘
'o ’ 1
1): 3- i i - 2- 2 1121291004”-
g 2 _ . i. Sample a _ l I ‘
L .131. N 1 _ _
1 1- L7,: - t
:a 9» CHZCIZ -
0 AA ...... %Al: 13; i; -: if ‘ .............. H 0 ..."... ..................................................... _.
—0.2‘0.0‘0.2‘0.4‘0.6081.0 0 A 6 T 10
Time (ps) TIme (PS)

 

 

 

Figure 3-52: Representative single wavelength kinetic traces for Cr(dbm)3 in

dichloromethane for excitation at 610 nm.

 

 

 

 

 

1.5.L , i l
12; A j 1 § 1
7’; 0.9» l
E; 05; 700nm probe j
P .
O.3~ -
00 1 1 1 1 1 1 1
16 17 18 19

Pump Energy (cm'1 x 103)

 

 

 

Figure 3-53: Lifetime values for various excitation energies across the 4T2

absorption for 700 nm probe.

241

Cr(hfac)3. This complex forms green powders and is quite difﬁcult to crystallize
due to its high solubility in most common solvents. The green powder will
decompose in air over the course of several weeks to give a colorless crystalline
material (presumably the ﬁ'ee ligand). The UV-Vis absorption spectrum is shown

in Figure 3-54.

 

 

_\
N
I
l

_s
O

a (x103 M'1 cm")
0)

 

#—

 

 

 

 

 

‘ 300 l 400 ‘ 500 I 600 ' 700% 800
Wavelength (nm)

 

 

 

Figure 3-54: UV-vis absorption spectrum for Cr(hfac)3 in dichloromethane.

242

As expected, it conforms with other members of this series with the exception that

most peaks are red-shifted relative to those of Cr(acac); due to the electron
withdrawing ability of the —CF3 substituent. In the ultraviolet one observes two

peaks at 273 nm (a ~ 8420 M" em") and 291(s ~ 8290 M" em") assigned as

ligand-based transitions and a more intense peak at 360 nm (8 ~ 9460 M'1 cm'l)
assigned as charge-transfer in nature. A broad, low-intensity band at 567 nm (61
M'1 cm!) is due to a transition to the 4T2 ligand ﬁeld state. The emission
maximum for this complex, at 12300 cm", has been reported before.76 We were
unable to observe any emission from this complex in an 80 K optical glass of 2-
methyltetrahydroﬁrran.

Preliminary ultrafast measurements showed complex kinetics at early
times, while long-time scans revealed a baseline offset aﬁer an ~ 400 ps
component, similar to what was observed for Cr(3-Methylac)3 in the previous
series. Furthermore, samples which underwent ultrafast experiments were shown
to have a different absorption spectrum (as well as a colorless microcrystalline
precipitate in some cases), indicative of photochemical decomposition to produce
the free ligand. Therefore, future experiments require ﬂowing of the sample to
reduce this photochemical damage. The presence of free ligand suggests some
source of protons in solution which may be due to some acidic impurities in the
dichloromethane solvent—future experiments using acetonitrile may avoid this

problem.

243

3.5 Comparison of Complexes. The initial goal of this research was to quantify
whether nonradiative decay theory was applicable to the ultrafast events occurring
after excitation of these systems. This quantitative aspect has proven quite
difﬁcult to implement, due to a number of chemical and technological hurdles,
namely 1) the emission spectra in many cases are too red to be reliable given the
detection limit of our instrument; 2) the energy gaps for these complexes are not
that disparate; 3) as shown for Cr(acac)3, the 4T2 state is likely not thermalizing
prior to intersystem crossing, making determinations of Franck-Condon overlap
speculative at best; 4) for the 3-substituted Cr(acac)3 series chemical modiﬁcation
did very little to change the dynamics, and in all cases the 4T2 —> 2E intersystem

crossing event was not observed for the ~ 100 fs pulses used. The result is that

qualitative differences dominate the discussion.
First the ground state recovery dynamics, i.e. those corresponding to the 2E
—+ 4A2 transition, are considered. Given the emission data presented above, one

would expect from nonradiative decay theory, considering only a two-state system
at this point, that the ground state recovery time constants would vary

systematically from Cr(dbm)3 (being the fastest) to Cr(acac)3, the slowest: for
nested potential energy surfaces, such as the 4A2 ground state and lowest energy
excited state 2E, a lowered zero point energy corresponds to greater vibrational

overlap of the component electronic potentials. In fact, from the data below, the

244

opposite trend appears to hold. It is possible that the breadth of emission data for
Cr(dbm)3 shows that the potential is not as nested, but breadth is likely due to an
extremely low response past 825 nm for the photomultiplier tube that is employed
in our ﬂuorescence experiment. In fact, as described in Chapter 1, quenching of

the 2E state can occur through a variety of mechanisms which are thermally

72,80,81

activated. Therefore, at ambient temperature one can expect not only a

decay component corresponding to nonradiative decay from the low-temperature
regime (dependent on factors such as potential well displacement and relative
energy which affect vibrational overlap), but also thermally-induced quenching
events (BISC, quenching via reaction, etc.). The fact that these events occur on a
timescale several times greater than the capabilities of the instrument also leads to
quite a bit of uncertainty: the dynamics were ﬁt with monoexponential decays, but
this model can not be conﬁrmed until the timescale capabilities of the instrument
are increased. Once this is achieved, the best way to parse out the low-
temperature dynamics from those of the possible thermally-activated pathways,
and to test the applicability of nonradiative decay theory to this ground state
recovery event, would be to carry out the experiment at low-temperature.

The ultrafast dynamics present an equal challenge. The 3-substituted
Cr(acac)3 series was originally chosen for comparison to the previous work on
Cr(acac); not only for potential ease of synthesis, but because it was speculated

that substitution at the 3-position would provide the most direct access to the inner

245

rt-system of the acetylacetonate ligand, thus modifying the electronic structure and
dynamics most dramatically. Clearly, this assumption proved to be incorrect: the
dynamics observed for the 3-substituted series are qualitatively all very similar,

lacking any biphasic kinetics, and are therefore given the same interpretation, i.e.

that of Cr(acac)3. Coherence data of Cr(acac)3 collected with ~ 50 fs optical

pulses provides some insight into why this may be. An oscillation of 164 cm],

assigned as vibrational coherence, was observed. This frequency may correlate to
a metal-ligand active vibrational mode observed in the ground state infrared
spectrum but in the egI-populated 4T2 sate. DFT frequency calculations of
Cr(acac)3 indicate that this mode, while M-L active, also includes a large
amplitude motion of the acetylacetonate methyl groups. If indeed this mode is
active in facilitating ultrafast intersystem crossing in the complex, then
perturbation to the ligand structure at the 3-position (especially the sterically small
substituents that were chosen for this study) is likely to not drastically change this
mode. Assuming that electronic differences normally considered for nonradiative
decay theory (E0 and AQ) are negligible for this series, the intersystem crossing
rate is expected to be same across the 3-substituted series based solely on the
promoting ability of the ligand backbone vibrational modes.

There are clear differences between the 3-substituted and the substituted
propanedione series: 1) different dynamics are observed for every member of this

series, despite electronic similarities, implying a role for the ligand structure in at

246

v 119—W

the 1,3-positions in nonradiative dynamics; 2) transient spectra from different
electronic states are observed; 3) there are varying timescales of dynamics across
the series. Given the previous results for the 3-substituted series, these differences
are not surprising, and in fact reinforce the interpretation of the Cr(acac)3
coherence data. The fact that biphasic kinetics are observed for both large (R =
cyclohexyl) and small (R = H) ligands reveals that the ultrafast kinetics are not
simply a matter of electronics, but also of vibrational motion of the molecule on
the timescale of the observed dynamics.

Several points must be made about the ﬁtting procedure employed
throughout this chapter. The single wavelength kinetic data were ﬁt with either a

mono- or biexponential function with a baseline offset, i.e
y = yo + A] exp(-%l)+ A,exp(—%2) (3.1)
For example, A1 “IX-%l) may correspond to depopulation of the 4T2 state (and

concomitant population of the 2E state) while AzeXP(‘%2)corresponds to

vibrational cooling the in 2E state. This model likely does not represent a correct

physical picture of the dynamics, as this ﬁtting model has both kinetic components
with amplitude at time zero, implying that both physical processes are initiated at
time zero, immediately following excitation. Assuming that the observed
dynamics are due to intersystem crossing followed by vibrational cooling in the 2E

state, a correct model of state evolution would involve l) a fast exponential decay

247

component with amplitude at time zero which decays to baseline, corresponding to
depopulation of the 4T2 state (this assumes that there is no thermalization of 4T2
prior to intersystem crossing); 2). a second component on the same timescale as the
ﬁrst, which is delayed and has no amplitude at time zero, and grows in
corresponding to population of the 2E; 3) a third component with a different
timescale than 1) and 2) corresponding to vibrational cooling of the 2E. A similar
model has been developed by Mathies and coworkers82 to account for the ultrafast
initial dynamics of ruthenium(II) polypyridyl chromophores, where the initial
1MLCT state passes through a second CT state before the lowest-energy 3MLCT is
formed (equation 3.2). From this kinetic picture the authors developed equation

3.3 which describes the population of the 3MLCT.

1‘1 k2
'MLCT —> X -) 3mm (3.2)
[3MLCT] = [IMLCT]O{1 — (k2 — kl)" [k2 exp(—k1t)— k1 exp(k2t)]} (3.3)

This basic picture can be applied to the two-state evolution of this system wherein

the intermediate state X and 3MLCT are replaced by initially populated and

thermalized vibrational levels of the 2E state, i.e.

4 leC 2 [WC 2
T 2 _> E0 _) Etherm

(3.4)

In this equation 2E0 represents the initially populated vibrational level (or, more

likely, levels) of the 2E electronic state while 2Ether,“ represents the thermalized 213

248

state. Substitution of the corresponding terms for this system to equation 3.3
results in equation 3.5. Of course, correct ﬁtting of the data requires convolution

of this quIation with an instrument response function.
[zEtherm] = A1 {1 “ (ch ‘“ kISC)_1[kVC exp(_k1SCt) ' kISC exp(kVCt )1} (35)
In this equation A] is essentially the long-time baseline offset of the kinetic trace.
While equation 3.5 accounts for the 2E population, the full evolution of the

dynamics can be described by including a second monoexponential term to

account for depopulation of the 4T2 state, and a third term corresponding to the

concentration of the “intermediate” 2E0 state, equation 3.6.
y = [RF 8 ([472] + [2150] + [ZEThermD (3.6)

Attempts are currently underway to implement this new ﬁtting model. One
drawback of this model is that is does not account for back intersystem crossing

from the 2E state to the 4T2 state, which Kunttu and coworkers have shown for
Cr(acac)3 to be important process (at least for high-energy excitations).83 This
model is appropriate for all systems, but necessary for those in which the RI and k2
are very different (Cr(dbm)3 and Cr(prop)3, as well as every member of 3-
substitued Cr(acac)3 series). For systems in which the intersystem crossing
dynamics are occurring on the same timescale as thermalization of the 2E state
(e.g. Cr(tbutylac)3), the [2E0] concentration can be assumed to be a steady-state.

Cleary, this model is not perfect for all systems, but represent a much more

249

physically sound model of intersystem crossing dynamics than a simple

biexponential function.

3.6 Final Remarks

The work presented in this chapter represents the initial efforts toward uncovering
mechanisms of ultrafast nonradiative decay in transition metal systems.
Qualitative trends were observed which seem to be indicating that certain low-
frequency metal-ligand active modes are facilitating the ultrafast intersystem
crossing event. Future work, described in Chapter 5, will beneﬁt from increased
spectroscopic capabilities, such as shorter pulses which will potentially enable
further coherence measurements and present a more quantitative analysis of the

series.

250

3.7 References

(1)
(2)

(3)
(4)

(5)
(6)
(7)
(8)

(9)

(10)

(11)

(12)

(13)

(14)

Meyer, T. J. Accounts of Chemical Research 1989, 22, 163-170.

Alstrum-Acevedo, J. H.; Brennaman, M. K.; Meyer, T. J. Inorganic
Chemistry 2005, 44, 6802-6827.

Forster, L. S. Chemical Reviews 1990, 90, 331-353.

Juban, E.; McCusker, J. Journal of the American Chemical Society 2005,
127, 6857—6865.

Zewail, A. H. Journal Of Physical Chemistry A 2000, 104, 5660-5694.
Parker, C. A.; Rees, W. T. Analyst 1960, 85, 587-600.
Juban, E., Ph.D. Thesis, University of California, 2006.

Fragnito, H. L.; Bigot, J. Y.; Becker, P. C.; Shank, C. V. Chemical Physics
Letters 1989, 160, 101-104.

Dexheimer, S. L.; Wang, Q.; Peteanu, L. A.; Pollard, W. T.; Mathies, R. A.;
Shank, C. V. Chemical Physics Letters 1992, 188, 61-66.

Schoenlein, R. W.; Peteanu, L. A.; Mathies, R. A.; Shank, C. V. Science
1991, 254, 412-415.

Zhu, L.; Sage, J. T.; Champion, P. M. Science 1994, 266, 629-632.

Rosca, F.; Kumar, A. T. N.; D., 1.; Ye, X.; Demidov, A. A.; Sjodin, T.;
Wharton, D.; Barrick, D.; Sligar, S. G.; Yonetani, T.; Champion, P. M.
Journal of Physical Chemistry A 2002, 106, 3540-3552.

Dillman, K. L.; Shelly, K. R.; Beck, W. F. Journal Of Physical Chemistry B
2009, 113, 6127-6139.

M. J. Frisch, G. W. T., H. B. Schlegel, G. E. Scuseria; M. A. Robb, J. R. C.,
J. A. Montgomery, Jr., T. Vreven; K. N. Kudin, J. C. B., J. M. Millam, S.
S. Iyengar, J. Tomasi; V. Barone, B. M., M. Cossi, G. Scalmani, N. Rega;
G. A. Petersson, H. N., M. Hada, M. Ehara, K. Toyota; R. Fukuda, J. H.,
M. Ishida, T. Nakajima, Y. Honda, 0. Kitao; H. Nakai, M. K., X. Li, J. E.
Knox, H. P. Hratchian, J. B. Cross; V. Bakken, C. A., J. Jaramillo, R.
Gomperts, R. E. Stratmann; O. Yazyev, A. J. A., R. Cammi, C. Pomelli, J.

251

(15)
(16)
(17)

(18)
(19)

(20)

(21)

(22)

(23)

(24)
(25)

(26)

(27)

(28)

W. Ochterski; P. Y. Ayala, K. M., G. A. Voth, P. Salvador, J. J.
Dannenberg; V. G. Zakrzewski, S. D., A. D. Daniels, M. C. Strain; O.
Farkas, D. K. M., A. D. Rabuck, K. Raghavachari; J. B. F oresman, J. V. O.,
Q.Cui, A. G. Baboul, S. Clifford; J. Cioslowski, B. B. S., G. Liu, A.
Liashenko, P. Piskorz; I. Komaromi, R. L. M., D. J. Fox, T. Keith, M. A.
Al-Laham; C. Y. Peng, A. N., M. Challacombe, P. M. W. Gill; B. Johnson,
W. C., M. W. Wong, C. Gonzalez, and J. A. Pople; Revision D.01 ed.;
Gaussian, Inc.: Wallingford CT, 2004.

Gaussview 2.1; Gaussian, Inc.: Pittsburgh, PA, 2001.
F emelius, W. C.; Blanch, J. E. Inorganic Syntheses 1958, 5, 130-131.

