Computational, Synthetic, and Spectroscopic Investigations of Molecular
Switches, their Effects on Pendant Groups, and their Abilities to Form and Affect
Measurable Changes in Supramolecular Assemblies
by
Miranda C. Andrews, B.S.
A Dissertation
In
Chemistry
Submitted to the Graduate Faculty
of Texas Tech University in
Partial Fulfillment of
the Requirements for
the Degree of
DOCTOR OF PHILOSOPHY
Approved
Anthony F. Cozzolino
Chairperson of the Committee
Dominick Casadonte
Kristin Hutchins
Joshua D. Howe
Mark Sheridan
Dean of the Graduate School
December 2019
c 2019, Miranda C. Andrews
Texas Tech University, Miranda C. Andrews, December 2019
ACKNOWLEDGMENTS
I would like to thank and acknowledge my advisor, Dr. Anthony Cozzolino for
his expertise and guidance throughout my time in graduate school and during the
writing of this dissertation. I would especially like to thank him for taking a chance
on someone with very little research experience and for always encouraging me to
learn new techniques and ask good questions.
I would like to thank Drs. Dominick Casadonte and Kristin Hutchins for serving
on my committee and also for their suggestions and help during my time at Texas
Tech. I would also like to thank Dr. David Birney for serving on my committee for
several years as well as for his help and encouragement, particularly with providing alternative perspectives for some of the questions I pursued.
I would like to thank my colleagues, both past and present, that have served as
excellent sounding boards whenever I encountered problems or new questions in
my research. I am very fortunate to have worked with fellow scientists who have
provided new perspectives and gentle criticism when I have asked for help.
There are several facilities and organizations that I would like to thank: Dr. Daniel
Unruh and the X-ray Facility and Drs. Piotr Dobrowolski and Mike Mayer and the
NMR Facility for the important assistance in structural characterization, and the
Chemistry Graduate Student Organization for assistance with traveling to conferences to present my work and to network.
I would like to thank my husband, Dr. Keller Andrews, whose help has been
invaluable in helping me to be a better communicator, a better scientist, a better
teacher, and a better person.
Lastly, I would like to thank W. Harold and Geneva Harrell Chambliss, Devon D.
Thrift, and Ozeala Crawford; I dedicate this dissertation to their memories. I was
encouraged to tinker and to get my hands dirty as well as the value of working
hard while I was growing up. My connections to this generation have been a crucial part of who I am.
ii
Texas Tech University, Miranda C. Andrews, December 2019
CONTENTS
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
ii
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
ix
List of Compounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
x
List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvi
List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvii
1. Introduction to Molecular Switches . . . . . . . . . . . . . . . . . . . .
1.1
1
Molecular Switches . . . . . . . . . . . . . . . . . . . . . . . . .
1
1.1.1 Photoswitches . . . . . . . . . . . . . . . . . . . . . . . . . .
2
1.1.1.1
Coordination Complexes Coupled with Photoswitches
5
1.1.2 Thermoswitches . . . . . . . . . . . . . . . . . . . . . . . .
7
1.1.3 Electroswitches . . . . . . . . . . . . . . . . . . . . . . . . .
9
1.1.4 Chemoswitches . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.1.5 Mechanoswitches . . . . . . . . . . . . . . . . . . . . . . . . 12
1.2
Effect of Pendant Groups on Photoswitchable Behavior of Fulgimides and Spiropyrans . . . . . . . . . . . . . . . . . . . . . . 13
1.3
Purpose, Scope and Overview of the Dissertation . . . . . . . . 14
2. Remote Switching of a Ligating Group . . . . . . . . . . . . . . . . . . 16
2.1
Synthesis and Structural Characterization . . . . . . . . . . . . 18
iii
Texas Tech University, Miranda C. Andrews, December 2019
2.2
Photochemical Behavior . . . . . . . . . . . . . . . . . . . . . . 21
2.2.1 Photochemical Stability . . . . . . . . . . . . . . . . . . . . 24
2.2.2 Thermochemical Behavior . . . . . . . . . . . . . . . . . . . 28
2.3
Electrochemical Behavior . . . . . . . . . . . . . . . . . . . . . . 30
2.3.1 Photoelectrochemistry . . . . . . . . . . . . . . . . . . . . . 30
2.3.2 Isolation of Photoswitched Compounds . . . . . . . . . . . 33
2.3.3 Spectroelectrochemistry . . . . . . . . . . . . . . . . . . . . 33
2.4
Future Work: Synthetic Targets and Strategies . . . . . . . . . . 34
2.4.1 Amidines, Aminopyridines, and Squaramides . . . . . . . 35
2.5
2.4.1.1
Amidines . . . . . . . . . . . . . . . . . . . . . . . . . 37
2.4.1.2
Aminopyridines . . . . . . . . . . . . . . . . . . . . . 38
2.4.1.3
Squaramides . . . . . . . . . . . . . . . . . . . . . . . 40
Synthetic and Experimental Methods . . . . . . . . . . . . . . . 45
2.5.1 Isolation of 1(c) . . . . . . . . . . . . . . . . . . . . . . . . . 45
2.5.2 Isolation of 2(o) . . . . . . . . . . . . . . . . . . . . . . . . . 45
2.5.3 Experimental Procedure and Data Fitting for Determination of Activation Energy of Isomerization of 2 . . . . . . . 48
2.5.4 Attempted Synthesis of 7(c) . . . . . . . . . . . . . . . . . . 50
2.5.5 Synthesis of 8 . . . . . . . . . . . . . . . . . . . . . . . . . . 51
2.5.6 Attempted Deprotonation and Subsequent Metallation of 8 51
iv
Texas Tech University, Miranda C. Andrews, December 2019
2.5.6.1
Single Crystal X-ray Diffraction . . . . . . . . . . . . 54
2.5.7 Synthesis of 9(c) . . . . . . . . . . . . . . . . . . . . . . . . . 56
3. Computational Studies of Molecular Switches as Ligands in Metal
Complexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
3.1
Metal Complexes . . . . . . . . . . . . . . . . . . . . . . . . . . 58
3.2
Calculations of M2+ Complexes . . . . . . . . . . . . . . . . . . 59
3.2.1 Monometallic Complexes . . . . . . . . . . . . . . . . . . . 60
3.2.2 Spin Crossover . . . . . . . . . . . . . . . . . . . . . . . . . 66
3.2.3 Future Endeavors . . . . . . . . . . . . . . . . . . . . . . . . 70
4. Towards Switchable Halogen Bonding . . . . . . . . . . . . . . . . . . . 71
4.1
Remote Photoswitching of Halogen Bonding . . . . . . . . . . 73
4.1.1 Crystal Structures . . . . . . . . . . . . . . . . . . . . . . . . 77
4.1.2 Switchable Behavior of Compound 11 . . . . . . . . . . . . 78
4.1.3 Future Endeavors . . . . . . . . . . . . . . . . . . . . . . . . 80
4.2
Toward Remote Electroswitching of Halogen Bonding . . . . . 80
4.2.1 BODIPY-based Switches . . . . . . . . . . . . . . . . . . . . 80
4.2.2 Computational Studies of BODIPY Halogen Bond Donors
81
4.2.3 Crystal Structures . . . . . . . . . . . . . . . . . . . . . . . . 82
4.2.4 Future Endeavors . . . . . . . . . . . . . . . . . . . . . . . . 85
4.3
Synthetic and Experimental Methods . . . . . . . . . . . . . . . 86
v
Texas Tech University, Miranda C. Andrews, December 2019
4.3.1 Synthesis of 10(c) . . . . . . . . . . . . . . . . . . . . . . . . 86
4.3.1.1
Single Crystal X-ray Diffraction . . . . . . . . . . . . 87
4.3.2 Synthesis of 11(c) . . . . . . . . . . . . . . . . . . . . . . . . 90
4.3.2.1
Single Crystal X-ray Diffraction . . . . . . . . . . . . 93
4.3.3 Synthesis of 12 . . . . . . . . . . . . . . . . . . . . . . . . . 97
4.3.3.1
Single Crystal X-ray Diffraction . . . . . . . . . . . . 100
4.3.4 Attempted Synthesis of 13 . . . . . . . . . . . . . . . . . . . 104
5. Towards Switchable Pnictogen Bonding . . . . . . . . . . . . . . . . . . 107
5.1
Pnictogen Bonding with Neutral and Charged Pnictogen Bond
Acceptors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
5.1.1 Antimony(III) Pnictogen Bond Donor Ability with Neutral and Charged Pnictogen Bond Acceptors . . . . . . . . . 108
5.1.2 Bismuth(III) Pnictogen Bond Donor Ability with Neutral
and Charged Pnictogen Bond Acceptors . . . . . . . . . . . 114
5.2
Pnicotgen Catecholates . . . . . . . . . . . . . . . . . . . . . . . 117
5.2.1 Photoswitchable Catechol . . . . . . . . . . . . . . . . . . . 120
5.3
Toward Photoswitchable Pnictogen Catecholates . . . . . . . . 120
5.4
Additional Titration Data . . . . . . . . . . . . . . . . . . . . . . 126
6. General Experimental Methods . . . . . . . . . . . . . . . . . . . . . . . 130
6.1
Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
6.1.1 IR Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . 130
vi
Texas Tech University, Miranda C. Andrews, December 2019
6.1.2 NMR Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . 130
6.1.3 Solution Ultraviolet-visible Spectroscopy . . . . . . . . . . 130
6.2
Solution-state Binding Studies . . . . . . . . . . . . . . . . . . . 131
6.2.1 Experimental Data Collection . . . . . . . . . . . . . . . . . 131
6.2.2 Data Reduction . . . . . . . . . . . . . . . . . . . . . . . . . 132
6.2.3 Data Fitting . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
6.2.4 1 to 1 Binding Model . . . . . . . . . . . . . . . . . . . . . . 134
6.2.4.1
Adaptation to UV-vis Spectroscopy (pathlength = 1
cm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
6.2.5 1 to 1 Binding Model with Ion Pairing . . . . . . . . . . . . 137
6.2.5.1
Adaptation to UV-vis Spectroscopy (pathlength = 1
cm) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
6.2.6 1 to 1 and 1 to 2 Composite Binding Model . . . . . . . . . 141
6.2.6.1
Adaptation to UV-vis Spectroscopy (pathlength = 1
cm) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
6.2.7 1 to 1 and 1 to 2 Composite Binding Model with Ion Pairing145
6.2.7.1
Adaptation to UV-vis Spectroscopy (pathlength = 1
cm) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
6.2.8 Computational Details . . . . . . . . . . . . . . . . . . . . . 150
7. Summaries and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . 152
7.1
Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . 155
vii
Texas Tech University, Miranda C. Andrews, December 2019
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
viii
Texas Tech University, Miranda C. Andrews, December 2019
ABSTRACT
The use of switchable molecules to modulate properties of pedant groups to affect a measurable change was investigated. A survey of the literature identified
many switchable systems, induced by stimuli such as light, heat, electrical current, and/or mechanical stress. Light is easily tuned and readily available, so photoswitchable systems were chosen to study. Switchable systems that could be synthesized in relatively few steps and could be incorporated into groups that showed
precedence for interacting with other species were proposed. These groups include amidines and aminopyridines, which can be used as ligands in metal complexes when deprotonated, and squaramides, which are well known to participate in hydrogen-bonding and ion-binding. Substituents known to participate in
halogen- and pnictogen-binding were also studied.
The ability of switchable moieties to modulate changes in the redox properties and
electrostatic potentials as they relate to halogen bond and pnictogen bond donor
ability was investigated computationally. For compounds that were successfully
synthesized, experimental and computational results correlate well. This modulation of pendant groups has potential applications in light harvesting and stimulus
responsive supramolecular assemblies.
ix
Texas Tech University, Miranda C. Andrews, December 2019
LIST OF COMPOUNDS
Compound 1
Compound 2
Compound 3
x
Texas Tech University, Miranda C. Andrews, December 2019
Compound 4
Compound 5
Compound 6
Compound 7
Compound 8
xi
Texas Tech University, Miranda C. Andrews, December 2019
Compound 9
Compound 10
Compound 11
Compound 12
xii
Texas Tech University, Miranda C. Andrews, December 2019
Compound 13
Compound 14
Compound 15
Compound 16
Compound 17
xiii
Texas Tech University, Miranda C. Andrews, December 2019
Compound 18
Compound 19
Compound 20
Compound 21
xiv
Texas Tech University, Miranda C. Andrews, December 2019
Compound 22
Compound 23
Compound 24
xv
Texas Tech University, Miranda C. Andrews, December 2019
LIST OF TABLES
2.1
Relative energies (kJ·mol−1 ) of different configurations of 1 and 2
before and after isomerization (PW91 functional, TZP(d) basis set). . 20
2.2
Redox potentials of 1 and 2 . . . . . . . . . . . . . . . . . . . . . . . . 32
2.3
Change in summed Hirshfeld charges for ligating groups of various
compounds upon ring-opening. . . . . . . . . . . . . . . . . . . . . . . 37
2.4
Crystal structure data for 8·CuCl(pyridine)3 . . . . . . . . . . . . . . . 55
3.1
Structural and energy changes upon ring-opening in various firstrow transition metal complexes containing a photoswitchable aminidate ligand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
3.2
Changes in calculated charge of metal ion and frontier orbitals of
first-row transition metal amindinates. . . . . . . . . . . . . . . . . . . 62
3.3
Energy information using TPSSh functional and def2-SVP basis set. . 68
4.1
Calculated bond distances for 10 and 11 with neutral halogen bond
acceptors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
4.2
Crystal structure data for 10. . . . . . . . . . . . . . . . . . . . . . . . . 89
4.3
Crystal structure data for 11. . . . . . . . . . . . . . . . . . . . . . . . . 94
4.4
Crystal structure data for 11·4,4‘-bipyridine. . . . . . . . . . . . . . . . 96
4.5
Crystal structure data for 12·4,4‘-bipyridine. . . . . . . . . . . . . . . . 101
4.6
Crystal structure data for 14·4,4’-bipyridine. . . . . . . . . . . . . . . . 103
xvi
Texas Tech University, Miranda C. Andrews, December 2019
LIST OF FIGURES
1.1
Examples of switchable organic molecules . . . . . . . . . . . . . . . .
1
1.2
Example of photoisomerization in photochromic lenses . . . . . . . .
2
1.3
Photodimerization and thermal decomposition of tetracene. . . . . .
3
1.4
Potential energy surfaces of possible photoreaction pathways . . . .
4
1.5
Notable examples of complexes that contain photoswitchable ligands that have measurable effects on the ligand field . . . . . . . . .
5
Photoisomerization-Induced SpinCharge Excited State (PISCES) Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6
1.7
Examples of thermochromic switches . . . . . . . . . . . . . . . . . .
8
1.8
NLO crystal thermoswitch . . . . . . . . . . . . . . . . . . . . . . . . .
8
1.9
Examples of electrochromic switches . . . . . . . . . . . . . . . . . . . 10
1.6
1.10 Supramolecular electrochromic switches . . . . . . . . . . . . . . . . . 11
1.11 Chalcone/flavanone chemosensor . . . . . . . . . . . . . . . . . . . . 11
1.12 Spiropyran chemoswitch . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.1
Molecular orbital diagram for general octahedral complex with πacceptor ligands. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.2
Molecular orbital diagram for general octahedral complex with πdonating ligands. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.3
Amidine ligating group. . . . . . . . . . . . . . . . . . . . . . . . . . . 18
xvii
Texas Tech University, Miranda C. Andrews, December 2019
2.4
Fulgimide and spiropyran-type amidine switches. . . . . . . . . . . . 18
2.5
Ball and stick representation of low quality crystal structure of 1(o). . 19
2.6
Possible isomers of amidine molecular switches, 1 and 2. . . . . . . . 20
2.7
UV-vis spectra of 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.8
Unsubstituted fulgimide and spiropyran molecular switches . . . . . 23
2.9
Frontier orbitals of 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.10 1 H NMR spectra of 1 in various solvents before and after 350 nm
irradiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.11 UV-vis spectra of 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.12 Frontier orbitals of 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.13 Multiple cycles of UV and visible irradiation of 1 and 2 monitored
by UV-vis spectrosopy . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.14 Eyring-Polanyi plot for both forward and reverse thermal reactions
of 2 in methanol. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.15 Energy diagram for the thermal conversion of 2(c) to 2(o). . . . . . . . 29
2.16 Bisphenyl amidine, 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.17 CV of 1 and 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.18 Charge distributions for 1 and 2 . . . . . . . . . . . . . . . . . . . . . . 32
2.19 Spectroelectrochemical spectra of 1 and 2 . . . . . . . . . . . . . . . . 34
2.20 SOMOs and ring-closing mechanism for 1 . . . . . . . . . . . . . . . . 35
2.21 Ligand scaffolds of interest for computational studies . . . . . . . . . 36
xviii
Texas Tech University, Miranda C. Andrews, December 2019
2.22 Summed Hirshfeld charges for ligating groups in (a) 6, (b) 7, and (c) 9. 37
2.23 Frontier orbitals of 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.24 Frontier orbitals of 7 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
2.25 Synthesis of bisarylsquaramide (8) and subsequent metallation attempt. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
2.26 Crystal structure of copper(I) chloride squaramide complex. . . . . . 41
2.27 Synthesis of 9(c) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.28 Frontier orbitals of 9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.29 Electrostatic potential maps of 9 . . . . . . . . . . . . . . . . . . . . . . 44
2.30 Di-ATR FT-IR spectra of 1(o) and 1(c) . . . . . . . . . . . . . . . . . . . 45
2.31 1 H NMR spectrum of isolated 1 in d6 -DMSO. . . . . . . . . . . . . . . 46
2.32 Di-ATR FT-IR spectra of 2(c) and 2(o). . . . . . . . . . . . . . . . . . . 46
2.33 1 H NMR spectrum of isolated 2 in d6 -DMSO. . . . . . . . . . . . . . . 47
2.34 Absorbance of a 0.100 mM solution of 2 in methanol at 537 nm as a
function of time at 323 K. . . . . . . . . . . . . . . . . . . . . . . . . . . 48
2.35 Absorbance of a 0.100 mM solution of 2 in methanol at 537 nm as a
function of time at 313 K. . . . . . . . . . . . . . . . . . . . . . . . . . . 48
2.36 Absorbance of a 0.100 mM solution of 2 in methanol at 537 nm as a
function of time at 303 K. . . . . . . . . . . . . . . . . . . . . . . . . . . 49
2.37 Scheme for attempted synthesis of 7(c) . . . . . . . . . . . . . . . . . . 50
2.38 1 H NMR spectrum of 8 in d6 -DMSO. . . . . . . . . . . . . . . . . . . . 51
xix
Texas Tech University, Miranda C. Andrews, December 2019
2.39 Di-ATR FT-IR spectra of 8 before and after deprotonation . . . . . . . 52
2.40 1 H NMR spectrum of deprotonated 8 in d6 -DMSO. . . . . . . . . . . . 53
2.41 1 H NMR spectrum of 9 in d6 -DMSO . . . . . . . . . . . . . . . . . . . 57
3.1
Generic octahedral metal complex . . . . . . . . . . . . . . . . . . . . 58
3.2
Examples of orbital overlap in octahedral metal complexes . . . . . . 59
3.3
Frontier orbitals of photoswitchable Cr(II) complex . . . . . . . . . . 63
3.4
Frontier orbitals of photoswitchable Mn(II) complex . . . . . . . . . . 64
3.5
Frontier orbitals of photoswitchable Cu(II) complex. The HOMO
shown is a singly-occupied alpha orbital (spin up). (0.04 a.u. isosurfaces) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
3.6
Frontier orbitals of photoswitchable Ni(II) complex . . . . . . . . . . 65
3.7
Frontier orbitals of low-spin photoswitchable Fe(II) complex . . . . . 67
3.8
Frontier orbitals of high-spin photoswitchable Fe(II) complex . . . . 68
3.9
Frontier orbitals of low spin photoswitchable Co(II) complex . . . . . 69
3.10 Frontier orbitals of high spin photoswitchable Co(II) complex . . . . 69
4.1
Compounds (a) 10 and (b) 11 . . . . . . . . . . . . . . . . . . . . . . . 74
4.2
Electrostatic potential map of 10 . . . . . . . . . . . . . . . . . . . . . . 75
4.3
Electrostatic potential map of 11 . . . . . . . . . . . . . . . . . . . . . . 76
4.4
Ball and stick models of 10 and 11 . . . . . . . . . . . . . . . . . . . . 77
4.5
Ball and stick model of 11(c)·4,4‘-bipyridine from single crystral structure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
xx
Texas Tech University, Miranda C. Andrews, December 2019
4.6
UV-vis spectra of 11 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
4.7
Thermal reversion of 11(o) to 11(c) . . . . . . . . . . . . . . . . . . . . 79
4.8
BODIPY core . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
4.9
Iodo-tetrathiafulvalene electroswitch. . . . . . . . . . . . . . . . . . . 81
4.10 Compounds 12 and 13 . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
4.11 Electrostatic potential map of 12 . . . . . . . . . . . . . . . . . . . . . . 83
4.12 Electrostatic potential map of 13 . . . . . . . . . . . . . . . . . . . . . . 84
4.13 Ball and stick model of 12·4,4‘-bipyridine from single crystal structure. 84
4.14 Structure of 14 and ball and stick model of a 14·(4,4’-bipyridine)4 . . . 85
4.15 Synthesis of 10(c) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
4.16 1 H NMR spectrum of 10 in d6 -DMSO. . . . . . . . . . . . . . . . . . . 87
4.17 Synthesis of 11(c) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
4.18 1 H NMR spectrum of 11 in CD3 OD. . . . . . . . . . . . . . . . . . . . 92
4.19 Synthesis of 12 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
4.20 1 H NMR spectrum of 8-phenyl-1,3,5,7-tetramethylBODIPY in CDCl3 . 98
4.21 1 H NMR spectrum of 12 in CDCl3 . . . . . . . . . . . . . . . . . . . . . 99
4.22 Synthetic approach to of 13 . . . . . . . . . . . . . . . . . . . . . . . . 104
4.23 1 H NMR spectrum of 8-chloroBODIPY. . . . . . . . . . . . . . . . . . 106
5.1
Graphic representation of strategy for achieving triple pnictogen
bond donor that prevents self-recognition. . . . . . . . . . . . . . . . . 107
xxi
Texas Tech University, Miranda C. Andrews, December 2019
5.2
Antimony(III) pnictogen bond donor, 15. . . . . . . . . . . . . . . . . 108
5.3
Electrostatic potential map of 15 . . . . . . . . . . . . . . . . . . . . . . 108
5.4
Ball and stick representation of crystal structure of 15·tris(2-pyridinylmethyl)amine109
5.5
Electrostatic potential maps of 15 with pyridine(s) bound . . . . . . . 109
5.6
Titration data for 15 with pyridine fitted to a 1:1 binding model . . . 110
5.7
Titration data for 15 with 4,4’-bipyridine fitted to a 1:1 binding model 111
5.8
Titration data for 15 with chloride fitted to a 1:1 binding model . . . . 112
5.9
Titration data for 15 with iodide fitted to a 1:1 and 1:2 composite
binding model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
5.10 Bismuth(III) pnictogen bond donor, 16. . . . . . . . . . . . . . . . . . 114
5.11 Titration data for 16 with 4,4’-bipyridine fitted to a 1:1 and 1:2 composite binding model . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
5.12 Titration data for 16 with chloride fitted to a 1:1 and 1:2 composite
binding model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
5.13 Titration data for 16 with bromide fitted to a 1:1 and 1:2 composite
binding model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
5.14 Titration data for 16 with iodide fitted to a 1:1 and 1:2 composite
binding model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
5.15 Antimony(III) catecholate, 17 . . . . . . . . . . . . . . . . . . . . . . . 117
5.16 Titration data for 17 with bromide fitted to a 1:1 and 1:2 composite
binding model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
5.17 Titration data for 17 with iodide fitted to a 1:1 and 1:2 composite
binding model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
xxii
Texas Tech University, Miranda C. Andrews, December 2019
5.18 Mechanism for anion binding and subsequent substitution of stibaindole . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
5.19 Spiropyran catechol, 18. . . . . . . . . . . . . . . . . . . . . . . . . . . 120
5.20 Compounds (a) 19 and (b) 20. . . . . . . . . . . . . . . . . . . . . . . . 121
5.21 Electrostatic potential map of 19 . . . . . . . . . . . . . . . . . . . . . . 122
5.22 Electrostatic potential map of 20 . . . . . . . . . . . . . . . . . . . . . . 122
5.23 Compounds 21 and 22 . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
5.24 Compounds 23 and 24 . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
5.25 Electrostatic potential map of 21 . . . . . . . . . . . . . . . . . . . . . . 124
5.26 Electrostatic potential map of 22 . . . . . . . . . . . . . . . . . . . . . . 124
5.27 Electrostatic potential map of 23 . . . . . . . . . . . . . . . . . . . . . . 125
5.28 Electrostatic potential map of 24 . . . . . . . . . . . . . . . . . . . . . . 125
5.29 UV-vis spectra of titration of 26.4 µM 15 with pyridine and 4,4’bipyridine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
5.30 Titration data for 15 with 4,4’-bipyridine fitted to a 2:1 and 1:1 composite binding model . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
5.31 UV-vis spectra of titration of 25.2 µM 15 with chloride and iodide . . 127
5.32 UV-vis spectra of titration of 20.3 µM 16 with 4,4’-bipyridine . . . . . 128
5.33 UV-vis spectra of titration of 26.6 µM 16 with chloride, bromide, and
iodide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
5.34 UV-vis spectra of titration of 22.7 µM 17 with chloride, bromide, and
iodide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
xxiii
Texas Tech University, Miranda C. Andrews, December 2019
5.35 Titration data for 17 with chloride fitted to a 1:1 and 1:2 composite
binding model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
7.1
Summary of electron density redistribution for photoswitchable amidines.
152
xxiv
Texas Tech University, Miranda C. Andrews, December 2019
CHAPTER 1
INTRODUCTION TO MOLECULAR SWITCHES
1.1
Molecular Switches
Molecular switches are molecules that contain moieties that can transition between
two or more states.1 In many cases these states are structurally different from one
another; switching from one state to another involves a rearrangement of covalent
chemical bonds as shown in Fig. 1.1.2–4 Many of these rearrangements involve a
ring-closing or ring-opening event upon exposure to some stimulus. As a result
of the different structures, the states also often have different chemical properties,
such as solubility,5 color,6 and/or reactivity.7
Figure 1.1. Examples of organic switchable molecules (a) azobenzene, (b) dithienylethylene, and (c) aryldicyanooazulene.
The stimulus required for switching among states can vary. Molecules can be
switched using light (photoswitches),8–11 pressure (mechanoswitches),12–22 heat (thermoswitches),23, 24 electric current (electroswitches),25–30 or the presence of some
other chemical species (chemoswitches).31, 32 In many cases a particular molecular switch may transition to a different state upon exposure to one of multiple
stimuli. With the exceptions of chemoswitches, the stimulus does not have any
associated waste, which makes molecular switches attractive for various applications. A familiar example of a molecular switch is the naphthopyran that is used
1
Texas Tech University, Miranda C. Andrews, December 2019
in eyeglasses and contacts that transition from clear when indoors to dark when
exposed to sunlight and thermally revert back to colorless again as depicted in Fig.
1.2.33 Another common use of molecular switches is in pH indicators, which are
compounds that change color when exposed to acids or bases. Switches have also
been used in molecular machinery,34–37 such as molecule-sized cars,38 for which the
2016 Nobel Prize in Chemistry was awarded to Stoddart, Sauvage, and Feringa.
Figure 1.2. Example of photoisomerization in photochromic lenses.33
1.1.1
Photoswitches
An important subclass of photoswitchable molecules are photochromic molecules.
Photochromic compounds change color upon irradiation with light. This can be
a result of a photoreaction, as in the case of the tetracene dimerization or of isomerization as shown in the examples in Fig. 1.1. In some cases, the terms “photochromic” and “photoswitchable” are used interchangably. These systems exhibit a
reversible light-induced change to their absorption spectra. One of the first known
examples of photochromism was reported by Fritzsche in 1867, when he noticed
that a solution of tetracene changed from orange to colorless in sunlight, and that
the color returned after the solution was kept in the dark.39 This was the result
of the photodimerization of tetracene and subsequent thermal decomposition as
seen in Fig. 1.3. Photochromic compounds have since been studied for their potential uses as photoswitches to modulate processes such as the activation of the
small molecule, ammonia,40 the catalysis of transesterifications and amidations,41
and the catalysis of the ring-opening polymerizations of lactones.42
In the case of photoswitchable molecules, the stimulus required to switch the molecule between states is light of a particular wavelength (energy). While many
2
Texas Tech University, Miranda C. Andrews, December 2019
Figure 1.3. Photodimerization and thermal decomposition of tetracene.
substituents can be added to the photoswitch to change which wavelengths of
light results in switching, in most cases, UV light results in switching to one state,
and visible light (usually of a particular color) switches back to the more thermodynamically stable state. It should be noted that photoswitching does not necessarily result in 100% conversion to one species. Under most circumstances, an
equilibrium state, called the photostationary state results. This state is defined as
“a steady state reached by a reacting chemical system when light has been absorbed by at least one of the components. At this state the rates of formation and
disappearance are equal for each of the transient molecular entities formed.”43 All
examples shown in Fig. 1.1 have some photoswitchable component as indicated,
but Fig. 1.1b is the only example shown that is only switched by light.
Most photoswitchable systems that result in photoreactions or photoisomerizations are one-photon systems. Typically, one photon results in an excited singlet
or triplet (less common) state of the starting material, which then follows some
reaction pathway to form the photoproduct as shown in Fig. 1.4. All of the photoisomerizations discussed in this work are diabatic, following a reaction pathway
like that shown in Fig. 1.4(a). In these cases, photoexcitation is followed by a vibrational relaxation to the ground state potential energy surface, which then proceeds
as shown to give the photoproduct.44
The other two pathways shown in Fig. 1.4(b) and (c) are more common in gas
phase reactions. In the case of an adiabatic pathway, photoexcitation is followed by
proceeding along the potential energy surface of the excited state as shown in Fig.
1.4(b). This is then followed by radiative or vibrational relaxation to the ground
state product as shown. The last photoreaction pathway, shown in Fig. 1.4(c), is a
hot-ground-state photoreaction pathway. These photoreactions typically occur in
3
Texas Tech University, Miranda C. Andrews, December 2019
the gas phase under reduced pressure. In this pathway, photoexcitation is followed
by relaxation to give a vibrationally excited state. The reaction then proceeds in a
manner that is very similar to that of a unimolecular thermal reaction, as shown in
Fig. 1.4(c).44
Figure 1.4. Potential energy surfaces of possible photoreaction pathways: (a) diabatic, (b) adiabatic, and (c) ”hot” ground state photoreactions.
The application of photochromic materials for data storage is particularly promising, especially for photoswitches that undergo thermally irreversible changes.45, 46
One such example is the diarylethylene shown in Fig. 1.1(b). Photoswitches that
do not undergo thermal reversibility are of particular interest to this application
as the only write-rewrite stimuli that will cause the necessary photoisomerization
are the two wavelengths of light responsible for the photochromic behavior. The
ability to store and read data on the molecular level is technologically relevant, as
this could lead to faster data processing and more dense data storage.
4
Texas Tech University, Miranda C. Andrews, December 2019
1.1.1.1
Coordination Complexes Coupled with Photoswitches
Molecular switches have already been used to modulate measurable changes in
the magnetic properties of a metal complex as a result of changes in the ligand
field in metal complexes. Boillot, et al. showed several examples of spin crossover, specifically, ligand-driven light-induced spin change (LD-LISC) in iron(II)
and iron(III) complexes upon cis-trans isomerization.47, 48 One example is shown
in Fig. 1.5(a), although it should be noted that this particular complex has a very
low working temperature of only 90 K and was later improved upon.48 More
recent examples that use azobenzene-type photoswitchable ligands indicate that
even at room-temperature, a measurable change in magnetic properties can be observed for certain orientations of the photoswitchable moiety relative to the ligating group.49
Figure 1.5. Notable examples of complexes that contain photoswitchable ligands
that have measurable effects on the ligand field.42, 50–52
The example shown in Fig. 1.5(b) uses a dithienylethene that undergoes ringclosing and -opening that results in a measurable change in the magnetic susceptibility.51, 53 In their discussion of this phenomenon, these authors used the term
5
Texas Tech University, Miranda C. Andrews, December 2019
light-induced excited spin-state trapping (LIESST). This would seem to be a misnomer for the effects observed. Instead of light directly inducing the excited spin
state of the metal ion, light induces a change in the structure of the ligand(s), which
in turn affects the ligand field and ultimately, the spin state of the metal ion.
Figure 1.6. Photoisomerization-Induced SpinCharge Excited State (PISCES) Process. (Adapted with permission from [54], c 2018, American Chemical Society54
The third example shown in Fig. 1.5(c) used IR spectroscopy to evaluate the direction and magnitude of changes in the ligand field upon photoisomerization of the
spirooxazine moiety.50 In this example, Paquette, et al. monitored the change in
ligand field by monitoring the carbonyl stretches of the compound shown in Fig.
1.5(c) upon irradiation with 568 nm light. Upon ring-opening of the spirooxazine
moiety, the carbonyl stretches shift to higher frequencies. This indicates that the
molybdenum(0) does not donate as much electron density when the spirooxazine
is in the open form as a result of the spirooxazine ligand becoming more electronwithdrawing. The solution was then given ample time to thermally revert to the
closed form, after which the IR spectrum matched that of the solution before irradiation. Further investigation into the use of spirooxazines yielded a cobalt(III)
complex that undergoes a reversible charge-transfer coupled spin-transition upon
irradiation with 550 nm light. Upon photoisomerization of the spirooxazine, the
6
Texas Tech University, Miranda C. Andrews, December 2019
low spin cobalt(III) is reduced to high spin cobalt(II). This transition occurs at elevated temperatures (325 K), and results in a relatively long-lived high spin cobalt(II)
complex (τ =10 seconds).54
Lastly, in Fig. 1.5(d), the catalytic activity of a rhodium(I) complex was modified
by using a dithienylethylene-substituted N-heterocyclic carbene (NHC) ligand, reducing its catalytic activity by up to an order of magnitude.42 Because the hydroboration mechanism is dependent on the electron-donating ability of the NHC, the
rate of hydroboration decreases upon ring-closing of the dithienylethylene. This is
a result of the carbene becoming less electron-donating upon ring-closing.
This thesis focuses on fulgimide and spiropyran switchable systems. These systems are well-established in the literature, and can be synthesized in few steps.
They are not known to be sensitive to ambient conditions (although light and ambient temperatures can and do affect the isomerization), so they can be studied
easily.
1.1.2
Thermoswitches
Thermoswitches transition among states upon heating and/or cooling. Thermoswitches have already been put to use in many decorative applications, such as
color-changing dyes,55 paints (Fig. 1.7(a)), or hair extensions. Heptafulvene/ dihydroazulene systems show faster rates of photo- and thermally induced ringclosure and -opening as well as different absorption (and therefore colors) in the
presence of Lewis acids (Fig. 1.7(b)).56 This could have potentially interesting
applications in sensing. A heptafulvene/dihydroazulene which contains two possible ring-closure sites shows an increase in thermal ring-closure, likely due to
the increased possibility of s-cis conformations, which are necessary for the ringclosure to occur (Fig. 1.7).57
Thermoswitchable non-linear optical (NLO) crystals (Fig. 1.8) have also been developed.58 These thermoswitches are less molecular switches, as the resulting
change in color and properties does not involve an isomerization. In this particular
7
Texas Tech University, Miranda C. Andrews, December 2019
Figure 1.7. Examples of thermochromic switches: (a) spirooxazine,55 and (b) heptafulvene/dihydroazulene,56 and (c) bisheptafulvene/dicyanoazulene.57
case, irradiation with 350 nm light or heating above 60 ◦ C results in a significant
decrease in the material’s second harmonic generation (SHG) intensity. While irradiation resulted in a more efficient switching, and therefore a greater loss in SHG
intensity, thermal switching did produce a measurable decrease in SHG intensity.
In both photo- and thermally induced switching of the NLO crystal, the loss in
SHG intensity was ascribed to a decrease in the molecular permanent dipole.
Figure 1.8. NLO crystal thermoswitch.58
8
Texas Tech University, Miranda C. Andrews, December 2019
Some photoswitches can also have thermoswitchable behavior, which is denoted
as “∆” in Fig. 1.1. The thermal behavior of switches is a result of the activation
energy barrier required to transition between states to low enough that excitation
to an excited state potential energy surface does not necessarily have to occur (as
shown in Fig. 1.4) in order to overcome that barrier.
1.1.3
Electroswitches
Some molecular switches respond to an electrical current as a stimulus. The removal (oxidation) or addition (reduction) of an electron can result in a molecule
switching to a new state. Because the mechanisms that drive photoswitchable behaviors involves the excitation of an electron from one orbital to another of higher
energy, the introduction or removal of one electron can initiate the transition from
one state to another.
Molecular switches such as photoswitches often show changes in structure and/or
absorption upon oxidation and/or reduction. Some molecular switches, such as
dithienylethylenes (Fig. 1.1(b)) display multiple modes of switching. In one such
case,25 the open isomer is electrochemically inert as observed by cyclic voltammetry (Fig. 1.9(a)). However, after photo-induced ring-closing upon irradiation
with UV light, the closed form can be be further electroswitched to give the quinone
form. Fluorans also show promise as electroswitches as the one-electron oxidation
results in lactone ring-opening (Fig. 1.9(b)). The resulting radical cation is fluorescent, where the neutral, ring-closed form is not.28
Molecular switches can also incorporate multiple types (stimuli) of switches. One
such example incorporates a reversible, electrochemically active group (tetrathiafulvene, TTF) and a photoswitchable vinylheptafulvene/dihydroazulene (VHF/
DHA) moiety (Fig. 1.9(c)). Upon oxidation of the TTF moiety to give the radical
cation and the dication, the absorption spectrum associated with the photoswitchable group changes. Upon oxidation to the radical cation, the photostationary state
of the photoswitchable moiety contains considerably less of the ring-opened (VHF)
state. Similar behavior was observed for the dication.30
9
Texas Tech University, Miranda C. Andrews, December 2019
Figure 1.9. Examples of electrochromic switches: (a) dithienylethylene,25 (b)
fluoran,28 and (c) tetrathiafulvene bound to vinylheptafulvene/dicyanoazulene.30
Supramolecular assemblies have also shown considerable electroswitchable activity. As mentioned previously, Sir Fraser Stoddart shared the Nobel prize in 2016
for his work on molecular machines, which often depend on switchable behavior
to function. The components of catenanes (represented in Fig. 1.10(a)) and rotaxanes (represented in Fig. 1.10(b)) can be manipulated by redox processes in such a
way that their relative conformations change. These molecule-sized machines can
have very interesting implications for nanotechnology and molecular devices in
the near future.29
1.1.4
Chemoswitches
Similar to electroswitches, chemical oxidation or reduction can result in switching
a molecule among states. Other secondary bonding interactions can also induce
the isomerization between states by introducing a change in electron density affecting the relative stability of the two states. This is commonly observed in the
case of pH indicators, but also shows promise in the field of sensing.
10
Texas Tech University, Miranda C. Andrews, December 2019
Figure 1.10. Supramolecular electrochromic switches. (Adapted with permission
from [29], c Angewandte Chemie, International Edition, 2004)29
Changes in pH can result in isomerization upon protonation or deprotonation (Fig.
1.11). Chemoswitches such as chalcone/flavanone systems undergo ring-closing
upon deprotonation (increasing pH). This is a reversible process; decreasing the
pH regenerates the ring-opened state. This is also somewhat tunable. The pH at
which the switch opens or closes can be increased or decreased by the addition of
electron-donating and -withdrawing substituents at various positions.31
Figure 1.11. Chalcone/flavanone chemosensor.31
11
Texas Tech University, Miranda C. Andrews, December 2019
Spiropyrans have been used for the detection of aromatic thiols (Fig. 1.12).32 Under the conditions reported, the spiropyran/merocyanine was selective for the detection of only aromatic thiols; other nucleophilic species did not result in ringopening to give the merocyanine isomer. However, under these conditions, the
sensor was not very sensitive, requiring 60 minutes and 30 equivalents of analyte
to produce significant changes in absorption.
Figure 1.12. Spiropyran chemoswitch.32
Molecular switches have been studied beyond just the idea that the introduction
of a particular species would change the color of the switch. There have also been
studies that pursue a more complex mechanism in which the spin state of a metal
ion in a porous framework that contains iron(II) can be switched upon the introduction of a guest molecule or species.59 This change in spin state was not particularly selective for compounds that would stabilize the high spin state. Many
donor molecules such as water, alcohols, pyridine, and furans as well as benzene
and toluene would all result in stabilization of the high spin state. The only reported guest molecule that would stabilize the low spin state was carbon disulfide.
1.1.5
Mechanoswitches
Mechanoswitches produce a measurable change in properties upon a change in
pressure. Because the switching event often involves a change in the molecules
shape, one can imagine initiating the property change by forcing the molecule to
change its shape first by either stretching or compressing the molecular switch.
Polymers and co-polymers containing mechanoswitchable molecules have potential applications in imaging stress,13, 20 modulating reactivity of neighboring groups,15
soft robotics,16 and wearable electronics.17
12
Texas Tech University, Miranda C. Andrews, December 2019
1.2
Effect of Pendant Groups on Photoswitchable Behavior of Fulgimides and
Spiropyrans
Functional groups are known to influence the switching behavior of spiropyrans
and fulgimides. It has been shown that electron donating groups on the indoline
moiety of spiropyrans and electron withdrawing groups on the opposing phenyl
ring result in a photostationary state that strongly favors one isomer while electron
donating groups on the indoline moiety can influence the rate of thermal isomerization to the point in which it competes with (or dominates over) the photochemical processes.60–63 Recent computational studies show that the photocoloration of
spiropyrans with an acceptor group on the pyran ring proceeds by a triplet state
for both CO cleavage and cis/trans isomerization. The rate-determining step for
both thermal and photocoloration is the subsequent cis-trans isomerization.64
In fulgides, the ability of functional groups to stabilize a highly zwitterionic excited state (charge transferred from the (hetero)aromatic group to the fulgide) plays
a crucial role in the photochromism. The introduction of electron-withdrawing
groups on the anhydride moiety and electron-donating groups on the heteroaromatic moiety leads to more stable zwitterionic states. This is particularly important
in designing systems that can participate in charge-transfer.65
Sterics can influence the ratio of isomers that make up the photostationary state
of fulgides. Fulgides have three possible states, closed, E-open and Z-open. As
certain substituents become bulkier, the photostationary states contain less (or no)
Z-open form. Bulkier substituents also lead to a photostationary state composed of
less closed form. Overall, steric bulk decreases quantum yield of fulgides, resulting
in less (if any) conversion from E-open to either Z-open or closed forms.66
Substituents also affect the absorption spectra of both fulgides and fulgimides, and
therefore the wavelengths that lead to photoswitching. Introduction of electrondonating groups on the 5-position of the indole ring of fulgides result in closed
forms that absorb at longer wavelengths. Incorporating additional electron-donating
groups results in a closed form that absorbs at even longer wavelengths. Moving
the electron-donating substituent to the 6-position leads to a larger molar extinc13
Texas Tech University, Miranda C. Andrews, December 2019
tion coefficient.67, 68 Electron-withdrawing substituents on the heteroaromatic moiety of fulgimides shift aborptions to longer wavelengths.69
Electronic effects play an important role in the absorption and photostationary
states of spiropyran/merocyanine systems. Strongly electron-withdrawing groups
on the chromene moiety lead to photostationary states that contain more ring-open
merocyanine isomers. Conversely, electron-donating groups on the indoline moiety also stabilize the merocyanine isomer.60 This is not unexpected, as the merocyanine structure contains a formal positive and formal negative charge on the
indoline and chromene moieties, respectively. An electron-donating group would
stabilize a formal positive charge, and an electron-withdrawing group would stabilize a formal negative charge.
There are fewer examples of studies that elucidate how the switchable component
affects the substituents. This thesis focuses on if, how, and to what extent switchable moieties can affect a measurable change in pendant groups.
1.3
Purpose, Scope and Overview of the Dissertation
Chapter 2 focuses on measuring the effects photo- and thermoswitchable groups
have on the neutral form of a remotely positioned ligating group. Particular emphasis is placed on measurements and the corresponding calculations that give insight into how the redox properties of the ligating group are changed upon switching of either a spiropyran or a fulgimide moiety. This leads into a computational
study of how the changes described in the neutral ligating group might extend
to or differ from the anionic ligand when bound to first-row transition metals in
Chapter 3.
Chapters 4 and 5 turn to main-group elements and how molecular switches might
affect supramolecular interactions such as halogen bonding and pnictogen bonding. Multiple photo- and electroswitches are discussed along with their abilities to modulate halogen bonding. The studies into pnictogen bonding begin by
establishing measurable properties, such as neutral molecule and anion binding
14
Texas Tech University, Miranda C. Andrews, December 2019
strength with the lesser explored pnictogen bonding motif. The chapter then moves
on to discuss computational studies that probe the abilities of various photoswitchable groups to modulate pnictogen bonding and to what extent those bonding interactions can be modulated by the photoisomerization(s). Chapter 6 describes the
more general experimental and computational methods employed for this work.
15
Texas Tech University, Miranda C. Andrews, December 2019
CHAPTER 2
REMOTE SWITCHING OF A LIGATING GROUP
This chapter will focus predominantly on the author’s contributions to ”Modulation of the carboxamidine redox potential through photoinduced spiropyran or
fuligmide isomersation.”70 For this study, synthetic work and structural characterization of the isolated molecular switches as the thermodynamically favorable
isomer were performed by Dr. Ping Peng and the electrochemical, photoelectrochemical, and spectroelectrochemical data were collected by Dr. Amit Rajput.
Figure 2.1. Molecular orbital diagram for general octahedral complex with πacceptor ligands.
Chapter 1 provided several examples of metal complexes that contain molecular switches. In all of the metal complexes discussed therein, the switchable ligand is a neutral, σ-donor ligand. In many cases, the ligand is also a π-acceptor
Fig. 2.1 shows the MO diagram for a generic octahedral complex which contains
π-acceptor ligands. In these cases, the switching of the ligand almost certainly
changes the relative energy of the the unoccupied π-orbitals of the ligand (Fig. 2.1,
purple arrow). This necessarily affects the relative energies of the t2g and eg ∗ molecular orbitals of the metal complex (Fig. 2.1, red arrow).
16
Texas Tech University, Miranda C. Andrews, December 2019
This approach is useful for some applications. However, in the case of σ-donor,
π-acceptor ligands, the ∆oct (or corresponding energy difference) is usually high,
generally leading to low spin complexes. Conversely, for π-donating ligands, the
∆oct is generally smaller (Fig. 2.2). By changing the relative energy of the π-orbitals
of π-donating ligands, one could potentially modulate the ∆oct to give rise to either
high spin or low spin complex. In cases like iron and cobalt which can undergo a
change in spin state, the use of π-donating ligands could lead to both high and low
spin cases in an octahedral geometry.
Figure 2.2. Molecular orbital diagram for general octahedral complex with πdonating ligands.
The ligating group initially chosen to test the hypothesis that a molecular switch
could be used to modulate a ligand field was an amidine, which when deprotonated to give the amidinate, is a π-donating ligand. Amidinates are well established
as ligands in the field of coordination chemistry71–73 and have not been studied as
pendant groups on photoswitches. Furthermore, literature searches did not produce any examples of a photoswitchable anionic ligand in a coordination complex.
Anionic ligands are useful as they support early transition metals and complexes
containing metal-metal bonds. Amidines are particularly versatile as there are up
17
Texas Tech University, Miranda C. Andrews, December 2019
to three opportunities to sterically or electronically tune the ligating group (Fig.
2.3, blue box).
Figure 2.3. Amidine ligating group.
The first generation of photoswitches had only one photoswitchable moiety (Fig.
2.3, R1 , red box). Fulgimide and spiropyran photoswitches were the switchable
moieties chosen (Fig. 2.4). The substituent opposite the photoswitchable moiety
(Fig. 2.3, R2 , yellow box) was a simple phenyl ring, and in an effort to force the
bite angle of the ligating group to be smaller and favor monometallic complexes,
a bulky tert-butyl group was chosen as the third substitution (Fig. 2.3, R3 , green
box).
Figure 2.4. (a) Fulgimide and (b) spiropyran-type amidine switches.70
2.1
Synthesis and Structural Characterization
The open form of 1 and the closed form of 2 were prepared in multigram scales.70
This followed the strategy to design these molecules in such a way that the pho18
Texas Tech University, Miranda C. Andrews, December 2019
toswitchable moiety electronically integrates with the amidine group, while utilizing a minimal number of simple/scalable synthetic steps. For 1, the E-isomer at
the fulgimide (Efulg ) was selectively isolated by starting from the pure (3E)-3-[(1,3dimethyl-1H-indol-2-yl)methylene]dihydro-4-(1-methylethylidene)-2,5-furandione.74
Both molecules were characterized by high-resolution mass spectrometry and infrared (IR), 1 H and 13 C nuclear magnetic resonance (NMR) spectroscopies.
Figure 2.5. Ball and stick representation of low quality crystal structure of 1(o).
A crystal structure was determined for 1(o); however, the quality was only sufficient to allow connectivity and configuration to be determined as seen in Fig.
2.5. The crystal structure shows only the Efulg -isomer which is consistent with
the synthetic approach as well as the UVvis absorption spectrum (the Zfulg -isomer
would be expected to absorb at longer wavelengths).74 1 H NMR spectra taken
in d6 -DMSO show that irradiation with UV light (350 nm) results only in 1(c) as
no peaks associated with either the Efulg -isomer or the Zfulg -isomer are observed.
Subsequent irradiation with green (530 nm) light leads only to the Efulg -isomer of
1.
19
Texas Tech University, Miranda C. Andrews, December 2019
Table 2.1. Relative energies (kJ·mol−1 ) of different configurations of 1 and 2 before
and after isomerization (PW91 functional, TZP(d) basis set).
Tautomer 1
1(o)
1(c)
∆Ephotoisomerization
Tautomer 2
1(o)
1(c)
∆Ephotoisomerization
Tautomer 1
2(c)
2(o)
∆Ephotoisomerization
Tautomer 2
2(c)
2(o)
∆Ephotoisomerization
Z
1.22
1.23
0.01
Z
3.10
1.85
-1.25
Z
0∗
9.80
9.80
Z
6.87
14.87
8.00
Z’
10.66
10.31
-0.35
E
1.47
-0.06
-1.53
Z’
11.23
23.13
11.90
E
2.21
14.79
12.58
E
0∗
0.31
0.31
Z’
10.17
7.13
-3.04
E
2.38
12.12
8.84
Z’
8.17
27.08
18.91
E’
40.10
40.97
0.87
E’
32.05
27.78
-4.27
E’
48.40
52.17
3.77
E’
29.00
47.19
18.19
Figure 2.6. Possible isomers of amidine molecular switches, 1 and 2.
∗
The lowest energy configuration for each photoswitch of Tautomer 1 (consistent with crystal
structure of 1(o)) was chosen as zero. The remaining values represent the energies relative to that
species for each compound.
20
Texas Tech University, Miranda C. Andrews, December 2019
In this crystalline phase, the isomer in which the aryl groups are both trans- to the
t-butyl group is observed. Furthermore, based on the markedly different bond
lengths and deviations from planarity in the crystal structure, the tautomer in
which the photoswitchable group is bound to the imine nitrogen of the amidine
can be confidently assigned to 1(o). Because single crystals of 1(c), 2(c), and 2(o)
could not be isolated, DFT calculations were used to elucidate the relative energies of the different amidine tautomers (Fig. 2.6) and configurations for each photoswitch in the gas phase. Several configurations were modelled, but given the
small differences in energy among most configurations, it is likely that all configurations are in dynamic equilibrium under experimental conditions, i.e., in solution at room temperature. In both tautomers, the Z-configuration is the lowest or
nearly the lowest energy confirmation for 1 and 2. The Z-configuration of Tautomer 1 (boxed species in Fig. 2.6) is used for all subsequent calculations discussed
herein for consistency, which is also consistent with the crystal structure of 1(o).
2.2
Photochemical Behavior
Substituents can affect the absorption spectra of both spiropyrans60 and fulgimides, and therefore the wavelengths that lead to photoswitching.67–69 The next step
was to determine how the introduction of the amidine functional group to the photoswitchable portion of the molecule would affect the photoswitchable behavior.
The effect that switching had on the amidine group was also probed.
Depending on the concentration, solutions of 1(o) and 2(c) were colorless to pale
yellow, and irradiation with a UV light source resulted in vibrant purple solutions.
Ultraviolet-visible (UV-vis) spectra are shown in Figs. 2.7 and 2.11. Upon irradiation of a methanol solution of 1 with 350 nm light, the peak at 368 nm diminishes as a peak at 531 nm grows in (Fig. 2.7). This is consistent with the behavior
of fulgimides upon ring-closing (1(o) to 1(c)).75 However, the substitution of the
amidine appears to have some effect on the relative energies of the frontier orbitals. The transition (associated with the peak at 368 nm) from the highest occupied
molecular orbital (HOMO) to the lowest unoccupied molecular orbital (LUMO),
21
Texas Tech University, Miranda C. Andrews, December 2019
Figure 2.7. UV-vis spectrum of 25.0 µM fulgimide 1(o) in methanol before and after
15 minutes irradiation at 350 nm to give 1(c).
(associated with the peak at 368 nm) of 1(o) is higher in energy than similar unsubstituted fulgimides.76–79 For instance, the structurally related unsubstituted
fulgimide 3 (Fig. 2.8(a)) absorbs at 400 nm.77 The HOMO-LUMO transition of
1(c) (associated with the peak at 535 nm) is lower in energy than some fulgimides
but nearly identical to that of the unsubstituted fulgimide (3).77 Solvatochromic
effects could be responsible for the differences observed, but DFT calculations predict similar, albeit small (∼0.06 eV), differences in orbital energies between the
amidine-substituted and unsubstituted fulgimides.
The spectra are consistent with the time-dependent DFT (TDDFT) calculated electronic transition energies and oscillator strengths.70 The HOMO-LUMO transition
of 1(c) is lower in energy than the HOMO-LUMO transition of 1(o), resulting in a
red-shift in the peak of interest in the absorption spectrum. Relevant orbitals are
shown in Fig. 2.9. 1 H NMR spectroscopy was used to determine the speciation
after irradiation. In deuterodimethyl sulfoxide (d6 -DMSO), quantitative conversion from 1(o) to 1(c) occurs upon UV-irradiation (350 nm) as seen in Fig. 2.10(a)
and (b).
22
Texas Tech University, Miranda C. Andrews, December 2019
Figure 2.8. Unsubstituted (a) fulgimide and (b) spiropyran molecular switches.
The absorption spectra of 2(c) has a λmax of 316 nm in methanol which is consistent
with other donor-substituted nitro-spiropyrans. For instance, a methoxy group in
place of the amidine in 2(c) has a λmax of 315 nm in ethanol.60 As 2(c) isomerizes to
2(o) under UV irradiation (Fig. 2.11), a peak grows in at 537 nm which is consistent with the behavior of nitro substituted spiropyrans with hydrogen or electron
donor groups in the R position (such as 4, Fig. 2.8(b)).60 No peak diminishes appreciably, which is again consistent with the known isomerizations of spiropyrans to
form merocyanines.60 Experimentally, no band associatated with a HOMO-LUMO
transition is observed for 2(c). Upon irradiation, a new peak grows in as 2(o) is
formed. TDDFT suggests that the HOMO-LUMO transition for 2(c) has a negligible oscillator strength (f ) and is therefore not observed experimentally.70 This
is likely because the HOMO is largely localized on the indoline side of the spirocarbon while the LUMO is entirely localized on the other side of the spiro-carbon
and these rings are orthogonal to each other (Fig. 2.12(a)). Upon opening, the
HOMO and LUMO become coplanar (Fig. 2.12(b)) which renders this transition
far more efficient than in the closed form leading to an observable absorption at
537 nm for 2(o). The HOMO-LUMO transition (537 nm) of 2(o) occurs at higher
energy as compared to the related unsubstituted spiropyran that lacks the amidine
group.80 This is consistent with TDDFT calculations which give an energy difference (transition of 2(o)-transition of 4(o)) of 0.05 eV. Fig. 2.12 illustrates a significant
reorganization of the electron density associated with the HOMO at the amidine
23
Texas Tech University, Miranda C. Andrews, December 2019
Figure 2.9. Frontier orbitals of (a) 1(o) and (b) 1(c) (0.04 a.u. isosurfaces).
group. Upon switching of 2, electron density appears to be pulled away from the
amidine as a result of electronically connecting the electron deficient nitro substituted ring to the rest of the molecule. This prompted us to probe for a measurable
change in the electronic structure of the redox-active amidine electrochemically
(vide infra). It should be noted that absorption spectra were calculated in the gas
phase, without accounting for solvent effects. This may contribute to the differences in λmax observed between the calculated and experimental spectra as shifts
in the λmax of up to 30 nm have been observed for related fulgides and fulgimides74, 81 and up to 100 nm for related spiropyrans depending on the solvent.62, 82
2.2.1
Photochemical Stability
In order to determine how robust the photoswitchable behaviors of 1 and 2 are,
solutions of the compounds were irradiated through ten cycles of UV and visible
light. It was found that both compounds efficiently switched under a wide range
of UV light (254, 350, and 378 nm in methanol). Compound 1(c) could be converted
back to 1(o) by irradiating with green light (530 nm). Fig. 2.13a shows the reversible photoswitchable behavior of 1 by monitoring the absorbance at 531 nm during
several cycles of alternating UV and visible light. Irradiation of the spiropyran 2(c)
24
Texas Tech University, Miranda C. Andrews, December 2019
Figure 2.10. 1 H NMR spectra of 1(o) in d6 -DMSO before (black) and after (grey) 10
minutes UV irradiation (350 nm) to give 1(c). (a) aromatic region (b) aliphatic region. 1 H NMR spectra of 1(o) in CD3 OD before (black) and after (grey) 10 minutes
UV irradiation (350 nm) to give 1(c). (c) aromatic region (d) aliphatic region.
with 254 nm light led to 25% 2(o) as seen in Fig. 2.13b, red triangles, but resulted
in decomposition upon repeated cycling. Irradiation of 2(c) with 350 nm light resulted in a lower conversion (only 11%) to 2(o) which can also be seen in Fig. 2.13b.
Irradiation at 419 nm or 530 nm could convert 2(o) back to 2(c) as depicted for ten
cycles in Fig.2.13b.
25
Texas Tech University, Miranda C. Andrews, December 2019
Figure 2.11. UV-vis spectrum of 25.0 µM spiropyran 2(c) in methanol before and
after 15 minutes irradiation at 350 nm to give 2(o).
Figure 2.12. Frontier orbitals of (a) 2(c) and (b) 2(o) (0.04 a.u. isosurfaces).
26
Texas Tech University, Miranda C. Andrews, December 2019
Figure 2.13. Absorbance of a solution of (a) 1(o) (25.0 µM in methanol) measured at
531 nm before and after irradiation with 350 nm light, then alternating irradiation
with 530 nm light and 350 nm light and (b) a solution of 2(c) (25.0 µM in methanol)
measured at 537 nm before and after irradiation with 350 nm light (black circle) or
254 nm light (red triangle), then alternating irradiation with 530 nm light and UV
light (right). Dashed lines are provided as a visual aid.
27
Texas Tech University, Miranda C. Andrews, December 2019
2.2.2
Thermochemical Behavior
As 2 was found to thermally isomerize, the activation energy (∆G‡ ) for this was
determined experimentally. The Eyring-Polanyi equation was used to plot (Fig.
2.14) the temperature dependence of the rate constant for the thermal conversion
of 2(c) to 2(o), which followed first order kinetics in methanol and to determine
the enthalpy (∆H‡ = 116 kJ·mol1 ) and entropy (∆S‡ = 0.04 kJ·K−1 ·mol−1 ) of activation. Literature values for related systems are on the order of 75-108 kJ·mol−1 , but
may be affected by the solvent,83 donor/acceptor groups64 or presence of suitable
nucleophiles.83, 84 This slightly higher activation energy is in line with the findings
of Leszczynski who determined that systems with small bond length alternation
(BLA) values, and therefore a greater contribution from the zwitterionic form of
the merocyanine, had a higher cis/trans isomerization activation energy than the
C-O cleavage and is the rate determining step.64 The BLA value for 2(o) is 0.004;
lower than the calculated value for 4(o) (BLA value = 0.019), which has an experimental activation energy for ring opening of 102 kJ·mol−1 . The ∆H‡ (90.7 kJ·mol1 )
and ∆S‡ (-0.02 kJ·K−1 ·mol1 ) for the reverse reaction were also determined. This
gives a ∆H of 25.3 kJ·mol−1 (Fig. 2.14) for the overall conversion of 2(c) to 2(o).
This is consistent with DFT calculations, which give a ∆H of 22.8 kJ·mol−1 for this
conversion. The absorption spectra of 1(o) and 1(c) do not change as a function of
time at accessible temperatures, so it is not possible to experimentally determine
an activation energy or change in enthalpy.
28
Texas Tech University, Miranda C. Andrews, December 2019
Figure 2.14. Eyring-Polanyi plot for both forward and reverse thermal reactions of
2 in methanol.
Figure 2.15. Energy diagram for the thermal conversion of 2(c) to 2(o).
29
Texas Tech University, Miranda C. Andrews, December 2019
2.3
Electrochemical Behavior
The amidine moiety is known to be redox active.85 If there is efficient electronic
communication between this group and the photoswitch, then a change in the
redox potential associated with the amidine should be observed upon photoinduced isomerization. Cyclic voltammetry (CV) was first performed on a simple
bisphenyl amidine, 5 (Fig. 2.16), to characterize the redox response of the amidine
group under these experimental conditions. Consistent with literature (1.24 V vs.
Ag/AgCl86 ), the only redox response observed was an irreversible oxidative wave
at 0.66 V vs. Fc/Fc+ . The CV of 1(o) (Fig. 2.17(a)) displays two redox processes:
one oxidation at 0.67 V and one reduction at E1/2 = 2.11 V. The oxidative response
is attributed to the amidine group. The reductive response is quasi-reversible in
nature (∆Ep = 180 mV, ipc /ipa ≈ 1) and consistent with the reductive responses of
fulgides which have been shown to induce cyclization as determine by Fox and
Hurst.87 The voltammogram of an CH3 CN solution of 2(c) (Fig. 2.17(b)) shows
three redox processes. There are two quasi-reversible oxidations at E1/2 values of
0.08 V (∆Ep = 180 mV) and 0.65 V (∆Ep = 160 mV). The first oxidation is attributed to the oxidation of the indole nitrogen,88, 89 and the latter is attributed to the
amidine group. There is one irreversible reduction at 1.72 V associated with the
reduction of the nitro group.89, 90
Figure 2.16. Bisphenyl amidine, 5
2.3.1
Photoelectrochemistry
Given that the amidine group is electrochemically active, photoelectrochemical experiments were performed to determine the extent to which the photoisomerisation would affect the oxidation potential of the amidine group (oxidation event
around 0.7 V) in 1(o) and 2(c). This would be an important measure of the ability
30
Texas Tech University, Miranda C. Andrews, December 2019
Figure 2.17. (a) CV (100 mV/s) of a 1.0 mM solution of 1(o) and (b) 2(c) before and
after UV-irradiation in CH3 CN (0.1 M Bu4 NPF6 ) at a platinum working electrode.
Indicated peak potentials are in V vs. Fc/Fc+ . Time interval indicated is time of
irradiation. Arrows indicate direction of scan.
of the photoswitch to influence the electronic structure of a ligating group and, ultimately, the ligand field in a manner analogous to redox switchable ligands.91–94
Solutions of 1(o) and 2(c) were irradiated with 254 nm UV light at 30-minute intervals. After each 30-minute interval, a voltammogram was recorded. Photoisomerization of 1(o) to give 1(c), results in a cathodic shift of the oxidative wave to
0.44 V and an anodic shift of the reductive wave to 1.81 V (Fig. 2.17a, Table 2.2).
This implies that the amidine group of 1(c) is easier to oxidize by 0.23 V and would
therefore be a more reducing ligand if bound to a metal than the open form, 1(o).