Nakamura, K.; Murakami, Y. Technology Reports of the Kyushu University
1965, 38, 329-336.

Harada, Y.; Girolami, G. S. Polyhedron 2007, 26, 1758-1762.

Bradley, D. C.; Hursthouse, M. B.; Newing, C. W.; Welch, A. J. Journal of
the Chemical Society, Chemical Communications 1972, 567-568.

Collman, J. R; Marshall, R. L.; William L. Young, 1.; Goldby, S. D.

Inorganic Chemistry 1962, 1, 704-710.

Collman, J. B; Moss, R. A.; Maltz, H.; Heindel, C. C. Journal of the
American Chemical Society 1960, 83, 531-534.

Kluiber, R. W. Journal of the American Chemical Society 1961, 83, 3030-
3032.

Grant, M. S.; Synder, H. R. Journal of the American Chemical Society
1960, 82.

Kuhr, M.; Bock, B.; Musso, H. Chemische Berichte 1976, 109, 1195-1203.
Inorganic Syntheses 1966, 8, 141.

Angela, M.; Jean-Pierre, G.; Se’bastien, J.; Delphine, B.; Virginie, R.-V.
Chemistry - A European Journal 1999, 5, 1160-1165.

Charles, R. G. Inorganic Syntheses, 8, 135-138.

Charles, R. Inorganic Syntheses, 8, 138-140.

252

(29)

(30)

(31)
(32)

(33)

(34)

(35)

(36)

(37)

(38)

(39)

(40)

(41)

(42)

(43)

Barnum, D. W. Journal of Inorganic and Nuclear Chemistry 1961, 21, 221-
237.

Fleming, C. A.; Thornton, D. A. Journal of .Molecular Structure 1975, 25,
271- 279.

Fatta, A. M.; Lintvedt, R. L. Inorganic Chemistry 1971, 10, 478-481.

Ohno, T.; Kato, S.; Kaizaki, S.; Hanazaki, I. Inorganic Chemistry 1986, 25,
3853- 3858.

Konig, E.; Herzog, S. Journal of Inorganic and Nuclear Chemistry 1970,
32, 585-599.

J uban, E. A., PhD. Thesis, University of California, Berkeley, 2006.

Serpone, N.; Jamieson, M. A.; Hoffman, M. Z.; Bolletaa, F.; Maestri, M.
Journal of the American Chemical Society 1979, 101, 2907-2916.

Hanazaki, I.; Hanazaki, F.; Nagakura, S. Journal of Chemical Physics
1969, 50, 265-276.

Hanazaki, 1.; Hanazaki, F.; Nagakura, S. Journal of Chemical Physics
1969, 50, 276-286.

Hanazaki, 1.; Nagakura, S. Inorganic Chemistry 1969, 8, 648-654.

Macoas, E. M. S.; Kananavicius, R.; Myllyperkio, P.; Pettersson, M.;
Kunttu, H. Journal of the American Chemical Society 2007, 129, 8934-
8935.

Shelly, K. R.; Golovich, E. C.; Dillman, K. L.; Beck, W. F. Journal of
Physical Chemistry B 2008, 1 12, 1299-1307.

Dillman, K. L. Journal of Physical Chemistry B 2009, I 13, 6127-6139.

Yoon, M.; Jeong, D.; Cho, S.; Kim, D.; Rhee, H.; Joo, T. Journal of
Chemical Physics 2003, 118, 164.

Gruia, F.; Ionascu, D.; Kubo, M.; Ye, X.; Dawson, J.; Osborne, R. L.;

Sligar, S. G.; Denisov, 1.; Das, A.; Poulos, T. L.; Temer, J .; Champion, P.
M. Biochemistry 2008, 4 7, 5156-5167.

253

(44)

(45)

(46)

(47)

(48)

(49)

(50)

(51)

(52)

(53)

(54)

(55)

(56)

(57)

Delﬁno, I.; Manzoni, C.; Sato, K.; Dennison, C.; Cerullo, G.; Cannistraro,
S. Journal of Physical Chemistry, B 2006, 110, 17252-17259.

Marin, T. W.; Homoelle, B. J.; Spears, K. G.; Hupp, J. T. Journal of
Physical ChemistryA 2002, 106, 1131-1143.

Reid, P. J.; Silva, C.; Barbara, P. F.; Karki, L.; Hupp, J. T. Journal of
Physical Chemistry 1995, 99, 2609-2616.

Consani, C.; Fremont-Schwarz, M.; ElNahhas, A.; Bressler, C.; van
Mourik, F.; Cannizzo, A.; Chergui, A. Angewandte Chemie, International
Edition 2009, 48, 7184-7187.

Pollard, W. T.; Dexheimer, S. L.; Wang, Q.; Peteanu, L. A.; Shank, C. V.;
Mathies, R. A. Journal of Physical Chemistry 1992, 96, 6147-6158.

Bardeen, C. J. Wang, Q.; Shank, C. V. Journal of Physical Chemistry A
1998, 102, 2759- 2766.

Champion, P. M.; Rosca, F.; Ionascu, D.; Cao, W.; Ye, X. Faraday
Discussions 2004, 127, 123-135.

Trushin, S. A.; Fuss, W.; Schmid, W. E.; Kompa, K. L. Journal of Physical
ChemistryA 1998, 102, 4129-4137.

Banin, U.; Waldmann, A.; Ruhmann, S. Journal of Chemical Physics 1992,
96, 2416.

Pugliano, N.; Palit, D. K.; Szarka, A. Z.; Hochstrasser, R. M. Journal of
Chemical Physics 1993, 99, 7273.

Rubstov, I. V.; Yoshihara, K. Journal of the Chinese Chemical Society
2000, 4 7, 673-677.

Chudoba, C.; Riedle, E.; Pfeiffer, M.; Elsaesser, T. Chemical Physics
Letters 1996, 263, 622-628.

Cheng, Y.; Fleming, G. R. Annual Review of Physical Chemistry 2009, 60,
241-262.

Engel, G. S.; Calhoun, T. R.; Read, E. L.; Ahn, T.; Mancal, T.; Cheng, Y.;
Blankenship, R. E.; Fleming, G. R. Nature 2007, 446, 782-786.

254

(58)

(59)

(60)

(61)

(62)

(63)

(64)

(65)

(66)

(67)

(68)

(69)

(70)

(71)

(72)

Lee, H.; Cheng, Y.; Fleming, G. R. Science 2007, 316, 1462-1465.

Fuji, T.; Jin Ong, H.; Kobayashi, T. Chemical Physics Letters 2003, 380,
135-140.

Takeuchi, S.; Tahara, T. Journal of Physical Chemistry A 2005, 109,
10119-10207.

Kobayashi, T.; Saito, T.; Ohtani, H. Nature 2001, 414, 531-534.

Lenderink, E.; Duppen, K.; Wiersma, D. A. Journal of Physical Chemistry
1995, 99, 8972-8977.

Nakamoto, K.; McCarthy, P. J.; Martell, A. E. Journal of the American
Chemical Society 1961, 83, 1272-1276.

Dismukes, J. P.; Jones, L. H.; Bailar, J. C. Journal of Physical Chemistry
1961, 65, 792-795.

Dial-Acosta, 1.; Baker, J.; Cordes, W.; Hinton, J. F.; Pulay, P.
Spectochimica Acta, Part A 2003, 59, 363-377.

Dial-Acosta, 1.; Baker, J.; Cordes, W.; Pulay, P. Journal of Physical
Chemistry, A 2001, 105, 238.

Tayyari, S. F.; Milani-nejad, F. Spectochimica Acta, PartA 2000, 56, 2679.

Tayyari, S. F .; Raissi, H.; Ahmaabadi, Z. Spectochimica Acta, Part A 2002,
58, 2669.

Thornton, D. A. Coordination Chemistry Reviews 1990, 104, 173-249.

Sato, H.; Taniguchi, T.; Nakahashi, A.; Monde, K.; Yamagishi, A.
Inorganic Chemistry 2007, 46, 6755-6766.

Nakamoto, K.; Udovich, C.; Takemoto, J. Journal of the American
Chemical Society 1970, 92, 3973-3976.

Perkovic, M. W.; Endicott, J. F. The Journal of Physical Chemistry 1990,
94, 1217.

255

(73)

(74)

(75)

(76)
(77)
(78)
(79)

(80)

(81)

(82)

(83)

Hare, P. M.; Crespo-Hemandez, C. E.; Kohler, B. Proceedings Of The
National Academy Of Sciences Of The United States Of America 2007, 104,
435-440.

Kang, H.; Jung, B.; Kim, S. K. Journal Of Chemical Physics 2003, 118,
6717- 6719.

Worth, G. A.; Cederbaum, L. S. Annual Review of Physical Chemistry
2004, 55, 127-158.

DeArmond, K.; Forster, L. S. Spectrochimica Acta 1963, 19, 1403-1406.
Riesen, H.; Dubicki, L. Journal of Physical Chemistry A 2008, ASAP.
DeArmond, K.; Forster, L. S. Spectrochimica Acta 1963, 19, 1687-1693.
Kirk, A. D. Chemical Reviews 1999, 99, 1607-1640.

Endicott, J. F.; Lessard, R. B.; Lynch, D.; Perkovic, M. W.; Ryu, C. K.
Coordination Chemistry Reviews 1990, 97, 65-79.

Endicott, J. F.; Tamilarasan, R.; Lessard, R. B. Chemical Physics Letters
1984, 112, 381-386.

Yoon, S.; Kukura, F.; Stuart, C. M.; Mathies, R. A. Molecular Physics
2006, 104, 1275-1282.

Macoas, E. M. S.; Kananavicius, R.; Myllyperkio, P.; Pettersson, M.;
Kunttu, H. Journal of the American Chemical Society 2007, 129, 8934-
8935.

256

Chapter 4: Electronic Structure and Nonradiative Dynamics of Heisenberg

Spin Exchange Complexes of Chromium(III)

4.1 Introduction

As described previously, the main theme of the research presented in this
dissertation is correlation of the electronic structure of transition metal complexes
to nonradiative dynamics in these systems. The previous chapters have been
devoted to exploring the electronic structure and nonradiative dynamics in quartet
Chromium(III) complexes, as well as the ground state electronic and magnetic
structure of gallium(III) semiquinone complexes (gallium(III) has an almost
identical charge-to-radius ratio as chromium(III), and was therefore chosen as a
diamagnetic analog of Chromium(III)). We now hope to combine the knowledge
of the previous chapters to be applied to the study of the magnetic and electronic
structures and nonradiative dynamics of spin-exchange complexes of
chromium(III), i.e. complexes with two or more interacting spin centers. This
chapter represents the initial efforts toward the synthesis and characterization of a
series of chromium-semiquinone complexes. This series will be employed to
study the effects of Heisenberg spin exchange interactions on the photophysical
properties of a transition metal system. These systems are of interest because
Heisenberg spin exchange is operative not only in the ground state (although most
examples in the literature only study the phenomena in the ground state), but also

in excited states, which has been shown conclusively by Gﬁdel and coworkers."4

257

It is thus apparent that introduction of spin exchange can tremendously affect the
photophysical as well as photochemical properties of molecular systems. Metal-
quinone complexes are ideal for the study of physicochemical properties of
exchange-coupled molecules because the redox activity of the quinone ligand
provides a facile mechanism for turning the exchange interaction on or off, i.e. in
the systems of study in this chapter one is able to selectively turn on spin exchange
with incorporation of the paramagnetic species Chromium(III) and semiquinone, or
turn off exchange interactions by substituting in the diamagnetic analogs
gallium(III) and catechol. In this manner a series of controls can be established,
which allows one to identify and differentiate properties endemic to the
constituents of the molecule and those which arise due to exchange interactions
between the constituents.

The main impetus for the dynamical study of these systems arose from a
kinetic study of [Cr(tren)(3,6-DTBCat)]+ (where DTBCat = 3,6-di-tert-butyl-l,2-
orthocatechol and tren is tris(2-aminoethyl) amine) and its semiquinone analog,

[Cr(tren)(3,6-DTBSQ)]2+, (vide infra). These data show the effect of spin

exchange on ground-state recovery dynamics in this system: the semiquinone and
catechol analogs return to baseline, indicating ground state recovery within ~3 ps
and ~6 ps, respectively. This is compared to the approximately 500 ps lifetime

observed for 2E -—> 4A2 relaxation in a low-symmetry analog, [Cr(tren)(acac)]2+.

Models for these dynamics are presented below, but it seems apparent from a ﬁrst

258

glance that inductive effects and spin exchange are playing a role in modifying
these dynamics from those of the related low-symmetry complex
[Cr(tren)(acac)]2+.

This chapter will focus primarily on the synthesis and characterization of
chromium(III)-catechol and Chromium(III)-semiquinone complexes. A series of
Chromium(III)-catechols (from which one can prepare the chromium(III)-
semiquinone complexes via a one electron oxidation) and was produced in order
to study the effects of modifying spin distribution in these systems. This series
utilizes the [Cr(tren)(3,6-R-Q)]+”+2 motif, where 3,6-R-Q represents an ortho-
quinone substituted at the three and six positions and either in the catechol (-2) or

semiquinone (-1) form (Figures 4-1 and 4-2).

 

 

 

 

Quinone Semiquinone Catechol
O + e' O‘
- e'
I O“ O“
O
O‘

 

 

 

Figure 4—1: The redox states of the ortho-quinone ligands.

259

“......“

 

— —+1/+2

NH2
0
d7l\< R: H, CN, N02, NH2
QB °

NH2

 

 

 

 

 

Figure 4-2: Members of the [Cr(tren)(3,6-R-Q)]+U+2 series employed in this

study.

While this chapter will focus primarily on Chromium(III) semiquinone complexes,
metal-metal as well as other spin-exchange complexes will be explored, such as
the spin-exchange complexes of the aforementioned [M1M2(tren)2(CA”')]m+ series,
where M = gallium(III) or chromium(III) and CA” is the chloranilate anion, the
bridging chelate between the two metal ions (Figure 4-3). This series was
introduced in Chapter 2, which explored the ground states of some of the
(magnetically) simpler members of this series via electron spin resonance
techniques: the doublet system 2 and the quartet system 3. While we have
characterized the simpler, “magnetically dilute” paramagnetic systems, i.e. those
that display no exchange coupling}6 we have yet to explore the effects of “turning

99

on exchange coupling within this series, namely

260

 

 

 

 

M, = M2 = Ga(III)

\
,Mzaren) M, = Ga(III), M2 = Cr(III)

tren =

M, = M2 = Cr(III)

 

3+
M, = M2 = Ga(III)

2(tren) M, = Ga(III), M2 = Cr(III)

M, = M2 = Cr(III)

 

{\NHZ
N/NNHZ

<\NH2

(1)
(3)
(5)

(2)
(4)
(6)

 

 

Figure 4-3: Members of the bimetallic [M1M2(tren)2(CA"')]m+ series.

lGaCr(tren)2(CAsq’°at)](BPh4)2(BF4) (4), [Cr2(tren)2(CA°m’Cm)](BPh4)2

261

(5), and
[Cr2(tren)2(CAsq’°at)](BPh4)2(BF4) (6). In this chapter, X-Band EPR spectroscopy
is the main tool utilized to study these magnetically complex ground states. An
approached of “building up” the physical interactions in these systems in a
piecewise fashion is applied: the physical properties of semiquinone complexes
and quartet complexes of Chromium(III) were extensively studied before moving

on to the study of spin-exchange systems. As such, much of the work presented in

this chapter remains incomplete; a section of chapter 5 is devoted to the future
study of these systems. EPR has been implemented to characterize the ground
states of magnetic systems similar to those of molecules within this series that
display exchange coupling: 1.) exchange-coupled metal-semiquinone complexes

including binuclear Crm-semiquinone species7’8 (like complexes 4 and 6) and

III 3. 10,

e "I o 0 ° 9 II ° '
binuclear Fe - and Co -sem1qu1none speCIes and Cu -sem1qumones , 2)

Ill “-13

exchange coupled metal centers including Cr dimers (like complexes 5 and

6). This fund of knowledge provides a useful starting point for the analyses of

these complexes.

4.2 Experimental Section

4.2.1 Synthetic Methods.

General. All reagents and materials were used as received unless otherwise
noted. Solvents were purchased from Aldrich Chemical Co. and distilled and
degassed by the freeze-pump-thaw method. The ligand tris(2-aminoethyl)-amine
(tren) was purchased from Aldrich and vacuum-distilled prior to use. All synthetic
procedures involving hydrochloranilic acid were performed under an inert
atmosphere. Hydrochloranilic acid was synthesized by a modiﬁcation of a

14.15

previously reported method. Ga(tren)(NO3)3 was prepared according to our

previous paper.16 [Cr(tren)Clz]Cl was prepared according to literature

262

procedures.l7 Synthesis of the bimetallic complexes 4 - 6 and the model complex

of 4 were synthesized according to previously published proceduresm’18

[Cr(tren)(pyrocatecholate)]BPha. This complex was made by adapting a
procedure which had been utilized previously to the synthesis of chromium(III)-
phenanthrenecatechol and —semiquinone systems.19 130 mg (0.43 mmol) of
[Cr(tren)C12]Cl is dissolved in 60 ml of 2:1 methanolzwater in a 2 neck 250 m1
roundbottom ﬂask equipped with a reﬂux condenser. This solution is rapidly
stirred and bubble de-gassed with N2 for 30 minutes. Meanwhile, a 20 ml
solution of 53 mg pyrocatechol (0.48 mmol) and 38 mg sodium hydroxide (0.96
mmol) is prepared under inert conditions in a 50 ml roundbottom ﬂask equipped
with a vacuum adaptor and septum. Note: if this solution is exposed to oxygen it
will turn green in a matter of minutes (semiquinone form) and eventually brown
(quinone). This solution is allowed to stir for approximately ten minutes and is
then cannula transferred to the pink solution of [Cr(tren)Clz]Cl. The solution
rapidly turns purple. The reaction is reﬂuxed under positive N2 pressure for
approximately 16 hours, over which time it turns a dark blue/green color. The
solution is ﬁltered under N2 onto 1.3 g (excess) NaBPh4, and nitrogen is allowed

to pass over the solution until it cools and most of the methanol has evaporated.
Blue/green microcrystals precipitate, which are collected on a frit and washed with

copious amounts of water. Yield: 160 mg (59.4%) Anal. Calc. for

263

C36H4zBCrN4Oz- 3 H20 expected: C, 63.62; H, 7.12; N, 8.24. Found: C, 63.50;
H, 6.20; N, 8.19. Mass spectroscopy (ESI, positive): m/z 306.1 (M-BPh4, 100%).
Crystals suitable for x-ray crystallography were grown by slow evaporation of a
methanol/NaBPh4 solution. The product appears to be stable in air as a solid and

in solution.

2,3-dihydroxyterephthalonitrile (3,6-cyano-l,2-orthocatechol). This catechol
was prepared according to a previously published method.20 Anal. Calc. for
C3H4N202- 0.2 H20 expected: C, 58.68; H, 2.71; N, 17.11. Found: C, 58.28; H,

2.68; N, 16.32. NMR(500 MHz, d6-DMSO): 7.19 (5), 10.89 (broad, -OH).

[Cr(tren)(3,6-CN-1,2-catecholate)]BPh4. This complex was prepared in a
similar fashion to [Cr(tren)(pyrocatecholate)]BPh4. 65 mg (0.22 mmol) of
[Cr(tren)Clz]Cl was dissolved in 30 ml of 1:1 MeOHszO and degassed with N2

for approximately 30 minutes. A 10 ml methanol solution containing 38.4 mg
(0.24 mmol) 3,6-cyano—l,2-orthocatechol and 19.0 mg (0.48 mmol) NaOH was
prepared under inert conditions. This dark orange solution was cannula
transferred to the pink [Cr(tren)C12]Cl solution, which instantly turns a dark
purple/brown color. After reﬂuxing for 30 minutes the solution turns orange, and
reﬂuxing is continued for another day. The hot solution is Schlenk ﬁltered under

N2 onto solid NaBPh4 and nitrogen was allowed to blow over the hot solution to

264

drive off methanol. Brown microcrystals precipitate which were washed with
water, dried in vacuo, and stored under N2, although the product does not appear
to be air sensitive. Anal. Calc. for C33H40BCrN602 expected: C, 67.56; H, 5.97; N,
12.44. Found: C, 67.89; H, 5.93; N, 12.30. Mass spectroscopy (ESI, positive):
m/z 356.1 (M-BPh4, 100%). Crystals suitable for x-ray crystallography were
grown by slow evaporation of an acetonitrile/NaBPh4 or methanol/NaBPh4

solution.

2-methoxy-3-nitrophenol (3-nitroguaiacol). This precursor to 2-methoxy-3,6-
dinitrophenol was prepared by a previously published method from ortho-
methoxyacetate (Alfa).21 Anal. Calc. for C7H7NO4 expected: C, 49.71; H, 4.17; N,
8.28. Found: C, 49.57; H, 4.09; N, 8.24. NMR(300 MHz, CDCl3): 3.95 (s, 3H,
methoxy), ~6 (broad, -OH), 7.09 (t, 1 H, aromatic), 7.21 (d, 1H, aromatic), 7.44 (d,

1H, aromatic).

2-methoxy-3,6-dinitrophenol (3,6-dinitroguaiacol). This precursor to 3,6-
dinitro-1,2-orthocatechol was prepared according to a published method from 2-
methoxy-3-nitrophenol.2' The product was extracted from the crude material
using hot petroleum ether. This solution was placed in a freezer and yellow
microcrystalline material precipitated, which was shown by NMR to contain only

the desired product and a small amount of water. NMR (500 MHz, CDCl3): 4.10

265

(s, 3H, methoxy), 7.26 (d, 1H, aromatic), 7.96 (d, 1H, aromatic), 10.86 (s, 1H, -

OH).

3,6-dinitrobenzene-l,2-diol (3,6-dinitro-1,2-orthocatechol). This catechol was
prepared from 2-methoxy-3,6-dinitrophenol using a published method.22

Recrystallization from ethanol resulted in a yellow crystals that were clean by

NMR. NlVIR (500 MHz, CDCl3): 7.71 (s, 2H, aromatic), 10.80 (s, 2H, -OH).

3,6-diaminobenzene-l,2-diol dihydrochloride (3,6-diamino-l,2-
orthocatechol-2H0). 88.3 mg (0.44 mmol) of 3,6-dinitro-1,2-orthocatechol was
dissolved in 50 m1 of concentrated HCl. With rapid stirring, 0.42 g (3.5 mmol) of
tin metal was added slowly over 10 minutes. The reaction was reﬂuxed for
approximately 45 minutes. During the reﬂux, the solution turns from a bright
yellow color to colorless. The volume of the solution is reduced slowly to
approximately 10 ml by blowing nitrogen over it, and a colorless microcrystalline
material precipitates which was collected, washed with ethanol, and dried in
vacuo. Yield: 68.9 mg (73.5%). Anal. Calc. for C6H3NzOz-2HC1 expected: C,
33.82; H, 4.73; N, 13.15. Found: C, 33.61; H, 4.66; N, 12.89. The product is
soluble in water but decomposes in solution over several hours to yield a brown

solution.

266

[Cr(tren)(3,6-dinitro-l,2-orthocatecholate)]BPh... 65 mg (0.22 mmol)
[Cr(tren)Clz]Cl was dissolved into 20 ml methanol and 10 ml water in a 2-neck
100 ml roundbottom ﬂask equipped with a reﬂux condenser. This solution was
bubble de-gassed with N2 for 30 minutes. Meanwhile, a solution of 48 mg (0.24
mmol) 3,6-dinitro-1,2-orthocatechol and 19 mg (0.48 mmol) NaOH in 10 ml

methanol was prepared in the drybox. Addition of NaOH to the yellow catechol
solution turns the solution red/orange. This solution was cannula transferred to the
pink/purple [Cr(tren)C12]Cl solution, which rapidly turns a dark red color. The
solution was reﬂuxed under nitrogen for approximately two days, at which point a
purple precipitate forms and the solution turns dark purple/brown. The reaction
was ﬁltered while hot under N2 onto solid NaBPh4 (0.7 g). Nitrogen was blown
over the solution for several hours to drive off methanol, and a red/brown
microcrystalline material precipitates. This product was ﬁltered and washed with
copious amounts of water. The product was dried in vacuo to yield a burnt orange
powder. Yield: 97 mg (63%). Anal. Calc. for C36H4OBCrN606 expected: C,
59.94; H, 5.86; N, 11.48. Found: C, 60.43; H, 5.63; N, 11.74. Mass spectroscopy

(ESI, positive): m/z 396.1 (M-BPh4, 100%).

[Cr(tren)(3,6-diamino-l,2-orthocatecholate)]BPha. This preparation was
Carried out in a similar manner to those of the other members of this series. A

solution of 86.3 mg (0.292 mmol) of [Cr(tren)C12]Cl in 40 ml methanol and 20 ml

267

water was degassed with nitrogen for approximately 30 minutes. In the drybox a
solution of 62.3 mg (0.292 mmol) of 3,6-diamino—1,2-orthocatechol-2HCI and
46.7 mg of NaOH (1.17 mol, 4 equivalents) in 20 ml of methanol was prepared.
This solution should be colorless. When exposed to air it rapidly turns blue
(semiquinone form) and eventually brown (quinone form). This solution was
removed from the drybox and transferred via cannula to the solution of
[Cr(tren)C12]Cl. The combined reaction solution was reﬂuxed under positive
nitrogen pressure for 24 hours, over which time the solution turns a dark green
color. While hot, the solution is ﬁltered under nitrogen onto solid NaBPh4 (~ 1 g),
and nitrogen is allowed to pass over the solution for several hours. As the
methanol is driven off the product precipitates as a purple solid, which is ﬁltered
and washed with water. The product is stored under inert conditions, and
solutions of the product are air sensitive, eventually decomposing to a brown
solution. Yield: 120 mg (0.183 mmol, 63%). Anal. Calc. for C3gH44BCrN602-05
NaBPh4-2 H20: C, 66.77; H, 6.78; N, 9.74. Found: C, 66.80; H, 6.21; N, 9.56.
Mass spectroscopy (ESI, positive): 336.2 m/z (M-BPh4, 85%), m/z = 656.4 (MHJ',
60%).

The UV-vis absorption and electrochemical properties of this complex (see
below) are not entirely consistent with the rest of the series. These data are

presented below, but until a crystal structure can be obtained these measurements

should be taken with a grain of salt.

268

4.2.2 Physical Measurements.
UV-vis spectra were acquired on a Cary-50. X-band EPR samples were prepared

in a dry, inert atmosphere (N2) by dilution in a mixture of butyronitrile and

propionitrile (9:2), both of which had been distilled from CaHz, degassed, and
stored under an inert atmosphere. The resulting sample concentration was
approximately millimolar. Continuous-wave X-band EPR spectra were obtained
with a Bruker ESP300E spectrometer using a perpendicular-mode cavity (Bruker
41028T/9516) with resonances around 9.458 GHz. For each compound an
appropriate power was chosen to avoid saturation of the signal. An Oxford ESR
A900 helium cryostat was employed to maintain the temperature at 4 K.

Cyclic voltammetry was carried out under inert conditions using a CHI
Electrochemical Workstation (model CHI620D). Solutions of the compounds
were dissolved in acetonitrile containing tetrabutylammonium
hexaﬂuorophosphate as the supporting electrolyte. A standard three-electrode
setup was employed with a platinum working electrode, graphite counter electrode

and Ag/AgCl reference electrode.
4.3 Theory of Heisenberg Spin-Exchange.

Exchange interactions essentially form weak bonds, in the same sense as spin

pairing is characteristic of bond formation.23 For example, two hydrogen radicals,

269

each having S = 1/2, come together to form a hydrogen molecule with singlet (S =
0) ground state. A triplet state is also formed, although at a much greater energy,
existing as an excited state in this system. Likewise, when two unpaired spins are
in close proximity a weak bond is formed. However, unlike exchange of 0- or rt-
bonds, these interactions are generally on the order of kT. Several criteria must
be satisﬁed for this interaction to occur: 1) the obvious observation that unpaired
electrons must reside on each substituent participating in the interaction; 2) an
orbital pathway for the exchange must exist; 3) energetic matching of the magnetic
orbitals (i.e. those containing unpaired spin). The theoretical physical picture of

this phenomenon is well established, and is described by Anderson’s Theory.24

Goodenoughzs’26 and Kanamori27 later gleaned criteria for spin-exchange
interactions from this formalism, which qualitatively express the outcome of a
spin-exchange interaction from a molecular orbital perspective. For two magnetic
orbitals (those containing unpaired spin) that have a large overlap integral, the
interaction is antiferromagnetic, while if the overlap integral is zero, the
interaction in ferromagnetic. Finally, if a magnetic orbital overlaps with an empty
orbital, the interaction is ferromagnetic. These rules are discussed more
thoroughly below in terms of metal-semiquinone systems.

In some instances paramagnetic centers are directly bound, resulting in a
through-bond spin exchange interaction that is appropriately named “direct
exchange.” However, in the majority of cases the exchange interaction is

facilitated by a diamagnetic bridge between the paramagnetic centers. This is

270

quite common in multinuclear transition metal systems, and is referred to as
“superexchange.” Considering the interaction of two centers of spin S, and S2, the

exchange interaction will, according to the angular momentum addition rules,

produce a “spin ladder” of states.

n+gHs+g—¢m,

 

s, — s,|

The energetic ordering of these states is determined by using the Heisenberg
Hamiltonian (also called the Heisenberg-Dirac Hamiltonian), which was originally
derived from the exchange Hamiltonian for the hydrogen molecule (S, = S2 =
'/2),28’29 but has been applied widely and successfully to many more complicated
spin systems. For two interacting spin centers S, and $2, the isotropic Heisenberg

Hamiltonian is given by equation 4.1.
H = lesl-s2 4.1

Here J ,2 is a coupling constant which gauges the interaction between the two spin
centers. This Hamiltonian has also been represented by various other coupling
constants (2J,2, -J,2, -2J,2), lending confusion to the application. In the formalism
presented above in equation 4.1, a negative value of J ,2 indicates ferromagnetic
exchange coupling, leading to stabilization of the highest total spin state, while a
positive value of 1,2 designates antiferromagnetic exchange, wherein the lowest

total spin state is stabilized.

271

In the direct (through-bond) exchange regime the lowest-lying total spin of
the system is predicted by examination of the symmetry of the constituent
magnetic orbitals. This is essentially a summation of the aforementioned
Goodenough and Kanamori rules. Considering only two spin centers, S, and S2, an
antiferromagnetic interaction will occur if the magnetic orbitals (i.e. those
containing unpaired electrons) of the spin centers are of the appropriate energy and

symmetry to interact. This interaction leads to an anti-parallel alignment of the
spin dipoles, resulting in the lowest total spin state, IS, ‘SZI , being lowest in

energy. This situation is realized in the diagram of Figure 4-4. On the other hand,
ferromagnetic exchange stems from complete orthogonality of magnetic orbitals.