This is consistent with the calculated change in the HOMO of 1, where the closed
form has a HOMO that is 0.19 eV less stable than the open form. For 2(c) at t = 0,
three redox responses were observed at 0.65, 0.08, and 1.72 V, respectively, but after
30 minutes of UV-irradiation to give 2(o), the redox responses shifted anodically
to 0.76, 0.01, and 1.62 V, respectively (Fig. 2.17b, Table 2.2). Here, the implication
is that the amidine group of 2(o) is harder to oxidize upon switching to 2(o) as
compared to 2(c) and would therefore be a more oxidizing ligand if bound to a
metal than the closed form. Since this is the second oxidative wave, the same analogy to the changes in the HOMO energy levels cannot be easily made. It should
be noted that 2(c) does not convert completely to 2(o) upon UV-irradiation (vide
31
Texas Tech University, Miranda C. Andrews, December 2019
supra). The observed redox responses are, therefore, a convolution of both 2(c) and
2(o) isomers.
Table 2.2. Redox potentials (V) of 1(o) and 2(c) before and after irradiation with 254
nm light for given time intervals. Calculations were done with B3LYP functional,
TZV(d) basis set.
1(o)
t=0
t = 30 min
Experimental Eox of 1(c) − Eox of 1(o)
Calculated EHOMO of 1(c) − EHOMO of 1(o)
2(c)
t=0
t = 30 min
t = 60 min
Experimental Eox of 2(o) − Eox of 2(c)
Calculated EHOMO of 2(o) − EHOMO of 2(c)
Eox of amidine
(V)
0.67
0.44
−0.23
−0.19 eV
E1/2ox
(V)
E1/2red
(V)
−2.11
−1.81
0.65
0.76
0.80
−0.08
0.01
0.01
0.09
0.05 eV
−1.72
−1.62
−1.63
Figure 2.18. Charge distribution from summed Hirshfeld charges for photoswitchable molecules. The color designates group of atoms to which the charge corresponds.
To corroborate the idea of using the switching event to push or pull electron density from the amidine group, the Hirshfeld charges were calculated and summed
across various groups within each molecule to provide a measure of the overall
change in electron density (or charge redistribution) upon switching. For 1, a net
32
Texas Tech University, Miranda C. Andrews, December 2019
increase in electron density was calculated for the amidine group following ring
closing, while a net decrease in electron density was calculated for the amidine
group in 2 upon ring opening (Fig. 2.18). These results are both consistent with the
changes observed by CV. For both 1 and 2, the implication is that the ligand field
generated by the amidine group could be modulated by affecting a cyclization or
ring opening event in a peripheral group.
2.3.2
Isolation of Photoswitched Compounds
Samples of 1(c) and 2(o) could be isolated. The removal of solvent from irradiated
solutions of 1(o) and 2(c) in DCM allowed for the isolation of pure 1(c) and 2(o) as
confirmed by 1 H NMR. CV‘s were recorded for the isolated switched forms (1(c)
and 2(o), and found to be virtually identical to the CVs that were recorded when
the compounds were irradiated in situ. The similar results from both experiments
on 2(o) are likely due to thermal equilibration that occurs upon dissolution.
2.3.3
Spectroelectrochemistry
Potentials were held constant and UV-vis absorption spectra were recorded over
time. Compound 1(o) shows a reversible reduction at 2.11 V, so the potential was
held at 2.35 V to ensure reduction occurred. Fig. 2.19(a) shows the absorption spectral changes during electrochemical reduction of 1(o). The resulting spectrum exhibits an absorption maximum at 521 nm as seen in Fig. 2.19(a). This is consistent
with the ring-closing observed from 1(o) to 1(c) via irradiation which gives rise to
a new absorption band centered at 531 nm, so we propose that this reduction event
leads to a ring closing to give [1(c)]•− similar to what has been previously reported
for related fulgide switches.87 The calculated singly occupied molecular orbital
(SOMO) of [1(o)]•− (Fig. 2.20(a)) illustrates how this electrochemically induced
ring closing mimics the photoinduced ring closing. Population of the LUMO of
1(o) gives a virtually identical SOMO with antibonding character at the CC double
bonds involved which facilitates conrotatory ring closing as shown in Fig. 2.20(b)
to give [1(o)]•− .
33
Texas Tech University, Miranda C. Andrews, December 2019
Figure 2.19. UV-vis spectra recorded during controlled-potential coulometry (2.35
V vs. Fc/Fc+ ) of a 1 mM solution at 298 K of (a) 1(o) in CH3 CN containing 0.1
M Bu4 NPF6 . Dashed grey line is a 25.0 µM solution of 1(c) in methanol scaled for
comparison to spectroelectrochemical spectra and (b) 2(c) in CH3 CN containing
0.1 M Bu4 NPF6 . Dashed grey line is a 0.1 mM solution of 2(o) in CH3 CN scaled for
comparison to spectroelectrochemical spectra.
Compound 2(c) undergoes oxidation at 0.08 V, so the potential was held at 0.20 V
to ensure oxidation occurred. Fig. 2.19(b), shows the changes to the absorption
spectrum during electrochemical oxidation of 2(c). A new peak is observed at 575
nm comparable to the new absorption band at 537 nm that appears upon thermal
or photochemical ring opening. This is consistent with oxidation and concomitant ring opening to give [2(o)]•+ in CH3 CN. Previous studies have proposed this
type of electrochemically induced ring opening of spiropyrans.84, 90, 95 Oxidative
CC coupling can occur at the position that is para to the indoline nitrogen when
that position is unsubstituted,88 but that position is substituted in 2 so no coupling
is expected. Given that irradiation induces a measurable change in the redox properties of the amidine group, the possibility that photoisomerization can influence
the redox properties of a coordinated metal seem likely.
2.4
Future Work: Synthetic Targets and Strategies
Several other potential molecular switches were probed computationally, and some
were synthesized. Although full photophysical studies have not yet been per34
Texas Tech University, Miranda C. Andrews, December 2019
Figure 2.20. (a) SOMO of [1(o)]•− and its rotamer (shown for clarification, 0.04
a.u. isosurfaces). (b) Conrotatory ring-closing mechanism for both 1(o) → 1(c) and
[1(o)]•− → [1(c)]•− .
formed, the computational results can give some insight into if and to what extent
molecular switches might be able to modulate the properties of appended groups.
These compounds will be studied more thoroughly in the future.
2.4.1
Amidines, Aminopyridines, and Squaramides
The previous study demonstrated, in part, that the changes in summed Hirshfeld
charges and frontier orbitals showed good correlation with what was observed experimentally for 1 and 2, particularly with regard to redox properties. This method
was applied to candidate molecules in order to determine if they are worth pursuing experimentally. Because 2 showed far greater changes both experimentally and
computationally, the spiropyran moiety was chosen as the model photoswitchable
moiety.
Because previous studies indicated that in the case of compound 2, changes in the
summed Hirshfeld charges of the amidine upon ring-opening of the spiropyran
correlated very well with the change in oxidation potential of the amidine group,
Hirshfeld charges were summed for other substituted spiropyrans. These substituents were chosen because there is literature precedence for similar groups in metal
complexes, which could have desirable properties that could be modulated by a
35
Texas Tech University, Miranda C. Andrews, December 2019
photoswitchable moiety. In the case of the squaramide, hydrogen bonding and ion
sensing could also potentially be modulated.
The changes in summed Hirshfeld charges for the compounds of interest are summarized in Table 2.3. In the case of both 1 and 2, ring-opening leads to an anodic
shift in the oxidation potential of the pendant amidine. This shift corresponds to a
net decrease in electron density at the amidine group, and therefore a net increase
in charge. In the case of 2, this change in charge of +0.011 corresponds to an anodic
shift of 150 mV under the conditions that CV of 2 were recorded.
Similar changes in charges of the methylamidine, aminopyridine, and squaramide groups should result in similar changes in redox potentials associtated with
those groups. Table 2.3 indicates that incorporating a photoswitchable moiety with
these proposed pendant groups should result in measurable changes in redox potential of these groups. The changes in summed Hirshfeld charges for 6, 7, and 9
upon ring-opening of the pendant spiropyran should result in similar if not greater
changes than those observed experimentally for 2.
Figure 2.21. Ligand scaffolds of interest for computational studies of (a) 6, (b) 7,
and (c) 9. Emphasis is placed on the groups highlighted in blue and red as these
are the pendant groups that a molecular switch could affect.
36
Texas Tech University, Miranda C. Andrews, December 2019
Table 2.3. Change in summed Hirshfeld charges for ligating groups of various
compounds upon ring-opening.∗ Refer to Fig. 2.21 for which atoms contribute.
Compound
1
2
6
7
9
Change in Summed Charges
+0.003∗
+0.011
+0.015
+0.023
+0.010†
+0.023‡
Figure 2.22. Summed Hirshfeld charges for ligating groups in (a) 6, (b) 7, and (c) 9.
2.4.1.1
Amidines
To make it more likely that 2 would be more stable in the conformation represented
by Fig. 2.6(d) and to simplify computational studies, the related methyl amidine
∗
For 1, this is the back-reaction as noted in previous figures
Group highlighted in blue in Fig. 2.21(c)
‡
Group highlighted in red in Fig. 2.21(c)
†
37
Texas Tech University, Miranda C. Andrews, December 2019
(R3 of Fig. 2.3) was probed computationally. This amidine (6, Fig. 2.22(a)) and
other functional groups that could be used as potential ligands metal complexes
and could be substituted with a spiropyran in few steps were modelled to determine if and to what extent a switchable moiety could be used to modulate the redox
properties of those functional groups.
Not unexpectedly, the frontier orbitals of 6 (Fig. 2.23) are very similar to those of
2. For 6(c) the HOMO and LUMO are orthogonal to each other, and any transition
associated with a HOMO-LUMO transition would likely not be observed in an absorption spectrum. Upon ring-opening, both the HOMO and the LUMO are more
spread out over the spiropyran structure due to increased conjugation upon photoisomerization. The HOMO of both closed and open states shows contribution
from the amidine ligating group, which indicates that the photoisomerization of 6
could be used to modulate the ligand field in a metal complex. It is worth noting
that the HOMO-1 of 6(c) matches better with the HOMO-2 of 6(o) as shown in Fig.
2.23. As summarized in Fig. 2.22(a) and Table 2.3, ring-opening of 6 would result
in a similar, perhaps greater change in electron density of the amidine group as
electron density is pulled away upon photoisomerization.
Figure 2.23. Frontier orbitals of 6(c) (top) and 6(o) (bottom) (0.04 a.u. isosurface
value).
2.4.1.2
Aminopyridines
Aminopyridinates form complexes with a variety of transition metals, particularly early transition metals in high oxidation states.96 Titanium and scandium
aminopyridinates have been used as polymerization catalysts,96, 97 and aminopyrid38
Texas Tech University, Miranda C. Andrews, December 2019
inates have also been used to stabilize quintuply-bonded chromium compounds
that can be used as starting materials for porous coordination polymers.98 Aminopyridinates can also be used as ligands in complexes with larger metal ions, including yttrium.97 Because the pyridyl group has low-lying π-orbitals with high
Np character, this group could function as a π-acceptor ligand. Complexes that
involve an aminopyridinate and electron-rich metal ions would likely show backbonding interactions. Due to the fact that the spiropyran molecular switches as
shown in Fig. 2.18 are more electron-withdrawing in the open state, it is very
likely that the back-bonding could be modulated by the photoswitch. Similar systems have shown that this effect on back-bonding can be measured indirectly using
IR spectroscopy.50
Attempts were made to synthesize 7, but isolation and purification proved to be
difficult. A procedure was adapted from literature procedures that used simpler
anilines and 2-bromopyridine to synthesize 2-aminopyridine compounds,99, 100 which
yielded bright blue and purple product mixtures that did not show any signals
other than solvent and water peaks by 1 H NMR spectra. If a procedure could be
developed for the synthesis of 7, the computational results discussed below could
inform behavior of 7 observed experimentally.
Similarly to previously discussed structures, the frontier orbitals of 7 give some additional insight into changes in electronic structure upon ring-opening. As shown
in Fig. 2.24, the potential ligating group contributes to the HOMO of both 7(c) and
7(o). The HOMO-1 of 7(c) and the HOMO-2 of 7(o) (which matches better than
the HOMO-1 of 7(o)) also show considerable contribution from the aminopyridyl
group. Upon ring-opening the electron density of these orbitals spreads out as a
result of increased conjugation. The LUMO‘s of 7(c) and 7(o) also show the increase in conjugation upon photoisomerization. These changes could be used to
explain changes in absorption spectra upon ring-opening of 7. Summed Hirshfeld
charges indicate that upon ring-opening, electron density is pulled away from the
aminopyridine group as seen in Fig. 2.22(b) and Table 2.3.
39
Texas Tech University, Miranda C. Andrews, December 2019
Figure 2.24. Frontier orbitals of 7(c) (top) and 7(o) (bottom) (0.04 a.u. isosurface
value).
2.4.1.3
Squaramides
Another system that was probed computationally was the squaramide group. Squaramides have been used as hydrogen-bond catalysts due to their chemical similarity
to urea and guanidine.101, 102 Squaramides owe their particularly good hydrogenbond acceptor ability to the fact that upon protonation, they gain considerable aromatic character.103 Squaramides have also been incorporated into metal-organic
frameworks (MOFs) as organocatalysts,104, 105 which provides an easily-isolated,
tunable, and thermally stable scaffold. Squaramides also show promise in the sensing of anions and cations.106, 107
There is no literature precedence for squaramides being used as dianionic ligands
in coordination complexes. In order to establish a procedure for synthesizing such
complexes and to establish a non-switchable control, attempts were made to deprotonate a simple bisaryl squaramide (8) and metalate with various first row
transition metals as well as with large main group metals, such as lead(II) and antimony(III). The large bite angle that would be associated with the deprotonated
squaramide would make it more likely that larger metals would coordinate, but
bimetallic complexes could also result from the squaramide(s) bridging the metals
or metal ions.
While 1 H NMR spectra indicated that the squaramide shown in Fig. 2.25 was,
at least partially deprotonated, attempts to subsequently metallate using divalent
metal chloride salts were unsuccessful. It should be noted that the 1 H NMR spectrum shown in Fig. 2.40 indicated that most of the resulting product mixture
40
Texas Tech University, Miranda C. Andrews, December 2019
Figure 2.25. Synthesis of bisarylsquaramide (8) and subsequent metallation attempt.
was squaramide that was successfully deprotonated at both nitrogens, but there
is some squaramide that is still protonated at one of the nitrogens (indicated by
peaks at 8.58, 7.79, and 7.40 ppm). IR spectra obtained under inert atmosphere
(Fig. 2.39) indicated that complete deprotonation was successful, so it is likely that
some protonation occurred between sample preparation and characterization by
1
H NMR spectroscopy.
An unexpected crystal structure was obtained from the treatment of the deprotonated ligand with CuCl2 . Fig. 2.26 shows the structure which, under the conditions used to attempt to deprotonate and metallate the squaramide, contains a
trispyridyl copper(I) chloride salt interacting with a neutral squaramide through
hydrogen bonds. This corroborates the ability of squaramides to act as anion
receptors. The mechanism of reduction of the copper(II) chloride salt is not yet
known.
Figure 2.26. Crystal structure of copper(I) chloride squaramide complex.
A squaramide with a pendant photoswitchable spiropyran group was designed
to evaluate the ability to control the the various interactions that squaramides can
undergo (vide supra). Compound 9(c) was synthesized and characterized, but ex41
Texas Tech University, Miranda C. Andrews, December 2019
perimentally, these squaramides proved difficult to study, as they showed poor
solubility in a number of non-coordinating organic solvents such as toluene, chloroform, or dichloromethane. This is likely due to the planarity of the squaramides
as shown in Fig. 2.26. The design of these switches could almost certainly be improved by the addition of alkyl groups, namely branched alkyl groups such as
t-butyl or isopropyl groups in place of the methyl groups in the 3- and 5-positions
that would decrease the planarity of the squaramide. There will also likely be a
change in the solubility of 9 upon photoswitching, based on the differences in solubility noted for 2 (vide supra).
Figure 2.27. Synthesis of 9(c)
The frontier orbitals of 9 were not as straightforward to interpret as those of 6 and
7. For some of the frontier orbitals, it was difficult to find orbitals that matched
well. As shown in Fig. 2.28, the HOMO-1 of 9(c) matches very well with the
HOMO−3 of 9(o). However, for the HOMO of 9(c), no obvious corresponding
orbitals were found among the frontier orbitals of 9(o), although the HOMO of
9(o) does share some features with that of 9(c). The increased conjugation can be
seen most easily in the HOMOs of 9(c) and 9(o). Changes in summed Hirshfeld
charges (Fig. 2.22(c) and Table 2.3) indicate that electron density is pulled away
from the squarmide groups upon ring-opening.
The use of molecular switches such as 9 could prove to be versatile. Because the
ring-opening of the spiropyran moiety makes that group more electron-withdrawing,
both the carbonyl groups and the amine groups become electron deficient relat42
Texas Tech University, Miranda C. Andrews, December 2019
Figure 2.28. Frontier orbitals of 9(c) (top) and 9(o) (bottom) (0.04 a.u. isosurface
value).
ive to the closed spiropyran state. This means that the photoswitching of the
spiropyran would likely result in turn-on functionality with regard to the squaramides hydrogen-bond donating ability. As seen in Fig. 2.29, the already considerable
Vmax of 9(c) increases from 283 kJ·mol−1 to 317 kJ·mol−1 . Photoswitching of 9(c)
would simultaneously lead to turn-off functionality with regard to its hydrogenbond accepting ability, as seen in Fig. 2.29. This could lead to a multi-functional
hydrogen-bond switch, that could be used for catalysis or perhaps sensing.
Change in summed charges for groups shown in Fig. 2.21 indicate that these
compounds are worth pursuing. These groups are proposed as ligating groups
for metal complexes that include switchable moieties similar to 4. In the case of
9, photoswitches could prove useful in the modulation of sensing or hydrogenbonding. It should be noted that only neutral molecules are being discussed in
this section. Changes in the neutral molecule may, but will not necessarily correlate with changes that might occur for an anionic ligand in a metal complex.
43
Texas Tech University, Miranda C. Andrews, December 2019
Figure 2.29. Electrostatic potential map depicting the carbonyl oxygens (top) and
amine protons (bottom) of 9 in (a) closed and (b) open states. Color map scale
shown is in kJ·mol−1 . Isosurface value is 0.001 a.u.
44
Texas Tech University, Miranda C. Andrews, December 2019
2.5
2.5.1
Synthetic and Experimental Methods
Isolation of 1(c)
Compound 1(o) (8.9 mg, 0.017 mmol) was dissolved in 0.50 mL of dichloromethane
in a large quartz test tube (15.5 cm × 2.5 cm) equipped with a stir bar. The test tube
was placed in a photoreactor and stirred for 30 minutes while irradiating with
350 nm light. After 30 minutes, the solvent was removed under vacuum while
continuing to irradiate. After the solvent was removed, the dry, red solid was
recovered quantitatively. The 1 H NMR shows that the resulting solid is a mixture
of 1(o) (6%) and 1(c) (94%).
Figure 2.30. Di-ATR FT-IR spectra of 1(o) and 1(c).
2.5.2
Isolation of 2(o)
Compound 2(c) (8.0 mg, 0.016 mmol) was dissolved in 0.50 mL dichloromethane
in a large quartz test tube (15.5 cm × 2.5 cm) with a stir bar. The test tube was
placed in a photoreactor and stirred for 15 minutes while irradiating with 350 nm
light. After 15 minutes, the solvent was removed using a Schlenk line while continuing to irradiate. After the solvent was removed, the dry, blue-purple solid was
recovered quantitatively. The 1 H NMR spectrum shows only one species, 2(o).
45
Texas Tech University, Miranda C. Andrews, December 2019
Figure 2.31. 1 H NMR spectrum of isolated 1 in d6 -DMSO.
Figure 2.32. Di-ATR FT-IR spectra of 2(c) and 2(o).
46
Texas Tech University, Miranda C. Andrews, December 2019
Figure 2.33. 1 H NMR spectrum of isolated 2 in d6 -DMSO.
47
Texas Tech University, Miranda C. Andrews, December 2019
2.5.3
Experimental Procedure and Data Fitting for Determination of
Activation Energy of Isomerization of 2
For the determination of activation energy, a solution (0.100 mM) of 2(c) in methanol (CH3 OH) was placed in the thermostated sample holder with stirring. The
thermal conversion of 2(c) to 2(o) was monitored by UV-vis absorption spectroscopy. Episodic data capture was used to ensure that spectra were taken at regular
intervals. Measurements were collected at intervals of approximately 30 seconds
over 2 hours at 323 K, intervals of approximately 1 minute over 4 hours at 313 K,
and intervals of approximately 3 minutes over 12 hours at 303 K.
Figure 2.34. Absorbance of a 0.100 mM solution of 2 in methanol at 537 nm as a
function of time at 323 K.
Figure 2.35. Absorbance of a 0.100 mM solution of 2 in methanol at 537 nm as a
function of time at 313 K.
k
1
*
In each case a generic reaction A −
)
− B, where A is 2(c) and B is 2(o), the forward
k2
reaction rate as written is represented by k1 and the reverse reaction rate by k2 ,
48
Texas Tech University, Miranda C. Andrews, December 2019
Figure 2.36. Absorbance of a 0.100 mM solution of 2 in methanol at 537 nm as a
function of time at 303 K.
was used to fit the data. The following form of rate law equations for first-order
reversible reactions were used:
k2 [A]0
k1 −(k1 +k2 )t
[A] =
1+ e
k1 + k2
k2
k1 [A]0
k1 −(k1 +k2 )t
[B] =
1− e
k1 + k2
k2
In order to use the absorbance data as obtained, the concentrations were then
multiplied by the molar extinction coefficients that were determined as described
above and the path length. The solver function of Microsoft Excel was then used
to minimize the sum of the squares of differences between the fitted and experimental data by varying the Keq (which is equal to k1 /k2 ) and k1 at each temperature. These values were then used to calculate k2 at each temperature. The
Eyring-Polanyi plot (Fig. 2.14) was used to determine the ∆H‡ and ∆S‡ for the
forward and for the reverse reactions per the linear Eyring-Polayi equation:
49
Texas Tech University, Miranda C. Andrews, December 2019
k
∆H ‡
ln = −
T
R
kB ∆S ‡
1
+ ln
+
T
h
R
where k is the rate constant in s−1 , T is the temperature in K, R is the ideal gas
constant in J·mol−1 ·K−1 , kB is Boltzmann‘s constant in J·K−1 , and h is Planck‘s
constant in J·s.
2.5.4
Attempted Synthesis of 7(c)
Figure 2.37. Scheme for attempted synthesis of 7(c)
This procedure was adapted from a literature procedure.100 2-Bromopyridine (30
mmol, 4.74 g) and 1,3,3-trimethyl-2-methyleneindolin-5-amine70 (3.0 mmol, 0.28 g)
was stirred at 170 ◦ C for 3.5 hours, then cooled to room temperature. The residue
was dissolved in ether and the solution was washed with aqueous saturated sodium bicarbonate and brine. The ether layer was separated and dried with magnesium sulfate. The ether was removed under vacuum to give the crude product,
which was purified by chromatography (silica gel, dichloromethane). This product
could not be isolated. 1 H NMR spectra were obtained in chloroform-d, d6 -DMSO,
and deutero-methanol, but in all cases, the only peaks observed were those of
solvent and water. The resulting solutions were shades of blue and purple, and
were not turbid, so it is not yet known what species was isolated.
50
Texas Tech University, Miranda C. Andrews, December 2019
2.5.5
Synthesis of 8
The neutral squaramide (8) was synthesized using a procedure adapted from literature.104 Diethylsquarate (4.0 mmol, 0.68 g) was stirred with 8 mmol (0.97 g)
of 3,5-dimethylaniline and 0.18 g (1.25 mol%) zinc trifluoromethane sulfonate in
methanol over 60-72 hours. The resulting squaramide was isolated by centrifugation and recovered in 83% (1.1 g) yield. 1 H NMR (400 MHz, d6 -DMSO) δ: 9.75 ppm
(2H, s) 7.08 ppm (4H, s), 6.72 ppm (2H, s), 2.26 ppm (12H, s).
Figure 2.38. 1 H NMR spectrum of 8 in d6 -DMSO.
2.5.6
Attempted Deprotonation and Subsequent Metallation of 8
In pyridine, 0.20 mmol (0.064 g) 8 was stirred with 0.42 mmol (0.21 mL) of lithium diisopropylamide (LDA) until the cloudy suspension became clear, indicating that 8 had been deprotonated. 0.20 mmol divalent metal chloride salt was
51
Texas Tech University, Miranda C. Andrews, December 2019
added to solution, and the solution was stirred overnight. Any solids produced
were collected by centrifugation, and the mother liquor was dried down. Resulting crude products were characterized by IR spectroscopy (Fig. 2.39) to determine
if the squaramide group was present in either phase and if it been deprotonated.
1
H NMR (400 MHz, d6 -DMSO) δ: 7.07 ppm (4H, m) 6.49 ppm (2H, s), 2.20 ppm
(12H, s). The peaks at 8.58, 7.79, and 7.40 ppm are possibly associated with a small
amount of singly-protonated squaramide 8 as similar signals were observed in the
products of mono-substitutions of diethylsquarate.
The single crystals of 8·CuCl(pyridine)3 were grown by the dropwise addition of
hexanes to the pyridine mother liquor after the addition of CuCl2 . Hexanes were
added until cloudiness persisted for a few seconds before redissolving upon stirring. This mother liquor was then placed in a freezer until crystals were observed.
Figure 2.39. Di-ATR FT-IR spectra of 8 before (solid black line) and after (hashed
grey line) deprotonation.
52
Texas Tech University, Miranda C. Andrews, December 2019
Figure 2.40. 1 H NMR spectrum of deprotonated 8 in d6 -DMSO.
53
Texas Tech University, Miranda C. Andrews, December 2019
2.5.6.1
Single Crystal X-ray Diffraction
The following crystal structure determination was performed by Dr. Daniel Unruh
and the Texas Tech Dept. of Chemistry and Biochemistry X-ray facility.
Data were collected on a Bruker PLATFORM three circle diffractometer equipped
with an APEX II CCD detector and operated at 1350 W (40 kV, 30 mA) to generate
(graphite monochromated) Mo Kα radiation (λ = 0.71073 Å). Crystals were transferred from the vial and placed on a glass slide in polyisobutylene. A Zeiss Stemi
305 microscope was used to identify a suitable specimen for X-ray diffraction from
a representative sample of the material. The crystal and a small amount of the oil
were collected on a MīTiGen cryoloop and transferred to the instrument where it
was placed under a cold nitrogen stream (Oxford) maintained at 100 K throughout
the duration of the experiment. The sample was optically centered with the aid
of a video camera to ensure that no translations were observed as the crystal was
rotated through all positions.
A unit cell collection was then carried out. After it was determined that the unit cell
was not present in the CCDC database a sphere of data was collected. Omega scans
were carried out with a 30 seconds per frame exposure time and a rotation of 0.50◦
per frame. After data collection, the crystal was measured for size, morphology,
and color. These values are reported in Table 2.4.
After data collection, the unit cell was re-determined using a subset of the full
data collection. Intensity data were corrected for Lorentz, polarization, and background effects using the Bruker program APEX 3. A semi-empirical correction
for adsorption was applied using the program SADABS.108 The SHELXL-2014,109
series of programs was used for the solution and refinement of the crystal structure. Hydrogen atoms bound to carbon and nitrogen atoms were located in the
difference Fourier map and were geometrically constrained using the appropriate
AFIX commands. The inconsistent reflections 8 1 3 and 4 2 9 were omitted.
54
Texas Tech University, Miranda C. Andrews, December 2019
Table 2.4. Crystal structure data for 8·CuCl(pyridine)3 .
Compound
Crystal color
Crystal habit
Empirical formula
Formula Weight
Temperature
Wavelength
Crystal system
Space group
Unit cell dimensions
Volume
Z
Calculated density
Absorption coefficient
F(000)
Crystal size
Theta range for data collection
Limiting indices
Reflections collected/unique
Completeness to theta = 25.242◦
Refinement method
Data/restraints/parameters
Goodness-of-fit on F2
Final R indices [I>2σ(I)]
R indices (all data)
Largest diff. peak and hole
8·CuCl(pyridine)3
Orange
Block
C35 H35 ClCuN5 O2
656.67 g·mol−1
100(2) K
0.71073 Å
Monoclinic
P21 /n
a = 17.4609(16) Å, α = 90◦
b = 8.3451(8) Å, β = 102.0840(10)◦
c = 22.090(2) Å, γ = 90◦
3147.4(5) Å3
4
1.386 g/cm3
0.819 mm−1
1368
0.360 × 0.145 × 0.110 mm
1.357 to 27.261◦
-22 ≤ h ≤ 22, -10 ≤ k ≤ 10, -28 ≤ l ≤ 28
35622/7021 [R(int) = 0.0497]
99.9%
Full-matrix least-squares on F2
7021/0/401
1.075
R1 = 0.0446, wR2 = 0.1159
R1 = 0.0576, wR2 = 0.1226
0.572 and -0.602 e·Å−3
55
Texas Tech University, Miranda C. Andrews, December 2019
2.5.7
Synthesis of 9(c)
Compound 9(c) was synthesized according to the scheme in Fig. 2.27 and based on
a similar literature procedure.104 2.0 mmol (0.34 g) diethylsquarate was stirred with
2.0 mmol (0.19 g) of 1,3,3-trimethyl-2-methyleneindolin-5-amine70 and 20 mol%
(0.4 mmol, 0.15 g) zinc trifluoromethanesulfonate in 2.0 mL methanol and 0.50
mL nitromethane for 16 hours at room temperature. 2.0 mL methanol and 2.0
mmol (0.19 g) aniline were added to the reaction mixture and was stirred for an
additional 24 hours. The solvent was removed under vacuum, and the resulting
solid was washed with ethanol and centrifuged 3 times to give the precursor to
9(c) (phenyl-indoline squaramide) in 95% yield (0.68 g). 1 H NMR (400 MHz, d6 DMSO) δ: 10.29 ppm (1H, very broad), 9.95 (1H, very broad), 7.50 ppm (3H, m),
7.38 ppm (3H, m), 7.14 ppm (2H, m), 3.45 ppm (2H, s), 2.09 ppm (3H, s), 1.64 ppm
(6H, s).
1 mmol (0.36 g) of the phenyl-indoline squaramide (vide supra) and 1 mmol 5nitrosalicylaldehyde was refluxed in absolute ethanol for 16 hours. After cooling
to room temperature, the resulting solid was filtered and washed with cold ethanol
to give 9(c) in 98% yield (0.50 g). 1 H NMR (400 MHz, d6 -DMSO) δ: 9.78 ppm (2H,
m), 8.24 ppm (1H, s), 8.03 ppm (1H, d), 7.50 ppm (2H, m), 7.38 ppm (2H, m), 7.31
ppm (1H, s), 7.24 ppm (2H, m), 7.19 ppm (1H, m), 7.08 ppm (1H, m), 6.95 ppm (1H,
d), 6.03 ppm (1H, d), 2.68 ppm (3H, s), 1.24 ppm (3H, s), 1.16 ppm (3H, s).
56
Texas Tech University, Miranda C. Andrews, December 2019
Figure 2.41. 1 H NMR spectrum of 9 in d6 -DMSO. Unlabeled peaks are associated
with nitromethane (4.42 ppm), ethanol (4.63, 3.44, and 1.06 ppm), and water (3.33
ppm).
57
Texas Tech University, Miranda C. Andrews, December 2019
CHAPTER 3
COMPUTATIONAL STUDIES OF MOLECULAR SWITCHES AS LIGANDS
IN METAL COMPLEXES
3.1
Metal Complexes
Organic compounds or moieties that interact with a metal or metal ion are known
as ligands. The ligand typically interacts with the metal by donating a pair of
weakly bound, sterically accessible electrons to the electron-deficient metal center.
The species that results from the interactions of the ligands and the metals or metal
ions are generally referred to as a metal complex or a coordination complex.
Figure 3.1. Generic octahedral metal complex
The interactions that occur between ligands and metal ions in complexes can vary
a great deal and depend on the electronic and steric properties of the ligand, the
d-electron count and oxidation state of the metal, as well as the resulting geometry
of the complex. When overlap occurs between the d-orbital of the metal and either
an s-orbital or a p-orbital of another species, the result is a σ-bond as shown in Fig.