The non-interacting nature of these orbitals stabilizes the highest possible total

spin state, '5, +52| , according to Hund’s rule, as depicted in Figure 4-5. The
spatial aspect is reﬂected, for example, in the well-understood role of orbital
symmetry of metal-semiquinone complexes of the ﬁrst row. For example,
complexes of [Ni(II)(tren)(psq)]+ (psq = phenanthrenesemiquinone) display
ferromagnetic exchange coupling, while the analogous complexes of
Chromium(III) display antiferromagnetic exchange coupling (vide infra).19’30’3l
The energetic consequences of Heisenberg spin exchange can be determined by

applying the so-called Kambe approximation, which deﬁnes the total spin operator

S, as a sum of spin operators, analogous to the total spin vector.32 For a two-

272

component spin exchange system, this approximation yields eigenvalues of the

form given by equation 4.2.

E = %[St (S, + l) — S1 (S1 +1)- S2 (S2 + 1)] (4.2)

 

Cr(III), s = 3r, so, 5 = 11,

* —7[

___eg

Ll—Lt

 

 

 

 

Figure 4-4: Antiferromagnetic exchange arises from non-orthogonality of

constituent spin orbitals, resulting in the stabilization of the lowest total spin state.

273

 

 

Ni(|l), s = 1 so, 5 = 1I2

4:

 

 

 

Figure 4-5: Ferromagnetic exchange arises from complete orthogonality of
constituent spin orbitals. The highest total spin state is stabilized according to

Hund’s rule.

Before moving on one last point of interest must be made. In transition
metal-semiquinone systems there is often confusion as to the oxidation states of
the constitutents of the system due the redox active nature of both transition metals
and quinoid ligands. An example (albeit somewhat unrealistic) that pertains to the
systems of interest in this chapter is the theoretical resonance formulations of the

chromium-semiquinone system to yield other redox states, Figure 4-6. Two of

274

 

these states would have no spin exchange interaction, as the ligands would be
diamagnetic in nature. Fortunately, due to the relative redox inertness of the
Chromium(III) ion, this issue does not cause a concern for these studies. In fact,
as discussed below, there are characteristic Spectrographic and structural
signatures that can be employed to very quickly determine whether the ortho-

quinone ligand is in its fully reduced catechol form or radical semiquinone form.

 

/o /o\ /o\
Cr(lV) -—- Cr(lll) -—- Cr(ll)
O O O

 

 

 

Figure 4-6: Possible redox formulations of the Chromium(III)-semiquinone
exchange-coupled species. Of these, the only realistic formulation is that of the

Chromium(III) species due to the relatively redox inert nature of this ion.

4.4 Results and Discussion

4.4.1 Previous Results: [Cr(tren)(3,6DTBCat)](PF6) and
[Cr(tren)(3,6DTBSQ)](PF6)2. As noted in the introduction to this chapter, the
impetus for the research presented in this chapter arose from an ultrafast transient
absorption study of [Cr(tren)(3,6DTBCat)](PF6) and its semiquinone analog,
[Cr(tren)(3,6DTBSQ)](PF6)2. This section will present information on these

systems which will prove relevant to the interpretation of similar systems

presented throughout this study. The UV-vis spectrum of

275

*QEE—ﬂ

[Cr(tren)(3,6DTBCat)](PF6) has been reported previously, and is reproduced in

Figure 4-7(a).18

 

 

01

A

 

 

8 (x103 M'1 cm'l)
OD

 

 

 

2
A.
1
O 3 L ‘2 A 1

Energy (x104 cm'l)

 

 

 

Figure 4-7: UV-vis absorption spectra of [Cr(tren)(3,6DTBCat)](PF6) (a) and its

semiquinone analog, [Cr(tren)(3,6DTBSQ)](PF6)2 (b), adapted from reference 19.

The spectrum is similar to that of Cr(acaC)3, Which exhibits a single broad, weak
absorption (epsilon ~60 M'l cm") corresponding to the spin allowed 4A2 —> 4T2
ligand ﬁeld transition, a medium intensity transition (~ 400 nm), and higher

energy absorptions of much greater oscillator strength corresponding to charge

transfer and ligand-based transitions. In this system, however, the term symbols

276

are different due to the lowered symmetry of the system: the ﬁrst spin-allowed
excited ligand ﬁeld state in octahedral symmetry, 4T2, correlates to the states 4B,,
4A2, and 43,, while the ground state has the term symbol 4B2. The effects of

lowered symmetry are shown in Figure 4-8.

 

 

‘. ’ ' . _4B1
4 o”’
T2 —::: ----- —4A2
~ ~ ~ ‘ —4A
1
.— ‘ g —282
2T1_:.E'E ..... —ZB1
~~~~~ 2A
— 2
_ 2
2E_:: ’’’’’ A1
- ~ ‘ _ZBZ
4A2—- ....... —4B2
0 C2,,

 

 

 

Figure 4-8: Energy-level diagram of a chromium(III) system under C2,,

symmetry, showing the three lowest-lying ligand ﬁeld excited states.

It should be kept in mind that the 2IE and 2r, states, which are usually referred to
collectively as the 2E, are separated by only a few hundred wavenumbers.

Therefore, the low-symmetry splittings of these states represents a high density of

277

doublet excited states. The fact that the width of the ligand ﬁeld absorption in this
system is not markedly different from that of Cr(acac)3 indicates that the splitting
of this transition due to lowered symmetry is not that signiﬁcant. Assignments in
the ultraviolet region of the spectrum were assigned as ligand-centered transitions
due to lack of Crm/CrII and Crm/CrlV redox couples and lack of solvatochromatic
behavior ruling out charge-transfer transitions.

While the absorption spectrum of the chromium(III)-catechol complex
displays many features similar to the corresponding spectrum for Cr(acac)3, the
spectrum of the chromium(III)-semiquinone complex is markedly different:
extinction coefﬁcients are much greater across the entire visible and near-infrared
spectrum, the transitions are, in general, much narrower and structured, and of
most interest is an extremely sharp transition that occurs near 15000 cm'], in the
energetic range of the lowest-lying ligand ﬁeld transitions. While an increase in
10Dq, and thus the energy of states derived from the 4T2 state, is expected due to
the weaker u-donating ability of the semiquinone ligand another effect is deemed
to be much more important. To begin to understand the differences between these
spectra, an appreciation and understanding of the requirements to obtain exchange
coupling is necessitated. These requirements, which were outlined above, can be
reiterated in one sentence-fragment: each moiety must have unpaired spins and
there must be an energetic and spatial matching of the orbitals involved. While

this formalism is usually thought of only in terms of the ground state, the

278

 

 

phenomenon is not limited to acting only on ground states as long as the
requirements listed above are met. This, therefore, allows for the drastic changes
observed in the absorption spectrum, as new spin-allowed transitions arise due to
the spin exchange interaction. The visible part of the spectrum can be understood
in terms of exchange coupling operating to split the ligand ﬁeld states under C2v
symmetry, as shown in the energy level diagram below (Figure 4-9).16 It is
important to note that coupling is expected to be weaker in states with chromium
eg“ occupation due to lowered overlap of the magnetic orbitals of these states with
those of the n-based semiquinone magnetic orbitals.

Experimental proof of this mechanism can be found in single-crystal

polarized absorption studies of spin-exchange complexes of chromium(III).2‘3’33’34

In general, as noted by McCarthy and Gl'idel,34 spin allowed transitions appear to
be less perturbed by spin-exchange interactions than spin-forbidden transitions.
One can easily imagine the transition derived from the 4A2 —> 4T2 transition to
persist under the features of the Cr-SQ absorption spectrum near 15000 cm". As
has been shown in other spin-exchange complexes of chromium(III), introduction
of spin exchange can essentially break spin restrictions as the spin-exchange
interaction produces both ground and excited states of new spin. In several cases,
transitions to 2E and 2T,-derived states have been observed, which usually present

as sharp absorptions due to the intraconﬁgurational nature of the transition. This

279

 

 

 

 

 

 

 

 

rats—-.- ----- ”—82
_ _ g _ ~__ 1A2
_——1 3
2T1< 281—-=::::—’ A1
5 . -—_- 1A
- - -—--- 3B
2 ——..-; ::: 2
K A2 ““““ _I___1B2
f2A1 X... :1 Z :::: 7;?— 381
- - _ - ""'"""1B1
2E <
23 _. ——————— KT” A2
\ 2 ....... 1" 1' 1A
.1--. 2
482 =: : :::: .___. 5A2
~~~~~ ..l... 3A2
02,, CZV-split by
.Is,s2

 

 

 

Figure 4-9: Splittings of the C2,, electronic components as a result of anti-

ferromagnetic coupling between the chromium ligand ﬁeld states (left side) and
the semiquinone It magnetic orbital. The diagram reveals that many spin-allowed
transitions are possible as a result of the spin-exchange interaction, whereas

energetically similar transitions in non-exchanged system are forbidden.

appears to be the case in this complex as well. In this system, the n-donating

ability of the quinoid ligand serves to raise the energy of the tzg orbital set,
lowering the energy of the 4Tz-derived states (of (tzg)2(eg*)l parentage) and raising

the energy of the 2E- and 2T,-based states (of (tzg)3 origin) relative to

280

‘5‘???“ ‘

acetylacetonate. The narrow breadth of the transition, as well as the n-donating
ability of the ligand, point to the fact that these transitions are to states derived
from 2E. The observation of these new features provides insight into the energy of
the 2E state in the catechol complex, which is non-emissive. Given that the low-
energy semiquinone absorption features are due to states derived ﬁom the 2E term,
then according to Figure 4-8 the nominal 4T2 state (or some component thereof) of
the catechol complex is lying below the 2E state.35 This is in agreement with the
lack of emission from the catechol complex: a lowest-energy quartet excited state
would provide excellent coupling to the ground state and thus facilitate rapid
nonradiative ground state recovery.

While this previously made assignment conforms with the non-emissive
character of this complex, the assignments of excited state energy and splittings
for both the catechol and semiquinone complexes remains uncertain without
polarized single crystal absorption spectra to quantify the state splittings. For
chromium systems this technique has been applied to many different systems,

37—39

including crystals of Chromium(III) oxide36 and chloride materials (which

have been studied as potential laser media, but mainly to explore fundamental
aspects of electronic structure), as well as many spin-exchange chromium(III)

dimers34 (vide infra). One recent example is that of trans-[Cr(HgO)4C12], studied

by Reber and coworkers.40 Using low-temperature polarized absorption and

281

 

luminescence spectroscopy the authors were able to quantify the splitting due to
lowered symmetry of states arising from the 4T2 state.

Preliminary ultrafast transient absorption studies on

[Cr(tren)(3,6DTBCat)](PF6) and [Cr(tren)(3,6DTBSQ)](PF6)2 have been carried

out.35 This study compared the nonradiative dynamics of these complexes to a

low-symmetry redox innocent analog, [Cr(tren)(acac)](PF6)2. Results for

[Cr(tren)(3,6DTBCat)](PF6) and [Cr(tren)(3,6DTBSQ)](PF6)2 after excitation at

333 nm are presented in Figure 4-10. In both cases the single-wavelength kinetic
traces were ﬁt with a monoexponential decay function, revealing a lifetime of ~
1.5 ps for the catechol complex and ~ 6 ps for the semiquinone analog. Both
traces decay back to the baseline (no long time y-offset), implying ground state
recovery that is complete within 10 ps for the catechol and within 25 ps for the
semiquinone. The dynamics of the catechol complex mesh well with the general
picture outlined above, namely that of a 4T2-derived electronic state lying lowest
in energy and providing a facile mechanism for ground state recovery. The
comparable dynamics of the semiquinone complex suggest a similar mechanism of
ultrafast ground state recovery, although, as discussed above, the electronic
structure is greatly affected by the presence of spin exchange. The major goals of
this research project are l) elucidation of electronic structure in these complexes
and 2) correlation of this structure to excited state dynamics. This work led us to

develop the [Cr(tren)(3,6-R-Q)]+“+2 series in the hopes of elucidating the excited

282

electronic structure, and thereafter aScertaining mechanisms of nonradiative decay

in these complexes.

 

 

_ . o [Cr(tren)(3,6-t-butylcat)]+
15_ e [Cr(tren)(3,6—t-butylSQ)]2+

 

 

 

 

 

 

 

-5 . . . .
0 5 1O 15 20 25
Time (ps)
Figure 4-10: Ultrafast transient absorption dynamics of

[Cr(tren)(3,6DTBCat)](PF6) and [Cr(tren)(3,6DTBSQ)](PF6)2 in acetonitrile

solution after excitation at 333 nm.

4.4.2 [Cr(trerr)(3,6-R-Q)]"""2 series

Synthesis. Only the chromium-catechol members of this series have been
synthesized to date. For all four catechol members of this series the complexes
were prepared using the same general procedure, namely reﬂuxing a solution of

the sodium catecholate complex with [Cr(tren)Clz]Cl in a methanol/water solution

283

 

under inert conditions. This was quite fortuitous given the electronic differences
across the series, and likely speaks to this method being a general preparation for
this class of molecules. Furthermore, high-quality crystals of all of these

complexes (with BPh4 as the counter-anion) can be grown from a methanolic
solution of the complex and NaBPh4. Unfortunately, only crystallographic data of
[Cr(tren)(pycat)]+ and [Cr(tren)(3,6-CNcat)]+ were available for this dissertation.

Also remarkable was the general stability of these complexes to oxidation under
ambient conditions, with the exception of the 3,6-amino—l,2-catechol complex,
which is discussed further below as the electrochemistry of these complexes is

explored.

[Cr(tren)(pyrocatecholate)]BPh4. This complex forms blue/green powders and
sky blue crystals. The crystal structure of this complex is presented in Figure 4-
11, which also shows some selected bond lengths and angles for this complex.
Full crystallographic data is presented in Appendix C. Starting with the metal
coordination bonds, the Cr-O bond lengths of approximately 1.92 A are within the

range of values reported for other chromium complexes (such as the high-
symmetry Cr(acac)3-type complexes of Chapter 3, with typical bond lengths of
1.95 A) while the Cr-N are also fairly typical of chromium-am(m)ine complexes

with bond lengths of ~ 2.10 A.41 The bond lengths of the catechol complex

correspond to the double bond delocalized 11: structure of the ligand, with all values

284

 

 

 

 

 

Cr1-O1 1.9244

Cr1-02 1 .9226

M-L Bonds Cr1-N1 2.093
(A) Cr1-N2 2.124
Cr1-N3 2.071
Cr1-N4 2.077

O1-Cr1-N1 179.28

02-Cr1-N4 176.46

M-L Angles N2-Cr1-N3 160.90

(°) O1-Cr1-02 85.21
O1-Cr1-N3 87.31
O1-Cr1-N2 85.77

02-C8 1 .364

01-07 1 .352

C7-C8 1 .400

Ligand Bonds CB-C9 1 .379
(A) C7-C12 1 .389
C9-C10 1 .396

C1 1-C12 1 .394

C10-C1 1 1 .375

 

 

 

 

Figure 4-11: Crystal structure and selected bond lengths and angles for

[Cr(tren)(pycat)]BPh4.

285

for bond lengths around 1.4 A. This is in contrast to the alternating bond lengths
of a semiquinone species, representing a more localized picture of the double
bonds, with values of either 1.45 A, corresponding to a single bond, or 1.35 A for
a double bond.'8 This crystallographic information for the ligand conﬁrms the
catechol oxidation state.