3.2(a) and (b). In these cases, there is no node along the bonding axis. If a d-orbital
and a p- or π-orbital overlap as shown in Fig. 3.2(c), the result is a π-bond, which
has one node along the bonding axis.
In the instances where the ligand π-orbitals are occupied, one can consider a molecular orbital diagram as shown in Fig. 2.2. In these cases, the ligand would
be referred to as π-donating. If one can imagine being able to design different ligands with varying energies associated with those ligand π-orbitals (Fig. 2.2, purple
arrow), this would necessarily change the relative energy associated with the co58
Texas Tech University, Miranda C. Andrews, December 2019
Figure 3.2. Examples of orbital overlap in octahedral metal complexes: (a) d- and
s-orbital overlap that results in a σ-bond, (b) d- and p-orbital overlap that results
in a σ-bond, and (c) d- and p-orbital overlap that results in a π-bond.
ordination complexs molecular orbitals resulting from overlap involving those ligand π-orbitals (Fig. 2.2, red arrow).
A similar phenomenon can occur when ligands have empty π*-orbitals (higher in
energy than the π-orbitals, Fig. 2.1), and the coordination complex involves a more
electron-rich metal ion. In these cases, the metal is able to donate electron density
back into the π*-orbital of the ligand. This interaction is referred to as backbonding,
and the ligands would be referred to as π-acceptor ligands.
Because many molecular switches undergo a reorganization of π-electrons, it was
hypothesized that utilizing a π-donating or accepting ligating group would result
in the most change in the ligand field, which would be measured as a change in
the redox properties, or even in the magnetic properties of the metal center.
3.2
Calculations of M2+ Complexes
In order to determine if and to what extent the photoswitchable moieties could
be used to modulate the redox potential of a metal upon coordination of the photoswitchable ligand, DFT calculations were employed. First-row transition metals
were first chosen to model as they are abundant and relatively cheap, making them
desirable as reagents and catalysts. First-row transition metals form complexes
with aminidate ligands very well,71 which were discussed in the previous chapter.
59
Texas Tech University, Miranda C. Andrews, December 2019
3.2.1
Monometallic Complexes
Initially, first row transition metal complexes that contained only one photoswitchable moiety were studied. A four-coordinate system containing the conjugate base
of 6 and simpler biphenyl amidinate (conjugate base of 5) was chosen to study first.
Geometry optimizations were performed using PW91 functional110 and def2-SVP
basis set.111 Optimizations were performed on multiple spin states to as amidines
seem to fall in the middle of the spectrochemical series and give rise to both high
spin and low spin complexes.112, 113 Because first-row transition metals can adopt
both tetrahedral and square planar geometries, several spin states should be considered.
Trends of the first-row transition metal amidinate complexes did not produce many
unexpected results. As amidinate complexes are fairly ubiquitous, many studies
have already been conducted on their structural and electronic properties.71 The
calculated geometries agree with similar aminidate complexes for which the crystal structures are known.112, 113 This agreement lends some confidence to the trends
seen in the calculated structures and the energies associated with them.
The chromium(II) and manganese(II) amidinate complexes were calculated to be
most stable in the highest possible spin state, adopting planar and tetrahedral geometries, respectively, in both closed and open states. This is consistent with crystral structures of similar bisaminidate complexes .112 Copper(II) and zinc(II) complexes each have only one spin state possible, so calculations performed on these
structures also resulted in predicted square planar and tetrahedral geometries, respectively. The nickel(II) could only exist as a singlet in a square planar geometry
(although a high-spin case would be a triplet), and must be a triplet in a tetrahedral geometry. Nickel(II) amidinates are diamagnetic, and therefore must be in a
square planar geometry.113 DFT calculations support this as seen in Table 3.1, as
the triplet for both the open and closed states are less energetically favored.
60
Texas Tech University, Miranda C. Andrews, December 2019
Table 3.1. Structural and energy changes upon ring-opening in various first-row
transition metal complexes containing a photoswitchable aminidate ligand using
PW91 functional and def2-SVP basis set.
Metal
ion
Cr(II)
Mn(II)
Fe(II)
Co(II)
Ni(II)
Cu(II)
Zn(II)
S
0
2
1/2
5/2
1
2
1/2
3/2
0
1
1/2
0
Angle (◦ ) at
metal center∗
(c)/(o)
88.5/89.6
22.6/1.31
30.4/87.9
87.9/87.7
0.79/5.95
88.1/87.0
1.32/5.58
89.4/88.8
0.90/10.7
42.6/83.7
36.6/33.6
87.7/87.6
∆E upon
ring-opening
(kJ·mol−1 )
41.9
39.7
34.0
40.9
51.8
41.0
45.4
41.1
59.7
49.3
41.2
40.9
E relative to
low-spin, (c)
(kJ·mol−1 )
0.0
−132.9
0.0
−124.4
0.0
9.3
0.0
22.3
0.0
51.9
−
−
E relative to
low-spin, (o)
(kJ·mol−1 )
0.0
−135.1
0.0
−117.6
0.0
−1.4
0.0
18.1
0.0
41.4
−
−
As discussed previously in Chapter 2, changes in summed Hirshfeld charges of
the amidine group correlated well with changes observed in the experimental oxidation potential of that group. This strategy was applied to the divalent metal
ions in the complexes studied computationally herein. The differences in direction and magnitude of the changes in electron density likely have consequences
for the redox behavior and therefore chemistry that could take place at the metal.
As summarized in Table 3.2, ring-opening of the pendant spiropyran does not necessarily only result in withdrawing electron density from the metal ion. Ringopening does result in a calculated electron density loss at the metal center for
high-spin manganese(II), copper(II), and zinc(II). Conversely for high-spin chromium(II) and low-spin nickel(II), electron density is gained by the metal ion. It is
worth noting that the examples in which electron density is gained by the metal
ion upon switching of the spiropyran moiety experience more dramatic changes in
geometry upon switching as seen in Table 3.1. These changes in geometry would
∗
Angle between planes formed by each amidinate ligand (N−M−N and N’−M−N’), where
angles closer to 0◦ would indicate a square planar geometry and those closer to 90◦ would indicate
a tetrahedral geometry.
61
Texas Tech University, Miranda C. Andrews, December 2019
necessarily affect the relative energies of the frontier orbitals for these compounds,
leading to the need for a more inclusive explanation than solely accounting for the
open form of the photoswitchable ligand being more electron-withdrawing.
For both the high-spin and low-spin cobalt(II) complexes, electron density is pulled
away from the metal center upon ring-opening. There is precedence for a lightinduced charge-transfer and simultaneous change in spin state for a cobalt(III)
complex containing a spirooxazine ligand,54 so the proposed cobalt(II) complex
or a related cobalt(III) complex could have some very interesting applications if
synthesized.
The direction of the change in electron density for the proposed iron(II) complexes
was calculated to depend on the spin state of the iron(II). The low-spin iron(II)
gains electron density upon ring-opening of the pendant spiropyran, while the
high-spin iron(II) loses electron density upon switching. For both spin states, it
appears that the oxidation potential could be modulated by photoswitching the
spiropyran moiety.
Table 3.2. Changes in calculated charge of metal ion and frontier orbitals of firstrow transition metal amindinates using PW91 functional and XX basis set. Only
the most stable spin state was shown except for those that might undergo spin
crossover.
∆Hirshfeld charge of metal
∆E of
upon ring-opening
HOMO (eV)
Cr(II)
−0.0024
−0.25
Mn(II)
+0.0016
−0.17
Fe(II), spin 1
−0.0030
−0.24
Fe(II), spin 2
+0.0019
−0.15
Co(II), spin 1/2
+0.0016
−0.27
Co(II), spin 3/2
+0.0020
−0.15
Ni(II)
−0.0028
−0.11
Cu(II)
+0.0061
−0.21
Zn(II)
+0.0012
−0.13
Metal ion
∆E of
LUMO (eV)
−0.90
−0.87
−0.87
−0.84
−0.86
−0.86
−0.40
−0.83
−0.87
The frontier orbitals for the most stable spin state of each complex were visualized. For the high spin chromium(II) and manganese(II) cases, there appears to
62
Texas Tech University, Miranda C. Andrews, December 2019
be considerable loss in contribution of the metal centers to the HOMO-1 and the
HOMO, respectively upon ring-opening as seen in Figs. 3.3 and 3.4. The copper(II)
complex also shows less contribution to the HOMO by the metal ion (Fig. 3.5). In
all three of these cases, the divalent metal ion and aminidate groups contribute
far less to the molecular orbital of interest when the spiropyran moiety is in the
open conformation. Upon photoswitching, the HOMO is more diffuse, spreading
out over the spiropyran moiety. This could have consequences in experimental
absorption spectra of these compounds, particularly if these compounds exhibit
metal-to-ligand charge transfer (MLCT), ligand-to-metal charge transfer (LMCT),
or d-d transitions.
Figure 3.3. Frontier orbitals of photoswitchable Cr(II) complex. Occupied orbitals
shown are singly-occupied alpha orbitals (spin up). (0.04 a.u. isosurfaces)
On the other hand, the nickel(II) complex showed the most change in the distribution of the LUMO (Fig. 3.6). When the spiropyran in closed, the LUMO is localized
on the nickel(II) and nitrogens of the amidinate ligating groups. Upon switching of
the spiropyran, the LUMO is no longer localized, but is instead spread out over the
spiropyran moiety. There is still some contribution to the LUMO by the nickel(II)
and ligating groups, but that contribution is smaller than for the closed form. It
should also be noted that it appears the nickel(II) contributes to both the LUMO
and the LUMO+1 of the open form. This could also be reflected in a change in
reduction potential upon ring-opening. Zinc(II) complexes modelled show little to
63
Texas Tech University, Miranda C. Andrews, December 2019
Figure 3.4. Frontier orbitals of photoswitchable Mn(II) complex. Occupied orbitals
shown are singly-occupied alpha orbitals (spin up). (0.04 a.u. isosurfaces)
no contribution by zinc to the frontier orbitals, which is not unexpected, as zinc(II)
would be the redox-innocent control in experimental studies of these compounds.
64
Texas Tech University, Miranda C. Andrews, December 2019
Figure 3.5. Frontier orbitals of photoswitchable Cu(II) complex. The HOMO
shown is a singly-occupied alpha orbital (spin up). (0.04 a.u. isosurfaces)
Figure 3.6. Frontier orbitals of photoswitchable Ni(II) complex (0.04 a.u. isosurfaces).
65
Texas Tech University, Miranda C. Andrews, December 2019
Calculations of the iron(II) amidinate complexes using the PW91 functional110 indicate that when the spiropyran moiety is closed, the low spin triplet is more
stable. However, upon ring-opening of the spiropyran, the high spin quintet is
more stable. The calculated energy differences are quite small (Table 3.1), but
they certainly indicate that even incorporating one photoswitchable moiety into
the metal complex could result in modulation of spin state. In order for this to occur in the case of iron(II), calculations indicate that the complex would necessarily
have to undergo a distortion that results in a more tetrahedral-like geometry. Previous studies also indicate that in the absence of a photoswitchable moiety, iron(II)
amidines have a very low energy difference between the two geometries,112 and
the DFT calculations presented herein agree. Calculations of the cobalt(II) amidine
complex indicated that the high spin quartet was more stable in both closed and
open states, but because both iron and cobalt complexes are known to undergo
spin crossover, calculations were performed on these amidinate complexes with a
functional shown to be more appropriate for spin crossover complexes (vide infra).114
In the case of both low- and high-spin iron(II) complexes, the HOMO of the open
state shows a smaller contribution from the iron(II) and aminidate ligating groups
(Figs. 3.7 and 3.8).
The low- and high-spin states of the cobalt(II) complexes indicate more nuanced
correlation between spin state and contribution to frontier orbitals. In the lowspin state (Fig. 3.9), there appears to be very little change in the contribution of
the cobalt(II) center to the HOMO. There is, however, considerable change in the
metal center’s contribution to the LUMO and LUMO+1. Conversely, in the highspin state (Fig. 3.10), the contribution of the cobalt(II) metal center to the HOMO
is reduced upon photoswitching of the spiropyran moiety.
3.2.2
Spin Crossover
While using computational methods to elucidate any trends that might exist for
first-row transition metal complexes containing 6 as a ligand, the relative single
66
Texas Tech University, Miranda C. Andrews, December 2019
point energies of iron(II) complexes indicated that upon the ring-opening of 6, the
quintet state would be lower in energy than the triplet state. Because the triplet
is more stable for 6(c), this indicates that the ring-opening and -closing could be
used to modulate the spin state of the divalent metal ion. This ability to modulate the spin state is particularly intriguing as materials that undergo such changes
show particular promise in a multitude of applications. Certain spin states are often required for the activation of typically stable metal-oxo bonds in biological and
biomimetic water photooxidation systems115 or the use of biomimetic systems for
more controlled oxidation reactions.116, 117 While a majority of compounds that undergo a spin state change contain iron, there have been some interesting advances
in the use of photoswitchable ligands to induce a spin state change in a cobalt
complex.54
Figure 3.7. Frontier orbitals of low-spin photoswitchable Fe(II) complex (0.04 a.u.
isosurfaces).
In order to probe this idea further, different computational methods were used.
These different methods were chosen because the TPSSh functional118 has been
previously shown to give relative enthalpies that correlate well with experimental
changes in energy upon spin crossover for select iron(II) and cobalt(II) spin-crossover
complexes.114 The energy differences determined with this new functional did not
correlate well with the results using PW91. The TPSSh functional gave single point
energies that indicated that the iron(II) complex would likely not change its spin
state upon the ring-opening of 6, with the triplet state of both the opened and
closed states being more stable by about 100 kJ·mol−1 . The calculations run on the
67
Texas Tech University, Miranda C. Andrews, December 2019
Figure 3.8. Frontier orbitals of high-spin photoswitchable Fe(II) complex. Occupied orbitals shown are singly-occupied alpha orbitals (spin up). (0.04 a.u. isosurfaces)
cobalt(II) complex with the TPSSh functional gave similar, albeit opposite, results
as calculations run with PW91 functional, indicating the the high spin state would
be more stable by roughly 25 kJ·mol−1 in both closed and open states. The values for the cobalt(II) complex indicate that it could perhaps be pursued along with
the iron(II) complex experimentally to determine if complexes like these could be
modulated between low and high spin using light.
Table 3.3. Energy information using TPSSh functional and def2-SVP basis set.
∆SPE upon SPE relative to
ring-opening low-spin, (c)
(kJ·mol−1 )
(kJ·mol−1 )
Fe(II), 1
57.4
0
Fe(II), 2
44.9
106.7
Co(II), 1/2
50.1
0
Co(II), 3/2
46.6
−24.2
Metal
ion, S
68
SPE relative to
low-spin, (o)
(kJ·mol−1 )
0
94.2
0
−27.7
Texas Tech University, Miranda C. Andrews, December 2019
Figure 3.9. Frontier orbitals of low spin photoswitchable Co(II) complex. Note that
it appears the LUMO of the closed state appears to match the LUMO+1 of the open
state, and the LUMO+1 of the closed state is consistent with the LUMO of the open
state (0.04 a.u. isosurfaces).
Figure 3.10. Frontier orbitals of high spin photoswitchable Co(II) complex (0.04
a.u. isosurfaces).
69
Texas Tech University, Miranda C. Andrews, December 2019
3.2.3
Future Endeavors
The next steps of this project would be to synthesize and fully characterize the
metal complexes discussed above. This would involve spectroscopy as well as
CV to determine the experimental extent to which a photoswitch can be used to
modulate the redox properties of the metal centers. Initial attempts to synthesize
such metal complexes of with photoswitches 1 and 2 involved the use of a nucleophilic base (n-butyl lithium), which resulted in the decomposition of the photoswitchable amidine. These syntheses would need to be pursued with the use of
a non-nucleophilic base, such as lithium diisopropyl amide. Because aminidates
also support bimetallic complexes, these systems should be pursued alongside the
monometallic complexes. Those that form metal-metal bonds could be of particular interest, and the types and lengths of those bonds could be give insight into the
effects the photoswitchable moieties have on the metal ions and the changes could
easily be followed by Raman spectroscopy.
70
Texas Tech University, Miranda C. Andrews, December 2019
CHAPTER 4
TOWARDS SWITCHABLE HALOGEN BONDING
Interactions between atoms and molecules can vary in strength and length. The
strongest and shortest interactions are those of covalent and ionic bonds. However, there are many weaker interactions that can and do occur. Despite the relative weakness of these interactions, they can be just as important to the form and
function of molecules and supramolecular structures. These types of interactions
include dipole-dipole, ion-dipole, and dispersion forces. One of the most important weak interactions is the hydrogen bond, a particularly strong dipole-dipole interaction. The International Union of Pure and Applied Chemisty (IUPAC) defines
hydrogen bonds as “ ... an attractive interaction between a hydrogen atom from
a molecule or a molecular fragment XH in which X is more electronegative than
H, and an atom or a group of atoms in the same or a different molecule, in which
there is evidence of bond formation.”119 This difference in electronegativity results
in the more electronegative atom having a partial negative charge and the hydrogen a partial positive charge. Strong non-covalent interactions between the partial
positive and negative regions take place that have considerable consequences for
the physical properties of the compound. In the case of water, which hydrogen
bonds very well, the resulting physical properties have played an important and
vital role in life developing and continuing on Earth.
Another important dipole-dipole interaction is the halogen bond. The IUPAC has
defined halogen bonding as occurring “ ... when there is evidence of a net attractive interaction between an electrophilic region associated with a halogen atom in
a molecular entity and a nucleophilic region in another, or the same, molecular entity.”120 This type of interaction typically involves a very polarizable halogen, iodine. In order for this interaction to take place, the iodine must be bound through a
σ-bond to a group that is strongly electron-withdrawing. Due to the polarizability
of iodine, electron density is withdrawn from the iodine along the σ-bond, leaving
a partially positively charged region opposite the σ-bond, known as a σ-hole. This
region of positive charge can interact with the full or partial negative charge of
71
Texas Tech University, Miranda C. Andrews, December 2019
another species. This type of interaction, like hydrogen bonding, is highly directional, having angles very near to 180◦ .
Because interactions like hydrogen and halogen bonds result from primary interactions, they are referred to as secondary bonding interactions. These interactions
are typically longer than covalent bonds but shorter than the sum of the van der
Waals radii of the species involved. Alcock proposed that such interactions are the
result of directed forces (the primary bonds) and as a result, these interactions are
regular, understandable, and predictable.121
Terms such as “donor” and “acceptor” can sometimes be used interchangibly,
depending on the system being referenced. Throughout this chapter, the term
“donor” will refer to the halogen bond donor, which as the name denotes, donates
a halogen to the halogen bond “acceptor.”
Recent work exploring halogen bonding for anion-binding involves some very
sophisticated large molecules. For example, Gilday, et al. developed a zinc(II)
porphyrin-based system that interacts with halides and other anionic species. Association constants for this compound with halides were measured on the order
of 103 M−1 in chloroform. Association constants for interactions with other anions
such as acetate and sulfate were measured on the order of 105 M−1 to greater than
106 M−1 in chloroform.122
The Beer group has also developed two families of cyclic imidazolium-based halogen bond donors, one with phenyl group spacers123 and one with naphthyl group
spacers.124 The first halogen bond donors bind halides on the order of less than
10 M−1 up to nearly 103 M−1 . These halogen bond donors bind most strongly
with bromide by nearly an order of magnitude, and could therefore be considered
selective for bromide. The halogen bond donor containing the naphthyl spacers
interacted with bromide and iodide more strongly, with association constants of
104 M−1 to 105 M−1 . This second family of halogen bond donors also showed that
upon binding of these two halides, fluorescence increases. This type of halogen
bond donor could be used for sensing purposes, particularly given the turn-on
functionality of the fluorescence. It should be noted that in both of these examples
72
Texas Tech University, Miranda C. Andrews, December 2019
discussed, the anion binding experiments were conducted in aqueous methanol
solutions, which certainly competed with the halides during these experiments.
The Diederich group has designed some very sophisticated and highly ordered
halogen bonded systems.125 One half of this system contains four halogen bond
donors, and the other half has four halogen bond acceptors. The two halves are
designed so that when they interact, they form pill-shaped capsules with cavities
large enough to enclose one or more species. The highly ordered nature of this
supramolecular assembly leads to association constants of 104 M−1 to 105 M−1 . The
ability to tune the halogen bond donor and the halogen bond acceptor could prove
to have interesting applications in crystal engineering or perhaps in biomimetic
transport systems.
Halogen bonding has also been put to use in the degradation of chemical warfare agent simulants in MOFs. The Cohen group developed a modified UiO-66
MOF that incorporated iodine into the structure. The resulting MOF shows better
degradation of dimethyl-4-nitrophenyl phosphate (a chemical warfare agent simulant) than MOFs that were previously developed and used in the degradation of
chemical warfare agents.126
4.1
Remote Photoswitching of Halogen Bonding
Two iodospiropyrans were proposed and synthesized (Fig. 4.1) according to literature procedures70, 127–130 (Figs. 4.15 and 4.17). These compounds were studied computationally to determine if and to what extent the photoswitching of the
spiropyran could influence the halogen bond-donating ability of the iodine. Surface analyses were done to determine the extent of and visualize the changes in the
σ-hole of the iodine.
In order to determine which photoswitchable molecules should be targeted for
synthesis and potential halogen-bonding applications, DFT calculations, including
surface analyses were performed. The first potential halogen bonding switch synthesized was 10. Surface analysis calculations suggest that 10 could participate in
73
Texas Tech University, Miranda C. Andrews, December 2019
Figure 4.1. Compounds (a) 10 and (b) 11
halogen bonding. The σ-hole is shown in Fig. 4.2(a). With an electrostatic potential
(Vmax ) of 59 kJ·mol−1 , the σ-hole is comparable to that of iodobenzene (calculated
to be 63 kJ·mol−1 at the same level of theory). However, upon ring-opening, there
is a decrease in the Vmax of 10 to 43 kJ·mol−1 . This indicates that 10 would exhibit
“turn-off” functionality with regard to halogen bonding. This is not wholly unexpected as the phenolate moiety is electron-rich upon ring-opening. Without an
electron-withdrawing group to inductively increase electron density on the iodine,
the σ- hole is to some extent deactivated.
Conversely in the case of 11, the iodine is bound to the indoline moiety of the
spiropyran, which upon ring-opening, becomes more electron-withdrawing. Particularly, with the iodine para- to the indoline nitrogen, which has a formal positive
charge in the open state, the halogen bond donor could exhibit “turn-on” functionality. This hypothesis is consistent with calculations, which are visualized in Fig.
4.3 and which show that the Vmax of 11 increases from 62 kJ·mol−1 to 101 kJ·mol−1
upon ring-opening. Preliminary calculations were also done that included neutral
pyridine halogen bond acceptors. These results are summarized in Table 4.1. The
changes in nitrogen−iodine distances are consistent with 10 becoming a worse
halogen bond donor while 11 becomes a better halogen bond donor upon ringopening.
74
Texas Tech University, Miranda C. Andrews, December 2019
Figure 4.2. Electrostatic potential map of 10 in (a) closed and (b) open states. Color
map scale shown is in kJ·mol−1 . Isosurface value is 0.001 a.u.
Table 4.1. Calculated bond distances for 10 and 11 with neutral halogen bond
acceptors. Calculations were performed using PW91 functional and def2-TZVPP
basis set. For reference, the van der Waal radii for N and I are 1.55 Å and 2.00 Å,
respectively (sum of 3.55 Å).
Complex
10·pyridine
10·4-DMAP
10·4,4‘-biyridine
11·pyridine
11·4-DMAP
11·4,4‘-bipyridine
N−I distance closed (Å)
3.018
2.966
3.030
3.019
2.950
3.032
75
N−I distance open (Å)
3.052
3.000
3.064
2.949
2.879
2.957
Texas Tech University, Miranda C. Andrews, December 2019
Figure 4.3. Electrostatic potential map of 11 in (a) closed and (b) open states. Color
map scale shown is in kJ·mol−1 . Isosurface value is 0.001 a.u.
76
Texas Tech University, Miranda C. Andrews, December 2019
4.1.1
Crystal Structures
Crystal structures were determined for 10(c) and 11(c). Several attempts were
made to co-crystallize 10 and 11 with neutral and anionic halogen-bond acceptors. In an effort to avoid competition, non-coordinating solvents, were used for
co-crystallizations. Solutions of 10(c) and 11(c) were made with a small excess of
halogen bond acceptor. In the case of 10(c), the solutions often had to be heated in
order for the halogen bond donor to dissolve. Depending on the solvent, solutions
of 10 and halogen bond acceptors ranged from colorless, to orange and yellow, to
deep blue-violet. Crystals were deposited in several cases, but all structures were
determined to either be pure 10(c) or the halogen bond acceptor used.
Figure 4.4. Ball and stick models of (a) 10(c) and (b) 11(c) from single crystal structures.
Solutions of 11 and halogen bond acceptors were colorless to pale yellow. Although attempts were made to co-crystallize 11 with several neutral and anionic
halogen bond acceptors, the only single crystal obtained was 11(c) with 4,4’-bipyridine
(Fig. 4.5), which shows the nitrogen-iodine distance to be shorter than the sum of
the van der Waals radii (1.55 Å for nitrogen and 2.00 Å for iodine).131 One might
expect that one molecule of 4,4’-bipyridine could bridge two molecules of 11(c),
however, the crystal structure indicates that this is not the case. This may be a
result of packing within the crystal, so in solution, it may be still be possible to
observe two 11(c) molecules interacting with one molecule of 4,4’-bipyridine.
77
Texas Tech University, Miranda C. Andrews, December 2019
Figure 4.5. Ball and stick model of 11(c)·4,4‘-bipyridine from single crystral structure.
4.1.2
Switchable Behavior of Compound 11
Of the two photoswitchable halogen bond donors, only 11 showed photoswitchable behavior in non-coordinating solvents. Compound 10 appears to show thermoswitchable behavior (based on observations during attempts at co-crystallizations,
vide supra), but only in coordinating solvents. These solvents likely stabilize the
open merocyanine isomer, which is better represented by the zwitterionic resonance structure. The use of non-coordinating solvents is preferable as that eliminates the possibility for the solvent to compete with the intended halogen bond
acceptor.
Preliminary studies show that 11(c) can isomerize to 11(o) upon irradiation with
254, 300, or 350 nm light in dichloromethane. Compound 11(o) can be switched
back to 11(c) with blue light (464 nm). The absorption spectra for 11 can be seen in
Fig. 4.6. The photoisomerization of 11(c) to 11(o) can be monitored by the bands
at 317 and 435 nm, which grow in upon UV-irradiation, while the peak at 265
nm diminishes. The transitions associated with each of these peaks have not yet
been determined. The extent to which 11(c) converts to 11(o) also needs to be
determined, which will be useful for determining activation energy as well as for
solution state binding studies.
Preliminary studies were also performed to determine if 11 exhibits any thermallyinduced switching behavior. No conversion from 11(c) to 11(o) was observed upon
heating. However, 11(o) can be converted to 11(c) upon heating (Fig. 4.7). It should
be noted that it appears in Fig. 4.7 that the peak at 317 nm associated with 11(o)
is growing in instead of diminishing as would be expected. This experiment was
78
Texas Tech University, Miranda C. Andrews, December 2019
Figure 4.6. UV-vis spectra of 25 µM 11(c) (red) and 11(o) (black) in dichloromethane.
performed in dichloromethane, so it is very likely that the concentration is changing as a result of the evaporation of solvent. This would give the appearance of
peaks growing in. The important peak to follow in this experiment is the one at
415 nm, which is only associated with 11(o). This peak diminishes, which indicates
that 11(o) thermally isomerizes to 11(c).
Figure 4.7. Thermal reversion of 11(o) to 11(c) at 30 ◦ C in dichloromethane monitored by UV-vis absorption over 16 hours.
79
Texas Tech University, Miranda C. Andrews, December 2019
4.1.3
Future Endeavors
The next steps of this project would be to study the photophysical behaviors of
10 and 11 as well as to determine the ability of 10 and 11 to bind with neutral
and anionic species in solution. It would be beneficial to evaluate the switchable
behaviors in solvents with various dielectric constants, especially considering the
observed behavior and various colors of 10 during co-crystallization attempts discussed previously. Recall the changes in color upon heating solutions of 10(c),
particularly in donor solvents. The binding constants would need to be determined experimentally to confirm the computational results that have been shown
herein. The data fitting would be performed according to the methods described
in Chapter 6 of this work. Control experiments should also be conducted in which
spiropyran switches with poor halogen bond donor (for example, chlorine) substituents are also titrated with neutral and anionic halogen bond acceptors. These
would serve as good baselines, making it easier to determine which interactions
observed are the results of the halogen bond interactions and not the results of
other electrostatic interactions.
4.2
4.2.1
Toward Remote Electroswitching of Halogen Bonding
BODIPY-based Switches
While most of the studies in this work focus on photoswitchable molecules, a few
other types of switches have been explored. Because many of the interactions and
potential applications of the molecular switches discussed herein depend on the
switching event resulting in withdrawing electron density in a pendant group,
it was hypothesized that taking advantage of a redox-active group should give
similar, if not greatly enhanced, results upon oxidation. In an effort to continue
using spectroscopic methods to assess binding, a fluorescent, redox-active group
was chosen as the switch. Boron-dipyrromethenes (BODIPYs, Fig. 4.8) are robust
and versatile fluorophores that have been used as catalysts,132 chemosensors,133
and in photodynamic therapy.134 BODIPYs are particularly versatile because they
80
Texas Tech University, Miranda C. Andrews, December 2019
can be easily substituted to tune the absorption or emission wavelengths as well
as quantum yield and solubility.135
Figure 4.8. BODIPY core (unsubstituted BODIPY is not known in the literature135 ).
There is no literature precedence for the use of halo-substituted BODIPYs as switchable halogen bond donors. There are recent examples of redox-active groups being
used to modulate halogen bonding. Tetrathiafulvalene groups have been successfully used to modulate the halogen bond donor ability of a pendant iodine (Fig.
4.9).136, 137 These studies also suggested the use of cyclic voltammetry to determine the strength (binding constants) of the halogen bond interactions (although it
should be noted that the nature of solvent and supporting electrolyte necessarily
have to be taken into account when using this technique). The strongly electronwithdrawing BODIPY core should make such a compound an excellent halogen
bond donor.
Figure 4.9. Iodo-tetrathiafulvalene electroswitch.136, 137
4.2.2
Computational Studies of BODIPY Halogen Bond Donors
Two iodo-BODIPYs, compounds 12 and 13 (Fig. 4.10), were proposed. Compound
12 was synthesized according to literature procedures138 (Fig. 4.19). Attempts to
synthesize 13 were not successful, but the compound was studied computationally.
81
Texas Tech University, Miranda C. Andrews, December 2019
Figure 4.10. Compounds (a) 12 and (b) 13.
Preliminary solution state studies were performed on 12. However, they have so
far been inconclusive.
Surface analyses were performed to determine the extent to which a one-electron
oxidation might affect the Vmax values of the σ-holes of compounds 12 and 13. The
results are summarized in Figs. 4.11 and 4.12. Even before oxidation, neutral 12
shows a relatively large Vmax of 99 kJ·mol−1 and should show a propensity to halogen bond. Upon oxidation, a very large increase in Vmax to 388 kJ·mol−1 is calculated. Such a large increase, particularly compared to the proposed photoswitchable halogen bond donors, 10 or 11, is not unexpected as the entire compound has
a formal positve charge.
A similar increase is calculated for 13, which is calculated to have a Vmax of 93
kJ·mol−1 in the neutral state. Upon oxidation, the Vmax of [13]•+ is calculated to
be 417 kJ·mol−1 . Again, this very large increase is halogen bonding ability is not
unexpected due to the formal positive charge of [13]•+ .
4.2.3
Crystal Structures
Preliminary calculations were performed on neutral pyridine adducts to approximate halogen bond capabilities of 12. As expected, the distances between the
pyridyl nitrogens of the halogen bond acceptors and the iodines of 12 decrease
upon oxidation to give [12]•+ , from 2.95 Å to 2.71 Å.
82
Texas Tech University, Miranda C. Andrews, December 2019
Figure 4.11. Electrostatic potential map of 12 in (a) neutral and (b) one-electron
oxidized states. Color map scale shown is in kJ·mol−1 . Isosurface value is 0.001
a.u.