The ligand oxidation state is further conﬁrmed by the UV-vis absorption
spectrum. As described above, the presence of spin exchange affects not only the
ground state electronic structure but also the excited state electronic structure,
leading to observation of previously spin-forbidden transitions. A quick glance at
the absorption spectrum of Figure 4-12 reveals that the spectrum has none of the
features associated with a chromium(III)-semiquinone complex, discussed above
for the case of [Cr(tren)(3,6-DTBSQ)]2+.

In the ultraviolet one observes a peak near 290 nm (8 ~2650 M 1 cm") and
a more clearly deﬁned absorption peak centered at 340 nm (8 =l650 M'l cm'l).
On the red shoulder of this transition is another peak. Finally, a broad, weak
featureless transition is centered at 612 (8 = 71 M‘I cm'l). Given the trends
observed for the high-symmetry systems of Chapter 3, is speculated that the
feature at 340 nm is a charge-transfer transition while the peak at 290 nm is an
intraligand transition. The low-intensity absorption at 612 nm is clearly a

transition into the lowest-energy 4T2 state (0 notation), which in this complex of

C2,, symmetry is split into three singly-degenerate term states (48,, 4A2 and 4A,).

286

This splitting is not observed in the room-temperature absorption spectrum of

 

 

 

 

 

     

 

 

 

Figure 4-12.
4 I I l _
r7; 3 —8 ‘
o — 8 X 20
'2
co 2 - -
S2
. 25
00
1 - ..
. 0 -; ......... i ................ .- ............ 1
400 600 800
Wavelength (nm)

 

 

 

Figure 4-12: UV-vis absorption spectrum of [Cr(tren)(pyrocatecholate)]BPh4 in

acetonitrile.

A cyclic voltammogram of this complex is shown in Figure 4-13. The
ﬁgure reveals a reversible process with Em = + 323 mV (vs Ag/AgCl) and a peak
separation (AEp) of 62 mV. This value corresponds to the one-electron catechol to

semiquinone oxidation. A second scan to more oxidizing potentials (not shown)

reveals a second, irreversible oxidation process at 873 mV, corresponding to

287

oxidation of the semiquinone complex to the quinone complex. It has been shown
that the neutral quinone ligand has very little binding afﬁnity for the
Chromium(III) metal center, and therefore this oxidation results in decomposition
of the metal complex, reﬂected by the irreversible nature of the oxidation. These
oxidative steps are shown in equations 4.3 and 4.4, and were described previously

for [Cr(tren)(3,6-DTBSQ)](PF6).'8

 

 

 

 

 

 

+e
[Cr(tren)(pySQ)] (Z) [Cr(tren)(pycat)] EV2 = + 0.323 V (4.3)
_e—
_e—

[Cr(tren)(pySQ)] —> [Cr(tren)(pr)] on = + 0.873 V (4.4)

1.0 ~ I I I I I I I I 1

2 0.5 ~ ~

3 ' j

45- 0.0 ‘

g -0.5 - -

0 -1.0 - ’i -

-1.5 ' ‘ ' : ' ‘ ‘
0.6 0.4 0.2 0.0
V (vs Ag/AgCl)

 

 

 

Figure 4-13: Cyclic voltammogram of [Cr(tren)(pyrocatecholate)]BPh4 in

acetonitrile. The values are references of Ag/AgCl.

288

1

[Cr(tren)(3,6-CN-l,2-catecholate)]BPh4. This complex is brown both as
powders and crystals. The crystal structure and selected bon lengths and angles
are shown in Figure 4-14. In many respects metal-ligand bond lengths and angles
are similar to those observed for [Cr(tren)(pycat)]+ and will not be elaborated on
here. Unlike the pyrocatecholate complex, the C-C bond lengths of the ligand do
have alternating character, but not to the extent observed for a semiquinone
system. These bond lengths are consistent with the delocalized rt-bonding of a
catechol complex, with deviations perhaps arising from the electron-withdrawing
capability of the neighboring substitutents, which weaken the bonds via
introduction of 8+ charges, as discussed in the literature.42

The lack of sharp features to the red of 650 nm in the UV-vis absorption
spectrum of this complex conﬁrms that a semiquinone species was not formed.
These features arise in chromium(III) complexes due to the presence of spin

exchange. In the UV one observes an intense transition with two sharp peaks near

365 nm (8 ~ 8530 M'1 cm”). The two sharp peaks are separated by about 1200
cm", the frequency of a C=C stretch. This feature is tentatively assigned as a

charge transfer feature given the intensity and observation of energetically similar
transitions in other chromium complexes explored throughout this dissertation.

On the low-energy shoulder of this feature, near 450 nm, are two peaks of
relatively low-intensity (8 ~ 200 M'1 cm'l), while further into the visible is a broad

transition centered at 567 nm (8 ~ 93 NT1 cm'l).

289

1'4'1-‘IM'hia
_l

 

 

 

 

 

 

Cr1-O1 1 .9568
Cr1-02 1 .9176
M-L Bonds Cr1-N1 2.083
(A) Cr1-N2 2.0942
Cr1-N3 2.087
Cr1-N4 2.060
O1-Cr1-N4 178.14
OZ-Cr1-N2 176.19
M-L Angles N1 -Cr1-N3 163.22
(°) O1-Cr1-02 84.73
O1-Cr1-N1 87.02
O1-Cr1-N3 88.22
02-C8 1 .325
01-07 1.336
C7-08 1 .424
Ligand Bonds 08-09 1.395
(A) C7-C13 1 .397
C9-C1 1 1.41 1
012-013 1 .423
C1 1 -C12 1 .363

 

 

 

Figure 4-14: Crystal structure and selected bond lengths and angles for

[Cr(tren)(3 ,6-CN- l ,2-catecholate)]BPh4.

290

 

 

 

 

 

 

 

 

 

 

 

 

300I400I500I600I700800900
Wavelength (nm)

 

 

 

 

Figure 4-15: UV-vis absorption spectrum of [Cr(tren)(3,6-CN-l,2-

orthocatecholate)]BPh4 in acetonitrile.

The peaks near 450 nm are likely 4T,-based ligand ﬁeld states; the spacing of
about 1500 cm"1 could be tied to effects of lowered symmetry, but it is more likely
given the energetic proximity to the nominally charge-transfer transition that this
splitting is in fact vibronic structure as a result of mixing of ligand- and metal-
based orbitals. The broad feature at 567 nm is clearly due to a transition into 4T2-

based ligand ﬁeld states. The splitting due to lowered symmetry is not observed,

291

but the expected trend of a blue-shift of this peak, a result of the stronger 1t

accepting ability (increasing 10Dq) relative to the unsubstituted catechol, is

observed.

The cyclic voltammogram of [Cr(tren)(3,6-CN-l,2-orthocatecholate)]BPh4
in acetonitrile is presented in Figure 4-16. All irreversible oxidation is observed at
+ 0.620 V, in close proximity to a reversible wave with Em = + 0.703 V and ABp

= 59 mV. This reversible process is easily assigned as originating from

 

catechol/semiquinone redox couple. The potential is shifted by approximately 380
mV relative the unsubstituted catechol complex, [Cr(tren)(pycat)]+ as a result of

strongly electron-withdrawing cyano substituents. The nature of the irreversible

 

 

 

 

 

07- I j I I I ' I

1

00- L/ 1

s -07- -

it -1.4- _
'5 1

o -21- -

-2.8- -

 

 

 

1.0 0.8 0I6 I 04 I 0.2 0.0
V (vs AglAgCl)

 

 

 

Figure 4-16: Cyclic voltammogram of [Cr(tren)(3,6-CN-1,2-

orthocatecholate)]BPh4 in acetonitrile. Potentials are referenced versus Ag/AgCl.

292

process observed at + 0.620 V is not apparent, but a similar peak observed for
[Cr(tren)(3,6-dinitro-1,2-orthocatecholate)]BPh4 (see below) suggests a common
impurity in both samples. A very irreversible oxidation process is observed at +
1.73 V, corresponding to the semiquinone/quinone redox couple, which results in
decomposition of the complex due to the low binding afﬁnity of the fully oxidized

quinone ligand.
+e‘
[Cr(tren)(3, 6 — CNSQ)] 2:2 [Cr(tren)(3,6 — CNCat)] E1 /2 = + 703 mV (4.5)
_e- '

[Cr(tren)(3,6 — CNSQ)]-—e)_ [Cr(tren)(3,6 — CNQ)] on = +1.72 V (4.6)

[Cr(tren)(3,6-dinitro-1,2-orthocatecholate)]BPh4. The UV-vis absorption
spectrum of this complex is presented in Figure 4-17. In general, two high-
intensity absorption features are observed with maximum absorption at 340 nm (8

= 7610 M" cm") and 450 nm (4600 M" em"). To the blue of the 340 nm feature

are a series of sharper peaks—these features do not show a constant spacing, and
therefore likely do not represent a vibronic progression. These peaks are
tentatively assigned as arising from intraligand transitions of the catechol ligand.
An absorption spectrum of the sodium salt of the catechol ligand reveals similar
features to those observed at 340 nm and 450 nm, and therefore the assignment of
these absorptions as due to intraligand transitions is made. However, there are

undoubtedly charge transfer transitions in this energetic range as a result of the

293

 

 

 

12 1 I 1 I T r ' I ' I

10

T
l

 

 

 

8 (x103 M‘1 cm")
0)

 

 

 

0 L4 1 1 l 1 l 1 I 1 1
300 400 500 600 700 800
Wavelength (nm)

 

 

 

 

Figure 4-17: UV-vis absorption spectrum of [Cr(tren)(3,6-dinitro-l,2-

orthocatecholate)]BPh4 in acetonitrile.

extended 7: structure and rt-accepting ability of the dinitrocatechol ligand. A
comparison of the absorption spectrum with a gallium(III) analog would perhaps
provide a means of parsing out charge transfer and intraligand transitions.An
absorption of much weaker intensity, centered near 560 nm (290 M‘1 cm”), resides
on the low energy onset of the more intense absorption features. This absorption

represents a transition to 4T2-derived ligand ﬁeld states, and the large intensity

294

 

(compared to the ~ 60 M'1 cm'1 observed for Cr(acac)3) is likely a result of
signiﬁcant mixing with charge transfer and intraligand transitions of similar
energy, i.e. intensity stealing from these states. The energy of the transition is
similar to that observed for the 3,6-CN-l,2-catecholate derivative, reﬂecting a
high degree of 1t accepting ability for both ligands.

The cyclic voltammogram of [Cr(tren)(3,6-dinitro-1,2-
orthocatecholate)]BPh4 in acetonitrile is presented in Figure 4—18. An irreversible
oxidation is observed at + 0.636 V, in close proximity to a reversible wave with
Em = + 0.807 V and AEp = 59 mV. This reversible oxidation is easily assigned as
the typical catechol to semiquinone oxidation. In this case, the potential is shifted
by approximately 480 mV relative the unsubstituted catechol complex,
[Cr(tren)(pycat)]+, as a result of strongly electron-withdrawing nitro substituents.
It remains unclear what physical process is occurring for the irreversible oxidation
at + 0.636 V, although it is similar to the irreversible process observed for
[Cr(tren)(3,6--CNcat)]+ (at + 0.620 mV vs Ag/AgCl), suggesting a common
impurity between both samples. A second irreversible oxidation is seen at 1.717
V, corresponding to the semiquinone to quinone one electron oxidation. This
value represents an 840 mV shift relative to the same value for [Cr(tren)(pycat)]+,
again as a result of the electron withdrawing nature of the nitro substituents.

These two one-electron processes are summarized in equations 4.8 and 4.9.

295

—" Vlw ..."..‘1

 

0.5

0.0
-O.5
-1.0
-1.5
-2.0
-2.5

Current (uA)

 

 

L

 

—

J

 

 

1.2‘1.0‘o.8‘o.6‘o.4‘o:2'o.o
V (vs Ag/AgCl)

 

 

Figure 4-18:

Cyclic

voltammogram

of

[Cr(tren)(3,6-NOz-1,2-

orthocatecholate)]BPh4 in acetonitrile. Potentials are referenced versus Ag/AgCl.

[Cr(tren)(3, 6 — N0,SQ)] +——_> [Cr(tren)(3, 6 — NOZCat)]
_e_
[Cr(tren)(3, 6 — N0,SQ)] 16) [Cr(tren)(3, 6 — NOZQ)]

[Cr(tren)(3,6-diamino-l,2-orthocatecholate)]BPh4. This complex was prepared
as a dark purple powder.
dramatic shift from those of the other members of this series. High intensity (>
103 M‘I cm’l) absorptions are observed across the entire ultraviolet, visible, and

out to nearly 850 nm. As a result, a solution of this complex has a much more

+8“

intense color than that of other members of this series.

three main, broad features centered at 29790 cm'1 (335 nm, 8 ~ 2500 M'1 cm'l),

296

,2 = + 0.807 V (4.8)

on = +1.717 V (4.9)

The absorption spectrum, Figure 4-19, represents a

In acetonitrile there are

“a; W .7? _m." S" 72751!

 

19305 cm" (518 nm, 8 ~ 1240 M'1 cm“) and 13200 cm" (758nm, 6 ~ 850 M" cm"
1). Given that the sodium salt of the diaminocatechol ligand is colorless, showing
only absorptions in the ultraviolet, the higher energy transitions at 335 nm and 518
nm are generally assigned as charge transfer in nature, although preparation of a
d10 analog complex would provide the best means of distinguishing charge transfer
and intraligand transitions. The shift to higher energy of the. feature near 750 nm
(in acetonitrile) to approximately 630 nm (in methanol) may be indicative of this
transition originating out of a non-bonding orbital (in this case on the amino
group): the higher hydrogen-bonding ability of methanol versus acetonitrile

would stabilize these orbitals, increasing the energy of the transition. This
transition is therefore assigned as an n —> 1:. transition, or perhaps a LMCT
originating out of the amino nonbonding orbitals. In the acetonitrile spectrum one
observes a ﬂat region between 12500 — 15000 cm". This is generally indicative of
multiple Gaussian lineshapes summing together. In fact, Gaussian deconvolution
of this region of the spectrum reveals a broad low-intensity (90 — 120 M'] cm'l)
absorption, centered at about 15470 cm'I (646 nm). This feature has the correct
intensity and energy, given the rt-donating ability of the amino substituents, to

correspond to the 4T2 absorption.

297

 

 

 

In Acetonitrile
In Methanol

 

e (x 103 M‘1 cm")

 

 

 

 

Energy (x103 cm'1)

 

 

 

Figure 4-19: UV-vis absorption spectrum of [Cr(tren)(3,6-diamino-1,2-

orthocatecholate)]BPh4 in acetonitrile.

The cyclic voltammogram of this complex in an acetonitrile solution is
shown in Figure 4-20. All potentials are versus Ag/AgCl. Not shown is a highly
irreversible oxidation process at 0.764 V, corresponding to the semiquinone —’
quinone oxidation. This value is approximately 1 V less than that of the nitro- and
cyano-substituted catechol complexes, and about 100 mV less than the

pyrocatechol complex. Given the n-donating ability of the amino substituents the

298

value of 764 mV reﬂects the expected trend. In the region of the reversible
catechol/semiquinone redox pair (Figure 4-20), one observes irreversible

processes at 219 mV and -449 mV. A reversible wave with E1 /2 = - 457 mV and

AEp = 95 mV is observed and assigned as the catechol/semiquinone redox pair.
This value represents an approximately 800 mV shift relative to the pyrocatechol
complex, and accounts for the air sensitive nature of this complex. No assignment
is made for irreversible processes, although these may be due to impurities in the
sample.

Given the data presented above, as well as the potential for multiple
binding modes of this ligand, it is suggested that future efforts be directed toward

other n—donating substituents (methoxy or dimethylamino).

 

 

 

 

 

 

l
1.0- -
A 0.5- -
l <1: .
3
_,_, 0.0- ,
C .
9
I— -0.5- _
3
o n
-1.0- A
_1.5 . 1 L 1 . 1 . 1 . 1 1 1 . 1 L 1 1 1
0.4 0.2 0.0 -0.2 -0.4 -O.6 -0.8 -1.0 -1.2
V (vs Ag/AgCl)

 

 

 

Figure 4-20: Cyclic voltammogram of [Cr(tren)(3 ,6-NH2- 1 ,2-

orthocatecholate)]BPh4 in acetonitrile. Potentials are referenced versus Ag/AgCl.

299

ﬁferg—q .1

 

[Cr(tren)(3,6 — NH 2SQ)] +262; [Cr(tren)(3,6 — NH 2Cat)] E1/2 = —457 mV (4.10)
_e_

[Cr(tren)(3,6 — NHZSQ)] :— [Cr(tren)(3,6 — NH2Q)] on = +764 mV (4.11)

4.4.3 Spin Exchange Members of the [M1M2(tren)2(CA"‘)]lm+ Series.
[GaCr(tren)2(CAsq’°at)](BPh4)2(BF4) (4). Magnetic susceptibility of 4 from 2 —

350 K indicates a temperature-independent value of neff = 2.72 i 0.05 113 (1.13pm-0n1y

6

for S = 1 is 2.83 113) , characteristic of an S = 1 state.1 This implies an

antiferromagnetic interaction between Chromium(III) and the trianionic CAM"can

bridge and a AE between S = 1 and S = 2 states exceeding kBT at 350 K (~240 cm'

1), leading to an essentially isolated triplet ground state, with the quintet state far

1,43 Ill

above it. This corroborates what we18 and others have observed for other Cr -

semiquinone complexes. As expected for 4 the X-band perpendicular— and
parallel-mode spectra revealed no features unique to this molecule (Figure 4-21).
Instead, the spectra reveal impurities corresponding to other semiquinone species,

111 dimer 2. This is in contrast to the ESI-MS

namely the Cr"I dimer 6 and the Ga
data for this complex presented previously,” which exhibited a single molecular

peak corresponding to complex 4, with no indication of complexes 2 or 6, and
testiﬁes to the higher sensitivity of EPR as compared to mass spectroscopy

techniques. Parallel-mode spectra indicate a transition around g’ = 2 (where g’ is

300

— V:

U i ‘C‘ 3'.”

the effective g-value and is given by g’= hv/BH, so that for X-Band frequency g’ =
2 corresponds to about 3300 G) that is likely a “leaking” of the perpendicular-

mode spectrum of the gallium(III) dimer. If this is the case, then zero-ﬁeld
splitting between the IO) and It 1) states within the triplet manifold is large

enough that no transitions are observed at X-band frequency, and the signs of the

zero-ﬁeld splitting parameters D and E are unknown. It is also possible that the

observed transition in the parallel mode spectrum is due to the I‘D—4+1)

transition. However, the parallel-mode signal is very weak, and has the same
temperature dependence as the main line of the perpendicular mode X-band
spectrum of 2, leading to the conclusion that this observed signal is in fact due to a
leaking of the perpendicular mode transition. High-frequency, high-ﬁeld
measurements will be required to address the magnitude of the zero-ﬁeld splitting
in this complex.

This series offers the possibility of fully characterizing the spin Hamiltonian
parameters for the Chromium(III)-semiquinone system. In this case the spin Hamiltonian

parameters can be considered in the strong exchange limit, i.e. where the effect of other
operators acts as a perturbation of the isotropic exchange Hamiltonian J 1231.52, the

following linear relationships hold for the various spin-exchange levels:
gs =clgA+cng (4.12)

D, =d1DA+d2DB+d12DAB (4.13)

301

where g3 and D3 are the g and zero-ﬁeld splitting tensors, respectively, gA and g;
are the g-tensors of the individual spin centers, and DA and DB are the zero-ﬁeld
splitting tenors of the individual spin centers. DAB is the asymmetric exchange
tensor, and arises via the exchange interaction. Bencini and Gatteschi3O provide
the scalar values c1, c2, (11, d2, and d12 for various spin pairs (d1 = 0 in this system
because the semiquinone doublet state does not have any zero-ﬁeld splitting (L =

9

0), so there is no “DSQ’ contribution). The strong exchange limit is a good
approximation for 4, given that the exchange coupling between the Chromium(III)
center and the semiquinone ligand has been shown to be much greater the kBT at
room temperature, and that other operators in the spin Hamiltonian operate at

much smaller scales (less than 1 cm'l). The g and D tensors for the S = 1 and S =

2 states of this system are given by:

g] = “)4 gSQ + %gc, (4.14)
D1 = %DCr ’ %DSQ—Cr (4°15)
g2 = %gSQ + %gC, (4.16)
D2 = yzDCr + %DSQO (4.17)

The tensors SSQ , 8’0 and Do have already been determined in the electron spin

~

resonance studies of chapter 2. DSQ—Cr is the lone tensor that needs to be

determined. Therefore, successﬁrl collection and simulation of an EPR spectrum

302

Wraiiﬁj

 

for this complex allows one to effectively reconstruct the zero ﬁeld splitting of a

spin-exchange complex in a piece-wise fashion, speaking to the utility of the

various “control complexes” in this series.

 

* (4) CrGaCA3'
(4)CrGaCA3' Par. Mode
11 1 (6) CrZCA3'

 

(2) GaZCA3'

 

A

A

2.10'205'200'195' 1.90

W‘ .J\ n N, c 1 ,rw‘lvrvlfy ‘1'"
V “WWAJ‘JUV W‘ W ”V V

  
 

 

 

12 1O

 

8 ' 6 ' 4 ' 2

Effective g-value

 

 

Figure 4-21: Comparison of bimetallic semiquinone complexes, showing that the

signals of both the parallel mode and perpendicular mode spectra of complex 4

can be attributed to impurities in the sample. Experimental conditions: Complex

6: 9.4595 GHz, 0.30 mW, 4K; Complex 4 perpendicular mode: 9.6250 GHz, 62.8

mW, 4 K; Complex 4 parallel mode: 9.3941 GHz, 63 mW, 4 K; Complex 6:

9.4595 GHz, 0.63 mW.

303

[Cr2(tren)z(CA“t’°“)](BPh4)2 (5). Exchange coupled dimers of Chromium(III)

have been explored extensively in the literature.34 Common motifs include ion-

44-48 2,3,49,50

pairs in doped crystals and triply bridged u-hydroxo and halo-

bridged33’51'54 systems consisting of face-shared octahedra. These halo-bridged
complexes fall into the well-known family of A3Cr2X9 compounds (where A is a
singly charged cation and X is the halogen anion). The system presented herein

represents one of the few exchange-coupled Chromium(III) dimer systems that

does not fall into one of the above mentioned categories, other notable complexes
being [Cr2L3]3+ (L = 2,6-bis(aminomethyl)-4-tert-butyl-thiophenolate)4 and the
quinone-bridged [Cr2(CTH)(DHBQ)]2+ complex reported by Dei and coworkers.55

[Cr2(tren)2(CA°at’°at)](BPh4)2 (5) consists of two Cr'" nuclei bound by the

cat.cat

diamagnetic CA bridging ligand. Fitting of the magnetic susceptibility data of

5 indicates an antiferromagnetic interaction on the order of 4 cm"l (for
H =J1251°52 ), resulting in a singlet ground state and higher-lying states that are

populated at low temperatures. The 4 K X-band EPR spectrum of this complex is
presented in Figure 4-22. The ﬁgure shows that a majority of the peaks in this
spectrum can be attributed to the other Chromium(III) dimer system of this series,

sq,cat

complex 6, wherein the bridging ligand is in its CA form. A number sharp
features near gefr = 2 can be assigned as arising from 5, as well as perhaps several

broader features at lower ﬁelds.

304

 

 

 

 

 

 

 

 

 

' 1
0-9 ’ —— (5) CrZCA4' . ‘
3.
E 0.6 _ —"—”" (6) CIZCA -
8 . ' =
9
s. 0.3 - ‘ ~ ‘
.5 f x ‘ )1
2 ll W
-0.3 - ' ‘ '
_ J
-0.6 . 1 ' ‘ ' ' ' ' ‘ ‘
1O 8 6 4 2
1 Effective g-value
i
l

 

Figure 4-22: X-band perpendicular mode EPR spectrum of 5 at 4 K. This
spectrum is plotted against the corresponding X-band spectrum for 6, revealing
that many of the observed features in the spectrum of 5 can be attributed to

complex 6 as an impurity.

Applying the Kambe approximation to the Heisenberg coupling Hamiltonian of
equation 4.1 and considering only the isotropic component yields eigenvalues of
the form given by equation 4.2, where S, = 3, 2, 1, 0 and S, = S2 = 3/2. This yields
the energy-level diagram in the absence of zero-ﬁeld splitting effects, i.e. splitting
of the m8 sublevels. However, these effects are likely important to the description

of the ground state electronic structure because the splitting is expected to be on

305

 

the order of the coupling constant due to the integer spin nature of the states. To
determine which states contribute to the observed X-band signal, a Boltzmann
distribution analysis was carried out by integrating the EPR spectra and
monitoring characteristic peaks over the temperature range 4 - 100 K, and
comparing the intensities with the anticipated Boltzmann distribution given the
calculated energy levels. Thus far the results remain unclear, which is
unsurprising given the complicated nature of this spin system.

Exchange-coupled transition metal systems with |J| < kT can be simulated

according to the spin Hamiltonian

~

H=,661§-§.§+§.15-§+§,-J-§2 (4.18)

Ill

Exchange coupling between the two Cr centers results in Heisenberg spin

manifolds of St = 3, 2, 1, 0, and each spin manifold can have its own unique g-
tensor and zero-ﬁeld splitting parameters, due to different spin-orbit coupling
contributions within each state. Furthermore, in the weakly coupled regime there
likely exists extensive mixing between the various spin sublevels. At the moment
simulation of the data to obtain this information is not possible, but it is our hope
that higher-ﬁeld (W-band) spectra of this complex will allow for higher resolution

and thus easier characterization of each of the Heisenberg spin levels.

306

. ‘03.? Si.- L.;-II
I

 

 

 

 

 

 

 

 

1.0 - -
0.8 - =0 -

: =1

-.‘-3 =

g 0.6 - =3 -

Q.

0

CL

1., 0.4 - -

.2

E

32 0.2 - g -
0.0 - 4

 

 

 

0 20 40 60 80 100
Temperature (K)

 

 

 

Figure 4-23: Boltzmann distribution of spin states as a function of temperature

for complex 5 assuming an antiferromagnetic coupling of J = 4 cm'1 and no ﬁlrther

zero ﬁeld splitting of the rnS sublevels.

For this system the strong exchange limit approximation can also be made,
but any results should be taken with a grain of salt: the value of 4 cm'1 for the
isotropic exchange value in this system still represents the largest magnitude
interaction in the spin system, but is less than an order of magnitude larger than
the other forces operating. For example, the axial zero ﬁeld splitting of each

Chromium(III) center is 0.4 cm'], only an order of magnitude smaller than the J

307

value. A rigorous treatment would involve diagonalization of the full Hamiltonian

matrix.

[Cr2(tren)2(CAsq’°at)](BPh4)2(BF4) (6). As discussed previously, exchange
coupling arising from the interaction of the unpaired spin on the bridging ligand
with the S = 3/2 ground state of chromium(III), as in complex 4, is

antiferromagnetic in nature. The same is true for exchange coupling between two
Chromium(III) centers, as in complex 5, again due to the shared symmetry of the
HOMO orbitals. For [Cr2(tren)2(CAsq’°at)](BPh4)2(BF4) (6), which has both
interactions, all values of exchange interactions are antiferromagnetic, resulting in

56-60

a so-called spin-frustrated S = 5/2 ground state. The perpendicular mode X-

Band spectrum is presented in Figure 4-24 (and has been invoked in previous
sections of this chapter to account for impurities observed in the EPR spectra of
complexes 4 and 5). This spectrum was simulated assuming a sextet spin state,
with simulation parameters gxx = 1.98, gavy = 1.975, gZZ = 1.97; D = 0.458 cm], E =
0.1008 cm". The magnetic susceptibility data from 10-100 K corroborate this
assignment with a magnetic moment of neff= 5.60 113(Pspin-on1y for S = 5/2 is 5.92
113). Like many high-spin ferric complexes this spectrum reveals a resonance near
g’= 4.2,61'65 conﬁrming the sextet nature of the ground state of this complex.
Using a procedure described previously for this system,16 the eigenvalues for the

various spin states are given by:

308

 

‘

_—

E : %[ST(ST +1)_SA(SA +1)_ 53(53 +1)“

152—[SAGA +1)--S1(S1 +1)—S2(52 +1)] (4.19)

where J = JCr-SQ is the coupling between the Chromium(III) centers and the
semiquinone bridge, J * = JC,.Cr is the coupling between the chromium centers, S3

is the spin on the CASq’Cat bridge, s,= 7/2, 5/2, 1/2 and SA= s,+ s2, where 31 = 32 =

3/2. Magnetic susceptibility of compounds 4 and 5 implies values of J * = 4 cm'1

and |J| > kBT at 350 K. The magnitude of J effectively isolates the sextet ground

state with respect to EPR transitions. Variable temperature EPR data reveal a
simple decay in the peak intensities as a function of temperature, indicating no
population of higher energy spin states as the temperature increases, and further
supporting the high value of J. Simulation of the spectrum at 4 K was carried out
using the spin Hamiltonian of equation 2.5 (including electronic Zeeman and zero
ﬁeld splitting terms) and assuming an isolated sextet ground state. Perpendicular
and parallel mode spectra were simulated in this manner, and using the same
parameters (Figures 4-24 and 4-25). The energy-level diagrams along the x, y,
and z axes, presented in Figure 4-26, indicate extensive mixing among the various
spin states of the sextet manifold as a result of the small zero-ﬁeld splitting (D =

0.458 cm", E = 0.1008 cm'l). This is evidenced most clearly by the observed
parallel mode transition which is formally a Am, = + 4 (s = .5/2 —> s = 3/2)

transition.

309

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

a
r— j
. — (6), 4 K '.
f. .'
l- w— Simulation
- {"13" ﬂ
_ l .
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

BIT]

 

 

 

Figure 4-24: Perpendicular mode X-band spectrum of compound 6 (9.458 GHz, 4
K, 0.3 mW). The spectrum was simulated assuming an isolated sextet ground

state using the following parameters: gxx = 1.98, gyy = 1.975, gZZ = 1.97; D = 0.458

cm'], E = 0.101 cm].

310

 

 

 

 

 

’

 

 

il

x s __
.1 ____

Z

 

 

 

 

 

 

 

 

.-‘ l 0 "'
———- Simulation // K 1
. __ I. 7 ’ ._ \’\_r~ ‘V/~ v ’ \\
Z .
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
3m

 

 

 

Figure 4-25: Parallel mode X-band spectrum of compound 6 (9.394 GHz, 4 K,

3.16 mW). The spectrum was simulated assuming an isolated sextet ground state
using the following parameters: gxx = 1.98, gyy = 1.975, gZZ = 1.97; D = 0.458 cm'

‘,E=0.101cm".