A co-crystal of 12 and 4,4‘-bipyridine was grown by slow evaporation of a dichloromethane solution of a 1:2 mixture of 12 and 4,4‘-bipyridine. This particular adduct was not calculated, but the intermolecular distances are shorter than
the sum of the van der Waals radii. This confirms that a halogen bonding interaction is taking place within the co-crystal as seen in Fig. 4.13. Contrary to the cocrystal grown of the photoswitchable halogen bond donor, 11, and 4,4’-bipyridine,
the crystal structure of 12 and 4,4’-bipyridine shows that one molecule of 4,4’bipyridine does bridge two molecules of 12. This is not shown in Fig. 4.13, but
the N−I bond distances are identical (2.982 Å).
While attempting to grow single crystals of 12 with halogen bond acceptors, an unexpected structure was obtained. While growing crystals of 12 and 4,4‘-bipyridine,
the experiment was initially done in ambient conditions, i.e., in air and ambient
light. BODIPYs are known sensitizers, and in the presence of oxygen and light,
can facilitate the generation of singlet oxygen.138–141 Although it has not yet been
confirmed beyond the crystal structure depicted in Fig. 4.14(b), it is hypothesized
that 12 was excited to a triplet state under ambient light conditions. Compound 12
should have a relatively long-lived excited state that could undergo inter-system
83
Texas Tech University, Miranda C. Andrews, December 2019
Figure 4.12. Electrostatic potential map of 13 in (a) neutral and (b) one-electron
oxidized states. Color map scale shown is in kJ·mol−1 . Isosurface value is 0.001
a.u.
Figure 4.13. Ball and stick model of 12·4,4‘-bipyridine from single crystal structure.
crossing to give an excited singlet state. This excited singlet state could then act
as a sensitizer, transferring energy to a molecule of oxygen, which would then
exist in an excited singlet state. Singlet oxygen is very reactive, and can act as a
mediator in oxidative coupling reactions as well as function as an efficient oxidant.140 It is hypothesized that such an oxidative coupling reaction occurred under
the conditions used to grow single crystals of 12·4,4’-bipyridine, which resulted in
the crystal structure of 14·4,4’-bipyridine depicted in Fig. 4.14(b).
This crystal structure also indicates that binding of one halogen bond acceptor to
one iodine of 14 may decrease the halogen bond donor ability of the opposite the
BODIPY core. The N−I distances for one BODIPY core differ as seen in Fig. 4.14.
84
Texas Tech University, Miranda C. Andrews, December 2019
Figure 4.14. (a) Structure of 14 and (b) ball and stick model of 14·(4,4’-bipyridine)4
from the single crystal structure, which may have resulted from a reaction with
singlet oxygen.
Further computational studies should be done to investigate this idea of negative
cooperativity.
4.2.4
Future Endeavors
The next steps for this project start with assessing the redox behavior of the iodoBODIPYs using cyclic voltammetry and perhaps spectroelectrochemistry. The solution-state binding of the BODIPYs would then need to be determined. Because
there are already scripts that utilize changes in UV-vis absorption to determine
binding constants, and BODIPYs are highly colored species, this would likely be
the best way to determine halogen bond-donating ability in solution, although it
may not be difficult to adapt the scripts to fluorescence spectroscopy. Titrations
85
Texas Tech University, Miranda C. Andrews, December 2019
could also be monitored by CV, but scripts would have to be written for this type
of titration.
4.3
4.3.1
Synthetic and Experimental Methods
Synthesis of 10(c)
Figure 4.15. Synthesis of 10(c)
The synthesis of 10 was adapted from literature procedures.127, 128 Salicylaldehyde
(82.0 mmol, 8.75 mL) added to solution of ICl (91 mmol, 4.6 mL) in glacial acetic
acid. The mixture was stirred at room temperature for 48 hours. The mixture
was then stirred an additional 24 hours at 40 ◦ C. The reaction mixture was concentrated under vacuum, and the residue diluted with water and dichloromethane. The organic layer was separated, and the aqueous layer was extracted with
dichloromethane. The combined organic layers were successively washed with
10% wt% aqueous sodium thiosulfate and brine. The dichloromethane solution
was dried with magnesium sulfate, filtered, and concentrated under vacuum. The
concentrated solution was then left overnight, and the pale yellow-brown solid
(5-iodosalicylaldehyde) precipitated out (12 g, 58% yield). 1 H NMR (400 MHz,
CDCl3 ) δ: 10.98 ppm (1H, s), 9.87 ppm (1H, d), 7.88 ppm (1H, d), 6.84 ppm (1H, d).
Equimolar (10.0 mmol) amounts of 5-iodosalicylaldehyde (2.50 g) and 1,3,3-trimethyl2-methyleneindoline (1.73 g) were refluxed in absolute ethanol for 16 hours. After
cooling to room temperature, the pale yellow solid was filtered (3.52 g, 87% yield).
1
H NMR (400 MHz, CDCl3 ) δ: 7.35 ppm (2H, m), 7.18 ppm (1H, m), 7.08 ppm (1H,
d), 6.85 ppm (1H, m), 6.76 ppm (1H, d), 6.50 ppm (2H, m), 5.72 ppm (1H, d), 2.71
ppm (3H, s), 1.24 ppm (3H, s), 1.15 ppm (3H, s).
86
Texas Tech University, Miranda C. Andrews, December 2019
Figure 4.16. 1 H NMR spectrum of 10 in d6 -DMSO.
Single crystals of 10(c) were obtained by heating 10 in ethanol and cooling slowly.
Upon heating, the colorless solid slowly dissolved to give a deep blue solution.
Upon slow cooling of the blue solution, which is believed to be 10(o) due to solubility and color changes, colorless crystals were deposited.
4.3.1.1
Single Crystal X-ray Diffraction
All crystal structure determinations included in this chapter were performed by
Dr. Daniel Unruh and the Texas Tech Dept. of Chemistry and Biochemistry X-ray
facility.
Data were collected on a Bruker PLATFORM three circle diffractometer equipped
with an APEX II CCD detector and operated at 1350 W (40 kV, 30 mA) to generate
87
Texas Tech University, Miranda C. Andrews, December 2019
(graphite monochromated) Mo Kα radiation (λ = 0.71073 Å). Crystals were transferred from the vial and placed on a glass slide in polyisobutylene. A Zeiss Stemi
305 microscope was used to identify a suitable specimen for X-ray diffraction from
a representative sample of the material. The crystal and a small amount of the oil
were collected on a Mı̈TiGen cryoloop and transferred to the instrument where it
was placed under a cold nitrogen stream (Oxford) maintained at 100 K throughout
the duration of the experiment. The sample was optically centered with the aid
of a video camera to ensure that no translations were observed as the crystal was
rotated through all positions.
A unit cell collection was then carried out. After it was determined that the unit cell
was not present in the CCDC database a sphere of data was collected. Omega scans
were carried out with a 5 seconds per frame exposure time and a rotation of 0.50◦
per frame. After data collection, the crystal was measured for size, morphology,
and color. These values are reported in Table 4.2.
After data collection, the unit cell was re-determined using a subset of the full data
collection. Intensity data were corrected for Lorentz, polarization, and background
effects using the Bruker program APEX 3. A semi-empirical correction for adsorption was applied using the program SADABS.108 The SHELXL-2014,109 series of
programs was used for the solution and refinement of the crystal structure. Hydrogen atoms bound to carbon atoms were located in the difference Fourier map
and were geometrically constrained using the appropriate AFIX commands. The
rigid-bond restraint RIGU was applied globally. There was a spurious peak 0.90 Å
from the atom site I2 that various absorption corrections did not resolve and was
left unidentified.
88
Texas Tech University, Miranda C. Andrews, December 2019
Table 4.2. Crystal structure data for 10.
Compound
Crystal color
Crystal habit
Empirical formula
Formula Weight
Temperature
Wavelength
Crystal system
Space group
Unit cell dimensions
Volume
Z
Calculated density
Absorption coefficient
F(000)
Crystal size
Theta range for data collection
Limiting indices
Reflections collected/unique
Completeness to theta = 67.679◦
Refinement method
Data/restraints/parameters
Goodness-of-fit on F2
Final R indices [I>2σ(I)]
R indices (all data)
Largest diff. peak and hole
10
Colorless
Blade
C19 H18 INO
403.24 g·mol−1
100(2) K
0.71073 Å
Monoclinic
P21 /n
a = 13.596(5) Å, α = 90◦
b = 14.444(5) Å, β = 93.460(11)◦
c = 16.977(6) Å, γ = 90◦
3328(2) Å3
8
1.610 g/cm3
1.926 mm−1
1600
0.480 × 0.210 × 0.195 mm
1.853 to 25.349◦
-16 ≤ h ≤ 16, -17 ≤ k ≤ 17, -20 ≤ l ≤ 20
32671/6089 [R(int) = 0.0436]
100.0%
Full-matrix least-squares on F2
6089/390/403
1.058
R1 = 0.0437, wR2 = 0.0866
R1 = 0.0578, wR2 = 0.0915
2.268 and -1.178 e·Å−3
89
Texas Tech University, Miranda C. Andrews, December 2019
4.3.2
Synthesis of 11(c)
Figure 4.17. Synthesis of 11(c)
The synthesis of 11(c) was adapted from literature procedures.70, 129, 130, 142 A solution containing 4-iodoaniline (10.0 g, 45.7 mmol) and hydrochloric acid (5.5 M, 10
mL) was cooled to -10 ◦ C and sodium nitrite (6.30 g, 91.5 mmol) in 25 mL of water
was added dropwise with continuous stirring. The suspension was allowed to stir
for 30 minutes and then an ice cold solution of Tin dichloride dihydrate (34.0 g,
151 mmol) in 15 mL of concentrated hydrochloric acid was added dropwise, keeping the temperature at -10 ◦ C. The reaction mixture was stirred at that temperature
for 1.5 hours and then for a further 18 hours at 5 ◦ C. The light brown precipitate
obtained was filtered and washed thrice with water. This solid was then stirred
with a saturated sodium hydroxide solution (25 mL) and extracted with ether (50
mL). The ether layer was washed with a saturated aqueous solution of sodium hydroxide (15 mL×2), sodium thiosulfate (15 mL×2), and finally with water (25 mL).
After drying over anhydrous magnesium sulfate, the ether layer was evaporated
to dryness to afford 4-iodophenylhydrazine (2.82 g, 26% yield) as a brown powder.
1
H NMR (400 MHz, CDCl3 ) δ: 7.48 ppm (2H, d), 6.63 ppm (2H, d), 5.18 ppm (1H,
s), 3.56 ppm (2H, s).
4-Iodophenylhydrazine (12.0 mmol, 2.82 g) was dissolved in absolute ethanol. 3Methyl-2-butanone (14.4 mmol, 1.55 mL) was added slowly with stirring. The
solution was refluxed for 1 hour. The ethanol was removed under vacuum and 50
mL of glacial acetic acid was added. The mixture was refluxed for 2 hours. The
acetic acid was removed, and 50 mL ether was added to the residue. Any resulting
90
Texas Tech University, Miranda C. Andrews, December 2019
precipitate was removed by filtration. The solution was washed with saturated
sodium hydroxide (2×25 mL), saturated sodium thiosulfate (2×25 mL), and then
with water (50 mL). The organic layer was dried with sodium sulfate, filtered, and
the solvent removed under vacuum to give 5-Iodo-2,3,3-trimethyl-3H-indole. (2.34
g, 61% yield). 1 H NMR (400 MHz, CDCl3 ) δ: 7.60 ppm (2H, m), 7.28 ppm (1H, d),
2.24 ppm (3H, s), 1.20 ppm (6H, s).
5-Iodo-2,3,3-trimethyl-3H-indole (8.22 mmol, 0.401 g) and methyl iodide (10.3 mmol,
0.630 mL) were refluxed in acetonitrile for 16 hours. Any precipitate was filtered
and washed with acetonitrile. The solvent was removed under vacuum. The resulting residue was stirred in solution of toluene/1 M aqueous sodium hydroxide
for 2 hours at room temperature. The mixture was then filtered to afford a colorless solid (1.14 g, 46% yield). 1 H NMR (400 MHz, CDCl3 ) δ: 7.52 ppm (2H, m), 7.26
ppm (1H, m), 3.88 ppm (2H, s), 2.98 ppm (3H,s), 1.33 ppm (6H, s).
Equimolar (3.80 mmol) amounts of 5-iodo-1,3,3-trimethyl-2-methyleneindoline (1.14
g) and 5-nitrosalicylaldehyde (0.635 g) were refluxed in absolute ethanol overnight.
After cooling to room temperature, the ethanol was removed. The crude product
was purified by filtration over a silica pad eluted with 1:1 dichloromethane/hexanes.
The solvent was removed under vacuum. to yield a pale yellow solid (1.26 g, 74%
yield). 1 H NMR (400 MHz, CDCl3 ) δ: 8.08 ppm (1H, s), 7.98 ppm (1H, m), 7.40
ppm (1H, m), 7.30 ppm (1H, d), 7.05 ppm (1H, m), 6.39 ppm (1H, d), 5.93 ppm (1H,
d), 2.68 ppm (3H, s), 1.22 ppm (3H, s), 1.14 ppm (3H, s).
Single crystals of 11(c) were grown by slow evaporation of an ethanol solution of
11. Single crystals of 11(c)·4,4’-bipyrdine were grown by slow evaporation of a
dichloromethane solution of equimolar amounts of 11 and 4,4’-bipyridine.
91
Texas Tech University, Miranda C. Andrews, December 2019
Figure 4.18. 1 H NMR spectrum of 11 in CD3 OD.
92
Texas Tech University, Miranda C. Andrews, December 2019
4.3.2.1
Single Crystal X-ray Diffraction
Data were collected on a Bruker PLATFORM three circle diffractometer equipped
with an APEX II CCD detector and operated at 1350 W (40 kV, 30 mA) to generate
(graphite monochromated) Mo Kα radiation (λ = 0.71073 Å). Crystals were transferred from the vial and placed on a glass slide in polyisobutylene. A Zeiss Stemi
305 microscope was used to identify a suitable specimen for X-ray diffraction from
a representative sample of the material. The crystal and a small amount of the oil
were collected on a MīTiGen cryoloop and transferred to the instrument where it
was placed under a cold nitrogen stream (Oxford) maintained at 100 K throughout
the duration of the experiment. The sample was optically centered with the aid
of a video camera to ensure that no translations were observed as the crystal was
rotated through all positions.
A unit cell collection was then carried out. After it was determined that the unit cell
was not present in the CCDC database a sphere of data was collected. Omega scans
were carried out with a 40 seconds per frame exposure time and a rotation of 0.50◦
per frame. After data collection, the crystal was measured for size, morphology,
and color. These values are reported in Table 4.3.
After data collection, the unit cell was re-determined using a subset of the full data
collection. Intensity data were corrected for Lorentz, polarization, and background
effects using the Bruker program APEX 3. A semi-empirical correction for adsorption was applied using the program SADABS.108 The SHELXL-2014,109 series of
programs was used for the solution and refinement of the crystal structure. Hydrogen atoms bound to carbon atoms were located in the difference Fourier map
and were geometrically constrained using the appropriate AFIX commands.
93
Texas Tech University, Miranda C. Andrews, December 2019
Table 4.3. Crystal structure data for 11.
Compound
Crystal color
Crystal habit
Empirical formula
Formula Weight
Temperature
Wavelength
Crystal system
Space group
Unit cell dimensions
Volume
Z
Calculated density
Absorption coefficient
F(000)
Crystal size
Theta range for data collection
Limiting indices
Reflections collected/unique
Completeness to theta = 67.679◦
Refinement method
Data/restraints/parameters
Goodness-of-fit on F2
Final R indices [I>2σ(I)]
R indices (all data)
Largest diff. peak and hole
11
Colorless
Plate
C19 H17 IN2 O3
448.24 g·mol−1
100(2) K
0.71073 Å
Monoclinic
P21 /c
a = 8.436(2) Å, α = 90◦
b = 19.173(5) Å, β = 107.721(2)◦
c = 11.371(3) Å, γ = 90◦
1751.9(7) Å3
4
1.699 g/cm3
1.848 mm−1
888
0.315 × 0.210 × 0.015 mm
2.124 to 26.404◦
-10 ≤ h ≤ 10, -23 ≤ k ≤ 23, -14 ≤ l ≤ 14
14599/3585 [R(int) = 0.0524]
100.0%
Full-matrix least-squares on F2
3585/0/229
1.029
R1 = 0.0434, wR2 = 0.0919
R1 = 0.0661, wR2 = 0.0998
0.975 and -0.874 e·Å−3
94
Texas Tech University, Miranda C. Andrews, December 2019
Data were collected on a Rigaku XtaLAB Synergy-i Kappa diffractometer equipped
with a PhotonJet-i X-ray source operated at 50 W (50 kV, 1 mA) to generate Cu
Kα radiation (λ = 1.54178 Å) and a HyPix-6000HE HPC detector. Crystals were
transferred from the vial and placed on a glass slide in polyisobutylene. A Zeiss
Stemi 305 microscope was used to identify a suitable specimen for X-ray diffraction
from a representative sample of the material. The crystal and a small amount of
the oil were collected on a MīTiGen cryoloop and transferred to the instrument
where it was placed under a cold nitrogen stream (Oxford) maintained at 100 K
throughout the duration of the experiment. The sample was optically centered
with the aid of a video camera to ensure that no translations were observed as the
crystal was rotated through all positions.
A unit cell collection was then carried out. After it was determined that the unit
cell was not present in the CCDC database a data collection strategy was calculated
by CrysAlisP ro .143 The crystal was measured for size, morphology, and color. These
values are reported in Table 4.4.
After data collection, the unit cell was re-determined using a subset of the full
data collection. Intensity data were corrected for Lorentz, polarization, and background effects using the CrysAlisP ro .143 A numerical absorption correction was
applied based on a Gaussian integration over a multifaceted crystal and followed
by a semi-empirical correction for adsorption applied using the program SCALE3
ABSPACK.144 The SHELXL-2014,109 series of programs was used for the solution
and refinement of the crystal structure. Hydrogen atoms bound to carbon atoms
were located in the difference Fourier map and were geometrically constrained
using the appropriate AFIX commands.
95
Texas Tech University, Miranda C. Andrews, December 2019
Table 4.4. Crystal structure data for 11·4,4‘-bipyridine.
Compound
Crystal color
Crystal habit
Empirical formula
Formula Weight
Temperature
Wavelength
Crystal system
Space group
Unit cell dimensions
Volume
Z
Calculated density
Absorption coefficient
F(000)
Crystal size
Theta range for data collection
Limiting indices
Reflections collected/unique
Completeness to theta = 67.679◦
Refinement method
Data/restraints/parameters
Goodness-of-fit on F2
Final R indices [I>2σ(I)]
R indices (all data)
Largest diff. peak and hole
11·4,4’-bipyridine
Pale yellow
Plate
C29 H25 IN4 O3
604.43 g·mol−1
100(2) K
1.54178 Å
Triclinic
P1̄
a = 7.67420(10) Å, α = 93.1300(10)◦
b = 10.194710(10) Å, β = 93.8800(10)◦
c = 16.9531(2) Å, γ = 104.3040(10)◦
12.78.71(3) Å3
2
1.570 g/cm3
10.155 mm−1
608
0.087 × 0.066× 0.017 mm
2.619 to 77.324◦
-9 ≤ h ≤ 9, -12 ≤ k ≤ 12, -20 ≤ l ≤ 21
37846/5302 [R(int) = 0.0532]
100.0%
Full-matrix least-squares on F2
5302/0/337
1.094
R1 = 0.0253, wR2 = 0.0647
R1 = 0.0264, wR2 = 0.0654
0.692 and -0.672 e·Å−3
96
Texas Tech University, Miranda C. Andrews, December 2019
4.3.3
Synthesis of 12
Figure 4.19. Synthesis of 12
The synthesis of 12 was adapted from a literature procedure.138 Under a nitrogen
atmosphere, 21 mmol of benzoyl chloride and 40 mmol 2,4-dimethyl pyrrole were
stirred in anhydrous dichloromethane at room temperature overnight. The reaction mixture was cooled in an ice bath, 20 mL (0.15 mol) triethylamine followed by
20 mL (0.16 mol) boron trifluoride etherate were added maintaining a temperature
of 0-5 ◦ C. The reaction mixture was stirred for an additional hour at this temperature. The mixture was then poured onto 200 mL of water. The organic layer was
separated and dried with magnesium sulfate. The dichloromethane was removed
under vacuum. The resulting crude material was purified by filtration through a
silica plug, eluted with 9:1 pentane/ethyl acetate. The isolated material (8-phenyl1,3,5,7-tetramethylBODIPY) was not pure by 1 H NMR but contained 2,4-dimethyl
pyrrole as seen by peaks at 5.88, 2.30, and 1.93 in Fig. 4.20. The remaining peaks
associated with the pyrrole starting material overlap with some of the peaks associated with the product in the aromatic region. 1 H NMR (400 MHz, CDCl3 ) δ: 7.48
ppm (3H, m), 7.29 ppm (2H, m), 5.98 ppm (2H, s), 2.56 ppm (6H, s), 1.37 ppm (6H,
s).
97
Texas Tech University, Miranda C. Andrews, December 2019
Figure 4.20. 1 H NMR spectrum of 8-phenyl-1,3,5,7-tetramethylBODIPY in CDCl3 .
98
Texas Tech University, Miranda C. Andrews, December 2019
The isolated 8-phenyl-1,3,5,7-tetramethylBODIPY was then stirred at room temperature in anhydrous dichloromethane with 0.100 mol N-iodosuccinimide for 30
minutes. The reaction mixture was then concentrated under vacuum, and the
crude product was purified by filtration over a silica plug in a 2:1 solution of hexanes/dichloromethane. The red band was collected and the solvent removed to
give 12, although the 1 H NMR of the isolated material shows some impurities
(Fig. 4.21). 1 H NMR (400 MHz, CDCl3 ) δ: 7.53 ppm (3H, m), 7.26 ppm (3H, m),
2.65 ppm (6H, s), 1.38 ppm (6H, s).
Figure 4.21. 1 H NMR spectrum of 12 in CDCl3 .
99
Texas Tech University, Miranda C. Andrews, December 2019
4.3.3.1
Single Crystal X-ray Diffraction
Data were collected on a Rigaku XtaLAB Synergy-i Kappa diffractometer equipped
with a PhotonJet-i X-ray source operated at 50 W (50 kV, 1 mA) to generate Cu
Kα radiation (λ = 1.54178 Å) and a HyPix-6000HE HPC detector. Crystals were
transferred from the vial and placed on a glass slide in polyisobutylene. A Zeiss
Stemi 305 microscope was used to identify a suitable specimen for X-ray diffraction
from a representative sample of the material. The crystal and a small amount of
the oil were collected on a MīTiGen cryoloop and transferred to the instrument
where it was placed under a cold nitrogen stream (Oxford) maintained at 100 K
throughout the duration of the experiment. The sample was optically centered
with the aid of a video camera to ensure that no translations were observed as the
crystal was rotated through all positions.
A unit cell collection was then carried out. After it was determined that the unit
cell was not present in the CCDC database a data collection strategy was calculated
by CrysAlisP ro .143 The crystal was measured for size, morphology, and color. These
values are reported in Table 4.5.
After data collection, the unit cell was re-determined using a subset of the full
data collection. Intensity data were corrected for Lorentz, polarization, and background effects using the CrysAlisP ro .143 A numerical absorption correction was
applied based on a Gaussian integration over a multifaceted crystal and followed
by a semi-empirical correction for adsorption applied using the program SCALE3
ABSPACK.144 The SHELXL-2014,109 series of programs was used for the solution
and refinement of the crystal structure. Hydrogen atoms bound to carbon atoms
were located in the difference Fourier map and were geometrically constrained
using the appropriate AFIX commands.
100
Texas Tech University, Miranda C. Andrews, December 2019
Table 4.5. Crystal structure data for 12·4,4‘-bipyridine.
Compound
Crystal color
Crystal habit
Empirical formula
Formula Weight
Temperature
Wavelength
Crystal system
Space group
Unit cell dimensions
Volume
Z
Calculated density
Absorption coefficient
F(000)
Crystal size
Theta range for data collection
Limiting indices
Reflections collected/unique
Completeness to theta = 67.679◦
Refinement method
Data/restraints/parameters
Goodness-of-fit on F2
Final R indices [I>2σ(I)]
R indices (all data)
Largest diff. peak and hole
12·4,4’-bipyridine
Dark red
Prism
C24 H21 BF2 I2 N3
654.05 g·mol−1
100(2) K
1.54178 Å
Monoclinic
C2/c
a = 16.5422(2) Å, α = 90◦
b = 15.96770(10) Å, β = 97.4530(10)◦
c = 18.2024(2) Å, γ = 90◦
4767.38(8) Å3
8
1.823 g/cm3
20.997 mm−1
2520
0.283 × 0.139 × 0.062 mm
3.863 to 77.218◦
-19 ≤ h ≤ 20, -19 ≤ k ≤ 16, -23 ≤ l ≤ 22
175688/4885 [R(int) = 0.0530]
99.7%
Full-matrix least-squares on F2
4885/0/293
1.004
R1 = 0.0459,wR2 = 0.1233
R1 = 0.0470, wR2 = 0.1244
1.522 and -1.537 e·Å−3
101
Texas Tech University, Miranda C. Andrews, December 2019
Data were collected on a Rigaku XtaLAB Synergy-i Kappa diffractometer equipped
with a PhotonJet-i X-ray source operated at 50 W (50 kV, 1 mA) to generate Cu
Kα radiation (λ = 1.54178 Å) and a HyPix-6000HE HPC detector. Crystals were
transferred from the vial and placed on a glass slide in polyisobutylene. A Zeiss
Stemi 305 microscope was used to identify a suitable specimen for X-ray diffraction
from a representative sample of the material. The crystal and a small amount of
the oil were collected on a MīTiGen cryoloop and transferred to the instrument
where it was placed under a cold nitrogen stream (Oxford) maintained at 100 K
throughout the duration of the experiment. The sample was optically centered
with the aid of a video camera to ensure that no translations were observed as the
crystal was rotated through all positions.
A unit cell collection was then carried out. After it was determined that the unit
cell was not present in the CCDC database a data collection strategy was calculated
by CrysAlisP ro .143 The crystal was measured for size, morphology, and color. These
values are reported in Table 4.6.
After data collection, the unit cell was re-determined using a subset of the full
data collection. Intensity data were corrected for Lorentz, polarization, and background effects using the CrysAlisP ro .143 A numerical absorption correction was
applied based on a Gaussian integration over a multifaceted crystal and followed
by a semi-empirical correction for adsorption applied using the program SCALE3
ABSPACK.144 The SHELXL-2014,109 series of programs was used for the solution
and refinement of the crystal structure. After the main atom positions were determined and refined, it was found that there was slight positional disorder about
the main molecule. This resulted in slight elongation of the ADPs for the two iodine atom positions. To help model this disorder, the positions were split (A and B)
and were allowed to free refine their occupancies (0.53 and 0.47, respectively) to
a total value of one. Hydrogen atoms bound to carbon atoms were located in the
difference Fourier map where possible and were geometrically constrained using
the appropriate AFIX commands. During the final refinements, the inconsistent
reflection 4 6 18 was omitted.
102
Texas Tech University, Miranda C. Andrews, December 2019
Table 4.6. Crystal structure data for 14·4,4’-bipyridine.
Compound
Crystal color
Crystal habit
Empirical formula
Formula Weight
Temperature
Wavelength
Crystal system
Space group
Unit cell dimensions
Volume
Z
Calculated density
Absorption coefficient
F(000)
Crystal size
Theta range for data collection
Limiting indices
Reflections collected/unique
Completeness to theta = 67.679◦
Refinement method
Data/restraints/parameters
Goodness-of-fit on F2
Final R indices [I>2σ(I)]
R indices (all data)
Largest diff. peak and hole
14·4,4’-bipyridine
Orange-red
Plate
C58 H48 B2 I4 N8 O
1478.26 g·mol−1
100(2) K
1.54178 Å
Monoclinic
C2/c
a = 32.3264(6) Å, α = 90◦
b = 10.2717(2) Å, β = 106.366(2)◦
c = 17.5709(3) Å, γ = 90◦
5597.96(19) Å3
4
1.754 g/cm3
17.993mm−1
2872
0.231 × 0.168 × 0.0.010 mm
2.849 to 77.549◦
-40 ≤ h ≤ 39, -12 ≤ k ≤ 11, -21 ≤ l ≤ 22
42858/5812 [R(int) = 0.0497]
99.8%
Full-matrix least-squares on F2
5812/0/370
1.086
R1 = 0.0401, wR2 = 0.1144
R1 = 0.0458, wR2 = 0.1191
0.888 and -0.791 e·Å−3
103
Texas Tech University, Miranda C. Andrews, December 2019
4.3.4
Attempted Synthesis of 13
Figure 4.22. Synthetic approach to of 13
The attempted synthesis of 13 was adapted from literature procedures.145, 146 A
solution of pyrrole (15 g, 0.22 mol) in dry diethyl ether (35 mL) was added dropwise to a stirred solution of thiophosgene (13 g, 0.11 mol) in dry diethyl ether (35
mL) at 0 ◦ C under nitrogen. After 30 minutes, methanol (25 mL) was added and
stirring was continued for a further 30 minutes. The solvent was removed under
reduced pressure and the residue dissolved in chloroform before filtering through
a short plug of silica. The solvent was removed under reduced pressure to give a
purple solid in 45% yield. 1 H NMR (400 MHz, CDCl3 ) δ: 9.76 (2H, s), 7.19 ppm
(2H, m), 7.04 ppm (2H, m), and 6.40 ppm (2H, m).
Hydrogen peroxide (35 mL, 30%) was added dropwise to a solution of potassium
hydroxide (10 g, 0.18 mol) and 2,2-dipyrrylthione (described above) (5.0 g, 2.5
mmol) in 95% aqueous methanol (200 mL) at 0 ◦ C. The mixture was refluxed for
5 minutes then cooled. Water (300 mL) was added and the solution chilled. The
product was filtered and dried to give the di(1H-pyrrol-2-yl)methanone in 20%
yield as pale yellow needles. 1 H NMR (400 MHz, CDCl3 ) δ: 10.09 ppm (2H, s), 7.16
ppm (2H, m), 7.08 ppm (2H, m), and 6.34 (2H, m).
Di(1H-pyrrol-2-yl)methanone (1.60 g, 10.0 mmol) was dissolved in 1,2-dichloro
ethane (50.0 mL). Phosphorus oxychloride (1.8 mL, 20 mmol) was added, and the
reaction mixture was heated to reflux for 3 hours, then cooled in an ice bath. Triethylamine (14.0 mL, 100 mmol) was added, and the reaction was stirred at 0 ◦ C
for 5 minutes. Boron trifluoride etherate (14.0 mL, 110 mmol) was added dropwise
104
Texas Tech University, Miranda C. Andrews, December 2019
while maintaining the temperature at 0 ◦ C. The reaction mixture was allowed to
warm to room temperature and stirred for an additional 2 hours. The resulting
solution was poured out in diethyl ether (300 ml) and extracted with water. After
drying over magnesium sulfate, filtration, and evaporation, the crude product was
purified by filtration over a silica pad (silica, DCM/petroleum ether; 1:1, v/v)
to give a red solid, 8-chloro-BODIPY, in 40%yield (1.02 g). 1 H NMR (400 MHz,
CDCl3 ) δ: 7.89 ppm (2H, s), 7.41 ppm (2H, s), 6.58 ppm (2H, s).
8-chloro-BODIPY (50.0 mg, 0.221 mmol) was dissolved in acetone (2.5 mL), and
sodium iodide (135 mg, 0.885 mmol) was added. The reaction mixture was purged
with nitrogen and heated to reflux for 15 min, showing only one compound by
TLC. The crude mixture was poured in diethyl ether, washed with water, and the
organic layer was evaporated to dryness. The 1 H NMR spectrum of the isolated
product did not match the literature spectrum.146 It did, however match the spectrum of 8-chloroBODIPY.
105
Texas Tech University, Miranda C. Andrews, December 2019
Figure 4.23. 1 H NMR spectrum of 8-chloroBODIPY.
106
Texas Tech University, Miranda C. Andrews, December 2019
CHAPTER 5
TOWARDS SWITCHABLE PNICTOGEN BONDING
Similar to heavy halogens, heavy pnictogens (group 15 elements) can be involved
in intermolecular interactions. Like iodine, if polarizable atoms such as antimony
or bismuth are bound to electron-withdrawing groups this leads to regions of partial positive charge (σ-holes) directly opposite from those groups. These regions
of electrophilicity can interact with negative or partially negative regions on other
species. This is illustrated in Fig. 5.1. Due to the ability for these atoms to form
multiple covalent (primary) bonds, these pnictogens can have multiple σ-holes
along the extension of those covalent bonds. This can, therefore, lead to multiple
secondary binding sites on the same atom that can further interact and form supramolecular structures.
Figure 5.1. Graphic representation of strategy for achieving triple pnictogen bond
donor that prevents self-recognition.
This chapter will focus on the author‘s contribution to solution binding behavior of
pnictogen bond donors. The synthesis and characterization of the pnictogen bond
donors and the elucidation of the solid state supramolecular structures they form
were performed by Shiva Moaven.
107
Texas Tech University, Miranda C. Andrews, December 2019
5.1
Pnictogen Bonding with Neutral and Charged Pnictogen Bond Acceptors
5.1.1
Antimony(III) Pnictogen Bond Donor Ability with Neutral and Charged
Pnictogen Bond Acceptors
The ability to form multiple secondary bonding interactions could lead to more
complex and interesting supramolecular assemblies than those formed by more
limited interactions, such as hydrogen and halogen bonding. In the Cozzolino
group, multiple two- and three-dimensional supramolecular assemblies and ion
receptors have been developed based on antimony(III) and bismuth(III) pnictogen
bond donors.147–150 One such example is shown in Fig. 5.2, where the antimony(III)
is bound to a rigid, tridentate indole-based ligand. The primary nitrogen-antimony
bonds are sufficiently polar to give rise to three electrophilic regions. This structure also takes advantage of the aromatic indole, which means that the lone pair
on the nitrogens is delocalized and is therefore not available to participate in intermolecular pnictogen bonding interactions.151
Figure 5.2. Antimony(III) pnictogen bond donor, 15.
Figure 5.3. Electrostatic potential of 15 mapped on the electron density surface
(0.001 a.u. isosurface). The red and black spheres indicated on the ESP are Vmax
and Vmin positions, respectively. Color map scale shown is in kJ·mol−1 .
108
Texas Tech University, Miranda C. Andrews, December 2019
Several crystal structures were grown that show the ability of 15 to form multiple secondary bonding interactions with several different pyridine-type pnictogen bond acceptors. In each case, however, only two interactions occur. Even
when the pnictogen bond acceptor is tripodal and chosen with the idea of a highly
organized heteromolecular assembly in mind (Fig. 5.4), only two of the pyridyl
groups interact with the pnictogen. DFT calculations indicate that is likely the result of negative cooperativity; the binding of one pnictogen bond acceptor reduces
the electrophilicity of the remaining two σ-holes as depicted in Fig. 5.5.151
Figure 5.4.
Ball and stick representation of crystal structure of 15·tris(2pyridinylmethyl)amine (hydrogens were omitted for clarity).
Figure 5.5. Electrostatic potential maps of 15 with (a) one pyridine and (b) two
pyridines bound. Electrostatic potential mapped on the electron density surface
(0.001 a.u. isosurface). The red and black spheres indicated on the ESP are Vmax
and Vmin positions, respectively. Color map scale shown is in kJ·mol−1 .
109
Texas Tech University, Miranda C. Andrews, December 2019
Solution state binding studies were also employed to further quantify the the pnictogen bond donor ability of 15. Binding constants were measured by fitting
the change in the UV-vis spectra accompanying titrations with pyridine and 4,4’bipyridine. Solutions of 15 in a non-coordinating solvent (dichloromethane) were
titrated with each pnictogen bond acceptor. Self-assembly was assumed to occur
stepwise, and the data extracted were first fit to a 1:1 binding model.
Figure 5.6. Titration data for 15 with pyridine fitted to a 1:1 binding model, monitored at 276 nm with residuals (left, solid line represents model) and speciation
diagram (right, solid line is 15, dotted line is 1:1 complex).
Titration data for solution state binding of 15 with pyridine fit very well with the
1:1 model as seen in Fig. 5.6. The experimental data agree very well with the
model, and the stochastic distribution of the residuals also indicates that the 1:1
binding model fits well. The speciation diagram shows indicates relatively strong
binding, with a binding constant of 1.75×104 M−1 . For comparison, binding constants of neutral halogen bond donors have been measured from 101 M−1 to 106
M−1 .152, 153
Titration data for the solution state binding of 15 with 4,4’-bipyridine proved to be
less straightforward to fit and interpret. Initially, the 1:1 binding model was tried
110
Texas Tech University, Miranda C. Andrews, December 2019
(Fig. 5.7). When the residuals indicated that this model may not be good (as seen
by the sinusoidal wave in the residuals of Fig. 5.7), composite models were fit to
the data. None of the composite models improved the fit. The only composite
model that gave a similar (albeit, slightly worse) fit was the 2:1 and 1:1 composite
model (Fig. 5.30). Because 4,4’-bipyridine is a ditopic ligand, the possibility exists
that early in the titration, there could be a supramolecular structure that consists
of two molecules of 15 and one bridging 4,4’-bipyridine.
Figure 5.7. Titration data for 15 with 4,4’-bipyridine fitted to a 1:1 binding model,
monitored at 296 nm with residuals (left, solid line represents model) and speciation diagram (right, solid line is 15, dotted line is 1:1 complex).
Because both models gave similar results, the simpler model was chosen. The negative cooperativity of the binding of pnictogen bond acceptors observed in crystal
structures is consistent with the 1:1 complex as the predominant species in solution. The 4,4’-bipyridine had a very similar binding constant to that of pyridine,
measured to be 1.22×104 M−1 .
Titrations were also performed with anionic pnictogen bond acceptors. For these
titrations, the tetra-N-butylammmonium salts of the halides were used. Compound 15 was titrated with chloride, bromide, and iodide, but the bromide titra111
Texas Tech University, Miranda C. Andrews, December 2019
Figure 5.8. Titration data for 15 with chloride fitted to a 1:1 binding model, monitored at 254 nm with residuals (left, solid line represents model) and speciation
diagram (right, solid line is 15, dotted line is 1:1 complex).
tions were not reproducible. Determined binding constants for bromide varied
from 102 to 106 M−1 . Chloride and iodide were more consistent from run to run,
but the data points did not lead to satisfactory fits. The best models are shown in
Figs. 5.8 and 5.9. For chloride, a 1:1 binding model seemed to be sufficient; composite models did not result in a better fit. The binding constant was determined
to be 5.71×105 M−1 . The fit for iodide was improved slightly by using a 1:1 and 1:2
(1 pnictogen bond donor to 2 iodide anions) composite model. The binding constant for the first binding event was calculated to be 1.26×105 M−1 , and the second
binding event was calculated to have a binding constant of 8.96×104 M−1 . This
is consistent with the negative cooperativity indicated by both crystal structures
and calculations. The binding constants are larger than those measured for neutral
pnictogen bond acceptors, which is to be expected. The full negative charge on the
anionic species will facilitate tighter binding than any partial negative charges in
neutral species due to the electrostatic nature of the interactions.
112
Texas Tech University, Miranda C. Andrews, December 2019
Figure 5.9. Titration data for 15 with iodide fitted to a 1:1 and 1:2 composite binding
model, monitored at 266 nm with residuals (left, solid line represents model) and
speciation diagram (right, solid line is 15, dotted line is 1:1 complex, hashed line is
the 1:2 complex).
113
Texas Tech University, Miranda C. Andrews, December 2019
5.1.2
Bismuth(III) Pnictogen Bond Donor Ability with Neutral and Charged
Pnictogen Bond Acceptors
Figure 5.10. Bismuth(III) pnictogen bond donor, 16.
A similar pnictogen bond donor with bismusth(III) has also been developed and
should be a stronger pnictogen bond donor. As the size (and therefore polarizability) of the atom increases, so does its ability to form secondary bonding interactions.154, 155 Compound 16 shows a better ability to bind with the neutral pnictogen
bond acceptor 4,4’-bipyridine, binding at two sites more definitively than 15. The
binding constants were measured to be 2.72×104 M−1 for the first binding event
and 4.29×103 M−1 for the second binding event. This is consistent with the idea
that bismuth(III) should be a better pnictogen bond donor and also with the idea
of negative cooperativity seen in the antimony(III) analogue.
As expected, anionic pnictogen bond acceptors bind more tightly than the neutral
pnictogen bond acceptor. Two chlorides bind with binding constants of 3.24×106
M−1 for the first binding event and 7.69×105 M−1 . These constants are quite high
for pnictogen bonds, but the better pnictogen bond donor and higher charge density of the pnictogen bond acceptor may contribute to this. Bromides bind with
constants of 8.48×105 M−1 and 1.19×105 M−1 . Lastly, two iodides bind with constants measured at 8.20×105 M−1 and 3.60×104 M−1 . The trend of decreasing binding constants is likely due the increasing size and decreasing charge density of the
pnictogen bond acceptors, although, it should still be noted, that these binding
constants are high for secondary bonding interactions.152, 153 This could likely have
some important consequences for further pnictogen-bonded system design.
114
Texas Tech University, Miranda C. Andrews, December 2019
Figure 5.11. Titration data for 16 with 4,4’-bipyridine fitted to a 1:1 and 1:2 composite binding model, monitored at 262 nm with residuals (left, solid line represents
model) and speciation diagram (right, solid line is 16, dotted line is 1:1 complex,
hashed line is the 1:2 complex).
Figure 5.12. Titration data for 16 with chloride fitted to a 1:1 and 1:2 composite binding model, monitored at 254 nm with residuals (left, solid line represents
model) and speciation diagram (right, solid line is 16, dotted line is 1:1 complex,
hashed line is the 1:2 complex).
115
Texas Tech University, Miranda C. Andrews, December 2019
Figure 5.13. Titration data for 16 with bromide fitted to a 1:1 and 1:2 composite binding model, monitored at 262 nm with residuals (left, solid line represents
model) and speciation diagram (right, solid line is 16, dotted line is 1:1 complex,
hashed line is the 1:2 complex).
Figure 5.14. Titration data for 16 with iodide fitted to a 1:1 and 1:2 composite binding model, monitored at 262 nm with residuals (left, solid line represents model)
and speciation diagram (right, solid line is 16, dotted line is 1:1 complex, hashed
line is the 1:2 complex).
116
Texas Tech University, Miranda C. Andrews, December 2019
5.2
Pnicotgen Catecholates
Other antimony complexes have been synthesized (Fig. 5.15), and solution-state
binding studies of these complexes have been performed. Previous studies show
that antimony catecholates show measurable interactions with halides and other
anionic species.147, 148 These compounds prove to be more robust towards hydrolysis than 15 and 16, although similar precautions in handling were taken to avoid
prolonged exposure to air and moisture.
Figure 5.15. Antimony(III) catecholate, 17
Like compounds 15 and 16, compound 17 shows promise as pnictogen bond donor.
Titrations with bromide and iodide show good binding constants, although better
models may be needed for these anions. Bromide binds with measured constants
of 8.83×105 M−1 for the first binding event and 1.20×105 M−1 for the second. The
first binding event with iodide showed a smaller binding constant (3.18×105 M−1 )
than that of bromide. However, the second binding event of iodide was measured
to be stronger than that of bromide with a binding constant of 1.54×105 M−1 . Both
anions show behavior consistent with negative cooperativity, similar to 15 and 16.
Titrations of 17 with chloride appeared to be outliers, giving binding constants of
5.90×109 M−1 for the first binding event and 2.30×107 M−1 . While these measured
binding constants would be very exciting, it is unlikely that chloride would have
a binding constant four orders of magnitude larger than bromide and iodide. This
experiment will need to be redone to confirm or disprove this first set of data.
The apparent significant increase in binding constants for chloride compared to
bromide and iodide could also be attributed to the fact that the chlorine bound to
117
Texas Tech University, Miranda C. Andrews, December 2019
Figure 5.16. Titration data for 17 with bromide fitted to a 1:1 and 1:2 composite binding model, monitored at 237 nm with residuals (left, solid line represents
model) and speciation diagram (right, solid line is 17, dotted line is 1:1 complex,
hashed line is the 1:2 complex).
the antimony(III) in 17 is quite labile.156 There would be no competition for forming a primary bond between the antimony(III) and the chloride when titrating with
a chloride solution. However, in the cases of bromide and iodide titrations, there
would be competition for forming the primary bond with antimony(III) between
the chloride and the heavier halide (Fig. 5.18). This competition could also help to
explain why the models tested were not particularly good fits; there are additional
interactions that might need to be accounted for.
118
Texas Tech University, Miranda C. Andrews, December 2019
Figure 5.17. Titration data for 17 with iodide fitted to a 1:1 and 1:2 composite binding model, monitored at 233 nm with residuals (left, solid line represents model)
and speciation diagram (right, solid line is 17, dotted line is 1:1 complex, hashed
line is the 1:2 complex).
Figure 5.18. Mechanism for anion binding and subsequent substitution of stibaindole.156
119
Texas Tech University, Miranda C. Andrews, December 2019
5.2.1
Photoswitchable Catechol
A spiropyran containing a catechol moiety (18, Fig. 5.19) has been synthesized
using an adapted literature procedure.70 One can imagine using this and related
catecholates to modulate the pnictogen bonding described previously. In an effort to begin studying this potential application, DFT calculations were performed
on antimony complexes containing photoswitchable catecholates. The spiropyran
catecholate compounds modelled were similar to simpler catecholates for which
binding constants have been determined experimentally by UV-vis spectroscopy
(vide supra).
Figure 5.19. Spiropyran catechol, 18.
5.3
Toward Photoswitchable Pnictogen Catecholates
Based on the previously studied behavior of spiropyrans, it was hypothesized
that the use of photoswitchable ligand bound to a heavy pnictogen could result in a measurable change in pnictogen bonding ability. The ring-opening of
a spiropyran results in a change in the distribution of electron density, and the
electron-withdrawing and -donating ability of moieties within the spiropyran. By
introducing a group that interacts with pnictogens, such as catecholates, this modulation of electronic properties could be used to modulate pnictogen binding ability. Calculations were performed on antimony(III) catecholates that contain spiropyran moieties to determine if and to what extent pnictogen binding could be modulated by a photoswitch.
Three different antimony(III) spiropyran-catecholates were modelled. Because the
catecholates are dianionic, another group is needed to complete the valence of
120
Texas Tech University, Miranda C. Andrews, December 2019
the antimony. Both phenyl-antimony and chloro-antimony catecholates have been
synthesized previously, so both cases were modelled as well to elucidate which
compound would exhibit the most change in the resulting σ-holes upon photoswitching.
Figure 5.20. Compounds (a) 19 and (b) 20.
The first compound modelled was based on the spiropyran-catechol (18). Calculations performed on 19 and 20 suggest that the change in electron density within the
spiropyran moiety upon ring-opening results in deactivation of a majority of the
σ-holes as seen in Figs. 5.21 and 5.22. This is not entirely unexpected as the phenolate moiety of the spiropyran becomes more electron-rich upon ring-opening. This
should result in the phenolate moiety being more electron-donating and therefore deactivates the antimony with regard to pnictogen bonding. Essentially, 19
demonstrates “turn-off” functionality with regard to pnictogen bonding. There is
one exception, however. In the case of 19, there does appear to be a slight increase
in one of the the Vmax values of 5 kJ·mol−1 upon photoswitching of the spiropyan
(from 156 to 161 kJ·mol−1 ). It is important to note that the optimized structures
as shown in Fig. 5.21 involve a more pronounced change in the orientation of the
of 19(o) compared to 19(c). This change in orientation means that as shown in
Fig. 5.21, the Vmax of 156 kJ·mol−1 in Fig. 5.21(a) corresponds to the Vmax of 161
121
Texas Tech University, Miranda C. Andrews, December 2019
kJ·mol−1 in Fig. 5.21(b), and the Vmax of 147 kJ·mol−1 in Fig. 5.21(a) corresponds to
the Vmax of 115 kJ·mol−1 in Fig. 5.21(b).
Figure 5.21. Electrostatic potential map of 19 in (a) closed and (b) open states.
Color map scale shown is in kJ·mol−1 . Isosurface value is 0.001 a.u.
Figure 5.22. Electrostatic potential map of 20 in (a) closed and (b) open states.
Color map scale shown is in kJ·mol−1 . Isosurface value is 0.001 a.u.
122
Texas Tech University, Miranda C. Andrews, December 2019
Because the indoline moiety of the spiropyran loses electron density upon ringopening, it was hypothesized that by incorporating the catecholate on the indoline,
the changes observed in the Vmax values would lead to “turn-on” functionality
with regard to pnictogen bonding. In order to maximize this effect, two different
spiropyrans (Figs. 5.23 and 5.24) were designed in which one of the catecholate
oxygens is para- to the indoline nitrogen, which can be drawn with a formal charge
of +1 when in the open state.
Figure 5.23. Compounds (a) 21 and (b) 22.
Figure 5.24. Compounds (a) 23 and (b) 24.
123
Texas Tech University, Miranda C. Andrews, December 2019
Figure 5.25. Electrostatic potential map of 21 in (a) closed and (b) open states.
Color map scale shown is in kJ·mol−1 . Isosurface value is 0.001 a.u.
Figure 5.26. Electrostatic potential map of 22 in (a) closed and (b) open states.
Color map scale shown is in kJ·mol−1 . Isosurface value is 0.001 a.u.
124
Texas Tech University, Miranda C. Andrews, December 2019
Figure 5.27. Electrostatic potential map of 23 in (a) closed and (b) open states.
Color map scale shown is in kJ·mol−1 . Isosurface value is 0.001 a.u.
Figure 5.28. Electrostatic potential map of 24 in (a) closed and (b) open states.
Color map scale shown is in kJ·mol−1 . Isosurface value is 0.001 a.u.
125
Texas Tech University, Miranda C. Andrews, December 2019
In the cases of compounds 21, 22, 23, and 24, each Vmax is increased by roughly 40
kJ·mol−1 upon photoisomerization of the spiropyran moiety. For some context, the
Vmax of iodobenzene is calculated to be 63 kJ·mol−1 , and the Vmax of perfluoroiodobenzene is calculated to be 125 kJ·mol−1 at the same level of theory.
The computational and solution state binding studies performed on 15 can serve
as a reference to estimate the solution state binding for the structures that have so
far only been studied computationally. Compound 15 had Vmax values calculated
to be 147 kJ·mol−1 and binding constants measured to be on the order of 104 to 105
M−1 . Using these data as a guide, the structures studied computationally should
have binding constants on the order of 105 M−1 or higher, as many of these structures have Vmax values calculated to be 81 to 223 kJ·mol−1 . The ability to modulate
secondary bonding interactions, particularly with light, could prove to be useful
in designing supramolecular structures. One can imagine using light to turn on or
off secondary bonding interactions that result in the controlled building or disassembling of supramolcular assemblies.
5.4
Additional Titration Data
Figure 5.29. UV-vis spectra of titration of 26.4 µM 15 with solutions of (a) pyridine
(2.03 mM, 7.50 mM, and 88.6 mM) and (b) of 4,4-bipyridine (2.21 mM, 7.22 mM,
and 49.5 mM) in dichloromethane.
126
Texas Tech University, Miranda C. Andrews, December 2019
Figure 5.30. Titration data for 15 with 4,4’-bipyridine fitted to a 2:1 and 1:1 composite binding model, monitored at 296 nm with residuals (left, solid line represents
model) and speciation diagram (right, solid line is 15, dotted line is 1:1 complex,
hashed line is the 2:1 complex).
Figure 5.31. UV-vis spectra of titration of 25.2 µM 15 with solutions of (a) chloride
(2.19 mM, 17.4 mM, and 55.6 mM) and (b) of iodide (1.73 mM, 18.7 mM, and 56.6
mM) in dichloromethane.
127
Texas Tech University, Miranda C. Andrews, December 2019
Figure 5.32. UV-vis spectra of titration of 20.3 µM 16 with solutions of 4,4’bipyridine (2.11 mM, 18.1 mM, and 53.6 mM) in dichloromethane.
Figure 5.33. UV-vis spectra of titration of 26.6 µM 16 with solutions of (a) chloride
(2.19 mM, 17.4 mM, and 55.6 mM), (b) bromide (2.24 mM, 19.0 mM, and 54.7 mM)
and (c) of iodide (1.73 mM, 18.7 mM, and 56.6 mM) in dichloromethane.
Figure 5.34. UV-vis spectra of titration of 22.7 µM 17 with solutions of (a) chloride
(2.19 mM, 17.4 mM, and 55.6 mM), (b) bromide (2.24 mM, 19.0 mM, and 54.7 mM)
and (c) of iodide (1.73 mM, 18.7 mM, and 56.6 mM) in dichloromethane.
128
Texas Tech University, Miranda C. Andrews, December 2019
Figure 5.35. Titration data for 17 with chloride fitted to a 1:1 and 1:2 composite binding model, monitored at 234 nm with residuals (left, solid line represents
model) and speciation diagram (right, solid line is 17, dotted line is 1:1 complex,
hashed line is the 2:1 complex). Note that the scale of the x-axis of the speciation
diagram is smaller in order to make the data easier to see.
129
Texas Tech University, Miranda C. Andrews, December 2019
CHAPTER 6
GENERAL EXPERIMENTAL METHODS
6.1
6.1.1
Spectroscopy
IR Spectroscopy
Typical IR spectra were collected using a Nicolet iS5 FTIR spectrometer equipped
with a diamond ATR accessory. Spectra were acquired between 4000 and 400 cm−1
with a resolution of 4 cm−1 .
6.1.2
NMR Spectroscopy
Typical NMR samples were prepared by dissolving 10-15 mg of a given compound
in 0.5-0.7 mL of deuterated solvent, ensuring a homogeneous soltution. NMR spectra were acquired on a JEOL ECS 400 MHz spectrometer. Typical aquisition parameters were 8 scans with a delay time of 5 s for 1 H NMR and 1000 scans with a
delay time of 2 s for 13 C NMR. Spectra were referenced using the residual proton signal or the 13 C signal of the deuterated solvent and setting it to a value that
corresponds to a Me4 Si signal of 0 ppm.
6.1.3
Solution Ultraviolet-visible Spectroscopy
UV-vis spectra discussed in Chapter 2 were acquired using a StellarNet SilverNova Super Range TEC spectrometer, with SL1 Tungsten Halogen Lamp (visible
and near-IR region) and SL3 Deuterium Lamp (UV region) as the light source. All
other UV-vis spectra discussed in this dissertation were acquired using a Mikropac
DH-2000 UV-vis-NIR source, with a halogen lamp (visible and near-IR region) and
a deuterium lamp (UV region). When not working in an inert atmosphere glovebox, a CUV-TEMP cuvette holder (qpod 2e) with a path length of 1.0 cm was used.
This sample holder allowed for temperature control and consistent stirring, typic-
130
Texas Tech University, Miranda C. Andrews, December 2019
ally at 1200 rpm. Spectra were acquired at 20 ◦ C, except where otherwise specified.
Spectra were recorded in a window from 185-1100 nm, but spectra are reported in
an appropriate window, showing only regions where absorbance is meaningful.
In all cases the pathlength of the cell was 1.0 cm. Quartz cuvettes were used in
all cases. For instances in which the samples were irradiated in situ, fluorescence
cuvettes were used, otherwise cuvettes with two frosted windows were used. Before each experiment, a dark spectrum and a blank spectrum were recorded. For
every spectrum obtained, the maximum integration time that did not saturate the
detector was used. This was determined at the beginning of each experiment to
ensure that any changes in experimental setup did not result in saturation. A good
signal to noise ratio was established by averaging multiple spectra together.
6.2
6.2.1
Solution-state Binding Studies
Experimental Data Collection
Changes observed by UV-vis spectroscopy were used to determine if and to what
extent secondary bonding interactions occur in solution. Before each data collection trial, dark and blank (of pure, dry solvent) spectra were collected. Under inert,
dry atmosphere, 2.00 mL of the analyte solution (approximate concentrations of 20
µM) was loaded via syringe into the cuvette that was charged with a stir bar and
capped with a custom Teflon cap, with a small hole drilled through it, just large
enough for the needle of the syringe (vide infra). The titrant solution was loaded
into a 25.0 µL Hamilton 80200 gastight syringe (with cemented needle). The syringe was used along with a Hamilton PB600-1 repeating dispenser. This assembly
dispenses the volume of the syringe in 50 identical increments; under these conditions, each press of the button dispenses 0.500 µL of the titrant solution. Solutions
of varying concentrations of titrant solutions were made such that three consecutive receptor solutions could be added per titration. This enabled control over the
number of data points, particularly fewer equivalents of titrant. Titrant solutions
were used in succession, starting with the least concentrated and working up to
the most concentrated to minimize the effect of contamination of the syringe. The
131
Texas Tech University, Miranda C. Andrews, December 2019
syringe was rinsed with clean DCM between segments and runs. The possibility of a competing monomor/dimer equilbrium was ruled out by performed by
monitoring the molar absorbtivity as a function of concentration (i.e., a dilution
experiment). The influence of the ionic strength of the solution was evaluated by
titrating with a non-coordinating anion, PF6 − .
6.2.2
Data Reduction
In order to ensure that any changes in absorption monitored during the titrations
were not due to any changes in the baseline, the raw data was baseline corrected by
applying an offset to the entire spectrum. This was done by subtracting the average
intensity of a region of 100 nm that remained flat throughout the duration of the
experiment from each spectrum. For example, if the region between 800 and 900
nm remained unperturbed throughout the titration, the average absorption over
that range for a given spectrum was subtracted from (only) that spectrum. Absorption data for a particular wavelength that maximized the observable change
without exceeding an absorbance of 1 au was extracted from the titration data
and plotted as a function of analyte concentration. The guest (titrant) concentration was determined by multiplying the addition rate (based on the autodispenser,
vide supra) by the time at which each spectrum was collected. The host (titrand)
concentration was determined at each point by considering the effects of dilution.
6.2.3
Data Fitting
The intensity data from the wavelength that has the largest change over the course
of the titration (without exceeding an absorbance of 1) was fitted against modeled
data. The models that were used are described below. For each set of data, the
1:1 binding model was first applied. In the cases where ion pairing should be accounted for, the appropriate binding constant of the titrant was included in the
model. One hundred possible solutions were obtained from randomly chosen
starting points for each parameter. In order to avoid biasing the model, nearly
all the parameters for the starting points were given large ranges. The exception to
132
Texas Tech University, Miranda C. Andrews, December 2019
this was the molar absorptivity of the titrand for the wavelength being monitored,
which was constrained to the absorbance divided by concentration of the titrand
in mol·L−1 and the pathlength (1 cm). The residuals were analyzed to ensure that
the model was correct as verified by a stochastic distribution.
In some cases, the residuals indicated a model mismatch. In these cases, the data
was then fitted to a composite binding model that included equilibria for 1:1 binding and 1:2 binding. In most cases, this model showed to be a good fit and no
further models were required as the residuals for this model had a stochastic distribution, indicating a strong match for this model. In any cases where these models
were insufficient, the particulars are described in the discussion of these experiments. Errors were determined by bootstrapping. The data was resampled 100
times for each trial with a sample size equal to the number of data points collected for each trial. The trials were then averaged together and the maximum error
(from bootstrap or trial averaging) was reported.
For all of the models described below, H = host (halogen or pnictogen bond donor),
G = guest (halogen or pnictogen bond acceptor), K = binding constant of interest
(M−1 , a special case of an equilibrium constant), KP = the ion pairing constant (M),
and A = the Lewis acid of the ion pair (tetra-n-butylammonium for the relevant
titrations discussed herein).
133
6.2.4
1 to 1 Binding Model
H +G
[HG]
[H][G]
(6.1)
134
[H0 ] = [H] + [HG], [HG] = [H0 ] − [H]
(6.2)
[G0 ] = [G] + [HG], [G] = [G0 ] − [HG]
(6.3)
Solve for [G], [H], or [HG] with Maple 2016 (outputs indicated by blue text) using equations 6.1-6.3.