311

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. = . -1 . E=0.100 cm'1 .
Isotroplc D 0458 c"; Ax1al = Rhomblc
remem— "“1 4g *- 4 E---__-__L__-
:2; a||0,(p 32 3:
5, 1? , 2? 2? a *
>. E l 1; 1 1
02>) Cl l 0E. 0 a: 111.11 .
LEI ‘11 j -1: ; _1 x '
0 0.2 0.4 026061“. ,___Aa.__..l 3,; a
B[T] A ; x-y plane 2; x
'7 t 1 - l
E -11. , -1E_
-21 . , -2‘
__ . L1 _LL ,_._______1
o 0.2 0.4 0.6 031.0 3:. ”an”,
BIT] A 2 Y ’
‘T 15 ,_
EDI-x *1 1*
3-1 N
I— ‘X
8-21 '
“J -3'e
0 0.2 0.4 0.6 0.81.0
Bfl'l

 

 

 

Figure 4-26: Energy level diagram of the ground state of complex 6 as a function

of magnetic ﬁeld, calculated from the simulation parameters.

4.5 Final Remarks.
Initial efforts toward the [Cr(tren)(3,6-R-Q)]+lH2 series, mainly synthetic in nature,

are presented in this chapter. This series was synthesized to systematically tune

the exchange coupling interaction while essentially retaining the same geometry

312

across the series. While some spectroscopic characterization of these complexes
has been carried out, the majority is yet to be done, and the reader is directed to
Chapter 5 for future directions on this project.

The ground states of the majority of the members of the bimetallic
[M1M2(tren)2(CA"')]m+ series have been characterized using X-band EPR
spectroscopy in this chapter as well as chapter 2. This series allows for a broad
diversity of ground states, ranging from the diamagnetic species 1 to the
paramagnetic sextet ground state of 6. Magnetic susceptibility indicate an
antiferromagnetic exchange interaction between the chromium spin centers of -2
cm", although this has not yet been conﬁrmed by EPR spectroscopy for
compounds 5 and 6. The exchange interaction between the chromium spin centers
and the CASq’cat bridge is large enough to isolate the ground state at X-band
frequency and the temperatures at which the experiments were carried out.
Magnetic susceptibility conﬁrms an isolated triplet state and an isolated sextet
state for 4 and 6, respectively, although this remains unconﬁrmed by X-band EPR
spectroscopy for complex 4 as a result of exceedingly large zero-ﬁeld splitting

within the triplet state. Future work on this area will likely require high-ﬁeld

high-frequency EPR techniques.

313

 

4.6 References

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

(14)

Benelli, C.; Dei, A.; Gatteschi, D.; Gudel, H. U.; Pardi, L. Inorganic
Chemistry 1989, 28, 3089-3091.

Schenker, R.; Weihe, H.; Gudel, H. U. Chemical Physics Letters 1999, 303,

229-234.

Schenker, R.; Weihe, H.; Gudel, H. U. Inorganic Chemistry 1999, 38,
5593-5601.

Schenker, R.; Weihe, H.; Gudel, H. U.; Kersting, B. Inorganic Chemistry
2001, 40, 3355.

Schrauben, J. N.; Guo, D.; McCracken, J. L.; McCusker, J. K. Inorganica
Chimica Acta 2008, 361, 3539-3547.

Schrauben, J .; Guo, D.; McCracken, J .; McCusker, J. Inorganic Chemistry,
Submitted 2010.

Bencini, A.; Daul, C. A.; Dei, A.; Mariotti, F.; Lee, H.; Shultz, D. A.;
Sorace, L. Inorganic Chemistry 2001, 40, 1582-1590.

Dei, A.; Gatteschi, D.; Pardi, L.; Russo, U. Inorganic Chemistry 1991, 30,
2589-2594.

Min, K. S.; DiPasquale, A. G.; Golen, J. A.; Rheingold, A. L.; Miller, J. S.
Journal of the American Chemical Society 2007, 129, 2360-2368.

Piskunov, A. V.; Cherkasov, V. K.; Druzhkov, N. 0.; Abakumov, G. A.;
Ikorskii, V. N. Russian Chemical Bulletin, International Edition 2005, 54,
1627- 163 1 .

Kremer, S. Inorganic Chemistry 1985, 24, 887-890.

Glerup, J .; Goodson, P. A.; Hodgson, D. J .; Masood, M. A.; Michelson, K.
Inorganica Chimica Acta 2005, 358, 295-302.

Bolster, D. E.; Giltlich, P.; Hatﬁeld, W.; Kremer, S.; Miiller, E. W.;
Wieghardt, K. Inorganic Chemistry 1983, 22, 1725-1729.

Graebe Justus Liebigs Annalen der Chemie 1868, 146, 31.

314

...-.._...,..1

 

(15)

(16)
(17)
(18>
<19)

(20)

(21)
(22)

(23)
(24)

(25)

(26)

(27)
(28)
(29)

(30)

(31)

Weider, P. R.; Hegedus, L. S.; Asada, H.; D'Andreq, S. V. Journal of
Organometallic Chemistry 1985, 50, 4276.

Guo, D.; McCusker, J. K. Inorganic Chemistry 2007, 46, 3257-3274.

Zipp, S. G.; Madan, S. K. Inorganic Chemistry 1976, 15, 587.

Wheeler, D. E.; McCusker, J. K. Inorganic Chemistry 1998, 37, 2296-2307.
Fehir, R. J ., Ph.D. Thesis, Michigan State University, 2009.

Loudon, J. D.; Steel, D. K. V. Journal of the Chemical Society 1954, 1163-
l 164.

Oxford, A. B. Journal of the Chemical Society 1926, 2004-2011.

 

Heertjes, P. M.; Nijman-Knape, A. A.; Talsma, H.; F aasen, N. J. Journal of
the Chemical Society 1955, 1313-1316.

Bencini, A.; Gatteschi, D. EPR of Exchange Coupled Systems; Springer:
New York, 1990.

Anderson, P. W. Solid State Physics-Advances In Research And
Applications 1963, 14, 99-214.

Goodenough, J. B. Physical Review 1955, 100, 564-573.

Goodenough, J. B. Journal Of Physics And Chemistry Of Solids 1958, 6,
287-297.

Kanamori, J. Journal Of Physics And Chemistry Of Solids 1959, 10, 87-98.
Heisenberg, W. Zeitschrift Fur Physik 1926, 38, 411-426.

Dirac, P. A. M. Proceedings Of The Royal Society Of London Series A-
Containing Papers Of A Mathematical And Physical Character 1926, 1 12,
661-677.

Bencini, A.; Gatteschi, D. EPR of Exchange Coupled Systems; Springer:
New York, 1990.

Benelli, C.; Dei, A.; Gatteschi, D.; Pardi, L. Inorganic Chemistry 1988, 27,
2831-2836.

315

(32) Kambe, K. J. Phys. Soc. Japan 1950, 5, 48.

(33) Schenker, R.; Heer, S.; Gudel, H. U.; Weihe, H. Inorganic Chemistry 2001,
40, 1482-1488.

(34) McCarthy, P. J .; Gudel, H. U. Coordination Chemistry Reviews 1988, 88,
69-13 1.

(3 5) Juban, 13., PhD. Thesis, University of California, 2006.

(36) Wang, G.; Long, X.; Zhang, L.; Wang, G.; Polosan, S.; Tsuboi, T. Journal
of Luminescence 2008, 128, 1556.

(37) Guedel, H. U.; Snellgrove, T. R. Inorganic Chemistry 1978, 17, 1617.

(38) Oliver, S. W.; Hans, U. G. The Journal of Chemical Physics 2001, 114,
5832-5841.

(39) Oliver, S. W.; Rafael, V.; Hans, U. G. The Journal of Chemical Physics
2001, 115, 3819-3826.

(40) Beaulac, R. m.; Reber, C. Inorganic Chemistry 2008, 4 7, 5048.

(41) Sofen, S. R.; Ware, D. C.; Cooper, S. R.; Raymond, K. N. Inorganic
Chemistry 1979, 18, 234.

(42) Wheeler, D. E.; Rodriguez, J. H.; McCusker, J. K. Journal Of Physical
Chemistry A 1999, 103, 4101-4112.

(43) Buchanan, R. M.; Claﬂin, J .; Pierpont, C. G. Inorganic Chemistry 1983, 22,
2552-2556.

(44) Naito, M. Journal Of T he Physical Society Of Japan 1973, 34, 1491-1502.
(45) Van der Ziel, J. P. Physical Review B 1974, 9, 2846-2862.

(46) Van der Ziel, J. P. Journal Of Chemical Physics 1972, 5 7, 2442-&.

(47) Van der Ziel, J. P. Physical Review B 1971, 4, 2888-&.

(48) Van Gorkom, G. G. P.; Henning, J. C. M.; Van Stapele, R. P. Physical
Review B 1973, 8, 955-973.

316

 

(49)

(50)

(51)

(52)

(53)

(54)

(55)

(56)

(57)

(58)

(59)

(60)

(61)

(62)

Glerup, J .; Larsen, 8.; Weihe, H. Acta Chemica Scandinavica 1993, 47,
1154-1161.

Riesen, H.; Gudel, H. U. Molecular Physics 1987, 60, 1221-1244.

Dubicki, L.; Ferguson, J .; Harrowﬁeld, B. V. Molecular Physics 1977, 34,
1545-1561.

Dean, N. J .; Maxwell, K. J .; Stevens, K. W. H.; Turner, R. J. Journal Of
Physics C-Solid State Physics 1985, 18, 4505-4519.

Johnstone, I. W.; Maxwell, K. J .; Stevens, K. W. H. Journal Of Physics C-
Solid State Physics 1981, 14, 1297-1312.

Briat, B.; Russel, M. F.; Rivoal, J. C.; Chapelle, J. P.; Kahn, 0. Molecular
Physics 1977, 34, 1357-1389.

Dei, A.; Gatteschi, D.; Pardi, L.; Russo, U. Inorganic Chemistry 1991, 30,
2589.

Guo, D.; McCusker, J. K. Inorg. Chem. 2007, 46, 3257-3274.

McCusker, J. K.; Vincent, J. B.; Schmitt, E. A.; Mino, M. L.; Shin, K.;
Coggin, D. K.; Hagen, P. M.; Huffman, J. C.; Christou, G.; Hendrickson,
D. N. Journal Of T he American Chemical Society 1991, 113, 3012-3021.
McCusker, J. K.; Christmas, C. A.; Hagen, P. M.; Chadha, R. K.; Harvey,
D. F .; Hendrickson, D. N. Journal Of T he American Chemical Society
1991, 113, 6114-6124.

Libby, E.; McCusker, J. K.; Schmitt, E. A.; Folting, K.; Hendrickson, D.
N.; Christou, G. Inorganic Chemistry 1991, 30, 3486-3495.

McCusker, J. K.; Jang, H. G.; Wang, S.; Christou, G.; Hendrickson, D. N.
Inorganic Chemistry 1992, 31 , 1874-1880.

Wickman, H. H.; Klein, M. P.; Shirley, D. A. Journal of Chemical Physics
1965, 42, 2113-2117.

Pilbrow, J. R. Journal of Magnetic Resonance 1978, 31, 479-490.

317

 

(63)

(64)

(65)

Troup, G. J .; Hutton, D. R. British Journal of Applied Physics 1964, 15,
1493-1499.

Dowsing, R. D.; Gibson, J. F. Journal of Chemical Physics 1969, 50, 294-
303.

Aasa, R. Journal of Chemical Physics 1970, 52, 3919-3930.

 

 

318

Chapter 5: Future Work

5.1 Introduction

The results presented in this dissertation represent only the beginning of a
potentially very rich ﬁeld of study. Future studies will beneﬁt from enhanced
capabilities in the McCusker group laser lab, namely a new ultrafast system that
will offer ~ 35 fs pulses in the visible, as well as tunabiltity into the infrared.
Furthermore, our current ~100 fs system is being enabled with a longer delay line
to achieve longer timescale studies and overcome the “dark window” of 1 — 10 ns
that has plagued our research. This chapter, much like the dissertation as a whole,
is broken into two general subsections: high-symmetry complexes and low-
symmetry, spin-exchange complexes. While study of the high-symmetry
complexes will beneﬁt most immediately from the enhanced technological
capabilities of our laser lab, much more ﬁmdamental work on understanding the
electronic structure and spin distribution of spin-exchange complexes is required
before meaningful dynamical studies can be performed. Speciﬁc future directions

of these projects are presented below.

5.2 High-Symmetry Complexes
Of the work on the high symmetry complexes (chapter 3), the most interesting and
potentially revealing data was uncovered employing 50 fs pulses, resulting in

coherent oscillations. These coherent oscillations may be providing insight into the

319

 

reaction coordinate of the ultrafast intersystem crossing in this system. In order to
more fully characterize this phenomenon, future work must expand upon the
pump/probe combinations employed in chapter 3. Vibrational spectroscopies,
namely resonance Raman spectroscopy, could be employed as well to more clearly
assign the nature of the coherence. While these measurements may be quite
difﬁcult due to the small enhancement afforded by the ligand ﬁeld absorption, this
method would provide the clearest characterization of the 4T2 vibrational structure.
In the absence of resonance Raman data, TD-DFT techniques must be employed
to gain access to the excited state vibrational frequencies. An example of this
technique is shown in Figure 5-1.

Of course, as shown in Chapter 3, synthetic modiﬁcations to the
acetylacetonate ligand backbone can change the dynamics. In order to test the
impact of ligand vibrational modes on ultrafast intersystem crossing dynamics,

coherence measurements must be carried out on heavier analogs of Cr(acac)3. The
ﬁrst “heavy” model for Cr(acac)3 will be Cr(d7-acac)3, synthesized via the route of
Figure 5-2. Other potential heavy models for Cr(acac)3, some of which have
already been studied, are shown in Figure 5-3. Ground state DF T frequency
calculations of some isotopically-enriched heavy Cr(acac)3 analogs at the
UB3LYP/6-311g** level, employing a CPCM solvent model of acetonitrile are
1

showin in Figure 5-4. Key vibrational modes between 150 cm'1 and 300 cm' ,

which may correspond to the vibrational mode observed in the coherence data of

320

F.;- I‘m-1-]

 

 

 

 

 

 

 

Figure 5-1: The lowest-energy spin-allowed transition for Cr(acac)3, as
calculated at the UB3LYP/6-311g** level (employing a CPCM solvent model for
acetonitrile) reveals that 920i and 9301 molecular orbitals (major contributors to the
HOMO) have lt-antibonding symmetry while the 9701 and 98 01 orbitals (major
contributors to this transition) have o-antibonding symmetry. This is qualitatively
in line with a transition resulting in population of an eg* orbital. Geometry
minimization of these MOs effectively constitutes the 4T2 geometry, upon which a

frequency calculation can be carried out.

321

warm

 

Cr(acac)3 presented in Chapter 3, are highlighted. These values are referenced to
the low-frequency portion of Raman spectrum of Cr(acac)3, collected by
Alexandre Rodrigue-Witchel of the Reber group at Université de Montreal.
Again, due to the likely low enhancement of resonance Raman, the key to
characterizing these complexes will the comparison of coherence data to low-

frequency Raman spectra, as well as DFT techniques.

 

OD

O O O
+
030 0 D30 co3

oac co3

 

 

D

 

Figure 5-2: Proposed synthesis of dg-acac.