Texas Tech University, Miranda C. Andrews, December 2019
K=
HG
(6.4)
eliminate({K · H · G − HG, H + HG − H0, G + HG − G0}, {HG, H});
2
{H = G − G0 + H0, HG = −G + G0}, {G K − G G0 K + G H0 K + G − G0}
(6.5)
135
eliminate({K · H · G − HG, H + HG − H0, G + HG − G0}, {H, G});
(6.6)
2
{G = −HG + G0, H = −HG + H0}, {G0 H0 K − G0 H0 K − H0 HG K + HG K − HG}
From equations 6.4-6.6:
1
[H] + [G0 ] − [H0 ] +
K
2
[H] −
[H0 ]
=0
K
(6.7)
Texas Tech University, Miranda C. Andrews, December 2019
eliminate({K · H · G − HG, H + HG − H0, G + HG − G0}, {HG, G});
2
{G = H − H0 + G0, HG = −H + H0}, {G0 H K + H K − H H0 K + H − H0}
[H] =
q
2
− [G0 ] − [H0 ] + K1 ±
[G0 ] − [H0 ] + K1 + 4 [HK0 ]
2
(6.8)
[H0 ] known experimental quantity (original concentration of host)
[G0 ] known experimental quantity (original concentration of guest)
136
6.2.4.1
Adaptation to UV-vis Spectroscopy (pathlength = 1 cm
Absorbancecalc = H [H] + HG [HG] + G [G]
H is the molar absorptivity of the host, a known experimental quantity
G is the molar absorptivity of the guest, a known experimental quantity
HG is the molar absorptivity of the complex, a refined parameter
(6.9)
Texas Tech University, Miranda C. Andrews, December 2019
K 1:1 binding constant, a refined parameter
6.2.5
1 to 1 Binding Model with Ion Pairing
AG
A+G
HG
137
K=
[HG]
[H][G]
(6.10)
Kp =
[A][G]
[AG]
(6.11)
[H0 ] = [H] + [HG], [HG] = [H0 ] − [H]
(6.12)
Texas Tech University, Miranda C. Andrews, December 2019
H +G
[AG0 ] = [G] + [HG] + [AG], [G] = [AG0 ] − [HG] − [AG]
(6.13)
[A] = [G] + [HG]
(6.14)
138
eliminate({K · H · G − HG, Kp · AG − A · G, H + HG − H0, G + HG − AGO + AG, A − G − HG},
{HG, G, AG, A});
AG0 H K + H 2 K − H H0 K + H − H0
H 2 K − H H0 K + H − H0
, AG =
,
A=−
KH
KH
H − H0
G=−
, HG = −H + H0 , {AG0 H 2 K 2 Kp + H 3 K 2 Kp − H 2 H0K 2 Kp − H 3 K+
KH
2
2
2
2
2
2H H0 K + H K Kp − H H0 K − H H0 K Kp − H + 2H HO − H0 }
(6.15)
Texas Tech University, Miranda C. Andrews, December 2019
Solve for [G], [H], or [HG] with Maple 2016 (outputs indicated by blue text) using equations 6.10-6.13.
eliminate({K · H · G − HG, Kp · AG − A · G, H + HG − H0, G + HG − AG0 + AG, A − G −HG},
(6.16)
139
eliminate({K · H · G − HG, Kp · AG − A · G, H + HG − H0, G + HG − AGO + AG, A − G − HG},
{H, G, HG, A});
−G02 Kp − G0 H0 + 2 G0 Kp + AG + H0 − Kp
AG
A=
, G = Kp(G0 − 1), H = −
,
G0 − l
G0 − 1
−G02 Kp + 2 G0 Kp + AG − Kp
HG =
, {G03 K Kp2 + G02 HO K Kp − 3G02 K Kp2 −
G0 − 1
AG G0 K Kp − 2 G0 H0 K Kp + 3 G0 K Kp2 + AG K Kp + G02 Kp + H0 K Kp − K Kp2 −
2 G0 Kp − AG + Kp}
(6.17)
Texas Tech University, Miranda C. Andrews, December 2019
{HG, H, AG, A});
AGO G K − G2 K − G H0 K + AG0 − G
G(GK + H0 K + 1)
, AG =
,
A=
GK +1
GK +1
H0
K G H0
H=
, HG =
, {AG0 G K Kp − G3 K − G2 HO K − G2 K Kp−
GK +1
GK +1
2
G H0 K Kp + AG0 Kp − G − G Kp}
eliminate({K · H · G − HG, Kp · AG − A · G, H + HG − H0, G + HG − AG0 + AG, A − G − HG},
{H, G, AG, A});
HG(G0 H0 K − G0 HG K − H0 K + HG K + G0 − 1
HG(H0 K − HG K + 1)
, AG =
,
A=
K(−HG + H0)
K(−HG + H0)
HG
G=
, H = −HG + H0 , {(G0 H02 K 2 Kp − 2 G0 H0 HG K 2 Kp + G0 HG2 K 2 Kp−
K(−HG + H0)
(6.18)
140
simplif y(AG0 H 2 K 2 Kp + H 3 K 2 Kp − H 2 H0 K 2 Kp − H 3 K + 2 H 2 H0 K + H 2 K Kp − H HO2 K−
H H0 K Kp − H 2 + 2 H HO − H02 );
(K 2 Kp − K)H 3 + (−1 − Kp(H0 − AG0)K 2 + (2 H0 + Kp)K)H 2 − (−2 + (H0 + Kp)K)H0 H − H02
(6.19)
simplif y(AG0 G K Kp − G3 K − G2 H0 K − G2 K Kp − G H0 K Kp + AG0 Kp − G2 − G Kp);
− G3 K + (−H0 K − K Kp − 1)G2 − (1 + (H0 − AG0)K)Kp G + AG0 Kp
(6.20)
Texas Tech University, Miranda C. Andrews, December 2019
H02 K 2 Kp + 2 H0 HG K 2 Kp − HG2 K 2 Kp + G0 H0 K Kp − G0 HG K Kp − H0 HG K−
2
H0 K Kp + HG K + HG K Kp − HG)HG}
[H0 ] known experimental quantity (original concentration of host)
[AG0 ] known experimental quantity (original concentration of acid/guest complex)
K 1:1 binding constant, a refined parameter
Kp acid/guest pairing association constant
Adaptation to UV-vis Spectroscopy (pathlength = 1 cm)
141
Absorbancecalc = H [H] + HG [HG] + G [G]
H is the molar absorptivity of the host, a known experimental quantity
G is the molar absorptivity of the guest, a known experimental quantity
HG is the molar absorptivity of the complex, a refined parameter
6.2.6
1 to 1 and 1 to 2 Composite Binding Model
H +G
HG
(6.21)
Texas Tech University, Miranda C. Andrews, December 2019
6.2.5.1
HG + G
HG2
(6.22)
K12 =
[HG2 ]
[HG][G]
(6.23)
β12 =
[HG2 ]
= K11 · K12
([H][G]2 )
(6.24)
[H0 ] = [H] + [HG] + [HG2 ]
(6.25)
[G0 ] = [G] + [HG] + 2[HG2 ]
(6.26)
Texas Tech University, Miranda C. Andrews, December 2019
[HG]
[H][G]
142
K11 =
Solve for [H] with Maple 2016 using equations 6.22-6.23, 6.25-6.26.
eliminate({K11 · H · G − HG, K12 · HG · G − HGG, H + HG + HGG − H0, G + HG + 2 · HGG − G0},
(6.27)
143
4 G0 H 2 K11 K12 + 4 G0 H H0 K11 K12 + H 3 K112 − 4 H 3 K11 K12 − H 2 H0 K112 +
2
2
8 H H0 K11 K12 − 4 H H0 K11 K12 − G0 H K11 − H + H0}
From equation set 6.27:
2
2
2
(K11
−4K11 K12 )[H]3 + ([G0 ]K11
− [H0 ]K11
+ 8[H0 ]K11 K12 − 4[G0 ]K11 K12 )[H]2 +
(4[G0 ][H0 ]K11 K12 − 4[H0 ]2 K11 K12 − [G0 ]K11 − [G0 ]2 K1 1K1 2 − 1)[H] + [H0 ] = 0
(6.28)
for: a[H]3 + b[H]2 + c[H] + d
2
a = (K11
− 4K11 K12 )
(6.29)
Texas Tech University, Miranda C. Andrews, December 2019
{HG, HGG, G});
G0 + 2 H − 2 H0
K11 H(G0 + 2 H − 2 H0)
G=−
, HG = −
, HGG =
H K11 − 1
H K11 − 1
G0 H K11 + H 2 K11 − H H0 K11 + H − H0
, {−G02 H K11 K12 + G0 H 2 K112 −
H K11 − 1
(6.30)
c = (4[G0 ][H0 ]K11 K12 − 4[H0 ]2 K11 K12 − [G0 ]K11 − [G0 ]2 K11 K12 − 1)
(6.31)
d = [H0 ]
(6.32)
144
[H0 ] known experimental quantity (original concentration of host)
[G0 ] known experimental quantity (original concentration of guest)
K11 1:1 stepwise binding constant, a refined parameter
K12 1:2 stepwise binding constant, a refined parameter
Texas Tech University, Miranda C. Andrews, December 2019
2
2
b = ([G0 ]K11
− [H0 ]K11
+ 8[H0 ]K11 K12 − 4[G0 ]K11 K12 )
6.2.6.1
Adaptation to UV-vis Spectroscopy (pathlength = 1 cm)
Absorbanceobs = H [H] + HG [HG] + HG2 [HG2 ] + G [G]
(6.33)
H is the molar absorptivity of the host, a known experimental quantity
HG is the molar absorptivity of the 1:1 complex, a refined parameter
145
HG2 is the molar absorptivity of the 1:2 complex, a refined parameter
6.2.7
1 to 1 and 1 to 2 Composite Binding Model with Ion Pairing
AG
H +G
A+G
HG
Texas Tech University, Miranda C. Andrews, December 2019
G is the molar absorptivity of the guest, a known experimental quantity
HG + G
HG2
(6.34)
K12 =
[HG2 ]
[HG][G]
(6.35)
Kp =
[A][G]
[AG]
(6.36)
β12 =
[H2 G]
= K11 · K12
[H][G]2
(6.37)
[H0 ] = [H] + [HG] + [HG2 ]
(6.38)
Texas Tech University, Miranda C. Andrews, December 2019
[HG]
[H][G]
146
K11 =
[AG0 ] = [AG] + [G] + [HG] + 2[HG2 ]
(6.39)
[AG0 ] = [AG] + [A]
(6.40)
147
eliminate({AG0 − AG − A, H + HG + HGG − H0, G + HG + 2 · HGG + AG − AG0, −A · G + AG · KD,
G · H · K11 − HG, G · HG · K12 − HGG}, {A, AG, H, HG, HGG});
G AG0
H0
K11 G H10
AG0 Kp
, AG =
,H = 2
, HG = 2
,
A=
G + Kp
G + Kp
G K11 K12 + K11 G + 1
G K11 K12 + K11 G + 1
K11 G2 H0 K12
, {AG0 G2 H0 K12 KD − G4 K11 K12 − 2 K11 G3 H0 K12−
HGG = 2
G K11 K12 + K11 G + 1
G3 K11 K12 Kp − 2 K11 G2 H0 K12 Kp + AG0 G K11 KD − G3 K11 − K11 G2 H0 − G2 K11 Kp−
2
K11 G H0 Kp + AG0 Kp − G − G Kp}
(6.41)
From equation set 6.41:
Texas Tech University, Miranda C. Andrews, December 2019
Solve for [G] with Maple 2017 using equations 6.34-6.35, 6.37-6.39.
simplif y({AG0 G2 K11 K12 Kp − G4 K11 K12 − 2 K11 G3 H0 K12 − G3 K11 K12 Kp−
2 K11 G2 H0 K12 Kp + AG0 G K11 Kp − G3 K11 − K11 G2 H0 − G2 K11 Kp−
K11 G H0 Kp + AG0 Kp − G2 − G Kp}, {power});
(6.42)
{−G4 K11 K12 + (−2 H0 K12 − K12 KD − 1)K11 G3 + (−1 + ((−1 + (−2 H0+
AG0)K12)Kp − H0)K11)G2 − Kp(1 + (H0 − AG0)K11)G + AG0 Kp}
[AG0 ] known experimental quantity (original concentration of guest with cation)
148
K11 1:1 stepwise binding constant, a refined parameter
K21 2:1 stepwise binding constant, a refined parameter
Kp dissociation constant for ion pair
6.2.7.1
Adaptation to UV-vis Spectroscopy (pathlength = 1 cm)
Absorbanceobs = H [H] + HG [HG] + HG2 [HG2 ] + G [G]
(6.43)
Texas Tech University, Miranda C. Andrews, December 2019
[H0 ] known experimental quantity (original concentration of host)
H is the molar absorptivity of the host, a known experimental quantity
G is the molar absorptivity of the guest, a known experimental quantity
HG is the molar absorptivity of the 1:1 complex, a refined parameter
HG2 is the molar absorptivity of the 1:2 complex, a refined parameter
Texas Tech University, Miranda C. Andrews, December 2019
149
Texas Tech University, Miranda C. Andrews, December 2019
6.2.8
Computational Details
The calculated structures discussed in this dissertation were fully optimized using
the ORCA quantum chemistry program packages. Calculations of compounds 1
and 2 and related ions discussed in Chapter 2 were performed using the ORCA
3.0 program package.157 For geometry optimizations the LDA and GGA functionals employed were those of Perdew and Wang (PW-LDA, PW91).110 In the case of
thermodynamic calculations, the hybrid functional (30% HF)158 of Becke (B88)159
was used for exchange and Perdew and Wang (PW91)110 for the correlation. The
geometry optimizations employed the Ahlrichs-TZV160 basis sets (def2-TZV111 ),
designated as Default-Basis-3160 in the Orca 3.0 program. The Ahlrichs (2d,2p)
polarization functions were obtained from the TurboMole basis set library under
ftp.chemie.uni-karlsruhe.de/pub/basen. Frequency calculations were performed
on gas-phase optimized structures and no negative frequencies were determined.
Electronic excitations were calculated with TDDFT using the B3LYP158, 161 functional and the Ahlrichs-TZV160 basis sets (def2-TZV111 ) designated as Default-Basis4160 in the Orca 3.0 program. The Ahlrichs (2d2fg,3p2df) polarization functions
were obtained from the TurboMole basis set library under ftp.chemie.uni-karls
ruhe.de/pub/basen.
Calculations performed on all other compounds discussed throughout this dissertations were performed using the ORCA 4.0 program package.157 Calculations
discussed in Chapter 3 were first optimized using the PW91 functional110 and the
def2-SVP basis set111 (def2/J auxiliary basis162 ). For the iron(II) and cobalt(II) complexes, additional calculations were performed using the TPSSh functional118 and
def2-SVP basis set.111 These additional calculations were performed on the structures that had be previously optimized using the PW91 functional.110
The compounds discussed computationally in Chapters 4 and 5 were performed
using the PW91 functional110 and the ZORA-def2-TZVPP basis set111 (SARC/J auxiliary basis162 ). Relativistic effects were accounted for with the use of the ZeroOrder Regular Approximation (ZORA).163, 164 Molecular electrostatic potential (ESP)
mapping was performed on the 0.001 a.u. isosurface of the electron density using
150
Texas Tech University, Miranda C. Andrews, December 2019
the Multiwfn program.165, 166 All molecular orbitals discussed in this dissertation
were calculated at a 0.004 a.u. isosurface value. All visualizations of the molecular orbitals and ESP maps were performed with the Gabedit graphical interface
software.167
151
Texas Tech University, Miranda C. Andrews, December 2019
CHAPTER 7
SUMMARIES AND CONCLUSIONS
The overarching theme of this work is to determine if, how, and to what extent
molecular switches can be used to affect a measurable change in a pendant group.
This was determined for several proposed and synthesized molecular switches
by a combination of DFT modeling and structural studies. DFT calculations and
experimental methods show that fulgimide and spiropyran moieties change the
oxidation potential of a pendant amidine group. The first generations of these
molecular switches show a potential for multimode switching as they can be modulated between open and closed by using light or electrochemically, and the spiropyran-based amidine can also be switched thermally. The color changes observed
in these compounds upon exposure to these stimuli could also possibly be developed into sensing technology. The ability for these photoswitches to modulate
electron density at the amidine group is summarized in Fig. 7.1.
Figure 7.1.
amidines.70
Summary of electron density movement for photoswitchable
Based on knowledge gained from the amidine-based molecular switches, several
other potentially switchable molecules were proposed and studied computationally. DFT studies indicate both aminopyridine- and squaramide-based switches
should result in similar behavior to the previously studied amidines. Attempts
to synthesize a spiropyran-substituted aminopyridine did not yield successful results, although DFT calculations suggest that this type of molecular switch may
yield more dramatic results in the modulation of redox potential at the pendant
group those that have already been developed. A spiropyran-substituted squaramide was synthesized, although solubility issues made it difficult to study ex152
Texas Tech University, Miranda C. Andrews, December 2019
perimentally. Further improvements could be made to the structure to increase
solubility in the future. Because squaramides are used as hydrogen bond catalysts and in ion recognition, these processes could be modulated upon irradiation
or exposure to other stimuli. If the binding of other species leads to switching of
the spiropyran moiety, and that is accompanied by a color change, the proposed
squaramide could be used for sensing purposes.
For all of the proposed switches discussed in this work, many also afford the possibility of incorporating more than one switch. This could lead to more dramatic
effects on the electronic structure of the pendant groups. Additionally, molecular
switches with different pendant groups could be developed and studied.
The studies of the modulation of the amidine group was then extended to firstrow transition metal complexes that include one photoswitchable aminidate ligand. These compounds were explored computationally. Many of these computational studies yielded expected results; upon ring-opening of a pendant spiropyran
group, there is electron density lost at the metal ion as well. Some metal complexes,
namely chromium(II), nickel(II), and low-spin iron(II) complexes, were calculated
to gain electron density upon ring-opening. This is unexpected and should be explored further. These changes in electron density would necessarily have an effect
on the reactivity of the metal ion.
Computational studies of the proposed metal complexes also indicated that molecular switches like spiropyrans could induce a spin state change, particularly
in iron(II), but also potentially in cobalt(II). There is already some literature precedence for a photoswitch-induced spin state change in a cobalt complex. This
could have potential applications in the field of optoelectronics and light harvesting. Synthesis and experimental studies of these compounds should follow to confirm what has been learned computationally.
The idea of modulating pendant groups was also applied to secondary bonding
interactions. Switchable halogen bond donors were synthesized and structurally
characterized. DFT studies and surface analyses indicate that the electrostatic potential of a pendant iodine can be modulated by the isomerization of spiropyran. A
153
Texas Tech University, Miranda C. Andrews, December 2019
crystal structure was determined that shows a halogen bond taking place between
one state of a photoswitch and 4,4’-bipyridine, with a bond distance well below
the sum of the van der Waals radii for the species involved. Solution state binding studies in various solvents (namely, varying the dielectric constants) would
further elucidate the potential halogen bond donor ability of iodospiropyrans that
have been synthesized.
Two potentially electroswitchable new halogen bond donors were proposed. Computational studies indicate the halo-BODIPYs would be excellent halogen bond
donors as synthesized. The diiodoBODIPY (12) that was synthesized shows halogen bonding in the solid state, as shown by the crystal structure determined and
presented. Additionally, the 12 may be investigated as a singlet oxygen generator,
as shown by a crystal structure of a BODIPY dimer that appears to have resulted from a reaction between 12 and singlet oxygen, which resulted in C-H bond
activation.
Additionally, binding constants were measured for antimony(III) and bismuth(III)
pnictogen bond donors with several neutral and anionic pnictogen bond acceptors. Both pnictogen bond donors show good solution state binding with neutral
and anionic species, with binding constants on the order of 104 to 106 M−1 for the
first binding events. The bismuth(III) compound study may show additional and
better binding than the antimony(III) analogue. The more robust antimony(III) catecholate that was studied also showed good binding with halides, despite possible
competition that may affect one of the primary bonds, and therefore the secondary
bonding interactions, of this compound.
Several antimony(III) catecholates that incorporate a spiropyran moiety were proposed and studied computationally. These calculated structures indicate that the
electrostatic potentials of the σ-holes of the antimony(III) could be modulated upon
ring-opening of the spiropyran moiety. The idea of modulating secondary bonding
interactions could be applied in the design of systems that assemble or disassemble
upon exposure to different wavelengths of light or other stimuli.
154
Texas Tech University, Miranda C. Andrews, December 2019
7.1
Concluding Remarks
Overall, this work has laid the foundation for the use of switchable molecules to
modulate measurable changes in the ligand field of metal complexes and measurable changes in pendant groups that can form secondary bonding interactions.
This work begins to describe the basic underlying concepts that drive these changes,
in the hopes that interesting and highly applicable systems can be further designed, synthesized, and fully characterized.
155
Texas Tech University, Miranda C. Andrews, December 2019
BIBLIOGRAPHY
[1] Jean-Pierre Sauvage. Molecular Machines and Motors (Structure and Bonding).
Structure and Bonding. Springer Berlin Heidelberg, New York, 2001.
[2] Estı́baliz Merino and Marı́a Ribagorda. Control over molecular motion using
the cistrans photoisomerization of the azo group. Beilstein Journal of Organic
Chemistry, 8:1071–1090, 2012.
[3] Masahiro Irie, Thorsten Lifka, Kingo Uchida, Seiya Kobatake, and Yuriko
Shindo. Fatigue resistant properties of photochromic dithienylethenes: byproduct formation. Chemical Communications, (8):747–750, January 1999.
[4] Vincent De Waele, Uli Schmidhammer, Thomas Mrozek, Jörg Daub, and
Eberhard Riedle. Ultrafast bidirectional dihydroazulene/vinylheptafulvene
(DHA/VHF) molecular switches: photochemical ring closure of vinylheptafulvene proven by a two-pulse experiment. Journal of the American Chemical
Society, 124(11):2438–2439, March 2002.
[5] Caihong Wang, Kei Hashimoto, Ryota Tamate, Hisashi Kokubo, and Masayoshi Watanabe. Controlled Sol-Gel Transitions of a Thermoresponsive Polymer in a Photoswitchable Azobenzene Ionic Liquid as a Molecular Trigger.
Angewandte Chemie International Edition, 57(1):227–230, 2018.
[6] Kazuki Urakawa, Michinori Sumimoto, Mitsuhiro Arisawa, Masaki Matsuda, and Hayato Ishikawa. Redox Switching of Orthoquinone-Containing
Aromatic Compounds with Hydrogen and Oxygen Gas. Angewandte Chemie
International Edition, 55(26):7432–7436, 2016.
[7] Sergi Garcia-Manyes, Jian Liang, Robert Szoszkiewicz, Tzu-Ling Kuo, and
Julio M. Fernández. Force-activated reactivity switch in a bimolecular chemical reaction. Nature Chemistry, 1(3):236–242, June 2009.
[8] Masakazu Morimoto and Masahiro Irie. A Diarylethene Cocrystal that Converts Light into Mechanical Work. Journal of the American Chemical Society,
132(40):14172–14178, October 2010.
[9] Søren Lindbæk Broman, Martyn Jevric, Andrew D. Bond, and Mogens Brøndsted Nielsen. Syntheses of Donor-Acceptor-Functionalized Dihydroazulenes. The Journal of Organic Chemistry, 79(1):41–64, January 2014.
156
Texas Tech University, Miranda C. Andrews, December 2019
[10] Helmut Goerner, Christian Fischer, Sebastian Gierisch, and Joerg Daub. Dihydroazulene/vinylheptafulvene photochromism: effects of substituents,
solvent, and temperature in the photorearrangement of dihydroazulenes to
vinylheptafulvenes. The Journal of Physical Chemistry, 97(16):4110–4117, April
1993.
[11] Stuart Aiken, Ross J. L. Edgar, Christopher D. Gabbutt, B. Mark Heron, and
Peter A. Hobson. Negatively photochromic organic compounds: Exploring
the dark side. Dyes and Pigments, 149:92–121, February 2018.
[12] Zachary S. Kean, Ashley L. Black Ramirez, Yufan Yan, and Stephen L. Craig.
Bicyclo[3.2.0]heptane Mechanophores for the Non-scissile and Photochemically Reversible Generation of Reactive Bis-enones. Journal of the American
Chemical Society, 134(31):12939–12942, August 2012.
[13] Gregory R. Gossweiler, Tatiana B. Kouznetsova, and Stephen L. Craig. ForceRate Characterization of Two Spiropyran-Based Molecular Force Probes.
Journal of the American Chemical Society, 137(19):6148–6151, May 2015.
[14] Jeremy M. Lenhardt, Ashley L. Black, and Stephen L. Craig.
gemDichlorocyclopropanes as Abundant and Efficient Mechanophores in Polybutadiene Copolymers under Mechanical Stress. Journal of the American
Chemical Society, 131(31):10818–10819, August 2009.
[15] Martin Krupička, Wolfram Sander, and Dominik Marx. Mechanical Manipulation of Chemical Reactions: Reactivity Switching of Bergman Cyclizations.
The Journal of Physical Chemistry Letters, 5(5):905–909, March 2014.
[16] Gregory R. Gossweiler, Cameron L. Brown, Gihan B. Hewage, Eitan SapiroGheiler, William J. Trautman, Garrett W. Welshofer, and Stephen L. Craig.
Mechanochemically Active Soft Robots. ACS Applied Materials & Interfaces,
7(40):22431–22435, October 2015.
[17] Meredith H. Barbee, Kunal Mondal, John Z. Deng, Vivek Bharambe,
Taylor V. Neumann, Jacob J. Adams, Nicholas Boechler, Michael D. Dickey,
and Stephen L. Craig. Mechanochromic Stretchable Electronics. ACS Applied
Materials & Interfaces, 10(35):29918–29924, September 2018.
[18] Takahiro Kosuge, Xiaolei Zhu, Vivian M. Lau, Daisuke Aoki, Todd J. Martinez, Jeffrey S. Moore, and Hideyuki Otsuka. Multicolor Mechanochromism
of a Polymer/Silica Composite with Dual Distinct Mechanophores. Journal
of the American Chemical Society, 141(5):1898–1902, February 2019.
157
Texas Tech University, Miranda C. Andrews, December 2019
[19] Jun Li, Chikkannagari Nagamani, and Jeffrey S. Moore. Polymer Mechanochemistry: From Destructive to Productive. Accounts of Chemical Research,
48(8):2181–2190, August 2015.
[20] Yoshimitsu Sagara, Marc Karman, Atsushi Seki, Mehboobali Pannipara,
Nobuyuki Tamaoki, and Christoph Weder. Rotaxane-Based Mechanophores
Enable Polymers with Mechanically Switchable White Photoluminescence.
ACS Central Science, April 2019.
[21] Audrey R. Sulkanen, Jaeuk Sung, Maxwell J. Robb, Jeffrey S. Moore,
Nancy R. Sottos, and Gang-yu Liu. Spatially Selective and DensityControlled Activation of Interfacial Mechanophores. Journal of the American
Chemical Society, 141(9):4080–4085, March 2019.
[22] Zachary S. Kean, Zhenbin Niu, Gihan B. Hewage, Arnold L. Rheingold, and
Stephen L. Craig. Stress-Responsive Polymers Containing Cyclobutane Core
Mechanophores: Reactivity and Mechanistic Insights. Journal of the American
Chemical Society, 135(36):13598–13604, September 2013.
[23] T. Karlessi, M. Santamouris, K. Apostolakis, A. Synnefa, and I. Livada. Development and testing of thermochromic coatings for buildings and urban
structures. Solar Energy, 83(4):538–551, April 2009.
[24] Michel Sliwa, Arnaud Spangenberg, Isabelle Malfant, Pascal G. Lacroix,
Rémi Métivier, Robert B. Pansu, and Keitaro Nakatani. Structural, Optical, and Theoretical Studies of a Thermochromic Organic Crystal with
Reversibly Variable Second Harmonic Generation. Chemistry of Materials,
20(12):4062–4068, June 2008.
[25] Stephen H. Kawai, Sylvain L. Gilat, Rachel Ponsinet, and Jean-Marie Lehn. A
Dual-Mode Molecular Switching Device: Bisphenolic Diarylethenes with Integrated Photochromic and Electrochromic Properties. Chemistry A European
Journal, 1(5):285–293, August 1995.
[26] Claire Fave and Bernd Schöllhorn. Electrochemical Activation of Halogen
Bonding. Current Opinion in Electrochemistry, 15:89–96, April 2019.
[27] Kenji Kanazawa, Kazuki Nakamura, and Norihisa Kobayashi.
Electroswitching of Emission and Coloration with Quick Response and High Reversibility in an Electrochemical Cell. Chemistry An Asian Journal, 7(11):2551–
2554, November 2012.
158
Texas Tech University, Miranda C. Andrews, December 2019
[28] Kenji Kanazawa, Kazuki Nakamura, and Norihisa Kobayashi. HighContrast Electroswitching of Emission and Coloration Based on SingleMolecular Fluoran Derivatives.
The Journal of Physical Chemistry A,
118(31):6026–6033, August 2014.
[29] David W. Steuerman, Hsian-Rong Tseng, Andrea J. Peters, Amar H. Flood,
Jan O. Jeppesen, Kent A. Nielsen, J. Fraser Stoddart, and James R. Heath.
Molecular-Mechanical Switch-Based Solid-State Electrochromic Devices.
Angewandte Chemie International Edition, 43(47):6486–6491, December 2004.
[30] Michael Åxman Petersen, Asbjørn Sune Andersson, Kristine Kilså,
and Mogens Brøndsted Nielsen.
Redox-Controlled DihydroazuleneVinylheptafulvene Photoswitch Incorporating Tetrathiafulvalene. European
Journal of Organic Chemistry, 2009(12):1855–1858, 2009.
[31] Brian M. Muller, Jesse Mai, Reid A. Yocum, and Marc J. Adler. Impact of
mono- and disubstitution on the colorimetric dynamic covalent switching
chalcone/flavanone scaffold. Organic & Biomolecular Chemistry, 12(28):5108–
5114, 2014.
[32] Yasuhiro Shiraishi, Kohei Yamamoto, Shigehiro Sumiya, and Takayuki Hirai.
Spiropyran as a reusable chemosensor for selective colorimetric detection of
aromatic thiols. Physical Chemistry Chemical Physics, 16(24):12137–12142, May
2014.
[33] Brett E. Erickson. Self-Darkening Eyeglasses. Chemical & Engineering News,
87(15):54, April 2009.
[34] Vincenzo Balzani, Miguel Clemente-León, Alberto Credi, Belén Ferrer, Margherita Venturi, Amar H. Flood, and J. Fraser Stoddart. Autonomous artificial nanomotor powered by sunlight. Proceedings of the National Academy of
Sciences, 103(5):1178–1183, January 2006.
[35] Nagatoshi Koumura, Robert W. J. Zijlstra, Richard A. van Delden, Nobuyuki
Harada, and Ben L. Feringa. Light-driven monodirectional molecular rotor.
Nature, 401(6749):152, September 1999.
[36] Laurence Raehm, Jean-Marc Kern, and Jean-Pierre Sauvage. A Transition
Metal Containing Rotaxane in Motion: Electrochemically Induced Pirouetting of the Ring on the Threaded Dumbbell. Chemistry A European Journal,
5(11):3310–3317, 1999.
159
Texas Tech University, Miranda C. Andrews, December 2019
[37] Jérôme Vachon, Gregory T. Carroll, Michael M. Pollard, Emile M. Mes, Albert M. Brouwer, and Ben L. Feringa. An ultrafast surface-bound photoactive molecular motor. Photochemical & Photobiological Sciences, 13(2):241–
246, January 2014.
[38] Tibor Kudernac, Nopporn Ruangsupapichat, Manfred Parschau, Beatriz
Maciá, Nathalie Katsonis, Syuzanna R. Harutyunyan, Karl-Heinz Ernst, and
Ben L. Feringa. Electrically driven directional motion of a four-wheeled molecule on a metal surface. Nature, 479(7372):208–211, November 2011.
[39] M. Fritzsche. Note sur les carrbures d’hydrogène solides, tirés du guodron
de houille. Comptes rendus de l’Académie des Sciences, 69:1035–1037, 1867.
[40] Aaron J. Teator, Yuan Tian, Mu Chen, Jeehiun K. Lee, and Christopher W. Bielawski.
An Isolable, Photoswitchable N-Heterocyclic
Carbene: On-Demand Reversible Ammonia Activation. Angewandte Chemie,
127(39):11721–11725, September 2015. 00006.
[41] Bethany M. Neilson and Christopher W. Bielawski.
Photoswitchable Organocatalysis: Using Light To Modulate the Catalytic Activities of N-Heterocyclic Carbenes. Journal of the American Chemical Society,
134(30):12693–12699, August 2012. 00041.
[42] Bethany M. Neilson and Christopher W. Bielawski. Photoswitchable MetalMediated Catalysis: Remotely Tuned Alkene and Alkyne Hydroborations.
Organometallics, 32(10):3121–3128, May 2013.
[43] J. W. Verhoeven. Glossary of terms used in photochemistry (IUPAC Recommendations 1996). Pure and Applied Chemistry, 68(12):2223–2286, 1996.
[44] Th. Förster. Diabatic and adiabatic processes in photochemistry. Pure and
Applied Chemistry, 24(3):443–450, 1970.
[45] Masahiro Irie. Diarylethenes for Memories and Switches. Chemical Reviews,
100(5):1685–1716, May 2000.
[46] Masahiro Irie, Tuyoshi Fukaminato, Kenji Matsuda, and Seiya Kobatake.
Photochromism of Diarylethene Molecules and Crystals: Memories,
Switches, and Actuators. Chemical Reviews, 114(24):12174–12277, December
2014.
[47] Marie-Laure Boillot, Cécile Roux, Jean-Paul Audière, Aurèlie Dausse, and
Jacqueline Zarembowitch. Ligand-Driven Light-Induced Spin Change in
160
Texas Tech University, Miranda C. Andrews, December 2019
Transition-Metal Complexes: Selection of an Appropriate System and First
Evidence of the Effect, in FeII (4-styrylpyridine)4 (NCBPh3 )2 . Inorganic Chemistry, 35(13):3975–3980, January 1996.
[48] Marie-Laure Boillot, Jacqueline Zarembowitch, and Angélique Sour. LigandDriven Light-Induced Spin Change (LD-LISC): A Promising Photomagnetic
Effect. In Spin Crossover in Transition Metal Compounds II, number 234 in Topics in Current Chemistry, pages 261–276. Springer Berlin Heidelberg, 2004.
[49] Alexander Bannwarth, Sven Olaf Schmidt, Gerhard Peters, Frank D.
Sönnichsen, Wulf Thimm, Rainer Herges, and Felix Tuczek. FeIII SpinCrossover Complexes with Photoisomerizable Ligands: Experimental and
Theoretical Studies on the Ligand-Driven Light-Induced Spin Change Effect.
European Journal of Inorganic Chemistry, 2012(16):2776–2783, 2012.
[50] Michelle M. Paquette, Brian O. Patrick, and Natia L. Frank. Determining the Magnitude and Direction of Photoinduced Ligand Field Switching in Photochromic Metal-Organic Complexes: MolybdenumTetracarbonyl Spirooxazine Complexes. Journal of the American Chemical Society,
133(26):10081–10093, July 2011. 00016.
[51] Masayuki Nihei, Yukihide Suzuki, Norihisa Kimura, Yosuke Kera, and
Hiroki Oshio. Bidirectional Photomagnetic Conversions in a Spin-Crossover
Complex with a Diarylethene Moiety. Chemistry A European Journal,
19(22):6946–6949, May 2013.
[52] Cecile Roux, Jacqueline Zarembowitch, Bernard Gallois, Thierry Granier,
and Renee Claude. Toward Ligand-Driven Light-Induced Spin Changing. Influence of the Configuration of 4 Styrylpyridine (stpy) on the Magnetic Properties of FeII(stpy)4(NCS)2 Complexes. Crystal Structures of the
Spin-Crossover Species Fe(trans-stpy)4(NCS)2 and of the High-Spin Species
Fe(cis-stpy)4(NCS)2. Inorganic Chemistry, 33(10):2273–2279, May 1994.
[53] Benedikt Rösner, Magdalena Milek, Alexander Witt, Benoı̂t Gobaut, Piero
Torelli, Rainer H. Fink, and Marat M. Khusniyarov. Reversible Photoswitching of a Spin-Crossover Molecular Complex in the Solid State at Room Temperature. Angewandte Chemie International Edition, 54(44):12976–12980, 2015.
[54] Michelle M. Paquette, Daniel Plaul, Aiko Kurimoto, Brian O. Patrick, and
Natia L. Frank. Opto-Spintronics: Photoisomerization-Induced Spin State
Switching at 300 K in Photochrome Cobalt-Dioxolene Thin Films. Journal of
the American Chemical Society, 140(44):14990–15000, November 2018.