 

O OH

0 OD \/U\/]\/

/

D30 CD3 0 0”
D /
18o 180D
0 OH
03c / co3
/
D

 

 

 

Figure 5—3: “Heavy” analogs of Cr(acac)3.

322

 

 

I ' I ' l

45.7 ‘

 

 

 

 

 

)—
Sym. Scissor _
‘ Breath. ' symmetrlc
f ‘ breathing \
, Exp 188 256 .
+F Gale
1
H 184 254
E - 2H 178 229
C . 18
9 o 181 250
E

 

 

 

 

 

100 200 300 400 500
1 Energy (cm'1)

 

Figure 5-4: Ground state DF T frequency calculations for Cr(acac)3 and

isotropically-enriched analogs of Cr(acac)3, referenced to the Raman spectrum.

Further studies could make use of more drastic synthetic modiﬁcations to
produce vibrationally constrained ligands, in the hope of experimentally
determining the coherent mode through suppression of this mode. One possible
series would be the hexaamine complexes which were thoroughly studied
throughout the latter third of the twentieth century, most notably by Endicott and
coworkers."4 For example, a series of increasingly constrained ligands can be
envisioned by systematically bonding the ligands to one-another (Figure 5-5).

One potential drawback of this series is the lack of ligand rt-structure, leading to

323

no charge-transfer transition for these complexes. This will likely lead to a very
small transient absorption signal: in Cr(acac)3-type complexes the excited state
absorption corresponds to a ligand ﬁeld to charge transfer transition, hence the
positive excited state absorption. To overcome this obstacle, one could envision
employing constraining ligands with an acetylacetonate motif so that the charge-
transfer manifold of states remains (Figure 5-6). The series presented in Figure 5-

6 systematically suppresses the scissor mode (observed near 250 cm”) while
affecting the symmetric breathing mode (near 190 cm'l) via the additional mass of

the ligand, lowering the frequency. While both modes are affected, the
suppression of the scissor mode should be more dramatic, allowing one to parse

out what mode is active in excited state dynamics.

 

NH NHHN

H N NH HN
2 \/\NH2 H NH NHHN
N

 

More constraining

 

 

 

Figure 5-5: [N6-chromium(III)]3+ series of variable ligand rigidity.

324

 

 

O OH

M /

More constraining

 

 

 

 

Figure 5-6: A series of increasing constrained acetylacetonate-type complexes.

Another potential future direction of this project is direct observation of
vibrational relaxation using ultrafast IR techniques. The best starting point would
be an attempt to reproduce the results of Kunttu5 on Cr(acac)3. Ultrafast IR
studies on substituted complexes may be especially fruitﬁll. Many of the 3-
substituted complexes would be useful for such a study as they offer unique
vibrational modes (C-Br, N02, SCN, etc) that could be used as IR tags to
potentially witness the vibrational redistribution, and address the role of the 3-
substituted position in the vibrational relaxation of these complexes. Once ligand
ﬁeld dynamics are worked out, excitation in charge-transfer and ligand localized
transitions can be explored. Rapid decay to the ligand ﬁeld manifold of electronic
states is expected, and many of the same questions outlined above must be again
asked, but with respect to ultrafast dynamics between charge-transfer (or ligand
localized) and ligand ﬁeld manifolds. Another potential avenue of study is to

address whether these modes of excited state deactivation are generally applicable

325

to other metal-acetylacetonate complexes, with potential applications to dye-

sensitized solar cells.

5.3 Low-Symmetry and Spin-Exchange Complexes. This project can be
broken up into two distinct yet interconnected studies: 1) ground state
characterization and manipulation of spin polarization and 2) excited state
electronic structure and dynamics. The work presented in Chapter 2 of this
dissertation, as well as the work of F ehir,6'8 provides the starting point for the ﬁrst
half of this project, revealing guidelines for manipulation of spin distribution in
these systems as well as the experimental techniques for veriﬁcation of spin
distribution.

The ground state characterization of the spin exchange complexes was
discussed in Chapter 4. For strongly coupled systems, a better characterization of
ground states will include high-ﬁeld high-frequency EPR, and temperature-
dependent EPR studies. Simulations of high-ﬁeld EPR spectra employing the spin
Hamiltonian parameters of low-symmetry chromium-catechol complexes suggest
that high-ﬁeld “ﬁeld sweep” experiments would be successful in elucidating spin
Hamiltonian parameters for these spin-exchange complexes. Of course, these
studies will likely do little to address the magnitude to the Heisenberg exchange

constant, which has been shown to be larger than kBT for [Cr(tren)(3,6-di-tert-

butylcatechol)](PF6).9 Variable temperature magnetic susceptibility measurements

326

must be carried out to address this. In the absence of conclusive variable
temperature magnetic susceptibility data, DF T methods must be applied to
estimate J.

The second, kinetic aspect of the project remains in its infancy. Further
efforts on this project must begin with characterization of the excited state
electronic structure of the low-symmetry chromium-catechol complexes. This can
most effectively be carried out using polarized single crystal absorption
spectroscopy. Initial studies of this sort can be carried out on the [Cr(tren)(3,6-R-
Q)]+1 series, which has already been initially characterized spectroscopically in
Chapter 4. Furthermore, it has been found that relatively large single crystals can
be grown for members of this series by slow evaporation of the complex in a
methanol/NaBPh4 solution. As discussed in Chapter 4, transitions between low-
symmetry components can be characterized by their allowedness in polarized
light. Once a grasp of the electronic structure is obtained, dynamical studies can
be carried out. This study, wherein electronic structure is varied while
maintaining the same coordination sphere about the chromium ion, would
represent a systematic study of the effects of lowered symmetry on ultrafast
dynamics of transition metal systems.

The semiquinone analogs of these complexes can be prepared via one
electron oxidation of the parent catechol complexes. The electrochemical data
presented in Chapter 4 should provide some clear choices of oxidants for this

purpose. Again, the ﬁrst crucial step to understanding the time-resolved dynamics

327

 

 

 

is elucidation of the excited electronic structure of these complexes. This will
likely involve polarized single-crystal absorption spectroscopy, a detailed
understanding of chromium-catechol electronic structures, and some additional
help from TD-DFT techniques. These same techniques can be applied to the spin-
exchange complexes of the [M1M2(tren)2(CA"')]m+ series. The study of these
spin-exchange complexes will in all likelihood represent one of the most well-
rounded projects to be found in the McCusker group, involving synthesis,

magnetism, computational work and time-resolved spectroscopies.

328

5.4 References

(1)

(2)

(3)
94,

(4)
30,

(5)

(6)
(7)

(8)

(9)

Endicott, J. F .; Tamilarasan, R.; Lessard, R. B. Chemical Physics Letters
1984,112, 381-386.

Endicott, J. F.; Lessard, R. B.; Lynch, D.; Perkovic, M. W.; Ryu, C. K.
Coordination Chemistry Reviews 1990, 97, 65-79.

Perkovic, M. W.; Endicott, J. F. The Journal of Physical Chemistry 1990,
1217.

Perkovic, M. W.; Heeg, M. J .; Endicott, J. F. Inorganic Chemistry 1991,
3140-3147.

Macoas, E. M. S.; Kananavicius, R.; Myllyperkio, P.; Pettersson, M.;
Kunttu, H. Journal of the American Chemical Society 2007, 129, 8934-
8935.

Fehir, R. J ., Ph.D Thesis, Michigan State University, 2009.

Fehir, R. J.; McCusker, J. K. Journal Of Physical Chemistry A 2009, 1 13,
9249-9260.

Fehir, R. J .; Krzyaniak, M.; Schrauben, J. N.; McCracken, J. L.; McCusker,
J. K. Journal of Physical Chemistry B 2010, to be submitted.

Rodriguez, J. H.; Wheeler, D. E.; McCusker, J. K. Journal of the American
Chemical Society 1998, 120, 12051-12068.

329

APPENDICES

330

Appendix A: Additional Figures for Chapter

1

 

 

 

 

 

 

 

 

80 4T“? 2A29
70. 2A1g
60. 4T19
50‘ 4T29
m
LU 40.
2F 2T2g
30- ,
94’! 2T19
20 . r77" 1 2
10 ‘
3F . . . . 4Azg
0 10 20 30 40 5O
AO/B

 

Figure A-l: Tanabe-Sugano diagram of a d3 system.

331

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3
8O 1’A‘293’A‘10 A295E91Eg
3 1AZg /
7 A
0 2g 3T 5
1 29 T29
3?:0 1T2g
O .1
L” 40‘ T29
16, 1| 3T1g
3F, 3%)
3H 20 .
10« 1A19
5D 51' '
O .' 2 . . 1A19
0 10 20 30 40 50
AO/B

 

Figure A-2: Tanabe-Sugano diagram of a (16 system.

332

 

Appendix B: Development of Characteristic Equations for Isolated Spin
Systems
Considering only the electronic Zeeman and zero-ﬁeld splitting interactions, the

following Hamiltonian is used:
H=neags+sbs

The tensor formalism can be avoided by expanding the Hamiltonian to include the
diagonal components of the tensor and employing the direction cosines. The zero
ﬁeld splitting tensor is split into D and E, as described in Chapter 2 of this

dissertation.

H = Be [ngzzSZ + lgxx Sx + mgnyy] +
D[S§ — %S(S +1)]+ %E[S§ + SE]
Where the direction cosines are:

n =cos6
l=sinl9cos¢
m =sin68in¢

The spin raising and lowering operators are then substituted into the Zeeman

component:

333

ll.

 

 

 

H 2 Be [ngzzSz + %lgxx S+ +%lgxx S_ —y2mgny+ +y2mgny_]

+ D[§§ — %S(S+ 1)] + %E[sﬁ +83]

 

This can be reduced to the ﬁnal, usable form

.23

A

H = ﬂe [ngzzSz + g_S+ + g-l-S :|+

D[SZ2 —%S(S +1)]+ %EF‘E + SE]

 

Where

8_ =%(lgxx-imgw)

8+ = %(lgxx‘i' imgyy)

Spin eigenfunctions are represented as comprised of the total spin and all of the

28+] spin components, i.e. IS, m) For example, for a triplet spin state, the three

eigenfunctions are:

1,1)
1,—1)
1,0)

 

334

The interactions of the various rnS levels are setup in a square matrix and the spin

operators are applied in the usual fashion to yield the secular determinant. For a

triplet state this equation (for the given Hamiltonian) is:

 

 

|1,l) |1,0) 1,—1)
(I’ll gzereBn + %D — 6‘ Tlg'g—ﬂeBm E
(1,0 $8.443"? -%D - 6 ﬁzz—ﬂew = 0
<1’_1| E $84,423"? -gzzﬂeBn +-;1i,-D - 8

 

 

For a quartet spin state the characteristic equation for this Hamiltonian is:

335

 

 

 

claimanamml
1 8 -ml
name
as

o

Amish

 

 

Jamaawms.
olelcmsaamwl
lance
mm,
«Tm _

mm,
ﬁance

elqlcmomamm.

61Q+=msman+ 61Q+cmsmswm+

ARE

o

as

Jase. .%

ARE

 

Kum
«ﬁx
«4.x
$8

V
336

VVV

Appendix C. Crystallographic Data for Chapters 3 and 4

 

C16

 

011 \
C13

 

 

 

 

Figure C-l: Crystal structure for Cr(3-NOzac)3.

Table C-l: Select bond lengths and angles for Cr(3-NOzac)3.

 

 

 

Cr1-O1 1 .9450

Cr-O Cr1-02 1.9414
Bonds Cr1-03 1.9428
(A) Cr1-O4 1 .9530
Cr1-O5 1 .9551

Cr1-O6 1 .9416

O1-Cr1-O4 179.63

02-Cr1-05 177.68

O—Cr-O O3-Cr1-O6 178.11
Angles O1-Cr1-05 91 .32
(°) O1-Cr1-O6 87.77
02-Cr1-O4 90.94

02-Cr1-O6 88.59

 

 

 

 

 

337

Table C-2: Crystallographic data for Cr(3-NOzac)3 and Cr(3-Phac)3.

 

Cr(3-No,ac)3

Cr(3-Phac)3

 

 

Formula
Mw
cryst syst
space group
T/ K
a/A
b/A
c/A

Reﬂns measured

Indepdendent reﬂns
Observed reﬂns [I>20'(I)]

p(Mo K01)/ cm-l
Rim
R1

00R2

GOF

C H CrN3o12

15 18
484.32

293
33.0109 (4)
33.0109 (4)
7.5242 (2)
90
90
90
8199.3 (3)
16
1.569
56.6
26701

5059
4786
0.71073
0.032
0.040

0.110
1.09

C991—199(31'3018
1732.78
Monoclinic
C2/c
173
19.0021 (4)
13.8617 (3)
34.0764 (7)

91.4740 (10)

8972.8 (3)
4
1.283
56.6
54629

1 1143
9402
0.71073
0.061
0.056

0.165
1.03

 

 

w = 1/[02(F02) + (0.0989P)2 + 0.8945P]

where P = (F02 + 2Fc2)/3

 

338

 

 

 

Table C-3: Crystallographic data for Cr(DBM)3 and Ga(DBM)3.

 

 

 

 

 

Cr(dbm)3 Ga(dbm)3
Formula C22.SOH16.50cr 0.5003 C45H33Ga06
Mw 360.86 739.43
cryst syst
space group
T/K 173(2) 173(2)
a/A 17.0031(2) 9.7821(3)
b/A 9.77730(10) 10.4784(3) ’
c/A 21.2772(2) 18.1320(5)
01/0 90 93.3500(10)
[3/0 92.6920(10) 93.9350(10)
y/0 90 106.223 0(1 0)
V/A3 3533.31(6) l774.48(9)
Z 8 2
Dc/g cm1 1.357 1.384
26*)”,ax 68.15 68.27
Reﬂns measured 27409 23613
Indepdendent reﬂns 6386 6260
Observed reﬂns
[I>20(I)] 5660 6113
p(Cu K01)/ cm:l 1.54178 1.54178
Rint 0.0308 0.025
R1 0.0376 0.0305
(0R2 0.1022 0.0804
GOF 1.032 1.052
w = 1/[02(F02) + (0.0989P)2 + 0.8945P]
where P = (F02 + 2F02)/3

 

339

 

 

Table C-4: Crystallographic data for [Cr(tren)(pycat)](BPh4) and [Cr(tren)(3,6-

1 ,2-orthocatecholate)](BPh4).

 

 

l

[Cr(tren)(pycat)]+ [Cr(tren)(CNcat)]+
Formula C38H42BCrN404 C 40H43BCrN6O2
Mw 681.57 716.62
cryst syst Monoclinic
space group P2(1)/c
T/K 293 173
a/A 9.8957 (2) 18.4129 (3)
b/A 10.0280 (2) 10.7008 (2)
c/A 19.2369 (3) 18.8264 (3)
01/0 92.7190 (10)°
[3/0 99.0650 (10)° 104.4090 (10)°
y/O 106.6950 (10)°
V/A3 1796.96 (6) 3592.73 (11)
Z 2 4
Dc/g cm'1 1.260 1.325
269max 55 55
Reﬂns measured 38895 40650
Indepdendent reﬂns 821 1 8242
Observed reﬂns
[I>26(I)] 7007 6789
p(Mo K01)/ A 0.71073 0.71073
Rint 0.057 0.084
R1 0.061 0.072
0)R2 0.183 0.202
GOF 1.01 1.08

 

 

w = 1/[02(F02) + (0.0989P)2 + 0.8945P]

where P = (F 02 + 2F c2)/3

 

340

 

 

 

MICHIGAN STATE UNIVERSITY LIBRARIE
3 1293 03306 0090