161
Texas Tech University, Miranda C. Andrews, December 2019
[55] Shah M. Reduwan Billah, Robert M. Christie, and Renzo Shamey. Direct
coloration of textiles with photochromic dyes. Part 1: Application of spiroindolinonaphthoxazines as disperse dyes to polyester, nylon and acrylic fabrics. Coloration Technology, 124(4):223–228, 2008.
[56] Christian Richard Parker, Christian Gregers Tortzen, Søren Lindæk
Broman,
Magnus Schau-Magnussen,
Kristine Kilså,
and Mogens Brøndsted Nielsen. Lewis acid enhanced switching of the 1,1dicyanodihydroazulene/vinylheptafulvene photo/thermoswitch. Chemical
Communications, 47(21):6102–6104, May 2011.
[57] Søren Lindbæk Broman, Anne Ugleholdt Petersen, Christian Gregers
Tortzen, Johan Vibenholt, Andrew D. Bond, and Mogens Brøndsted Nielsen.
A Bis(heptafulvenyl)-dicyanoethylene Thermoswitch with Two Sites for
Ring Closure. Organic Letters, 14(1):318–321, January 2012.
[58] Xiu-Shuang Xing, Rong-Jian Sa, Pei-Xin Li, Ning-Ning Zhang, Zhong-Yuan
Zhou, Bin-Wen Liu, Jie Liu, Ming-Sheng Wang, and Guo-Cong Guo. Secondorder nonlinear optical switching with a record-high contrast for a photochromic and thermochromic bistable crystal. Chemical Science, 8(11):7751–
7757, October 2017.
[59] Masaaki Ohba, Ko Yoneda, Gloria Agustı́, M. Carmen Muñoz, Ana B. Gaspar, José A. Real, Mikio Yamasaki, Hideo Ando, Yoshihide Nakao, Shigeyoshi Sakaki, and Susumu Kitagawa. Bidirectional Chemo-Switching of Spin
State in a Microporous Framework. Angewandte Chemie International Edition,
48(26):4767–4771, 2009.
[60] Edward I. Balmond, Brandon K. Tautges, Andrea L. Faulkner, Victor W. Or,
Blanka M. Hodur, Jared T. Shaw, and Angelique Y. Louie. Comparative Evaluation of Substituent Effect on the Photochromic Properties of Spiropyrans
and Spirooxazines. The Journal of Organic Chemistry, 81(19):8744–8758, October 2016.
[61] Craig J. Roxburgh, Peter G. Sammes, and Ayse Abdullah. Steric and electronically biasing substituent effects on the Photoreversibility of novel, 3’, 5’- and 3-substituted indolospirobenzopyrans. Thermal evaluation using
1
H NMR spectroscopy and Overhauser enhancement studies. Dyes and Pigments, 83(1):31–50, October 2009.
[62] Helmut Görner. Photochromism of nitrospiropyrans: effects of structure,
solvent and temperature. Physical Chemistry Chemical Physics, 3(3):416–423,
January 2001.
162
Texas Tech University, Miranda C. Andrews, December 2019
[63] Alexander K. Chibisov and Helmut Görner. Photochromism of spirobenzopyranindolines and spironaphthopyranindolines. Physical Chemistry Chemical Physics, 3(3):424–431, January 2001.
[64] Yinghong Sheng, Jerzy Leszczynski, Antonio A. Garcia, Rohit Rosario, Devens Gust, and Joseph Springer. Comprehensive Theoretical Study of the
Conversion Reactions of Spiropyrans: Substituent and Solvent Effects. The
Journal of Physical Chemistry B, 108(41):16233–16243, October 2004.
[65] Gaia Tomasello, Michael J. Bearpark, Michael A. Robb, Giorgio Orlandi, and
Marco Garavelli. Significance of a Zwitterionic State for Fulgide Photochromism: Implications for the Design of Mimics. Angewandte Chemie International Edition, 49(16):2913–2916, April 2010.
[66] Soichi Uchida, Shintaro Yamada, Yasushi Yokoyama, and Yukio Kurita.
Steric Effects of Substituents on the Photochromism of Indolylfulgides. Bulletin of the Chemical Society of Japan, 68(6):1677–1682, June 1995.
[67] Soichi Uchida, Yasushi Yokoyama, Jitsuo Kiji, Tamon Okano, and Hitoshi
Kitamura. Electronic Effects of Substituents on Indole Nitrogen on the Photochromic Properties of Indolylfulgides. Bulletin of the Chemical Society of
Japan, 68(10):2961–2967, October 1995.
[68] Yasushi Yokoyama. Fulgides for Memories and Switches. Chemical Reviews,
100(5):1717–1740, May 2000.
[69] Rohith Chindam, Hoque M. Hoque, Amar S. Ali, Fatima Z. Rafique, and
Jonathan D. Gough. Theoretical assessment of indolylfulgimides and novel
asymmetric di-indolylfulgimide photochromes. Journal of Photochemistry and
Photobiology A: Chemistry, 279:38–46, April 2014.
[70] M. Crawford Andrews, Ping Peng, Amit Rajput, and Anthony F. Cozzolino. Modulation of the carboxamidine redox potential through photoinduced spiropyran or fulgimide isomerisation. Photochemical & Photobiological
Sciences, 17:432–441, March 2018.
[71] James Barker and Melvyn Kilner. The coordination chemistry of the amidine
ligand. Coordination Chemistry Reviews, 133:219–300, July 1994.
[72] F. Albert Cotton, Lee M. Daniels, John H. Matonic, and Carlos A. Murillo.
Highly distorted diiron(II, II) complexes containing four amidinate ligands.
A long and short metal-metal distance. Inorganica Chimica Acta, 256(2):277–
282, March 1997.
163
Texas Tech University, Miranda C. Andrews, December 2019
[73] F. Albert Cotton, Xuejen Feng, and Carlos A. Murillo. Electronic structure
of dinuclear trigonal-lantern amidinato compounds of iron and cobalt. Inorganica Chimica Acta, 256(2):303–308, March 1997.
[74] Yongchao Liang, A. S Dvornikov, and P. M Rentzepis. Photochemistry of
photochromic 2-indolylfulgides with substituents at the 1’-position of the
indolylmethylene moiety. Journal of Photochemistry and Photobiology A: Chemistry, 146(12):83–93, December 2001.
[75] Yongchao Liang, A. S. Dvornikov, and P. M. Rentzepis. Synthesis of
novel photochromic fluorescing 2-indolylfulgimides. Tetrahedron Letters,
40(46):8067–8069, November 1999.
[76] Joakim Andréasson, Stephen D. Straight, Thomas A. Moore, Ana L. Moore,
and Devens Gust. Molecular All-Photonic Encoder-Decoder. Journal of the
American Chemical Society, 130(33):11122–11128, August 2008.
[77] A. S. Dvornikov, Y. Liang, C. S. Cruse, and P. M. Rentzepis. Spectroscopy and
Kinetics of a Molecular Memory with Nondestructive Readout for Use in 2D
and 3D Storage Systems. The Journal of Physical Chemistry B, 108(25):8652–
8658, June 2004.
[78] Li Li, Fu-Quan Bai, Jian Wang, and Hong-Xing Zhang. Theoretical investigation on the spectroscopy prosperities of four isomers of an encoder molecule
FGDTE. Dyes and Pigments, 107:108–117, August 2014.
[79] Gaowa Naren, Shiming Li, and Joakim Andréasson. One-Time Password
Generation and Two-Factor Authentication Using Molecules and Light.
ChemPhysChem, pages 1726–1729, January 2017.
[80] Ayse Abdullah, Thomas G. Nevell, Peter G. Sammes, and Craig J. Roxburgh.
Unusual thermo(photo)chromic properties of some mononitro- and dinitrosubstituted 3’-alkyl indolospirobenzopyrans. Dyes and Pigments, 121:57–72,
October 2015.
[81] Nithyanandan Sivasankaran and Kannan Palaninathan. Photochromic
switchable pendant indolyl fulgimide polypyrrole. Polymer Degradation and
Stability, 98(9):1852–1861, September 2013.
[82] James T. C. Wojtyk, Adnaan Wasey, Peter M. Kazmaier, Shmaryahu Hoz, and
Erwin Buncel. Thermal Reversion Mechanism of N-Functionalized Merocyanines to Spiropyrans: A Solvatochromic, Solvatokinetic, and Semiempirical
Study. The Journal of Physical Chemistry A, 104(39):9046–9055, October 2000.
164
Texas Tech University, Miranda C. Andrews, December 2019
[83] Alexander K. Chibisov and Helmut Görner. Complexes of spiropyranderived merocyanines with metal ions: relaxation kinetics, photochemistry
and solvent effects. Chemical Physics, 237(3):425–442, October 1998.
[84] Michele Zanoni, Simon Coleman, Kevin J. Fraser, Robert Byrne, Klaudia
Wagner, Sanjeev Gambhir, David L. Officer, Gordon G. Wallace, and Dermot
Diamond. Physicochemical study of spiropyran-terthiophene derivatives:
photochemistry and thermodynamics. Physical Chemistry Chemical Physics,
14(25):9112–9120, June 2012.
[85] Benoit Daoust and Jean Lessard. Electrochemical behavior of amidine hydrochlorides and amidines. Canadian Journal of Chemistry, 73:362–374, 1995.
[86] Hanan E. Abdou, Ahmed A. Mohamed, and John P. Fackler. Synthesis,
Characterization, Luminescence, and Electrochemistry of New Tetranuclear
Gold(I) Amidinate Clusters: Au4 [PhNC(Ph)NPh]4 , Au4 [PhNC(CH3 )NPh]4 ,
and Au4 [ArNC(H)NAr]4 . Journal of Cluster Science, 18(3):630–641, September
2007.
[87] Marye Anne Fox and John R. Hurst. Electrochemically induced pericyclic reactions. A radical anionic cyclization. Journal of the American Chemical Society,
106(24):7626–7627, November 1984.
[88] Oleksii Ivashenko, Jochem T. van Herpt, Petra Rudolf, Ben L. Feringa,
and Wesley R. Browne. Oxidative electrochemical aryl C-C coupling of
spiropyrans. Chemical Communications, 49(60):6737–6739, July 2013.
[89] Mylene Campredon, Gerard Giusti, Robert Guglielmetti, Andre Samat,
Gerard Gronchi, Angelo Alberti, and Massimo Benaglia. Radical ions
and germyloxyaminoxyls from nitrospiro[indoline-naphthopyrans]. A combined electrochemical and EPR study. Journal of the Chemical Society, Perkin
Transactions 2, (11):2089–2094, 1993.
[90] Georgina Armendáriz-Vidales, Eduardo Martı́nez-González, Denhy
Hernández-Melo, Jorge Tiburcio, and Carlos Frontana. Electrochemical
Characterization of Spiropyran Structures. Procedia Chemistry, 12:41–46,
2014.
[91] José A. Mata, F. Ekkehardt Hahn, and Eduardo Peris. Heterometallic
complexes, tandem catalysis and catalytic cooperativity. Chemical Science,
5(5):1723–1732, April 2014.
165
Texas Tech University, Miranda C. Andrews, December 2019
[92] Toshikazu Hirao. Conjugated systems composed of transition metals
and redox-active π-conjugated ligands. Coordination Chemistry Reviews,
226(1):81–91, March 2002.
[93] Alan M. Allgeier and Chad A. Mirkin. Ligand Design for Electrochemically
Controlling Stoichiometric and Catalytic Reactivity of Transition Metals. Angewandte Chemie International Edition, 37(7):894–908, April 1998.
[94] Volodymyr Lyaskovskyy and Bas de Bruin. Redox Non-Innocent Ligands:
Versatile New Tools to Control Catalytic Reactions. ACS Catalysis, 2(2):270–
279, February 2012.
[95] Michael J. Preigh, Mark T. Stauffer, Fu-Tyan Lin, and Stephen G. Weber. Anodic oxidation mechanism of a spiropyran. Journal of the Chemical Society,
Faraday Transactions, 92(20):3991–3996, January 1996.
[96] Awal Noor, Winfried P. Kretschmer, Germund Glatz, and Rhett Kempe.
Highly Active Titanium-Based Olefin Polymerization Catalysts Bearing
Bulky Aminopyridinato Ligands. Inorganic Chemistry, 50(10):4598–4606, May
2011.
[97] Jad El Haj Hassan, Vasily Radkov, Vincent Dorcet, Jean-François Carpentier, and Evgueni Kirillov. Scandium and yttrium complexes of an hybrid
phenoxy-amidopyridinate ligand. Use in ROP of racemic lactide. Journal of
Organometallic Chemistry, 823:34–39, November 2016.
[98] Awal Noor, Emmanuel Sobgwi Tamne, Benjamin Oelkers, Tobias Bauer, Serhiy Demeshko, Franc Meyer, Frank W. Heinemann, and Rhett Kempe. From
Chromium-Chromium Quintuple Bonds to Molecular Squares and Porous
Coordination Polymers. Inorganic Chemistry, 53(23):12283–12288, December
2014.
[99] Neeraj Kumar Mishra, Miji Choi, Hyeim Jo, Yongguk Oh, Satyasheel
Sharma, Sang Hoon Han, Taejoo Jeong, Sangil Han, Seok-Yong Lee, and
In Su Kim. Direct C-H alkylation and indole formation of anilines with
diazo compounds under rhodium catalysis. Chemical Communications,
51(97):17229–17232, November 2015.
[100] Ai Ito, Hiroto Fujino, Keiko Ushiyama, Eriko Yamanaka, Ryu Yamasaki, and
Iwao Okamoto. Acid-controlled switching of conformational preference of
N,N-diarylamides bearing pyridine. Tetrahedron Letters, 57(42):4737–4741,
October 2016.
166
Texas Tech University, Miranda C. Andrews, December 2019
[101] R. Ian Storer, Caroline Aciro, and Lyn H. Jones. Squaramides: physical properties, synthesis and applications. Chemical Society Reviews, 40(5):2330–2346,
April 2011.
[102] Pankaj Chauhan, Suruchi Mahajan, Uǧur Kaya, Daniel Hack, and Dieter Enders. Bifunctional Amine-Squaramides: Powerful Hydrogen-Bonding Organocatalysts for Asymmetric Domino/Cascade Reactions. Advanced Synthesis & Catalysis, 357(2-3):253–281, February 2015.
[103] David Quiñonero, Antonio Frontera, Pau Ballester, and Pere M Deyá. A
theoretical study of aromaticity in squaramide and oxocarbons. Tetrahedron
Letters, 41(12):2001–2005, March 2000.
[104] Seth M. Cohen, Zhenjie Zhang, and Jake A. Boissonnault. Toward “metalloMOFzymes”: Metal-Organic Frameworks with Single-Site Metal Catalysts
for Small-Molecule Transformations. Inorganic Chemistry, 55(15):7281–7290,
August 2016.
[105] Xiaoping Zhang, Zhenjie Zhang, Jake Boissonnault, and Seth M. Cohen. Design and synthesis of squaramide-based MOFs as efficient MOFsupported hydrogen-bonding organocatalysts. Chemical Communications,
52(55):8585–8588, 2016.
[106] Nathaniel C. Lim, Svetlana V. Pavlova, and Christian Bruckner. Squaramide
Hydroxamate-Based Chemidosimeter Responding to Iron(III) with a Fluorescence Intensity Increase. Inorg. Chem., 48(3):1173–1182, December 2008.
[107] Ali Rostami, Alexis Colin, Xiao Yu Li, Michael G. Chudzinski, Alan J. Lough,
and Mark S. Taylor. N,N-Diarylsquaramides: General, High-Yielding Synthesis and Applications in Colorimetric Anion Sensing. The Journal of Organic
Chemistry, 75(12):3983–3992, June 2010.
[108] Lennard Krause, Regine Herbst-Irmer, George M. Sheldrick, and Dietmar
Stalke. Comparison of silver and molybdenum microfocus X-ray sources
for single-crystal structure determination. Journal of Applied Crystallography,
48(Pt 1):3–10, February 2015.
[109] G. M. Sheldrick. Crystal structure refinement with SHELXL. Acta Crystallographica Section C: Structural Chemistry, 71(1):3–8, January 2015.
[110] John P. Perdew and Yue Wang. Accurate and simple analytic representation
of the electron-gas correlation energy. Physical Review B, 45(23):13244–13249,
June 1992.
167
Texas Tech University, Miranda C. Andrews, December 2019
[111] Florian Weigend and Reinhart Ahlrichs. Balanced basis sets of split valence,
triple zeta valence and quadruple zeta valence quality for H to Rn: Design
and assessment of accuracy. Physical Chemistry Chemical Physics, 7(18):3297–
3305, August 2005.
[112] Christian A. Nijhuis, Erica Jellema, Timo J. J. Sciarone, Auke Meetsma,
Peter H. M. Budzelaar, and Bart Hessen. First-Row Transition Metal
Bis(amidinate) Complexes; Planar Four-Coordination of FeII Enforced by
Sterically Demanding Aryl Substituents. European Journal of Inorganic Chemistry, 2005(11):2089–2099, June 2005.
[113] Yoshiaki Nishibayashi. Recent Progress in Transition-Metal-Catalyzed Reduction of Molecular Dinitrogen under Ambient Reaction Conditions. Inorganic Chemistry, 54(19):9234–9247, October 2015.
[114] Kasper P. Jensen and Jordi Cirera. Accurate Computed Enthalpies of Spin
Crossover in Iron and Cobalt Complexes. The Journal of Physical Chemistry A,
113(37):10033–10039, September 2009.
[115] Rupal Gupta, Taketo Taguchi, Benedikt Lassalle-Kaiser, Emile L. Bominaar,
Junko Yano, Michael P. Hendrich, and A. S. Borovik. High-spin Mn-oxo
complexes and their relevance to the oxygen-evolving complex within photosystem II. Proceedings of the National Academy of Sciences, 112(17):5319–5324,
April 2015.
[116] Theodore A. Betley, Yogesh Surendranath, Montana V. Childress, Glen E.
Alliger, Ross Fu, Christopher C. Cummins, and Daniel G. Nocera. A ligand
field chemistry of oxygen generation by the oxygen-evolving complex and
synthetic active sites. Philosophical Transactions of the Royal Society B: Biological
Sciences, 363(1494):1293–1303, March 2008.
[117] Mian Guo, Teresa Corona, Kallol Ray, and Wonwoo Nam. Heme and Nonheme High-Valent Iron and Manganese Oxo Cores in Biological and Abiological Oxidation Reactions. ACS Central Science, 5(1):13–28, January 2019.
[118] Jianmin Tao, John P. Perdew, Viktor N. Staroverov, and Gustavo E. Scuseria.
Climbing the Density Functional Ladder: Nonempirical Meta–Generalized
Gradient Approximation Designed for Molecules and Solids. Physical Review
Letters, 91(14):146401, September 2003.
[119] Elangannan Arunan, Gautam R. Desiraju, Roger A. Klein, Joanna Sadlej,
Steve Scheiner, Ibon Alkorta, David C. Clary, Robert H. Crabtree, Joseph J.
168
Texas Tech University, Miranda C. Andrews, December 2019
Dannenberg, Pavel Hobza, Henrik G. Kjaergaard, Anthony C. Legon, Benedetta Mennucci, and David J. Nesbitt. Definition of the hydrogen bond
(IUPAC Recommendations 2011). Pure and Applied Chemistry, 83(8):1637–
1641, July 2011.
[120] Gautam R. Desiraju, P. Shing Ho, Lars Kloo, Anthony C. Legon, Roberto
Marquardt, Pierangelo Metrangolo, Peter Politzer, Giuseppe Resnati, and
Kari Rissanen. Definition of the halogen bond (IUPAC Recommendations
2013). Pure and Applied Chemistry, 85(8):1711–1713, 2013.
[121] N. W. Alcock. Secondary Bonding to Nonmetallic Elements. Advances in
Inorganic Chemistry and Radiochemistry, 15:1–58, 1972.
[122] Lydia C. Gilday, Nicholas G. White, and Paul D. Beer. Halogen- and
hydrogen-bonding triazole-functionalised porphyrin-based receptors for anion recognition. Dalton Transactions, 42(44):15766–15773, October 2013.
[123] Antonio Caballero, Nicholas G. White, and Paul D. Beer. A Bidentate
Halogen-Bonding Bromoimidazoliophane Receptor for Bromide Ion Recognition in Aqueous Media. Angewandte Chemie International Edition,
50(8):1845–1848, 2011.
[124] Fabiola Zapata, Antonio Caballero, Nicholas G. White, Tim D. W. Claridge,
Paulo J. Costa, Vı́tor Félix, and Paul D. Beer. Fluorescent Charge-Assisted
Halogen-Bonding Macrocyclic Halo-Imidazolium Receptors for Anion Recognition and Sensing in Aqueous Media. Journal of the American Chemical
Society, 134(28):11533–11541, July 2012.
[125] Oliver Dumele, Benedikt Schreib, Ulrike Warzok, Nils Trapp, Christoph A.
Schalley, and François Diederich. Halogen-Bonded Supramolecular Capsules in the Solid State, in Solution, and in the Gas Phase. Angewandte Chemie
International Edition, 56(4):1152–1157, 2017.
[126] Mark Kalaj, Mohammad R. Momeni, Kyle C. Bentz, Kyle S. Barcus,
Joseph M. Palomba, Francesco Paesani, and Seth M. Cohen. Halogen bonding in UiO-66 frameworks promotes superior chemical warfare agent simulant degradation. Chemical Communications, 2019.
[127] Young Jin Cho, Kee Yoon Rho, Sam Rok Keum, Sung Hoon Kim, and
Cheol Min Yoon. Synthesis of Conjugated Spiropyran Dyes via PalladiumCatalyzed Coupling Reaction. Synthetic Communications, 29(12):2061–2068,
June 1999.
169
Texas Tech University, Miranda C. Andrews, December 2019
[128] Young Jin Cho, Seung Hwan Lee, Jong Woo Bae, Hyung-Jung Pyun, and
Cheol Min Yoon. Fischer’s base as a protecting group: protection and deprotection of 2-hydroxybenzaldehydes. Tetrahedron Letters, 41(20):3915–3917,
May 2000.
[129] Jordan Atchison, Sukanta Kamila, Heather Nesbitt, Kieran A. Logan,
Dean M Nicholas, Colin Fowley, James Davis, Bridgeen Callan, Anthony
P McHale, and John F Callan. Iodinated cyanine dyes: a new class of sensitisers for use in NIR activated photodynamic therapy (PDT). Chemical Communications, 53(12):2009–2012, 2017.
[130] Gergely Knorr, Eszter Kozma, Judith M. Schaart, Krisztina Németh, György
Török, and Péter Kele. Bioorthogonally Applicable Fluorogenic CyanineTetrazines for No-Wash Super-Resolution Imaging. Bioconjugate Chemistry,
29(4):1312–1318, April 2018.
[131] A. Bondi. van der Waals Volumes and Radii. Journal of Physical Chemistry,
68(3):441–451, March 1964.
[132] Ying Quan, Qiu-Yan Li, Quan Zhang, Wen-Qiang Zhang, Han Lu, JunHao Yu, Jian Chen, Xinsheng Zhao, and Xiao-Jun Wang. A diiodo-bodipy
postmodified metal-organic framework for efficient heterogeneous organophotocatalysis. RSC Adv., 6:23995–23999, 2016.
[133] Zhong-Hua Pan, Geng-Geng Luo, Jing-Wei Zhou, Jiu-Xu Xia, Kai Fang, and
Rui-Bo Wu. A simple BODIPY-aniline-based fluorescent chemosensor as
multiple logic operations for the detection of pH and CO2 gas. Dalton Transactions, 43(22):8499–8507, May 2014.
[134] Anyanee Kamkaew, Siang Hui Lim, Hong Boon Lee, Lik Voon Kiew,
Lip Yong Chung, and Kevin Burgess. Bodipy dyes in photodynamic therapy. Chem. Soc. Rev., 42:77–88, 2013.
[135] Aurore Loudet and Kevin Burgess. BODIPY Dyes and Their Derivatives:
Syntheses and Spectroscopic Properties. Chemical Reviews, 107(11):4891–
4932, November 2007.
[136] R. Oliveira, S. Groni, C. Fave, M. Branca, F. Mavré, D. Lorcy, M. Fourmigué,
and B. Schöllhorn. Electrochemical activation of a tetrathiafulvalene halogen
bond donor in solution. Physical Chemistry Chemical Physics, 18(23):15867–
15873, June 2016.
170
Texas Tech University, Miranda C. Andrews, December 2019
[137] Raquel Oliveira, Sihem Groni, Antoine Vacher, Frédéric Barriére, Dominique Lorcy, Marc Fourmigué, Emmanuel Maisonhaute, Bernd Schöllhorn,
and Claire Fave. Electrochemical Activation of TTF-Based Halogen Bond
Donors: A Powerful, Selective and Sensitive Analytical Tool for Probing a
Weak Interaction in Complex Media. Chemistry Select, 3(31):8874–8880, 2018.
[138] Wanhua Wu, Huimin Guo, Wenting Wu, Shaomin Ji, and Jianzhang Zhao.
Organic Triplet Sensitizer Library Derived from a Single Chromophore
(BODIPY) with Long-Lived Triplet Excited State for Triplet-Triplet Annihilation Based Upconversion. The Journal of Organic Chemistry, 76(17):7056–7064,
September 2011.
[139] Takatoshi Yogo, Yasuteru Urano, Yukiko Ishitsuka, Fumio Maniwa, and
Tetsuo Nagano. Highly Efficient and Photostable Photosensitizer Based
on BODIPY Chromophore.
Journal of the American Chemical Society,
127(35):12162–12163, September 2005.
[140] Ling Huang, Jianzhang Zhao, Song Guo, Caishun Zhang, and Jie Ma. Bodipy Derivatives as Organic Triplet Photosensitizers for Aerobic Photoorganocatalytic Oxidative Coupling of Amines and Photooxidation of Dihydroxylnaphthalenes. The Journal of Organic Chemistry, 78(11):5627–5637,
June 2013.
[141] Ling Huang, Wenbo Yang, and Jianzhang Zhao.
Switching of the
triplet excited state of styryl 2,6-diiodo-bodipy and its application in acidactivatable singlet oxygen photosensitizing. The Journal of Organic Chemistry,
79(21):10240–10255, 2014.
[142] Fadi M. Jradi, Daniel O’Neil, Xiongwu Kang, Jinsze Wong, Paul Szymanski,
Timothy C. Parker, Harry L. Anderson, Mostafa A. El-Sayed, and Seth R.
Marder. A Step Toward Efficient Panchromatic Multi-Chromophoric Sensitizers for Dye Sensitized Solar Cells. Chemistry of Materials, 27(18):6305–6313,
September 2015.
[143] CrysAlisP ro , 2018.
[144] SCALE3 ABSPACK, 2005.
[145] M. John Plater, Stuart Aiken, and Grant Bourhill. A new synthetic route to
donor-acceptor porphyrins. Tetrahedron, 58(12):2405–2413, March 2002.
[146] Volker Leen, Peijia Yuan, Lina Wang, Noël Boens, and Wim Dehaen. Synthesis of Meso-Halogenated BODIPYs and Access to Meso-Substituted Analogues. Organic Letters, 14(24):6150–6153, December 2012.
171
Texas Tech University, Miranda C. Andrews, December 2019
[147] Jinchun Qiu, Daniel K. Unruh, and Anthony F. Cozzolino. Design, Synthesis,
and Structural Characterization of a Bisantimony(III) Compound for Anion
Binding and the Density Functional Theory Evaluation of Halide Binding
through Antimony Secondary Bonding Interactions. The Journal of Physical
Chemistry A, 120(46):9257–9269, November 2016.
[148] J. Qiu, B. Song, X. Li, and A. F. Cozzolino. Solution and gas phase evidence of
anion binding through the secondary bonding interactions of a bidentate bisantimony(III) anion receptor. Physical Chemistry Chemical Physics, 20(1):46–
50, December 2017.
[149] Shiva Moaven, Jingze Yu, Jason Yasin, Daniel K. Unruh, and Anthony F.
Cozzolino. Precise Steric Control over 2D versus 3D Self-Assembly of Antimony(III) Alkoxide Cages through Strong Secondary Bonding Interactions.
Inorganic Chemistry, 56(14):8372–8380, July 2017.
[150] Shiva Moaven, Jingze Yu, Maythe Vega, Daniel K. Unruh, and Anthony F.
Cozzolino. Self-assembled reversed bilayers directed by pnictogen bonding
to form vesicles in solution. Chemical Communications, 54(64):8849–8852, August 2018.
[151] Shiva Moaven, Miranda C. Andrews, Thomas J. Polaske, Brian M. Karl,
Daniel K. Unruh, Eric Bosch, Nathan P. Bowling, and Anthony F. Cozzolino. Supramolecular Capsule Self-Assembly via Pnictogen Bonding. Journal
of the American Chemical Society, page In Peer Review, 2019.
[152] Thomas M. Beale, Michael G. Chudzinski, Mohammed G. Sarwar, and
Mark S. Taylor. Halogen bonding in solution: thermodynamics and applications. Chemical Society Reviews, 42(4):1667–1680, January 2013. 00145.
[153] Oliver Dumele, Dino Wu, Nils Trapp, Nancy Goroff, and François Diederich.
Halogen Bonding of (Iodoethynyl)benzene Derivatives in Solution. Organic
Letters, 16(18):4722–4725, September 2014.
[154] Peter Politzer, Jane S. Murray, and Timothy Clark. Halogen bonding: an
electrostatically-driven highly directional noncovalent interaction. Physical
Chemistry Chemical Physics, 12(28):7748–7757, July 2010.
[155] Peter Politzer, Jane S. Murray, and Timothy Clark. Halogen bonding and
other σ-hole interactions: a perspective. Physical Chemistry Chemical Physics,
15(27):11178–11189, June 2013.
172
Texas Tech University, Miranda C. Andrews, December 2019
[156] Anna M. Christianson and François P. Gabbaı̈. A Lewis Acidic, /piConjugated Stibaindole with a Colorimetric Response to Anion Binding at
Sb(III). Organometallics, 36(16):3013–3015, August 2017.
[157] Frank Neese. The ORCA program system. Wiley Interdisciplinary Reviews:
Computational Molecular Science, 2(1):73–78, 2012.
[158] A. D. Becke. Densityfunctional thermochemistry. III. The role of exact exchange. The Journal of Chemical Physics, 98(7):5648–5652, April 1993.
[159] A. D. Becke. Density-functional exchange-energy approximation with correct asymptotic behavior. Physical Review A, 38(6):3098–3100, September
1988.
[160] Ansgar Schafer, Hans Horn, and Reinhart Ahlrichs. Fully optimized contracted Gaussian basis sets for atoms Li to Kr. The Journal of Chemical Physics,
97(4):2571–2577, August 1992.
[161] P. J. Stephens, F. J. Devlin, C. F. Chabalowski, and M. J. Frisch. Ab Initio Calculation of Vibrational Absorption and Circular Dichroism Spectra
Using Density Functional Force Fields. The Journal of Physical Chemistry,
98(45):11623–11627, November 1994.
[162] Florian Weigend. Accurate Coulomb-fitting basis sets for H to Rn. Physical
Chemistry Chemical Physics, 8(9):1057–1065, February 2006.
[163] E. van Lenthe, J. Baerends Evert, and J. G. Snijders. Relativistic regular twocomponent Hamiltonians. Journal of Chemical Physics, 99:4597–4610, 1993.
[164] J.-L. Heully, I. Lindgren, E. Lindroth, S. Lundqvist, and A.-M. MartenssonPendrill. Diagonalisation of the Dirac Hamiltonian as a basis for a relativistic
many-body procedure. Journal of Physics B: Atomic and Molecular Physics,
19(18):2799, September 1986.
[165] Tian Lu and Feiwu Chen. Multiwfn: A multifunctional wavefunction analyzer. Journal of Computational Chemistry, 33(5):580–592, 2012.
[166] Tian Lu and Feiwu Chen. Quantitative analysis of molecular surface based
on improved Marching Tetrahedra algorithm. Journal of Molecular Graphics
and Modelling, 38:314–323, September 2012.
[167] Abdul-Rahman Allouche. Gabedit-A graphical user interface for computational chemistry softwares. Journal of Computational Chemistry, 32(1):174–182,
2011.
173